Sunday, January 13, 2008

Music as a language

We all have heard that, and I know we all cringe. Even more if "language" is preceded by the word "universal." On last Thursday I went to a talk by David Pesetsky from MIT about the syntax of music and of language. It was a linguistics talk, and that was (for me) the most interesting part, learning about language. It's nice to reaffirm that there are many ways of looking at things, not to mention good to meet other people (not that I interacted with anyone beyond the enclave of us musicians in the back that much). And I don't want to criticize Dr. Pesetsky too much, working as he is from a layman's understanding of music theory (although I can't help but wince at the fact that all his musical definitions came from wikipedia). But the talk wasinteresting, if not always clear.

The gist was that language often gets rearranged ("I wonder [which person] Mary was talking to" as an example where the object of "to" is displaced) and the same syntactical possibilities are present in suspensions and appagiaturas. The largest gap, I think, in his theory is in his broadest application, viewing the fact that humans like consonance but listen to music that features dissonance as a paradox. And he's right that dissonance requires a syntax, but the connections he draws between whether a melody comes before or after (or both) a harmony change felt incomplete. Part of this was his wide array of linguistic examples, but only one musical example involving a melodic G-D-G-F-E against a V-I cadence. I would have loved to ask a question about whether more motion of elements was possible within the melody. For instance, E cannot go to F, but F and D could be switched. I do think he's on to something about layers of syntactic organization in plausible melodies, but I remain unsold that it's either as direct as he says (and he says they are exactly the same), or that this has as much to do with solving the "dissonance paradox" as he claims. Me, I like dissonance.

Actually, the most interesting part was a side tangent about the rhythms of lists, where each unit gets an equal beat, subdivided. So when you list "carrots, bread, cinnamon, yogurt, wine, watermelon..." each item gets a beat, subdivided into 2, 1, 3, 2, 1, 5. Similarly, if Peter and Susan are a couple, they will similarly fit into one beat when listing party invitees. A particularly telling moment was putting "and" between monosyllabic list elements, had been written as two even beats, but all of us independently turned it into a dotted eighth and sixteenth, and turned to each other. Funny what gets our attention.


  • Dear whoever you are,

    Thank you for coming to my talk at UMich, and for your interest in it. Since I started to present this work in talks last Spring, I have always benefited from the questions and comments of audience members more expert than me in music theory. What a shame that you didn't ask your questions during the discussion period, or afterwards.

    You make three substantive remarks that I'd like to reply to.

    1. You characterize what I called the "dissonance paradox" as "the fact that humans like consonance but listen to music that features dissonance", and respond with the objection: "Me, I like dissonance". I'm afraid this is not really what I was claiming at all. The claim was rather that tonal music may be perceived subjectively as consonant even when it is actually full of dissonances. See the 1984 paper by Barucha, cited in the bibliography of my handout, for a similar point.

    2. In discussing my canned example of "a melodic G-D-G-F-E against a V-I cadence", you write "I would have loved to ask a question about whether more motion of elements was possible within the melody. For instance, E cannot go to F, but F and D could be switched." I wish you had asked the question! In any case, part of the answer did come up in response to a different question. The theory predicts that the linear order of melodic units can be changed by syntactic movement only in the first structural domain in which the units were merged. (This was the reference to my own linguistic work co-authored with Fox.) So the transposition of F and D might be possible (though the result would be indistinguishable from a derivation in which they were just introduced in the opposite order), but I believe the same would not be possible with E and F.

    3. Finally, you make the interesting remark that a dotted eighth plus sixteenth-note pattern would be appropriate for a monosyllabic list like "wine and beer and cheese and...". On this topic, I would have had a *lot* to say...had you but asked. To begin with, I would have disagreed with your tablemates that the dotted rhythm is the *only* possibility (or even the preferable one) -- but I would have agreed that it is definitely available as an alternative. And it is the device normally used by speakers to disambiguate logically ambiguous conjunct structures such as those that mix "and" and "or" (e.g. "[Wine or juice] and beer" vs. "Wine or [juice and beer]"). I would also have mentioned some excellent work by the linguist Michael Wagner on just this topic.
 My suspicion, of course, is that Wagner's observations are relevant to the right theory of dotted rhythms in music, but I would have mentioned some problems that arise in making the proposal work correctly on both the language side and the music side.

    The work that I presented is preliminary, which is why it benefits so much from the questions and criticisms that I receive. (Some excellent and very important points were made by some of your colleagues in informal discussions afterwards.) Still, your posting includes one or two criticisms that I don't think the talk deserves. For example, it's just not true that I presented a "wide array of linguistic examples, but only one musical example". And as for my wince-provoking citation of Wikipedia, it was actually light-hearted, as you might recall. My mention of Wikipedia broke the ice, just as it was intended to. But while the audience was chuckling, the non-musicians among them were painlessly learning what a suspension is -- which was the intent. The more professional prose of, say, the New Grove would not have accomplished this particular task nearly as well.

    By Blogger kto-to, At January 14, 2008 at 10:23 PM  

  • Thanks (if you're reading this) for responding. The truth is I would have loved to ask questions and talked more, but I had to run to a meeting.

    I want to thank you for presenting some interesting work, teaching me something about linguistics, and continuing to think about music.

    On your comments
    1) I think that the a piece with dissonances is perceived as having dissonances, and in fact listeners will seize upon those. Maybe I just want a different word there. Because a twelve tone piece that consists entirely of consonant diads and triads will still sound consonant, but in a nonsensical way.

    2)Thanks for the answer!

    3)I'm sure it's not the only way, but it was the way all of us at the table felt it should be. Not statistically significant, but anecdotally so.

    Again, I don't mean to duck out of a Q&A only to comment elsewhere. I just wasn't able to stay, but wanted to get a chance to mention it. It's not often I get to see talks outside my department, and this proved quite engaging.

    By Blogger Dan B., At January 17, 2008 at 9:53 PM  

Post a Comment

Subscribe to Post Comments [Atom]

<< Home