An eye tracking study of cognitive effort allocation across translation subtasks

Abstract

As new technologies have been transforming the landscape of translation industry, translation has become a particular field of interest for research aiming to improve its workflow and optimize its working environment. The objectives of this project are to identify distribution patterns of translators' cognitive effort to different subtasks of translation during the translation process with the aid of a video-based eye-tracking system. This project report first reviews a number of models used in previous researches, as well as the mechanism and interpretation of eye tracking systems and data. The empirical investigation in this project attempts three major analyses: (1) analysis of generalized patterns of attentional distribution to compare the levels of cognitive effort required in respective subtasks; (2) analysis of the fixation duration and pupil size to investigate the cognitive workload among different subtasks under investigation and also the relationship between key events and fixation durations; (3) analysis of translators' working styles of with respect to cross interest area saccades and duration of follow-up fixations. The analysis results show that most attentional shifts take place between ST and TT, while no significantly different patterns of dictionary lookup activities prompted by the needs for comprehension versus production are identified. The analyses are presented together with discussions and references to previous models, and possible avenues to future research are also sketched.