This study examines the nature of peer interaction, in particular pair interaction, of adult learners engaged in a range of activities in naturalistic EFL lessons. Six pairs engaged in seven different activities followed by a stimulated recall interview. Data from the transcripts of the audio-recorded pair talk and interviews were analysed to describe different types of pair interaction according to mutuality and equality. Five different types of pair interaction were observed during pair work: collaborative, cooperative, dominant/passive, expert/passive and expert/novice. Only one pair displayed a single type of pair interaction while the others showed more than two types of pair interaction. However, on the last activity, all six pairs showed collaborative interaction. These findings suggest that pair interaction can be collaboratively enhanced in a socially developed relationship between learners over time.

The purpose of this paper is to examine the effectiveness of corrective feedback on second language (L2) composition from both theoretical and practical perspectives. While it has been claimed that corrective feedback on L2 composition should be provided to student writers to prevent their errors from becoming fossilized, it has also been claimed that corrective feedback on L2 composition is useless because student writers make the same mistakes repeatedly even when they are corrected. Before we conclude that corrective feedback is wholly ineffective, however, a close reexamination seems warranted. This paper will approach the issue of corrective feedback on L2 composition from the perspectives of both those in favor of and those against it. By clarifying various issues, this paper seeks to provide educators with a better understanding of L2 composition teaching and learning.