Acoustic cues and obstruent devoicing in Minnesotan English

Abstract

Previous research shows that the amount of glottal pulsing in American English obstruents varies by dialect, with some dialects characterized by high rates of devoicing. The present study is an examination of word-final obstruent devoicing in the Twin Cities metro area of Minnesota. A production task with 30 native speakers in their 20s, 50s, and 80s shows robust devoicing, which is conditioned by multiple linguistic factors. Nearly equal proportions (30%) of obstruents surface as fully voiced or fully devoiced, and the ∼rest surface as partially de-voiced. These results show that obstruent devoicing in the Twin Cities is fundamentally different from devoicing in the Iron Range of Minnesota. Secondary cues to phonological voicing are further examined, and in all cases, the cues are significant factors in the amount of glottal pulsing present in an obstruent, regardless of underlying voicing. However, the cues do not have a compensatory relationship with glottal pulsing. Finally, there are generational differences in the use of cues, suggesting a possible cue re-weighting: older generations rely more on glottal pulsing to signal underlying voicing, while younger generations equally use glottal pulsing and preceding vowel duration. In sum, devoicing in this region is phonologically stable, but phonetically changing.