Music and emotion

Many scientific disciplines deal with the topic of music and emotion, including philosophy, musicology and psychology. The perspective presented here is mainly a psychological one, yet some theoretical and philosophical considerations will be made to clarify prevailing concepts about music and emotions and how they can be connected.

Expressiveness of music – philosophical problems

Claiming that music is expressive of emotions and that it can elicit emotions in the listener does not seem highly disputable at first glance. However, this claim gives rise to a number of questions.

1. How can a piece of music (when we consider purely instrumental music without any vocals, text or title) appear emotional, as a piece of music is no psychological agent?
2. Why would we respond emotionally to music knowing that there is nobody undergoing the emotion expressed?
3. What are psychological mechanisms that lead to the emotional reaction in the listener?
4. What is the nature of these emotions?

The first question deals with how emotions are transported in the music, questions 2-4 with emotions in the listener. (Not mentioned here are emotions in the composer or the performer.) However, perceiving a piece of music as to be emotional and being moved by this emotion mostly go in hand.

We don’t find it hard to explain why and how we respond emotionally to something expressing an emotion, e.g. a person expressing joy or sadness (or indirectly to an event like an earthquake that affects people as to express an emotion, which ends up being the same). A stone rarely moves us to tears, so why would music do that? Thus, the core of this problem is the question how music can be expressive at all. This problem is examined by the field of aesthetics.

Appearance emotionalism

Two of the most influential philosophers in the aesthetics of music are Stephen Davies and Jerrold Levinson. Without going into the depths of the philosophical argument, this view mainly follows Davies’ position. He terms his concept the expressiveness of emotions in music appearance emotionalism. Appearance emotionalism holds that music is for example sad in the same way the posture of a person is sad or a weeping willow is sad. A piece of music is not sad because it feels sadness, but because it expresses sadness, it is sad in appearance.

Why does something (that is not a person) appear sad? Because we can identify in its structure certain characteristics that we know from a person’s expression of sadness. We would sometimes call an old hunchbacked lady sad (although we don’t doubt that she might feel completely differently) because she looks like someone sad we’ve already seen. In the same way we would call a piece of music sad because its dynamic character resembles a person’s expression of sadness. “The resemblance that counts most for music’s expressiveness […] is between music’s temporally unfolding dynamic structure and configurations of human behaviour associated with the expression of emotion.” If a person does not give verbal account of his or her feelings , the observer can still note them from the person’s posture, gait, gestures, attitude, and comportment. Music recalls an appearance of sadness e.g., according to Davies, by a slow and quiet downward movement, underlying patterns of unresolved tension, dark timbre , heavy or thick harmonic bass textures.

Not everybody associates the same musical features with the same emotion. Appearance emotionalism does not claim that movement in music generally resembles human behaviour but that many listeners have this perception of similarity, and that this is the crucial connection that constitutes the expressiveness of music. This perception of similarity can be widely common among listeners or highly individual. Which musical features are more commonly associated with certain emotions is left over to the testing of music psychology (see next paragraph). Davies claims that expressiveness is an objective property of music and not subjective in the sense of being projected into the music by the listener. Music’s expressiveness is certainly response-dependent, i.e. it is realized in the listener’s judgement. However, suitably skilled listeners display a high degree of agreement in attributing emotional expressiveness to a certain piece of music. Although this is an empirical finding, it indicates according to Davies (2006) that the expressiveness of music has to be somewhat objective. If there was no expressiveness in the music, no expression could be projected into it as a reaction to the music.

Psychological Methods

The expressive qualities of music have been studied for years, the foremost of which has been the expression of emotion. Studies have shown that music is not only emotionally expressive but that there is high agreement among listeners about what type of emotion is being expressed. Psychologists study how music conveys or elicits emotions using one primary method. Clips of music are chosen based on certain structural features that are known to convey certain emotions. Participants listen to these clips and make judgments about the emotions they elicit or convey either during or directly after the clip. Many differing scales are used; however, a bipolar happy-sad scale is the most common scale given immediately after the clip to adults, and a choice between four emotions is the most common given to children. In studies on music that conveys or elicits mixed emotions, bipolar scales are separated so emotions like happiness and sadness are judged independently. When participants are asked to make judgments while listening to the music clips, they press one button for when the music is sad, and another when the music is happy, or both when responses are mixed.

Conveying Emotion Through Music

The ability to perceive conveyed emotion is said to develop early in childhood, and improve significantly throughout development. Empirical research has looked at which emotions can be conveyed as well as what structural factors in music help contribute to the perceived emotional expression. There are two schools of thought on how we interpret emotion in music. The cognitivists' approach argues that music simply displays an emotion, but does not allow for the personal experience of emotion in the listener. Emotivists argue that music elicits real emotional responses in the listener.

It has been argued that the emotion experienced from a piece of music is a multiplicative function of structural features, performance features, listener features and contextual features of the piece, shown as:

Experienced Emotion = Structural features x Performance features x Listener features x Contextual features

where:

Structural features = Segmental features x Suprasegmental features

Performance features = Performer skill x Performer state

Listener features = Musical expertise x Stable disposition x Current motivation

Contextual features = Location x Event

Structural Features

Structural features are divided into two parts, segmental features and suprasegmental features. Segmental features are the individual sounds or tones that make up the music; this includes acoustic structures such as duration, amplitude, and pitch. Suprasegmental features are the foundational structures of a piece, such as melody, tempo and rhythm.

Performance Features

Performance features refers to the manner in which a piece of music is executed by the performer(s). These are broken into two categories, performer skills and performer state. Performer skills are the compound ability and appearance of the performer; including physical appearance, reputation and technical skills. The performer state is the interpretation, motivation, and stage presence of the performer.

Listener Features

Listener features refers to the individual and social identity of the listener(s). This includes their personality, knowledge of music, and motivation to listen to the music.

Contextual Features

Contextual features are aspects of the performance such as the location and the particular occasion for the performance (i.e., funeral, wedding, dance).

These different factors influence expressed emotion at different magnitudes, and their effects are compounded by one another. Thus, experienced emotion is felt to a stronger degree if more factors are present. The order the factors are listed within the model denotes how much weight in the equation they carry. For this reason, the bulk of research has been done in structural features and listener features.

Specific Structural Features

There are a number of specific musical features that are highly associated with particular emotions. Within the factors affecting emotional expression in music, tempo is typically regarded as the most important, but a number of other factors, such as mode, loudness, and melody, also influence the emotional valence of the piece.

Tempo

Tempo is the speed or pace of a musical piece. Studies indicate an association between fast tempo and happiness or excitement. Slow tempo may be associated with sadness or serenity.

Mode

Mode, or the major or minor tonality in a piece often indicates happiness or sadness. Major tonality often conveys happiness or joy, while minor tonality is associated with sadness.

Loudness

Loudness, or the physical strength and amplitude of a sound, may be perceived as intensity, power, or anger; while soft music is associated with tenderness, sadness, or fear. Rapid changes in loudness may connote playfulness or pleading, whereas few or no changes can indicate peace and sadness.

Melody

In melody, a wide range of notes can imply joy, whimsicality, or uneasiness; a narrow range suggests tranquility, sadness, or triumph. Consonant, or complementing harmonies, are connected with feelings of happiness, relaxation, or serenity; dissonant, or clashing harmonies may imply excitement, anger, or unpleasantness.

Rhythm

Rhythm is the regularly recurring pattern or beat of a song. A smooth, consistent rhythm may be associated with happiness and peace. A rough, irregular rhythm may be associated with amusement and uneasiness, while varied rhythm implies joy.

Conflicting Cues

Which emotion is perceived is dependent on the context of the piece of music. Past research has argued that opposing emotions like happiness and sadness fall on a bipolar scale, where both cannot be felt at the same time. More recent research has suggested that happiness and sadness are experienced separately, which implies that they can be felt concurrently. One study investigated the latter possibility by having participants listen to computer-manipulated musical excerpts that have mixed cues between tempo and mode. Examples of mix-cue music include a piece with major key and slow tempo, and a minor-chord piece with a fast tempo. Participants then rated the extent to which the piece conveyed happiness or sadness. The results indicated that mixed-cue music conveys both happiness and sadness; however, it remained unclear whether participants perceived happiness and sadness simultaneously or vacillated between these two emotions. A follow up study was done to examine these possibilities. While listening to mixed or consistent cue music, participants pressed one button when the music conveyed happiness, and another button when it conveyed sadness. The results revealed that subjects pressed both buttons simultaneously during songs with conflicting cues. These findings indicate that listeners can perceive both happiness and sadness concurrently. This has significant implications for how the structural features influence emotion, because when a mix of structural cues are used, a number of emotions may be conveyed.

Specific Listener Features

Development

Studies indicate that the ability to understand emotional messages in music starts early, and improves throughout child development. Researchers play a musical excerpt for children and have them look at pictorial expressions of faces. The children are asked to select the face that best matches the music's emotional tone. Studies have shown that children are able to assign specific emotions to pieces of music; however, there is debate regarding what age this ability begins.

In one study, preschool and elementary-age children heard twelve short melodies, each in either major or minor mode, and chose between four pictures of faces: happy, contented, sad, and angry. All the children, even as young as three years old, performed above chance in assigning positive faces with major mode and negative faces with minor mode.

In another group of experiments, children labelled musical excerpts with the affective labels "happy", "sad", "angry", and "afraid". The results of the first study showed that four-year-olds did not perform above chance with the labels "sad" and "angry", nor the five-year-olds with the label "afraid". The second study found conflicting results, where five-year-olds performed much like adults. However, all ages sometimes confused "angry" and "afraid".

Eliciting Emotion Through Music

Along with the research that music conveys an emotion to its listener(s), it has also been shown that music can produce emotion in the listener(s). This view often causes debate because the emotion is produced within the listener; and thus, hard to measure. In spite of this controversy, studies have shown observable responses to elicited emotions, which reinforces the Emotivists' view that music does elicit real emotional responses.

Responses to Elicited Emotion

The structural features of music not only help convey an emotional message to the listener, but also may create emotion in the listener. These emotions can be completely new feelings or may be an extension of previous emotional events. Empirical research has shown how listeners can absorb the piece's expression as their own emotion, as well as invoke a unique response based on their personal experiences.

Basic Emotions

In research on eliciting emotion, participants report personally feeling a certain emotion in response to hearing a musical piece. Researchers have investigated whether the same structures that conveyed a particular emotion could elicit it as well. The researchers presented excerpts of fast tempo, major mode music and slow tempo, minor tone music to participants; these musical structures were chosen because they are known to convey happiness and sadness respectively. Participants rated their own emotions with elevated levels of happiness after listening to music with structures that convey happiness, and elevated sadness after music with structures that convey sadness. This evidence suggests that the same structures that convey emotions in music can also elicit those same emotions in the listener.

In light of this finding, there has been particular controversy about music eliciting negative emotions. Cognitivists argue that choosing to listen to music that elicits negative emotions like sadness would be paradoxical, as listeners would not willingly strive to induce sadness. However, emotivists purport that music does elicit negative emotions, and listeners knowingly choose to listen in order to feel sadness in an impersonal way, similar to a viewer's desire to watch a tragic film.

Researchers have also found an effect between one's familiarity with a piece of music and the emotions it elicits. In one study, half of participants were played twelve random musical excerpts one time, and rated their emotions after each piece. The other half of the participants listened to twelve random excepts five times, and started their ratings on the third repetition. Findings showed that participants who listened to the excerpts five times rated their emotions with higher intensity than the participants who listened to them only once. This suggests that familiarity with a piece of music increases the emotions experienced by the listener.

Emotional Memories and Actions

Music may not only elicit new emotions, but connect listeners with other emotional sources. Music serves as a powerful cue to recall emotional memories back into awareness. Because music is such a pervasive part of social life, present in weddings, funerals and religious ceremonies, it brings back emotional memories that are often already associated with it. Music is also processed by the lower, sensory levels of the brain, making it impervious to later memory distortions. Therefore creating a strong connection between emotion and music within memory makes it easier to recall one when prompted by the other. Music can also tap into empathy, inducing emotions that are assumed to be felt by the performer or composer. Listeners can become sad because they recognize that those emotions must have been felt by the composer, much as the viewer of a play can empathize for the actors.

Listeners may also respond to emotional music through action. Throughout history music was composed to inspire people into specific action - to march, dance, sing or fight. Consequently, heightening the emotions in all these events. In fact, many people report being unable to sit still when certain rhythms are played, in some cases even engaging in subliminal actions when physical manifestations should be suppressed. Examples of this can be seen in young children's spontaneous outbursts into motion upon hearing music, or exuberant expressions shown at concerts.

Comparison of Conveyed and Elicited Emotions

Evidence for Emotion in Music

There has been a bulk of evidence that listeners can identify specific emotions with certain types of music, but there has been less concrete evidence that music may elicit emotions. This is due to the fact that elicited emotion is subjective; and thus, it is difficult to find a valid criterion to study it. Elicited and conveyed emotion in music is usually understood from three types of evidence: self-report, physiological responses, and expressive behavior. Researchers use one or a combination of these methods to investigate emotional reactions to music.

Self-Report

The self-report method is a verbal report by the listener regarding what they are experiencing. This is the most widely used method for studying emotion and has shown that people identify emotions and personally experience emotions while listening to music. Research in the area has shown that listeners' emotional responses are highly consistent. In fact, a meta-analysis of 41 studies on music performance found that happiness, sadness, tenderness, threat, and anger were identified above chance by listeners. Another study compared untrained listeners to musically trained listeners. Both groups were required to categorize musical excerpts that conveyed similar emotions. The findings showed that the categorizations were not different between the trained and untrained; thus demonstrating that the untrained listeners are highly accurate in perceiving emotion. It is more difficult to find evidence for elicited emotion, as it depends solely on the subjective response of the listener. This leaves reporting vulnerable to self-report biases such as participants responding according to social prescriptions or responding as they think the experimenter wants them to. As a result, the validity of the self-report method is often questioned, and consequently researchers are reluctant to draw definitive conclusions solely from these reports.

Physiological Responses

Emotions are known to create physiological, or bodily, changes in a person, which can be tested experimentally. Some evidence shows one of these changes is within the nervous system. Arousing music is related to increased heart rate and muscle tension; calming music is connected to decreased heart rate and muscle tension, and increased skin temperature. Other research identifies outward physical responses such as shivers or goose bumps to be caused by changes in harmony and tears or lump-in-the-throat provoked by changes in melody. Researchers test these responses through the use of instruments for physiological measurement, such as recording pulse rate.

Expressive Behavior

People are also known to show outward manifestations of their emotional states while listening to music. Studies using facial electromyography (EMG) have found that people react with subliminal facial expressions when listening to expressive music. In addition, music provides a stimulus for expressive behavior in many social contexts, such as concerts, dances, and ceremonies. Although these expressive behaviors can be measured experimentally, there have been very few controlled studies observing this behavior.

Strength of Effects

Within the comparison between elicited and conveyed emotions, researchers have examined the relationship between these two types of responses to music. In general, research agrees that feeling and perception ratings are highly correlated, but not identical. More specifically, studies are inconclusive as to whether one response has a stronger effect than the other, and in what ways these two responses relate.

Conveyed More Than Elicited

In one study, participants heard a random selection of 24 excerpts, displaying six types of emotions, five times in a row. Half the participants described the emotions the music conveyed, and the other half responded with how the music made them feel. The results found that emotions conveyed by music were more intense than the emotions elicited by the same piece of music. Another study investigated under what specific conditions strong emotions were conveyed. Findings showed that ratings for conveyed emotions were higher in happy responses to music with consistent cues for happiness (i.e., fast tempo and major mode), for sad responses to music with consistent cues for sadness (i.e., slow tempo and minor mode,) and for sad responses in general. These studies suggest that people can recognize the emotion displayed in music more readily than feeling it personally.

Sometimes Conveyed, Sometimes Elicited

Another study that had 32 participants listen to twelve musical pieces and found that the strength of perceived and elicited emotions were dependent on the structures of the piece of music. Perceived emotions were stronger than felt emotions when listeners rated for arousal and positive and negative activation. On the other hand, elicited emotions were stronger than perceived emotions when rating for pleasantness.

Elicited More Than Conveyed

In another study analysis revealed that emotional responses were stronger than the listeners' perceptions of emotions. This study used a between-subjects design, where 20 listeners judged to what extent they perceived four emotions: happy, sad, peaceful, and scared. A separate 19 listeners rated to what extent they experienced each of these emotions. The findings showed that all music stimuli elicited specific emotions for the group of participants rating elicited emotion, while music stimuli only occasionally conveyed emotion to the participants in the group identifying which emotions the music conveyed.

Based on these inconsistent findings, there is much research left to be done in order to determine how conveyed and elicited emotions are similar and different.

The nature of musical emotions (some aspects)

The last problem to be considered here is the nature of the emotions elicited by music. Take the kind of sadness one may feel hearing the Rachmaninoff’s second piano concerto. The listener in this case feels a different kind of sadness than for loss of a loved person. He neither feels regret for the music as for an unfortunate event, nor compassion for a sentient being experiencing tragedy. Juslin and Västfjäll (2008)state that the listener’s emotional reaction lacks the cognitive appraisal of a “real” emotion, i.e. the subjective evaluation of an event, in this case the music, in relation to goals and needs of the individual. This led some theorists to the conclusion that music does not elicit emotions at all or that music can only elicit moods, i.e. affective states with lower intensity than emotions and without a clear object (cf. Juslin and Västfjäll 2008). Stephen Davies rejects this view. Although the emotional response does not take the music as its intentional object, music is the “perceptional object and the cause for this response”. The listener’s response of sadness is “not about the music, but to the music” (Davies 2006). The emotion enfolds over the course of the music and as a consequence of it. Juslin and Västfjäll (2008)compiled evidence that most of the psychological processes that lead to a real-life emotion can also be found in the emotional responses to music. These include subjective feeling, physiological arousal, brain activation, action tendency, and emotion regulation. Thus, they argue, music-induced emotions are of the same quality as “normal” emotions are, and do not just represent moods.

Still, emotions elicited by music have some characteristics that make them different from real-life emotions. Coming back to the example of the piano concerto, the sadness felt at hearing it does not only lack the regret of the sadness at the death of a loved person, but it is also certainly less intense. There seem to be emotional intensities of real-life events that the experience of an artwork cannot reach. In addition, most emotions, particularly the negative ones, felt when listening to music seem to have a positive tinge. Why do we seek the experience of a negative emotion as in a sad piece of music? One reason is that we appreciate, in an artistic, aesthetic way, the music as an artwork that manages to create the expressiveness. Another reason is given by Kendall Walton (see http://plato.stanford.edu/entries/music/): Sadness is not negative in itself. Rather the life situation that causes it, e.g. the death of a loved person, is negative. “Thus, though we would not seek out the death of a loved one, given the death we ‘welcome’ the sorrow.” Music gives the listener the possibility of self-experience through real emotions, without the consequences of real-life circumstances, just as any art and play does.

References

This article incorporates material from the Citizendium article "Music and emotion", which is licensed under the Creative Commons Attribution-ShareAlike 3.0 Unported License but not under the GFDL.