Examining cognitive processing in simultaneous interpreting with text: What happens when speaker deviates from text?

Abstract In simultaneous interpreting with text (SIMTXT), speaker's deviations from their prepared script often bring challenges to interpreters' cognitive processing. However, empirical research on this input variable and how interpreters cope with it are limited. It is also unclear whether more extensive script pre-reading would mitigate the impact of such deviations. This study explores two essential factors concerning the working conditions of SIMTXT: auditory-script incongruence (congruent vs. incongruent) and durations of preparatory script reading (long-time vs. short-time). Twenty postgraduate interpreting trainees participated in the experiment of English-to-Chinese SIMTXT comprising two stages: preparatory script reading and actual interpreting. Utilizing eye-tracking, we gauged participants' cognitive effort in script processing and visual attention distribution during SIMTXT. Our findings reveal the following: (1) Lexical attributes of the source text (word frequency, lexical density) exerted more pronounced influences on script processing effort than incongruence or preparatory script reading duration. (2) Speech-script incongruence significantly influenced visual attention allocation, manifesting as increased proportions of dwell time and fixation count on the video rather than the script. Furthermore, occurrences of numbers and proper names also influenced attentional processing. (3) Longer preparatory script reading amplified the (in)congruence effect on attention allocation.