Many theorists claim that to say music is expressive of emotion is simply to attribute to the music “expressive qualities.” Others claim that music can be an expression of emotion in a more full-blooded way. In this chapter I will be defending the idea that at least some music can be a genuine expression of emotion in the sense that it can be a manifestation of emotion that someone (although perhaps a fictional someone) actually feels. I will not be talking directly about the emotions music arouses in listeners, although what the music arouses and what it expresses, if anything, are clearly connected. And I will not be arguing that all music expresses emotions. The mature compositions of Milton Babbitt, for example, exhibit little interest in emotion. My discussion will be focused on Western art music that is clearly emotionally expressive, most notably, music in the Romantic and post-Romantic style.
For many people, to say that a piece of music “expresses sadness” simply means that the music has a certain quality that is named by an emotion word: the music “is sad.” (See, for example, John Hospers 1955; Tormey 1971.) Expression in this view is simply a matter of possessing expressive qualities, and expressive qualities are simply “aesthetic qualities” like any others, such as dynamism or freshness. But music can be sad by virtue of conventions (it is in the minor key) or cultural associations (it is used at funerals) without expressing much, if any, emotion. Like the upside-down smiley-face, music can be sad without being very expressive.
According to Stephen Davies, the expressiveness of music consists in its “presenting emotion characteristics in its appearance” (1994: 228). Just as the face of a basset hound is called “sad” because that is the way sad people typically look when they are expressing their sadness, so music is called “sad” because it sounds or moves like a person who is sad. Music is expressive of sadness without being an expression of anyone’s sadness, that is, without revealing anything about anyone’s actual state of mind. Similarly, in The Corded Shell (1980), Peter Kivy argues that music is expressive of emotion by virtue of sharing the “contours” of vocal or behavioral expressive gestures made by human beings when in the throes of emotion. Like Davies, Kivy compares musical expressiveness to the expressiveness of a dog’s face, in his case the St. Bernard. (Both Kivy and Davies also recognize the role of conventions in musical expressiveness. See also Kivy 2002, which partially repudiates his earlier view.)
This “doggy” theory of musical expressiveness emphasizes how a musical line can be heard as expressive of grief by virtue of its resemblance to the “contour” or intonation pattern of a grief-stricken voice, as in the famous “weeping figure” at the beginning of Monteverdi’s Arianna’s Lament, or by virtue of how musical movement mimics expressive behavior, especially “the gait, attitude, air, carriage, posture, and comportment of the human body” (Davies 2006: 182). For Davies, “the resemblance that counts most for musical expressiveness ...is that between music’s temporally unfolding dynamic structure and configurations of human behavior associated with emotion” (2006: 181). We experience movement in music not only in terms of “progress from high to low or fast to slow,” but also in “the multistranded waxing and waning of tensions generated variously within the harmony, the mode of articulation and phrasing, subtle nuances of timing, the delay or defeat of expected continuations, and so on” (2006: 181–2). Davies thinks that “this movement is like human behavior in that it seems purposeful and goal-directed” (2006: 182).
To those who object that there is no greater “objective” resemblance between musical movement and emotions than between musical movement and various natural phenomena – the weather, the moods of the sea – Davies responds that the degree of resemblance is beside the point: listeners simply do experience a resemblance between the music and “the realm of human emotion.” Listeners make the connection between music and emotion by an “experience of similarity” (2006: 182), not a mere recognition that there is a similarity. And our interests shape how we experience the world. As he says, we are more likely to see a weeping willow as a downcast person than as a frozen waterfall, even if the similarity between the willow and the waterfall is no less than that between the willow and the droopy person. We hear music as expressive of emotions because in listening to music, we anthropomorphize or “animate” it so that we hear it as expressive of emotion.
One limitation of the doggy theory is that it allows for music to express only those emotional states that exhibit characteristic vocal intonations or expressive behaviors. This has three important consequences. First, it is hard to see how music can express patterns of feeling, such as the way in which despair is with difficulty overcome and transforms gradually into resignation. Second, and relatedly, it seems to follow that cognitively complex emotions cannot be expressed by music: there are no distinguishing vocal or behavioral marks of resignation, for example. Third, the theory does not explain why listeners are so powerfully moved by emotional expression in music. We are not particularly moved (except perhaps to laughter) by the sad doggy faces of the St. Bernard and the basset hound. Why, then, should we be moved by the sad appearance of music?
Davies has responded to all three objections. First, he has argued that a pattern of feeling can be expressed by an appropriate sequence of musical gestures. Thus, “just as music might present the characteristic of an emotion in its aural appearance, so too it might present the appearance of a pattern of feelings through the order of its expressive development” (Davies 1994: 263). But if what we are listening to is a sequence of expressive “contours” without any underlying psychological reality, there is no organic connection between one expressive “appearance” and the next: they are simply concatenated. It is like watching a series of expressions moving across someone’s face. If there is a pattern, it is only because of the thoughts, desires, intentions and so on that underlie the sequence. If it is just a series of facial contortions, why call this a pattern of expressions?
Second, Davies has defended the idea that music can express cognitively complex emotions, arguing that a piece of music can express hope, for example, if the “emotion characteristics in appearance” of a longish piece or passage of music are judiciously ordered (1994: 262–4). But again, a mere sequence of expressive gestures is not enough to distinguish a cognitively complex emotion such as hope, whatever the order in which these gestures occur. If all you have to work with are expressive gestures, then the best you can do to express hope in music is to have a cheerful passage followed by a sad one or a passage in which cheerfulness and sadness somehow intermingle or something of this sort. But the expression of hope requires the expression of desires and thoughts. A hopeful person is one who wishes for something to happen that he construes as good. Hope cannot be expressed merely by a succession of bodily gestures and vocal intonations. (See Karl and Robinson 1995 for a detailed discussion of this point.)
More recently, Davies has conceded that only a few emotional types “can be individuated solely on the basis of observed bodily comportment” (2006: 183). His candidates for expressible emotions include sadness and happiness, timidity, anger, “swaggering arrogance, the mechanical rigidity that goes with repression and alienation from the physicality of existence, ethereal dreaminess, and sassy sexuality” (2006: 183). Notice, however, that apart from sadness, happiness and anger, the rest of these examples are not strictly speaking emotions at all, but rather behaviors that could but need not be indicative of some emotion. As for more complex emotions, Davies is cautious: “where deep sadness gives way gradually to joy and abandonment, it may be reasonable to regard the transition as consistent with acceptance and resolution” (2006: 185). But notice here that “acceptance” and “resolution” are inner states, requiring beliefs, desires, and intentions. It is implausible that the transformation of a deeply sad appearance (such as a grieving facial expression) into a joyful appearance (such as a smile) is capable of expressing a complex shift in one’s inner states, involving thoughts of acceptance, an intention to be courageous, a wish that things had been different conquered by a desire for the capacity to deal with things as they are, and so on. In general, if all musical expression could be explained according to the doggy theory, then music would be able to express very little about our inner life.
The final problem concerns why expressive music should be moving, if the doggy theory is correct. Here Davies relies on the idea that expressive music is “contagious” (1994: 279–307, forthcoming). There is indeed evidence that music can affect the motor system and to some degree change people’s behavior and mood (see Robinson 2005: ch. 13). But we are not typically moved by an expression of emotion in a musical “appearance” in the way in which we are moved by an expression of genuine emotion. Even if I am affected physiologically and motorically by a piece of expressive music, this does not explain the power of our emotional responses to expressive music. After all, I am powerfully moved not because my friend has a sad-looking face, but only because that sad-looking face is a sign that she really is sad. In Bill Viola’s slow-motion video installation, The Quintet of Remembrance, five actors perform different emotions (sadness, anger, and so on) via gradually changing facial expressions and gestures. The people in the group do not appear to interact, and there is no hint as to why they are expressing these emotions. The result is that the piece is both lifeless and melodramatic. Yet this installation is supposed to get its expressiveness in just the same way as the doggy theory claims music does.
Jerrold Levinson propounds a variation of the “animation theory,” which, unlike the doggy theory, accepts that what we experience as musical expressiveness is an experience as of someone genuinely expressing his or her emotions. In Levinson’s formulation, “a passage of music P is expressive of an emotion E if and only if P, in context, is readily heard, by a listener experienced in the genre in question, as an expression of E” (Levinson 2006: 193; see also Levinson 1996). It is crucial to Levinson’s view that expression “requires an expresser” (Levinson 2006: 193). He believes that when we hear music as expressive of emotion, we hear or imagine an agent or persona in the music, the “owner” of the states expressed. Now, when we listen to a lyric song such as “Gute Nacht” from Schubert’s Winterreise, we naturally hear it as emanating from a person or character in the music who is expressing his gloomy state of mind. In Levinson’s view, however, we also hear all purely instrumental music (“absolute music”) that is expressive of emotion in the very same way, namely, as emanating from a persona in the music, which may be a “character,” or the composer himself, or a persona of the composer.
There is much to be said in favor of Levinson’s view. It allows for musical expressiveness to be treated as the genuine expression of emotion. It permits us to hear extended passages of music as expressing unfolding psychological states, rather than as mere sequences of expressive “appearances.” And once we hear the music as genuinely expressing a sequence of emotions, it is possible to find “patterns of feeling” in the music as well as the expression of cognitively complex emotions such as hope. If we hear a persona in the music, we can hear him as seeking or striving toward certain goals (as fragments of a theme struggle to transform into another theme with a different character), as desiring certain things and rejecting others (as a sequence of harmonies yearns toward resolution but is turned aside into an alien key which it then struggles to resist), or as remembering past events with nostalgia or bitterness (as when an early sunny theme is recalled later in a piece with reassuring or troubling effect). Emotion characteristics in appearances do not strive or seek or desire or remember, but people do. Through positing a persona in the music, Levinson allows us to hear the music as expressing the inner states of this persona. Finally, because it allows us to hear the music as a genuine expression, it makes sense that we would be moved by music’s expressiveness. (See Karl and Robinson (1995) for a case study of Shostakovich’s Tenth Symphony. For a recent study that emphasizes how the listener not only hears what the music expresses but also enacts virtual expressive behaviors afforded by the music, see Nussbaum 2007.)
Despite its many virtues, however, there are problems with Levinson’s theory: in some respects it goes too far and in other respects it does not go far enough. First, Levinson means his theory to be a general account of expressiveness in music. But there are many pieces which in common parlance are said to “express melancholy” even though we have no inclination to posit a melancholy persona in the music. As we have seen, a piece can be “sad” or “cheerful” for diverse reasons: associations or conventions may play the major role. Other pieces can be explained simply by reference to the doggy theory: we hear a piece as sad because of its sad “contours.” Perhaps we should stipulate that the term “musical expression” should be confined to those pieces that fit Levinson’s theory, but then we need to know how to determine which those are.
This brings me to my second objection to Levinson’s theory: in some respects it does not go far enough. For Levinson, like Kivy and Davies, expression in music is primarily something determined by the experience of listeners or audiences, not primarily something achieved by artists. Now, it is true that emotional expression in ordinary life is a means of communication – looking at your gait and posture tells me how you are feeling – but it is also true that the reason why expression is such a good means of communication is that, when it is sincere, it accurately reveals genuine inner states. In other words, expression is primarily something achieved by expressers, not something noticed or experienced by spectators or audiences.
In conclusion, there is much expressiveness that does not need Levinson’s persona, and there is some expressiveness that does require the persona but as a genuine (dramatic) protagonist genuinely expressing his or her emotions, not merely as something imagined or postulated by listeners. (For further discussion of Levinson on expression and expressiveness see Robinson 2007b.)
The Romantic movement at the end of the eighteenth and beginning of the nineteenth centuries spawned the idea that one of the main goals of the arts is to express the emotions of artists. One of the most carefully worked out versions of the expression theory comes from the philosopher R. G. Collingwood (1963). Collingwood claims that all “art proper” is expression. But I will treat his view as a theory of expression in art, not a theory of art in general. According to Collingwood, in both the expression of emotion in ordinary life and the expression of emotion in artworks someone who is in an emotional state communicates that state to other people. But artistic expression also differs from what we call expression in ordinary contexts in at least three ways.
First, to “express” an emotion in real life means that you manifest or show this emotion by means of facial or vocal expressions, by the visible concomitants of autonomic arousal (trembling, weeping, blushing), or through “action tendencies” (fist-clenching, hiding, caressing). But Collingwood says that expression in music (and the other arts) is quite distinct from displaying symptoms of emotion (as he calls blushing and fist-clenching and so on). A flood of tears betrays an emotion willy-nilly; a symphony that expresses emotion is an object intentionally constructed so as to express that emotion.
Second, an artistic expression is distinguished from merely describing or labeling an emotion: when I say “I love you,” that would seem to be a paradigm expression of love in ordinary life, but it is not an expression at all in Collingwood’s sense, because describing my emotion as “love” generalizes it; my words do not capture the specificity of my love for you and distinguish it from all other loves. Artistic expression, on the other hand, individualizes an emotion. If the funeral march of Beethoven’s “Eroica” Symphony expresses sorrow, this is a quite distinct sorrow from that expressed by the funeral march in Chopin’s B-flat minor Piano Sonata. (See Ridley 1995 for one way of explaining the difference.)
Third, Collingwood notes that expression in art cannot be identified with the arousal of emotion in audiences: an artist “proper” should not be aiming to arouse emotions in audiences, because that would be manipulating other people’s emotions rather than sincerely expressing his own. However, if a composer genuinely succeeds in expressing an emotion in a piece of music, then the audience should, as a kind of by-product, be able to experience it for themselves.
What really makes the difference between ordinary expression and the expression of emotion in music and the other arts for Collingwood is that artistic expression is essentially a cognitive process, a matter of articulating an emotion in such a way that the nature of the emotion is clarified for the understanding. Here we see indirectly the influence of Hegel, who thought of the arts as a mode of understanding distinct from both religion and philosophy. Collingwood’s main examples are literary: the poet who wants to express his emotions in a poem but does not know exactly what emotions he is feeling, yet who, in writing the poem, reflects upon and thereby comes to understand that emotion. An emotion that was unclear in the poet’s mind is clarified once it has been articulated in a structure of words, imagery, rhythm, and other poetic devices. As for the reader, Collingwood claims that in order to understand what a poem expresses, the reader should experience it for herself and come to grasp what is expressed by recreating in herself the emotions of the artist that are expressed in the poem. So the poet is not aiming to arouse our emotions, but if he does a good job, he will have created a poem that will in fact enable us to recreate his emotions and feel them for ourselves. Thus Keats’s “Ode to a Nightingale” expresses the poet’s longing for an unchangeable world of art and beauty far away from “the weariness, the fever, and the fret” of our mundane world, and as we read the poem, we imagine the poet’s situation and come to experience the emotions with which he responds to it.
It is important to remember that the concept of art as a personal expression of emotion originated in Romanticism. Keats’s Ode is a paradigm of expression because in it the poet – or his persona – is expressing some complex emotional state that he is actually experiencing and there is development in this emotional state from the beginning to the end of the poem. This is what expression is in its fullest sense: an achievement by an artist, not a mode of experiencing by a reader or listener.
But how can music express in this full-blooded way? The doggy theory rightly suggests that we can experience music as resembling the vocal expressions and the motor activity – including expressive bodily gestures and action tendencies – that characterize particular emotions. But music can also to some extent express the appraisals in emotion: we can hear in the music when things are going along in a regular, pleasant way, and when they take a turn for the worse. There are also ways in which music can express desire, aspiration, or striving: a theme may struggle to achieve resolution, fail, try again, and finally achieve closure; or one theme may gradually and with apparent difficulty transform into a theme with a different character. There are many different strands in our emotional life, as different emotions ambiguously intertwine, morph from one to another, or blend to make a new emotional state. It would seem, then, that music, which is also woven of many strands, is peculiarly well suited to mirror our emotional life.
In a Romantic lied, such as “Gute Nacht” words and music collaborate to express the protagonist’s unhappiness at having been rejected by his beloved and his sense of defeat and abandonment. The Winterreise is of course both an actual and a psychological journey, but even this one song is a mini-drama in itself: the wanderer’s emotions shift and change from the beginning to the end. From the first bars, the funereal D minor harmonies, the descending notes of the piano accompaniment, and the harsh dissonance on the penultimate harmony of the cadence tell us that we are in a dark, cold world both physically and psychologically. We hear the wanderer trudging along in the repeated chords of the piano accompaniment, which continue throughout the piece. The repetitive character of the accompaniment seems to mirror his obsessive thinking about what he has left behind. But in the fourth and final verse, D minor changes to D major with the words “Will dich im Traum nicht stören, wär Schad’ um deine Ruh.” Suddenly a hopeful vista seems to open that had been closed off before. The wanderer nostalgically remembers his beloved and in his imagination tenderly tells her that he will not awaken her but will instead inscribe “Gute Nacht” on the gate as he departs, so that she will know that he was thinking of her. But as he repeats “An dich hab’ ich gedacht” a second time, the piece sinks back into the darkness and despair of the tonic D minor.
The words and music of “Gute Nacht” articulate the development of the protagonist’s emotions in just the same way as in Keats’s Ode. The lied illustrates how music can convey the way things seem to be going from good to bad or from bad to good, a sense that desires have been gratified or disappointed, and a sense that memories have engulfed a person or been swept away. What is even more interesting, however, is that some “pure” or “absolute” music can express the emotions of a protagonist in a very similar way.
Every piece of music, says Edward T. Cone, has an “expressive potential” (1974: 171) able to be realized in different ways in different contexts, but with broad limits on what it can express. Thus the expressive potential of a piece can include a movement from grief to joy, from being oppressed by difficulties to overcoming them, or from dreading a direful fate to resignation. The possibilities are extensive, but they do not include just anything. In particular they do not permit joy turning into grief, or a sunny life that turns sour.
Why should we interpret music as “mirroring” emotional processes rather than processes in inanimate nature: clouds followed by the sun or a stormy sea gradually calming down? In the case of “Gute Nacht,” it is clear from the words that the song is about the protagonist’s emotions. But what about “pure” instrumental music? The answer is that in the nineteenth-century Romantic tradition, it was thought normal and reasonable for music without words to express the emotions of characters or composers. Indeed, new forms or adaptations of old ones – nocturnes, impromptus, tone poems, and program music of all sorts – were created partly in order to increase the possibilities of emotional expressiveness. When Schumann wrote music expressing the conflict between his two personae, Florestan and Eusebius, when Shostakovich imprinted his signature motif on symphonies and string quartets, when Mahler composed symphonies that morphed into mini-operas or oratorios, they were following a Romantic tradition of expressing the self (and its various personae) in their music.
Not all expressive music is populated with personae who are expressing their emotions, however. If we are listening to an Impressionist work of program music (e.g. La Mer), we know we should not be looking for a persona in the music (although one could interpret this piece as somebody’s impression of the sea, rather than a straightforward pictorial characterization of the sea). If we know we are listening to a Baroque character piece, such as Couperin’s “La Superbe,” then it is reasonable to hear a particular type of person in the music, but not reasonable to think we are experiencing an outpouring of emotion by that person. Sometimes, we will know we are entitled to find a persona in a work of instrumental music because the composer has given us an evocative title, such as Schubert’s Wanderer Fantasy. But even where there is no special hint, it is reasonable to interpret certain kinds of Romantic instrumental music as expressions of emotion in a persona, because that was how composers of the time thought of (some of) their compositions. (For further discussion and defense of this view, see Robinson 2007a. For excellent examples of this type of criticism see Newcomb 1984 and 1997. For a recent full-bore attack on this approach see Kivy 2009.)
The Wanderer Fantasy is not the only late work of Schubert’s in which we find the theme of the “wanderer,” who is an outcast from the world just like the protagonist of Winterreise. Cone has argued that the A-flat Moment Musical, Op. 94 No. 6, “dramatizes the injection of a strange, unsettling element into an otherwise peaceful situation” (Cone 1986: 26). This idea has great “metaphorical resonance” in Anthony Newcomb’s phrase, suggesting the idea of the stranger or outsider, the “Fremdling” of Georg Philipp Schmidt von Lübeck’s poem “Der Wanderer,” which Schubert set to music as a song that later he used as the theme for the Adagio of the Fantasy. Newcomb has christened these kinds of story structures in music “plot archetypes” (Newcomb 1984).
Charles Fisk (2001) has made a particular study of the trope of the wanderer or outcast in Schubert’s late music. For example, in the first movement of the Piano Sonata D960 in B-flat there is a harmonic “outsider,” embodied in the strange trill on G-flat which interrupts the cheerful ambulatory music that opens the piece. Fisk describes how the music seems to dramatize a search for reintegration of this “alien” element, as the music wanders into far distant keys, and he tells a psychologically convincing tale in which the wanderings are those of a persona, whom he identifies for various reasons with the composer himself, who is seeking to be integrated into the “normal” group. Fisk’s underlying premise is that there are suggestions in Schubert’s cyclic forms and tonal structures of larger dramatic structures, in which there are agents or personae expressing complex emotions and desires.
Once we hear the structure of a piece of music as a psychological as well as a musical structure, then we are able to hear in it not just specific emotions but patterns of emotion. Moreover we can hear in it not only the effects noticed by the doggy theorists but also more complex emotions such as yearning, nostalgia, and resignation, all prime examples of Romantic emotions. And it is no surprise that we are moved by such expressions, because they are not just emotional appearances but have psychological reality, although the psychology in question may be that of a fictional persona.
See also Arousal theories (Chapter 20), Music’s arousal of emotions (Chapter 22), and Resemblance theories (Chapter 21).
Collingwood, R.G. (1963) The Principles of Art, Oxford: Clarendon Press.
Cone, E.T. (1974) The Composer’s Voice, Berkeley: University of California Press.
—— (1986) “Schubert’s Promissory Note: An Exercise in Musical Hermeneutics,” in W. Frisch (ed.) Schubert: Critical and Analytical Studies, Lincoln: University of Nebraska Press, pp. 13–30.
Davies, S. (1994) Musical Meaning and Expression, Ithaca: Cornell University Press.
—— (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, pp. 179–91.
—— (forthcoming) “Infectious Music: Music-listener Emotional Contagion,” in A. Coplan and P. Goldie (eds) Empathy: Philosophical and Psychological Perspectives, New York: Oxford University Press.
Fisk, C. (2001) Returning Cycles: Contexts for the Interpretation of Schubert’s Impromptus and Last Sonatas, Berkeley: University of California Press.
Hospers, J. (1955) “The Concept of Artistic Expression,” Proceedings of the Aristotelian Society 55: 313–44.
Karl, G. and Robinson, J. (1995) “Shostakovich’s Tenth Symphony and the Musical Expression of Cognitively Complex Emotions,” reprinted in Robinson 1997: 154–78.
Kivy, P. (1980) The Corded Shell: Reflections on Musical Expression, Princeton: Princeton University Press.
—— (2002) Introduction to a Philosophy of Music, New York: Clarendon.
—— (2009) Antithetical Arts: On the Ancient Quarrel between Literature and Music, Oxford: Oxford University Press.
Levinson, J. (1996) “Musical Expressiveness,” in The Pleasures of Aesthetics, Ithaca: Cornell University Press, pp. 90–125.
—— (2006) “Musical Expressiveness as Hearability-as-expression,” in M. Kieran (ed.) Contemporary Debates in Aesthetics and the Philosophy of Art, Oxford: Blackwell, pp. 192–204.
Newcomb, A. (1984) “Once More ‘Between Absolute and Program Music’: Schumann’s Second Symphony,” 19th Century Music 7: 233–50.
—— (1997) “Action and Agency in Mahler’s Ninth Symphony, Second Movement,” in Robinson (1997), pp. 131–53.
Nussbaum, C. (2007) The Musical Representation: Meaning, Ontology, and Emotion, Cambridge: MIT Press.
Ridley, A. (1995) Music, Value and the Passions, Ithaca: Cornell University Press.
Robinson, J. (ed.) (1997) Music and Meaning, Ithaca: Cornell University Press.
—— (2005) Deeper than Reason: Emotion and its Role in Literature, Music, and Art, Oxford: Clarendon Press.
—— (2007a) “Can Music Function as a Metaphor of Emotional Life?” in K. Stock (ed.)
Philosophers on Music: Experience, Meaning, and Work, Oxford: Oxford University Press, pp. 149–77.
—— (2007b) “Expression and Expressiveness in Art,” Postgraduate Journal of Aesthetics 4: 19–41, available at www.british-aesthetics.org/uploads/Expression%20and%20Expressive ness%20in%20Art.pdf.
Tormey, A. (1971) The Concept of Expression: A Study in Philosophical Psychology and Aesthetics, Princeton: Princeton University Press.
Budd, M. (1985) Music and the Emotions, London: Routledge and Kegan Paul. (An examination of several of the classic accounts such as those by Schopenhauer and Langer.) Goodman, N. (1968) Languages of Art, Indianapolis: Bobbs-Merrill. (Chapter 2 contains an important account of musical expression as metaphorical exemplification. For Goodman more than just emotional properties can be expressed in works of art and music.) Langer, S. (1957) Philosophy in a New Key, 3rd edn, Cambridge: Harvard University Press. (A classic theory of musical expression as a kind of symbolism.)