(This is the continuation of a Twitter discussion with @RafalMista , whose is the opening quote.)
I’ll concentrate on “If there is a feedback between preferences and musical variants (transformed by preferences which are transformed by available repertoires of musical variants etc.)…”
I think we have to distinguish between various theories, e.g. the Darwinian theory and the CES theory which, as Nathalie Gontier says, has yet to be formulated. Per Darwin, musical variants can be melody, harmony, rhythm and timbre among others. A preference in a group is an observed tendency to, where there is a choice, do one thing and not the other, as play a syncopated rather than a non-syncopated triplet. That is an act of selection, which in a human being can only get out of the brain by muscle contraction, as Richard Dawkins notes in The Extended Phenotype. What I think you refer to as a cultural preference is not a unitary thing. A village band might consensually dislike syncopation but the clarinet player may slip a bit in nonetheless. The bandmaster frowns, but the clarinetist does it again at a dance and the kids cheer. That is an act of selection by the kids which may become ancestral to a new variant of the tune. Here at a stroke the trope “preference” disintegrates into what is actually happening. There is replication with fidelity (the band plays the tune with no syncopation time without number), there is variation within a very limited envelope (the clarinet player syncopates a triplet), there is selection (by the cheering young people at the dance), the clarinet player is encouraged to do it again at the next dance, at the third dance he is forbidden to do so by the bandmaster, he doesn’t syncopate, the kids jeer, the fourth time he does it again, cheers, the syncopated triplet moves towards fixation as an acknowledged variant. Maybe the bandmaster induces a split, taking the older non-triplet-syncopating conservatives with him. The syncopaters are the ones employed for dances, the non-syncopaters give sparsely attended concerts for a few elderly villagers. The syncopated version has now reached fixation in that village.
“I’ve read how people couldn’t dance to music played by a band from another village, as they had strong local preferences.” Perhaps they couldn’t dance because their dances (learned patterns of muscle contraction) didn’t fit the music (learned patterns realised by rapid muscle contraction of the musicians). This is an entirely different use of the vague word “preference”.
I think “perceptual niche” is a good term. But it refers to a limitation, a confinement to a narrow range of auditory experience of replication and selection delimited by the music that lies within that range—niche music in fact.
Sperberian tropes seem to me to belong to a world of make-believe. With Darwinism, Modern Synthesis, there has always been a conflict for those non-materialists who practise, or pretend to practise, science, but also believe in some sort of reality beyond science, to which science has no access. That’s fine, but the attempt to try to reconcile science and any form of “spirituality”, when the literal big money is behind the “spirituality” is pernicious, because it deliberately distorts science.
That’s why if it would save a lot of confusion if evolution meant Darwinism (MS). Anything else might more properly be called metaphysical culturalism.