Abstract
This work tests the effect of heritage language background on imitation and discrimination of prevoicing in word-initial stops. English speakers with heritage languages of Spanish (where prevoicing is obligatorily present) or Cantonese (where prevoicing is obligatorily absent), as well as monolingual English speakers, imitated and discriminated pairs of stimuli differing minimally in prevoicing, both in English (participants’ dominant language) and Hindi (a foreign language), and they also completed a baseline word reading task. Heritage speakers of Spanish were expected to show the highest performance on both imitation and discrimination, given the contrastive status of prevoicing in Spanish. Spanish speakers did indeed show more faithful imitation, but only for Hindi, not English, sounds, suggesting that imitation performance can differ based on language mode. On the other hand, there were no group differences in imitation of prevoicing in English or in discrimination in either language. Imitation was well above chance in all groups, with substantial within-group variability. This variability was predicted by individual discrimination accuracy, and, for Cantonese speakers only, greater prevoicing in baseline productions corresponded with more faithful imitation. Overall, despite an expectation for differences, given previous evidence for the influence of heritage languages on production and perception of English voiced stops, our results point to a lack of cross-language influence on perception and imitation of English prevoicing.
Funder
Natural Sciences and Engineering Research Council of Canada
Subject
Linguistics and Language,Language and Linguistics