Subject: Minor third calls From: "Jeremy Day-O'Connell" <jdoc(at)UCHICAGO.EDU> Date: Sat, 29 Jan 2005 16:29:20 -0600Dear Colleagues, As the author of the original query, allow me to clarify. Linguists who study the intonation patterns of English have widely reported use of the descending minor third in chants and calls (e.g., Lieberman; Bolinger) or more generally, in "stereotypical" or "predictable" speech (Ladd). Similar phenomena have been reported across many languages (e.g., in Hirst and DiChristo, eds., 1998). It is frustrating, however, that such claims appear to be mostly anecdotal rather than empirical. ("Air Ball" is one exception: see musicologist Cheryl Heaton's article in _PMS_ 1992.) It's more frustrating still that many linguists tend not to measure pitch with any precision (and among those who do, some use pitch _difference_ instead of pitch _ratio_!). Nevertheless, the phenomenon is extremely recognizable to this English-speaker, and I find the universality claim entirely plausible. Hence my plea: If anyone knows of any responsible research that bears on this purported universality, I'd be grateful for the reference. Now as for my hypothesis. Perhaps "scene analysis" conjured up the wrong ideas. As Al Bregman writes (here and in his excellent and well known publications), grouping by pitch proximity is a common strategy in forming auditory Gestalts. What I'm imagining isn't grouping but something more like salience or conversely, habituation. Let's assume a natural environment in which most sounds are noise, and the few relatively "tonal" sounds (the whistle of the wind, the creaking of trees) change in pitch only slightly and only slowly. Then a good way to be heard, or rather, "noticed," would be to project two clear tones at some significant interval; listeners would be liable to infer that something non-random (i.e., information-laden) had happened. But of course, as Bregman would also point out, grouping by _timbre_ stands to preserve the unitary identity of the sound source. Since both the timbre and the loudness of the (amateur) human voice vary with pitch, the caller faces an opposite constraint as that described in the previous paragraph. Hence, make the interval big, but not too big. (I had been tempted by the critical band but couldn't imagine its relevance to what is, after all, a _melodic_ phenomenon; I think Christian Kaernbach's idea is a good one.) Hope that's clearer. Thank you all for your thoughts, and I welcome more. Jeremy -- Jeremy Day-O'Connell Postdoctoral Fellow in Music Theory Music Department University of Chicago 1010 E. 59th St. Chicago, IL 60637 W (773) 702-8668