Cognitive load

Template: KeyGuides, Font: Bembo Date: 03/06/2015; 3B2 version: 10.0.1465/W Unicode (Dec 22 2011) (APS_OT) Dir: //integrafs1/KCG/2-Pagination/TandF/REIN_RAPS/ApplicationFiles/9780415634328_text.3d

COGNITIVE LOAD

largely based on observational evidence. But while some see this phenomenon as a natural part of language comprehension (Lederer 1981; Seleskovitch 1976/2001), others see in it a strategy specifically deployed by simultaneous interpreters to overcome task-inherent constraints such as syntactic differences between source and target languages. Behavioral evidence collected for SI from languages with verb-final structures (e.g. German) into languages with verb-initial structures (e.g. English) shows that sentence-final verbs are anticipated correctly in 50% of cases (Jörg 1997; Seeber 2001). Interestingly, Van Besien’s (1999) re-analysis of Lederer’s (1981) data reveals an even higher incidence of (verb-final to verb-initial) verb anticipation in SI, suggesting that it might indeed be a language-specific phenomenon. As for the cues enabling anticipation, the current consensus is that both top-down and bottomup processes guide this kind of processing. Recent psychophysiological evidence suggests that top-down processes do not offset the load generated by bottom-up (e.g. syntactic) constraints (Seeber & Kerzel 2012); rather, it is contextual constraints (i.e. semantic and syntactic cues in context) and not transitional probability (i.e. the statistical likelihood of a verb occurring with the rest of the sentence) that guide interpreters in verb anticipation (Hodzik 2013). Among the many aspects of BILINGUALISM that have a bearing on interpreting, the issue of DIRECTIONALITY, that is, whether conference interpreters (should) work into their ‘A’ language (mother tongue) or from their A into a so-called B language (second/acquired language), was once the dividing line between two opposing ideological camps (Gile 2005b). Informed mainly by theoretical argument, the debate still shows considerable discord. Greater strides towards understanding the implications of directionality in interpreting have become possible thanks to neurophysiological insights into the cognitive processes involved during SI: increased activation in the right hemisphere is found when interpreters work into a foreign language (Petsche et al. 1993). Additional evidence (Rinne et al. 2000) shows that SI into a foreign language (and in certain language combinations) engenders more extensive frontotemporal activation increases, and is thus conceivably more demanding of neural resources than SI into the mother tongue. KILIAN G. SEEBER COGNITIVE LOAD ↑ COGNITIVE APPROACHES → WORKING MEMORY ↓ INPUT VARIABLES

Cognitive load is a multidimensional construct used in the field of psychology. It is based on the assumption that the human processor, the brain, is limited in the number of operations it can carry out and in the amount of information it can keep available for processing at a given time (Miller 1956). Cognitive load can refer to the processing load imposed on the performer by a particular task (Paas & van Merriënboer 1993), or the perceived mental effort the performer invests in a task (Yin et al. 2008). This load can be further distinguished as follows: ‘instantaneous load’ is the load at a given moment in time, which fluctuates throughout the performance of a task; ‘peak load’ reflects the maximum value of instantaneous load reached; ‘accumulated load’ is the total amount of load experienced; and ‘average load’ is the mean intensity of load (Xie & Salvendy 2000). Whether the cognitive resources that are depleted by this load, or that fuel this effort, are unitary or multiple is a matter of debate, as is the very need for a concept that may simply reflect processing structures. 60

Template: KeyGuides, Font: Bembo Date: 03/06/2015; 3B2 version: 10.0.1465/W Unicode (Dec 22 2011) (APS_OT) Dir: //integrafs1/KCG/2-Pagination/TandF/REIN_RAPS/ApplicationFiles/9780415634328_text.3d

COHERENCE

The explanatory potential of this concept for the interpreting process, which combines language COMPREHENSION and language production in real time and is generally accepted as one of the most difficult linguistic skills, is manifest. The notion of cognitive load was first introduced to the field of interpreting studies by Gile (1985), in an attempt to explain information loss observed in professional interpreters. His EFFORT MODELS, which describe interpreting in terms of listening and analysis, production and MEMORY, were originally devised as a pedagogical tool but have long evolved beyond their initial purpose and are among the most widely used theoretical frameworks to explain cognitive processes in interpreting. However, the conceptual framework of the Effort Models has yet to find unequivocal empirical support (Seeber 2011). Similarly, the notion that interpreters work at a level close to cognitive saturation most of the time, reflected in Gile’s (1999a) ‘tightrope hypothesis’, is yet to be empirically corroborated. Indeed, one of the major challenges in applying the construct of cognitive load to research in interpreting has been the difficulty of measuring the phenomenon. For example, performance measures, particularly error scores during a secondary task, are often used to quantify cognitive load in psychology (e.g. Power 1986), but are ill suited to assess cognitive load in interpreting: as both CONSECUTIVE INTERPRETING and SIMULTANEOUS INTERPRETING already involve multiple concurrent tasks (comprehension and note-taking in consecutive interpreting; comprehension and production in simultaneous interpreting), combining them with further tasks is impracticable. Moreover, individual differences in TIME LAG and regulatory focus make it impossible to establish a clear cause-and-effect relationship between cognitive load, performance speed and performance ACCURACY (Seeber 2013). On the other hand, psychophysiological measures, like cognitive pupillometry using EYE TRACKING, albeit more costly and arguably more complex, appear to be better suited for the measurement of cognitive load during the interpreting process (Seeber & Kerzel 2012) and hold considerable potential for future research. KILIAN G. SEEBER COHERENCE ↑ LINGUISTIC/PRAGMATIC APPROACHES → COHESION, → QUALITY CRITERIA

In simple terms, ‘to cohere’ means ‘to hang together’; it stems from the Latin cohaerere, literally ‘stick’ (haerere) ‘together’ (co-). The word ‘text’ has its origins in a similar metaphor, the Latin textus, literally ‘something woven’. For linguists, what distinguishes a passage of speech or writing as a text rather than a random sequence of utterances or sentences is precisely its coherence as a unified whole (Halliday & Hasan 1976). Coherence is generally seen as having two dimensions. The first is semantic and concerns the relationship of a text with the ‘real’ world. From this perspective, a text is a unified discourse that relates consistently to a picture of the extra-linguistic universe. In this respect, some researchers stress the role of mental schemata – expectations about common situations and events – in achieving coherence in both the production and COMPREHENSION of texts (e.g. Gernsbacher & Givón 1995). Thus, coherence can also be seen as the result of the interaction between a text and the text users, for whom the concepts and relations underlying the text must be ‘mutually accessible and relevant’ (Beaugrande & Dressler 1981). In other words, understanding a text requires INFERENCING based on an individual’s knowledge and expectations. The second dimension of coherence concerns the lexical and grammatical features of the surface text that signal the underlying concepts and relations. This is usually referred to as 61