Terry Teachout had an interesting post the other day about musical keys and their emotional characters, and Fred Himebaugh responds (after initially writing about the topic here). The idea is that D-minor, for example, has a different emotional "feel" than, say, F-major.
I confess that I know very little about the internal theory behind this idea, but I've always been slightly skeptical about it: with the exception of people who suffer from synesthesia, I've never believed that people perceive musical tones in the same way that they perceive colors, and even then, colors themselves don't seem to me to do much at all emotionally if they are presented to the eye devoid of context. This seems to partly be Fred's point, as he wonders if people with "perfect pitch" would fare well when presented tones that are synthesized, i.e., totally divorced from the physical characteristics of an instrument. That would be an interesting experiment to perform. I once owned an electric metronome that also would sound, for tuning purposes, an A at 440. This latter feature was not nearly as useful as I originally thought it might be, because the machine's tuning note was a very harsh electronic sound that I found unsuited to the task of tuning my trumpet.
Another experiment might be to digitally "transpose", if such a thing were possible, an musical work into another key entirely. Everything would be the same about the work -- tempo, instrumentation, et cetera -- except for the key. Say, lower the entirety of Mozart's Symphony No. 40 by a minor third, so that instead of being heard in G-minor, it would be heard in E-minor. Would the resulting version of Mozart's symphony have the same character? I don't think it would, but as Fred notes, is that because the melodies would sound different, being pitched in a key where the tones sound different than the ones we're accustomed to hearing, or is that because E-minor actually does have a different "emotional character" than G-minor?
Anyway, it seems to me that it would be perfectly possible for a composer to take D-minor, that "saddest" of keys, and write a perfectly happy little minuet in it. But I might be wrong.
On a related note, Greg Sandow makes some interesting points about early music and our knowledge of early-music performance practices. It's interesting stuff, although I have to admit that I've never really been one to get worked up about whether a piece is being performed in a way that the composer him or herself, at whatever time in which they worked, would recognize it. I own very few "original instruments" recordings -- a full cycle of Beethoven and Berlioz symphonies, and a few Bach works -- and I'm perfectly happy with my recording of Handel's Messiah in which Sir Thomas Beecham commands musical forces that Handel himself would probably blush to hear. I'm not really a purist, so I've never really considered Sandow's point: that "purism" might just be an impossible proposition in the first place.