Quantifying the manipulation of F0 as an acoustic correlate of stress in Rebkong Amdo Tibetan

Abstract

Quantifying the manipulation of F0 as an acoustic correlate of stress in Rebkong Amdo Tibetan Rebkong Amdo is a non-tonal variety of Tibetan spoken in Rebkong County, Qinghai Province, PRC. Speakers draw on varied acoustic resources to convey stress in disyllabic words, which are highly frequent in the language. Non-verbs (nouns, adjectives, and numerals) are stressed on the second syllable (S2); the primary acoustic correlate of stress is F0. Verbs are stressed on the first syllable (S1); stress is conveyed by both intensity and F0. For non-verbs, speakers manipulate F0 in different ways to convey S2 prominence. The pitch contour may (a) be flat in both syllables but higher in S2; (b) be flat or slope gently in S1 but fall sharply in S2; or (c) be generally flat in S1 but define a prominent arcing curve in S2. These three patterns, which occur in free variation, require different analytical approaches in order to quantify relative acoustic prominence across syllables: (a) mean F0 (Hz) is measured for each syllable; (b) pitch slope (in Hz/100 ms) is measured for each syllable; (c) F0 curves are characterized using a principal components analysis, and by determining the tonal center of gravity. In each approach, measurements are paired to determine the magnitude of the F0 contrast across syllables in individual words; these are then aggregated, demonstrating that speakers’ use of F0 to convey S2 prominence is statistically as well as perceptually significant.Quantifying the manipulation of F0 as an acoustic correlate of stress in Rebkong Amdo Tibetan Rebkong Amdo is a non-tonal variety of Tibetan spoken in Rebkong County, Qinghai Province, PRC. Speakers draw on varied acoustic resources to convey stress in disyllabic words, which are highly frequent in the language. Non-verbs (nouns, adjectives, and numerals) are stressed on the second syllable (S2); the primary acoustic correlate of stress is F0. Verbs are stressed on the first syllable (S1); stress is conveyed by both intensity and F0. For non-verbs, speakers manipulate F0 in different ways to convey S2 prominence. The pitch contour may (a) be flat in both syllables but higher in S2; (b) be flat or slope gently in S1 but fall sharply in S2; or (c) be generally flat in S1 but define a prominent arcing curve in S2. These three patterns, which occur in free variation, require different analytical approaches in order to quantify relative acoustic prominence across syllables: (a) mean F0 (Hz) is measured for eac...