Lexical Semantics and Pragmatics

LexicalSemanticsandPragmaticsReinhardBlutner,BerlinAbstractInthispaperIdiscusssomegeneralproblemsoneisconfrontedwithwhentryingtoanalyzetheutteranceofwordswithinconcreteconceptualandcontextualsettingsandtogobeyondtheaspectsofmeaningtypicallyinvestigatedbyacontrastiveanalysisoflexemeswithintheKatz-Fodortraditionofsemantics.Afteremphasizingsomeimportantconsequencesofthetraditionalview,severalphenomenaarecollectedthatseemtoconflictwiththetheoreticalsettingsmadebyit.Someextensionsofthestandardtheoryareoutlinedthattakeabroaderviewoflanguageinterpretationandclaimtoincludepragmaticaspectsof(utterance)meaning.ThemodelscriticallyconsideredincludeBartsch'sindexicaltheoryofpolysemy,Bierwisch'stwo-levelsemanticsandPustejovsky'sgenerativelexicon.Finally,Iargueinfavorofaparticularaccountofthedivisionoflaborbetweenlexicalsemanticsandpragmatics.Thisaccountcombinestheideaof(radical)semanticunderspecificationinthelexiconwithatheoryofpragmaticstrengthening(basedonconversationalimplicatures).1IntroductionIntheviewofKatz&Fodor(1963)thescopeofalanguagedescriptioncoverstheknowledgeofafluentspeakeraboutthestructureofhislanguagethatenableshimtouseandunderstanditssentences.Thescopeofasemantictheoryisthenthepartofsuchadescriptionnotcoveredbyatheoryofsyntax.ThereisasecondaspectwhichKatzandFodormakeuseofinordertoboundthescopeofsemantics.Thisisthepragmaticaspectoflanguageanditexcludesfromthedescriptionanyabilitytouseandunderstandsentencesthatdependsonthesettingofthesentence.Setting,accordingtoKatz&Fodor(1963)canrefertopreviousdiscourse,socio-physicalfactorsandanyotheruseofnon-linguisticknowledge.Anicedemonstrationoftheessenceofnon-linguisticknowledgeintheunderstandingofsentenceshasbeenprovidedbypsychologistsinthe70's(e.g.Kintsch1974,Bransfordetal.1972).Let'sconsiderthefollowingutterance:(1)Thetonessoundedimpurebecausethehemwastorn.Iguesswedonotreallyunderstandwhatthissentencemeansuntilweknowthatthissentenceisaboutabagpipe.ItisevidentthatthisdifficultyisnotduetoourinsufficientknowledgeofEnglish.Thesyntaxinvolvedisquitesimpleandtherearenounknownwordsinthesentence.Instead,thedifficultyisrelatedtotroublesinaccessingtherelevantconceptualsetting.Theideaofbagpipingissimplytoounexpectedtobederivedinaquasi-neutralutterancecontext.Theexampledemonstratesthatwehavetodistinguishcarefullybetweenthelinguisticaspectsofrepresentingthe(formal)meaningofsentencesandthepragmaticaspectsofutteranceinterpretation(speaker'smeaning).InthispaperIrestrictmyselftothesemanticsoflexicalunitsandintendtoexplaintheinteractionoflexicalmeaningwithpragmatics.AlreadyKatz&Fodor(1963)havestressedthepointthatafullaccountoflexicalmeaninghastoincludemoreinformationthanthatwhichallowsonetodiscriminatethemeaningsofdifferentwords.Inoneoftheirexamplestheyarguethattakebackisusedinverydifferentwaysinthesentences(2a,b),althoughtherelevantlexicalentriesaresemanticallyunambiguous.(2)a.Shouldwetakethelionbacktothezoo?b.Shouldwetakethebusbacktothezoo?Anobviousdifferencebetweenthesesentencesisthatthelionistheobjecttakenbacktothe2zooin(2a),butthebusistheinstrumentthattakesusbacktothezooin(2b).Theproblemforthepragmaticcomponentofutteranceinterpretationistoexplainthedifferenceintermsofdifferentconceptualsettings(worldknowledge),startingfromalexiconthatdoesn'tdiscriminatethetwooccurrencesoftakebacksemanticallyandfromasyntaxthatiscompletelyparallelforthetwosentences.Asathirdintroductoryexamplelet'sconsidertheperceptionverbsinEnglish(cf.Sweetser1990).Thisexamplemaybehelpfulfordemonstratingawayofhowtodiscriminatebetweenthepurelylexicalsemanticcomponentoflanguageandthepragmaticcomponentinmorepracticalterms.IfSaussureisright,thereisanessentiallyarbitrarycomponentintheassociationofwordsormorphemeswithwhattheymean.Consequently,thefeatureofarbitrarinesscouldbetakenatleastasasufficientconditionforthepresenceofsemanticinformation.ItiscertainlyanarbitraryfactofEnglishthatsee(ratherthan,say,buyorsmell)referstovisualperceptionwhenitispartoftheutterance(3a).Giventhisarbitraryassociationbetweenaphonologicalwordanditsmeaning,however,itisbynomeansarbitrarythatseecanalsohaveanepistemicreadingasin(3b).(3)a.Iseethetree.b.Iseewhatyou'regettingat.Moreover,itisnotrandomthatothersensoryverbssuchassmellortastearenotusedtoexpressanepistemicreading.Sweetser(1990)triestosketchanexplanationforsuchfactsandinsiststhattheyhavetodowithconceptualorganization.Itisourknowledgeabouttheinnerworldthatimplicatesthatvisionandknowledgearehighlyrelated,incontrastto,saysmellandknowledgeortasteandknowledge,whichareonlyweaklyrelatedfornormalhumanbeings.Ifthisclaimiscorrect,thentheinformationthatseemayhaveanepistemicreadingbutsmellandtastedonotmustnolongerbestipulatedsemantically.Instead,thisinformationispragmaticinnature,havingtodowiththeutteranceofwordswithinaconceptualsetting,andcanbederivedbymeansofsomegeneralmechanismofconceptualinterpretation.Considerationsofthiskindraiseastandardpuzzleforlexicalsemanticswhenweaskhowtoseparatethe(mental)lexiconfromthe(mental)encyclopedia.Howshouldweseparateinformationaboutthemeaningofwordsfrominformati