Examining readers' use of machine translation through eye tracking

This study examined the use of Google Translate among 50 Japanese learners of English engaged in task-based reading. There was one novel word per paragraph, and the study analyzed the frequency at which the participants accessed machine translation (MT) as opposed to the use of other strategies (e.g., dictionary use). Using eye tracking, it also examined how learners viewed translated paragraphs on MT pages. Finally, the study analyzed the relationship between the MT usage data and reading outcomes (content and vocabulary recall). A sub-group of participants provided both bilingual dictionary and MT access more often used the dictionary than MT. However, participants that were not provided dictionary access tended to be overreliant on MT, using it for irrelevant passages and when clear context cues were present. Eye movement data revealed that most participants read the English text first before checking MT, but some participants rushed to access translations. Participants displayed various strategies when viewing MT. The study found that aiming to work out the meaning from the text first and then checking the translation to search relevant novel lexical items was associated with positive outcomes. Potential implications for L2 education are discussed.