Presupposition - Stanford Encyclopedia of Philosophy

We discuss presupposition, the phenomenon whereby speakers mark linguistically information as being taken for granted, rather than being ... StanfordEncyclopediaofPhilosophy Menu Browse TableofContents What'sNew RandomEntry Chronological Archives About EditorialInformation AbouttheSEP EditorialBoard HowtoCitetheSEP SpecialCharacters AdvancedTools Contact SupportSEP SupporttheSEP PDFsforSEPFriends MakeaDonation SEPIAforLibraries EntryNavigation EntryContents Bibliography AcademicTools FriendsPDFPreview AuthorandCitationInfo BacktoTop PresuppositionFirstpublishedFriApr1,2011;substantiverevisionThuJan7,2021 Wediscusspresupposition,thephenomenonwherebyspeakersmark linguisticallyinformationasbeingtakenforgranted,ratherthan beingpartofthemainpropositionalcontentofaspeechact. Expressionsandconstructionscarryingpresuppositionsarecalled “presuppositiontriggers”,formingalargeclassincluding definitesandfactiveverbs.Thearticlefirstintroducesasampleof triggers,thebasicpropertiesofpresuppositionssuchasprojection andcancellability,andthediagnostictestsusedtoidentifythem. Thereaderisthenintroducedtomajormodelsofpresuppositionfrom thelast50years,separatedintothreeclasses:Frege-Strawson derivedsemanticmodels,pragmaticmodelssuchasthatofferedby Stalnaker,anddynamicmodels.Finallywediscusssomeofthemain currentissuesinpresuppositiontheory.Theseinvolveaccommodation, whichoccurswhenahearer’sknowledgestateisadjustedtomeetthe speaker’spresuppositions;presuppositionfailure,whichoccurswhena presuppositionis(knowntobe)false;theinteractionbetween presuppositionsandattitudes;andvariabilityinthebehaviorof triggersandtheirpresuppositions. 1.CharacterizingPresupposition 1.1Introduction 1.2Projection 1.3Cancellability 2.TheFrege-Strawsontradition 3.Pragmaticpresupposition 4.Localcontextsandthedynamicturn 4.1Presuppositionandanaphora 4.2Satisfactioninlocalcontexts 4.3Satisfactiontheories 5.Accommodation 5.1GlobalandLocalAccommodation 5.2AccommodatingPresuppositionsinSatisfactionFramework 5.3ResolvingandAccommodatingPresuppositionsinDRT 6.Borntofail,orunborn? 7.Presuppositionsandattitudes 8.Presuppositionvariability Bibliography AcademicTools OtherInternetResources RelatedEntries 1CharacterizingPresupposition 1.1Introduction Speakerstakealotforgranted.Thatis,theypresuppose information.Aswewrotethis,wepresupposedthatreaderswould understandEnglish.Wealsopresupposedaswewrotethelastsentence, repeatedin(1),thattherewasatimewhenwewroteit,forotherwise thefrontedphrase“aswewrotethis”wouldnothave identifiedatimeinterval. (1) Aswewrotethis,wepresupposedthatreaderswouldunderstand English. Further,wepresupposedthatthesentencewasjointlyauthored,for otherwise“we”wouldnothavereferred.Andwepresupposed thatreaderswouldbeabletoidentifythereferenceof “this”,i.e.,thearticleitself.Andwepresupposedthat therewouldbeatleasttworeaders,forotherwisethebareplural “readers”wouldhavebeeninappropriate.Andsoon. Notethatsomeofthesepresuppositionsarisebydefaultfromspecific wordsthatweused.Theexistenceofatimewhenwewrotethearticle isarequirementassociatedwithouruseof“as”.Itisa requirementbuiltintothemeaningofthetemporalpreposition “as”,whichhasasimilarmeaningtotemporal"while", thatinaphrase“asX”,the “X”hastoholdatsometime.Wesaythat “as”isapresuppositiontrigger.Similarly, “this”isapresuppositiontriggerrequiringsomething salienttoreferto,thebarepluralisapresuppositiontrigger requiringexistenceofmultipleindividuals,and“would” isapresuppositiontriggerrequiringasalientfutureorhypothetical circumstance. Incontrast,someofthepresuppositionsabovehavenothingtodowith themeaningsofanyofthosewords.Forexample,wecansaythatthe presuppositionthattheaddresseespeaksEnglish,likethe presuppositionthattheaddresseeisinterestedinwhatthespeaker (orwriter)hastosay,isaconversationalpresupposition or,followingStalnaker(1972;1974),speakerpresupposition orpragmaticpresupposition.Thepresuppositionsassociated withspecifictriggersaresaidtobeconventionalor semantic.Theterminologicaldistinctionbetweensemanticand pragmaticpresuppositionisoftheoreticalimport:aswewillsee later,sometheoristsregarditasanopenquestionwhetherthereare anypurelyconventionalpresuppositions.Ahalfwayhouse,suggested forexamplebyKarttunen(1973)andSoames(1982),istodefinea notionofutterancepresupposition,thusinvolvingaspecific formthatisuttered,butallowingthatwhatisactuallypresupposed maydependalsoontheattitudesofthespeakerwhouttersit. Itisimportanttonotethattocallpresuppositionalexpressions “conventional”or“semantic”isnot necessarilytoimplythatthepresuppositionstheytriggerdon’t dependonthecontextinanyway.Forexample,although “this”maybeviewedasaconventionalpresupposition trigger,theinterpretationverymuchdependsonthecontext,andthe presupposition,althoughtypicallyregardedasconventional,is normallyseenpreciselyasaconstraintontheutterancecontext. Whatmakespresuppositionsspecial?Thatis,totheextentthat presuppositionsarejustapartoftheconventionalmeaningofsome expressions,whatmakesthemsufficientlydistinctivethattheymerit theirownentriesinhandbooksandencyclopedias,aswellasmany hundredsofotherarticlesandbookchapterselsewhere?First, presuppositionsareubiquitous.Andsecond,therearevariousrespects inwhichthebehaviorofpresuppositionsdifferssharplyfromother aspectsofmeaning. Asregardstheubiquityofpresuppositions,atleastthefollowing lexicalclassesandconstructionsarewidelyagreedtobe presuppositiontriggers: factives(KiparskyandKiparsky,1970) BerlusconiknowsthatheissigningtheendofBerlusconism. →BerlusconiissigningtheendofBerlusconism. aspectualverbs(“stop,continue”)(Simons, 2001;Abusch,2002;Lorenz,1992) Chinahasstoppedstockpilingmetals. →Chinausedtostockpilemetals. temporalclausesheadedby“before”, “after”,“since”,etc.(Beaverand Condoravdi,2003;Heinämäki,1974) Thedudereleasedthisvideobeforehewentonakillingspree. →Thedudewentonakillingspree. manneradverbs(Abbott,2000) Jamieduckedquicklybehindthewall. →Jamieduckedbehindthewall. sortallyrestrictedpredicatesofvariouscategories(e.g., “bachelor”)(Thomason,1972) Juliusisabachelor. →Juliusisanadultmale. cleftsentences(Delin,1995;Prince,1986) ItwasJesuswhosetmefree. →Somebodysetmefree. quantifiers(Roberts,1995;Gawron,1995;Abuschand Rooth,2000;Cooper,1983) IhavewrittentoeveryheadmasterinRochdale. →ThereareheadmastersinRochdale. definitedescriptions(Strawson,1950) ThePrimeMinisterofTrinidadandTobagostoodupandwaggedhis finger. →TrinidadandTobagohavea(unique)primeminister. names(vanderSandt,1992) TheauthorisJuliusSeidensticker. →JuliusSeidenstickerexists. intonation(e.g.,focus,contrast)(Jackendoff,1972; GeurtsandvanderSandt,2004;Roberts,1998) HEsetme free.[1] →Somebodysetmefree. Andthisisonlyasmallsampleofthewordsandsyntactic constructionsthathavebeenclassifiedaspresuppositiontriggers,so evenifinsomecasestheremaybedoubtsaboutthisdiagnosis,itcan hardlybedoubtedthatpresuppositiontriggersaboundineveryday language.Inthefollowingsectionswewilldiscussthebehaviors whichmarkoutpresuppositionsfromordinaryentailments,andthen introducesomeofthetheoriesthathavebeendevelopedtoaccountfor thosebehaviors. 1.2Projection Thehallmarkofpresuppositions,aswellasthemostthoroughly studiedpresuppositionalphenomenon,isprojection (LangendoenandSavin,1971).Consider(2).Thishasallthe presuppositionsin(3).Thesepresuppositionsallfollowfrom utterancesofthebasesentencein(2),asdotheregularentailments in(4):someonewhosincerelyuttered(2)wouldcertainlybeexpected toacceptthetruthof(3)and(4),aswell: (2) It’stheknavethatstolethetarts. (3a) Thereisa(salientandidentifiable)knave. (3b) Therewere(salientandidentifiable)tarts. (3c) Somebodystolethetarts. (4a) Theknavedidsomethingillegal. (4b) Theknavetookpossessionofthetarts. Nowconsiderthesentencesin(5): (5a) Itisn’ttheknavethatstolethetarts.(negation) (5b) Ifit’stheknavethatstolethetarts,hewillbepunished. (antecedentofaconditional) (5c) Isittheknavethatstolethetarts?(question) (5d) Maybe/Itispossiblethatit’stheknavethatstolethetarts. (possibilitymodal) (5e) Presumably/probablyit’stheknavethatstolethetarts. (evidentialmodal,probabilityadverb) (5f) Thekingthinksit’stheknavethatstolethetarts.(belief operator) Inalltheseexamples,sentence(2)isembeddedundervarious operators.Whatisnotableisthatwhereasthestatementsin(4)do notfollowfromanyoftheseembeddings(andwouldnotbeexpectedto followaccordingtoclassicallogics),thepresuppositionsdofollow. Wesaythatthepresuppositionsareprojected.Certainly,the inferenceismorerobustinsomecasesthaninothers:whileitis hardtoimaginesincerelyuttering(5a)withoutbelievingsometarts tobesalient,itiseasiertoimagineacircumstanceinwhich(5f) couldbeutteredwheninfactthetartswerenotstolen,buthidden. Butintheabsenceofspecialfactors,towhichwewillturnshortly, someonewhosincerelyutteredanyofthesentencesin(5)mightbe expectedtobelieveallofthepresuppositionsin(3a)–(3b). Projectionfromembeddings,especiallynegation,isstandardlyusedas adiagnosticforpresupposition(hencetheterm“negation test”).Howeveritisimportanttotryseveraltypesof embeddingwhentestingforpresuppositionforafewdifferentreasons. Firstofall,itisnotalwaysclearhowtoapplyagivenembedding diagnostic.Forexample,althoughitiswidelyagreedthatthe additiveparticle“too”isapresupposition-inducing expression,thenegationtestisawkwardtoapplybecauseitisnot immediatelyclearwhetherthenegationoutscopestheadditiveor viceversain(6b),andit’snotobvioushowthenegation interactswithfocus: (6a) FredkissedBETTY,too. (6b) Freddidn’tkissBETTY,too. Weleaveitasanexerciseforthereadertoshowthatembedding(6a) underadifferentoperatore.g.,underamodalorintheantecedentof aconditional,providesevidencethatthissentencepresupposesthat someoneotherthanBettywaskissedbyFred. Additionally,someinferencesseemtoprojectfromnegationbutnot otherembeddings.Forexample,both(7a)and(7b)seemtopresuppose thatthespeakerhassensoryexperiencewiththepie,andthistypeof datahasledsometoconjecturethatusesofpredicatesofpersonal tasteareassociatedwithapresuppositionthatthejudgmentoftaste resultedfromdirect experience.[2] However,wedon’tgetanyprojectionofthisinferencefromthe conditionalin(7c),andexample(7d)seemstoimplythatthespeaker hasnottastedthepiethemselves.Thisdatasuggeststhatanyaccount claimingapresuppositionofdirectexperienceforpredicatesof personaltasteneedstobehighlynuanced. (7a) Thepieistasty. (7b) Thepieisn’ttasty. (7c) Ifthepieistasty,itwillgoquickly. (7d) Isthepietasty? Moregenerally,suchexamplesdemonstratewhyit’simportanttolook atawiderangeofembeddingswhenidentifyingpresuppositionsonthe basisoftheirprojectionbehavior. 1.3Cancellability Whatmakesthe“projectionproblem”problematic?Ifsome partofthemeaningofanexpressionαwasneveraffectedbythe linguisticcontextinwhichαwasembedded,thatwouldbe philosophicallyinteresting,andwoulddemandatheoretical explanation,butitwouldatleastbetrivialtocompletelydescribe thedata:allpresuppositionalinferenceswouldsurviveanyembedding, endofstory.Butthatisn’twhathappens.Presuppositionstypically project,butoftendonot,andmostoftheempiricalandtheoretical workonpresuppositionsincethe1970shasbeentakenupwiththetask ofdescribingandexplainingwhenpresuppositionsproject,andwhen theydon’t. Whenapresuppositiondoesnotproject,itissometimessaidtobe “cancelled”.Theclassiccasesofcancellationoccurwhen thepresuppositionisdirectlydenied,asinthefollowingvariantsof someofthesentencesin(5): (8a) Inthiscourt,itisn’ttheknavethatstealsthetarts:theking employsnoknavespreciselybecausehesuspectstheyareresponsible forlarge-scaletart-lossacrosshiskingdom. (8b) Ifit’stheknavethatstolethetarts,thenI’maDutchman:there isnoknavehere. (8c) Isittheknavethatstolethetarts?Certainlynot:thereisno knavehere. (8d) Thekingthinksit’stheknavethatstolethetarts,buthe’s obviouslygonemad,sincethereisnoknavehere. Presuppositionalinferencesaretypicallysubjecttocancellationby directdenialonlywhenthepresuppositiontriggerisembeddedunder someotheroperator.Whenthepresuppositionisnotembedded,such cancellation(bythesamespeaker)isusuallyinfelicitous,justasis cancellationofentailedcontentwhichisnotembedded.Thusthe denialofapresuppositionin(9)andthedenialofanordinary entailmentin(10)bothleadtopragmaticallyinfelicitousutterances (markedbya“#”). (9) #It’stheknavethatstolethetarts,butthereisnoknave. (10) #It’stheknavethatstolethetarts,buthedidn’tdoanything illegal. Thefactthatpresuppositionsassociatedwithunembeddedtriggersare notcancellable,exceptinanoutrightretractionofwhatthespeaker haspreviouslyclaimed,isoneofthefeaturesthatdistinguishesmost presuppositionsfromGriceanconversationalimplicatures(Grice, 1989).Forexample,anutteranceof(11a)mightordinarilyleadtothe so-calledscalarimplicaturein(11b).Butwhilethisimplicatureis cancellable,asin(11c),thepresuppositionthatthereisaknave, onceagain,isnotcancellable,asshownbytheoddityof(11d). (11a) Theknavestolemostofthetarts. (11b) Theknavedidnotstealallofthetarts. (11c) Theknavestolemostofthetarts—infact,hestolethem all. (11d) #Theknavestolemostofthetarts,buttherewasnoknave. Wecansummarizethetypicalbehaviorofentailments,presuppositions, andconversationalimplicaturesasfollows:   Entailments Presuppositions Implicatures Projectfromembeddings no yes no Cancellablewhenembedded — yes — Cancellablewhenunembedded no no yes Becausepresuppositionsaretypicallyonlycancellablewhenembedded, Gazdar(1979a,1979b)arguesthatpresuppositionsareusuallyentailed whenthetriggerisnotembedded. Theliteratureischoc-a-blocwithexamplesofpresuppositional inferencesapparentlydisappearing.Whethersuchexamplesare appropriatelydescribedasinvolvingcancellationispartlya theoreticaldecision,and,aswewillsee,manyscholarsavoidusing theterm“cancellation”forsomeorallsuchcases.One reasonforthisisthattheterm“cancellation”appearsto suggestthataninferencehasbeenmade,andthenremoved.Butinmany casestherearetheoreticalreasonsnottoregarditasanapt characterization,andinconsideringearlyviewsonpresuppositionin thenextsection,wewillcomeacrossoneclassofsuchcases (specifically,casesinvolvinghypotheticalreasoningin conditionals). 2.TheFrege-Strawsontradition Theearlyliteratureonpresuppositionalmostexclusivelyrevolved arounddefinitedescriptions,whicharesaidtopresupposethe existenceofauniquereferent.Aproblemariseswhenadefinite description,liketheKingofFrance,failstorefer.Russell (1905)claimedthatsentenceslike"TheKingofFranceisbald"are falsebecausethelogicalformofdefinitedescriptionscontainsa falseexistentialclaim.However,Strawson(1950)famouslyargued againstRussell’stheorybyproposingthatwhenadefinitedescription failstorefer,theresultcanbeasentencewhichlacksatruth value.Thuspresuppositionsareunderstoodasdefinednessconditions, necessaryrequirementsforanexpressiontohaveameaning. Strawson’sintuition,whichcanbetracedbacktoFrege(1892),leads tothefollowing definition:[3] Definition1(Strawsonianpresupposition) Onesentencepresupposesanotheriffwheneverthefirstistrue orfalse,thesecondistrue. Anotherdefinitionthatisoftenusedisthis: Definition2(Presuppositionvianegation) Onesentencepresupposesanotheriffwheneverthefirst sentenceistrue,thesecondistrue,andwheneverthenegationofthe firstsentenceistrue,thesecondsentenceis true.[4] hesetwodefinitionsareequivalentifnegationmapstrueontofalse, falseontotrue,andisundefinedwhenitsargumentisundefined. However,theseconddefinitionisnotableinthecontextoftheabove discussionofprojection,becauseitseemstodirectlyencodethe projectionpropertiesofatleastoneoperator:negation. Specifically,itsaysthatpresuppositionsareinferencesthatsurvive embeddingunder negation.[5] Itisclearthatiftheaboveassumptionsaboutpresuppositionare made,thenthepresuppositionsofasentencewillbethesameasthe presuppositionsofthenegationofthesentence.Butwhatabout projectionfromembeddingsotherthannegation?Averysimpleaccount ofprojectionisbasedonthecumulativehypothesis,first discussedbyMorgan(1969)andLangendoenandSavin(1971).Thisis theideathatpresuppositionsalwaysprojectfromembedding,asif therewerenoeffectslikecancellation.Atrivalentsemanticsthat yieldsthisbehaviorisobtainedbyusingtheWeakKleene connectives(Kleene,1952).Assume(forallthepartial/multivalent semanticsgiveninthisarticle)thatforclassicallyvalued arguments,theconnectivesbehaveclassically.ThenWeakKleene connectives(alsoknownastheBochvarInternalconnectives)are definedasfollows: Definition3(WeakKleene) IfanyargumentofasentencewithaWeakKleeneconnectivelacksa classicaltruthvalue,thenthesentenceasawholelacksatruth value. WeakKleenefailsasatheoryofpresuppositionbecauseitentails thatpresuppositionsprojectuniformly,whereasinfacttheydonot. AnothersystemofKleene’s,theStrongKleeneconnectives,doesnot havethisproperty: Definition4(StrongKleene) Iftheclassically-valuedargumentsofasentencewithaStrongKleene connectivewouldsufficetodetermineatruthvalueinstandardlogic, thenthesentenceasawholehasthatvalue;otherwiseitdoesn’thave aclassicalvalue. Forexample,inclassicallogicaconjunctionisboundtobefalseif oneofitsconjunctsisfalse,andthereforethesameholdsforStrong Kleene“and”.Similarly,sinceinclassicallogica disjunctionmustbetrueifoneofitsdisjunctsistrue,thesame holdsforStrongKleene“or”.Weobtainthefollowing truthtablesforthemainbinaryconnectives: Conjunction φ ∧ ψ t f ✭ t t f ✭ f f f f ✭ ✭ f ✭ Disjunction φ ∨ ψ t f ✭ t t t t f t f ✭ ✭ t ✭ ✭ Implication φ → ψ t f ✭ t t f ✭ f t t t ✭ t ✭ ✭ Nowconsiderthefollowingexample: (12) Ifthereisaknave,thentheknavestolethetarts. Let’signoreallpresuppositiontriggersin(12)save“the knave”,andshowthatStrongKleenepredictsthatthesentence asawholedoesnotpresupposethatthereisaknave.UsingDefinition 1,itsufficestofindatleastonemodelwhere(12)hasaclassical truthvalue,butinwhichthereisnoknave.Thisiseasy:insucha model,theantecedentisfalse,andinspectionoftheaboveStrong Kleenetableshowsthatwhentheantecedentofaconditionalisfalse, theconditionalistrue,aswouldbethecaseclassically.Infact, StrongKleenepredictsnopresuppositionfor(12).Thisisin contradistinctiontoWeakKleene,whichwouldfailtogive(12)a classicalvalueinknave-lessmodels,andhencepredictthat(12) presupposestheexistenceofaknave. ThereareothercaseswhereStrongKleenedoespredicta presupposition,andthepresuppositionpredictedisnotwhatwemight haveexpected.ThusStrongKleenegives(13a)aclassicaltruthvalue inallmodelswherethereisaknave,andinallmodelswherethere wastrouble.Sowhilewemighthaveexpectedthepresuppositionin (13b),StrongKleenepredictsthepresuppositionin (13c).[6] Wewillreturntothisissueshortly. (13a) Iftheknavestolethetarts,thentherewastrouble. (13b) Thereisaknave. (13c) Iftherewasnotrouble,thenthereisaknave. Muchofthediscussionofpartialandmultivalentapproachesto presuppositionoverthelastthreedecadeshascenteredonthe treatmentof negation.[7] Specifically,theissuehasbeenthetreatmentofcancellation exampleslike(14). (14) Thetartswerenotstolenbytheknave:thereisnoknave. Astandardapproachistoproposethatnegationisambiguousbetweena presupposition-preservingnegationandapresupposition-denying negation;seee.g.,thediscussionbyHorn(1985,1989).The presupposition-preservingnegation(akachoicenegation)we havealreadyseen,anditisfoundinboththeWeakandStrongKleene systems.Thepresupposition-denying(orexclusion)negation istypicallytakentomaptruetofalseandfalsetotrue,asusual, butalsotomapanargumentlackingaclassicalvaluetotrue.Thusif (14)isinterpretedinamodelwherethereisnoknave,but “not”isunderstoodasapresupposition-denyingnegation, then“thetartswerestolenbytheknave”wouldlacka classicalvalue,but“Thetartswerenotstolenbythe knave”,and(14)asawhole,wouldbetrue. 3.Pragmaticpresupposition Probablythemostsignificantphilosophicalcounterpointtothe Frege-Strawsonapproachtopresupposition,otherthantheoriginal non-presuppositionalworkofRussell,isduetoStalnaker(1972,1973, 1974),andlaterclarifiedinStalnaker (1998).[8] Stalnakersuggeststhatapragmaticnotionofpresuppositionis needed,sothattheproperobjectofphilosophicalstudyisnotwhat wordsorsentencespresuppose,butwhatpeoplepresupposewhenthey arespeaking.Apragmaticpresuppositionassociatedwithasentenceis aconditionthataspeakerwouldnormallyexpecttoholdinthecommon groundbetweendiscourseparticipantswhenthatsentenceis uttered.[9] OneconsequenceofStalnaker’sviewisthat,contrasemantic accountsofpresupposition,presuppositionfailureneednotproducea semanticcatastrophe.Thereare,however,twoweakertypesoffailure thatcanoccur:(i)aspeakerutteringsomesentence ScanfailtoassumethatsomepropositionPisin thecommonground,eventhoughmostutterancesofSwouldbe accompaniedbythepresuppositionthatP;and(ii)a speakercanpresupposesomethingthatisnotinthecommonground.We seethesetwotypesoffailurerespectivelyin(15a),wherethereis nopresuppositionthatMullahOmarisalive,and(15b)whereitis presupposedthatLukewasalive. (15a) Idon’tknowthatMullahOmarisalive.Idon’tknowifhe’sdead either.(GeneralDanMcNeill,Reuters,19May2008) (15b) Vaderdidn’tknowthatLukewasalive,sohehadnointentionsof convertingLuketotheSith.(Webexample) Theseexamplesinvolveasubclassoffactiveverbscalled "semifactives,"whichKarttunen(1971b)concludedonlytriggera presuppositioninsomepersonandtenseforms.AsKarttunenhimself realized,suchastipulationisunmotivated.ForStalnaker’spragmatic accountofpresupposition,theseexamplesarenotproblematic;the verb"know"neednotpresupposethatitscomplementistrue.Whenan addresseehearsthefirstsentenceof(15a),hewillrealizethatif itwereinthecommongroundthatMullahOmarwasalive,thenthe speakerwouldknowthis,andsothespeaker’sclaimwouldbefalse. Thereforethehearercanreasonthatthespeakerisnotpresupposing thecomplementof“know”tobetrue.Ontheotherhand, whenahearerisconfrontedby(15b),itisconsistenttoassumethat Lukewasalive.Sincespeakersusing“know”typically presupposethetruthofthecomplement,wecanassumethatthisisthe casehere. Stalnaker’sworkwaspartofanavalancheofpragmaticattacksonthe semanticconceptionofpresupposition.Workingintheimmediate aftermathofGrice’s1967WilliamJames lectures[10], accountslikeAtlas(1976;1977;1979),AtlasandLevinson(1981), Kempson(1975),Wilson(1975),andBöerandLycan(1976)all presentdetailedargumentsthatpresuppositionsshouldbeunderstood assomethingakintoconversationalimplicatures.Generallyspeaking, theseapproachesjustifypresuppositionalinferencesbyusingthe maximsofrelevanceandquantity.Thus,forexample,Atlas(1976) suggeststhatanembeddingofadefiniteunderanegationwilltendto produceameaningthatisruledoutasinsufficientlystrongto satisfythemaximofquantity,unlessitisstrengthenedbytreating thedefiniteasifithadwidescopeandcouldactreferentially. ContemporarydescendantsofthispragmatictraditionincludeAbbott (2000;2006;2008),Simons(2001;2003;2004;2006;2007),and Schlenker(2007;2008).BothAbbottandSimonsareatpainsto distinguishbetweendifferentpresuppositiontriggers,ratherthan lumpingthemalltogether.ThusSimons,forexample,makesacasefor derivingpresuppositionalinferencesassociatedwithfactivesand aspectualadverbsusingacombinationofStalnakerianandGricean reasoning,allowingthattypicallyanaphorictriggerslikethe additive“too”mightfunctionconventionally.Ontheother hand,Schlenker’spragmaticderivationofprojectionproperties,using bothstandardmaximsandatleastonerulespecificto presuppositions[11], doesnotmakefine-graineddistinctionsbetweenpresupposition triggers. Thereisacontrastamongpragmaticapproachestopresuppositionas regardswhetherthesourceofpresuppositionsisdiscussed.The approachesmentionedintheprecedingparagraph,whichattemptto derivepresuppositionalinferencesfromgeneralconversational principles,aimtoexplainboththesourceofpresuppositions,andthe phenomenonofprojection.ButStalnakermadenoattemptwhatsoeverto explainwherepresuppositionscamefrom,beyondindicatingthatthey areinferentialtendenciesthatmightormightnotbeassociatedwith semanticpresuppositions.Thisemphasisontheprojectionof presuppositionsratherthantheirsource,whichholdsalsoofthe contemporaneousworkbyKarttunen(1974;1973),towhichweshallturn shortly,livedoninmuchoftheworkinfluencedbythesetheories.It isparticularlyobviousinwhatwecancollectivelyterm cancellation-basedtheoriesofpresupposition,ledbyGazdar (1979a;1979b),andincludingSoames(1979;1982),Mercer(1987; 1992),Gunji(1981),Marcu(1994),Horton(1987),HortonandHirst (1988),Bridge(1991),andvanderSandt(1982;1988). CancellationaccountscanbetracedbackinspirittoStalnaker’s accountofsemifactives,discussedabove,inwhichpresuppositionsare defeatedbycompetingconversationalinferences:thegeneralideais simplytomakepresuppositionsintodefaults,andwipethemout whenevertheywouldcausepragmaticembarrassment.Gazdarprovideda remarkablystraightforwardformalizationofthisaccount,aswellas extendingtomanyotherprojectionphenomena,basedonageneral principlehecharacterizesas“Allthenewsthatfits”.In Gazdar’smodel,thestrategyforahearerisfirsttoidentifysetsof entailments,conversationalimplicatures,andpresuppositions,and thentotryaddingthemtothespeaker’ssetofcommitments. Definition5(Gazdar:cancellation) Implicaturesandentailmentsdefeatpresuppositions,soaheareradds tohisorhercommitmentsonlythosepresuppositionsthatare compatiblewithbothimplicaturesandentailments.Allremaining presuppositionsarecancelled. Consider(16a),andassumetherearenorelevantpre-existing commitments: (16a) Ifthekingisangry,thentheknavestolethetarts. (16b) Ifthereisaknave,thentheknavestolethetarts. AccordingtoGazdar,(16a)entailsthatifthereisanangrykingthen thereisaknaveandhestolesomesetoftarts.(Thismuchall theoriesagreeon;sometheoriesmaypredictstrongerentailments.) Thesetofimplicatureswouldincludetheclausalimplicaturethatthe speakerdoesn’tknowwhetherakingisangry,anddoesn’tknowwhether aknavestoletarts.Thepresuppositions(or“potential presuppositions”,inGazdar’sterms)arethatthereisaunique king,auniqueknave,andauniquesetoftarts.Thehearerproceeds byaddingtheentailmentsto(theirrepresentationof)thespeaker’s commitmentset,thenaddingwhateverimplicaturesfitin,andthen addingthepresuppositionsthatfitafterthat.Inthiscase,allthe entailments,implicatures,andpresuppositionsareconsistent,andall canbeaddedwithoutanybeingcancelled. Butnowconsider(16b),repeatedfrom(12).Herethereisan implicaturethatthespeakerdoesn’tknowwhetherthereisaknave. Theheareracceptsthisandotherimplicatures,andthenconsidersthe presuppositionsthatthereisaknaveandthattherearesometarts. Thepresuppositionthattherearetartsisunproblematic,andis added,butthehearercannotconsistentlyaddthepresuppositionthat thereisaknave.Sothispresuppositioniscanceled,and(16b)does notpresupposethatthereisaknave.Hence,accordingtoGazdar, presuppositionsaresometimesblockedbyconversational implicatures. Withinthespaceofcancellation-basedaccountsofpresupposition,it ishardtobeatGazdar’sforitsconceptualandtechnicalsimplicity, anditsempiricalcoverage.Someconceptualquestionsremain,however, suchaswhyitshouldbethatpresuppositionsarethelast thingstobeaddedintheprocessofupdatingcommitments.Vander Sandt’s(1982,1988)reformulationofthecancellationmodelgivesus analternativewaytothinkaboutthis,bymodelingprojectionin termsofwhetherpresuppositionscouldhavecomefirst. Definition6(VanderSandt:cancellation) Projectonlythosepresuppositionsthatcouldbeconjoinedtothe beginningofthesentencewhileleavingtheutteranceconsistentwith (neo-Gricean)conversational principles.[12] TheintuitiveideaunderlyingvanderSandt’sproposalisthat presuppositionsaregiveninformation,andinthissense “precede”theircarriersentences,ifnotde factothenatleastdejure,inthesensethatthe contextwouldhavelegitimatedtheirpresence.Inthecaseof(16), frontingthepresuppositionthattherearesometartsyieldsthe sentencesin(17). (17a) Therearesometartsandifthekingisangrythentheknavestole thetarts. (17b) Therearesometartsandifthereisaknave,thentheknavestole thetarts. Thefrontingofthepresuppositiondoesnotleadtoaclashwithany Griceanprinciples,sothepresuppositionsarepredictedtoprojectin (16a)and(16b).Similarly,frontingthepresuppositionthatthereis aknaveto(16a),asin(18a),producesnoclash,so(16a)presupposes thatthereisaknave.Butaddingthepresuppositionthatthereisa knaveto(16b),asin(18b),doesresultinaclash:since(18b)is truth-conditionallyequivalenttothesimpleconjunction“there isaknaveandtheknavestolethetarts”,itisredundant, conflictingwiththerequirementofbrevityinGrice’sMaximof Manner.OnvanderSandt’sanalysis,iffrontingapresupposition wouldproducearedundantresult,thenthatpresuppositioncannot project.So(16b)iscorrectlypredictednottopresupposethatthere isaknave. (18a) Thereisaknaveandifthekingisangrythentheknavestolethe tarts. (18b) Thereisaknaveandifthereisaknave,thentheknavestolethe tarts. Itshouldbenoted,however,thatevenif(18b)isredundant,itis arguablyafelicitousdiscourse,andthereforesomesubtletyisneeded inapplyingvanderSandt’scancellationprincipleinthesimplified formabove.Theissueisnotsimplywhetheradiscourseisfelicitous, butwhetherthereisanyclashwiththemaxims.Andthiswillof coursedependonhowexactlythemaximsareformulated.Butforthe purposesofunderstandingtheintentionofvanderSandt’sanalysis, wecantakeitthatthoughanutteranceof(18b)couldbefelicitous, itwouldbeacaseofflouting(inGrice’ssense),acase whereamaximisdisobeyedinordertopreservesomegreater conversationalgoal. Amorerecentpragmaticapproachsituatespresuppositionsinawider taxonomyofprojectivecontentbasedonthepropertyof at-issueness. Oncewe’veidentifiedtheQUD,wecanidentify contentthatismeanttoberelevanttowardsansweringit.Inother words,contentthatis at-issue.Simonsetal(2010)proposethat thispragmaticpropertyisrelevanttoprojectionsuchthatonly not-at-issuecontentprojects. Definition7(Cancellationviaat-issueness) Projectonlythosepresuppositionsthatarenot-at-issue. Considerexamples(19)and(20)fromBeaver(2010).Theseexamples involvethefactiveverb"discover",andhenceareexpectedtotrigger apresuppositionthatitscomplementistrue,i.e.thatthe addressee’sworkisplagiarized.In(19),where"discover"is stressed,thispresuppositionseemstoproject,withtheimplication that,absenttheT.A.findingoutwhat’sgoingon,thespeakeris preparedtooverlooktheplagiarism.Inthiscase,wemightsaythat whatisat-issueisthequestionofwhethertheT.A.findsoutabout theplagiarism.However,in(20),wherestressinon"plagiarized",it isfarlessclearwhetherthespeakerbelievestheaddressee’sworkis plagiarized.Thiscanbeexplainedifstresson"plagiarized"is markingthepropositionthattheworkisplagiarizedasat-issue.So, thegeneralizationinDefinition7predictsthatthispropositiondoes notproject. (19) IftheT.A.disCOVersthatyourworkisplagiarized,Iwillbe forcedtonotifytheDean. (20) IftheT.A.discoversthatyourworkisPLAgiarized,Iwillbe forcedtonotifytheDean. Pragmaticapproachestopresuppositiontaketheburdenofprojection offconventionaltriggeringandallowfordiscoursestructureto influenceprojectionwhentriggersareplacedindifferentcontexts. Oneconsequenceofthisviewisthatitcanbeapplied cross-lingustically(e.g.Tonhauseretal2013)withoutrunninginto theissueoflanguage-specificconventionalizedtriggers.Additionally itallowsustoaccountfortheprojectionofexpressionsthatdon’t displayotherpropertiesthatweassociatewithpresupposition,for exampleappositives,expressivesandhonorifics(Potts2005).Inother words,there’sareasonwhysomethingsthataren’tpresuppositions stillbehavelikethem. Anewtrendofpsycholinguisticresearchpotentiallyofferssome promisingnewevidenceaboutthedegreetowhichpresuppositionsare processedlikepragmaticinferencesorsemanticentailments.The methodsinthesestudiestypicallyrelyonidentifyingatwhatpoint duringtheprocessingofanutterancepresuppositionsbecomeavailable tothelistener.Thelineofreasoninggoeslikethis:if presuppositionsareconventionallyencodedintheirtriggersandare conditionsforcontextstoupdate,theyshouldariseimmediatelywhen atriggerisused.However,ifpresuppositionsaretheresultof pragmaticreasoning,thereshouldbeadelayintheirprocessing. Whiletheresultsofthesestudieshavebeensomewhatinconsistent, theylargelysuggestthatpresuppositionsareprocessedimmediately. Forexample,oneoftheexperimentsinTiemannetal.(2011)finds thatpresuppositiontriggersthemselvescarryaprocessingcost relativetoothernon-presuppositionalexpressions.Inaneye-tracking study,Kim(2008)findsthatthereareshiftsineyemovementsquite rapidlyaftertheutteranceofthetriggeronly.These studiesandthemethodologiesusedinthemarestillrelatively new. However,empiricalworkontheprocessingofpresuppositions remainsapromisingstrandofresearchindeterminingtheextentto whichwecanconsiderpresuppositionstobeconventional 4.Localcontextsandthedynamicturn Instatictheoriesofmeaning,suchasthetrivalentaccountsof presuppositiondiscussedabove,whenasentenceisinterpretedina givencontext,everyexpressioninthatsentenceisinterpreted relativetothatsamecontext.Indynamictheoriesofmeaning,the contextisallowedtochangeasasentenceisinterpreted.The earliestandperhapsstillclearestmotivationforthisdynamismis whatmightbetermedtheanaphoricasymmetry.Theanaphoric asymmetryconsistsinpronounsacrosslanguagesbeingusuallyresolved toexpressionsthathavealreadyoccurred,andonlybeingresolved cataphoricallytoexpressionsthatoccurlaterinthesentencein quitespecialconfigurations.Thedynamicexplanationforthisisthat pronounsareresolvedtoreferentsviathecontext,andthatthe contextisupdatedduringprocessingsothatittypicallyonly containsinformationaboutthingsthathavealreadybeenmentioned. Thequestionmightthenbeasked:howdoestheprocessingofanaphora relatetotheprocessingofpresuppositions?Itturnsoutthatthere areremarkablystrongparallelsbetweenphenomenainvolvinganaphora andpresupposition,andthisinturnprovidesamotivationfor consideringnotonlyanaphorabutalsopresuppositiontobean intrinsicallydynamicphenomenon.Inthissection,wewillfirststudy theparallelsbetweenpresuppositionandanaphora,seenasapossible motivationforconsideringcontextchangeinanalyzingpresupposition, andthendescribeinlargelyhistoricaltermsthesequenceof developmentsthatledtocontemporarydynamictheories. 4.1Presuppositionandanaphora Whileanumberofauthorshavenotedthatpresuppositionsbehavein somerespectslikeanaphors(e.g.,Kripke2009andSoames1989),it wasvanderSandt(1989;1992)whobroughtouttheconnectionthemost forcefully.Henotedthatforeveryconfigurationofanaphorsand antecedentswhereapronounisinterpretedanaphoricallybutisnot interpretableasaboundvariable,asimilarconfigurationispossible with presuppositions.[13] Ineachofthefollowingquadruples,weillustratesomeconfiguration inthe(a)and(b)examples,whilethe(c)and(d)casesshowthata slightdivergencefromtheoriginalconfigurations(i.e.thosein(a) and(b)),producesinfelicity.Notethatthe(a)and(c)examples includeananaphoricpronoun(“it”),andthe(b)and(d) examplesincludethefactiveverb“knows”,whichtriggers apresuppositionthatitspropositionalcomplementistrue(i.e.,that Fredleft).So,anaphoraisfelicitousinconfigurationsinwhich presuppositionsarefelicitous,andanaphoraisinfelicitousin configurationsinwhichpresuppositionsareinfelicitous.Itisclear thatconnectionsbetweenanaphoricpronounsandtheirantecedentson theonehand,andpresuppositiontriggersandtheirantecedentsonthe other,aresensitivetoverysimilarconfigurationalrequirements. Inter-sentential(discourse)anaphora (21a) Therewasastorm.Itwasfierce. (21b) Fredleft.MaryknowsthatFredleft. (21c) #Itwasfierce.Therewasastorm. (21d) #MaryknowsthatFredleft.Fredleft. Donkeyanaphora (22a) Ifafarmerownsadonkeythenhebeatsit. (22b) IfFredleftthenMaryknowsthatFredleft. (22c) #Ifafarmerdoesn’townadonkey,thenhebeatsit. (22d) #IfFreddidn’tleaveleftthenMaryknowsthatFredleft. Modalsubordination (23a) Awolfmightcometothedoor.Itwouldeatyou. (23b) Fredmighthaveleft.MarywouldknowthatFredhasleft. (23c) #Awolfmightcometothedoor.It’sbrown. (23d) #Fredmighthaveleft.MaryknowsthatFredhasleft. Bathroomanaphora (24a) Eitherthere’snobathroominthishouse,orelseit’sina funnyplace. (24b) EitherFreddidn’tleave,orelseMaryknowsthatheleft. (24c) #Eitherthereisabathroom,orelseit’sinafunnyplace. (24d) #EitherFredleft,orelseMaryknowsthatheleft. 4.2Satisfactioninlocalcontexts Aspreviouslynoted,earlyworkonpresuppositionwasprimarily focusedondefinitedescriptions,anddefinitedescriptionshave continuedtobetheparadigmaticexampleofapresuppositiontrigger inphilosophicalliteratureoverthelastfiftyyears.However,bythe early1970s,morelinguisticallyorientedworkhadexpandedthe empiricaldomainofpresuppositiontheoryfromdefinitedescriptions toothertriggertypes,includingfactives(KiparskyandKiparsky, 1970),implicatives(Karttunen,1971a),focusparticles(Horn,1969), verbsofjudging(Fillmore,1971)andsortalconstraints(Thomason, 1972).Stalnaker’sdiscussionofKarttunen’ssemifactivesprovidesan earlyexampleofhowthislinguisticexpansionoftheempiricaldomain hasimpactedphilosophicalwork.Alsobytheearly1970s,linguists hadexpandedtheempiricaldomaininanotherdirection.The philosophicalliteraturewaslargelyorientedtowardsunembedded presuppositiontriggersandtriggersundernegation,butaswehave alreadymentioned,Morgan(1969)andLangendoenandSavin(1971) generalizedtheissuebyconsideringarbitraryembeddings.However,it wasnotuntilKarttunen(1973)thatthefullcomplexityofthe projectionproblembecameapparent.Bymethodicallyconsidering projectionbehaviorconstructionbyconstruction,Karttunenshowed thattherewasmorevariationinprojectionbehaviorthanhadbeen previouslydescribed,makingitquiteclearthatnoneoftheextant Frege-Strawsonderivedsystemscouldhopetocovereverycase,and thisultimatelyledtotheapplicationofdynamicmodelsofmeaningto theproblemofpresuppositionprojection. Karttunen(1973)presentedataxonomyofembeddingconstructionsthat dividedthemintothreeclasses:plugs,holesand filters.Plugscompriseaclassofpredicatesandoperators whichKarttunenclaimedblocktheprojectionofpresuppositions,while holesareaclassofpredicatesandoperatorswhichallow presuppositionstoprojectfreely.So,forexample,since“told that”isaplug,accordingto Karttunen,[14] (25)ispredictednottopresupposethatthereisaKingofFrance. Ontheotherhand,since“perhaps”isahole,(26)is predictedtopresupposethatthereisaKingofFrance. (25) MarytoldJimthattheKingofFrancewasbald. (26) PerhapstheKingofFranceisbald. Karttunen’sfiltersincludethebinarylogicalconnectives“if then”,“and”,and“or”.Theintuition behindthefiltermetaphoristhattheseconstructionsallowonlysome presuppositionstoproject,othersbeingcaughtinthefilter,andwe havealreadyseenexamplesofthisphenomenon.Thusexample(12) showedthatsometimesapresuppositionintheconsequentofa conditionaldoesnotproject:herethepresuppositionthattherewasa knaveisfilteredout.Butthesameexampleincludesanoccurrenceof thedefinite“thetarts”intheconsequent,andthe presuppositionthatthereare(oratleastwere)sometartsprojects fromtheconditional.Karttunenconcludedthattheconsequentofa conditionalactsasaholetosomepresuppositions,butfiltersout allthosepresuppositionswhichareentailedbytheantecedent,or, moregenerally,byacombinationoftheantecedentandcontextually suppliedbackgroundinformation. Karttunen’skeyexampleshowingtheroleofcontextbears repetition: (27) EitherGeraldineisnotamormonorshehasgivenupwearingher holyunderwear. Thesecondhalfof(27)contains(atleast)twopresupposition triggers:thedefinitedescription“herholyunderwear” andtheaspectualverb“giveup”,whichtriggerthe presuppositionsthatGeraldineusedtohaveandwearholyunderwear, respectively.Karttunen’sfilteringconditionfordisjunctionsremoves fromtherightdisjunctanypresuppositionsthatareentailedbya combinationofthecontextandthenegationoftheleftdisjunct.Now consideracontextsupportingthepropositionthatallmormonshave holyunderwearwhichtheywearregularly.Itfollowsfromthis propositionandthenegationoftheleftdisjunct,i.e.,the propositionthatGeraldineisamormon,thatGeraldinehasholy underwearandhaswornitregularly.Buttheseareexactlythe presuppositionstriggeredintherightdisjunct,sotheyarefiltered out.Itfollowsthat(27)hasnopresuppositions. Karttunen’s(1973)accountisofinterestnotonlyforitstriptychof plugs,holesandfilters,butalsobecauseitsetsthebackgroundfor acrucialshiftofperspectiveinKarttunen(1974),andthencetothe dynamicapproachestopresuppositionthathavebeendominantinrecent years.Whatremainedunclearinthe1973paperwasthemotivationfor filtering,i.e.whypresuppositionsshouldbefilteredoutwhen entailedbyothermaterial.Karttunen(1974)suggestsanalternative conceptionbasedontheideaoflocalcontextsofevaluation. Theideaisthatthepartsofasentencearenotnecessarilyevaluated withrespecttothesamecontextasthatinwhichthesentenceasa wholeisevaluated:alocalcontextmaycontainmore informationthantheglobalcontext.Forexample,when evaluatingaconjunction,thesecondconjunctisevaluatedinalocal contextwhichcontainsnotonlytheinformationintheglobalcontext, butalsowhateverinformationwasgivenbythefirstconjunct. Karttunen(1974)defineslocalcontextsofevaluationforarangeof constructions,andsuggeststhefollowingrequirement:presuppositions alwaysneedtobeentailed(or“satisfied”,asheputit) inthelocalcontextinwhichthetriggerisevaluated.Giventhis requirement,theoverallpresuppositionsofasentencewilljustbe whateverpropositionsmustbeinacontextofanutteranceinorderto guaranteethattherequirementsassociatedwithpresupposition triggersaresatisfiedintheirlocalcontextsofinterpretation. Karttunenspelledouthowlocalsatisfactionshouldbecalculated separatelyforeachconnectiveandoperatorheconsidered.However, recentdevelopmentsinSchlenker(2008)provideageneralwayof calculatingwhatthelocalcontextshouldbe.Inthefollowing reformulationofKarttunen’smodelweincorporateSchlenker’sinsights alongthelinesproposedbyBeaver(2008). Letussaythatsomeclauseinacomplexsentenceis redundantrelativetosomecontextofutteranceifyoucan replacethatclausebyatautologywithoutaffectingtheamountof factualinformationconveyedbythesentenceinthatcontext.For example,in(28),thefirstconjunctisredundantinanycontextof utterance.Here,thesamefactualinformationwouldbeconveyedby “MaryisMaryandMaryownsasheep”,wherethefirst conjunctisreplacedbythetautology“MaryisMary”. (28) MaryownsananimalandMaryownsasheep. Nowletussaythataclauseisleft-redundantifitis possibletotellbylookingatthematerialinthesentencetothe leftoftheclausethattheclauseisredundant.So“Maryowns ananimal”isnotleft-redundantin(28)(exceptifthecontext ofutterancealreadyentailsthatMaryownsananimal),becausethere isnomaterialbeforethatclause,implyingthatitisimpossibleto tellbylookingatmaterialtotheleftoftheclausethattheclause isredundant.Ontheotherhand,thesamesententialfragment, “Maryownsananimal”,isleft-redundantin(29)andalso in(20): (29) MaryownsasheepandMaryownsananimal.(Truth-conditionally equivalentto:MaryownsasheepandMaryisMary) (30) IfMaryownsasheepthenMaryownsananimal. (Truth-conditionallyequivalentto:IfMaryownsasheepthen MaryisMary) Nowweusethisideaofleftredundancy,whichechoestheabovevan derSandtanalysisoftheeffectoffrontingpresuppositions,to definethecrucialnotioninKarttunen’s(1974)account. Definition8(Karttunen/Schlenker:Presuppositionvia satisfaction) ApresuppositionPissatisfiedatpointX inSiffPwouldbeleft-redundantifaddedatpoint X.Asentencepresupposeswhateverpropositionsmustholdin globalcontextsofutterancesuchthateachlocallytriggered presuppositionissatisfiedwhereitstriggeroccurs. Asanexample,letusconsiderthepresuppositionspredictedfor(27), repeatedbelow: (27) EitherGeraldineisnotamormonorshehasgivenupwearingher holyunderwear. Notefirstthatforallsentencesoftheform“Aor B”,thenegationofAissatisfiedwithinthe rightdisjunct.So“Geraldineisamormon”issatisfiedin therightdisjunctof(27).Andmoregenerally,anythingentailedbya combinationofpropositionsinthecontextandthenegationofthe leftdisjunctwillbesatisfiedintherightdisjunct.Now,letus considertheclause“shehasgivenupwearingherholy underwear”:wetakethistotriggerthepresuppositionthat Geraldinehashadholyunderwearthatshewore.Thispresupposition willbesatisfiedprovidedtheglobalcontextofutterance,combined withthenegationoftheleftdisjunct,entailsthatshehashadholy underwearthatshewore.Andclassicallythiswillbethecaseifand onlyifthecontextsupportstheconditional“ifGeraldineisa mormon,thenshehashadholyunderwearthatshewore”(which wouldalsobethecaseifthecontextsupportedsomethingstronger). Hence,thisconditionalisthepresuppositionKarttunen(1974) predictsfor(27). OnenotablepropertyofKarttunen’s1974treatmentofexampleslike (27),apropertynotfoundinhis1973model,isthatthe presuppositionpredictedisconditionalized.Thatis,(27)isnot predictedtopresupposethatGeraldinehashadholyunderwearthatshe wore,butthatifsheisamormonthenshehashadsuchunderwear.We alreadyencounteredsuchconditionalizedpresuppositionsinour discussionofStrongKleene;infact,StrongKleenepredictsexactly thesameconditionalizedpresuppositioninthiscase.Karttunen’s1974 modelalsopredictsconditionalizedpresuppositionswhenthe presuppositiontriggerisintherightconjunctofaconjunction,or intheconsequentofa conditional.[15] Thusin(16a),repeatedbelow,thepresuppositionspredictedarethat thereisaking(sincepresuppositionstriggeredintheantecedentare notconditionalized),andthatifthekingisangry,thenthereisa knave.In(16b),theconditionalpresupposition(thatifthereisa knave,thenthereisaknave)istrivial,soineffectthereisnonet presupposition.[16] (16a) Ifthekingisangrythentheknavestolethetarts. (16b) Ifthereisaknave,thentheknavestolethetarts. AlthoughKarttunen’s(1974)modelisdistinctconceptuallyfromanyof itspredecessors,wehavealreadynotedthatitsharesatleastsome predictionswithStrongKleene.AnobservationmadebyPeters(1979) showedthatthe1974modelissurprisinglycloselyrelatedtothe semanticaccountsofpresuppositiondiscussedabove.Inparticular, PetersshowedthatKarttunen’swayofcalculatingpresuppositionsfor thetruthconditionalconnectivesisequivalenttowhatwouldbe obtainedwithinathree-valuedlogic,butwithspecialnon-symmetric connectives.HereisageneralwayofdefiningthePetersConnectives, inspiredbothbySchlenker(2009;2008)andGeorge(2008): Definition9(MiddleKleene/Petersconnectives) Gofromlefttorightthroughthesentence.Foreachargument Xthattakesanon-classicalvalue,checkwhetheronthe basisofmaterialonitsleft,assigninganarbitraryclassicalvalue toXcouldconceivablyhaveaneffectontheoverallvalue. Ifso,thesentenceasawholelacksaclassicaltruthvalue.Ifnot, justassignXanarbitraryvalue,andcarryon.Ifthis procedureallowsallnon-classicalvaluestobefilledinclassically, thenthesentencecanbeassignedaclassicalvalue. Forexample,thisproceduremakesaconjunctionclassicalifbothits argumentsareclassical,falseiftheleftconjunctisfalse,and undefinedotherwise.Thusundefinednessoftheleftconjunctforces undefinednessoftheentireconjunction,whereasundefinednessofthe rightconjunctonlysometimesyieldsundefinednessoftheentire conjunct,asseeninthefollowingcomparisonoftruthtablesin varioussystems.Theneteffectisthatpresuppositionsoftheleft conjunctprojectintheMiddleKleenesystem,justasintheWeak Kleenesystem,butpresuppositionsoftherightconjunctare conditionalized,justasintheStrongKleenesystem.Thisbehavior preciselymirrorsthatoftheKarttunen(1974)model. Definition10(Trivalenttruthtablesfor conjunction) WeakKleene φ ∧ ψ t f ✭ t t f ✭ f f f ✭ ✭ ✭ ✭ ✭ MiddleKleene/Peters φ ∧ ψ t f ✭ t t f ✭ f f f f ✭ ✭ ✭ ✭ StrongKleene φ ∧ ψ t f ✭ t t f ✭ f f f f ✭ ✭ f ✭ TheequivalencebetweenPeters’connectivesandKarttunen’smodel pavedthewayforamorecompletereformulationoftheKarttunenmodel inKarttunenandPeters(1977;1979),wherecertaintypesof presupposition(whichKarttunenandPetersregardasconventional implicaturesratherthanpresuppositions)aretreatedina compositionalgrammarfragment.Thisfragmentusestwodimensionsof meaning,oneforpresuppositionandoneforassertion,andis effectivelyanimplementationofthePetersconnectivesina four-valuedlogic;seeKrahmer(1994);Krahmer(1998),Beaver(2001), andBeaverandKrahmer(2001)fordiscussion,andthelatterfora fragmentthatmirrorsthatofKarttunenandPeters,butallowsfor richerinteractionsbetweenpresuppositionsand quantifiers.[17] 4.3Satisfactiontheories AlthoughKarttunen’s(1974)modelturnedouttobeequivalenttoa systemwhich,fromapurelytechnicalpointofview,isinthe Frege-Strawsontradition,Karttunen(1974)wasoneoftheseminal papersofthedynamiczeitgeistthatsweptthroughsemanticsand pragmaticsinthelastdecadesofthetwentiethcentury.Alsorelevant hereareHamblin(1970),Stalnaker(1972;1974),Gazdar(1979)and Lewis(1979),allofwhomadvanceddynamicmodelsofpragmaticsin whichthe(joint)commitmentsofspeakersandhearersevolveasnew assertionsaremadeandtheircontentbecomespartofthelinguistic contextavailableforfutureutterances.Itisagainstthisbackground thatHeim(1982;1983)offeredthefirstdynamicsemanticaccountof presupposition.Heim’smodelutilizesStalnaker’snotionofacontext asasetofallpossibleworldscompatiblewithwhathasbeen establishedatthatpointinaconversation,butinvolvesacrucial twistadaptedfromKarttunen.InStalnaker’smodel,asingle globalcontextisupdatedeachtimenewinformationis asserted,butinHeim’smodelthecontextisupdatedlocally intheprocessofcomputingthemeaningsofsubpartsofacomplex expression.[18] WecandefineasimplifiedversionofHeim’ssystemasfollows: Definition11(DynamicSemantics) AssumingthatthecontextsetCisasetofpossible worldsandSandS′aresentences: C+S=thesubsetofworldsinCthat arecompatiblewithS,butthisisdefinediff S’spresuppositions(ifany)aretrueinallworldsin C. C+¬S=C−(C+ S) C+S∧S′=(C+ S)+S′ C+S○S′,where○ issometruthfunctionaloperator,isgivenbythesimplestclassical definitionof○intermsof¬and∧thatpreservesthe orderofthetwosub-clauses. SissatisfiedinacontextCiff C+S=C(i.e.,updatingCwith Shasnoeffect). SpresupposesS′iff S′issatisfiedinallcontextswhereupdatewith Sisdefined. Clause(iv)entailsthatupdatewithaconditionalisdefined viatheequivalenceA→B≡¬ (A∧¬B)(providedthatthesubordinate clauseprecedesthemainclause).Toseehowthiswillwork,let’s considerthefollowingexample: (31) Ifthekingisangry,thentheknavestolethetarts.(= (16a)) Inordertoupdateacontextwith(31),wemustdothe equivalentofupdatingwith(32a).Nowclause(ii)saysthat toupdateacontextwith(32a),wemustfirsttryupdatingwith(32b), andsubtracttheresultfromtheoriginalcontext(soastoleave behindwhicheverworldsarenotcompatiblewith(32a)).But (32b)isaconjunction,sowemustfirstupdatewiththeleftconjunct (32c),andthenwiththeright(32d).Updatingwith(32c)isonly definedifthepresuppositionthatthereisakingissatisfiedinall worldsinthecontextset.Weimmediatelyseethat(31)and(32a), (32b),and(32c)allhavethisrequirement,i.e.,theypresupposethat thereisaking.Providedthispresuppositionissatisfied,updating with(32c)producesasubsetofworldswherethekingisangry.Weuse thisreducedcontextsetforupdatewith(32d).Butupdatewith(32d) againusesthenegationclause(ii)oftheabovedefinition. Sowestartedoffwithasetofworldswherethereisaking,we reducedittoasetofworldswherethekingisangry,andnowwemust updatethatcontextwith(32e),anupdatewhichwillonlybedefined ifthereisaknave. (32a) It’snotthecasethat[thekingisangryandtheknavedidn’t stealthetarts]. (32b) Thekingisangryandtheknavedidn’tstealthetarts. (32c) Thekingisangry. (32d) Theknavedidn’tstealthetarts. (32e) Theknavestolethetarts. Theupshotoftheaboveargumentationisthatupdateofacontextwith (31)isonlydefinedforcontextsinwhich(a)thereisaking,and (b)alltheworldswherethekingisangryareworldswherethereisa knave.Followingthedefinitionsthrough,itturnsoutthat,once again,theoriginalsentencecarriesbothanon-conditionalized presupposition,thatthereisaking,andtheconditionalized presuppositionthatifthekingisangry,thenthereisaknave. Thesatisfactionbasedmodelhasseenconsiderablefurther development—seee.g.,Beaver(1992;2001),Chierchia(1995), Heim(1992),Zeevat(1992),and,foraratherdifferentformalization ofadynamicsemanticapproach,vanEijck(1993;1994; 1995).[19] 5.Accommodation Themostimportantfeatureofthesatisfactionmodelnotcoveredin thedescriptionaboveisaccommodation.Accommodationwasfirst discussedbyKarttunen(1974)andStalnaker(1974),thoughonlynamed assuchbyLewis(1979).Karttunenintroducestheconceptas follows: Ordinaryconversationdoesnotalwaysproceedintheidealorderly fashiondescribedearlier.Peopledomakeleapsandshortcutsbyusing sentenceswhosepresuppositionsarenotsatisfiedinthe conversationalcontext.Thisistheruleratherthantheexception […]Ithinkwecanmaintainthatasentenceisalwaystakento beanincrementtoacontextthatsatisfiesitspresuppositions.If thecurrentconversationalcontextdoesnotsuffice,thelisteneris entitledandexpectedtoextenditasrequired.(Karttunen1974:191) Ifthislooksreasonablystraightforward,thereadershouldbewarned thataccommodationisamongthemorecontentioustopicsin presuppositiontheory. [20] Tobeginwith,therearevariousnotionsofaccommodation,someof whicharestricterthanothers.Toexplain,considerthefollowing examplebyHeim(1982): (33) JohnreadabookaboutSchubertandwrotetotheauthor. Inordertodeterminetheintendedmeaningof“the author”,thehearerhastoinfer(i)thatthereisan authorand(ii)thatthesaidauthorwrotethebookreadby John.Whereasonabroadunderstandingofaccommodation,allofthis isaccommodated,onastrictconstrualonly(i)is,and (ii)isabridginginference.Thisisnotjustamatterof terminology.Ifwechoosetobestrict,wecanarguethatthereis somethinglikean“accommodationmodule”,whichassuch hasnothingtodowithworldknowledge;whereasifthenotionis construedmorebroadly,accommodationisofapiecewithbridging.To facilitatethefollowingdiscussion,wewilladoptastrictnotionof accommodation,andtakethenaiveviewthatwhatisaccommodatedis thepresuppositionastriggeredby,e.g.,adefiniteNPorfactive verb. 5.1GlobalandLocalAccommodation Withthesepreliminariesoutoftheway,weturntothefirstmajor question:Wherearepresuppositionsaccommodated?Thoughit mayseemoddatfirst,thisquestionisinescapableifweassume,as isstandardindynamicsemantics,thatanexpressionmayoccurin severalcontextsatthesametime(cf.Section4.3).ForHeim(1982), followingLewis(1979),accommodationisaprocesswherebycontexts areadjustedsoastomakeupdatepossiblewhenpresuppositionsare notsatisfied.Intermsofhertreatmentofaccommodation,Heim’s majorinnovationoverLewiswastoallowthisprocesstotakeplace notonlyintheglobalcontextofutterance,butalsoonlocal contextsfoundmidwaythroughanupdate.Toillustrate,considerthe following: (34a) (c0)Maybe(c1)Bettyis tryingtogiveupdrinking. (34b) (c0)Maybe(c1)Wilma thinksthat(c2)herhusbandishavingan affair. Herec0referstotheglobalcontextinwhicha givensentenceisuttered,andc1and c2areauxiliary,orlocal,contexts.In(34a), themodal“maybe”createsanauxiliarycontextofpossible statesofaffairsinwhichBettyistryingtogiveupdrinking;the same,mutatismutandis,for(34b).Thepresuppositiontriggeredin (34a),thatBettyusedtodrink,canbeaccommodated globally,i.e.,inc0,or locally,inc1.Intheformercase,the utteranceisconstruedasmeaningthatBettyusedtodrinkandmaybe tryingtokickthehabit;inthelatter,itconveysthat,possibly, Bettyusedtodrinkandistryingtogiveupdrinking.Likewise,in (34b),thepresuppositionthatWilmaismarriedmaybeaccommodated globally,orlocallyinthemostdeeplyembeddedcontext.Buthere thereisathirdoption,aswell:ifthepresuppositionis accommodatedinc1,thesentenceisreadas “MaybeWilmaismarriedandshethinksthatherhusbandis havinganaffair”,andwespeakofintermediate accommodation. Itiswidelyagreedthatthefollowingempiricalgeneralization,made explicitbyHeim(1983),iscorrect: PGA:Globalaccommodationispreferredtonon-global accommodation. WhilethePGAwasinitiallybasedsolelyonintuitions,quantitative studiesonthespeedofprocessinglocalandglobalinterpretations havesinceprovidedempiricalsupportforit(e.g.ChemlaandBott 2013andRomoliandSchwartz2014). Intheexamplesin(34)thePGA(preferenceforglobalaccommodation) clearlyholds:non-globalinterpretationsmaybepossible,butthey requirespecialcontexts.Onesuchcontextmaybethatthe presuppositioncontainsavariablewhichisboundbyaquantifier: (35) MostGermanswashtheircarsonSaturday. In(35),withthepossessivepresuppositiontrigger“their cars”,thereisaglobalcontext(outsideofthescopeof “most”),alocalcontextcorrespondingtothescopeofthe quantifier(occupiedbytheVP“washtheircarson Saturday”),andalsoanintermediatecontextintherestrictor ofthequantifier(occupiedby“Germans”).Themost naturalinterpretationofthissentencesurelyisthatmostGermans whoownacarwashitonSaturday.Sointhiscaseintermediate accommodationseemstobethepreferredoption,andthismightbe explained,followingvanderSandt(1992),bysupposingthatthe possessivepronouncontainsavariableboundbythequantifier. Thereareothercaseswhereintermediateaccommodationisvirtually impossible: (36) (c0)If(c1)Fredis comingtothereception,(c2)hemay (c3)bringhiswife. Itisquiteunlikelythatthismaybeconstrued,withintermediate accommodationinc1,whichistheantecedentof theconditional,as“IfFredismarriedandiscomingtothe reception,hemaybringhiswife.”Moregenerally,wedon’tknow ofanyclear-cutcases(i.e.,casesinwhichaccommodationisnot forcedbyindependentcontextualfactors)inwhichapresupposition triggeredintheconsequentofaconditionalisaccommodatedinthe antecedent. Thepictureisratherconfusing.Whileinsomecases,e.g.,(34b)or (35),intermediateaccommodationseemspossibleandsometimeseven preferred,inothercasesitdoesn’tseempossibleatall.Andthings getevenmoreconfusedthanthis.Thusfar,wehavetakenourexamples atfacevalue,butsomeauthorshavearguedthatweshouldn’tbecause, asamatteroffact,intermediateaccommodationdoesn’texist.For instance,accordingtoBeaver(2001),thepresuppositionin(35)is takenasevidencethatthetopicofconversationiscar-owning Germans,anditisthistopicthatrestrictsthedomainofthe quantifier,makingintermediateaccommodationredundant.Seealsovon Fintel(1995)andGeurtsandvanderSandt(1999)fordiscussion. ReturningtotheuncontestedPGA,letusaskhowitcanbeexplained. Heim(1982)wasnotexplicitaboutexactlyhowaccommodationshould workandwhatshouldbeaccommodated.Andoddlyenough,although virtuallyalltheoriesofpresuppositionprojectionacceptthatthe PGAholds,therehaven’tbeenthatmanyattemptsatexplanation.One candidateiswhatBeaver(2001)callsthe“Atlas Principle”,afterAtlas(1976): AP:Oneaccommodationalternativeispreferredto anotheriftheformeryieldsastrongermeaningthanthelatter(i.e., ifthefirstmeaningunilaterallyentailsthesecond). AdvocatesoftheAtlasPrincipleinclude,besidesitseponym,Yeom (1998),Zeevat(1999),andBlutner(2000).Onethingtonoteaboutthe AtlasPrincipleisthatitdoesnotnecessarilyvindicatethePGA acrosstheboard:whilethepredictionsmadebytheAtlasPrinciple willtendtocomplywiththePGA,theydon’thaveto.However,itis surprisinglydifficulttosaywhereexactlytheAtlasPrinciple deviatesfromthePGA,becausethisdependsonvariousextraneous factors;seeGeurts(2000)fordiscussion. AmajorworryabouttheAtlasPrincipleisthatitisadhoc. DespiteitssoothingresemblancetoGrice’ssecondQuantityMaxim,it isquestionablewhetherhearersgenerallypreferstronger interpretationstoweakerones.ThissuggeststhattheAtlasPrinciple maynotbejustifiedasaninstanceofamoregeneralconstraint,and isthereforestipulative.Thesadandsomewhatembarrassingtruth seemstobethat,thusfar,wedon’treallyknowwhythePGAshould hold. Onelastissuewewouldliketomentionisthataccommodationisn’t alwaysequallyeasy(orhard).Forexample: (37a) Fredislookingfortheperson. (37b) Wilmaispregnant,too. In(37a)thepresuppositiontriggeredby“theperson” clearlyrequiresasalientdiscoursereferenttohookonto;the sentencewouldbeinfelicitouswhenutteredoutoftheblue.Thesame goesforthepresuppositiontriggeredby“too”in(37b), viz.thatsomesalientpersondifferentfromWilmaispregnant.This sentencewouldbeverypeculiarwhenutteredoutoftheblue,orin anycontextwherenosalientpersonwasunderdiscussionwhocouldbe pregnant.Putotherwise:unlikethepresuppositionswehaveseenthus far,itisveryhardtodealwiththepresuppositionsof“the person”or“too”byaccommodation alone.[21] Whyshouldthisbeso? VanderSandt(1992)proposesthatpresuppositionswhosedescriptive contentisrelativelypoorarehardtoaccommodate.This generalizationisborneoutbypronouns,names,andsemantically attenuatedefiniteNPslike“theperson”.However,itis notveryclearwhat“relativelypoor”means.DefiniteNPs like“thewatermolecule”or“thelonely carpenter”don’tstrikeusasparticularlypoor,butmaybehard tointerpretbywayofaccommodation.Similarly,thepresuppositions associatedwith“too”maybequiterich,andnevertheless theyaregenerallyhardtoaccommodate.GeurtsandvanderSandt (2004)proposetoaccountforthelatterbyadoptingHeim’s(1992) ideathatthepresuppositiontriggeredby“too”containsa pronominalelement,whichblocksaccommodation,butthisproposalhas beencriticisedbyBeaverandZeevat(2007).Bethisasitmay,it seemsclearthatvanderSandt’sgeneralizationcannotbereadas biconditional:evenifpoorpresuppositionsaredifficultto accommodate,thereversedoesn’talwayshold. AnotherproblemwithvanderSandt’sgeneralizationisthatitisnot clearhowitcanbejustified.Evenifitsomehowmakessensethat lackofdescriptivecontentshouldmakeaccommodationhard,wewould liketoknowmorepreciselywhythisshouldbeso.Apossibleanswer tothisquestionmaybefoundintheworkondefinitesbyHawkins (1978),ClarkandMarshall(1981),Heim(1982),andothers.AsHeim observes,“accommodationinresponsetodefinitesisnot normallyamatterofaddingjusttheminimalamountofinformation thatwouldrestorefelicity.”(Heim1982,p.372)Itseems plausiblethatthisholdsforaccommodationgenerally,anditarguably followsfromthenatureofaccommodation.Ifapresuppositionistobe interpretedbywayofaccommodation,newinformationispresented asifitweregiven,andithasoftenbeenobservedthatthis willonlyworkiftheinformationisnotcontentiousorotherwise remarkable.Thatistosay,itshouldalwaysbepossibleto integratethenewinformationintothecommonground:ithas tobelinkedtowhatisalreadygiven.Thiswillbehardwith semanticallyattenuateNPslike“thething”orpronouns like“he”,butalsowithricherdefiniteswhosecontent cannotreadilybelinkedtoanythinginthecommonground. Nevertheless,itwillgenerallybeeasywithricherpresuppositions, notbecausetheyhavemorecontent,butsimplybecausethey aremorelikelytocontainanchorsintothecommonground. 5.2AccommodatingPresuppositionsinSatisfactionFramework Localsatisfactionisattheheartofdynamictheoriesof presupposition,butwe’vealreadyseenthatitfacessomeempirical challengesinaccountingforconditionalizedpresuppositions. Specifically,sometimessatisfactiontheoriesgenerateweaker conditionalpresuppositionswhenintuitivelythespeakeris presupposingastronger,unconditionalizedpresupposition.Geurts (1996,1999a)deemsthis"TheProvisoProblem,"andthedata surroundingthisproblemhasleftquiteamessforproponentsof satisfactiontheoriestoclean up.[22] Accordingtosatisfactiontheories,onlyweakconditional presuppositionsshouldfollowfromembeddingpresuppositionsin certainconstructions,notablyconditionalsthemselves.Forexample, imaginethatyouareabouttogosurfingforthefirsttimeinthe coldwatersofNorthernCalifornia,andyouandyourinterlocutor suspectthatacommonacquaintance,Theo,isascubadiver,andthat hemaybeabletohelp.Inthiscontext,theuseofhisin (38a)intuitivelyleadstotheconditionalinferencein(38c): (38a) IfTheoisascubadiver,thenhewillbringhiswetsuit. (38b) Theohasawetsuit. (38c) IfTheoisascubadiver,thenhehasawetsuit. Contrastthiswith(39a),whichstronglysuggeststhatthespeaker believesthatTheohasawetsuit(39b),andnottheweaker conditionalin(39c): (39a) IfTheoisinagenerousmood,thenhewillbringhiswetsuit. (39b) Theohasawetsuit (39c) IfTheoisinagenerousmood,thenhehasawetsuit. Whatevermechanismisusedtodeterminewhatisaccommodated,itmust ensurethatsometimesconditionalpresuppositionsarestrengthened(to becomeunconditionalized),andsometimestheyaren’t.Whatexamples like(38)and(39)suggestisthatthismightbeexplainedintermsof plausibility:itcouldbethattheconditionalin(39c)is strengthenedbecauseitseemsimplausiblethataspeakerwouldbe assumingthatTheo’sgenerositywaslinkedtohisownershipofaquatic paraphernalia,whereasitseemsmoreplausiblethataspeakerwould simplybeassumingthatTheoownedsuchequipment. Severalproposals(e.g.Beaver2001,Singh2007,2009;Schlenker2011) trytomotivatethiskindofstrengtheningmechanismtoexplain when exactlyaconditionalizedpresuppositionshouldbe strengthened.Forexample,Beaver’smodelofaccommodation,first publishedinBeaver(1992)treatsaccommodationasatypeoffiltering operation.Beaversuggeststhatduetouncertaintyaboutwhatthe speakertakesthecommongroundtobe,thehearerhastoentertain multiplealternativecontextsets,withsomerankingofwhichisthe mostplausible.Allthesealternativecontextsareupdated simultaneously.Accommodationisthenwhathappenswhentheupdateis notdefinedonwhatwaspreviouslyconsideredtobethemostplausible context,inwhichcasethehearerdropsthatcontextfromcontention. Whatremainsisanewsetofcontextsinwhichthemostplausibleone isacontextthathasbeensuccessfullyupdated. Unfortunatelyforthisstyleofexplanationofpresupposition strengthening,plausibilityappearstohaveitslimits.Considerthe twoexamplesin(40a,b).Ifweassumethatacleft"itisX thatY-ed"presupposesanexistential,roughly "someone/somethingY-ed"theninadynamictheoryof presupposition(40b)willgeneratethesameconditionalized presuppositionasresultsfrom(40a),whichinvolvesthefactive "knows",namelytheconditionalin(40c).However,thereisaclear differenceintheinferenceswewouldtendtodrawfromutterancesof (40a)and(40b).Whereas(40a)licensesaninferenceonlytothe conditionalpresuppositionin(40c),example(40b)leadstothe stronger,unconditionalizedpresuppositionin(40d).Underthe assumptionswehavemade,namelythat(40a)and(40b)generateexactly thesameunderlyingconditionalizedpresupposition,theplausibility approachusedinBeaver’sframeworkfailstopredictadifferencein whatisaccommodatedinthesecases. (40a) Peterknowsthatiftheproblemwaseasy,someonesolvedit. (40b) IftheproblemwaseasythenitwasKristiewhosolvedit. (40c) Iftheproblemwaseasy,someonesolvedit. [Conditionalpresuppositionfor(40a,b),accommodatedfor(40a)] (40d) Someonesolvedtheproblem. [Unconditionalizedpresupposition,accommodatedfor(40b)butnot (40a)] Onerelevantfactorherethatmightinfluenceaccomodationis informationstructure.In(40b),thecleftstructuremightbesaidto reflectthepresenceofaQuestionUnderDiscussion(QUD),in thesenseofRoberts(2012),concerningwhosolvedtheproblem.Tothe extentthatsuchaquestionisunderdiscussion,theinterlocutors willpresumablyknowthatthisisso.Andif"Whosolvedtheproblem?" isindeedthequestiontheinterlocutorsaretryingtoanswer,then theyarelikelytomutuallyacceptthatsomeonesolvedtheproblem, thusexplainingprojection.Crucially,thislineofreasoningdepends onthecleftin(40b)havingaparticularinformationstructural function,andasimilarargumentcouldnotbemadeinacaselike (39a),wherethereisnoconstructionalmarkingofinformation structure,andinparticularnoindicatorthat"Whosolvedthe problem?"istheQUD. (41a) IftheproblemwaseasythenPeterknowsthatsomeonesolvedit. (41b) IftheproblemwaseasythenPeterKNOWSthatsomeonesolvedit. (41c) IftheproblemwaseasythenevenPETERknowsthatsomeonesolvedit. Wemayhavereasonablemotivationsforwhenwemightexpecta conditionalpresuppositiontobestrengthened.Butevenwithsuch mechanisms,satisfactiontheoriesstillfallshortwhenconfronting certaindata.Forexample,consider(42)below. (42) IfwhenSamgetsbackfromvacationhefindsmoldyfruitinthe kitchen,he’llbeannoyedthathehadfoodinthehouse. Here,theconsequentpresupposesthatSamhadfoodinthehouse,which isalsoentailedbytheantecedent.Sincesatisfactiontheoriesassume thatpresuppositionsinacomplexsentencewillbesatisfiedlocally, we’releftwithatautology.So,thelocalsatisfactionconstraint generatespresuppositionswhichareredundant. Theprovisoproblemisstillarelevanthurdleforproponentsof modernsatisfactiontheories.Mandelkern(2016)presentsaversionof satisfactiontheorydesignedtoavoidtheprovisoproblementirely. Romolietal(2011)explorethisissueexperimentallybyaskingwhich presupposition-theconditionalorunconditionaloption-ismore basic.But,ultimatelytheprovisoproblemislessproblematicfor otheraccountsofpresupposition,whichdon’trelyonlocalcontext updates. 5.3ResolvingandAccommodatingPresuppositionsinDRT Inordertoaccountforbothpresuppositionprojectionfactsandthe parallelsbetweenpresuppositionandanaphora,vanderSandtproposed anaccommodation-basedmodelasanextensionofDiscourse RepresentationTheory(seetheSEPentryon discourserepresentationtheory foranintroduction).Presupposedinformationisinformationthatis presentedasgiven,andinvanderSandt’stheorythismeansthat presuppositionswanttohavediscoursereferentstobindto.However, whereasanaphoricpronounsarerarelyinterpretableintheabsenceof asuitableantecedent,thesamedoesnotholdforall presupposition-inducingexpressions.Forinstance,aspeakermay felicitouslyassertthathemet“Fred’ssister”evenifhe knowsfullwellthathisaudienceisn’tawarethatFredhasasister. Insuchcases,presuppositionsaregenerallyaccommodated,whichisto saythattheheareracceptstheinformationasgiven,andreviseshis representationofthecontextaccordingly.Accommodation,thus understood,isaformofexploitationinGrice’ssense:thepurposeof presuppositionalexpressionsistosignalthatthisorthat informationisgiven,andifsomeinformationisnewbutnot particularlyinterestingorcontroversial(likethefactthatsomebody hasasister)thespeakermaychooseto“getitoutofthe way”bypresuppositionalmeans. VanderSandt’stheoryincorporatesthenotionofaccommodationas follows.Presuppositions,accordingvanderSandt,introduce informationthatpreferstobelinkedtodiscoursereferentsthatare alreadyavailableinthehearer’srepresentationofthediscourse,and inthisrespecttheyarelikepronouns.VanderSandtinfactusesthe term"binding"torefertoconfigurationsinwhichpresuppositions haveantecedentsintheDiscourseRepresentationStructure(DRS),thus generalizingthestandardnotionofaboundpronountocasesinvolving multiplediscoursereferents.However,ifasuitablediscourse antecedentisnotavailable,anewonewillbeaccommodated,andthe presuppositionislinkedtothat.Generallyspeaking,accommodationis notanoptionintheinterpretationofpronouns,andonereasonthat hasbeensuggestedforthisisthatapronoun’sdescriptivecontentis relativelypoor(seeSection5.1fordiscussion).Beingtoldthat “she”iswonderfulisnotparticularlyhelpfulifitisn’t clearwhothepronounismeanttoreferto.Bycontrast,ifthe speakerrefersto“Pedro’ssister”thereismoretogoon, andaccommodationbecomesfeasible.Hence,vanderSandthypothesizes thatpronounsareaspecialclassofpresuppositionalexpressions: whileallpresuppositiontriggersprefertobelinkedtoantecedents, pronounsalmostalwaysmustbelinkedtoantecedentsbecausetheyare descriptivelyattenuated,andthereforecannotbeconstruedbywayof accommodation. Togetabetterideahowthisissupposedtowork,letusconsideran examplewithseveralpresuppositiontriggers: (43) IfFredisprotesting,thenhissonisprotesting,too. ThissentencecontainsthedefiniteNP“hisson”,whichin itsturncontainsthepronoun“his”,andthefocus particle“too”.Assumingthepronoun’santecedentis “Fred”,thedefiniteNPtriggersthepresuppositionthat Fredhasason,whilethefocusparticletriggersthepresupposition thatsomeoneotherthanFred’ssonisprotesting.Notethatinthis examplethepresuppositiontriggeredbythedefiniteNPis “inherited”bythesentenceasawhole,whiletheone triggeredby“too”isnot:normallyspeaking,anutterance of(43)wouldlicensetheinferencethat(accordingtothespeaker) Fredhasason,butnotthatsomeoneelsebesidesFred’ssonis protesting. VanderSandt’stheoryaccountsfortheseobservationsasfollows.We supposethatthegrammarassigns(43)theintermediatesemantic representationin(44a).Here[u1,…,um: φ1,…,φn]isasimple DiscourseRepresentationStructureinlinearform,withu1,…, umalistofdiscoursemarkers,and φ1,…,φnalistof conditionsonthosemarkers;connectiveslike⇒areusedtobuild upcomplexconditions.Weassumeforconveniencethatmost interpretativeproblemshavebeenclearedoutofthewayalready,and thattheonlythingthatremainstobedoneisresolvethe presuppositionstriggeredby“his”,“hisson” and“too”,whichareflaggedbyBeaver’s(1992)trigger symbol,∂. (44a) [x:Fred(x),[:protesting(x)]⇒[:∂[z:∂[y:],z isy’sson],protesting(z),∂[u:u≠z,protesting(u)]]] …bindingytox… (44b) [x:Fred(x),[:protesting(x)]⇒[:∂[z:zisx’s son],protesting(z),∂[u:u≠z,protesting(u)]]] …accommodatingx… (44c) [x,z:Fred(x),zisx’sson,[:protesting(x)]⇒[: protesting(z),∂[u:u≠z,protesting(u)]]] …bindingutox… (44d) [x,z:Fred(x),zisx’sson,[:protesting(x)]⇒[: protesting(z),x≠z]] (44a)istheinitialsemanticrepresentationassociatedwith(43),in whichthreepresuppositionsremaintoberesolved.Thefirstofthese, triggeredbythepronoun“his”,isboundtothediscourse referentrepresentingFred,whichresultsin(44b).Thesecond presupposition,thatFredhasason,cannotbebound,andtherefore mustbeinterpretedbywayofaccommodation.VanderSandt’stheory, likeHeim’s(Heim,1983),stipulatesthataccommodationattheglobal level,asshownin(44c),ispreferredtoaccommodationatother sites.Finally,thepresuppositiontriggeredbythefocusparticlecan beboundintheantecedentoftheconditional;aftersimplification, thisresultsin(44d),whichrepresentsthemostnaturalwayof interpreting(43). 6.Borntofail,orunborn? Whathappenswhenapresuppositionisfalse?Thetextbook proto-historyofanswerstothisquestiongoesasfollows.According toFrege(1892),ifanexpressionAsuffersfrom presuppositionfailure,thenanysentencecontainingAwill lackatruthvalue;Russell(1905)famouslydeniedthis,holdingthat suchasentencewillalwaysbetrueorfalse;andthenStrawson(1950) reaffirmedFrege’sposition,moreorless.Whatislesswellknown,at leastinsofarasitusuallydoesn’tmakeitintointroductoryclasses onphilosophyoflanguage,isthatinsubsequentwork,Strawsonpartly recantedhisinitialviewandcametodoubtthatpresupposition failureinvariablyentailslackoftruthvalue. Takingacloserlookathowspeakersactuallyassessasentence, Strawson’s(1964)paperarguesthatpresuppositionfailuremaybut neednotcauseasentencetobeinfelicitous.Twoofhisexamplesare thefollowing: (45a) Jonesspentthemorningatthelocalswimmingpool. (45b) TheexhibitionwasvisitedyesterdaybythekingofFrance. Ifthereisnoswimmingpoollocally,itis“natural enough”,accordingtoStrawson,tosaythat(45a)isfalse,and sincethekingofFrancedoesn’texist,thesameappliesto(45b).And ifthesesentencesarefalse,theirnegationsmustbetrue.So,if thesesubtlejudgmentsareaccepted,therearecasesinwhich presuppositionfailuredoesnotpreventusfromsayingthatasentence istrueorfalse.ButStrawsonhasn’tchangedhismindaboutRussell’s example: Confrontedwiththeclassicalexample,“ThekingofFranceis bald”,wemaywellfeelitnaturaltosay,straightoff,that thequestionwhetherthestatementistrueorfalsedoesn’tarise becausethereisnokingofFrance.(Strawson1964:90) Strawsongoesontoobserve,however,thatspeakerswhosubscribeto thisjudgmentmaywanttoreconsidertheirverdictifthecontextis setuptherightway.Forinstance,ifRussell’ssentenceisusedto answerthequestion,“Whatexamples,ifany,arethereoffamous contemporaryfigureswhoarebald?”,wemaybemoreinclinedto saythattheanswerissimplyfalse. Strawson’sexplanationforthesefactsisgivenintermsof topicality.Themostlikelypurposeofasentencelike(45a)isto describewhatJoneshasbeendoinginthemorning,ratherthan,say, whothelocalswimmingpoolwasvisitedby.Thatis,intheabsenceof furtherinformationaboutthecontextinwhichthissentenceis uttered,itstopicwillbeJones’sexploits.Similarly,asentence like(45b)willnormallybeusedtoconveyinformationaboutthe exhibition.Ifso,althoughthesentencepurportstorefertotheking ofFrance,itisnotabouthim;thekingofFranceisnotthetopicof discourse,norpartofthetopic.Inotherwords,theexistenceofthe kingofFranceisnotat-issue.Strawson’ssuggestionisthat thiscircumstanceinfluencesthewaypresuppositionfailureisdealt with.Onthisview,presuppositionfailureresultsininfelicityonly ifitaffectsthetopicofasentence;otherwisethesentencewillbe judgedtrueorfalse,asappropriate. Oneoftheappealingfeaturesofthisanalysisisthatittakesinto accountthecontext-dependenceofspeakers’intuitions.AsStrawson notes,Russell’ssentence(46)willbydefaultbeconstruedasbeing aboutthekingofFrance,whenceastrongtendencytojudgethe sentenceinfelicitous. (46) ThekingofFranceisbald. If,however,thediscourseisaboutroyalbaldnessingeneral,for instance,thegrammaticalsubjectof(46)isusedtosaysomething aboutthattopic,andStrawson’saccountpredictsthatthesentenceis morelikelytobejudgedfalse,whichseemscorrect.Another observationthatneatlyfallsintoplaceisthatwordordermayhave aneffectonspeakers’intuitionsaboutpresuppositionfailure.As Strawsonobserves,ifwecompare(45b)with(47),wherethedefective descriptionisinsubjectposition,wewouldbe“ashademore squeamish”tosaythatthesentenceissimplyfalse(p.91). Thisispreciselywhatoneshouldexpectifspeakers’intuitionswere topic-dependent. (47) ThekingofFrancevisitedtheexhibitionyesterday. AssumingthatStrawson’sobservationsarecorrect,shouldwesay(a) thatnon-topicaldefinitesarenon-presuppositional,or(b)thatthey dohavepresuppositions,whosefailurehappensnottoaffectspeakers’ truth-valuejudgments?Someauthorsarguefortheformer(e.g., Reinhart1982,Horn1989);thisisStrawson’sview,aswell.Von Fintel(2004)arguesforthelatter:topicalornon-topical, “thekingofFrance”alwaystriggersthepresupposition thatthereisakingofFrance;it’sjustourtruth-valuejudgments thatfluctuate. VonFintel’spositionisinlinewithwhathasbeenaworking hypothesisformanyyearsintheoriesofprojection.Accordingto thesetheories,presuppositionsareneverreallycancelled:ifa presuppositionseemstodisappear,itisbecauseitprojectstoa non-globalcontext.Itbearsemphasizingthatthisviewisnot irrevocablylinkedtothesetheories.It’sjustthatthereisnoneed toassumethattherearepart-timepresuppositiontriggers(inasense weexpanduponinSection8below),becausenon-globalprojectioncan bereliedupontoexplainwhypresuppositionssometimesseemto disappear.However,thisworkinghypothesismayhavetobereassessed ifwewanttoaccountforStrawson’sobservations,whichmightbe takentoshowthatpresuppositionsarecancellablenotonlyincomplex sentencesbutinsimplesentences,aswell. Sofarwe’veonlytalkedaboutdefinitedescriptions.Someother triggersshowsimilarvariability,butnotallofthemdo.Ontheone hand,thedomainpresuppositionsassociatedwithstrongquantifiers like“all”and“most”behaveverymuchlike definitepresuppositions(deJongandVerkuyl1985,Lappinand Reinhart1988,Geurts2007).Forexample,whenutteredoutofthe blue,(48)willsoundodd,butwhenprofferedinresponseto Strawson’squestion,“Whatexamples,ifany,arethereoffamous contemporaryfigureswhoarebald?”,itseemsmorelikelytobe judgedfalse: (48) AllSouth-Americanmonarchsarebald. Ontheotherhand,considerthefollowing: (49a) #BENEDICTXVIistheincumbentpope,too. (49b) #CarnapmanagedtobebornonMay18,1891. Giventhatthereisonlyoneincumbentpopeandthatitcan’thave takenCarnapanyefforttobebornonMay18,1891,both(49a)and (49b)sufferfrompresuppositionfailure.Butifsomeonewhoknewthat therewasonlyoneincumbentpopeuttered(49a),itwouldnotseem entirelynaturaltosaythattheyhadlied.Onemightevensaythat thesesentencesaretrue,andthisintuitiondoesnotseemto betopicdependent.Thereasonforthismaybethatthe presuppositionsinquestionareincidentaltotheprimary contentofthesesentences. 7.Presuppositionsandattitudes Fornearlyfourdecades,theHolyGrailofpresuppositionresearchhas beentoexplainthebehaviorofpresuppositionalexpressionsoccurring inembeddedpositions.Aparticularlychallengingmodeofembedding ariseswithattitudeverbs.Ifweembedapresuppositiontrigger under,e.g.,“believe”,weobservetwotypesofinference. Thisisseeninthefollowingexamples,where“→” indicatesthatthereisaninference,butofunspecifiedtype: (50) Barneybelievesthathissisterisdrunk. →Barneyhasasister. →Barneybelieveshehasasister. (51) WilmabelievesthatBarneyknowsthathissisteris drunk. →Barney’ssisterisdrunk. →WilmabelievesthatBarney’ssisterisdrunk. Bothinferencesarefairlyrobust,andbothseemtoexhibitprojection behavior,asweillustrateherebyapplyingvariousembeddingteststo (50): (52a) Barneydoesn’tbelievethathissisterisdrunk. (52b) PerhapsBarneybelievesthathissisterisdrunk. (52c) IfBarneyhasasister,thenhebelievesthathissisteris drunk. (52d) IfBarneybelievesthathehasasister,thenhealsobelieves thathissisterisdrunk. Itisnaturaltoinferfromboth(52a)and(52b)thatBarneyhasa sisterandthathebelievesthathehasasister,andtheseinferences canbeblockedinsentencessuchasthosein(52c)and(52d),where “hissister”occursintheconsequentofaconditional whoseantecedentmakesoneortheotherinferenceexplicit.Itmay seemoddatfirstthat(52c)shouldblocktheinferencethatBarney believesthathehasasister,while(52d)blockstheinferencethat Barneyhasasister.Notehoweverthat,generallyspeaking,Barneyhas asisteriffhebelievesthathehasasister.We’llreturn tothisthemebelow. Interestingly,literallythesamepatternofobservationsholdsfor otherattitudeverbs,like“want”forexample: (53) Barneywantstophonehissister. →Barneyhasasister. →Barneybelieveshehasasister. Thepuzzlingthingisthat(53)doesnotlicensetheinferencethat Barneywantstohaveasister,butratherthathebelieves thathehasone. So,inmanycasesatleast,apresuppositionφtriggeredwithin thescopeof“x VA…”,where VAisanattitudeverb,givesrisetotwo inferenceswithacandidatureforpresuppositionalstatus:(a)that φand(b)thatxbelievesφ.Hence,wehavethree possiblewaysofproceeding,allofwhichhavebeendefendedinthe literature: Bothinferencesarepresuppositions(Zeevat1992,atleastfor certaintriggers). Onlyφisapresupposition(Gazdar1979a;vanderSandt1988; Geurts1998). Only“xbelievesthatφ”isa presupposition(Karttunen1974;Heim1992). Thefirststrategyseemstorequireastipulationthatcertain presuppositionshavetobeaccommodatedtwice,onceinsideabelief context,andonceoutside,andsuchastrategyisdifficultto motivate.Ontheotherhand,for(ii)and(iii), thereistheobviousproblemthatifweadopteitheroneofthem,we onlyaccountforhalfoftheobservedinferences.Howtoexplainthe otherhalf?Threepossibleanswerstothisquestionhavebeen discussedintheliterature,mainlybyHeim(1992): Dereconstrual(Heim1992)Whatatfirstlookslikea presuppositionprojectingtotheglobalcontextmayinfactbedueto adereconstrualofthepresuppositiontrigger.This solutionhasseveralseriousdrawbacks.Inparticular,it’shardto seehowthisproposalcangiveusbothinferencesatthesametime, ratherthanone(dere)ortheother(presupposition).For thisandotherreasons(seeGeurts1998),wewillnotconsideritany further. Exportation(Karttunen1974,Heim1992,Kay1992)If Barneybelievesthathehasasister,thenitmayplausiblybe inferredthathehasasister.Therefore,ifitispresupposedthat Barneybelievesthathehasasister,thenitmayplausiblybe inferredthathehasasister. Importation(Heim1992,Geurts1998)IfBarneyhasa sister,thenitmayplausiblybeinferredthathebelievesthathehas asister.Therefore,ifitispresupposedthatBarneyhasasister, thenitmayplausiblybeinferredthathebelievesthathehasa sister. Ifourpresuppositiontheorypredictsthattheinferencesprojected from“xVA”areof theform“xbelievesthatφ”,thenwecan appealtoexportationtoexplainwhyφisinferable,aswell. Viceversa,atheorywhichpredictsthatφispresupposed canuseimportationforderiving“xbelievesthat φ”.Sowehavetwooptions:   xbelievesthatφ φ OptionA importation presupposition OptionB presupposition exportation Whichisitgoingtobe?That’sahardquestion,whichraisesvarious issues,onlysomeofwhichwecanmentionhere.First,itshouldbe notedthat,whereasfortheoriesofthesatisfactionfamilyitishard toavoidmakingthepredictionthatpresuppositionsprojected fromattitudecontextsareoftheform“xbelievesthat φ”,DRT-styletheoriesaremoreflexible,andcanpredict eitherthisorthatthepresuppositionissimplyφ.Inother words,satisfactiontheoriesaremoreconstrained(whichisgood),but thereforepracticallyforcedtoresorttoOptionB. Oneoftheissuesthatneedtobeaddressedisthatofthetruenature ofimportationandexportationinferences.Itseemsreasonableto assumethat,generallyspeaking,people’sbeliefsareconsistentwith thefacts,andthatwetendtoassumebydefaultthatthisisthe case.Butevenifthismuchistrue,itismostunlikelythat,by default,peoplewillinferφfrom“xbelieves φ”(forarbitraryxandφ),orviceversa. Whateverimportationandexportationare,theyaren’tgeneral heuristics;rather,theseinferencesarealmostcertainlydependenton thecontexttosomedegree. 8.Presuppositionvariability Muchcontemporaryworkonpresuppositionhasbeendevotedtocase studiesofparticulartriggersorfamiliesoftriggers.Moreoften thannot,thesestudiesendupgrantingsomesortofspecial considerationtothetrigger(s)inquestion.Wehavealreadyseen instancesoftriggerdifferentiationabove; forexample, anaphoricaccountsidentifyaclassoftriggerswhosecontentis "descriptivelypoor"toexplainwhytheirpresuppositionsaremore difficulttoaccommodatethanthoseofothertriggers(seeSections 5.1&5.3).Researchershavenotedmoreandmorewaysinwhichthe presuppositionsoftriggersdiffer,mostlyfocusedaroundquestionsof howrobustlythepresuppositionprojectsandinwhatsortsofcontexts thepresuppositiondisappears. Forsome,theblurryboundariesbetweenthesortsofinferencesthat arecoveredbypresuppositiontheorieshaveopenedupthedoorfornew triggersandpresupposition-likethingsthatmightnotbecoveredby traditionallabels.Forexample,Schlenker(2015,2018)appliesa presuppositionalframeworktoiconicco-speechgestures,toidentify whathecalls"cosuppositions."Therearealsoinferencesthatseemto challengetraditionaldiagnosticslikeprojection.Forinstance, non-restrictiverelativeclauses("Fred,whoIlike,...")produce projectionlikepresuppositions,buttypicallyintroducenew informationtothediscourse(Chierchia&McConnell-Ginet 1990). Finally,thewaythatsometheoriesmodelpresuppositionscangive risetonewtypesofrelatedinferences.Forexample,Chemla(2008) notesthatthesentencein(54a)stronglysuggeststhatthespeaker doesnothaveasister.AccordingtoHeim(1991),thisinference arisesfromthefactthatthespeakerdidnotutterthealternativein (54b),whichstronglypresupposesthatthespeakerdoeshaveasister. ThispredictionfollowsfromHeim’sMaximizePresupposition! principle,thataspeakerwillusethefelicitoussentencewiththe strongestpresuppositionamongasetofalternatives.Thesetypesof inferenceshavebeenreferredtointheliteratureas ’anti-presuppositions,’atermoriginatingwithPercus(2006). (54a)JohnbelievesIhaveasister. (54b)JohnknowsIhaveasister. Oneobserveddifferencebetweensometriggersisthattheydon’tseem toreliablygeneratepresuppositionsincertaincontexts.Aswe mentionedinSection6,theoriesofpresuppositionprojectiontendto adopttheworkinghypothesisthatpresuppositionsareassociatedwith expressiontypes:ifanexpressionαtriggersapresupposition φ,thenφwillalwaysbetriggeredbyα.Thisisjusta matterofeconomy:sinceaprojectiontheoryalreadyprovidesa mechanismforexplaininghowpresuppositions,oncetriggered,canseem todisappear,thereisnoneedforsupposingthattherearepart-time presuppositiontriggers.Indeed,itwouldbeabitofanuisanceifit turnedoutthatthatsomeexpressionsorconstructionstriggertheir presuppositionsonapart-timebasis,becausethenwewouldhavetwo waysofexplainingwhyagivenpresuppositionfailstoappear:echoing ourdiscussioninsection6,eitherithasbeencanceled(saythrough beingresolvedinanembeddedcontext)oritwasn’ttriggeredinthe firstplace. Aretherecompellingreasonsforbelievingthattherearepart-time triggers?Notasfarasweknow.Therearesuggestivefacts,though. InthefollowingexamplefromFauconnier(1985),Aand Baresittinginabarobservingastrangerwhodoesn’tseem tobetoohappy: (55) A:Iwonderwhythatguyislookingsoglum. B:Maybehisgirlfriendjiltedhim. ThemostnaturalinterpretationofB’sanswer,inthis context,isthatitispossiblethatthestrangerhasagirlfriendwho hasjiltedhim;thisconstrualwillrequirelocalaccommodationofthe girlfriendifweassumethat“hisgirlfriend”always triggersthepresuppositionthatthepronoun’sreferenthasa girlfriend.[23] Theproblemisthatifthisislocalaccommodation,thenit’stoo easy.Localaccommodationissupposedtobedispreferred.If weallowthatinthiscasethepresuppositionislocallyaccommodated, whycan’tallembeddedpresuppositionsjustbelocallyaccommodated awayandnotproject?Allowinglocalaccommodationtooccurwithout constraintwouldpotentiallydestroyanyclaimtohaveexplainedthe observationthathasdefinedthephenomenonofpresuppositionsince itsinception,theobservationthatpresuppositionstendto project. Ifwecouldexplainwhythisorthatexpressiontriggerssuch andsuchapresupposition,oneofthecorollariesmightbethatthe expressioninquestionisapart-timetrigger.Toseehowsuchan argumentmightruninoutline,considerthepresuppositionsassociated withaspectualverbslike“start”or “stop”: (56a) Jillhasstartedsmokingpot. →Jilldidn’tusetosmokepot. (56b) Jillhasstoppedsmokingpot. →Jillusedtosmokepot. Aspectualverbsdescribeatransitionfromonestatetoanother,and itisalwaysthefirststatethatispresupposed,asstandard projectiontestsconfirm.Anexplanationmightbesoughtintheidea thatinterlocutorsaremoreinterestedinwherethestoryisgoing thanwhereitcamefrom,andthereforetendtotakethepastasgiven. Apurelypragmaticanalysismightinvolvethehearerreasoningthat sincethespeakerislikelytointendtoprovidetheinformationthat isofgreatestinterest,theyareprobablyassumingacontextinwhich thepre-stateassociatedwithatelicsentenceisalreadyinthe commonground,i.e.presupposed.Itmightfurtherbearguedthatsuch reasoningonlyapplieswhenthehearerhasreasontothinkthatthe speakerhasexpertiseonthesituationbeingdescribed.Apart-time triggeranalysismightbebasedontheideathattheregularitywith whichpre-statesoftelicsentencesareinthecommongroundhasled tothispragmaticreasoningbeingconventionalized.Apart-time triggeranalysismightinvolveaspecialtriggeringcondition,e.g. involvingthespeaker’sexpertise.Sothepart-timetriggeranalysis mightamounttotheclaimthatwhenthecontextdeterminesthatthe speakerhasexpertiseaboutanevent,thepre-stateispresupposed, butwhenthecontextdeterminesthatthespeakerlacksexpertise, thereisnosuchpresupposition. Thedisappearanceofapre-statepresuppositioninexample(57)could beexplainedinthisway.Onamorestandardview,theaspectualverb triggersaquantifiedpresupposition,roughlythatxusedto smoke.Thishastobeaccommodatedlocally,perhapsbecauseas suggestedbyvanderSandt(1992),boundpresuppositionscanneverbe accommodatedoutsideofthescopeofthebinder,whathecalls trapping.Butitisalsoconceivablethatthereisno presuppositiontobetrapped.Theideawouldbethatthespeakerlacks expertiseonthehabitsofthearbitraryindividualsquantifiedover bythequantifier"anyone",andthatthisisjustthesituationin whichapresuppositionisnottriggered.Theeasewithwhichthe existencepresuppositionofthepossessiveinexample(55)disappears couldbeexplainedalongthesamelines. (57) IfanyonestopssmokingbeforeJuly1,theyareeligiblefora paymentfromtheTobaccoIndemnityFund.(Exampleadaptedfrom Abusch2002.) Adifferentperspectivehasemergedinrecentyearstoexplainthewhy certaintriggersseemtoonlyresultinapresuppositionpartofthe time.Ratherthanproposingthatthereisvariationintriggering, somelinguistsinterpretthisbehaviorastheresultofvariationin theinherentcancellabilityamongtriggers.Abusch(2002,2010) distinguishesbetween‘soft’triggerswhichareweakand context-dependentand‘hard’triggers,whichareresistentto cancellation.Forexample,sheclassifiestheachievementverb win,in(58a),asasofttrigger,sincethepresupposition thatJohnparticipatedintheraceiscontextuallydefeasible,whereas thecleftconstructionin(58b)isconsideredahardtriggerbecause itcannotbecancelledbythespeakerbyexplicitlyexpressing ignorance. (58a) IhavenoideawhetherJohnendedupparticipatingintheRoad Raceyesterday.Butifhewonit,thenhehasmorevictoriesthan anyoneelseinhistory. (58b) ??Ihavenoideawhetheranyonereadthatletter.Butifitis Johnwhoreadit,let’saskhimtobediscreetaboutthecontent. Abusch’ssoft/harddistinctionsuggeststhatdegreesofcancellability areintrinsicpropertiesofthetriggersthemselves,andthosethat buyintothistheorymustmotivatewhythatis.ForAbusch, hardpresuppositionsarestrictlysemantic,whereassoft presuppositionsarisefromalternativesetsgeneratedinthecontext. Romoli(2011,2014)takesthisastepfurtherbyarguingthat presuppositionsfromsofttriggersaren’tpresuppositionsatall,but ratherentailmentsorscalarimplicatures.Incontrast,Abrusán (2011a,2011b,2016)arguesthatthedifferencebetweensoftandhard triggersemergesfromhowbothtypesinteractwiththecontext,for examplewithregardtotheirfocussensitivity. Despitetheargumentsmadeforadistinctionbetweensoftandhard triggers,empiricalstudieshavechallengedtheviewthattriggers differintheirconventionalproperties.Jayezetal(2014) find thatprojectiondifferencesbetweendifferenttriggers dependmoreoncontextthanontheclassificationofthetrigger itself.Schwarz(2014)comparesprocessingtimesbetweenhardandsoft triggersandfindsthatthereisnodifferencebetweenthetwo. Withtheriseofargumentsformorefine-grainedvariabilityamong presuppositions,inferencesthatwemayhaveconfidentlycalled ’presuppositions’havebecomehardertothinkofasforminganatural class.Forsome,theinconsistencyintriggerbehaviorhasfedinto somethingofapresuppositionidentitycrisis,acrisisdatingat leastasfarbackasBöerandLycan’s(1976)descriptionof semanticpresuppositionasa"myth".Karttunen(2016)reflectsonthe stateofthefieldwithanairofregret,sayingthatthe"veritable zoo"oftriggersthatwerelumpedintothesamepresuppositionalcage shouldneverhavebeenconsideredthesamespeciesinthefirstplace. Ultimately,inordertoeithersalvageaunifiednotionof presupposition,orelseargueinagrandtradition(followinge.g. BöerandLycan1976orKarttunenandPeters1977)forrejecting thepossibilityofaunifiednotionofpresupposition,ithasbecome pressingforlinguiststounderstandhowtriggersandtheirinferences varyandwhy,fuelingatrendinrecentyearstowardssystematic empiricalwork,lookingforevidenceofcross-linguisticvariation, anddevelopingmethodsforexperimentallytestingthepredictionsof differenttheories. Bibliography Abbott,Barbara,2000,“Presuppositionsas nonassertions”,JournalofPragmatics,32(10): 1419–1437. Abbott,Barbara,2008,“Issuesinthesemanticsand pragmaticsofdefinitedescriptionsinEnglish”,in Reference:InterdisciplinaryPerspectives,J.K.GundelandN. Hedberg(eds.),OxfordUniversityPress. Abbott,Barbara,2006,“Wherehavesomeofthe presuppositionsgone?”,inDrawingtheboundariesof meaning:Neo-Griceanstudiesinpragmaticsandsemanticsinhonorof LaurenceR.Horn,BettyBirnerandGregoryWard(eds.), Philadelphia:JohnBenjamins. Abrusan,Marta,2011a,"Predictingthepresuppositionofsoft triggers",LinguisticsandPhilosophy,34(6):491–535. Abrusan,Marta,2011b,"Anoteonquasi-presuppositionandfocus", JournalofSemantics,30(2),257–265. Abrusan,Marta,2016,"Presuppositioncancellation:explainingthe ’soft-hard’triggerdistinction",NaturalLanguageSemantics, 24(2):165–202. Abusch,Dorit,2002,“Lexicalalternativesasasourceof pragmaticpresuppositions”,inJackson,Brendan(ed.), ProceedingsofSALTXII,Ithaca,NY:CLCPublications, 1–20. Abusch,Dorit,2010,"Factivityinexclamativesisa presupposition",StudiaLinguistica,64(1):141–157. Abusch,DoritandRooth,Mats,2000,“Empty-domaineffects forpresuppositionalandnon-presuppositionaldeterminers”,To appearinHansKampandBarbaraPartee(eds),ContextDependencyin theAnalysisofLinguisticMeaning. Atlas,Jay,1976,OntheSemanticsofPresuppositionand Negation:anEssayinPhilosophicalLogicandtheFoundationsof Linguistics,Ph.D.thesis,PrincetonUniversity. Atlas,Jay,1977,“Negation,ambiguityand presupposition”,LinguisticsandPhilosophy,1: 321–336. Atlas,Jay,1979,“Howlinguisticsmatterstophilosophy: Presupposition,truth,andmeaning”,inOh,C.K.andDineen,D. A.(eds.),Presupposition,NewYork:AcademicPress,vol.11 ofSyntaxandSemantics,265–281. Atlas,JayandLevinson,Stephen,1981,“It-clefts, informativenessandlogicalform:Radicalpragmatics”,inCole, Peter(ed.),RadicalPragmatics,AcademicPress, 1–61. Beaver,David,1992,“Thekinematicsof presupposition”,inDekker,P.andStockhof,M.(eds.), ProceedingsoftheEighthAmsterdamColloquium,Universityof Amsterdam:ILLC,17–36. Beaver,David,1997,“Presupposition”,invanBenthem, J.andterMeulen,A.(eds.),TheHandbookofLogicand Language,Amsterdam:Elsevier,939–1008. Beaver,David,2001,PresuppositionandAssertioninDynamic Semantics,Stanford,CA:CSLIPublications. [Availableonline] Beaver,David,2008,“Asbriefaspossible(butno briefer)”,TheoreticalLinguistics,34: 213–228. Beaver,D.(2010)."HaveYouNoticedthatYourBellyButtonLint ColorisRelatedtotheColorofYourClothing?",inR.Bauerle,U. Reyle,andT.Zimmermann(eds.)Presuppositionsanddiscourse: EssaysofferedtoHansKamp,Brill,65–100. Beaver,DavidandCondoravdi,Cleo,2003,“Auniform analysisof”before“and”after“”,in ProceedingsofSemanticsandLinguisticTheoryXIII,Cornell, Ithaca:CLCPublications,37–54. Beaver,DavidandDenlinger,Kristin,2020,"Negationand Presupposition",inDeprez,V.andEspinal,M.T.(eds.),The OxfordHandbookofNegation,369–390. Beaver,DavidandKrahmer,Emiel,2001,“Apartialaccount ofpresuppositionprojection”,JournalofLogic,Language andInformation,10:147–182. Beaver,DavidandZeevat,Henk,2007,“Accommodation”, inRamchand,G.andReiss,C.(eds.),TheOxfordHandbookof LinguisticInterfaces,OxfordUniversityPress, 533–538. Blamey,Steven,1989,“Partiality”,inGabbay,D.and Guenther,F.(eds.),TheHandbookofPhilosophicalLogic, Dordrecht:Reidel,vol.IV,553–616. Blutner,Reinhard,2000,“Someaspectsofoptimalityin naturallanguageinterpretation”,JournalofSemantics, 17(3):189–216. Bochvar,DmitriiAnatoljevich,1939,“Obodnomtrehznachom iscisleniiiegoprimeneiikanalizuparadoksovklassicskogo rassirennogofunkcional‘nogoiscislenija’”,Matematiciskij Sbornik,4:87–112.EnglishTranslation:Historyand PhilosophyofLogic2,1981,“OnaThree-valuedCalculus andItsApplicationstotheAnalysisoftheParadoxesoftheClassical ExtendedFunctionalCalculus”. Böer,StevenE.andLycan,WilliamG.,1976,“Themyth ofsemanticpresupposition”,inZwicky,A.(ed.),Papersin nonphonology,Columbus,Ohio:DepartmentofLinguistics,Ohio StateUniversity,vol.21ofWorkingPapersinLinguistics, 1–90. Bridge,Derek,1991,ComputingPresuppositionsinan IncrementalNaturalLanguageProcessingSystem,Ph.D.thesis, UniversityofCambridgeComputerLaboratory.AvailableasTechnical ReportNo.237. Chemla,Emmanuel,2008,"Ananti-introductiontopresuppositions", inProceedingsofMIT-Franceworkshoponscalarimplicatureand presupposition,60. Chemla,EmmanuelandBott,Lewis,2013,"Processing presuppositions:Dynamicsemanticsvs.pragmaticenrichment", LanguageandCognitiveProcesses,28(3):241–260. Chierchia,Gennaro,1995,DynamicsofMeaning,Chicago: UniversityofChicagoPress. Chierchia,Gennaro&SallyMcConnell-Ginet.1990.Meaningand Grammar.Cambridge,Mass.:MITPress. Clark,HerbertH.andMarshall,C.,1981,“Definite referenceandmutualknowledge”,inJoshi,A.andWebber,B.L. andSag,I.(eds.),ElementsofDiscourseUnderstanding, CambridgeUniversityPress. Cooper,Robin,1983,QuantificationandSyntacticTheory, Dordrecht:Reidel. deJong,FranciskaandVerkuyl,Henk,1985,“Generalized quantifiers:thepropernessoftheirstrength”,interMeulen, A.andvanBenthem,J.(eds.),GeneralizedQuantifiers:Theoryand Applications,Dordrecht:Foris,vol.4ofGRASS, 21–43. Delin,Judy,1995,“Presuppositionandsharedknowledgein it-clefts”,LanguageandCognitiveProcesses,10(2): 97–120. vanEijck,Jan,1993,“Thedynamicsofdescription”, JournalofSemantics,10:239–267. vanEijck,Jan,1994,“Presuppositionfailure:acomedyof errors”,FormalAspectsofComputing,6(6): 766–787. vanEijck,Jan,1995,“Presuppositionsanddynamic logic”,inPapersfromtheSecondCSLIWorkshoponLogic, LanguageandComputation,MakotoKanazawa,Christopher Piñon,andHenriettedeSwart(eds.),Stanford:CSLI Publications. Fauconnier,Gilles,1985,MentalSpaces:AspectsofMeaning ConstructioninNaturalLanguage,Cambridge,Massachusetts:MIT Press. Fillmore,Charles,1971,“Verbsofjudging:Anexercisein semanticdescription”,inFillmore,C.andLangendoen,D.T. (eds.),StudiesinLinguisticSemantics,NewYork.:Holt, ReinhardtandWinston,272–289. vonFintel,Kai,1995,RestrictionsonQuantifier Domains,Ph.D.thesis,UniversityofMassachussettsat Amherst. vonFintel,Kai,2004,“Wouldyoubelieveit?Thekingof Franceisback!Presuppositionsandtruth-valueintuitions”,in Reimer,M.andBezuidenhout,A.(eds.),Descriptionsand Beyond,OxfordUniversityPress,269–296. vonFintel,Kai,2008,"What’sPresuppositionAccommodation, Again?",PhilosophicalPerspectives,22:137–170. Fraassen,BasC.van,1968,“Presupposition,implication, andself-reference”,TheJournalofPhilosophy,65(5): 136–152. Frege,Gottlob,1892,“UberSinnundBedeutung”, ZeitschriftfurPhilosophieundphilosophischeKritik,C: 25–50.EnglishTranslation:“OnSenseandMeaning”, inMcGuinness,B.(ed),Frege:collectedworks,Oxford:Basil Blackwell,157–177. Gamut,L.T.F.,1991,Logic,LanguageandMeaning,vol. 1,UniversityofChicagoPress. Gawron,JeanMark,1995,“Quantification,quantificational domainsanddynamiclogic”,inHandbookofContemporary SemanticTheory,ShalomLappin(ed.),Oxford:Blackwell, 247–267. Gazdar,Gerald,1979a,Pragmatics:Implicature,Presupposition andLogicalForm,NewYork:AcademicPress. Gazdar,Gerald,1979b,“Asolutiontotheprojection problem”,inOh,C.-K.andDineen,D.(eds.),Syntaxand Semantics11:Presupposition,NewYork:AcademicPress, 57–89. George,BenjaminR.,2008,PresuppositionRepairs:aStatic, TrivalentApproachtoPredictingProjection,Master’sthesis, UCLA. Geurts,Bart,1996,“Localsatisfactionguaranteed”, LinguisticsandPhilosophy,19:259–294. Geurts,Bart,1998,“Presuppositionsandanaphorsin attitudecontexts”,LinguisticsandPhilosophy,21: 545–601. Geurts,Bart,1999a,PresuppositionsandPronouns,vol.3 ofCRiSPI,Amsterdam:Elsevier. [Availableonline] Geurts,Bart,1999b,“Specifics”,inGeurts,B.and Krifka,M.andvanderSandt,R.(eds.),Focusandpresupposition inmulti-speakerdiscourse,Utrecht:ESSLLI99, 99–129. Geurts,Bart,2000,“Buoyancyandstrength”, JournalofSemantics,17(4):315–333. Geurts,Bart,2007,“Existentialimport”,in Comorovski,I.andvonHeusinger,K.(eds.),Existence:syntaxand semantics,Dordrecht:Springer,253–271. Geurts,BartandvanderSandt,Rob,1999,“Domain restriction”,inBosch,P.andvanderSandt,R.(eds.), Focus:Linguistic,Cognitive,andComputationalPerspectives, Cambridge:CambridgeUniversityPress,268–292. Geurts,BartandvanderSandt,Rob,2004,“Interpreting focus”,TheoreticalLinguistics,30(1): 1–44. Grice,H.Paul,1989,StudiesintheWayofWords, Cambridge,Massachusetts:HarvardUniversityPress. Groenendijk,JeroenandStokhof,Martin,1991,“Dynamic PredicateLogic”,LinguisticsandPhilosophy,14: 39–100. Gunji,Takao,1981,TowardsaComputationalTheoryof Pragmatics—Discourse,PresuppositionandImplicature,Ph.D. thesis,OhioStateUniversity. Hamblin,CharlesL.,1970,Fallacies,London:Methuen &Co. Hawkins,J.A.,1978,DefinitenessandIndefiniteness:AStudy inReferenceandGrammaticalityPrediction,CroomHelm. Heim,Irene,1982,OntheSemanticsofDefiniteandIndefinite NounPhrases,Ph.D.thesis,UniversityofMassachusettsat Amherst. Heim,Irene,1983,“Ontheprojectionproblemfor presuppositions”,inBarlow,M.andFlickinger,D.andWestcoat, M.(eds.),SecondAnnualWestCoastConferenceonFormal Linguistics,StanfordUniversity,114–126. Heim,Irene,1992,“PresuppositionProjectionandthe SemanticsofAttitudeVerbs”,JournalofSemantics,9: 183–221. Heinämäki,Orvokki,1974,SemanticsofEnglish TemporalConnectives,Ph.D.thesis,UniversityofIndianaat Bloomington. Horn,Laurence,1969,“Apresuppositionalanalysisof onlyandeven”,inPapersfromtheFifth RegionalMeetingoftheChicagoLinguisticsSociety,University ofChicago,98–107. Horn,Laurence,1972,Onthesemanticpropertiesoflogical operatorsinEnglish,Ph.D.thesis,UniversityofCalifornia,Los Angeles. Horn,Laurence,1985,“MetalinguisticNegationandPragmatic Ambiguity”,Language,61:121–174. Horn,Laurence,1989,ANaturalHistoryofNegation, UniversityofChicagoPress.Reissuedwithnewintroduction(2001), DavidHumeSeries,CSLIPublications,Stanford. Horton,Diane,1987,“Incorporatingagents’beliefsina modelofpresupposition”,Tech.Rep.CSRI-201,ComputerSystems ResearchInstitute,UniversityofToronto. Horton,DianeandHirst,Graeme,1988,“Presuppositionsas beliefs”,inProceedingsoftheInternationalConferenceon ComputationalLinguistics,COLING,255–260. Jackendoff,Ray,1972,SemanticInterpretationinGenerative Grammar,MITPress. Jayez,Jacques,Mongelli,Valeria,Reboul,AnneandVanDerHenst, Jean-Bapteiste,2014"Weakandstrongtriggers",inExperimental perspectivesonpresuppositions,Cham:Springer,173–193. Karttunen,Lauri,1971a,“Implicativeverbs”, Language,47(2):340–358. Karttunen,Lauri,1971b,“SomeObservationson Factivity”,PapersinLinguistics,5:55–69. Karttunen,Lauri,1973,“PresuppositionsofCompound Sentences”,LinguisticInquiry,4:167–193. Karttunen,Lauri,1974,“PresuppositionsandLinguistic Context”,TheoreticalLinguistics,1: 181–194. Karttunen,LauriandPeters,Stanley,1977,“Requiemfor presupposition”,inBLS3,ProceedingsoftheThirdAnnual MeetingoftheBerkeleyLinguisticSociety,Berkeley:California: 266–278. Karttunen,LauriandPeters,Stanley,1979,“Conventional ImplicaturesinMontagueGrammar”,inOh,C.-K.andDineen,D. (eds.),SyntaxandSemantics11:Presupposition,NewYork: AcademicPress,1–56. Karttunen,Lauri,2016,"Presupposition:Whatwentwrong?",in SemanticsandLinguisticTheory,26:705–731. Kay,Paul,1992,“Theinheritanceofpresuppositions”, LinguisticsandPhilosophy,15:333–381. Kempson,Ruth,1975,PresuppositionandtheDelimitationof Semantics,Cambridge:CambridgeUniversityPress. Kim,Christina,2008,"Processingpresupposition:verifying sentenceswith‘only’",UniversityofPennsylvania WorkingPapersinLinguistics,14(1):17. Kiparsky,PaulandKiparsky,Carol,1970,“Fact”,in Bierwisch,M.andHeidolph,K.(eds.),Progressin Linguistics,TheHague:Mouton,143–173. Kleene,Stephen,1952,IntroductiontoMetamathematics, Amsterdam:North-Holland. Krahmer,Emiel,1994,“Partialityanddynamics”,in Dekker,P.andStokhof,M.(eds.),ProceedingsoftheNinth AmsterdamColloquium:Volume2,UniversityofAmsterdam: ILLC,391–410. Krahmer,Emiel,1998,PresuppositionandAnaphora, Stanford:CSLIPublications. Kripke,Saul,2009,“Presuppositionandanaphora:Remarkson theformulationoftheprojectionproblem”,Linguistic Inquiry,40:367–386. Langendoen,D.TerenceandSavin,Harris,1971,“The projectionproblemforpresuppositions”,inFillmore,C.and Langendoen,D.T.(eds.),StudiesinLinguisticSemantics, NewYork:Holt,ReinhardtandWinston,373–388. Lappin,ShalomandReinhart,Tanya,1988,“Presuppositions ofquantifiers:aprocessingaccount”,Linguistics,26: 1021–1037. Lewis,David,1979,“Scorekeepinginalanguagegame”, JournalofPhilosophicalLogic,8:339–359.Alsoin Bäuerle,R.andEgli,U.andvonStechow,A.(eds.),1979, SemanticsfromDifferentPointsofView,Springer, Berlin. Lorenz,Sven,1992,“Ontheroleofreasoningaboutchange intheprojectionofpresuppositions”,Report234,IWBS,IBM Germany. Mandelkern,Matthew,2016,"Anoteonthearchitectureof presupposition",SemanticsandPragmatics,9(13):1–24. Marcu,Daniel,1994,AFormalisationandanAlgorithmfor ComputingPragmaticInferencesandDetectingInfelicities,Ph.D. thesis,UniversityofToronto,ComputerSystemsResearchInstitute. AvailableasTechnicalReportCSRI-309,ComputerResearchInstitute, UniversityofToronto. McCawley,James,1981,EverythingThatLinguistsHaveAlways WantedtoKnowaboutLogic(butWereAshamedtoAsk),University ofChicagoPress. Mercer,Robert,1987,ADefaultLogicApproachtothe DerivationofNaturalLanguagePresuppositions,Ph.D.thesis, UniversityofBritishColumbia. Mercer,Robert,1992,“DefaultLogic:Towardsacommon logicalsemanticsforpresuppositionandentailment”, JournalofSemantics,9:223–250. Morgan,Jerry,1969,“OntheTreatmentofPresuppositionin TransformationalGrammar”,inPapersfromtheFifthRegional Meeting,ChicagoLinguisticSociety,UniversityofChicago, 167–77. Nerlich,G.,1967,“PresuppositionandClassicalLogical Relations”,Analysis,27(3):104–106. Pearson,Hazel,2013,"AJudge-FreeSemanticsforPredicatesof PersonalTaste",JournalofSemantics,30:103–154. Peters,Stanley,1979,“Atruth-conditionalformulationof Karttunen’saccountofpresupposition”,Synthese, 40(2):301–316. Percus,Orin,2006,"Antipresuppositions",inA.Ueyama(ed.) Theoreticalandempiricalstudiesofreferenceandanaphora, JapanSocietyforthePromotionofScience:ReportoftheGrant-in-Aid forScientificResearchB. Potts,Christopher,2005,Thelogicofconventional implicatures,NewYork:OxfordUniversityPress. Prince,Ellen,1986,“Onthesyntacticmarkingof presupposedopenpropositions”,inFarley,A.M.andFarley,P. T.andMcCullough,K.-E.(eds.),PapersfromtheParasessionon PragmaticsandGrammaticalTheory,22ndRegionalMeetingofthe ChicagoLinguisticSociety,UniversityofChicago,208–222. Reinhart,Tanya,1982,“PragmaticsandLinguistics:An AnalysisofSentenceTopics”,Philosophica,27: 53–94. Roberts,Craige,1995,“DomainSelectioninDynamic Semantics”,inBach,E.Jelinek,E.andKratzer,A.andPartee, B.(eds.),QuantificationinNaturalLanguages,Dordrecht: Kluwer,661–700. Roberts,Craige,1998,“Focus,theflowofinformation,and universalgrammar”,inCulicover,P.andMcNally,L.(eds.), TheLimitsofSyntax,NewYork:AcademicPress,109–160. Roberts,Craige.,2012,"Informationstructure:Towardsan integratedformaltheoryofpragmatics",SemanticsandPragmatics,5: 1–69. Romoli,Jacopo,2011,"Thepresuppositionsofsofttriggersaren’t presuppositions",SemanticsandLinguisticTheory,21: 236–256. Romoli,Jacopo,2014,"Thepresuppositionsofsofttriggersare obligatoryscalarimplicatures",JournalofSemantics,32(2): 173–219. Romoli,Jacopo,Sudo,YasutadaandSnedeker,Jesse,2011,"An experimentalinvestigationofpresuppositionprojectioninconditional sentences",SemanticsandLinguisticTheory,21: 592–608. Romoli,Jacopo&Schwarz,Florian(2014),"Anexperimental comparisonbetweenpresuppositionsandindirectscalarimplicatures", inF.Schwarz(ed.),ExperimentalPerspectiveson Presuppositions,Cham:Springer,215–240. vanRooij,Robert,2007,“Strengtheningconditional presuppositions”,JournalofSemantics,24: 289–304. Rothschild,Daniel,2008,“Presuppositionprojectionand logicalequivalence”,PhilosophicalPerspectives, 22(1):473–497. Russell,Bertrand,1905,“OnDenoting”,Mind, 14:479–493. Sæbø,KjellJohan,1996,“Anaphoric presuppositionsandzeroanaphora”,Linguisticsand Philosophy,19:187–209. vanderSandt,Rob,1982,KontekstenPresuppositie:Een StudievanhetProjektieprobleemendePresuppositioneleEigenschappen vandeLogischeKonnektieven,Ph.D.thesis,NijmegenInstituteof Semantics. vanderSandt,Rob,1988,ContextandPresupposition, London:CroomHelm. vanderSandt,Rob,1989,“Presuppositionanddiscourse structure”,inBartsch,R.vanBenthem,J.andvanEmdeBoas,P. (eds.),SemanticsandContextualExpression,Dordrecht: Foris,287–94. vanderSandt,Rob,1992,“Presuppositionprojectionas anaphoraresolution”,JournalofSemantics,9: 333–377. Schlenker,Philippe,2007,“Anti-dynamics:Presupposition projectionwithoutdynamicsemantics”,JournalofLogic, LanguageandInformation,16(3):325–256. Schlenker,Philippe,2008,“Bearticulate:Apragmatic theoryofpresupposition”,TheoreticalLinguistics,34: 157–212. Schlenker,Philippe,2009,“Localcontexts”, SemanticsandPragmatics,2(3):1–78. Schlenker,Philippe,2011,"Theprovisoproblem:anote", NaturalLanguageSemantics,19(4):395–422. Schlenker,Philippe,2015,"Gesturalpresuppositions", Snippets,30:9–11. Schlenker,Philippe,2018,"Gestureprojectionand cosuppositions",LinguisticsandPhilosophy,41: 295–365. Schwarz,Florian(ed.),2014,Experimentalperspectiveson presuppositions,Cham:Springer. Sellars,W.,1954,“Presupposing”,Philosophical Review,63:197–215. Simons,Mandy,2001,“Ontheconversationalbasisofsome presuppositions”,inHastings,R.andJackson,B.andZvolensky, Z.(eds.),ProceedingsofSemanticsandLinguisticsTheory 11,Ithaca,NY:CLCPublications,431–448. Simons,Mandy,2003,“Presuppositionandaccommodation: UnderstandingtheStalnakerianpicture”,Philosophical Studies,112(3):251–278. Simons,Mandy,2004,“Presuppositionandrelevance”, inZoltánSzabó(ed.),Semanticsvs. Pragmatics,OxfordUniversityPress,329–355. Simons,Mandy,2006,“Foundationalissuesin presupposition”,PhilosophyCompass,1(4): 357–372. Simons,Mandy,2007,“Observationsonembeddingverbs, evidentiality,andpresupposition”,Lingua,117(6): 1034–1056. Simons,Mandy,Tonhauser,Judith,Beaver,DavidandRoberts, Craige,2010,"Whatprojectsandwhy",SemanticsandLinguistic Theory,20:309–327. Singh,Raj,2007a,“Formalalternativesasasolutiontothe provisoproblem”,inGibson,M.andFriedman,T.(eds.), Proceedingsofthe17thConferenceonSemanticsandLinguistic Theory,Ithaca,NY:CLCPublications,CornellUniversity, 264–228. Singh,Raj,2009,"Symmetricandinteractingalternativesfor implicatureandaccommodation",inN.KlinedinstandD.Rothschild (eds.),ProceedingsoftheESSLLI2009Workshop:Newdirectionsin thetheoryofpresupposition. Soames,Scott,1979,“Aprojectionproblemforspeaker presuppositions”,LinguisticInquiry,10: 623–666. Soames,Scott,1982,“Howpresuppositionsareinherited:a solutiontotheprojectionproblem”,Linguistic Inquiry,13:483–545. Soames,Scott,1989,“Presupposition”,inGabbay,D. andGuenther,F.(eds.),HandbookofPhilosophicalLogic, Dordrecht:Reidel,vol.IV,553–616. Stalnaker,Robert,1972,“Pragmatics”,inDavidson,D. andHarman,G.(eds.),SemanticsofNaturalLanguage,Reidel: Dordrecht,389–408. Stalnaker,Robert,1973,“Presuppositions”,The JournalofPhilosophicalLogic,2:447–457. Stalnaker,Robert,1974,“Pragmaticpresuppositions”, inMunitz,M.andUnger,P.(eds.),SemanticsandPhilosophy, NewYorkUniversityPress,197–214. Stalnaker,Robert,1998,“Ontherepresentationof context”,JournalofLogic,LanguageandInformation, 7:3–19. Strawson,PeterF.,1950,“Onreferring”, Mind,59:320–44. Strawson,PeterF.,1952,IntroductiontoLogicalTheory, London:Methuen. Strawson,PeterF.,1954,“AReplytoMr.Sellars”, PhilosophicalReview,63:216–231. Strawson,PeterF.,1964,“Identifyingreferenceand truth-values”,Theoria,30:96–118.Reprintedin: Logico-linguisticPapers.Methuen,London,1974, 75–95. Tiemann,Sonja,Schmid,Mareike,Bade,Nadine,Rolke,Bettina, Hertrich,Ingo,Ackermann,Hermann,Knapp,Julia,andBeck,Sigrid, 2011,"Psycholinguisticevidenceforpresuppositions:On-lineand off-linedata",inProceedingsofSinn&Bedeutung, 15(15):581–595. Tonhauser,Judith,2012,"Diagnosing(not-)at-issuecontent", ProceedingsofSemanticsofUnder-representedLanguagesofthe Americas(SULA),6:239–254 Tonhauser,Judith,Beaver,David,Roberts,Craig,andSimons, Mandy,2013,"Towardataxonomyofprojectivecontent", Language,89(1):66–109. Thomason,Richmond,1972,“Asemantictheoryofsortal incorrectness”,JournalofPhilosophicalLogic,1: 209–258. Veltman,Frank,1996,“Defaultsinupdatesemantics”, JournalofPhilosophicalLogic,25:221–261. Wilson,Deidre,1975,PresuppositionandNon-truth-conditional Semantics,London:AcademicPress. Yeom,J.I.,1998,APresuppositionalAnalysisofSpecific Indefinites:CommonGroundsasStructuredInformationStates, GarlandPublishing. Zeevat,H.,1999,“Explainingpresupposition triggers”,inProceedingsoftheTwelfthAmsterdam Colloquium,ILLC,Amsterdam,19–24. Zeevat,Henk,1992,“Presuppositionandaccommodationin updatesemantics”,JournalofSemantics,9: 379–412. AcademicTools Howtocitethisentry. PreviewthePDFversionofthisentryatthe FriendsoftheSEPSociety. Lookuptopicsandthinkersrelatedtothisentry attheInternetPhilosophyOntologyProject(InPhO). Enhancedbibliographyforthisentry atPhilPapers,withlinkstoitsdatabase. OtherInternetResources Workonpresuppositionatthe SemanticsArchive. Workonpresuppositionidentifiedby GoogleScholar. RelatedEntries assertion| discourserepresentationtheory| pragmatics| reference| semantics:dynamic| speechacts| Strawson,PeterFrederick Acknowledgments TheauthorswouldliketothankLeahVellemanforinitiallyconverting thearticletohtml,andtheeditorialteamatSEPforextensivehelp andadvicewithbothcontentandform.Beaver’sworkwassupportedby NSFaward0952862SemanticsandPragmaticsofProjectiveMeaning acrossLanguages,byNSFaward0904913ModelingDiscourseand SocialDynamicsinAuthoritarianRegimes,byBAADCHC-BAA-000101 TrackingSecretiveBehaviorthroughWordUse,andbytheNYCT TeXITproject. Copyright©2021by DavidI.Beaver BartGeurts KristieDenlinger OpenaccesstotheSEPismadepossiblebyaworld-widefundinginitiative. TheEncyclopediaNowNeedsYourSupport PleaseReadHowYouCanHelpKeeptheEncyclopediaFree Browse TableofContents What'sNew RandomEntry Chronological Archives About EditorialInformation AbouttheSEP EditorialBoard HowtoCitetheSEP SpecialCharacters AdvancedTools Contact SupportSEP SupporttheSEP PDFsforSEPFriends MakeaDonation SEPIAforLibraries MirrorSites Viewthissitefromanotherserver: USA(MainSite) Philosophy,StanfordUniversity Infoaboutmirrorsites TheStanfordEncyclopediaofPhilosophyiscopyright©2021byTheMetaphysicsResearchLab,DepartmentofPhilosophy,StanfordUniversity LibraryofCongressCatalogData:ISSN1095-5054