ERROR ANALYSIS OF GOOGLE NEURAL ENGLISH TO ARABIC MACHINE TRANSLATION: A COMMUNICATIVE COMPETENCE PERSPECTIVE

Abstract

This study examined Google neural machine translation from the communicative competence perspective. The communicative competence linguistic framework consists of linguistic, sociolinguistic, discourse, and strategic competences. The communicative competence framework was implemented to explore what interlingual error patterns are more likely to impede and subvert bilingual translation and communication. The findings of this study indicated that Google’s online translation tool produced errors with various degrees of severity in the four aspects of communicative competence. However, most of Google translation errors appeared to be linguistic and discourse-based errors. Additionally, the findings indicated that the quality of Google’s translation was adequate but needed moderate human’s editing to appear comprehensible. The author of this study concluded that Google tended to produce better results while translating texts that were created to address a public audience and written in a serious tone of writing.