Philosophers on Music: Experience, Meaning, and Work

Philosophers on Music

MIND ASSOCIATION OCCASIONAL SERIES This series consists of occasional volumes of original papers on predefined themes. The Mind Association nominates an editor or editors for each collection, and may cooperate with other bodies in promoting conferences or other scholarly activities in connection with the preparation of particular volumes. Publications Officer M. A. Stewart Secretary R. D. Hopkins

recently published in the series: Desert and Justice Edited by Serena Olsaretti Leviathan after 350 years Edited by Tom Sorell and Luc Foisneau Strawson and Kant Edited by Hans-Johann Glock Identity and Modality Edited by Fraser MacBride Impressions of Hume Edited by Marina Frasca-Spada and P. J. E. Kail Ramsey’s Legacy Edited by Hallvard Lillehammer and D. H. Mellor Transcendental Arguments Problems and Prospects Edited by Robert Stern Reason and Nature Essays in the Theory of Rationality Edited by Jos´e Luis Berm´udez and Alan Millar Values and Virtues Aristotelianism in Contemporary Ethics Edited by Timothy Chappell

Philosophers on Music Experience, Meaning, and Work

Edited by

Kathleen Stock

1

1

Great Clarendon Street, Oxford ox2 6dp Oxford University Press is a department of the University of Oxford. It furthers the University’s objective of excellence in research, scholarship, and education by publishing worldwide in Oxford New York Auckland Cape Town Dar es Salaam Hong Kong Karachi Kuala Lumpur Madrid Melbourne Mexico City Nairobi New Delhi Shanghai Taipei Toronto With offices in Argentina Austria Brazil Chile Czech Republic France Greece Guatemala Hungary Italy Japan Poland Portugal Singapore South Korea Switzerland Thailand Turkey Ukraine Vietnam Oxford is a registered trade mark of Oxford University Press in the UK and in certain other countries Published in the United States by Oxford University Press Inc., New York  The Several Contributors 2007 The moral rights of the authors have been asserted Database right Oxford University Press (maker) First published 2007 All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, without the prior permission in writing of Oxford University Press, or as expressly permitted by law, or under terms agreed with the appropriate reprographics rights organization. Enquiries concerning reproduction outside the scope of the above should be sent to the Rights Department, Oxford University Press, at the address above You must not circulate this book in any other binding or cover and you must impose the same condition on any acquirer British Library Cataloguing in Publication Data Data available Library of Congress Cataloging in Publication Data Philosophers on music : experience, meaning, and work / edited by Kathleen Stock. p. cm. Includes bibliographical reference and index. ISBN-13: 978-0-19-921334-4 (alk. paper) 1. Music--Philosphy and aesthetics. I. Stock, Kathleen. ML3800.P46 2007 781’.1--dc22 2007012568 Typeset by Laserwords Private Limited, Chennai, India Printed in Great Britain on acid-free paper by Biddles Ltd., King’s Lynn, Norfolk ISBN 978–0–19–921334–4 10 9 8 7 6 5 4 3 2 1

Preface K AT H L E E N S T O C K

The idea for a volume of new work on issues in the Philosophy of Music emerged after a conference entitled ‘Aesthetics from an Analytic Point of View’, which I co-organized with Julian Dodd. This was held at the University of Manchester in May 2003, and kindly sponsored by the Mind Association, the British Society of Aesthetics, the Analysis Trust, and the University of Manchester. Three of the contributions here (those of Paul Boghossian, Julian Dodd and Michael Morris) are versions of papers delivered at that conference. The rest have been commissioned specially, with the exception of Jenefer Robinson’s paper, which is a largely revised version of a paper which first appeared in a special issue of ´ the Revue Fran¸caise d’Etudes Americaines 86 (2000): 77–89. Together these papers constitute some of the best new work in what is an exciting field of research, and one which has much to engage, not just aestheticians, musicologists and music practitioners, but metaphysicians and philosophers of language as well. I wish to thank all of the contributors for getting involved in the project, for their readiness to read and comment upon the work of fellow authors, and for producing such excellent work themselves. Particular thanks are due to Julian Dodd for motivating me to take the volume on. I would also like to thank Sandy Stewart for his helpful guidance and encouragement on editorial matters. Kathleen Stock University of Sussex

This page intentionally left blank

Contents Preface

v

Notes on Contributors

ix

List of Musical Examples

xi

Introduction Kathleen Stock

1

Part I. Musical Ontology

21

1. Sounds, Instruments, and Works of Music Julian Dodd

23

2. Doing Justice to Musical Works Michael Morris

52

3. Versions of Musical Works and Literary Translations Stephen Davies

79

Part II. Musical Expression

93

4. Expression in Music Derek Matravers

95

5. Explaining Musical Experience Paul Boghossian

117

6. Persona Sometimes Grata: On the Appreciation of Expressive Music Aaron Ridley

130

Part III. Musical Meaning

147

7. Can Music Function as a Metaphor of Emotional Life? Jenefer Robinson

149

viii

contents

8. The Structure of Irony and How it Functions in Music Eddy Zemach and Tamara Balter

178

Part IV. New Issues

207

9. Music and Electro-sonic Art Gordon Graham

209

10. Thoughts on Rhythm Roger Scruton Index

226

257

Notes on Contributors Tamara Balter is a doctoral student in Music Theory at Indiana University, Bloomington. Paul Boghossian is Professor of Philosophy at New York University. He has published many articles on a variety of topics, including colour, rule-following, eliminativism, naturalism, self-knowledge, a priori knowledge, analytic truth, realism, and the aesthetics of music. He is currently at work on a book on the notion of objectivity. Stephen Davies is Associate Professor in Philosophy at the University of Auckland. His most recent books are Themes in the Philosophy of Music (Oxford University Press, 2003) and The Philosophy of Art (Blackwell, 2005). Julian Dodd is Professor in Philosophy at the University of Manchester. He works mainly in metaphysics, the philosophy of language and the ontology of art, and his recent publications include Truthmakers: The Contemporary Debate (co-edited with Helen Beebee, Oxford University Press, 2005), and Works of Music: An Essay in Ontology (Oxford University Press, 2006). Gordon Graham is Luce Professor of Philosophy and the Arts at Princeton Theological Seminary. Art and Re-enchantment, his 2005 Stanton Lectures at the University of Cambridge, will be published by Clarendon Press (Oxford) late 2007. Derek Matravers is Senior Lecturer in Philosophy at the Open University, and a Fellow of Jesus College, Cambridge. He is currently working on a book on the nature of value (forthcoming with Acumen Press). Michael Morris is Professor in Philosophy at the University of Sussex. Apart from his work on aesthetics, his principal interests are in the philosophy of language. Currently, he is completing an introduction to philosophy of language for Cambridge University Press, as well as working on a monograph on describing the world. He is also collaborating with Julian Dodd on the Routledge Philosophy Guidebook to Wittgenstein’s Tractatus. Aaron Ridley is Professor of Philosophy at the University of Southampton. He has recently published The Philosophy of Music: Theme and Variations (Edinburgh

x

notes on contributors

University Press, 2004), and is currently completing a book, Nietzsche on Art and Literature, to be published by Routledge. Jenefer Robinson is Professor of Philosophy at the University of Cincinnati. She is the co-editor of a 2004 special issue of the Journal of Aesthetics and Art Criticism on Art, Mind, and Cognitive Science, and the author of Deeper than Reason: Emotion and its Role in Literature, Music and Art (Oxford University Press, 2005). Roger Scruton is visiting Professor in Philosophy at the University of Buckingham. He has published more than thirty books, including philosophy, political and cultural commentary, criticism, and novels. His most recent work on music is Death-Devoted Heart: Sex and the Sacred in Wagner’s Tristan and Isolde (Oxford University Press, 2003). Kathleen Stock is Lecturer in Philosophy at the University of Sussex. Her main area of research interest is the imagination. She has published several articles in aesthetics, and is currently writing a monograph on imagination and fiction. Eddy Zemach is Ahad Ha’am Professor Emeritus of Philosophy at the Hebrew University of Jerusalem. His areas of research interest include metaphysics, aesthetics, epistemology, and philosophy of language, and his most recent book in aesthetics is Real Beauty (Penn State Press, 1997).

List of Musical Examples 1. Haydn Quartet in E Major Op. 54 No. 3, opening bars (mm. 1–14).

169

2. Beethoven: String Quartet in F Major, Op. 135, i, mm. 1–10.

184

3. Beethoven: String Quartet in C-sharp Minor, Op. 131, v, mm. 30–48.

186

4. Mozart: Le nozze di Figaro, Act IV, Scene I, No. 23, Cavatina, L’ho perduta, mm. 29–36.

188

5. Mozart: Le nozze di Figaro, Act III, Scene VIII, No. 19, Recitative, mm. 24–25.

191

6. Beethoven: 33 Ver¨anderungen u¨ ber einen Walzer von A. Diabelli, Op. 120, variation 1, mm. 1–5.

197

7. Beethoven: 33 Ver¨anderungen u¨ ber einen Walzer von A. Diabelli, Op. 120, variation 13, mm. 1–16.

198

8. Haydn: String Quartet in G Major, Op. 76, No. 1, iv, mm. 179–88.

200

9. Beethoven: String Quartet in F Major, Op. 135, iv, mm. 240–58.

201

10. Diana Deutsch: pattern hearing.

228

11. Wagner: Prelude to Parsifal.

230

12. Wagner: Prelude to Parsifal.

231

13. Bart´ok: Music for Strings, Percussion and Celesta, 1st movement.

232

14. Boulez: Le marteau sans maˆıtre.

232

15. Az-Z´eloub: Le maˆıtre sans marteau.

233

16. Elgar: Violin Sonata, 1st movement.

235

17. Dvoˇra´ k: New World Symphony, 3rd movement.

236

18. Sibelius: Violin Concerto, 3rd movement.

237

19. Rhythmic augmentation.

245

20. (a) Delirio (Peruvian folk song). (b) Messiaen: Harawi no. 2 ‘Bonjour toi, colombe verte’.

245

21. Non-retrogradable rhythms.

245

xii

list of musical examples

22. Messiaen: Cant´eyodjayˆa.

246

23. Bulgarian Christmas carol.

247

24. Wagner: Tristan and Isolde, motive of ‘The Look’.

249

25. Wagner: Tristan and Isolde, elaboration of ‘The Look’ motive.

249

Introduction K AT H L E E N S T O C K

This volume presents ten contributions to the philosophy of music, nearly all wholly new, from leading thinkers in the analytic tradition.¹ Some are well known for work in aesthetics, while others are better known for work elsewhere in philosophy. Indeed, it is striking that the art form of music seems like no other to attract philosophers from diverse areas of analytic philosophy. One part of the explanation is undoubtedly that, given the love many philosophers have for music, and in particular for classical music, their attention is specially drawn to its philosophical possibilities. Perhaps equally significantly, though, analytic philosophers in general have an urge to keep things tidy, and are drawn to music because of the suspicion that it cannot be easily tidied away. That is, it seems to present some special philosophical problems not posed by other art forms. Often these overlap with or are germane to problems in more general philosophical areas. For instance: musical works are thought to present special ontological problems, and are often of interest as such to the metaphysician; the issue of what it is to experience music as expressive tends to interest those working on issues in the philosophy of emotion; the question of musical meaning tends to attract those active in the philosophy of language; and so on. At the same time, perhaps as a corrective to the tendency of certain philosophers to aim, apparently exclusively, at the illumination of problems in the philosophy of music by reference to more general philosophical views, others have stressed the need to attend to music as music, and, if not to treat it as entirely sui generis, then at least to avoid those generalizations which do not do full justice to its nature, and, ¹ An exception is the contribution of Jenefer Robinson, a much modified version of a paper which ´ first appeared in a special issue of the Revue Fran¸caise d’Etudes Americaines 86 (2000: 77–89).

2

kathleen stock

in particular, to the experience of individual works.² Of course, ideally, these two goals need not be in conflict. The best works in the philosophy of music combine, as the contributions here do, a deep understanding of music with a keen sense of the relevance of more general philosophical issues to the problems it raises.

1. Musical Ontology One of the most prominent of the problems facing philosophers of music is that of identifying what sort of thing a musical work is. On the face of it, musical works present some puzzling features which makes their ontological categorization less than straightforward. Namely, a musical work apparently can exist as a performed work, or as a score, or perhaps even (just) as an intentional object, thought of by a composer. It can be performed, or a recording played of it, on several different occasions, and even in several places simultaneously, or need never be performed at all. Similarly, the score of a work can be multiply reproduced or never produced in the first place. Meanwhile, two performances of the same work can sound very different to one another. A historically dominant approach to facts such as these is to see musical works as norm-types, repeatably instantiable as particular spatiotemporal tokens,³ thereby categorizing them in terms already familiar from other debates in philosophical ontology. Opinions differ as to what should count as the relevant tokens: they may be construed as particular dateable, locatable performances, or sound sequence-events (that is, sequences of sound, located spatiotemporally), and perhaps scores too. That a work may be performed or played on multiple occasions, but need not be, nor even ever written down, supposedly reflects the fact that a type can have many tokens existing consecutively or simultaneously, but need not have any. Meanwhile, that two performances of the same work can sound differently is supposedly explained by their being more or less properly formed tokens of the same type. ² This need is stressed explicitly by Michael Morris and Aaron Ridley in their respective contributions: Chapters 2 and 6. ³ See, for instance, Wolterstorff (1980).

introduction 3 Julian Dodd defends a version of this view, arguing that a work is a norm-type, whose tokens are sound sequence-events (2000, 2002, 2004). In ‘Sounds, Instruments and Works of Music’ he defends, as part of this view, sonicism: the claim that one work is identical to another if and only if the two are sonically indistinguishable. He claims that sonic properties are the only kind of property normative for a musical work, understood as a type, in that they alone can determine what count as properly or improperly formed tokens of it; this amounts to sonicism, in effect. In embracing sonicism, Dodd rejects ‘performance-means essentialism’—the claim that, as well as or instead of certain sonic properties, a particular instrumentation is essential to a work’s identity. The version of sonicism that Dodd favours is timbral sonicism, according to which the relevant sonic properties normative within a musical work include timbral ones. Dodd agrees with a claim made elsewhere by Jerrold Levinson that to classify timbral properties as inessential to a work’s proper formation is potentially to allow its aesthetic character to be compromised (Levinson 1980: 73–8). He denies, however, that this supports performance-means essentialism, since, he argues, the connection between timbral properties and instrumentation is contingent rather than necessary; thus, though an intended timbral aspect might be indicated by a particular instrumentation, it could in principle be produced via other means. In fact, Dodd suggests, the employment of a particular instrumentation in a work should not be understood as an end in itself for a composer, but rather as a means of specifying a particular kind of sound with particular timbral properties as normative for that work, based on what instruments are available to the composer at the time. Amongst the objections to his position that Dodd rejects is Levinson’s claim that some aesthetic and expressive properties normative for a given work depend on performance means, thereby supposedly showing that in such cases the performance means are likewise normative for the work (Levinson 1980: 76–7, 1990a: 396–401). Dodd responds by insisting that the aesthetic and expressive content of a work is a product of its sound rather than its connection to a performance means. He argues that it is possible, with imagination, to hear a piece as played on a given instrument without it being played on that instrument; and that the aesthetic and expressive properties which, Levinson claims, depend on performance means depend instead on imaginatively hearing the work as performed in a certain way.

4

kathleen stock

In the rest of the paper, Dodd defends this move against some further anticipated objections from Levinson. The type theory of musical works, of which Dodd is an exponent, comes under powerful attack in the second contribution to this section. In ‘Doing Justice to Musical Works’ Michael Morris makes two powerful objections to the view that musical works are types; and, indeed, with certain adjustments, to any view which classes musical works as general entities. First, the type theory cannot account for the fact that works of art, generally, are essentially meaningful; which is to say what they are ‘there for’ is to be understood. This is not just to say that meaning can be found in them, as it is in the natural world, for instance, but something stronger: that, in some sense, their nature prescribes that one try to understand them, so that to fail to do so is to fail ‘to do justice to’ them. The type theory is naturally understood as entailing that musical works cannot be created, but are instead discovered; indeed, this is acknowledged by its adherents (Dodd, for example). Yet, Morris argues, if musical works are eternal existents which cannot be created, then they cannot be essentially meaningful in the stronger sense just articulated. The type theorist cannot acknowledge that there is anything in a work to be understood in this sense; though she may acknowledge that musical works have characteristic psychological effects on listeners, she cannot think of them as products of understanding. Nor may she account for the meaning in a musical work by thinking of it as arising subsequent to the work’s ‘discovery’ by a composer; for this would be inconsistent with the point, entailed by the claim that musical works are essentially meaningful, that the meaning of a work is constitutive of the entity that it is. Morris’s second objection is to a particular version of the type theory: that musical works are types whose tokens are performances (though with slight adjustments he claims that it will apply to any conception of musical works as general entities). Performances, insofar as they let works be heard, allow them to be understood. But more than that, a performance only counts as of a given work if the performer understands the work, to some extent. This creates a problem for the type theory, insofar as it purports to be a reductive theory, aiming to specify the relation between performance (token) and work (type) in terms of some common property shared by both, without any reference to the work as such.

introduction 5 Morris concludes by casting some doubts on certain common assumptions of type theorists and others: that musical works must in some sense be ‘repeatable’; and that an explanation of what it is for a thing to be repeatable in the way a musical work is must lie in the assimilation of works and performances to more familiar ontological kinds, such as generals and particulars. Not all questions of musical ontology concern the nature of musical works per se. There are equally interesting questions to be raised about how we should think of those musical objects generated by the various stages of the process of composition. In ‘Versions of Musical Works and Literary Translations’, Stephen Davies is concerned with the ‘ontologically provocative’ issue of works, musical and otherwise, whose completion apparently precedes aspects of their composition. He urges that, in many cases, we should categorize a musical work of which it is true that aspects of composition succeed its completion, not as distinct from that work completed earlier, but as a ‘version’ of it. A work version, understood in Davies’s special sense, is produced wherever features of the work, which otherwise would be constitutive of its identity as such, are intentionally and ‘moderately’ altered, either by the author or, in the case of works whose composition is unfinished or ambiguous, by someone else. Davies carefully distinguishes the class of work versions from others with which it might be confused. First, work versions are distinct from drafts, which predate completion of a work, and pose no particular ontological puzzle; presumably because, at the time of their production, one cannot yet identify what will turn out to be identity-constitutive features of the work, and so there can be no question of such features being altered in production of the draft. Second, work versions are distinct from performance interpretations of a work: performances of a work which (pace Morris) reproduce its identity-constitutive features though which may deviate from what is prescribed by the composer in other ways. Third, they are distinct from ‘interpretation versions’: performance interpretations of a work, produced by a performer in order to be performed repeatedly, by her or others. A performance interpretation of a work, like a work version, may post-date the work in question’s moment of completion, but, unlike it, is not intended as a recomposition of that work, and—if it is accurate—nor does it alter any of its identity-constitutive features. Fourth, work versions are distinct from transcriptions: that is, adaptations of

6

kathleen stock

a medium-specific work to some new medium. These, like work versions, alter identity-constitutive features of a work: specifically, they change a work’s (identity-constitutive) instrumentation, such that the medium of the work, and perhaps also its contents, are then changed. Here, unlike in the case of a work version, the alterations wrought are sufficiently radical to license our calling their product a new, derivative work, rather than a version or interpretation of the original one. Having delineated the work version as a distinct entity, Davies examines its relevance as a category to other art forms, noting that work versions can be found elsewhere in the arts. One such place is in the visual arts, insofar as, for instance, an artist might moderately rework a painting after its official completion time. From this he draws the interesting consequence that, in one sense, paintings and other handmade visual artworks, traditionally thought of as ontologically singular, can exist in multiples. He denies, however, that no meaningful distinction can be preserved between works in these forms and those which are more usually thought of as multiply instantiable: for, while the former can exist only in successive versions (since a change in one version signals the end of it and the production of another), the latter can exist in multiple versions or instances simultaneously. Finally, Davies looks to the literary arts, urging us to think of translations as work versions rather than transcriptions of the original work, at least insofar as artistically relevant features of the original work are preserved in translation.

2. Musical Expression Another central philosophical problem raised by music, apparently separate from any ontological puzzles it raises, is that of saying what is meant by claims that attribute expressive or emotional properties to music, such as (rather, basically) ‘the music is sad’. This, broadly speaking, is the problem with which contributors in this section of the volume are concerned. As a question about the language we use to describe music, it tends to attract philosophers actively concerned with questions of meaning. Meanwhile, since many think the question is best tackled by giving an account of what it is to hear music as expressive, it also attracts those interested in analysing musical experience, often in terms of other more general sorts of experience (for instance, emotional experience of some sort).

introduction 7 One philosopher who takes the latter approach is Derek Matravers. In ‘Expression in Music’ he stresses that a relevant explanation of what it means, for instance, to say that music is sad (taken as a representative case), should describe what it is to hear music as sad, phenomenologically speaking; and, moreover, should make clear what such an experience has in common with cases in which the epithet of ‘sad’ is literally applied. It need not take the form of an account of the properties that cause the experience, since, generally, an account of the cause of a given experience need not illuminate the nature of that experience. Better, he argues, to characterize the experience directly. For his own part, he advocates a ‘dispositional’ theory, according to which music is heard as sad if it tends to induce in the listener some non-cognitive feeling(s) associated with a typical reaction to the expression of sadness (Matravers 1998). In his paper here, he indirectly defends this view by comparing it to other leading accounts, and showing that, despite some well-known objections, it is not obviously inferior to them. Along the way, he provides a helpful map of the territory. Broadly speaking, Matravers categorizes prominent attempts to characterize the experience of sad music into two kinds. The first has it that music is experienced as apt for description as ‘sad’ because it is heard in terms of behaviour expressive of sadness. This can be cashed out in several ways. One such way, roundly rejected by Matravers, is to say that music is sad in virtue of an experienced resemblance to sad behaviour. Another, proposed by Jerrold Levinson, is that sad music is heard by a suitably informed listener as the sui generic expression of sadness by a musical persona (1996: 107 et passim). Matravers initially objects that, whether this is understood simply as a redescription of the experience of sadness or as a positing of an imaginative activity supposedly explanatory of it, it is unhelpful, since in both cases, the supposedly antecedent notion of music being heard as the sui generic expression of a persona is dependent upon one’s grasp of what it is to hear music as expressive (but see Ridley: Chapter 6, this volume). In addition to critical discussion of Levinson’s responses on this point, Matravers launches a more general objection against his account: not everyone who hears music as sad would agree that they hear it as the expression of a musical persona, and in the absence of any consensus, it is unclear what resources Levinson may call upon to persuade dissenters that this is the right account. The second sort of answer considered by Matravers focuses, not on thinking of sad music in terms of behaviour expressive of sadness, but

8

kathleen stock

on the sad feelings produced in the viewer on listening. Matravers’s preferred dispositional account falls into this category. In defence of his view, Matravers sketches out the lines he takes against some wellknown objections to arousal theories. To the objection that feeling sad, understood as a full-blooded emotional state with cognitive components, is not necessary to the experience of hearing music as sad, he agrees, stressing that it is a non-cognitive component of a typical reaction to sadness which is necessary to such an experience, and that, moreover, it need not be as ‘strongly present’ in the experience as full-blooded sadness would be. To the complaint that such feelings are not sufficient for the experience of music as sad, he responds that the relevant feelings are those which closely ‘track’ the nature of the music, and hence which cannot be experienced in other contexts. To the objection that, even so, the posited feelings are insufficiently integrated with one’s aural experience of the music to be equivalent to the experience of expression, he denies this, arguing that the fact that two co-occurring mental states can be logically distinguished from one another does not show that they cannot be experienced as one unified state, phenomenologically. Towards the end of his rich paper, Matravers raises an issue affecting both his account and others; namely, that hearing music as sad can involve either hearing the sadness, not as a product of oneself but emanating ‘from without’, or, alternatively, hearing it ‘from within’, as the expression of one’s own sadness. Levinson’s persona account can accommodate the former fact, but needs to cite something like empathy on the listener’s part to accommodate the latter. Meanwhile, to accommodate such facts, Matravers suggests that his own view be supplemented by the claim that the feelings sad music induces in a listener can be empathetic or sympathetic. Assuming, like Matravers, that certain pieces of absolute music characteristically provoke emotion in the listener, in ‘Explaining Musical Experience’ Paul Boghossian reiterates that what we are after is not a mere causal explanation of this fact. Rather, he urges, we should want an explanation of why it is rational for us so to respond. To him, the most promising-looking such account is one which attributes expressive meaning to music. The problem of expressive meaning he then characterizes as that of identifying those properties of sounds which enable one to hear a musical work as expressive of some emotion. The relevant properties are musical properties: those of pitch, harmony, melody, rhythm, and so on. Hearing a work as having

introduction 9 certain musical properties (yet to be fully specified) makes the attribution of certain expressive meanings apt. He examines a potential problem for this claim. Roger Scruton has argued that musical properties are not literally possessed by sounds, claiming that they are metaphorically possessed instead ((1999); see Chapter 10, this volume, for an articulation of this view with respect to rhythm in particular). It might seem to follow, against Boghossian, that we cannot explain the presence of an expressive property in terms of the presence of musical properties. Indeed, Scruton argues that both musical properties and expressive properties are metaphorically rather than literally possessed by musical works, and concludes on this basis that there is no explanation available of why a particular musical or expressive description is apt for a given piece of music. In response, Boghossian denies that the attribution of such properties to music must be metaphorical. For one thing, he claims, the use of metaphor to describe music must be intentional, whereas the propensity to hear sound as having musical or expressive properties need not be. For another, for a metaphorical description to be applicable to an experience, the experience must possess some literal content, which the former is supposed to illuminate; yet musical experience does not possess such content. If expressive properties are not metaphorically possessed, then we need not yet deny that the expressive meaning of a work is potentially explicable in terms of some other set of properties. Meanwhile, Boghossian claims, even if musical properties turned out not to be real properties of sounds, this would not compel us to reject his claim about expressive meaning: namely, that a work is heard as having certain expressive properties in virtue of being heard as having certain musical properties. For to hear a work as having a certain property is compatible with its not having it. Though he does not defend it directly here, Boghossian is most attracted to the sort of view which says that a work is heard as having certain expressive properties in virtue of being heard as having those musical properties which sound ‘the way a person would sound who was expressing E vocally’ or ‘the way a person would look who was expressing E gesturally’. (He acknowledges, but leaves to one side the difficulties, raised by Matravers in Chapter 4, about identifying resemblances across sensory modalities.) As we have seen, an alternative to resemblance views has been presented by Jerrold Levinson. The rest of Boghossian’s paper is spent addressing,

10

kathleen stock

first, aspects of Levinson’s positive account of expression, and second an objection of Levinson’s to the resemblance view. I shall concentrate here on the first set of points. Boghossian focuses on a claim of Levinson’s presented in a recent version of his view (Levinson 2005): a work is expressive of some property E iff the work is disposed to be heard, by a suitably informed listener, as if it were an expression of E. (This is referred to here as a ‘dispositional account’ but is to be distinguished from the similarly-named sort of arousal view endorsed by Matravers.) It apparently would follow from this claim that hearing a work as expressive of E is equivalent to hearing it as being disposed to be heard by a suitably informed listener as if it were an expression of E. Yet, Boghossian objects, it is not clear that the equivalence in fact obtains; indeed, it is not clear how the latter property could be audibly detected at all. To this Levinson has responded that, at least for suitably qualified listeners, reference to such a complicated and non-audible dispositional property is not required in an analysis of what it is to hear a work as expressive, since for such listeners hearing a work as if it were an expression of E just is equivalent to hearing it as expressive of E (2005). Boghossian worries that this response does not yet show how a qualified listener (and so nor any other listener, presumably) can be justified in finding a work expressive of some particular emotion, which, as we have seen, is what he thinks an account of expressive meaning should demonstrate. In addition, he expresses two further worries. One concerns the potential vacuity of Levinson’s reference to a suitably qualified listener. The other is about his analysis of what it is for a qualified listener to hear a work as an expression of E in terms of her imagining something, given the relatively unconstrained nature of imagining generally. In ‘Persona Sometimes Grata: On the Appreciation of Expressive Music’ Aaron Ridley defends a version of a position rejected by Matravers in Chapter 4: that the experience of expressive music may involve the imaginative postulation of a musical persona in the relevant emotional state. Unlike Levinson, Ridley is not concerned to argue that such a postulation is necessary to any experience of expressive music whatsoever, and, unlike other commentators, for instance Stephen Davies (1997), nor is he concerned to argue directly that it is not necessary for such an experience. Rather he argues that sometimes it is necessary for full appreciation of a particular piece of expressive music. This is not simply a compromise position, but a principled stance emerging from his view that any conclusion

introduction 11 about what is involved in the experience of expressive music should be motivated primarily by considerations drawn from critical practice in music appreciation, rather than from facts about expressive properties in other objects, or generally. Ridley identifies the following as a possible issue between members of the pro- and anti-persona camps. In the most general version of each view, a member of the pro-camp might think that perception of a feature of a bearer (whether musical or otherwise) as expressive of an emotion E is apt iff the bearer is imaginatively perceived as a person; while a member of the opposing camp might think that perception of a feature of a bearer as expressive of E is apt iff the feature would be expressive of E were it a feature of a person. In the former case, the postulation of a persona is supposedly prior to the experience of expression; in the latter, it is supposedly posterior to it. Ridley notes that both positions, in such general forms, have little or nothing to say about the nature of the bearer: it might be a willow tree, or a musical work, or something else entirely. Yet this matters, because, although the dispute is usually treated as a conceptual issue, Ridley argues that, for many possible bearers, in fact its resolution is an empirical matter which cannot be decided in advance, since for many possible bearers (willow trees, for instance, that might be seen as sad) there are no normative constraints governing the precise grounds on which we should see them as expressive, even when we all agree that they can indeed be seen as expressive. He then turns to music, where, obviously, there are normative constraints governing the perception of musical works, and so where the discussion about whether a persona is involved or not is, to some extent, a conceptual one. Even so, he notes, participants in the debate tend to talk about musical works at their most general level as bearers of expressive properties, without paying attention to the distinctive nature of the bearer in particular cases. This, he thinks, makes it the case that the debate is bound, eventually, to become an impasse between those who think that the postulation of a persona is prior to the perception of expressive features in a work, and those who think it posterior to such perception. Instead, Ridley advocates paying attention to the nature of a given musical work in deciding whether postulation of a persona is required to properly understand its expressive character. He notes that this move is consistent with some of the objections made against the claim that involvement of a persona is

12

kathleen stock

necessary for perception of expression: for instance, even if Davies is right that perception of a feature of a musical work as expressive of E is apt iff the feature would be expressive of E were it a feature of a person (something he takes to be prior to the postulation of any persona), this would be consistent with Ridley’s own claim that a particular piece of music may require the postulation of a persona in order to fully appreciate its expressive nature; that is, to respond to it in a way which goes beyond what is simply ‘apt’. In support of the latter claim, and against generalist accounts, Ridley discusses two pieces of music, one for which the postulation of a persona enhances appreciation, and one where it does not.

3. Musical Meaning The third section of the volume contains two papers each concerned with musical meaning, in some sense. The first contribution, by Jenefer Robinson, is concerned with the relation between music and metaphor. Meanwhile, in their joint contribution, Tamara Balter and Eddy Zemach address the issue of music which apparently has ironic meaning by presenting a semantic theory of irony, understood as an aesthetic property which may be possessed by many kinds of music, and in many different guises. In ‘Can Music Function as a Metaphor of Emotional Life?’ Robinson examines whether music can function as metaphor: while she eventually concludes that strictly speaking, it cannot, she also urges that we see pieces of music, potentially, not just as having aspects apt for metaphorical description in expressive terms, but, more radically, as metaphorically exemplifying expressive qualities, by referring to extra-musical aspects of the world via certain expressive aspects possessed by them. (Following Goodman (1976), ‘expressive’ is used here in a generous sense to include, not only emotional qualities, but also those picked out by adjectives such as ‘weighty’, ‘watery’, ‘sunny’, ‘twittering’ and so on.) Robinson accepts a point rejected by Boghossian in Chapter 5: namely, to say that a piece of music has expressive properties is to make a metaphorical claim. In this, she follows Nelson Goodman. Goodman has argued that musical expression is a kind of exemplification, where (roughly, but accurately enough for our purposes) the exemplified qualities of an object are those which are possessed by the object; and to which the

introduction 13 object refers. Specifically, expression is metaphorical exemplification: the expressive qualities of a musical work are those which are metaphorically possessed by the work; and to which the work refers (Goodman 1976: 95). Though critical of aspects of this account, Robinson is sympathetic to it insofar as it draws attention to the important role played by metaphors in the description of music’s expressive qualities. However, she thinks that Goodman does not fully exploit the idea of metaphorical exemplification, restricting his attention to the metaphorical possession of mere qualities such as sadness or heaviness. To remedy this, Robinson makes a distinction between two kinds of metaphorically possessed qualities: those that belong to the ‘musical surface’ of a work, such as sadness and heaviness, and those deeper metaphors that characterize the unfolding of extended musical passages, perception of which structures one’s experience of the work in a more radical way. Further examples of the former include breeziness and fieriness; examples of the latter are ‘perseverance through adversity’ and ‘struggle leading to victory’. Goodman is concerned only with the former sort in his discussion of expression; yet, Robinson notes, if surface features of a work are apt for description in metaphorical terms it does not yet mean that the work refers to qualities of breeziness or fieriness, understood as extra-musical phenomena. So breeziness and fieriness need not be exemplified by the work, in a strong sense. In contrast, she suggests, where a work, understood as a whole, has expressive qualities such as perseverance through adversity, this phenomenon, understood extra-musically, is genuinely referred to by the work. In this sense, we can say that perseverance through adversity is (part of) the expressive meaning of the work, and is metaphorically exemplified by it. This is not, however, to say that the listener’s attention is distracted away from the work, Robinson argues, because where a work refers to some extra-musical phenomenon in this way, it has a structure isomorphic to it, and so reflection upon the relevant phenomenon leads one back towards the relevant aspect of the work and to enhanced understanding of it. A well-known objection to this sort of view is that the structure of a work must underdetermine what extra-musical features it refers to. Robinson thinks this objection can be seen off if we acknowledge that metaphorical talk of music is guided by ‘umbrella’ conceptual metaphors, which determine the course of our experience of a given work and what further metaphors we think of as appropriate to apply to it. Hence, we

14

kathleen stock

should say in response to the objection, not that a given work refers to some extra-musical phenomenon absolutely, but that it does so relative to some conceptual metaphor applied either to the work, or to the music of which the work is an example. Meanwhile, what conceptual metaphor it is appropriate to apply to a work will depend upon which conceptual metaphors it was conceived under at the time of composition; this is illustrated by Robinson with an extended discussion of the metaphor of absolute music as ‘discourse’, current in the eighteenth century as the form emerged. She goes on to suggest that the noticing of the way a piece of music metaphorically exemplifies, say, a conversation amongst interlocutors can enable us to hear emotional expression in the voices, and hence, the piece. None of this, however, forces us to admit that musical works themselves can be metaphors, in a strict sense. They cannot be, since they have no literal meaning to start with. In ‘The Structure of Irony and How it Functions in Music’, Eddy Zemach and Tamara Balter are concerned with musical meaning of a different kind. According to them, irony, whether verbal, situational, dramatic, general, parodic, or romantic, has certain necessary features. Namely, where a situation, whether in art or reality, is perceived as ironic it (i) causes some person, usually the perceiver, to project (that is, think of) a possible situation, where the possible situation thought of (ii) is thought of as ‘genidentical’ to the actual one considered (i.e. the two constitute versions of a single ‘transworld individual’ and as such are ‘counterparts’ of one another). In this sense, irony is essentially modal. A further characteristic feature of most varieties of irony is (iii) that the person concerned is caused to think unfavourably of the actual situation which confronts him, in comparison to its projected counterpart. Hence, on this account, the use of words is not essential to irony: what is essential is the projection of a possible situation against which reality may be compared, usually unfavourably, which is something which might be accomplished by music as well as words. According to Balter and Zemach, many forms of irony, including musical irony, fit this basic template. Much of the argumentative weight of the authors’ claims depends on their close readings of particular musical works, to which I am unable to do justice here. Instead I shall focus on the basic claims they make about different varieties of irony, and about how each may be manifested in music of a certain character.

introduction 15 The first sort of irony discussed is verbal irony, which involves understanding a sentence as ironic. Understanding a sentence ‘p’ as ironic, we are told, is equivalent to its causing one to project a possible situation that satisfies ‘p’, where this situation (i) is understood to contrast ‘sharply’ with its ‘real-world counterpart’, i.e. the relevant situation which obtains in the actual world, and with which it is thought of as genidentical; and (ii) is understood as superior in nature to that counterpart, so that it ‘mocks’ it. Since music does not have the kind of content that a sentence does, unless it includes lyrics, verbal irony is rarely found in music. However, it is argued, other types of irony, understanding of which does not depend on understanding sentences, are more frequently found in music, and it is to these that the authors turn next. Broadly speaking, these can be divided into two categories: on the one hand, those in which a possible situation is projected, and which is understood as superior to the actual situation of which it is a counterpart; and on the other, those in which a situation is projected on to the actual one and which is understood as unsatisfactory or deformed in some way. As well as verbal irony, situational and dramatic irony fall into the first category. Situational irony, occurs, it is claimed, when an actual situation is viewed by an onlooker with the thought of an idealized possible situation in mind, against which the former is unfavourably compared. The idealized situation comes to mind as the result of the onlooker’s apprehension of the actual situation, in conjunction with internalized norms which bring to mind how the situation should be, ideally. The category of ‘situations’ is here as elsewhere construed broadly, as potentially including musical events: hence, music can convey situational irony by including an actual compositional feature which is ‘mocked’ or ‘contradicted’ by the feature one expects to find, according to familiar norms governing composition. In the same category is dramatic irony, which can occur in fictions, including musical fictions such as opera, as well as in ‘real life’. In dramatic irony, a superior counterpart to the actual situation (or actual situation as depicted by the fiction) is projected, not, this time, by the onlooker who perceives the situation as ironic, but by some other (fictional) protagonist who believes the counterpart to be actual, whereas the onlooker sees that in fact this is not the case. The effect of this is to make the projected situation, and so too the projector, look ridiculous. Dramatic irony can be

16

kathleen stock

signalled in music, not only in the fictional narrative of an opera, but also insofar as music can signal to the listener, via her understanding of features of the work’s composition, that a fictional world is not as a protagonist takes it to be. Meanwhile, into the second category falls parody, and general and romantic irony. A parodic work is one intended to cause its spectator to realize some unpleasant truths about a particular actual situation, by causing her to project an only slightly deformed counterpart which she initially mistakes for the actual situation in question. Realization of her mistake then causes her to identify the deformation in the projected situation as present in the actual one. In this way, the work powerfully conveys some critical point about actuality. Likewise, in music, a work which is a parody of some other musical work or feature presents itself as a deformed version of its target, in order to persuade the listener that some of the unflattering aspects of the parody actually belong to the target as well. Romantic irony is similar to parody, in that a romantically ironic work is intended to cause its spectator to project a deformed version of some actual entity, so as to mock the latter; the difference being that in this sort of irony, the actual entity which is the target of the mockery is the work itself, or at least, some aspect of it. A piece of music displays romantic irony where it contains some aspect which is presented as genidentical to some other aspect of the same work, to which it is juxtaposed or in close proximity, and to which it contrasts sharply in order to mock it. Finally, general irony occurs wherever a situation causes an onlooker to project a number of counterpart situations, all of which she considers to be as deformed as the actual one. An art work manifests general irony when its narrative takes the audience to a number of ‘worlds’, each of which is successively or simultaneously ‘rejected’ without any accompanying sense of being rejected for something better. In the musical sphere, general irony occurs, for instance, in the simultaneous presentation, in a single piece, of diverse ‘levels, styles and topics’ which cannot be made to cohere melodically, harmonically, rhythmically, thematically or stylistically; as, for instance, in Stravinsky’s Three Pieces for String Quartet, which is illuminatingly discussed.

introduction 17

4. New Issues The two papers in this final section are distinguished by their careful attention to important issues which tend to be somewhat left in the philosophical shade besides those of musical ontology, expression and meaning. In his contribution, Gordon Graham addresses himself to developments in contemporary composition, enquiring, by way of a comparison with our experience of classical music, whether electro-sonic art can be called music at all, properly speaking. Meanwhile, in his paper, Roger Scruton offers us a detailed analysis of a crucial yet neglected aspect of musical experience: musical rhythm. In ‘Music and Electro-sonic Art’ Graham is concerned with the provocative question of whether a particular genre of composition should be categorized as music at all. Namely, he asks whether electro-sonic art works—digitally produced sequences of atonal sounds, by composers such as Lutoslawski and Var`ese—can be called music, properly speaking; and, furthermore, whether they afford the same opportunities for value as do musical works, either in terms of listening experience, and in terms of active engagement in their production. His answer depends on a prior understanding of the value of music. Music, according to Graham, is ‘intentionally organized sound’, and, as such, is necessarily to be listened to: not simply as a means to some non-aural experience only contingently available via aural experience (say, that of an expressed or felt emotion, or represented image or idea), but rather as an end in itself. His is not the claim that the content of music is essentially sui generis; rather, it is the claim that, whatever the content of music, it is necessarily accessed aurally, so that its sonic character is part of its value. In this respect, it contrasts with most speech tokens, whose valuable properties are only contingently aural ones, usually to do with the communication of information. Meanwhile, another part of music’s value which issues from its essentially aural nature are the multiple opportunities it affords to engage in its production: most obviously, by composing or performing, but also by the invention and production of instruments, the management of acoustics, and so on. Graham then moves to discuss electro-sonic art works, understood (roughly) as composed sound sequences which mainly employ as their basic

18

kathleen stock

constituents sounds other than the twelve tones of the chromatic scale. He notes that works in this form are, like musical works, intentionally organized sound sequences necessarily orientated towards the enrichment of aural experience, and not simply intended as a contingent means to valuable non-aural experience: however, he denies that such works count as music, since, in their wholesale abandonment of tones, they differ so radically from existing music. Such a denial is not meant to entail any particular evaluative stance towards electro-sonic art. This is made clear by subsequent discussion, where Graham gives a partly positive response to the question of whether electro-sonic art can enrich aural experience to the extent that music can, granting that it may share many structural features with music, and may turn out to share expressive features with it too. However, he notes a further difference between electro-sonic art and music which, as well as underlining the difference between the two, has implications for the comparative value of the former against the latter: namely, that electro-sonic art does not provide as many opportunities as music does for active engagement in its production, and hence cannot match music’s value in this respect. In particular, he suggests, the production of electro-sonic art allows little conceptual space for performance of it, distinct from composition. The second contribution to this section is that of Roger Scruton. In it, he focuses on musical rhythm, analysing it first and foremost as a feature of musical experience and distinguishing it from any purely structural features of composition, such as metre. He grants that the rhythm of a work is often established via its metre, which organizes a work into repeatable segments, and can lead us to hear certain notes within each segment as stressed or unstressed. However, the existence of metre in a work is not yet the experience of rhythm, Scruton argues. That it is insufficient for rhythm is demonstrated by works without a discernible rhythm from non-classical metric traditions, both non-Western and modern; meanwhile, that it is unnecessary is shown by rhythmic non-musical phenomena, such as speech and dancing. Furthermore, within music, Scruton notes, rhythm can be generated or at least enhanced, not just by accents and stress, but by melody and harmony as well. Such facts lead Scruton to distinguish between the ‘beat’ of a work—the pattern of time values and accents it contains—and its ‘rhythm’. Unlike beat, rhythm can emerge from melodic and harmonic factors as well as metric

introduction 19 ones, and the interplay between them, and is a way of hearing a work as moving in a certain way. Whereas the beat of a work can be conveyed via symbols, experience of musical rhythm can only be characterized phenomenologically, and is identified by the ear, and perhaps also by the body, in dance. Such claims reflect Scruton’s wider view that the experience of music is an ‘acousmatic’ one, in the sense that it is heard as governed by a ‘virtual causality’, linking one sound to the next via such metaphorical aspects as movement, position and distance, and not wholly dependent on the organization of causal features of its constituent sounds. An important distinction is drawn by Scruton between ostinato rhythm, in which a rhythm is imposed via factors ‘external’ to the melodic and harmonic lines, and rhythm which is internal to, and emerges from, melodic and harmonic movement. An example of ostinato rhythm is that typically found in pop and dance music, where the rhythm is imposed via a percussive beat wholly distinct from melodic or harmonic constituents. In contrast, in the predecessors of pop music, jazz and early rock, as in most classical music, percussion is typically used to enhance a pre-existing rhythm generated from ‘within’, either by instrumental or vocal aspects of melody or harmony. In atonal music, Scruton suggests, rhythm of this kind is difficult to produce, since usually it relies on an impression of movement and return, which in turn usually depends on the use of harmony and melody. Often composers of atonal music are forced to compensate by introducing rhythm from outside, as it were, via ostinato, or some metrical device. An exception here is Messiaen, and Scruton discusses his techniques for generating rhythm internally in some depth. He identifies Wagner as a precursor to Messiaen in this respect, as an innovator in his ability to generate rhythm as an internal feature without reliance on techniques familiar from the classical tradition. Another way in which the atonal composer can generate rhythm is by incorporating speech rhythms into a work, and in this regard, also, Wagner is identified as an innovator. Meanwhile, from the distinction between rhythm which is generated as part of the ‘musical surface’ of a work, and ostinato imposed from without, Scruton extracts some practical and even moral consequences for the nature of dances associated with each. For one thing, he suggests, the prevalence of modern dance music employing ostinato has perniciously narrowed musical preferences towards music of a similar kind. Perhaps more seriously, he argues that, whereas dance music which generates rhythm from within

20

kathleen stock

guides people as to how they should move, thereby facilitating social interaction and directing their attention towards other people, dance music which uses ostinato, in contrast, does not facilitate and even inhibits such interaction, encouraging the individual to become isolated from others, and ‘lost’ in the music in a morally detrimental way.⁴ References

Davies, S.(1997) ‘Contra the Hypothetical Persona in Music’, in M. Hj¨ort and S. Laver (eds), Emotion and the Arts, Oxford: Oxford University Press, 95–109; reprinted in his Themes in the Philosophy of Music, Oxford: Oxford University Press (2003), 152–68. Dodd, J. (2000). ‘Musical Works as Eternal Types’, British Journal of Aesthetics 40: 424–40. (2002) ‘Defending Musical Platonism’, British Journal of Aesthetics 42: 380–402. (2004). ‘Types, Continuants and the Ontology of Music’, British Journal of Aesthetics 44: 342–60. Goodman, N. (1976) Languages of Art, Indianapolis: Hackett. Levinson, J. (1980). ‘What a Musical Work Is’, The Journal of Philosophy 77: 5–28; reprinted in his 1990b, 63–88. (1990a) ‘Authentic Performance and Performance Means’, reprinted in his 1990b, 393–408. (1990b) Music, Art and Metaphysic, Ithaca, NY: Cornell University Press. (1996) ‘Musical Expressiveness’, in his The Pleasures of Aesthetics, Ithaca, NY: Cornell University Press, 90–125. (2005). ‘Musical Expressiveness as Hearability-as-Expression’, in M. Kieran (ed.), Contemporary Debates in Aesthetics and the Philosophy of Art, Oxford: Blackwell. Matravers, D. (1998) Art and Emotion, Oxford: Oxford University Press. Scruton, R. (1999) The Aesthetics of Music, Oxford: Oxford University Press. Wolterstorff, N. (1980) Works and Worlds of Art, Oxford: Clarendon Press.

⁴ Thanks to those contributors who offered helpful criticism of an earlier draft of this Introduction.

PA RT I

Musical Ontology

This page intentionally left blank

1 Sounds, Instruments, and Works of Music J UL I A N D O D D

1 A satisfying account of the ontology of works of music must do two things.¹ First of all, it must enlighten us as to musical works’ ontological nature: that is, tell us to which ontological category such works belong. Building on this, it must then explain how musical works are individuated: that is, provide an informative account of when we have one and the same work of music and when we have numerically distinct such works. What I shall term the simple view does just this.² Its first component—what I shall call the type/token theory—has it that works of music are types whose tokens are sound sequence-events: datable, locatable patterns of sounds.³ Its second component—sonicism—has it that work W = work W ∗ if and only if W ¹ By ‘works of music’ I mean works of pure, instrumental music. This restriction should be understood to apply throughout this paper. ² The simple view (or, at least, something approaching it) is defended in my 2000, 2002, 2004 and 2007. ³ Such sound sequence-events will tend to be either performances: sound sequence-events produced by someone performing the work; or else playings: sound sequence-events that involve either the reproduction of such performances via CDs and the like, or the use of a playback artefact (such as a player piano) in the production of the sounds. However, I see no reason to rule out the possibility that a work could have tokens that were neither performances nor playings. The wind blowing through the trees could produce a token of ‘Greensleeves’. ( This example is Wolterstorff ’s (1980: 86), though he is less happy about accepting it as a consequence than I am. According to Wolterstorff, whilst there is no conclusive reason to restrict a work’s tokens to performances and playings, such a restriction is desirable since it accords with our pre-theoretical intuitions about what kinds of thing do, and do not, count as instances of works. I am unpersuaded by his discussion (1980: 86–8), but will say no more about this local dispute.)

24

julian dodd

and W ∗ are sonically indistinguishable. However, both of these claims need to be further clarified and rendered more precise. To begin with, the type/token theorist should treat the types that are works of music as norm-types: types that admit of properly and improperly formed tokens. In this respect, works of music resemble types such as The Polar Bear and The Piano:⁴ something can be a polar bear and yet be missing an ear; something can be a piano and yet be out of tune; and something can genuinely count as a performance of a work and yet contain mistakes here and there. As Nicholas Wolterstorff has explained (1980: 54–8), granting this latter point is essential, if we are to allow for the fact that a work can be improperly performed. Such incorrect performances are flawed but performances of the piece nonetheless. With this in mind, let us follow Wolterstorff (1980: 58) in saying that a property F is normative within a type φ just in case (i) φ is a norm-type; and (ii) it is impossible for there to be something that is a properly formed token of φ and which lacks F. Consequently, for any work of music W , there is a set  that comprises the properties normative within W . In order for a performance to be a properly formed token of W , it must have every member of ; to count as a performance of the piece at all, a sound-event must not lack too many of ’s members. This claim is fine as far as it goes, but merely serves to raise the question of how such types are individuated, albeit in a new form. For the question now becomes: what is the nature of the properties that are members of ? The second component of the simple view—sonicism—has a straightforward answer: the members of  are all purely acoustic in character, and, hence work W = work W ∗ if and only if W and W ∗ have the same—purely acoustic—properties normative within them. According to a sonicist, whether a sound sequence-event is a well-formed instance of W is solely determined by how it sounds. Nothing else matters.

2 Having distinguished the two components of the simple view, I shall now put the first to one side and focus squarely on the second. In particular, I want ⁴ I capitalize names of types.

sounds, instruments, and works of music

25

to consider whether the sonicist is right in thinking that the properties normative within works of music comprise merely acoustic properties and not, in addition to these, performance-means properties: properties concerning how a performance’s constituent sounds are produced. Boiled down to its bare essentials, my question is this: can a properly formed token of a work W only be produced by means of the particular instrumentation specified in W ’s score? According to the sonicist, of course, the answer to this question is a resounding ‘No’. Since the only members of  are acoustic properties, a properly formed instance of W may be produced by other means of sound production than those specified by the composer. A properly formed performance of the Hammerklavier Sonata, for example, need not have its constituent sounds produced by the playing of a piano; the sounds constituting such a performance could also be produced by the operations of a space-age Perfect Timbral Synthesizer: a machine that emitted acoustic facsimiles of piano-sounds. For the sonicist, whether a sound sequenceevent faithfully represents the piece is determined by its acoustic appearance alone. How these sounds are actually produced is by the by. Swimming against a strong intellectual current, I wish to defend a version of this sonicist claim. And in so doing, I reject what I shall call instrumentalism: the view that a properly formed token of W must be produced by the instrumentation called for by W ’s score.⁵ The instrumentalist insists that a synthesized performance of the Hammerklavier —even if acoustically indistinguishable from a performance she acknowledges to be properly formed—would nonetheless be a defective presentation of the piece. And her saying this amounts to a claim that the sonicist’s description of the properties that constitute  is incomplete. For the instrumentalist has it that the properties normative within W include not only acoustic properties, but precisely those performance means-properties that the sonicist ignores. Clearly, if the instrumentalist is right about this, then the sonicist has given, at best, only a necessary condition for work-identity. For once it is granted that performance means-properties are normative within works, we have no choice but to admit that W and W ∗ —even if acoustically indistinguishable—are distinct, if their respective scores specify different means of performance. ⁵ Those who argue vigorously and ingeniously against sonicism from the perspective of instrumentalism include Stephen Davies (2001: 60–71), Stan Godlovitch (1998), Jerrold Levinson (1980, 1990a, 1990b) and Kendall Walton (1988).

26

julian dodd

In what follows I shall argue that a suitably nuanced sonicism can repel the challenge presented by instrumentalism. And my route to this conclusion will be via two claims. First, I shall present one particular variant of sonicism—timbral sonicism—as the face-value theory when it comes to the question of musical works’ instrumentation. It is, I shall claim, the theory that we should accept unless it is defeated. Second, and crucially, I shall argue that the arguments offered by instrumentalists are insufficient to defeat it. To be sure, sonicism is an unfashionable doctrine. But then again, aestheticians, of all people, should not be fashion victims.

3 Let us treat  as the set of properties normative within the Hammerklavier: the set of properties a sequence of sounds must have to be a properly formed token of the piece. The sonicist, we have seen, takes these properties to be purely acoustic in nature: features concerning the pitch and duration of notes, melodic, harmonic and articulational features, and such like. Such a view, I have claimed is prima facie correct. Why is this so? For two reasons. First of all, sonicism—or a version thereof—is the default position when it comes to musical works’ individuation. There is something deeply intuitive about the idea that works of music are pure types of sound-event: in other words, that what makes a certain work that work is simply that its performances should sound like that. And I am not alone in thinking this. Even Jerrold Levinson, one of sonicism’s most articulate and ingenious opponents, accepts that, when it comes to the types that are musical works, ‘the most natural and common proposal ... is that a musical work is a sound structure—a structure, sequence, or pattern of sounds, pure and simple’ (1980: 64). Ultimately, I doubt whether adequate sense can be made of the claim that a type has parts or constituents,⁶ which is why I have set up the sonicist thesis without any talk of ‘sound structures’ or the like. But whatever view we take on the ontological nature of types, Levinson’s suggestion that sonicism is intuitive and natural is surely correct. Someone unencumbered by developments in analytical aesthetics will most likely ⁶ For an attack on the idea that types are structured, see my 2007: Ch. 2.

sounds, instruments, and works of music

27

accept that a performance of the Hammerklavier on a Perfect Timbral Synthesizer, indistinguishable to the ears from a performance on a piano, would be a no less satisfactory presentation of the piece. True enough, if such a synthesizer were easier to play than a piano, we might think less of the performer’s achievement, but the respective patterns of sounds the performers produced—qua representations of the piece—would surely be regarded as being on a par. Intuitively, the vehicles of musical meaning are the sounds we hear—to appreciate a work of music we need only use our ears—and in the example we are considering, the performances in question are acoustic doppelg¨angers. Here, then, we have a bare intuition pointing towards sonicism. But further weight may be lent to the sonicist cause by considering how well it coheres with a plausible account of the nature of musical experience. Roger Scruton describes our musical experience as acousmatic (1997: 2–3); and what he means by this is that, in hearing sounds as music, we attend purely to the sounds themselves, and not to their causal origin. Scruton believes that our experience of sounds as music sees us detach these sounds from the circumstances of their production, and hear them as organized according to pitch, rhythm, melody and harmony (1997: 20). We hear the sequence of sounds as an organized whole, something that develops and progresses, so that, rather than focusing on the fact that a middle C is produced by an oboe, for example, we hear it as the response to the B that preceded it, and as calling for the E that follows it (Scruton 1997: 19). Obviously, if this view of our musical experience is correct, instrumentalism takes on a hugely uncompulsory air. For if hearing the music as it ought to be heard requires us to attend merely to its qualitative nature, and not to the ways in which the sounds are produced, then nothing about the nature of our musical experience threatens to dislodge us from the default position that is sonicism. Having said this, Scruton’s account of our experience of music is flawed. For, as we have seen, Scruton believes the only essential features of our musical experience to be organizational features: pitch, rhythm, harmony and melody. As a result, he comes to regard timbre as an inessential feature of our musical experience, even though he accepts that it can contribute to musical meaning (1997: 77). In my view this is a mistake: timbral properties, as we shall see in a moment, are normative within works, for the reason that their presence determines many of a work’s aesthetic properties; so even though timbre is not itself

28

julian dodd

an organizational feature of sound sequences, an appreciation of timbre is essential to our experience of sounds as music. What follows, if we depart from Scruton’s position this way? As Scruton himself notes (1997: 77), timbral features are specified by reference to their customary origin: if we are to avoid recourse to metaphor, we have no choice but to conceive of a certain timbre as ‘the sound of a middle C played on an oboe’, for example. Consequently, it follows that Scruton is wrong in thinking that our musical experience can always be described without making any reference to sounds’ usual origin. But, as it happens, this turns out to be a relatively small amendment, for what we might term a modified acousmatic account of our musical experience is still compatible with a denial of instrumentalism. Whilst it is true that we must make reference to certain sounds’ customary causal origins in order to put our finger on their timbral aspects, what does not follow from this is that we need be cognizant of the sounds’ actual origins; and this leaves it open for these sounds to have been produced by non-standard means. We do ‘not have to identify [the sounds’] cause in order to hear them as they should be heard’ (Scruton 1997: 3); we need only characterize them in terms of their customary means of production. And if this is right, and if we grant the highly plausible thesis that how a musical work should be heard is revelatory of its ontological character, then our modified acousmatic account of the nature of musical experience strongly suggests that we need not depart from sonicism. In order for an audience to hear a performance as it ought to be heard, the performers need not use certain specific instruments; they need only produce sounds with the sonic appearance of sounds produced by such instruments. Be this as it may, the sonicist is obliged to explain what she takes the ‘auditory character’ definitive of a work’s identity to consist in. Which acoustic properties comprise set : the set of properties normative within the Hammerklavier? According to the view that Stephen Davies calls ‘pure sonicism’ (2001: 60), the members of  are the sorts of properties one would expect—melodic, rhythmic, harmonic, articulational, tempic properties, and the like—but with one notable exception: timbral properties. On such a view, a performance of the Hammerklavier would be in no way defective, if produced on a Hammond organ, as long as the resultant sound sequence had the tonal structure and other (non-timbral) acoustic properties indicated by its score. In my opinion, though, such a position is too extreme to

sounds, instruments, and works of music

29

be acceptable. Indeed, as I shall now explain, it is implausible in itself, inadequately motivated, and subject to a serious objection. However, an appreciation of pure sonicism’s demerits opens up the possibility of there being a version of the doctrine that avoids these pitfalls. That the pure form of sonicism is counter-intuitive (almost) goes without saying. A performance of the Hammerklavier on a Hammond organ or, perhaps, by a flute quartet,⁷ might capture the piece’s tonal make-up, but such a performance would entirely ignore the composer’s specification of a piano in the score. As I shall argue presently, the fact that a score instructs performers to use specific instrumentation does not entail that a properly formed performance must use the instrumentation specified. Such an instruction is essentially context-bound and provisional, its purpose being that of instructing performers how to produce a properly formed sound sequence given the instruments available at the time. The goal of such an instruction—the production of sound sequences of a certain qualitative character—would be met even if such a sound sequence were produced without such instrumental instructions having been followed. Having said this, Beethoven’s specification of a piano on the score of the Hammerklavier does at least demonstrate that a correctly formed performance of the piece should be constituted by piano-like sounds. His scoring of the piece for piano was not completely arbitrary; in specifying a piano he was making it clear that performances of it would be inadequate, if constituted by the kinds of sounds made by organs, violins or flutes. The fact that Beethoven specified that a piano be used, whilst not entailing that being produced by a piano is normative within the work, does serve to characterize the qualitative nature of the sounds that must constitute a proper instance. Such sounds should have the timbral quality typical of sounds produced by such an instrument. Saying this, of course, serves merely to invite the sonicist to retreat from the pure version of the doctrine to its timbral counterpart: the version that takes timbral properties to figure in . As Levinson has noted, it is a decisive objection against pure sonicism that it renders some of a composer’s choices—for instance, trumpet or oboe, or bassoon or cello—motiveless and arbitrary (Levinson 1990a: 244). But the timbral sonicist precisely avoids this charge by ensuring that timbral properties are normative within ⁷ This example is Levinson’s (1990a: 240).

30

julian dodd

musical works. Elsewhere, Levinson recognizes that timbral properties, though generally specified via performance means properties, are in principle ‘physically and logically separable from them’ (1990a: 220); and an appreciation of this possibility enables us to formulate the weaker, timbral version of sonicism that can account for the importance of the choices that composers, in practice, make. It is true, of course, that a performance of the Hammerklavier on a Hammond organ would still be recognizable as a performance of the piece; and this, perhaps, is the source of the pull that pure sonicism exerts over some philosophers.⁸ But, having set out the type/token theory in the way in which I have, such a motivation for this version of sonicism is swiftly revealed as misconceived. A performance using the Hammond organ could, indeed, count as a performance of the piece; but acknowledging that this is so is quite compatible with recognizing that a properly formed performance must be made up of piano-like sounds. A work of music is a norm-type, remember: its identity is determined by the properties that are normative within it. Consequently, the mere fact that there could be a performance of the Hammerklavier on a Hammond organ (if such a performance had a sufficient number of the other properties normative within the work) does not demonstrate that timbral qualities are not members of . It is a familiar fact that a work may have instances that are improperly formed. After all, that there can be performances (albeit imperfect ones) of the piece that miss out certain passages, play wrong notes, or misconstrue the rhythm does not show that a properly formed token of the work need not obey the score in these respects. Pure sonicism is thus an under-motivated doctrine. But it also faces a decisive objection whose origin lies in the work of Levinson (1980: 73–8). For Levinson draws our attention to the fact that many of a work’s aesthetic and expressive properties depend upon their timbral, in addition to their more broadly acoustic, features. The Hammerklavier, Levinson notes, is a ‘sublime, craggy and heaven-storming piece of music’ (1980: 76), but a sound sequence would fail to represent it as such, if it were produced by a Hammond organ or (worse still) by a recorder ensemble.⁹ Likewise, John Cage’s In a Landscape would lose much of its gentle, exploratory, ⁸ Over Peter Kivy (1988), for example. ⁹ The latter example is Stephen Davies’s (2001: 64).

sounds, instruments, and works of music

31

contemplative character, if played on a trombone. What examples such as these reveal is that a properly formed performance of a piece—one that, among other things, does justice to the work’s aesthetic content—must have a specific timbral character. It is insufficient merely for such a performance to be properly formed with respect to its other audible features. This being so, it is my contention that a philosopher appreciative of sonicism’s status as the face-value theory, and yet sensitive to the problems besetting sonicism in its purest form, should adopt timbral sonicism. But be this as it may, the instrumentalist insists that a proper understanding of the role played by scores, and of the features determinative of a work’s aesthetic content, reveals that performance means-properties, and not merely timbral properties, are normative within the types that are works of music. In the next two sections I explain why such a conclusion is unwarranted.

4 As we have seen, the crux of dispute between the timbral sonicist and the instrumentalist concerns whether performance means-properties are normative within works. If they are, then, pace timbral sonicism, W and W ∗ may yet be numerically distinct even if sonically indistinguishable (if they differ with respect to the performance means-properties normative within them). Consequently, given timbral sonicism’s status as the facevalue theory of musical works’ individuation, we need to consider whether the instrumentalist’s arguments for performance means-properties being normative within works are sufficiently strong to warrant our giving up what amounts to the default position. The first such argument concerns the nature of a work’s score. In §3 I charged the pure sonicist with disregarding the fact that composers specify the use of certain instruments to produce sounds. Could it not be argued that this manoeuvre equally demonstrates the falsehood of even timbral sonicism? Does not such an appeal to the specification of instruments in scores throw one directly into the arms of instrumentalism? Levinson, for one, thinks that it does. In his view, scores are definitive of musical works, and so, if Beethoven’s Quintet Op. 15 calls for a clarinet to be played, this means that ‘the Quintet Op. 15 without a clarinet is not the same

32

julian dodd

piece—even if all sound-structural characteristics (including timbre) are preserved’ (1980: 75). Levinson’s point, in other words, is that Beethoven’s specification of a clarinet means that the work essentially involves the clarinet, and hence that a sound-event cannot be a properly formed instance of the piece unless a clarinet has been used in the production of the sounds.¹⁰ If Levinson is right, being produced by a clarinet is normative within the work. However, as I briefly explained in §3, this way of understanding the instrumental instructions in scores is not obligatory. Indeed, once one accepts timbral sonicism as the prima facie, natural way of individuating works of music, one will interpret scores’ instrumental specifications in another way entirely. This alternative reading has it that the way to make sense of the fact, for example, that the Hammerklavier’s score calls for a piano is to see this apparently exceptionless instruction as having an altogether more provisional nature. For the philosopher impressed by timbral sonicism’s intuitiveness, a work’s score specifies how a sound sequence-event must sound, if it is to be a properly formed instance of the work; and the score’s specification of certain instrumentation should be seen in this context: as exclusively serving the aim of facilitating the production of sound sequence-events with certain specific timbral features. True enough, the Hammerklavier’s score contains an instruction to produce such sounds in certain specific ways, i.e. by means of a piano. But this does not entail that a properly formed performance must involve the specified instrumentation, as opposed to some other aurally indistinguishable means of sound production. For this instruction is aimed at a particular constituency—namely, potential performers of the piece at the time the piece was composed—and is wholly in the service of the production of sound sequence-events of a certain qualitative nature. In short, the score was instructing performers in 1818 to use a piano simply because no other means of producing sounds with the requisite timbral features existed at that time. However, since the purpose of such an instruction is merely to specify the tone colour that must be had by a well-formed instance of the work, it is an instruction that can, in principle, be overruled as long as sounds with the right timbral qualities are produced. ¹⁰ This position is endorsed by Currie, who claims that ‘[a] performance that violates the composer’s directions as to how the sounds are to be produced is not a correct performance of it’ (1989: 49).

sounds, instruments, and works of music

33

A predictable objection to this is to point out that composers often believe their own instrumental instructions to have the status of exceptionless instructions for performers. It may well be the case, for instance, that Beethoven, had he been given the choice, would have preferred, for whatever reason, a performance of the Hammerklavier on the piano rather than one that made use of a Perfect Timbral Synthesizer. But the crucial point is that such a consideration cuts little ice against the timbral sonicist’s arguments. Convinced by her doctrine’s status as the default position, and impressed by the way in which it coheres with the modified acousmatic account of our musical experience, the timbral sonicist concludes that performance means-properties cannot be normative within works, and hence that instrumental specifications in scores should be interpreted in the alternative way suggested. Indeed, it seems to me that our objector may well have been misled by a misunderstanding of the sonicist’s claim about the instrumental specifications in scores. That such specifications serve solely to characterize the qualitative nature of the sounds constituting a properly formed occurrence of works is not a psychological claim concerning the nature of composers’ intentions; it is a claim concerning how such specifications should be interpreted in the light of mature philosophical theory concerning both the individuation of works of music and the nature of our musical experience. The timbral sonicist denies that the proper way to perform a work can simply be read off from the composer’s actual (or hypothetical) preferences, and hence allows that composers can misinterpret their own works. Naturally, saying this will not of itself placate the instrumentalist. Davies, for example, would be tempted to reply by drawing our attention to kinds of scores that seem especially concerned with the production of sounds on specific kinds of instrument. One such example is provided by sixteenth-century lute tablature, in which the notation tells the performer where to place his hands on the neck of the lute. Such a score, Davies suggests can only be ‘[i]nstrument specific. Rather than thinking of his piece as an abstract sound structure and being indifferent to the manner of realizing it, [the composer] is addressing himself expressly to a musician who is holding a lute in his hands’ (Davies 2001: 62). But such an example provides no new challenge to the timbral sonicist. Right enough, the producer of such a score addresses himself to someone holding a lute, but it does not follow from this that a properly formed token of the piece must

34

julian dodd

be produced by a lute rather than by an artefact that produces lute-like sounds. The same kind of response as before suffices to make the point: although the composer addresses himself to lute players, his instructions are (dare one say it) to be interpreted instrumentally;¹¹ they function merely as a means of specifying the qualitative nature of the sounds that performers must produce. So, given that sounds with such a qualitative nature could also be produced by non-lutes, a performer may ignore the letter of the score’s instrument-specific instructions, just as long as she is able to produce sounds indistinguishable from those produced by a lute player. Whether he recognizes it or not, how the sounds are produced is only of practical, de facto interest to the composer: an interest in performance means subserves the aim of specifying the acoustic appearance of the sounds themselves. Levinson makes a similar objection to Davies’s, but, ultimately, to no greater effect. Levinson too objects to the sonicist’s conception of a score as a recipe for providing an instance of a type of sound-event for its own sake. As he sees it, ‘[w]hen Beethoven writes a middle C for the oboe, he has done more than require an oboelike sound at a certain pitch—he has called for such a sound as emanating from that quaint reed we call an ‘‘oboe’’’ (1980: 74). And one reason why Levinson takes this line is that he believes it to be the only position that makes sense of the way in which composers actually think of the tone colours that they wish to see produced. As he himself puts it: [c]omposers 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 canvasses; it would obviously be pointless or at least frustrating to do so. (Levinson 1980: 74)

To my mind, though, this objection sees Levinson mistake an epistemological fact for a metaphysical one. For the fact that a work’s composer describes the kinds of sounds he wants produced by referring to their customary origins does not entail that only sounds with such origins can figure in a correct performance. A timbral sonicist, we noted in §3, accepts that sounds ¹¹ This joke is Kivy’s (1988: 85).

sounds, instruments, and works of music

35

must be thought of in terms of their customary origin, if their timbral aspects are to be characterized. What she denies is that this connection in thought between sound and customary origin need actually obtain for a performance to be properly formed. As we have just seen, there is a defensible interpretation of scores available to the timbral sonicist, according to which instrumental specifications serve merely as a means for characterizing how correct performances should sound. In the light of this, pointing out what we all know—namely, that we think of tone colours by means of thinking of their customary origin—is beside the point. Once this observation has been made, Levinson’s attempts to bolster what we might call the argument from scores against timbral sonicism are doomed. No doubt, he is right to point out that composers often conceive of a use of instrumentation prior to conceiving of pure types of sound: double-stopping is an example of this, as is pizzicato. We may agree with Levinson that ‘ideas for pizzicato passages do not first occur to composers and then, only later, what will serve, for the nonce, to realize them’ (1990a: 244). But, to reiterate, this evident fact does not have the implications for the conditions of correct performance that Levinson takes it to have. Facts about how composers conceive of sounds do not determine which kinds of properties are normative within works. What, then, is left of the instrumentalist’s appeal to the authority of scores? Not much. One might try to bolster it by adopting what Stan Godlovitch describes as a ‘socio-historical perspective’ on the question (1998: 56). According to Godlovitch, once one understands that musical communities are akin to guilds—defining their own conditions for membership and rank—one comes to appreciate that such communities will automatically be hostile to the idea of performing the Hammerklavier upon a Perfect Timbral Synthesizer, even if such a performance were an exact acoustic replica of a performance on a piano. Guilds, reports Godlovitch, ‘systematically resist instrumental innovations in order to preserve their own structure which requires the establishment and maintenance of a skill hierarchy based on handicaps legitimised by the Guild’ (1998: 53). According to the music-making establishment, then, the value of a performance is not simply determined by how it sounds, but by how it is achieved. In particular, the performer must demonstrate skill in overcoming the handicaps presented by playing the piece on the kind of instrument sanctioned by the musical community (1998: 57).

36

julian dodd

Now, I see no need to disagree with any of this. Godlovitch, it seems to me, makes a convincing case for his sociological account of the values held dear by musical communities. But the crucial point is that such a descriptive account of how music-making communities come to the opinions that they do on matters of instrumentation can in no way support a philosophical thesis concerning the individuation of works of music. That the practice of musicians embodies a commitment to instrumentalism does not entail that this doctrine is correct. To make good this latter thesis we need to come up with sound philosophical arguments, not a sociology of belief. Merely reporting the beliefs embedded in musical practice does not help us decide whether these beliefs are true. Having abandoned such a sociological approach, it might be tempting to try to buttress the argument from scores by arguing that the norm of authenticity demands that performers use the instruments specified by the composer. As Davies sees it, ‘[i]f the composer determinatively instructs that certain instruments should be used, they must be played if the performance is to be fully authentic’ (2001: 69). And Davies thinks this because he believes that there exist conventions in music that necessarily connect the scored sounds with the specified instrumentation: [I]mprovisation, the creation of works, the public specification of works, the performance of works, the reception of works—all these assume, involve, and rely on practices and conventions, some of which are predominantly musical and others which have more to do with social structures and purposes. As a result, connections that otherwise might be merely contingent take on a different status because they become normative within the relevant practice ... . Within the relevant practices, playing the appropriate instruments is not merely a useful means to the production of the desired result, which is supposedly the creation of an abstract sound structure. Instead, the use of those instruments is part of the end, which is to articulate the specified sound structure on the required instruments. To do otherwise is to ‘cheat’ and thereby to undermine the playing’s status as a performance. (Davies 2001: 64–5)

I remain unconvinced, however. For one thing, and as I have explained already, the composer’s instruction to use certain instruments can be viewed as provisional. To be true to the piece, a performer need only produce sounds with the same timbral features as sounds produced using the specified instruments. Authenticity in performance is, indeed, a desideratum, but what

sounds, instruments, and works of music

37

I am suggesting is that the features required for a performance to be authentic cannot be simply read off from the score in the way in which Davies suggests. The simple specification that a piano be used, for example, must be interpreted in the light of the nature of our musical experience and the fact that sonicism is the face-value theory when it comes to the question of the relation between works of music and the instrumentation specified in their scores. Davies will remain unsatisfied by this, no doubt. Indeed, at this point he will surely seek to press his claim that the conventions and practices underlying the making and appreciation of music give composers, musicians and listeners an essential interest in how sounds are produced, not just their qualitative appearance. But the case that Davies presents as an example of such a convention is nothing of the kind, as we shall now see. Davies claims that we regard a performer who uses alternative instrumentation to that specified in the score as having cheated, and that such a decision on the part of the performer undermines the playing’s status as a performance (2001: 65). This claim, though, needs careful handling. To begin with, the timbral sonicist, by contrast with his pure sibling, will agree that a performance of a piece on instruments that produce sounds of a different tone colour to those produced by the specified instruments is inauthentic and deficient. So let us focus on a case in which the instrumentalist’s and timbral sonicist’s positions diverge: our imagined example in which the Hammerklavier is performed on Perfect Timbral Synthesizer. Davies argues that such a performance would be a less adequate performance of the piece than a sonically indistinguishable performance making use of a piano; and he takes this to be demonstrated by the fact that we would regard the producer of the synthesized sounds as having cheated in some way. As a matter of fact, it seems to me that we can distinguish two kinds of case here, neither of which have the implications that Davies believes them to have. In the first kind of case, the synthesized performance is produced with the aim of deceiving the audience into thinking that a piano has been used. Clearly, in this case, if we discovered the deception we would, indeed, regard the performer as having cheated; but, equally, the source of our resentment would not be that the performer had misrepresented the piece by playing it on a synthesizer, but that the performer had, in effect, lied to us about what she was doing. In the second kind of case, by contrast, the audience is well aware that the performer is using a synthesizer, but

38

julian dodd

is also aware that such an instrument is easier to play than the piano, thus making the performer’s task less demanding than is usual. Here too we might regard the performer as having cheated, but, once more, our sense of having been short-changed does not have its origin in such a performance’s being inauthentic; this time, the audience would feel cheated because the performer had not been tested to the degree that other performers of the piece, using conventional instrumentation, have been. In this instance, the performer’s decision to use a synthesizer does, indeed, ‘undermine the playing’s status as a performance’ (Davies 2001: 65); but, once more, this is not because such a performance lacks authenticity, but because the performer’s feat is less impressive than it would have been had traditional instrumentation been employed. So my point is this: although we could feel cheated by synthesized performances, the fact that this is so is explicable in ways that do not entail the truth of instrumentalism. At this juncture, Davies may seek to reply by insisting that there is a possible case of our feeling cheated by a synthesized performance that I have not considered, and that does entail the falsehood of timbral sonicism. He might, in other words, claim that we would still feel cheated by a synthesized sonic doppelg¨anger even if we were aware of the sounds’ causal origin and even if such an instrument were as difficult to play as a piano. But I simply deny that we would feel in any sense cheated by such a performance. The performance would be a sound-sequence indistinguishable from a performance using a piano; and it would make demands of the performer that are on a par with the demands faced by pianists. What would there be for an audience to feel resentful about? Short of simply begging the question, I cannot see a reply for Davies here. There is no case in which our sense of having been cheated by a synthesized performance tells against timbral sonicism.

5 So much for variants of the argument from scores. Such arguments by no means force us to admit that performance means-properties are normative within works, and hence do nothing to undermine the timbral sonicist’s identity criterion. The timbral sonicist is not yet home and dry, however. For as I noted in §3, an instrumentalist may also argue for her position by

sounds, instruments, and works of music

39

claiming that a performance’s transmission of a work’s aesthetic content and artistic import requires not only that it sound a certain way, but that the sounds be produced in a certain way. Arguing along these lines, Levinson claims that a work’s aesthetic content and artistic properties are ‘determined not only by its sound structure ... but also in part by the actual means of production chosen for making that structure audible’ (Levinson 1980: 76). According to Levinson, the kind of thinking about the determination of aesthetic content that prompted my move from pure to timbral sonicism ultimately commits us to instrumentalism; and so, if he is right, the line cannot be held where the timbral sonicist wishes to hold it. A consideration of what is required of a performance, if it is to transmit a work’s aesthetic and artistic content, reveals that performance means-properties—and not just acoustic properties—are normative within works. And this can only mean that—contrary to the sonicist’s identity criterion—there can be numerically distinct works that are sonically indistinguishable. Some examples should help Levinson’s style of argumentation to stick. First, consider Paganini’s Caprice Op. 1, No. 17. This piece, Levinson claims, has the artistic property of being virtuosic and, hence, a correct performance of it should have a virtuosic quality too. But, argues Levinson, such a performance could only have this quality, if it made use of a violin: 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. (Levinson 1980: 77)

For Levinson, then, the Caprice Op. 1, No. 17 is virtuosic, but a performance could only transmit this artistic quality, if the sounds were actually produced by a violin rather than by some kind of violin-substitute. As for artistic properties, so for aesthetic properties. For, according to Levinson: [t]he aesthetic qualities of the Hammerklavier Sonata depend in part upon the strain that its sound structure imposes on the sonic capabilities of the piano; if we are not hearing its sound structure as produced by a piano, then we are not sensing this strain, and thus our assessment of aesthetic content is altered. (Levinson 1980: 76–7)

40

julian dodd

And the same goes, if Levinson is right, for works’ expressive content too. Mozart’s Serenade in E Flat, K. 375 begins with a passage which has an ‘assertive, attention-getting quality’ (Levinson 1990b: 396); and the reason why the piece’s opening has this expressive content is that the sounds produced by the winds have a honking aspect: that is to say, air is forced through narrow openings in tubes in the same way in which a goose—a literal honker—forces air through its windpipe (Levinson 1990b: 396). Now, according to Levinson, in order to appreciate the passage’s assertive content, the audience must hear the sounds as honkings; but they can only do the latter, if they hear these sounds as produced in the way in which honkings are produced; which is to admit that the passage’s expressive content is dependent upon its being played by wind instruments and not, for example, by Perfect Timbral Synthesizers: Would those sounds—i.e. those produced on the Perfect Timbral Synthesizer—be a honking in the quasi-literal sense in which the accented outputs of oboes and clarinets are such? I suggest not. If it is aesthetically appropriate to take sounds in performance for what they are, and not for what they aren’t, and if the opening’s particular assertiveness is partly a matter of its properly coming across as a honking, in the sense described, then part of the aesthetic character of that opening is distorted or undercut if the performing means that Mozart directly specifies are bypassed and only the resulting sound (or timbral complex) that he indirectly specifies is adhered to. (Levinson 1990b: 396–7)

So, if Levinson is to be believed, it turns out that being performed using wind instruments (and not simply being performed by means that exactly mimic the sounds made by wind instruments) is normative within the piece because a performance on Perfect Timbral Synthesizers would fail to produce genuine honking sounds at the piece’s beginning, and hence would fail to deliver the assertive content Mozart intended. Needless to say, further examples of this kind are easily constructed. The appreciation of a piece’s expressive content, so it is claimed, requires us to hear passages as making musical gestures akin to human expressions of emotion: we must hear phrases as sighing, crashing, booming, or such like (Levinson 1990b: 399–401). But in Levinson’s view, in order to hear gestures in music, one must hear the sounds produced as made with the relevant gestures; and this, in turn, requires that the sounds

sounds, instruments, and works of music

41

really are produced by means of an instrument with which such gestures can be made.¹² Hence, if a work is to be performed properly, such a performance must not merely perfectly ape the sounds made by the specified instrumentation; it must use those very instruments to produce those sounds. So how should a timbral sonicist reply? Two points need to be made at this stage. First, and contrary to the spin that Levinson himself occasionally places on these examples, they cannot of themselves demonstrate a work’s instrumentation to be essential to it. Even if it is shown, for example, that a performance of Paganini’s Caprice No. 17 can only transmit the work’s aesthetic content, if its constituent sounds are produced on a violin, it would only follow that the work essentially involves the violin, if the work’s aesthetic content were essential to it, and this is a thesis that Levinson does not defend (1990a: 221–2). At best, then, the examples can only demonstrate that performance means-properties are normative within works, not that they are so normative in each possible world in which the works in question exist. Having made this point, such examples are still a threat to the sonicist’s account of the identity conditions of musical works, of course. The sonicist’s distinctive claim is that acoustic properties alone are normative within works, so she repudiates the suggestion that a properly formed performance of the Caprice, for example, must make use of a violin. However, and this is my second point, Levinson’s examples do not even show this much, as we shall now see. To begin with, it would be as well to focus on Levinson’s argumentative strategy. Levinson takes his discussion of his various examples to demonstrate the truth of the following claim: that it is a necessary condition of a work W ’s having certain aesthetic, expressive or artistic properties that certain performance means-properties are normative within W . In other words, he points to certain features that works are taken to have indisputably, and then claims that works can only have these features, if instrumental properties are normative within such works. However, in what follows I argue that Levinson’s examples are insufficient to justify ¹² One example of this putative phenomenon should suffice. Drums can contribute a pounding or battering content to passages of music; and one might be tempted to follow Levinson (1990b: 400) in supposing that this expressive content can only be fully transmitted if the sounds are actually made with a pounding or battering gesture.

42

julian dodd

this claim. To this end, my strategy will be disjunctive. First of all, I shall argue that the kinds of artistic properties to which Levinson appeals are not genuinely possessed by works at all. When it comes to artistic properties, I thus deny the thesis whose presumed truth supposedly requires the truth of instrumentalism. Aesthetic and expressive properties, on the other hand, certainly are genuinely possessed by works of music; but in this case, my complaint against Levinson is that his examples do not demonstrate that a work’s having such a property is genuinely contingent upon instrumental (as opposed to merely timbral) properties being normative within it. Let us start by considering artistic properties and, in particular, by returning to Levinson’s own example. Paganini’s Caprice Op. 1, No. 17, he says, deserves the attribution ‘virtuosic’, and, hence, a performance—if it is do justice to the piece’s virtuosic quality—must make use of the violin rather than, say, an acoustically indistinguishable instrument that, perhaps, demanded less of a performer. But at this point, I think that the starting point of Levinson’s discussion of this example—namely, his presumption that the work literally possesses the property being virtuosic —will simply be denied by a sonicist. True enough, a performance may be virtuosic (if the performer plays a violin rather than a space-age sound-alike that is easier to play); but a sonicist will not accept that being virtuosic is normative within the work: i.e. that any well formed performance must be virtuosic. For, since the sonicist takes the Caprice to be a pure type of sound-event—a type whose normative properties are all acoustic in nature—she precisely repudiates the suggestion that a property concerning how sounds are produced, as opposed to their qualitative nature, can be normative within it. Of course, if it were true that the piece itself were virtuosic, it would follow that it could only be played on an instrument that required technical excellence of the performer. But it is the antecedent of this conditional that no self-respecting sonicist will accept. It is performances of works that can be virtuosic, not the works themselves. A similar response can be made to other putative examples of how a work’s artistic properties determine performance means-properties as normative. Levinson asks us to imagine a piece written for violin to be played in such a way as to make certain passages sound as if they were played by a flute. Would not this piece, by contrast with a sonically indistinguishable piece written for violin and flute, be original and unusual? And would not this demonstrate that a correct performance of the former

sounds, instruments, and works of music

43

piece—one that preserved its artistic import—would have to be played on the violin alone (Levinson 1980: 77–8)? But to this we can make what is essentially the same reply as before, albeit with a slightly different twist. Once again, Levinson’s mistake is to suppose that a sonicist will agree that works themselves possess artistic properties such as unusualness and originality. Since the sonicist takes works of music to be pure types of sound-event, she denies that they can be analytically tied to instruments, and hence denies that such works can themselves be unusual by virtue of involving surprising instrumentation. By contrast, a sonicist will insist that it is composers and their compositional acts that, strictly speaking, have the artistic properties concerned. What is unusual and original in Levinson’s example, is not the work itself, but a particular composer’s scoring of such a work for solo violin rather than violin and flute. The work itself is a pure sound event-type, and hence cannot have properties such as these. One thing is clear, however: the same sonicist strategy will not work for examples taken to illustrate the fact that performance means-properties are normative within works by virtue of determining such works’ aesthetic or expressive qualities. Works certainly can be sublime, craggy and assertive, so a timbral sonicist cannot reply by denying this. Nevertheless, replies are available, and we shall now see that Levinson is mistaken in thinking that a work’s having such an aesthetic or expressive quality may be dependent upon performance means-properties being normative within it. Consider, once more, the Hammerklavier’s sublimity and cragginess. It is Levinson’s contention that unless we hear the piece as played on the piano, we will have no sense of the strain that the piece places on the instrument, and hence will fail to pick up on its true awesomeness. In this sense, according to Levinson, ‘[t]he aesthetic qualities of the Hammerklavier sonata depend in part on the strain that its structure imposes on the sonic capabilities of the piano’ (1980: 76–7). But this conclusion is resistible. Whilst it is true that we must hear a performance of the piece as played on the piano, if we are to grasp the full gamut of its aesthetic qualities, it does not follow that a properly formed performance must actually make use of the piano. A truth about aspect perception is surely this: to see or hear o as (an) F does not entail that o really is (an) F. I can see a smile as friendly, and yet the smile be contemptuous; I can hear what is really a cry of surprise as a cry of pain. So, taking her cue from facts such as these, a timbral sonicist will insist that all that is required for a performance of the

44 julian dodd Hammerklavier to transmit the piece’s cragginess and awesomeness is that its constituent sounds sound as if they originate from a piano. As long as the sounds produced sound just like those that a piano would produce, it is of no consequence whether they are actually produced by a piano or not. All that is required is that sounds seem to have, rather than actually have, that origin. Exactly the same response is available in the case of works’ expressive qualities. Let us return to the assertive character of the opening of Mozart’s Serenade in E Flat, K. 375. Levinson’s claim, remember, is that this passage would not have this content, were it not for the fact that it has a honking quality; and that the sounds would not have this honking quality—i.e. would not be a honking in the required ‘quasi-literal sense’—unless they were actually made by oboes and clarinets (Levinson 1990b: 396–7). However, this second claim is false since a sound can have a honking quality and yet not be a literal—or even quasi-literal—honking. Just as long as a sound seems for all the world to be a genuine honking sound, it has a honking quality; it is not necessary for the sound to be the result of air having been forced through tubes. And such a reply to this case yields a generalizable moral. We may grant Levinson the fact that understanding a piece’s expressive content requires us to hear gesture in music (1990b: 398); but we should deny that ‘expressive content in music is not detachable from the means of performance that are written into musical compositions’ (1990b: 399). As long as we hear the music as embodying a certain gesture, it does not matter whether the sounds have actually been produced with that gesture. So, for example, if we hear a run of tones as a genuine glissando on the piano, it will succeed in conveying the insouciance associated with such a flicking gesture, even if the sounds are produced by pressing buttons on a Perfect Timbral Synthesizer. The music’s insouciant quality requires only that we imaginatively construe the sounds to have been produced by a flicking or sweeping gesture; these sounds need not have been genuinely produced by a glissando-proper. Levinson’s reply to such a sonicist move ultimately cuts no ice. Envisaging a case in which a sound sequence exactly duplicates a glissando but is produced by other means of performance, he says this: If this imaginative construal [of the sounds as having been produced by a glissando] is in fact unsupported by performance in the appropriate manner, that is, on a

sounds, instruments, and works of music

45

keyboard, and we are aware of this, the resultant effect is not that proper to the music, but rather some degree of cognitive dissonance. (Levinson 1990b: 399)

But the crucial clause here is ‘and we are aware of this’. Once this clause is inserted, the sonicist can acknowledge that such cognitive dissonance may occur. Granted, if the sounds are glissando-like, and yet we know them not to have been produced by such a gesture, the music might —though need not¹³—fail to have the abandon for us that it would have had otherwise. But such a difference in perceived expressiveness admits of an explanation that avoids any commitment to instrumentalism. Rather than being explained by the fact that the sounds are not produced by a glissando, an audience’s failure to hear the passage as expressive of abandonment is caused by a failure of imagination on their part: a failure to hear the sounds as if made by a glissando. Putting things this way, sonicism is unscathed. For the fact remains that a sound sequence produced on a Perfect Timbral Synthesizer could be heard as a glissando, and hence could have the same expressive content as a sonically identical sequence produced in the characteristic way. All Levinson has shown is that we might fail to hear a sound sequence as making a certain gesture, if we know it not to have been produced by means of such a gesture; and this need not be denied by a sonicist. Ultimately, then, Levinson faces a dilemma. If, in the glissando example and others like it, he argues that the music would fail to have the appropriate expressive content, if we knew it to be produced by nonstandard performance means, the sonicist can grant such a possibility without damaging her position. If, on the other hand, Levinson claims that the music would fail to have the appropriate content even if we took the sounds to have been produced by a genuine glissando, this claim would seem to be false. For as we have seen already, sounds can be heard as embodying a certain gesture, and hence have a certain expressive content, even if they are not made with that gesture. All that is required is that the sounds are qualitatively indistinguishable from sounds produced with that gesture. Levinson has failed to make a case for the claim that the performers ¹³ This qualification is crucial. Levinson seems to regard it as being extremely difficult for an audience to imagine sounds to have a certain source whilst knowing these sounds to have another origin (1990b: 403). As we shall see in §6, Levinson overestimates the difficulty in what is an everyday phenomenon.

46

julian dodd

must actually be doing what we imagine them to be doing, if we are to hear such gestures in the sounds made.

6 As we have just seen, and as Levinson himself recognizes (1990b: 402–3), the sonicist responds to his argument by claiming that a performance of a piece can transmit its aesthetic and expressive content, just as long as the sounds produced are imagined to be produced in certain specific ways. To insist that the sounds should be so produced is to insist on more than suffices. Levinson, though, is dissatisfied with this reply for three reasons. First, he claims that its cogency depends upon the audience’s not learning the real origin of the sounds they are hearing, or else being willing to be fooled concerning their source, ‘neither of which is a viable stance vis-`a-vis a performance’ (1990b: 403). Second, he states that a sonic doppelg¨anger of, say, a glissando will present merely a simulacrum of the expressive gestures that belong to the passage, rather than the gestures that really belong to it (1990b: 404). Finally, he contends that such a sonic doppelg¨anger cannot count as authentic because it would effectively stymie our ability to compare and evaluate performances (1990b: 404–5). I shall consider these objections in reverse order. Let us, then, start by examining Levinson’s third objection. A performance of the Hammerklavier on a Perfect Timbral Synthesizer, even if indistinguishable from a properly formed performance on a piano, would be inauthentic, Levinson claims, because it would present us with a severe critical quandary. In short, we would not know how to evaluate it. As Levinson puts it, if a piano were not involved: we would be pretty much completely at sea in regard to assessing the particular expressiveness of the performance ... , its particular manner of bodying forth [the work’s] inherent expression. For an important dimension of assessment would have been removed: how have the instrumentalists, given their control over and way of internalising the gestural capacities of their instruments, related themselves, at each turn, to the demands of this music, which is conceived for and referred to those capacities? (Levinson 1990b: 405)

But inasmuch as there is a genuine worry here, it would seem to be irrelevant to the point at issue, which is, remember, whether a performance

sounds, instruments, and works of music

47

of the work on a Perfect Timbral Synthesizer could transmit the full range of the piece’s aesthetic content. It may well be difficult to evaluate such a performer’s achievement in transmitting such content, given that we tend to make such evaluations according to how well the performer has overcome the limitations of the instruments for which the work was written. It would, indeed, be easier said than done to compare the achievements of an Ashkenazy with those of someone who produced a qualitatively indistinguishable sound sequence on a synthesizer: we are accustomed to the problems posed by playing the piece on a piano—our familiarity with both the instrument itself and the attempts made by pianists down the years give us a reasonably clear sense of this—but until we had a clear idea of the problems and constraints faced by an operator of a Perfect Timbral Synthesizer, any such comparison with performances by pianists would be obstructed. Having said this, it is far from clear why the existence of such a puzzle would demonstrate the synthesized performance to be inauthentic, i.e. to fail to present the full extent of the performed work’s aesthetic content. For the puzzle to which Levinson draws our attention in fact concerns how we would evaluate a performer’s achievement in transmitting a work’s aesthetic content; it is irrelevant to the question of whether a synthesized performance could deliver this content. And, of course, to this latter question, the timbral sonicist has a clear answer: such content is determined wholly by how a performance sounds, and to no extent by how these sounds are produced. What of Levinson’s second objection to the sonicist’s reply to his argument? Must a synthesized performance inevitably fail to convey a work’s aesthetic substance? I don’t think so. Naturally, a timbral sonicist agrees with Levinson that ‘an authentic performance should seek to present a piece’s aesthetic substance as it is’ (1990b: 404). But her claim is precisely that this aesthetic substance is not dependent upon specific means of performance in the way that Levinson supposes: as long as a sound sequence sounds exactly like a glissando on a piano, so that the audience imagines the sounds to have been produced with a sweeping gesture along a piano keyboard, the resultant sounds will have exactly the same aesthetic substance as a genuine glissando. The glissando-like sounds express the sweeping gesture, and hence give the music an insouciant quality, even though they were not produced with such a gesture. Aesthetic substance is determined by how the music sounds, not how it is produced. To put it another way, the vehicle

48

julian dodd

of a work’s musical meaning is not, as Levinson assumes (1990b: 404), sounds-as-produced-by-specific-instruments, but those sounds themselves, whatever their origin. His claim that a sonic doppelg¨anger of a glissando ‘gives the appearance, but not the reality of the complex of soundings that are the true vehicle of a piece’s musical meaning’ (1990b: 404) sees him, once more, beg the question against the timbral sonicist. Nonetheless, Levinson will insist that a synthesized sonic doppelg¨anger—even if it manages to convey the right expressivity to an audience—inevitably ‘falsifies or disguises the basis for that expressivity’ (1990b: 404), and, as a consequence, cannot be authentic. And this, in essence, is his one remaining objection to the kind of sonicist response to his original argument that I outlined at the end of §5. This objection emerges once the following kind of case is considered. Imagine that an audience attends a performance of Mozart’s Serenade in E Flat, K. 375, well aware that it is performed using a Perfect Timbral Synthesizer rather than wind instruments. As a result, in order to hear the assertive quality of the piece’s opening passage, an audience must hear the sounds produced as a sequence of honkings whilst at the same time knowing the sounds not to be honkings, i.e. not to have been produced by forcing air through narrow holes in tubes. But this, Levinson suggests, is none too easy a thing to do: we must imagine the sounds to be honkings while knowing them not to be (1990b: 403); that is, we must allow ourselves to be fooled concerning the sounds’ origin. And, in any case, even if an audience could succeed in doing this, a performance that required such a thing of its auditors could not be counted as authentic, and hence could not be properly formed: A performance that enforces mental acrobatics on listeners in order that an intended expressiveness should emerge, which expressiveness should emerge effortlessly and unconsciously, can hardly be thought to further authenticity. On the contrary, to ensure that a listener’s experience of a performance be informed by the thought of certain performing actions in the right way, the performance should actually involve those very actions and the very instruments that make them possible. (Levinson 1990b: 403–4)

But I have two replies to this reasoning. First of all, an audience’s hearing sounds it knows to be synthesized sounds as honkings is not thereby allowing itself to be fooled as to the sounds’ origin. On the contrary, the audience is imagining these sounds to have an alternative source to that which it knows

sounds, instruments, and works of music

49

them to have actually. The audience knows the sounds to be synthesized but imagines them to be produced by clarinets and oboes. And before we are tempted to look askance at such a phenomenon, we should remember that imaginative engagement of this kind is not unfamiliar to audiences of the performing arts: the way in which an audience imaginatively engages with a play—imagining that the people on stage are in real situations, but knowing the participants to be actors—is another example of the same genus. Second, such a reflective exercise of the imagination is not an instance of ‘mental acrobatics’. As I have just explained, knowing that p but imagining that not-p is an everyday phenomenon par excellence. True enough, for the expressiveness of the music to emerge in a synthesized performance, the audience must employ greater imagination than if the sounds are produced by the wind instruments that the composer had in mind. But the exercise of such a capacity is both effortless and perhaps even unconscious: one just sits back in one’s seat, shuts one’s eyes, and lets the sounds, what one knows about such sounds’ customary origin, and one’s imagination do the rest. I fail to see why a synthesized performance of a piece that required this of its audience would not be authentic. After all, this requirement, so easy to meet, ensures that such a performance would succeed in transmitting the piece’s aesthetic content.

7 Given timbral sonicism’s status as the face-value theory of how musical works are individuated, we should only give it up once it has been shown to be untenable. But there is nothing in the instrumentalist’s arsenal of arguments that forces us to take such drastic action. The fact that a work’s score instructs performers to use certain instruments does not entail that there could not be a properly formed performance of the piece produced by other means. As long as these other means of sound production were used to produce sounds aurally indistinguishable from sounds produced by the instruments specified in the score, the demands of authenticity would be satisfactorily met. Equally, the attempt to argue that a work’s possession of certain aesthetic and artistic properties is dependent upon its having performance-means properties normative within it falls short of its target. All that is required for works to have the content they have is that they sound as they do. How these sounds are produced is incidental.

50

julian dodd

So can we now embrace timbral sonicism—and, with it, the simple view—with a clear conscience? Not just yet. Some philosophers take the sonicist version of the type/token theory to have Platonic consequences that require us to identify musical works with types of a more complex kind.¹⁴ Others, self-styled contextualists, argue that sonicism should be rejected because composers occupying differing musico-historical contexts inevitably compose distinct works, even if the said works are acoustically indistinguishable.¹⁵ Finally, the other component of the simple view—the type/token theory—by no means commands universal acceptance. Works of music have been claimed to be continuants (Rohrbaugh 2003), actiontypes (Currie 1988), and even action-tokens (Davies 2004). Ultimately, I doubt whether the arguments for these various positions add up, but making the case for this claim is beyond the scope of this paper.¹⁶

References

Currie, G. (1989) An Ontology of Art, Basingstoke: Macmillan. Davies, D. (2004) Art as Performance, Oxford: Blackwell. Davies, S. (2001) Musical Works and Performances: A Philosophical Exploration, Oxford: Oxford University Press. Dodd, J. (2000) ‘Musical Works as Eternal Types’, British Journal of Aesthetics 40: 424–40. Dodd. J. (2002) ‘Defending Musical Platonism’, British Journal of Aesthetics 42: 380–402. Dodd. J (2004) ‘Types, Continuants and the Ontology of Music’, British Journal of Aesthetics 44: 342–60. Dodd, J. (2007) Works of Music: An Essay in Ontology, Oxford: Oxford University Press. Godlovitch, S. (1998) Musical Performance: A Philosophical Study, London: Routledge. ¹⁴ See, for example, Levinson’s justly celebrated construal of musical works as indicated types (1980, 1990a). ¹⁵ Levinson, once more, is in the vanguard of this camp (1980, 1990a), though versions of this view have become more widely accepted than his own positive ontological proposal. See, for example, Currie (1989) and Davies (2001). ¹⁶ An early version of this paper was given at the British Society of Aesthetics conference in Oxford in September, 2004. Many thanks to all those who attended. Special thanks are due to David Davies and Kathleen Stock, who gave me extremely helpful written comments.

sounds, instruments, and works of music

51

Kivy, P. (1983) ‘Platonism in Music: A Kind of Defence’, Grazer Philosophische Studien 19: 109–29. Kivy, P. (1987) ‘Platonism in Music: Another Kind of Defence’, American Philosophical Quarterly 24: 245–52. Kivy, P. (1988) ‘Orchestrating Platonism’, reprinted in his 1993, 75–94. Kivy, P. (1993) The Fine Art of Repetition, Cambridge: Cambridge University Press. Levinson, J. (1980), ‘What a Musical Work Is’, reprinted in his 1990c, 63–88. Levinson, J (1990a) ‘What a Musical Work Is, Again’, reprinted in his 1990c, 215–66. Levinson, J (1990b) ‘Authentic Performance and Performance Means’, reprinted in his 1990c, 393–408. Levinson, J (1990c) Music, Art and Metaphysics: Essays in Philosophical Aesthetics, Ithaca, NY: Cornell University Press. Rohrbaugh, G. (2003) ‘Artworks as Historical Individuals’, European Journal of Philosophy 11: 177–205. Scruton, R. (1997) The Aesthetics of Music, Oxford: Oxford University Press. Walton, K. (1988) ‘The Presentation and Portrayal of Sound Patterns’, in Jonathan Dancy, J. M. E. Moravcsik and C. C. W. Taylor (eds), Human Agency: Language, Duty and Value, Stanford: Stanford University Press, 237–57. Wolterstorff, N. (1980) Works and Worlds of Art, Oxford: Clarendon Press. Further Reading

When it comes to the ontological nature of works of music, competing versions of the type/token theory are propounded by Levinson (1980, 1990b) and Kivy (1983, 1987) respectively. I take Kivy’s Platonist side in my 2000, 2002 and 2007. Alternatives to the type/token theory are suggested by Currie (1989), David Davies (2004) and Rohrbaugh (2003). The second aspect of the simple view—sonicism—is criticized by Levinson in his 1990b, and by Stephen Davies in his 2001: 60–71. Kivy responds to such arguments in his 1988.

2 Doing Justice to Musical Works M I C H A E L M O RRI S

1. Introduction There is something puzzling about musical works. The puzzlement is ontological in a broad sense: we find it difficult to understand what kind of thing musical works are. We note, for example, that a musical work cannot be stored in a cellar,¹ though copies of a score can be. Musical works can be performed on several occasions, though they can exist unperformed. Some musical works are written, but not all are. The leading response to this ontological puzzlement is to attempt to remove it by assimilating musical works to other, supposedly more familiar, kinds of thing. I shall call this the assimilating response, and those who adopt it assimilators. The dominant version of the assimilating response aims to remove our puzzlement by denying that musical works are particulars. Instead, they are types (Wollheim 1980, §35; Dodd 2000) or kinds (Wolterstorff 1975) or universals.² If they are types or kinds, what are they types or kinds of? If they are universals, what are their instances?³ There are more and less reductive answers to these questions within the dominant tradition. According to the more reductive answers, musical works are types or kinds of sound sequence; or else they have particular sequences of sounds as their instances. According to the less reductive, musical works ¹ Compare Heidegger: ‘Beethoven’s quartets lie in the storerooms of the publishing house like potatoes in cellar’ (Heidegger 1956: 145). But this looks like a rare case where Heidegger seems to accept uncritically the kind of ontology which elsewhere (including elsewhere in this work) he is concerned to make us aware of. ² Levinson (1980 and 1990b) is standardly cited as someone who holds that musical works are types, but there are features of his view which make that classification a little complicated. ³ Although it is often uncritically said that musical works have instances, we should recall that the word ‘instance’ is correlative with ‘universal’ or ‘property’, just as ‘token’ is correlative with ‘type’.

doing justice to musical works 53 are types or kinds of performance; or else they have particular performances as their instances. The assimilating response is motivated by a certain ontological conservatism. It takes certain kinds of things—generally things with clear spatio-temporal boundaries—to be familiar and unpuzzling; and it aims to acknowledge the existence of as few things other than these supposedly familiar items as possible. This general ontological conservatism is applied at two points to produce the view that musical works are universals, types, or kinds. In the first place, it allows us to acknowledge only those particulars which have relatively clear spatio-temporal boundaries: the relevant particulars here are therefore sound sequences or performances. And then it lets us accept just those non-particular things which we have general theoretical reasons for acknowledging. At this point, assimilators see themselves as having to argue for the acknowledgement of universals, types or kinds, against even more parsimonious views, which only accept classes in addition to spatio-temporally well-defined particulars (see Goodman 1981). I shall argue that this approach is altogether wrong. Musical works are not types or kinds or universals, and the assimilating response in general is the wrong response to our puzzlement about the nature of musical works. I will take as my principal target the specific version of the common view which holds that they are types of performance. I choose the idea that they are types (rather than kinds or universals) because that seems to be the best worked out. And I choose the idea that they are types of performance (rather than sound sequence) because that is less reductionist, and the basic objection I will make to the general view is, in effect, that it is impossibly reductionist. This is how I’ll proceed. First, I’ll lay out what seems to me to be the basic argument for the common view. My objection to this view depends on claiming that it cannot accommodate the fact that musical works, like all works of art, are essentially meaningful, so next I’ll explain that. I will then argue that there are two ways in which the common view cannot accommodate the meaningfulness of works of art. Finally, I will look again at the basic argument for the common view, in order to see where it goes wrong, and I will say something about how we should respond to our puzzlement about the nature of musical works.

54

michael morris

2. The Basic Argument for the Type-Token View I will call the view that musical works are types, of which their performances are tokens, the type-token view. The basic argument for it is that it provides the best explanation of the following facts: (i) musical works can be heard; (ii) the existence of a musical work does not depend on any particular performance; (iii) the same musical work can be performed many times. Fact (i) seems undeniable. We might doubt the truth of (ii), if there are any performances which can be regarded as musical works in their own right, or of (ii) and (iii), if there are any musical works which are essentially improvisations, but (ii) and (iii) remain plausible for the large bulk of what we ordinarily regard as musical works, and might be thought to be essential to anything which can properly count as being both musical and a work.⁴ Fact (i) requires an intimate relation between musical works and performance, because it is only in being performed that a musical work can be heard. But that relation cannot be identity with any particular performance, because of facts (ii) and (iii). Facts (ii) and (iii) make it difficult to regard musical works as spatio-temporally located objects of any familiar kind, in which case it is natural to describe them as ‘abstract’.⁵ There is a natural aversion (among many philosophers, including assimilators) to admitting the existence of abstract particulars, so facts (ii) and (iii) encourage us to deny that musical works are particulars. Once we deny that musical works are particulars, it is natural to suggest that they are types. And when we suggest that they are types of performance, we seem to be able to explain how they can be heard—by being tokened in a particular performance; how their existence is independent of any particular performance—because the existence of a type is independent of the existence of any particular one of its tokens; and how the same musical work can be performed many times—because it is of the nature of types to be capable of having many tokens. And the explanation is ontologically conservative, which is always ⁴ In fact, I think this thought is wrong, since I think that performances of musical works are themselves both musical (with luck) and works; but there is no inconsistency in making the performable work, and not the performances of it, the centre of our interest here. ⁵ Note that ‘abstract’ is not an innocent word: abstraction is most famous as a theory of generality.

doing justice to musical works 55 a virtue in the eyes of an assimilator: we do not need to introduce any radically new categories specially for the case of musical works, and we do not need to admit any surprising combinations (such as abstract particulars). Types are individuated by the conditions which things have to meet to count as tokens of them. The claim that musical works are types would be empty (at best), if no such condition could be found to unify the things which count as performances of a given work. What might such a condition be? There is just one natural candidate. In order for a musical work to be capable of being performed repeatedly, it must in some sense bring with it a prescription which indicates how it is to be performed. The prescription can be more or less precise in various ways: the notes and their values can be prescribed more or less precisely, as can tempi and instrumentation—and so on. Let me gather together all the various respects in which ways of performing a work might be prescribed under the crude phrase the notes. Then we can say that a performance is note-perfect if it meets all the conditions (however precise) which are prescribed. It seems that the very fact that a work can be performed repeatedly requires there to be such a thing (as it were) as a note-perfect performance. The concept of a note-perfect performance can be used in one of two ways to define the condition which performances must meet in order to count as performances of the same work. First, the type might be defined non-normatively, simply as the type of note-perfect performances: this is a non-extensionalist version of Goodman’s famous proposal (Goodman 1981: 186). If we follow this line, it will turn out that only those performances which are note-perfect are really performances of the relevant work. That will have the surprising consequence that some familiar, but technically difficult works may never actually have been performed. We could just brazen this out—claiming, perhaps (with an eye to fact (i)), that such works have never really been heard—but we are unlikely to win many people over. The other alternative is to define the type normatively, as the type of performances for which being note-perfect is a requirement or an ideal. (This is a variant of the view proposed in Wolterstorff (1975 and 1980).) On this view, to be a performance of a given work is to be a performance of which it can strictly be said either that it ought to be note-perfect or else that if it is not note-perfect that is some kind of defect in it. This is surely the natural way for someone to flesh out the type-token view.

56

michael morris

We seem in this way to be able to substantiate the claim that musical works are types of performance. We have also—apparently incidentally, though we may claim it as a virtue of the view—found ourselves endorsing a particular approach to the criticism of performances. We will be tempted to think that good performances will be at least close to note-perfect; if we are inclined to include instrumentation, tempi and pitch among the things which are prescribed of a note-perfect performance—which certainly seems natural in the case of many musical works—we will find ourselves giving high value to authenticity in performance. (This line is most explicitly adopted by Davies (2001: Ch. 5); Levinson (1990a) clearly values authenticity, though he does allow that a good performance may not be authentic in Levinson (1987).) I think this is the principal argument for the type-token view (see Dodd 2004). It can be adapted, mutatis mutandis, to suit any version of the view that musical works are not particulars but in some way general. All versions of the view, however, are subject to the same basic problem: they cannot properly accommodate the fact that musical works, like all works of art, are essentially meaningful.

3. Musical Works are Essentially Meaningful Some things are true of musical works just in virtue of their being works of art. Among these is the claim that musical works are meaningful: this ought to be a truism, since it seems bound to be a truism that works of art are meaningful. I will offer a brief consideration in support of the latter truism, before dealing briefly with a simple objection. First of all, it is clear that works of art are meaningful because it is clear that one cannot properly consider a work of art as a work of art without considering it as meaningful. To treat a work of art as not being meaningful would be to treat it as some kind of found object, a curious feature of the natural world, and this is precisely not to treat it as a work of art. When we consider works of art as works of art, we regard them as things which call for interpretation—or at the very least, as things of which interpretation is legitimate—and an interpretation is a presentation of meaning. When we reflect on this, we can see that no objection to the claim that works of art are meaningful can be provided by cases where we are inclined to say that

doing justice to musical works 57 what is presented as a work of art is meaningless. For example, someone might declare, in a reactionary spirit, that a painting was a ‘meaningless smear’, or a piece of music was nothing but ‘meaningless noise’. But, first, the very fact that what we take to be a meaningless smear is exhibited as a painting, or that what we take to be meaningless noise is presented as a piece of music, seems to give meaning to what might antecedently have been meaningless. And, second, even if we finally conclude that what we have before us really is nothing but a meaningless smear or meaningless noise, all this will mean is that what we have before us is not really a work of art. So the examples cannot threaten the claim that works of art must be meaningful. We might feel some unease, however, because of a certain kind of preconception about the notion of meaning. The problem can be expressed quite roughly and intuitively as follows: nothing could be meaningful without having a meaning, but any attempt to say what a work of art means seems inevitably to trivialize it; art always seems to transcend any interpretation. There is indeed a strong temptation to associate the notion of meaning with the possibility of a complete, explicit, and decisive interpretation, but we need to be careful about the argument involved here. It has two crucial premises: (1) (2)

If something is meaningful, it is possible to say what it means. To say what a work of art means is to trivialize it.

Neither of these premises is obviously true; (1) seems to depend on a significant claim about language, which we might express like this: (1a)

Everything is describable.

(After all, if the strong general claim of (1a) were not true, why should one think that meaning in particular can always be specified?) And (2) depends upon a claim about works of art, which we might express like this: (2a)

Works of art transcend all interpretations.

In fact, I think both (1a) and (2a) are true. The real problem lies with something else which (2) depends on: (2b)

To say what a work of art means is to capture its meaning (finally and completely).

58

michael morris

We can understand the link with (2a) if we paraphrase (2a) as saying just this: (2a∗ ) No interpretation (no paraphrased saying, for example) can capture the meaning of a work of art (finally and completely). If this is how the argument works, there are two problems with it. First, it depends on certain assumptions ((1) and (2b)) which are fundamentally assumptions about language, rather than about either art or the general concept of meaning. And, secondly, it seems impossible to formulate the argument without presupposing that works of art are meaningful: (2a) has to be interpreted in the manner of (2a∗ ), for example. Works of art are not only meaningful, but essentially meaningful: that is to say, those things which are in fact works of art could not exist without being meaningful (or, indeed, having the meaning they have). This can be seen by considering what we naturally take to be essential to a work of art. In any work of art (including works of music), we are able to distinguish between essential and non-essential parts and features. In the case of a painting, for example, it might be essential that a brush was used in a particular way on a certain part of the canvas, but inessential that the brush in question was made by a particular person in his workshop one Monday morning. In the case of a poem, it will be essential that those very words are used, but it will probably be inessential that the poem was first written with a fountain pen. What this suggests is that the essential parts and features of a work of art include those which are relevant to its meaning, or at least central to its meaning. We will happily tolerate quite extensive variations while still regarding ourselves as having the same work of art, provided that the variations do not affect the meaning of the work, or don’t affect it centrally. (There are tricky questions about when we are to count revised editions of a musical or literary work as being different versions of the same work, and when we should think of them as different works.) Incidentally, this fits nicely with an explanation of why the meaning of a work of art cannot be captured in paraphrase. A paraphrase will always differ from the original work in a significant number of its essential features: otherwise it wouldn’t be a paraphrase. If those features are relevant to meaning, we can understand why the paraphrase cannot capture the meaning.

doing justice to musical works 59 If it is true that the essential features of a work of art include in particular those which are relevant, or central, to its meaning, it would be miraculous if works of art were not essentially meaningful. For it seems that for works of art not to be essentially meaningful, it would have to be the case that a scheme of individuation which made no reference to artistic meaning could nevertheless, quite coincidentally, select as essential parts and features of an object all those parts and features which are counted as essential when we are concerned with meaning. But the idea of such a miraculous coincidence is itself incoherent, on most conceptions of possibility and necessity, which allow that what is possible or necessary is necessarily possible or necessary. Within such a conception of possibility and necessity, the idea of this kind of miraculous coincidence between schemes of individuation becomes the idea of the mere coincidence of the essential properties identified by the different schemes across all possible worlds; and nothing could merely coincide across all possible worlds. So a natural conception of what is essential to works of art forces us to accept that works of art are essentially meaningful. And this applies to musical works just as much as to works of art of other kinds.

4. To be Meaningful is to be there to be Understood There is a familiar slogan in the philosophy of language, that a theory of meaning is a theory of understanding (see, e.g., Dummett 1975: 99). The truth at the heart of this slogan is this: the meaning of words is what is to be understood in them. The notion of meaning is correlative with the notion of understanding. We can use this point to explain what it is for a musical work (or a work of art more generally) to be meaningful. A mere correlation with understanding is not enough, however. Someone can understand how weather systems work, but this fact on its own doesn’t make weather systems meaningful; nor is there any obvious sense in which what such a person understands is the meaning of weather systems. This point seems to apply to the natural sciences in general: the understanding provided by the natural sciences is not an understanding of the meaning of their objects, so there is no direct requirement that an object of natural-scientific enquiry be meaningful. At the very least, some further argument is required to show that natural sciences presuppose the meaningfulness of their objects.

60

michael morris

What might be missing here? The obvious thought is that there is too indirect a connection between natural science and teleology. It might be that the natural sciences presuppose some teleology, general or specific. That is, it might be that the natural sciences in general presuppose the general purposive orderliness of nature; or it might be that certain natural sciences presuppose that their particular objects have purposes of some kind—perhaps the notion of a biological species presupposes the notion of a proper development of a member of a species, and the notion of a bodily organ presupposes some idea of function. But insofar as these things are studied by natural sciences, the teleology is not the object of our interest. And this seems to explain why it is that the objects of natural sciences are not thought of as meaningful by those sciences, nor is it their meaning (if they have any) which those sciences aim to understand. So how does something have to be for a correlation with understanding to make room for the notion of meaning? Here is a simple suggestion. Something is meaningful if it is, in some sense, there to be understood. The gerundive construction shows the teleology: being understood is what it is for. And the meaning of a meaningful thing is what is to be understood in it: that is to say, it is that part or aspect of the thing the understanding of which is the understanding which the thing is for. This simple suggestion seems to be confirmed by the two contrasting cases of words and the natural world. Words are there to be understood: that’s what they’re for. And the meaning of a word is what is to be understood about a word. On the other hand, consider the natural world. Many may doubt that the natural world is meaningful, but even they can surely agree with this: to regard the natural world as meaningful is to think it is there to be understood; and to look for the meaning of the natural world is to look for what it is about the natural world the understanding of which would be understanding what is there to be understood about it. This analysis of meaningfulness is confirmed by the case of works of art: works of art are there to be understood. I shall now offer an independent argument for this claim. Here are three ways of approaching what is, I think, the same feature of the concept of a work of art. First, we think that works of art demand, or deserve, a certain kind of attentiveness. In some cases (in galleries, or at concerts, for example), the careful adoption of an appropriately attentive attitude can be comic; but it nevertheless shows that at least those who

doing justice to musical works 61 so carefully adopt the attitude believe that attentiveness is demanded, and that they take themselves to be following a common assumption. Second, it often seems natural to talk about the ‘appreciation’ of art. The etymology of this word—which suggests putting the proper price on something—reveals something both inadequate and ugly about this natural talk (as if price were what matters). But the word itself, as I understand it, and the idea that appreciation is appropriate to art, both again suggest that works of art demand or deserve something. And third, to adapt a Kantian thought about morality,⁶ we want to say that to think of something as a work of art is to think of it not merely as a means. We can use a painting as a duckboard, or a sculpture as a hat-stand, but in these uses we are not treating the works as works of art. That implies that to think of something as a work of art is to think of there being something which is due to the thing itself, and not merely as a consequence of any use it might have. So for something to be a work of art is for it to be right to think that there is something which is due to the thing itself. This adaptation of what is originally a Kantian thought about moral demands makes it natural to introduce the concept of justice in explaining what is important about works of art. It is natural to say that there is such a thing as doing justice to works of art. Not only that: there is a sense in which justice to works of art is demanded of us; in some sense, it seems, we ought to do them justice. But this looks at first sight to be a very peculiar sense of ‘ought’. There is something of the precious aesthete in the attitude which takes us to have a moral obligation to attend to works of art. And on the other hand too much is lost if the imperative amounts to no more than this: ‘You ought to attend to works of art if you like that kind of thing.’ This fails to distinguish works of art, and other essentially meaningful things, from anything else. The natural solution is to understand this ‘ought’ as relating to the demands of the work itself: you ought to attend to the work, if you are to do what it asks. And this idea of the demands of the work itself is naturally explained teleologically: it is to do with what the work of art is for. What a work of art is for is being done justice to: it is there to be done justice to. This is the sense in which it asks to be done justice to. The feature of the ⁶ Note that the Kant I make use of is the Kant of the Groundwork and the second Critique, not the Kant of the third Critique. In fact, I think that the view of works of art which I arrive at is in tune with the view of the third Critique, but reaching it by means of the third Critique would have been more complicated and controversial.

62

michael morris

concept of art which we understand in those three ways I have mentioned is just this, then: to think of something as a work of art is to think of it as being there to be done justice to. From which it seems to follow that if something really is a work of art, it really is there to be done justice to. The formulation in terms of justice is derived from that adaptation of Kant’s conception of moral obligation, but I think it is also true to the experience of considering works of art. Think again of the experience of being in a gallery or at a concert, for example. It seems to me impossible to see a picture in a gallery, and understand even the tiniest part of the significance of its location there, without thinking that the picture is there (in the world, not just in the gallery) to be given its due. Similarly, it is impossible to hear a piece of music in a concert without thinking that it is there (in the world, not just in the concert) to be given a proper and just attention. It is just this kind of ‘due attention’ (Hume 1757: 232) which the comic self-styled aesthete wants to be thought to be giving when she frowns so elaborately in the gallery; and it is in order to find out what such attention should be directed at that the earnest apprentice goes to classes in musical appreciation. What is actually involved in doing justice to a musical work (or a work of art, more generally)? It is surely at least a necessary condition of doing justice to a work of art that one understand it, at least partially. Consider what a response to a work of art would have to be like, if it did not involve understanding. We would have to count it as a merely causal, or merely psychological response. Such a response is one which could, in principle, have been produced by another means—by a drug, for example. Just as it could only be a joke to say that someone did justice to a drug by succumbing to it, so it could only be a joke to say that someone did justice to a work of art by responding to it in a way which could equally well have been a response to a drug. (Compare Wittgenstein (1966: 29): ‘Would a syringe which produces these effects on you do just as well as the picture?’) An important part of denying that a work of art is to be treated merely as a means is to insist that to take it seriously as a work of art is to attend to what is peculiar to it. Nor would matters be much improved if we required merely that understanding be a precondition for any response which could count as doing justice to a work of art. This would, in effect, be to treat the work of art as like a drug which could only work if taken after a meal: it would

doing justice to musical works 63 be saying that the merely psychological response induced by a work of art is only available to someone who has understood it. On this view, it is a contingent and empirical fact that the relevant response can only be produced in someone who has understood the work in question, although it is a conceptual truth—definitional, in effect—that something only counts as a work of art if the response depends on such an empirical precondition. But it is surely not merely a contingent and empirical fact that doing justice to a particular work of art requires understanding it: that is to say, for each particular work of art, nothing could count as doing justice to it which did not involve understanding it. The proper response to a work of art must be one which is not merely preceded or accompanied by understanding: it must be a response made with understanding, an understanding response. This seems to be enough to show that if works of art are there to be done justice to, they are there to be understood. Even if more than understanding is required for one to do justice to a work of art, understanding is so intimately involved in what it is to do justice to a work of art that we can say that works of art in general, and musical works in particular, are there to be understood. (It is even tempting to suggest that, given a proper recognition of what is really involved in understanding, a work of art can ask for nothing more than to be understood: that doing justice to a work of art just is understanding it. But this is more than we need for our purposes.)

5. Musical Works must be Created If musical works are types, or any kind of non-particular, it is natural to think that they cannot be created.⁷ A type is individuated by the condition which things have to meet to count as tokens of it; it is natural to think that ⁷ This conditional is a central feature of an argument presented by Levinson (1980: 7). It is elaborated, in order to draw the conclusion that musical works are not created, by Dodd (2002); that conclusion is anticipated, e.g., by Wolterstorff (1980) and Kivy (1983 and 1987). Levinson himself uses the conditional to argue in the other direction, against the view that musical works are ‘pure sound structures’; but he thinks there are such things as ‘initiated types’: see Levinson (1980 and 1990b). I suspect that the apparent plausibility of this idea depends upon an ambiguity of construal: if we say that a work is a type of performance initiated by a composer, do we mean that the composer initiated the type (surely not a natural thing to say), or do we mean that the work is a type of performance in a tradition or practice initiated by the composer (a more natural thing to say—though, of course, I don’t think that works are types at all)?

64

michael morris

conditions exist atemporally (some say ‘eternally’); so it is natural to think that types also exist atemporally.⁸ The same applies, mutatis mutandis, to universals: for redness to exist, for example, is for there to be such a thing as being red, a condition which things have to meet to count as red. Some adopt an ‘Aristotelian’ view of universals, according to which universals cannot exist without being instantiated (see, e.g., Armstrong 1989: Ch. 5), and a similar view could in principle be adopted in the case of types. But, for one thing, being instantiated or tokened is not the same as being created; so this ‘Aristotelian’ proposal would not on its own entitle us to think that musical works are created. And, for another, it is implausible to think that written music depends for its existence on being performed. The point remains: if musical works are types, the natural thing to think is that they are not created but discovered. If this is the natural thing to think, it has a special significance in relation to the basic argument for the type-token view. For that argument works by suggesting that a certain assimilating explanation is the natural response to a puzzlement we feel about musical works. That explanation is only natural if it is itself understood in the natural way. It may be possible to understand types as created entities, but if we do this, we lose the clarity which the type-token view promised to bring to the issues. Whether the view that types are eternal represents the core of the typetoken theory, or only a dominant form of it, it faces a simple objection. This is that nothing can be meaningful, in anything like the way works of art are, without having been created.⁹ The point is intuitive enough, but is underlined by my analysis of meaningfulness as being there to be understood. Nothing can have that specific teleological purpose without having been created. It’s important to note that this does not involve denying natural teleology. It is reasonable to think, for example, that, as a matter of natural fact, hearts ⁸ Those who think it is natural to think of types as coming into existence tend to understand the notion of a type by example, rather than on the basis of any explanation or definition. Thus Wollheim (1980, §35) explains the notion of a type by means of the examples of the Red Flag and the word ‘red’. But this kind of explanation is not only unhelpful because it doesn’t actually tell us what types are; the examples are themselves all open to question (it is not obvious to me, for example, that the Red Flag and the word ‘red’ are not particulars). The moment you attempt to explain what types are, in terms of conditions which things have to meet to count as tokens of them, you immediately make it natural to take types to be eternal entities. ⁹ Note the contrast between this argument, which is in essence a transcendental argument, and the kind of appeal to intuition and common practice which is found in Levinson (1980 and 1990b).

doing justice to musical works 65 are there to pump blood. That’s what they are for, and their being for that is quite independent of the hypothesis that they have been created. The problem is rather with the specific teleology involved in works of art. Being there to be understood cannot be a natural purpose, unless we suppose that nature itself is created. In fact, considering what would be involved in thinking that the world as a whole is there to be understood lends further support to the claim that nothing can be meaningful without having been created. Of course, every natural scientist (perhaps also every artist) must think that the world is there to be understood in some sense, and this need not involve thinking that the world was created. But the sense in which the natural scientist or artist, just as such, must think of the world as being there to be understood is similar to that in which a cleaner must think of dirt as being there to be cleaned up: the attitude is induced by the job of being a natural scientist or artist—their job just is to understand the world, we might say—and is not due to the nature of the world itself (as its raison d’ˆetre, perhaps). On the other hand, to think that the world itself is there to be understood, in virtue of the nature of the world and not because of the special job of the natural scientist or artist—to think, perhaps, that it is part of the raison d’ˆetre of the world that it be understood—really does seem unintelligible without thinking of the world as created. Interestingly, the creation in question is naturally thought of as creation as by an artist: think of the God of the opening pages of Genesis who after each day steps back, as it were from an easel, and sees that it is good. Might someone respond on behalf of those who deny that works of art are created, and say that works of art themselves are not there to be understood in virtue of their own nature? Such a person might say that the idea of works of art being there to be understood is really just a projection of the demands of a particular job, just as it is only as an object of the demands of the job of being a natural scientist or an artist that we say that the natural world is there to be understood. This suggestion is barely intelligible, I think. What is the job whose business is to understand works of art, and for which works of art are there to be understood? The job of artist or critic? But these jobs themselves are only intelligible if we have an idea of what works of art are: after all, the artist is someone who produces works of art, and the critic is someone who interprets and appraises them. Which, then, of the things there are, are the works of art? At least they

66

michael morris

are things which are essentially meaningful. But that means that they are things which are there to be understood, in some sense which is more fundamental than the sense in which anything is there to be understood by the person whose job is to understand it. This abstract point, that meaningfulness of the kind which works of art possess requires creation by something like an artist, is reinforced by considering what it would be to understand a work of art, if works of art were not created. We would need to suppose that artistic media have a special kind of property, which I shall call being intrinsically affective. To be intrinsically affective a medium has to have an effect on a hearer or viewer which is entirely independent of the hearer or viewer assuming that anything that partakes of that medium was produced by an artist. So it might be thought that particular kinds of sequences of sound tend to make people feel sad, or uplifted, or relaxed; or that colours occurring in particular kinds of combination with other colours tend to make people uneasy, or calm, or cheerful. We might suppose that intrinsic affectiveness was dependent upon historical, cultural, and social factors: people brought up in different conditions might be susceptible to artistic media in different ways, independently of their assuming that anything that partakes of the medium was produced by an artist. And this might even allow us to grant that languages can be intrinsically affective: certain phonetic combinations might gradually tend to have particular effects on people of particular cultures. What is there to understand in a work of art, if we concentrate just on this kind of intrinsic affectivity and ignore any assumption that the work is produced by an artist? We can, of course, as psychologists, understand and predict the effect produced in viewers and hearers by viewing and hearing the relevant works; in this way, and to this extent, we understand the medium. And as hearers and viewers, we might become psychologically informed, and notice and predict these effects on us; in a certain sense we end up understanding ourselves. But neither of these kinds of understanding is understanding the work of art. And neither kind of understanding can be required to do justice to a work of art. After all, we were led to introduce the notion of understanding precisely to distinguish what works of art do from merely psychological effects; but appeal to what I’ve called intrinsic affectivity seems capable of explaining nothing but merely psychological effects.

doing justice to musical works 67 There is a broader version of what is at root the same problem: if we deny that works of art are created we are forced into a suspect conception of artistic media. There is, at best, a very limited range of effects of any artistic medium which are felt quite independently of the assumption that someone put those parts or features there. And artistic media are, in general, precisely media in which artists work to produce works of art: it is unclear that we could have a conception of an artistic medium as such which does not assume that it is a medium in which works of art are made. The type-token view cannot be defended against these objections by claiming that although the things which are works of art are eternal entities, they become open to aesthetic understanding only after the intervention of an artist. For I have argued that works of art are not only meaningful, but essentially meaningful. We could not have the very same things without their being meaningful. Consequently, we could not have had the very same things before the intervention of an artist. (This means that so-called ‘ready-mades’ are not properly so-called; but that is hardly a counterintuitive claim: it is quite natural to think that taking an already-made object and placing it in a gallery creates a work of art which did not exist before.) Insisting that the meaningfulness of a work of art depends on its having been created exposes us to a possible objection. Surely, it may be suggested, the meaningfulness of a work of art can only depend on its having been created if what is there to be understood in a work of art is the reason why it has been created as it is. That is, we will be driven to suppose that the meaning of a work of art is to be explained in one or both of two ways: by explaining the intentions of the artist, or by explaining the historical circumstances of various kinds which gave rise to the work, and which the work may be understood to express. It may be very useful for understanding the life and mind of the artist to work out her intentions, and it may be important for understanding the history of a period to understand what is shown in a work about the circumstances of its production, but neither of these things—the objection goes—is understanding the work of art itself. As it happens, I agree with the objector that neither historical nor biographical understanding counts as understanding the work itself.¹⁰ But ¹⁰ But biographical and historical information can help us to understand the medium, and, as will be clear in a moment, I think understanding the medium is crucial to understanding the work.

68

michael morris

it is not true that the meaningfulness of a work of art can only depend on its having been created if what is there to be understood in a work is the reason why it has been created as it is. So insisting that works of art be created does not force us into the unacceptable position we are here offered. But if what is there to be understood in a work of art is not just the reason why it has been created, what difference does creation make? How can creation be essential to a work of art’s being meaningful? I will offer only the beginning—indeed, the caricature of a beginning—of an answer. For the simplest (a ludicrously simple) kind of case, let’s begin with the conception of an artistic medium which is available to someone who thinks that works of art are not created: artistic media are intrinsically affective, in the sense which was introduced before. Suppose—to keep things ludicrously simple—that we have a sequence of sounds which tends to make us sad followed by one which tends to make us cheerful. What is the difference between such a sequence of sequences (the sad-making sequence followed by the cheerful-making sequence) appearing as a natural occurrence, and one appearing as the result of the work of an artist who understands this intrinsic affectiveness? Of the latter, but not the former, we can say this: this sequence of sequences is there because it is intrinsically affective in this way. That looks as if it is enough to ensure that we have something which is there to be understood; we have something which is meaningful. Nothing in any real work of art presents itself like this, of course. It is even doubtful that we can isolate a level of intrinsic affectivity which underlies whatever is not intrinsically affective—which underlies, that is, whatever depends on assuming that what we have was created. The crucial parts and features of a work of art strike us, at first sight or hearing, as meant through and through: we never react to them as if to something which has occurred naturally. Moreover, the medium of any work of art is always formed within a tradition over time, and is always understood within that tradition. The tradition itself is a tradition of using the medium intentionally, given the properties the medium has acquired as an intentionally exploited medium within a tradition of using it intentionally. The whole thing is so shot-through with intentionality that we cannot begin to make sense of the intrinsically affective qualities of that medium. For all that, the caricature case does enough to show what we need to do to make sense of the contribution of creation to the meaning of

doing justice to musical works 69 works, without lapsing into thinking that meaning is to be found in the artists’ intentions or the historical circumstances of the works’ creation. The medium of the work of art, whatever it is, has properties of its own. These properties are not the intrinsically affective properties of the caricature: they are properties of a medium whose very nature involves their being exploited intentionally. They may (surely will) include qualities which are commonly described as expressive and representational. The crucial point is that at the time of the creation of any given work of art, there are properties of the medium which are there independently of the intentions and beliefs of the would-be creator. It is the artist’s business to understand these independent properties and to know how to work with them. If she does understand them, and can handle them, she can produce a work of art within that medium of which the following can be said: its essential parts and features are there (in the work of art) because of the properties of the medium. If this is what the artist does, we have the following account of the difference creation makes, which is compatible with denying that understanding a work of art is a matter of understanding the artist’s intention. I claimed in Section 3 that the essential parts and features of a work of art are those which are essential or central to its meaning. If they can be said to be there (in the work of art) because of the properties of the medium, then it seems that the meaning of a work of art must be due to those properties of the medium. More precisely: the meaning of the work of art is due to the intentional exploitation of the independent properties of the medium. That there is something there to exploit is quite independent of the artist’s intentions: no examination of the artist’s intentions will show us what lies in the medium, ready for intentional exploitation. And the contribution of the artist’s intention is just this: it brings it about that what lies ready for intentional exploitation is intentionally exploited. The artist’s intention produces something meaningful, but it does not determine the particular meaning it has—that is due, rather, to what lies in the medium, independently of her intentions. So we can see why understanding a work of art is not a matter of understanding the artist’s intentions; rather, it is a matter of understanding the medium which is intentionally exploited. But the difference made by the artist’s intention is still crucial. Without that intention, we could not say that the features of the work were there (in the work) because of the properties of the medium. It is only when

70

michael morris

we can say this that we can say that there is something which is there to be understood, something which would not exist without existing in order to be understood. It is only when an artist intentionally exploits those independent properties of the medium that we have something which is meaningful at all.

6. Musical Works cannot be Types of Performance Musical works are distinctive (if not unique) among works of art in their relation to performance: they are there to be performed. What can we say about the relation between performance and work, if musical works are meaningful in the way all works of art must be? First of all, there cannot be a merely accidental connection between the two fundamental teleological facts about musical works, that (as works of art) they are there to be understood, and that (as works of music in particular) they are there to be performed. Nor is the connection with hearing incidental here: musical works are there to be heard, and they are only heard in performance. The first thing we need to say about the relation between musical works and performance is just this: performance lets a musical work be understood—it opens it to understanding, we might say—by letting it be heard. But this doesn’t mark a fundamental distinction between musical works and mechanically reproducible works. Take the case of an etching, for example. The process of pressing the copper plate on paper lets the work be seen as it is supposed to be seen, and hence opens it to the kind of understanding which it was designed for. This seems entirely analogous to what I have said about the relation between musical works and performance, allowing for the fact that one is a visual, and the other an aural, medium. But pressing the copper plate on paper is not performing the etching: etchings are not performed. What is the difference? The natural suggestion is this: in the case of a performable art, the process by which the work is opened to understanding itself depends on understanding. Performers themselves have to understand the work, even if only in performing it. What a performer does is guided all the time by her understanding of the work. A performance seems to be connected by understanding to the work of which it is a performance—and not once, but twice. First, its nature as a

doing justice to musical works 71 performance is defined functionally, as being there to open the work to understanding. And secondly, the performance only counts as a performance (rather than a reproduction) by itself depending on an understanding of the work. We may say that a performer doesn’t understand the work she’s performing, but we surely don’t mean that she doesn’t understand it at all: all we mean is that she has a poor or superficial understanding. But (someone may object) mightn’t someone just play the notes in a score, without any conception of how they should be played—just as, perhaps, someone could read out loud words in a language she could not understand—and wouldn’t this count as performance without understanding? I think not: we cannot really understand playing as uncomprehending as this, except as a kind of mentally disengaged automatic behaviour; and then it cannot be thought of as a performance of a work. It seems to me, then, that the notion of a performance cannot be explained without the notion of understanding; and the object of understanding is nothing other than the work itself. The most fundamental objection to the type-token view of musical works is simply that a token cannot be related to a type of which it is a token in this kind of way. This objection can be elaborated in a number of ways, given the conception of what performances are which I’ve just spelled out. Here is one of them. The claim that the relation between works and performances is the relation between types and their tokens is a reductive claim: the ‘of’-relation which holds between performances and the works they are performances of is supposed to be nothing but the antecedently understood ‘of’-relation which holds between tokens and the types of which they are tokens. But the reduction which is demanded, if performances are involved with understanding in the way I have claimed, looks impossible. A type is individuated by a common property which every one of its tokens must share in order to count as tokens of that type. The fundamental thing which makes tokens of a given type tokens of that type is their possession of the relevant property. The type-token view claims, then, that the fundamental thing which makes performances of a given work performances of that work is their possession of some common property. And since the view is a reductive view—aiming to specify the relation between performances and a work without using the notion of a work itself—the relevant common property must be specifiable without reference to the work. What performances of a given work have in

72

michael morris

common, if my account of the nature of performance is even roughly right, is that they are doubly involved in understanding of that work. So the type-token view has to suppose that an account of understanding that work can be given without reference to the work. But it is just not credible that a suitable reductive account of such understanding can be given. To suppose that it can is just to enter the familiar minefield which surrounds all forms of intentional reductionism. Someone might object that this argument depends on my having chosen to attack the particular version of the type-token view which holds that musical works are types of performance, but in fact the argument will apply in any case, given that musical works are there to be heard and are only heard in performance. Suppose, for example, that it were suggested that a musical work is a type of sound sequence. Let us ask: accepting the view for the moment, for the sake of argument, which sound sequences count as tokens of a given work? Given that musical works need to be performed to be heard, the answer has to be: all and only those which are realizations of performances of that work. Thinking of the relation between a work and realizations of performances of it as the relation between a type and its tokens faces at least the same difficulties as those which face the attempt to treat the relation between a work and its performances in the same way. Nor do the problems depend on my having chosen to attack the typetoken view, as opposed to a universal-instance view or a kind-member view of the relation between musical works and their performances. The relation between a universal and its instances cannot be doubly dependent upon understanding in the way in which the relation between a work and its performance is;¹¹ nor can the relation between a kind and its members. We may also note that once we take account of the double link between performance and understanding, we have space to introduce a plausible liberalization in our approach to the criticism of performances. Once we recognize that the connection between a performance and the work of which it is a performance depends on understanding, we can accept that it need be no defect in a performance that it is not a note-perfect performance, in the sense introduced in Section 2. This is helpfully intuitive: we have ¹¹ We should therefore be suspicious of Levinson’s choice of the word ‘instance’ as a quasi-technical term for ‘a sound event, intentionally produced in accord with the determination of the work by the composer, which completely conforms to the work’s sound and instrumental structure as so determined’: Levinson (1987: 377).

doing justice to musical works 73 all heard great performances which pull the music about a little, or disobey certain performance directions, or change the original instrumentation, and it is intuitively just silly to regard these deviations from the letter (as it were) of the score as defects. But more importantly, it allows room for what can be seen a priori to be a possibility: a composer can, intentionally and perfectly properly, write a piece of music which is literally unperformable; the whole point of a piece of music may depend on the fact that there can be no actual note-perfect performance of it. Given that, it can be no defect in any performance of it that it fails to be note-perfect.

7. Puzzlement and the Assimilating Response If the type-token view is wrong, there must have been something wrong with the basic argument for it. I suggested that the basic argument is that it offers (so it is claimed) the best explanation of those three facts: (i) musical works can be heard; (ii) the existence of a musical work does not depend on any particular performance; (iii) the same musical work can be performed many times. The obvious weakness of arguments to the best explanation is that they may not have considered all of the possible explanations. Are there other explanations available? There are. Much more needs to be said about these other explanations than I can say here, but their very existence is enough to undermine the initial case for the dominant version of the assimilating response. The assimilating response to the puzzlement we feel about musical works depends on two assumptions about these three facts: (a) the basic problem manifested in the three facts is the possibility of a single thing being ‘repeatable’; (b) the best explanation of the possibility of a single thing being ‘repeatable’ is provided by assimilating such things to things of a familiar kind. I have used the word ‘repeatable’ in this interpretation of the diagnosis offered by the assimilating response, because it is used by some of its champions (Rohrbaugh 2003; Dodd 2004; Wolterstorff 1975: 122); but I

74

michael morris

leave it in scarequotes, because there is some uncertainty about what it means (the dominant version of the assimilating response wants it to be heard as meaning something like multiply instantiable, but without quite saying that it does mean that). It is still possible to deny (a), on any natural interpretation of the word ‘repeatable’. For just because a performance depends on the performer understanding the work, a performance cannot be regarded as a mere ‘repetition’ of the work. And thinking a little about the importance of the notion of understanding here in fact suggests an alternative explanation of our three facts. That is simply that a musical work is an objective, and objectively meaningful, thing, whose existence and meaningfulness is independent of any understanding of it. It can be performed again and again for the same reason as that for which it is possible to look at the same objectively existing flower (for example) on more than one occasion. As an explanation of facts (i)–(iii), this looks as plausible as that provided by the type-token view. What assumption (a) does is assimilate musical works to other kinds of works of art which might be more simply regarded as ‘repeatable’:¹² for example, novels (because there can be many copies of the same novel), etchings (because there can be many offprints from the same etchingplate), or photographs (because there can be many prints from the same negative). Does assumption (b) hold even of these? This is far from clear. An alternative explanation is suggested by the fact that it is not even initially intuitive to regard works of art of all kinds as types, or universals, or kinds: it is counter-intuitive (even if bravely defensible) to say that paintings and stone sculptures are types, for example. We might then think that the fundamental account of the ‘repeatability’ of ‘repeatable’ works lies in an understanding of what is special about the media in which they are produced. What is distinctive about the media involved in ‘repeatable’ works is that a single act of creation is the source of a number of things (the copies or prints, for example) which have the same meaning. Why should we count this as one thing? Just because we have a single act of creation and the same meaning. It is important not to be misled here. We never need to say that the different copies of the same novel just are the same novel: we need only ¹² The assimilation is clearly at work in Wolterstorff (1975).

doing justice to musical works 75 say that they are copies of the same novel. We do not need to say that the different prints of the same photograph just are the same photograph: we can always insist on saying simply that they are prints of the same photograph. The assimilator may try to press us on this, by asking how two people can literally be reading the same novel unless there is some sense (to be explained, for example, by a type-token view) in which the copies are the novel. But we can always resist this pressure: what it is for two people literally to be reading the same novel just is for them to be reading copies of the same novel. We can always resist the attempt to assimilate the relation between a copy and what it is a copy of to that between particular and general. And it looks as if we will have some version of the argument of the last section to support us in this: the relation between a copy and what it is a copy of must always be an intentional relation (however mechanical the means of reproduction), which the relation between particular and general (token and type, instance and universal, member and kind) cannot be. I have claimed that it is far from obvious that assumption (b) is true, that, even where we have ‘repeatable’ works of art, the best explanation of their ‘repeatability’ is the one provided by the assimilating response. So far I have only illustrated this with versions of the assimilating response which take ‘repeatable’ works to be in some way general. But an assimilating response is one which aims to resolve the puzzlement we feel about these works of art by assigning the works in question to some familiar ontological category, and the candidate categories are not all general. Someone might claim, for example, that ‘repeatable’ works are continuants, of which the various copies are ‘stages’, perhaps, or ‘embodiments’ (see Rohrbaugh (2003), who takes the copies to be ‘embodiments’). This proposal in effect assimilates ‘repeatable’ works to familiar things like tables, ships, and people. But we should ask why this is a better explanation of unity despite ‘repeatability’ than the simple one I have offered already, in terms of a medium which allows a single act to generate multiple copies which have the same meaning. And as before, it seems extremely doubtful that the intentional relation which must hold between a copy and what it is a copy of can be modelled intelligibly in terms of the relation between the stages or embodiments of a continuant and that continuant itself. No doubt assimilators, of whatever kind, will have something to say in response to these points: after all, I have offered no more than sketches of alternative explanations of facts (i)–(iii), rather than any finished theory.

76

michael morris

Nevertheless, even these brief remarks seem to me enough to take the shine off the assimilators’ arguments to the best explanation. And given the fundamental defects of the views which the assimilators end up proposing, I think we should feel no temptation to adopt any version of the assimilating response. In the face of the problems which confront the assimilating response, it is tempting to say that we should just regard musical works (or performable works in general, perhaps) as sui generis. This may be right, but I think it is not enough just to say this and have done with it. For saying that musical works (or performable works in general) are sui generis may be understood as trying to do something which the assimilating response also aims to do: it may be understood as trying simply to remove the puzzlement which we feel about the nature of musical works. Accept that musical works are sui generis, it might be taken to say, and stop worrying about them. There are two ways in which this seems inadequate. First, it seems to forestall the general overhaul of our everyday ontological assumptions which acknowledging the true nature of musical works seems to demand. The supposedly familiar, spatio-temporally well-defined items like sound sequences or performances turn out not to be more basic than musical works: we need, then, to revise our usual assumptions about what should be taken as familiar and unpuzzling in ontology, and to examine our reasons for holding them. And, second, it is not at all obvious that the right response to an ontological puzzlement, like the one we feel about musical works, is to attempt to remove it. It seems to me that to get a serious sense of the nature of musical works, we don’t need just to acknowledge and record the fact that they are sui generis: we need to keep their peculiarity alive. It is tempting here to draw a parallel between the task of the philosopher of music and that of the musical performer. The philosopher of music, we might say, needs to be attentive to the true nature of musical works, as such and in general, just as a performer needs to be attentive to the true character of a particular work she wants to perform. And in both cases that attentiveness cannot exist without something which is at least akin to puzzlement: the philosopher’s attentiveness to the nature of musical works depends on keeping alive a sense that musical works are not just any old things, just as the performer’s attentiveness to the character of a particular piece of music depends on not letting any feature of it seem merely routine.

doing justice to musical works 77 This seems to be required for the philosopher to do justice to the nature of musical works, just as it is for a performer to do justice to the piece she is playing.¹³ References

Armstrong, D. M. (1989) Universals: An Opinionated Introduction, Boulder, CO: Westview Press. Davies, S. (2001) Musical Works and Performances, Oxford: Clarendon Press. Dodd, J. (2000) ‘Musical Works as Eternal Types’, British Journal of Aesthetics 40: 424–60. Dodd, J. (2002) ‘Defending Musical Platonism’, British Journal of Aesthetics, 42: 380–402. Dodd, J. (2004) ‘Types, Continuants, and the Ontology of Music’, British Journal of Aesthetics 44: 342–60. Dummett, M., (1975) ‘What is a Theory of Meaning?’, in S. Guttenplan (ed.), Mind and Language, Oxford: Clarendon Press. Goodman, N. (1981) Languages of Art, 3rd edn, Brighton: Harvester Press. Heidegger, M. (1956) ‘The Origin of the Work of Art’, trans. A. Hofstadter; reprinted in M. Heidegger, Basic Writings, ed. D. F. Krell, 2nd edn, London: Routledge (references to the reprint). Hume, D. (1757) ‘Of the Standard of Taste’, reprinted in D. Hume, Essays Moral, Political, and Literary, ed. E. Miller, Indianapolis: Liberty Fund. Kivy, P. (1983) ‘Platonism in Music: A Kind of Defence’, Grazer Philosophische Studien 19: 109–29. Kivy, P. (1987) ‘Platonism in Music: Another Kind of Defence’, American Philosophical Quarterly 24: 245–52. Levinson, J. (1980) ‘What a Musical Work Is’, Journal of Philosophy 77: 5–28. Levinson, J. (1987) ‘Evaluating Musical Performance’, Journal of Aesthetic Education 21: 75–88; reprinted in Levinson (1990c) (references to the reprint). Levinson, J. (1990a) ‘Authentic Performance and Performance Means’, in Levinson (1990c). ¹³ I am very grateful to Julian Dodd for provoking me to think about the issues of this paper, to Paul Davies for many conversations about them, and to Stephen Davies, Terry Diffey, Jerrold Levinson, Stefano Predelli, Roger Scruton and Kathleen Stock for their comments on an earlier draft.

78

michael morris

Levinson, J. (1990b) ‘What a Musical Work Is, Again’, in Levinson (1990c). Levinson, J. (1990c) Music, Art, and Metaphysics, Ithaca, NY: Cornell University Press. Rohrbaugh, G. (2003) ‘Artworks as Historical Individuals’, European Journal of Philosophy 11: 177–205. Wittgenstein, L. (1966) Lectures and Conversations on Aesthetics, Psychology and Religious Belief, ed. C. Barrett, Oxford: Blackwell. Wollheim, R. (1980) Art and Its Objects, 2nd edn, Cambridge: Cambridge University Press. Wolterstorff, N. (1975) ‘Toward an Ontology of Art’, Noˆus 9: 115–42. Wolterstorff, N. (1980) Works and Worlds of Art, Oxford: Clarendon Press. Further Reading

A clear statement of the kind of view which is my target is Wolterstorff (1975), where musical works are said to be kinds. The view that musical works are types is often associated with the groundbreaking work of Levinson (1980), although his version has a number of awkward complexities. A bold, clear, and austere version of that kind of view (nearer perhaps to Wolterstorff) is Dodd (2000).

3 Versions of Musical Works and Literary Translations S T E P H E N DAV I E S

The composition of a musical piece can fall short of the work’s completion. In 1821, Schubert sketched an E major symphony he did not finish. The following year, he abandoned another symphony after writing two movements and outlining its scherzo. We call it the ‘‘Unfinished.’’ Mahler notated much of his Tenth Symphony in short score, but only the first movement was fully orchestrated before his death. A less often remarked fact is that a work’s composition can overshoot its completion. It is the description apt for these cases that is the topic of this chapter. But before I get to that, it is useful to describe some of the signs that show a work to be finished.

1. What Signifies the Completion of a Musical Work? There are a number of indicators that a musical work is complete. (1) The composer declares that it is so. In a letter to his father dated April 10 of 1784, Mozart writes of his piano concerto in G, K. 453: ‘‘I have finished today another new concerto for Fr¨aulein Ployer’’ (Anderson 1966: 874). (2) Or the composer marks the score or a catalogue. In the case of K. 453, Mozart entered the opening theme into the thematic catalogue he had started in

80

stephen davies

February 1784. ¹ (3) Or the score is copied for performance. Along with other manuscripts, that of K. 453 was dispatched to Mozart’s father in Salzburg on May 15, 1784. Mozart wrote: ‘‘I am not particular about the symphony, but I do ask you to have the four concertos copied at home, for the Salzburg copyists are as little to be trusted as the Viennese’’ (Anderson 1966: 876). (4) Or the work receives a public performance designated as its premier. K. 453 was played publicly for the first time with Fr¨aulein Ployer as the soloist on June 13, 1784 (Deutsch 1966: 225). (5) Or an authorized version of the score is printed and sold. The published score of K. 453 was advertised for sale in Vienna in August and September of 1785 (Deutsch 1966: 249–52). None of these signs of a work’s completion is necessary. The unsigned but finalized manuscript of many an unperformed and unpublished symphony has languished in its composer’s bottom drawer. Also, because they are temporally spread, the combination of these markers can leave the precise moment of completion somewhat uncertain. Writing to his father on June 12 of 1784, Mozart observes of the autograph score of his piano concerto in D, K. 451, that there is ‘‘something missing’’ in the solo passage in C in the Andante. He continues: ‘‘I will supply the deficiency as soon as possible and send it with the cadenzas’’ (Anderson 1966: 880). Yet the concerto was ‘‘finished’’ about three months earlier, on March 22 of 1784, and premiered with Mozart as soloist on March 31 at a subscription concert. It was fairly common for Mozart not to write out the detail of the soloist’s part if he was its player. He would later record the part in full for the concerto’s publication. Nevertheless, if all these signs are in place, we can have a high level of confidence that the work is finished. (Cases in which an unfinished work bears all these markers are conceivable but, in practice, more or less non-existent.)

2. Versions by the Composer As I indicated at the outset, the process of composition sometimes outlives the work’s completion. In other words, the signs of completion, including ¹ Bach penned the letters ‘‘S. D. G.’’ in the margins of his scores, which translates as ‘‘solely to the glory of God.’’ At the end of his scores, Haydn wrote: ‘‘Fine. Laus Deo.’’ Other composers have signed and dated their scores.

versions of musical works/literary translations 81 public premier and publication, are plainly in place, and the composer designates that the work is finished, yet he later changes it in ways that should affect its identity. As I discuss below when considering (selfauthored) transcriptions, sometimes the result is a new, derivative work. My concern, though, is with the case in which the composer’s post-completion efforts result in a change or addition to the finished work, not in a new one. I call the result a version. Here are some examples: Bruckner revised and altered his early symphonies after their first publication. For instance, in 1891 he recomposed his First Symphony of 1866.² Published editions and manuscripts of Chopin’s works differ, and not only as a result of editorial error. Chopin was responsible for most of the variants. This has led some musicologists to claim that Chopin did not share our modern concept of musical works as completed, fixed, re-identifiable individuals. I disagree (see Davies 2001: 92–7, 119–23). Whereas composers of earlier eras quickly moved on to new pieces, even if they frequently inserted into these material borrowed from prior works written by themselves or others, Chopin and his contemporaries became more reluctant to relinquish control of their compositions, even after publication. I do not interpret this as showing that the work concept had yet to take shape. Instead, I see it as reflecting a pragmatic concern within a tough but unregulated market, both to do what was necessary to cater to the taste of the purchasing public and to exploit any work for all that could be wrung from it.³ Hardheaded pragmatism can be seen also to lie behind the creation of work versions in other situations. Mozart had a great success with Don Giovanni in the provincial center of Prague, but when he took it to Vienna the court singers expected him to add numbers tailored for them, which he did. Throughout the nineteenth century, composers who exported their operas to Paris had to conform to the French passion for ballet by inserting dances. When Stravinsky lost access to royalties for Firebird and ² To complicate matters, the status of many editions of Bruckner’s symphonies is controversial. Other composers and editors mangled the early editions and the twentieth-century scholarly editions of Robert Haas and Leopold Nowak by no means agree. My comments do not rely on these uncertainties, however. ³ The same phenomenon is evident with new music that is posted on the Internet and frequently remixed. Composers of such pieces seem to have a robust idea (the standard one) of what a work is and of when it gets released to the community, but they take advantage of the chance to play around with it after ‘‘completion.’’

82

stephen davies

Petrushka, following the revolution in Russia, he re-scored them and had the new versions published in Europe and America. The alterations are modest—the notations of certain rhythms are simplified and there is a reduction in the numbers of flutes and such like. As was plainly intended, the changes did not depart so far from the originals that they counted as new pieces, but they did make it more likely that the new incarnations would be played and that the revenue could again go to the composer. In all these cases, what resulted were versions of old works, not new ones. By contrast, consider Stravinsky’s composition of Les Noces. With the sound structure more or less intact, he struggled with the orchestration. He scored the piece for a large orchestra and, again, for mechanical pianos, before settling on the instrumentation that is now familiar, which uses percussion and four pianos played live. On the account presented here, the work does not have three versions because the various changes antedated the work’s completion. Instead, it had several drafts. They were phases in its initial composition, not alterations made after its finalization. More generally, it is not unusual for a composer to tinker with the work during rehearsals or after its premier, as it is prepared for publication. Though the completion date is not always clear-cut, these changes usually concern the work’s finalization. Authors’ versions, if there are any, come after that point. As I use the term, a version is produced by the work’s composer if he changes elements that should be constitutive of the work’s identity after the piece’s completion, and where the alterations intentionally and moderately alter identity-relevant features of the original, but without resulting in the production of a new but derivative work. Drafts are like versions, except they are made prior to the work’s completion.⁴ Multiple versions of the same musical work can co-exist. Though the final version sometimes has a special authority because it indicates the composer’s definitive thoughts about the work, this is by no means always the case. For instance, the opera composers who adapted their works for Paris productions often found the process distasteful even if pragmatically desirable. The order in which ⁴ Where one studies the psychology or history of the work’s creation, drafts may be more interesting than versions, but versions are more philosophically intriguing, which is why I focus on them. The production of succeeding drafts is a predictable part of the piece’s development and completion. By contrast, versions are ontologically provocative, since, with them, the identity of the work apparently survives alteration in the kinds of features we would normally think of as work-constitutive.

versions of musical works/literary translations 83 versions are produced implies nothing on its own about their aesthetic importance. The introduction of work versions to our musical ontology is messy, but so what? It respects a lack of neatness shown by composers themselves in the labeling and dissemination of their music. And, apart from the discomfort that untidiness among our classifications can bring, this ontological profligacy is supportable. All that is required is that the performers (analysts, historians) make clear that they are dealing with the version published in Paris, or the version of 1863, or the version for Prague as against the one for Vienna, or whatever. In other words, even if works with multiple versions are equivocal, we can usually individuate their versions clearly enough.

3. Versions by Others A version of a work can be produced by a person other than its composer. This happens, for example, when an unfinished piece is completed by a third party. Deryck Cooke produced a performing version of Mahler’s Tenth Symphony. Each of the many composers who continued the last, unfinished fugue of Bach’s Art of Fugue produced a different version of the work. Mozart’s Requiem exists only in a version finished by three of his pupils. For these cases, there is no definitive, finalized composition. There is the piece left unfinished by the composer and one or more versions indicating a way of completing it. Versions of a work are also produced by music editors if the composer’s text, or extant copies of it, are ambiguous or conflicting. If the editor chooses between several possibilities—for instance, she decides that the C in measure 40 should be sharpened, though the sharp is indicated in only some sources for the work and is not mandated by the performance practice or conventions—the result is a version.

4. Work Versions versus Performance Interpretations Work versions should be distinguished from performance interpretations. Because the prescriptions addressed to the work’s performers underdetermine the detail of the performance, the players must go beyond what is

84

stephen davies

work determinative in rendering the work. There are many ways of doing this while remaining faithful to the piece. The differences that result mark distinctions in the interpretations the work receives.⁵ When a composer interprets his own piece in performance, the outcome is an interpretation, not a new work version. The production of a work version requires changes in features that are usually constitutive for pieces of the kind in question. By contrast, accurate interpretations respect what is work-constitutive. They vary qualities, features, or levels of detail that are not identity determining for the work. For instance, it may be prescribed that the performer decorates the melody when it is repeated, but the detail of that decoration might be left open and the player’s choices add to her performance interpretation. Though the work includes decoration, no particular set is mandated and many choices (even if they are not all equally tasteful or effective) can be consistent with respecting the work’s identity. And even where the interpretation deliberately deviates from what the composer instructs, the acts involved are ones of performance, not composition. Where one draws the line between work versions and work interpretations is likely to depend on the level of detail at which works of the relevant kind are specified. In other words, it depends on the work’s ontological character. If the work is ‘‘thin,’’ wide variety among its instances is likely to be a function of its interpretation, not of its re-composition into versions. If the work is very ‘‘thick,’’ significant differences between its instances, where they are not accidental, may be indicative of versions. Interpretations within some performance traditions share important characteristics with work versions, however. I have in mind cases in which interpretations attain an autonomous standing, with the expectation that they are to be repeated and preserved, and in this resemble work versions. Also, where the piece being interpreted is ontologically thin, its interpretation involves creative decisions of the kind that composers make, and these decisions post-date the work’s completion, which again makes them similar to work versions. The result can be called an interpretation version. It is not a work version, as I have characterized that notion, because it ⁵ This is not to say that interpretation comes into the picture as something added only after faithfulness has been achieved. Interpretation is not reserved solely for the ‘‘gap’’ between the work and the concrete detail of its performance. It reaches all the way down, as it were.

versions of musical works/literary translations 85 aims to establish a repeatable way of performing the source work, not to recompose that work. Interpretation versions are common and well known, but let me introduce an out of the way example: in the Balinese gamelan gong kebyar tradition, Teruna Jaya is a piece that has existed in its modern form for more than fifty years. Though much is common between renditions of the work by different ensembles, there is also considerable variety. It is expected both that different performance groups will create their own interpretation and that each performance group will stick closely to the interpretation it establishes as its own.⁶ Because these interpretations are worked out, are preserved and repeated, and are associated with particular groups, they are paradigms of interpretation versions. They do not involve post-completion changes to features that the tradition treats as work-constitutive, but neither are they as ephemeral as interpretations that are specific to a single performance. A more familiar example of an interpretation version is the rock ‘‘cover.’’ Because rock songs are typically rather spare as regards their workidentifying elements, re-recordings can remain faithful while differing in many respects from the original.⁷ Joe Cocker’s account of ‘‘With a Little Help from my Friends’’ differs considerably from the Beatles’, but is not thereby unfaithful as a performance, I maintain. As a recording, the cover attains a status in its own right as a repeatable individual—not only can it be re-played, it can be performed again and again by Cocker or by his emulators—so it is a version, not a one-time-only interpretation. Because the cover does not involve the re-composition of work-constitutive elements of the original, despite differing in many respects from the Beatles’ source, it qualifies as an interpretation version, however, not as a work version.⁸ ⁶ Over the years, a group’s interpretation can evolve and alter along with wider aspects of the prevailing style, but these alterations are more gradual and less self-conscious than is the creation of the initial interpretation. ⁷ I defend this account of rock’s ontology in Davies (2001). Gracyk (1996) takes the different view that the works in rock are electronic compositions that are presented via recordings and, hence, that are not for performance (see also Fisher 1998). In his theory, the cover is a new but derivative piece, not a new performance of the same song, as I maintain. ⁸ In Davies (2001: 180–1), I suggest that Benny Goodman’s band arrangement of George Gershwin’s Fascinating Rhythm and Glenn Miller’s rendering of Joe Garland’s In the Mood qualify as independent but derivative works, because they achieved the status of autonomous, repeatable versions preserved via recordings and notations. It is no less plausible to regard them as interpretation versions, as I imply here, however. Unlike transcriptions, which are discussed in the next section, these adaptations generate performances that are faithful to the original, ontologically thin, sources.

86

stephen davies

For reasons that have been already acknowledged, interpretation versions can be like work versions in many respects, and the terminologies I have specified are not clearly observed in ordinary discourse about music. Nevertheless, it is worthwhile to mark their differences. We expect completed works to receive differing interpretations and we regard interpretational variety and creativity as consistent with faithfulness to the work being performed, so long as its work-constitutive elements are preserved. Also, where works are thin, we appreciate that interpreters are creative after the manner of composers. As well, there is nothing surprising about the fact that interpretations can achieve standing in their own right as repeatable, persevering entities. By contrast with all this, it is disconcerting to have to accept that a work’s identity can survive alteration to its work-constitutive features following its authorization as a completed piece, and to allow that a single work can exist in multiple incarnations. These latter characteristics are the hallmarks of what I have called work versions, as opposed to interpretation versions.

5. Composers’ Versions versus Transcriptions Just as work versions should be distinguished on the one side from performance interpretations and interpretation versions, so, on the other, they must be separated from derivative yet distinct new works, a prominent example being transcriptions. Where a medium-specific piece is adapted to a new medium, as when a nineteenth-century symphony is rewritten for the piano, the result is a distinct work, I claim (Davies 1988). Usually the transcription postdates the original, but it can anticipate it if, for example, the composer creates a symphonic work on the piano. The first outing of The Rite of Spring came when Stravinsky and Debussy played the composer’s transcription for two pianos. Work versions involve changes to work-constitutive properties. Their creation can include alterations to pieces’ orchestrations. As mentioned above, Stravinsky reduced the number of wind players in later versions of Firebird and Petrushka. These modifications did not depart from the medium in which the original was written, however. For these, the medium is the symphony orchestra, but such an orchestra can be treated flexibly as regards its forces; in particular, with regard to the number of woodwind,

versions of musical works/literary translations 87 brass, and percussion lines. When he created new work versions of these early ballets, Stravinsky lightly reworked them in the same medium. Transcriptions also involve changes to work-constitutive instrumentations.⁹ I regard transcriptions as new works, not as new work versions, however, because they involve changes in instrumentation that alter the work’s medium and its medium-specific contents. When one instrumental medium is replaced by a significantly different one, the work’s contents must be adapted accordingly. Among other matters, this usually involves changes to the piece’s notes or their relative dispositions, and imposes different technical requirements and constraints on the piece’s new performers. When a work’s features are filtered through a new medium, the impact is usually sufficiently radical that a new piece results. A musician might begin to compose, using a finished work as her source. If she carries the process of re-composition far enough, she writes a new piece. The new work is influenced by the original, and perhaps audible traces of this inspiration remain detectable in quotations or allusions. In a different scenario, the composer does not carry the process very far and she conceives of herself as revising the source rather than going beyond it. The product is what I have called a work version. The practice of transcription lies between these extremes.¹⁰ The audible relation with the original is preserved, as is the sound-structural outline and much else, yet the change in instrumental medium distances the transcription from its model, with the result that a new work is produced. At what point does the process of composition leave the original work behind? When do alterations to the original cross the border between a work version and an entirely new, albeit derivative, piece? There is no simple answer, of course. And as long as we are all clear about the degree of arbitrariness involved, perhaps we can say what we like. Nevertheless, my intuition is that transcriptions usually achieve a degree of independence from their sources that versions do not. This is intended and reflects differences in what typically motivates the act of reconstruction in the ⁹ If the instrumentation of a piece is not work-determinative, varying its orchestration would result in an interpretation, not a transcription. Many musical pieces—folk songs, Bach’s Art of Fugue, most pop songs, much music composed prior to the standardization of the orchestra—do not have work-determinative instrumentations. ¹⁰ On the border between transcriptions and entirely new pieces is the ‘‘fantasia after’’ and ‘‘homage to.’’ Sometimes these are labeled as transcriptions, an example being Percy Grainger’s ‘‘transcription’’ of Tchaikovsky’s ‘‘Waltz of the Flowers.’’

88

stephen davies

first place. Work versions re-compose their sources, but in a deliberately restrained way, so that they do not threaten to count as new pieces. Also relevant is this: if we were to count transcriptions as work versions, not as new works, there would be the implication that the composer of the transcription is a co-author of the work. We would have, not Liszt’s transcription of Beethoven’s Fifth, but Beethoven’s Fifth as composed by Liszt. When one composer completes another’s work, the final result should be credited to both. When one composer transcribes another’s piece, she deserves acknowledgment as the transcription’s author, but not as the co-author of (a version of) the original.¹¹

6. Versions in other Artforms Much of what I have written about work versions in music could be applied to works produced in other artforms. The novelist can rewrite her already published book. The movie can be re-released as the director’s cut, the modified-for-TV print, and the DVD supplement, which might include deleted scenes and an alternative ending. The artist can add to a previously completed oil painting and the sculptor can cut bits from a marble statue that was designated as finished long before. These last cases provoke the following thoughts: it is common to distinguish artforms with works that must be singular from those in which multiple instances of the work are possible. Handmade sculptures, drawings, and paintings are singular; prints, cast statues, novels, plays, songs, movies, and musical works are potentially multiple. Some philosophers, Currie (1988) being one, reject this division, however. They regard all artworks as potentially multiple. A doppelganger of Mona Lisa would instance the same work as Leonardo’s painting, provided it was accurate enough. Does the earlier observation—that oil paintings and marble sculptures can exist in more than one version—show that ontological multiplists are right? Handmade paintings and sculptures can be made in sets. An artist might paint more or less identical copies of a single scene, and sell one to the ¹¹ Sometimes transcribers deserve and get equal billing with the composer of the original, as Ravel does for his orchestral transcription of Mussorgsky’s Pictures at an Exhibition, which was written for the piano. This equal acknowledgment is justified by the level of creativity and originality displayed by Ravel, but we do not identify him as co-author of the work.

versions of musical works/literary translations 89 palace, one to the church, and so on. We might regard these paintings as presenting different versions of the same subject matter, but there is no reason to treat the members of the set as versions of a single, generic painting. Members of the set can be identified as discrete but related works. When it was said earlier that there can be more than one version of an oil painting, the relevant example was not this but a different one, which I now describe. The artist completes a work but returns (much later, let us assume) to the same canvas or statue and takes to it again with brush or chisel. In doing so, he does not obliterate the original, though he does alter it in ways that should be relevant to its identity. He adds an angel or two in the sky, trims fat from the statue’s thighs, or whatever. As a result, the work exists through time in more than one version, and to that extent it appears to be multiple. I deny that this counts against the commonsense view that handmade paintings and sculptures are singular, however. The multiple versions of plays and novels are numerically distinct and coexistent. This is possible because they are types or kinds indicated as such by an artist. By contrast, the painter in oils or the sculptor in marble can create a version only by laboring directly and concretely on the original work and its material. The new version supplants the earlier one; in other words, such works exist in only one version at a given time. The work preserves its identity through the process of change, but this does not make it a universal or type. Instead, it might be compared to a person, who can be stooped and gray now yet be the same as the child who formerly was blond and upright. Oil paintings and hewn statues can have successive versions, but unlike plays and novels, they cannot have coexistent versions (and instances), which is what would be required if they were to be multiple in the relevant sense. Rather than deducing from the possibility of their having versions that paintings and sculptures are all potentially multiple, it is more sensible to observe that the ontological differences between singular and multiple works dictate a corresponding difference in what is involved in creating versions of them.

7. Translations in Literature and Poetry Here is a new issue. What is the status of translations of literary works to a language other than that in which they are written? Are translations

90

stephen davies

equivalent to work versions or to transcriptions? Is a translation a variant that instances the original work, or is it a distinct but derivative piece? If we ask a person if she has read War and Peace, she is likely to answer ‘‘yes’’ if she has read an English translation. It would be odd if she replied: ‘‘No, only the translation.’’¹² But it would be hasty to conclude from this that the translation thereby counts as an English-language version of Tolstoy’s novel. It is likely that the respondent interprets the question as asking if she is familiar with Tolstoy’s work via its translation, not as asking if she accepts their co-identity. People turn to a translation usually because they do not read the language of the original and therefore, for the case of War and Peace, cannot access Tolstoy’s work as he wrote it. In the past, musical transcriptions were the same. Before there were radios, recordings, or affordable subscription concerts, piano transcriptions provided the main route to many orchestral works because households frequently possessed a piano and a competent pianist. It would have been natural in those times to claim acquaintance with Beethoven’s Fifth on the basis only of Liszt’s transcription. Nowadays we are liable to insist that a person does not really know Beethoven’s Fifth until she hears the work as it was written for orchestra. A person’s claim to have read Tolstoy’s work by reading a translation might seem natural, then, only because the original remains comparatively inaccessible to most native speakers of English.¹³ But in that case, it could be that the translation is best regarded as a distinct piece, which is the way Liszt’s transcription should be considered. Notwithstanding these observations, there are good reasons for thinking of the translation as a work version. It is intended and represented as a variant of the original, not as a new, though derivative, piece. And this is plausible, to the extent that it is possible to achieve a high degree of ¹² That answer might be more likely, though, if the question had been ‘‘Have you read Voiynah ee meer?,’’ supposing the person to understand this as the Russian title. ¹³ Compare the role of literary translations with that played by mass reproductions of singular or limited edition paintings and sculptures. Reproductions can range from photographs to duplicates done in the same medium as the original. We are liable to regard these as standing in for the original but not as work versions. Perhaps this is because they are often less than fully faithful to the original. If we had a matter replicator that could reproduce physical objects down to the molecular level, we might be inclined to accept clones of the originals as work equivalents. They are not work versions, because they preserve rather than modify work-identifying features. And neither are they genuine instances, if the works in question necessarily are singular or limited in number.

versions of musical works/literary translations 91 accuracy (as regards content and general mood) in converting one language to another. Admittedly, it can be difficult for a foreigner to appreciate the nuances and associations of a story in translation, but that problem may be cultural rather than linguistic. A foreigner who could read the work in its original language might miss exactly the same features through a lack of sensitivity to the conventions and values of the culture in which the story finds its home. If the work is the story, and if the story survives its translation, the translation is best regarded as a version, not as an autonomous work. Though I stand by it, this conclusion should be viewed warily. The more resistant to translation are linguistic and semantically relevant features of the work, the nearer translation approaches the case of musical transcription; that is, to being an autonomous though related work. An example might be that of translating from a tone language (such as Mandarin) to a non-tone language (such as English). The relevance of pitch to meaning may be exploited in the original to present artistically relevant possibilities that cannot survive in translation. A different, more obvious case is that of poetry or rhyming verse. It is notoriously difficult to reproduce in another language the subtle semantic content and ease of expression of something written in rhyming iambic pentameter while respecting the same formal constraint. The nearer language comes to being merely a vehicle for meaning, so that the content of the story is ‘‘transparent’’ to the language in which it is written, the nearer translations approach work versions so long as the translator aims successfully to produce a literal rendering of the original. But the more the language of writing becomes a medium that draws attention to itself and, prism-like, modulates, filters, and inflects the message expressed through it—or alternatively, the freer and more ‘‘creative’’ the translator tries to be—the more translations are analogous to musical transcription. In this latter case, translations must function like transcriptions, either because they lose the original’s interplay between medium and content or because they reproduce an appropriate effect via the different resources and possibilities of the alternate language. In other words, where the treatment of the language of the original is such that a translation cannot convey what is required simply by telling the same story under similar formal constraints, the translation might better be seen as a new but derivative piece than as

92

stephen davies

a version of the original. But where the audience is not multilingual, the translation then provides a useful substitute, if not a work version.¹⁴ References

Anderson, E. (1966) The Letters of Mozart and his Family, 2nd edn, London: Macmillan. 2 vols. Currie, G. (1988) An Ontology of Art, London: Macmillan. Davies, S. (1988) ‘‘Transcription, Authenticity and Performance.’’ British Journal of Aesthetics 28/3: 216–27. Davies, S. (2001) Musical Works and Performances, Oxford: Clarendon Press. Deutsch, O. E. (1966) Mozart: A Documentary Biography (2nd edn), trans. E. Blom, P. Branscombe and J. Noble, London: Adam & Charles Black. Fisher, J. A. (1998) ‘‘Rock ’n’ Recording: The Ontological Complexity of Rock Music,’’ in P. Alperson (ed.), Musical Worlds: New Directions in the Philosophy of Music, University Park, PA: The Pennsylvania State University Press, 109–23. Gracyk, T. (1996) Rhythm and Noise: An Aesthetics of Rock Music, Durham, NC: Duke University Press. Further Reading

Davies, S. (2003) Themes in the Philosophy of Music, Oxford: Oxford University Press. Dilworth, J. (2001), ‘‘A Representational Theory of Artefacts and Artworks,’’ British Journal of Aesthetics 41/4: 353–70. Livingston, P. (2003) ‘‘Pentimento,’’ in B. Gaut and P. Livingston (eds), The Creation of Art, Cambridge: Cambridge University Press, 89–115. Rohrbaugh, G. (2003) ‘‘Artworks as Historical Individuals,’’ European Journal of Philosophy 11/2: 177–205. ¹⁴ I am grateful to Paisley Livingston for challenging me to clarify my views on this topic and to Jonathan McKeown-Green and others too numerous to list conveniently for comments and suggestions on drafts of this paper. The versions will all be mine.

PA RT I I

Musical Expression

This page intentionally left blank

4 Expression in Music D E RE K M AT RAV E RS

Progress in enquiry depends as much on getting the right questions as it does on getting the right answers. This is nowhere better illustrated in recent philosophy than in the discussion of the expression of emotion by instrumental music. What I hope to do in this paper is to argue for framing the question in one particular way, and, having done so, evaluate the answers that have been given to that question (even if the answers were not originally given in answer to that question). It has become more or less traditional to approach this through attempting to make sense of the claim ‘the music is sad’. Clearly this sentence is simply a placeholder; nobody should take it to be an example of serious criticism (Kivy 1989: 181). The narrowness of focus consequent on taking this as the placeholder does, however, sideline several issues. In particular, it directs our attention to the expression of sadness by a brief passage of music, rather than the expression of sadness by an entire piece (although the latter may well emerge from relations between the former). However, if we can first identify and second solve the philosophical problem this simple placeholder raises, that would itself be progress. There have been many attempts to identify the problem, most along the lines of pointing out that, in the central case, we operate with a loose equivalence between ‘being sad’ and ‘feeling sad’. As music is insensate it cannot feel sad, and hence, given the equivalence, cannot be sad. This, however, does not seem the most perspicuous way of stating the problem, as we are happy to use ‘sad’ of insensate objects or events (such as ‘a sad occasion’, ‘sad news’, ‘a sad book’). Rather, I think, we should think first about what we could expect philosophy to contribute to the debate and frame the question so as to draw on this expectation. It is part of philosophy

96 derek matravers (at least) to explicate what we mean by our claims: that is, to provide a perspicuous account of their content. Hence, the problem would be to explicate the content of ‘the music is sad’: what would someone who made that claim mean? The answer to that question, at least as a preliminary, seems to me reasonably obvious: they mean that the experience of the music has a certain phenomenology—when one listens appropriately, one will have an experience of it the right description of which is ‘the music is sad’. Thus, to explicate the content of the claim will be (at least in part) to throw light on the character of this experience so as to make it clear why it is rightly described in terms drawn from the emotions. Compare this with the philosophical work on pictorial representation. An approach here, perhaps popular enough to be regarded as the orthodox approach, is to say what it is for an object to be a picture by providing a characterization of the experience to which it gives rise. This view is most closely associated with Richard Wollheim, for whom a picture of an x is an object that gives rise to a certain experience: that of seeing an x in the picture (provided also the artist intends us to see an x in the picture). A great deal of effort has been expended on clarifying the nature of the experience of this ‘seeing in’. Whatever the virtues of this approach in the case of pictorial representation, it is not obviously correct. That is, it is not obvious why the answer to the question ‘What makes it the case that this object is a picture of an x?’ is ‘Because of the nature of the experience to which it gives rise’. After all, unless one is a sort of idealist, one would not think that form of answer appropriate in attempting to define ‘car’ or ‘bed’. However, the train of reasoning given above seems to provide a good reason for this form of answer in the case of expression. That is, what one appears to be claiming when one claims that a piece of music is sad is that the experience of hearing it has a certain phenomenological characteristic. If the foregoing is correct, the problem is to throw light on the character of the experience of hearing the music. I said above that any solution should also make it clear why the music is rightly described in terms drawn from the emotions. This, however, may have been too hasty. Inspired by the thought that the experience of a passage of sad music is not the experience of the natural expression of sadness, one might think that the content of the experience is a purely musical property—one of a certain limited set of musical constructions, perhaps—or a sui generis aesthetic property. This approach has been criticized by Roger Scruton:

expression in music

97

Consider the application of an emotion term—such as ‘sad’—to a work of art (or, for that matter, to an event, or a letter, or anything that cannot literally be in the emotional state of sadness). To understand the word ‘sad’ is to know how to apply it to people in order to describe their emotional state. The criteria for the application of the term ‘sad’ concern the gestures, expressions and utterances of people on the basis of which I describe them as sad, and to grasp the concept of sadness is to know how to apply it on the basis of these criteria. When we apply the concept to art, however, it is arguable that these criteria are not, or need not be, present. Does this mean that the term ‘sad’ is ambiguous? (1982: 38)

Scruton is right to think that the entailment that the term is ambiguous is a reductio of the view. It is worth noting a small complexity here. Scruton claims that difference in criteria of application, rather than difference in property being referred to, is sufficient to ensure ambiguity. Thus it is not clear that his point can be used, as I have attempted to use it, to criticize the notion that the property being referred to is distinct from that to which ‘sad’ would refer in the central case. However, we can state the point more generally. The content of the claim ‘the music is sad’ should overlap in the central and the aesthetic cases. Realists and non-realists typically differ as to what they take the content to be: the former take it to be attributing a certain property, the latter to be expressing a certain experience. However, unless the realist can show that the property is not distinct from the property that would be referred to in the central case, and unless the non-realist can show that the experience expressed is not distinct from an experience that would be had in the central case, the danger of ambiguity will loom. A second thought might be to characterize the experience of expression by looking at its causes in the music (this is arguably the focus of Kivy (1989)). We might try to justify this approach by noting that, in the central case, we frequently account for an emotion by citing its cause: the woman is afraid because she has seen the tiger running towards her. Clearly, there is an interesting question as to what the properties of music are that cause a listener to hear it as expressive. Two further matters, however, are not so clear: first, the question of whether this is a concern for philosophy and second, whether, even if we did have an answer to our question, the nature of the experience would be illuminated. I think the answer to both of these questions is ‘no’.

98

derek matravers

The herb Puck puts on Titania’s eyes causes her love of Bottom. The identity of the cause would be a matter of empirical investigation, to which Philosophy would have nothing to contribute. Furthermore, the identification of the cause would not be an account of Titania’s feelings, nor explain what is meant by ‘What angel wakes me from my flow’ry bed?’ If this were right, however, it would appear to make mysterious the above case, in which citing the cause of the emotion (the charging tiger) did seem to throw light on the nature of the experience. The appearance is deceptive, however, as the two cases are relevantly different. In the case of the tiger, the cause of the emotion is closely related to its object. The tiger is the cause of the woman’s fear, and her fear is of the tiger. Citing the object of the emotion—the state of affairs towards which the emotion is directed (the charging tiger)—does something to account for the subjects’ feelings, and does something to explain what is meant by ‘I am scared of the tiger’. Citing the cause looks to be illuminating because we assume the cause will lead us to the object, and thus are able to grasp the object. Indeed, we are usually right to do so because the cause of an emotion usually is a belief (the tiger is charging) the content of which refers to the object of the emotion (the charging tiger). It is uncontroversial that some properties of music (including properties such as the context in which it is usually heard) cause the experience of expression in a listener, which results in the listener uttering ‘The music is sad.’ The task we have set ourselves is to throw light on the nature of the experience. The above discussion suggests finding the properties of the music that cause the experience will aid us in our task only if the cause leads us to the object. Geoffrey Madell has argued just that: dissonant properties of music arouse a desire for the resolution of dissonance, and a frustration if that is not achieved or a satisfaction if it is (2002: 28). However, the frustration at the failure to resolve, or satisfaction at the resolution of, dissonance is not yet the experience of expression (although it might be the cause of, or part of, such an experience). Madell seems to face the dilemma of either having an emotion with a bona fide object that does not amount to expression, or a case of expression but with the wrong object (2002: 128). Putting Madell’s view to one side, I cannot see that we have any reason to think that the cause of our experience of expression should also be its object. The object of the experience of expression would, presumably, have to fall within a reasonably narrow range fixed a priori. By contrast,

expression in music

99

there is no reason to suppose that there are a priori limits on what properties of music can cause the experience of expression, nor that some of those properties will not turn out to be quite unexpected, even bizarre. The right approach, then, is to characterize the experience directly. It might be thought that this is a task unsuited to philosophy. That is, the tools of philosophy, or at least, analytic philosophy, are not suited to phenomenological description. Furthermore, the task seems pointless. If I introspect, on hearing a piece of sad music, and report what it is like to have that experience, who can gainsay what I report? Matters are not quite that bad, however. In the first place, the task is not simply to describe what I experience, but to come up with a description with which everyone who has the experience can agree. This is not always an easy task, as we can see if we consider the gustatory case. It sometimes takes an expert to identify what we taste in a bottle of wine. The second constraint is that the elements that make up the description fit into a naturalistic account of the experience being caused by the music. Third, which is a special case of the second, the elements themselves must not be philosophically dubious. The second constraint is a qualification of the above claim that the search for causes is irrelevant. If a description of the experience fits the known facts about the relevant musical properties, then so much the better for it. This is compatible with the view that philosophical theories do not owe us any account of the causation itself. It is optimistic to think that an introspective awareness of the phenomenon is going to yield a unique correct description. This means that in adjudicating between competing descriptions, much weight is going to fall on the second criterion. This does pose a problem. Let us say that an account goes something like this. The elements of the music that enter into the experience are a, b and c, and the description of the experience puts those elements (and perhaps some others) into a particular configuration. Let us say that we have two competing accounts—one puts the elements into configuration A and the other into configuration B (they need not agree on the list of elements—different philosophers might have different views as to what is plausible). How are we to know which is correct? There is no independent check whether it is configuration A or configuration B that we are experiencing: both accounts would claim that what it is to experience configuration A (or B) is simply to have the experience of expression.

100

derek matravers

This is, I think, a problem and should dampen our optimism about the robustness of philosophical accounts of expression. However, any such account is no worse off in this respect than at least some accounts of representation. Consider, for example, the competing accounts of representation proposed by Kendall Walton, Richard Wollheim and Robert Hopkins (Walton 1992; Wollheim 1987; Hopkins 1998). In none of these cases is there any independent check on whether the analysis given of the experience of representation is as they say it is: their claim is, rather, that to experience the elements of their analysis in the configuration claimed is simply to have an experience of representation. The burden, in deciding between them, rests on the dubiousness (or not) of the proposed elements and the plausibility (or not) of the purported manner in which they are combined. There are a number of attempted descriptions of the experience of expression in the literature. I shall divide them into two sorts. The first sort is characterized by taking us to be imagining the music, or imaginatively conceptualizing the music, in such as way as to provide us with a reason to apply an emotion term to it. I shall examine three versions of this: those of Peter Kivy, Stephen Davies and Jerrold Levinson. The second sort describes our experience in a way that stops short of imaginatively characterizing the music. I shall examine two versions of this: the standard dispositional (or ‘arousal’ theory) and that of Kendall Walton. There is much of merit in the literature that I shall not discuss, including Aaron Ridley’s account, which is a hybrid of the two sorts, and Malcolm Budd’s account, which is similar to that of Walton (Ridley 1995; Budd 1989). In addition, I shall not discuss Roger Scruton’s important contribution (Scruton 1997). Finally, I would draw the reader’s attention to Budd’s Music and the Emotions, a work that cleared the ground of previous theories and did much to generate recent interest in this topic (Budd 1985). Peter Kivy’s The Corded Shell (later published with additional essays as Sound Sentiment) was an important early contribution to the recent debate (Kivy 1989). Perhaps because it was discovering (or rediscovering) the ground of the debate, it elides distinctions prominent in later literature. As already intimated, it does not clearly distinguish the search for the properties that cause the expressive experience from the character of that experience. Furthermore, even when focused on the latter, a number of characterizations are suggested. The most prominent is that we ‘animate’ the

expression in music

101

music, and experience it as animated (1989: 59). The notion of animation is not, however, any more immediately perspicuous than expression. Kivy’s attempt to elucidate the notion by providing examples is not, I think, successful. His principal example is a circle in which three lines are drawn that we will ‘inevitably’ see as a face. However, that is a simple case of pictorial representation, and it is unclear what is supposed to generalize to the case of expression. A further claim that he makes is that animation is the same phenomenon as ‘seeing as’ and ‘seeing in’ (1989: 172). As the comparison with aspect perception appears in several other places in the discussion of expression, it is worth a brief digression. The phenomenon of aspect perception was introduced into contemporary philosophy by Wittgenstein (Wittgenstein 1976: 193–216). He discusses various cases in which the object that we see remains unchanged, although our experience of it changes. The most well-known case is that of the ambiguous figure, ‘the duck/rabbit’, in which the same set of lines can be seen as a duck-picture and as a rabbit-picture. It is unclear, without further explanation, what in this phenomenon is supposed to be analogous to expression. It cannot be the relation between the figure and what it is seen as: in both cases, that is the relation of pictorial representation, and all (including Kivy) agree that expression is not an instance of that. It must be, rather, that the way in which we process the figure (either as a duckpicture or a rabbit-picture) has a phenomenological upshot. Analogously, the way in which we process the music has a phenomenological upshot: we hear it as expressive. However, we hardly needed the analogy with aspect perception to tell us that. We might try to provide a fuller or more illuminating description of the experience of expression than its being ‘animated’. Elements in Kivy’s work suggest that he might conceive of expression as a matter of ‘experienced resemblance’ (Kivy 1989: 53, 142). This is a notion familiar from Hopkins’ account of pictorial representation. Consider, for example, hearing a foghorn. One might experience this sound as resembling that of a whale. Hearing the music as sad might be a matter of experiencing a resemblance between the music and a person expressing sadness. This theory has at least three things to recommend it. First, it is the right sort of experience to make sense of there being a connection between music and the emotions, and to justify the utterance ‘the music is sad’. Second, it fits with a plausible account of how expression is caused. Several philosophers

102 derek matravers have claimed that what one could broadly call the ‘movement’ properties of music are isomorphic to the movement properties of natural expression. Putting the matter at its crudest, sad people tend to move slowly and this slowness is mirrored in some music. If there is such an isomorphism, that is, if certain music does resemble the natural expression of sadness, then it is a plausible that, in such cases, resemblances cause the experience of resemblance. Although the theory fits with a plausible causal story, it is not conceptually tied to it. However, the antecedent—the claim that sad music is music that resembles sad people—is difficult to defend. Such a resemblance is not necessary—the major chord differs in expressive quality from the minor chord, yet both resemble the natural expression of emotion as much as, or as little as, the other (this claim is discussed further below). Neither is resemblance sufficient—there is plenty of low, slow music that is stately rather than sad. Indeed, if there is a resemblance between the natural, non-audible, expression of emotion and music it is trivial. This does not undermine the claim that the experience of expression is the experience of resemblance. Some music is experienced as resembling the natural expression of emotion (through whatever cause), and it is this music that is expressive. If we find that such music also exhibits some isomorphism to the natural experience of emotion that can be characterized independently of experience, that would be no more than a step along the way to part of a causal account. Experienced resemblance is not, however, a good account of the experience of expressive music. The claim that a resemblance can be experienced across sense modalities makes doubtful sense. I can understand experiencing a sound (the foghorn) as resembling another sound (that made by a whale), but it is difficult to see how this generalizes to cross-modal experienced resemblance: experiencing a sound (the music) as resembling a sight (a person moving slowly). However, if we stay with a single modality experienced resemblance will fail our first test: it is not a description of the experience of expressive music on which competent listeners will agree. Sad music is not experienced as resembling the sounds made by sad people (except in certain special cases). Stephen Davies has provided an account that attempts to overcome these problems (Davies 1994: 228–40).

expression in music

103

Our experience of musical works and, in particular, of motion in music is like our experience of the kinds of behaviour which, in human beings, gives rise to emotion characteristics in appearances. The analogy resides in the manner in which these things are experienced rather than being based on some inference attempting to establish a symbolic relation between particular parts of the music and particular bits of human behaviour. (1994: 239)

An ‘emotion characteristics in appearance’ is a type: a perceivable property of a person that is the criteria by which we would judge that person to be feeling an emotion. We can use emotion terms to refer to such appearances: for example, we might say that the St Bernard dogs have ‘sad faces’, where such a comment is free from implications concerning the dogs’ psychology. Davies’s claim is that we experience expressive music in a way that is like our other experiences of emotion characteristics in appearance. I am not sure the introduction of ‘emotion characteristics in appearance’ does solve the problem identified above. That is, I am not sure whether there is a type of appearance, tokens of which are exhibited in the face of the person expressing sadness, the face of the St Bernard, and in a piece of sad music. That the sad person and the St Bernard present the same type appearance is made more plausible by the fact that in each case the token is something visible. Can, however, the same type appearance be presented as something experienced through a different sense modality altogether (Levinson 1996: 105)? Davies’s analogue of the move made above (that the ‘emotion characteristic in appearance’ presented by music is a token of an audible type—the sound characteristically expressed by a person in the throes of an emotion) is rejected as not being true to the phenomenology. Instead, Davies affirms that the expressiveness of music depends mainly on a resemblance we perceive between the dynamic character of the music and human movement, gait, bearing and carriage. (Davies 1994: 229)

However, the appearance exhibited in human movement involves change of location, whilst the appearance exhibited in the relevant sense of musical movement does not, which makes it difficult to see how they can each

104 derek matravers exhibit the same type appearance. Davies’s solution to this is to point out that ‘many widespread, mundane descriptions of non-musical phenomena’ are described in terms that apply to movement yet do not involve any change of location: time flows, stock exchange indices plunge, governments lurch to the right (Davies 1994: 235–6). We can readily concede the appropriateness of such characterizations; however, this falls short of what Davies requires. For the worry was not with whether emotion terms could be applied appropriately to ‘objects’ detected by different sense modalities, but whether there was a type of emotion characteristic in appearance that crossed the sense modalities. Davies’s examples support only the former claim. A second worry with Davies’s account is whether emotion characteristics in appearance that are centred on movement properties cover all cases. Sometimes, as in the Dies Ira of Mozart’s Requiem, it seems to do with timbre. Davies could concede that, at times, we experience the music as being of a type with vocally expressive utterance. The principle remains the same: the attribution is made on the grounds of experiencing the music as exhibiting an emotion characteristic in appearance; it needs to remain ‘emotion characteristic’ even if the type of appearance might change. However, to return to a point brought up above, what of the minor and the major chords? They seem to exhibit the same appearance, although they have different expressive properties. In a private communication, Davies has argued that combinations of notes (A-C-E) are heard within a tonal context. That combination of notes is a minor chord, the tonic of A minor, but can also feature in a sequence as part of a piece of music in C major. When it does so, it might mark a point of relative tension, but not sadness. Hence, in isolation, it is simply a combination of notes. To hear it in the minor key is to contextualize it in a certain way, and thus make it part of a (virtual) emotion characteristic in appearance. However, it is not clear why this would help. The two chord sequences (one in A minor and the other in C major) will themselves only exhibit an emotion characteristic in appearance (with respect to movement) as much (or as little) as each other and yet they will afford different expressive experiences; it is still the difference between major and minor (except this time a difference in key) that is bringing the experience about. The problem has been moved rather than solved.

expression in music

105

Jerrold Levinson has argued for a richer description of the experience of emotion: A passage of music P is expressive of an emotion or other psychic condition E iff P, in context, is readily and aptly heard by an appropriately backgrounded listener as the expression of E, in a sui generic, ‘musical’ manner, by an indefinite agent, the music’s persona. (Levinson 1996: 107)

Imagining a mechanism for expression (such as playing the piano) is not part of the experience: rather, we simply hear the music as the expression of an emotion by a persona through musical output. This relieves Levinson of any obligation to try to tie, through similarity claims, the nature of our experience of the music to the nature of our experience of the expression of emotion in the central cases. There are two ways in which we could interpret Levinson’s claim (Levinson seems to hover between the two (1996: 118)). The first would take Levinson to be attempting to throw light on the experience by providing us with a perspicuous description of it. That is, not all descriptions of the same event are equally enlightening. The same event can be a moving of my arm, a flicking the switch, an illuminating the room and a scaring of the burglar. Thus, Levinson is proposing that instead of describing our experience as hearing music or even hearing expressive music, we describe it as the hearing of the music as the unmediated expression of an emotion by a persona. The second reading is that what it is to hear music as expressive is to make the music the object of some kind of imaginative endeavour: we ‘hear the music with imagination’—that is, in however attenuated way, we intentionally imagine the music to be the unmediated expression of emotion by a persona. Each of these readings can be seen to face the same root problem. In the first case, the problem is whether the re-description is enlightening. In the case of the moved arm, some of the descriptions are enlightening because (roughly) they refer to the intention behind the action, or fit the action into a context. If we look at an example that is closer to expression, the account of pictorial representation in terms of ‘seeing in’ is only enlightening to the extent that we have an understanding of seeing something in something else that is broader than, and prior to, its application to the pictorial case. In the second case, the problem is whether my grasp of the project of imagining

106

derek matravers

the music to be an unmediated expression of emotion by a persona is independent of my grasp of expressive music. If it is not, the claim that I engage in the project of imagining the music to be the expression of emotion by a persona would add nothing to my understanding of the experience of expressive music. This is related to the problem expressed above. If we ask what the experience of music as the direct, unmediated, expression of emotion by a persona would be like, or how we would know if we had succeeded in our imaginative project, Levinson would—I assume—say that the experience is like the experience of expressive music and we would know if we had succeeded were we to have that experience. We have no check as to whether we have succeeded in our imaginative project independent of our hearing the music as expressive. Is my grasp of what it would be for an experience of instrumental music to be the unmediated expression of emotion by a persona independent of my grasp of the experience of expressive music? Levinson argues that it is: [T]he extension of ‘sounding like’ beyond the range of actual human behavioural expression, in the form of music sounding like, or as if it were, an alternate, specifically ‘musical’ mode of expression of emotion, is both natural and imaginatively unproblematic, suitably understood. (1996: 116)

This seems to me unduly optimistic. If the account was that expressive music was music that was heard as if it were within the range of actual human behavioural expression of emotion, it would be enlightening were it true—although obviously it is not true. The problem of expression surely is just that we do not hear expressive music as falling within the range of actual human behavioural expression of emotion. This is a version of an objection considered by Levinson, owed to Hubert Eiholzer. To this objection Levinson argues for two possible responses. The first is to ‘maintain that listeners need not know, in any detailed fashion, what music’s being a natural mode of expression would amount to but only be disposed to posit, in imagination, that music was such a mode’ (1996: 120). The second is that we drop the claim that the music is the form of expressing, and instead maintain that the experience is of expressing tout court on which, if we reflect upon it, we would decide that it constituted a sui generis mode of expressing (1996: 121). Both responses retreat from the claim that music is experienced as a sui generis mode of expression.

expression in music

107

However, as it is the experience we are trying to characterize, it is unclear that Levinson’s account can survive intact (to be fair, Levinson allows that both responses are a concession). The first seems to involve a dilemma. Either our positing, in the imagination, that music is a natural mode of expression results in our experiencing it as a natural mode of expression or it does not. If it does, then Eiholzer is not answered: we still need some grasp of what it is to hear music as a natural mode of expression. If it does not, then Levinson will have failed to capture what it is to experience music as expressive, for it is not to experience music as not expressive, and to imagine, of it, that it is the natural mode of expression. The second response aims to remove the music as the vehicle of the persona’s emotion from the experience, and place it (if it is to appear at all) in the listener’s later reflection. All that remains of the attempted elucidation of the content of ‘the music is sad’ is that it is ‘readily and aptly heard by an appropriately backgrounded listener as the expression of sadness’. Levinson might object that this neglects the claim that further reflection will drive the listener to the conclusion that the mode of expression is the sui generis mode of expression by a persona. However, this raises an issue that threatens to undermine the entire account. What are the grounds of Levinson’s claims? How does he know that further reflection will take this form? More generally, how does he know an appropriately backgrounded listener will hear the music in the way he describes? He might argue that anyone who is familiar with the experience of expressive music will accept the description. This again seems unduly optimistic: according to Stephen Davies at least, ‘it seems straightforwardly false that there are public conventions or consensus regarding these matters’ (Davies 1997: 101). This does seem to be a problem: Levinson is giving a naturalistic account of the claim that music is sad, in terms of how an appropriately backgrounded listener would hear the music. If, however, the backgrounded listener claims not to hear the music like that, it is unclear what the resources are to which Levinson could appeal in order to convince him. The theories I have considered so far all claim that we experience the music as, in some way, appropriate for description in emotion terms. It is experienced as resembling the expression of emotion, or as manifesting the appearance of the expression of emotion, or as emotion expressed in a musical manner. There is, however, another way of answering this question: that sad music is music that makes us sad. Stated in this bald

108

derek matravers

form, the (so-called) ‘arousal theory’ has been the subject of much scorn. In retrospect, however, we can see that it got two points roughly right. First, let us consider a slightly sophisticated version: a passage of music expresses an emotion if, among the mental states caused by the music, is some non-cognitive state (a feeling) which stands in the right relation to the appropriate reaction to the expression of that emotion in the central case (Matravers 1998: 146). To give an example, music expresses sadness if, among the mental states it arouses is something like the feeling component of the normal reaction to the expression of sadness in the central case. The first point it got roughly right is that it never concerned itself with the causation behind this. That is, it had no commitments as to the nature of the properties that caused the experience of expression. The second point it got roughly right was to focus on the nature of the experience of expression. Both of these, I have argued, are the right stance for a philosophical theory to take. The question, then, is whether the account of the experience is correct. On the face of it, it appears foolish to equate the experience of music as sad with a conjunction of the experience of music and an experience of sadness. I will return to that after considering two more obvious problems. An aroused feeling is not necessary in that some people appear to experience expression whilst claiming not to be experiencing the relevant feeling. It is not sufficient in that plenty of things arouse our feelings, without thereby being expressive. I cannot fully enter into the matter here, but a defence of the arousal theory would run along the following lines. First, the claim that one was not experiencing the relevant feeling may be compatible with experiencing some kind of appropriate non-cognitive state. That is, the listener might be disavowing the mental state usually referred to by the term ‘feeling’, or terms such as ‘sad’ or ‘joyous’. Those terms are usually used in a context in which the mental state is dominant: one does not usually claim to have those mental states unless they are strongly present in experience. The arousal theorist does not require such strong presence, only the presence, among the mental states caused by the music, of some non-cognitive state that stands in the right relation to the appropriate mental state in the central case. There is a danger here of reducing the feeling to a cipher, in which case it loses any explanatory weight. A balance needs to be struck between having little enough of the non-cognitive state

expression in music

109

to make sense of the listener claiming that he is not experiencing a feeling, but enough of it to play the required role in the account. To rebut the claim that the aroused non-cognitive state is not sufficient, the arousal theorist needs to pick out a psychological role filled only by those aroused feelings relevant to expression. In short, the listener experiences the feeling as ‘tracking’ the music; it varies as the music varies. This is also a more plausible account of the experience of expression than the simple conjunction proposed above. I say ‘more plausible’, although for most it would not be plausible enough, for the analysis remains in terms of (at least) two separate mental states: the music and the aroused non-cognitive state. The question is how this gap is to be closed. It is not clear whether this is a problem and, if it is a problem, it is not clear whether it is peculiar to the arousal theory. The analysis provided is in terms of two separate mental states (the music and the non-cognitive state) which we do not experience as separate. The relation between the terms of a philosophical analysis of an experience and its phenomenology is not, however, obvious. The two separate mental states are co-instantiated in consciousness in the particularly intimate causal relation of the one tracking the other. Is it absolutely clear that this could not be an account of the phenomenology—especially given the fact that so many listeners attest to there being some effect on their feelings and emotions by expressive music? In other words, these are elements we have anyway—it is obvious that this is not the way in which they are arranged. Kendall Walton has proposed a more nuanced account that shares with the arousal theory the thought that the solution to the problem lies not in imagining the music to be a certain way, but in the way in which our response to the music is characterized: I propose that, although music does not in general call for imaginative hearing or imaginative perceiving, it often does call for imaginative introspecting. We mentioned the possibility that music is expressive by virtue of imitating behavioural expressions of feeling. Sometimes this is so, and sometimes a passage imitates or portrays vocal expressions of feelings. When it does, listeners probably imagine (not necessarily consciously and certainly not deliberately) themselves hearing someone’s vocal expressions. But in other cases they may instead imagine themselves introspecting, being aware of, their own feelings. (Walton 1988: 359)

110 derek matravers In a later paper, he elucidates this a little: Anguished or agitated or exuberant music not only induces one to imagine feeling anguished or agitated or exuberant, it also induces one to imagine of one’s auditory experience that it is an experience of anguish or agitation or exuberance. (Walton 1994: 55)

Let us grant that claim that some music is expressive by virtue of imitating behavioural expressions of feeling and focus on the claim that what it is to experience expression is for the hearer to imagine of his awareness of auditory sensations, that it is an introspective awareness of states of mind. I say this is in the arousal tradition, because, first, there is no commitment to any causal account of how this state comes about, and second, because we are not imagining of the music that it be appropriate for description in emotion terms. This account faces a familiar difficulty. Is our grasp of what it would be to imagine of one’s auditory experience that it is an experience of anguish or agitation or exuberance independent of our grasp of the experience of expressive music? Is there any independent check that it is this, rather than any other imaginative project, that we are pursuing? One problem with Walton’s account (a general problem with a number of Walton’s accounts, and one of which he is aware) is the absence of a robust account of the imagination. In answer to the question of whether this is the way an appropriately backgrounded listener would hear the music, I would be inclined, like Stephen Davies, to scepticism: there is simply no consensus on these matters. For Jerrold Levinson, the main problem for Walton’s account (which looks as if it will generalize to being a problem for the arousal theory) is that it ‘casts the activity of perceiving musical expressiveness in too egocentric a light; it represents expression in music as in effect the expression of the listener’s own, albeit imaginary, feelings’. By contrast, Levinson maintains, ‘expressiveness in music ... is something we encounter fundamentally as residing ‘out there,’ as existing anterior to our own minds’ (Levinson 1996: 94). The argument, however, is double-edged as we can hear music as being either an expression of emotions from the outside, or as a vehicle for our own emotion. The distinction was marked by R. K. Elliott concerning lyric poetry, in a way that generalizes to music: If a work is experienced as expression, experiencing it from within involves experiencing this expression after a certain imaginative manner as one’s own.

expression in music

111

Experiencing it from without is experiencing it as expression, but not experiencing this expression as if it were one’s own. (1967: 146)

Walton might claim that hearing music from without is already accounted for in his theory by his admission that sometimes ‘listeners imagine (not necessarily consciously and certainly not deliberately) themselves hearing someone’s vocal expressions’, reserving the account discussed here for hearing from within. What of Levinson’s own account? As the experience is one of hearing the music as the expression of a persona, can it account for hearing music from within? Levinson’s definition of expression states a necessary as well as a sufficient condition. Thus, it seems as if he is committed to expression being the experience of an emotion being expressed by someone else (the musical persona). This, however, seems contradicted by claims Levinson has made in an earlier paper (‘Music and Negative Emotion’): If one begins to regard music as the expression of one’s own current emotional state, it will begin to seem as if it issues from oneself, as if it pours forth from one’s innermost being. (1982: 328)

Is the claim that our experience can be of the music seeming to ‘issue from oneself’ compatible with the claim that, necessarily, to hear music as expressive is to hear it as being expressed by the musical persona? The answer to this depends on the relation between the listener and the musical persona. In explicating what it is to hear music from within, Levinson could construe the relation in at least two ways. The first would be that the listener experiences the music as described in the later essay, but then empathizes with the persona and takes on, at least in imagination, the persona’s emotion (1982: 327). That is, the relation is one of ‘identifying with’ the persona in the way in which one might, for instance, identify with a friend who is experiencing some emotion. The second would be that the relation between listener and persona is one strict identity: we experience the music as a sui generis expression of our own emotion and there is no further figure involved. Which of these two options should we favour? The issue depends on whether hearing from within and hearing from without are on a par, or

112

derek matravers

whether we take hearing from without to be primary. Levinson would appear to favour the latter, in which case he would take the first option. If so, the claims in the two essays are compatible. If we take hearing from within and hearing from without to be on a par, and thus take the second option, the claims in his two essays would not be compatible. If this were the case, Levinson would need to weaken the definition in ‘Musical Expressiveness’ to the claim that it captured only one (albeit important and central) experience of expression. Empathy and sympathy look to be a resource to help the arousal theorist make the distinctions necessary here. Recall the version of the arousal theory considered above: a passage of music expresses an emotion if, among the mental states caused by the music, is some non-cognitive state (a feeling) which stands in the right relation to the appropriate reaction to the expression of that emotion in the central case. The appropriate reaction to the expression of emotion can take two forms. Consider a situation in which you are faced with someone expressing sadness. One reaction would be empathize—that is, to adopt the type of emotion being expressed—that is, to feel sadness. Another would be to sympathize—that is, to feel an emotion that responds to the emotion being expressed. One virtue of the arousal theory is its simplicity: amongst the mental states aroused by the music is some non-cognitive state. For this to help the arousal theorist, he would first have to make sense of the notion that the non-cognitive state aroused could be either empathetic or sympathetic, and then show how this distinction could solve the problem of experiencing expression from within or without. The first looks to be a problem because the difference between empathy and sympathy is not in the quality of the phenomenological state, but in the context. A necessary condition for empathy, for example, is that the observer experiences the same emotion as the expresser, which is difficult to accommodate in the musical case. Can we be sure, however, that the phenomenology of empathy does not differ from that of sympathy? Does the sadness we feel towards someone feel the same as the sadness we feel with someone? It is not clear that this is the best way to put the question or even, put like this, where we might look for an answer. Much work in the theory of the emotions recently has been to explore the complexities of the intentionality of feelings. Animals and pre-linguistic infants can certainly direct their attention on things outside them, yet we do not attribute to them the kind of cognitive apparatus required by traditional accounts of

expression in music 113 intentionality (Deigh 1994). It might be that there is a type of expressive music that arouses a feeling that feels as if it is directed on objects, and a type of expressive music that arouses a feeling that feels like it is our own feeling (although brought about in virtue of the expression of others). The second issue is more difficult to deal with, if only because of the complexity of the phenomenon to be explained. The arousal theory being explored relies on there being a ‘right relation’ between the non-cognitive state aroused, and the feeling state involved with the response to expression in the central case. However, there is nothing internal to the arousal theory to restrict the right relation to central-case emotions felt to other people (it is only that hearing from without is the usual case). There might be cases in which the right relation is to central case emotions felt ourselves. The point, put at its most general, and covering both the issues raised, is that it is not clear that the feeling components of emotions are simple and homogenous. What sadness feels like might differ depending on whether it is our sadness per se, an empathetic sadness or a sympathetic sadness. Whatever complexity that can be found can be used by the arousal theorist to add complexity to the response to the music. This is, of necessity given the shakiness of a science of feelings, speculative. However it has some intuitive force and is the obvious area for any dispositional theory to explore. I turn now to a problem for all of these accounts, and that is normativity. Each of these could be an account of what it is for a listener to experience music as expressive. However, what they aspire to be is what it is for music to be expressive. A passage of music is expressive if it is appropriate that it bring about a certain experience. The most plausible way of meeting this aspiration seems to be to use some version of that empiricist standby, the qualified observer in the right perceptual circumstances. However, even here there are grades of difficulty. For example, one could give up claims to truth or falsity, and instead rely on the fact that some music is rewarding when heard in one way rather than another. Unless one is secure in one’s own judgement about which music is to be heard how, one will be open to guidance from ‘expert’ listeners. Terms such as ‘the appropriate way to listen to this’ will have a natural role in that discourse. The stonier road is trod by the realist, who needs to provide determinate truth conditions to underlie the claim that sad music is music that possesses the (real) expressive property of sadness. For the realist, which would seem to include Levinson (2001), the judgement of the ‘appropriately backgrounded listener’ has a

114

derek matravers

constitutive role. The objection (here made to Levinson in particular) has been forcefully put by Roger Scruton: The invocation of a ‘reference class’ of listeners opens the way to a radical scepticism: how is this class to be defined and by whom? The natural gloss on Levinson’s definition is to identify the reference class as the class of those who are able to discern the expressive content of a work of music. But this would be to reduce the definition to vacuousness. (Scruton 1997: 353)

Apart from reducing the commitment to realism, and accepting the consequent restrictions on the aspirations of criticism, I can see no way around this problem. As it is hardly a problem for expression in particular, and as it does nothing to help us weigh one account against another (as opposed to the further commitments of advocates of the accounts) I shall say no more about it. What will help in weighing the accounts against each other is their plausibility in capturing the phenomenon, whether we have a grasp of the elements of the accounts independently of expression, and whether these elements can be fitted together in a perspicuous way. It is not obvious to me that the dispositional account is in a worse position than that of its rivals.¹

References

Budd, M. (1985) Music and the Emotions, London: Routledge and Kegan Paul. Budd, M. (1989) ‘Music and the Communication of Emotion’, Journal of Aesthetics and Art Criticism 47 2: 129–37. Davies, S. (1994) Musical Meaning and Expression, Ithaca, NY: Cornell University Press. Davies, S. (1997) ‘Contra the Hypothetical Persona in Music’, in M. Hjort and S. Laver (eds), Emotion and the Arts, Oxford: Oxford University Press, 95–109. ¹ Stephen Davies and Jerrold Levinson were (as usual) generous and helpful in their comments on this essay. It has also been much improved by suggestions from the editor. Jerrold Levinson addresses the problems raised in the final section (on normativity) in an exchange with Stephen Davies in Kieran (2005) that came out too late to be dealt with in this paper.

expression in music

115

Deigh, J. (1994) ‘Cognitivism in the Theory of Emotions’, Ethics 104 4: 824–54. Elliott, R. K. (1967) ‘Aesthetic Theory and the Experience of Art’, in H. Osborne (ed.), (1972), Aesthetics, Oxford: Oxford University Press. Hopkins, R. (1998) Picture, Image and Experience, Cambridge: Cambridge University Press. Kieran, M. (ed.) (2005) Contemporary Debates in Aesthetics and the Philosophy of Art, Oxford: Blackwell. Kivy, P. (1989) Sound Sentiment: An Essay on the Musical Emotions, Philadelphia, Penn.: Temple University Press. Levinson, J. (1982) ‘Music and Negative Emotion’, in Music, Art and Metaphysics: Essays in Philosophical Aesthetics (1990), Ithaca, NY, and London: Cornell University Press, 306–35. Levinson, J. (1996) ‘Musical Expressiveness’, in The Pleasures of Aesthetics, Ithaca, NY: Cornell University Press, 90–125. Levinson, J. (2001) ‘Aesthetic Properties, Evaluative Force and Differences of Sensibility’, in E. Brady and J. Levinson (eds), Aesthetic Concepts: Essays After Sibley, Oxford: Oxford University Press, 61–80. Madell, G. (2002) Philosophy, Music and Emotion, Edinburgh: Edinburgh University Press. Matravers, D. (1998) Art and Emotion, Oxford: Oxford University Press. Ridley, A. (1995) Music, Value and the Passions, Ithaca, NY: Cornell University Press. Scruton, R. (1982) Art and Imagination: A Study in the Philosophy of Mind, London: Routledge and Kegan Paul. Scruton, R. (1997) The Aesthetics of Music, Oxford: Oxford University Press. Walton, K. (1988) ‘What is Abstract about the Art of Music?’, Journal of Aesthetics and Art Criticism 46 3: 351–64. Walton, K. (1992) ‘Seeing-In and Seeing Fictionally’, in J. Hopkins and A. Savile (eds), Psychoanalysis, Mind and Art: Perspectives on Richard Wollheim, Oxford: Blackwell, 281–91. Walton, K. (1994) ‘Listening with Imagination: Is Music Representational?’, Journal of Aesthetics and Art Criticism 52 1: 47–62. Wittgenstein, L. (1976) Philosophical Investigations, Oxford: Blackwell. Wollheim, R. (1987) Painting as an Art, London: Thames and Hudson.

116

derek matravers

Further Reading

Budd, M. (1985) Music and the Emotions, London: Routledge and Kegan Paul. Davies, S. (2006) ‘Artistic Expression and the Hard Case of Pure Music’, in M. Kieran (ed.), Contemporary Debates in Aesthetics and the Philosophy of Art, Oxford: Blackwell, 179–91. Kivy, P. (1989) Sound Sentiment: An Essay on the Musical Emotions, Philadelphia, Penn.: Temple University Press. Levinson, J. (1996) ‘Musical Expressiveness, in The Pleasures of Aesthetics, Ithaca, NY: Cornell University Press, 90–125. Levinson, J. (2005) ‘Musical Expressiveness as Hearability-as-Expression’, in M. Kieran (ed.), Contemporary Debates in Aesthetics and the Philosophy of Art, Oxford: Blackwell, 191–204. Scruton, R. (1997) The Aesthetics of Music, Oxford: Oxford University Press.

5 Explaining Musical Experience PAUL B O G H O S S I A N

Music and the Emotions 1 I start with the observation that we often respond to a musical performance with emotion—even if it is just the performance of a piece of absolute music, unaccompanied by text, title or program. We can be exhilarated after a Rossini overture brought off with subtlety and panache; sombre and melancholy after Furtw¨angler’s performance of the slow movement of the ‘‘Eroica’’. And so forth. These emotions feel like the real thing to me—or anyway very close to the real thing. When one experiences them, it takes time for them to wear off, and one gets irritated with the companion who, because not similarly moved, wants immediately to start discussing where to go for dinner. Like many others, I am drawn to the philosophy of music by a need to understand how such emotional responses are possible. How can absolute music move us in the way that it does, and to the extent that it does? 2 In seeking an answer to this question, we do not just seek any sort of answer. For example, we would not be satisfied with a brute physiological explanation along the following lines: In listening to music, we are exposed to sounds. Sounds are vibrations in the air. These vibrations cause our ear drum to vibrate, which in turn causes nerve impulses to travel up the auditory nerve to the brain. In the brain, these nerve impulses cause certain neurons to fire, leading to the perception of sound. In certain cases, the firing of the

118

paul boghossian neurons that constitute the perception of sound triggers the firing of certain other neurons, leading to the experience of specific emotions.

I take it that, even if an explanation along these lines is actually true, it is not the sort of explanation we are looking for. Why not? What is missing from the brute physiological explanation just outlined? 3 What is missing, I think, is any explanation of the rationality of our emotional response. For all that the brute explanation cares to claim, music might induce emotional states in us in just the way that a drug might: certain chemicals cause us to feel things, and so do certain sounds. End of story. On this view, it is pointless to ask whether it makes sense for us to respond to those sounds in the way that we do, in just the way in which it is pointless to ask whether it makes sense for us to respond to Prozac or to marijuana in the way that we do. And yet we think that it does make sense for us to be moved by music, that it isn’t just a matter of a chemical response. In fact, we not only think it’s rational to be moved in this way, we are especially admiring of those who are capable of the right emotional response, and critical of those who aren’t. We take the presence of the right emotional response to be indicative of understanding. We recommend music appreciation classes to those who profess not to see what the fuss is about. So the question becomes: how could we explain the rationality of our emotional response to music? 4 Peter Kivy thinks that he can justify a minimal sense in which it can make sense to be moved by absolute music, and that is the sense in which one can be moved by the sheer beauty of the music, full of appreciation for the excellence of its craftsmanship and so forth (Kivy 2001: 92–118). But that doesn’t come close, in my view, to explaining the diversity of feeling that is aroused by our experience. Think of how differently you feel after listening to the witty elegance of the Mozart Piano Concerto K. 271

explaining musical experience 119 as opposed to the brooding gloom of the Tchaikovsky Sixth Symphony. They are both examples of craftsmanship and beauty. 5 I think our only hope of vindicating the range of emotional responses that we have to absolute music depends on our being able to see absolute music as telling us things, in much the way in which we see characters in opera or fiction as telling us things. In other words, we need to be able to assimilate the problem of absolute music to the problem of ‘feeling for fiction’ by finding meaning in absolute music. That is not to deny that there is a genuine problem of explaining the rationality of our feeling for fiction. But everyone is agreed that there must be a solution to that problem. Kivy says: The question I am raising is how we are emotionally aroused by what the nineteenth century called absolute music ... It is important to remember this because when the resources of language are added to the musical work, the terms of the argument are radically changed. I have no quarrel, for example, with someone who says that when he attends a performance of La Traviata, he experiences real and intense sorrow over the death of Violetta, ... This is not to say that there is no philosophical problem in just how the emotions of sorrow and love can be aroused by the fates of fictional characters ... But the presence of language, with all its potential for conveying concepts, and the presence of delineated characters, such as Violetta and Alfredo ... provide materials for arousal of garden variety emotions far exceeding anything that can reasonably be postulated in absolute music. And that is why absolute music poses a problem far beyond opera, oratorio, song and programme music to those who wish to claim that it arouses the garden variety emotions. (Kivy 2001:101–2)

It makes sense to be moved by opera, Kivy tells us, because characters in opera can tell us sad things. The problem for absolute music, he says, is that, lacking language, title or text, it can’t tell us anything; a fortiori, it can’t tell us anything sad, happy or whatever. Hence, it cannot make sense to be moved by absolute music. My thought is a modus tollens on Kivy’s. Since it so obviously does make sense to respond to music with emotion, there must be a sense in which music is capable of telling us things. The question is how.

120

paul boghossian

6 The fact that we can rationally respond to music with real emotions, I have been saying, is indirect evidence that there must be musical meaning. Not so much, perhaps, representational meaning—propositions that tell us how things are—though some, myself included, would be prepared to allow a limited role for such meanings; but, rather, expressive meaning, the capacity, principally, to express emotions. But an inference from the legitimacy of our emotional responses is not the only sort of evidence that we have for the existence of musical meaning. Let me mention three other sources. First, there is the sheer phenomenology of listening. Any reasonably experienced listener will hear the Tristan Prelude as suffused with the expression of yearning, whether or not they knew anything about the opera to follow, or even whether they knew that there is an opera that follows. A second important source for our conviction that music possesses audible expressive properties derives from our evaluative thinking about music. We value some pieces over others because of what we take to be their greater expressive power. Derivatively, we value some musicians over others for their greater ability to unlock that expressive power. (A criticism that is often made, rightly in my view, of a famous New York opera conductor, is that he routinely sacrifices expressive meaning in favor of lush beauty, leading to performances that may be marvelously sensuous but are otherwise superficial and unsatisfying. We understand this criticism.) A final source for our conviction in the expressive capacity of music derives from the role that music plays in opera or film. Clearly, music enhances the expressive impact of opera. The ceremony of the grail at the end of Parsifal wouldn’t be half as powerful as it is if it were accompanied by ‘‘Baa Baa Black Sheep’’, would it? But how could music contribute in this way to the opera’s expressive capacity if it didn’t have an expressive capacity all its own that it was bringing to the scene? Philosophers routinely concede the expressive impact of opera. But part of that expressive power is provided by the music, something it couldn’t do unless it had some expressive power all its own to contribute.

explaining musical experience 121

Explaining Musical Expression 7 In the final analysis, the real problem for musical meaning is not to justify belief in its existence, but to explain its possibility. How is it possible for mere sound, lacking speaker intention, or any of the other resources which make linguistic meaning possible, to express meanings? What properties of the sounds could constitute their expressive capacity? Clearly, we would be looking at the musical properties of the sounds, properties like that of pitch, rhythm, harmony and melody. Sounds are heard as having the expressive properties they have because they are heard as having certain musical properties: it is something about the shape of the melody which opens the fourth movement of Tchaikovsky’s Sixth Symphony that makes us hear it as sombre. So what we are asking is: how could possessing certain musical properties amount to expressing a state of mind? Roger Scruton (1999) has argued that, in a sense, there is nothing to explain. His argument is distinctive and it will be worthwhile to linger over it. Scruton’s starting point is the observation that mere sound is not the intentional object of musical perception. Even when sound is understood not merely as a physical phenomenon but as a ‘‘secondary object’’—a mental object—which exists only when it is heard, it is not what we hear when we hear music. The intentional object of musical perception is rather tone, where tone is characterized by such musical variables as pitch, rhythm, harmony and melody. But nothing can literally have such properties and so nothing can literally be a tone. A melody must rise or fall. But there is nothing either out there or in here that rises or falls in the way that a melody does. As applied to sound, therefore, he concludes, the concepts of pitch, rhythm, harmony and melody are metaphorical. To describe music, says Scruton: we must have recourse to metaphor, not because music resides in an analogy with other things, but because the metaphor describes exactly what we hear, when we hear sounds as music. (Scruton 1999: 96)

And the use of any metaphor cannot ultimately be explained.

122

paul boghossian

It does not seem strained to suggest that Smetana’s music expresses the shining and silken qualities that we hear in it. Smetana’s music is not literally shining or silken. But its expressive power is revealed in its ability to compel these metaphors from us, and to persuade us that they fit exactly. Of course, it is a mystery that they fit. But the mystery is immovable. Every metaphor both demands an explanation and also refuses it, since an explanation would change it from a metaphor to a literal truth, and thereby destroy its meaning. (Scruton 1999: 141)

8 Let me first get rid of the distracting claim, that Scruton ultimately takes back, that sounds are not the intentional objects of musical experience. Surely, we hear sounds when we hear music! In an afterthought to his chapter on ‘‘Tone,’’ Scruton shows some awareness of this. He says: Finally, we should not think of sounds and tones as distinct individuals—as though tones really existed apart from sounds. Perhaps the best way of understanding the relation between the two is in the way Spinoza understood the relation between mind and body. For Spinoza reality can be conceptualized in two ways: as mental or as physical. But that which we conceptualize in these two ways is one. (Scruton 1999: 79)

Invoking Spinoza’s obscure account of the relation between mind and body is not likely to help here. If sounds and tones are the same individuals, then sounds are equally the intentional objects of musical experience. At any rate, it is intuitively overwhelmingly clear that sounds are heard when music is heard. (Imagine Dick Cheney dressed up as Santa Claus to entertain (scare?) the children at a White House Christmas party. Is Dick Cheney the intentional object of the children’s experience? Santa Claus? Surely, the right answer is: Dick Cheney, seen as Santa Claus.) If we set the strange claim about intentionality to one side, we are left with three other theses, which jointly ground Scruton’s claim that there is no explaining musical expression: First, the thesis of musical anti-realism: sounds do not literally have the musical properties we hear them as having. Rather, and this is the second thesis, the musical and expressive descriptions under which we hear sounds when we hear them as music are metaphors. And, finally, although it is true that metaphors can be more or less apt,

explaining musical experience 123 there is no explaining any particular metaphor’s aptness. Every metaphor both demands an explanation and also refuses it. I think all three theses are deeply problematic, though here I will settle only for questioning the second.¹ Suppose we agree with Scruton that the musical descriptions under which we hear sounds are literally false. Still, is it really appropriate to call these perceptually exercised descriptions metaphors? I would have thought that it is essential to metaphor that our use of it be intentional. Having experienced something and wishing to illuminate some truth about it, we intentionally use a false description to bring that truth to light. ‘‘Jimi’s on fire’’ we say, for example, about a particularly exciting performance by Jimi Hendrix.² But could that really be what’s going on when I hear the theme that launches the Goldberg variations? Do I have an experience of the sounds qua sounds, notice something about them and then intentionally decide to illuminate what I heard by hearing them as constituting that famous melody? That is surely absurd, for two reasons. First, the tendency to hear it as that melody is as far from an intentional mental act as anything gets. It is completely ineluctable. Scruton mentions several examples of musical perception that do seem subject to the will. For example, the four opening drum beats of the Beethoven Violin Concerto can be heard either as part of the ensuing melody or just a preamble to it. But if his thesis were right, it would have to hold quite generally; any musical metaphor would have to be optional from the standpoint of the will. But they don’t seem to be. Second, if the use of the notion of metaphor is to be justified, we would have to be aware of some layer of musical experience with a perfectly literal content that our musical metaphors would be designed to illuminate. But there doesn’t seem to be such a layer of experience; for the most part, our experience is always already musical. If this is right, then we can no longer ground the claim that there can be no explaining music’s expressiveness by appealing to the general claim that there can be no explaining any metaphor’s aptness. At any rate, the anti-explanatory thesis is independently implausible. The point is that the expressive properties of music are clearly grounded in its purely musical properties. It is because a passage has certain musical ¹ This represents a change of heart about musical anti-realism, to which I had been inclined to be more sympathetic. See Boghossian (2001). ² The example is Stephen Yablo’s, though not the point it is being used to illustrate.

124

paul boghossian

properties that it is heard as having certain expressive properties, as Scruton in effect acknowledges in discussing the ‘‘Todesklage’’ from Act 2 of Wagner’s Die Walk¨ure: This theme contains a tragic, and yet questioning expression. It is a normal exercise of the critical intelligence to look for the features which are responsible for so powerful an effect: the accumulated suspensions, and the final Neapolitan cadence finishing on a seventh chord, with its ‘‘unsaturated’’ and yearning character. Remove the suspensions and the tension goes. Alter the final cadence and we have (with a slight change of rhythm) the serene introduction to Mendelssohn’s Scottish Symphony in A minor, Op. 56. (Scruton 1999:164–5)

Even if the musical properties are not themselves real properties of the sounds, but are only experienced as though they are real properties of the sounds, there would still be the question: how could the fact that sounds are heard as having certain musical properties explain why they are heard as having certain expressive properties? So, even if we were to concede the musical anti-realism, that would not invalidate the question with which this section began: How is it possible for mere musical sound, lacking speaker intention, or any of the other resources which make linguistic meaning possible, to express meaning? 9 The view I am most inclined to favor is a version of the resemblance theory, a view that has been very ably developed by Malcolm Budd (1995) and Stephen Davies (1994). I would put it like this: A passage P is expressive of E just in case P sounds the way a person would sound who was expressing E vocally, or sounds the way a person would look who was expressing E gesturally. This kind of account leans, of course, on the idea that persons have characteristic ways of expressing their inner states. For example, there are typical—and typically natural—ways that sad people look, typical ways that they sound, and so forth. The account I favor, then, seeks to explain a passage’s being expressive of sadness by equating this property with the property of sounding like the way sad persons sound (leaving out for now the difficult cross-modal case).

explaining musical experience 125 Jerrold Levinson (2005) has an interesting objection to this idea. If you say that sad music is music that resembles sad sound, he observes, you have to say to what degree, since everything resembles everything else to some degree. But the degree of resemblance required cannot be specified in terms of some fixable degree of resemblance between the two. It can only be specified as whatever resemblance is sufficient to induce appropriate listeners to hear the music as sad. Levinson’s conviction that the resemblance relation cannot be spelled out in substantive terms—in terms which avoid what Crispin Wright has called, in a different connection, a ‘‘whatever it takes formulation’’—lies behind his own dispositional account of musical expression: P is expressive of E iff P, in context, can readily be heard, by a listener experienced in the genre in question, as an expression of E. (Levinson 2005)

He tentatively glosses a listener readily hearing P as an expression of E as: the listener readily imagining that P is an expression of E. I think this is a challenging argument; but I would like to resist it because I have qualms about the dispositional account of musical expressiveness to which it seems inexorably to lead. That dispositional account is itself a version of a no-explanation view of musical expressiveness. 10 Most of the objections that worry me are explicitly considered by Levinson; but I don’t think they have been sufficiently answered. First, I think the gloss—hearing P as an expression of E in terms of imaginings—is ill-advised. It’s just too easy to imagine a particular passage as being the expression of a variety of different—even contradictory—emotions. Imagining is too unconstrained and so the resulting content too indeterminate. Second, Levinson’s definition is of ‘‘P’s being expressive of E.’’ This is a property that a passage can literally have. One of the motivations for thinking that music has such expressive properties is, as we have seen, that we seem to be able to hear them in the music. But can our hearing reveal to us that a particular musical passage is expressive of sadness, for example, on Levinson’s view? To do so, it would seem, our hearing would have to reveal to us that P is a passage that has the dispositional property to induce

126

paul boghossian

listeners who are appropriately versed in the genre to hear the passage as though it were a literal expression of sadness. I don’t see how that could be an audible property of the music. Levinson has a reply to this objection. He says that qualified listeners tacitly assume that they are indeed qualified listeners, and so, for such listeners, hearing an expression of sadness in a passage becomes tantamount to hearing the passage’s being expressive of sadness. Qualified listeners arguably tacitly assume, while listening to music that they are qualified listeners, listening appropriately. Thus on the view I favor, hearing an expression of E in a stretch of music becomes, for such listeners, tantamount to hearing the music’s expressiveness of E. All a qualified listener need do to readily hear the expressiveness of the music is to readily hear expression of emotion in it. (Levinson 2005)

Whether this reply is adequate depends on what one takes the relevant datum to be. What are we claiming when we say that an appropriately qualified listener can just hear the sadness in the music? On a strong interpretation, it is to say that an appropriately qualified listener can know, on the basis of hearing alone, that the music is expressive of sadness, in just the way in which it is natural to say that an appropriately qualified taster can know, on the basis of taste alone, that a particular substance is sweet. On a weaker interpretation, the claim would be not that a listener can know on the basis of hearing alone that a particular passage expresses sadness, but only the psychological claim that most listeners will feel no need to make an overt inference from their hearing to the attribution of expression. Levinson’s observations can accommodate the weaker claim but not the stronger one. However, it is arguable that it is the stronger claim that needs respecting. As a matter of psychology, there may be no inference when an experienced observer sees some dark clouds and believes that it will rain. But as a matter of justification, the knowledge would have to be based on inference. Something is a perceptual justification of the belief that P just in case transitions from perceptions as of P (to the extent that there are any) to the belief that P are justified, without need of additional inference. Levinson’s listener is like the experienced observer, moving swiftly from his hearing to the attribution of expression. But he cannot be said to know, purely on the basis of hearing, that a particular passage has a certain expressive property.

explaining musical experience 127 Arguably, though, I never need to infer that some music I hear is expressive of sadness, let alone an inference that proceeds from some complicated premise about how others are likely to react. Rather, I hear the expression in the music much as I taste the sweetness in the wine. Third, I worry that Scruton’s objection that reference to the appropriate class of listeners is doomed to vacuity will stick. There is a possible world in which there are incredibly sophisticated musicians who write highly mathematically intricate music, perhaps a` la Babbitt, but who profess to hear no expression in any of the nineteenth century music we play them. They have emotions, so they are not Spocks (as in Star Trek). They express those emotions as we do. But they simply don’t see the resemblances between musical sadness and human sadness that underlie judgments of expressiveness. Of course, we will judge them incompetent in the genre. But will we have any basis for this judgment other than their failure to hear expression in the music. This is, of course, a standard problem for this sort of ideal observer theory of a given fact. You have to make sure that your conception of who counts as an ideal observer isn’t controlled by a prior conception of the fact in question, for the aim is to reduce the fact to the responses of the ideal observer.

11 What, then, should we say about Levinson’s argument? He makes it sound as though there is a perfectly general problem here: everything resembles everything to some extent; so there is no saying in what respects music needs to resemble the expression of emotion if it is to express that emotion. However, there is reason for doubting this general argument. Levinson agrees that listeners will often use judgments of resemblance to make judgments of expression. If we ask one of Levinson’s ideal observers whether some passage P is expressive of E, he will give an answer and that answer won’t be a stab in the dark. He will have reasons and those reasons will consist in judgments of resemblance. We could raise a similar puzzle for this use of judgments of resemblance: given that everything resembles everything else to some degree or other, how could a judgment of resemblance help a listener to make a justified judgment of expression? To what extent should the musical passage sound the way a person would

128

paul boghossian

sound who was expressing sadness before we could be justified in saying that the passage expresses sadness? Well, it’s obviously hard to say, but there had better be a solution to that problem if judgments of resemblance are to do what virtually everyone agrees they must be able to do: namely, underlie judgments of expression. But if we can solve that problem in its epistemological guise, then I don’t see a problem of principle in solving it in its analytic guise. If we can say to what extent a passage P must sound like the natural expression of E for it to be a good guide as to whether P is expressive of E, then we ought to be able to say to what extent P must sound like E for it to be the case that P is expressive of E. If there are problems with this analytic claim, they won’t stem from an inability to specify the degree of resemblance that’s required.

Conclusion In this short paper, I have tried, very schematically, to indicate how I would be inclined to navigate my way through this very difficult terrain. The requisite details must await another occasion. References

Boghossian, P. (2001) ‘‘On Hearing the Music in the Sound,’’ Journal of Aesthetics and Art Criticism, November 2001. Budd, M. (1995) Values of Art, London: Penguin Press. Davies, S. (1994) Musical Meaning and Expression, Ithaca, NY: Cornell University Press. Kivy, P. (2001) New Essays on Musical Understanding, Oxford: Oxford University Press. Levinson, J. (2005) ‘‘Musical Expressiveness as Hearability-as-Expression,’’ in Matthew Kieran (ed.), Contemporary Debates in Aesthetics and the Philosophy of Art, Oxford: Blackwell. Scruton, R. (1999) The Aesthetics of Music, Oxford: Oxford University Press. Yablo, S. (2000) ‘‘Apriority and Existence,’’ in P. Boghossian and C. Peacocke (eds.), New Essays on the A Priori, Oxford: Oxford University Press.

explaining musical experience 129 Further Reading

Boghossian, P. (2001) ‘‘On Hearing the Music in the Sound,’’ Journal of Aesthetics and Art Criticism, November 2001. Levinson, J. (2005) ‘‘Musical Expressiveness as Hearability-as-Expression,’’ in Matthew Kieran (ed.), Contemporary Debates in Aesthetics and the Philosophy of Art, Oxford: Blackwell. Scruton, R. (1999) The Aesthetics of Music, Oxford: Oxford University Press, especially Chapters 3 and 6.

6 Persona Sometimes Grata: On the Appreciation of Expressive Music A A RO N RI D L E Y

Several writers have discussed the possibility that, in hearing pieces of music as expressive, listeners imaginatively construct for themselves a musical ‘persona’ whose affective states the music is then heard as expressing. This thought has polarized opinion: Jerrold Levinson, for instance, claims that the construction of such a persona is always necessary if music is to be heard as expressive (1996); Stephen Davies, by contrast, claims that it is never necessary (2003). Apparently floating around in between, I have suggested that the construction of a persona is sometimes necessary, sometimes not (Ridley 1995). I want to defend that view here. But I want to make clear, in doing so, why that view is not the compromise position that it might seem to be. Levinson and Davies appear to disagree, first and foremost, about whether the construction of a persona is necessary if music is to be heard as expressive at all. My claim, by contrast, is a critical claim: whether the construction of a persona is necessary or not depends upon the particular character of individual pieces of music. And this isn’t a compromise because what interests me is the appreciation of expressive music, rather than the bare possibility of hearing music as expressive—the issue that primarily exercises Levinson and Davies. Indeed, or so I will argue, my issue is prior to theirs, and a correct treatment of it indicates that, while Davies is in one sense right about the question of bare possibility, Levinson’s intuitions are, in another sense, closer to the mark.

persona sometimes grata: expressive music 131

1 Let’s begin, though, with a case that has nothing directly to do with music at all, although it is very familiar from the literature on musical expression. Consider a weeping willow. It droops; it slowly sways; it does all of the usual willowy things. And it very naturally attracts epithets such as ‘sad’, ‘melancholy’ or ‘downcast’, even though no one thinks that these terms capture anything about the willow’s frame of mind, since no one thinks that there is anything of that sort to be captured. Nor do these terms necessarily reflect the states of mind induced by the willow in its beholders: I may perfectly reasonably characterize the tree as ‘melancholy’ whatever the sight of it makes me feel, including nothing. And yet it doesn’t seem to be hard to see why the relevant expressive ascriptions might be thought to apply. The willow is ‘sad’ or ‘downcast’ just because it looks as if it is. On the face of it, this observation might be meant in either of two senses. It might be taken to mean, first, that we see the tree as if it were a sad person: we imaginatively construct a human figure out of it, and diagnose its posture as one of sadness. Or it might be taken to mean, simply, that we notice—and label ‘sad’—features of the tree (droopiness etc.) that, were they the features of a human being, would betoken sadness. In this second case, it is tempting to say, we do not see the tree as if it were a sad human being; rather, we recognize features of it that it shares with sad human beings, but see those features as features of the tree. In the first case, we have constructed a persona; in the second case we have not. In either case, however, it appears that we have seen the tree as ‘sad’, as expressive of sadness. This might seem to suggest that, at least as far as willows are concerned, the construction of a persona is not necessary for the experience of expressiveness. We can see the droopiness of the willow as ‘sad’, it appears, without imputing either the droopiness or the sadness to the human figure whom we might see in the tree. The tree itself, as one might want to say, is droopy and hence, without further ado, appropriately to be seen as sad. But there are reasons to wonder whether this might not be a bit quick. For one thing, the adjective ‘droopy’ seems already to be laden with emotive connotations of just the sort—of ‘sadness’, of ‘melancholy’—that it is being enlisted to justify when those terms are applied to the willow.

132

aaron ridley

What if, instead, one saw and described the willow simply as possessed of thin and pliable parts which, given gravity, tend to hang downwards? There are some grasses like that—not to mention party-streamers—and one would strain to describe these as either droopy or sad. And this would seem to make the ‘droopy, hence sad’ move in the case of the willow look rather under-motivated. Nor does the most obvious sort of attempt to remotivate it look terribly promising. The willow, it might be said, unlike grasses and streamers, has a fundamental uprightness of build that makes the downward tendency of its outer parts especially and strikingly droopy, and so that much more likely to be seen as sad. But this, clearly enough, is to come perilously close to conceding that we see the tree in that way only because we relate its basic uprightness to the uprightness of a human figure, and see the downwardness of its branches as ‘droopy-hence-sad’ in the context of that gestalt, specifically. In which case, it is going to be tricky to steer off the thought that we attribute sadness to the tree only because we see it as if it were a sad person—only, in other words, because we construct a persona. Yet this is just what was to be denied. So does the persona story win out here, after all? I doubt it. For it is clearly open to the opponent of that story to argue that the cues that invite the construction of a persona in the first place must—if we are not out-and-out animists—be a feature or a pattern of features of the tree that is already expressively suggestive, or else nothing in the situation should have prompted us to see the tree as if it were any sort of person at all, let alone a sad one. And that, surely, suggests that expressive attributions are logically prior to the construction of the persona that was, ex hypothesi, supposed to explain them. So it would seem that, despite the considerations offered in the previous paragraph, no persona is necessary if the willow is to be seen as sad. But then again ... One could go on in this vein, I think, pretty well indefinitely, alternating between plausible enough defences of either position without, in the end, arriving at anything conclusive. Do we see willows as expressive only because we anthropomorphize them? Or do we (can we) anthropomorphize them only because we see them as expressive? I find it hard to imagine what might settle the question either way. And there are reasons for this—reasons that matter if we are to get clear about the apparently analogous debate in the philosophy of music.

persona sometimes grata: expressive music 133

2 We can summarize the disagreement between the positions discussed in the previous section in the following rather unattractive way. The pro-persona position holds that: (i)

a feature, F, of a bearer, B, is aptly to be seen as expressive iff B is seen as if it were a person.

The anti-persona position holds that: (ii)

a feature, F, of a bearer, B, is aptly to be seen as expressive iff F would be expressive were F a feature of a person.¹

One striking thing about these positions, so summarized, is the fact that neither concerns itself with the character of B, beyond its possession of F: (ii) has nothing whatever to say about the matter, while (i) offers only the thought that B must be seeable as if it were a person, which, given that anything can in principle be seen as if it were anything else, is not informative.² It might be suggested, I suppose, that their relative muteness about B is actually a strength of these positions, since it promises a certain generality that might otherwise be lacking. But any mileage that this move might seem to have evaporates, surely, once one notices that neither position can deal convincingly with the case in which B is, in fact, a human being. And this means that neither can pretend to offer an account of expressiveness in general (should such a thing be possible at all). This is not, however, the most striking thing about the indifference of these positions to the character of B. The most striking thing, rather, connects to the reason why the disagreement between them appears, in the case of weeping willows at least, to be irresolvable. For it matters, surely, that a willow is a tree—that it is a natural object whose character is ultimately independent, as one might put it, of any fact or facts about ¹ It is worth mentioning that the openings of each of these formulations might, with some plausibility, have been reversed: i.e., both might have started ‘The bearer, B, of a feature, F ... ’, so making B rather than F what is aptly to be seen as expressive. I have chosen the formulations given in the main text simply because they seem to me to work more convincingly for willow trees; but—as we will see in Section 6, below—it may be difficult to justify any general preference for one formulation over the other. ² Cf. Goodman 1976: 5.

134

aaron ridley

human beings.³ Or, to put the point another way, it matters that there is nothing in the character of the tree, qua natural object, that warrants our taking it as being expressive for these reasons rather than those. The upshot of this—if it’s right—is that the disagreement sketched in the previous section was always bound to be irresolvable, at least as it was introduced there. The disagreement was set up so as to have the appearance of a conceptual dispute about the precise application-conditions of certain expressive predicates, an impression strengthened by the place allotted in the exposition to normative-sounding notions such as ‘aptness’ and ‘appropriateness’. But really the dispute was only ever, and at most, an empirical one, and so not usefully to be addressed in the way that I went through the motions of trying to. The real question comes to this: Do you, as a matter of fact (about you), find that you have to imagine that the tree is a human figure in order to see it as sad? Or do you, as a matter of fact (about you), find that you can see the sadness without having to do that? And in this context the notion of ‘aptness’ is to be understood not as normative, and hence as connected to the conceptual, but as referring to the normal —that is, to whatever, given human beings or some subset of them, should turn out to be statistically commonplace. So the dispute, such as it is, never did have anything to do with the specific grounds on which one would be warranted in seeing the tree as sad, however it might have looked at first. In saying this, I am not of course denying that a weeping willow is aptly to be seen as sad. It looks as if it is sad, after all, and that’s enough. Nothing that I have said affects that. I am denying only that the finergrained distinctions at issue between (i) and (ii) can get any conceptual purchase in this context. The dispute between the two could be turned into a conceptual one, of course. It could be turned into one if the pro- and anti-persona positions each had something—perhaps different things—to say about the character of trees that made it plausible to think that our expressive ascriptions were constrained in a relevantly fine-grained way. And perhaps this is possible; I don’t know.⁴ But the dispute as I’ve framed it doesn’t even try to do this. And the important point, for present ³ I set aside here any issues to do with humanly orchestrated breeding, planting, cultivation, etc.: the willow that I have in mind is, as it were, all its own work. ⁴ I can think of nothing in the recent literature on the aesthetics of nature that points very promisingly in this direction.

persona sometimes grata: expressive music 135 purposes, is that without trying to do this nothing of any philosophical moment can be at issue. What we have here is only the appearance of a philosophical disagreement, not the substance of one.

3 Music, by contrast, obviously does have something to do with us. We make it, we play it and, if it is understood at all, it is so by us. There is, moreover, a long tradition, in the West at least, of hearing music as expressive and—correspondingly—of composing and performing music with an ear to its expressive effect. In these as in other respects, it is quite difficult to imagine anything much less like a willow tree than a piece of music is. One might guess from this that the persona-debate, transplanted into what is after all its proper context, music, would come into its own as the site of a genuinely pointful philosophical disagreement. And there is, as we shall see, a measure of truth to this. But it is still worth noting that nothing in the basic form of the disagreement changes. In particular, and as we will also see, it is worth noting that neither position, proor anti-, has a lot to say about the individual character of the bearers of the features aptly to be heard as expressive. What both do, in effect, is to help themselves to the general cultural background gestured at in the previous paragraph—which does at least bring with it a certain ballast of normativity—and then move directly to their competing claims about the application-conditions of expressive predicates to music tout court. And this, or so I’ll suggest, simply isn’t enough. Let’s look at Davies’s anti-persona position first. In his book, Musical Meaning and Expression, he advances an account of musical expression that imputes expressiveness to pieces of music in virtue of their sharing ‘emotioncharacteristics in appearance’ with human beings—that is, in virtue of their bearing features which would be expressive were those features the features of a person (Davies 1994). The position, in other words, is precisely a variety of (ii), as set out above. And Davies’s project in the essay that concerns us is to defend that account by discrediting the pro-persona position, which he takes to be its most serious rival (Davies 2003: 154).⁵ ⁵ On the same page, Davies glosses his own account as follows: ‘musical materials can be literally expressive as a result of presenting to audition sounds with emotion characteristics ... Music

136

aaron ridley

He attempts to do this, primarily, by arguing that the reasons that people have had for proposing the pro-persona position are not very good ones. He devotes most attention to two such reasons. One is that music, or so it is sometimes claimed, is capable of expressing the so-called ‘higher’ emotions, that is, emotions whose character goes beyond ‘what can be presented in appearances’ because, unlike for instance ‘sadness and happiness’, they involve cognitively complex content. Examples of higher emotions include: ‘patriotism, shame, pride, embarrassment, envy, and hope’. If it were true, Davies suggests, that music could convey these states, this would count in favour of the pro-persona position: To make believe someone personified in the music is to imagine that person as possessing beliefs, desires, intentions, and attitudes, even if the contents or objects of these are not transparently conveyed. To entertain that such a person is present in the work is also to make believe a context where cognitively complex emotions might be musically expressed. If it is appropriate to hear music as expressing not solely happiness and sadness but also more subtle, cognitively rich emotions, [the pro-persona position] provides for this possibility (Davies 2003: 156).

But Davies doubts that it is true that music can convey such cognitively complex states, and he has a reasonably easy time of it showing that one attempt to establish a claim to the contrary is not persuasive (2003: 161–6).⁶ I am more impressed than Davies is by the stronger defences of the capacity of music to express the higher emotions.⁷ But for the sake of argument let’s agree with him that these states lie beyond music’s compass, and that the capacity to express them cannot be enlisted on behalf of the pro-persona position. This still leaves open the possibility that to hear music as expressive of (some of) the other emotions might (sometimes) require the construction of a persona. And certainly it is hard to see how these emotions’ alleged lack of cognitively complex content might be supposed to rule this possibility out. The other main reason in favour of the pro-persona position that Davies considers is that a listening conducted in its terms may ‘explain is sad-sounding in much the way that basset-hounds ... are sad-looking. Though it takes imagination to hear music’s expressiveness, it does not take more than is needed to see ... willow trees as downcast’ (2003: 154). ⁶ His target in these pages, incidentally, is Robinson 1994, plus, on occasion, Robinson and Karl 1995. ⁷ In particular, I am impressed by Levinson 1990.

persona sometimes grata: expressive music 137 the structure and coherence of [a] work where a purely technical account will be inadequate’ (2003: 166). His grounds for rejecting this claim, however, seem to be merely stipulative. He insists, without argument, that any persona-construction ‘comes after the recognition of musical unity and the like and, therefore, does not account for that experience’ (2003: 167). But this, clearly enough, is to presuppose, rather than to give a reason for accepting, the absolute priority of the technical. Indeed, he suggests that ‘If the work strikes us as episodic and disjointed, it is not plain that we should prefer a coherent’ experience of it that depends upon the imaginative construction of a persona; and it is clearly his view that we should not prefer such an experience (2003: 168). Yet it is hard to see how this (negative) preference is to be justified: for, on the persona account envisaged, which does not presuppose the absolute priority of the technical, the work in question will precisely not strike us ‘as episodic and disjointed’; and so the issue, in this form at any rate, would not arise.⁸ But—again—let’s concede the point for the sake of argument, and accept that pieces of music are unified or not irrespective of their expressive properties, and so that the construction of a persona can never trump the findings of technical analysis. How much would this show? Only, surely, that the possibility of structural integrity cannot be taken as a sign that the pro-persona position is right. And this hardly shows that it’s wrong. Davies presents himself as discussing these two pro-persona reasons under the umbrella of a larger pro-persona reason, namely, that ‘one cannot understand ... music (fully) without invoking the presence of [a persona] within it’ (2003: 160)—a reason that, as I hope that the foregoing will have made clear, cannot be taken to have been defeated by the reflections offered on the alleged impossibility of music’s expressing the higher emotions or of its having unity imparted to it by its expressive character.⁹ But Davies’s principal conclusion, that apparent emotion-characteristics are all that is needed, at a general level, to explain our experience of music as expressive, still has the room, as it were, to be completely right as far as it goes. If his account is persuasive, after all—as it might well be—it establishes that ⁸ See also Davies 2003: 155–6, for further arguments of the same kind. I have tried to say something about the sources of this sort of question-begging elsewhere: see Ridley 2004, passim. ⁹ I try to say something more positive than this in Section 4, below.

138

aaron ridley

it is a sufficient condition of our expressive ascriptions’ being at least apt that they are grounded in the possession by pieces of music of features that would be expressive if they were the features of a person. And this thought, clearly enough, would survive the truth of a claim to the effect that this piece of music, or some particular detail of a piece of music, does in fact require the construction of a persona if its expressiveness is to be grasped fully, that is, in a way that goes beyond being merely apt.¹⁰ It is also, of course, a thought that comes more or less for free once one reflects for a moment on the general cultural background, as I put it a moment ago, against which music is heard. We are highly attuned to hearing and talking about music as expressive; we are alert for and proficient at spotting the general expressive character of pieces of music; and this, in effect, is all that Davies’s account trades upon. And his account would stay in the business proper to it even if it could be shown, which I doubt, that detecting emotion characteristics in appearance was only made possible by some prior but now largely discarded habit (whether at an individual or a cultural level) of hearing expressiveness in music as the expression of a constructed persona. What I draw from this brief discussion of Davies’s position is threefold. First, the grounds that he gives for denying that it is (ever) true that ‘one cannot understand ... music (fully) without invoking the presence of [a persona] within it’ are not sufficient to establish that conclusion, even when one lets through the (surely over-generous) concessions made above. Second, Davies’s account should be regarded as evincing his commitment to, rather than as offering a defence of, the view that expressiveness is relatively insignificant to musical understanding (hence his rather quick way with the claims that music might express the higher emotions, or that expressiveness might impart musical unity). And third, and following on from the previous point, there is little sign that Davies’s position is driven by close critical inspection of individual pieces of expressive music: instead, it seems to be motivated by quite general considerations about expressive ascriptions, of a sort that sometimes take us rather close to weeping willow territory. ¹⁰ That is, accurate as far as it goes. So, for example, someone might quite aptly, in this sense, describe the ending of King Lear, or of Mahler’s Ninth Symphony, as ‘sad’. (I leave it open whether these examples might not work as a virtual reductio of this sense of ‘aptness’.)

persona sometimes grata: expressive music 139

4 I turn now—still more briefly—to Jerrold Levinson’s pro-persona position. Levinson, like Davies, is concerned to tell a story about all musical expressiveness; and, as the following should make clear, he is an advocate of a version of (i) (see Section 2, above): A passage of music P is expressive of an emotion or other psychic condition E iff P, in context, is readily and aptly heard by an appropriately backgrounded listener as the expression of E, in a sui generis, ‘musical’, manner, by an indefinite agent, the music’s persona. (Levinson 1996: 107)

There are of course some additional details here which are important if one is to arrive at a rounded grasp of Levinson’s view. But for our purposes we can set these aside, and note, simply, that he regards the presence of a persona in the listener’s experience as a necessary condition of music’s being heard as expressive at all. But what, one might reasonably ask, if—like Davies—I attend to my musical experience (either in general, or in some particular case) and find that it includes the apprehension of expressiveness, but not the imagination of a persona? Can I (or Davies) just be wrong about something like that? Levinson’s answer, in effect, appears to be yes, we can. He says: the imaginative hearing-as that the view suggests typically goes on when such expressiveness is registered need not be highly foregrounded, nor ... need it be very determinate in regard to the nature or identity of the expressing agent. (Levinson 1996: 118)

So: we may believe that we are not imagining a persona, but the explanation for that is simply that we have not noticed that we are doing so (although we are). I complained above about the places in Davies’s account where argument seemed to have given way to stipulation, and it is only fair to raise the same complaint here. Why, exactly, should I accept what Levinson is telling me? Or, to put the question another way, what would he say to me if, in a spirit of good will, I listened again and really paid attention this time, but could still find no persona-imaginings in my experience (although I did hear the

140

aaron ridley

music’s expressiveness)? Surely we’d have reached a stand-off.¹¹ And the sort of stand-off we’d have reached, it seems to me, is non-coincidentally similar to the one that—perhaps inevitably—ended the parallel ‘debate’ about willow trees.

5 Such an outcome is not inevitable in the case of music, however. There are at least two ways of avoiding it. One would be to stick at the level of generality that Davies and Levinson appear to prefer, and to say something like this: that (a) Davies’s account tells us what must be minimally true whenever music is heard as expressive (for which persona-construction is never necessary); while (b) Levinson’s account tells us what must be true when the expressive character of any piece of music is understood, not merely in a minimally apt way, but fully (for which persona-construction is always necessary). Neither party would be likely to welcome this resolution, of course. For Davies, the cost of conceding that persona-construction is always required for a full understanding would surely be too high; while Levinson, I imagine, ought to baulk at having to concede that, for it to be true that someone aptly hears pieces of music as expressive of what they express, persona-construction need never go on.¹² But, actually, it seems to me that none of us should welcome this resolution. It is possible, as I’ve indicated on a couple of occasions, that Davies’s account may be more successful in locating the minimally necessary conditions for expressive hearing to be possible at all; and part (a) of the resolution does at least preserve a version of that. But the problem, in my ¹¹ But what about Levinson’s positive arguments for the necessity of a persona? Am I just neglecting these? Yes and no. I’m not neglecting them insofar as they feature in the essay under discussion, since I can find none there. I am, however, neglecting them as they (implicitly) feature elsewhere in Levinson’s writings. The reason for this, though, is that they seem to be very much like the pro-persona arguments canvassed in the previous section. And these, while they may in fact be persuasive (within limits), I have decided—for the sake of argument—to set aside, as Davies would have us do. The larger strategic point of proceeding in this way should become clear in Section 5. ¹² In fact—and at odds with the main thrust of his essay—Levinson might, at one level, just about countenance this concession. He writes: ‘it may not, indeed, be strictly necessary that the listener ... imagine [a persona], as long as he sees what sort of imagining is prompted by the music’ (Levinson 1996: 118). I’m not sure exactly what this means, but it does seem to gesture in the direction that I have mentioned.

persona sometimes grata: expressive music 141 view, is that both parts of the resolution inherit from the original two positions their least attractive, and most gratuitous, feature—namely, the determination to move directly from the background cultural conditions sketched at the beginning of Section 3 to an utterly general account of what makes for apt, or as it might be ‘full’, understanding of expressive music under any circumstances whatever. And while the (usually implicit) inclusion of the relevant cultural background is the reason why this debate doesn’t merely collapse into a version of the one about willows, the direct move made by both parties from background to generality nonetheless ensures that the threat of some such collapse is never far away. The other way of avoiding irresolvable impasse, at least of the sort that has so far arisen, is therefore, it seems to me, to repudiate any premature ambition to generality, and even to suspend judgement (maybe for a very long time indeed) about the very possibility of generality in this context. In Section 2—perhaps cryptically—I attributed the futility of the willow ‘debate’ to the failure of either position to have much to say about the character of the bearers of the features at issue; and, at the beginning of Section 3, I observed that the parallel disagreement about music carried over that indifference. It should now be clear what I was getting at. Levinson’s and Davies’s respective versions of (i) and (ii) depend only minimally, if at all, upon any sort of critical engagement with individual pieces of music: the premature demand for generality precludes that. And what I am suggesting now, of course, is that any prospect of philosophical progress in this area depends precisely upon, and must be rooted thoroughly in, an engagement of just that kind—namely, an engagement that pays proper critical attention to the character of particular pieces of music. In my previous efforts to say something about musical personas this is how I tried to proceed (Ridley 1995: Chapter 8; 2004: Chapter 5). And this is why what I came up with—the thought that persona-construction is sometimes necessary, sometimes not—does not represent any kind of compromise between the two positions that we have been discussing.¹³ My concern, unlike theirs, was to tease out some of the philosophical consequences of the critical appreciation of music. And that, for the reasons ¹³ If anything would represent such a compromise, it would be the (unappetizing) ‘resolution’ sketched at the beginning of the present section.

142

aaron ridley

I have given, is a project that must come before any grander, and perhaps more metaphysically glamorous, undertaking is attempted.¹⁴

6 Let me conclude with two musical examples. The first, which is relatively Davies-friendly, is the G Major Fugue from Book II of J. S. Bach’s The Well-Tempered Clavier. There is no question that this piece has expressive features; nor any question, I think, that these are of a predominantly up-beat kind. Does this work call for the construction of an expressive persona? I think not. Of course, there is nothing to prevent one from constructing a persona if one wants to (I’ve just tried—it’s easy). But, to my ear, nothing is gained by doing so: the music doesn’t emerge the richer for it, nor is there anything about its expressive character, specifically, that appears to reward this way of listening. In this case, it seems to me, the piece is aptly to be heard as possessing up-beat expressive features, and that’s it. The reason for this, I would suggest, is simply that the fugue’s expressiveness is not the most interesting, or even a very interesting, thing about it; so that the kind of imaginative effort that goes into the construction of a persona actually distracts attention from aspects of the work that deserve it more. And this should alert us to one of the many pitfalls of generalizing too soon. For if not every work that is expressive is interesting primarily in virtue of that fact, there can be no grounds for assuming that all expressive works are to be listened to in the same way, with the same pattern and balance of attention.¹⁵ The second example is of a work—Samuel Barber’s Adagio for Strings— which does, I think, call for the construction of a persona. I have chosen this piece for several reasons, but two of the more prominent ones are these: first, it is at least plausible to say that the Adagio is expressive of sadness, say, or depression, and not of any of the higher emotions; and, ¹⁴ I should note that my approach is obviously not immune to impasses of its own. Some critical disagreements may be simply irresolvable. But these, at least, are impasses that come with the territory—with art, with music, with appreciation: they are not the creatures of a particular philosophical project. ¹⁵ I noted at the end of Section 3 that Davies seems to regard expressiveness as relatively incidental to musical understanding. And this, of course, is why the Bach example is ‘Davies-friendly’. Davies’s mistake is to assume that examples of this sort have the field to themselves.

persona sometimes grata: expressive music 143 second, the work can undoubtedly be considered unified and coherent on ‘technical’ grounds alone. It is, in other words, exactly the sort of work that should, if Davies is right, least reward being listened to with a persona in mind. On Davies’s account, of course, the Adagio would aptly be heard as exhibiting, merely, a succession of sadness- or depression-characteristics in appearance;¹⁶ and his faith in this view is bolstered by the following reflection: A work that might be heard as laying out the developing gloom of a depressed persona could be experienced no less convincingly as indicating the unconnected moods of a series of personas, each of whom is (independently) more depressed than the last. (Davies 2003: 167)

This seems to leave the idea of a persona (as opposed to a succession of emotion-characteristics in appearance) doing no work. Now I’ve tried listening to the Adagio—surely an appropriate test-case—in the way that Davies suggests, and find that any sort of experience of it, let alone a convincing one, as having to do with the ‘unconnected moods of a series of personas’ is just not to be had (or at least not in a way that leaves the faintest trace of a musical experience intact). And the sense of outlandishness provoked by this suggestion is only increased when Davies floats the idea that someone might defend a preference for one persona over several, or vice versa, as ‘a form of inference to the best explanation. Where evidence underdetermines theory’, he says, ‘we may still prefer some theories over others’ on the basis of, e.g., ‘predictive power, economy of elements, elegance of structure, and the like’ (2003: 167).¹⁷ This brings out, I think, the gulf that separates Davies’s project not only from mine, but from any project that is at all rooted in the actual experience of listening to music: no one concerned with the latter, surely, would have it cross their mind that what might be at issue was an ‘inference to the best explanation’.¹⁸ Rather, the issue would present itself as a critical one.¹⁹ Is Barber’s Adagio, like Bach’s fugue, most rewardingly or revealingly ¹⁶ Davies doesn’t discuss this piece, but what he would have to say about it is clear. ¹⁷ Davies goes on to reject this idea, although not on the grounds that most people would. ¹⁸ Nor, I suspect, would many think that what they were after was a theory. There is a distinct whiff of scientism here. ¹⁹ As Levinson, to his credit, also recognizes. He considers a worry that is outwardly similar to Davies’s—one persona or several?—and suggests that ‘the answer will vary with the genre and

144

aaron ridley

to be experienced as (merely) possessing expressive features? Or is it best experienced as the expression of a persona—or, indeed, as the expression of a series of personas? I have already indicated that the Adagio strikes me as impervious to a multiple-persona experience.²⁰ And, despite the absence from it of the sorts of cues—having to do with the higher emotions and structural unity—that Davies regards as pointing to the principal, if in his view spurious, reasons for the attraction of the pro-persona position, I have suggested that the Adagio is, indeed, to be heard as the expression of a (single) persona. So why do I say this? What does this way of hearing the work have to offer that is missing from the experience of hearing it as a string of ‘emotioncharacteristics in appearance’? The answer, I think, is depth. Experienced in Davies’s preferred way (which, unlike the multiple-persona experience, seems to me to be perfectly possible), the work exhibits a sustained character of sadness and a structural unity that can be accounted for in purely technical terms—that is, without reference to the work’s expressive features. Indeed, heard thus, those features are essentially a side-effect of the aspects of the work to which a technical account would appeal; and this is to make those features incidental, and, in one important sense, superficial. Heard as the expression of a persona, by contrast, the work’s expressive dimension moves centre-stage, and affords an absorbing portrait of a very particular frame of mind, the character and unity of which are not heard as any sort of side-effect of something else, but as presented directly. And this provides the possibility of a richness and depth of musical experience that Davies’s alternative, in the case of this particular work, does not. The Adagio is simply better, more interesting, when experienced as the expression of a persona. And since none of the relevant cultural background or music history counts against such a listening, we should—on ordinary criticalappreciation grounds—prefer it. (It is an idea somewhat like this, I take it, that underlies, and is then over-stated in, Levinson’s pro-persona position.) One final thought, and one—by now predictable—caveat. The thought is this: if what I have said about the Bach fugue and the Barber Adagio is style of the work involved, the particulars of its formal construction, and the way in which the local expressiveness of its parts relate to any global expressive aim that might appear to be in view’ (Levinson 1996: 107, fn.55). ²⁰ This is not of course to say that all works are similarly resistant. For example, it seems to me that Elgar’s Enigma Variations might very plausibly be heard in this way (and not just because Elgar said so).

persona sometimes grata: expressive music 145 persuasive, this has consequences for what, precisely, we hear as expressive. For, while in the former case it seems plausible to say that the upbeatness resides in certain features of Bach’s music, in the latter case it is altogether more tempting to say that it is the music itself that is heard as sad, and not just this or that about it. And this, if it’s right, means that any attempt to give a general formulation (however persona-tolerant) of the conditions under which music is to be heard as expressive runs into a problem. For either of the obvious ways of framing such a formulation—namely, ‘a feature, F, of a bearer, B, is aptly to be heard as expressive iff ... ’ and ‘the bearer, B, of a feature, F, is aptly to be heard as expressive iff ... ’—deforms the landscape in advance, as it were, and places arbitrary limits upon the ways in which pieces of music are to be heard. This thought confirms me in the view that generality is not a valuable goal in this area of philosophy. And now the caveat, which is related. I said a moment ago that it doesn’t follow from a work’s being expressive that it is interesting primarily in virtue of that fact—and so that it doesn’t follow that all expressive music is to be listened to in the same way. A corollary of this is that it doesn’t follow from the fact that a work is interesting primarily in virtue of its expressiveness (as Barber’s Adagio, I think, is) that it is therefore to be heard in a given way—for example, as the expression of a persona. It may be; it may not be.²¹ One just has to listen and see. So, again, there is no place here, at the level that matters, for generalizing in advance of the specific experiences that individual pieces of music have to offer.²² References

Davies, S. (1994) Musical Meaning and Expression, Ithaca, NY: Cornell University Press. Davies, S. (2003) ‘Contra the Hypothetical Persona in Music’, in his Themes in the Philosophy of Music, Oxford: Oxford University Press, 152–68. Goodman, N. (1976) Languages of Art, Indianapolis: Hackett. Levinson, J. (1990) ‘Hope in ‘‘The Hebrides’’ ’, in his Music, Art, and Metaphysics, Ithaca, NY: Cornell University Press: 336–75. ²¹ Delius’s Summer Night on the River, for example, seems to me to be a piece that is primarily interesting in virtue of its expressive character, but which is not best heard as the expression of a persona. ²² My thanks to Maria Alvarez, Chris Janaway, Alex Neill and Kathleen Stock for comments on earlier versions of this essay.

146

aaron ridley

Levinson, J. (1996) ‘Musical Expressiveness’, in his The Pleasures of Aesthetics, Ithaca, NY: Cornell University Press: 90–125. Ridley, A. (1995) Music, Value and the Passions, Ithaca, NY: Cornell University Press. Ridley, A. (2004) The Philosophy of Music: Theme and Variations, Edinburgh: Edinburgh University Press. Robinson, J. (1994) ‘The Expression and Arousal of Emotion in Music’, Journal of Aesthetics and Art Criticism 52: 13–22. Robinson, J., and Karl, G. (1995) ‘Shostakovitch’s Tenth Symphony and the Musical Expression of Cognitively Complex Emotions’, Journal of Aesthetics and Art Criticism 53: 401–15. Further Reading

Davies, S. (1994) Musical Meaning and Expression, Ithaca, NY: Cornell University Press. Davies, S. (2003) ‘Contra the Hypothetical Persona in Music’, in his Themes in the Philosophy of Music, Oxford: Oxford University Press, 152–68. Levinson, J. (1990) ‘Hope in ‘‘The Hebrides’’’, in his Music, Art, and Metaphysics, Ithaca, NY: Cornell University Press, 336–75. Levinson, J. (1996) ‘Musical Expressiveness’, in his The Pleasures of Aesthetics, Ithaca, NY: Cornell University Press, 90–125. Ridley, A. (1995) Music, Value and the Passions, Ithaca, NY: Cornell University Press. Ridley, A. (2004) The Philosophy of Music: Theme and Variations, Edinburgh: Edinburgh University Press.

PA RT I I I

Musical Meaning

This page intentionally left blank

7 Can Music Function as a Metaphor of Emotional Life? J E NE F E R RO B I NS O N

Introduction Over the past thirty years or so there has been a significant change in Anglo-American music theory. The idea of ‘‘pure’’ instrumental music as a formal, quasi-mathematical structure with no reference to anything beyond itself has given way to more contextual approaches, in which instrumental music is treated as having meaning and expressive qualities which link it to wider human concerns. Instead of focusing solely on the intricate technicalities of harmonic analysis, music theorists are turning to the study of musical semiotics, so-called musical ‘‘topoi,’’ and, more generally, the psychological and emotional meanings to be found in even the purest of pure music.¹ However, as formalist critics persist in pointing out, it is not obvious in what sense pure music without words can mean anything at all. Meaning involves some kind of reference of a sign to a signified, and according to the formalists instrumental music has no ‘‘extra-musical’’ reference at all. Some formalists are happy to grant that music can express lots of things—melancholy, joy, serenity, restlessness—but deny that expression is a form of meaning (Kivy 1989, 1990). On this view, to possess an expressive quality is not to refer to anything. Others disagree. Most notably, Nelson Goodman has suggested that expression is indeed a kind of reference, but of a metaphorical sort (Goodman 1968, Ch. 2). Some theorists go further, ¹ There has also been much written on the political meanings of music, but I will not address this issue directly.

150

jenefer robinson

arguing that music can itself function as a metaphor of emotional life (Ferguson 1960). In this essay I argue that Goodman’s analysis of expression as metaphorical exemplification contains the germs of two different conceptions about musical meaning and expression. One of these is the familiar idea that musical expression is just a matter of a musical piece or passage metaphorically having or possessing certain expressive qualities, such as sadness, tenderness, jollity, serenity, or nervousness. The other is the more intriguing notion that musical expression sometimes goes beyond the mere possession of qualities and can be understood in terms of the exemplification by an extended piece of music of some more complex psychological or emotional drama, such as a struggle to victory or a nostalgic attempt to recover a lost past. It is this second set of cases that have tempted some people to say that music can function as a metaphor for complex human emotional and other psychological states. I begin by laying out Goodman’s view. Then I discuss a critique of the view by Stephen Davies, which seems to hint at this distinction, but in fact makes a different distinction that I find problematic. Then I turn to Anthony Newcomb’s more sympathetic interpretation of Goodman’s theory, according to which a piece of music as a whole can sometimes be understood as metaphorically exemplifying some large-scale story or narrative. The trouble with this proposal is that it suggests no way to distinguish between appropriate and inappropriate candidates for what is being exemplified. In the final two sections I develop Newcomb’s view by suggesting that some pieces of music are appropriately heard in terms of a particular overarching structural metaphor, in particular the metaphor of music as discourse, and by proposing criteria of appropriateness for such metaphors. Finally, I suggest that if a piece of music is appropriately understood as metaphorically a discourse of one sort or another, then it may also be appropriate to hear such music as expressing something about the emotional life of the characters participating in the discourse. A piece of music as a whole can, then, in a sense, function as a metaphor of emotional life, to the extent that it can appropriately be experienced in terms of a large-scale structural metaphor—music as oration, music as drama, music as narrative etc.—in which there are characters whose emotional lives are exemplified by the music.

can music function as metaphor of emotional life?

151

1 The basic premise of Goodman’s Languages of Art is that works of art are symbols in symbol systems. A novel is in a language, which is one kind of symbol system; pictures are in pictorial symbol systems; musical notation is a third kind of symbol system. Since art works are symbols in systems, our appreciation of art is largely cognitive. Traditionally, art works have been discussed aesthetically in terms of three important features: their subject matter, their formal features and the feelings, thoughts or attitudes that they express. Much of the force of saying that works of art are symbols in symbol systems is that for Goodman these three aspects of art can be analyzed as different kinds of symbolic relation or reference relation between the work and its subject, its formal features and its expressive features. So the work represents or denotes its subject matter;² it exemplifies its formal features; and it metaphorically exemplifies or expresses its expressive features. Aesthetic appreciation is a matter of being able to discriminate among what the work represents, exemplifies, expresses, and so on. Exemplification, whether literal or metaphorical, is a kind of reference (of a sample to a label): a sample a exemplifies a (label) predicate b iff b denotes a and a refers to b. Expression is metaphorical exemplification: a sample a expresses a (label) predicate b iff b metaphorically denotes a and a refers to b. Goodman summarizes his view of expression as follows: if a expresses b then: (1) a possesses or is denoted by b; (2) this possession or denotation is metaphorical; and (3) a refers to b. (Goodman 1968: 95)

As is well known, Goodman uses the tailor’s swatch as his example of exemplification: the swatch possesses many qualities (yellow, tweedy, ² This is an oversimplification. Not every subject exists and those which do not cannot be denoted, according to Goodman.

152

jenefer robinson

square), but it refers to only a subset of them (yellow and tweedy, but not square). The swatch exemplifies those qualities that it both possesses and refers to. In the symbol system of swatches, a swatch refers to its color and texture but not to its size and shape. Similarly, a work of art possesses many qualities, only some of which it refers to and hence exemplifies. Expressive qualities are qualities possessed and referred to by a work of art, but they differ from exemplified qualities in being metaphorically rather than literally possessed.³ According to Goodman’s theory of metaphor, when a term from one realm, such as the word ‘‘sad’’ which literally applies to the realm of humans and their psychological states, is transferred from its home realm (human emotion) to a new realm (music), there is a transfer not only of the term itself (‘‘sad’’) but also of the whole family of terms (emotion words) to which it belongs. The new realm (music) is then organized and structured by the application of the new set of terms. What counts as sad and as happy in the new realm is partly a function of how these terms functioned in the old realm. Terms which literally apply to human psychological states are not the only terms which can be metaphorically applied to music. Goodman explicitly allows for the expression of non-psychological qualities such as ‘‘weight’’ (said to be expressed by Daumier’s Laundress) (Goodman 1968: 87) and ‘‘fluidity’’ (offered as an example of what a building could express) (Goodman 1968: 91). Goodman’s theory is at once suggestive and schematic. Notoriously, he says very little about what he means by a work’s ‘‘referring to’’ the qualities it exemplifies and expresses. I myself think that he means that certain qualities in a work are more aesthetically significant (although not necessarily more salient) than others and that it is these qualities that the work exemplifies and expresses rather than merely possesses.⁴ On this interpretation, if a piece of music expresses sadness, this means for Goodman that it is metaphorically sad and that its sadness is one of its more aesthetically significant properties ³ Some commentators have thought that Goodman says that expressed properties are metaphorically possessed and metaphorically referred to, but I think it is clear that Goodman thinks the reference involved is real and literal; he never characterizes a work of art as ‘‘metaphorically referring to’’ what it expresses. ⁴ Goodman characteristically talks as if languages and symbol systems can be understood in abstraction from their users, so he says that it is the swatch and the work of art that refer, not the person handling the swatch or experiencing the artwork. Furthermore, it is strictly speaking inaccurate to say, as I have done, that on Goodman’s view a work exemplifies and expresses qualities. For Goodman, what is expressed and exemplified is always only a label, such as a predicate. For more discussion of these issues, see Robinson (2000).

can music function as metaphor of emotional life?

153

(it ‘‘refers to’’ its sadness in this sense). In short, the expressive qualities of music are qualities metaphorically possessed by the music which are aesthetically significant and which are describable by predicates which are at home in some realm of human experience beyond the music. At the same time, however, his suggestion that the music refers to or signifies those qualities that it metaphorically possesses, (or, strictly, the predicates that denote it), also encourages a rather different idea of musical expression that Goodman would probably reject: the idea that expressive music refers to or signifies not just predicates or musical qualities, but other phenomena—events, actions, ‘‘plots’’—in realms of human experience ‘‘beyond the music.’’ As we will see later, this idea is in many ways more fruitful than Goodman’s ‘‘official’’ theory of expression.

2 Goodman’s theory of expression has never been very popular. Nevertheless, it must be admitted that he has pinpointed something very important in critical discourse: the way that critics often resort to metaphorical descriptions in order to characterize a work of art. The trouble is that critics use metaphors in their critical discourse in order to describe many aspects of a work, not just what it expresses. In a discussion of the language used by art historians, Carl Hausman has pointed out that figurative language is often used to ‘‘direct appreciation’’ (Hausman 1991: 125)⁵ of individual works, and gives as an example Svetlana Alpers’ description of the point of view in a particular Dutch painting: When figures enter they are captives of the world seen, entangled Gulliver-like in the lines of sight that situate them.⁶

Similarly, in describing the color in a Chardin painting, Michael Baxandall says that among the striking color devices, ‘‘the most obvious is the red-lacquered table assertive, but almost unstable.’’⁷ Critics standardly use metaphors to pick out important qualities in works of art. However, the qualities they pick out may be representational, as in the Alpers example, or ⁵ Notice that the use of metaphor is one of the methods suggested by Sibley (1959) in his classic paper ‘‘Aesthetic Concepts’’ for directing attention to aesthetic qualities in works of art. ⁶ Cited by Hausman (1991: 124). ⁷ Cited by Hausman (1991: 119).

154

jenefer robinson

formal, as in the Baxandall case. They are not necessarily what are usually called expressive qualities. Consequently, it would seem that there is no necessary connection between being described metaphorically and being a quality that is expressed by a work of art. Stephen Davies has criticized Goodman on different grounds. He thinks that the qualities expressed by works of art are possessed not metaphorically but literally. Thus for Davies descriptions such as ‘‘sad music’’ and ‘‘mournful music’’ are literal rather than metaphorical, and furthermore they are ‘‘ineliminable’’ in the sense that the properties they refer to cannot be indicated in any other way. He distinguishes such ineliminable metaphors, which are ‘‘mandated by the nature of the music itself’’ (Davies 1994: 146), from metaphorical descriptions of features of music that could be eliminated in favor of a literal description of the same feature. Davies suggests that Goodman is committed to the view that there is a distinction between ‘‘the case in which the music’s possession of a feature is metaphoric and that in which a description of the music is metaphoric’’ (Davies 1994: 146). He thinks that the idea that expression is metaphorical exemplification ‘‘suggests that its being metaphoric is a feature of the manner in which music possesses or presents the property in question’’ (Davies 1994: 146), and that this in turn suggests that perhaps music itself can be metaphoric rather than the language used to describe it. Davies then argues that this way of making the distinction is misleading; it should rather be thought of as a distinction between two kinds of metaphorical description of music, ineliminable descriptions (which later turn out to be literal, not metaphorical at all) and eliminable descriptions: ‘‘for musical metaphor the use of the metaphor is ineliminable, which is not so for (merely) metaphoric descriptions of music’’ (Davies 1994: 147). So, for example, the term ‘‘sad,’’ which appears to be metaphorical when applied to music, is ‘‘ineliminable,’’ whereas other metaphorical descriptions of music are ‘‘gratuitous’’ or ‘‘eliminable.’’⁸ Davies gives as an example the entry of the soloists in the first movement of Mozart’s Symphonia Concertante for Violin and Viola, K.364, which can be metaphorically described as ‘‘homing pigeons, first perceived as dots in the distance, which, on drawing near, flutter to their dovecote’’ ⁸ Budd (1989) has argued that ‘‘sad’’ in ‘‘sad music’’ is ineliminable, and for this very reason cannot be metaphoric.

can music function as metaphor of emotional life?

155

(Davies 1994: 146). The difference between the two sorts of case might be ‘‘dramatized,’’ says Davies, as a difference between ‘‘the metaphoric description of a literally possessed property and the literal description of a metaphorically possessed property’’ (Davies 1994: 148). Davies says: ‘‘Any musical feature might be characterized metaphorically,’’ but ‘‘some musical features seem to require metaphoric description in expressive terms, for they cannot be indicated readily except by such words as ‘sad’’’ (Davies 1994: 146). The idea that the term ‘‘sad’’ is ineliminable when applied to music suggests that only this term can adequately capture a particular quality in the music: ‘‘sad’’ is the unique term which can do this. By contrast, the use of the homing pigeons metaphor ‘‘might be regarded as gratuitous, since there are many other, equally informative accounts’’ that could be offered of the passage (Davies 1994: 146). There are a number of problems with this discussion. First of all, there is no evidence that Goodman wants to distinguish between ‘‘a musical property’s being metaphoric and a musical property’s being described metaphorically’’ (Davies 1994: 146). Since Goodman thinks that only labels can be exemplified or expressed, there is no room in Goodman’s ontology for such a distinction. A ‘‘metaphoric musical property’’ for Goodman can’t be anything other than a property named by a label used metaphorically. Second, even if there is a distinction between ineliminable and eliminable metaphorical descriptions of music, I do not think that Davies has given a plausible account of this distinction. It is true that if the music expresses sadness, then on Goodman’s account, given his nominalism, the term ‘‘sad’’ is ‘‘ineliminable,’’ since the music both is sad and refers to the label ‘‘sad.’’ It does not for Goodman refer to any synonyms of ‘‘sad,’’ since Goodman is suspicious of the whole notion of synonymy. Davies, however, is not committed to any such view. If we admit synonymy, then I think we have to say that the term ‘‘sad’’ when applied to music is not strictly ‘‘ineliminable.’’ If the music is sad, then its quality of sadness may be captured by several synonymous or near-synonymous terms such as ‘‘mournful,’’ ‘‘melancholy,’’ and ‘‘miserable,’’ by similes (it is like a woman weeping and sighing) and by further metaphors (it is a cry of woe and despair). Each of these terms picks out—with greater or less precision and appropriateness—the sad quality in the music. So the term ‘‘sad’’ itself is not ‘‘ineliminable,’’ although the range of suitable replacements is confined to synonyms or near synonyms that refer to a similar kind of emotional experience.

156

jenefer robinson

What is arguably ‘‘ineliminable’’ is that each of these descriptions is couched in terms of human emotional experience. Each of them invites the listener to hear the music in terms of human psychological states. It does not matter very much whether the term is literal (‘‘sad’’ maybe) or metaphorical (it is a cry of woe); both these descriptions invite the listener to hear the music in terms of human emotion. We cannot ‘‘eliminate’’ either description by a translation into non-emotional terms: meaning is lost when we say ‘‘the music is played slowly in the minor key on the cellos.’’ Nor can we eliminate ‘‘sad’’ in favor of a different kind of metaphor: ‘‘dark music,’’ for instance, invites us to hear it in terms of a different realm of experience. Davies also wants to say that descriptions such as ‘‘sad music’’ are literal rather than metaphorical. Certainly if ‘‘sad music’’ is indeed a metaphor, it is arguably a pretty moribund one. Even if Davies were right about ‘‘sad music,’’ however, we cannot generalize this conclusion to all descriptions of expressive qualities. Many of the terms we use to describe the expressive qualities of music are by anyone’s standards used metaphorically: in describing music as weighty or fluid, fearful, anxious, or angry, metallic, watery, or ethereal, roaring or twittering, sunny or breezy, stealthy or heavy-footed, we use metaphors to bring our experience of the music into contact with other areas of our lives. All of these predicates are describing the musical surface in terms drawn from some other realm of human experience, and can bring us to see some new and unexpected quality in the music. If I am brought to hear the belligerent quality in the music, I have been brought to hear something new (to me) in the music through the application to music of a term from a ‘‘foreign realm.’’⁹ Davies contrasts the term ‘‘sad’’ which he says is ineliminable, with his homing pigeons story, which, he claims, is ‘‘eliminable’’ or ‘‘gratuitous,’’ since ‘‘there are many other, equally informative accounts [he] might offer of the passage’’ (Davies 1994: 146). But the two examples are more alike than Davies acknowledges. The mere fact that there are other ways of describing the same passage does not entail that the homing pigeons metaphor is ‘‘eliminable.’’ It is only if there were other equally effective ways of picking ⁹ Here I skirt a number of issues about aesthetic qualities that are not germane to my main purpose in this essay. For example, I have not discussed whether moribund metaphors such as ‘‘sad’’ when applied to music have acquired secondary literal meaning rather than metaphorical meaning. Nor am I committed to any particular position on the question of realism vis-`a-vis aesthetic qualities.

can music function as metaphor of emotional life?

157

out the same quality that the metaphor would arguably be ‘‘gratuitous.’’ However, saying that the soloists enter softly together, gradually get louder, and then stop at the same time is not as informative as the metaphorical description. The quality picked out by the metaphor is the way in which the soloists gradually approach from a distance and eventually land gently together. Perhaps the doves and dovecotes are eliminable (although the choice of bird is not gratuitous, since pairs of doves have often been taken as models of love and fidelity, as in La Fontaine’s fable, Les Deux Pigeons), but what is not eliminable—or otherwise describable—is the companionable quality of the soloists’ journey and their arrival together. The description invites us to hear the music in terms of the human experience of travelling and coming home. Indeed, much music can be conceived of in this way, as either literal or psychological journeyings. Any attempt to get at this quality in the music must involve the idea of journeying. If the two soloists do not arrive by motor-boat or pull up in a pick-up truck, this is because the music suggests a certain quality in the arrival which is captured by the image of gentle, companionable flight. I conclude that Davies is wrong to assume that sadness is somehow ‘‘in’’ the music, whereas the gentle arrival of the soloists is just a fanciful way of describing a fact about the music that could be described in many other ways. Both are descriptions of how we experience the music; both can be replaced without loss of meaning only by synonyms or near synonyms. The equation of musical metaphor with ineliminable and hence literal linguistic descriptions of music fails because many clearly metaphorical descriptions (‘‘fiery,’’ ‘‘breezy,’’ and ‘‘belligerent’’ as well as the gentle arrival) are just as ineliminable as moribund metaphorical or literal descriptions such as ‘‘sad.’’¹⁰ There is an interesting moral to be drawn from Davies’s homing pigeon story, however. I suspect that the reason why it seems adventitious to him is that he is talking about a piece of ‘‘pure’’ music by Mozart, rather than, say, a tone-painting by Vivaldi. One can easily imagine a musical representation of the flight of two fond birds by a violin and a viola playing in the way Davies describes. What is really ‘‘gratuitous’’ about the story of the homing ¹⁰ Frank Sibley has distinguished three kinds of metaphorical description, which are often applied to music: (1) ‘‘‘scenarios’, more or less extended pictures, programs, or narratives: e.g. ‘A L¨andler danced by ogres’’’; (2) ‘‘descriptions in terms of substances and processes’’; and (3) descriptions ‘‘in terms of qualities’’ (Sibley 1993: 167). According to this categorization, the homing pigeon story is a ‘‘scenario.’’

158

jenefer robinson

pigeons in the Mozart example is that, while the metaphor does indeed get us to hear qualities present in the music —the quality of cordial companionship on a journey, the sense of a gentle landing—there is no particular reason for hearing the piece as a whole in terms of the journey metaphor. There is no reason for thinking that the piece is about or refers to traveling homing pigeons. The metaphor can get us to hear qualities in the musical surface, but it does not explain the underlying structure of the piece in any way. It is difficult to describe the way music sounds, just as it is difficult to describe the way a painting looks. When critics say such things as ‘‘This sounds like two homing pigeons coming from afar and landing gently together,’’ or ‘‘This passage sounds silken, even though it’s played on the bassoons,’’ they are trying to help listeners identify how the music sounds, particular qualities of melody, harmony, rhythm, dynamics, orchestration or timbre that may be hard to describe literally. Significantly, however, such remarks carry no implications that the music is about or refers to or exemplifies homing pigeons or silk, or, more generally, birds or materials. Indeed formalist critics are as inclined to describe music in such terms as anyone else.¹¹ If, however, a piece of music really is about two homing pigeons on a journey or about materials or textures, that will affect the kinds of description that seem to be appropriate to the music. A piece entitled ‘‘Textures’’ invites description as ‘‘silken’’ or ‘‘rough-hewn.’’ A piece about journeying birds invites description in journey-metaphors. In general, what the structure of a piece of music is conceived of as exemplifying as a whole may well affect the choice of metaphorical descriptions it invites of its parts. If a piece of music is thought of as metaphorically an outpouring of feeling, then it might seem peculiarly appropriate to call it ‘‘mournful’’ or ‘‘tender.’’ If it is thought of overall as metaphorically a struggle between opposing forces, then ‘‘belligerent’’ might be an appropriate term to use. If it is thought of as an actual or psychological journey, then it would seem just right to describe it in terms of gradual approach and gentle landing.¹² ¹¹ Even arch-formalist Eduard Hanslick uses very flowery, metaphorical language in his music criticism, while at the same time insisting that he is merely describing the character of the music itself, not saying that the music represents or expresses any quality in the ‘‘extra-musical’’ world. E.g. in criticizing Bruckner’s Eighth Symphony, Hanslick writes of its ‘‘muddled hangover style’’ and of ‘‘the sheer weight and monotony of this interminable lamentation.’’ (Hanslick 1978: 288–9). ¹² It could be argued, however, that almost any piece of Western tonal music lends itself to interpretation in terms of the journey metaphor. Contrast, say, Balinese gamelan music, which is not

can music function as metaphor of emotional life?

159

Of course many pieces of ‘‘pure’’ music do not announce themselves as about any specific things or events. Nevertheless there may be good reasons why we conceive of a piece of music as a whole as metaphorically exemplifying some specific theme or other. How we decide which overarching metaphors are appropriate to a particular piece is an issue I address at length in Section 4.

3 Anthony Newcomb (1984a) sees some important advantages in Goodman’s theory of expression. He emphasizes that Goodman’s account maintains the ‘‘transparency’’ of the musical symbol: in calling a piece of music sad or fiery, we are focusing on properties of the music, even though the music is being described in terms of properties of experience ‘‘beyond’’ the music. The two-way reference posited by exemplification seems to fit much better the flow of reference back and forth between musical objects and expressive meaning—between extramusical and intramusical patterns—that is at the heart of musical expression. (Newcomb 1984a: 623–4)

So while there is indeed reference by the music to the ‘‘extramusical,’’ this does not mean that our attention is diverted from the music to something else. The music refers to an ‘‘extramusical’’ process, but we can also hear this ‘‘extramusical’’ process in the musical process. Another advantage that Newcomb sees in Goodman’s theory is that it is not only emotional qualities that can be expressed. Musicologists have often found philosophers’ exclusive concern with sadness and other emotional terms to be too narrow, and have pointed to the wealth of terms actually employed by music critics to describe music. As we have just seen, many different kinds of qualities can be exemplified by music. Besides being sad, happy, serene and sentimental, music can exemplify qualities of the weather (stormy, sunny, misty, breezy), natural substances (silken, metallic, fiery, ethereal, watery), animal sounds (twittering, growling, roaring), human and readily describable as traveling towards or seeking a goal. Again, many pieces of Western tonal music lend themselves to being heard as expressive of emotion, hence perhaps the ubiquity of emotion words to describe musical qualities.

160

jenefer robinson

animal movement (stealthy, clumsy, limping, on tiptoe) and many more. There would seem to be no theoretical limit on the kinds of qualities that music can exemplify. Newcomb’s own view of expression is that it results from: the metaphorical resonances or analogies that a viewer-listener-reader finds between properties that an object possesses and properties of experience outside the object itself. Thus expression results from intrinsic properties of an artwork but also from the metaphorical resonances these properties may have for the perceiver. (Newcomb 1984a: 625)

These ‘‘resonances’’ are ‘‘often broadly intersubjective’’ among a ‘‘class of listeners,’’ those familiar with music in this style; ‘‘expressive interpretation ... is concerned with how the [intrinsic] properties are connected to the resonances for a class of listeners’’ (Newcomb 1984a : 625). As Goodman insists, it is the critic’s task to decide which of a work’s properties it exemplifies and expresses and, further, to bring ‘‘these qualities into the relationships and patterns that make the best argument for the work at hand’’ (Newcomb 1984a: 626). Although Newcomb seems to endorse Goodman’s theory of expression, I do not think that the theories are entirely consistent. First of all, Goodman would have nothing to do with the idea of expressive ‘‘resonances’’ suggested to listeners who then interpret a piece of music in terms of these perceived ‘‘resonances.’’ For Goodman, interpreters are restricted to finding the qualities that a work of art actually has. By contrast, Newcomb seems to think that different classes of listeners will find different resonances in a piece and thereby arrive at different interpretations, all of which can be justified and argued about.¹³ Secondly, Newcomb emphasizes the importance of large-scale musical movement, or as he puts it, ‘‘process or overall form’’ (Newcomb 1984a: 626) rather than musical detail, so that a whole symphony or movement of a symphony might express perseverance through adversity or struggle to victory.¹⁴ Possibly Goodman’s theory can cope with these examples, ¹³ This seems to me an advantage to Newcomb’s way of thinking about these things, but I will not pause here to discuss rival theories of critical interpretation. ¹⁴ These are examples of ‘‘plot archetypes’’ in Newcomb’s terminology. See Newcomb 1984b.

can music function as metaphor of emotional life?

161

but this is not obvious. Goodman’s own examples suggest that the labels expressed by artworks are mostly simple terms like ‘‘sad.’’ Thirdly, although Newcomb does not develop this theme, it seems to me that Newcomb’s discussion of how an extended piece or passage of music can be expressive of some large-scale process points the way to a more important distinction between two somewhat different ways in which music can express or metaphorically exemplify. On the one hand, music can exemplify—in Goodman’s sense—aesthetically significant expressive qualities that are emergent qualities of the ‘‘musical surface,’’ and on the other hand, some music also exemplifies what Newcomb calls ‘‘expressive meaning,’’ the overall meaning of a work taken as a whole. If we look at Goodman’s examples, we find him talking only about properties of the ‘‘musical surface.’’ Terms like ‘‘mournful’’ or ‘‘growling,’’ typically describe details of the way the music sounds, or, more generally of the character of the music, but they carry no implication that the music in question is about human emotion or animal noises. Sometimes, however, a piece of ‘‘pure’’ music does have an overall ‘‘expressive meaning.’’ For example, Beethoven’s Fifth can be described as being about ‘‘perseverance through adversity’’ or ‘‘struggle leading to victory.’’ On this view, the music is not just a surface that has expressive qualities; the music as a whole refers to and is about a process of perseverance through adversity or struggle leading to victory. This does not mean that the music points away from itself and towards perseverance and that paying attention to perseverance distracts us from the ‘‘music itself.’’ The music is a ‘‘transparent symbol.’’ As Newcomb says, when music is truly expressive of some meaning, such as ‘‘struggle leading to victory,’’ that meaning is (metaphorically) exemplified by the music. The music has the same structure as the ‘‘extra-musical’’ reality it refers to; it is isomorphic to these ‘‘extra-musical’’ processes. The music does not merely refer to struggle leading to victory, as the words ‘‘struggle to victory’’ do;¹⁵ the music itself metaphorically struggles and we can experience struggle in the music itself. Indeed we experience the drawnout process of a struggle to victory both as a musical and as a psychological process. For example, one particular heroic-sounding theme might be heard as struggling for predominance over another darker theme in a minor ¹⁵ In Goodman’s terminology, ‘‘struggle to victory’’ denotes a struggle that leads to victory.

162

jenefer robinson

key, and then over a period of interaction, the ‘‘hero’’ theme can be heard as triumphing over the darker minor theme. In such cases there is not just one simple quality being exemplified such as ‘‘fiery’’ or ‘‘heroic,’’ although perhaps these might be appropriate terms to describe the hero theme itself. Rather, what is (metaphorically) exemplified is an extended drama which gives expressive meaning to the piece as a whole. In short, the music metaphorically exemplifies [a certain kind of] perseverance through adversity, or struggle leading to victory. Newcomb believes that the same piece of music has a wide ‘‘expressive potential,’’¹⁶ which can be described—verbally—in many different ways by different listeners. Different expressive interpretations have to make sense of the music as a whole; each one has to illuminate transitions and developments; each one has to fit together the many diverse elements of the whole. One-word ‘‘labels,’’ such as ‘‘fiery’’ or ‘‘heroic,’’ although useful for pointing out expressive qualities of the music that might need to be mentioned in the context of an expressive interpretation, will never be adequate all by themselves to explain large-scale expressive meaning in music. The distinction I am drawing may be what Leo Treitler has in mind when he contrasts what most people think of as one-word metaphorical descriptions of music (‘‘sad music’’) with the ‘‘metaphorlike effects’’ (Treitler 1997: 47) possible when one considers a large-scale piece of music as a whole.¹⁷ The second movement of Schubert’s E flat Piano Trio, D.929, opens with a ‘‘long discursive theme in C minor, accompanied by a trudging figure’’ that ‘‘registers as the tattoo of a funeral march’’ (Treitler 1997: 47). Later a second soaring theme enters, which through its ‘‘vigorous, affirmative development,’’ seems ‘‘as though it might be capable of lifting the movement out of the hopelessness of the C-minor music but in the end always slumps back into it’’ (Treitler 1997: 47). Towards the end of the final movement, Allegro moderato, a rondo, in which the themes keep repeating in ‘‘an endless chatter of mindless energy,’’ unexpectedly the solemn tattoo theme from the second movement reappears on the cello. Treitler comments: ‘‘Suddenly it is as though this is what all that patter had been intended to suppress’’ (Treitler 1997: 48). Treitler does not give an analysis of these ‘‘metaphorlike effects’’ in music, but I would ¹⁶ This is a phrase Newcomb borrows from Cone (1974: 166). ¹⁷ I say ‘‘most people think’’ of ‘‘sad music’’ as a metaphorical description, because Treitler, like Davies, thinks ‘‘sad’’ is really literal when applied to music.

can music function as metaphor of emotional life?

163

suggest that what he is talking about is an example of ‘‘expressive meaning’’ in Newcomb’s sense: the Trio as a whole is experienced as having a ‘‘narrative’’ dimension, in which the piece exemplifies a wider pattern of human experience.

4 In his classic work on aesthetics, Aesthetics: Problems in the Philosophy of Criticism, Monroe Beardsley articulated what is still the most central difficulty in the theory that music exemplifies or is an iconic sign of extramusical processes, whether psychological, physical, political, or whatever. Using Ravel’s Bolero as his example, Beardsley says that if it is: an iconic sign of those psychological processes that have all its qualities, then of course it is a sign only of itself. If it is taken to be a sign of processes that are similar to it in being characterized by increasing intensity, then we cannot tell simply from hearing the music that it is these processes it signifies rather than others. The music must be supplemented by a verbal rule that prescribes which of its qualities are the iconically significant ones. And in the absence of such a rule, we cannot decide among the innumerable possible qualities, so that if the music is a sign at all, it is ambiguous. (Beardsley 1958: 336)

Here in a nutshell is the problem with theories such as Newcomb’s: the very same musical passage or movement may apparently exemplify hopping up the stairs and down again, rising to heaven and falling to hell, feeling uplifted and then feeling dejected, a sunny day in the mountains followed by a gloomy evening in the motel. All these possibilities are consistent with the ‘‘expressive potential’’ of the music. How are we to know which of them (if any) is the most appropriate in any given case? Without a ‘‘verbal rule,’’ as Beardsley puts it, how can this question be decided? The short answer is that the question is decided by reference to the larger context in which the music is embedded. If it is a whimsical piece of tone-painting describing the capering of small children, then hopping up and down the stairs may be what the piece is about; if it is a cantata about the death and resurrection of Christ, then perhaps it should be heard as exemplifying rising to heaven and falling to hell; and if it is an expression of

164

jenefer robinson

personal feeling it might be appropriate to hear it as exemplifying feelings of uplift followed by dejection. But then the question arises: if there is no helpful title to guide us, how do we know whether the piece should be heard as a tone-painting, a meditation on the resurrection or a piece of emotional expression? My suggestion is that music as an art has historically been conceived of in terms of certain large-scale or umbrella metaphors, what Lakoff and Johnson have called ‘‘conceptual metaphors,’’ that structure our musical experience (Lakoff and Johnson 1980). In their well-known example, ‘‘Argument is war,’’ Lakoff and Johnson show how we structure our experience of argument in terms of this conceptual metaphor, thinking of the participants in an argument as taking sides, as defending and attacking, as winning or losing, and so forth. Similarly, music—or at least music in the Western tradition—has always been understood in terms of one or another conceptual metaphor. The music Newcomb discusses, for example, comes from the repertory of nineteenth- and twentieth-century Romanticism. In various essays Newcomb has discussed Schumann’s Third symphony, Wagner’s music drama, Siegfried, and Mahler’s Ninth. The ruling metaphors for such works include music as psychological drama or story. A symphony is not literally a play, but in the wake of the invention of opera, symphonies could be thought of as wordless operas or dramas. The big question, however, is when, if ever, is it appropriate to experience ‘‘pure’’ music as metaphorically a drama or a story of some sort? My answer is that we need to find the answer in the history of music and music reception. Different metaphors are current at different periods of music history for thinking about music. Formalism is itself a theoretical position that arose in particular historical circumstances and is more appropriate to some musical pieces than to others. We best understand a piece of music if we understand how composers and their culture conceived of their music, in terms of what conceptual metaphors. To illustrate, I will briefly consider the period in the mid- to lateeighteenth centuries when the baroque period gave way to the classical, the polyphonic style gave way to the newly invented sonata form, and the suite—derived from dance forms, such as the sarabande, minuet and gigue familiar from Bach’s suites—gave way to the symphony and the string quartet. Scholars have defended a number of different views about how this new ‘‘absolute’’ music was conceived of by its contemporaries, each

can music function as metaphor of emotional life?

165

of which relies on a different conceptual or structural metaphor, such as music as mathematical structure and music as discourse. My goals here are twofold. First I want to point out how all these different views assume that the musicians and theorists of this period thought of music in terms of certain recurrent structural metaphors.¹⁸ Second, I want to emphasize that one of the most widely accepted of these structural metaphors at this period was the metaphor of music as discourse. For most Enlightenment thinkers purely instrumental music was despised as trivial, a ‘‘mere twittering,’’ written ‘‘for the entertainment and idle enjoyment of the few’’ (Fubini 1991: 253). Before 1800 music was inextricably connected to words, and only music with words—especially the recently invented form of opera—could have any serious content or could express ideas and emotions: The older idea of music, against which the idea of absolute music had to prevail, was the concept, originating in antiquity and never doubted until the seventeenth century, that music, as Plato put it, consisted of harmonia, rhythmos, and logos. Harmonia meant regular, rationally systematized relationships among tones; rhythmos, the system of musical time, which in ancient times included dance and organized motion; and logos, language as the expression of human reason. Music without language was therefore reduced, its nature constricted: a deficient type or mere shadow of what music actually is. (Dahlhaus 1989: 8)

Given such assumptions, the new ‘‘absolute’’ music without language required justification. John Neubauer has argued that the roots of this aesthetic shift lie in an old tradition going back to Pythagoras, according to which music is (literally) a brand of mathematics. In the medieval university curriculum, the quadrivium consisted of astronomy, geometry, arithmetic, and music. Neubauer argues that the revitalization of Pythagoreanism ‘‘motivated the change from an aesthetics of affects and expression [the baroque] to an aesthetics of structure [the classical]’’ (Neubauer 1986: 194). He cites thinkers such as Rameau, who emphasized ‘‘a combinatorial and permutational principle [of music], which allows the generative construction of chords and their cadential progression within the framework of the major-minor ¹⁸ Cf. Cook (1990: 4): ‘‘a musical culture is, in essence, a repertoire of means for imagining music.’’

166

jenefer robinson

tonal system’’ (Neubauer 1986: 83), as well as a deductive principle according to which music derived from natural (i.e. non-conventional) harmonic laws. Music, as Pythagoras claimed, reflects the mathematical constitution of the universe. According to Neubauer, this strand of thinking continued through the eighteenth century and laid the groundwork for the idea of music as an autonomous art form, distinct from poetry and the poetic goals of expression and imitation. I find this account unconvincing—or at least incomplete—for a number of reasons. First, although it is true that the idea of music as fundamentally connected to mathematics is present throughout the history of Western music, having more or less influence at different periods, yet at the time of the new ‘‘classical’’ style, the ideas of Enlightenment figures such as Rousseau were in the ascendant, and Rousseau explicitly denigrated music without words because he thought that, unlike poetry, it had no content. Such proto-Romantic figures had little interest in music’s mathematical properties. Second, if we think of early masterpieces in the new classical style, such as Haydn’s early string quartets, although they do exemplify symmetry and proportion, their dramatic, human qualities are just as striking as their carefully crafted form. A theory I find more plausible has been developed by Mark Evan Bonds, who argues that the most important metaphor for understanding early ventures in sonata form is the oration: a speech, a piece of rhetoric. Music was thought of as a discourse that seeks to persuade and emotionally affect its audiences. In his book, Wordless Rhetoric: Musical Form and the Metaphor of the Oration, Bonds traces the idea of music as rhetoric throughout the eighteenth and early nineteenth centuries, and shows how this temporal metaphor—of a development of thoughts over time—eventually, in the nineteenth century, gives way to the spatial metaphor of music as an organism: music as a living thing rather than a series of events. In the eighteenth century: The instrumental work was seen as a wordless oration, and its form was viewed not so much as a harmonic or thematic plan but as an ordered succession of thoughts. (Bonds 1991: 53)

Bonds argues that sonata form was understood by contemporaries on the model of an oration with a Hauptsatz, a main idea, which was

can music function as metaphor of emotional life?

167

then elaborated in various ways; the way the central idea was elaborated determined the structure of the movement as a whole. Whether or not sonata form was thought of on the model of an oration, it is clear from numerous eighteenth-century sources that music was indeed regarded (metaphorically) as a rhetorical discourse of some sort. Enrico Fubini notes that the idea of sonata form as a reasoned discourse was a commonplace of contemporary thinking. He suggests, however, that rather than thinking of it as an oration, it is more illuminating to think of the rise of sonata form as a narrative form comparable to the newly burgeoning literary art form of the day, the novel. Haydn, who more than anyone else, was responsible for this new musical language, wrote music that can perhaps best be conceived in this way. ‘‘Haydn had the gift of creating musical discourse and conversation’’ (Fubini 1991: 255), which is heard to best advantage in his string quartets. In contrast to baroque counterpoint and polyphony, ‘‘a substantially repetitive style in which variety was obtained through ornamentation and contrasting timbres and dynamic levels,’’ Haydn’s sonata form ‘‘is narrative and not confrontational; it seems to be offering in modern, Enlightenment terms the ideal of a solid musical language which might lead to the most exciting spiritual adventures, and which is capable of allowing the expression of the most sophisticated emotions within a logical structure which can be understood by any reasoning person’’ (Fubini 1991: 257). Similarly, in discussing the development of the classical symphony, Charles Rosen emphasizes the influence of opera, and comments that the classical symphony should be thought of as (via the metaphor of) a drama: The application of dramatic techniques and structure to ‘‘absolute’’ music was more than an intellectual experiment. It was the natural outcome of an age that saw the development of the symphonic concert as a public event. The symphony was forced to become a dramatic performance, and it accordingly developed not only something like a plot, with a climax and a denouement, but a unity of tone, character, and action it had only partially reached before. Unity of action was, of course, one of the classical requirements of tragedy, and the symphony as drama gradually abandoned every trace of the looseness of the [baroque] suite. (Rosen 1972: 155)

With Beethoven and on into the next century the dramatic possibilities of sonata form were fully exploited, and symphonies and string quartets

168

jenefer robinson

became wordless mini-operas, full of dramatic conflict in which different characters are pitted against each other. Moreover, with the Romantic movement came the notion that art in general involves the outpouring of feeling; the metaphor of pure instrumental music as drama fits nicely with the metaphor of music as psychological drama, and the metaphor of music as psychological exploration. In summary, the various theories about how the new ‘‘absolute’’ music was conceived of each rely on an overarching conceptual or structural metaphor, such as music as mathematical structure or music as discourse, and among these metaphors the most widely accepted seems to have been music as discourse, including music as oration, music as novel or narrative, music as conversation, and music as drama.¹⁹ I now want to suggest that unless we grasp the metaphors that structure a piece of music, we do not hear it as it was meant to be heard and we are likely to miss some of its most important expressive qualities. Thus, if we think of Haydn’s music as merely pretty tunes or pleasingly organized musical structures, we miss some of its most important qualities, which depend upon our hearing it in terms of the metaphor of ‘‘string quartet as civilized conversation.’’ Haydn’s string quartets are often said to be ‘‘good-humored,’’ ‘‘conversational,’’ and ‘‘full of give and take.’’ If we do not hear them as conversations, however, it is unlikely that we will hear those expressive qualities in the music. It is only because we hear the music in terms of the metaphor of conversation (we experience it as exemplifying—referring to and sharing a structure with—a particular kind of conversation) that we are able to detect the good humor and the conversational quality of the music. We have to hear the instruments as ‘‘voices,’’ as interrupting each other, as denying or confirming what each other ‘‘says,’’ as repeating with emphasis or repeating with irony; as tossing ideas back and forth; as commenting amusingly on each other’s ideas, and so on and so forth. This whole way of listening to a string quartet is governed by the conceptual metaphor: ‘‘this music is a conversation.’’ And given the wit with which Haydn develops the idea of string quartet as ¹⁹ Romanticism and post-Kantian idealism also spawned other conceptions of music. Some postKantian idealist philosophers thought of music as having metaphysical meaning. Schelling held that art is the expression of the Absolute or the Infinite, and Schopenhauer claimed that pure instrumental music is the representation of the Will, the force or energy that constitutes metaphysical reality and forms the basis of all of nature.

can music function as metaphor of emotional life?

169

Allegro

5

10

Example 1. Haydn Quartet in E Major Op. 54 No. 3, opening bars (mm. 1–14)

conversation, it is highly unlikely that the reference to conversation in this music is merely fortuitous. So, for example, we find Charles Rosen describing the opening of Haydn’s E Major Quartet Op. 54 No. 3 (Example 1) in terms of a ‘‘sociable comedy’’ (Rosen 1972: 141). The second violin and viola begin but are interrupted by the first violin with a new phrase; the second violin and viola try again and are again interrupted. Then, ‘‘the second violin and viola, resigned, give up their phrase and accept the first violin’s melody; begin it—and are again comically interrupted’’ (Rosen 1972: 142). The point is that there would be no comedy if we did not hear the music as discourse or conversation among four voices. Haydn’s wit consists in part in how he exploits the idea of the string quartet as conversation. Now, it may be that formalists will claim that music can be heard as witty and conversational even if they deny that it refers to or exemplifies a conversation. Certainly, we can hear music in terms of any kind of scenario we choose, just as Davies suggests it is possible to hear a passage in the Mozart Symphonia Concertante in terms of the companionable flight

170

jenefer robinson

of two homing pigeons. The homing pigeon story gets us to hear a certain quality in the music even though it is not appropriate to hear the music as exemplifying or referring to the journey of two homing pigeons. However, it is hard to imagine hearing a conversational quality in a quartet without experiencing the quartet as a conversation. Moreover, it is impossible to hear Haydn’s treatment of the conversational aspect of the quartet as witty unless you hear this conversational aspect. More importantly, the difference between the homing pigeons in Mozart and the conversation in Haydn is that whereas it is inappropriate to hear a story about homing pigeons in the Mozart piece, hearing a Haydn String Quartet as metaphorically a conversation is an appropriate way to hear it, authorized by the way in which it was received by contemporary and later audiences and was probably thought of by its composer. In short, the witty, conversational quality of the Quartet Opus 54 No. 3 is part of the expressive meaning of the whole, not just an adventitious description of something in the musical surface. The piece itself is (partly) structured by the metaphor of conversation and (metaphorically) exemplifies a particular conversation, one in which, among other things, two of the participants are constantly interrupted by a third.

5 What does it mean to say that music is metaphorically a discourse? Without discussing and defending in detail any particular theory of metaphor, it is perhaps fair to say that there are two main types of theory: on the one hand theories which claim that metaphors involve a transfer of meaning and, on the other hand, the view associated with Donald Davidson,²⁰ that metaphor is primarily a pragmatic rather than a semantic affair, that metaphors keep their primary meaning but simply induce us to change our attitudes to what is being described metaphorically. If we are told that music is metaphorically a conversation, we are being invited to think of music as a conversation, just as Romeo’s calling Juliet the sun invites us to think of Juliet as the sun. We are also invited to organize our experience of music in a new way: to experience music as a conversation is to hear it as an ²⁰ See especially Davidson (1978).

can music function as metaphor of emotional life?

171

interchange of voices, as a means of social communication, in which each voice listens to the others, comments on the others, interrupts the others, echoes the others, accepts or denies what the others assert and so on. Second, in being told that music is metaphorically a conversation, we are being encouraged to experience the same feelings and attitudes to the music that we experience towards conversation, just as calling Juliet the sun invites us to experience the same attitudes to Juliet that we already have towards the sun. Romeo’s calling Juliet the sun reflects his new and different experience of Juliet: she is the center of the universe, the brightest thing in all the heavens, the giver of light and life, and so on. Similarly, our experience of music as a conversation reflects a new and different experience of the music; for example, we experience it as civilized and civilizing, as witty, as a synthesis of reason and imagination (Fubini 1991: 256). Our attitude towards the music changes when we think of it as a conversation. When we say that a is metaphorically b, the focus is on a and our changed attitude to a. At the same time, our attitude to b also shifts a little by being brought into juxtaposition with a (Black 1962). So the sun seems to be a less distant entity, more intimately concerned with human affairs, perhaps, when considered as an appropriate description of a young and charming girl. Similarly, when we consider music as some sort of discourse, our main attention is on the music, but we may also notice our attitudes to discourse changing as well, as we think of it as potentially musical. Third, metaphors, as Arthur Danto has pointed out, are intensionally opaque (Danto 1981: 188–9). To say that ‘‘No man is an island’’ is not to say that no man is a body of land completely surrounded by water. Similarly, to say that music is a discourse does not mean that music is a set of sentences in a natural language uttered by a speaker or speakers. It is quite beside the point to object, as the formalist might, that a string quartet isn’t really a conversation, because in the quartet no-one is really uttering sentences or making assertions. We all know this. The music is only metaphorically a conversation. The metaphorical description simply encourages us to have a certain sort of experience of the music. If I am right to accept the view that at the beginning of the classical period, ‘‘pure’’ instrumental music in sonata form was widely thought of as a kind of discourse—as an oration, or a story, a conversation or a drama—and that these ways of thinking about music were embedded

172

jenefer robinson

in corresponding metaphors, the question arises: is it correct to say that particular pieces of music can be thought of as themselves metaphors for ‘‘extra-musical reality’’? If Beethoven’s Fifth is metaphorically a drama, can it perhaps be a metaphor for struggle leading to victory? If Haydn’s Opus 54 quartets are metaphorically conversations, can they also serve as metaphors for friendly companionship? The short answer to this question is ‘‘No.’’ Because a piece of ‘‘pure’’ or absolute music does not have any literal meaning, it cannot function as a metaphor. Why, then, do some people have such a powerful inclination to treat music as metaphor? One reason is probably that we think of music in terms of the metaphors that apply to it, in the ways I have been discussing: on the one hand broad over-arching conceptual metaphors—music is a drama, music is a conversation—and on the other hand metaphors that characterize details of the musical surface, such as music as ‘‘anxious,’’ ‘‘stormy,’’ ‘‘belligerent’’ and so on. However, to describe music metaphorically is not the same thing as to treat a piece of music as itself a metaphor. We talk metaphorically about lots of things without wanting to say that those things are themselves metaphors. Just because a friend is a jewel, we don’t feel inclined to say that the friend herself is a metaphor for jewelry. When people say that music is a metaphor for something, or—more circumspectly, like Newcomb—that music has ‘‘metaphorical resonances,’’ I think they mean simply that the piece, when conceived of metaphorically as a certain kind of process, thereby exemplifies a process or processes in the ‘‘extra-musical’’ world, i.e., it refers to those extra-musical processes and it is structurally isomorphic to them.²¹ For example, if we think of a Haydn string quartet as a conversation, then we can experience the instruments of the quartet as exemplifying voices in conversation, agreeing, disagreeing, commenting wittily, uniting and resolving their disagreements, and so on. The quartet is not a metaphor for conversation; it is metaphorically a conversation, and it exemplifies a particular (kind of) conversation: it refers to this conversation and is isomorphic with it. Similarly, if a Beethoven symphony is metaphorically a drama, then it can be experienced as a particular (kind of) drama in which, for example, ²¹ Even the formalistically inclined Malcolm Budd suggests a space for exemplification in music when he allows that ‘‘the appreciation of music is infused with the perception of relations between parts as such relations, and ... these relations are not specific to music but obtain outside it’’ (Budd 1995: 170).

can music function as metaphor of emotional life?

173

there is a struggle to victory. The symphony exemplifies struggle to victory: it is about struggle to victory and it has the same structure as struggle to victory. Instead of allowing, like Newcomb, that each listener may bring quite different metaphorical resonances to a piece, however, I am insisting that there are better and worse ways of metaphorically conceiving of a particular musical work, and restrictions on what it is appropriate to hear it as exemplifying. It is only if the music is metaphorically a drama or a narrative, that it makes sense to hear it exemplifying a struggle to victory. At the same time, of course, there is room for multiple interpretations of the same piece, within the confines of the metaphor ‘‘music as drama.’’ So far I have suggested that Goodman’s account of expression as metaphorical exemplification can be read both as an account of expressive qualities of parts of the musical surface, and—in Newcomb’s version—as an account of the ‘‘expressive meaning’’ of a piece as a whole, the way in which some musical works metaphorically exemplify some story or sequence of events or process in the ‘‘extra-musical’’ world. Now, at least some works of art also contain expressions of emotion in a more literal way. Thus King Lear, expostulating with the elements on the heath, thereby expresses his rage and grief. The protagonist of Schubert’s Winterreise expresses his forlornness and despair. In concluding this essay I would like to point out that if pure music is conceived of as metaphorically a kind of discourse, it too can contain expressions of emotion of this sort. If a string quartet is conceived of as metaphorically a conversation, then it refers to voices in conversation. Voices are expressive and they can express particular thoughts, emotions, feelings and attitudes. Similarly, if a symphony is metaphorically a drama, then it refers to characters and a ‘‘plot;’’ characters too can express particular thoughts, emotions, feelings, and attitudes. Moreover, composers can express their emotions, attitudes and so on through the interplay of voices or characters that they create. Edward T. Cone and Jerrold Levinson have each suggested that in music we can hear a musical persona or perhaps a set of characters who behave somewhat like the protagonists in a drama.²² If a symphony expresses nostalgia for an Arcadian past, which is threatened by a desire for a glitzy urban life (as in Newcomb’s reading of Mahler’s Ninth, second ²² Cone (1974); Levinson (1996); see also Karl and Robinson (1995). My most recent attempt to explain musical expression is in Robinson (2005).

174

jenefer robinson

movement),²³ it is a persona or character in the music who is expressing this nostalgia and sense of conflict. The character may be, as Cone says, some particular musical element—a theme, a key, or a set of instruments—rather than the music as whole. So when we say that a (whole) piece of music expresses nostalgia for an idyllic past, we are probably saying something much more complicated than that the music (as a whole) is metaphorically nostalgic and refers to that particular quality, as in Goodman’s theory. In Newcomb’s interpretation of the Mahler movement, he hears the music as representing a character who is conflicted between the desire for a past state of Arcadian innocence (represented by a theme from the first movement, which is recalled in the second), and the desire for a more sophisticated and exciting urban life (represented by a seductive waltz theme).²⁴ In his reading, this character expresses nostalgia for the past, and a desire for the glitzy present. Whatever you think of the details of his account, it shows how, if music is metaphorically a drama in which there are characters, these characters can express their emotions and attitudes. To conclude: formalists complain that we should not treat music as anything other than itself, not as language, and not as painting. If we describe music in metaphorical terms, it is only because it is so difficult to capture the character of a piece of ‘‘pure’’ music. I have argued, however, that although metaphorical descriptions of the ‘‘musical surface’’ play an important role in music criticism, there are more fundamental metaphors which do not just describe the surface of music but structure our whole experience of it. To hear music as metaphorically a drama is to hear it as capable of referring to characters and exemplifying a plot; to hear music as metaphorically a conversation is to hear it as capable of referring to voices in conversation and as exemplifying the unfolding of a particular conversation. And once we hear characters and voices in the music, we can also hear the emotions and attitudes that they express.²⁵ ²³ Newcomb (1997). ²⁴ Leonard Ratner has described how music exploits ‘‘topoi’’ which aid in expressive interpretation, such as the l¨andler, an ‘‘innocent’’ peasant dance, versus the waltz, a sophisticated ‘‘urban’’ dance. See Ratner (1980), especially Chapter 2. ²⁵ I am grateful to Stephen Burton, Russell Dancy, Lydia Goehr, Kim Lockwood, Mary Sue Morrow, Edward Nowacki, Kathleen Stock, Leo Treitler, and Nick Zangwill for useful comments and/or discussion of these issues. This essay is a much revised and extended version of an article that ´ appeared in a special issue of the Revue Fran¸caise d’Etudes Americaines 86 (2000): 77–89. I would like to thank the editors of the Revue for permission to adapt this article.

can music function as metaphor of emotional life?

175

References

Beardsley, M. (1958) Aesthetics: Problems in the Philosophy of Criticism, New York: Harcourt, Brace and World. Black, M. (1962) ‘‘Metaphor,’’ in J. Margolis (ed.), Philosophy Looks at the Arts, New York: Scribner’s, 218–35. Bonds, M. E. (1991) Wordless Rhetoric: Musical Form and the Metaphor of the Oration, Cambridge: Cambridge University Press. Budd, M. (1989) ‘‘Music and the Communication of Emotion,’’ Journal of Aesthetics and Art Criticism 47: 129–38. Budd, M. (1995) Values of Art: Pictures, Poetry and Music, London: Penguin. Cone, E. T. (1974) The Composer’s Voice, Berkeley, Calif.: University of California Press. Cook, N. (1990) Music, Imagination and Culture, Oxford: Clarendon Press. Dahlhaus, C. (1989) The Idea of Absolute Music, trans. R. Lustig, Chicago: University of Chicago Press. Danto, A. (1981) The Transfiguration of the Commonplace, Cambridge, Mass.: Harvard University Press. Davidson, D. (1978) ‘‘What Metaphors Mean,’’ Critical Inquiry 5: 31–47. Davies, S. (1994) Musical Meaning and Expression, Ithaca, NY: Cornell University Press. Ferguson, D. N. (1960) Music as Metaphor, Westport, Conn.: Greenwood Press. Fubini, E. (1991) The History of Music Aesthetics, trans. M. Hatwell, London: Macmillan. Goehr, L. (1992) ‘‘Writing Music History,’’ History and Theory: Studies in the Philosophy of History 31: 182–99. Goodman, N. (1968) Languages of Art, Indianapolis: Bobbs-Merrill. Hanslick, E. (1978) Hanslick’s Music Criticism, trans. and ed. H. Pleasants, New York: Dover. Hausman, C. R. (1991) ‘‘Figurative Language in Art History,’’ in S. Kemal and I. Gaskell (eds.), The Language of Art History, Cambridge: Cambridge University Press, 101–28. Karl, G. and Robinson, J. (1995) ‘‘Shostakovich’s Tenth Symphony and the Musical Expression of Cognitively Complex Emotions,’’ Journal of Aesthetics and Art Criticism 53: 401–15.

176

jenefer robinson

Kivy, P. (1989) Sound Sentiment: An Essay on the Musical Emotions, Philadelphia, Penn.: Temple University Press. Kivy, P. (1990) Music Alone: Philosophical Reflections on the Purely Musical Experience, Ithaca, NY: Cornell University Press. Lakoff, G. and Johnson, M. (1980) Metaphors We Live By, Chicago, Ill.: University of Chicago Press. Levinson, J. (1996) ‘‘Musical Expressiveness,’’ in The Pleasures of Aesthetics : Philosophical Essays, Ithaca, NY: Cornell University Press, 90–125. Neubauer, J. (1986) The Emancipation of Music from Language, New Haven, Conn.: Yale University Press. Newcomb, A. (1984a) ‘‘Sound and Feeling,’’ Critical Inquiry 10: 614–43. Newcomb, A. (1984b) ‘‘Once More ‘Between Absolute and Program Music’: Schumann’s Second Symphony,’’ 19th Century Music 7: 233–50. Newcomb, A. (1997) ‘‘Action and Agency in Mahler’s Ninth Symphony Second Movement,’’ in J. Robinson (ed.), Music and Meaning, Ithaca, NY: Cornell University Press, 131–53. Ratner, L. (1980) Classic Music : Expression, Form, and Style, New York: Schirmer Books. Robinson, J. (2000) ‘‘Languages of Art at the Turn of the Century,’’ Journal of Aesthetics and Art Criticism 58: 213–18. Robinson, J. (2005) Deeper than Reason: Emotion and its Role in Literature, Music, and Art, Oxford: Oxford University Press. Rosen, C. (1972) The Classical Style, New York: Norton. Sibley, F. (1959) ‘‘Aesthetic Concepts,’’ Philosophical Review 68: 421–50. Sibley, F. (1993) ‘‘Making Music Our Own,’’ in M. Krausz (ed.), The Interpretation of Music, Oxford: Clarendon Press, 165–76. Treitler, L. (1997) ‘‘Language and the Interpretation of Music,’’ in J. Robinson (ed.), Music and Meaning, Ithaca, NY: Cornell University Press, 23–56. Further Reading

Bonds, M. E. (1991) Wordless Rhetoric: Musical Form and the Metaphor of the Oration, Cambridge: Cambridge University Press. Budd, M. (1995) Values of Art: Pictures, Poetry and Music, London: Penguin Books.

can music function as metaphor of emotional life?

177

Cone, E. T. (1974) The Composer’s Voice, Berkeley, Calif.: University of California Press. Davies, S. (1994) Musical Meaning and Expression, Ithaca, NY: Cornell University Press. Goodman, N. (1968) Languages of Art, Indianapolis: Bobbs-Merrill. Kivy, P. (1989) Sound Sentiment: An Essay on the Musical Emotions, Philadelphia, Penn.: Temple University Press. Levinson, J. (1996) The Pleasures of Aesthetics: Aesthetics Philosophical Essays, Ithaca, NY: Cornell University Press. Robinson, J. (ed.) (1997) Music and Meaning, Ithaca, NY: Cornell University Press. Robinson, J. (2005) Deeper than Reason: Emotion and its Role in Literature, Music, and Art, Oxford: Oxford University Press.

8 The Structure of Irony and How it Functions in Music E D DY Z E M A C H A ND TA M A RA B A LT E R

1 Irony is an aesthetic property, that is, a feature of objects that is necessarily relevant to their aesthetic evaluation.¹ Musicologists hold that many works of music have that property, and we agree. However, contemporary studies of irony in music² tend to focus on one kind of irony³ or on a single composer.⁴ Usually, they offer no general semantic theory of irony, showing how music can be ironic.⁵ In this paper we present a general theory that shows why there are many kinds of irony and how music displays them, analyzing one example of each musically realizable kind of irony. Like other analyses of irony in music we use texts, too, but only as supporting material, for if music can be ironic that property must lie in the music itself. Our study also differs from that of most musicologists in that we ground it in today’s philosophy (e.g., possible-world semantics) and not in philosophies of the composer’s time. We do not doubt that historical material is valuable in art interpretation, yet historical material, ¹ The concept of an aesthetic property is due to F. N. Sibley, mainly in Sibley ([1959] 1978); also see Brady and Levinson (eds.) (2001). A non-aesthetic property may be relevant to aesthetic evaluation (e.g. that the author is six years old is a good reason for believing his work is not very deep. However, tender age is only inductively linked to lack of emotional depth, and so the former fact is only contingently relevant to the latter). ² Among the latter Longyear (1970), Bonds (1991), and Sheinberg (2001) are especially noteworthy. ³ For example, romantic irony in Longyear (1970) and Dill (1989). ⁴ For example, Schumann’s lieder to Heine’s texts in Brauner (1981), Dill (1989), and Suurp¨aa¨ (1996). ⁵ Sheinberg’s work (2001), though extensive and illuminating, is no exception. It presents an array of edifying quotes on various kinds of irony and sophisticated analyses of certain specimen, but nothing like a semantic theory is attempted.

the structure of irony/how it functions in music

179

even when its relevance is not in doubt,⁶ can only show that the piece was thought (by its author or his contemporaries) to be ironic; it cannot warrant an attribution of an aesthetic property. What aesthetic properties an item has is an objective fact about it. One would not use a vintage 1800 medical theory to diagnose Schubert’s illness; similarly, one should not exclusively use a vintage 1800 aesthetic theory to analyse his music. A theory of music may use historical sources, but must also be conversant with analytic aesthetics.

2 We start, not with musical, but with verbal irony. Suppose that speaker S, to express his opinion that Joe is stupid, says: ‘‘Joe is very wise.’’ Since S thinks that Joe is not very wise, does he lie? No; S’s interlocutors readily understand that S paid Joe no compliment. S’s token sentence is ironic. What does that mean? To deny that S lied, one may suggest that ‘‘wise’’ has an additional sense: it also means ‘‘stupid,’’ and S used that sense in his token-sentence. Some writers do hold that view about metaphor: metaphors, they say, extend the literal sense of words⁷. But stupid is not an extension of the sense wise; no dictionary lists it under ‘‘wise.’’ Further, since all terms may be ironically used, that view would imply that every term also means the opposite of its usual sense. If that were so, contracts were worthless. We therefore agree with Davidson (1978) and Scruton (1997: 84) that an expression retains its literal meaning in all contexts. What, then, about S’s sentence ‘‘Joe is very wise’’? Davidson bites the bullet: he would say that what S said is false or truth-valueless. But if what S says is false, why do all S’s interlocutors, who share his assessment of Joe’s intellectual stature, say that what he said is true? Do they lie, too? No doubt, what S says implies that Joe is not wise, but this cannot be all that it means. If to speak ironically were, as Quintillian alleged, to say one thing and mean another, that is, if irony were an operator that ⁶ Cf. Daverio (1990), among others. Most of those who wrote about irony in Beethoven cite Schlegel and Tieck. Longyear (1970) does so too, yet he is careful to note we have no evidence that Beethoven was familiar with the writings of Schlegel and Tieck on irony. ⁷ G. Lakoff and M. Johnson (1980); S. Glucksberg and B. Keysar (1993).

180

eddy zemach and tamara balter

took a sentence ‘p’ into its negation, ‘‘∼p,’’ then it would be a most tedious and pointless device: why risk error and confusion instead of plainly saying ‘‘not-p’’? Yet, as we know, irony does not make conversation tedious but enlivens it, rendering it more entertaining. So, irony is not just a lugubrious way of achieving an effect that could be achieved more frugally by negation. We hold that ‘‘p’’ keeps its literal sense even when ironically used: it makes us project a situation, p, which literally satisfies ‘‘p.’’ Some writers, who, like us, hold that metaphorical (including ironical) phrases retain their literal meanings do so by stipulating special undefined operators. Searle (1978, 1979) holds that to speak metaphorically is to use an irreducible kind of illocutionary force. Walton (1990) regards the sentential operators ‘‘Make-believe (p)’’ and ‘‘Fictional (p)’’ as basic and irreducible. We think that leaves them unintelligible. Our most basic concepts must indeed be left undefined, but surely these are not among our most basic concepts? We offer a definition: ‘‘p is fictional’’ means ‘‘p obtains in a possible world other than the real one’’ and ‘‘S makes believe that p’’ means ‘‘S projects p in some world, when some real situation Cp is taken to be its counterpart.’’ Metaphors, one of us suggested (Zemach 1999b; 2001) are modal terms: the sentence ‘‘Metaphorically (Fx)’’ is true iff x is F, not in the real world, but in x’s home world: a world where x’s nature is apparent.⁸ A similar semantic analysis is developed below for irony.

3 Let us build a definition gradually, starting with a linguistic competence condition. To understand a sentence is to project a situation⁹ that satisfies it. For example, one who understands the sentence ‘‘Joe is wise’’ projects (envisages) the situation, that Joe is wise, in some world. Let ‘‘p’’ be true in w. Call the situation in w that satisfies ‘‘p’’, ‘‘pw.’’ So, ‘‘p’’ makes us project pw (‘‘→’’ stands for ‘‘projects’’): 1. (∃w)(‘‘p’’→ pw). ⁸ A common objection is that a metaphor may be logically impossible; for a reply see Zemach (2001). ⁹ Scruton (1997: 93) calls it an ‘‘unasserted thought’’; his discussion is illuminating, but that term is not, for it conflates the thought with the situation that would satisfy it.

the structure of irony/how it functions in music

181

Now, pw has a counterpart in the real world (W0); call it ‘‘CpW0.’’ CpW0 and pw are genidentical: they make one trans-world individual. For the sentence ‘‘p’’ to be ironic, however, these situations must be qualitatively opposed to each other. In symbols (‘‘x|y’’ stands for ‘‘x is the inverse of y’’): 2. (∃w)(pw|CpW0). In words: A situation p (satisfying ‘‘p’’) in some world w is qualitatively the inverse of its counterpart in the real world. Now put (1) and (2) together: 1+2. (∃w)[(‘‘p’’→ pw) & (pw|CpW0)]. In words: ‘‘p’’ makes us project a situation that satisfies it in a world, where it sharply contrasts with its real-world counterpart. Finally, we need to add that the projected situation is superior to its real counterpart. Envisaging pw, it becomes evident that CpW0, the real-world counterpart of pw, is far inferior to it. Compared to its ideal counterpart pw, the real situation looks deformed. Thus, irony is bitter-sweet: it is amusing, but also painful, to imagine intellectually challenged Joe as wise. The projected situation pw, where Joe is very wise, shames and hence mocks the real situation CpW0. Ironically to say ‘‘p’’ is therefore more rhetorically effective than to assert ‘‘not-p,’’ for the contrast demonstrates the ludicrous deficiency of the real situation. With all the conditions combined, we get (‘‘>’’ stands for ‘‘much better than’’): 3. (∃w)[(‘‘p’’→ pw) & (pw|CpW0) & (pw>CpW0)] ⊃ I‘‘p’’. The last conjunct of the definition is important. Irony is not a mere tonguein-cheek use of language: it bites. If Joe is very wise and his friend S fondly calls him ‘‘silly,’’ that is no irony: it is a kind gesture, an understatement used to compliment. A similar understatement is used by the basketball players who call their tallest ‘‘shorty.’’ Such expressions neither bite nor mock, so the trope involved is not irony. An amicable understatement, or an exaggeration, used as endearment, or as an underhanded, coy, praise, is not irony. Irony involves criticism, a dramatic exposition of depravity. That a given token is ironic is linguistically and extra-linguistically indicated. In our example, the highly laudatory words ‘‘very wise’’ put us on guard. The situation that satisfies it is projected and, having had its effect (shocked amusement, due to the known facts about Joe) we see that our interpretation must be further elaborated. The methodological constraint

182

eddy zemach and tamara balter

of charity is invoked: interpreting S’s words literally is uncharitable to S, making him a liar, or dead wrong. We cannot ignore the literal meaning of ‘‘very wise,’’ so, we take it to have a modal rider. Another hint that the sentence is ironic is given by S’s tone of voice. It is taken to have a semantic role, indicating the modal operator, irony. Music is not a language, so verbal irony in music is extremely rare. However, we hope to have shown that use of words is not essential to irony. Its essence lies in projecting a situation that makes reality look deformed, and such projection need not be prompted by verbal means. It may be evoked by certain real-life situations, too.

4 Joe is a clumsy and reckless driver but never had an accident. Jane is an adept and scrupulous driver, yet she is killed in a traffic accident. That situation, we say, is ironic. What is an ironic situation? Like verbal irony, situational irony invokes an ideal situation that is compared with the one at hand; the two situations are genidentical yet dissimilar, the real one being a deformed version of its projected counterpart. In verbal irony a speaker uses language to indicate an ideal situation in a possible world; in situational irony the ideal situation is indicated by our norms. We have strong intuitions on how things should be: what is right, just, and fair in a particular case; how the world should then be. It is as if we all jointly do what a speaker does in verbal irony: we indicate a counterpart of the real situation at hand. The present situation looks like a wizened, degenerate version of its projected counterpart. It looks funny yet terribly wrong. The contrast between it and its possible counterpart makes us regard it with a sad smile. In our example, the real situation CpW0 makes us project a just and proper situation pw, where people get what they deserve. There, it is Joe and not Jane who is killed in a motor accident. The projected situation pw makes us see the real situation CpW0 as a grotesque and deformed version of itself. Thus, the possible situation serves vividly to denigrate its real counterpart. An ironic situation is modal since it requires interworld comparisons. A sentence is modal if its evaluation in reality depends on its truth in other

the structure of irony/how it functions in music

183

worlds. Similarly, a situation is modal if appreciating it requires that we compare it to a counterpart; if it is ironic, that comparison finds it wanting. The real situation CpW0 (the good driver is killed and the bad comes to no harm) is ironic in that it projects a just counterpart, pw, which puts it to shame. Let us define: 4. (∃w){[CpW0→(pw)] & (pw|CpW0) & (pw>CpW0)} ⊃ I(CpW0). In one of the very few articles devoted to situational irony Lee Miller (1976) requires an additional condition to hold for an ironic situation. An ironic situation, he says, though it runs contrary to the hopes and efforts of those involved, displays a particular propriety: a kind of poetic justice. But, to use Miller’s own examples, what poetic justice is there in the wealthiest woman winning the lottery, or in a father’s constantly yelling at his children to be quiet? Miller’s general diagnosis, we think, is correct: ironic situations are uniquely apt; they are characterized by extra significance. But the aptness, we think, is that which makes us regard the two situations as counterparts, and the extra significance is the demonstration of the wretched nature of reality. We see the ironic situation as profoundly instructive, almost prooflike, for it shows the reality as a degenerate version of what there should be. Theorists of music do not often discuss situational irony, either. This is strange, for two central, universally agreed upon, characteristics of music make it amenable to situational irony. First, music raises expectations: competent listeners project situations they consider right for the (musical) conditions at hand. Second, listeners compare and contrast the anticipated situation (a musical event) with the one that does occur in the work. These traits, however, are necessary, but not sufficient, conditions for musical irony, for contrasting situations and frustrating expectations have many other uses too: they create tension, drama, or simply add interest to the work. When does the presence of these features indicate situational irony? A situation may challenge our assumptions on what is proper and right, expressing fundamental contradictoriness, by defying our expectations about the proper location of musical material of a certain kind. Thus, a piece that begins with a closing gesture flies in the face of our expectations from a musical event proper to the beginning of a work (in the given style). Such ‘‘closing-beginnings’’ exist in the music of Haydn and later composers, mainly in string quartets. Kramer (1973) and Lochhead (1979) discuss the contrast between these two states, the projected and the actual

184

eddy zemach and tamara balter Allegretto

Violino1

Violino II

Viola pizz. Violoncello

5

poco cresc.

poco cresc.

poco cresc.

Example 2. Beethoven: String Quartet in F Major, Op. 135, i, mm. 1–10.

(Kramer names them ‘‘clock-time’’ and ‘‘gestural-time’’), in Beethoven’s String Quartet Op. 135 (Example 2). The musical event at the beginning of that work suggests closing or cadencing. It thus abuses its actual (clock-time) place at the head of the piece. At that hallowed location we expect (hence project) normative tonic stability and clear periodicity. Instead, we get a compound ten-measure sentence that closes with a final cadence (in m. 10 all four parts have an F in four different registers). Two surprising Phrygian cadences open the piece, each of which is followed by a sprightly three-note motive that mocks its serious mood and questioning gesture (mm. 1–4). A nonchalant sequence follows and further develops the mocking three-note motive. A terminal cadence occurs in m. 10, and then there starts a new, unrelated musical idea. That beginning makes a mockery of the projected normative beginning. It undermines the belief in the good order of things and includes comic elements. Thus, we say that the actual beginning of Op. 135 is ironic. One of the few who investigate situational irony in art, Lars Ellestr¨om, uses Magritte’s painting In Praise of Dialectic as an example. It shows the inside of a room through an open window, but that room only contains the fa¸cade of another building (1996: 202). Instead of a family scene we get an inferior

the structure of irony/how it functions in music

185

situation: a forbidding externality.¹⁰ The frustration of our expectations by the exterior wall that stands where an inside should have been parallels the above closing-beginning in Beethoven’s Op. 135. In Magritte’s work ‘‘the inside is the outside’’ (ibid.); likewise, Beethoven shows that the beginning is the end.¹¹ The competent listener’s tendency to project the norm creates a paradox. Beethoven seems to resolve it at the end of the work, but that solution only intensifies the irony. The ‘‘closing’’ sentence that opened Op. 135 reappears in its end, doing, so to speak, its ‘‘duty’’ to conclude the piece. Yet this ending redoubles the irony: after we consented to forgo our previous conceptions and accept a new musical idiom wherein that musical event has the role of a commencing gesture, the new norm is forfeited. We must abandon our newly acquired norm and go back to the old one. The opening sentence concludes the very work that taught us not to expect it there. The destructive irony here is radical and most effective. Another way to express situational irony is by frustrating a meticulously prepared expectation for harmonic resolution or a key area. The sudden dislocations of the apparent tonal stability in the finale (sonata-rondo) of Beethoven’s violin sonata Opus 30, No. 3 are ironic, says Longyear (1970: 656). There are two tonally surprising shifts in that finale. First, after the finale’s second episode there is a false reprise of the theme; that false reprise, in the major mediant (mm. 133–6) had been carefully prepared, yet is abruptly abandoned. The movement proceeds to a preparation for a correct recap (m. 141). Second, in the coda there is a surprising shift to the flat submediant (m. 177) after a prolonged dominant; this unexpected tonality is ‘‘stated in a ‘vamp’ accompaniment’’ and precedes another statement of the main theme in a wrong key (Longyear 1970: 656–7). Longyear believes that these shifts show romantic irony, for they destroy an illusion. Leaving aside, for now, the question of romantic irony (but see Section 8 below), we ask: is it situational irony? We think not. Even if we expect neither the ¹⁰ Although the situation that contrasts with the projected one is not real, but merely painted, it is one we see in front of us, so it serves here as ‘‘real.’’ Other substitutes for reality are situations we see in front of us on the stage. (We thank Kathleen Stock for pointing it out to us.) ¹¹ Following Muecke (1969), Ellestr¨om writes ‘‘An ironic situation is a situation that displays a striking incongruity between an expectation and an event’’ (p. 210). But that condition, while necessary, is not sufficient, for it fails to distinguish an ironic situation from a merely surprising one. Ellestr¨om himself goes beyond it when he claims that in music, ‘‘when we feel that two contrasting moods are mutually exclusive, and yet in a way make sense when jumbled together, it is irony that tickles our ears’’ (p. 205). This is reminiscent of Miller’s (1976) notion of the overall appropriateness of the ironic situation.

186

eddy zemach and tamara balter

major mediant nor the flat-submediant, their surprising appearance does not amount to contradiction, as required in situational irony. Listeners are led to project a more normative, perhaps a banal, tonic area, which does not materialize; but the projected situation (tonic area) does not mock the realized situation (major mediant) and does not contradict it. There is, however, a late movement by Beethoven, the scherzo in String Quartet Op. 131, where unexpected modulations to the mediant may be retrospectively interpreted as cases of situational irony.¹² In that movement Beethoven modulates to the mediant very early in the scherzo sections, then directly moves from the dominant seventh chord of the mediant (marked with a fermata) to the tonic, E major. That shift brings to mind the deceptive move in the coda of the violin sonata Op. 30, no. 3, from a dominant seventh chord with a fermata to the flat submediant. In Op. 131 the deception is double, since what may sound as the submediant in G# (an E major chord) is, in fact, a correct (tonic) reprise of the A section of the scherzo (Example 3). A deceptive move (to bVI) turns out to be a move to the tonic. Molto poco adagio

30

A

Tempo1

più

più

dim .

più

dim .

un poco più adagio

più

dim . 39

ritard.

in Tempo

Example 3. Beethoven: String Quartet in C-sharp Minor, Op. 131, v, mm. 30–48.

¹² Longyear (1970) briefly mentions this movement, calling it, too, a case of romantic irony.

the structure of irony/how it functions in music

187

These early modulations to the key of the mediant are ironic because, once it becomes clear that the E major chord is a correct beginning of the reprise, the modulation to G# minor looks wrong. Beethoven makes it look as if this (unconventional) modulation to the key of the mediant was a mistake by writing no modulation back to the home key. Instead, he makes the players slow down, and seem utterly confused. They hesitate, as if they wonder how they arrived at this G# minor area and got stuck there, not knowing how to return to E major. So what they do is just give up and start all over again, without modulation. That makes it look as if the modulation to the G# minor area is an error they made. Now, to clinch the point, it does make sense to modulate to G# minor toward the next movement, which is in G# minor, and is immediately linked to the present movement without a pause; but there, no such modulation occurs. It thus appears as though things got mixed up; the said modulation came at the wrong time. The composer has, then, intentionally made the actual situation to look like an aberration, a deformation of another, normative, one.

5 Dramatic irony is a modification of situational irony. A situation in a drama is said to manifest dramatic irony if its protagonist does not comprehend the situation she is in, while the audience does. For example, the scene where Rigoletto cheers on the gang of courtiers not realizing that the lady they abduct is not Countess Ceprano but his own daughter, Gilda; or, in Wilhelm Tell, the scene where Governor Gessler boasts how he will subdue the Swiss, not knowing that his Swiss assassin is already in place, ready to kill him. Analysing this we see that in dramatic irony, too, a counterpart to the present situation is projected, which is the opposite of the situation at hand. That projected situation, however, is good for its projector only; it is not (as in situational irony) recognized as good by us all. The situations Rigoletto and Gessler project are good for them, but sharply contrast with the actual situations, which we now see as their negative counterparts. Dramatic irony is modal, for it interprets the real situation via a possible counterpart. Now, in dramatic irony the projected situation is the one that looks ridiculous; does this mean that dramatic irony deviates from the basic structure of irony, where reality is mocked and criticized? No, for

188

eddy zemach and tamara balter

the inadequacy of the situation projected makes the present protagonist, who projected it, look ludicrous. The unrealistic protagonist is, therefore, the butt of the dramatic irony. It is pathetic to be so na¨ıve as to believe an idealized counterpart of grim reality. Let us define: 5. (∃w){[CpW0→S (pw)] & (pw>S CpW0) & [BS (w = W0)] & (w =W0)} ⊃ I(CpW0). In words: A situation is fraught with dramatic irony if its protagonist s projects a counterpart situation (that is much better for s than the real one) and falsely believes it is real. Dramatic irony is the most prevalent kind of irony used in music. In every B-movie the music forewarns the spectators, before the hero has any inkling of it, that the new man is the villain, or that the heroine is about to fall in love. Music is a forecaster, interpreter, and prompter in all operas, so dramatic irony is ubiquitous in them (see Burnham 1994 and Sawyer 1999). The example we chose (the cavatina of Barbarina from Le nozze di Figaro) is, however, subtler; it is threefold. The protagonists see it as a vulgar joke; the audience perceives it as dramatic irony, but, upon reflection, may also take it as meta-poetic criticism. Barbarina’s cavatina is the only aria in Le nozze di Figaro written in a minor key from start to finish (Example 4). Suffused with appoggiaturas and chromatic chords (diminished seventh and augmented sixth chords) it expresses deep sorrow, yet it is sung by a minor character. Dent (1960: 112) thinks the cavatina has a structural role: it signals the opera’s transition from a light to a somber atmosphere; but this does not explain why Barbarina, 29

- du

-

ta! ah, chi sa do - ve sa - rà?

E mia cu - gi-na?

e ilpa - dron,

co- sa di - rà?

co - sa di - rà?

Example 4. Mozart: Le nozze di Figaro, Act IV, Scene I, No. 23, Cavatina, L’ho perduta, mm. 29–36.

the structure of irony/how it functions in music

189

an unlikely figure to express a deep sentiment, is chosen to introduce that somber tone. Also, as Kivy (1980) notes, the crying motive in the cavatina is appropriate to a greater loss than that of a pin. What is the point of that conspicuous incongruity? The mystery deepens when we note that the entire section was added by Da Ponte: it is not found in Beaumarchais’ play. Da Ponte also gave Barbarina her name (Beaumarchais calls her ‘‘Fanchette’’). Since the libretto is quite faithful to the play (it omits, but does not add scenes!) and only one other very minor character (Don Curzio) is renamed, one cannot but wonder why this scene was added. Is it only for the sake of the joke? This cannot be the whole story. Had Mozart only wished to mock Barbarina for the weight she attaches to her tiny loss he could have set her words to a recitative; he would not have written for them dark, heart-rending, music that makes all jokes go sour. A pin may be the most fungible object there is. It is (even in the eighteenth century) a mass-produced, cheap, common object (the Count says women have them everywhere), hence most easy to replace. Indeed, Figaro immediately replaces the pin Barbarina lost by taking one off Marcellina’s dress. To him, her moaning seems silly: pins are indistinguishable; they are an entirely generic product. A dirge for a lost pin is absurd, and may be funny. But the music disallows that interpretation: it tells us that this cavatina has a serious import that is incompatible with the frivolous spirit of the opera-buffa. The pin must have a special significance here that neither Figaro nor Barbarina grasp, but the audience should. The pin Barbarina lost was the seal of Susanna’s letter, setting out (on her wedding night!) a tryst at which the Count will deflower her. She pretends to promise him the prize he wanted to take by force, that is, the right to break her seal. Thus the seal of Susanna’s letter symbolizes her virginity. That symbol is further developed when, upon contact with the Count, it draws blood. Ironically, the roles are here reversed: it is not the virgin Susanna, when her seal is broken, who bleeds, but the man who tries to break her seal. In symbolic, ironic, revenge, the seal (the pin that seals her letter) pricks his finger. Virginity is the subject of the entire opera: it starts when Almaviva wants to reinstitute his feudal droit de seigneur to the virginity of his vassals. So, if the seal is virginity, then in mourning the loss of the seal Barbarina really mourns the loss of virginity. What is lost is not just a pin: it is virginity. Indeed, the Count also embraced Barbarina, kissed her, and promised to give her whatever she wants if she would love

190

eddy zemach and tamara balter

him (no. 21). So it may be that it is her virginity too, not only Susanna’s, that (unknowingly) she mourns. Barbarina, of course, never realizes the full meaning of her lament over the pin, but there is no question that Figaro does. To Marcellina he explicitly says: ‘‘Ah, that ‘pin’, mother, is the same one the Count lately renounced.’’ The right the Count lately renounced is, of course, the infamous Jus prima noctis. That is one level of dramatic irony; but there is a higher one. All of Mozart’s operas present a view of love according to which love is generic: the identity of the beloved is of little importance. In Cos`ı fan tutte the wise cynics (Don Alphonso and Despina) cause the girls to switch their lovers—and then switch them back again. In Die Zauberfl¨ote, Papageno admits he is attracted to all women; he wants, as he often says, a (generic) woman (Ein M¨adchen oder Weibchen ... ). In Figaro, Cherubino voices that sentiment when he candidly describes (e.g., in Voi che sapete) the love he feels as generic: it may shift objects or pick one at random. Susanna is the lady with the seal, while Cherubino is one whose patente has no seal (end of Act II). Most blatant in his generic attitude to erotic objects is, of course, Don Giovanni. In Figaro that attitude is manifested by the lewd Count, by Figaro’s condemnation of all women (the aria Aprite un po’ quegli occhi), by the fact that every lover in this opera accuses his or her partner of infidelity, and by the ease with which one person can pass, even during an assignation, for another. Such is the society depicted in Mozart’s operas. Here, in a statement that transcends that society, Mozart criticizes it; it lost, he says, something valuable. People in that society cannot see themselves as tragic, but we can, so the music advises us of the composer’s judgment. Barbarina, then, speaks for the composer. That is why her cavatina is so profoundly sad, so out of step with the carefree levity of the musical material in Figaro. Figaro, for instance, even when given the most awful news, is assigned light and cheerful music. No one in this opera has notes as sad as those in Barbarina’s lament over a lost pin. Thus it is not only the plight of women that this plaintive music laments, but the entire notion of romantic, non-generic, individually ‘‘sealed’’ love; it is non-fungible love that Mozart yearns for. Why is Barbarina chosen to express that yearning? Because she is the proverbial Ing´enu, too na¨ıve to play the games of dissimulation and deceit that all others in that society

the structure of irony/how it functions in music

191

play. Muecke’s example of Ing´enu irony is the child who cries out loud (in Andersen’s The Emperor’s New Clothes) that the King is naked. In Muecke’s words, ‘‘even simple innocence or ignorance may suffice to see through the woven complexities of hypocrisy’’ (1969: 91). Simple-minded Barbarina is thus the perfect choice for noting how sad the world of her sophisticated betters and elders is. Muecke says: ‘‘the ing´enu is himself sometimes a victim of irony, exposed as well as exposing’’ (92). Surely, that is true of Barbarina; moaning over a lost pin is indeed farcical, but it exposes the far greater farce of the Almaviva court, which, essentially, is Mozart’s own society. Another contribution of the music to the sense of this cavatina is by alluding to the Countess’s big recitative and aria in Act III (no. 19). Barbarina’s cavatina ends abruptly with exactly the same half cadence—a sustained Italian augmented sixth chord resolving to a dominant—that ends the Countess’s recitative, ‘‘fammi or cercar da una mia serva aita!’’ (‘‘[he] makes me seek help of my servant!’’) (Example 5). The reference is to Susanna, but the musical allusion makes us think of Barbarina, who also expresses the agony of the Countess. Another link between Barbarina and the Countess is structural: the cavatina of the countess, Porgi amor, opens Act II of the opera; Barbarina’s cavatina opens Act IV. Barbarina’s loss is thereby associated with the loss of the Countess, a loss of her ‘‘treasure’’ (mio tesoro): love. Finally, note that the main aria of the Countess (Dove

24

fam - mi or

cer- car

da u

-

na mia ser - va

ai - ta!

Example 5. Mozart: Le nozze di Figaro, Act III, Scene VIII, No. 19, Recitative, mm. 24–25.

192

eddy zemach and tamara balter

sono, which mourns the loss of happy moments and the forfeiture of oaths sworn by her husband) is written to the theme of Mozart’s Agnus Dei in the ‘‘Coronation’’ Mass, K. 317 (completed in 1779). Mozart was not in the habit of pillaging his own works for melodies. The rare hyper-textual connection shows, then, an affinity of moods. A critical message hidden behind the jocular face: that is the gist of dramatic irony.

6 Another kind of irony is total rejection: treating all alternatives to reality as flawed. Muecke (1969) calls that kind of irony, made famous by Kierkegaard ([1841] 1989), ‘‘general irony’’; others call it ‘‘cosmic irony,’’ ‘‘infinite irony,’’ etc. It is a negative attitude to all possibilities, whatever they may be. General irony is therefore anti-modernist, for modernism believes that a better alternative to reality can be envisaged. Like all other kinds of irony, general irony, we say, is modal, for its evaluation in reality depends on its evaluation in alternative worlds. A situation that displays general irony does so by invoking its counterpart. In the kinds of irony discussed above the projected situation pw is considered better than the real situation. General irony, however, implies that no alternative to the real situation can make a difference; all possibilia are equally flawed. That attitude is typical to fin de si`ecle symbolists bored by everything and to existentialists for whom the depravity of a situation lies not in its quality but in its very being (as Sartre puts it, it is its facticity that makes a situation unbearable). Nineteenth-century works like Goethe’s Die Leiden des jungen Werthers, turn-of-the-century Decadence works such as Picasso’s Les Demoiselles d’Avignon, and mid-twentieth-century works such as Sartre’s La Naus´ee and Celine’s Voyage au bout de la nuit, all flaunt general irony. The irony of Romanticists and Modernists presented the world as socially, i.e., redeemably, flawed; general irony regards every situation with equal ennui or disgust.¹³ According to Kierkegaard, that attitude is essential ¹³ Several theorists assimilate general irony to romantic irony. Esti Sheinberg is one: romantic irony, for her, has three kinds: aesthetic distance (romantic irony proper), infinite negation, and existential irony (2001: 61–2). That classification has some merit, yet it misleads. General irony condemns all alternatives, while romantic irony is corrective: it criticizes an element in the work, or, by shattering the illusion, alerts us to the nature of the work as an artefact.

the structure of irony/how it functions in music

193

to achieve religiosity: only those who sense the pull of nihilism make the ‘‘jump’’ to religion. Wittgenstein, in the Tractatus, holds a similar view.¹⁴ We define that kind of irony as follows (Where ‘‘D’’ is ‘‘depraved’’): 6. (∀w){(CpW0→ pw) ⊃ [ ∼(pw>CpW0) & D(CpW0)]} ⊃ I(CpW0). In words: a situation is ironic if all the counterparts it projects are as flawed as it is. Patricia Herzog (1995) claims that Beethoven’s Diabelli variations express an ironic-sceptic attitude to life. Her characterization of that attitude makes it a case of general irony. These variations, Herzog says, take the listener to diverse and contrary worlds; just as one begins to get accustomed to a world, expecting it to endure, one is ejected into its opposite. It is a narrative of successive rejections that has no goal and no end; so (as we put it), it manifests general irony. William Kinderman (1987) interpreted that work in a very different way: according to him, these variations do have a goal; their unfolding is a teleological process through which Beethoven overcomes the waltz.¹⁵ The finale of Beethoven’s Op. 101 (The Deed) generates a similar dispute. Wilfred Mellers (1983) finds in it cosmic irony; Robert Hatten disagrees. Although Hatten, too, thinks the movement contains a ‘‘Romantic-ironic projection of the persona,’’ he takes it to enhance the achievement of ‘‘a positive outcome’’ (1994: 186). As these debates show, it is hard to find works that consist of incongruities between segments, all of which are equally presented as wrong, and therefore may be said to express equally inadequate views. We saw that even a work filled with incongruities and shifts in level of discourse can be seen as a teleological narrative; alternatively, the negation may be construed as a device for creating artistic distance. Perhaps only synchronic presence of diverse levels, styles and topics can clearly embody general irony. The music of Mahler and Stravinsky often has concurrent presentations of noncongruent layers, though not throughout entire movements. Stravinsky, in the first miniature piece from Three Pieces for a String Quartet (1914) does that, too. ¹⁴ ‘‘If there is any value that does have value, it must lie outside the whole sphere of what happens and is the case. For all that happens and is the case is accidental. What makes it non-accidental cannot lie within the world, since if it did it would itself be accidental. It must lie outside the world’’ (6.41). ¹⁵ Kinderman’s interpretation is that the clashing variations in the middle section of the work are indeed ironic, as evidenced by the direct parodies of the waltz. The final variations, however, go beyond irony to uplift the original waltz; there, the waltz is used to create lyrical, richly textured, variations.

194

eddy zemach and tamara balter

Conversation and communication between partners is, perhaps, the commonest metaphor for the string quartet genre.¹⁶ We suggest that in the first of his Three Pieces for String Quartet (later named ‘‘Dance’’) Stravinsky goes all out to combat it. Like many pieces by Stravinsky, ‘‘Dance’’ has stratified layers, but here stratification is radical: each layer includes an independent, musical idea that repeats as an ostinato over and over again, to the very end. The first violin repeats five times a 23-beat phrase (the fifth repetition is shortened) consisting of a major tetrachord (G-A-B-C). It is reminiscent of a naigrysh, the Russian folk instrumental ‘‘dance-untildrop’’ tune (Watkins 1994: 260); yet here the phrase is much longer than in typical naigryshi. The viola and cello repeat many times a much more monotonous 7-beat ostinato; the second violin alternates between single and double statements of a descending minor tetrachord (F#-E-D#-C#) that sound like impulsive interruptions. These independent layers clash with each other. Clashes occur within layers, too; for instance, although the ostinati of the viola and of the cello are of the same length, the viola’s D clashes with both the Eb and the Db of the cello. The relation between the layers varies due to their different phrase lengths. A three bar dissonant drone (a minor ninth) in the viola part starts and ends the piece; added to the extended folk tune in the first violin and to the repeating ostinati it renders the piece a grotesque folk dance. A performance of that piece creates an impression of miscommunication, where each interlocutor repeats, ad nauseam, an unrelated idea. It pictures incommensurable non-coalescing world versions that cancel each other out. It is a world that does not and cannot make sense. In brief, it carries general irony. Watkins argues, in detail, that this piece is influenced by cubism, in particular, the cubist appeal for simultaneism.¹⁷ We accept that, but note that cubism itself has a nihilistic streak; it criticized classical painting for giving a definite, unambiguous, ‘‘literary,’’ sense to the visual world. Cubism had another streak, too, a constructive, modernist streak exemplified, e.g., by Fernand Leger. But it is the first, nihilistic, streak found in Decadence art from Alfred Jarry (Ubu roi) to Jean Cocteau and Dadaism that is relevant to Stravinsky’s work of that period. ¹⁶ For a comprehensive discussion of that metaphor, see Mara Parker (2002). ¹⁷ ‘‘Dance’’ was conceived by Stravinsky after discussing a possible collaboration with Jean Cocteau on his ballet David (Watkins, 1994: 258–65). We thank Marianne Kielian-Gilbert for this reference.

the structure of irony/how it functions in music

195

What most inclines us to interpret the Stravinsky piece as divulging general irony is the original choice of the string quartet as its medium, while flagrantly defying the tradition of the string quartet as a paradigm of cooperation. Stravinsky’s piece retains enough of the early eighteenthcentury tradition, for instance, a longer and more interesting melody in the first violin part, to make the link with it evident.¹⁸ It thus seems to be a deliberate attempt to ironize and negate the notion of a joint effort to make sense of the world. Cooperation is ruined precisely where a musician would most look for it: in the string quartet.

7 Instead of showing the depravity of the real situation by projecting an ideal situation that sharply contrasts with it, one may project a situation that is much worse than the real one, but does not sharply contrast with it. This kind of irony is not immediately recognized as such: we first take the projected situation pw to be real. So, when we realize our mistake, we see the real situation in a new light. Since the projected bad situation pw could so readily be mistaken for the real situation CpW0, we now see the latter in the light of the former, that is, as depraved. That is parody: a projected worse counterpart reveals the real situation as also flawed. Swift’s ‘‘A Modest Proposal’’ proposes to relieve the poverty in Ireland by selling Irish children as meat to be ‘‘Stewed, Roasted, Baked or Broiled.’’ Since this piece is ironic, does it mean the opposite of what it seems to say, i.e., that the Irish should not be treated like cattle? No; that semantics makes irony otiose. The text projects a situation where cannibalism is advocated. Only later we realize that, by hyperbole, Swift criticizes a real situation (cf. Booth 1974: 105–20), i.e., that a tacit modal operator precedes the text. In the possible situation pw the proposal is seriously made. The depravity of the real CpW0 is revealed when one notes its proximity to the projected pw. Let us define: 7. (∃w){(‘p’→ pw) & ∼(pw|CpW0) & (pw
196

eddy zemach and tamara balter

In words: a text that projects a situation which is far worse than the real one yet does not contrast with it, is ironic (a parody). Elsewhere, one of us argued that pictures are like binoculars for possibilia: they enable us to inspect items from other possible worlds.¹⁹ When the picture is a caricature, the possible item seen in (i.e., via) the cartoon is assumed genidentical with some real item. In visual parody an item is projected via a cartoon and is grasped as a counterpart of a real item, say, a famous politician. Usually, a caricature mocks a well-known person by projecting an object that is seen in the cartoon, capitalizing on their assumed genidentity and perceived similarity. Since the two are similar, we tend to impute to the real person properties of his projected counterpart. In parody, these are negative properties; once you see (thanks to the artist) that the projected object has them, you see them in the real object, too. Take a cartoon that projects a man with a huge nose. You say ‘‘voil`a, General de Gaulle!’’ De Gaulle in reality was not so deformed, he did not have such a huge nose; but in the possible world you now descry, he does. The counterparts, the one with monstrous proboscis, pw, and his real counterpart, CpW0, are similar enough for you to impute to the real one the deformities that only his projected counterpart has. The funny, possible, projected De Gaulle illuminates the real one, so that you now see the real De Gaulle in the light of his unreal, deformed counterpart. Unlike literary and visual parodies, musical parody is not a possible object we project but a real piece of music which suggests that another real object, being its counterpart, also has its manifest drawbacks. Call the parody ‘‘pW0,’’ for it occurs in the real world, and its alleged counterpart ‘‘CpW0’’; pW0 says of itself that, essentially, it is CpW0 (an older work, style, or genre). By presenting itself as CpW0’s true self, pW0 mocks it and puts it to shame, since if CpW0 is as flawed as pW0 evidently is, then it is not good at all. Musical parody is thus immediate, funny, and vicious. Not every musical borrowing is a parody, but every musical parody is a borrowing—one that sheds unfavourable light on its source. It may be a mere irreverent joke that entertains by casting aspersions on a venerated predecessor, or a serious attempt to reform our taste by exposing the older model. ‘‘Look at me,’’ the parody says, ‘‘see how flawed I am? Well, for all intents and purposes I am a doppelg¨anger of the august CpW0; it is no better ¹⁹ For a discussion of seeing in, cf. Wollheim (1986, 1987, 1998) and Zemach (1999a, 2002).

the structure of irony/how it functions in music

197

than I am.’’ The spoof pW0 tries to persuade us that the grand old work (or style, or genre) is really, deep down, not better than it, its counterpart, is. The Theme-and-Variations form is particularly suited to parody since it provides clear reference to the item parodied: its theme. Usually, a claim that a piece of music pW0 parodies another requires weighty evidence; in the case of theme-and-variations that claim is obvious, for each variation is assumed genidentical with the theme that it is supposed to elaborate. When the theme is by another composer, the variations, functioning as explicit quotations, have an obvious external counterpart which they paraphrase and may ridicule. Consider Beethoven’s Diabelli variations. In some of them Beethoven abandons many features of Diabelli’s waltz: its structure, its melodic and harmonic framework, etc. Those variations that do maintain salient qualities of the theme such as its contour and the oft-repeated chord or single note, parody it. Indeed, some traits of the theme seem gross; no wonder Beethoven considered it banal and ludicrous (cf. Schindler [1860] 1996 and Kinderman 1987). Yet, as Tovey (1972) showed, Beethoven found in it a rich vein for a profusion of diverse variations, and so only some of them reflect his low opinion of the theme. We shall confine the illustration of parody to two variations, nos. 1 and 13. The first variation in the cycle is a pompous march (Example 6). Apparently, that variation was one of the latest to be written, c.1823 (Kinderman 1987: 29–34), Beethoven placed it first only later, perhaps in order to declare his ironical intentions at the outset. The ironical treatment of the theme is manifested here, first, by its being a radical antithesis of a waltz; it is a ponderous, pompous march in quadruple meter. This is a deliberate defiance of convention, which demands that one begins a work in this form with a variation that closely resembles the theme. Another indication that the Diabelli waltz is here made fun of is Beethoven’s adherence to its most ludicrous features: eleven repeats of the tonic triad, a Alla Marcia maestoso

Example 6. Beethoven: 33 Ver¨anderungen u¨ ber einen Walzer von A. Diabelli, Op. 120, variation 1, mm. 1–5.

198

eddy zemach and tamara balter

monotonous rhythm, and repeats of the strange dynamics, including a great many sforzandi. These sforzandi may have added a mischievous quality to the waltz, but in Beethoven’s variation they amplify the heavy pretentiousness of a march. That is a veritable cartoon: contempt for Diabelli’s waltz is expressed by the variation’s being cast in a contrasting form (a march) while adhering to the waltz’s ridiculous features, magnified and exaggerated. In reality the old CpW0 (the waltz) looks much better than the new pW0 (variation no. 1) makes of it. This does not imply that the variation is of no aesthetic value: it is a good cartoon, a successful parody, brilliantly highlighting the depravity of Diabelli’s waltz. Variation 13 is the most biting parody in the cycle (Example 7). It reduces the theme more than any other variation in the cycle, and negates (brings ad absurdum) almost every element of the waltz. It shows its contempt to Diabelli’s theme by ignoring it. The many chord-repeats, which are, perhaps, the most preposterous feature of the waltz, are mocked by being reduced to a thunderous silence in the form of surprising general pauses.²⁰ Half of the variation (46 beats out of 96) consists of silences only! While the first variation in the cycle ridicules Diabelli’s chords by exaggerating them, variation 13 hints at them briefly with a dotted rhythm that suggests a march, raising expectations for yet another pompous march, but instead Vivace

1.

2.

9

cresc.

Example 7. Beethoven: 33 Ver¨anderungen u¨ ber einen Walzer von A. Diabelli, Op. 120, variation 13, mm. 1–16.

²⁰ Replacing chord-repeats with GPs is reminiscent of Laurence Sterne’s ploy in Tristram Shandy. Sterne fills a page with asterisks or blank spaces to make fun of a prevalent literary convention that used such devices to avoid description of sexual matters. Mark E. Bonds (1991) draws a similar analogy between the irony of Sterne and Haydn’s irony. Cf. Imeson (1996).

the structure of irony/how it functions in music

199

we get a long rest followed by two notes in a weak small voice. Finally, the gentle turn motif from the waltz becomes a short, grotesque march gesture.

8 An interesting case of irony is when the projected pw and its real counterpart CpW0 are not only genidentical (the same item in different worlds), but fully identical: the very same situation. In that case the work mocks some or all of its own features: it may deride its own harmonies, melodies, voice leading, and emotive content; it may poke fun at its poetics and deny its own candour and aesthetic value. Such self-denigration and self-degradation is called Romantic irony. According to the argument of the present article romantic irony is the epitome of all irony, since, as we argued, all irony is, essentially, self-directed. Irony, we said, is always based on the claim of one item that another item is, deep down, genidentical with it. Only a counterpart of a given item can impute its own traits to that given item. Irony would be neither funny nor clever without that dramatic assumption of a trans-world, or (in musical parody) at least inter-temporal, genidentity. Self-ridicule, then, is the basis of irony as such. Romantic irony only takes it to the extreme by making the mocking situation not only genidentical, but plainly identical, with its target. We define: 8. (∃w)(CpW0→ pw) & D(pw) & [CpW0= pw) ] ⊃ I (CpW0) In words: a situation that projects a depraved situation which turns out to be none other than it, itself, is ironic. We mentioned that some writers compared works of Haydn and Beethoven to novels of Sterne; these writers attributed romantic irony to both composers.²¹ The coda that ends Beethoven’s string quartet Op. 95 is often regarded as an archetype of romantic irony in music.²² An earlier example, however, one that has not been noted, is Haydn’s coda in string quartet Op. 76, No. 1, in G Major (Example 8). Though the entire quartet is in the major mode, its final movement is mostly dramatic, in the Sturm und Drang style, in the minor mode with orchestral sound effects (e.g., ²¹ Such comparisons date back to the eighteenth century. See Bonds (1991) and Kinderman (1996). ²² See Longyear (1970); Hatten (1994, chapter 7).

200

eddy zemach and tamara balter

179

pizz.

pizz.

pizz.

184

cresc. arco

pizz.

cresc. arco

pizz.

cresc. arco

pizz.

cresc.

Example 8. Haydn: String Quartet in G Major, Op. 76, No. 1, iv, mm. 179–88.

the opening tutti in unison) and sigh gestures. Only in m. 139 does the mode change from G-minor to its parallel key. In m. 180 a buffa-like coda starts: the violin has a banal, music-box-like, melody while the other instruments accompany it in pizzicato. That melody repeats twice and ends in a laughing gesture: a quick descending scale marked staccato played simultaneously by three instruments (mm. 187, 195). The coda starts on the last beat of m. 180 in a delay, after the expectation for a tonic on the downbeat of this measure is frustrated (in the first violin part in m. 179 the leading tone is left in the air, unresolved). It thus seems that the coda pokes fun at the very lofty sentiment the movement has just so pathetically avowed. Such humorous, twaddle codas are found in other movements by Haydn, too, but the debunking effect of this coda is similar to the effect of the widely discussed coda (the addendum, as Hatten (1994: 187) calls it) that closes Beethoven’s Op. 95: deflating bathos that sharply contrasts with the preceding noble high drama. In both works the abrupt switch to the comic

the structure of irony/how it functions in music

201

is a drop from a zenith to a nadir level of discourse. That is romantic irony. Unlike in Beethoven’s Op. 95, the coda in Haydn’s Op. 76 No. 1 does not directly follow the serious, dramatic minor music; a few measures in major precede it. Yet the precipitous drop from the high genre suggests that here Haydn mocks, hence defiles, his own grandiose gesture in this movement. It represents the serious, pathetic portion in minor, indeed, all Sturm und Drang music, as a vain exaggeration. It is as if Haydn confesses here to perjury, as if he says ‘‘Do not believe me. My sighs were phoney, my tribulations trivial, and my heroic posture ludicrous. It was all a sham.’’ In his last quartet, Op. 135, Beethoven follows the general scheme of Haydn’s Op. 76, No.1. As in Haydn’s work, all four movements are in the major mode, but the finale begins in minor. None of Beethoven’s 240

1.

2.

Si repete la seconda parte al suo piacere 247

Poco adagio

Tempo I. pizz.

pizz.

pizz.

pizz.

253

Example 9. Beethoven: String Quartet in F Major, Op. 135, iv, mm. 240–58.

202

eddy zemach and tamara balter

earlier quartets has such a plan. The resemblance to that work of Haydn becomes even more pronounced in the ironic coda that comes after the movement has cadenced in the tonic. That coda lightly dismisses both the question (‘‘Muss es sein?’’ in minor, marked Grave) and the answer (‘‘Es muss sein!’’ in major, marked Allegro) in the finale (Example 9). It starts by presenting the answer motive, for the first time, in a mock-pathetic tone; due to the slow tempo, fermatas, and endings on diminished seventh chords, it now sounds like a question. The core of the coda, we suggest, gives the real answer: the na¨ıve melody, played pizzicato, which (as in Haydn’s Op. 76, No.1) sounds like a music-box tune, derides all that preceded it, including, most notably, the lofty assurance of the ‘‘Es muss sein!’’ answer. George Edwards regards such music-box endings in Haydn as ‘‘nonsense endings,’’ claiming that Haydn’s works ‘‘search for resolution or closure more often than they find it’’ (1991: 228). We suggest that the common scheme of the above examples divulge romantic irony. The item that the composer mocks is juxtaposed to the item that mocks it. That immediate (or almost immediate) succession, where the two items are thrust against each other, makes them look like counterparts. In romantic irony the projector and the projected are one; it is self-reflection that borders on self annihilation. Had the music been funny, ridiculous or gross from the start, it would have been a parody (denigrating something else) or a mere (non-referential) joke. If the buffoonish situation is relegated to another part in the work, not immediately adjacent to the lofty situation it mocks, we would still see them as counterparts, but not as identical, hence it would be parody rather than romantic irony. It is therefore the amalgamation that generates romantic irony in music. Romantic irony is found, then, where there is, first, a sharp contrast between the two items, and second, a motivic and physical proximity between them. These features make us hear the later musical event as referring to the former, and, through it, to the work as a whole. Reference to an earlier segment of a work is usually achieved through motivic connection.²³ When enhanced by physical proximity, we cannot but regard it as identity. We therefore must see the work as ferociously turning on itself in a grand act of self-immolation. ²³ For instance, the beginning of the addendum in Beethoven’s Op. 95, F-F#-G, hints at the previous function of these pitches, mainly the pathos that the F-Gb projects from the very start of the quartet.

the structure of irony/how it functions in music

203

References

Bonds, M. E. (1991) ‘‘Haydn, Laurence Sterne, and the Origins of Musical Irony,’’ Journal of the American Musicological Society 44/1: 57–91. Booth, W. C. (1974) A Rhetoric of Irony, Chicago, Ill.: University of Chicago Press. Brady E. and Levinson, J. (eds.) (2001) Aesthetic Concepts, Oxford: Clarendon Press. Brauner, C. S. (1981) ‘‘Irony in the Heine Lieder of Schubert and Schumann,’’ The Musical Quarterly 67/2: 261–81. Burnham, S. G. (1994) ‘‘Mozart’s Felix Culpa: Cos`ı Fan Tutte and the Irony of Beauty,’’ The Musical Quarterly 78/1: 77–98. Daverio, John (1990) ‘‘Reading Schumann by Way of Jean Paul and His Contemporaries,’’ College Music Symposium 30/2: 28–45. Davidson, D. (1978) ‘‘What Metaphors Mean,’’ Critical Inquiry 5: 31–47. Dent, E. J. (1960) Mozart’s Operas: A Critical Study, London: Oxford University Press. Dill, H. J. (1989) ‘‘Romantic Irony in the Works of Robert Schumann,’’ The Musical Quarterly 73/2: 172–95. Edwards, G. (1991) ‘‘The Nonsense of an Ending: Closure in Haydn’s String Quartets,’’ Musical Quarterly 75/3: 227–54. Ellestr¨om, L. (1996) ‘‘Some Notes on Irony in the Visual Arts and Music: The Examples of Magritte and Shostakovich,’’ Word & Image 12: 197–208. Glucksberg, S., and Keysar, B. (1993) ‘‘How Metaphors Work,’’ in A. Ortony (ed.), Metaphor and Thought (2nd edn), Cambridge: Cambridge University Press, 401–24. Hatten, R. S. (1994) Musical Meaning in Beethoven, Bloomington, Indiana: Indiana University Press. Herzog, P. (1995) ‘‘The Practical Wisdom of Beethoven’s Diabelli Variations,’’ The Musical Quarterly 79/1: 34–54. Imeson, S. (1996) ‘‘The time gives it proofe’’: Paradox in the Late Music of Beethoven, New York: Peter Lang Publishing. Kierkegaard, S. ([1841], 1989) The Concept of Irony, With a Continual Reference to Socrates, ed. and trans. H. V. Hong, and E. H. Hong, Princeton, NJ: Princeton University Press.

204

eddy zemach and tamara balter

Kinderman, W. (1987) Beethoven’s Diabelli Variations., Oxford: Clarendon Press. Kinderman, W. (1996) ‘‘Beethoven’s High Comic Style in Piano Sonatas of the 1790s, or, Beethoven, Uncle Toby, and the ‘Muckcart-Driver’,’’ Beethoven Forum 5: 119–38. Kivy, P. (1980) The Corded Shell: Reflection on Musical Expression, Princeton, NJ: Princeton University Press. Kramer, J. D. (1973) ‘‘Multiple and Non-linear Time in Beethoven’s Opus 135,’’ Perspectives of New Music 11/2: 122–45. Lakoff, G., and Johnson, M. (1980) Metaphors we live by, Chicago, Ill.: University of Chicago Press. Lochhead, J. (1979) ‘‘The Temporal in Beethoven’s Opus 135: When Are Ends Beginnings?,’’ In Theory Only 4/7: 3–30. Longyear, R. M. (1970) ‘‘Beethoven and Romantic Irony,’’ in P. H. Lang (ed.), The Creative World of Beethoven (essays originally published in The Musical Quarterly), New York: W. W. Norton & Co. Inc., 647–64. Mellers, Wilfrid (1983) Beethoven and the Voice of God, New York: Oxford University Press. Miller, C. L. (1976) ‘‘Ironic or Not?,’’ American Philosophical Quarterly 13/4: 309–13. Muecke, D. C. (1969) The Compass of Irony, London: Methuen. Parker, M. (2002) The String Quartet: Four Types of Musical Conversation, 1750–1797, Aldershot: Ashgate. Sawyer, J. E. (1999) ‘‘Irony and Borrowing in Handel’s Agrippina,’’ Music and Letters 80/4: 531–59. Schindler, A. F. ([1860], 1996) Beethoven As I Knew Him, ed. D. W. MacArdle, trans. C. S. Jolly, Mineola, New York: Dover. Scruton, R. (1997) The Aesthetics of Music, Oxford: Oxford University Press. Searle, J. (1978) ‘‘Metaphor,’’ in A. Ortony (ed.), Metaphor and Thought (2nd edn), Cambridge: Cambridge University Press, 83–111. Searle, J. (1979) Expression and Meaning, Cambridge: Cambridge University Press. Sheinberg, E. (2001) Irony, Satire, Parody and the Grotesque in the Music of Shostakovich: A Theory of Musical Incongruities, Aldershot: Ashgate. Sibley, F. N. ([1959], 1978) ‘‘Aesthetic Concepts,’’ in J. Margolis (ed.), Philosophy Looks at the Arts, Philadelphia, Penn.: Temple University Press, 64–78.

the structure of irony/how it functions in music

205

Stravinsky, I. and Craft, R. (1960) Memoirs and Commentaries, London: Faber and Faber. Suurp¨aa¨, L. (1996) ‘‘Schumann, Heine, and Romantic Irony: Music and Poems in the First Five Songs of Dichterliebe,’’ Int´egral 10: 93–123. Tovey, D. F. (1972) Essays in Musical Analysis: Chamber Music, London: Oxford University Press. Walton, K. (1991) Mimesis as Make-Believe, Cambridge, Mass.: Harvard University Press. Watkins, G. (1994) Pyramids at the Louvre: Music, Culture, and Collage from Stravinsky to the Postmodernists, Cambridge, Mass.: Harvard University Press. Wittgenstein, L. ([1922], 1961) Tractatus Logico Philosophicus, trans. D. F. Pears and B. F. McGuinness, Oxford: Blackwell. Wollheim, R. (1986) ‘‘Imagination and Pictorial Understanding,’’ Proceedings of the Aristotelian Society 60 (suppl. vol.): 45–60. Wollheim, R. (1987) Painting as an Art, Princeton, NJ: Princeton University Press. Wollheim, R. (1998) ‘‘On Pictorial Representation,’’ Journal of Aesthetic and Art Criticism 56: 217–26. Zemach, E. M. (1999a) ‘‘Look, This is Zeus!,’’ in K. Krausz, and R. Shusterman (eds.), Interpretation, Relativism and the Metaphysics of Culture, Amherst, NY: Humanity Books, 311–33. Zemach, E. M. (1999b) ‘‘Metaphors and Ways of Life,’’ in J. Hintikka (ed.), Aspects of Metaphor, Dordrecht: Kluwer, 243–54. Zemach, E. M. (2001) ‘‘A Modal Theory of Metaphor,’’ Theoria 67: 60–74. Zemach, E. M. (2002) ‘‘What Do We See On TV?,’’ in R. Lorand (ed.), Television: Aesthetic Reflections, NY, Washington: Peter Lang Publishers, 89–106.

Further Reading

Bribitzer-Stull, M. (2004) ‘‘Did You Hear Love’s Fond Farewell? Some Examples of Thematic Irony in Wagner’s Ring,’’ Journal of Musicological Research 23/2: 123–57. Diener, B. S. (1992) ‘‘Irony in Mozart’s Operas’’ (Ph.D. dissertation), New York: Columbia University.

206

eddy zemach and tamara balter

Finson, J. W. (1994) ‘‘The Intentional Tourist: Romantic Irony in the Eichendorff Liederkreis of Robert Schumann,’’ in R. L. Todd (ed), Schumann and His World, Princeton, NJ: Princeton University Press, 156–70. Hoeckner, B. (1997) ‘‘Schumann and Romantic Distance,’’ Journal of the American Musicological Society 50/1: 55–132. London, J. (1996) ‘‘Musical and Linguistic Speech Acts,’’ The Journal of Aesthetics and Art Criticism 54/1: 49–64. Lowe, M. (2002) ‘‘Falling from Grace: Irony and Expressive Enrichment in Haydn’s Symphonic Minuets,’’ The Journal of Musicology 19/1: 171–221. Nestrovski, A. (c.2000) ‘‘Beethoven’s Ironies,’’ in D. Greer, I. Rumbold, and J. King (eds.), Musicology and Sister Disciplines: Past, Present, Future, New York: Oxford University Press, 428–38. Spitzer, M. (1996) ‘‘The Retransition as Sign: Listener-Oriented Approaches to Tonal Closure in Haydn’s Sonata-Form Movement, Journal of the Royal Musical Association, 121/1: 11–45.

PA RT I V

New Issues

This page intentionally left blank

9 Music and Electro-sonic Art G O RD O N G RA H A M

In this paper I want to discuss two questions. First, is electro-acoustic music, music? And, second, how important is the answer we give to this question? Though the term ‘electro-acoustic music’ now has a wide currency, using it to explore this issue is clearly a recipe for confusion, and so I shall use the term ‘electro-sonic art’ instead. The questions thus become: is sonic art the same as music? And does it matter whether it is or not? Or, to put the second question in a slightly different way, is music the whole of sonic art? I shall begin with some reflections on the nature of music.

1 Music is for listening to as nothing else is. Such an assertion is neither startling nor novel and may even sound banal. In fact, I mean it to be a commonplace and not to be confused (at least at this stage) with any sophisticated contention in the philosophy of music. For example, Peter Kivy has identified a ‘prescriptive listening code embodied by the aesthetic attitude’, and argued persuasively that music which is specially meant to be listened to in accordance with this code—the music of the concert and the concert hall—is not music per se, but one phase of its history and development, albeit an enormously important one (Kivy 2001: 55). My proposition, by contrast, means only to assert that (pace John Cage perhaps, to whom further reference will be made at a later stage) an essential, indispensable feature of music is that it is heard (rather than seen, touched, smelt, tasted or otherwise apprehended).

210

gordon graham

The proposition that music is for listening to as nothing else is, should not be confused with the false proposition that music is only for listening to. On the contrary, music is one of the performing arts, and most composed music has been written first and foremost for performance. However, the dominance of ‘art music’ (that has an audience) over folk music (that generally does not) has given listening to music a certain prominence and this is what makes the indispensability of listening worth stressing. Because, in view of what people often say about music, that music has to be heard is one of those propositions illustrative of the fact that truisms are not always obvious. There is a recurrent tendency for both theoretical explanations of music in general and musicological interpretations of particular pieces to pass beyond the experience of listening, and to think of it as primarily a means to an experience of some other kind. There are two familiar candidates for what other kind of experience this might be—the emotion the music expresses is one, and the images or ideas represented in it another. Of these two candidates, emotion is the more dominant, despite the early onslaught of Hanslick (Vom Musicalisch-Sch¨onen, 1854) and the much more recent arguments mounted by Kivy (1989) and others. It is rare to find a programme note that does not try to identify both the emotional origins and the emotional content of a composition. The origins are usually sought in the composer’s biography (or autobiography), and the content of the music is analysed in terms of emotional states expressed within it and likely to be experienced by the audience. When formalism about music was pre-eminent, such commentary tended to be dismissed as mere ‘music appreciation’, but this appeal to emotion is not an idiosyncratic feature of those who write programme notes; almost everyone’s first attempt at explaining the nature and value of music is couched in terms of emotional expression, and it is an explanation with which sophisticated philosophies of music have often persisted. Sometimes, this kind of analysis is combined with an account of representation in sound—the use of music to paint pictures and tell stories as in (supposedly) programme music. Occasionally, the intellectual content of music is construed, even, as going beyond mere representation to the incorporation of moral, political, or religious ‘ideas’. Generally, though, expressivism remains the dominant view because even when some element of representation or philosophizing is thought to be present, it is usually in the feelings associated with the pictures, stories or ideas that the meaning

music and electro-sonic art

211

of music is ultimately said to lie. In some cases such intellectual content is almost coincidental—the peasants in Beethoven’s Pastoral symphony, for example; and in some it is highly intentional—the paintings depicted by Mussorgsky, for example, or the idea of resignation expressed in Beethoven’s late quartets. But the essential idea is the same: that music, like pigment, is a material medium through which something else is realized. This thought was expressly articulated by Wassily Kandinsky, who tells us that ‘all the arts are identical. The difference [between them] manifests itself by the means of each particular art ... Music expresses itself by sound, painting by colours etc.’ (quoted in Chipp 1968: 346–7). This idea that sound is the means of musical expression whose real content lies elsewhere seems to me an important mistake. Part of the purpose of this paper is to trace some of its implications, but to do so with an eye to the issues surrounding tonality and the innovations of atonality and electro-acoustic composition.

2 There is a powerful argument against any suggestion that musical composition and performance are means to some other form of experience, emotional or intellectual. This may be termed ‘the redundancy argument’ because it turns on the idea that construing music in either way makes listening to it redundant in principle. Suppose that when we listen to (say) the second movement of Elgar’s Cello Concerto, we are hearing ‘an old man’s melancholy’ (as Elgar himself declared); or that we are ourselves experiencing the emotion of sadness (as many commentators on this movement have asserted), and that it is in this experience that the meaning or value of Elgar’s composition lies. There are clearly other ways of experiencing the same thing—having Elgar himself tell us of his melancholy in words, for example, or feeling sad in the same way by recalling something that has happened to us in the course of life. If such possibilities are alternative means to the same end, it follows that, though Elgar’s music may be one route to this experience, it is entirely dispensable should these other means be available to us. The same point may be made about other supposed musical meanings. Wilfred Mellers has claimed that Beethoven’s late quartets ‘say’ the same

212

gordon graham

as T. S. Eliot’s Four Quartets (Mellers 1988: 657–8). If so, the music is not uniquely valuable. We can read Eliot’s poems instead of listening to Beethoven’s music. In this way (quite contrary to his intention) Mellers’s attribution of meaning makes the music redundant. So too with the strictly representational. If Mussorgsky’s music truly is a depiction of (for example) the Great Gate at Kiev, then we have a choice; listen to the music or look at the picture. Neither is essential, in which case we can ignore the music so long as we have the picture (and vice versa of course). ‘The redundancy argument’ is my name for a long established and familiar line of thought. Its first articulation is to be found (in a slightly different form) in Hanslick. But in my view its force has not always been properly recognized. This is perhaps the result of its being construed as an unnecessary demonstration of a self-evident proposition—that sound is indispensable to music. This is, rather, the incontestable proposition on which it is based. The argument itself is intended to show that any attribution to music of nonaural meaning or content—emotion or depiction for example—carries the implication that this self-evident proposition is false. The identification of musical significance with emotional experience or intellectual content leads to the absurd conclusion that in principle we can get everything music has to offer without actually listening to it. Anyone who wants to resist this implication has available at least two sorts of reply. The first is to deny that the proposition from which the implication is drawn is as truistic as I have alleged. The fundamental contention, it might be said, is not about sound, but about music—that the content of music is itself music and nothing but music. Far from being a truism, this contention is just one of a number of theories of musical content. However, it is only the second half of this contention that conflicts with other theories of content. While there can be competing theories about what the content of a composition can coherently be said to be, its being a piece of music is an indispensable part of its having the value it does. Roger Scruton, for instance, argues at length in favour of the view that ‘The expressive qualities of a work of music form the most important part of its content’ (Scruton 1997: 344). A competing theory (L. B. Meyer’s perhaps) would be that the structural properties of a work are the most important part of its content. But both theories rightly take it for granted that it is musical expression and musical structure that we are talking about, and the important point is that this musicality is essentially, not contingently

music and electro-sonic art 213 related to its value or importance. Whatever it is we get from it, we cannot get it in non-musical form. A second response to the redundancy argument (with respect to expressivism) would be that the relation between music and emotion is more sophisticated than I have allowed. Music, it might be claimed, is not merely a means to the expression and apprehension of emotion. But what else could the relation be? One familiar line of thought is that the redundancy argument does not rule out the expression of emotion as realized in music. What exactly does this mean, though? What it must not mean, obviously, is ‘realized in a way not susceptible to the redundancy argument’, since this would simply beg the question. What it could mean is that emotional terms seem to be directly applicable to music, and not simply to the states of mind of composers and audiences. A minor chord is naturally described as ‘sad’, irrespective of the feelings of those who play or hear it, and people spontaneously describe much more extended musical pieces in a similar way. However, all this shows, I think, is that the language of emotion, whose home is in the context of human feeling, can (perhaps surprisingly) be applied to music as well. It does not actually show that the full implications of the originating context are preserved in the extension of that language to music. And if there were reason to think that they are, we would still have to assert that emotion ‘realized’ in this particular way requires to be heard. This contrasts with those instances in which, for example, music is used merely as a signal. Famously, the opening bars of Beethoven’s Fifth meant something special as D-Day drew near, and were intended to do so. A non-aural signal could have performed the same role. But if there is some emotion denoted that is intimately connected with those opening bars, the realization of it in this form is uniquely valuable. In short, music is for listening to as nothing else is—the proposition with which I began.

3 What could give music this indispensably aural character? The answer I want to explore is that music precisely consists in intentionally organized sound. This is a familiar contention, but it needs both explication and refinement. Jerrold Levinson writes as follows:

214

gordon graham

It has often been suggested that music be defined as organized sound. But this is patently inadequate. While sonic organization has some plausibility as a necessary feature of music, it is hardly a sufficient one. The output of a jackhammer, the ticking of a metronome, the shouts of a drill sergeant during a march ... are all organized sounds, but not instances of music. (Levinson 1990: 269–70)

Levinson himself is willing to accept the ‘necessary but not sufficient’ interpretation. For present purposes, his intuitive rejection of alternative types of sonic structure as music is a matter yet to be discussed. However, the important point is this: even if we do interpret intentional sonic organization as only a necessary condition of something’s being music, it can still be the case that its sonic character is an essential part of its value. In the same essay Levinson invites us to ask why music is important to us. His answer is that ‘actively engaging with organized sounds, intentionally produced for such purpose, can enrich experience’ and he identifies the following forms of enrichment: conveying emotional content, stimulating the ordering faculties of the mind, providing insights into the mysteries of the psyche, making one pleasurably aware of one’s body, securing permanent alterations in outlook and attitude (ibid.: 278). Now all of these ends can be obtained in other ways and cannot therefore secure the unique value of music. What the intentional organization of sound alone secures (and what is not expressly mentioned here) is the enrichment of strictly aural experience. It is for this reason that listening is essential, and for this reason that the sonic properties of music are indispensable. It is this fundamental thought that the redundancy argument rests upon. Let us agree that music may prompt in us emotional experiences or intellectual insights. Yet if and when it fails to do so, it can still be valued as an extension and exploration of aural experience itself. Moreover, its essentially aural nature allows us to explain important differences between the creations and explorations of different composers. National anthems can (and often do) stir people deeply, and the music of ‘avant garde’ composers may be full of ‘ideas’. But the ultimate mark of a superior composition is that it is more worth listening to. What makes it ‘more’ worth listening to is not a simple matter of quantity of notes or instruments, but the creation of sounds that we naturally have to describe in metaphorical or analogically extended language. This is a point worth stressing. The fact that pure or

music and electro-sonic art

215

absolute music can be described as bright or architectonic or mournful is specially interesting not because the music is in some way connected with lights, or buildings or even, as is so commonly supposed, with emotions (even if it is so connected), but because it shows that sound—pure sound—can have properties far more sophisticated than those of volume (loud or soft), pitch (high or low), and tempo (fast or slow). A further, similar and crucial point is to be made about performance whose role is to reveal the sonic properties of a musical composition in a way that no musicological analysis of the ‘meaning’ of the piece or study of the emotional life of the composer could do. A gifted performer enables us to discover the existence of interesting and unusual sonic properties that left to ourselves we might have missed. That is how a particular performance can be specially worth listening to. Composers imagine and performers realize sonorities that it may well be natural to describe as exhilarating, melancholic, agitated and so on, but in the experience of listening it is nonetheless the musical sound itself that is the principal object of the attentive mind. By contrast, in speaking, noise-making can be a mere means of expression or communication. What matters is the meaning of what is said, not the aural quality of its saying. It is true that focus on phonetic properties may be too narrow. Sameness of ‘what is said’ does not make two expressions interchangeable. For example, ‘rabbits are animals’ and ‘bunnies are animals’ could be said to say exactly the same thing, but the latter would be inappropriate in a zoology textbook (an example I owe to Stefano Predelli). Plainly, there are cases in which what is communicated depends not only on what is said, but on how it is said. Moreover, there is poetry, in which aural properties do matter, and there are some natural languages in which pitch makes a difference to meaning. Still, for present purposes these complications can be set aside. The irrelevance of accent, rhythm and pitch in ordinary English is enough to make the point. While the high-pitched utterance of a sentence spoken quickly in a Scottish accent says exactly what the same sentence slowly uttered in a low-pitched Welsh accent says, it is precisely such variables that matter in music. Some of Steve Reich’s compositions are interesting in this connection. Reich has used his celebrated technique of ‘phasing’ to separate the sound of speech from its meaning, and thus turn it into a form of music. He contends that, so transfigured, language offers us a ‘window on the human soul’. But a more natural thought would be that untransfigured language is a

216

gordon graham

window on the soul (when we let it be), and that the interest in transfigured speech lies in the special aural rather than semantic properties it has. Listening to music is a peculiarly human phenomenon. Animals other than humans can listen as well as hear, of course, but while we have every reason to think that their being equipped with auditory powers is enough to let them hear music, its sophisticated structural properties mean that only human beings can listen to it. This is why reflection and analysis can help us to hear the properties of a piece of music. Indeed where the music is of a highly developed sort we may need the vocabulary of technical analysis to isolate and describe the structures of sound to be found within it. Nevertheless, if we are to know that these sounds have that property, listening is inescapable. And so it should be of course. The problem with the familiar alternative explanations of music’s value is (as I argued) that they make the music redundant. Emotional ‘highs’ are to be had elsewhere; what the music ‘says’ can be said more intelligibly in words; what it is said to depict can be painted; any values that it exemplifies, such as beauty or gracefulness, are exemplified in other arts. But once we locate the value of music in its unique ability to extend and explore aural experience, we have attributed to it a distinctive value. Since we cannot have that experience non-aurally, the activity of actually listening to music is ineliminable. However, in the business of listening both composer and performer have a crucial role. Just as the painter directs our visual perceptions, so the composer and performer direct our aural perceptions. Listening to music is not just a matter of sound pouring into a receptor, but of the mind being directed through a series of perceptions. We are, so to speak, steered through our experience. It is as though the composer were saying ‘It must be heard this way’ by actually making us hear it that way. An analogy might be this: we enter a series of underground caverns where our journey can take alternative routes through spaces of differing shape, dimension, and atmosphere, lighted by different means. Each composer is the guide who decides upon the lighting and the order in which we pass from cave to cave; the performer is the guide who leads us through the caverns. This analogy touches upon an important debate in the ontology of music: is composition a matter of creation or discovery? The parallel with the underground caverns would seem to favour the sort of Platonism in music that Peter Kivy has defended (Kivy 1993: Ch. 2). The shapes and dimensions of the caves are there for all to see, but they can be seen only

music and electro-sonic art 217 this way or that. The way we see them and the ways that are especially worth seeing, are matters for which we rely upon our guides. Applying this to music it seems that the melodies and harmonies are ‘there for all to hear’, and only await the composer and the performer to point them out. Now, while I am sympathetic to Platonism in music (and to Kivy’s defence of it), it does not seem to me that the centrality of listening to music requires us to think in this way. The parallel with the caverns is only intended heuristically, so to speak. It is consonant with thinking of composition (and improvisation) as pure invention, the creation ex nihilo of musical structures through which we are led. Either way, we can correctly be said to enjoy enriched aural experience. But whether we think of musical composition as invention or discovery, there is a further question. Is it only music that can enrich aural experience? Is music truly unique in this regard? Hitherto the answer has perhaps seemed obvious, but recent technological innovations can be construed as raising a doubt on this point. There are now other possibilities, and that is why we have to deploy the term ‘sonic art’. By ‘sonic art’ I mean any auditory construction whose purpose is to explore and enrich the world of aural experience. Does this necessarily mean music?

4 It might be thought that the answer to this question depends upon our definition of music. Clearly, if music is defined as any sonic construction, then the answer is an uninteresting ‘yes’. But to include any piece of intentionally organized sound under the category of music begs certain questions. Levinson (as noted earlier) thinks that such a definition has counter-intuitive implications—‘The output of a jackhammer, the ticking of a metronome, the shouts of a drill sergeant during a march ... are all organized sounds, but not instances of music.’ It is not clear, however, that this differentiation is supported by his own definition of music, namely: sounds temporally organized by a person for the purpose of enriching or intensifying experience through active engagement (e.g.) listening, dancing, performing with the sounds regarded primarily, or in significant measure, as sounds. (1990: 273)

218

gordon graham

In recognition of this fact, Levinson amplifies his definition with a distinction between what is music and what can be treated or regarded as music: One way to ignore this distinction is to claim, in the spirit of John Cage’s Zeninspired reflections, that any and all sounds are music. This is simply false ... . What Cage shows perhaps is that any sounds can be listened to as if they were music ... It does not follow that all sounds are at present music. The whirr of my blender and the whistling of the wind are not instances of music. (1990: 275)

The qualification ‘at present’ may be of some significance. Levinson’s counter instances are plausible because most people perceive an evident difference between music properly so called and ‘the whirr of my blender and the whistling of the wind’, even if the difference is one they find hard to articulate. On the other hand, it is precisely these other types of sound that composition using digital technology is keenest to deploy and exploit. As digital compositions become more widely heard (as film scores, for example), it may be that the sounds classified as ‘music’ extend beyond those so regarded ‘at present’. The main difference between music traditionally understood, and the sounds that characterize more recent compositions (intuition tells us) is tonality, a difference equally detectable by the tutored and untutored ear. Most people incline to Hanslick’s view that ‘[t]he material out of which the composer creates, of which the abundance can never be exaggerated, is the entire system of tones, with their latent possibilities for melodic, harmonic and rhythmic variety’ (Hanslick 1854: 28). Accordingly, they reserve the label ‘music’ for tonal sounds. Furthermore, most people (I imagine) share Roger Scruton’s opinion that ‘[t]onality provides a paradigm of musical organization ... [a]nd attempts to depart from tonality, or to discard it entirely, seem only to confirm its authority over the musical ear’ (Scruton 1997: 239). That is why the expression ‘electro-acoustic music’ is confusing, and the term ‘electro-sonic art’ to be preferred. Two further issues arise, however. First, given the history of composition in twentieth-century music, it is highly contentious to make tonality the paradigm of music. This is precisely what the advocates of atonality, Schoenberg chief among them, expressly denied. Second, even if the perceptible difference between music and electro-sonic art is sufficient warrant for using different terms to refer to them, this tells us

music and electro-sonic art

219

nothing about their respective aesthetic value. I shall discuss these issues in turn. Scruton, as is well known, defends tonality against atonality and he argues that if those composers who were (are) persuaded that tonality is exhausted do succeed in extending composition and performance in new directions: it is either because tonality can be extended or because its effects can be preserved through the kind of oblique tonal thinking shown by Skriabin, Stravinsky, Britten and Berg. The possibility remains that tonal music is the only music that will ever really mean anything to us, and that, if atonal music sometimes gains a hearing, it is because we can elicit within it a latent tonal order. (Scruton 1997: 308)

A preference for tonality is widely regarded as indefensibly conservative and is usually countered by the observation that popular responses to atonal composers simply repeat the experience of the greatest of tonal composers. If near riots greeted the first performance of Stravinsky’s Rite of Spring, this is no more adverse a reaction than Beethoven’s late Quartets prompted. Yet while episodes from the history of music undoubtedly provide a salutary counter to unthinking conservatism, there does seem to be some basis for regarding the twentieth century as innovatory to a degree unmatched by earlier centuries. As Gerald Abraham has remarked: [e]ven such a drastic innovation as the Florentine recitativo was gradually assimilated to the melodic tradition in ariosos and finally in Wagner’s scores. Monteverdi could employ his prima prattica and seconda prattica side by side; the Western tradition was strong enough to absorb both. [But then] the first signs of a more serious crisis showed early in the twentieth century. A sense that the great tradition was approaching a dead end was reflected in Debussy’s non-functional harmony, in Skryabin’s experiments with a completely new harmonic system, and in Schoenberg’s revolt against tonality. (Abraham 1979: 820)

However, it is not with Schoenberg or Skryabin that the most radical innovations are to be found. If we include in the definition of ‘tonal’ music (as Scruton does) the existence of a tonic, a dominant tone, then atonal music is indeed atonal. But odd though this may sound, atonal music is continuous with tonal. That is why it is less misleadingly called ‘serialism’.

220

gordon graham

The difference is the method of composition. Scales and chords are abandoned in favour of a self-imposed system of composition and organization. But even if the end result sounds repellent to the conventionally educated ear, it is still tones that are being serialized. A more complete break with tonality is to be found in composers such as Var`ese and Lutoslawski. Var`ese declared himself to be engaged in a ‘fight for the liberation of sound and for my right to make music with any sound and all sounds’ (www.zakros.com/mica/soundart/s04/varese− text.html). Accordingly, his compositions Deserts and Poem electronique consist in amalgamated sounds of the sort that have become much more familiar with the advent of digital technology. Not only is there nothing we could recognize as melody or harmony, and no key or mode, if the individual sounds out of which it is composed are isolated, they are not found to be tones either, but more like Levinson’s jackhammer or blender. Var`ese seeks ‘a whole new world of unsuspected sounds’. Similarly, in the second movement of Lutoslawski’s Trois Po`emes d’Henri Michaux singing is abandoned for speaking and shouting. Lutoslawski has devised a notation which instructs the ‘choir’ about pitch and volume, a notation that looks a little like ‘music’, and another special notation for the percussion that accompanies it. The effect is highly dramatic and aurally impressive, but as the need for special notation indicates, the composition signals a move away from music, precisely because it seeks to go beyond the confines of resonant tones. To repeat, serialism devises deliberate rules of composition that bear no relation to established harmonic structures. This makes its outcomes strange to the ear raised on Bach and Beethoven, which is why it is readily described as atonal. But it still takes the twelve tones of the chromatic scale as the basic sonic material. It is with compositions such as Deserts and Trois Po`emes, not atonalism, that we move to non-musical sonic art, a move that finds its full realization with the advent of recording and digital technology. Var`ese is widely regarded as a founding figure in electro-sonic art, but with hindsight the devices at his disposal seem exceptionally primitive. Still, the gap between his compositions and those of even the most innovative serialists seems to me radical enough to warrant replacing the term ‘music’ with that of ‘sonic art’. It has been widened almost immeasurably by the deployment of new technology. The most telling difference between (for instance) serial composition and Steve Reich’s ‘phasing’ is that the

music and electro-sonic art

221

latter is not the application of rules of composition, but a technological process for combining recorded sounds. Digital composition widens the gap still further; in preference to using actual recordings it manufactures and manipulates its own sounds. At the end of this development, it seems to me, we are left with things so different that it makes for confusion to bracket music and electro-sonic art together. But if the two are not the same, how do they compare? Var`ese was a classical musician by training who saw his work standing in the long tradition and denied that he had any desire ‘to disparage and even to discard the great music of the past’. ‘No matter how original, how different a composer may seem,’ he says, ‘he has only grafted a little bit of himself on the old plant’ (www.yale.edu/oham/CD1transcript.html). But notwithstanding the sincerity of this assertion, the continuity is hard to hear, the graft and the plant hard to identify. In my view, however, this does not matter, or to be more precise, it does not matter from an evaluative standpoint. The nature and the status of music and electro-sonic art are separate questions. We can and should conclude, I think, that the aspirations of Var`ese, now that they have finally been realized with digital technology, have indeed resulted in something wholly new. A more difficult question is whether this wholly new thing should be accorded the kind of value that hitherto music properly so called has uniquely enjoyed.

5 The first point to be made of course is that electro-sonic art fits Levinson’s definition of music—sounds regarded primarily as sounds and temporally organized for the purpose of enriching or intensifying experience through active engagement (which I have modified to read ‘aural experience’). The evaluative question, accordingly, is twofold: does electro-sonic art offer as much as music does by way of exploring and enriching our aural experience? And does it offer the same or similar opportunities for active engagement? One way of answering the first of these questions is to consider how much of the language that we use to analyse and describe music can also be used in the analysis of electro-sonic art. As far as analysis is concerned, it seems to me clear that the two share a good many of the same structural

222

gordon graham

properties; theme, variation and recapitulation are obvious cases in point. Counterparts to a tune and its accompaniment are harder to discern, but insofar as some such relationship can be characterized in terms of relative aural importance, it seems plausible that electro-sonic art can contain similarly related aural sequences. On the matter of description, the position is perhaps less clear. Here we return to the matter of expressiveness. I have argued against the familiar idea that music is a medium of emotional experience. This seems to me a post-Romantic fallacy. But as I noted earlier, the mistake lies in a fallacious inference from an undoubted fact. The undoubted fact is that musical sound can be described in affective language. To repeat, a minor chord is rightly and naturally described as ‘sad’, but we cannot infer from this alone that composing, playing or hearing a minor chord is in any wider way an experience of sadness. It is the same mistake, though far less obvious, as one that would be made by anyone who tried to compare the noise levels of loud and soft colours. The language is right; the reasoning is wrong. But it is precisely because the language is right that we are able to marvel at the richness of music, whose strictly aural properties can be described in such diverse ways. Can the sound constructions of electro-sonic art sustain a similarly rich descriptive vocabulary? I do not know what the answer to this question is, partly because electro-sonic art is in relative infancy. Electro-sonic constructions are rightly described as captivating, intriguing, exciting, even haunting. Whether they are, or can also be moving, tranquil, transcendent, humorous, and so on, are matters waiting to be resolved. It is upon their resolution that the question of electro-sonic art’s significance will ultimately turn. Enthusiasts will say ‘of course they can’, but more compelling evidence will come from the ways in which, eventually, a much wider cross-section regularly makes use of electro-sonic compositions. In addition to the applicability of structural analysis and expressive description to electro-sonic art, there is a further substantial question about active engagement. Music is a communal as well as a natural product. Asked to name the really ‘big’ figures in the world of music, we immediately come up with composers—Bach, Beethoven, Mozart. Asked to identify outstanding ‘musicians’, we naturally think of performers—Pavorotti or Jacqueline du Pr´e, say. Yet even the greatest names will have had teachers whose names are virtually unknown, and the meaning and value of the composer’s and performer’s work requires realization in the ears of the

music and electro-sonic art

223

listener. Music exists to be heard, and neither its composition nor its performance can simply spring into existence by virtue of the ingenuity of individuals. Even the categories of composer, performer, teacher and listener do not exhaust all the forms of participation. There are also those, known and unknown, who have invented, made and perfected the huge variety of instruments whose distinctive sounds, both alone and in combination, produce (so to speak) the matter out of which this world is made. Is the same range of participation available in electro-sonic art? There is some reason to think that the answer might be ‘no’. I shall consider only the case of performance. Before the advent of digital technology modern composers employed a number of devices to extend the range and character of sounds that appear in their compositions. For instance, the ‘prepared’ piano has nuts and bolts and other items inserted into the instrument, and is played as much by striking and plucking inside the piano itself as by depressing the keys, which are in any case thumped by arms and elbows as well as being played by fingers in the normal way. Some composers have included noise producers that are not instruments, for example, vacuum cleaners, bat and ball, and gun shots. Others have included passages of screaming and shouting as well as singing. These innovations are sometimes derided. Whether they should be or not is of no special concern here. Rather, the important point to note in the present context is that such innovations are child’s play in comparison to what can now be done by means of digital technology. But unlike its forerunners, the composition of this kind of sound is inextricably connected with the technology which makes it possible. Composers like Lutoslawski had to devise systems of notation that would instruct singers in the production of sounds; composition which deploys digital technology has no need to do this. More tellingly, there is no place for it, because the composition and its realization are one and the same. In other words, there is no gap, either temporal or conceptual, between composition and performance, the sort of gap that there is between a musical score and its realization. What this implies, however, is that electro-sonic art has no counterpart to what has been an essential feature of music, namely performance, and by eliminating performance, electro-sonic art eliminates interpretation. Why does this matter? It matters not just because for most people active engagement beyond listening means performance and not composition, but

224

gordon graham

also because interpretation in performance is a significant further avenue for the exploration of sound. Conventional music, in other words, provides for a collaborative exploration of sound by composer and performer (and arranger, sometimes). In digital composition, by contrast, the composer does everything. This can be true of musical sounds as well as non-musical ones. The music of Vangelis, for instance, cannot be performed; it can only be listened to. But this is a non-standard case. By contrast, directly recorded composition, like Reich’s, is characteristic of non-musical sonic art. What this means is that electro-sonic art is more limited than music in the opportunities it offers for active engagement. This contention is likely to be disputed by electro-sonic composers, who will point out that increasingly electro-sonic art is not simply listened to over headphones, but ‘performed’ in the sense that the digital product is played to an audience through different arrangements of speakers in different acoustics. It is true that for the most part it is composers themselves who perform in this way, but that might change. As it reaches new maturity, people may emerge who are distinguished performers of electro-sonic art but not themselves composers. Perhaps so, though the nature of the medium lends some credence to the view that performance and composition can never be separated in the way that they are in conventional music. Many great composers were fine performers, but by no means all, and there is nothing to prevent someone composing a violin sonata who cannot play the violin. By contrast, a digital composer must have mastered the very same technology that needs to be deployed in performing it, at least in the sense described. Digital technology is a source of enrichment insofar as it allows composers to overcome the aural limits of conventional musical sounds, as it unquestionably does. Nevertheless it may bring with it a still greater limitation yet. Since its products are to an important extent fixed and final, they give relatively little scope for something we value—participation in active music making. If so, then, whether we call it music or not, a world in which digital composition gradually came to dominate, though technically much more sophisticated, would in fact be a poorer one. But just for that reason, those who subscribe to the view that the value of music lies in the world of sound that it enables us to explore can persist with a high estimation of its value even in the world of digitally generated and manipulable sounds. And for just the same reason, it is unlikely that people will ever cease to value conventional composition.

music and electro-sonic art

225

References

Abraham, G. (1979) The Concise Oxford History of Music, Oxford and New York: Oxford University Press. Chipp, Herschel B (ed.) (1968) Theories of Modern Art, Berkeley, Los Angeles and London: University of California Press. Hanslick, E. (1854) Vom Musicalisch-Sch¨onen, trans. G. Payzant (1986), On the Musically Beautiful, Indianapolis: Hackett. Kivy, P. (1989) Sound Sentiment, Philadelphia, Penn.: Temple University Press. Kivy, P. (1993) The Fine Art of Repetition: Essays in the Philosophy of Music. Cambridge: Cambridge University Press. Kivy, P. (2001) New Essays on Musical Understanding, Oxford and New York: Clarendon Press. Levinson, J. (1990) ‘The Concept of Music’, in his Music, Art and Metaphysics, Ithaca, NY, and London: Cornell University Press. Mellers, W. (1988) Man and His Music (revised edn), London: Barrie and Jenkins. Scruton, R. (1997) The Aesthetics of Music, Oxford and New York:Clarendon Press. Further Reading

Kivy, P. (2002) Introduction to a Philosophy of Music, Oxford: Oxford University Press. Madell, G. (2002) Philosophy, Music and Emotion, Edinburgh: Edinburgh University Press. Sharpe, R. A. (2000) Music and Humanism, Oxford: Oxford University Press.

10 Thoughts on Rhythm RO G E R S C RUT O N

1 Perceptual information must be assembled into comprehensible units if it is to guide us around the world. This is as true of the ear as of the eye; hence Gestalt principles operate in the auditory as in the visual sphere, though applied to temporal rather than spatial configurations (Bregman 1990; Gelfand 1998). Some philosophers might wish to speak of a ‘nonconceptual’ or ‘preconceptual’ unity in the auditory Gestalt —even of a nonconceptual content (Peacocke 1983: Ch.3). For the unities and wholes that we hear are, in the first instance, presented under no description. When, for example, a regular sequence of scale steps is interrupted by noises that mask individual pitches, we hear a continuous sequence, interrupted by foreground noise, rather than discontinuous sections of a scale. This perception does not require us to conceptualize the sequence as a scale, as musical movement, or indeed as anything else. It requires only that we group together the sounds that precede the noise with those that follow it. It is clear, therefore, that we could group sounds into coherent temporal figures without hearing them as music. Hence there is a real question as to what must be added to transform an experience of sound into an experience of music. To say that we must hear the sounds as music is either vacuous or false (false if it implies that we apply that very concept, a concept which many a singing, listening, dancing infant has yet to acquire). I argue elsewhere that sounds become music when organized rhythmically, melodically or harmonically—with the implication that each form of organization is sufficient to provide an experience of music. But I also suggest that these forms of organization pertain to the intentional rather

thoughts on rhythm 227 than the material object of perception. Melody is something that we hear in a sequence of sounds, and is not something that would be mentioned in a description, however complete, of the sounds themselves, judged as items in the material world (Scruton 1997).

2 Our ways of grouping and streaming individual sound events—both in everyday perception and in musical attention—reflect a peculiar metaphysical feature of sounds, which is that they are ‘pure events’. Unlike a car crash, a sound is not a change in something else. It is a self-subsisting object. When a sound occurs we assume that there is some physical cause. But we do not need to identify that cause in order to identify the sound, and the sound is fully intelligible as sound without reference either to the cause or to any other spatio-temporal particular. (Sounds should be compared in this respect with tastes and smells.) One consequence of this is that auditory streams can be organized internally, by reference to audible features of the sound events, and without invoking any order in the objects that produce them. An illustration is provided by the octave illusion studied by Diana Deutsch (Deutsch 1982). In this experiment headphones are placed over the subject’s ears, and the notes of a descending scale and an ascending scale played in each ear, in the patterns shown on lines 1 and 2 (Example 10). The two ears therefore receive, respectively, the inputs shown on lines 3 and 4. What they hear, however, are the two continuous sequences of lines 5 and 6. It is as though the sounds gravitate towards neighbours, where ‘neighbourhood’ is defined not by the physical proximity of the causative events, but by adjacent places on the pitch spectrum. Yet the sequences as heard are played into neither ear, and represent no causally unified process in the physical world. The auditory Gestalt is not merely incongruous with the physical events that produce it. It is organized according to principles that are intrinsic to the world of sounds, and which would be operative even if there were no physical events that could be identified as the causes of the individual sounds. The streaming that occurs in Deutsch’s experiment is not yet musical streaming. This kind of grouping by pitch proximity and sequential

228

roger scruton

Example 10. Diana Deutsch: pattern hearing.

‘flow’ could occur even in the experience of the wholly unmusical person—perhaps even in the experience of an animal. Musical experience, I suggest, involves the importation of a spatial framework, and the organization of the auditory field in terms of position, movement and distance. Those spatial concepts do not literally apply to the sounds that we hear. Rather they describe what we hear in sequential sounds, when we hear them as music. In other words the concepts that provide the fundamental framework for musical perception are applied metaphorically, and I have elsewhere tried to give an account of what this means and why it is important (Scruton 1997). The relation between auditory ‘streaming’ and the perception of musical movement can be likened to the relation between the perception of shapes and the perception of their figurative content. In the second case you see something in the shape that in the first case you merely see as a shape. In like manner the musical experience arises when you hear movement in a stream that you might otherwise hear merely as a

thoughts on rhythm 229 stream.¹ (The word ‘stream’ here is not a metaphor, but a primitive term denoting a continuous temporal Gestalt.) The musical order emerges when we adopt the ‘acousmatic’ attitude to the world of sounds,² attending to sounds without focusing on their material causes. There is a virtual causality that governs musical movement, as when one note in a melody is heard to bring its successor into being, even though sounded on another instrument in another place, and this virtual causality organizes the acousmatic Gestalt. The acousmatic experience is associated with a certain kind of listening, in the concert hall or at home. But it occurs also when musicians play together and become united in the first-person plural of the band, the music flowing through them as though with a force of its own (Schutz 1964: 159–78). It occurs too when we dance to music, and match the movement of our body to the movement that we hear. Dancing is both a response to musical movement and a way of understanding it as movement. And it is partly through its connection with the movement of the body and its social meaning that music acquires its moral character, a point made in other terms by Plato.

3 My subject in what follows is rhythm, and the distinction between rhythms imposed by metre, and rhythms generated by musical movement. By metre I mean the measuring and parcelling out of the temporal sequence. Not all music has a metre, and not every metre is like the metres familiar in Western music, which govern the divisions and subdivisions that correspond to time signature and bar-line. There are musical traditions that derive metrical patterns by adding note-values and not, as we do, by dividing larger units symmetrically. In classical Arab music, for example, rhythmic cycles are composed of time units added together to make often asymmetrical patterns, which do not permit whole-number division (Wright 1980: Sections 1–4). ¹ On the distinction between ‘seeing in’ and ‘seeing as’, see Richard Wollheim’s account of ‘representational seeing’, in Wollheim (1987). For reasons given in Ch. 5 of Scruton (1997), I do not regard ‘hearing in’ as sufficient foundation for musical representation, and in general prefer to discuss these phenomena as varieties of ‘double intentionality’, rather than ‘representational perception’. ² The term ‘acousmatic’, introduced in this sense by Pierre Schaeffer in Schaeffer (1966), is explained in Scruton (1997, Ch. 1).

230

roger scruton

The metres (tˆalas) classified by Sharngadeva, the thirteenth-century Indian theorist, and co-opted in our time by Messiaen, are likewise formed through addition rather than division, so that the rhythmic unit is assembled from note-values, rather than deduced by whole-number division within a bar.³ Yet more complicated are the patterns of African drum music, in which multiple metres are sustained indefinitely, producing patterns which cannot be assigned a single measure. (See the examples notated by Jones (1959), who convincingly argues that drum music is composed of repeated metrical units, with bar lines that fail to coincide.) Metre is—as its name implies—a form of measurement, in which timevalues are ordered into repeatable segments. Christopher Longuet-Higgins (1987: 150–68) has argued that we understand metrical organization by a kind of ‘generative grammar’, through which smaller units are derived in a rule-governed way from larger. Thus I hear the first note of Parsifal (Example 11) as the second of four crotchet values, and therefore locate it on a weak beat, after the strong but silent beat that begins the bar. The off-beat experience is intensified by the tie, which pushes the second note on to the second quaver value of a further subdivision within the bar. I do not carry out this calculation consciously. Nevertheless by unconsciously latching on to the generative hierarchy I am able to assign a measure, a beat and a temporal value to the notes that I hear, and thereby strain after the music as it flees my attempt to fix it to a downbeat, catching up with it, so to speak, only at the beginning of the third bar. Whatever the plausibility of that account, we should recognize that it applies only to the kind of metrical organization with which we are familiar from our own tradition, founded on the division and subdivision of the bar.

Example 11. Wagner: Prelude to Parsifal. ³ See Grosset (1913), and the discussion of modern Indian tˆala in Fox Strangways (1914: 191–224).

thoughts on rhythm 231 The oriental traditions referred to above introduce metres that are generated by addition, not division, and which therefore impose quite another order on the musical line—one in which syncopation cannot easily take root. Moreover, even if we confine ourselves to works in our own tradition, we must acknowledge that metrical order and rhythm are distinct. In grasping the rhythm of the opening phrase of Parsifal I am distinguishing accented from unaccented beats; I am responding to differences in stress, hearing certain notes as part of an ‘upbeat’, others as initiated by a downbeat, and grouping the tones into separate rhythmical units, as in Example 12.⁴ These other aspects of rhythm belong not to number but to life: they are features of the virtual energy that flows through the music, and which causes me to move with it in sympathy. The ancient musicologist and pupil of Aristotle, Aristoxenus of Tarentum, argued, in this connection, that rhythm is a temporal order imposed upon, though not inherent in, the sequence in which it is heard.⁵ The sequence is ‘rhythmized’ by the perceiver, by dividing it into up-beat and down-beat, and assigning a variable duration to each (the first always shorter than the second). It is clear from Aristoxenus’s account that he has dancing primarily in mind, up-beat and down-beat being explicitly connected to the arsis and thesis—lifting and falling—of the foot; nevertheless, his discussion reminds us that rhythm is a phenomenal, not a mathematical, property of a sequence, and that our capacity to perceive it is dependent upon our wider ability to respond to movement. Often the metre is an ad hoc attempt to place bar-lines across an organism that has no such divisions, as in certain parts of the Danse sacrale from the Rite of Spring (which, incidentally, is measured out somewhat differently in

Example 12. Wagner: Prelude to Parsifal. ⁴ The term ‘up-beat’ is contentious, and can mean one of at least two things: the entire preparatory phrase that precedes some musical emphasis, or the short intake of breath that leads directly into it. The first use began with J. J. de Momigny and was elaborated into a theory by H. Riemann, in Riemann (1903), to be further adapted by Edward T. Cone in Cone (1968). ⁵ Aristoxenus, Elementa Rhythmica, surviving fragments of Book 2 translated in Barker (1989), vol. 2.

232

roger scruton

Example 13. Bart´ok: Music for Strings, Percussion and Celesta, 1st movement.

Example 14. Boulez: Le marteau sans maˆıtre.

the four-hand piano score and the orchestral score). Consider the opening theme of Bart´ok’s Music for Strings, Percussion and Celesta (Example 13). Here the bar-lines and time-signatures seem like a tentative analysis of stress patterns, rather than an authoritative division into beats. This impression is even more evident in the measures from Le marteau sans maˆıtre (Example 14), where time-values are established by addition and not division within the musical line. (A comparable example is given in Example 15, from Le maˆıtre sans marteau by Nabil Az-Z´eloub, a poignant lament over the death of Jacques Derrida.) In such works the time-signature is like a piece of

thoughts on rhythm 233

Example 15. Az-Z´eloub: Le maˆıtre sans marteau.

speculative commentary placed above the score, and identifies no feature of the musical process. An analysis of Boulez’s rhythms, for example, would show them to be derived by the superposition of two or more metres, each formed by addition of time-values, rather than division of the bar. There is a real question as to whether we hear the result as a rhythm at all: paradoxically, the attempt to escape from metrical division has led, in Le marteau, to the cancellation of rhythm by metre, though a metre hidden behind the time-signatures that straddle the bars. Here the mathematical order seems to arrest the forward movement, rather than to guide it. (Likewise with Az-Z´eloub, who uses in this piece the himar metre of his native Algeria, transformed by successive augmentations which are then rearranged in a permutational sequence. The result is again entirely a-rhythmical.)

234

roger scruton

In the classical tradition metrical divisions coincide with significant junctures in the rhythm. However, as I suggest below, this coincidence can occur in two quite distinct ways. In one kind of case the metre is laid across the movement like a grid. In another kind of case it emerges from the movement, as though ‘precipitated out’, in the way that the hexagonal cells of the honeycomb are precipitated out of the honey-making within them.

4 The musical phenomena that we group together under the rubric of rhythm have their counterparts in other areas of human activity. Stress and measure occur in dancing and also in the movements required by certain kinds of physical work, such as those remembered in sea-chanties, in the occupation songs of India, and in the Negro spirituals (B¨ucher 1909). Stress, accent, metre and grouping all occur in speech, and speech-rhythms are both patterns and constraints when set to music. The rhythms of hymn tunes are classified according to their syllabic rather than their metrical structure since, from the liturgical point of view, it is more important to know what texts can be sung to them than to assign a time-signature to their musical line. And folk rhythms reflect syntactical features of the associated languages (Bart´ok 1931). I suspect that there might be a useful contrast to be made between composers whose rhythmic organization primarily reflects dance patterns, and those for whom speech patterns are more important. Tchaikovsky, Dvoˇra´k, Elgar and Stravinsky belong to the first kind; Mussorgsky, Wagner, Jan´acˇ ek, Puccini, Britten and Schoenberg to the second. (In Stravinsky’s Les Noces, for example, the composer uses nonsense syllables, precisely in order to reconstitute speech-rhythms as dance-rhythms. In Stravinsky it is not words that give the meaning of the dance, but dance that gives the meaning of the words.) A. H. Fox Strangways has argued that classical Indian music uses time-value rather than accent to emphasize a note because that is how you emphasize a syllable in Sanskrit. He explains the complex Indian tˆalas as derived from the verse metres of Sanskrit liturgy and poetry, in the way that the metres danced by the Greek chorus reflect the quantities exhibited in the verse.⁶ ⁶ See the illuminating discussion in West (1992, 129–59).

thoughts on rhythm 235 Music supporting movement and music supporting words will differ in measure, accent and stress. In dancing and marching small-scale repetitions are necessary, whereas in song the strophe overrides small-scale regularities and encourages their variation. If there were nothing to rhythm save measure each piece of music could be assigned a clear and unambiguous rhythmic character simply through tempo and metre. But this would be to ignore the effects, not only of grouping and stress, but also of melodic and harmonic organization which, in our tonal tradition, exert their own gravitational pull over the rhythmical movement. A very obvious instance of this is the device of suspension, in which a phrase is held back from its melodic or harmonic closure across a metrical closure, as in the straightforward example of repeated suspensions from Elgar in Example 16. The result is a tie across the bar line—hence a form of syncopation—and also a clash of movements, as metrical closure anticipates the harmonic closure that follows. Atonal music, which (officially at least) does not admit suspensions, cannot generate this kind of rhythmic pulse. Nor can it reinforce rhythmical closure by creating melodic and harmonic closures that coincide with it. The rhythmic effect of harmony is not confined to cases where the harmony is explicitly stated. It can be witnessed in an unaccompanied line, such as that from Parsifal in Example 11. Here the first seven notes (all off the beat) arpeggiate the triad of A flat, to which they add first the sixth and then the major seventh, creating an implied dissonance that is resolved on to the C minor triad, as G, C and E flat are sounded successively in an emphatic down-beat. The intense up-beat experience here is dependent on the cadence implied in the melodic line and could not be reproduced merely by reproducing the note values.

Example 16. Elgar: Violin Sonata, 1st movement.

236

roger scruton

5 The considerations raised in the previous section suggest a use (though not the only possible use) for the distinction between ‘beat’ and ‘rhythm’. ‘Beat’ denotes a pattern of time-values and accents, while ‘rhythm’ denotes the movement that can be heard in that pattern, and which may be influenced by harmony and melody so as to reach across metrical closures and establish contrary motions of its own. A piece of music may have a strong beat but little or no rhythm, and some of the most rhythmical pieces in our tradition are characterized by a light beat and a refusal to emphasize the bar-line. Often a composer will present accompanying figures which shift the accent sideways through the music, as in the trio from Dvoˇra´k’s New World Symphony (Example 17), where a rhythmical cell, replete with melodic and harmonic meaning, imposes its own micrometre on the over-arching triple time. To see the point of this it is sufficient to imagine how uninteresting Ravel’s Bolero would be, were it to consist only of the underlying beat, overlayed by a melody that exactly ran in its groove. The rhythm of this piece is generated within the never-ending syncopated melody, which plays against the beat like a squash-player against a wall. The beat is what makes the rhythm possible, but in itself it is without rhythmic interest. When Jazz musicians distinguish beat from ‘swing’ it is partly this that they have in mind. It has been plausibly argued by Gunther Schuller (1968) that swing involves a feel for the residue of African polyrhythm that underlies Jazz syncopation. But it should not be thought that swing can be reduced to a metrical outline. Swing is a phenomenological feature of the music. It is not understood by counting but felt in the musical line, and Louis Armstrong’s much quoted remark, that if you don’t feel it you’ll never know what it is, conveys a profound truth about the nature of rhythm. Metre is the frame; rhythm the life that grows on it. Jazz idioms like ‘swing’ and ‘groove’ are

Example 17. Dvoˇra´ k: New World Symphony, 3rd movement.

thoughts on rhythm 237

Example 18. Sibelius: Violin Concerto, 3rd movement.

therefore adopted in order to remind us that rhythm is not reducible to metre. In this connection we should draw a contrast between ostinato rhythm, in which a relentless beat subjects the music to a discipline that might have nothing much to do with its melodic and harmonic movement, and rhythm which adapts to and takes its accents from the musical movement. Stravinsky in Oedipus Rex sustains the chorus with an ostinato 6/8 beat, on G and B flat, and the result is rhythmical but static. The beat is like an external force, constraining the music from outside. It is the steady march of fate that can be deflected by nothing in the action. This does not mean that ostinato is a substitute for, or denial of, rhythm, or that it can be understood without those other features—accent, stress and grouping—which I earlier mentioned as fundamental to the rhythmical order. On the contrary, even the simplest ostinato can be heard in competing ways, if accent and grouping are left ambiguous by the melodic line. (Witness the sustained ostinato of the last movement of Sibelius’s Violin Concerto, a fragment of which is given in Example 18.) Such examples show that measure is never sufficient in itself to determine rhythm, even when made fully explicit to the ear.

6 There is an extreme case of the ostinato phenomenon, in which rhythm seems to become detached from harmonic and melodic organization, so as to be fired at them from outside, as it were. I refer to rhythmic ‘backing’,

238

roger scruton

as exemplified by a certain style of pop. This might have a mechanical ‘tic-toc’ character, as in ‘No Son of Mine’ by the group Genesis (from the appropriately named album, We Can’t Dance); or it may depend upon mixing, in which melody and harmony are smeared together so as to become indistinct, leaving only percussive ostinato to establish some kind of measure—as in ‘Be Here Now’ by the group Oasis. Here rhythm has fallen away from the music altogether, leaving a bare shell of melody and a harmonic progression without cogent voice leading, both overwhelmed by the percussive noise from next door.⁷ We encounter in the idiom adopted by Oasis a disaggregation of music into beat and pitch, neither giving true support to the other. The drumkit has the role of marshalling the music to its stride, forbidding all deviation, while adding nothing to the melodic or harmonic development. (Contrast the Dvoˇra´k; here the lilting phrase that generates the rhythm is also replete with melodic and harmonic implications, which subsequently unfold through the melodic line.) The use made of the drum-kit in contemporary pop is to a great extent an innovation. Classical jazz introduced the drum-kit as an embellishment to a pre-existing rhythm, often sounded on the off-beat and hidden, as it were, behind the strumming of the banjo. The rhythm was generated by the syncopated voices of the instruments, each of which played its part in breathing rhythmic life into the bar-lines. Strictly speaking, New Orleans jazz has no need of percussion, which it uses—if at all—purely ornamentally, and in deference to the African polyrhythms which, according to Schuller’s plausible view of the matter, are remembered in the syncopations (Schuller 1968). In modern pop percussion has a constitutive, rather than an ornamental use: without it, there would be no music, since the beat—on which everything depends—would not exist. (All that, it seems to me, is already implied in the word ‘backing’.) In the early days of rock you find a jazz-like use of the drum-kit—not to impose a rhythm synthesized outside the melodic line, but to emphasize and vary a rhythm generated within it. The locus classicus of this is Elvis Presley, whose extraordinary voice, with its barely perceptible microrhythms and tremors, produces melody and rhythm together, so that the ⁷ The prohibitive cost attached to the copyright of pop music prevents me from giving examples. I have been interested to discover that the worse the music, the more expensive it is to reproduce it.

thoughts on rhythm 239 one is inseparable from the other. In ‘Heartbreak Hotel’, for example, the rhythm is compellingly announced by the solo voice, and the bass and percussion seem merely to take it up and prolong it.⁸ Equally impressive in this respect is Eric Clapton, who uses the guitar rather as Elvis uses his voice, to set the music in motion before the drum-kit enters, so establishing the rhythmical identity of the piece as an internal feature (‘Lay Down Sally’ is an effective instance). It is interesting to note that Clapton, in his days as leader of Cream, was criticized by Jimi Hendrix as rhythmically incompetent—a criticism taken up by David Henderson in his biography of Hendrix (Henderson (1983) ). Cream was disbanded in 1969, and Clapton turned his back on Hard Rock, an idiom which in any case destroys the need for rhythmic competence. The guitarist that has emerged from this is surely one who has ‘got rhythm’ in just the sense that Gershwin intended.

7 The dance-forms adopted by Bach and Handel were attached to elaborate rituals and courtesies, and required complicated steps and formations from the dancers. In the ancient dances to which Debussy, Ravel and Respighi looked back with such poignant emotions, partners were assigned by courtesy and exchanged by rule, with people of all ages participating without embarrassment in a dance which could at any moment place them side by side and hand in hand with a stranger. In a very real sense the dancers were generating the rhythm that controlled them, and generating it together, by attentive gestures governed by a ritual politeness. The experience of dancing as a ‘dancing with’ (usually with a group, subsequently with a single partner) survived right down to the days of rock ’n roll. This ‘withness’ of the dance is captured by the baroque and classical idioms, in which rhythmic organization is not imposed but extracted by the metre, as song-like phrases weave around and embellish one another, moving to closures of their own.⁹ ⁸ Presley’s vocal style—heavily influenced by both R&B and Gospel—is a subject in itself, addressed in part by Henry Pleasants in Pleasants (1974: 270 ff ), and by Peter Guralnick, in Guralnick (1982, 1987). ⁹ On the ‘withness’ of the dance see Scruton (2000: 164–76).

240

roger scruton

The human need for this kind of dancing is still with us, and explains the current craze for Salsa as well as the periodic revivals of ballroom dancing and Scottish reels. The ‘withness’ of the reel was noticed and commented upon by Schiller, who regarded what he called ‘English’ dancing as confirming the connection between beauty and gentility. His words are worth quoting: The first law of gentility is: have consideration for the freedom of others, the second; show your freedom. The correct fulfilment of both is an infinitely difficult problem, but gentility always requires it relentlessly, and it alone makes the cosmopolitan person. I know of no more fitting image for the ideal of beautiful relations than the well danced and multiply convoluted English dance. The spectator in the gallery sees countless movements which cross each other colourfully and change their direction wilfully but never collide. Everything has been arranged so that the first has already made room for the second before he arrives, everything comes together so skilfully and yet so artlessly that both seem merely to be following their own mind while never impeding the other. This is the most fitting picture of a maintained personal freedom, which respects the freedom of others. (Schiller 2003: 173–4; I have slightly amended the translation)

It is undeniable that, for many if not most young people, the experience of ‘withness’ is absent from their dancing, which typically involves neither complicated steps nor formations. The normal dancing of the disco floor involves little or no contact with or recognition of a partner, and may occur with no partner at all. You dance to Heavy Metal by head-banging, slam-dancing or ‘mashing’ (pushing people around in the crowd). Such dancing is not really open to people of all ages, but confined to the young and the sexually available. Of course, there is nothing to forbid the old and the shrivelled from joining in: but the sight of their doing so is an embarrassment, all the greater when they themselves seem unaware of this. The social impoverishment of disco has a rhythmical cause. The pulse of disco music sets the dance in motion and controls its beat, but it does little to suggest how or with whom you should move. The dance, like the rhythm, remains external to the music, a kind of generalized ‘setting in motion’ rather than a balletic commentary on the musical line. You see this at its most extreme in techno music, especially when embellished with strobe lights and similar psychedelic effects. Such dancing is like throwing oneself into a pool of collective emotion, to be swept away in its frenzy.

thoughts on rhythm 241 There is nothing you can do, either to create or to embellish the rhythm. And communication with a partner is rendered impossible by the noise, the lights and the sheer formless press of the crowd to every side. Schiller saw ‘English’ dances as emblematic of personality, freedom and the civic community. If we were to view disco dancing through the same Kantian spectacles we should describe it as ‘pathological’—an event in which freedom is displaced by empirical causality, personality by nature and will by desire. Sarabandes, galliards and reels were social dances in the very real sense of being society-forming dances. Modern rock is crowd-forming, rather than society-forming.

8 Dancing shapes the body rhythms and trains the ears of those who engage in it, and changes in the dance-culture will lead of their own accord to changes in the rhythmical organization of music—even the music of the concert hall. This is what we witnessed in the nineteenth century, when gipsy rhythms affected the music of Brahms (explicitly, in works like the G minor piano quartet, but implicitly in many of the other chamber works, songs and concerti), and when the waltz exerted an ubiquitous fascination that was to begin with Schubert’s drawing-room waltzes for piano, and to culminate in Ravel’s monumental La Valse. This is one of the factors that must be borne in mind when we consider the relation between serious and popular music today. People have become used to the ostinato beat of pop, which throbs in the background of life and shapes the expectations of all of us, like it or not. It is hard to attract modern audiences to music in which rhythm is either melodically generated or measured out in Messiaen’s way, by addition rather than division of time-values; it is comparatively easy to attract them to music with an ostinato propulsion, regardless of its melodic or harmonic invention. Hence the popularity of John Adams, whose ‘The Chairman Dances’ (from Nixon in China) and ‘Short Ride in a Fast Machine’ typify a new kind of ostinato writing, with the melodic instruments assigned essentially percussive tasks, and with continual repetition of elemental rhythmical cells. For the pop-trained ear this music is easy to listen to, since its rhythmic structure does not have to be deciphered by following

242

roger scruton

the melodic line but is imposed by a regular external emphasis. Composers used to be wary of ‘the tyranny of the bar-line’; in Adams, however, the tyranny is accepted, as a benign dictatorship which gives us all what we want.

9 This bears on the controversy surrounding atonal music. Much music theory conceives tonality as a melodic and harmonic system, in which both harmony and melody are goal-directed. Chords in this system are not simultaneities, but nodes in a web, bound by functional relations to the chords that precede and follow them. Musical order is achieved when voices move together towards closures that are harmonically driven and melodically complete. Theories of tonality—such as that defended by Schenker—see these features as giving the essence of tonal music. Hence they have a tendency to leave rhythm out of the picture (Schenker 1979; Salzer 1952–62). And those who propose atonality as a genuine alternative to the tonal tradition often describe it in similar terms, as replacing the horizontal and vertical relations of pitches with a new order based on the permutation of pitch-class sets. Rhythm is again left out of the picture, often with counter-intuitive results, since the experience of rhythm is an experience of grouping, and will therefore affect the segmentation of the musical surface on which the pitch-class structure depends.¹⁰ It seems to me, however, that we should see tonality as in part a rhythmical system, and recognize that the difficulties of atonal music are experienced as much at the rhythmical as at the melodic and harmonic level. The harmonic and melodic principles of tonal music arose out of the desire to make satisfying sequences, in which movement begins, continues and comes to a conclusion—not only on the large scale but also region by region and voice by voice. Rhythm is not an addition to this, but a part of it, both sustaining and sustained by the harmonic and melodic relations. Measure is useful not only because it corresponds to familiar movements of the body but also because tonality is founded in repetition, ¹⁰ See Nicholas Cook’s devastating demolition of Allen Forte’s set-theoretic analysis of Stravinsky’s Excentrique, in Cook (1987: 138–51).

thoughts on rhythm 243 with the music constantly returning to identifiable places, and constantly departing on some new but related journey of its own. Take away the old tonal order, however, and rhythm has a tendency either to collapse entirely—since it is not driven by the forward movement of the musical line—or else to become external to the music, a frame supporting an inert display of musical fragments. The atonal composer therefore has a serious dilemma. To suspend organized ‘pitch class sets’ on an ostinato frame is to take a big step backwards into banality. On the other hand, to reconstitute rhythm outside the expectations fostered by the tonal order—expectations of grouping, closure, accent and so on—is to lose the connection between rhythm and the life-phenomena that give sense to it. In response to this problem Milton Babbitt has produced ‘serialized’ rhythms in which the twelve pitch intervals are set in relation with twelve equal time intervals, and both subjected to systematic permutation. However, the very procedure shows that it is not rhythm, but measure, that is being reconstructed. Serialized rhythm is really serialized metre, and is as little part of the rhythmic surface of the music as is the drum-kit in synthetic pop. It seems to me that Messiaen’s rhythmical experiments were prompted at least in part by an awareness of this problem. Messiaen sought to reinvent rhythm as an internal feature of the musical line, even when the melodic and harmonic language had abandoned all the old forms of closure. Messiaen’s experimental approach was taken further by Stockhausen, who attempted to reconstruct the rhythmic dimension of music by localizing movement in competing orchestral units. Stockhausen’s large-scale music shifts great blocks of sound through musical space, often with an indifference to human life and bodily sensations that reminds one of a slave-minder building a pyramid. The result is a meticulous, though hidden, metrical order, which generates an entirely a rhythmical surface. If you were to introduce rhythmic motion into a piece like Gruppen, for example, it would have to be an external rhythm, like a pop-ostinato, laid on top of the musical structure but generated outside it. In itself the music has no enduring pulse; the emphasis generated within one episode does not survive to the next, and therefore cannot combine with it in a dance step. There is a curious parallel here, between Stockhausen and John Adams: both think of rhythm in ostinato terms, the one therefore rejecting it, as an extra-musical device, the other accepting it as the sole

244

roger scruton

organizing principle of a musical surface that in all other respects is wholly monotonous.

10 Like his teacher Olivier Messiaen, Stockhausen uses metres formed by the addition rather than division of time-values. The result may be a-rhythmical, but it is meticulously measured. Messiaen’s music, on the other hand, is not merely highly rhythmical but organized rhythmically, so that rhythm permeates the music with the same intensity as it permeates a symphonic movement by Beethoven. Messiaen has himself written of this subject, emphasizing the way in which rhythm can be altered through ‘added values’ and ‘non-retrogradeable rhythms’ so as to defy metrical symmetries, and compel the listener to cling to the musical movement, rather than the bar-line, in following the score (Messiaen 1956).¹¹ Messiaen took the idea of additive metre from the de¸cˆı-tˆalas of Sharngadeva (see above, Section 3), in which the notion of a beat is replaced by that of the shortest note-value (mˆatra), and metres are classified as repeatable sequences, of which Sharngadeva lists 120. Some of the tˆalas are extremely long—no. 35, simhanandana, has 21 separate ‘beats’, comprising 62 mˆatras—and could not, therefore, be heard as self-contained rhythmical units, on the model of the bars in Western music.¹² And Messiaen’s use of them does not compel us to hear rhythmic groups in his music that correspond exactly to the metrical order: for example, in the Turangalˆıla Symphony at its most rhythmical, we tend to hear curtailed polyrhythms based on division of the bar, rather than the underlying additive order. Messiaen’s principal device is illustrated in Example 19, showing the method of ‘added values’. The first sequence (Example 19 a, b and c) shows three bars in standard metre; the second (d, e and f) shows the transformation of these bars either by adding a rest or lengthening one of the notes by a fraction. Example 20 illustrates the method in practice: Messiaen’s transformation of a Peruvian folksong, Delirio (Example 20a) in the second of the Harawi songs (Example 20b). ¹¹ Messiaen’s theories and practice are lucidly presented in Johnson (1989: Ch. 4). ¹² The tˆalas are listed in Appendix 2 to Johnson (1989).

thoughts on rhythm 245

Example 19. Rhythmic augmentation. (a)

(b)

Example 20. (a) Delirio (Peruvian folk song). (b) Messiaen: Harawi no. 2 ‘Bonjour toi, colombe verte’.

In such examples time-signature is clearly of little importance. Example 19f, for example, can be assigned a time-signature; indeed the performer will be helped, up to a point, if the bar is prefaced with the metrical sign 13/16. But this measure misrepresents the rhythmical movement: the bar has six beats, each subdivided and one of them stretched. By recursively adding or subtracting note-values in this way, Messiaen attempts to build character into his motives, which develop over time by acquiring or losing note-values, while retaining the same number of ‘syllables’ overall. (Compare the way in which a face changes over time, as age first strengthens and then weakens the flesh, always retaining a certain recognizable outline.) Example 21 illustrates the method of ‘non-retrogradable rhythms’. Here Group B is the retrograde of Group A, with the central five semiquaver note shared between them. The whole sequence forms a rhythmic

Example 21. Non-retrogradable rhythms.

246

roger scruton

palindrome which is unchanged when played in retrograde, and hence ‘non-retrogradeable’. Again this metre can be assigned a time-signature, but the result (19/16) is even less informative than in the previous examples. What we hear is a rhythm which seems to flow back into itself, constantly returning the energy that propels it to its source. Messiaen’s rhythms are completely emancipated from the bar-line, and the metrical markings of his scores say little or nothing either about the underlying additive metre, or about the movement that can be heard in it. The result owes its rhythmic vitality not to any regular beat that can be mapped on to bar-lines, but to the energy that flows from note to note and which is inseparable from accent and grouping, both enforced by the melodic line. A work like Cant´eyodjayˆa (Example 22) offers the supreme instance of an organization in which rhythmical order has been emancipated completely from measure, with accent and grouping becoming the unaided vehicles of the rhythmic movement. Certain commentators have described Messiaen’s music of this period as conveying a new, or at any rate unconventional, experience of time: the forward-moving temporal order of the classical style, which leads to the relentless drive of a Beethoven symphony, has been replaced by a kind of circularity, in which everything returns into itself as though in a state of rapt contemplation (Davidson 2001). One can see the point of such descriptions, whether or not one endorses the claim that it is time, rather than musical sequence, that is being experienced differently.

Example 22. Messiaen: Cant´eyodjayˆa.

thoughts on rhythm 247 However, a word of caution is in order. Messiaen, like Babbitt and Boulez, often uses metre in ways that make no contact with rhythm, as this is perceived by the listener. The metrical schemes deployed in Chronochromie, for example, involve permutations of tˆala-like sequences, of a kind that cannot be followed by the ear, still less by the body.¹³ If we hear rhythms in these permutations it is not because of the mathematical order but in spite of it. Fred Lerdahl has argued that musical understanding looks for elaboration rather than permutation in its object (Lerdahl 1988), and therefore that serial organization will always be irrelevant to the organization that we actually hear in the musical surface. Even if not true of all musical phenomena, Lerdahl’s observation is surely true of rhythm, which we understand through patterns, accents and repetitions, creating a movement in which we can join. The passage quoted from Cant´eyodjayˆa in Example 22 can be supplied with a full metrical analysis, in terms of fundamental sequences and their successive permutations. But the rhythm resides in the groupings of semi-quavers, the pauses, and the accented notes, which shake us around as we try to fit them to patterns that cannot quite contain them. The experience of rhythm here cannot be captured by metre, and certainly not by the arcane metrication that preoccupied the composer.

Example 23. Bulgarian Christmas carol.

¹³ See the table of permutations set out in Johnson (1989: 177).

248

roger scruton

11 Although Messiaen deserves credit for recognizing that melodic and harmonic innovation require new ways of grouping tones, and that this, in turn, requires new forms of rhythmic organization, we should recognize that the practice of adding value to notes has been widespread in folkmusic, and is by no means confined to oriental traditions. Example 23 gives a Bulgarian Christmas carol, in which the lengthening of certain syllables has led to the stretching of a bar of four semiquavers to 9/16 and even 10/16 time, while retaining the same basic rhythm of four beats to the bar.¹⁴ Fox Strangways, in his closely observed study of Indian classical music as it was before the disease of Western harmony, persuasively argues that measure is prior to rhythm in the Indian raga because melodies originated as hymns in the Sanskrit language, which emphasizes a syllable by lengthening rather than stressing it, as in ancient Greek prosody (Fox Strangways 1914). Moreover, the metaphysically conceived tˆala system of Sharngadeva does not represent the rhythmic organization that we hear in Indian classical music, as will be evident to anyone who has swayed to Ravi Shankar. Most of the melodies written down by Fox Strangways can be divided into bars of equal length, with the stretching of syllables marked by a pause. Structurally speaking, it is true, additive measure dominates rhythmical division. But this can be accounted for not only by the peculiarities of the Sanskrit language, but also by the early emergence in India of a listening culture, with dancing as an art to be observed rather than an event to join.

12 The need for rhythmical organization to change with the change in harmonic language was widely recognized in the wake of Messiaen’s teaching. But it was not Messiaen who first observed the connection. Wagner was already aware that genuine melodic and harmonic innovation require a new approach to rhythm. If Wagner is not often thought of as a rhythmic innovator, this is because the ostinato conception of rhythm has to a great measure obscured the origins of rhythm in the grouping ¹⁴ Harmonized by Raina Katzarova, and quoted in Bart´ok (1976: 48).

thoughts on rhythm 249 experience, and obscured the roots of that experience in speech, song and dance. In fact the path-breaking character of Tristan und Isolde is shown as much in the rhythmic organization of the musical line as in melody and harmony. This is so in spite of the fact, and also because of the fact, that Wagner hardly ever uses percussion as a rhythm-generating device. The timpani play an enormous role in Tristan, but it is a melodic role, consisting largely of prolonged tremolandi which swell the melodic line and suggest reserves of unexplored emotion beneath it. And the other instruments of the orchestra are used melodically rather than percussively, even in the most frenetic passages such as the scene in which Tristan tears the bandages from his wound, and the music enters a kind of rhythmic catastrophe as it strives to keep pace with his delirium. Wagner’s motives in Tristan have pronounced rhythmical contours, but, unlike the motives in Beethoven say, their emphasis is rarely on the downbeat. Consider the motive sometimes known as the ‘look’ (Example 24), which sets the Prelude in motion, after the interrupted cadence on to F major. This establishes a complex rhythm in 6/8 time, with a fractured triplet followed by a crotchet-quaver sigh. The melodic line here, and the chromatic movement in the bass which lifts the harmony from F major to G major, endow the rhythmic pattern of this phrase with a kind of completeness. This is a rhythm that has been ‘sung out’, and which henceforth bears the memory of the phrases that sang it. Wagner promptly repeats it, incorporating it into no less than five subsequent motives, which he draws together with the original ‘look’ to form a continuous sequence (Example 25 illustrates the musical process).

Example 24. Wagner: Tristan and Isolde, motive of ‘The Look’.

Example 25. Wagner: Tristan and Isolde, elaboration of ‘The Look’ motive.

250

roger scruton

In the classical style the closures imposed by tonal melody and diatonic harmony are aligned with rhythmical closures reinforced by bar-lines. In Tristan all three forms of closure are minimized or avoided entirely: for they are symbols of a law-governed order that passion has undermined. Hence the motives that take up the rhythmic organization of the ‘look’ involve ties across bar-lines, off-beat accents, and the ubiquitous fracture in the triplet, that together create a rhythmic profile that cannot be understood apart from the melodic pulse in which it originates. The prelude to Tristan is already moving in the direction of Messiaen, with the bar-line as the effect, rather than the cause, of the rhythmic organization, which deploys highly energized and repeatable cells, and in which accent and grouping are often at variance with those suggested by the metre.

13 Rhythm is a property of dancing and also of speech. As Aristotle pointed out, we hear speech differently from other sounds,¹⁵ since speech is an expression of soul, and is organized semantically. The surface order of speech may be highly irregular and punctuated by silences; but it is heard as flowing from a single source. Its movement is the movement of meaning, and its pulse is the pulse of life. Speech is therefore a paradigm for us of a rhythmical organization generated not by measure and beat but by internal energy and the intrinsic meaningfulness of sound. Hence there is a great distinction to the ear between speech and chant, in which speech sounds are suspended at a single pitch and on a single pulse. A chanting voice sounds from another region, a place of unseen powers. The ostinato chant represents the inexorable order that controls us and which forces our words to move in time to its commands. (One of the most remarkable developments within the sphere of pop is the ‘rap’ artist, who speaks toneless rhyming prose along the groove of a relentless percussive rhythm, so replacing speech-rhythms with chantrhythms. This is an extreme form of the ostinato experience, in which not only harmony and melody but speech itself are annihilated by metre.) ¹⁵ De Anima, 420b.

thoughts on rhythm 251 Often Jan´acˇ ek is given credit for being the first composer to build speech rhythms into his melodic line. It seems to me, however, that speech rhythms have been part of phrasing in classical music at least since Monteverdi and the stile rappresentativo, and that the conscious attempt to give them musical form began with Wagner, not Jan´acˇ ek. Wagner himself discusses the matter in Opera and Drama, referring to the accents that are natural to the spoken language as underpinning the musical divisions within the bar (Wagner 1887–8: 103–41). His use of Stabreim arose directly from his search for a language that could be sung out without losing the inflections of natural speech. Stabreim derives from two musically important features of Germanic languages, namely their preference for accent over quantity as a form of emphasis, and their wealth of plosive consonants. Schoenberg was heir to Wagner in this as in other things. Unlike so many of his successors on the path of atonality, Schoenberg recognized that atonal music would succeed only if it could emerge as a rhythmical system. His early atonalism was therefore as much a rhythmical as a harmonic experiment. Schoenberg did not wish to repudiate the connection between rhythm in music and rhythm in life. Instead he wished to replace dancerhythm by speech-rhythm as the organizing principle of the musical line. The instrumentation and harmony of a work like Pierrot Lunaire are dictated in part by the desire to generate speech-like rhythms in all the instrumental voices. Tonal harmony compels voices to move together, to magnetize each other, to work towards the same points of closure and stasis—as in a dance. Hence in tonal music, even when speech-rhythm inspires the melodic line, a reminiscence of dance-rhythm will inhabit the harmony. In order to free the music entirely from this dance-like togetherness, Schoenberg pulls the harmony apart, so that what we hear is a simultaneity of utterances rather than a sequence of harmonizing songs. The instruments speak to us, in rhythms that match the Sprachgesang of the voice. The later development of the serial technique likewise has a pronounced rhythmical meaning. The serialization of the pitch sequence permits a kind of polyrhythmic structure, in which accents in the various voices seldom coincide, and the unified chorus gives way to a conversation (examples of interest: the chorus of the Israelites wandering in the desert in Moses and Aaron—contrast the flat and ineffective dance rhythms that accompany the worship of the golden calf).

252

roger scruton

14 Rhythmic organization is fundamental to musical meaning and intimately connected with the moral character of music. The externalized rhythm of much contemporary disco music invites people to move in crowd-like rather than conversation-like ways. You do not ‘move with’ rhythm of this kind; you surrender to it, are overwhelmed by it. You lose yourself in it, as in a drug. And that is a feature not only of the way disco music is received, but also of the moral character that we hear in it. By contrast the ostinato rhythms in the Bart´ok piano sonata, for example, have no reality apart from the sharpened harmonies and emphatic melodic line that produce them. You do not lose yourself in these rhythms: rather you discover something other than yourself—an idealized form of human life, in touch with the soil and with the natural world. The Darmstadt orthodoxy (as propagated by Stockhausen, Boulez and their immediate followers) effectively killed off rhythmical organization as an intrinsic feature of the musical line. Somehow the dancing musical line that Messiaen discovered through the tˆalas disappeared from the music of his followers. The idea of additive metre became institutionalized and mathematized, and accents lost all connection with the body. People can listen to Stockhausen’s Gruppen and move ‘along with’ it, even be moved by it. But the experience of rhythm is absent from Stockhausen’s blocklike sounds, as it is absent from the meticulously metrical music of Brian Fernyhough. In reaction, composers like John Adams have tried to address our need for rhythm by adding rhythm to melodic lines and harmonies that do not have the strength to generate rhythm out of their own inner movement. The rhythm is pumped in from outside, not breathed out by the melody. It seems to me that we are placed by this music in the vicinity of the ‘crowd’ experiences that dominate the world of pop. There are, by contrast, composers still writing who invite us to join in rhythms that are intrinsic to the melodic and harmonic life of their music: David Matthews, for example, in his string quartets and symphonies; John Borstlap in ‘Psyche’; Michael Berkeley in his concertos; Oliver Knussen in his two operas on childhood themes, Higgledy-Piggledy-Pop and Where the Wild Things Are. And these works make systematic use of tonal relations, illustrating the thesis that tonality is a rhythmical system, arising when voices

thoughts on rhythm 253 ‘move with’ each other, as in a dance. They illustrate both the depth and the importance of the controversy surrounding musical modernism. This is not a controversy about harmonic systems only; it is primarily a controversy about the nature of listening, about the role of music in the shaping of our emotions, and about the connection between music and life.

References

Barker A. (1989) Greek Musical Writings, London, vol. 2. Cambridge Readings in the Literature of Music (Cambridge: Cambridge University Press, 1984–9). Bart´ok, B. (1931) Hungarian Folk Music, trans. M. D. Calvocoressi, London: Oxford University Press. Bart´ok, B. (1976) ‘The So-called Bulgarian Rhythm’, in his Essays, ed. B. Suchoff, New York: St Martin’s Press. Bregman, A. S. (1990) Auditory Scene Analysis: The Perceptual Organization of Sound, Cambridge, Mass.: MIT Press. B¨ucher, K. (1909) Arbeit und Rhythmus, Leipzig und Berlin: BG Teubner. Cone, E. T. (1968) Musical Form and Musical Performance, New York: Norton. Cook, N. (1987) A Guide To Musical Analysis, London: Dent. Davidson, A. E. (2001) Olivier Messiaen and the Tristan Myth, Westport, Conn., and London: Praeger. Deutsch, D. (1982) ‘Grouping Mechanisms in Music’, in D. Deutsch (ed.), The Psychology of Music, New York: Academic Press. Fox Strangways, A. H. (1914) The Music of Hindostan, Oxford: Oxford University Press. Gelfand, S. A. (1998) Hearing: An Introduction to Psychological and Physiological Acoustics, New York: Marcel Dekker. Grosset, J. (1913) ‘Histoire de la musique: Inde’, in A. Lavignac (ed.), Encyclop´edie de la musique et dictionnaire du conservatoire, vol. 1, pt. 1, Paris: Delgrave, 287–324. Guralnick, P. (1982) Lost Highways: Journeys and Arrivals of American Musicians, New York: Harper and Row. Guralnick, P. (1987) Liner notes to Elvis Presley: The Sun Sessions CD (BMG/RCA 6414-2-R).

254

roger scruton

Henderson, D. (1983) ’Scuse Me While I Kiss the Sky: The Life of Jimi Hendrix. New York: Bantam Books. Johnson, R. S. (1989) Messiaen, 2nd edn, London: Dent. Jones, A. M. (1959) Studies in African Music, 2 vols, London: Oxford University Press. Lerdahl, F. (1988) ‘Cognitive Constraints on Compositional Systems’, in J. A. Sloboda (ed.), Generative Processes in Music, Oxford: Oxford University Press. Longuet-Higgins, C. (1987) Mental Processes: Studies in Cognitive Science, Cambridge, Mass.: MIT Press. Messiaen, O. (1956) Technique de mon langage musical, Paris: Leduc. Peacocke, C. (1983) Sense and Content: Experience, Thought and their Relations, Oxford: Oxford University Press. Pleasants, H. (1974) The Great American Popular Singers, New York: Simon and Schuster. Riemann, H. (1903) System der musikalischen Rythmik und Metrik, Leipzig: Mayer und Wigand. Salzer, F. (1952–62) Structural Hearing, 2 vols, New York: Dover. Schaeffer, P. (1966) Trait´e des objets musicaux, Paris: Editions du Seuil. Schenker, H. (1979) Free Composition, trans. E. Oser, New York: Longman. Schiller, F. von (2003) ‘Kallias or Concerning Beauty: Letters to Gottfried Keller’, in J. M. Bernstein (ed.), Classic and Romantic German Aesthetics, Cambridge: Cambridge University Press. Schuller, G. (1968) Early Jazz: Its Roots and Musical Development, New York: Oxford University Press. Schutz, A. (1964) ‘Making Music Together’, in Collected Papers, vol. 2, The Hague: Martinus Nijhoff: 159–78. Scruton, R. (1997) The Aesthetics of Music, Oxford: Oxford University Press. Scruton, R. (2000) Perictione in Colophon, South Bend Indiana: St Augustine’s Press. Wagner, R. (1887–8) Oper und Drama, dritter Theil: Dichtkunst und Tonkunst im Drama der Zukunft, in Gesammelte Schriften und Dichtungen, vol. 4, 2nd edn, Leipzig: Fritzsch. West, M. L. (1992) Ancient Greek Music, Oxford: Clarendon Press. Wollheim, R. (1987) Painting as an Art. Princeton, NJ: Princeton University Press.

thoughts on rhythm 255 Wright, O. (1980) ‘Arab Music’, Sections 1–4, in Stanley Sadie (ed.), The New Grove Dictionary. London: Macmillan. Further Reading

Bregman, A. S. (1990) Auditory Scene Analysis: The Perceptual Organization of Sound, Cambridge, Mass.: MIT Press. Cone, E. T. (1968) Musical Form and Musical Performance, New York: Norton. Cook, N. (1987) A Guide To Musical Analysis, London: Dent. Deutsch, D. (1982) ‘Grouping Mechanisms in Music’, in D. Deutsch (ed.), The Psychology of Music, New York: Academic Press. Messiaen, O. (1956) Technique de mon langage musical, Paris: Leduc. Scruton, R. (1997) The Aesthetics of Music, Oxford: Oxford University Press. Scruton, R. (2000) Perictione in Colophon, South Bend, Indiana: St Augustine’s Press.

This page intentionally left blank

Index Adams, John 241, 243, 252 Alpers, Svetlana 153 Aristotle 250 Aristoxenus of Tarentum 231 Armstrong, David 64 Armstrong, Louis 236 Babbit, Milton 243, 247 Bach, Johann Sebastian 80 n., 83, 87 n., 142–5, 239 Barber, Samuel 142–5 Barker, Andrew 231 n. Bart´ok, B´ela 232, 234, 248 n., 252 Baxandall, Michael 153 Beardsley, Monroe 163 Beatles, the 85 Beethoven, Ludwig van 88, 90, 172, 179 n., 184–7, 193, 197–9, 201, 211–2, 213, 219, 249 Berkeley, Michael 252 Black, Max 171 Boghossian, Paul 123 n. Bonds, Mark Evan 166, 178 n., 198 n., 199 n. Booth, Wayne 195 Borstlap, John 252 Boulez, Pierre 232–3, 247, 252 Brady, E. and Levinson, J. 178 n. Brahms, Johannes 241 Brauner, Charles S. 178 n. Bregman, Albert S. 226 Britten, Benjamin 234 Bruckner, Anton 81 B¨ucher, K. 234 Budd, Malcolm 100, 124, 154 n., 172 n. Burnham, Scott G. 188 complete, indicators that a musical work is 79–80 Cage, John 209 Celine, Paul 192 Chopin, Fr´ed´eric 81 Clapton, Eric 239 Cocker, Joe 85

Cocteau, Jean 194 Cook, Nicholas 242 n. Cooke, Deryck 83 Cone, Edward T. 162 n., 173–4, 231 n. Currie, Gregory 32 n., 50 n., 88 Dahlhaus, Carl 165 Darmstadt orthodoxy, the 252 dance 239–42 Danto, Arthur 171 Davidson, Donald 170, 179 Davidson, Audrey Ekdahl 246 Davies, Stephen 25 n., 28, 33, 36–8, 50 n. 56, 81, 85 n., 86, 100, 102–4, 107, 124, 130, 135–9, 142–4, 154–7, 162 n., 169 Daverio, John 179 n. Debussy, Claude 86, 239 Delius, Frederick 145 n. De Momigny, J.J. 231 n. Dent, E.J. 188 Deutsch, Diana 80, 227–9 Diabelli, Anton 197–9 Dill, Heinz J. 178 n. Dodd, Julian 56, 63 n., 73 Dummett, Michael 59 Dvoˇra´ k, Antonin 234, 236, 238 Edwards, George 201 Eiholzer, Hubert 106–7 Elgar, Edward 144 n., 211, 234–5 Eliot, T.S. 212 Ellestr¨om, Lars 184, 185 n. Elliot, R.K. 110–11 electro-sonic art 220–4 emotional responses to music 108–9, 112–13, 117–20 are evidence of musical meaning 119–20 see also musical expression, as arousal Fernyhough, Brian 252 Fox Strangways, A.H. 234, 248 Fubini, Enrico 165, 167 Furtw¨angler, Wilhelm 117

258 index Garland, Joe 85 n. Genesis 238 Gershwin, George 85 n. Glucksberg, S. and Keysar, B. 179 n. Gracyk, Theodore 85 n. Grainger, Percy 87 n. Godlovitch, Stan 25 n., 35–6 Goethe, Johann W. von 192 Goodman, Benny 85 n Goodman, Nelson 53, 55, 133 n., 149, 161, 173–4, 174 Guralnick, Peter 239 n. Handel, George Frideric 239 Hanslick, Eduard 158 n., 210, 212, 218 Hass, Robert 81 n. Hatten, Robert 193, 199 n., 201 Hausman, Carl 153 Haydn, Joseph 80 n., 166, 167–70, 172, 183, 199–202 Heidegger, Martin 52 n. Heine, Heinrich 178 n. Henderson, David 239 Hendrix, Jimi 123, 239 Herzog, Patricia 193 Hopkins, Robert 100, 101 Hume, David 62 Imeson, Sylvia 198 n. instrumentalism 25–6 Jan´acˇ ek, Leoˇs 234, 251 Jarry, Alfred 194 Johnson, Robert Sherlaw 244 n. Johnson, Mark, see Lakoff, George and Johnson, Mark Kandinsky, Wassily 211 Kant, Immanuel 61 n. Karl, Gregory 173 n. Katzarova, Raina 248 n. Keysar, B., see Glucksberg, S. and Keysar, B. Kierkegaard, Søren 192–3 Kinderman, William 193, 197, 199 n. Kivy, Peter 34 n., 63 n., 95, 97, 100–1, 118–19, 149, 189, 209–10, 216–17 Knussen, Oliver 252 Kramer, Jonathan D. 183

La Fontaine, Jean de 157 Lakoff, George and Johnson, Mark 164, 179 n. Leger, Fernand 194 Lerdahl, Fred 247 Levinson, Jerrold 25 n., 26, 29, 30–2, 34–5, 39–49, 50 n., 52 n., 56, 63 n., 64 n., 72 n., 100, 103, 105–7, 110–12, 114, 125–8, 130, 136 n., 139–40, 143 n., 173, 178 n., 213–14, 217–18, 221 see also Brady, E. and Levinson, J. Liszt, Franz 88, 90 Lochhead, Judy 183 Longuet-Higgins, Christopher 230 Longyear, Rey M. 178 n., 179 n., 185, 186 n. 199 n. Lutoslawski, Witold 220, 223 Madell, Geoffrey 98 Magritte, Ren´e 184–5 Mahler, Gustav 79, 83, 164, 174, 193 Matthews, David 252 Mellers, Wilfred 193, 211–12 Messiaen, Olivier 230, 241, 243–8, 250, 252 metre, varieties of 229–34 Meyer, Leonard B. 212 Miller, Glenn 85 n. Miller, Lee 183, 185 n. Monteverdi, Claudio 251 Mozart, Wolfgang Amadeus 79–80, 81, 83, 118, 154, 157 169–70, 188–92 Muecke, D.C. 185 n., 190–1, 192 music acousmatic nature of 27, 229 atonal 218–20, 242–3 created or discovered? 216–17 electro-acoustic, see electro-sonic art indispensably aural character of 209–17 musical expression 95–146, 120–8 and empathy/ sympathy 112–13 and normativity 113–14 as ‘arousal’ 108–9, 112 as imaginative characterisation of music 100–7, 110–13, 139–45 as metaphorical 121–4, 150–74 as metaphorical exemplification 150–74 cannot be sui generis 96–7 comparison with aspect perception 101 comparison with pictorial representation 96, 100

index 259 in terms of a musical persona, see persona account of musical expression is not adequately characterised in terms of causes 97–9, 117–18 Levinson’s dispositional account of 125–8 resemblance account of 101–5, 124–5, 127–8, 135 whether it is a form of meaning 119–24, 149–74 musical irony 178–203 comparison with verbal irony 179–82 dramatic 187–92 general 192–5 parodic 195–9 romantic 199–203 situational 182–7 musical meaning 149–206 musical ontology 23–92 musical performance its relation to the work 70–2 see also performance means essentialism musical works are they sui generis? 75–7 as norm-types, see type-token theory in what sense are they ‘repeatable’? 73–5 the essential meaningfulness of 56–63 the essentially created nature of 64–70 Mussorgsky, Modest 88 n., 211, 212, 234 Neubauer, John 165–6 Newcomb, Anthony 159–63, 172–4 Nowak, Leopold 81 n. Oasis 238 Parker, Mara 194 n. Peacocke, Christopher 226 performance interpretations 83–6 performance means essentialism 39–41 persona account of musical expression 105–7, 111, 130–45, 173–4 relevance of individual character of music to 136–45 Picasso, Pablo 192 Plato 229 Pleasants, Henry 239 n.

Presley, Elvis 238–9 Puccini, Giacomo 234 Rameau, Jean-Philippe 165 Ratner, Leonard 174 n. Ravel, Maurice 88 n., 163, 239, 241 ‘redundancy argument’, the 211–16 Reich, Steve 215–16, 220, 224 Resphighi, Ottorino 239 rhythm 226–55 and morality 252–3 different from beat 236 ostinato 237–9 which reflects dance patterns 234–5, 239–42 which reflects speech patterns 250–2 Riemann, H. 231 n. Rohrbaugh, Guy 73, 75 Rosen, Charles 167, 169 Ridley, Aaron 100, 130, 137 n., 141 Robinson, Jenefer 136 n., 152 n., 173 n. Rousseau, Jean-Jacques 166 Rossini, Gioachino 117 Salzer, Felix 242 Sartre, Jean-Paul 192 Sawyer, John E. 188 Schaeffer, Pierre 229 Schelling, Friedrich von 168 n. Schenker, Heinrich 242 Schiller, Friedrich von 240 Schindler, Anton F. 197 Schlegel, Karl W.F. 179 n. Schoenberg, Arnold 195 n., 218–19, 234, 251 Schopenhauer, Arthur 168 n. Schubert, Franz 79, 162, 173, 179, 241 Schuller, Gunther 236, 238 Schumann, Robert 164, 178 n. Schutz, Alfred 229 Scruton, Roger 27–8, 96, 114, 121–4, 127, 179, 180 n., 212, 218–19, 229 n., 239, 227 Searle, John 180 Shankar, Ravi 248 Sharngaveda 230, 245, 248 Sheinberg, Esti 178 n., 192 Sibelius, Jean 237 Sibley, Frank 157 n., 178 n.

260

index

Skryabin, Alexander 219 sonicism 23–51 pure 27–31 timbral 31–50 arguments from scores against timbral 31–8 Stravinsky, Igor 81–2, 86–7, 193–5, 219, 231, 234, 237 Sterne, Laurence 198, 199 Stockhausen, Karlheinz 243–4, 252 Suurp¨aa¨, Lauri 178 n. Swift, Jonathan 195 Tchaikovsky, Pyotr Ilyich 87 n., 119, 234 Tieck, Ludwig 179 n. Tolstoy, Leo 90 Tovey, Donald F. 197 transcriptions 86–8 translations in literature and poetry 89–92 Treitler, Leo 162 type-token theory 23–4, 52–6, 63–4, 67, 71–2, 73 universals, musical works as 52–3 Vangelis 224 Var`ese, Edgar 220–1 verbal irony, semantic analysis of 180–2 Vivaldi, Antonio 157

Wagner, Richard 124, 164, 230–1, 234, 235, 248–50, 251 Walton, Kendall 25 n., 100, 109–11, 180 Watkins, Glenn 194 Webern, Anton 195 n. West, M.L. 234 Wittgenstein, Ludwig 62, 101, 193 Wollheim, Richard 64 n., 96, 100, 196 n., 229 n. Wolterstorff, Nicholas 23 n., 24, 55, 63 n., 73, 74 n. work versions by others 83 characterisation of 82 contrasted with performance interpretations 83–6 contrasted with transcriptions 86–8 examples of 81–2 in art forms other than music 88–9 paintings and sculptures as successive 88–9 translations are a form of 90–1 Wright, Crispin 125 Wright, Owen 229 Yablo, Stephen 123 n. Zemach, Eddy 180, 196 n.