Individual differences in first and second language sentence processing: evidence from statistical learning

Abstract

This dissertation investigates why individuals differ in their success in learning language--in particular, why some second language (L2) learners learn more quickly and more successfully than others--by examining the correlation between ability in statistical learning tasks involving a nonadjacent dependency and success in processing filler-gap constructions in English and Korean. The results show (a) that statistical learning ability influences English native speakers' reading times in processing English object relative clauses, (b) that an effect of statistical learning ability is observed in L2 learners' comprehension accuracy and reading times when processing English object relative clauses, and (c) that Korean native speakers' reading times when processing topicalized object sentences in their native language are associated with statistical learning ability. These findings shed light on the role of statistical learning in first and second language processing, implying that individual variation in this area may be one factor that explains why individual differences appear in language learning ability, at least in the case of comprehending filler-gap dependencies.