Asymmetrical patterns of formant variability in English vowels

Abstract

Previous studies have claimed that lower formants should be weighted more than higher formants in a perceptual model of vowel perception (e.g., Schwartz et al., 1997). Given this formant weighting hypothesis, and if vowels have acoustic targets, vowels should be more variable in higher formant frequencies. Here, we examined within-speaker variability for five English vowels /æ, ʌ, ɔ, ɛ, ɪ/ in various contexts as produced by 32 speakers in the x-ray microbeam database (Westbury, 1994). For variabilities of the first three formants, only /ɔ/ follows this prediction (i.e., variability: F3 > F2 > F1) while /æ/ exhibits the opposite pattern; if we ignore F3 (as being less reliably measured), most vowels conform to the prediction (i.e., variability: F2 > F1), except for /æ/. Although the F2 variability is generally consistent with the perceptual model of vowel perception, it is also consistent with a possibly greater effect of consonant coarticulation (which is extensive here) on F2 relative to F1; this requires more testing. Further, while these results do not fully conform to the prediction made by the hypotheses, coproduction effects arising from the diverse contexts likely interact with the expected tendency. Correlation with observed kinematic variabilities will also be discussed. [Work supported by NIH grant DC-002717.]