In this presentation, I review the perceptual research related to auditory representation for music. The research suggests that multiple representations exist concurrently in the auditory system, and that the dominant representation is shaped by the specific auditory environment. I note that the research is consistent with theories of competitive representations, such as Edelman's neural Darwinist approach. I propose that the difference in predictive accuracy for different representations provides the feedback mechanism by which competing representations are selected. Repercussions for cognitive modeling of music are discussed.