An Applicability Analysis Of Machine Translation And Post-editing Methods From Functional Translation Perspective

There are differences in the applicability of machine translation to different types of documents.Starting from the perspective of comparative study,previous study has argued for the applicability of machine translation for three types of text.Therefore,a new mode combining "machine translation" and"post-editing" was raised.The emerging mode has attracted attention,and several studies have discussed the specific method of post-editing.However,differences in the applicability and the ways of post-editing between various text types attract little attention.Based on a case study of two translations,namely the translation of report text and the translation of clinical process text,this thesis discusses the applicability of machine translation in the Chinese-English translation and the differences in the methods of post-editing between these two text types.The source text are first sent to the Deep-L translating engine and then checked for translational errors manually.In the present thesis,the Multidimensional Quality Metrics(MQM)framework is employed to identify and classify the extracted errors.The frequency of error and the distribution of error types are the main metrics to detect machine translation’s applicability.The applicability of machine translation on two text types is compared and then analyzed based on the text features.The present thesis presents some specific post-editing strategies for both text types.Differences in the methods of post-editing are also listed.In conclusion,machine translation is not equally applicable to both selected text types.The Independent Sample T-test for error frequency reveals that machine translation is more practical to Laboratory Text than report text.The present thesis found that using subtitles and non-subject sentences of report text is the two most substantial reasons causing the difference.After focusing on the post-editing methods for both types of text,it is found that the post-editing method of clinical process text focuses only on the lexical level,while the post-editing method of report text focuses on both lexical and syntactical level.The present thesis also concludes that the pre-editing method could effectively reduce the frequency of machine translation errors when dealing with report text.The results of the present thesis could benefit future research from several perspectives.From a theoretical perspective,the present thesis will help translators further understand the influence of text type on translation strategies and point out the shortcomings in today’s machine translation technology.Secondly,from a pedagogical perspective,the present thesis’s translating materials could provide fruitful teaching examples in the translation classes.Finally,from a practical perspective,the present thesis shows the differences in post-editing methods between report text and clinical process text,which could significantly increase translators,eficiency when dealing with these two text types.