This study explores nine EFL learners' writing processes with MT when writing in English and discusses the results of a questionnaire survey of 123 EFL learners about their use of MT strategies in FL writing. The findings suggest how and why EFL learners might use MT in FL writing.
The application of neural-network-based artificial intelligence to machine translation (MT) technology has improved MT to the point where it may now be a useful tool for people learning and using foreign languages. So far, however, few researchers or educators have investigated how EFL (English as a Foreign Language) learners might use MT as strategies in foreign language (FL) writing. This study first explores learners' writing processes with MT based on data collected using screen capture software and stimulated recall in case studies of nine EFL learners writing in English. It also discusses the results of a questionnaire survey of 123 EFL learners, regarding their use of MT strategies in FL writing. By integrating quantitative and qualitative data within these studies, the results of the analysis showed that (1) differences were observed in terms of the number of words in participants' writing with MT and writing directly in English, (2) a total of 8 strategies specific to FL writing with MT were identified in the case of intermediate EFL learners, and (3) there were three strategies clusters within the respondents of questionnaire surveys about their use of MT in FL writing, which we called the top strategy cluster using machine translation with a variety of strategies, the middle strategy cluster, and the lowest strategy cluster without adjusting input source and using back-translation. Based on the results of these studies, this study attempts to discuss how and why EFL learners might use MT in FL writing.
