English Possessive s Clitic and Ax John J - PDF document

EnglishPossessive's:CliticandAxJohnJ.LoweUniversityofOxford1.IntroductionThesynchronicsyntacticanalysisofthePresent-DayEnglish(PDE)possessive'shasbeenthesubjectofconsiderabledebate;mostrecentlyBörjarsetal.(2012)withearlierreferences.Somescholars,e.g.Zwicky(1987),Lapointe(1990),Miller(1991),Payne(2009),seektoanalysetheEnglishpossessiveasan`edgeax',i.e.anaxattachednottowordsbuttosyntactic constructionexistingsidebyside.Morphologically`irregular'(i.e.lexicallyspecic)genitiveformsarestillfound,butallornearlyallnounshavetheoptionofusing-(e)s.Atthesametimethephrasalgenitivebecamefullyestablishedasapossibility(thoughitremainedrare).Whatmany(includingAllen)assumewasareanalysisofaxascliticmusthaveoccurredgraduallyoveralongperiodoftime;betweenthe14thcenturyandtheearly17thcenturybothphrasal(i.e.clitic)andcombined(i.e.axal)possessivesexistsidebyside.Evenmoretellingly,thephrasalpossessiveoriginatedincertainsyntacticcontexts,andlaterspread,asdiscussedbyRosenbach(2004).Itoriginatedincontextswherethepossessor`phrase'wasmosteasilyanalyzableasasinglesyntacticunit,namelywithappositionalphrases,coordinatedphrasesandxed,lexicalizedexpressions(esp.ofthetype[TITLEofPLACE]'sX).Itisclear,then,thatwecannottalkaboutanabsolutechangefromaxtoclitic,butmustacknowledgeagradualprocessofchangeovertime.Assuming,therefore,thattheexistenceofthephrasalpossessiveissucienttoadmitaclitic-s,therewasastageinearlyMiddleEnglishwherethepossessiveexistedbothasacliticandasanax.Thatis,therewasatimewhenthephrasal(clitic)possessivecouldbeusedincertainsyntacticcontexts,whileinothercontextsthenon-phrasal(axal)formwasrequired.IfthiswaspossibleinMiddleEnglish,itstandstoreasonthatthesameisatleastpossibleinPDE.3.LFGand`LexicalSharing'TreatingthePDEpossessiveasbothcliticandaxcouldbeconsideredtheoreticallyundesirable.Ifthepossessiveisbothcliticandax,notonlyisitlessclearhowacoherent,uniedanalysiscouldbeundertaken,buttherearealsopracticalproblems,suchashowtopreventacliticfromappearingwhentheaxappears,orviceversa.ByutilizingLexical-FunctionalGrammar(LFG;Bresnan,2001;Dalrymple,2001;Falk,2001),andinparticularthetheoryof`LexicalSharing'developedbyMichaelWescoat(2002,2005,2007,2009)inthisframework,boththetheoreticalandtechnicalproblemsoftreatingthepossessiveasbothcliticandaxcanbeovercome.InthefollowingIadaptWescoat'sLexicalSharingmodelinvariouswaystotmoremainstreamLFGarchitecturalassumptions;spacedoesnotpermitmetoexplaintheseadaptionsinanydetail,butthefundamentalsofWescoat'sapproacharepreserved.LFGassumesaseparatelevelofgrammaticalrepresentationbesidesthehierarchicalconstituentstructure(andbesidesthefunctionalstructure),thes-string(Kaplan,1987),whichessentiallyrepresentsasequenceofwordsasastring,i.e.asetofsyntacticelementslinearlyorderedinaccordancewiththeir`surface'order.Theconstituentstructureisthenprojectedfromthes-string,informaltermsbymeansofaprojection.Usually,oneiteminthes-stringmapstooneiteminthec-structure.Wescoatshows,however,thatcertainsetsofwordsinsomelanguagesarebestanalysedbyassumingthatoneiteminthes-string(Wescoat's`l-structure')correspondsrathertotwonodesinthec-structure.Thisisthecase,forexample,withpronoun-auxiliarycontractionsinPDE,andpreposition-determinercontractionsinFrenchandGerman.Bypermittingaone-to-twocorrespondence,aconsistenthierarchicalrepresentationcanbeobtained,forexample,forallFrenchprepositionalphrases,despitethefactthatinsomeofthesephrasestheprepositionandfollowingdeterminerareclearlyseparatewords(e.g.àla),whileinotherstheyareclearlyasinglelexicalitem(e.g.au,du).(4)PP P  DP D  NP N  auto.thegarçonboy(5)PP P  DP D  NP N  àtolathellegirl4.TheLexicallySharedPossessiveWenowhavethreepossibilitiesforthePDEpossessive,inrelationtowhichtheregularpossessiveformation(1)isinprincipleambiguous:2 (6)Clitic:DP DP D" NP D  NP N  N  Henry'scars(7)LSax:DP DP D" NP D  NP N  N  Henry'scars(8)Ax:DP DP D" NP NP N  N  Henry'scarsWehaveseenthatsomecontextsprovideclearsupportforacliticpossessive,whileothersprovidegoodsupportforanax,suchthatwecanassumethatbothformationsexistsidebysideinPDE.Aphrasalpossessivesuchas(3b)requiresthecliticanalysis(9),sinceotherwisewewouldhavetoassumethateveryniteverbform(etc.)hasanaxal`possessive'formusedonlyinthesecontexts.Sincethephrasalpossessiverequiresaclitic,itismostparsimonioustoassumethecliticconstructionalsoforallregularpossessives,i.e.(6)inpreferenceto(7)or(8).Incontrast,nounswiththeaxal`zero'possessive(2c)utilizeLexicalSharing(10)suchthat,whilethesepossessiveformsaresinglelexicalitems(sincethepossessivemarkerisanax),theirsyntacticparallelismwiththetwoword(nounpluspossessiveclitic)sequencesispreserved,yieldingaconsistentsyntacticanalysisofPDEpossession.(9)SomeoneIknow'sbrother.DP DP D" D" D  NP D  CP N  IP SomeoneIknow'sbrother(10)Thespecies'immunity.DP DP D" D" D  NP D  NP N  N  thespecies'immunitySimilarly,whenthelastwordofaphrasalpossessiveisanounthattakesanaxal`zero'possessive,lexicalsharingisutilizedinpreciselythesameway(11).(11)Thefemaleofthespecies'deadliness.DP DP D" D" D  NP D  NP N  N" N  PP P" P  DP D" D  NP N  thefemaleofthespecies'deadliness3 In(11)thelinearsequenceofnounspeciesfollowedbypossessiveDsurfacesasthelexicallyidiosyncraticpossessiveformofthenoun,species',eventhoughspeciesitselfisnotthepossessornounandhasnodirectsyntacticrelationwiththepossessiveDnode.Wecanassumethatthe`possessive'formofthenoun,species',islistedinthelexicon.However,thisformisnotusedwhenthenounisitselfthepossessor,butratherwhenthenounappearsattheendofapossessorDP,directlypreceding,inlinearterms,theheadDnode,i.e.whenthe`next'nodeinlineartermsistheD0nodeofthesuperordinateDP.ThiscanbespeciedalongthelinesofWescoat'slexicalinstantiationrules,whichinourtermscanbeincludedinthelexicalentryinthefollowingway.(12) species':=NDN(#PRED)=`species'D"POSS Thepartiallexicalentryin(12)statesthatthewordspecies'mapstoasequenceofNfollowedbyDinc-structure,themeaningoftheNis`species',andtheDmusthaveapossessoratf-structure.Thisissucienttoensurethatlexicallyspeciedpossessiveslikespecies'areusedonlywhennalinthepossessorDP(whethertheheadofthatDPornot).SothePDEpossessiveexistsinbothclitic(6)andLSax(7)forms,butthereisnoevidencefora`simple'axal(8)possessive.Spacedoesnotpermitfurtherdiscussion,butitispossibletotreatallthedata,muchofwhichhasnotbeenmentionedhere,inthismodel.5.ConclusionLinguisticcategorizationisrarelyneat,insofaraslinguisticphenomenararelytabsolutelyintotheboxesweconstructforthem.WhilethereissomethingtheoreticallyelegantaboutbeingabletosaythatthePDEpossessiveispurelyaclitic,orpurelyanax,theevidencedoesnotsupporteitherabsoluteanalysis.Ihave,Ihope,shownthatitisinfactpossibletoprovideacoherent,consistentanduniedanalysisofthePDEpossessive,whileatthesametimerecognizingthatitdisplayspropertiesof,andtoanextentis,bothcliticandax.TheLexicalSharingLFGanalysisutilizedherepermitsamorene-grainedrepresentationofsuchdual-naturecategories,makingtheanalysiscloser,perhaps,totherealityofnon-discretelinguisticphenomena.ReferencesAllen,CynthiaL.(1997).`TheOriginsofthe`GroupGenitive'inEnglish'.TransactionsofthePhilologicalSociety95(1),pp.111131.ˆAllen,CynthiaL.(2003).`DeexionandthedevelopmentofthegenitiveinEnglish'.EnglishLanguageandLinguistics7(1),pp.128.ˆAllen,CynthiaL.(2008).GenitivesinEarlyEn-glish:TypologyandEvidence.OxfordUniversityPress.ˆAnderson,StephenR.(2008).`TheEnglishGroupGenitiveisaSpecialClitic'.EnglishLinguistics25,pp.120.ˆBörjars,Kersti,DavidDenison,andAlanScott(eds.)(2012).MorphosyntacticCategoriesandtheExpressionofPossession.Benjamins.ˆBresnan,Joan(2001).Lexical-FunctionalSyntax.BlackwellPublishing.ˆDalrymple,Mary(2001).LexicalFunctionalGram-mar.AcademicPress.ˆFalk,YehudaN.(2001).Lexical-FunctionalGrammar:anIntroductiontoParallelConstraint-BasedSyntax.CSLIPublications.ˆKaplan,RonaldM.(1987).`ThreeSeductionsofComputationalPsycholinguistics'.InWhitelocketal.(eds.),LinguisticTheoryandComputerApplications,AcademicPress,pp.149181.ˆLapointe,StevenG.(1990).`EdgefeaturesinGPSG'.InZiolkowskietal.(eds.),Papersfromthe26thRegionalMeetingoftheChicagoLinguisticSociety,vol.1:TheMainSession,ChicagoLinguisticSociety,pp.221235.ˆMiller,PhilipH.(1991).Cliticsandconstituentsinphrase-structuregrammar.Ph.D.thesis,RijksuniversiteitteUtrecht.ˆPayne,John(2009).`TheEnglishgenitiveanddoublecase'.TransactionsofthePhilologicalSociety107(3),pp.322357.ˆQuirk,Randolf,SidneyGreenbaum,GeoffreyLeech,andJanSvartvik(1985).AComprehensiveGrammaroftheEnglishLanguage.Longman.ˆRosenbach,Anette(2004).`TheEnglishs-genitive:Acaseofdegrammaticalization?'InFischeretal.(eds.),UpandDowntheCline:TheNatureofGrammaticalization,Benjamins,pp.7396.ˆWescoat,MichaelThomas(2002).`OnLexicalSharing'.Ph.D.Thesis,StanfordUniversity.ˆWescoat,MichaelThomas(2005).`EnglishNonsyllabicAux-iliaryContractions:AnAnalysisinLFGwithLexicalSharing'.InOnlineProceedingsoftheLFG05Conference,CSLIPublications,pp.468486.ˆWescoat,MichaelThomas(2007).`Preposition-DeterminerContractions:AnAnalysisinOptimality-TheoreticLexical-FunctionalGrammarwithLexicalSharing'.InOnlineProceedingsoftheLFG07Conference,CSLIPublications,pp.439459.ˆWescoat,MichaelThomas(2009).`UdiPersonMarkersandLexicalIntegrity'.InOnlineProceedingsoftheLFG09Conference,CSLIPublications,pp.604622.ˆZwicky,ArnoldM.(1987).`SuppressingtheZs'.JournalofLinguistics23,pp.133148.4