A Study On Student Translators’ Performance In Post-Editing

With the utilization of machine translation,post-editing has become a common practice in the translation industry.However,previous studies primarily focus on its didactic and practical aspects.Few investigate student translators’ post-editing performance.Therefore,based on the error typology by Temnikova and Farrus,this study categorizes machine translation errors into errors at lexical,phrasal and syntactic levels and investigates student translators’ post-editing performance from two dimensions of error detection and error correction.In this study,27 student translators are selected and divided into two groups of beginners and advanced learners,according to their CATTI levels and their postediting experience.The study adopts a mixed-method approach by combining postediting task,screen-recording,cued retrospective interview and semi-structured interview to analyze the discrepancy of two groups’ overall post-editing performance and their performance in detecting and correcting errors at the three linguistic levels.The reasons for the discrepancies are then elaborated.Data analysis shows that in error detection,the overall performances of the beginners and the advanced learners were satisfying(detection rate≥86%),no significant difference was found between the two groups.In error correction,the correction rates of the two groups were found much lower than their detection rates,and the advanced learners performed significantly better than the beginners.In addition,regarding student translators’ performance in detecting and correcting errors at lexical,phrasal and syntactic levels,the advanced learners were found better than the beginners only at the lexical level in error detection.However,in error correction,the advanced learners were better at all three levels.The reasons for the discrepancies of the two groups’ performance in error correction can be attributed to student translators’ staying at different stages in translation skill acquisition,the differences in their bilingual competence and tool & research sub-competence,and their application of varied revision procedures.At the theoretical level,the error classification applied in this study enriches the English-Chinese machine translation error typology.At the methodological level,the mixed-method approach adopted in this study provides referential value for postediting studies.At the practical level,this study offers a new perspective for empirical studies on post-editing and translation practice by focusing on translators’ post-editing performance in error detection and error correction;additionally,it provides pedagogical implications for translation teaching and translator training.