A Revisit to Recast Efficacy in L2 Learning: An Alignment Perspective

ABSTRACT This study examined recast efficacy in L2 learning from the perspective of alignment. As the hallmark of interaction, alignment facilitates L2 learning when models of linguistic structures proceed and prime subsequent output. This input‐before‐output sequence runs counter to the output‐before‐input array associated with the way of delivering recasts which may induce weaker alignment effects and hence limited recast efficacy. To verify this reasoning, 72 Chinese learners of English were randomly assigned to three interaction modes: output‐first recast group, input‐first alignment group, and meaning‐oriented control. A pretest–posttest‐delayed posttest design was employed to detect improvement in post‐interaction production of ‐ ed , followed by an interview that looked into alignment with ‐ ed during interaction. Results showed that recast and control groups achieved less post‐interaction gains than the alignment group in oral and written production; furthermore, during interaction they demonstrated more content orientation whereas the alignment group showed more aligned language use. These findings suggest that providing input models before output can combine with recast delivery to increase the alignment effect on language learning.