Epistemic modality

©TheAuthorJournalcompilation©2006BlackwellPublishingLtdEpistemicmodalitymarkersinresearcharticlesw61InternationalJournalofAppliedLinguisticswVol.16wNo.1w2006©TheAuthorJournalcompilation©2006BlackwellPublishingLtdEpistemicmodalitymarkersinresearcharticles:across-linguisticandcross-disciplinarystudyEvaThueVoldUniversityofBergenResearchontheuseofhedgingstrategiesinresearcharticleshasreceivedincreasingattentionduringthelastfewdecades,butfewhavecomparedtheuseofhedgesacrosslanguagesanddisciplines.Thisarticleexplorestheuseofepistemicmodalitymarkers–animportantandfrequentlyusedtypeofhedges–inresearcharticleswritteninthreedifferentlanguages(English,FrenchandNorwegian)andbelongingtotwodifferentdiscip-lines(linguisticsandmedicine).Genderdifferencesarealsoexamined.ThematerialiscompiledwithintheelectroniccorpusoftheKIAPproject(CulturalIdentityinAcademicProse).Statisticalanalysesofthedata(Kruskar-WallisandMann-Whitneytests)showthatNorwegian-andEnglish-speakingresearchersusesignificantlymoreofthesehedgesthantheirFrench-speakingcolleagues.Disciplinaryaffiliationandgenderseemtohavelittleinfluenceontheproportionofepistemicmodalitymarkersinatext,butthereareinterestingdifferencesbetweendisciplinesasregardsthetypeofmarkersused.Thesefindingsmayhaveimplicationsfortheteachingofacademicwritingaswellasforcross-culturalunderstandingbetweenacademics.Keywords:epistemicmodality,hedging,academicdiscourse,cross-linguisticandcross-disciplinaryvariationCesdernièresdécennies,larecherchesurl’emploid’atténuateursdansdesarticlesderechercheareçuunintérêtcroissant.Cependant,ilexistepeud’étudescomparantl’emploid’atténuateursàtraversleslanguesetlesdisciplines.Laprésenteétudeexaminerauntypefréquentd’atténuateurs,àsavoirlesmodalisateursépistémiques,etleuremploidansdesarticlesderechercherédigésentroislangues(l’anglais,lefrançaisetlenorvégien)etappartenantàdeuxdisciplines(lalinguistiqueetlamédecine).Lesdifférencesdesexeserontexaminéesaussi.LesmatériauxsonttirésducorpusélectroniqueduprojetKIAP(Identitéculturelledanslediscoursscientifique).Lesanalysesstatistiques(lestestsdeKruskar-WallisetdeMann-Whitney)révèlentquelesauteursanglaisetnorvégiensutilisentplusdemodalisateursépistémiquesqueleurscollèguesfrançais.L’appartenancedisciplinaireetlesexedel’auteurnesemblentpasinfluencersubstantiellementlafréquencedecesmarqueurs,maisonobservecertainesdifférencesintéressantesentrelesdisciplinesquantauxtypesdemarqueursutilisés.Lesrésultatsobtenuspeuventavoirdesimplicationspourl’enseignementdelarédactiondetextesscientifiquesaussi62wEvaThueVold©TheAuthorJournalcompilation©2006BlackwellPublishingLtdbienquepourlacompréhensioninterculturelleentrechercheursdeculturesdiverses.Motsclés:modalitéépistémique,atténuation,discoursacadémique,variationinterlangagièreetinterdisciplinaireIntroductionFollowinganincreasingfocusinrecentyearsonacademicwriting,numerousacademicwritingclasseshavebeenestablishedinmanydifferentcountries.Unfortunately,theteachingpracticesofsuchcoursesoftenseemtobebasedontraditionalnormativeprinciplesratherthanonsolidempiricalevidence.Itisimportantthatthecontentofsuchclassesbebasedonempiricalresultsfromanalysesofactuallanguageuseratherthanontraditionalnormativeprinciplesthatinmanycasesneedtobereconsidered.1Academicwritingbecomesespeciallychallengingwhenthetextistobewritteninaforeignlanguage.Englishhasbecomethelinguafrancaofacademicdiscourse,andnovicesaswellasestablishedresearchersmustbeabletoexpressthemselvesinthatlanguageiftheywanttobefullyacceptedmembersoftheinternationalacademiccommunity.ThishasbecomemoreandmoreimportantoverrecentyearsasthepressuretoproduceworkinEnglishandpublishinternationallyhasincreased.DuetothispositionofEnglishasalinguafrancainacademicsettings,numerousAcademicEnglishcoursesforresearchersandstudentsatalllevelsarebeingofferedincountriesalloverEurope.Theseclasses,too,needtobasetheirteachingpracticesonempiricalresearch.Forthispurpose,analysesofcontemporaryusageofacademicEnglishareessential.Thisarticleaimsatcontributingtothestockofempiricalevidencethatcanbeusedforsuchpedagogicalpurposes.Thefeaturefocusedonherewillbehedging,morespecificallytheuseofepistemicmodalitymarkers,whichconstituteaspecificandfrequenttypeofhedge.ThearticledealswiththeuseofepistemicmodalitymarkersnotonlyinEnglishbutalsoinFrenchandNorwegian.Thechoiceofalanguage-contrastiveapproachismotivatedbytheviewthatnotonlyEnglishacademiclanguagewithitsspecificfeaturesshouldbedescribedbutalsoitssimilaritiesanddifferenceswithotherlanguages,sothatstudentsandresearchersfromnon-EnglishbackgroundscaneasilycompareandcontrasttheacademiclanguageoftheirownmothertonguewithEnglishacademiclanguage.Hedging,orthemitigationofclaims,isoftenseenasarhetoricaldeviceusedtoconvinceandinfluencethereader.Itisanargumentativestrategyconsideredtobecrucialtothewriterofscientifictexts(cf.Myers’1989:13assumptionthatallstatementsconveyingnewknowledgearehedged,becausetheyhavenotyetgainedacceptanceinthescientificcommunity).Thisisrelatedtotheviewthatscientificdiscourseisnotonlyinformativeandcontent-orientedbutalsoaimsatconvincingthereader(seeMarkkanen©TheAuthorJournalcompilation©2006BlackwellPublishingLtdEpistemicmodalitymarkersinresearcharticlesw63andSchröder1997:9).Thetraditionalviewoftheresearcharticleasanobjective,neutral,informativeandfactual-orientedgenrehasbeenquestionedbynumerousstudies(seee.g.Bazerman1988;Swales1990;Hyland1998,2000;Fløttum,DahlandKinninpreparation;Vassileva1998).MasteringEnglishacademicdiscourseisthereforenotrestrictedtomasteringEnglishvocabulary,syntax,morphologyetc.andthegenreschemataofthedisciplineinquestion.ItisalsocrucialtomastertherhetoricalstrategiesandgenrepracticesspecifictoEnglishacademicdiscourse.AsVassileva(1997:203)pointsout:“Itis...essentialforanyscientistanxioustobecomeorremainamemberoftheinternationalacademiccommunitytomastertherhetoricalorganizationofEnglishacademicdiscourse.”Asculturaldifferencesinargumentationstrategiesandrhetoricalmeansareembodiedinlanguageuse,itisessentialtohavesomeknowledgeofthesedifferenceswhilewritinginaforeignlanguage.Awarenessofculturaldifferenceswithinacademicdiscourse,suchasthedifferencesintheuseofhedges,isimportantforresearcherswhowanttoexpressthemselvesandreadacademictextsinlanguagesotherthantheirown.Inadditiontothelanguageaspect,disciplinarydifferencesintheuseofepistemicmodalitymarkerswillbeexaminedhere.Muchoftheresearchinvestigatingvariationacrossdisciplineshasfocusedondifferencesintextstructure,referringtoformatssuchastheIMRADmodelinthehardsciencesandSwales’(1990)CARSmodelforintroductions.ButasFløttum(2005a:30)pointsout,“thereismoretothegenreoftheresearcharticle...thanamoreorlessstraightforwardstructuretobeidentifiedandimitatedbystudents”.Althoughgenreschemataareimportant,itisequallyimportanttofocusontherhetoricalorganizationthattakesplacewithinthetext.Ashedgingisanimportantelementintherhetoricalorganizationofatext,thestudyofhedgesacrossdisciplinescantellussomethingabouttheargumentativestrategiesusedindifferentdisciplines.Itisimportanttobeawareofdisciplinarydifferences,becauseresultsfromonedisciplinecannotautomaticallybetransferredtootherdisciplines.Everydisciplinehasitsownterminologyandalso,itseems,itsownpreferredrhetoricalstrategies.Thecontentofanacademicwritingcourseshouldthereforebeadjustedtotheappropriateresearchfield.Quiteaconsiderableamountofresearchonhedgesandtheircom-municativefunctionsinEnglishacademicdiscoursehasbeencarriedoutduringthelastfewdecades(seeforexampleVihla2000;Varttala1999,2001;Salager-Meyer1994;Salager-Meyeretal.2003;Hyland1994,1998,2000;Skelton1997;Meyer1997).Formanyotherlanguages,ontheotherhand,thereisclearlyaneedformoreresearch.Generally,verylittleresearchhasbeendoneonhedginginlanguagesotherthanEnglish(exceptperhapsforGerman,seeMarkkanenandSchröder1997:252–3forreferences),andfewhavecomparedhedgingstrategiesacrosslanguages.Asregardsvariationbetweendisciplines,Varttala(2001)hascomparedthehedgingstrategiesofthreedifferentdisciplinesandfoundthatthefrequencyofhedgesmayvary64wEvaThueVold©TheAuthorJournalcompilation©2006BlackwellPublishingLtdconsiderablyfromonedisciplinetoanother.MarkkanenandSchröder(1997:10)suggest,ontheotherhand,thatdisciplinaryvariationintheuseofhedgesisnotasgreatashasoftenbeenassumed.AccordingtoVarttala(2001:41–2),verylittleempiricalresearchexistscomparinghedgingindifferentdisciplines,andwethereforestillknowverylittleaboutdisciplinaryvariation.Inordertocontributetofillingthesegaps,thisarticleaimsatcomparingtheuseofaspecifictypeofhedges,viz.epistemicmodalitymarkers,inEnglish,FrenchandNorwegianresearcharticlestakenfromthedisciplinesofmedicineandlinguistics.Theoverallaimistoexploretowhatextenttherearedifferencesbetweenthesetwodisciplinesandthethreelanguageswhenitcomestotheuseofepistemicmodalitymarkersinresearcharticles.ThepresentstudyiscarriedoutwithintheframeworkoftheNorwegianKIAPproject(KulturellIdentitetiAkademiskProsa=‘CulturalIdentityinAcademicProse’)which,throughanalysesoflinguisticdevicesintexts,seekstoanswerthequestionofwhetherculturalidentitiescanbeidentifiedinacademicprose,and,ifso,whethertheseidentitiesarenationalordiscipline-specificinnature.Beinganimportantrhetoricaltool,epistemicmodalitymarkersconstituteoneofthelinguisticdevicesthatmaycontributetoananswertothisquestion.ThematerialdiscussedinthisarticleisasubcorpusoftheelectronicKIAPcorpus(seemoreabouttheKIAPprojectonwww.kiap.aksis.uib.no).EpistemicmodalitymarkersDespitethegrowingbodyofresearchconcerninghedgingstrategiesinacademicdiscourse,theresearchfieldsuffersfromalackofagreementonimportantterminologicalquestions.Noadequateandprecisedefinitionoftheterm‘hedge’hasbeenagreedupon,andtherehasbeenconsiderabledebateonhowtounderstandtheterm.(cf.EnglishforSpecificPurposes16.4and17.3).AcommonlyuseddefinitionisHyland’s(1998:5),whichstatesthathedgesare“themeansbywhichwriterscanpresentapropositionasanopinionratherthanafact”.2Thisfairlybroaddefinition,whichhasbeencriticisedforbeingtoovague(Crompton1997),allowstheinclusionofaseriesofdifferenttypesofmarkers,rangingfrommarkersofuncertaintyandpointerstomethodologicallimitationstowhatSalager-Meyer(1994:154)calls“emotionally-chargedintensifiers”,i.e.expressionslikeextremelyinterestingandsurprisingly.Whetheroneoptsforabroadorarestricteddefinitionoftheterm‘hedge’,thegroupofmarkersselectedforthispaper–i.e.epistemicmodalitymarkersexpressinguncertainty–remainsadomin-antandbasictypeofhedge(seeSalager-Meyer1994;Varttala1999;Vihla2000).InspiredbyLyon’sdefinitionofepistemicmodality,“[a]nyutteranceinwhichthespeakerexplicitlyqualifieshiscommitmenttothetruthofthe©TheAuthorJournalcompilation©2006BlackwellPublishingLtdEpistemicmodalitymarkersinresearcharticlesw65propositionexpressedbythesentenceheutters...isanepistemicallymodal,ormodalizedutterance”(Lyons1977:797),Iwilldefineepistemicmodalitymarkersaslinguisticexpressionsthatqualifythetruthvalueofapropositionalcontent(forexampleperhaps,probably).Epistemicmodalitymarkersthusmarktowhatextentonecanrelyontheinformationwhichisbeingconveyedbytheproposition.Take,forexample,theproposition‘smokingcauseslungcancer’.Thetruthvalueofthispropositionismarkedindifferentwaysin(a–c):a)Itispossiblethatsmokingcauseslungcancer.b)Smokingprobablycauseslungcancer.c)Weknowthatsmokingcauseslungcancer.Thepropositionismarkedasapossibilityin(a),aprobabilityin(b),andacertaintyin(c).Thedomainofepistemicmodalityrangesinprinciplefromtotaluncertaintytoabsolutecertainty(LeQuerler1996:71).Inthisarticle,however,Iwillconcentrateonexpressionsofuncertaintyonly.Markersthatstateorconfirmthetruthofaproposition(likeexamplecdoes)arenothedgesandwillnotbeconsideredhere.ThecriteriaIusedtoidentifyepistemicmodalitymarkersinthetextswerethefollowing:i)Themarkerhadtoexplicitlyqualifythetruthvalueofacertainpropositionalcontent.ii)Themarkeralsohadtobealexicaloragrammaticalunit(thusIamnottalkingaboutentirephrasesorparagraphsthatareusedtotonedownthefindings).Seebelowformoreinformationaboutthemethodusedtodistinguishepistemicandnon-epistemicsenses.Thecriterionofexplicitqualificationisimportant:verbssuchasproposeandclaim,whichhavebeenincludedincertainstudiesofepistemicmodalitymarkers(e.g.inVarttala1999:185),havenotbeenincludedinthisone.Withthesemarkers,theremightbeanimplicitqualificationofthetruthvalueofthepropositionalcontentasquestionable,buttheirmainfunctionisthatofareportingverb,introducingthepropositionalcontentwithoutreallyqualifyingitstruthvalue.Itshouldalsobeemphasizedthatepistemicmodalityisextra-propositional,i.e.anepistemicmodalitymarkeralwaysmodifiesapropositionalcontent.Thelexicalverbassume,whichinmanyothercontextswouldbeclassifiedasanepistemicmodalitymarker,isnotclassifiedassuchinexample(1)becausethereisnopropositionalcontenttobemodified:1)...explicitcontentis‘muchmoreinferentialandhenceworthyofpragmaticinvestigation’thanisassumedbypragmatistsintheGriceantradition(engling05).366wEvaThueVold©TheAuthorJournalcompilation©2006BlackwellPublishingLtdInordertoillustratewhatkindofmarkershavebeenincluded,letuslookatsomevalidexamples,takenfromtheKIAPcorpus:2)Ashasbeendiscussedintheliteratureonrelativeclauses,gerundiverelativesseemtohaveareducedclausestructurewhencomparedtofullrelatives(seeWilliams1975,Stowell1982).(engling01)Herewehaveanidentifiablelexicalunit,thesemi-auxiliaryseem,whichmodifiesapropositionalcontent,namely‘gerundiverelativeshaveareducedclausestructurewhencomparedtofullrelatives’.Inthenextexample,wehavetwoepistemicmodalitymarkers,thefirstonebeingthelexicalverbsuggestandthesecondagrammaticalunit,themodalverbmight:3)Ourfindingsinnon-smokingmotherssuggestthatthepositiveassociationspreviouslyreportedamongfirstbirthsmightsimplyreflectinadequateadjustmentforconfoundingvariables.(engmed01)Thepropositionthatisbeingmodifiedis‘positiveassociationspreviouslyreportedamongfirstbirthsreflectinadequateadjustmentforconfoundingvariables’.Itshouldbenotedthatmarkerslikemightandsuggestarepolysemousanddonotalwaysfunctionasepistemicmodalitymarkers.Iwillreturntothequestionofpolysemousmarkersinthesectionconcerningtheclassificationoftheoccurrences.ResearchquestionsandhypothesesOneofthepurposesofthisarticleistoinvestigatewhichfactorsinfluencetheuseofhedgesor,moreprecisely,theuseoftheselectedepistemicmodalitymarkersinresearcharticles.Itisreasonabletobelievethatfactorssuchastheobjectofstudy,typeofstudy,researchfield,language,theauthor’sageorpositioninthescientificcommunityetc.influencetheauthor’suseofsuchhedges.Itisofcourseimpossibletocheckforallpossiblefactorsinonesinglestudy.Inthisarticle,Iexaminethreefactorsthatmightinfluencetheauthor’suseofepistemicmodalitymarkers:discipline,languageandgender.ThefirsttwofactorsarethevariablesstudiedintheKIAPprojectandtheonesthattheKIAPcorpusselectionisbasedon.Firstly,weshallhavealookatthedisciplinefactor.Doesdisciplinaryaffiliationinfluencethefrequencyofepistemicmodalitymarkersused?Itcouldbeassumedthatonewouldfindmorehedgesinthelinguisticarticlesthaninthemedicalones,becausethislatterdisciplineisthoughttobeclosertothepositivisticideal,andthustheclaimsmadewouldbemorefactual-oriented(cf.MarkkanenandSchröder1997:10).However,medicineisa©TheAuthorJournalcompilation©2006BlackwellPublishingLtdEpistemicmodalitymarkersinresearcharticlesw67disciplinewithatraditionallyhighfrequencyofhedges(Varttala1999:177).Secondly,weshallhavealookatthelanguagevariable.Doesthefrequencyofepistemicmodalitymarkersvaryacrosslanguages?Ifso,whataretheexplanationsforthis?Thirdly,weshallexaminethepossiblefactorofgender.Aretheredifferencesbetweenmaleandfemaleauthorswhenitcomestotheuseofepistemicmodalitymarkers?Somestudiesshowthatwomenusemorehedgesthanmen,makingroomforsuggestionslike“womenarelesssureofthemselvesasscientists”or“womenaremorepoliteandmoreattentivetothereader”,whileotherstudiesshownosuchdifferences(seeMarkkanenandSchröder1997:8–9).Finally,itwouldbeinterestingtoinvestigatewhetherthefrequencyofhedgescorrelateswithindividualmeasurablefactorssuchasageandpositioninthescientificcommunity.However,Ihavenotcheckedforthesehere.Infutureresearchthiscouldbeafruitfulcoursetopursue.MaterialandmethodMaterialInordertoanswerthequestionsposedabove,Istudiedthefrequencyofepistemicmodalitymarkersinaselectionof120researcharticleswritteninEnglish,FrenchandNorwegianandbelongingtothedisciplinesofmedicineandlinguistics.ThematerialisasubcorpusoftheKIAPcorpus,whichisanelectronicdatabaseconsistingof450researcharticlesfromthreedifferentdisciplinesandwritteninthethreelanguages.InbuildinguptheKIAPcorpus,articleswrittenbynativespeakerswereselectedtotheextentthatitwaspossibletodeterminethis.However,itwasnotalwayspossibletoestablishthenationalityoftheauthor(s).4ForEnglish,bothAmericanEnglishandBritishEnglisharerepresented,butpotentialdifferencesbetweenthetwowillnotbeconsideredhere.TheresearcharticlesconstitutingtheKIAPcorpusarealltakenfromwell-recognized,high-qualityrefereedjournals.Forthisstudy,thearticlescomefromthesourceslistedinTable1.Allthearticleswerepublishedduringthelastdecade,mostofthemduringthefivemostrecentyears.Theresearcharticlesareevenlydistributedoverthesixsubgroups–Englishlinguistics(engling),Englishmedicine(engmed),Frenchlinguistics(frling),Frenchmedicine(frmed),Norwegianlinguistics(noling)andNorwegianmedicine(nomed).Thematerialincludesthebodyofthearticles,i.e.thecompleterunningtextofthearticles,excludingabstracts,notes,biblio-graphies,quotations,linguisticexamples,tablesandfigures.Sincethenumberofwordsisveryunevenlydistributedoverthesixsubgroups(considere.g.thelengthoftheEnglishlinguisticarticles),attentionshouldbepaidtotherelativefrequencyofmarkersratherthantothenumberofoccurrences.68wEvaThueVold©TheAuthorJournalcompilation©2006BlackwellPublishingLtdMethodTheselectionofmarkerswasprimarilybasedonfrequencyinanexploratorycorpusconsistingof30articles.Inordertoensure(totheextentthatitispossible)therepresentativenessoftheexploratorycorpus,itreflectedinitscompositionthelargercorpus’distributionoverdifferentjournals,disciplinesandlanguages.Forthearticlesconstitutingtheexploratorycorpus,allepistemicmodalitymarkerswerewrittendownandcounted.Asmentionedabove,inordertobeconsideredanepistemicmodalitymarker,themarkerhadtoqualifyexplicitlythetruthvalueofaparticularpropositionalcontentandbealexicalorgrammaticalunit.Themostfrequentepistemicmodalitymarkersweresubmittedtoaquantitativeanalysisofthecorpusasawhole.ThesemarkersarelistedinTable2.5Inaddition,somemarkers(Frenchpeut,peut-être,indiquerandpossible,Englishprobably,couldandpossible)wereselectedeventhoughtheywerenotparticularlyfrequentintheexploratorycorpus.Thesemarkerswereselectedbecausetheyortheirequivalentsintheotherlanguagesareoftenconsideredasprototypesofepistemicmodalitymarkers,and/orbecausetheyseemedparticularlyinterestingtostudyfromacontrastiveperspective.Englishprobably,forinstance,wasquiterareintheEnglishexploratorycorpus(2occurrences),whiletheNorwegianequivalentsannsynligviswasthethirdmostfrequentmarkerintheNorwegianexploratorycorpus.EpistemicindiquerTable1.SourcesusedforthisstudyLanguageDisciplineWordsSourcesNo.ofarticlesFrenchLinguistics68,727TravauxdeLinguistique18MargesLinguistiques2Medicine61,180AnnalesdeMédecineInterne15MaladieschroniquesauCanada5NorwegianLinguistics90,579Norsklingvistisktidsskrift20Medicine43,234TidsskriftforDenNorske20LægeforeningEnglishLinguistics170,981JournalofLinguistics10EnglishforSpecificPurposes5Language4LinguisticInquiry1Medicine59,410JournaloftheAmerican10MedicalAssoc.B