RoadstoReference
AnEssayonReferenceFixing inNaturalLanguage
MARIOGÓMEZ-TORRENTE
GreatClarendonStreet,Oxford,OX26DP, UnitedKingdom
OxfordUniversityPressisadepartmentoftheUniversityofOxford. ItfurtherstheUniversity’sobjectiveofexcellenceinresearch,scholarship, andeducationbypublishingworldwide.Oxfordisaregisteredtrademarkof OxfordUniversityPressintheUKandincertainothercountries
©MarioGómez-Torrente2019
Themoralrightsoftheauthorhavebeenasserted
FirstEditionpublishedin2019
Impression:1
Allrightsreserved.Nopartofthispublicationmaybereproduced,storedin aretrievalsystem,ortransmitted,inanyformorbyanymeans,withoutthe priorpermissioninwritingofOxfordUniversityPress,orasexpresslypermitted bylaw,bylicenceorundertermsagreedwiththeappropriatereprographics rightsorganization.Enquiriesconcerningreproductionoutsidethescopeofthe aboveshouldbesenttotheRightsDepartment,OxfordUniversityPress,atthe addressabove
Youmustnotcirculatethisworkinanyotherform andyoumustimposethissameconditiononanyacquirer
PublishedintheUnitedStatesofAmericabyOxfordUniversityPress 198MadisonAvenue,NewYork,NY10016,UnitedStatesofAmerica
BritishLibraryCataloguinginPublicationData Dataavailable
LibraryofCongressControlNumber:2019944138
ISBN978–0–19–884627–7 DOI:10.1093/oso/9780198846277.001.0001
PrintedandboundinGreatBritainby ClaysLtd,ElcografS.p.A.
LinkstothirdpartywebsitesareprovidedbyOxfordingoodfaithand forinformationonly.Oxforddisclaimsanyresponsibilityforthematerials containedinanythirdpartywebsitereferencedinthiswork.
Contents
Preface vii
1.Reference:ProblemsandPromises1 WhatIsReference?1 WhyStudytheFixingofReference?7 RoadstoReference:APreview11 OnReferenceandMeaning16
2.DemonstrativesandConflictingIntentions19 IntroducingDemonstratives19 Kaplan’sClassicalTheories21 TheSearchforDescriptivistConditionsforDemonstrative Reference27 ReferentialIndeterminacyasaFundamentalProblem fortheDescriptivistProject36 APictureofDemonstrativeReferenceFixing49
3.ProperNamesandReferentialIndeterminacy60 SomeClassicalKripkeanandEvansianAdvancesin OurUnderstandingofProperNames60 RecentDescriptivistViewsaboutNameReferenceFixing64 ReferentialIndeterminacyasaFundamentalProblemfor Descriptivism73 RootsofReferentialIndeterminacy80 RecentTheoriesofJointlyNecessaryandSufficient ConditionsforNameReference82 APictureofNameReferenceFixing90
4.ArabicNumeralsandtheProblemofMathematical Sophistication108 VerbalNumerals,ArabicNumerals,andReference108 ExistingDescriptivismsaboutArabicNumeralsand TheirProblems111 APictureofReferenceFixingfortheArabicNumerals122 WhattheReferentsoftheNumeralsCouldBe130
5.NounsforNaturalKindsandtheProblemofArbitrariness140 The “Kripke-PutnamOrthodoxy ” aboutNounsfor NaturalKinds140 Challengestothe “Kripke-PutnamOrthodoxy,” with SpecialAttentiontoArbitrarinessObjections148
SomeUnsatisfactoryResponsestotheChallenges154 APictureofReferenceFixingforNaturalKindNouns159 TheProblemofArbitrarinessandtheReferentsofNouns forNaturalKinds168
6.WordsforSensibleQualitiesandtheProblemofPerceptual Variation184
The “SecondaryQuality ” ViewofColorandSome ArgumentsforandagainstIt184 IntersubjectivePerceptualVariationArgumentsand SomeImplausibleObjectivistResponsestoThem190 AnAccountofReferenceFixingforColorAdjectives195 PerceptualVariationandtheReferentsofColorAdjectives203
7.ConcludingNotes213
TheStrongestReasonsforAnti-Descriptivism,and ItsProperReach213 TheReferentsattheEndoftheRoads215
References 221 Index 229
Preface
Thisbookismypersonalattempttocontributetothetheoryofreference,and specificallytosomeofthecentralissuesthatarisewhenoneaskshowthe conventionalreferenceofsomelinguisticexpressionsorexpressionusesis fixedordetermined(whenitis).Assumingthatreferenceisasemantic propertyoflinguisticexpressions(orexpressionuses),thisisthusabook aboutwhatsometimesgoesbythename “metasemantics,” thestudyofhow expressionscometohavethesemanticpropertiestheyhave.Insomecases,the bookbasicallytakesforgrantedviewsaboutwhatthereferentsofavarietyof expressionsorexpressionusesare,andsetsouttoexplainhowtheycouldhave acquiredthosereferents;alternativeviewsofwhatthereferentsoftherelevant expressionsarearefeasible,butitishopedthattheaccountsofreference fixing inthebookwouldprovidethebasisforinterestingsupplementsevenofthose alternativeviews.Inothercases,thebookofferssubstantiveargumentsin favorofparticularviewsofwhatthereferentsofcertainexpressionsor expressionusesare,aswellasofhowtheyhaveacquiredthosereferents.But evenintheselattercasesthemainfocusisonissuesofreference fixing,rather thanonproperlysemanticorothernon-“metasemantic” linguisticissues abouttherelevantexpressions.
Theissuesaboutreference fixingthatthebookfocuseson,andthetreatmentsofthemdevelopedbelow,arevariedinnature,butmanyofthemare relatedbyseveralconnectingthreads.Thus,inchapters2and3,Iconsider demonstrativesandpropernames,andseektorejectsomerecentviewson whichcertainkindsofdescriptivismareunavoidableasaccountsofreference fixingforthoseexpressions.Infact,Iseektoprovideastronganti-descriptivist considerationaboutreference fixingfornamesanddemonstratives,basedon casesofreferentialindeterminacy,thathasnot,tomyknowledge,been exploitedinthepreviousliteratureonthetopic.Ithendevelopanaccount ofreference fixingfortheseexpressionswhichiscompatiblewithantidescriptivismandwhichembracestheidea,hintedatbyKripkeandothers, thattherelevantreference-fixingconventionsneednotprovidenecessaryand sufficientconditionsforreference,butonlyimpreciseroughlysufficientconditionsforreferenceandreferencefailure.Arelatedaccountofreference fixing,involvingonlyimpreciseroughlysufficientconditions,isprovidedfor
ordinarynaturalkindnounsinchapter5.Chapter4,onArabicnumerals, arguesagainstsomeviewsthatseethereferenceofthenumeralsasunavoidably fixedbycertainsophisticateddescriptivestructures,anddevelopsinstead alessdemandingaccountofreference fixingforthenumerals.Chapters4,5, and6allseektorejectavarietyofeliminativistargumentsintendedtoshow thatreferencetonumbers,naturalkinds,andcolorpropertiesisimpossible atleastbymeansofnaturallanguagenumerals,ordinarynaturalkindnouns, andcoloradjectives.Accordingtothesechapters,thereachesofreferenceand, accordingly,oftruthextendwellbeyondwhatcurrentlypopulareliminativist considerationsallow reference fixingcantakemany(bumpy)roadstoits destinationsandsurmountobstaclesofallkinds.Ontheotherhand,chapters 2,3,and6allseektorejectinsomewayorotherviewswhich,whileakintothe viewsofthisbookinbeingbroadlyanti-descriptivistand/ornon-eliminativist aboutreferenceintherelevantareas,defendsuchanti-descriptivismandnoneliminativismatthecostofpostulatingepistemicmysteriesinthetheoryof reference fixing.
Onelesstheoreticalbut,tomymind,nolesssubstantivethreadunifying manyofthediscussionsbelowistheirKripkeancharacter.AsIthinkwillbe clearafterreadingthem,someofthemainconsiderationsinpracticallyallof thechaptersseektoestablishbroadlyKripkeanviews,oftenviaextensionsor refinementsofbroadlyKripkeaninsights.Thisistrueinparticularofchapters 2,3,and5andtheirunderlyingKripkeandistrustoftheoriesofnecessaryand sufficientconditionsforreference;ofchapters4,5,and6andtheirrealistand objectivistviewsconcerningnumbers,ordinarynaturalkinds,andcolor properties;oftheskepticism,presentinallthechapters,towardviewsthat seektoexplaintheexistenceofordinaryreferenceintermsofsurprising epistemicresourcesorunexpectedmetaphysicaltargets;andofmanyother smallerthingshereandthere.
Ihavemanypeopleandinstitutionstothankfortheirreactionstoancestors ofthismaterialandforotherkindsofinspirationorsupportforthebook,and theattempttoprovidealistofthemwouldbeanagonizinglyincomplete exercise.Imustneverthelessagonize,mentioningexplicitlymygratitudeto onlysomeofthem.IthanktheMexicanCONACyTresearchprojectCCB 2011166502andtheSpanishMINECOresearchprojectFFI2015-70707-Pfor support,andJohnWileyandSonsforpermissiontouseapartofGómezTorrente(2016)asthebasisforapartofchapter6.Thanksarealsoduefor theirhelptothreeanonymousrefereesforOxfordUniversityPress,andto PeterMomtchiloff,myeditoratthePress.Iwishtoexpressaspecialgratitude totheteachersfromwhomIlearnedmuchofwhatIhopeIknowabouthowto
dophilosophy,especiallySaulKripkehimself,ManuelGarcía-Carpintero,and ScottSoames.IalsowanttothankmycolleaguesattheInstitutodeInvestigacionesFilosóficasoftheUniversidadNacionalAutónomadeMéxico (UNAM),theseinstitutionsthemselves,andmystudents,fromwhomIhave learnedsomuch.
Reference
ProblemsandPromises
WhatIsReference?
AsadvancedinthePreface,thisbookisabouthowtheconventionalreference ofsomelinguisticexpressionsorexpressionusesis fixedordetermined (whenitis).Butwhatisreference,inthesenseinwhichwewanttotalk aboutithere?
Asinsomanyothercaseswhereaquestionastowhatsomethingisis asked,inthiscasewecanlistexamplesthatwewant,atleastat firstsight,to countasexamplesofconventionallinguisticreference,andwecangive examplesthatwewant,atleastat firstsight,tocountasexampleswhere referencetosomeparticularthing,orreference toutcourt,doesnottake place.However,givingageneral,sufficientlyinformative,butexceptionless characterizationseemselusive,andmaywellbeimpossible infact,thiswill beoneoftheideasmadeplausiblebythediscussionsinthisbook.Still,itis possibletostatesomeplausiblegeneralideasaboutreferenceinthesense thatinterestsushere,ideasthatwewilltakeasguidingintheremainderof thebook.
Likesomanyothertheoristsinthetraditionofthetheoryofreference,we wanttoincludeascasesofreferencesomecaseswhereauseofalinguistic expressionintuitivelystandsforacertainthingasaneffectatleastinpartof linguisticconventionsconcerningtheexpression,andinthosesamecaseswe wanttosaythatthereisnoreferencetothingsdifferentfromthatthing;and also,wewanttosaythatthereisnoreferenceatallinsomecaseswhere intuitively,asaneffectatleastinpartoflinguisticconventions,itturnsout thatthereisnothingthatarelevantuseofanexpressionstandsfor.Wewant tosay,forexample,that “Aristotle,” asnormallyusedinanancientphilosophy class,refers,andthatitreferstoAristotle,thegreatancientGreekphilosopher. WewanttosaythatthemadmanwhobelievesofhimselfthatheisAristotle doesnotuse “Aristotle” sothatitreferstothemadmanhimself,muchasthe madmanbelievesthatheisAristotle andmuchas he (asopposedtothe
RoadstoReference:AnEssayonReferenceFixinginNaturalLanguage.
expression “Aristotle”)refers,inthesenseof “refers” inwhichapersonrefers, tohimselfwith “Aristotle.” WewanttosaythatifIutter Thatisabeautiful mountain,whileintentionallypointingandreferring(inthepersonalsenseof “referring”)toamountaininplainview,withnoothermountaininsight,my useof “that” referstothemountaininquestion.Wewanttosaythatthe agonizingmanwhoinhisdeliriumhallucinatesademon,andutters Heis comingtogetme,doesnotuse “he” sothatitreferstoanyone.Inallthesecases, thereference,whenthereissuch,istheobjectthat,inanintuitivesense,the relevantexpressionusestandsfor,asamatteratleastinpartofwhatseemto bethelinguisticconventionsapplyingtothatexpressionuse.Andwhenthere isintuitivelynoreference,orreferenceisnottosomeparticularthing,itwould seemthatthisisbecausesomehowthosesameconventionsimply(together withthesurroundingcircumstances)thesefactsabouttherelationofstanding forfailingtoholdbetweenusesoflinguisticexpressionsandobjects.
Butbeyondexamplesofthissort,involvingusesofnamesanddemonstrativeswherewhatwewanttosayabouttheirreferentsorlackthereofismoreor lessintuitivelyclear,theremaynotbeageneral,informative,butexceptionless characterizationthatwewanttoaccept,onedeterminingforanarbitrary expressionusewhetherithasalinguisticreferenceornot,andwhatthat referenceistobeifthereistobeone.Wewillseeinchapters2and3thatina fairnumberofactualandpotentialcasesofusesofdemonstrativesandnames, itisunclearwhetherthereissomethingthatarelevantexpressionuse conventionallystandsfor,andthusunclearwhetherreferenceoccursornot. Andwhenwemovebeyondnamesanddemonstrativesaltogether,andconsiderotherexpressions,suchasadjectives,commonnouns,Arabicnumerals, etc.,itseemsasifwecan’tevenrelyontheintuitiveideaofthethingthatan expressionusestandsforasaguidetothequestionofreference.Forexample, isthereanythingthatauseofaverb,auseof “snored,” say,intuitivelystands for?Well,evenifthereneednotbeageneral,informative,andexceptionless characterizationofreferencethatdefinitelysolvesquestionssuchasthis,there areprobablyafewthingsthatcanbesaidinthisarea,whichsuggestthatthe answermaylieinacertaindirection.
Frege,towhomtheintroductionofthemoderntheoreticalnotionof referenceatstakeherecanbeattributedinallessentialrespects,postulated thatavarietyofmeaningfulexpressionsofdiversegrammaticaltypesare susceptibleofhavingreferents.Andhepostulated,morespecifically,thatthe referenceofanexpressionuse(whenithasone)isthethingwhichconstitutes theexpressionuse’scontributiontothetruthconditionofthesentencein whichitappears,thethingspecificallycontributedbytheexpressionuseon
whichthetruthorfalsehoodofthesentencedepends.¹This,whichwewillcall theBasicFregeanIdea,isanappealingideaprovidingageneralcriterionfor reference,oneconsistentwiththedesiredconnectionsbetweenthetheoretical notionofreferenceandtheintuitivenotionofstandingforanobjectinthe caseofnamesanddemonstratives,andanideathathelpsinthesearchfor referentsforexpressionswhicharenotdemonstrativesornames.Ifonesays Aristotlesnored intheancientphilosophyclass,presumablythethingspecificallycontributedby “Aristotle,” onwhichthetruthorfalsehoodof Aristotle snored depends,isAristotle.AndFregesupposed,notimplausiblyat firstsight, thatthecontributionof “snored” tothetruthconditionof Aristotlesnored wasthefunction s assigningthetruth-valuetruthtoathingthatsnoredand thetruth-valuefalsehoodtoathingthatdidn’tsnore.Onecanreasonably thinkthatthethingcontributedby “snored” onwhichthetruthorfalsehood of Aristotlesnored dependsissomethingthatencodestheinformationwhether Aristotlesnoredornot,andthefunction s surelydoesthat ifAristotleis assignedthetruth-valuetruthby s,thatmeanshesnored;ifheisassignedthe truth-valuefalsehood,thatmeanshedidn’t.Butfurthermore,itisreasonable tothinkthatthethingcontributedby “snored” inthecaseof Platosnored oughttobethesamething,andoughttoencodetheinformationwhether Platosnoredornot(and mutatismutandis forsimilarpredicationsof “snored” ofotherthings).Andagainthefunction s surelydoesthat.Fregethus postulatedthatthereferenceof “snored” in Aristotlesnored isthefunction s, orequivalentlythesetofthingsthatsnored,alsoknownasthe extension of “snored.”
Buttheideaanditsassociatedcriteriondonotsolveallthequestionsabout thereferenceof “snored” thatonecouldthinkof.ItwouldseemthattheBasic FregeanIdeaimpliesthat “snored” doesnotrefertothesetofthingsthat
¹AccordingtoarelativelyuncontroversialwayofreadingFrege,whichIamadoptinghere,itishis viewthateachwell-formedcomplexexpressionarisesfromtheapplicationofasymbolreferringtoa functiontosymbolsreferringtoargumentsofthatfunction.Then,inthecaseofacompleteassertoric sentence,itsreference,whichisatruthvalue,isafunctionofthereferentsofitsparts,obtainedby successiveapplicationsoffunctionstothereferentsofsimplerconstituents.Thisisaprincipleof compositionalityforreferents.Inthissensethetruthorfalsehoodofasentencedependsonthe referentsofitsconstituentexpressions,whichthuscontributetheirreferentstothedeterminationof thesentence’struthvalue.Forexample,supposethatAristotleisthereferentof “Aristotle” andthatthe referentof “snored” isthefunction s assigningthetruthvaluetruthtoathingthatsnoredandthetruth valuefalsehoodtoathingthatdidn’tsnore;andsupposethatthesentence Aristotlesnored arisesfrom theapplicationof “snored” toitsargument “Aristotle.” Thenthetruthvaluetruthisobtainedasthe referenceofthesentence Aristotlesnored byapplicationof s toAristotle.Likemanyothers,Itakethe attributionofaprincipleofcompositionalityforreferentstoFregeaswell-grounded,e.g.givenhis acceptanceoftheprinciplethatsubstitutionofco-referentialexpressionsdoesn’talterasentence’ s referent(inFrege(1892));seee.g.Pelletier(2001).Butthereareofcoursedissentingvoices(seee.g. Janssen(2001)).
snoredwhenusedinanutteranceof Aristotlemightnothavesnored,say.For thesetofthingsthat(actually)snoredwouldnotseemtoencodeenough informationtodetermine,togetherwithwhateverreasonablereferentswe mightassigntotheotherexpressionsinthesentence,whetherAristotle mightnothavesnored,i.e.whetherhewouldnothavesnoredinotherpossible circumstances.Itwouldseemthatif “snored” istohaveareferencehere,it shouldbearicherthingthanthesetofthingsthatactuallysnored;perhaps itshouldbethesetofallpairsformedbyapossiblecircumstanceandthesetof thingsthatsnoredinthatcircumstance(athingalsoknownasthe intension of “snored”).AndthisthingdoesalsoencodetheinformationwhetherAristotle (oranyotheractuallyexistingthing,suchasPlato)actuallysnoredornot. Sowasthismorecomplexthingthereferenceof “snored” in Aristotlesnored, afterall?Andthenwemayalsoask,wastheintensionof “Aristotle” (thesetof allpairsformedbyapossiblecircumstanceandthethingthatisAristotlein thatcircumstance)thereferenceof “Aristotle” in Aristotlesnored,afterall?
Fregehimselfpointedoutthatitseemsasifthereferenceof “Hesperus” cannotbethesameasthereferenceof “Phosphorus” in TheGreeksinHomer’ s timesbelievedthatHesperuswasnotPhosphorus,despitethefactthatthey wouldappeartostandforthesamething,theplanetVenus,whichwouldseem tobethethingtheycontributeonwhichthetruthorfalsehoodofsimple sentencescontainingthemdepends.For TheGreeksinHomer’stimesbelieved thatHesperuswasnotHesperus seemsclearlyfalse,butif “Hesperus” and “Phosphorus” contributethesamethingtothedeterminationoftruthor falsehoodin TheGreeksinHomer’stimesbelievedthatHesperuswasnot Phosphorus,thenthissentenceshouldhavethesametruthvalueas TheGreeks inHomer’stimesbelievedthatHesperuswasnotHesperus,wherethetwouses of “Hesperus” certainlymustcontributethesamething.Andyet TheGreeksin Homer’stimesbelievedthatHesperuswasnotPhosphorus seemsintuitively true.So,whatisthereferenceof “Hesperus,” or “Phosphorus”?Frege,aswe know,postulatedthattheirreferenceinsimplesentencesistheplanetVenus, butalsothattheirreferentswhenembeddedinverbsofpsychologicalattitude areotherthings,appropriate senses containingmodesofpresentationofthe planetVenus,eachincludingenoughinformationastodeterminewhetherthe subjectoftheattitudeholdstheattitudetowardVenusunderthatmodeof presentation.Thesamegoesfor “Aristotle” whenembeddedunderverbsof psychologicalattitude,andasenseforAristotledoesencodetheinformation thatitisasense for Aristotle it determines Aristotle,infact,onFrege’ s theory.Sowhyshouldn’tasense,thislessfamiliarthingconjuredupby Frege,havebeenthereferenceof “Aristotle” in Aristotlesnored,afterall?
Sensesareverycontroversialthings.Manyphilosophers,includingperhaps mostofthephilosophersinthetraditionofthe “newtheoryofreference,” probablythemostinfluentialintherecentliteratureonreference,don’tgive sensesmuchofaroleinthestudyoflanguage,andthinkthatonecando semantics,andexplainthephenomenathatFregeusedsensestoexplain,using justordinaryindividualobjects,properties,intensions,andthelike some evendoubtsensesexistassuch.Thisbooksharesmanybasicviewsand assumptionswiththetraditionofthe “newtheoryofreference,” soperhaps wecanjumpontothewaveherealsoandjustforgetaboutsenses.Buteven then,indeterminacieswillremain.Forexample,whyshouldintensionsnotbe referents,giventhattheyseemtobecontributedbypredicativeexpressions and(atleastinmanycases)theyseemtoencodetheinformationneededto determinethetruthorfalsehoodofthesentencesinwhichthoseexpressions occur?(Justabouteverybodyseemstothinkintensionsarerespectablethings.)
ItisnaturaltoaddoneguidingideaaboutreferencetotheBasicFregean Idea,onethatappearstogetridoftheintensionsofpredicativeexpressionsas thereferentsofthosesameexpressions.Thereferenceofanexpression(when itexists)mustpresumablybe,atleastintypicalcases,somethingofwhichit canreasonablybesaidthatsomeordinaryspeakerhasatsomepoint(oftenin theearlyhistoryofhumankind)intendedtorefer(inthepersonalsenseof “refer”)toitwiththeexpression(orwithsomeancestororrelativeofthe expression).Linguisticreferentsare,wearesupposing,determinedinpartby linguisticconventions.Butthesearepresumablyconstituted,atleastinsome typicalcasesandinpart,bytheexistenceofwidespreadagreementsregarding howspeakers’ referentialintentionsgiverisetoreferentsforexpressionsor expressionuses(agreementswhichinmanycasesdateinsomeformorother fromtheearlyhistoryofhumankind).²Andifthisisso,itisunreasonableto supposethatthereferentofanexpressionisitsownintension,foritis unreasonabletosupposethat(typically)anordinaryspeakerhasintendedto refertosuitableintensionswhensheusedexpressionssuchas “Aristotle” or “snored” (letalonethatsomeoneintheearlyhistoryofhumankindintended torefertotheappropriateintensionswiththeappropriateancestorsofthese words).(Though,ofcourse,oneisfreetogiveanametoanintensionifone knowswhatanintensionisandwishestogiveanametoone.)Foronething,
²Forourmainpurposeinthisbook,whichistoexplainhowreferenceisachievedforcertainclasses ofexpressions,wewillessentiallyfocusonthisaspectofthenatureofreferentialconventions.Wewill notneedtosaymuchaboutotheraspectspresumablyessentialtoconventionsingeneral,suchasthe factthattheyaresustainedviacoordinationmechanismsofcertainkindsandinvirtueoftheir successfulroleintheattainmentofcommonaims.
intensionsarecomplicated,highlytheoreticalthings,andordinarypeoplejust don’tseemtobeabletohaveintentionsaboutthemassuch,certainlynot explicitly,butpresumablynotimplicitlyeither.
Ifthisplausibleideaisaccepted,namesandverbsdonotrefertotheirown intensions(atleastinrun-of-the-millcases).Butintensionsmustsomehowbe encodedbytherealreferents,ifwewanttorespecttheBasicFregeanIdea connectingreferenceandtruth,andtorespectaswelltheintuitionthatthe truthorfalsehoodof Aristotlemightnothavesnored dependson(something thatencodes)theintensionsof “snored” and “Aristotle.” Fortunately,atleast fortypicalnamesandverbs,andfortheotherexpressionswewilldealwithin thisbook(demonstrativesandArabicnumeralsinthecaseofapparent singularterms;andordinarynaturalkindnounsandadjectivesforsensible qualities,whenweturntoapparentpredicativeexpressions),itisreasonableto supposethatourplausibleideaandtheBasicFregeanIdeaconnectingreferenceandtruth(aswellastheinitialdesideratumthatthereferenceofa demonstrativeoranameshouldbethethingitintuitivelystandsfor)are jointlysatisfiedbyappropriatethings.
Ontheonehand,itisreasonabletosupposethatthereisawidespread,if implicit,intentiontorefertoAristotlewith “Aristotle” andtosomethingsuch asthe property ofsnoring(orhavingsnored)with “snored,” anintentionthat mayhavedevelopedinpartviatheformationandtransmissionofearlier intentionstorefertothosethings.Itisjustverynaturaltosaythatordinary peoplehaveatsomepointintendedtorefertoAristotleandsnoringwithsome words,inawayinwhichitisnotnaturaltosaythatordinarypeoplecanhave intendedtorefertointensions.Thatthisisnaturalisdue,atleastinpart,tothe factthatitisclearhowAristotleandsnoringcanhavebeentheobjectsof ordinarypeople’sintentions;inanintuitivelyclearsense,bothAristotleand snoringhavebeentheobjectsofrepresentationsinthemindsofordinary peopleinawayinwhichintensionshavenot.Ordinarypeopledidatsome pointhaveaperceptualrepresentationofAristotlethatformedthebasisofa linguisticreferentialintention,latertransmittedtopeoplewhodidnotgetto seeAristotle;andordinarypeopledidanddohaveaperceptualrepresentation ofsnoringthatcanformthebasisoflinguisticreferentialintentionsconcerning “snored.” Butordinarypeopletypicallyhavenomentalrepresentationof intensions,whetherperceptual,linguistic,orofotherkinds.
Andontheotherhand,itisreasonabletosupposethattheintensionsof “Aristotle” and “snored” areencodedbyAristotleand(thepropertyof) snoring,respectively.First,theintensionof “Aristotle” canbesaidtobe encodedbyAristotle,ifweacceptthelargelyuncontroversialviewthat
“Aristotle” isrigid,whichimpliesthatitsintensionisthesetofallpairsformed byapossiblecircumstanceandAristotle himself.(Andthesamecanbesaidof propernamesingeneral,demonstratives,andArabicnumerals.)Andsecond, theintensionof “snored” canbesaidtobeencodedbythepropertyofsnoring, astheintensionof “snored” isthesetofallpairsformedbyapossible circumstanceandthesetofthingsthatinstantiatethepropertyofhaving snoredinthatcircumstance.(Andthesamecanbesaidofverbsingeneral, andofotherpredicativeexpressions.)Accordingly,wewillsupposethatthings ingeneral(Aristotlebeingaparadigmaticexample)arethereferentsofusesof singularterms(whentheyhavereferents),andthatpropertiesarethereferents ofusesofpredicativeexpressions³(whentheyhavereferents).
Thisisnottosaythatpredicativeexpressionsrefertopropertiesinjustthe samewayinwhichsingulartermsrefertothingsingeneral.Presumably singularterms,inanintuitivesense, standfor thethingstheyreferto,but thepropertiesreferredtobypredicativeexpressionsarenotthingsthatinthe sameintuitivesensetheseexpressionsstandfor.Themodesofreferenceof singulartermsandofpredicativeexpressionsarepresumablydifferent.There areperhapsdifficultiesspellingoutthedifferenceinanunequivocalway,as Fregeconcludedinhisreflectionsonthedistinctionbetweenobjectsand concepts.ButIthinkwecaninanycasesay,withoutfearoferringtoo much,thatwhenreferredtobyapredicativeexpression,apropertyisinvoked initscapabilityorpotentialityofapplyingtothings,whilewhenreferredtoby asingularterm,athingingeneral(includingaproperty)isinvokedsimplyasa thingaboutwhichsomethingcanbesaid,itscapabilityofapplyingtothings(if any)notbeinginvoked.
WhyStudytheFixingofReference?
Wehavenowageneralideaofwhatreferencemustbe,inthesensethatwill concernushere evenifwhatwehavesaiddoesnotamount,byanystretch, toafulltheoryofwhatreferenceis.Sowenowhaveanideaofthesortofthing whose fixingordeterminationforlinguisticexpressionsorexpressionuseswe wanttostudyinthisbook.But,itmightlegitimatelybeasked,whyshoulda philosopher,orevenaphilosophyoflanguagespecialist,careabouthow referencegetsdetermined?AsIseethings,therearetwomainreasons.
³Intakingpropertiestobethereferentsofpredicativeexpressions(whentheyhavereferents),we arealsofollowingtheleadofseveralauthorsfromthe “newtheoryofreference” (andfromother traditions).Seee.g.Salmon(1981),(1986);Soames(2002).
The first,perhapslessimportantreason,isthataphilosopher,andin particularaphilosopheroflanguage,shouldcareabouthowcommunication ispossible,andreferenceseemstohaveanimportant,ifperhapssomewhat negative,roleintheexplanationofcommunication.Fregeemphasizedthat communication,atleastinthecaseofsentencesinvolvingnames butsuitablyanalogouspointsholdforeveryotherexpression would not workif thereferentofonesamename,asusedbydifferentpeople,werenottypically thesame.⁴ Why?Becauseifpeople’sknowledgeofwhateveritisthat fixesthe referenceofdifferentusesofonesamenamedirectedthemtodifferentthings, orjusttonothingatall,thepossibilityofcommunicationwouldbehampered, aspeoplewouldtheninterpretotherpeopleastalkingofdifferentthings,orof nothingatall,withtheresultingfailuresofcoordination.Inordertoexplain whythissituationdoesnottypically(seemto)arise,Fregethoughtitwas enoughtosupposethatthedifferentsensesdifferentpeopleattachedtothe samenameturnedouttocontainmodesofpresentationofthesamething,the thingcommonlyreferredtobyusesofthenamemadebydifferentpeople.
Alongwiththe “newtheoryofreference,” Idon’tthinkit’splausibleto supposethatitissomedescriptivesenseattachedbyaspeakertoanamethat fixesthename’sreference.ThedescriptivesensesthatFregehadinmindjust couldnotdothejob,aswell-knownargumentsbyKripke,Donnellan,and othershaveshown.AndinthisbookIwillarguethatotherdescriptivesenses thathavemorerecentlybeenthoughttodothejobinfactcannotdoiteither. Ithink,however,andthisisnotapointalwaysclearlyspelledoutorperhaps evenconcededby “newtheoristsofreference,” thatitisplausibletosuppose thattheexistenceofsharedreference-fixingconventionsplaysaroleinthe explanationofthepossibilityofcommunication.Thismightbethoughtto leadtodescriptivism,butitdoesn’t,because(asIwillargue)theexistence ofobservedreference-fixingconventionsdoesnotimplytheexistenceof reference-fixingdescriptivesenses.Ingeneral,itisreasonabletosuppose thattheabilitythatspeakershaveoffollowingthecommonlinguisticconventionsthat fixreferencemustplayaroleintheexplanationoflinguistic communication.Whenaspeakeruseslanguageinordertoattempttocommunicatethatacertainthing(suchasAristotle)hasordoesnothavea
⁴ Intalkinghereofthereferentofusesof “ onesamename, ” I,likeFrege,amnotpresupposingthat thenotionoftwousesbeingusesofonesamenamecanbeunderstoodindependentlyoftheideathat theyshouldhavethesamereferent.Rather,I’musingtheideaofusesofonesamenameinanintuitive sense,deferringthequestionofwhentwousesareofonesamenametoulteriortheorization, theorizationwhichshouldvindicatetheFregeanideainthemaintext.Theconsiderationsin chapter3willactuallyprovidethebasisforasuitabletheoreticalelucidationoftheideaoftwouses beingusesofonesamename.
certainproperty(e.g.thepropertyofhavingsnored),whichthingsandwhich propertiestheseareissomethingthatherhearerswillplausiblyoftenguessat leastinpartinvirtueoftheirhavingdevelopedanabilitytofollowthe commonconventionsgoverningwhatthingsarebeingtalkedaboutby meansoftheexpressionsusedbythespeaker.Theseconventions,asthe “newtheoryofreference” plausiblyargued,andaswewillemphasizemany timesinthisbook,neednotbeintegratedintothesemanticstructureofthe expressionsthatdothereferring.Nevertheless,anabilitytoabidebythe conventionsgoverningparticularexpressionsmustbedevelopedsomehow byspeakers,onpainoffailuresofcommunication thisbeingthethesisnot alwaysclearlyspelledoutby “newtheoristsofreference.” Inthisbookwe willproposeanumberofreference-fixingconventionsasunderlyingtheuse ofvariouskindsofexpressionsinlinguisticexchanges,hopingthatitwill appearplausibletothinkthattheabilitytofollowtheseconventionsis manifestedinrelativelysimplelinguisticbehaviors,anddoesn’ trequirethe attributionofunlikelyknowledgetoanormalspeaker.Bycontrastwith descriptivism,whichthe “newtheoryofreference” andthenewconsiderationsinthisbookshowtoinvokeimplausiblydemandingreference- fixing descriptiveassociationsonthepartofspeakers,itwillappearplausibleto thinkthatcompetencewiththerelativelysimpleconventionspostulatedby thepictureofthisbookdoesnotexceedinanimplausiblewaythecapacities thatwecanreasonablyattributetoanormalspeaker.
Thesecond,perhapsmoreimportant,butatanyraterelatedreasonwhya philosopher,regardlessofspecialty,shouldcareabouthowreferencegets determined,arisesagainfromaFregeanview,theBasicFregeanIdeaabout theconnectionbetweenreferenceandtruth.Onanatural,muchreviled,but nevervanquishedconceptionofhumanendeavorsandaccomplishments, truthissomethingwepursueandveryoftenachieve,ofwhichweoftenfail togetasmuchaswewant,butthatweconstantlygetmoreof.Nowtruth dependsonreference:whattruthswecometogetdependsonwhatitisthatis referredtowhenweexpressthosetruths;andifwhatwethinkaretruthswe havegotturnouttocontainexpressionswithoutreference,theywillhave turnedoutnottobetruthsatall.Ifthenaturalconceptionofhuman endeavorsandaccomplishmentsconcerningtruthistoholditsground,it mustbeaccompaniedbyacongenialvieworseriesofviewsabouthow referenceis fixed,onethatmakesitreasonabletobelievethatreference couldinfactbe fixedthatway,consistentlywithhumancapacitiesand dispositions;andonethatmakesitalsoreasonabletobelievethatourlanguage isnotplaguedbywidespreadfailuresofreference.
Therearemanypuzzlesindiscussionsofreference fixingthatconstitute obstaclestoaviewcongenialwiththenaturalconceptionofhumanendeavors andaccomplishmentsconcerningtruth.A firstkindofpuzzlesmakeithardto understandhowreferencecouldtakeplaceconsistentlywithhumancapacities anddispositions.Someofthesepuzzlesexploitthelimitationsofourepistemic capacitieswhenitcomestosinglingoutthingsbymeansoflanguageand thought;othersexploitourdispositionstocontradictourselves,whichwould oftenseemtogetinthewayofourabilitytosingleoutthingsbymeansof languageandthought.Asecondkindofpuzzlesdon’tnecessarilyexploitour epistemiclimitations,butmakeithardtounderstandhow,eveninsomecases whereourepistemiccapacitiesarenotworkingparticularlybadly,andwhere wearenotbeingutterlyinconsistent,theworldcouldbesuchthatour reference-fixingconventionsmanagetosingleoutappropriateitemsinit.
Whenweconsiderpuzzlesofthesetwokinds,wearenolongerexclusively concernedwithguessingwhatconventionsmaygoverntheassignmentof referencetoexpressions,orwiththerolethatreferentialintentionsplayin suchconjecturedconventions.Wemustengageinmoredefinitelyepistemologicalandmetaphysicalphilosophicaltasks.Atemptingthoughtisthenthat suchtasksgobeyondthelegitimateareaofconcernofthespecialistin reference fixingorinmetasemantics.However,thetemptingthoughtmust beresisted,asthereisatleastastrongreasonforthespecialistinreference fixingtoconcernherselfwithanappropriateresolutionofthoseepistemologicalandmetaphysicalpuzzles.Ifthepuzzlesarenotresolved,wewillbeleft withunchallengedreasonstobelievethatourlanguage,orlargeandimportant partsofit,isplaguedbywidespreadfailuresofreference.Andthenthiswill immediatelyconstituteachallengetowhateverconjecturedconventionsabout reference fixingthespecialistmayhavecomeupwith,fortheseconventions willnothavebeenshowntobecompatiblewiththenaturalconceptionof humanendeavorsandaccomplishmentsconcerningtruth.Accordingly,from theperspectiveadoptedinthisbook,whichItaketobetheperspective adoptedbyillustriousclassicsofthetraditionofthoughtaboutreference fixing,fromFregetoKripke,thetheoryofreference fixing,ormetasemantics, mustconcernitselfwitharelativelybroadrangeofepistemologicaland metaphysicalissues.Foronlyinthiswaycanthespecialistinreference fixing bereasonablysatisfiedthatherpreferredtheoryofthemechanismsofreference fixingdoesnotinappropriatelydisconnectreferencefromtruth.
Inthisbookwewillconsideragoodnumberofpuzzlesofthementioned kinds,asproblemsforatheoryofreference fixing,andwewilloftenprovide attemptsatsolutionsthat,ifcorrect,willtogetherconstituteaseriesofviews
aboutreference fixingthatwillbecongenialwiththenaturalconceptionof humanendeavorsandaccomplishmentsconcerningtruth.Someofthese viewsarebrieflypreviewedinthesummaryofthebook’scontentsgivenin thenextsection.
RoadstoReference:APreview
Demonstrativessuchas “this,”“that,”“he,”“she,”“it,” and “they” areparadigmatic,andperhapsinmanywaysthemostbasic,instrumentsoflinguistic reference.Theyalsoturnouttogiverisetophenomenacloselyrelatedtothose involvingpropernames,whichareprobablythemoststudiedinstrumentsof linguisticreference.Butinthecaseofdemonstratives,unlikeinthecaseof propernames,andundertheinfluenceespeciallyofKaplan’sgroundbreaking work,acertainkindofdescriptivismatthereference-fixinglevelhasnot seemeddifficulttoswallowtooriginatorsandfansoftherecentadvancesof the “newtheoryofreference.” Justabouteverybodywritinginthisareaseems toimply,insomemoreorlessexplicitway,thatthereshouldbeadescription, providedbythereference-fixingrulefor,say, “that” (bytheKaplanian character of “that”)which fixesthereferenceofauseof “that” inagivencontext, andwhichisknowninsomemoreorlessimplicitorinchoatewaybya competentspeaker.Butthedescriptioninquestionhasturnedouttobe exceedinglydifficultto find,asattestedbythereviewoftheliteratureon reference fixingfordemonstrativesofferedinchapter2.Thisliterature,as wewillsee,hascorrectlyidentifiedseveraldifficultproblemsfordemonstrativereference fixing,includingaparticularlyvexingonecausedbythefrequent existenceof conflicting referentialintentionsinthesamespeakerandoccasion ofuseofademonstrative.Thechapterwillpropose,inlinewithsimilarideas aboutpropernamestobedevelopedlaterinchapter3,thatthereasonfor thesedifficultiesisthatthereference-fixingconventionsfordemonstrative referencedonotamounttonecessaryandsufficientconditionsforreference, butonlytoalistofroughlysufficientconditionsforreferenceandreference failuretotakeplaceinselectedsituations afactthatismanifestedin theexistenceofafairnumberofcaseswhereitisuncertainandpresumably conventionallyindeterminatewhetherauseofademonstrativereferstoa particularthingornot.Ifthisisso,thechapterargues,thereislittlehopethata normalcompetentspeakercouldingeneralandinvariablyassociate(in howeveranimplicitorinchoateway)withauseofademonstrativean appropriatereference-fixingdescriptionthatheisawareof;forsucha
descriptionwouldessentiallyamounttoageneralnecessaryandsufficient conditionforathingtobethereferentofthatuse,aconditionthatwould yieldinappropriatelydeterminateandnotreallyknownverdictsofreference failureinmanycasesforwhichtherealmerelysufficientconditionsdonot yieldadeterminateverdict.Thus,thedemonstrativedescriptivismatthe reference-fixinglevelembracedbyoriginatorsandfansoftherecentadvances ofthe “newtheoryofreference” isjustaswrongasthecorresponding descriptivismaboutpropernamespresumablyis.Thechapterendsbydevelopingapictureofthereference-fixingconventionsthatstatetheconjectured roughlysufficientconditionsfordemonstrativereferenceandreferencefailure,andarguesthatthepicturesquareswellwithalltheelementsappealedto intheprecedingdiscussion.
Propernameslike “Aristotle” arealsoparadigmaticinstrumentsofreference,anditcanbesaidthatthereismuchthatweknowaboutthem,including muchabouthowitmustbethattheirreferentsget fixed(whensomereferents get fixedforthem).However,onefundamentalworryhassubsisted,Ithink, afteralltheconsiderableadvancesinourunderstandingofnamesthathave beenmadeinthelast fiftyyearsorso.Someofthesefundamentaladvances, dueespeciallytoKripke,havepointedtotheconclusionthatthereferentsof names,atleastintypicalcases,arenot fixedbytheordinarydescriptive informationthatspeakersassociatewiththem.Theinformationthataspeaker associateswithaname N thatseemstocomeoutofhismouthwithacertain intuitivereferent r,notesKripkeamongotherthings,willoftenbemostlyfalse of r;andifitismostlytrue,itwillstillbe,toallappearances,ofteninsufficientlydetailedtopickout r.Butthenhowisitthat r comestobethereferent of N inourspeaker’smouth?Anotheroftherelativelyrecentadvancesinour knowledgeofnames,againdueespeciallytoKripke,isourcurrentunderstandingthat r willoftenhavecometobethereferentof N inthemouthofa speaker s asaresultoftheexistenceofachainoftransmissionofthename, fromsomespeakerswhooriginallystartedusing N (oranancestorofit)with theintentiontoreferto r,downtoitsuseby s.Thisseemsright.Butsurely,if thishappensitisasaneffectoflinguisticconventionsthatspeakersshould havesomekindoffamiliaritywith,howeverimplicitandinchoate.Socan’tit beheld(onceaccountistakenalsoofotherconventionsfornamereference thatwouldappeartobemoreeasilyaccessible,suchasconventionsaboutthe possibilityofnamingstipulationsor “baptisms”)that s mustafterallknowin somewaysomesemi-occultdescriptiveinformation theinformation encodedbyhisfamiliaritywithreferentialnamingconventions fixing r as thereferentof N (asitcomesoutofhismouth)?Chapter3considersthis
question,madewithincisivenessespeciallybyproponentsofso-called “causal descriptivism,” andanswersitnegatively.Theproblemwiththiskindofideais againthatournamereference-fixingconventionsdonotseemtoamounttoa setofnecessaryandsufficientconditionsfornamereference;theconventions atworkseemtogiveroughlysufficientconditionsforreferenceorreference failuretotakeplaceinparticularcases,butnomore.Hereagaincasesof presumablereferentialindeterminacywillplayaroleinanumberofargumentsforthesuggestionthatonlyroughlysufficientconditionsareprovided bytherelevantlinguisticconventions.Theupshotisthatno “causaldescriptivist” reference-fixingdescriptionknownbynormalcompetent speakers,notevenoneknowninamerelyimplicitorinchoateway,canbe plausiblybuiltfromthoseroughlysufficientconditions.Thechapterconsiders alsothedamagingimplicationsoftheseandotherfactsforsomerecent theoriesofnamereference fixingthatarenotsoepistemicallydemandingas “causal” andotherdescriptivismsbutthatproposenecessaryandsufficient conditionsfornamereference.Finally,thechapterdevelopsapictureofthe reference-fixingconventionsstatingroughlysufficientconditionsforname referenceandtransmission,arguingthatitexplainsthephenomenaappealed tointhechapter.
Therearemanyotherinstrumentsofsingularreferenceinnaturallanguage asidefromnamesanddemonstratives,mentionsofwhichsurfaceinthe discussionsoftheseminalauthorsofthe “newtheoryofreference,” but whicharerarelytreatedindepthbythoseauthors.Amongtheseinstruments one findsArabicnumerals,towhichchapter4isdevoted.Inthiscase,asitwill turnout,acertainkindofdescriptivismwillnotseemsoimplausibleasa consequenceofwhatwillappearastheprobablerelevantconventionsinvolved inreference fixing.However,therightkindofreference-fixingconventions anddescriptionscanbebroughttolightonlyafteraconsiderableworkof clarificationofpuzzlesanddebunkingofmisleadingalternatives.Thereisa commonideathat(atleastthecomplex,decimal)Arabicnumeralsmusthave adescriptivesemanticstructure,inthiscaseonegivenapproximatelybythe polynomialsthatdetailthecontributionofeachdigitinanumeralasthe multipleofthecorrespondingpoweroften inthesumofmultiplesofpowers oftenthatthenumeralwouldappeartocodify.Eventhosewhoadopttheview thattheArabicnumeralsaresemanticallynon-descriptivetypicallythinkof theirreferentsasconventionally fixedbythosesamepolynomials.But,as explainedinchapter4,thereisjusttoomuchevidenceagainsttheseand otherideaswhichseektoexplainreference fixinginthisareabyappealto sophisticatedmathematicaldescriptions.Thechapterproposesanalternative
view,onwhichthereferenceof(atleastthecomplex)Arabicnumeralsis fixed (“merely fixed”)bymeansofsimpledescriptionswhichsingleoutthereferents oflaternumeralsintermsofthereferentsofearliernumerals;ontheproposal, anumeral(oratanyrateatleastacomplexone)willtypicallygetitsreferentas theresultoftheoperationofaddingoneasappliedtothenumberdenotedby theprecedingnumeralinthenaturalseriesofArabicnumerals(whichis generatedbyprinciplesindependentofsemanticinterpretation).Thepossibilityofreferencefornumerals,however,facesabasicproblem:ifnumerals refertoanything,theyoughttorefertonumbers,andnumbershaveabad pressamongphilosophers.Somemetaphysicalorepistemologicalprinciples areoftenthoughttoimplythatthingswiththepropertiesofthenumbersjust couldnotexist,whichleadsdirectlytotheviewthatarithmeticaltruths apparentlycontainingreferencetonumbersvianumerals,regardlessoftheir simplicity,justcannotbeliterallytrue.Whatcouldthenumbersbe,iftheyare tobe fixedasthereferentsoftheArabicnumerals(whenusedasnouns)bythe reference-fixingmechanismpostulatedinchapter4?Thechapterarguesthat someofthefeaturesofthisreference-fixingmechanismactuallypointtoward theviewthatthereferentsofthenumerals,hencethenumbers,arethe finite pluralcardinalityproperties.
Chapter5concernsordinarynounsfornaturalkinds,substances,and phenomena,whichwereshownbyKripkeandPutnamtosharemanycharacteristicswithpropernames,includingcharacteristicsrelevanttothe fixingof theirreferents.Inthisarea,amajorchallengeforthenaturalconceptionof humanendeavorsandaccomplishmentsconcerningtruthanditsFregean rootscomesfrommanyrecentargumentsinthephilosophyofbiologyand thephilosophyofchemistry.Theseargumentscastseriousdoubtsonthe abilityofthepresumablereference-fixingconventionsforordinarylanguage nounsfornaturalkinds,substances,andphenomenato fixordeterminereal kinds,substances,andphenomenathatthosetermscouldreferto.Asinthe caseofnames,inthecaseofnounsfornaturalkindsthereismuchwehad wrongorthatwejustdidn’tknowabout fiftyyearsago,butmanyfundamental advanceshavebeenmadeintheinterveningyears,dueespeciallytoPutnam andagaintoKripke.However,thementionedargumentsfromthephilosophiesofchemistryandbiologysuggestthatmanyaspectsoftheKripkePutnampicturecannotberight,andinparticularmakeithardtobelieve thatordinarynounsfornaturalkinds,substances,andphenomenacould refertothekinds,substances,andphenomenathatKripkeandPutnam thoughtweredeterminedbythereference-fixingconventionstheypostulated forthem.Atemptingposition,inviewofthearguments,isagainan
eliminativistposition,onwhichordinarylanguagediscourseaboutnatural kinds,substances,andphenomenaisnomorethanapicturesqueandpleasing fiction,buta fictionnonetheless,falseortruth-valueless thesortofviewthat dragsalongwithitanundesirableconflictwiththenaturalconceptionof humanendeavorsandaccomplishmentsconcerningtruth.Theissueshereare complicatedandrequireveryextensivediscussion,butchapter5sketches whatIthinkisasensiblenon-eliminativistviewaboutthereference-fixing potentialitiesofthepresumablereference-fixingconventionsforordinary termsfornaturalkinds,substances,andphenomena.Thesereference-fixing conventions,whichagainamounttolistsofroughlysufficientconditionsfor reference,arenottoodifferentfromtheconventionssomewhatimprecisely envisionedbyKripkeandPutnam.Butcarefulattentiontotheiraccurate formulation,andtoanumberofdetailssimplyoverlookedbytheseseminal authors,leadstothementionednon-eliminativistview,onwhichthereferents fortypicaltermsfornaturalkinds,substances,andphenomenaturnouttobe interestinglydifferentfromthereferentsassumed,somewhatuncritically,by KripkeandPutnam.
The finalchapter6dealswithamorelocalized,butneverthelessimportant, challengetothepossibilityofreferenceinlargeareasofdiscourse.The challengeariseswithspecialforceforadjectivesforcolorproperties,butif correctitmustquicklygeneralizetoalltermsforwhatmostphilosophershave traditionallyconsideredassecondaryqualities,suchassound,heat,etc.There ismuchevidenceindicatingthatdifferentapparentlynormalpeopleperceive thesamecoloredthingviaqualitativelydifferentexperiences.Thisleadsthem tomakewhatappeartobeincompatiblecolorjudgments,judgmentswhich, giventhatthedifferentpeopleinquestionseemperfectlynormal,would neverthelessappeartobeequallyfaultless.Thesefactshaveprovokeda tripartitesetofimplausiblereactions.Ona firstkindofresponse,thejudgmentsofsomeoftheapparentlynormalpeopleareright,andtheother judgmentsarewrong,whichiscompatiblewiththereferenceofacolorterm beingacertainobjectivepropertyorphenomenon;butthisbringsalongan epistemologicalmysteryabouthowpreciselythatpropertyorphenomenon hascometobereferredtoconventionallybyallpeople.Asecondkindof responseisprovidedbyerror-theoreticoreliminativistviewsonwhichcolor languageisjustnotasuitablevehicleofreferenceandtruth.Andathird responseisprovidedbylessradicalbutsomewhatadhocandlinguistically improbableviewsonwhichcolorjudgments(andjudgmentsaboutsounds, heat,etc.)involvereferencetosubjective,secondaryqualities.Chapter6developsanalternativepictureonwhichcolortermsoftenrefertoslightlydifferent
