Philosophy of art assignment

The Expression and Arousal of Emotion in Music

JENEFER ROBINSON

THIS ESSAY IS ABOUT the relation between the expression and the arousal of emotion by music. I am assuming that music frequently ex- presses emotional qualities and qualities of human personality such as sadness, nobility, aggressive- ness, tenderness, and serenity. I am also assum- ing that music frequently affects us emotionally: it evokes or arouses emotions in us. My question is whether there is any connection between these two facts, whether, in particular, music ever ex- presses emotion by virtue of arousing emotion. Of course, what it means to say that music ex- presses emotion is a contentious issue and I shall not be directly addressing it here, although what I say will have implications for any theory of mu- sical expression. Nor will I be examining all the possible contexts in which music can be said to arouse emotion. My focus in this essay will be narrower. The question I shall try to answer is this: Are the grounds on which we attribute the expression of emotion to music ever to be iden- tified with the arousal of that same emotion in listeners?

According to some theories of musical expres- sion, the grounds on which we attribute expres- sive qualities to music have nothing to do with

the arousal of emotion in the audience. Accord-ing to Peter Kivy's account in The Corded Shell, a musical element such as a melody, a rhythm, or a chord expresses a feeling not because it arouses that feeling in anyone but for two quite different reasons. (1) It has the same "contour" as expres- sive human behavior of some kind and thus is "heard as expressive of something or other be- cause heard as appropriate to the expression of something or other" (for example, the "weep- ing" figure of grief in Arianna's lament from Monteverdi's Arianna) or it contributes in a par- ticular context to the forming of such an expres- sive contour (as the diminished triad in a suitable context can contribute to a restless quality in the music, although all by itself it does not express anything). (2) The musical element is expressive by virtue of some custom or convention, which originated in connection with some expressive contour. The minor triad, for example, is "sad" by convention, although it may have started life as part of some expressive contour.

There are many examples of musical expression for which Kivy's argument is convincing. Thus it does seem to be true that Arianna's lament mir- rors the passionate speaking voice expressing grief, that Schubert's "Gretchen am Spinnrad" mirrors Gretchen's monotonous, leaden gestures

From The Journal of Aesthetics and Art Criticism 52:1 (Winter 1994). Reprinted by permission of the journal.

Jenefer Robinson: The Expression and Arousal of Emotion in Music 181

at the spinning wheel and her correspondingly dejected, leaden heart, and that the "Pleni cunt coeli" from Bach's B Minor Mass maps "bodily motion and gesture . . . of tremendous expansive- ness, vigor, violent motion," thus mirroring the exuberance of "`leaping' joy." At the same time, as Renee Cox, among others, has pointed out, virtually all the musical examples in Kivy's book are examples of music with a text, and it is rela- tively uncontroversial that a text can specify a par- ticular feeling or object which is characterized by the music. Moreover, when we look closely at Kivy's examples of particular emotions said to be expressed by music we find mainly varieties of joy, sorrow, and restlessness. The vast majority of musical examples in The Corded Shell can be char- acterized as expressions of either positive or neg- ative emotion ( joy or sorrow) of various sorts. Thus although what Kivy says seems to be true as far as it goes, it does not go very far, and leaves a great deal of expressiveness in music unexplained.

Kivy holds that music can express particular emotional states such as sorrow and joy, restlessness and serenity. Susanne Langer, while agreeing that emotional qualities are to be found in the music, rather than in the li ste n e r, follows Hanslick in arguing that since only the dynamic qualities of anything (including emotional states) can be expressed by music, no particular emo- tions can be expressed by music, but only the felt quality of our emotional life and its dynamic development:

[There] are certain aspects of the so-called "inner life"—physical or mental—which have formal properties similar to those of music— patterns of motion and rest, of tension and re- lease, of agreement and disagreement, prepara- tion, fulfillment, excitation, sudden change, etc.

[Music] reveals the rationale of feelings, the rhythm and pattern of their rise and decline and intertwining, to our minds. . . .

In contrast to Kivy's view that the words of a text supply the "fine shadings" to otherwise only grossly expressive musical meanings, Langer holds that musical meanings are inherently richand significant yet cannot be linked to any particular

words. Langer's theory emphasizes the development of structures of feeling throughout a lengthy piece of music, which Kivy ignores, but she in turn ignores the expression of particular emotional qualities which Kivy emphasizes. Both theorists have insightful things to say about musical expression but neither tells the whole story.

A very different view of musical expression has recently been presented by Kendall Walton in a paper called "What Is Abstract about the Art of Music?" Walton proposes that one important way in which music is expressive is by virtue of the fact that in listening to music we imagine ourselves introspecting, being aware of, our own feelings. As he puts it, we imagine "of our actual introspective awareness of auditory sensations" that "it is an experience of being aware of our states of mind." Thus the expressiveness of music has to do with its power to evoke certain imagi- native emotional experiences. Moreover, Walton says that if this is right, then:

music probably can be said to "portray particu- lars" in the sense that figurative paintings do, rather than simply properties or concepts. Pre- sumably the listener imagines experiencing and identifying particular stabs of pain, particular feelings of ecstasy, particular sensations of well- being, etc., as in viewing a painting one imag- ines seeing particular things.

However, whereas one perceives the psychologi- cal states of other people, as in figurative paint- ings, one "introspects one's own psychological states. "

There are at least two problems I see with Walton's account. (1) First, suppose someone denies that this is what she does when listening to expressive music; we should be able to explain to her why this is what she should be doing. What reason is there why we should imagine our awareness of auditory sensations—experienced sequences of musical tones—to be an experience of our feelings and other inner states? True, there are similarities between the two: the experience of auditory sensations is an introspectible state,

182 PART TWO: THE ANALYTIC CRITIQUE OF THE AESTHETIC

and so is awareness of our feelings. True, part of what we are aware of in these auditory sensations is, as Langer points out, their ebb and flow, and our feelings too have ebb and flow. But beyond these points of resemblance there seems to be little explanation why we should be inclined to imagine our awareness of musical sounds to be an awareness of our feelings. Imagination re- quires some guidance if it is not to be merely free association: I can imagine the tree at the end of the garden to be a witch because it has a witch- like appearance, but it is unclear what it would mean for me to imagine the snowdrop at my feet to be a witch if there is nothing about the snow- drop to set off my imagination. Similarly, in order for me to imagine my awareness of musical sounds to be awareness of my feelings, some- t h i n g i n t h e mu s i c a l s o u n d s mu s t g u i d e m y imagination. However, if the only points of re- semblance between feelings and sounds is intro- sp ectibility and ebb and flow, th en I would suggest that this is insufficient to ground an imaginative identification between the two. There are, moreover, striking differences between the two which would seem to preclude any such imaginative identification. In particular, whereas our feelings clearly rise up inside us (as we say), musical sounds as clearly rise up at a distance from us: even when listening to music over good earphones—when the music is experienced with peculiar immediacy—we still experience the au- ditory sensations as coming from an external source, such as trombones and the like. That is why although we can perhaps imagine these sounds as feelings welling up inside the composer, or perhaps in some character described by the music, it is not obvious to me that we can imag- ine them as feelings welling up inside ourselves.

(2) There is a second problem related to this one. I am willing to grant that there are indeed movements in music which it is appropriate to call "stabbing" or "surging." According to Wal- ton, however, the music induces me to imagine myself feeling a particular ecstatic surge or stab of pain. He says that the music portrays these par- ticulars (it picks them out or refers to them). Anumber of questions need to be

distinguished here. (1) Can the stab be identified as a stab of feeling rather than the stab of a dagger or some other kind of stab? (2) If the stab is a stab of feeling, can it be identified as a stab of pain rather than some other feeling such as excitement or jealousy? (3) If the stab is indeed a stab of pain, can it be identified as a stab of pain which I imagine myself experiencing rather than a stab of pain attributed to someone else, such as Othello or the composer? If the music portrays my imagined stab of pain, as Walton suggests, then the music must be able to distinguish my imagined stab of pain from all these other possible alternatives. Can music do this? Can music portray this particular stab of pain and no other?

If the music were accompanied by an appro- priate verbal text, then perhaps it could. As we listen to the music we hear in it particular tones, rhythms, harmonic modulations, phrases, melo- dies, counterpoint sections, etc., etc. We might also hear particular movements that we charac- terize as "stabbing" or "surging." Given a par- ticular accompanying text, we might then be able to identify the stabbing as the stabbing of Mer- cutio rather than a stab of pain. With a different text, we might be able to identify the stabbing as Othello's stabbing pains of jealousy rather than his stabbing pains of remorse. And so on. How- ever, in the absence of a text, Walton suggests no good reasons for identifying the stabbing in the music (1) with a stab of feeling (rather than some other kind of stab), (2) with a stab of pain (rather than some other kind of feeling), or (3) with my imagined stab of pain (rather than yours or Othello's real or imagined pain). Walton claims earlier in his paper that musical character- ization is inherently general. It would seem to follow that without the specification of a particu- lar context we cannot specify that the stab is even a stab of emotion, let alone the stab of a particu- lar emotion, let alone an imagined stab of my particular emotion. However, Walton gives us no guidance as to how a particular context could be specified. In short, although we can hear a stab- bing movement in a piece of music, Walton does

Jenefer Robinson: The Expression and Arousal of Emotion in Music 183

not show us how to tell from the music alone— without any accompanying text—what, if any, particular stabbing is occurring.

Although Walton's theory does not identify musical expression with the straightforward arousal of feelings, he does try to explain expres- sion in terms of the arousal of imaginary feel- ings. I am not actually feeling a stab of pain as I listen to the stabbing music; I am imagining ex- periencing a stab of pain, so it would seem that the pain is an imaginary feeling. In his paper "Music and Negative Emotions," Jerrold Levin- son makes a similar point. Levinson's paper deals with the problem of why people enjoy music when it evokes negative emotions such as sadness in them. While the paper does not develop a the- ory of musical expression, it does make certain assumptions about what often happens when people listen to music which we would charac- terize as sad. In particular, he assumes that it is a normal response for people to have a sadness- reaction to music.

W hen a per son h as a "deep emotional re- sponse" to music, this is "generally in virtue of the recognition of emotions expressed in music," but recognition then leads to a kind of empathic identification: we "end up feeling as, in imagina- tion, the music does." Such empathic emotional responses to music consist in "something very like experience of the emotion expressed in the music" but not exactly like It. In both cases the physiological and affective components of emo- tion are present and in both cases there is cogni- tive content, but the "empathic" response lacks determinate cognitive content:

When one hears sad music, begins to feel sad, and imagines that one is actually sad, one must, according to the logic of the concept, be imag- ining that there is an object for one's sadness and that one maintains certain evaluative beliefs (or attitudes) regarding it. The point, though, is that this latter imagining generally remains indeterminate.

I feel sad but my sadness has no determinate ob- ject; it is directed only to "some featureless ob- ject posited vaguely by my imagination."

Levinson illustrates his view with various kinds of negative emotion: "intense grief, unrequited passion, sobbing melancholy, tragic resolve, and angry despair." Suppose, for example, that the music evokes in me an empathic response of unrequited passion. On Levinson's view, this means that I recognize unrequited passion in the music, I imagine that I am experiencing unrequited passion, and I actually experience the physiological and affective components, of unrequited passion. My imagined unrequited passion has a cognitive content which is "etiolated by comparison to that of real-life emotion"; however, since I am not really suffering the pangs of unrequited pas- sion, and in particular there is no special person for whom I am languishing.

I am sympathetic to some of Levinson's as- sumptions: I think he is right to stress that the detection of emotional qualities in music has something to do with the arousal of emotion by music, and I think he is right also to stress the role of the imagination in the appreciation of emotional qualities in music. However, the the- ory as it stands will not do. First of all, it is far from clear that every emotional state has identifi- able physiological and affective components. For example, real-life unrequited passion might on different occasions be accompanied by a great va- riety of inner feelings (love, grief, longing, jeal- ousy, wretchedness, despair, self-contempt, etc., etc.). For another thing, the particular feelings I experience on a given occasion of unrequited passion may be just the same as I have felt on oc- casions of angry despair or intense grief. The truth of the matter is that there may be very little difference between the affective and physiological components of very different emotions: I may feel the same mixture of grief and rage when I am jealous or when I am grieving (without jealousy), I may have very similar feelings whether angrily despairing, tragically resolving, or suffering from the pangs of unrequited passion. The difference between these emotions lies not so much in their affective and physiological components as in their cognitive content. The chief

184 PART TWO: THE ANALYTIC CRITIQUE OF THE AESTHETIC

difference between unrequited passion, tragic re- solve, and angry despair is how I view or con- ceive of the situation.

But now we come to a second set of difficul- ties. Levinson argues that I can recognize unre- q u i t e d p a s s i o n ( s a y) i n t h e " e mo t i o n - l a d e n gestures embodied in musical movement" and by virtue of this recognition respond empathically with feelings of unrequited passion of my own, since I identify with the music or perhaps "with the person whom we imagine owns the emotions or emotional gestures we hear in the music." However, he fails to tell us how we detect or em- pathically feel the unrequited passion in the music. Although we all have some idea of what sad music is like, I suggest that it is much less clear what a piece of music is like in which we can r e c o g n i z e , a n d h e n c e e mp a th iz e w i th , u n r e - quited passion (always assuming, of course, that there is no accompanying verbal text to help us out). If I am right and there are no distinctive af- fective or physiological components of unre- quited passion, then the obvious way to clarify the nature of music in which we can detect unre- quited passion would be to specify its cognitive content. Now, Levinson claims that the cognitive content of an emotional response to music is normally "etiolated." This could mean simply that my imagined feelings of unrequited passion are not directed to any particular individual. While it is a little odd to say that one can feel un- requited passion for someone I know not whom, we can perhaps make sense of this suggestion since on Levinson's view the unrequited passion I feel empathically belongs to the music itself or to someone whom we imagine feels unrequited passion, so that we merely empathize with this imagined person's unrequited passion.

Even if we grant, however, that there need be no specific object for the unrequited passion I de- tect in the music and empathize with, it would seem that there must be some identifiable cog- nitive content, however etiolated, which is de- t e c tab le in the mu si c in order to j usti f y th e attribution of this particular emotion. I would suggest that if my response is to count as a re- sponse of unrequited passion rather than some

other emotion, then I must imagine that there is someone whom I care about deeply, that this per- son does not care deeply about me, and that I care deeply that this person does not care deeply about me (or something of this sort). It is a seri- ous problem for Levinson's account that he does not tell us how such conceptions can be em- bodied in music and hence how we can either recognize or empathize with the corresponding emotion. We find the same problem with tragic resolve and angry despair: we cannot clearly dis- tinguish these emotional responses by their affec- tive and physiological components alone, but only by their cognitive content. However, Levinson gives us no clue as to how their cognitive content can be recognized in or induced by music.

In a la ter paper, "Hope in T h e H e b r i d e s , " Levinson claims that perhaps it is possible for music to express "higher" emotional states, and that in addition to the affective and physiological components of an emotion, music might even be able to convey part, at least, of its cognitive con- tent. He points out that emotions are normally intentional but that music can convey a general "sense of intentionality (aboutness)." He also notes that just as ordinary extramusical emotions are often individuated by their context of occur- rence, it might perhaps be the case that musical context can play a similar role for emotions in music. When he illustrates his thesis by reference to the emotional state of hope, which he claims to be able to distinguish in Mendelssohn's He- brides overture, he remarks that "perhaps some of the pure conceptual content of hope—its fa- vorable assessment of future in relation to pres- ent"—can be suggested by the position of the hopeful passage in its musical context. Levinson does not develop this idea very far, however, and what he does say along these lines is very tenta- tive. Certainly he gives us no clue as to how the three marks of unrequited passion that I distin- guished above could be adequately conveyed by music.

Recently Levinson's view has been criticized by Peter Kivy on the grounds that the expression

Jenefer Robinson: The Expression and Arousal of Emotion in Music 185

of emotion in music is entirely independent of the arousal of that emotion. Kivy argues that to h a v e o n e ' s e m o t i o n s a r o u s e d b y a p i e c e o f music—in particular, to be moved by a piece of music—is quite distinct from perceiving a partic- ular emotional quality in that piece. Music that is sad or expresses sadness is music with a sad ex- pressive contour or music that is sad by conven- tion, not music that arouses or evokes sadness. L e v i n s o n a r g u e s t h a t a " d e e p e mo t i o n a l r e - sponse" to sad music consists in the arousal of a kind of imaginative but cognitively truncated sadness. Kivy rightly attacks this claim, arguing on the one hand that sad music may or may not make me feel anything, depending on how great the music is (the "yards and yards of mournful music" written by Telemann may fail to make me feel anything much at all), and on the other hand that there are important emotions aroused by music which are full-blown, ordinary, real-life emotions, not "truncated" or "imaginary" in any sense. He illustrates his point by reference to a performance of Josquin's "Ave verum virginitas" which, he says moves him deeply.

When listening to the "Ave verum virginitas" I may simply be moved by "the sheer beauty of the sound as it unfolds in its ebb and flow." If my sophistication increases, however, I may also be moved by "the incomparable beauty and crafts- manship of Josquin's counterpoint" and by the fact that despite its seeming effortlessness, the music is written in a particularly difficult canonic form, "a canon at the fifth, with the voices only one beat apart." This, then, is the cognitive com- ponent of the emotion aroused by the music, my being moved by the music. It is not a truncated emotion in any way. It is a genuine emotional ex- perience, arising out of my perception of the music and its qualities. Furthermore, this emo- tion might be directed at emotional, expressive qualities in the music, such as sadness, but it does not follow that the emotion aroused by the music is the emotion detected in the music. Part of what I may be moved by in a piece of music may be its sadness, but I can be moved by joyful, by ener- getic, and by serene music just as well, as well asby music which does not have any marked emotional

character. The expressive qualities, if any, which I detect in the music are entirely independent of the emotions I feel as I listen to the music.

Now, Kivy is certainly right to claim that when I am moved by a piece of music, my emo- tion may be independent of the emotional quali- ties, if any, that the music happens to have. When I appreciate a piece of music I may indeed be moved in the way Kivy describes. On the other hand, Kivy has not succeeded in showing that the expression of emotion by a piece of music is always and entirely unconnected to the arousal of emotion. Kivy makes this claim based on an analysis of just one emotion, "being moved," and it may well be true that we can be equally moved by music with different emotional quali- ties, as well as by music which has no marked emotional qualities. However, I believe that music arouses other feelings as well and that some of these may indeed be connected to the expressive qualities that music has. Furthermore, I think Kivy is wrong to insist that all the feel- ings aroused by music have to have a complex cognitive component as in his example from Josquin. It may be true that being moved by music involves complex evaluative judgments, but being moved is not the only emotional or feeling response which music can arouse.

Let me summarize the results of my discus- sion so far. Walton argues that expressive music evokes the imaginative experience of the emo- tion expressed: more precisely, music expressive of sadness, say, induces the listener to imagine herself experiencing sad feelings. Levinson simi- larly claims that sad music has the power to evoke a kind of truncated sadness-response: the listener feels certain symptoms of sadness, has an "inde- terminate" idea that there is something or other to be sad about and imagines that she in fact feels sad. Both writers find a connection between the presence of an emotional quality in music and the arousal of that emotion in the listener's imagination. I have urged, however, that neither Walton nor Levinson has shown how complex

186 PART TWO: THE ANALYTIC CRITIQUE OF THE AESTHETIC

feelings such as unrequited passion, stabs of pain, or even sadness can be aroused by music whether in fact or in imagination. Furthermore, Kivy is clearly right to hold that to have a deep emo- tional response to music is not necessarily to mirror the feelings that the music expresses.

At the same time, however, I believe that Walton and Levinson are right to stress the connection between the expression and the arousal of emotion in music, and that Kivy is quite wrong to think that his analysis of the one emotion "being moved" demonstrates that no such connection exists. In what remains of this paper I will try to sketch a more adequate account of what this connection really is.

None of the writers I have discussed in this essay has focused on the way in which music can directly affect our feelings. For both Walton and Levinson the arousal of feeling is imaginative and it relies on a good deal of cognitive activity on the part of the listener. For Kivy the emotion of being moved is a real emotion, not an imagined one, but it too relies on cognitive activity, such as rec- ognizing the clever partwriting, etc. However, some music has the power to affect our feelings without much, if any cognitive mediation. In par- ticular, music can induce physiological changes and a certain quality of inner feeling (what Levin- son calls respectively the "phenomenological" and "sensational" aspects of the "affective" com- ponent in emotion). Music can make me feel tense or relaxed; it can disturb, unsettle, and startle me; it can calm me down or excite me; it can get me tapping my foot, singing along, or dancing; it can maybe lift my spirits and mellow me out.

Emotions vary in degree—and perhaps in kind—of cognitive content. At one end of the scale there is the startle response, which is an in- nate response, found in human neonates as well as throughout the phylogenetic scale. At the other end of the scale there is unrequited passion which, by contrast, is found only in humans with their highly developed cultural norms. What I want to suggest is that in addition to the sophis- ticated emotions of appreciation, which Kivyidentifies as "being moved" by certain

perceived aspects of the music, there are more primitive emotions aroused by music, perhaps requiring less developed cognitive mediation. There are, after all, moments in music which make us jump or startle us. Similarly, the perception of certain rhythms may be enough— without any further cognitive mediation—to evoke tension or relaxation, excitement or calm. If the melodic and harmonic elements in a piece of music affect our emotions, this would seem to require familiarity with the stylistic norms of the piece, but no further cognitions need be required in order for us to feel soothed, unsettled, surprised, or excited by developments in the music. Certainly we need not notice that we are listening to a canon at the fifth in order for that canon to soothe us.

We have seen that to feel unrequited passion necessarily involves a certain fairly complicated conception of one's situation. By contrast, to feel disturbed or calm does not require having a con- ception of one's situation in this way. Music can make me feel disturbed or calm just by perceiving it (listening to it). The feeling is a result of a per- ception and to this extent it has "cognitive con- ten t," bu t it is not the full-blown cogn itiv e content required for tragic resolve, angry despair or unrequited passion. The sense of relaxation we feel at the end of "Tristan and Isolde," for ex- ample, is the result of the long-awaited resolution, after over four hours of constant modulation without resolution. The feeling is the result of a perception, but we may not even be aware why we feel as we do: the effect of the constantly shifting harmonic pattern affects us "directly" without conscious cognitive mediation (except, of course, what is required by our understanding of Wag- ner's style). There is some psychological evidence (from Berlyne and others) that people seek high levels of arousal in order to have them drop after- wards: "excitement and complex, conflicting in- formation are sought because of the 'arousal jag." The effect of the final Tristan chord may be partly accounted for in these terms.

Now, the feelings evoked "directly" by music explain some of the cases of musical expressive-

Jenefer Robinson: The Expression and Arousal of Emotion in Music 187

ness that the contour theory finds hard to deal with. Music that disturbs and unsettles us is dis- turbing, unsettling music. Modulations that sur- prise us are surprising. Melodies that soothe us are sooth ing. Furthermo re, unexp ected har- monic shifts excite us and are exciting; protracted stay in a harmonic area distant from the home key makes us uneasy and produces uneasy music; the return to the home key after a protracted stay in a distant harmonic area relaxes the tension in us and produces relaxing music. And so on. In short, as against Kivy's position, it seems to me that the expression of a feeling by music can some- times be explained straightforwardly in terms of the arousal of that feeling. However, the feel- ings aroused "directly" by music are not stabs of pain or feelings of unrequited passion, but more "primitive" feelings of tension, relaxation, sur- prise, and so on. These feelings do, therefore, in a sense have an "etiolated" cognitive content, in the way that Levinson specifies in "Music and Negative Emotions," but it is not an etiolated, imaginary version of an emotion which normally has a complex cognitive content (such as unre- quited passion), but rather a feeling such as sur- p r ise , wh ich b y it s n a tu re j u s t h as—o r can have—a relatively simple cognitive content.

Even more interesting, however, is the way in which the simple feelings "directly" aroused by music can contribute to the imaginative expres- sion of more complex emotions such as those discussed by Levinson. When we listen to a piece of music in a relatively familiar style, a succession of feelings is aroused in us: in a pattern typical of Classical sonata form, we may first be made to feel relaxed, then jolted into uncertainty, then made to feel uneasy for a prolonged period be- fore experiencing relief and final release of ten- sion. Now, something that most philosophical theorists of musical expression have either ig- nored or underemphasized is the fact that the musical expression of complex emotions is not a function of a few isolated measures here and there, as in Kivy's examples in The Corded Shell; rather it is very often a function of the large-scale formal structures of the piece as a whole. We

cannot understand the expression of complex emotions in music apart from the continuous development of the music itself None of the philosophical writers I have discussed has fully appreciated this point. Langer has indeed stressed the importance of large-scale movements of ebbing and flowing, tension and relaxation in musical expression, but she denies that any particular emotions can thereby be expressed. Levinson suggests at times that we need to look at the total musical context before we can say what particular emotions are being expressed. But he does not explore this idea very far. In order to explain how particular cognitively complex emotions can be expressed musically, we need to look at the overall structure of a piece and at the feelings aroused by the piece as it develops in time.

In his celebrated book, Emotion and Meaning in Music, Leonard Meyer showed how the formal structure of works in the Classical and Romantic styles could be analyzed in terms of the emotional responses of the practiced listener: his was a kind of "Reader-Response" or rather "Listener-Response" theory of musical structure. In order to understand a piece of music, on this view, the listener has to have her feelings aroused in a certain way. If we are experienced in the style of the piece, then we have certain expectations about the way the music will develop; in a meaningful piece of music these expectations will be either frustrated or satisfied in unexpected ways. As we listen new expectations are constantly being aroused and we are just as constantly being surprised by novel developments, relieved by delayed resolutions, made tense by the delays, etc., etc. In short, understanding musical structure, according to Meyer, is not just a matter of detached analysis; rather, it is impossible without the arousal of feeling in the listener.

Now, just as the formal structure of a piece of music can be understood in terms of the arousal of such feelings as uncertainty, uneasiness, relax- ation, tension, relief, etc., so too can we understand the expressiveness of that piece of music in terms of the arousal of those and similar

188 PART TWO: THE ANALYTIC CRITIQUE OF THE AESTHETIC

feelings. After all, as Anthony Newcomb has put it: "Formal and expressive interpretations are in fact two complementary ways of understanding the same phenomena." Emotional expressiveness in music frequently corresponds to or mirrors its formal structure. The "direct" arousal of cognitively " s i mp l e " e mo t i o n s s u c h a s b e i n g ma d e s u r prised, disturbed, satisfied, relaxed, etc. is a clue not only to the formal structure of a musical piece, as Meyer showed, but also to its structure of emotional expressiveness. If a piece of music is heard as successively disturbing and reassuring, or as meandering uncertainly before moving forward confidently, or as full of obstacles which are with difficulty overcome, this is at least in part because of the way the music makes us feel. Disturbing passages disturb us: reassuring ones reassure. Passages that meander uncertainly make us feel uneasy: it is not clear where the music is going. Passages that move forward confidently make us feel satisfied: we know what is happening and seem to be able to predict what will happen next. Passages that are full of obstacles make us feel tense and when the obstacles are overcome, we feel relieved. It is important to notice that the feeling expressed is not always the feeling a r o u s e d : an uncertain, d iffid en t p assage may make me uneasy; a confident passage may make me feel reassured or relaxed.

Now, of course we are still a long way from show in g how unrequited passion can be ex- pressed by a piece of music, but we can perhaps begin to see how the development of a complex piece of music can mirror the development of a complex emotional experience, and how we can become aware of both the formal development and the corresponding emotional development by means of the relatively "simple" feelings that are aroused in the listener as she follows that de- velopment. As I listen to a piece which expresses serenity tinged with doubt, I myself do not have to feel serenity tinged with doubt, but the feel- ings I do experience, such as relaxation or reas- surance, interspersed with uneasiness, alert me to the nature of the overall emotional expressiveness in the piece of music as a whole. Consider, for

example a piece of music in sonata form in which the two chief themes in their initial formulation are respectively lively and ponderous (we can suppose that the contour theory accounts for these characterizations). Now, suppose that the initially lively theme (in the major) gets gradually but relentlessly overwhelmed by the ponderous (minor) theme in such a way that the first theme is never allowed to return to its initial lively for- mulation but gets increasingly distorted, be- comes darker and is finally heard in a truncated form in the same minor key as the ponderous theme. Such music might well make me feel in- creasingly nervous and tense, even disturbed, as it develops. On the view I am suggesting, the emotional experience aroused by the music is es- sential to the detection of the emotional expres- siveness in the music itself. At the same time, the emotions aroused in me are not the emotions ex- pressed by the music. I feel nervous, tense, and disturbed: the m u s i c expresses cheerful confi- dence turned to despair, or something of this sort. If this account is correct, then it shows that Kivy is wrong to suppose that expressiveness in music is just a matter of contour and convention, even if some expressive passages in music can be explained in such terms. In my example, it is not enough to spot the respective lively and ponder- ous contours of the initial statements of the two themes; the expressiveness of the piece as a whole can only be grasped if the listener's feelings are aroused in such a way that they provide a clue to both the formal and the expressive structure of the piece as it develops through time.

0

r

We can now see that Levinson and Walton are right to insist on a connection between the arousal and expression of emotion in music. How- ever, neither of them has succeeded in showing how music can actually arouse, even in imagina- tion, the complex emotional states that music sometimes expresses. In my example, I did not myself have to feel cheerful confidence turning to despair in order to detect that emotion in the music. The feelings I felt, which were evoked "directly" by the music, were less cognitively complicated, such as unease, tension, and

Jenefer Robinson: The Expression and Arousal of Emotion in Music 189

disturbance. At the same time, we can see why Levinson is tempted to say that we empathize imaginatively with the feelings expressed by the music, for in order to detect these feelings in the music I am myself emotionally involved in listening to the music: I feel genuine feelings of unease, disturbance, and so on. Moreover, if I imagine that the themes are themselves characters in a kind of musical drama, then perhaps I can empathize with the fate of the lively theme, feeling sorrow and pity for it as I might for a character in a drama, and maybe I can even feel anger and frustration at the ponderous theme.

Walton wants to say that I imagine of my in- trospective awareness of auditory sensations that they are an experience of particular states of my own psyche, such as particular stabs of pain. Again we can see why Walton is tempted by this i d e a , s i n c e o n t h e o n e h a n d t h e mu s i c d o e s arouse feelings in me, although not usually the ones expressed, and on the other hand I may per- haps imagine that the feelings expressed by the music do belong to me. However, I think this view is more problematic than Levinson's. In my example, must I imagine of both themes that they are an experience of my own emotions? In this case my imagination must take both sides in

the conflict as it were. Why cannot I identify en- tirely with the suffering lively theme, or even gloatingly—entirely with the powerfully insistent ponderous theme? Why, more fundamentally, should I imagine these musical events as belong- ing to my own psyche at all? When I watch a per- formance of King Lear I do not imagine the drama to be taking place inside my own head; it seems to me that the same is just as true of the King Lear overture.

In this essay I have tried to confine my atten- tion to the question of how the expression of emotion by music is related to the arousal of emotion in the listener. Obviously I have left many questions unanswered. In particular, I have given only a skeletal account of how music can express cognitively complex emotions such as the "cheerful confidence turning to despair" of my example. I have not attempted to show how cog- nitive content can get expressed by music nor whether particular emotions such as unrequited passion can be so expressed. What I have tried to do, however, is to indicate how such analyses might proceed. And the point I have urged above all is that any such analysis must begin with the emotions that are aroused by the music in the listener.