A Comparative Study On The Quality Of Human And Machine English Translation Of Mao Zedong’s Anthology

The study of corpus-based translation quality has advanced rapidly over the last two decades,yielding fruitful results in the field of corpus translation studies both at home and abroad.Nevertheless,there is still a need to further improve and enhance the research content and research methods,which will serve as inspiration for the future development of a more perfect research paradigm for the study of corpus-based translation quality.With the advancement of artificial intelligence technology,machine translation has gradually been applied to the field of translation of literary works,which greatly improves translation efficiency and gradually draws attention to the quality of its translations.The parallel corpus in the study includes the official English translation of Mao Zedong’s Anthology(1965 edition by Foreign Language Press),the machine translation obtained by the leading machine translation algorithm m BART and the original Chinese text,this study constructs an aligned Chinese-English corpus and selects the translation of the first volume of Mao Zedong’s Anthology for quantitative and qualitative comparative study on translation quality,including the linguistic and non-linguistic aspects.Furthermore,this study applies Julian House’s Translation Quality Assessment Model as a theoretical framework.With the help of corpus and statistical tools and methods,the overall translation quality and the mistranslations in the machine translation are summarized,and the reasons for the mistranslations are analyzed from multiple perspectives,providing targeted suggestions for the machine translation improvement.The comparative study reveals that the quality of official translation is superior to machine translation in three dimensions: language expression,translation strategy and mistranslation.In general,the official translation uses expressions that flow naturally,domesticated and covert translation strategies,and has fewer cases of mistranslation when compared to machine translation.Machine translation,on the other hand,is more faithful to the original language,and the translation strategy is more alienating and overt,with a greater emphasis on preserving the characteristics of the original language.However,machine translation has certain mistranslations,due to the lack of sufficient knowledge about the translation process,the lack of a detailed summary of language rules,and the constraints of context.Based on the classification and analysis of the causes,this study proposes strategies to cope with the shortcomings of machine translation,such as improving and expanding the corpus on relevant contexts,human-computer interaction and collaboration,and the integration of multiple technologies,in order to effectively improve the quality of machine translation and better optimize the level of artificial intelligence for future translation work.The significance of this study lies in the research model,research method and research content,which provides ideas and methods for more scientific and comprehensive corpus-based studies of translation quality.Furthermore,the differences in translation quality of political papers and leaders’ works between official translation and machine translation also provide some implications for literature translation research and the performance improvement of machine translation models.