Chapter 2: Disciplinary context
2.2 Conference interpreting review
2.2.2 Research profile
2.2.2.1 Cognitive issues
2.2.2.1.6 Interpreting strategies
One of the most important aspects of the interpreting process is the strategic dimension which has enjoyed a great deal of interest by many IS researchers. The interpreting process from comprehension to production, seen by some professionals as a ‘crisis management’ situation (cf. Gile 1995d: 191), is a task that imposes a high processing load and certain cognitive constraints (such as simultaneity, SL input rate or text complexity) on the interpreter who, in order to cope with these constraints, must apply certain strategies or
tactics to achieve the task successfully (cf. Pöchhacker 2004: 132).
A strategy is
goal-oriented, so that the goal determines the amount and thoroughness of processing. It may be consciously used but may also have become automatic in so far as the processor will not have to make any cognitive decision (Kalina 1992: 253, emphasis in original).
Pöchhacker (2004: 132) distinguishes between overall task-related strategies, process- oriented strategies and communicating content strategies. As Pöchhacker (ibid) rightly argues, the line between the various classifications of different strategies is ‘hard to draw’. Pöchhacker’s way of presenting the different strategies, which is borrowed here, is used only as a convenient example among some others (Kirchhoff 1976a: 114-117; Van Dam 1989; Kohn and Kalina 1996, etc.) including a more detailed treatment of the issue (Gile 1995d: chapters 6-8) and is thus not intended to be viewed as a model to be followed11.
2.2.2.1.6.1 Overall task-related strategies
Overall task-related strategies have to do with the interpreting task in general. Thus, a distinction is made between off-line and on-line strategies (Pöchhacker 2004: 132f).
11 See also Lee (2006) for a recent proposal of SI strategies categorization and review of other categorizations
2.2.2.1.6.1.1 Off-line strategies
Off-line strategies have to do with creating and consulting sources for specific knowledge acquisition (glossaries, reference documents, encyclopaedias, etc.) and preparing for interpreting assignments by highlighting specific terms, names, numbers or segments when written material is available in the booth, archiving and indexing documents for future use, etc. (Gile 1995d: 146-151; Pöchhacker 2004: 133; 2.2.2.3.2 Documentation; 2.2.2.5.2.4.2 Training on SI with text; 7.5.4.3.3 Off-line procedures).
2.2.2.1.6.1.2 On-line strategies
Pöchhacker (ibid) classifies these according to interpreting modes. Good examples are note-taking in CCI (3.2.2 The strategy of note-taking) and anticipation in SI (see below). Others include using the help of the booth mate (‘passive interpreter’) in SI for noting down a technical term, name or number (Gile 1995d: 193) and asking the speaker for help in CCI if the interpreter misses something he/she thinks is important. Jones’s (1998: 53f) advice in this respect is to reflect the missing information on the notes so that the interpreter can locate it quickly on the notepad when asking the delegate and put the delegate in context to help him/her find the answer quickly. Other examples include consulting written material in the booth if any documentation is provided by conference organisers/speakers, noting down a number or name during SI to reduce memory or processing load, etc. (Gile ibid: 194).
2.2.2.1.6.2 Processing strategies
Processing strategies can be applied during a certain phase in the interpreting process to achieve a given task effectively under certain processing conditions. The following sections are dedicated to the discussion of the following most commonly cited processing strategies: waiting, stalling, segmentation and anticipation.
2.2.2.1.6.2.1 The strategy of waiting
When faced with temporarily vague strings of meaning, interpreters may wait for a few seconds until further constituents contribute to the disambiguation of the incoming message (Kirchhoff 1976a: 114; Gile 1995d: 192; Setton 1999: 50; Pöchhacker 2004: 133).
2.2.2.1.6.2.2 The strategy of stalling
padding expressions’ or hesitation fillers (Kirchhoff, 1976a: 116) when searching for or trying to remember a missing item, word or even a sentence structure (Gile 1995d: 197).
2.2.2.1.6.2.3 The segmentation strategy
According to Kirchhoff (1976a: 114f), proper segmentation of the SL text into function units is one of the prerequisites of information processing. A function unit is defined as the smallest possible SL decoding unit for which a one-to-one relationship may be established with a TL segment and is thus determined not only by the structure of the SL, but also by this relation of SL-TL equivalence. The lower limit for the size of the function unit is determined by the criterion of equivalence to the TL segment while the upper limit by the interpreter’s processing capacity (cf. also Wilss 1978: 346). The information value of a segment therefore must not go beyond the interpreter’s processing capacity in relation to the time available. The interpreter’s choice of timing thus lies between a minimum time lag and a maximum EVS of 10 seconds, and his/her optimum start should lie in a point that ensures the greatest amount of certainty and a minimum load on capacity. Comparisons between SL input and TL output show that experienced interpreters endeavour to adjust their segmentation strategies to the speaker’s rhythm (Kirchhoff ibid).
Kirchhoff’s argument confirms Barik (1969, quoted in Gerver 1976: 178) and Gerver’s (cf. 1976: 178ff) results that SL pauses assist interpreters in segmenting the SL message.
However, Goldman-Eisler (1972: 74) has experimentally arrived at a contradictory result in that in 90% of the cases interpreters have been more inclined to ignore the SL chunking and impose their own segmentation of the input either by starting to translate before the SL chunk has been finished or by delaying translation and storing more than one chunk. Moreover, Goldman-Eisler (ibid: 73f) suggests that whether interpreters begin to translate before the end of the SL chunk or delay translation and store more than one chunk, is by and large an issue of the nature of the particular language, citing the example of German which forces interpreters to store more than one chunk while waiting for the predicate which comes at the end of the proposition (cf. also Wilss 1978: 350). Goldman-Eisler’s finding that interpreters decide on a different structure from that of the input is supported by Kalina (1992: 254) who suggests that the reason for this is to avoid interference between the SL and TL syntactic systems, thereby achieving a better TL style.
For proponents of the theory of sense, the segmentation of input is based on ‘units of meaning’ (Lederer 1978: 330) or ‘subunits of sense’ (Seleskovitch and Lederer 1995: 125) which are a combination of a number of words in the STM and prior cognitive experiences and occur whenever the interpreter grasps what the speaker intends to say (ibid: 227-231).
2.2.2.1.6.2.4 The anticipation strategy12
Anticipation is described as ‘a fundamental feature of strategic discourse processing’ (Kohn and Kalina 1996: 124) and one of the most essential factors that explain simultaneity in SI (Chernov 1994: 140). Setton (1999: 52) defines13 anticipation as the production of ‘a sentence constituent – a main verb, for example – before any equivalent constituent has appeared in the SL input’.
Almost all researchers who have discussed anticipation (e.g. Kirchhoff 1976a: 115; Wilss 1978: 349; Lederer 1978: 330ff; Gile 1995d: 176ff) distinguish between two types of anticipation: linguistic/syntactic and extralinguistic anticipation. The former refers to prediction from knowledge of ‘syntactic and semantic regularities in the SL’ (Kirchhoff 1976a: 115); formulas, such as debate openers or gambits, and collocations (Wilss 1978: 348/350; Lederer 1978: 331); ‘predictor words’ such as connectives, function words, or subordinators (Setton 1999: 52); ‘selectional categories and case morphology’ or generally speaking from ‘transitional probabilities related to phonological, grammatical, stylistic and other language-related rules’ (Gile 1995b, quoted in Setton 1999: 52).
Extralinguistic anticipation, which is based on ‘sense expectation’ (Lederer 1978: 331), refers to prediction from extralinguistic knowledge or, in the words of Lederer (1990) and Seleskovitch and Lederer (1995: 23), ‘cognitive complements’ (verbal, situational and cognitive contexts and world/encyclopaedic knowledge).
According to Setton (1999: 52-57) and Pöchhacker (2004: 134), the study of anticipation by various scholars with different research orientations reflects a deeper controversy over the issue of language-specific factors in interpreting. While proponents of the theory of
12 See also 2.2.2.5.2.3.10 Anticipation exercises.
13 For a recent and more comprehensive review of research into anticipation and a survey of various
sense, or the ‘universalists’ (Setton 1999: 53), have generally maintained a position that minimises the importance of the role of (strategies for) syntactic asymmetries, followers of the tradition of the Leipzig School and scholars in the information processing paradigm, or the ‘bilateralists’ (ibid), have generally insisted on the importance of ‘language-pair- specific’ factors (Wilss 1978: 350; see also Riccardi 1996), especially in languages with ‘left-branching’ or ‘verb-last structures’ (Setton ibid).
The findings of Setton’s (1999) cognitive-pragmatic analysis, which has been based on authentic corpora involving two left-branching languages (Chinese and German) do not seem to support the bilateralists’ word-order-strategy or ‘strategies-for-structures’ approach (ibid: 126); Setton concludes that ‘marked syntactic structure alone does not obstruct SI’ (ibid: 282) and stresses the need for recognising ‘individually and culturally variant expressions of intentionality’ (ibid: 283; see also Setton 2003a: 55; 2005: 72-75). Setton’s conclusion confirms Seleskovitch’s argument on ‘the reflex nature of language’ that:
If … we look at language in communication, we note that, at least in our mother tongue, we are not aware of language as such; we are aware of our purpose, of what we intend to say when speaking, of the contents of the message we receive, but rarely do we choose our word deliberately or remember specific words uttered by a speaker (1977: 31).
2.2.2.1.6.3 Communicating content strategies
According to Pöchhacker (2004: 134), communicating content strategies, which mainly concern ‘reductive’ or ‘adaptive’ strategies, have been proposed as forms of coping tactics that can be applied as a response for processing constraints.
2.2.2.1.6.3.1 Reductive strategies
In response to either one, some or all of such extreme input variables as high SL presentation rate (cf. e.g. Gerver 1969), ‘excessive time lag and overload’ (Kirchhoff 1976a: 116), linguistic complexity or syntactic and semantic difficulty (Tommola and Helevä 1998: 179), the interpreter can apply one or more of several reductive strategies, which refer more or less to the same operation but are given various denominations by different researchers. Chernov (1994: 145f) suggests lexical or syntactic compression. He argues that redundancy, which is usually present in the thematic or topical part of utterances, allows for compressing speech by reducing the number of syllables, words or
semantic elements or simplifying syntactic structures. Thus, Chernov (1994: 146; 1992: 156) argues that the interpreter can compress a speaker’s announcement such as:
I now give the floor to the distinguished delegate of the United Republic of Tanzania!
to only
Tanzania!
by deleting the theme and conveying only the rheme of the utterance.
Using data from a 588-word text corpus from Spanish-Danish CCI involving ten MA translation and interpreting students and two professional interpreters, Dam (1993) proposes ‘text condensing’ through substitution and omission. Dam argues that this is ‘a perfectly admissible and even good interpreting strategy’ (1993: 299) as long as the end product allows for getting the message across to the TL audience (ibid).
Sunnari (1995) has concluded that macroprocessing of the SL message (microstructure), which can be achieved by Van Dijk’s (1977) macrorules (selection, deletion, generalization and construction), is one of the most important strategies for fluent and high-quality interpreting between languages with different syntactic structures. Sunnari’s data have been generated from three recorded presentations and their translations by four professionals in a real conference and have rendered three different situations: ideal, counter-ideal and pseudo-ideal. Part of the pseudo-ideal original has been interpreted in a laboratory by four trainees whose performance has been compared with the professionals’. Results have not revealed any loss of relevant information in all cases because professionals have known ‘when and how to apply macroprocessing’ (Sunnari 1995: 118). Besides, it has been revealed that macroprocessing has been an effective strategy without which interpreting would have been impossible in the counter- and pseudo-ideal situations and that it has helped make the output smoother and more natural in the ideal situation. It seems, however, that the trainees have not mastered the strategy and thus their output has been incoherent.
The results of an empirical analysis of 357 speeches and their interpretations in three languages (English, Italian and Spanish) in all possible combinations and directions from
the European Parliament Interpreting Corpus confirm the above arguments on reductive strategies. Russo et al. (2006: 248) have found, inter alia, that there has been a ‘general tendency to text compression in interpreted speeches’ as revealed by the lower number of words in the interpreted speeches compared to their corresponding originals.
2.2.2.1.6.3.2 Adaptive strategies
Kohn and Kalina (1996: 127f) suggest adaptation strategies in the production phase in accordance with the TL conventions of expression, in particular cultural adaptations which interpreters may apply to bridge gaps brought about by differences between the SL and TL cultures or when the SL discourse contains facts or events known to the SL audience but not necessarily known to the TL audience. Thus, the interpreter might have to go beyond linguistic mediation. Inevitably, this has implications for the interpreter’s role as to whether or not he/she can, should or must play the intercultural-mediator role, and if yes, through what means can this be achieved? This is the main subject of this thesis and is fully investigated within the context of CCI (Chapters 5-8).