THE MANY FACES OF TRUTH: A RESPONSE TO SOME

Transcription

1 Forthcoming:InternationalJournalofPhilosophicalStudies. NOCITATIONWITHOUTPERMISSION **************UNCORRECTEDDRAFT*********** THEMANYFACESOFTRUTH:ARESPONSETOSOME CRITICS MICHAELPATRICKLYNCH UNIVERSITYOFCONNECTICUT INTRODUCTION InTruthasOneandMany,(hereafterTOAM)Iarguethattruth isbestdefinedasafunctionalproperty.todefinetruth functionallyistodefineitbywayofitsconnectionstoother relatedconcepts.theseconnectionsareembodiedincertain commontruismsthathaveplayedacentralroleinthe historicaldiscussionsovertruth.threeofthesetruismsareas follows. 1 ForeverypropositionP, OBJECTIVITY:PistrueifandonlyifwerePtobebelieved, thingswouldbeastheyarebelievedtobe. NORMOFBELIEF:ItisprimafaciecorrecttobelievePif andonlyifpistrue. ENDOFINQUIRY:Otherthingsbeingequal,ifPistrue, thenbelievingpisaworthygoalofinquiry. Theideaisthatthesetruisms,oronesverymuchlikethem, togetherwithcertainobviousplatitudesconnectingtruthwith validity,knowledgeandthelike,jointlypickoutconceptually essentialfeaturesofthetruthproperty.wemightputitlike this:truthjustisthepropertythathasthesefeatures essentially.havingthatpropertyiswhatconstitutesa proposition sbeingtrue.itisinthissensethatthefunctionalist thinksthattruthisone. 1 Thereareavarietyofwaysonecanstatethesetruisms.HereIfollow Pedersen s(xxx)regimentationofmymoreinformalpresentationintoam.

2 TheManyFacesofTruth:Draft Truth sbeingafunctionalpropertyisconsistentwith whatissometimescalledtruthpluralism,ortheideathatthere ismorethanonepropertyinvirtueofwhichpropositionsare true. 2 Thatis,itisconsistentwiththeideathatwhetheragiven propositionistrueisdeterminedbyitshavingsomeother, ontologicallydistinctproperty perhapsarepresentational property,likecorrespondence,orperhapsanepistemically constrainedpropertylikesuperwarrant.accordingtothe functionalist,wecanpickoutapropertythatdetermines whetherapropositionistruebyseeingwhetheritplaysthe truth role.suchpropertiesshouldnotbeconfusedwithtruth itself,buttheycouldbedescribedaspropertiesthat,under certainconditions,andforcertainkindsofcontent,realize truth.toampresentsatheoryofthekindofrealizationin question:whaticallmanifestation.iftruthisafunctional property,thentruthisopentomultiplemanifestation.itisin thissensethattruthismany. Functionalismabouttruthopensthedoortosome interestingnewsolutionstooldproblems.forexample:in somedomains,weareinclinedtothinkboththat(a)beliefsin thatdomainarecapableofbeingtrue;but(b)suchbeliefs don trepresentreality.thedomainofmoralbeliefsisa plausible ifcontroversial example.ontheonehand,we tendtothinkthatatleastsomeofourmoralbeliefscanbetrue. Wecanhavereasonsforsuchbeliefs,andwecanhavemoral knowledge.ontheotherhand,itisdifficulttoseeourbeliefsas representingdistinctivemoralpropertiesofactions.ifwe adoptfunctionalismabouttruth,itseemspossibleto accommodatebothintuitions.wecanallowthatsomebeliefs (andtheircontents)aretrue,butnottruebyvirtueof representinganything.andthatinturnallowsustoexplain howwecanhave,e.g.moralknowledgewithoutneeding recoursetorepresentationorontologicallysuspiciousmoral properties. Inthispaper,Irespondtosomeobjectionstothese views,recentlypublishedinthisjournalbybenjaminjarvis, DouglasEdwards,NikolajPedersenandNiallConnelly. RESPONSETOEDWARDSANDJARVIS DouglasEdwardsandBenjaminJarvisoffercriticismsthat whiledifferingincertaindetails runlargelyparallel. Consequently,I lldiscusstheirpaperstogether. 2 EarlystatementsofthisviewcanbefoundinCrispinWright(1992)andin Lynch

3 M.P.Lynch EdwardsandJarvisare,invaryingrespects, sympatheticwithideathattruthcouldbeplural.andtheyeach notethatthefunctionalistversionofthisideahassomeaffinity withstandarddeflationaryviewsoftruth.afterall, functionalistsanddeflationistsagreethattruthitselfcanbe definedintermsofitsfunction.thisdistinguishesboth functionalismanddeflationismfrommoretraditionaltheories oftruth.butthestandarddeflationisttakesthatfunctiontobe exhaustedbytruth sexpressiverole.accordingtothe deflationist,truth ssolepointistoactasalogicaldevicefor generalization,andthereforeitcanbedefinedsolelyinterms ofasingletruism:theequivalenceprincipleort schema.inmy view,andintheviewofedwardsandjarvis,thisconceptionof truthistoolimited:truthalsoactsasthenormofbeliefandas agoalofinquiry.consequently,theconcepthasnormative content.asjarvisputsit,thismeans(amongotherthings)that tobelievethatitistruethatrosesareredcommitsonetothe correctbelievabilityofthecontentthatrosesarered(xx).by holdingthattruthhasanormativedimensioninthisway, functionalismgoesbeyondthestandarddeflationaryaccount oftheconcept.moreover,asjarvisandedwardsalsoagree,it alsogoesbeyondthestandarddeflationaryaccountofthe propertyoftruthaswell.forthefunctionalist,asnotedabove, truthcanbeplurallymanifested. Buthowpluralisplural?Inthebook,Ientertaintwo theoriesofpropertiesthatmightmanifesttruth.oneisa successortheoryofsortstocorrespondencetheoriesofold. Theotherisasuccessortotraditionalepistemictheoriesof truth,suchasthecoherenceorpragmatistaccounts.both EdwardsandJarvisthinkthattherearefarmoreproperties thatcanplaythetruth rolethanthesetwo.asjarvisputsit,if truthisplural,itis maniacallyplural.thereisnotjustmore thanonemanifestingproperty.rather,thereareasmanyas therearepropositions.consequently,functionalismisfarmore deflationaryinspiritthani veallowed. EdwardsandJarvisofferslightlydifferentarguments forthisclaim.accordingtoedwards,icommitmyselftoa moredeflationaryviewmerelybyholdingthattruthcanbe definedfunctionally. AsInotedabove,todefinetruthfunctionallyistodefine itbythesumofthoserelationalfeaturespickedoutbythe truismsobjective,normofbeliefandendofinquiry.callthe featuressopickedoutthetruishfeatures.thuspropositions aretruewhentheyhaveapropertythatplaysthetruth role. AndapropertyTplaysthatrolejustwhenithasthetruish features,oris,aswemightsay,truish. 3

4 TheManyFacesofTruth:Draft InTOAM(p.72),Ipointoutthatthiscommitsoneto thinkingthat,(where TR standsforapropertythatplaysthe truth role): (S)PisTRif,andonlyif,wherePisbelieved,thingsare astheyarebelievedtobeiffpistr;otherthingsbeing equal,itisaworthygoalofinquirytobelievepifpis TR;anditiscorrecttobelievePiffPisTR. Edwardsnotesthatifweputtogether(S)withtherelevant instanceof(es) (ES)PistrueifandonlyifP, wecaninfer,viasubstitution,foranyarbitrarypropositionp, (SW)PisTRifandonlyif,wherePisbelieved,things areastheyarebelievedtobeiffp;itisaworthygoalof inquirytobelievepifp;itiscorrecttobelievepiffp. 3 Thequestioniswhatthisshows.Edwardstakesittoshowthat foreverysingleproposition,therewillbeadifferent manifestationoftruth (Edwards,xx).Iamnotsosure. Onereasontodoubtisthat,sofar,wedon'tactually haveanyrivalcandidatepropertiesonthetable.inorderfor thattobethecase,we dhavetofollowsomethinglikejarvis s suggestion,andsaythatthetruthmanifestingpropertyforthe propositionthatgrassisgreenisthepropertyofbeingsuch thatgrassisgreen,andthetruthmanifestingpropertyforthe propositionthatsnowiswhiteisbeingsuchthatsnowiswhite andsoon.thethoughtisthatforeachproposition,thereisa differentpropertythatcouldplaythetruth role:itisa propertyoftheform:beingsuchthatp.callthisajarvisproperty.somewillraisetheireyebrowsatjarvis properties. Andwelltheymight.Muchhangsonhowonereadsthe being suchthat.supposeweask:whatisit,ingeneral,fora propositiontobesuchthattheworldisacertainway?that seemsa lotlikeaskingwhatitisforapropositiontobetrue.so Jarvispropertiesdon tseemtogetusveryfar. 4 3 Inordertosecuretheinference,we dneedtoconsidersubstitution instances.butthepointremainsthesame. 4 Ofcourse,Jarvismightsaythatthereisnogeneralanswertotheabove question thatis,afterall,justthedeflationist spoint.butthenwhydowe needrecoursetojarvispropertiestomakeit? 4

5 M.P.Lynch 5 Butputthataside;I llgranttheexistenceofjarvisproperties.let sthinkaboutwhetherthefunctionalistshould feelanymorepressureherefromedwards argument.note thataverysimilarpointmightbemadeagainstother substantivetheoriesoftruth.consider,forexample,a correspondencetheorythatholdsthatforeverypropositionp, (CT)PistrueifandonlyifPCORRESPONDStothefacts. HereCORRESPONDENCEisunderstoodasametaphysicallyrich relation.nowpresumablyouradvocateof(ct)alsoaccepts (ES).Ifso,thenitseemsforanyarbitrarypropositionP,wecan infer PCORRESPONDStothefactsifandonlyifP. Wouldthisbyitselfshowthatthisversionofthe correspondencetheoryismaniacallypluralordeflationary? Well,inatrivialsenseitwould.Afterall,whichfacta propositionmustcorrespondtoinordertobetruedependson thepropositionitis.butthiswouldpresumablynotconcern ourimaginarycorrespondencetheoristverymuch.norshould it.theyneedn tdeny(es)oritsconsequences.theirpointis thatthistruismdoesn texhaustalltheimportantfactsabout truth.indeed,theythinkthatanappealtoageneralpropertyof correspondenceisrequiredtoexplainthattruism.thatiswhat makestheview substantive.thethoughtisthat correspondenceisthepropertythatalltruthshaveincommon, anditisthehavingofthispropertythatexplainswhytrue propositionsobjective,correcttobelieve,aworthyaimof inquiryand,ofcourse,(es). 5 Inthesamevein,itisnotpartofmyviewtodeny(SW) anditscousins.(sw),asi venoted,followsdirectlyfromour functionaldefinitionoftruthintermsofthetruismstogether with(es).therealquestioniswhether,forcertainsortsof propositions,theremustbeafurtherproperty,distinctfrom truth,thehavingofwhichmakesitthecasethatthis propositionssatisfiesthetruisms thatexplains,inshort,why thepropositionistrue.thatthisissoisakeypartofthe functionalistview.supposetherereallyarejarvisproperties andsomeofthepropositionsibelieveareluckytohavethem. Thatis,supposethatthe(true)propositionthatsnowiswhite 5 Whynotmention(ES)asoneofthecoretruismsabouttruth?InTOAM,I arguethatitisentailedbyobjectivityalongwithsomeobviouspremises aboutwhatbearscontent(seetoam,9).

6 TheManyFacesofTruth:Draft hasthepropertyofbeingsuchthatsnowiswhite.wouldthe factthatthisbeliefofminehasthispropertyexplainwhyitis correcttobelieve,objectiveandsoon?iseenosenseof explanation inwhichitwould.putanotherway,thequestion athandiswhetherthefunctionalistmustacceptthatthata truebelief sjarvis propertyisthepropertythatmanifests truth,orwhetheritonlyhasitsjarvis propertyinvirtueof havingsomemoresubstantiveproperty.edwardsargumentis consistentwitheitherhypothesis. Nowifonehasdeflationistsympathies,itwon tmatter whetherapropositionorbelief sjarvis propertyisexplanatory ornot.deflationistsdon tthinkthereisanythingtoexplain abouttruthotherthanitsroleasanexpressivedevice.but whetherdeflationismistrueisnottheissuehere.theissue hereiswhether,evenifwegrantthateachtrueproposition hasajarvis property,thefunctionalistmustacceptthatatrue belief sjarvis propertyisapropertythatmanifestsor determineswhetherthatbeliefistrue,orwhether,likeany robusttheoristoftruth,shecaninsistthereismoretosay. Jarvisacknowledgesthatsomephilosophersthinkthere ismoretosay,andadvocatesofrobustcorrespondenceviews (ofcorrespondencethatis)areamongthem.but,hesays,there isnogeneral,non trivialtheoryofcorrespondencetoappeal to.intoam,iagreethattraditionalcorrespondencetheories like(ct)abovearenotveryplausible.theyeitherendup beingmereplatitudes(notexplanations)ortheyrequirethere tobesentence shapedfactsintheworld.butthatdoesn tmean wemustgiveuponthebasiccorrespondenceintuition.in TOAM,Iarguedthattherearepromisingsuccessorstothe traditionalcorrespondencetheories.thesetheoriesarethe productoftherepresentationaltheoriesofmindandlanguage foundincognitivescience. Themostplausibleextenttheoriesthatbuildonthe correspondenceplatitudearecomposedofthreeelements. First,theytakethetargetofexplanationtobetruebeliefs. 6 Thisdoesnot,surprisingly,requirethemtodenythethought thatpropositionscanbetrueorfalse,norneedtheydenythe thoughtthatpropositionsarethecontentsofattitudeslike beliefs(seetoam,28 29foranargument)Whattheyare committedtoisheideathatwhatourtheoriesoftruthshould betryingtoexplainishowpropositionswebelievearetrue. Thesecondelementreducesthetruthofabelieftothe representationalrelations(orreferentialrelations)ofits 6 Ortruesentences.EarlyversionsoftheviewsI mdiscussinghereare foundindevitt*1997)andfield(1972). 6

7 M.P.Lynch 7 componentconcepts.thethirdelementofanysuchview, obviously,isatheoryofwhatrepresentationis.themost common,andithink,mostreasonabletheoriesof representationholdthatfactsaboutrepresentationare identicalto,orsuperveneon,naturalfacts.theguiding metaphoristhemap:ourbeliefsrepresenttheworldasmaps representtheenvironment.hereisatoyversion: REPRESENT:ThebeliefthataisFistrueifandonlyifthe objectcausallymappedby a hasthepropertycausally mappedby F. Somemightsaythattheorieslikethisare really theoriesof contentdetermination,nottheoriesoftruth.jarviswisely demurs,notingthat, thetheoryoftruthisjustwhateverthe theoryofattitude,contentandtruthconditionsneedsittobe (xx).injarvis s,mineanddavidson sview,whatdeterminesa giventheoreticalelementisproperlyconsideredanelementof atheoryoftruthdependsonwhetherweneedtoappealtothat theoreticalelementtodotheworkweneedatheoryoftruthto do:thatis,tohelpusexplainthenatureofcontent,attitudes andtruthconditions.onthiswayoflookingatit, representationaltheoriesthatimplysomethinglikerepresent aretheoriesoftruthandcontent. SoJarvisagreesthattheorieslikeREPRESENTaretheories oftruth.butheobjectsthattheyarenotsubstantive.why? BecauseaviewlikeREPRESENTisjustarecipe: itdependson whatwesubstituteinfortheschematic a s and F s ;andso whatrepresentationalcorrespondenceis justdependsonthe content typejustasmaniacalpluralismsaysthatitdoes. JarvisiscertainlyrightthatREPRESENT,asstated,isnota generalization.inordertoquantifyoverit,we dneeda substitutionalquantifierorfunctionallysimilardevice;thatof coursecouldbedone.butwhydoweneedto?whyshouldthe factthatrepresentisarecipeprecludeusfromthinkingthat thereisageneralaccountofrepresentationorcorrespondence intheoffing?afterall,representitselfisn tmeantasatheoryof whatrepresentationis;thatwillbegivenbyasubstantive accountofhowmentalstatescausallymapobjectsinthe environment.indeed,suchtheories beingastheyare, theoriesofhowmentalstatesrelatetotheworld will presumablybepartlyempiricalinnature,bestdevelopedin conjunctionwithcognitivescience. Putdifferently,REPRESENTismeantasatheoryofwhat makesabelieftrue.butthattheoryclaimsthatthereisa dependencerelationbetweenabelief sbeingtrueandhowthe

8 TheManyFacesofTruth:Draft partsofabeliefcausallymaptheworld.thatis,represent, whileabiconditional,istoberead Socratically.Whatmakes beliefstrueisthattheycausallymaptheworldasitis;they don tmaptheworldasitisbecausetheyaretrue.suchviews maybefalse.theymightbemisguided.buttheyarecertainly substantive,preciselybecauseitisasubstantivequestionwhat causalmappingconsistsin. Edwardsalsomakesafurtherargument.Theviewhe callsdeflationarypluralismissuperiortothemoresubstantive versionbecauseitavoidstheproblemof mixedcompounds. Thisistheproblemofhow,iftruthisinsomewayplural,to understandthetruthofapropositionlike: MIX:Murderiswrongandgrassisgreen. Supposethatthepropertythatmanifeststruthformoral propositionsisdistinctfromthepropertythatmanifeststruth forpropositionsconcerningthephysicalworld.nowask:what manifeststruthformixitself? Avarietyofanswerstothisproblemhavebeen proposed. 7 InTOAM,Iarguethatasimpleanswertheproblem dropsrightoutofmytheoryofmanifestation.iwon trehearse thatanswerhere,however.edwards thoughtisthatwhathe callsdeflationarypluralismissuperiortootherformsnot becauseithasabetteranswer,butbecauseitneedsnoanswer. Accordingtodeflationarypluralism,eachcompound proposition,mixedornot,willspecifyitsownmanifestationof truth.whatmanifeststruthformixisitsjarvis property:being suchthatmurderiswrongandgrassisgreen.consequently, thereisnopuzzleaboutwhichpropertymanifeststruthfora mixedcompound. Ihavetwopointstomakeinreply.Thefirstisthatthe problemofmixedcompoundsisn tjustaproblemspecificto functionalistorpluralisttheoriesoftruth,ortheoriesoftruth generally.itissimplyoneaspectofalargerproblem what I vecallintoamtheproblemoftheuniversalityofreason. Thislargerproblemarisesforanyviewwhich(a)takesthere tobedistinctivekindsofcontent(saynormativeand descriptive);and(b)takesitthatsomecontentsarenottruein virtueoftherepresentationalrelationsthosecontents(andthe 7 7 AsamplingoftheliteraturehereincludesWilliamson,1994;Tappolet 1997;2000;N.Pedersen2006,Edwards2008,2009,Cotnoir2009,Lynch, 2001,2004,

9 M.P.Lynch 9 mentalstatesdirectedatthem)beartotheworld.anyview likethis suchasexpressivism,forexample mustsayhowit isthatwecanreasonandmakeinferencesacrossdifferent typesofcontent.consequently,theabovepoint thattheories ofcontentandtheoriesoftruthtravelintandem comesinto playhereaswell.toavoidtheproblemofmixedcompounds, oneneedsnotjustadeflatedtheoryoftruth,butadeflated theoryofcontent onewhichdoesnotdistinguishbetween differentkindsofcontent. Hence,mysecondpoint.Perhapsthedeflationary pluralistadoptsadeflatedtheoryoftruthandcontent.ifthey do,theniagree:thereisnopuzzleaboutmixedcompoundson suchaview.indeed,thereisnopuzzleabouthowreasoncan beuniversaleither.butitdoessoonlyatthecostofadoptinga bloodlessquietismandignoringtherealdifferencesbetween ourthoughts. REPLYTOPEDERSEN WhereJarvisandEdwardsaimatmyaccountoftheproperties thatplaythetruth roleordeterminetruth,n.j.l.l.pedersen aimsatmyviewofthepropertyoftruthitself. Whatisthatview?Howdothepropertiesinvirtueof whichpropositionsaretruerelatetothepropertyoftruth itself.intoam,igivethefollowinganswer.recallthata propertyplaysthetruth rolejustwhenithasthetruish features.forsomepropositions,anepistemicpropertylike superwarrantplaysthetruth role.whenitdoes,ithasthe truishfeatures.butsuperwarrantdoesn tplaythetruth role foreverykindofproposition.sosuperwarrant,ifithasthe truishfeatures,hasthemaccidentally(toam,78). Consequently,superwarrantisn ttruth.thepropertyoftruth itselfisthepropertythathasthetruishfeaturesessentiallyor whichplaysthetruth roleassuch.itisthepropertythatis, essentially,hadbybelievedcontentsjustwhenthingsareas theyarebelievedtobe;hadbypropositionsbelievedattheend ofinquiryandwhichmakespropositionscorrecttobelieve. Pedersenarguesthatthisaccountofthemetaphysicsof truthisconsistentwithholdingthattruthisactuallya disjunctiveproperty. Supposethatsomepropositionsaretruewhentheyare superwarrantedandallotherpropositionsaretruewhenthey representthingsastheyare.ifso,thenwemightsaythatthe functionalconceptoftruthjustpicksouttheaproperty definedlikethis:

10 TheManyFacesofTruth:Draft Anatomicpropositionistrue D justwhenitiseither superwarrantedorrepresentsthingsaretheyare. Isthepropertyofbeingtrue D acandidateforbeingthe propertyoftruthitself?pedersenpresentsaningeniously simpleargumentintendedtoshowthatitisbylightsofthe metaphysicalpicturegivenabove.theargumenthingesona principlethatpedersenascribestome: (SAT):Thedomain specific(orlocal)properties T1.TNsatisfythetruisms(Pedersen,p.x). Theargumenthastwoparts.Thefirstpart:Takesome propositionp,andsupposeitistrue D. Ifso,thenweknowit willalsobetrueinsomespecificway:sayitissuperwarranted. Ifsuperwarrantplaysthetruth role,weknow,by(sat)that propositionswhichhavethatpropertyarecorrecttobelieve. SoPiscorrecttobelieve.SoifPistrue D, itiscorrectbelieve. Secondpart:AssumethatPiscorrecttobelieve.(SAT), Pedersenclaims,entailsthatPissuperwarrantedif,andonlyif, Piscorrecttobelieve.Itfollowsfromthisandourassumption thatpissuperwarranted.butifpissuperwarranted,pistrue D. SoifPiscorrecttobelieve,Pistrue D. Putthetwopartstogether,Pedersenclaims,andweget: ForanyP,Pistrue D ifandonlyifpiscorrecttobelieve,and thusourdisjunctivepropertysatisfiesoneofthetruisms.put differently,p smakesithavethetruishfeatureofbeingcorrect tobelieve. Pedersenrightlynotesthatsimilarargumentscanbe madeforeachoftheothertruisms.and,heargues,because suchargumentsonlyrelyonthedefinitionofthedisjunctive property,basiclogicalinference and(sat),whichhetakesto beanecessarytruth,heconcludesthatthedisjunctive propertyshouldberegardedbymeasanequallyviable candidateforthereferentof truth. Idisagree,fortworeasons.First,asIargueinTOAM, andaspedersennotes,thedisjunctiveviewhasaparticularly hardtimedealingwiththeproblemofmixedcompounds. ConsideramixedconjunctionlikeMIXabove.Again,the questioniswhatmakestheconjunctionitselftrue.andthe conjunctionitselfissurelynottruebecauseithastheproperty of,say,beingeithersuperwarrantedorrepresentingthefacts. ForthattobethecaseMIXmusthaveoneofthe disjunct properties ;butitisnotclearwhatpropertythatwouldbe. Thepropositionthatgrassisgreenmightrepresentsomefact (orobject/propertypair).itisfarfromclearthatthe 10

11 M.P.Lynch propositionthatgrassisgreenandmurderiswrongitself representsanyfact. Pedersenarguesthatthispointshouldbeseenas supplemental.itrulesoutthedisjunctivepropertyviewby virtueofpointsseparatefrommyownaccountofwhattruthis. Assuch,hethinksInowmustsaythattruthjustisthenondisjunctivepropertythathasthetruishfeaturesessentially. Andthat,hesays,makesmyview impure.perhapsitdoes(it wouldnotbethefirsttimemylackofpurityhasbeennoted). ButIthinkwecanputthatpointaside.ForPedersen s argumentcanberesisted.idon tthinkitshowsthatbeing true D isaviablecandidateforbeingthethepropertyoftruth bymyowntheory slights. Here swhy.pedersenclaimsthatiamcommittedtothe necessarytruthof(sat).butjustwhatdoes(sat)mean?it claimsthatanypropertythatplaysthetruth rolesatisfiesthe truisms.asinotedabove,thisamountstosayingthatthe propertiesthatplaythetruth rolearetruish.butwemustbe carefulabouthowweparsethisclaimmodallyspeaking.there areleasttwointerpretations: (DAT)Necessarily,ifapropertymanifeststruth,thenit istruish. Butnecessitydededictoisdistinctfromnecessitydere.And thustheaboveclaimisdistinctfrom: (WAT)Ifapropertymanifeststruth,thenitis necessarilytruish. (DAT)istrueonmyview.But(WAT),whichconcernswhether apropertyhasthetruishfeaturesessentially,issimplyfalse.as Inotedabove,itisnotpartofmyviewthatpropertieslike superwarrantareessentiallytruish.theyonlyhavethetruish featuresaccidentally:thatiswhenpossessedbypropositions forwhichtheyplaythetruth role. TheproblemhereisthatitseemslikePedersen s argumentwouldrequirethetruthof(wat).thatis,inorderto showthatthedisjunctivepropertyisanequallygood candidateforbeingthepropertyoftruth,pedersenhasto showthatthepropertyofbeingtrue D isessentiallytruish.but hisargumentdoesnotshowthat.indeed,itisnotclearhowit could.thefactthatitisnecessarythat Pcorrespondsoris superwarranted istruishdoesn tentailthatit Pcorresponds orissuperwarranted isnecessarilytruish.indeed,ifaisnot necessarilyf,andbisnotnecessarilyf,thenhowcouldthe 11

12 TheManyFacesofTruth:Draft propertyofbeingeitheraorbbenecessarilyf? ResponsetoConnolly Inhispaper,NiallConnollyfocusesonthissecondaspectof alethicfunctionalismthatimentionedinmyintroductory remarks:thethoughtthatanappealtotruth smultiple manifestationallowsforthepossibilityofnon representational truth.heraisesanumberofinterestingpoints,twoofwhich strikemeascentral.i lltaketheminorder. Connolly sfirstpointisthat,inhisview,i failto establish thatsomeversionofthecorrespondencetheory doesn texplainthetruthofallofour(true)beliefs(p.xx).ifby establish wemean, prove thenofcoursethatisright.asi seeit,determiningthebesttheoryoftruthforanydomainof inquiryisgoingtobeamatterofbestexplanation.asissooften thecase,settlingonthebestphilosophicaltheoryisamatterof weighingcostsandbenefits.conceivedofinthatway,istill thinkitisfairlyclearthatthecorrespondencetheoryisnotthe bestexplanationofwhattruthingeneralconsistsinbecause, asinotedabove,thebestcurrenttheoriesofcorrespondence arebuiltoutoftheoriesofrepresentation,andthebest theoriesofrepresentationarenaturalist. Ofcoursethatisdebatable.Butitisadominant assumptionofcontemporarycognitivesciencethatsomething likethisiscorrect.ifitis,thenitplacesconstraintsonwhat ourmindscanrepresent.asconnollyconcedes,these constraintsarenotimplausible.rejectingthemrequires supplyingatheoryofrepresentationthatallowsourmental statestorepresentthingsthatarenotpossibleobjectsof causalinteraction.butifwethinkourmentalstateseitherare, orsuperveneonourbrainstates,itisnotclearhowthiscould bethecase. Connollyhimselfdoesnotwishtoquarrelwiththe representationaltheoryperse.instead,hewonders,wherewe findourselvesunabletoexplainhowakindofcontentbearsa naturalisticrepresentationalrelationtotheworld sayinthe caseofmorality,ormathematics whynotsimplysaythat beliefswithsuchcontentsaresimplyinerror?theerror theorist,itseemstome,facesthreeburdens.thefirstburden liesinconvincingusthatwecanevenmakesenseoftheidea thatabroadswathofourthought suchasthoseabout morality arereallyjustfalse.thesecondtaskistoexplain thepersistenceofpracticesthatemploysuchthoughts,despite theirpervasiveerror.andthethird,conceptuallypriortothe firsttwo,istoprovideareasonforthinkingthatinforming 12

13 M.P.Lynch beliefswiththetargetkindofcontent,weactuallydoaimat representingtheworld. Connollyisnodoubtawareoftheseburdens.Buthe thinksthatthealternativeofferedintoam thatatleastsome ofourbeliefsaretruebutnottrueinvirtueofrepresentingthe world isimplausible.inparticular,heisconcernedtoargue thatwhaticallconcordanceisnotaplausibletheoryofmoral truth. Connolly sobjectionagainstconcordanceasapossible manifestationofmoraltruthisthatitistoopermissive:it doesn truleoutaperniciousformofmoralrelativismbeing true. InTOAM,Iacknowledgethatifconcordancemanifests truthforourmoraljudgments,thenthetruthofsuch judgmentsmightberelativeinacertainsense.idon tthink thisstraightawayentailsthattherelativismispernicious, however.nordoithinkthatatheoryoftruthisgoingtosave usfromthesensibleknavesandthrasymacusesoftheworld, asi llnowexplain. Whetheramoralbeliefisconcordantdependsontwo factors.first,thebeliefmustsupercohereswithaframework ofmoralandnonmoralbeliefs wherethenonmoralbeliefs aretruebyvirtueofrepresentinghowtheworldis.inshort,it isnotenoughforamoraljudgmenttobetrueforittobea memberofasupercoherentframeworkofbeliefs.italsohasto supercoherewiththenon moralfacts,asitwere. Consequently,thefactthattherecouldbetwoinconsistentbut equallycoherent(orevensupercoherent)setsofmoralbeliefs isnotitselfsufficienttomakethebeliefswhichcomposethe framework equallytrue.moraltruthisconstrainedbythe nonmoraltruth. Ofcourse,asConnollypointout,thisdependson whetherwecanmakesenseofanonmoralbelieflending evidentialsupportofsomekindtosomemoralbeliefsrather thanothers.idon tseethisasdifficultmyself,nordoitakemy viewtobeunusualinthisrespect. AsIseeit,anonmoralbeliefevidentiallysupportsa moralbeliefinconjunctionwithfurthermoralbeliefs.one formsuchevidentialsupportcantakeisentailment.but nonmoralbeliefscanalsobeepistemicallyrelevanttomoral beliefsinotherways:byplayingaroleincomparative reasoningbetweencases.comparativereasoningofthissortis commoninpracticalreasoning:onenotices(orthinksone notices)comparisons,andthen,ontheprinciplethatlikecases betreatedalike,drawssomenormativeconclusion(correctly orincorrectly).iseenoreasontothinkthattheseallsuch 13

14 TheManyFacesofTruth:Draft 14 inferencesmustbeformallytractable(thatis,itneedn tbea materoflogicalconsequence).buttheycansurelybe evidential,orperceivedtobesuch. Toseehowthesepointsrelatetoconcordance,consider the19 th centurywhiteslave ownerwhoclaimsthathisafrican slavesarenothuman,andonthatbasis,togetherwiththe moralclaimthatitispermissibletoenslavenon humans, arguesthatispermissibletoenslaveafricans.herewehavea casewhereanonmoraljudgment,togetherwithamoral principle,entailsafurthermoraljudgment.theslave owner s positionisfalsebecauseitisnotconcordant.onereasonitis notconcordantisthatitisfalsethatafricansaren thuman; thatmisrepresentsthebiologicalfacts.butnowconsiderthat ourslaveownermayalsobelieveitispermissibletotreathis wifeasproperty.andhemayholdvariousfalsenonmoral beliefsaboutwomenconcerning,e.g.theirrelativeaverage intelligencetomen,theirsimilaritiestochildrenetc.these judgmentsmaywell(inconjunctionwithotherpremises) entailfurtherfalsemoralbeliefsaboutwomen.butthesefalse nonmoralbeliefsaboutwomenmayalsobepartofthegrounds onwhichhebaseshisviewsaboutthepermissibilityofslavery, simplybecauseontheirbasishedrawscomparisonsbetween themoralstatusofwomenandslavestakenfromafrica.thus hismoralbeliefswillnotbeconcordant:notonlyaretheyof dubiousinternalcoherenceinandofthemselves,theyarenot supercoherentwiththenonmoralfacts. Connollyarguesthatthisconstraintonmoraltruthis tooweak.heasksustoconsidertwosupercoherent frameworksfandf*suchthateverymoralbeliefcontainedin FhasitsnegationcontainedinF*.Couldthesesystemsbe equallyconcordantaswell?ifso,thenmoralbeliefsband~b wouldhavetobeequallysupercoherentnotjustwithother moraltruths,withanypossiblenonmoraltruththatcould becomeamemberofeithersystem.connollythinksthisisvery possible,indeed,heseemstosuggestthatforanymoralmoral beliefb,bothitanditsnegationcouldbesupercoherentwiththe verysamenonmoraltruths.hisreasoningisthis.supposef containsthebeliefsthathumansshouldbetreatedbetterthan squidandthatmoreintelligentbeingsshouldbetreatedbetter thanlessintelligentbeings.supposef*containsthebeliefsthat itisnotthecasethathumanbeingsaretobetreatedbetter thansquidanditisnotthecasethatintelligentbeingsshould betreatedbetterthanunintelligentbeings.bothframeworks, hesays,arepresumablyequallycoherentwiththefactthat humanbeingsaremoreintelligentthatsquid.hence

15 M.P.Lynch supercoherencewithnonmoralfactsdoesnothingtoconstrain whichmoralframeworksareplausiblythoughttobetrue. Idon tthinkitisaseasyforrivalmoralsystems let alonecontradictoryones! tobeequallysupercoherent,or evencoherent,withallthesamenonmoralfacts.takethe examplejustcited.itistruethat humanbeingsshouldbe treatedworsethansquid isconsistentwith humanbeingsare moreintelligentthansquid.butisitcoherent,letalone supercoherentwithotherrelevantnonmoraltruths?istheidea thathumanbeingsshouldbetreatedworsethansquid coherentlybelievableinconjunctionwithwhatisknownabout theirrelativecapacitiesforpain?considerwhatonemight havetobelieveabouttheworldinordertoactuallyholdsucha view.perhapsonemightbelievethatasquidisasupernatural being.butthatofcoursewouldrequireothernonmoralbeliefs: that,e.g.therearesupernaturalbeings.andsoon.moreover, considerwhatotherpracticallyrationaldecisionswouldbe engenderedbysincerelybelievingthatunintelligentbeings shouldbetreatedbetterthanintelligentones.couldone coherentlyapplythattooneself?perhaps,butitishardto shakethethoughtdoingsowouldrequirebelievingthatthe worldiscertainwaysitisnot. AmuchbetterexampleforConnelly spurposesisthe intelligent,consistentegoist thethrasymachusofplatoor thesensibleknaveofhume.suchcharacterscertainlyseem possible.moreover,inimaginingthemweneedn tconceiveof themasholdingthenegationofallourmoraljudgments.so Connellyasks:mightitbepossiblethattheknave smoral judgmentsareconcordant?thatis:wouldtheyremain coherentwithallfurtherincreasesofinformation,moraland nonmoral? Beforeyouanswerthisquestion,thinkaboutthis: supposeyouthinkthatmoraltruthisamatterof correspondencewithreality.youareamoralrealist.now supposeyoualsothinkthattheknave sviewsarefalse. Presumablyyoumustsaywhyyouthinkthisisthecase.Itwill notbesufficienttosaytheyfailtocorrespondtothefacts,for tosaythatwouldbetosimplyrepeatyourviewthattheyare false.whateverevidenceyoucite,whateverfailureof coherenceormistakeyounotetheknavetomake,thatvery evidencewillbeavailabletothefriendofconcordancetociteas well.therealistwilltakeittobeevidencethattheknave s judgmentsfailtocorrespond;theconstructivistwilltakeitas evidenceforthinkingthattheknave sjudgmentsfailtobe concordant. 15

16 TheManyFacesofTruth:Draft 16 Therelevanceofthispointisthis.Ourconstructivist mightwellgrantthatitislogicallypossiblethatthereare incompatiblemoraljudgmentseachofwhichisconcordant withadistinctframework.butitisafurtherpointtograntthat theknave sviewis,orcouldbeconcordantinthisway.to showthat,itwouldneedtobethecasethattheknave s positionwillneversufferarationaldefeatofanykind,no matterhowmuchnewinformationandexperiencescome downthepike.ifthatdreadfulpossibilityturnsouttobe realized,thenitisnotcleartomewhatedgetherealisthas gained.shewillbeabletosaythattheknave sviewcouldstill befalse,butwon tbeabletosaywhyweshouldthinkso.the realistandtheconstructivistwillbothbeinthesameboat, rationallyspeaking:forcedtoconfronthim,bothwillfindthat reasonshavefled,savethefivethatonehasinone sright hand. 8 References Connolly,Niall.(2012).Truth,AtMost,One.Thisjournal.xxx Cotnoir,A.(2009). GenericTruthandMixedConjunctions: SomeAlternatives.Analysis69.3: Devitt,Michael.(1997)RealismandTruth,Princeton,NJ: PrincetonUniversity. EdwardsD.(2008) HowtoSolvetheProblemofMixed Conjunctions.Analysis,68.2, (2009)TruthConditionsandTruth:ResolvingtheProblem ofmixedconjunctions.analysis69.4: (2012).Alethicvs.DeflationaryFunctionalis.This journal.xxx Field,Hartry.(1972) Tarski stheoryoftruth.journalof Philosophy,69: (1998)Truth2ndedn.,Oxford:OxfordUniversityPress. Jarvis,Ben.(2012).TruthasOneandVeryMany.Thisjournal xxx. Lynch,MichaelP.(2001) AFunctionalistTheoryofTruth in TheNatureofTruth(ed.)M.Lynch,Cambridge,Mass.: MITPress. (2004a) TruthandMultipleRealizability Australasian JournalofPhilosophy,82: Thankstomycritics,NiallConnolly,NikolajPederesen,BenJarvisand DouglasEdwards,fromwhosepapersI velearnedagreatdeal,tothe philosophydepartmentatuniversitycollegedublin,andtomaria Baghramian,forconversationsandpatience.

17 M.P.Lynch (2004b)TruetoLife.Cambridge:MITPress. (2005)AlethicFunctionalismandourFolkTheoryofTruth. Synthese,145: (2006)ReWrightingPluralism.TheMonist,89: (2009)TruthasOneandMany.Oxford:OxfordUniversity Press. Pedersen,Nikolaj,(2006)WhatCantheProblemofMixed InferencesTeachusaboutAlethicPluralism?The Monist,89: (2010)StabilizingAlethicPluralism.PhilosophicalQuarterly 60, (2012)TrueAlethicFunctionalism?Thisjournalxxx. Tappolet,Christine(1997)Mixedinferences:AProblemfor PluralismaboutTruthPredicates.Analysis,57: (2000)Truth,PluralismandMany valuedlogic:areplyto Beall.PhilosophicalQuarterly50: Wright,C.D.(2005) OntheFunctionalizationofPluralist ApproachestoTruth.Synthese145:1 28. (2010).Truth,RamsificationandthePluralist srevenge. AustralasianJournalofPhilosophy,88. Wright,Crispin(1992)TruthandObjectivity.Cambridge,Mass.: HarvardUniversityPress. 17