Speech recognition and word learning in 24-month-olds: The roles of non-native speech and familiar words

Abstract

After 18 months of age, infants’ lexical representations are sufficiently flexible to recognize acoustically unfamiliar productions as variants of familiar words (Best et al., 2009; Mulak et al., 2013). For novel words, 24-month-olds still have trouble generalizing from native to non-native pronunciations, although exposure to the accent improves recognition (Schmale et al., 2011, 2012; White & Aslin, 2011). In the present study, we tested how quickly 24-month-olds could use exposure and familiar words to learn the meanings of novel words, on-line, when listening to Spanish-accented speech. In the exposure phase, twenty-four-month-olds heard a novel word embedded in a dialogue that either contained linguistic cues to the referent’s animacy (The vep is eating) or was uninformative (The vep is right here). Infants then saw two pictures at test, one animate and one inanimate, and were asked to find the vep. Looking time to the two potential referents here was compared to performance with native-accented speech (Ferguson et al., under review). Infants hearing Spanish-accented speech struggled to learn the referents of novel words, but differences also emerged between the accent conditions for familiar target words. Infants are slower to access the meanings of familiar words when hearing non-native speech.