摘要:Within the theoretical and methodological framework of Conversation Analysis, the present study explores the nature of the native speaker (NS) and nonnative speaker (NNS) identities in repair practices of English conversation. It has identified and analyzed in detail repair sequences in the dataand has also conducted quantitative analyses in order to investigate the relationship between NS/NNS identities and repair practices. The results show that the categories of NS and NNS identities are social products that are invoked during and through the participants’ ongoing interaction. First ofall, the participants did not invariably make their NS/NNS identities relevant to their repair work; specifically, more than half of the repair practices were irrelevant to these identities. Secondly, when the participants’ NS/NNS identities were made relevant to repair, both the NS and the NNS participants initiated repair in order to solve interactional problems that had resulted from the NNS participants’ insufficient linguistic abilities. The participants’ orientation to their NS/NNS identitieswas frequently triggered by the NNS participants’ repair-initiation while the NS participants invoked these linguistic identities only in a limited range of interactional environments. The findings of this study provide a further empirical basis for arguing the importance of an emic approach to NS-NNS interactions and the potential to learn new vocabulary through repair.