Preliminaryresults
3. stage one: the survey 3.1 P urPose of the survey
Theaimofthesurveyistobroadenourknowledgeofsimultaneous interpreting with text
(si+t)byexaminingattitudesandstrategiesofinterpreterswitha languagecombinationthat
includesHungarian.
Theresearchquestionandthecorrespondinghypothesisarethefollowing:doesthewrit-ten text help, rather than hinder interpreters’ performance in simultaneous interpreting?
Hypothesis:Whentheappropriatestrategiesareused,visualinformationcomplementsaudi-toryinformationandhelpsconferenceinterpreterstorenderthemessagemoreaccuratelyand
ina moresuitableform.
Borbálarohonyi
Thesurveyconductedinthefirststageoftheresearchreliedona studycarriedoutby
a groupoffourresearchers:Cammoun,davies,Ivanovandnaimushin(2009).Thepresent
projectadaptedtheirquestionnairebeingspeciallydesignedtorevealinterpreters’practices
andattitudestowardssi+t totheHungarianmarket.Fora detaileddescriptionofthemethod
andthequestionnaireusedinthesurvey,seeannexe1.
3.2 resultsofthesurvey – discussion
Surprisingly,almosthalfoftherespondents(15interpreters)tookpartinsomekindof si+ttrain-ing.13amongthemhadsi+texercisesduringtheirinterpretertraining.Thisresultconfirmsthe
earlierfindingsofCammounetal.(2009)andprovesthattraininginsi+t isnotmissingfrom
interpretertraining.Thequestionnairedidnotlookintothenatureandscopeofthetraining.In
termsofbasicstrategies,thequestionnaireinquiredfirstwhethertheinterpreterchoosestouse
thewrittencopyofthespeechornot.Thevastmajorityofrespondents(91%)alwaysusethewrit-tentextwhentheyreceiveitwellinadvance(scenario1);a majority(79%)alwaysuseitifthey
receivethewrittencopyatleast15minutesbeforethestartofinterpreting(scenario2).Inthecase
theyaregiventhewrittentextonly5minutesbeforethestart(scenario3),lessthanhalfalways
useit(42%),alsolessthanhalfsometimesuseit(45%).The‘never’optionforthesequestionsis
worthmentioning:nobodydiscardsthetextinscenario1,itcamesomewhatasa surprisethat
onerespondent(3%)discardsthewrittentextinscenario2,and3respondents(9%)prefernot
tousethetextwhentheyreceiveitlessthan5minutesinadvance(scenario3).
Inordertodeepentheanalysis,Pearsonchi2(4)correlationanalysiswascarriedoutto
showtherelationofsi+ttrainingandtheuseofthewrittentextforthethreescenarios.si+t
trainingdoesnotcorrelatesignificantlywiththedecisiontousethemanuscriptinthefirst
andsecondcase(Pr=0.927;Pr=0.130).However,Table 1showsthatsignificantcorrelationwas
foundinthethirdscenariobetweentraininganduseoftext(Pr=0.063).Thosewhoreceived
si+ttrainingwerelessinclinedtousethetextintheconstrainedsituationwheretheyhave
lessthanfiveminutestopreparethetext,comparedwiththosewhodidnothaveanysuch
training.a possibleexplanationforthisunexpectedresultcouldbethatinterpretersprefer
todiscardthetextincasetheycannotprepareit“properly”.However,thishypothesisneeds
furtherstudytobeconfirmed.
Table 1:relationofSI+Ttraininganduseofthewrittentextinscenario3
do you use the text if you receive it less than 5 minutes before the start of interpreting?
Training Yes,always Yes,sometimes never Total
no 10 7 0 17
Yes 4 8 3 15
Total 14 15 3 32
Pearson
chi2(4) 5.5347Pr=0.063
Insituationswheretheyhavetimetopreparethewrittentext(scenarios1and2combined),
respondentsfocusonnamesandtitlesinthefirstplace,followedbyterms,numbers,sentence
structureandlinks,and,finally,references.Theeffectofsi+ttrainingcanbeshown:those
whoreceivedtrainingconsiderthepreparationofnames,titles,numbers,referencesandterms
(fouroutoffivecategories)tobemoreimportantthanthosewhodidnothavetraining.
Figure 1showsthatwhenrespondentshavelessthanfiveminutes,mostprefertogetthe
generalideaofthetextbyskimmingthroughit,followedcloselyinimportancebycircling/
underlying/highlightinga fewimportantelements,includingnamesandnumbers,andfi-nally,readingthefirstandlastparagraph.Thisisnotanunexpectedresult,yettheorderof
preferenceofthetwofirstoptionsisreversedcomparedwiththefindingsofCammounetal.
(2009).
Figure 1:Scenario3–Lessthan5minutestopreparethewrittentext
The literature review elaborated the four benefits and five constraints conceptualised by
Cammounetal.(2009).eightofthesewereusedinthepresentsurvey,anda separateques-tionreferredtodualinput.
Interpretersrankedtheconceptsintermsoftheirimportanceinsi+t(theorderofprior-itywas1to8where1wasthemostimportantand8theleastimportant).Firstofall,they
considerthewrittentexttobehelpfulforanincreasedaccuracy ofdatain the output(2.73).
Secondly,inthecasewherethedeliveryspeedisaccelerated(3.16),thirdlyasithelpsthem
anticipate(3.85).Thesewerefollowedbythelackofredundancy(3.89),copingwitha heavy accent(4.03),tohavea broaderknowledgeofthetopicofthemeeting(4.43),interference
fromthesourcetext(5.81),andfinally,whenthedeliveryismonotonous(6.12).relativetothe
conceptofinterferencethesamedoubthasarisenasintheresearchperformedbyCammoun
etal.(2009).Interferencerefersto“thedirectimpactofthesourcetextonthetargettext”
(g. Láng2002:194);or“speakersapplyingknowledgefromthefirst/sourcelanguageto
a second/target language” (Cammoun et al. 2009: 60), and the negative effect might be
strongerandmoreapparentwhenthevisualsignsarepresentinfrontofthem(gile2009).
respondentsmighthaveconfusedinterferencewiththedifficultyofdualinput.Therefore,
thisaspectwasnotfurtheranalysed.
Thecomparison,withthehelpofa meantest(Wald)oftheaveragerankingsoftheeight
constraintsandbenefitswithinthesi+ttrainingsubgroupandothersubgroupsperyears
ofprofessionalexperience,provedtobefruitful.amongtheeightaspectsonestoodout:
Borbálarohonyi
therewereonlysignificantdifferencesbetweenthesubgroupsintherankingof“havingthe
textwhenthespeakerhasa heavy accent”.Itsrankingwassignificantlyhigher(witha 95%
confidencelevel)amongthosewhohavenotreceivedtrainingandthosewhohavemorethan
tenyearsofprofessionalexperience.Inviewofthesesurprisingresults,accentseemedtobe
a conceptthatrequiresfurtherstudy.
Figure 2:relyingontheauditory/visualinputinSI+T
Figure 2showsparticipants’preferencesintermsofthesighttranslationcomponentofSI+T.
Theratioofthosewhomainly‘followtheirears’wasslightlyhigherthanthosewhofavour
thedominanceofsight translation(10versus8respondents,30%versus24%).Butbasedon
theresults,theoutstandingratiooftheoption“Itdepends”(17respondents,51%)isworth
mentioning.Itindicatesthatmanyinterpretersdonothavea one-and-onlystrategy,orgeneral
preferenceofa strategythatworksfortheminsi+tmode.Theirstrategychangesdepending
–presumably–onhowmuchtimetheyhavetopreparethetext,whattypeoftexttheyhave
todealwithandhoweasilytheycanunderstandthespeaker.Furtheranalysiswascarriedout
inthesecondstageoftheresearchtofindoutwhattheirstrategydependson.
Foranextendedanalysisoftheresultsofthequestionnaireandmoredetaileddescription
ofwhatdifferenceswerefoundfromfindingsofCammounetal.(2009),seerohonyi(2016).