Translators’ ocular measures and cognitive loads during translation (Abstract)

Revised Hierarchical Model (Kroll and Stewert, 1994) predicts that translating into one's passive language is cognitively more demanding than translating into the dominant language. Such ocular measures as pupil size, number of fixations, blink frequency, and fixation frequency have been shown to be powerful indicators of a task's difficulty. The study aims to use these measures to test the assumption behind the Model. Two eye-tracking experiments were carried out. Experiment One's results showed translating into their passive language was more cognitively demanding than translating into their dominant language, as differences between pupil sizes, numbers of fixations, blink frequencies and fixation frequencies for the two tasks within subjects were statistically significant, using multivariate repeated measures analysis. Experiment Two also showed the same results as predicted by the Model, like Experiment One. For Experiment One, principal component analysis reduced the dimensionality of five ocular measurements to three components to describe the variability of the dataset: 1) number of fixations vs. task time; 2) fixation frequency vs. blink frequency; and 3) pupil size. Based on this, it would be advisable for future researchers to focus on number of fixations, blink frequency and pupil size, should they wish to replicate similar studies. Longitudinal analysis on the pupil size data demonstrated that there was a ‘familiarity effect’ (Hyönö et al.,1995) during both tasks, as the pupil size decreased in the middle of the tasks. However, this analysis also revealed that for both tasks ‘familiarity effec’ disappeared toward the end of both tasks, as the pupil size increased significantly in the end. This could be interpreted as sudden alertness to completion of the tasks or urge to proofread the translations.