The present study examines L2 reading proficiency effects on the relative contribution of vocabulary knowledge and grammar knowledge to L2 reading comprehension for Korean high school EFL learners. To this end, 200 high school students were asked to take a vocabulary knowledge test, a grammar test, and a reading comprehension test. The participants were divided into three sub-groups by L2 reading ability in order to examine L2 proficiency effects. Multiple regression analyses for the sub-groups indicated the relationships among the three variables as distinctive. The results showed that syntactic knowledge had a predictive power for reading performance in the high reading group, but vocabulary had the same quality in the intermediate reading group. For the low reading group, neither vocabulary nor grammar could significantly account for the L2 reading variance. Theoretical implications and directions for further studies are discussed.

This study looks at the employment of negotiation about form by a pair of advanced English L2 users engaged in collaborative composition tasks, and compares their negotiation with that of a beginner English L2 pair. Contrary to the increasing interest in negotiation for meaning within the L2 literature, there is little research that investigates how learners interact in negotiation about form contexts, where learners are required to explicitly talk about the form that they encounter. In particular， few studies have been conducted with learners at different proficiency levels in such contexts. Recognising this paucity, the study presents a holistic analysis of learners' negotiation about form generated by learners at different proficiency levels. This means that first, the negotiation about form was quantified in terms of language-related episodes (LREs); second, the same data was examined via an in-depth, descriptive analysis; third, delayed post-tests were conducted on specific linguistic items produced via negotiation about form. The study does not find much difference in LREs between the two proficiency levels of learners or convincing evidence that LREs lead to L2 learning at all. The results also reveal limitations in the relationship between the interactions engaged in and eventual learning. (196 words)