SUMMARY
Interaction between learners is proven to be beneficial for second and foreign language acquisition. This article reports on a quasi-experimental study conducted in a German as a foreign language (GFL) classroom in a university in Spain. The study explored the effect of input modality on the provision of recasts in oral peer interactions. Two intact classes of GFL, one with 12 and the other with 16 learners, participated in six oral interaction tasks. The researchers divided the two classes into two groups: one group was only exposed to aural and visual input, while the other also received written input. Audio recordings and full written transcripts of learners’ oral peer interactions in the two groups were made and the frequency and nature of recasts were analysed quantitatively in order to examine the effect of input modality on the production of this corrective feedback type. Results suggested that input modality impacted the way learners interacted with each other. Findings showed that learners who were not provided with written input provided significantly more recasts to each other. Results also revealed that the nature of the recasts (i.e., form-, lexical- and pronunciation-focused recasts) that learners produced was input modality dependent. Since the provision of recasts is beneficial for second and foreign language acquisition, the results indicate that teachers and task designers should consider input modality as a relevant task design variable.