The communicative phenomena of tonal music and speech observed in all human societies differ qualitatively from other human sound expressions. This difference consists mainly of the fact that both tonal music and speech are generative, i.e., they are composed of a limited number of discrete, perceptual units organized according to some tacit rules. In the case of tonal music, these units are experienced as pitch classes ordered in time. Listening to tonally organized pitch classes leads to the experience of specific (...) and relative emotions. These emotional qualities are called tonal qualia (often described as different tinges of tension and relaxation, or instability and stability), which are elicited thanks to our predictions concerning musical pitch structure. However, the emotional component of the recognition of musical structure seems to be incomparable with the experience of speech structure or other generative phenomena. Therefore, I propose that emotional reactions to mentally discrete pitches became an important part of the cognitive mechanism designated for music perception. I suggest that the general mechanism of prediction which acts permanently during the perception of the external world was specifically incorporated into the processing of musical pitch structure. As a result, a new, cortico-subcortical loop evolved that enables the recognition of musical pitch structure. The connection and exploitation of separate, evolutionarily old mechanisms for a new adaptive function is accordant with the evolutionary operational rule called “tinkering.” I will also discuss possible adaptive functions of pitch structure and the evolution of the ability to recognize tonal qualia by means of the Baldwin effect. (shrink)