Raising language awareness in peer interaction: a cross-context, cross-methodology examination

In this paper, both a cognitive and a sociocultural theoretical perspective are used to bring together findings from two studies that investigated the effects of instruction designed to enhance the potentially positive effect of peer interaction on L2 development. Despite differences between the studies’ learning contexts, participants’ age, and research design, a comparative examination of their findings promotes our overall understanding of whether and how instruction that teaches students to engage in reciprocal L2 learning can be effective. In Study 1, Japanese university students were taught how to provide corrective feedback to each other during communicative peer interaction activities. Pre- and post-tests were administered and statistical analyses were conducted on the change in frequency of corrective feedback and self-initiated modified output as well as in overall accuracy of spontaneous production. In Study 2, Grade 3 and 4 Canadian French immersion students were given strategy instruction to collaborate on task- and language-related problems. Their audio-recorded interactions were qualitatively analysed for contextual factors affecting peer corrective feedback. Based on the combined findings, it is concluded that language awareness can be enhanced through peer interaction but a reciprocal mindset among learners plays a significant role in deciding its outcome. Keywords: peer interactioncorrective feedbackawareness-raisingcollaborative learningcontextual factorscognitive and sociocultural approaches Notes 1. We do not separately discuss prompts and recasts based on the results that (1) prompts and recasts were comparable in terms of frequency, (2) both CF types were followed by comparable amounts of modification moves, and (3) the prompt and recast groups showed comparable developmental patterns (see Sato and Lyster [Citationin press] for more details). 2. Note that Levelt's speech production model draws on adult monolingual speakers and, thus, lacks developmental perspectives. The current discussion follows SLA research that employed Levelt's model (e.g., Crookes, Citation1991; Kormos, Citation2000; Towell et al., Citation1996).