As you can kind of tell by reading between the lines of the articles about music in that New Scientist issue, music scientists don't really have a clue as to what this thing called "music" (which they are studying) actually is, let alone what it might be for, if indeed it is for anything.

My own "super-stimulus" theory is based on the assumption that music is a super-stimulus for an otherwise hidden component of speech perception. The most recent development of my theory, as explained in Music: A Drug, Which Used To Be Stronger Than It Is Now, Musical Immunity and Auditory Super-Cheesecake and Music: The Hidden Sense and the Three Ways of Knowing, is that this hidden component is responsible for our ability to slowly but steadily acquire a useful "worldview" directly from what other people tell us.

In the past, our response to music, as a type of "false speech", was so strong that it compromised our perception of reality, and this may even have led to our near extinction as a species. We have since evolved partial resistance to the effects of music, which explains our somewhat ambiguous emotional response to music, i.e. music makes us "feel" emotions, but we know that they are not real, especially once the music has stopped.