Our next presentation will be this Tuesday, October 28, from 1-2 pm in room 002 and on Zoom. Massimo will present his work.
Title: Patterns of variation in sibilant acoustic dynamics
Abstract: According to conventional wisdom, sibilant fricatives require relatively long articulatory–and thus, acoustic–steady states due to their complexity. Recent studies have called this into question, however, finding gradual, continuous change in the acoustics over the time course of sibilant fricatives. These dynamics have been reported to vary across languages and possibly also across speakers of a same language, but methodological differences make it hard to compare results across studies. In particular, it is not clear to what extent differences in the acoustic dynamics of a same sibilant across speakers or languages may be due to confounds like speech rate differences or differing phoneme inventories. It is also unclear whether the dimensions of variation (and the amount of variation on each of these dimensions) differ between speakers and languages: the answer to this question could help determine whether dynamic information forms part of the target for sibilants (as it does in vowels) or has some other underlying source.
In this work, we examine the acoustic dynamics of /s/ across 32 typologically diverse languages. We find dynamic differences between languages even after controlling for prosodic and phonological factors, whereas inter-speaker differences in dynamics are comparatively constrained. We interpret these results prima facie as evidence that speakers are learning a dynamic target for /s/ in their language and adhering closely to it. We however observe a certain degree of family resemblance between related languages, which may conversely suggest that one or more missing variables play a role in determining patterns of acoustic dynamics
The full schedule and all relevant documents are available on this Google Drive.