ofasclaimsaboutsensedata,orasclaimsaboutmaterialobjects.Intuitively,it - PDF document

ofasclaimsaboutsensedata,orasclaimsaboutmaterialobjects.Intuitively,itishardtoseehowmaterialobjectstatementscouldcountasobservationsentencesifonebuysasensedatumtheoryofperception(asAyerandmanyoftheotherlogicalpositivistsdid);butifonetakessensedatumstatementsastheobservationstatements,thenonerunsthedangerofmakingthematerialobjectstatementsof,e.g.,sciencecomeoutmeaningless.ThisisnotaresultthatAyerandtheotherlogicalpositivistswerepreparedtoaccept;intheirview,scienticandcommonsenseclaimsmadeonthebasisofsensoryexperienceweretheparadigmcasesofmeaningfulutterances,andthequestionwaswhetherotherclaimsofphilosophyshouldalsobecategorizedasmeaningful.Inwhatfollowswe'llabstractawayfromthenatureofobservationsentences;wecantakethemlooselytoincludebothsensedatumstatementsandthematerialobjectstatementscorrespondingtothem,andfocusinsteadontherelationshipbetweenasentenceswhicharenotobservationstatementsineithersenseandobservationrequiredfortheformertobemeaningful.3`Inprinciple'veriabilityBeforegoingintothedetailsofattemptstodothis,itisworthdistinguishingbetweenclaimswhichhavebeenveried{i.e.,whichhavebeeninvestigatedviatheircorrelatedobservationsentences{andclaimswhichcan,inprinciple,beveried.AsAyersays,plausibly,\Plainlyweallunderstand,inmanycasesbelieve,propositionswhichwehavenotinfacttakenstepstoverify.Manyofthesearepropositionswhichwecouldverifyifwetookenoughtrouble....Asimpleandfamiliarexampleofsuchapropositionisthepropositionthattherearemountainsonthefarthersideofthemoon.Norockethasyetbeeninventedwhichwouldenablemetogoandlookatthefarthersideofthemoon,sothatIamunabletodecidethematterbyactualobservation.ButIdoknowwhatobservationswoulddecideitforme,if,asistheoreticallyconceivable,Iwereonceinapositiontomakethem.AndthereforeIsaythatthepropositionisveriableinprinciple,ifnotinpractice,andisaccordinglysignicant.Ontheotherhand,suchametaphysicalpseudo-propositionas\theAbsoluteentersinto,butisitselfincapableof,evolutionandprogress",isnoteveninprincipleveriable.ForonecannotconceiveofanobservationwhichwouldenableonetodeterminewhethertheAbsolutedid,ordidnot,enterintoevolutionandprogress."(36)Themoralofthisisthatifwearetryingtoestablishthemeaningfulnessofsomesentencebyrelatingittosomesetofobservationsentences,weneednotrequirethatwehaveactuallymadetheobservationscorrespondingtothoseobservationsentences;allthatisrequiredisthatwecould,inprinciple,makethoseobservations.3 thatnonitesetofobservationsentencescanentailthattheuniversalgeneralization\Allswansarenon-red"istrue.4.3ConclusiveveriabilityorfalsiabilityThisresultsuggestsanobviouswayofextendingourrsttwoattemptstodenemean-ingfulnessintermsofstrongverication:wecanclaimthatasentenceismeaningfulifandonlyifitiseitherconclusivelyveriableorconclusivelyfalsiable.Thisseemstodealwithsimpleuniversalgeneralizations,sincetheyareconclusivelyfalsiable,andwithsimpleexistentialgeneralizations,sincetheyareconclusivelyveriable.Ayerdoesnotexplicitlyconsiderthispossibility.Butitseemstorunintothefollowingthreeproblems(fromSoames,PhilosophicalAnalysisintheTwentiethCentury,pp.280-282).4.3.1MixedquanticationThereisstillaproblemindealingwithsentenceswhichcontainbothuniversalandexis-tentialquantication,likeForeveryquestion,thereisananswer.whichmaybeformalizedas8x(xisaquestion!9y(yisananswertox))Weknowthatthisclaimisnotconclusivelyveriable,forthesamereasonas`Allswansarewhite'isnotconclusivelyveriable:nonitesetofobservationstatementsentailsthetruthofaclaimaboutallthequestions.Soifitistobemeaningful,ithadbetterbeconclusivelyfalsiable.Butitisn't.Tocon-clusivelyfalsifyaclaimistobeabletoderiveitsnegationfromanitesetofobservationstatements.Inthiscase,we'dhavetobeabletoderive:8x(xisaquestion!9y(yisananswertox))whichisequivalentto9x:(xisaquestion!9y(yisananswertox))i.e.,9x(xisaquestion&:9y(yisananswertox))5 whichisequivalentto9x(xisaquestion&8y:(yisananswertox))Soforouroriginalsentencetobeconclusivelyfalsiable,wehavetobeabletoderivetheabovefromsomesetofobservationsentences.Butwecan't:toderivethisfromasetofobservationsentences,we'dhavetobeabletoderivefromsuchasettheuniversalclaimthatthereissomequestionsuchthateveryanswerfailstobeananswertothatquestion.Butwecannomorederivethisfromasetofobservationsentencesthanwecanderive\Allswansarewhite."fromsuchaclaim.(Theproblemherestemsfromthefactthatconclusivefalsiabilityofasentenceisjustconclusiveveriabilityofitsnegation,andifasentencecontainsbothexistentialanduniversalquanticationintheaboveway,thenbothitanditsnegationwillbeuniversalclaims.)Sotheclaimthatasentenceismeaningfuliitiseitherconclusivelyfalsiableorcon-clusivelyveriableentailsthatallsentencesofmixedquanticationliketheabovearemeaningless.Butagain,thepositivistswerenotwillingtoacceptthisresult.(Andtheywererightinthinkingthatthisresultwasabsurd.)4.3.2`Most'andotherquantiersAnotherproblemstemsfromcertainquantiersotherthantheuniversalandexistentialquantiers,like`most.'Consider,e.g.,`Mostravensareblack.'Howwouldyougoabouttryingtoeitherconclusivelyverifyorfalsifythissentenceonthebasisofobservationsentences?Oneplausibleidea:wecouldconclusivelyverifythissentence,itseems,ifwecouldaddtothelistofobservationsentencesageneralclaimlike`Thesearealltheravens.'Thisgeneralclaimisconclusivelyfalsiable,andhencemeaningful.Soonewantstoformulateamoregeneralcriterion,whichgivesastep-by-stepanalysisofmeaningfulness,alongthefollowinglines:Sismeaningfulieitheritoritsnegationareentailedbyasetofsentencescontainingonlyobservationsentences,sentenceswhichareconclusivelyveriableorfal-siable,andothersentencesalreadyqualiedasmeaningfulbythistest.ThisresemblesthetestthatAyereventuallysuggests;we'llreturntoitlater.Butthiskindofstepwisedenitionintermsofstrongvericationisalsochallengedbyathirdproblemfacingourproposedanalysis.4.3.3StatementsaboutunobservablesConsider,forexample,claimsaboutelectrons.Electronsarenotdirectlyobservable;rather,theyarepostulatedtoexplainphenomenawhichareobservable.Butjustbecause6 thepostulationofelectronsexplainsobservablephenomena,itdoesnotfollowthatclaimsaboutelectronsmaybederivedfromobservationsentences.Butthepositivistswere,forgoodreason,reluctanttoconsignclaimsaboutelectronstothesamecategoryasmetaphysicalclaimslike\TheAbsoluteislazy."5WeakvericationThisproblemsuggeststhatanentirelynewapproachisinorder,andthisisinfactwhatAyersuggests.Hesays:\Accordingly,wefallbackontheweakersenseofverication.Wesaythatthequestionthatmustbeaskedaboutanyputativestatementoffactisnot,Wouldanyobservationsmakeitstruthorfalsehoodlogicallycertain?butsimply,Wouldanyobservationsberelevanttothedeterminationofitstruthorfalsehood?Anditisonlyifanegativeanswerisgiventothissecondquestionthatweconcludethatthestatementunderconsiderationisnonsensical."(38)Thismovefromfocusingonwhatcanbederivedfromobservationclaimstofocusingonwhatobservationclaimsmightberelevanttomarksanimportantshiftbetweentwodierentwaysofthinkingaboutthecriterionofveriability.But,asAyernotes,weneedtobeabitmorespecicthanthisinformulatingaweakcriterionofveriability.Hegivestwoprecisewaysofformulatingthebasicintuitionhere,oneinChapter1,andthesecondinthePrefacetothe2ndeditionofLanguage,Truth,&Logic.5.1Ayer'srstdenitionInthersteditiontothebook(1935),Ayergavethefollowingaccountofveriability:\Letuscallapropositionwhichrecordsanactualorpossibleobservationanexperientialproposition.Thenwemaysaythatitisthemarkofagenuinefactualproposition,notthatitshouldbeequivalenttoanexperientialpropo-sition,oranynitenumberofexperientialpropositions,butsimplythatsomeexperientialpropositionscanbededucedfromitinconjunctionwithcertainotherpremiseswithoutbeingdeduciblefromthoseotherpremisesalone."(38-39)Wecanformulatethisasfollows:SismeaningfulifthereissomesetofsentencesP1...PnandsomeobservationsentenceOsuchthat(i)OfollowsfromStogetherwithP1...Pnbut(ii)OdoesnotfollowfromP1...Pnalone.7 Whilethisformulaisfairlyabstract,thereisasimpleandplausiblethoughtbehindit.Ayer'sideawasthat,giventhefailureofstrongvericationanalysesofmeaningfulness,veriabilitycannotamounttoequivalenceorentailmentrelationsbetweenasentenceandsomesetofobservationsentences;rather,whatwewantisaclearwayofstatingthethoughtthatasentence,evenifnotequivalenttoasetofobservationsentences,makessomedierencetowhatisobservable.ThisiswhatAyer'sdenitiontriestocapture.Itsays,plausibly,thatasentencehasempiricalconsequences,andhenceismeaningful,ifaddingittosomestockofpropositionschangeswhichobservationsentencesfollowfromthatstockofpropositions.Thisdenitionseemstodowellwiththecaseswhichprovedproblematicabove.Let'sexaminethem:Sentencesofmixedquantication,like`Everyliquidhasaboilingpoint.'Considerthesentence`xisaliquid.'Byitself,itdoesnotfollowfromthissentencethat`xhasaboilingpoint'istrue.Butitdoesfollowthissentence,alongwithourexamplesentence,than`xhasaboilingpoint'istrue.Sowegetthegoodresultthatsentenceslikethisonecanbemeaningful.Otherquantiedsentences,like`Mostapplesarered.'Considerthesentence`Theseveitemsarealltheapples.'Fromthis,itdoesnotfollowthat`Threeoftheseitemsarered.';butitdoesfollowifweaddthepremise`Mostapplesarered.'Hencethelatterclaimqualiesasmeaningful.Statementsaboutunobservables.Toseehowthismightwork,notethatwecanhaveconditionalsconnectingclaimsaboutunobservableswithobservationstatements.Forexample,consider`Ifasubstanceiscom-posedprimarilyofH2Omolecules,itwillboilat100C.'Fromthis,itdoesnotfollowthat,forsomespeciedx,`xwillboilat100C.'Butitdoesfollowifweaddastatementaboutunobservables:`xiscomposedprimarilyofH2Omolecules';hencewecancountthisstatement,correctly,asmeaningfulbyAyer'snewcriterion.Thereis,however,aproblemimplicitinthewaythatwejusthandledthecaseofstate-mentsaboutunobservableswhichisdevastatingtoAyer'sdenition.ThiswaspointedoutbyIsaiahBerlin.Theproblemis,intuitively,thatwhenwegotodecidewhetheragivensentenceSismeaningful,wetrytondsomepropositionorsetofpropositionsSsuchthat,whenweaddStoS,wecanderivesomeobservationsentencenotderivablefromSalone;buttherearenorestrictionsatallputonwhatScanbe.TheresultisthatwecanderivefromAyer'srstdenitiontheunwantedresultthateverysentenceisveriable,andhencemeaningful.ConsiderthesentenceTheabsoluteislazy.8 Isthereanysentencewhich,combinedwiththissentence,entailsanobservationsentenceitdoesnotentailbyitself?Ofcourse;theconditionalsentenceIftheabsoluteislazy,thenthisisred.issuchasentence.Ingeneral,foranysentenceS,wecanprovethatSismeaningfulbyderivingfromittogetherwiththeconditionalS!OtheobservationsentenceO.5.2Ayer'sseconddenitionAccordingly,inthesecondeditionofthebook(1946)Ayergaveanewaccountoftheprincipleofveriability.Hisaimwastostickwiththespiritofhis1935denition,whileplacingrestrictionsontheclassof`supplementarypropositions'whichcouldbeusedtoderiveanobservationstatementfromagivensentence.Tothisend,hedistinguished(p.13)betweendirectandindirectveriability,wherethesenotionsaredenedasfollows:Sisdirectlyveriabledf(i)Sisanobservationstatementor(ii)Sentailsinconjunctionwithasetofobservationstatementssomeobservationstatementnotentailedbythatsetalone.Sisindirectlyveriabledf(i)SentailsinconjunctionwithasetSofstatementssomeobservationstatementnotentailedbythatsetaloneand(ii)thereisnostatementinSwhichisnoteither(a)directlyveriable,(b)analytic,or(c)capableofbeingindependentlyshowntobeindirectlyveriable.Thecentralclaimofthetheoryisthen:Asentenceismeaningfulifandonlyifitiseitherdirectlyorindirectlyveri-able.5.3Hempel'sobjectionHempel'sobjectionisstatedinhis\TheEmpiricistCriterionofMeaning."Hisargumentassumesthefollowingtwoplausibleclaims:(i)Asentenceismeaningfulifandonlyifitistrueorfalse.(ii)ForanysentenceS,Sistruei:Sisfalse.Nowtakeanytruemeaningfulsentence(itcanbeanobservationsentence,ornot)like`Thisisred.'Considertheconjunction9 ThisisredandtheAbsoluteislazy.Thissentencemustbemeaningful,sinceitentailstheobservationsentence`Thisisred.'So,byassumption(i),itmustbeeithertrueorfalse.Supposerstthatthesentenceistrue.Thenitfollowsthat`TheAbsoluteislazy.'mustbetrueaswell,sincethetruthofaconjunctionentailsthetruthofitsconjuncts.Butthen,byprinciple(i),itmustbemeaningful.Supposethenthatthesentenceisfalse.Thenbyprinciple(ii)itsnegationmustbetrue;and,ifboth`Not(ThisisredandtheAbsoluteislazy)'and`Thisisred'aretrue,weknowthat`Not(TheAbsoluteislazy)'mustbetrue.Butthenweknowby(ii)againthatitsnegation,`TheAbsoluteislazy'mustbefalse.Butthenitfollowsfromprinciple(i)that`TheAbsoluteislazy.'ismeaningful.Soagainaformulationoftheprincipleofveriabilityleadstotheresultthatallsentencesaremeaningful.WhataretheprospectsfordenyingoneorbothofthepremisesonwhichHempel'sargumentrests?5.4Church'sobjectionInhisreviewofthesecondeditionofLanguage,Truth,&Logic,AlonzoChurchprovidedanargumentforthesameconclusionwhichmadedowithlesssubstantialassumptionsthan(i)and(ii).(ThefollowingexpositionoftheargumentfollowsthatinSoames,pp.289-90).Assumethefollowing:Thereareatleastthreeobservationsentencesp,q,r,noneofwhichentaileitheroftheothers.Nowwecantakeanysentences{evenanonsensesentencelike`TheAbsoluteislazy'{andformthefollowingcomplexsentence:(1)(:p&q)_(r&:s)Nownotethatsentence(1)countsasdirectlyveriable.Proof:pisanobservationsentence;theconjunctionofsentence(1)withpentailsr;andrisanobservationsentencewhich(byhypothesis)isnotentailedbypalone.qiseitherentailedbysentence(1),oritisnot.Eitherway,swillendupcountingasmeaningful.Supposerstthatqisnotentailedby(1).Then,sinceqisentailedbythecombinationof(1)ands,and(1)isdirectlyveriable,scountsasindirectlyveriable.10