管辖约束理论(语言学) 34页

  • 526.57 KB
  • 2022-08-23 发布

管辖约束理论(语言学)

  • 34页
  • 当前文档由用户上传发布,收益归属用户
  1. 1、本文档由用户上传,淘文库整理发布,可阅读全部内容。
  2. 2、本文档内容版权归属内容提供方,所产生的收益全部归内容提供方所有。如果您对本文有版权争议,请立即联系网站客服。
  3. 3、本文档由用户上传,本站不保证质量和数量令人满意,可能有诸多瑕疵,付费之前,请仔细阅读内容确认后进行付费下载。
  4. 网站客服QQ:403074932
GovernmentBindingTheoryRobertN.St.ClairUniversityofLouisville IntroductionTherewereaseriesofmajordevelopmentsleadingtotheriseofGovernmentandBindingTheory.IntheatavisticversionofsyntactictheorywhichwaspublishedasSyntacticStructures(1957),Chomskyarguedforsyntaxasthebasisforcorrelatingthelinguisticmeanings(thesemanticcomponent)withlinguisticforms(thephonologicalcomponent).Atthattime,syntaxconsistedofthreekindsofrules:(1)Thephrasestructurerulesorderedthepartsofasentenceintolinguisticcategoriesandprovidedthelexicalformsfornouns,verbs,prepositions,andadjectives;(2)thetransformationalrulesoperatedondeepstructuresandreorderedthephrasestructuralforms;and(3)morphophonemicrulesmerelychangedlexicalformswherenecessary(e.g.go+past=went).ThephrasestructureRulescreatedthedeepstructuresofsentences.Thetransformationalrulesoperatedonthesedeepstructurestoproducethesurfacestructuresofthelanguage.Thesetransformationalruleswereordered.\nBy1965,themodelwasgreatlyexpandedandrevised.AspectsoftheTheoryofSyntaxbyChomskyincludedadetailedaccountofthesemanticcomponent(projectionsrules,globalinsertionrules,selectionalrestrictions,rulesfortheinterpretationofsubcategorization,andsemanticdistinguishers).ThemostinnovativeworkonthesemanticcomponentwasdonepriortothistimeasevidencedinthepublicationofAnIntegratedTheoryofLinguisticDescriptions(1964)byJerroldKatzandPaulPostal.Thiswasenhancedfurtherbytheco-authoredworkonTheStructureofLanguage(1964)byJerroldKatzandJerroldFodor.Thephonologicalcomponentwasalsogreatlyenhancedtheinclusionofunderlyingphonemicforms,orderedrules,andphoneticoutputs.ThemostdefinitiveworkonthephonologicalcomponentcanbefoundinTheSoundPatternsofEnglish(ChomskyandHalle,1968).Thispublicationprovidedthetheoreticalframeworkforauniversaltheoryofphonology(distinctivefeaturetheory,theprincipleofstressrules,thephonologicalcycle,phonologicalconstraints,etc.).\nTheaforementionedmodelbecameknownasTheStandardTheory.By1972morerevisionsinthemodeltookplaceandthisledtoarenamingoftherevisedstandardtheory.ItbecameknownasTheExtendedStandardTheory.Therewereseveralreasonsfortheserevisions.WhatwasreferredtoastheSemanticRepresentationofasentencewasnolongerseenasasingle,uniformstructure.Thesyntacticcomponentinteractedwiththesemanticcomponentofthelanguagemanytimesduringtheprocessingofsyntacticstructures.Beforetheapplicationoftransformationalrules,forexample,thedeepstructurewenttothesemanticcomponentwhereitwasinterpretedintermsofitsfunctionalstructures.Atthistime,thesemanticcomponentprovidedinformationontheinterpretationofthevarioussemanticrolesusedinthesentencesuchasagent,patient,goal,experiencer,etc.Thisdeepstructurecontinuestobeprocessedsyntacticallyasitgoesthroughthevariouscyclesoftransformationalrules.Thisinformationisagainturnedovertothesemanticcomponentforfurtherinterpretation.Thistimethesemanticcomponentprovidesinformationonmodalstructuressuchasthescopeofnegation,andtheinterpretationofquantifiersinthe\nlanguage.Anotherkindofsemanticinformationdevelopedatthisstageincludestheestablishmentofatableofcoreferencesinthesentencebeinganalyzed.  Marysawherselfinthemirror.(co-referential)Marysawherinthemirror.(notco-referential)Marythinksthatsheisattractive.(nospecificationofreferentiality)Theserules,itshouldbenoted,couldnotbeprocessedatthedeepstructurelevelandtheprocesshadtobedelayeduntilcertainmodificationsandruleapplicationshadtakenplacewithinthetransformationalcomponentofthelanguage.Finally,atthesurfacestructurelevel,theoutputofthefinaltransformationalcycleissenttothesemanticcomponentforfurtherprocessing.Thistimethesemanticrepresentationisanalyzedforfocus,presuppositions,topicalization.Thepresuppositionofadeclarativesentencehastodowithwhatthespeakerassumesthehearerknows.Focus,ontheotherhand,hastodowithwhatthespeakerassumesthatthehearerdoesnotknow.Focusissymbolizedinthefollowingsentencesbymeansofbold-facedcapitalizedwords.Suchwordshaveextraheavystress.MarydranktheMILKPresuppositions:Maryexists,Marydidsomething,Marydranksomething.Focus:Shedrankmilkandnotwater.MARYdrankthemild.Presuppositions:sameasbeforeFocus:MarydrankthemilkandnotJohn.JohnsawMaryOldInformation:WeknowthatJohndidsomething.NewInformation:WedidnotknowthathesawMary.MarywasseenbyJohnOldInformation:WeknowthatMarywasseen.NewInformation:Wedidnotknowwhosawher.RayJackendoffhasbeeninstrumentalinthedevelopmentoftheExtendedStandardTheory.TheargumentsforthecurrentrevisionsarestatedinSemanticInterpretationinGenerativeGrammar(1972).Oneoftheconcernsthathevoicedduringtheserevisionswasthe\nneedforautonomoussyntax.Heproposedthattransformationsshouldbeappliedwithouthavingtomentionsemanticinformationsuchasreferentialitywithinatableofcoreference,theuseofindexmarkers,etc.Thiscouldbedone,heargued,bychangingthekindofinformationallowedinthedeepstructureintheStandardTheory.Consider,forexample,therulesofreflexivization,pronominalization,andEqui-NPdeletion.Ineachofthesecases,fulllexicalformsareassumedtoexistatthedeepstructurelevel.Inreflexivization,JohnsawJohnistransformedintoJohnsawhimself.Thisrequiressemanticinformationoncoreferentiality.Inthecaseofpronominalization,coreferentialityisalsoneededtotransformMarywondersifMarywillbehappyintoMarywondersifshewillbehappy.InthecaseofEqui-NPDeletion,coreferentialityisalsoneedtotransformMaryexpectsMarytowinintoMaryexpectstowin.Onecouldavoidthesereferencestothesemanticcomponent,Jackendoffargued,ifpronounsanddummysubjects(gaps)alreadyexistatthedeepstructurelevel.Thisintroductionofabstractelementsandemptycategoriesintothedeepstructureofsentencesmarkedanimportantturningpointinlinguistictheory.ItledtotheemergenceofGBTheory.Nevertheless,beforeturningtothismajorrevisionbasedonthepremiseofautonomoussyntax,itisimportanttoreviewtherepresentationofthemodeloftheExtendedStandardTheory.Thisisthemodelofgrammarwithoutautonomoussyntax.\nThedevelopmentofGovernmentandBindingTheoryasamodularmodelreallybeganin1977whenChomskyandLasnikproposedsomemajorrevisionsintheExtendedStandardTheory.Inthisarticleon"FiltersandControl"(LinguisticInquiry8.3),theyquestionedthenecessityofphrasestructurerulesinthecontextoflexicalsubcategorizationalrulesinthelexiconwhichprovidedsimilarinformation.Theyalsoaddressedthequestionoftransformationalruleswithregardtostylisticrulesversusmeaningchangingrules.TheissuesaddressedinthisearlyarticlewerefurtheraddressedbyNoamChomskyinhiswritingsonLecturesonGovernmentandBinding(1981)andSomeConceptsandConsequencesoftheTheoryofGovernmentandBinding(1982).HebecamelessconcernedwithregardtothebasegenerabilityintransformationalgrammarandmoreconcernedwithStructure-PreservingConstraintsonlanguage.Therevisionsledtoamodularmodeloftransformationalgrammar. \nSincetheconceptsofDeepStructureandSurfaceStructureweresubstantiallyrevisedbytheinclusionofabstractelements,traceelements,andemptycategories,theyweregivenanewnomenclature,d-structureands-structurerespectively.Thesearerelatedtoeachotherbymeansofmovementrules.S-structureswerefurtherdevelopedintoPhoneticFormsorLogicalForms.Whatiscalledphoneticforminvolvesmorethanmereacousticandarticulatoryinformation.Itincludedsemanticproperties,andlow-leveltransformationalrules(suchasstylisticrules),deletionrules,contractionrulesandphonologicalrules.TheLogicalFormComponentdealswiththemeaningofsentences.ThedistinctionbetweenPFandLFcomponentswasnecessarybecausetheyrepresentedverydifferentaspectsandconcernswithingrammaticaltheory.TheLFcomponentdealswithsemanticinformation. PFComponent:JohnlikesMaryLFComponent:Foreveryx,xisaperson,Marylikesx.ConcernsaboutBindingConditionsaredealtwithintheLFComponentbecausetheyinvolvesemanticissues(referentialdependencies,coreferences,etc.).Similarly,quantifierraisingdealtwithsemanticissueswithinthecontextoftheLFComponent.Thiscomponentbeganbyrepresentinginformationintermsofstandard\nlogic,butwasmodifiedtoincorporateconstituentcommand,tracetheory,andotherlinguisticissues.Thedifferencesbetweenad-structureandans-structurecaptureswhatchangeshavetakenplaceonceamovementrulehasapplied.ButGBTheoryinvolvesalotmorethanmerelyrevisingthedeepstructureofagrammar.Inincludesmanynewfeatures: D-Structure,FullyRecoverableD-Structures,S-Structure,X-BarTheoryMovementRules,TraceTheory,EmptyCategoriesCaseFilters,ThetaTheory,Well-FormednessConditions,Considerhowthesurfacestructuresentence"DidJohninviteMary?"wasdealtwithintheextendedtheoryofsyntax. DeepStructure:JohnpastinviteMaryQuestionpastJohninviteMaryAffixHoppingDoSupportdo+pastJohninviteMarySurfaceStructure:DidJohninviteMary?IntheGBTheory,movementrulesareseverelyrestricted.Onecanonlymoveelementstounfilled(empty)ortocertaindesignatedcategories.Thesentenceisnowrepresentedasaninflectionalphrase(IP).\nOneofthemovementrulestakestheINFLform(past)andmovesitintoonlyopenandemptycategoryofComplement.IntheeventthattheINFLisnotmoved,DoisdeletedbymeansofaDoDropRule.Withinthismodel,therearethreekindsofmovementrules.OnethatmovesNPs(maximalprojections),anotherthatmovesHeadsofPhrases,andathirdthatmovesWh-constructions.AnothermajorchangethattookplaceinthetransitionfromtheExtendedStandardTheorytoGBTheoryhastodowiththeshiftfromrulestorepresentations.Intheearliermodelsoftransformationalgrammar,oneaccountedforgrammaticalchangesbymeansofsyntacticrules. DeepStructure:JohnpastseeMaryPassiveMarypastbe+enseebyJohnAffixHopMarypast+besee+enbyJohn\nSurfaceStructure:MarywasseenbyJohnSucharulehadtobelanguagespecific.OnlyEnglish,forexample,insertbe+enbeforethemainverbandinsertbybeforetheagentivenounphrase.Iftransformationalrulesaretobeuniversal,theycannotalsobelanguagespecific.Hence,thesystemhastobechanged.Whatisneeded,Chomskyargued,isatheorythatreliesonconditionsontheserepresentations.Theseconditionsoperateatalllevelsandemploygeneralrulessuchas: MOVENPThischangebecameknownasthePrinciplesandParametersApproach.Later,itcametobecalledGovernmentandBindingTheory.Thisshiftfromrulestoconditionsonrepresentationsavoidedlanguagespecificrules.Italsostrengthenedtheconceptofthelanguageacquisitiondevice(LAD)inpsycholinguistics.Intheearliermodel,achildwasgivenatoolboxandwasaskedtoconstructagrammarforhimself.Hehassomeideaofwhatagrammarshouldlooklike,buthehasnoideaaboutuniversalprinciplesgovernallgrammars.Thiscouldhappenonlyifhewereprovidedwithasetofprincipleswhichdefinedtheparametersandtheconditionsoperatingonsyntacticrepresentations.ThisisexactlywhatGBTheoryattemptstodo.Thisiswhytheshiftfromrulestoconditionsonrepresentationshadtotakeplace.ItprovidesachildwiththeprinciplesandtheconditionsonnaturallanguagesandallowshimtosettheparametersinthatlanguagesuchasSOVorVSOwordorder,etc. CATEGORIESANDPHRASESTRUCTURESTheStandardTheoryofSyntaxcontainedabasecomponent.Inthisbaseweretwosubcomponents,onethatgeneratedthecontext-freePhraseStructureRulesforthedeepstructuresofthelanguageandtheotherwhichgloballyinsertedlexicalitemsintothefinalstringofthedeepstructureofthesentence.TheproblemwiththismodelisthatitalsogeneratedmanyPSrulecombinationswhichdonotoccurinlanguage.WhatwasneededwasaPScomponentthatavoidedthisproblem.Thisledtoare-examinationofPSrulesandseveralthingswerediscovered.Oneofthemisthatthelabelsforthesecategoriesarebasedonthetraditionalnotionofthepartsofspeech.Outsideofthistradition,theylackedcogentreasonsfortheirnomenclatureandtheirexistence.Therevisionofthisinformationbecameknownas\nStructureDependentConstructions,andX-BarTheory.Anotherdiscoverywhichcameoutofthisre-examinationisthattheheadofaphraseturnsouttobeanimportantlinguisticconcept.Thisledtoacloserlookatendocentricandexocentricconstructions.TherevisionsinthiscasewereknownastheHeadParameter,HeadandMaximalProjections,TheHierarchicalOrganizationPhrases,andX-BarTheory.Ineveryphrase,thereisalwaysanobligatoryelement.Thiselementiscalledtheheadofthecategory.TheheadofaNounPhraseistheNoun.TheheadofaVerbPhraseistheVerb.TheheadofanAdjectivalPhraseistheAdjective.TheheadofaPrepositionalPhraseisthePreposition.Fromthisawareness,theynotedasimilarityamongallphrases,viz.X(N,V,A,P)istheheadofXP(NP,VP,AP,PP).  NisheadofNPVisheadofVPAisheadofAPPisheadofPP______________________Therefore,XisheadofXPThisconceptbecomeknownasX-BarTheorywhereXstandsfortheheadofXP(phrase).Themoreinterestingquestionstoemergehavetodowiththeheadofasentenceandtheheadofacomplementphrase.Therationalefortheheadofasententialphraseistheinflectionalmarker,INFL.Theinflectionalmarkerincludesmuchmorethanmeretensemarkers.Italsohadtodowithagreementfeatures,etc.GiventheanalogyofXistheheadofXP,onemustreinterpretINFLastheheadofthesentenceorIP.Asfortheheadofthecomplementphrase,itisobviously,thecomplement.Hence,CistheheadofCP.INFListheheadofIP,orIistheheadofIPCOMPistheheadofCP,orCistheheadofCPThisfocusontheheadanditsprojectionsbecameknownasthecoregrammar.Thismeansthatanyotherelementsaddedtothesephrasesareadjuncts(add-ons). ThedestructionofthecityinAprilof1945.\nHeadofNPAdjunct JohnsawMarythisafternoon.HeadofIPAdjunct whetherJohnwillseeMarytomorrowafternoon.HeadofCPAdjunctItshouldbeobviousfromtheaforementionedexamplesthatwhathasbeennormallycalledtemporaladverbsarenowtreatedasadjunctstophrases.Theyarenolongerconsideredtobeperipheralpartofthesentenceanddoesnotbelongtothecoregrammar.SomeoftheterminologyandinsightsusedinX-BarTheoryarerelatedtosimilarresearchdoneintheareaofmorphology.Linguistshavealwaysstudiedhowwordsareputtogethertoformmorphologicalunitsorcompounds.Thesecompoundsfunctionasthoughtheyweresinglewords:blackmail,skyscraper,bathroom,bookshelf,football,in-crowd,etc.Twocombinationsofwordswithincompoundshavedrawnthemostattentionandcommentaryinlinguisticanalysis.Oneofthesehastodowithconstructionsinwhichoneoftheelementsisaheadword(endocentric)andtheotherelementactsamodifierofthatheadword.Suchacombinationisknownasanendocentricconstruction.Bookshelf,forexample,isanendocentricconstructionbecause"books"modifytheheadword"shelf".Suchcompoundsareeasytodetectbecausethecombinedelementsfunctiontodefineasubclassofoneofthenounsintheconstruction.Abookshelfisashelfforbooks.Similarly,anosebleedisakindofbleedingfromthenose.Theotherkindofconstructionwhichmorphologistsarefoundofdiscussingisconstructionswithcontrastwithendocentriccompounds.Thesearecalledexocentricconstructionsandtheyaredefinednegativelybecausethecombinedelementsdonotshareaheadwordandmodifierrelationship.Scapegoat,forexample,isnotakindofagoat.Itisakindofaperson.Hence,thisisanexocentricconstruction.WhatisinterestingtoGBTheoryaboutthisinvestigationofmorphologicalconstructionsisthefactthatendocentricconstructionshavemanyinterestingparallelswithX-Barphrases.Theheadwordfunctionsastheheadofthecompoundnounjustasthenounfunctionsastheheadofanounphrase.TheconclusionthatonedrawsfromthisinformationisthatallXPconstructionsareendocentric.Thisdiscoveryhasbeenstatedinevenstrongterms:therearenoexocentricconstructionsamongXPs.Itisnotenoughtonotethataphraseconsistsofaheadandamodifyingelement.Itisimportanttoalsoconsiderhowthese\nrelationshipsarestructured.Whenallofthecategorieswithinaphrasearestrungoutinalinearpattern,thisiscalledaflatstructure.Inaflatstructure,alloftheelementsareconnectedbymeansofsisteradjunction.Theyallbelongtothesamecategory(NP).Thereisabundantevidencethatphrasesarestructuredhierarchically.Thisevidencecomesfromthestudyofpro-forms.Thisiswhenacategoryisreplacedbyanotherform(pro-noun,pro-verb,pro-sentence,etc.).Whatisactuallytakingplacewithpro-formsubstitutionisthatsomehierarchicalstructureisreplacedbyanotherform,apro-form.Thesesubstitutionsoperateonstructures.Theyarestructure-based.Thesestructureoccurinlayers.Theyarelayeredstructures.  NP:IsawJohntalkingtoMary.Isawhimtalkingtoher.PP:JohnisgoingtoLondonnextweek.Johnisgoingtherenextweek.VP:JohnsawMaryandBillsawMary.JohnsawMaryandsodidBill.AP:JohnsawtheprettychildandMarysawtheprettychild.JohnsawtheprettychildandMarysawhertoo.XPsconsistofaheadelementwithinaphrasalconstruction.Theheadisalexicalcategoryandthephrasalcomponentiscalledaprojection.Sincetheseprojectsmayoccurinlayers,onespeaksofthetopmostlayerasthemaximalprojectionofaphrase.NPisthemaximalprojectionofN.VPisthemaximalprojectionofV.Thereareintermediateprojections,however,thatmustbeaccountedforin\ndescribingthestructureofphrases.ThesearecalledX-Barprojections. HEADMAXIMALPROJECTION X-BARPROJECTION VVPV-BAR NNPN-BAR AAPA-BAR PPPP-BAROnewouldnaturallyaskhowmanylevelsorlayersofstructurecanoccurfromthemaximalprojectionofaphrasetoitshead.InGBTheory,thereareonlytwosuchprojectionsallowed-themaximalprojection(XP)andtheintermediateprojection(X-Bar).Itshouldbenotedthatinadditiontoaheadelementanditsprojection,thereisanotherelementwithinaphrase.Theelementisknownasthespecifier(SPEC).AsimilarelementcanbefoundattheCPlevelanditisknownasthecomplementC.Bothspecifiersandcomplementsarenotsyntacticcategories.InthisregardtheydifferfromX-DoubleBar(XP),X-Bar,andXcategories.Theseelementsareemptycategories.Theyfunctionaslocationsforpartsofthestructureofaphrasethatmaybefilledinbyactualsyntacticcategoriesintheprocessoftheapplicationofamovementrule.BothSpecandCoccuratthesecondlevelalongwithX-Barcategories.Giventhisawarenessofendocentricconstructions,onecannowvisualizehowanXPstructureisrepresented.\n\n \n C-COMMANDANDGOVERNMENTWhenlinguistsbegantoreallystudyhowpronounsfunctioninlanguage,theyquicklylearnedthatwhereanounphraseoccurswithinaphrasestructuretreewilldeterminewhetherornotitispronominalized.ThefirstonetnotethiswasEdwardKlimaatMITwhowasthenworkingonthescopeofnegationinEnglish.KlimawenttoteachattheUniversityofSanDiego.Atthatsamecampus,RonaldLangackerbecameinterestedinthisphenomenonandwroteanarticle"OnPronominalizationandtheChainofcommand."ThisresearchwaspublishedinabookofreadingsbyDavidReibelandSanfordSchane(ModernStudiesinEnglish,1969).Langackerexploredtheparametersoftheruleofpronominalizationbystudyingitinthesentenceconjunction,verbphraseconjunction,nounphraseconjunction,andinembeddedsentences.Hisparametersoftheapplicationofthisrulewerecalledthechainofcommand.Later,ChomskywouldmakesomerevisionsandcallitConstituentCommand,orC-Command.Langackerfoundthatifanounphraseoccurredinamatrixsentence\nanditwasidenticaltothenounphraseinanembeddedsentence,thematrixNPwouldbeabletocommandandpronominalizetheembedNP.NP'sinlowersentences,ontheotherhand,couldneverpronominalizeidenticalnounphrasesinhighersentences.Ralphinthematrixsentencecommandsthepronounintheembeddedsentence.Noticethatthematrixsentenceisalsolocatedontheleftandpriortotheembeddedsentence.Languagefoundthatanotherkindofrelationshipneededtobemadeexplicitandhecallthisprecedence.Hence,heconcludedfromthestudyofthedatahewasworkingwiththatifanounphraseoccursinalowersentence,,butitprecedesanotheridenticalnounphraseinthematrixsentence,itwillpronominalizetheNPontheright.Whatthismeans,inessence,isthattheorderoftheelementsinpronominalizationissignificant.\nPronominalizationoccursifanidenticalNPprecedesanotherNP(theprecedencerelationship).ItalsooccursifanidenticalNPishigherinthephrasemarker(commandrelationship).AsLangackerinvestigatedthisphenomenon,henotedthatsomethingelsebesidesprecedenceandcommandtoaccountforsomeanomaliesthatheencounteredamongmoredeeplyembeddedsentences.Herealizedthatpronominalizationdidnotworkinthemoredeeplyembeddedphrasemarkers.\nThereissomekindofbarrierbetweentheidenticalNPsandtheirpronominalizedforms.Itisnotenoughtomerelystatethatoneofthemcommandsanotherorthatoneofthemprecedesanother.Thechainofcommandhastobemoredirect.Langackerturnedtotheconceptofdominancetoexplaintheaforementionedsentences.HenotedthatanidenticalNPmustdirectlydominateitspronoun.Ifsomekindofbarrierpreventsthedominancefrombeingdirect,pronominalizationwillnottakeplace.ChomskywastotakethisconceptofbarriersanduseitasoneofthebasicconditionswithinhisGovernmentandBindingTheory(cf.BarrierbyChomsky,1986).Klimausedtheterm"inconstructionwith"toexplainthescopeofnegation.GiventwoconstituentsAandB,BissaidtobeinconstructionwithAifnodeCthatdirectlydominatesAalsodominatesB.\nNegationisinconstructionwiththesubjectNP(John).NegationisisnotinconstructionwiththeobjectNP(Mary)sincethenode(VP)thatdirectlydominatestheobjectNP(Mary)doesnotalsodominateNegation.Obviously,therelationshipsbetween"inconstructionwith"and"chainofcommand"(dominance,commandandprecedence)aresimilar.However,thesearenotequivalentconceptswithregardtoconstituentstructure.TheframeworkthatLangackerdevelopedincludedmoreinformationandcouldexplainmuchmorewithregardtodominancethanthemodelproposedbyKlima.Langacker'sconceptcouldexplainbothnegationandpronominalization,Klima'scouldonlyaccountforthescopeofnegation.ShehadsomemoneyShedidnothaveanymoneyYouwillmarrysomeoneYouwillnotmarryanyone.Considerwhathappenswhenthesesentencesareembedded.Theresultofwhatquantifierssurfacesdependsonwhetherornotthenegationmarkerislocatedinthematrixorintheembeddedsentence.\nThenegationmarkerchangesthequantifier(someone)intoanyone.Inthiscase,thenegationmarkerdirectlydominatesthequantifier.Now,considerwhathappenswhenthenegationmarkerdoesnotdirectlydominatethequantifier.\nSincethenegationmarkerisinthematrixsentence,itdoesnotdirectlydominatethequantifierintheembeddedsentence.Hence,someoneremainsunaffectedbythescopeofnegation.Whatmakesthescopeofnegationinterestingisthatitforcesonetorecognizetheconceptofgrammaticalbarrierthatlimitthescopeofcertaintransformationalrules.Noticethatbarrierscanbefoundwhenoneisdealingwithinfinitivalcomplements.Thequestionthatonemayaskiswhetherornotothergrammaticalconstructionsarealsowithinthescopeofnegation?Thiscanbeascertainedbylookingatthefollowingexamplesofthat-clausesandrelativeclauses.Ididn'trealizethatshehadanymoney.Ididn'trobthewomanwhohadsomemoney.*Ididn'trobthewomanwhohadanymoney. Itturnsoutthatthenegationmarkercanreachdownintoathat-clause,butitcannotreachdownintoarelativeclause.Itcontrolsonlyuptoacertainpositiononthephrasemarkertree.Hence,relativeclausesarenotwithinthescopeofnegation.Theyprovidedbarrierstotheruleofthescopeofnegation.Thefollowingsentenceisambiguouswithregardtothescopeofnegation. Thecitizencouldnotvotefortheproposition. Itcouldmean:ItwouldnotbepossibleforthecitizentovoteforthepropositionItwouldbepossibleforthecitizennottovotefortheproposition. Inordertodisambiguatethesesentencesandensureaproperinterpretationofthesesentences,oneonlyneedstolocatethenegationmarkeronthephrasemarker.Thecitizen[couldnot]votefortheproposition.Thecitizencouldvote[notfortheproposition]\n Themodal(could)andtheverbphrase(votefortheproposition)formtwoseparateconstituentgroups.Ifthenegationmarker(not)isattachedtothemodel,itcommandsoneinterpretation.IfitisattachedtotheVP,itcommandsanotherinterpretation.Linguistsarealwayslookingforsomekindofverificationthattheiranalysesarecorrect.Inthiscase,theplacementoftheadverb"simply"providescorroboratingevidence.TheselexicalitemscanonlybeplacedatthebeginningofaVP.Therearetwoverbphraseitemsinquestionintheaforementionedsentences.Oneofthemis[votefortheproposition]andtheotheris[notvotefortheproposition].Byplacing"simply"attheleftmostpositionintheVP,oneconfirmsthat"not"inonesentencebelongstothemodal,and"not"intheothersentencebelongstotheVP.Thecitizen[couldnot][simplyvotefortheproposition]Thecitizencouldvote[simplynotfortheproposition]AnotherformofsyntacticcorroborationforthisplacementofnegationinthephrasemarkercomesfromthetransformationknownasPseudo-Cleft:WhatXdidwasY.Hence,thefollowingexamplesofpseudo-cleftingcorroboratethefactthatnegationcontrolsthemodalinoncesentenceandtheverbphraseintheother.WhatthecitizencouldnotdoisvoteforthepropositionWhatthecitizencoulddoisnotvoteforthepropositionFinally,furtherevidencecomesfromthenegationcontractiontransformation.Thisruleonlyoperatesonnegativeformsthatmodifyamodal.WhatthecitizencouldnotdoisvoteforthepropositionWhatthecitizencouldn'tdoisvotefortheproposition.Thecitizen[couldnot]simplyvoteforthepropositionThecitizen[couldn't]simplyvotefortheproposition\nJustwhenoneishappytolearnthatthechainofcommandcanexplainthephenomenonofpronominalizationandthescopeofnegation,itturnsoutthattheseprinciplesofcontrolcanalsoexplainwh-formationwithinthecontextwherewh+someformsthequestionword"what.":TheWHMarkercommandsthequantifier"some."Consequentlywh+somebecomeswhat.Thismarkercontrolselementsintheembeddedsentence.Itdoesnotcontrolotherelementsinthematrixsentence.Hence,itcouldnotproducethefollowingsentencesbecausesomereasonandsomemanarenotwithinthecontrolofthewh-marker.*SomemanaskedIwantedsomebookforwhatreason.*WhatmanaskedIwantedsomebookforsomereason.Onceoneisawareofthefactthatthereisachainofcommandwithinaphrasemarker,thenextconcernforlinguistshadtodowithwhysomeruleswereblockedbybarrierswhichalsoisolatedcertainportionsofthephrasemarkerintoislands.Notallconstructionsworkedthesameway.Infinitivalcomplementsandthat-clauseswereopentothescopeofnegation.Relativeclauseswerenotopentothescopeofnegation.Barriers,theComplexNPConstraintofRoss,protectedsomeexamplesofwh-formation.Inothercases,thebarriersdidnotprotectwh-formation.Whatthismeanisthatsyntacticresearchhadanewsetofproblemstosolve.ItwouldnotbeuntiltheriseofGovernmentBindingTheorythatmostoftheseproblemscouldbesatisfactorilydealwith.WhentheseissuesaredealtwithininGBTheory,however,theyarereferredtowithinaverydifferentcontextandwithadifferentvocabulary.Dominatingnodes,forexample,willbecalled"mothernodes."Dominatednodeswillbe\ncalled"daughternodes."Therelationshipsofcommand,precedenceanddominancearealltreatedasexamplesofC-Command(ConstituentCommand). C-COMMAND,M-COMMANDANDGOVERNMENTChomskyhasnotedthataphrasemarkerisessentiallyagraphinwhichthenodes(thesetsofpoints)areconnectedbybranches(solidlines).Thelabelsusedinthenodesmaybemaximalprojections(X-Double-Bar),intermediateprojections(X-Bar),orfinalprojections,i.e.,non-terminalnodes(NP,M,V,P,andothercategorylabels).Withintherevisedmodel,someofthesenodesmaybeempty(e).Alsowithintherevisedmodel,whenelementsaremoved,theyleavebackatraceelement(t).AnotherdistinctionmadeinGovernmentBindingTheoryhastodowiththenatureofdominanceandprecedence.TheSnodeoccurshigheronthephrasemarkertreeanditdominatesallothernodes.TheVPnodesdominatestheVnodeanditsterminalnode(stay).Onenodeimmediatelydominatesanothernodeifitisthenexthighernodeinthephrasemarkertreethatisconnectedbyasolidline.Hence,theSnodeimmediatelydominatesNP,M,andVPnodes.TheSnodedominatesthePPnode,butitdoesnot\nimmediatelydominatethatnode.Anotherrelationshipthatmeritscommentwithregardtothephrasemarkertreeisthatofprecedence.Onnodeprecedesanotherifitislocatedontheleftoftheothernode.Hence,theMnodesprecedeVP,V,PP,NP,N(hence,theterminalnodesstayathome).TheMnodeimmediatelyprecedestheVPandtheVnodes(theterminalnode(stay).WhatisimportantabouttheserelationshipsofdominanceandprecedenceisthatChomskyusedthemtodefinethetermsofexhaustivedominanceofconstituentsandimmediateconstituents. CONSTITUTENTS(CHOMSKY)Nodesformaconstituentifandonlyif(iff)theyareasetofnodesthatareexhaustivelydominatedbyacommonnode.Intheaforementionedexample,NP,M,andVPallbranchoutofthesamesinglecommonnode(S).Therefore,SexhaustivelydominatesNP,M,andVP.Therefore,ingeneraltermsonecouldsaythatXisaconstituentofYiffXisdominatedbyY.Similarly,onnotesthatXisanimmediateconstituentofYiffXisimmediatelydominatedbyY. ThemothernodeAdominatesnodesB,C,D,E,andF.ThemothernodeAimmediatelydominatesthenodesBandC.ThesisternodesD,E,andFareimmediatelydominatedbythedaughternodeC.Mothers[A]dominatetheirdaughters[BC]andnieces[DEF].Mothers[A],however,canonlyimmediatelydominatetheirowndaughters[BC].Daughters[C]havetheirownchildren[DEF]andimmediatelydominatethem.Nieces[DEF]arenotexhaustivelydominatedbythe\nmothernode[A],butthey[DEF]areexhaustivelydominatedbythedaughternode[C].Chomskywasaskedmanytimeswhyhedidnotusethetraditionalnotionofdividingasentenceintoasubjectandapredicate.HewasaskedthisquestionagainwhenconfrontedwithFillmore'sCaseGrammar.Hisreplywasthatsuchrelationshipshavetodowithgrammaticalfunctionsandnotgrammaticalforms.Further,thesefunctionscanbereadilyascertainedbylookingatthedominanceandprecedencerelationshipswithinaphrasemarker.  SubjectoftheSentence=[NP,S]DirectObjectoftheSentence=[NPa,VP]IndirectObjectoftheSentence=[NPb,VP]PrepositionalObject=[NP,PP]ThesubjectofasentenceistheNPwhichisimmediatelydominatedbyS.TheobjectofasentenceistheleftmostNPimmediatelydominatedbyVP.IftwoNPsoccurtotheleftoftheVP,thefirstisthedirectobjectandthesecondistheIndirectobject.Theobjectofaprepositionisdefinedstructurally.ItistheNPwhichisimmediatelydominatedbyaPP.Theuseofdominanceandprecedencetocharacterizethepartsofspeech,however,isnottheonlyreasonwhytheserelationshipsareimportant.Theyplayimportantrolesinmanyotherareasofsyntaxandsemantics.ChomskyreferredtotheseotherusesofdominanceandprecedenceasConstituent-Command(C-Command).WhenoneconsidersthenotionofC-Commandintermsofitsmother/daughterrelationships(NotaBene:fortheItaliansthiswouldbethefather/sonrelationships),theconceptiseasy.AnodeC-Commandsitssisters,itsnieces,anditsgrandnieces.\nThenodesDandEaresisters.ThefirstbranchingnodeabovethesesistersisthedaughternodeC.ThismeansthatCimmediatelydominatesDandE.ThenodeCalsodominatesnodesFandG.Now,accordingtoChomsky,onenodec-commandsothernodesifandonlyifthatnodeisnotsubordinatetoanothernodeandbotharecommandedbyacommonnode.ThenodeCcommandsDandE.ThenodeDistotheleftofE.ThenodeDdoesnotcommandEandviceversa(ThenodeEdoesnotcommandD).ThismeansthatthenodeDnowc-commandsE,F,andG.ThismeansthattheleftsisternodeDisinapositiontoc-commanditssistersandtheirnieces.ThenodeDc-commandsE,F,andGbecauseitisatthesamelevelasnodeEanditishigheronthetreethannodesFandG.Now,nodeEisalsoasisternodewithDandthecommonnodeCcommandsit.ThismeansthatnodeEc-commandsnodeD.ThemothernodeAc-commandsnothing.Itisnotasister.ThedaughternodeBc-commandsC,D,E,F,G(daughters,sisters,nieces)ThedaughternodeCc-commandsE,F,G(fellowsisterandnieces)ThesisternodeEc-commandsD(theothersister)TheniecenodeFc-commandsG(theotherniece)TheniecenodeGc-commandsF(theotherniece).Thenotionofc-commandisusedtoexplainhowanaphoraworks.Asonehasnotedearlier,inGBTheory,pronounsareco-indexed.Ananaphortakesitsreferencefromanotherexpressioninthesentence.Theelementthatservesasareferenceisknownastheantecedentoftheanaphor.\n JohnandMarylovethemselves(reflexives)[antecedent][anaphor]JohnandMaryloveeachother.(reciprocals)[antecedent][anaphor]Whattheseexamplesdemonstrateisthatananaphormusthaveanantecedentwhichc-commandsit.Iftheantecedentc-commandstheanaphor,theconditionsarerightfortheformationofreflexivesorreciprocalsinalanguage.Inthesentence"theylovedthemselves,"thesubject"they"c-commandstheobject"themselves."Thesubjectistheantecedent(they)ofthereflexiveobject(themselves).Consequently,theantecedentc-commandsitsanaphor.ChomskycallsthistheC-CommandCondition.ThereareseveralconditionsunderwhichoneNPisinterpretedasbeingcoreferentialwithanother.Oneofthesehasalreadybeendiscussed,viz.,anaphors(reflexivesandreciprocals).But,pronouns(pronominals)andotherkindsofreferenceexpressions(R-expressions)alsorequirethiskindofinterpretation. Marythinksthatsheisgoingtobeamother.(Maryandshereferstothesameperson)MarythinksthatJohnisthefather.(MaryandJohnareR-expressionswithoutcoreference)TheconditionsunderwhichNPsareinterpretedasbeingcoreferentialwithotherNPsisknownasBindingTheory.INadditiontothe\nconditionofcoreferentiality,bindingtheoryalsohasalottosayabouthowtheseelementsarebound.Thismeansthatbindingtheoryhastoexplainhowoneconstituentorargumentcommandsanother(i.e.,thenotionofc-commandisrelevanthere).Whatisimportantinthiscase,however,isthatnoconstituentcanc-commandelementsoutsideofthemaximalprojectionthatcontainsit.Themaximalprojectionisknownasthegoverningcategoryandthec-commandingargumentiscalledthegovernor.ThesubjectNP(Mary)c-commandstheobjectNP(me).NotethatSisthegoverningcategoryoftheobjectpronoun(Me).ThegovernorofmeisVanditiscontainedwithinthemaximalprojection(S).sofar,oneknowsthatbindingtheoryhastodowiththeestablishmentofagoverningcategoryandagovernor.Onealsoknowsthatbindingtheoryusestheconceptofc-commandtospecifytheconditionunderwhichagovernorcanbindaNP.Oneotherconceptisneededtoexplainbindingtheory.Thishastodowiththeargumentofabindingcategory.Theterm"argument"impliesthatasemanticrolehasbeenassignedtotheNPunderdiscussion.Thisassignmentiscalledatheta-role.SuchassignmentsarewithinthedomainofThetaTheory.Intheaforementionedexample,theNP(me)isanargumentbecauseithasbeenassignedthethetaroleofObject.ThisassignmentoftheObjectRoletotheNPwasaccomplishedinThetaTheory.Inparticular,theVoftheVPPgovernstheboundNPandassignsittheroleofObject.Inorderforagovernor(V)toassignsucharole,itmustalsobewithinthegoverningcategory(VP)oftheNP(me).Later,casetheorywillassignthisanaccusativeformiftheconditionsareright.Finally,itshouldbenotedthatthesubjectNPandthereflexiveobjectNPareinthesameclause.Theyareclausemates.Theysharethesamelocal\ndomain.Thismeansthattheyhavethesamebindingdomain.Thismakesthereflexivelocallybound. [COMP[Harryjloveshimselfj]VPisthegoverningcategoryV(loves)isthegovernorNPhimself)isassignedathetaroleVc-commandsNPNP(Subject)iscoindexedwithReflexivePronoun(Object)ThecoindexedNPs(theantecedentandtheanaphor)areinthesamebindingdomain.TheanaphoricNPislocallybound.VgovernsandboundsNPThisassignmentundergovernancepreventsthewrongassignmentofthetarolesandcasemarkingstopronouns.Forexample,"*mesawher."Bindingtheoryguaranteesthatonlysentenceswithanaphorsthatareproperlygovernedwillbepredictedasbeinggrammatical.BindingoccurswhentheantecedentNPc-commandstheanaphorandtheantecedentandtheanaphorarecoindexed.Now,itisthetimetoconsiderwhathappenswithinanembeddedsentencewhenthesubjectofthematrixsentenceisnotcoindexedwiththeobjectoftheembeddedsentence.Inthiscase,onefindsbothgovernmentandconstituentcommand,butbecausetheantecedentandtheanaphorarenotcoindexed.Asaconsequence,nobindingoccurs.Thepronominalisfree.Itisfreetorefertoanyindividualwhichisnotreferredtowithinitsgoverningcategory.Itisnotboundbythec-commandingargument.Thissituationcanbefoundintheaforementionedsentence:Marylikeme. S-Structure:COMP[Mary-slikeme]ThesubjectNPcommandstheobjectNPSisthegoverningcategoryofmeV-c-commandsmeMaryandmearenotcoindexedNPsTheanaphorisfree.Itisnotbound.Therefore,nobindingoccurs.Inthefollowingexample,thesubjectofthesentenceisassignedathetarolebyINFL.WhatusedtobecalledModal,Aux,Tense,etc.isnowreferredtoasthecategoryofinflection(INFLorI).JustasNis\ntheheadofNP,VistheheadofVP,PistheheadofPP,itfollowsthatINFListheheadofasentenceorIP(inflectionalphrase).TheINFLinthegoverningcategory(sentence)ismarkedfortense(-s).However,INFLintheembeddedsentenceisnotmarkedfortense.Intraditionalgrammars,tenseisdefinedasfinitetime.Itislimitedtimesuchastimeinthepast,present,orfuture.Whentimeisnotlimited,itbecomesnon-finite.Thisiswheretheterminfinitivecomesfrom.Itmeansnon-finite.Nevertheless,whenINFLisnotmarkedfortenseasevidencedintheembeddedsentence,theverbformemergesasaninfinitivalform.ThenextareaofinvestigationthatmeritsconsiderationistheMinimalistTheory.  \nBibliographyChomsky,Noam.(1993).LecturesonGovernmentBinding:ThePisaLectures.NewYork:MoutondeGruyter.Haegeman,Liliane.(1994).IntroductiontoGovernmentandBindingTheory.SecondEdition.Oxford:BlackwellPublishing.

相关文档