You are on page 1of 10


Music, Metaphor, and Emotion

i. the issues We describe music in terms of emotion. How should we understand this? Some say that emotion descriptions should be understood literally. Let us call those views literalist. By contrast nonliteralists deny this and say that such descriptions are typically metaphorical.1 This issue about the linguistic description of music is connected with a central issue about the nature of music. That issue is whether there is any essential connection between music and emotion. According to what we can call emotion theories, it is essential to music to be somehow related to real emotion. Prominent examples of such theories are these: it is the main function of all or most music to express emotions, to arouse emotions, or to represent emotions.2 In my view, such theories have little plausibility, and they face a battery of powerful objections. In particular, these theories are objectionable on the grounds that essential features of emotion preclude such essential relations between music and emotion.3 Yet to argue against various specific emotion theories of music, of which there is a large variety, does not address the reasons that draw people to emotion theories. I think that there are two main reasons. The first is that the most obvious explanation of why we describe music in emotion terms is that emotions, or relations to emotions, are part of what music is. The second reason is introspective, or phenomenologicalthat much music moves us when we listen to it, so it seems that music generates emotions in us, which we project onto the music when we describe it in emotion terms. In this article, my negative purpose is to dissolve these two reasons. My positive purpose is to argue for a particular nonliteralist view of linguistic descriptions of music in terms of emotion, a

view according to which emotion descriptions are metaphorical descriptions of aesthetic properties. I call this the aesthetic metaphor thesis. This view is naturally conjoined with a view of the nature of music according to which music has the function of sustaining aesthetic properties. On this view, music itself has nothing essential to do with emotion. If emotion descriptions are not literal, then that removes the first reason that people have for turning to emotion theories. I shall also show how we can nevertheless allow a sense in which music may move us, even though it does not generate emotions. That undermines the second reason. Furthermore, we will see that in many respects, the precise role that emotion descriptions play supports the aesthetic metaphor thesis over its rivals.

ii. against the obviousness of literalism Literalism is not obviously correct. I begin with some observations that show that literalism should not be assumed to be the default view. First, we describe nature in emotion terms. As I mentioned, one reason that people are drawn to emotion theories of the nature of music is that they seem to be the most obvious way to explain the indisputable fact that we so often describe music in emotion terms. This is understandable. What could be simpler than the idea that we describe music in emotion terms because music is in some way connected with emotion? However, this step should seem far less obvious if we bear in mind that we also describe nature in emotion terms. For example, we talk of proud rocks, shy flowers, or angry clouds. Since this is so, it is clear that there is no general link between the prevalence of emotion descriptions and the possession, expression, or representation of emotion. For inanimate

The Journal of Aesthetics and Art Criticism 65:4 Fall 2007

392 natural things cannot possess, express, or represent emotion. They cannot have emotions or any other thoughts or feelings and, not being artifactual, they cannot express or represent emotions. That leaves open the possibility that nature arouses emotions. But it is at least clear that we should be cautious about postulating emotions corresponding to emotion descriptions. Therefore, it is not obvious that the description should be taken literally. Second, we also describe other sorts of works of art beside music in emotion terms. For example, it is common to give emotion descriptions of abstract paintings and sculptures. We might talk of the joyful colors of a painting or of the angry forms of a sculpture. However, in the case of the representational arts, our purpose in giving emotion descriptions is clearly different, since some person is represented as having an emotion or some situation is represented in way that invites an emotional response. But there are also emotion descriptions of abstract paintings and sculptures that do not depend on representation. It is not obvious that there is anything special about music in this respect. Third, many writers assume that it is uncontroversial to claim that music has expressive properties. If they mean by this merely that music is appropriately described in emotion terms, then that is unobjectionable, since it leaves open the question of whether they are literal or metaphorical. But they often aspire to say more than this. They say something controversial and disputable if they use the word expressive so that ordinary emotion descriptions are somehow literally construed as referring to emotional mental states. If nonliteralism is right, that is a mis-construal: ordinary folk are right to use emotion descriptions when talking about music; the theorists are wrong in their literal understanding of those descriptions. We need to be able to discuss whether emotion descriptions are literal or metaphorical without that question being closed by speaking of expressive properties. These observations have force given an intuitive distinction between metaphorical and nonmetaphorical descriptions.

The Journal of Aesthetics and Art Criticism literalism. However, I now want to give a positive argument in favor of nonliteralism. What I shall call the parity argument for nonliteralism about emotion descriptions begins from the fact that:
We give many descriptions of music (and of the way music sounds) that are not emotion descriptions, yet which are obviously metaphorical.

For example, we describe music as delicate or balanced. In these cases, no one rushes to say that the words serve to refer to some emotion that the music is connected with. (Feeling delicate? Feeling unbalanced?surely not.) Something that is aesthetically delicate need not be liable to break, and something that is aesthetically balanced need not have an equal distribution of weight among its parts. So these uses of delicate and balanced are metaphorical. Furthermore, Roger Scruton, following Victor Zuckerkandl, has emphasized that our descriptions of music in terms of height and motion are metaphorical.4 Although this comes as something of a surprise when one first encounters the claim, Zuckerkandl and Scruton are clearly right about this, and importantly so. Yet metaphors of height and motion are not emotion metaphors. Philosophers of music need to keep these descriptions in mind as well as emotion descriptions. The dominance of emotion descriptions can be exaggerated. We should not fixate on emotion descriptions. But then we need a very good reason not to treat emotion descriptions as we do these nonemotion metaphorical descriptions of music. Surely, a unitary account of such descriptions is desirable. Since delicate, balanced, high, and moving from note to note are clearly metaphorical descriptions of music, we should treat angry similarly. Treating emotion and nonemotion descriptions in the same way is common sense, and it has the virtue of simplicity. Furthermore, it is not merely that we offer both emotion and nonemotion descriptions, but that emotion descriptions and nonemotion descriptions are interwoven. What we can call the interweaving thesis is this:
Emotion descriptions of music are intimately connected with other descriptions that are obviously metaphoricalin particular, emotion descriptions are often offered because of obviously metaphorical nonemotion descriptions, and vice versa.

iii. the parity argument Above, I drew attention to emotion descriptions of nonmusic in order to remove bias in favor of

Zangwill Music, Metaphor, and Emotion For example, we might say that some music is serene because it is delicate. Yet if these nonemotion descriptions obviously cannot be treated except as metaphors, it would be odd to treat emotion descriptions differently. According to the interweaving thesis, the use of metaphors of emotion in music is often closely connected with the use of other metaphors and cannot be understood without them. Angry music is usually violent or jagged. Emotion descriptions of music are only appropriate to absolute music because other metaphorical descriptions are appropriate. We should have a unitary account of these intimately interwoven descriptions. In this parity argument, I have appealed to the obviousness that nonemotion linguistic descriptions of music are metaphorical without giving a general account of what makes a description a metaphorical description. The parity argument has some force even without such an account, but we do indeed need to say something more if we are to pursue the discussion further.

393 a criterion for telling whether an emotion word is used metaphorically without a full-scale theory of metaphor.6 It also means that we can defend the thesis that emotion descriptions are metaphorical without depending on a controversial general theory about the nature of metaphor. Stephen Davies has claimed that descriptions of music, such as sad, delicate, and high, are not metaphorical but are literally applied to music.7 He thinks that these words are ambiguous, having both primary and secondary literal meanings. On Daviess view, emotion descriptions of music are literal, like talk of table legs and river mouths. It is true that some emotion descriptions of music are very familiar and cliched. So there is an understandable temptation to think that these metaphors have died and gone to literal heaven. But as a general thesis, Daviess thesis seems implausible when we consider that these descriptions are of the same sort as those we can freshly coin for describing music, which are obviously metaphorical. For example, I do not recall having heard the following descriptions of music, but I can think of music that I would want to call perturbed, remorseless, hesitant, nervous, insistent, vibrating, and so on. These are clearly metaphors and are of the same sort as cliched descriptions of music as sad, delicate, and high. Do they suddenly acquire secondary meanings? There is surely a lot to be said for a unitary account of both fresh and cliched emotion descriptions of music. Saying something quite different about fresh and cliched emotion descriptions is ad hoc.8 What is more important, however, is that regardless of whether the emotion words that are used of music are metaphors or dead metaphors, they do not describe states with the essential characteristics of emotions in the primary sense of these words. Even if Davies were right, and some emotion descriptions of music are dead metaphors and are now literal, with a secondary sense, it changes little. It is still not the case that descriptions of music in emotion terms refer to real emotional mental statesthat is, states that are had by a person, that have intentional content, that have qualitative character, and that are rationalized in distinctive wayswhich the words taken in their primary literal sense would suggest. So we can stipulate that a dead metaphor view of emotion descriptions counts as a nonliteralist view, since it invokes secondary meanings, not primary

iv. criteria for emotion metaphors I propose to avoid giving a quite general account of what metaphor is (and thus of the metaphorical/nonmetaphorical distinction), but I shall propose criteria specifically for the metaphorical and nonmetaphorical uses of emotion terms. This will enable us to make progress with the issue we are interested in, without an overly controversial general theory. I assume that emotions, or at least central cases of emotions, are essentially mental states that a person has, which have content, which (when occurrently manifested) have qualitative character, and that have distinctive rationalizations.5 For example, anger, fear, and pride are had by people, they are about something, they feel a certain way, and they are irrational if had without certain beliefs. Now if someone applies an emotion word to something that does not have these characteristics, or that does not stand in any obvious relation to something that has these characteristics, thenunless the person is mad or deludedthat linguistic use is probably metaphorical, or at least an extended use of the word. For example, if an emotion word is applied to a person, it is likely to be literal, but if it is applied to the sky (angry sky), it is likely to be metaphorical. This gives us

394 meanings that refer to real emotions, where real emotions have the essential characteristics that I listed.9 A dead metaphor view is in principle compatible with the view defended here, and Davies could view my criteria as distinguishing primary and secondary meanings.10 I stipulate that a literal use of emotion words is one where they refer to emotional mental states or something that stands in relation to emotional mental states. The issue is whether emotion descriptions of music refer to such states. Thus far we have seen some reasons to think that they do not. More reasons will follow. v. literalism, nonliteralism, and the nature of music How exactly is the issue about the linguistic description of music connected with the issue over the nature of music? As I have characterized literalism, emotion descriptions of music refer to real emotions, which have the features that I listed. Hence literalism, if true, has implications for the nature of music: it implies that music stands in relation to emotion. And given the prevalence of emotion descriptions, it also invites the stronger thesis that music is essentially related to emotionthe emotion descriptions describe something that is essential at any rate to the music being described. And perhaps most or all music invites such essential emotion description. So literalism implies an emotion theory of music. What about the other direction? What if nonliteralism is right? I think that nonliteralism at least strongly encourages a nonemotion view. If emotion descriptions do not literally refer to emotions, then surely music itself has nothing essential to do with emotion. Thus, there would seem to be a twoway connection between the linguistic theses and the theses about the nature of music: it is natural to think that literalism goes hand in hand with emotion theories, and nonliteralism goes hand in hand with the rejection of emotion theories. There is a complication, however, which is that it may be possible to combine nonliteralism with a more sophisticated emotion theory according to which music is related not to emotions but to thoughts about emotions. We might call that an intentional emotion theory. On Roger Scrutons view, for example, we imagine emotions when we listen to music.11 Jerrold Levinsons view is in the same genre, since he also appeals to imagining

The Journal of Aesthetics and Art Criticism emotions.12 Imagining an emotion is a thought about emotion that is not a belief. Nevertheless, on Scrutons view, descriptions of music in emotion terms are not to be taken literally.13 It is not clear whether such theories should be classified as emotion theories, since real emotions are not invoked, only thoughts about them. The classification does not matter much. However, the combination of nonliteralism with an intentional emotion theory is puzzling: if thoughts about emotions are involved in the experience of music, surely we might expect linguistic descriptions of music literally to mean the emotions that are thought of in those experiences. Moreover, intentional emotion theories are as badly off as nonemotion theories in doing justice to the introspective and phenomenological reasons that people have for favoring straightforward emotion theories, since there is no reason to think that a thought about emotion will share the phenomenology of the emotion itself.14 vi. the aesthetic metaphor thesis and the aesthetic theory of music Thus far, I have talked of nonliteralism and nonemotion theories. But what are the positive options? What I shall call the aesthetic metaphor thesis is this:
Emotion descriptions of music are (mostly) metaphorical descriptions of its aesthetic properties.15

In the rest of this article, I shall defend and argue for this thesis.16 However, in this section I shall make some comments on the view of the nature of music that this aesthetic metaphor thesis is naturally conjoined with. The aesthetic metaphor thesis is not itself a theory about the nature of music, but it does, however, invite an aesthetic theory of music according to which the function of music is to generate aesthetic properties that depend on sounds.17 Such aesthetic properties include beauty, elegance, daintiness, and, crucially, a host of aesthetic properties that we describe metaphorically. For the aesthetic theory of music, ordinary emotions are neither part of either the immediate creation nor of the immediate reception of music. Ordinary emotions may sometimes play a role as the more distant causes of creative activity, and emotions may sometimes be among the further effects of the experience of music. But that is of

Zangwill Music, Metaphor, and Emotion little significance. Primary musical experience and creation does not involve ordinary emotion. By primary musical experience of music I mean the experience that is of that music, and by primary musical creation I mean the creative thought about music that brings it into being. Is the aesthetic theory of music formalist? Eduard Hanslick does not have too much to say about his positive formalist view of music in his great work On the Musically Beautiful. Hanslick concentrated on his array of devastating attacks on emotion theories. He does say, however, that the elements of music are tones in their artistic combination.18 The aesthetic theory of music is perhaps necessary for fleshing out a formalist view of music. However, it is compatible with certain anti-formalist viewssuch as those that emphasize the role of the artistic context or of the artists intentions in partly determining the aesthetic characteristics of music. It is not compatible with anti-formalist views according to which aesthetic properties are partly constituted by relations to ordinary emotions (or thoughts about them). As it happens, I am a formalist about much music. The aesthetic theory of music is certainly comfortably compatible with formalism. It is also compatible with many anti-formalist views. Either way, the aesthetic theory of music can embrace the aesthetic metaphor thesis and say that the aesthetic properties of music are often metaphorically described in emotion terms. The formalist will say that the aesthetic character of music is internal to the sounds, and that is what is described metaphorically in emotion terms. Although we may need training and acculturation in order to appreciate those formal values, that does not mean that what we appreciate, armed with that training and acculturation, is not internal to the sounds we hear. Anti-formalists, for their part, will say that the aesthetic characteristics of music depend on its musical and nonmusical contexts, and that our appreciation of its aesthetic properties depends on knowing those contexts and hearing the music in the light of that knowledge. I am skeptical of such claims, certainly for much music.19 But perhaps some musical appreciation is like this. The aesthetic metaphor thesis would apply just the same to such contextually determined aesthetic properties of music, if there are such aesthetic properties. Contextually determined aesthetic properties might be metaphorically described. In some versions, an aesthetic theory of music

395 might allow that aesthetic appreciation of music involves a specifically aesthetic or musical emotion. On such a view, the emotions we feel when we appreciate music are not the ordinary emotions that are usually referred to by the words used in emotion descriptions of music. If there are special musical or aesthetic emotions, then our emotion descriptions of music are not literal descriptions of the aesthetic or musical emotions. As we shall see, this is dialectically important because of the introspective or phenomenological reason that people have for embracing emotion theories of music, which is that they are aware of being moved by music. Music affects them phenomenologically. It is true that music affects themit makes them feel a certain waybut what they feel are not the emotions that correspond to emotion description of music. As Peter Kivy has urged, whatever they are, they are not what he calls garden variety emotions. They might be specifically aesthetic feelings that do not coincide with the ordinary emotions that we often ascribe to music, such as anger and pride.20 I return to this theme in Sections IX and X.

vii. robinsons visceral emotions Jenefer Robinson argues that some emotions, such as exhilaration and the startle reaction, do not have intentional content, and she claims that music can provoke such content-less emotions.21 I have elsewhere commented that Robinsons account will not cover enough of the cases where we describe music in terms of emotions, such as pride, which do have intentional content.22 I argued that even if she is right about the phenomena of musical experience to which she draws our attention, that analysis does not cover many other cases of emotion descriptions. We deploy terms for sophisticated thought-dependent emotions as well as the more visceral emotions and moods that she focuses on. However, I now want to argue that Robinsons analysis of even those kinds of cases that her account best fits is not adequate and that the aesthetic metaphor thesis does them better justice. Consider, for example, that we might plausibly describe some passages of Shostakovitchs Fifth Symphony as exhilarating, where that is meant in the ordinary sense in which we feel exhilarated after running. It is true that this exhilaration is

396 not a sophisticated thought-dependent emotion. However, this is only because there is also exhilaration in the music that is not felt exhilaration of a Robinsonian sort. If we feel Robinsonian exhilarationthough it is surely contingent that we dothen that feeling is consequential on our recognition of the exhilaration that inheres in the music. The music is not exhilarating because we are exhilarated; rather, we are typically exhilarated because we hear exhilaration in the music. The exhilaration we feel derives from the exhilaration we hear. The plausibility of this analysis is enhanced when we consider that we hear the aesthetic exhilaration in the music only because there are other qualities in the music that we also perceive and react to. The interweaving thesis is important here. The exhilaration is part of the tension and energy of the music that is the intentional object of musical perception and musical response. The music is exhilarating because it is strident, because it is charged, because it is tumultuous, confident, brash, and so on. But these are all aesthetic properties. We should avoid an atomism of properties and responses. We hear a complex of characteristics of the musiccharacteristics that give it aesthetic value (as we shall see in the next section). That experience may then generate Robinsonian felt exhilaration in listeners. Or it may not. But where it does, that feeling lies downstream from primary musical experience, which is the hearing of qualities, such as exhilaration, and the other features interwoven with it, in the music.23 Similarly, there may be objectless feelings of unease caused by musical perception and response; but these objectless feelings lie downstream from primary musical experience. The interweaving thesis puts added pressure on Robinsons literalist emotion thesis because it is not plausible that the other interwoven features are ascribed to music on the basis of a Robinsonian content-less emotion. Thus when we reflect on Robinsons visceral responses to music, we are led away from literalist and emotion theories to the aesthetic metaphor thesis.

The Journal of Aesthetics and Art Criticism ative direction or polarity of emotion descriptions of music. What we can call the evaluative direction thesis is this:
Emotion descriptions of music function as explanations and justifications of judgments of aesthetic value.

viii. evaluative direction A major advantage of the aesthetic metaphor thesis is that it fits comfortably with, and is the best explanation of, an important fact about the evalu-

For example, we might say that some music has positive aesthetic value because it is angry or proud or serene. Emotion metaphorical descriptions of music are typically linked to judgments of aesthetic valuepositive or negative.24 According to the aesthetic metaphor thesis, we use emotion descriptions to describe the substantive aesthetic features of music that make it aesthetically good or bad.25 It is difficult to see how literalist theories can explain the evaluative direction thesis. In particular, theories other than the aesthetic metaphor thesis have a problem with explaining the direct way that emotion descriptions typically function as reasons for evaluations. The evaluative reasongiving aspect is intrinsic to the emotion description because what is described is an evaluatively charged feature. But any literalist or emotion theory must deny this. It might be objected that some emotion descriptions seem to be nonevaluative, but this is only because we are considering the descriptiontype, out of the context of some concrete application of it to a particular work or stretch of music. In such contexts, particular emotion descriptions function as reasons for evaluations.26 However, even if this is not universally correct and there are some emotion descriptions that are evaluatively neutral, the argument still goes ahead, for there are many other emotion descriptions that by themselves yield evaluations. An emotion theorist might perhaps offer an account of how relations between music and emotions (or thoughts about emotions) are evaluatively relevant. But such theories will fail to capture the immediacy of evaluative directionality. In folk aesthetic thought, the emotion description is directly or intrinsically evaluatively relevant; it is not relevant because of the sort of extraneous facts to which emotion theorists appeal. Consider an analogy: in folk morality, it seems that a bond such as love, friendship, or family sets up an immediate moral requirement; it does not seem to do so in virtue of the fact that it is on the whole a good thing if people have such feelings or bonds. A person who feels bound by filial

Zangwill Music, Metaphor, and Emotion duty, such as Cordelia of King Lear, would surely reject an indirect consequentialist account as untrue to the nature of the obligation that she felt applied to her. We think of it as holding in virtue of her bond and nothing morecertainly nothing to do with the consequences of the respect for bonds of that kind. Ergo it seems that indirect conseqentialism fails to match folk morality, where folk morality, crucially, does not just include verdicts about cases, but also grounds for verdicts. Similarly, the ordinary musical listener typically thinks of emotion descriptions of music as describing something that is of aesthetic value in itself. Emotion theorieswhether straightforward or intentionalare at variance with the ordinary musical listeners experiences of, and judgments about, that in virtue of which music has its aesthetic value. Of course, folk aesthetics, like folk morality, might be wrong. But I doubt that this is how emotion theorists want to win the debate. Thus the evaluative direction thesis is a major consideration in favor of the aesthetic metaphor thesis.

397 the listener is experiencing ordinary emotions of grief, melancholy, and despair that the terms taken literally would suggest, rather than experiencing these enjoyable features of the music that we describe metaphorically in those terms. Again, literalism and emotion theories lead to problems and paradoxes, whereas the aesthetic metaphor thesis leads to explanation and understanding. On this view, the issue of negative emotion in music should be firmly distinguished from issues about negative emotion in representational arts such as literature or painting where genuine emotional responses of a sort may be required of the reader or spectator. When we feel pity for a fictional or depicted person, it is an emotion that is not quite normal pity, since we do not believe that the person exists, but it is certainly not an experience of an aesthetic characteristic that is described metaphorically as pitiful. The issues over our emotional responses to literature and representational painting are complicated and controversial; but if the aesthetic metaphor thesis is right, they are completely different issues from those surrounding emotion and music. In literature and painting, it seems that we should have emotional responses that are necessarily related to beliefs or entertained thoughts about fictional or depicted people. For the aesthetic metaphor thesis, nothing analogous is the case in our experience of music.28 Jerrold Levinson very briefly considers something like the resolution of the paradox that I propose.29 He advances what seems like a good objection to it: that the view implies that listening to music would be cool and detached. The objection is that the appeal to our cognition of aesthetic properties of the music, which we just happen to describe metaphorically in emotion terms, cannot do justice to the fact that we respond with powerful feelings to some music. This is the introspective and phenomenological point mentioned at the beginning of this article. I concede some force to this objection. My response is to say that primary musical experience involves various experiences, such as perceptual experiences of the music and experiences of pleasure in the music. These experiences are not emotions, such as grief, anger, passionate longing, or despair. If we have such emotions as a consequence of listening to music, they are quite accidentally connected with our perceptual or pleasurable experience of music: either there is a very variable correlation between listening to music

ix. negative emotion and pleasure Let us now consider what has come to be called the paradox of negative emotion. We will see that the aesthetic metaphor thesis allows us to avoid this paradox and also that it allows a superior interpretation of the phenomenology. There seems to be something paradoxical about the fact we enjoy music that we describe in terms of negative emotions, such as grief, melancholy, or despair.27 Surely, we would be crazy to want to experience music with these characteristics. Do we want to make ourselves miserable? Is the appreciation of such music a masochists pastime? However, this paradox disappears once we accept the aesthetic metaphor thesis together with the evaluative direction thesis. It is not that we relish feeling grief, melancholy, or despair, and we want to experience more of such feelings. That would indeed be crazy. Instead, we experience characteristics of the music that we describe metaphorically as sad, melancholy, and desperate, and the experience of these characteristics of the music yields pleasure (of a certain sort). These metaphorically described aesthetic features of the music are what we enjoy. They ground our pleasure. There is therefore no paradox. The paradox arises only if we think that

398 and such emotional reactions, or else, even if there are some typical music-reaction correlations, they are not part of the nature of music or musical experience. Our pleasurable experiences of music prompt us to describe the music metaphorically in emotion terms, and there is no reason to suppose that those pleasurable experiences are cool and detached. Suppose that someone nevertheless intransigently insisted that musical experiences are emotions. In reply, I would concede that it is true that musical experiences have some features in common with many uncontroversial cases of emotionthey have intentional content and they are felt. That is, they are directed toward properties or things in the world and they have a qualitative aspect. So we could call them emotions (even though that category would now include all pleasures that have intentional contents). The notion of emotion may be somewhat elastic. Other cultures have notions of emotion that group mental states differently from our notion. What is clear, however, is that whatever this mental state is that we have when we hear (or make) music, it is quite unlike the standard uncontroversial cases of emotion, such as anger, grief, pride, or passionate longing, which emotion descriptions of the music, taken literally, would suggest. Irrespective of definitions of emotion, that is surely an error.30 x. ecstasy and substantive aesthetic properties The issue of negative emotion in music is in fact not just a problem that the aesthetic metaphor thesis can overcome, but something that weighs significantly in favor of the aesthetic metaphor thesis and against emotion theories. Thus far, I have talked of the pleasure and displeasure that we take in music. Perhaps speaking of mere pleasure does not do justice to the rapt and intense and ecstatic nature of some of our experience of music. Some of our experience of some music has a peculiar intensity. Of course, much musical experience is disappointingly mundane; but some is not. What about musical ecstasy? When we are listening intently to musicwhen we are rapt and ecstaticwhat exactly is going on in us? That is, what is the nature of these special and intense experiences of the music? It might be argued that to do justice to these intense experiences, we must see the experiences as involving emotion rather

The Journal of Aesthetics and Art Criticism than mere pleasure. (Note that thoughts about emotion will not do.) It is true that we must go beyond a simple appeal to pleasure. But this is because to appeal to pleasure is necessary but not sufficient. It is the sort of pleasure that is important. It is not mere pleasure, but it is not ordinary emotion either; rather, it is a special kind of pleasure. The experience of music need not be cool and detached. There is a venerable tradition in aesthetics, running through Hume and Kant, according to which judgments of aesthetic value or beauty involve a special kind of pleasure and displeasure. I would suggest that we add that substantive aesthetic judgments require more specific aesthetic responses, with an experiential character corresponding to substantive aesthetic features, just as pleasure corresponds to beauty and displeasure to ugliness. Metaphysicallysubstantive aesthetic properties determine the beauty or aesthetic value of the music; experientiallythe specific felt aesthetic reaction justifies pleasure and displeasure. But there is no reason to believe that specific aesthetic reactions are emotions. Consider our experiences of substantive aesthetic features of music, such as those we might describe as angry or passionately longing. These experiences are not emotions, like real anger or real passionate longing. For one thing, the essential normative constraints on such emotions are lacking. For example if we are really angry, we should believe or at least entertain the thought that someone has wronged us, and if we imagine anger, we should imagine a state with such normative characteristics. But nothing like this is true of the experience of angry music. It is more natural to describe the experience as the experience of anger or passionate longing in the music; but that is a metaphorical description of the content of our musical experience. If we are lucky, the music is wonderfully angry and listening to it gives us great pleasure. So I do not want to imply that our experience of listening to music is somehow cool and detached, which Levinson would find objectionable. We need not be driven to an emotion theory for fear that otherwise we cannot allow that we are deeply moved by music. The aesthetic metaphor thesis can allow that listening to music causes feelings of pleasure and displeasuresometimes intense and ecstatic pleasures and displeasures, and metaphors can describe what generates such pleasures. Such pleasures and displeasures have the

Zangwill Music, Metaphor, and Emotion music as their intentional objects, and such pleasures are the grounds of judgments of taste (or aesthetic value) about music. However, such pleasures and displeasures are quite different from emotions, such as anger or pride, which emotion descriptions would describe if taken literally. For, except in very unusual cases (such as when I am the composer), anger or pride would not have the music itself as its object, unlike genuine musical experience; and anger or pride involves normative relations to other mental states, such as beliefs and desires, that genuine musical experience lacks. If listening to music yields these feelings of pleasure in the beauty of the musicpleasure that is of an especially valuable sortthen it is understandable that we value it. And, if substantive features of music are those that make it beautiful, then it is understandable that we value those substantive features and our experiences of them. Anger and grief in the music, and our experience of those features of the music, are valuable in this way. No wonder then that we want to listen to music that has these characteristics. Literalist and emotion theories, by contrast, have a huge difficulty saying something believable about why we want to experience anger and grief, and even if they tell a semi-plausible story, perhaps invoking catharsis, or some such idea, the aesthetic metaphor thesis is a simpler and more plausible explanation.31

399 able to discover what it is about music that we truly care about and what it is about music that we value. The nature and value of music will elude us so long as we are mired in emotion. Once we are liberated from emotion we can see music as a world quite unto itself, a world with features that we describe with emotion metaphors, which may give the music a value that we can experience with intense delight and even ecstasy.32

Department of Philosophy Durham University Durham DH1 3HN, United Kingdom internet:

xi. coda The aesthetic metaphor thesis explains prominent features of emotion descriptions of music features that are hard to explain without it. In particular, the aesthetic metaphor thesis has the best explanation of the role of visceral responses to music and of the evaluative direction of emotion descriptions. Furthermore, it allows us an easy way round the paradox of negative emotion and yields a good explanation of our being moved by music with apparently negative characteristics. It is easy to explain how emotion theorists come to be misled: they take metaphorical descriptions literally, and they confuse the feelings involved in the immediate experiencing and creation of music with emotions, whereas in fact they are another sort of mental state with some but not all of the features of emotions. We must turn our backs irrevocably on emotion theories of the nature of music. Only then will be

1. Metaphors are, of course, not the only nonliteral linguistic device, but to simplify matters it will help to assume that they are. 2. For an example of an expression theory, see Aaron Ridley, Music, Value and the Passions (Cornell University Press, 1995); Derek Cooke, The Language of Music (Oxford University Press, 1990). For an example of an arousal theory, see Derek Matravers, Art and Emotion (Oxford University Press, 2001). For an example of a representational theory, see Susan Langer, Philosophy in a New Key (London: Routledge and Kegan Paul, 1953). 3. Nick Zangwill, Against Emotion: Hanslick was Right about Music, The British Journal of Aesthetics 44 (2004): 2943. 4. Victor Zuckerkandl, Sound and Symbol (New York: Prometheus, 1956); Roger Scruton, Understanding Music, in his The Aesthetic Understanding (Manchester: Carcanet, 1983); Roger Scruton, The Aesthetics of Music (Oxford University Press, 1997), chs. 1, 2. 5. To some extent, this is stipulative but, as we shall see, not so that important questions are begged thereby. 6. That said, I confess that I incline toward Donald Davidsons view in his What Metaphors Mean (1978), in Inquiries into Truth and Interpretation (Oxford: Clarendon, 1982). 7. Stephen Davies, Musical Meaning and Expression (Cornell University Press, 1994), pp. 162165; see also Leo Treitler, Language and the Interpretation of Music, in Music and Meaning, ed. Jenefer Robinson (Cornell University Press, 1993). 8. If it were nevertheless insisted that all emotion metaphors applied to music are dead metaphors, then, given that we can coin fresh ones that we have never heard before, then they must die very quickly! They must die as soon as they leave our lips! 9. Treitlers example of sad news (in his Language and the Interpretation of Music) is different since, although this use is not a straightforward one, real emotions are involved in a roundabout way. Sad news is news that causes people to feel sad, or perhaps should cause them to feel sad. It is not a secondary meaning as in table leg or river mouth.

10. The dead metaphor view, like the metaphor view, incurs the onus of explaining the point of the secondary usage. 11. Scruton, The Aesthetics of Music. 12. Jerrold Levinson, Musical Expressiveness as Hearability-as-Expression, in his Contemplating Art (Oxford University Press, 2006). 13. I suspect that Levinson would be inclined to follow Davies in appealing to secondary senses of emotion words. 14. The arguments of my paper Against Emotion were directed at straightforward rather than intentional emotion theories. Those arguments may or may not carry over to intentional emotion theories. However, the arguments of Sections VIIIX of this article apply to both kinds of emotion theory. 15. I shall not defend the stronger thesis that many aesthetic properties of music must be described metaphorically, just that they very often are described metaphorically. I defend the stronger view in my Metaphysics of Beauty (Cornell University Press, 2001), ch. 10. Malcolm Budd criticizes that view in Aesthetic Realism and Emotional Qualities of Music, The British Journal of Aesthetics 45 (2005): 111122. I intend to reply to his interesting arguments in that paper in the future. 16. Nelson Goodman has an idea that may be thought to be like the one being advanced here. He says that sad music metaphorically exemplifies sadness, in his Languages of Art (Indianapolis: Bobbs-Merrill, 1968). But I find it is hard to understand what he means by the idea that metaphoricality qualifies the exemplification relation rather than uses of language. 17. For an articulation and defense of this kind of theory across many arts, see Nick Zangwill, Aesthetic Creation (Oxford University Press, 2007). 18. Eduard Hanslick, On the Musically Beautiful (Indianapolis: Hackett, 1986), p. 28. 19. See my Metaphysics of Beauty, chs. 48. 20. Peter Kivy, Music Alone (Cornell University Press, 1990), ch. 8; Peter Kivy, Experiencing the Musical Emotions, in his New Essays on Musical Understanding (Oxford University Press, 2001). 21. Jenefer Robinson, The Expression and Arousal of Emotion in Music, The Journal of Aesthetics and Art Criticism 51 (1993): 1322; see also her Deeper than Reason: Emotion and its Role in Music, Literature, and Art (Oxford University Press, 2005). 22. See Zangwill, Against Emotion, p. 36. 23. There might be some unusual cases where we are exhilarated without perceiving an exhilarating quality in the music, in the way that electronic music can be irritating, irrespective of what we perceive in it. But I think that such cases are not musically significant. Such pathological effects are usually mediated by our awareness of qualities of the music. 24. Roger Scruton and Peter Kivy have emphasized this

The Journal of Aesthetics and Art Criticism

important point. See, for example, Scruton, The Aesthetics of Music, ch. 6; Peter Kivy, The Corded Shell (Princeton University Press, 1980), ch. XI. See also Zangwill, Metaphysics of Beauty, ch. 2. 25. On substantive aesthetic properties, see the first two chapters of my Metaphysics of Beauty. 26. Evaluations range between positive and negative and can be neutral. 27. See Jerrold Levinson, Music and Negative Emotion in his Music, Art, and Metaphysics (Cornell University Press, 1990). 28. Distinguishing these two issues enables us to make sense of Hanslicks argument against the musical expression of emotion from the reuse of music to set different texts that literally and uncontroversially have emotion content. Hanslick argued that this reuse shows that there is no literal expression of emotion in music (Hanslick, On the Musically Beautiful, ch. 2). For if the music literally expressed emotion, the use of music to set emotion texts would be what we might call a rigid matter: sad music should express sad texts, proud music should express proud texts, and so on. But if music does not literally express emotion, then there is a looser relation of appropriateness between music and text. (See Zangwill, Metaphysics of Beauty, ch. 4, 5.) 29. Levinson, Music and Negative Emotion, pp. 317 318. 30. See Kivy, Music Alone, ch. 8, and Experiencing the Musical Emotions. On rereading Kivy, it occurs to me that some points that I make here overlap with Kivys work. If so, I am happy with that! 31. Emphasizing pleasure, albeit pleasure of a certain sort, as I have done, invites or naturally accompanies a quite general view about the nature and value of musicthe aesthetic theory of music. It might be objected that this view is overly essentialist about the nature and value of music. Musicologists sometimes ask: Is there really one thing that all music does, in all cultures and eras? It is common to think not. But in fact there is a lot to be said for an essentialist theory. We have a sense modality of hearing, and we take a particular kind of pleasure in hearing sounds. It is true that music does plenty of other things, too. Granted. But it does those other things in virtue of what music does essentially, which is to give a certain kind of pleasure in hearing sounds. We should allow a plurality of musical purposes that have a common core. The philosophical problem is to give an adequate account of the kind of pleasure that is in question; and this is something that Hume, Kant, and many others have pursued. 32. Many thanks to Jenefer Robinson, Kathleen Stock, and Andy Hamilton for very helpful and insightful comments. This article was given as a talk at an aesthetics conference at Manchester University in 2003, and at the University of La Trobe, the University of Tokyo, and the University of Rome.