On Student’s Self-repair In Computer-aided English-Chinese Simultaneous Interpreting

The application of artificial intelligence in the field of interpreting is manifested in Computer Aided Interpreting(CAI)and Artificial Intelligence Interpreting(machine interpretation).Though the Computer-Aided simultaneous interpreting system has been basically constructed,it will not replace human interpreters for the time being.Therefore,probing the computer-aided interpreting is of significance for complying with the interpreting model that combines human with machine.By comparing self-repairs in English to Chinese simultaneous interpreting with computer-assistance(CASI)and that in simultaneous interpreting without computer-assistance(SI),this study aims to answer the following questions: 1)How do the frequency and type of self-repairs differ in SI with computer-assistance and in SI without computer-assistance? 2)What are the reasons for the differences and how can the frequency of self-repairs be reduced? 3)Do the strategies summarized from the interpreting performance and retrospective interview reduce self-repair frequency for student interpreters after one-month practice?On the basis of mixed-approached research,eleven postgraduate students of interpreting studies were recruited as subjects to perform two rounds of experiments with a 4-week interval.Each experiment comprises a SI,a CASI and a post-interpreting retrospective interview.Strategies for coping with self-repairs in CASI were proposed after first-phase retrospective interview whose validity was examined in the second round of experiment.Findings of this study are as follows: 1)The frequency of self-repairs in CASI is much higher than that in SI.Self-repairs demonstrate themselves prominently as interruption with an editing term and the repeat of one or more lexical items,which fall into Covert repairs.2)Two reasons account for such disparities.Partly due to the greater demand for interpreters of cognitive capacity and better coordination between listening to source speech and referring to subtitles.Partly due to the malfunction of computer-assistant embodied in its delay,erroneous recognition,dynamic subtitles and suddenly shifted display board.3)The strategies deriving from the retrospective interview had been proven valid during the second experiment.Strategies are: A)Practice more to improve the energy distribution and cooperation with the machine,so that interpreter can render simultaneous interpreting by listening to the source language and referring to the subtitles at the same time.B)Slightly extend EVS to adapt to the delayed response of the computer-assistant.C)If interpreters intend to improve the accuracy of acquired source language and shorten the delay,they may mainly refer to the English transcription.D)If interpreters intend to reduce memory load and analyzing effort and better identify unfamiliar vocabulary,they may mainly refer to the Chinese transcription.The findings of this study try to provide implications for SI Teaching design that couples humans and machines.