Sad Songs Say So Much: The Paradoxical Pleasures of Sad Music: Sizer Sad Songs Say So Much

ABSTRACT

In this article I revisit the question of why people like to listen to sad music. If music can induce genuine sadness in listeners, why would we deliberately seek out such negative experiences? Drawing from work in both the philosophy and psychology of music, as well as work in the philosophy and science of affect, I argue to shift the focus of the question to music‐induced moods, not emotions. This reframes the debate but does not dissolve the puzzle. To understand what is appealing about the affective experience of listening to sad music, I suggest we take into account the unique features of music‐induced sad mood. I argue that sad mood and a certain sort of focused music listening are mutually reinforcing in ways that differ from other mood/music interactions. Sad mood and sad music are, in a sense, made for each other.

Listening to music can be an intensely moving experience. Many people love music in part because of its power to alter or amplify their moods and turn to music for inspiration, comfort, or therapy. It is a puzzle, then, why many of us spend so much time listening to sad music. If music can influence our moods, and assuming that most people would prefer to be happy, not sad, why choose to listen to sad music? Philosophers have wrestled with this question and similar ones about our taste for tragic, horrifying, or otherwise unpleasant art, leading Jerrold Levinson to call it “one of the hoariest [questions] in aesthetics” (1997, 216). In this article I revisit the question of why we like sad music, drawing from work in both the philosophy and psychology of music as well as work in the philosophy and science of affect. My contributions to this debate are, first, to shift the focus of the question to music‐induced moods, not emotions. This reframes the debate but does not dissolve the paradox. There is still a puzzle about why listeners would listen to music that brings about a sad mood state. In order to understand what is appealing about that particular musical experience I suggest we take into account the unique features of music‐induced sad mood. I argue that sad mood and a certain sort of focused music listening are mutually reinforcing in ways that differ from other mood/music interactions. Sad mood and sad music are, in a sense, made for each other.

I. GETTING CLEAR ABOUT THE NATURE OF THE PARADOX

The paradox about why we like to listen to sad music is an instance of a more general puzzle about why we like any art that deals with negative or unpleasant themes. It also gets tangled up in questions about how we come to perceive and appreciate the affective content communicated or expressed by artworks. The version of the paradox I am interested in is about affective experiences: why do we deliberately seek out experiences that induce negative affects in us? In particular, why do many people spend so much time listening to music that induces or amplifies negative affective experiences?¹ Levinson gives this problem the colorful moniker, the “paradox of musical masochism” (1997, 217).

In talking broadly about affective experiences I am deliberately expanding the scope of the problem to consider affective states other than emotions. The debate has tended to focus on musical emotions, but, as I argue, there is a more interesting and productive discussion to be had by thinking about moods.²

Let us begin by setting out the paradox in a set of claims that make clear the tension at the heart of the matter:

• 1.

  Sad music can induce or amplify sadness in listeners;

• 2.

  Sadness is an unpleasant, negative affective experience;

• 3.

  We strive to reduce or avoid unpleasant, negative experiences; and yet,

• 4.

  Many people frequently and deliberately listen to sad music.

In his discussion of it, Stephen Davies summarizes the issue this way:

It seems that one is trapped on the horns of a dilemma. If we enjoy the sadness that we claim to feel, then it is not plainly sadness that we are talking of, because sadness is not an enjoyable experience. On the other hand, if the sadness is unpleasant, we would not seek out, as we do, artworks leading us to feel sad. (1994, 307)

II. POSSIBLE SOLUTIONS TO THE PARADOX

A number of possible solutions to the paradox present themselves, focused on rejecting one or the other of the assumptions.

• 1.

  Reject the claim that music induces or amplifies sadness in listeners. Perhaps music qua music (that is, divorced from personal associations and memories) cannot induce real sadness in listeners. Perhaps music elicits only a sort of quasi‐sadness, which is not unpleasant.

• 2.

  Reject the idea that the sad affects engendered by music are purely unpleasant or aversive. Perhaps there are pleasurable aspects to the sadness, or the sadness is mixed with other more positive affects. Or perhaps listening to sad music brings about other pleasant or positive effects or outcomes that makes the negative experience worthwhile.

In what follows I proceed by first rejecting option (1) and demonstrating that music can induce or amplify real sad moods in listeners. While the sad moods elicited by music are perhaps more unalloyed versions of “everyday” moods in being abstracted from specific cares and consequences, they are not different in kind.

I then go on to discuss the role of affect in music listening more broadly and take up option (2), arguing that the affective experiences brought about by music listening have both positive and negative features. I show that all (good) music listening involves some moments of pleasure. Listening to sad music, therefore, is not purely unpleasant. While this helps explain why listening to music in general is pleasurable, it does not yet explain what might be particularly attractive about the experience of listening to sad music—why sad music seems to have a special hold on some. To address this, I turn to the nature of sad mood itself and argue that sad moods and sad music are a particularly potent and self‐reinforcing combination. In other words, the pleasures of sad music are, in part, because of, not in spite of, its capacity to elicit sad mood.

III. THE CASE FOR MUSICAL MOODS

Earlier versions of the debate tended to focus on whether music can induce emotions. Philosophers such as Eduard Hanslick (1986), Peter Kivy (1990, 2002), and Stephen Davies (1994) have argued strenuously that music cannot induce what Kivy refers to as the “garden variety emotions.” While I think that these arguments are based on a problematic theory of emotions and are therefore not successful (see Robinson 2005 for a thorough discussion of this), a more fruitful strategy is to refocus the debate on moods. In this section I argue that we should talk about musical moods, not emotions. Then I go on to argue that music can induce or amplify our moods.

III.A. From Musical Emotions to Musical Moods

I begin with a brief review of the arguments that music cannot induce emotions. Arguments for the claim often rely on a particular theory of emotions, the cognitive theory of emotions, that gives a defining role to cognitions or propositional attitude states. Since music cannot furnish specific cognitions necessary for emotions, it cannot cause emotions. Hanslick, for instance, states that

Only on the basis of a number of ideas and judgments (perhaps unconsciously at moments of strong feeling) can our state of mind congeal into this or that specific feeling. … [W]ithout this cognitive apparatus, we cannot call the actual feeling “hope” or “melancholy”; … If we take this away, all that remains is an unspecific stirring, perhaps the awareness of a general state of well‐being or distress. (1986, 9)

Kivy (1990, 2002) picks up and expands the argument by saying that music lacks the “logical machinery” to give rise to the specific beliefs, desires, and other propositional attitudes viewed as necessary for generating emotions in listeners. Furthermore, Kivy argues, music cannot provide an appropriate intentional object for emotions. I am always angry at or sad about something in particular; but music cannot furnish us with such objects. Therefore, Kivy concludes, music cannot induce the “garden variety” emotions in listeners.³

There are well‐rehearsed good reasons to reject the cognitive theory of emotions as an adequate theory of emotions in general (see, for example, Griffiths 1989). Other philosophers have used other, more biopsychological, theories of emotion to argue for the possibility of music‐induced affect (Robinson 2005, for example). These theories of emotion allow that emotions can be induced fairly automatically and without the involvement of specific cognitive states. However, while I agree that these are more successful theories of emotion, I think that the debate over affective responses to music is better framed in terms of moods, not emotions. Moods better capture our affective experiences of music.

Both emotions and moods are complex and dynamic states that have felt, physiological, cognitive, and behavioral elements. Moods are diffuse affective states that are not focused on particular intentional objects. Too, unlike a sad emotion, a sad mood is detached from specific beliefs or particular courses of action. As Richard Davidson puts it,

The primary function of emotion is to modulate or bias action. … The primary function of moods, on the other hand, is to modulate or bias cognition. Mood serves as a primary mechanism for altering information‐processing priorities and for shifting modes of information processing. (1994, 52)

In other words, as I have argued elsewhere, moods influence “how we think, not what we think” (Sizer 2000, 754). A sad mood, therefore, is a diffuse, objectless affective state that subtly modulates our thoughts, but not toward a particular topic or subject matter and not inspiring us to take a particular course of action.

Moods are nonetheless reflective of and responsive to the subject's overall state of well‐being (Morris 1992; Prinz 2004; Sizer 2000; Thayer 1989). Edith Jacobson (1957) described moods as “barometers of the ego.” Prinz argues that moods, like emotions, involve what he calls “embodied appraisals.” An embodied appraisal registers patterns of bodily changes, both conscious and unconscious, that “represent an organism‐environment relation that bears on well‐being” (Prinz 2004, 77). These appraisals are patterns of physiological change that “convey how we are faring in the world” (78).⁴ Different patterns of bodily change come to reflect and be associated with typical general eliciting situations: having what one needs and desires (happy mood) or a loss (sad mood). In this way the embodied appraisals of our moods represent general states of well‐being.

These features of mood have led several philosophers to conclude that when talking about affective responses to music, we should focus on moods, not emotions. Noel Carroll (2003) proposes that shifting to consider musical moods allows us to better capture the ways that music listening is deeply affecting without running into the problems with musical emotions. Colin Radford proclaims himself more of a “moodist” than an “emotivist” in response to Kivy's arguments (1991, 250). And Jenefer Robinson points to the objectless quality of moods along with their detachment from specific goals and actions as reasons to cast affective responses to music in terms of moods, not emotions (2005).

By shifting from emotions to moods I sidestep some (but not all, as I discuss below) of the prior debate on whether music can induce real “garden variety” affects.⁵ However, the shift to mood does not avoid the pull of the paradox. Sad moods are thought to be negative, unpleasant experiences. Therefore, we can still ask why someone would listen to music that can bring about that sort of affective experience.

III.B. How Music Can Induce Moods

In his later work Kivy extends his arguments against musically‐induced emotions to moods. In response to Carroll's (2003) proposal, Kivy wrote, “the prima facie evidence for absolute music's actually arousing or engendering moods in musical listeners [is] at best very thin. … Moods are in the music, not the man” (Kivy 2006, 275).

Kivy's principal criticisms here are, first, that there is not sufficient evidence establishing a mechanism by which music could induce moods. In particular, Kivy argues that while music might stir up some vague feelings, “there is no evidence absolute music has specific magic bullets for specific mood arousals” (276). Second, he argues that, to the extent that music could arouse moods, it must be because listeners are engaging in some “mind wandering” that takes them away from the music itself and brings in personal associations and memories that are responsible for the mood change. His worry is that it is not, then, the music qua music that brings about a change in mood.

Kivy's concerns set up two important conditions that a claim about music‐induced moods should meet.

• 1.

  Sad music can give rise to a real and specific sad mood, not a vague inchoate feeling or quasi‐mood.

• 2.

  It does so in virtue of musical properties of the music, not personal memories or associations. This makes sure that we are addressing the question of why it is sad music that people enjoy, not simply an occasion or prompt for personal reflection.

While one can always quibble over what counts as sufficient evidence, and these questions remain areas of active research on the empirical side, a strong claim can be made that sad music can induce or enhance sad moods. There appear to be multiple mechanisms or pathways by which music can influence affect. These range from fast, automatic, and cognitively impenetrable processes to more reflective and consciously elaborate ones (Bharucha et al. 2006; Juslin and Västfjäll 2008; Robinson 2013). In the next section I review the evidence and show how Kivy's conditions can be met.

III.C. Behavioral and Physiological Evidence

A number of philosophers have suggested that music listening creates an urge to move and that this is one way that music induces affect (see Davies 1994; Carroll 2003). We sway in time to music, tap our feet, and so on. Indeed, before the creation of that peculiar Western music tradition, the concert hall, which expects music listeners to sit still while they listen to music, music and movement were regularly linked.

Metaphors of movement and gesture certainly make their way into how we talk about music. As Robinson notes, “music can sigh and wail; it can freeze or frolic. It can creep menacingly or stride angrily. It can also mirror autonomic changes, as when an agitated or irregular rhythm mimics an agitated heart or irregular breathing” (Robinson 2005, 311). Aaron Ridley argues that affect‐expressive music includes “melismas” or “melismatic gestures” that mimic human gestures and movements. We sympathetically mirror the gestures in our own bodies, triggering an affective response (Ridley 1995). Davies has likewise described a “musical contagion” or “mirroring response” to music (1994, 2013).⁶

Robinson's claim is that music does more than resemble affective expression and action; it actually elicits bodily and brain changes that are similar to those associated with affective responses. These can be achieved by creating physiological changes in the listener unmediated by cognition or conscious attention. She labels this the “Jazzercise effect” (2005).

Robinson cites significant empirical support for this claim. For example, psychologist Carol Krumhansl and her collaborators found that listening to music induced a suite of physiological changes in listeners that are similar to the physiological changes associated with different affective states. Krumhansl (1997) found that “happy,” “fearful,” and “sad” music brought about distinct patterns of changes in heart and breathing rate, skin temperature, galvanic skin responses, and blood pressure. These changes were dynamic and cued to affectively salient moments in the musical pieces. In particular, sad music brought about relatively slower heart rates, slight elevation in blood pressure, decreased skin conductance, and decreased finger temperature.⁷ Krumhansl notes that while some aspects of these patterns of physiological change matched the physiological changes associated with full expressions of those emotions found in other studies (Ekman 1992, for example), in other respects they did not match. A study by Nyklicek et al. (1997) met with better results, finding that patterns of physiological change caused by sad music listening did mirror those of sad mood.

Discriminating between basic emotions or moods on the basis of physiology alone has been met with mixed results across multiple different induction methods, however. Better results are found when facial expression is included in the analysis. Witvliet and Vrana (2007) found that while listening to affectively valenced music, listeners made minute alterations in their facial expressions, consistent with the affects expressed in the music. In addition to the physiological changes described above, they found that sad music induced greater activity in the corrugator muscles—the muscles used for frowning. Subjects are not necessarily aware of these micro‐expressive changes, but they can be measured with facial electromyography. The facial expression results are particularly interesting as other research has established that facial expressions, including unconscious micro‐expressions, are particularly powerful means of influencing and inducing affect (Ekman et al. 1983; Strack et al. 1988).

Another study found that listeners also unconsciously mimic the vocal expression of music. Using fMRI, Stefan Koelsch et al. (2006) found that while listening to music, those areas of the brain related to the formation of premotor representations for vocal sound production became active. In other words, even though participants were not actually singing, some areas of their brains were, so to speak, going through those motions.

III.D. Neuroscientific Evidence

If we peer inside the brain, we see more evidence that processes associated with both motion and affect are engaged by music listening, including the early processing of features of musical structure such as meter, rhythm, and contour. This suggests not only that music engages affect, but also that these affective responses are vital parts of how we hear and understand both musical structure and meaning.

Daniel Levitin (2006) has found that the cerebellum, a brain structure associated with functions such as timing, gait, and other feats of coordinated motion, is active during music listening. The cerebellum, Levitin argues, is crucial for perceiving and understanding rhythm in music. It also projects to areas of the brain involved in affect, including the amygdala, frontal lobe, and nucleus accumbens.

Many of the brain systems involved in affect processing are also engaged during music processing, such as limbic system structures of the amygdala, hippocampus, and cingulate cortex (Blood et al. 1999; Juslin and Västfjäll 2008; Peretz 2001; Trost and Vuilleumier 2013). Brain regions associated with music and affect processing are located throughout the brain and overlap and interact. Music processing in the brain involves ongoing feedback loops between brain areas associated with musical structure, affect, and motion (Bharucha et al. 2006).

One of the most influential theories about affective responses to music has been Leonard Meyer's (1956) theory of musical expectancy. This claims that even untrained listeners come to music with a set of expectations—about typical chord progressions or how a passage of music tends to resolve within a particular musical style or genre. These do not require special training; they are acquired simply through experiences listening to music. Listeners are therefore very sensitive to the ways that music can violate, delay, or confirm their expectations about how the music will go, and these violations, delays, and confirmations can elicit affective responses. A violation of expectations can elicit surprise, bewilderment, and tension, which then gets dissipated or released pleasantly when the music resolves. Listeners tend to like music that contains some surprises and puzzles—is not too predictable—but also resolves those puzzles, leaving the listener with the sense that the mystery is resolved.

Psychologist David Huron (2006) has recently revived Meyer's theory and found evidence that the brain engages in the automatic generation of expectation, reaction to violation, and reaction to confirmation while listening to music and that each of these is accompanied by affective states of tension, release, and, as he calls it, “sweet anticipation,” making music listening a process that is both driven by and saturated with affect. These automatic micro‐affects caused by anticipation, violation, and resolution of expectations use the same brain areas implicated in positive and negative affective experiences and are part of how we attend to and process both the structural and macro‐properties of the music. These affects contribute to our experience of the music as sounding overall happy or sad, but they are not necessarily fully conscious to us; they unfold very rapidly (in concert with the music) and simply become part of our experience of the music.⁸

III.E. Cognitive Evidence

Evidence that music can influence affect also comes from psychology studies on the patterns of biases in judgment, attention, memory, and other cognitive systems associated with different moods. Indeed, this area of research assumes that music can induce particular moods. Psychologists studying affect often rely on the Musical Mood Induction Procedure (MMIP) to induce the desired affective state in subjects participating in an experiment. The MMIP, as the name suggests, uses short pieces of usually instrumental music to induce positive, negative, or neutral affective states in subjects before asking them to perform a task. The MMIP is considered to be one of the most effective methods of mood induction and, because the MMIP usually involves purely instrumental music, it avoids unwanted semantic or cognitive priming that may arise from other mood induction procedures (Västfjäll 2001). The use and apparent success of the MMIP is predicated on the idea that happy music can induce positive affects, and sad music can induce negative affects, and so on.

This research has found that positive/happy and negative/sad moods bias cognition in distinctive ways. Happy moods are correlated with more wide‐ranging, creative thinking, more far‐flung associations, greater use of heuristic and stereotype reasoning, and greater ease in recalling positively valenced concepts and memories. Sad moods are associated with more narrowly and inwardly focused attention, more detail‐oriented and specific reasoning, and greater ease in recalling negatively valenced concepts and memories.⁹ This research provides more detail to Davidson's claim that moods are generally associated with modulations in cognitive and information processing styles and not with particular thoughts or action tendencies.

III.F. The Multiple Musical Mechanisms for Mood

The micro‐affects generated by the music's playing with our expectations is one layer of our affective response to music. As Robinson (2005, 2013) points out, they help us understand what the music is conveying, but they do not determine the entirety of our conscious experience of the music. There are multiple mechanisms and layers of affective appraisal involved. We can reflect on and react to the dynamic unfolding of micro‐affects in a variety of ways, pulling from our knowledge (or lack thereof) of the particular piece, reflecting on the narrative flow of the piece as a whole. We can choose to attend more to certain structural properties of a piece (listening for a theme to be picked up and repeated by different instruments, say) or focus on the affective content (the way the piece moves back and forth between sadness and longing). Likewise, Juslin and Västfjäll (2008) underscore that these multiple mechanisms might be activated simultaneously at different levels or may feed back and feed forward to each other. Our conscious experience of the music is a result of the interplay of these different mechanisms and levels of appraisal.

Pulling together the different kinds of evidence one can see that there are multiple pathways by which music can influence or induce moods in listeners. While all music listening engages a mix of affective responses at multiple levels of activation and analysis, it can generate distinguishable physiological, neurological, and cognitive profiles for different moods, with the differences between happy and sad moods being the most robust and prominent. Furthermore, the affective responses to music are closely linked to intrinsic, structural features of the music itself. Indeed, the initial processing of musical structure, rhythm, and tempo seem to necessarily and automatically involve affect. Affective processing helps us hear and integrate these features of the music. Processing of larger scale structures and musical meaning also involve the sorts of embodied appraisals associated with moods.

III.G. When We Remain Unmoved

One possible counterargument to the claim that music can induce affect is to point to those times when it fails to do so. If music can quickly and automatically induce or amplify affects in us, then why are we not always and instantly saddened or gladdened by sad or happy music? But merely being in the presence of sad or happy music will not automatically bring about mood congruent changes that the listener consciously feels. We are not Frankenstein's monster, in thrall to the music.

To be saddened by sad music, we must first of all be attending to the music and be open to it. But even then a listener's overall affective state may not mirror that expressed by the music. As Robinson and others point out, our conscious experience of a piece of music is a result of the interplay of a number of affective and cognitive factors. The paradox of sad music arises for those cases when listeners are saddened by sad music. The fact that there are cases where listeners are not saddened (or only mildly saddened) or the fact that sadness is not an inevitable consequence of being exposed to sad music is beside the point.

Indeed, some cases where a listener's mood fails to mirror the affective content of a piece can serve to strengthen the argument that music can affect us. Listening to music that does not “match” your mood can sometimes be an irritating experience. Consider the last time you were in a sad or irritable mood and found yourself in the presence of upbeat, cheerful music; in my experience holiday shopping reliably brings about this set of circumstances. While you might, through careful attention to the music, find your mood momentarily lifted, it is more likely that you will feel increased irritation.¹⁰

I suggest that what one experiences here is what I call “affective dissonance.” As I noted above, listening to music and being in a particular mood state are both fully body and brain activities. To be sad is to be thinking, moving, and breathing in certain ways. Happy music plays upon our body and brain in ways that jostle and irritate our current set, however slightly; we are in the grips of opposing forces and this feels uncomfortable. The cases where we are irritated by happy music can be partly explained by and underscore the descriptions of music listening as involving affective and motion processing systems of the brain and body.

If one is happy, however, sad music feels like a drag, a downer. Perhaps it is slightly easier to bear, though, because one is happy. Perhaps the happy person is less drawn into the music and the feelings in his or her own body—a point I elaborate on later. So, while sad music can make the listener feel sad, and happy music can make the listener feel happy, it does not do so necessarily. One must be attending to and listening to the music and be open to the affective content expressed by the music.

This means that the manner and degree to which each person is moved by the music can differ. While affective processing is involved at the earliest stages of music processing, someone who is engaged in Kivy's “canonical listening”—listening focused on aspects of musical structure and form—may not feel that they are being affectively moved. Aspects of their affective response might be dampened and not amplified through attention, whereas other responses (an emotion of awe at the beauty of the artwork, as Kivy allows) might be heightened. How and whether we attend to music, and the features we choose to focus on, will influence our felt experience.¹¹

I conclude that Kivy's two criteria for musical mood induction have been met. The affects that are induced by music listening are not vague undifferentiated feelings or “unspecified stirrings”; they are specific and identifiable mood states. Most relevant to our inquiry, it is clear that music can induce sadness in a way that is distinct and distinguishable from other moods. Too, when music influences mood, it does so in virtue of formal musical properties of the music itself. A listener may, of course, bring in personal associations and memories, but these are not necessary to induce mood. To paraphrase Kivy, sad music can induce sad moods in virtue of the music's being sad.

IV. MIXED MOODS

Now I turn to the second strategy for resisting the paradox: rejecting the claim that the sad moods caused by music are purely negative or unpleasant. I argue that sad music induces a dynamic mix of affects that have positive and negative qualities to them. In so doing I will have to return to the question of whether these are, then, real sad moods. If it is not an entirely unpleasant experience, is it real sadness?

If Meyer and Huron are correct, then all music listening (at least, to music we find sufficiently engaging) gives rise to patterns of tension, anticipation, and the pleasures of “sweet release.” This explains why listening to good music is pleasurable; it does not explain why sad music is pleasurable or appealing.¹² After all, if all good music gives rise to these moments of pleasure, why not listen to good happy music?

We still need to understand why, if sad music induces sad mood, this is nonetheless an appealing or pleasurable experience. Here is where some version of the second approach is more fruitful. Sad music can induce or enhance sadness in listeners, but the experience is not wholly negative. Indeed, it can be uniquely positive and comforting.

IV.A. The Particular Pleasures of Musical Sadness

Some previously offered explanations for the pleasures of sad music point to various moral or psychological benefits that can accrue to the listener from the experience. Enduring the sad, negative feelings caused by sad music can be cathartic (Aristotle 2013), give one a sense of control (Eaton 1982), or be an avenue for exploring and understanding the human condition (Davies 1994). All of these are interesting suggestions, and some may contribute to making music‐induced sad mood a net positive experience. I do not think they adequately capture the experiences of listening to sad music, however. They all point to goals or outcomes that are a result of enduring the negative experience of music‐induced sad mood. But they do not say anything about why that experience might be pleasurable or attractive. Many people find that the experience of listening to sad music is not something to simply endure in order to get to a good outcome; the experience itself as it unfolds is attractive or positive. This is what we need to understand.

Levinson (1997) suggests that the sadness induced by music is more pure or unencumbered than the sadness we experience in response to everyday events. He argues that sad music, and other forms of sad or negative art, allows us to experience and explore negative affects uncoupled from real world consequences. We experience sadness, but a sadness divorced from our everyday cares and concerns. We are free to get swept up in the sadness without the anxieties of having to solve a problem or live with a loss.

This is supported by work psychologist Thalia Goldstein (2009) did looking at how subjects respond to real sad stories versus fictional ones. Narrative artforms (films and stories) can mirror real life events in many more ways than can music. And yet, Goldstein found that subjects responded to fictional sad stories differently than real ones. Subjects reported similar levels of sadness when watching sad films and recalling sad events from their own lives, but in the latter case their sadness was conjoined with anxiety—even when recalling sad events that were long past, such as the loss of a parent or a difficult divorce. Like Levinson, she argues that when we read sad stories or watch sad films we are able to experience what she calls “pure sadness”—sadness without the anxiety that usually attends unhappy events in our own lives.

Furthermore, when one is consuming a tragic story or film one approaches the task with a different attitude than one would take to, say, reading a newspaper story. Rolf Zwaan (1994) argues that consumers of fictions mute the appraisal system that demands that we ask skeptical questions and integrate the information with other facts. When our information gathering and fact checking approaches are dialed down, we are free to engage with and respond to the story and characters. When we approach stories in this way, Zwaan argues, we pay more attention to the structure and form, the “surface structure of the text” itself, and not just the information it delivers (1994, 922).

Sad and tragic artworks invite the consumer to experience affects that are decoupled from the messy complexities of everyday life. We are free to be drawn into the artwork and engage with its formal aesthetic properties and to experience a pure, unalloyed form of sadness—sadness with all the aching melancholy of everyday sadness but divorced from its anxieties and complexities.

A skeptic might object that the fact that musical sadness differs in this way from everyday sadness is a reason to think of it as mere quasi‐sadness, not real sadness. In response, let us begin by recalling what the “mere quasi‐sadness” move is supposed to accomplish. It was proffered as a way to resolve the paradox, but only insofar as quasi‐sadness differs from real sadness in being a not‐unpleasant experience. But the arguments and evidence just reviewed supports the claim that the sad mood brought on by music listening is unpleasant in many of the ways that real sadness is unpleasant. It feels like sadness; it aches and hurts like sadness. So, calling this quasi‐sadness would not dissolve the paradox.

Davies makes a similar point:

The artistic context of appreciation might strip the emotional response of some of its aspects—desires and the need to act on them, the vitality of the feeling tone of the experience, and so on—but it cannot remove altogether the unpleasantness that is in part constitutive of negative emotions. (1997, 244–245)

There are other reasons for thinking of musical sadness as real sadness, not quasi‐sadness. The moods generated by music share the same physiological and brain mechanisms as everyday moods (Juslin and Västfjäll 2008). Too, the mood congruent cognitive biases brought on by music listening (as with the MMIP) are similar to those brought on by other mood induction procedures. Here again the difference lies not in the sadness itself but what it is usually conjoined with in everyday situations. Everyday sadness—sadness caused by perceptions of real loss or failure—is usually mixed with other affects such as fear or anxiety. This is consistent with what psychologists using the MMIP have found: MMIP influences subjects’ sadness and despondency scores, but not their anxiety scores (Västfjäll 2001, 185). This supports the point made by Levinson, Davies, and others: music can bring on a real (not quasi‐) sad mood. It differs from our everyday experiences of sadness only in not being mixed with other affects such as anxiety.

The suggestion so far is that the unalloyed sadness brought on by listening to music allows the listener to stay with and attend to the music and get caught up in both its structure and meaning. This leads to the play of micro‐affects described by Meyer and Huron that attend the processing of musical expectations and consummations. This account so far gives an explanation of why the sad mood generated by music does not interfere in enjoyment; we are not motivated to respond to real‐world concerns and instead are free to engage with and respond to the music itself. We still do not have an explanation for why sad music is particularly enjoyable for some people. For this I turn to the nature of sad mood.

IV.B. Sad Moods and Sad Music: A Powerful Feedback Loop

My central claim is that the cognitive and attentional biases associated with sad mood actually encourage and reinforce the sort of focused attention on the music that allows us to engage with it fully. Sad moods, as compared to happy or neutral moods, are associated with a narrowing of attention, a turn inward with a greater focus on the self and one's feelings as opposed to events in the outside environment (Wood et al. 1990). Sad moods also promote an increase in more detail‐oriented, analytic thinking as opposed to more wide‐ranging heuristic thinking (Isen and Daubman 1984; Schwarz and Bless 1991). Sad moods, therefore, encourage and promote the sort of focused music listening that enhances the listener's engagement with and experience of the music. Sad mood and sad music listening, therefore, are mutually reinforcing. Sad music enhances the very processes and behaviors that keep one wrapped up in and engaged with the music itself.

This, I argue, is unique to the sad mood/sad music relationship. Insofar as happy music induces a happy mood, it encourages the listener to move, to explore, and to reach out and make connections with others. One's attention is expanded, and thought processes are more likely to move away from the music itself or one's experience of it, to envelop outside events and other people. This does not mean that one cannot stay engaged with and listen closely to happy music or that happy music resists more formal analysis or attention. Our affective responses to music do not dictate all of the elements of our engagement with it. We can choose to attend to the music (or not) in many different ways.

The mood responses to music exert a subtle pull on how we continue to engage with and hear the music. A happy mood brought on by listening to happy music has a tendency to nudge our attention away from the intricacies of the music; perhaps the up‐tempo or expansive, soaring qualities of the music encourage the listeners to move, dance, get “out of their heads” and into the world. But sad moods can pull us in and keep us wrapped up in the music itself and the beautiful unalloyed sadness it conveys. Sad music and sad mood can engage a unique feedback process. This helps explain the findings that people who report the most intense affective responses to sad music also report liking it the most (Vuoskoski et al. 2012).¹³

There is some interesting neuroendocrine activity that adds another layer to this feedback hypothesis. Huron argues that for some people (people who particularly like sad music), listening to sad music brings about increases in levels of prolactin. Prolactin is a hormone that is released, most notably, during lactation (as the name suggests) but is produced in both males and females after sex and at lower levels during grief, stress, and also during empathetic sadness.¹⁴ Prolactin is present in tears, but only during “psychic crying”—crying due to psychological, not physical, distress. Prolactin produces feelings of calm, tranquility, well‐being, and consolation (Gračanin et al. 2014). Prolactin's release during these events is thought to play a homeostatic function similar to the release of endorphins to physical pain—a mechanism to regulate and return the self to equilibrium.

Huron suggests that the sad moods caused by listening to sad music can trick the body into a homeostatic response—releasing prolactin as a way of moving toward consolation and well‐being. Therefore, feeling sad while listening to sad music can induce feelings of consolation and relaxation that are comforting. This is consistent with people's reports that they listen to sad music for comfort or consolation (Hanser et al. 2016).

The arguments and evidence suggest that sad music is simultaneously saddening and comforting. It allows the listener to experience a pure, unalloyed sadness that modulates the processing styles of the listener in ways that keep attention engaged with the formal and structural properties of the music itself. It also engages mood repair mechanisms that help listeners feel comforted and consoled, encouraging them to stay with and within the music. The engagement of mood repair mechanisms that bring on feelings of comfort and consolation seem to be unique to the sad mood/sad music relationship.

V. CONCLUSION

I have argued that sad music is uniquely pleasurably affecting in a number of ways. Sad music that plays with our anticipations and relief in the ways that Meyer and Huron described is pleasurable. The sort of sadness we can experience while listening to it is unalloyed in the sense that it is divorced from the stress and anxieties that attend real world negative events. And yet it is sufficient to engage the homeostatic mechanisms that work to make us feel consolation and tranquility. The sad mood brought on or enhanced by listening to sad music also tends to keep us focused on and engaged with the music itself. This is the sense in which sad music and sad mood are made for each other. They engage feedback loops that reinforce each other. We can be drawn into the music and kept there by our moods, which further enhances our engagement with the music. Listening to sad music therefore is a uniquely rich, affectively complex, and comforting experience.¹⁵

REFERENCES