Testing the Effort Models tightrope hypothesis insimultaneous interpr - PDF document

1 Testing the Effort ModelsÕ tightrope hyp
Testing the Effort ModelsÕ tightrope hypothesis insimultaneous interpreting - A contribution mode, errors and omissions (e/oÕs) were found to affect different source-speech seg-ments, and a large proportion among them were only made by a small proportion of thesubjects. In a repeat performance, there were some new e/oÕs in the second versionversion. These findings strengthen the Effort ModelsÕ Òtightrope hypothesisÓ that manye/oÕs are due not to the intrinsic difficulty of the corresponding source-speech segments,1.The nature of the Effort Modelsoperation of simultaneous interpreting. More recently, Setton (1997),Paradis (1994) and Mizuno (1994, 1995) have developed their ownments in cognitive psychology, neurolinguistics and linguistics. On

2 thepractical side, however, over the pas
thepractical side, however, over the past two decades, they have not beenthe mental operations involved (see for instance Lambert 1995, Moser- *Daniel GileUniversitŽ Lumire Lyon 2Home address: In the early eighties, a set of models were developed in a differentbe easily attributed to deficient linguistic abilities, insufficient extra-Effort Models (see for example Gile 1995, 1997) poolEffortsL - the Listening and analysis EffortP - the Production Effort (speech production in simultaneous, and noteM - the short-term Memory Effort essentially dealing with memoryreformulated in the target speech or disappears from memory.The Effort Models (EM) are models of operational constraints, nottal structure and information-processing flow, as is the cas

3 e of the otherdation and component and/o
e of the otherdation and component and/or flow additions and corrections. Moreover,models cannot make operational predictions. However, though theyassumption of a strong correlation between task difficulty and task-The most fundamental architectural assumption in the Effort Modelsis that in spite of the sharing of some cognitive resources, and in partic- 154 ular long-term memory, there are enough unshared components in in-terpreting to justify the distinction between the three Efforts, one re-comprehension is in one language and production in another, and injustifies separation of these two Efforts. The definition of the MemoryEffort as a distinct component is less evident, since both sensoryproduction (in the latter, for self-monitoring). T

4 hus, it could be arguedproduction phase
hus, it could be arguedproduction phase would be sufficient and more representative of actuala.The necessary co-existence in short-term memory (including sen-target-speech elements in simultaneous interpreting, and of source-target speech, inhibition and activation, links with the mental lexicon inone language or another, etc.), which are not usual in the non-inter-s memory.b.In strategic terms, interpreters make specific decisions on their EVS(Ear-Voice Span, or how much they lag behind the speaker) on the basis and Fabbro1994, Osaka 1994, Padilla 1995, Gran and Bellini 1996, Chincotta and a.Each of the three Efforts has non-automatic components. Therefore,matic only if it involved automatic word-for-word replacement, whichis clearly not the

5 case. As to the short-term memory effor
case. As to the short-term memory effort, it is non-automatic insofar as it involves storing and retrieving ever-changingmatic nature of the three Efforts, see Gile 1995). This first operationalb.The three Efforts are at least partly competitive, meaning that even can be repre-sented in the following way, with the total processing capacity con-sumption TotC associated with interpreting at any time represented as a (not in the pure arithmetic sense) of consumption for L, con-sumption for M and consumption for P, with further consumption for (C) between the Efforts, that is, the management ofcapacity allocation between the Efforts:(1)TotC = C(L) + C(M) + C(P) + C(C)(2)C(i) (3)TotC (4)TotC C(i) + C(j) i,j = L, M, P and i different from j- in

6 equality (2) means that each of the thre
equality (2) means that each of the three Efforts requires some pro-equal to that of any single Effort equal to that of any two Efforts performed in conjunction (in otherwords, adding a third Effort means adding further capacity consump-tioners, is explicit in many anecdotal accounts of difficulties encounter-to cognitive scientists at various interdisciplinary meetings. However, itc.The idea that most of the time, interpreters work near saturation2.Previous theorizing and testingOn the basis of the Effort Models, some further theoretization waspossible. Firstly, the existence of was hypothesized,CliffConversely, low density segments in a speech can lower cognitive pres- of up to more than 10syllables in length. In terms of the Effort Models

7 , such a pressure dropwhen interpreting
, such a pressure dropwhen interpreting from Japanese. Unfortunately, no on-line testing ofthis potential effect could be done at this stage. diverted from one Effort to another where was neces-sary, thus one speech segment but jeopardizing an ulterior seg- (Gile 1995). In a simple interpretation tasktified in the source speech and looked at the target speech rendition ofprobably involved. However, the design of the experiment was toothe conclusions explicitly.The Effort Models also predict higher attentional requirements whenworking from syntactically different languages. Starting with this as-In a recent doctoral dissertation, Lamberger-Felber (1998) also test-ed a number of hypotheses which she derived from the Effort Modelsregarding dif

8 ferent types of errors and omissions (in
ferent types of errors and omissions (in numbers and prop-From a slightly different angle, the EM for consecutive predicts adisruptive effect of note-taking in untrained interpreters for the follow- -extra processing capacity is involved in deciding what to write and-extra processing capacity is involved in controlling the writing ope--writing generally takes much longer than uttering the same speech3.The tightrope hypothesis1, one that is more holistically associated with the Effort Models, name-s total available capacity, so that any increase in process-attentional deficit (in one of the Efforts) and consequent deterioration is crucial in ex-ed in interpreting even when no particular technical or other difficultieswhen significant difficult

9 ies came up in the source speech.s) affe
ies came up in the source speech.s) affecting segments that present no evident intrin-sic difficulty. If there are, it is likely that they can be explained in termsthe segments affected in the sample (at the level of words or proposi-interpreting difficulty of the relevant segments (too spe- 159 cialized, poorly pronounced, delivered too rapidly, too difficult to ren-der in the target language, etc.), even if available descriptive tools arenot sensitive enough to identify such difficulty beforehand. If howeveronly a few subjects fail to render them correctly in the target language,esis that processing capacity deficits are involved. An analysis of inter-ject is asked to interpret the same speech twice in a row. Having becomeof interpreting pe

10 rformance from the first to the second t
rformance from the first to the second target-languageare involved. It is difficult to find another explanation: the fact that thesegments affected in the second target-language version were interpret-and knowhow to reexpress them in the target language.4.Methodrence given by George Fisher when his position as KodakExecutive Officer was announced. It was interpreted by myself from aby George Fisher to a question put to him by a journalist. It is 1 minutesimultaneous interpreting working day, and always after they had time 160 with one or two turns of interpreting in the booth within into French (argentThe experiment was carried out in interpretation booths, with the sourcephones and the target speech being recorded on a portable cassette re-

11 corder. When they finished interpreting,
corder. When they finished interpreting, subjects were asked to startinterpreting again. Ten subjects were recruited over three distinct inter- bilinguals. All had regular working expe-international organizations, in particular OECD and UNESCO. TheyTarget speeches were transcribed, and transcripts were scanned forof high inter-rater variability in the perception of what is and what is notsentation of the speech (as demonstrated in Gile 1999). To avoid thesedissenting opinions. Moreover, the likelihood that the e/o - e/o ncorrected in the second version of the target speech by at least some of (mistaking text manipulations considered acceptable by thea.How many subjects in the sample made an e/o for each affectedb.What e/os were corrected in

12 the second version of the target speech.
the second version of the target speech.(However, due to a technical problem, no second version could be rec-5.Results5.1.List of e/om sure my...). Type ofe/o: error. Corrected in the subject). Error. Corrected - ). Error. Corrected in the second version t know...: omission. Uncorrected in the second version.for the foreseeable futureas far as capture goes captif11.highest resolution cause mine sure does with mere really excitingninety percentten percent - no indication in the target speech that - no indication in the tar- 5.2.New e/os found in the second version of the targetje connais suffisamment les scientifiques et les ingpy... - omissionmais je connais suffisamment bien...but I know sufficiently well in perhaps ways that are totally dif

13 ferentde les diffuser de mani-fait difft
ferentde les diffuser de mani-fait diffto disseminate them in a way that may be totally different...-fait diffto distribute them in a totally different way - the idea expressed in je ne connais suffisamment bienfisamment bienÒnotÓ is missing, but the sentence is negativein perhaps ways that are totally differentre totalement diffperhaps in a totally different waytement diffcertainly be completely different - the idea expressed in is miss-b.omission b.omissionre really excitingb.omission5.3.Quantitative analysisTable 1:Errors and omissions in the first and second renditions.(0:correct; 1: error or omission)Table 1 summarizes the quantitative aspects of the analysis. It turns outtarget speech, 8 e/os (47%) were made by one or two interpreters

14 only. SubjectABCDEFGHIJ Source-speech se
only. SubjectABCDEFGHIJ Source-speech segmentrendition 1I'm sure my000001-000001 2I don't even know these people yet00-10-111-01-001-001-15 3scientists and engineers01-00000001-00-12 4well enough0-10-1001-11-100002 5but since I don't know what the products are1-11-00101-11-11-01-107 6I can speak loosely01-01-000000002 7the imaging side of Kodak00011-00001-01-04 8let's concentrate on that0-11-01-110000-100-13 9for the foreseeable future001-00001-10002 10as far as capture goes1-0000000001-02 11highest resolution1-11-01-1101-01-11-1007 12cause mine sure does with me000001-01-101-003 13the last two nights01-1011-01-01-11-0006 14and they're really exciting01-101001-101-00-14 15ninety percent, ten percent0001000001-02 16my ideas001000001-002 17that

15 will be killers1-01-1011-11-01-101-01-1
will be killers1-01-1011-11-01-101-01-18 perhaps (2nd version only)00-100000-1000 total e/o's in 1st and 2nd rendition4-48-65-395-28-27-84-27-15-5 number of "new" e/o's (in 2nd rendition only)231001103 167 Moreover, among them are interpreters who only made a total of 5 e/o in the sample. It may therefore be conservatively assumeds, no intrinsic difficulty of the affected source-specific difficulty is involved will appear clearly to readers from the listment in the second version of the target texts was confirmed: the num-s decreased for 5 subjects (B,E, F,H, I), remained the same foranalysis (as mentioned earlier, only parts of the recording were availa-their target text whereas the relevant source-text segment had been bet-6.Discussion and

16 conclusionused here in order to reduce
conclusionused here in order to reduce to the largest possible extent the number of means that other phenomena that could have been usedto measure cognitive load were not exploited. In particular, no attemptoutput quality, or at changes in the prosody or the quality of the inter-be designed, and reliability could have become a problem. Fortunately,the tool proved to be sufficient, illustrating the idea that at the beginningpothesis and thus give some support to the Effort Models as a concep- cognitive-constraints-based limita-tions. However, the usefulness of the EM as an operational tool and 168 be found for the Effort Models. Meanwhile, I can only agree that theinterpreting(as opposed to cognitive models andBibliographical referencesChinco

17 tta, Dino & Geoffrey Underwood (1998). S
tta, Dino & Geoffrey Underwood (1998). Simultaneous Interpreters and theeffect of concurrent articulation on immediate memory. In Interpreting, Valeria & Franco Fabbro (1994). Verbal memory during simultaneous interpre-tation: Effects of phonological interference. In Word Order in Chinese-English Simultaneous Interpretation:. Unpublished M.A. thesis, Fu Jen University.Gerver, David (1976). Empirical studies of simultaneous interpretation: a review and amodel. In R.W. Brislin (ed.) (1976). Translation: Applications and ResearchYork: Gardner. 165-207.Frauenfelder, Uli & Herbert Schriefers (1997). A psycholinguistic perspective on Sim-InterpretingFukuii, Haruhiro & Tasuke Asano (1961). preter. Tokyo. Kenkyusha. The InterpreterBasic Concepts and

18 Models for Interpreter and Translator Tr
Models for Interpreter and Translator Train-Cognitive Processes in Translation and InterpretingGile, Daniel (1999). Variability in the perception of fidelity in simultaneous interpre-Gran, Laura & Beatrice Bellini (1996). Short-Term memory and Simultaneous Interpre-tation: An Experimental Study on Verbatim Recall. In The Interpreter7. 103-112.Granacher, Martin (1996). Gile. Versuch einer Evaluierung auf der Grundlage einer Fallstudiet Heidelberg.Lamberger-Felber, Heike (1998). Der Einfluss kontextueller Faktoren auf das Simul-Lambert, Sylvie (1995). Foreword. In Lambert, Sylvie and Barbaras Moser-MercerBridging the Gap. Empirical research in simultaneous interpretationMassaro, Dominic W. & Miriam Shlesinger (1997). Information processing and

19 a com-InterpretingInterpreting ResearchS
a com-InterpretingInterpreting ResearchStudy on the Convergence of Translation Patterns. In Interpreting ResearchMoser, Barbara (1978). Simultaneous Interpretation: a Hypothetical Model and itsPractical Application. In Gerver, D. & H. Wallace Sinaiko (eds) (1978). Interpretation and CommunicationFactors. New York and London: Plenum Press. 353-368.Moser-Mercer, Barbara (1997). Methodological issues in interpreting research: An in-Interpreting2(1-2). 1-11. ron, P. and H. Nanpon (1965). Recherches sur la traduction simultanshinri gakkai dai 58 kai taikai ronbunshuu (Proceedings of the 58th Convention ofProcesos de memoria y atencin en la interpretaci. Tesis doctoral, Universidad de Granada, Departamento de FilologParadis, Michel (1994). Toward

20 a neurolinguistic theory of simultaneous
a neurolinguistic theory of simultaneous translation:Comprehension difficulties in simultaneous interpretationfrom non-standard EnglishSchjoldager, Anne (1996). Simultaneous Interpreting: Empirical Investigation intoTarget-text Source-text RelationsA Pragmatic Theory of Simultaneous Interpretation Appendix - Source speech. answer only. You suggested that through Kodak you can manipulate technology and m sure my... I donhalide capture media is probably the most cost-effective, highest reso-could ask for. So to me, you want to put that in the context of being avery effective way of getting the information to begin with, then youare totally different than people envision today, than Ination run off with you, cause mine sure does with me. I laid