Purely instrumental musical passages and works without words or an associated program or story are often experienced, by many laypersons and musicians, as being sad, happy, calm, angry, and so on. However, as something that has neither life nor consciousness, music cannot itself possess such mental states. And this leads to the philosophical problem of musical expressiveness, the problem of how something inanimate and insentient such as music can be, and be heard as, sad, happy, and the like; other formulations of the problem ask how music can be described as sad (Kivy 1989: 6–10), or how it can possess or have sadness “inhering” in it (Kivy 2002: 31–2), or how emotions could be expressed in it (Davies 1994: x, 2001: 169, 173), but let us focus on many people’s ready and immediate experience of music as sad rather than descriptions of this experience, though the positive view advanced in this chapter can also answer these other formulations of the problem, as we will see later.
To begin, let us address a couple of clarifications before proceeding further. First, at least since Alan Tormey (1971), philosophers have distinguished between expression and expressiveness. To express a mental state is to display outwardly an actual occurrent state in one’s psychology, whereas being expressive of a mental state involves merely displaying outwardly features typically associated with that state, without necessarily having or feeling that state; the performance of actors, for example, is usually expressive of mental states that actors do not actually feel while acting. Second, one might ask about the truth of claims about musical expressiveness: why is it true, or what makes it true, that Samuel Barber’s Adagio for Strings, for example, is sad or mournful (or something in that ballpark)? One might give an error-theory in answer, claiming that such truth-judgments involve an error for music cannot be literally sad. Or one might say they are metaphorically true (Scruton 1997), though it is unclear what the alleged metaphor ultimately amounts to (Davies 1994: 150–62; Levinson 1996: 105–6). Alternatively, it might be claimed that such truth-judgments are literally true but in a secondary sense (Davies 1994: 162–6), though here one might doubt if the literal/metaphorical distinction ultimately illuminates much (Budd 2003: 220), and also whether appeals to it are too influenced by the linguistic turn in analytic philosophy some decades back. Other possible answers may involve the suggestion that such judgments are only imagined to be true, or that they are true in virtue of resemblance between music and something to do with mental states, or that the truth-maker here is the consensus of competent (but fallible) listeners, or some combination of these. One might also step back from the entire question of truth, and claim as above that the experience of music in terms of mental states has primacy over linguistic descriptions of the experience and the truth of these, and so we should focus on that experience instead.
Peter Kivy and Stephen Davies, amongst others, have tried to solve the problem of musical expressiveness by appealing to various perceived or experienced resemblances between music and the vocal, bodily, and behavioral expression of various mental states (Kivy 1989, 2002; Davies 1980, 1994, 2001, 2006), though Kivy has recently distanced himself from the resemblance theory, and now claims it is unknown how music possesses the emotions we hear in it (Kivy 2002: 47–8). In this chapter I will first briefly summarize these resemblance theories, and then I discuss criticisms of these views, and some possible replies to these criticisms. I will conclude by sketching a resemblance-plus-imagination, or imaginationist, view of musical expressiveness, which builds on the many insights of resemblance theories, instead of throwing away the baby with the bath water. Progress in intellectual inquiry of many sorts, including philosophy, usually involves building on the achievements of one’s predecessors; Newton, for example, famously claimed that if he had seen further than others, it was only by standing on the shoulders of giants, referring thereby to such physicists before him as Kepler and Galileo.
My summary of resemblance theories of musical expressiveness begins with Peter Kivy’s theory, which he sometimes calls the contour-convention view (Kivy 1989: 71–83). Kivy claims that expressive properties are “objective” qualities that are recognized or perceived in the music just as we recognize sadness in a St. Bernard dog’s face, rather than being something the music only has in virtue of arousing or evoking mental states in listeners. Musical expressiveness is a complex, emergent quality. We hear musical sounds as expressive of sadness because we hear them as human utterances, as structurally similar to our voices when we express sadness vocally. Additionally, Kivy says musical contour or shape can also resemble our expressive behavior – movement, gesture, posture, and the like. We hear sadness in music because we hear it resembling the gestures and bearing of sad people. Likewise, happy music is heard as such because it resembles the motion and gestures of happy people in being expansive, vigorous, “leaping,” and so on.
Kivy also claims that we tend to animate all kinds of sights and sounds, and cannot but perceive expressiveness in them, in ways that are not always conscious or noticed (1989: 57–9, 2002: 41–3). A piece of cloth tied around a wooden spoon will be taken by children to be a doll; a circle with three short lines in it (two on top, adjacent to each other, and one below and parallel to them) is seen as a face. Likewise, claims Kivy, we see figures in clouds, and hear gesture and utterance in music, even though we are not conscious of our animation of it that allows us to hear it as expressive. We may, he suggests, be evolutionarily hard-wired to animate things, as this is conducive to our survival; for example, seeing a stick as a snake puts us on our guard, whereas doing the reverse would be disastrous. Similarly, we may animate sounds subliminally.
The final element in Kivy’s resemblance theory is his appeal to musical conventions (1989: 80–3). He claims it is only due to the customs or conventions of the Western musical tradition that the major scale, triad, and third are heard as upbeat, while minor keys, chords, and the minor third are heard as expressive of grief, sorrow, etc. Likewise, musical conventions account for why chromaticism is heard as expressive of sorrow, pain, and the like. Thus, claims Kivy, contour, or resemblance, and convention together explain musical expressiveness, sometimes separately and sometimes jointly.
Stephen Davies’s resemblance theory is quite similar to Kivy’s (Davies 1994: 221–67). Davies claims that inanimate and insentient things such as weeping willows, cars, and St. Bernards may display features that resemble what he terms “emotion characteristics” of human sadness in their overall bearing, posture, or appearance, and are thus seen as expressive. Similarly, argues Davies, music presents emotion characteristics associated with human expression of emotions in its aural appearance or sounds, and thus is expressive of emotions it does not itself possess. Musical expressiveness, claims Davies, is a public, objective property of the music, one that it possesses literally, and which mainly depends on perceived or experienced resemblances between the dynamic character of music and the demeanor of the human body – its movement, gait, bearing, carriage, and so on. In sum, in Davies’s view, music is expressive in virtue of presenting the outward features associated with sadness or happiness in general. Music is expressive in resembling the bodily stance, gait, bearing, carriage, and gestures typically expressive of particular emotional states. Just as sad people often walk slowly, hang their heads low, droop in their bodily stance, and are generally subdued, similarly sad music is often slow, has a downward tendency, is quiet, and so on. Likewise, just as happy people tend to skip and leap quickly and lightly and make expansive gestures, happy-sounding music is often similarly lively and exuberant.
A different kind of resemblance theory that there is not enough space here to discuss at length but should be mentioned at least briefly has been offered by Malcolm Budd (1995: 133–57) who claims, following the American psychologist Carroll Pratt (1931), that music sounds the way emotions feel: there are cross-categorial similarities between music and the emotions, as music mirrors our inner lives in having tension and resolution, in having intermediate and final goals that it strives toward, and so on. Budd’s view has been criticized elsewhere (Trivedi 2001) on grounds very similar to those offered below against Kivy and Davies. Another view that should be mentioned here briefly in passing is that of Suzanne Langer (1942), who claimed that music is an iconic symbol of the emotions on account of isomorphisms between music and the emotive life in general. Langer’s view has been criticized at length by Stephen Davies (1994: 123–34).
Let us now consider four criticisms of resemblance theories, as well as possible replies to some of these criticisms. To begin with, one might doubt if music really resembles the emotions, or something to do with them such as emotional behavior (Madell 2003). It should not be too hard for resemblance theorists to reply to this concern, appealing to two moves. As a first move, they can point to various resemblances between music and something to do with mental states, either their vocal or bodily or behavioral expression or their affective tones. A lot of music seems to sound like human vocal expression: think of rapid runs and glissandi on clarinets, saxophones, and electric guitars which often sound like someone crying or wailing, the opening clarinet glissando of Gershwin’s Rhapsody in Blue being one example of this. In addition, a lot of music is readily and immediately experienced by many as resembling the way sad people often walk slowly: the music is slow in tempo, low in pitch, and soft in volume, just as sad people hang their heads low, droop in their physical stance and gait, and talk softly. The opening passages of the second movement of Beethoven’s “Eroica” Symphony provide a well-known example of this. Also, along the lines of Budd’s suggestions briefly mentioned above, musical passages are often heard right away, both by musicians and by laypersons, as having tension, which may or may not be resolved later, and as having points of repose as well as final resting-points or goals (such as the tonic chord or key) which may be arrived at after intermediate goals (such as the dominant chord or key) have been reached, mirroring the way our lives often have tense moments, which may or may not be resolved, and the way we strive for and arrive at our intermediate and final goals.
Additionally, there is a second move resemblance theorists can make in reply, borrowing a leaf from those who criticize appeals to resemblance (especially when it comes to pictorial depiction). It is sometimes said that resemblance is a very broad (and vague) notion, so broad that just about anything can resemble anything else in some respect; for example, unicorns and Alpha Centauri might be said to resemble each other in that they are both mentioned in this sentence. Even if their critics are right about this point, resemblance theorists can go on to claim that it should not surprise us then that music resembles mental states in some way, such as the ways briefly discussed above.
Here is a second criticism, this time specifically against Kivy’s version of the resemblance theory. It might be doubted if we really animate sounds (Kivy 2002: 46–7). In reply, the resemblance theorist can offer the following two scenarios as examples of our animating sounds (Trivedi 2006). Very often, while walking down quiet, empty city streets late at night, one might hear a noise. Immediately, one is on guard, thinking that the sound might be coming from another person (perhaps a potential mugger) or some creature (such as a vicious dog on the loose). It turns out, however, that the sound is only that of a leaf rustling in the wind. Similar things happen when one hears a sound while going round the bend on a quiet, lonely mountain trail. Once again, one is on guard immediately, fearing the sound might be coming from a creature (such as a bear) or another person (perhaps someone dangerous). It turns out, however, that the sound is only that of a branch breaking off a tree. Both these cases provide clear sonic analogs of Kivy’s example of animating the stick in the forest as a snake, as this helps our survival. Now it certainly seems to be the case, as Kivy has suggested before, that as a species we depend more on sight than on hearing for survival; and it is also true that our noise-filled modern lives are rarely filled with silence for very long. Add to this the fact that the animation of sounds may be very dim or subliminal, and you begin to get some sense of why it is hard to detect the animation of sounds, making some skeptical of this.
A third criticism of the resemblance theory seems more pressing. Besides the fact that resemblance and expressiveness are philosophically and logically quite distinct as concepts, perceived resemblances by themselves are not sufficient for expressiveness, nor for hearing it, though resemblance may be causally necessary for expressiveness. All kinds of things may resemble how we vocally or physically or behaviorally express various mental states or the affective tones of these mental states, but they are not thereby expressive of these mental states, even if we perceive these resemblances. For example, turtles move slowly, with their heads hung low, and their bodies very close to the ground, resembling the way sad people often walk. But such resemblances and perceptions of them do not by themselves necessarily lead to our seeing turtles as sad, or as expressive of sadness. To see turtles as sad, we need to add to the account something more than merely these resemblances that we perceive.
Now, Kivy and Davies are aware of the concern that resemblance is not a sufficient condition for expressiveness. Kivy characterizes the sufficiency objection to resemblance theories as follows: according to resemblance theories, music should be expressive of everything it resembles, such as ocean waves, the rise and fall of the stock market, and so on, which is clearly not the case (Kivy 1989: 61–2). In reply, Kivy claims that it makes no sense to say that music is expressive of ocean waves or the stock market. Expressiveness must be of mental states, thus the objection flouts a “logical” condition of expressiveness. It is important to see here, however, that Kivy has not stated or addressed our objection above that perceived resemblances are not sufficient for expressiveness, even if he may have stated and answered a related objection. Our objection is not that music should be expressive of everything it resembles, such as ocean waves and the stock market. Rather, our objection is that all kinds of things, such as turtles, may resemble our vocal or bodily or behavioral expressiveness, or the affective feel of mental states such as emotions, moods, and feelings, and we may perceive these resemblances, but that alone does not make them expressive. The same holds for music.
Davies also tries to answer the concern that perceived resemblances are not sufficient for expressiveness (2001: 184). As he states this worry, it is that resemblance alone cannot ground musical expressiveness or explain why we experience music as expressive, for resemblances can be found between music and many things in addition to the resemblances between music and expressive appearances. Davies replies that we can simply say that “this is how we hear” the music (as expressive), without being committed to explaining what mechanisms underlie and trigger this response. Many insentient things, such as pictures of the human face, crude masks of tragedy and comedy, and Edvard Munch’s “scream” face, are likewise experienced as being expressive. The resemblance theory is no worse on this count, asserts Davies, than other theories, which he claims are in no better position to go beyond perceived resemblances in explaining expressiveness. Once again, it is worth noting here that, like Kivy, Davies has not quite addressed our concern. Our worry is not about things resembling music in their expressivity, as Davies puts it. Instead, it is about things such as turtles resembling our vocal or bodily or behavioral expression or the affective feels of mental states, which are not thereby expressive, even though we may perceive these resemblances consciously or otherwise. The concern, then, is why the case of musical expressiveness should be any different, why perceived resemblances alone should suffice to make music expressive. To be sure, Davies claims that this is just how we are psychologically, “this is how we hear” the music (as expressive), thus making the question not one for philosophers to answer. But contra Davies, it is not clear that we have here a brute fact not amenable to further philosophical explanation, and one might instead be able to dig deeper and say more, building on the notion of perceived resemblances and adding something more to the picture, as is attempted in the next section of this chapter.
I turn now to a fourth, and arguably the most formidable criticism of resemblance theories of musical expressiveness in general. The resemblance theories of Kivy, Davies, and Budd, even when combined, give us the causal grounds or mechanisms underlying musical expressiveness. They may tell us what causes or allows music to be, and to be heard as, expressive, to wit, perceived resemblances between music and something to do with mental states such as emotions, moods, and feelings. Put differently, these views tell us why we hear music as expressive: we hear music as sad, happy, etc., because or in virtue of various resemblances we consciously or otherwise hear between the music and something to do with such mental states. However, merely giving us this causal story underlying musical expressiveness does not tell us how something inanimate and insentient such as music can be, and be heard as, sad, happy, and the like, which is the basic problem of musical expressiveness. How can music, a sequence or set of sounds without life, consciousness, or mental states be sad or somehow have sadness “in” it, and be experienced to be so? This question is not adequately answered by resemblance theories. There must thus be doubt about whether resemblance theories even address let alone solve the basic problem of musical expressiveness, instead of giving us a mere causal story about what makes music expressive (Levinson 1996: 106; Scruton 1997: 147).
I will now sketch a resemblance-plus-imagination, or imaginationist, view of musical expressiveness, taking the resemblance theory as the causal foundation of the imaginationist view, and adding an imaginative component that shields it from the objections discussed above.
The imaginationist grants three claims made by resemblance theorists: (1) that there exist various sorts of resemblances between music and something to do with mental states such as emotions, moods, and feelings; (2) that listeners may hear these resemblances in not always highly foregrounded or conscious ways; and (3) that these resemblances may provide the causal basis or ground of why we hear music as expressive.
Here is a very brief, rough statement of the resemblance-plus-imagination view of musical expressiveness, also argued for at length elsewhere (Trivedi 2001, 2003, 2006): music is willy-nilly, readily, and immediately imagined by listeners in various, not always highly conscious, ways to be sad, happy, and so on, because it is consciously or otherwise perceived to resemble something to do with mental states such as emotions, moods, and feelings, such as their vocal or bodily or behavioral expression, or their affective feel or tones. Note in passing that this view can also answer the other formulations of the problem of musical expressiveness that we saw at the very start of this chapter: music is not literally or really sad but is rather only imagined to be so; it is only imagined that sadness “inheres” in it; it is only imagined to express sadness, which it cannot really do.
What follows is a non-exhaustive list of various, not always highly conscious, ways in which we imagine the music is sad, happy etc. because we consciously or otherwise perceive it to resemble something to do with mental states. One kind of imagining involves our animating the music, imaginatively projecting life and life-like qualities, including mental states, onto it willy-nilly, readily, and immediately. This kind of imagining may happen especially when we listen to very intense music, such as passages in Beethoven’s late string quartets. In such cases, we may hear the music itself – not something besides it, such as the composer or performer or the musically aroused listener or an indeterminate, imagined persona in the music, or something else – as the very thing that is emotionally expressive. Animating the music is very similar to the kind of animation that we as a species engage in when we imaginatively see faces in clouds or rocks, as our pagan ancestors did in seeing life and gods in the sun, thunder, the ocean, and so on. And animating the music is similar to what we do when we see comic strips and imagine within the world of the comic strip that the talking and expressive cars, trains, trees, or sun we see in them are themselves sad, happy, etc. Animation films provide an even better example, for they consist of changing images, just as musical passages are dynamic processes. Animating music involves a very similar, if not the same kind, of imagining, except that it is harder to detect musical animation due to both the abstract nature of music as an art and the fact that we engage in various imaginings without always noticing at the time that we are doing so (Trivedi 2001). Note incidentally that this notion of animation is “thicker” than the one which Kivy appeals to, for it involves not just Kivy’s idea that we hear gesture and utterance in the music but also requires in addition that we imaginatively project life and mental states onto the music. Note also that our animating the music in this manner provides the simplest and most natural solution to the problem of musical expressiveness: we hear something inanimate and insentient, such as music, as sad for we imaginatively project life and mental states onto the music, imagining that it is alive and possesses the mental states we hear in it.
Alternatively, we may sometimes imagine of the music that it is the expression of a mental state by an indeterminate, imagined persona in the music, as claimed by Jerrold Levinson and Jenefer Robinson, amongst others (Levinson 1996, 2006; Robinson 2005). In such cases, we may form an auditory image and imagine that someone or something, we know not exactly who or what, is crying or laughing or dancing or expressing themselves somehow in the music. Note that imagining a musical persona is different from the animation of music described above (Trivedi 2001: 416): The persona is someone or something “in” the music and is thus philosophically distinct (even if not detached) from the music rather than being the music itself; and in imagining a musical persona, the persona is imagined to have the mental states heard in the music, whereas in animating the music, the music itself is imagined to have the mental states heard. Note also that to imagine the music itself is experiencing mental states need not involve imagining the music is an indeterminate persona or a product thereof, though of course the music itself is imagined as something capable of having mental states.
Third, we may sometimes imagine in ways not highly foregrounded that it is the musical instrument(s) that are sad, happy, and the like. Witness, in this vein, talk of wailing violins, weeping guitars, etc. Likewise, one might also sometimes imagine that the composer(s) or performer(s) are expressing their emotions musically.
A fourth kind of imagining involves imaginative identification, and this can happen in various ways. Sometimes we may imagine of our auditory experience of hearing the music that it is an experience of our feeling the mental state we hear the music as expressive of (Walton 1988, 1994). In such cases we imaginatively identify one experience with another experience, imagining our having the feeling that we hear the music as expressive of. On other occasions, we may imaginatively identify with the music, imagining that it is expressive of our own emotion; in doing so, we may feel as if we are the music (Budd 1995: 168). Alternatively, one might imaginatively identify with the performer(s), or with the musical persona, or with some (fictional) persona of the composer, and so on.
There may be ways of imagining musical expressiveness besides those adumbrated above. This should not surprise us, given the many ways in which we imagine things, and the fact that we may often imagine things without being aware at the time that we are engaged in certain imaginings that are not very highly foregrounded.
Resemblance theories of musical expressiveness appear to get a lot of things right. It seems there are resemblances of various sorts between music and something to do with mental states; that we perceive these resemblances consciously or otherwise; and that resemblances account for the causal story underlying what allows music to be heard as expressive. However, resemblance theories have some drawbacks, two of which seem especially troublesome. First, besides resemblance and expressiveness being distinct concepts, mere resemblance does not seem sufficient for expressiveness. This is partly what motivates adding imagination to resemblance to complete the picture. Second, while resemblance may give us the causal story behind expressiveness, it does not explain by itself how something inanimate and insentient such as music can be and be heard as sad, unless one also claims, as the positive view advanced above does, that we imagine the music is sad, often animating it, imagining that the music itself is alive and possesses the mental states we hear in it.
See also Analytic philosophy and music (Chapter 27), Arousalist theories (Chapter 20), Expression theories (Chapter 19), Hanslick (Chapter 33), Music and imagination (Chapter 11), and Music’s arousal of emotions (Chapter 22).
Budd, M. (1995) Values of Art, London: Penguin.
—— (2003) “Musical Movement and Aesthetic Metaphors,” British Journal of Aesthetics 43: 209–23.
Davies, S. (1980) “The Expression of Emotion in Music,” reprinted in Davies (2003), pp. 134–51.
—— (1994) Musical Meaning and Expression, Ithaca: Cornell University Press.
—— (2001) “Philosophical Perspectives on Music’s Expressiveness,” reprinted in Davies (2003), pp. 169–91.
—— (2003) Themes in the Philosophy of Music, Oxford: Oxford 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.
Kivy, P. (1989) Sound Sentiment, Philadelphia: Temple University Press.
—— (2002) Introduction to a Philosophy of Music, Oxford: Clarendon Press.
Langer, S. (1942) Philosophy in a New Key, Cambridge: Harvard 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.
Madell, G. (2003) Philosophy, Music and Emotion, Edinburgh: Edinburgh University Press. Pratt, C. (1931) The Meaning of Music, New York: McGraw-Hill.
Robinson, J. (2005) Deeper than Reason, Oxford: Clarendon Press.
Scruton, R. (1997) The Aesthetics of Music, Oxford: Oxford University Press.
Tormey, A. (1971) The Concept of Expression, Princeton: Princeton University Press.
Trivedi, S. (2001) “Expressiveness as a Property of the Music Itself,” Journal of Aesthetics and Art Criticism 59: 411–20.
—— (2003) “The Funerary Sadness of Mahler’s Music,” in M. Kieran and D. Lopes (eds) Imagination, Philosophy, and the Arts, New York: Routledge, pp. 259–71.
—— (2006) “Imagination, Music, and the Emotions,” Revue Internationale de Philosophie 60: 415–35.
Walton, K. (1988) “What is Abstract about the Art of Music?” Journal of Aesthetics and Art Criticism 46: 351–64.
—— (1994) “Listening with Imagination: Is Music Representational?” Journal of Aesthetics and Art Criticism 52: 47–61.
Budd, M. (1985) Music and the Emotions, London: Routledge. (The classic treatment of several earlier theories about music and the emotions.)
Davies, S. (1999) “Response to Robert Stecker,” British Journal of Aesthetics 39: 282–87. (Defends Davies’s resemblance theory against Stecker 1999, below.)
Hjort, M. and Laver, S. (eds) (1997) Emotion and the Arts, New York: Oxford University Press. (Useful anthology of fifteen essays.)
Matravers, D. (1998) Art and Emotion, Oxford: Clarendon Press. (Criticizes resemblance theories and other theories of musical expressiveness, and advances a moderate arousalism.)
Ridley, A. (1995) Music, Value, and the Passions, Ithaca: Cornell University Press. (Criticizes resemblance theories, and advocates a moderate arousalism that mediates between resemblance theories and strong arousalism.)
Robinson, J. (ed.) (1997) Music and Meaning, Ithaca: Cornell University Press. (Ten interdisciplinary essays on musical meaning, language, metaphor, imagination, emotion, and drama.)
Stecker, R. (1999) “Davies on the Musical Expression of Emotion,” British Journal of Aesthetics 39: 273–81. (Criticizes Davies 1994, above.)