Machine Translation Post-Editing Behavior of Undergraduate EFL Students Across Two Genres

Numerous studies have explored to what extent and in what ways second language (L2) learning or L2 writing quality is impacted by the use of machine translation (MT) tools or post-editing the MT output, as well as how L2 learners perceive the effectiveness of MT tools. However, the potential of MT, more specifically post-editing machine-translated texts, to improve L2 learners’ translation skills remains unexplored. This study reports on how the MT post-editing behavior of undergraduate students of English as a foreign language (EFL) differs across two genres – short story and research article. Turkish learners of L2 English (n = 25) who were enrolled in a second-year online translation course post-edited Google-translated short stories and sections of research articles (from Turkish to English) for better translation. Using the Wilcoxon signed-rank test, we compared their post-edits in the two tasks for overall frequencies, linguistic levels of the post-edits (morphological, syntactic, semantic, and pragmatic), the success of the post-edits, and the types of post-edits (addition, deletion, substitution, reorder, rewrite). Our findings revealed significant genre effects on several aspects, with the short story genre promoting more student engagement with the post-editing task than the research article genre. Pedagogical implications are discussed.