Dave Munger reports on a new music cognition study. Andrea Halpern et al, have shown that non-musicians can identify the difference between major and minor melodies, but only when they are labeled as "happy" or "sad" rather than "major" or "minor." This gets at the difference between perception and cognition. Musicians and non-musicians (I hate that label) alike can perceive the modal differences, but interpret them differently. This difference remained even when trained on either strategy. Now, the training was in the form of playing major and minor melodies and saying "this is major" and "this is minor", with no other theoretical underpinning. So it makes sense that nonmusicians would be more comfortable with antonym-pairings that make sense outside of musical jargon. The brain electrical activity maps are cool, showing typical denial-of-expectation reactions for the minor music among the musicians, and no expectations among the non-musicians (I still don't like that label).
Dave has written about other music cognition research here, here, here, and here.