Gender differences in voice onset time perception

Abstract

It has been known for a while that a gender difference in voice onset time (VOT) production exists such that males typically produce shorter VOTs than females (Swarts, 1992; Robb et al., 2005). However, potential gender differences in VOT perception have not been systematically investigated. This study addresses this question and asks if male and female listeners of English exhibit a different crossover point in a VOT continuum between a voiceless and voiced sound. 11-step VOT continua between /t/ and /d/ (VOT ranges from 70 ms to 5 ms) were created and a set of identification tasks were performed. Results show female listeners switch from a /t/-percept to a /d/-percept at an earlier VOT step (i.e., with a longer VOT) than male listeners do. That is, when female listeners start to categorize a token as a /d/-sound, male listeners still categorize the same token as a /t/-sound. This difference in the perceptual pattern between males and females mirrors their difference in production. VOTs for /t/ are typically shorter for male than female speakers, and male listeners categorize shorter VOT tokens as /t/ more often than female listeners do. This finding highlights the role of social factors like gender in the production-perception link.