Methodological and other challenges in the use of the concept of effort in TIS
Scientific research is effective when it discovers/uncovers facts, when it produces theories that help explain them and predict what is yet uncovered, and when it develops concepts and methods that help uncover more facts and develop and test theories. The feeling of eftortfulness in translation and interpreting is part of reality.The concept of cognitive effort has demonstrated its usefulness in developing theories that explain phenomena observed in translation and interpreting. How predictive these theories are depends on the level of accuracy and detail sought. For instance, the Effort models and the associated tightrope hypothesis have been shown to successfully predict holistic phenomena such as increases in the rate of errors, omissions and infelicities in interpreting under certain working conditions (in particular high delivery speed) and in the presence of well-identified problem triggers. However, the precision of their predictions is limited because of their very nature as a holistic conceptual framework. The tightrope hypothesis did not quantify the interpreters’ “closeness” to saturation. It only indicated functionally that they tended to work close enough to saturation to become vulnerable to unexpected hikes in attention requirements and to attention management errors. More specific cognitive theories could do better, provided researchers find a way to offset the high variability arising from the large number of highly influential parameters, which are determinants of performance and are difficult to control in ecologically valid experiments.
Measurements of cognitive effort per se are also of potential use in uncovering interesting correlations; for example, if it could be shown, beyond general principles as postulated by existing theories, that certain tactical behaviour patterns and certain types ofTranslation quality deteriorations tend to occur at specific thresholds of cognitive load or cognitive effort. Many indicators provide relative data but not interval scales against which particular Translational phenomena can be identified. What is the level of effort at which errors of specific types occur, and what is the level of effort at which Translators give up trying hard and adopt less effortful but also less efficient strategies and tactics as regards the quality of the final product? Answers may come from triangulation studies measuring Translation behaviour and cognitive effort while eliciting retrospective reports (as in Gumul, 2018, in which many retrospective comments on 240 interpretations by students turned out to point to processes described by the Effort models).
Yet another limitation of many quantitative cognitive effort indicators such as pupillometry, gaze duration and hesitation pauses in their use to test ideas and theories about interpreting cognition (the problem may be less conspicuous in translation, where source-text reading is easier to separate from target-text writing) is that while they may be able to detect variations in cognitive effort intensity, they are not specific enough to point to the particular sub-processes or process components that caused them. How does one know whether increased cognitive effort is caused by a comprehension difficulty, a production difficulty or by the retrieval of information from memory, and what is the triggering speech segment? When considering ear-voice span, or earpen span in the case of CI, as in Chen (2017b), lag time can be taken as an indicator of cognitive effort possibly arising from processing difficulty associated with a previously heard source-speech segment. However, it can also indicate a tactical pause, which is not linked to any such difficulty, for instance when the interpreter reviews previous notes to make sure s/he will be able to read them when reformulating the speech.
Brain imaging techniques may provide one solution to the problem, if they show that distinct processes tend to mobilize specific areas of the brain or generate specific activity. However, in many cases, a far simpler solution can also help, again in the form of retrospective reports. Cued retrospection, as already used with keylogging techniques a long time ago, inaccurate and incomplete as it is, could well help interpret more specifically causes of increased cognitive effort as detected by objective indicators.
In human Translation, effort cannot be reduced to one parameter in a productivity or profitability equation. On the professional side and on the didactic side, it is a determinant of quality,
and as such, an ethical obligation—up to a certain point. It is also, for at least a decade or so in every Translator s early career, an investment in future expertise (Ericsson, 1996).
On the research side, it is also a useful concept for the investigation of Translation behaviour, including tactics, strategies and sub-optimal performance (e.g. errors, omissions and infelicities) and also socio-affective behaviour determinants.
It is particularly useful in research into Translation cognition, especially with the help of theories from cognitive sciences and with relevant technology that provides physiological evidence of mental activity. Pupillometry and brain imaging techniques, which are becoming less and less invasive, hold much promise, but when aiming for the best results, it is important to take on board the specific technical and social environment ofTranslation and to consider triangulating these “objective” techniques with more qualitative techniques, in particular retrospection, which can be of much help when seeking to interpret quantitative data correctly.