Bilingual Infants' Accommodation of Accented Speech

Abstract

Infant word recognition is sometimes hindered by variability in the speech input. Previous research has shown that, at 9 months, monolinguals do not generalize wordforms across native- and accented-speakers (Schmale & Seidl, 2009). In the current study however, it was predicted that bilingual infants would be advantaged in accommodating for accented speech due to experience with phonetic variability across their two phonological systems. It was also predicted that this hypothesized ability would be restricted to accommodating for an accent derived from a familiar language (e.g., French-English bilinguals would accommodate for French-accented English but not Mandarin-accented English), since this type of variability would be consistent with the language sounds to which infants were regularly exposed. Study 1 set the experimental stage by identifying native and non-native speakers with similar voices, as perceived by a group of adults. This was done in order to restrict variability across speakers to differences in accent, rather than biological differences in voice (e.g., a higher or lower pitched voice). Following speaker selection, acoustic measurements of vowels and word stress placement were taken to compare native and non-native speakers and confirmed several expected deviations between native and accented speech. Study 2 tested the hypothesis that bilingual infants would be advantaged in accommodating for these deviations when the accent is derived from a familiar phonology. Using a headturn preference procedure (HPP), 9- and 13-month-old English-learning monolinguals and French-English learning bilingual infants were tested on their ability to recognize familiarized English wordforms across a native- and French-accented speaker. Bilinguals in both age groups succeeded in generalizing wordforms across speakers, however monolingual infants failed regardless of age. Study 3 tested whether bilinguals’ success would persist when the accented speaker’s first language was unfamiliar. Infants in this study failed as a group to generalize across native- and Mandarin-accented productions of English wordforms. However, bilinguals who received balanced exposure to their two languages performed better in accommodating for Mandarin accented speech than unbalanced bilinguals. This hints at a general ability to ignore irrelevant phonetic information, perhaps due to an advantage in cognitive control.