Metaphor in converstional context toward a connect

METAPHORANDSYMBOL,79(4),265-287Copyright©2004,LawrenceErlbaumAssociates,Inc.MetaphorsinConversationalContext:TowardaConnectivityTheoryofMetaphorInterpretationDavidRitchieDepartmentofCommunicationPortlandStateUniversityThemodelproposedinthisarticlemergesthecognitiveandsocialelementsofmeta-phoruseandinterpretationinthecognitiverepresentationofmutualcognitiveenvi-ronment(Sperber&Wilson,1986/1995)orcommonground(Clark,1996).Thepro-posedprincipleofmetaphorinterpretationisbasedontheinteractionofbothvehicleandtopicwiththecommonground.Becausecommongroundisinherentlyproblem-atic,itisrarelyaccuratetodiscussthemeaningofametaphor,asifmetaphorsmusthaveasinglewell-specifiedmeaning.Eachmetaphorisinterpretedintheparticularcommunicativecontextinwhichitisencountered,andindividualinterpretationswillnotnecessarilymatchunlesstheindividuals'cognitiverepresentationsofthecom-mongroundaresimilar.Theproposedmodelisconsistentwithseveralothertheoriesofmetaphorinterpretation,butitexplicitlyacknowledgestheneedforinterpretiveprocessestobecontextspecificandneurologicallyembodied,andavoidstheprob-lemofcircularity(Ritchie,2003b).Thestudyofmetaphorshasoftenproceededinanabstract,top-downway,basedonthelogicalorganizationoflanguageratherthanonthewaycognitiveprocessesareembodiedinthehumanbrainandnervoussystem,anddivorcedfromthesocialandcommunicativecontextsinwhichmetaphorsoccur.Althoughsuchtop-downtheoriesareoftenquiteelegantfromalogicalpointofview,theultimatetestmustalwaysbewhetherthestipulatedprocessesarecompatiblewiththewayhumansactuallythinkandcommunicate.Giventherecentadvancesinunderstandingtheneurologicalprocessesofperceptionandthought,itissomewhatsurprisingthatlittlesystematicattempthasbeenmadetodeveloptheoriesofmetaphorinterpreta-RequestsforreprintsshouldbesenttoDavidRitchie,DepartmentofCommunication,PortlandStatetJniversity,Portland,OR97214.E-mail:[e-mail]266RITCHIEtionconsistentwithwhatwealreadyknowabouttheneurologicalfunctioningofthebrain.TheoutlineofsuchamodelwasproposedbyRitchie(2004b),asanalternativetoFauconnierandTurner's(1998)metaphorsofconceptualspaceandconcep-tualblending.InthisarticleIdeveloptheseideasmoresystematically,begin-ningwithSperberandWilson's(1986/1995)theoryofrelevanceandClark's(1996)modelofconversation.IsuggestaneuralembodimentofSperberandWil-son'sconceptsofcognitiveenvironmentandcontextandClark'sconceptofcom-mongroundintermsofrepresentationinworkingmemory(Kintsch,1998).Ithenproposeamodelofmetaphorprocessingbasedonalterationofconnectionslink-ingotherelementspresentinworkingmemory,includingelementsactivatedbythemetaphortopic,tofeaturesofthemetaphorvehiclethatarerelevanttothesepreviouslyactivatedelements.Theresultingmodelisconsistentwithmanyofthefeaturesofextanttheoriessuchasconceptualmetaphor(Lakoff&Johnson,1980),aswellaswithwhatweknowabouttheneurologicalprocessesinvolvedinlan-guagecomprehension.Italsoprovidesafirmerfoundationforaneurologicallybasedcognitivetheoryofmetaphor.CONTEXT,RELEVANCE,ANDCOMMONGROUNDSperberandWilson(1986/1995)distinguishedbetweencodemodelsofcommuni-cation,whichassumethatutterancesandgesturesconveyparticularandreadilyidentifiable,decodablemeanings(cf.Reddy,1993)andinferencemodels,whichassumethatutterancesandgesturesareoftenambiguousandrequireinference,basedonthecontext.Muchoftheirdiscussionfocusesonostensivecommunica-tiveacts(indirectspeech,gestures,etc.),buttheirdiscussionappliestomostin-stancesofeverydaylanguageuse,withtheprimaryexceptionofinstancesinwhichtermshavebeengivenpreciselyprescribedtechnicalmeanings,asinmilitary,sci-entific,andlegallanguage.Intheirmodel,inferenceisbasedonasearchforacon-textinwhichthecommunicativeactwillberelevant,thatistosay,willhavemaxi-mumeffectwithminimumcognitiveeffort.CognitiveEnvironmentSperberandWilson(1986/1995)arguedthatcommunicationisachievedbychangingthemutualcognitiveenvironment.Theydefinedcognitiveenvironmentasthesetofallfactsthataremanifest(pp.39,151)toanindividual,thatistosay,allfactstheindividualcanatthetimerepresentmentally.Acognitiveenvironmentismutualtotheextentthateachpersonassumesthatitismanifesttoallpartiestothecommunicativeevent.Bothpartsofthisdefinitionposepotentialproblems,whichSperberandWilsonwenttoconsiderablelengthstoresolve.IarguethatCONNECTIVITYANDMETAPHOR267mostoftheseproblemsarisefrombeginningwithanapriori,top-downmodelofcommunication,ratherthanwithaneurallyembodied,bottom-upmodel.Throughoutmostoftheirdiscussionofinterpretation,SperberandWilson(1986/1995)reliedontheconceptofcontext,firstdefinedonasThesetofpre-misesusedininterpretinganutterance,...asubsetofthehearer'sassumptionsabouttheworld(p.15).Theypointedout:Acontextinthissenseisnotlimitedtoinformationabouttheimmediatephysicalen-vironmentortheimmediatelyprecedingutterances:expectationsaboutthefuture,scientifichypothesesorreligiousbeliefs,anecdotalmemories,generalculturalas-sumptions,beliefsaboutthementalstateofthespeaker,mayallplayaroleininter-pretation,(p.15)Acontext,thus,wouldseemtoconsistofsomesubsetofthethoughtsaccessibletoahearer(oraspeaker),andthecognitiveenvironmentofape