Species counterpoint is entirely mathematical, as is (to a great extent) common-practice partwriting. Yet, I guarantee you that most people who listen to that music would think it is entirely musical, when in fact very little is more than rule-following. (And yes, this is for a lot of genres! Sonata form, pop structure, etc., etc.)
Also, involving the work of David Cope, his rule-based composition systems repeatedly passed the turing test.
(I'm hoping, once I get to grad school, for my master's thesis to be around this exact idea: the computational aspect of creativity, and how the idea of "musicianship" is really just a human attribute we hold on to.)
I agree that music theory is mainly descriptive rather than prescriptive, but the fact that many works of genius center around patterns, well... it's hard for me to believe that we aren't heavily dependent, on some level, on math.