Enjoy reading? Consider buying the framework as an e-book on Gumroad:
https://sonatasecrets.gumroad.com/l/howmusicworks

8. Emotional Charge

We have seen how elements along the musical dimensions convey their part of an expression. When they come together, as is the normal case in music, they often form an expression we perceive as coherent, and moreover we often attribute an emotional charge to that expression. The big question here is why we do that – the music itself cannot feel any emotions, because it’s just vibrations. Philosophers have grappled with this question for a long time, and adjacent to it, how we also feel emotions when listening to music. Let’s start by separating the two issues – emotions perceived in the music on one hand and emotions felt or aroused in the listener by the music on the other. It’s perfectly natural to be able to identify or perceive emotion in another entity without feeling it oneself. Here already, many earlier theories have failed to grasp the difference and produced confusing accounts of how emotion in music works under the banner of Arousal theories of expression (see Matrevers 2011; Davies 1994 chapter 4; Robinson 2005). We will look at the arousal part, or emotional responses, in chapter 9. Even Persona theories of expression (Levinson 2006; see Robinson 2011), which presupposes an imaginary persona expressing emotions in music, is in my opinion only an unnecessary addition.

A Linking Mechanism?

When it comes to the expression and perception of an emotional charge of the music, the question is how these features, and their valence are linked to the emotional states. The philosopher Stephen Davies presents a Resemblance theory of “Emotional characteristics in appearance” that provides a robust groundwork for an explanation (Davies 1994). He submits that one way of our talking about emotions is precisely only the emotional characteristics in the appearance of other entities, and not automatically combined with any “intentional objects” that is otherwise essential for emotions. He gives the example of the face of a basset hound dog who looks said, even though it isn’t, to illustrate the difference between the two ways of talking about emotion. A person can look sad, or convey sadness in their appearance, without us knowing if he really is sad or if he’s just acting, and it still makes sense for us to say that he is sad, and act accordingly until we have further information. When we interact with each other, facial expression is a primary source for conveying this information, but two other important sources are tone or infliction in speech (vocal expression) and our way of moving (walking, gesturing). Facial expression cannot be transferred to music since it only occurs in a visual medium. Davies draws the musical emotional characteristics instead primarily on movement and gestures similarities as what he calls “configurations of human behavior associated with emotions” (Davies 2006: 181), in how we have specific bodily movement characteristics that are shared by musical movement. For example, happy movement features higher tempo and rising gestures, and sad movement slower tempo and falling gestures. The theorists Peter Kivy and Patrik Juslin on the other hand promote vocal expression to be the primary connecting link in their respective works that take slightly different routes (Kivy 1989; Juslin 2019 chapter 11.2). This makes sense intuitively since both music and speech are auditory sources of expression. The similarities are the same as with the movement connection, but seemingly even more pronounced.

Basic Emotions

This is however valid only for those “basic” emotions that are possible to convey using applicable characteristics. More complex emotions usually require intentional objects, but the expression of those emotions will still mostly look like one of the basic emotions. For example someone feeling love will seem happy to some degree (joyful or content), and it’s only possible to determine the emotion as love once you know the reason for the expression characteristic. With the characteristics applicable to music, we have basically five basic emotions in play: Happiness, Sadness, Tenderness, Anger and Fear. These are easily differentiated by only different characteristics in vocal expression in speech, and they have evolved in humans for us to be able to communicate effectively in a social context.

“Selection pressures throughout evolution have favored efficient and robust communication of basic emotion categories, because they concern our most fundamental life issues.” (Juslin 2019: 160)

There is one more important basic emotion that serves a clear evolutionary purpose: Disgust. But this is not commonly represented in music, and I can think of two reasons why not. (1) It’s signifying something we want to avoid so it’s naturally not a popular emotional topic for music; and/or (2) it lacks a specific characteristic in vocal expression or movement that the other basic emotions have, and so cannot be represented in music by this feature.

Now, the five basic emotions can be placed within a plane along the two axes of valence (positive/negative) and energy or arousal. The only problem is that Anger and Fear appear in the same quadrant, even though they still have different characteristics. I think it’s fitting here that we talk about the fight-or-flight response when encountering danger, which will trigger either of the two high-energy negative emotions. Anger summons energy towards something, and fear summons it to get away. If we insert the emotions on this plane, we get what I would like to call the Matrix of Emotional Characteristics, the MEC.

Drawing from other domains we find some interesting parallels for the four quadrants. First the reductionist use of Emojis to convey emotion through facial expression. Typical emojis for the basic emotions follow two features along the same axes: Eyes are relaxed for low energy and open for high energy, and the mouth is smiling for positive valence and forming an inverted smile for negative. This is not applicable to music at all, but I find it fitting that the most fundamental aspects of facial expression overlaps. Secondly, we have the historical, pre-scientific peculiarity of classifying human personalities in four categories that was thought to be dependent on the prominence of different Bodily fluids: Choleric, Melancholic, Sanguine and Phlegmatic. They fit pretty neatly onto the matrix’s two axes as well! (although not perfectly)

Returning to music we still need to answer the question why specifically emotional characteristics are perceived as primary in music, and not shared characteristics from other domains, for example nature or mathematics. Here Davies brings up one important feature: the intentionality, or agency, we perceive in music. Music seems like it has a will, it does things that are expressed in its gestures and has a direction in its movement. This is a feature that is very human – it’s essential for how we interact with other people and an important basis for any moral system. Infants and animals also show agency, but it is limited to a lesser extent and not as fully developed as in adult humans, so we treat them according to such limitations. The argument with music is that because we sense this agency or intentionality (it is of course part of the illusion; the sound vibrations themselves have no agency), we go so naturally to the emotions that are an indispensable part of human agency, since they are its very regulators. Now we finally have a model of emotional valence, or charge, in place. Juslin calls this “Iconic coding:”

“The ‘core’ layer of expression … consists mainly of iconically coded basic emotions based on vocal expression and human gesture. … This layer explains universal recognition of basic emotions in musical expression. … Whenever we observe strong listener agreement in empirical studies, it is mostly thanks to iconic coding which due to its characteristics tends to render emotional expressions broad (some might say imprecise), yet robust and reliable.” (Juslin 2019: 182-183)