Dictating Translations With Automatic Speech Recognition:Effects On Translators’ Performance

Technological aids play a critical role in improving translators’ productivity and reducing their workload.Previous studies on translators’ interaction with technologies have found a productivity increase brought about by the use of automatic speech recognition(ASR)tools for translation dictation.However,these studies have primarily focused on European language pairs and had relatively small sample sizes.They have also largely overlooked other important aspects of translators’ performance,such as translation quality and cognitive effort.The current study aims to investigate the impact of ASR input on translators’performance in terms of productivity,time allocation,editing operations,cognitive effort,and translation quality,as well as whether source-text difficulty influences these factors.To achieve this,the study tested the performance of 60 Chinese translation trainees when using ASR input and keyboard input.Participants were randomly assigned to either a dictation group or a typing group,and they each completed two English-Chinese translations of different levels of source-text difficulty.The typing group used only the keyboard and mouse for text input,while the dictation group used ASR software to dictate a draft of their translation,and then used the keyboard and mouse to make error corrections and revisions until they were satisfied with the text.Data were collected using keylogging,subjective ratings,screen recording,and a questionnaire.The study found that using ASR for drafting reduced participants’ typing effort,without compromising the quality of their translation outputs or leading to a significantly higher level of cognitive effort.Source-text difficulty did not have a significant impact on these results.The dictation group and the typing group had comparable productivity but showed significantly different time allocation patterns during their translation process.The analysis of the revisions made by the dictation group and the results of the post-test questionnaire shed light on ways to optimize ASR systems for translation purposes.This study offers insights into the applicability of ASR support in English-Chinese translation and contributes to a better understanding of translators’ behavior under different input modes.