Abstract

Aims and Objectives/Purpose/Research Questions: This paper investigates the extent to which current formal models of third language (L3) grammatical acquisition extend to sequential child L3 acquisition. We examine whether heritage speakers learning a foreign language as an early L3 transfer grammatical properties from the heritage language or the dominant second language (L2). Design/Methodology/Approach: We used a sentence repetition task and a picture story retelling task. The tasks focussed on grammatical phenomena that were either different between English and German, that is, verb-second and adverb order, or between English and German, on the one hand, and Turkish, on the other, that is, verb-complement order as well as subject and article realization. Data and Analysis: We tested matched groups of 31 Turkish-German and 31 monolingual German children learning English in grades 3 and 4, and we compared sentence repetitions as well as oral sentence production across different grammatical phenomena using parametric statistics. Findings/Conclusions: In both tasks, the two groups perform indistinguishably from each other, and both groups show selective transfer of grammatical properties from German. These findings suggest L2 transfer from a typologically related language in sequential child L3 acquisition. Originality: This paper breaks new ground by testing the applicability of formal models of adult L3 acquisition of grammar to sequential child L3 learners. It uses aural comprehension and oral production tasks with carefully matched groups of L2 and L3 learners of English to isolate the source of grammatical transfer in L3 acquisition. Significance/Implications: The research advances our understanding of cross-linguistic influence and unravels the dynamics of grammatical transfer in early child multilingualism. It adjudicates between current models of transfer in L3 acquisition in a multiple-methods design, it shows that these models apply to early L3 acquisition of heritage speakers, and it highlights that these models need to be expanded to include factors such as dominance and proficiency in prior languages.

Full Text
Paper version not known

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call