下载

1下载券

加入VIP
  • 专属下载特权
  • 现金文档折扣购买
  • VIP免费专区
  • 千万文档免费下载

上传资料

关闭

关闭

关闭

封号提示

内容

首页 句法结构_乔姆斯基

句法结构_乔姆斯基.doc

句法结构_乔姆斯基

baicai229
2012-12-28 0人阅读 举报 0 0 暂无简介

简介:本文档为《句法结构_乔姆斯基doc》,可适用于高等教育领域

ANELEMENTARYLINGUISTICTHEORY一个初级的语言理论AAssumingthesetofgrammaticalsentencesofEnglishtobegiven,wenowaskwhatsortofdevicecanproducethisset(equivalently,whatsortoftheorygivesanadequateaccountofthestructureofthissetofutterances)WecanthinkofeachsentenceofthissetasasequenceofphonemesoffinitelengthAlanguageisenormouslyinvolvedsystem,anditisquiteobviousthatanyattempttopresentdirectlythesetofgrammaticalphonemesequenceswouldleadtoagrammarsocomplexthatitwouldbepracticallyuseless一种语言是一个极其庞杂的系统。如果想要把这个语言中符合语法的音位序列的集合直接描述出来这种语法将会复杂而没有实际用处。Forthisreason,linguisticdescriptionproceedsintermsofasystemof“levelsofrepresentations”(多平面描述代表平面)Thelinguistsetsupsuch“higherlevel”elementsasmorphemes,andstateseparatelythemorphemicstructureofsentencesandthephonemicstructureofmorphemesWeaskwhatsortofgrammarisnecessarytogenerateallthesequencesofmorphemesthatconstitutegrammaticalEnglishsentences,andonlytheseOnerequirementthatagrammarmustcertainlymeetisthatitbefiniteHencethegrammarcannotsimplybealistofallmorphemesequences,sincethereareinfinitelymanyoftheseAcommunicationtheoreticmodel(通讯理论模型)forlanguageTHEFINITESTATEGRAMMAR(有限状态语法)Supposethatwehaveamachinethatcanbeinanyoneofafinitenumberofdifferentinternalstates,andsupposethatthismachineswitchesformonestatetoanotherbyproducingacertainsymbol(letussay,anEnglishword)Oneofthesestatesinaninitialstate(起始状态)anotherisafinialstate(终端状态)Thenwecallthesequenceofwordsthathasbeenproduceda“sentence”StateDiagramEachnodeinsuchadiagramcorrespondtoastateofthemachineEachsuchmachinethusdefinesacertainlanguagenamely,thesetofsentencethatcanbeproducedinthiswayFinitestatelanguage(有限状态语言)Byaddingclosedloops(封闭圈)wecanextendthismachinetoproduceaninfinitenumberofsentenceGivenastatediagram,wecanproduceasentencebytracingapathfromtheinitialpointonthelefttothefinialpointontheright,alwaysproceedinginthedirectionofthearrowsWecanallowtransitionformonestatetoanotherinseveralways,andwehaveantnumberofclosedloopsofanylengthMarkovProcesses马尔科夫过程ThisconceptionoflanguageisanextremelypowerfulandgeneraloneInproducingasentence,thespeakerbeginsintheinitialstate,producesthefirstwordofthesentence,therebyswitchingintoasecondstatewhichlimitsthechoiceofthesecondwordEachstatethroughwhichhepassesrepresentsthegrammaticalrestrictionthatlimitthechoiceofthenextwordatthispointintheutteranceBAnyattempttoconstructafinitestateforEnglishrunsintoseriousdifficultiesandcomplicationsattheveryoutsetHowever,itisunnecessarytoattempttochowthisbyexample,inviewofthefollowingmoregeneralremarkaboutEnglish:Englishisnotafinitestatelanguage(英语不是有限状态语言)Itisimpossibletostatethemorphemicstructure(语素结构)ofsentencesdirectlybymeansofsuchdevicesasastatediagram,andthattheMarkovprocessconceptionoflanguagecannotbeaccepted,atleastforthepurposeofgrammarBeforeinvestigatingEnglishdirectly,letusconsiderseverallanguageswhosealphabets(字母)containjustthelettersa,b,andwhosesentencesareasdefinedin:()ab,aabb,aaabbb,…,andingeneral,allsentencesconsistingofnoccurrencesofafollowedbynoccurrencesofb()aa,bb,abba,baab,aaaa,bbbb,aabbaa,abbbba,…,ingeneral,allsentencesconsistingofastringxfollowedbythe‘mirrorimage(镜像)’ofx()aa,bb,abab,baba,aaaa,bbbb,aabaab,abbabb,…,andingeneral,allsentencesconsistingofastingxofa’sandb’sfollowedbytheidenticalstringxWecaneasilyshowthateachofthesetreelanguagesisnotafinitestatelanguageButitisclearthattherearesuchsubpartsofEnglishwiththebasicformof()and()LetS,S,S,…bedeclarativesentences(陈述句)inEnglishThenwecanhaveEnglishsentencesas:()IfS,thenS()EitherS,orS()ThemanwhosaidS,isarrivingtoday在()中不能用or代替then在()中不能用then代替or在()中不能用are代替is上述各句中逗号两侧的词之间都存在着相互依存的关系。if…theneither…orman…is但是在相互依存的词之间我们都可以嵌入一个陈述句S,S,S在()中我们将()作为S将()作为S我们可以得到这样一个句子()if,either(),orS,thenS在()中的S可以又是()、()、()中的任何一个句子……Asetofsentencesthatisconstructedinthiswaywillhaveallofthemirrorimageproperties(全部的镜像性质)of()whichexclude()fromthesetoffinitestatelanguageThuswecanfindvariouskindsofnonfinitestatemodelswithinEnglishCTheconceptionofgrammarwhichhasjustrejectedrepresentsinawaytheminimallinguistictheory(最小限度的语言理论)thatmeritsseriousconsiderationAfinitestategrammaristhesimplesttypeofgrammarwhich,withafiniteamountofapparatus,cangenerateaninfinitenumberofsentencesWeareforcedtosearchforsomeabstractformoflinguistictheoryPhrasestructure短语结构ACustomarily,linguisticdescriptiononthesyntacticlevel(句法平面上的语言描写)isformulatedintermsofconstituentanalysis(成分分析)。Wenowaskwhatformofgrammarispresupposebydescriptionofthissort(要进行这样的描写要预先假定出哪一种形式的语法)。Asasimpleexampleofthenewformforgrammarsassociatedwithconstituentanalysis,considerthefollowing:()iSentence→NPVPiiNP→T(指示词)NPiiiVP→VerbNPivT→thevN→man,ball,ectviVerb→hit,took,ectSupposethatweinterpreteachruleX→Yof()astheinstruction“rewriteXasY”OnlyonesingleelementcanbewritteninanysingleruleWeshallcall()aderivation(推导式)ofthesentence“themanhittheball”,wherethenumberattherightofeachlineofthederivationrefertotheruleofthegrammar()usedinconstructingthatlineformtheproceedingline()Wecanrepresentthederivationinanobviouswaybymeansofthefollowingdiagram:()Thediagram()conveyslessinformationthanthederivation(),sinceitdoesnottellusinwhatordertheruleswereappliedGiven(),wecanconstruct()uniquely,butnottheviceversa,sinceitispossibletoconstructaderivationthatreducesto()withadifferentorderofapplicationoftherulesThediagram()retainsjustwhatisessentialin()forthedeterminationofthephrasestructureofthederivedsentence(导出句)“themanhittheball”Asequenceofwords(语词序列)ofthissentenceisaconstituentoftypeZ(Z型成分)ifwecantracethissequencebacktoasinglepointoforiginin(),andthispointoforiginislabeledZThus“hittheball”canbetracedbacktoVPin(),hence‘hittheball”isaVPinthederivedsentencesWesaythattowderivationareequivalenttheyreducetothesamediagramoftheform()Occasionallyagrammarmaypermitustoconstructnonequivalentderivations(不等值推导)ofagivensentenceUnderthesecircumstances,wesaythatwehaveacaseof“constructionalhomonymity(同音结构)”andifourgrammariscorrect,thissentenceofthelanguageshouldbeambiguousOnegeneralizationof()isclearlynecessaryWemustbeabletolimitapplicationofaruletoacertaincontext(语境):()ZXW→ZYW(contextZW)()NPsingVerb→NPsinghitsIndicatingtheVerbisrewrittenhitsonlyinthecontextNPsingBWecannowdescribemoregenerallytheformofgrammarassociatedwiththetheoryoflinguisticstructurebaseduponconstituentanalysisEachsuchgrammarisdefinedbyafinitesetof∑ofinitialstrings(初始符号链的有限集合)andafinitesetFof“instructionformulas(指令公式)”oftheformX→Yinterpreted“rewriteXasY”Inthegrammar(),theonlymemberoftheset∑ofinitialstringwasthesinglesymbolSentence,andFconsistedoftherule(i)(vi)butwemightwanttoextend∑toinclude,forexample,DeclarativeSentenceInterrogativeSentenceasadditionalsymbolsGiventhegrammar∑,F,wedefineaderivationasafinitesequenceofstrings,beginningwithaninitialstringsof∑,andwitheachstringinthesequencebeingderivedfromtheprecedingstringbyapplicationofoneoftheinstructionformulasofFCertainderivationsareterminatedderivation(终止推导式)inthesensethattheirfinalstringcannotberewrittenanyfurtherbytherulesofFIfastringisthelastlineofaterminatedderivation,wesaythatitisaterminalstring(终端语符链)Thusthemanhittheballisaterminalstringformthegrammar()Asetofstringsisaterminallanguageifitisthesetofterminalstringsforsomegrammar∑,FGivenaterminallanguageanditsgrammar,wecanreconstructthephrasestructureofeachsentenceofthelanguagebyconsideringtheassociateddiagramsoftheform()WecanalsodefinethegrammaticalrelationsintheselanguageinaformalwayintermsoftheassociateddiagramCFinitestatelanguagesTerminallanguages:Theorem(定理):Everyfinitestatelanguageisaterminallanguage,butthereareterminallanguageswhicharenotfinitestatelanguagesTheimportofthistheoremisthatdescriptionintermsofphrasestructureisessentiallymorepowerfulthandescriptionintermsoftheelementarytheorypresentedaboveAsexamplesofterminallanguagesthatarenotfinitestatelanguageswehavethelanguage(),()Thusthelanguage(),consistingofallandonlythestringsab,aabb,aaabbb,…canbeproducedbythe∑,Fgrammar:()∑:ZF:Z→abZ→aZbThisgrammarhastheinitialstringZandithastworulesItcaneasilybeseenthateachterminatedderivation(终止推导式)constructedfrom()endsinastringofthelanguage(),andthatallsuchstringsareproducedinthiswayWepointedoutthatthelanguages(),()correspondtosubpartofEnglish,andthatthereforethefinitestateMarkovprocessmodelisnotadequateforEnglishWenowseethatthephrasestructuremodeldoesnotfailinsuchcasesNotethatinthecaseof(),wecansaythatinthestringof(),abisaZ,aabbisaZ,andaaabbbitselfisaZItisimportanttoobservethatindescribingthislanguagewehaveintroducedasymbolZwhichisnotcontainedinthesentencesofthislanguageThisistheessentialfactaboutphrasestructurewhichgivesitits‘abstract’characterDSupposethatbya∑,FgrammarwecangenerateallofthegrammaticalsequencesofmorphemesofalanguageInordertocompletethegrammarwemuststatethephonemicstructureofthesemorphemes,sothatthegrammarwillproducethegrammaticalphonemesequencesofthelanguageMorphophonemics(语素音位)ofthelanguagecanalsobegivenbyasetofrules()Notethatordermustbedefinedamongtheseruleseg,(ii)mustprecede(v)or(vii),orwesillderivesuchformasteyktforthepasttenseoftakeInthesemorphophonemicrulesweneednolongerrequirethatonlyasinglesymbolberewrittenineachruleWecanextendthephrasestructurederivationbyapplying(),sothatwehaveaunifiedprocessforgeneratingphonemesequencefortheinitialstringSentenceItcaneasilybeseenthatitwouldbequiteadvantageoustoordertherulesofthesetFsotatcertainoftherulescanapplyonlyafterothershaveapplyForexample,weshouldcertainlywanttherule()NPsingVerb→NPsinghitstoapplybeforeanyrulewhichenableustorewriteNPasNPPreprsitonNPOtherwisethegrammarwillproducesuchnonsentencesas“themennearthetruckbeginsworkateight”LIMITATIONSOFPHRASESTRUCTUREDESCRIPTION短语结构描写的局限AIthinktherearegroundsforrejectingthetheoryofphrasestructureasinadequateforthepurposeoflinguisticdescriptionThestrongestpossibleproofofinadequacyofalinguistictheoryistoshowthatitliterallycannotapplytosomenaturallanguageAweaker,butperfectlysufficientdemonstrationofinadequacywouldbetoshowthatthetheorycanapplyonlyclumsilythatis,toshowthatanygrammarthatcanbeconstructedintermsofthistheorywillextremelycomplex,and“unrevealing”,thatcertainverysimplewaysofdescribinggrammaticalsentencescannotbeaccommodatedwithintheassociatedformsofgrammar,andthatcertainfundamentalformalpropertiesofnaturallanguagecannotutilizedtosimplifygrammarsBOneofthemostproductiveprocessesforformingnewsentencesistheprocessofconjunctionIfwehavetwosentencesZXYandZYW,andifXandYareactuallyconstituentsofthesesentences,wecangenerallyformanewsentenceZXandYW()thesceneofthemoviewasinChicagothesceneoftheplaywasinChicago()thesceneofthemovieandplaywasinChicagoIfXandYarenotconstituents,wegenerallycannotdothis:()thelinersaileddowntheriverthetugboatchuggeduptheriver()thelinersaileddowntheandthetugboatchuggeduptheriverSimilarly,ifXandYarebothconstituents,butareconstituentsofdifferentkinds(inthediagram()theyeachhaveasingleorigin,butthisoriginislabeleddifferently),thatwecannotingeneralformanewsentencebyconjunction:()thesceneofthemoviewasinChicagothescenethatIwrotewasinChicago()thesceneofthemovieandthatIwrotewasinChicagoInfact,thepossibilityoftheconjunctionoffersoneofthebestcriteriafortheinitialdeterminationofphrasestructureWecansimplifythedescriptionofconjunctionifwetrytosetupconstituentsinsuchaway:()IfSandSaregrammaticalsentences,andSdiffersformSonlyinthatXappearsinSwhereYappearsinS,andXandYareconstituentsofthesametypeinSandS,respectively,thenSisasentence,whereSistheresultofreplacingXbyXYinSButwenowfacethefollowingdifficulty:wecannotincorporateintherule()oranythinglikeitinagrammar∑,Fofthephrasestructure,becauseofcertainfundamentallimitationsonsuchgrammarTheessentialpropertyofrule()isthatinordertoapplyittothesentencesSandStoformanewsentenceS,wemustknownotonlytheactualformofSandSbutalsotheirconstituentstructure(成分结构)wemustknownotonlythefinalshapeofthesesentences,butalsotheir“historyofderivation(推导史)”ButeachX→Yofthegrammar∑,FappliesorfailstoapplytoagivenstringbyvirtueoftheactualsubstanceofthisstringIfthestringcontainsXasasubstring,theruleX→Ycanapplytoitifnot,therulecannotapplyRule()isalsofundamentallynewinadifferentsenseItmakesessentialreferencetotwodistinctsentences,butingrammarofthe∑,Ftype,thereisnowaytoincorporatesuchdoublereferenceBInthegrammar()wegraveonlyonewayofanalyzingtheelementVerbButevenwiththeverbalrootfixed(动词词根是确定的),therearemanyotherformsthatthiselementcanassume(承担),eg,takes,hastaken,willtakehasbeentakenectThestudyofthere“auxiliaryverb(助动词)”turnsouttobequitecrucialinthedevelopmentofEnglishgrammarWeshallseethattheirbehaviorisveryregularandsimplydescribablewhenobservedformapointofviewthatisquitedifferentformthatdevelopedaboveWecanstatetheoccurrenceoftheseauxiliariesindeclarativesentencesbyaddingtothegrammar()thefollowingrules:()(i)Verb→AuxV(ii)V→hit,take,walk,readect(iii)Aux→C(M)(haveen)(being)(been)C里面的成分是必须用到的至于括号里的成分则可用可不用但必须按照规则制定的顺序排列。(iv)M→will,can,may,shall,must(modal)()(i)C→SinthecontextNPsing(单数)(inthecontextNPpl(复数)Past(过去式)(ii)LetAfstandforanyoftheaffixes(词缀)past,S,(,en,ingLetvstandforanyMorV,orhaveorbe(动词结构中的非词缀成分)Then:Afv→vAf#(#词界)(iii)Replaceby#exceptinthecontextvAfinsert#initiallyandfinally(除去在vAf的上下文中都以#号代替号把它加在开头和结尾的地方。)现在写出一个省略了前面几步的推导式(derivation)()themanVerbthebookthemanAuxVthebook(i)themanAuxreadthebook(ii)themanChaveenbeingreadthebook(iii,在这里我们选取了成分C,haveen,being)themanShaveenbeingreadthebook(i)themanhaveS#been#reading#thebook(运用ii三次)#the#man#haveS#been#reading#the#book(iii)运用之前提到的语素音位规则我们就可以将推导式的最后一列转化为()themanhasbeenreadingthebook(phonemicdescription)Noticethatinordertoapply(i)inwehadtousethefactthatthemanisasingularnounphraseThatis,wehadtoreferbacktosomeearlierstepinthederivationinordertodeterminetheconstituentstructureofthethemanHence(i),justlike(),goesbeyondtheelementaryMarkoviancharacterofgrammarsofphrasestructure,andcannotbeincorporatedwithinthe∑,FgrammarRule(ii)wiolatetherequirementsof∑,Fgrammarsevenmoreseverelyitalsorequiresreferencetoconstituentstructureandinaddition,wehavenowaytoexpresstherequiredinversionwiththetermsofphrasestructureThereadercaneasilydeterminethattoduplicatetheeffectof(iii)and()withoutgoingbeyondtheboundofasystem∑,Fofphrasestructure,itwouldbenecessarytogiveafairlycomplexstatementIntheauxiliaryverbwereallyhavediscontinuouselements(不连续成分)butdiscontinuitiescannotbehandledwith∑,FgrammarsIn(iii)wetreatedtheseelementsascontinuous,andweintroducedthediscontinuitybytheverysimpleadditionalrule(ii)CConsidertheactivepassiverelation:Passivesentencesareformedbyselectingtheelementbeeninrule(iii)Butthereareheavyrestrictionsonthiselementoftheauxiliaryphrase(及物动词及物动词后不能再接名词性成分*lunchiseatenJone前置词短语byNP共现时强制选择)ButthiswholenetworkofrestrictionsfailscompletelywhenwechoosebeenasapartoftheauxiliaryverbInfact,inthiscasethesomeselectionaldependencieshold,butintheoppositeorder(同样限制的依赖条件依然存在只不过换了一个方向)Thatis,foreverysentenceNPVNP,wecanhaveacorrespondingsentenceNPisVenbyNP()IfSisagrammaticalsentenceoftheformNPAuxVNPthenthecorrespondingstringoftheformNPAuxbeenVbyNPisalsoagrammaticalsentenceWecannowdroptheelementbeen,andallofthespecialrestrictionsassociatedwithit,form(iii)But()iswellbeyondthelimitof∑,Fgrammars,itrequiresreferencetotheconstituentstructureofthestringtowhichitappliesanditcarriesoutainversiononthisstringinastructurallydeterminedmannerDWehavediscussedthreeruleswhichmateriallysimplifythedescriptionofEnglishbutwhichcannotincorporatedintoa∑,FgrammarIfweexaminecarefullytheimplicationsofthesesupplementaryrules,weseethattheyleadtoanentirelynewconceptionoflinguisticstructuregrammaticaltransformation(语法转换)AgrammaticaltransformationoperatesonagivenstringwithagivenconstituentstructureandconvertsitintoanewstringwithanewderivedconstituentstructureItisclearthatwemustdefineanorderofapplicationonthesetransformationsCertaintransformationsareobligatory(强制的)whereasothersareonlyoptional(可选的)。(ii)mustbeappliedtoeverysentence,but()mayormaynotbeappliedinanyparticularcaseThedistinctionbetweenobligatoryandoptionaltransformationsleadsustosetupafundamentaldistinctionamongthesentencesofthelanguageSupposethatwehaveagrammarGwitha∑,Fparta

用户评价(1)

  • hover0823 只有有一部分,并不是全文

    2013-08-04 17:52:55

关闭

新课改视野下建构高中语文教学实验成果报告(32KB)

抱歉,积分不足下载失败,请稍后再试!

提示

试读已结束,如需要继续阅读或者下载,敬请购买!

文档小程序码

使用微信“扫一扫”扫码寻找文档

1

打开微信

2

扫描小程序码

3

发布寻找信息

4

等待寻找结果

我知道了
评分:

/22

句法结构_乔姆斯基

VIP

在线
客服

免费
邮箱

爱问共享资料服务号

扫描关注领取更多福利