This study investigates whether the usage patterns of resultative constructions (e.g., The lake froze solid; Tom painted the wall white) in argumentative essays written by Korean-speaking learners of English follow the usage-based principles influencing first language acquisition. Specific research questions addressed in this study are: a) whether resultative constructions with high complexity and low frequency are more difficult to learn for Korean-speaking learners of English than other types of argument structure constructions, and b) whether learners expand their choice of verbs in the resultative constructions as their proficiency increases. Multiple regression analyses showed that the occurrence of resultative constructions explains the variability of learner proficiency better than that of other constructions. In addition, learners employed less frequent and more varied types of verbs in the resultative constructions as their proficiency was higher. Our findings suggest that usage-based theories of language development hold true in the context of foreign language learning. ). For example, an initial stage of learning argument structure constructions is characterized by an item-based process. The repertoire of constructions that children use at this stage is mostly restricted to individual tokens, including simple lexical words, chunks, and some highly frequent argument structure patterns, such as intransitive (e.g., Tom cried), transitive (e.g., Jane ate pizza), and ditransitive constructions (e.g., Give me some cookies). In contrast, the acquisition of the complex and infrequent constructions, such as the transitive resultative construction (e.g., Tom painted the wall red) is shown to set in considerably later (e.g., Snyder 2001).The usage-based acquisition of progressing from acquiring frequent and salient items to building abstract knowledge of complex and less frequent constructions also manifests itself in the use of verbs. In the early stages of item-based learning, the verbs children use are mostly restricted to highly frequent words that are closely associated in meaning with the constructions that they appear in, e.g., want in [V-NP], give in [V-NP-NP], and put in [V-NP-PP] (Goldberg, Casenhiser and Sethuraman 2004; Ninio 1999). As children accumulate a larger repertoire of