Usage-based approaches to language acquisition explain language development as a gradual process of generalizing constructions through language experience. This study investigated second language learners’ development of constructional knowledge from the perspective of usage-based language development. A total of 169 Korean EFL students at five grade levels completed a sentence-sorting and a translation task. Results of the sorting task showed stronger constructional sorting as the learners’ grade level increased. Additionally, the sorting of intermediate-level learners was influenced by verb semantics such that the sentences including light verbs were more strongly clustered according to constructions than the sentences with heavy verbs, suggesting learners’ reliance on light verbs in the early stages of constructional development. Results of the translation task demonstrated a higher translation accuracy with increasing proficiency, but with a significant amount of variation across individual constructions contingent on the constructions’ syntactic and semantic complexity. Overall, our findings confirm the usage-based development of L2 learning.

This paper replicates the sorting task conducted on Bencini and Goldberg (2000) to examine whether Korean learners of English only concentrate on the meaning of verbs or they also pay attention to constructional meanings in English sentence processing. The results were broadly consistent with the original study―Korean learners of English were able to recognize the existence of English argument structure constructions and utilize the configurations in sentence interpretation. However, the experiment detected discrepancy in learners' understanding constructions according to the types. Language learners generally have a sense of English argument structure constructions and try to make full use of the representations, but their cognitive ability has not been sufficiently trained to access the internal relation among the constructions. Consequently, the learners feel difficulty in interpreting other types of argument structure constructions except transitive constructions and they tend to stick to the representations. Constructionists regard argument structure constructions as impetus and facilitator in language acquisition. Based on the perspective, the materials and the instructions provided to language learners need to be developed in a way that promotes the learners' access to the characteristics of English argument structure constructions.

This study explores the frequency distribution of the verb seem together with its three alternating complement structures: to-infinitive, that-clause, and adjectival/nominal complements. Based on an analysis of sample data selected from COCA, we investigate the conditions that may influence the choice among these argument structures. Our analysis has shown the following results: First, of the three complement structures, the that-clause complement was rarely used, a pattern consistent with the findings of previous studies. Second, discourse pragmatic factors such as information status and topichood of the subject, and judgement/evaluative semantics of complements were shown to be involved in the selection of the argument structure. Third, the preference for simpler structures evidenced in the usage data suggests the economy principle as a possible force behind the argument selection. The preliminary findings of this study should be complemented by future research with bigger sample data.

The current research aims to show that a transitive predicate/verb in one language may not necessarily be transitive in another. An alternative argument structure, Alternative Argument Structure Hypothesis (AASH), is proposed and advocated in this paper in order to account for erroneous production patterns noted and observed in Kim (2001), Park (2013), and Hong (2015a, b). Korean L1ers learning English as an L2 insert illicit prepositions between an English transitive V such as ‘marry’,‘kiss’,‘answer’etc. and its complement, yielding fairly high inaccuracy. It is argued that the erroneous patterns may be attributed to the asymmetries in the lexical argument structures between the Korean Vs and its English counterparts. The Korean counterparts of these English Vs are of the Sino-Korean origin light verbs (Han and Rambow, 2000, Choi and Wechsler 2002, Bak 2011) accompany‘-hata’. Under this proposal, it is the argument structure rather than morphological case as Montrul (1997, 2000), Ahn (2013), and Brown and Iwasaki (2013) have argued for that transfers to the acquiring process of the transitivity of the English Vs by Korean L1ers. The ramification of this study is that L1 grammar of argument structure transfers to L2 acquisition more fully and noticeably than has been assumed in the literature.

It has been a common view in the literature since Yang (1984, 1988) that caki ‘self' is a long distance anaphor, subject to parameterizaton of the binding domain across languages. I suggest that caki ’self' is a local anaphor with a DP structure including pro. I also suggest that a null argument in Korean has a DP structure with pro, too, based on the wide range of similarities with caki ‘self.' The present proposal for caki ’self' and the null argument has a nontrivial theoretical implication for anaphors and null arguments in languages typologically akin to Korean.