Can one get a bit deeper into how this bit of music produces its effects? Others think so, so I’ll look at a book I admire, even though it has limitations.
David Huron’s Sweet Anticipation: Music and the Psychology of Expectation (2006) is pretty ambitious. It’s more than fifteen years old now, so he wasn’t able to take in recent neuroscience However, as the subtitle says, his focus is psychology – in this case observable effects on perception and response – rather than underlying brain mechanisms so that doesn’t matter too much.
He’s intent on accounting for pleasure in music. His take is that there are lots of musical pleasures, and they pose intriguing problems. Why, for example, do we find miserable music pleasing, sometimes, when it gives us feelings that we normally want to avoid?
He offers ideas that might help us understand why some of the pleasure arises. They are basically an elaboration of the expectation met or denied, tension-and-release model that is commonplace in talk about music’s effects.
Huron deals with the possible contradictions that dog this position – it usually comes across as claiming that we get a buzz from having our predictions about what comes next in a piece of music confirmed, but also enjoy it when they aren’t. Huron’s solution: a multi-part theory of musical expectation.
The theory has a distinctly unsnappy acronym, ITPRA. The letters label the five stages of forming an expectation and reacting to the ensuing events. These, he reckons, are completely general, even though he’s going to focus on music. The theory is crafted, he says, “to explain how expectations evoke various feeling states, and why these evoked feelings might be biologically useful.” It is inspired, though, by earlier work on music, especially Leonard Meyer’s 1950s book Emotion and Meaning in Music, which claimed that managing expectations was a composer’s main way of tapping into emotion.
Here’s what each letter stands for.
I – Imagination: we have evolved to imagine future events, put a probability on them, and on whether they will feel good or bad.
T – Tension: the stress we get from anticipating the bad thing, or the mood lift from waiting for a good one, in the moments just before the imagined event comes to pass.
P – Prediction response. This comes just after the event. What happened? Was it what we thought was coming? If so, that feels good, even when the actual event is bad. This is a quick one, though the timescales the outline of the theory talks about still seem much slower to me than what’s happening in any piece of music.
R – Reaction response. Getting the prediction right may be positive, but immediate responses relate to the actual event, and can be positive or negative. Not great to find you have successfully predicted that the fierce dog is going to sink its teeth into your leg.
A – Appraisal. A slower response, after thought or conscious awareness kicks in. How we feel about the event may change depending on lots of enviromental, cultural or social influences.
This version of the theory proposes that each of the five bits is biologically functional, in the sense of having some survival value. That’s the conventional evolutionary assumption, if still a large one. He also suggests each one is the work of a “functionally distinct neurophysiological system”. That feels like a pretty big addition to me. Are either of these things true? Let’s suspend judgment on that: I don’t really know. But anyhow, how might all this relate to music, and why it’s feels good?
To start with, the theory has enough parts to provide escape routes from the dead-ends that loom when the source of pleasure is taken to be successful prediction. Example: Getting a musical prediction right, as for any other stab at forecasting the future, is easiest when things just repeat. And Huron cites a bunch of psychology experiments that indicate people do like to hold on to what is familiar.
But, as he says, that would make “the most enjoyable music… entirely predictable – utterly banal?” Don’t worry, he assures us that “proposing that predictability evokes pleasure does not preclude the existence of other psychological phenomena that tend to limit the effect of repetition”. There might be separate rewards for novelty-seeking, or boredom avoidance. And other parts of ITPRA, namely the reaction and appraisal responses, can also counter positive effects of repetition and predictability.
Well, I suppose so. But I’m getting uneasy now. The theory, which seems reasonable as an elaborate, somewhat orderly, redescription of what is going on for a listener, seems to have a set of separate components that can be combined in suitably ad hoc ways to explain pretty well any result. In which case, to my mind, it explains nothing much.
I can, I think, apply this framework to sitting and listening again to Ramblin’. I can imagine what it will be like pretty well. I know the piece by heart. And when I imagine, I imagine liking it. That creates a nice anticipation of the pleasures to come. And that prediction proves true, or always has so far. Reaction and appraisal then follow, and are similarly positive.
That doesn’t seem to add much to just owning up to the fact that I’ve grown to love the piece and when I hear it again I like it a lot, again. Part of the problem, again, seems to be timescales. This broad brush approach is hard to relate to the experience of listening to Coleman, Cherry, Haden and Higgins create each moment of the music. So can we adjust the view to bring those details into focus through the lens of Huron’s theory?
Maybe a bit, though the view remains blurry. The approach is a little indirect, via general preferences that have been uncovered in experiments. The constraint is that, on the whole, the better designed the experiment, the simpler the aspect of music that is isolated for study. Do lots of experiments like that, and you end up with a clutch of bits and pieces, tendencies to like, or at least pay attention to, certain features of sounds. They are things like what pitch will follow one just heard, whether it will move up or down a scale, and how far. The “preference” is revealed by various stand-ins that can be measured, such as reaction time (for adults) or how long a gaze is directed toward a sound source (for infants).
This kind of work, mainly with subjects who knew Western music, seems to show that listeners’ fastest, unconscious preferences are attuned to the statistical distribution of tones in the music they hear, which is presumably how such predictions are made. That leaves open whether they acquired the preference by learning, or because the music conforms to an innate liking for certain intervals or particular rhythmic devices. Opinions differ there, but it seems safe to assume that it’s a bit of each.
Huron’s own group, for example, gathered evidence for four different kinds of expectation relating to tones. People, Western people in the lab, respond measurably differently to scale tones, and to whether they are heard in a major or minor key. And they have expectations about the frequency of particular tones at the end of a musical phrase, also depending on whether the context is major or minor.
A different strand of work gives results a bit like this that relate to rhythm. The downbeat feels good for the same reason some tones do: it’s where the brain has predicted the next sound will happen.
And the whole thing is rounded off by a gesture toward the fact that bad things, soundwise, sometimes go with good feelings. Ah, says Huron, That’s because of contrastive valence. Or to put a bit of low value-added jargon back into plain English, when a good thing follows a bad one, it feels even better than usual. The first drink of harmony after you are rescued from a musical desert is the sweeter for it.
That’s enough to give some idea about what one mostly positive reviewer called “a somewhat idiosyncratic perspective on music cognition”. It seems an impressive effort to take the analysis of expectation in music a bit further, although the specific examples of anticipation are still pretty simple compared with, you know, an actual piece of music such as this one.
It’s less helpful, as far as I can see, in explaining quite how this anticipation ties in with pleasure, although it surely is. I’m left a bit confused whether I am enjoying the instant, unconscious musical predictions my brain offers when I listen again to Ramblin’, and whether the surprising bits are still surprising, and indeed whether I ought to be (enjoyably) surprised that I can still be surprised – in some sense – by a piece that is now so deliciously familiar. It’s hard to introspect about all this, as about many aspects of music, as the immediate cognitive processing of sound happens so fast, and remains inaccessible to my listening self.
And does this account for why there’s a particular frisson attaching, say, to the elongated blurry alto sax shriek just before the last phrase of Ornette’s solo, before he hands over to Don Cherry? Or a reliable tingle when Billy Higgins begins a pattern on the cymbal stem behind Charlie Haden’s bass solo? Could we stage an experiment to check? I don’t know…
I’ll settle, though, for the underlying proposition – that part of the pleasure derived from music is bound up with a continuous series of small predictions, whether momentarily proved correct or incorrect. If so, that raises another question. It seems our brains have developed to respond that way, so that predictions, specifically musical predictions, can access pleasure. The pleasure is there: we can trace the release of the neurotransmitter dopamine in areas associated with feeling good when we hear the right music. But why? The pleasure reinforces the desire to listen to music some more. From an evolutionary viewpoint, there ought to be some advantage we can figure out that explains why that happens. There are quite a few possibilities – but first, time for another listen…
This is No 9 in a series of posts that all centre on a single piece of music – it starts here, and may go on for some time.