As the curtain fell on the first day of CIOL 2020, information overload fatigue set in and pangs of hunger began to override my concentration. We’ve all been there, half-heartedly listening to the last presentation of the day whilst resisting the urge to clock-watch and yearning for some fresh, unconditioned air. Luckily for me, Friday’s closing keynote speech, So you are a translator. What’s going on in your brain?, was anything but dull and sent me away from the conference with lots to ponder over dinner.
In his talk, Dr Binghan Zheng, Associate Professor of Translation Studies at Durham University, gave a fascinating overview of his cutting-edge research into the physical processes occurring in the brain as linguists perform translation tasks. Citing James Holme’s map, Dr Zheng began by explaining that, in comparison to the subject areas of both pure and applied translation, which have been fairly well studied, process-oriented studies are relatively few and far between. By using a range of scientific techniques, Dr Zheng and his team hoped to uncover the actual physical processes occurring inside our “little black box”.
By way of introduction, he began by discussing thinking aloud studies which, as the name suggests, rely on the verbal reporting of subjects about their thought processes at the point of translation. Known to yield intensely subjective results, Dr Zheng then pointed out a range of issues with the thinking aloud approach and moved to discuss other technological approaches which have produced increasingly insightful and scientifically reliable data.
Key-stroke logging, which involves the deployment of a computer programme, namely Translog, to track keyboard key selections during a task, was one such approach which offers greater insight into internal translation processes. The software produces a linear representation of how the translation task was carried out and helps identify areas of text where, for example, the translator, having already selected a word or phrase, goes back and deletes it in favour of another. While key-stroke tracking does offer more insight, Dr Zheng also pointed out it’s main limitations, such as how an individuals keyboard competence can potentially sque the data.
The next approach discussed by Dr Zheng was eye-tracking, which is used for measuring the point gaze or motion of the eye relative to the head. This is a well validated tool and has been employed across a number of scientific fields including cognitive linguistics, market research and behaviour studies. Dr Zheng described a range of experiments in which eye tracking revealed how the eyes move across a given text in strikingly different ways depending upon the task being performed. For example, while reading a portion of native language text, hot spot visualisation revealed that metaphors often require more cognitive effort to decipher. When studying the difference between reading for comprehension, versus for translating, gaze spot visualisation showed how, while reading for translation, the eye moves back and forth both horizontally along sentences and vertically between paragraphs in comparison to the steady left-to-right linear motion noted during comprehension reading. Going still further, Dr Zheng then showed gaze spot visualisations for reading during sight translation, versus while typing a translation. While there was marked eye movement in both cases, it was clear that, while typing a translation, the translator’s eye jumps dramatically in all directions across different areas of a given text compared with sight translation; constituting a veritable spiders web of blue dots and lines. While eye-tracking, like key-stroke logging, provides a helpful window into the cognitive processes involved, Dr Zheng was nevertheless keen to point out that tracking gaze does not necessarily give any information about other aspects of the subject’s cognitive processes, i.e. just because we know where a person was looking doesn’t tell us what they were thinking while they were looking there. Dr Zheng then noted that triangulation of all these investigations has, up until recently, been required in order to make sense of the possible cognition intimated by the results.
At this point, Dr Zheng moved to discuss his own research using the tools of cognitive neuroscience as first described by Maria Tymoczko, Professor of Comparative Literature at the University of Massachusetts Amherst. Dr Zheng’s research centred chiefly around two main areas of interest: identifying the different areas of the brain activated during reading and translating, and the different areas activated during L1 (target/native language) to L2 (source language) translation, versus L2 to L1 translation. He and his team deployed fMRI imaging while participants where asked to perform a series of randomised L1-L2, L2-L1 translation and reading tasks.
His results have shown increased activity in the left inferior temporal (visual word form) area of the brain, whereas during the translation tasks, there was increased activity in the dorsal prefrontal (attention, working memory), ventral prefrontal (language) and occipital (visual) regions. During translation, the basal ganglia is also activated, which is strongly linked with goal-directed tasks. Drawing his findings to a conclusion, Dr Zheng noted that the task of translation is indeed more effortful and recruits some domain general neural networks beyond domain specific (language) regions than standard reading. He also added that, of the two directions, forward translation (L1-L2) is more demanding and requires further cognitive resources and additional motor mechanisms. Dr Zheng concluded by stating that forward translation has more activation in the middle temporal lobe which functions greatly in conceptual mediation and representation as opposed to a more lexical mediation required during L2-L1 translation.
To stay up to date with all the lastest WMTS blog posts, please remember to SUBSCRIBE using the form at the bottom of the page.
All photographs by William Maitland ©2020