Individual differences in cue weighting of stop consonant voicing in perception and production

Amanda A Shultz, Purdue University

Abstract

The purpose of the present study is to investigate individual variation in native English speakers' relative weighting of two voicing cues, the time between the release of a stop and the beginning of the vocal fold vibration (voice onset time or VOT) and the fundamental frequency at the onset of phonation (onset f0) in both a production task and a perception task. If perception and production representations are tightly linked, then there should also be a correspondence between the individual weighting that listeners give to specific cues in perception and the degree to which they use those same cues to instantiate a phonetic contrast in production. In the production task, participants repeated words drawn from 4 monosyllabic, monomorphemic minimal pairs with CVC syllable structure, differing in terms of their initial stop consonant (/b/ or /p/). For each production, VOT and onset f0 were measured and the relative weighting of each of these acoustic cues were computed for each participant using discriminant function analysis (SAS Institute Inc., 2002; Kondaurova & Francis, 2008). In the perception task, the same participants performed a labeling task of synthesized tokens along a two-dimensional continuum ranging in terms of VOT and onset f0 between /ba/ and /pa/ (Francis, Llanos, Dmitrieva, & Chapman, 2011). Individual participants' weighting of these two cues were calculated using logistic regression (Kondaurova & Francis, 2008), and the results of the two tasks were compared using correlation analyses. Results were expected to show that individuals' use of onset f0 and VOT in production corresponds to their relative weighting of onset f0 and VOT in perception. The results from the production task show a significant negative correlation, r (23) = -0.42, p = 0.037, of VOT and onset f0, while the results from the perception task show a nearly significant trend toward a positive correlation, r (23) = 0.36, p = 0.076, between VOT and onset f0. Across perception and production, there is a slight positive but not significant correlation between perception and production for VOT, r (23) = 0.1, p = 0.623, and a slight negative but not significant correlation for onset f0, r (23) = -0.34, p = 0.101. Possible explanations of the results of this study and implications for future research will be discussed as well as additional findings concerning prevoicing in production.

Degree

M.A.

Advisors

Francis, Purdue University.

Subject Area

Linguistics

COinS