Gender features in English

Shortnote

GenderfeaturesinEnglish*

KENICHINAMAI

1.Introduction

Inthestandardminimalistframework(Chomsky1993,1995),genderisconsideredasoneofthew-featuresthatparticipateinsyntacticfeaturechecking;thus,itisa‘‘formal’’(asopposedtosemanticandphonological)featureintheterminologyofChomsky(1995).Notalllanguages,how-ever,seemtopossessthisfeature,Chinesebeingonesuchexample.1Ontheotherhand,itiswidelyknownthatmanyEuropeanlanguagesdopossessthisfeature.Italian,forexample,clearlyshowsitspresence,asiswitnessedin(1)below(m=masculine,f=feminine).

(1)a.

b.il/*lalibro‘the(m)/*the(f)book(m)’la/*iltavola

‘the(f)/*the(m)table(f)’

Noticein(1)thatobjectslikebookandtable,whichhavenoreal-worldsex,arestill‘‘referredtobyNPsthathaveamorphologicalgender’’(Napoli1993:19)and,moreimportantly,thattheytriggergrammaticalgenderagreementontheaccompanyingdeterminers.Inthisregard,there-fore,thenotionofgenderinItalianispurelygrammatical,notdirectlyre ectingthesemanticcorrelationbetweentheobjectbeingtalkedaboutanditsbiologicalsex.2

HowaboutEnglish?GreenbaumandQuirk(1990:99)statethat‘‘[i]nEnglish,genderisnotafeatureofnounsthemselves(asinsuchlanguagesasGermanorRussian).Rather,itrelatesdirectlytothe ofnouns,withparticularreferencetobiologicalsex.’’Ifso,Englishshouldnotexhibitanygrammaticalagreementintermsofgender,butexamplessuchasthefollowingmightappeartoargueagainstthispredictionat rstsight(SC=smallclause):

(2)*Hehitherself.ii(3)*Iconsider[himmygirlfriend].SCii

Infact,fromtheunacceptabilityofsentenceslike(2),Radfordarguesthatgenderfeatures(masculine,feminine,andinanimate)areindeedformalfeaturesinEnglish,sincethey‘‘playaroleinthesyntaxofre exiveLinguistics38–4(2000),771–7790024–3949/00/0038–0771

©WalterdeGruyter

772K.Namai

anaphors’’(1997:172).Likewise,Chomsky(1993:8–9),whoanalyzesasmallclauseastheprojectionofAgr,wouldalsoattributetheoddityof

(3)tothemismatchofformalgenderfeatures(i.e.him=masculinevsgirlfriend=feminine)inthesmallclauseofthesentence.3

Inthisshortnote,however,Iwillshowthattheagreementinthesesentencescannotbesyntactic(i.e.mediatedbyformalgenderfeaturechecking)butisrathersemanticinnature,whichinturnsupportsGreenbaumandQuirk’ssemanticcharacterizationofEnglishgender.Intheabsenceofevidencetosuggestotherwise,IwillthenconcludethatthetermsexismoreappropriateforEnglishgender,avoidingtheunneces-saryconfusionbetweengender(=formal)andsex(=semantic),whichmightarisefromtheoveruseoftheformerterm(seeTrask1993:115).

2.Re exives

Regardingre exives,ChomskyandLasnik(1993:553)writethat‘‘itisplausibletoregardtherelationbetweenare exiveanditsantecedentasinvolvingagreement.’’Inordertoelaboratethisideaintheirframework,theyfurthersuggestthat‘‘[s]inceagreementisgenerallyastrictlylocalphenomenon,there exivemustmovetoapositionsu cientlynearitsantecedent’’(1993:553).4TranslatingthisintotheminimalistframeworkofChomsky(1995),wemightthensaythatthere exivemovestothecheckingdomainofitsantecedentandentersintoacheckingrelationwithit.Inthecaseof(2)above,thismovement,illustratedin(4),musthappenatLF.

(4)herselfhehittii

In(4),herselfisinthecheckingdomainofitsantecedentheandthereforefeaturecheckingcannowtakeplacebetweenthetwoitems.Yet,theirgenderfeaturesdonotmatch,andasaresult,thesentenceisdeemedungrammatical.ItseemsthatthislineofargumentmustbebehindRadford’scontentionthatgenderfeaturesareformalfeatures,sincehesuggeststhat‘‘theyplayaroleinthesyntaxofre exiveanaphors,’’aswasnotedabove.

However,thisargumentdoesnotseemtoholdup.First,considerthefollowing:

(5)a.

b.Thestudentswant/*wantstoseeme.Thestudentwants/*wanttoseeme.

Fromthesesentences,wecanconcludethatnumber(andperson)isaformalfeatureinEnglish,sinceitscheckingdoesmanifestitselfinthe

Shortnote:GenderfeaturesinEnglish773

formof-sontheverbwhenthefeatureisspeci edassingular,althoughthereisnothingintrinsicallysingular(orplural)abouttheverbitself.Inthisrespect,therefore,numberinEnglishiscomparabletogenderinItalian,whichwesawin(1)above.Inshort,then,theenvironmentforordinarysubject–verbagreementinEnglishmustbepurelysyntactic.5NowconsidertheexamplesbelowfromNapoli(1993:21):

(6)a.

b.Anybodywhowantstheirpaperbackshouldtellme.Ifapersoncomesinandtheywanttoseeme,pleasetellthem

I’veleftforPanama.

Asshownin(6),thepronounsthey,their,them,etc.,canbeusedwithasingularantecedent‘‘whentheirantecedentisaninde nite’’(Napoli1993:21).6However,thisdoesnotmeanthatthepluralpronounsinthisusagehaveasingularnumberfeature,http://wendang.chazidian.compare(6b)with(7)below.

(7)...andtheywant/*wantstoseeme,...

Withthismuchinmind,lookat(8).

(8)Everybody(inthelaw rm)killsthemselves(byworkingovertime).Fromfactsabout(6)and(7),wecannowassumethatthemselves,whichhasananaphoricrelationwithasyntacticallysingularinde niteevery-body(cf.Everybodykills/*kill...),neverthelesshasapluralnumberfeaturein(8).Then,ifthere exiveweretomoveandenterintoacheckingrelationwithitsantecedentatLF,afeaturemismatchwouldoccurinthisexample,renderingthesentenceungrammatical,contrarytofact.Thus,incontrasttotherelationbetweenthesubjectandtheverbinanordinary niteclause,theonebetweenare exiveanditsantecedentcannotlogicallyinvolvesyntacticagreementrealizedbyfeaturechecking.7Suppose,instead,thattheagreementbetweenare exiveanditsante-cedentissemanticinnature.Itthenexplainswhy(8)ispossible;thatis,everybodyentails‘‘pluralmeaning’’(GreenbaumandQuirk1990:121),anditthussemantically‘‘agrees’’withthepluralre exivethemselvesintermsofnumber.Asfortheotherinde nitesin(6),asimilarexplanationcanbegivenfromasemantic/discourseperspectiveaswell;see,amongothers,BrownandYule(1984)andCornish(1986,1986–1987)inthisrespect.

Inanycase,thegrammaticalityof(8)showsthatthereisnosyntacticnumberfeaturecheckingbetweenare exiveanditsantecedent.Itshouldbenotedherethatgenderfeaturecheckingcannottakeplacealoneinthisrelationwithoutcausingaderivationalcrash,sincefeaturescanbecheckedas‘‘freeriders’’(Chomsky1995:268),anassumptionthatall

774K.Namai

featuresarecheckedimmediatelyoncetheyareincheckingrelations.In

(8),therelevantnumberandgenderfeaturesenterintocheckingrelationsatLFatthesametime,aswasexempli edin(4).Thus,thefeature-checkingaccountforthere exive-antecedentrelationseemsuntenable,andifso,theoddityof(2)cannotbeduetotheunsuccessfulcheckingoftheformalgenderfeature.Rather,itseemstoresultfromasemanticcon ictbetweenthere exiveanditsantecedentintermsofsex,whichisinaccordancewithoursemanticaccountofthe‘‘agreement’’betweenthemselvesandeverybodyin(8).

3.Smallclause

Aswasmentionedintheintroduction,itisoftensuggestedthatasmallclauseisaprojectionofAgr.8Hence,thesmallclauseinasentencelike(9a)isusuallyanalyzedasin(9b)below(masc=masculine,sg=singular).

(9)a.Iconsider[thetaxidriverentirelyinnocent].

SC

Inthisanalysis,thesubjectofasmallclauseisviewedtotakepartinsyntacticfeaturecheckingagainstAgr,towhichthepredicateofthesmallclauseadjoins.Noticethatinthisaccountoftheagreementbetweenthesubjectandthepredicateofasmallclause,allthew-features(includinggender)arepresupposedtobeformalfeatures.

However,asHeycock(1994),amongothers,pointsout,thereisampleevidencetosuggestthattheagreementinasmallclausecannotreallybe

Shortnote:GenderfeaturesinEnglish775

syntactic.Lookatthefollowing‘‘nonagreement’’betweensmallclausesubjectsandpredicates(pl=plural,sg=singular,1st= rstperson,2nd=secondperson,3rd=thirdperson,f=feminine,m=masculine):9

(10)Number

a.Diamondsareagirl’sbestfriend.

b.are[diamonds(pl)agirl’sbestfriend(sg)].SCa∞.I’mfriendswithJohn.

b∞.(a)m[I(sg)friends(pl)withJohn].SCPerson

a.Iamateacher.

b.am[I(1st)ateacher(3rd)].SCa∞.Youaremygirlfriend.

b∞.are[you(2nd)mygirlfriend(3rd)].SCGender

a.Jelacroyaismon/*madernierespoir.

Iherbelievedmy(m)/my(f)lasthope(m)

‘Iconsiderhermylasthope.’

b.Jecroyais[la(f)mon(m)dernierespoir(m)].SC(Heycock1994:123)(11)(12)

(10b)and(10b∞),and(11b)and(11b∞)representrespectivelythestruc-turesof(10a)and(10a∞)and(11a)and(11a∞)beforesubjectraising.(12b)showsthepre–clitic-movementstructureof(12a),whichisfromFrench,alanguageattestedtoshowgenuinesyntacticgenderagreement.Noticethatnumberagreement(i.e.[10]),personagreement(i.e.[11]),andgenderagreement(i.e.[12])arenotrespectedintheseexamples.10Now,ifthesubjectandthepredicateofasmallclauseweretoinvolvesyntacticfeaturecheckingtorealizeagreement,alltheseexampleswouldbedeemedungrammaticalbecauseoffeaturemismatch;yet,theyareallperfectlygrammatical.Therefore,itisveryunlikelythatthesubjectandthepredicateofasmallclauseundergosyntacticfeaturechecking.Thisstateofa airsthusformsasharpcontrasttotheagreementbetweenasubjectandaverb,whichispurelysyntactic,aswasshownin(5)above.Forourpurposesnow,thissu cestoexcludethefeature-checkingaccountofthesmallclausein(3),repeatedhereas(13).

*Iconsider[himmygirlfriend].SCii

Thatis,theoddityof(13)cannotstemfromtheunsuccessfulcheckingoftheformalgenderfeaturesbornebyhimandmygirlfriend,sinceasmallclausedoesnotseemtoprovideasyntactic-feature-checkingenvi-ronmentinthe rstplace.ThisdoesnotdirectlyargueagainsttheexistenceofformalgenderfeaturesinEnglish,butintheabsenceof(13)