Music and Emotion
Three

When German poet Heinrich Heine personifies his artistic creations as birds, he introduces a metaphor. But what about his claim that his emotional distress is the material of his songs? Is this to be taken literally, as a theory of artistic inspiration? Or is it another metaphor?

Although I have thrown cold water on the idea that birds and German poets are kin in their ability to make music, I face a long history of opposition. Some of that opposition arises from the longstanding view that music expresses emotion, and this expressive capacity is the essence of musical art. Many bird lovers endorse it, as well. They believe that birds sing for some reason besides communicating location, staking out territory, and attracting a mate. Like people, birds sing in order to express emotion. In “Ode to a Nightingale,” John Keats contends that the bird’s music expresses ecstasy. More recently, philosopher David Rothenberg argues that the variations that birds introduce into their songs demonstrate that they are doing more than communicating. Sometimes, it seems as if “birds burst into song out of pure joy,” and there is no scientific way to prove otherwise. But why take it as an expression of joy? Because, Rothenberg argues, “Evolution is not supposed to produce beauty for the sake of loveliness alone.” In sum, if a nightingale’s song is beautiful, and its actual design exceeds what is required by evolutionary adaptation, then it must be an expression of joy, for it must also be doing what people do when they make music.

There is nothing special about Rothenberg’s version of the argument that beautiful bird song is a species of expressive art. Many people endorse it. From the perspective of influence, the most important version might be that of Jacques Delamain, the French ornithologist who trained composer Olivier Messiaen to recognize a variety of birdsongs. In the book Why Birds Sing, Delamain writes,

Delamain advances three distinct claims in this passage. First, the bird’s song displays or reveals its emotional life. Second, the songs display distinct, recognizable emotions that parallel human emotions. In this case, the song displays the bird’s joy. Third, the bird cannot contain or control itself. This, Delamain maintains, is how “musical art is born.”

This chapter argues that, on the contrary, the natural expressiveness of sound does not support the position that the expression of emotion is the essence of music. If the art of music is born anywhere, it emerges from a society’s conventions for adapting music’s expressive qualities to its specific purposes. Art reflects culture. Unchecked natural outpourings of emotion are not art.

I. Symptoms and Symbols

Understanding the debate about musical expression requires a detour into the general topic of expression. Emotions permit expression because emotions embody beliefs and attitudes about events.

Consider a particular sound, made on two different occasions. On Saturday evening, at a chamber music recital, a quartet is about to launch into Henryk Górecki’s third string quartet. The first movement begins very softly. Fifteen seconds into it, Frida coughs loudly. People who are seated nearby turn and glare at her. The woman seated beside Frida hands her a cough drop. Two days later, at Monday morning’s staff meeting, Frida and her colleagues are discussing impending changes in their office. One of Frida’s co-workers begins to say, “We’ll have to hire someone for . . .,” only to be interrupted when Frida coughs loudly. Frida realizes that her co-worker is about to blurt out a possible plan to re-assign some work duties. Frida thinks that it is not yet time to share those plans. Frida’s co-workers get the message and no one pursues the subject of hiring. No one hands Frida a cough drop.

This example illustrates two things. First, there is a difference between someone who is coughing and someone who is imitating or representing themselves as coughing. When Frida coughs on Saturday, the cough is a symptom of discomfort in her throat. The sound that she makes is caused by the discomfort. If she could contain herself, she would not cough, because she knows that it will disturb the audience for the music. She, and everyone else, wants to hear the music. On Monday, she can contain herself. There is no discomfort in her throat. Instead, she does not want her co-worker to continue talking. Her cough is staged to disrupt a line of conversation. The co-worker’s understanding of the message depends on understanding that it is not a symptom of a sore throat.

Like many philosophers today, I think that expression requires a mental state with an intentional object. That is, the mental state is directed at something, which is the object of that mental state. I endorse Jesse Prinz’s formulation, “The objects of our emotions are the real or imagined conditions that elicit them.” Emotions fall into the class of things that can be expressed, then, because they are within the realm of psychological states that judge or evaluate something. An expressive behavior is about something because it reflects a mental state that is about something. If we ask what mental state is expressed by Frida’s fake cough, we will probably agree that it expresses her annoyance at, and disapproval of, the coworker’s contribution to the conversation. So the staged cough expresses emotion. In contrast, Frida’s cough at the concert is not directed at anyone or anything. Consequently, it is a mistake to think that it expresses the feeling of dryness or irritation in her throat.

Like a cough or sneeze, many emotions are intimately related to the body. As Peter Goldie summarizes the crucial point, “Many emotions, especially short-term emotions such as fear, anger, and disgust, involve characteristic involuntary bodily changes—muscular reactions, hormonal changes . . . and so on.” These involuntary changes are the basis for an emotion’s feeling-component. Anger increases the flow of the hormones adrenaline and noradrenaline, causing an angry person to experience an increased heart rate, contributing one element to the cluster that characterizes feeling angry. Furthermore, there may be changes that other people can observe, permitting others to detect the emotion. As you become angry, an observer may see bulging neck muscles, flushed red cheeks, and a clenched fist. However, emotions are more than bodily changes that are felt and observed. Although many people use “emotion” and “feeling” as synonymous terms, what follows depends on using the criterion of intentionality or “aboutness” to make a distinction between them. Emotions include feelings, but not every feeling has an intentional object.

In Chapter 2, I explained how emotion involves a belief about how one’s past and present are linked to a possible future. For example, when Frida coughs at the concert, the anger of other audience members reflects a judgment that Frida is spoiling things. Given this requirement, there is an important difference between having the feeling that is characteristic of anger and actually being angry. Given this requirement, the mere presence of a characteristic feeling does not qualify as an emotion. Furthermore, emotions are frequently distinguished from moods, which are longer-term feelings that lack the directed “aboutness” of emotions. Feeling sad following a significant loss is very different from feeling depressed for no particular reason. For purposes of my argument, it is not important whether this distinction is genuine, or whether Jesse Prinz is correct that both emotions and moods have intentional content. However, moods illustrate that there are many complications that arise when we move beyond the most basic emotions, which are my focus here.

Psychologists and philosophers who study emotion generally agree that there are seven simple, core emotions. They are fear, anger, happiness, contempt, surprise, disgust, and sadness. Because they involve involuntary bodily changes, all seven have natural and therefore universal symptoms. The most notable symptoms are our distinctive facial expressions. These are not universal symptoms in the sense that they always occur and are always recognized. Rather, they are universal in the sense that people who can recognize them are equally adept at identifying them cross-culturally. Furthermore, they are natural symptoms, for individuals who are born without sight display the same core repertoire of facial expression as everyone else despite their inability to learn them by watching others. Like everyone else, babies born without sight express happiness with a smiling mouth and a downward shift of the outer corners of the eyebrows, and they express disgust with a wrinkled nose and raised upper lip. These natural facial expressions are our most straightforward, unproblematic examples of human expression of emotion. As a consequence, we are not always the best judges of our own emotions. People have subjective, "internal," first-person access to their own emotions byway of their characteristic feelings. However, the accompanying belief element is often fleeting and it does not always rise to consciousness, So if I am not clear about the judgment I am making, I may be unclear about which emotion is present. Suppose a friend tells me that a mutual acquaintance has been promoted at work. I feel surprised. My friend, watching my face, sees an expression of disgust. In this case, my friend may know better than I do how I feel about the news.

Although facial expression is the paradigm case of expression, it is not always trustworthy. Expression of emotion is culturally influenced—some even say culturally “scripted.” Each society develops a set of expectations about proper and improper emotional expressions, including rules concerning how, when, to whom, and to what degree various expressions are permitted. Facial expression is heavily policed by cultural display rules. In many Asian societies, “natural” emotional expression is highly restrained. With many Japanese adults, a strong emotional response is only detectable by watching for fleeting changes around the eyes, if it is to be seen at all. In contrast, Americans feel far freer than almost everyone else to indiscriminately display the core emotions through facial expression.

So far, I have concentrated on basic emotions, which are short-lived responses to immediate circumstances. We also have cognitively complex emotions, such as shame and jealousy. Cognitively complex emotions can involve judgment about complex social relationships that are themselves “scripted” by social rules governing behavior. Cultural context heavily informs both the emotion’s arousal and its outward expression. Suppose that Fred, one of Frida’s co-workers, starts to make belittling jokes about her job performance after she receives the promotion that he sought. If the jokes are only directed at Frida and only after her promotion, then they are not expressions of his otherwise humorous, teasing nature. They express Fred’s jealousy. Yet notice the enormously complex set of concepts that Fred must bring to bear on their respective social standings in order to feel jealous, and the additional complex understanding of workplace politics that directs its outward expression into belittling jokes.

Cultural scripting often converts short-lived emotions into extended episodes of expression. And of course expression becomes even more complicated as different expressive scripts are linked to class and gender. In Victorian England, for example, a respectable widow wore black mourning dress for at least a full year. This full mourning was followed by a year of second mourning, followed by a period of half-mourning. During these two periods, the amount of black fabric was reduced in turn. As an additional expression of her grief, a widow could not remarry for at least two years. In contrast, a widower could express his grief by wearing black gloves and perhaps a black armband for a few months. If he chose to remarry within two years of his wife’s death, his new bride was expected to adopt mourning garb for her predecessor. These Victorian rules for the expression of grief can seem overly elaborate, arbitrary, and generally insincere to us now, in the twenty-first century. However, it would be a great error to conclude that they do not count as genuine expressions of grief. Most expressions of emotion are more like Victorian mourning clothes than smiles on faces. If Fred’s jokes can express jealousy, then Queen Victoria’s decision to wear black for forty years following Prince Albert’s death can serve as a sincere expression of her grief.

Three closely related consequences follow from the cultural scripting of expression. Since music is also shot through with cultural scripting, these three points should inform our thinking about music’s capacity to express emotion. First, an expression of emotion does not have to express an occurrent feeling. Once we acquire a non-natural method for expressing emotion, the expressive gesture can be detached from the immediate circumstances and emotion, allowing individuals to “express” emotions that they do not possess. The full mourning “weeds” of a Victorian widow continue to express her grief during moments when something makes her happy. Second, once detachability is in place, the expression can be insincere. Who hasn’t said “Thank you” after receiving a gift that one is not thankful for? Third, there will be occasions when what looks to be an expression will not be expression at all. On any given occasion, a behavior or “expression” that is a characteristic symptom or conventional expression of an inner psychological state might be something else altogether. The symptom or convention might be present in the absence of any intention to communicate an emotional state. Lacking that intention, it does not even count as expression at all, not even as an insincere one. (To judge that a smile or another facial expression is insincere, one must believe that the person is intentionally misleading others, and that it is wrong to do so in the relevant circumstances.) The next section explores the idea that music is also a heavily “scripted” expression of emotion, and therefore music that seems to express emotion might not be expressing anything at all.

II. Expression and Expressive Qualities

Let us, at long last, return to music. Consider the Scottish fiddle tune, “The Soldier’s Joy.” It is appropriately named, for it is expressively joyful. Although several sets of lyrics have been added to it over the years, neither the title nor the lyrics have any bearing on my description of it as joyful music. Given a forced choice from the standard list of core emotions (fear, anger, happiness, contempt, surprise, disgust, and sadness), anyone who hears “The Soldier’s Joy” and who does not pick happiness is simply not competent with respect to the basics of Western music. Furthermore, it is very happy music, hence joyful.

But how can that be? It cannot experience joy, for it is a rhythmic and melodic pattern. It is no more capable of being joyful—of having the emotion of joy—than is a heap of pebbles or the equation "2 × 2 = 4." Nor can it express joy, for nothing can express joy unless it satisfies two conditions: it can experience the emotion of joy and it can provide an outward symptom or conventional symbol of that inner state. Since music is not a sentient thing with "inner," psychological states, music cannot satisfy the first of these two conditions for expression. Therefore "The Soldier's Joy" does not express emotion. There is only one way to avoid the conclusion that we talk nonsense whenever we say that the music is joyful or we say that it expresses joy. Our alternative is to deny that we mean what we seem to mean. It will surprise no one when I report that music lovers and philosophers prefer to save the day in this way, by interpreting "The music is joyful" non-literally.

The favored reinterpretation focuses on the meaning of “is” in “‘The Soldier’s Joy’ is joyful.” That word can be reinterpreted as short-hand for “is an expression of,” so that “‘The Soldier’s Joy’ is joyful” really means “‘The Soldier’s Joy’ is some person’s expression of joy.” It is just like seeing a festively decorated house in the weeks leading to Christmas and saying, “That house has the Christmas spirit.” What we really mean is that whoever decorated the house has the Christmas spirit, and the decorations express that emotional connection to the holiday. (Again, one lesson from the previous section is that the expression might be insincere. The decorations might be the work of someone who despises Christmas but who hopes to win the cash prize given annually to the most festively decorated home. However, the possibility of insincerity arises precisely because we treat it as the expression of a particular person’s emotion. Insincerity is a species of expression.) Notice, on this account, that an attribution of emotion to music necessarily requires attributing the emotion to the historical individual(s) responsible for the design. So “‘The Soldier’s Joy’ is joyful” really means “‘The Soldier’s Joy’ is its composer’s expression of joy.” On this account, “The Soldier’s Joy” is emotionally expressive in roughly the same way that Delamain says that the thrush’s song is emotionally expressive.

Many complications disrupt this appealingly simple theory. The important objections arise from the fact that the musical pattern does all the work of communicating the emotion. One might suppose that certain sound patterns are naturally expressive, as the natural, universal analogue to the natural behavior of moving one’s mouth into a smile to express happiness. However, just as the mouth can do more than smile and frown, expressive musical patterns have other functions besides expressing emotion. Except in the very limited case of a child’s spontaneous facial expression that betrays occurrent emotion, what appears to be an expression of emotion is frequently something other than an expression of emotion. What follows is an explanation of Alan Tormey’s important observation, made four decades ago, that the presence of expressive qualities is never sufficient for expression.

My argument depends on separating two features of “The Soldier’s Joy” that fulfill two different social functions. First, it is a reel, and I assume that whoever composed it intended it as such, which is to say that it was composed with the intention of functioning to support a specific kind of dancing. As I argued in Chapter 2, that intention reflects a great many concepts about music and its uses. Since this particular reel seems to date back to the middle of the eighteenth century, we can assume that the composer knew the difference between a jig and a reel, two dances with very different time signatures. Let’s assume, therefore, that the composer intended to compose a reel and succeeded (admirably!). The music’s second feature is its expressive quality. If you are going to maintain that “'The Soldier’s Joy’ is joyful” means that it expresses joy, then I assume you are maintaining that the composer intended to express his own personal happiness by creating and sharing this particular piece of music. My alternative proposal is that the composer’s success in achieving his rhythmic goal generates the conditions that create the impression that the composer must also have been expressing happiness. Therefore the music’s expressive quality of happiness may have been quite unintentional, and therefore the music does not necessarily express happiness.

Because I am contrasting expressive and non-expressive functions, I should say a little more about the functional value of expressing emotion. In particular, consider the point that the functional value of expressing emotion is not identical with the value of experiencing emotion. The capacity to experience emotion is functionally useful to the individual who has the emotion. Emotion guides behavior by alerting us to the significance of changing situations and by motivating actions in response. We can imagine sentient creatures who have emotions but who do not show them expressively in their bodies. (In societies that suppress the open expression of emotion, this might be regarded as a biological advantage!) So what advantage is gained by our natural proclivity to exhibit emotions? The obvious advantage is that we are social creatures. Consequently, it is useful to be able to “read” the expressions of emotion of people we encounter. That is why children with Asperger’s syndrome—characterized by a failure to notice and recognize facial changes that express emotion— are laboriously taught what other children learn spontaneously. However, looking at faces and body language is a very limited mode of access. It limits us to knowledge about occurrent emotions of people near us. So there is a social benefit to designing public signals that communicate emotions. Music is one of the media we employ for this purpose. Steven Feld has documented the social use of a particular falling melodic phrase by the Kaluli people of Papua New Guinea.

In a ritualized use of music, a Kaluli woman who is distressed by a death or another significant loss begins to weep. Then she begins to sing the distinctive descending phrase, transforming her weeping into singing. This musical phrase is adapted from the song of the fruitdove, which Kaluli interpret as genuine expressions of grief by birds, some of whom possess the souls of dead humans. Next, the woman improvises a song. This sung-weeping and singing sets off a prolonged event that ignores the performer–audience boundaries typical in modern musical culture. Other women respond with weeping and melodic wailing. Men, who are normally the musicians of Kaluli society, do not respond with melody. Instead, they weep hysterically, without singing. Feld explores multiple social benefits to this use of music, including the way it reverses normal Kaluli social patterns by demanding public confirmation of women’s emotions.

The functional value of conventionalized expression requires members of a society to agree on its expressive character. In contemporary life, for example, music is routinely attached to mass communication. A political candidate might use “The Soldier’s Joy” as background music in a television advertisement with a positive message about better times to come. The same tune will never be used as the background music for an “attack” advertisement that presents an opponent in a negative light, for that would send a mixed message that would confuse potential voters. Yet in order for music to be expressively useful in mass media, something in the general pattern of the music has to communicate its expressive character. It cannot always depend on some prior social conditioning that associates a particular melody with a particular expressive function—the way, for example, that the red octagon of the stop sign is a purely conventional symbol for stopping. If it is to be useful in an advertisement, “The Soldier’s Joy” has to be univocally expressive for millions of people who do not otherwise listen to bluegrass and Scottish folk music. Similarly, there would be little point to hiring composers to write original film scores if music does not possess distinctive expressive qualities that supplement and enhance the visual and narrative material. Suppose the ominous quality of the “Imperial March”—Darth Vader’s theme in the Star Wars films—is established through association, because it accompanies Vader’s appearance. If that is why the music is ominous, then there’s no point in hiring John Williams to write the music and then hiring an orchestra to play the theme, for the film audience will already have to know how they feel about Vader before the music functions expressively. The same point applies to songwriting. The melody that Woody Guthrie supplied to “Dear Mrs. Roosevelt” does not express his sadness about President Roosevelt’s death simply because the words express grief; if Guthrie had somehow shoehorned a lyric about Roosevelt’s death into the music of “This Land Is Your Land,” the song would simply create cognitive dissonance due to conflicting emotional expression in the words and the melody.

Examples like these show that music has features that make it useful for presenting expressive qualities and that people are generally very good at recognizing these musical features and the resulting expressive character. Previously, I argued that the production and reception of music is shaped by cultural tradition. If we combine these three points, they undercut the idea that “The music is joyful” always attributes an expressive function to the music in the strict sense of providing an externalization of a mental state. They demonstrate that an expressive quality of joy may be nothing but a byproduct of a culturally entrenched musical pattern, created for reasons other than expressing emotion. The musical features that prompt us to say that “The Soldier’s Joy” is joyful might be present for reasons that have absolutely nothing to do with the expression of joy.

“The Soldier’s Joy” was already at least a century old when Thomas Hardy described a performance of it in the novel Far from the Madding Crowd, written in 1874.

In rural England in the nineteenth century, the attraction of “The Soldier’s Joy” is its capacity to support dancing. At the farm dance described by Hardy, its expressiveness is not the point.

If we could go back in time and interview the anonymous composer of this reel, we would probably discover a Scottish fiddler. He might confirm my hunch that expressiveness was never the point. The three fiddlers at Weatherbury farm must have been exhausted—and perhaps bored, too—after playing the same reel for forty-five minutes. Perhaps the composer of “The Soldier’s Joy” was simply very bored of playing whatever reel was then popular in the Scottish highlands, and wanted something new to play at dances. Perhaps he just wanted to show off his musical skills. However, in creating a new reel, he also created something with a joyful expressive character.

As a general rule, a reel has a lively tempo in common time, with a heavy emphasis on the downbeat. Reels are commonly played with a bit of lilt, which is created by playing the downbeats a fraction early, in front of the anticipated beat. Although some reels are associated with sad titles and texts, such as “The Hangman’s Reel” (in French-Canada, “Le Reel du pendu”), they are inherently happy-sounding. I defer to Scottish poet Robert Fergusson:

So I challenge anyone to write a decent medium tempo or up-tempo reel with a Highland rhythm that is expressively sad. Given the inherent tendency of such reels to sound happy, the “happiness” of “The Soldier’s Joy” is no reason to conclude that the expression of happiness was an element of the compositional process. Because its social purpose in that society at that time placed limits on the composer’s creative process, and those cultural limitations directed it expressively, then some highly expressive music is expressive without being the expression of emotion.

III. The Song Thrush

In the last section I argued that the presence of the expressive quality of happiness in a Scottish reel is adequately explained by noting that it was created as an instrumental tune to support a particular kind of dancing. In turn, it follows that it is simply wrong to think that expressive qualities of music are always present as an expression of emotion.

My argument collapses if we assume that certain musical patterns are just like facial expressions. On this assumption, certain musical patterns are the natural, universal expressions of our basic simple emotions. Someone who is sufficiently aware of these patterns might choose to actively suppress personal expression, creating emotionally neutral music, the way that adult Japanese often suppress facial expressions. Or she might use the patterns to generate insincere expressions of emotion. But if a composer does not actively avoid expressing her own emotions, her musical compositions will employ patterns that express her emotions. Or, recognizing that the extended time required to compose music makes it unlikely that a piece of music expresses an occurrent mental state, the advocate of natural expression might endorse William Wordsworth’s proposal that poetry arises at a later time, as “the spontaneous overflow of powerful feelings: it takes its origin from emotion recollected in tranquility.”

On this revised account, “‘The Soldier’s Joy’ is joyful” means that the music reflects its composer’s recollection of some earlier happy occasion or the composer had a miserable life but knew how to be insincere. Fans of Ludwig van Beethoven’s “Pastoral” symphony (Op. 68) will point out how neatly this model fits Beethoven’s program for his “Recollections of Country Life” and its warning that it is “more an expression of feeling than painting.” Thus, the goal of Beethoven’s first movement is to express “the awakening of cheerful feelings upon arriving in the country.” Although I grant that Beethoven aimed at genuine expression in the sixth symphony, there are serious reasons to resist generalizing from the musical practices of nineteenth-century Vienna. It is a historically recent use of music. Projecting it back onto pre-modern music would be like thinking that the mourning garb of Victorian England is a natural response to death that was somehow repressed for all those preceding centuries. I will expand this objection in this chapter’s final section. Until then, I will discuss more obvious problems with attributions of expression.

If the presence of distinctive patterns justifies the attribution of joy to music, then there is no reason to stop with music. If this position is supposed to provide a general account of the expressive qualities of music, then it should be applied to all expressive qualities. We are justified in assigning a capacity for emotional response wherever we find patterns that “naturally” convey emotive expressiveness. But that invites endless, unjustified anthropomorphism about the world around us. Our tendency toward anthropomorphic perception invites further scrutiny of seemingly expressive non-human music: the “music” of birds. Perhaps birds make joyful-sounding vocalizations for some non-expressive reason.

The heart of the argument is that music is not the only realm of experience that inspires application of the names of basic emotions to non-sentient things. Spontaneous anthropomorphism, the attribution of human characteristics to non-human objects and events, is a basic strategy that we use for locating patterns in complex perceptual information. We engage in a mild anthropomorphism when we say that our car “refuses” to start on a cold morning, that the weather isn’t “cooperating” when it rains during our weekend at the lake, and that the stock market is “nervous.” From a scientific perspective, the Ojibwa people of central North America engage in a strong anthropomorphism when educated members of the tribe endorse the traditional belief that the four winds are living creatures and that thunder and lightning are the observable effects of the thunder birds, persons with avian form, traveling through the air.

Activities that unfold in time are strong targets of anthropomorphism. Activity suggests life, encouraging anthropomorphic perception of non-human forms as embodiments of purposeful activity. Since the detection of emotion is central to our perception of human agency, it is not surprising to find that basic emotions are routinely attributed to musical patterns that unfold in time. This would happen even if no music were ever used to express emotion. And it is not that music is particularly begging to be treated in this way (itself an anthropomor phism). Steven Feld notes that the Kaluli people typically describe musical patterns as waterfalls, pools of water, and other water phenomena. If that same metaphor was widely used in Western musical aesthetics, we might think about “The Soldier’s Joy” as a bubbling hot spring, totally side-stepping the question of what emotion it expresses.

Here we have our troubling double standard. Although scientifically literate people routinely use emotive vocabulary for cars, household appliances, the weather, and a host of other things, most of us have abandoned the theory that a “sad” willow tree and an “angry” rain squall express the emotions of some sentient agent. “Sad” music confronts us with an interesting problem concerning a widespread habit of thought and language, not a problem that is specific to music. The widespread belief that music is a special case is largely due to the continuing influence of a nineteenth-century Romantic aesthetic. I expand on this point in the final section of this chapter and then again in Chapter 4.

My position faces the obvious objection that music just happens to be a very special case. Granted, it’s a mistake to think that the weather is a person with human-like emotions. However, birds express themselves musically. Therefore their activity is independent evidence for the thesis that music is a natural method for expressing emotion, and therefore we are dealing with a very different phenomenon than people who say that the stock market is nervous.

My argument against this special pleading for music is another appeal to our tendency to anthropomorphize what we perceive. Animals are so-named because they are animate, self-moving, and so we find it very difficult to avoid assigning them “inner” psychological states and motivations that parallel our own. My quarrel is not with the idea that birds, reptiles, and non-human mammals have mental states, including emotions. I am quite confident that my dog is emotionally frightened by the sound of thunder and is happy when I return home from a trip. I have no quarrel with the idea that monogamous birds feel differently about their mates than about other birds. My objection is based on the obvious point that non-human species do not, for the most part, express their emotions in the manner of humans. I am hardly the first person to say that if you want to know if a Saint Bernard is feeling happy, its facial “expression” is irrelevant. All Saint Bernards have sad-looking faces. That is, they are sad-looking by human standards. But dogs don’t express happiness and sadness with facial expressions. They express emotion in their tails, stance, and mode of movement. If you’re looking at their faces, you’re falsely anthropomorphizing them. This warning is not specific to dogs. We cannot determine how any non-human species displays an occurrent emotion without carefully studying its patterns of behavioral response to different stimuli.

My argument against a universal, natural, cross-species proclivity to use music to express emotion is therefore quite simple. If you believe that Kaluli engage in anthropomorphism when they attribute grieving human souls to fruitdoves, then you should agree with me that the fruitdove’s sad song is not a symptom of genuine sadness. Granted, the song of the fruitdove sounds sad to human perceivers. But there is no more reason to think that fruitdoves are routinely sad and express that sadness with their songs than to think that Saint Bernards are naturally sad and express it with their faces. Since we cannot reliably determine how a dog feels by looking at its face, then we cannot reliably determine how a song thrush or a fruitdove feels by listening to its song. A “mournful” or “joyful” birdsong means no more than a “sad” dog’s face. Until we have very good evidence otherwise, the expressive quality of a bird song is like the expressive quality of a Saint Bernard’s face. We should resist anthropomorphic perception as a guide to their respective emotional states.

Rothenberg’s appeal to the beauty of birdsong is subject to the same censure. Peacocks cannot see most of the colors in the male peacock’s elaborate tail, and the male peacock’s horrible screaming may well be more beautiful than their tails to females of the species. Our sense of beauty may be very different from theirs.

We should be equally cautious about using familiar cultural practices as a basis for making conclusions about the “natural” proclivities of our own species. The fact that it became all the rage to express emotion in instrumental music in nineteenth-century Europe is no reason to think that music was, from the dawn of our evolutionary history, our natural tool for expressing emotion. We are so accustomed to the thesis that art has the function of expressing the emotions of artists that it can be very difficult to acknowledge that the idea is of recent vintage. Informed appreciation of Beethoven’s “Pastoral” symphony should acknowledge that it reflects the composer’s recollected feelings. But it doesn’t follow that the same holds for “The Soldier’s Joy.”

IV. Arousal Theory

Until now, I have focused on the thesis that expressive-sounding music expresses emotion by making someone’s emotion outwardly apparent. I have mentioned that this theory is of recent European origin. In contrast, the earliest Greek philosophers emphasize a different dimension of the expressive qualities of music. For example, when Plato discusses musical expression, he is far less concerned with the composer’s externalization of feeling than with the listener’s internalization of the music’s expressive quality. Naturally sorrowful musical modes should be banished from the state, for they make us self-indulgent, weak, and prone to feel sad in the presence of misfortune. The Confucian tradition of Chinese philosophy also warns about music’s expressive capacity and its attractions. Unlike Plato, Confucianism emphasizes a complete reciprocity between expression and reception. If musical choices reflect inner states, then hearing music will shape and control the inner states of listeners. Both Plato and Confucius conclude that the good of society demands censoring music that supports socially destructive emotions.

The thesis that sorrowful music makes listeners sorrowful and angry music makes them angry is logically independent of the idea that “‘The Soldier’s Joy’ is joyful” attributes joy to its composer. Instead, the music is joyful in the same way that an overcast day is gloomy, that is, as a cause of that emotional state in observers. As you may already have noted about Robert Fergusson’s praise of the Scottish reel, he endorses the reel as a superior cause of joy. We literally feel its influence. This theory about the meaning of “is joyful” in “the music is joyful” is traditionally known as the arousal theory, on the grounds that “is joyful” can be cashed out as meaning “arouses joy.” Arousal theory comes in two versions. The first, less plausible version holds that expressive descriptions indicate that the speaker is reporting his or her own response. The second version denies that “the music is sad” is necessarily a first-person report. However, neither version is attractive as a general theory, for there are many situations in which neither version captures what we are communicating when we expressively label music.

The first-person version might capture what some people mean when they talk about sad and joyful music, but it falls apart when we consider people who have something more to say than “it makes me sad.” Consider boredom, an emotion that I have mentioned only in passing. Suppose Professor Harmonious teaches music appreciation and she asks her students, “What did you think of ‘Trio,’ by King Crimson?” “It was peaceful,” one says. “It was boring,” another responds. It is plausible that the first student is reporting that the music had a calming effect and the second is reporting that it caused boredom. The difficulty is that, if Professor Harmonious continues the discussion and asks the second student whether the first student’s description is appropriate, the second student might very well endorse it. An individual can be bored to tears by a piece of music and yet know perfectly well that it is not expressing boredom. (There is nothing special about boredom. Annoyance is another emotion. I know people who find country music very annoying and yet who have no difficulty distinguishing between sad tunes, happy ones, humorous ones, and so on.)

To deal with the objection that the bored student is not reporting her own peacefulness, a second version of the arousal theory is available to preserve the theory’s core insight. When she agrees that “Trio” is peaceful, she is saying that the music typically or frequently produces a peaceful feeling in listeners. I will grant that this analysis captures what some speakers intend to communicate on some occasions, as when a record label packages music together as The Most Relaxing Classical Music in the Universe. It is a prediction that it will relax people. However, as a general account of what it means to say that a piece of music “is sad” or “is joyful,” this analysis fails on the grounds that it requires us to overlook much of the evidence that we have about music’s typical tendencies to arouse listeners. When it is not in actual conflict with our real estimation of the effects of listening to a particular piece of music, it reflects unjustified speculation. For example, if you listen to King Crimson’s “Trio” on the Starless and Bible Black album and to “Blue in Green” on the Miles Davis Kind of Blue album, and you find them boring, as a very great many listeners do, then what is your evidence for their calming effect on others? If you actually go looking for that evidence, I wager that you will discover that these two pieces of music bore most people—most people are not fans of these styles of music and are quickly bored by the placidity of these two pieces. If one needs another example, “amusing” appears to be a solid candidate for the revised arousal theory, as a label that indicates an expectation that a piece of music will amuse most people. Wolfgang Amadeus Mozart’s Ein musikalischer Spaß (K. 522, “A Musical Joke”) and Led Zeppelin’s “D’yer Mak’er” are two highly amusing pieces of music, but I say this with full knowledge that nine out of ten randomly chosen people will not hear the musical jokes.

To summarize the two great difficulties for arousal theory, many people assign expressive labels that do not reflect their personal response and without waiting to observe how others actually respond to it. However, if neither personal response nor general pattern of response is consulted in order to assign joyfulness to music, then the music itself must display an aural property that is appropriately described as “joyful.” The joyfulness of “The Soldier’s Joy” must be a property that emerges from its musical structure. Applied to music, an expressive label is a description of an emergent, gestalt property that the particular piece of music shares with other music that is described with the same term. It is like seeing that a vase has a fragile design without having to break it. On that basis the music may be predicted to be the source of emotional contagion for some listeners, or regarded as a suitable vehicle for self-expression if used in that way. However, theorizing that music’s expressive qualities are essentially bound to these functional uses of music is to let the tail wag the dog.

Some defenders of arousal theory will respond that I am being unfair, because I am ignoring basic ideas that I defended in the last chapter. Ignorant listening is not relevant. The responses that matter are those of listeners with sufficient familiarity to grasp the relevant musical structures. Most people have not acquired the right concepts for hearing what’s funny in Mozart’s “A Musical Joke,” and I have argued elsewhere that what is amusing about the Led Zeppelin piece is equally dependent on the right mix of musical and cultural information. The proper question, therefore, is whether informed listeners have the responses required by either version of arousal theory. My response has two prongs. First, if arousal theory is supposed to account for the natural expressiveness of sound, then the test group should be randomly selected people, drawn from people who listen to all sorts of music. Focusing on the responses of culturally informed audiences would be as misleading as demonstrating that there is a “natural” and universal sense of justice by questioning college graduates in the United States. Second, if we are going to defend arousal theory while abandoning the theory of natural expressiveness, then we introduce several new problems. For example, there is the problem of audiences who respond with “been-there, done-that, no-longer-interesting.” In other words, well-informed respondents are generally able to identify the emotional tone of a piece of music based on very little information. Then recall Brahms’s famous insult (“Every jackass hears that!”) when he wanted to make the point that what’s obvious about a piece of music is unlikely to be what’s musically interesting in it. Informed listeners frequently regard the music’s expressivity as a mere starting point for appreciation. They are generally more interested in the musical skill displayed in the composing and performing than in the act of expressing emotion. Eduard Hanslick appears to have been on the right track: if expression is elemental in music, it will be the same thing, over and over, and it will not sustain the aesthetic interest of more informed listeners.

V. Kaluli Grief, American Jazz, Hindustani Rasa

For the remainder of this chapter I return to the topic of music as art. For the sake of argument, suppose that there is a pure, original music and that three things are true about it. Suppose that we have some natural proclivity to perceive basic expressive qualities in many things, including musical patterns. (Many theorists push the story back a step and say that the true origin is the expressivity of vocal tones, which is then developed into music.) Suppose that, as a general rule, expressive musical patterns originate in the emotional lives of those who create those patterns. And suppose that, as a general rule, listeners respond to those patterns by experiencing, either genuinely or imaginatively, the emotions that are expressed. We now have a theory of why “The Soldier’s Joy” exhibits joy and we have an account of why there is such a strong consensus that it is joyful. However, we are not yet making any connection between expressivity and art. Therefore it is seriously misleading to point to a beautiful and “expressive” bird song and identify it as a species of musical art. Exhibiting symptoms is different from creating art.

The most telling difference between exhibiting an emotion and artistically expressing it is that art invites appreciation. Appreciation is different from mere liking or admiration. It is a complex evaluation of a human achievement within a particular tradition. In Chapter 2, I mentioned the Beatles’ “In My Life.” Liking the “harpsichord” sound is a much simpler response than appreciating its studio virtuosity. To clarify my argument, let’s consider a seemingly simple case of appreciating a musical performance.

Let’s return to Feld’s fieldwork in Papua New Guinea. Although Kaluli explain their use of the grieving weeping-into-song by pointing to its origins in the song of the fruitdove, Feld reports their interest in the recordings that he made of particular rituals. He was intrigued by the fact that many members of the village visited his house in order to request playback of a specific improvisation, in which a woman, Hane, addresses the spirit of Bibiali, a male cousin. The improvised song lasts just under five minutes. Why, Feld wondered, did so many people want to hear that particular song? As Feld interviewed people who sought out the recording, he found that they were drawn to its aesthetic achievement: “Kaluli found its construction controlled, deliberate, crafted, and almost composed like a song.” Although improvised, it was a strikingly superior articulation of grief. In other words, Kaluli do not treat all musical self-expression as equally valuable natural expressions. They listen for, and appreciate, the artistry of coherent articulation “within the constraints of an improvised form.” In short, Kaluli apply standards to Hane’s public grieving that are not applied to non-artistic expression of grief. Kaluli admire Hane’s performance in much the way that jazz fans applaud some performances of “Goodbye Pork Pie Hat,” Charles Mingus’s music to commemorate Lester Young, as aesthetically better than others.

Kaluli admiration of Hane’s song of grief is strikingly like the European Romantic preference for expressive specificity and particularization. However, artistic individualization is a criterion for artistic success that we do not apply to natural outpouring of basic emotion. When people praise the expressive improvisations of Hane and jazz musicians as direct and spontaneous outpourings of pure feeling that they cannot contain, the evaluation is condescending. Ted Gioia points out that aesthetic admiration of “primitive,” natural self-expression is a mainstay of jazz criticism. Yet what it really does, he notes, is treat the musician as “practitioner of an art which he himself scarcely understands. . . . Presented in such terms, the jazz performance seems hardly a cultural event and more like a medical affliction, akin to epilepsy.” An aesthetic evaluation of someone’s epileptic seizure objectifies and dehumanizes them. Evaluating Hane’s highly expressive music as an uncontrolled symptom of emotion is no better.

However, I must caution against generalizing too widely from the examples of Hane’s song and jazz improvisation. Both examples conform to expectations about musical expression that are associated with the European Romantic account of art. Earlier in this chapter, I quoted Wordsworth’s influential formulation about expressive poetry. Romantic ideals continue to influence art. As Jenefer Robinson points out, a generic “smiley face” has an expressive quality, but that does not make it an expression of emotion in the Romantic sense. The Romantic aesthetic endorses a high degree of individuation of emotion. Artistic expression is valuable when it precisely articulates a unique version of an emotion. Even when the emotions are those of a fictional character who is clearly not the artist, as in Bob Dylan’s “North Country Blues,” the expressivity is expected to reveal something new about the human capacity for emotion. Robinson observes that Romantic expression requires more than skilled making, for artistic expression requires the articulation “of new and unique emotional states.” Whether we find it in Kaluli appreciation of Hane’s song or in English poet Philip Larkin’s appreciation of Mingus, Romantic expression places a high value on the individuality of each person.

More importantly, assigning a shared criterion for success to Hane and Mingus assumes that personal expression is a functionally distinctive approach to expressive qualities of artworks. Appreciating the expressivity of a particular piece of music requires appreciating choices made against the background of musical and therefore cultural constraint. But we can also step back from the particular case and appreciate the tradition as an evolving collective choice to endorse expressive art. That choice is subject to revision. There would be no point to identifying and debating the Romantic theory of expression if it was the only game in town. Romantic expression is a cultural preference, not a musical universal. Some traditions endorse high levels of musical individualization without interpreting it as unique, personal expression.

In Chapter 2, I discussed audience confusion about the categories of tuning and performing at the start of Ravi Shankar’s performance during the 1971 benefit concert for Bangladesh. That mistake was easily corrected by Shankar’s joke, for the audience already possessed (but had incorrectly applied) the concept of performance. However, it is considerably more difficult to appreciate the expressivity of Hindustani classical music correctly, because it rejects the core value of Romantic expression theory. It questions the individuality of each person. The goal of musical performance is to exhibit rasa, which is a refined, de-personalized exhibition of the flavor of a basic emotion. When Shankar engages in an extended improvisation of a classic raga, his artistic goal is completely distorted by approaching the performance as personal expression. Nor is it offered as mere musical virtuosity. The Hindustani aesthetics sees nothing valuable about pure, expressionless music. Technical skill has a spiritual purpose. Standard Western interpretations, Shankar warns, are simply wrong.

Most Westerners cannot formulate—or, upon having it explained, cannot take seriously—the philosophical doctrine that individuality is a problem in need of a solution and part of the solution is to purge it from music. Although the point is often ignored by secular musicians and audiences in India, Hindustani classical art reflects the view that art is a path to enlightenment. In Hinduism, enlightenment is an escape from the “self” that loves, hates, suffers, and laughs. The central aesthetic principle of Hindustani art, Raj Kumar explains, is the presentation of impersonal, universalized emotion, so “the final effect that is left on [audience members] is not that of the passion depicted but is an impersonal absorption in the aesthetic mood.” Artistic expressivity is highly valued as a method of purging personal emotions. In this tradition, instrumental music displays rasa rather than bhāva. Crudely summarized, bhāvas are mental states, including the important subset of sthāyibhāvas, the basic emotions. Sorrow and anger are sthāyibhāvas. Their display is common in representational art. Michelangelo’s Pietà displays the bhāva of sorrow and it may arouse empathetic sorrow in viewers. But that is a superficial first step for the viewer.

Each of the basic emotions has a corresponding rasa (literally, “sap” or an extract or flavor that we can taste and savor). The point of listening to music is to “taste” its rasa. However, a rasa is not a bhāva. The difference is sometimes expressed with the metaphor that bhāva is to rasa as grapes are to wine. Hindustani aesthetics highlights the distinction between bhāva and rasa by assigning them different names. Sorrow (shoka) is an emotion, but its rasa is pathos (karuna). Although it is rarely found in music, the emotion of disgust (jugupsā) is allied with a rasa of repulsiveness (bābhatsa). More than four dozen bhāvas are recognized in Hindustani aesthetics, but only nine rasas.

However “expressively” complex a work might be in representing emotion, the real goal of art is to present the audience with a single, unified rasa. Established musical structures are merely blueprints. Music is energy, and so the musician must energize the melodic framework. To present rasa, Shankar explains, “The musician must breathe life” into the melody, “mak[ing] the bare notes vibrate, pulsate, come alive.” From a Western or Kaluli perspective, the resulting particularity of one of Shankar’s performances might appear to be his expression of a unique and personalized emotion. However, to “appreciate” it in that manner is to admire it without taking an interest in the performance’s true goal. Transcendence requires a skilled performer to extract and present a rarefied essence of an emotion, rather than mere self-expression.

We can tell all the stories we like about the natural evolutionary sources of music as an account of music’s expressive potentials. However, we must also recognize the difference between Jaco Pastorius’s gesture in opening his debut album with “Donna Lee” and Ravi Shankar’s aim of revealing the essence of erotic-romantic longing when he performs Raga Bhimpalasi. Unlike Shankar but like Hane, Pastorius was engaged in self-expression. In Chapter 1, I explained how Pastorius’s expressive gesture depends on its place in a musical tradition. He relies on cultural norms and expectations about music. When music is used as Romantic expression of a highly individualized emotion, that functional use relies on cultural expectation that does not apply to Shankar’s performances of traditional music. The contrast between Hindustani and “Romantic” improvisation illustrates that the expressive function of art does not lie in a universal human proclivity to use music in just one way. There are multiple musical arts with distinct expressive functions. Knowing which function is present in a performance requires a judgment about the music’s origins in a particular system of symbolic interactions and social preferences.

We can go one more step in the analysis. Individual composers and musicians can deploy expressive qualities in different ways on different occasions. Pastorius’s fretless electric bass is the musical highlight of one of my favorite records, Joni Mitchell’s Hejira. (Like Jaco Pastorius, Hejira was released in 1976.) Although his contribution is crucial to Hejira’s power, Pastorius is supporting Mitchell’s expressive gestures. His own emotions are not the point. So his musical exploration of emotion is functionally different from the personal gesture of arranging “Donna Lee.” The Hejira performances are more like Shankar’s relationship to Raga Bhimpalasi. Conversely, Shankar spent part of his childhood in Paris and a good deal of the middle period of his musical career working with non-Indian musicians. No stranger to Western music, Shankar’s three concertos for sitar and Western orchestra are, I suggest, best approached as his personal expressions of pride in his Indian heritage. Again, music is art because its meaning is inseparable from its cultural contexts.