University of Pennsylvania Working Papers in Linguistics


This paper explores how vocalic outliers are perceived in the context of an ongoing sound change, with the prior understanding that listeners integrate vocalic outliers in establishing speakers’ intended vowel output, but they do so with a bias (Labov, Baranowski, and Dinkin 2010). We focus on the short-e (DRESS) vowel that is undergoing generational change in lower Michigan. We conducted an experiment to determine whether local community members preferentially incorporate progressive outliers (the outliers in the direction of change) over other outliers into their perception of the vowel's mean position. The results indicate that vocalic outliers were integrated into mean calculation of speakers output, concurrent with the prior study (Labov, Baranowski, and Dinkin 2010). Listeners also displayed bias, as the low short-e progressive outliers were accorded more weight when compared with the high short-e outliers that are in the opposite direction of change. Yet this preference was almost exclusively displayed by the female listeners, who also produced a more lowered short-e in a reading task. This suggests that vocalic outliers are important indicators of the direction of sound change in progress, and that women show special perceptual sensitivity to diachronically progressive outliers.