How to Empower Machine-Translation-to-Web Pre-Editing From the Perspective of Grice’s Cooperative Maxims

Abstract

This article takes pre-editing to a new level, shifting its purpose from improving the linguistic accuracy of the machine-produced translation (MT) to boosting the communication effect of MT. Adopting the method of MT textual analysis, this research asked ten university students to read and assess two sets of MT versions generated from pre-edited and non-pre-edited texts retrieved from the web pages of five companies in Taiwan. Students were interviewed to express their preferences and make comments. The findings showed that 90% (9/10) of student respondents perceived the MT texts, which were generated from pre-edited text with a shorter length, as more communicative than the MTs produced from non-pre-edited texts. Additionally, 100% (10/10) of student respondents agreed that the MTs, which were generated from pre-edited texts with concise textual contents and thematic relevance to subheadings, were more communicative than the MTs produced from non-pre-edited texts. The respondents’ feedbacks are compatible with Grice’s maxims of quantity, quality, and relation. This article concludes by highlighting the new direction of revising the information amount, content, and thematic relevance of a web text, and supports extra-linguistic pre-editing from the perspective of Grice’s cooperative maxims.