As I suspected I would.... here goes:

Ears are not calculators.

Some are better at it than others, I suppose.

When we speak to each other we process a very complex set of *equations that are quite quickly and accurately (for the most part ) resolved into meaning, intent, accent and perhaps even *direction. Language is quite firm in most of those four divisions (I'm sure there are plenty other ways to analyze that I'm missing, but... and notwithstanding what we'all doing here)

Music doesn't make hardly any of those as clear as speech, yet in spite of this can be transcribed about as easily (with training, of course). Scribing the semantics of music is futile even if meaning is often prescribed (eg "fate"=Beethoven's 5th Symph. theme), but the meanings of 'musical' sounds are a personal overlayment. Intent, like language, draws from context, but unlike language (ie: without consistant meaning) is again strictly a personal endeavor. An "accent", comparing music and language, is probably the most mathematically based of those as the ear can tell the style of music (again, if trained) with about as much of a sample as one would need to tell what general accent a person is speaking... an "aural statistic" (so-to-speak).

The rules for music are much less stringent and at the same time, within this universality, we only listen/hear a fraction of the possibilities it offers and therefore create more anticipation of understanding of it than we do a wonder of its newness. This may speak toward my suggestion of 'direction'...

... Yada, yada, yada... etcetera, etcetera, etcetera...

********

OK. So... three more *sets of comparison do I offer. These I give as, again, part of what a composer of words (a poet, perhaps) and a composer of music does mold as they do thier thing:

Choice
Emphasis
Arrangement (props to T.S.Elliot, I believe)

With these in mind, are we not composing as we speak?

Granted: communication through sound is a primary goal *there, but other than 'trees falling in woods', what ain't?

Calculating (or perhaps 'deconstruction') is exactly what we do... and, of course, rather quickly reconstructing as a response/resolution, as well.

Just thinkin' out loud.