Aesthetics and the philosophy of art the analytic tradition: an anthology [2 ed] 9781119222446, 1119222443

Citation preview

Aesthetics and the Philosophy of Art

BLACKWELL PHILOSOPHY ANTHOLOGIES Each volume in this outstanding series provides an authoritative and comprehensive collection of the essential primary readings from philosophy’s main fields of study. Designed to complement the Blackwell Companions to Philosophy series, each volume represents an unparalleled resource in its own right, and will provide the ideal platform for course use. 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 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40

Cottingham: Western Philosophy: An Anthology (second edition) Cahoone: From Modernism to Postmodernism: An Anthology (expanded second edition) LaFollette: Ethics in Practice: An Anthology (third edition) Goodin and Pettit: Contemporary Political Philosophy: An Anthology (second edition) Eze: African Philosophy: An Anthology McNeill and Feldman: Continental Philosophy: An Anthology Kim and Sosa: Metaphysics: An Anthology Lycan and Prinz: Mind and Cognition: An Anthology (third edition) Kuhse and Singer: Bioethics: An Anthology (second edition) Cummins and Cummins: Minds, Brains, and Computers – The Foundations of Cognitive Science: An Anthology Sosa, Kim, Fantl, and McGrath Epistemology: An Anthology (second edition) Kearney and Rasmussen: Continental Aesthetics – Romanticism to Postmodernism: An Anthology Martinich and Sosa: Analytic Philosophy: An Anthology Jacquette: Philosophy of Logic: An Anthology Jacquette: Philosophy of Mathematics: An Anthology Harris, Pratt, and Waters: American Philosophies: An Anthology Emmanuel and Goold: Modern Philosophy – From Descartes to Nietzsche: An Anthology Scharff and Dusek: Philosophy of Technology – The Technological Condition: An Anthology Light and Rolston: Environmental Ethics: An Anthology Taliaferro and Griffiths: Philosophy of Religion: An Anthology Lamarque and Olsen: Aesthetics and the Philosophy of Art – The Analytic Tradition: An Anthology (second edition) John and Lopes: Philosophy of Literature – Contemporary and Classic Readings: An Anthology Cudd and Andreasen: Feminist Theory: A Philosophical Anthology Carroll and Choi: Philosophy of Film and Motion Pictures: An Anthology Lange: Philosophy of Science: An Anthology Shafer‐Landau and Cuneo: Foundations of Ethics: An Anthology Curren: Philosophy of Education: An Anthology Shafer‐Landau: Ethical Theory: An Anthology Cahn and Meskin: Aesthetics: A Comprehensive Anthology McGrew, Alspector‐Kelly and Allhoff: The Philosophy of Science: An Historical Anthology May: Philosophy of Law: Classic and Contemporary Readings Rosenberg and Arp: Philosophy of Biology: An Anthology Kim, Korman, and Sosa: Metaphysics: An Anthology (second edition) Martinich and Sosa: Analytic Philosophy: An Anthology (second edition) Shafer‐Landau: Ethical Theory: An Anthology (second edition) Hetherington: Metaphysics and Epistemology: A Guided Anthology Scharff and Dusek: Philosophy of Technology – The Technological Condition: An Anthology (second edition) LaFollette: Ethics in Practice: An Anthology (fourth edition) Davis: Contemporary Moral and Social Issues: An Introduction through Original Fiction, Discussion, and Readings Dancy and Sandis: Philosophy of Action: An Anthology

Aesthetics and the Philosophy of Art The Analytic Tradition An Anthology Second Edition

Edited by

Peter Lamarque and Stein Haugom Olsen

This edition first published 2019 © 2019 John Wiley & Sons, Inc. Blackwell Publishing Ltd. (1e, 2004) 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, electronic, mechanical, photocopying, recording or otherwise, except as permitted by law. Advice on how to obtain permission to reuse material from this title is available at http://www.wiley.com/go/permissions. The right of Peter Lamarque and Stein Haugom Olsen to be identified as the authors of the editorial material in this work has been asserted in ­accordance with law. Registered Office John Wiley & Sons, Inc., 111 River Street, Hoboken, NJ 07030, USA Editorial Offices John Wiley & Sons, Inc., 111 River Street, Hoboken, NJ 07030, USA 9600 Garsington Road, Oxford, OX4 2DQ, UK For details of our global editorial offices, customer services, and more information about Wiley products visit us at www.wiley.com. Wiley also publishes its books in a variety of electronic formats and by print‐on‐demand. Some content that appears in standard print versions of this book may not be available in other formats. Limit of Liability/Disclaimer of Warranty While the publisher and authors have used their best efforts in preparing this work, they make no representations or warranties with respect to the accuracy or completeness of the contents of this work and specifically disclaim all warranties, including without limitation any implied warranties of merchantability or fitness for a particular purpose. No warranty may be created or extended by sales representatives, written sales materials or p­ romotional statements for this work. The fact that an organization, website, or product is referred to in this work as a citation and/or potential source of further information does not mean that the publisher and authors endorse the information or services the organization, website, or product may provide or recommendations it may make. This work is sold with the understanding that the publisher is not engaged in rendering professional services. The advice and strategies contained herein may not be suitable for your situation.You should consult with a specialist where appropriate. Further, readers should be aware that websites listed in this work may have changed or disappeared between when this work was written and when it is read. Neither the publisher nor authors shall be liable for any loss of profit or any other commercial damages, including but not limited to special, incidental, consequential, or other damages. Library of Congress Cataloging‐in‐Publication data is available. Paperback ISBN: 9781119222446 Cover Design: Wiley Cover Image: Landscape with Apollo and the Muses, Claude Lorrain, National Galleries of Scotland. Purchased with the aid of the Art Fund and a Treasury Grant 1960. www.nationalgalleries.org Set in 9.5/11.5pt Bembo by SPi Global, Pondicherry, India Printed and bound in Singapore by C.O.S. Printers Pte Ltd 10 9 8 7 6 5 4 3 2 1

Contents

Acknowledgmentsx Extracts from the General Introduction to the First Edition (2004) xiii General Introduction to the Second Edition xvii

Part I  Identifying Art

Introduction3 1 The Artworld Arthur C. Danto

7

2 The New Institutional Theory of Art George Dickie

15

3 An Aesthetic Definition of Art Monroe C. Beardsley

22

4 “But They Don’t Have Our Concept of Art” Denis Dutton

30

5 Nobody Needs a Theory of Art Dominic McIver Lopes

43

6 Art: What it Is and Why it Matters Catharine Abell

54

Part II  Ontology of Art

1

67

Introduction69 7 What a Musical Work Is Jerrold Levinson

71

8 Defending Musical Platonism Julian Dodd

84

vi contents 9 Against Musical Ontology Aaron Ridley 10 The Ontology of Art and Knowledge in Aesthetics Amie L.Thomasson

Part III  Aesthetic Properties and Aesthetic Experience

98 108

117

Introduction119

11 Aesthetic Concepts Frank Sibley

121

12 Categories of Art Kendall L.Walton

134

13 In Defence of Moderate Aesthetic Formalism Nick Zangwill

149

14 How to Be a Pessimist about Aesthetic Testimony Robert Hopkins

159

15 Recent Approaches to Aesthetic Experience Noël Carroll

170

Part IV  Intention and Interpretation

183

Introduction185

16 Intentions and Interpretations: A Fallacy Revived Monroe C. Beardsley

187

17 The Literary Work as a Pliable Entity: Combining Realism and Pluralism Torsten Pettersson

197

18 Authors’ Intentions, Literary Interpretation, and Literary Value Stephen Davies

208

Part V  Values of Art

223

Introduction225

19 Originals, Copies, and Aesthetic Value Jack W. Meiland

229

20 Artistic Value Malcolm Budd

236

21 The Ethical Criticism of Art Berys Gaut

247

22 Artistic Value and Opportunistic Moralism Eileen John

258

23 What’s Wrong with the (Female) Nude? A Feminist Perspective on Art and Pornography A.W. Eaton

266

contents

Part VI  Art and Knowledge

vii

283

Introduction285

24 On the Cognitive Triviality of Art Jerome Stolnitz

289

25 Art and Moral Knowledge Cynthia A. Freeland

295

26 Reading Fiction and Conceptual Knowledge: Philosophical Thought in Literary Context Eileen John

310

27 Cognitive Values in the Arts: Marking the Boundaries Peter Lamarque

326

Part VII  Fictionality and Imagination

337

Introduction339

28 Fearing Fictions Kendall L.Walton

343

29 The Logical Status of Fictional Discourse John Searle

355

30 The Expression of Feeling in Imagination Richard Moran

363

31 The Puzzle of Imaginative Resistance Tamar Szabó Gendler

378

32 Anne Brontë and the Uses of Imagination Gregory Currie

393

33 Fiction as a Genre Stacie Friend

402

Part VIII  Pictorial Art

417

Introduction419

34 On Pictorial Representation Richard Wollheim

421

35 Pictorial Realism Catharine Abell

431

36 Telling Pictures: The Place of Narrative in Late Modern ‘Visual Art’ David Davies

441

Part IX  Photography and Film

451

Introduction453

37 Photography and Representation Roger Scruton

457

38 Photography and Causation: Responding to Scruton’s Scepticism Dawn M. Phillips

472

viii contents 39 Cinematic Art Berys Gaut

483

40 Theses on Cinema as Philosophy Paisley Livingston

496

41 Narration in Motion Katherine J.Thomson‐Jones

503

Part X  Literature

511

Introduction513

42 Style and Personality in the Literary Work Jenefer M. Robinson

517

43 Literary Aesthetics and Literary Practice Stein Haugom Olsen

527

44 Fictional Characters and Literary Practices Amie L.Thomasson

537

45 The Elusiveness of Poetic Meaning Peter Lamarque

549

Part XI  Music

561

Introduction563

46 The Profundity of Music Peter Kivy

567

47 Against Emotion: Hanslick Was Right about Music Nick Zangwill

574

48 Listening with Emotion: How Our Emotions Help Us to Understand Music Jenefer Robinson

583

Part XII  Popular Arts

601

Introduction603

49 Defining Mass Art Noël Carroll

607

50 Just a Song? Exploring the Aesthetics of Popular Song Performance Jeanette Bicknell

623

51 Comics as Literature? Aaron Meskin

632

52 The Vice of Snobbery: Aesthetic Knowledge, Justification and Virtue in Art Appreciation Matthew Kieran

647

contents

Part XIII  Aesthetics of Nature and Everyday Aesthetics

ix

659

Introduction661

53 Appreciation and the Natural Environment Allen Carlson

665

54 Scientific Knowledge and the Aesthetic Appreciation of Nature Patricia Matthews

673

55 Aesthetic Character and Aesthetic Integrity in Environmental Conservation Emily Brady

684

56 Everyday Aesthetics Yuriko Saito

695

57 The Pervasiveness of the Aesthetic in Ordinary Experience Sherri Irvin

700

Index710

Acknowledgments

1 Arthur C. Danto, “The Artworld,” Journal of Philosophy 61 (1964), pp. 571–84. Reproduced by permission of The Journal of Philosophy. 2 George Dickie, “The New Institutional Theory of Art,” Proceedings of the 8th Wittgenstein Symposium 10 (1983), pp. 57–64. Public domain. 3 Monroe C. Beardsley, “An Aesthetic Definition of Art,” in Hugh Curtler (ed.), What Is Art? (New York: Haven Publications, 1983), pp. 15–29. Reproduced by permission of the editor. 4 Denis Dutton, “But They Don’t Have Our Concept of Art,” in Noël Carroll (ed.), Theories of Art Today (Madison: University of Wisconsin Press, 2000), pp. 217–38. © University of Wisconsin Press. 5 Dominic McIver Lopes, “Nobody Needs a Theory of Art,” Journal of Philosophy 105 (2008), pp. 109–127. Reproduced by permission of The Journal of Philosophy. 6 Catharine Abell, “Art: What it Is and Why it Matters,” Philosophy and Phenomenological Research 85 (2012), pp. 671–91. Reproduced by permission of Wiley. 7 Jerrold Levinson, “What a Musical Work Is,” Journal of Philosophy 77 (1980), pp. 5–28. Reproduced by permission of The Journal of Philosophy. 8 Julian Dodd, “Defending Musical Platonism,” British Journal of Aesthetics 42 (2002), pp. 380–402. Reproduced by permission of Oxford University Press. 9 Aaron Ridley, “Against Musical Ontology,” Journal of Philosophy 100 (2003), pp. 203–20. Reproduced by permission of The Journal of Philosophy. 10 Amie L. Thomasson, “The Ontology of Art and Knowledge in Aesthetics,” Journal of Aesthetics and Art Criticism 63 (2005), pp. 221–9. Reproduced by permission of Wiley.

11 Frank Sibley, “Aesthetic Concepts,” in Joseph Margolis (ed.), Philosophy Looks at the Arts: Contemporary Readings in Aesthetics, rev. edn. (Philadelphia: Temple University Press, 1978), pp. 64–87; reproduced with “extensive minor revisions” from Philosophical Review 68 (1959), pp. 421–50. Public domain. 12 Kendall L.Walton, “Categories of Art,” Philosophical Review 79 (1970), pp. 334–67. Public domain. 13 Nick Zangwill, “In Defence of Moderate Aesthetic Formalism,” Philosophical Quarterly 50 (2000), pp. 476–93. Reproduced by permission of Oxford University Press. 14 Robert Hopkins, “How to Be a Pessimist about Aesthetic Testimony,” Journal of Philosophy 108 (2011), pp. 138–57. Reproduced by permission of The Journal of Philosophy. 15 Noël Carroll, “Recent Approaches to Aesthetic Experience,” Journal of Aesthetics and Art Criticism 70 (2012), pp. 165–77. Reproduced by permission of Wiley. 16 Monroe C. Beardsley, “Intentions and Interpretations: A Fallacy Revived,” in Michael J.Wreen and Donald M. Callan (eds.), The Aesthetic Point of View (Ithaca, NY: Cornell University Press, 1982), pp. 188–207. Reproduced by permission of Cornell University Press. 17 Torsten Pettersson, “The Literary Work as a Pliable Entity: Combining Realism and Pluralism,” in Michael Krausz (ed.), Is There a Single Right Interpretation? (University Park: Penn State University Press, 2002), pp.211–30. Reproduced by permission of Penn State University Press. 18 Stephen Davies, “Authors’ Intentions, Literary Interpretation, and Literary Value,” British Journal of

acknowledgments Aesthetics 46 (2006), pp. 223–47. Reproduced by ­permission of Oxford University Press. 19 Jack W. Meiland, “Originals, Copies, and Aesthetic Value,” in Denis Dutton (ed.), The Forger’s Art (Berkeley: University of California Press, 1983), pp. 115–30. © University of California Press. 20 Malcolm Budd, Ch 1, Values of Art (London: Penguin, 1995), pp. 1–16 (up to line 4), p. 38 (from line 12)–p. 43 (excluding last 4 lines), endnotes 1‐16 (pp. 173–6), 52–9 (pp. 182–4). Reproduced by permission of Penguin Books Ltd. 21 Berys Gaut, “The Ethical Criticism of Art,” in Jerrold Levinson (ed.), Aesthetics and Ethics (Cambridge: Cambridge University Press, 1998), pp. 182–203. Reproduced by permission of Cambridge University Press. 22 Eileen John, “Artistic Value and Opportunistic Moralism,” in Matthew Kieran (ed.), Contemporary Debates in Aesthetics and the Philosophy of Art (Oxford: Blackwell, 2006), pp. 331–41. Reproduced by permission of Blackwell Publishing Ltd. 23 A.W. Eaton, “What’s Wrong with the (Female) Nude? A Feminist Perspective on Art and Pornography,” in Hans Maes and Jerrold Levinson (eds.), Art and Pornography: Philosophical Essays (Oxford: Oxford University Press, 2012), pp. 278–307. Reproduced by permission of Oxford University Press. 24 Jerome Stolnitz, “On the Cognitive Triviality of Art,” British Journal of Aesthetics 32 (1992), pp. 191–200. Reproduced by permission of Oxford University Press. 25 Cynthia A. Freeland, “Art and Moral Knowledge,” Philosophical Topics 25 (1997), pp. 11–36. Reproduced by permission of University of Arkansas Press. 26 Eileen John, “Reading Fiction and Conceptual Knowledge: Philosophical Thought in Literary Context,” Journal of Aesthetics and Art Criticism 56 (1998), pp. 331–48. Reproduced by permission of Wiley. 27 Peter Lamarque,“CognitiveValues in the Arts: Marking the Boundaries,” in Matthew Kieran (ed.), Contemporary Debates in Aesthetics and the Philosophy of Art (Oxford: Blackwell, 2006), pp. 127–39. Reproduced by permission of Blackwell Publishing Ltd. 28 Kendall L. Walton, “Fearing Fictions,” Journal of Philosophy 75 (1978), pp. 5–27. Reproduced by permission of The Journal of Philosophy. 29 John Searle,“The Logical Status of Fictional Discourse,” in Expression and Meaning (Cambridge: Cambridge

xi

University Press, 1979), pp. 58–75. Reproduced by permission of Cambridge University Press. 30 Richard Moran, “The Expression of Feeling in Imagination,” Philosophical Review 103 (1994), pp. 75–106. Reproduced by permission of Duke University Press. 31 Tamar Szabó Gendler, “The Puzzle of Imaginative Resistance,” Journal of Philosophy 97 (2000), pp. 55–81. Reproduced by permission of The Journal of Philosophy. 32 Gregory Currie, “Anne Brontë and the Uses of Imagination,” in Matthew Kieran (ed.), Contemporary Debates in Aesthetics and the Philosophy of Art (Oxford: Blackwell, 2004). Reproduced by permission of Wiley. 33 Stacie Friend, “Fiction as a Genre,” Proceedings of the Aristotelian Society 112 (2012), pp. 179–209. Reproduced by permission of Oxford University Press. 34 Richard Wollheim, “On Pictorial Representation,” Journal of Aesthetics and Art Criticism 56 (1998), pp. 217–26. Reproduced by permission of Wiley. 35 Catharine Abell, “Pictorial Realism,” Australasian Journal of Philosophy 85 (2007), pp. 1–17. Reproduced by permission of Taylor & Francis. 36 David Davies, “Telling Pictures: The Place of Narrative in Late Modern ‘Visual Art,’” in Peter Goldie and Elisabeth Schellekens (eds.), Philosophy and Conceptual Art (Oxford: Oxford University Press, 2007), pp. 138–56. Reproduced by permission of Oxford University Press. 37 Roger Scruton, “Photography and Representation,” in The Aesthetic Understanding: Essays in the Philosophy of Art and Culture (London: Methuen, 1983), pp. 102–26. Reproduced by permission of the author. 38 Dawn M. Phillips, “Photography and Causation: Responding to Scruton’s Scepticism,” British Journal of  Aesthetics 49 (2009), pp. 327–340. Reproduced by ­permission of Oxford University Press. 39 Berys Gaut, “Cinematic Art,” Journal of Aesthetics and Art Criticism 60 (2002), pp. 299–312. Reproduced by permission of Wiley. 40 Paisley Livingston, “Theses on Cinema as Philosophy,” Journal of Aesthetics and Art Criticism 64 (2006), pp. 11–18. Reproduced by permission of Wiley. 41 Katherine J. Thomson‐Jones, “Narration in Motion,” British Journal of Aesthetics 52 (2012), pp. 33–43. Reproduced by permission of Oxford University Press. 42 Jenefer M. Robinson, “Style and Personality in the Literary Work,” Philosophical Review 94 (1985), pp.  227–47. Reproduced by permission of Duke University Press.

xii acknowledgments 43 Stein Haugom Olsen,“Literary Aesthetics and Literary Practice,” in The End of Literary Theory (Cambridge: Cambridge University Press, 1987), pp. 1–19. Reproduced by permission of Cambridge University Press. 44 Amie L. Thomasson, “Fictional Characters and Literary Practices,” British Journal of Aesthetics 43 (2003), pp. 138–57. Reproduced by permission of Oxford University Press. 45 Peter Lamarque, “The Elusiveness of Poetic Meaning,” Ratio 22 (2009), pp. 398–420. Reproduced by permission of Wiley. 46 Peter Kivy, “The Profundity of Music,” in Music Alone: Philosophical Reflections on the Purely Musical Experience (Ithaca, NY: Cornell University Press, 1990), pp. 245–55. Reproduced by permission of Cornell University Press. 47 Nick Zangwill, “Against Emotion: Hanslick Was Right about Music,” British Journal of Aesthetics 44 (2004), pp. 29–43. Reproduced by permission of Oxford University Press. 48 Jenefer Robinson, “Listening with Emotion: How Our Emotions Help Us to Understand Music,” in Deeper than Reason (Oxford: Oxford University Press, 2005), pp. 348–78. Reproduced by permission of Oxford University Press. 49 Noël Carroll, Section 3 of “The Nature of Mass Art,” from A Philosophy of Mass Art (Oxford: Clarendon Press, 1998), pp. 184–211. Reproduced by permission of Oxford University Press.

50 Jeanette Bicknell, “Just a Song? Exploring the Aesthetics of Popular Song Performance,” Journal of Aesthetics and Art Criticism 63 (2005), pp. 261–70. Reproduced by permission of Wiley. 51 Aaron Meskin, “Comics as Literature?” British Journal of Aesthetics 49 (2009), pp. 219–39. Reproduced by permission of Oxford University Press. 52 Matthew Kieran, “The Vice of Snobbery: Aesthetic Knowledge, Justification and Virtue in Art Appreciation,” Philosophical Quarterly 60 (2010), pp. 243–63. Reproduced by permission of Oxford University Press. 53 Allen Carlson, “Appreciation and the Natural Environment,” Journal of Aesthetics and Art Criticism 37 (1979), pp. 267–76. Reproduced by permission of Wiley. 54 Patricia Matthews, “Scientific Knowledge and the Aesthetic Appreciation of Nature,” Journal of Aesthetics and Art Criticism 60 (2002), pp. 37–48. Reproduced by permission of Wiley. 55 Emily Brady, “Aesthetic Character and Aesthetic Integrity in Environmental Conservation,” Environmental Ethics 24 (2002), pp. 75–91. Reproduced with extensive minor revisions, by permission of the author. 56 Yuriko Saito, “Everyday Aesthetics,” Philosophy and Literature 25 (2001), pp. 87–95. Reproduced by permission of The Johns Hopkins University Press. 57 Sherri Irvin, “The Pervasiveness of the Aesthetic in Ordinary Experience,” British Journal of Aesthetics 48 (2008), pp. 29–44. Reproduced by permission of Oxford University Press.

Extracts from the General Introduction to the First Edition (2004)

This anthology has a number of clearly statable aims: •• to present in a single volume some of the key texts from the analytic tradition in aesthetics and philosophy of art; •• to display the development of this tradition from its beginnings in the 1950s to the present day; •• to illustrate the broad range of topics and problems addressed by analytic aestheticians, from general issues of a theoretical nature to more specific issues relating to particular art forms; •• to provide a valuable reference resource for teaching and research purposes. In selecting articles for inclusion we have tried to strike a balance on many fronts: between “classic” contributions and more recent developments, between topics, between art forms, between the needs of undergraduate teaching and the needs of a scholarly archive, between the desire for comprehensive coverage and the constraints of manageability. We hope the volume will act as something of a showcase for the considerable achievements of analytic ­ ­aesthetics over the past fifty years. But above all, we have sought to put together a selection which will be of practical usefulness for those working in the field, at ­ all levels. Why “analytic”? This volume is a companion to Blackwell’s Continental Aesthetics: Romanticism to Post­ modernism: An Anthology, edited by Richard Kearney and David Rasmussen, and, we believe, nicely complements it, in ­showing the distinctive treatment of sometimes not dissimilar topics by those working in the Anglophone tradition and from the perspective of analytic philosophy. Together the two volumes give an excellent overview of the full range of  philosophical thinking about the arts in the twentieth

century. It has often been remarked how inappropriate are the designations “Continental” and “­analytic” in marking different approaches to philosophy. For one thing, the former is a geographical indicator, the latter a methodological one, so they are already incommensurate. But more strikingly, many leading figures in analytic philosophy – Gottlob Frege, Ludwig Wittgenstein, Friedrich Waismann, Moritz Schlick, and other members of the Vienna Circle  –  came from Continental Europe, and currently in Germany, France, Spain, Scandinavia, and Italy there is extensive interest in analytical philosophical methods. However, these two volumes on aesthetics do show a pronounced difference in methodology and it is worth reflecting on the characteristics distinctive of the analytic tradition. Clearly the idea of “analysis” is central to analytic ­philosophy. But the aims and methods of analysis differ markedly in the various incarnations of the analytic school. In the early years of the twentieth century, under the direct influence of Frege, Russell, and Wittgenstein (as author of Tractatus Logico‐Philosophicus), the logical analysis of propositions was paramount, with the aim of displaying their “logical form,” as distinct from their surface grammatical form. Russell’s Theory of Descriptions was held to be paradigmatic in this regard. Superficially his theory might seem like a mere paraphrasing of sentences containing definite descriptions into a logical notation; in fact it had profound repercussions for traditional problems in philosophy, notably the problem of nonexistence, the relations between meaning and truth, and the manner in which false propositions relate to the world. Analytic aestheticians were to draw heavily on Russell’s achievement in analyzing fictionality. The uncovering of logical forms developed into a more general program in philosophy: the use of logic to “regiment” language, in W.V.O. Quine’s terms, into a

xiv

extracts from the general introduction to the first edition

“canonical notation,” with the aim of eliminating vagaries in common usage and delivering a streamlined vehicle for science. An even grander ambition lay behind this species of analysis, encouraged by early ideas in Russell and Wittgenstein, namely that logical analysis could reveal the vacuity of much traditional philosophy. The highpoint of this ambition came with the Logical Positivists’ sweeping denunciation of metaphysics as meaningless. But analytic philosophy did not restrict itself to the logical analysis of propositions. Specific, problematic, concepts were also subject to analysis. Sometimes this took the form of seeking definitions for troublesome terms: “knowledge,” “freedom,” “truth,” “good,” “existence,” and  –  later on  – “art.” Arguably this was an extension of an approach ­originating with Socrates, but the emphasis on “necessary and sufficient conditions” for the true application of a concept was a peculiarly modern  –  and “analytic”  – ­ ­phenomenon. However, not all analytic philosophers took definition to be the aim of conceptual analysis. Some, the Ordinary Language Philosophers from Oxford in the 1940s and 1950s, preferred the analogy with geography, proposed by Gilbert Ryle, seeing their task as “mapping out” concepts or finding their “logical geography.” Ryle’s Concept of Mind (1949) was paradigmatic in this regard, owing much to the later work of Wittgenstein. By the late 1960s the optimistic thought that logical analysis or the study of ordinary usage could alone solve – or dissolve – the major issues in philosophy, sweeping away centuries of metaphysical confusion, was being questioned. The interest in language and logic became focused into a relatively new form of inquiry, also traceable to Frege, namely “philosophy of language,” which sought a clearer understanding of such concepts as meaning, truth, reference, and indeed language itself, but without any ­programmatic ambition toward solving all philosophical problems. By the 1970s few philosophers styled themselves as “linguistic philosophers” or “ordinary language philosophers,” yet significantly the term “analytic philosophy” grew in popularity. The Fregean tradition continued to inform philosophy of language but the original linguistic turn lost its “revolutionary” edge and settled down merely into a style of philosophizing. Analytic philosophy now became distinctive for its methodology and its theoretical presuppositions. Characteristic of the analytic methodology are: •• the prominent application of logic and conceptual analysis; •• the commitment to rational methods of argument;

•• the emphasis on objectivity and truth; •• the predilection for spare, literal prose, eschewing overly rhetorical or figurative language; •• the felt need to define terms and offer explicit formulation of theses; •• the quasi‐scientific dialectical method of hypothesis / counter‐example / modification; •• the tendency to tackle narrowly defined problems, often working within on‐going debates. Notable among presuppositions, although not universally held, are: •• the treatment of scientific discourse as paradigmatic; •• a tendency toward ontological “parsimony,” realism about science, and physicalism about mind; •• the belief that philosophical problems are in some sense timeless or universal, at least not merely constructs of history and culture. It is perhaps the latter presupposition that distinguishes the analytic tradition most obviously from the “Continental.” Analytic philosophers tend not to historicize their debates; there is little reference to the historical development of problems or the history of ideas and a widespread skepticism about the value of historically contextualized study of earlier philosophers. A consequence is that analytic philosophers have little interest in the social, political, or ideological underpinnings of their work and tend to treat the problems they address as timeless, ahistorical, and solvable, if at all, by appeal to logic rather than to observations about external cultural factors. Analytic philosophy came relatively late to aesthetics. It was not until the 1940s and 1950s that philosophers trained in analytic methods turned their attention to issues in aesthetics and these were mostly philosophers who had established their reputations in different areas of the subject. Typical in this regard was the highly influential anthology, Aesthetics and Language, edited by William Elton in 1954, which collected papers published in the preceding decade from prestigious journals like Mind, Proceedings of the Aristotelian Society, and The Philosophical Quarterly, with contributors of the caliber of Gilbert Ryle, Stuart Hampshire, O.K. Bouwsma, John Passmore, and Arnold Isenberg. The editor was frank about the missionary purpose of the collection: “to diagnose and clarify some aesthetic confusions, which it holds to be mainly linguistic in origin” and “to provide philosophers and their students with



extracts from the general introduction to the first edition

a number of pieces that may serve as models of analytical procedures in aesthetics.” It had many targets associated with less enlightened times: “obfuscatory jargon,” the “pitfalls of generality,” the “predisposition to essentialism,” “misleading analogies” (e.g., between the aesthetic and the moral), and “irrefutable and non‐empirical” theories. […] Analytic contributions to aesthetics […] soon took off and Frank Sibley saw no need in his classic paper “Aesthetic Concepts” from 1959 to keep disparaging earlier efforts. In fact in 1958 the analytic school of aesthetics came of age with the publication of Monroe C. Beardsley’s Aesthetics: Problems in the Philosophy of Criticism, which provided a sustained treatment of a wide range of problems illustrated by examples from an equally wide range of art forms. By the 1980s and 1990s the felt need to apologize for, or be defensive about, working in aesthetics had long subsided. Philosophers of the highest caliber  –  Nelson Goodman, Richard Wollheim, Arthur Danto, Kendall Walton, Martha Nussbaum, Roger Scruton – were not only writing in aesthetics but were introducing debates in the subject to philosophers from quite different areas. In fact to the extent that aesthetics has been integrated into the mainstream analytic tradition this is because of movement in two directions. The first is through the appearance of ostensibly aesthetic issues in debates on quite other kinds of topics, often by philosophers who have no deep concern with aesthetics for its own sake. Thus, for example, in recent times, John McDowell, Crispin Wright, and Philip Pettit, among others, have used aesthetic properties as a test case for realism; David Lewis has applied possible world semantics to fiction; Peter van Inwagen ­ and Nathan Salmon have written on fictional objects and ontology; David Wiggins has defended subjectivism in relation to aesthetic judgement. Many similar examples could be given. Discussions of realism and anti‐realism, supervenience, ontology, secondary qualities, and relativism will not infrequently allude to the aesthetic realm. But these as it were incidental incursions into aesthetics are not the only measure of the standing of aesthetics in the analytic community. Of more central concern, moving in the opposite direction, is the recharacterization of traditional questions within aesthetics in an idiom drawn from other branches of philosophy. Treating aesthetics as a special case for metaphysics, ontology, epistemology, theory of meaning, value theory, and social or political philosophy has served, p­ erhaps above all else, to entrench aesthetics – and aestheticians – in the analytic mainstream. Work of aestheticians has made an

xv

impact beyond aesthetics back to the very areas from where the original issues grew up. One thinks of Goodman on symbolism, Walton on make‐believe, Sibley on aesthetic concepts, Danto on indiscernibles, Levinson on ontology, Margolis on interpretation, Scruton on aesthetic culture, Currie on fiction. These are efforts which could never be deemed marginal in philosophy. We take up later, in the different sections of the anthology, the story of how analytic aesthetics developed in its own right. Let us end this introduction, though, with a few more observations of a general nature about its characteristic features. First of all, as just noted, analytic aesthetics has tended to give priority to topics arising from concerns elsewhere in philosophy. The emphasis on logic and philosophy of language, for example, led inevitably to an interest in questions about meaning and truth in aesthetics. One notable aspect of this is the attention given to fictionality. We saw how work by Frege and Russell raised problems about nonexistence and reference in the context of seeking logical forms. It did not escape the notice of aestheticians that this had a bearing on fictional narratives of all kinds in the arts. […] When speech act theory developed in the 1960s, initially through J.L. Austin’s work, later by John R. Searle and others, it too was soon applied to aesthetics. Searle’s speech act analysis of fiction had considerable impact. When Monroe Beardsley returned to the question of intention and literary meaning later in life he also appealed to speech acts. Indeed the debate about intention – to a large extent initiated by Beardsley in his original attack on the Intentional Fallacy – is typical of the analytic tradition, drawing both on theories of meaning and philosophy of mind. The influence of Wittgenstein, especially his views on language, can be felt throughout the development of analytic aesthetics. [… and] it appears [for example] in the thought that there is a distinctive “practice” associated with the arts and perhaps also in the appeal to “games” to illuminate our interactions with art.The recent revival of interest in metaphysics among analytic philosophers has led to a substantial amount of work on the ontology of art. Again the logical emphasis – exploring the categorizations of objects, properties, types, instances – is distinctive of the analytic approach and marks it off from ontological enquiries in the “Continental” tradition, notably that of Heidegger, even though the problems are ostensibly similar. Finally, among characteristic topics, is the discussion of aesthetic concepts, initiated by Sibley. … Given the nature of analytic philosophy it is not surprising to find the kinds of topics just mentioned – meaning,

xvi

extracts from the general introduction to the first edition

reference, ontology, concepts, definition, fictionality, representation – but a notable aspect of recent work in analytic aesthetics has been the attention given to particular art forms, painting, literature, music, film, photography. We have acknowledged this development with sections on the different arts and the anthology contains contributions on all the arts just listed. In becoming more specialized in this way analytic aesthetics might be seen as falling into line with other meta‐enquiries in philosophy. No competent researcher in philosophy of mind, for example, can now show the kind of ignorance of empirical psychology ­provocatively flaunted by Gilbert Ryle half a century ago. And philosophy of science is commonly divided into ­distinct specialities – philosophy of physics, philosophy of biology – just as ethics has a growing normative strand in medical or business ethics. It is a strength of current ­analytic aesthetics that it too focuses on the individual arts. For one thing it brings aesthetics closer to actual critical practice and encourages links with subject specialists, in musicology, film studies, literary theory, and art history. For another it puts salutary constraints on the grand designs of aesthetics, particularly attempts to develop overarching or all‐embracing theories of the arts. It is, however, worth signaling potential dangers too. One danger is that aesthetics becomes more culture bound. When philosophers talk about music or film or literature it is usually a pretty narrow band of works that are taken as paradigmatic – inevitably these are works that the writers know. More often than not they are canonical works in the Western tradition. Generalizations about these works and their properties might not always carry over to works from different cultural traditions. Nevertheless, even if it were the case that discussion of particular arts had a relatively narrow frame of reference there is benefit to be gained just to the extent that philosophers can shed light on an important canon of works. In fact more characteristic of the analytic approach is a balance between proposing genuinely universal claims about a particular art form and illustrating those claims by reference to individual works. We have selected papers on particular art forms – including painting, music, literature, photography, film, and other popular arts – which seem especially effective in this regard. […]

Analytic aesthetics has sometimes been identified with “philosophy of art” or even meta‐criticism (Beardsley’s Aesthetics made the identification explicit). But not all work by analytic aestheticians concerns the aesthetics of art. We have included a section on the aesthetics of nature. […] The idea that aesthetic descriptions apply to all kinds of objects, not only works of art, is as important as is the recognition of other cultural traditions when speaking of art. We mentioned at the beginning one feature characteristic of analytic philosophy, the tendency to tackle narrowly defined problems.The complaint of William Elton was that aesthetics, prior to the advent of the analytic school, was overly ambitious, inclined to generality and given to untestable theories. We will end with a word about the scope of these essays. Individually, they do tend to stick to the proposing and defense of specific, even limited, theses. Such is the manner of the analytic enterprise. Indeed it has long been thought a merit of this enterprise that it favors slow meticulous work  –  finding strong arguments to support precise, clearly defined theses – over generalizations weakly or imprecisely defended. This can give the impression of pedantry and lack of ambition. It is, however, a false impression. Certainly debates by analytic aestheticians seem to move slowly, but that is because attention to detail is highly valued. There is a sense of community among contributors to these debates however overtly critical analytic philosophers can seem of each others’ work. Progress comes through criticism, often in the form of unexpected counterexamples to general theses. […] The cumulative effect of such debates is a sense of concentrated effort on carefully circumscribed ground. When we stand back and survey all the micro‐debates two features stand out: the seriousness of purpose and the difficulty of the issues. However narrowly defined the topics of individual papers and however small the steps taken, there is no disputing the centrality and resonance of  the underlying questions: What is the nature of art? What is the place of art in human life? How do meaning and truth and representation arise in the arts? What is the scope of the aesthetic? These essays make a substantial impact on such questions.

General Introduction to the Second Edition

The overall rationale for this anthology has not changed from that of the first edition (2004) and accordingly we have included a substantial extract from the General Introduction to the First Edition which spells out our aims and gives the context within analytic philosophy. But times move on and it has become clear that a newer, revised version is needed. The new edition has grown in size, now including 57 articles, where before there were 46; and there are more individual sections, 13 instead of 11. There are new sections on Art and Knowledge and on Photography and Film. Some sections have been renamed; some items have been moved around. But the basic structure is the same. Also we have retained 19 of the original articles, comprising what we see as a core of well‐established pieces that provide paradigms of the analytic approach. The authors of these are just the names one might expect as the leading lights of analytic aesthetics from a time when it was establishing itself as a strong and autonomous branch of philosophy: figures such as Monroe Beardsley, Malcolm Budd, Allen Carlson, Arthur Danto, George Dickie, Peter Kivy, Jerrold Levinson, Jenefer Robinson, Roger Scruton, Frank Sibley, Kendall Walton, and Richard Wollheim. But while we are retaining 19 articles from the First Edition we have also introduced no fewer than 38 new ones. So the anthology will have a new, we hope fresh, feel to it. Although the majority of the new articles were published in the period after the publication of the first edition, this is not the case with all. Occasionally we have felt that articles published prior to the first edition – for example, those by Denis Dutton, Cynthia Freeland, Nick Zangwill, Eileen John, Richard Moran, Tamar Gendler, and Yuriko Saito – usefully lay the ground for topics that have grown in importance since 2004. It is wrong to fetishize dates.

For the most part we have deliberately not chosen articles that have appeared very recently (e.g., the past two or three years). It can take time for articles to bed down and prove their usefulness in ongoing debates. We have sought out work that has something lasting and significant to say, that adds to prominent debates, and that fits well with other pieces in the anthology. None of the papers selected has been abridged: all are in their original published form. Many, but not all, come from academic journals, but some are from collections of papers. One point of an anthology is to gather in one place items that might be scattered and perhaps not easily obtained. As we said in the Introduction to the First Edition we have had to strike a balance between “the needs of undergraduate teaching and the needs of a scholarly archive, between the desire for comprehensive coverage and the constraints of manageability.” As always, selection has been the most ­difficult challenge, not from a shortage of well‐qualified contenders but from a surfeit. We are sometimes asked: why give focus to the “analytic tradition”? Perhaps the term is indeed becoming less significant. In Anglophone aesthetics arguably this is just the mainstream way of proceeding and it is far from obvious that there is a unique style of philosophizing across all  the  articles included in the anthology. On the other hand the features we identified in the General Introduction to the First Edition do seem to be present in nearly all the articles, whatever their style of writing. For example, there is not much emphasis given to earlier historical figures by way of contextualizing a discussion: usually topics are introduced directly as problems to be addressed, even if they do draw on recent debates. (There are exceptions: Zangwill appeals to Hanslick, Moran to Hume.) Also there is a marked tendency in most of the articles to state a thesis

xviii

general introduction to the second edition

more or less explicitly and defend it against potential counterexamples, crucially the defence resting on argument rather than appeal to authority. And the idea of clarifying, even analysing, concepts is well in evidence. So even if it is correct to say that “analytic aesthetics” is a broad‐brush term covering different methods and different styles, it is by no means empty.Those of us for whom the analytic method in philosophy is important have little difficulty recognizing it in practice and above all noticing when it is absent. We think the articles presented here are good exemplars, albeit exhibiting a healthy diversity of writing styles, interests, and ideas advanced. So what is the state of the field? On the evidence here, it is very strong.There is a substantial body of serious and original work in aesthetics, recent and continuing to grow, right across the Anglophone world. What is encouraging is that it is dynamic not static. Subtle shifts are noticeable  –  indeed apparent from the two editions of this anthology.Think how reflections on art itself, what it is, how it might be defined, have changed, not least through the work of Dominic Lopes (Part I). Or consider the renewed interest in relations between fiction and imagination (Part VII); or the continuing work on art and ethics (Part V); or the new kinds of

interest in aesthetic testimony (Hopkins, Part III); or the ever‐growing aesthetics of film and photography (Part IX); or poetry (Part X); or comics (Part XII); or everyday experiences (Part XIII). Many of these topics simply wouldn’t have occurred to those pioneering analytic aestheticians in the 1950s and 1960s. We take it that counts as progress. Numerous people have offered advice on this second edition, not least through a survey conducted by Blackwell, and we are most grateful to them all, specifically to Greg Currie for his detailed comments. We haven’t quite been able to accommodate every suggestion but we have done our best. The Blackwell team, as always, has been patient and constructive throughout and we thank them for their help and forbearance, in particular Marissa Koors, Rachel Greenberg, Manish Luthra, Natasha Wu, and Janey Fisher, the latter our extremely efficient copy‐editor. Above all, I suppose it is our students who have offered the most helpful guidance; it was from them that we have learnt what worked well or less well in the first edition and what they thought might be improved. So thank you to them too. Peter Lamarque Stein Haugom Olsen

Further Reading Beardsley, Monroe C. (1981) Aesthetics: Problems in the Philosophy of Criticism (2nd edn.). Indianapolis: Hackett. Budd, Malcolm (1995) Values of Art: Pictures, Poetry and Music. Harmondsworth, UK: Penguin Books. Carroll, Noël (ed.) (2000) Theories of Art Today. Madison: University of Wisconsin Press. Danto, Arthur C. (1983) The Transfiguration of the Commonplace. Cambridge, MA: Harvard University Press. Eaton, Marcia Muelder (1988) Basic Issues in Aesthetics. Belmont, CA: Wadsworth. Elton,William (ed.) (1954) Aesthetics and Language. Oxford: Blackwell. Feagin, Susan L., and Maynard, Patrick (eds.) (1997) Aesthetics. Oxford: Oxford University Press. Freeland, Cynthia (2001) But Is It Art? Oxford: Oxford University Press. Gaut, Berys and Lopes, Dominic McIver (eds.) (2013) Routledge Companion to Aesthetics (3rd edn.). London: Routledge. Goodman, Nelson (1976) Languages of Art, 2nd edn. Indianapolis: Hackett. Kelly, Michael (ed.) (2014) Encyclopedia of Aesthetics (2nd edn.), 4 vols. Oxford: Oxford University Press.

Kieran, Matthew (ed.) (2006) Contemporary Debates in Aesthetics and the Philosophy of Art. Malden, MA: Blackwell. Korsmeyer, Carolyn (2004) Gender and Aesthetics: An Introduction. London: Routledge. Levinson, Jerrold (ed.) (2005) The Oxford Handbook of Aesthetics. Oxford: Oxford University Press. Margolis, Joseph (1980) Art and Philosophy: Conceptual Issues in Aesthetics. Atlantic Highlands, NJ: Humanities Press. Neill, Alex and Ridley, Aaron (eds.) (2008) Arguing About Art: Contemporary Philosophical Debates (3rd edn.). New York: McGraw‐Hill. Sheppard, Anne (1987) Aesthetics: An Introduction to the Philosophy of Art. Oxford: Oxford University Press. Shusterman, Richard (ed.) (1989) Analytic Aesthetics. Oxford: Blackwell. Walton, Kendall L. (1990) Mimesis as Make‐Believe: On the Foundations of the Representational Arts. Cambridge, MA: Harvard University Press. Wollheim, Richard (1980) Art and Its Objects (2nd edn.). Cambridge: Cambridge University Press.

Part I Identifying Art

Introduction

It might reasonably be thought that the starting point for any serious philosophy of art must be the question “What is art?” After all, so the argument goes, without knowing what art is we could not know what enquiry we are pursuing. Furthermore, on this line of thought, nothing short of a list of essential or defining properties of art  –  those ­properties both necessary and sufficient for something to be a work of art – will satisfy the need to identify art apart from other human endeavors. The proliferation of avant‐garde art movements in the twentieth century, from Dadaism to Performance Art, Abstract Expressionism to Conceptual Art, made it all the more difficult to isolate any such defining essence of art. The often‐heard skeptical question “But is it art?” increased the pressure on art theorists and philosophers of art to tackle the definitional issue. But when analytic philosophers entered the fray in the 1950s they introduced a distinctive twist to the debate, not only seeking to formulate an adequate ­definition but, more characteristically, asking what might count as an adequate definition, if anything at all. Some, like Morris Weitz (1956), following Wittgenstein, argued that no definition of art, conceived as a set of necessary and sufficient conditions, was possible, given that the concept of art is an “open concept” and that any definition would compromise the creativity of art. In the end few came to agree with Weitz in rejecting the definitional ­project but it was acknowledged that the key to the debate would lie in what kinds of properties might be sought as the “essence” of art. Out of this debate there arose two

different conceptions of what kind of definition might count as adequate. These have been labeled (by Stephen Davies, 1991) as “procedural” and “functional” definitions. These approaches are discussed in detail in the papers by Dominic McIver Lopes (Chapter 5) and Catharine Abell (Chapter 6). Broadly a procedural definition, which can take different forms, is one that identifies as an essential feature of art certain facts about how works of art come to be accepted as such in the first place, notably the procedures (practices, institutions) that make it possible for something to be art. In contrast functional definitions, again of different kinds, focus on the functions that works of art serve, such as, for example, inviting aesthetic interest or expressing emotions. The two approaches are well exemplified, indeed paradigmatically so, in the papers below by Arthur Danto, George Dickie, and Monroe Beardsley, the former two offering ­procedural accounts, the latter a functionalist account. Danto (Chapter 1) takes the first crucial step – remarkable in the history of aesthetics – by directing our attention to features of artworks that are not perceptible. In his often quoted words (in the essay here): “To see something as art requires something the eye cannot descry – an atmosphere of art theory, a knowledge of the history of art: an artworld”. To help us grasp this revolutionary insight he uses thought‐experiments about what he calls “indiscernibles.” In his later work The Transfiguration of the Commonplace (1981) the idea is finely honed but it figures strikingly also in “The Artworld,” where he uses the example of Andy Warhol’s Brillo Boxes. If Warhol’s

Aesthetics and the Philosophy of Art – The Analytic Tradition: An Anthology, Second Edition. Edited by Peter Lamarque and Stein Haugom Olsen. © 2019 John Wiley & Sons, Inc. Published 2019 by John Wiley & Sons, Inc.

4 introduction Brillo boxes are indistinguishable in appearance from ­ordinary Brillo boxes, how could the former be art but not the latter? The stark answer Danto offers (an answer that gets more refined in later work) is that what makes all the difference is “a certain theory of art.” The artworld at particular stages of its development makes possible certain kinds of works. Only against the background of Pop Art – with its theoretical presuppositions – do works like Brillo Boxes, or in Danto’s other example, Rauschenberg’s bed (which from a certain perspective looks like a quite ordinary bed), become possible. For Danto, then, there is something essential to all art – contra Weitz – namely embeddedness in an artworld along with its supporting theories. Dickie (Chapter 2), in his “institutional theory” of art, developed and formalized certain aspects of Danto’s account. Going further than Danto, Dickie proposed a definition of art, employing the notion of an “artworld”. It is striking in its simplicity: “A work of art is an artifact of a kind created to be presented to an artworld public.” The crucial feature is that what makes something an artwork is not what it looks (sounds, etc.) like or what materials it is made of, or any of its intrinsic qualities, but how it relates to a loosely characterized institution (the “artworld”) or set of practices. Without the institution of art there would be no artworks even if there were things that looked exactly like what we call artworks. Dickie is content to acknowledge a circularity in his definition (“art” and ­“artworld” are inter‐definable) but insists that it is non‐ vicious circularity. Beardsley’s starting point (Chapter  3) is different: that the artistic is essentially connected, not to an institution, but to the aesthetic. He boldly presents a definition of art in these terms: “An artwork is something produced with the intention of giving it the capacity to satisfy the aesthetic interest.” As an analytic philosopher he carefully analyzes  –  and thus explains and justifies  –  each of the main terms in the definition. The definition is important, not least for presenting a clear contrast with institutional theories. The emphasis is on the function or aim of art rather than on institutional endorsement. Beardsley is much less tolerant of avant‐garde art – such as Duchamp’s “readymades” – than either Danto or Dickie but his insistence on “aesthetic interest” does, controversially, make problematic those species of “conceptual art” which overtly repudiate the aim Beardsley identifies. It should probably be conceded, though, that Beardsley captures an intuition about the nature of art which still has widespread support, if not among philosophers certainly among the general public.

The papers by Danto, Dickie and Beardsley provide the classic background to the definitional debate about art in analytic aesthetics, not least in exemplifying the procedural and functional approaches. The remaining papers in this section show the different kinds of direction that analytic discussion has taken in more recent years. Denis Dutton, in “But They Don’t Have Our Concept of Art” (Chapter 4), addresses the question whether “art” is a genuinely cross‐ cultural, universal concept (i.e., applicable across cultures and across time). He argues that it is and challenges the views of certain anthropologists that “art” is essentially a Western, and modern, concept. He identifies a set of characteristics that he believes are central to art even if they do not strictly count as necessary and sufficient conditions, arguing that any human practice which had none of the features enumerated would not be art, and that any human practice which possessed most of them would be art … in the sense that ­characterizes it through the whole of human history.

Dutton’s conception of art happily embraces a wide variety of practices, objects, and human activities, including those of “small‐scale, nonliterate tribal societies” that are a world away from Western twentieth‐century avant‐garde works, which Dutton believes have distorted how philosophers think about art. Dominic McIver Lopes, in his essay (Chapter 5), returns directly to the question of definition and has much to say about how the problematic avant‐garde cases are to be ­handled – the sorts of cases that helped to motivate institutional theories. Lopes’s view, as his provocative title “Nobody Needs a Theory of Art” suggests, is radical. Stop looking for a theory of art in general and seek only “theories of the arts.” What are the arts? Well, things like music, dance, theatre, literature, film, painting, architecture, and so on. He calls this “buck passing” (from “art” to “the arts”) and it is encapsulated, paradoxically perhaps, in a definition of “art” as: “item x is a work of art if and only if x is a work in activity P and P is one of the arts.” To settle the art status of the avant‐garde cases we need to appeal not to a theory of art but to theories of art forms. That, he thinks, is where the problems reside but also where they can be solved: after all, where the puzzling cases do not fit established art forms they often serve to create new art forms. Catharine Abell, in “Art:What it Is and Why it Matters” (Chapter  6), seeks to draw on both institutional and functional theories, and brings the idea of the values of art to the fore. She grounds her account in John Searle’s

introduction 5 conception of institutions and institutional facts. Her definition of art appeals not only to art institutions but to the functions that art institutions perform (or are perceived to perform by participants in the institutions). Hence: “Something is an artwork … [if and only if] it is the product of an art institution, and it directly affects how effectively that institution performs the perceived

functions to which its existence is due.” The artistic value of artworks is then explained in terms of their “tendency to improve how well the institutions of which they are products perform the functions that make them art institutions.” As always with analytic philosophy, the convincingness of the definitions lies in the detail with which they are explained and supported.

References Davies, Stephen (1991) Definitions of Art. Ithaca, NY: Cornell University Press.

Weitz, Morris (1956) “The Role of Theory in Aesthetics,” Journal of Aesthetics and Art Criticism 15: 27–35.

Further Reading Carroll, Noël (ed.) (2000) Theories of Art Today. Madison: University of Wisconsin Press. Currie, Gregory (2010) “Actual Art, Possible Art, and Art’s Definition,” Journal of Aesthetics and Art Criticism 68: 235–41. Danto, Arthur (1981) The Transfiguration of the Commonplace. Cambridge, MA: Harvard University Press. Dickie, George (1984) The Art Circle. New York: Haven. Dean, Jeffrey T. (2003) “The Nature of Concepts and the Definition of Art,” Journal of Aesthetics and Art Criticism 61: 29–35. Dutton, Denis (2006) “A Naturalist Definition of Art,” Journal of Aesthetics and Art Criticism 64: 367–77. Gaut, Berys (2000) “‘Art’ as a Cluster Concept,” in Noël Carroll, ed., Theories of Art Today, Madison: University of Wisconsin Press.

Iseminger, Gary (2004) The Aesthetic Function of Art, Ithaca, NY: Cornell University Press. Levinson, Jerrold (1990) “Defining Art Historically,” in Music, Art and Metaphysics. Ithaca, NY: Cornell University Press. Margolis, Joseph (1999) What, After All, is a Work of Art? University Park: Pennsylvania State University Press. Shiner, Larry (2001) The Invention of Art: A Cultural History, Chicago: Chicago University Press. Shiner, Larry (2007) “Western and Non‐Western Concepts of Art” [a reply to Dutton], in A. Neill & A. Ridley, eds. Arguing about Art (3rd edn.). London: Routledge. Stecker, Robert (1997) Artworks: Definition, Meaning, Value. University Park: Pennsylvania State University Press. Zangwill, Nick (2007) Aesthetic Creation, Oxford: Oxford University Press.

1

The Artworld Arthur C. Danto

Hamlet:  Do you see nothing there ? The Queen:  Nothing at all; yet all that is I see. Shakespeare: Hamlet, Act III, Scene IV

Hamlet and Socrates, though in praise and deprecation respectively, spoke of art as a mirror held up to nature. As with many disagreements in attitude, this one has a factual basis. Socrates saw mirrors as but reflecting what we can already see; so art, insofar as mirrorlike, yields idle accurate duplications of the appearances of things, and is of no cognitive benefit whatever. Hamlet, more acutely, recognized a remarkable feature of reflecting surfaces, namely that they show us what we could not otherwise perceive – our own face and form – and so art, insofar as it is mirrorlike, reveals us to ourselves, and is, even by socratic criteria, of some cognitive utility after all. As a philosopher, however, I find Socrates’ discussion defective on other, perhaps less profound grounds than these. If a mirror‐image of o is indeed an imitation of o, then, if art is imitation, mirror‐images are art. But in fact mirroring objects no more is art than returning weapons to a madman is justice; and reference to mirrorings would be just the sly sort of counterinstance we would expect Socrates to bring forward in rebuttal of the theory he instead uses them to illustrate. If that theory requires us to class these as art, it thereby shows its inadequacy: “is an imitation” will not do as a sufficient condition

for “is art.” Yet, perhaps because artists were engaged in imitation, in Socrates’ time and after, the insufficiency of the theory was not noticed until the invention of photography. Once rejected as a sufficient condition, mimesis was quickly discarded as even a necessary one; and since the achievement of Kandinsky, mimetic features have been relegated to the periphery of critical concern, so much so that some works survive in spite of possessing those virtues, excellence in which was once celebrated as the essence of art, narrowly escaping demotion to mere illustrations. It is, of course, indispensable in socratic discussion that all participants be masters of the concept up for analysis, since the aim is to match a real defining expression to a term in active use, and the test for adequacy presumably consists in showing that the former analyzes and applies to all and only those things of which the latter is true. The popular disclaimer notwithstanding, then, Socrates’ auditors purportedly knew what art was as well as what they liked; and a theory of art, regarded here as a real definition of ‘Art’, is accordingly not to be of great use in helping men to recognize instances of its application. Their antecedent ability to do this is precisely what the adequacy of the theory is to be tested against, the problem being only to make explicit what they already know. It is our use of the term that the theory allegedly means to capture, but we are supposed able, in the words of a recent writer, “to separate those objects which

Arthur C. Danto, “The Artworld,” Journal of Philosophy 61 (1964), pp. 571–84. Reproduced by permission of The Journal of Philosophy.

Aesthetics and the Philosophy of Art – The Analytic Tradition: An Anthology, Second Edition. Edited by Peter Lamarque and Stein Haugom Olsen. © 2019 John Wiley & Sons, Inc. Published 2019 by John Wiley & Sons, Inc.

8

arthur c. danto

are works of art from those which are not, because … we know how correctly to use the word ‘art’ and to apply the phrase ‘work of art’.” Theories, on this account, are somewhat like mirror‐images on Socrates’ account, showing forth what we already know, wordy reflections of the actual linguistic practice we are masters in. But telling artworks from other things is not so simple a matter, even for native speakers, and these days one might not be aware he was on artistic terrain without an artistic theory to tell him so. And part of the reason for this lies in the fact that terrain is constituted artistic in virtue of artistic theories, so that one use of theories, in addition to helping us discriminate art from the rest, consists in making art possible. Glaucon and the others could hardly have known what was art and what not: otherwise they would never have been taken in by mirror‐images.

I Suppose one thinks of the discovery of a whole new class of artworks as something analogous to the discovery of a whole new class of facts anywhere, viz., as something for theoreticians to explain. In science, as elsewhere, we often accommodate new facts to old theories via auxiliary hypotheses, a pardonable enough conservatism when the theory in question is deemed too valuable to be jettisoned all at once. Now the Imitation Theory of Art (IT) is, if one but thinks it through, an exceedingly powerful theory, explaining a great many phenomena connected with the causation and evaluation of artworks, bringing a surprising unity into a complex domain. Moreover, it is a simple matter to shore it up against many purported counterinstances by such auxiliary hypotheses as that the artist who deviates from mimeticity is perverse, inept, or mad. Ineptitude, chicanery, or folly are, in fact, testable predications. Suppose, then, tests reveal that these hypotheses fail to hold, that the theory, now beyond repair, must be replaced. And a new theory is worked out, capturing what it can of the old theory’s competence, together with the heretofore recalcitrant facts. One might, thinking along these lines, represent certain episodes in the history of art as not dissimilar to certain episodes in the history of science, where a conceptual revolution is being effected and where refusal to countenance certain facts, while in part due to prejudice, inertia, and self‐interest, is due also to the fact that a well‐established, or at least widely credited theory is being threatened in such a way that all coherence goes. Some such episode transpired with the advent of post‐ impressionist paintings. In terms of the prevailing artistic

theory (IT), it was impossible to accept these as art unless inept art: otherwise they could be discounted as hoaxes, self‐advertisements, or the visual counterparts of madmen’s ravings. So to get them accepted as art, on a footing with the Transfiguration (not to speak of a Landseer stag), required not so much a revolution in taste as a theoretical revision of rather considerable proportions, involving not only the artistic enfranchisement of these objects, but an emphasis upon newly significant features of accepted artworks, so that quite different accounts of their status as artworks would now have to be given. As a result of the new theory’s acceptance, not only were post‐impressionist paintings taken up as art, but numbers of objects (masks, weapons, etc.) were transferred from anthropological museums (and heterogeneous other places) to musées des beaux arts, though, as we would expect from the fact that a criterion for the acceptance of a new theory is that it account for whatever the older one did, nothing had to be transferred out of the musées des beaux arts – even if there were internal rearrangements as between storage rooms and exhibition space. Countless native speakers hung upon suburban mantelpieces innumerable replicas of paradigm cases for teaching the expression ‘work of art’ that would have sent their Edwardian forebears into linguistic apoplexy. To be sure, I distort by speaking of a theory: historically, there were several, all, interestingly enough, more or less defined in terms of the IT. Art‐historical complexities must yield before the exigencies of logical exposition, and I shall speak as though there were one replacing theory, partially compensating for historical falsity by choosing one which was actually enunciated. According to it, the artists in question were to be understood not as unsuccessfully imitating real forms but as successfully creating new ones, quite as real as the forms which the older art had been thought, in its best examples, to be creditably imitating. Art, after all, had long since been thought of as creative (Vasari says that God was the first artist), and the post‐impressionists were to be explained as genuinely creative, aiming, in Roger Fry’s words, “not at illusion but reality.” This theory (RT) furnished a whole new mode of looking at painting, old and new. Indeed, one might almost interpret the crude drawing in Van Gogh and Cézanne, the dislocation of form from contour in Rouault and Dufy, the arbitrary use of color planes in Gauguin and the Fauves, as so many ways of drawing attention to the fact that these were non‐imitations, specifically intended not to deceive. Logically, this would be roughly like printing “Not Legal Tender” across a brilliantly counterfeited dollar bill, the resulting object (counterfeit cum inscription) rendered incapable of deceiving



the artworld

anyone. It is not an illusory dollar bill, but then, just because it is non‐illusory it does not automatically become a real dollar bill either. It rather occupies a freshly opened area between real objects and real facsimiles of real objects: it is a non‐facsimile, if one requires a word, and a new contribution to the world. Thus, Van Gogh’s Potato Eaters, as a  consequence of certain unmistakable distortions, turns out to be a non‐facsimile of real‐life potato eaters; and ­inasmuch as these are not facsimiles of potato eaters, Van Gogh’s picture, as a non‐imitation, had as much right to be called a real object as did its putative subjects. By means of this theory (RT), artworks re‐entered the thick of things from which socratic theory (IT) had sought to evict them: if no  more real than what carpenters wrought, they were at least no less real. The Post‐Impressionist won a victory in ontology. It is in terms of RT that we must understand the artworks around us today. Thus Roy Lichtenstein paints comic‐strip panels, though ten or twelve feet high. These are reasonably faithful projections onto a gigantesque scale of the homely frames from the daily tabloid, but it is precisely the scale that counts. A skilled engraver might incise The Virgin and the Chancellor Rollin on a pinhead, and it would be recognizable as such to the keen of sight, but an engraving of a Barnett Newman on a similar scale would be a blob, disappearing in the reduction. A photograph of a Lichtenstein is indiscernible from a photograph of a counterpart panel from Steve Canyon; but the photograph fails to capture the scale, and hence is as inaccurate a reproduction as a black‐and‐white engraving of Botticelli, scale being essential here as color there. Lichtensteins, then, are not imitations but new entities, as giant whelks would be. Jasper Johns, by contrast, paints objects with respect to which questions of scale are irrelevant. Yet his objects cannot be imitations, for they have the remarkable property that any intended copy of a member of this class of objects is ­automatically a member of the class itself, so that these objects are logically inimitable. Thus, a copy of a numeral just is that numeral: a painting of 3 is a 3 made of paint. Johns, in  addition, paints targets, flags, and maps. Finally, in what I hope are not unwitting footnotes to Plato, two of  our pioneers  –  Robert Rauschenberg and Claes Oldenburg – have made genuine beds. Rauschenberg’s bed hangs on a wall, and is streaked with some desultory housepaint. Oldenburg’s bed is a rhomboid, narrower at one end than the other, with what one might speak of as a built‐in perspective: ideal for small bedrooms. As beds, these sell at singularly inflated prices, but one could sleep in either of them: Rauschenberg has expressed the

9

fear that someone might just climb into his bed and fall asleep. Imagine, now, a certain Testadura – a plain speaker and noted philistine – who is not aware that these are art, and who takes them to be reality simple and pure. He attributes the paintstreaks on Rauschenberg’s bed to the slovenliness of the owner, and the bias in the Oldenburg to the ineptitude of the builder or the whimsy, perhaps, of whoever had it “custom‐made.” These would be mistakes, but of rather an odd kind, and not terribly different from that made by the stunned birds who pecked the sham grapes of Zeuxis. They mistook art for reality, and so has Testadura. But it was meant to be reality, according to RT. Can one have mistaken reality for reality? How shall we describe Testadura’s error? What, after all, prevents Oldenburg’s creation from being a misshapen bed? This is equivalent to asking what makes it art, and with this query we enter a domain of conceptual inquiry where native speakers are poor guides: they are lost themselves.

II To mistake an artwork for a real object is no great feat when an artwork is the real object one mistakes it for. The problem is how to avoid such errors, or to remove them once they are made. The artwork is a bed, and not a bed‐illusion; so there is nothing like the traumatic encounter against a flat surface that brought it home to the birds of Zeuxis that they had been duped. Except for the guard cautioning Testadura not to sleep on the artworks, he might never have discovered that this was an artwork and not a bed; and since, after all, one cannot discover that a bed is not a bed, how is Testadura to realize that he has made an error? A certain sort of explanation is required, for the error here is a curiously philosophical one, rather like, if we may assume as correct some well‐known views of P. F. Strawson, mistaking a person for a material body when the truth is that a person is a material body in the sense that a whole class of predicates, sensibly applicable to material bodies, are sensibly, and by appeal to no different criteria, applicable to persons. So you cannot discover that a person is not a material body. We begin by explaining, perhaps, that the paintstreaks are not to be explained away, that they are part of the object, so the object is not a mere bed with – as it happens – streaks of paint spilled over it, but a complex object fabricated out of a bed and some paintstreaks: a paint‐bed. Similarly, a person is not a material body with  –  as it happens  –  some thoughts superadded, but is a complex entity made up of a body and some conscious states: a conscious‐body. Persons,

10

arthur c. danto

like artworks, must then be taken as irreducible to parts of themselves, and are in that sense primitive. Or, more accurately, the paintstreaks are not part of the real object – the bed – which happens to be part of the artwork, but are, like the bed, part of the artwork as such. And this might be generalized into a rough characterization of artworks that happen to contain real objects as parts of themselves: not every part of an artwork A is part of a real object R when R is part of A and can, moreover, be detached from A and seen merely as R. The mistake thus far will have been to mistake A for part of itself, namely R, even though it would not be incorrect to say that A is R, that the artwork is a bed. It is the ‘is’ which requires clarification here. There is an is that figures prominently in statements concerning artworks which is not the is of either identity or predication; nor is it the is of existence, of identification, or some special is made up to serve a philosophic end. Nevertheless, it is in common usage, and is readily mastered by children. It is the sense of is in accordance with which a child, shown a circle and a triangle and asked which is him and which his sister, will point to the triangle saying “That is me”; or, in response to my question, the person next to me points to the man in purple and says “That one is Lear”; or in the gallery I point, for my companion’s benefit, to a spot in the painting before us and say “That white dab is Icarus.”We do not mean, in these instances, that whatever is pointed to stands for, or represents, what it is said to be, for the word ‘Icarus’ stands for or represents Icarus: yet I would not in the same sense of is point to the word and say “ That is Icarus.” The sentence “That a is b” is perfectly compatible with “That a is not b” when the first employs this sense of is and the second employs some other, though a and b are used nonambiguously throughout. Often, indeed, the truth of the first requires the truth of the second. The first, in fact, is incompatible with “That a is not b” only when the is is used nonambiguously throughout. For want of a word I shall designate this the is of artistic identification; in each case in which it is used, the a stands for some specific physical property of, or physical part of, an object; finally, it is a necessary condition for something to be an artwork that some part or property of it be designable by the subject of a sentence that employs this special is. It is an is, incidentally, which has near‐relatives in marginal and mythical pronouncements. (Thus, one is Quetzalcoatl; those are the Pillars of Hercules.) Let me illustrate. Two painters are asked to decorate the east and west walls of a science library with frescoes to be respectively called Newton’s First Law and Newton’s Third Law. These paintings, when finally unveiled, look, scale apart, as follows:

A

B

As objects I shall suppose the works to be indiscernible: a black, horizontal line on a white ground, equally large in each dimension and element. B explains his work as follows: a mass, pressing downward, is met by a mass pressing upward: the lower mass reacts equally and oppositely to the upper one. A explains his work as follows: the line through the space is the path of an isolated particle. The path goes from edge to edge, to give the sense of its going beyond. If it ended or began within the space, the line would be curved: and it is parallel to the top and bottom edges, for if it were closer to one than to another, there would have to be a force accounting for it, and this is inconsistent with its being the path of an isolated particle. Much follows from these artistic identifications. To regard the middle line as an edge (mass meeting mass) imposes the need to identify the top and bottom half of the picture as rectangles, and as two distinct parts (not necessarily as two masses, for the line could be the edge of one mass jutting up – or down – into empty space). If it is an edge, we cannot thus take the entire area of the painting as a single space: it is rather composed of two forms, or one form and a non‐form. We could take the entire area as a single space only by taking the middle horizontal as a line which is not an edge. But this almost requires a three‐ dimensional identification of the whole picture: the area can be a flat surface which the line is above (Jet‐flight), or below (Submarine‐path), or on (Line), or in (Fissure), or through (Newton’s First Law) – though in this last case the area is not a flat surface but a transparent cross section of absolute space. We could make all these prepositional qualifications clear by imagining perpendicular cross sections to the picture plane. Then, depending upon the applicable prepositional clause, the area is (artistically) interrupted or not by the horizontal element. If we take the line as through space, the edges of the picture are not really the edges of the space: the space goes beyond the picture if the line itself



the artworld

does; and we are in the same space as the line is. As B, the edges of the picture can be part of the picture in case the masses go right to the edges, so that the edges of the picture are their edges. In that case, the vertices of the picture would be the vertices of the masses, except that the masses have four vertices more than the picture itself does: here four vertices would be part of the art work which were not part of the real object. Again, the faces of the masses could be the face of the picture, and in looking at the picture, we are looking at these faces: but space has no face, and on the reading of A the work has to be read as faceless, and the face of the physical object would not be part of the artwork. Notice here how one artistic identification engenders another artistic identification, and how, consistently with a given identification, we are required to give others and precluded from still others: indeed, a given identification determines how many elements the work is to contain. These different identifications are incompatible with one another, or generally so, and each might be said to make a different artwork, even though each artwork contains the identical real object as part of itself – or at least parts of the identical real object as parts of itself. There are, of course, senseless identifications: no one could, I think, sensibly read the middle horizontal as Love’s Labour’s Lost or The Ascendency of St. Erasmus. Finally, notice how acceptance of one identification rather than another is in effect to exchange one world for another. We could, indeed, enter a quiet poetic world by identifying the upper area with a clear and cloudless sky, reflected in the still surface of the water below, whiteness kept from whiteness only by the unreal boundary of the horizon. And now Testadura, having hovered in the wings throughout this discussion, protests that all he sees is paint: a white painted oblong with a black line painted across it. And how right he really is: that is all he sees or that anybody can, we aesthetes included. So, if he asks us to show him what there is further to see, to demonstrate through pointing that this is an artwork (Sea and Sky), we cannot comply, for he has overlooked nothing (and it would be absurd to suppose he had, that there was something tiny we could point to and he, peering closely, say “So it is! A work of art after all!”).We cannot help him until he has mastered the is of artistic identification and so constitutes it a work of art. If he cannot achieve this, he will never look upon artworks: he will be like a child who sees sticks as sticks. But what about pure abstractions, say something that looks just like A but is entitled No. 7? The 10th Street abstractionist blankly insists that there is nothing here but white paint and black, and none of our literary identifications need apply.

11

What then distinguishes him from Testadura, whose philistine utterances are indiscernible from his? And how can it be an artwork for him and not for Testadura, when they agree that there is nothing that does not meet the eye? The answer, unpopular as it is likely to be to purists of every variety, lies in the fact that this artist has returned to the physicality of paint through an atmosphere compounded of artistic theories and the history of recent and remote painting, elements of which he is trying to refine out of his own work; and as a consequence of this his work belongs in this atmosphere and is part of this history. He has achieved abstraction through rejection of artistic identifications, returning to the real world from which such identifications remove us (he thinks), somewhat in the mode of Ch’ing Yuan, who wrote: Before I had studied Zen for thirty years, I saw mountains as mountains and waters as waters. When I arrived at a more intimate knowledge, I came to the point where I saw that mountains are not mountains, and waters are not waters. But now that I have got the very substance I am at rest. For it is just that I see mountains once again as mountains, and waters once again as waters.

His identification of what he has made is logically dependent upon the theories and history he rejects.The difference between his utterance and Testadura’s “This is black paint and white paint and nothing more” lies in the fact that he is still using the is of artistic identification, so that his use of “That black paint is black paint” is not a tautology.Testadura is not at that stage. To see something as art requires something the eye cannot decry  –  an atmosphere of artistic theory, a knowledge of the history of art: an artworld.

III Mr Andy Warhol, the Pop artist, displays facsimiles of Brillo cartons, piled high, in neat stacks, as in the stockroom of the supermarket. They happen to be of wood, painted to look like cardboard, and why not? To paraphrase the critic of the Times, if one may make the facsimile of a human being out of bronze, why not the facsimile of a Brillo carton out of plywood? The cost of these boxes happens to be 2 × 103 that of their homely counterparts in real life – a differential hardly ascribable to their advantage in durability. In fact the Brillo people might, at some slight increase in cost, make their boxes out of plywood without these becoming artworks, and Warhol might make his out of cardboard without their ceasing to be art. So we may forget questions of

12

arthur c. danto

intrinsic value, and ask why the Brillo people cannot manufacture art and why Warhol cannot but make artworks. Well, his are made by hand, to be sure. Which is like an insane reversal of Picasso’s strategy in pasting the label from a bottle of Suze onto a drawing, saying as it were that the academic artist, concerned with exact imitation, must always fall short of the real thing: so why not just use the real thing? The Pop artist laboriously reproduces machine‐ made objects by hand, e.g., painting the labels on coffee cans (one can hear the familiar commendation “Entirely made by hand” falling painfully out of the guide’s vocabulary when confronted by these objects). But the difference cannot consist in craft: a man who carved pebbles out of stones and carefully constructed a work called Gravel Pile might invoke the labor theory of value to account for the price he demands; but the question is, What makes it art? And why need Warhol make these things anyway? Why not just scrawl his signature across one? Or crush one up and display it as Crushed Brillo Box (“A protest against mechanization …”) or simply display a Brillo carton as Uncrushed Brillo Box (“A bold affirmation of the plastic authenticity of industrial …”) ? Is this man a kind of Midas, turning whatever he touches into the gold of pure art? And the whole world consisting of latent artworks waiting, like the bread and wine of reality, to be transfigured, through some dark mystery, into the indiscernible flesh and blood of the sacrament? Never mind that the Brillo box may not be good, much less great art. The impressive thing is that it is art at all. But if it is, why are not the indiscernible Brillo boxes that are in the stockroom? Or has the whole distinction between art and reality broken down? Suppose a man collects objects (ready‐mades), including a Brillo carton; we praise the exhibit for variety, ingenuity, what you will. Next he exhibits nothing but Brillo cartons, and we criticize it as dull, repetitive, self‐plagiarizing – or (more profoundly) claim that he is obsessed by regularity and repetition, as in Marienbad. Or he piles them high, leaving a narrow path; we tread our way through the smooth opaque stacks and find it an unsettling experience, and write it up as the closing in of consumer products, confining us as prisoners: or we say he is a modern pyramid builder. True, we don’t say these things about the stockboy. But then a stockroom is not an art gallery, and we cannot readily separate the Brillo cartons from the gallery they are in, any more than we can separate the Rauschenberg bed from the paint upon it. Outside the gallery, they are pasteboard cartons. But then, scoured clean of paint, Rauschenberg’s bed is a bed, just what it was before it was transformed into art. But then if we think this matter

through, we discover that the artist has failed, really and of necessity, to produce a mere real object. He has produced an artwork, his use of real Brillo cartons being but an expansion of the resources available to arists, a contribution to artists’ materials, as oil paint was, or tuche. What in the end makes the difference between a Brillo box and a work of art consisting of a Brillo Box is a certain theory of art. It is the theory that takes it up into the world of art, and keeps it from collapsing into the real object which it is (in a sense of is other than that of artistic identification). Of course, without the theory, one is unlikely to see it as art, and in order to see it as part of the Artworld, one must have mastered a good deal of artistic theory as well as a considerable amount of the history of recent New York painting. It could not have been art fifty years ago. But then there could not have been, everything being equal, flight insurance in the Middle Ages, or Etruscan typewriter erasers. The world has to be ready for certain things, the Artworld no less than the real one. It is the role of artistic theories, these days as always, to make the Artworld, and art, possible. It would, I should think, never have occurred to the painters of Lascaux that they were producing art on those walls. Not unless there were neolithic aestheticians.

IV The Artworld stands to the real world in something like the relationship in which the City of God stands to the Earthly City. Certain objects, like certain individuals, enjoy a double citizenship, but there remains, the RT notwithstanding, a fundamental contrast between artworks and real objects. Perhaps this was already dimly sensed by the early framers of the IT who, inchoately realizing the nonreality of art, were perhaps limited only in supposing that the sole way objects had of being other than real is to be sham, so that artworks necessarily had to be imitations of real objects. This was too narrow. So Yeats saw in writing “Once out of nature I shall never take/My bodily form from any natural thing.” It is but a matter of choice: and the Brillo box of the Artworld may be just the Brillo box of the real one, separated and united by the is of artistic identification. But I should like to say some final words about the theories that make artworks possible, and their relationship to one another. In so doing, I shall beg some of the hardest philosophical questions I know. I shall now think of pairs of predicates related to each other as “opposites,” conceding straight off the vagueness of  this demodé term. Contradictory predicates are not



the artworld

­ pposites, since one of each of them must apply to every o object in the universe, and neither of a pair of opposites need apply to some objects in the universe. An object must first be of a certain kind before either of a pair of opposites applies to it, and then at most and at least one of the opposites must apply to it. So opposites are not contraries, for contraries may both be false of some objects in the universe, but opposites cannot both be false; for of some objects, neither of a pair of opposites sensibly applies, unless the object is of the right sort. Then, if the object is of the required kind, the opposites behave as contradictories. If F and non‐F are opposites, an object o must be of a certain kind K before either of these sensibly applies; but if o is a member of K, then o either is F or non‐F, to the exclusion of the other. The class of pairs of opposites that sensibly apply to the (ô)Ko I shall designate as the class of K‐relevant predicates. And a necessary condition for an object to be of a kind K is that at least one pair of K‐relevant opposites be sensibly applicable to it. But, in fact, if an object is of kind K, at least and at most one of each K‐relevant pair of opposites applies to it. I am now interested in the K‐relevant predicates for the class K of artworks. And let F and non‐F be an opposite pair of such predicates. Now it might happen that, throughout an entire period of time, every artwork is non‐F. But since nothing thus far is both an artwork and F, it might never occur to anyone that non‐F is an artistically relevant predicate. The non‐F‐ness of artworks goes unmarked. By contrast, all works up to a given time might be G, it never occurring to anyone until that time that something might both be an artwork and non‐G; indeed, it might have been thought that G was a defining trait of artworks when in fact something might first have to be an artwork before G is sensibly predicable of it – in which case non‐G might also be predicable of artworks, and G itself then could not have been a defining trait of this class. Let G be ‘is representational’ and let F be ‘is expressionist’. At a given time, these and their opposites are perhaps the only art‐relevant predicates in critical use. Now l­ etting ‘+’ stand for a given predicate P and ‘−’ for its opposite non‐P, we may construct a style matrix more or less as follows:

F + + – –

G + – + –

13

The rows determine available styles, given the active critical vocabulary: representational expressionistic (e.g., Fauvism); representational nonexpressionistic (Ingres); nonrepresentational expressionistic (Abstract Expressionism); nonrepresentational non‐expressionist (hard‐edge abstraction). Plainly, as we add art relevant predicates, we increase the number of available styles at the rate of 2n. It is, of course, not easy to see in advance which predicates are going to be added or replaced by their opposites, but suppose an artist determines that H shall henceforth be artistically relevant for his paintings. Then, in fact, both H and non‐H become artistically relevant for all painting, and if his is the first and only painting that is H, every other painting in existence becomes non‐H, and the entire community of paintings is enriched, together with a doubling of the available style opportunities. It is this retroactive enrichment of the entities in the Artworld that makes it possible to discuss Raphael and De Kooning together, or Lichtenstein and Michelangelo. The greater the variety of artistically relevant predicates, the more complex the individual members of the Artworld become; and the more one knows of the entire population of the Artworld, the richer one’s experience with any of its members. In this regard, notice that, if there are m artistically relevant predicates, there is always a bottom row with m minuses. This row is apt to be occupied by purists. Having scoured their canvasses clear of what they regard as inessential, they credit themselves with having distilled out the essence of art. But this is just their fallacy: exactly as many artistically relevant predicates stand true of their square monochromes as stand true of any member of the Artworld, and they can exist as artworks only insofar as “impure” paintings exist. Strictly speaking, a black square by Reinhardt is artistically as rich as Titian’s Sacred and Profane Love. This explains how less is more. Fashion, as it happens, favors certain rows of the style matrix: museums, connoisseurs, and others are makeweights in the Artworld. To insist, or seek to, that all artists become representational, perhaps to gain entry into a specially prestigious exhibition, cuts the available style matrix in half: there are then 2n/2 ways of satisfying the requirement, and museums then can exhibit all these “approaches” to the topic they have set. But this is a matter of almost purely sociological interest: one row in the matrix is as legitimate as another. An artistic breakthrough consists, I suppose, in adding the possibility of a column to the matrix. Artists then, with greater or less alacrity, occupy the positions thus opened up: this is a remarkable feature of contemporary art, and for those unfamiliar with the matrix, it is hard, and

14

arthur c. danto

perhaps impossible, to recognize certain positions as occupied by artworks. Nor would these things be artworks without the theories and the histories of the Artworld. Brillo boxes enter the Artworld with that same tonic incongruity the commedia dell’arte characters bring into Ariadne auf Naxos.Whatever is the artistically relevant predicate in virtue of which they gain their entry, the rest of the

Artworld becomes that much the richer in having the opposite predicate available and applicable to its members. And, to return to the views of Hamlet with which we began this discussion, Brillo boxes may reveal us to ourselves as well as anything might: as a mirror held up to nature, they might serve to catch the conscience of our kings.

2

The New Institutional Theory of Art George Dickie

The version of the institutional theory that I worked out in 1974 in Art and the Aesthetic1 was defective in several respects, but the institutional approach is, I think, still viable. By an institutional approach I mean the idea that works of art are art as the result of the position they occupy within an institutional framework or context. I have tried in […] The Art Circle, to work out a revised version of the theory. In this paper, I shall attempt to give a summary account of the new version of the institutional theory of art. It should be made clear here at the beginning that the theory of art I am trying to work out is a classificatory one. Some theories of art have assumed that a work of art is necessarily a good thing, but this assumption would leave unaccounted for all the mediocre, bad, and worthless art. It is the wider class of objects which contain the worthless, the indifferent, the mediocre, the good, and the masterpieces about which I am concerned to theorize. Traditional theories of art place works of art within ­simple and narrowly focused networks of relations. The imitation theory, for example, suspends the work of art in a three‐place network between artist and subject matter, and the expression theory places the work of art in a two‐place network of artist and work. The institutional theory attempts to place the work of art within a multi‐placed network of greater complexity than anything envisaged by the various traditional theories. The networks or contexts of the traditional theories are too “thin” to be sufficient.

The institutional theory attempts to provide a context which is “thick” enough to do the job. The network of relations or context within which a theory places works of art I shall call “the framework” of that theory. Despite my reservations about the traditional theories of art, they were, I believe, on the right track about the group of objects they focus on. All of the traditional theories assume that works of art are artifacts, although they differ about the nature of the artifacts. There is, then, a sense in which the institutional approach is a return to the traditional way of theorizing about art for it too maintains that works of art are artifacts. By the way, what is meant by “artifact” here is the ordinary dictionary definition: “an object made by man, especially with a view to subsequent use.” Furthermore, although many are, an artifact need not be a physical object: for example, a poem is not a physical object, but it is, nevertheless, an artifact. Still further, things such as performances, for example, improvised dances, are also “made by man” and are, therefore, artifacts. In the 1950s, first Paul Ziff and then Morris Weitz challenged the assumption of artifactuality, claiming that being an artifact is not a necessary condition of art. Although Ziff ’s and White’s views differ somewhat, they have in common the claim that there is no necessary condition for something’s being art, not even artifactuality. Their common view can be called “the new conception of art.” This new view conceives of the members of the class of works

George Dickie, “The New Institutional Theory of Art,” Proceedings of the 8th Wittgenstein Symposium 10 (1983), pp. 57–64. Public domain.

Aesthetics and the Philosophy of Art – The Analytic Tradition: An Anthology, Second Edition. Edited by Peter Lamarque and Stein Haugom Olsen. © 2019 John Wiley & Sons, Inc. Published 2019 by John Wiley & Sons, Inc.

16

george dickie

of art as having no common feature of any theoretical ­significance. The members of the class are related only by means of similarities: work of art A resembles work of art B and work of art B resembles work of art C, but A does not have to resemble C. According to the new view, an object becomes a work of art by sufficiently resembling a prior‐ established work of art. The new conception speaks of sufficient resemblance as the only way that a work of art can come into being. An examination of the new view reveals, however, that it entails that there must be another way than sufficient resemblance to a prior‐established work of art for a work of art to come into being. That two ways of becoming art are required by the new conception of art can be shown in the following way. Suppose that work of art A had become art by sufficiently resembling prior‐established work of art B. Work of art B would have had to become art by sufficiently resembling an earlier prior‐established work of art, call it C. If resemblance to a prior‐established work of art is the only way of becoming art, then the way back in time from work of art A to work of art B to work of art C generates an infinite regress of works of art receding into the past. If resemblance to a prior‐established work of art were the only way of becoming art, there could be no first work of art and, consequently, there could not be any art at all. Some way of becoming art other than resemblance to a prior‐established work is required for resemblance to a prior‐established work to function as a way of becoming art. Works of art which become art by sufficiently resembling prior‐established works may be called “similarity art.” In order for there to be similarity art there must be at least one work of art which did not become art in virtue of its similarity to a prior‐established work of art. Consequently, the new conception of art really requires two ways of becoming art: the similarity way and some nonsimilarity way. The new conception is an unacknowledged “double” theory of art. What is the nature of the nonsimilarity art required by the new conception? Since neither Ziff nor White was aware that their view requires nonsimilarity art, it is not surprising that they said nothing about it. The nature of nonsimilarity art will have to be inferred from the stated theory. First, nonsimilarity art is primary within the theory  –  there could not be similarity art unless there is first nonsimilarity art. Second, the class of works of art, according to the new conception, consists of two distinct subsets of which one (nonsimilarity art) is more basic than the other (similarity art). Finally, there is nothing in the new conception of art or outside of it which requires non-

similarity art to be a one‐time sort of thing the only function of which is to block the regress and get the art process going. Although nothing in the new conception entails that it is, the only plausible account of the nature of ­nonsimilarity art that I can think of is that it is art which is art as the result of someone’s creating an artifact. This, of course, does not prove that nonsimilarity art is to be identified with what may be called “artifactual art,” but artifactual art seems to be the only real contender. The new conception of art involves two distinct kinds of art – artifactual art and similarity – with the former being primary. Artifactual art is clearly not confined to the beginning of the art process, because such art is being created at the present and has been created throughout the history of art. Ziff and Weitz demand that if one is to theorize about art, one must produce a theory which encompasses all members of the class of works of art. And according to their view, the members of the class have no common feature or features. Consequently, they claim that one cannot theorize about art in the traditional manner of discovering necessary and sufficient features. The closest they can come to theorizing about art is to say that there is a class of objects to which the terms “art” and “work of art” meaningfully apply and that this class cannot be theoretically characterized further. The earlier examination of the new conception of art has shown that the class of objects to which the terms “art” and “work of art” meaningfully apply divides into two distinct subclasses of art. This division shows that the class can be theoretically characterized further. The first thing to be noted about the subclasses is that the two activities which generate the two subclasses are very different. Artifactual art is generated by the human activity of making. Similarity art is generated by the human activity of noticing similarities. The strikingly different activities which generate the two subclasses suggest that the two classes are not literally subclasses of a single class. The two classes seem more like a class picked out by the literal uses of a term and a derivative class picked out by the metaphorical uses of the same term. I will not, however, pursue this point here. Even if one were to agree with Ziff and Weitz that ­artifactual art and similarity art are both literally art, why should this persuade philosophers to abandon their traditional concern with theorizing about what is in effect ­artifactual art? From Plato’s time forward, philosophers of art have been concerned to theorize about the class of objects which is generated by a particular kind of human making. Philosophers have been interested in these objects precisely because they are human artifacts. The fact that there is another class of objects which is in some way



the new institutional theory of art

derivative by means of similarities from the class of objects they have traditionally been interested in is not surprising and is no reason to divert philosophers of art from their traditional activity.That traditional activity is the attempt to describe correctly the nature of the making of artifactual art and, consequently, the nature of the objects made. Artifactuality is, in effect, a “built‐in” characteristic of the interest of p­ hilosophers in works of art. On the surface anyway, there is no mystery about the making of the great bulk of works of artifactual art; they are crafted in various traditional ways – painted, sculpted, and the like. (Later, I will attempt to go below the surface a bit.) There is, however, a puzzle about the artifactuality of some relatively recent works of art: Duchamp’s ready‐mades, found art, and the like. Some deny that such things are art because, they claim, they are not artifacts made by artists. It can, I think, be shown that they are the artifacts of artists. (In Art and the Aesthetic I claimed, I now think mistakenly, that artifactuality is conferred on things such as Duchamp’s Fountain and found art, but I will not discuss this here.) Typically an artifact is produced by altering some preexisting material: by joining two pieces of material, by cutting some material, by sharpening some material, and so on. This is typically done so that the altered material can be used to do something. When materials are so altered, one has clear cases which neatly fit the dictionary definition of “artifact”  –  “An object made by man, especially with a view to subsequent use.” There are other cases which are less clear‐cut. Suppose one picks up a piece of driftwood and without altering it in any way digs a hole or brandishes it at a threatening dog. The unaltered driftwood has been made into a digging tool or a weapon by the use to which it is put. These two cases do not conform to the nonnecessary clause of the definition “especially with a view to subsequent use” because they are pressed into service on the spot. There does seem to be a sense in which something is made in these cases, but what is it that has been made if the driftwood is unaltered? In the clear cases in which material is altered, a complex object is produced: the original material is for present purposes a simple object and its being altered produces the complex object – altered material. In the two less clear‐cut cases, complex objects have also been made – the wood used as a digging tool and the wood used as a weapon. In neither of the two less clear‐cut eases is the driftwood alone the artifact; the artifact in both cases is the driftwood manipulated and used in a certain way. The two cases in question are exactly like the sort of thing that anthropologists have in mind when they speak of unaltered stones found in conjunction with human or human‐like

17

fossils as artifacts. The anthropologists assume that the stones were used in some way. The anthropologists have in mind the same notion of a complex object made by the use of a simple (i.e., unaltered) object. A piece of driftwood may be used in a similar way within the context of the artworld, i.e., picked up and displayed in the way that a painting or a sculpture is displayed. Such a. piece of driftwood would be being used as an artistic medium and thereby would became part of the more complex object  –  the‐driftwood‐used‐as‐an‐artistic‐medium. This complex object would be an artifact of an artworld system. Duchamp’s Fountain can be understood along the same lines.The urinal (the simple object) is being used as an artistic medium to make Fountain (the complex object) which is an artifact within the artworld – Duchamp’s artifact. The driftwood would be being used and the urinal was used as artistic media in the way that pigments, marble, and the like are used to make more conventional works of art. Thus far, I have talked of artifactuality as a necessary condition of art, but this discussion does not distinguish the institutional theory from the traditional theories, as the ­latter have assumed or implied that being an artifact is a necessary condition of art. In the last paragraph, however, I introduced without explanation the notion of the artworld, and it is now time to turn to a discussion of the artworld, for it is this notion which lies at the heart of the institutional theory. Perhaps the best way to begin a discussion of the artworld is to quote the now‐abandoned definition of “work of art” from the earlier version of the institutional theory. “A work of art in the classificatory sense is (1) an artifact (2) a set of the aspects of which has had conferred upon it the status of candidate for appreciation by some person or persons acting on behalf of a certain social institution (the artworld).”2 Monroe Beardsley has observed that in the ­discussion which surrounds the definition in the earlier version of the theory I characterized the artworld as an “established practice” which is to say, an informal kind of activity He then goes on to point out that the quoted definition makes use of such phrases as “conferred status” and “acting on behalf of.” Such phrases typically have application within formal institutions such as states, corporations, universities, and the like. Beardsley correctly notes that it is a mistake to use the language of formal institutions to try to describe an informal institution as I conceive the artworld to be. Beardsley queries, “… does it make sense to speak of acting on behalf of a practice? Status‐awarding authority can center in [a formal institution], but practices, as such, seem to lack the requisite source of authority.”3

18

george dickie

Accepting Beardsley’s criticism, I have abandoned as too formal the notions of status conferral and acting on behalf of as well as those aspects of the earlier version which connect up with these notions. Being a work of art is a status all right, that is, it is the occupying of a position within the human activity of the artworld. Being a work of art is not, however, a status which is conferred but is rather a status which is achieved as the result of creating an artifact within or against the background of the artworld. The claim is then that works of art are art as the result of the position or place they occupy within an established practice, namely, the artworld. There are two crucial questions about the claim. (1) Is the claim true and (2) if the claim is true, how is the artworld to be described? The claim is a claim about the existence of a human institution, and the test of its truth is the same as for any other claim about human organization – the test of observation. “Seeing” the artworld and the works of art embedded in its structures, however, is not as easy as “seeing” some of the other human institutions which we are more accustomed to thinking about. Arthur Danto has invented an argument which helps somewhat in “seeing” the structure in which works of art are embedded. (I must note, however, that what Danto himself “sees” with the use of his argument is quite different from what I “see,” but I will not here attempt to rebut Danto’s theory.) My version of Danto’s argument runs as follows. Consider a painting and another object which looks exactly like it but which was produced accidently and is, therefore, not a work of art. Or consider Fountain and a urinal which is its twin but is not a work of art. Here are two pairs of objects with visually indistinguishable ­elements, but the first element in each pair is a work of art and the second element is not. The fact that the first ­element of each pair is a work of art and the second element is not although the elements of each pair are visually indistinguishable shows that the first object in each pair must be enmeshed in same sort of framework or network of ­relations in which the second element is not. It is the first element’s being enmeshed in the framework which accounts for its being a work of art, and it is the second element’s not being enmeshed in the framework which accounts for its not being a work of art. The framework in question is not, of course, visible to the eye in the way that the colors of the two objects are. Some will argue that the Fountain/urinal pair does not show anything because Fountain is not a work of art. Fortunately, the other hypothetical pair is sufficient to get the argument off the ground. The Fountain/urinal pair,

however, can also be shown to suffice even if Fountain is not a work of art. Fountain does not actually have to be a work of art to show the necessity of a context or framework. It is sufficient for the argument that at some time some person mistakenly thought Fountain to be a work of art. The framework within which Fountain apparently had a place would in this case explain the mistake. And, some persons have thought Fountain to be a work of art. Danto’s argument shows that works of art exist within a context or framework, but it does not reveal the nature of the elements which make up the framework. Moreover, many different frameworks are possible. Each of the traditional theories of art, for example, implies its own particular framework. For one example, Susanne Langer’s view that “Art is the creation of forms symbolic of human feeling” implies a framework of artist (one who creates) and a specific kind of subject matter (human feeling). And as I noted at the beginning of the paper, the imitation theory and the expression theory each implies a particular framework. Langer’s theory and the other traditional theories, however, fall easy prey to counterexamples, and, consequently, none of the frameworks they imply can be the right one. The reason that the traditional theories are easy prey for counterexamples is that the frameworks implied by the theories are too narrowly focused on the artist and various different of the more obvious characteristics which works of art may have rather than on all the framework elements which ­surround works of art. The result is that it is all too easy to find works of art which lack the properties seized upon by a particular traditional theory as universal and defining. The frameworks of the traditional theories do lead in the right direction in one respect. Each of the traditional theories conceives of the making of art as a human practice, as an established way of behaving. The framework of each of these theories is conceived of, then, as a cultural phenomenon which persists through time and is repeatable. The persistence of a framework as a cultural practice is enough, I think, to make the traditional theories themselves quasi‐ institutional. That is to say, each of the traditional theories purports to describe an established cultural practice. In every one of the traditional theories, however, there is only one established role envisioned and that is the role of the artist or the maker of artifacts. And in every case, the artist is seen as the creator of an artifact with a property such as being representative, being symbolic, or being an expression. For the traditional theories the artist role is envisaged as simply that of producing representations, producing symbolic forms, producing expressions, or some such thing. It is this narrow conception of the artist role which is



the new institutional theory of art

responsible for the ease with which counterexamples can be produced. Since the traditional theories are inadequate, there must be more to the artist role than the producing of any, or even all, of these kinds of things which the traditional theories envisage. What an artist understands and does when he creates a work of art far exceeds the simple understanding and doing entailed by the traditional theories. Whenever art is created there is, then, an artist who ­creates it, but an artist always creates for a public of some sort. Consequently, the framework must include a role for a public to whom art is presented. Of course, for a variety of reasons many works of art are never in fact presented to any public. Some works just never reach their public although their makers intended for them to do so. Some works are withheld from their publics by their creators because they judge them to be in some way inferior and unworthy of presentation. The fact that artists withhold some of their works because they judge them unworthy of presentation shows that the works are things of a kind to be presented, otherwise, it would be pointless to judge them unworthy of ­presentation. Thus, even art not intended for public presentation presupposes a public, for not only is it possible to present it to a public (as sometimes happens), it is a thing of a type which has as a goal presentation to a public. The notion of a public hovers always in the background, even when a given artist refuses to present his work. In those cases in which works of art are withheld from a public, there is what might be called a “double intention” – there is an intention to create a thing of a kind which is presented, but there is also an intention not to actually present it. But what is an artworld public? Such a public is not just a collection of people. The members of an artworld public are such because they know how to fulfill a role which requires knowledge and understanding similar in many respects to that required of an artist. There are as many ­different publics as there are different arts, and the knowledge required for one public is different from that required by another public. An example of one bit of knowledge required of the public of stage plays is the understanding of what it is for someone to act a part. Any given member of a public would have a great many such bits of information. The artist and public roles are the minimum framework for the creation of art, and the two roles in relation may be called “the presentation group.” The role of artist has two central aspects: first, a general aspect characteristic of all ­artists, namely, the awareness that what is created for presentation is art, and, second, the ability to use one or more of a wide variety of art techniques which enable one to

19

create art of a particular kind. Likewise, the role of a public has two central aspects: first, a general aspect characteristic of all publics, namely, the awareness that what is presented to it is art and, second, the abilities and sensitivities which enable one to perceive and understand the particular kind of art with which one is presented. In almost every actual society which has an institution of art‐making, in addition to the roles of artist and public, there will be a number of supplementary artworld roles such as those of critic, art teacher, director, curator, conductor, and many more. The presentation group, i.e., the roles of artist and public in relation, however, constitutes the essential framework for art‐making. Among the more frequent criticisms of Art and the Aesthetic was that it failed to show that art‐making is ­institutional because it failed to show that art‐making is rule‐governed. The underlying assumption of the criticism is that it is rule‐governedness which distinguishes institutional practices such as, say, promising from noninstitutional ones such as, say, dog‐walking. And it is true that Art and the Aesthetic did not bring out the rule‐governedness of art‐ making and this requires correcting.There are rules implicit in the theory developed in the earlier book, but unfortunately I failed to make them explicit. There is no point in discussing the rules governing art‐making implicit in the earlier theory, but those of the present revised theory can be stated. Earlier in this paper I argued that artifactuality is a necessary condition for being a work of art. This claim of necessity implies one rule of art‐making; if one wishes to make a work of art, one must do so by creating an artifact. Also earlier in this paper I claimed that being a thing of a kind which is presented to an artworld public is a necessary condition for being a work of art. This claim of necessity implies another rule of art‐making; if one wishes to create a work of art, one must do so by creating a thing of a kind which is presented to an artworld public. These two rules are jointly sufficient for making works of art. The question naturally arises as to why the framework described as the institutional one is the correct essential framework rather than some other framework. The framework of the traditional theories are clearly inadequate, but their inadequacy does not prove the correctness of the framework of the present version of the institutional theory. Proving that a theory is true is notoriously difficult to do, although proving that a theory is false is sometimes easy to do. It can be said of the present version of the institutional theory that it is a conception of a framework in which works of art are clearly embedded and that no other plausible framework is in the offing. For lack of a more conclusive

20

george dickie

argument that the institutional theory’s framework is the right one, I shall have to rely on the description of it I have given to function as an argument as to its rightness. If the description is correct, or approximately so, then it should evoke a “that’s right” experience in the listener. In the remainder of the paper I shall, in effect, continue my description of the essential framework for the creation of art. In Art and the Aesthetic I talked a great deal about conventions and how they are involved in the institution of art. In that book, I tried to distinguish between what I called “the primary convention” and other “secondary conventions” which are involved in the creation and presentation of art. One example of the so‐called secondary conventions discussed there is the Western theatrical convention of ­ ­concealing stagehands behind the scenery. This Western convention was there contrasted with that of classical Chinese theater in which the stagehand (called the property man) appears on stage during the action of the play and rearranges props and scenery. These two different ­theatrical solutions for the same task, namely, the employment of stagehands, brings out an essential feature of ­conventions. Any conventional way of doing something could have been done in a different way. The failure to realize that things of the kind just discussed are conventions can result in confused theory. For example, it is another convention of Western theater that spectators do not participate in the action of a play. Certain aesthetic‐attitude theorists failed to realize that this particular convention is a convention and concluded that the nonparticipation of spectators is a rule derived from ­aesthetic consciousness and that the rule must not be violated. Such theorists are horrified by Peter Pan’s request for the members of the audience to applaud to save Tinkerbell’s life.The request, however, merely amounts to the introduction of a new convention which small children, but not some aestheticians, catch on to right away. There are innumerable conventions involved in the creation and presentation of art, but there is not, as I ­ claimed in my earlier book, a primary convention to which all the other conventions are secondary. In effect, in Art and the Aesthetic I claimed that not only are there many conventions involved in the creation and presentation of art, but that at bottom the whole activity is completely conventional. But theater, painting, sculpting, and the like, are not ways of doing something which could be done in another way, and, therefore, they are not conventional. If, however, there is no primary convention, there is a primary something within which the innumerable conventions that there are have a place. What is primary is the understanding shared

by all involved that they are engaged in an established activity or practice within which there is a variety of roles: artist roles, public roles, critic roles, director roles, curator roles, and so on. Our artworld consists of the totality of such roles with the roles of artist and public at its core. Described in a somewhat more structured way, the artworld consists of a set of individual artworld systems, each of which contains its own specific artist and public roles plus other roles. For example, painting is one artworld system, theater is another, and so on. The institution of art, then, involves rules of very different kinds. There are conventional rules which derive from the various conventions employed in presenting and creating art. These rules are subject to change. There are more basic rules which govern the engaging in an activity, and these rules are not conventional. The artifact rule – if one wishes to make a work of art, one must do so by creating an ­artifact – is not a conventional rule; it states a condition for engaging in a certain kind of practice. As I remarked earlier, the artifact rule and the other nonconventional rule are sufficient for the creating of art. And, as each rule is necessary, they can be used to formulate a definition of “work of art”: A work of art is an artifact of a kind created to be presented to an artworld public.

This definition explicitly contains the terms “artworld” and “public,” both of which have been discussed but not defined in this paper.The definition also involves the notions of artist and artworld system, both of which have been d­ iscussed but not definitionally characterized in this paper. I shall not attempt to define either “artist,”“public,”“artworld,” or “artworld system” here, as I do in my book ­manuscript, but the definition of “work of art” given here and the definitions of these other four central terms provide the leanest possible description of the institutional theory of art. To forestall an objection to the definition, let me acknowledge that there are artifacts which are created for presentation to the artworld publics which are not works of art: for example, playbills. Such things are, however, parasitic or secondary to works of art. Works of art are artifacts of a primary kind in this domain, and playbills and the like which are dependent on works of art are artifacts of a ­secondary kind within this domain. The word “artifact” in the definition should be understood to be referring to ­artifacts of primary kind. The definition of “work of art” given in Art and the Aesthetic was, as I affirmed there, circular, although not



the new institutional theory of art

viciously so. The definition of “work of art” just given is also circular, although again not viciously so. In fact, the definitions of the five central terms constitute a logically circular set of terms. There is an ideal of noncircular definition which assumes that the meaning of terms used in a definition ought not to lead back to the term originally defined, but rather ought to be or lead to terms which are more basic. The ideal of noncircular definition also assumes that we ought to be able to arrive at terms which are primitive in the sense that they can be known in some nondefinitional way, say, by direct sensory experience or by rational intuition. There may be some sets of definitions which satisfy this ideal, but the definitions of the five central terms of the institutional theory do not. Does this mean that the institutional theory involves a vicious circularity? The circularity of the definitions shows the interdependency of the central notions.These central notions are inflected, that is, they bend in on, presuppose, and support one another. What the definitions reveal is that art‐making involves an intricate, co‐relative structure which cannot be described in the straight‐forward, linear way envisaged by the ideal

21

of noncircular definition. The inflected nature of art is reflected in the way we learn about art. This learning is sometimes approached through being taught how to be an artist  –  learning how to draw pictures which can be displayed, for example. This learning is sometimes approached through being taught how to be a member of an artworld public  –  learning how to look at pictures which are presented as the intentional products of artists. Both approaches teach us about artists, works, and publics all at the same time, for these notions are not independent of one another. I suspect that many areas within the cultural domain also have the same kind of inflected ­ nature that the institution of art has. For example, the area which involves the notions of law, legislature, executive, and judiciary. The ideal of noncircular definition holds also that sets of circular definitions cannot be informative.This may be true of some sets of definitions, but it is not, I think, true of the definitions of the institutional theory. For these definitions just mirror the mutually dependent items which constitute the art enterprise, and, thereby, inform us of its inflected nature.

Notes 1 G. Dickie, Art and the Aesthetic (Ithaca and London, 1974). 2 Ibid., p. 34.

3 M. Beardsley, “Is Art Essentially Institutional?,” in Culture and Art, ed. Lars Aagaard‐Mogensen (Atlantic Highlands, NJ, 1976), p. 202.

3

An Aesthetic Definition of Art Monroe C. Beardsley

Like other questions of the same syntactic form, the question “What is art?” invites analysis of words and ­ thoughts as well as the phenomena they refer to. But, being a philosophical question, it will not be satisfied by lexicography or psychology. I don’t suppose that my task here is to inquire how various people commonly, or uncommonly, use the word “art” in English, or corresponding words in other languages, nor to canvas popular, or unpopular, opinion about what art is or ought to be. Taken philosophically, the question calls for decisions and proposals: What are the noteworthy features of the phenomena to which the word in question seems, however loosely, to call our attention? What are the significant distinctions that need to be marked for the purposes of theoretical understanding, and that the word “art” or one of its cognates (“artwork,” “artistic,” “artistry,” etc.) is most apt and suitable for marking? How does art, defined in a comparatively clear, if somewhat unorthodox, way, differ from closely related things? Of course it will be a merit in any proposed definition of art if it matches reasonably well at least one fairly widespread use of the term, but it will not necessarily be a merit if the match is close, since philosophical reflection is expected to yield definitions and distinctions rather more valuable – to philosophy – than casual familiar ones. I proclaim these assumptions or biasses at the start, to show my conception of my task – for which no further defense or apology will be forthcoming. What I will try to defend is

a definition of art – actually, a rather old‐fashioned one in essentials – that I have come to regard as best adapted to the requirements of a sound philosophy of art.

I It should not be necessary to argue that we have need for a definition of art, but since this has been vigorously denied, some argument must be given. The point of a definition is, of course, to fix a meaning – to establish and stabilize it for some range of contexts  –  and thus mark out a class of things to be referred to by some group of people. One would think the philosopher of art could use a definition, since he should be curious to know what he is philosophizing about. Not that he can begin with an adequate definition – he only needs some uncontested examples of art, which are surely not hard to come by – but at some point he will want to say why these are examples of art, and how we are to tell what future things his philosophical conclusions are supposed to apply to.The critic of an art may be a similar case; it should be useful to him to have criteria for deciding what sorts of thing he is to criticize.The historian of an art – of dance, drama, architecture, or whatever – will surely have use for a definition to tell him what belongs in his history and what does not: Should the dance historian

Monroe C. Beardsley, “An Aesthetic Definition of Art,” in Hugh Curtler (ed.), What Is Art? (New York: Haven Publications, 1983), pp. 15–29. Reproduced by permission of the editor.

Aesthetics and the Philosophy of Art – The Analytic Tradition: An Anthology, Second Edition. Edited by Peter Lamarque and Stein Haugom Olsen. © 2019 John Wiley & Sons, Inc. Published 2019 by John Wiley & Sons, Inc.



an aesthetic definition of art

deal with parades and with cavorting bears? Should the architecture historian deal with igloos and Macdonald’s eateries? If so, why? If not, why not? To answer such questions well requires a defensible definition. Even the practical legislator or administrator may have use for a definition, in deciding, for example, which imported objects to exempt from duties, or which allegedly artistic projects should be funded by the National Endowment for the Arts. But we must especially note the needs of the anthropologist, and indeed as we move on we should keep in mind this broad cross‐cultural perspective. Essential to our understanding of any culture is a grasp of the various forms of activity that it manifests, and of distinctions that are most significant to the members of the society that has that culture. When we observe someone carving wood or moving about in a circle with others, we must ask whether the activity is religious, political, economic, medical, etc. – or artistic. What function do the participants think of themselves as fulfilling, and how does it relate to other activities in which they engage? Even when the same act of carving or dancing in a ring has more than one character  –  it is both religious and artistic, say  –  we don’t understand it unless we make this distinction and see that both descriptions apply. Distinction does not entail separation, and it is the distinction that is basic – though of course it may also be important to note that in one society these two forms of activity are always combined (the artists are the priests), while in another society they are kept apart and assigned to  specialized persons. In a very few societies, horribly deprived or nearing destruction, we may be able to discern no activities to label “artistic,” but there seems to be an ingredient or dimension of culture that runs across most societies, however varied, and whose nature we would like to articulate as well as we can. In other words, we want to define. There are two definition questions that spring up when we take this anthropological point of view.We may, as I said, observe activities, and want to know which activities are artistic ones (allowing that they may also be political or pedagogical, etc.). We may notice certain objects that seem to engage the attention of many persons in the society from time to time, and want to know which of them, if any, are artworks (again allowing that they may also be sacred objects or economic objects, etc.). Here are two paired concepts, that of artistic activity and that of artwork; and already we are faced with one of the definitional decisions we must make. If we have an independent definition of “artwork,” it is easy to define “artistic activity” as “activity that involves dealing with artworks.” But it may turn out to

23

be impossible, or inconvenient, to give a satisfactory definition of “artwork” without making reference to some form of artistic activity, either on the part of those who create artworks or those who receive them or both. We can hold off this problem for a moment. If our definition of art must be designed to serve the varied purposes, both theoretical and practical, that I have reviewed, we obviously cannot accept some proposals that have been made in recent years and that still have adherents. It will not be wise to take as our basic notion that of an art as a skill of some kind. Of course this is a common use of the term, and has its value: we speak of the art of medicine, of salesmanship, of motorcycle maintenance, of cooking, of war. In a society in which various forms of activity, such as those just named, have been developed to a fairly high degree, they are done with art, artfully, in this broad sense. But when we want to distinguish the kind of activity that, in our society, results in oil paintings on museum walls, poems in books, sounds in concert halls, we are clearly after a narrower concept of art, rather remote perhaps from salesmanship, motorcycle maintenance, and war. It is this narrower concept that we must aim to capture in our definition – or at least some concept that will make a significant distinction and in so doing show us how to gather up into a single class a great many paintings, poems, and musical compositions and performances. Nor will it help us, I think, to introduce the concept of institution at this point, even though this concept, when carefully delimited, is of great importance in the study of culture. First, it does not seem prudent to make noninstitutional artistic activities impossible by definition. Much of the artistic activity that we are familiar with in our society or observe in others is indeed institutionalized (for example, the Philadelphia Orchestra and New Directions Press). Some artistic activity occurs in the form of general practices (such as rural whittling and bluegrass musicmaking), without being, in a usefully strict sense, institutionalized at all. And it may be that still other artistic activities (such as amateur film‐making and Sunday painting) should not even be called practices, though they may recur. At least, we should want our definitions to leave open the possibility of new forms of artistic activity appearing before they become encompassed by institutions. To define any form of activity in terms of the concept of institution, rather than the other way about, seems to be to invert logical order: how can we conceive of religious, political, artistic, and other institutions except in terms of the forms of activity that they sponsor and regularize? When, in studying an unfamiliar society, we try to identify the commercial institutions, say, we must

24

monroe c. beardsley

surely begin by looking for sustained cooperative group activities (such as bartering and hoarding) that we acknowledge already to be economic in character. Even more obviously, we cannot follow those who hold that in order to justify the current avant‐garde we must allow that anything is art if anyone says it is. If being called “art” by someone is all it takes for something (say, an earthquake) to be an artwork, then when we say something is art we are saying no more than someone has applied this label to it. But of course our saying is also a performative, since in saying it is art we ourselves are applying the label to it; so the sentence “that is art” can never be false, no matter what it is said of or who says it. Sentences that become true merely by uttering them can be interesting if there are some rules to follow in uttering them, and hence some restrictions on when and by whom they can take effect. But sentences governed by no such rules have no point; on this proposal, the word “artwork” simply loses all content and becomes empty. No doubt there is sociological significance in the fact that there has been this effort in our society, in our age, to get us to cease to make any distinction between artistic and nonartistic activities, and between artworks and other things. This effort, I think, partly reflects a confused notion that if we allow a distinction we will encourage a separation; and it is partly a confused way of pleading for a far wider range of artistic activity than has ever before been recognized. But differences do not disappear just because we resolve to ignore them, and the avant‐ garde effort at erasure has no bearing on the philosophical task of finding out what differences are significant and deserve to be marked by appropriate terms.

II Once we know what things are artworks in a particular society (and anything that is an artwork in some society is an artwork tout court), we can identify artistic activities by discovering which activities involve interaction with artworks. Artistic activities, so conceived, may be extremely varied, especially in a complex society, though we might want to draw some useful lines to separate interactions from more remote or indirect relationships. Thus editors of novels and dance coaches are plainly engaged in artistic activities, but typesetters of novels and doctors who strive to keep dancers in a state of physical repair are plainly not. Interesting as it might be to pause for a discussion of this distinction, and how to make it, we must stick to more central matters. And highly central are two artistic activities

in particular, for they enable us to define “artwork,” which can be used in turn to define the other artistic activities. The first of these central activities is art‐making or art‐ creating  –  what I shall call, very broadly, art‐production. Again, tempting problems turn up, but we have to set them aside: they concern the concept of art‐creation and the difference between making and creating. I use “production” generously, trying to slip around these puzzles: it includes making, altering, assembling, joining, arranging, and other distinguishable actions, including certain kinds of doing (as in dancing). What is produced, I think, is always something physical (an object or event) and perceptual, in that it has some properties that can be perceived. Often what is produced also has properties that are not perceptual or physical  –  such as meanings, messages, the capacity to evoke emotions or images, etc.Thus even an alleged work of conceptual art – one that is no more than a closed art gallery with a sign on it saying that the artwork being exhibited that week is just the closed art gallery itself – is a product in my sense (though whether it is an artwork is a question of the sort I expect my definition to clarify and help resolve). Since art‐production is a species of production in general, if we are going to distinguish it from the rest, we shall need to specify a differentia. One plausible candidate for such a differentia, is mode of production. But this is barred to us, by the fact that the same process of production, as we said, may result in an object that is both an artwork and a religious object. Another plausible candidate is intention: that it is the presence of a certain kind of intention that makes production, art‐production. This suggestion is immune to the counter‐argument just given, for something can be produced with more than one intention – say, both an artistic and a religious intention. Another plausible candidate is result, that it is the achievement of a product of a certain kind that makes the production artistic. But this seems to lead to saying that production is artistic when it results in an artwork; and I do not want to adopt this definition since I want to be able to define “artwork” (partly) in terms of its production. Thus we seem to be beckoned toward adopting a definition of “art‐production” that will make use of the concept of intention. But to characterize the specific type of intention needed we must turn to the second central artistic activity, which I have (for want of a better word) called reception. Reception comprises a variety of activities engaged in, in the presence of – or perhaps in response to reproductions or reports or memories of – sculpture, oral performances of literature, films, operas, etc. We view, listen to, contemplate, apprehend, watch, read, think about,



an aesthetic definition of art

peruse, and so forth. Sometimes in this receptive interaction we find that our experience (including all that we are aware of: perceptions, feelings, emotions, impulses, desires, beliefs, thoughts) is lifted in a certain way that is hard to describe and especially to summarize: it takes on a sense of freedom from concern about matters outside the thing received, an intense affect that is nevertheless detached from practical ends, the exhilarating sense of exercising powers of discovery, integration of the self and its experiences.When experience has some or all of these properties, I say it has an aesthetic character, or is, for short, aesthetic experience. Much more, of course, ought really to be said to fill out this mere sketch; and much more could be said, were room available. To say no more leaves, I am aware, a soft place in my definitional scheme, but if it proves to be the softest place, I shall have to be content. When we voluntarily receive those things that are the result of art‐production, we often do so with the intention of obtaining aesthetic experience – in other words, we have an aesthetic interest in those things. (We can have an aesthetic interest in other things as well – such as products of nature or technology.) I use the term “interest” in its fuller sense, which includes two ideas: (1) we take an interest in something, when approaching it and interacting with it, and we take an interest in the aesthetic character of the experience we hope to gain from it; but also (2) we have an interest in that experience, in the sense that it is in our interest to obtain it, since it is worthwhile. I do not defend this claim here, but I believe myself to be justified in making the assumption that aesthetic experience is desirable, has value, satisfies a genuine human interest. We have now the makings of the definition I am after, and I propose that: An artwork is something produced with the intention of ­g iving it the capacity to satisfy the aesthetic interest.

Admirably terse, but therefore in need of exegesis, which I hope will guard it against misreading. It is an aesthetic definition of art, that is, a definition using the concept of the aesthetic, and though it is not the only possible aesthetic definition, it is the one I shall henceforth refer to. To begin at the end, I have to assume that the concept of aesthetic interest is good enough for present purposes, even though it is defined in terms of aesthetic experience, which is the subject of lively and unabated controversy. It is important that the word “capacity” appears where it does: there are of course intimate cases of artistic activity in which the artist (i.e., artwork‐producer) directly aims to

25

provide aesthetic experience to some particular person or group of persons; but I say that it is enough if (as is more widely the case) he or she intends to fashion something that aesthetic experience can be obtained from. The artist may have no idea who, if anyone, will be able to obtain aesthetic experience from it, and maybe no one will (for example, if it is immediately destroyed). Appropriate interaction with the artwork may be extremely demanding or may depend on rare talents and extensive knowledge. The artist may be content to put his painting away or paint it over, once he is satisfied with what he has done – that is, believes that if someone with the requisite qualifications were to take an aesthetic interest in it, that interest would be satisfied to some degree. “Satisfaction,” it should be noted, allows of degrees; it does not mean complete satisfaction. The artist who works solely for his or her own enjoyment is an extreme case, no doubt; art‐production is normally a social activity in which the producer and receiver are different persons.The impulse to make something capable of satisfying the aesthetic interest and to share it must be a very elemental one in very many societies, and must be present before there is a further desire to establish a continuing group (an institution) with an explicitly acknowledged group aim to foster artistic activity. In time there will emerge a tradition, a conviction that there are approved and disapproved ways, or better and worse ways, of carrying on the activity. But I can’t see how tradition is essential to it, and I see no good reason to withhold the label “artwork” from art‐production undertaken before or independently of a tradition. Artistic innovations involve rejection of at least some elements of a tradition, a striking out on one’s own, but they must be counted within art as much as less adventurous works. When I use the word “intention,” I mean a combination of desire and belief: intending to produce a work capable of satisfying the aesthetic interest involves both (a) desiring to produce such a work and (b) believing that one will produce, or is in the process of producing, such a work. So normally the artist has a serious purpose: there is something he wants to do in producing what he produces; and he will have reason to believe that success is possible. If, for example, it is unlikely that a painter could believe that by closing the art gallery and placing a sign on it he would produce something capable of satisfying the aesthetic interest, then it is unlikely that in closing the art gallery and placing a sign on it he was producing an artwork. We must of course allow for far‐out beliefs, however unreasonable; there are painters quite capable of believing that closing the art ­gallery and placing a sign on it might provide aesthetic experience to someone who came to the gallery, found the

26

monroe c. beardsley

sign, and meditated on the symbolic significance of this higher order of art that is so self‐effacing, so sublimely self‐ sacrificial that it denies itself its own existence (like Keats’s unheard melodies that are sweeter than the heard ones). If this was indeed an intention the painter had in closing the gallery and placing the sign, then I am prepared to classify the closed and labeled gallery an artwork. There is, of course, no implication here about the degree of success; it is enough that something was intended. This places the painter’s activity in a significant class of human activities, the artistic ones; it decides the kind of social enterprise he is engaged in, and it determines the primary (not the only) way in which his product is to be judged. The example points up a problem in the application of this definition, though not, in my opinion, a fatal one. To identify the artistic activities and the artworks of a society we must make correct inferences about intentions. And intentions, being private, are difficult to know. But artistic activities are no different in this respect from all other significant activities of a society; if the anthropologist cannot understand what the observable behavior means to the people so behaving, what their desires and beliefs, purposes and motives, are, then he does not understand their culture. We must make use of available verbal testimony, but inferences can legitimately reach beyond that. Once we discover that people in a given society have the idea of satisfying an aesthetic interest, and once we know at least some of their ways of satisfying this interest, we can reasonably infer the aesthetic intention (that is, the intention to produce something capable of satisfying the aesthetic interest) from properties of the product. A painting with a religious subject and evident power to move believers to religious devotion may also give evidence of extreme care in the composition, color harmony, subtle variations in light and texture; then we have a good reason to believe that one of the intentions with which the painter worked was the aesthetic one. The fact that a product belongs to a genre that already contains indubitable artworks also counts as evidence of aesthetic intention: hence, for example, some rather inferior statues of the Buddha sometimes find their way into art museums. According to the definition I propose, and am defending, the aesthetic intention need not be the only one, or even the dominant one; it must have been present and at least to some degree effective – that is, it played a causal or explanatory role with respect to some features of the work. Again, even if we know that a Chinookan story, such as “Seal and Her Younger Brother Lived There,” is told mainly to children to teach them a lesson (and so has primarily a pedagogical intention), the presence of aesthetically satisfying formal fea-

tures and its success in satisfying the aesthetic interest is enough to stamp it as a work of literary art. It is sometimes said that Paleolithic cave‐drawings were not produced with any aesthetic intention at all, or aesthetically enjoyed, because they had a magical or religious function. One argument seems to be that no one in the culture could have had an aesthetic interest because the culture left no pure artworks (that is, works produced solely with an aesthetic intention). My view is that we know far too little about what was going on in the minds of Paleolithic people to be at all dogmatic about this, or to use their drawings as a counter‐example to the aesthetic definition of art. Moreover – although this is a matter of interesting dispute, too – it seems highly probable to me that early human beings developed a capacity for aesthetic experience and a relish for it before they deliberately fashioned objects or actions for the purpose of providing aesthetic experience – just as they must have learned to use fire they found before they learned to make fires, and used found rocks as tools before they shaped rocks into better tools, and caves as housing before they built houses. It is easy to point out vaguenesses in my definition; borderline cases can be found. This is true of most definitions, especially of cultural activities and objects. How pronounced must the aesthetic intention be? (It may be very subordinate to other intentions.) How much activity on the part of the artist is required in order to say that he is producing? (He may merely bring the snowflakes indoors and let them melt as they will.) How much of a sense of integration must an experience have to be aesthetic? (It may involve only a momentary pulling together of pleased attention to a particularly pregnant pothole.) There will inevitably be things we can cover by the definition if we are charitable but exclude if we want to be severe. Fuller explanation might tighten up the definition in some respects, but a rigorous line of demarcation is not to be expected – or even desired, perhaps, except by the legislator or administrator faced with practical decisions, and even he may have to decide some cases a bit arbitrarily, thinking mainly of precedents that will be set and untoward legal consequences to be avoided. Of course, if my definition is deemed hopelessly vague, so that it marks no useful distinction at all, it must be rejected; but surely, even as it stands, it clearly and decisively admits a large number of things and rules out a large number of other things. I would incline toward generosity and a welcoming attitude toward novelty  –  but I would look for evidence of some aesthetic intention, and I see no reason to twist my definition to make room for something like, say Edward T. Cone’s one hundred metronomes running down with nobody silly



an aesthetic definition of art

enough to wait around for them  –  even if this “musical composition” is titled “Poème symphonique.” There will always be room for debate about such distinctions as that between producing art and kidding art – and perhaps the closed art gallery is a case in point. Kidding is a kind of message, and when done well certainly does not disqualify the kidding object from being itself an artwork; the enjoyment of wit can be an aesthetic experience. But common sense should not be abandoned along with philosophical acumen in these matters. The fuss that has been made about Duchamp’s Fountain has long amazed me. It does not seem that in submitting that object to the art show and getting it more or less hidden from view, Duchamp or anyone else thought of it either as art or as having an aesthetic capacity. He did not establish a new meaning of “artwork,” nor did he really inaugurate a tradition that led to the acceptance of plumbing figures (or other “readymades”) as artworks today. If there was a point, it was surely to prove to the jury that even their tolerance had limits, and that they would not accept anything  –  at least gracefully. This small point was made effectively, but the episode doesn’t seem to me to provide the slightest reason to regard the aesthetic definition as inadequate. Many objects exhibited today by the avant‐garde evidently do make comments of some kind on art itself, but these objects may or may not be artworks. To classify them as artworks just because they make comments on art would be to classify a lot of dull and sometimes unintelligible magazine articles and newspaper reviews as artworks, and where is the advantage of that? To classify them as artworks just because they are exhibited is, to my mind, intellectually spineless, and results in classifying the exhibits at commercial expositions, science museums, stamp clubs, and World’s Fairs as artworks.Where is the advantage of that? To classify them as artworks just because they are called art by those who are called artists because they make things they call art is not to classify at all, but to think in circles. Perhaps these objects deserve a special name, but not the name of art.The distinction between objects that do and those that do not enter into artistic activities by reason of their connection with the aesthetic interest is still vital to preserve, and no other word than “art” is as suitable to make it.

III It remains to consider some of the reasonable objections that may occur to the reader – partly to dispose (if I can) of the more obvious ones, thus encouraging potential critics

27

of the aesthetic definition to concentrate on raising more surprising and far‐reaching ones, and partly to suggest some guidelines for constructing alternative proposals to replace the definition with a better one. My method will be to consider briefly some propositions that may be suspected of being untoward consequences of the definition; some of them, I argue, do not really follow from it and the others are quite acceptable. 1. If we take artworks as I have characterized them, then it is possible for children, even quite young children, to create artworks. This consequence is unacceptable to philosophers whose definitions assign to artworks an esssential dependence on institutions, traditions, or aesthetic theories, for we may suppose that seven‐year‐ olds, say, when they draw pictures, write poems, or make up songs, do not yet participate in what is called the “artworld.” But it seems to me invidious to deny children this capability, especially on rather a priori grounds. The basic social activity, as I see it, is one in which a person produces something that he or she finds aesthetically enjoyable and shares it with others who may be able to appreciate it. Once this concept arises, we can have artworks, however simple and crude they may be. This does not mean making pictures of Thanksgiving turkeys according to the directions of a teacher, and seeking commendation for doing the assignment right. But developmental psychologists who have studied the “artful scribbles” of children (notably Howard Gardner) discern a frequent ability to use pictorial symbol systems for aesthetic purposes. 2. The question arises whether, on the aesthetic definition, forgeries of artworks can themselves be artworks. My answer is that some are and some are not: it all depends on the effective presence of an aesthetic intention. To begin with one extreme, a mechanical reproduction of an artwork (such as a picture‐postcard sold in an art‐museum shop) should probably not be classified as an artwork, even though it may be capable of affording some aesthetic satisfaction. At the other extreme, we have free copies (such as a Reubens copy of a Titian) that are plainly painted with an aesthetic intention. But of course neither of these is a forgery, strictly speaking, because there is no intent to deceive, no misrepresentation of producership. Now we can imagine a person making an exact copy of a painting by Chirico for the purpose (that is, with the intention) of passing it off as a genuine Chirico; and we can

28

monroe c. beardsley

imagine that his method of working is tediously atomistic: he uses instruments to make sure that each small area matches the original perfectly. In that case he might execute the forgery without any aesthetic intention at all. This is hard to believe; it is far more likely that he will paint with an eye to capturing the peculiar quality of the empty, lonely, ominous space in the Chirico in order to make sure the forgery is good enough as a painting to fool a connoisseur. In that case the forger is producing an artwork. This conclusion will strike art historians as a reductio ad absurdum of the proposed definition, for they make no room in their history for forgeries. I agree that forgeries do not belong to the history of art, any more than any other copies or reproductions, because they have no significance for the development of art, But I do not see why a careful copyist, though thoroughly unoriginal, cannot be producing an artwork; and I do not see how the addition of another intention (to deceive someone) makes the product less an artwork – especially since the very success of the deception may depend on the painter’s having the intention to make his work as capable as the original one of satisfying the aesthetic interest. 3. It follows from my definition that once an artwork, always an artwork. Anyone who holds that something can become an artwork and cease to be an artwork will object to my definition, and be inclined to adopt an institutional definition. It is sometimes said that, for example, a spinning wheel or snow shovel may begin life as a nonartwork, tied down to its lowly function; that at some point the artworld (prodded by an enterprising museum curator or by a maverick painter like Duchamp) may open its arms to embrace it, placing it on display and thus converting it into an artwork; and that at a still later time this designation or status may be withdrawn, so that the spinning wheel or snow shovel reverts to its original state. If some versions of the institutional theory are right, and there is an implicit or explicit performative act by which nonart can be given art‐status, then it seems the process should be reversible by a converse act, as marriages can be annulled, names changed, contracts declared void, cabinet appointments rescinded, votes overturned, priests defrocked, and so forth. But the notion of taking away an object’s property of being art is prima facie puzzling, and this puzzlingness is accounted for on the aesthetic definition. Something that is not an artwork

can of course be exhibited, can become the object of aesthetic interest, can fall from fashion as people lose aesthetic interest in it. But since what makes art, art is an intention with which it was produced, nothing can be art that was not art from the start, and nothing that is art can cease to be art. And this seems to me a more intelligible and less misleading way to talk than to say, for example, that the Victorian paintings now stored in the art‐museum’s basement have ceased to be art, just because they are no longer much admired. 4. If it is the intention that counts in art‐production, does it follow that one cannot fail to produce an artwork, if one intends to? That might seem an unfortunate consequence of the proposed definition, if it is a consequence. To straighten out the difficulty here requires some careful attention to certain features of actions and intentions. I do not say that the aesthetic intention is the intention to produce an artwork, but the intention to produce something capable of satisfying the aesthetic interest.To intend to produce an artwork is to intend to produce something produced with the intention of producing something capable of satisfying the aesthetic interest; and this second‐order intention, though it does occur, is not the usual one, I think.Thus in a certain rather innocuous sense artworks are often produced unintentionally – that is, without the intention to produce an artwork. But of course an artwork cannot be produced without any intention and in this sense there can be no unintentional artworks. Paint may be spilled and pottery cracked unintentionally, and the pattern of paint or of the cracks may be capable of satisfying the aesthetic interest; but that alone does not make an artwork. One can fail in various ways in trying to produce an artwork: the work may not be as original as was hoped (the box of Kitty‐Litter submitted to the avant‐garde show turns out to be the second one submitted), or the work may not be the kind of work hoped for (the poem is not nearly as ironic as had been expected). But as far as I can see, the only way one can fail to produce an artwork after setting out to produce something capable of satisfying the aesthetic interest is by failing to make the physical object one tries to make (the clay is defective or the temperature of the kiln not right, so the pot falls apart), or to do the deed one tries to do (the dancer slips and falls instead of completing the pirouette). As long as something is produced with the aesthetic intention, an artwork is produced.



an aesthetic definition of art

5. It is a consequence of my aesthetic definition, then, that tawdry and negligible objects will be classified as artworks. In such a sorting there is no implication of worth or value. Admittedly, this insistence on preserving a value‐neutral sense of “artwork,” for the purposes sketched at the start of this essay, runs counter to a familiar use of “work of art” for artistic praise. But we need a term for the classification; indeed, without it we will not even be able to make sense of the evaluative expression “a good artwork.” And there are other terms to do the job of evaluating  –  indeed, a large number, because we call artworks not only “good” but many other things, both kind and unkind. Before the anthropologist could hope to discover which objects in a society are its “artworks” in the normative sense (that is, what I would call its good or great artworks), it seems that he or she will first have to gain some notion of the class of things in reference to which some of its members are judged superior or supreme. And the anthropologist (since he is not necessarily an art historian) may be as interested in the failed products of a society’s artistic activities as in its successes – especially if he can come to understand the causes and the consequences of these failures. Moreover, when we take an anthropological interest in our own society, or others of a comparable level of civilization, we have to study a vast and highly significant phenomenon that is generally labeled “popular

29

art.” Part of my aim in providing a definition of art that has no built‐in value‐judgments ranking decorative designs, stories, dances, songs, etc. is to encompass the popular arts as well as those for which we lack a convenient label – I suppose because they have been thought of as the real thing: portraits in oil, epic poems, ballet, lieder, etc. To gather popular and “esoteric” artworks (using the latter term with some diffidence) in the same broad class is not, of course, to deny differences in value, but it is to invite study of continuities and degrees of accessibility, of complexity, of training in taste, of seriousness in affect, and so forth. I don’t see any good reason for not regarding Guys and Dolls and The Pirates of Penzance as artworks, along with Tannhäuser and The Marriage of Figaro. When it comes to other sorts of product now much studied by scholars  –  bed‐quilts, cigarstore Indians, Dixieland jazz, old Tarzan movies – it may strike some people that we have wandered to the edge of art, if not beyond.There will, of course, always be a question whether something was indeed produced with an aesthetic intention, but even when this intention was probably minimal and the skill to carry it out deficient, the social function served may be the same as that of clearly artistic activities, or closely related to it. So a broad definition of art that still retains its essential connection with the aesthetic interest has much to recommend it.

4

“But They Don’t Have Our Concept of Art” Denis Dutton

I In the current discourse on cross‐cultural aesthetics, an oft‐ repeated formula has it that understanding the art of another people may be difficult or impossible because “they have a concept of art different from ours,” or “they don’t have art in our sense.” But what does “the concept of art” denote here? One way to approach this question is to ask what “concept of ” adds to the claim that another culture has “a concept of art different from ours.” If the claim were merely that they have “an art different from ours,” there would be no issue. The added “concept of ” seems to want to extend the claim, as though to say that despite outward appearances to the contrary, the meaning art has for these people differs radically from the meaning art has for us, that we may be mistaken even to call it “art.” Similarly with the claim that their “sense” of “art” is different from ours. Such claims represent a style of thinking that has deeply marked cross‐cultural aesthetics for the last generation. Whether the area of investigation is the artistic life of small‐ scale, non‐literate societies (so‐called tribal or ethnographic arts) or the arts of non‐Western civilizations, such as India, the frequent contention is that the aesthetic forms of these cultures are wholly other, and cannot be understood in terms familiarly applied to the arts of the West. I shall turn to tribal cultures later, but I begin by considering an essay

in the influential collection The Traffic in Culture, edited by George E. Marcus and Fred R. Myers. Lynn M. Hart writes about large decorative paintings on mythological themes made by Hindu women in Uttar Pradesh, sometimes individually, sometimes in groups together, which are part of the activity surrounding marriage celebrations.1 She describes the women artists in their working environment; then the appearance of one such painting in a North American dining room; thence to the exhibition of another of these jyonthi paintings in the Magiciens de la terre show in the Pompidou Center in Paris in 1989. Despite the fact – or perhaps because of it – that jyonthi paintings are straightforward, colorfully stylized depictions of Hindu mythological themes (Ganesh, Laksmi and Vishnu, sun and moon, lovebirds, etc.), Hart insists on using “producer” instead of “artist” and “visual image” instead of “art” to refer to this work (if it is “work”). Hart is determined, she explains, to avoid “inappropriate Western terminology.”This is important, she thinks, because otherwise Westerners might have trouble appreciating that “the images and patterns themselves are based on religion, ritual, and mythic themes and derive their meaning – and their power – from the religious contexts of their production and use.”The indigenous aesthetic principles of this art, or visual image production, are “different from standard Western aesthetics.” The excellence of the works from an indigenous perspective, she explains, “is

Denis Dutton, “But They Don’t Have Our Concept of Art,” in Noël Carroll (ed.), Theories of Art Today (Madison: University of Wisconsin Press, 2000), pp. 217–38. © University of Wisconsin Press.

Aesthetics and the Philosophy of Art – The Analytic Tradition: An Anthology, Second Edition. Edited by Peter Lamarque and Stein Haugom Olsen. © 2019 John Wiley & Sons, Inc. Published 2019 by John Wiley & Sons, Inc.



“but they don’t have our concept of art”

seen to lie in the closeness of the central symbol’s approximation to an ideal image, with special attention paid to the style, technique, and materials used. It is important to re‐ present the symbols used in an adequate way, not to improve upon them, though at the same time the image on the wall should be as beautiful and pleasing as possible” – and so on, all “quite distinct from Western aesthetic canons” (131). In point of fact, there is nothing in Hart’s descriptions that is distinct from Western canons and concepts of art, which variously include in many Western genres and historic epochs the colorful approximation of images from religious mythology, produced with attention to style, technique, and materials. The conservatism of jyonthi painting, its prohibition on “improving” on the traditional iconography, may not characterize the Western avant‐garde, but it is a feature found through much of the history of European art in the Middle Ages, as well as traditional religious folk arts and women’s arts of Europe for the three centuries prior to the present [twentieth] one. The theology might be different, but there’s not one thing Hart describes that can’t be found in “Western aesthetics.” This last point is worth dwelling on, for it seems to me that often when it is said that some other culture has a “different concept of art” from ours, there is implicit in the claim an extremely circumscribed and historically specific definition of the art denoted as “ours.” Hart has made no effort to probe the history and traditions of “our” art to see if analogues or similarities might exist for the Uttar Pradesh example. Hart’s claim that jyonthi painting cannot be understood by applying to it categories or concepts of Western art is in the end either trivial or false. If the claim means that Western painting does not traditionally include elements of Hindu mythology, is not painted on whitewashed mud walls by fluent speakers of Hindi as part of the celebration of marriage rites, then indeed, jyonthi painting is quite beyond Western categories. But Hart wants more than that; she would have us believe that jyonthi painting is not  art “in our sense,” a claim which is demonstrably false.  At one desperate point she attempts to dramatize the cultural difference between jyonthi image producers and European artists: The Western producer of a painting destined (he or she hopes) for the wall of an art gallery and possibly for the wall of a great art museum is conscious of him or herself as “artist” making an object that is contrived, posed, set apart from everyday life, just as the short stories and novels of contemporary fiction are  contrived, posed, and separate from everyday life. These products proclaim, “Look at me, I’m art!” The producer of the

31

ritual images in a Hindu village is not conscious of herself in this particular way. She is producing an image that derives its meaning from the part it plays in life, rather than as a contrived, posed object. (144)

While I would challenge the adequacy, indeed the competence, of this as a description of Western art, it is at least clear that Hart is comparing two very different categories of activity. On the one hand, the ambitious Western artist operating in a professional market of agents, dealer galleries, and museums; against this familiar image she pits Indian women who decorate the walls of their houses with conventionalized religious designs as part of making a special occasion of a wedding. Hart says that beyond the careful, conscious use of aesthetic judgment in producing the paintings, there is a further human dimension absent from the Western point of comparison: “A woman, a mother, lovingly creates beautiful, emotion‐filled, auspicious, important images for her own children for the purpose of helping them, of supporting them so they can succeed and be happy in the next stage of their lives.” Hart has chosen a false comparison. In fact, the history of the West is replete with countless mothers and prospective mothers‐in‐law who have labored at embroidery, knitting, and sewing, “producing” beautiful artifacts for their children’s weddings, either as part of a trousseau or as decorative elements (e.g., decorated cakes) for the wedding day. These beautiful – or beautified – objects can be as lovingly created by European as by Indian women. Much of this output is cloth or fiber art, but it also would include decorated ceramics and items of household furniture. Some of these objects would embody religious themes. Why has Hart failed to mention comparable Western traditions of dowry or trousseau arts to place in relation to the jyonthi paintings of Uttar Pradesh? It is because she is guilty of the very ethnocentrism she accuses others of. She studies a genre of folk art in one culture and, seeing that it is a type of painting, looks within Western culture to discover an analogue. Her mistake is to imagine that the comparison will be painting in Western culture. But if you want a comparison for a jyonthi painting, it is absurd to look at, say, a Diebenkorn hanging in a Western gallery. Jyonthi paintings belong with domestic and dowry arts of cultures worldwide, from beautifully woven Maori feather cloaks for infants to embroidered samplers to knitted blankets and painted cradles in European folk traditions. Elsewhere in her essay, Hart complains of the West’s tendency to place a greater value on High Art traditions than on craft traditions. In fact, Hart does exactly that herself: she is so

32

denis dutton

impressed by these Indian forms as painting that she fails to acknowledge the women’s craft traditions associated with marriage celebrations and trousseaus in her own culture. They too can involve the loving and devoted exercise of skill and aesthetic judgment, and produce objects to help celebrate auspicious occasions and contribute to the success of the next generation.

II Hart’s analysis is an example of a widespread tendency to try to exaggerate cultural difference far beyond reality, to try to make a foreign art form seem more alien than it actually is. In her case the strategy entails taking what should seem to us a familiar art form and estranging it by finding an inappropriate practice in our own culture with which to compare it. Here is another strategy for achieving a similar end: to describe an art form strictly in terms of one of its aspects, omitting to mention features which would render it comprehensible to a Western audience, and from this drawing large conclusions about non‐Western or tribal arts. In an essay entitled “The Technology of Enchantment and the Enchantment of Technology,” Alfred Gell defends a general thesis about ethnographic arts in terms of the analysis of a single example, Trobriand Island canoe decoration.2 Gell begins by noting anthropology’s general lack of regard for art, but he says this should be expected: social anthropology ought by its very nature to be anti‐art. The aesthetic awe afforded by objects in the Museum of Mankind demonstrates what “is an unredeemably ethnocentric attitude, however laudable in other respects.” Gell argues that the anthropological study of art should be carried out under the assumption of a “methodological philistinism” analogous to the “methodological atheism” required of the study of religion.“I would suggest that the study of aesthetics is to the domain of art as the study of theology is to the domain of religion.” Just as anthropologists of religion must set aside their religious predilections, so anthropologists of art must ignore the aesthetic attractiveness of the objects and practices they study  –  the anthropology of art requires “a complete break with aesthetics” (42). With this in mind, Gell invites us to consider the arts as components of “a vast and often unrecognized technical system, essential to the reproduction of human societies ….” He proposes that art be thus understood as a “technology of enchantment,” where enchantment is seen not as peculiar only to art, but as a potentiality “immanent in all kinds of technical activity.”

This potentiality is essentially magical, and Gell uses as his central example the stunning prow configurations of Trobriand Island canoes that are used for Kula expeditions. With their bright red and white paint and intricate carving sometimes resembling a mushroom, or recalling the scroll‐ like appendages of an Ionic capital, the prows are designed to dazzle and disorient the spectator, giving a possible trading advantage to the party which arrives in such a decorated canoe. So much is uncontroversial; one thinks of not only the psychological warfare of Kula transactions, but combat equipment, such as the fighting shields of Sepik and Highlands warriors of New Guinea, which often display horrific faces designed to frighten an enemy. It is not the bold effect of such work that impresses Gell, however. Instead, he emphasizes what he calls “the halo effect of technical difficulty” in Trobriand art. As a child, Gell tells us, he was deeply impressed by a matchstick model of Salisbury Cathedral: “from a small boy’s point of view it was the ultimate work of art, much more entrancing in fact than the cathedral itself ….” He draws from this a very large conclusion about the reaction of all of us to works of art: “I am impressed by works of art in the extent to which I have difficulty … in mentally encompassing their coming‐into‐ being as objects in the world accessible to me by a technical process which, since it transcends my understanding, I am forced to construe as magical” (49). Works of art become objects of mystery and fascination by virtue of their incomprehensibility as technical feats. Gell attempts to reinforce this view by referring to J.F. Peto’s 1894 trompe‐l’oeil painting popularly known as Old Scraps, a highly realistic still life of letters, paper scraps, drawing pins, and faded ribbons tacked to old board. The fascination of this work, he claims, is that its audience cannot comprehend how mere paint could be used to create such a realistic representation. This “technical magic” gives the painting its prestige and value (a value no similar photograph could attain). Moreover, the meaning of Old Scraps in our aesthetic lives has analogies in the art of small‐scale traditional societies. In the case of a Trobriand canoe splashboard, “it is very difficult to acquire the art of transforming the root‐buttress of an iron‐wood tree, using the rather limited tools which the Trobrianders have at their disposal, into such a smooth and finished product” (54). Magic is the ideal technology of such societies as the Trobrianders’; it enables one to accomplish a task instantly and effortlessly – rather than with uncertainty and effort. Art also exhibits technological mastery; hence Gell argues that there is a “convergence” between the aims of ordinary technology, magic, and art – the last two being enchanted



“but they don’t have our concept of art”

versions of the first. Like conjurers, artists who defy ordinary technical understanding are given the ambiguous status of being “half‐technician and half‐mystagogue.” While this puts artists at a disadvantage in modern market societies, Gell claims, it gives them a special status in traditional societies such as the Trobriands’. Gell concludes with a description of Trobriand horticultural magic.The Trobriand garden is “a system of technical knowledge and at the same time a work of art, which produces yams by magic.” The technology of enchantment is manifest in garden layout and poetry: “Just as when, confronted with some masterpiece, we are fascinated because we are essentially at a loss to explain how such an object comes to exist in the world, the litanies of the garden magician express the fascination of the Trobrianders with the efficacy of their actual technology which, converging towards the magical ideal, adumbrates this ideal in the real world” (62). There is no doubt Gell’s argument throws light on how some forms of ethnographic art might profitably be understood. However, the connecting of art with magic is plausible only so long as he attends to the general awe felt by audiences; but to appropriate in general the logic of artistic technique to magical technology is wrong. Considering Western art alone, the claim does not stand up. While there are many works of art which fascinate audiences as technical display (Peto’s painting is a perfect example), and while technique is for many people virtually the only criterion for artistic value (hence the familiar abuse of modernism: “My kid could do that!”), technical excellence is not the main reason most European audiences are interested in art. At the present time the most popular art period, judged by print and art book sales and exhibition attendances, is French Impressionism, which is not a historical school particularly marked by technical display. Gell mentions that Rembrandt is admired for technical skill, but so are many other seventeenth‐century painters who are rated much lower as artists. That even Gell is uncomfortable with this position is indicated by his strained attempts to expand his conception of technique to include Picasso’s bronze baboon whose head is a toy car (1955), and even Duchamp’s Fountain (1917). This strategy only compounds Gell’s error: Baboon and Young is a humorous but technically uninteresting piece, while Fountain is in part a direct attack on the very idea of technique in art.The point of Duchamp’s readymades is that they are ordinary objects to which the artist does nothing except to present them in a gallery. If Gell can include this particular piece of plumbing in the class of technically accomplished works, then he has expanded the definition of “technique” to encompass witty or original gestures which

33

involve no making‐skill on the part of the artist. This is not a load the term can intelligibly carry; Fountain is a famous work, but not because it exhibits extraordinary technique. Rather, it purports to show that it is possible for an object to be a work of art while demonstrating no technique whatsoever. The inclusion of Picasso and Duchamp here is not a minor confusion on Gell’s part, but in fact undermines his whole effort to show that admiration for technical mastery – and with it a sense of magic and enchantment – is the central component of the aesthetic response. Turning specifically to ethnographic arts, we encounter further uncertainties. Technical skill is perhaps more obviously admired in, say, Oceanic art traditions than in many European modernist exercises, but not always. Virtuoso carving, such as is seen on Trobriand splashboards, fits Gell’s case very well, as would much Maori and Polynesian carving. But consider Sepik: in northern New Guinea wild expressiveness, rather than elaborate finish or virtuoso facility, is frequently the criterion of aesthetic excellence and cultural power. The same could be said of New Guinea Highlands shields, which are powerful as works of art through bold visual impact rather than noticeable technical accomplishment. Moreover, while some cultures treat artists as a virtual priestly class, as possessors of special magical/aesthetic knowledge, others do not. This suggests another consideration contrary to Gell. He stresses that we are amazed, wondering of the artwork, “How was it done?” True indeed, especially with well‐developed European technical traditions: as a sometime pianist, I have trouble conceiving what it takes to perform Liszt’s Transcendental Etudes the way Georgy Cziffra does, and realistic painting in the style of J.F. Peto is impressive in a similar manner. But many small‐scale traditional societies, lacking either the vast population or the extreme specialization of art‐technical labor that makes possible the emergence of the likes of Cziffra or Peto, treat the artist not as a master of a kind of technical magic, but as a trained craftsman performing tasks anyone could learn. In the end, despite his universalizing ambitions, Gell fails to establish an acceptable way of looking across the whole range of art in traditional societies, falling into a kind of Trobriand localism.

III I turn next to Susan M.Vogel, whose writing on tribal arts displays an eloquence and intellectual sophistication considerably exceeding that of the previous authors.Yet in her

34

denis dutton

wonderful book Baule: African Arts,Western Eyes, she begins with statements reflecting a point of view similar to theirs: This book is inspired by my enjoyment of certain objects of Baule material culture as works of art in a Western sense, but it seeks to explore what “artworks” mean in Baule thinking and in individual Baule lives. For almost a century, Baule art has been recognized in Europe (and later in America) as one of Africa’s most significant sculpture traditions. Although Baule art is important in the Western view of African art, the people who made and used these objects do not conceive them as “art,” and may equate even the finest sculptures with mundane things, devoid of any visual interest, that have the same function and meaning …. “Art” in our sense does not exist in Baule villages, or if it does villagers might point to modern house decorations, rather than famous traditional sculptures still made and used in villages and evoked by the term “African art.”3

Her support for this contention includes the following observations, among others. First, the Baule will “merge and equate” (a) spirits and unseen powers, (b) ordinary physical objects in which they dwell, such as a lump of clay, and (c) superb sculptures which they may also inhabit. However, only the last are works of art in the Western sense. Second, the Baule “attribute great powers to their artworks – powers that Western culture would mainly relegate to the realm of superstition.… Enormous powers of life and death are integral parts of the sculptures we admire in museums, and Baule people do not consider them apart from those powers.” Third, and especially emphasized by Vogel, many of the most important artworks of the Baule are not meant to be seen by large audiences, or by just anybody, but are normally hidden from view, “kept in shuttered or windowless rooms that few people enter” or wrapped in cloth and taken out only infrequently. This sharply contrasts with the Western ethos of aesthetic objects which invite “intense, exalted looking” from a large audience (83). Looking itself is for the Baule a privileged and risky act, as the very sight of a sculpture can be fatal for the wrong person. This in turn has to do with the special place held by sight in Baule culture, where “seeing something is potentially more significant, more dangerous and contaminating, than touching or ingesting something” (110). (Thus, Vogel says, a woman inadvertently seeing a sacred men’s mask might die from the event, whereas a blind woman who laid her hand on it but didn’t realize what she was touching would not necessarily be so threatened; men might find the sight of a woman’s genitals fatal.) Do such considerations as these support the view that the Baule have a different conception of art from the West,

that “art” in our sense cannot be found in Baule villages? No, they do not, as Vogel’s subsequent account makes abundantly, repeatedly clear. She begins her account by describing masks and figure sculptures that have profound spiritual and intense personal significance to the Baule. These include personal portrait masks and so‐called spirit spouses. Among those pieces, spiritual, magical, or personal aspects certainly loom larger in the minds of their owners than their aesthetic qualities, a fact which Western observers must take into account. But the relationship of an art genre to a spiritual world is a consideration that applies to the arts of Western culture as well. Thus a majority of believers whose religious sentiments were inspired by Giotto’s frescos at Padua might have been just as moved by similar frescos which did not approach Giotto’s artistry; in other words, the original audience might have possessed little or no appreciation of the comparative artistic value, let alone historical importance, of Giotto’s frescos, and would have been responding to them as religious narratives. Part of understanding the cultural importance of Giotto for his original audience and its local descendants is grasping the place of his work in a specific economy of religious thought, and religion, though often intermingled with art, need not be confused with it.That acknowledged, it is perfectly valid for an art historian to discuss the aspects of Giotto’s work which form part of art history – technique, formal excellence, modes of representation  –  rather than religious or social history. Nor are the aesthetic qualities of Giotto’s paintings and frescos accidental by‐products of religion, however closely tied to religion that art may be.Their status as works of art is not threatened by their having been treated by most of their audience as mere biblical illustrations, or as colorful backdrops, barely to be noticed, for religious ceremonies. But even taking into account the privacy and magical properties of Baule spirit carvings – or at least many of the ones most prized by Europeans – they are nevertheless subject by the Baule themselves to the same kinds of aesthetic characterizations applied to art carvings elsewhere. In fact, aesthetic appreciation of Baule carving is,Vogel admits, one of the points of agreement between Baule people and Western connoisseurs: “Baule artists, and the individual owners of objects, certainly sometimes enjoy the beauty of these objects and the skill it took to produce them” (29). Following Herbert Cole, Vogel says that Baule language points away from the “thingness” of art as noun, and emphasizes adverbial forms applied to carvings elegantly made to enhance, embellish, or empower in experience. The nounish sense of the English notion of “art” is not



“but they don’t have our concept of art”

entirely appropriate in the Baule context, where adjectives and adverbs relevant to artistic experience are used as modifiers attached to personal life, moral and physical struggles, and, Vogel says, “the drabness of daily existence” (292). Nevertheless, Baule will refer to outstanding sculptures in Baule equivalents of sweet, pleasing, beautiful, and good. A common phrase is to praise something or someone as “beautiful as a statue” – recalling the English “pretty as a picture.” Conversely, English has no hesitation to apply aesthetic modifiers to nonmaterial objects of appreciation: dances and musical performances, for example. Nor can a vast cultural gap be made of the fact that some of the spirit carvings are neither well nor often seen. As Vogel acknowledges in a note, “many works of European art (ceiling frescoes, books of hours, hinged altarpieces) and numerous objects from other traditions (Japanese netsuke, Egyptian and other tomb furnishings, Chinese scrolls, Russian icons) were created in the full knowledge that they would be seen in low light, partially or at a distance, or only rarely, or privately by only a few people.” Moreover, beyond the personal and highly charged artworks which dominate the first half of Vogel’s book, Vogel explains in a separate chapter that the Baule have a voluminous, purely secular decorative art. This includes doors, gold weights, stools, fans, combs, gong mallets, beautifully carved weaver’s pulleys, and other decorated utilitarian objects. Because these sculpted artifacts are sold on their visual appeal, rather than being privately commissioned and kept out of view, they are very often of better technical quality than the more deeply important spiritual carvings. Their aesthetic quality also serves to advertise the skill of their makers, many of whom specialize in specific kinds of domestic object, such as ointment pots. Although increasingly replaced by machine‐made objects today, they were, Vogel explains, “once very common, satisfying the basic desire for a pleasing, aestheticized environment” (270). Through much of her discussion, Vogel is attempting to defamiliarize Baule art in the minds of her Western readers – requiring them to stop and think about the presuppositions they may bring to any appearance of the word “art” – in order that they might see the Baule objects as the magical and spiritual objects they are in the minds of many Baule people. In itself, this demand for a certain kind of “unlearning” of cultural habits is entirely laudable: it vastly extends the Western reader’s understanding and appreciation – and, by the way, it is a strategy that could with profit be more often applied to Giotto as well. But it is a strategy that can encourage the false notion that the Baule do not have works of art and that we are ethnocentrically mistaken

35

in calling their works “art.” In fact, Vogel does not believe this herself, which is why, having tried to establish the strangeness of the Baule approach to art, she turns around near the end of her book to assure readers of its familiarity: “Nothing described in this book is completely unique to the Baule. In fact, the greatest interest of a tightly focused art study like this one may lie precisely in how much light it can shed on the place of art in other, distant cultures.”

IV How much different from a familiar practice in our culture must an alien practice, x, be in order to merit the designation,“They have a different concept of x from ours”? There is one extreme answer to this question, held earnestly and systematically to my knowledge by no ethnographers, though it is often hinted at or suggested informally: it is that version of cultural relativism (sometimes called contextualism) which claims that since the meaning of any concept is constituted by the other concepts and cultural forms in which it is embedded, concepts can never be intelligibly compared cross‐culturally. As every cultural system/context is different from every other, it follows therefore that any item within a system is strictly incomparable to any item in another system. Although counter to ordinary cross‐cultural experience, the kind of thinking suggested by this incommensurability thesis  –  a rhetoric of cultural uniqueness  –  is attractive to some ethnographers who have specialized in specific cultures: it affords them a privileged standpoint, as they alone possess superior knowledge of the conceptual world of “their” tribe. The cultural interpretations of an ethnographer who knows the local language of a tribe, and has a grasp of the tribe’s web of rarified or esoteric meaning, cannot easily be challenged or criticized by outsiders. And since no concept in any culture could embody exactly the meaning of any concept in any other culture, it follows that the translation of not only poetic language, but any language – along with comparison of political forms or social structures, judicial structures, cooking and eating practices, warfare, and especially works of art – would therefore be impossible. In the actual realm of day‐to‐day ethnography, where comparison and the cross‐cultural application of concepts are constantly practiced, such incommensurability is never actually advocated or viewed as given fact. Nevertheless, ethnographers will occasionally claim that a tribe “does not have our concept” of some practice or other. It is my contention that the notion of “a different concept” is stretched

36

denis dutton

beyond intelligibility in most such contexts, and I have yet to see it used validly in connection with art. In the first place, the claim that a cultural form is unique, or that the concept which denotes it in our culture is useless or inapplicable in another culture, requires that the person making the claim have a firm command of the potentially comparable practices or meanings in Western culture with which the alien meaning might be analogized. This is not a purely theoretical issue, for it suggests a practical line of interrogation which ought to be applied to any ethnographer claiming cultural uniqueness for an alien meaning: Are you confident you know enough about your own culture to make an incomparability claim? This problem is at the core of the essay by Hart: through either ignorance or oversight, she fails to find the proper comparison for jyonthi painting, which is not European High Art gallery painting, but traditional religious folk painting practiced in the context of trousseau arts. Broadly speaking, this is a general deficiency of the anthropology of art. Too often, it has transpired that young anthropologists, possessing limited familiarity with the vast range of arts of Western history, perhaps on their first overseas, let alone ethnographic, experience, set out to explain the subtle and intricate arts of remote tribal cultures. Some anthropologists may achieve descriptive accuracy and aesthetic insight in such an ethnographic exercise; many, however, are simply inadequate to the job. With the Trobriand and Baule examples presented by Gell and Vogel, on the other hand, the issue is different.The magical powers associated with these arts do not commonly find a literal analogue in contemporary Western art practice (though they remind me in some respects of weeping or healing religious statues that periodically appear even today in Europe and the Americas, or outposts of Christianity, such as the Philippines). Nevertheless, we have no trouble appreciating the carving skill and aesthetic characteristics of Trobriand splashboards and Baule spirit spouse sculptures; we can also understand magical technologies (and economic objectives) as described by Gell as well as comprehend, thanks to Vogel, the psychological utility of the notion of the spirit spouse for the Baule. Combining our general ideas of art  –  even our nounish concept of a work of art or art‐sculpture – with these other aspects of a foreign artistic/magical/religious practice is hardly an insurmountable task for the Western intellectual imagination. Vogel in particular paints a lucidly coherent picture of the world of Baule belief and art. Understanding what she says does not require even the slightest stretching or adjustment of “our concept” of art, however much she extends the category of objects we call art.

Consider by way of comparison another human practice, cooking. Suppose there existed a tribe whose only way of cooking food  –  any food, ever  –  was to boil it in water. Everything this people ever prepared and ate was either raw, unheated in any manner, or boiled. Would we say, “They have a different concept of cooking from us”? No; they cook food, though within a more limited repertoire of techniques than ours. But a greater range of techniques to carry out a practice does not in itself change the concept of that practice. The invention of the microwave oven did not change the concept of cooking; it provided a new way to do it. Our grandparents had our concept of cooking, even if they cooked different food and never used microwave ovens. Suppose, however, that we discovered a tribe that never heated food, had never heard of heating it, but always passed a spirit wand over it before eating it. Would we say they had “a different concept of cooking from ours”? Again, no; whatever else they are doing  –  blessing food, sanctifying it, warding off poisons, authorizing the occasion of its being eaten – they are not cooking it with a spirit wand (although the wand could act to “cook” symbolically if they already knew what cooking was; but the tribe would have to have the concept of cooking for that). In parallel fashion, suppose some culture’s concept of art included objects which, although sculpted out of wood, were never looked at with amazement, pleasure, or fascination of any kind, in public or in private (or were expected to be looked at even by nonhuman entities, as by gods), were the subject of no critical vocabulary whatsoever, did not represent anything numerically, were crafted in no discernibly regular style, and although employed as doorstops, were never accorded any attention beyond what was required to place them before open doors or to remove them in order to shut doors, Could we say that this tribe “has a different concept of art from ours”? No; on the evidence so far supplied, whatever else these objects are (doorstops, evidently), they are not “art in a sense different from the Western sense.” They are not works of art at all. In order to qualify as works of art, in whatever attenuated, distant, strange, or obscure sense we might want to capture, the objects would have to share in some of those aspects – sensuous pleasure in experience, created in (or against) a traditional style, involving intense imaginative attention, skillfully made or performed, being symbolic or representative, expressing emotion or feeling, and so forth – that art shares not only in Western culture, but in the great art traditions of Asia and the rest of the world, including tribal cultures of Africa, the Americas, and Oceania, If there is no discernible connection with this established complex of



“but they don’t have our concept of art”

ideas, it is not a new kind of art form; it is rather a category of object or practice distinct from art. Art is not a technical concept (like “endorsing a check”) confined to one culture – ours – and by either patronizing generosity or imperialist ambition extended to others. From a cross‐cultural, transhistorical perspective, art is a vast assemblage of related practices – most probably ephemeral, some resulting in material objects, or recorded as texts – which can be connected in terms of analogues and homologies between all known human societies.The similarities and analogies are not difficult to see in comparing one culture with another, and in fact the anthropological literature leaves no doubt that all cultures have some form of art in a perfectly intelligible Western sense of the term. As Francis Sparshott says, the word “art” gestures vaguely “toward an immense, indeterminate, and disparate body of practice and theory with a dense and much‐studied history.”4

V I note Sparshott’s remark from its appearance in an eassy by David Novitz, in which Novitz interprets Sparshott as wanting to emphasize the extent to which the concept of art is constructed differently by different cultures.5 Our decisions about what is and is not art, Novitz argues, do not derive from some “essential nature of art but from certain historical and social contingencies.” Such identifications on our part would not be “straightforward” or “undemanding” but would require that we understand “the history and theory that pervades a tribal culture” (24). As works of art are “cultural, rather than natural kinds,” the identification of something as a work of art presupposes cultural knowledge, rather than the noting of mere similarities (Novitz remarks on what a mistake it would be to classify Baule spirit spouses as “art” because they resemble Cubist sculpture). Novitz insists that there is no one way that an artifact must be in order to be a work of art; there are shades, degrees, nuances, and subtleties bred of social life, all of which defy straightforward empirical investigation and so cannot be captured in precise formulations and rigorous definitions; still less by appeal to artistic laws or aesthetic universals. Rather, the decision to treat an artifact as a work of art is made in terms of criteria that have much to do with the historically‐shaped life of a society; criteria that are of significance only because of their social location – the beliefs, preferences, values, and social arrangements that prevail within a society at a given time, and which make these features (rather than those) a mark of arthood. (26)

37

It follows for Novitz that we could not identify a work of art as such without first identifying it as belonging to a culture. As for such objects as the twin surrogate carvings of the Yoruba, they are “difficult to identify as works of art in the prevailing sense of the term,” while “it would be at best misleading, at worst inaccurate” to describe Baule spirit spouse carvings as works of art “in any full‐blooded sense of the term”; they are not “works of art in our sense of this word.” What, however, is “our” sense of the word “art”? Novitz does not say precisely, though his Western examples  – standard paintings and sculptures, Van Eyck, Picasso, Michelangelo  –  suggest that for him at least the Western sense of “art,” at least insofar as it pertains to visual artifacts, refers primarily to conventional museum works. This feeling is reinforced by his passing denial that in our culture banknotes, vintage cars, and postage stamps are works of art.At one point he discusses how we might know “whether a tribal artifact is art in our sense of this word,” and then adds, “that is, in the only sense of the word we understand.” This strikes me as a very odd remark. Even if we accept that our sense of the English word art is the only sense we understand, what does that come to? My sense of art, the sense I imagine is shared by most educated contemporary speakers of modern European languages (and certainly not only Europeans), does not refer exclusively to European art, but to all things in human history to which the term might reasonably refer, including art objects and artistic activities of non‐Western cultures and distant historical times – objects, practices, and performances I’ve not experienced yet, but will someday. Similarly, we would intend that our concept of language does not refer to our language alone – for example, English – but to all languages, regardless of whether we can speak them, know what they are called, or even yet know they exist. (This is so even if the first thing we might think of as an example of a language is our own language; asked to imagine a bird, I might well think first of a sparrow or a robin, but still realize that penguins, kiwis, dodos, condors, and ruby‐throated hummingbirds are birds as well.) The “only sense of ‘art’ we understand” cannot be a sense that refers only to art we already know. Art, in European thinking in any event, is an open concept, and like the concepts of religion, government, or sport, art stands ready to cover new instances and incarnations. In a famous remark in Art, Clive Bell says that “either all works of visual art have some common quality, or when we speak of ‘works of art’ we gibber.”6 He meant, of course, that they have in common more than being referred to by the same word – there must be some deep reason why that word is applied to such apparently different objects.

38

denis dutton

This fundamental truth, as Bell realized, has at least as much pertinence in the discourse of cross‐cultural aesthetics as it has for disputes about visual art within Bell’s (and our) culture. I have the impression many of the theorists who have written of “art in our sense” suppose the meaning of the term is a function of its class of referents; even if they might deny it as a bald assertion, they write implicitly as though “our sense” of the term is governed by “our” referents, “the only ones we know.” The two problems suggested by this are, first, that if our sense of “art” were determined by its referents, that sense would therefore be constantly changing, as it is extended daily to refer to objects and performances offered both from within our culture and from beyond it. But, second, how would we even known when to extend the application of “art,” if we didn’t have some principle of application which validates bringing new objects and performances under it? There must be stable elements in its meaning; to deny this entails that we go about arbitrarily calling anything art. Although he does not provide a full‐blooded definition of art to accompany his mention of the “full‐blooded sense of the term,” David Novitz does hint at the existence of at least one fundamental criterion for art, basing his view on a suggestive remark by Monroe Beardsley: “in creating works of art we humanize the earth as we can in no other way, we warm it for ourselves, make a place where we belong….”7 This has a nice, almost Heideggerian ring to it, but as Novitz notes, it does not tell us very much, and in any event it invites the refutation that some paradigmatic art makes its audience feel rather more alienated than at home on the humanized earth. I interpret Beardsley’s statement as pointing vaguely toward the rather un‐Beardsleyean notion of art as the affirmation of cultural identity. This construal would be consistent with the emphasis Novitz places on culturally constructed ways in which art affects us: “If the capacity of an artifact to enrich particular lives is not merely incidental to it but derives from its form and content, and if the artifact can be seen to instantiate the values that people live by, so that it somehow legitimates their existence and enhances their sense of who and what they are, and if, furthermore, the artifact is valued for this sort of complex reason, we would, given the present moral ethos in which we live, be inclined to endorse the claim (should it be made) that it is a work of art” (25, italics added). Novitz is clearly right about many art objects and the activities associated with them, viewed cross‐culturally: they enrich lives and amplify a sense of personal and cultural identity. But many objects and activities which are not art by Novitz’s own description – collecting the stamps of one’s

homeland, perhaps, or tooling about the countryside in one’s vintage car – accomplish the same goal nonartistically. Novitz distinguishes art identity‐building from nonart identity‐building with his qualification that the identity‐ building capacity of art should derive from its “form and content,” rather than content alone. Here he invokes a familiar and quite traditional Western criterion for art: that art’s value derives from the fusion of form and content, rather than from content (mere information or practical communication) alone. Yet over all, it seems to me that despite the fact that “the present moral ethos” of contemporary multicultural society may stress the enhancement of a sense of “who and what we are” as an important function for art, there is vastly more going on in the production and enjoyment of art cross‐culturally than is even suggested, let alone captured, by such a formulation. Noël Carroll has remarked on the way that art theories, despite claims to universality, are often rooted in aesthetic issues and debates of their own times.Thus, Carroll says, the theories of Clive Bell and R.G. Collingwood “are defenses of emerging avant‐garde practices – neoimpressionism, on the one hand, and the modernist poetics of Joyce, Stein, and Eliot on the other.” Susanne Langer can be read as providing a justification for modern dance, while the initial version of George Dickie’s institutional theory “requires something like the presupposition that Dada is a central form of artistic practice” in order to gain intuitive appeal.8 Arthur Danto’s near‐obsessional theorizing about indiscernible art/nonart objects, such as Warhol and supermarket Brillo boxes, derives from a special, and recent, theoretical problem, and I would add that Novitz’s implicit conception of art seems to grow, directly or indirectly, from current preoccupations with personal and cultural identity. None of these kinds of theoretical approach, emerging as they do from the concerns of their originating cultures, seem to me especially appropriate to the arts of small‐scale, nonliterate tribal societies, though each has partial relevance.9 I would not contend that art theory is explained away by being historicized and relativized to the social conditions or preoccupations of the theorist. But there is no denying Carroll’s gentle suggestion that the artistic interests and preoccupations of a theorist have the potential to affect, intentionally or not, the scope and substance of a theory, and that much of the art theory of this century has been connected to justifications of avant‐garde European art, and therefore, I would add, might be of only marginal relevance to understanding the arts of small‐scale, nonliterate cultures. Moreover, the insistence by philosophers on trying to hone definitions to the greatest scope combined



“but they don’t have our concept of art”

with the greatest simplicity, understood traditionally as a perspicuous and finite set of necessary and sufficient conditions, may work against understanding in a domain, in particular a domain as ragged and multilayered as that of tribal art. This, in any event, is the conclusion to which I have been forced by my own practical fieldwork and literature research into tribal arts, and in this respect at least I find myself in full agreement with Novitz’s remark that “precise formulations and rigorous definitions” are of little help in capturing the meaning of art cross‐culturally. Still, just because, as Novitz says, there is “no one way” to be a work of art, in tribal society or any other, it does not follow that the converse “many ways” are so hopelessly numerous as to be unspecifiable. In fact, that they are specifiable, however disputatiously, is required by the very existence of a literature on cross‐cultural aesthetics, arguments which make it possible for Novitz or me to publish views on the subject in aesthetics journals. These considerations persuade me to approach tribal arts as a subject for philosophic inquiry through an indefinite list of features characteristic of art in tribal, small‐scale, nonliterate cultures, While I do not claim that any one feature on this list is indefeasibly criterial for art in a tribal context, this list, or something close to it, is what makes possible cross‐cultural discourse about art in general. Granting that there may be marginal cases, by arts I mean artifacts (sculptures, paintings, and decorated objects, such as tools or the human body) on the one hand, and performances (dances, music, and the composition and recitation of stories) on the other hand. Features on this list can be found in the work of such writers as Richard L. Anderson and the ethnographer and philosopher H. Gene Blocker,10 and are even informally discussed by the philosopher Julius Moravcsik.11 Although it is intended to apply to every known tribal society, it has larger relevance, since every known society, tribal or large‐scale, makes and appreciates some form of art object or artistic performance. Not every element on the list can be associated with or incorporated into every art of a small‐scale, nonliterate culture, but most can be. 1. The art object, narrative story, crafted artifact, or visual and aural performance, is a source of pleasure in itself, rather than a practical tool or source of information. Its material embodiment may be a tool (a shield, a knife) or a source of information (a sacred poem), but aspects of the embodiment give pleasure in experience aside from these practical or information/communication considerations. (This pleasure is often called aesthetic

2.

3.

4.

5.

6.

39

pleasure; but I avoid the word aesthetic as implicitly circular, and in any event unnecessary, in this context.) The making of the object or the performance requires the exercise of a specialized skill.This skill is learned in an apprentice tradition in some societies or in others may be picked up by anyone who finds that she or he “has a knack” for it. Where the skill is acquired by virtually everybody in the culture, such as with communal singing or dancing, there are still to be individuals who stand out by virtue of special talents.Technical artistic skills are noticed in small‐scale societies and are generally admired. Art objects and performances (including oral narratives) are made in recognizable styles, according to rules of form and composition. The degree of stylistic determination varies as much in tribal cultures as in the arts of literate civilizations, with some sacred objects and performances being tightly circumscribed by tradition, with others open to free, creative, individual variation. The style may be the culture’s, or a family’s, or be the invention of an individual; styles involve borrowing and sudden alteration, as well as slow, evolving changes. There exists some kind of indigenous critical language of judgment and appreciation, simple or elaborate, that is applied to tribal arts. This may include the shop talk of art producers or evaluative discourse of audiences. Unlike the arts themselves, which can be immensely complicated, it has often been remarked that this critical discourse is in oral cultures sometimes rudimentary compared with the art discourse of literate European history. It can, however, be elaborate. In widely varying degrees of naturalism, art objects, including sculptures, paintings, and oral narratives, represent or imitate real and imaginary experience of the world. The differences between naturalistic representation, highly stylized representation, and nonimitative symbolism are understood by artists and their audiences in ways directly intelligible to Western observers. (Thus Danto’s view that there is “no distortion” in African art is certainly false from an indigenous perspective.12 Africans understand the distinction between highly realistic representations and stylistically distorted images or symbols; the distinction between naturalistic realism and stylized distortion is not a cultural construction.) The pleasures afforded by the arts of small‐scale societies to their indigenous audiences are consciously intended by the makers of such objects, even if the object’s indigenous meaning or importance is primarily

40

denis dutton

utilitarian or nonartistic. Aesthetic or artistic pleasure as an accidental by‐product of nonartistic activities is as common or as rare in tribal societies as it is in our society. The suggestion that tribal peoples might generally create things beautiful (to them or to us) without realizing it, or that they make things which are beautiful to us but to which they are wholly indifferent, is certainly false. 7. Works of art and artistic performances are frequently bracketed off from ordinary life, made a special and dramatic focus of experience.While there are plenty of mundane artistic objects and performances (such as decorated parts of Baule looms, or communal singing done to pass the time while mending fishing nets), every known culture has special artworks or performances which involve what Ellen Dissanayake calls “making special.”13 These occasions are often imbued with intense emotion. 8. Finally, and among the more important characteristics, the experience of art in tribal societies is an imaginative experience for both producers and audiences. The carving may realistically represent an animal, but as a sculpture it becomes an imaginative object. The same can be said of any story well told, whether mythology or personal history. The costumed dance by firelight, with its intense unity of purpose among the performers, possesses an imaginative element which transcends mere group exercise. In tribal cultures, as elsewhere in the history of human life, art happens in the theater of the imagination. There are other potential candidates for this list, items which, though perhaps more marginal or controversial, might warrant inclusion. Blocker, for example, thinks there is a near universality in tribal societies of the idea that the artist is considered “eccentric, or a bit socially awkward.”14 He also thinks the inherent tension between artistic tradition and novelty is a general aspect of tribal arts. I find both of these features appealing as candidates for the list, because they accord strikingly with my experience in the New Guinea village of Yentchenmangua. David Novitz’s notion of art as affirming the identity of a culture is also a relevant potential general characteristic, and could be applied especially to the more theatrical, large‐scale ceremonial occasions of tribal society, which often seem rather like patriotic rallies. But I shall for now limit the list to the central eight characteristics I have recounted. In their introduction to The Adapted Mind, Jerome R Barkow, Leda Cosmides, and John Tooby contend that for

the last few generations anthropologists have been prone systematically to overemphasize the differences between world cultures at the expense of recognizing similarities and pancultural universals. They quote with approval Maurice Bloch’s remark that anthropologists are guilty of a form of “professional malpractice” in the extent to which they have tried “to exaggerate the exotic character of other cultures.”15 This tendency, as I have noted throughout, has certainly infected the anthropological approach to art. Such mystification in ethnographic aesthetics often consists in focusing attention on what in any event is a marginal instance of art in another culture, or perhaps not art at all, and treating it as though it were characteristic of some exotic aesthetic form undreamt of in our philosophy, and therefore assaulted our aesthetic ethnocentrism. The standard strategy is: find a putative art object in a tribal society about which early ethnographers were wrong, or one which confounds any simple attempt at understanding, and you’ve demonstrated that “they don’t have art in our sense.” Among the Yoruba, twins are minor deities, and there is a genre of wood carvings to honor deceased twins, whose spirits in the older religion inhabit the sculptures. As Susan Vogel explains, however, this tradition is in decline, particularly among Muslim and Christian families.16 The older carvings, of which there are many stunning examples, are being replaced in some instances by simplified carvings of low relief, and in others by cheap, imported, mass‐produced, plastic dolls (with European features). Increasingly, no sculpture appears at all in the twin cult, but rather photographs, where the surviving twin often stands in for its deceased sibling. Both Vogel and David Novitz are impressed by the alacrity with which Yoruba people have been willing to supplant wood sculptures with cheap plastic dolls. Vogel sees these practices as “an updating of the tradition without rupture,” as “an imaginative use of imported items as replacements for traditional artworks.” Novitz draws from this phenomenon a more radical conclusion: the ibeji sculptures, since they are so easily replaced by mass‐produced dolls that “most assuredly would not be considered art in our culture,” are therefore “appreciated not for their originality, nor for their beauty, nor yet for their proportions; they are appreciated primarily as quasi‐ religious artifacts that allow the beneficial influence of the deceased twin to persist in the parents’ lives” (27). The ibeji carvings, Novitz says, “occupy a social space in Yoruba society that is remote from the social space occupied by works of art in our society.” I can see no argument in any of this showing why either the older or more recent ibeji carving is not art. Particularly



“but they don’t have our concept of art”

the older ibeji sculptures are (a) skillfully made objects, (b) produced in a recognizable, conventional style, (c) subject to a critical vocabulary among carvers and owners, (d) treated as very special objects, though in a private sense rather than for public display, (e) mimetic representations of the figure of a child, with conventional oversized head, and (f) imaginative objects  –  that is, they stand for the dead child and are inhabited by its spirit, but do not literally replace it.Taken together, these features are sufficient to call ibeji carvings works of art. In this respect I cannot share Vogel’s bland acceptance of these changes as an imaginative updating of a tradition. Like the replacement of Pueblo pottery by cheap (and more practical) tin pots in the nineteenth‐century Southwest, the invasion of the Yoruba ibeji cult by Taiwanese plastic toys does not constitute the further development of an artistic tradition, but its very death. In any event, none of this is relevant to whether historic or contemporary specimens of ibeji carving are art. There may be many reasons for the ready acceptance of plastic dolls as ibeji. Certainly, the Christianizing of Yoruba life is a major factor. Perhaps there are Yoruba mothers who are too poor to commission carvings, or are simply uninterested in ibeji statues as distinctly Yoruba art (thus, incidentally, casting doubt on whether the enrichment or enhancement of Yoruba identity with art makes any difference to them, at least in this case). The brightly colored plastic dolls may even have sheer novelty appeal. But in general, that there are people in any culture who do not care for an indigenous art, or who lose interest in it long enough for it to die out, so far tells us nothing about whether it actually is an art form. To construe the importation of plastic dolls into Yoruba life as showing that they have a different concept of art from us, or that their ibeji carvings are not art in our

41

sense, is yet another confounding, exoticizing, and mystifying digression in ethnographic aesthetics. The concentration by theorists of ethnographic art on dubious cases drawn from the ambiguous margins of the artistic life of tribal peoples (areas where art disappears, or is gradually replaced by ritual, religion, or practical concerns), on misleadingly described artistic practices, or on needless attempts to make foreign arts exotic, has inclined many aestheticians to give up the search for artistic universals, or at best to remain silent on the subject. But neither the universality of art nor the universality of its central features is endangered by the existence of marginal or disputed cases of art in tribal (or European) culture. The investigation of ethnographic arts is only impeded by the dogmatic refusal to discuss and debate their general features.The list I have provided does not insist that each of its eight characteristics will be present in every work of art. I do claim that any human practice which had none of the features enumerated would not be art, and that any human practice which possessed most of them would be art; not “art in our sense,” but art in the sense that characterizes it through the whole of human history. If this seems an unacceptably vague conclusion, that may be because the evolved, universal tendency of human beings to have art of some description in their lives does not produce a body of practice and artifact that is amenable to definition in terms precise enough to satisfy some theorists. But as anyone who has attempted figure drawing will attest, the human body, marvelous mechanism that it is, did not evolve in order to be an easy subject for artists. Nor, pace the simplifying impulses of theory, did the arts evolve in order to make life easy for philosophers.

Notes 1 Lynn M. Hart, “Three Walls; Regional Aesthetics and the International Art World,” in The Traffic in Culture: Refiguring Art and Anthropology, ed. George E. Marcus and Fred R. Myers (Berkeley: University of California Press, 1995), 127–50. 2 Alfred Gell, “The Technology of Enchantment and the Enchantment of Technology,” in Anthropology,Art, and Aesthetics, ed. Jeremy Coote and Anthony Shelton (Oxford: Clarendon Press, 1992), 40–63. 3 Susan Mullin Vogel, Baule: African Art, Western Eyes (New Haven:Yale University Press, 1997). 4 Francis E. Sparshott, “Art and Anthropology,” Journal of Aesthetics and Art Criticism 55 (1997): 239.

5 David Novitz, “Art by Another Name,” British Journal of Aesthetics 38 (1998): 19–32. 6 Clive Bell, Art (1913; New York: Capricorn, 1958), 17. 7 Monroe C Beardsley, The Aesthetic Point of View (Ithaca, NY: Cornell University Press, 1983), 370. 8 Noël Carroll, “Identifying Art,” in Institutions of Art: Reconsiderations of George Dickie’s philosophy, ed. Robert J. Yanal (University Park: Pennsylvania State University Press, 1994), 15. 9 I have in particular objected to Arthur C. Danto’s analysis of tribal arts in terms of indiscernible artifacts of the contemporary Western artworld. His essay “Artifact and Art” was first published in Art/Artifact, ed. Susan M. Vogel (New York:

42

denis dutton

Center for African Art, 1988). My response is “Tribal Art and Artifact,” Journal of Aesthetics and Art Criticism 51 (1993): 15–29. 10 Richard L. Anderson, Calliope’s Sisters: A Comparative Study of Philosophies of Art (Englewood Cliffs, N.J.: Prentice Hall, 1990). H. Gene Blocker, The Aesthetics of Primitive. Art (Lanham, MD: University Press of America, 1993). 11 Julius Moravcsik, “Why Philosophy of Art in a Cross‐ Cultural Perspective?” Journal of Aesthetics and Art Criticism 50 (1992): 233–49. 12 “David Hockney once told me that he believed that there is no such thing as distortion, and while I think him wrong in

general, he is right about African art, where there are no distortions.” Danto, “Artifact and Art,” 32. 13 Ellen Dissanayake, Homo Aestheticus (New York: Free Press, 1992), chap. 3. 14 Blocker, Aesthetics of Primitive Art, 148. 15 Jerome H. Barkow, Leda Cosmides, and John Tooby, The Adapted Mind: Evolutionary Psychology and the Generation of Culture (New York: Oxford University Press, 1992), 43–44. 16 Susan M.Vogel, “Elastic Continuum,” in Africa Explores: 20th Century African Art, ed. Susan M.Vogel (New York: Center for African Art, 1991), 32–55, 88–89.

5

Nobody Needs a Theory of Art Dominic McIver Lopes

It would never have occurred to the painters of Lascaux that they were producing art on those walls. Not unless there were neolithic aestheticians. Arthur Danto

What is art? More precisely, what makes any item a work of art? The question is “probably the most venerable in aesthetics,” according to Jerrold Levinson,1 and Monroe Beardsley adds that it is “the most vexing.”2 In fact, the ­following answer to the question should be obvious: (R) item x is a work of art if and only if x is a work in activity P and P is one of the arts.

The arts, of course, include music, dance, theater, literature, film, painting, architecture, and the like. Yet (R) has seemed so far from obvious that, until now, nobody has given it a moment’s thought.The trouble is not that anyone might seriously deny its truth, but rather that they will find it uninformative. After all, the vexing question is pressed upon us by radical changes in art of the avant‐garde, and (R) offers no resources to address these changes. With that in mind, here is the case for (R). The challenges posed by the avant‐garde are real and they need to be addressed, but the vexing question is the wrong question to address them. It does not follow that the question has no good answer. On the contrary, (R) is all the answer we need, if we do not

need an answer that addresses the challenges posed by the avant‐garde. Moreover, (R) points to questions that we do need answered. So, not only is it true but, in addition, (R) is as informative as we need.3

I.  From Kristeller to Testadura A theory of art is an answer to the question of what makes an item a work of art. (R) is one such theory among many.4 The case against (R) is that it fails to address the challenges posed by the avant‐garde.The case for it is that theories of art are not needed to address the challenges of the avant‐garde. However, the case for (R) runs immediately into a road block. The avant‐garde is not much older than a century, but the vexed question is venerable, and so is the search for a theory of art. It therefore seems that we expect a theory of art to do more than cope with developments in the [twentieth] century. Clearing this road block means doing a little history, and the history supplies materials useful in making the case for (R). As far back as 1914, Clive Bell identified the “central problem of aesthetics” with the search for “the quality that distinguishes works of art from all other classes of objects.”5 Morris Weitz began his famous attack on Bell’s project by admitting that “theory has been central in aesthetics and is still the preoccupation of the philosophy of art. Its main

Dominic McIver Lopes, “Nobody Needs a Theory of Art,” Journal of Philosophy 105 (2008), pp. 109–127. Reproduced by permission of The Journal of Philosophy.

Aesthetics and the Philosophy of Art – The Analytic Tradition: An Anthology, Second Edition. Edited by Peter Lamarque and Stein Haugom Olsen. © 2019 John Wiley & Sons, Inc. Published 2019 by John Wiley & Sons, Inc.

44

dominic mciver lopes

avowed concern remains the determination of the nature of art which can be formulated into a definition of it” (op. cit., p. 27). Responses to Weitz play along, conceding that “underlying every traditional aesthetic theory is the essentialist presumption that that the expression ‘work of art’ applies to the entities that it does in virtue of some shared essential property or properties.”6 A different story is told by Paul Kristeller in his classic study, “The Modern System of the Arts.”7 According to Kristeller, European texts deploy no concept of art until the eighteenth century. The Greek techne and Latin ars, for instance, refer to all kinds of crafts and sciences. The eighteenth‐century innovation is to group some activities apart from others. For the first time in history, painting, sculpture, architecture, music, and poetry are grouped (sometimes with decorative art, landscape architecture, dance, theater, or prose literature) and also distinguished from the liberal arts, the practical arts, and the sciences. Fishing, anatomical dissection, poetry, philosophy, football, travel, eating, architecture, fashion, trading, astronomy, dance, painting, plumbing… which of these belong together? A grouping is salient for you and me that was not salient for contemporaries of Aristotle and Abelard. Kristeller writes that “the various arts are certainly as old as human civilization, but the ­manner in which we are accustomed to group them and to assign them a place in our scheme of life and of culture is comparatively recent” (ibid., p. 45). Kristeller identifies some of the factors that, over the centuries, drove the innovation. While subsequent scholarship challenges some of the details, Kristeller’s account illustrates the kinds of forces that were in play. The humanists gave poetry top spot over grammar and rhetoric in their new curriculum. Painting, sculpture, and architecture g­ radually gained prestige from the fourteenth century onwards. A new body of literature in the sixteenth century compared poetry and painting. The seventeenth century saw the founding of the French academies, which ­sponsored the first treatises on painting, sculpture, and architecture to stand alongside texts on poetics. Contributors to the quarrel between the ancients and the moderns distinguished activities where success depends on accumulated knowledge (science, where the moderns win) from activities where success depends on individual talent (the arts, where the ancients win).Thus the new concept of science helped spur the new concept of art. Finally, the newly‐drawn distinction between beauty and moral goodness suggested a domain for each. By the early eighteenth century the “problem of the arts” was a topic of learned discussion in Paris and by mid‐century a consensus settled on the m ­ odern system of the arts.8

Several lessons are typically drawn from Kristeller’s essay. Some note that the concept of art is historically conditioned: it is acquired at a certain time in a certain place and it spreads from there. However, this is neither surprising nor interesting, for many concepts are historically conditioned in this sense. Others go further, concluding that there is no art before the eighteenth century.9 This implies that there is no art at a time unless people have a concept of art at that time  –  a controversial claim.10 (Kristeller holds that art is ancient even if the concept of art is modern.) A third lesson is that the concept of art begins as a theoretical concept. Like the concepts of polymers and logical completeness, it is introduced by means of a theory, indeed a definition. The concept of art and theories of art are equally venerable twins. Easily missed is a fourth lesson, which qualifies the third lesson. True, the concept of art is from the get‐go a theoretical concept devised by intellectuals. Kristeller is clear about their goals. They aimed to establish the unity of the arts – to make salient a scheme grouping together painting, sculpture, architecture, music, and poetry (and some ­others). They also sought to identify what sets these arts apart from other activities, especially the sciences and practical arts.11 They hoped to understand the distinctive features of each of the arts.12 Finally, some expected to ground rankings of the arts. This is not at all the project of Bell and his successors. Recent theories of art state what makes any given item a work of art. The early moderns wanted to know what makes any given ­activity one of the arts. They sought a theory of the arts. To see the difference, consider the form of each. A ­theory of art completes the schema item x is a work of art if and only if….

A theory of the arts completes the schema, activity P is an art form if and only if….13

The two types of theory are not the same, so evidence of the venerability of the one is not automatically evidence of the venerability of the other. Of course, the theories might be linked. In developing theories of the arts, the early moderns might have drawn implicitly on theories of art. They might have reasoned: 1. P is an an art form if and only if the works of P are works of art; 2. x is a work of art if and only if x is φ; 3. so P is an art form if and only if the works of P are φ.



nobody needs a theory of art

Here, (2) states a theory of art in the work‐oriented sense which delivers (3), a theory of the arts in the activity‐­ oriented sense. This is exactly how contemporary ­philosophers think up theories of the arts.14 Did the early moderns reason on the model of (1) to (3)? The absence of any explicit statement like (2) is inconclusive: it might be implicit. Better evidence lies in the kinds of counterexamples the early moderns considered. For example, Charles Batteux proposed that the arts imitate beauty in nature, and versions of the imitation theory gained dominance for several decades (op. cit.).When critics pointed out that music need not imitate beauty in nature, theorists spoke to the counterexample.15 Given that the imitation theory was tested against counterexamples, it is striking that nobody worried that Homer’s battle scenes and Grünewald’s Crucifixion are works of art but do not imitate beauty in nature. Counterexamples to (3) come in for debate, but not counterexamples to (2). The best explanation of this is that the early modern imitation theory is a theory of the arts that does not imply an imitation theory of art. That an art aims to imitate natural beauty does not imply that every work in the art is an imitation of natural beauty. Kristeller’s history ends with Kant, and it is interesting to consider his art‐theoretical concerns.16 In section 43 of the third Critique, Kant distinguishes art “as human skill” first from nature, then from science, and finally from craft. In the next section, he distinguishes the agreeable arts that aim at mere enjoyment from fine or schön art that furthers those mental powers that facilitate social communication. Sections 51 and 52 contain an extended discussion of each of the arts and their value. For Kant, there are connections between genius and the fine arts and also between the fine arts and beauty; but he never attempts to define the fine arts with reference to a theory of art. Hegel also inherited the early modern project.17 He writes that “the real world of art is the system of the individual arts” (ibid., p. 614), which function to “find for the spirit of a people the artistic expression corresponding to it” (ibid., p. 603), where an artistic expression is in a sensuous medium (ibid., p. 621). Thus sculpture expressed the spirit of the Greeks and music that of the Germans. Other nineteenth‐century writers have other fish to fry. Many speculate about the elements of greatness in art and hence (as they see it) the character of artistic genius, a Kantian theme. By taking what is great in art to be the work of the genius, these writers inch towards theories of art. Indeed, the search for a theory of art requires a grip on the idea that the products of the arts are works – an idea

45

that may only have developed at this time.18 At any rate, recent philosophers are not interested in art defined as an expression of genius. For them, art includes Bunyan as well as Milton, Boilly as well as Poussin, and Mahavishnu Orchestra as well as Miles Davis. The upheavals of the [twentieth] century launch the contemporary project of defining art.The historian Michael Fried writes that the avant‐garde “called into question … the already somewhat dubious concept of a ‘work of art’” for it “equipped one to treat virtually any object as a work of art.”19 Turning to the philosophers, Beardsley confesses that his search for a definition of art sprung from “the enormous and even ridiculous variety of objects, events, situations, texts, thoughts, performances, refrainings from any performance, and so on that have, in recent times, drawn the label ‘artwork’ from their authors, admirers, or patient endurers.”20 Likewise, George Dickie pays due to “the strange and startling innovations of Duchamp and his latter‐day followers” as the inspiration for his theory of art.21 The Kristeller of our times is Noël Carroll: it is in the twentieth century that the theoretical task of coming to terms with virtually continuous revolutions in artistic practice has become urgent. That is, it is in the twentieth century that the problem of identifying art has become persistently unavoidable. … The recurring task of the philosophy of art … has been to provide means to identify new and emerging work, particularly work of a revolutionary sort, as art.22

Whereas Batteux and Diderot seek a theory of the arts, Beardsley and Dickie seek a theory of art. Consider Weitz’s landmark anti‐essentialism (op. cit.). Weitz argues that there is no correct definition of art, where a definition of art states individually necessary and jointly sufficient conditions for an item’s being a work of art. Weitz assumes that art cannot be defined if the concept of art is an open concept. An open concept, first, has application conditions that are in principle emendable and, ­second, is not closed by stipulation. The application conditions of the concept of art are emendable because new cases “have constantly arisen and will undoubtedly arise … which will demand decisions on the part of those interested, usually professional critics, as to whether the concept should be extended or not” (ibid., p. 32). Moreover, the concept cannot be closed by stipulation since that “forecloses on the very conditions of creativity in the arts” (ibid., p. 32). Therefore art cannot be defined. Commentators quickly complained that artistic creativity is not incompatible with the concept of art being closed, unless open

46

dominic mciver lopes

c­ oncepts question‐beggingly entail creativity.23 Be that as it may, the argument also implies the empirical claim that tomorrow’s avant‐garde artworks falsify today’s definitions of art.Weitz targets definitional theories of art, not theories of the arts; and what drives his anti‐essentialism is avant‐ garde art production. A second landmark argument also bears out Carroll’s observation that the contemporary concern is with theories of art. In his 1964 paper “The Artworld,” Arthur Danto argues as follows.24 If art is defined in respect of perceptible features alone, then no artwork is perceptually indiscernible from a nonartwork. But some artworks are perceptually indiscernible from nonartworks. So art is not defined only in respect of perceptible features. The situation promotes a specific kind of error: Danto imagines a fictional Mr Testadura who visits his local MoMA and cannot tell the art from the nonart – he mistakes Robert Rauschenberg’s Bed for the kind of item sold at Sears. Again, what drives this argument are works like Bed that are perceptually indiscernible from “real objects” – and that the Testaduras of the world mistake for nonart. A case can be made for (R) if we expect theories of art to address the challenges of the avant‐garde and if, in addition, theories of art are not in fact needed to address these challenges. We do indeed expect theories of art to address the challenges of the avant‐garde. Are we wrong to do so?

II. The State of the Art Philosophy often obeys Newton’s third law of motion. Fifty years ago Weitz tried to scotch the project of definition forever (op. cit.);25 but his efforts led instead to an explosion of new theories of art, most of them definitions  –  a recent ­bibliography runs to more than four hundred items.26 This activity has not been wasted: we know more about art than we did before. However, we now face an impasse, a deep disagreement about avant‐garde works like Rauschenberg’s Bed. The impasse does not impeach the new theories, but it warrants doing as (R) recommends and looking elsewhere for tools to cope with the avant‐garde. The standard typology divides theories of art into functionalist and proceduralist varieties.27 On functionalist ­theories, what makes something a work of art is a function it has. Poisons are defined functionally, since what makes something a poison is its having the function of causing harm when absorbed into the body. Recent functional theories of art assign works of art an aesthetic function such as “affording an experience with marked aesthetic character.”28

On proceduralist theories, what makes an item an artwork is a matter of the “procedures, rules, formulas, recipes, or whatever by which artworks are generated.”29 Knighthood is defined procedurally: Sir Mick is made a knight by the Crown’s conferring the status upon him. In the case of artworks, the “whatever” may be an act of someone who is authorized within an informal institution to confer art status.30 The dispute between functionalism and proceduralism is not a dispute about whether artworks have a function. A proceduralist might argue that since nothing is a function of all and only artworks, functionalism is false, so proceduralism is true. However, that is just one route to proceduralism. Procedural theories of art are consistent with works of art having a function. By the same token, functionalism is consistent with there being procedures for generating all and only artworks. A functionalist might argue that since there are no procedures for generating artworks, proceduralism is false, so a functional theory is true; or she might just as well reject the opening premise of this argument. Functional and procedural theories render contrary ­verdicts when there is a gap between function and procedure – they represent a disagreement over puzzle cases. Some puzzle cases are imagined,31 but plenty are real, though a few get recycled in the literature. Marcel Duchamp’s Fountain of 1917 is a urinal from a plumbing supply store. Andy Warhol’s 1964 Brillo Boxes are plywood boxes painted to match Brillo soap pad boxes stocked at grocery stores. In 1953, Robert Rauschenberg erased a drawing by Willem de Kooning and showed it as the Erased de Kooning Drawing. A year before, Rauschenberg’s friend, John Cage, had premiered 4′33″, a work “for piano.” Functionalists reason that unless these works have the function that makes for artworks, they are not art. Proceduralists counter that they are art if they are generated by the required procedures. Dickie claims as a virtue of his procedural definition that it counts the puzzle cases as art; Beardsley claims as a virtue of his functional definition that it withholds them status as art.32 As long as our goal is to cope with the puzzle cases, we seem to have two options. We can ride with functionalists, who stick to their guns and deny that the puzzle cases are artworks. Or we can ride with the proceduralists and carry the flag for the puzzle cases. We must decide which class of theory gets the extension of art right. This would not be the first time in philosophy that a choice between theories represents a clash in intuitions. A dilemma is not by itself an impasse – it might be the starting point for making progress.Theories of art might rectify intuitions when there are independent grounds for theory choice – grounds that go beyond extensional adequacy.



nobody needs a theory of art

No theory is worth choosing unless it has content, coherence, and internal consistency. Critics fault versions of functionalism and proceduralism by all three standards. For example, some say that Dickie’s institutional theory of art is viciously circular, hence without content.33 Functionalism faces a similar charge. Beardsley claims that all and only works of art afford experiences with a marked aesthetic character: one gets the experience from Bob Marley’s “Duppy Conqueror,” T.S. Eliot’s “Ash Wednesday,” and Martha Schwartz’s Bagel Garden, but not from Olympic hockey, the essays of David Lewis, or a Macy’s window display. The worry is that no notion of an experience with marked aesthetic character both gets this sorting right and can be characterized independently of the sorting. Gallons of ink have been spilled making and answering these charges. Let us acquit the accused just to see where we get. Stipulate that versions of functionalism and proceduralism are contentful, coherent, and internally consistent. Apart from the general virtues of any theory worth having, some propose a trait that specifically marks a good theory of art.34 Art has value as a whole and a theory of art might explain art’s value. Moreover, individual artworks have value and a theory of art might ground critical principles that determine what value accrues to any given ­artwork.Thus one might prefer a theory of art if it explains the value of art as a whole or determines the principles attributing value to individual works of art. The trouble is that this criterion of theory choice is not common ground among functionalists and proceduralists. It favors functionalism and functionalists favor it; it disfavors proceduralism and proceduralists reject it in turn. This is explicit in exchanges between Beardsley and Dickie. Insisting that a theory of art should solve the mystery of art’s value, Beardsley notes that his own theory implies that “it is good for us to experience, at least occasionally and to a degree seldom made possible except by artworks, the immediate sense (say) of inclusive self‐integration and complex harmony with phenomenal objects.”35 Dickie agrees that we need an explanation of the value of art as a whole and a grounding for critical principles, but he denies that an adequate theory of art must supply the explanation. He awards his own theory a green star for hiving off the evaluative tasks to a separate theory of art value.36 No general principles of theory choice decide this dispute between Beardsley and Dickie. Analogies with other domains cut both ways. It seems reasonable to define life in a manner that does not imply an account of the value of life. It also seems reasonable that a theory of an institution like money should explain the value of money. One

47

s­uspects that it is intuitions about the extension of art that drive not only the choice of one class of theories over the other but also that drive the selection of criteria for theory choice. If the suspicion is accurate, then recent art theorists endorse incompatible types of theories because they rely (1) on different intuitions about the puzzle cases and (2) on different criteria of theory choice, where (2) is also determined by (1). That is a genuine impasse. The impasse does not impeach functionalist or proceduralist theories of art – it is methodological, not probative. But it brings up this question: What if we do not look to theories of art to help cope with the puzzle cases?

III.  What We Need History distinguishes theories of the arts, which were pursued by the early moderns, from theories of art, which are now used to address the challenges of the avant‐garde. But why look to theories of art to address these challenges? Why not theories of the arts instead? Or theories of the art forms – that is, theories of music, dance, theater, literature, film, painting, architecture, and the like.We have three types of theory to choose from: theory of art: item x is a work of art if and only if… theory of the arts: activity P is an art form if and only if… theory of an art form: item x is a work of art form P if and only if…. If (R) is our theory of art, then we need theories of the arts or the art forms to cope with the avant‐garde. Carroll observes that “avant‐garde mutations often strike the public and some critics as unintelligible, and therefore, as not art.”37 The scenario should ring a bell. The public is presented with an item – a urinal or a shark in a tank of formaldehyde – which defies expectations and induces bafflement. Sometimes the bafflement feeds a desire to know more, which feeds a cottage industry of slim primers on the nature of art. Sometimes it feeds a suspicion that we are victims of a hoax. Theories of art are crafted to address the bafflement. This bafflement can take one or both of two forms. On one hand, it might be conceptual. That is, avant‐garde works might confound a concept – a concept of art, a concept of the arts, or a concept of a specific art form. On the other hand, the bafflement might be theoretical. Avant‐ garde works might put pressure on a theory of art, a theory

48

dominic mciver lopes

of the arts, or a theory of a specific art form.The distinction between the two kinds of bafflement is important. It is possible that to have a concept of art or the arts or an art form is to know a theory of art or the arts or the art form. However, we cannot assume that having any one of these concepts consists in knowing the corresponding theory. Just as a person can have a concept of water without knowing that it is H2O, perhaps a person can have a concept of art without knowing what makes an item a work of art. This would mean only that having the concept involves deference: one may have a concept of art and defer to the experts about the nature of art. Begin with conceptual bafflement. Suppose that to have a concept of art is, at a minimum, to be pretty reliable at sorting items that are works of art from items that are not works of art. William Kennick set a test for having this concept: we have a concept of art just in case we mostly succeed when sent into a warehouse under orders to bring out all and only works of art (op. cit., pp. 321–2). Mr Testadura fails Kennick’s test. Sent into the warehouse, he mistakes art for nonart: he leaves behind avant‐garde works like Fountain, Brillo Boxes, and 4′33″. And we are all Testaduras. Avant‐garde works confound our ability to identify works of art, showing us to be poor art detectors. Hence our bafflement. In point of fact, we are not one bit like Testadura. In normal contexts, anyone minimally schooled in art can reliably tell art from nonart. Nobody normally mistakes Rauschenberg’s Bed for an ordinary bed or 4′33″ for an unscheduled intermission. That is, nobody has any trouble identifying artworks in museums or concert halls, set atop pedestals, or performed by people sitting at pianos. In normal conditions, Fountain and 4′33″ are no less markedly artworks than Winged Victory or The Magic Flute. The “normally” is obviously crucial.The capacity to identify an item of a kind often requires training and depends on context. I can tell red from orange in normal conditions, not any conditions. I can identify my keys in part because I know what they look like and also because they are in the dish in the front hall. Placed in context, the urinal or four and a half minutes of not playing the piano are blatantly works of art. Special conditions, like Kennick’s warehouse, do subvert art recognition, but that is not surprising. Many concepts are confounded in some conditions. Two facts, one about recent art and one about our concepts, drive the point home. First, it is possible to create artworks whose identification requires picking up on contextual cues, and a great deal of recent art plays on context in this way. As an art critic observes, although much recent

art “is rooted in an anti‐museum attitude … it is museums and art galleries that are the primary locus for such art… [which] needs an institutional context to be seen.”38 Second, people often defer to experts who know what makes an item a work of art. Moreover, the two points are connected. Such facts about an item as its having been placed on a pedestal trigger deference by indicating the experts’ verdict. Concepts of art forms such as music, dance, and painting are not in quite the same boat as the concept of art. First, context seems to play a lesser role in identifying items in at least some art forms. Second, most people defer less to experts on what items count as works of music, dance, or painting, for example. If so, then perhaps bafflement at the avant‐garde springs not from a challenge to the concept of art but instead from a challenge to art form concepts. Music is patterned sound and we recognize works as works of music by hearing patterned sound. Nobody attending a performance of 4′33″ doubts that they are in the presence of art.That is what they know for sure and will bet money on. The trouble is that they are told that they are listening to music and yet they identify music by listening for patterned sound. Likewise, nobody takes Erased de Kooning Drawing for a mere sheet of paper for it is displayed as art. What is baffling is how it is a drawing, for it does not fit the concept of a drawing. Conceptual bafflement has an evaluative side: artworks are made to be appreciated, not identified. So if Testadura does not exist, his slightly more sophisticated cousins do. They recognize artworks and yet are baffled because they do not know how to appreciate them. They are apt to remark that a small child could do the same. The usual explanation of their bafflement is that they do not employ up‐to‐date principles of art criticism. They judge Fountain by the same lights as Winged Victory and 4′33″ by the same lights as a Chopin étude. Here is another hypothesis: Testadura’s cousins do not need principles of art criticism, because they need critical principles pertaining to an art form. Kennick remarks that “only a man corrupted by aesthetics would think of judging a work of art as a work of art in general, as opposed to this poem, that picture, or this symphony” (op. cit., p. 329). Testadura’s cousins recognize Erased de Kooning Drawing as pictorial art, but remain baffled. Having attended Richard Wollheim’s Mellon lectures,39 they appreciate pictures as surfaces marked up so as to sustain experiences of seeing‐ in, experiences that reveal what the markings meant to the artists who made them. Erased de Kooning Drawing yields nothing along these lines, for there is nothing to see in it.



nobody needs a theory of art

The hypothesis is that avant‐garde works challenge traditional principles of music appreciation, picture appreciation, dance appreciation, and the like.They do not challenge generic principles applying equally to Middlemarch, “My Aim Is True,” Brillo Boxes, and Sagrada Familia. None of this plays down the challenge of the avant‐ garde. Avant‐garde works are baffling. However, the bafflement stems from a challenge to art form concepts rather than the concept of art, or to art‐form specific principles of criticism rather than generic principles of criticism. Turn next to theoretical bafflement. As we saw, having a concept of art (or an art form) need not amount to knowing a theory of art (or the art form). Most people have no trouble identifying art, but they defer to the experts about the nature of art. Still, wondering what the experts know can induce bafflement. On the standard story, we find ourselves in Kennick’s warehouse and exclaim, “Ugh, which of these is art?” This story is fiction. Still, we might truly find ourselves in the MoMA and exclaim, “This is art! Ugh, what are these people thinking?” We want to know what theories the experts wield  –  the theories knowledge of which makes them experts. Such bafflement is theoretical rather than conceptual. Theoretical bafflement is bafflement that is dispelled by revising one’s theories. Grant that works like Fountain or 4′33″ induce theoretical bafflement. The question is what kind of theory is needed to address the bafflement. According to the standard story, we need a theory of art. We wonder: If Fountain or 4′33″ is art then what in blazes makes anything art? We look to a theory of art to settle the question. Some adopt procedural theories that rule in Fountain and 4′33″ and others turn to functional theories that rule them out. But maybe all we need is a theory of the arts. We should wonder: If Fountain is sculpture then what in heck makes anything sculpture? Or if 4′33″ is music then what, pray tell, makes anything music? Theories of sculpture and music settle these questions. The two kinds of questions and the kinds of theories that settle them are asymmetric. Take Fountain. There is nothing left to settle when a theory of sculpture settles the status of Fountain as sculpture. A theory of sculpture that settles whether Fountain is a sculpture leaves nothing to be settled by asking whether Fountain is art.The converse does not hold. A theory of art that settles whether Fountain is art leaves unsettled the question whether it is sculpture. Perhaps it belongs to a new art form like conceptual art.40 Thus we do not need a theory of art except to address cases where an answer to the art form question leaves open the answer to the art question.We need a theory of art only

49

if it is settled by a theory of each art form whether a work belongs to the art form and yet it is not settled whether the work is art. In particular, we do not need a theory of art to cope with the philosophers’ puzzle cases. It is not settled by a theory of sculpture that Fountain is sculpture or by a theory of music that 4′33″ is music. What about other cases? It is not settled by a theory of landscape architecture that Christo’s wraps are architecture, by a theory of painting that Ben Vautier’s “Regardez moi cela suffit je suis art” is painting, by a theory of theater that a performance modeled on Augusto Boal’s “theater of the invisible” is theater, or by a theory of poetry that bp nichols’s “Cold Mountain” is poetry. We need a theory of art to cope with the avant‐ garde only if there are works that contribute to our bafflement whose art forms are settled while their status as art is open. There are no such works as long as every work of art belongs to at least one art form. The theoretical challenge posed by any given work can be settled in four ways. The first finds that a theory of an art form admits the work and it is an artwork. The second finds that no theory of any art form admits the work, but it is not an artwork. The third finds that no theory of any art form admits the work, but it is an artwork.The fourth finds that no theory of any familiar art form admits the work, but the work pioneers a new art form. Only the third case shows that we need a theory of art. However, any reason we have to say that a work is an artwork that does not belong to any art form is reason to say that it pioneers an unfamiliar art form. The move is not ad hoc. It represents how things really stand. It is a fact about the arts that their number grows  –  notable recent additions being photography, film, video, computer art, and conceptual art.That a work belongs to no familiar art form and yet is an artwork is reason enough to project a new art form. To sum up, the avant‐garde is baffling conceptually and theoretically. A case can be made that the conceptual bafflement is bafflement about art form concepts – so long as we rely less on context and defer less to experts in applying art form concepts than the generic concept of art. Moreover, we need a theory of art to address the avant‐garde only if theories of the arts leave it some work to do. The avant‐ garde works that baffle philosophers do not meet this condition. If attachment to traditional theories of art is hard to shake, then maybe the culprit is a category mistake. Each art form is obviously a topic of study which implicitly embeds a theory of that art form, so we have theories of the

50

dominic mciver lopes

arts. Then, just as we wonder what men and women make up the division apart from the battalions, batteries, and squadrons, we wonder what is art apart from each of the arts.41 The mistake is to treat art like another one of the arts – to look for a theory of art to place alongside theories of music, pictures, drama, and the rest.

IV.  What We Can Have If the avant‐garde is baffling, then what is the right description of the bafflement? The assumption has been for several decades that it only makes sense to describe the bafflement as a challenge to concepts or theories of art. But maybe the bafflement is a challenge to concepts and theories of the art forms  –  or some of them, anyway. How do we choose between the two descriptions? Intuitions can hardly be expected to respect the required distinctions between concepts and theories of art and the art forms.We must choose the option that helps us best understand what we need to understand. As long as we do not need a theory of art to address the avant‐garde, we can have (R). (R) is no less informative than it need be. It passes the buck to theories of the individual art forms and theories of the arts as a whole, and it is from these that we should expect answers to our questions. Of course, the prospects for these theories had better look good. In fact, their prospects improve if (R) is true. The buck passes first to theories of the art forms. These tell us what makes some items but not other items works in a given art form. Here is a hurdle. Painting is one of the arts. If a work of painting is a product of that activity, then a dirty rag is a work of painting. But it is not an artwork. The reply is that works in an activity are not merely products of the activity; some activities have products only some of which are works. A start is this: x is a work in P just in case x is a primary product of P – that is, a product of a kind upon which the existence of P conceptually depends. A dirty rag is not a primary product of painting because there could be painting without dirty rags. Since there can be no painting without paintings, paintings are primary products of painting. So paintings are works of painting. Another objection targets theories of the arts conjoined with (R). Grant that every work of painting, music, and theater is a work of art – that is contentious enough. Surely some pots are works of art, but not the ones I buy from

Walmart, and some buildings are works of art, but not my garden shed. And if ceramics and architecture are arts, and pots are ceramics and sheds are buildings, then any pot or shed is an artwork. Better reject (R) than accept such an absurd result. The reply displays some advantages of (R). Nothing requires the identification of works in an art form with works in a medium like ceramics or building. There is a distinction between ceramics and works in the ceramic art form, between buildings and works of architecture. However, the distinction can be made differently for different arts. Works of architecture may have a function that mere buildings lack, and works in the ceramic art form may be generated in accordance with special procedures (maybe they are made by MFAs). Traditional theories of art require the very same distinction, but attempt to make it ­uniformly – so that what distinguishes Sagrada Familia from a garden shed also distinguishes bizen ware from Walmart ware. Theories of the arts conjoined with (R) need not take a uniform approach. First, functional theories of music and painting might assign to each a different function. Maybe paintings are imitative and musical works are expressive. Second, it is possible that some art forms are best characterized functionally and others procedurally. Moreover, functionalist theories of the art forms are more promising than functionalist theories of art.The latter posits a function shared by “Duppy Conqueror” and “Ash Wednesday,” but not by a Rangers game or a window display at Macy’s. Dance has it, but not ice dance; ballet has it, but not gymnastics; novels have it, but not biographies. By contrast, a functional theory of music may characterize the function of musical works as specific to music. Think of Eduard Hanslick’s account of a “specifically musical kind of beauty” in “tonally moving forms.”42 Thinking of Hanslick suggests where to look for the function, if there is one, for it will be implicit in musicology, music theory, music criticism, and the psychology of music. Functional theories of the arts are in development,43 and so are theories of the aesthetic specific to art forms.44 Finally, functional theories of the arts are less vulnerable to avant‐garde counterexamples than functional theories of art. Suppose that a functional theory of theater implies that a work of “invisible theater” is not theater. This result is not an impasse: it is a fruitful discovery. It does not ­follow that the work is not art, and if the work is art then it follows from that fact together with (R) that the work is  a  work in another, perhaps unfamiliar, art form.



nobody needs a theory of art

The innovations of the avant‐garde are represented on this view as innovations. The buck also passes to theories of the arts as a whole. Not all activities are arts. Painting, poetry, and architecture are arts; window dressing, philosophy, and garden shed ­fabrication are not. Which activities are arts? This is the question, now roundly ignored, that troubled the early moderns. They sought functional theories of the arts. Kristeller himself was skeptical, remarking that “the system of the fine arts is hardly more than a postulate and most of its theories are abstracted from particular arts, usually poetry, and more or less inapplicable to the others” (op. cit., p. 46). Many will agree. Do painting, video art, landscape architecture, and literary prose writing have any function that is not a function of tattooing, fly tying, gymnastics, philosophy, and perfumery? Why count dance and calligraphy as arts but not ice dance and paper‐making; or mime but not clowning; sculpture but not body building; opera but not World Cup soccer? Skeptics about functional theories of the arts have an option not open to the early moderns – procedural theories. Nobody has developed such a theory, though it is worth trying. A fascinating, unexplored body of literature is ready at hand. During the past hundred years, several new art forms have been recognized following debates about their admissibility. Proponents of photographic art, conceptual art, and video art might have appealed to the same general criteria for any activity to be an art form. If so, their ­reasoning might suggest a functional theory of the arts. Alternatively, their reasoning might show that only a ­procedural theory of the arts has any plausibility. Either way, (R) is consistent with functional and procedural ­theories of the arts. The two bucks are passed independently. (R) allows that theories of the art forms can be developed without waiting for a theory of the arts. And if there be worthy questions about why some activities are arts, we need not wait upon answers in developing theories of the art forms. A final worry. One might accept all this buck passing and still reject (R). After all, (R) is implicitly disjunctive and so implies that there is no unified phenomenon of art. That may be fine as a bit of philosophy, but the concept of art is deployed in theoretical contexts – in scientific and historical hypotheses, for instance – and we need a theory of art to make explicit the nature of art as an object of study outside philosophy. Beardsley writes that “the definition of

51

‘artwork’ should be of the greatest possible utility to inquirers in other fields besides aesthetics” and he is surprised that nonphilosophers are “wont to spurn the friendly offers of aestheticians to help them out.”45 The response of art scholars makes perfect sense if they make no use of a concept of art in their work on art (even if they use a concept of art in discussing our concept of art). Art is not the object of any study. There are no serious psychological, anthropological, sociological, or historical hypotheses about art, although there are serious hypotheses about pictures, music, and the novel. Moreover, there is no reason to expect such uniformity across studies of the arts as might imply a uniformity of theories. Film historians need not think of their object of study in a way that is consistent with the way music theorists think of their object of study, for there are no hypotheses about movies and songs as a single phenomenon. Or, if there are hypotheses about works in all of the arts they are also hypotheses about a lot of other phenomena besides.46 Beardsley almost sees this when he writes that “it stands to reason that someone starting out to write a history of (visual) art would want to have a reasonably definite idea of what it is he is writing a history of.”47 Drop the parentheses around “visual” and the claim is obvious, for a historian of visual art perforce deploys a concept of visual art. It does not follow, however, that she employs a concept of art in any sense except that given in (R). This is no accident if it reflects the fact that there is nothing left to explain when you have theories of the art forms and the arts.

V. Conclusion Although avant‐garde puzzle cases create demand for theories of art, they are best addressed by theories of the individual arts and theories of the arts as a whole. All we need in a theory of art we get from (R). Happily, (R) breathes new life into functionalism, because functional theories of the arts stand a fighting chance. It also tips a hat to proceduralism since procedural theories of the arts are especially plausible. Perhaps, in the end, (R) is less radical than it seems. Its purpose is to change the subject; but the new subject is one that is already getting attention from philosophers who are crafting theories of the arts. True, we must also revive the question that troubled the early moderns, but having more work to do is a blessing unless it cannot be done.

52

dominic mciver lopes

Notes 1 Levinson, “Defining Art Historically,” British Journal of Aesthetics, xix (1979): 232–50, here p. 232. 2 Beardsley, “Redefining Art,” in The Aesthetic Point of View (Ithaca: Cornell UP, 1982), pp. 298–316, here p. 298. 3 This paper deflects the vexed question by defending (R) while taking seriously the challenges of the avant‐garde. Others dismiss the vexed question. For example, Kendall Walton, “Review of Art and the Aesthetic, by George Dickie,” Philosophical Review, lxxxvi (1977): 97–101; Richard Wollheim, Painting as an Art (New York:Thames and Hudson, 1987); and Peter Kivy, Philosophies of Arts (New York: Cambridge UP, 1997). 4 Notably Morris Weitz, “The Role of Theory in Aesthetics,” Journal of Aesthetics and Art Criticism, xv (1956): 27–35; Monroe Beardsley, “An Aesthetic Definition of Art,” in Hugh Curtler, ed., What Is Art? (New Haven: Yale UP, 1983), pp. 15–29 [reproduced as Chapter 3 in this Anthology]; Arthur Danto, Transfiguration of the Commonplace (Cambridge: Harvard UP, 1981); George Dickie, The Art Circle (New York: Haven, 1984); Levinson,“Defining Art Historically”; Stephen Davies, Definitions of Art (Ithaca: Cornell UP, 1991); Berys Gaut, “‘Art’ as a Cluster Concept,” in Noel Carroll, ed.,  Theories of Art Today (Madison: Wisconsin UP, 2000), pp. 25–44; and Gary Iseminger, The Aesthetic Function of Art (Ithaca: Cornell UP, 2004). 5 Bell, Art (London: Chatto and Windus, 1914), p. 3. 6 Robert Matthews, “Traditional Aesthetics Defended,” Journal of Aesthetics and Art Criticism, xxxviii (1979): 39–50, p. 39. 7 Kristeller, “The Modern System of the Arts,” Journal of the History of Ideas, xii (1951): 496–527, and xiii (1952): 17–46. See also Larry Shiner, The Invention of Art: A Cultural History (Chicago: Chicago University Press, 2001). 8 Especially Charles Batteux, Les Beaux arts réduits à un même principe (Paris, 1746); and Denis Diderot, ed., Encyclopédie ou Dictionnaire Raisonné des sciences, des arts, et des métiers (Paris, 1751). 9 For example, Stephen Davies, “Definitions of Art,” in Berys  Gaut and Dominic Mclver Lopes, eds., Routledge Companion to Aesthetics (New York: Routledge, 2005, 2nd ed.), pp. 227–39, here p. 234. 10 Lopes, “Art without ‘Art’,” British Journal of Aesthetics, xlvii (2007): 1–15. 11 In The Invention of Art, Shiner stresses the distinction between the arts and the practical sciences, or crafts. This emphasis might project a later concern onto the earlier period, when it was just as important to distinguish the arts from the sciences. 12 For example, G.E. Lessing, Laocoon, E.A. McCormick, trans. (Indianapolis: Bobbs‐Merrill, 1962 [1766]). 13 Theories are definitions if they give individually necessary and jointly sufficient conditions. They are not definitions if

they provide disjunctive lists which are either closed or open to future revision. 14 For example, Dickie, The Art Circle. 15 Kivy, Philosophies of Arts, chapter 1. 16 Kant, Critique of the Power of Judgement, Paul Guyer and Eric Matthews, trans. (New York: Cambridge UP, 2000). 17 Hegel, Aesthetics: Lectures on the Fine Arts, T.M. Knox, trans. (New York: Oxford UP, 1975). 18 Lydia Goehr, The Imaginary Museum of Musical Works (New York: Oxford UP, 1995). 19 Fried, Three American Painters (Cambridge: Harvard UP, 1965), p. 47. 20 Beardsley, “Redefining Art,” p. 298. 21 Dickie, “The Institutional Conception of Art,” in Benjamin Tilghman, ed., Language and Aesthetics (Lawrence: Kansas UP, 1973), pp. 21–30, here p. 27. 22 Carroll, “Historical Narratives and the Philosophy of Art,” Journal of Aesthetics and Art Criticism, li (1993): 313–26, pp. 313–14. 23 Maurice Mandelbaum, “Family Resemblances and Generalizations Concerning the Arts,” American Philosophical Quarterly, ii (1965): 219–28. See also Kivy, Philosophies of Arts, pp. 36–37. 24 Danto, “The Artworld,” Journal of Philosophy, lxi (1964): 571–84. 25 See also Paul Ziff, “The Task of Defining a Work of Art,” Philosophical Review, lxii (1953): 58–78; and William Kennick, “Does Traditional Aesthetics Rest on a Mistake?,” Mind, lxvii (1958): 317–34. 26 Jinhee Choi, “Bibliography,” in Theories of Art Today, pp. 241–62. 27 Davies, Definitions of Art. 28 Beardsley, “Redefining Art,” p. 299. See also James Anderson, “Aesthetic Concepts of Art,” in Theories of Art Today, pp. 65–92; and Iseminger, The Aesthetic Function of Art. 29 Davies, Definitions of Art, p. 30. 30 Dickie, The Art Circle. 31 Davies, Definitions of Art, pp. 100–6. 32 Beardsley, “An Aesthetic Definition of Art,” p. 25. 33 Wollheim, Painting as an Art; Gaut, “‘Art’ as a Cluster Concept.” 34 Beardsley, “Redefining Art.” 35 Beardsley, “In Defense of Aesthetic Value,” Proceedings and Addresses of the American Philosophical Association, lii (1979): 723–49, here p. 743. 36 Dickie, Evaluating Art (Philadelphia: Temple, 1988). 37 Carroll, “Historical Narratives and the Philosophy of Art,” p. 316. 38 Michael Rush, New Media in Late Twentieth Century Art (New York: Thames and Hudson, 1999), p. 116. 39 Wollheim, Painting as an Art.



nobody needs a theory of art

40 Lopes, “Conceptual Art Is Not What It Seems,” in Peter Goldie and Elizabeth Schellekens, eds., Conceptual Art and Philosophy (New York: Oxford UP, 2007), pp. 238–56. 41 Gilbert Ryle, The Concept of Mind (London: Hutchinson, 1949), pp. 16–17. 42 Quoted in Kivy, Philosophies of Arts, p. 28. 43 For example, Wollheim, Painting as an Art; Peter Lamarque and Stein Haugom Olsen, Truth, Fiction and Literature (New York: Oxford, 1994); and James Hamilton, The Art of Theater (Malden, MA: Blackwell, 2007).

53

44 For example, Lopes, Sight and Sensibility: Evaluating Pictures (New York: Oxford UP, 2005). 45 Beardsley, “Redefining Art,” p. 304. 46 For example, Howard Gardner, Art, Mind, and Brain (New York: Basic, 1982); and Ellen Winner, Invented Worlds: The Psychology of the Arts (Cambridge: Harvard UP, 1982). 47 Beardsley, “Redefining Art,” p. 304.

6

Art: What it Is and Why it Matters Catharine Abell

Introduction We do not, in general, expect a definition of some concept to elucidate the value of the things to which it applies. For example, we do not generally expect a definition of knowledge to explain why it is valuable to know things rather than merely to truly believe them. Nevertheless, as Edward Craig notes, an adequate definition of knowledge should be regarded as a prolegomenon to a further inquiry into what general human needs served by the concept of knowledge account for its widespread use. To the extent that existing definitions of knowledge ignore this further question, the complexities they introduce to accommodate our intuitions about the extension and the intension of the concept may impede subsequent attempts to explain its value (Craig 1990: 2). Likewise, definitions of art that ignore the question of its value may impede subsequent attempts to illuminate that value. Clive Bell claimed that the problem of identifying the quality that distinguishes artworks from all other kinds of objects is the “central problem of aesthetics” (Bell 1914: 3). Jerrold Levinson likewise claims that the problem is “probably the most venerable in aesthetics” (Levinson 1979: 232). However, definitions that ignore the question of why we value art may rob this question of much of its philosophical significance.

Evaluative definitions of art pursue the projects of definition and of value elucidation simultaneously, by defining artworks as things with value of a certain kind. However, they have the undesirable consequence that to be art is necessarily to be good art. By so closely associating what it is to be art with its value, they preclude the possibility of bad art.What is needed is a descriptive definition of art that is able to accommodate the existence of bad art, while illuminating the value of good art. Craig (1990) proposes to address the task of defining knowledge by beginning with a hypothesis about the human needs served by the concept of knowledge, and then asking what conditions would govern the application of a concept that played that role. My approach in this paper is somewhat similar. Starting with the hypothesis that artworks are the products of institutions that serve certain human social needs, I then go on to identify the conditions something must meet in order to be an artwork, and then examine what these conditions reveal about the value of good art.

1.  Functionalist Approaches Historically, the most common approach to the task of defining art has been functionalist. In their simplest form, functionalist theories hold that something is art iff it has a

Catharine Abell, “Art: What it Is and Why it Matters,” Philosophy and Phenomenological Research 85 (2012), pp. 671–91. Reproduced by permission of Wiley.

Aesthetics and the Philosophy of Art – The Analytic Tradition: An Anthology, Second Edition. Edited by Peter Lamarque and Stein Haugom Olsen. © 2019 John Wiley & Sons, Inc. Published 2019 by John Wiley & Sons, Inc.



art: what it is and why it matters

certain function. Proposals for the function definitive of art include having aesthetic value, having valuable formal characteristics, expressing thoughts or feelings, and embodying meanings. Functionalist accounts promise not only to explain what it is for something to be art, but also to explain its value.They typically hold that the value of art consists in its performance of the function in terms of which they define art. A problem that faces all these proposals is that many ­artworks simply do not seem to have the functions at issue. Avant garde works in particular pose a problem for many functionalist accounts. In addition to such works as Titian’s Venus of Urbino, Bach’s GoldbergVariations, and Michelangelo’s David, which perform at least some of the above functions, the category of artworks includes works such as Duchamp’s Fountain, Manzoni’s Merde d’Artiste, Damien Hirst’s The Physical Impossibility of Death in the Mind of Someone Living, and John Cage’s 4′33″. Such avant garde works comprise counterexamples to functionalist accounts which claim that all artworks have aesthetic value, or valuable formal characteristics, or express thoughts or feelings. One response to this problem is to loosen the connection between something’s being an artwork and its having a function. For example, one could claim that something is art iff: it belongs to a type that is typically intended to have a certain function or it is of a kind such that something is good of that kind iff it has that function. Such accounts define art in terms of a function without insisting that every artwork must have the function at issue. This enables accounts that define art in terms of functions such as the above to accommodate avant garde works. However, there is a cost to adopting this response. Functionalist accounts that deny that all artworks perform a certain function are typically presented as descriptive definitions according to which what distinguishes good art from bad is that good artworks fulfil the function at issue, whereas bad artworks do not. Insofar as they are able to illuminate the value of art, accounts that take this line accommodate avant garde works only by construing them as bad art. However, some avant garde works seem to be good art. An alternative response is to insist that, contrary to appearances, there is some function common to all artworks. Advocates of this approach face the problem of identifying some function that both traditional and avant garde works perform (or that good traditional and good avant garde works perform) but that things that are not artworks do not perform (or are not typically intended to perform, or are not of a kind such that they are good of that kind if

55

they perform that function). What makes this problem difficult to surmount is that any function broad enough to accommodate both traditional and avant garde works is likely either to be performed not just by artworks but also by things that are not artworks, or to be specified too imprecisely either to clearly include all artworks or all good artworks, or to rule out things that are clearly not artworks or are not good artworks. In both cases, because the function at issue does not distinguish clearly between artworks and things that are not artworks, the proposed functionalist definition fails. For example, Arthur Danto claims that both traditional and avant garde artworks have meanings. In response to counterexamples proposed by George Dickie, he attributes meanings to the works in question and claims “give me an example, and I will deal with it” (Danto 2000: 133). This response is inadequate. That it is possible to attribute meanings to all artworks does not suffice to show that they actually have such meanings, or that the meanings they have differ in kind from the meanings of newspaper reports, conversations, instruction manuals and other things that are not artworks. To do this, one must provide an account of what it is for an artwork to have meaning of the kind distinctive of (good) artworks, and then show both that all (good) artworks meet the account’s criteria, and that all things that are not (good) artworks do not. Danto fails to do this.

2.  Procedural Approaches Procedural accounts construe the procedure by which things are produced, rather than the function they serve, as essential to determining whether or not they are artworks. Things produced by a single procedure may not share any common function. Stephen Davies notes that practices or procedures that are developed to produce things with a certain function may, once established, continue even when their products cease to serve that function (S. Davies 1991: 31).When this happens, the products of a procedure will no longer have a common function, but will be united only by the shared procedure by which they are produced. While procedural accounts are consistent with the claim that ­artworks once had a common function, they construe the practices or procedures by which artworks are made to be their only essential feature. Institutional and historical definitions are the two most prominent procedural approaches. Institutional definitions hold that art making is an essentially institutional activity, such that whether or not something is art depends on

56

catharine abell

whether or not members of the institution of art deem it art, or are prepared to accept it as art (Dickie 1969). They hold, therefore, that the practice of art making necessarily occurs within the institutional context of the artworld. Historical definitions hold that something is an artwork iff it is intended to be regarded in at least some way in which artworks of the past were correctly or standardly regarded (Levinson 1979). New ways of correctly regarding artworks may emerge because new artworks need be intended to be regarded only in part in ways correct or standard for artworks of the past (Levinson 1979). Because there may be other ways in which such works are intended to be regarded or are correctly or standardly regarded, the introduction of new artworks introduces new regards able to ground sufficient conditions for arthood. This account is essentially historical in attempting to define art at a given time by reference to an uncontroversial body of past art. Both these procedural accounts allow enormous latitude in the sorts of things that can be art. Because it is generally assumed that members of the institution of art may deem art or accept as art anything whatsoever, the institutional approach allows things of any sort to be art. Likewise, on the historical approach, works that are intended to be regarded in one of the ways in which past art is correctly regarded may also be intended to be regarded in almost any further respect, with the consequence that this further respect then becomes one of the ways in which past art is correctly regarded. This latitude enables these procedural accounts to accommodate avant garde works, and thus gives them an important advantage over functionalist accounts. Nevertheless, procedural accounts also suffer some important disadvantages relative to functionalist accounts. Firstly, they tend to be circular. Institutional definitions define art by appeal to the institution of art, but do not go on to define that institution independently of the notion of art. Historical definitions define art by appeal to past art, without providing a reductive account of first art. This ­circularity does not deprive these accounts of explanatory value altogether. Dickie claims that his non‐reductive institutional account explains the relations between the concepts of an artwork, an artist, the artworld, an artworld system, and an artworld public (Dickie 2000: 102), and Levinson claims that his historical definition enables us to identify from a body of past artworks those works that are art now (Levinson 1979). Whether or not one considers such circularity a fatal flaw, there remains a further respect in which procedural accounts are inferior to functionalist accounts. The very arbitrariness that enables them to accommodate avant

garde artworks precludes them from elucidating the value of art. Presented as evaluative definitions, they are implausible. The mere fact that something bears the requisite ­relation to either the institution of art or past art does not suffice to make it good art, since a great many things we consider bad art meet the proposed criteria. Such definitions are plausible only when presented as descriptive. However, unlike descriptive functionalist definitions, they lack even the potential to explain what makes some artworks good. They are utterly unilluminating about either why we care about art, or why it plays the role that it does in our lives. Such accounts therefore make it difficult to see why we should be interested in the phenomenon of art in the first place. Instead, by construing art status as arbitrary, they call into doubt the claim that it does matter. My aim, in this paper, is to develop a non‐circular procedural account of art that is able to explain the value of good artworks. In the next section, I evaluate a range of recent responses to the problem of defining art. In section four, I outline John Searle’s account of institutions and institutional facts. In section five, I argue that Searle’s account shows that the existence of all institutions is due to their being perceived by their participants to perform some humanly valuable function, and identify the functions to which the existence of art institutions is due. I then use these functions, in section six, to provide a reductive institutional definition of art. Finally, in section seven, I examine the account’s consequences for the value of good art.

3.  Recent Approaches The claim that art status is conferred by agents operating within art institutions is central to many previous institutional accounts. According to Dickie’s earliest elaboration of the institutional approach, art is produced by a process whereby artefacts have the status of candidate for appreciation conferred on them by people acting on behalf of the social institution of the artworld (Dickie 1971) (see also (Dickie 1969)). Stephen Davies objects that this account places no restrictions on who can confer art status on objects. He suggests that a non‐circular institutional definition of art could be provided by explaining what it is for an agent to have the requisite institutional authority to confer art status (S. Davies 1991: 112). An adequate institutional definition, he claims, should provide an account of the institutional roles that carry that authority, identifying their boundaries and limitations, conditions for occupancy, and the circumstances under which they change (S. Davies 1991: 94).



art: what it is and why it matters

Stephen Davies implicitly assumes that art making, like marrying people or sentencing them to jail, involves the performance of what Searle calls declarative illocutionary acts (Searle 1979). Declarations are alone among illocutionary acts in having the power to bring about an alteration to the status or condition of the things to which they refer simply in virtue of their successful performance (Searle 1979: 17). When a celebrant says to a man and a woman “I pronounce you man and wife”, they are thereby married. If a judge utters the words “I sentence you to three years’ jail” in the appropriate circumstances, you are thereby sentenced to three years’ jail. The performance of declarations involves the exercise of authority. One cannot perform the act of marrying unless one is a priest or a celebrant, and one ­cannot sentence someone to jail unless one is a judge or a magistrate. Similarly, Davies assumes, there is some act whose successful performance enables those with the requisite authority to make something a work of art. Declarations depend on the existence of extra‐linguistic institutions in order to bring about these alterations to the status of their objects. Only in virtue of the institution of marriage does a celebrant’s saying “I pronounce you man and wife” effect a marriage, and only in virtue of the institution of the law does a judge’s saying “I sentence you to three years’ jail” bring it about that you are sentenced. The claim that art status conferral involves the performance of declarations therefore construes art as an essentially institutional phenomenon. Stephen Davies does not actually carry out the task of analysing the institutional roles that carry the authority to confer art status. Nevertheless, even if this task were undertaken, it is unlikely to result in an adequate non‐circular definition of art. Even assuming, with Davies, that art is the product of declarations, explaining what it is for an agent to have the requisite institutional authority to confer art status will not result in an adequate institutional definition of art. The authority required to perform a declaration can change over time. While it used to be that only priests could marry people, for example, celebrants have now been granted this authority. It is implausible that all art institutions at all times share the same authority roles. This problem can be overcome, Stephen Davies believes, by explaining how the current structure and authority roles of the institution of art are determined by its previous structure and the wider social context (S. Davies 1991: 95). However, the way in which the prior structure of an art institution interacts with the wider social context to determine its current structure and authority roles is a contingent matter. The wider social factors that affect the development

57

of art institutions include the other institutions with which they interact. For example, throughout their histories, the structures and authority roles of art institutions have been influenced by their interactions with both property and religious institutions.The nature and effects of these interactions depend, not just on the historical structure of art institutions, but on the historical structures of these other institutions. Like those of art institutions, these structures are contingent. It is implausible, therefore, that it will prove ­possible to provide any general account of how the successive structures and authority roles of art institutions are determined by its previous structures and the wider social context. Both Robert Stecker and David Davies attempt to combine aspects of both functionalism and proceduralism in a way that overcomes the limitations of each. Stecker argues that something is a work of art at a certain time, where that time is no earlier than the time at which it was made, iff either (a) it is in what counts as one of the central art forms at that time and is made with the intention of fulfilling a function that art has at that time, or (b) irrespective of whether it is in a central art form or was intended to fulfil such a function, it is an artefact that achieves excellence in fulfilling such a function (Stecker 1997: 50). This account is like historical accounts and unlike ­functionalist accounts in accommodating the fact that the functions artworks perform change over time. By allowing that the functions that traditional works perform may enable them to meet the criteria for being art relative to one time and that the functions that avant garde works perform may enable them to meet the criteria for being art relative to another time, Stecker’s account explains why both kinds of works have a claim to art status despite lacking a common function. Moreover, he suggests that both may count as art relative to a single time because, in determining what counts as art now, we “tend to be maximally charitable” by admitting both works that meet the criteria for being art relative to the present time and works that meet the criteria for being art relative to the time at which they were produced (Stecker 1997: 53). However, the account is like functionalist accounts and unlike Levinson’s historical account in allowing the functions works actually perform to take precedence over the intentions with which they were produced in determining whether or not they are art, as the second disjunct of Stecker’s account makes clear. While it arguably overcomes functionalism’s inability to identify some feature or features distinctive of both traditional and avant garde art by appealing to different functions that art has at different times, and overcomes proceduralism’s

58

catharine abell

inability to explain why art matters by appealing to its ­various functions, Stecker’s account does not overcome the circularity of existing procedural accounts. It attempts to define art by appeal to central art forms and to functions that art has. Stecker claims that this circularity is merely apparent, on the basis that it is possible, in principle, to say what the central art forms and functions of art are for any past or present time, because we do not need a definition of art in order to recognise things as works of art (Stecker 1997: 51). That is, he claims, the right hand side of his ­definition involves the extension of the concept of art, not its intension. This is also true of Levinson’s historical definition. Nevertheless, this does not undermine the charge of circularity. As Stecker himself accepts, an adequate definition of art should give us the extension of “art” (Stecker 1997: 14). We measure the adequacy of any account of the intension of art according to its ability to accommodate our intuitions about its extension.We then adjust our hypotheses about its intension and its extension in order to establish a reflective equilibrium between the two. However, this ­process of mutual adjustment of hypotheses can proceed only on the assumption that the two hypotheses are independent of one another. Neither Stecker’s nor Levinson’s accounts of the intension of art can be evaluated in this way because both rely on a prior grasp of the extension of “art”. David Davies argues that the circularity of existing institutional accounts of art can be avoided by providing an account of the function that art institutions perform, such that the resultant account appeals, not to art institutions, but to institutions that perform a certain function (D. Davies 2004). Because an institution can perform a function without every one of the individual works to which it gives rise performing that function, this approach also has the potential to explain the value of art by elucidating the function that art institutions perform, without adopting the implausible view that there is any single function that all artworks (or all good artworks) perform. On David Davies’s account, artworks are the products of institutional processes, not because they are the result of declarative acts of art status conferral, but because artworks are made by manipulating an artistic medium, which is institutional in the broad sense that it consists in a system of shared understandings embodied in the practices of a community of receivers (D. Davies 2004: 245). In elaborating what is artistic about this medium, he draws on Nelson Goodman’s claim that all artworks function as symbols of a kind characterised by syntactic density, semantic density, relative repleteness, exemplification, and multiple and complex reference (Goodman 1976: 67–8). Davies denies that

all artworks function as symbols of this kind and thus that Goodman’s account is correct as it stands (D. Davies 2004: 250), but claims instead that a medium counts as artistic in virtue of facilitating the production of such symbols (D. Davies 2004: 253). That is, an artistic medium consists in a set of shared understandings concerning the symbolic ­significance attributable to certain features, such that members of the relevant community of receivers can draw on those understandings to produce symbols of the kind characterised by Goodman. The resultant descriptive account avoids the functionalist claim that all artworks function as symbols of this kind by allowing that not every work produced in a medium that facilitates the production of such symbols need be such a symbol (D. Davies 2004: 251). While this account does not require all artworks to function as symbols of the kind Goodman identifies, it does require them to function as symbols of some kind or another. However, there are some artworks that appear to lack symbolic content altogether. This problem is compounded by the fact that Davies grounds the symbolic functioning of artworks in the shared understandings of a community of receivers. An adequate response to the problem must not only show that all artworks have symbolic content, but also that their contents are accessible to the community of receivers on whose shared understandings they rely. Moreover, as David Davies acknowledges, this account does not sharply distinguish artworks from all other things (D. Davies 2004: 253).There are various different systems of shared understandings that facilitate the production of symbols, and the symbols produced by manipulating any such system will possess the features Goodman identifies to varying degrees. It is not clear to what degree they must do so in order for the system at issue to count as an artistic medium. Davies claims that such vagueness is desirable given our uncertainty regarding whether or not such practices as carpet weaving and pottery count as artistic (D. Davies 2004: 253). However, while “art” is undoubtedly a vague term, it is not clear that Davies correctly locates this vagueness. Insofar as woven carpets and pottery function as symbols, they rarely possess any of the features Goodman identifies. Moreover, the account construes a variety of other, clearly non‐artistic practices as falling within the vague extension of the term “art”. For example, as Goodman makes clear, all pictorial symbols are both syntactically and semantically dense, and relatively replete (Goodman 1976). They also often exemplify the properties they represent, as with coloured pictures that represent colours by instantiating them. Nevertheless, it is implausible



art: what it is and why it matters

that every product of a pictorial system is an artwork. An adequate account of art should reflect this fact, by distinguishing artworks from mere coloured pictures. Berys Gaut claims that it is impossible to provide individually necessary and jointly sufficient conditions for being a work of art (Gaut 2000). Art, he claims, is a “cluster concept”, admitting of various different sufficiency conditions, but no non‐disjunctive necessary conditions. He claims that possession of each of the following ten properties counts towards something’s being an artwork: having positive aesthetic properties, such as beauty, grace or elegance; ­ expressing emotion; being intellectually challenging; having formal complexity or coherence; being able to convey complex meanings; exhibiting an individual point of view; being original; being the product of a high degree of skill; belonging to an established artistic form; and being the product of an intention to make an artwork (Gaut 2000: 28). Because many of these are the properties of performing certain functions, Gaut’s approach promises to help elucidate the value of art, while avoiding the claim that there is any single function in which its value resides. However, the final two properties make reference to art, thus rendering Gaut’s account circular. Moreover, it is not obvious that it will be possible to identify all the different sufficiency ­conditions for being an artwork by appeal solely to the properties Gaut identifies. Arguably, there are further properties to which some such conditions will have to appeal. Even if it were possible to elucidate a disjunctive necessary condition for being an artwork by appeal solely to the properties Gaut describes, it is not obvious possession of which particular subsets of the properties just listed suffices for something’s being an artwork. A work of philosophy, for example, can be elegant, convey complex meanings, be intellectually challenging, be original and be the product of a high degree of skill. Because he does not provide a comprehensive list of the different sufficiency conditions for art status, Gaut cannot tell us whether or not such a work of philosophy is a work of art. Nevertheless, if he is right about the nature of the concept, the way to answer this question is by providing such a list, rather than by identifying necessary and sufficient conditions for being an artwork. I agree with Gaut that art is characterised by a variety of different functions, none of which is individually necessary for art status, but think that it is nonetheless possible to identify individually necessary and jointly sufficient conditions for being an artwork. In what follows, I will propose an institutional account which provides such conditions. Like David Davies, I will try to avoid the circularity of existing institutional accounts by identifying a feature

59

d­ istinctive of art institutions, which can be specified independently of the notion of art. To see the form that an adequate institutional definition of art must take, it is ­crucial to begin by briefly examining the nature of institutions and institutional facts.

4.  Institutions and Institutional Facts According to John Searle, any fact involving the collective intentionality of two or more agents is a social fact (Searle 1995). Collective intentional states are intentional states— including intentions, beliefs and desires—that two or more people share. For example, the members of an orchestra may collectively desire to play a piece of music. Institutional facts are a subclass of social facts that meet two further conditions. Firstly, they result from the collective assignment of function to an object that does not have that function intrinsically, but has an observer‐relative function as a result of that function’s being assigned. For example, nothing has the function of being an ironing board intrinsically. Rather, things are ironing boards because we collectively assign that function to them. Secondly, the functions assigned are status functions: functions that the objects to which they are assigned cannot perform simply in virtue of their physical structure. Whereas something can perform the function of being an ironing board because it has a certain physical structure, things can perform status functions only because a community collectively accepts that they have the requisite status. For example, pieces of paper with a certain pattern on them function as money, not because of their physical structure, but only because a community collectively accepts that those pieces of paper have the status of money. Status functions typically have the form X counts as Y in context C, where the X term identifies certain features of an object, person or state of affairs, and the Y term assigns a special status to those features. For example, putting a ball through a hoop during a game of basketball counts as scoring a goal in certain contexts. Searle argues that money, private property, and political leadership all involve the assignment of status functions—in all these cases, things acquire a function which can be performed only because the corresponding function is collectively accepted. When the practice of assigning a certain status function becomes regularised, Searle argues, the practice of counting X as Y in C becomes a rule of the form X counts as Y in C. These are called constitutive rules. These rules are constitutive of institutions: an institution, on Searle’s account, is simply a system of such rules. These systems may have an iterative

60

catharine abell

structure: the Y term in one status function can serve as the X term in another, the Y term of which can serve as the X term in yet another status function, and so on. Constitutive rules contrast with regulative rules, which regulate activities whose existence is independent of the rules in question. Whereas the rule in a fast food restaurant requiring diners to dispose of their own rubbish regulates the antecedently existing activity of dining, the rule according to which placing one’s chess pieces in a certain relation to one’s opponent’s king counts as checkmate does not regulate an antecedently existing activity of chess playing, but is instead partly constitutive of playing chess. Without the system of constitutive rules of which this rule is a part, there would be no such thing as chess. This account construes the essential role of human institutions as being to create deontic powers. Institutions create rights, duties, obligations, authorizations, permis­ sions, and empowerments. These powers can be directly assigned, as when an agent is given a certain authority, or they can be indirectly assigned, as when something is deemed a five dollar bill, such that those who possess it are thereby empowered to use it to buy things (Searle 1995: 98). Institutions therefore create desire‐independent reasons for action. When we recognise something as a duty or obligation, we recognise a reason for doing that thing that is ­independent of our current inclinations.

5. The Function of Art Institutions Searle’s account of institutions shows that their very existence depends on their being seen to perform some function. In order for people collectively to assign a set of status functions with sufficient regularity for a system of constitutive rules to emerge, with its concomitant deontic powers, they must see that system of rules as performing some function they value. If they did not do so, they would not accept the rules at issue or recognise them as imposing duties and obligations. People may be prepared to accept individual rules that they do not see as performing any valuable function, and may recognise them as creating desire‐independent reasons for action, so long as those rules belong to a system that they perceive, as a whole, as serving some valuable function. However, a system of ­constitutive rules will not emerge unless it is perceived as performing such a function. The different participants in an institution may perceive it as performing different functions. For example, some people see the law as method of social control, while others

perceive it as promoting distributive justice. An institution may emerge even though its various participants attribute quite different functions to it, so long as they all attribute the functions at issue to the same system of constitutive rules. What is essential is not that an institution have ever actually performed a humanly valuable function, but that its participants believed it to do so. An institution that never actually performs any valuable function may nonetheless come into existence so long as its participants believe it to perform such a function or functions. Once an institution has come into existence, it may ­persist even when it ceases to be perceived by its participants to perform any valuable function. It may do so, for example, if it is artificially propped up by physical threat. A non‐coercive institution may also persist in such circumstances if the duties and obligations it imposes are not too onerous and its participants are apathetic. Nonetheless, an institution that ceases to be perceived to perform the functions to which its existence is due will often cease to exist, because in such a case its participants will no longer have any reason to accept the duties and obligations it imposes. There is thus a general connection between an institution’s existence and its perceived performance of certain functions: it would not have come into existence in the first place unless its participants had seen it as performing some valuable function or functions, and its continued existence is often due to its continuing to be perceived to perform those functions. The perceived functions to which an institution’s existence is due determine the type of institution it is. An institution is a sporting institution iff its existence is due to its participants’ belief that it promotes competition in the exercise of a physical skill. Sporting institutions may persist long after they cease to be perceived to perform such a function (consider the institution of wrestling). However, if wrestling had never been perceived to perform this function, it would not be a sporting institution. This suggests that the features distinctive of all art institutions at all times are the perceived functions to which their existence is due. David Davies’s approach to defining art is right insofar as he seeks to avoid the circularity of existing institutional accounts by identifying the function of art institutions. However, he is wrong both to suppose that the actual, rather than the perceived functions of an institution are what determine its type and that there is a single such function art institutions perform. The existence of an art institution may be due to its being perceived to perform a variety of different functions.



art: what it is and why it matters

This suggests the strategy of providing a non‐circular institutional definition of art by analysing art institutions as institutions whose existence is due to their being believed by their participants to perform various functions. Following Gaut’s specification of the functions the performance of which counts towards something’s being an ­artwork, this strategy yields the following definition: Something is an art institution iff it is an institution whose existence is due to its being perceived to perform certain functions, and these functions form a significant subset of the ­following: promoting positive aesthetic properties; promoting the expression of emotion; facilitating the posing of intellectual challenges; promoting formal complexity and coherence; facilitating the communication of complex meanings; promoting the exhibition of individual points of view; promoting originality; and promoting the exercise of a high degree of skill.

Whereas Gaut must specify the various specific subsets of properties possession of which suffices to make something an artwork if he is to distinguish artworks from works of philosophy and other things that possess a subset of the properties he identifies, I do not need to do so. Let us call those things—such as artworks—that result from the collective assignment of status functions according to the constitutive rules of some institution products of that institution. Construing artworks as the products of art institutions, as defined above, suffices to distinguish artworks from works of philosophy. Unlike artworks, works of philosophy do not result from the collective assignment of status functions, and thus are not the products of an institution. Furthermore, even supposing that philosophy is an institution of which there are products, its existence is not due to its participants’ belief that it performs a significant subset of the functions identified above. Rather, it owes its existence to its participants’ belief that it performs such functions as helping to answer fundamental questions about the nature of reality and our place in it. This is not to deny that, on the above definition, it is vague whether or not certain institutions are art institutions. However, unlike David Davies’s account of what it is to be an art institution, the definition above correctly locates the vagueness of the term “art”. It is unclear whether or not the institutions of carpet weaving and pottery are art institutions because, while their existence is due to their participants perceiving them to promote positive aesthetic properties, formal complexity and coherence, and the exercise of a high degree of skill, it is not clear whether this set of functions is a sufficiently significant subset of the functions identified above for them to qualify as art institutions.

61

6.  A New Institutional Definition of Art On the approach I am advocating, the notion of an artwork is essentially institutional, because facts concerning which things are artworks are institutional facts. However, one cannot define artworks simply as the products of institutions whose existence is due to their participants’ belief that they perform the functions identified above. Such a definition would distinguish artworks from exhibition catalogues and instruction manuals for artists, since such things are not the products of art institutions in the relevant sense. They do not result from the collective assignment of status functions according to the constitutive rules of art institutions. Whether or not something is an exhibition catalogue or an instruction manual for artists depends, not on the collective intentional states of participants in art institutions, but on its semantic features and the intentions with which it was produced. Nevertheless, art institutions have a variety of products other than artworks. For example, artists and art critics may both figure as Y terms in the constitutive rules of art institutions. To provide a reductive definition of art, rather than of any of these other institutional notions, we need a means of distinguishing institutional facts about a­rtworks from the other institutional facts to which art institutions give rise, without appeal to the notion of art. The variety of different functions that art institutions are perceived to perform suggests that they incorporate a range of different constitutive rules, each with the form X counts as an artwork in C, with different X and C terms. There is a plurality of different sufficient conditions for being an artwork, grounded in this plurality of constitutive rules. Just as football incorporates a variety of different constitutive rules assigning the status function of scoring a point to players’ performance of different tasks, so too there are different constitutive rules of art institutions that assign the status function of being an artwork to things with different properties. While the existence of one rule may be due to its assigning the status function of being an artwork to things that are perceived to have one function, another may assign that status to things that are perceived to have some other function. Institutional facts about artworks directly affect how well art institutions perform the functions in terms of which they are characterised. Depending on the kinds of things that count as artworks according to its constitutive rules, an art institution will do a better or worse job of performing those functions. Artworks that perform one or more of the functions at issue have a positive effect on how well the art

62

catharine abell

institution of which they are products performs these ­functions, whereas artworks that do not do so have a negative effect on how well it performs them. Moreover, the effects that institutional facts about artworks have on how well art institutions perform these functions are direct: they do not depend on further institutional facts. On the assumption that, according to the constitutive rules of at least some art institutions, artists are people who make ­artworks, what counts as an artwork affects who counts as an artist according to the constitutive rules of those art institutions.What counts as an artwork may therefore affect the further institutional facts to which the institutions of which they are products give rise. Nevertheless, simply in virtue of either performing or of failing to perform the functions in terms of which art institutions are characterised, all artworks affect how well the institutions of which they are products perform those functions, whether or not they affect the further institutional facts to which those institutions give rise. By contrast, institutional facts about artists and art critics do not directly affect the ability of art institutions to ­perform these functions. Who counts as an artist and who counts as a critic according to the constitutive rules of an art institution has no direct bearing on how well that institution performs the functions that make it an art institution. Institutional facts about artists and critics affect its ability to perform those functions only indirectly, by affecting the further institutional facts to which it gives rise. More specifically, they do so by affecting what counts as an artwork according to the constitutive rules of that institution. Who counts as a critic may influence what counts as an artwork either in virtue of the existing constitutive rules of an art institution (if, for example, those rules hold that artworks are things that art critics deem worthy of appreciation), or because critics influence the nature of those of its constitutive rules that determine what counts as an a­ rtwork. This happens when, for example, the opinions critics express regarding which artworks are good and which of their features account for their being good influence subsequent practices within that institution, such that its participants come regularly to count works with these features as artworks, leading to the emergence of a new constitutive rule providing sufficient conditions for being an artwork. In either case, who counts as an art critic according to the constitutive rules of that institution affects the institution’s ability to perform the functions characteristic of art institutions only indirectly, by influencing what counts as an artwork according to its constitutive rules.

I propose the following institutional definition of art: Something is an artwork iff it is the product of an art institution, and it directly affects how effectively that institution ­performs the perceived functions to which its existence is due.

This is a reductive definition of art. Its reference to an art institution can be removed simply by replacing “art institution” with the definition provided earlier. This definition accommodates both readymade artworks and artworks comprising objects that are deliberately manufactured so as to fall within the scope of the constitutive rules of an art institution. To say that something is the product of an art institution is not to say anything about how it came to have its physical composition or structure. Duchamp’s ­artwork Fountain is the product of an art institution, despite comprising a readymade urinal that was manufactured by people who were not acting as participants in an art institution because, according to the constitutive rules of that institution, the urinal in question counts as an artwork. It also accommodates the existence of artworks that do not perform any of the functions in terms of which I characterised art institutions. Artworks that do not perform any of these functions have a direct adverse effect on how well the institution of which they are products performs those functions. Institutional theories of art are often criticised for failing to accommodate as artworks artefacts produced by isolated individuals operating independently of any institutional context. Like other institutional theories, the present account denies that such works are artworks. Just how counterintuitive is this? Note that we are not tempted to take just any artefacts produced by an isolated individual to be artworks. The most plausible candidates for extra‐institutional artworks are those we believe to perform a significant subset of the functions in terms of which I defined art institutions. However, as I noted when discussing Gaut’s position, it is very difficult to specify just which of these functions something must perform in order to be an artwork, as the example of the work of philosophy demonstrates. The present account explains this difficulty: something’s performing some subset of the functions identified by Gaut never suffices to make it art. It must also be embedded in an appropriate institutional context. This strikes some as counterintuitive because they assume that the institutional conditions required for it to acquire art status are unduly onerous. Previous institutional accounts have implied that, for proto‐art to yield art proper, a formalised institutional structure with explicitly encoded rules and authority roles must be established. However, Searle’s account of institutions shows that institutions do



art: what it is and why it matters

not require explicitly encoded rules, but merely regularities in the collective assignment of status functions. Although an isolated individual who produces artefacts that perform a subset of the functions Gaut identifies does not produce art, because she does not engage in an institutional practice, if others became aware of her activity, came collectively to assign status functions to the artefacts she produced and to do so with sufficient regularity for constitutive rules to emerge, and did so because they saw those artefacts as performing the functions at issue, they would count as artworks.

7.  Artistic Value Artworks are good as artworks—they have artistic value— insofar as they have the tendency to improve how well the institutions of which they are products perform the functions that make them art institutions. The greater the improvement they have the tendency to make to these institutions’ performance of those functions, the greater their artistic value. While the constitutive rules of art institutions are determined by their participants’ beliefs about the nature of the functions at issue, a work’s artistic value is determined by its actual tendency to improve how effectively the institution of which it is a product performs the functions at issue. A work’s artistic value therefore depends on what it actually is for something to perform the functions at issue, rather than on widespread beliefs about what it is for it to do so. An artwork’s tendency to improve how effectively an institution performs these functions is determined by both its direct and its indirect effects on the institution’s performance of them. All artworks have a direct positive or negative influence on how effectively the institutions of which they are products perform those functions. However, they may also have a range of either actual or potential indirect effects on its ability to perform them. Institutional facts about which things are artworks may indirectly influence how effectively an institution performs these functions by influencing what ­ other things count as artworks according to its constitutive rules. For example, an artwork may indirectly improve how effectively the institution of which it is a product promotes positive aesthetic properties because it introduces novel techniques for producing artefacts with positive aesthetic properties that are subsequently widely adopted within that institution, leading to the emergence of constitutive rules which accord art status to works produced by those techniques. Whether or not a given artwork actually has an indirect positive influence on an institution’s effectiveness in ­performing the functions characteristic of art institutions

63

depends on contingent facts about how that artwork influences the institution’s subsequent development. However, its tendency to influence how effectively that institution ­performs those functions—and thus its artistic value—is independent of such contingencies. The history of a work’s effect on the institution that produced it may illuminate its artistic value by providing evidence of its tendency to influence that institution’s performance of those functions, but its artistic value is independent of that history, which may be determined by a range of things other than the work’s tendency to influence the institution’s ability to ­perform those functions. This account of what it is for an artwork to be good as an artwork does not construe only works that perform the functions characteristic of art institutions as artistically ­valuable. An artwork may indirectly improve how effectively the institution of which it is a product performs these functions without itself performing any of them. For example, a work that introduces novel techniques that can be used to produce artefacts with positive aesthetic properties need not itself have such properties, or indeed perform any of the other functions characteristic of art institutions. Nevertheless, because it has the tendency positively to influence how effectively the institution of which it is a product performs the function of promoting positive ­aesthetic properties, the work has artistic value. Neither all artworks nor all good artworks need possess any of the properties Gaut identifies. The only property that unites all good artworks is that of having the tendency to enhance an institution’s performance of the functions in terms of which I have characterised art institutions. Avant garde works may have a range of indirect positive influences on how effectively the institutions of which they are products perform these functions. For example, they may help to clarify the nature of the functions at issue. The nature of the constitutive rules of an art institution depends on its participants’ beliefs about the nature of those functions, and these beliefs may be incorrect. Avant garde works may have artistic value although they do not perform any of the functions at issue because they shed light on the nature of those functions, and thereby indirectly improve the effectiveness with which art institutions perform them. Likewise, self‐reflexive works about the nature of art itself may help to illuminate the nature of art institutions and clarify the means by which they can perform their characteristic functions. In so doing, they may indirectly improve the effectiveness with which the institutions that produced them perform the functions at issue, even though they do not themselves perform any of these functions.

64

catharine abell

Every artwork that performs one or more of the functions characteristic of art institutions improves the effectiveness with which the institution that produced it performs those functions. Consequently, every such artwork has some degree of artistic value. Nevertheless, their artistic value may be negligible. For example, an artwork with positive aesthetic properties produced using hackneyed techniques will have little tendency indirectly to influence how effectively the institution of which it is a product performs the functions characteristic of art institutions and will therefore have very limited artistic value. By contrast, an artwork which lacks positive aesthetic properties but introduces a new technique that can be used to produce artefacts with positive aesthetic properties and could give rise to new constitutive rules assigning art status to works with positive aesthetic properties produced using those techniques may have much greater artistic value. A work’s artistic value is not a straightforward measure of its possession of the properties Gaut identifies, but rather a measure of its ability to improve the institutionalised pursuit of such properties.

Conclusion The problem of defining art has seemed intractable because it has consistently been approached in the wrong way. Searle’s account of institutions shows that their existence depends on the internal points of view of their participants. Institutions depend on collective intentionality and the assignment of status functions, and such facts are ontologically subjective: their existence depends on the first person points of view of those who collectively assign the status functions at issue (Searle 1995: 98). From the external point of view of someone who does not accept the desire‐independent reasons for action provided by the deontic structure of an institution, such as that of an anthropologist examining the institutions of an alien culture, the existence of the institution can be discerned only by identifying the collective intentional states of its participants. Only if we recognise the collective intentional states of participants in art institutions can we determine either what it is for something to be art, or what it is for it to be good art.

Previous functionalist accounts have assumed that the function that good artworks perform must be identifiable from an external point of view, independently of the intentional states of participants in art institutions. They have thus failed to identify any function shared by all examples of good art. By acknowledging the dependence of art institutions on the collective intentional states of their participants and the dependence of those states on the perceived function of art institutions, I have identified a function that all good artworks have in common: that of tending to ameliorate the institutionalised pursuit of the functions to the perceived performance of which art institutions owe their existence. Previous institutional accounts have failed to acknowledge that art institutions come into existence because their participants believe them to perform valuable functions. This has prevented them both from providing a reductive analysis of art, and from explaining the value of art. The definition that I have proposed overcomes the circularity of previous institutional accounts of art by appealing to the functions art institutions are perceived to perform. Moreover, it explains the value of good art by appeal to its capacity to improve art institutions’ performance of these functions. Why should one accept the duties and obligations imposed by art institutions, when one could instead pursue these functions independently of any institution? The answer resides in the enormous social power that institutions can achieve. An individual who pursues such functions in an extra‐institutional context can at best make a small contribution to their prominence. By contrast, one who participates in an institution that performs them can influence both whether or not and how whole societies are educated about these functions; can induce governments to fund programmes dedicated to their performance; and can influence what vast groups of people in different geographical and temporal contexts believe about their nature and importance. Although they may perform the functions in terms of which they are characterised imperfectly, art institutions can nonetheless perform them more effectively than individuals can do.

References Bell, Clive. 1914. Art. London: Chatto and Windus. Craig, Edward. 1990. Knowledge and the State of Nature: An Essay in Conceptual Synthesis. Oxford: Clarendon Press. Danto, Arthur. 2000. “Art and Meaning”. In Carroll, ed, Theories of Art Today. Madison: University of Wisconsin Press.

Davies, David. 2004. Art as Performance. Malden, MA: Blackwell. Davies, Stephen. 1991. Definitions of Art. Ithaca: Cornell University Press. Dickie, George. 1969. “Defining Art,” The American Philosophical Quarterly 6: 253–6.



art: what it is and why it matters

Dickie, George. 1971. Aesthetics:An Introduction. Indianapolis: Pegasus. Dickie, George. 2000. “The Institutional Theory of Art”. In Carroll, ed, Theories of Art Today. Madison: University of Wisconsin Press. Gaut, Berys. 2000. “‘Art’ as a Cluster Concept”. In Carroll, ed, Theories of Art Today. Madison: University of Wisconsin Press. Goodman, Nelson. 1976. Languages of Art (2nd edn.). Indianapolis: Hackett Publishing Company.

65

Levinson, Jerrold. 1979. “Defining Art Historically,” British Journal of Aesthetics 19: 232–50. Searle, John. 1979. Expression and Meaning: Studies in the Theory of Speech Acts. Cambridge: Cambridge University Press. Searle, John. 1995. The Construction of Social Reality. London: Penguin. Stecker, Robert. 1997. Artworks: Definition, Meaning, Value. Pennsylvania: Pennsylvania State University Press.

Part II Ontology of Art

Introduction

Among works of art are paintings, sculptures, prints, ­photographs, films, poems, novels, plays, buildings, ballets, symphonies, operas, and other more exotic species of ­multimedia productions, installations, and conceptual art. Is there any single ontological category to which all such items belong? Do they all share au fond the same “mode of existence,” in virtue of being works of art? On the face of it, it might seem impossible to find unity amidst such apparent diversity. Even the ways we talk about different kinds of works suggest the differences go deep. Of dance, music, and drama we speak of performances and we recognize that there could be strikingly different performances of the same work. Poems and novels have copies which allow us to read and enjoy them without ever seeing the original manuscript. Manuscripts, including musical scores, can be lost without the works themselves being lost. Engraved prints and photographs have multiple reproductions, just as films can be reproduced and distributed. Yet paintings and carved (as opposed to cast) sculptures seem to be unique physical objects. Certainly they can be copied but to see a copy is not to see the real thing, which we call the original. Buildings seem sometimes to be unique objects— York Minster, the Taj Mahal—but sometimes they can have multiple instances, as in designs for housing estates. Out of this complexity and diversity analytic philosophers have not been reluctant to propose suitably complex and diverse ontological theories, often of a highly revisionary kind. In a not untypical move, Gregory Currie, for example, has argued that “No work of art [including ­paintings or

sculptures] is a physical object. … A work of art is rather an action type” (Currie 1989, p. 7). But what is the status of claims of that kind? On what basis is their truth determined? Amie Thomasson in her paper here (Chapter 10) addresses such questions and the methodological underpinning of the ontology of art. She rejects what she calls the “discovery model of knowledge,” ­sometimes associated with the natural sciences, in ontological enquiries about art. Rather than searching for mind‐independent facts about what kinds of entities paintings, symphonies, or poems are, we should instead undertake a “conceptual analysis that teases out from our practices and things we say the tacit underlying ontological conception of those who ground the reference of the term.”We must accept too that there will be indeterminacies, questions that admit of no determinate answer (When exactly has a poem been created? How much paint on a canvas must be removed before the painting is destroyed?). These methodological considerations are important but our principal focus in this section is on the ontology of music, to which analytic philosophers have given much attention and where characteristic problems in ontology arise. (We return to other aspects of the aesthetics of music in Part XI.) It is far from straightforward to give a precise account of what kind of entity a work of music is. There seem good reasons not to identify such works with relatively tractable entities like written scores, actual performances, sound waves, or psychological states in the minds of composers or listeners.The thought that they must be abstract entities of a certain kind seems compelling. Jerrold Levinson in his

Aesthetics and the Philosophy of Art – The Analytic Tradition: An Anthology, Second Edition. Edited by Peter Lamarque and Stein Haugom Olsen. © 2019 John Wiley & Sons, Inc. Published 2019 by John Wiley & Sons, Inc.

70 introduction much discussed paper “What a Musical Work Is” (Chapter 7) brings the full weight of analytic p­ hilosophy to bear on the issue, arguing that a musical work cannot be merely a “sound‐structure type” tout court for then no music would be created (pure abstract structures are timeless entities), and there would be no essential link between a work and the person who composed it or the historical context of composition. Instead, he suggests that musical works are sound structure types of a highly c­ omplex kind, not timeless, and tied to the conditions of their origin. Julian Dodd has challenged some fundamental arguments in Levinson’s account, including the idea that musical works are created and thus not timeless. Dodd, in “Defending Musical Platonism” (Chapter  8), argues that musical works exist eternally and “composition can only be a kind of creative discovery.” He engages in a detailed debate with several philosophers in defence of his view, a debate that gives clear focus to the differences between his view and that of Levinson. The issues are complex but Dodd lays out the dialectic plainly and fairly so the two sides in the debate are clearly delineated.

In his paper, “Against Musical Ontology” (Chapter 9), Aaron Ridley also reflects in passing on Levinson’s theory but, as his title suggests, his target is wider, namely, the very sort of enquiry that Levinson and Dodd, like many others, are engaged in. Ridley rejects as “absolutely worthless” philosophical enquiries into the “identity conditions” of a musical work: in particular, in relation to the question whether a performance is truly a performance of that work. The focus on identity conditions, according to Ridley, ignores all the important factors about how musical performances are appreciated (aesthetically), understood, and evaluated. Ridley’s paper can helpfully be paired with Thomasson’s in that both demand a closer look at what might be expected of ontological enquiries into art, what the constraints are, and what gains, if any, are to be had. These are salutary questions for analytic aestheticians as the ontology of art—particularly the ontology of music—has attracted a considerable amount of attention, not least in bringing the philosophy of art closer to mainstream investigations in contemporary (analytical) metaphysics.

References Currie, Gregory (1989) An Ontology of Art, London: Macmillan.

Further reading Bartel, Christopher (2012) “Music Without Metaphysics?” British Journal of Aesthetics 51: 383–98. Davies, David (2004) Art as Performance, Malden, MA: Blackwell. Dodd, Julian (2007) Works of Music: An Essay in Ontology, Oxford: Oxford University Press. Dodd, Julian (2012) “Defending the Discovery Model in the Ontology of Art: a Reply to Amie Thomasson on the Qua Problem,” British Journal of Aesthetics 52: 75–95. Goodman, Nelson (1981) Languages of Art (2nd ed.). Indianapolis: Hackett. Kania, Andrew (2008) “Piece for the End of Time: In Defence  of  Musical Ontology,” British Journal of Aesthetics 48: 65–79. Kivy, Peter (1983) “Platonism in Music: A Kind of Defence,” Grazer Philosophische Studien 19: 109–29.

Lamarque, Peter (2010) Work and Object. Oxford: Oxford University Press. Levinson, Jerrold (1990) Music, Art and Metaphysics. Ithaca, NY: Cornell University Press. Margolis, Joseph (1980) Art and Philosophy: Conceptual Issues in Philosophy. Atlantic Highlands, NJ: Humanities Press. Stecker, Robert (2009) “Methodological Questions about the Ontology of Music,” Journal of Aesthetics and Art Criticism 67: 375–86. Thomasson, Amie (2014) Ontology Made Easy. Oxford: Oxford University Press. Wollheim, Richard (1980) Art and Its Objects (2nd ed.). Cambridge: Cambridge University Press. Wolterstorff, Nicholas (1980) Works and Worlds of Art. Oxford: Clarendon Press.

7

What a Musical Work Is Jerrold Levinson

What exactly did Beethoven compose? That is the question I will begin with. Well, for one, Beethoven composed a quintet for piano and winds (flute, oboe, clarinet, horn) in E‐flat, Opus 16, in 1797. But what sort of thing is it, this quintet which was the outcome of Beethoven’s creative activity? What does it consist in or of? Shall we say that Beethoven composed actual sounds? No, for sounds die out, but the quintet has endured. Did Beethoven compose a score? No, since many are familiar with Beethoven’s composition who have had no contact with its score.1 Philosophers have long been puzzled about the identity or nature of the art object in nonphysical arts, e.g., music and literature. In these arts  –  unlike painting and sculpture – there is no particular physical “thing” that one can plausibly take to be the artwork itself. This puzzlement has sometimes led philosophers (e.g., Croce) to maintain that musical and literary works are purely mental – that they are in fact private intuitive experiences in the minds of ­composers and poets. But this does not seem likely, since experiences can be neither played nor read nor heard. More generally, the Crocean view puts the objectivity of musical and literary works in dire peril  –  they become inaccessible and unsharable. Fortunately, however, there is a way of accepting the nonphysicality of such works without undermining their objectivity. Those familiar with recent reflection on the ontological question for works of art will know of the widespread

c­ onsensus that a musical work is in fact a variety of abstract object – to wit, a structural type or kind.2 Instances of this type are to be found in the individual performances of the work. The type can be heard through its instances, and yet exists independently of its instances. I believe this to be basically correct. A piece of music is some sort of structural type, and as such is both nonphysical and publicly available. But what sort of type is it? I aim in this paper to say as precisely as I can what structural type it is that a musical work should be identified with. The most natural and common proposal on this question is that a musical work is a sound structure – a structure, sequence, or pattern of sounds, pure and simple.3 My first objective will be to show that this proposal is deeply unsatisfactory, that a musical work is more than just a sound structure per se. I will do this by developing three different objections to the sound‐structure view. In the course of developing these objections, three requirements or desiderata for a more adequate view will emerge.The rightness – or at least plausibility – of those requirements will, I think, be apparent at that point. My second objective will then be to suggest a structural type that does satisfy the requirements, and thus can be identified with a musical work.4 At the outset, however, I should make clear that I am confining my inquiry to that paradigm of a musical work, the fully notated “classical” composition of Western culture, for example, Beethoven’s Quintet for piano and winds in

Jerrold Levinson, “What a Musical Work Is,” Journal of Philosophy 77 (1980), pp. 5–28. Reproduced by permission of The Journal of Philosophy.

Aesthetics and the Philosophy of Art – The Analytic Tradition: An Anthology, Second Edition. Edited by Peter Lamarque and Stein Haugom Olsen. © 2019 John Wiley & Sons, Inc. Published 2019 by John Wiley & Sons, Inc.

72

jerrold levinson

E‐flat, Opus 16. So when I speak of a “musical work” in this paper it should be understood that I am speaking only of these paradigm musical works, and thus that all claims herein regarding musical works are to be construed with this implicit restriction.

I The first objection to the view that musical works are sound structures is this. If musical works were sound structures, then musical works could not, properly speaking, be created by their composers. For sound structures are types of a pure sort which exist at all times. This is apparent from the fact that they – and the individual component sound types5 that they comprise  –  can always have had instances.6 A sound event conforming to the sound structure of Beethoven’s Quintet Opus 16 logically could have occurred in the Paleozoic era.7 Less contentiously, perhaps, such an event surely could have taken place in 1760  –  ten years before Beethoven was born. But if that sound structure was capable of being instantiated then, it clearly must have existed at that time. Beethoven’s compositional activity was not necessary in order for a certain sound‐structure type to exist. It was not necessary to the possibility of certain sound events occurring which would be instances of that structure. Sound structures per se are not created by being scored – they exist before any compositional activity. Sound structures predate their first instantiation or conception because they are possible of exemplification before that point.8 So, if ­composers truly create their works – i.e., bring them into existence – then musical works cannot be sound structures. We can also defend the pre‐existence of pure sound structures (i.e., existence prior to any instantiation or ­conception) in a somewhat different manner.We need only remind ourselves that purely sound structures are in effect mathematical objects  –  they are sequences of sets of sonic elements. (Sonic elements are such as pitches, timbres, durations, etc.) Now if the pre‐existence of simple sonic element types be granted  –  and I think it must be  –  it ­follows automatically that all sets and all sequences of sets of these elements also pre‐exist. Therefore pure sound structures are pre‐existent. But if pure sound structures pre‐exist, then it is not open for them to be objects of creational activity. So again, if composers are truly creators, their works cannot be pure sound structures.9 But why should we insist that composers truly create their compositions? Why is this a reasonable requirement?

This question needs to be answered. A defense of the desideratum of true creation follows. The main reason for holding to it is that it is one of the most firmly entrenched of our beliefs concerning art. There is probably no idea more central to thought about art than that it is an activity in which participants create things  – these things being artworks. The whole tradition of art assumes art is creative in the strict sense, that it is a godlike activity in which the artist brings into being what did not exist ­beforehand – much as a demiurge forms a world out of inchoate matter. The notion that artists truly add to the world, in company with cake‐bakers, house‐builders, law‐ makers, and theory‐constructers, is surely a deep‐rooted idea that merits preservation if at all possible. The suggestion that some artists, composers in particular, instead merely discover or select for attention entities they have no hand in creating is so contrary to this basic intuition regarding artists and their works that we have a strong prima facie reason to reject it if we can. If it is possible to align musical works with indisputably creatable artworks such as paintings and sculptures, then it seems we should do so. A second, closely related reason to preserve true creation vis‐à‐vis musical works is that some of the status, significance, and value we attach to musical composition derives from our belief in this. If we conceive of Beethoven’s Fifth Symphony as existing sempiternally, before Beethoven’s compositional act, a small part of the glory that surrounds Beethoven’s composition of the piece seems to be removed. There is a special glow that envelops composers, as well as other artists, because we think of them as true creators. We marvel at a great piece of music in part because we marvel that, had its composer not engaged in a certain activity, the piece would (almost surely) not now exist; but it does exist, and we are grateful to the composer for precisely that. Ecclesiastes was wrong  –  there are ever some things new under the sun, musical compositions being among the most splendid of them – and splendid, at least in part, in virtue of this absolute newness. Shall we then accept the creatability requirement as suggested? Before we do so a last qualm should be addressed. It is open for someone to admit the importance of musical composition being characterized by true creation and yet waive the creatability of works themselves. Such a person will point to entities associated with the compositional process which composers unequivocally bring into existence – e.g., thoughts, scores, performances – and claim that true creation need be extended no further. Now it is certainly true that these entities are strictly created, and we may also accord composers some recognition of their c­ reativity in



what a musical work is

regard to these things. But the fact of the matter remains that works are the main items, the center and aim of the whole enterprise, and that since musical works are not identical with scores, performances, or thoughts,10 if those are the only things actually created, then much is lost. “Composers are true creators” acquires a hollow ring. Creation in music shrinks to an outer veneer with no inner core. I propose then that a most adequate account of the musical work should satisfy the following requirement, that of creatability11:

73

II

least the following: (a) the whole of cultural, social, and political history prior to t,13 (b) the whole of musical development up to t, (c) musical styles prevalent at t, (d) dominant musical influences at t, (e) musical activities of P’s contemporaries at t, (f) P’s apparent style at t, (g) P’s musical repertoire14 at t, (h) P’s oeuvre at t, (i) musical influences operating on P at t. These factors contributing to the total musico‐historical context might be conveniently divided into two groups, a–d and e–i. The former, which we could call the general musico‐historical context, consists of factors relevant to anyone’s composing at t; the latter, which we could call the individual musico‐­ historical context, consists of factors relevant specifically to P’s ­composing at t. In any event, all these factors operate to ­differentiate aesthetically or artistically musical works ­identical in sound structure, thus making it impossible to identify those works with their sound structures. I now provide several ­illustrations of this.15

The second objection to the view that musical works are sound structures is this. (1) If musical works were just sound structures, then, if two distinct composers determine the same sound structure, they necessarily compose the same musical work. (2) But distinct composers determining the same sound structure in fact inevitably produce different musical works.12 Therefore, musical works cannot be sound structures simpliciter. The rest of this section is devoted to supporting and elucidating the second premise of this argument. Composers who produce identical scores in the same notational system with the same conventions of interpretation will determine the same sound structure. But the musical works they thereby compose will generally not be the same. The reason for this is that certain attributes of musical works are dependent on more than the sound structures contained. In particular, the aesthetic and artistic attributes of a piece of music are partly a function of, and must be gauged with reference to, the total musico‐historical context in which the composer is situated while c­ omposing his piece. Since the musico‐historical contexts of composing individuals are invariably different, then even if their works are identical in sound structure, they will differ widely in aesthetic and artistic attributes. But then, by Leibniz’s law, the musical works themselves must be non‐ identical; if W1 has any attribute that W2 lacks, or vice versa, then W1 ≠ W2. I will not attempt to give a strict definition of musico‐­ historical context, but will confine myself to pointing out a large part of what is involved in it.The total musico‐historical context of a composer P at a time t can be said to include at

1. A work identical in sound structure with Schoenberg’s Pierrot Lunaire (1912), but composed by Richard Strauss in 1897,would be aesthetically different from Schoenberg’s work. Call it ‘Pierrot Lunaire*’. As a Straussian work, Pierrot Lunaire* would follow hard upon Brahms’s German Requiem, would be contemporaneous with Debussy’s Nocturnes, and would be taken as the next step in Strauss’s development after Also Sprach Zarathustra. As such it would be more bizarre, more upsetting, more anguished, more eerie even than Shoenberg’s work, since perceived against a musical tradition, a field of current styles, and an oeuvre with respect to which the musical characteristics of the sound structure involved in Pierrot Lunaire appear ­doubly extreme.16 2. Mendelssohn’s Midsummer’s Night Dream Overture (1826) is admitted by all to be a highly original piece of music. Music of such elfin delicacy and feel for tone color had never before been written. But a score written in 1900 detailing the very same sound structure as is found in Mendelssohn’s piece would clearly result in a work that was surpassingly unoriginal. 3. Brahms’s Piano Sonata Opus 2 (1852), an early work, is strongly Liszt‐influenced, as any perceptive listener can discern. However, a work identical with it in sound structure, but written by Beethoven, could hardly have had the property of being Liszt‐influenced. And it would have had a visionary quality that Brahms’s piece does not have. 4. The symphonies of Johann Stamitz (1717–1757) are generally regarded as seminal works in the development of orchestral music.They employ many attention‐­getting

(Cre)

Musical works must be such that they do not exist prior to the composer’s compositional activity, but are brought into existence by that activity.

74

jerrold levinson

devices novel for their time, one of which is known as the “Mannheim rocket” – essentially a loud ascending scale figure for unison strings. A symphony of Stamitz containing Mannheim rockets and the like is an exciting piece of music. But a piece written today which was identical in sound structure with one of Stamitz’s symphonies, Mannheim rockets and all, would not be so much exciting as it would be exceedingly funny. Stamitz’s symphony is to be heard in the context of Stamitz’s earlier works, the persistence of late Baroque style, the contemporary activities of the young Mozart, and the Napoleonic wars. “Modern Stamitz”’s symphony would be heard in the context of “Modern Stamitz”’s earlier works (which are probably dodecaphonic), the existence of aleatory and electronic music, the musical enterprises both of Pierre Boulez and of Elton John, and the threat of nuclear annihilation. 5. One of the passages in Bartok’s Concerto for Orchestra (1943) satirizes Shostakovitch’s Seventh Symphony (“Leningrad”) of 1941, whose bombast was apparently not to Bartok’s liking. A theme from that symphony is quoted and commented on musically in an unmistakable manner. But notice that if Bartok had written the very same score in 1939, the work he would then have composed could not have had the same property of satirizing Shostakovitch’s Seventh Symphony. Nor would the work that would have resulted from Shostakovitch’s penning that score in 1943. These examples should serve to convince the reader that there is always some aesthetic or artistic difference between structurally identical compositions in the offing in virtue of differing musico‐historical contexts. Even small differences in musico‐historical context  –  e.g., an extra work in P’s ouevre, a slight change in style dominant in P’s milieu, some musical influence deleted from P’s development as a composer – seem certain to induce some change in kind or degree in some aesthetic or artistic quality, however d­ ifficult it might be in such cases to pinpoint this change verbally. For example, suppose there had been a composer (call him “Toenburg”) in 1912 identical with Schoenberg in all musico‐historical respects  –  e.g., birthdate, country, style, musical development, artistic intentions, etc. – except that Toenburg had never written anything like Verklarte Nacht though he had in his oeuvre works structurally identical with everything else Schoenberg wrote before 1912. Now suppose simultaneously with Schoenberg he sketches the sound structure of Pierrot Lunaire. Toenburg has not produced the same musical work as Schoenberg, I maintain, if

only because his work has a slightly different aesthetic/ artistic content owing to the absence of a Verklarte‐Nacht‐ ish piece in Toenburg’s oeuvre. Schoenberg’s Pierrot Lunaire is properly heard with reference to Schoenberg’s oeuvre in 1912, and Toenberg’s Pierrot Lunaire with reference to Toenburg’s oeuvre in 1912. One thus hears something in Schoenberg’s piece by virtue of resonance with Verklarte Nacht that is not present in Toenberg’s piece  –  perhaps a stronger reminiscence of Expressionist sighs? Before formulating a second requirement of adequacy, as suggested by the fatal problem that contextual differentiation poses for the equation of musical works with pure sound structures, I must confront an objection that may be lurking in the wings. The objection in short is that the aesthetic and artistic differences I have been discussing are not really an obstacle to equating works and sound structures, because these supposed differences between works due to compositional context really just boil down to facts about their composers, and are not attributes of works at all. The objection is understandable, but I find it rather unconvincing for several reasons which I will briefly detail. 1. Artistic and aesthetic attributions made of musical works are as direct and undisguised as attributions typically made of composers. It seems to be as straightforwardly true that the Eroica symphony is noble, bold, original, revolutionary, influenced by Haydn, and reflective of Beethoven’s thoughts about Napoleon, as it is that Beethoven had certain personal qualities, was a genius, changed the course of western music, studied with Haydn, and at one point idolized Napoleon. 2. Whereas we may admit some plausibility to reducing artistic attributions (e.g., ‘original’, ‘influenced by Haydn’) to attributes of persons, there is no plausibility in so reducing aesthetic attributions; it is absurd to maintain that “W is scintillating,” for example, is just a way of saying “W’s composer is scintillating.” 3. Finally, in the case of artistic attributions, not only do they appear as entrenched and legitimate as parallel attributions to composers, but, if anything, they often seem to be primary. Consider originality, for example, and imagine a composer and oeuvre that possess it. Surely the composer is original because his works are original; his works are not original because he is. I thus propose a second requirement  –  that of fine ­individuation – to which any acceptable theory of the ­musical work should conform:



what a musical work is (Ind)

Musical works must be such that composers composing in different musico‐historical contexts17 who determine identical sound structures invariably c­ ompose distinct musical works.

III The third objection to the view that musical works are sound structures is this. If musical works were simply sound structures, then they would not essentially involve any ­particular means of performance. But the paradigm musical works that we are investigating in this paper, e.g., Beethoven’s Quintet Opus 16, clearly do involve quite ­specific means of performance, i.e., particular instruments, in an essential way.The instrumentation of musical works is an integral part of those works. So musical works cannot be simply sound structures per se. Arguments in defense of the claim that performance means are an essential component of musical works now follow. 1. Composers do not describe pure sound patterns in qualitative terms, leaving their means of production undiscussed. Rather, what they directly specify are means of production, through which a pure sound ­pattern is indirectly indicated.The score of Beethoven’s Quintet Opus 16 is not a recipe for providing an instance of a sound pattern per se, in whatever way you might like. Rather, it instructs one to produce an instance of a certain sound pattern through carrying out certain operations on certain instruments. When Beethoven writes a middle C for the oboe, he has done more than require an oboe‐like sound at a certain pitch – he has called for such a sound as emanating from that quaint reed we call an “oboe.” The idea that composers of the last 300 years were generally engaged in composing pure sound patterns, to which they were usually kind enough to append suggestions as to how  they might be realized, is highly implausible. Composers are familiar with tone colors only insofar as they are familiar with instruments that possess them. We do not find composers creating pure combinations of tone color, and then later searching about for instruments that can realize or approximate these aural ­canvases; it would obviously be pointless or at least frustrating to do so. Composers often call for complex sounds that they have never heard before and can scarcely imagine – e.g., the sound of two trombones

75

and three piccolos intoning middle C while four ­saxophones and five xylophones intone the C‐sharp a half‐step above; it is obvious here that what is primarily composed is not a pure untethered sound but an instrumental combination.18 2. Scores are generally taken to be definitive of musical works, at least in conjunction with the conventions of notational interpretation assumed to be operative at the time of composition. It is hard to miss the fact that scores of musical works call for specific instruments in no uncertain terms. When we read in Beethoven’s score the demand ‘clarinet’ (rather, ‘Klarinett’) we may wonder whether a clarinet of 1970 vintage and construction will do as well as one of 1800, but we have still been given a fairly definite idea of what sort of instrument is required. There is nothing in scores themselves that suggests that instrumental specifications are to be regarded as optional – any more than specifications of pitch, rhythm, or dynamics. Nor does the surrounding musical practice of the time encourage such a way of regarding them.19 If we are not to abandon the principle that properly understood scores have a central role in determining the identity of musical works, then we must insist that the Quintet Opus 16 without a clarinet is not the same piece – even if all sound‐structural characteristics (including timbre) are preserved. To feel free to disregard as prominent an aspect of scores as performing means is to leave it open for someone to disregard any aspect of a score he does not wish to conform to  –  e.g., tempo, accidentals, accents, articulation, harmony  –  and claim that one nevertheless has the same work.20 The only way it seems one could justify regarding performing‐means specifications as just optional features of scores is to simply assume that musical works are nothing but sound structures per se.   Consider a sound event aurally indistinguishable from a typical performance of Beethoven’s Quintet Opus 16, but issuing from a versatile synthesizer, or perhaps a piano plus a set of newly designed wind instruments, two hundred in number, each capable of just two or three notes. If performance means were not an integral aspect of a musical work, then there would be no question that this sound event constitutes a performance of Beethoven’s Quintet Opus 16. But there is indeed such a question. It makes perfect sense to deny that it is such a performance on the grounds that the sounds heard did not derive from a piano and four standard woodwinds. We can count something as a performance

76

jerrold levinson

of Beethoven’s Quintet Opus 16 only if it involves the participation of the instruments for which the piece was written – or better – of the instruments that were written into the piece. 3. To regard performing means as essential to musical works is to maintain that the sound structure of a work cannot be divorced from the instruments and voices through which that structure is fixed, and regarded as the work itself. The strongest reason why it cannot be so divorced is that the aesthetic content of a musical work is determined not only by its sound structure, and not only by its musico‐historical context, but also in part by the actual means of production chosen for making that structure audible.The character of a musical composition, e.g., Beethoven’s Quintet Opus 16 for piano and winds, is partly a function of how its sound structure relates to the potentialities of a certain instrument or set of instruments designated to produce that structure for audition. To assess that character ­correctly one must take cognizance not only of the qualitative nature of sounds heard but also of their source of origin. Musical compositions, by and large, have reasonably definite characters; that is to say, we can and do ascribe to them many fairly specific aesthetic qualities. But if prescribed performing forces were not intrinsic to musical compositions, then those compositions would not have the reasonably definite characters we clearly believe them to have. The determinateness of a work’s aesthetic qualities is in peril if performing means are viewed as inessential so long as exact sound structure is preserved.   Consider a musical work W with specified performing means M which has some fairly specific ­aesthetic ­quality ϕ. The sound structure of W as produced by different performing means N, however, will invariably strike us either as not ϕ at all, or else as ϕ to a greater or lesser degree than before. Therefore, if means of sound production are not regarded as an integral part of musical works, then W cannot be said determinately to have the attribute ϕ. So if we wish to preserve a wide range of determinate aesthetic attributions, we must recognize performing means to be an essential component of musical works. I now provide two ­illustrations of this point.21 (a) Beethoven’s Hammerklavier sonata is a sublime, craggy, and heaven‐storming piece of music. The closing passages (marked by ascending chordal trills) are surely among the most imposing and awesome in all music. However, if we understand

the very sounds of the Hammerklavier sonata to originate from a full‐range synthesizer, as opposed to a mere 88‐key piano of metal, wood, and felt, it no longer seems so sublime, so craggy, so awesome. The aesthetic qualities of the Hammerklavier sonata depend in part on the strain that its sound structure imposes on the sonic capabilities of the piano; if we are not hearing its sound structure as p­ roduced by a piano then we are not sensing this strain, and thus our assessment of aesthetic content is altered. The closing passages of the Hammerklavier are awesome in part because we seem to hear the piano bursting at the seams and its keyboard on the verge of exhaustion. On a 10‐octave electronic synthesizer those passages do not have quite that quality, and a hearing of them with knowledge of source is an aesthetically different experience. The lesson here applies, I believe, to all musical works (of the paradigm sort).Their aesthetic attributes always depend, if not so dramatically, in part on the performing forces understood to belong to them. (b) Consider a baroque concerto for two violins, such as Bach’s Concerto in D minor, BWV1043. In such pieces one often finds a phrase (A) assigned to one violin, which is immediately followed by the very same phrase (B) assigned to the other violin. Now when one hears such passages as issuing from two violins (even if in a given performance there are no discernible differences between A and B in timbre or phrasing), a sense of question‐­ and‐answer, of relaxation and unhurriedness is communicated. But if one were to construe such passages as issuing from a single violin, that quality would be absent, and in its place the passages would assume a more emphatic, insistent, and repetitive cast. 4. The dependence of aesthetic attributes on assumed or  understood performing forces should now be apparent. The dependence of artistic attributes is even more plain. (a) Consider Paganini’s Caprice Opus 1, No. 17. This piece surely deserves and receives the attribution ‘virtuosic’. But if we did not conceive of the Caprice No. 17 as essentially for the violin, as inherently a violin piece (and not just a violin‐sounding piece), then it would not merit that attribution. For, as executed by a computer or by some novel string instrument using nonviolinistic technique, its sound structure might not be particularly ­difficult to get through.



what a musical work is (b) Imagine a piece written for violin to be played in such a way that certain passages sound more like a flute than they do like a violin. Such a piece would surely be accounted unusual, and to some degree, original as well. Understood as a piece for violin and occasional flute, however, it might have nothing unusual or original about it at all. Retaining the sound structure while setting actual performance means adrift completely dissolves part of the piece’s artistic import. (c) According to one respected critic, Beethoven in the Quintet Opus 16 was interested in solving problems of balance between piano and winds – a nominally incompatible array of instruments – and succeeded in his own individual way.22 It is not hard to agree with this assessment; thus, ‘solves the problem of balance between piano and winds’ is an attribution true of Beethoven’s Quintet. It is difficult to see how this would be so if the Quintet is purely a sound structure, if piano and winds are not strictly part of the piece at all.23

I thus propose a third requirement for any account of the musical work: inclusion of performance means: (Per)

Musical works must be such that specific means of performance or sound production are integral to them.

IV If musical works are not sound structures simpliciter, then what are they? The type that is a musical work must be capable of being created, must be individuated by context of composition, and must be inclusive of means of performance.The third desideratum is most easily met, and will be addressed first. I propose that a musical work be taken to involve not only a pure sound structure, but also a structure of ­performing means. If the sound structure of a piece is basically a sequence of sounds qualitatively defined, then the performing‐means structure is a parallel sequence of ­performing means specified for realizing the sounds at each point. Thus a musical work consists of at least two structures. It is a compound or conjunction of a sound structure and a performing‐means structure. This compound is itself just a more complex structure; call it an “S/PM” structure,

77

for short.24 Beethoven’s Opus 16 Quintet is at base an S/ PM structure; the means of producing the sounds belonging to it are no more dispensable to its identity as a composition than the nature and order of those sounds themselves. This satisfies requirement (Per). To satisfy the first and second requirements of adequacy we arrived at, it is necessary to realize that a musical work is not a structure of the pure sort at all, and thus not even a S/PM structure  simpliciter. A S/PM structure is no more creatable or context‐ individuated than a sound structure is. I propose that we recognize a musical work to be a more complicated entity, namely this: (MW)

S/PM structure‐as‐indicated‐by‐X‐at‐t

where X is a particular person – the composer – and t is the time of composition. For the paradigmatic pieces we are concerned with, the composer typically indicates (fixes, determines, selects) an S/PM structure by creating a score. The piece he thereby composes is the S/PM structure‐as‐indicated by him on that occasion. An S/PM structure‐as‐indicated‐by‐X‐at‐t, unlike an S/ PM structure  simpliciter, does not pre‐exist the activity of composition and is thus capable of being created. When a composer θ composes a piece of music, he indicates an S/ PM structure  ψ, but he does not bring ψ into being. However, through the act of indicating ψ, he does bring into being something that did not previously exist – namely, ψ‐ as‐indicated‐by‐θ‐at‐t1. Before the compositional act at t1 no relation obtains between θ and ψ. Composition establishes the relation of indication between θ and ψ. As a result of the compositional act, I suggest, the world contains a new entity, ψ‐as‐indicated‐by‐θ‐at‐t1. Let me call such entities indicated structures. And let me represent indicated structures by expressions of form “S/PM*x*t.” It is important to realize that indicated structures are entities distinct from the pure structures per se from which they are derived. Thus, in particular, ψ*θ*t1 is not just the structure ψ with the accidental property of having been indicated by θ at t1 – ψ*θ*t1 and ψ are strictly non‐identical, though of course related. ψ*θ*t1 unlike ψ, can be and is created through θ’s composing. Thus requirement (Cre) is satisfied. Indicated structures also serve to satisfy our second requirement (Ind). If musical works are indicated structures of the sort we have suggested, then two such works, ψ*θ*t1 and α*ϕ*t2 are identical iff (i) ψ = α, (ii) θ = ϕ, and (iii) t1 = t2. But if musical works are necessarily distinct if composed either by different people or at different times, then it ­certainly follows that works composed in different musico‐historical

78

jerrold levinson

contexts will be distinct, since any difference of musico‐historical context from one work to another can be traced to a difference of composer or time or both. Put otherwise, musico‐historical context (as explained in s­ection II) is a function of time and person; given a time and person, musico‐historical context is fixed. So requirement (Ind) is satisfied. That it is satisfied by our proposal with something to spare is a matter I will return to in s­ection V. I now endeavor to increase the reader’s grasp of what indicated structures are. Indicated structures are a different class of type from pure structures. Types of the latter class we may call implicit types, and those of the former class initiated types. Implicit types include all purely abstract structures that are not inconsistent, e.g., geometrical figures, family relationships, strings of words, series of moves in chess, ways of placing five balls in three bins, etc. By calling them “implicit types” I mean to suggest that their existence is implicitly granted when a general framework of possibilities is given. For example, given that there is space, there are all the possible configurations in space; given there is the game of chess, there are all the possible combinations of allowed moves. Sound structures simpliciter are clearly implicit types. Given that there are sounds of various kinds, then all possible ­patterns and sequences of those sounds must be granted existence immediately as well. For a sound structure, in company with all pure structures, is always capable of instantiation before the point at which it is noticed, recognized, mentioned, or singled out. And thus its existence must predate that point. The same goes for a performance‐ means structure  simpliciter. Given performing means (i.e., instruments) of various kinds, then all possible combinations and sequences of such means exist as well.The compound of these two, a sound/performance‐means structure, thus of course also counts as an implicit type. The other class of types, initiated types, are so called because they begin to exist only when they are initiated by an intentional human act of some kind. All those of interest can, I think, be construed as arising from an operation, like indication, performed upon a pure structure. Typically, this indication is effected by producing an exemplar of the structure involved, or a blueprint of it. In so indicating (or determining) the structure, the exemplar or blueprint ­inaugurates the type which is the indicated structure, the structure‐as‐indicated‐by‐x‐at‐t. All indicated structures are, perforce, initiated types. Initiated types include such types as the Ford Thunderbird, the Lincoln penny, the hedgehog. The Ford Thunderbird is not simply a pure structure of metal, glass, and plastic. The

pure structure that is embodied in the Thunderbird has existed at least since the invention of plastic (1870); there could certainly have been instances of it in 1900. But the Ford Thunderbird was created in 1957; so there could not have been instances of the Thunderbird in 1900. The Ford Thunderbird is an initiated type; it is a metal/glass/plastic structure‐as‐indicated (or determined) by the Ford Motor Company on such and such a date. It begins to exist as a result of an act of human indication or determination. The instances of this type are more than just instances of a pure structure – they are instances of an indicated structure. The Lincoln penny is similarly not a pure structure, an abstract pattern tout court, but a structure‐as‐indicated, a pattern‐as‐ denominated‐by‐the‐US Government. Objects conforming to the pattern tout court but existing in ad 100 in Imperial Rome would not be instances of the Lincoln penny. Even the hedgehog is probably best understood, not as a pure ­biological structure, but rather as a biological structure‐­ ­as‐determined‐or‐fixed by natural terrestrial evolution at a particular point in history.The creatures we call “hedgehogs” possess a certain structure and stand in certain causal relations to some particular creatures which came into existence at a given past date. The biological structure of the hedgehog might have been instantiated in the Mesozoic era, or on Uranus, but nothing existing at that time, or at that place, could be an instance of the hedgehog as we understand it. Musical works, as I have suggested, are indicated structures too, and thus types that do not already exist but must instead be initiated. The same is true of poems, plays, and novels – each of these is an entity more individual and temporally bound than the pure verbal structure embodied in it. The distinction between indicated structure and pure structure can perhaps be made clearer by analogy with the distinction between sentence and statement long enshrined in the philosophy of language.25 These distinctions are ­motivated in similar ways. Statements were recognized partly in response to the need for entities individuated in some respects more finely than sentences, in order to provide bearers for the varying truth values that turned up in connection with a given sentence on different occasions.26 Just so, indicated structures are recognized in response to the need for entities more finely individuated than pure structures, in order to provide bearers for various incompatible sets of aesthetic, artistic, cultural, semantic, and genetic ­properties. We allow that a given sentence can make different statements when uttered in different circumstances. Similarly, we realize that a given sound/performance‐means structure yields different indicated structures, or musical works, when indicated in different musico‐historical contexts.27



what a musical work is

V I have proposed that musical works be identified with rather specific indicated structures, in which a particular person and time figure ineliminably.The proposal (MW) was made, recall, in order to satisfy the creatability and individuation requirements. However, as I noted at that point, MW ­satisfies the individuation requirement with logical room to spare. Perhaps both requirements can be satisfied without invoking types that are quite so particularized? The obvious alternative is that a musical work is this sort of type: (MW′) S/PMstructure‐as‐indicated‐in‐musico‐historical‐ context‐C

Such types would be both creatable and sufficiently ­individuated. A type of this sort, like an MW type, comes into existence through some actual indication of an S/PM structure by a person at a time  –  a person who at that ­certain time is situated in a particular context. But the type’s ­identity is not inherently tied to that of any individual as such. Thus, two composers composing simultaneously but independently in the same musico‐historical context who determine the same S/PM structure create distinct MW types, but the same MW′ type. Given these two proposals, then, which satisfy all our desiderata, do we have reason to prefer one or the other? I will discuss one consideration in favor of MW′, and three considerations in favor of MW. 1. On the MW′ proposal, it is at least logically possible for a musical work to have been composed by a person other than the person who actually composed it. If  A  is the actual composer of a musical work, ψ‐as‐indicated‐in‐C1 then all we need imagine is that someone other than A was the person to first indicate the S/PM structure ψ‐in musico‐historical context C1. On the MW proposal, however, it becomes logically impossible for a work to have been composed by other than its actual composer. Could someone else have composed Beethoven’s Quintet Opus 16, according to MW? For example, could Hummel have done so? No, because if ψ is the S/PM structure of the Quintet Opus 16, then all that Hummel might have composed is ψ‐as‐indicated‐by‐Hummel‐in‐1797, and not ψ‐as‐indicated‐by‐Beethoven‐in‐1797.28 It must be admitted to be somewhat counterintuitive for a theory to make the composer of a work essential to that work. 2. We can turn this consequence upside‐down, however. One might cite as a virtue of the MW proposal that it

79

gives a composer logical insurance that his works are his very own, that no one else has or ever could compose a work identical to any of his. If A’s musical work is an MW type, then even a fellow composer situated in an identical musico‐historical context determining the same S/PM structure composes a distinct musical work. It seems to me this is a desirable consequence, from the point of view of preserving the uniqueness of compositional activity. Why should a composer have to fear, however abstractly, that his work is not exclusively his, any more than a painter painting a painting or a sculptor sculpting a sculpture need be troubled about whether his work is at least numerically distinct from anyone else’s? Why not adopt a construal of ‘musical work’ (and of ‘poem’, ‘novel’, ‘dance’, etc.) which, while maintaining musical works as abstract types, guarantees this individuation by artist for them as well? Considerations (1) and (2) thus appear to fairly well cancel each other out. 3. A more decisive reason, however, for ensuring by proposal MW that composers A and B who determine the same S/PM structure in the same musico‐historical context yet compose distinct works W1 and W2, is that, although W1 and W2 do not, it seems, differ structurally or aesthetically or artistically at the time of composition t, differences of an artistic sort are almost certain to develop after t. So, unless we wish to embrace the awkwardness of saying that two musical works can be identical when composed, but non‐identical at some later point, we have a strong incentive to adopt MW. W1 and W2 will almost certainly diverge artistically because of the gross improbability that A and B will continue to be subject to the exact same influences to the same degree and that A’s and B’s oeuvres will continue to appear identical after the composition of W1 and W2. If A’s and B’s artistic careers do exhibit these differences after t, then and W2 will acquire somewhat different artistic significance, since W1 will eventually be seen properly against A’s total development, and W2 against B’s total development. W1 may turn out to be a seminal work, whereas W2 turns out to be a false start. Or W1 may turn out to be much more influential than W2, owing to the fact that A comes to be much better known than B. In any case, there will be some divergence in artistic attributions, if not always so marked, unless A and B remain artistic duplicates of one another throughout their lives (and thereafter). Since circumstances subsequent to a work’s composition are not comprised in musico‐historical context of composition, proposal MW′ leaves us open

80

jerrold levinson

for the awkwardness mentioned above. MW forestalls this problem completely.29 4. A last consideration inclining us to MW comprises ­certain intuitions concerning what would count as a performance of what. It seems that, in order for a performance to be a performance of W, not only must it fit and be intended to fit the S/PM structure of A’s work W; there must also be some connection, more or less direct, between the sound event produced and A’s creative activity. Whether this is primarily an intentional or causal connection is a difficult question,30 but, unless it is present, I think we are loath to say that A’s work has been performed. Consider two composers, Sterngrab and Grotesteen, who compose quartets with identical S/PM structures; suppose even that they share the same musico‐historical context. Now imagine that the Aloysious Ensemble, who are great friends of Sterngrab, give the ill‐attended première of Sterngrab’s Quartet Opus 21. Clearly, the Aloysious have performed Sterngrab’s Quartet Opus 21  –  but have they also performed Grotesteen’s Quartet Opus 21? I think not. Why? For several reasons: they don’t know Grotesteen; they weren’t using Grotesteen’s scores; they didn’t believe themselves to be presenting Grotesteen’s work – in short, there was no connection between their performance and Grotesteen the c­ reator. Grotesteen’s creating his Opus 21 Quartet had nothing whatever to do with the sound event ­produced by the Aloysious Ensemble on the afore‐mentioned occasion. Now, if Sterngrab’s Quartet has performances that Grotesteen’s does not, and vice versa, then, again by Leibniz’s law, Sterngrab’s and Grotesteen’s quartets cannot be identical. On proposal MW′, Sterngrab and Grotesteen have composed the same musical work; on proposal MW, their works are distinct. That MW squares with this intuition regarding identification of performances is thus one more point in its favor. I therefore rest with the account of musical works represented by MW. In the next section I offer some remarks on performances and transcriptions in light of this account.

VI 1. On my view, the following must all be distinguished: (a) instances of W; (b) instances of the sound structure of W; (c) instances of the S/PM structure of W; (d) performances of W. An instance of a musical work W is

a sound event which conforms completely to the sound/ performance‐means structure of W and which exhibits the required connection31 to the indicative activity wherein W’s composer A creates W. An instance of W is typically produced, either directly or indirectly, from a score that can be causally traced and is intentionally related by the performer, to the act of creation of W by A. Thus, all instances of W are instances of W’s sound structure, and instances of W’s S/PM structure – but the reverse is not the case. Instances are a subclass of the set of performances of a work.A performance of a musical work W is a sound event which is intended to instantiate W  –  i.e., represents an attempt to exemplify W’s S/PM structure in accordance with A’s indication of it32 – and which succeeds to a reasonable degree.33 Since one cannot instantiate a musical work – an S/PM structure‐as‐indicated‐by‐X‐at‐t  –  without intending to, because instantiating that demands conscious guidance by instructions, memories, or the like, which one regards as deriving from A’s indicative act at t, it follows that the instances of W are all to be found among the performances of W. However, not all performances of W count as instances of W; many if not most attempts to exemplify S/PM structures fail by some margin. So these cannot count as instances of W, but they are performances – namely incorrect performances. (Of course, that they are strictly incorrect by no means entails that they are bad.) There are not, however, any incorrect instances of W; the correct performances of W are its instances, and no others.34 Finally, let me note that musical works as I understand them can be heard in or through their performances. One hears an S/PM structure‐as‐indicated‐by‐X‐at‐t whenever one hears an instance of that S/PM structure produced by performers who, roughly speaking, are guided by X’s indication of the S/PM structure in question. And one knows precisely what musical work, i.e., structure‐as‐indicated, one is hearing if one knows what creative act is in effect the guiding source of the sound event being produced. 2. On my view of what a musical work (of the paradigm sort) is, it follows immediately that a transcription of a musical work is a distinct musical work, whether it involves alteration of the sound structure (the normal case), or even of just the performance‐ means structure. It is a virtue of my view that it gives a clear answer to this question, which is often thought to be only arbitrarily decidable. If we want such pieces to have the definite aesthetic qualities we



what a musical work is take them to have, instrumentation must be considered inseparable from them. Thus, we need not ­rely, in endorsing the distinctness position on transcriptions vis‐à‐vis original works, merely on the principle of fidelity to the composer’s intended instrumentation. Rather we are also constrained by higher‐order considerations of preserving the aesthetic integrity of such pieces.   In conclusion, let me stress some obvious consequences of accepting the theory of the musical work that I have proposed. First, composers would retain the

81

status of creator in the strictest sense. Second, musical composition would be revealed as necessarily personalized. Third, musical composition could not fail to be seen as a historically rooted activity whose products must be understood with reference to their points of origin. Fourth, it would be recognized that the pure sound structure of a musical work, while graspable in isolation, does not exhaust the work structurally, and thus that the underlying means of performance must be taken into account as well if the work is to be ­correctly assessed.35

Notes 1 There are of course several other objections to these proposals, and to the Crocean proposal mentioned below. I do not mean  to suggest that those I recall are clearly decisive by themselves. 2 See, for example, C.L. Stevenson, “On ‘What Is a Poem?’,” Philosophical Review, lxvi, 3 (July 1957): 829–862; J. Margolis, The Language of Art and Art Criticism (Detroit: Wayne State UP, 1965); R. Wollheim, Art and Its Objects (New York: Harper & Row, 1968). 3 It should be understood at the outset that sound structure includes not only pitches and rhythms, but also timbres, dynamics, accents – that is, all “purely aural” properties of sound. 4 The present paper owes a debt to two recent theories of the musical work: N. Wolterstorff, “Toward an Ontology of Artworks,” Noûs, IX, 2 (May 1975): 115–142; and K. Walton, “The Presentation and Portrayal of Sound Patterns,” In Theory Only (February 1977): 3–16. These writers are aware of some of the considerations that I adduce pointing to the complexity of a musical type. However, I believe they do not take them seriously enough, and thus are inclined to acquiesce in the view that musical works are or may be just sound structures. The present paper aims squarely to reject that view and to formulate one more adequate. 5 E.g., F# minor triad, three‐note French‐dotted rhythmic ­figure, middle C of bassoon timbre, etc. 6 This point is made by Wolterstorff, op. cit., p. 138. 7 Though of course lack of suitable production facilities made this impossible in some nonlogical sense. 8 I am aware that someone might hold that in saying that a certain novel sound instance is possible at t, all we are committed to is that the sound structure of which it would be an instance might possibly come into existence at t, simultaneously with its first instance. But I do not think this a plausible view; in saying that a certain sound event could occur at t we are saying something stronger than that the structure it would exemplify might come into existence – we are saying that that structure is right then available.

9 Some who yet resist the idea that pure sound structures pre‐ exist compositional activity are possibly failing to distinguish between structure and construction. It is true that constructions need to have been constructed in order to exist; it does not follow that structures need to have been constructed – i.e., actually put together from parts  –  in order to exist. The Brooklyn Bridge is a construction, and embodies a structure. The Brooklyn Bridge did not exist before its construction. But the geometrical structure it embodies, which required and received no construction, has always existed.   Given that there will still be some who are attracted to the view that pure sound structures are in some way created by composers, presumably through mental activity, and that these are their works, I will take this occasion to point out briefly two untoward consequences of such a view. The first is that instances of pure sound structures can always have been sounded accidentally before any composer thinks them into existence by directing his attention on the realm of sounds. In which case we would then be countenancing compositions that have instances before those compositions begin to exist. The second is that a person who conceives or sketches a sound structure new to him has no (logical) assurance that he has in fact composed anything. For if composing is bringing sound structures into existence, one may fail to do so in writing a score, provided someone else has conceived the same structure earlier. Notice that this is not a matter of the latecomer having composed the same work as his predecessor, but rather  –  what he and we would surely find incredible  –  a ­matter of his having composed no work at all. 10 Though composers compose their works by writing scores, having thoughts, or, less typically, producing performances. 11 It would be well to note here that, even if one rejects the requirement of creatability, abandonment of the sound‐­ structure view in favor of something like the view I eventually propose will be demanded by the second and third requirements developed. And those requirements strike me as being nonnegotiable.

82

jerrold levinson

12 Notice that if we assume that composing musical works is strictly creating them, it follows immediately that two ­composers cannot compose the very same musical work (no matter what sound structures they determine) unless they are either composing jointly or composing independently but simultaneously. This is just a consequence of the fact that the same thing cannot be created both at t and at a later time t’. (The same goes for a single composer on temporally separate occasions; if composing is creating, a composer cannot compose the same work twice.) I will not, however, in this section assume that composing is strict creation. 13 Cf. J.L. Borges, “Pierre Menard, Author of the Quixote” (in Labyrinths [New York: New Directions, 1962]) for a fictional demonstration of the dependence of artistic meaning on the historical context of creation. 14 Cf. Wollheim, op. cit., pp. 48–54, for a discussion of the dependence of a work’s expression on the artistic repertoire of the artist.The notion of “repertoire” is roughly that of a set of alternative decisions or choices within which an artist appears to be operating in creating his works. Wollheim extracts this idea from E.K. Gombrich’s discussions of artistic expression in Art and Illusion and Meditations on a Hobby‐Horse. 15 The convincingness of these examples depends crucially on accepting something like the following principle: “Works of art truly have those attributes which they appear to have when correctly perceived or regarded.” I cannot provide a defense of this principle here, but it has been well argued for by C. Stevenson, “Interpretation and Evaluation in Aesthetics” (in W.E. Kennick, Art and Philosophy (New York: St. Martin’s, 1964), and Walton, “Categories of Art,” Philosophical Review, lxvi, 3 (July 1970): 334–367 [reproduced as Chapter 12 in this Anthology], among others. 16 It is a mistake to regard this illustration as concerned with what Pierrot Lunaire would have been like if it had been composed by Strauss. (I am not even sure what that supposition amounts to.) The illustration rather concerns a possible musical work that possesses the same sound structure as Pierrot Lunaire, but is composed by Strauss in 1897. This work would be distinct from Pierrot Lunaire, because aesthetically divergent. But if musical works were identified with sound structures it could not be distinct.   Another way of casting the argument using this example would be as follows. Consider a possible world Q in which both Schoenberg’s Pierrot Lunaire and Strauss’s Pierrot Lunaire* exist, and call the sound structure they have in common “K.” In Q, the works diverge aesthetically and hence are non‐ identical Clearly, the works cannot both be identified with their common sound structure, but to so identify only one of them would be perfectly arbitrary. So in Q, Pierrot Lunaire ≠ K. But then in the actual world as well, Pierrot Lunaire ≠ K. Why? Owing to the necessity that attaches to identity and difference. If two things are non‐identical in any possible world, they are non‐identical in every possible world in which they exist. Put otherwise, statements of identity and

17 18

19

20

21 22 23

24

difference involving rigid designators are necessary. ‘Pierrot Lunaire’ and ‘K’ designate rigidly; they are proper names, not definite descriptions. Thus ‘Pierrot Lunaire ≠ K’ is necessarily true, since true in Q. Therefore, in the actual world, Pierrot Lunaire ≠ K. (The argument can be recast in this way, mutatis mutandis, for illustrations (2)–(5) as well.) This includes a single composer on separate occasions. It is inevitable that someone will object at this point that certain composers, in certain periods, did not compose with definite instruments in mind and did not make specific instrumentation integral to their works. This may be true to some extent. But two points must be noted. First, I have set out to define the nature of the paradigmatic musical composition in Western culture, of which Beethoven’s Quintet Opus 16 is an example. It is enough for my purpose that most “classical” compositions, and effectively all from 1750 to the present, integrally involve quite definite means of performance. Second, even in a case such as J.S. Bach, where controversy has long existed as to exactly what performing forces Bach intended, called for, or would have allowed in such compositions as The Well‐Tempered Clavier or the Brandenburg Concerto No. 2, it is clear there are still more restrictions as to performing forces which must be considered part of those compositions. Thus, The Well‐Tempered Clavier may not be a work belonging solely to the harpsichord (as opposed to the clavichord or fortepiano), but it is clearly a work for keyboard, and a performance of its sound structure on five violins would just for that reason not be a performance of it. And although the performance component of the Brandenburg Concerto No. 2 may be indeterminate between a trumpet and a natural horn in that prominent instrumental part, it certainly excludes the alto saxophone. Finally, a composition such as Bach’s The Art of Fugue, for which perhaps no means of sound production are either prescribed or proscribed, is in this context merely the exception that proves the rule. This should not be confounded with the fact that many composers were ready and willing to adapt their works in response to exigencies – in short, to license transcriptions. This is not to say that everything found in scores is constitutive of musical works. Some markings do not fix the identity of a work but are instead of the nature of advice, inspiration, helpful instruction, etc. However, the suggestion that instrumental specifications are of this sort is totally insupportable. Cf.Walton, “Categories of Art,” op. cit., pp. 349/50, for related examples. James Lyons, liner notes, phonograph record Nonesuch 71054. The best one could say would be that the Quintet achieved a satisfactory blending of pianoish sounds and woodwind‐ish sounds. One could alternatively speak of a single structure which, construed rightly, entails both the required sounds and the required means of sound production. This would be a structure of performed sounds, as opposed to “pure” sounds. For example, one such performed sound would correspond to the



what a musical work is

following specification: “Middle C of half‐note duration played on oboe.” Clearly this implies both a certain sound qualitatively defined and a means of producing it.   The main reason I favor the S/PM formulation is that it is more transparent. It preserves some continuity with the sound‐structure view which it supersedes, and displays more clearly than the performed‐sound formulation that, although a musical work is more than a sound structure, it most definitely includes a sound structure. 25 This analogy was brought to my attention by Warren Ingber. 26 See, for example, J.L. Austin’s “Truth,” Proceedings of the Aristotelian Society, supp. vol. xxiv (1950): 111–128. 27 The analogy might even be reversed, so as to illuminate the nature of statements. If musical works are structures‐ as‐indicated…, then possibly statements just are: sentences‐as‐uttered… 28 I am assuming, of course, that Hummel could not possibly have been Beethoven. If he could have, then I suppose that, even on MW, Hummel might have composed Beethoven’s Quintet. 29 I will take this opportunity to point out that although aesthetic and artistic attributes have played a large role in this paper, I have not insisted on them as essential to musical works, but only as relevant  –  in common with all other attributes – to individuating them.The argument has nowhere required as a premise that such attributes are essential attributes. It has assumed only that aesthetic/artistic attributes truly belong to works in a reasonably determinate fashion. As for what attributes are essential to musical works, given MW, it seems that certain structural and genetic attributes would have to be admitted: S/PM structure, composer, date of composition. But it is not obvious that aesthetic/artistic attributes will turn out to be essential, i.e., possessed by a work in all possible worlds it inhabits. Consider a possible world in which Schoenberg determines the S/PM structure of Verklarte Nacht during 1899 but in which Wagner had never existed. The resultant work might still be Verklarte Nacht, though some of its aesthetic/artistic attributes would be subtly different. 30 Quandaries arise when these considerations conflict, which I will not attempt to deal with here. For example, suppose the  Aloysious Ensemble are actually reading copies of Grotesteen’s score while believing themselves to be playing Sterngrab’s score. Do they perform Sterngrab’s Quartet, Grotesteen’s Quartet, or both?

83

31 I will assume here that the required connection is primarily, if not wholly, intentional. 32 And thus an attempt to exemplify an S/PM‐as‐indicated‐ by‐X‐at‐t. 33 What constitutes a “reasonable degree,” and thus what differentiates poor or marginal performance from nonperformance, is for many compositions perhaps marked by the ability of an informed and sensitive listener to grasp, at least roughly, what S/PM structure is struggling to be presented. For example, even an especially informed and sensitive listener would grasp approximately nothing of the Hammerklavier sonata from my attempt to present its structure,since my ­facility at the piano is next to nil—no performance (much less an instance) of the Hammerklavier sonata can issue from me or my ilk. 34 Thus I am in opposition to Wolterstorff ’s suggestion, in “Toward an Ontology of Artworks,” op. cit., that musical works be construed as norm‐kinds, i.e., as having correct and incorrect, or proper and improper, or standard and defective instances. What we say about musical works can, I think, be more perspicuously interpreted in terms of the distinction between instance and performance. Further, construing instance as requiring full conformity to score (i.e., as an all‐ or‐none proposition) has the virtue, as Nelson Goodman pointed out in Languages of Art (Indianapolis: Bobbs‐Merrill, 1968), of assuring preservation of a work’s identity from work to instance and from instance to work. But by also distinguishing between instance and performance (which Goodman does not do) one can sweeten the judgment, say, that Rubinstein’s playing of the Chopin Ballade No. 3 with two mistakes is not an instance of the work, with the willing admission that it is surely a performance of it (and possibly a great one). 35 It is worth observing that, if the position developed in this paper is correct, it may have interesting implications not only for the identity of other sorts of art work (this I take to be obvious), but for the identity of abstract cultural objects of various sorts – e.g., scientific theories, speeches, laws, games. A physical theory, for example, can’t be simply a set of ­sentences, propositions, or equations if it is in fact the possessor of properties such as brilliance, revolutionariness, derivativeness, immediate acceptance. For that very set of sentences, propositions, or equations might be found in another theory occurring fifty years earlier or later which lacked those properties.

8

Defending Musical Platonism Julian Dodd

I.  Introduction: The Simple View and the Argument from Creatability In my article, ‘Musical Works as Eternal Types’,1 I began a project of rehabilitating a simple view of the ontological nature of works of music. This conception has it that such works are sound structures: structured types that have only sound‐types as constituents. Two features of the simple view are worth highlighting at this stage. First, it is the default position when it comes to the ontology of musical works. Even its most influential critic, Jerrold Levinson, admits that it is ‘the most natural and common proposal on this question’.2 Second, the simple view is a robust version of musical Platonism. Abstract patterns of sound‐types, it is agreed, cannot come in or out of existence; so, to identify a work of music with a sound structure is to commit oneself to the thesis that the work exists at all times, and hence cannot have been created by its composer.3 The dominant attitude towards the simple view has been to regard its Platonism as a consequence so counterintuitive as to trump its initial appeal. So much so, indeed, that philosophers are apt to think that the demise of the simple view is assured by what I have termed ‘the argument from creatability’:

(1) Sound structures exist at all times. So (2) If musical works were sound structures, they could not be created by their composers. (3) But musical works are created by their composers. So (4) Musical works are not sound structures.4 Naturally, if the soundness of this argument is granted, the project of attempting to reconstrue musical works as creatable types becomes pressing. And to this end, Levinson has made the most significant suggestion. A work of music, Levinson claims, is not a sound structure, but an ‘indicated type’: a sound structure‐as‐indicated‐by‐a‐composer‐at‐a‐ time. According to Levinson, because indicated types are ‘person‐and‐time‐tethered’ entities (because, that is, they are individuated by relation to the composer’s compositional act),5 it follows that they are only in existence from the time to which they are tethered. It is thus supposedly a virtue of Levinson’s ‘qualified Platonism’ that he is able to allow for the fact that musical works are brought into existence by the composer’s act of composition. So much for the orthodoxy. My response to all this was twofold. First, I argued against the argument from creatability’s third premise.6 In my view, what matters for our concept of composition is that we recognize the composer’s creativity. The suggestion that composers literally bring into being their compositions is inessential: an intuition

Julian Dodd, “Defending Musical Platonism,” British Journal of Aesthetics 42 (2002), pp. 380–402. Reproduced by permission of Oxford U ­ niversity Press.

Aesthetics and the Philosophy of Art – The Analytic Tradition: An Anthology, Second Edition. Edited by Peter Lamarque and Stein Haugom Olsen. © 2019 John Wiley & Sons, Inc. Published 2019 by John Wiley & Sons, Inc.



defending musical platonism

that we can harmlessly drop. Composition, I said, is a kind of creative discovery.7 A proponent of the simple view should reject the (frankly absurd) view that musical composition is akin to the unimaginative tracing of an abstract pattern. My second move was to turn the spotlight on Levinson’s alternative proposal: his contention that his person‐and‐ time‐tethered types are brought into existence by acts of indication. To this my response was that types of any kind are eternal existents, and my argument for it, what we might call ‘the argument from the eternal existence of properties’, may be represented as follows: (5) The identity of any type K is determined by the condition a token meets, or would have to meet, in order to be a token of that type. (6) The condition a token meets, or would have to meet, in order to be a token of K is K’s property‐ associate: being a k. So (7) The identity of K is determined by that of being a k. So (8) K exists if and only if being a k exists. (9) Being a k is an eternal existent. So (10) K is an eternal existent too.8

My point against Levinson is that this argument applies to any kind of type, pure or indicated. All properties are eternal existents, so, by (8), all types exist eternally. So it follows that, even if we were persuaded that musical works are created by their composers, Levinson’s proposal would be in no better shape to do justice to this requirement than is the simple view.9 If a work of music is a type, even an indicated type, it is not the sort of thing that can be brought into existence. There might be arguments that could convince us to jettison the simple view in favour of Levinson’s counterproposal, but the argument from creatability is not one of them. Not wholly unexpectedly,‘Musical Works as Eternal Types’ has attracted a good deal of trenchant, though good‐natured criticism. Forthright objections to my view have been raised by Robert Howell,10 R.A. Sharpe,11 and Saam Trivedi.12 These objections have been invaluable to me inasmuch as they have challenged me to elaborate and explain the position I am defending. It is my contention, however, that, in the wake of such elaboration, my position emerges unscathed.

II.  Composition as Creative Discovery If my argument is sound, musical composition cannot be a matter of a composer bringing an entity into existence. Musical works, qua types, exist eternally. Consequently,

85

composition can only be a kind of creative discovery. In composing In This House, On This Morning,Wynton Marsalis came across something that was already there, but his manner of coming across the work required him to think creatively and deploy a musical imagination of great depth. In this section I shall reply to objections to this conception of composition raised by Trivedi and Sharpe. As it happens, some of the objections pressed by Trivedi were discussed, and to my mind refuted, in my original article.13 The same goes for another of Trivedi’s complaints, namely that ‘Not regarding composers as creators but instead as discoverers would result in a false divide in this respect between music and the other arts’ (AMW, p. 75). But in this latter case, it will be useful to discuss this once more for two reasons: first of all, to clarify my position; but secondly, to lay bare the reliance on questionable intuitions which forms a leitmotif of Trivedi’s paper. Let us, then, focus on the suggestion that theoretical unity demands that we treat musical works as created by their composers. The assumption here is that paintings, sculptures, and the like are created entities, and hence that our account of works of music should fall into line. But as Trivedi is aware, my reply to this point is to ask what my opponent supposes the ontological status of paintings and sculptures to be. If the objector assumes that paintings and sculptures, unlike works of music, are concrete particulars, then the appeal to theoretical unity cuts no ice. For if works of music are abstract types and paintings concrete particulars, it is far from clear that theoretical unity demands that they both be regarded as literally created. On this view of paintings and sculptures, The Haywain, since it is a concrete particular, is so different in kind from the abstract type that is In This House, On This Morning that we have no right to expect them to have creatability in common. If, on the other hand, the objector takes paintings and sculptures to be types rather than particulars,14 then any appeal to theoretical unity will work against, rather than for, the creationist. For the considerations that tell against the creatability of works of music (including the argument from the eternal existence of properties) will apply equally to paintings and sculptures, and we will thus be pressed into arguing that these art works, like works of music, are eternal existents. To this,Trivedi makes three kinds of reply. First, he claims that the fact ‘that musical works and paintings are ontologically different in some ways (in that the former are types and the latter concrete particulars) does not preclude their being similar with respect to being created’ (AMW, p. 75). Second, he takes the question of how and where supposedly eternal types exist to be unanswerable. And, finally, Trivedi

86

julian dodd

raises problems for the thesis that paintings and sculptures are types. Here he suggests that we would only have a good reason to take this view of paintings and sculptures, if (contrary to what is currently the case) we had the technology that allowed for their perfect reproduction; adds that a painting’s identity is in part determined by contextual properties; and ends his critique by noting the supposed additional counter‐intuitiveness of suggesting that the Mona Lisa is a type which is eternal, and hence existed before the painting’s subject was born. This demonstrates, Trivedi claims, that ‘Dodd must either abandon his view that paintings and carved sculptures are types, or he must jettison his view that all types are eternal existents, or he must give up both these views, as I would urge; at any rate, something has to give’ (AMW, p. 76). In my view, if we consider these points individually, it is plain that they fall some way short of their intended target. To begin with, Trivedi begs the question in stating that a conception of works of music as types does not preclude their being created. The argument from the eternal existence of properties is an argument for the negation of just this claim. Trivedi can only earn the right to say that types are creatable by demonstrating my argument to the contrary to be unsound, and this is a challenge he fails to take up. If Trivedi’s first reply begs the question, his third, alas, is irrelevant. For his spirited rebuttal of the conception of paintings and sculptures as eternal types is a denial of a thesis to which I am careful not to commit myself. Whilst acknowledging that this Strawsonian doctrine is ‘by no means absurd’,15 I did not go so far as to sign up to it. Indeed, the conception of paintings and sculptures as types only figures in my discussion in a characterization of the second horn of the dilemma facing Levinson, namely, that if he were to take paintings and the like to be types, then the same considerations which tell against creationism with regard to musical works would tell against creationism with regard to paintings. My point is that whatever Levinson says about the ontology of paintings and sculptures, his appeal to theoretical unity cannot help him; and the making of this point does not require me to (and did not see me) take a stand on this ontological question. Of Trivedi’s objections to my attempt to explain away the supposed ‘false divide’ between music and the other arts, this just leaves us with his claim that I cannot explain how and where eternal types exist. But such a claim could only convince someone who already shared Trivedi’s overtly nominalist sympathies.To someone lacking Trivedi’s prejudice against abstracta, the questions where and how such eternal types exist look either malformed or else

s­uspect. For to ask where such entities exist is to presume that they must exist somewhere, which they do not. Eternal types, in common with sets, numbers, propositions, and facts, have no spatial location. And to ask, as Trivedi does, ‘how did these allegedly pre‐existent musical works exist eternally … before the advent of humans and the human activity of music making?’ (AMW, p. 76), is to ask a question which, if it make sense at all, can only reveal the questioner as deaf to the arguments of my original article.Types must be eternal because they inherit their existence conditions from their eternally existent property‐associates. This can only mean that the human activity of music‐making sees people creatively discover, rather than create, works of music. But this conclusion, far from distorting our concept of composition, sits comfortably with those features of it worth preserving. That concludes my response to Trivedi’s objections to the conception of composition as creative discovery. What of Sharpe’s critique? Sharpe, unlike Trivedi, does not rely on contested nominalist intuitions; he offers a self‐standing argument against the thesis that composers creatively discover their works. In Sharpe’s view, if we pay attention to the concept of discovery, it becomes evident that, to borrow Sharpe’s own example, Beethoven’s composition of the Archduke Trio does not fall under this concept. In the remainder of this section I shall explain why Sharpe’s argument is unsound, and then take the opportunity to say a little more about the nature of composition itself. Sharpe begins by drawing our attention to a contrast between certain central cases of discovery, such as Columbus’s discovery of the Americas or Andrew Wiles’s discovery of a proof of Fermat’s last theorem (AT, p. 326), and Beethoven’s composition of the Archduke Trio. Columbus, for example, whilst certain that he had indeed discovered the Americas, might have been mistaken in his belief. He could have got it wrong. Beethoven, on the other hand, could not have been mistaken in an analogous way. Of course, the work composed by Beethoven might have been other than it is. Beethoven might have ended up composing a work other than the Archduke Trio (although he might have given it the same name). But in such a possible world Beethoven would not have made a mistake in failing to compose the Archduke Trio. He would just have composed a different work. Put more formally, Sharpe’s thinking here takes the form of a commitment to the following argument: (11) For a person A to discover an entity o, it must be possible for A to have been mistaken.



defending musical platonism

So (12) For Beethoven’s composition of the Archduke Trio to have been a discovery of the work, it must have been possible for Beethoven to have been mistaken in his compositional act. (13) But Beethoven could not have been so mistaken. So (14) Beethoven’s composition of the Archduke Trio was not a discovery of the work. What are we to make of this? I do not wish to question (13) here, although it could perhaps be queried. My quarrel is with (11). For (11) sees Sharpe seeking to elevate an observation concerning certain cases of discovery into a conceptual truth. And it is nothing of the kind, as we shall now see. In fact, counterexamples to (11) are easily constructed. Let us consider a case in which someone discovers something by, as it were, stumbling across it. So imagine that my daughter comes across a coin whilst pottering in the flowerbeds, picks it up, and says ‘Look what I’ve discovered!’ Now it is clear that she might not have found the coin. But it is also clear that it would be senseless to say that she might have been mistaken: for what could such a mistake have consisted in? Because she did not set out to find such a thing, merely coming across it by accident, there are no criteria of correctness here that she could have failed to meet. None the less, it would be equally implausible to deny that my daughter had indeed discovered the coin. (One would not ‘correct’ her by saying, ‘No, Eleanor, you didn’t discover the coin; you see, you couldn’t have made a mistake.’) This being so, we have a clear counterexample to (11). Pace Sharpe, an agent may discover an entity and yet it be inappropriate to describe the situation as one in which she might have been mistaken. To explain the nature of Sharpe’s mistake we must return to the cases of discovery with which he started: those of Columbus and Wiles. Why is it possible for Columbus and Wiles to have been in error? Sharpe’s own answer to this question would seem to be this: the possibility of being mistaken is a corollary of the fact that what is discovered is a genuinely mind‐independent entity. As Sharpe himself puts it, ‘what we discover exists and with that goes the possibility of being mistaken’ (AT, p. 326). But this is not quite right, as our counterexample illustrates. The point is not that discovery is conceptually linked to the possibility of making a mistake; it is that in certain cases of discovery the agent’s enquiry is constrained by criteria of success, criteria that, as we shall see, are not present when a discovery is made which is other than the end product of a process of enquiry.

87

When it comes to Columbus’s discovery of the Americas and Wiles’s discovery of his mathematical proof, the existence of criteria of success is a corollary of the fact that both agents were engaged in a search for a particular thing: both Columbus and Wiles were trying to find something fitting a certain description; and, hence, each might have failed to find what they were looking for. In both cases, the language of possible error is applicable because what would count as a mistake is determined by the point of the search. Consequently, the moral to take from our counterexample to (11) is that it is only a sub‐ species of discovery, what we might call discovery by enquiry, which has a conceptual connection to the possibility of being mistaken. When a discovery is something other than the culmination of an enquiry – when the purpose of  the agent’s activity is something other than that of uncovering some entity or some fact – there is no such conceptual connection. Having made this point, the consequences for the case of musical composition should be obvious. To be sure, composers do not stumble across their compositions like happy‐ go‐lucky gardeners, but Beethoven’s composition of the Archduke Trio has at least this much in common with my daughter’s discovery of the coin: although he came across something that existed prior to his compositional act, he did not set out to uncover an entity fitting a certain description. The piece was there anyway, but Beethoven did not form a conception of it and then compose it. Indeed, to view Beethoven’s achievement in this way would be to misrepresent the nature of composition. Beethoven’s coming to form a clear conception of the Archduke Trio (by way of a process of indication) just was his compositional act, so there existed no gap between conception and search as there was for Columbus and Wiles. It is, therefore, fair to say that Sharpe mistakes norms of enquiry for norms of discovery. The fact that Beethoven could not have been mistaken in his composition of the Archduke Trio illustrates, not that he did not discover it, but that his discovery was not the consummation of an enquiry. And I take it that this conclusion coheres with the way in which we ordinarily think about the purpose of composition. Clearly, Beethoven was not setting out to uncover something; his purpose was that of expressing himself musically and thereby adding value to his life and the lives of others. It is for this reason, and not because composers do not make discoveries at all, that we cannot say of Beethoven that he might have been mistaken in his composition of the Archduke Trio. Composition is a species of creative, not inquisitive, discovery.

88

julian dodd

There is thus no mileage in Sharpe’s objection to my claim that composers discover their works. However, his objection serves to issue an important challenge to the proponent of the simple view: namely, to flesh out a Platonist view of musical composition. With this challenge in mind, we should first of all confront the suspicion that the Platonist inevitably both downplays the composer’s achievement and ignores the fact that a composer’s work is determined in part by the musico‐historical context of his compositional act.16 But such an underestimation of the determining part played by the compositional act’s musico‐ historical context is more imagined than real, although, admittedly, contextual factors play a different role for the Platonist than they do for philosophers such as Levinson, Howell, Trivedi, and Sharpe. For these thinkers, a musical work is ontologically tethered to the composer’s compositional act, and, as a consequence, it is metaphysically impossible for works to have been composed by anyone else at any other time. The Platonist, by contrast, treats musical works as sound structures pure and simple, and thus admits the bare metaphysical possibility that the Archduke Trio was composed earlier than it in fact was, and by another composer. However, to admit this metaphysical possibility is not to ignore the fact that Beethoven’s musico‐historical situation played a part in his composition of the work. Musical works are not composed in a vacuum. Beethoven came to compose the Archduke Trio as a result of his having been immersed in a particular musico‐historical context: his composition of the work was the product of his location at a set of coordinates in musico‐historical space. But in saying this, we need not commit ourselves to the thesis that the context of composition is essential to the work, and hence to the view that it is impossible for a given work to have been composed by someone else at some other time. To write a composer’s position in musico‐historical space into the identity of his compositions is to overreact to the truism that a composer’s compositions reflect his place in history. And that it is an overreaction is evident in the position’s counterintuitive consequences: namely, that Beethoven could not have composed the Archduke Trio any earlier or later than he did; and that a Twin‐ Beethoven, in a qualitatively identical musico‐historical context, who indicated the Archduke Trio’s sound structure, could not have composed the piece. So how can a proponent of the simple view allow for the influence of contextual factors on the output of a composer? By taking such factors to be, as a matter of contingent fact, causally relevant to the composer’s discovery of the work, as opposed to items that are written into the work’s

very identity. Although it is metaphysically possible that the Archduke Trio was composed in the sixteenth century, a possible world in which this occurred is a very distant world. And the reason why this is so is that it is the composer’s position in musico‐historical space that enables him to discover the works that he does.The fact that Beethoven had internalized much of the history of music up until that point, and was part of a social and artistic milieu, inevitably shaped the nature of his compositions. Consequently, for someone else to have composed the work in the sixteenth century, without the input of the musico‐historical context that enabled Beethoven to compose it, something anomalous and astounding (though not impossible) would have had to have happened. Such a composer’s genius and originality would have to have been peerless; he would have to have somehow broken out of the epoch in which he was working. There would have to have been a huge imaginative leap: an occurrence that somehow skipped chapters in musical history. But notice that this appeal to the part played by context in musical composition falls short of tying the work to Beethoven’s compositional act.We appeal to context in order to shed light on how composers come to creatively discover the works they do; but this is compatible with acknowledging that anomalous compositional acts – acts such as the counterfactual one that we have just been imagining – are metaphysically possible. To admit the bare metaphysical possibility that a work could have been composed at a different time, and by a different person, is not to deny that Beethoven composed the works that he did because he occupied the position in musico‐historical space that he did. Making use of the clear analogy between musical composition and scientific theorizing helps us to understand how the Platonist need not downplay the composer’s achievement, nor ignore the contextual factors causally relevant to composition. Einstein discovered the facts that constitute the Special Theory of Relativity; he did not invent them. This, however, does nothing to undermine our sense of Einstein’s peerless brilliance and creative thought: only someone very much like Einstein could have uncovered such facts. Equally, a possible world in which the Special Theory of Relativity emerged in 1850 is a world light years away from the actual world: Einstein’s discovery was contingent upon other discoveries before it, and upon Einstein’s internalization of a vast amount of physical theory, so all of this background would have to have been in place too. But, of course, it all might have been. The Platonist, it seems to me, should say exactly analogous things about musical composition. It is no insult to



defending musical platonism

Beethoven to compare his discoveries with the mould‐ breaking, though historically conditioned, discoveries of Einstein. And the Platonist may make this comparison safe in the knowledge that an acknowledgement of such historical conditioning is compatible with a recognition of its metaphysical contingency. Beethoven creatively discovered the Archduke Trio partly as a result of his occupancy of a position in a specific cultural and artistic milieu. But it is not metaphysically necessary that the work was composed by Beethoven at that time, any more than it is metaphysically necessary that Einstein discovered the Special Theory of Relativity when he did. Despite my disagreement with Sharpe, however, my reading of Sharpe’s article has enabled me to see that the analogy between musical composition and scientific discovery should not be pushed too far.True enough, highly creative, yet historically determined, discoveries can be made in both disciplines; but the crucial difference (which I failed to stress in my original article17) is that musical composition is not a form of enquiry. Einstein was aiming to find things out; Beethoven was not. It is this difference, and this difference alone, which explains why it makes sense to say of Einstein, but not of Beethoven, that he could have been mistaken. Both made discoveries; indeed, if musical works are types, and if my argument from the eternal existence of properties is sound, we have no choice but to accept this.

III.  Properties and Types This last remark nicely brings me on to what I consider to be the most important criticisms of ‘Musical Works as Eternal Types’: those levelled at my views on the metaphysics of types and, in particular, my argument from the eternal existence of properties. A truly satisfying response to my view must tackle my ontological claims head on, and it is to attempts to do just this that I now turn. The moral of the argument from the eternal existence of properties is this: types inherit their existence conditions from their property‐associates, so, since properties are eternal, so are types. In Trivedi’s view, however, I am guilty of conflating types with properties (= universals): I would urge that there are differences between the type– token distinction and the universal–particular distinction, which Dodd seems to be obliterating in assimilating types to a Platonist conception of universals when he claims, ‘If indicated structures are genuine types, they cannot be created.’ (AMW, pp. 73–74)

89

And Trivedi goes on to set out three cruces of the distinction between types and universals, as he understands it. First, claims Trivedi, all and only the properties possessed necessarily by tokens (qua tokens of their respective types) may be shared by those types; whilst the equivalent thesis does not hold when it comes to universals and particulars. So, for example, whilst the type The Union Jack is itself coloured and rectangular, the universal being red is not itself red. Second, types are abstract particulars, whilst universals are properties: predicative features of objects. And finally, Trivedi claims, types, but not universals, are creatable. Appealing to the work of Wollheim, Trivedi asserts that types are postulated where we can correlate human inventions with a class of elements, whereas this need not be true of universals. Types, in other words, are creatable abstract particulars, while universals, though abstract, are neither creatable nor particulars. (AMW, p. 74)

I have two replies to all this. To begin with, Trivedi’s remarks about types and tokens seem to me to contain errors. If, for example, The Union Jack is really an abstract object, Trivedi cannot blithely state that this object can share with its tokens the properties of being coloured and rectangular. Abstract objects have no spatial location, so cannot be a certain shape or a certain colour. Furthermore, Trivedi is not entitled to assume that types are, by definition, creatable. The Elephant is a type, but only idealists would insist that we brought it into being. And, in any case, I have an argument for the thesis that types, like universals, exist eternally. Unless Trivedi can explain where this argument fails, his objection is powerless. Most importantly, though, Trivedi’s claim that I conflate types and universals is simply false. All I am committed (and wish to be committed) to is (8): the thesis that K exists if and only if being a k exists. That is not to treat K and being a k as identical; it is merely to say that there is no possible world in which only one of these entities exists.18 Properties and types, I have agreed, differ in their natures: types are abstract particulars whilst properties are predicative features of objects. Trivedi, I am thus forced to say once more, fails to engage with the detail of my argument. The same criticism cannot be levelled at Howell. Confronting head on the ontological claims made in ‘Musical Works as Eternal Types’, Howell argues that the line of thought represented by the argument from the eternal existence of properties is unsound. Defending Levinson’s conception of musical works as indicated types, Howell

90

julian dodd

argues that such works form a subset of types that are brought into being by human practices. The remainder of this article will be devoted to replying to his metaphysically heavyweight objections. It is my contention, however, that his punches can be slipped. Pace Howell, all types exist eternally. Howell’s first objection concerns the relationship between types and properties. In his view, my (9) is false. There are, he claims, many properties, notably including those property‐associates of indicated types, which do not exist eternally. The properties Howell has in mind are impure properties: properties whose specification includes reference to another entity, such as being a son of Abraham Lincoln and being older than Alistair Cooke. Crucially, if Levinson is right, the property‐associates of works of music are impure properties. To use Howell’s own example, Beethoven’s Fifth Symphony is sound structure S‐as‐indicated‐by‐Beethoven‐in‐1804–1808, and hence its property‐ associate is being an instance of S that is produced in a way that is properly connected to Beethoven’s 1804–1808 acts of indication. So if Howell is correct in his views about impure properties, and if we grant Levinson’s account of works of music, such works are not eternal existents after all. And such a result, of course, could tempt us to replace the simple view with Levinson’s qualified Platonism. The pressing question is this: why does Howell deny that impure properties are eternal existents? The answer lies in the following, apparently intuitive, thought: that such properties only exist once the entities they essentially concern have come into existence. Consequently, according to Howell, being a son of Abraham Lincoln only came into existence once Lincoln himself was born, and, crucially for our present discussion, a similar moral applies to being an instance of S which is produced in a way that is properly connected to Beethoven’s 1804–1808 acts of indication. Of this latter property, Howell asks How can it already have existed in, say, 1600 – or at the moment of the Big Bang – when the specific concrete entities to which it essentially relates had themselves not yet come into existence? To suppose that it can would be like supposing that your signature – not just ink marks geometrically congruent to it, but actual marks that attest to you, to your own personal identity – could exist a million years before you do or that the set consisting of last night’s thunderstorm and today’s gusting of the wind could pre‐exist both these events. (TII, pp. 112–113)

But we should not lose our heads here. Before we get carried away by Howell’s charge of counterintuitiveness, we should note that the threat seemingly posed by Howell’s

remarks about signatures and sets is easily deflated. First of all, the suggestion that it is obviously absurd to take my signature to pre‐exist me may be countered once we remind ourselves that by ‘signature’ here we mean signature‐type: the abstract pattern which my signature‐tokens instantiate. Presumably, this type is an indicated type: X‐as‐produced‐by‐ Julian Dodd (where X is the inscriptional type). Now, what would be absurd would be to claim that this type could have been tokened before I was born. Only marks made by me can count as tokens of my signature. But it does not follow from this that the signature‐type did not exist before I did. To think that such a conclusion follows is to confuse the question of what it is for a type to be tokened with the distinct question of what it is for the type itself to exist. However, it seems to me that the force of Howell’s question – how ‘actual marks’ that attest to me could pre‐ exist me – relies upon precisely this conflation. For what are such ‘actual marks’ but signature tokens? Of course, no ‘actual marks’ – that is to say, concrete tokens – made a million years ago could be examples of my signature. But pointing this out does not undermine the claim that the type existed a million years ago. Furthermore, the analogy between properties and sets is misleading. The reason why the set consisting of last night’s thunderstorm and today’s gusting of the wind could not pre‐exist those events is simple: those events are essential constituents of the set (since sets have their members essentially). But it is far from obvious that impure properties have constituents in an analogous way. Being a son of Abraham Lincoln is a condition which something has to meet in order to be a token of a certain type: namely, a son of Abraham Lincoln. What is obscure to me is why we should regard this condition as a complex entity that has the man himself as a part. I shall return to this in a while. For now, however, let us delve a little deeper into why it might be thought that being a son of Abraham Lincoln did not exist until Lincoln himself existed. I think there are two candidate reasons for taking this line, neither of which is convincing. The first such reason is not explicitly formulated in Howell’s article, but is, perhaps, lurking beneath the surface. To begin with, it is undeniable that being a son of Abraham Lincoln exists at t if and only if there exist at t objective satisfaction conditions for whether something is a son of Lincoln or not. There must be a fact of the matter as to whether something meets these conditions or fails to meet them. But, the argument continues, it is precisely these satisfaction conditions that are lacking before Lincoln’s birth. For, to borrow Howell’s own way of putting it, before



defending musical platonism

Lincoln was born, Lincoln’s identity was not ‘available’ in the world (TII, n. 13). And what Howell means by this is, I think, that before Lincoln existed there could have been no question of whether or not something was Lincoln, or indeed a son of Lincoln.These questions could have had no content at this time. But, given the fact that a property F can only exist at t if there exist at t objective satisfaction conditions for being F, and given that the existence of Lincoln is required for such satisfaction conditions to exist, being a son of Abraham Lincoln (and the other impure properties concerning Lincoln) could not have existed before Lincoln himself did. With the aim of putting some more flesh on the bone of this notion of ‘availability’, it might help to explain why it is that Howell believes Lincoln’s identity to be available after his death but not before his birth. According to Howell, Lincoln’s identity is available in the world in 2002 because there are causal chains linking events involving Lincoln to events in 2002. The existence of such causal chains means that there are, as it were, traces of Lincoln in 2002 even though the man himself has long since ceased to exist. And it is because there exists such a causal link between Lincoln and events in the present moment that being a son of Abraham Lincoln exists now, or, to put it another way, that the question of whether something is a son of Lincoln or not presently has content. Were determinism about Lincoln true, there would, of course, have been analogous causal chains linking events before Lincoln’s birth to later events involving Lincoln. The existence of such a causal link would have ensured Lincoln’s availability (and, hence, the existence of any impure property involving Lincoln) from the moment at which Lincoln’s existence was determined. However, says Howell (TII, n. 13), given the fact that the universe is not deterministic in this way, when it comes to the availability of Lincoln’s identity, there is a clear disanalogy between the time after Lincoln’s death and the time before his birth. Lincoln’s identity only became available at his birth, and hence no impure property concerning Lincoln could have existed before Lincoln did. In my view, however, the causal chains Howell describes have an epistemological, rather than a metaphysical, relevance. At first blush, availability looks like an epistemological notion  –  ‘availability to whom?’, we might ask  –  and appearances are not deceptive in this instance. It is, no doubt, true that before Lincoln’s birth his identity was not available in Howell’s sense; but what this means is that it was not available to us: that we could have had no knowledge of Lincoln then. And one reason why this is important is that it entails that, before Lincoln’s birth, it was

91

impossible for anyone to refer to, or otherwise have a singular thought about, Lincoln. But it does not follow from this, as Howell seems to think it does, that being a son of Abraham Lincoln did not exist before Lincoln’s birth, and hence that there was no fact of the matter as to whether an entity was Lincoln or not at such a time. For in 399 bc it was a fact, albeit an ungraspable fact then, that no entity was Lincoln, and hence that no entity was Lincoln’s son. Howell has thus misinterpreted what is really an important fact concerning the possibility of reference and singular thought: we can only refer to, or have a singular thought about, things that are available in the world for us. We cannot refer to, or entertain singular thoughts about, things of which we know nothing, and the possibility of having such knowledge would seem to require the existence of the kind of causal link Howell envisages between the thinking subject and the object of knowledge. True enough, the question of whether an entity is Lincoln’s son or not could not have arisen before Lincoln’s birth: in 399 bc ‘Abraham Lincoln’ could not have been used to refer to the man who eventually became US President. But admitting this is not to deny that properties such as being Abraham Lincoln and being a son of Abraham Lincoln existed before Lincoln did. The existence conditions of properties and the conditions necessary for reference and singular thought to be possible are distinct matters. It is my belief, then, that it is a misplaced conflation of matters metaphysical and epistemological that makes it appear problematic to believe impure properties involving Lincoln to pre‐exist Lincoln himself. But more than this, it is the denial that such impure properties pre‐exist Lincoln that leads to trouble. For example, if being a son of Abraham Lincoln did not pre‐exist Lincoln, then it follows that before Lincoln’s birth no entity could either have had or failed to have had that property. (If, at t, there is no condition F that something must meet to be of a certain kind, then, at t, there is nothing which success or failure in meeting F can consist in.) But if this is so, then it follows that in 399 bc the fact that Socrates was not a son of Lincoln did not yet exist. Presumably, Howell would say this: the fact in question relates Socrates in the fourth century bc to a property which first exists when Lincoln is born, and hence it is a fact that itself only comes into existence at Lincoln’s birth.19 But, to begin with, this contention is, I suggest, extremely counterintuitive. It did not become a fact that Socrates was not Lincoln’s son only once Lincoln was born: it was always a fact. Only someone distracted by the proposition’s only becoming expressible at Lincoln’s birth could fail to be convinced of this.

92

julian dodd

But the problems do not end here, for it would seem to involve a category mistake to suppose, as Howell seems to, that facts can come into existence.20 Let us stipulate that Bill’s cross‐channel swim  –  the event  –  took place on August 10th, 2001. At what time did the fact that he swum the channel begin to exist? This question sounds bizarre to an English‐speaker, and its paradoxical flavour reflects the fact that facts do not originate. A sentence such as 15. On August 10th it became a fact that Bill swum the channel is nothing but a clunky transform of the correct 16. It is a fact that Bill swum the channel on August 10th.21 Well, the fact that Socrates was not Abraham Lincoln’s son is no different in kind from the fact that Bill swum the channel. The items the fact concerns occupy periods of time, but only someone in the grip of a malign philosophical theory could think that there was a moment when this fact began, was established, or otherwise came to be. So much for the first reason for denying that impure properties pre‐exist the particulars they involve. The second such reason is, like the first, not explicitly acknowledged, but differs in being driven by a claim about the ontological nature of impure properties. Let us focus once more on Howell’s rhetorical question. Speaking of an impure property, Howell asks How can it already have existed in, say, 1600  –  or at the moment of the Big Bang – when the specific concrete entities to which it essentially relates had themselves not yet come into existence? (TII, pp. 112–113)

We may notice immediately that Howell takes impure properties themselves to be essentially related to concrete entities (in the present example, to Abraham Lincoln). The ‘it’ in ‘to which it essentially relates’ refers to the same thing as the first occurrence of ‘it’: the property. So what is the essential relation between being a son of Abraham Lincoln and Lincoln himself supposed to be? Just how does the property ‘involve’ the man?22 Howell’s answer is that which I expressed a certain scepticism about earlier: the relation holding between being a son of Abraham Lincoln and Lincoln himself is nothing other than that which obtains between a complex object and one of its constituents. The property ‘involves’ the concrete particular in the sense that the latter is an essential part of the former, a conception demonstrated by Howell’s explanation of how being a son of Abraham Lincoln comes into being:

the wholly general relational property being a son of someone already exists (and, for all I know to the contrary, is eternal). When Lincoln comes into existence, he automatically fills the open, ‘someone’ slot in this property. In consequence, the monadic property being a son of Lincoln … automatically itself comes to be. (TII, p. 115)

So now we have an answer to the question of why impure properties can only come into existence when the particulars they involve do.This conclusion is entailed by Howell’s account of the nature of impure properties: the thesis that impure properties are complex objects that have the particulars they involve as constituents. Such properties cannot exist until all of their constituents do. Having said this, it seems to me that Howell’s conception of the structure of impure properties is itself inadequately motivated. A property, after all, is nothing but a condition something meets, or would have to meet, in order to be of a certain kind. Such properties may vary in complexity in the sense that a particular’s having a certain property may or may not consist in its entering into relations with other particulars. Being a son of Abraham Lincoln is one such example. But what I fail to see is why such complexity should be ontologized: why, in other words, we should read the complexity into the structure of the property itself by having the entities to which a particular must be related in order to have the property as very constituents of that property. It can only be a mistake to think that the property itself – that entity – is essentially related to Lincoln (in the sense that Lincoln is an essential part of it). The thing that must bear a certain relation to Abraham Lincoln (if it is to have the property in question) is the concrete particular. To specify a property is to specify what is required for a particular to be a certain way.And to have the property being a son of Abraham Lincoln, a particular must instantiate a certain relation to Lincoln. The thing that must be related to Lincoln is a particular that has the property, not the property itself. The property is not complex in an ontological sense. Howell is sliding from a thesis concerning the instantiation of impure properties to a thesis concerning their ontological nature. My charge could perhaps be put another way.To be sure, in specifying an impure property, we refer to concrete particulars. But we refer to such particulars only to spell out the nature of the condition that must be met by something, if it is to have the property in question. It does not follow that we should regard the concrete particulars themselves as parts of the property. Hence, since Howell’s objection to the eternal existence of properties rests upon just this assumption, this objection is toothless.



defending musical platonism

Needless to say, this is not yet to explain how an impure property can exist at a time before the concrete entity that it involves exists. But this, it appears to me, is a philosophical problem more manufactured than real. Once more, we need to distinguish carefully the question of whether the property exists from the issue of whether it can be instantiated. It would indeed be crazy to suppose that being a son of Abraham Lincoln is a property that could have been possessed by anything before there was such a person as Abraham Lincoln. Likewise, it would be extremely odd to think that my grandfather, Jim Clark, had the property being an ancestor of Julian Dodd before I was born. But such facts do not entail that the properties themselves did not exist then. My grandfather only became an ancestor of mine when I was born, but what happened then was, to my mind, not a case of my birth bringing about the birth of a property, but an already existent property coming to be instantiated. It is, of course true that in 1600 no one would have been able to express the property of being an ancestor of mine, because no one would have been able to refer to me. But, as we noted earlier, this is a point about reference, not the metaphysics of properties. The property existed then, even though it could not have been introduced into a language. But, as with facts, there would seem to be a category mistake involved in thinking that properties of any kind can come in or out of existence. I take properties to be abstract entities, and Howell appears to agree with me:23 it is far better, I think, to hold that concrete particulars instantiate abstract universals than to hold that the same universal is, at a given time, wholly present in more than one place. But, having said this, it is quite unclear what it would be for an abstract entity to come in or out of existence, unless, like certain sets, it possessed concrete entities as essential constituents. For the problem is this: as I said in my original article,24 the coming into existence of an abstract entity would have to be a causal process, and it is a plausible principle that abstracta cannot enter into causal relations. Of course, once the Eiffel Tower was built, the singleton containing the Eiffel Tower thereby came into existence, but the fact that such sets can come in and out of existence does not violate the principle of the causal inertness of abstracta: the causal process in this case involved people and bits of metal, the coming to being of the set being an ontological free lunch. When it comes to properties, however, things are very different. Properties are not sets: they do not have concreta as essential constituents. Consequently, if a property could come into existence, there would have to be at least one causally efficacious abstract entity. And there can be no such thing.

93

At this juncture, Howell might seek to reply by claiming my account of properties to be obscure. Just what are these abstract, and yet ontologically simple, entities? To which I can make three replies. First, properties, as I have said, are conditions that particulars may meet. The mistake, in my view, is to assume that properties must have a structure isomorphic to the predicates that express them. This assumption that language mirrors the structure of the world, though beloved by the logical atomists, is little more than a dogma. Second, if Howell were to charge me with obscurity, I could reply with ‘tu quoque’. The claim that Abraham Lincoln fills a ‘slot’ in being a son of someone is a metaphor as uncashable as the Fregean declaration that properties are ‘unsaturated’ things that may be completed by ‘saturated’ objects. In what sense does an abstract entity have a slot? How could a concrete particular fill such a slot in what is an abstract entity? And why not think of particulars as having the slots (to be filled by properties)? I remain to be convinced that these questions are anything but unanswerable. But, finally, I would dispute that my remarks about properties are obscure. They only seem so if one is in the grip of the presumption that properties must be complex objects, and that the philosopher’s task is to analyse this ontological complexity. Properties are sui generis abstract entities that cannot be assimilated to complex objects such as facts or states of affairs.This view may be favourably contrasted with Howell’s position on impure properties. Presuming them to be structured entities, he says nothing non‐metaphorical about the nature of their structure. I dispute the need for the kind of theory that, in the end, even Howell cannot deliver.

IV. Types, Indicated Types and Patterns So much for Howell’s objection to (9): the claim that properties are eternal. His claim that impure properties are a counterexample to this thesis is based upon a misconceived way of thinking about such things. But Howell’s critique does not stop here. He also argues that (8) – my claim that type K exists if and only if the property being a k exists – is false. Naturally, if Howell were correct in this, the failure of his objection to the thesis that properties are eternal would be of no consequence: the argument from the eternal existence of properties would be unsound, albeit for one reason rather than two. None the less, the thinking behind (8) is simple and, so it seems, persuasive. To begin with, it is unproblematic that the biconditional holds left to right: if K exists, then there

94

julian dodd

must also be a condition (that is to say, a property) which something must meet to be a k. Equally, it would appear to be true that if being a k exists, then so does K. If, for example, the property being an elephant exists, and hence it is clear what it would be for something to be an elephant, then the type The Elephant surely exists also. If there is a condition that something must meet to be an elephant, then nothing else is needed for the corresponding type to exist. But however plausible this sounds, Howell nevertheless denies (8). In his view, (8)’s biconditional fails to hold from right to left. The existence of a property does not, he claims, entail the existence of its corresponding type. However, this conclusion is a consequence of Howell’s invocation, and motivation, of an unfamiliar ontological framework. So it would be as well to begin with a brief outline of Howell’s ontology of properties, patterns, and types. According to Howell, the existence of being a k does indeed entail the existence of an abstract entity, but this is the K‐pattern, not the type K (TII, p. 116). The pattern, explains Howell, is the mere ‘arrangement of parts or features that is possessed by anything having the associated property’ (ibid.). The associated type, meanwhile, is not a simple arrangement of parts; it is this arrangement once it plays an ‘important rôle in nature or human life’ (ibid.). So even if we presume that the properties underlying types are eternal, it does not follow that all types are eternal too. We may say that patterns are eternal existents, but the same is not true of every type. Indeed, Howell argues that linguistic, technological, and cultural types (such as word‐types, games, and The Ford Thunderbird) all come to be initiated only once a network of human practices has come into existence, and hence are the literal products of the relevant practices. Needless to say, if Howell could persuade us that The Ford Thunderbird  –  the type – was man‐made, the analogous case for In This House, On This Morning would be that much easier to make. At the very least, if The Ford Thunderbird is an indicated type, and if Howell can assure us that it was a human product, then Levinson’s corresponding suggestion for musical works will not be susceptible to my original objection. Howell will have shown us that indicated types are not eternal. What, then, does Howell have to say about such types seemingly bound up with human practices? According to Howell, the relation between human practices and a type of this kind is one of production. Types such as these, says Howell, are all initiated. They exist only once the practices in question do.… All these practice‐related types are themselves to be distinguished from mere patterns that are not currently involved in practices. (TII, p. 124)

So Howell’s thought is this: a word‐type, such as ‘dine’, only exists once the communal practice of using tokens of it with its meaning exists. Before there existed this communal practice (indeed, even before there were any English‐speakers) there existed the property being a /dine/phonemic sequence used in English to mean dine, and hence there existed the corresponding pattern; but it would be absurd, supposes Howell, to think that the word‐type itself existed in the Jurassic age. There were no words before there were any people, and this is because a word only comes into being once ‘some community establishes a practice of producing (and recognizing) concrete sound and visual items … that instantiate the relevant pattern and so have the property that underlies it’ (TII, p. 118). Howell then delivers the coup de grâce with what he considers to be a reductio of his opponent: Words do not exist in a natural language merely because properties exist that specify sound‐and‐meaning patterns that are admissible in that language. ‘Glank’ does not exist as an actual word (type) in English, meaning ‘joyful jeep’, simply because there exists, uninstantiated, the property being a /glank/ phonemic sequence in English to mean joyful jeep …. ‘[G]lank’ was not, before I wrote this paper, an uninstantiated but existent word in English. It was no such word at all, and it remains no word now. (TII, p. 117)

A word‐type, so Howell suggests, only comes to exist once a practice of using the word with a particular meaning is established. To Howell, it is absurd to suppose that a word‐type could pre‐exist such practices. But if Howell’s intuitions about word‐types are correct, how could the corresponding thesis for works of music be any less absurd? What price on In This House, On This Morning pre‐existing Marsalis’s compositional act: the act that established the practice of producing and recognizing instances of the said work? The dialectic of Howell’s challenge is clear. We are supposed to accept that cultural and linguistic types are temporally initiated, see musical‐works‐as‐construed‐by‐Levinson (and indicated types quite generally) as analogous to the former entities, and hence come round to Levinson’s opinion that these indicated types are temporally initiated too. But in my view, the first step in this reasoning is mistaken. The claim, for example, that word‐types are temporally initiated is eminently resistible. Indeed, in my view Howell is wrong about word‐types, and hence is not entitled to the premise from which he seeks to derive the conclusion that works of music are temporally initiated too. To see this, we should, first of all, note that, although the word‐type ‘glank’ does not presently exist in English, it does



defending musical platonism

not follow that it does not exist at all. I take it that for a word to exist in a language L at t is for it actually to have been used by speakers of L at t or some time before t. Only word‐types that have actually been tokened by speakers of L count as being words of L. A word only enters into a language once a practice of producing tokens of it, and recognizing such tokens, has been established. Communal practices, then, bring word‐types into languages. So it is, indeed, absurd to suggest that ‘glank’ is presently an actual word‐type of English. But having located the source of the absurdity in the picture Howell objects to, we are able to distinguish the truly absurd from the harmless. For although Howell has demonstrated that ‘glank’ is not presently a word‐type in English, he has not shown that the word‐type does not exist. The Platonist’s point is not, as Howell would have him say, that ‘glank’ is an uninstantiated but existent word in English: a word‐type currently present in English but which no one has yet used. It is that ‘glank’ is an existent word‐ type that has not entered into English (or any other language). If a practice of actually using the word were to become established, it would thereby become a part of English. But this has not yet happened and, I dare say, never will. For the Platonist, human practices are crucial to understand language‐use, but their function is not that of bringing word‐types into existence; it is that of bringing (eternally existent) word‐types into languages. Howell’s example has not, of itself, subverted this idea. Nevertheless, Howell might respond by claiming that his distinction between patterns and types enables us to give the best explanation of what happens when, for example, a word‐type first enters the language. As Howell sees it, what happens is this: when, through the coming into existence of communal practices, the word‐type enters the language, ‘the pattern becomes a type’ (TII, p. 119). And we have no choice but to regard the type as a new entity, distinct from the pattern, because the type has a property essentially which the mere pattern does not have essentially: namely, the ‘property of actually being used in the community in order to carry those qualities’ (TII, p. 119), or, as Howell also describes it, ‘the property of being used in the way specified by the practice’ (TII, p. 120). Howell’s thought is this: a word‐type such as ‘dine’, unlike the pattern, would not be that entity unless it was used in that way (and thus had that meaning). The same goes, Howell claims, for the hooked cross used in pre‐Columbian Amerindian cultures used to signify, amongst other things, fire. Although it shares a geometric pattern with the Nazi swastika, it is not identical with it because the two symbols have distinct uses, and hence differ in meaning (TII, pp. 119–120).

95

But once more, Howell’s discussion of examples, interesting as it is, does not succeed in making his case. According to Howell, the hooked cross and the swastika share a common (eternal) geometrical pattern, but the types themselves, with their different meanings, only came into existence when their respective communities established their distinct meaning‐determining practices. But if Howell supposes that we can only explain the difference between two symbol‐types in this way, he is mistaken. For here is another explanation. Let us call the eternal geometrical pattern shared by the hooked cross and the swastika θ. The two symbol types have the following distinct property‐associates: being a θ mark used to signify fire and being a θ mark used to signify the German Nazi Party. The condition that a particular must meet to be a token of one is distinct from the condition that must be met by a particular, if it is to be a token of the other. Consequently, since the properties that underlie the respective types are different, so are the types themselves. Now, on my view, both of these properties are eternal and, as we saw in Section IV above, Howell does not succeed in refuting this view. Consequently, given (8) – the fact that a type exists if and only if its property‐associate exists  –  it follows that both types are eternal too. Naturally, the Nazi swastika is individuated in such a way that it could not have been tokened before the Nazi Party existed, but to suppose that this entails that the type itself did not exist before this time is to confuse the existence conditions of a type with those of its tokens. Of course, Howell disputes (8), but if his reason for doing so is that the Platonist cannot adequately make out the difference between the hooked cross and the swastika, he is wrong.The hooked cross and the swastika are distinct indicated types, but this difference is not located in the putative fact that respective communities produced them. They differ because the properties underlying them are distinct properties. This is not to deny that types are in some sense bound up with human practices. Our specification of the swastika just is a specification of a way of using the geometrical pattern.This symbol‐type is nothing more nor less than a kind of use that the pattern can be put to in a certain kind of practice. But the establishment of the practice is not required for the symbol‐type to exist, only for the already existent symbol‐type to be discovered and tokened. Actual communal practices are the means by which we come across indicated types such as this. This is unlikely to convince Howell, however. As we have seen, he regards a symbol‐type as possessing a property essentially which it could only come to have had once the relevant communal practice was in place.This property, ‘the pattern’s property of actually being used in the community’

96

julian dodd

(TII, p. 119; my italics) to carry semantic or expressive qualities, is what supposedly distinguishes the type from its mere pattern, and which entails that the type only exists once the practice exists. But, once more, there is no need to interpret the place of communal practices in this way. According to the view I favour, it is not an essential property of the swastika – the type – that it actually be used to symbolize the Nazi Party. For this type might never have been tokened. Being actually used to symbolize the Nazi Party, rather than being a property essentially possessed by the swastika, is a property that must be possessed by any token, if it is to count as a token of the swastika. The type merely has the following property essentially: being such that its correct use is to symbolize the Nazi Party. But a type can have this property without actually being so used, or even being in a position to be so used. So it follows that the type was not brought into existence once the practice of using it to symbolize the Nazi Party was set up. To my mind, in taking the property of actually being used in a certain way by the community to be an essential property of a symbol‐type, Howell once again confuses types with tokens. We should agree with Howell that cultural, technological, and linguistic types are distinct from their corresponding patterns. The types, but not the patterns, are specified by making reference to communal practices. But it does not follow that types only come to exist once the relevant practices are in place. Howell has thus failed to provide counterexamples to premise (8) of the argument from the eternal existence of properties. And now the implications for Howell’s attempt

to show how a musical work, if an indicated type, could be created by its composer, should be clear. Musical works, if indicated types, are cultural types that we specify by making reference to a particular individual’s compositional act. But if a word‐type exists prior to the practice that makes it actual use possible, then we have no reason to deny that the indicated type with which Levinson and Howell identify a work of music exists prior to its composer’s compositional act. Even if In This House, On This Morning were an indicated type (which I deny for the reasons given above in Section I), it existed before Marsalis composed it. Types – all types – are eternal existents, and Howell has said nothing to undermine this claim.

V. Conclusion Having considered the objections raised by Howell, Sharpe, and Trivedi, I remain more convinced than ever of the truth of the conclusion of ‘Musical Works as Eternal Types’. The simple view of musical works as sound structures is the most natural ontological proposal on offer.What many take to be a counterintuitive and paradoxical consequence of it – namely, that musical works are eternal entities – is, in fact unproblematic. There is no harm in taking composition to be a kind of creative discovery. Given all this, and given the fact that Levinson’s counterproposal cannot allow for the literal creation of a musical work either, we should only abandon the simple view if a more convincing kind of objection to it is forthcoming.

Notes 1 Julian Dodd, ‘Musical Works as Eternal Types’, British Journal of Aesthetics, vol. 40 (2000), pp. 424–440. 2 Jerrold Levinson, Music, Art and Metaphysics (Ithaca, NY: Cornell U.P., 1990), p. 64. 3 The musical Platonism I defend has much in common with the respective positions set out and defended by Nicholas Wolterstorff and Peter Kivy, respectively. See the former’s Works and Worlds of Art (Oxford: Clarendon Press, 1980), Part Two, and the latter’s ‘Platonism in Music: A Kind of Defense’, Grazer Philosophische Studien, vol. 19 (1983), pp. 109–129 and ‘Platonism in Music: Another Kind of Defense’, American Philosophical Quarterly, vol. 24 (1987), pp. 245–252. 4 See, for example, Levinson, Music, Art and Metaphysics, p. 65, and Stefano Predelli, ‘Musical Ontology and the Argument from Creation’, British Journal of Aesthetics, vol. 41 (2001), pp. 279–292.

5 Levinson, Music, Art and Metaphysics, p. 79. 6 Dodd, ‘Musical Works as Eternal Types’, pp. 427–434. 7 Ibid., p. 428. 8 My commitment to this argument is evident in ‘Musical Works as Eternal Types’, pp. 434–440. I formulate it here as steps (5)–(10) for the first time. 9 Predelli comes to a similar conclusion. See his ‘Musical Ontology and the Argument from Creation’, p. 289. 10 Robert Howell, ‘Types, Indicated and Initiated’, British Journal of Aesthetics, vol. 42 (2002), pp. 105–127. Hereafter TII. 11 R.A. Sharpe, ‘Could Beethoven Have “Discovered” the Archduke Trio?’, British Journal of Aesthetics, vol. 41 (2001), pp. 325–327. Hereafter AT. 12 Saam Trivedi, ‘Against Musical Works as Eternal Types’, British Journal of Aesthetics, vol. 42 (2002), pp. 73–82. Hereafter AMW.



defending musical platonism

13 For example,Trivedi’s claim that our artistic practice embodies an assumption that musical works are created (AMW, p. 75), and his suggestion that my denial of this assumption is incompatible with the special status accorded to composers (pp. 76–77), were objections countered in ‘Musical Works as Eternal Types’ at pp. 428–433. 14 As has been argued by: P.F. Strawson, ‘Aesthetic Appraisal and Works of Art’, reprinted in his Freedom and Resentment and Other Essays (London: Methuen, 1974), pp. 183–184; Gregory Currie, An Ontology of Art (Basingstoke: Macmillan, 1989), ch. 4; and Eddy Zemach, ‘No Identification with‐out Evaluation’, British Journal of Aesthetics, vol. 26 (1986), pp. 239–251. 15 Dodd, ‘Musical Works as Eternal Types’, p. 433. 16 See Levinson, Music, Art and Metaphysics, p. 67. 17 See the discussion in ‘Musical Works as Eternal Types’ at pp. 428–429. 18 We fail to distinguish between these two kinds of claim at  our peril. My writing desk and the singleton with this object as its member exist at the same possible worlds, but  to  admit this is not to identify my writing desk with the  singleton. My writing desk is a concrete particular, not a set.

97

19 Howell seems to suggest such a reading when he says that ‘Using that property retrospectively, we in 2002 can then consider Socrates, who died in 399 bc, and note that he was not a son of Lincoln’ (TII, n. 13).The implication here is that the fact noted in 2002 did not exist in 399 bc. 20 It is because we describe facts as neither beginning to exist nor as going out of existence that it sounds so odd to describe facts as in any sense lasting. If something has no beginning or end, then it cannot be said to last. 21 This point is well made by Zeno Vendler in his ‘Facts and Events’, published in his Linguistics and Philosophy (Ithaca, NY: Cornell U.P., 1967), p. 144. 22 This talk of impure properties ‘involving’ concrete particulars is Howell’s own (TII, p. 126). 23 This would seem to be evident in Howell’s talk of ‘a property (or, more generally, … any sort of metaphysical abstractum)’ (TII, p. 112). It is also suggested by the fact that, while he denies that being a son of Abraham Lincoln is ‘brought into being by someone’s generating an abstract entity out of nothing’ (ibid., p. 115; my italics), he does not deny that such properties are man‐made abstracta. Trivedi too takes properties to be ‘abstract, generic entities’ (AMW, p. 74). 24 Dodd, ‘Musical Works as Eternal Types’, pp. 431–432.

9

Against Musical Ontology Aaron Ridley

Many recent philosophers of music have thought it ­interesting, and possibly even useful, to try to say what sort of thing a musical work is, to specify, for example, the conditions that a performance must satisfy if it is to count as a performance of the work it purports to instantiate. They have thought it self‐evidently worthwhile, in other words, to raise questions about musical ontology; and they have thought that answers to those questions might facilitate, or even be required for, reflection on apparently adjacent areas of interest, such as musical performance. It is my contention here that all such thoughts are mistaken, and that a serious philosophical engagement with music is orthogonal to, and may well in fact be impeded by, the pursuit of ontological issues, and, in particular, that any attempt to specify the conditions of a work’s identity must, from the perspective of musical aesthetics, be absolutely worthless.

I Philosophical enthusiasm for musical ontology was fueled, if not actually created, by the publication of Nelson Goodman’s Languages of Art,1 where it was claimed that a musical work is identical to the score in which it is notated, and so that the score “defines” the work (ibid., p. 178). It followed from this, on Goodman’s view, that “complete compliance with the score is the only requirement for a

genuine instance of a work”—or, to put it rather more starkly, that even “the most miserable performance without actual mistakes [counts] as such an instance, while the most brilliant performance with one wrong note does not” (ibid., p. 186). This, not surprisingly, has struck most readers as wildly implausible – as Goodman, of course, knew that it would. “Could we not,” he asks, “bring our theoretical vocabulary into better alignment with common practice and common sense by allowing some limited degree of deviation in performances admitted as instances of a work?” But the answer, he holds, is no, we could not. For this is one of those cases where ordinary usage gets us quickly into trouble. The innocent‐seeming principle that performances differing by just one note are instances of the same work risks the consequence  –  in view of the transitivity of identity  –  that all performances whatsoever are of the same work…. Thus while a score may leave unspecified many features of a performance, and allow for considerable variation in others within prescribed limits, full compliance with the specifications given is categorically required (ibid., pp. 186–87).

No one, so far as I am aware, has thought that Goodman is right about this. The general feeling has been that any theory this seriously at odds with the intuitions shared by more or less everyone who has ever listened to or played a  piece of music must be (and whatever “the transitivity of  identity” might suggest) mistaken. It should be noted,

Aaron Ridley, “Against Musical Ontology,” Journal of Philosophy 100 (2003), pp. 203–20. Reproduced by permission of The Journal of Philosophy.

Aesthetics and the Philosophy of Art – The Analytic Tradition: An Anthology, Second Edition. Edited by Peter Lamarque and Stein Haugom Olsen. © 2019 John Wiley & Sons, Inc. Published 2019 by John Wiley & Sons, Inc.



against musical ontology

however, that Goodman is driven to conclude as he does by reasons that are, in the end, peculiar to his own highly distinctive theoretical priorities  –  in this case, most proximately, by a markedly rarefied and abstract conception of what it is for something to count as a “notational system” (ibid., pp. 148–57). It is his thinking about “notational systems,” that is, that both leads him to the views about works and performances that he espouses, and leaves him relaxed in the face of their extremely counterintuitive character. This suggests that the natural move, if one disagrees with his conclusions, must be to declare oneself motivated by different priorities – by musical ones, say – and pass by on the other side; and I think that this is in fact the right move to make.There is no need, it seems to me, to be drawn into argument here. That is not, however, how things have gone. What the literature shows instead is a quite remarkable willingness to take Goodman as having thrown down the gauntlet. What, it is asked, if one disagrees with Goodman’s conclusions, is a musical work identical to (if not the score in which it is notated)? What is it for a performance to be of the work it purports to be of (if not getting all the notes right)? Thinking about musical performance has been assimilated, that is, to asking what might constitute the essence of a musical work, and what conditions anything that is to be taken as an “instance” of a given work  –  for example, a performance of it – must satisfy. Ontological questions, in other words, have been placed firmly in the front line. For instance, and to tease out just one strand in the post‐ Goodman literature, a number of broadly Platonist positions have been advanced, in conscious opposition to Goodman’s own nominalism. Peter Kivy,2 for example, has argued for a fairly full‐blooded, if forgivably broad‐brush, version of Platonism, suggesting that “works are universal, or types, or kinds, performances related to them as particulars, tokens, or instances” (ibid., p. 75); that a “sound structure” (considered as a universal, type, or kind) is “a concrete identity criterion” for a work’s being the work that it is;3 and that such sound structures are notated in scores, where a score, “in a loose sense of ‘uniquely’,” uniquely determines a correct performance, under a given set of implicit conventions for interpreting the score—conventions which may be quite different in different historical contexts—and a performance uniquely determines a score under a similar set of historically bound conventions.4

Kivy’s Platonism, then, identifies a musical work with an (abstract) “sound structure” and regards as “correct” any

99

performance that allows that structure  –  or something “loosely” like it – to be heard. He also expressly embraces the consequence of his Platonism, namely, that if “works are universals, or types, or kinds” they can neither come into nor go out of existence, and hence that the “composing” of works, so construed, must be a matter of discovery rather than creation (ibid., pp. 38–47). It is partly out of a reluctance to accept this sort of conclusion that Jerrold Levinson5 has proposed a different kind of Platonism.6 According to this, a work is to be regarded, not merely as a “sound structure,” that is, as “a sequence of sounds qualitatively defined,” but as “a compound…of a sound structure and a performing‐means structure,” where the latter is “a parallel sequence of performing means specified for realizing the sounds at each point” (ibid., p. 78). Levinson suggests that we call the resultant compound “an S/PM structure, for short.” S/PM structures, described by Levinson as “implicit types,” include every logically possible combination of sounds in every logically possible combination with every logically possible combination of performing means. Such types are Platonic through and through: they can neither come into nor go out of existence. But not all S/PM structures are works; the overwhelming majority of the possibilities that they represent never have been, and never will be, realized. What distinguishes works from other S/PM structures is that they are what Levinson calls “initiated types”—that is, types that “begin to exist only when they are initiated by an intentional human act”: works, then, as “initiated types,” can be said to be brought into being, created rather than discovered, by the intentional act of composing (ibid., pp. 80–81). A performance of a given work is then specified as “a sound event that is intended to instantiate” the work, which “represents an attempt to exemplify” the work’s “S/PM structure in accordance with” the composer’s “initiation” of it – “and which succeeds to a reasonable degree” (ibid., p. 86). These examples should give some indication of the sorts of effort that have been made to address the ontological questions I mentioned a moment ago, while also attempting to avoid some of the counterintuitive consequences of Goodman’s position. There is, so far as I am aware, no consensus – nor even a hint of it on these matters, except perhaps for a growing feeling that the sheer diversity of things that might, in various times and places, be thought of as music or as performance makes it unlikely that any monolithic account, however tolerant of deviation, will suffice. A more catholic approach to ontology has emerged accordingly – an approach perhaps most comprehensively evinced by Stephen Davies’s recent book, Musical Works and Performances: A Philosophical Exploration.7 This does, as I say,

100

aaron ridley

represent a development, but it does not represent any kind of departure: ontological questions still dominate. My view, as I said at the outset, is that the whole move to ontology in thinking about music is a mistake. Before saying why, though, I should make clear how I will be taking the term ‘performance’ from now on. By ‘performance’ I will mean not only the playing of a work by an individual or group before an audience, but also recordings, transcriptions, arrangements, versions and, in general, renditions of every sort. I take my lead in this from Kivy,8 who argues, first, that a performance is “an arrangement or version of the musical work that has been performed, and, as such, a subject of the kind of evaluation and aesthetic satisfaction that artworks support and provide,” and, second, that the “performer is an artist, somewhat akin to a composer or, better, ‘arranger’ of musical works” (ibid., p. 261). The relatively capacious conception of performance that results will allow me, I hope, to make the argument I want to make in a reasonably direct manner, without getting bogged down in (or indeed traducing) such finer‐grained details as might or might not be at issue.

II When was the last time you came away from a performance of a piece of music – live or recorded – seriously wondering whether the performance had been of it? My guess is, never. Even if you were to hear the Bach‐Busoni Chaconne, say, as played by me, or by some comparably giftless pianist, it still would not occur to you to doubt that the victim of the musical murder you had witnessed was, indeed, the Chaconne. In fact, if it were played any worse, in such a way that the victim was actually unidentifiable, there would be no reason to think that you had witnessed any sort of performance at all – no reason to think, that is, that you had witnessed a performance of anything. Indeed, it is precisely because doubts of the relevant sort have next to no tendency to arise that muzak’d versions of, say, the Ode to Joy theme have the power to enrage and depress that they do. One would not suffer as one does in elevators and supermarkets if the doubts were real. One would not be reduced to misanthropic cursings, to mutterings of “How could they? How dare they?” if it really did strike one as a serious possibility that the miserable, denatured pap oozing from the speakers was not the Beethoven after all. It clearly is the Beethoven, and that is why it makes one feel so low and vicious. In cases like these, then, one’s doubts are not about what a rendition is of  –  far from it  –  but about

whether any penalty could feasibly reflect the gravity of the offense. These thoughts, admittedly much understated, suggest two conclusions. The first is that, in our ordinary – indeed in our actual – aesthetic encounters with renderings of pieces of music, our primary concern, or at the very least one of our most prominent concerns, is whether a given rendition is any good, or, if it is not, whether it is so bad as to merit further action. As listeners, that is, we are chiefly alert and sensitive to issues about the value of what we are  hearing. The second conclusion is that, since these ­sensitivities are operative, and even virulent, against the background of an apparently rather robust sense of work identity, issues concerning work identity can hardly be very urgent if what we are chiefly interested in is our aesthetic experience of renditions of pieces of music. If we are doing aesthetics, that is, ontological questions deserve a place in the back row, at best. I think that both of these conclusions are correct. Neither, however, commands evidently overwhelming support from contemporary philosophers of music, as a moment’s reflection on the material mentioned in section I will confirm. What we find there is, essentially, a range of views about the room for maneuvering that a performance has before it ceases to be a performance of what it was supposed to be of, backed up by a range of views about what constitutes a work’s essence or identity. From this sort of starting point, it is hardly surprising that evaluative issues, where they get a look in at all, are confined to questions of legitimacy—to asking whether this or that performance of a given work is faithful enough to whatever happens to be the favored set of identity conditions to count as acceptable, legitimate instances of performances of it. The question whether this or that performance, or style of performance, is actually any good, or is minimally worth listening to, is scarcely raised.9 If one is serious about the philosophy of music, this last fact should strike one as scandalous. Not as inexplicable, perhaps: philosophers are drawn to metaphysics, and the chance to go on about identity conditions, and related ontological matters, has possibly been found too good to pass up. But one really ought to be shocked, even so. The most direct way to bring out why, perhaps, is to note that an indifference to genuinely evaluative issues  –  to issues that animate the experience of actual listeners when they listen to actual pieces of music – presupposes a sharp distinction between what it is to take a philosophical interest in music and what it is to take a critical interest in it. It is true that such a distinction can be drawn. It is true, that is, that the philosophy of music is not identical to music



against musical ontology

criticism. But the distinction is not, and cannot be made to be, a sharp one, for unless one’s philosophical engagement with music is driven by, and is of a sort that might pay dividends for, one’s musical experience  –  including one’s evaluative experience – there is no obvious sense in which one is engaged in philosophical aesthetics at all. At best, and instead, one might be engaged in the metaphysics of music, and that is a very different activity. I can imagine two kinds of response to what I have just said. The first is simply to grant that what is being done is indeed musical metaphysics rather than musical aesthetics: If one has made no attempt to do aesthetics, after all, how much of a scandal can it be that one’s results are aesthetically inert? I propose, for the moment, to accept this response – although I will be coming back to it. It is the second response, however, that most self‐avowed philosophers of music are likely to reach for. According to this, one cannot hope to do justice to the kinds of evaluative concern that I have mentioned unless one has got one’s metaphysics – one’s ontology – into good shape first.There is no point, that is, in asking whether a performance of such‐ and‐such is of any value until one has determined that that performance is, indeed, of such‐and‐such; and this means that one must first work out what it is to be such‐and‐such, that is, determine the relevant identity conditions, and then satisfy oneself that the putative performance meets them. Only after that has been done, the response concludes, can one ask whether the performance is, aesthetically, any good – first things first. This way of thinking strikes me as quite profoundly mistaken; and the examples I have already given, the muzak example especially, show why. Imagine yourself seized with a furious sadness at some particularly vile, beat‐enhanced, edges‐smoothed‐over version of the Ode to Joy theme. And now imagine yourself thinking as the second response invites you to. On the face of it, things could go in either of two ways, depending on whether one takes one’s ontology prescriptively or descriptively. Taken prescriptively, one might decide that what one was hearing did not meet one’s own rather stringent conditions for work identity, that is, that the muzak was not really, and despite appearances, a rendition of Beethoven’s theme at all. But in which case, why the rage? Or  –  and still prescriptively  –  one might decide that the muzak did meet one’s conditions, in which case it is hard to see how that decision, given one’s enraged response, has left one any the wiser. And taken descriptively the result is the same: one’s enraged response shows either that one’s identity conditions are wrong (if the muzak does not satisfy them) or that they are superfluous (if it

101

does)  –  superfluous in the sense that their deployment is not any sort of precondition of pointfully raising questions of aesthetic value. This way of thinking is surely a matter of putting misleading or redundant things first, not first things. It will be objected, no doubt, that I have stacked the deck in my favor by stipulating that what one hears fills one with rage and depression. It is only this stipulation, after all, that makes preemptive ontological beavering look either misleading or redundant. But this objection has force, clearly enough, only if ontological considerations, in musical contexts, are allowed to trump one’s experience of what one actually hears. And the single case in which that possibility has any potential bite is the first one I considered  –  where one decides, on prescriptive work identity grounds, that what one is hearing is not the Ode to Joy theme at all, and therefore, presumably, that one should not (necessarily) mind it. But here, I have to say, I begin to lose touch with what might be being said. In the sort of example I am imagining – which is to say, a case where the Ode to Joy theme is clearly being murdered – one could only, surely, decide that it was not that theme after all if it had first occurred to one that it might be, if it sounded enough like it to be worth giving it the ontological once‐over for that possibility, specifically. But now it looks as if the decision to deny the title Ode to Joy to what one is hearing (assuming that that is the decision one arrives at) cannot be anything more than an elliptical and rather unhelpful way of expressing one’s disapproval at what has been done to Beethoven’s original.The idea of wondering, perfectly seriously, and from scratch, as it were, whether the prevailing ontological conditions are such as to warrant one’s disapproval is either impossible to understand, it seems to me, or else a symptom of something close to aesthetic autism. But perhaps I am tilting at straw men here. Perhaps the position I should be considering is not, as I have been assuming, that one must always, in every case, get one’s ontology straight before evaluative questions can pointfully be raised, but rather that evaluative questions will, in general, be more perspicuously framed if raised against the backdrop of a convincing ontology. And if this is right, it might be thought, metaphysically inclined philosophers of music need neither be flummoxed in the face of their muzak‐inspired paroxysms, supposing they have them, nor fearful for their status as aestheticians. On the face of it, there is no doubt that this modified and more modest position is an improvement on its predecessor. But it is still deeply unpersuasive. How, exactly, is a convincing ontological backdrop supposed to lend perspicuity to evaluative questions? No one, so far as I am aware,

102

aaron ridley

has actually asked this: certainly no one has given any sort of explicit answer. But we can begin to see how such an answer might go from something that Davies says: To be of a work, a performance must satisfy three conditions. There must be a suitable degree of matching between the performance and the work’s contents, the performers must intend to follow most of the instructions specifying the work in question…, and there must be a robust causal chain from the performance to the work’s creation, so that the matching achieved is systematically responsive to the composer’s work‐determinative decisions (op. cit., p. 5).

Davies is here addressing the metaphysical‐cum‐evaluative question – what conditions must a performance of a work satisfy if it is to count as a legitimate performance of it? And his answer – that the performance must “match” the contents of the work, where those contents are determined by the composer’s decisions as embodied in the instructions (as found, for instance, in the score) which mark out the work as the work that it is – gives us what I have been calling the “ontological backdrop.” According to this, then, a performance is legitimate if it “matches” the work’s contents. How might one attempt to understand this backdrop as lending perspicuity to the framing of genuinely evaluative questions? What, in other words, if this is the backdrop, does it suggest that a question about the value of some particular performance ought to look like? The only possible answer is this: How well, how closely does the performance “match” the contents of the work?  –  with the presumption that the better it “matches” the better it is. One can see how someone might offer this sort of formulation in an effort to capture what we mean when we commend a performance for its faithfulness to the work performed; and indeed, one can see how someone might think that, in order to make or to understand such a commendation, something appropriately ontological must be going on in the background. But a moment’s reflection should be sufficient to show that the notion of “matching” is incapable of doing any such work. The problem, of course, is that to make the value of a performance a function of how well it “matches” the contents of the work performed, it would be necessary to be able to specify, quite independently of this or that performance, what, precisely, those contents were – what exactly, in other words, the performance was supposed to “match.” It would be necessary, that is, to be able to give an exhaustive prescription for the production of a performance that was, not merely legitimate, but excellent, first‐rate, admirable. And that is, of course, quite impossible, in just the same

way and for just the same reasons that it is impossible to give an exhaustive prescription for the production of a great work of art. However apt Davies’s talk of “matching” might be in the context of legitimacy, then, it certainly fails to carry over into genuinely evaluative contexts. His ontological reflections, that is, do nothing  –  as backdrop  –  to assist in the perspicuous framing of evaluative questions. Indeed, quite the reverse: they actively encourage the framing of those questions in a misleading way. Nor is this a problem peculiar to Davies. I have chosen to discuss Davies explicitly only because he, explicitly, refers to “matching.” But the point is quite general: to the extent that one’s ontological reflections are geared to the specification of work identity conditions  –  that is, to conditions that might be satisfied by a range of different performances – one is necessarily committed to the specification of conditions that are, in the operative sense, not specific to any performance in particular  –  that is, to the specification of conditions of precisely the sort that “matching” talk is most obviously suited to. And this means, clearly enough, that the kinds of discrimination that one is able to make within the relevant range will not include discriminations between good performances and merely legitimate ones—a point brought out forcefully by Peter F. Strawson:10 Objects primarily of aesthetic assessment [for instance, performances] have plenty of shareable properties: there are plenty of ways in which we find resemblances between them. But in naming these, we do not name, in non‐evaluative terms, those features directly on account of which we make aesthetic judgments of the individual bearers of those properties; for either these names of shareable general properties are themselves evaluative, or, if they are not, then, in applying them, we leave our listeners in the dark as to what evaluations to make of the individual [performances] to which they are applied (ibid., p. 188).

The thought that a well‐ordered ontology is any sort of precondition of perspicuously framed evaluative questions founders, and founders decisively, on this point.11 Things get even worse for the purveyor of identity conditions, as we can see if we ask what is meant by a performance’s being “faithful” to a work  –  the genuinely evaluative notion, as I have just suggested, that no talk of “matching” can hope to capture. Much might be said about this, and a full treatment of the question would take us far afield. But it is sufficient for present purposes to note that, whatever the details of a full account should turn out to be, this much, at least, would have to be included among them – that a performance of a work cannot be “faithful” to it unless it evinces an understanding of it. And if a



against musical ontology

performance’s faithfulness is, minimally, a matter of the understanding it shows, then a performance is, in that much, to be valued in proportion to the richness, depth, insight, subtlety and so on of the understanding it evinces.12 But if this is right, evidently enough, much of the “content” of a given work is only revealed in the understandings that faithful performances of it evince. And that means that any attempt to specify that content – the content to which a good performance is faithful – in advance of evaluative judgments about particular performances of it, or independently of such judgments, must be futile and self‐defeating. The fact is, in other words, that since faithfulness is a matter of understanding, and since understanding is not a matter of bringing off a “match” of any sort with some independently specifiable content, nothing at all can be said in advance of any particular faithful performance about what “faithfulness,” in its case, will amount to. This is why some performances, in evincing an original or especially penetrating understanding of the work performed, and so in bringing to light aspects or dimensions of its content that other performances have not, are described as “revelatory.” But why is this bad news for the ontologist? Or why, rather, is it more bad news, rather than just a gloating restatement of the bad news already announced, namely, that the specification of identity conditions cannot hope to do any evaluative work? Here is why. Until just now, it might have seemed that the evaluative impotence of identity conditions at least left in place their claim to be identity conditions – that is, their claim to be the arbiters of legitimacy, even if not of quality. It might have seemed, in other words, as if aspects of a work’s content not revealed exclusively through performance might be specified independently, and taken, if a performance “matches” them, to show that the performance is indeed a legitimate performance of the work. But even this toehold is now denied. For if, as I have argued, a revelatory performance is one that is faithful to a work in a way that necessarily defies specification in advance, then, first, as a performance that is faithful to the work it is, trivially, of it, and second, amongst its unforeseeable qualities may well be the disregard of any, or even of all, of the independently specifiable bits of “content” that it was supposed, as a legitimate performance, to have to match. This is not to deny that such bits of “content” may indeed be specifiable independently. But it is to deny that their being so specifiable accords to them any special authority. It is just as futile, in other words, to hope that one’s identity conditions will trump the evidence of one’s ears in this case, where the performance strikes one as masterly, as tremendous, as it is in the case

103

discussed earlier, where the performance strikes one as an infuriating, depressing travesty. And this means that one’s “identity conditions,” whichever ones they happen to be, cannot be identity conditions, or even conditions. If an exemplary performance can violate them, then they are, at best, an expression of some more or less reasonable set of expectations, or, as one might rather more grandly put it, a set of defeasible criteria. And this means, in turn, that they lose any claim to report the fundamental ontology of the matter: if some – or perhaps even all – of a work’s allegedly identity‐confirming properties can fail to be reflected in a faithful performance of it, then those properties can have nothing to do with the work’s being what, essentially, it is. Rather, if what I have been arguing is right, one finds out what a work is, what properties it has, by experiencing performances of it, or by giving performances of it, and that is a process of discovery that may well have no determinate end. The upshot of all this is not only that neither of the possible ways of cashing out the second response to the charge of “scandal” is able to deliver, but that retreat to the first response  –  to the claim that what ontologically inclined philosophers of music are doing is musical metaphysics, not musical aesthetics – is doomed to failure too. For what the foregoing shows is that ontological questions about pieces of music are only perspicuously to be framed, if they are to be framed at all, against an aesthetic backdrop of already‐ answered questions about the value of performances of them. Strawson is again helpful here: To use a fashionable phrase, the criterion of identity of a work of art is the totality of features which are relevant to its aesthetic appraisal…. Perhaps I could also express the point in this way: the only method of describing a work of art which is both… adequate for the purpose of aesthetic appraisal, and does not use evaluative language, is to say, ‘It goes like this’ – and then reproduce it. And, of course, this is not a method of describing at all (op. cit., p. 185).13

If the second response to the charge of “scandal” put the ontological cart before the evaluative horse, and so forfeited its claim to be engaged in either ontology or aesthetics, although not its claim to be engaged with music, albeit in a thin way, this first response, in attempting to put the horse and cart in different fields entirely, forfeits any claim to be concerned with music, with a “method of describing” music, at all. Its proponents must cut themselves off from what Davies calls a work’s “content” – that is, from what can be understood and appreciated in it. And in doing so they confine themselves, at best, to the metaphysics of

104

aaron ridley

sounds that might, conceivably, be heard as music, or maybe to the metaphysics of scores. Musical metaphysics is predicated on musical aesthetics; and in musical aesthetics, ontology comes last (at the end of time, perhaps).

III One might wonder, if the argument of the previous section is right, why it is that the temptation to begin thinking about performance in terms of ontology and identity conditions should have been felt at all. In part, and I am convinced of this, it does have to do with the lure of metaphysics: metaphysics is grown‐up philosophy, and insecure aestheticians may reach for it too readily. It also has to do with the unwittingly baneful influence of Goodman, as I suggested at the beginning of this essay. But I can also think of two less ad‐ or ab‐hominem sorts of reason, both superficially understandable, but neither, when one looks at all closely, compelling. The first concerns “hard cases”; the second, first performances. I will take them in turn. “Hard cases”: the examples I have used have tended to suppress the possibility of genuine borderline cases, cases in which it really is not clear whether a given “performance” is of a given work. It is all very well to concentrate on obvious outbreaks of musical murder, but sometimes, at least, things are not so straightforward. What is one to make, for instance, of a case in which one is genuinely puzzled by what one is hearing – not enraged or bilious, just puzzled? It is possible, one thinks, that one is hearing a performance of such‐and‐such. But, then again, one might not be. Surely, it will be said, this sort of case, a genuine borderline case, shows precisely why one needs to have some ontological views to draw upon, some identity conditions that will, at least in principle, allow one either to settle the question, or else to characterize the borderline on which the case sits in a perspicuous way. And surely neither result would be aesthetically worthless: ontology cannot, then, be written off so quickly. I am perfectly happy to agree, at least for the sake of argument, that there might be genuine borderline cases. I am ready to agree that it might not always be clear, or may be simply undecidable, whether a particular “performance” is, indeed, of a given work. But this fact, if it is one, does not have any of the redemptive consequences for ontology envisaged. For one thing, and in light of the argument of the previous section, it would be a bit odd, to put it no higher, if ontology suddenly turned out to be splendid for borderline cases while being irrelevant or misleading for

uncontentious ones. But be that as it may, the problem with the present suggestion is that it simply ignores the argument I have offered. Suppose that one were able to decide, in some particular borderline case, that a performance was, in fact, of such‐and‐such a work. This would indeed be because one had decided that it satisfied certain conditions – conditions relating, presumably, to the “content” of the work in question. But since, as I have argued, and I think shown, that content is crucially revealed by the understandings evinced in faithful performances of the work, one’s capacity to decide in favor, as it were, of some particular putative performance represents a triumph, not for ontology, but for aesthetics. And the same goes if what is sought is a perspicuous characterization of the borderline that a “performance” occupies: that is going to be yielded, not, say, by a list of the identity conditions that the “performance” matches or fails to match, but by evaluatively driven critical inquiry. As far as I can see, then, “hard cases” offer no comfort at all to those who would like to see ontology done first. At most, it needs to be done last. First performances: according to me, much of the “content” of a given work is revealed through the understandings evinced of it by faithful performances. But what if there have been no performances, faithful or otherwise? What if this performance is the first one? There is now no “aesthetic backdrop,” as I put it a moment ago, against which ontological questions might be raised. And yet there is surely an answer to the question: Is this (first) performance a performance of the work it purports to be of? And if there is an answer to that question, that surely shows that one can determine the identity of a work on independently specifiable grounds – on the grounds, that is, of precisely the sorts of ontological consideration that I have been so busy maligning. So again, the suggestion goes, ontology is not to be seen off so quickly. My response to this point begins with what is, essentially, an expression of puzzlement. What, exactly, is the difficulty supposed to be here? One shows up at a concert hall expecting to hear, as the program promises, the first performance of such‐and‐such by so‐and‐so; the orchestra have their scores open in front of them, the conductor does the usual conductorly things, and they play a piece one has never heard before. How, precisely, is the question whether this (first) performance is a performance of the work it purports to be of meant to arise? To what dark suspicions would one have to be prey for this to strike one, genuinely, as a pressing or pointful question? What I am trying to suggest, in other words, is that if all of the surrounding circumstances are exactly as one would expect, if everything is



against musical ontology

outwardly consistent with the performance’s being what it purports to be, namely, the first performance of such‐and‐ such, the question whether it is of what it purports to be of simply does not come up. The context does not invite the question, and the motivation for asking it is, as things stand, wholly inscrutable. Skepticism for its own sake is as difficult to make sense of here, in other words, as it is anywhere else. For the question to be worth asking it must, first, be raised in acknowledgment of ordinary contextual considerations and, second, be motivated by factors that might, given those considerations, be intelligible grounds for doubt. So there really do have to be some dark suspicions in play – some genuine reasons to wonder whether what one is hearing really is what it was said to be – before the question, “Is the performance of the work?” is so much as askable. Let us imagine some cases. First, imagine that one is familiar with all of the composer’s previous works, and that what one is hearing sounds as if it could not possibly have come from the same hand. This might well be perplexing, and it might well make one wonder what was going on. Possible explanations? Perhaps one is not as good as one had thought at making judgments about composers’ styles. Or perhaps the composer’s style has changed, or he has simply written something uncharacteristically good or bad. Or perhaps the program has been altered at the last minute. Or perhaps the performance is a hoax, and not of a work by the advertised composer at all. The question we are concerned with can be answered or dealt with in each case. In the first two, and in the absence of further complicating details, the answer is yes, the performance is indeed of such‐and‐such. In the third case, where the program has been changed, the state of affairs originally purported has ceased, the relevant way, to be purported, and so the question does not arise. And in the final case, the answer is no, the performance is a hoax – it is not what it presents itself as being. How have these answers been arrived at? Certainly not through the specification of identity conditions. All that is required for any of them is the wholly uncontentious observation that composers standardly compose their own works – an observation that is consistent with any minimally plausible ontology, and so presupposes none. Second, imagine that one has had a sneak preview of the score, and that what the orchestra is playing seems to have nothing whatever to do with it. Again, one would wonder what was going on. And, again, there might be several explanations. It might be that one is very bad at reading scores. Or it might be that the score one saw was a forgery. Or it might be that the program has been changed at the last minute. Or

105

it might be that the performance is a hoax. The answers to the question whether the performance is of what it purports to be of are exactly the same as in the previous case: yes, yes, not an issue, and no. And this time, all that is required to arrive at them is the humdrum truth that works are standardly notated in and performed from scores, again not a thought that either presupposes or suggests any particular set of ontological commitments rather than any other. Finally, and even less interestingly, suppose that one has seen the score in advance, and that one creates for oneself a vividly imagined performance in one’s head. And then suppose that the performance one hears in the concert hall is not like that at all. But this is hopeless. First, if the performance one hears is not like that at all, this example collapses into the previous one. But if it is enough like it to make one concerned about the divergence, then a dilemma presents itself. If, on the one hand, it is granted that the imaginary rendition is equivalent to a genuine performance, then (a) the performance one hears in the concert hall is not, in the relevant sense, a first performance, and so cannot prompt the question we are supposed to be worrying about; and (b) the most, anyway, that the divergence at issue could indicate is that the actual performance is either significantly better or significantly worse than the imagined one (and not, that is, that the actual performance was not of what it was meant to be of). Whereas if, on the other hand, the imaginary rendition is denied the status of “performance,” its divergence from the actual performance can indicate only (a) that actual performances are different from nonperformances; and (b) that that difference is sufficient to entail the collapse of this example into the previous one, as before. Neither alternative can give the example even the appearance of bite. However one looks at it, it seems to me, first performances pose no threat whatever to my contention that musical ontology is, at best, superfluous if what one is interested in is musical aesthetics. But have I not, it might be asked, committed myself to a whole set of ontological claims throughout the course of my own argument? Have I not claimed, in effect, that works are identical to faithful performances of them? And have I not nailed my colors to a firmly realist view of the properties of works that faithful performances reveal? And have I not just helped myself  –  a bit conveniently  –  to thoughts about the relations between composers and works and between works and scores that look, for all the world, like covertly specified identity conditions? Have I not, in short, shot myself repeatedly in the foot? It will come as no surprise to learn that I think the answer is no, I have not. I have nowhere, in effect or

106

aaron ridley

­ therwise, claimed or assumed that works are identical to o (some? all?) faithful performances of them, and nor can I see why anyone would want to. All that I have argued is that performances can show us things about works; and that requires nothing more than the thought that (some) performances are interpretations of works – not, I surmise, a proposition likely to provoke a storm of protest, and certainly not one that involves or presupposes (or should prompt) the slightest flicker of ontological reflection. With respect to the second charge, it is true that I have cast a number of my sentences in undeniably realist terms. It is true, that is, that I have spoken of aspects of a work’s content as being revealed by faithful performances of it, rather than as being brought into existence by them, say. But I have put it like that largely as a matter of convenience, and also out of deference to the convention that, unless there is some good reason not to, one should assume that what one is talking about is, in a straightforward and commonsensical way, real. I do not insist upon this, though. If someone feels more comfortable with the thought that performances

“create” content rather than “discover” it, that is fine by me: nothing in what I have said or want to say depends on the distinction. And finally, to the suggestion that I have really just helped myself covertly to a raft of identity conditions, I simply deny that I have. At most, I have helped myself to some perfectly neutral, pre‐theoretical thoughts. And these are, first, of a sort that no one, whatever their ontological views, could possibly object to, and second, of a sort which, if they were (by someone else) misconstrued as quasi‐theoretical items, would not thereby become identity conditions, but – as I have already argued – expressions, merely, of some more or less reasonable sets of expectations. My position, it seems to me, is steadfastly devoid of ontological commitments; and it is, for the reasons I have given, in much the better shape for that. Musical ontology may, at first blush, appear to be an entirely harmless and even a proper philosophical pursuit. But if it is the philosophy of music that one is interested in doing, then ontology really is an idle distraction, or worse, and I propose that we should have nothing further to do with it.

Notes 1 Indianapolis: Hackett, 1976. Goodman was not of course the first to discuss musical ontology or the relation between works and performances, but his conclusions certainly heated the discussion up. 2 “Orchestrating Platonism,” in his The Fine Art of Repetition (New York: Cambridge University Press, 1993), p. 75. For a harder‐nosed Platonism, see Nicholas Wolterstorff, Works and Worlds of Art (New York: Oxford University Press, 1980). 3 Kivy, “Platonism in Music: Another Kind of Defense,” in The Fine Art of Repetition, p. 63. 4 Kivy, “Platonism in Music: A Kind of Defense,” in The Fine Art of Repetition, p. 56. Kivy’s sense of ‘uniquely’ has to be a “loose” one, otherwise his characterization of the relation between score and performance would be identical to Goodman’s characterization of the relation between a score, understood as “a character in a notational system,” and its “compliance‐class,” a class which includes performances – see Goodman, pp. 177– 78. Kivy needs to keep his distance from Goodman, not only because he rejects Goodman’s nominalism, but because he wants his own conclusions to be intuitively plausible. 5 “What a Musical Work Is,” in his Music, Art, and Metaphysics (Ithaca: Cornell, 1990), pp. 63–88. [Original article reproduced as Chapter 7 in this Anthology.] 6 Or, more strictly, had already proposed a different kind. Levinson’s work on the topic pre‐dates Kivy’s by several years. 7 New York: Oxford University Press, 2001. 8 Authenticities: Philosophical Reflections on Musical Performance (Ithaca: Cornell University Press, 1995).

9 There are exceptions, of course—the most honorable recent one, in my view, being the raising of just that question by Kivy in his generally splendid Authenticities, see pp. 155–61. 10 “Aesthetic Appraisal and Works of Art,” in his Freedom and Resentment (London: Methuen, 1974). 11 In fact, what often happens, I think, is that evaluative considerations are simply mistaken for identity conditions. Consider a dispute about the merits of Busoni’s piano transcription of Bach’s great Chaconne for solo violin.The reasons that someone might offer for disparaging Busoni’s transcription include, in no particular order, the following (with the “identity conditions” they might wrongly have been stated as in brackets afterwards): The Busoni wholly fails to capture the violinistic quality of the original (being for violin is an essential property of the Chaconne). Bach cannot possibly have intended his work to have anything like the big, late‐Romantic feel that Busoni has imparted to it (Bach’s intentions are partly constitutive of what it is for the Chaconne to be the Chaconne). No audience made up of Bach’s contemporaries would recognize in Busoni’s transcription a faithful rendition of Bach’s original (historically appropriate listening habits contribute to the determination of  the Chaconne s essential properties). And so on. From the other side, by contrast, those who are impressed by Busoni’s transcription might offer as reasons in its favor, again in no particular order, these (and again with the ­“identity conditions” they might have been misstated as in brackets): Busoni illuminates the progression of Bach’s musical argument in a particularly telling way (the Chaconne is, essentially, an intellectually



against musical ontology articulated structure of sound). Busoni highlights facets of Bach’s work that could not have been appreciated in the original version (the essential properties of the Chaconne are not identical to or exhausted by the properties revealed in “historically appropriate” renditions of it). Busoni creates Bach’s work anew (a rendition of a work can, in effect, determine new conditions of what it is to be that work). And so on. In each case, and on either side of the disagreement, the unbracketed remarks represent, with varying degrees of persuasiveness, the kinds of

107

r­easons that might feature in a genuinely pointful critical debate; whereas the remarks in brackets are decent examples of the sorts of thing that reliably stifle any such debate at source. 12 The notion of “understanding” cannot, of course, be reduced to the notion of “matching,” for standard Wittgensteinian reasons – see Philosophical Investigations, §§527–37. 13 I take it that Strawson’s “totality” of relevant features is exactly what the process of discovery I refer to would, at some notional end point, have brought to light.

10

The Ontology of Art and Knowledge in Aesthetics Amie L.Thomasson

The ontology of art has provided one of the richest areas of discussion in recent aesthetics, yielding a variety of carefully articulated and well‐argued positions about the ontological status of works of music, literature, even painting and sculpture. In some ways the variety of positions seems to be an embarrassment of riches, for it is not clear how we are to decide among these apparently mutually incompatible and often surprising views about whether works of art of some or all kinds are physical objects, abstract objects, action types, and so on. In other respects, work on the ontology of art seems to be embarrassingly impoverished, for there seems to be no natural and nonarbitrary way of answering other questions in the ontology of art, such as how many mistakes a performer may make and yet still perform a work of music; how much restoration a work of painting or sculpture may survive; or even what the exact criteria are for creating a work of literature. If we hope to resolve these questions and decide among the competing theories, we must step back from the particular debates about the status or identity conditions of a given kind of work to address issues in meta‐ontology, particularly as applied to issues in the ontology of art.What are we doing when we argue about the ontological status of works of art? What are the proper methods and criteria of success to be used in answering and evaluating answers to these questions? What kinds of answers can we legitimately expect and demand in questions about the ontology of art?

An influential paradigm of what it is to acquire knowledge has come from a certain (perhaps naïve) view of how the natural sciences and other empirical investigations work. According to this paradigm  –  call it the discovery view – the world contains a broad range of fully determinate, mind‐independent facts about which everyone may be ignorant or in error, but (some of) which the scientist seeks to discover by substantive empirical investigations. Thus, one acquires knowledge about, say, the biological nature of whales by ostensively applying the term ‘whale’ to this kind of thing and undertaking substantive empirical investigations about them (their internal structure, genetics, etc.) in order to discover the real truth about whales’ biological nature, which may overturn our common‐sense views about them. Moreover, on this view, there is a complete range of mind‐independent facts to be discovered, so that, for any empirical proposition P we could formulate about whales, either P or not‐P is the case; the only challenge lies in discovering which. So, similarly, knowledge claims in the ontology of art are often presented as discoveries of fully determinate, mind‐ independent facts about the ontological status of works of art of various kinds, about which everyone may be ignorant or in error – so that we should not be surprised if the “right” view turns out to be that works of art are discovered rather than created, action‐types rather than objects, and so on, and so that we may rightly demand that theories

Amie L.Thomasson, “The Ontology of Art and Knowledge in Aesthetics,” Journal of Aesthetics and Art Criticism 63 (2005), pp. 221–9. Reproduced by permission of Wiley.

Aesthetics and the Philosophy of Art – The Analytic Tradition: An Anthology, Second Edition. Edited by Peter Lamarque and Stein Haugom Olsen. © 2019 John Wiley & Sons, Inc. Published 2019 by John Wiley & Sons, Inc.



the ontology of art and knowledge in aesthetics

provide precise answers to any questions we care to invent about, for example, the creation, survival, and identity of works of art. I am not concerned here to either defend or attack this discovery view of knowledge, but rather to argue that  –  whatever its merits as an understanding of scientific or other empirical investigations – thinking of the process of acquiring knowledge about the ontology of art on that model (as, I think, many have been inclined to do) leads us badly astray. This paper, then, is an exercise in meta‐ontology, as applied to issues in the ontology of art, designed to examine what it is we are doing when we formulate theories about the ontology of art, how we can adjudicate among competing theories, and what the limits of knowledge are in this area. In fact, the points to be made seem to have quite general application to much of what goes under the heading of “ontology” both inside and outside of aesthetics. Thus here as elsewhere, I think that careful study of issues in aesthetics can lead to progress in other areas of philosophy.1 Nonetheless, the ontology of art provides a particularly useful case study, since (I will argue) issues come to prominence here that might otherwise be overlooked.The consequences of rejecting the discovery model of knowledge for ontological issues are also particularly important for aesthetics since this is an area in which debates about the ontological status of the objects concerned (works of art) play a prominent role, with most beginning from the presumption that there are such objects, the only issue being what sorts of things they are. As a result, the field is flooded with all manner of diverse, and often revisionary, proposals about the ontological status of works of art, as well as attempts to answer all sorts of question about works’ identity, creation, and survival. A proper understanding of what we are doing in the ontology of art, I will argue, can lead us to reevaluate this whole ontological side of discourse in aesthetics.

I. The Ontology of Art The discovery view of knowledge is bolstered by causal theories of reference, which ensure that we have some independent way of picking out kinds like whales or gold (e.g., by ostending a sample), so that we may then go on to investigate their true nature, which may turn out to be at odds with our concepts or initial presuppositions. But the qua problem has made it clear that this sort of view has crucial limitations.2 For any sample will include entities of  a great many kinds  –  physical, chemical, biological, ­functional, and so forth – so that without some disambigu-

109

ating concept specifying the sort of kind to be picked out, we cannot unambiguously ground reference to any kind. In fact, although causal theories of reference have most often been discussed with natural kinds in view, the full virulence of the qua problem becomes evident only when we recognize that our general terms may name not only diverse natural scientific kinds, but also social and cultural kinds, so that in a single situation, would‐be grounders of a general term may be confronted with members of a great many kinds: physical and chemical kinds involving the kind of canvas painted on or the chemical structure of the paints used; observable kinds like red thing or square thing, which all may exemplify; institutional kinds like object worth more than $1,000 or entry in the First Annual Ocean Bank Young Artist’s Competition, and so on. This places at least some constraints on a causal theory of reference and on the possibilities for discovery: at least some frame‐level disambiguating concept must be involved to stipulate the kind of kind being named, by specifying the sorts of feature (physical, biological, chemical, etc.) that are to be relevant to unifying the kind. As a result, to disambiguate, grounders of the reference of a term such as ‘whale’ must, for example, intend it as a biological kind (species) term, and so cannot discover themselves to have been wrong about this (though they can discover that there is no biological kind before them and so discover that their term does not refer).3 Nonetheless, they can still discover much about the precise biological nature of their kind, for example, that it is a kind of mammal, not fish, that a certain DNA structure is essential to it, that it evolved in certain ways, and so on. As a result, the constraints at most appear around the edges, and so it seems that the discovery theory can, by and large, be retained for cases of empirical knowledge about the kinds referred to. But can the discovery theory nonetheless largely be retained for issues of ontology, allowing, for example, for surprising discoveries about the ontological status of works of art of various sorts (say, that works of art are discovered not created, that sculptures are action‐types, that works of music cannot be transcribed, nor paintings restored, etc.), paralleling surprising empirical discoveries about whales? I will argue that it cannot – indeed that a careful study of the ontology of art suggests that the discovery theory is nowhere plausible as applied to ontological (as opposed to empirical) issues. I argued elsewhere that the disambiguating conception of the sort of kind involved in grounding the reference of art‐ kind terms must include a nascent concept of the ontological status of the kinds of work involved.4 As a result, at least

110

amie l. thomasson

a background concept of the ontology of the work of art is needed to establish the reference of terms like ‘painting’ or ‘symphony.’ Such concepts determine the ontological kind, if any, picked out by the term, and so the ontology of the work of art must be something we learn about through conceptual analysis of the associated concepts of people who competently ground (and reground) the reference of terms like ‘symphony’ and ‘painting,’ not something we can seek to discover through investigations into mind‐independent reality. Moreover, competent grounders cannot (as a whole) be massively ignorant of or in error about the ontological nature of the art‐kind they refer to since their concepts are determinative of this. Thus  would‐be grounders have some forms of epistemic privilege regarding the ontological status of the art‐kinds (if any) they refer to that everyone lacks regarding, say, the biological nature of whales or the chemical nature of soy protein. There has, however, been a lot of discussion of similar issues for the case of artifacts generally, some of which might form the basis for an interesting reply to this line of thought. In Section I.i I will show how one might develop such an objection, and then in Section I.ii I will attempt to demonstrate why this sort of objection ultimately fails to undermine the idea that facts about the ontology of paintings, symphonies, novels, and so forth (and indeed facts about ontology generally) are determined by the beliefs and practices of those who ground the use of the term. In Sections II and III I will then go on to consider some consequences of this conclusion for methods of knowledge acquisition and the limits of possible knowledge about the ontology of art.

I.i.  Direct Reference Theories and Art‐Kind Terms Hilary Kornblith has argued that artifactual‐kind terms function in ways parallel to natural‐kind terms, since, for example, Martian anthropologists could simply point at a sample of apparent earth artifacts and coin a term ‘glug’ to refer to that kind of thing, thereby establishing reference to an artifactual kind even though the Martians who coin the term may have no idea what the nature of that artifactual kind is.5 Nonetheless, to overcome the qua problem and establish ‘glug’ as a would‐be artifactual kind term (rather than as a would‐be physical‐, chemical‐, or biological‐kind term), the Martians must at least associate the term with some sort of kind (e.g., artifactual kind), with a broad concept of what sorts of features are relevant to membership in a kind of that sort as opposed to those features relevant to

membership in, for example, a given chemical kind.6 Kornblith seems to accept this, and takes the relevant sorts of features for artifactual kinds to be functional. (Others have argued against the view that functional features are what are relevant to artifactual‐kind membership.7) One might try to argue, analogously, that those who would ground the reference of an art‐kind term may do so merely by pointing to a sample (say, of objects from a very different culture) and coining a term for works of art of that kind. But again, to do so, such would‐be grounders must at least have some base‐level conception of what sort of kind an art‐kind is, and what sorts of features are relevant to belonging to the same art‐kind (as opposed to those relevant to belonging to the same chemical‐, biological‐, or physical‐kind). What general sort(s) of feature(s) could be appealed to in order to unify objects into art‐kinds such as paintings, symphonies, or novels? Clearly no common chemical, physical, or biological features are necessary or sufficient for membership in such kinds. Nor is common function a plausible way of delineating artkinds, since, for example, works of music may be created to serve (and may actually serve) many different functions, or none. Resemblance is also not a plausible demarcating feature, since everything resembles everything in so many ways that this is useless as a unifying criterion without specifying the sorts of resemblance that are relevant to belonging to the same art‐kind. But the sort of resemblance that is relevant seems to vary by art‐kind (e.g., visual resemblance being relevant to paintings but not to works of literature), and so seems to presuppose rather than aid in demarcating these basic art‐kinds. One common feature that seems to distinguish members of any art‐kind, or any artifactual‐kind, from mere members of natural kinds is being the product of human intentions. The products of human activities may of course be classified and referred to in any number of ways (economically, causally, physically, etc.). But central ways of classifying them as artifacts of particular kinds involve classifying them according to the intentions involved in their creation (rather than according to their physical or chemical structure, etc.).Thus Paul Bloom has suggested that the core of artifact concepts is the idea that, in each case, the things were successfully created with the intention that they belong to that very (artifactual) kind; for example, something is a chair only if it is the product of a successful intention to make a chair (as opposed to being a fortuitously shaped tree).8 Similarly, one promising proposal is that at least one sort of feature necessary to unify objects into such art‐kinds such as paintings, symphonies, and novels is that they be the



the ontology of art and knowledge in aesthetics

products of the same sorts of intentions (namely, intentions to make something of that kind).9 It does seem apt to say that something is a painting only if it is the product of an intention to make a painting (as opposed to the mere unintended byproduct on a drop cloth of a process of painting walls), and that something is a work of music only if it is the product of an intention to make a work of music (thus distinguishing the normal sound of people adjusting radios from that comprising a performance of Cage’s 1956 “Radio Music”). On this view, an artist’s intentions would determine what sort of work (if any) he or she creates  –  for example, whether a dance or a painting is being created by certain movements that involve the spread of pigments, thereby also determining which of its features are essential (e.g., having roughly this arrangement of pigments), and which accidental (e.g., having been created by way of roughly these gestures). (This, of course, does not entail that an artist’s intentions to create a work of a certain kind always succeed, or that an artist’s intentions determine the proper interpretation or value of his or her work, or the style or value kinds to which it belongs.) For this proposal to be plausible, we would have to allow that intending to make something of the relevant kind does not require that all makers have use of the same word, or that they must be able to refer back transparently to extant members of the kind. We would also need to allow that intentional creation may involve only intentional selection and appropriation rather than constructing or reshaping an object. But it seems plausible enough that intending to create a chair by appropriating and not changing anything about this stump, or intending to create a sculpture by appropriating and not altering this piece of plumbing, is an act of intentional creation just as, on Alvin Goldman’s view, negations of basic act‐types (e.g., intentionally not raising one’s hand during a vote) are also act‐types.10 So understood, being created with the same sort of intention does seem plausible as at least a necessary condition for entities to belong in the same art‐ kind. But can one then establish the reference of an art‐kind term semi‐ostensively as “whatever was created with the same sort of intention as this was” and go on to investigate the ontological status of members of the kind (about which all speakers may be ignorant or in error)?

I.ii. The Need for an Ontological Conception No – even if one accepts all of the above, grounders still cannot establish the reference of an art‐kind term with no  further concept than that (or even than that plus

111

r­esemblance among kind members), and go on to investigate the kind’s ontological status. For “being created with the same sort of intention” is not specific enough to disambiguate among the many ontologically different sorts of thing that may be present before grounders, and so the qua problem arises again and prevents would‐be grounders from establishing a univocal reference for their term if that is all they have in mind. To see this, consider the analogous case of action‐ kinds – for, as has often been noted, there are many similarities between actions and artifacts, as both are (in a sense) products of human intentions.11 Just as actions performed may be layered so that one may intend to raise one’s hand, and also intend to thereby cast a vote for a particular bill, so may artifacts (among them works of art) be layered so that one may intend to create a particular musical performance and thereby to also create a new (repeatable) musical work, or intend to perform an act of story telling and also intend thereby to create a certain literary work.12 Similarly, one may intend to create three paintings and thereby to create an altarpiece, or intend to write twenty‐seven installments for a magazine and thereby to create a novel. Thus it is not necessarily or even normally the case that, in a given context, a would‐be grounder of an action‐kind term has members of exactly one action‐kind before him or her, or that a would‐be grounder of an art‐kind term has members of exactly one art‐kind before him or her. On the contrary, it is entirely typical that an artist intends to create a song by playing a tune or a triptych by painting three panels. Suppose a would‐be grounder of an action‐kind term is standing before a bunch of people in a field and points to one of them saying “that kind of action I will call zaybing.” Such an attempted grounding is hopeless since it is entirely ambiguous which of the many types of (intended) actions the person is performing (say, moving his or her right arm, touching another person, tagging someone, making an “out,” helping his or her team to win the game, etc.) is being named; it is even ambiguous when the supposed sample action of that type begins and ends (taken over a longer spread of time, relevant actions could include playing baseball, passing the time on a Sunday afternoon, improving one’s athletic skills, etc.) Similarly, without a further conception than “work of art like this,” it will remain ambiguous which of the many intended art‐kinds instanced before a would‐be grounder (each of which may have strong internal resemblances among members) the term refers to. So it seems that in order to unambiguously ground the reference of a general term to name a kind of work of art, the grounder must not only have the idea that the reference

112

amie l. thomasson

of his or her term will be an art‐kind, but must also have a background conception of what ontological sort of art‐kind he or she means the term to refer to, establishing existence conditions and identity conditions for works of that kind. Such an ontological conception then disambiguates potential reference by determining the ontological kind referred to by the art‐kind term (if it succeeds in referring at all), establishing, for example, whether the relevant art‐kind is to be a kind of activity or object, a concrete individual or the abstract pattern exemplified by these concrete individuals, and establishing where a work’s spatial and/or temporal boundaries lie and the conditions under which one and the same work survives. Having such a background conception of course does not require that speakers (or makers) have detailed and sophisticated ontological concepts in mind when they attempt to refer to paintings, songs, or stories. Normally, such ontological disambiguation is achieved not by a philosophical and explicit decision on the part of grounders about what their term will refer to, but rather by appeal to background practices already in place that co‐evolve with the use of the art‐kind term. Thus, whether we are intending to refer to a concrete or abstract individual is linked to factors like whether we consider it essential to go to a particular place at a given time to see the work of art (as, e.g., we do for paintings but not for works of literature).Whether we intend to refer to an activity or an object is similarly linked to whether it is appropriate to ask, for example, when the work occurs, begins, or ends (as is suitable for plays and musical performances, but not sculptures). What level of work is intended is expressed by practices regarding, for example, whether it is possible to physically move, buy, or sell the work, or only to buy or sell, for example, performance or reproduction rights, and whether we would count the work as destroyed if some particular physical object were destroyed. The extent and limits of the object are similarly reflected in practices such as when we arrive and when we clap, and what is and is not properly remarked upon in reviewing the work (e.g., comments on the frame or dirty back of the canvas are treated as inappropriate to reviewing a work of traditional painting, and comments on the drinks available at intermission are treated as inappropriate to reviewing a play). Such background practices thus embody a tacit ontological conception of what sorts of things works of that kind are, which (understood by grounders) disambiguates the reference of the relevant kind term, and determines ontological features of the members of the kind (if any) picked out by the term. These features may then be drawn

out in more formal philosophical theories of the ontology of the work of art – explicitly describing, for example, their relevant formal category (object, event, etc.), existence conditions and relations to human intentions and physical objects and processes, and their boundaries and individuation conditions.13 The need for an ontological conception to disambiguate the potential reference of our terms seems to be quite general, for whenever a grounder attempts to establish the reference of a term, there will in fact be a great many ontological kinds of thing available for reference, from biological individuals, to collections of particles, to abstract types, created types, qualities, events, and so on. For reference to be established, we must disambiguate among them. If so, then it is not just for art‐kind terms, but for all terms, that grounders’ concepts establish the ontological status of the referent (if any) of their new term. Nonetheless, this general point has often been overlooked. This can be attributed at least in part, I think, to the fact that discussions of causal theories of reference have largely been undertaken with rather precise sorts of natural‐kind terms in mind, for example, species terms, chemical‐kind terms, and the like. These terms are each specific enough that they are already uniquely associated with background ontological assumptions that provide the needed disambiguation among the diverse ontological kinds present in a grounding situation, and so the specific role played by ontological conceptions is easily missed. But if we consider instead art‐kind terms, it becomes obvious that to disambiguate reference, it is essential that we not just associate our term with some or other sort of kind supposed to be named (e.g., art‐kind), but with an ontological conception that provides at least some of the necessary disambiguation. Studying art‐kind terms is thus particularly helpful, since it helps reveal the role of specifically ontological conceptions in establishing reference where this might otherwise be overlooked.

II.  Epistemic and Methodological Consequences If the above is something like the right account of what determines the ontological status of works of art of various kinds, then study of the ontology of art (indeed, of ontology generally) cannot proceed along the discovery model with which we began. We cannot discover the ontological status of paintings, symphonies, or novels first by referring directly to this kind of thing and then



the ontology of art and knowledge in aesthetics

investigating its true ontological nature. Instead, the background ontological conception of grounders determines the ontological status of members of the art‐kind referred to by the term (if the term refers at all). The ontological status of paintings, symphonies, or works of literature is thus not something we can discover by investigations into the mind‐independent world; instead, we must follow the method of analyzing the conception embodied in the practices of those competent speakers who ground and reground reference of the term. As a result, although competent grounders of the term’s reference may not have an explicit ontological view stated in formal philosophical terms, their background ontological conception of the sort of entity they are talking about is not subject to the kinds of massive error to which everyone’s beliefs are subject, according to the discovery model. These results again seem to be quite general whenever we are talking about the ontological status, identity, and survival conditions of works of art, persons, social institutions, artifacts, and so on. If they are correct, they show that ontological knowledge is always to be acquired by way of this kind of conceptual analysis, and cannot be conceived on the model of an empirical inquiry about mind‐independent facts that may turn out to radically overturn our prior beliefs and concepts. Perhaps this is already tacitly recognized by those working in the ontology of art, who generally do attempt to evaluate theories at least in part by their consistency with ordinary beliefs and practices. Made explicit, though, this observation has some important practical consequences – especially for aesthetics. For recent aesthetics has produced something of a boom in diverse views about the ontology of art, many of which are radically revisionary views, for example, that despite popular belief, all works of art (including paintings and noncast sculptures) are action‐ types rather than individuals;14 or that literary works are eternal abstracta that can neither be created nor destroyed; or that plays do not merely present fictional individuals and events, but are themselves fictional objects;15 or that despite common practice, any transcription of a musical work must be a different work (as Jerrold Levinson somewhat tentatively suggests);16 or that no painting or sculpture can ever be repaired or restored by replacing broken or worn‐out parts, however minor.17 If the above arguments are correct, such radical views cannot be presented as discoveries about the “real truth” of the ontology of works of art that may overthrow commonsense and show that we are mistaken in treating works as individuals capable of being bought and sold,

113

things created at a certain time, or as being transcribable or restorable. For the only way to find out the truth about the ontology of the work of art is by way of conceptual analysis that teases out from our practices and things we say the tacit underlying ontological conception of those who ground the reference of the term, perhaps making it more explicit, smoothing out any apparent inconsistencies, and showing its place in an overall ontological picture. Radical solutions cannot be seen as discoveries about what the ontological standing of any art‐kind really is, but only as proposals about how we should change our practices – not because they are wrong in the sense of being based on tacit views that are inconsistent with the real facts, but only, perhaps, because the proposed change would be clearer, less prone to vagueness, and so forth. (Most often, I think, such proposals are based in following some single strand of current practice through to its extreme limit, at the expense of others.) But although such massive mistakes about the ontology of art are not possible, the limits of epistemic privilege must be carefully acknowledged: the relevant epistemic privilege is only collective, not individual, and concerns only ontological features of the kind referred to. The above alone does not entail that there is any protection from error about, for example, the causal role of the works of art in the relevant culture, any functions they can or do serve, their histories of production, aesthetic properties, and so forth.18 All grounders are assured of (collectively) is that, if there is any art‐kind referred to by the terms they attempt to ground the reference of, it has the ontological standing they commonly (if tacitly) understand and treat it as having.

III.  Ontological Shallowness and the Limits of Knowledge The determination of facts about the ontology of art by human concepts, beliefs, and practices has consequences not only for the methods of acquiring knowledge in the ontology of art, but also regarding the nature of the facts involved, and the limits of possible knowledge in this area. The discovery view has it that the world is fully determinate, so that for any proposition P, either P or not‐P is the case, with one being made determinately true and the other determinately false by independent facts of the world so that there is, at least in principle, the possibility of discovering the truth or falsehood of any scientific claim.

114

amie l. thomasson

But this does not seem to be the case in matters to do with the ontology of art, since many awkward questions arise for which it is hard to give any precise and determinate answer without it seeming arbitrary and inappropriate. For example, what percentage of the paint in a painting may be replaced in restoration while preserving the same painting? What sorts of editing or translation can and cannot a T.S. Eliot poem survive? Is a poem created if the poet thinks it up to himself or herself and is promptly run over by a bus? How many mistakes, and of what sort, may be made in a performance if it is to count as a performance of the relevant musical work? We can now see why there seems to be no good way to answer many such questions and why we will err if we try to treat the ontology of art on the discovery model. Even if the mind‐independent world is fully determinate (so that, for any proposition P involving mind‐independent facts, either P or not‐P is the case), it is certainly not the case that human beliefs and conceptual systems are complete (so that, for any proposition P, either P or not‐P is believed, accepted, or treated as being the case). As a result, anywhere the criteria for applying a predicate are determined by human beliefs and practices, risks of indeterminacies and vagueness inevitably arise as a result of the intensionality of human beliefs. To borrow an example from Kit Fine, suppose we define the predicate ‘nice’ as applied to numbers as follows: n is nice if n >15, n is not nice if n  15’, ‘n is not nice if n