ScholarWorks@숙명여자대학교

초록

This study investigates how statistical learning ability affects the processing of Japanese accusative relative clauses among native Korean learners of Japanese. Statistical learning―defined as the implicit acquisition of recurring patterns―is widely recognized as a key cognitive factor in second language acquisition, particularly in the processing of structures involving long-distance dependencies. Although previous research has primarily focused on head-initial languages such as English, considerably less attention has been given to head-final languages like Japanese. In the present study, participants completed a non-adjacent statistical learning task and a self-paced reading task designed to target accusative relative clauses in Japanese. The results showed that reaction times in the statistical learning task were significantly correlated only with the segment containing the verb of the relative clause; no significant correlations were found for the remaining segments. Furthermore, accuracy in the statistical learning task did not correlate significantly with reading performance. No significant accuracy differences emerged between relative-clause sentences and filler sentences in the self-paced reading task. Taken together, these findings suggest that, in contrast to English, statistical learning exerts only a limited influence on the processing of Japanese relative clauses. This restricted effect may stem from typological characteristics of Japanese, including (1) its head-final structure and (2) the absence of relative pronouns. The study underscores the need to consider both cognitive factors and language-specific structural properties when examining second language learning.

키워드