LOOKING INTO BILINGUALISM THROUGH THE HERITAGE SPEAKER'S MIND

Abstract

Due to their unique profile as childhood bilinguals whose first language (L1) became weaker than their second language (L2), heritage speakers can shed light on three key issues in bilingualism - timing, input, and cross-linguistic interaction. The heritage speakers of focus in this dissertation are Korean second generation immigrants mainly exposed to their heritage language (HL) when young but who became more dominant in their L2 later in life. The ability of Korean heritage speakers in both their HL (Korean) and L2 (English), including speech perception, translation priming, and grammatical intuition were examined. Six psycholinguistic tasks, a bilingual experience questionnaire, and Korean and English proficiency tests were administered. Data were collected from 48 Korean heritage speakers, 36 English speakers learning Korean as adults and 36 Korean speakers learning English as adults. The two L2-learner comparison groups also served as native speaker controls for their respective native languages. The Korean heritage speakers raised in an English-speaking country, despite having been exposed to Korean first and throughout their lives, exhibited significant weaknesses in their Korean competence while exhibiting (near-)native-like competence in English. It is thus argued that the input-dominance switch that occurred before the critical period ended caused a dramatic reorganization of early/first established linguistic representation, which challenges some previous views on the implasticity of human language representation (e.g., Pallier et al, 1997). When compared to adult L2 learners of Korean, heritage speakers exhibited a slight advantage in speech perception and translation priming while showing no advantage in the grammaticality judgment of locative alternation. It is therefore suggested here that heritage speakers may have an advantage over adult L2 learners with early-acquired linguistic features and with implicit processing capacity. Another notable finding is that Korean heritage speakers showed less-than-nativelike performance in locative alternation in both Korean and English, a finding that highlights cross-linguistic interaction in bilingualism. The standard practice of comparing bilinguals to monolingual competence in SLA studies is thus called into question. Finally, although individual differences among the heritage participants in the current study were best predicted by language aptitude and amount of instruction, no conclusive claim regarding the role of language aptitude or instruction in early bilingualism is proposed here because it is unclear whether such effects influenced the childhood bilingual development or re-learning during adulthood of the current heritage participants. In short, timing, input, and cross-linguistic interaction all seem to contribute significantly to the development of bilingual competence. The heritage speakers examined in this dissertation turned out to be an excellent testing ground for all three of these ingredients of language acquisition.