The Music Instinct, By Philip Ball

"Writing about music is like dancing about architecture." That statement, which is most often attributed to Elvis Costello, is quoted in this bright, complex and occasionally profound book, and it highlights just what a tricky task the science writer Philip Ball has set himself. Because here, he's trying to examine what music is, how it works and why it exists. It might seem strange to ask the seemingly basic question of what music is, given the ubiquity of it in the world (no human culture has developed without some form of musical expression), yet it is incredibly difficult to define what we mean by music.

"Organised sound" was one definition proposed by the avant-garde composer Edgar Varèse, but that doesn't quite cover it. As Ball argues, music isn't really created by composers and musicians at all – the real music-making occurs in the listener's brain, in the unbelievably complicated mental processing which can take a collection of vibrations in the air and transform them into something which can make us cry or laugh or tremble, or feel something we simply can't put into words. "Experiencing music is an active affair," writes Ball, "no matter how idly we are listening. If it wasn't, we would not be hearing music at all."

Ball takes as his starting point the assertion by the cognitive scientist Steven Pinker that music is "auditory cheesecake" – that it developed, in evolutionary terms, only as a by-product of language, and that we now use it simply for gratification. The debate about the evolutionary origins of music rages in academic circles, with every new piece of research announced by ethno- musicologists seeming to contradict the last one. But Ball does a great job of sifting through this mass of data to get to what we really know, as well as, more often than not, what we still don't.

The Music Instinct is chock-full of fascinating little questions which Ball, as far as he can, attempts to answer. How do we choose the notes we use in composing, and why? How many of the factors affecting what we like to listen to are innate, and how many are learnt culturally? How are we so good at distinguishing timbre, the rather vague tonal aspect of sound that makes Tom Waits and Ella Fitzgerald singing the same note sound so obviously different?

Many of these questions were unanswerable a few years ago, but brain-mapping techniques are beginning to show how much of an interactive process listening to music is. And it seems that the best music is the stuff that challenges us just enough.

Along the way, Ball explodes plenty of myths. There is apparently no evidence for the "Mozart effect" of increasing babies' brainpower through exposure to classical music. On the other hand, genuine tone deafness is actually extremely rare (and, bizarrely, is not incompatible with having perfect pitch).

If there is one small gripe, it's that The Music Instinct does sometimes get bogged down in technical stuff. The jacket blurb claims that no specialist knowledge of music or science is needed, but this reviewer (PhD in physics and amateur musician) occasionally struggled, especially on the musical side, where talk of cadences, tonics, chroma and modes once or twice brought Costello's quote to mind.

Overall, this is a truly fascinating and eye-opening account of a phenomenon so commonplace we barely think about it, yet one which is also mind-bogglingly complicated. Once you've read The Music Instinct, you'll never listen to music the same way again.