What cutting-edge developments in cognitive neuroscience reveal about the translating brain. By Binghan Zheng
In this increasingly globalised world, language mediation has become widespread – from daily communication and business negotiation to international affairs. Over the past three decades, researchers have undertaken in-depth investigation into the cognitive process of translation by applying methods such as keystroke logging and eye-tracking in order to address a key question: ‘What goes on in the translator’s mind?’ This has greatly contributed to our understanding of the behavioural and cognitive processes of translation. However, the core question regarding the neurobiological mechanisms underlying the complex mental processes involved has remained one of the chief ‘known unknowns’ in Translation Studies.
Translation scholars and neurolinguists share an increasing interest in the interface between the cognitive process and neurophysiological evidence, which has been primarily supported by the development of powerful neuroimaging technologies such as EEG, PET and fMRI.
It is the general view of neurolinguistic researchers that for most right-handed individuals, the left hemisphere of the brain controls the functions relating to grammar, vocabulary and literal meaning. The two major areas associated with language comprehension and speech production (Broca’s area and Wernicke’s area) are also located in the left hemisphere. Confirming this, recent research has found that translation routes are strongly left-lateralised, and this is supported by empirical evidence that damage to this language-dominant hemisphere leads to severe dysfunction in translation tasks.
Forward vs backward translation
Translation directionality studies look into the differences between forward translation (from L1 into L2) and backward translation (from L2 into L1).1 Since there has been a ‘golden rule’ that forward translation should be discouraged, researchers have created experimental projects to test the evidence for this assertion. Kroll and Stewart2 were among the first researchers to study translation directionality empirically. Based on a study in which 16 bilingual participants conducted single-word translation tasks, they proposed a Revised Hierarchical Model (RHM) to account for the “translation asymmetry”: forward translation (FT) is more likely to involve time-consuming conceptual processing, while backward translation (BT) can be done by the simpler process of lexical association.
The RHM has triggered huge debates and a considerable amount of follow-up research, part of which cited behavioural evidence in its favour. Reaction time was found to be significantly longer for FT than BT irrespective of task type, language combination and the cognateness, concreteness or familiarity of the word stimuli. Some neuroimaging studies reported that FT involves significantly enhanced activation in areas of the brain associated with language production (Broca’s area), and also in areas associated with regulating movements.
Examining specific brain areas
Professional translation requires the development of specialised expertise through task-specific training, so researchers have asked whether translation requires a particular cognitive processing style that utilises specific brain areas. They have compared the cognitive load of reading-for-repetition and reading-for-translation, and found that the latter takes more time and requires more working memory, most likely due to the parallel activation of target and source lexical entries.
Neuroimaging technologies have enabled researchers to pinpoint the locations in the brain activated when a task is being carried out (known as neural correlates) and create an image of them as translators perform reading and translation tasks. By identifying the neural correlates associated with translation but not associated with reading, they have been able to track translation-specific processes or routes.
PET, fMRI and fNIRS evidence indicates that translation entails more cognitive effort than reading. The traditional ‘localisationist’ view is that language switching is mainly controlled by the frontal brain regions.3 However, this new evidence shows that translation engages neural networks in different areas of the brain, which implies that translation is embedded in more general linguistic and executive systems.
Simultaneous interpreting and extreme control
Another challenging question that has attracted researchers is how simultaneous interpreters deal with the extreme language use involved, and what sort of neural mechanisms enable them to control (perceive and produce) two languages simultaneously. Research in this area led by Alexis Hervais-Adelman and Barbara Moser-Mercer revealed that two regions associated with decision making and executive control (the caudate nucleus and putamen) are key to this demanding cognitive process.4 There is no single brain centre devoted exclusively to the control of interpreting, and the brain areas that control the process are generalist, contradicting the traditionalist view.
The team also conducted a longitudinal study to see whether interpreting training could lead to a change in brain structure or functioning.5 They found a reduced recruitment of the right caudate nucleus, which controls all sorts of skilled actions during simultaneous interpreting, as a result of training. This means that the impact of such training on the brain is not specifically linguistic, but involves a variety of domain-general executive functions. In other words, the effect of simultaneous interpreting training is to make the brain’s executive functioning in the caudate more efficient, and this can be measured in lower activation levels in the brain of a trained interpreter.
In collaboration with neuroscientists at Cambridge and Durham universities, my research focuses on the neural basis of translating individual sentences between English and Mandarin Chinese – languages which differ greatly in many respects. We used fMRI to investigate differences in the brain area of activation when 25 female mother-tongue Chinese trainee translators read aloud (as the baseline), carried out sight translation, and performed FT (Mandarin to English) and BT (English to Mandarin). We discovered that reading or translating English caused more activation in the visual word form area of the brain (left inferior temporal lobe) than reading or translating Chinese. This suggests that English, as the L2, complicated both tasks at the reading and recognition stage, causing participants to recruit more cognitive effort and increase activation in the visual word form area. In addition, sight translation proved to be more demanding than reading, and activated more regions in both the domain-specific and domain-general areas. This result is consistent with the findings of previous research.
Finally, we found that FT is more demanding and requires further cognitive resources, causing more activation in the cortex areas responsible for conceptual mediation and representation. This finding supports the RHM from the perspective of English-Chinese sentence translation.
Challenges and future prospects
Cognitive neuroscience, like a scalpel, has opened up an effective way to probe the mechanisms of the translating brain. The findings from research in this area can provide additional insights into the developmental context of translation, and reveal some links between translation competencies and the fundamental neuropsychological functions.
For instance, the finding that domain-general areas of the brain feature extensively in translation supports the idea that improving executive processing and attention/concentration ability may help to enhance lexical retrieval and information processing during extreme language use, such as simultaneous interpreting.
Just as Maria Tymoczko predicted, the cognitive neuroscience of translation will be one of the most important areas of future Translation Studies.6 However, due to its comparatively short history, the high demand for cutting-edge facilities, the unnatural research environment, and the expertise required to set up neuropsychological experiments and interpret the imaging data, such research remains limited.
The future of this kind of research will broadly lie in interdisciplinary projects, in which translation scholars are more proactive, working closely with neuroscientists and molecular biologists to overcome the difficulties of data collection and analysis in order to illuminate the ‘black box’ of human translation. Many more questions and hypotheses await examination. These include topics such as perception and memory in translation, translator training and brain plasticity, translation disorders in brain-lesioned bilinguals, and the interpreter advantage hypothesis.
1 L1 refers to one’s first/native language; L2 to one’s second/most proficient foreign language
2 Kroll, JF and Stewart, E (1994) ‘Category Interference in Translation and Picture Naming: Evidence for asymmetric connections between bilingual memory representations’. In Journal of Memory and Language. 33: 149-174
3 Lei, MM et al (2014) ‘Neural Basis of Language Switching in the Brain: fMRI evidence from Korean-Chinese early bilinguals’. In Brain and Language. 138: 12-18
4 Hervais-Adelman, A, Moser-Mercer, B, Michel, CM and Golestani, N (2015) ‘fMRI of Simultaneous Interpretation Reveals the Neural Basis of Extreme Language Control’. In Cerebral Cortex. 25(12): 4727-4739
5 Hervais-Adelman, A, Moser-Mercer, B and Golestani, N (2015) ‘Brain Functional Plasticity Associated with the Emergence of Expertise in Extreme Language Control’. In NeuroImage. 114: 264-274
6 Tymoczko, M (2012) ‘The Neuroscience of Translation’. In Target. 24(1): 83-102
Dr Binghan Zheng is Associate Professor of Chinese Translation at Durham University, and By-fellow at Churchill College, Cambridge. His research interests include cognitive neuroscience of translation and translation process research.