Effects Of Recasts And Elicitations On Chinese Efl Learner Repair In Dyadic Interaction

The role of interactional feedback in second language acquisition (SLA) has recently received considerable research attention. Some observational studies have demonstrated that recasts lead to a high amount of immediate repair of erroneous utterances (e.g., Ellis et al., 2001; Mori, 2002; Sheen, 2004), while others have shown that recasts lead to a limited amount of immediate repair (e.g., Lyster & Ranta, 1997; Panova & Lyster, 2002; Lyster, 2004). These discrepancies indicate a need for further research into the role of such feedback in student-teacher interaction and the various factors that influence its effectiveness.Based on the Noticing Hypothesis, the Interaction Hypothesis and the Output Hypothesis, this study investigated the usefulness of two major types of interactional feedback (recast and elicitation) in dyadic interaction. To be specific, this study addressed the following three questions: (1) Do recasts and elicitations have differential effect on learner repair in dyadic interaction? If so, (2) Is there a relationship between learner repair as a result of recasts and elicitations and the type of errors? (3) Is there a relationship between learner repair as a result of recasts and elicitations and type of tasks?40 second-year students from a key university in Chongqing were arranged into two groups (recasts group and elicitations group). Each of the groups had 20 students. The study included a one-way task and a two-way task which were implemented between an individual student and the teacher in the way to complete the story. After the interaction, the audio-recordings were transcribed. The results showed that: (1) Recasts and elicitations do have differential effect on learner repair in dyadic interaction. Recasts were much more effective than elicitations in leading to learners'successful repair. (2) There is no relationship between learner repair as a result of recasts and elicitations and type of errors. Despite error types, recasts were remarkably more effective than elicitations on learner repair. (3) There is no relationship between learner repair as a result of recasts and elicitations and type of tasks. Recasts were remarkably more effective than elicitations on learner repair regardless type of tasks.Some theoretical and pedagogical implications can be drawn from the findings. Theoretically, the findings affirmed the claim that recasts are potentially more effective and further support the Noticing Hypothesis, the Interaction Hypothesis and the Output Hypothesis (Carroll & Swain, 1993; Nassaji, 2007). Pedagogically, the findings also suggested teachers should provide recasts with a falling intonation outside of the context on one-to-one bases in classrooms, which might be helpful for L2 learning. In addition, once drawing learners'attention to form, recasts were effective in leading to all three kinds of error repair. Finally, as long as recasts can draw learners'attention to form, recasts were effective in both types of tasks.