Sympathy for the devil

I have two half-Italian friends who have independently decided to flee that country, partly in despair at the state it’s in. The science magazine Sapere is trying to restore a little intellectual culture, and I'm glad to contribute a regular column on music cognition. Here is the latest installment.

_____________________________________________________________________

Many people who dislike the atonal music of composers such as Arnold Schoenberg say that it’s because their works are full of harsh dissonances: notes that sound horrible together. Schoenberg argued that dissonance is just a matter of convention: there’s nothing intrinsically wrong with it, it’s just that we’re not used to it.

The truth is a bit of both. Some dissonance really is convention: in the Middle Ages, a major third chord (C and E, say) was considered dissonant, but by Mozart’s time it was perfectly harmonious. But there’s also a “sensory dissonance” that stems from the basic physics of sound. If two pure tones very close in acoustic frequency are played together, the sound waves interfere to create a rattle-like sensation called roughness, which is genuinely grating. This seems to imply that any notes should sound okay as long as they’re not close in pitch. But because instruments and voices produce overtones with a whole range of frequencies, you have to add up all the possible combinations to figure out how “rough” two notes will sound together. The nineteenth-century German scientist Hermann von Helmholtz was the first to do this, and modern calculations confirm his findings: perfect fifth chords (C-G) and octaves have very little sensory dissonance, but all other two-note combinations have much the same roughness except for the minor second (C-C#), which has a lot.

So maybe Schoenberg was right! As long as we don’t play notes that are directly adjacent on the keyboard, shouldn’t any chord sound fine? Not so fast. Some researchers claim that we have an innate preference for the chords that are conventionally labelled consonant – that we like a fourth (C-F), say, more than a tritone (C-F#, often called the ‘devil’s interval’ and used to represent the demonic). These claims come from studies of very young infants, whose preferences about sounds can be judged from their attention or agitation. The idea is that, if the children are young enough, their preferences haven’t been conditioned by hearing lots of consonant nursery rhymes.

But is that so? Babies can hear sounds in the womb, and they learn voraciously. So it’s extremely hard to know whether any preferences are truly innate even in newborns. One study claimed to find a slight preference for consonance in two-day-old babies of deaf parents, who wouldn’t have heard their parents sing in the womb. But the evidence either way is marginal at best.

In any case, culture seems to over-write any innate tastes in harmony. The ganga folk songs of Croatia use harmonized minor seconds that are usually deemed intrinsically dissonant, while Indonesian gamelan music uses tunings that sound jarring to Western ears. In comparison, Schoenberg doesn’t seem to be asking so much.