Second language acquisition in 6- to 8-year-old native Spanish-speaking children : ERP studies of phonological awareness, semantics, and syntax

Abstract: Most people in the world and about a fifth of all school-aged Americans speak at least two languages. Nevertheless, little is known about second language (L2) processing in development, even though language proficiency is strongly related to success in almost all domains. Whereas behavioral studies of L2 acquisition in children are abundant, neurocognitive studies of L2 processing typically are limited to adults with several years of exposure, who may use general cognitive mechanisms to compensate for any difficulties in L2 processing. Research on bilingual adults suggests that age of acquisition (AoA) and proficiency have different effects on different aspects of L2 processing. The present study therefore recorded event-related brain potentials (ERPs) in order to index processes of phonological awareness (Rhyming effect: RE), semantics (N400), and syntax (LAN, P600) in bilingual and monolingual children 6-8 years of age. Even though behaviorally, bilingual children with an average AoA of 4 years had lower English proficiency than monolingual children, proficiency predicted similar differences in ERPs across groups: greater proficiency was linked with shorter latencies and higher amplitudes of all ERP components. Latency in these cases represents speed of processing while amplitude of ERP effects in children can v be thought of as an indication of detection of the introduced violations. The appearance of the anterior rhyming effect, latency of the posterior rhyming effect, along with the distribution of the anterior ERP effect for phrase structure violations were related to AoA. More specifically, bilingual 6- to 8-year olds of higher English proficiency processed rhyming nonwords slower than 3- to 5-year-old monolingual children, which could have a strong impact on later vocabulary acquisition. Differences across lingualism groups in distribution of the anterior negativity elicited by phrase structure violations could indicate different neural generators for processing of syntax. Noteworthy is that differences in processing as illustrated by these ERP effects were recorded even though in both these cases bilingual children’s English proficiency were within the normal range expected of monolingual children of similar age. Early acquisition was thus important for processing of rhyming and for more automatic syntactic processing as revealed by differences in the anterior negativity.