Dear me! You expect a scholarly answer!
I'll refer you back to a conversation I had with Mike Keiser in Chicago a little over a year ago. He came to our pre-dinner get-together with a paper which basically set out the options for canons, mirror-image canons and so on. It sounds very academic, but I was able to point him to lots of musical examples from Machaut to Mozart and contemporary music. We had a fruitful conversation, during which, I think, he was looking for a justification for a reversible course. My musical answer was that in musical terms the listener should not be aware of it.
The analyst can, if he/she searches hard enough, find these things, but the success of the music depends upon its being satisfactory in its own way. It may be marvellous that you have written a fugue on the notes A BAD EGG, as it is easily possibly to do, but is it a good piece of music in its own right?
This is where the GCA analogy begins to make sense. Every designer can be given A BAD EGG with which to work. Only a few make a work of lasting musical enjoyment out of that theme.
It really doesn't matter what musical form we take. That is only - in golfing terms - a few givens:18 holes, links/heath/parkland/highland etc etc. The structure is largely given to you. What you do with it is the genius. There are great symphonies, sonatas, rhapsodies, canons, etc etc. The form is largely immaterial. What matters is what you do with it.