Re: CIMESP-Results Fwd:

Subject: Re: CIMESP-Results Fwd:
From: Richard Wentk (
Date: Sun Nov 06 2005 - 09:53:21 EST

At 23:50 04/11/2005, you wrote:
>Well, that's an old theory, but I think it's wrong. The problem is that we
>listen to
>pieces we like over and over again and still like them. I know the bach a-
>concerto pretty well -- I first heard it as a kid an it inspired me to
>take up the violin.
>I've done a ton of computer analysis of it and discovered many things
>about it. I just
>listened to it the other day and I felt "jeez -- this performance is way
>too fast." I still
>get a tremendous kick out of the piece.
>So the goodness can't be that I can predict what's happening or not, since
>I can actually predict
>everything that happens, assuming I've noticed it before. There must be a
>different explanation.

I think you're confusing structural and perceptual prediction here.

Structural prediction happens at the level of the score. If you already
know how the notes go, there will indeed be no structural surprises.

But perceptual prediction still has a momentum of its own. Knowing that the
leading note will be resolved by a step up to the tonic doesn't mean the
*experience* of that move is any less novel just because you've heard it
thousands of times before. So a familiar piece can still surprise at the
perceptual level, because e.g. V VI will always be experienced as more
surprising than V I, even if you know perfectly well what's going to happen

Musically it's a step away from the default momentum. Or, if you like, from
the expectation built up in the perceptual model in the listener's mind
which tries to predict what happens next.

Bach is particularly good at adding surprises at every level, so the music
is more information rich than that of a hack composer who would be more
likely to compose using cliches and defaults. In terms of an analogy to
information theory a new symbol is required every time the music avoids a
default. So it's hard to see how the theory is wrong - although it's also
hard to argue that it's complete. It could only be complete if the
model-making perceptual process were fully understood.

Still - this process of model-building and prediction seems to be key to
musical perception. If it didn't happen we could only ever perceive music
as isolated sound moments.


This archive was generated by hypermail 2b27 : Sat Dec 22 2007 - 01:46:14 EST