Since my senior colleagues and I are trying to do something about emotions in music from computer science angle, it’s better to see how people in other disciplines(mainly humanity) handle this problem before starting our own project. So this is a very short introduction to the music and emotion study from humanity perspective. And this is not an academic survey, so I will share with you using as simple language as I can.
When listening to the music, usually people will have some kinds of feelings. And the behavior of treating music as a medium of conveying emotion dates back to ancient Greek times; there is no doubt that music is a way of expressing and generating emotion at least in non-academic world (our normal world).
However, in academic world, theorists are arguing about how music arouse our feelings. Does music itself contain emotions? Or it’s our recognition process that is responsible for the emotion so music is purely a complex stimuli.
How Music and Emotion relates to each other?
Nowadays there are mainly two types of view about how music and emotion relates. The first is that music do elicit some specific feelings by using several secret ingredients; the second is that music itself have no direct relationship with human emotions and music simply act as an incentive; it is a complicated mechanism including physiological and cognitive factors that help us translate incentive into specific emotions.
First View – Music itself convey emotion
Cooke1 thinks that melodic intervals and patterns have their own respective emotions. That is to say, music itself contains emotion, different intervals have different emotions. For example, an ascending major third is joyful while major second is less suggestive compared with major third. In his theory, he suggests that melodic interval and patterns are responsible for emotions in music all around the cultures. Yet non-tonal music like India music has no explicit form of major third and so on, which derives from the western music system. Though his theory is restricted in western music system, there are many useful data he’s collected for future emotion evaluation. In later part I will list some data he collected.
Kivy2 holds the same opinion as Cooke, and they both think music itself represents some feelings like anger, sadness and delight. Compared with Cooke’s work, Kivy takes more factors into consideration, not just intervals and patterns. He thinks that tempo and mode are also important if composers want to convey specific feelings in their music.
Unlike Cooke and Kivy, people who hold the second view regard music as a complex stimuli and don’t think music itself contains specific and precise feelings. The representative is Meyer, Huron and Juslin.
Meyer3 argued that music denote no specific emotions. The arousal from music is undifferentiated, and we can’t get any specific feelings naturally. The emotion from music is closely related to our expectation, that is to say, this is an acquired skill to appreciate music. When the music sequence do not meet our expectation, the arousal happens. Unlike Cooke’s theory, we have specific feelings when arousal happens due to many learned factors in reality, such as individual custom and tradition. Hence, according to Meyer’s theory, if we want to analyze the music revealed by the music, we need to take personal information into account, which is nearly impossible in most conditions. 😦
In general, people who believe in the second view think that music can only convey some undifferentiated emotional arousal and we have to consider more factors when processing these arousal into specific feelings. Hence, they accept that expectancy is a crucial factor of music emotion, as well as that some musically evoked emotions are not simply from derived expectancy system. They believe that there is a complex system which is responsible to emotions in music.
Huron4 identifies five categories of expectancy music response: Imagination, Tension, Prediction, Reaction and Appraisal. Imagination and tension are pre-outcome responses, while prediction, reaction and appraisal are post-outcome responses. Pre-outcome responses happen prior to an event and the post-outcome responses happen after an event. For example, along the music performing, people may have their own imagination prior to the next note or chord and may feel the tension, while after one note or chord, prediction may be wrong, reaction system may be activated and appraisal may be given. This system sometimes explains a very complex situation. Five systems response respectively will result a very complex emotion. For example, people feel dangerous when tension is tight, prediction always fails and imagination is highly activated. In essence, the central to this theory is the natural physiological responses of human being when stimulated by music, which composes of complicated elements that simulates almost everything.
As to Juslin and Västfjäll’s5 multiple mechanism theory, they also raise a complex system composed of 6 sub-system to explain the emotions when people listen to music. The sub-systems are brain stem flex, evaluative conditioning, emotional contagion, visual image, episodic memory and expectancy. The brain stem flex is the basic physiological system that responses to all the acoustic characteristic of sound; the evaluative conditional means that some emotion will be expressed only in some specific conditions or we can say condition is a factor determines the emotion; emotional contagion refers to that perceiving an emotion will sometimes induce the same emotion to oneself or simply saying empathy; visual Image means that listening to music leads to visual imagination; episodic memory implies that your before experience will evoke or boost some emotions when you are listening to specific music; expectancy is very similar to Meyer’s idea, that is, an unexpected event in music leads to emotion arousal. This theory integrates many factors to explain the complex relation of music and emotion. In my opinion, this theory has strong bias on personal previous experience and outside environment when listening to music.
Music and emotion have close relationship, and many researchers have done a lot of work in exploring the relationship. There are two major opinions. One is that there are some secret ingredients in the music that directly leads to or express some feelings, and the other view suggests that it is many other factors that response to the specific emotions aroused by music. This conclusion is far from satisfying, but at least we gain an unified sense that emotion aroused music is not simply from music itself and we have to consider a lot of other factors case by case. As a result, now it is almost impossible to design an algorithm to precisely analyze the emotions hidden in the complex music. Yet we still have a lot of work to do. Many evidences imply that music do have some concrete patterns we can follow to inject our feelings. Though these composed feelings may not apply to everyone, we can still analyze the strength level and whether this is positive.
Next step I will try to explore as many ingredients as I can for later computer analysis use. Also, I will post my result here to share with you! Hope one day we decode the secret of musical emotions!
1 Cooke, D. (1959) The language of music. London: Oxford University Press.
2 Kivy, P. (1980) The corded shell: Reflections on musical expression. Princeton, NJ: Princeton University Press.
3 Meyer, L. B. Emotion and meaning in music. Chicago: University of Chicago Press.
4 Huron, D. Sweet anticipation: Music and the psychology of expectation. Cambridge, MA: MIT Press.
5 Juslin, P. N. and Västfjäll, D. Handbook of music and emotion: Theory, research, application. New York: Oxford University Press.