Harmony of the neurons
December 9, 2008 2 Comments
I’ve just finished reading This is your Brain on Music, by Daniel Levitin. It’s a good read, even though it sometimes feels like it was written using an outliner, with sudden changes of topic and facts seemingly thrown in at random.
I’m a non-musician, so I found the parts on the structure of music really interesting. I was quite surprised to find out how arbitrary things like the Western scales are. I always assumed music would tell us a lot about the brain – for instance why should a minor key sound sad – but I didn’t know enough about the subject to find the answers. It does seem to me that music is playing on our innate sense of dynamics, and we interpret melodies using the same mechanisms we use to predict the movement of projectiles and prey, although Levitin seems to think the major/minor thing is mostly just cultural conditioning. He does support this dynamics hypothesis when he talks about the “trajectories” of melody and how different composers use this knowledge to exploit our expectations. Expectation seems to be the key, since he points out that we like music that’s challenging enough to be unpredictable yet simple enough not to sound random. It’s a kind of exercise for the brain that tunes our ability to predict in general.
For me, the most interesting parts relate to memory and classification. He demonstrates that we store cortical schemas consisting only of properties that are invariant. For example, if you try to sing Bohemian Rhapsody you’re very likely to sing it in the right key and at the right tempo. Because you’ve only ever heard the one rendition of it, you store all of its properties in memory. But sing Happy Birthday and it could come out in any old key, because you’ve heard it in so many keys and tempos that your brain only stores the relative notes (or, at least, all the distinct memories interfere to cancel out the key). Key isn’t an invariant property for Happy Birthday but it is for Bohemian Rhapsody, and key is stored separately from melodic shape. This is interesting because sensory invariance is the most striking aspect of perception, and up until now I’ve only thought about it from the angle of vision (e.g. we can recognize an object regardless of its location, scale and orientation). But invariance applies to hearing and touch too, and I hadn’t thought of it in terms of the construction of schemas for categorizing the world at large. It gave me pause for thought.