An acoustic and phonological description of /z/-devoicing in Southern American English

In this paper, we investigate whether z-devoicing is a feature of Southern American English. 37 students from around Virginia, half of whom identified as Southern, half of whom did not, were recorded completing a picture naming task and a reading task designed to elicit /z/-final tokens. They also answered questions about their orientation and attitudes toward their hometown, Appalachia, and the South in general. Participants’ final /z/ tokens were automatically categorized as being [z] or [s] using the FAVE Aligner (Rosenfelder et al. 2011) and then acoustically analyzed in terms of duration, spectral energy measures, and voicing. The results of the categorical analysis and voicing analysis suggest that people who identify as Southern /z/-devoice at higher rates than those who do not, but only in pre-pausal environments. In fact, there appears to be significantly more voicing in Southern speech in pre-voiced environments. Additionally, within the Southern identifying participants, those who express more ...