Slide Slam C10 Sandbox Series
Differences in Processing of Voice Onset Time in Bilingual Italian-German Children and Monolingual German controls.
Theresa Bloder1, Valerie Shafer2, Tanja Rinker1; 1Catholic University Eichstaett-Ingolstadt, 2Graduate Center, City University of New York
Exposure to more than one language influences children’s phonemic development (e.g. Kehoe & Kannathasan, 2021). Bilingual children’s neural commitment to their two languages and hence their phoneme processing abilities vary in accordance with the relative amount of exposure to each language (Sebastián-Gallés & Bosch, 2002). Mismatch Responses (MMRs) are known to index automatic neural discrimination of speech contrasts (Yu et al., 2020). By combining neurophysiological and behavioral measures we examine the perception and production of Voice Onset Time (VOT) in the two languages of bilingual Italian-German speaking five-year-olds and a control group of monolingual German peers. VOT is a salient phonological feature in many languages that has often been used to study the impact of bilingualism on speech production and perception as its specific realization differs across languages (e.g. Kupisch & Lleó, 2017; Lein, Kupisch, & van de Weijer, 2015). German and Italian differ in that respect, with German contrasting short lag VOT with long lag VOT and Italian contrasting short lag VOT with voicing lead. Twenty-two five-year-olds (12 monolingual German and 10 bilingual Italian-German children) were tested using the neurophysiological potential Mismatch Response (MMR) to speech-sound contrasts. Stimuli consisted of bilabial stop consonants. The short lag stop (0 ms VOT), common to both languages was used as the standard. Four deviant VOTs were selected: 92 ms (ASPeasy) and 36 ms (ASPdifficult) lag for German; and 112 ms (PREVeasy) and 36 ms (PREVdifficult) lead for Italian. MMRs were derived by subtracting the standard from each deviant (Yu et al., 2020). A picture-naming task was used to elicit productions of 10 initial-stop words in each of a child’s language/s. The acoustic measurements were undertaken in Praat (Boersma & Weenink, 2013). Language background was assessed using a parent questionnaire. MMRs were greater for the easy compared to the difficult condition. Preliminary analysis revealed that both groups of children showed MMRs of an initial positivity followed by a negativity, similar to Shafer and colleagues (2010). Mixed ANOVAs revealed a significant group difference for ASPeasy (p=0.017). The monolinguals showed a greater positive MMR at 100-150ms. For ASPdifficult a significantly greater negative MMR (p=0.013) was observed for the monolingual group around 200-250ms. No group difference was observed for neither of the prevoiced stimuli. Examination of the means show that generally bilinguals showed a greater negativity compared to the monolinguals. A repeated measures ANOVA revealed that bilinguals produced significantly distinct VOT categories in their two languages (p=0.001). Post hoc tests revealed that bilinguals, in fact, had formed a total of four distinct VOT categories. Our data show that a greater acoustic difference between the standard and the deviant enhances the MMR. Further, our findings are consistent with Yu and colleagues (2020) suggesting that bilinguals may pay more attention to the speech signal thereby enhancing their negative MMR. It an unexpected result that we did not find a significant advantage for the bilingual children when processing the prevoiced contrast. Explanations for this will be discussed within the context of the children’s language background and their VOT production.