The Acquisition of Korean Reflexive Caki by Heritage Speakers and L2 Learners

Abstract

This study investigates acquisition of the Korean long-distance binding anaphor caki by heritage speakers of Korean, compared to English-speaking L2 learners of Korean and Korean native speakers. In English, a reflexive should be locally bound by its antecedent (e.g., Johni thinks that [Billj likes himself*i/j]), whereas it can be bound either by a local antecedent or a long-distance antecedent in languages like Korean (e.g. Johni-un [Billj-i cakii/j-lul cohaha-n-ta]-ko sayngkakha-n-ta). An experimental study using a sentence-picture matching judgment test with auditory stimuli was conducted with 32 heritage speakers, 20 English-speaking L2 learners, and 14 native speakers of Korean, comparing accuracy scores of judgments. The results indicated overall native-like competence throughout different oral proficiency scales in a range from ILR 2 (85% accuracy) to ILR 4 (100% accuracy), still obtaining statistically significant correlations between proficiency scores and individual, as well as mean accuracy scores. Comparing heritage speakers and L2 learners, a similar pattern was found in both groups for long-distance binding conditions, correctly accepting the long-distance antecedent in matched conditions and correctly rejecting the long-distance antecedent in unmatched conditions, with slightly higher accuracy scores for heritage speakers. However, differences were found in local-binding conditions. Heritage speakers accepted the local antecedent only 40% of the time in matched conditions, similar to native speakers, whereas L2 learners accepted the local antecedent 70% of the time even though both groups correctly rejected the local antecedent in unmatched conditions. The results of the study add to the pool of evidence showing the advantages of heritage speakers in the acquisition of syntax in SLA.