Dialect Differences in the Perception of English Stops as a Function of Prosodic Conditions by Korean Learners of English

Abstract

The current study investigated how tonal and non-tonal dialect speakers of Korean perceived the voicing contrast of English stops across different prosodic boundaries and accent conditions. Voice onset time and fundamental frequency of target segment were systematically manipulated across IP and Wd levels under accented and unaccented conditions. Both non-tonal Seoul and tonal Kyungsang listeners required a longer VOT for the perception of English voiceless stop in the IP level than in the Wd level. The ambiguous VOT range was also found to be different between the IP and Wd levels and a relatively higher onset F0 was required in the perception of voiceless stop for the stimuli in the ambiguous VOT range. The results from logistic regression analyses revealed that for both dialect groups, the effect of position was found to be significant, while the effect of accent was not found to be significant. The coefficient value of the VOT was greater than that of onset F0 for both dialect groups, meaning that when distinguishing the voicing contrast of English stop across different prosodic conditions, both dialect listeners used VOT as a primary and F0 as a secondary perceptual cue. However, the two dialect groups showed differences in the weight of each VOT and onset F0: Kyungsang listeners showed greater coefficient values of VOT and onset F0 than Seoul listeners. When the different prosodic conditions were given in the perception of the voicing contrast of L2, the preference and weightings of the perceptual cues were affected by the dialect of L2 learner’s native language.