Quality Assessment of Machine Translation from Translation Criticism Perspective—A Case Study of Chinese Translation of a Speech by Neural Machine Translator DeepL

Speech discourse, as one of the most important forms of discourse in human society, has dialogical features and is essential for understanding the viewpoints and positions. Meanwhile, the technology of machine translation (MT) has developed unprecedentedly, and the quality of MT has been continuously improved, which may facilitate the translation performance of speech discourse. At present, a paucity of systematic discussions converges on the performance of MT in the translation of speech discourse. To this end, from the perspective of translation criticism, this study adopts a translation quality assessment model to analyze the translation results of "Gettysburg Address" by the neural machine translator DeepL. The findings detect two ideational deviations and two interpersonal deviations without any negative equivalence. Hence, it can be safely concluded that the translation quality of DeepL is satisfactory, and it is feasible to be applied in the practice of English to Chinese translation of speech discourse although it still depends on certain context management and pre- and post-translation editing. The results of this study are enlightening for both English learning and teaching.