L2 acquisition of number marking: a bidirectional study of adult learners of Korean and Indonesian

Abstract

This study investigates the L2 acquisition of the Korean and Indonesian number systems by adult learners in light of the Feature-Reassembly Approach (henceforth FRA, Lardiere 2009). In considering this approach, Montrul and Yoon (2009) further suggest that subtracting or unlearning existing L1 features may pose greater difficulty than the addition of new L2 features, while Hwang & Lardiere (2013) propose that features more deeply embedded in a feature co-occurrence hierarchy, such as the [±human] feature, will be developmentally acquired later.

This dissertation presents the findings of two experiments examining the restrictions on quantifiers with classifiers and/or plural marking by L1 Indonesian-speaking learners of L2 Korean (Study1) and L1 Korean-speaking learners of L2 Indonesian (Study 2). Seventy learners at three L2 Korean proficiency levels along with 40 native controls participated in Study 1, and 61 learners at three L2 Indonesian proficiency levels along with 39 native controls participated in Study 2. The participants of both experiments completed three tasks--a Sentence Completion Task, a Grammaticality Judgment Task and a Multiple-choice Task--designed to probe knowledge of features and restrictions associated with Korean or Indonesian number marking.

The results of Study 1 showed that, although there was improvement in L2 Korean learners' performance with increasing proficiency, the advanced learners had still not acquired all the co-occurrence restrictions of the L2 Korean number system; specifically, L2 Korean learners, regardless of proficiency level, showed difficulty in acquiring the more deeply-embedded [human] feature associated with Korean plural marking. On the other hand, the results of Study 2 indicated that while low-proficiency L2 Indonesian learners initially showed some difficulty in subtracting L1-Korean-based features in L2 Indonesian plural marking, the advanced learners successfully acquired the restrictions and required feature matrices of number marking in the target language.

The current study suggests that the particular L1-L2 pairing will certainly impact what kind of feature restructuring is required in order to acquire new feature matrices of lexical items in a second language, which will in turn affect the developmental process of acquisition, thus lending additional overall support to the FRA.