Why are we moved by music? For any attempt to understand how our minds grasp and process music, this is perhaps the hardest question of all. There is something uniquely intangible about the way music works its alchemy. Many great paintings literally depict feelings, whether in faces, gestures or circumstances.

Even abstract art can evoke associations with its shapes and colours: Yves Klein’s celestial, infinite azure; Mark Rothko’s sombre horizons; Jackson Pollock’s spattered traces of frenzied intensity. Literature stimulates emotion through narrative, characterisation and allusion, and even if the ‘meaning’ can be somewhat elastic, the limits are relatively narrow: no one suggests, say, that Great Expectations is a story about coal mining.

But music is invisible and ephemeral: it sighs and roars for a moment, and then it is gone. Save for the occasional moments of intentional mimicry, it refers to nothing else in the world. Particular phrases and tropes had conventionalized ‘meanings’ in the Classical era, and some have argued that Western instrumental music more generally has specific meanings that can be objectively decoded.

But those arguments always run aground in arbitrary assertion. The answers that they offer are too easy, too superficial. For it is hard enough to understand why we should be able to make the slightest sense of these eliding acoustic signals, let alone why they should move us to tears and laughter, make us dance or rage.

It is only rather recently that the question of emotion has been raised at all by cognitive scientists and musicologists. Eduard Hanslick, whose 1854 book The Beautiful in Music was one of the first serious modern studies of musical aesthetics, complained that previously music had tended to be discussed either in the “extremely dry and prosaic language” of technical theory or by aesthetes “wrapped in a cloud of high-flown sentimentality”.

His focus was not on emotion per se but on the aesthetic effect, particularly the sense of beauty, that arises where feeling and intellect meet. In the eighteenth and nineteenth centuries, said Hanslick, it was all but taken for granted that music is a crafting of sound with the primary purpose of expressing and exciting passions. And yet he doubted that there was any real correspondence between a musical composition and the feelings it elicits. After all, Beethoven was considered in his time to be passionate in comparison to the cool clarity of Mozart, yet Mozart was likewise once deemed fervent in comparison to Haydn. “Definite feelings and emotions,” Hanslick claimed, “are unsusceptible of being embodied in music.”

But no one can doubt that some music is capable of exciting some emotion in some people some of the time. Indeed, many might still agree with the old view that music’s raison d’être is to inspire emotion. The question is how.

Hanslick insisted that this would always be mysterious. “The physiological process by which the sensation of sound is converted into a feeling, a state of mind, is unexplained, and will ever remain so,” he wrote. “Let us never appeal to a science for explanations which it cannot possibly give.” That was unduly pessimistic (or, one suspects, hopeful). And yet, someone coming to the psychology of music hoping to learn how it inflames the heart may well be disappointed. At present, such studies often seem woefully inadequate, even simple-minded, on the topic of emotion. When neuroscientists examine how and when people classify particular musical extracts as ‘happy’ or ‘sad’, the music lover might reasonably see that as traducing the emotive qualities of music – as though we sit through a piano concerto either beaming or moping.

Some people might react with horror to the thought that the emotive effects of music can be explained and classified with the cold instruments of science. Doesn’t this threaten to reduce the composer to a technician manipulating the responses of a passive audience? It might even conjure up the spectre of music being used for social manipulation and behavioural control: muzak making us bovine, or Wagner’s ‘Ride of the Valkyries’ turning us homicidal.

But one needn’t look very deeply for these worries to recede. For one thing, although music probably can, in the right contexts, encourage violence or calm, so can language, drama, perhaps even colour or the weather. We do not need cognitive psychology to tell us what kinds of music tend to relax or arouse people, nor is it at all likely to show us how to make music that achieves these things significantly more effectively than is already possible.

And music is not a pill that, when swallowed, inevitably produces a prescribed state of mind. Music psychologist John Sloboda warns that psychological studies of musical emotion run the risk of seeming to reduce much-loved pieces “to a set of ‘effects’ such as might similarly be realized by the colour of paint on a wall, or the administration of caffeine.” He calls this the ‘pharmaceutical model’ of music.

Hanslick was doubtless right to assert that no amount of scientific and musicological study will tell us all there is to know about the link between music and emotion. Yet neither must it remain inscrutable. We have good reason to think that there are some general underlying principles of musical affect, and that it is by no means some magical property beyond the reach of rational analysis. I doubt that any composer or musician has ever truly believed that.

Even if we don’t necessarily feel the emotion that we imagine a piece of music seeks to express, most people will agree that music does sometimes move them. This need not be (as Hanslick feared) because they are determined it should do so, in some egotistical display of ‘sensitivity’. It’s frequently quite the opposite – who hasn’t struggled to hold back embarrassed tears when the manipulative soundtrack swells in some corny movie? It’s almost disturbing how susceptible we are to this emotional compulsion. “Music operates on our emotional faculty with greater intensiveness and rapidity than the product of any other art,” wrote Hanslick. “The other arts persuade us, but music takes us by surprise.” How right neuroscientist Isabelle Peretz is when she says, “My intuition is that musical emotions typically occur without consciousness or willingness.”This is the central mystery we need to explain.

How might music create these feelings? Think of Ann Ronell’s great 1932 song ‘Willow Weep for Me’, sung by everyone from Billie Holiday to Nina Simone. Willows ‘weep’ because they look like a weeping person, with a slouched, drooping physiognomy. It seems likely that some broad characteristics of musical emotion stem from a similar source: they mimic in music the physicality of that emotion in people, particularly in speech and gesture.

When most listeners agree that music is sad, it is likely to be slow and soft – rather like the way sad people tend to move and speak. This doesn’t mean that all slow, soft music is sad, but that few people would find Japanese Kodo drumming or Handel’s Music for theRoyal Fireworks sad. Similarly, music commonly designated ‘happy’ tends to be relatively fast-paced and jaunty, and at a moderately loud volume. Such purely acoustic properties allow simple emotional messages to be easily discerned on the ‘surface’ of the music.

But these mimetic qualities often account only for relatively superficial aspects of our emotional response to music. They set the mood, but offer little nuance. Moreover, we seem reliably able to recognise in music only a very small number of basic emotions this way – what, for example, would we expect in the ‘physiognomy’ of music that sounded ‘hopeful’ or ‘resentful’?

In the 1950s, the American music theorist and philosopher Leonard Meyer presented a plausible account of how aesthetic affect in music can arise from the expectations the music creates. What Meyer realised is that the emotional effect of music doesn’t come from having our expectations met, but from having them more or less violated.

We don’t simply want to feel happy or satisfied by music because it has gone the way we expected; indeed, that would be more likely to leave us only mildly affected by it, if not in fact positively bored. No, we like to listen to music because it sounds exciting, vigorous, poignant, beautiful, ennobled, sexy, and too many other things to list. And it achieves these things, Meyer argued, not in spite of, but because of, the mismatch between our expectations and the reality.

This is how Meyer imagined a musical cadence working. Cadences – sequences of chords that end a music phrase – may bring about a feeling of partial or total closure by carrying the harmony on to a relatively ‘stable chord’, especially the tonic. Western listeners have learnt to expect, even to long for, the final cadence of a piece of music to move to the tonic.

In Meyer’s view a perfect cadence tossed off at the end of a phrase is unexceptional, even mildly dissatisfying in its predictability. But when it is delayed, we experience tension, because we sense what is coming but don’t know when it will come. When it does, our satisfaction is greater. Even the slightest deviation from the most predictable outcome – a slowing down or rallentando of the kind that musicians commonly apply, and composers often specify, at the final cadence – is enough to create an emotional effect by toying with our expectations, creating hints of uncertainty about the outcome.

According to Meyer, music is full of moments like this in which we are given clues about what will come next but can’t be sure that we’ve guessed right. Composers manipulate those expectations with more or less calculated gestures, pulling on the strings of our emotions. There are many ways of doing that, with pitch, melody, harmony, rhythm and timbre.

But there are difficulties in imagining that musical emotion is solely a product of expectation and its violation. It is easy to enumerate the various ways in which these ingredients might create emotion; the problem is that, once you apply these to real music, you find these features everywhere. Almost nothing in music can happen without violating some plausible expectation, because the possibilities are too great. Perhaps, as some neuroscientists have suggested, we maintain a battery of different types and representations of expectations, and are constantly shifting between them depending on which seems at any moment to be the most successful – a notion dubbed ‘neural Darwinism’.

Yet I believe we need a wider general framework for constructing a picture of musical emotion – one that reflects more precisely and explicitly the true nature of what we feel when we take pleasure in music. Where most current cognitive analysis seems to fall short most egregiously is in presenting musical emotion as some kind of oscillation between positive and negative emotional responses. Obviously, the positive must largely win out, because we enjoy music so much (that’s to say, we enjoy the music that we enjoy!).

But I wonder whether a considerable part of that positive response comes not from pleasure at having expectations (eventually) met, but from the sheer pleasure of mental arousal, which itself carries a basically neutral valence. In other words, the music itself produces the aural equivalent of fireworks, and we delight in this sense of activity. No one, I think, would argue that fireworks are enjoyed because of the tension between our expectation of darkness and its temporary violation by explosives. There is an aesthetic beauty in the way each sparkling trail combines with the others to produce a transient pattern, and in the rich contrasts of glowing colours against velvety black.

There is probably an optimal level for this sort of stimulation: fireworks filling the entire canopy of night would be overwhelming and confusing. This position has been argued most cogently by philosopher Peter Kivy. He points out that some of the finest music can’t meaningfully be described in terms of ‘garden-variety’ emotions such as happiness, sadness, anger or ‘lovingness’. Think, he says, of the sacred music of the Middle Ages or the Renaissance, of Josquin des Prez and Thomas Tallis. This is beautiful and magnificent to a degree that can easily move us to tears – and yet the music is in a real sense not expressive, certainly not in terms of attempting (as Mozart’s music might) to elicit a particular emotional state.

Yet this is not, Kivy reassures us, because the music is invoking some abstract ‘aesthetic emotion’ without a name. We know what this feeling is: it is ‘excitement’, or ‘exhilaration’, or ‘wonder’, or ‘enthusiasm’. I would suggest that it is, more precisely, “the excitement that you feel when listening to music” that we can best understand it with explicit reference to its cause or object, just as we can describe love as the feeling we get in relation to our loved ones. This doesn’t necessarily mean that the fundamental feeling is music-specific, but it does suggest that it is not well captured by reference to the ‘garden-variety’ emotions that are the focus of a great deal of music psychology.

One could also object that Kivy’s notion of exhilaration or excitement sounds suspiciously close to psychologist Steven Pinker’s notorious suggestion that music is a kind of sonic cheesecake, a mere massaging of the auditory sense. But this essentially hedonistic quality is not like the satisfying stimulus of sugar and fat, or the sensual nature of a warm bath (at least, it is not just that). It holds within it something of the joy of being alive and in community with others. It is partly a kind of wonder at realising what other minds are capable of creating.


This is an edited extract from The Music Instinct: How Music Works and Why We Can’t Do Without It, 2010.

Image credit: Mait Juriado