Talkin' Blues

By a random coincidence, two explorations of language and semitonal pitch crossed my path this week. First, Mark Liberman over at Language Log had a long post on pitch analysis of everyday speech. A team of Dutch linguists analyzed recordings of volunteers reading excerpts from Winnie-the-Pooh—first, some dialogue of Tigger (the happy, hyperactive tiger), and then Eeyore (the pessimistic donkey). Their conclusion: the readers tended to use major intervals for Tigger and minor intervals for Eeyore. The results, needless to say, are highly sketchy; it's a small sample and a rather suggestive methodology. But it's a neat concept, the idea that people associate major and minor with happy and sad even in speech. (Go ahead and open up another can of nature-vs.-nurture over this one.) Liberman then took the next step and did pitch analyses of various other types of recorded speech, finding that most examples did tend to revolve around two or three approximate pitches. (I think that just might be a matter of the natural range of non-trained speakers. I'd be curious to see more results for really well-modulated voices.)

And then, by way of gracious commenter Valdemar Jordan's long-dormant blog, I found Joe Monzo's microtonal analysis of Robert Johnson's "Drunken-Hearted Man." Fascinating, especially his efforts to pin down the exact nature of the "blue" flatted third:

After analyzing this song, I believe I have found a possible reason for hearing the "halfway between" 3rd. Johnson sings a common blues figure in

in the second half of the first two lines. It gives an interval of 11/10 [= (2^(2/12) - 35 cents) = 1.65 semitones] between G+ 11¹ [= 5.51 semitones] and F# 5¹ [= 3.86 semitones]. If G+ 11¹ is interpreted casually as the 12-equal G 2^(5/12), it makes the F# 5¹ sound like it could be either F 2^(3/12) or F# 2^(4/12), or somewhere between.

Short version: the equal-tempered scale makes for a pretty clunky approximation of speech and vernacular singing.

But since my ears are more accustomed to music than speech, I think I actually tend to hear things the other way around. Here's an example. Most American and British speakers tend to end their sentences on the lowest pitch of the sentence. You can hear this in newscasters: there's a baseline "tonic" pitch; most of the sentence is pitched higher than that, but the full-stop punctuation is signaled by a return to that low tonic. (Peter Jennings, in particular, used to do a neat gloss on this: he'd deliberately undersell the final punch line of a story by putting the whole thing on the baseline pitch, like a closing tonic pedal point.)

My current favorite news voice, though, is a gentleman who's been filing reports from India and Pakistan for the NPR morning news roundup. (I haven't been able to catch his name, and the NPR website is useless in this regard. He was on two or three days in a row last week. Can anybody help me out?) His voice, which sits in a tenor-ish range, is very well-modulated: he uses a huge range of pitches, and, like a lot of Anglo-Indian and Anglo-Pakistani accents, the sound is quite melodic. What's interesting, though, is that he finishes his sentences in what, to my ears at least, is a typically subcontinental way—his final one or two syllables are higher than the "tonic" pitch of the sentence, which instead comes on the penultimate or antepenultimate syllable. If I had to put it in equal temperament, it might look like this:


I don't know if this is a feature of Indian or Pakistani languages, or if it's something that might have been picked up from indigenous musical style. To me, though, it sounds rather elegantly sly, and infectiously charming, regardless of context. Why? Probably because it reminds me of a jazz inflection—an ornamental blue note after the final tonic chord. It's completely unfounded judgment—my favorable opinion of this man's speech patterns is due to a totally unrelated and culturally irrelevant coincidence derived from my own idiosyncratic musical tastes. But we all form our impressions of people, at least in part, on the basis of the way they talk, and until now, I, for one, never thought very carefully about what that impression might be based on. I wonder how many other intuitive, snap judgments I'm making on the basis of my mental record collection?