Perceptual switching in Spanish/English bilinguals

Abstract

Previous studies have shown that bilingual native speakers of languages in which /p t k/ are realized with short-lag voice onset time (VOT) values identify stops with short-lag VOT as voiceless more often in an L1 perceptual “set” than in an L2 (English) set. The purpose of this study was to explore the basis of language set effects. The listeners in Experiment 1 identified nine short-lag Spanish /t/ tokens as “t” or “d”. The Spanish /t/s were identified as “t” more often when they were presented along with short-lag English /d/ tokens than when they were presented along with long-lag English /t/ tokens. Much the same small “phonetic context” effect was observed for Spanish and English monolinguals and for two groups of Spanish/English bilinguals. As expected, native Spanish listeners identified the Spanish /t/s as “t” more often than the native English listeners. The fact that native English listeners often identified the Spanish /t/s as “t” indicated that VOT was not an overriding cue to the voicing feature for them. In Experiment 2, the short-lag Spanish /t/ tokens were presented in Spanish and English perceptual sets. Much the same small “language set” effect was observed for the monolingual and bilingual listeners. The size of the “phonetic context” effect in Experiment 1 and the “language set” effect of Experiment 2 was much the same for most tokens. Detailed acoustic analyses failed to reveal acoustic dimensions that would reliably differentiate the short-lag Spanish /t/ tokens that were predominantly identified as “t” from those that were ambiguous between “ t” and “d”. Taken together, the results indicate that the so-called language set effects observed for bilinguals in previous studies probably do not arise from the application of different language-specific criteria for voicelessness to naturally produced stops with short-lag VOT.