From 'Is' to 'Ought'-Laws, Norms

From'Is'to'Ought':Laws,NormsandStrategiesinTranslationStudiesAndrewChestermanUniversityofHelsinkiAbstract:Translationstudiesneedtocaterforbothdescriptionandevaluation.Thiscanbeachievedviathestudyoftranslationnorms.Thenormsgoverningtranslationare:(a)professionalnormsconcerningthetranslationprocess(=normsofaccountability,communicationandtarget-sourcerelation);and(b)expectancynormsconcerningtheformofthetranslationproduct,basedontheexpectationsoftheprospectivereadership.Whilegeneraltranslationlawsaccountforthebehaviouroftranslatorsingeneral,normativelawsdescribethetranslationbehaviourofasubsetoftranslators,namely,competentprofession-als,whoestablishthenorms.Normativelawsoriginateinrational,norm-directedstrategieswhichareobservedtobeusedbyprofessionals.Theselawsareempirical,spatio-temporallyfalsifiable,probabilistic,predictiveandexplan-atory.Résumé:Latraductologiedoitprendreenchargeladescriptionetl'évaluation,parlavoied'uneétudedesnormesquigouvernentlatraduction:(a)desnormesprofessionnelles,quiconcernentleprocessusdelatraduction(normesderes-ponsabilité,decommunication,derelationcible-source);(b)desnormesdel'attenteinspiréesparleslecteurspotentielsrelativementàlaformedutextetra-duit.Tandisquelesloisgénéralesdelatraductiondécriventlecomportementdetraducteursengénéral,lesloisnormativesdécriventceluid'unsous-groupe,lestraducteursprofessionnelscompétents,quiétablissentlesnormes.Lesloisnor-mativestrouventleuroriginedansdesstratégiesrationnellesadaptéesauxnor-mes,etutiliséespardesprofessionnels.Cesloissontempiriques,falsifiablessurleplanspatio-temporel,probabilistes,prévisionnellesetexplicatives.Target5:1(1993),1-20.DOI10.1075/target.5.1.02cheISSN0924-1881/E-ISSN1569-9986©JohnBenjaminsPublishingCompany2ANDREWCHESTERMAN1.DescriptivePlusEvaluativeTranslationisincreasinglyseenasaprocess,aformofhumanbehaviour.Atheoryoftranslation,therefore,shouldseektoestablishthelawsofthisbehaviour,ase.g.Touryhasarguedrecently(1991;inpress)."Law"inthissensemaybeglossedsimplyas"observableregularity".Suchlawswouldbepurelydescriptive,givinganempiricalaccountofactualtranslationbehaviour.Theywouldtakethegeneralform:UnderconditionsABC,translators(tendto)do(orrefrainfromdoing)X.Iwillrefertosuchlawsas"generaldescriptivelaws".Thisstressonadescriptiveapproachisnodoubtpartlyduetothelongtraditionofconfusionintranslationstudies,betweendescriptiveandprescriptiveaims.Generaldescriptivelawsoftranslationbehaviourcouldbesetup(i.e.behaviouralregularitiescouldbestated)foranyandeverytypeoftrans-lator,howevercompetent,andforanykindoftranslation—ononecondi-tion:thatthekindofbehaviourstudiedissuchthatitcan(appropriately)becalledtranslating,i.e.thattheresultingproductisacceptedas"atransla-tion".Thisdescriptiveview,bydefinition,wouldconsiderboththedegreeofproficiencyofthetranslatorandindeedthequalityofthetranslationirrelevant.Whatonestudiesisthebehaviourof(peoplewhocallthem-selvesorarecalled)translators,togetherwiththeend-resultsoftheirbehaviour,whichtheycalltranslationsandwhichareacceptedassuch.Thislatterpointisworthstressing.Whohastherighttodesignateagiventext"atranslation"?First,thetranslator:itisatranslationifI,theproducerofthetext,sayso.Thetranslatormaybetheonlypersonwithaccesstobothsourceandtargettexts,andthustheonlypersoninapositiontoestablishthetranslationstatusofthetargettext.Second,atextisatrans-lationifitsreceiversindeedacceptitasatranslation;theynormallydoso,ofcourse,inaccordancewiththetranslationstatusofthetextasclaimedbythetranslator,i.e.theyaccepthisorherword.Incasesofconflictbetweenthewriter'sclaimandthereaders'judgement,thetranslationstatusofthetextisindispute.Macphersonmighthavethought"theythinkmyOssianpoemsaretranslations,butactuallytheyaren't";andwecanalsoimagineasituationinwhichreaderswoulddenythatagiventextwasatranslation,despitethewriter'sclaim(orhope)—forinstanceifthetextwereunac-ceptablydistantfromthepurportedsourcetext.Butifthereadersacceptthewriter's(honest)claimthatatextisatranslation,thisisnormallyenough.FROM'IS'TO'OUGHT'3Itisthusbynomeansthecasethatsomemythical"perfectequiva-lence"(oreven"adequateequivalence")hastoexistbeforeatextcanbeappropriatelycalledatranslation.Sometranslationsareappalling,buttheyareneverthelessappallingtranslations.Othertextsareclaimedandacceptedtobetranslationseventhoughtherelationbetweentargettextandsourcetextistenuous.Recall,forinstance,theexperimentalsymbolisttranslationsofphoneticformalone,asinJandl's(1966)GermantranslationofWordsworth'sMyheartleapsupwhenIbehold/Arainbowintheskyas"Maihartliebzapfeneibehold/errennbohrinseeskai"(quotedatgreaterlengthbyToury(1980:44)).Thisisundeniablyatranslation,ofasort(apartfrombeingaLautgedichtinitsownright).Infactitseemsthattheonlynecessaryandsufficientconditionforatexttobeappropriatelycalled"atranslation"isthattheremustbesomeperceivedrelationbetweentargettextandsourcetext.Thisconditionismetif(a)thetranslatorclaimsthatsucharelationexists,and(b)thereceiv-ersofthetext(intersubjectively)acceptthatsucharelationexists(cf.alsoToury1985).Theclaimoftranslationstatusisobviouslythestronger,thelargertheproportionofreceiverswhoagree.Ifreceiversdisagree,allwecansayisthatthetranslationstatusofthetextinquestionisindoubt,asitwouldbeifthe(majorityofthe)receiversdisagreedwiththetranslator'sclaim.Descriptivetranslationstudieswillthereforesetouttodescribeandexplainanybehaviourwhichleadstosomethingthatcanbeappropriatelycalledatranslation.Obviousformsofdescriptiveresearcharethusthestudyofindividualtranslationproducts,andalsothestudyofthetransla-tionprocessusingforinstancetheprotocolmethod.Inthisrespect,thestudyoftranslationbehaviour,andthegeneraldescriptivelaws(statementsofregularities)thatemergefromthisstudy,aresimilartothosepertainingtoanyotherformofhumanbehaviour.Suchlawswouldbeprobabilistic,ofcourse,andnotuniversal.Theywoulddescribewhatpeople(atvariouslevelsoftranslationcompetence)tendtodo,undercertaingivencircumstances.Butnowaproblemarises.Apurelydescriptiveapproachwillnotincorporateanyevaluativeelements,sothatastudyofthetranslationbehaviourofabeginneris,quatranslation-theoreticalstudy,justasvalidasthestudyofthebehaviourofacompetentprofessional.Asarguedabove,goodtranslationsandbadtranslationsareneverthelessbothtranslations.Unfortunately,ofcourse,suchanapproachnecessarilyoverlooksmuchof4ANDREWCHESTERMANthemotivationforstudyingtranslationbehaviourinthefirstplaceandinevitablyleadstoaratherone-leggedtheory.Whatweneedtoknow,notinsteadofbutinadditiontothesegeneraldescriptivelaws,iswhatmakesagoodtranslation.OnwhatcriteriadowewanttosaythattranslationbehaviourAisbetterthantranslationbehaviourB?Constitutivelawsofbehaviour-that-can-be-defined-as-translation-behaviourmustsurelybesupplementedbyregulatorylawsofgoodtranslationbehaviour.Translationtheory,ifitistotaketheformofatheoryoftranslationbehaviour,mustincludebothadescriptiveandanevaluativeelement.Onewayofdoingthisisvianormtheory.Indeed,theconceptofnormshasplayedapartinthedevelopmentoftranslationtheorysincethe1970's.Delabastita(1991)hasrecentlyarguedthatitispreciselyviathestudyoftranslationnormsthathistoricalstudiescanbelinkedtopurelytheoreticalresearch.2.LawsandNormsToury(1980:51)seestranslationnormsasfallingbetween"objective,rela-tivelyabsoluterules"and"fullysubjectiveidiosyncrasies".Normsgovernbehaviouraltendencies,orratheracontinuumofmoreorlessmandatorybehaviouraltendencies.Further,adistinctionismadebetweennormsand"universals"oftranslationbehaviour:universalsareactualtendenciesthatmayormaynotbedesirable—suchasthetranslators'well-knownhabitoftendingtomakeexplicitinthetargettextwhatwasimplicitinthesourcetext,orToury'sownexampleoftextemestendingtobetranslatedasreper-toremes(Toury1991,andespeciallyinpress)."Universals"thereforedonothaveprescriptiveforce;normsdo.Touryreturnstothisdistinctioninalaterpaper(1991).Speakingnowof"laws"ratherthan"universals",but(Iassume)withthesameintendedsense,hestressesagainthatthelawsoftranslationbehaviourarenotpre-scriptive,notdirectives."Theyaredesignedtofacilitatethepredictionof'realworld'phenomenaand/ortheirexplanation"(p.187,emphasisorigi-nal);however,"inthemselvestheydonotobligeanybody,unlesstheyareacceptedasbindingnormswithina(recipient)culture"(ibid.,emphasisadded).Somelawsthusbecomenorms:inotherwords,certainbehaviouralregularitiesareaccepted(inagivencommunity)asbeingmodelsorstan-dardsofdesiredbehaviour.FROM'IS'TO'OUGHT'5Wemaythereforeanalysenormsdescriptivelyasobjectsofstudy.But,insofarastheyareindeedacceptedbyagivencommunityasnorms,theybydefinitionhaveprescriptiveforcewithinthatcommunity.Bartsch(1987:4)definesnormsasthe"socialreality"of"correctnessnotions".Withrespecttolanguagenorms,peopleinagivensocietyhave,intersubjectively,certainnotionsofwhatconstitutescorrectorappropriatelinguisticbehaviour(ingivensituations)inthatsociety;thesenotionsresideinthesocialconscious-nessasnorms."Residinginthesocialconsciousness"entailsthatnormsbeconsciouslyknown(oratleast,thattheybepotentiallyaccessibletocon-sciousknowledge)byindividualsinasociety:toputitbluntly,"normsnotknowntoexistdonotexist"(Itkonen1983:73).Normsfunctionbyvirtueoftheirsocialexistenceplustheirinternaliza-tionbyindividualmembersofagivensociety.Thefunctionofnormsistoregulatebehaviour,toestablish"acceptablemarginsofdeviation"(Bartsch1987:70);theyalsocreateandmaintainsocialorder,andtheysavetimeandeffort.Norms"reducethecomplexityofperceivingandevaluatingstatesofaffairsandbehaviourandthusmakeeffectiveactionpossible"(Bartsch1987:173).Normsfunctionbasicallyasexpectations:Successfulwaysofperceivingandactingbecomepersistentandthus'fro-zenpatterns'oforientation.Theyaremorethansocialhabitsastheyacquireanormativeforceinthepopulation.Asnormkernels[i.e.theactualcontentsofnorms],theregularitiesinthesepatternsprovidetheindividualsin[apopulation]Pwithanorientationtowardsreality(facts,possiblestatesofaffairs)andaction;thisorientationiscoordinatedforthemembersofP.Itconsistsbasicallyofexpectationsaboutsociallyrelevantthingsandevents,ofexpectationsaboutthebehaviourandintentionsofothers,andofexpectationsaboutothers'expectationsaboutone'sownbehaviourandintentions.(Bartsch1987:173)Normsthatrelatetolanguageandtheuseoflanguageareprimarilyasubclassoftechnicalnorms(otherrelevantnormsaresocialandethicalnorms)(Bartsch1987:170f).Technicalnormscomprise(a)productionnorms,havingtodowithmethodsandprocesses;and(b)productnorms,havingtodowiththeformoftheend-resultsofprocesses.3.ValidationofNormsHowarenormsvalidated?Somenormsarevalidatedbyanormauthority:theimperativetheoryofnormstakestheseascentral(cf.Bartsch1987:6ANDREWCHESTERMAN76f).Thepracticetheoryofnorms,ontheotherhand,acknowledgesthatsomenormsarevalidatedbytheirveryexistencealone(seealsovonWright1968).Accordingtothisview(cf.Raz1975:53,citedinBartsch1987:76),avalidnorm(or"rule"inRaz'sterminology:thisisanunfortunateconfu-sionofterms,butIamtakingRaz's"rule"asasynonymfor"norm"here,asdoesBartsch)existstotheeffectthatxoughttodoHunderconditionCinsocietySifandonlyifthefollowingconditionshold:1.MostmembersofSregularlydoHunderC.2.Ifsomebodydoesnotcomplywiththerule[i.e.norm],heorshewillbecriticizedbyothermembersofSandsuchcriticismwillbelookedonasjustifiedbyothermembersofS,suchthatthiscriti-cismisnotcriticizedbythem.3.MembersofSrefertotherule[norm]byexpressionslike'AnxoughttodoHwhenCor'Itisarule[norm]thatxoughttodoHwhenCinordertojustifytheiractions,ordemandsmadeofothers,orcriticismofbehaviour.Notethewayinwhichmembersofasocietycanthusappealtonormstojustifytheiractions,criticismetc.Thisshowsthedifferencebetweennormsandcustomsorconventions:breachingaconventiongivesnocauseforsanctionifnonormisinvolved.Thisdifferencebetweennormsandconventionsisnotalwaysmain-tained.FollowingLewis(1969),Nord(1991)stressesthatconventionsarearbitraryregularitiesofbehaviour,arbitraryinthesensethattheyarenotnecessarilymotivated.AsNordpointsout,normsarebinding,andtheirviolationusuallyarousesdisapprovalofsomekindamongthecom-munityconcerned.Butconventionsarenotbinding,theyembodyonlypre-ferences.Nord(1991:100)alsodistinguishesbetween"regulative"transla-tionconventions(which"refertothegenerallyacceptedformsofhandlingcertaintranslationproblems")and"constitutive"conventions(which"de-terminewhataparticularcultureacceptsasatranslation"(emphasisorigi-nal)).Now,itseemstomethatNord'sconventionsareactuallynorms,notconventions.Theyarenormsprecisely,becausetheirviolationgivesrisetosomecriticalcomment—herown.InNord'sownwords,"theconventionalconceptoftranslation,consistingofthesumtotalofconstitutiveconven-tions,determinestheregulativeconventionsthetranslatormayhavetoobserve..."(1991:100,emphasisadded).TheseconventionslookprettyFROM'IS'TO'OUGHT'7binding.Bartsch(1987:110)addsafurtherclarifyingcomment:normsoflanguagemaybeconventionsfromaphylogeneticpointofview,withrespecttotheirorigins,but"fromtheontogeneticpointofview,i.e.rela-tivetotheindividualwhohasbeenbornintoaspeechcommunity,theyarenorms"—thatis,theyareexperiencedasbindingbyeverynewgenerationandeverynewcomer.Agoodexampleofnormswhicharevalidatedbytheirveryexistence,beinggenerallyconsideredrationalanddesirablebehaviourbutnotlaiddownbyanyauthority,arethenormsgoverningqueuebehaviourinmanysocieties.Oneobservesthenormofrecognizingone'sownandotherpeople'splaceinthequeue,perhapsalsothenormof"keepingaplace"forsomeoneelse;anddeviationsfromthenormareusuallymetwithcriticismintheformofsardonicorirritatedcomments.Ontheotherhand,acceptedclassroombehaviouratschool(asdistinctfromactual"schoolrules",whichareregulationsratherthannorms)con-cerninghand-raising,movement,talkingetc.isguidedbynormsofanotherkind,givenbyanauthority,theteacher.Somenormsarevalidatedbothbyauthorityandbyacceptedusage:therearenotices(byauthorities)ontheHelsinkiundergroundaskingpeoplewaitingfortrainstoletpassengersdisembarkfirstbeforetheythem-selvesenterthetrain.This"official"normofcoursedoesnomorethanmakeexplicitwhatmostpeopledoanyway:inthiscasethenormalreadyexistsbeforetheauthorityvalidatesit.4.TranslationNormsUllmann-Margalit(1977:9;citedinBartsch1987:104)pointsoutthatsomenormsare"solutionstoproblemsposedbycertaininteractionsituations".Theyhaveaproblem-solvingfunction,infact.Translationnormswouldappeartobepreciselyofthistype.Theyexistinorderforcommunicationtotakeplaceinasituationwhereitwouldotherwisebeimpossible.(SeealsoHermans(1991),whosimilarlydiscussestheapplicabilityofnormtheorytotranslationstudies.)Onwhatbasisdoesasocietyestablishnormsfortranslationbehaviour?Inthefirstplace,asubsetofindividualswhotranslatecanbeisolated,asubsetwhosetranslationbehaviourisacceptedtobestandard-setting;thisisthesubsetof"competentprofessionaltranslators".(Ibegthequestion8ANDREWCHESTERMANhereofhowtodefine"competent"and"professional"precisely:thisisanimportantissue,butnotcentraltothepresentargument.)Inthesecondplacewecanisolateasubsetof(translated)textswhicharesimilarlyacceptedtorepresenta"model"ofthedesiredquality:theyarepreciselythosethathavebeen,or(intheopinionoftheirreaders)couldhavebeen,translatedbycompetentprofessionaltranslators.Bothsubsetsare,Ithink,bestdefineddefacto:anindividualoratextwillcountasbelongingtotherelevantsubsetifthereisaconsensusofopinionintherelevantsectionsofsociety(i.e.sectionsthatsocietyacceptsasbeingrelevant)thatsuchisthecase:quitesimply,suchisthecaseifitis("generally")acceptedtobethecase.Thesetwosubsets—onebehaviouralandonetextlinguistic—arethesourcesoftranslationnorms.Ishallcallthetwokindsofresultingnormsprofessionalnormsandexpectancynorms.Botharefactual,notideal,norms.(CompareagainNord's(1991)regulativeandconstitutiveconven-tions.)4.1.ProfessionalNormsProfessionalnormsarethenormsconstitutedbycompetentprofessionalbehaviour.Theyare,ineffect,kindsofproductionnorms,governingtheacceptedmethodsandstrategiesofthetranslationprocess.Someprofes-sionalnormscontroldetailedaspectsoftranslationalbehavioursuchassourcetextanalysis,needsanalysisoftheprospectivereadership,profes-sionaluseofreferencematerialsandsoon.However,Iclaimthatallpro-fessionalnormscanbesubsumedunderthreehigher-ordernormswhichcanbeformulatedasfollows.(i)Theaccountabilitynorm:atranslatorshouldactinsuchawaythatthedemandsofloyaltyaremetwithregardtotheoriginalwriter,thecom-missioner,andtheprospectivereadership.Thisisthusanethicalnormrequiringprofessionalstandardsofintegrityandthoroughness.Byhisorheractionthetranslatorshowsthatheorsheacceptsresponsibilityforthetranslation.(Cf.theconceptofloyaltyinNord1991,andthe"trueinterpre-ter"norminHarris1990.)(ii)Thecommunicationnorm:atranslatorshouldactinsuchawayastooptimizecommunicationbetweentheoriginalwriterand/orcommis-sionerandtheprospectivereadership.Thisisasocialnormspecifyingthetranslator'ssocialroleasacommunicationexpert.Thecommunicationnormcanbederived,accordingtoBartsch(1987:194),"byapplyingtheFROM'IS'TO'OUGHT'9principleofrationalitytotheoverallgoalofcommunication,understand-ing".Itrequiresthatweshouldcommunicateinsuchawaythatothersrec-ognizeandinterpretthemeansofcommunicationinthewayweintendthemto,inthesamewayaswedoourselves.Ireturntotheprincipleofrationalitybelow.(iii)Therelationnorm:atranslatorshouldactinsuchawaythatanappropriaterelationisestablishedandmaintainedbetweentargettextandsourcetext.Thenatureofthisrelation—thetypeanddegreeofequiva-lence,inotherwords—isdeterminedbythetranslator,onthebasisofhisorherunderstandingoftheintentionsoftheoriginalwriterand/orcommis-sioner,thetypeandskoposofthetext,andthenatureoftheprospectivereadership.(Recallthediscussionofwhatcountsasatranslation,section1,above.)Beingthusdefinedinter-textually,thenormisalinguisticone.Considernowhowtheseprofessionaltranslationnormsarevalidated.Sometranslationbehaviourisindeedregulatedbynormssetupbyanauthority—byprofessionalsofonesortoranother.Translationteachers,examiners,translationcritics,evenprofessionalswhocheckthedraftsofotherprofessionals—allareimplicitnorm-authorities,whoareacceptedashavingnorm-givingcompetence.Professionalnorms,then,areatleastinpartvalidatedbynormauthorities.Buttheyalsoconstitutetheactualprac-ticeofcompetentprofessionaltranslators:theyareacceptedasexisting,asbeingtheguidelinesthatsuchtranslatorstendtofollow.4.2.ExpectancyNormsYetthisnorm-constitutingbehaviourisitselfgovernedbyhigher-ordernorms:theexpectancynorms.Theseareestablishedbythereceiversofthetranslation,bytheirexpectationsofwhatatranslation(ofagiventype)shouldbelike,andwhatanativetext(ofagiventype)inthetargetlan-guageshouldbelike.Forinstance,ifatranslationisexpectedtobecovert(House1981),itshouldmatchnon-translatedparalleltextsinthetargetcul-ture.Forcoverttranslations,then,theexpectancynormswillbeessentiallythesameasthosethatholdfornativetextsofthesametype.Ifatranslationisexpectedtobeovert,ontheotherhand,theexpectancynormswillbedifferent.Whetheratranslationisexpectedtobeovertorcovert,andhowthisexpectationthenaffectsthetranslationitself,willbedeterminedpartlybythetranslationtraditioninthetargetculture(seee.g.Hermans1991).Expectancynormsareineffectkindsofproductnorms.Theyarenot10ANDREWCHESTERMANvalidatedbyanyactualnorm-authority(unlessthetotalsetofreceiversissodesignated),butarevalidbyvirtueoftheirexistenceinthetargetlanguagecommunityandinthespecificcommunicativesituation.Aprofessionaltranslator,inotherwords,seekstodesignatargettextinsuchawaythatitwillmeettheexpectancynormspertainingtoit.Thus,atranslationentersthe"translationalsubsystem"(Hermans1991:159)anddefactobecomespartofthatsystem,whichitsel