Beruflich Dokumente
Kultur Dokumente
LostinConceptualSpace:
MetaphorsofConceptualIntegration
L.DavidRitchie
PortlandStateUniversity
Ritchie,D.(2004)Lostinspace:metaphorsinconceptualintegrationtheory.MetaphorandSymbol,19,31
50.
Abstract
ConceptualIntegrationTheory(Fauconnier&Turner,19982002)isreexaminedinthe
lightofrecentcriticisms(e.g.,Gibbs,20002001).Itisarguedthattheassumptionoffour
independentcognitivespacesentersthemodelprimarilyasanentailmentofthespace,
blending,andconstructionmetaphors,leadstounnecessaryambiguity,andworksagainst
statementofthetheoryinaformthatsupportsderivationoftestablehypotheses.Several
examplesfromFauconnierandTurner(2002)areanalyzedtoshowthattheycanbeinterpreted
moresimply,withoutneedoffourseparatespaces.Suggestionsaremadeforreformulating
ConceptualIntegrationTheorywithouttheproblematicmetaphors.
LostinConceptualSpace:
MetaphorsofConceptualIntegration
ConceptualIntegrationTheory(CIT),alsoknownasConceptualBlendingTheory
(Fauconnier&Turner,19982002)hasbeenrecognizedasapowerfulmodeloflanguage
processingandapotentialsolutiontoanumberofproblemsincognitivetheory(Grady,2000
Gibbs,2000Harder,2003).Itpromisesanintegratedmodelofcognitiveactivitythatwillcombine
explanationsoflinguisticcreativitywithexplanationofotherlanguagebehavior,aswellasof
variousproductsofhumanimagination(Gibbs,2000).YetConceptualBlendingTheoryhasalso
beencriticizedbecauseithasnotbeenmadeclearhowaspectscouldbeformulatedfortesting
http://web.pdx.edu/~cgrd/Metaphors%20of%20Conceptual%20Integration.html 1/24
20171125 MetaphorsofConceptualIntegration
andpotentialfalsification(Gibbs,20002001)andbecauseitseemstointroduceneedless
complexityintorelativelysimplelinguisticprocesses(Harder,2003).
InthisessayIexaminecrucialassumptionsofConceptualBlendingTheory,includingthe
assumptionthatintegrationalwaysinvolvesatleastfourdistinctconceptualspaces,includinga
genericspacecontainingelementscommontothetwoinputspaces.Iexaminetheentailments
ofthecentralmetaphors,mentalspaces,conceptualpackets,andconceptualblending,and
showthatthesemetaphorsobscureprocessesspecifiedbythemodel.IreviewFauconnierand
Turners(2002)responsestosomeofthecriticismsofConceptualBlendingTheoryandshowthat
theprincipalcriticismshavenotbeenadequatelyaddressed.Ishowhowtypicalexamplesusedto
explainandjustifyConceptualBlendingTheorycanbereadilyexplainedinlesscomplicatedand
morestraightforwardterms.Iconcludethatthemetaphorsusedtoexplainconceptualintegration
requireclosescrutinytoseparatemetaphoricalentailmentsfromtheactualrequirementsofthe
theory.
ConceptualBlending:TheModel
FauconnierandTurner(2002)explainConceptualBlendingTheoryintermsofmental
spaces,smallconceptualpacketsconnectedtolongtermschematicknowledgecalled
frames,aswellastolongtermspecificknowledge(p.40).Thementalspacesareillustratedby
circles,withrelevantcontentseitherdisplayediconicallyorlistedinabbreviatedform(seeFigure
1).Themodelpositsaminimumoffourmentalspaces:Twoinputspaces,agenericspacethat
containswhatthetwoinputshaveincommon,andablendedspacethatcontainssomeelements
fromeachinputspace.Theblendedspacemayalsocontainadditionalelements(emergent
structure)thatcanincludenewelementsretrievedfromlongtermmemoryorresultingfrom
comparisonofelementsdrawnfromtheseparateinputs,orfromelaborationontheelementsinthe
blendedspace(runningtheblend).
http://web.pdx.edu/~cgrd/Metaphors%20of%20Conceptual%20Integration.html 2/24
20171125 MetaphorsofConceptualIntegration
Figure1
ThisdiagramcreatedusingInspiration7.5byInspirationSoftware,Inc.
Example:Amonkclimbingamountain.FauconnierandTurner(2002)frequentlyreturn
toalogicalpuzzledrawnfromKoestler,inwhichamonkclimbsamountainononeday,beginning
atdawnandarrivingatsunset,thenreturnsdownthemountainonasubsequentday,again
beginningatdawnandarrivingatsunset.Thetaskistoshowthatthereissomeplacealongthe
paththatthemonkoccupiesatthesamehourofthedayonthedifferentjourneys.Thesolution
preferredbyFauconnierandTurneristoimaginethemonkwalkingbothupanddownonthesame
day,ascenarioconstructedbytakingelementsoftwoseparateinputspaces,oneforthemonk
walkingupthemountainononedayandoneforthemonkwalkingdownthemountainona
differentday,andblendingthemintoasingleimage(blendedspace).Agenericspace,
containingeverythingthetwoinputspaceshaveincommon(themonk,themountainpath,anda
daybeginningatdawnandendingatsunset),isrequiredtosupporttheblend(Figure1).Inthe
blend,themountainslopeandthetwoseparatedaysarefusedintoasinglemountainslopeanda
singleday,butthetwomonkimagescannotbefused,becausetheymoveinoppositedirections,
http://web.pdx.edu/~cgrd/Metaphors%20of%20Conceptual%20Integration.html 3/24
20171125 MetaphorsofConceptualIntegration
sotheymapintothefourth,blendedspace,astwoseparateindividuals.Whenwerunthe
blend,byimaginingthetwoindividualmonksprogressingalongthepath,weseethattheymust
inevitablymeet.
Example:MargaretThatcherforPresident.FauconnierandTurner(2002)citea
commonplaceobservationamongconservativecommentatorsintheearly1990s,thattheUnited
StatesneedsaMargaretThatcherbutthatshewouldnevergetelectedherebecauseof
oppositionfromlaborunions.FauconnierandTurnerclaimthat,tocomprehendthisargument,we
mustimagineThatcherrunningforU.S.presidentanddevelopthisscenarioinsufficientdetailto
perceivetheobstaclestoherelection.Thepointismadebysettingupasituation(theblend)
thathassomecharacteristicsofGreatBritain,somecharacteristicsoftheUnitedStates,andsome
propertiesofitsown(p.19).Inthisexample,thegenericspacecontainselementscommonto
MargaretThatcherandtheU.S.presidencysuchasworldleader.Theinputspace
correspondingtoMargaretThatchercontainsinformationaboutThatcher,includingthequalities
thatendearhertoconservativecommentators,alongwithherfamousstruggleswithBritishlabor
unions.TheinputspacecorrespondingtotheU.S.presidencycontainsinformationaboutthat
office,includingtheinfluenceoflaborunionsinelectoralpolitics.Therelevantinformationis
mappedfromthetwoinputspacesintoafourth,andentirelynew,blendedspace.Whenwerun
theblendwefirstperceivethatMargaretThatcherwouldmakeagreatpresidentbecauseofher
fabledqualitiesoftoughnessetc.thenweperceivethatshewouldfailasapresidentialcandidate
becausetheunionswouldblockhercandidacy.
ApproachestoConceptualIntegration.Itseemsobviousthatexamplesofreasoning
suchasthemonkpuzzleandtheMargaretThatcherforPresidentquiprequirethatdisparate
concepts(includingimages,schemata,andnarratives)becombinedinsomeformofconceptual
integration.However,conceptualintegrationcouldbeaccomplishedinmanyways.Considerfirst
thedegreeofconnectedness:Integrationmightmeanthatseparateconceptsbecomeconnected
orlinkedinmemory,justasonemightconnectastereosettoacomputer.Alternatively,features
orelementsoriginallyassociatedwithoneconceptmightbeaddedinmemorytoadifferent
concept,justasstereocomponentssuchastheFMreceiverandcdplayermightbewiredtoa
computer.Finally,anentirelynewconceptmightbeconstructedinmemory,combiningelements
similartosomeelementsofeachinputconcept,justasonemightpurchaseseveralentirelynew
stereoandcomputercomponentsandconnectthemtogetherinaseparateroom,apartfromthe
http://web.pdx.edu/~cgrd/Metaphors%20of%20Conceptual%20Integration.html 4/24
20171125 MetaphorsofConceptualIntegration
roomscontainingtheoriginalstereoandcomputer.Neurologically,theseconnectionsmightbe
accomplishedinvariousways,suchasalteringthepatternofsynapticconnectionsbetween
neurongroups(Calvin,1996Deacon,1997Phillips,1997),changingthelevelofactivationin
connectedneurongroups,orsynchronizingthespiketrainsofsomeneurongroupswiththoseof
others(Fotheringhame&Young,1997Phillips,1997).Constructionofanoutputspace,distinct
fromtheinputspaces,wouldinvolveconstructingparallelneuralcircuitsinseveralpartsofthe
brain,sincelanguageprocessingisusuallydistributedacrossareasassociatedwithmotoractivity
andvarioussensorymodalitiesaswellaswithsemanticmemory(Deacon,1997).
Conceptualintegrationmighttaketheformoflinksamongdiscreteconceptualelements,
justasoneuseswiresorcablestoconnectcomponentsofastereoset,or,astheblending
metaphorsuggests,conceptualelementsdrawnfromeachmightbemerged,inthefashionofthe
compositedrawingsofasuspectcreatedbypoliceartists,orinthefashionofcertainproductsin
whichcircuitsfromaGPS(GlobalPositioningSystem)receiverordigitalcameraandcircuitsfrom
atwowayradioorcellulartelephoneareimprintedonasinglechip,actuallyfusedintoasingle
newproductthatduplicatessomebutnotallfeaturesoftheoriginalproducts.Althoughtheactual
mechanismsareunclear,thistypeofblendingisconsistentwiththeobservationthatoneneuron
mightbeinvolvedinrepresentingseveralfeatures(Fotheringhame&Young,1997)andwith
Calvins(1996)speculationsaboutDarwinianprocessesincognition.
Calvin(1996)pointsoutthatneuronsarebundledinminicolumnsofabout100,and
suggeststhatthespatiotemporalfiringpatternsofagroupofneuronsmightrepresentaconcept,
word,ormetaphor.Basedonamixtureofresearch,mostlyinthevisualcortex,Calvinarguesthat
lateralinhibitioncreatesaringofdepressedactivityaroundanactivatedneuron,andthatoncetwo
cellsorsetsofcellsarefiringrepeatedly,inaparticularfiringpattern,thereshouldbeatendency
torecruitanothercellabout0.5mmaway,leadingtoaroughlyhexagonalpatternofspreading
activation.Thesepatternsmaydieawayasexcitationweakens,maybesuppressedbyinhibiting
signals,ormaybereinforcedbyconsonantsignals,accordingtoDarwinianprinciplesofselection.
IfindeedtheseDarwinianprinciplesare,asCalvinbelieves,responsiblefortherefinementand
clarificationofperceptionsandconcepts,then,itisalsopossiblethattwopatternsoffiring,
representingdistinctconcepts,couldformanoverlayorcomposite(Calvin,1996,pp.115120).
ThisprocesscouldprovideaneuralmechanismforFauconnierandTurners(2002)modelof
conceptualblending.
http://web.pdx.edu/~cgrd/Metaphors%20of%20Conceptual%20Integration.html 5/24
20171125 MetaphorsofConceptualIntegration
CriticismsofConceptualBlendingTheoryandResponsestoCriticisms
InresponsetotheobjectionthatConceptualBlendingTheoryhasyettobespecifiedwith
sufficientprecisionthatitcanbeempiricallytested,FauconnierandTurner(19982002)andtheir
variouscolleagues(seeforexampleCoulson&Matlock,2001)haveproducedempiricalevidence
thatisconsistentwithConceptualBlendingTheory.However,theevidencetodatestillfailseither
tojustifythecomplexitiesofConceptualBlendingTheoryortodifferentiatebetweenConceptual
Blendingandothertheorieswithwhichitisalsoconsistent.Themoredetailedexpositionof
ConceptualBlendingTheoryinFauconnierandTurner(2002)goesalongwaytowardspecifyinga
unifiedtheory,butitdoesnotsatisfactorilyaddresstheissuesoffalsificationandofeliminatingrival
hypotheses.Tothecontrary,FauconnierandTurner(2002)dismissthecallforfalsifiabilityby
claimingacomparisontoscienceslikeevolutionarybiologythatarenotaboutmakingfalsifiable
predictionsregardingfutureevents.Theythenshiftthegroundsoftheargument,claimingthat
wehavealreadyfalsifiedexistingaccountsofcounterfactualsbyshowingthecentralityof
counterfactualsliketheIronLady,whichsuchtheoriesareonprincipleunabletohandle(pp.54
56).
Whileitistruethatevolutionarybiologyandcosmologydonotmakepredictionsthatare
directlyfalsifiable,bothdependontheoriesthatdomakefalsifiablepredictions.Moreover,
FauconnierandTurner(2002)failtoconsiderotherpossibleexplanationsfortheIronLady
(MargaretThatcherforPresident)exampleinanydetailand,asIwillshow,theIronLadyexample
isreadilyexplainedbyasimplerandmorestraightforwardintegrationprocess,withnoneofthe
metaphoricalbaggageassociatedwithmentalspaces.AtestofwhetherFauconnierand
Turnersaccountmakespredictionsaboutactualhumanbehaviorthatareuniquelysupportedby
empiricalevidenceisstilllacking,andrelianceonthespaceandblendingmetaphorsseemsto
workagainstthekindofprecisespecificationthatwillsupportmeaningfulempiricaltests.
Computability.Apotentialcriticismofanymodeloflanguageisbasedonwhetherit
specifiesoperationsthatcanbecarriedoutinsuchawaythattheywillactuallyproduceresults
consistentwiththemodel(computability).VealeandODonoghue(2000)demonstratethat
FauconnierandTurners(1998)modeliscomputable,inthesensethatacomputerprogram
basedonitproducesresultsconsistentwiththemodel.However,theprimarysenseinwhicha
modelmustbecomputationalisthatanymodelofhumanlanguageusemustbecomputablebya
humanbrain.AsVealeandODonaghuesdiscussionillustrates,computationalmodelsof
http://web.pdx.edu/~cgrd/Metaphors%20of%20Conceptual%20Integration.html 6/24
20171125 MetaphorsofConceptualIntegration
languagearethemselvesoftenmetaphorical(notmerelymetonymical),inasmuchastheydifferin
fundamentalwaysfromhowhumanbrainsprocessinformation(Winograd&Flores,1986).
Anexampleofhowfarafieldcomputermodelingcantakethetheoristappearsinthe
discussionofsemanticnetworksinVealeandODonoghue(2000).InFigure1(page257)they
diagramsemanticrelationsbetweenNetscapeandMicrosoftandinFigure2(page258)they
diagramsemanticrelationsbetweenCokeandPepsi.(CocaColacreatesCokeCanscreate
CokeMarketaffectsPepsiMarketaffectsPepsiCo.)Itisnotclearfromwhencetheauthors
obtainedthedatarepresentedinthesediagrams,butitseemsunlikelythateitherdiagram
representsthenetworkofassociationsbywhichtheseconceptsarelinkedinatypicalcustomers
mindor,forthatmatter,eveninamarketingmanagersmind.Ontheotherhand,amorerealistic
setofassociationswouldbeverydifficulttorepresentinadiagramthatwouldfitonasinglepage
ofanacademicjournal,ortoprogramintoacomputerdatabase.
AsVealeandODonoghue(2000)pointout,anycomputationalalgorithmthatistorunona
computerrequiresaclearandprecisestoppingrule(seealsoSperber&Wilson,1986),andthat
ispartofwhatCITprovides.However,itisnotevidentthatthehumanbrainworksquitethatway:
Itisdifficulttopredictfullytheextenttowhichanindividualwillelaborateonafigureofspeechor
onasituation,althoughfactorssuchashaste,involvement,andneedforcognitionallcontribute
(Petty&Cacioppo,1981).Oftenahearerwillnotevenbothertoteaseoutthemeaningofa
metaphor,especiallyifitissomewhatobscureatothertimes,thesamepersonmaymulloverthe
possibleentailmentsandimplicationsofaninterestingmetaphor,joke,orstoryforseveralhours(a
cognitivelinguistmaythinkaboutaturnofphraseformanyyears).Itisquitepossiblethatsome
studentsinthephilosophyseminarinFauconnierandTurners(1998)examplewillprocessthe
professorsdebatewithKantpurelyintermsoftheevidenceandcounterevidence,entirely
ignoringthesupposeddebate,whileotherswillbetakenwiththeamusingideaofImmanuelKant
sittingtherenexttoProfessorJones,andelaborateonthestoryuntiltheyquitelosethethreadof
thelecture.Thehumanbrainprobablyhasseveralwaystostopitselffromfallingintothesortof
endlessloopthatcomputerprogrammersworryabout,includingboredomanddistractionbya
moreinterestingidea.Noneofthisistodenytheusefulnessandinterestofeffortslikethose
reportedbyVealeandODonoghue(2000),butitdoesraisethequestion:Howmuchdoesthis
sortofmodelingcontributetoourunderstandingofhumanlanguageprocessing?
http://web.pdx.edu/~cgrd/Metaphors%20of%20Conceptual%20Integration.html 7/24
20171125 MetaphorsofConceptualIntegration
DoesConceptualBlendingapplytoSimpleCognitiveProcesses?Accordingto
Fauconnier(1994),thecognitivespacemodelwasoriginallydevelopedinreactiontoasetof
formalisttheoriesofmeaning,theoriesthatassumethatnaturallanguagesemanticscanbe
adequatelystudiedwiththetoolsofformallogic(Lakoff&Sweetser,1994,p.ix),andassuchit
hasprovidedausefulbridgetoamorecognitivemodeloflanguage.Harder(2003)applaudsthe
conceptualspaceapproachbecausepreviouslydisparatepropertiescanbebroughttocoexistin
thesamementalspacewithpropertiesthatwerefoundinneitheroftheoriginalspaces(p.91),
butobjectstoattemptstoextenditfromcomplextosimplecognitiveprocesses.
Harder(2003)givestheexampleoffakegun,claimingthat,unlessitisaguninonespace,
itcannotbecalledafakeinanother(p.91).Butthisseemsdisingenuousanextendedindex
fingerisnotaguninanyspace,butisfrequentlyusedasaniconicrepresentationofagun,and
occasionally,hiddeninacoatpocket,serves(atleastinthemovies)toconvinceavictimthatareal
gunispresent.Norisitapparentwhyfakewouldrequireatreatmentdifferentfromblack,
which,accordingtoHarder,doesnotrequireblendingfromseparateconceptualspaces.Bothare
qualifiers:Blackaddsinformationaboutthecolorofthemetal,fakeaddsinformationabout
constraintsontheobjectsaffordances.Inadditiontodescriptorsofcolor,shape,andmaterial
(blackgun,snubnosegun,pearlhandledgun,plasticgun),thereisacontinuumofdescriptors
thatspecifyoperationalcharacteristicsoftheobject(singleshotgun,dartgun,startersgun,flare
gun,capgun,watergun,toygun,andfakegun).Somedescriptors(handgun,shotgun,machine
gun)specifybothphysicalappearanceandoperationalcharacteristics.
Thecontinuumbetweenrealandfakeissimilartothecontinuumofmetaphoricity(Ritchie,
2003)andraisessimilarissues.Inthecaseofatoygun,theexpectationofbeingtreatedasifit
werearealguniscreatedbythecontainedreality(orlayer)ofplayinthecaseofafakegunused
inarobbery,itiscreatedbytheconstructedreality(orlayer)ofdeception(cf.Clark,1996).Abb
guncanbeusedtokillsmallanimals,anditcanoccasionallybedeadlytohumansisitarealgun
oratoygun?Ifblackgundoesnotrequireblendingfromtwoseparatespaces,thenitisdifficultto
seeexactlywherealongthiscontinuumblendingwouldbegintoberequired.Bycontrast,if
conceptualintegrationisaccomplishedthroughasimplerprocessofalteringthestrengthofneural
connections,allthatneedhappenasweprogressalongtherealfakeorliteralmetaphorical
continuumisthattheneuralconnectionsbecomedenser,moreextended,andmorecomplex.
http://web.pdx.edu/~cgrd/Metaphors%20of%20Conceptual%20Integration.html 8/24
20171125 MetaphorsofConceptualIntegration
Harders(2003)objectiontoconceptualblendingtheoryisprimarilyoneofscope:he
proteststhatblendingcannotbetheexplanationforsimplerprocessesofgrammar.Tosupporthis
objection,hecitesevidencefromexperimentsshowingthatchildrenbegintodistinguishbetween
theactualcolorofagreencatandtheapparentcolorofagreencat,coveredbyaredfiltersoasto
lookblack,onlyataboutagefourwellaftertheyhavelearnedbasicsyntax.Fromthis,heinfers
thatchildrenlearnsyntaxbeforetheylearntodistinguishtheseparateconceptualspacesoccupied
bytheactualgreenimageofthecatandtheperceptuallyalteredimageperceivedthroughthe
filter.Butveryyoungchildrenseemquitecapableofdistinguishingbetweenadoll(toybaby)anda
realbaby,astuffedanimal(toypuppy)andarealanimal,etc.,evenastheyconstructelaborate
playscenariosandcarryonapparentconversationswiththesetoys.Theprobleminthegreen
catexperimentsmaybewiththeunfamiliarityandartificialityoftheexperimentaltask.Thisisnot
todenythatcognitivetasksdifferinthecomplexityofconceptualintegrationtheyrequire:Tothe
contrary,Iwouldsuggestthattheyrangecontinuouslyfromconnectionssosimpleastoescape
notice(integrating,orbindingacolorsuchasblack,alongwithothersensoryperceptionssuchas
weightandhardness,withanobjectsuchasagun)throughtoconnectionssocomplexthatwe
marvelatthem(e.g.,themonkpuzzlecitedbyFauconnier&Turner,[2002],AliceinWonderland
[Carroll,1923],andimaginarynumbers).
Myownquestionshavetodowithwhetherthefullconceptualapparatus,includingatleast
fourdistinctspacesforeachinstanceofconceptualintegration,andmanymoreforcomplex
operationsthatrequirelongsequencesofintegrations,isnecessaryforconceptualintegration.I
willsuggestthatmanyoftheseelementsderive,notfromtherequirementsofneurologically
embodiedcognitiveprocesses,butfromtheentailmentsofthefocalmetaphors,spacesand
blending.
MetametaphorsofCognitiveProcesses
FauconnierandTurner(19982002)useavarietyofmetaphorstodescribeconceptual
integration,andsomeofthesehaveconflictingentailments.Mentalspacesmightbeinterpreted
asametaphoricalconnectiontoamodelsuchasmultidimensionalscaling,oravectormodelsuch
asthatofKintsch(1998Kintsch&Bowles,2002),inwhichconcepts,perceptions,motoractivities,
etc.areunderstoodasvectors,meaningsareunderstoodasdimensionsinsemanticspace,
andsimilarity,connectedness,andcoactivationbetweentwoconceptsareunderstoodintermsof
proximityincognitivespace.Kintsch(1998)linkshisvectormodeloflanguageprocessingtothe
http://web.pdx.edu/~cgrd/Metaphors%20of%20Conceptual%20Integration.html 9/24
20171125 MetaphorsofConceptualIntegration
strengthofconnectionsamongtheneuralrepresentationsofcognitiveelementsthedimensionsof
avectorassociatedwithaword,phrase,orconceptcorrespondinsomewaytothestrengthofthe
neuralconnectionstohundredsofotherwords,phrases,andconcepts,and,inprinciple,to
embodiedexperience(perception,motoraction,etc.).However,becauseKintschhasbeenunable
tooperationalizelinkagestononlanguageexperiences,Kintschsmodelremainsungrounded
(Gibbs,1999).
FauconnierandTurneralsolinktheirmodelexplicitlytoanetworkorconnectionistmodel:
Intheneuralinterpretationofthesecognitiveprocesses,mentalspacesaresetsofactivated
neuronalassemblies,andthelinesbetweenelementscorrespondtocoactivationbindingsofa
certainkind(2002,p.40).However,theydonotexplainhowtheelementsoftheirmodelmightbe
realizedinconnectionsamongneuronalassemblies,andmuchoftheirdiscussionisinconsistent
withanetworkmodel.
Inthesamesection,FauconnierandTurneralsolinktheideaofmentalspacestolong
termschematicknowledgecalledframes(2002,p.40)and,byimplication,tothekindof
organizednetworksofgeneralknowledgeabouttheworldoftenreferredtoasschemasand
scripts.However,conceptualpacketsandmentalassemblies,alongwithFauconnierand
Turnersuseofcirculardiagramstoillustratementalspacesandtheirdescriptionoftheprocess
throughwhichrelevantcontentsoftwoinputspacesarecopiedintoaseparateblendedmental
space,suggestsmoreofaconduitorcontainermetaphor(Reddy,1993),inwhichmeanings
areconceptualizedasobjectsthatcanbeputintowordsandphrasesandconveyedtoareader
orlistenerwhogetsthemeaningoutofthem.
AlthoughFauconnierandTurner(2002)statethatconceptualintegrationisrelatedtoa
networkmodeloflanguageprocessing,andattempttolinktheirmodeltoneuralprocesses,
metaphorssuchasspace,packets,andblendingworkagainstanetworkorconnectionist
understandingoflanguage.Thecreationofanentirelynewblendedspace,whichcontainsthe
relevantconceptsduplicatedfromtheinputspacesandmappedontotheblendedspace,isa
directentailmentofthespacemetaphor,asisthepositingofagenericspaceastheinput
connectingprinciple,ratherthanaprocessofselectionandsuppressioninworkingmemory
(Gernsbacheretal.,2001Kintsch,1998).Theuseofcirclesandboxesinillustrationsofthe
model(seeFigure1)reinforcestheideaofboundariesseparatingthevariousconceptual
http://web.pdx.edu/~cgrd/Metaphors%20of%20Conceptual%20Integration.html 10/24
20171125 MetaphorsofConceptualIntegration
elementsandtheneedforreplicationofelementswithinaseparatespaceratherthanconnection
ofexistingelementsinanewcompositepattern,aswouldbeentailedbyanetworkmetaphor.
Allofthisimpliesthat,whenconceptsareintegratedduringlanguagecomprehension,the
patternsassociatedwitheachconcept(howevertheyarerepresentedinthebrain)mustbe
duplicatedinanewpatternwithitsownindependentrepresentation.Butconsider(1):
(1)Heusedwhathethoughtwasafakegunintheholdupbutitturnedoutthatitreallywasagun
andtheclerkbehindthecounterwasanundercoverpoliceofficersohewaschargedwitharmed
robberyandassaultingapoliceofficer.
Interpretingacomplexnarrativesuchas(1)wouldrequirecreationofalongsequenceofentirely
separatespaces(representations),eachfullyduplicatingtherelevantfeaturesofthepreceding
spaces.Unlesstheseindependentmentalspacesaredissolvedasnewonesaregenerated,the
loadoncognitivecapacitymustexpandquiterapidly.
Bycontrast,aschemadrivenconnectionistmodelmightpositthataculturallylearned
robberyschema(embodiedasaparticularpatternofneuralconnections)isactivated,thenaltered
asthenarrativeprogressesbyaddingorchangingconnectionswithotherschemas(e.g.,forfake
gunandundercoverpoliceman)andfinallywithacriminaltrialschema.Thesechangesinthe
connectionsaretemporaryduringanyparticularnarrative,butcanbestrengthenedwithrepeated
useuntiltheybecomepartofthebasicschemaandare,ineffect,lexicalizedasundercover
policemanprobablyisformanycrimefictionbuffs.Theactivationanddeactivationofsynaptic
connectionsrequiredbyaconnectionistmodelseemsinprinciplesimplerthanthereallocationof
entiregroupsofneuronsentailedbythecognitivespacemetaphor.
Thespacemetaphorisparticularlytroublesomeinanexamplesuchasthemonkpuzzle,
becausethemonkinthestoryactuallymovesthroughasinglephysicalspace,butthenarrative
positstwodistinctlocationsinspacetime,thuscreatinganapparent,butunnecessary,paradox.
Theextensionofphysicalspace(themountainwithitspath)toconceptualgenericspace(the
elementscommontothetwohalvesofthestory)seemsnaturalandunproblematic.Oncethis
connectionismade,itseemstofollowthatthetwojourneysinphysicalspacecanbecompared
onlybymatchingthesharedelementsinthegenericspace(thecognitiverepresentationofthe
mountainandthepathleadingupthemountain),andmergingthetwoconceptualspaces,
includingrelevantelementsuniquetoeachoftheinputconceptualspaces(therepresentationof
themonktravelingupthephysicalspaceofthemountainandtherepresentationofthemonk
http://web.pdx.edu/~cgrd/Metaphors%20of%20Conceptual%20Integration.html 11/24
20171125 MetaphorsofConceptualIntegration
travelingdownthephysicalspaceofthemountain).Butiftheseparateconceptualspacesare
mergedintoanewspace,thepreviousspacesarelost,sothespacemetaphorrequiresthatan
entirelynewblendedspacebecreated,duplicatinginformationfromtheinputspaces.
Allofthisisaconsequenceofconflatingphysicalspace(themountain)withconceptual
space,anddoesnotnecessarilyhaveanythingtodowithactualneurologicalprocessesor
eventsinthepuzzlesolversbrain.Onlybystrippingthetheoreticalaccountofitsmetaphorical
languagecanweseethatprocessingthepuzzleneednotinvolveentirelyseparatesetsof
activatedneuronalassemblies(Fauconnier&Turner,2002,p.40)butmayrequirenomorethan
creatingnewlinkagesorchangingtheactivationlevelsofexistinglinkageswithinandbetweenthe
preexistingactivatedneuronalassembliesfortheinputconcepts.
Thespacemetaphoralsocontributestoconfusionaboutlevelsofanalysis.Fauconnier
andTurnerdefinementalspaceintermsofactivatedneuronassemblies.Buttheyalsodefineit
asasocialandculturalphenomenon,forexample,Inculturalpractices,theculturemayalready
haverunablendtoagreatlevelofspecificityforspecificinputs,sothattheentireintegration
networkisavailable,withallofitsprojectionsandelaborations(2002,p.72).Evenasidefromthe
questionofwhetherablendisacognitiveoraculturalphenomenon,itisnotatallclearwhatit
wouldmeanforaculturetorunablendortomakeavailableanentireintegrationnetwork.The
phrasesuggestsacomputermetaphor,e.g.,blendedspaceiscomputersoftware,butitisnot
clearhowtheblendedspace/softwarewouldactuallyberuninanindividualmind,muchlessin
aculture.
ItappearsherethatFauconnierandTurner(2002)areattemptingtoachievemuchthe
sameobjectiveasmemetheorists,namely,aunitaryaccountoftheculturaltransmissionofideas
(Dawkins,1993Blackmore,1999forcriticalappraisalsseeAunger,2000Kuper,2000Sperber,
2000).Intuitivelyweknowthatintricateconceptualcombinationsarelearnedfromothers:Difficult
asstudentsmayfindittolearncalculus,forexample,itwasmuchmoredifficulttoinventit.But
simplypositingaconceptsuchasablendedspacethatcrossestheindividual/sociallevelsof
analysisscarcelyprovidesanexplanationofculturaltransmission,anymorethanitdoestoposita
genelikeunitofmeaningandcallitameme.
AlternativeInterpretations.
http://web.pdx.edu/~cgrd/Metaphors%20of%20Conceptual%20Integration.html 12/24
20171125 MetaphorsofConceptualIntegration
ManyoftheexamplesprovidedbyFauconnierandTurner(2002)relyonaneedlessly
awkwardoridiosyncraticinterpretationoffigurativelanguage.Itisusefultoconsideralternative
viableinterpretationsoftheseexamples(Ritchie,2003),bothtoseewhethertheexplanatory
apparatusofconceptualblendingisstillrequiredandtounderstandhowtheunderlyinglogicofthe
modelworksinpractice.
TheMonkClimbingtheMountainRevisited.Forexample,amorestraightforward
solutiontotheriddleofthemonkistodrawagraphwiththemonkselevation(ordistancefrom
home)ononeaxisandthetimeofdayontheother,andplotthetwojourneysonthegraph.The
completedgraphwilllooklikearectangle,withthejourneysrepresentedbydiagonals(not
necessarilystraight).Nomatterhowthetwodiagonalsaredrawn,theymustintersect.Thereisno
needforafancifulnarrativeinwhichthemonkmeetshimselfmidway,andthetroublesome
ambiguityoftheconcept,space(physicalspace,conceptualspace,ormathematicalspace)
isavoidedalongwiththequestionofwhetherthemonkhastostepasidetolethimselfpass.
Thesameresultcanbeachievedwithinanarrativeframebyaddingaqualifyingmarkerto
partofthenarrative.Consider,
(2)WhenItouredthebattlefieldatGettysburgIwashauntedbyallthesoldierswhodied
there.
In(2)thehearereasilydistinguishesbetweenthespeakersvisittoGettysburg,representedas
real,andtheghostsofdeadsoldiers,representedasmetaphorical,withinasinglenarrative
thereisnoapparentreasonwhydistinguishingbetweenfactualandfancifulrequireseitherspeaker
orlistenertoconstructaseparatespace.Similarly,onthewaybackdownthemountain,wecan
readilyimaginethatatsomepointthemonkmightremarktohimself,
(3)Atthishourafewdaysago,Iwaspassingthisverysamerock.
In(3)bothspeakerandhearereasilydistinguishbetweenthemonkscurrentlocation,represented
aspresentandthemonkspriorlocation,representedasremembered.Allthatisneededisa
connectiontothemonksmemoryofthejourneyupthemountain,withamarkertodistinguish
presenteventfromrememberedevent.(Anactualexampleofsuchamarkedconnectionbetween
twojourneysisprovidedbyKrakauer[1998],whenheobserves,atonepointinhisflighttoNepal,
thattheairplaneisactuallycruisingatanaltitudebelowthatofthetopofMt.Everest.)Tobring
http://web.pdx.edu/~cgrd/Metaphors%20of%20Conceptual%20Integration.html 13/24
20171125 MetaphorsofConceptualIntegration
thesetwosolutionstogether,wecanalsoimaginethemonk,ashejourneysbacktowardhiscell,
recallingthetimethathepassedeachlandmarkalongthetrailonthewayupthemountain:Four
p.m.,threethirtyp.m.,threep.m.,andsoon,untilheuttersatimethatmatchesthereadoutofhis
digitalwatch.
Itisnotclearwhatismeantbyablendedspace,asdistinctfrommerelyaddingor
strengtheninglinksbetweentwoneuronalnetworks(orbetweentwopartsofthesamenetwork),or
whatconstructionofathird,independentnetworkwouldaddtothecomprehensionprocess.Noris
itclearwhatismeantbyagenericspace.Atfirstglance,itseemsclearenoughinthemonk
climbingamountainexample,butasIpointedoutinthediscussionofmetaphoricentailments,the
apparentobviousnessoftheneedforagenericspaceisatleastinpartaresultofconfounding
themetaphor,mentalspacewiththeliteralspaceoccupiedbythemonkonthemountain.
Comparethisactofconceptualintegrationwiththeactthatoccurswhenanobjectisdescribed:
Thegunisblackconnectstheperceptualconcept,black,withtheobjectconcept,gun.There
isnoreasontosupposethattheconcepts,blackandgun,needtohaveanyelementsin
commonasabasisforintegratingthem.Noristhereanyreasontothinkthatconnectingsuch
elementsleadstothecreationofanewrepresentationthatistotallyindependentofthediscrete
concepts(ablendedspace).Studiesofbrainactivationduringlanguageprocessingshowthat
processingapassageinwhichinformationfrommultiplesensorymodalitiesisintegratedentails
activationofseparateareasofthebrainassociatedwitheachmodality(Deacon,1997).Tobe
sure,itisasyetunknownhowtheinformationfromtheseseparateareasisboundtogether
(Phillips,1997),butthereisnoreasontothinkthatsuchbindingrequiresthereplicationof
conceptualstructurefromeachinanovelstructure.Giventheabilityofmosthumanstospinout
fancifulnarratives,andtoconstructcomplexlogicalarguments,thecreationofanentirelynew
blendedspaceforeachactofconceptualintegrationwouldmultiplyconceptualrepresentationsin
thebraintothepointthatmemorycapacitywouldquicklybeexhausted.
MargaretThatcherforPresidentRevisited.AccordingtoFauconnierandTurners
analysis,ablendedspaceisrequiredtoprocessthecounterfactualelementsofthescenario
impliedbythequipaboutMargaretThatcherThatcherisnotaU.S.citizen,hencecouldnot
legallybeelectedPresident.Butitisnotevidentthatthecounterfactualelementsneedenterinto
thecomprehensionprocessatall.Itissimplerandmorestraightforwardtoviewthisasacaseof
metonymy,inwhichMargaretThatcherstandsmetonymicallyforasetofqualitiesassociatedwith
http://web.pdx.edu/~cgrd/Metaphors%20of%20Conceptual%20Integration.html 14/24
20171125 MetaphorsofConceptualIntegration
thepersonofthatname(cf.Glucksberg&McGlone[1999],CambodiahasbecomeVietnams
Vietnam).
Considerhowtheconversationamongagroupofpoliticallyconservativeactivistsmightgo,
shouldoneofthemdrawattentiontothecounterfactualnatureofthequip:
(4)WhatthiscountryneedsisaMargaretThatcherbutshecouldneverbeelected
herebecausetheunionscantstandher.
(5)ButMargaretThatchercouldneverbeelectedintheUnitedStatesanyway,
becausesheisntaU.S.citizen.
Sentence(5)wouldberegardedaseitherdeliberatesabotageoftheconversationorevidenceof
politicalnaivete(Grice,1975Sperber&Wilson,1986).Similarly,in
(6)ImgoingtowriteinHomerSimpsonforpresident.
(7)Butacartooncharactercantbeelectedpresident!
Sentence(7)wouldberegardedasevidenceeitherofadesiretosabotagetheconversation,orof
atotallackofhumor.
Onthefaceofit,(4)simplypositsthatcertaincharacteristicsassociatedwithMargaret
ThatcherwouldfitwellwiththerolerequirementsofPresidentoftheUnitedStates.Likewise,(6)
invitesthehearertotakecertaincharacteristicsassociatedwiththe(fictional)personofHomer
SimpsonandassociatethemwiththeroleofPresidentoftheUnitedStates.ConsiderthePat
PaulsenforPresidentcampaignontheoldSmothersBrothersTVprogram,beginningin1968.A
nativebornU.S.Citizen,Paulsenmettheconstitutionalrequirementsfortheoffice,sotherewas
nothingcounterfactualaboutthescenario.YetPaulsensactualdeclaredcandidacywasnomore
seriousthantheimaginedcandidacyofMargaretThatcher.Neitherproposalwasintendedtobe
takenseriouslybothproposalswereintendedasacommentaryonthestateofU.S.politics.
Althoughmentionofafamiliarconcept(likeMargaretThatcherorPatPaulsen)willordinarilylead
toincreasedactivationofmultiplelinkstootherinformation(suchasThatchersnationalityandher
famoushandbag,orPaulsensdrollpoliticalcommentaries),allbuttherelevantlinkswillbe
suppressedbythecontextualinformationinshorttermmemorybeforetheyrisetoconsciousness
(Deacon,1997Gernsbacheretal.,2001Kintsch,1998).TheconstitutionalbarrierstoThatchers
http://web.pdx.edu/~cgrd/Metaphors%20of%20Conceptual%20Integration.html 15/24
20171125 MetaphorsofConceptualIntegration
electionandthepracticalbarrierstoPaulsenselectionareunlikelytobeprocessedunlessthey
arerelevanttotheconversation.Thereisnoneedtoposittheexistenceoffourdistinct
representationsormentalspaces.
Peoplearecapableofentertainingelaboratelycounterfactualscenarios,andsometimesthe
counterfactualnatureofthescenarioispartofthefunaswiththecandidacyofPatPaulsenor
HomerSimpson.Evenforwellelaboratedimaginativescenarios,however,allthatisneededisa
processinwhichrelevantknowledgeisactivatedinshorttermmemory,irrelevantconnections
suppressed,andnewconnectionsaremadeorstrengthenedamongrelevantaspectsof
simultaneouslyactivatedknowledgenetworks.Thesenewconnectionsthenremainactivein
workingmemoryforatime,andinfluencethewaysubsequentinformationisprocessed(Deacon,
1997Gernsbacheretal.,2001Kintsch,1998).
Example:DiggingyourownGrave.Anotherusefulexampleisthefamiliaridiom,You
arediggingyourowngrave.FauconnierandTurner(2002)claimthatthismetaphorical
expressionrequiresamorecomplexinterpretation(adoublescopenetwork),becausedigginga
gravedoesnotcauseonesdeath,andinanyeventonesgraveisordinarilydugbysomeone
else.Theyclaimthatthemetaphorcombinesthecausal,intentional,andinternaleventstructure
ofanunwittingfailurescenariowiththeconcretestructureofgraves,digging,andburial,froma
diggingthegravescenario(pp.132133).Buttheymaysimplyhavemisconstruedthemetaphor.
Diggingyourowngraveisrelatedtootherfamiliarexpressions,includingdiggingyourself
intoahole,diggingitdeeperanddeeper,gettinginoveryourhead,andgettingintoodeep.
Whileitistruethatpeoplerarelyknowinglydigtheirowngraves,withtherareoccasionof
prisonersinatotalitarianregime(Fauconnier&Turner,2002,p.132),andthatthemereactof
havingagravereadydoesnotcausedeath,itisalsotruethataholedugforanotherpurposecan
spectacularlyandunintentionallybecomethediggersgrave,asinthecaseofmineandtunnel
caveins.Sucheventshappenalltoofrequentlyand,whentheyhappen,theyremainsalientinthe
newsfordaysorevenweeks,andsometimesbecomememorializedinmusicorfolklore(Dickens,
1972Seeger,1958).Theexampleofanoffspringbeingwarnedaboutriskyfinancialventures
(Fauconnier&Turner,2002,p.131133)seemsdirectlyrelatedtothissortofdisaster.Throughhis
riskyinvestments,theoffspringismetaphoricallydiggingaholeforhimselfwhich,although
intendedtobecomeagoldminemayinsteadprovetobehisgrave.Closelyrelatedmetaphors
http://web.pdx.edu/~cgrd/Metaphors%20of%20Conceptual%20Integration.html 16/24
20171125 MetaphorsofConceptualIntegration
includehisplansfelldownaroundhim,hewasburiedunderamountainofdebt,andthe
poignanttitletoFriedmans(1992)biographyofJanisJoplin,BuriedAlive.
Digginginpursuitofwealthisacommonplacemetaphorinoursociety,anddeathornear
deathbycaveinsissufficientlyfrequentandsufficientlynotoriousastobereadilyaccessibleinthe
longtermmemoryofmostpeople.Somewhatlessspectacularly,itisrelativelyeasytodigahole
sufficientlydeepthatitisdifficulttoclimboutofwithoutexternalassistance.Comprehensionof
diggingyourselfintoaholeordiggingyourowngraveiseasilyexplainedintermsofcontextual
activationandsuppressionofcommonknowledgebyelementspresentinshorttermworking
memory.Muchofthiscommonknowledgehasbecomemoreorlesslexicalizedthroughrepeated
reference:Yourediggingyourselfintoaholemayormaynotevokeavisualimageofsomeone
standinginapitanddeepeningitbythrowingevenmoredirtoutofthebottom,andstuckinarut
mayormaynotevokeavisualimageofapersondrivingavehicledownamuddyanddeeply
ruttedroad.
ADifferentApproachtoConceptualIntegration.
Analysisofthediggingmetaphorssuggestsadifferentapproachtoconceptual
integration,atleastwithrespecttofigurativelanguage.Digginginthegroundinsearchofwealth
andtheassociatedrisksofcaveinsandotherdisasterhasbecomecloselyassociatedwiththe
immoderatepursuitofwealthaswellaswithselfgeneratedrisk.Moremundanely,peoplealso,at
onetime,routinelydugwells,basements,androotcellarsandsometimesexperienceddifficulty
gettingoutofaholetheyhaddug.Thus,diggingbecameassociatedinmemorywithabstract
conceptsincludingthesearchforwealth,dangerfromunexpectedevents,andtheironyof
committingactions(digging)thatledtolaterdifficulties(requiringhelptoclimboutofadeephole).
Thesescenariosweresufficientlycommon,eitherindirectexperienceorvicariouslythroughthe
mediaofnewspapers,entertainmentmedia,andsoforth,andtheyarousedsufficientlycommon
emotionsandexpectations,thattheyprovidedareadyexpressionformoreabstractideasand
experiencesthatarousesimilaremotionsandexpectations.
Thevividimagesofdigging(arootcellaroracoalmine)becomeconnectedwiththe
expectations,motoractions,perceptions,andemotionsassociatedwith,respectively,ironicself
entrapmentandsudden,catastrophicdeath.Thisconnection,strengthenedthroughrepeated
exposure,becomesreadilyavailabletoexpressthehumiliationofspeakingwithexcessivecandor
http://web.pdx.edu/~cgrd/Metaphors%20of%20Conceptual%20Integration.html 17/24
20171125 MetaphorsofConceptualIntegration
orneedinghelptoovercomeasetback(Yourediggingyourselfintoahole)ortheexpectationof
catastrophiclossassociatedwithariskyventure(Yourediggingyourowngrave.)Asthe
metaphoricalassociationsareused,thedirectneurologicalconnectionsbetweenthesemantic
expressions,thesocialsituationstowhichtheyapply,andtheexpectationsandemotionsthey
evokebecomeincreasinglystrong,untiltheexpressionsbecomelexicalized,andthevividimagery
thatlentpowertothesemetaphorsinthefirstplacemayfadeaway.Often,theimagesthat
underliemetaphoricalidiomsbecomedisassociatedbecausetheexperienceonwhichtheyare
basedbecomesuncommon.Itisunlikelythatmanypeopleinmodernsocietyhavedrivena
vehicledownamuddyanddeeplyruttedroad,makingitallbutimpossibletoturnthevehicle:For
manypeople,inarutmayhaveentirelylostitsmetaphoricalgrounding,andthusbecomeopento
alternativeandevenoppositeinterpretations,inawaysimilartothemetaphoricalidiomsstudied
byKeysarandBly(1999).
Similarly,diggingyourowngravemayhavebecomelargelylexicalized,tothepointthat
peoplewhousethisidiommaynotthinkof,ormaynotevenbeawareof,theconnectionto
hazardousoccupationssuchasmining,welldigging,andtunnelconstruction.Thereisanentire
setofidiomsassociatedwithdiggingandrelatedmetaphors,anditseemslikelythatthe
metaphorsinthisgroup,likesomeoftheconceptualmetaphorsdiscussedbyLakoffandJohnson
(1980),areatleastlooselyassociatedwitheachother(Ritchie,2003).Totheextentthatthese
expressionshavebecomelexicalizedandhavelosttheirgrounding,theassociationislikelytobe
ratherweak,buttheremainingexpectations,emotions,perceptions,etc.thatareassociatedor
linkedtoconceptssuchasdigging,hole,andgravepreservethepoweroftheseexpressionsto
evokestrongimagesandemotionalresponses.Thatisnottosaythatmetaphoricalidioms
necessarilyimplicatesingularunderlyingconceptualmetaphors:AsVervaekeandKennedy(1996)
havepointedout,almostanymetaphoricalexpressionsupportsawidevarietyofinterpretations.
Forexample,thegrassisalwaysgreenerontheothersideofthefencecanbeinterpretedin
termsofcattleinafencedpastureorsuburbanitescomparingtheirneighborslawntotheirown
(Ritchie,2003).Evenstuckinarutmightconceivablybeinterpretedintermsofthereproductive
biologyofelkandotherbiggameanimals,whichareoftenunableeventoeatduringmating
season.
Interpretationofmetaphorsandidioms,liketheotherexamplescitedbyFauconnierand
Turner(2002),certainlyinvolvesintegratingconceptsinsomeway,andFauconnierandTurner
http://web.pdx.edu/~cgrd/Metaphors%20of%20Conceptual%20Integration.html 18/24
20171125 MetaphorsofConceptualIntegration
provideawealthofusefulinsightsaboutthisprocess.Moreover,althoughIhavearguedthat
manyoftheirexamplescanbeanalyzedmorestraightforwardly,itisnonethelesstruethathumans
arecapableofspinningelaboratecounterfactualtales,suchasthescenarioofamonkmeeting
himselfonatripdownamountain,amodernyachtracingwiththeghostofa19thcenturyclipper
oraphilosophyprofessorengaginginanargumentwithKant.Myfatherhasbeendeadforover
fifteenyears,butIcaneasilyimagineaconversationwithhimaboutIraqoraboutthedirectionmy
careerhastakensincehisdeath.Allofthesefeatsofimaginationrequiretheintegrationof
complexnarrativesandscenariosinawaythatdefieseverydaylogic,andmuchofFauconnierand
Turnersaccountofconceptualintegrationishelpfulinseeinghowweaccomplishthem.However,
thespaceandblendingmetaphorsintroduceentailmentsthatareinatleastsomecases
misleading,andthatneedlesslycomplicatethemodel.Metaphorsbasedonnetworksorwebs
mightalsoleadtoproblematicentailments,buttheywouldatleastrequirefewerunsupported
assumptionsabouttheactualfunctioningofsynapticconnectionsinthebrain.
ConcludingRemarks
ConceptualIntegrationTheory(Fauconnier&Turner,19982002)providesapotentially
powerfulmodelforexplainingawiderangeofcognitiveprocessesthatbuildconnectionsamong
differentsortsofideassuchasamonkclimbingupamountain,thesamemonkdescendinga
mountain,andtwolinesonagraph.However,theauthorsrelianceonthemetaphoricallanguage
ofspaceandblendingobscuresasmuchoftheunderlyinglogicoftheprocessasitilluminates,
andleadstoentailmentsthatareambiguousandevencontradictory,suchastheentailmentthat
conceptualintegrationrequiresthepresenceofagenericspaceindependentofthetwoinputs,
andleadstocreationofaseparateblendedspace.Formanyoftheexamplesanalyzedby
FauconnierandTurner(2002),asimplerandmorestraightforwardanalysisseemssufficient.No
independentconceptualstructure(blendedspace)isneeded.Somethinglikeagenericspace
mayplayaroleinsomeinstances,butsometimesitinvolveselementscommontotheinputs(asit
doesinFauconnierandTurnerssolutiontothemonkpuzzle),andatothertimesitinvolvesother
connectingprinciples.Forexample,inthegraphicsolutiontothemonkpuzzlethemonks
journeysonthemountainmustbemappedontotwolinesonagraph.Theconnectingprincipleis
betterdescribedasalearnedschemaorconventionforspatialrepresentationofsequentialevents,
andisnotwellcapturedbythemetaphorofagenericspace.
http://web.pdx.edu/~cgrd/Metaphors%20of%20Conceptual%20Integration.html 19/24
20171125 MetaphorsofConceptualIntegration
ThemetaphoricallanguageusedbyFauconnierandTurner(19982002)mayhelpto
facilitateinitialunderstandingofthetheory,butitmayalsoleadthenoviceastray.Whatis
necessaryistorecognizethelimitationsandpitfallsofsuchmetaphoricallanguage,andto
elaboratethetheorywithouttheirrelevantandevencontradictoryentailmentsofthemetaphors.
Reconceptualizedwithouttheentailmentsofthespaceandblendingmetaphorsandtheir
entailments,ConceptualIntegrationTheorypromisestocontributetomajoradvancesinour
understandingoflanguageprocessing,andofmetaphorinparticular.Aformulationofthetheoryin
termsofneuralconnectionsorotherphysiologicallyrealizablemechanismswouldhelpprovidea
basisforspecifyingfalsifiablehypotheses,andforlinkingthemodeltotheworkofothercognitive
theoristsandtoresearchonneurologicalprocessesinthehumanbrain.
TheHazardsofEvocativeMetaphors.Asimilarcriticismcanbeleveledagainstthe
metaphoricalentailmentsofKintschs(1998)connectionistmodeloflanguagelearningand
processing.ItappearsfromaclosereadingofthetextthatKintschoriginallyintroducedavector
representationofconceptmeaningsasawayofmetaphoricallyrepresentingthestrengthof
synapticconnectionsamongneuronsorneurongroupsincomputableform.However,inlater
sectionsofthesamebook,andinotherwritings(Kintsch,2000Landauer&Dumais,1997),
Kintschandhiscolleaguesseemtoreifythevectorrepresentationbytreatingitasamoreorless
literalmodelofwordmeanings.
Hutchins(1995)hassharplycriticizedtheentirecomputermetaphorforcognition,arguing
thatithasledcognitivescientiststofocusonanartificiallyconstrainedmodelofinformation
processingthatdisregardsthecrucialcontributionsofsocialandculturalstructuretocognitionin
thewild.Wearestillalongwayfromunderstandinghowcomplexconceptsarerepresentedin
humanmemory,andconsequentlyfromunderstandinghowcomplexconceptsarecombinedand
appliedinprocessingdiscourse,figurativeorotherwise.Untilwehavedetailedknowledgeofthese
processesattheneuronallevel,wewillalmostcertainlyhavetorelyonevocativemetaphorsto
helpusunderstandandreasonaboutcognitiveprocesses.Rigorouscriticismofourown
metaphorswillberequiredifwearenottobecomeenmeshedintheirunintendedbutseductive
entailments.
Bibliography
http://web.pdx.edu/~cgrd/Metaphors%20of%20Conceptual%20Integration.html 20/24
20171125 MetaphorsofConceptualIntegration
Aunger,R.(2000).Conclusion.InR.Aunger,ed.,Darwinizingculture:Thestatusof
memeticsasascience(pp.205232).Oxford,England:OxfordUniversityPress.
Blackmore,S.(1999).Themememachine.Oxford,England:OxfordUniversityPress.
Calvin,W.H..(1996).Thecerebralcode:Thinkingathoughtinthemosaicsofthemind.
Cambridge,MA:MITPress.
Carroll,L.(1923).AliceinWonderland.NYC,NY:Holt,Rinehart,andWinston.
Clark,H.H.(1996).Usinglanguage.Cambridge,England:CambridgeUniversityPress.
Coulson,S.,andMatlock,T.(2001).Metaphorandthespacestructuringmodel.Metaphor
andSymbol,16(3&4),295316.
Dawkins,R.(1993).Virusesofthemind.InB.Dahlbom,ed.,Dennettandhiscritics:
Demystifyingmind(pp.1327).Oxford,England:Blackwell.
Deacon,T.W.(1997).Thesymbolicspecies:Thecoevolutionoflanguageandthebrain.
NYC,NY:W.W.Norton.
Dickens,H.(1972).DisasterattheManningtonMine.Boston,MA:Rounder.
Fauconnier,G.(1994).Mentalspaces:Aspectsofmeaningconstructioninnatural
language.Cambridge,England:CambridgeUniversityPress.
Fauconnier,G.,andTurner,M.(1998).Conceptualintegrationnetworks.Cognitive
Science,22(2),133187.
Fauconnier,G.,andTurner,M.(2002).Thewaywethink:Conceptualblendingandthe
mindshiddencomplexities.NewYork,NY:BasicBooks.
Fothringhame,D.K.,andYoung,M.P.(1997).Neuralcodingschemasforsensory
representation:Theoreticalproposalsandempiricalevidence.InRugg,M.D.,Ed.Cognitive
Neuroscience,pp.4776.Cambridge,MA:MIT.
Friedman,M.(1992).Buriedalive:ThebiographyofJanisJoplin.Boston,MA:Harmony
Books.
http://web.pdx.edu/~cgrd/Metaphors%20of%20Conceptual%20Integration.html 21/24
20171125 MetaphorsofConceptualIntegration
Gernsbacher,M.A.,Keysar,B.,Robertson,R.W.,andWerner,N.K.(2001).Theroleof
suppressionandenhancementinunderstandingmetaphors.JournalofMemoryandLanguage,
45,433450.
Gibbs,R.W.,Jr.(1999).Propositionsinmind:Aparadigminpractice.APAReviewof
Books,44,310312.
Gibbs,R.W.,Jr.(2000).Makinggoodpsychologyoutofblendingtheory.Cognitive
Linguistics,11,347358.
Gibbs,R.W.,Jr.(2001).Evaluatingcontemporarymodelsoffigurativelanguage
understanding.MetaphorandSymbol,16(3&4),317333.
Glucksberg,S.andMcGlone,M.S..(1999).Whenloveisnotajourney:Whatmetaphors
mean.JournalofPragmatics,31,15411558.
Grady,J.(2000).Cognitivemechanismsofconceptualintegration.CognitiveLinguistics,
11,335345.
Grice,H.P.(1975).Logicandconversation.InP.ColeandJ.L.Morgan,(Eds.),Syntaxand
Semantics,(vol.3):SpeechActs(pp.4158).NewYork:AcademicPress.
Harder,P.(2003).Mentalspaces:Exactlywhendoweneedthem?CognitiveLinguistics,
14,9196.
Hutchins,E.(1995).Cognitioninthewild.Cambridge,MA:MITPress.
Keysar,B.,andBly,B.M.(1999).Swimmingagainstthecurrent:Doidiomsreflect
conceptualstructure?JournalofPragmatics,31,15591578.
Kintsch,W.(1998).Comprehension:Aparadigmforcognition.Cambridge,England:
CambridgeUniversityPress.
Kintsch,W.(2000).Metaphorcomprehension:Acomputationaltheory.Psychonomic
Bulletin&Review,7,257266.
Kintsch,W.,andBowles,A.R.(2002).Metaphorcomprehension:Whatmakesa
metaphordifficulttounderstand?Metaphorandsymbol,17,249262.
http://web.pdx.edu/~cgrd/Metaphors%20of%20Conceptual%20Integration.html 22/24
20171125 MetaphorsofConceptualIntegration
Krakauer,J.(1998).Intothinair:ApersonalaccountoftheMountEverestDisaster.New
York,NY:Villard.
Kuper,A.(2000).Ifmemesaretheanswer,whatisthequestion?InR.Aunger,ed.,
Darwinizingculture:Thestatusofmemeticsasascience(pp.176188).Oxford,England:Oxford
UniversityPress.
Lakoff,G.,andJohnson,M.(1980).Metaphorsweliveby.Chicago,IL:Universityof
ChicagoPress.
Lakoff,G.,andSweetser,E.(1994).ForwardtoFauconnier,G.,Mentalspaces:Aspects
ofmeaningconstructioninnaturallanguage.Cambridge,England:CambridgeUniversityPress.
Landauer,T.K.,andDumais,S.T.(1997).AsolutiontoPlatosproblem:Thelatent
semanticanalysistheoryofacquisitioninduction,andrepresentationofknowledge.Psychological
Review,104,211240.
Petty,R.E.,andCacioppo,J.T.(1981).Attitudesandpersuasionclassicand
contemporaryapproaches.Dubuque,Iowa:W.C.BrownCo.Publishers.
Phillips,W.A.(1997).Theoriesofcorticalcomputation.InRugg,M.D.,Ed.Cognitive
Neuroscience,pp.1146.Cambridge,MA:MIT.
Reddy,M.J.(1993).Theconduitmetaphor:Acaseofframeconflictinourlanguageabout
language.InOrtony,A.(ed.),MetaphorandThought,2ndEd,pp.164201.Cambridge,England:
CambridgeUniversityPress.
Ritchie,L.D.(2003).ARGUMENTISWAROrisitagameofchess?Multiple
meaningsintheanalysisofimplicitmetaphors.MetaphorandSymbol,18,125146.
Seeger,P.(1958).SpringhillMinedisaster.Bethlehem,PA:SingOutMusic.
Sperber,D.(2000).Anobjectiontothememeticapproachtoculture.InR.Aunger,ed.,
Darwinizingculture:Thestatusofmemeticsasascience(pp.163174).Oxford,England:Oxford
UniversityPress.
Sperber,D.,andWilson,D.(1986).Relevance:Communicationandcognition.
Cambridge,MA:HarvardUniversityPress.
http://web.pdx.edu/~cgrd/Metaphors%20of%20Conceptual%20Integration.html 23/24
20171125 MetaphorsofConceptualIntegration
Veale,T.,andODonoghue,D.(2000).ComputationandBlending.CognitiveLinguistics,
11,253281.
Vervaeke,J.andKennedy,J.M.(1996).Metaphorsinlanguageandthought:Falsification
andmultiplemeanings.MetaphorandSymbolicActivity,11(4),273284.
Winograd,T.,andFlores,F.(1986).Understandingcomputersandcognition:anew
foundationfordesign.Norwood,N.J.:Ablex.
AuthorsNote
IwishtothankRayGibbs,EricJensen,andananonymousreviewerfortheirinsightfuland
helpfulcommentsonearlierdraftsofthisessay.Allremainingmistakesandomissionsare,of
course,entirelymyownresponsibility.
http://web.pdx.edu/~cgrd/Metaphors%20of%20Conceptual%20Integration.html 24/24