Current Emotion Research in Music Psychology

Universal Features of Emotion in Music

Does music from different cultures share common features? Evidence that there are universal cues to emotion in music comes from several key findings. One is that general acoustic cues (i.e., those that are not specific to music or to a specific genre of music) contribute to the emotional expressiveness of music. In fact, some cues to emotion appear to be relevant to all forms of human vocal communication, such that there are notable commonalities in the acoustic features of emotionally expressive speech and music (Juslin & Laukka, 2003). For example, angry speech is characterized by high vocal intensity, just as angry sounding music tends to be loud. High-arousal emotions (anger, fear, and happiness) are expressed using faster rates compared to low-arousal emotions (sadness and tenderness/love) in speech and in music. Moreover, happy sounding speech in English or Tamil has larger pitch intervals than sad sounding speech, just as happy sounding music in South Indian and Western music incorporates larger pitch intervals (Bowling, Sundararajan, Han, & Purves, 2012). One exception is that a lower overall pitch makes music sound less pleasant (Jaquet, Danuser, & Gomez, 2014), such that low pitches are used to express emotions with negative valence. In speech, negative emotions such as fear and anger tend to have a relatively high pitch (Juslin & Laukka, 2003).

Young children use these general acoustic cues to make judgments about emotions expressed through music before they start using culture-specific cues. Consider tempo (speed). Faster pieces are judged to be happy or positive sounding, whereas slower pieces are deemed to sound sad or negative (e.g., Gagnon & Peretz, 2003; Juslin & Lindström, 2010). Even 5-year-olds perceive and successfully decode tempo as a cue to emotion (Dalla Bella, Peretz, Rousseau, & Gosselin, 2001). They are not sensitive to changes in mode (major or minor), however, because mode is a culture-specific cue in Western music that becomes meaningful after 6 years of age (Dalla Bella et al., 2001). When asked to convey emotions by singing, 5- to 8-year-olds also modulate basic acoustic cues such as tempo, pitch, and loudness in their performances (Adachi & Trehub, 1998).

Culture-general cues allow listeners to identify emotions in music from other cultures with modest accuracy. For example, Western listeners are able to identify happiness, sadness, or anger in Hindustani (North Indian) pieces, but they have difficulty recognizing peacefulness (Balkwill & Thompson, 1999). Similarly, Japanese listeners can identify joy, sadness, and anger in unfamiliar sounding Hindustani music, as well as in more familiar sounding Japanese and Western music (Balkwill, Thompson, & Matsunaga, 2004). Listeners who have never heard Western music (the Mafa from Cameroon) can identify happiness, sadness, and fear in excerpts of Western music at above-chance levels (Fritz et al., 2009). Finally, Indian, Japanese, or Swedish listeners can identify emotions expressed in Western, Hindustani, Japanese, or Swedish music, although performance is more accurate for basic emotions (e.g., anger, fear, happiness, sadness) than for complex emotions such as solemnity and spirituality (Laukka, Eerola, Thingujam, Yamasaki, & Beller, 2013).

Culture-Specific Features of Emotion in Music

Different musical systems have different structures and ways of conveying musical meaning. In Western music, for example, listeners associate major and minor modes with positive and negative emotions, respectively (e.g., Gagnon & Peretz, 2003). Enculturation to a particular culture’s music is a developmental process in which associations and regularities are internalized, at least implicitly, through repeated exposure. One view holds that the experience of emotion depends on whether culturally learned musical expectancies are fulfilled or violated (Huron, 2006; Meyer, 1956), which implies that listeners should exhibit biases favoring the music of their native culture.

Indeed, listeners are better at perceiving emotions conveyed in the music of their own culture than in unfamiliar music (Balkwill et al., 2004; Fritz et al., 2009; Laukka et al., 2013). In the cross-cultural studies noted above, Mafa listeners performed markedly poorer than Western listeners at identifying emotions expressed in Western music (Fritz et al., 2009), and Swedish and Indian listeners exhibited performance advantages for music from their native culture (Laukka et al., 2013). Listening to music from a foreign culture may also be less emotionally rewarding than listening to culturally familiar music. For example, American participants rate Indian music as “tenser” than Western music (Wong, Chan, Roy, & Margulis, 2011; Wong, Roy, & Margulis, 2009), whereas participants from rural India rate Western music as “tenser” than Indian music (Wong et al., 2009).

Emotions that are perceived reliably in music change over development, which implicates a role for learning and exposure, which would vary across cultures. For example, 3- to 4- year-olds are more accurate at perceiving happiness and sadness compared to anger, fear, and neutrality in music, whereas 5- to 7-year-olds can identify happiness, sadness, fear, and neutrality (but not anger) at above-chance levels (Stachó, Saarikallio, van Zijl, Huotilainen, & Toiviainen, 2013). In general, high-arousal emotions such as happiness and fear are easier for young children to identify than low-arousal emotions such as sadness or peacefulness, but by 11 years of age children are as good as adults at identifying emotions expressed musically (Hunter, Schellenberg, & Stalinski, 2011). Older adults exhibit impaired recognition of negative emotions such as sadness, but they recognize positive emotions in music as well as younger adults do (Laukka & Juslin, 2007). The ability to identify negative emotions in music begins to decline in middle age, whereas the ability to recognize positive emotions is stable across the lifespan (Lima & Castro, 2011). Most of the developmental research on music and emotion has focused on Western music, however, which makes it impossible to confirm whether developmental patterns differ across cultures.

In any event, these findings confirm that affective meaning in music is influenced by exposure and learning. In the past, some scholars took the extreme position that there is nothing inherently or naturally meaningful in music and that all affective meaning in music is a product of cultural interpretation (e.g., Walker, 1996). This view is belied, however, by evidence that people identify emotions in music from other cultures with some degree of accuracy. In short, the emotional expressiveness of music arises from a combination of culture-specific and universal cues.

Mechanisms of Emotion Induction

Do affective responses to music result from something intrinsic to the music, or are they the result of cognitive appraisals, which are deemed to be central to emotional responding in general (Ellsworth & Scherer, 2003)? Some theorists suggest that music induces emotions without the need for cognitive appraisals. For example, one view holds that musically induced emotions are the result of automatic mirroring by listeners of the motor movements musicians use to communicate emotion (Molnar-Szakacs & Overy, 2006). Automatic mirroring is perhaps served by the mirror-neuron system, which spans cortical parieto-frontal motor regions (Rizzolatti & Sinigaglia, 2010; cf. Hickok, 2009). In line with this view, activations in cortical motor regions are evident during music listening (e.g., J. L. Chen, Penhune, & Zatorre, 2008; Koelsch, Gunter, Zysset, Lohmann, & Friederici, 2002; Maess, Koelsch, Gunter, & Friederici, 2001). As noted, however, the emotion listeners perceive in response to a musical piece is not always identical to the emotion experienced. Thus, mirroring cannot be the only mechanism of emotion induction.

Juslin and Västfjäll (2008) acknowledge the role of cognitive appraisals in emotional responding to music, but they propose six additional mechanisms that could lead to emotional responses without the need for cognition: (a) automatic brainstem reflexes to musical events such as loud, unexpected, or dissonant sounds, (b) evaluative conditioning—certain pieces are associated with pleasurable or negative events and evoke the same emotion as the original event, (c) emotional contagion—the emotion perceived by the listener is also experienced, (d) visual imagery associated with music, (e) associations between a piece and emotionally charged episodic memories, and (f) violation or fulfillment of musical expectations. Although Juslin and Västfjäll (2008) claim that these six mechanisms circumvent the need for cognitive appraisal, phenomena such as mental imagery and episodic memory are not independent of cognition (Konečni, 2008b). Whether emotional reactions to music can occur without cognitive appraisals is reminiscent of the long-standing debates in emotion research (e.g., Lazarus, 1982; Zajonc, 1980) and unlikely to be resolved in the near future.

Which Emotions Does Music Evoke?

One relatively common emotional response to music is the pleasant experience of chills, shivers, or piloerection (i.e., goosebumps; Grewe, Katzur, Kopiez, & Altenmüller, 2010). For example, half of the sample studied by Goldstein (1980) reported experiencing music-induced chills. Musicians are more likely than nonmusicians to experience the phenomenon (Sloboda, 1991; but see Grewe, Kopiez, & Altenmüller, 2009), and people who score high on openness-to-experience are more likely than other individuals (McCrae, 2007; Silvia & Nusbaum, 2011). Huron (2006) proposes that chills are an automatic fight response in reaction to surprising stimuli that are judged initially to be threatening. Because a piece of music is non-threatening (unless it is deafeningly loud), the response turns to one of pleasure. Indeed, chills are often reported during unexpected or sudden musical events, such as sustained high pitches, unexpected harmonies, or sudden changes in loudness (Guhn, Hamm, & Zentner, 2007; Panksepp, 1995; Sloboda, 1991). Chills are also reported more often in response to slow rather than fast-tempo music (Guhn et al., 2007).

Chills that are evoked in response to self-selected music are accompanied by physiological changes indicative of emotional arousal, such as increases in electro-dermal activity (EDA), heart rate, and respiration rate, but decreases in body temperature (Grewe et al., 2009; Rickard, 2004; Salimpoor et al., 2009). The experience of chills also coincides with patterns of limbic and mesolimbic activity associated with reward processes (Blood & Zatorre, 2001; Salimpoor, Benovoy, Larcher, Dagher, & Zatorre, 2011).

Although chills represent a particularly intense emotional response to music, emotional responses to music vary widely in terms of quality and quantity (or intensity). Much debate centers around whether music evokes the same emotions as other stimuli such as faces or stories. In other words, does music induce utilitarian or everyday emotions (happiness, sadness, fear, anger, surprise) or does it induce strictly aesthetic emotions (such as wonder, awe, nostalgia, etc.)? Utilitarian emotions involve goal-relevant cognitive appraisals that motivate adaptive action tendencies, whereas aesthetic emotions are not goal-relevant and involve feelings of subjective pleasure in response to the structural characteristics of the stimulus itself (Scherer, 2004).

In a series of studies, Zentner et al. (2008) factor-analyzed participants’ reports of felt emotions in response to music. If music induces utilitarian emotions, the solution should have revealed factors corresponding to such emotions as joy, anger, fear, disgust, and contempt. Instead, the analysis revealed nine dimensions that are more aesthetic than utilitarian in nature: wonder, transcendence, nostalgia, tenderness, peacefulness, joyful activation, tension, sadness, and power. Nevertheless, a majority of the participants were listeners who preferred classical music, a genre that may be more strongly associated with aesthetic emotions than other genres (e.g., heavy metal music may be more strongly associated with anger, the blues with sadness, etc.).

Aesthetic emotions can also occur in everyday situations. For instance, descriptions of aesthetic emotions appear to overlap in some respects with descriptions of flow experiences (Csikszentmihalyi, 1990), which arise in aesthetic (e.g., viewing a painting; Csikszentmihalyi & Robinson, 1990) as well as everyday (e.g., formal education; for review see Nakamura & Csikszentmihalyi, 2002) contexts. By definition, the goal of an activity becomes irrelevant during flow experiences, such that simply engaging in an activity (e.g., listening to music) gives rise to pleasure and intense absorption. Nonetheless, the constructs of flow (a diffuse positive affective state) and aesthetic emotions (feelings such as transcendence or tenderness) are not identical. In short, it is unclear whether music induces utilitarian or aesthetic emotions, and even whether aesthetic emotions are uniquely aesthetic.

A final issue is the ability of music to induce more than one emotion in the listener simultaneously. Fast-tempo music is considered to sound happier than slow-tempo music, just as major and minor modes are happy and sad sounding, respectively. By varying tempo and mode independently, it is possible to create music with conflicting cues (songs with fast tempo and minor mode or slow tempo and major mode), which elicits mixed emotions more so than music with consistent cues (fast and major, slow and minor; Hunter, Schellenberg, & Schimmack, 2008, 2010; Ladinig & Schellenberg, 2012; Larsen & Stastny, 2011). When participants’ responses are recorded continuously as they listen to music with conflicting cues, “happy” and “sad” response keys are pressed at the same time, which suggests that they are experiencing happiness and sadness simultaneously and not simply vacillating between the two emotions over time (Larsen & Stastny, 2011).

Emotional responses are often measured along two continuous dimensions, one assessing arousal (low to high), and the other assessing valence (negative to positive), in line with the circumplex model, which posits that all emotions can be represented as points in two-dimensional (arousal by valence) space (Russell, 1980; Russell & Carroll, 1999). Because positive and negative emotions lie at opposite ends of the bipolar valence continuum, they cannot co-occur. Unlike the circumplex model, the evaluative space model (Cacioppo & Berntson, 1994; Cacioppo, Gardner, & Berntson, 1997) proposes that negative and positive valence can be activated independently and therefore simultaneously. Findings of mixed affective responses to music are thought to provide support for the evaluative space model.

Self-reports of mixed emotions may be contaminated by confusions between perceived and experienced emotions, although listeners have been shown to distinguish mixed perceptions from mixed feelings (Hunter et al., 2010). Future research could determine whether self-reports of mixed emotions are reflected in physiological responses. EMG is likely to be particularly useful in this regard because it is sensitive to variations in valence (Khalfa et al., 2008; Witvliet & Vrana, 2007).

Liking Sad Sounding Music

Although adult listeners tend to prefer happy sounding pieces, people often seek out and enjoy sad sounding pieces. Individuals who score high on measures of empathy (Garrido & Schubert, 2011; Kreutz, Schubert, & Mitchell, 2008) or openness-to-experience (Ladinig & Schellenberg, 2012; Vuoskoski, Thompson, McIlwain, & Eerola, 2012) are particularly likely to enjoy sad sounding music, as are introverts (Ladinig & Schellenberg, 2012), and people who score high on a measure of “absorption” (Garrido & Schubert, 2011, 2013a; Kreutz et al., 2008). Music training is also predictive of intense emotional responding to sad-sounding music (Park et al., 2014).

When individuals are in a sad mood, they show mood congruency effects: increased liking for sad sounding music, and increased perceptions of sadness in music that is selected to sound neutral (Hunter, Schellenberg, & Griffith, 2011). Individuals with clinical depression may be especially likely to listen to music that expresses negative valence because it matches their chronic mood state (Wilhelm, Gillis, Schubert, & Whittle, 2013). In some instances, participants in a sad mood are less likely than participants in happy or neutral moods to choose to listen to happy sounding music, presumably because it is a poor match to how they are feeling (Friedman, Gordis, & Förster, 2012; Taylor & Friedman, 2014). A contrast effect has also been reported—preferences for sad sounding music increase after participants listen to multiple (but different) happy sounding pieces consecutively (Schellenberg, Corrigall, Ladinig, & Huron, 2012).

In general, people avoid situations associated with negative affect, which makes liking sad sounding music a puzzling phenomenon. One view holds that negative affect is inhibited in aesthetic contexts, such that it is pleasurable to experience any emotion when listening to music (Garrido & Schubert, 2011; Schubert, 1996). In line with this view, pieces that are perceived to sound sad also tend to be judged as beautiful (Eerola & Vuoskoski, 2010), and exposure to sad sounding music can improve sad mood states (Tsai, Chen, & Tsai, 2014; van den Tol & Edwards, 2014). Conflicting evidence emerges from behavioral and physiological studies, however, which often report that listening to sad sounding music does indeed induce negative emotions (e.g., Blood et al., 1999; Garrido & Schubert, 2015; Green et al., 2008; Juslin, Liljeström, Laukka, Västfjäll, & Lundqvist, 2011; Mitterschiffthaler, Fu, Dalton, Andrew, & Williams, 2007; Vuoskoski et al., 2012).

Huron (2011) argues that sad sounding music induces genuine sadness accompanied by elevated levels of prolactin, a hormone that is associated with comfort and consolation, and released during episodes of sadness. Another possibility is that people enjoy listening to sad sounding music not because negative emotions are inhibited but because sad music simultaneously evokes both positive and negative emotions. In line with this view, participants report positive emotions such as nostalgia and wonder in addition to sadness when they listen to sad sounding music (Vuoskoski et al., 2012).