科学哲学现状

科学哲学现状

\nThePresentSituationinthePhilosophyofScience\nProceedingsoftheESFResearchNetworkingProgrammeTHEPHILOSOPHYOFSCIENCEINAEUROPEANPERSPECTIVEVolume1SteeringCommitteeMariaCarlaGalavotti,UniversityofBologna,Italy(Chair)DiderikBatens,UniversityofGhent,BelgiumClaudeDebru,EcoleNormaleSupérieure,FranceJavierEcheverria,ConsejoSuperiordeInvestigacionesCientificas,SpainJanFaye,UniversityofCopenhagen,DenmarkOlavGjelsvik,UniversityofOsloGerdGrasshoff,UniversityofBern,SwitzerlandTheoKuipers,UniversityofGroningen,TheNetherlandsLadislavKvasz,ComeniusUniversity,SlovakRepublicAdrianMiroiu,NationalSchoolfoPoliticalStudiesandPublicAdministration,RomaniaIlkkaNiiniluoto,UniversityofHelsinki,FinlandDemetrisPortides,UniversityofCyprus,CyprusWlodekRabinowicz,LundUniversity,SwedenMiklósRédei,LondonSchoolofEconomics,UnitedKingdomFriedrichStadler,UniversityofViennaandInstitutWienerKreis,AustriaGregWheeler,NewUniversityofLisbon,FCT,PortugalGereonWolters,UniversityofKonstanz,Germanywww.pse-esf.org\nFriedrichStadlerGeneralEditorTogetherwith:DennisDieks,WenceslaoJ.González,StephanHartmann,ThomasUebelandMarcelWeberThePresentSituationinthePhilosophyofScience123\nEditorFriedrichStadlerInstituteViennaCircleUniversityofViennaDepartmentofPhilosophyandDepartmentofContemporaryHistoryCampusderUniversitätWien,Hof1A-1090ViennaAustriaFriedrich.Stadler@univie.ac.atISBN978-90-481-9114-7e-ISBN978-90-481-9115-4DOI10.1007/978-90-481-9115-4SpringerDordrechtHeidelbergLondonNewYorkLibraryofCongressControlNumber:2010926739cSpringerScience+BusinessMediaB.V.2010Nopartofthisworkmaybereproduced,storedinaretrievalsystem,ortransmittedinanyformorbyanymeans,electronic,mechanical,photocopying,microfilming,recordingorotherwise,withoutwrittenpermissionfromthePublisher,withtheexceptionofanymaterialsuppliedspecificallyforthepurposeofbeingenteredandexecutedonacomputersystem,forexclusiveusebythepurchaserofthework.Coverdesign:eStudioCalamarS.L.Printedonacid-freepaperSpringerispartofSpringerScience+BusinessMedia(www.springer.com)\nTABLEOFCONTENTSFRIEDRICHSTADLER,Editorial:OnthePresentSituationinthePhilosophyofScience.....................................................................................7PartI(TeamE)THOMASUEBEL,SomeRemarksonCurrentHistoryofAnalyticalPhilosophyofScience.................................................................13THOMASMORMANN,HistoryofPhilosophyofScienceasPhilosophyofSciencebyOtherMeans?CommentonThomasUebel.............................................29CRISTINACHIMISSO,AspectsofCurrentHistoryofPhilosophyofScienceintheFrenchTradition............................................................................................41ANASTASIOSBRENNER,ReflectionsonChimisso:FrenchPhilosophyofScienceandtheHistoricalMethod.............................................................................57MICHAELHEIDELBERGER,AspectsofCurrentHistoryof19thCenturyPhilosophyofScience..............................................................67MASSIMOFERRARI,Well,andPragmatism?CommentonMichaelHeidelberger’sPaper.......................................................................75PartII(TeamA)VINCENZOCRUPIANDSTEPHANHARTMANN,FormalandEmpiricalMethodsinPhilosophyofScience...................................................................................87VINCENTF.HENDRICKS,TheBaneofTwoTruths..................................................99THOMASMÜLLER,FormalMethodsinthePhilosophyofNaturalScience.........111FRANZDIETRICHANDCHRISTIANLIST,TheProblemofConstrainedJudgmentAggregation.................................................................................................125GABRIELLAPIGOZZI,AggregationProblemsandModels:WhatComesfirst?.....141PartIII(TeamB)MARCELWEBER,LifeinaPhysicalWorld:ThePlaceoftheLifeSciences.......155CLAUDEDEBRU,CommentsonMarcelWeber’s“LifeinaPhysicalWorld:ThePlaceoftheLifeSciences”..................................................................169THOMASA.C.REYDON,HowSpecialaretheLifeSciences?AViewfromtheNaturalKindsDebate.................................................................................173\n6TableofContentsMILESMACLEOD,TheEpistemology-onlyApproachtoNaturalKinds:AReplytoThomasReydon........................................................................189MEHMETELGIN,ReductionisminBiology:AnExampleofBiochemistry.........195RAFFAELLACAMPANER,ReductionistandAntireductionistStancesintheHealthSciences.................................................................................205PartIV(TeamC)WENCESLAOJ.GONZÁLEZ,TrendsandProblemsinPhilosophyofSocialandCulturalSciences:AEuropeanPerspective................................................221ARTOSIITONEN,StateoftheArt.ACommentaryonWenceslaoJ.Gonzalez’Contribution,“TrendsandProblemsinPhilosophyofSocialandCulturalSciences:AEuropeanPerspective”...............................................243MATTISINTONEN,ScientificRealism,theNewMechanicalPhilosophers,andtheFriendsofModelling......................................................................257DANIELANDLER,IsNaturalismtheUnsurpassablePhilosophyfortheSciencesofManinthe21stCentury?..........................................................283ANTÓNIOZILHÃO,WhatDoesitMeantoBeaNaturalistintheHumanandSocialSciences?ACommentonDanielAndler’s“IsNaturalismtheUnsurpassablePhilosophyfortheSciencesofManintheTwenty-firstCentury?”............305PartV(TeamD)DENNISDIEKS,ReichenbachandtheConventionalityofDistantSimultaneityinPerspective............................................................315MAURODORATO,OnVariousSensesof“Conventional”andtheirInterrelationinthePhilosophyofPhysics:SimultaneityasaCaseStudy......................335ROMANFRIGGANDCARLHOEFER,DeterminismandChancefromaHumeanPerspective....................................................................................351LÁSZLÓE.SZABÓ,WhatremainsofProbability?................................................373HOLGERLYRE,HumeanPerspectivesonStructuralRealism..............................381F.A.MULLER,TheCharacterisationofStructure:DefinitionversusAxiomatisation............................................................................................399IndexofNames..................................................................................................417\nFRIEDRICHSTADLEREDITORIAL:ONTHEPRESENTSITUATIONINTHEPHILOSOPHYOFSCIENCETheUniversityofViennasawwitnesstotheOpeningConferenceoftheESF-ResearchNetworkingProgramme“ThePhilosophyofScienceinaEuropeanper-spective”(PSE)whichwasorganisedbytheViennaCircleInstitute1andtookplaceonthe18–20DecemberattheCampusoftheUniversityofVienna,2008.2Itsoverarchingaimwastosetthebackgroundforacollaborativeprojectorganis-ing,systematising,andultimatelyforginganidentityfor,EuropeanphilosophyofsciencebycreatingresearchstructuresanddevelopingresearchnetworksacrossEuropetopromoteitsdevelopment.Assuchunderthegeneralrubricof‘thepresentsituationinthephilosophyofscience’,theemphasiswasonasafirststepidentifyingtraditionsandresearchstructuresalreadypresent,andthedirectionsinwhichthisresearchwasleading.Thisvolumepresentsthepapersoftheopeningconferenceaccordingtofivepre-establishedgroupings,eachrepresentedbyspeakersfromateam:TeamE:FoundationalandMethodologicalDebates(teamleaderThomasUebel)TeamA:FormalMethodsandtheirApplicationstothePhilosophyofScience(teamleaderStephanHartmann)TeamB:ApproachestotheFoundationsofScience:thePlaceoftheLifeSciences(teamleaderMarcelWeber)TeamC:ThePresentSituationofthePhilosophyoftheCulturalandSocialSci-ences:The“NaturalistTurn”,the“SocialTurn”,andtheDiscussiononScientificRealism(teamleaderWenceslaoJ.González)TeamD:PhilosophicalFoundationsofthePhysicalSciences(teamleaderDennisDieks).Tostartwithabroadoverviewofthetalksthemselves,andtheaccompanyingdis-cussion,wecancharacterisethegeneralthemesthatwerepursuedandissuesthatemergedwithintheframeof‘Europeanphilosophyofscience’.Oneaspectthatcamereadilytotheforewasthestronghistoricalaspectsofthe‘Europeanperspec-tive.’Astheopeningconferenceitwasofcourseimportantforthephilosophers1Cf.theactivitiesoftheInstitutWienerKreis/InstituteViennaCircle:www.univie.ac.at/ivc.2Seethedetailedconferencereport:FriedrichStadler,DonataRomizi,MilesMacLeod,“ThePresentSituationinthePhilosophyofScience:OpeningConferenceoftheESF-Research-NetworkingProgramme‘ThePhilosophyofScienceinaEuropeanPerspec-tive’”,in:JournalforGeneralPhilosophyofScience(forthcoming).PublishedonlineAugust13,2009:DOI10.1007/s10838-009-9088-y.\n8FriedrichStadlerinvolvedtodefinewhatcouldbemeantbyaEuropeanidentity,andnaturallytheyfocusedonthedeephistoricalrootsandtheircontinuingtraditionsinmodernphi-losophyofscience.AssuchtherewasastrongattentiontothehistoricalrelationsandoriginsinEuropeanhistoryofmodernissues,andhowinfactthiscontextpresentedanongoinginfluenceonthemodernpractice.TheconferencealsorevealedthatthishistoricaldimensioniscomplementedbytheevidentbroadscopeofEuropeanphilosophyofsciencewhichembodiesnotonlyastrongtraditionofhistoryandphilosophyofscience,historyofphilosophyofscience,butalsophilosophywithrespecttotheculturalandsocialsciencesaspartof(notseparateto)thediscipline,combinedwithmoretraditionalphilosophi-calissuesandapproaches,suchastheapplicationofformalmethods,theproblemofrealism,determinismandchanceorthenaturalkindsdebate.Thisconsidera-tionofgeneralphilosophicalquestionsinscienceismarriedtoastrongtraditionofengagingnaturalisticallywiththeparticularphilosophicalissuesinindividualscienceswherethereexistsaprerogativeofbeingcloselyschooledintherelevantscientifictheoryandresearchcontext.Additionally,onecanrefertoparticularpo-sitions,like‘structuralrealism’,as‘European’,havingtheiroriginandtheircentreofpursuit,andindeedtheirhistoricallinks,inthecontextofEuropeanresearch.CombiningtheseelementsisthequintessentiallyEuropeanself-reflectionontheaimsandvaluesofphilosophyofscienceinitselfandtherightmethodologywithwhichtodoit.Thisisanever-presenttheme,whichtracesitsrootsstronglyalsointhehistoryofEuropeanphilosophy.Thiswasraiseddirectlywithrespecttodiscussionson‘naturalism’butaroseinthecontextofdiscussionoverformalmethods,naturalkinds,andtherelationsbetweensocialandculturalsciencesandthenaturalsciences.Theteamleadersorganisedthe5sections,elaboratinggenerallyonitstheme.Eachofthepapersarecomplementedbypreparedcommentariesfromoneinvitedcommentator,occasionallytakingintoaccountthegeneraldiscussion.Thepre-historyandresearchbackgroundoftheprogrammerunningfor5yearsupto2013isthebeginningofapromisinginterdisciplinarynetworkingandcooperationinthephilosophyofsciencealloverEuropewith17participat-ingcountriesandstructuredin5teamswithmorethan60scholarscomingfrom22countries–renownedscientistsaswellasyoungergiftedphilosophersofsci-ence.3ThiscollaborativeenterpriseisbasedontwopreviousESF-researchnet-workson“HistoricalandContemporaryPerspectivesofPhilosophyofScience”andon“ThePhilosophyofPhysics”.Togetherwiththe“EuropeanPhilosophyofScienceAssociation”(EPSA),4whichwasfoundedin2006thisprogrammeseemsapromisingforumtoimprovethecooperationandinteractionintheflourishingphilosophyofscience.Thereisalongandpowerfultraditioninthisresearchfieldandlateronwiththeemergenceofphilosophyofscienceasadisciplinesince3www.pse-esf.org4www.epsa.ac.at\nEditorial9thebeginningofthe20thcenturyinthecapitalsofBerlin,Prague,WarsawandinVienna(withtheBerlinGrouparoundReichenbachtotheViennaCirclearoundSchlick),butalsoinFrance,ItalyandGreatBritain.TheforcedmigrationofthemovementsofLogicalEmpiricismbeforeWorldWarIIledtoaradicaltransferandcognitivetransformation,whichcanbecharacterizedasaturn“fromWissen-schaftslogik(LogicofScience)toPhilosophyofScience”.5Theinstitutionalandacademicdisciplinewasaresultofthistransatlanticinteractionandtransition.InNorthAmericathismovehasledtoanearlyinstitutionalizationandprofessionali-zationofphilosophyofscience,asbecamemanifestin1934withthefoundingoftheJournalPhilosophyofScienceandlateronthesocietyPhilosophyofScienceAssociation(PSA).Onlydecadeslater,inthe1960swecanrecognizeareturnofthesecurrentsbacktoEuropeandapleasingmutualcommunicationbetweenEuropeandtheUSA.InthemeantimethereexistsalivelycooperationwiththeNorthAmericancommunity,asisdocumentedpartlywiththePSAmeetings,theInternationalSocietyforHistoryandPhilosophyofScience(HOPOS),evenifthiswasnotalwaysasymmetricinteraction.ThereforetherecentdevelopmentsandinceptionsinEuropeareseenasawelcomedscholarlycounterpartandacollaborativeresearchactivitywhichisre-vivingtheEuropeantraditionaswellasfosteringtodays’increasingeffortsandpotentialsinEuropean’sphilosophyofscience,butwithoutaimingatanexclusive“Euro-centric”approach.AlreadytheheritageoftheViennaCirclewasnottheoreticaluniformitybutplurality,e.g.,withtheprincipleoftoleranceandtheacknowledgmentofanin-tegratedhistoryandphilosophyofscienceasaheuristicstrategycompletingthelinguisticandsemanticturnswithapragmatic-historicaldimension.Thereformu-lationanddiversificationofphilosophyofsciencewasthuspre-coined,evenifnotyetfullyelaboratedbecauseofhistoricalreasons.GiventhisprehistoryandintellectualcontexttheESF-Programmeisdesignedasafurtherdevelopmentbasedonearlierconceptionsbutalsochallengingsomewellknownhistoricallydetermineddogmas(likeanalytic/synthetic,theory/obser-vation,contextofdiscovery/contextofjustification,induction/deduction)incor-poratingrecentEuropeanandglobalresearchresults.Accordingly,theopeningconferenceaimedataddressingthecurrentsituationofphilosophyofscienceinEuropewithreferencetothemaintopicsandrecentresultsasasortofdescriptionandcriticalaccountofthestateoftheartregardingthemainfoundationalandmethodologicalissues.Intheforthcomingyearsthefiveteamswillfocusinseparateworkshopsre-latedtoannualtopicsfromtheirspecificperspective:“Explanation,predictionandconfirmation”(2009),“Probabilityandstatistics”(2010),and“TheSciencesthat5cf.FriedrichStadler,“HistoryofthePhilosophyofScience.FromWissenschaftslogik(LogicofScience)toPhilosophyofScience:EuropeandAmerica,1930-1960”,in:TheoKuipers(Ed.),HandbookofthePhilosophyofScience–FocalIssues.Amster-dametc.:Elsevier2007,pp.577-658.\n10FriedrichStadlerphilosophyhasneglected”(2011).AlargeclosingconferencewillbeheldinBo-lognaon“NewDirectionsinthePhilosophyofScience”.Astronginteractionandcooperationbetweenthesegroupswithinvitedspeakersinadditionisintendedandwillleadtoaseriesof5booksasproceedingsofthewholeprogrammeofwhichthisvolumeispublishedasthefirstone.Inherclosingremarksattheconference,MariaCarlaGalavotti,thechairoftheprogramme,summarizedwiththemethodologicallessonthatatypicaltraitofEuropeanphilosophyofscienceisattentiontohistoricalcontextandtheuseofhis-torytoidentifytrendsinargumentation,andprovideperspectivesandinterpreta-tionsoncontemporarydebates.Especially,shenotedatleast3importantinsights:(1)theimportanceofthehistoricalresearch,(2)therootsofempiricisminEurope,and3)thesignificanceofpragmatismforthefutureinvestigations.Thestatusquoofphilosophyofscienceafterthe“historicalturn”ischaracterizedbypluralityandspecializationallovertheworld.TheEuropeanperspectiveinphilosophyofscienceistheinclusionofthehistoricalrootsofcurrentdebatesandthefocusonmethodologicalproblemsthatcrossthevarioussub-disciplines.ThisvolumeisaseriousattempttoopenupthesubjectofEuropeanphiloso-phyofsciencetorealthought,andprovidethestructuralbasisfortheinterdiscipli-narydevelopmentofitsspecialistfields,butalsotoprovokereflectionontheideaof‘Europeanphilosophyofscience’.ThiseffortsshouldfosteracontemporaneousreflectiononwhatmightbemeantbyphilosophyofscienceinEuropeandEurope-anphilosophyofscience,andhowinfactawarenessofitcouldassistphilosophersinterpretandmotivatetheirresearchthroughastrongercollectiveidentity.Vienna,October2009Acknowledgment:ThanksgotothechairMariaCarlaGalavottiandsecretaryCristinaPaoletti,thefiveteamleadersandtoallcontributors,themembersoftheSteeringCommitteeofthePSE-programfortheircooperation.IamgratefultoDonataRomiziandMilesMacLeod(UniversityofVienna,DoctoralProgram“TheSciencesinHistoricalContext”)fortheiressentialcontributiontothebasicconferencereport,andRobertKaller(InstituteViennaCircle)forhislayoutandeditorialworkwiththisfirstvolumeintheseriesofthePSE-programmewithSpringerPublisher(Dordrecht).UniversityofViennaDepartmentofPhilosophyandDepartmentofContemporaryHistoryInstituteViennaCircleUniversitätscampus,Hof1A-1090Wien,AustriaFriedrich.Stadler@univie.ac.at\nPartI(TeamE)(teamleaderThomasUebel)FoundationalandMethodologicalDebates\n\nTHOMASUEBELSOMEREMARKSONCURRENTHISTORYOFANALYTICALPHILOSOPHYOFSCIENCEForthisfirstplenaryconferenceofournetwork,thehistoryofphilosophyscienceteamispresentingits“overviewplus”ofthecurrentstateofthedisciplineunderfairlytraditionalheadings.Ishouldsayrightawaythereforethatthepurposeofdoingsoisthatofgivingafairlyclearidentificationofeachcontribution’sstart-ingpoint.Thecategorieschosenaremeanttobeneitherexclusivenorexhaustive–nor,indeed,evaluative.ILetmebeginbynotingthatthehistoryofphilosophyofsciencehasmadetremen-dousprogressoverthelasttwodecades.Thattherenowexistsaninternationalscholarlysociety(called“HOPOS”)dedicatedtoworkinthisfieldwithbiennialconferencesandaplannedjournalisevenaninstitutionalindicatoroftheprogressmade.Butbeforewecanask“Sowhereareweinourdiscipline?”,wemustasknotonly“Whatmadethisdevelopmentpossible?”butalso“Whatpreciselyisitthathasdevelopedhere?”Bothofthepreliminaryquestionshavequickanswers,butthereareconsiderablecomplexitieshidingbehindtheirsuperficialplausibility.Oneofthethingsthatmadethegrowthofhistoryofphilosophyofsciencepossibleistheever-increasingdistanceofphilosophyofsciencefromitsbegin-nings.Whilesuchadistanceinevitablyencouragesdisciplinaryself-reflectioninahistoricalvein,thereisinthepresentcaseanadditionalpoignancywhichderivesfromthefactthat,atleastasremembered,analyticalphilosophyofscienceorigi-nallyprofessedtocarelittleabouthistory.Sowegettheapparentdialecticthat,withgrowingage,aoncea-orevenself-consciouslyun-historicaldisciplineac-quireshistoricalconsciousness.Onemayremarkthatthisishardlysurprisingnorinitselfnewsworthy,butthisisnottheonlydevelopmenttobetakennoteof.HereI’mthinkingofachangeinmethodologicalattitudethatlate20thcen-turyphilosophyofscienceprideditselfon,achangesometimescharacterisedasanaturalisticturnorevenaturntoscientificpractice:eitherwayitinvolvestheself-consciousrejectionofapriorireflectionaboutgrandphilosophicalthemesrelatedtoscienceandinsteaddemandsdetailedknowledgeofcurrentscientifictheoriesandexperimentalpractices.Soifhistoryofphilosophyofsciencewantstopartakeinthischange–whichIpresume–itspractitionersmustengrossthemselvesalsoinhistoryofscienceanddetailsofpastscientificpracticesinparticularscientificF.Stadler(ed.),ThePresentSituationinthePhilosophyofScience,ThePhilosophyofScienceinaEuropeanPerspective,DOI10.1007/978-90-481-9115-4_1,©SpringerScience+BusinessMediaB.V.2010\n14ThomasUebeldisciplines.Andgiventhedevelopmentsinhistoryofscienceinturn,thistypicallyinvolves,ataminimum,alsoanawareness,notonlyofdevelopmentsinternaltothescientifictheorisingsatissue,butalsoofthesocio-politicalandculturalcon-textsofthesetheorisings.Aqualificationisnecessaryhere.Theturntodetailsoffirst-orderscientifictheoryandpracticeisdemandedofhistoryofphilosophyofsciencetodifferentdegrees,namely,asappropriatetothecasesathand.Incaseswherethephiloso-phiesunderconsiderationthemselvesconcernedverygeneralthemesandweredevelopedatarmslengthfromscientificpractice,extensiveattentiontodetailsofthelatterislessimperativethanincaseswherethephilosopherinquestionwasapracticingscientist.Parallelqualificationsappearindicatedconcerningtheconsid-erationofthesocio-politicalandculturalcontextsofphilosophiesofscience,butherethedynamicseemsreversed:thegreatertheirdistancefromscientificpracticeandthemoregeneralthetopicsandconclusion,themoresuchcontextualisingseemsrequiredtounderstandthepresuppositionsofthesephilosophies.Againitisdifficulttogeneralise,butcertaintendenciesseemclearenough.Historyofphilosophyofsciencethen,atleastinitsnoblestintent,aspirestobeaninterdisciplinaryundertakingtoanevengreaterextentthancurrentphilosophyofscience.Itseekstocomprehenddevelopmentsinthephilosophyofscienceinrelationbothtothetechnicalcontextoffirst-orderscientifictheorisingandinrela-tionthegeneralsocio-culturalcontext.NotethatIdidnotsay“thedevelopmentofphilosophyofscience”,forsofarsuchgrandnarrativeshavenotbeennotontheagenda.Thisisnottosaythatideasofhistoricaltrajectoriesdonotinformworkinthehistoryofphilosophyofscience,butitseemstomethatitspractitionersaretooawareofthepitfallofreachingforviewsofthedevelopmentofthephilosophyofsciencesubspecieaeternitatis.Theirworkconcernsmoreorlessextendedepisodesandconsidersthemfromaparticularphilosophicalviewpoint—bethatproblemorprogramme-based.IITurningnowtoabriefandsomewhatroughcatalogueofworkdoneinrecenthistoryofphilosophyofscience,IwilldrawsomeminormoralsbearingonthetaxonomicmisgivingsIalludedtoearlier.Iwillthengoontooffersomethoughtsonwhyworkinhistoryofphilosophyofsciencemaybeturnedtophilosophicalgain.Stickingtomybriefforthe20thcenturyandtheanalytictradition,itispossi-blecategorisetheworkdoneintermsofphilosophicalmovementsandperiodsand\nSomeRemarks15intermsofthescientificdisciplineswithwhichitisconcerned.1(Thatwithsomeofthefigureswereachbackintothelate19thcenturyseemsinevitable:thosewithotherfociarewelcometoplundertheearlyandlate20thcenturyinturn.)Withoutanyclaimtocompletenesswefindacertainconcentration,intermsofmovements(withinorfadingintotheanalytictradition):•Austro-Germanpositivism(Mach,Petzold);•Frenchconventionalism(Poincaré,Duhem);•Britishempiricism(Russell);•logicism(Frege,Russell);•formalism(Hilbertandhisschool);•earlylogicalempiricism(ViennaCircleandBerlinSociety);•Neo-Kantianphilosophyofscience(Cassirer);•Lvov-Warsawschooloflogic(Twardowski,Lukasiewicz,Tarski);•orthodoxlogicalempiricism(post-WWIINorthAmerica);•criticalrationalism(Popper);•post-WWIIAustro-Germanphilosophyofscience(Stegmüller);•theemergenceofpost-positivistphilosophyofscience(Feyerabend,Kuhn,Hanson);intermsofdisciplines:•logic(predicatelogicandmodeltheory);•mathematics(foundationalissues);•probabilitytheory;•space-timetheories;•quantumphysics;•chemistry;•biology;•psychology;•socialscience.Needlesstosay,theseconcentrationsdocross-cutinvariouswaysbetweenmovementsanddisciplines.Forinstance,•workonMachtendstofocusonhisphilosophyofphysicsandpsychol-ogy;•workonRusselltendstofocusonhisphilosophyoflogicandarithmetic,ofphysicsandofpsychology;•workonPoincaréandDuhemtendstofocusontheirphilosophyofphys-icsandtheformer’sphilosophyofarithmetic;•workonearlylogicalempricismpredominantlyfocussesonitspropo-nents’philosophiesoflogicandmathematicsandofphysics,esp.space-time,withlittleattentiontoitsphilosophiesofsocialscience;1Giventhespaceconstraintsofthisessay,evenasamplingoftherepresentativelitera-tureisoutofthequestion.Readersmaycomparemylistwiththetopicsofrelevantpa-persgivenatpastHOPOSconferences(seehttp://www.hopos.org/conferences.html).\n16ThomasUebel•workonNeo-Kantianphilosophyofsciencetendstofocusonitsphiloso-phyofphysics,withitssouth-westernwing’sphilosophyofsocialsciencecomingintoview;•workonorthodoxlogicalempiricismandcriticalrationalismonitspro-ponents’philosophiesofphysics,nowoftenquantumphysics,andprob-ability,withsomeattentiontotheformer’sphilosophyofpsychology;•workontheemergingpost-positivistphilosophyofsciencetypicallyfo-cussesonitsproponents’philosophiesofphysics.Notsurprisingly,oftheempiricalsciences,physicstakesthelion’sshareofatten-tionwheretheinvolvementofpastphilosophiesofsciencewithfirst-ordertheoriesisconcerned.Workonpastphilosophyofbiologyandchemistryiscatchingupbutwouldappeartoescape,withonlyafewexceptions,associationwiththemove-mentsmentioned.Workonpastphilosophyofsocialscienceseemstocontinuetoplayadistinctlyminorrole.2Consideringeveninbriefestoverviewsomeofthevarietyandbreadthofworkinhistoryofphilosophyofsciencemakesclearthatitdefiesaninformativesum-mary.Butwecanask:howdoesitbearontheself-imageofanalyticalphilosophyofscience?First,wecanseethattheboundariesoftheanalyticaltraditionarebynomeanssharp.Thereis,forexample,thefruitfulexchangebetweenMach(whoalsostandsontheborderlineof19thand20thphilosophyofscience)andDuhemand,evenmoreimportantly,thelastinginfluenceoftheentireschoolofFrenchconventionalism–Poincaré,Duhem,AbelRey,lesssoLeRoy–onlogicalem-piricism.3WhatthelatterpointstoisthatinphilosophyofsciencewhateverbreakoccuredbetweenFrenchphilosophyofscienceandwhatwenowcalltheanalytictradition,itoccurredonlywellintothe20thcentury.(ThisbreakisnotablylaterthanthebreakwhichaccordingtoDummettsplitanalyticalandcontinentalphi-losophyoflanguagebetweenHusserlandthelaterFrege.)4AsecondexampleoftheboundariesoftheanalyticaltraditionbeingbynomeanssharpispresentedbyNeo-Kantianphilosophyofscience.Thuseventhoughthesouthwesternwingofthatmovement(Windelband,Rickert)couldberegardedasprototypically“continental”,itturnsoutthatCarnap’searliestmajorworkwasnotinsignificantlyinfluencedbyRickert.5Ofgreatestimportancehere,however,2What’sstrikingisthevirtualabsenceofworksonthehistoryofphilosophyofhistoryinoursub-discipline:thatthereisnodearthofsuchworkcanbeseen,e.g.,bythecontributionstothejournalHistoryandTheory.3See,e.g.,RudolfHaller,‘DerersteWienerKreis.’Erkenntnis22(1985)341-358,trans.‘TheFirstViennaCircle’inT.Uebel(ed.).RediscoveringtheForgottenViennaCircle.Dordrecht:Kluwer,1991,pp.95-108;AnastasiosBrenner(ed.),Interférencesettransformationsdanslaphilosophiefrancaiseetautrichienne.PhilosophiaScien-tiae3(1998-99)Cahier2.4MichaelDummett,“OriginsofAnalyticalPhilosophy”,LinguaeStile23(1988),3-49,171-210,repr.asamonographCambridge,mass.:HarvardUniversityPress,1993.5SeeThomasMormann,‘Carnap’sLogicalEmpiricsm,ValuesandAmericanPragma-tism.”JournalforGeneralPhilosophyofScience38(2007),127-146.\nSomeRemarks17isthattheheiroftheMarburgwingofthemovement,Cassirer,mustbecountedamostsignificantphilosopherofphysicsinhisownrightwhoseworkprovidedanindispensiblereferencepointforthephilosophyofspace-timeofearlylogicalempiricism.6Inthisway,earlyanalyticalphilosophyofsciencealsotapsintotheKantiantradition.Sohistoryofphilosophyofsciencehelpstobreakdownacertaininsularitywithwhichsomepractitionersofanalyticalphilosophyofsciencelookupontheirtradition.Noneofthisshows,however,thattheterm“analyticaltradition”iswith-outmeaning,butitstronglysuggeststhatattemptstodefineitintermsofnecessaryandsufficientconditionsformembershiparemistakenanditremindsusthat,likealltraditions,theanalyticonecanonlybeindividuatedintermsofhistoricalline-age–suchthatconnections,influencesandoverlapsastheonesjustmentionedcanonlybeexpected.Butthedispellingofeasyillusionsaboutthepastalsohaseffectsonthestorythathistoryofanalyticalphilosophyofsciencetellsaboutitself.Take,forexample,myearlierclaimthat“asremembered,analyticalphilosophyofscienceoriginallyprofessedtocarelittleabouthistory”.Well,isthatrememberedcorrectly?Whenwelookbeyondthelogicalempiricistorthodoxythatinstitutionaliseditselfbythemiddleofthe20thcentury,itbecomesobviousthatearlyanalyticalphilosophyofscienceitselfwasnotasuniformlyunhistoricalasreceivedwisdomhasit.Whatfollowsisthatwehistoriansofphilosophyofsciencealsomusttakecarethatinpresentingourowncasewedonottradeontheveryillusionsthatwe(partly)makeitourbusinesstounmask.Moreover,ourownenterpriseisbynomeansasnovelasitmightappear.Thushistoriansofancientandearlymodernphilosophywouldberighttoremindusthatafairamountoftheirworkcanbealsoclassifiedashistoryofphilosophyofscience,giventhatmuchoftheirsubjects’philosophicalthoughtdistilledreflectiononthescienceoftheirday.IIIIturntothequestionwhethertheconcernwithhistoricalfactsexhauststheinter-estthathistoryofphilosophyofsciencepossessesandwhatfollowsifitdoesnot.Asweshallsee,thisiscloselyrelatedtothequestionwhyourworktendstoberelativelylocalandshiesawayfromtheviewfromnowhere.Letmeapproachthequestionviathecommonplacethatit’sprettymuchim-possibletotellhistory“wieeswirklichgewesen”,withoutanysuperadditionordistortionwhatsoever.Likehistoryingeneral,historyofphilosophyofsciencecannotdowithoutwhatArthurDantocalled“narrativesentences”–sentencesthat6ErnstCassirer,SubstanzbegriffundFunktionsbegriff,Berlin:BrunoCassirer,1910,trans.inSubstanceandFunctionandEinstein’sTheoryofRelativity,1923,repr.NewYork:Dover,1953,pp.3-346.\n18ThomasUebeldescribeaneventbyreferencetoalaterone–andtheseintroducethehistorian’sperspective.7(Evenamerechronicle–aslongasitisnottheutopian“ideal”one–couldbeseentointroduceperspectiveduetheinevitableselectionofwhatitincludesandexcludes.)Importantly,thesenarrativesentencesdonotimpugnthefactualityoftheindi-vidualeventsthathistorydiscusses,northatofanycausalclaimsmadeaboutthem,buttherecognitionoftheinevitableperspectivalnatureofhistoryhasnonethelessgivenrisetoworriesaboutitsobjectivity.(Thisperspectivalnaturemayexpressit-selfnotonlyinconflictingevaluationsoftheeventsordevelopmentsunderdiscus-sion,butalsointhedifferenttemporalframeworkswithinwhichtheyaredescribedasunfolding.)Whatvalidatesthenarrativeshistorianslayout,giventhatthereisnotrans-historicalstandpointwemaytakeonthem?Theanswerwouldseemtobethattheresimplyisnovalidationavailablethatsinglesoutoneandonlyonenarra-tiveaslegitimatefromallthedifferentonesthatdealwithatleastsomeofthesamebasicfacts,sothechallengehereistoholdonbothtothefactualityofhistoryandtheplasticitythatcomeswithhistorybeingtoldfromaperspective.8Totheideathatnarrativityunderminestruthitisrightlyobjectedthattheex-istenceofdistinctbutcompatiblenarrativesdoesnosuchthing.9Obviouslyoneeventmayfigureindifferentstories.(Thefactscitedareeithertrueoffalseandtheselectionofeventsaspertinenttothehistorytoldissubjecttoobjectivecri-teriaofrelevanceetc.)Butwhatofnon-compatiblenarratives?Canwebesurethatthenarrativesthatwedeemacceptableinprinciplearemutuallycompatiblewhenweturntohistoryofphilosophy(includingphilosophyofscience)?CouldweguaranteethatanarrativeofagivenepisodeinthehistoryofphilosophyofsciencefromabroadlyKantianperspectiveiscompatiblewithonefromabroadlyempiricistperspective?Ihavenodefiniteanswerhere,butIamsceptical.ConsiderthedebateabouttheapplicabilityoftheKantianframeworktotheepistemologyofgeneralrelativity.Beyondagreementonthefactsofwhosaidandwrotewhatandwhen,thereseemstobelittlecompatibilitybetweenthenarrativesinvolvingSchlick,ReichenbachandCassirerinthisrespectastold,ontheonehand,byarecentattempttorevivetheresearchprogrammeoftranscendentalidealismforthephilosophyofphysicsand,ontheotherhand,astandardaccountoftherise7ArthurDanto,AnalyticalPhilosophyofHistory.Cambridge:CambridgeUniversityPress,1965.Rev.andenlargedasNarrationandKnowledge.NewYork:ColumbiaUniversityPress,1985,Ch.8.8Seethediscussionsin,e.g.,FredAnkersmithandHansKellner(eds.),ANewPhiloso-phyofHistory.Chicago:UniversityofChicagoPress,1995,orinBrianFay,PhilipPomperandRichardVann(eds.),HistoryandTheory.ContemporaryReadings.Ox-ford:Blackwell,1998.9SeeNoelCarroll,“Interpretation,HistoryandNarrative.”TheMonist1990,repr.inFayetal.,op.cit.,pp.34-56,arguingagainstHaydenWhite,Metahistory.Baltimore:JohnsHopkinsUniversityPress,1974.\nSomeRemarks19oflogicalempiricism.10What’sagenialcuttingofaGordianknotforthelatterisaprofoundmisunderstanding,anoutrightbeggingofthequestionfortheformer.Notablythesearenotmerevaluejudgementsbutpertaintothequestionofwhattheproblemsarethatwereatissue.Sothechallengetotheobjectivityofhistoryofphilosophyofsciencedoesnotappeartobeaseasilydeflectableasinotherhistoricalinquiries.InresponseI’dliketoarguethatinourfieldhistoryisnotouronlyconcernandthatthereforeworriesaboutobjectivitycanbecounterweightedappropriately,fortheplasticityofhistorybeingtoldwithaneyetotheopenfuturecaninthiscasebesuitably“dis-ciplined”.(Thisisnotsaythatallhistoryofphilosophyofsciencemustexemplifysuchperspective-taking,butthatitisnotillegitimatewhenthishappens,especiallyifitremainssuitablylocal.)AsIunderstandit,historyofphilosophyofscience,despiteitsstrongemphasisonrelevantaspectsofthefirst-orderscienceasprac-ticedandontheculturalcontextofitstheories,isstillpredominantlyphilosophyofscience.Itsmodusoperandisishistorical,yetitsaimisbynomeansexclusivelytoestablishthemerefactsofwhatwasthoughtwhenandbywhom(thoughtheymustbeestablishedassecurelyaspossibleasamatterofcourse).Somypointisthatsincehistoryofphilosophyofscienceisphilosophyofsciencebyothermeans,thedangerinherentintheperspectivalnatureofitshistoricalmethodologyisneu-tralisedbythenatureofitssubjectmatter.Letmeexplainfurther.Theabsenceofapresuppositionlessstartingpointmeansthatinphilosophyuncontestablefacts(beyondhistoricalonesaboutwhoclaimedwhatwhen)arefarhardertocomebythaninthesciences.Whenweascendtothemeta-level,thisbe-comesstillmoreobvious:whetheraphilosophicalresearchprogrammeisprogres-siveorregressiveismuchlessclearcutthanwhetherascientificresearchagendais.Itseemswemustaccepttheextremecontestabilityofitsdomain-specificfactsasafactaboutphilosophy.Considernowwhethertheinevitableperspectivalna-tureofhistoryofphilosophyofscienceintroducesanythingradicallynew,giventhatitisphilosophyofsciencewhatthehistoryishistoryof.Mysuggestionisthatitdoesnot.Whilehistoryintroducesanarrativeperspective,philosophyalreadycomeswithitsownpotentiallyincompatiblephilosophicalperspectives.Whatdoesthismeanforthenarrativesofthehistoryofphilosophyofsci-ence?11Doesthismeanthatheretwoincompatiblenarrativescanbothbetrue–orthatneithercanbe?Ordinarylogicpushesustotheconclusionthat,indeed,oneofthemwouldhavetobefalse.Ifoneisunwillingtourgetheabandonmentofbivalence(ortheembraceofmetaphilosophicalrelativism)thenoneisreducedtoaskingwhat,ifanything,makesfalseoneoftheseincompatiblenarrativeswhich,10CompareThomasRyckman,TheReignofRelativity.PhilosophyinPhysics1915-1925,Oxford:OxfordUniversityPress,1995,and,e.g.,HansReichenbach,TheRiseofScientificPhilosophy.Chicago:UniversityofChicagoPress,1951.11Similarconsiderationsmaywellholdforhistoryofphilosophygenerally,butIwillkeepmyremarkslimitedtooursubdiscipline.\n20ThomasUebelafterall,aresupposedtorespecttherelevanthistoricalfacts.Clearly,itmustbethephilosophicalpresuppositions.Whatfollowsforourhistoricalnarratives?Sincephilosophicaldisputes,givensufficientdepth,tendtoresistsettlement,itfollowsthatdisputesaboutincompat-iblenarrativesinthehistoryofphilosophyofsciencemaylikewisetendtoresistsettlement.Thatisnotundulyproblematic,however,oncewerecognisethathis-toryofphilosophyispartoftheongoingphilosophicaldebateandcontestation.Somysuggestionisthat,inthecaseofhistoryofphilosophyofscience,thequestionofwhatfalliblyvalidatesthehistorian’snarrativeperspectivefindsitsanswerinthecontestablephilosophicalperspective.Thenarrativeisvalidated(totheextentthatitisbeyondthedataitdealswith)bythephilosophicalperspectiveitunravels.Inotherwords,thatdifferentnarrativesinthehistoryofphilosophyofsciencewhichrespecttherelevanthistoricalfactsmaystillbeincompatibleneednotberegardedasimpugningtheobjectivityofhistorybutasinsteadpointingtothephilosophicalnatureofthehistoryattempted.Sohistoryofphilosophyofscienceisnotjustlikeotherhistory,norisitjustlikeotherphilosophy.Thishasconsequencesfortheuseofanepithetthathis-toriansarefondoflabelingopponentswith:“presentism”.Wecanallagreethatreadingcontemporaryconcernsintothepastisabadideawhendonewithoutso-phisticationandespeciallywithoutanybackingfromthematerialatissue.ButwemustequallyagreethatalreadythephenomenonofWeberianvalue-relevance(thatatheorist’svaluesinfluenceherchoiceofresearchprogramme)banishestheideaofwhollydisinterestedhistoricalresearch.YetthepointInowwanttopressonthebackofDanto’saboutnarrativityisstillstronger.Itisthat,oncethehistoricalfactsofthecaseareestablished,historyofphilosophyofsciencecanbeaspartisanasphilosopyitself.Providedtheproperprecautionsaretaken,itssearchforcon-temporaryrelevancecannotbedismissedas“presentist”,normustitstendencybeirenic(thoughitmaydelightinfindingsthatmaybeconsideredironic).IVLet’sconsiderfurtherthenthelegitimatepartisanshipofhistoryofphilosophyofscienceintheformoftheexplorationof“pathsnottaken”.Letmeleadintothisbyaskingwhethertherearestillmoresubstantialillusionsaboutitspastthanitseasyseparabilityfromothertraditionsthatinformtheself-imageofcontemporaryana-lyticalphilosophyofscienceandthatareoverturnedbyworkonitshistory?IthinkthereareandtheycanbesummarisedintermsofbonmotbyJohannNestroythatWittgensteinadoptedasthemottoforhisPhilosophicalInvestigations,namely,thatprogresstendstolookbiggerthanitis.Hereonemaypointtothecommonconceittopaintthelogicalempiricists–typicallyunderthelabel“logicalpositivists”–asphilosophersofastonishing\nSomeRemarks21naivity,inshort,asnegativeposterboysforwhatevernewthesisapost-positivistphilosopherisoffering.Inthatcapacity,logicalempiricistphilosophyisportrayedaspursuantofepistemologicalfoundationalismandontologicalreductionism,committedtoafact-fetishisinginstrumentalismthatrefusestograntrealitybeyondthegivenandconsequentlyasdenying,withparticularlydisastrouseffectinthesocialsciences,thecriticalforceofinvestigatingalternativepossibilities–allthis,moreover,fromaperspectiveonscientifictheorisingthatabsolutisesitsformalis-ableaspectsandabstractsfromallsocio-historicalaspectsofscientificpractice.Needlesstosay,everyoneofthesefeaturesiseminentlycriticisable,notonlysystematicallybutespeciallyasattributedtologicalempiricism.Nowthereisnodenyingthatsomeofthecriticismsjustsurveyeddofindaproperobjectamongthetheoriesandtheoriststhatcanbeassembledundertheum-brellaoflogicalempiricism.Forinstance,thereceivedviewofscientifictheoriesdidgiverisetoallsortsofuncomfortablequestionsaboutthestatusoftheoreticalentities.Moroever,orthodoxlogicalempiricismgenerallyappearstohavebeencaughtupinakindofformalisationfrenzy,ignoring,undertheguiseofobservingdisciplinarystricturesregardingthedistinctionbetweenthecontextsofdiscoveryandjustification,whateverresistedsuchtreatment:ithadnoroomforconsidera-tionofscientificpracticeanditssocialcontextandhistoricaldimensions.Buteveninearlylogicalempiricismwecanfindperiodswheretheparty-line,asfarasitwasobserved,invitescriticismofunduereductionismandwhere,oncetherewasnolongerauniformparty-linetospeakof,wecanfindtheoristswhoseemedtobepursuingthetypeoffoundationalistprogrammethatpost-positivistsrejectsovigorously.Evenso,however,therearealsoaspectsofearlylogicalempiricismthatcontradictthecommonstereotypemoststarklyandthoroughly.12Sothefirstfactuallessonthathistoryofphilosophyofsciencehashereisthatearlylogicalempiricismwasbynomeansamonolithicmovementandmuchmorevariedinphilosophicaloutlookthanthepost-WWIIorthodoxy.Thesecondfactuallessonisthatmuchoftheanti-foundationalismandanti-reductionismthatpost-positivismpridesitselfoncanalreadybefoundinsomeversionsofearlylogicalempiricism.Thethirdisthateventhedemandtopayattentiontothehistoricalandsocialdimensionofscientifictheorisingheldnosurpriseforproponentsofthisver-sionoflogicalempiricismwhosoughttoovercometherigiddichotemybetweentherationalandthesocial.Thatthisversionthusanticipatedsomeofthecriticismsthatlaterwereraisedagainsttheorthodoxlogicalempiricismofthe’50sthussug-gests,thatitislargelyimmunetostandardanti-positivistcriticisms.Sowhat?youmaybetemptedtoreply.Typicallytheversionoflogicalempiri-cismjustinvokedwassomewhatheterodoxfromthestartandcertainlywasmar-ginalisedbythetimetheorthodoxycametoholdsway.Thepointofpost-positivist12Forarecentstock-takingofresearchonlogicalempiricism,seeAlanRichardsonandThomasUebel(eds.),TheCambridgeCompaniontologicalEmpiricism.Cambridge:CambridgeUniversityPress.2007.\n22ThomasUebelpriderightlyliesinthatwhatwasheterodoxthenisnowcommonconsensus:thisprogressmaynotbeasradicalasclaimed,butsurelythat’sprogressenough.Ishouldstressthatthehistorian’splaintandtheunconvincedreactiontoitarenotuniquetothiscase.Indeed,asomewhatparallelrevisionistargumentmaybedevisedfortranscendentalidealismoftheMarburgvariety.FarfrombeingstuckinthetraditionalKantianmould,Cassirerdevelopedthenotionofthesyntheticaprioriandde-apodicticisedit,renderingitrelativetoahistoricallygivenframe-workoftheorisinginsteadofconceivingofitasaconstantpresuppositionofhu-manreasonassuch(without,however,conventionalisingit,aslogicalempiricistswerewonttodo).13GivenhowthisconceptionoftheapriorichimeswithKuhnianphilosophyofscience,againhistoryofphilosophyofsciencemaylayclaimtoshowingthataverytopicalandcurrentphilosophicalthesisisnotreallyterriblynew.Inthiscaseitisthethesisthatsignificantscientifictheoriestendtoproceedagainstabackgroundofsubstantiveassumptions(oftenhiddenbytheiradoptionofnewformsofmathematicalrepresentation)thattheythemselvesareinnopositiontoredeem.Sowhat?youmaybetemptedtoreplyagain.Thepointofpost-Kuhnianpriderightlyliesinthatwhatoncewasamaverick’sviewisnowmuchmorewidelyheld:thisprogresstoomaynotbeasradicalasclaimed,butsurelythat’sprogressenough.Thismaybeconceded.Yetconcedingthisdoesnotunderminethesig-nificanceofdispellingtheillusionoftheagetohavemaderadicalprogress.Foronething,itcounselsacertainmodestyandthatisnobadthing.Foranother,therecognitionthatnoteverything“logicalpositivist”oreverything“transcendentalidealist”belongsinthedustbinofhistorymayalsopromptthequestionwhetherthatheterodoxversionoflogicalempiricismorthatlatedevelopmentoftranscen-dentalidealismdonotalsoholdothersuggestionsorinsightsthatmaybefruitfullybroughttocurrentdebates.Thisbringsustotheexplorationof“pathsnottaken”.First,letmenotethatthisisoptional,notobligatory.Second,thatpointingoutthatexploringpathsnottakenmayholdprofitablevistasneednotrepresentmessianichubrisleadingusbacktosome“goldenage”.Andthird,thatthisdoesnotmeancrossingthelinebetweenhistoryofphilosophyofscienceandsystematicphilosophyofscience.Evenhistoryofphilosophyofsciencethatisphilosophyofsciencebyothermeansstillusesthoseothermeansessentially.Itdoesnotexplorephilosophicaltopicsabinitio,asitwere,fromfirstprinciples,butwithinframeworksorfromassumptionsthatarehistoricallygiven.Topursuepathsnottakeninthiscontextmeanseithertothinkthroughtotheirconclusioncertainphilosophicalhypothesesmorethoroughlythantheywereinthehistoricalsituationinwhichtheyaroseortovarycertainelementsinthatoriginalconfigurationofideasandprobetheplau-sibilityofthetargethypothesisunderthesestrictlycounterfactualcircumstances.13SeeMichaelFriedman,“ErnstCassirerandthePhilosophyofScience.”InGaryGut-ting(ed.),ContinentalPhilosophyofScience.Oxford:Blackwell.2005,pp.71-83.\nSomeRemarks23(Importantly,asIunderstandithere,pursuingpathsnottakendoesnotmeanim-aginingthecourseofhistoricaldebatestohavebeendifferentandimaginingtheconsequencesofdifferentoutcomes:itdoesnotmeanpursuingcounterfactualhis-toryassuch.)Yettheexplorationofpathsnottakenmayalsohaveresultsforphilosophyofsciencethatareofsignificanceforitsself-understanding–inawaystilldifferentfromthemoralssuggestedearlier.There’sonecasewhichIthinkraisestheissuesinaparticularlystarkway.(Letmealsostressthatthisisanexamplefrommyownworkandisnotherepresentedastheonlycaseofapathnottakenthatisworthexploring.)Intheearly1930sthereobtainedadivisionintheViennaCirclecharacterisedsomewhatmischievi-ouslybyCarnap,followingNeurath,as“leftwing”and“moreconservative”or(misleadingly)“rightwing”.14Theso-calledleftwingwascomprisedofCarnap,Neurath,Hahn(d.1934)andFrank,themoreconservativewingcentredonSchlickandWaismann.Thoughtheterm“leftwing”invokedasharedpoliticaloutloookamongstitsmembers,itsmaindesignationwasphilosophical.TheleftwingwascharacterisedatfirstbyitsoppositiontoWittgenstein’sradical(namelycomplete)verificationism.Mostimportantly,itpioneeredthethoroughrejectionofepiste-mologicalfoundationalisminthephilosophyofscience.(Needlesstosay,enoughtopicsofdisagreementremainedamongstthem.)Oneofthemoststrikingofitssharedphilosophicaltheseswasthattraditionalphilosophyhadoutliveditsusefulnessandthatwhatremainedusefulofphiloso-phywashappilyconceivedasmeta-theoryofscience,asasecond-orderinquiryofitselfscientificnature,inotherwords,asscienceintheself-reflexivemode.What’sparticularlyinterestingistheformofthemetatheorythatwecanascribetothem.Importantly,itcameintwoforms.TherewasCarnap’s“logicofscience”andNeurath’s“behavioristicsofscholars”,anaturalistic“pragmaticsofscience”,inFrank’slaterterminology.Thelogicofscienceinvestigatedscientifictheories,theirinternalstructureandtheirrelationtotheirevidentialbaseinpurelylogicalterms(deductiveandinductive).Thepragmaticsofscienceinvestigatedscientificpracticebymeansoftheempiricalsciencesofscience,thepsychologyandsoci-ologyaswellasthehistoryofscience.Sowhiletheformerinvestigatedabstractrelationsofevidentialsupport,thelatterinvestigatedconcretetheorychoiceandchange.Whatwastherelationbetweenthetwometatheories?Clearly,theyarediffer-entbothinmethodology–oneusingformalapriori,theothermaterialaposteriorireasoning–andintermsoftheirobject.Itisperhapsastandardviewtoregardthemasstanding,ifnotinoutrightopposition,atleastinconsiderabletensionwitheachother.GivenacomparisonoftheexemplaryclaritywithwhichCarnap’sin-quiriesproceededwithNeurath’sdecidelylessclearexplorations,itisperhapsno14SeeRudolfCarnap,“IntellectualAutobiography”,inP.A.Schilpp(ed.),ThePhiloso-phyofRudolfCarnap,LaSalle,Ill.:OpenCourt,1963,p.57.Alreadyin“TestabilityandMeaning”,PhilosophyofScience3(1936)422,Carnapspokeofthe“morecon-servativewing”.\n24ThomasUebelsurprisethatthisviewwontheday,moreover,thatthelatter’s“behavioristicsofscholars”wasnottakenupintheburgeoningmovementoflogicalempiricism.Indeed,Carnap’sclaimthat“oncephilosophyispurifiedofallunscientificelements,onlythelogicofscienceremains”iscommonlyreadassayingthatle-gitimatephilosophycomprisedonlythelogicofscience,nothingelse.15HislogicofscienceistheneasilyassimilatedtoReichenbach’s“analysisofscience”withintheso-calledcontextofjustification,withallconcernwithhistoryandsociologybanishedtotheunphilosophicalcontextofdiscovery.Suchareadingoverlooks,however,thatReichenbachallowedintohisanalysisofsciencenotonlytheprob-lemsoflogic,probabilitytheory,butalso“allthebasicproblemsoftraditionalepistemology”.16Bycontrast,Carnapstressedthattodesignatehislogicofscienceas“theoryofepistemology(orepistemology)”is“notquiteunobjectionable,sinceitmisleadinglysuggestsaresemblancebetweentheproblemsofourlogicofsci-enceandtheproblemsoftraditionalepistemology”.17Carnap’sviewoftheroleoflogicofscienceassuccessortophilosophyanditsrelationtothepragmaticsofsciencewasquitedifferentfromReichenbach’s.NowforNeuraththroughoutbutexplicitlysosincethemid-’30s,therelationbetweenthelogicofscienceandthebehaviouristicsofscholarswasoneofcoexist-enceandcomplementation.18Likewise,alreadyintheearly’30sFranksuggestedthatunderstandingsciencerequiredaddingthesociologicaldimensiontoatheoryofsciencethatwasconductedmostlyintermsofthe(atthetimesyntactic)analysisofthesymbolsystemitused.19Thisofcoursemeantaddingsomethinglikeabe-haviouristicsofscholarstothelogicofscience.(Inthe1950sFrank’sworkmainlyconcernedthepragmaticsofscience,butwaswidelyignored.)BothFrankandNeurathcanthusbecountedassupportersofaconceptionofthesuccessordisci-plinetophilosophyasconsistingofboththelogicofscienceandthepragmaticsofscience,thatis,abipartitemeta-theoryofscience.SowhataboutCarnap?Carnaprecognisedasperfectlylegitimate,“inadditiontothelogicofscience…alsotheempiricalinvestigationofscientificactivity,suchashistorical,sociological,and,aboveall,psychologicalinquiries”andgroupedbothofthemtogetherundertheheading“theoryofscience”.20Thisamountstoa15RudolfCarnap.LogischeSyntaxderSprache.Vienna:Springer,1934.Revisedandtrans.TheLogicalSyntaxofLanguage.London:Kegan,Paul,TrenchTeubner&Cie,1937.Repr.Chicago:OpenCourt,2002,p.279.16HansReichenbach.ExperienceandPrediction.Chicago:UniversityofChicagoPress,1938.Repr.NotreDame:UniversityofNotreDamePress,2006,p.8.17Carnap,op.cit.,p.280.18SeeOttoNeurath.“PhysikalismusundErkenntnisforschung.”Theoria2,97-105,234-7.Trans.“PhysicalismandtheInvestigationofKnowledge”inNeurath,PhilosophicalPapers1913-1946(ed.byR.S.CohenandM.Neurath),Dordrecht:Reidel,1983,pp.159-67.19SeePhilippFrank.DasKausalgesetzundseineGrenzen.Vienna:Springer.Trans.TheCausalLawanditsLimits,Dordrecht:Kluwer,1998,p.14.20Carnap,op.cit.,p.279.\nSomeRemarks25bipartitescientificmetatheorywhichlikewisedividesintologicalandempiricalinquiries.Ofcourse,Carnap,forhispart,workedonlyonthelogicofscience.Butheexplicitlyendorsedthebipartitenatureofmeta-theorybyacknowledgingtheimportanceoftheempiricalaspectsofthetheoryofscienceinhiseulogytoFrankandnotedinacomprehensivereviewofhisownworkthat“unfortunatelyadivisionoflaborisnecessary,andthereforeIamcompelledtoleavethedetailedworkinthisdirectiontophilosophicallyinterestedsociologistsandsociologicallytrainedphilosophers”.21Carnap’sviewiscommensurablewithNeurath’sandFrank’s.Thissuggeststhatthe“LeftViennaCircle”assuchembracedtheideaofabipartitemetatheoryofscience.(Needlesstosay,onoccasionthornyquestionsofrelativepriorityandgeneralrelevancecanarisebetweenthetwopartsofscientificmetatheory:infact,justthose,withsomepersonalissuesadded,didariseandledtotheestrangementinlateryearsbetweenCarnapandNeurath,butthisdoesnotaffecttheprinciple.)Moreover,sinceFrankhadrecommendedthatitwaspreciselytherecoursetothe“morecomprehensivestudyoftheconnectionsthatexistbetweentheactivityoftheinventionoftheoriesandtheothernormalhumanactivities”22thatobviatedtheneedformetaphysicalphilosophy,onemaystatethatthatfortheLeftViennaCir-cle,itwasthebipartitemeta-theoryofsciencethatwasthesuccessoroftraditionalphilosophy.Sowhat?Well,itstrikesmeasnotimpossiblethatthisconceptionofabipar-titemeta-theoryshouldproveusefulinthefurtherdevelopmentofwhatnowa-dayswouldbecalled“deflationary”philosophyofscience.Sohere’sone“pathnottaken”thatmight,forcolleagueswithdeflationistsympathies,meritfurtherconsideration.Yetthatisnotallthatthecasemayholdforus.ForitmightbearguedthatIwentonesteptoofar.Itmightbeconcededthatinprinciplethelogicofscienceandthepragmaticsofsciencearecompatibleandcombinable,andyetcontestedthattheyweresoputforwardbytheleftwingoftheViennaCircle,indeed,thattheveryideaofthatleftwingasaphilosophicallycoherentgroupandofCarnap,FrankandHahnsharingacommonprogrammeisjusttoofancifultobebelieved.ThisobjectionraiseswhatIthinkisanimportantpointnotonlyaboutthephilosophyofsciencehereatissue,butalsoaboutphilosophyofscienceingeneral.Considerfirstthatlogicalempiricistsgenerally–thoughnotuniversally–sawthemselvesengagedintheprojectofrenderingphilosophyofscienceitselfsci-21See,respectively,Carnap,“AFewWordstoPhilippFrank,fortheSecondVolumeoftheBostonStudiesinthePhilosophyofScience.”InR.C.CohenandM.Wartofsky(eds.).InHonorofPhilippFrank.BostonStudiesinthePhilosophyofScienceVol.2,NewYork:HumanitiesPress1965,pp.xi-xii.AndCarnap,“CommentsandReplies.”InP.A.Schilpp(ed.),ThePhilosophyofRudolfCarnap,LaSalle:OpenCourt,1963,p.868.22Frank,op.cit,p.14.\n26ThomasUebelentific.23AccordingtotheCircle’sinofficialmanifesto,“thescientificworldcon-ception”endeavours“tolinkandharmonisetheachievementsofindividualin-vestigatorsintheirvariousfieldsofscience.Fromthisaimfollowstheemphasisoncollectiveefforts,andalsotheemphasisonwhatcanbegraspedintersubjec-tively…”24Prominentherearetwocentralfeaturesofphilosophyasascientificactivity:co-operationandintersubjectivity,theformeramethodofinquiry,thelatteraconditiononitsresults.ItistheformerthatIwanttoconcentrateonhere.(Thelatterfeatureprovedcrucialinbringingouttheleftwing’sagreementonanti-foundationalismby1932whichhadnotyetbeenachievedin1929.)25Carnap,alongwithHahnandNeurath,hadbeenoneofthesignatoriesofthemanifestowhileFrankwastheorganiseroftheconferenceatwhichitwasfirstpresentedtothepublic.Sowhen,aswesaw,Carnapexplicitlyinvokedthedivisionoflabourin1963,hegaveexpressiontoalong-heldview.Thissuggeststhatnotonlywerethelogicandthepragmaticsofsciencecompatibleaseachwasconceivedoffromabout1932onwards,butthattheywererecognisedassuchalreadythenbythemembersoftheLeftViennaCircle.Howthentoexplaintheundeniablefact,whichadetractormightfocusupon,thatCarnap,FrankandNeurathremarkeduponthissharedmetaphilosophicalcon-ceptionsorarelyintheirpublications?Toclaimthatanycompatibilityinprinci-plebetweenthesestheyembracedwassopatentlyobvioustothemthatnowordsneededtobewastedonit,wouldperhapsbeabitrash.(ParticularlybetweenCar-napandNeurath–thoughmainlyduringtheprotocolsentencedebatethatledtotheirsharedanti-foundationalism–therearenumerouscaseswheredisagreementlurksevenunderconfessedagreement.)Allthat’srequiredthoughisthatbetweenexplicitagreementandunnoticedcomplementaritylieavarietyofpossibilities,eachofwhichsufficetounderwritemyclaimabouttheLeftViennaCircle’sviewonscientificmetatheory.Especiallyagainsttheirexplicitlysharedprogrammeofrenderingphilosophyofsciencescientific,thedocumentedconvergenceoftheirremarksonthenatureofscientificmetatheoryasthesuccessortotraditionalphilosophycanbetakentohavebeenreadilyunderstoodbyCarnap,FrankandNeurath.Itisanessentialfeatureofthemanifesto’sprogrammethatphilosophyofsciencebeacollectiveundertaking,requiringmanymindsandhandsindifferentcapacitiesandindiffer-entdepartmentsofunifiedscience,asitwere.ThedivisionoflabourthatCarnap23Thesignificanceoftheprojecttorenderphilosophy“scientific”hasbeenexploredbyAlanRichardsoninseveralpapers;see,e.g.,his“ScientificPhilosophyasaTopicforHistoryofScience”,Isis99(2008)88-96.24RudolfCarnap,HansHahnandOttoNeurath.WissenschaftlicheWeltauffassung-DerWienerKreis.Vienna:Wolf,1929.Trans.“TheScientificConceptionoftheWorld:TheViennaCircle”.InNeurath,EmpiricismandSociology(ed.byM.NeurathandR.S.Cohen),Dordrecht:Reidel1973,p.306.25SeemyEmpiricismattheCrossroads.TheViennaCircle’sProtocolSentenceDebate.Chicago:OpenCourt,2007.\nSomeRemarks27remarkeduponretrospectivelyin1963wasalreadyprojectedintoforthcomingworkin1929,thoughnotinthepreciseformthatitwastotake.IfitiscorrectthatwemayindeedattributeasharedperspectivetotheLeftViennaCirclealongtheselines,thenthereareconsequencesthatinprinciplegofarbeyondthecaseathand.Historyofphilosophyofsciencemayhavetotakesomeoftheprogrammaticstatementsofthethinkersitstudiesmoreseriouslythanitoftendoes,inparticulartheirclaimstobeinvolvedinacollectiveundertaking.Theremaybemoretoanindividual’sphilosophyofsciencethanwhatisapparentinthatindividual’swork.Thereasonisthatthisindividualmayhaveseenhim-orherselfasworkingina“community”whichprovidedessentialextensionsorcom-plementationtotheworkheorsheperformedandwhichwasverymuchmeanttobeunderstoodinthiswidersetting.Forinstance,tostaywithCarnap,itmaybegravelymistakentocriticisehisphilosophyofscienceasbeingpurelyformalistandunreflectiveofthepracticalandsocio-historicaldimensionsofscience.Tobesure,thelogicofscienceis,butconsidered“intheround”thatisnotallthathisphilosophyofscienceamountsto.WhatjustifiesusinrejectingsuchacriticismofCarnapandtakingawiderviewisprecisely,ofcourse,hisdocumentedcommit-menttophilosophyofscienceasacollectiveenterprise:anyoneworkerwillonlyeverbeabletoworkonsomepartorother,ifluckyandparticularlygiftedperhapsonseveral,butnever(orhardlyever)onallaspectsofit.VI’mafraidI’vetalkedabitmoreaboutdoinghistoryofphilosophyofsciencethanI’vedoneit,thoughIhopeyouwillagreethatwhatI’vedonesupportswhatIsaidaboutdoingit.InparticularIhopetohaveshownthatthehistoryofanalyticphilosophyofscienceoffersupaninterestingsuggestionofhowphilosophyofsci-encemaybethoughttodifferfromotherfieldsofphilosophy.Ofcourse,whetherweallwouldwishtoagreethatphilosophyofscienceisasignificantlycollectiveundertakingandthatinthisregarditdiffersfrommetaphysics,say,isyetanothermatter.Eitherway,however,bypresentinguswithsuchacasestudy,thehistoryofanalyticalphilosophyofsciencepresentsphilosophyofsciencegenerallywithyetanotheroccasionforreflectiononwhatitisandwantstobe.PhilosophySchoolofSocialScienceUniversityofManchesterManchesterM139PLUnitedKingdomthomas.uebel@manchester.ac.uk\n\nTHOMASMORMANNHISTORYOFPHILOSOPHYOFSCIENCEASPHILOSOPHYOFSCIENCEBYOTHERMEANS?COMMENTONTHOMASUEBEL1.INTRODUCTIONAdornoonceremarkedthatthehistoryofphilosophyisthehistoryofforgetting.Problemsandideasonceexaminedfalloutofsightandoutofmindonlytoresur-facelaterasnovelandnew.Ontheotherhand,understandinghistoryashistorio-graphy,thehistoryofphilosophymaybedescribedastheorganizedandinstitu-tionalizedattemptofovercomingtheever-growingthreatofhistoricalamnesia.Evenif„historyofscienceisnotjustlikeotherhistory“(Uebel),Adorno’sdictummaywellbetruealsoforthe(stillratherbrief)historyofphilosophyofscience:problemsandideasonceexaminedfalloutofsightonlytoresurfacelaterasnovelandnew.Examplesareeasilyfound:historicalandsociologicalaspectsofsciencethatweretreatedbyauthorssuchasDuhem,Bachelard,Cassirer,NeurathorFleck,fellintooblivionandwereignoredbymainstreamphilosophyofscience,onlytoberediscovereddecadeslater.Inadifferent,perhapsevenworsemanner,theleadingideasofsomehistoricalcurrents,forinstanceclassicallogicalempiri-cismoftheViennaCircleorNeokantianphilosophyofscience,sufferedfrompar-tialhistoricalamnesia:seriouslydistortedbylaterauthorstheysometimesservedastheirowncaricatures.Thereby,toborrowNestroy’swell-knownbonmottheprogressinphilosophyofscienceoftentendedtolookbiggerthanitreallywas.Adornomighthaveexaggeratedthedangerofhistoricalamnesiathatthreatensphilosophy,andcertainlyhedidnotspendathoughtonthesituationinphilosophyofscience.Buthisclaimmightserveasanantidoteagainstthenaiveideathatthehistoryofphilosophyofsciencehasbeenahistoryofpermanentandunilateralprogressinthesensethattodayweobviouslypossessabetterphilosophicalun-derstandingofthesciencesthanourforefathers.Rather,themoreprogresshistoryofphilosophyofsciencemadeinrecentyearsthemoreitbecameclearhowmuchremainstobedoneforachievinganadequateunderstandingofthestillyoungdisciplinecalledphilosophyofscience.Thisnotewillnotdealwithhistoryofphilosophyofscienceingeneral,itonlyhasthemodestaimtomakesomecommentsonUebel’ssurveyarticleonthehistoryofanalyticalphilosophyofscience,inparticularonhisproposaltounder-standthetaskofhistoryofphilosophyofscienceasdoing„philosophyofsciencebyothermeans“conceivingitasanarsenalofabandonedorforgottenconceptualpossibilities,or,asheputit,asthe„explorationofpathsnottaken“.F.Stadler(ed.),ThePresentSituationinthePhilosophyofScience,ThePhilosophyofScienceinaEuropeanPerspective,DOI10.1007/978-90-481-9115-4_2,©SpringerScience+BusinessMediaB.V.2010\n30ThomasMormannTheoutlineofthispaperisasfollows:InthenextsectionI’dliketopointoutinwhatsenseUebel’stopic–thehistoryofanalyticalphilosophyofscience–maybeconsideredasaparticularlydifficultbutalsoparticularlyinterestingtopicofhistoryofphilosophyofscience.Theaimofsection2istoshowthattheunderstandingof„historyofphilosophyofscienceasphilosophyofsciencebyothermeans“maybeconceivedasanexpedientstrategyofforestallaprofusionofundesiredmeta(meta)disciplineswhichthreatentheconceptualunityofaninter-disciplinaryresearchdealingwiththehistoryandphilosophyofscientificculture.Insection3IdealwithsomeproblemsthatmayariseforUebel’sfavoriteexampleofa„pathnottaken“,towit,whathecallsthe„bipartitemetatheory“ofCarnap,Neurath,andFrank.Insection4IarguethatUebel’sproposalofacombinedme-tatheorymaybecomparedwithMorris’searlierproposalofasynthesisoflogicalempiricismandAmericanpragmatisminthe1930s.Morris’sattemptfailed,anditmaybeusefultoinquireintothereasonsofthisfailuretobeinabetterpositiontoassesstheprospectsofUebel’sproposal.2.THEMANYISSUESONTHEAGENDAOFHISTORYOFPHILOSOPHYOFSCIENCELetmestartwithanobservationconcerningtheprogrammeof„HistoryofPhi-losophyofScience“ofthiscongress.Fromalogicalpointofviewthecollectionsoftopics•HistoryofAnalyticalPhilosophyofScience•HistoryofPhilosophyofScienceintheFrenchTradition•Historyofthe19thCenturyPhilosophyofSciencemayappearabitmysterious.ComparingitwithBorges’sfamousclassificationof„animalsinacertainChineseencyclopedia“intendedtodefyeveryattemptofexplainingitrationally,maybeexaggerated,butIthinkhistoryofphilosophyofscienceshouldmakesomeeffortstoexplicateasclearlyaspossiblethedomainofpossibleissuesthatitwantstodealwith.Atleastforsomeoftheabovemen-tioneditemsthisseemseasyenough.„HistoryofPhilosophyofScienceintheFrenchTradition“and„Historyofthe19thCenturyPhilosophyofScience“maybecharacterizedas„local“subdisciplinesinageographicalandintemporalsense,respectively.Otherexamplesforlocalsubdisciplinesinthissenseeasilycometomind:„HistoryofPhilosophyofScienceinthePolishTradition“or„intheGer-manTradition“,or„Historyofthe18thor17thCenturyPhilosophyofScience“andsoon.However,a„dimensional“classificationofthiskindbynomeansexhauststhepossibletopicsofourdiscipline:InarecentnewslettertheeditorsoftheSocietyforHistoryofPhilosophyofScience(HOPOS)callforaseriesof„stateoftheart“essaysdealingwith„HOPOSfigures“(sic)includingAristotle,Descartes,\nHistoryofPhilosophyofScience31Newton,Leibnizandmanyothers.Thereisnoreasontocriticisesuchapersonal-izedapproachofhistoryofphilosophy,butitiscertainlyofaquitedifferentkindthantheonedealingwith„local“topicsofvarioustype.StillanotherpossibilityisexemplifiedbyUebel’schoiceof„AnalyticalPhilosophyofScience“asatopicofhistoryofphilosophyofscience.WithoutarguingforitIthinkthatUebel’schoiceis„morephilosophical“andmoreinteresting,atleastifweconceivehistoryofphilosophyofscienceasanotherwayofdoingphilosophyofscience.Nevertheless,thetopicofanalyticalphilosophyofscienceisasomewhatdelicatechoice.Somephilosophers,rootedintheanalyticaltradition,stillbelievethatanalyticalphilosophyofscienceistheonlyphilosophyofsciencethatistobetakenseriously.Allothereffortsundertakenbyphilosophersinthecourseofhistorytocometotermswithsciencemaysimplybedisqualifiedasmetaphysicalrubbish.Asimilar,slightlylesspretentiousstanceistoconsiderallnon-analyticalphilosophyofscienceasnothingbutahistoricalprecursoroftherealthing,i.e.an-alyticalphilosophyofscience.Thiswayofinterpretingtheplaceofanalyticalphi-losophyofscienceisnotamereremotepossibility,ithasbeencommonusageinmanyquarters.Forinstance,KitcherandSalmon,intheiranthologyOnScientificExplanation,classifyDuhem’sLaTheoriePhysiqueexplicitlyasbelongingtothe“ModernPrehistory”ofthesubject(cf.KitcherandSalmon1989,ChronologicalBibliography,196).Uebeldoesnotsubscribetotheseradicalwaysofdetermingtheplaceofanalyticalphilosophyofscienceinthehistoryofthesubject.Rather,heisatpainstopointoutthattheboundariesbetweentheanalyticalandothercurrentsofphilosophyofsciencearebynomeanssharp.Nevertheless,theplaceofanalyticalphilosophyofscienceremainsremainsdifficulttodetermine,andinanycaseitissomewhatspecialcomparedwiththeapparentlyeasilylocatedlocaltopicsmentionedaboveandthemoretraditionalwayofdoingthingsbydealingwiththephilosophyofscienceofthegreatdeadphilosophersofthepast.Thisbe-comesevidentwhenoneattemptstoimagineacoherent“historyofnon-analyticalphilosophyofscience”or,perhapsbetter,a“historyofcontinentalphilosophyofscience”.Ifsuchahistoryexisted,itwouldbearathermixedbag.Froman“Americanperspective”theattitudeofconceivinganalyticalphi-losophyofscienceastheculminatingpointofthehistoryofphilosophyofsci-encemaybetemptingandappearsperhapsalmostnatural,butthisperspectiveiscertainlynotwithoutpresuppositions,andonemayaskwhetherfroma“Europeanperspective”theseshouldbetakenforgranted.Itleadstocertaindifficultiesevenifwerestrictourattentiontolocaltopicsofthehistoryofphilosophyofscienceas,say,philosophyofscienceintheGermanortheAustriantradition.Forinstance,takeSchlick’sempiriocriticismofAllgemeineErkenntnislehre.Itcanhardlybeclassifiedas“analytical”,butIdoubt,whethercharacterizingitas“pre-analytical”or“proto-analytical”isreallyfair.Asimilar,evenmoreimportantproblemarisesforNeokantianphilosophyofscience,inparticularCassirer’sasUebelcorrectlyremarks.AnalogousremarksholdforconventionalistphilosophyofscienceintheFrenchtraditionandothercurrentsofEuropeanphilosophyofscience.\n32ThomasMormannWilly-nilly,then,theanalyticalperspectiveoftentendstoplaydowntheprop-ervalueofothercurrentsofphilosophyofscience.Ittendstoignorethelossesphilosophyofsciencehassufferedonitswaytowarditsanalyticalrealization.Ihavenoquickrecipehowtoovercomethesedifficulties,theonlythingIwanttosayisthatthechoiceofthetopic„historyofanalyticalphilosophyofscience“isnotwithoutpresuppositions.Uebelseekstoavoidthesedifficulties.HerightlyobservesthatCassirer’sNeokantianphilosophyofscienceisimportantinitsownright:itisproblematictoconceivehisaccountsolelyfromtheanalyticalperspec-tiveaskinghowitinfluencedCarnaponhiswaytoanalyticalmaturity.Uebelattributesthegrowingimportanceofhistoryofphilosophyofscienceforphilosophyofsciencetoanaturalisticturninphilosophythatself-consciouslyrejectsanyapriorireflectionaboutgrandphilosophicalthemesrelatedtoscience.Thisiscertainlycorrect.Moreover,IthinkUebelisrightinassertingthatitisatrendtobeheartilywelcomed.Ageneralnaturalisticperspectiveseemstobelessfondofproducingintellectualfashions,necessarilyaccompaniedbycomplemen-taryblindspots,thanaprioristicaccounts.Inthiswaynaturalismmayseemanantidotetothethreatofamnesiamentionedinthebeginningofthisnote.3.HISTORYOFPHILOSOPHYOFSCIENCEASPHILOSOPHYOFSCIENCEBYOTHERMEANSIntheongoingprocessofnaturalization,which,ofcourse,notonlycomprisesthehistoricaldimension,butalsothesociological,thepsychologicalandotherones,alotofnewmeta-disciplinespopup:•Historyofphilosophyofscience•Psychologyofphilosophyofscience•Sociologyofhistoryofphilosophy•Historyofhistoryofscienceandsoon.Thismayleadustoconceive“Historyof…”,“Psychologyof…”assortofoperatorsanalogoustothemodaloperatorssuchas„possibly“,„necessar-ily“,„obligatory“,andsoonthatareusedinalethicordeonticmodallogics.Itera-tionsofsuchoperatorsmakesenseformally,butbecomemoreandmoreopaqueconceptually.Givenapropositionponemayformexpressionssuchas£◊◊por◊££◊pbutprobablyveryfewpeoplehaveanintuitiveideawhatthesemightmean.Anal-ogously,themeaningsofnewmetadisciplinessuchas„historyofphilosophyofsociology“,„philosophyofhistoryofsociology“tendtobecomeobscure.1Uebel1Actually,thingsaremorecomplicatedthantheoperatoranalogysuggests:forinstance,itisfarfromclearhowtherelationbetweenthe“historyofscience”and“philosophyofscience”istobeconceived.One,notoverlyconvincing,optionistounderstandthefirstas“descriptive”andthesecondas“normative”.Anotheroptionistoclaimanunspecified“complementarity”and“collaboration”ofsomekindbetweenthem.\nHistoryofPhilosophyofScience33putsforwardanimportantthesisthatcanbeusedtocutoffthisundesiredmul-tiplicationofpossiblemeta-disciplines:“Historyofphilosophyofscience…is…predominantlyphilosophyofscience.…Historyofphilosophyofscienceisphilosophyofsciencebyothermeans…”Thisthesisis,ofcourse,notnew.Invariousforms,ithasbeenbroughtfor-wardbymanyauthors,usuallynotrestrictedtophilosophyofscience,butclaimedtoholdforphilosophyingeneral.OnemightrecallthatalreadyWindelband,morethanonehundredyearsago,inhisinfluentialLehrbuchderGeschichtederPhilos-ophie(Windelband1889)ponderedontherelationbetweenphilosophyandhis-toryofphilosophy.Accordingtohim,historyofphilosophyshouldbeconsideredasanintegralpartofphilosophy,namelyasits„organon“(ibid.,567).2Neverthe-less,heinsistedthatbothdisciplinesshouldnotbemixedup,philosophersshouldnotforgetthephilosophicaloverthehistorical(ibid.,iii).Itisacommonplacethattraditionallythehistoricalwasastrongholdofcon-tinentalphilosophy,whiletheinterestofanalyticalphilosophyinhistoryofphi-losophywaslessfullydeveloped.Recently,thisstateofaffairsischangingasisevidenced,forinstance,bythemeanwhilewell-developedhistoryofthelogicalempiricismoftheViennaCircleandsimilarcurrentsthatareusuallypursuedfromananalyticalperspective.Evenifthereisarathergeneralconsensusamongphilosophersthathistoryofphilosophyshouldplayacertainroleforphilosophy,itisfarfromclearwhatpreciselythisroleistobe.3Forinstance,somecontendthatwehavetoknowthehistoryofphilosophy(ofscience)soasnottoavoidimportantalternativestocomtemporaryproposals(cf.Curley1986).ThisseemstobethestanceofUebel,Hardcastle,Richardson,andotherphilosophersofscience.Inthecaseofphiloso-phyingeneral,someauthorswanttogofurther,claimingthatphilosophyises-sentiallyanhistoricistendeavor(cf.Cohen1986).Howfarthisattitudemaybeapplicablealsotohistoryofphilosophyofscience,remainstobeinvestigated.Inanycase,ifwetakeintoaccountsomethinglikeUebel’sthesistheprofu-sionofmeta-disciplinesbecomeslessdisturbing.Doinghistoryofphilosophyofscience,orsociologyofphilosophyofscience,justmeansdoingphilosophyofscienceinspecificways.Conceivinghistoryofphilosophyofscienceasoneofthewaysofdoingphilosophyofscience,itisnaturaltoaskwhyweshouldpursueFurther,asisexemplifiedbythelogicalempiricismandothercurrentsof“scientific”philosophy,thereareinterestingrelationsbetween“historyofphilosophyofscience”and“historyofscientificphilosophy”thatrendertheagendaof“historyof(philosophyof)science”rathercomplicated(cf.Richardson2008).2Regrettably,Windelbandnowhereexplainedexactlywhatheunderstoodby“organon”here.Thereisnoreasontoexpectthattherolethathehadinmindforhistoryofphi-losophyresembledverymuchtothatwhichUebelisthinkingof.3Asitseems,therehasnotbeenmadetoomuchprogressinthisissuesincethetimesofWindelband.\n34ThomasMormannthehistoricalwayofdoingphilosophyofscience,andwhatachievementswecanexpectfromthisendeavor.Philosophersofsciencehavedealtwiththisquestionforsometimenow.RoughlyinlinewithUebel’snaturalisticturn,someyearsagoHardcastleandRi-chardsonspokeofa„historicistturn“inphilosophyofsciencethatmighthelptoovercomethecrisisthatplaguesphilosophyofscience.BythistheydidnotmeantheturninauguratedbyKuhn’sTheStructureofScientificRevolutions:Werefertoamorerecentdevelopmentinwhichphilosophershavebeguntorecovertheproblems,solutionsandmotivationsofearlierprojectsinthephilosophyofscience,payingattentiontohowthehistoricalfiguresengagedintheseprojectsunderstoodthem.…Adapt-ingwhatisperhapsthemostfamoussentenceinthephilosophyofscienceofthesecondhalfofthetwentiethcentury,wecanassertthatthehistoryofthephilosophyofscienceiscomingtobeviewedasmorethanarepositoryforanecdoteandchronology,andcan,ifweallowit,produceadecisivetransformationinthephilosophyofsciencewenowpossess.“(HardcastleandRichardson2003,vii).Tobeexplicit,forHardcastleandRichardsonthe„mostfamoussentenceinthephilosophyofscienceofthesecondhalfofthe20thcentury“isKuhn’sdictum:“History,ifviewedasarepositoryformorethananecdoteorchronology,couldproduceadecisivetransformationintheimageofsciencebywhichwenowarepossessed.”(Kuhn1962,1)InagreementwithUebel,then,HardcastleandRich-ardsoncontendthathistoryofphilosophyofscienceisphilosophyofsciencebyothermeans.Moreover,theyclaimthatphilosophyofscienceurgentlyneedsthisnewmeans.Accordingtothem,contemporaryphilosophyofscienceisentangledinadeepconceptual,almostexistentialcrisis,andhistoryofphilosophyofsciencemighthelpovercomeit.Indeed,theyinviteustotapthespiritualsourcesofthepast:[I]tmaywellbetimetoreturntothesocialspiritofphilosophyofscienceofthe1930.Per-hapsthatisourbestphilosophicalventureinaworldofanxioussocialandtechnologicalMaybes.“(ibid.,xxvi)Uebeldoesnotspeakofacrisis.Lessalarmistheistocontenttopointoutthatthehistoricalwayofdoingphilosophyofsciencemayhelpustorecoverconceptualresourcesthatwelostfromsight.Asherightlycontends,not„everything„logi-calpositivist“oreverything„transcendentalidealist“belongsinthedustbinofhistory.Inotherwords,heproposestousehistoryofphilosophyofscienceasasourceforexploringhithertoundevelopedorunderestimatedconceptualpossibili-ties.InthisrespectIfullyagreewithhim.Historyofphilosophyofsciencemightbeawayofdoingphilosophyofsciencethatovercomesthewide-spreadhistoricalamnesiareigninginmanyquartersofphilosophy,asalreadyAdornolamented.Ontheotherhand,historyiscertainlynotafoolproofwayofdoingphilosophyofsci-ence.ItmayleadusastraybyinvitingustoindulgeintheideaofanGoldenAge\nHistoryofPhilosophyofScience35ofPhilosophyofSciencewhenallegedlyourdisciplineflourishedineveryrespect.Duetohistoricalcontingencies,thisnarrativeclaimsthatlaterphilosophyofsci-encedeviatedfromtherightpathandendedupinthemorasswherewepresentlyfindit.UebeldoesnotconsiderhimselfasanaficionadooftheGoldenAge.4.THEBIPARTITEMETATHEORY:APATHTOBETAKEN?Eveniftheconceptionofhistoryofphilosophyofscienceasaspecialmeansfordoingphilosophyofscienceisnottheonlywayofpursuingit,forphilosophersofscienceitisperhapsthemostnaturalattitude.Uebel,apparentlyinlinewithRichardson,Hardcastle,Howard,andmanyothersconceivethereflexivedimen-sionthatisopenedupforphilosophyofsciencebypayingattentiontoitshistoryasameansformakingproposals.Heinvitesustoemploythehistoricistoptiontodophilosophyofscienceandtoexplore„pathsnottaken“.Moreprecisely,heproposestoreconsiderapaththatinthefirsthalfofthe20thcenturyallegedlywaspersuedbyCarnap,Neurath,andFrankbutlaterwasabandonedbytheirsuc-cessors.Thispathhedubsthepathofa„bipartitemetatheoryofscience“.Uebelconsidersitasapromisingstrategyforcontemporaryphilosophyofscience.This„bipartitemetatheory“hastwocomponents:–alogicalcomponentinthesenseofaCarnapianlogicofscience;–anempiricalpartroughlyinthesenseofNeurath’s„behavioristicsofscholars“orFrank’s„pragmaticsofscience“.UebelcontendsthatthisbipartitetheoryshouldbeconsideredasthecommonlegacyofthreegreatfiguresoftheViennaCircle,towit,Carnap,Neurath,andFrank.Imustconfess,thatIamskepticalabouttheprospectsofpursuingfur-therthe“bipartitemeta-theory”,evenifthegreatfiguresoftheGoldenAgeofphilosophyofsciencecouldbereadashavingsubscribedtoitasadesideratum,althoughintheirlatercareerstheydidnotundertakeseriouseffortstorealizeit.Itseemsnottotallyunreasonable,asUebeladmits,todoubtthatsuchatheoryasamoreorlesscoherentconceptualenterprisehaseveractuallyexisted.Tome,themerecoexistenceandallegedcomplementationofCarnapianlogicofscienceandFrank-Neurathianpragmaticsofscience,donotjustifytheclaimthattherewasakindoftheorythatcomprisedthelogicandpragmaticsastwosub-theories.Afterall,atheoryshouldhaveacertainamountoftheoreticalunity,whichthebipartitemetatheoryisclearlylacks.Uebelhimselfadmitsthatthetwohalvesofhisenvis-agedbipartitetheoryneithershareacommonmethodologynoracommonobject.InlinewithNeurath,forhimitsufficesthattherelationbetweenthelogicofsci-enceandthepragmaticsofsciencewasone–orcouldhavebeenone–of„coex-\n36ThomasMormannistenceandcomplementation“.4Inmyopinion,Uebelisratherindulgentwiththeoftenvagueprogrammaticalannouncementsofthefoundingfathersofclassicallogicalempiricism.5Forinstance,heemphasizesthatitwasanessentialfeatureofthemanifesto’sprogrammethatphilosophyofsciencebeacollectiveundertakingbasedonawellorganizeddivisionoflabour.Byhistoricalhindsight,however,theresultsofthisenvisaged“newway”ofdoingphilosophyofsciencehavebeenlessthantotallyconvincing.Afterall,uptonow,thescoreofreallycollectiveworkinphilosophyofsciencehasnotbeentooimpressive.Ingeneral,then,itisanimportanttaskforhistoryofphilosophyofsciencetoidentifytheillusionsandunfulfillabledreamsthatpastphilosopherscherished.Letuscomebacktotheissueofthebipartitemetatheory.Onemayreadilyadmitthatitis,ofcourse,alwayspossibletoweakentherequirementsatheoryhastosatisfyinsuchawaythatanyiuxtapositionofmoreorlessunrelatedtheoreti-calendeavorscountsasatheory.Thequestionis,whetheritisusefultoconceivesuchathingasatheory.Forthetimebeing,theexpression“bipartitemetatheory”seemstobememoreanameofaproblemratherthanaframeworkforatrulycomprehensivephilosophyofscience.Bethisasitmay,mynegativeassessmentshouldnotbeconsideredasanobjectiontoUebel’smainthesis,namely,thathis-toryofphilosophyofscienceisanexpedientmeansforexploringconceptualpossi-bilitiesforcontemporaryphilosophyofscience.HavingexpressedmydoubtsaboutthefeasibilityofUebel’sproposalofex-ploringfurtherthepaththebipartitemetatheorymaybeinterpretedassortofanobligationtomakemyselfaproposalofaninterestingpathnottakenthatwouldbeworthtobeexploredinthepresentsituation.Ithink,however,thatitwouldbemoreappropriatetothecontextofhistoryofphilosophyofsciencetopointoutthatUebel’sproposalofasynthesisoflogicalandpragmaticalcurrentsofphiloso-phyofsciencehadaninterestingprecursorsomesixtyorseventyyearsago.Thisattemptofsynthesisfailedforreasonsthatwedonotfullyunderstanduptonow.Thus,itmaybejustifiedtorescuethatunifyingattemptfromoblivion–nottheleastofthesereasonstheonethatthisexampleperhapscouldshedsomelightonthefeasibilityofUebel’sproposal.4Iconsiderthisconditionastooweakasthoughitcoulddistinguishclearlythesci-entificcomplementationandcollaborationofpartisansoftheprojectofabipartitemeta-theoryfromdubiousendeavorssuchastherecentfashionableattemptofcertainphilosopherstoconstrue“religion”and“reason”as“complementingandcooperatingelements”ofthemoderncondition.5Hesharesthisattitudewithmanyscholarsengagedinthehistoryofearlylogicalem-piricism.Thisattitudemightbeunderstandableinviewofthefactthatthe“true”his-toryoflogicalempiricismandrelatedcurrentshasbeenundulyneglectedforalongtime,andisstillneglectedinsomequartersofanalyticphilosophyperhapsevennow.Nevertheless,Iseecertaindangersinthisattitude.\nHistoryofPhilosophyofScience375.ACOMPREHENSIVEMETATHEORYTHATFAILED:MORRIS’SSCIENTIFICEMPIRICISMAlreadyinthethirtiesofthelastcenturyMorrishadurgedthelogicalempiricistsoftheViennaCircletothinkovertheirnarrowconceptofscientificphilosophyassyntaxofthelanguageofscience.AgainstthisoverlynarrowconceptionofphilosophyMorrisarguedforapragmatistscientificphilosophythatcomprisedfourdifferentstages:PaintingwithabroadbrushMorrisidentifiedfourrealmsofscientificphilosophizinglabelingthemwiththenamesofCarnap,Peirce,Dewey,andWhitehead(cf.Morris1937,8ff.)6•Philosophyaslogicofscience(Carnap)•Philosophyasclarificationofmeaning(Peirce)•Philosophyasempiricalaxiology(Dewey)•Philosophyasempiricalcosmology(Whitehead)Inthisschema,Carnap’spurelytheoreticalaccountofscientificphilosophyassyntaxofthelanguageofsciencefiguredasthefirstandmostrestrictedlevelofacomprehensivescientificphilosophywhichwhouldtakeintoaccountnotonlythelogicalbutotherdimensionsofofascientificcultureaswell(cf.Morris1937,8ff).Morrisreadilyadmittedthatmovingfrom“Carnap”to“Whitehead”amountedtoloweringthestandardsofexactnessandcertainty(ibid.,19).Buthewascon-vincedthatscientificphilosophyhadtopaythisprice,ifitwantedtoberelevantforlifeinacomprehensivemannerthattookintoaccounttheoryandpracticeofhumanexistence.Moreover,hegaveacompellingnaturalistreasonwhyitmightbeunscientificorevenunreasonabletoinsistonCarnapianstandardsofexactnessthroughout:Sciencerevealsnoabsolutebreakbetweentheoryandpractice,andthereisnoclearreasonwhythesituationshouldbedifferentinphilosophy.Meaningatthelevelofphilosophicalgeneralityhasitspragmaticdimensionjustahavethemeaningsatotherlevels....Itwouldbeasignalinstanceofethicalirresponsibility...toturntheworldovertotheexclusivecontrolofdreamers,adventurers,menofaction,andtechnicians.(ibid.,20).Carnapnevershowedmuchsympathywiththepragmatistunificationprogrammeevenifhedidnotmilitateagainstitexplicitly.Rather,hetriedtoeschewitinsomewayorother.InhisreplytoMorris’sproposal(intheSchilppvolume)attheendofthedayhehadnotmoretoofferthantheblandassertion:„Iaminclinedtoagree6Uebeldoesnotmentionpragmatistphilosophyofscienceinhislistof„movements(withinorfadingintotheanalyticaltradition)“,probablybecauseEuropeanphiloso-phershavenotcontributedmuchtoaphilosophyofsciencefromapragmatistpointofview.Bethisasitmay,historyofpragmatistphilosophyofsciencewouldcertainlyanimportantissueontheagendaofhistoryofphilosophyofscience.\n38ThomasMormannwithMorristhatthedifferencebetweenmyviewandthatofthepragmatistsisnotaslargeasitmightappearatfirstglance“(Carnap1963,862).Notwithstand-ingthisconciliaryassessmenthestucktohisethicalnon-cognitivismclingingtotheexistenceof„pureoptatives“andrefusingthethepragmatist„mean-endcontinuum“.DespiteCarnap’shalf-heartedconciliationtheconceptualdifferencesbetweenCarnapianlogicalempiricismandAmericanpragmatismofDewey,Lewis,andMorris(tosaynothingaboutPeirceandWhitehead)essentiallyremainedastheywere.Probablytheycanbeattributedtothequitedifferentconceptionsofscienceunderlyingthiscurrentsofscientificphilosophy.ForCarnap,sciencewasasystemoftheoreticalknowledge–asetofconsistentandrationallyjustifiablestatements(Carnap1935,32).ForDewey,totakehimasthemostoutspokenrepresentativeofagenuinepragmatistphilosophyofscience,sciencewasratheraprocessoractivity.Science,accordingtohim,wasnotknowledge,butaprocessforsolv-ingproblems.ThisentailedthatDeweyandtheotherpragmatistscontendedthatvaluationwasessentialtotheproductionofscientificknowledge,whereasCarnapinsistedonaradicalseparationbetweenknowledgeandvaluation.Thegapbe-tweenthesefundamentallydifferentphilosophicalperspectivesonsciencewasneverreallyovercomeasisshownbythedifficultandfinallyratherunsatisfyingcoexistencebetweenempiricistandpragmatistcurrentsofscientificphilosophyinthesecondhalfofthe20thcentury.Thefailureofconstructingacomprehensive„scientificempiricism“inMor-ris’ssenseshouldbetakenintoaccountwhenweseektoassessthechancesofUebel’s„bipartitemetatheory“,eveniftheparallelismbetweenUebel’sandMor-ris’sproposalsislimited.Atfirstlook,itmightbetemptingtoassociateNeurath’s„behaviorismofscholars“andFrank’s„pragmaticofscience“atleastgrossomodowiththepartthatMorrisreservedforthepragmatistsinhissketchyprogramofacomprehensivephilosophyofscience.ButIamnotsurehowfarthisgoes.IfNeurath’sandFrank’saccountscouldserveasapragmatic(orpragmatist?)complementationofCarnapianlogicofsciencesomehowanalogoustoclassicalAmericanpragmatismthatMorrishadenvisagedsomedecadesagothenUebel’sbipartitemetatheorywouldbeconfrontedwithsimilardifficultiesthatledtotheabandonmentofMorris’sprogram.7Ontheotherhand,ifitwouldturnoutthatthecomplementationenvisagedbyUebelwereofaquitedifferentkindthanthatwhichMorrisenvisaged,interestingproblemsconcerningtherelationbetweengenuineAmericanpragmatistphilosophyofscienceandtheVienneseersatzprag-matismofNeurathandFrankwouldarise.Inanycase,therearestillalotofissuesontheagendaofhistoryofanalyticalphilosophyofsciencethatdeservetobestudiedinthefuture.7Actually,IthinkthatNeurath’sandFrank’s“pragmatism”or“pragmatic”wouldmakearatherpoorsubstituteoftherealthing,butthisisnotanissuetobediscussedhere.\nHistoryofPhilosophyofScience39REFERENCESCarnap,R.,1935,PhilosophyandLogicalSyntax,London,Kegan,Trench,andTrubner.Carnap,R.,1963,“ReplytoMorris”,inP.A.Schilpp(ed.)ThePhilosophyofRu-dolfCarnap,LaSalle,OpenCourt.Cohen,L.,1986,“DoingPhilosophyisDoingitsHistory”,Synthese67,51–55.Curley,E.,1986,“DialogueswiththeDead”,Synthese67,33–49.Giere,R.,1973,“HistoryandPhilosophyofScience:IntimateRelationshiporMarriageofConvenience”,TheBritishJournalforthePhilosophyofScience24,282–297.Kitcher,P.,Salmon,W.,1989,ScientificExplanation,MinnesotaStudiesinthePhilosophyofScienceXIII,Minneapolis,UniversityofMinnesotaPress.Kuhn,T.S.,1962,TheStructureofScientificRevolutions,Chicago,ChicagoUni-versityPress.Morris,C.,1937,LogicalPositivism,Pragmatism,andScientificEmpiricism,Ac-tualitésscientifiquesetindustrielles449,Paris,Hermann.Morris,C.,1963,“PragmatismandLogicalEmpiricism”,inP.A.Schilpp(ed.),ThePhilosophyofRudolfCarnap,LaSalle,OpenCourt,87–98.Richardson,A.W.,Hardcastle,G.L.,2003,“LogicalEmpiricisminNorthAmer-ica”,inG.L.HardcastleandA.W.Richardson(eds.),LogicalEmpiricisminNorthAmerica,MinnesotaStudiesinthePhilosophyofScienceXVIII,Uni-versityofMinnesotaPress,Minneapolis,London,vii–xxix.Richardson,A.W.,2002.“EngineeringPhilosophyofScience:AmericanPragma-tismandLogicalEmpiricisminthe1930s”,PhilosophyofScience69,S36–S47.Richardson,A.W.,2008,“ScientificPhilosophyasaTopicforHistoryofScience”,Isis99,88–96.Uebel,T.,2007,EmpiricismattheCrossroads,LaSalleandChicago,OpenCourt.Windelband,W.,1892(1912,199318),LehrbuchderGeschichtederPhilosophie,J.C.B.Mohr(PaulSiebeck),Tübingen.DepartmentofLogicandPhilosophyofScienceUniversityoftheBasqueCountryUPV/EHUAvenidadeTolosa70200.80DonostiaSanSebastianSpainylxmomot@sf.ehu.es\n\nCRISTINACHIMISSOASPECTSOFCURRENTHISTORYOFPHILOSOPHYOFSCIENCEINTHEFRENCHTRADITIONFRENCHPHILOSOPHYOFSCIENCEAND‘MAINSTREAM’PHILOSOPHYOFSCIENCEWhenThomasUebelinvitedmetowriteapaperonthecurrentsituationofhis-toryofFrenchphilosophyofscience,ImustadmitthatIfoundthetaskalittledaunting.Idonotthinkthatitispossibletodojusticetothediverseresearchpro-grammesthatscholarsindifferentcountriesaredeveloping,ortopresentthemasacoherentwhole.Iwouldlike,however,tomakesomeremarksonthestateofthisparticularfieldofstudy,withtwoprovisos:oneisthatmyperspectiveissomewhatcentredinmyexperienceinGreatBritain,althoughitisnotlimitedtoit,theotheristhatIdonotaimatanoverallpresentationofthecurrentstateofthestudyofhistoryofFrenchphilosophyofscience.Inevitably,myremarkswillmainlyrefertothatpartofFrenchphilosophyofsciencethatistheobjectofmyownresearch.Iamconfident,however,thatAnastasiosBrennerinhiscommentarywillcorrectmynecessarilypartialpresentation.EspeciallyfromthepointofviewofsomebodyworkingintheEnglish-speak-ingworld,Frenchphilosophyofscienceappearstobeanareaofstudywithclearerboundariesthanothernationaltraditions.Thereseemstobeageneralunderstand-ingthatFrenchphilosophyofscienceisdifferentfrom‘mainstream’philosophyofscience:thisdifferencehasbeenmadeofficial,asitwere,inreferenceworksandEncyclopaedias.Inthis,theRoutledgeEncyclopediaofPhilosophyisparadig-matic:ithastwoentries,onefor‘PhilosophyofScience’,andanother,contributedbyGaryGutting,for‘Frenchphilosophyofscience’.FrenchphilosophyofscienceisnotperceivedasautonomousonlybyEng-lish-speakingphilosophers.Indeed,thesamedistinctionasthatoftheRoutledgeEncyclopediaofPhilosophyhasbeenproposedbyFrench-languagescholars.Do-miniqueLecourt,forinstance,inhisoverviewofthephilosophyofthesciences,haspresentedthisdisciplinefirstasalargelyAustrianandAnglo-Americanaffair(althoughAugusteComteispresentasafounderfather),andthenhasintroducedthe‘Frenchtraditionofphilosophicalreflectiononthesciences’asautonomousfromthetraditionoflogicalpositivismanditslegacy.LecourthasexplainedthatthedistinctiveidentityofthistraditionmainlyrestsonitsconstantlinkbetweenF.Stadler(ed.),ThePresentSituationinthePhilosophyofScience,ThePhilosophyofScienceinaEuropeanPerspective,DOI10.1007/978-90-481-9115-4_3,©SpringerScience+BusinessMediaB.V.2010\n42CristinaChimissohistoryandphilosophyofscience,andontherejectionofempiricismandofa‘certainlogicalformalism’.1Infact,thereislittledifferencebetweenGutting’sandLecourt’schoicesofillustriousnamesforthepantheonofFrenchphilosophyofscience,andbothplaceGastonBachelardandGeorgesCanguilhematthecentreoftheirpresentations.UnsurprisinglyforaphilosopherwhohaspromotedBachelard’sideasarguablylikenoother,Lecourthasdeclaredtheformertobethe‘emblematicfigure’ofthe‘Frenchtraditionofphilosophyofscience’,andhaspresentedGeorgesCan-guilhemasdevelopingBachelard’sphilosophy.InLecourt’saccount,FrançoisDagognetisthedirectinheritorofthistradition,whichforhimhasalsoproducedthinkerswhodonotsitcompletelycomfortablyundertheheadingofphilosophyofscience:MichelFoucaultandLouisAlthusser.2LikeLecourt,Guttinghasdedi-catedin-depthanalysesnotonlytoBachelardandCanguilhem,butalsotoMichelFoucaultandMichelSerres.3ThepresenceofFoucaultisparticularlyimportantforanEnglish-speakingreadership,whoismuchmorelikelytobefamiliarwithhiswritingsthanwiththoseofeitherBachelardorCanguilhem,tothisdaynotalltranslatedintoEnglish.4Indeed,somereaderswouldhaveheardofthembecauseofFoucault,notleastduetoGuttinghimself:hisbookonFoucaultopenswithachapterdedicatedtothesetwophilosophers.5BothLecourtandGuttingprovidebackgroundsforthemajorphilosophersofthe‘Frenchtradition’inphilosophyofscience:theformerpresentsasthe‘fathers’ofthistraditionCondorcet,AugustinCournotandAugusteComte,butmainlyfocusesonPierreDuhem,HenriPoincaré,EmileMeyerson,AbelRey,LéonBrunschvicgandAlexandreKoyré.GuttingintroducesthemainpartofhisarticlebysketchingahistoryofFrenchphilosophyofscienceinwhichthemaincharactersareDescartes,theEnlightenment,AugusteComte,PierreDuhem,EmileMeyersonandHenriPoincaré.However,thecentralityofBachelardand1DominiqueLecourt,Laphilosophiedessciences.Paris:PressesUniversitairesdeFrance2001,p.90.2Ibid,pp.113-4.3GaryGutting,“Frenchphilosophyofscience,”inCraig(Ed),RoutledgeEncyclopediaofPhilosophy.London:Routledge,RetrievedMarch09,2009,fromhttp://www.rep.routledge.com/article/Q0381998.4Canguilhem’sworksthathavenotasyetbeentranslatedintoEnglishincludeevenGeorgesCanguilhem,LaformationduconceptderéflexeauxXVIIeetXVIIIesiècles.Paris:PressesUniversitairesdeFrance1955;Bachelard’sbooksnotavailableinEng-lishinclude:GastonBachelard,L’activitérationalistedelaphysiquecontemporaine.Paris:PressesUniversitairesdeFrance1951;GastonBachelard,Lematérialismera-tionnel.Paris:PressesUniversitairesdeFrance1972[1953];GastonBachelard,Lerationalismeappliqué.Paris:PressesUniversitairesdeFrance1986[1949];GastonBachelard,Lepluralismecohérentdelachimiemoderne.Paris:Vrin1973[1932];GastonBachelard,L’intuitiondel’instant.Paris:Stock1992[1931].5GaryGutting,MichelFoucault’sArchaeologyofScientificReason.Cambridge:Cam-bridgeUniversityPress1989.\nAspectsofCurrentHistoryofPhilosophyofScience43CanguilhemishardtomissintheseasinotherpresentationsofFrenchphiloso-phyofscience.Indeed,BachelardandCanguilhem’s‘historicalepistemology’hascometobesynonymouswithFrenchphilosophyofscience.Thetrademarkofthesephilosophers,andwhatarguablymostsharplydistin-guishesthemfromtheirAnglo-Americancounterparts,istheimportanceofhis-toryfortheirphilosophy–andindeedofphilosophyfortheirhistory.Infact,LecourtalsocallsFrenchphilosophyofscience‘philosophicalhistoryofthesci-ences’.6Asaconsequence,oftenitspractitionershavebeenvariouslydesignatedasphilosophersorhistorians:PierreDuhemandAlexandreKoyréhavebeencon-sideredsometimeshistoriansandsometimesphilosophers;AbelReypublishedbothphilosophicalandhistoricalworks;7HélèneMetzgermaybechieflyknownforherworksonseventeenthandeighteenth-centurychemistry,butshealsowroteaphilosophicalbookonscientificconcepts,andmanyhistoriographicalpapers;8LéonBrunschvicgconsideredhimselfaphilosopher,butthisdidnotstophimbe-ingregardedbysome,includingGeorgeSarton,asahistorianofscience,suchwashisinterestinhistoryofscienceandhisattentiontothehistoricaldetail.9Indeed,Bachelard’sandCanguilhem’sdoctrineshavenotonlybeencalled‘historicalepistemology’,butalso‘epistemologicalhistory’,notablybyMichelFoucault;10DominiqueLecourthasdistinguishedbetweenBachelard’s‘historicalepistemol-ogy’andCanguilhem’s‘epistemologicalhistory’,adistinctionfurtherdeveloped6Lecourt,Laphilosophiedessciences,p.90.7AbelRey’sphilosophicalworkscomprisemonographs(e.g.AbelRey,Leretouréter-neletlaphilosophiedelaphysique.Paris:Flammarion1927),textbooks(AbelRey,Elémentsdephilosophiescientifiqueetmorale.Paris:Cornely1903)andarticles(e.g.AbelRey,“Surlepositivismeabsolu”,inRevuephilosophique34,no.681909,AbelRey,“Verslepositivismeabsolu”,inRevuephilosophique34,no.671909,AbelRey,“Pourleréalismedelascienceetdelaraison”,inRevuedemétaphysiqueetdemorale19,no.41911);hishistoricalworksincludeafive-volumehistoryofancientscience:AbelRey,Lasciencedansl’antiquité,vol.1:Lascienceorientaleavantlegrecs.Pa-ris:LaRenaissancedulivre1930;AbelRey,Lasciencedansl’antiquité,vol.2:Lajeunessedelasciencegreque.Paris:LaRenaissancedulivre1933;AbelRey,Lasciencedansl’antiquité,vol.3:LamaturitédelapenséescientifiqueenGrèce.Paris:LaRenaissancedulivre1939,AbelRey,Lasciencedansl’antiquité,vol.4:L’apogéedelasciencetechniquegreque:Lessciencesdelanatureetdel’homme,lesmathé-matiquesd’HippocrateàPlaton.Paris:LaRenaissancedulivre1946,AbelRey,Lasciencedansl’antiquité,vol.5:L’apogéedelasciencetechniquegreque:L’essordelamathématique.Paris:LaRenaissancedulivre1948.8HélèneMetzger,Lesconceptsscientifiques.Paris:Alcan1926;herhistoriographi-calpapershavebeenre-publishedin:HélèneMetzger,Laméthodephilosophiqueenhistoiredessciences.Textes1914-1939,réunisparGadFreudenthal.Paris:Fayard1987.9SeeBrunschvicg,lettertoSartonof2February1923(HoughtonLibrary,SartonPa-pers,bMSAm1803/1803.1).10MichelFoucault,TheArchaelogyofKnowledge.London:Tavistock1972[1969],p.190.\n44CristinaChimissobyJeanGayonandHans-JörgRheinberger.11Alongwithhistory,anotherresearchfocusofextremeimportanceformanyleadingFrenchphilosophersofsciencehasbeenthestudyofthemind.UNDERSTANDINGSCIENCEANDUNDERSTANDINGTHEMINDINHISTORYTheintegrationofhistoryandphilosophydidnotstartwithBachelardorCanguil-hem,althoughthesetwophilosophershaveoftenbeenpresented,orperceived,asinitiatorsofatradition,especiallyinEnglish-languagecriticism.12Fromthepointofviewofpureintellectualhistory,thephilosophicaltraditionofinterpretinghis-toricaldevelopmentwasobviouslynotnewtoFrenchphilosophy.Morespecifi-cally,thereexistedatraditionthatwasaimedatsketchingthehistoryofthemind,whichincludedsuchclassicmodelsoftheprogressofthemindinhistoryasCon-dorcet’s,Cournot’sandComte’s.Twentieth-centuryprojectsofstudyingthemindinhistory,whichwerecentraltoFrenchphilosophyofscience,certainlydidnotignorethoseillustriousmodels(BachelardevenrecalledthelawofthethreestagesatthebeginningofLaformationdel’espritscientifique,13withoutneedingtomen-tionComte).However,twentieth-centuryscholarsatthecentreofthereflectiononsciencetookhistorymoreseriously,insomecasesextremelymoreseriously.Moreover,mostofthemopposedpositivisticandmechanicisticviewsofhistoricalprogress.Inthisrespect,HélèneMetzger’sremarksonComteareemblematic:shearguedthathisnumerousexamplesfromthehistoryofscienceinevitably‘prove’hislawofthethreestagesofthedevelopmentofthemind,butonlybecausehepostulatesthislawasan‘inviolabledogma’.14Forher,insuchphilosophicalrep-resentationsofhistoryasComte’s,historicaleventsarechosenandinterpretedtoillustrateatheory,ratherthanbeingthebasisforthetheory.NotonlythehistorianMetzger,butalsophilosophers,notablyBrunschvicg,insistedthatthemindshouldratherbestudiedaposteriori.Twentieth-centuryscholarsdidcarefullyconsiderpreviousmodelsofthehistoryofthemind,butofteninapolemicalway.11DominiqueLecourt,Pourunecritiquedel’épistémologie.Paris:Maspero1972;JeanGayon,“TheConceptofIndividualityinCanguilhem’sPhilosophyofBiology”,inJournaloftheHistoryofBiology311998.;Hans-JörgRheinberger,“ReassessingtheHistoricalEpistemologyofGeorgesCanguilhem,”inGutting(Ed),ContinentalPhi-losophyofScience.Oxford:Blackwell2005.12ThereceptionofBachelardandCanguilheminEnglish-languagecriticismpresentedthemasthebeginningofatraditionthatcontinuedwithAlthusserandFoucault,asIshalldiscussbelowinthisarticle.AnexampleofthisisGutting’sbookonFoucaultcitedabove(note5).13GastonBachelard,Laformationdel’espritscientifique:contributionàunepsychana-lysedelaconnaissanceobjective.Paris:Vrin1993[1938],p.8.14HélèneMetzger,“Tribunaldel‘histoireetthéoriedelaconnaissancescientifique,”inGadFreudenthal(Ed),Laméthodephilosophiqueenhistoiredessciences,textes1914–1939.Paris:Fayard1987[1935],p.27.\nAspectsofCurrentHistoryofPhilosophyofScience45AsIhavediscussedatlengthinmybookWritingtheHistoryoftheMind,15thereisamuchmorehistorically-situatedstorytotellinordertounderstandthetwentieth-centuryprojectsthatareseenasthecoreofFrenchphilosophyofsci-ence,andthatmadepossiblethedevelopmentofhistoricalepistemology.Theseprojectsdevelopedinanintellectualandinstitutionalcontextthatmadeamean-ingfuldialoguebetweenhistoryandphilosophypossible.Theimportancethathistoryofphilosophycametoacquireinthefirsthalfofthetwentiethcentury,especiallyattheSorbonne,createdtheidealenvironmentforthedevelopmentofhistoricalepistemology.Theanalysisofprofessorships,courses,doctoraldisserta-tionsandpublications,aswellastheviewsofcontemporaries,alldemonstratethestrongdevelopmentofhistoryofphilosophyinhighereducationinthefirsthalfofthetwentiethcentury,inparticularaftertheFirstWorldWar.16Notonlydidhistoryofphilosophycometoberegardedasanimportantsubject,butitwasalsore-gardedasaphilosophicalsubject.Manyscholars,farfromconsideringhistoryasirrelevanttophilosophy,believedthathistorywas,inafortunateexpression,‘thelaboratoryofphilosophy’,thatistosaythedisciplinethatprovidestheempiri-caldatatophilosophy,andthatallowsphilosophicaldoctrinestobetested.Thisisparticularlytrueofthespecificresearchprogrammes,whichwereelaboratedduringtheThirdRepublic,whosequestionsBachelardandCanguilheminherited.ThehistoriansofphilosophyLucienLévy-BruhlandLéonBrunschvicgwerecen-traltotheseresearchprogrammes.LucienLévy-BruhlandLéonBrunschvicg,whobetweenthemoccupiedtheSorbonnechairofhistoryofmodernphilosophyforalmostthewholefirsthalfofthetwentiethcentury,17bothaimedtostudythemind.Theybothbelievedthatthemindwasnotfixed,butratherchangedindifferenttimesandplaces.Asaconsequence,itwasnotpossibleforthemtostudyitapriori,withoutrecourseto15CristinaChimisso,WritingtheHistoryoftheMind:PhilosophyandScienceinFrance,1900to1960s.Aldershot:Ashgate2008.16Forananalysisofthisdevelopmentintheearlytwentiethcentury,seeIbid.,Ch.1andJean-LouisFabiani,LesphilosophesdelaRépublique.Paris:EditionsdeMinuit1988.Nineteenth-centuryhistorywasratherdifferent(aboutitseeJohnI.Brooks,TheEc-lecticLegacy:AcademicPhilosophyandtheHumanSciencesinNineteenth-CenturyFrance.NewarkandLondon:UniversityofDelawarePressandAssociatedUniversityPress1998),anditisimportantnottoconfusethetwoperiods.17Lévy-BruhlwasappointedtotheSorbonnechairofhistoryofmodernphilosophyin1908andretiredin1927(hischairwasof‘histoiredelaphilosophiemoderne’,whichincludeswhatinEnglishisgenerallycalledhistoryofearlymodernphilosophy.Therewas,however,disagreementabouttheexactchronologicallimitsof‘philosophiemod-erne’).BrunschvicgreplacedLévy-Bruhlin1927andretiredin1940.Lévy-BruhlhadbeenappointedmaîtredeconférencesattheSorbonnein1899,andwasactiveafterhisretirement.BrunschvicgwasmadeprofessorEmeritusathisretirement,buthadtogointohiding,wherehediedin1944,astheGermanswereoccupyingFrance.Forthedetailsoftheirappointments,seeAlbertGuigue,LaFacultédesLettresdel’UniversitédeParisdepuissafondation(17mars1808)jusqu’au1erJanvier1935.Paris:Alcan1935.\n46CristinaChimissoempiricalresearch.Indeed,theyintendedtostudythemindaposteriori,thatistosaythroughdocumentsthatwouldshowthewayitworks.Theyhoweverelectedtoemploydifferentsources.Brunschvicgchosetoanalyseintellectualhistory,inparticularhistoryofphilosophyandhistoryofscience,whereasLévy-Bruhlturnedtothestudyofethnologists’reportsonthewayofthinkingofpeoplesinPapuaNewGuinea,AfricaandSouthAmerica.However,hehadalreadydevel-opedhisresearchaimsandmethodsinhisworksonhistoryofphilosophyandethics.Moreover,asheexplicitlyexplained,evenhisworkonprimitivementalitywasnotmeanttobeacontributiontoethnologybutrathertothestudyofhumannature;hisaimwastoinvestigatethetruthofHume’sandComte’sclaimsthathu-mannatureisuniversal.18TheimportanceofBrunschvicg’sandLévy-Bruhl’sdoctrinesfortheFrenchtraditioninphilosophyofsciencecannotbeoverstated.BrunschvicgwasBache-lard’smentorandsupervisorononeofhisdoctoraldissertations,19andsupportedhiminhiscareer.WhenBachelard’sdissertationswerepublished,Brunschvicg,whohadjustbeenappointedtotheprestigiouschairofhistoryofmodernphiloso-phyattheSorbonne,immediatelyreviewedtheminoneofthetwomajorphiloso-phyjournal,theRevuedemétaphysiqueetdemorale.Inoneofhisreviews,hesalutedhisformerstudentasa‘thinkerofthefirstorder’;20hisvalidationcouldnotfailtoproduceaprofoundimpressiononthephilosophicalestablishment.Moreimportantly,manyaspectsofBrunschvicg’sphilosophywerethestartingpointofBachelard’sphilosophy,includingtheaimtounderstandthemindbyexaminingintellectualhistory,theviewthatthemindchangesthroughhistory,andtheideathattheobjectsofknowledgearenotmind-independent.ItishardlysurprisingthatBachelard’sphilosophyhasbeenpresentedasanoriginaldevelopmentofBrunschvicg’s.2118JeanDuvignaud,Lelangageperdu.Essaisurladifférenceanthropologique.Paris:PressesUniversitairesdeFrance1973,p.126;LucienLévy-Bruhlandal.,“Lamen-talitéprimitive.Séancedu15février1923.”inBulletindelaSociétéfrançaisedePhilosophie231923.19GastonBachelard,Etudesurl’évolutiond’unproblèmedephysique:lapropagationthermiquedanslessolides.Paris:Vrin1973[1927].20LéonBrunschvicg,“Etudesurl’évolutiond’unproblèmedephysique.Lapropagationthermiquedanslessolides,parGastonBachelard”,inRevuephilosophique541929,p.94;LéonBrunschvicg,“Essaisurlaconnaissanceapprochée,parGastonBachelard”,inRevuephilosophique541929.21JeanWahl,Tableaudelaphilosophiefrançaise.Paris:Gallimard1962,p.114;GaryGutting,FrenchPhilosophyintheTwentiethCentury.Cambridge:CambridgeUni-versityPress2001,pp.85-6;FrançoisDagognet,“M.BrunschvicgetBachelard”,inRevuedemétaphysiqueetdemorale701965,pp.43-54;GaryGutting,“Introduction:WhatisContinentalPhilosophyofScience?,”inGutting(Ed),ContinentalPhilosophyofScience.Oxford:Blackwell2005,p.14.;JacquesGagey,GastonBachelardoulaconversional’imaginaire.Paris:Rivière1969,pp30,54;CarloVinti,Ilsoggettoqualunque:GastonBachelardfenomenologodellasoggettivitàepistemica.Napoli:Edizioniscientificheitaliane1997,pp.168,427–52;TeresaCastelão-Lawless,“Gas-\nAspectsofCurrentHistoryofPhilosophyofScience47Lévy-Bruhl’simpactonavarietyofdisciplineswasremarkable;hereitissufficienttonoticetheimportantrolethatthereceptionofhisworkplayedinthedoctrinesofphilosophersandhistoriansofscience.Justtomentionafew,HélèneMetzgerdevelopedherconceptsofmentalaprioriandexpansivethought,aswellashertheoryofactiveanalogy,withdirectreferencetoLévy-Bruhl’stheoryofprimitivementality;22AbelReyextensivelycitedLévy-Bruhlinhisdiscussionoftheoutillagemental,ormentaltool,inthefirstarticleoftheEncyclopédiefrançaise;23bothLéonBrunschvicgandGastonBachelardreferredtoLévy-Bruhl’stheory,acceptingsomeaspectsandrejectingothers,inordertodefinetheirownviewsofpastintellectualhistoryandtheirconceptionsofthemind.24Theuseofhistoryinordertoanswerphilosophicalquestions,andinparticularquestionsaboutthefunctioningofthemind,waswhatcharacterizedtheworknotonlyofphilosophers,butalsoofhistorians–firstofallthehistoriansofmental-ités,includingLucienFebvre,andthehistoriansofscience,includingAlexandreKoyré.However,thepresentdisciplinarydistinctionsdidnotholdintheinter-warperiodinFrance.Newdisciplines,includingethnology,25sociology,experimentalpsychologyandgeneralhistoryofthescienceshadstronglinkswithphilosophy,fromwhichtheyoriginated.Theirpractitioners,todifferentdegrees,didaimtodifferentiatetheirdisciplinesfromphilosophy,buttheykepttheirinstitutionalandintellectuallinkswithitverywellalive.Philosophers,sociologists,ethnologists,psychologists,historiansofscienceandothersdiscussedthemind,history,societyandsciencetogetherattheSociétéfrançaisedephilosophie,theCentredesyn-thèse,andinternationalconferences,andsharedstudentsandprojects.Thisdoesnotmeanthattheynecessarilyagreedwithoneother’sperspectivesandmethods,butevenwhentheydisagreed,theydidsoreflectively,referringtotheotherschol-ars’approaches.tonBachelardetlemilieuscientifiqueetintellectuelfrançais,”inPascalNouvel(Ed),ActualitéetpostéritésdeGastonBachelard.Paris:PressesUniversitairesdeFrance1997,pp.101-15.22HélèneMetzger,„LucienLévy-Bruhl,L’âmeprimitive”,inIsis91927,p.486,n.1.;HélèneMetzger,“L’aprioridansladoctrinescientifiqueetl’histoiredessciences”,inArcheion181936,p.37.;HélèneMetzger,“LaphilosophiedeLévy-Bruhletl’histoiredessciences”,inArcheion121930;HélèneMetzger,AttractionuniverselleetreligionnaturellechezquelquescommentateursanglaisdeNewton.Premièrepartie.Introduc-tionPhilosophique.Paris:Hermann1938.23AbelRey,“L‘évolutiondelapensée:Delapenséeprimitiveàlapenséeactuelle,“inFebvre(Ed),Encyclopédiefrançaise.Paris:1937.24LéonBrunschvicg,Lesétapesdelaphilosophiemathématique.Paris:Alcan1912;Ch.1;LéonBrunschvicg,“Nouvellesétudessurl’animeprimitive”,inRevuedesdeuxmondes521932;GastonBachelard,Lapsychanalysedufeu.Paris:Gallimard1949[1938].25Forthesakeofsimplicity,Itranslateethnologiewith‘ethnology’;however,ethnologieismorecorrectlytranslatedas‘culturalanthropology’.\n48CristinaChimissoItwouldbefartoolongheretodiscussthedisciplinary,institutionalandper-sonalnetworksofFrenchacademia,butitmaybeinterestingtorecallAbelRey’scareerasanexampleofthedisciplinaryfluiditythatwasstandardatthattimeinFrance.AbelReywasBachelard’sothersupervisor,andwasalsohispredeces-sorintheSorbonnechairofhistoryandphilosophyofthesciences.BeforehisSorbonneappointment,hehadbeenprofessorofphilosophyatDijonwherehefoundedthelaboratoryofexperimentalpsychology.AttheSorbonnehefoundedtheInstitutd’histoiredessciencesettechniques,and,outsideacademia,closelycollaboratedwiththehistoriansoftheCentredesynthèse.LucienFebvre,whogenerallyspeakingdidnotparticularlylikephilosophers,hadneverthelessaclosecollaborationwithRey;indeedheentrustedthelatterwiththefirstvolumeoftheEncyclopédiefrançaise,dedicatedtotheoutillagemental.Onceagain,thestudyofthemindthroughhistorywaswhatlinkedmanyofthesescholars.Frenchintel-lectualandinstitutionalhistorycreatedafertilesoilforthedevelopmentofadis-tinctivetraditioninphilosophyofscience,thatproducedhistoricalepistemology.HASHISTORICALEPISTEMOLOGYSTOLENTHESHOW?OTHERASPECTSOFFRENCHPHILOSOPHYOFSCIENCEThecentralityaccordedbyDominiqueLecourtandGaryGuttingtohistoricalepis-temologyintheirpresentationofFrenchphilosophyofscienceissharedbythelargemajorityofrepresentationsofthistradition,bothinsideandoutsideFrance.ThereislittledoubtthathistoricalepistemologyhasbeenthedominantimageofFrenchphilosophyofscience.Indeed,foralongtimeinFranceBachelard’sphilosophyhasrepresentedasortoforthodoxy.AsClaudeDebruhasputit,thecentralconceptofBachelard’sphilosophyofsciencebecamethe‘catechism’ofthephilosophyofscience‘madeinFrance’.26Thisimageisnowbeingchallengedaspartial.Moreover,historiansofphilosophyofsciencearepresentingworksthatshowthatitisnotcorrecttoseeacompleteseparation,indeedanopposition,be-tweenFrenchphilosophyofscienceontheonehand,andlogicalpositivism,itslegacy,andcurrentmainstreamphilosophyofscience,ontheother.AnexcellentexampleoftheseattemptsisAnastasiosBrenner’sLesoriginesfrançaisesdelaphilosophiedessciences.27BrennerexplicitlypointsouttwoassumptionsthatareoftenmadeconcerningFrenchphilosophyofscience:oneisthatitsstartingpointisthephilosophyofGastonBachelard,andthesecondisthattheFrenchtraditioninphilosophyofscienceisautonomousandirremediablydifferentfrommainstreamphilosophyofscience,logicalpositivismandpost-positivism.Hehasshownthat26ClaudeDebru,GeorgesCanguilhem,scienceetnon-science.Paris:Editionsrued’Ulm/Pressesdel’Ecolenormalesupérieure2004,p.67.27AnastasiosBrenner,Lesoriginesfrançaisesdelaphilosophiedessciences.Paris:PressesUniversitairesdeFrance2003.\nAspectsofCurrentHistoryofPhilosophyofScience49Frenchconventionalism,especiallyinPoincaré’sandDuhem’sversions,butalsointhoseofEdouardLeRoyandGastonMilhaud,playedanimportantroleintheformationofcurrentphilosophyofscience.Insodoing,Brenneraimstorevisethewidespreadviewthatcurrent(onecouldaddanalytical-oriented)philosophyofsciencehasitsrootsonlyinAustriaandinlogicalpositivism.ItisverywelcomethatscholarshavebeenworkingtowardsshowingthattheFrenchtraditioninphilosophyofsciencehasbeenfarricher,andmorecom-plexthanthestandardimagewouldallow.Forexample,thephilosophyofoneofBachelard’scriticaltargets,EmileMeyerson–the‘forgottenphilosopher’asbothJeanLargeaultandEvaTeklès-Kleinhavecalledhim28–isbeingbroughttotheattentionofscholarsagain,forinstanceintheprojectofBernadetteBen-saude-VincentandFrédéricFruteaudeLaclos.29SomescholarshavealsoturnedtheirattentiontoMeyersonoutsideFrance:amongthelattergrouptheyhavebeenintellectualhistoriansandhistoriansofphilosophy,suchasMarioBiagioliandMichaelHeidelbergerrespectively,butalsophilosophersofscienceintheanalyti-caltradition,suchasElieZaharandPeterLipton.30Theepistemologyofotherscholars,forinstanceofHélèneMetzger,hasbeenre-discovered;GadFreudenthalhasgivenatremendousimpulsetothestudyofherwork;manyothercriticshaveanalyzedherphilosophicalwork,includingtwomembersofourTeamE,MichaelHeidelbergerandI,andmanyothers,suchasIanGolinski,JohnChristie,IlanaLöwy,MichelBlay,ChristineBlondel,PietroRedondi,LuciaTosiandBernadetteBensaude-Vincent.31Metzgerhadnotbeenforgotten,butshehadbeenmainlyrememberedasahistorianofchemistry.28JeanLargeault,“EmileMeyerson,philosopheoublié”,inRevuephilosophique,no.31992;EvaTelkès-Klein,“EmileMeyerson:AGreatForgottenFigure”,inIyyun522003;seealsoEvaTelkès-Klein,“EmileMeyerson,d’aprèssacorrespondence.Unepremièreébauche”,inRevuedesynthèse5esérie2004.29SeeBernadetteBensaude-Vincent,“ChemistryintheFrenchTraditionofPhilosophyofScience:Duhem,Meyerson,MetzgerandBachelard”,inStudiesinHistoryandPhi-losophyofScience362005;FrédéricFruteaudeLaclos,‘Laphilosophiedel’intellectd’EmileMeyerson.Del’épistémologieàlapsychologie’(thesis:UniversitédeParis-XNanterre2004).FruteaudeLaclos’dissertationhasnowbeenpartlyturnedintoabook:FrédéricFruteaudeLaclos,L’épistémologied’EmileMeyerson.Uneanthropo-logiedelaconnaissanceParis:Vrin2009.SeealsoAnastasiosBrenner,“Lestatutdel’épistémologieselonMeyerson”,inArchivesdephilosophie70,no.32007.Iunder-standthatanumberofpublicationsonMeyerson,includingofvolumeofunpublishedprimarysources,iscomingoutin2009,andaconferenceisbeingorganizedtomarktheevent.30ElieZahar,“Meyerson’s‘RelativisticDeduction’:EinsteinVersusHegel”,inTheBrit-ishJournalfortheHistoryofScience381987;P.Lipton,“ExplanationintheSciences–Meyerson,E”,inAnnalsofScience51,no.21994.31SeeforinstancethearticlesinGadFreudenthal(Ed.),Etudessur/StudiesonHélèneMetzger.Leiden:Brill1990:J.R.R.Christie,“NarrativeandRhetoricinHélèneMetzger’sHistoriographyofEighteenthCenturyChemistry”;JanGolinski,“HélèneMetzgeretl’historiographiedelachimieduXVIIIesiècle;BernadetteBensaude-\n50CristinaChimissoBecauseoftheestablishedtraditionofstudyingthehistoryofphilosophy,thestudyofthehistoryofphilosophyofsciencehasalsobeenlively,andwecansaymainstreaminFrance.Moreover,althoughinFrancethereceptionofFrenchphilosophyofsciencehasgonewellbeyondphilosophy–itwouldbeenoughtorecallhowcentralmanyconceptsofhistoricalepistemologyhavebeentotheso-ciologyofPierreBourdieu–thestudyofitshistoryisaperfectlystandardfieldofstudyforphilosophers.Anumberofdistinguishedphilosophershaveworkedonit,includingDominiqueLecourt,FrançoisDagognet,JeanGayon,EtienneBalibar,BernadetteBensaude-Vincent,Jean-FrancoisBraunstein,MichelFichant,Anasta-siosBrenner,ClaudeDebru,GuillaumeLeBlancandDidierGil,justtomentionafew,innoparticularorder.32InItalytooseveralphilosophershaveworkedonVincent,“Unessaidevulgaritation:Lachimiedansl’Histoiredumonde”;GadFreu-denthal,“Epistémologiedessciencesdelanatureetherméneutiquedel’histoiredessciencesselonHélèneMetzger”;GadFreudenthal,“HélèneMetzger:Elémentsdebiographie”;GadFreudenthal,“Epistémologiedessciencesdelanatureetherméneu-tiquedel’histoiredessciencesselonHélèneMetzger”;IlanaLöwy,“Constructivistepistemologies:MetzgerandFleck”;MichelBlay,“LéonBlochetHélèneMetzger:Laquêtedelapenséenewtonienne”;ChristineBlondel,“HélèneMetzgeretlacristal-lographie:delapratiqued’unescienceàsonhistoire”;MartinCarrier,“SomeaspectsofHélèneMetzger’sphilosophyofscience”;PietroRedondi,“HenriBerr,HélèneMetzgeretAlexandreKoyré:lareligiond’HenriBerr”;inAgnèsBiard,DominiqueBourelandEricBrian(Eds.),HenriBerretlacultureduXXesiècle.Histoire,scienceetphilosophie.Paris:AlbinMichel/Centreinternationaldesynthèse1997;LuciaTosi,“HéleneMetzgerylahistoriadelaquímica”,inSaberyTiempo92000;CristinaChimisso,“HélèneMetzger:thehistoryofsciencebetweenthestudyofmentalitiesandtotalhistory”,inStudiesinHistoryandPhilosophyofSciencePartA32,no.2200;CristinaChimissoandGadFreudenthal,“AMindofHerOwn:HélèneMetzgertoEmileMeyerson,1933”,inIsis94,no.32003;Bensaude-Vincent,“ChemistryintheFrenchTraditionofPhilosophyofScience:Duhem,Meyerson,MetzgerandBa-chelard”,inStudiesinHistoryandPhilosophyofScience362005;GadFreudenthal,“HélèneMetzger(1888–1944)”,inBitbolandGayon(Ed.),L’épistémologiefrançaise,1830–1970.Paris:PressesUniversitairesdeFrance2006.32SomeofthesescholarshaveextensivelypublishedonthehistoryofFrenchphilosophyofthesciences,andIwillnotattempttogiveafulllistoftheirpublicationshere.Else-whereinthisarticleIcitesomeoftheworksonthissubjectbyLecourt,Debru,Bensau-de-VincentandBrenner;IhavealsocitedanarticlebyDagognetonBrunschvicgandBachelard,butIwouldliketoaddhereatleasttwootherworks:FrançoisDagognet,GastonBachelard:savie,sonœuvre.Paris:PressesUniversitairesdeFrance1965,andFrançoisDagognet,GeorgesCanguilhem:Philosophedelavie.LePlessis-Robinson:InstitutSynthélabo1997.BalibarisbetterknownforhisworksonAlthusser,buthehasalsopublishedonCanguilhem:seeEtienneBalibar,“ScienceetvéritédanslaphilosophiedeGeorgesCanguilhem,”inE.Balibaretal.(Ed.),GeorgesCanguilhem:Philosophe,historiendessciences.Actesducolloque(6-7-8décembre1990).Paris:AlbinMichel1993.SeealsoGuillaumeLeBlanc,Canguilhemetlesnormes.Paris:PressesUniversitairesdeFrance1998,GuillaimeLeBlanc,Laviehumaine:anthro-pologiesetbiologiechezGeorgesCanguilhem.Paris:PressesUniversitairesdeFrance2002.DidierGil,Bachelardetlaculturescientifique.Paris:PressesUniversitairesde\nAspectsofCurrentHistoryofPhilosophyofScience51thehistoryofFrenchphilosophyofscience,includingFrancescaBonicalzi,PietroRedondi,GasparePolizziandCarloVinti.33ThepresentdiscussionabouttheidentityofFrenchphilosophyofscienceispossiblethankstotraditionsthatvaluethestudyofthehistoryofphilosophy,includingasatooltostimulatefurtherphilosophicalresearch.Inthiscontext,thisdiscussionhasbothhistoricalandphilosophicalmeanings.THECURIOUSFATEOFFRENCHPHILOSOPHYOFSCIENCEINTHEENGLISH-SPEAKINGLANDSThefortunesofFrenchphilosophyofscienceinEnglish-speakingcountrieshavebeenuneventosaytheleast.Inthe1970s,AnglophonereadershipsdidshowaconsiderableinterestinBachelard’sandCanguilhem’sideas,whichtheyreceivedmainlybyreadingDominiqueLecourt’sworks.34ThroughLecourt,Bachelard’sFrance1993;MichelFichant,“L’épistémologieenFrance,”inChatelet(Ed),Histoiredelaphilosophie:le20esiècle.Paris:Hachette1973;J.-F.Braunstein,“CanguilhemavantCanguilhem”,inRevued’histoiredessciences53,no.12000,andJean-FrançoisBraunstein,“AbelReyetlesdébutsdel’Institutd’histoiredesscienceettechniques(1932–1940),”inM.BitbolandJ.Gayon(Ed.),L’épistémologiefrançaise,1830–1970.Paris:PressesUniversitairesdeFrance2006.ArticlesonCanguilhembyBraunstein,Lecourt,DebruandDelaporte(andIanHackingandArildUtaker)arecollectedinJean-FrançoisBraunstein,(Ed.),Canguilhem:Histoiredessciencesetpolitiqueduvivant.Paris:PressesUniversitairesdeFrance2007;infact,someoftheseauthorshavepublishedonFrenchphilosophyofscienceinothereditedvolumes,includingFrançoisBing,Jean-FrançoisBraunsteinandElisabethRoudinesco,(Eds.),ActualitédeGeorgesCanguilhem.Lenormaleetlepathologique.Paris:Synthélabo1998.andJeanGayonandJean-JacquesWunenburger,(Eds.),Bachelarddanslemonde.Paris:PressesUniversitairesdeFrance2000.33FrancescaBonicalzi,LeggereBachelard:leragionidelsapere.Milano:JacaBook2007;GasparePolizzi,Formedisapereeipotesiditraduzione;materialiperunasto-riadell’epistemologiafrancese.Milano:Angeli1984;GasparePolizzi,TraBachelardeSerres:aspettidell’epistemologiafrancesedelNovecento.Messina:ArmandoSicili-ano2003;PietroRedondi,Epistemologiaestoriadellascienza.LesvolteteoretichedaDuhemaBachelard.Milano:Feltrinelli1978,PietroRedondi,“Sciencemoderneethistoiredesmentalités.LarencontredeLucienFebvre,RobertLenobleetAlexandreKoyré”,inRevuedesynthèse1041983;PietroRedondiandP.V.Pillai,(Eds.),TheHistoryofScience:TheFrenchDebate.London:SangamBooks1989;PietroRedondi,“HenriBerr,HélèneMetzgeretAlexandreKoyré:lareligiond’HenriBerr,”inAgnèsBiard,DominiqueBourel,EricBrian,(Eds).HenriBerretlacultureduXXesiècle.Histoire,scienceetphilosophie,Paris,AlbinMichel/Centreinternationaldesynthèse1997;SabyasachiBhattacharyaandPietroRedondi,(Eds.),TechniquestoTechnology:AFrenchHistoriographyofTechnology.London:SangamBooks1990;CarloVinti,Ilsoggettoqualunque:GastonBachelardfenomenologodellasoggettivitàepistemica.34DominiqueLecourt,MarxismandEpistemology.Bachelard,CanguilhemandFoucault.London:NLB1975.ThisbookcomprisesthetranslationoftwoFrenchbooks:Lecourt,\n52CristinaChimissoandCanguilhem’sphilosophieswerereadinrelationtoAlthusser;withtheless-eningoftheinterestinAlthusser,theattentiontoBachelard’sandCanguilhem’sphilosophies,andIdaresaytoFrenchphilosophyofscience,alsowaned.Inthe1980s,MaryTilesmadeanexemplaryefforttorenderBachelardcomprehensibleandacceptabletoAnglo-Americanphilosophersofscience.35Despitethequalityofherwork,herefforthadalimitedeffect,apartfromsporadicarticles,likeforinstanceonebyMaryTijattaswhoattemptedtoreconcileBachelard’sideaswithaviewofscientificrealismcommonintheanalyticaltradition,andonebyDanMcArthur,whonineyearslaterrespondedtoher,36orrecentattemptstobringtogetheranalyticphilosophyofsciencewith‘Continental’philosophyofscience,suchasChristopherNorris’.37WhereasinFranceBachelard’sphilosophymayhavebeenasortof‘orthodoxy’,inEnglish-languagecountriesithasbeenre-gardedasanicheinterest.Thelittleattentionthat‘mainstream’philosophersofsciencehavepaidtotheFrenchtraditionislargelyduetoitsimage,dominatedbyhistoricalepistemology,whichsuggestsanintimateintegrationofhistoryandphilosophy.Intruth,thisim-ageonthewholeisnotmisleading.Althoughnotallphilosophersofscienceinthistraditionhaveputhistoryatthecoreoftheirdoctrines,itisundeniablethathistoryhasplayedamajorroleinFrenchphilosophyofscience,andnotonlyinhistoricalepistemology,butalsoinotherdoctrines.Evencriticswhoaimtoshowtherichtraditioninphilosophyofscienceoutsidehistoricalepistemology,neverthelessstresstheimportanceofhistoryforotherphilosophersofscience,asBrennerdoesinrelationtoDuhem.38ThiscentralityofhistoryhasbeenanobstacleforthereceptionofFrenchphi-losophyofsciencebyanalyticalphilosophers,whohavebeenbyandlargelittleinterestedinhistory.IdonotmeantoignoretheimpactthatFrenchphilosophyofsciencehashadinEnglish-languagephilosophyofscience.Theusethatphiloso-phersofsciencewritinginEnglishhavemadeofworksintheFrenchtraditionhasevenpromptedcriticslikeDenisVernanttoincludeunderthe‘historicalepiste-mology’headingnotonlyKoyré,BachelardandCanguilhem,butalsoPaulFeyer-Pourunecritiquedel’épistémologie,DominiqueLecourt,L’épistémologiehistoriquedeGastonBachelard.Paris:Vrin1969.35MaryTiles,Bachelard:ScienceandObjectivity.Cambridge:CambridgeUniversityPress1984.36M.Tijiattas,“BachelardandScientificRealism”,inPhilosophicalForum221991;DanMcArthur,“WhyBachelardisnotaScientificRealist”,inPhilosophicalForum33,no.22002.37ChristopherNorris,MindingtheGap:EpistemologyandPhilosophyofScienceintheTwoTraditions.Amherst,Mass.:UniversityofMassachusettsPress2000,ChristopherNorris,Epistemology.London:Continuum2005.38AnastasiosBrenner,“TheFrenchConnection:ConventionalismandtheViennaCir-cle,”inMichaelHeidelbergerandFriedrichStadler(Eds.),HistoryofPhilosophyofScience:NewTrendsandPerspectives.Dordrecht–Boston–London:KluwerAcademicPublishers2002.\nAspectsofCurrentHistoryofPhilosophyofScience53abendandThomasKuhn.39Itiswell-knownthatintheprefacetohisTheStructureofScientificRevolutions,KuhnacknowledgedtheimportancethatthewritingsofAlexandreKoyré,EmileMeyersonandHélèneMetzgerhadforthedevelopmentofhisownviewofscience.40ItisalsotruethatfromKuhnonwardstherehasbeenareception,althoughselective,ofFrenchphilosophyofscienceintoAnglo-Amer-icanphilosophyofscience,andakeenerattentiontohistory,suchasintheworksofIanHackingandGeraldHolton.41However,thereceptionofFrenchphilosophyofsciencehasnotbecomemainstream.Ifanyproofwerenecessary,itwouldsuf-ficetoseethelittlespacethattheFrenchtraditioninphilosophyofsciencefindsinmainstreamEnglish-languagepublicationsdedicatedtothephilosophyofsci-ence.Tomentionanexample,inthejournalPhilosophyofScienceintheyearsbetween1934and2008IcouldonlyfindonefullarticlededicatedtoBachelard,TeresaCastelão-Lawless’pieceonphenomenotechnique,42andnoneaboutCan-guilhem.Duhemfaresbetter,relativelyspeaking,asheismentionedinelevenarticles,althoughoftenwithintheexpression‘Duhem’sproblem’,withlittleornodirectreferencetoDuhemhimself.ThisofcoursedoesnotmeanthattherearenoacademicswhoaremembersofphilosophydepartmentsinEnglish-speakingcountriesandwhoatthesametimeworkonFrenchphilosophyofscienceanditshistory:theydoexist,andindeedIamoneofthem.However,thereisnoescapingthefactthatweareinasmallminority.Infact,thereceptionanduseofFrenchphilosophyofscienceintheEnglish-speakingworldseemstobestrongeroutsidephilosophyofscience.Forinstance,afterCanguilhem’sdeathin1995,severalvolumesandjournals’specialissuesonthisphilosopherwerepublished.InEnglish,itwasEconomyandSociety,asocialsciencesjournal,thatdedicatedadoubleissuetohim.43Theguesteditors,NikolasRoseandThomasOsborne,aresociologists,asareothersamongtheEng-lish-languagecontributorstothevolume,namelyMonicaGreco,LornaWeirand39D.Vernant,“Epistémologie,”inS.Auroux(Ed.),Encyclopédiephilosophiqueuni-verselle.Vol.2:Lesnotionsphilosophiques.Paris:PressesUniversitairesdeFrance1990.Onthistheme,seealsoA.Brenner,“Whichhistoricalepistemology?Kuhn,Fey-erabend,HackingandBachelard’sschool”,inRevuedemétaphysiqueetdemorale,no.12006.40ThomasS.Kuhn,TheStructureofScientificRevolutions.Chicago:TheUniversityofChicagoPress1996[1962],pp.vii-viii.41SeeforinstanceGeraldHolton,ThematicOriginsofScientificThought:KeplertoEinstein.Cambrige,Massachussets:HarvardUniversityPress1973;GeraldHolton,“EinsteinandtheCulturalRootsofModernScience”,inScienceinCulture127,no.11998;IanHacking,TheEmergenceofProbability.Cambridge:CambridgeUniversityPress1975;IanHacking,TheTamingofChance.Cambridge:CambridgeUniversityPress1990;IanHacking,HistoricalOntology.Cambridge,Mass.:HarvardUniversityPress2002.42TeresaCastelão-Lawless,“PhenomenotechniqueinHistoricalPerspective:ItsOriginsandImplicationsforPhilosophyofScience”,inPhilosophyofScience621995.43EconomyandSociety27,2/3(1998).\n54CristinaChimissoMikeGane;PaulRabinow,whoalsocontributedanarticle,isaculturalanthro-pologist.44TheonlycontributorbasedinanEnglish-languagephilosophydepart-mentappearedtobeIanHacking,wholatermovedtotheCollègedeFrance.Bycontrast,thearticlesbyFrancophoneauthors,althoughcomprisingapaperbyPierreBourdieu,weremainlybyphilosophers,includingDominiqueLecourt,AlainBadiouandFrançoisDelaporte.45AtleastinBritain,thefewscholarswhoworkonFrenchphilosophyofsciencehaveavarietyofbackgrounds,includingFrench,asinthecaseofMaryMcAllester.46SocialscientistshaveemployedinparticulartheworkofMichelFoucault,who,forthescholarofhistoryofphiloso-44AsisoftenthecaseinEnglish-speakingscholarship,heretheinterestinCanguilhemislinkedtoapreviousinterestintheworkofFoucault;indeedseveraloftheEng-lish-speakercontributorstotheEconomyandSociety’sspecialissueonCanguilhemhaveextensivelyworkedonFoucault;seeforinstance:HubertL.DreyfusandPaulRabinow,MichelFoucault:BeyondstructuralismandHermeneutics.Brighton:Har-vester1982;MichelFoucault,TheBirthoftheClinic:AnArchaeologyofMedicalPerception.trans.A.M.Sheridan,London:Routledge2003[1963];MichelFoucault,TheEssentialWorksofMichelFoucault,1954–1984.Vol.2,Aesthetics,editedbyPaulRabinow.London:Penguin2000;MichelFoucault,TheEssentialWorksofMichelFoucault,1954–1984.Vol.3,Power,editedbyPaulRabinow.London:Penguin2002;PaulRabinow,(Ed.),TheFoucaultReader.London:Penguin,1991[1986];AndrewBarry,ThomasOsborneandNikolasRose,(Eds.),FoucaultandPoliticalReason:Liberalism,Neo-liberalismandRationalitiesofGovernment.ChicagoandLondon:ChicagoUniversityPressandUCLPress1995.RabinowalsowrotetheintroductionoftheEnglish-languageanthologyofCanguilhem’swork:PaulRabinow,“Introduction:AVitalRationalist,”inF.Delaporte(Ed),AVitalRationalist:SelectedWritingsfromGeorgesCanguilhem.NewYork:ZoneBooks1994.45SeeT.OsborneandN.Rose,“Introduction”,inEconomyandSociety27,no.2-31998;andthearticlesinthesamedoubleissue:AlainBadiou,“IsthereaTheoryoftheSub-jectinCanguilhem?”;PierreBourdieu,“GeorgesCanguilhem:AnObituaryNotice”;GeorgesCanguilhem,“TheDeclineoftheIdeaofProgress”;FrançoisDelaporte,“Foucault,EpistemologyandHistory”;MonicaGreco,“BetweenSocialandOrganicNorms:ReadingCanguilhemand‘Somatisation’”;IanHacking,“CanguilhemamidtheCyborgs”;DominiqueLecourt,“GeorgesCanguilhemontheQuestionoftheIndi-vidual”;DavidMacey,“TheHonourofGeorgesCanguilhem”;PaulRabinow,“FrenchEnlightenment:TruthandLife”;NikolasRose,“Life,ReasonandHistory:ReadingGeorgesCanguilhem”;M.Gane,“CanguilhemandtheProblemofPathology”;C.Gordon,“Canguilhem:Life,HealthandDeath”;L.Weir,“CulturalIntertextsandSci-entificRationality:TheCaseofPregnancyUltrasound”.46MaryMcAllesterhasworkedonBachelard’sphilosophyasawhole,ratherthanonlyonhisphilosophyofscience(seeforinstanceMaryMcAllesterJones,GastonBachelard,SubversiveHumanist:TextsandReadings.Madison,Wisc.:UniversityofWisconsinPress1991;MaryMcAllesterJones,“Bachelard’sMetaphorsoftheSelf”,inFrenchStudies54,no.12000;MaryMcAllesterJones,“TheRedemptiveInstant–BachelardontheEpistemologicalandExistentialValueofSurprise”,inPhilosophyToday47,no.52003.);shehasalsoworkedonCanguilhem:MaryMcAllesterJones,“GeorgesCanguilhemonscienceandculture:learningbiology’slessons”,inFrenchCulturalStudies11,no.31,2000.\nAspectsofCurrentHistoryofPhilosophyofScience55phyofscience,istheinheritorofthetraditionofhistoricalepistemology.Withinthesocialsciences,Foucault’sworkhasgenerallynotbeenconsideredphilosophy,butrather‘theory’,thatisthetheoreticalpartofthesociologists’work.Foucaulthasalsobeenvindicatedforsciencestudies,asMartinKuschhasdone.47Thelargemajorityof‘Continentalphilosophers’donotfocusonphilosophyofsci-ence.However,thereareexceptions;forinstanceDavidWebb,whomainlyworksonMichelFoucaultandMichelSerres,andalsoonBachelardandCavaillès.48Ishallnotgoonwiththislist;mygeneralpointisthatFrenchphilosophyofsciencehasbeenreceivedacrossseveraldisciplines,butatthesametimehasplayedaminorrolewithinphilosophyofscience.Inadditiontothedifficultiesoftheencounterbetweentraditionsthathaveregardedsciencefromdifferentperspectives,anotherobstaclehasbeenthelittleattentionthatanalyticalphilosophers,whoarestillverydominantintheEnglish-speakingworld,onthewholepaytothehistoryoftheirowndiscipline.NotonlyhistoryofFrenchphilosophyofscience,buthistoryofphilosophyofscienceingeneralappearstobeaminorityinterest;indeedthereisaratherweakpresenceofhistoryofphilosophy,letaloneofhistoryofphilosophyofscience,intheuni-versitiesofEnglish-speakingcountries,especiallyintheUK.Thispresenceisparticularlyweakwithinphilosophy.AsIhavediscussed,historyofphilosophy,andintellectualhistoryingeneral,playedacrucialroleinthephilosophicaldebateinFrance,indeedinthedebateacrossmanydisciplines.Historyofphilosophy,notjusthistoryofphilosophyofscience,hasnotbeenasintegraltophilosophyintheEnglish-speakingworldasithasbeeninFrance,and,inadifferentway,inItaly.AtleastinBritain,itisnotunusualforphilosophydepartmentsnottohaveevenasinglehistorianofphilosophy.Moreover,historyofphilosophyhaslargelybeenhistoryofearlymodernphilosophy,asevenacursoryglanceatthepublicationsinjournalsdedicatedtothehistoryofphilosophywouldshow.Thisnarrowfocushasleftoutthebulkofphilosophyofscience.However,thesituationischanging.TheBritishJournalfortheHistoryofPhilosophyhasrecentlyappointedtwoassoci-atededitors,oneforancientphilosophy,andoneforhistoryofphilosophyfromKantonwards,becauseithasnowbeenrecognizedthathistoryofphilosophydoesincludethesetwoperiods.Thecreation,inthemid-1990s,ofHOPOS,thelearnedsocietyspecificallydedicatedtothestudyofthehistoryofthephilosophyofsci-ence,hasalsocreatedaninterestinginternationalcontextforthissubdiscipline.Iprobablypresentedaratherdividedimageofthecurrentsituationinthehis-toryofFrenchphilosophyofscience,oratleastofthatpartofthisfieldwithwhich47MartinKusch,Foucault’sStrataandFields:AnInvestigationintoArchaeologicalandGenealogicalScienceStudies.Dordrecht:Kluwer1991.48SeeforinstanceDavidWebb,“Microphysics–fromBachelardandSerrestoFoucault”,inAngelaki-JournaloftheTheoreticalHumanities10,no.22005;DavidWebb,“TheComplexityoftheInstant:Bachelard,Levinas,Lucretius,”inR.Durie(Ed.),TimeandtheInstant:EssaysinthePhysicsandthePhilosophyofTime.Manchester:ClinamenPress2000.\n56CristinaChimissoIammostfamiliar.IamafraidIhavealsopresentedasomewhatbleakimageofthestateofhealthofthissubdisciplineintheEnglish-speakingworld.Manyofushistoriansofphilosophyofsciencethinkthatthestudyofthehistoryofphilosophyofscienceshouldbepartofphilosophyofscience,oratleastveryrelevanttoit,butthisisnotauniversallysharedview,tosaytheleast.However,althoughIthinkthatthehistoryofFrenchphilosophyofsciencedoesnotreceivetheattentionitdeserveswithinphilosophyofscience,Ialsothinkthatthepresentsituationisopentochange,andindeedchanging,andthatprogrammesandinitiativeslikethepresentonewillhaveasignificantimpact.DepartmentofPhilosophyArtsFacultyTheOpenUniversityWaltonHallMiltonKeynesMK76AAUnitedKingdomc.chimisso@open.ac.uk\nANASTASIOSBRENNERREFLECTIONSONCHIMISSO:FRENCHPHILOSOPHYOFSCIENCEANDTHEHISTORICALMETHODINTRODUCTIONOverthepastseveralyearstherehasemergedacollectiveandconsciouseffortaimingtounderstandthehistoryofphilosophyofscience.ThishasledtotherenewedexaminationoftheViennaCircleandlogicalpositivism,whichiscon-sideredasoneofthemainsourcesofphilosophyofscienceintheEnglish-speak-ingworld.Yettherehavealsobeenattemptstoexplorethedevelopmentofotherschoolsofthought.AstudyoftheFrenchtraditionraisesseveralquestions,inparticularthereceptionofthistraditionanditssalientrecoursetoahistoricalap-proach.Attentionhasturnedfromthewell-knownBachelardianschooltoearlierphilosophers.CristinaChimissohasprovideduswithabroadandstimulatingpictureofFrenchphilosophyofscience.ShecandrawonherrecentbookWritingthehis-toryofthemind:philosophyandscienceinFrance,1900to1960s.Chimissotakesusbackpriortothosedoctrinesthatcontinuetopervadecurrentviews,thatispostpositivisminEnglish-speakinglandsandhistoricalespistemologyinFrench-speakingcountries.The1960smarkashift,andwhatliesbeforeisnowpartofhistory.Chimisso’sinterestsarenotmerelyantiquarian;sheleadsustophilosoph-icalissues.Inparticular,shedrawsourattentiontoLucienLévy-BruhlandLéonBrunschvicg,who,althoughprominentintheirtime,havelongbeenneglected.Chimissotherebypointstoworksthatarebeginningtoreceiveinterestagain.Sheincludesseveralotherthinkerswhoplayedaroleinthedevelopmentofphiloso-phyofscience,includingHenriBerr,AbelRey,HélèneMetzgerandAlexandreKoyré.Awholecommunitymakesitsre-appearance.BytakingusbackbeforeWorldWarTwo,Chimissodirectsustoatimeofintensephilosophicaldebate.Philosophyofscienceascarriedoutatthistimeappearshoweverquitedifferentfromwhatwepracticetodayunderthesameheading.Thishastheeffectofmak-ingussensitivetothehistoricaldimension:wemaymeasurethedistancecovered,evaluatethepersistentcoreofourdisciplineandscrutinizedthebackgroundwithrespecttowhichnewmethodsandthesesarose.Howtojustifyhistoricalstudy?Ibelievethatreturningtotheprimarysourcesalreadyprovidesananswer.Thepictureofearlierphilosophyofscienceasitwashandeddowntousbywayofretrospectivetestimonyorinthegeneralsurveysorintroductionstophilosophyofsciencedoesnotcorrespondtothehistoricalrecord.Chimisso’sstudy,Ibelieve,translatesanewsensitivity:theneedtopushfurtherasF.Stadler(ed.),ThePresentSituationinthePhilosophyofScience,ThePhilosophyofScienceinaEuropeanPerspective,DOI10.1007/978-90-481-9115-4_4,©SpringerScience+BusinessMediaB.V.2010\n58AnastasiosBrennerregardsourmethods,oursituationasobserversandtheevolutionoftheconstitu-tivenotionsofourdiscipline.Suchasensitivityhasbeentermedvariously:historyofphilosophyofscience,historicalsemanticsandmeta-epistemology.MyaimhereistoreflectuponChimisso’sresultsandtobringinfurthermate-rial.HowdidLévy-BruhlandBrunschvicgcontributetophilosophyofscience?Whatweretheirrelationswithotherscholarsworkinginthefield?Howtounder-standtheirmarkedlyhistoricalapproachwithrespecttotheapplicationoflogictophilosophythatcametodominateEnglish-languagephilosophy?Iwishalsotoinquireintothenatureofhistoricalmethodasputtophilosophicaluseaswellasthedifferencebetweenthephilosophicaltraditions.1.LÉVY-BRUHL,BRUNSCHVICGANDTHEAPOSTERIORIEXPLORATIONOFTHEMINDChimissodevotesherselftophilosophicalreflectiononscienceproducedduringthefirsthalfofthe20thcentury.OnecouldofcourseextendherinquiryfurtherbackintimetothefoundingfathersoftheFrenchtradition.AcompletehistorywouldcertainlyincludeAugusteComte.HisCoursdephilosophypositiveprovid-edanimpressivepictureoftheentirespectrumofthesciencesandinitiatedseveralmajortopicsofthisnewfieldofstudies,suchastheclassificationofscientificdis-ciplines,theroleofhypothesesandtheempiricalcriteriaofmeaning.1ComtesettheagendainseveralrespectsforphilosophyofscienceinFrance.Positivism,inoneformoranother,dominatedherethephilosophicalsceneuntilWorldWarOne,andevenlaterthinkerswhohadrelinquishedpositivismcontinuedtopaytributetohim,mostnotablyCanguilhem.FirstandforemostisComte’sdecisiontofavorahistoricalapproachoveralogicalone.Philosophyofscience,hecontinuallyasserts,mustbegroundedonhistoryofscience.ThistrendwastocharacterizeFrenchphilosophyofsciencegenerally.Asanattempttodirectphilosophicalre-flectiontowardscienceandtomakescientificknowledgeamodel,positivism,initsvariousformshasbeenintimatelyboundupwithalargeportionofphilosophyofscienceeitherasasourceofinspirationorasatargetforcriticism:fromComtianpositivismtologicalpositivismandeventopostpositivism.Itisthusimportanttocometogripswiththesignificanceandroleofthisdoctrine.Therewereothersignificantfiguresofthetime:André-MarieAmpère,Antoine-AugustinCournot,ClaudeBernardandCharlesRenouvier.Theyallmadesignificantcontributionstothephilosophyofscienceandhelpedtoshapetheearlystageofthefield.Myspaceishoweverstrictlymeasured,andIshallkeeptoChimisso’smainfocus.Lévy-Bruhlcameofagein1875andBrunschvicgadecadelater.TheybothdiedaroundthetimeofthesecondWorldWar.TheiractivelifespanswhatIshallcharacterizeasthesecondstageinthedevelopmentofphilosophyofscience.The1Thissix-volumeworkwaspublishedbetween1830and1842.\nReflectionsonChimisso59Franco-PrussianWarin1870notonlysignaledachangeofpoliticalsystem–theendoftheSecondEmpireandthebeginningoftheThirdRepublic–butledawholegenerationtoreflectonFrenchscienceandtoseektoemulatetheGermanuniversitysystem.Wemaymarkouthereafifty-yearperiodrunninguntiltheendofthefirstWorldWarin1918.Itischaracterizedbytheearlyinstitutionalizationofthediscipline.Thereafterfollowedtheinterwarperiod,whichrepresentsanewphase,thatofexpansionofthedisciplineanddevelopmentofareflectiononthelatestscientificdiscoveries.IshallthustakethestorybacktotheformativeyearsofLévy-Bruhlandthefactorsthatexplainthenewdeparturesoftheearly20thcen-tury.BachelardandKoyré,whostartedtheircarriersduringtheinterwarperiod,willbeconsideredhereonlyinsofarastheywereinfluencedbythetheoriesoftheirpredecessors;theirworkhasindeedreceivedagooddealofattention.TheimportanceofLévy-BruhlandBrunschvicgintheconstitutionofphiloso-phyofscienceinFranceisduetoseveralfacts.Theformerpavedthewayforthelatter:Lévy-BruhlhadstartedteachingattheSorbonnein1902andwaselectedtothechairofhistoryofmodernphilosophyin1908;hegaveanewdirectiontothediscipline,studyingphilosophersofthepastinrelationtothecontextoftheirepochinitsvariousaspects,withparticularemphasisonthescientificbackground.Tobesure,ÉmileBoutrouxhadalreadyinitiatedachangewithrespecttotheliteraryapproachcharacteristicoftheschoolofVictorCousin,whichhavebeeninfluentialuntilthen.2ButthearrivalattheSorbonneofLévy-BruhlfollowedbyBrunschvicgandMilhaud,allofwhominsistedonbringingsciencetobearonphilosophy,markedadecisiveshift.Lévy-Bruhlisresponsibleforhavingforgedthemodernnotionofmentality.3Thisnotionwastoplayacentralrolenotonlyinphilosophybutalsoinhistory;theFrenchhistoricalschoolinthe20thcanbecharacterizedinthemainbyitsrecoursetomentalities.ThisprovidestheinterpretativethreadofChimisso’sstudy,cen-teredonthe“historyofthemind”.Shunninglogic,Frenchphilosophyofsciencemadeextensiveuseinitsinvestigationsofthesocialsciences(sociology,anthro-pologyandpsychology),oftencombinedwithhistory.Anti-psychologismdidnothaveastrongholdinFrance,exceptingphenomenologists.ThisleadstoseveraldifferenceswithrespecttophilosophyofscienceinGerman-speakingcountries.Chimissosinglesoutseveralendeavorsthatarecloselyrelatedmethodologically:Theunderlyingassumptionsthatunitedtheseprojectswerethatthemindcouldnotbestudiedapriori,andthatwaysofthinkingweredifferentindifferentcivilizations.Asaconsequence,historywasasaruleanessentialcomponentofresearch.Pastphilosophyandpastsciencewereexpectedtorevealworldviewsandmentalprocessesthatdifferedfromcurrentones.42BoutrouxreplacedPaulJanet,adiscipleofVictorCousin,in1888.3ThisisnotunrelatedtoAugusteComte’snotionofmindoresprit,whichcorrespondstothethreestatesofhumanity:theological,metaphysicalandpositive.4Chimisso,2008,p.3.Cf.p.73,168.\n60AnastasiosBrennerIndeed,Lévy-BruhlandBrunschvicgelaboratedanaposteriorimethodofexplo-rationofthemind,basedonthehistoricaldocumentsthatityieldedinitsaimtounderstandtheworld.LetusnowturntoBrunschvicg.HehadbeenteachingattheSorbonnesince1905,andin1927hereplacedLévy-Bruhlinthechairofhistoryofmodernphi-losophy.Ashecametoelaboratehisphilosophicalposition,heacknowledgedhisdebttohispredecessor.5Therearemanyconnectionsbetweenthetwothinkers;theywereassociatedinmanynetworks,andtogethertheyrepresentastronglineofdevelopment.BrunschvicgwastoexertanascendancyoverFrenchphiloso-phy,establishingaparticularbrandofrationalismandidealismaswellasformingmanystudents.InparticularhewasBachelard’sdoctoralsupervisor,andChimissostressesthemanysimilaritiesoftheirphilosophies.62.THEMOMENT1900,SCIENTIFICREVOLUTIONSANDPHILOSOPHICALREFLECTIONIhavesuggestedthatseveralfactorsexplain“themoment1900”.7WeshouldnotforgetthatLévy-Bruhlbelongstothesamegenerationasanumberofotherimpor-tantfiguresforthephilosophyofscience:Poincaré,Duhem,MilhaudandMeyer-son.ItseemsthatLévy-Bruhl’sshiftfromarathertraditionalhistoryofphilosophytoanewapproachoccurredatatimewhenhewasworkingonhisbookonComte.8TheoriginalityofthisbookistodepartfromthehagiographicwritingsofComte’sdisciplesandtoprovideamoredistancedreading,bysettinghisdoctrinemorepreciselywithinitshistoricalcontext.Oneshouldnotforget,however,asynchro-nousattemptbyMilhaudtoevaluateComte’slegacy:Lepositivismeetleprogrèsdel’esprit:étudescritiquessurAugusteComte.9Milhaudlikewisewasproposingacriticalevaluationofthisthinker.BothLévy-BruhlandMilhaudneverthelessretainedsomethingoftheattitudethatComtehadinitiated.Furthermore,Lévy-Bruhl’sethnologyoranthropologyisnotwhollyunrelatedtoComte’ssociology,whichaimstodevelopa“positive”studyofhumankinddrawinglargelyonhis-tory.InasenseComte’spositivismgaverisetoseveralparalleldevelopments:so-ciology,anthropologyandhistoryofscience.Theseweretoreplacemetaphysics.PaulTannery’shistoryofsciencewasonesuchoutcome.Brunschvicgcouldcallonbothanthropologyandhistoryofscience.5See“L’idéedelavéritémathématique”,inBrunschvicg,1958,vol.3.6SeeChimisso,2008,p.141.7IamreferringheretotheconceptualisationgivenbyFrédéricWorms,Lemoment1900enphilosophie.8Lévy-Bruhl,Laphilosophied’AugusteComte,1900.9Thisworkwaspublishedin1902,butMilhaudalreadycriticizesComteinhisfirstbook,1893,p.205.\nReflectionsonChimisso61Itisworthtopointouttheunderlyingcontroversies;thesehelpedtoshapethemovementweareinterestedin.OnemaynotethatallourauthorsdevelopedtheirmethodsinoppositiontotheschoolofVictorCousin.MilhaudbringsthisoutclearlyinspeakingofPaulTannery’scontributiontothehistoryofphilosophy:YouknowwhatacademicphilosophywaslikeforalongtimeinFrance,ImeanthatkindofnaïveandbanalcatechismwhichtheschoolofCousinhadresultedin;andyouknowtowhatextentrhetoric,towhichwasgivenfreerein,hadinevitablydivorcedphilosophyfromscience.10Oneshouldnotomitthescientificfactorscomingintoplay.Asuccessionofrevo-lutionsinsciencehadtakenplacethatcalledforareworkingofthepictureofknowledge,insuccession:non-Euclidiangeometry,thetheoryofevolution,ther-modynamicsandelectromagnetism.OneoftheleadingfiguresofthetimewasHenriPoincaré.Hisresearchinmathematicsconvincedhimthatnon-Euclidiangeometrywasnotamerefictionbutafruitfulconceptualconstruction.Meditatingonthenatureofgeometricalhypotheses,Poincaréadvancedtheideathattheyareconventions.PierreDuhemformulatedasimilarideawithrespecttophysics.Hypothesesarenotdirectlyderivedfromexperience;theyarefoundedonthefreechoiceofthetheorist.Experimentalrefutationismorecomplexthanitwasgenerallybelieved.Thisledtotheholistthesis,whichNeurath,followedbyQuine,wastotakeupinthecontextofalogicalanalysisofscience.Thesestrikingresultswereseizeduponbyseveralphilosophersandscientists.ÉdouardLeRoyperceivedheretheriseofanintellectualmovementthathelabeled“anewpositivism”.GastonMilhaudwentsofarastospeakoflogicalpositivismorpositivismelogiqueasearlyas1905.11ThisreformulationofpositivismattractedtheattentionofyoungAustrianscholarswhoweretofoundtheViennaCircleandprovidesuswithanoteworthyconnectionbetweenthephilosophicaltraditionsofFranceandAustria.LeRoyemphasizedthenoveltyofthesereflectionsonscience;hewasoneofthefirsttomakeuseofthetermépistémologieorepistemology.Thetermdes-ignatesinFrenchusagephilosophyofscienceratherthantheoryofknowledge.Whatwasbeingproposedwasaninvestigationpreciselycenteredonscientificactivity.Thiscarriedanimplicitcriticismofearlierphilosophyofscience,asprac-ticedbyComte,andsignaledashiftinthediscipline.Inconnectionwiththesedebatesoverthenatureofscientifictheoriesearlyattemptsweremadetointroducephilosophyofscienceintotheuniversitycurricu-lum.In1892achairof“Generalhistoryofscience”wasinstitutedattheCollègedeFrance.In1909,achairof“Historyofphilosophyinitsrelationtoscience”wascreatedforMilhaudattheSorbonne.Hethuscametoworkinthesameuniversity10Milhaud,1911,p.2.SimilarcriticismisvoicedbyBrunschvicginthesecondeditionofhisthesis.ApointalsomadebyBougléasquotedbyChimisso,2008,p.73.11Milhaud,1927,p.55,reproducinganarticlepublishedin1905.\n62AnastasiosBrennerasLévy-BruhlandBrunschvicg.Milhaud’schairwastoplayapivotalroleinthefutureofthefield,beingheldsuccessivelybyAbelRey,BachelardandCanguil-hem.12TakingupPoincaré’sideas,AbelReywascarefultoemphasizethetendencytowardrealism.Hewasinparticularstruckbytherecentdiscoveriesofatomictheory,andwasledtoelaborateahistoricalapproachemployingtechniquesde-velopedinthesocialsciences.Histhesis,asyntheticpresentationoftheturn-of-the-centurydebates,wasseizeduponbythelogicalpositivists.AbelReywasfur-thermoreincludedamongNeurath’scollaboratorstotheEncyclopediaofUnifiedScience.However,thispromisingconnectionbetweenFrenchconventionalismandAustrianpositivismwascutshort13.Bachelard,whosucceededtoReyin1940,canbecreditedwithhavingforce-fullydirectedphilosophicalattentiontothelatestscientifictheories.AlongwithAlexandreKoyré,hewasconvincedthatthesuccessionofrevolutionsthathadshakensciencesincethediscoveryofnon-Euclidiangeometrycalledfora“philo-sophicalrevolution”.BorrowingaphrasefromReichenbach,Bachelardspokeofa“conflictofgenerations”,andhewasquicklyledtospellouttheinadequaciesofthephilosophicalconceptionsofhispredecessors.Thuswasbroughttoacloseaparticularphaseinthedevelopmentofphilosophyofscience.Howeverthehisto-rianmayquestionthisportrayalandseekdeeperlinksandtransmissions.3.ONHISTORICALMETHODWhatcharacterizesalargeportionofFrenchphilosophyofscienceistheimpor-tanceallottedtohistory.ThisisapparentintheearlyformulationofthedisciplinebyComteaswellasitslaterinstitutionalestablishment.Ofcourse,ahistoricalapproachcanbepursuedinmanyways.Onedirectionconsistsingroundingphi-losophyofscienceuponthehistoryofscience.Intheabsenceofempiricaltest-ing,historyofscienceprovidesameansofassessingphilosophicalconceptionsofscience;itprovidestheanalogueofthelaboratory14.ThisisparticularlyclearinDuhem.HisAimandStructureofPhysicalTheoryfurnishedananalysisofthestagesinvolvedintheconstructionofascientifictheory.Butthis“logicalanaly-sis”15,ashetermedit,wastobefollowedbyahistoricalstudy,andthenumerousvolumeshedevotedtotheevolutionofsciencesinceAntiquitybearwitnesstothispreoccupation.Suchamethodwasfollowedbymanyofhiscontemporaries,forexampleMeyerson.Postpositivistswerelatertocallonthistraditionintheireffort12ConcerningthefiliationbetweenTannery,MilhaudandRey,seeBrenner,2005.13Contingenthistoricalfactorsenterhere.14ThismetaphorusedbyBrunschvicgisquotedbyChimisso,2008,p.73.Cf.p.168.15Duhem,1906,p.XV.Cf.Duhem,1913,p.115.\nReflectionsonChimisso63toreasserttheimportanceofhistory,andthiswasoneofthetrendsoftheFrenchtraditionthatreceivedthemostsustainedinterestabroad.“Historyofphilosophyinitsrelationtothesciences”,tousethetitleofthechaircreatedforMilhaud,constitutesanothersignificantlineofresearch.Inin-troducingphilosophyofsciencewithintheuniversitycurriculum,Milhaudwascarefultolinkthisspecialitywiththehistoryofphilosophy,whichoccupiedanimportantroleinFrance.BrunschvicgandReyfitwellintothisprogram16.Suchaninterdisciplinaryapproachallowedforvariouscollaborationsandmanytopicsofinquiry.ItcharacterizestheinstitutionalsituationinFranceandmarksadiffer-encewithrespectforexampletophilosophyinGreatBritain.InasenseBachelardandKoyrébuiltontheseantecedentefforts,theformerinthedirectionofahistoricalphilosophyofscienceandthelatterinthesenseofaphilosophicalhistoryofscience.Buttheygaveanewtwisttothisapproach.Bothhadmisgivingsoverearlierconceptionsofscientificgrowthasacontinuousproc-ess.Theysetabouttoelaboratewhathasbeennameda“historicalepistemology”.Studyofpastsciencestillretaineditsimportance.Butitwastobeplacedwithinaclearlydiscontinuistconception,inspiredbytherecentdiscoveriesinscience.Scientificrevolutionsareaccompaniedbybreaksbetweencommonknowledgeandscientificknowledge.Bachelardespeciallymadeexplicitthepositionfromwhichthephilosopherobservesthepast:readingisnecessarilyretrospectiveorrécurrent.OnemaycallhereonIanHacking,whothrowslightonthisissue.HebringsoutclearlythedifferencebetweenBachelardandFoucault,inotherwordstheevolutionundergonebyhistoricalmethod.HehimselftakesFoucault’shistoricalepistemologyorhistoricalontologyastepfurtherandgivesexpressiontoawholetrendofresearchbeingdonetoday.Althougheducatedintheanalytictradition,HackingdoesnothesitatetocallonFrenchhistoryandphilosophyofscience.Foucault,enlargingonBachelard’sperspective,hadmadeabroaderandmoresystematicuseofhistory,whichheinduecoursenamed“archeologyofknowl-edge”or“historicalontology”.Hackingtakesupthisapproach,applyingitmorespecificallytophilosophyofscience.Inparticularhegivesaconcretemeaningtotheattempttorelatediscoursetoitscontextofformulation.AndHackingoffersacarefulanalysisofthesitesofproductionofexperimentalscience:thelaborato-ries,theobservatoriesandtheresearchcenters.Heclaimsthatitisquitepossibletorecovertherebytheconcernsofanalyticphilosophy.Historicalontologyisjustanotherwayofpursuinganalysis:thecon-ceptualusagesarereferredchronologicallytotheirsiteofenunciation.Thisishowhepresentshisprogram:16AlthoughReyobtainedthechangeofthischairto“Historyandphilosophyofsci-ence”,heneverthelessadmittedtopursuingthepathopenedbyMilhaud.Formore,seeBrenner,2005.\n64AnastasiosBrennerHistoricalontologyisaboutthewaysinwhichthepossibilitiesforchoice,andforbeing,ariseinhistory.Itisnottobepracticedintermsofgrandabstractions,butintermsoftheexplicitformulationsinwhichwecanconstituteourselves,formulationswhosetrajectoriescanbeplottedasclearlyasthoseoftraumaorchilddevelopment,or,atoneremove,thatcanbetracedmoreobscurelybylargerorganizingconceptssuchasobjectivityorevenfactsthemselves17.Onecanthensubmittheconstitutivenotionsofsciencetoahistoricalanalysis,recordingthediscursiveformulationsandmappingouttheirdevelopment.CONCLUSIONTheperiod1870–1920thatIhavesingledoutforexaminationisverydifferentfromthefoundingyearsofphilosophyofscience;manynewobjectsofinquiryarose,andtheanalysisofscientificknowledgeprovidedwasrich,originalandfruitful.Itisworthwhiletoreturntothisepochinordertosharpenourtoolsandtoenlargeourlistofproblems.Furthermore,acompletepictureofphilosophyofsciencerequiresustounderstandthetransformationthatbroughtaboutthecon-ceptionsofthemid-twentiethcentury.Ibelievethatonewaytomoveaheadistobeclearastotheobjects,methodsandaimsofourinquiry.Itisessentialthatweplotthetrajectoriesofthetoolsofourtrade.InthepasttwentyyearsseveralFrenchphilosophersofscienceoftheperiodpriortotheSecondWorldWarhavebecometheobjectofamorethoroughandsystematicinvestigation:firstDuhemandPoincaré,thenMeyersonandMetzger.ChimissohasconvincinglyarguedinfavorofaddingtoourlistLévy-BruhlandBrunschvicg.Wenowhaveawholeseriesofphilosophers,amongwhomthecon-nectionsarenumerous.Historicalresearchhasnotonlyfocusedonindividuals;workiscurrentlybeendirectedtowardthecontentofjournalsaswellassocietiesandinstitutions.Networksofrelationsamongscholarsarebeingextensivelyex-plored.Inconsequence,ourpictureofthefieldandourunderstandingofthenatureofphilosophyofscienceisbeingdeeplymodified.BIBLIOGRAPHYBachelardGaston(1934),Lenouvelespritscientifique,Paris,PUF,1971.—(1972),L’engagementrationaliste,Paris,PUF.BitbolMichelandJeanGayon(2006)(eds),L’épistémologiefrançaise:1830-1970,Paris,PUF.BrunschvicgLéon(1958),Écritsphilosophiques,Paris,PUF.17Hacking,2002,p.23.\nReflectionsonChimisso65BrennerAnastasios(2003),Lesoriginesfrançaisesdelaphilosophiedessciences,Paris,PUF.—(2005),“Réconcilierlessciencesetleslettres:lerôledel’histoiredessciencesselonPaulTannery,GastonMilhaudetAbelRey”,Revued’histoiredessci-ences,vol.58,p.433-454.BrennerAnastasiosandJeanGayon(2009)(eds),FrenchStudiesinthePhiloso-phyofScience,Vienna,Springer.CanguilhemGeorges(1966).Lenormaletlepathologique,Paris,PUF.—(1968),Étudesd’histoireetdephilosophiedessciences,Paris,Vrin,1970.ChimissoCristina(2008),WritingtheHistoryoftheMind:PhilosophyandSci-enceinFrance,1900to1960s,Aldershot,Ashgate.ComteAuguste(1830–1842),Coursdephilosophiepositive,M.Serres,F.Da-gognetA.SinaceurandJ.P.Enthoven(eds),newed.revised,2vols,Paris,Hermann,1998.DuhemPierre(1906),LaThéoriephysique,sonobjet,sastructure,Paris,Vrin,1981.—(1913),Noticesurlestitresettravauxscientifiques,Bordeaux,Gounouilhou.Hacking(2002),HistoricalOntology,Cambridge,Mass.,HarvardUniversityPress.KoyréAlexandre(1966),Étudesd’histoiredelapenséescientifique,Paris,Gal-limard,1973.LecourtDominique(2008),GeorgesCanguilhem,Paris,PUF.LeRoyÉdouard(1901),“Unpositivismenouveau”,Revuedemétaphysiqueetdemorale,9,p.138-153.Lévy-BruhlLucien(1900),Laphilosophied’AugusteComte,Paris,Alcan.MeyersonÉmile(2009),Lettresfrançaises,B.Bensaude-VincentandEvaTelkesKlein(eds),Paris,CNRSÉditions.MilhaudGaston(1893),Leçonssurlesoriginsdelasciencegrecque,Paris,Al-can.—(1902),Lepositivismeetleprogrèsdel’esprit:ÉtudescritiquessurAugusteComte,Paris,Alcan.—(1911),Nouvellesetudessurl’histoiredelapenséescientifique,Paris,Alcan.—(1927),LaphilosophiedeCharlesRenouvier,Paris,Vrin.PoincaréHenri(1902),LaScienceetl’hypothèse,Paris,Flammarion,1968.WagnerPierre(2002)(ed.),Lesphilosophesetlascience,Paris,Gallimard,2002.WormsFrédéric(2005)(ed.),Lemoment1900enphilosophie,Lille,Septentrion.UniversitéPaulValéry-MontpellierIIIRoutedeMende34199MontpellierCedex,Franceanastasios.brenner@wanadoo.fr\n\nMICHAELHEIDELBERGERATHSPECTSOFCURRENTHISTORYOF19CENTURYPHILOSOPHYOFSCIENCE“Throughhistory,philosophyseeks,initspast,itseternalpresent.”(Bréhier1940,44)Theattitudeofphilosophytowardsitsownhistoryandhistoriographyhasbeensubjecttochangeovertheages.Periodsofdeepinterestinthehistoryofphi-losophyandperiodswherethisinterestfadesintothebackgroundalternateinthecourseoftime.Onereasonforthisvariationmightbeanaturallife-cycleinthesuccessionofgenerations:theyoungmustbeiconoclastsanddestroytheirheritagetoacertainextentinordertodevelopanewandoriginalone.Whentheygrowold,however,theysoonerorlaterreachthepointwheretheystartwonder-ingwhethertheirpromiseshavebeenkeptandwhethertheirownachievementsreallyfulfilwhattheyintendedatthestart–andanewinterestintherecentpastsetsinagain.Ithinkthat,ifappearancesarenotdeceptive,wearewitnessingtodayare-newedinterestinthehistoryofthephilosophyofscience,andespeciallyofthe19thcentury.Thisinterestmightbetosomeextenttheresultofthelife-cyclejustdescribed,butthereseemtobeadditionalanddeepermotivesforit.Inthefol-lowing,Iwouldliketoaskwhatthesedeeperreasonsmightbe,wherewestandtodayinrelationto19thcenturyphilosophyofscience,andhowwecanfruitfullydevelopthisinterestfurtherandgainbetterinsightintotheworkofourforebearsand,withthis,intoourownsituation.Theinterestinabygoneperiodofphilosophicalthoughthasfirstofallanemotionaldimension:Ontheonehandonecanseethepastwithsomewistfulnessasanagethatalreadyrealizedwhatonewishesforone’sownpresentandfuture,andwhatonelamentsasbeinglost,maybeevenforever.Inthissensethephilo-sophicalpastcanappearasagoldenageandserveasaprojectionscreenofhopesandidealsforthefuture,butalsoasadreamlandinwhichonecanseekrefugefromthecruelpresent.Ontheotherhand,thepastcan,ofcourse,alsoappearasabarrenagewhoselimitationsandmistakescanandmustbeovercomebyanewphilosophicalspiritthatisimmunetothefailuresofthepast.Thereisalsothethirdpositionthat,withouttakingsides,sotospeak,usestheformerageasamirror,inwhichthepresentcanbereflectedandreappraised.Thiscanleadtoabetterunder-standingofone’sownsituationandtoastrengtheningofone’sownidentity.F.Stadler(ed.),ThePresentSituationinthePhilosophyofScience,ThePhilosophyofScienceinaEuropeanPerspective,DOI10.1007/978-90-481-9115-4_5,©SpringerScience+BusinessMediaB.V.2010\n68MichaelHeidelbergerIthinkthatourattitudetodaytowardsthephilosophyofscienceofthe19thcenturyisfedbyallthreeattitudes:Wecanadmirethesublimeinnocenceandsimplicitywithwhichmany19thcenturythinkerslaidthefoundationsforcount-lessphilosophicalviewsinthephilosophyofsciencewhoseoriginshaveoftenbeenforgottenandburiedundertheirelaborationbylatergenerations.Wecan,ofcourse,alsooftenseein19thcenturythought“bewitchmentsofourintellect,”asWittgensteinwrote,thatstillexertadeleteriousinfluenceonustoday.Boththeseattitudescanhelpusbetterunderstandourownstandpointinthepresent.Thisisallthemorepossibleasagreementamongphilosophersofscienceontheirsubjectmatterhasdefinitelydiminishedsincethe1960s.HerbertFeiglcouldstillclaimin1964thatinthefieldofphilosophyofscience“thereisperhapsalargermeasureofagreement[…]thaninanyotherareaofphilosophy”(Feigl1964,467).Thestric-turesoflogicalempiricismhavebeenloweredsincethenandhaveledtoamorelarge-mindedattitudeasaresultofwhichthevarietyofindividualapproachesandprocedureshasdefinitelyincreased.Foralongtime,19thcenturyphilosophyofscienceappearedtomanyasamereoutgrowthofabygoneagethateitherundulypraisedscienceastheredemp-tionfromallevilor,equallyunduly,condemneditasgodlessmaterialismandtheoriginofmodernsociety’sfaults.Inanycase,thephilosophyofscienceofthe19thcenturycarriedwithitanovertoneofWeltanschauungthatmadeitoftenindigest-ibleforthephilosophicaloutlookofmostofthe20thcentury.Thenegativeimageofthe19thcenturyallowedforonlyahandfulofexceptions,mostly“forerunners”likee.g.thetoweringphilosopher-scientistsMachorHelmholtz,orsomeisolatedgeniuseslikePoincaréorEinstein,whobecamehouseholdgodsoflogicalempiri-cism,themajorschoolofphilosophyofscienceinthe20thcentury.Theimageofthe19thcenturyasanobsoleteandoutdatedagewithwhichphilosophyofsciencehaslittleornothingtodoanymorewasalsoduetothedev-astatingexperienceoftheFirstWorldWar.Inthisvein,HansReichenbachwroteinthefirstissueofErkenntnisofthe“breakdownoftraditionalemotionalworlds”(Gefühlswelten),of“disenchantment”(Entzauberung)and“desoulment”(Entsee-lung).Thegreatdisillusionment(Ernüchterung)ofthetimeisforReichenbach,ashewrote,notonly“adominantfeatureofscientificresearch,butitisalsothemainfeatureofourdailyexistence,thecategoryunderwhichwehavetoseeourpresent”(Reichenbach1930/31,70).Anotherreasonforleavingthe19thcenturybehindatthetimecanbeseenintherevolutionarydevelopmentsinphysics:withrelativitytheoryandquantummechanicstheoldmechanistworldviewofthe19thcenturycanbethrownintothedustbinofhistoryandwithittheassociatedphi-losophyofscience.Inthemeantime,however,itseemsthat,ifIamnotmistaken,amorepositiveattitudehassetintowardsthe19thcentury.19thcenturyphilosophyofscienceisincreasinglyseeninamuchmorepositivelightoratleastfeltasasounding-boardthatstillresonatesintoday’sphilosophicalwork.Thereasonsforthisdevelop-mentaremanifold.Afirstreasonliesinthefactthatinthewakeofpost-positivist\nAspectsof19thCenturyHistoryofPhilosophyofScience69philosophyofscience,i.e.thephilosophicaldevelopmentafterKuhn,Feyerabend,Lakatosandothers,philosophersincreasinglylefttheirarmchairsandtriedtodojusticetoconcretehistoryofscienceintheirwork.Asaresultofthis,philosophersofsciencehavedevelopedaninterestinthehistoryoftheirownfield.Oneofthefirststepswastohaveacloserlookatthe“official”forerunnersofenlightened20thcenturyphilosophyofscienceastheyarecodifiedinthe1929manifestooftheVi-ennaCircle,forexample.Thiscloserlookrevealedthatthese“heroes”weremuchlessanexceptionoftheirtimeandmuchmoreembeddedintothecontemporarymainstreamofthoughtthaniswidelyassumed.Inmyownwork,IhavefoundoutandshownthatErnstMach,oneofthefoundingfathersofourfield,isdeeplyrootedinthepsychophysicsofG.T.Fechnerandtheintricaciesofthemind-bodyproblemasitposeditselfatthetimeofthenewlyrisingfieldofexperimentalpsy-chology(Heidelberger2004,chs.4and6).Thistrendtowardsmorehistoricalcontextissometimescorrelatedwithagrowingawarenessthatsomeofthehotlydebatedissuesintoday’sphilosophyofsciencearetakingonamoreweltanschaulichdimensionaswellthatisgettingclosertostandpointsofthe19thcenturyinsomerespects.Withtheanti-metaphys-icsofCarnapandNeurathandothersitseemedforawhileasifgenuinephilo-sophicalcontroversieshadcometoanend.Therecannotreallybeanydifference,e.g.,betweenidealismandrealism,Carnapheld,otherthanametaphysicalone.Andifmetaphysicsisnonsense,thedifferencebetweenidealismandrealismdis-appears.Yetthedevelopmentofphilosophyofsciencehasshownthatthishopewasmuchtoorash.Eventhemosthardboiledfolloweroflogicalempiricismhastoadmitthatthereisatleastadifferencebetweenanempiricistandarealistat-titudetowardsempiricaltheories.Therearemanymoresubstantialphilosophicalproblemsthatarenotjustpseudo-problemsandarenottobesolvedjustbyad-ditionalempiricalinformationorlogicaltricks.Thenewlyfoundappreciationofsubstantivephilosophicalproblemsinthephilosophyofscienceiscomplementedbyatrendinsocietyatlargetoconnectsciencewithmorethanjusttechnicalexpertise.Inanextremescenario,thereisthepossibilitythattoday’scontroversiesaboutscienceriskconjoiningwiththeconflictofMuslimculturewiththeWestsomeday.Aglimpseofthiscanalreadybefeltinthedisputesaboutintelligentdesign.Asaresultofthisdevelopment,philosophyofsciencefindsithasadeeperaffinitywiththewayphilosophicalproblemsweredebatedinthe19thcenturythanwiththedeflationaryattitudeofmuchof20thcenturyphilosophy.Take,forexample,thesecondpartofFriedrichAlbertLange’sHistoryofMa-terialismfromthe1870s,andreadwhathehastosayaboutthe“philosophicalmaterialismsincethetimeofKant”(Lange1873-75).Yougeta“déjàvu”exceptthatthematerialistprophetsoftodaynolongerteachatGermanorSwissuniver-sitiesbut,rather,atsomeremoteAustralianorTexanone.Orflipthesamebookopentothechapteron“brainandsoul”andyouwillalmostgettheimpressionthatyouarereadingapresent-dayphilosopherofsciencerebuttingtheclaimofsome\n70MichaelHeidelbergerneuroscientistthatLibet’sexperimentshavedoneawaywithfreedomofthewillonceandforall.AnotherexampleisLange’schapteron“DarwinismandTeleol-ogy”whereyoufindmaterialthatcouldeasilybeusedtodayasargumentsagainstthebeliefinintelligentdesign.Also,Lange’spassionateoutlookonthenewlyemergingscientificpsychol-ogyofhistimehasmuchmoresimilaritywithourdiscipline’senthusiasmwithcognitivesciencethanwiththemethodologicalgamesphilosophyofscienceplayedwithlogicalorempiricalbehaviourismintheearly20thcentury.Lange,however,isnottheonly19thcenturyauthorwhocanspeaktousdirectlywithoutmuchhermeneuticsinvolved.IthinkthatthereishardlyabetterwaytointroduceastudenttotheenigmasofrealismandantirealisminphilosophyofsciencethantoreadwithherorhimachapterofVaihinger’sPhilosophyofthe‘AsIf’ortocomparesomesuitableparagraphsoftheMechanicsofMach,theantirealist,withacorrespondingtextfromatalkofHelmholtz,thearch-realist.ItsometimeslooksasifthetwoWorldWarsandtheiraftermathdistractedphilosophyofsciencefromitsrealtopicsandthatthedeclineofLogicalEmpiri-cism–atleastintheperspectiveproposedbyGeorgeReisch–andthemoreliberalattitudetowardsmetaphysicsandotherschoolsofthoughthasbroughtusbackagaintothepointfromwhereLogicalEmpiricismstarteditsjourney.Historianssometimescallthe20thcenturythe“shortcentury”andcompareittothepreceding“long”onebecausethedeepchangesthattookplaceinthe1900sonlyspanfrom1914to1989,whereasthe19thcenturyasanepochmustbeconsideredfrom1789to1914.Itseemstomethatthe20thcenturywasshortalsointhesecondsensethatthenumberofdurablerevolutionaryachievementsinphilosophyofscienceturnsouttobelessthanoneoriginallythought.EventhecalltoreverttoKantthatcouldbeheardinphilosophyfromthe1830sonwardandthatdominatedsomuchofthe19thcentury,alsoinphilosophyofscience,hasitsseriouspresent-daycounterpart.MichaelFriedmantirelesslytriestoconvinceusthattherelativismthataroseinthewakeofThomasKuhn’sworkcanbeovercomebymakinguseofReichenbach’sdistinctionbetweentwomeaningsoftheKantianapriori,necessaryandunrevisableontheonehand,andconstitutiveoftheconceptofthescientificobjectanditsknowledgeontheother.Einstein’srevolutionhasshown,saysFriedman,thatthefirstmeaningmustberelativizedinordertoallowforhistoricalchangewhereasthesecondmeaningcanandmustberetained.QuestionsarisinginsideascientificparadigmcanbelikenedtoCarnap’s“internalquestions”andtherelativeapriori,whereasexternalques-tionsthatconcerntheframeworkitselfcorrespondtoKuhn’sscientificrevolutionsandtotheconstitutiveapriori.Inthissense,Friedmanisabletoarguethatthepartingofthewaysofso-called“analytical”and“continental”philosophy(ofsci-ence)wasnotaninescapablefatebutmoretheresultofthecontingenciesofthetwisted20thcentury.Intheend,Carnapappearsasthecentralfigureof20thcenturyphilosophyofscienceandphilosophytoutcourtbecausehisprinciplesfitbestintotheneo-neo-KantianismproposedbyFriedman(Friedman2001).\nAspectsof19thCenturyHistoryofPhilosophyofScience71ThecalltoKantisnottheonlynostalgicreturntothe19thcenturythatcanbenotedinthephilosophyofsciencetoday,althoughitisperhapsthemostnotableone.WearealsowitnessestoarevivalofanAmericanandotherbrandsofprag-matismthatseems,however,alittlelesshomogeneousthantheneo-Kantianone.Andwecanfinallynoteagrowinginterestinso-called“continentalphilosophyofscience”generally.Thisisnolongerjustafringeinterestindubiousandfrivolousideas,astheword“continental”originallywassupposedtosignify,butingenuineandviablealternativestologicalempiricism.Theuseoftheterm“continental”hasledtoacuriousdevelopmentinthisre-spect.FromtheearlydaysofFrege,RussellandWittgenstein,analyticalphiloso-phyhasinheritedastronganti-naturalisttendency.Sincethephilosophyofsciencegrownoutoflogicalempiricismalsopartiallysympathizedwithaphilosophicalprogrammeofnaturalisation,butwasandisatthesametimepartoftheanalytic,andthusanti-naturalist,movement,atensiondeveloped,fromthetimeofNeurathandCarnaponward,betweenamorelogicalandamorenaturalistoutlookinphi-losophyofscience.Asaresult,someonelikeErnstMach,whoisjustlycalledtheforefatheroflogicalempiricism,isdiscussedtodayunderthelabelof“continentalphilosophyofscience”–i.e.ofamovementthatissupposedtobeinoppositiontoanalyticalphilosophy(Babich1994,Norris1999).AlogicalconsequencewouldbetocategorizeQuineinthesameway:Althoughhewasoneoftheforemostlogiciansofthe20thcentury,his“epistemologynaturalized”madehimacriticoflogicalempiricismandthusof“analytic”philosophyofscience.Inaddition,hiscritiquewasdevelopedwiththehelpofatleasttwotypical“continental”philoso-phers:DuhemandMeyerson(cf.Laugier2009).Theresultingobliterationofcat-egorieshasasalutaryeffect:Onecanadmitthatphilosophyofsciencehasneverbeenintrinsicallyoressentiallyanalytic.Thisinturnallowsustogiveupmoreeasilytheahistoricalpostureentertainedby“analytical”philosophyandadmittingouraffinityto19thcenturyphilosophyofscience.AsimilarremarkcanbemadeabouttheFrenchphilosophyofsciencetradi-tionwithBachelard,Canguilhemandothers.Theyseemevenmoretoconstituteaparadigmofa“continental”approachthanMachdoes.ButthisviewneglectsthefactthatwithouttheconventionalismofHenriPoincaré,whichdeeplyinformed,andstillinforms,Frenchépistémologie,logicalempiricismwouldnothavede-velopedinthewayitdid.Curiouslyenough,theFrenchphilosophyofsciencetraditionseemsformanytoberootedmuchmoreinthe19thcenturythanCarnaporNeurathandotherlogicalempiricists,althoughbothsideslivedandworkedintheearly20thcentury.Thereturnofthe19thcenturyintoday’sphilosophyofscienceisnotonlyvis-ibleinpublications,butalso,Ithink,inteaching,atleastinmyownexperience.Ifyoureallywanttoenthusebeginningstudentsforthesubject,itisveryefficienttotakesomescientist-philosopherofthelate19thcenturyasastartingpoint–think-erslikeMach,Helmholtz,Boltzmann,Darwin,Hertz,Poincaré,Duhem,EinsteinorevenOstwald–orascientificallyenlightenedphilosopherlikeF.A.Lange,\n72MichaelHeidelbergerAloisRiehlorHansVaihinger.EvenSchopenhauerandNietzscheand,yes,HenriBergson,candothesamejobtosomeextent.Incidentally,mostoftheseauthorsservedasinspirationsforourlogicalempiricistheroeswhentheyattendedtheGymnasium.Thatisalreadyenoughreasonnottoforgetaboutthemtoday.IfindithardtoseducerawrecruitsdirectlytoCarnaporQuineorNeurath,althoughtheyshould,ofcourse,getexcitedaboutthematalater,moreadvanced,stage.Theremightbeaslightdifferencebetweenteachingphilosophyofsciencetobeginningphilosophersandteachingittobeginningsciencestudents.Thelatterdonotneedasmuch19thcenturyflavourbecausetheycanmuchmoreidentifywith20thcenturysciencethanmost“pure”philosophystudents,buttheyseeitasachallengetobemasteredandtobeconnectedrightawaywithaphilosophicaloutlook.Butevenamongsomephysicsstudentsoftodayonecanfindafeelingoffindesiècleasitwaspresentattheendofthe19thcentury:Mostofthefunda-mentalproblemsofphysicsseemtohavebeensolvedbynow,andthosethatareleftmightbeunsolvable,sothattheonlyjobleftistoexpandtherealmofapplica-tionoffundamentalscientificinsightsandnotsomuchtofindnewrevolutionarybeginnings.Perhapsyouhavebecomesuspiciousbynowofmyenthusiasmforthe19thcenturyandtendtoregarditasmeresentimentalismthatshiesawayfromad-dressingthechallengesofthescienceofthe21stcentury.Ithink,however,thatifwewanttounderstandwherewestandtodayasphilosophersofscience,andasEuropeanphilosophersofscienceatthat,wedefinitelyhavetocometotermswithourownhistory.AlanRichardsonhasrecentlyputitinasuccinctway:Mostphilosophersofscienceengagewithlogicalempiricismonlyinsofarastheyarecon-cernedtoclaimthattheyhavegonebeyondit.[…]Yetthequestionofwhatcontemporaryphilosophyofscienceowestologicalempiricismandhowithasadvancedbeyonditcanbeadequatelyansweredonlywith[…]ahistory[oflogicalempiricismasaprojectintwen-tiethcenturyphilosophy].Onlythroughsuchahistorycanphilosophersfullyunderstandboththeirsenseofwhatisphilosophicalintheirownprojectandhowtheyoughttoengageinphilosophicalinquiry”(Richardson2008,96).Icanverymuchagreewiththis,butIthinkthatasEuropeanphilosophersofsci-encetryingtofindourownwaytoday,wecannotlimitourselvesandbecontentwithjustahistoryoflogicalempiricism.ForNorthAmericanphilosophyofsci-enceandperhapsforthosepartsoftheinternationalcommunitythatareexclu-sivelyinfluencedbyanAnglo-Saxonoutlook,thismightbeenoughforsometime.ButevenfromthisoutlookitisnecessarytoconsideralsothehistoryofAmericanpragmatismaswell:someonelikeQuinecannotbeunderstoodwithouttakingac-countofapragmatistinfluence.FromagenuinelyEuropeanperspective,however,thereis–andshouldbe–aspecialinterestinthedevelopmentofphilosophyofsciencethatgoesbeyondtheformativeperiodoflogicalempiricism.Europeanphilosophyofsciencecomprisesmore:Therearemanydifferentandrichtradi-\nAspectsof19thCenturyHistoryofPhilosophyofScience73tionsofreflectingonsciencethatalreadyexistedbeforetheriseoflogicalempiri-cismandstillawaitreappraisal.Thehistoryofthephilosophyofsciencetobere-consideredandunderstoodinordertodevelopaEuropeanidentityforphilosophyofsciencedefinitelyextendsbeyondtheformationperiodoflogicalempiricismandbeyondtheofficialforerunnersofthismovementfurtherintothepast.SoIwouldcountertheallegationofbeingsentimentalistaboutthepastagainstthosewhodisdainthehistoryoftheirfieldwiththechargeofperhapsblithefulbutmind-lessignoranceofone’sownidentity.Inconclusion,letmereflectalittleontheconsequencesthatresultfromthisdefenceofthehistoryof19thcenturyphilosophyofscienceforourfuturework.Forme,ithasbecomeevidentbynowthatweshouldstarttoundertakecompara-tivestudiesinthehistoryofthephilosophyofscience.Ithinkthatlittleworkhasbeendonesofarinthisdirection,regardlessofhowwedefinetheprecisemeaningof“comparative”.Itcanmeanthatwestartstudyingthedifferenttransfersofideasfromonecountryorculturetoanotherduringthe19thcenturyandbeyond.Itcanalsomeancomparingthephilosophicaloutlookonthedifferentsciencesfromonecountryorsocietytoanother.Anditcanlikewiseimplythestudyofthedifferenceorsimilarityamongconceptsthatplayed–andstillplay–aroleinthephilosophyofscienceingeneral.Thisshouldnotbemisunderstoodasapleafornational(ist)histories.Onthecontrary:itshouldandcanmakevisiblethesubtlebuteffectiveinterdependenceofaphilosophicalspiritinrelationtosciencethathasafteralloutlivedmanywarsandhostilitiesamongthedifferentEuropeansocietiesandhasintheendledtocross-culturalmovementssuchaslogicalempiricism.EvenifitmightintheendbefarfetchedtoassumeacommonEuropeanspiritinallthedifferentapproachesthathavebeendevelopedinourhistory,itwouldbewrongjusttoforgetthesedifferentattitudes.Recently,IhavereadaFrenchintro-ductiontothephilosophyofsciencethatdrawsaveilofsilenceoverthepeculiartraditionofFrenchépistémologie(Barberousse2000).ItotherwisegivesaquiterespectableoverviewofthefieldasitstandstodayintheAnglo-Saxonworld.ItistruethatpartsoftheFrenchacademicscenestillhavetocatchupwiththein-ternationaldevelopmentandaresometimestoowrappedupin(oftenexclusivelytheirown)history,therebyneglectingsystematicapproaches.Butthisshouldnotbeareasontohushupone’sownroots.Andevenifonefeelsalienatedfromone’sowntradition,itisstillworththeefforttolearnfromthefailuresofthetradition–iftherereallyareany.Butinordertolearnfromsomething,onehasatleasttotakenoteofit.REFERENCESBabich,BabetteE.(1994),“Philosophiesofscience:Mach,Duhem,Bache-lard”,inTwentieth-CenturyContinentalPhilosophy,ed.byRichardKearney.\n74MichaelHeidelberger(RoutledgeHistoryofPhilosophy,VolumeVIII)London:Routledge,144-183.Barberousse,Anouk,MaxKistlerandPascalLudwig(2000),LaphilosophiedessciencesauXXesiècle.Paris:Flammarion.Bréhier,Émile(1940),Laphilosophieetsonpassé.Paris:PressesuniversitairesdeFrance.Feigl,Herbert(1964),“PhilosophyofScience”,inPhilosophy,ed.byRoderickM.Chisholmetal.EnglewoodCliffs,NJ:Prentice-Hall,465-540.Friedman,Michael(2001),DynamicsofReason:The1999KantLecturesatStan-fordUniversity.Stanford:CSLIPublications.Heidelberger,Michael(2004),NaturefromWithin:GustavTheodorFechner’sPsychophysicalWorldview.Pittsburgh:TheUniversityofPittsburghPress.Lange,FriedrichAlbert(1873-75),GeschichtedesMaterialismusundKritiksein-erBedeutunginderGegenwart.2ndrev.andexpandeded.2vols.,Iserlohn:Baedeker.Vol.1:1873;vol.2:1875.[Reprintin2vols.,ed.AlfredSchmidt.FrankfurtamMain:Suhrkamp1974.]Transl.ofthe2ndGermaned.asTheHistoryofMaterialismandCriticismofitsPresentImportancebyErnestChesterThomas.3vols.,London:TheEnglishandForeignPhilosophicalLi-brary1877-79.Reprintofthe3rded.of1892inonevol.withintrod.byBer-trandRussell.London:Kegan,Paul1925.AlsoNewYork:Humanities1950,andLondon:Routledge2000.1stGermaned.,Iserlohn:Baedeker1866.]Laugier,Sandra(2009),“ScienceandRealism:TheLegacyofDuhemandMeyer-soninContemporaryAmericanPhilosophyofScience”,inFrenchStudiesinthePhilosophyofScience,ed.byAnastasiosBrennerandJeanGayon.(Bos-tonStudiesinthePhilosophyofScience,276)Dordrecht:Springer,91-112.Norris,Christopher(1999),“ContinentalPhilosophyofScience”,inTheEdin-burghEncyclopediaofContinentalPhilosophy,ed.bySimonGlendinning.London:Routledge,402-415.Reichenbach,Hans(1930-31),“DiephilosophischeBedeutungdermodernenPhysik”.Erkenntnis1:49-71.Richardson,Alan(2008),“ScientificPhilosophyasaTopicforHistoryofSci-ence”.Isis99(1):88-96.UniversitätTübingenPhilosophischesSeminarBursagasse1.D-72070TübingenGermanymichael.heidelberger@uni-tuebingen.de\nMASSIMOFERRARIWELL,ANDPRAGMATISM?COMMENTONMICHAELHEIDELBERGER’SPAPERMichaelHeidelbergersuggeststhat“wearewitnesstodayofarenewedinterestinthehistoryofphilosophyofscience,especiallyofthe19thcentury”.ThisstatementinvolvesabroadhistoricalperspectiveandasksnotonlythatwetakeadeeperlookintothephilosophyofscienceoftheageofHelmholtzandMach,ofPoincaréandHertz,butalsothatweconsiderthemandotherleadingfiguresofthetimeinthecontext–asHeidelbergercorrectlysays–ofnationaltraditions(suchasthatofFranceorevenItaly)andwithinamorearticulatedhistoricalbackground.1Iagreeinparticularwiththeproposalfor“comparativestudiesinthehistoryofphiloso-phyofscience”.Thetransferofideasfromonecountrytoanotherduringthe19thcenturyrepresentsacrucialhistoricalissueforourresearchandmaycontributetoanewinterpretationofthehistoryofphilosophyofscience,notonlyduringthe“longcentury”,butalsoinregardtothe“short”one,i.d.lookingforward–broadlyspeaking–toourpresentdebatesandourphilosophicalagendas.Heidelbergeroffersverystimulatingconsiderationsaboutthe“nostalgicre-turn”toKantthatcanbenotedtodayandsuggeststhatphilosopherssuchFrie-drichAlbertLangecanbereadinadifferentperspective.HisgreatworkontheHistoryofMaterialismwasactuallyaStandardwerkforalmosttwogenerationsofphilosophersofscience;andasWilhelmDiltheywrotein1877,Lange’sopusmagnumwasdestinedtoremainabookmarkingaturnpointinthephilosophicaldebateoflate19thcentury.2ButtheHistoryofMaterialismwasaveryimportantreferenceforCarnap,ReichenbachandSchlickstillintheageofearlyLogicalEmpiricism.Thoughitmayseemstrange,wecanreadinthethirdissueofErkennt-nisanenthusiasticportraitofLange,oneofthefewphilosopherswhowasable–accordingtotheeditorsofhiscorrespondencewithAntonDorn–toengageinadialoguewiththenaturalsciencesandtoacquireinthissensethegreat«merit»(quitesimilartotheMarburgSchoolofNeo-Kantianism)tohaverediscovered“Kantasanaturalscientist”,inoppositiontothemetaphysicalinterpretationof1AmongrecentpublicationsonthisissueIwouldliketomentionthestimulatingcollec-tionsofessaysJean-ClaudePont/LaurentFreland/FlaviaPadovani/LiliaSlavinskaia(Eds.),PourcomprendreleXIXe.Histoireetphilosophiedessciencesàlafindusiè-cle.Firenze:Olschki2007andMichaelHeidelberger/FriedrichStadler(Eds.),HistoryofPhilosophyofScience.NewTrendsandPerspectives,Dordrecht–Boston–London:Kluwer2002.2WilhelmDilthey,GesammelteSchriften,vol.XVII,ZurGeistesgeschichtedes19.Jahrhunderts,ed.byU.Hermann,Göttingen:Vandenhoeck&Ruprecht1974,p.101.F.Stadler(ed.),ThePresentSituationinthePhilosophyofScience,ThePhilosophyofScienceinaEuropeanPerspective,DOI10.1007/978-90-481-9115-4_6,©SpringerScience+BusinessMediaB.V.2010\n76MassimoFerrariKant’sphilosophyendorsedby“professionalphilophers”.Itwillbeusefultore-member,thatsuchapraiseofLangewasdrawnfromnolessthanRudolfCarnapandHansReichenbach.3OtherwiseitiswellknownthatNeokantianism,firstofalltheMarburgNe-okantianismofCohen,NatorpandCassirer,butalsotheNeokantianismofAloisRiehlor,tosomeextent,ofHansVaihinger,hadagreatinfluenceonthephilosophyofsciencewhichstartsitsjourneyattheViennaStation.Historicalandsystematicreconstructions–inthecaseofCarnapitwillbeenoughtoremindyouofthecon-tributionsofAlanRichardsonandAndreW.Carus–showinaveryexcitingwayhowthereceivedviewandthecurrentgeneaologiesofLogicalEmpiricismmustbecorrectedwithintheframeworkofanotherstory.4Thatstorystartsfromscien-tificNeokantianismandreformulatessomecrucialaspectsofthistraditioninare-allyrevolutionarynewperspective.Butthebreakdownoftheolderworldineveryrevolutionismorecomplicatedthanamerefarewelltothepreviousageor,inthiscase,topreviousconceptualtools.Inmyopinion,Neokantianismwasinthiscontextnotonlyaphilosophicalstream,butabranchoflate19thcenturyphiloso-phyofsciencethat,firstofallintheGermanspeakingworld,wasveryinfluentialon,andatsomelengthwaselaboratedfrom,theleadingscientistsintheirownwork.5Therewasatime,toputitdifferently,inwhichagreatphysicistasHeinrichHertzwasabletoreadKant’sCritiqueofPureReasonorhisFirstMetaphysicalPrinciplesofNaturalScienceafterlonghoursofhardlaboratorywork,obviouslynotasa“moralholiday”fromhisscientificengagement.6Ifourhistoricalandphilosophicaltaskisthecontextualizationofphilosophyofscienceaswellasofepistemologicalframeworkselaboratedinthelate19thcentury,itseemsunavoid-abletoelucidatetheassimilitationwithinphilosophyofscienceoftheKantianandNeo-Kantianheritage,or–toquoteMichaelFriedman’sstatement–todescribe3Ireferto„DokumenteüberNaturwissenschaftundPhilosophie.Briefwechselzwi-schenFriedrichAlbertLangeundAntonDohrn“,in:Erkenntnis3,1932/33,pp.262-300(quotationfrompp.262-263).4AlanW.Richardson,Carnap’sConstructionoftheWorld.The“Aufbau”andtheEmergenceofLogicalEmpiricism,Cambridge:CambridgeUniversityPress1998andAndreW.Carus,CarnapandTwentieth-CenturyThought.ExplicationasEnlighten-ment,Cambridge:CambridgeUniversityPress2007.5ForanexcellentoverviewonthistopicsseeMichaelFriedman/AlfredNordmann(Eds.),TheKantianLegacyinNineteenth-CenturyScience,Cambridge(Massachu-setts)–London:TheMITPress2006.SeealsoMassimoFerrari,“IlKantdegliscien-ziati:immaginidellafilosofiakantiananeltardoOttocentotedesco”,in:GiusppeMicheli(Ed.),MomentidellaricezionediKantnell’Ottocento,Milano:FrancoAngeli2006,pp.183-201.Forthephilosophyelaboratedbyscientistsmoregenerally,seethenoteworthybookbyErhardScheibe,DiePhilosophiederPhysiker,München:Beck2007.6HeinrichHertz,Erinnerungen,Briefe,Tagebücher,ed.byMathildeHertzandCharlesSusskind,SanFrancisco:PhysikVerlag1977,p.190.\nWell,andPragmatism?77“howtheoriginalKantianpositionwassuccessivelytransformedbyalongtradi-tionofscientificthinkersleadingallthewayuptothepresentday”7.MichaelHeidelbergerremindsusthatwearenowalsowitness“ofarevivalofAmericanpragmatism”,although“lesshomogeneousthantheneo-Kantiancamp”.ThisisaveryinterestingpointandIwouldliketodevelopsomereflectionsaboutthetransferofAmericanPragmatismfromHarvardtoEuropeand,particularly,toItalyontheonesideandtoViennaontheother.Ithinkthatthelistofscientist-phi-losophersorphilosophersofthelate19thcenturythatHeidelbergcitesasfruitfulteachingmaterial(Mach,Helmholtz,Poincaré,Duhem,LangeorevenBergson)canbeenrichedbythenameofaleadingfigureofPragmatism:WilliamJames.Thestandardviewofthetopic“PragmatismandEuropeanPhilosophyofSci-ence”iswellknown.Accordingtoit,theemigrationofLogicalEmpiricismfromGermany,AustriaandCentralEuropebetweenthewarsandtheintricateprocessofitsalterationinthe“newworld”createdacontextinwhichEuropeanphilosophyofsciencewascontaminatedbyNorthAmericanwaysofthinking,especiallythetraditionofpragmatism.Thisstandardviewhasindeedoverlookedtwoaspects.Ontheoneside,recentscholarshiphasshowedthatthetransferofLogicalEmpiri-cismintheU.S.Ainvolvedanincreasingprofessionalizationofphilosophyofsci-enceand,atthesametime,thelostofthetypicalpoliticalandculturalengagementofitsheydayinVienna.8Ontheotherside,andthisismuchmoreimportantforourpresentperspective,arelationshipbetweenEuropeanphilosophyofscienceandPragmatismwasestablishedlongbeforetheintellectualemigrationfromEuropebetweentheWorldWars.EspeciallyJames’pragmatisticinsights–certainlymoreJames’versionofPragmatismthanPeirce’s–travelledfromAmericatoEuropeattheverybeginningof20thcenturyinpreciselytheoppositedirectionofthelater,morewell-knownjourneyfromWeimarGermanyand‘red’ViennatoAmericandepartmentsofphilosophy.Inotherwords,thereisanotherversionofthestoryoftherelationshipbe-tweenPragmatismandLogicalEmpiricismwhichstartsattheendof19thcenturyandwhosedirectionis–paraphrasingthetitleofGeraldHolton’scontributionon“theAmericanizationoftheWissenschaftlicheWeltauffassung”–fromHar-vardSquaretotheViennaCircle.9Abriefaccount,particularly,ofthereceptionofWilliamJames’pragmatismwithinEuropeanphilosophyofsciencewouldun-7MichaelFriedman,“HistoryandPhilosophyofScienceinaNewKey”,in:Isis99,2008,p.133.8SeetheilluminatingreconstructionofferedbyGeorgeA.Reisch,HowtheColdWarTransformedPhilosophyofScience.TotheIcySlopesofLogic,NewYork:CambridgeUniversityPress2005.ImportantcontributionsonthistopiccanbefoundalsoinGaryL.Hardcastle/AlanW.Richardson(Eds.),LogicalEmpiricisminNorthAmerica,Min-neapolis-London:UniversityofMinnesotaPress20039GeraldHolton,“FromtheViennaCircletoHarvardSquare:TheAmericanizationofaEuropeanWorldConception”,in:FriedrichStadler(Ed.),ScientificPhilosophy:Ori-ginsandDevelopments,Dordrecht–Boston–London:Kluwer1993,pp.47-73.\n78MassimoFerraridoubtedlydealwithMachandhisentourageinVienna.AsHoltonpointsout,James’“philosophyofPragmatism,developedinthefirstinstanceasawayoutofapersonalstrugglethathasbeencalledJames’‘Kantcrisis’,overlappedwithMachianempiricistpositioninmanyways,forexample,infindingthemeaningofideasinthesensationsthatmaybeexpectedfromtheirrealization”.10Tobesure,JameswaswellacquaintedwithMach’sworkswhichhehadreadcarefully,mak-ingannotations,marginalia,queriesandsoon;and,particularly,JameswasdeeplyinterestednotonlyinMach’sAnalysederEmpfindungen,butalsoinhisbookonMechanik,especiallyMach’sfamousdiscussionofNewton’sviewsontime,spaceandcausality11.Forhispart,MachwasindeedaconvincedsupporterofJames’workonThePrinciplesofPsychology,butontheotherhandhisdisagreementwithPragmatismasphilosophicalorientationwasquiteclear:aninterestingproofofhiscriticalevaluationmaybefoundinalettertoAntonThomsenfromJanuary1911.12Nevertheless,theconnectionJames-Machsuggestsfirstofallanothercon-nectionwhichhastodowithboththeAmericanthinkerandtheViennesescien-tist.WemeantheItalianphilosopherofscienceandlanguageGiovanniVailati,aformercollaboratorofGiuseppePeano’sFormulariomathematicoandacon-vincedsupporterofMach’shistoricalandepistemologicalwork,whowasalsoengaged,attheverybeginningofthecentury,toendorsea“logicalpragmatism”quitedifferentfromthe“magicpragmatism”ofhisfriendGiovanniPapini.VailatihadagreatadmirationforPeirceandhispragmaticruleofmeaning(i.e.theruleformulatedbyPeirceinhisseminalessay“HowtoMakeOurIdeasClear”),buthealsowasawareimmediatelyoftheepistemologicalrelevanceoftheJamesianpragmatism.InhisreviewsbothofTheWilltoBelieveandsomeyearslaterofJames’famousPragmatism,VailatiemphazisesJames’greatmeritofhavingof-feredacertainrehabilitationof«theconstructiveandanticipatingactivitiesofhumanunderstanding».AccordingtoVailati,Jameswasrighttocriticiseasthecommonunderstandingscientificandphilosophicaltruthhasunderestimatedthisaspectandconsequentlyhasendorsedanimageofmentalactivitywhichislimitedtoamereclassificationand,sotospeak,arecordingofempiricaldata.InVailati’sopinion,Jamesisinthisrespectperfectlyinagreementwiththerecent“logicofscience”,namelywiththeanalysesdevelopedbyMach,Cliffordandothersofthemethods,historyandprinciplesofmodernscience.Ontheotherhand,VailatiunderlinestheepistemologicalimportanceofJames’criticalassessmentofposi-tivismaswellasofthesometimes«narrow-minded»philosophynourishedbythe10Holton,“FromtheViennaCircletoHarvardSquare”,loc.cit.,p.50.11Holton,“FromtheViennaCircletoHarvardSquare”,loc.cit.,p.51.12ErnstMachalsAußenseiter.MachsBriefwechselüberPhilosophieundRelativitäts-theoriemitPersönlichkeitenseinerZeit,ed.byJ.BlackmoreandK.Hentschel,Wien:Braumüller1985,p.86(„DerSchwerpunktseinerArbeitliegtgewißinseinerausge-zeichnetenPsychologie.MitseinemPragmatismuskannichmichnichtganzbefreun-den“).\nWell,andPragmatism?79scientists.AccordingtoVailati,Jamesisperfectlyrightinemphasizingthecrucialroleinthescientificinquiryofaudaciousformulationsofhypotheses;13similarly,hepointsoutthatJameshasrecognizedbetterthananyotherphilosopherofsci-encethefunctionofbeliefforthescientificmethod.14Broadlyspeaking,Vailatiappreciatesthepragmaticviewaccordingtowhichscientificknowledgeisalwaystheresultofamentalconstruction,whereastheempirical,factualbasisseemstobenotasfoundationalandunavoidableasthe(positivistic)standardviewtendstosuggest.15ThegreatmeritofVailatiseemstohavebeentohaveunderstood,quiteun-likehiscontemporaries,thatJameswaselaboratingaversionofPragmatismthatwasinnowaytobethoughtofasamerevoluntaristicoreven“irrationalistic”philosophy.AndwemayrecognizethatVailati’ssuggestionsarecorrect.InhisbookonPragmatism,indeed,Jamesoffersashortbutveryilluminatingaccountofcontemporaryphilosophyofscience.Mach,DuhemandPoincaré–saysJames–are“teachers”,accordingtowhich“nohypothesisistruerthananyotherinthesenseofbeingamoreliteralcopyofreality.Theyareallbutwaysoftalkingonourpart,tobecomparedsolelyfromthepointofviewoftheiruse.”16Moreover,Jamesgivesanholisticaccountofwhatmeanstheacquisitionandgrowthoftruthwithinthehistoricalprocessofknowledgewhichseemsundoubtedly‘uptodate’toareaderwellacquaintedwiththefollowingphilosophyofsciencefromNeurathtoQuine17.Jamessays,forinstance:[A]newideais[…]adoptedasthetrueone.Itpreservestheolderstockoftruthswithaminimumofmodification,stretchingthemjustenoughtomakethemadmitthenovelty,butconceivingthatinwaysasfamiliarasthecaseleavespossible[…]Newtruthisalwaysago-between,asmoother-overoftransitions.Itmarriesoldopiniontonewfactsoasevertoshowaminimumofjolt,amaximumofcontinuity.Weholdatheorytruejustinproportiontoitssuccessinsolvingthis“problemofmaximaandminima”18.Jamesisfullyconvincedthatananti-foundationalistaccountofknowledgeisre-quiredwhenwewanttotakeintoaccountthatourthinkingdevelopsinquiteadifferentwayfromthatofferedbytraditionalphilosophysinceDescartes:13GiovanniVailati,Scritti,Firenze:Seeber&Barth1911,p.270.14RegardingVailati’spositionwithinEuropeanphilosophyofsciencebetween19thand20thcenturyIwouldliketorefertomybookNonsoloidealismo.FilosofiefilosofieinItaliatraOttocentoeNovecento,Firenze:LeLettere2006,pp.141-164.15Vailati,Scritti,op.cit.,p.283.16WilliamJames,Pragmatism.ANewNameforSomeOldWaysofThinking,ClevelandandNewYork:MeridiansBooks1963,p.125.17OnJamesandQuineseeI.Nevo,“James,Quine,andAnalyticPragmatism”,in:R.Hollinger/D.Depew(Eds.),Pragmatism.FromProgressivismtoPostmodernism,Westport(Connecticut)–London:Prager1995,pp.153-16118James,Pragmatism,op.cit.,pp.50-51.\n80MassimoFerrariTobeginwith,ourknowledgegrowsinspots.Thespotsmaybelargeorsmall,buttheknowledgenevergrowsallover:someknowledgealwaysremainswhatitwas[…]Ourmindsthusgrowinspots;andlikegrease-spots,thespotsspread.Butweletthemspreadaslittleaspossible:wekeepunalteredasmuchofouroldknowledge,asmanyofouroldprejudicesandbeliefs,aswecan.Wepatchandtinkermorethanwerenew.Thenoveltysoaksin;itstainstheancientmass;butitisalsotingedbywhatabsorbsit.Ourpastap-perceivesandco-operates;andinthenewequilibriuminwhicheachstepforwardintheprocessoflearningterminates,ithappensrelativelyseldomthatthenewfactisaddedraw.Moreusuallyitisembeddedcooked,asonemightsay,orsteweddowninthesauceoftheold.Newtruthsthusareresultantsofnewexperiencesandofoldtruthscombinedandmu-tuallymodifyingoneanother19.WemayconsequentlyaffirmthatVailatiwasrightinemphazisingtheepistemo-logicalcoreofJames’Pragmatism:thismakeshimanexcellentexceptioninthephilosophicallandscapeatthebeginningof20thcenturyinEurope.Butthereisan-othermeaningfulhistoricalcircumstancethatsupportstherelevanceofVailatiinthiscontext.InSeptember1908VailatiwasinHeidelberginoccasionoftheThirdInternationalCongressofPhilosophy.TheEuropeanquarrelaboutpragmatismstartedjustthere,inthesectionoftheCongressdevotedtothediscussionofFerdi-nandSchiller’stalkaboutthepragmatictheoryoftruth.ThecriticalreactionoftheGermanphilosophicalestablishmenttowardsthe“yankee”philosophyjustarrivedinEuropewasextremelyunfavourableandthedebatefollowingSchiller’slecturewas,accordingtothecongressreport,verylively.20Itisnoteworthy,however,thattheonlyparticipantstotheCongressbeinginagreementwiththepragmaticmeth-odinphilosophywereVailatiandaphilosophicaloutsiderfromVienna,WilhelmJerusalem.InthesameyearastheCongressinHeidelbergJerusalempublishedaverygoodGermantranslationofJames’Pragmatismandwroteahighlyinterest-ingprefacetoit.FirstofallJerusalemexpressedthehopethatJames’contribu-tioncouldbewelcomeinGermanyandbeabletorenewitsphilosophicspirit.InthesecondplaceheunderlinedthatPragmatismwasnotasystem,butamethod,whichfindsitscentreofgravityintherefusalofapriori,asacredplaceforGermanphilosophers.Finally,Jerusalemclaimedthatthepragmatistviewoftruth–whichisherebynomeansassociatedwiththe“yankee”spiritofdollarpursuit–oughttobeintegratedintothehistoricalinvestigationsofthegrowthofknowledgeandintohis“sociologyofknowledge”–whichstudiestruthasa“socialcondensation”–thusachievingaconvergenceofPragmatismandsociology.21Inthesameyearof19James,Pragmatism,op.cit.,pp.112-113.20TheodorElsenhans(Ed.),BerichtüberdenIII.InternationalenKongressfürPhiloso-phiezuHeidelberg,Heidelberg:Winter1909,pp.711-740.21WilhelmJerusalem,„VorwortdesÜbersetzers“,in:WilliamJames,DerPragmatis-mus.EinneuerNamefüralteDenkmethoden,übersetztvonW.Jerusalem,Leipzig:Klinkhardt1908pp.V,VIII-IX.Wemustalsorememberhispaper„SoziologiedesErkennens“publishedinMay1909inDieZukunft(andavailablealsoinWilhelmJerusalem,GedankenundDenker.GesammelteAufsätze.NeueFolge,WienundLeip-zig:Braumüller1925,pp.140-153).\nWell,andPragmatism?811908,JerusalemtookupsuchanalternativeviewofPragmatismsupportingitinapaper,whichrepresents,sotospeak,themissedroadofGermanreceptionofPrag-matism.HeconfirmedhisstruggleagainstapriorismandpresentedJamesianPrag-matismastheirreplaceableallyinordertoofferanalternativesolutiontoKant’stheoryofknowledge.Furthermore,hestronglyinsisted-onthebasisoftheircom-monviewofbiologicrootsofhumanmind-onJames’andMach’saffinities,thusdrawinganidealaxisbetweenViennaandUnitedStates,amovewhichappearedtoaimatavoidingtheencumberingdefensivewallofGermanGeist22.AcloseraccountofJerusalem’scontributiontothediscussionaboutthephilos-ophyofpragmatismaswellasaboutitstheoryoftruthandknowledgeintheGer-manspeakingcultureatthebeginningot20thcenturygoesbeyondthelimitsofthepresentcomment.ItmustneverthelessbeemphazisedthatJerusalemrepresentedtheessentialconnectionbetweenAmericanPragmatismandthefutureVienneseLogicalEmpiricism,notonlyduetohismediationbetweenJamesandGermanspeakingculture,butmorespecificallyduetohisrelationshipwithOttoNeurath,acrucialfigureinthehistoryoftheViennaCircle.Iftheirpersonalconnectionsarestilltobedocumentedindetail,itisnothardtosupposethatJerusalem–whowasactiveinViennanotonlyinthestrictlyacademicenvironment,butalsoinwiderintellectualcircles,inthepressandinculturalassociationswellrepresentedintheAustriancapitalduringNeurath’searlyyears–waswell-knownalsotothefuturepromoterofthe“leftViennaCircle”.23ItwasnotbyaccidentthatNeurathmen-tionsJerusalemnotonlyinalateworkof1935,LedéveloppementduCercledeVienneetl’avenirdel’empirismelogique,whereheplacedhiminthemainstreamofanti-KantianismtypicalofbothAustrianphilosophyandtheViennaCircle,butparticularlyinabrieftextthatfollowedshortlyafterwards.TherehedepictsJerusalemasthe“pioneer(Vorkämpfer)ofapragmatistconception”,underlyinghismembershipofthecharacteristicstreamofHabsburgthoughtandespeciallyofViennaUniversitytradition.24ThankstoJerusalem’smediation,therefore,aconnectionseemstohavetakenplacebetweenPragmatismandLogicalEmpiri-cism.Whilewell-knowninitsgeneraloutlines,itwouldbebetterdescribedinNeurath’scasebythelightofacertainidealfiliationJames-Jerusalem-Neurath,as22SeeWilhelmJerusalem,„DerPragmatismus.EineneuephilosophischeMethode“,in:DeutscheLiteraturzeitung,29,25.Januar1908,coll.197-206(republishedin:Gedan-kenundDenker,op.cit.,pp.130-139).OnJerusalemandPragmatismseeLudwigNagl,„WilhelmJerusalemsRezeptiondesPragmatismus“,in:MichaelBenedikt/ReinholdKnoll/CorneliusZehetner(Eds.),VerdrängterHumanismus–verzögerteAufklärung,vol.V,ImSchattenderTotalitarismen,Wien:FakultasVerlags-undBuchhandelsAG2005,pp.344-353.23SeethedocumentationavailableinThomasUebel,VernunftkritikundWissenschaft:OttoNeurathundderersteWienerKreis,Wien–NewYork:Springer2000,esp.pp.164-167,292-295.24OttoNeurath,DerLogischeEmpirismusundderWienerKreis,in:OttoNeurath,Ge-sammeltephilosophischeundmethodologischeSchriften,ed.byR.HallerandH.Rutte,Wien:Hölder-Pichler-Tempsky1981,vol.II,p.742.\n82MassimoFerrarilongastheconvergenceofhisanti-fondationalistepistemologyandtheoutcomesofAmericanPragmatismfromPeircetoDeweyisrecognised.25Itshort,itwouldnotbeimplausibletoclaimthatmanyissuescharacterisingNeurath’sphilosophy(mainlyinthe1930s)areatleastinagreementwithbothJames’Pragmatismandits“enlargements”proposedbyJerusalemsubspeciethesociologyofknowledge.TherewasaplacealsoforJamesandfortheonewhohasbroughthimtolightintheGerman-speakingphilosophicalcultureatthebeginningoftwentiethcenturyonNeurath’sfamousboat,tousehismetaphorforinquiryandknowledgeasal-waystravellingthroughtheseaofhistoryunabletoassumeatabularasaorbuildonacertainfoundationonceandforall.26Allthishasobviouslytodowithan“image”ofJamesquitedifferentfromtheimagethatwaswidelydominantinearly20thcentury.Hewasinnowaythephilosophersupportingtheyankeewayofthinkingdeploredbyhismostpromi-nentGermancolleaguesatthetimeoftheInternationalCongressofHeidelberg.Jameswasratheraphilosopheroflate19thcenturywhowasperfectlyawareofhiscommitmenttorecentphilosophyofscience.InhisessayHumanismandTruth(1904)Jamespointedouthowdeeplythepragmatisticwayofthinkingwascon-nectedwiththeincreasingtransormationsinexactandnaturalscienceduringthelastdecades.AsIunderstandthepragmatistwayofseeingthings,itowesitsbeingtothebreak-downwhichthelastfiftyyearshavebroughtaboutintheoldernotionsofscientifictruth.“Godgeometrizes”,isusedtobesaid;anditwasbelievedthatEuclid’selementsliterallyrepro-ducedhisgeometrizing.Thereisaneternalandunchangeable‘reason’;anditsvoicewassupposedtoreverbeateinBarbaraandCelarent.Soalsoofthe“lawsofnature”,physicalandchemical,soofnaturalhistoryclassification–allweresupposedtobeexactandex-clusiveduplicatesofpre-humanarchetypesburiedinthestructureofthings,towhichthesparkofdivinityhiddeninourintellectenablesustopenetrate.Theanatomyoftheworldislogical,anditslogicisthatofauniversityprofessor,itwasthought.Uptoabout1850almosteveryonebelievedthatsciencesexpressedtruthsthatwereexactcopiesofadefinitecodeofnon-humanrealities.Buttheenormouslyrapidmultiplicationoftheoriesintheselatterdayshaswell-nightupsetthenotionofanyoneofthembeingamoreliterallyobjec-tivekindofthingsthananother.Therearesomanygeometries,somanylogics,somany25OnNeurathandPragmatismseeThomasMormann,„NeurathsanticartesischeKon-zeptionvonSpracheundWissenschaft“,in:ElisabethNemeth/RichardHeinrich(Eds.),OttoNeurath:Rationalität,Planung,Vielfalt,Wien–Berlin:OldenbourgVer-lag-AkademieVerlag,1999,pp.32-61(MormannhoweverignoresJames’influenceonNeurath).ForabriefmentionoftheconnectionJerusalem-JamesseeNancyCart-wright,JordiCat,LolaFleck,ThomasUebel,OttoNeurath:PhilosophybetweenSci-enceandPolitics,Cambridge:CambridgeUniversityPress1996,p.94n.10.26OttoNeurath,“Protokollsätze”,in:Erkenntnis,III,1932,p.206.RegardingNeurath’s“anti-fondationalisticPragmatism”seeThomasUebel,VernunftkritikundWissen-schaft,op.cit.,pp.88,101aswellasThomasUebel,“OttoNeurath,theViennaCir-cleandtheAustrianTradition”,in:A.O’Hear(Ed.),GermanPhilosophysinceKant,Cambridge:CambridgeUniversityPress1999,pp.257,267.\nWell,andPragmatism?83physicalandchemicalhypotheses,somanyclassifications,eachoneofthemgoodforsomuchandyetnotgoodforeverything,thatthenotionthateventhetruestformulamaybeahumandeviceandnotaliteraltranscripthasdawneduponus.Wehearscientificlawsnowtreatedassomuch‘conceptualshorthand’,truesofarastheyareusefulbutnotfarther.Ourmindhasbecometolerantofsymbolinsteadofreproduction,ofapproximationinsteadofexactness,ofplasticityinsteadofrigor27.Wemaywellaskifthisandsimilarstatementscanbereadasprovidinganotherreasonforlookingat19thcenturyphilosophyofscienceinthenostalgic,butalsofruitfulwayproposedbyMichaelHeidelberger.IwouldliketosuggestthatJamesandsomeofitssupporterssuchasVailatiorJerusalemprovidetheoccasionforastimulatingcasestudythatofferstousagoodopportunitytoachievenewinsightsintothepastand,startingfromareconsiderationofthisneglectedinteraction,intothefutureofthehistoryofphilosophyofscience.UniversitàdegliStudidiTorinoDipartimentodifilosofiaViaS.Ottavio,2010124TorinoItalymassimo.ferrari@unito.it27WilliamJames,“HumanismandTruth”,in:TheMeaningofTruth.ASequeltoPrag-matism,Cambridge-Massachusetts:HarvardUniversityPress1975,p.206\n\nPartII(TeamA)(teamleaderStephanHartmann)FormalMethodsandtheirApplicationstothePhilosophyofScience\n\nVINCENZOCRUPIANDSTEPHANHARTMANNFORMALANDEMPIRICALMETHODSINPHILOSOPHYOFSCIENCEABSTRACTThisessayaddressesthemethodologyofphilosophyofscienceandillustrateshowformalandempiricalmethodscanbefruitfullycombined.Specialemphasisisgiv-entotheapplicationofexperimentalmethodstoconfirmationtheoryandtorecentworkontheconjunctionfallacy,akeytopicintherationalitydebatearisingfromresearchincognitivepsychology.Severalotherissuescanbestudiedinthisway.Intheconcludingsection,abriefoutlineisprovidedofthreefurtherexamples.1.INTRODUCTIONPhilosophersofscienceuseapluralityofapparentlydivergentmethods.Thisclaimcaneasilybesubstantiatedbylookingintooneoftherelevantjournals:onerealizesthatsomeauthorsusethetraditionalmethodofconceptualanalysis,otherengageinformalmodelling,conductcasestudiesand–morerecently–experi-ments,orconsultthehistoryofscienceinconsiderabledetail.Buthowdothesemethodsrelatetoeachother?Isoneofthemtherightone?Pluralisticcautionswouldsuggestthatmultiplemethodologicalapproachesarelegitimate.Infact,wewouldliketostressthatacombinationoftwoormoremethodsmaybeparticularlyfruitfulinsomecases.Carnap,forexample,com-binedformalmethods(i.e.,basedonlogicandprobabilitytheory)withconceptualanalysistoarriveatanexplicationofthenotionofconfirmation.AndauthorsinthetraditionofKuhnandFeyerabendusecasestudiesfromthehistoryofsci-encetochallengephilosophicalmodelsofscientificreasoningsuchasPopper’sfalsificationism.Inthisessaywewouldliketoexplorehowformalmethodsandexperimentscanbecombined.Experimentsarealltherageincontemporaryphilosophy(KnobeandNichols2008,Stotz2009).Inepistemology,people’sintuitionsaboutGettiercaseshavefamouslybeentested.Inethics,aspectsofthefreedomofwilldebatearestudiedexperimentally.Philosophersoflanguagealsotestourintuitionsabouttherefer-enceofpropernames.Thislistcouldeasilybecontinued.Interestingly,theresultsofthesestudiesareoftensurprisingwhencomparedwiththecorrespondingintui-tionsofprofessionalphilosophers.Whilemanyoftheseexperimentsareusedtotestphilosophers’intuitions(orhypotheses),itisworthnotingthatexperimentshaveotherfunctionsbesidestest-F.Stadler(ed.),ThePresentSituationinthePhilosophyofScience,ThePhilosophyofScienceinaEuropeanPerspective,DOI10.1007/978-90-481-9115-4_7,©SpringerScience+BusinessMediaB.V.2010\n88VincenzoCrupiandStephanHartmanning,asHacking(1983)remindedus.Experimentsmay,forexample,inspirenewhypotheses,andthisholdsforexperimentalresearchintraditionaldomainsaswellasinphilosophy.Usually,thesehypothesesarenotputforwardinatheoreticalvacuum:theymayrelatetoanexistingtheoreticalframework,andsosometinker-ingmayhavetobedonetofitthenewhypothesis(oramodifiedversionofit)intothetheoreticalframework(oramodifiedversionofit).Inshort,experimentaldatamayprovideguidanceandinsightintheory-constructioninanumberofways.Thisessayismeanttoillustratetheclaimsabove.Itfocusesonexperimentsandexperimentalphenomenawhicharedirectlyrelatedtoworkdonebyformalepistemologists.Morespecifically,wewilllookattwocasestudies.Section2focusesonconfirmationtheoryandtherecentempiricalworkinthisfield.Sec-tion3discussestheconjunctionfallacy,whichisofconsiderableimportanceastherationalitydebatelurksinthebackground.Finally,wewilloutlinealistofopenproblemssuggestingpromisinglinesofresearchtobepursuedfurtherinthefuture.2.CASE-STUDYI:CONFIRMATIONHypothesistestingandconfirmationhavebeencentralissuesinthephilosophyofsciencefordecades.Earlyaccountsbasedonlogicandessentiallyqualitativenotionshavestruggledtodealwithanumberofpuzzles,includingthe“tacking”problem,Hempel’sparadoxes,Goodman’snewriddle,thevarietyofevidence,andtheDuhem-Quinethesis.Importantly,suchissueshavebeenshowntore-ceiveamoreeffectivetreatmentinquantitativetermswithinaBayesianapproachtoconfirmationandscientificreasoning(seeEarman1992,pp.63-86,foranowclassicaldiscussioninthisvein).Aquantitativeapproachalsoseemstobeuptoageneralreal-worldchallenge:judgmentsconcerningtheamount(ordegree)ofsupportthatapieceofinformationbringstoahypothesisarecommonlyrequiredinscientificresearchaswellasinotherdomains(medicine,law).Thus,acentralaimofphilosophyofscienceandepistemologyistoprovideaproperfoundationtosuchjudgments.Bayesianismarguablyisamajortheoreticalperspectiveincontemporarydis-cussionsofreasoninginscienceaswellasinotherdomains(e.g.,BovensandHartmann2003,HowsonandUrbach2006,OaksfordandChater2007).Bayesiantheoristspostulateaprobabilisticanalysisofmanysortsofordinaryandscientificreasoningbyendorsingasubjectivereadingofprobability,i.e.,byusingprobabili-tiestomodeldegreesofsubjectivebelief.Withinthisframework,contemporaryBayesianscommonlyidentifyconfirmationwithanincreaseintheprobabilityofahypothesishprovidedbyapieceofevidenceeascomparedtotheinitialprobabil-ityofh(i.e.,withevidenceenotbeinggiven).Anaturalwaytomeasureconfirma-tionalstrengththenamountstoafunctionmappingrelevantprobabilityvaluesofh\nFormalandEmpiricalMethodsinPhilosophyofScience89andeontoanumberwhichiseitherpositive,nullornegativedependingonp(h|e)beinghigher,equalorlowerascomparedtop(h).Amongtraditionalproposalsmeetingthisbasicconstraintarethefollowing:1–thedifferencemeasure:D(h,e)=p(h|e)–p(h)–the(log)ratiomeasure:R(h,e)=log[p(h|e)/p(h)]–the(log)likelihoodratiomeasure:L(h,e)=log[p(e|h)/p(e|¬h)]Morerecentvariantsincludethefollowing:2S(h,e)=p(h|e)–p(h|¬e)p(h|e)–p(h)ifp(h|e)≥p(h)1–p(h)Z(h,e)=p(h|e)–p(h)ifp(h|e)ymeansthatxis(strictly)preferredtoy:Voter1:x>y>zVoter2:y>z>xVoter3:z>x>y\n144GabriellaPigozziSincetherearetwopeoplewhopreferxtoy,twovoterswhopreferytozandtwopeoplewhopreferztox,weobtainthefollowingsocialordering:x>y>z>x.Thisoutcomecannotbeacceptedasitiscyclic.Thelessonisthat,whenwecombineindividualchoicesintoacollectiveone,wemaylosesomethingthatheldattheindividuallevel,liketransitivityorlogi-calconsistency.Kornhauser[11]noticesthatthedoctrinalparadoxresemblestheCondorcetparadox,butclarifiesthatthetwoparadoxesarenotequivalent.Indeed,asstatedbyListandPettit:[W]hentranscribedintotheframeworkofpreferencesinstancesofthediscursivedilemmadonotalwaysconstituteinstancesoftheCondorcetparadox;andequallyinstancesoftheCondorcetparadoxdonotalwaysconstituteinstancesofthediscursivedilemma.([16],pp.216-217)Besidestheimpossibilityofaone-to-onemappingbetweeninstancesofthediscur-sivedilemmaandthoseoftheCondorcetparadox,forKornhauserandSager[12]thedistinctivedifferencebetweenpreferenceandjudgmentaggregationisthat,whenanindividualexpressesapreference,she“speaksonlytoherownvaluesandadvantage”([12],p.85).Sotwopeoplemaydisagreeintheirpreferenceswithouteitherofthembeingwrong.Ontheotherhand,whenapersonmakesajudgment,sheisstatingheropinionaboutthetruth.Iftwopeopledisagreeintheirjudgments,theyacknowledgethattheymaybewrong.Ifpreferencesareneithertrueorfalse,judgmentsare.Thus,expressingpreferencesandjudgmentsarediverseactivities.Indeed,judgmentaggregationprocedurescanbeevaluatedwithrespecttotheirtruth-trackingproperties,asin[14,2,9]Itisworthmentioningthat,adecadebeforejudgmentaggregationmadeitsappearance,Sencriticizedtheclassicalframeworkofsocialchoiceforbeingtoogeneralandabstract,andthusunabletocapturethespecificitiesofdifferenttypesofaggregation:Itcanbearguedthatsomeofthedifficultiesinthegeneraltheoryofsocialchoicearisefromadesiretofitessentiallydifferentclassesofgroupaggregationproblemsintooneuniformframeworkandfromseekingexcessivegenerality.Analternativeistoclassifytheseproblemsintoanumberofcategoriesandtoinvestigatetheappropriatestructureforeachcategory.([24],p.53)Inparticular,Sendistinguishesbetweentheaggregationofindividualinter-estsandofindividualjudgments.Acommitteedecisionisconcernedwiththeaggregationoftheviewsofitsmembersonsomealternativeproposals,whereastheexerciseofinterestaggregationisconcernedwiththeaggregationoftheper-sonalwelfarelevelsofthedifferentpeopleinvolved.Sen’snotionofindividualjudgmentsisthusthatofjudgmentrankingsofalternativepolicies,ratherthanoflogicallyconnectedpropositions.Senclaimsthatessentiallydiverseapproachesareneededfordifferentclassesofaggregationproblems.Theconclusionisare-examinationoftheimpossibilityresultsinsocialchoicetheory,whichwouldgobeyondthepurposeofthisarticle.Whatinterestsushereisthewarningtoconsider\nAggregationProblemsandModels145thekindsofinformationtobecombinedinseveralcontextsbesidestheformalsim-ilaritiesthatsuchsituationsmaydisplay.Preferenceandjudgmentaggregationarefundamentallydiverseexercisesinthenatureoftheinformationtheycombineandalsobecausetheirinputshavedifferentstructures,asweshowinthenextsection.2.1TheindependenceconditionArrow’sfamousimpossibilitytheorem[1]showsthattheculpritoftheCondorcetparadoxisnotthepairwisemajorityvoting.Indeedtheproblemismoregeneralasthereexistsnofunctionthatassignsacollectivepreferenceorderingtoasetofindividualpreferenceorderings,andthatmeetssomeminimalconditions.Simi-larly,ListandPettit[15,16]showthatthedoctrinalparadoxdoesnotdependonthespecificchoiceoftheaggregationprocedure.Rather,byrephrasingArrow’spropertiesinthelogicbasedmodelofjudgmentaggregation,theyproveageneralimpossibilitytheoremstatingthatthereexistsnoaggregationfunctionthatsatisfiesaminimalsetofdesirableconditions.3TherelationsbetweenthetwoaggregationframeworkshavebeenexploredbyListandPettit[16]and,laterandmorethoroughly,byDietrichandList[7],whoprovethatArrow’stheorem(forstrictpreferences)isacorollaryofoneoftheimpossibilityresultsinjudgmentaggregation.SoDietrichandListconcludethattheArrowianpreferenceaggregationisaspecialcaseofjudgmentaggregation.AproposalforaunificationofthetwoframeworkshasbeenlatelyputforwardbyGrossi[8],whoshowsacorrespondencebetweenpreferenceaggregationandasubclassofjudgmentaggregationproblemsinamany-valuedlogic.Undeniablythetwoframeworkssharesomefeatures.Nevertheless,formaltranslationsofonemodelintotheothershouldnotovershadowthedifferencesoftheaggregationproblemsthetwodisciplinesaimtocapture.AsrecalledinSection2,notonlytheCondorcetparadoxandthediscursivedilemmaarenotequivalent,buttheyalso(asmaintainedbyKornhauserandSager)combinedifferenttypesofinformation.Inaddition,wenowwanttoillustratethattheinputsofpreferenceandjudgmentaggregationhavedifferentstructures.Hence,conditionsthatcanbereasonablyimposedinoneframeworkresulttobeoddintheother.OneofthepropertiesinArrow’stheoremistheindependenceofirrelevantalternatives,thatwarrantsthatthegrouprankingoveranypairofalternativesde-pendssolelyonthepersonalrankingsoverthesamepairofalternatives.Theintuitionisthatthesocialrankingover,forexample,xandyshouldbedeter-minedexclusivelyonhowtheindividualsrankxcomparedtoyandnotonother(irrelevant)alternatives,likez.Thisrequirementhasbeenintroducedinjudgmentaggregationastheindependencecondition.Thisensuresthatthecollectivejudg-3Forotherimpossibilitytheoremsthatstrengthenedandexpandedtheoriginalformula-tion,see[17].\n146GabriellaPigozzimentoneachpropositiondependsexclusivelyontheindividualjudgmentsonthatproposition(andnotonother–assumedtobeindependent–propositions).4Yet,webelievethat(besidethedifferenttypesofinformationtobeaggre-gated)thefeaturethatmostdistinguishesjudgmentaggregationfromotheraggre-gationexercisesamountstomakingadecisionontheconclusionwhileprovidingreasonsinsupportofthatdecision.Noequivalentdistinctionbetweenpremisesandconclusioncanbetracedinpreferenceaggregation.Thetwoaggregationprob-lemsarestructurallydifferent,besidesthefactthattheycombinedifferenttypesofinformation.Importingtheindependencepropertyintothemodelsforjudgmentaggregationcontributestoneglectthisfeature.Indeed,thedistinctionbetweenpremisesandconclusiondoesnotplayaroleinthelogicalmodelsofjudgmentaggregation[22].Suchadivisioniscrucialandunderminestheindependenceconditionsincepremisesareoftenindependentfromeachother,buttheyareneverindependentfromtheconclusion(andviceversa).5Inordertoaggregatepremisesandconclusions,weneedtobetterunderstandtheirrelations.Thisamountstoin-vestigatethejustificationfortheindependenceconditionandeventuallytoproposeweakened(oralternative)conditions,whichmayprovideescaperoutesfromtheimpossibilityresultsthatplaguethediscipline.63THEDOCTRINALPARADOXANDOTHERAGGREGATIONPARADOXESThediscursivedilemmadoesnotresembleonlytheCondorcetparadox.Listre-markstheanalogiesbetweenthedoctrinalparadoxandotherwell-knownvotingparadoxes,andboundstheessenceofthediscursivedilemmatothelogicalrela-tionsamongthepropositions:ThedoctrinalparadoxisrelatedtoAnscombe’sparadox,orOstrogorski’sparadox[...].Likethedoctrinalparadox,theseparadoxesareconcernedwithaggregationovermultiplepropo-sitions.Unlikethedoctrinalparadox,theydonotincorporateexplicitlogicalconnectionsbetweenthepropositions.([14],p.4)InthissectionthediscursivedilemmaiscomparedwiththeOstrogorski’spara-doxandtheparadoxofmultipleelections.Theaimofthefirstcomparisonisto4Herewebrieflyrecallalsotheothertwoconditionsimposedonanyjudgmentaggre-gationfunctionF:universaldomainandanonymity.ThefirstensuresthatFacceptsasinputsallconsistentandcompleteindividualjudgmentsets,whileanonymityguar-anteesthatallindividuals’judgmentsaregivenequalweightintheaggregation.5TheurgeforatheoryofjudgmentaggregationonnormativelydefensibleconditionshasbeenclaimedbyMongin[18].Monginrecognizesthatpropositionalformulasarenotindependentwhentheysharepropositionalvariables.ThisleadsMongintoweakentheindependencecondition.Nevertheless,hisnewindependenceconditionisnotweakenoughtoensurepossibilityresults.6Howthedistinctionbetweenpremisesandconclusioncanbeintroducedinaweakenedindependenceconditionhasbeenstudiedin[22].\nAggregationProblemsandModels147showthat,byfocusingexclusivelyonthelogicalconnectionsbetweenproposi-tions,otherrelevantaspectsofthejudgmentaggregationproblemmaybemissed.Aswehaveseen,thedoctrinalparadoxisthebenchmarkexampleofjudgmentaggregation.However,asillustratedbythemultipleelectionparadox,collectiveirrationalityshouldnotbetheonlyworryofjudgmentaggregation.LetusstartwiththeOstrogorski’sparadox.7Ostrogorskiwasconcernedwiththedemocraticgovernance.Hearguedthatindividualsshouldvotedirectlyforpoliciesandnotforpoliticalparties.TheOstrogorski’sparadoxhintsatthedis-tortionsthatpoliticalpartiesproducewhenindividualsarenotallowedtoexpresstheiropinionsdirectlyonthepolicies[23,10].Consideratwo-partycontest(governmentandopposition)andthreeissues(economic,environmental,international).Thetwopartieshaveoppositeviewsontheissues,andeachindividualcastsavote(yesorno)dependingonwhethershewishesapolicychangeonthatissue(sosheagreeswiththeoppositiononthatissue),orthegovernmentrepresentsheropiniononthatmatter(nopolicychange).Eachvotervotesforthegovernment(resp.opposition)ifsheagreeswiththegovernment(resp.opposition)onamajorityoftheissues.SupposetherearefivevotersandthattheyvoteasinTable2.EconEnvIntPartyVoter1noyesnonoVoter2noyesnonoVoter3yesnononoVoter4yesyesyesyesVoter5yesyesyesyesMajorityyesyesnonoTable2:Ostrogorski’sparadoxLikethedoctrinalparadox,theOstrogorski’scaseispuzzlingbecause,despitetheindividualsbeingrational,thecollectiveoutcomeisinconsistent.Ifeachvotervotesforthepartywithwhichsheagreesonamajorityofissues,thegovernmentwins.However,theoppositionrepresentstheviewsofthemajorityofthevotersontheissues,specificallyontheeconomicandtheenvironmentalpolicies.Theinterdependenceofthepropositionsinthetwoparadoxesarenotofthesamekind.IntheOstrogorski’sparadoxitisacompoundmajoritydecisionthatbindspartiesandissues,whereasinthediscursivedilemmathepropositionsarelogicallyconnected.Nevertheless,thepolicies-partydivisionintheOstrogorski’sparadoxissimilartothepremises-conclusiononeinthedoctrinalparadox.Thesupportedparty/conclusiondependsontheopinionsexpressedonthepolicies/7Theanalysisisbasedon[21],thoughthescopetherewasbroader:anaggregationoperatorwasrecommendedtodealwiththeOstrogorski’sandthedoctrinalparadox.\n148GabriellaPigozzipremises,whereasthepolicies/premisesareindependentfromeachother(anycombinationofjudgmentsisadmissible).Thedifficultywiththeaggregationproblemsisthatthesetofpropositionsonwhichmostgroupmembersagreeisnotguaranteedtobeacandidateforthecollectivedecision.Thesetcanfailtosatisfythedependencerelationsamongtheitemseventhougheachmemberconsistentlyexpressedherjudgmentsorvotes.ThecomparisonwiththeOstrogorski’sparadoxshowsthatthekindofdependencebetweentheitemsoftheagendatobevotedondoesnotneedtobeofalogickindtoleadtoaparadoxicaloutcome.Howcompellingarethedependencerelationsamongtheissuesforaparadoxtoarise?Toanswerthisquestionitisinstructivetoconsideralastvotingparadoxtraditionallystudiedinsocialchoicetheory:theparadoxofmultipleelections.Inamultipleelectionvotersarerequestedtovoteonseveralissues,andtheymayvoteonanissuewithoutknowingtheresultoftheelectiononthepreviousone.Brams,KilgourandZwicker[4]introducetheparadoxasfollows:ConsiderareferenduminwhichvoterscanvoteeitherYorNoneachpropositionontheballot.Theparadoxofmultipleelectionsoccurswhenthecombinationofpropositionsthatwinsreceivesthefewestvotes,oristiedforfewest.([4],p.213)AninstanceofthisphenomenonisinTable3.Ifvotesareaggregatedseparatelyoneachissue,thewinningcombinationisyes-yes-yes.Althoughyes-yes-yesissociallyacceptable,nosingleindividualvotedthatcombination.Twopeoplevotedyes-yes-no,twovotedyes-no-yes,fourindividualsvotedno-yes-yesand,finally,twovotedyes-no-no.Thus,thesocialoutcomeisasequenceofalternativesthatdoesnotcorrespondtothevoteofanyindividual.Toputitmoredramatically,theoutcomedoesnotreflectthewillofanymemberoftheelectorate.Voter1yesyesnoVoter2yesyesnoVoter3yesnoyesVoter4yesnoyesVoter5noyesyesVoter6noyesyesVoter7noyesyesVoter8noyesyesVoter9yesnonoVoter10yesnonoMajorityyesyesyesTable3:TheparadoxofmultipleelectionsEvenmorealarmingisthe“complete-reversalparadox”where,forexample,theissuebyissueaggregationselectsyes-yes-yes-yeswhereasthecombinationthat\nAggregationProblemsandModels149receivesthemostvotesisthe‘opposite’no-no-no-no(andyes-yes-yes-yesreceivesnovote).8Brams,KilgourandZwickerprovidenecessaryandsufficientconditionfortheparadoxtoappearandwereferthereadertotheirpaperforthefullanalysis.Herewewanttodrawtheattentiontothefollowingquestion:shouldanoutcomeliketheonesintheparadoxofmultipleelectionsbeacceptableinjudgmentaggregation?Supposethatpropositionwisemajorityvotingselectsacollectivejudgmentsetthatsatisfiesthedecisionrule,butthatdoesnotcorrespondtoanyoftheindividuals’judgmentsets.Shouldthisbeasourceofconcernforjudgmentaggregation?Ifwefindthediscrepancybetweenthewinnerunderpropositionaggregationandtheonechosenbycombinationaggregationdisconcertingintheparadoxofmultipleelections,weshouldevenmorefindsowhentobeaggregatedarejudgmentsetsinsteadof(independent)issues.Webelievethatcollectiveirrationalityshouldnotbetheonlyconcernofjudg-mentaggregation,andthediscussiononhowtoaggregateindividualjudgmentsetsshouldnotbelimitedtoavoidthediscursivedilemma.Aswehaveseen,judgmentaggregationwouldnotrejectaconsistentcombinationofreasonsandconclusionthatnomembervotedfor.Yet,likeintheparadoxofmultipleelections,thismaynotbeadesirablesolution.SupposethatthethreejudgesofSection1convictthedefendantforreasonsthatnoneofthemsubmitted.Wouldsuchaverdictbeac-ceptable?Inordertotacklethisproblem,amoreholisticnotionofresponsivenessisrequired.Thisobservationwasoneofthemotivationsforaninvestigationofjudgmentaggregationinabstractargumentation[5],wherethreeaggregationop-eratorsthatguaranteeasocialoutcome‘compatible’withtheindividualjudgmentsareintroduced.A‘compatible’groupdecisionissuchthatanygroupmemberisabletodefendthegroupdecisionwithouthavingtoargueagainsthisownopin-ions.Notonlyacollectiveinconsistentoutcomewouldnotbeaccepted,butneitherwouldbeagroupoutcomethatisnotcompatiblewithallthegroupmembers.Thenewmodelsshouldaimatcapturingthespecificityoftheseriddlesratherthanflatteningthemtopreferenceaggregationproblems.4CONCLUSIONInthispaperwehaveconsideredarecentaggregationproblemanditsparadox:thediscursivedilemma.Thedoctrinalparadoxappearswhenseeminglynatu-ralaggregationrulesareappliedtoindividualjudgmentsonlogicallyconnected8Nurmiconceivesthattheparadoxofmultipleelectionsisparticularlytroublesomewhentheissuesarerelatedtoeachother(i.e.“nonseparable”,intechnicalterms).Yetheclaimsthat“Iftheissuestobevoteduponareseparable,thentheparadoxjustamountstopointingoutthatnoindividualisexactlyliketheelectorateconsideredasawhole.ItcouldevenbearguedthattheoccurrenceoftheparadoxamountstotherebeingnoArroviandictator.Thiscertainlyshouldnotbotherusverymuch.”([19],p.85)\n150GabriellaPigozzipropositionstoobtainacollectivejudgmentonthesamepropositions.Judgmentaggregationisanoveldisciplinewhoserelationswithaggregationproblemsstud-iedinsocialchoicehavebeenexploredintheliterature.BecauseofthesimilaritieswithpreferenceaggregationandtheresemblanceofthedoctrinalparadoxwiththeCondorcetparadox,theformalmodelsdevelopedsofarforjudgmentaggregationcombinelogicwithanaxiomaticapproachintheArrowianspirit.Herewehavemaintainedthatthedifferencesbetweenpreferenceandjudg-mentaggregationdeservemoreattentioninthehopetodefinemodelsforjudg-mentaggregationthatcapturethespecificitiesoftheproblem.“Allmodelsarewrong,butsomeareuseful”,saidthestatisticianGeorgeBox[3].Modelsnec-essarilyprivilegesomeaspectsoverothers,soacertainamountofidealizationisunavoidable.Theytrytocapturewhatthemodelerbelievestobetheessenceofacomplexphenomenon,oratleastitsrelevantaspects.However,evenifallmodelswerewrong,somearemoreadequatethanothers.Thefactthatanaggregationprocedurethatisnormativelydefensibleisnotmeaningfulbecausemayproduceirrationaloutcomesissharedbytheparadoxesinjudgmentandinpreferenceaggregation.Wearguedthatthesimilaritiesbe-tweenthetwoareasdonotjustifytheimpositionofthesameconditionsontheiraggregationprocedures.First,judgmentaggregationcombinesindividualevalua-tionsonpropositionsratherthanindividualpreferences.AsKornhauserandSagerobserved,unlikeforthecaseofjudgments,thelatterdonotconveyanytruthvalue.Second,judgmentandpreferenceaggregationdifferinthetypeofinputstheyac-cept.Judgmentsetsarevectorsofyes/noitemsonpremisesandconclusion,wherethepremisesaretypicallyindependentoneachotherwhiletheconclusiondependsonthepremises.Suchadistinctionismissinginthepreferencerealmand,bycon-sequence(andunfortunately),intheformalmodelsofjudgmentaggregationthathaveborrowedtheindependenceconditionfromsocialchoicetheory.Thirdandfinally,weraisedthequestionwhetherthelogicalrelationsamongpropositionsconstitutethemainfeatureofjudgmentaggregation.Astheparadoxofmultipleelectionsshows,agroupoutcomethatdoesnotviolateanylogicalconstraintmaybeperceivedasunacceptableifnomembersubmittedit.Thus,focusingexclu-sivelyonthedoctrinalparadoxmayovershadowotherrelevantissuesforjudgmentaggregation.Judgmentaggregationandpreferenceaggregationaredifferentclassesofprob-lemsandtoexpectthemtofitintoauniformframeworkimposesexcessivegener-ality.Webelievethatthedifferencesthatwehavehighlightedherecallfordistinctmodels.Judgmentaggregationproblemsdisplayfeaturesthatneedtobeproperlymodeledwithoutimposingexogenousframeworks.\nAggregationProblemsandModels151REFERENCES[1]K.Arrow.SocialChoiceandIndividualValues.CowlesFoundationMono-graphSeries,1963.[2]L.BovensandW.Rabinowicz.Democraticanswerstocomplexquestions.Anepistemicperspective.Synthese,150:131-153,2006.[3]G.Box.Robustnessinthestrategyofscientificmodelbuilding.InR.LaunerandG.Wilkinson,editors,RobustnessinStatistics,pages201-236.Aca-demicPress:NewYork,1979.[4]S.Brams,D.Kilgour,andW.Zwicker.Theparadoxofmultipleelections.SocialChoiceandWelfare,15(2):211-236,1998.[5]M.CaminadaandG.Pigozzi.Onjudgmentaggregationinabstractargumen-tation.AutonomousAgentsandMulti-AgentSystems,forthcoming.[6]F.Dietrich.Ageneralisedmodelofjudgmentaggregation.SocialChoiceandWelfare,28(4):529-565,2007.[7]F.DietrichandC.List.Arrow’stheoreminjudgmentaggregation.SocialChoiceandWelfare,29(1):19-33,2007.[8]D.Grossi.Unifyingpreferenceandjudgmentaggregation.In:ProceedingsoftheEighthInternationalConferenceonAutonomousAgentsandMulti-AgentSystems(AAMAS’09),Budapest,Hungary,2009.[9]S.Hartmann,G.Pigozzi,andJ.Sprenger.Reliablemethodsofjudgmentaggregation.Workingpaper,2009.[10]J.Kelly.TheOstrogorskiparadox.SocialChoiceandWelfare,6:71-76,1989.[11]L.Kornhauser.Modelingcollegialcourts.II.Legaldoctrine.JournalofLaw,Economics,andOrganization,8:441,1992.[12]L.KornhauserandL.Sager.Unpackingthecourt.YaleLawJournal,96:82-117,1986.[13]L.KornhauserandL.Sager.Theoneandthemany:Adjudicationincollegialcourts.CaliforniaLawReview,81:1-51,1993.[14]C.List.Theprobabilityofinconsistenciesincomplexcollectivedecisions.SocialChoiceandWelfare,24:3-32,2005.[15]C.ListandP.Pettit.Aggregatingsetsofjudgments:Animpossibilityresult.EconomicsandPhilosophy,18:89-110,2002.[16]C.ListandP.Pettit.Aggregatingsetsofjudgments:Twoimpossibilityre-sultscompared.Synthese,140:207-235,2004.[17]C.ListandC.Puppe.Judgmentaggregation:Asurvey.In:P.Anand,C.Puppe,andP.Pattanaik,editors,OxfordHandbookofRationalandSocialChoice.OxfordUniversityPress,2009.[18]P.Mongin.Factoringouttheimpossibilityoflogicalaggregation.JournalofEconomicTheory,141(1):100-113,July2008.[19]H.Nurmi.VotingParadoxesandHowtoDealwithThem.Springer,1999.[20]P.Pettit.Deliberativedemocracyandthediscursivedilemma.PhilosophicalIssues,11:268-299,2001.\n152GabriellaPigozzi[21]G.Pigozzi.Twoaggregationparadoxesinsocialdecisionmaking:TheOs-trogorskiparadoxandthediscursivedilemma.Episteme,2(2):119-128,2005.13[22]G.Pigozzi,M.Slavkovik,andL.vanderTorre.Conclusion-basedprocedureforjudgmentaggregationsatisfyingpremiseindependence.InG.Bonanno,B.Lowe,andW.vanderHoek,editors,ProceedingsoftheEighthInter-nationalConferenceonLogicandtheFoundationsofGameandDecisionTheory,LOFT2008,p.35,Amsterdam,TheNetherlands,2008.[23]D.RaeandH.Daudt.TheOstrogorskiparadox:Apeculiarityofcompoundmajoritydecision.EuropeanJournalofPoliticalResearch,4(4):391-398,1976.[24]A.Sen.Socialchoicetheory:Are-examination.Econometrica:JournaloftheEconometricSociety,pp.53-89,1977.[25]A.K.Sen.CollectiveChoiceandSocialWelfare.HoldenDay,1970.IndividualandCollectiveReasoningUniversityofLuxembourgComputerScienceandCommunicationrueR.Coudenhove-Kalergi6L-1359LuxembourgLuxembourggabriella.pigozzi@uni.lu\nPartIII(TeamB)(teamleaderMarcelWeber)ApproachestotheFoundationsofScience:thePlaceoftheLifeSciences\n\nMARCELWEBERLIFEINAPHYSICALWORLD:THEPLACEOFTHELIFESCIENCES1.PHYSICALISM,BIOLOGY,ANDREDUCTIONISM:STATEOFTHEARTDebateabouttheplaceofthelifescienceswithintheempiricalscienceshasoftencenteredaroundtheissuesofphysicalismandreductionism.1Giventhatsomeformofphysicalismiscorrect,whyisbiologicalsciencenotphysicalscience?Whydobiologicaltheoriesappeartobeautonomousandirreducibletophysicaltheories?Andwhatisthenatureofbiologicallawsorregularities,assumingthatthefundamentalinteractionsthatgovernthephysicalworldalsoareatworkinlivingorganisms?Thesearesomeoftheoldestandmostextensivelydiscussedquestionsconcerningthebiologicalsciences.WhilephilosophersofscienceofaLogicalEmpiricistbentfirsttriedtodefendtheviewthatbiologicaltheoriessuchasthoseofclassicalMendeliangeneticsareinprinciplereducibletophysi-cal-chemicaltheories,2ananti-reductionistconsensusemergedduringthe1970s.3Thisconsensuswasmainlybasedontheargumentthatgeneticconceptssuchasdominanceorthegeneconceptitselfcannotberedefinedinanextensionallyequivalentwayintermsofmolecularconcepts.Thereasonforthisisthoughttolieinthefunctionalcharacterofbiologicalconcepts.Thismeansthatcertaintheoreticallysignificantpropertiesinbiologyareindividuatedbytheircausalrole,notsomeintrinsicstructuralproperty.Butthemolecularrealizersofthesecausalrolesarehighlyheterogeneousatthemolecularlevel;inotherswords,therealizersdon’thaveatheoreticallysignificantmolecularpropertyincommonthatcouldbeusedtoeliminatethehigher-levelterms.Therefore,higher-levelconceptsinbiologyremainexplanatorilyindispensable;theyhaveautonomousexplanatoryvaluethatcannotbereproducedbymoleculartheoriesalone.Thus,onthisby1See,e.g.,ErnstMayr,ThisisBiology.Cambridge,Mass.:HarvardUniversityPress1997.2KennethF.Schaffner,“ApproachestoReduction”,in:PhilosophyofScience34,1967,pp.137-147;KennethF.Schaffner,“TheWatson-CrickModelandReductionism”,in:BritishJournalforthePhilosophyofScience20,4,1969,pp.325-48.3DavidL.Hull,PhilosophyofBiologicalScience.EnglewoodCliffs:PrenticeHall1974;PhilipKitcher,“1953andAllThat.ATaleofTwoSciences”,in:ThePhilo-sophicalReview93,3,1984,pp.335-373;AlexanderRosenberg,TheStructureofBio-logicalScience.Cambridge:CambridgeUniversityPress1985;cf.C.KennethWaters,“WhytheAnti-ReductionistConsensusWon’tSurvivetheCaseofClassicalMende-lianGenetics”,in:PSA1990,EastLansing,Mich.:PhilosophyofScienceAssociation1990,pp.125-139.F.Stadler(ed.),ThePresentSituationinthePhilosophyofScience,ThePhilosophyofScienceinaEuropeanPerspective,DOI10.1007/978-90-481-9115-4_12,©SpringerScience+BusinessMediaB.V.2010\n156MarcelWebernowreceivedviewinphilosophyofbiology,biologicaltheoriesareirreducibleforbasicallythesamereasonthatmostphilosophersacceptasthedefinitiverefutationofmind-brainreductions.Thus,philosophersofbiologyhavereachedsimilarconclusionsasphiloso-phersofmindhaveinregardoftheissueofmind/brain-reductionism.4However,JaegwonKim5hasarguedthat,inthephilosophyofmind,thisconsensusisbasedonaninadequatemodelofreduction,namelyErnestNagel’s.6Heproposedanalternativeschemeaccordingtowhichreductiondoesnotconsistinfirstconnect-ingthetermsofthetheorytobereducedtothoseofthereducingtheorybywayofbiconditionalbridgeprinciples(asNagel’smodelassumesoriswidelytakentoassume),followedbythederivationofthelawsofthetheorytobereducedfromthelawsofthereducingtheory.Instead,Kimarguesthatsuccessfulreductionsmustfirstgiveafunctionalcharacterizationofthereferentsofthetermsofthetheorytobereduced.Suchacharacterizationspecifiesthesetofthingsthatcomeunderaconceptbystatingthecausesand/ortheeffectsthatthesethingshaveintheircontainingsystem.Next,scientistsmustidentifythethingsthatplaythesecausalrolesatthelowerlevel.Forexample,Kimthinksthatthecaseofgeneticsprovidesaparadigmforthiskindofreduction.Geneswerefirstidentifiedbythecausalrolestheyplayinlivingorganisms,namelycausingheritablecharacterdif-ferences,beingsegregatedandassortedinaccordancewithMendel’slaws,etc.etc.Later,itwasdiscoveredthatthesecausalrolesareactuallyfulfilledbyDNAsequencesthatcodeforproteinand/orRNAmolecules.7Thisisareduction;noth-ingmoreisrequired.ItakeitthatKimdoesnotrequirethatscientistbeabletostatenecessaryandsufficientphysical(molecular)conditionsforsomethingtoinstantiateatheoreticallysignificanthigher-levelproperty,forifhedid,hismodelwouldbasicallycollapseintoNagel’s(orwhatisusuallytakentobeNagel’s).Allthatherequiresisthattherealizersofthecausalrolethatdefinesthehigher-levelpropertybesomehowdescribablefromthephysicallevel(headmitsthatthismaynotalwaysbepossible,forexample,hethingsitisnotpossibleforqualia).Kim’ssuggestionhasnotsucceededindisplacingtheanti-reductionistcon-sensusinthephilosophyofbiology;infact,itwashardlynoticedbyphilosophersofbiology.However,itisclearwhattheirresponsewouldbe:EvenifKim’snewmodelofreductionisaccepted,thatthemolecularrealizersofsomefunctionally4DonaldDavidson,“MentalEvents”,in:L.Foster/J.Swanson(eds.),ExperienceandTheory.Amherst,Mass.:UniversityofMassachusettsPress1970;JerryA.Fodor,“SpecialSciencesortheDisunityofScienceasaWorkingHypothesis”,in:Synthese28,1974,pp.97-115.5JaegwonKim,MindinaPhysicalWorld:AnEssayontheMind-BodyProblemandMentalCausation.Cambridge,Mass.:MITPress1998.6ErnestNagel,TheStructureofScience.ProblemsintheLogicofScientificExplana-tion.London:RoutledgeandKeganPaul1961.7C.KennethWaters,“GenesMadeMolecular”,PhilosophyofScience61,1994,pp.163-185.\nLifeinaPhysicalWorld157individuatedbiologicalconceptcanbedescribedatthemolecularlevelaloneisexactlywhatisnotpossibleaccordingtotheanti-reductionistconsensus.Onthestandardargumentfrommultiplerealizability,suchadescriptionwouldinvolveanungainlydisjunctivepredicatewithoutanyexplanatoryforce.Thisiswhythehigher-leveltheoriesareexplanatorilyindispensable.Tothisreply,aKim-stylereductionistcouldretortthattheidentificationofclassicalgeneswithprotein-andRNA-codingDNAsequencesisnotungainlyatall.Understandingwhatgenesareatthemolecularlevelispreciselywhatmo-lecularbiologyhasdoneforgenetics,andifthisdoesnotaccountasareduction,thennothingdoes.However,thisreductionistresponsemissesthatreductionissupposedtoatleastconservetheexplanatoryachievementsofthetheorytobereduced,inadditiontoprovidingexplanationsthatexceedthoseofthetheorytobereduced.Butthisisnotthecaseinthegenetics/molecularbiologycaseaccord-ingtoantireductionists.8Classicaltransmissiongeneticsoffersexplanationsofinheritancepattersthatbasicallycitethepairingandseparationofchromosomes.Theseexplanationsabstractawayfromthe“gory”moleculardetailsthatconstitutetheseprocessesatbottom.Kitcher9drawsananalogyheretoPutnam’swellknownsquarepeg-inaroundhole-argument.Accordingtothisargument,thereisaper-fectlyfineexplanationofwhysquarepegsdon’tfitinroundholesthatappealsonlytotheseobjects’geometricalshape.Thisexplanationabstractsawayfromthecompositionoftheobjectsandfromanyphysicallawsthatthesemayobey.Infact,thisexplanationismoregeneralthananyexplanationthatappealstotheobjects’composition.Kitchersuggeststhatthisisanalogoustotheexplanationsthatclassicalgeneticsgiveofinheritancepatterns.Therearedifferentrepliesthatareductionistcangivetothisargument.First,itcanbearguedthatthetheoreticalcontentofclassicalgeneticsisnotexhaustedbypatternsofgenetransmission.ClassicalgeneticistsdescribedgeneticstructureswiththehelpofelaboratemapslongbeforemoleculartechniquessuchasDNA-sequencingbecameavailable.Forinstance,itwaspossibletoshowthatgenesmustbelinearstructures,afindingwhichwasconfirmedbythediscoveryofthewayinwhichDNAencodesgeneticinformation.10ThisfitsnicelywithKim’smodelofreduction.AsecondpossibleresponseisthatKitcher’sargument–justlikePutnam’s–isamanifestationofatheoreticallyunfounded“explanatoryProtagoreanism”,11accordingtowhich“somehumanorotheristhemeasureofallputativeexplana-tions,ofthosewhichdoexplainandofthosewhichdonot.”WhileKitcher’s8Kitcher,op.cit.9Kitcher,op.cit.,p.35010MarcelWeber,“RepresentingGenes:ClassicalMappingTechniquesandtheGrowthofGeneticalKnowledge”,in:StudiesinHistoryandPhilosophyofBiologicalandBiomedicalSciences29,2,1998,pp.295-315.11AlexanderRosenberg,DarwinianReductionism.Or,HowtoStopWorryingandLoveMolecularBiology.Chicago:TheUniversityofChicagoPress2006,p.35.\n158MarcelWeberchromosomalmechanicsexplanationsorPutnam’ssquarepeg-in-a-round-hole-explanationmayseemperfectlysatisfactorytosomepeople,perhapsrelativetocertainpragmaticcontexts,itisnotanexplanationthatwouldsatisfyaphysicistoramolecularbiologist.Scienceoughttodobetterthanthat,forexample,byshowingexactlywhatforcespullthechromosomesapartbeforeacelldivides,orwhatforcesrepelthepegfromthehole,takingintoaccounttheircomposition,ofcourse.Here,thereductionism/antireductionismdebateturnsondivergentas-sumptionsastowhatconstitutesagoodexplanationofaphenomenon–amatteronwhich,naturally,reductionistsandantireductionistshavedifferentintuitions.12Theseargumentsandcounter-argumentsarewellknownandhavebeendis-cussedintheliteratureadnauseam.Itisbeyondthescopeofthispapertopresentallthetwistsandturnsofthereductionism/antireductionismdebate,oreventolayoutthevariouspositionsthathavebeendefended,fromstrongformsofreduction-ismtonon-reductivephysicalism,emergentism,scientificpluralism,andsoon.13Instead,whatIwouldliketodohereistoexaminesomenovelarguments,whichhavereceivedlittleattention.Ithinkthatbothattempts,whileperhapsnotsuccess-ful,containsomegenuineinsightswithrespecttotheplaceofthelifesciencesintheconceptuallandscapeofthenaturalsciences.ThefirstviewIwanttocriticallyreviewisanattempttodefendofastrongformofreductionismaboutbiologythatcanbefoundinAlexRosenberg’srecentbook.14IwillshowwhyRosenberg’saccountfails,eventhoughitcontainsavalu-ableinsightconcerningtheroleoftheconceptoffunctioninbiology,namelyintheindividuationoftraits.Rosenbergthinksthatthismakesallofbiologycon-ceptuallydependentonevolutionarytheory,whichisnotgenerallythoughttobereducibletomorefundamentaltheories.Asaresult,anunbridgeablegapthreatensbetweenbiologyandphysicaltheories.Rosenbergtriestoclosethisgapbytryingtoshowthatevolutionarytheory,atleastnaturalselectiontheory,isfundamental.IshallcriticizeRosenberg’spositionontwocounts:First,Iwillshowthattheideathatnaturalselectiontheoryisfundamentalisproblematic(Section2).Second,Iwillarguethattherewasnoproblemforthereductionistinthefirstplace,becausetherearewaysofindividuatingorganismictraitsthatdonotdependontheconceptofnaturalselection(Section3).ThesecondviewIwilldiscussherecomesfromoutsidethephilosophyofbi-ology,namelyfromgeneralmetaphysicsandisitisnotveryrecent,butithasbeenhardlynoticedbyphilosophersofbiologyandofscience:theviewofbiological12PaulHoyningen-Huene,“EpistemologicalReductionisminBiology:Intuitions,Ex-plicationsandObjections”,in:P.Hoyningen-Huene/F.M.Wuketits(eds.),Reduc-tionismandSystemsTheoryintheLifeSciences.Dordrecht:KluwerAcademic1989:pp.29-44.13AnimportantstrandofthisdebateiscriticallyreviewedinThomasReydon’scontribu-tioninthisvolume,namelytheissueofnaturalkindsanditsimplicationsforreduc-tion.14Rosenberg2006,op.cit.\nLifeinaPhysicalWorld159lawsthathasbeendevelopedbyMichaelThompson.15Hethinksthatbiologicallawsdifferfundamentallyfromphysicallaws.Whilethisclaimishardlynew,thespecificdifferencesthatThompsonseesbetweenthetwoclassesoflawshave,tomyknowledge,notbeennoticedinthephilosophyofbiology.EventhoughIdisagreewithsomepartsofThompson’saccount,Ibelievethatitmeritsseriousdiscussion,whichIshallattemptinSection4.2.ROSENBERG’SDEFENSEOFREDUCTIONISMANDWHYITFAILSInhisrecentbook,16Rosenbergfirmlyadherestotheviewthat“nothinginbiol-ogymakessenseexceptinthelightofevolution.”EvolutionarybiologistssuchasErnstMayr17orTheodosiusDobzhansky,18whohavedefendedthisview,basedtheirargumentsontheassumptionthatafullunderstandingoforganismsrequirestheidentificationoftheultimatecausesoftheircharacteristicproperties.TouseMayr’sfavoriteexample,evenifwefullyunderstandthephysiologicalmecha-nismsthatinducemigratorybirdstoflocktogetherandembarkonalongjourneytowardsawarmerclimatezone–i.e.,theproximatecause–afullunderstand-ingofthisbehaviorrequiresanaccountofwhatitwasselectedforinthebirds’evolutionarypast–i.e.,theultimatecause.Onthisreceivedview,proximateandultimateexplanationsarecomplementaryandconceptuallyindependent.Thiscon-ceptualindependenceallowsforthepossibilityofendorsingbothreductionismaboutproximatebiologyandantireductionismaboutevolutionarybiology.Thelatterkindofantireductionismisusuallyjustifiedongroundsofthemultiplereal-izabilityoffitness.19However,accordingtoRosenberg,ultimateandproximatebiologyarenotconceptuallyindependent.Howcouldthisbe?Whycan’tbiologistspinpointanorganism’smolecular,physiological,developmentaletc.mechanismsindepend-entlyofitsevolutionaryhistory?ForRosenberg,thishastodowiththewayinwhichbiologistspicktheexplananda,inotherwords,thatwhichtheywanttoexplainbydiscoveringtheunderlyingmechanisms.Letussay,forexample,thatbiologistswanttounderstandhowchickembryosformwings.‘Wing’isafunc-tionalconcept.Inotherwords,theclassificationofsomestructureasawing,in-cludingitsexactdelimitationfromneighboringstructures,involvesanappealto15MichaelThompson,“TheRepresentationofLife”,in:R.Hursthouse/G.Lawrence/W.Quinn(eds.),VirtuesandReasons.PhilippaFootandMoralTheory.Oxford:Claren-don1995,pp.247-296.16Rosenberg2006,op.cit.17ErnstMayr,“CauseandEffectinBiology”,in:Science134,1961,pp.1501-1506.18TheodosiusDobzhansky,“Biology,MolecularandOrganismic”,in:AmericanZoolo-gist4,4,1964,pp.443-452.19ElliottSober,TheNatureofSelection.EvolutionaryTheoryinPhilosophicalFocus.CambridgeMass.:MITPress1984.\n160MarcelWeberfunction(flightinthiscase).Rosenbergarguesthatthesalientconceptoffunctionheremustbethatofproperfunction,20thatis,functionasselectedeffect.Awingisastructurethatwasselectedbecauseitconferstheabilitytofly.Itisafunctionaltype,and“function”meansproperfunctionaccordingtoRosenberg.Therealizersofthisfunctionaltypeareheterogeneousbecausedifferentstructureswithdiffer-entevolutionaryoriginscanconfertheabilitytofly.Thisiswhytherearealsononaturalkinds(essences)inthetraditionalsenseinbiology,Rosenbergargues.Forselectionisblindtoessences(intrinsicstructure).21Theupshotisthatthewayinwhichanorganismisdividedintopartscruciallydependsonthetheoryofnaturalselection.Sinceproximatebiologytakesitsrequestsforexplanationfromsuchdivisions(“whatmechanismscontrolthedevelopmentofthechickwing?”),itisconceptuallydependentonevolutionarybiology.ThispositionwithrespecttofunctionsandproximatebiologyseemstoputRosenberginthedifficultpositionthat,inordertomaintainhisreductionism,hemustshoweitherthatthetheoryofnaturalselectionisreducibletomorefunda-mentaltheoriesorthatitisitselfafundamentaltheory.Hechoosesthesecondpath:Hearguesthatwhathecallsthe“principleofnaturalselection”isitselfafundamentallaw.Hereisoneformulationofthisalleged“principle”:22∀x∀y∀E[Ifxandyarecompetingorganismsingenerationn,andxisfitterthanyinE,thenprobably(thereissomegenerationn’,atwhichxhasmoredescendantsthany)]Therearealternativeformulations,andRosenbergisawarethatthismaynotbethemostgeneralwayofstatingtheprinciple.Rosenbergtakesthistobeanempiricallaw(incontrasttoSober,23whothinksthattheprincipleofnaturalselectionisapriori)andheunderstandsfitnessintermsofaprobabilisticpropensity.NowforwhatisprobablyRosenberg’sboldestclaim:Hearguesthattheprin-cipleofnaturalselectionisaphysicallaw,orperhapsachemicallaw(orboth).20RuthG.Millikan,“InDefenseofProperFunctions”,in:PhilosophyofScience56,1989,pp.288-302.21ThomasReydon(inthisvolume)arguesthatselectedeffectfunctionsarenotmulti-plyrealizable,becausetheyrequirethatthefunctionbearersstandinanappropriatehistorical(genealogical)relationship,whichmeansthatevensomethingwhichplaystheexactsamecausalroletodaywouldnotcountasaninstanceofthefunctionifitevolvedindependently.Tothis,itcouldberepliedthatnothingpreventsacertainorgantochangeitsinternalstructure(itsessence)inevolutionwhileitcontinuestobenefitfromnaturalselection,sothesetofthingsthathasthesameactivityandstandsintheappropriategenealogicalrelationswouldcountasinstancesofthefunction.Thiswouldcountasmultiplerealization.However,Reydon’spointdoesseemtolimitthemultiplerealizabilityofselectedeffectfunctions.22Rosenberg(2006),op.cit.,p.16023ElliottSober,“TwoOutbreaksofLawlessnessinRecentPhilosophyofBiology”,in:PhilosophyofScience(Proceedings)64,1998,pp.S458-S467.\nLifeinaPhysicalWorld161Insupportofthisclaim,hearguesthateventhingsthatarenotconsideredtobealiveobeythisprinciple,forexample,self-replicatingmolecules.Healsooffersastorywhytextbooksofphysicalchemistrydonotnormallycitethislaw,namely,becausephysicalchemistsnormallyaskdifferentquestions.Butthisdoesn’tprovethatthisisn’tafundamentallawofnatureaccordingtoRosenberg.Rosenbergneedsthisclaiminordertomake“naturalselectionsafeforre-ductionism.”Thereasonis,asIhavealreadyshown,isthatRosenbergthinksthatnaturalselectionviatheconceptofproperfunctionprovidestheexplanandaforbiologicalexplanations,evenoutsideofevolutionarybiology.IwouldliketoaddresstwocriticalpointsatRosenberg’sargument.Thefirstconcernshisclaimthatthereexistsa“principleofnaturalselection”whichisaphysicallaw.Thesecondpointchallengestheclaimthatnaturalselectiontheoryisneededforidentifyingtheexplanandaforbiologicalexplanations.First,letusconsiderRosenberg’salleged“principleofnaturalselection”.Asstated,itisonlyapplicabletopopulationswithdiscretegenerations.Evolutionarytheoristsusedifferentfitnessmeasuresforpopulationswithdiscretegenerationsandforage-structuredpopulationswithoverlappinggenerations.Inoneofmyownworks,Iarguethatifthereisageneralprincipleofnaturalselection,thenitishighlyabstractandneedstobeinstantiatedbyspecificmodels.24Onthisview,thetheoryofnaturalselectionisafamilyofmodels(“semanticview”oftheories)anditscontentisnotappropriatelyexpressedbyauniversallyquantifiedclaim.Universallyquantifiedclaimsonlycomeinwhenitcomestostatingclassesofnaturalsystemstowhichthemodelsapply.Thegeneraltheoryismerelysomesortofaguidelineforbuildingspecificmodels.Rosenbergcouldreplythat,perhaps,hehasnotcorrectlystatedthefunda-mentalprincipleofnaturalselection,butthathispointthatthereexistssuchaprincipleandthatitisafundamentallawofnaturestands.However,Idon’tthinkthathecansustainthisview.Thereasonisthattherearenoreasonstobelievethatthereisafundamentalmeasureofevolutionaryfitness.“Fitness”meansdifferentthings,dependingontheevolutionaryproblemthatbiologistsaretryingtosolve.Sometimes,fitnessisanabsolutegrowthrate.Sometimesitisanabsolutenumberforthesurvivingoffspring.Sometimesitisacoefficientinapopulationgeneticmodelthatmakesexplicitassumptionthegeneticsystem(e.g.,Mendelianinherit-ance).Fitnessispredicatedofgenes,genotypes,individuals,andgroups.Sofar,thereisnounifyingframeworkforevolutionarytheory,andtherearenoreasonstothinkwhythereshouldbeone.Therearedifferentevolutionaryprocessesanddifferentquestionsthatonecanaskaboutthem.Anyfitnessmeasurecanbeusefulforansweringonekindofquestion,butnotanother.24MarcelWeber,DieArchitekturderSynthese.EntstehungundPhilosophiedermod-ernenEvolutionstheorie.Berlin:WalterdeGruyter1998,Ch.6.\n162MarcelWeberIfthereisnofundamentalfitnessmeasure,itfollowsthatthereisnogeneral“principleofnaturalselection”.Andafortiorithereisalsonofundamentallawofnatureaboutnaturalselection.25IthinkthatthisfailureofRosenberg’sattemptisexemplaryforthewholeofbiology.Biologyisnotconcernedwithidentifyinglawsofnatureinthetraditionalsense.Itsgoalisrathertoanswerspecificwhy-questionsbyusingvariouscon-ceptualtools,includinginsomecasesmathematicalmodels.Theanswerstosuchwhy-questionscannotgenerallybeincorporatedintosomeunifiedframework.26Aswehaveseen,theultimatemotivationforRosenberg’saccountofbiologi-callawswashisgoalofshowingthatbiologicaltraitscouldbebothfunctional,intheproperrolesense,andyetphysical.Inthefollowingsection,Ishallexamineiftherearenootherwaysofhowbiologicaltraitscanbeindividuated.3.ANALTERNATIVEACCOUNTOFFUNCTIONSANDTRAITINDIVIDUATIONAswehaveseen,Rosenbergbasedhisdefenseofreductionismontheviewthatbiologicaltraitsareindividuatedfunctionally,where“function”isunderstoodinthesenseofselectedeffectfunctionor“proper”function.Ithinkthefirstpartofthisclaimiscorrect,however,thereisaproblemwiththesecond.Onthisview,someitemXhasafunctionFinorganismSexactlyifXdoesFandthefactthatsomeearliertokensofXhavedoneXisacauseofX’spresenceinS.Thewayinwhichearliertokenscancausethepresenceofsomeiteminlatergenerations,ofcourse,isnaturalselection.Thus,Rosenberg’sviewisthatnaturalselectionisnotonlyneededtoexplainwhysomeorganismScametohaveapartX,buttospeakofXashavingsomekindofunityinthefirstplace.ItisforthisreasonthatRosenbergthinksthatthetheoryofnaturalselectionisfundamentalforthewholeofbiology.This,ofcourse,includesbehavioralbiology.AccordingtoRosenberg,thedescriptionofbehavioraltraitsisladenand/oroughttobeladenbytheoreticalhypothesesaboutselectionhistory.Atraitsuchasawingisindividu-atedbythefactthatitwasselectedforflying,nomatterwhatothercapacitiesitmayhave(forinstance,it’scapacityofbeingflappedsoastodistractorattractsomeotheranimal).Onthisview,descriptionsofanorganism’straitsareladenbythetheoryofnaturalselectionandassumptionsabouttheevolutionarypast.25DanielSirtes(personalcommunication)objectsthatthisargumentatbestprovesthatthereisnosinglefundamentalprinciple;therestillcouldbeoneforeverytypeofevo-lutionaryprocess.However,itseemstomethatsuchahodgepodgeofprinciples–andtherewouldhavetoquitealotofthem–wouldnotdeservethestatusof“fundamental”principles,becausetheywouldallonlybeapplicabletosomerestrictednumberofcases.26Thisclaimisgenerallyknownasscientificpluralism,seeStephenH.Kellert/HelenE.Longino/C.KennethWaters(eds.),ScientificPluralism.MinnesotaStudiesinPhi-losophyofScience,Vol.XIX.Minneapolis:UniversityofMinnesotaPress2006.\nLifeinaPhysicalWorld163PaulGriffiths27hasarguedthatthisviewputsthecartbeforethehorse.Thepartsoforganismsandtheircausalcapacitiesmustbeunderstandableindepend-entlyofnaturalselection.Otherwise,thefollowingregressthreatens:1.Selectedeffectfunctionsareascribedbycausalanalysisofthecapacitiesofthepartsofancestralorganismsandadeterminationoftheirfitnesscontribu-tion.2.Thus,wemustalreadybeabletoindividuatetheparts.Thiscannotbedoneonthebasisoftheancestorstotheancestralorganisms,becausethiswouldgeneratearegress3.Butifweareabletoindividuatepartsforancestralorganismsindependentlyoftheirselectionhistory,thenthisispossibleforlivingorganismsSoifnaturalselectionisnotfitfortheindividuationoforganismicparts,whatis?Thisturnsouttobeaverydifficultquestion,andIcanansweritonlyinoutline.Inessence,Idonotthinkthatthereisageneralanswertothisquestion.Inotherwords,thereisnouniqueprincipleofcuttingupanorganismintopartsinthewaythatPlatosuggestedintheinfamouspassageofthePhaedrus,accordingtowhichagoodscientistshouldcarvenatureatherjoints.Clearly,Socrates’sadvicefromthePhaedrustoproceedbytryingnottosplinteranyparts,“asabadbutchermightdo,”28isnothelpfulatall,forwehavenotheory-independentwayofknow-ingwhenwehavesplinteredsomething.Theexplanandumisalmostneverneutralwithrespecttotheexplanans.Dif-ferenttheoreticalmodelsoftencomewithdifferentwaysofclassifyingthephe-nomena.Thishaslongbeenrecognizedforthephysicalsciences,forexample,byKuhnandFeyerabend,butfewpeople(exceptingRosenberg)havenoticedthatthesameholdsforbiology.Developmentalbiology,evolutionarybiology,evo-devo,physiology,cellbiologyandsoonhavedifferentwaysofindividuatingphenom-ena.However,Idowanttoarguethattheconceptofbiologicalfunctionisoftenin-volvedwhenbiologistscutupanorganismintoparts,includingmechanisms.Butthesalientconceptoffunctionneednotbethatofselectedeffectfunctions.Thereareotherconceptsoffunction,andtheycanalsofulfilltherolethatRosenbergthinksonlyselectedeffectfunctionscanplay.Asanalternative,Isuggestamodifiedversionofcausalrolefunctions.29ThisaccountstartswithCummins’s30analysisaccordingtowhichfunctionsaresuch27PaulGriffiths,“Function,Homology,andCharacterIndividuation”,in:PhilosophyofScience73,1,2006,pp.1-25.28Plato,CompleteWorks,J.M.Cooper(ed.).Indianapolis:Hackett1997,265e(p.542).29MarcelWeber,PhilosophyofExperimentalBiology.Cambridge:CambridgeUniver-sityPress2005;MarcelWeber,“Holism,Coherence,andtheDispositionalConceptofFunctions”,in:AnnalsintheHistoryandPhilosophyofBiology10,2005,pp.189-201.30RobertCummins,“FunctionalAnalysis”,in:JournalofPhilosophy72,1975,pp.741-\n164MarcelWebercapacitiesthatarecapableofexplainingacapacityofsomecontainingsystem.Theparadigmistheheart’scapacitytopumpbloodfiguringinanyadequateex-planationofthecirculatorysystem’scapacitytotransportnutrients,oxygenandbloodcellsthroughthebody.AccordingtoCummins,thepertinentcapacityofthecontainingsystemisamatterofaninterest-basedchoicetobemadebytheinvestigator.Ihavemodifiedthisaccountbysuggestingthatthissystemscapacityshouldbemadedependentnotontheinvestigator’sinterests,butontherolethatthecontainingsystemitselfplaysintheself-reproductionofthewholeorgan-ism.IarguethatthisiswhatturnsCummins-functionsintobiologicalfunctions.Cummins-functionscanbeappliedtoanykindofsystem.Butonlybiologicalsystemsarecapableofself-reproduction.Inorderforself-reproductiontooccur,anorganism’sfunctionsmustworktogether.Thespecificcontributionthatsomeorgan’scausalcapacitiesmaketoself-reproductionmakeswilldependonwhatotherorgansdo.Forexample,ifthereweresubsystemsofanorganismthatwouldusetheheart’sheatproductiontowardssomethingthatitselfmakesacontributionofself-reproduction,thentheheartwould(also)havethefunctionofproducingheat.Itistheplacethatsuchacausalcapacityplaysinawholenetworkthatgivesititsfunction(perhapsmuchinthewayinwhichalinguisticexpression’smean-ingisgivenbytheinferentialrolethattheexpressionplaysinanetworkofotherexpressions,asclaimedbyinferentialistsandsemanticholists).Ihavearguedthatintroducingsuchaglobalconstraintonasystemoffunctionsmightmaketheinterest-dependencevanish,providedthatthereisexactlyonewayoflayinganetworkofcooperatingfunctionsoveranorganism.Ofcourse,thisishardtoprove;butIsuggestthatitmightbepossiblebyusinganotionofmaximalexplanatorycoherence.31Thus,contrarytowhatRosenbergclaims,dividingupanorganismintodif-ferentpartsortraitscanbedoneindependentlyofitsselectionhistory.Whetherthereisonecorrectornaturalwayofdoingthis,however,isverydifficulttosay.32Whatseemsclearisthatfunctionshaveaholistic33character:Somethingonlyhasafunctionifitisconnectedtomanyotherthingsthatalsohavefunctionsand765.31Weber,“Holism,Coherence,andtheDispositionalConceptofFunctions”.32Alothangsonthewayinwhichtheexplanadumofsuchanetworkoffunctionsisconstrued.Itistemptingtosuggestthatithastobe“self-reproductionoftheindi-vidual”(asIhavedoneinmyPhilosophyofExperimentalBiology),however,thisnotionsuffersfromacertainindeterminacythatisintroducedbythereflexiveterm“self.”Whatisthat“self”thatisbeingreproduced?Andwhatdoesits“reproduction”or“maintenance”involve,inotherwords:whatareitspersistenceconditions?Notethattheanswer“theindividual”doesn’treallyhelpbecauseofthenotionofbiologi-calindividualisnotoriouslydifficult(seeJackWilson,BiologicalIndividuality-TheIdentityandPersistenceofLivingEntities.Cambridge:CambridgeUniversityPress1999).Thiscouldmakesomeroomforpluralism.33SeeMichaelEsfeld,HolisminPhilosophyofMindandPhilosophyofPhysics.Dor-drecht:Kluwer2001.\nLifeinaPhysicalWorld165thatconspiretomaintaintheorganism’sform.Furthermore,whatsomething’sfunctioniscandependonwhatotherthingsdotowhichitisconnected.How-ever,thisholismneednotnecessarilybeanobstacletoreductionism,unlesstherequirementsforsuccessfulreductionaremadeexcessivelystrong.Forinstance,itmightstillbepossiblethatKim’srequirements(seeSection1)canbesatisfied.Ofcourse,ontheviewoffunctionsthatIhavementioned,something’sfunctionmaynotonlydependonhowthisthinginteractswithitsimmediateinteractionpartners(Kim’s“causalrole”)butalsoonwhattheroleofthatthingisinthewholeorganism.Butoncethisroleisknown,therearenoobstaclestothenidentifyingtherealizersofthesefunctions.Inthefinalsection,Ishallcriticallydiscussanaltogetherdifferentchallengetoreductioninbiology.4.MICHAELTHOMPSON’SACCOUNTOFBIOLOGICALREGULARITIESAftermuchdebateonceterisparibuslawsandvarious“outbreaksoflawless-ness”34inbiology,manyphilosophersofbiologyincludingmyselfhavefoundJimWoodward’saccountofcausationandexplanation35veryhelpfultocometotermswithcausalregularitiesinbiology.However,thereissomethingthatthisaccountdoesnotquitecapture,andthisisthequestionofwhatmakesacertaincausalgen-eralizationabiologicalgeneralizationasopposedtomerelyaphysicalorchemicalone.Ithinkthefollowinganswerisnotreallysatisfactory:“Acausalgeneraliza-tionisbiologicalifitconcernslivingorganismsorpartsthereof.”Forthereareendlesslymanycausalgeneralizationsaboutanypartofanorganismthatcouldjustaswellbedescribedasphysicalorchemical,forexample,“bloodvesselswithahighcontentofelastinexpandasinternalfluidpressureincreases.”36Aninterestinganswertothequestionofwhatcharacterizesbiologicalgener-alizationscanbefoundintheworkofMichaelThompson.37Itcomesfromgen-eralmetaphysicsandhasthereforerarelybeennotedbyphilosophersofscience.Thompsonwritesforexample:NowsupposeIsay,‘Bobcatsbreedinspring’:itisobviousthatthisisn’tgoingtohappeninanyparticularcaseunlesscertainconditionsaresatisfied.Perhapsaspecialhormonemustbereleasedinlatewinter.Andperhapsthehormonewillnotbereleasedifthebobcatistooclosetosealevel,orifitfailstopassthroughtheshadeofacertainsortoftallpine.Butnow,toarticulatetheseconditionsistoadvanceone’steachingaboutbobcats.[…]The34Sober(1998),op.cit.35JamesWoodward,MakingThingsHappen:ATheoryofCausalExplanation.Oxford:OxfordUniversityPress2003.36C.KennethWaters,“CausalRegularitiesintheBiologicalWorldofContingentDistri-butions”,BiologyandPhilosophy13,1998,pp.5-36.37Thompson,op.cit.\n166MarcelWeberthoughtthatcertainhormonesarereleased,orthattheyliveinsuch-and-suchaltitudesandamidsuch-and-suchvegetation,isathoughtofthesamekindasthethoughtthatthybreedinthespring.[…]Theseconditionsarepresupposedbythelife-formitself.38Thompsonthinksthatthereisanimportantdifferencebetweenbiologicalgeneral-izationssuchas‘bobcatsbreedinspring’andpurelyphysicalgeneralizationssuchas‘waterboilsat100°C’.Butthedifferenceisnotthatonerequiresceterisparibusclauseswhiletheotherdoesn’t.Theybothdo.i.e.,bothgeneralizationsaresubjecttocertainconditionsthatmustobtainforthegeneralizationstobemanifested.Inthefirstexample,itisnecessarythatcertainenvironmentalcuesthattriggermat-ingbehaviorinbobcatsoccur(e.g.,longerdays,mildertemperatures)andthatnothinginterferes(e.g.,ashortageofprey).Inthesecondexample,itisnecessarythatnormalatmosphericpressureobtainsandthatthewaterhasnotbeensalted.ButaccordingtoThompson,inthebiologicalcaseitisitselfafactaboutthisspe-ciesthattheseconditionsobtain.Bobcatswillseekanenvironmentwherethecon-ditionsforbreedingarefavorable,suchthattheregularitywillobtain.Bycontrast,thereisnolawaboutwaterthatsaysthatallwatertendstooccurunderconditionssuchthattheregularity“waterboilsat100°C”oranyothersuchregularitywillobtain.Infact,thelattergeneralizationhasapurelyhypotheticalcharacter:Itonlysays,waterboilsifthetemperatureis100°Cormore.Bycontrast,thebiologicalgeneralizationiscategoricalinnature.Itreadsasitiswritten:bobcatsbreedinspring.Thatbobcatsliveinplaceswherethereisaseasonalchangeintemperatureanddaylengththattriggerstheirbreedingispartofthenatureofbobcats.ItisclearthatThompsonhasquiteadifferentconceptionofregularitiesorlawsthancontemporaryphilosophyofscience,infact,itisclosertoAristotelianformsthantolawsofnatureinthemodernsense.AccordingtoThompson,eachorganisminstantiatesacertain“life-form”thatischaracterizedbysuchcaterogicallawsastheonesaboutbobcatsinhisexample.Hisnotionoflife-formseemstobeoneofacomplexirreducibleessence,muchlikeAristotle’sconceptofeidos.Ofcourse,assuchthisconceptionisproblematic,especiallyinlightofalltheargu-mentsagainstbiologicalessentialismthathavebeenproducedinrecentyears.39However,theremightbesomemeritinThompson’ssuggestionthatwhatcharac-terizesbiologicalgeneralizationsisinpartthewayinwhichdifferentgeneraliza-tionsconspiretoensureeachother’sbeingmanifestedbyindividualorganisms.Theremightperhapsevenbeananalogytowhatsomephilosophersofsciencehavesaidaboutnaturalkindsinbiology,forinstance,RichardBoyd’stheoryofhomeostaticpropertyclusters.4038ibid.p.287.39JohnDupré,TheDisorderofThings:MetaphysicalFoundationsoftheDisunityofSci-ence.CambridgeMass.:HarvardUniversityPress1993.40RichardBoyd,“Homeostasis,Species,andHigherTaxa”,in:RobertA.Wilson(ed.),Species:NewInterdisciplinaryEssays.Cambridge,Mass.:MITPress,pp.141-185.\nLifeinaPhysicalWorld167AccordingtoThompson’saccount,whatmakescertainregularitiesbiologi-callysalientisthattheyensurethatotherregularitiesareinstantiated,regularitiesthatarethemselvesimportantforthesurvivaloftheindividual,andsoon.Thisisquitereminiscentofmytentativeanswertothequestionofwhatmakescertainactivitiesinanorganismfunctionallyrelevant(seetheprecedingsection).OnthisaccountoffunctionsaswellasonThompson’saccountofbiologicallaws,thereexistsahighlycomplexrelationbetweenthedifferentpartsofanorganism,are-lationthatobtainsexactlyifthepartsareorganizedsuchthatthesystemsustainsitself.Thiskindoffocusonself-reproductioniswhatdistinguishesbiologyfromothernaturalsciences.WhereImustpartwithThompsonishere:Iseenoprincipledwayofdrawingalinebetweenessentialandnon-essentialpartsofanorganism.Whichlawsareassociatedwiththeformandwhichonesaren’t?Furthermore,IseenoreasonwhyThompson’saccountofbiologicallawsshouldbeinconsistentwithanadequateformofreductionism.Evenifitistheirrelationtotheinstantiationconditionsofotherlawsthatmakescertainlawsaboutbiologicalentitiessalient,thereisnoreasonwhytheserelationscannotbefullyunderstoodandexpressedinphysical-chemicallanguage.Onefinalpoint:ItshouldalsobenotedthatThompson’scategoricallawsareonlyvalidforthelivingstateandiftheorganismsoverwhichtheyrangeliveintheirnormalenvironment.Theycontainnoinformationwhatwouldhappen,forinstance,ifNorthAmericanbobcatsweretransferredtotheTropics.Wouldtheystillbreedinspring?Toanswerthiskindofquestionrequiresgoodold-fashionedcausallawsthatrangenotonlyoverasetofactualstates,butovercounterfactualsituationsaswell.Biologistscandiscoversuchcausallawsaswell,buttheywillbeoftheordinary,hypotheticalsort.Inthisrespect,biologyisnodifferentfromothernaturalsciences.5.CONCLUSIONSTherehavebeenmanyattemptstoshowthatbiologyoccupiessomespecialplaceinthenaturalsciences,andmostofthemhaveattemptedtoshowthatbiologicaltheories(orlaws)areirreducibletophysical-chemicaltheories.Thisisobviouslycorrectif“reduction”isunderstoodinastrong,derivationalsense,butfarlessobviousifaweakersenseofreductionsuchasKim’sisassumed.Oneofthemostpopularargumentsagainstreduction,theargumentfrommultiplerealizability,isnotconvincingonsuchaweakerview.Additionalargumentstotheeffectthatsomehigher-levelexplanationsdosomeexplanatoryworkthatcannotberecov-eredatthelowerlevelrelystronglyonintuitionsastowhatconstitutesagoodexplanationandarenotconvincingtothosewhodon’tsharetheseintuitions,fortheintuitionsofreductionistsandanti-reductionistsnotoriouslydiffer.\n168MarcelWeberFurther,IhaveconsideredRosenberg’sargumentthat(1)eventhoughprox-imatebiologyneedsevolutionaryconcepts(properfunction)toindividuatethepartsofanorganism,this(2)isnoproblemforthereductionistbecausethesalientevolutionaryprinciplesarefundamentalphysicallaws.Thelatterclaimfailsbe-causenaturalselectiontheoryisnotaunifiedtheory;itconsistsofawidevarietyofspecificmodelsthatdeploydifferentfitnessmeasures.Furthermore,therearealternativewaysofhowbiologistscanindividuatethepartsofanorganism,forexample,bycausalrolefunctions.Ihavediscussedarichversionofcausalrolefunctionsthatmightyieldanaturalsystemoffunctionsforeachtypeoforganism.Eventhoughfunctionsareinasenseholisticpropertiesonthisaccount,reduction-istsneednotworryaboutthis.Finally,IhavecriticallyexaminedM.Thompson’sessentialisticaccountofbiologicallawsaccordingtowhichthelatterdevelopanirreduciblelife-formforeachspeciesoforganism.Iarguethat,whilethisaccountgivesagoodanswerofwhatmakescertainregularitiesbiologicallysalient(orbiologicalatall),italsoprovidesnoargumentsagainstasuitablyunderstoodreductionism.UniversitätKonstanzFachbereichPhilosophiePostfachD978457KonstanzDeutschlandMarcel.Weber@uni-konstanz.de\nCLAUDEDEBRUCOMMENTSONMARCELWEBER’S“LIFEINAPHYSICALWORLD:THEPLACEOFTHELIFESCIENCES”MarcelWeber’scontributionisanextremelyaccurateandsyntheticviewofthestateofsomediscussionswhicharetypicalofwhatismeanttodayby“philosophyofbiology”.Beforecommentingmorecloselyhiscontribution,Iwouldliketotakesomedistanceandtocomeclosertoactualscience(atleasttosomepartsofthebiologicalsciences).Lookingatbiologyasasetofdifferentsciencesismoreandmorefrequent.Frommathematicalbiologytomedicalsciencesandtobiotech-nologies,thefieldofbiologybecamemoreandmorediversified,althoughsomecommonfeaturesremainatthemostgenerallevel.Biologyasawholeremainsba-sicallyanempiricalscience,orasetoflargelyempiricalresearches.Philosophersdealingwithbiologyshouldnotunderestimatebiologicalempiricismandbiologi-calexperimentalism.Theextenttowhichthisisthecase,whichisrecognizedbyMarcelWeber,createsarealdifficultyforphilosophicalthinking,becauselookingatbiologyasalargelyempiricalandbasicallyexperimentalscience,leadstotheconclusionthatbiologyismuchlessstabilized,atleastinimportantparts,thanitcouldbesupposed.Ifyoudonotlookatthingsfromapurelyconceptualpointofview,ifyouavoidprojectingonbiologyasascienceinprogressorratherasasetofcloselyrelatedresearchdisciplines,theideaofthestructureofbiologicalknowledgeasthemajorproblemofthephilosophy“of”biology,thenyougetamuchmorerealisticandricherpictureofthebiologicalsciences.Pluralismisnowcompletelyintegratedinbiologicalthinking,andexperimentalismremainsamostgeneralfeature.AsfarasIcansee,MarcelWebercouldagreeentirelyontheseremarks,sincehewroteabookonbiologicalexperimentalism.Nowletusgotothefirstitemunderdiscussion,reductionism.Thissubjecthasbeendiscussedunderpreciselythisterm,reductionism,sinceatleastonehun-dredandfiftyyears–andwestillcontinue.InhisIntroductiontothestudyofex-perimentalmedicine(1865)ClaudeBernardaskedthequestion:canwe“reduce”biologytophysicsandchemistry.Hisanswerwaspartlyyesandpartlyno,andwasbasedonthestateofthephysiologyofmetabolism.Thecatabolicpartofme-tabolismwasunderstandableintermsofthemolecularactivityofenzymes,whichwereknownasproteins.Thebiosyntheticpartofmetabolismwasnotunderstand-ableintermsofthechemicalactivityofenzymes.Lateritbecameunderstandable,whenitwasrealizedthatenzymesacteduponequilibriumreactions,whichcouldbeacceleratedinbothdegradationandsynthesisdirections.AtaboutthesametimeasBernard’stheoreticalreflections,ErnstMachaskedthesamequestion:canwereduce(zurückführen)?AsaphysicisthewasextremelysuccessfulinhisF.Stadler(ed.),ThePresentSituationinthePhilosophyofScience,ThePhilosophyofScienceinaEuropeanPerspective,DOI10.1007/978-90-481-9115-4_13,©SpringerScience+BusinessMediaB.V.2010\n170ClaudeDebruapproachesofexperimentalpsychologyandpsychophysiology.Ifwecontinueinthishistoricalaccount,wecanobservethatClaudeBernard’stheoreticalviewontheconstancyoftheinternalenvironmentbecameentirelyunderstandable,regard-ingthechemicalpropertiesofblood,intermsofbiophysicalchemistry,thankstotheAmericanphysiologistLawrenceHenderson,whodevisedtheequationsgov-erningbloodequilibria,certainlyoneofthemajortheoreticalresultsinclassicalbiology.Thesamekindofdemonstrationcouldapplytomanyfieldsofbiology.Sowhydowecontinueaskingthesamequestion?Isitbecauseofourdeepreluctancetochangeourmentalhabits,tomodifyourcherishedattitudes–asan-othergreatViennesescientist,ErwinSchrödinger,oncesaidintheconcludingpar-agraphsofhisNobelLecturein1933“DerGrundgedankederWellenmechanik”:“mitsolchenalten,liebenundunentbehrlichscheinendenBegriffenwiewirklichoderblossmöglich”etc.(withourold,cherishedandapperentlyindispensablecon-ceptslikerealandsimplypossibleetc.)?1Schrödingerwascertainlyanextremelycriticalmind.Oneofthereasonswhywecontinueaskingthesamequestionaboutreductionismisperhapsacognitiveone:thiswouldbebecausewepresentlydonothaveanysatisfactorymeantoreconstruct,ortovisualizesimultaneouslyallaspectsofagivenbiologicalreality,togofromaverypartialviewofthepartstoafullyintegratedviewofthewhole.Anotherreasonwouldbemoreontological:thiswouldbethepeculiarmultilevelstructureoforganismswhichrequiresamuchmoresophisticatedaccountthantheold,nineteenthcenturytypeofreduction.MarcelWeberasksthequestion:“WhyRosenberg’sdefenceofreductionismfails”,inspiteofstrongarguments?Theargument,thattheprincipleofnaturalse-lectionshouldbeconsideredasaphysicallawshouldbeconsideredveryclosely.Indeed,suchaviewhasbeendiscussedinclassicalcontributionstobiophysi-calchemistry.In1979,ManfredEigenandPeterSchusterpublishedtheirfamousmemoirTheHypercycle.APrincipleofNaturalSelf-Organization.Inthismemoirtheydevisedthemathematicaltheoryofanewkindofbiochemicalcycle,thehypercycle.True,biochemicalaswellaschemicalcycleswerewell-knownmuchbeforeEigenandSchuster’swork.ThecomplicationintroducedbyEigenandSchusteristhatinsteadofdealingwithcatalyticsystems,whicharealreadycyclic,theydealtwithautocatalytic,self-replicativesystems,whichtheycalledhypercy-clic.Ahypercycleisacyclicalarrangementofcyclicunits.Acatalytichypercycleis“asystemwhichconnectsautocatalyticorself-replicativeunitsthroughacycliclinkage”.2Thismeansclearlyanewleveloforganization.Hypercycleshavepar-ticular,emergingproperties.Theysharewithself-replicativeunitsthepropertyofconservingacertainamountofinformation,whichisaprerequisiteforDarwinianselectionandevolution.Theyhaveadditionalintegratingproperties,whichen-hancetheirselectivepower.Indeed,accordingtoEigenandSchuster,“theycom-1ErwinSchrödinger,WasisteinNaturgesetz?BeiträgezumnaturwissenschaftlichenWeltbild.München–Wien:R.Oldenbourg1967p.100.2ManfredEigenandPeterSchuster,TheHypercycle.APrincipleofNaturalSelf-Or-ganization,Berlin–Heidelberg–NewYork:SpringerVerlag1979,p.6.\nCommentsonMarcelWeber171peteevenmoreviolentlythanDarwinianspecieswithanyreplicativeentitynotbeingpartoftheirown.Furthermore,theyhavetheabilityofestablishingglobalformsoforganizationasaconsequenceoftheir‘once-forever-selectionbehavior,whichdoesn’tpermitacoexistencewithotherhypercyclicsystems,unlessthesearestabilizedbyhigherorderlinkages”.3Thesehypercyclesaredynamicalsys-temswhicharedescribedbynon-lineardifferentialequations,withtheirtypicalbehaviorsintermsoflimitingcyclesandattractors.AccordingtoEigenandSchuster,hypercyclesarenecessaryprerequisitesofDarwiniansystems.Doesthismeanthatnaturalselectionisagenerallawofna-ture?Certainlynot.“Whatisthemolecularbasisofselectionandevolution?Ob-viously,suchabehaviorisnotaglobalattributeofanyarbitraryformofmatterbutratheristheconsequenceofpeculiarpropertieswhichhavetobespecified”.Thesepropertiesaredynamicalones.Wecanconcludethat“systemsofmatter,inordertobeeligibleforselectiveself-organization,havetoinheritphysicalproper-tieswhichallowformetabolismandforself-reproduction.Theserequirementsareindispensible”.4ItisperfectlytruethatEigenandSchusterdevelopedtheconceptofmolecularDarwinism.Itiscertainlynottruethattheymadenaturalselectionagenerallawofnature.Atthebiochemicallevel,thesystemswhicharedescribedbyEigenandSchusterarehighlyelaborated,likeenzymesetc.SoIamperfectlyinagreementwithWeber’scriticismofRosenberg.NowIwouldliketomakeanadditionalcommentonthenotionoffunctionandontheproblemof“individuatingtheparts”.Biochemistry,physiology,andevenmoresoneurophysiologyarecharacterizednotonlybyfunctionality,asyoumaydefineitinaveryabstractway,butmainlybypolyfunctionalityandlink-agebetweenfunctions,higherorderlinkages(likehypercyclesorotherkindsoflinkageslikeallosterism)whichyoumayfindalreadyatthelevelofsinglemol-ecules.Theselinkagesmaybeperfectlywelldescribedinmanyways,includingfundamentalthermodynamics(thiswastheworkofJeffriesWyman,abiophysicalchemist,inthesixtiesandseventies).5Polyfunctionalityisakindofrule.Proteinsofcoursearethemostintenselystudiedexamplesofthesebehaviors.Anditcan-notbeotherwise,becauseifnolinkagebetweenfunctions,nofunctionatall(asweknowalreadyfromthehypercyclemodel).Thismakesthequestionofdefiningfunctionsevenmorecomplicated,andthenecessityofamoreglobalviewmoreevident.IagreeonthispointwithMarcelWeberratherthanwithRosenberg.Ifwelookatthequestionsunderdiscussion,reduction,function,andlaw,allthesetermsaretypicalofnineteenthcenturyscience.Morerecently,biologistsstartedtodiscussemergingpropertiesofvariouskinds,levelsoforganization,polyfunctionality,modelsandsimulations,computationalcomplexity,allfeatures(sometimesunsolvedlikethecomputationalcomplexityofbiologicalstructures)3Ibid.,p.6.4Ibid.,p.8.5SeeClaudeDebru,L’espritdesprotéines.Histoireetphilosophiebiochimiques,Paris:Hermann1983.\n172ClaudeDebruwhicharetheresultofamuchdeeperunderstandinginbiologicalsciences.Phi-losophersshouldbecomemoreacquaintedwiththatandgotothelab.Indeed,phi-losopherscouldbeextremelyusefuliftheywouldcopewithreal,particularbio-logicalproblems(whichtheycertainlycandoinacreativefashion)–forinstance(speakingoffunctions)theycouldcopewiththeproblemofunknownfunctions.Indeedtherearephysiologicalprocesseswhosefunctionsarestillunknown,andthisisparticularlythecaseinneurophysiology,arapidlychangingfieldofinvesti-gation.Thecaseofparadoxical(orrapideyemovement)sleep,whichcorrespondstodreaming,isagoodexampleofthat.6Paradoxicalsleepiscertainlymostim-portantforphysiologicalregulationsinthebrain.Butitsfunctionsremainpres-entlyentirelyunknown–althoughtherearemanyspeculations.Itscomplicatedbiochemicalmechanismsaremoreorlessentirelyoralmostentirelyknownatthemolecularandcellularlevels,buttheoutputoftheseprocessesisstillunknown,sothatphysiologistscannotdefinethefunctionsofparadoxicalsleepmechanisms.Thisremainsasubjectforfuturephysiology.Philosopherscouldbeextremelyuse-fulinthisenquiry.Theycouldcollaboratewithneuroscientistsastheydidalreadyinthepastwithmuchsuccess(thebestknownexampleisPopperandEccles).Butthisisnotamatterof“philosophyof”.Ratherthisisphilosophypureandsimple.Speculatingaboutpossiblefunctionsforparadoxicalsleepisnotdoing“philoso-phyof”biology,butratherdoingphilosophywithinphysiology.Weshouldputagainphilosophypureandsimpleontheagenda–togetherwithepistemology,whichismoremodestandconveysasenseofmodifyingquestionsandvaryinghypothesesandmodels.Idon’tthinkwecanlookatbiologyasanentirelyunifiedscience,evenfromaphilosophicalpointofview.Naturalselectioniscertainlyamajorpartofthegame.Chemistryisanotherpartofthegame,withitsownrulesandconstraints.IonceheardaStrasbourgchemistdiscussingFrançoisJacob’sideaofmoleculartinkeringandmakingpreciselythispoint.Certainlynaturecan-notplayanygameinbiology.Weshouldtrytocombineseveralkindsofcausesandlookmorecloselyatthestabilizingroleofemergingproperties.DepartmentofPhilosophyEcoleNormaleSupérieure45rued’Ulm75005ParisFranceclaude.debru@ens.fr6SeeClaudeDebru,Neurophilosophiedurêve,Paris:Hermann2006.\nTHOMASA.C.REYDONHOWSPECIALARETHELIFESCIENCES?AVIEWFROMTHENATURALKINDSDEBATE1.INTRODUCTIONPhilosophersofthespecialsciencesseemtofinditimportanttoaskwhetherornotparticulargroupingsofthingsthatfeatureinparticularspecialsciencescanbeconceivedofasnaturalkinds.1Forexample,aquicksearchofthephilosophicalliteratureofthepastdecadescomesupwithseveraldozensofpaperstargetingthequestion“Is…anaturalkind?”,manyoftheseconcerningkindsofemotionsandtheemotioncategoryinpsychology2andthecategoryofconceptsinpsycholo-gy/cognitivescience.3Motivatingthesepapersoftenareconcernsaboutthescientificstatusofthefieldthatstudiesthekindinquestion–i.e.,concernsaboutissuessuchastheplacethatthefieldoccupieswithinthewholeofscience,itsindependencefromandrelationshipstootherfields,itsreducibility,itsexplanatorypowerandautonomy,itsbeingaunifiedandself-containedfieldofwork,etc.Ifitcanbemadeplausiblethatthekindorkindsthatconstitutethespecificobjectsofstudyofaparticularfieldofinvestigationarenaturalkinds,thethoughtis,thiscanbetakenasanindicationthatthefieldisacomparativelyindependent,autonomousandself-con-tainedsciencethathasitsownproperplaceamidsttheothersciences,thatstud-1Iuse‘specialsciences’inthesenseofJerryA.Fodor,‘Specialsciences(or:thedisu-nityofscienceasaworkinghypothesis)’,in:Synthese28,1974,pp.97-115.2E.g.,LouisC.Charland,“Thenaturalkindstatusofemotion”,in:BritishJournalforthePhilosophyofScience53,2002,pp.511-537;PaulE.Griffiths,“Emotionsasnaturalandnormativekinds”,in:PhilosophyofScience71,2004,pp.901-911;PaulE.Griffiths,“Isemotionanaturalkind?”,in:RobertC.Solomon(Ed.),ThinkingAboutFeeling:PhilosophersonEmotions.NewYork:OxfordUniversityPress,2004,pp.233-249;JesseJ.Prinz,GutReactions:APerceptualTheoryofEmotion,NewYork:OxfordUniversityPress,2004,Chapter4;LisaF.Barrett,“Areemotionsnatu-ralkinds?”,in:PerspectivesonPsychologicalScience1,2006,pp.28-58;AlexandraZinck&AlbertNewen,“Classifyingemotion:Adevelopmentalaccount”,in:Synthese161,2008,pp.1-25.3E.g.,EdouardMachery,“Conceptsarenotanaturalkind”,in:PhilosophyofScience72,2005,pp.444-467;EdouardMachery,“Howtosplitconcepts:AreplytoPiccininiandScott”,in:PhilosophyofScience73,2006,pp.410-418;EdouardMachery,“100yearsofpsychologyofconcepts:Thetheoreticalnotionofconceptanditsoperation-alization”,in:StudiesinHistoryandPhilosophyofBiologicalandBiomedicalSci-ences38,2007,pp.63-84;GualtieroPiccinini&SamScott,“Splittingconcepts”,in:PhilosophyofScience73,2006,pp.390-409.F.Stadler(ed.),ThePresentSituationinthePhilosophyofScience,ThePhilosophyofScienceinaEuropeanPerspective,DOI10.1007/978-90-481-9115-4_14,©SpringerScience+BusinessMediaB.V.2010\n174ThomasA.C.Reydoniesitsownspecificdomainofphenomenaandthatdevisesitsownexplanatorytheoriestoaccountforthesephenomena.ConsiderthestatementbypsychologistZenonPylyshynthatthepresenceofnaturalkindsspecifictocognitivesciencewasoneofthefactorsthatcouldopenupan“(…)excitingpossibility:theprospectthatcognitivescienceisagenuinescientificdomainlikethedomainsofchemistry,biology,economics,orgeology”.4Butthereareclearproblemswiththiswayofanalyzingthescientificstatusoffields,astheconclusionsthatonewilldrawaboutthescientificstatusofafieldofworkdependontheparticularaccountofnaturalkindhoodthatoneadopts.Eveniftherewereagenerallyacceptedtheoryofnaturalkinds,itwouldnotbeclearexactlyinwhichwaysthescientificstatusofafieldofinvestigationisaffectedbyits“having”naturalkindsornot.Andaspresent-dayphilosophyisstilllackingagenerallyacceptedtheoryofwhatitmeansforagroupofthingstobeanaturalkind,itseemsthatthenotionof‘naturalkind’cannotbeofmuchusewhenassess-ingthestatusofaparticularfieldoraclusterofsciences,suchasthelifesciences.Indeed,Hackingrecentlyarguedthatthenotionof‘naturalkind’won’tbeofmuchusewhenaddressinganyissueofphilosophicalinterest,becausenotonlythereisnogeneraltheoryofnaturalkindsbutthereisnoagreementonwhattheproblemisthatatheoryofnaturalkindsissupposedtoresolve.5Rather,Hackingcontended,framingphilosophicalquestionsintermsof‘naturalkinds’onlyservestomakethingsunnecessarilycomplicated.Againstthisbackground,myquestioniswhethertherelationsbetweenthelifesciencesandotherscientificfields–i.e.,theplaceofthelifesciencesamongtheothersciences–canbemeaningfullycharacterizedusingthenotionof‘naturalkinds’.6ContraHacking(whoprovidedageneralargumentanddidnottalkabout4ZenonW.Pylyshyn,ComputationandCognition:TowardaFoundationforCognitiveScience.Cambridge(Mass.):MITPress,1984,p.xi;emphasisadded.Formorerecentstatementstothisextent,involving‘emotion’and‘concept’asdenotingnaturalkindsofpsychologyorcognitivescience,seeCharland,loc.cit.,p.512,Griffiths,“Emotionsasnaturalandnormativekinds”,loc.cit.,p.901,orMachery,“100yearsofpsychologyofconcepts:Thetheoreticalnotionofconceptanditsoperationalization”,loc.cit.,p.66.5IanHacking,“Naturalkinds:rosydawn,scholastictwilight”,in:AnthonyO’Hear(Ed.),PhilosophyofScience(Philosophy–RoyalInstituteofPhilosophySupplement61),Cambridge:CambridgeUniversityPress,2007,pp.203-239.Cf.PaulM.Church-land,“Conceptualprogressandword/worldrelations:insearchoftheessenceofnatu-ralkinds”,in:CanadianJournalofPhilosophy15,1985,pp.1-17(p.1).6Talkaboutthescientificstatusofafieldanditsplaceinscienceisoftenthoughtofasreferringtothesamesetofissues–see,e.g.,Mayr’slongdiscussionoftheplaceofthebiologyamongthesciences(ErnstMayr,TheGrowthofBiologicalThought:Diversity,Evolution,andInheritance.Cambridge(Mass.):HarvardUniversityPress,1982,pp.21-82).Here,thisquestionwaspre-given,asthispaperwaswrittenforthefirstconferenceoftheESFresearchnetworkThePhilosophyofScienceinaEuropeanPerspective,inthecontextofTeamB’sworkonthetheme“ApproachestotheFounda-tionsofScience:ThePlaceoftheLifeSciences”.Workonthispaperwassupportedby\nHowSpecialAretheLifeSciences?175naturalkindsinspecificrelationtoscientificstatus),Iholdthatnaturalkindtheorycanbeausefultoolwhenattemptingtocharacterizetheplaceofascientificfieldofworkwithinthewholeofscienceanditsrelationstoothersciences.Myaim,then,isprogrammatic:toachieveclarityabouthownaturalkindtheoriescouldbeusedfruitfullyinthiscontextbycriticallyreviewingtheprincipalwaysinwhichtheyhavebeenusedbyothers,andinsodoingsetthestageforfuturework.Ishallbeginbyexaminingtwodifferentwaysinwhichphilosophershavetriedtodrawconsequencesaboutthescientificstatusoffieldsofinvestigationfromconsiderationsaboutthekindsthatfeatureinthem(Sections2and3).Whiletherearequiteafewtheoriesofnaturalkindsavailableintheliterature,thetheo-riesthatIshalllookatrepresentthetwomostbasicviewsofwhatnaturalkindsare.Thediscussionwillshowthatframingquestionsabouttheplaceofascienceintermsof‘naturalkinds’hasfailedtoyieldunequivocalconclusions,insteadcomplicatingthephilosophicaldiscussiononthetopic.So,thispartofmydiscus-sionsupportsHacking’spessimisticviewoftheusefulnessofnaturalkindtheory.Ishall,however,goontoarguethatthesituationisnotasproblematicasitlooksandthattheremightbewaysofmakingnaturalkindtheoryusefulinthepresentcontext(Section4).InSection5,Ishallconcludewithsomeremarksonthe“placequestion”forthecaseofthelifesciences(broadlyconceived)againsttheback-groundoftheviewsuggestedinSection4.2.THEESSENCE-VIEWOFNATURALKINDSPerhapsthemostwidespreadaccountofnaturalkindsiswhatI(forobviousrea-sons)shallcallthe“essence-view”ofnaturalkinds.Onthisview,acleardistinc-tioncanbemadebetween“good”naturalkindsthatarecharacterizedbykindessencesandotherkindsofkindsthatarenotcharacterizedbykindessences.Thisdichotomyofkinds,inturn,isoftentakenasunderwritingaviewofsci-enceaccordingtowhichtherealmofsciencedividesintotwodistinctdomains,thosesciencesthatarebuiltaround“good”naturalkindsandthosethataren’t.Theformerareoftentakentobethesciencespitchedatthemostfundamentallevelsoforganization(physicsandpossiblysomefieldsofchemistry),whilethelatterarethoughtofasbeingthosefieldsofworkthatstudyphenomenaathigherlevelsoforganization(chemistry,thelifesciences,psychology,cognitivescience,thesocialsciences,economics,etc.).Thedichotomybetweennaturalkindsandotherkindsofkindsiscommonwithauthorswhodefendsomeformoftraditionalessentialismaboutkinds.Thetraditioncanroughlybetracedasfollows.ItoriginatedwithPlatoandAristotle,isprominentin,amongothers,Locke’sdiscussionof“realessences”inhisEssayConcerningHumanUnderstandingandMill’sdiscussionofkindsintheSystemofgrants(RE2613/1–1,RE2613/1-2)fromtheGermanResearchCouncil(DFG).\n176ThomasA.C.ReydonLogic,wasrevivedincontemporaryphilosophyinKripke’sandPutnam’sworksonthecausaltheoryofreferenceandhasitsmostrecentmanifestationsinmeta-physicalpositionssuchasEllis’sScientificEssentialismandOderberg’sRealEs-sentialism.7Accordingtothistradition,everynaturalkindischaracterizedbyitsownspecifickindessence,akindessencebeingconceivedofasasetofproperties(orpossiblyasingleproperty)thatallandonlythemembersofthekindinques-tionpossess.Possessionofeachofthepropertiesincludedinakind’sessenceisseparatelynecessaryandpossessionofallofthemisjointlysufficientforbeingamemberofthekind.Onthisview,naturalkindsarespatiotemporallyunrestrictedinthesensethatanaturalkindcaninprinciplehavemembersanywhereandatanytimeintheuniversewheretherightconditionsfortheirexistenceobtain.8Howev-er,becauseoftheirstrictmembershipconditionsnaturalkindscanonlybefoundatthoselevelsoforganizationwherethingsreallyareneatlyorderedintodisjointgroupsinsuchawaythatforeverygroupofthingsthereisakindessence.Asonecontemporaryproponentofkindessentialismputit,inorderfortheretobenaturalkinds,thebordersbetweenkindsmustbedrawnbynature,notbyus.9Evenifinpracticewemightbeunabletoidentifythekindessencesofputativenaturalkinds,weshouldatleasthavegoodreasonstoassumetheirexistence.Onthisviewofwhatitmeanstobeanaturalkind,genuinenaturalkindsarehardtofind.Givenwhatthesciencestellusaboutwhatsortsofthingsexistintheworld,wehavereasonstobelievethat“proper”kindessencescanbeidentifiedonlyforkindsonthemostfundamentallevelsoforganization,thatis,presumablytheorganizationallevelsofelementaryparticles,atomsandcomparativelysimplemolecules(formacromoleculesitalreadybecomesunclearwhethertherearesuf-ficientlystrictbordersdrawnbynature).Inaddition,manyofthekindsthatfea-tureinscientificallyimportantgeneralizationsinthespecialsciencesaresuchthat7BrianEllis,ScientificEssentialism.Cambridge:CambridgeUniversityPress,2001.DavidS.Oderberg,RealEssentialism.London&NewYork:Routledge,2007.Forahistoricalsketch,seee.g.,JohnDupré,“Naturalkinds”,in:WilliamH.Newton-Smith(Ed.),ACompaniontothePhilosophyofScience.Oxford:Blackwell,2000,pp.311-319.Irealizethatsomeauthorsdisagreewithmyroughhistoricalsketchandtracethephilosophicaltraditionofdiscussionaboutnaturalkindsthroughthehistoryofphilosophyinadifferentmanner(e.g.,IanHacking,“Atraditionofnaturalkinds”,in:PhilosophicalStudies61,1991,pp.109-126;RichardN.Boyd,“Realism,anti-foun-dationalismandtheenthusiasmfornaturalkinds”,in:PhilosophicalStudies61,1991,pp.127-148).Furthermore,Irealizethatthereareimportantdifferencesbetweenthevariouspositionsthathavebeendefendedandaredefendedbyauthorswhostandintheessentialisttradition.ButtheseareissuesthatIcannottakeupinthepresentpa-per.8ThisiswhyMillikancallssuchnaturalkinds“eternalkinds”,distinguishingthemfromkindsthatcannothavemembersatanytimeorplace(RuthG.Millikan,“Histori-calkindsandthe‘specialsciences’”,in:PhilosophicalStudies95,1999,pp.45-65;RuthG.Millikan,OnClearandConfusedIdeas:AnEssayAboutSubstanceConcepts.Cambridge:CambridgeUniversityPress,2000).9Ellis,loc.cit.,pp.19-20.\nHowSpecialAretheLifeSciences?177theirmemberscannotoccuranywhereandatanytimeintheuniverse,butonlyatspecifictimesandplaces.Biologicalspeciesareacaseinpoint.Whilespecieshaveoftenbeenconsideredoneoftheparadigmexamplesofnaturalkinds,themembersofaspeciescannotoccuratanytimeorplaceintheuniversewheretheconditionsfortheirexistencehappentoobtain,astheymuststandinonecontinu-ousreproductivelineageinordertocountasmembersofthesamespecies.Moreo-ver,biologicaltheorygivesusreasonstobelievethattherearenokindessencesforbiologicalspecies,asvariationamongthemembersofaspeciesiscrucialforevolutiontooccur.Thefactthattheorganismsofaspeciescaninprinciplevaryinanyoftheirtraits,however,conflictswiththeideathatthereshouldbeoneorseveralessentialpropertiesthatallorganismsofaspecieshaveincommon.Althoughonthisviewmany(ormost,orall)ofthekindsthatthespecialsciencesstudyseemnottobegenuinenaturalkindsatall,assessmentsofthescientificstatusoffieldsbasedonwhethertheystudyessentialistnaturalkindstendtoremaininconclusive.Still,aconsequencethatisoftendrawnisthatthespecialsciencesarequalitativelydifferentfromthosefundamentalfieldsofsci-encethatcenterroundnaturalkinds.Forexample,ontheviewthatoneoftheprimaryaimsofscienceistoprovideuswithaninventoryofthenaturalkindsthatmakeupthefurnitureoftheworldandtouncovertheirinnermostnatures,10oneshouldconcludethatthefieldsthatdonotstudynaturalkindsaremoremarginaltorealizingtheaimsofsciencethanthefundamentalsciences.Ifonethinksthatcitingthenaturalkindmembershipofthingshasaspecialexplanatoryforce–forexample,ifoneagreeswithEllis’spositionthatthelawsofnaturemustultimatelybegroundedinnaturalkinds–,oneshouldconcludethatthesciencesthatdonotstudynaturalkindslackthisspecialexplanatoryforce.Forthelifesciences,thisissometimestakentoimplythatmostofthemare“sciencesofcasestudies”,i.e.,fieldsofinvestigationthatdonotdevisegeneralexplanatorytheories,butratheraccountforindividualcasesbyrelyingontheexplanatorypowerofothersci-ences.11However,traditionalkindessentialismcanbecriticizedonatleasttwocounts.First,itrestsonquestionableaprioriassumptionsaboutwhatitisforagroupofthingstobeanaturalkind.Apartfrombeingrootedinavenerable,long-standingphilosophicaltradition,theideathatthefurnitureoftheworldcomesinnatural10Aviewthatusedtobeheldmorewidelybypastphilosophersthanitistoday(cf.Du-pré,loc.cit.,p.311;StathisPsillos,PhilosophyofScienceA-Z.Edinburgh:EdinburghUniversityPress,2007,p.156).11Rosenbergarguedthatmostofbiologyconsistsofsciencesofcasestudiesthatdonothavetheirownexplanatorypowerbutusetheexplanatorypowerofthephysi-calsciences,theprincipleofevolutionandthelawsofmolecularbiology(AlexanderRosenberg,TheStructureofBiologicalScience.Cambridge:CambridgeUniversityPress,1984,pp.202,211,219-225).Forasimilarpointaboutecology,seeKristinS.Shrader-Frechette&EarlD.McCoy,MethodinEcology:StrategiesforConservation.Cambridge:CambridgeUniversityPress,1993.\n178ThomasA.C.Reydonkinds,eachofwhichindividuatedbyitsownkindessence(inthestrictsensediscussedabove)doesnotseemtohavemuchtosupportit.Ultimately,thisidearemainsanaprioriassumptionaboutwhattheworldislikethatcanberejectedaseasilyasitcanbeaccepted.Inaddition,traditionalkindessentialismfailstoprovidesomethinganad-equatetheoryofnaturalkindsshouldprovide,namelyanexplanationofwhatitisthatmakesmembersofthesamekindsimilar.Letmeuseawell-wornexampletoillustratethis.Accordingtotraditionalessentialism,thekindessenceofgoldistheatomicpropertyofhaving79nuclearprotons.Thisproperty,itisheld,explainstheotherpropertiesthatgoldatomsexhibit:thevariousmaterialpropertiesofgoldatomsandtheircharacteristicbehaviorsinvariouscircumstancescausallyflowfromtheirhaving79nuclearprotons.Butthisdoesnotexplainwhydifferentgoldatomshavesimilarproperties–rather,itshiftsthequestionfromwhygoldatomshavesimilarobservablepropertiesandbehaviorstowhygoldatomshaveasimilaressentialpropertythatunderliestheirobservablepropertiesandbehaviors.Pre-sumably,thisproblemtracesbacktotherootsofthetraditionofkindessentialism.ForbothPlatoandAristotle,kindessencesweregivenaspectsofnaturethatwerenotthemselvesinneedofdeeperexplanation.Kindessenceswereconsideredtobebasicfeaturesoftheworld,eitherexistingasPlatonicideasthatareimperfectlyinstantiatedbymaterialthingsorasAristotelianformsthatareimmanentinmate-rialthingsasoneoftheirconditionsofexistence.However,asMillikanrightlypointedout,ataskofnaturalkindtheoriesisnottoexplainwhythingshavetheobservablepropertiesthattheyhave(whichiswhatkindessencesexplain)butwhymembersofthesamekindhavethesameorhighlysimilarproperties.12Andhere,explainingsimilarityinhigher-levelpropertiesintermsofsamenessofun-derlyingessenceamountstobeggingthequestion.Insum,theessence-viewofnaturalkindsgivesrisetoaviewofscienceasdi-videdinto“genuine”sciencesbuiltaroundnaturalkindsandotherfieldsnotbuiltaroundnaturalkindsandhencenottobeconsidered“genuine”sciences.Butastheessence-viewisproblematicinanumberofways,theensuingviewofscienceshouldbedoubted.3.THELAW-VIEWOFNATURALKINDSAsecondwidelyheldviewofwhatnaturalkindsarelinksnaturalkindstolawsofnature.PerhapsthemostprominentexampleofanargumentthatusesthisviewtoderiveconclusionsaboutthescientificstatusoffieldsofinvestigationisFodor’sargumentagainsttheideathatscienceisunified.13Inhisargument,Fodor12RuthG.Millikan,“ResponsetoBoyd’scommentary”,in:PhilosophicalStudies95,1999,pp.99-102(p.100).13Fodor,loc.cit.\nHowSpecialAretheLifeSciences?179startsfromthebasicassumptionthateveryscientificfieldofworkisorganizedaroundoneoranumberofnaturalkindsthatarespecifictothefieldinquestion.AccordingtoFodor,aparticularscientificfieldisindividuatedtoalargeextentbythepredicatetermsthatfeatureinthefieldinquestionbutnotinotherfields.Thosepredicatetermsthatfeatureinthelawsthatarepropertothefieldinques-tion,Fodorholds,denotethefield’sownnaturalkinds:“roughly,thenaturalkindpredicatesofasciencearetheoneswhosetermsaretheboundvariablesinitsproperlaws”.14Thenotionof‘law’istakenhereinasuitablybroadsense:thelawsofafieldofworkareitsexplanatorygeneralizations,wherebothnon-universalgeneralizationsanduniversallawsarecounted.15Fodor’sconceptionofnaturalkindsiswidespread:itisoftenclaimedthatnaturalkindsandlawsofnatureareinseparablyconnectedinthatnaturalkindsarethosekindsthatarementionedinlawsofnatureand,viceversa,lawsofnaturearethosegeneralizationsthatreachovernaturalkinds.Onthisview,asciencelikebiologywouldbeindividuatedbythepredicatetermsthataretypicalforbiologicaldiscourse,whichincludegeneralkindtermssuchas‘species’or‘gene’,aswellasspecifickindtermssuchas‘Pantroglodytes’or‘Drosophilamelanogasterantennapediagene’.Fodor’sargumenthingesontherelationsthatobtainbetweenthenaturalkindsofaparticularspecialscienceandthoseoflower-levelsciences,typicallyphysics.Anecessaryrequirementforafieldofsciencetobereducibletophysics,Fodorholds,isthateachofthepropernaturalkindsofthefieldcanbereducedtoanaturalkindofphysicsbywayofa1–1mapping.16Thatis,eachnatural-kindde-notingpredicatetermSofthefieldthatistobereducedmustbeconnectedwithnanatural-kinddenotingpredicatetermPofthereducingfield(here,physics)bynmeansofabridgelawoftheformSx↔Px.17But,Fodorargues,formanyofthennkindsthatfeatureinspecialsciencesthisrequirementisnotsatisfied:thespecialsciencescommonlymakeimportantgeneralizationsoverkindswhosemembersarenotallofthesamephysicalkind.Fodorobservesthatmanyspecialsciencesusefunctionallydefinedkindsasthebasesofgeneralizationsandarguesthatsuchkindstypicallycannotbemappedina1–1wayontophysicalkindsbecausefunc-tionstendtobemultiplyrealizable.Thatis,inmanycasesthefunctionthatindi-viduatesaspecial-sciencekindcanbeperformedbyvariousentitieswithdifferentmaterialstructuresthat–becauseoftheirstructuraldifferences–donotconstituteasinglenaturalkindofphysics.1814Ibid.,pp.98,102.15Millikan,“Historicalkindsandthe‘specialsciences’”,loc.cit.,p.55.16Ibid.,pp.102,104.17SandParepredicatesspecifictothespecialscienceandtophysics,respectively,nnandSxandPxaresentencesoftheform“xisaS”and“xisaP”.ThebridgelawnnnnSx↔PxthusstatesthatallthingsthatareS’sarealsoP’sandviceversa,inthisnnnnwaymappinganaturalkindofthespecialscienceunderconsiderationontoanaturalkindofphysicsina1-1manner.18Functionsaresaidtobemultiplyrealizableinthesensethataparticularfunctioncan\n180ThomasA.C.ReydonConsequently,thefieldsthatfocusonsuchkindsarenotfullyreducibletophysicsandaretobeconsideredastosomedegreeself-containedfieldsofworkthatoccupyanautonomouspositionamongtheothersciences.Asmanyofthekindsthatfeatureinthelifesciencesarefunctionallydefined,thiswouldhaveim-portantconsequencesforthestatusofthelifesciences:incontrasttotheanalysispresentedinSection2,Fodor’sanalysiswouldleadtotheconclusionthatthelifesciencesareautonomoussciencesthatdonotdependfortheirexplanationsonthephysicalsciences.Eventhoughmanyspecial-sciencekindsaredefinedbymeansofdifferentsortsofpropertiesthanphysical-sciencekinds(functionalvs.non-functionalproperties),allarepropernaturalkindsthatconstitutethefocalpointsoftheirownsciences.Hence–andthisisFodor’sgeneralconclusion–,scienceisnotaunifiedphenomenon:therearespecialsciences[…]becauseofthewaytheworldisputtogether:notallnaturalkinds(notalltheclassesofthingsandeventsaboutwhichthereareimportant,counterfac-tualsupportinggeneralizationstomake)are,orcorrespondto,physicalnaturalkinds.19AsageneralcritiqueofFodor’sargumentIwanttopointtosomeassumptionsthatunderlieit.Fodorsimplypositsthattheboundvariablesinthepredicativesen-tencesspecificforaparticularfieldofscience,importantlyincludingthoseboundvariablesthatrefertofunctionalproperties,denotethatfield’snaturalkinds.20Butitisunclearwhetherthisassumptionholds:mostphilosophersagreethatbecausefunctionsareoftenmultiplyrealizable,functionallydefinedkindsgenerallyaren’tnaturalkinds.21Thus,notallpredicatetermsthatappearinscientificgeneraliza-tionsnecessarilyrefertonaturalkinds.Ifthisconsensusviewiscorrect,theques-tioniswhetherthe(ir)reducibilityofaparticularscience’sfunctionalkindstooftenberealizedbyentitieswithdivergingmaterialstructures.Conversely,materialentitiestendtobemultiplyfunctionalinthatanentitywithaparticularmaterialstruc-tureusuallyisabletoperformdifferentfunctionswhenplacedindifferentcontexts.19Ibid.,p.113.Inadditiontothismetaphysicalargument,Fodormakestheepistemologi-calargumentthatitisnotevenimportantwhetherornotthephysicaldescriptionsofthekind’smembersarethesame,asthisdoesnotaffecttheepistemicimportanceofthegeneralizationsthataremadeoverthekindinthespecialsciencethatusesit(Ibid.,p.103).20Forasimilarassumptioninthecontextofasimilarargument,seeHaroldKincaid,IndividualismandtheUnityofScience:EssaysonReduction,Explanation,andtheSpecialSciences.Lanham,MD:RowmanandLittlefield,1997,p.75.21Accordingtosomeauthors,however,thereisnointrinsicdifferencebetweenfunc-tionalandnaturalkindsandatleastsomefunctionalkindsshouldberecognizedasnaturalkinds(e.g.,RichardN.Boyd,“Kinds,complexityandmultiplerealization”,in:PhilosophicalStudies95,1999,pp.67-98(pp.92-96);IngoBrigandt,“Naturalkindsinevolutionandsystematics:Metaphysicalandepistemologicalconsiderations”,in:ActaBiotheoretica57,2009,pp.77-97;ThomasA.C.Reydon,“Howtofixkindmem-bership:AproblemforHPC-theoryandasolution”,PhilosophyofScience,forthcom-ing).\nHowSpecialAretheLifeSciences?181naturalkindsofphysicshasanybearingontherelationshipbetweenthatscienceandphysics.Forreducibilitypresumablyisanissuethatarisesbetweenthesamekindsofelementsoftwodifferentsciences:thequestioniswhetherthetheories,laws,explanatorygeneralizations,naturalkinds,etc.ofonesciencearereducibletothetheories,laws,explanatorygeneralizations,naturalkinds,etc.ofanotherscience.Iffunctionalkindsarecategoricallydifferentfromnaturalkinds,tryingtoreducethefunctionalkindsofonesciencetothenaturalkindsofanotherscienceamountstoacategorymistake–andnothingmuchcanbeconcludedfromtheirreducibilityofthefunctionalkindsofonefieldtothenaturalkindsofanotherfield.TheupshotisthataslongasithasnotbeenestablishedthatFodoriscorrectinadmittingfunctionallydefinedkindstothenaturalkindfold,hisargumentrestsonaquestionableassumption.Asecond–andinthecontextofthesciencesoflifeandmindprobablymoreforceful–criticismofFodor’sargumentationisthatitcruciallydependsonas-sumptionsaboutthenatureofthefunctionsthatindividuatekindsinthespecialsciences.Forthedegreetowhichfunctionsactuallyaremultiplyrealizableseemstodependonwhatonetakesfunctionstobe.Ononeimportantnotionofbiologicalfunction,thatoffunctionasselectedeffect,functionalkindsseemmultiplyreal-izableonlyinaveryweaksense,asanentity’sfunctionisinseparablylinkedtothespecificselectionalhistoryofthatparticularentity.Anentity’sselected-effectfunctionisthatactivitythatitsancestorswereoriginallyselectedfor–thus,aselected-effectfunctionbydefinitionisafunctionthatcanonlybeperformedbyevolutionarilyrelatedentitiesthatsharethesameancestors.22Andevolutionarilyrelatedentitiestendtohaveverysimilar(thoughnotexactlythesame)materialstructurespreciselybecausetheystandinonelineofdescent.Thisisdifferentwhenoneadoptsanotionoffunctionasactivityorascausalrole,wherewhetherdifferententitiescanperformthesamefunctionisindependentoftheirrelated-ness.2322Whataboutconvergentevolution,i.e.,casesinwhichthesamefunctionhasevolvedindependentlyinanumberoflineages?AccordingtoMillikan(“Historicalkindsandthe‘specialsciences’”,loc.cit.,pp.59-60)akindindividuatedbyaconvergentlyse-lectedfunctionwouldnotconstituteanaturalkindoverwhichscientificallyusefulgeneralizationshold,asthemerefactthatthingsperformthesamefunction(causalrole)doesnotmakeitthecasethattheyaresimilarinanyotherrespects.Iwouldsuggestthatincasesofconvergentevolutionofthesamecausalrolefunctiontherearemultiplefunctional(asselectedeffects)kinds,eachindividuatedbyoneparticularselectionalhistoryofoneparticularlineage.Thus,‘vertebratelenseye’,‘insectfaceteye’and‘trilobitecrystallineeye’woulddenotedifferentfunctionallydefinednaturalkindsofbiologythatpresumablywouldneedtobefurthersubdividedintomorespe-cificnaturalkinds.23Foranin-depthdiscussionofthesedifferentnotionsoffunction–notconsideringmul-tiplerealizability,however–,seeArnoG.Wouters,“Fournotionsofbiologicalfunc-tion”,in:StudiesinHistoryandPhilosophyofBiologicalandBiomedicalSciences34,2003,pp.633-668;ArnoG.Wouters,“Thefunctiondebateinphilosophy”,in:Acta\n182ThomasA.C.ReydonSo,letmeagainturntothelifesciences.Whatkindsoffunctionsplaycentralrolesinthedifferentfieldsoflifescience?Accordingtosomeauthors,thelifesciencesareprimarily–orperhapsexclusively–concernedwithselected-effectfunctions.24Othersholdthatthefunctionsthatindividuatekindsinthelifescienc-esoftentimesarecausal-rolefunctionsoractivity-functions,sothatoneneedstodistinguishbetweendifferentnotionsoffunctionthatbiologistsuse,dependingontheparticularfieldthattheyworkin.25Itisunlikelythatthisdebatecanberesolvedbymeansofaglobalargumentaboutthenotionof‘function’inthelifesciences–andevenlesslikelywhenthenotionof‘function’throughoutthewholeofscienceisconsidered.Rather,acase-by-caseanalysisseemsmoreappropriate,explicatingthenotionoffunctionthatisinvolvedineveryindividualcaseinwhichlifescien-tiststalkaboutthefunctionsoftraitsorentitiesandgroupthingsintofunctionallydefinedkinds.ThismeansthatanoverallanalysisàlaFodorofthereducibilityoffunctionallydefinedlife-sciencekindstokindsofphysicsmustfail.Insum,thelaw-viewofnaturalkindsgivesrisetoaviewofscienceasencom-passingmanydistinctfields,physicalaswellasspecialsciences,thatallcenteraroundtheirpropernaturalkinds(featuringintheirproperexplanatorygeneraliza-tions)thatcannotbemappedontooneanotherina1–1manner.Allthesedifferentsciencesthushavetheirownplacewithinthewholeofscience,standingsidebysideascomparativelyself-containedandautonomousfieldsofinvestigation.Butasthelaw-viewisproblematicinanumberofways,theensuingviewofscienceistobedoubted.4.APOSSIBLEWAYOUTIhaveoutlinedandcriticizedtwobasicviewsofwhatnaturalkindsareandpoint-edtotheirdisparateimplicationsforassessmentsoftheplaceofafieldofworkwithinthewholeofscience.Apparently,theconclusionmustbethatanswerstoquestionsoftheform“Is…anaturalkind?”andconsequencesdrawnfromthemforthescientificstatusofaparticularfieldofworkarevacuousaslongasthereisnogenerallyacceptedtheoryofnaturalkinds.Differentviewsofwhatitistobeanaturalkindleadtodifferentanswersandhencetodifferentassessmentsofthesta-tusofthefieldinquestion.Moreover,ifdifferentauthorsbasetheirassessmentsondifferentaccountsofnaturalkindhoodthiswillyieldincompatibleresults:whileBiotheoretica53,2005,pp.123-151.24E.g.,KarenNeander,“Functionsasselectedeffects:Theconceptualanalyst’sdefense”,in:PhilosophyofScience58,1991,pp.168-184;Millikan,“ResponsetoBoyd’scom-mentary”,loc.cit.25Itseemsthatatleasttwonotionsof‘function’playaroleinbiology(KimSterelny&PaulE.Griffiths,SexandDeath:AnIntroductiontoPhilosophyofBiology.Chicago&London:UniversityofChicagoPress,1999,pp.223-224andreferencestherein),butprobablymore(Wouters,loc.cit.).\nHowSpecialAretheLifeSciences?183oneauthorusesviewAtoarguethatfieldFisagenuine,autonomousscience,anotherauthorusesviewBtoconcludethatfieldGmerelyconsistsinonecasestudyafteranother.Itisunlikelythatthiswillyieldmuchclarityabouttheplacesofdifferentfieldswithinthewholeofscience.UnderthesecircumstancesitseemsthatHackingwasrighttostatethatframingphilosophicalquestions(inthiscasethe“place/statusquestion”)intermsof‘naturalkinds’doesn’tdomuchbesidescomplicatingtheproblemandmakingithardertofindgoodanswers.However,thetwoviewsdiscussedabovearenotasdisparateasitmightseem.Bothattempttodojusticetowidespreadintuitionsaboutwhatitmeanstobeanaturalkind:theideathatwegroupthingsintokindsinsuchawaythatthesegroupingsrepresentsomeaspectsofthenaturalstateofaffairsandthatpreciselybecauseofthis,thesegroupingsaresuitableforgeneralizingabout.Theessence-viewgivesprioritytothemetaphysicalintuitionthatnaturalkindsrepresentas-pectsofwhateverorderthereisintheworld,orthatnaturalkindsinsomesenseexist“outthere”innature.26Thatwecanmakegeneralizedstatementsaboutnatu-ralkindsissecondary,asitisaconsequenceofwhatmakesagroupofthingsintoanaturalkindinthefirstplace.Thelaw-viewgivesprideofplacetotheepiste-mologicalintuitionthatnaturalkindsarejustthosegroupingsofthingsthatusefulgeneralizationsreferto.27Thelaw-viewencompassestheoftenassumedintimatelinkbetweenlawsofnatureandnaturalkinds,accordingtowhichlawsreachoverkindsandkindsarereferredtobythepredicatesthatfeatureinlaws,butgoesfurtherbyconceivingoflawsinanon-strictsense(recallthataccordingtoFodorthelawsofasciencearejustitsexplanatorygeneralizations).The“specialness”ofnaturalkindsthusliesinthefactthatwerefertotheminimportantgeneraliza-tions;whatmetaphysicallysupportsthisepistemicsuccessissecondary,asthereprobablyareagreatmanyfactorsinnaturethatcandothejob.Thisdifference,Iwanttosuggest,ismoreadifferenceinemphasisthanadeepincompatibilityofpositions.28Ultimately,thethoughtunderlyingthetwo26Cf.Dupré,loc.cit.,p.311;KathrinKoslicki,“Naturalkindsandnaturalkindterms”,in:PhilosophyCompass3,2008,pp.789-802(p.789).27Forexample:“naturalkindsreflectastrategyofdeferringtonatureinthemakingofprojectabilityjudgments”(Boyd,“Realism,anti-foundationalismandtheenthusiasmfornaturalkinds”,loc.cit.,p.139),“Naturalkindsarescientificcategoriespositedbyourtheoriesasepistemologicaldevices;insofarastheyhaveontologicalstatus,itisasfeaturesofthewaysinwhichcausalstructuresintheworldinteractwithourclas-sificatorypracticesinsuchawayastosupportreliableinductionandexplanation.Thenaturalnessofnaturalkindsconsistsintheiraptnessforinductionandexplanation.”(RobertoA.Keller,RichardN.Boyd&QuentinD.Wheeler,“Theillogicalbasisofphylogeneticnomenclature”,in:BotanicalReview69,2003,pp.93-110;p.102).SeealsoPhilipKitcher,TheAdvancementofScience:SciencewithoutLegend,ObjectivitywithoutIllusions.NewYork:OxfordUniversityPress,1993,p.172;Boyd,‘Kinds,complexityandmultiplerealization’,loc.cit.28SeealsoThomasA.C.Reydon,“Naturalkindtheoryasatoolforphilosophersofsci-ence”,in:MauricioSuárez,MauroDorato&MiklósRédei(Eds):EPSA07:Launchof\n184ThomasA.C.Reydonviewsisthesame.Afterall,ifsuccessfulgeneralizationswouldnotreflectaspectsofwhateverorderthereisintheworld“outthere”,theirsuccesswouldremainamiracle.(Onecouldacceptthatoccasionallyageneralizationissuccessfulbyac-cident,butnotthatthishappensallthetime.)Conversely,ifthereissomeorderintheworld,itwouldbesurprisingifthiswouldn’tbereflectedtosomeextentinourgeneralizations.Thequestion,then,iswhereemphasisshouldbeplacedwhenaskingwhetheraparticulargroupingisanaturalkind–onmetaphysicalorepiste-mologicalcriteriaforbeinganaturalkind.Thisquestioncanbeansweredbyexaminingtheprospectsforsuccessofbothapproaches.Conceivingofnaturalkindsasthosegroupingsofthingsforwhichakindessenceexistsis,asexplicatedabove,problematicfortworeasons:itrestsonapriorimetaphysicalassumptionsaboutwhatitistobeanaturalkindanditfailstoadequatelyexplainoneofthemainexplanandaofnaturalkindtheory(namely,whythemembersofakindaresimilar).Itaketheseproblemsasfatalforanymetaphysics-firstapproach.Forbeginningwithsettingupametaphysicsofnaturalkindswillalwaysinvolveadoptingsomeaprioricriteriaofwhatitistobeanaturalkindthatincludesomegroupingsandexcludeothersfrombeingfurtherconsidered.Themetaphysics-firstapproachwill,therefore,sufferfromacertaindegreeofmyopiaaboutnaturalkindhood.Takingnaturalkindsassimplythosegroupingsoverwhichwesuccessfullygeneralize,however,entailsexactlytheoppositeproblem.Onthisview,anygen-eralizationthatissufficientlystable(i.e.,hasproventoholdwithoutconfrontingtoomanyexceptions)shouldbetakenasreferringtoanaturalkindandanygrouptermfeaturinginasuccessfulgeneralizationthusshouldbetakenasanaturalkindterm.29Butthisgoesagainsttheintuitionthat,askindsrepresentwhateverorderthereisintheworld,anyentityisatmostamemberofalimitednumberofkinds–thatis,thatthequestion“Whatkindofthingisthis?”issusceptibleonlytoalimitednumbercorrectanswers.30Andastheretypicallyareinnumerablymanysimilaritiesbetweenentities,countingeverysimilaritythatappearsinagenerali-zationasindividuatinganaturalkindwouldalloweveryentitytobeamemberofinnumerablymanynaturalkinds.ThinkingofnaturalkindsassimplythosegroupsreferredtoinsuccessfulgeneralizationsentailstheriskofbeingtoogenerouswiththeEuropeanPhilosophyofScienceAssociation.Dordrecht:Springer,forthcoming.29ThatthisisindeedwhatsomeadvocatesofthisapproachhaveinmindcanbeseenfromBoyd’sremarksthat“weshouldalwaysrequirethesortofsemanticmachineryindicatedbythetheoryofnaturalkindswhenouraimisinductionandexplanation[…].Kindsusefulforinductionorexplanationmustalways“cuttheworldatitsjoints”inthissense:successfulinductionandexplanationalwaysrequirethatweaccommo-dateourcategoriestothecausalstructureoftheworld.”(Boyd,“Realism,anti-foun-dationalismandtheenthusiasmfornaturalkinds”,loc.cit.,p.139;emphasisadded).30Inthecaseoforganisms,forexample,“It’saDrosophilamelanogaster.”,“It’sanin-sect.”,etc.,butnot“It’sablackthing.”,“It’saflyingobject.”\nHowSpecialAretheLifeSciences?185respecttoallowinggroupsintothenaturalkinddomain:iftakentoitsextreme,ultimatelytherewillbenonon-naturalkindsleft.31Still,theepistemology-orientedviewofnaturalkindshasbetterprospectsforsuccessthanthemetaphysics-orientedview.Forone,aswedonothavedirectac-cesstothenaturalkindstructureoftheworld(ifthereissuchastructure),thebestwecandoisexaminethosekindsthatfeatureinsuccessfulattemptstogeneralize,systematizeknowledge,explainphenomena,predictfutureevents,etc.–i.e.,insuccessfulepistemicpractices.32Theproblemiswheretostop:shouldallusefulgeneralizationsbeconsideredasbeingaboutnaturalkinds,ordoonlygeneraliza-tionsofaprivilegedsortrefertonaturalkindsand,ifthelatter,whatmakesthesegeneralizationsprivileged?Thequestionishowtotruncatethedomainofgener-alizationsthatrefertonaturalkindsinanon-arbitrarymanner.Ausefulheuristicmightbethefollowing.AsMillikanpointedout,itsurelyisnotthecasethateverysingle“good”generalizationreachesoveritsownnaturalkind.Ifwetaketheintuitionseriouslythatnaturalkindsinsomewayrepresentthefurnitureoftheworld,itisplausiblethateverynaturalkindshouldlieattheintersectionofmanygeneralizations.33Theunderlyingthoughtisthatkindsthatcometofeatureinmoreandmoredistinctgeneralizationsastimeprogresseswillbecomeestablishedintheontologiesofthefieldsthatusethem.Incontrast,kindsofwhichtheusefulnesspersistentlydoesnotgobeyondfeaturinginjustoneortwogeneralizationswillultimatelybecomeontologicallysuspect.Althoughthisdoesnotholdexclusivelyforthesciences,butgenerallyinallcasesinwhichthingsaregroupedintokinds,thekindsthatfeatureinthesciencesaretypicallyofthesortoverwhichmanydifferentgeneralizationscanbemade:“[a]sciencebeginsonlywhen,atminimum,anumberofgeneralizationscanbemadeoverinstancesofasinglekind”.34MillikantracedthisinsightbacktoJohnStuartMill,whowroteaboutkindsthatahundredgenerationshavenotexhaustedthecommonpropertiesofanimalsorplants…nordowesupposethemtobeexhaustible,butproceedtonewobservationsandexperi-ments,inthefullconfidenceofdiscoveringnewpropertieswhichwerebynomeansim-pliedinthosewepreviouslynew.3531Forasimilarpoint,seeMillikan(OnClearandConfusedIdeas:AnEssayAboutSub-stanceConcepts,loc.cit.,p.15).AnexampleisDupré’s“promiscuousrealism”aboutnaturalkinds(JohnDupré,TheDisorderofThings:MetaphysicalFoundationsoftheDisunityofScience.Cambridge(Mass.):HarvardUniversityPress,1993;‘Naturalkinds’,loc.cit.,p.318).32Boyd,“Kinds,complexityandmultiplerealization”,loc.cit.,p.69.33Millikan,“Historicalkindsandthe‘specialsciences’”,loc.cit.,p.48;OnClearandConfusedIdeas:AnEssayAboutSubstanceConcepts,loc.cit.,p.15,17.34Millikan,“Historicalkindsandthe‘specialsciences’”,loc.cit.,p.48.35Mill,quotedinMillikan,OnClearandConfusedIdeas:AnEssayAboutSubstanceConcepts,loc.cit.,p.16.ThisisalsothegistofRussell’sassertionthatanaturalkindisaclassofthingsthatpossessanumberofpropertiesthatarenotknowntobelogically\n186ThomasA.C.ReydonThesuggestion,then,istoconceiveofnaturalkindsasgroupsofentitiesthataresimilarinnumerousrespects,wherethesesimilaritiesarenotaccidentalbutduetocausalfactorsinnature.Thatis,themembersofanaturalkindaresimilarforareason.36Thenaturalkindsofaparticularfield,then,arethekindsthatlieattheintersectionsofmanyofthatfield’sgeneralizations–notsimplythosethatfeatureinanyofthescience’sgeneralizations(asonFodor’saccount),northosethatmeetstrictmetaphysicalcriteriaforbeinganaturalkind.Naturalkindscanthusbefoundbysearchingforkindsthatfeatureinmanygeneralizationsandinvestigat-ingwhichcausalfactorsinnatureunderliethesimilaritiesbetweenthemembersofaputativekind.Aswhetherakindtermfeaturesinmanygeneralizationsisaheuristicfactor,thequestionhowmanygeneralizationscountas“many”isnotaseriousproblem.Thiswayofthinkingaboutnaturalkindsavoidstheproblemsconfrontedbytheessence-view(apriorismaboutthenatureofnaturalkinds)andthelaw-view(countingallkindsreferredtoingeneralizationsasnaturalkinds)whiledoingjusticetothecentralintuitionsthatunderlieboththeseviews.5.HOWSPECIALARETHELIFESCIENCES?So,howdoesthisviewofnaturalkindsworkoutwhenanalyzingtheplaceofthelifescienceswithinthewholeofscience?Afirstthingtonoteisthatiffeaturinginmanygeneralizationsistakenastheprincipalindicationofnaturalkindhood,itremainsunclearhowkindsandfieldsofworkarerelated.Asonekindcanfeatureingeneralizationsinanumberoffieldsofinvestigation,foreverykindthequestionarisestowhichfielditshouldbeallo-cated.Conversely,foreveryfieldofinvestigationthatisrecognized,thequestionariseswhichareitspropernaturalkindsandwhicharemoremarginalbecausetheyactuallyareproperkindsofanotherfield.Forinstance,does‘Drosophilamel-anogasterantennapediagene’denoteapropernaturalkindofgenomics,genetics,developmentalbiology,orsimplyofbiology?Presumably,therelationbetweenfieldsandkindsneedstobeassessedonacase-by-casebasis.SuchquestionsdonotariseonFodor’saccount,asFodorassumesthatafieldisindividuatedbyitsownproperpredicates.Butifkindstypicallyfeatureinmanygeneralizations,thiscannotgenerallybethecase,asthereisnoreasontoassumethatthegeneraliza-tionsthatrefertoakindKwillallfeatureonlyinonefield.interconnected(BertrandRussell,HumanKnowledge:ItsScopeandLimits.London:GeorgeAllen&Unwin,1948,p.317).Millikan(“Historicalkindsandthe‘specialsciences’”,loc.cit.,p.57)thinksofthe“manygeneralizations”ideaasspanningacontinuumofclearly“good”naturalkindstoclearlynon-naturalkinds.NotethatMil-likan(ibid.,pp.47-48)presentedherapproachasafurtherelaborationofFodor’s.36Millikanspeaksofa“realground”,“ontologicalground”,ora“groundinnature”thatunderliesthesimilaritiesbetweenthingsofakind;Millikan,ibid.,p.50;OnClearandConfusedIdeas:AnEssayAboutSubstanceConcepts,loc.cit.,pp.16-18,25.\nHowSpecialAretheLifeSciences?187Beforeexaminingthekindsofthedifferentfieldsofworkthatmakeupthelifesciences,then,thesefieldsthemselvesneedtobeindividuatedfirst.Butnowanobviousproblemisthatthereisnostandardlistofallthevariousfieldsoflifescience.Simplyusingcommonsenseandlistingthefieldsthatarementionedinthestandardtextbookswon’tdo,asthismightresultinatoofine-grainedortoocoarse-grainedlist.Shouldourlistread:ecology,physiology,genetics,…–orshoulditbe:organismalecology,populationecology,communityecology,ecosys-temecology,evolutionaryecology,animalphysiology,plantphysiology,devel-opmentalgenetics,comparativegenomics,…?Shouldphilosophersexaminetheplaceofecologyamongthesciencesbylookingatecologicalkindsingeneral,ortheplacesoforganismal,population,community,etc.ecologybyexaminingeachofthesefields’properkindsseparately?Orperhapsboth?Andwhathappensincasesinwhichonekindcanbeequallylegitimatelyallocatedtoseveralfields?Theanswersdependonthecriteriathatoneadoptsforindividuatingscientificfields;andtheredonotseemtobeanyunequivocalcriteriaavailable.Ononeaccount,forexample,afieldisanareaofscienceconsistingofthefollowingelements:acentralproblem,adomainconsistingofitemstakentobefactsrelatedtothatproblem,generalexplanatoryfactorsandgoalsprovidingexpectationsastohowtheproblemistobesolved,techniquesandmethods,and,sometimes,butnotalways,concepts,lawsandtheorieswhicharerelatedtotheproblemandwhichattempttorealizetheexplanatorygoals.Aspecialvocabularyisoftenassociatedwiththecharacteristicelementsofafield.[…]Fieldsemergeinscience,evolve,sometimesevenceasetobe.[…]Althoughanyoralloftheelementsofthefieldmayhaveexistedseparatelyinscience,theymustbebroughttogetherinafruitfulwayforthefieldtoemerge.Suchanemergenceismarkedbytherecognitionofapromisingwaytosolveanimportantproblemandtheinitiationofalineofresearchinthatdirection.37Tobesure,thisisaratheropen-endedwayofindividuatingfieldsthatdoesnotprovideanystrictcriteriabymeansofwhichfieldsofinvestigationcanbedelim-itedoritcouldbedeterminedwhichkindsaretobeallocatedtowhichfields.Theconclusion,then,mustbethatonthecurrentstateofaffairsinphilosophythequestion“Howspecialarethelifesciences?”cannotbeanswered–atleastnotbyrelyingonananalysisoftheputativenaturalkindsthatfeatureinthevari-ousfieldsoflifescience.Theanswercouldemergeafteraconsensusemergesonwhichdifferentfieldsofinvestigationmakeupthelifesciencesandtheproperkindsofeachfieldhavebeenestablished.Quitelikely,itwillbedifferentfordifferentfields.Thereis,afterall,noreasontoassumethatscientificstatusisanall-or-nothingaffair(statusoftencomesindegrees,afterall)orthattheplaceofafieldwithinthewholeofsciencewillbethesameforallfieldsoflifescience.To37LindleyDarden&NancyL.Maull,“Interfieldtheories”,in:PhilosophyofScience44,1977,pp.43-64(pp.44-45).Forasociologicalaccountofscientificfields,seeRichardWhitley,TheIntellectualandSocialOrganizationoftheSciences(secondedition).Oxford:OxfordUniversityPress,2000.\n188ThomasA.C.Reydonbesure,thisconclusionisunsatisfactory,asitmeansthatusingnaturalkindtheorytoanalyzetheplaceofthelifescienceswithinthewholeofsciencewillinvolvealotofwork,mostofwhichstillremainstobedone.Inaddition,theprospectsofultimatelyreachinganunequivocalanswerarenotclear.ButIhopethatatleastsomeclarityhasbeenachievedaboutwhataprojectthatattemptstousenaturalkindtheoryinthiscontextshouldlooklike–andthatithasbecomeclearwhyask-ing“Is…anaturalkind?”isnottherightquestiontostartfrom.CenterforPhilosophyandEthicsofScience(ZEWW)LeibnizUniversitätHannoverImMoore21D-30167HannoverGermanyreydon@ww.uni-hannover.de\nMILESMACLEODTHEEPISTEMOLOGY-ONLYAPPROACHTONATURALKINDS:AREPLYTOTHOMASREYDONWhatI’mgoingtocommentonheremostlyaretheissuesofnaturalkindsthatThomasReydonraiseswithrespecttohis‘epistemology-first’strategy,ratherthanspecificallytheproblemofthespecialsciences,althoughisclearthatitiswithrespecttothespecialsciencesandtheproblemoffielddemarcationthatanynotionofnaturalkindhasitsmostpressingapplication.I’mgoingtosuggestaperhapsratherpragmaticextensionofthisstrategywhichwouldattempttoprovideade-greeofphilosophicalusefulnessinthisregardbutavoidthetraditionalproblemswhichIdon’tthinkweescapeunlessthisstrategyispushedtothelogicalconclu-sionofanepistemology-onlyapproach.AsReydonmentionsinhistext,thereistodayaratherlargeshadowloomingoverthosewhotradeinnaturalkinds.ThishasbeencastbyIanHacking,whohasinhisinimitablewaydeclaredtheconcepttobearbitraryinitsmanyinconsistentformulations,tohavefailedinitschiefambitionsofprovidingageneralaccountofcategorisationinscience,andultimatelytobeofnoconsequencetophilosophyofscience.1ButifwereadHackingcloselywhatIthinkwefindheisreallyobjectingtoisaunifiednotionofnaturalkind;namely,thatsomesingleunitarylogicalormetaphysicaldefinitiongroundstherelationsandpropertiesofseeminglydiverseclassstructuresinscience.“Itistheideaofawell-definedclassofnaturalkindsthathasself-destructed.…”2Yetitispreciselyinthedirectionsofmorepluralisticandmorenaturalisticunderstandingsofnaturalkindsawayfromessentialismthattheoristshavebeenheading.Hackingshowshissympathytothis.InthepaperIamcitingforinstancehecriticisesalmosteverybodybutseemswillingtoenter-tainRichardBoyd’sHomeostaticPropertyClusterconceptionforspeciesandthelike,althoughhethinksitnotrelevantforallpotentialnaturalkinds(whichIdon’tbelievewaseverBoyd’sclaim).3Hedoeshoweverprefertoregardthesemorepluralisticapproachestonaturalkindsassimpleevidenceofthevacuousnessof1IanHacking,“NaturalKinds:RosyDawn,ScholasticTwilight”,in:RoyalInstituteofPhilosophySupplement,82,2007,pp.203-239.2Ibid.,p.209.3TheclustertheorygoesbacktopaperssuchasRichardBoyd,“HowtobeaMoralRe-alist”,inG.Sayre-McCord(Ed.),EssaysonMoralRealism,Ithaca:CornellUniversityPress1988,pp.181-228;RichardBoyd,“WhatRealismImpliesandWhatitDoesNot”,Dialectica,43,1989,pp.5-29;RichardBoyd,“Realism,Anti-Foundationalism,andtheEnthusiasmforNaturalKinds”,PhilosophicalStudies,61,1991,pp.127-148.F.Stadler(ed.),ThePresentSituationinthePhilosophyofScience,ThePhilosophyofScienceinaEuropeanPerspective,DOI10.1007/978-90-481-9115-4_15,©SpringerScience+BusinessMediaB.V.2010\n190MilesMacLeodthenaturalkindconcept.Idon’tthinkthat’sright.Iftheimperativeisthatnaturalkindsbeshowntobephilosophicallyusefulinouraccountsofscienceandsolvingtheproblemsthatemergetherein,thenIthinkthebesthopeisanapproachthatiswillingtoacceptthatitisnotconfinedtosomeunitarymetaphysicalaccountofwhatnaturalkindsare.WhileHackingobjectstotheepithet‘natural’asappliedtokinds,hedoesappreciatethatkindsthemselveshavecertainepistemicvalueinscientificpracticeinforinstanceinductionandexplanation,andit’sexactlyIbelievealongthelinesoftheseobservationsthatnaturalkindscanbegivensenseandseentobeusefulconceptually,ifnotessential,toourunderstandingofscien-tificpractice.Theapproachtonaturalkindswhichseemsmorepromisinginthisrespect,andIwillgivesomegoodreasonsforthinkingthisbelow,treats‘naturalkinds’asfirstandforemostepistemicdevices,asepistemicallyrelevantgroupings:thatis,toolsofinductivegeneralisationandexplanation.Itletssciencedecideineachinstancethroughitsowninvestigativeprocessesandinitsownlanguagewhattheunderlyingreductivebasisofthissuccessmightbe.Thisisoftenanaturalandcen-tralpartofafield’sinvestigativeprocesses(investigatingitsownnaturalkinds),butitneedn’tbe.Scientistsmightworkatalevelofexplanationwherereductioninthiswayissimplynottheaim.Thesuccessitselfjustifiestheirbeliefinitandinturn,relianceonit.Reydon’sepistemology-firstapproachfitshimwithinthislineofthought.SoletmestartbyexpressingwhyIthinkitmakessensetoseekananswertotheproblemsoffieldsandthestatusofspecialsciences,andotherproblemstoo,throughthisepistemicallybasedcriterionforidentifyingkinds.AsReydonpointsoutarealproblemwithusingnaturalkindstoresolvephilo-sophicalproblemsuchasthestatusoffieldsinthespecialsciences,isthestarkdependencethishasonthenotionofnaturalkindonechoosestosupportandtheextremesthatresult.Essentialismincludesalmostnothinginthespecialsciences,aFodorianapproachalmosteverything.4Mostwouldagreehoweverthatdespitetheinconclusivenessofthedebatethereisasensetotheconceptthatstemsfromitsroleandsuccessinscientificpractice,evenifapreciseontologicalformulationofwhatitistobeanaturalkindcan’tbegiven.Thisintuitiongetslostperhapsbytheattributiontonaturalkindsofcontentandstructurethattriestoclassifyoraccountforthemintermsforinstanceof‘essences’or‘functions’whichturnoutultimatelytobeproblematicnotions.Butthefactisthereremainsapracticeofusingkindsgenerallyinscienceforepistemicpurposes.Anditisnotjustexplanationandgeneralisationthatgivesthe4SeeJerryFodor,“SpecialSciences(or:theDisunityofScienceasaWorkingHy-pothesis)”,Synthese,28,1974,pp.97-115.OnFodor’sapproach“thenaturalkindpredicatesofasciencearetheoneswhosetermsaretheboundvariablesinitsproperlaws”(p.102).Thisofcourseleavesopenthenotionofwhata‘properlaw’iswhichFodoradmitshecan’tsay,exceptthatlaterinthetexthearguesthat‘lawlikeness’itselfdependsonitsrelatabeingnaturalkindterms.Itappearshehassomekindofequiva-lencerelationinmindbetweenlawsandnaturalkinds.\nTheEpistemology-onlyApproachtoNaturalKinds191useofthesekindstheirepistemiccharacter.Theyformtheveryconceptualbasisofscientificsystemsofthoughtandreasoningandthebasisuponwhichresearchisorganisedanddirected,sinceitisintermsofkindsthatresearchquestionsandissuesofinvestigationsareoftenrealisedandposed,andatthesametimeresolvedbyassigningnewpropertiestothemandformingnewrelationsbetweenthem.Theyareepistemicinsofarastheyareabasisuponwhichscientificknowledgeissoughtfor,obtainedandformulated.Nowifultimatelyourintentionistoun-derstandthebasisofvariouselementsofscientificpractice,suchasthepracticeoforganisingscienceinfields,includingwhatmakesthemdistinctfromoneanother,thenIthinkitisnaturaltoperceivethatnaturalkindswhenevaluatedandidenti-fiedintheseepistemictermsarecriticaltothis.Afterallascientificfieldisinprac-ticeitself‘epistemic’.Onewayoflookingatafieldisasagroupingofphenomenaonthebasisofconnections,patternsandrelationsthatsuggestthereexistreveal-inggeneralisationsandunderlyingexplanationstothisorder.Apersistingfieldisoneinwhichsuchgeneralisationoccursandsuccessfulexplanatoryframeworkshavedeveloped.Discoveringwhatdefinesormakesafieldthendependsonwhatunderliesthesegeneralisationsandexplanatorysuccesses,andhereofcourseiswherewewouldanticipatetheprimaryandactiveepistemicroleofnaturalkindsasthebasesofthese.Itakeitthatthisisamuchmoreepistemicapproachtofieldsandnaturalkind.ButIthinkthisistheway‘naturalkind’emergesasarelevantconcept.Thegoverningpresuppositionisthatthestatusofafieldisnotdirectlyunderwrittenbymetaphysicalconsiderations,butbyepistemicones,suchasitsinvestigativeandexplanatoryactivity,andsuccessinthisregard.ThisisaperspectivewhichaddssupporttowhatReydonisdoing.OnthisbasisIthinkheisrighttothinkthattherearegoodprinciplesatleasttofavourtheepistemologicalapproach,ashecallsit,intheattempttodiscoveramoreastuteandusefulnotionof‘naturalkind’.Hisnoveltyinthisregardistoposeheuris-tically,whatwemightplausiblythinkisalsoastandardforscientificpracticeforidentifying‘naturalkinds’,whichisthesuccessofakindinmultiplediffer-entgeneralisations.Letmegivesomebroaderreasonsforthinkingthatkindsforwhichtherearemultiplegeneralisationsshouldbeexactlythebasiswhichisreliedoninpracticetoidentify‘naturalkinds’sinceitunderwritesthekindsofspecialepistemicrolestheyhave,whichincludestheorganisationoffields.Afterallifakindcanbeassociatedwithvariousdifferentgeneralisationsthenthereisanobvi-ousestablishedusefulnesstotheconceptasasystemisingelementforwhichtheroleasaunifyingprinciplenodoubtgivesitexplanatoryvalue…thinkofthekind‘acid’forinstancewhichisexplanatorilyusefulinmanydifferentcircumstancesbecauseofthenumberofgeneralisationsthataremadeaboutit.AsReydonputsitnaturalkindsarethosekindsthatlieatthe‘intersection’ofmanyofthefield’sgeneralisations,whichsuggestsacentralitytothosekindsinjustthisregard.Atthesametime,apointnoticedbyMillikanbutnotquiteputintheepistemictermsI’mputtingit,thesuccessofmanygeneralisationsunderwritestheexpectation\n192MilesMacLeodthatthekindisasubjectofinvestigation(byhavingareality)andthushasfurtherdiscoverableproperties,relations,ultimatelygeneralisations…andtherebyfur-therexplanatoryapplications.5Infactit’sonthebasisthatwecanexplicatewhatthesenseof‘natural’contributestotheconceptinepistemicterms.Partofwhat‘natural’impliesofakindlikephosphorousforinstanceisthatthereisalwaysmoretosayofitanddiscoverofit:thatitisnotexhaustedbyanysetofpropertiesoronedescription.AsMillikanputsit,“sciencebeginsonlywhen,atminimum,anumberofgeneralisationscanbemadeoverinstancesofasinglekind”.6Thisidentificationof‘natural’with‘manygeneralisations’thusreflects,Iwouldpos-tulate,thedecisionsscientiststhemselvesmakeinmanycasesaboutwhatcountsasanaturalkindwhichinturngovernshowtheyemployitandrelyuponit(par-ticularlyif,likeFodor,wethinkreductionismisnotnecessarilyaregulativecon-straintontheestablishmentofnaturalkindsinscientificpractice).Ifsothenthe‘multiplegeneralisations’testforidentifyingnaturalkindspicksoutasignificantaspectoftheepistemicbasisofsuchkinds,andbyvirtuepicksoutthegroupingsfromwhichfieldsthemselvesacquireepistemicvalueandsuccess,andpresentasproductivescientifically.Themeaningof‘natural’herethenplaysoutintermsofthecentralbeliefscientistshaveinthekindandthecorrespondingapplicationthatismadeofitasaresultofthosebeliefs.Ofcourseallthisismostlysupportive,butIthinkitcompelsadifferentbutnonethelesslogicaloutcomefromtheoneReydonaspiresto.IsuggestthattotakethekindofapproachI’mexpressingonereallyhastobewillingtoconcernoneselfsolelywiththepracticeofnaturalkindsassodefined.AttemptingtoprescribeanykindofontologicalcriterionIthinkcompromisesthisusefulness,becauseitwillinevitablycutacrossthispracticaldimension,especiallywithrespecttothespe-cialsciences.Reydonexpressesthatthe‘manygeneralisations’viewpointistobetreatedasaheuristicstrategyforfindingnaturalkindcandidates,butnotitselfasawayofdemarcatingnaturalkinds.Tofind‘real’naturalkindswestillmustturnourselvestodiscoveringwhetherornotthereisacausalstructureunderlyingthosekinds.Theremustbeinotherwordssomenaturalisticexplanationforthesuccessofapotentialnaturalkindterm,i.e.somebasisfortreatingitsmembersasagroupwithwhichonecanprojectandexplain,toconsideritanaturalkind.Iwanttoputhoweverthatifouraimisausefulnotionofnaturalkindforfathomingscientificpracticethenthisaddedconditioniscounterproductive.Firstlyinmanyinstancesfieldsemploytermsasnaturalkinds,relyingontheminthesesenseabove,withoutbeingabletoreducethemorexplainthematadifferentleveltheirsuccess.Itisoftentakenforgrantedthatthereisanunderlyingbasiswhichmightbecomplexormultiply-realisablebuthasthecoherencenonethelesstoprovidethissuccess5RuthMillikan,“HistoricalKindsandtheSpecialSciences“,PhilosophicalStudies,95,1999,pp.45-65.6Ibid.,p.48.\nTheEpistemology-onlyApproachtoNaturalKinds193andmakeitaconceptthatcanberelieduponfurther.Thussettingthesekindsofconditionsthreatensourchanceofaccountingforonehighlyimportantaspectofpractice;usingkindsasnaturalkindswithoutreduction.SecondlyIwouldtakeithoweverthatthereisstilltheproblemofwhattodoaboutmultiple-realisability.Howdowetreatkindsthatarediscoveredtohavemorethanonecausalbasis?Aretheynaturalkindsornot?Eventuallythisisthekindofquestiononepresumablyhastoresolveifonewishestodemarcatethespecialsciences‘metaphysically’bynaturalkinds.Whichmeanswearerightbacksomewhatwherewestartedwiththeseoldfamiliarmetaphysicalproblems.Thesehoweverrunadversetothecon-ceptionofausefulnotionofnaturalkind.Certainlyifyourgoalisdescribingtheboundariesofafield,asopposedtosaythescientificrecognitionofafield,thentakingapositionseemstopredeterminetheanswer.Yetintermsoftheirepis-temicrolesgeneratinginductionsandexplanationsforinstancethenaturalkindsfunctioninsimilarwaysinthosedeemedtobefieldsandthosedeemednotbyametaphysicalstandard.Surelytheanswertowhatdemarcatesfieldsisaquestionofpracticeinrespectoftheuseofconceptsandthebeliefsinvolved,butnotoneofmetaphysics.AsReydonpointsoutFodor’sviewpoint,whichtreatsfunctionally-definedkindsasnaturalkinds,iscriticisedbecauseafunctionalkindseemslikeadifferentkindofthingfromanaturalkindandbecause‘functional’isnotinterpretableinanyonepreciseway.ButofcourseReydon’sheuristicitselfwouldseemtoac-ceptthatatleastattheoutsetfunctionalkinds,howeveronedefinesfunctional,aspotentialnaturalkindsbecausefunctionalkindsmightwellalsobekindsthataresuccessfullyinvolvedinmultiplegeneralisations.Butifweattempttoparekindsdownbytheircausalbasisinsuchawaythenpresumablywelosetheconnectionwemightotherwisebeabletoestablishinageneralwayinscientificpracticebetweentheepistemicvalueofnaturalkindsandthedemarcationandstatusofafield.SoIwanttoputIsupposethefollowingquestion.IfouraimiscontraHackingtohaveausefulconceptofnaturalkind,whynotjustemployitasanepistemiccategoryforwhichwecacheout‘natural’intermsofthewayinwhichscientistsplacetheirbeliefsandusetheconceptbyvirtueofthosebeliefs.Naturalkindsinthiswayareseenastoolsofpracticeandareexplanatoryforphilosophersasdevicesthatexplainhowscientificpracticefunctions,includingitsdivisionintofields.Dowereallyneedtomorethanthis?Isitjustcounterproductivetoexpectmore?Wenotethisviewisn’taswide-openasFodor’sbecausetheepistemiccri-terionofmany-generalisationsisstricterandamorecompellingbasisforbeliefsinanunderlyingrealitytothekind.Butwedon’tneedtospecifywhatthisrealityneedstoconsistin,ortoputitanotherway,it’snotourintereststodoso.YettheratherobviousobservationshouldbemadethateventhoughIthinkthere’sgoodreasontopursuethisepistemologicalapproach,asReydondoes,itseemstofailinthetasksetforitinhisproposalshere.Relyingona‘manygeneral-isations’heuristicwon’tinfactalonehelpwiththetaskofdemarcatingfieldsand\n194MilesMacLeodgivingthemstatusthereby.Itdoesn’tseemlikefieldsarethemselvessimplybuiltatopasetofrefinednaturalkinds.Afterallidentifyingnaturalkindsby‘manygeneralisationsofafield’requiresfirstpickingoutwhatthegeneralisationsofafieldare.LetmesaythatIthinktheimpressionthatafieldinthespecialsciencesismorecomplexthanasetofnaturalkindsissurelytrue.Onecan’tdefineafieldsimplyintermsofthem.Butthisdoesn’tstopusmaintainingasIhaveabove,thatnaturalkindsareanessentialpartofthewayafieldoperatesasanepistemicunitandessentialtoanyunderstandingofthis.Thisiswheretheusefulofthenaturalkindconceptlies.Andhoweverwechoosetodefinefields,naturalkindswillbeintegratedessentiallyintothisdefinition.Afterallafieldmaywellhavecentralproblemsbutthoseproblemsthemselvesmaywellbeproblemsofthenaturalkindstructure,oratleastexpressedinthevocabularyofthesekinds.Themethodologyofthefieldwillitselforganiseitselfaroundthekindsitconsidersfundamentalandsoon.Naturalkindsrepresentacentralpartofthecategoricalstructurewithandthroughwhichafield’sscientistsinterprettheworld,organiseandunderstandthephenomena.Theyguidehowtheworldisfurtherinvestigated.Obviouslythetheoryweneedmustbesensitivetothecomplexityinvolved,butnaturalkindsaspartofthebasicepistemicstructureofafieldwillsurelybepartofthetoolsofphilosophersforunderstandingthe‘field’asaunitofscientificpracticeandfind-ingwhatgivesonestatus,atleastforthescientistsinvolved.Initiativkolleg‚NaturwissenschaftimhistorischenKontext‘UniversitätWienRooseveltplatz10/91090WienAustriamiles.macleod@univie.ac.at\nMEHMETELGINREDUCTIONISMINBIOLOGY:ANEXAMPLEOFBIOCHEMISTRYABSTRACTInthispaper,Iarguethatthemultiplerealizabilityargumentagainstreductionismdoesnotworkinbiochemistryandthatbiochemistryasareductionistprojectisaprogressiveresearchprogram.Sincetheanti-reductionistargumentthatappealstothemultiplerealizabilitythesisdoesn’tworkandsincebiochemistrythatincorpo-ratestheprinciplethatbiologicalfunctionsofbiomoleculesinlivingcellscanbeunderstoodintermsofchemicalandphysicalpropertiesofthosemoleculesisaprogressiveresearchprogram,Iconcludethatplausibilityofreductionismisstillworthyoffurtherstudy.IReductionisminbiologyisconcernedwiththerelationbetweenbiologicalknowl-edgeandchemicalorphysicalknowledge.1Thereistheideaoftheory-reduction,whichconcernswithwhetherandhowabiologicaltheorycanbereducedtoachemicaloraphysicaltheory.Thereistheideaofexplanatoryreduction,whichconcernswithwhetherorhowbiologicalrepresentationscanbeexplainedbychemicalorphysicalrepresentations.2Inthispaper,Iwillonlyfocusonbiochem-istryandIwillnotdiscussthenatureofreductionrelation.Instead,Iwillarguethatthemultiplerealizabilitythesisdoesnotshowthattype-typereductionisnotpossibleatleastinbiochemistry.Second,Iwanttoaddresstheissueofreduction-isminadifferentwaybylookingatascience(biochemistry)thatformsareduc-tionistresearchprogram.Todoso,Iwouldliketoanswerthequestionofwhetherthisreductionistresearchprogramleadstonewempiricalknowledgeaboutthebiologicalsystemsorwhetheritdistortsourunderstandingofthosesystems.Toanswerthisquestion,IwouldliketouseImreLakatos’TheoryofScientificRe-searchPrograms.ThereasonIchooseLakatos’Theoryis:1.Itspecificallyad-dressestheissueofwhetherresearchtraditionsareprogressivei.e.,whethertheyleadtonewempiricalknowledge.2.Othertheoriesofsciencesuchasfalsifica-tionismorinductivism(withtheexceptionofThomasKuhn–IcouldaswelluseKuhn’sTheoryandmakesimilarpointsaboutthissciencesinceKuhn’snotion1AlanLoveandIngoBrigandt,“ReductionisminBiology”,EntryinStanfordEncyclo-pediaofPhilosophy.2Ibid.F.Stadler(ed.),ThePresentSituationinthePhilosophyofScience,ThePhilosophyofScienceinaEuropeanPerspective,DOI10.1007/978-90-481-9115-4_16,©SpringerScience+BusinessMediaB.V.2010\n196MehmetElginofprogressintermsofincreaseineffectivenessofpuzzlesolvingwoulddothesamejob)arenotfittedtoevaluateresearchtraditionsasintegratedwholes.Iac-knowledgethatconclusionsIreachaboutBiochemistryResearchProgramonthebasisofLakatos’Theorywillbesensitivetotheappropriatenessofthattheoryinunderstandingscience.However,Iwillnotaddressthebroaderissuesaboutwhichtheoryofscienceisbetter.IIThemultiplerealizabilitythesishasbeenintroducedintophilosophybyPutnamandFodor.3Theideaisthis:thesamekindsofhigherlevelpropertiescanbereal-izedbydiversekindsofphysicalproperties.Ifthereisahigherlevelgeneraliza-tionoftheform“IfMthenB”whereMandBarehigherlevelkinds,therearemultiplephysicalpropertiesPtoPthatwouldrealizeMandtherearemultiple1nphysicalpropertiesP’toP’thatwouldrealizeB,wherePsnotonlyneednotbut1nalsotypicallyarenotequaltoP’s.Therelationofthisthesistoreductionismisasfollows:ifahigherleveltypecanberealizedbydiversephysicaltypesatalowerlevel,thetype-typereductionisnotpossiblesincethereisnouniquelowerleveltypetowhichahigherleveltypecanberelatedto.Inphilosophyofbiologyandphilosophyofmind,thecorrectnessofthethesishasbeentakenforgrantedbe-causeithasbeenthoughtthatexamplesareeverywhere(althoughitsimplicationshavebeendebated,seeSober4):justlikeonecanmakeanautomobilefromverydifferentphysicalmaterialsandyetrealizethefunctionofbeingacar,inthesamewaymindsandlivingorganismscanbebuiltupofdiversephysicalpropertiesandyetdonotlackanythingintermsoftheirfunctionsathigherlevels.TheclaimthatmultiplerealizabilityiscoherentwascriticizedbyLarrySha-piro.5Tosaythatakindismultiplyrealizableistosaythattherearedifferentwaystobringaboutthefunctionthatdefinesthekind.But,iftwoparticularsdifferonlyinpropertiesthatdonotinanywayaffecttheachievementofthedefiningcapacityofakindthenthereisnoreasontosaythattheyaretokensofdifferentrealizationsofthekind.Differentlycoloredcork-3HilaryPutnam,“PsychologicalPredicates”,in:W.CapitanandD.Merril(Eds.),Art,MindandReligion.Pittsburgh:UniversityofPittsburghPress1967,pp.37-48.HilaryPutnam,“PhilosophyandOurMentalLife,”in:Mind,LanguageandReality.Cam-bridge:CambridgeUniversityPress1975,291-303.JerryFodor,PsychologicalEx-planations.Cambridge,MA:MITPress1968.JerryFodor,TheLanguageofThought.NewYork:ThomasCrowell1975.4ElliottSober,“TheMultipleRealizabilityArgumentagainstReductionism”,in:Phi-losophyofScience66,1999,pp.542-564.5LarryA.Shapiro,“MultipleRealizations”,in:TheJournalofPhilosophy97,12,2000,pp.635-654.\nReductionisminBiology197screws,alikeineveryotheraspect,arenottokensofdifferentrealizationsofacorkscrewbecausedifferencesincolormakenodifferencetotheirperformanceasacorkscrew.6Wecanextractthefollowingcriterionfromthispassage:Iftherearetwodifferentkindsofrealizersofthesamehigherlevelkind,thentheymustdifferintheircausalpowersthatarerelevanttothefunctionofamultiplyrealizedstate.Contrapositiveofthisconditionalstatementrunsasfollows:Ifdifferentrealizersofthesamehigherlevelkinddon’tdifferintheircausalpowersthatarerelevanttothefunctionofamultiplyrealizedstate,thenthoserealizersarenotdifferentkindsofrealizersofthishigherlevelkind.Thiscriterionimpliesthatnomatterhowdiverserealizersofagivenstatemayseem,aslongastheyshareacommoncausalpowerrelatedtothefunctionatahigherlevel,withrespecttothathigherleveltype,realizersdon’tfallunderdistinctkindsatthelowerlevel.Theymaydifferinotherphysicalorchemicalcharacteristics.Itishighlydubious,accordingtoShapiro,thatrealizersatalowerlevelwoulddifferintheirrelevantcausalcapacitiesinbringingaboutthehigherlevelpropertiesandwewouldstillcallthemtherealizersofthesamehigherleveltype.Realizersofahigherleveltypemayfallunderdistinctkindswithrespecttomanyproperties.Forexample,wecanbuildpendulumsfrommanydifferentkindsofphysicalmaterialbutifalltheserealizersobeypendulumlaw,theydonotcon-stitutedifferentkindsofrealizerswithrespecttothefunctionofpendulums.7Inthesameway,physicalorchemicalpropertiesthatrealizebiologicalpropertiesofalivingcellmayormaynotfallunderdistinctkindsdependingonwhichcharacter-isticsweareinterestedinthem.Ifweareinterestedintheirrelationtobiologicalpropertyinquestion(theirabilitytobringaboutbiologicalproperties),theymaynotfallunderdifferentkinds.Ifontheotherhand,weareinterestedinclassifyingtheirotherphysicalpropertiesthatarenotrelevanttotheirabilitytobringaboutthosebiologicalproperties,theymayfallunderdistinctkinds.Ifthisistrue,itisthenatleastpossibilitythatrealizersofagivenbiologicalfunctionmayhaveatleastsomechemicalorphysicalcharacteristicsthatarerelevanttotheirabilitytobringaboutthatfunction.Then,thetype-typereductionisatleastapossibility.Thereisnoimplicationinthisthesisthatallbiologicalknowledgecanbereducedtochemicalandphysicalknowledge.Itmaybethatsomepartsofbiologyresistthis.However,itentailsthattheclaimthatreductionismisinprinciplenotpos-sibleduetomultiplerealizabilityisfalse.Itfollowsthatthefactthatthestateis“multiplyrealized”doesnotentailthattherecannotbekindgeneralizationsaboutlowerlevelphysicalorchemicalpropertiesrelevanttotherealizationofahigher6Ibid.,p.644.7RobertBatterman,“MultipleRealizabilityandUniversality”,in:BritishJournalforthePhilosophyofScience51,2000,pp.115-145.\n198MehmetElginleveltypebecause(again)distinctrealizersofasystemmayexhibitsomefeaturesthatareuniversalwithrespecttothebehaviorunderconsideration.Considertherelationbetweenproteinfunctionanditsstructureandsequence.Tomakesenseofthisrelation,weshouldspecifyexactlywhatwecallahigherleveltype(inthiscaseaboutthespecificfunctionofaprotein).Then,wemustbeclearabouttherelevantlowerlevelpropertythatwemaysayresponsibleforthatspecificfunction.Forexample,thesameproteinmayberesponsibleforseveralfunctionsoraproteinhavingdifferentsequencesmayrealizethesamefunction.Doesitfollowfromthisthatmultiplerealizabilitythesisisrightandconsequentlyreductionismisfalse?No.Ifthetwosequencesaredifferentandyetstillreal-izethesamefunction,thentherelevantquestioniswhatpartofthesequenceisresponsibleforthefunction.Whatisakind?Shouldwetaketwosequencesasdifferentlowerlevelkindsbecausetheydifferonlyinoneplaceormany?Some-timesonlyonechangeinthesequencemaybeenoughtocallthemdifferentkindsbutsometimeseveniftheydifferinmanyplacesitmaynotbe.Thisdependsonwhatfunctionweareinvestigating.Sowhethersomethingcountsasarelevantkindatthelowerleveldependsonthefunctionweareinterestedin.Ifthereisapartinthesequencethatmakesadifferenceintherealizationofthefunctionthatistherelevantpropertyweshouldbefocusingonasalowerlevelkind.Ifwecanidentifysuchapartandifthatpartiscommoninalloftherealizersofthefunction(inallthedifferentsequencesthatrealizethefunction)thenwehaveidentifiedakindatthelowerlevelwithrespecttothefunctioninquestion.Ifahigherlevelgeneralizationis“allproteinshavesomebiologicalfunctions”andwewanttore-ducethistoalowerlevelkind,thenweshoulddirectourattentiontotheirchemi-calorphysicalattributesthatenableallofthemtodosomebiologicalfunctions.Ifahigherlevelgeneralizationis“ProteinXhasabiologicalfunctionY”andwewanttoreducethistoalowerlevelkind,thenweshouldlookatthechemicalorphysicalfeaturesofthisproteinthatwecanassignresponsibilityforthatspecificfunction.Soahigherleveltypecanbegeneralizationsaboutasingleproteinorallproteins;butdependingonaspecifickindofhigherlevelgeneralization,thelowerlevelphysicalorchemicalgeneralizationmaybedifferent.Inbothcloselyanddistantlyrelatedproteinsthegeneralresponsetomutationisconforma-tionalchange.Variationsinconformationinfamiliesofhomologousproteinsthatretainacommonfunctionrevealhowthestructuresaccommodatechangesinaminoacidsequence.Residuesactiveinfunction,suchastheproximalhistidineoftheglobinsorthecatalyticserine,histidineandaspartateoftheserineproteinases,areresistanttomutationbecausechangingthemwouldinterfere,explicitlyanddirectly,withfunction8(Italicsaremine).Whetherthissituationiswidespreadorrearisirrelevant.However,thisexampleillustrateshowdespitemanydifferencesinthesequencewecanstillcallallof8ArthurM.Lesk,IntroductiontoProteinArchitecture.Oxford:OxfordUniversityPress2001,p.172.\nReductionisminBiology199themthesamekindofsequencewithrespecttoacommonpropertythatisrelevanttothespecificfunction.TheaboveclaimbyLesksaysthatthereisapartinthesequencethatispreserveddespitethefactthatotherpartsshowvariations.Theexplanationisthat,thepartthatispreservedplaysavitalroleintherealizationofthefunction.Sowithrespecttothisspecificfunction,thereisacommonlowerlevelkindtowhichthatfunctioncanbereduced.ThisiswhatShapiro’scriterionofmultiplerealizationspredicts.IIIThefailureofaprioriargumentsagainstreductionismpavesthewayforadefenseofmethodologicalreductionism.Thebasicideahereisthatifaresearchprogramwithreductionisttenetssuchasbiochemistryleadstonewempiricalknowledgeaboutbiologicalphenomena,theissueofwhetherreductionismisjustifiedcanbeaddressedonthesemethodologicalgrounds.Heretheissuedoesnotconcernwiththetruthofreductionistthesis;itmainlyconcernswiththeheuristicvalueofit–i.e.,whethersciencessuchasbiochemistry,biophysicsetc.arejustifiedinfol-lowingreductionisttenets.Itisimportant,however,tonotethatifareductionistresearchprogramconstantlysucceedsinthediscoveryofnewempiricalknowl-edge,thiswillprovideplausibilityforthereductionistthesisevenifitdoesnotjustifyitstruth.AccordingtoImreLakatos,9thebasicunitofscienceisaresearchprogram.ScientificResearchPrograms(SRP)consistofnegativeandpositiveheuristics:negativeheuristicsdeterminewhatisnotallowedinSRPandpositiveheuristicsdeterminewhatispermitted.SRPalsoconsistsoftwosetsofassumptions:hardcoreandprotectivebelt.HardcoreassumptionsarethefundamentalprinciplesofSRP(forexample,inNewtonianphysicstheywouldbethreelawsofmotionplusgravitationallaworinevolutionarybiologytheywouldbeformulationsofprinci-plesthatdefinehowevolutionaryforcesaffectgeneticstructureofapopulation)andprotectivebeltassumptionsareanythingthatmaybeneededtorelatethesehardcoreassumptionstotheworld.Whenthereisamismatchbetweentheoreti-calresultsandtheactualobservations,negativeheuristicssaythatnochangeinhardcoreassumptionsisallowed.Positiveheuristicssaythatonlynon-adhocchangesareallowed.Inthecaseofagapbetweentheoreticalresultsandactualmeasurements,thereisnorecipeaboutwhatkindofchangescanclosethegap.However,sometimesmethodologicalprinciplesmayleadustomakechangesincertaindirections.Forexample,inNewtonianphysics,commitmenttotheidea9ImreLakatos,“FalsificationandtheMethodologyofScientificResearchProgrammes”,in:ImreLakatosandAlanMusgrave(Eds.),CriticismandtheGrowthofKnowledge.NewYork:CambridgeUniversityPress1965.ImreLakatos,TheMethodologyofSci-entificResearchProgrammes.PhilosophicalPapersVolume1.NewYork:CambridgeUniversityPress1978.\n200MehmetElginthatnatureshouldbeexplainedinmechanicaltermswillleadscientiststolookfornewmodelsthatwillnotviolatethismaxim(sowhenonemechanicalmodelfailstheywilllookforanothermechanicalmodelthatwilldothejob).Inthissense,whenweassessanSRP,wearealsoassessingthesemethodologicalmaximsaswellbecausesuchprinciplesmaysometimesberesponsibleforthefailureofanSRP.AccordingtoLakatos,failureorsuccessofSRPcannotbeputinabsoluteterms.Forhim,successfulSRPmakesbothempiricalandtheoreticalprogress.FailedSRPisdegenerativeinthesensethatitfailstomakeempiricalprogress.SincethereisnorecipehowtomakeappropriatechangesinSRPwhenthereisagapbetweentheoreticalresultsandactualmeasurements,inmostpartcreativeandimaginativescientistsdeterminethefaithofSRP.Inthissense,sometimeseventhedegenerativeSRPmaymakeacomeback.Inordertodefinetheoreticalandempiricalprogress,followingMalcolmForster,10letmeintroducetheterm‘model’.Amodelisbasicallythecombinationofhardcoreassumptionsandprotectivebeltassumptionsfromwhichweobtaintheoreticalresultsthatcanberelatedtotheworld.Whenatheoreticalresultweobtainfromamodeldoesnotmatchtotheactualobservations,thenanothermodeliscalledfor.AccordingtoLakatos,whenwemakechangesintheoldmodel,thesechangesshouldnotbeadhoc;i.e.,suchchangesshouldleadtonewpredictionsandtheyshouldbeindependentlytestable.Ifthesenewpredictionsareempiricallyconfirmed,thenSRPismakingempiricalprogress.Ifweconstructmoreandmoremodelsandtheymakeempiricalprogress,thenSRPthatthesemodelsbelongtoisprogressive.However,ifmoreandmoremodelsbelongingtoSRPstartfailing,thenitisdegenerative.Doesbiochemistrycontainreductionisttenets?IsitaprogressiveresearchprograminLakatos’sense?Biochemistrystudieschemicalprocessesandreactionsthattakeplaceinlivingcells.Therearevarietiesofdifferentmoleculesinlivingcells.MoleculesthatBiochemistrystudiesarecarbohydrates,proteins,enzymes,lipidsandnucleicacids.Theguidingideaofbiochemistryisthatprocessesofliv-ingcellscanbeunderstoodintermsofthechemicalpropertiesofthesemoleculesthatformalivingcell.Onetextbookdefinesthegoalofbiochemistryasfollows:Theoverallgoalofbiochemistryistodescribelife’sprocessesusingthelanguageofmol-ecules,thatis,applyingtheprinciplesandmethodsofchemistrytodeterminemolecularstructurefromwhichitisoftenpossibletoexplainbiologicalfunction.1110MalcolmR.Forster,„TheHardProblemsinthePhilosophyofScience“,in:R.NolaandH.Sankey(Eds.),AfterPopper,Kuhn&Feyerabend:RecentIssuesinTheoriesofScientificMethod,AustralasianStudiesinHistoryandPhilosophyofScience,KluwerAcademicPublishers2000,pp.231-251.11RodneyF.Boyer,ConceptsinBiochemistry.Hoboken,NJ:Wiley&SonsInc.2006,p.2.\nReductionisminBiology201Itisimportanttodistinguishbetweentwosetsoftheoriesthatwemaycallhardcoreassumptionsofbiochemistryresearchprogram:1.Therearebackgroundtheories,suchaschemical,physicalandbiologicaltheories.2.Therearesetofprinciplesaboutthenature,functionandinteractionsofbiomoleculesthatarebuildingblocksoflife.Itisthesecondonethatisdistinctiveaboutbiochemistryandthereductionistnatureofthisresearchprogramliesinthesesecondkindsofprinciples.Sincethesecondclaimisdistinctiveofbiochemistryresearchprogram,itplaysavitalrolewhetherthissciencesucceedsinrealizingitsgoals.Lehninger,NelsonandCox12write:Themoleculesofwhichlivingorganismsarecomposedconformtoallthefamiliarlawsofchemistry,buttheyalsointeractwitheachotherinaccordancewithanothersetofprin-ciples,whichweshallrefertocollectivelyasthemolecularlogicoflife.Theseprinciplesdonotinvolveneworyetundiscoveredphysicallawsorforces.Instead,theyaresetofrelationshipscharacterizingthenature,function,andtheinteractionsofbiomolecules.Thelistoftheprinciplesconcerningthemolecularlogicoflifeare:13Alivingcellisaself-contained,self-assembling,self-adjusting,self-perpetuatingisother-malsystemofmoleculesthatextractsfreeenergyandrawmaterialsfromitsenvironment.Thecellcarriesoutmanyconsecutivereactionspromotedbyspecificcatalysts,calleden-zymes,whichitproducesitself.Thecellmaintainsitselfinadynamicstadystate,farfromequilibriumwithitssurround-ings.Thereisgreateconomyofpartsandprocesses,achievedbyregulationofthecatalyticactivityofkeyenzymes.Self-replicationthroughmanygenerationsisensuredbytheself-repairing,linearinforma-tion-codingsystem.GeneticinformationencodedassequencesofnucleotidesubunitsinDNAandRNAspecifiesthesequenceofaminoacidsineachdistinctproteinwhichulti-maltelydeterminesthethree-dimensionalstructureandfunctionofeachprotein.Manyweak(noncovalent)interactions,actingcooperatively,stabilizethethree-dimension-alstructuresofbiomoleculesandsupramolecularcomplexes.Thecommonthemeinalltheseprinciplesistheideathatlifecanbeunderstoodintermsofchemicalorphysicalpropertiesofbiomoleculesandtheirinteractions.ItisbecauseofthiscommonthemethatIclaimbiochemistryisareductionistre-searchprogram.Biochemistryresearchprogramalsoincludesmethodsandtech-niquesabouthowtoidentifystructureandhowtorelatethisstructuretoaspecificfunction.Thisinvolvestheuseofinstruments,forexample,NMRspectroscopy,X-raycrystallography,Cryoelectronmicroscopyandelectroncystallography.Italsoinvolvescertainheuristicsaboutrelatingstructuretofunctionandinterpreta-tionofdataprovidedbytheseinstruments.Thus,biochemistryhasfeaturesofaresearchprograminLakatos’sense.12AlbertL.Lehninger,DavidL.NelsonandMichaelM.Cox,PrinciplesofBiochemis-try.NewYork:WorthPublisher1993,p.4.13Ibid.,p.19.\n202MehmetElginThequestionnowiswhetherbiochemistryasaresearchprogramisempirical-lyprogressive.Toanswerthisquestion,wehavetolookatthehistoricalrecordofbiochemistrywhetheritsmodelsconstructedfromhardcoreassumptionstogetherwithprotectivebeltassumptionshaveledtonewempiricalknowledgeaboutliv-ingcellsandwhethertherearecuesthatpointtoexpectationsaboutfurthernewempiricalknowledge.Toshowthatbiochemistryresearchprogramhasrealizeditsgoalstosomeextent,itsufficestolistjustsomemajordiscoveriesaboutthestructureofDNA,thestructureandfunctionofproteins,discoveriesaboutthecausesofmanydiseases,developmentsofnewtechniquesandinstrumentsinsolvingproblemsinbiochemistryresearchprogram.TherearemorediscoveriesinthefieldthanthenumberofNobelprizesawardedbuttheselectedlistofnobleprizeawardswillgivesomeideaaboutitsprogresstowardprovidingnewem-piricalknowledgeaboutbiologicalfunctions:Fisherforenzymeaction,Buchnerfordescriptionoffermentation,Summerforcrystallizationofurease,Krebsfordescriptionofcitricacidcycle,WatsonandCrickforDNAdoublehelix,PerutzforX-rayofproteincrystals,Smithforrestrictionenzymes,CechandAltmanforcatalyticRNA,Mullisforpolymerasechainreaction,Horvitzforbiochemistryofprogrammedcelldeath,Wüthrich,FennandTanakaforNMR,MSstructureofproteins,MackinnonandAgreforAquaporinsandmembranechannelsandHershko,RoseandCiechanoverforubiquitin-mediatedproteinbreakdown.14Furthermore,justlookingthroughpaperpublicationsrelatedtobiochemistrywillshowthatmoreknowledgeisbeingproducedandontheway.2005JCRScienceEditionreportsthatthereare261journalslistedunderthecategoryof“biochemistryandmolecularbiology”between2003and2005.Inthesejournals236,517paperspublishedbetween2000and2005andthetotalnumberofcitationsthesepapersproducedwas511,212.15Between1971and1990,thepercentageofbiochemistryarticlesinchemistrypaperspublishedinthejournalNatureisfoundtobe83butthepercentagedropsto73in1990s.16Aroundthesameyearstheper-centageofchemistryarticlesare13andthepercentageofbiologyandmedicinearticlesare49.17Moreinformationabouttheperformanceofscientificfieldsisavailableinjournalsrelatedtoscientometricandbibliometricstudiesofperform-anceevaluationofthesefields.ThefiguresIcitedabovepointtosomeroughandreadyideasabouttheprogressofbiochemistry.Thesefiguresmayincluderepeti-tivepublicationsandnotsignificantdiscoveries.However,evensmallpercentageofthesefigureswillshowthatthisresearchprogramisempiricallyprogressiveinLakatos’sensesinceitisnottooharmfultoassumethattopjournalsinthefielddonotpublishpapersthatarenotoriginalcontributiontothefield.Thereisalsointer-14Boyer,ConceptsinBiochemistry,p.6.15NanMa,JianchengGuan,andYiZhao,“BringingPageRanktotheCitationAnalysis”,in:InformationProcessingandManagement44,2007,p.802.16D.B.Arkhipov,“ScientometricAnalysisofNature,TheJournal”,in:Scientometrics46,1,1999,p.62.17Ibid.,p.59.\nReductionisminBiology203estingstatisticsaboutthenumberofpublicationsrelatedtothesubfieldsofbiol-ogy.Between1991and1998,intermsofaverageannualnumberofpapersamongthesubfieldsofbiology,with320papersmolecularbiologyranksfirst,with155papersmedicinecomesnext,with109papersbrainranksthird,with21papersnaturalhistoryranksfourth,andwith20papersagricultureranksfifth.18Eventhesefiguresgivesomeapproximateideaaboutthedirectionbiologicalsciencesheading.Thesefiguresshould,ofcourse,bedetailedandshouldbesubjectedtoseriousanalysistoprovidemoredetailedanswertothequestionofwhethersci-encesareheadinginthedirectionofreductionism;but,asastartinghypothesisfromthesefigures,wecansaythatbiochemistryisprogressiveresearchprogram.Itmustbenoted,however,thatLakatos’SRPdoesnotallowustomakejudg-mentsaboutthefinalfaithofaresearchprogram.For,SRPcanbeprogressiveatonetimeandthenmaybecomedegenerativelater;itmaybedegenerativeatonetimebutthenwiththeimaginativeandcreativeabilitiesofresearchersworkinginthefielditmaybecomeprogressiveagain.Soourevaluationsoftheoverallsuccessofaresearchprogramwillalwaysberelativetotheinformationavailabletousatagiventimeperiod.Accordingly,myclaimhereisthatthebestevidenceavailabletousnowleadstotheconclusionthatbiochemistryresearchprogramisempiricallyprogressive.IVThefundamentalquestionofbiochemistryisdefinedasfollowsinoneofthemostinfluentialtextbookinthefield:“Biochemistryaskshowthethousandsofdifferentbiomoleculesformedfromtheseelementsinteractwitheachothertoconferthere-markablepropertiesoflivingorganisms.”19Throughtheapplicationsoftechniquesandconceptsfromchemistry,biochemistshopetounderstandthisfundamentalquestion.Giventheirtrackrecordinashorttime,weshouldbeoptimisticabouttheirpossiblesuccessinansweringthisquestion.Soweshouldbeoptimisticaboutthisreductionistproject.Ibelievewewillmakemoreprogressinourphilosophi-calprojectsaboutreductionismbystudyingindetailasciencewhoseprojectistounderstandbiologicalphenomenaintermsofchemicalandphysicalconcepts.DepartmentofPhilosophyMuğlaÜniversitesiTürkiyemelgin@mu.edu.tr18Ibid.,p.67.19Lehninger,NelsonandCox,PrinciplesofBiochemistry,p.1.\n\nRAFFAELLACAMPANERREDUCTIONISTANDANTIREDUCTIONISTSTANCESINTHEHEALTHSCIENCESReductionismandantireductionismareamongthemostlargelyandhotlydebatedtopicsinphilosophyofbiologytoday.Inthissectionofthevolume,aimingtocon-veythecurrentsituationinthephilosophyofthenaturalandlifesciences,thesetopicsarespecificallyaddressedinMehmetElgin’spaper,focusingonbiochemis-try.Elginstronglysupportsreductionism,firstbyclaimingthatthenowclassicalargumentbasedonmultiplerealizabilitydoesnotentailanti-reductionismandsec-ondlyhighlightinghowtheversionofmethodologicalreductionismthatbiochem-istryhasbeenadopting–centeredon“theprinciplethatbiologicalfunctionsofbiomoleculesinlivingcellscanbeunderstoodintermsofchemicalandphysicalpropertiesofthosemolecules”1–hasprovedlargelysuccessful,teaching“usnewknowledgeaboutthebiologicalsystems”.Takentogether,thesetwoargumentsaredeemedtoprovidegoodgroundsforathoroughdefenceofreductionism.WhileElginchoosesbiochemistryashisprivilegedstandpointontheissue,Ishalldwellonthestancesemerginginthehealthsciences.Althoughthehealthsciencesarecloselyintertwinedwith–amongstothers–biologyandbiochemistry,theyalsohavetheirownpeculiarfeatures.Referringtosomeexamplestakenfromdiffer-entmedicaldisciplines,Iwillquestionwhetherreductionismcanberegardedasnotonlyaviable,option,butasthebestsolutiontogainnewknowledgeaboutbiomedicalsystems.Ishallsuggestthatconsiderationsarisingfromcurrentmedi-calresearchandpracticesupportapluralisticapproachtothetopic,inwhichbothreductionistandantireductioniststancescanbeaccommodatedindifferentways.1.REDUCTIONISMANDEXPLANATIONThecontroversyoverreductionismandantireductionismismultifaceted,andthedevelopmentofanumberofversionsofeachhornofthedebateinthephilosophi-callandscapemakesitparticularlyintricate.Ingeneralterms,Ishalltakeherethereductionistprogramofmolecularbiologytoclaimthatallbiologicalphenomenamustinprinciplebefullyreducibletophysicochemicalentitiesandtheirorganiza-tion,andtothelawsidentifiedbyphysicsandchemistry.Henceitisheldthattounderstandsomethingonehastolookatthesetofphysicalandchemicalentities1MehmetElgin,“ReductionisminBiology:AnExampleofBiochemistry”,thisvol-ume,p.x.F.Stadler(ed.),ThePresentSituationinthePhilosophyofScience,ThePhilosophyofScienceinaEuropeanPerspective,DOI10.1007/978-90-481-9115-4_17,©SpringerScience+BusinessMediaB.V.2010\n206RaffaellaCampanerofwhichitiscomposed,attheirpropertiesandarrangement:theleveldescribedbyphysicsandchemistryisregardedasthelowestleveltowhicheverythingthatexistscanbereducedbyvirtueofthefactthateverythingthatexistsisultimatelycomposedofphysicochemicalentities.Thereductionistmindsethaslargelypervadedmolecularbiologyintheconvictionthat“becausebiologicalsystemsarecomposedsolelyofatomsandmolecules[…],itshouldbepossibletoexplainthemusingthephysicochemicalpropertiesoftheirindividualcomponents,downtotheatomiclevel”2.Oneoftheversionsinwhichreductionismhasbeendevelopedisthusexplanatoryreduction-ism,withreductionismbeingconsideredthemostappropriatekeytoexplanation.“Ifitisafactthattheentitiesandpropertiesofarealmcanbeidentifiedwithcertainarrangementsofbasictypesofcomponents,thenourunderstandingofthisrealmshouldimprove,tosomedegree,byourbecomingawareoftheexistenceandnatureofthesecomponents.”3Thistargetwillbereachedbyappealingtophysicsandchemistry,whichhavebeenprovedtoexplainavastrangeofthingsbyasmallsetoflaws.Oneofthemostforcefulsupportersofexplanatoryre-ductionismwithrespecttobiology–morespecifically,toexperimentalbiology–hasrecentlybeenMarcelWeber.Maintainingthatmodernexperimentalbiologyisthoroughlyreductionist,heemphasizesthat,whilephysicsandchemistryareinterestedinidentifyinganddescribingnaturallaws,experimentalbiologyjustaimsatapplyingthemtobiologicalphenomenaforexplanatorypurposes.If,ontheonehand,biologycanberecognizedasanautonomousscienceasfarasitsspecificconcepts,methodologicalstandardsandresearchmethodsareconcerned,Weberstressesthat,ontheotherhand,itdependsonphysicsandchemistryasfarasexplanationisconcerned.Biologicaltermssuchas“axon”or“synapse”playadescriptiverole,whereasalltermsfiguringingenuineexplanationsarephysicalandchemicalterms,referring,forinstance,tomolecules,macromolecularaggre-gatesandpurelyphysicalentities.Biologicaltermsareadoptedsimplytoidentifythekindofsystemstowhichphysicalandchemicaltermsaretobeapplied.Doesthisholdforthehealthsciencesaswell?Doessuchaperspectiveexhaustthestancesemerginginthehealthsciences?Thepatient’sbodycanbeseenascomposedofdifferentanatomicalsystems,suchastherespiratoryorcardiovascularsys-tems.Thesesystemsare,inturn,composedofvariousorgans,suchaslungsandhearts,whicharemadeupofepithelial,muscular,nervousandglandulartissues.Finally,tocom-pletethereduction,thesetissuesarecomposedofdiversecellulartypesthataremadeupofavarietyofmolecules.42MarcH.V.VanRegenmortel,“ReductionismandComplexityinMolecularBiology”,in:EuropeanMolecularBiologyOrganizationReports5,2004,pp.1016-1020,quot.exp.1016.3ToddJones,“ReductionismandAntireductionism:RightsandWrongs”,in:Metaphi-losophy35,5,2004,pp.614-647,quot.exp.616.4JamesMarcum,“BiomechanicalandPhenomenologicalModelsoftheBody,theMeaningofIllnessandQualityofCare”,in:Medicine,HealthCareandPhilosophy7,\nReductionistandAntireductionistStancesintheHealthSciences207Itiswithrespecttotheiraffectingsuchsystemsthatdiseasesareinvestigated,andtheirfundamentalphysicochemicalfeaturesidentified5.Accordingtoreduction-ists,biologicalsystems–inthecasesofthehealthsciences,bodiesaffectedbydiseases–arecomposeduniquelyofatomsandmolecules,andcanthereforebefullydescribedandexplainedintermsofthephysicochemicalpropertiesoftheirconstituentparts.Thesepartsareusuallycombinedinarticulatedsystems,whichhavetobebrokendownintosimplerpiecestodeterminethepropertiesandcon-nectionsbetweentheparts.Afterdescribingacomplexsystemintermsofitsconstituents,abiologistmaybeledtobe-lievethathehas“reduced”somethingcomplextoitssimplercomponents[…]Whencellsandorganellesaredescribedintermsoftheirmolecularconstituents,itmay,indeed,seemplausiblethatbiologicalentitiesarenothingbutphysicochemicalsystemsandthatbiologyshouldbereducibletochemistryandphysics.6Muchoftherecentphilosophicalliteraturehasemphasizedthecomplexityofbiologicalsystemsandthemulti-levelcharacteroftheiranalyses.Thepossiblerelationsbetweenreductionisttenetsandamechanisticperspectivearebeingcon-sideredindifferentways.Ontheonehand,mechanisticinvestigationshavebeenopposedtoreductionism.CarlCraver,aneo-mechanist,forinstance,haspointedoutthatadequatemechanisticexplanationsofbiologicalsystemsmusthaveaninterfieldcharacter.Hestresseshowtypicallyreductionistsreducetheoriesaboutphenomenaatahigherlevel(e.g.,gases,lightning,andlife)[…]totheoriesaboutphenomenaatlowerlevels(e.g.,molecules,electrons,andphysiologicalsystems).[…]Themechanisticapproach[he]develop[s][instead]suggestsareasonablewaytoun-derstandthelevelrelationshipandinsistsonrecognizinginterfieldrelationsthatoscillateupwardanddownwardinahierarchyoflevels.73,2004,pp.311-320,quot.exp.313.5Forsomeexamples,seeIainH.McKillop,DiarmuidM.Moran,etal.,“MolecularPathogenesisofHepatocellularCarcinoma”,in:JournalofSurgicalResearch136,1,2006,pp.125-135;MaryannE.Smela,SophieS.Currieretal.,“TheChemistryandBiologyofAflatoxinB:fromMutationalSpectrometrytoCarcinogenesis”,in:Car-1cinogenesis22,4,2001,pp.535-545.6MarcH.V.VanRegenmortel,“PitfallsofReductionisminImmunology”,in:MarcH.V.VanRegenmortelandDavidHull(Eds),PromisesandLimitsofReductionismintheHealthSciences,Chichester:Wiley&Sons2002,pp.47-66,quot.exp.47.7CarlCraver,“BeyondReduction:Mechanisms,MultifieldIntegrationandtheUnityofNeuroscience”,in:StudiesinHistoryandPhilosophyofBiologicalandBiomedicalSciences36,2,2005,pp.373–395,quot.exp.376.Craverdwellsonthecaseoflong-termpotentiationandmemoryconsolidation,concludingthatthegoalofreduction,especiallypursuedinthe1950sand1960s,hasbeenreplacedbythatofidentifyingmultilevelmechanismswithanexplanatorypurpose.\n208RaffaellaCampanerOntheotherhand,WilliamWimsattclaimsthatmechanicismandreductionismcanactuallygohandinhand,stressingthatmuchoftheexplanatoryweightinexplanatoryreductionsisbornebytheorganizationofsmallestpartsintomecha-nismsaccountingforthebehaviourofthesystematstake.8Clarifyingsuchanorganizationisactuallyregardedasthemajortaskreductionsaresupposedtodo:areductionistanalysisoffersalowerlevelmechanisticaccountofahigher-levelphenom-enon,entity,orregularity.Todoso,onecommonlydecomposesacomplexsystemintoitsparts,analyzestheminisolation,andthenre-synthesizesthesepartsandtheexplanationsoftheirbehaviourintoacompositeexplanationofsomeaspectofthebehaviourofthesystem.9Anauthorthat,asweshallseebelow,takespartinthedebatewithaspecificfocusonmedicine,MarcH.V.VanRegenmortel,criticisesreductionistthinkingforanalysing“complexnetworkinteractionsintermsofsimplecausalchainsandmechanisticmodels”andforfavouring“causalexplanationsthatgiveundueex-planatoryweighttoasinglefactor”.10Thisistakentooverlookthefactthatclini-calstatesarealwaystheoutcomesofmanybiochemicalpathwaysandnetworks,andthatdiseasesaretheresultsofalterationstocomplexsystems.Contemporarymechanisticviews,whichdevoteaparticularattentiontobiologicalandbiomedi-calsystems,stronglyemphasizethecomplexityofthephenomenainvestigated,suchasthefunctioningofthehumanbodyandthedevelopmentofdiseases.Farfromsearchingfor“simplecausalchains”orfor“asingle[explanatory]factor”,theyinsistonoutcomesbeingbroughtaboutbyapluralityofcauses,combinedasdifferentcausalnets.FollowingWimsatt’ssuggestion,areductionistcanthereforebeseenasunderstanding“thecharacter,properties,andbehaviourofthestudiedsystemintermsofthepropertiesofitspartsandtheirinterrelationsandinterac-tions”,11wheresuchpartscanbeconceivedofascomponentsinamechanism.The8ForapositionthatassociatesreductionwiththesearchformechanismsseealsoSa-hotraSarkar,“ModelsofReductionandCategoriesofReductionism”,in:Synthese91,5,1992,pp.167-194.9WilliamC.Wimsatt,“ReductionismandItsHeuristics:MakingMethodologicalRe-ductionismHonest”,in:Synthese151,3,2006,pp.445-475,quot.exp.466.10VanRegenmortel,“ReductionismandComplexityinMolecularBiology”,loc.cit.,p.1018,italicsadded.SeealsoMarcH.V.VanRegenmortel,“PitfallsofReductionismintheDesignofPeptide-BasedVaccines”,in:Vaccine19,17-19,2001,pp.2369-2374:“Causalexplanationsarereductiveinthesensethatonefactorissingledoutforatten-tionandgivenexcessiveexplanatoryweight”(p.2370).11Wimsatt,“ReductionismandItsHeuristics:MakingMethodologicalReductionismHonest”,loc.cit.,p.467.Onthis,seealsoWilliamC.Wimsatt,“Aggregate,ComposedandEvolvedSystems:ReductionistHeuristicsasMeanstoMoreHolisticTheories”,in:BiologyandPhilosophy21,5,2006,pp.667-702:“Suchmechanisticexplanationsarealsoreductionistexplanationsofthebehaviourorapropertyofasystemintermsoftheinteractionsofitspartsandproperties.Suchareductionneednotdenythecausalimportanceofhigher-levelphenomena,regularities,entities,structures,andmecha-\nReductionistandAntireductionistStancesintheHealthSciences209mechanisticaccountemphasizeshowsuchrelevantcomponentsworktogether,howtheyinteractandareorganizedinproductivecontinuityfromsomebegin-ningtosomefinalconditions.Atthesametime,neo-mechanicismcountsamongitsadvantagesthecapacitytoprovideinsightsinto“interlevelformsofinterfieldintegration”,largely“neglectedinmuchoftheliteratureonreduction”.122.ANTIREDUCTIONISTSTANCESANDTHEHEALTHSCIENCESDoreductionisttenetssufficeforafullunderstandingofthebehaviourofthesys-temsinvestigatedbythehealthsciences,orissomethingdifferentorsomethingmorerequired?Somelimitationsarisefromourinabilitytoknowtheexactdetailsthephysicochemicalentitiesinvolvedinthedevelopmentofadisease,or–ifwearedealingwithhealth–inthestandardfunctioningofthehumanbody.Theacknowledgmentthatwecannotidentifyandkeeptrackofthebehavioursofthefundamentalmoleculesthatcollectivelyactandinteractintoanorganismdoesnotexcludeexplanatoryreductionism:apurelyphysicochemicalexplanationcaneventuallybereached,andreductiveexplicabilitycanstillbemaintainedinprin-ciple.Isareductionistexplanation,though,theonlygoaltobereached,inpracticeorjustinprinciple,inthebiomedicalsciences?Criticismsofareductionistapproachhavebeenformulatedinvariousfieldsofthehealthsciencesondifferentmatters.Someofthemaretargetedagainstthegeneralconceptionofthehumanbodyanddoctor-patientrelationshipthatistakentousuallyunderliethereductionistperspective.JamesMarcumcriticisesthe“biomechanicalmodel”ofthepatient’sbody,consideredthepredominantmodelinmodernWesternmedicine.Advocatesofthismodel,whichbasicallyconstruesthehumanbodyasamachine,“reducethepatienttoseparate,individualbodypartsinordertodiagnoseandtreatdisease”,andfailtoconsideritasanembodiedperson:oncethoughtofasgovernedbyphysicalandchemicallaws,thebodyisreducedtoanabstractthingandtotally“strippedofitslivedcontext”.Thebodyisconceivedasamechanizedmaterialobject,reducibletoacollectionofphysicalparts,that,whenbrokenormalfunctioning,canberepairedorsubstituted.Accord-ingtothismodel,illnessisconstruedintermsof“diseasedordysfunctionalbodynismsbuiltuponthem”(p.669);“areductiveexplanationofabehaviouroraprop-ertyofasystemisonethatshowsittobemechanisticallyexplicableintermsofthepropertiesofandinter-actionsamongthepartsofthesystem”(pp.670-671).Wimsattsuggeststhata“multilevelreductionisticanalysis”canpickout“theappropriatelevelsforobjects,processes,andphenomena,andarticulatesandexplicatestheirrelationstocompletetheexplanatorytaskwithnofurthermystery.Itistobedistinguishedfromapocalypticreductionismthroughtherecognitionofrelevantdynamicsatmultiplelev-els(ibid.,p.672).12Craver,“BeyondReduction:Mechanisms,MultifieldIntegrationandtheUnityofNeuroscience”,loc.cit.,p.388.\n210RaffaellaCampanerpartsseparatefromtheoverallintegrityofthepatient’sbodyandlivedcontext”,13andthephysician’sroleiscomparabletothatofatechnicianoramechanic.Asaconsequence,Marcumstresses,thequalityofmedicalcarehasbeenundergoingacrisis.Theneedforaproperconsiderationofthepatientassuchinnon-reductionistterms,withthe“restitutionoftheintactpersontohisorherfullpersonhood”,14isalsoadvocatedbyAlfredTauber,whoemphasizesthesocio-psychologicalaspectsofthehealingprocess,andbyElizabethLloyd,whodiscussescross-populationalcorrelationsbetweensocioeconomicstatusandmorbidityandmortality.Lloydre-viewsanumberofsurveysaimedatdemonstratingthatamajorincomeinequalityinagivensocietyisstronglyrelatedtothatsociety’slevelofmortality:themoreseverlythepoorarepoorerthantherichinacertainsociety,thelowertheaveragelifeexpectancyeverybody–notjustthepoor–has:“socioeconomicfactorsturnouttobepowerfulpredictorsofhealthoutcomes[…],andthesefactorscannotbeinvestigatedifallresearchfundsareconcentratedatproblemsconceivedatthemolecularlevel.”15Shetakesthistoshowthatnotallmedicalresearchcanbeexhaustedbyananalysisatthemolecularlevel.Furtherantireductioniststancescanbefoundincontemporarymedicalre-search,withquiteanimpactonthemethodologyadoptedanditspracticalimplica-tions.Somesuchstanceshavebeenexpressedwithrespecttothediscoveryofnewdrugsandthedevelopmentofvaccines.Ithasbeenclaimedthatthereductionistmindset,whichislargelyadoptedinthesecontexts,actuallyturnsouttohaveseverelimitations.Morespecifically,asfarasdrug-discoveryprogrammesareconcerned,VanRegenmorteltracesthedeclineintheapprovalofnewdrugsperyearinthelastdecadebacktounmitigatedreductionism.Heclaimsthatdissect-ingthehumanbodyintoitscomponentsunderratestheimportanceofregardingorganismsandpatientsaswholes,andlosesawealthofusefulinformationontheirworkings.Withregardtovaccineresearch,theideathatabiologicalphenomenonsuchasprotectionagainstinfectioncanbereducedtothelevelofchemistry,al-thoughfashionable,iscriticisedandaccusedofseriousshortcomings.Vaccinationandprotectiveimmunityhaveameaningonlyatthelevelofthewholebiologicalorganism,giventhat“molecules,tissuesandorganscannotbevaccinated”,andareanchoredinthebiologicalrealm:13Marcum,“BiomechanicalandPhenomenologicalModelsoftheBody,theMeaningofIllnessandQualityofCare”,loc.cit.,pp.311-313.14AlfredI.Tauber,“TheEthicalImperativeofHolisminMedicine”,in:MarcH.V.VanRegenmortelandD.Hull(Eds.),PromisesandLimitsofReductionismintheHealthSciences,Chichester:Wiley&Sons2002,pp.261-27,quot.exp.268.Taubergoesasfarastoarguethateventhemoralandspiritualaspectsofhumanbeingsaretobetakenintoaccount.15ElisabethA.Lloyd,“ReductionisminMedicine:SocialAspectsofHealth”,in:MarcH.V.VanRegenmortelandD.Hull(Eds.),PromisesandLimitsofReductionismintheHealthSciences,Chichester:Wiley&Sons2002,pp.67-82,quot.exp.66.\nReductionistandAntireductionistStancesintheHealthSciences211immunogenicitydependsonthebiologicalpotentialofthehostthatisbeingimmunised;[…]antibodiesactinacollectivemannerand[…]theneutralizingsynergybetweenvariousantibodiescannotbereducedtothesimpleadditiveeffectsofindividualmolecules.16Henceitisstressedthatvaccinationandimmuneresponsecanbeelucidatedonlybystudyingthedynamicsoftheentiresystemofpathogen,antibodyandhostcell.Reductionistanalysesintermsofthephysicochemicalprinciplesunderlyingimmunologicalrecognitionareclaimedtobeinsufficienttodesignnewvaccines,andattemptstoreducebothvaccinationandautoimmunephenomenatodiscretemolecularfeaturesofindividualcomponentsoftheimmunesystemareboundtofail.Furthermore,ifallofthedetailsoftheincrediblycomplexworkingsoftheimmunesystemarespelledoutbutnomentionismadetotheeffectthatallthesemechanismshaveonthehealthoftheorganism,somethingdesperatelyimportanthasbeenleftout.17Anotherexampleofantireductionistthinkingisprovidedbysomereflectionsonthehistoryofcancerresearch.Ithasbeennoticedthattheestablishedreduc-tionistapproachknownasthe“somaticmutationtheory”(SMT)expoundedbyRobertWeinbergandcolleagues,aimingtodisentanglethegeneticandmolecularcircuitryofcarcinogenesis,hasbeenchallengedinthepastfewyearsbyanorg-anicistapproach,the“tissueorganizationfieldtheory”(TOFT)developedbyAnaSoto,CarlosSonnenscheinandcolleagues,focusingontissuesandthecomplexarraysofrelationshipsrangingoverseverallevelsofentities.Fromareductionistperspective,cancerisconsideredastheproductofmutatedgenes.Fromanorganicistperspective,accordingtoSonnenscheinandSoto,thearchitectureofthetissuecannotbereducedsimplytotheunderlyingindividualcellsandbiomacromoleculesthatcomposethem.Rather,tissuearchitectureisanemergentpropertyof‘thesocietyofcells’,i.e.,itisnotafunctionsimplyofthecollectivepropertiesofthecellsthatmakeupthetissue.1816VanRegenmortel,“ReductionismandComplexityinMoelcularBiology”,loc.cit.,p.1019.Inthefieldofvaccinology,thereductionistclaimthat“itwillsoonbepos-sibletorationallydesigneffectivesyntheticvaccinesonthebasisofourconsiderableknowledgeofthemolecularconstituentsinvolvedinimmunologicalinteractions.[…]arisesfromanunwarrantedfaithinthepowersofreductionismand[…]overlooksthefactthatprotectionagainstdiseaseachievablebyvaccinationisa[…]conceptthatismeaninglesswhenexpressedonlyinmolecularorchemicalterms”(VanRegenmor-tel,“PitfallsofReductionismintheDesignofPeptide-BasedVaccines”,loc.cit.,p.2369).17DavidHullandMarcH.C.VanRegenmortel,“Introduction”,in:DavidHullandMarcH.C.VanRegenmortel(Eds),PromisesandLimitsofReductionismintheHealthSci-ences,Chichester:Wiley&Sons2002,pp.1-13,quot.exp.11.18JamesMarcum,“MetaphysicalPresuppositionsandScientificPractices:ReductionismandOrganicisminCancerResearch”,in:InternationalStudiesinthePhilosophyof\n212RaffaellaCampanerWhilethereductionistperspectivefocusesonmolecularentitiesandexpressionsoffaultygenes,theorganicistperspectivefocusesonthedisruptionoftissueor-ganization,withoutreachingthelowestphysicochemicallevel.TOFTseescancer“asaproblemakintohistogenesisororganogenesisgoneawry,andthustoaproblemofdevelopmentalbiology”19.Thestrictlyreductionistapproachtocancerresearch–althoughlargelysuccessfulandfarfrombeingdefeated–hasthusbeencriticisedasinsufficient.Evenifafinaltriumphisalongwayoff,andjudgementsarethereforehardtoformulate,ithasbeenacknowledgedthattheorganicistper-spectivehasyieldedsomeimportantobservationsonthedifferentlevelsinvolvedincarcinogenesis(e.g.socialandenvironmental,organismal,tissuelevel).Differ-entapproachestothetopicaffectresearchbyorientingitindifferentdirections.Considerationsontheadequacyofreductionismandantireductionismasgeneralapproachesintheachievementofknowledgeshallalsotakeintoaccountwhichkindofknowledgewearethinkingof.Tostartwith,differentconceptionsmayunderliethesearchforexplanations.Ifthevarietyandrichnessofbiomedicalscientificactivityistakenintoaccount,itcanberevealedhowforawhilecertainlevelswillnotbeinvestigated,butadvancesatonelevel(andnotalwaysalowerlevel)willsometimesopenupanotherlevelforinvestigation,sothatthroughtimethethrustofscientificinvestigationwillwanderfromleveltolevelintheorganizationalhierarchy.Norisitthecasethatthelower-levelexplanationsarealwaysmoreusefulthanhigher-levelexplanations.Itdependsonwhatyouwanttodo,20andontheconceptionofscientificexplanationyouareembracing.Sometimeswecannotphysicallyexplainwhythingsarehappeningacertainwaybecauseweareignorantofwhichlawsorboundaryconditionsareinvolved.“Perhapsthecorrectphysicalexplanationiscomputationallyintractableforus[…].Theremayevenbephysicallaws,particlesorpropertiesthatwedon’tyetknowabout.”21Ifthesecasescanwelloccur,theadoptionofanon-reductionistapproachcanalsobedictatedbyadifferentconceptionofscientificexplanation,sincetheverytopicofexplanationcanbeaddressedindifferentways.If,ingeneral,itcanbeclaimedthatmostscientificeffortsinbiomedicalresearchareaimedattheconstructionofmostlydetailedcausalexplanationsintermsoffundamentalphysicochemicalentitiesandlaws,differentconceptionsofexplanationcanleadustooptforadif-ferentapproach.AshighlightedbysomeofthelatestphilosophicalliteratureonScience19,1,2005,pp.31-45,quot.exp.37.19ChristopherMalaterre,“OrganicismandReductionisminCancerResearch:TowardsaSystemicApproach”,in:InternationalStudiesinthePhilosophyofScience21,1,pp.57-73,quot.exp.60.20DavidHull,“VarietiesofReductionism:DerivationandGeneSelection”,in:DavidHullandMarcH.C.VanRegenmortel(Eds),PromisesandLimitsofReductionismintheHealthSciences,Chichester:Wiley&Sons2002,pp.161-173,quot.exp.171.21Jones,“ReductionismandAntireductionism:RightsandWrongs”,loc.cit.,pp.618-619.\nReductionistandAntireductionistStancesintheHealthSciences213scientificexplanation,higher-levelexplanationswith–forinstance–apragmaticorunificatorycharactercanbebroughtforthandfavoured:iftoexplainistakentobethesearchforverygeneralprinciplesgoverningasystem–astheunificatoryapproachholds–ortoincludemostlycontextualfactors–assuggestedbythepragmaticapproach–outliningfundamentalphysicochemicalpropertiesandlawsmaynotbethebestoption.Thechoicecandependon“whatcountsasarelevantquestion,anadequateexplanationandasufficientdegreeofunderstanding.Reduc-tionistsandantireductioniststendtodisagreeaboutwhattherelevantquestionsareandaboutwhatconstitutesanadequateanswertothesequestions.”22Evenwhenlower-levelexplanationsareavailable,appealtomoregeneralbiomedicalfeaturescanproveessentialand/ormoreadequateintheacquisitionand/ortransmissionofmedicalknowledge.Ifanexplanationofadiseaseistobeprovided,forexam-ple,toapatient’srelatives,oranexplanationofthecontributionsofriskfactorsintheemergenceofapathologyaretobeoutlinedtoapopulationforapreven-tioncampaign,anantireductionistappealtogeneralbiomedicalorenvironmentalprinciplescanworkperfectly.Explanationscanbebothdownward-lookingandupward-looking,atdifferenttimesandindifferentcontexts,andpluralismasfarasthereductionism/antireductionismissueisconcernedcanthereforegotogetherwithsomeformofexplanatorypluralism.Furthermore,explanationisnotnecessarilythefinalgoalofallmedicalen-quiry.Explanatoryknowledgeisnottheonlykindofknowledgewemayachieve:descriptiveknowledge,unifyingknowledge,classificatoryknowledge,predictiveknowledge,orsketchyknowledgeaimedatpracticalinterventions(forexample,targetedatcuresandtherapeuticstrategies)belongtoscientificactivityandcon-tributetoscientificprogress,andmaywellnotbeformulatedinphysicochemicalterms.Thesekindsofknowledgemaybeduenotonlytotemporarylimitationsinourabilitytocarryourreductions,buttospecificinterestsandaimsinagivencon-text.Inbothcases,theyprovidesomerepresentationsofbiologicalsystemsandhencenewknowledgeaboutthem.Newproblemsmaycallforanumberofdiffer-entanswersandneedanumberofdifferentepistemicstrategies.Theadoptionofanon-reductionistperspectivemaybedrivenbystructuralfeaturesofthekindsofknowledgesoughtfor,and/orbythestagewehavereachedintheelaborationofscientifictheorizing.Iftheexplanationsreachedbybiologicalandbiomedicalre-search,whicharenodoubtoneofthemaintargetsofsuchresearch,mostlyfollowareductioniststrategy,andtheirsuccesscanbetakentoprovidegoodgroundsforsupportingreductionism,notallprogressivescientificknowledgeisexplanatory,nordoesitstartoffwithcompleteexplanations.Weoftendonotknowhowtoexplainalivingsystemstraightawayintermsofitsphysicochemicalmicrocom-22VanRegenmortel,“PitfallsofReductionismintheDesignofPeptide-BasedVaccines”,loc.cit.,p.2370.SeealsoClaudeDebru,“FromNineteenthCenturyIdeasonReduc-tioninPhysiologytoNon-ReductiveExplanationsinTwentieth-CenturyBiochemis-try”,in:MarcH.V.VanRegenmortelandDavidHull(Eds.)PromisesandLimitsofReductionismintheHealthSciences,Chichester:Wiley&Sons2002,pp.35-46.\n214RaffaellaCampanerponents,ormaynotbeinterestedindoingso.Antireductionisttenetsmayorientscientificresearch,identify,defineandcircumscribethesystemunderinquiry,thepropertiesweaimtodiscoverandtherelevantboundaryconditions,aswellasprovidefirststagedescriptions,generalizationsandcrucialquestions.Thedifferencesbetweenexplanatoryontheonehand,andunifying,descrip-tive,predictivestrategiesontheother,mayunderliethefactthatmolecularre-searchmethodologiesflanktheexplorationofnon-reductionistresearchstrategies.Ifwetakealookatepidemiology,forinstance,wecanseehowthemaininterestthereisprevention:itisthroughmanipulationofknownvariablesandcorrelationsthattheappearanceandspreadofpathologiesisavoided.Thiscanbeachievedevenintheabsenceofexplanatoryknowledge.Inmanycasesadequateexplana-toryknowledgeisactuallyreachedonlylongafterknowledgeaimedateffectiveprevention.Descriptionsofthesystematstakeandmeanstoexercisesomeformofcontroloveritaresought,andenquiriesconcerningitsrelationswiththeenviron-mentarelargelypursuedforariskassessment.AshighlightedbyPaoloVineisandMicaelaGhisleni,thehistoryofepidemiologyshowsanumberofcasesinwhicheffectivemeasureswereestablishedevenmanydecadesbeforetheachievementofadetailedunderstandingofthefunctioningofthedisease.Guidelinestostartpreventiveactionsweredevised,forinstance,forpathologiessuchasscurvy,pel-lagra,cancerofthescrotum,smallpox,cholera,yellowfever,andothers.Awaitingacompleteexplanationofthediseasewouldhaveentailedseriousdelaysintheelaborationofthepreventivemeasures.Insteadofbeingprimarilyinterestedinthefundamentalentitiesconstitutingthebiomedicalsystematstakeandinrelationsgoverningitsinternalworking,epidemiologyishencestronglyorientedtowardstheimplementationofsuchpreventivestrategies.Itistargetedatpromotingmeas-uressuchas–forexample–abalanceddietandabetterqualityoftheenviron-mentevenintheabsenceofphysicochemicalsatisfyingexplanations.2323SeePaoloVineisandMicaelaGhisleni,“Risks,CausalityandthePrecautionaryPrinci-ple”,in:Topoi23,2,2004,pp.203-210;PaoloVineis,“Laconfusionetracauseemec-canisminell’insorgenzadellemalattie”,in:LaNuovaCiviltàdelleMacchineXXIII,3,2005,pp.113-118.AspointedoutbyVineisandGhisleni,thingsaredifferentforfieldssuchastoxicology,whichisstronglyorientedtowardstheevaluationofchemicalsub-stances.“Traditionaltoxicologyisessentiallycharacterizedbyananalyticalapproach(eachchemicalsubstanceisevaluatedinisolation)andbasedonstrongtheoreticalpremises(inparticular,athresholdoftoxicity)”(VineisandGhisleni,“Risks,Causal-ityandthePrecautionaryPrinciple”,loc.cit.,p.207).Epidemiologicalobservationshaveprovedthatthelattercanbeseriouslymisleading.Seeibid.,pp.207-208.\nReductionistandAntireductionistStancesintheHealthSciences2153.REDUCTIONISTANDANTIREDUCTIONISTSTANCES:FORAPLURALISTICAPPROACHHowdoreductionisttenetsfarewithrespecttosuchfieldsandlinesofresearchaspresentedinsectiontwo?Isitreallynecessarytolodgeoneselffirmlyatoneex-tremeofthecontroversybetweenreductionistsandantireductionists,oraretherewaystopreserveboththestancestheyexpress?Inthelightoftheexamplesmen-tionedabove,webelievethatwithrespecttothehealthsciencesthedisputecanbelargelyseenasoneoveremphasis,conditionalonthecontextandtargetsoftheenquiry.Ontheonehand,thedissectionofbiomedicalsystemsinto[their]componentshasgivenusconsiderableinsightintotheworkingsofthesystem[s],afactthatevenstaunchanti-reductionistsdonotdeny.Whatiscontroversial,however,istheextenttowhichdescriptionsoftheisolatedcomponentsinmoleculartermsareabletoprovidethetypeofexplanation,thelevelofunderstandingandthepredictiveability24thatscientistssuchas–aswehaveseen–immunologistsandepidemiologistswouldliketohave.Weshallconcludethatreductionistandnon-reductioniststanc-es,and,moreprecisely,reductionistexplanationsandinvestigativenon-reduction-ism,canbeadoptedjointlyforatentativelyexhaustiveunderstandingofthephe-nomenaatstakeinthehealthsciences.“Thereductionofvitalphenomenatoirreduciblephysicalandchemicalele-mentsand/orrelationshipsisbothanecessaryandaninsufficientcondition”25inthehealthsciences.Explanatoryreductionismcanbeadmittedtogetherwithanantireductionistapproachtoboththegeneraladvancementofmedicalknowledgeandtheconsiderationoftheembodieddisease,i.e.thesickperson.Explanatoryreductionismcanprovideanessentialtoolintheunderstandingofdiseasesandtheelaborationofgeneralmodelsoftheirfunctioning,but,giventhepeculiaritiesofthehealthsciences,thereductionistmindsetdoesnotseemtosuffice.Withregardtoclinicalpractice,viewingandtreatingthepatientasabiologicalentirety,notsimplyinmolecularterms,andthinkingintermsofaconcreteillness,notanabstractdisease,isafundamentaldemand,tobemetinordertoestablishacorrectdiagnosisandprognosis.Theadoptionofreductionisminanexplanatoryperspectivedoesnotimplyareificationofthemedicalobjectasinvolvedinclini-calpractice,noradenialofotherformsofknowledgeinadditiontoexplanatoryknowledge.Aswehaveseen,antireductionism,or,better,non-reductionism,canproveaveryimportantepistemicstrategyinotherfieldstoo,suchasimmunol-ogyandepidemiology.However,thisisnottosaythatextremeholismhastobeembraced.Itcertainlydoesnotprovideaviablemethodologicalalternative.Thata24VanRegenmortel,“PitfallsofReductionismintheDesignofPeptide-BasedVaccines”,loc.cit.,p.2370.25Debru,“FromNineteenthCenturyIdeasonReductioninPhysiologytoNon-Reduc-tiveExplanationsinTwentieth-CenturyBiochemistry”,loc.cit.,p.40.\n216RaffaellaCampanerreductionistperspectivehasbeenandcontinuestobeverysuccessfulinattainingexplanatoryknowledgecanalsoberecognisedbynon-reductionists,whoadmitthatreductionismiseffectiveinexplainingthechemicalbasisofalargenumberoflivingprocesses.Furthermore,theimportanceofmedicineasasciencelargelyrestsonitscapacityofhealing,whichinturnlargelyrestsonaphysicochemicalbasis.Theeffectivenessofpharmaceuticaldrugs,variouskindsofcuresanddi-agnostictechniques(see,forinstance,electronicmicroscopes,chemotherapyandradiotherapy,positronemissiontomography,tomentionbutafew)undeniablydependsonphysicochemicalinvestigations.Withoutdenyingthatreductionismshouldbegivenachance–asElginclaims–and,rather,acknowledgingthatitisthemostsuccessfulexplanatorystrategy,Isuggestthatexaminingthe“presentsituation”inthebiomedicalsciences–whichistheaimofthisvolumeandtheconferenceitoriginatesfrom–wecannotignorethatsomeimportantnon-reductioniststancesarealsopresent.Referringtovariousquotationsfrombiochemistrytextbooks,Elginconcludes“thecommonthemeinalltheseprinciplesistheideathatlifecanbeunderstoodintermsofchemicalorphysicalpropertiesofbiomoleculesandtheirinteractions”.26Itisclaimedthatthesuccessofbiochemistryislargelyduetothiscommontheme,andthatitisonthesegroundsthatexplanatoryreductionismcanbesupported,holdingthattoexplainabiologicalsystemistoprovideanaccountofthecombinationandarrangementofmorebasic,physiochemicalentitiesandproperties.Thisdoesnotclashwithmethodsinspiredbynon-reductionism,indicating,forinstance,howtooperatethefirstcutswhenoneisinvestigatingaverycomplexsystem,andhowtoselectthefeaturesthatwillproverelevant.Withoutneglectingthepopularityofreductionasascientificstrategyandwithoutadvocatinganythinglikea“phenomenological”modelofthebody,27areciprocalintegrationofbothreductionandnon-reduction-isttenetscanbehypothesized,whichcouldultimatelybeacceptedinthekindsofbiomedicalexamplesmentionedabove.Suggestingthatreductionismasacrucial,successfulstrategycanbeaccompaniedby,andbenefitfrom,anantireductionistepistemologymighthelpavoidingthelimitationsduetoassumingjustreduction-ism,whichlookslike“aninsufficientpresuppositionformedicalknowledgeandpractice”.28WhilesympatheticwithElgin’sreductionismasfarasexplanationisconcerned,Ifinditplausiblethatbothreductionistandantireductionisttenetscanpointtoexpectationsaboutfurtherknowledge,andthuscontribute,indifferentways,toprogressinthelifesciencesandtoadeeperconsiderationofthedifferentdisciplinaryfieldsbelongingtothehealthsciences.26Elgin,“ReductionisminBiology:AnExampleofBiochemistry”,loc.cit.,p.x.27SeeMarcum,“BiomechanicalandPhenomenologicalModelsoftheBody,theMean-ingofIllnessandQualityofCare”.28JamesMarcum,“ReflectionsonHumanizingBiomedicine”,PerspectivesinBiologyandMedicine51,3,2008,pp.392-405,quot.exp.395.\nReductionistandAntireductionistStancesintheHealthSciences217Neitherreductionismnoranti-reductionismshouldbeseenastheonlygameintown,andpluralismisnottobearrivedatas“thelastresortoflosers”.29Acare-fullookatactualscientificpracticerevealsahighsensitivityofmethodologiestocontexts,whichreductionismgenerallytendstode-emphasize.Inevaluatingwhetherreductionismshouldbeprivilegedinourtoolboxtoenhancescientificknowledge,whatfieldofbiomedicalresearchwearethinkingofoughttobeclari-fied.Medicinehasthestatusofastronglymultidimensionalscience,whichex-ploitsthesupportofmultipledisciplines,andpresentsmultipleattitudestowardsthereductionism/antireductionismissue.Insteadofbeingaterrainforconflictbetweenreductionistandantireductioniststances,thehealthsciencescanbere-gardedasastageinwhichdifferentemphasiscanbeputondifferentknowledgeprocessesindifferentcontexts.Notanypluralismwilldo,butbothreductionistexplanatorymethodsandnon-reductionistattitudestothediseaseandtothesickpatientmeritconsideration.Notonlyarereductionistandantireductionisttenetscompatibleoncetheyareproperlyunderstoodandcontextualised,buttheycanbetakenasatooltobetterclarifyandhighlightspecificfeaturesofsinglemedi-calfields.Toembraceapluralisticapproachinthissenseisnottofragmenttheanalysis,butrathertoacknowledgeandtakeintodueaccountthemethodologicalpeculiaritiesofthedifferentfieldsofmedicalresearchandtheirfinalaims,whichincludeunderstandingthediseaseprocess,preventingitandtakingcareofthesufferingpatient.Themultifariousadoptionofreductionistandantireductioniststancesinfieldssuchasbiomedicalresearchandphysiopathology,epidemiology,immunologyandclinicalmedicinecanhencebetakenashintstowardstheneedtodevelopdifferentphilosophiesofmedicine,eachtoaccountforthefoundationalandmethodologicalpeculiaritiesofthesefields.FURTHERBIBLIOGRAPHICREFERENCESPhilipKitcher,“ExplanatoryUnificationandtheCausalStructureoftheWorld”,in:PhilipKitcherandWesleySalmon(Eds),ScientificExplanation,Minne-apolis:Univ.ofMinneapolisPress1989,pp.410-505.PeterMachamer,LindleyDardenandCarlCraver,“ThinkingaboutMechanisms”,in:PhilosophyofScience67,1,2000,pp.1-25.WesleySalmon,ScientificExplanationandtheCausalStructureoftheWorld,Princeton:PrincetonUniversityPress1989.WesleySalmon,“FourDecadesofScientificExplanation”,in:PhilipKitcherandWesleySalmon(Eds),ScientificExplanation,Minneapolis:Univ.ofMinne-apolisPress1989,pp.3-219.WesleySalmon,CausalityandExplanation,Oxford:OxfordUniversityPress1998.29VanRegenmortelandHull,“Introduction”,loc.cit.,p.1.\n218RaffaellaCampanerSahotraSarkar,“GenesversusMolecules:HowTo,andHowNotTo,BeaRe-ductionist”,in:MarcH.V.VanRegenmortelandDavidHull(Eds),PromisesandLimitsofReductionismintheHealthSciences,Chichester:Wiley&Sons2002,pp.191-206.KennethSchaffner,DiscoveringandExplanationinBiologyandMedicine,Chi-cago:TheUniversityofChicagoPress1993.KennethSchaffner,“Reduction:theCheshireCatProblemandaReturntoRoots”,in:Synthese,151,3,2006,pp.377-402.BasvanFraassen,TheScientificImage,Oxford:OxfordUniversityPress1989.MarcelWeber,PhilosophyofExperimentalBiology,Cambridge:CambridgeUni-versityPress2005.DeptartmentofPhilosophyUniversityofBolognaViaZamboni2840126BolognaItalyRaffaella.Campaner@unibo.it\nPartIV(TeamC)(teamleaderWenceslaoJ.González)ThePresentSituationofthePhilosophyoftheCulturalandSocialSciences:The“NaturalistTurn”,the“SocialTurn”,andtheDiscussiononScientificRealism\n\nWENCESLAOJ.GONZÁLEZTRENDSANDPROBLEMSINPHILOSOPHYOFSOCIALANDCULTURALSCIENCES:AEUROPEANPERSPECTIVE1.TheLevelsofAnalysisDealingwiththeCulturalandSocialSciences2.TheKindofApproach:“AEuropeanPerspective”3.The“NaturalistTurn”,the“SocialTurn”,andtheDiscussiononScientificRealism4.Explanation,Prediction,andConfirmation:RealmandLimits5.TheDebateonMathematicalModelingintheSocialSciencesandConsequencesforExperimentation6.TheSciencesthatPhilosophyhashithertoIgnored:TheSciencesofDesign7.NewDirectionsinthePhilosophyofScience8.FinalRemarksForthisinitialconferenceoftheprogramonThePhilosophyofScienceinaEu-ropeanperspective,theSteeringCommitteesuggestedanoverviewofthetopicsofTeamCtobediscussedduringthefiveyearsofthisproject.Thebroadtitle–“TrendsandProblemsinPhilosophyofSocialandCulturalSciences:AEuropeanPerspective”–wasawayofmeetingtheaimsuggested.Thus,theoriginalfocuswastooffersomephilosophicalremarksintermsofpossiblelinesofdiscussioninthiscommonendeavor.1Butthistaskrequirespreviousreflectiononthegeneralframework:thekindofanalysistobedevelopedandonwhatmightbeconsideredas“aEuropeanperspective”.ThetwofoldconsiderationtakesintoaccounttheelementsoftheproposalsenttotheEuropeanScienceFoundation.2Thepapertriestomakethemmoreexplicitaswellastodrawsomeconsequences.Thus,theoverviewofthetopicstobeexaminedbyTeamC,orientedtowards“ThePhilosophyofSocialandCul-turalSciences”,startswithsomeremarksonwhatthephilosophicalcontextunderdiscussionhereis.Thisinvolvesanexplicitreflectionontwodifferentaspects:thelevelsofanalysisandthekindofapproach.1SomeadditionaltopicsofdiscussionwhichalsohaveapresenceinEuropecanbefoundinHaroldKincaid,“SocialSciences”,in:PeterMachamerandMichaelSilber-stein(eds.),TheBlackwellGuidetothePhilosophyofScience,Oxford:Blackwell2002,pp.290-311;especially,pp.306-307.2Cf.SteeringCommittee,ThePhilosophyofScienceinaEuropeanperspectivePropos-alofan“àlacarteProgramme”tobesubmittedtotheEuropeanScienceFoundation,24February2006.F.Stadler(ed.),ThePresentSituationinthePhilosophyofScience,ThePhilosophyofScienceinaEuropeanPerspective,DOI10.1007/978-90-481-9115-4_18,©SpringerScience+BusinessMediaB.V.2010\n222WenceslaoJ.GonzálezOntheonehand,thereareinfactseverallevelsofanalysistobeconsideredwhendealingwiththeculturalandsocialsciences.Amongthem,thisphilosophi-cal-methodologicalstudyhasimportantconnectionswithtendenciesofgeneralphilosophyandmethodologyofscience.3Ontheotherhand,thekindofapproachusedtoundertakethisparticularstudy–a“Europeanperspective”–isparticularlyrelevantandneedsalsosomeclarification,insofarasthereareseveralwaystocharacterizeit.Afterthistwofoldframeworkofthephilosophicalcontext,thereisapre-sentationofthemaintopicstobeconsideredineachyearoftheprogram:1)the“naturalistturn”,the“socialturn”,andthediscussiononscientificrealism;2)explanation,prediction,andconfirmation:realmandlimits;3)thedebateonmath-ematicalmodelinginthesocialsciencesandconsequencesforexperimentation;4)thesciencesthatphilosophyhashithertoignored:thesciencesofdesign;and5)newdirectionsinthephilosophyofscience.Thiscontentisaccompaniedbysomeadditionalremarksontheculturalandsocialsciencesinordertomakecertainphilosophical-methodologicalpointsexplicit.Commonly,theseremarksbelongtothe“internal”perspectiveonscience(language,structure,knowledge,method,etc.).4Theyassumethatthefoundation-alandmethodologicaldebatehasacentralroleintheconfigurationoftheculturalandsocialsciences.Anadditionalstudy,whichisbeyondthelimitsofthepresentpaper,couldincludeadetailedanalysisofthe“external”pointofview:social,po-litical,economic,…aspects.ThiscontextualviewrequirestheempiricalresearchontheinstitutionsthatworkinEuropeonthesetopics.5Nonetheless,somepieces3Onthegeneralphilosophical-methodologicaltendenciessince1980,seeWenceslaoJ.Gonzalez,“NoveltyandContinuityinPhilosophyandMethodologyofScience”,in:WenceslaoJ.GonzalezandJesusAlcolea(eds.),ContemporaryPerspectivesinPhilosophyandMethodologyofScience,ACoruña:Netbiblo2006,pp.1-28.Italsoincludessomeremarksonthestatusof“methodologyofscience”anditspresentrela-tionswith“philosophyofscience”.4Onthephilosophicalapproach:from“internal”to“external”,seeWenceslaoJ.Gonza-lez,“ThePhilosophicalApproachtoScience,TechnologyandSociety”,in:WenceslaoJ.Gonzalez(ed.),Science,TechnologyandSociety:APhilosophicalPerspective,ACoruña:Netbiblo2005,pp.13-20.5TheproposalitselfsubmittedtotheEuropeanScienceFoundationlistedintheIn-troductionanumberofresearchcentersinEuropefocusedonphilosophyandmeth-odologyofscience:“InstituteViennaCircle,Vienna,Austria;CentreforLogicandPhilosophyofScience,GhentUniversity,Belgium;MaxPlanckInstituteforHumanDevelopment,Berlin,Germany;ZentrumfürPhilosophieundWissenschaftstheorie,KonstanzUniversity,Germany;ZIF,ZentrumfürinterdisziplinäreForschung,Uni-versityofBielefeld,Germany;EuropeanCulturalCentreofDelphi,Delphi,Greece;CIRESS,InterdisciplinaryResearchCentreforEpistemologyandHistoryofScience,UniversityofBologna,Italy;InstituteforHistoryandFoundationsofScience,UtrechtUniversity,Netherlands;SCASS,SwedishCollegiumforAdvancedStudiesintheSocialScience,Uppsala,Sweden;CPNSS,CentreforPhilosophyoftheNaturalandSocialScience,LSE,London,UK”.Manyofthemdealinonewayoranotherwith\nTrendsandProblemsinPhilosophyofSocialandCulturalSciences223ofinformationofthecontextualviewarealsocontainedhereinconnectionwiththe“internal”perspective,whichisthemainstreamofthepresentpaper.1.THELEVELSOFANALYSISDEALINGWITHTHECULTURALANDSOCIALSCIENCESThereareatleastthreedifferentlevelsofanalysisrelatedtothesesciences.First,thegeneralscientificstatusoftheculturalandsocialsciences,whichrequiresustoconsidersuchdiverseelementsofscience(mainlylanguage,structure,knowledge,method,activity,aimsandvalues).Second,thescientificstatusoftheculturalandsocialsciencesascomparedtothatofthenaturalsciences.Amongotherproblems,thisconcernsmethodologicalcontroversiessuchasErklären-Verstehen.6Third,thespecificissuesonthescientificstatusofeachculturalandsocialdiscipline(economics,psychology,sociology,archaeology,anthropology,law,etc.).Thesespecificaspectscanbeseenineveryscienceofthisrealm(e.g.,thoserelatedtopredictionineconomics).Thisanalysisincludesthosesciences(somenew)thathavebeenhithertoignoredinthephilosophicalliterature(suchasthesciencesoftheartificialunderstoodassciencesofdesign).7Withinthisthirdlevelofanalysis,therearetwophilosophical-methodologicaloptions:thebroadapproachandtherestrictiveposition.Inthecaseofthebroadapproach,thereisinterestinconnectingthephilosophicalreflection(i.e.,thecon-siderationofthesemantical,logical,epistemological,methodological,ontological,axiological,ethical,etc.,aspects)ontheparticulardiscipline(economics,psychol-ogy,sociology,etc.)withtopicsonscienceingeneral.Inthecaseoftherestrictiveposition,thefocusofattentionisontheproblems(mainlymethodological)ofthedisciplineanalyzedwithoutanyrealinterestinscienceingeneral.Dependingonthedegreeofgeneralityintheanalysis,thereisascaleofpro-gressivespecializationoftheresearchmadeinphilosophyandmethodologyoftopicsrelatedtothesocialsciences.Obviously,thelistisnotexhaustive,andrequiresmanyUniversityDepartmentsinHistoryandPhilosophyofSciencethathavebeenestablishedacrossEuropetobetakenintoaccount.6Cf.WenceslaoJ.Gonzalez,“FromErklären-VerstehentoPrediction-Understanding:TheMethodologicalFrameworkinEconomics”,in:MattiSintonen,PetriYlikoskiandKarlMiller(eds.),RealisminAction:EssaysinthePhilosophyofSocialSciences,Dordrecht:Kluwer2003,pp.33-50.InthecaseoftheErklären-Verstehen,themainnamesinvolvedinthesevenconceptionsofthismethodologicalcontroversyareEuro-peans:WilhelmDilthey,MaxWeber,CarlGustavHempel,PeterWinch,HansGeorgGadamer,GeorgHenrikvonWright,andKarlOttoApel.7Commonly,thecompanionsandthebooksofreadingsmadebyEuropeanpublish-inghousesarefocusedonthesecondandthirdlevel.SeeforexampleStephenP.Turner/PaulA.Roth(eds.),TheBlackwellGuidetothePhilosophyoftheSocialSci-ences,Oxford:Blackwell2003.(Theinterestingthingisthat,inthiscase,onlythreeoutofthirteenoftheauthorsworkataEuropeanuniversity.)\n224WenceslaoJ.Gonzáleztheculturalandsocialsciences.Atthefirstlevel,thereisadirectrelationwithphilosophersofsciencewhoworkedingeneralconceptionsofscience(suchasKarlPopper).Frequently,thesegeneralapproachesonscience(Popperian,etc.)haveadirectincidenceonthefollowinglevelsofanalysis,eventhoughsomeofthoseviews(Kuhnian,Lakatosian,etc.)werenotinitiallythoughtfortheculturalandsocialsciences.Withinthesecondlevel,theinterestcancomefromscientistsandphiloso-phersinterestedmostlyintheculturalandsocialsciences(CarlMenger,MaxWe-ber,FriedrichvonHayek,HansGeorgGadamer,KarlOttoApel,etc.).Thus,itisverycommonthattheydiscusstopicsrelatedtothesimilaritiesanddifferencesbetweenthenaturalsciencesandthesocialsciences.Thesethemesincludethemethodologicaldifferencesbetween“explanation”and“understanding”,whichalsohaverepercussionsforcentralissuesofthesocialsciencessuchasprediction(possibility,reliability,etc.).Atthethirdlevel,thephilosophicalpointofviewisfocusedoneachscience(economics,psychology,etc.).Sotheycanfollowabroadapproachonthosesci-ences,whentheyareopentogeneralissuesofscience(suchasinthecaseofphilosophersofeconomics:MarkBlaug,UskaliMäki,etc.);ortheycanadoptarestrictiveposition(e.g.,whentheanalysisisspecificallymethodological),andwhentheyarescientiststhatdonothavearealinterestforphilosophyofscience(asfrequentlyhappenswithexpertsinstatisticaleconomicsandeconometrics).2.THEKINDOFAPPROACH:“AEUROPEANPERSPECTIVE”Thesethreelevelsofanalysis–general,comparative,andspecific–canbecon-sideredwithin“aEuropeanperspective”.Butitisacomplexissuetoestablishtheprecisesenseandreferenceofthatexpression.Inthisregard,theinitialproblemisthequestionoftheexistenceofaspecificEuropeanapproachindealingwithphil-osophicalquestionsofscience:isthereatoutcourt“Europeanperspective”ratherthanamereviewpointheldbysomerepresentativeEuropeans?Thisontologicalquestionisundoubtedlylinkedtotheproblemofanadequatecharacterizationof“aEuropeanperspective”inphilosophyofscience.Thiscategorizationshouldgobeyondsemanticalconsiderationstodealwithepistemologicalandmethodologi-calaspects.Obviously,thepresentprogramoftheESFassumesasamatteroffactthatthereis“aEuropeanperspective”inphilosophyofscience.Inaddition,thepro-posalsenttotheEuropeanScienceFoundationdepictssomefeaturesofthatperspective.Thus,itappearstobethecombinationoftheamountofphilosophi-calcontributionsonsciencemadeindifferentcountriesofthiscontinent,bothhistoricallyandthematically.Inaddition,the“Europeanperspective”seemsto\nTrendsandProblemsinPhilosophyofSocialandCulturalSciences225besomethingdevelopedwithspecifictraits,somehowdifferenttoviewpointsofotherplacesintheworld.Certainlywecansee“aEuropeanperspective”fromdifferentangles,amongthemisthehistoricalapproachandthethematicview.Thefirstonecanbecon-nectedwiththerealityofpast,whereasthesecondonepaysmoreattentiontothepresentandtothefuture.Thismeansthatwecanconsiderwhat“aEuropeanper-spective”isaswellashowitoughttobe.Inthisregard,thephilosophicaloutlookneedstobeanalytical(i.e.,theconsiderationofthepastandthepresent)aswellasprescriptive(i.e.,themeta-reflectiononthefuture).1)Historically,wecanunderstandthatthereisdefactoalongtraditionofdo-ingphilosophyinEurope.8Thus,amongotherphilosophicalbranches,“philoso-phyofscience”–asweconceiveitnow–hasbeendevelopedinonewayoran-otherinmanyEuropeancountries.Butthiscouldalsobeseenasmererecognitionofafactumratherthanaformofdescribingsomethingparticularorcharacteristicofthiscontinent.2)Wecanthinkofthisissueinamorethematicview:thereisawayofdoingphilosophyofsciencethatitissomehowdistinctiveofEuropeand,therefore,differentfromthephilosophyofsciencemadeinAmericaorinothercontinents.Thus,inthelatter–thecontentsandstyleofthinking–theissueof“identity”becomesmorerelevantthanintheformer(i.e.,inthementionedhistori-calapproach).ThisrouteofsearchingforaEuropeanperspectiveintermsof“identity”canleadtoseveralpositions,amongthemarethe“integrative”position,thespecificviewbasedonthehistoricalbackground,andamorerigorousconceptionuponthematicterms.Allthreeoptionsareopen,inprinciple,tothethreelevelsofphilo-sophical-methodologicalanalysisalreadypointedout:general,comparative,andspecific.a)An“integrative”positionisadoptedwhenthefocusofattentionisonthesearchforacommonground,bothinhistoricaltermsandinthematicones.Sothe“Europeanperspective”istheresultofthecontributionsmadebyinstitutionslocatedinEuropeancountriesaswellastheworkofphilosophersthateitherareofanyEuropeannationality(insofarasthereisnoactual“Europeannationality”)orareresidentsinanynationofthiscontinent.b)Amorespecificviewthanthepreviousonecomesfromemphasizingthatthe“Europeanperspective”isbasedonahistorical-methodologicalapproach.Accordingtothisstandpoint,the“Europeanperspective”isonlyvalidforthose8Ifwefocusoncontemporaryphilosophy,therearesomephilosophicalconceptionsthatstartedonEuropeansoilandhavehadinfluenceonthephilosophyoftheculturalandsocialsciences.Theseviewsincludethefollowing“stylesofthought”:a)theana-lyticalphilosophyoflanguagedevelopedbyG.Frege,B.Russell,L.Wittgenstein,etc.;b)thelogicalempiricistviewsbasedonR.Carnap,H.Reichenbach,etc.;c)thehermeneuticaltraditionofthinkersinterestedinhistoricalaspects,suchasH.G.Ga-damer;andd)thepostmodernapproachesofstructuralistrootsthathavehadparticularrelevanceinFrenchauthors.\n226WenceslaoJ.Gonzálezphilosophicalmovementsthatoriginallystartedinthiscontinentorhavebeende-velopedmostlyinEurope,eventhoughtheinitialoutsetswereestablishedabroad.Inthiscasehasaspecialweighttherecognitionoftheresearchtradition.c)Ifthereisaparticularstressonthethematicelements,thenamorerigorousconceptionofthe“Europeanperspective”canbeadoptedthatassumestheexist-enceofstrictboundaries.Thismeanstakingintoaccountonlythosephilosophi-calmovementsrootedinEuropeanddevelopedinthiscontinent,otherwiseitisconsideredthattheydonothavethespecificfeaturestobeEuropeanorthattheybelongtotheso-called“internationalphilosophy”.ThesethreeoptionsrequirestotakeintoaccounttheProposalofan“àlacarteProgramme”thatwassubmittedtotheEuropeanScienceFoundation,entitledThePhilosophyofScienceinaEuropeanperspective.Thedocumentgivesaguideonhowtheissueistobeunderstoodinthisproject.9Moreover,theguidelinedirectlyaffectsthepresenttopicoftrendsandproblemsinphilosophyofsocialandcul-turalsciencesbecauseitgivesa“roadmap”toembracetheEuropeanperspectiveatstake.Thiscanbesummarizedinfourelements:I)TheanalysisoftheproposalislinkedtorelevantscientiststhatwerebornandlivedinEuropeancountries(suchasCarlMenger,LudwigvonMises,OttoNeurath,etc.).II)Aspecialemphasisisputontheinstitutionalcontributiontothistopic,bothinhistoricalterms(theViennaCircle,theBerlinSchool,etc.)andinourtimes,pointingouttheexistenceanincreasingnumberofcentersdevotedtophilosophyofscienceinEurope(someofthemexplicitlyfocusonthesocialsci-ences).10III)Thegeneralapproachcanbeseenasan“integrative”position:thein-terestismainlyinwhatisbeingdoneacrossEuropewithoutmorespecifications.11IV)Thereisaclearrecognitionoftheproblemsregarding“identity”,inspiteoftheamountofworkdoneinthiscontinentonthissubject-matter:“althoughthere(...)isasolidgrowthininterestandnumbersofpeopleworkinginthesubject,thereisalackofcoherenceintheEuropeanresearcheffort”.12Therefore,inthispapertheoverviewofthetrendsandproblemsfromaEu-ropeanperspectiveisawareofgeographicalandinstitutionalcomponents.Item-phasizesan“integrative”positioninthephilosophyofsocialandculturalsciences9Cf.SteeringCommittee,ThePhilosophyofScienceinaEuropeanperspectivePropos-alofan“àlacarteProgramme”tobesubmittedtotheEuropeanScienceFoundation,Introduction.10AmongothersaretheSwedishCollegiumforAdvancedStudiesintheSocialScience(SCASS),Uppsala(Sweden),andtheCentreforPhilosophyoftheNaturalandSocialScience(CPNSS)attheLondonSchoolofEconomics(LSE),London,UK.11“EventhoughEuropeisnolongeraloneinsettingtheparametersfordiscourseinandaboutscience,duringthelastfewdecadesarenewedandincreasinginterestinphilo-sophicalissueshasbeenshownagainbyscholarsalloverEurope”,SteeringCommit-tee,ThePhilosophyofScienceinaEuropeanperspectiveProposalofan“àlacarteProgramme”,Introduction.12SteeringCommittee,ThePhilosophyofScienceinaEuropeanperspectiveProposalofan“àlacarteProgramme”,Introduction.\nTrendsandProblemsinPhilosophyofSocialandCulturalSciences227opentoanidentityfactor,duetotheexistenceofawidevarietyoforientations.Consequently,theremarkswilltakeintoaccountthethreelevelsofanalysispoint-edoutonthescientificstatusofthesocialsciences–general,comparative,andspecific–aswellastheguidanceoftheproposalapprovedbytheEuropeanSci-enceFoundation.Putdifferently,theremarksaremadehereonthebasisoftheexistenceoffourmainelements:(i)therelevantscholarsrelatedtoEurope,becausetheirnational-ityorplaceofresidence;(ii)theinstitutionsthatdevelopresearchonphilosophyofsocialandculturalsciences;(iii)the“integrativeview”;and(iv)thelackofcoherenceinordertohaveaneatpictureoftheworkdoneinthecontinent.Becauseofthespaceavailablefortheseremarksandtheexplicitcharacterofoverviewofthispaper,whichmakesclearthatitisnotanexhaustivestudy(ofnames,tendencies,institutions,etc.),thefocuswillbemainlyinthecentralprob-lemsofthephilosophyofsocialandculturalsciencesratherthaninotheraspects.Thesequenceoftopicswillbepresentedaccordingthescheduleforthefiveyearsoftheprogram.Obviously,theattentionwillbepaidinthemainaspectsandthosethinkerswithmoreinfluencesofarinthetopicstobediscussed.3.THE“NATURALISTTURN”,THE“SOCIALTURN”,ANDTHEDISCUSSIONONSCIENTIFICREALISMFollowingthethreelevelsofanalysispointedout–general,comparative,andspecific–,somekeytendencies–the“naturalistturn”,the“socialturn”,andthevarietiesofscientificrealism–areapproachesthatcanbeseeninthefirstandsecondlevels.Thesephilosophical-methodologicalviewpointsaredirectlycon-nectedtothedebatesonthescientificstatusofthesocialsciencesaswellastothecomparisonbetweenthenaturalandsocialsciences.Thesetrendsthatbelongtoa“post-historicalturn”,andtheyhavebeenveryinfluentialformorethantwodecades.Allofthem–theviewsthatconform“naturalisticturn”,“socialturn”,andtherealistconceptualizations–haveanover-archingviewpointonscience.Theyacceptageneralframethatiscommonforempiricalsciences.Fromthatframecanbereachedanyculturalorsocialscience,eventhougheachoneofthoseview-pointsinvolvesasetofdifferentconceptions(i.e.,therearedifferenttypesofnaturalism,adiversityofrealisms,avarietyofsocialorientations…).Moreover,someofthemcanbecombined,becauseitispossiblethatathinkermightbenaturalistinsomeissues(e.g.,epistemologicalandmethodological)andrealistinotherpoints(e.g.,semanticalandontological).Inanycase,thethreegeneralviewpointsmentioned–naturalist,social,andrealist–involvesomekindof“foundational”approachonsocialsciences–anepistemologicalgrounding–aswellasamethodologicalcharacterizationoftheresearchonsocialsciences.\n228WenceslaoJ.González1.Naturalisminthesocialsciences–thesecondlevelofanalysis:thecom-parativeposition–presupposesnaturalisminscienceingeneral(thefirstlevelofanalysis).Inthisregard,wecanfindseveralkindsofnaturalismsinsciencethathaverepercussionsonthenaturalismsinthesocialsciences,amongthosegeneralphilosophical-methodologicalviewsonsciencearethese:(i)Semanticnaturalism,wherethereisanacceptanceofmeaningaslinguis-ticuse,becausemeaningisbasedonapracticethatcanbedescribedratherthanprescribed.(ii)Epistemologicalnaturalism,whichacceptsthathumanknowledgeiswellorientedandassumesacontinuitybetweenscienceandphilosophy(and,then,thatametaphysicalfoundationofanyofthemisnotneeded).(iii)Methodo-logicalnaturalism,wheretheprogressofscience(includingthecaseofthesocialsciences)canbemadethroughprocessesempiricallytestedaccordingtocriteriausedinnaturalsciences.(iv)Ontologicalnaturalism,whichonlyacceptsenti-tiesthatinonewayoranothercouldbeobservable(i.e.,itdeniesthelegitimacyofunobservableentitiessuchas“mind”,“consciousness”,andthelike).(v)Axi-ologicalnaturalism,wherethescientificvaluesarethosethatcomefromscientificpractice.13Undoubtedly,naturalisminthesocialscienceshasalongtraditioninEurope.Thisviewassumes,inprinciple,thatsocialsciences,bothasawholeandsingu-larly,aregroundedonanepistemologicalandmethodologicalbasisanalogous–orevenidentical–tothatofthenaturalsciences.Thus,keyelements–laws,models,andregularities–canbeanalyzedsimilarlytowhatisdoneinphysics,biology,etc.Eachdiscipline(economics,psychology,sociology,archaeology,an-thropology)offersexamplesofthenaturalisticperspective,suchasthedefenseofeconomiclaws,reductionistmodelsinpsychologyandinsociology,orcausalexplanationsinarchaeology.Averyinfluentialversionofnaturalismcomesfromtheuseofanevolutionaryframeworkforexplainingandpredictinginthesocialsciences.Whenanalternativeviewtonaturalisminthesocialsciencesisadopted,therecouldbedifferentoptions,eitherstrong(i.e.,anti-naturalism)ormoderate(i.e.,aninterpretativeperspective).Thesepossibilitiesleadtostresshumanaction–in-dividualandsocial–asmeaningfulandruledbyintentionality.Therefore,theinternalfactors–intentions,beliefs,goals,etc.–andtheexternalaspects–social,cultural,political,etc.–ofhumanactioncanmodifythebasicmethodologicaltools(laws,modelsandregularities).Asaconsequence,adualistmethodologyisassumed,eitherinstrictterms(anti-naturalismrejectsaconvergencewithmeth-odsofthenaturalsciences)orinamoreflexibleway(ininterpretativeversions).Eachdisciplinefeaturesexamplesofalternativeapproachestonaturalism:rhetoricineconomics,interpretativesociology,narrativeconceptionofhistory,symbolicsocialanthropology,etc.Inthisregard,thelegacyofWittgenstein’s13Cf.WenceslaoJ.Gonzalez,“NoveltyandContinuityinPhilosophyandMethodologyofScience”,Op.cit.,p.5.\nTrendsandProblemsinPhilosophyofSocialandCulturalSciences229viewsonthelanguageofaction(eitherdirectly14orthroughsomeauthorssuchasGeorgHenrikvonWright15)ortheimpactofhermeneuticalelements(eitherinaGadamerianflavor16orbasedonanyoftheproposalsoftheFrankfurtSchool17)haveworkedoutinfavorofanalternativetonaturalisminthesocialsciences.Intentionalityandinterpretationareusedtodistinguishculturalandsocialsci-encesfromnaturalsciences.TheseviewscanbeseenintunewithsomeformofVersteheninsteadofsupportinganErklären.2.Meanwhilethe“socialturn”inthephilosophyofthesocialsciencesisfrequentlyconnectedtoKuhnianschemes,18eventhoughKuhnhimselfdisagreeswithsomeoftheseinterpretations19.AfterthepublicationofTheStructureofScientificRevolutions,therewasinEuropea“sociologicalswitch”basedonanexpansionofKuhnianideas.Theyhavefolloweddifferentviews,whichcanbesummarizedinthefollowingways:a)Thefinalizationthesis(theconceptionoftheFinalizierungderWissen-schaft)developedbythegroupthatworkedattheMaxPlanckInstitutatStarn-berg;20b)thestrongprogrammeinsociologyofscienceoftheEdinburghSchool14AnexampleisPeterWinch,TheIdeaofaSocialScience,London:RoutledgeandK.Paul1958(2ndedition,1990).15Cf.GeorgHenrikvonWright,ExplanationandUnderstanding,Ithaca:CornellUni-versityPress1971;GeorgHenrikvonWright,“Replies”,in:JuhaManninen/RaimoTuomela(eds.),EssaysonExplanationandUnderstanding.StudiesintheFoundationsofHumanitiesandSocialSciences,Dordrecht:Reidel1976,pp.371-413;andGeorgHenrikvonWright,“ProblemedesErklärensundVerstehensvonHandlungen”,in:Conceptus,19,1985,pp.3-19.16Cf.HansGeorgGadamer,WahrheitundMethode,Tubingen:J.C.B.Mohr(P.Sie-beck)5thed.,1986(1sted.,1960).17Cf.KarlOttoApel,“CausalExplanation,MotivationalExplanation,andHermeneuti-calUnderstanding”,in:GilbertRyle(ed.),ContemporaryAspectsofPhilosophy,Stockfield:OrielPress1976,pp.161-176;andKarlOttoApel,DieErklären-Verste-henKontroverseinTranszendental-PragmatischerSicht,Frankfurt:Suhrkamp1979.TranslatedintoEnglishbyGeorginaWarnke:UnderstandingandExplanation.ATran-scendental–PragmaticPerspective,Cambridge,MA:TheMITPress1984.18WenceslaoJ.Gonzalez,“ThePhilosophicalApproachtoScience,TechnologyandSo-ciety”,Op.cit.,pp.3-49.19“Iamamongthosewhohavefoundtheclaimsofthestrongprogramabsurd:anexam-pleofdeconstructiongonemad”,ThomasS.Kuhn,“TheTroublewiththeHistoricalPhilosophyofScience”,lectureattheUniversityofHarvardon19November1991.PaperreprintedinThomasS.Kuhn,TheRoadSinceStructure:PhilosophicalEssays,1970-1993,withanAutobiographicalInterview,editedbyJamesConantandJohnHaugeland,Chicago:TheUniversityofChicagoPress2000,p.110.20Cf.WolfSchäfer(ed.),FinalizationinScience.TheSocialOrientationofScientificProgress,Dordrecht:Reidel1983.\n230WenceslaoJ.GonzálezledbyBarryBarnes21(nowattheUniversityofExeter)andDavidBloor;22c)theEmpiricalProgrammeofRelativism(EPOR)endorsedbyacademicsoftheUniversityofBathsuchasHarryCollins;23andd)theethnomethodologyofaconstructivistorientation,connectedtotheworksofBrunoLatour24andSteveWoolgar,25whohaveapurelysociologicalvisionofscientificknowledge(objects,factsanddiscoveries).26Manyoftheseauthorsaresociologistsratherthanphilosophers,buttheyhavehadinfluenceonphilosophy,mainlyinthediscussionsonscience,technologyandsociety.Ontheonehand,theyhavestressedtheanalysisofthe“externalperspec-tive”onscience,emphasizingsocialvalues(e.g.,onaims,processesandresults);andontheotherhand,theyhavecalledattentiontosociologyandsocialanthro-pologyinordertounderstandlaboratoryexperimentationandscienceasahumanpracticewithinasocialmilieu.Themainstreamtendencyinthosecasesisasocialconstructivism,apositioncommonlycriticizedbymanyrealists.21Cf.BarryBarnes,InterestsandtheGrowthofKnowledge,London:RoutledgeandK.Paul1977;BarryBarnes,T.S.KuhnandSocialScience,London:Macmillan1982(N.York:ColumbiaUniversityPress1982);andBarryBarnes/DavidBloor/JohnHenry,ScientificKnowledge.ASociologicalAnalysis,Chicago:TheUniversityofChicagoPress1996.22Cf.DavidBloor,“WittgensteinandMannheimontheSociologyofMathematics”,in:StudiesinHistoryandPhilosophyofScience,4,1973,pp.173-191;DavidBloor,“Popper’sMystificationofObjectiveKnowledge”,in:ScienceStudies,4,1974,pp.65-76;DavidBloor,KnowledgeandSocialImagery,London:RoutledgeandK.Paul1976(2nded.,Chicago:TheUniversityofChicagoPress1991);andDavidBloor,Wittgenstein:ASocialTheoryofKnowledge,London:Macmillan1983.23Cf.HarryM.Collins,“AnEmpiricalRelativistProgrammeintheSociologyofSci-entificKnowledge”,in:KarinD.Knorr-Cetina/MichaelMulkay(eds.),ScienceOb-served:PerspectivesintheSocialStudyofScience,Sage,London,1983,pp.85-100;andHarryM.Collins/TrevorPinch,TheGolem:WhatEveryoneShouldKnowAboutScience,CambridgeUniversityPress,Cambridge,1993.24Cf.BrunoLatour,ScienceinAction,MiltonKeynes:OpenUniversityPress1987.BrunoLatour,LesMicrobes:guerreetpaix,suivideIrréductions,Paris:A.-M.Mé-taillé1984;revisedandexpandedEnglishversion,translatedbyA.SheridanandJ.Law:ThePasteurisationofFrance,Cambridge,MA:HarvardUniversityPress1988.BrunoLatour,Nousn’avonsjamaisétémodernes–Essaid’anthropologiesymétrique,Paris:LaDécouverte1991;revisedandaugmentededition,translatedintoEnglishbyCatherinePorter:WehaveNeverbeenModern,Brighton:Harvester1993.25Cf.StevenWoolgar,“CritiqueandCriticism:TwoReadingsofEthnomethodology”,in:SocialStudiesofScience,11,4,1981,pp.504-514;StevenWoolgar,Science:TheVeryIdea,London:Tavistock1988;StevenWoolgar(ed.),KnowledgeandReflexitiv-ity:NewFrontiersintheSociologyofKnowledge,London:Sage1988;andMichaelLynchandSteveWoolgar(eds.),RepresentationinScientificPractice,Cambridge,MA:TheMITPress1990.26Cf.WenceslaoJ.Gonzalez,“LasrevolucionescientíficasylaevolucióndeThomasS.Kuhn”,in:WenceslaoJ.Gonzalez(ed.),AnálisisdeThomasKuhn:Lasrevolucionescientíficas,Madrid:Trotta2004,pp.15-103;especially,pp.36-43.\nTrendsandProblemsinPhilosophyofSocialandCulturalSciences2313.Aparalleltendencyisscientificrealisminsocialsciences,whichiswithinthesecondlevelofanalysis(thecomparativeposition)andpresupposestheexist-enceofrealismonscienceingeneral(thefirstlevelofanalysis).Inthisregard,followingthelistofcharacteristicelementsofascience,wehavedefactoalargenumberofpossibilitiesforscientificrealism:semantical,logical,epistemological,methodological,ontological,axiological,andethical.Commonly,thediscussiononrealismpaysmoreattentiontosomeofthesepossibilities,mainlythoseversionsrelatedtolanguage,knowledge,method,andreality.Thesefactorsinvolveavisionoftheworldaswellasvalues(inthedif-ferentrealms:cognitive,ethical,social,economic,etc.)thatinfluencescientificpractice.Inaddition,everypossibilityofapproachtoscientificrealism(e.g.,se-mantical,epistemological,methodologicalorontological)canreceivedifferentinterpretationswithinarealistframework.Altogether–thelistoflevelsandthenumberofinterpretationsineach–leadstoaplethoraofversionsof“scientificrealism”.27Amongthecontemporarydiversityofscientificrealisms,therearesomethathaverootsinEuropeorhaveacharacteristicversiondevelopedbyaEuropean.JohnWorrall’s“structuralrealism”28andIlkkaNiiniluoto’s“scientificcriticalreal-ism”aretwoexamplesofgeneralconceptionsofscienceproposedbyEuropeansthatcanhaverepercussionsonthesocialsciences.Inthecaseofthestructuralrealism,theinitialfocusonthenaturalscienceshasbeenenlargeddefacto.Thus,therearesomenewviewsrelatedtosocialaffairs,eventhoughtherelationisratherindirect:thesocialconcernappearsthroughtheanalysisofmedicalprob-lems.29Meanwhile,theconceptionofthescientificcriticalrealismalreadyhasageneralcharacterizationofthesocialsciences.Inadditiontoepistemologicalandmethodologicalclaims,thispositionisconnectedwithanontologyofthethreeworlds.3027Cf.WenceslaoJ.Gonzalez,“NoveltyandContinuityinPhilosophyandMethodologyofScience”,Op.cit.,pp.12-13.28Worrall’sideashavebeenthestartingpointfornewreflectionsonthesocialsciences,cf.HaroldKincaid,“StructuralRealismandtheSocialSciences”,in:PhilosophyofScience,75,5,2008,pp.720-731.29Cf.JohnWorrall,“WhyRandomize?EvidenceandEthicsinClinicalTrials”,in:Wen-ceslaoJ.GonzalezandJesusAlcolea(eds.),ContemporaryPerspectivesinPhilosophyandMethodologyofScience,pp.65-82;andJohnWorrall,“WhyThere’sNoCausetoRandomize”,in:TheBritishJournalforthePhilosophyofScience,58,3,2007,pp.451-488.30Cf.IlkkaNiiniluoto,“Realism,Wordlmaking,andtheSocialSciences”,in:IlkkaNi-iniluoto,IsScienceProgressive?,Dordrecht:Reidel1984,pp.211-225.(Symposiumon“ScientificProgressandtheSocialSciences”,UniversityofTampere,April,1980);IlkkaNiiniluoto,“Finalization,AppliedScience,andSciencePolicy”,in:IlkkaNiini-luoto,IsScienceProgressive?,pp.226-243;IlkkaNiiniluoto,CriticalScientificReal-ism,Oxford:ClarendonPress1999;andIlkkaNiiniluoto,“World3:ACriticalDe-fence”,in:IanJarvie,KarlMilfordandDavidMiller(eds.),KarlPopper:ACentenary\n232WenceslaoJ.GonzálezBothapproachesonrealism–structuralandcritical–haverelevantEuropeanroots–HenriPoincaréandImreLakatos–,inthecaseofWorrall,andKarlPop-per’sthirdworldview,inNiiniluoto’sproposal.Theirviewsgobeyondtheorigi-nalsourcesinsofarastheyarenewareasofphilosophicalresearch,suchassocialconsequencesofclinicaltrials(includingethicalproblems)31andsustainablede-velopment.32Theirrealistconceptionscanbeusedtoanalyzeacharacteristicprob-lemofthesocialsciences,suchasthediscussiononobjectivityand,thereafter,thedebateontruth.Thephilosophicaldiscussiononrealisminthesocialsciencesdependsonthepossibilityofobjectivity(mainlyinlanguage,knowledge,method,andvalues)asacrucialstartingpointtofacetheproblemoftruth(historical,eco-nomic,sociological,psychological,anthropological,etc.)4.EXPLANATION,PREDICTION,ANDCONFIRMATION:REALMANDLIMITSSubsequenttothereflectiononthenaturalist,social,andrealistviews,theissuesfortheresearchinthesecondyeararethoseinthissection.Inthisregard,focusingonbasicscience,therearenowadaystwocentraltopicsinthephilosophyoftheculturalandsocialsciencesrelatedtoexplanation,prediction,andconfirmation.Ontheonehand,thereisamethodologicalcontroversyonexplanationandde-scriptionintheculturalsciences,whichaffectsaboveallsocialanthropologyandhistoricalsciences;and,ontheotherhand,theproblemsoftherealmandlimitsofexplanationandpredictioninthesocialscienceshaveaclearrelevance,especiallyineconomics,butalsowithrepercussionsonothersocialsciences,suchassociol-ogyorpoliticalscience.“Explanation”versus“description”isamethodologicaldisputethatcanbeseencommonlyintheresearchofculturalsciences,suchassocialanthropologyandhistoricalsciences(e.g.,inuniversitydepartments).Defacto,manyauthorsassumethat“description”isgoodenoughinordertohave“science”oreven“ma-turescience”inculturalsubjectmatters.Atthesametime,thosesocialscientiststhathavebeentrainedingeneralmethodologyofsciencerejectthat“description”couldbegoodenoughinordertohave“science”,becausetheyconsiderthatsci-entificundertakingsrequire“explanation”,andthatscienceshouldthereforegivegroundedrepliesto“why?”questions.33Assessment,vol.II,Aldershot:Ashgate2006,pp.59-69.31Cf.JohnWorrall,“WhyRandomize?EvidenceandEthicsinClinicalTrials”,Op.cit.,pp.76-80.32Cf.IlkkaNiiniluoto,“Nature,Man,andTechnology–RemarksonSustainableDevel-opment”,in:LassiHeininen(ed.),TheChangingCircumpolarNorth:OpportunitiesforAcademicDevelopment,ArcticCentrePublications6,Rovaniemi,1994,pp.73-87.33Cf.WenceslaoJ.Gonzalez,“Caracterizacióndela‘explicacióncientífica’ytiposdeexplicacionescientíficas”,in:WenceslaoJ.Gonzalez(ed.),Diversidaddelaexpli-\nTrendsandProblemsinPhilosophyofSocialandCulturalSciences233Therealmandlimitsofexplanationandpredictioninthesocialsciencesraisedanothermethodologicalquestions,someofthemconnectedwiththepre-viousdiscussion.Ontheonehand,thereisatwofoldproblem,concerningthecharacteristicsof“scientificexplanations”inthesocialsciences,andwhatkindofscientificexplanationsistobefoundinthissphere(causal,noncausal,etc.).Ontheotherhand,socialscientists–includingNobelPrizelaureates–deeplydisagreeaboutthepossibilityofelaboratingscientificpredictionsintherealmofhumanaffairs.34Inaddition,fromthepointofviewofconfirmation,thereisadebateonwhetherexplanationorpredictionhasagreaterepistemologicalandmethodologicalimpact.Regardingscientificexplanationinthesocialsciences,evenwhenacausalapproachisbeingadopted,thenotionof“cause”isnotalwaysunderstoodinthesameway.IneconomicstherearecleardifferencesbetweenCliveG.Granger’sviewsandotherperspectives.35Inpsychologytherearealsodifferenceson“causa-tion”inconnectionwiththelinksbetweenreasons,actionsandcauses.Inarchae-ology,wecanfindattemptstocombinetwodifferentkindsofcauses(connectedwithmeansandends).Predictionisakeyissueintheepistemologyandmethodologyofsocialsci-ences.Itcanbeusedasatestofbasicscience(e.g.,ofeconomictheory)andasatoolforsocialpolicyinappliedscience(e.g.,ofappliedeconomics).Therealmandlimitsofpredictionsarecrucialtopicsforappliedsciences,36suchaseconom-cacióncientífica,Barcelona:Ariel2002,pp.13-49.Inthisregard,seeAmparoGomez,FilosofíayMetodologíadelasCienciasSociales,Madrid:AlianzaEditorial2003,andRobertC.Bishop,ThePhilosophyofSocialSciences,London:Continuum2007,chapter15,especially,pp.326-331.34AparamountexampleisSirJohnHicks,NobelPrizeineconomicsin1972.Cf.JohnHicks,“ADisciplinenotaScience”,in:JohnHicks,ClassicsandModerns.CollectedEssaysonEconomicTheory,v.III,Cambridge,MA:HarvardUniversityPress1983,pp.364-375;andJohnHicks,“IsEconomicsaScience?”,in:MauroBaranzini/Ro-bertoScazzieri(eds.),FoundationsofEconomics.StructuresofInquiryandEconomicTheory,Oxford:B.Blackwell1986,pp.91-101.Onthisissue,cf.WenceslaoJ.Gonzalez,“PredictionasScientificTestofEconom-ics”,in:WenceslaoJ.GonzalezandJesusAlcolea(eds.),ContemporaryPerspectivesinPhilosophyandMethodologyofScience,pp.83-112;especially,pp.88-90and98-100.SeealsoDieterHelm,“PredictionsandCauses:AComparisonofFriedmanandHicksonMethod”,OxfordEconomicPapers,newseries,36,1984,pp.118-134.35Cf.CliveW.J.Granger,“TestingforCausality,aPersonalViewpoint”,in:JournalofEconomicDynamicsandControl,2,1980,pp.329-352;CliveW.J.Granger,“WherearetheControversiesinEconometricMethodology?”,in:CliveW.J.Granger(ed.),ModellingEconomicsSeries:ReadingsinEconometricMethodology,Oxford:Claren-donPress1990,pp.1-23;andCliveW.J.Granger,“TimeSeriesAnalysis,Cointegra-tion,andApplications”,in:ToreFrängsmyr(ed.),FromLesPrixNobel.TheNobelPrizes2003,NobelFoundation,Stockholm:NobelFoundation2004,pp.360-366.36Cf.IlkkaNiiniluoto,“TheAimandStructureofAppliedResearch”,in:Erkenntnis,38,1993,pp.1-21.\n234WenceslaoJ.Gonzálezics,sociologyorsocialpsychology.Theimportanceofpredictionisevengreaterinsofarasitisconnectedtoprescription(i.e.,whatoughttobedone),37asisthecaseineconomics(astheongoinginternationalfinancialcrisesthatbreakoutin2008hasemphasized).Buttheneedforpredictioninsocialsciencesisusuallycounterbalancedbyitsunreliablenature,asNobelPrizesineconomics(HerbertSimon,AmartyaSen)haverepeatedlypointedout,38eventhoughsomesocialevents(e.g.,publictrans-port)canhaveveryreliablepredictions.Inaddition,thereisalwaysthepossibilitythatasocialpredictioncanaffectthesocialphenomenonitself,aproblemthataffectssociologyaswellaspoliticalscience.Alltheseissuesrelatedtoexplanation,prediction,andconfirmation,theirrealmandlimits,shouldbediscussedfromaEuropeanperspective.Theapproachshouldtakeintoaccountan“integrative”viewpoint.Thisinvolvesthecontribu-tionsmadebyinstitutionslocatedinEuropeancountriesandtheworkofphi-losophersofanyoftheEuropeannationalities.Inaddition,thethreelevelsofthephilosophical-methodologicalanalysis–general,comparative,andspecific–areatstakehere.5.THEDEBATEONMATHEMATICALMODELINGINTHESOCIALSCIENCESANDCONSEQUENCESFOREXPERIMENTATIONAtraditionalcontroversialareaofdiscussionfortheculturalsciencesistheroleofprobabilityandstatisticsinthestudyofhumaneventsintheculturalmilieu,becausethepredominantviewisinfavorofqualitativeaspects(e.g.,trends,sym-bols,etc.)andagainsttheuseofexperimentsonculturalfactors.Meanwhile,inthesocialsciencesthereiscommonlyadifferentattitudeonmathematicalmodelingandontheroleofexperiments.Thethirdyearresearchofthisprogramwilldealpreciselywiththedebateonmathematicalmodelinginthesocialsciencesanditscontributiontoexperimentsonsocialissues.Thismethodologicaldebateonmathematicalmodelingistobeanalyzedtak-ingintoaccountatleastthreekindsofproblems:a)theverystatusofmathematicalmodeling(i.e.,mathematicsasa“language”usedtoestablishknowledge,whichincludestheproceduresofproof,ormathematicsasa“heuristictool”connected37SeeHerbertA.Simon,“PredictionandPrescriptioninSystemsModeling”,in:Opera-tionsResearch,38,1990,pp.7-14;reprintedinHerbertA.Simon,ModelsofBoundedRationality.Vol.3:EmpiricallyGroundedEconomicReason,Cambridge,MA:TheMITPress1997,pp.115-128.38Onthisissue,cf.WenceslaoJ.Gonzalez,“PredictionandPrescriptioninEconomics:APhilosophicalandMethodologicalApproach”,in:Theoria,13,32,1998,pp.321-345.SeealsoHerbertA.Simon,“ForecastingtheFutureorShapingit?”,in:IndustrialandCorporateChange,11,3,2002,pp.601-605.\nTrendsandProblemsinPhilosophyofSocialandCulturalSciences235todiscoverythroughabstractrepresentation);39b)theissueofsymmetryorasym-metrybetweenmathematicalmodelinginthenaturalsciencesascomparedtothesocialsciences;andc)themethodologicalcontroversybetweenqualitativeandquantitativemodelswithinthesocialsciences.Usually,experimentalapproachesinthesocialsciencesneedtodealwiththesekindsofproblemsonmodeling(status;symmetryorasymmetry;andqualitativeversusquantitative).Certainlyitisthecaseinexperimentaleconomics,whereNobelPrizes(suchasReinhardSeltenorVernonSmith)areexpertsingamethe-ory.40Itisalsothecaseinexperimentalpsychology,wherestatisticaltechniques(ClassicalorBayesian)haveacrucialrole.Inaddition,thereareinterdisciplinarystudiesbaseduponthosecriteria(e.g.,inthecombinationofpsychologyandeco-nomics).41Eventhoughmathematicalmodelingisusedinmostsocialsciences(eco-nomics,psychology,history,sociology,politicalscience),thereisalsoanongo-ingcontroversyaboutitsrealmandlimits.Thedifferenceintheresponsesontheroleofmathematicalmodelingcanbeseeninmanyways,suchasthediversityofmethodologicalapproachesinhistory(e.g.,American“NewHistory”versus“narrativehistory”)orinpsychology(e.g.,inpsychobiologyandpsychologyofpersonality).39HerbertSimonusedtodistinguishtwousesofmathematics.Ontheonehand,mathe-maticsasalanguageofproof,whererigorisessential,aguaranteethatconclusionsarecorrect(TjallingKoopmans,GerardDebreu,KennethArrow,etc.).And,ontheotherhand,mathematicsasalanguageofdiscovery,asatooltoarriveatnewideas,wheresolutionsreachedwithitshelpshouldbecheckedforcorrectness.ThesecondoneisSimon’spreference,andheconsidersthatitisphysicists’mathematicsandengineers’mathematicsinsteadofmathematicians’mathematics.Cf.HerbertA.Simon,ModelsofmyLife,N.York,NY:BasicBooks1991,pp.106-107.40Cf.WenceslaoJ.Gonzalez,“TheRoleofExperimentsintheSocialSciences:TheCaseofEconomics”,inTheoKuipers(ed.),GeneralPhilosophyofScience:FocalIssues,Amsterdam:Elsevier2007,pp.275-301.RobertAummanisanotherNobelPrizeineconomicsexpertingametheoryandinterestedinepistemologicalandmeth-odologicalmatters,cf.RobertJ.Aumann,“RationalityandBoundedRationality”,in:GamesandEconomicBehavior,21,1-2,1997,pp.2-14.41Cf.DanielKahneman/JackKnetsch/RichardThaler,“Experimenta1TestsoftheEn-dowmentEffectandtheCoaseTheorem”,in:JournalofPoliticalEconomy,98,1990,pp.1325-1348;andDanielKahneman,“MapsofBoundedRationality:PsychologyforBehavioralEconomics”,in:AmericanEconomicReview,93,5,2003,pp.1449-1475.\n236WenceslaoJ.González6.THESCIENCESTHATPHILOSOPHYHASHITHERTOIGNORED:THESCIENCESOFDESIGNAmongthefieldsthathavebeenhithertoignoredbyphilosophyarethesciencesoftheartificial,understoodas“sciencesofdesign”.42Ithappensthatanimportantpartofwhatisusuallycalled“socialsciences”belongs,defacto,totherealmofthe“sciencesoftheartificial”(e.g.,libraryscience,communication,oreconom-ics).Inaddition,thereareotherscienceswithintheterritoryofthe“sciencesofdesign”,suchaspharmacology,thatarealsosciencesoftheartificial–inthesenseproposedbyHerbertSimon–andhaveaclearlinktosocialsciences.Commonly,thesescienceshavebeenhithertoignoredbyphilosophy(atleast,inthemain-streamphilosophyofscience),yettheyhaverelevantlinkswiththeculturalandsocialsciencesthatshouldbeanalyzed.Thesesciencesoftheartificial,whichareusuallydesignsciences,comefromwhatIlkkaNiiniluotocalls“scientification”,i.e.,aprocesstochangeaprofes-sionalpracticeintoascientificdiscipline.43Thus,theycomefromasocialpracticethatrequiresscientificsupportintheformofadesign.Habitually,theyare“ap-pliedsciences”,andthisfeatureisrelevanttostudythem,becausetraditionallyphilosophyofsciencehasfocusedmoreonbasicsciencethanonappliedscience.Followingthedifferencesandsimilaritiesbetweenthesciencesoftheartificialandtheculturalandsocialsciences,itispossible–inthefourthyearofresearchofthisprogram–toshedlightonsomesciencesthatwerenotcommonlystudiedinthisarea.Thesearelibraryscience,communication,pharmacology,economics,etc.,whichcanhaveepistemologicalandmethodologicalproblemsinsofarastheyareappliedsciencesofdesign.Theseproblemsarenotpurelycognitive,becausetheyinvolveothervalues(social,cultural,economic,ecological,etc.).7.NEWDIRECTIONSINTHEPHILOSOPHYOFSCIENCEAEuropeanperspectiveshouldbeopentothefuture.TheresearchofTeamCin-cludesanexplicitreflectionofnewdirectionsinthephilosophyofscience.Primafacie,thismaybeunderstoodindifferentways,accordingtowhichitmightbeconsideredas“novelty”,whichincludesatleastthreepossibilities.(i)Thereisanoveltyinontologicalterms,whensomethinghashappenaftertheoriginalmo-ment;inthiscase,newconceptionsaftertheapprovaloftheprogrambytheEu-ropeanScienceFoundation.(ii)Thereisanoveltyinepistemologicalterms,whensomethingisgoingondefactobutwearenotawareoftheinnovativecharacterof42Cf.HerbertA.Simon,TheSciencesoftheArtificial,3rded.,Cambridge,MA:TheMITPress1996(1sted.,1969;2nded.,1981).43Cf.IlkkaNiiniluoto,“TheAimandStructureofAppliedResearch”,Loc.cit.,pp.8-11.\nTrendsandProblemsinPhilosophyofSocialandCulturalSciences237itscontents.(iii)Thereisalsoaheuristicnovelty,whentherearenewpatternsororiginalpathstoaddressoldproblems.Lookingatfiveyearsfromnow,itseemsratherobviousthattherewillbesomenewdirectionsinthephilosophyofscienceinanyofthesethreepossibleop-tionsof“novelty”.Inadditiontothestudyofnewbranchesofthesocialsciences(suchasrisingsubdisciplines),theseemergingphilosophiesofsciencecanbeonnewapproachesinthesocialconcernonscience.Theymaybeconnectedtothesocialconstructivismandtherealismontheculturalandsocialsciences.Thus,inthefinalyearoftheprogram,TeamCwillinvestigatethenovelviewsontheinflu-enceofsocialconstructivismandrealismonthesocialsciences.Fromamethodologicalviewpoint,theoppositionbetweensocialconstructiv-ismandrealismaffectsthesocialsciencesasawholeaswellaseachoneofthem(anthropology,archaeology,economics,etc.).Itisadebatethathasnotonlytheo-reticalconsequences,butalsopracticaleffects.Oneexampleishowtounderstandmedicine,duetonewapproachestomedicineasaculturalandsocialscience:thealternativemedicine,whichisarelevantaspectofthesocialconcernonmedicineasascience.44ThisissueshouldbestudiedwithaneyetotheEuropeanperspec-tiveunderstoodinthe“integrative”termspointedout.Ontheonehand,socialconstructivismisawidespreadtendencyinthesocialsciences,asIanHackinghasemphasized.45Understoodasanepistemologicalandmethodologicalapproach–andoftenalsoasanontologicalconception–,socialconstructivismisadoptednowadaysbyvirtuallyeverysocialdiscipline.Socialconstructivismcouldbeopentomethodologicalindividualism(e.g.,incognitiveapproachesinpsychology)ormightadoptaclearholisticmethodology(e.g.,insocialpsychologyorinsocialpolicy).Ontheotherhand,realismoffersavarietyofalternativestosocialconstructivism.Therevivalof“scientificrealism”inre-centyearsemphasizesthepossibilityofobjectivityregardingsocialphenomena.8.FINALREMARKSMypreviouslinessuggestthatanyattempttoanalyzethetrendsandproblemsinphilosophyofsocialandculturalsciencesfromaEuropeanperspectiveshouldtakeintoaccountseveralaspects.First,thelevelofanalysisinthephilosophical-methodologicalapproachused(general,comparative,orspecific).Second,whatmightbereasonablyconsideredfromaphilosophicalpointofviewas“aEuro-peanperspective”,whichcanbeunderstoodin“integrative”terms.46Third,the44Cf.DonaldA.Gillies,“ElproblemadelademarcaciónylaMedicinaalternativa”,in:WenceslaoJ.Gonzalez(ed),KarlPopper:Revisióndesulegado,Madrid:UniónEditorial2004,pp.197-219.45Cf.IanHacking,TheSocialConstructionofWhat?,Cambridge,MA:HarvardUni-versityPress1999.46Ifwefocusourattentiononjournalsthatdealwithtopicsrelatedtophilosophyof\n238WenceslaoJ.Gonzálezcurrentstageofthephilosophicalconsiderationsafterthedeclineofthe“histori-calturn”.Inthisregard,thereisnowanactivecompetitionamongthenaturalistconceptions,thefollowersofthe“socialturn”,andthesupportersofthescientificrealism.Thesegeneralphilosophical-methodologicalapproacheshavedirectre-percussionsontheviewsoftheculturalandsocialsciences.Theyappearastopicsofdiscussionforthefirstyear.Thereafter,someoth-errelevanttopics,whichhavebeenpointedout,arescheduledforthefollowingyears.Myproposalshavebeenpredominantlywithinlevelstwoandthreeofthephilosophicalanalysis.AndtheyhavebeenfocusedonwhatshouldbethecontextofdiscussioninordertomovealongthelinesofthesuggestionsofsubmittedtotheEuropeanScienceFoundation.Inthisregard,Ihavesuggestedthat“Euro-pean”canbeunderstoodfroman“integrative”viewpoint,whereawiderangeofconceptionscanbeincluded(insofarasthereisasearchforacommonground,bothinhistoricaltermsandinthematicones).Moreover,thecontributionsmadebyinstitutionslocatedinEuropeancountriesandtheworkofphilosophersofanyoftheEuropeannationalitiescanbeusedtobuildupan“identity”forthephiloso-phyoftheculturalandsocialsciencesinEurope.47BIBLIOGRAPHYKarlOttoApel,“CausalExplanation,MotivationalExplanation,andHermeneuti-calUnderstanding”,in:GilbertRyle(ed.),ContemporaryAspectsofPhiloso-phy,Stockfield:OrielPress1976,pp.161-176.KarlOttoApel,DieErklären-VerstehenKontroverseintranszendental-pragma-tischerSicht,Frankfurt:Suhrkamp1979.TranslatedintoEnglishbyGeorginaWarnke:UnderstandingandExplanation.ATranscendental-PragmaticPer-spective,Cambridge,MA:TheMITPress1984.RobertJ.Aumann,“RationalityandBoundedRationality”,in:GamesandEco-nomicBehavior,21,1-2,1997,pp.2-14.BarryBarnes,InterestsandtheGrowthofKnowledge,London:RoutledgeandK.Paul1977.economics,thereisarelevantlistofjournalsdirectedbyEuropeaneditorsorrunbyEuropeanpublishinghouses.ThelistincludesEconomicsandPhilosophy,JournalofEconomicMethodology,andJournalfortheHistoryofEconomicThought.Inaddi-tion,thereareotherpublicationswithaclearpresenceofEuropeansintheeditorialboard,suchasPolitics,PhilosophyandEconomics.Inan“integrativesense”,thesejournalsmightbeconsideredas“Europeans”.But,ifwefollowrestrictedcriteria,itishardtosaythattheyaretoutcourt“Europeans”.47ThisissueisnotdisconnectedfromthephilosophicaldiscussionontheEuropeanunity.Inthisregard,therearemanycontributions.Amongthemarethoseinitiatedinthefirsthalfofthetwentiethcentury,cf.HaroldC.Raley,JoséOrtegayGasset:PhilosopheroftheEuropeanUnity,Alabama:UniversityofAlabamaPress1971.\nTrendsandProblemsinPhilosophyofSocialandCulturalSciences239BarryBarnes,T.S.KuhnandSocialScience,London:Macmillan1982(N.York:ColumbiaUniversityPress1982).BarryBarnes/DavidBloor/JohnHenry,ScientificKnowledge.ASociologicalAnalysis,Chicago:TheUniversityofChicagoPress1996.RobertC.Bishop,ThePhilosophyofSocialSciences,London:Continuum2007.DavidBloor,“WittgensteinandMannheimontheSociologyofMathematics”,in:StudiesinHistoryandPhilosophyofScience,4,1973,pp.173-191.DavidBloor,“Popper’sMystificationofObjectiveKnowledge”,in:ScienceStud-ies,4,1974,pp.65-76.DavidBloor,KnowledgeandSocialImagery,London:RoutledgeandK.Paul1976(2nded.,Chicago:TheUniversityofChicagoPress1991).DavidBloor,Wittgenstein:ASocialTheoryofKnowledge,London:Macmillan1983.HarryM.Collins,“AnEmpiricalRelativistProgrammeintheSociologyofScien-tificKnowledge”,in:KarinD.Knorr-Cetina/MichaelMulkay(eds.),ScienceObserved:PerspectivesintheSocialStudyofScience,London:Sage1983,pp.85-100.HarryM.Collins/TrevorPinch,TheGolem:WhatEveryoneShouldKnowAboutScience,Cambridge:CambridgeUniversityPress1993.HansGeorgGadamer,WahrheitundMethode,Tubingen:J.C.B.Mohr(P.Sie-beck)5thed.1986(1sted.1960).DonaldA.Gillies,“ElproblemadelademarcaciónylaMedicinaalternativa”,in:WenceslaoJ.Gonzalez(ed.),KarlPopper:Revisióndesulegado,Madrid:UniónEditorial2004,pp.197-219.AmparoGomez,FilosofíayMetodologíadelasCienciasSociales,Madrid:AlianzaEditorial2003.WenceslaoJ.Gonzalez,“PredictionandPrescriptioninEconomics:APhilosophi-calandMethodologicalApproach”,in:Theoria,13,32,1998,pp.321-345.WenceslaoJ.Gonzalez,“Caracterizacióndela‘explicacióncientífica’ytiposdeexplicacionescientíficas”,in:WenceslaoJ.Gonzalez(ed.),Diversidaddelaexplicacióncientífica,Barcelona:Ariel2002,pp.13-49.WenceslaoJ.Gonzalez,“FromErklären-VerstehentoPrediction-Understanding:TheMethodologicalFrameworkinEconomics”,in:MattiSintonen,PetriYlikoskiandKarlMiller(eds.),RealisminAction:EssaysinthePhilosophyofSocialSciences,Dordrecht:Kluwer2003,pp.33-50.WenceslaoJ.Gonzalez,“LasrevolucionescientíficasylaevolucióndeThomasS.Kuhn”,in:WenceslaoJ.Gonzalez(ed.),AnálisisdeThomasKuhn:Lasrevolucionescientíficas,Madrid:Trotta2004,pp.15-103.WenceslaoJ.Gonzalez,“ThePhilosophicalApproachtoScience,TechnologyandSociety”,in:WenceslaoJ.Gonzalez(ed.),Science,TechnologyandSociety:APhilosophicalPerspective,ACoruña:Netbiblo2005,pp.3-49.WenceslaoJ.Gonzalez,“NoveltyandContinuityinPhilosophyandMethodologyofScience”,in:WenceslaoJ.GonzalezandJesusAlcolea(eds.),Contem-\n240WenceslaoJ.GonzálezporaryPerspectivesinPhilosophyandMethodologyofScience,ACoruña:Netbiblo2006,pp.1-28.WenceslaoJ.Gonzalez,“PredictionasScientificTestofEconomics”,in:WenceslaoJ.GonzalezandJesusAlcolea(eds.),ContemporaryPerspectivesinPhiloso-phyandMethodologyofScience,ACoruña:Netbiblo2006,pp.83-112.WenceslaoJ.Gonzalez,“TheRoleofExperimentsintheSocialSciences:TheCaseofEconomics”,in:TheoKuipers(ed.),GeneralPhilosophyofScience:FocalIssues,Amsterdam:Elsevier2007,pp.275-301.CliveW.J.Granger,“TestingforCausality,aPersonalViewpoint”,in:JournalofEconomicDynamicsandControl,2,1980,pp.329-352.CliveW.J.Granger,“WherearetheControversiesinEconometricMethodolo-gy?”,in:CliveW.J.Granger(ed.),ModellingEconomicsSeries:ReadingsinEconometricMethodology,Oxford:ClarendonPress1990,pp.1-23.CliveW.J.Granger,“TimeSeriesAnalysis,Cointegration,andApplications”,in:ToreFrängsmyr(ed.),FromLesPrixNobel.TheNobelPrizes2003,NobelFoundation,Stockholm:NobelFoundation2004,pp.360-366.IanHacking,TheSocialConstructionofWhat?,Cambridge,MA:HarvardUni-versityPress1999.DieterHelm,“PredictionsandCauses:AComparisonofFriedmanandHicksonMethod”,OxfordEconomicPapers,newseries,36,1984,pp.118-134.JohnHicks,“ADisciplinenotaScience”,in:JohnHicks,ClassicsandModerns.CollectedEssaysonEconomicTheory,v.III,Cambridge,MA:HarvardUni-versityPress1983,pp.364-375.JohnHicks,“IsEconomicsaScience?”,in:MauroBaranzini/RobertoScazzieri(eds.),FoundationsofEconomics.StructuresofInquiryandEconomicTheo-ry,Oxford:B.Blackwell1986,pp.91-101.DanielKahneman/JackKnetsch/RichardThaler,“Experimenta1TestsoftheEn-dowmentEffectandtheCoaseTheorem”,in:JournalofPoliticalEconomy,98,1990,pp.1325-1348.DanielKahneman,“MapsofBoundedRationality:PsychologyforBehavioralEconomics”,in:AmericanEconomicReview,93,5,2003,pp.1449-1475.HaroldKincaid,“SocialSciences”,in:PeterMachamerandMichaelSilberstein(eds.),TheBlackwellGuidetothePhilosophyofScience,Oxford:Blackwell2002,pp.290-311.HaroldKincaid,“StructuralRealismandtheSocialSciences”,in:PhilosophyofScience,75,5,2008,pp.720-731.ThomasS.Kuhn,“TheTroublewiththeHistoricalPhilosophyofScience”,lectureattheUniversityofHarvardon19November1991.PaperreprintedinThomasS.Kuhn,TheRoadSinceStructure:PhilosophicalEssays,1970–1993,withanAutobiographicalInterview,editedbyJamesConantandJohnHaugeland,Chicago:TheUniversityofChicagoPress2000,pp.105-120.BrunoLatour,ScienceinAction,MiltonKeynes:OpenUniversityPress1987.\nTrendsandProblemsinPhilosophyofSocialandCulturalSciences241BrunoLatour,LesMicrobes:guerreetpaix,suivideIrréductions,Paris:A.-M.Métaillé1984.RevisedandexpandedEnglishversion,translatedbyA.SheridanandJ.Law:ThePasteurisationofFrance,Cambridge,MA:Har-vardUniversityPress1988.BrunoLatour,Nousn’avonsjamaisétémodernes–Essaid’anthropologiesymé-trique,Paris:LaDécouverte1991.Revisedandaugmentededition,translatedintoEnglishbyCatherinePorter:WehaveNeverbeenModern,Brighton:Har-vester1993.MichaelLynchandSteveWoolgar(eds.),RepresentationinScientificPractice,Cambridge,MA:TheMITPress1990.JamesW.McAllister,“Editorial:ContoursofaEuropeanPhilosophyofScience”,in:InternationalStudiesinthePhilosophyofScience,22,1,2008,pp.1-3.IlkkaNiiniluoto,“TheAimandStructureofAppliedResearch”,in:Erkenntnis,38,1993,pp.1-21.IlkkaNiiniluoto,“Realism,Wordlmaking,andtheSocialSciences,”in:IlkkaNiiniluoto,IsScienceProgressive?,Dordrecht:Reidel1984,pp.211-225.(Symposiumon“ScientificProgressandtheSocialSciences”,UniversityofTampere,April,1980.)IlkkaNiiniluoto,“Finalization,AppliedScience,andSciencePolicy,”in:IlkkaNiiniluoto,IsScienceProgressive?,Dordrecht:Reidel1984,pp.226-243.IlkkaNiiniluoto,“Nature,Man,andTechnology–RemarksonSustainableDevel-opment,”in:LassiHeininen(ed.),TheChangingCircumpolarNorth:Oppor-tunitiesforAcademicDevelopment,ArcticCentrePublications6,Rovaniemi,1994,pp.73-87.IlkkaNiiniluoto,CriticalScientificRealism,Oxford:ClarendonPress1999.IlkkaNiiniluoto,“World3:ACriticalDefence”,in:IanJarvie,KarlMilfordandDavidMiller(eds.),KarlPopper:ACentenaryAssessment,vol.II,Aldershot:Ashgate2006,pp.59-69.HaroldC.Raley,JoséOrtegayGasset:PhilosopheroftheEuropeanUnity,Ala-bama:UniversityofAlabamaPress1971.WolfSchäfer(ed.),FinalizationinScience.TheSocialOrientationofScientificProgress,Dordrecht:Reidel1983.HerbertA.Simon,“PredictionandPrescriptioninSystemsModeling,”in:Opera-tionsResearch,38,1990,pp.7-14.ReprintedinHerbertA.Simon,ModelsofBoundedRationality.Vol.3:EmpiricallyGroundedEconomicReason,Cam-bridge,MA:TheMITPress1997,pp.115-128.HerbertA.Simon,ModelsofmyLife,N.York,NY:BasicBooks1991.HerbertA.Simon,TheSciencesoftheArtificial,3rded.,Cambridge,MA:TheMITPress1996(1sted.,1969;2nded.,1981).HerbertA.Simon,“ForecastingtheFutureorShapingit?”,in:IndustrialandCor-porateChange,11,3,2002,pp.601-605.\n242WenceslaoJ.GonzálezSteeringCommittee,ThePhilosophyofScienceinaEuropeanperspectivePro-posalofan“àlacarteProgramme”tobesubmittedtotheEuropeanScienceFoundation,24February2006.StephenP.Turner/PaulA.Roth(eds.),TheBlackwellGuidetothePhilosophyoftheSocialSciences,Oxford:Blackwell2003.PeterWinch,TheIdeaofaSocialScience,London:RoutledgeandK.Paul1958(2ndedition,1990).GeorgHenrikvonWright,ExplanationandUnderstanding,Ithaca:CornellUni-versityPress1971.GeorgHenrikvonWright,“Replies,”in:JuhaManninen/RaimoTuomela(eds.),EssaysonExplanationandUnderstanding.StudiesintheFoundationsofHu-manitiesandSocialSciences,Dordrecht:Reidel1976,pp.371-413.GeorgHenrikvonWright,“ProblemedesErklärenundVerstehensvonHandlun-gen,”in:Conceptus,19,1985,pp.3-19.StevenWoolgar,“CritiqueandCriticism:TwoReadingsofEthnomethodology”,in:SocialStudiesofScience,11,4,1981,pp.504-514.StevenWoolgar,Science:TheVeryIdea,London:Tavistock1988.StevenWoolgar(ed.),KnowledgeandReflexitivity:NewFrontiersintheSociol-ogyofKnowledge,London:Sage1988.JohnWorrall,“WhyRandomize?EvidenceandEthicsinClinicalTrials”,in:WenceslaoJ.GonzalezandJesusAlcolea(eds.),ContemporaryPerspectivesinPhilosophyandMethodologyofScience,ACoruña:Netbiblo2006,pp.65-82.JohnWorrall,“WhyThere’sNoCausetoRandomize”,in:TheBritishJournalforthePhilosophyofScience,58,3,2007,pp.451-488.UniversityofACoruñaFacultyofHumanitiesDr.VazquezCabreraStreet,w/n15403FerrolSpainwencglez@udc.es\nARTOSIITONENSTATEOFTHEARTACOMMENTARYONWENCESLAOJ.GONZALEZ’CONTRIBUTION,“TRENDSANDPROBLEMSINPHILOSOPHYOFSOCIALANDCULTURALSCIENCES:AEUROPEANPERSPECTIVE”INTRODUCTIONScientificresearchproceedsfromproblemstotheirsupposedsolutions,whichinturnraisefurtherproblems.Atypicalresearchprocessgoesthroughthestagesofinventionandtesting.Itsresultsarepresentedintheories;thesejustifytheresultsandmakethemunderstandable.Theoriesoftenhaveapplications,forinstanceintechnology,commerce,socialorganizationsandinteaching.Researchisaculturalactivitypursuedininternationalco-operationamongindividualsandgroups.Itbothfollowstraditionsandrenewsthem,andtakesitsdirectionaccordingtobothitsinternallogicandalsoexternalfactors.Theassumeddirectionatacertaintimemakesupthetrendthatischaracteristicforscienceingeneral,orforagivendisci-pline.Trendsaretypicalwaysofapproachingproblemsanddesigningtheories.Culturalandsocialsciencesinparticulararereflective,inthattheyarepartoftheprocessestheyinvestigate,andinthatthesesciences’cultivationcontributestothedevelopmentofcultureandsociety.Eachofthesegroupsofscienceshasinitsbackgroundcertainbranchesofphilosophy–i.e.,philosophyofculturalsciencesandphilosophyofculture;philosophyofsocialsciencesandsocialphilosophy.Moreover,theideaofhumanismcanbeseenbehindtheculturalsciences,andbe-hindphilosophyofculture.Onemaywonderwhetherthedoctrineofsocialismhasasimilarroleinthephilosophyofsocialsciencesandinsocialphilosophy.Thereisanimportantdifference,however.Humanisminvolvesthecultivationoflitera-tureaswellasdedicationtolearningandteaching,whereastheword‘socialism’hasastronglypoliticalconnotation.Culturalsciencesarecalled“humanisticdis-ciplines”,butsocialsciencescannotbecharacterizedas“socialisticdisciplines”.Sometimestheculturalandsocialsciencesareclassifiedtogetherashumanormoralsciences,sincetheresearchobjectofbothareasconcernshumanactionanditsconsequences.Inthiscase,theterm‘humansciences’differsfromtheterms‘humanisticsciences’and‘humanities’,thelatterbeingtantamountto‘culturalsciences’.Moreover,thehumanitiesarealsocalled‘arts’or‘artsubjects’(cf.thetitles‘BachelorofArts’,‘MasterofArts’).Theboundariesbetweenculturalsci-F.Stadler(ed.),ThePresentSituationinthePhilosophyofScience,ThePhilosophyofScienceinaEuropeanPerspective,DOI10.1007/978-90-481-9115-4_19,©SpringerScience+BusinessMediaB.V.2010\n244ArtoSiitonenencesandsocialsciencesarenotsharp.Forinstance,insomecaseshistoryiscon-sideredasocialratherthanaculturalscience.1Cultureandlanguagedifferencesintroduceafurthercomplication,becausetheGermantraditionoftheGeisteswissenschaftendiffersfromtheAnglo-Saxonconceptionofsciences.Inthelatter,theterms‘science’and‘naturalscience’areimplicitlyunderstoodassynonyms.Thus,theGermanword‘Wissenschaft’hasamuchbroaderscopethantheEnglishword‘science’.Inrespecttothisawkwardsituation,onemaysuggestthatthereisadistinctionbetweenhumanitiesandcul-turalsciences.Thedistinctivemarkofthelatteristhattheyfollowacceptedmeth-odsofscience–constructhypothesesandtestthem,usingdeduction,inductionandabduction.Besidesidentifyingvariousbranchesofscience,wemayalsoaskwhereandhowthesehavebeencultivated.Thesearequestionspertainingtotraditions,stylesandgeography.Termssuchas‘Mesopotamian’,‘Egyptian’,‘Greek’,‘American’,etc.tomodifytheword‘science’areneededinordertoanswersuchquestions.ThefollowingcommentsconcernWenceslaoJ.Gonzalezanalysisofthefieldsofculturalandsocialsciencesandtheproblems,trends,turnsanddirectionsintheirphilosophy,asseenfromaEuropeanpointofview.1.TASKSAHEADGonzalezthoroughlyandinsightfullyreviewsandmapsthesocialandculturallandscapeinadditiontothesciencesdedicatedtoclarifyingit.Hecompleteshisreviewwithaplanoftaskstobeaccomplished,yearbyyear.Therelevantdisci-plines–socialandculturalsciencesandtheirphilosophy–aretobeexaminedinrespecttotheirbasicconceptsandmethods.Heemphasizesthatthereareseverallevelsofanalysisandsetshimselfthetaskofclarifyingwhatcanbeconsideredas“aEuropeanperspective”.Socialsciencesmakeupaspectrumthatstretchesfrompsychology,socialpsychologyandsociologytothestudyofinternationalrelations.Socialsciencesarededicatedtosociety,politicsandeconomy.Ontheotherhand,culturalsciencescompriselinguisticsandallphilologies(thereareinprincipleasmanyoftheseastherearenaturallanguages),archaeologyandsocialorculturalanthropology.Lit-eraryandfolkloricstudies,history,musicologyandthestudyoffineartsbelongtohumanisticdisciplines.Culturalsciencesandhumanisticdisciplinesarededicatedtosounds,signsandsymbolsandthespokenandwrittenword.Thebasicmethodsofculturalsciencesandhumanitiesaredocumentation,interpretationandexplana-tion.1“History,orsomeaspectsofhistory,aresometimesclassifiedasasocialscience.”Ber-nardWilliams,PhilosophyasaHumanisticDiscipline.PrincetonandOxford:Prince-tonUniversityPress,2006,p.180.\nStateoftheArt245Astomathematicsandnaturalsciences,thesebelongmostlytothebackgroundthantothecontentoftheplannedstudy.Mathematicsandlogicprovidetoolsforanalysingformalstructures.Statisticsandprobabilisticreasoningarecentraltothemethodologyofeconomics,sociologyandothersocialsciences.Theresearchareasofnaturalsciencesandofsocialandculturalsciencessometimesoverlap,whereasthesciences’approachesdiffer.Thesecommonfactorswillreceiveadueconsiderationonthefirsttwolevelsofstudy.Thus,leveli)ofGonzalez’sanalysisconcerns“thegeneralscientificstatusoftheculturalandsocialsciences”,whereasinlevelii)hefocusesonclarifyingof„thescientificstatusoftheculturalandso-cialsciencesascomparedtothatofthenaturalsciences“(cf.Gonzalez´text).Suchastatus-determinationandcomparisonarehighlydemandingtasksbythemselves.Nevertheless,leveliii)oftheanalysisvirtuallyexceedsthegeneralandthecomparativelevelsintermsofdetailstobeconsidered:itconcerns“thespecificissuesonthescientificstatusofeachculturalandsocialdiscipline”.Thiscanbeseeninboth“thebroadapproach”andin“therestrictiveposition”.Thebroadapproachinvolvesphilosophicalreflectiononagivenparticulardisciplineanditsrelationstoscienceingeneral,whereastherestrictiveapproachfocusesspecificallyon“theproblemsofthedisciplineanalyzed”.Thus,Gonzalez’smaindistinctionsare:i)general,ii)comparativeandiii)specific.Herethespecificcanbethebroadapproachandtherestrictedposition.Onemaydrawafigureofthisbyfirstdistinguishinggeneral–specific,andsecondmakingasubdivisionofthespecific-categoryintobroad–restrictive.Moreover,anextralinecanbedrawnbetweenthegeneralandthespecifictomarkthecom-parisonofthegroupofculturalandsocialsciencestothatofthenaturalsciences.Inanygivendiscipline,thevariousproblems’identification,analysisandpos-sibleresolution,aswellastheclarificationofmethodsandtheorieswiththeirramifications,arehugetasksinthemselves(cf.therestrictiveapproach).Society,economy,politics,jurisdictionandculturecontinuouslygeneratenewproblems,andsodothesocialandculturalsciencesthemselvesthatarededicatedtothoseproblems’analysisandresolution.SomereliefisgivenbyGonzalez’squalifica-tionthatthefocuswillmainlybeonmethodologicalproblems.Thebroadap-proachwillhavetoconnecttheanalysesofproblemstoanoverallphilosophicalstudyoftheuniversalfeaturesofscientificresearch–andofthespecificcharacterofagivensocialorculturalscience.Allofthisistoberelatedtoananalysisofthegeneralstatusofculturalandsocialsciences,andtobecomparedwithnaturalsci-ences.Thus,therequirementsaheadaredemandingbutbynomeansimpossibletomeet.\n246ArtoSiitonen2.ONEUROPEAND‘EUROPEAN’Historiansofscienceorphilosophytypicallyraisefivekindsofquestion:where?,when?,who?,what?,andhow?Thedesiderataofthesequestionsareplace,time,persons,contentandstyle.Concerningthedesideratumofthe‚where‘question,itshouldbenotedthatlocalizationofsuchentitiesas‘Europe’,‘America’,‘Asia’,etc.firstrequiresaclarifyingoftheirboundaries,thenaspecifyingofanareawithintheseboundaries–forinstance:whereisEurope?,whereinEurope?Thesearegeographicissues.Chronologicaldeterminationinturnvariesinitsscope,dependingonwhethertheresearchobjectis,say,astreamofthought,aninstitute,auniversity,aresearchgroup,orcertainperiodsintheirdevelopmentorinthelivesoftheirmembers.Thewho-questionconcernsindividualscientistsandphi-losophers,theirlifeandworks.Thequestion‘what?’isaddressedtotheachieve-mentsofthesepersons.Thehow-question,finally,concernsmethodsandstylesofthought.AccordingtoGonzalez,analysesonallthreeabove-mentionedlevels–i)gen-eral,ii)comparativeandiii)specific–shouldbedonewithin,orfromtheintegra-tivepointofview,ofa“Europeanperspective”thatisassumeddefactointhepro-grammeoftheEuropeanScienceFoundation.Thisraisescomplexquestionsasto“theprecisesenseandreferenceofthatexpression.”Onemaythinkthattherefer-enceoftheword‘European’istheconcretegeographicentitywithitspast,presentandfuture,whilethesenseisconstitutedbythefeaturesthatmakethephilosophycultivatedonthatcontinentaspecificallyEuropeanone.Thus,Gonzalezspeaksofanhistoricalandathematicapproachthatcomplementeachother.TheformerisaddressedtothetraditionofphilosophyontheEuropeancontinent,whereasthelatterisconcernedwiththestyleofthinkingthathascreated,andcreates,identity,andisthuspeculiartoEurope.Logicallyspeaking,severalpositionscanbeassumedinrespecttotheexpli-cationofaEuropeanperspective.Gonzalezdistinguishesthreepossiblestand-points.Andafourthonecouldbeadded;butifitwereassumedthenitwouldbeincompatiblewithanyconsolidatedpictureofEuropeanphilosophyofscience.Thewidestalternativehecalls“theintegrativeposition”.Itoutlinesageneralap-proachtoactualphilosophicalworkdoneinEuropewiththegoaloffindingcom-monground.Thesecondviewismorespecific,sinceitrequiresthatonemaycall‘European’onlythosestreamsthatoriginatedormainlydevelopedinEurope.Thethirdviewisevenmorerigorous,insubstitutingthe‘or’byan‘and’,sothattheboundariesareespeciallystrict.Finally,accordingtothefourthview,thereisnotenoughcoherenceamongthevariousphilosophicaloccupationsinEuropetosus-tainaunifiedpicture.Onemaynotethatchoosingamongthesefouralternativesisnotonlyanissueofcontentsormerelyamatteroffactsofculture,butisalsobasedonvoluntarydecisionandcommitment.Accordingly,whyshouldwenotadopttheintegrative\nStateoftheArt247positionanddesignourargumentssothatwecandefendit?ThisappearstobethechoicemadebyGonzalezandencouragedbytheESFsteeringcommittee.Letusalsomentionageographic-historicalconsiderationofEurope,andtwoearlieranalysesoftheconceptofEuropeanphilosophy.‘Europe’isageographicconcept.EuropeconsistsofacertainpartoftheEurasiancontinent,ofNorthSeaisles(Iceland,GreatBritain,etc.)andofMediterraneanisles.Itisamatterofcon-ventionwheretheeasternboundariesofEuropearedrawn;customarilythesearethoughttorunacrosstheUralianmountainsandBosporus.Present-dayEurope,asapolitically,economicallyandculturallyconceivedentity,maybecharacter-izedintermsoftheEuropeanUnionagenda(cf.theareaofthepresentEUanditsexpansionprogramme).TheveryconceptofEuropeanphilosophy,initshistoricalmeaning,seemstoincludesuchcitiesasMiletus,EphesosandAlexandria,andthuspartsofMiddleEast.EuropeanphilosophyformsapartofWesternphilosophy.ThistraditionwasinauguratedinthecitystatesofAncientGreece,continuedinmedievalcitiesandsites(alongsidetheEastern,Arabictradition),andmodernizedsincetheRenais-sance.Linguistically,itwasrootedinClassicalGreek,andcontinuedinLatinandinvariousnationallanguages(Romanic,Germanic,Anglo-Saxon,Slavonic,etc.)Thegeographicandchronologicalcharacterizationof‘Europe’iscompletedby‘Europe’asexperiencedpsychologicallyandsocio-culturally,andasexpressedsymbolically.AspecificallyEuropeanphilosophyaswellasperspectivecanbegleanedfromtheseelements.Europeanphilosophyanditsperspectiveareclearlydistinguish-ablefromtheIndianandChinesephilosophicaltraditions,albeitfarlessdefinitelyfromtheAmericanandAustralianphilosophies.However,althoughthelattercanbetracedbacktotheEuropeantradition,theyarecultivatedelsewherethaninEurope.Inthissense–butonlyinthissense–doestheAustrianphilosophy,forinstance,belongtoEuropeanphilosophy,whereastheAustralianphilosophydoesnot.WearethusallowedtofixsomeboundariesbytracingalinefromPlato’sAcademythroughMedievalcloistersanduniversitiestomodernandcontempo-raryEuropeanuniversities,researchinstitutes,groupsandprojects,aswellasbynamingrepresentativeindividualsandworks.AlloftheseareidentifiableasthedefiningfactorsofEuropeanphilosophyandEuropeanscience.AmongexplicitpresentationsoftheEuropeanphilosophicalandscientificheritage,onemaymentionInnocentM.Bochenski’swork,ContemporaryEuro-peanPhilosophy,2andEdmundHusserl’sculturalcriticisminhiswork,DieKri-sisdereuropäischenWissenschaftenunddietranszendentalePhänomenologie.3Husserl’sworkisbasedonhislectureinViennaonMay7th,1935.Thatyearsaw2BerkeleyandLosAngeles:UniversityofCaliforniaPress1957.(Originally:Europä-ischePhilosophiederGegenwart.1947).3Haag:MartinusNijhoff1962(HusserlianaBandVI).\n248ArtoSiitonenmigrationofmanyintellectuals,andGermanscienceandphilosophywerelosingtheirleadingpositionintheworld.Bochenskidividesthestreamstobeconsideredintosixmainclasses:philoso-phyofmatter,mind,idea,life,essence,existenceandbeing.Heincludestheprag-matismofWilliamJamesandJohnDeweyinthechapteronphilosophyoflife,justifyingthisbyadmittingthatthey“areAmericans,buttheirinfluenceuponEu-ropeanthoughthasbeensoimportantthattheyatleastdeservetouchingupon”.4IfoneaddsthatthesephilosophersnotonlyinfluencedEuropeanthoughtbutwerestronglyinfluencedbyit,theintegrativeposition(cf.above),widelyapplied,maybeseentoaccommodatethem.HusserlisstressingthefreedomofscientificandphilosophicalthinkingthatwasgainedinRenaissanceasarediscoveredgiftoftheClassicalculture.Theactual(i.e.,until1935)crisisofEuropeanphilosophyandsciencehasresultedfromblindformalisticthinkingandone-sidedservicetotechnology.Thewayoutbecomesavailable,whenformalizationiscomplementedbytranscendentalreflec-tion,andexternalanalysisbythestudyofintentionality.53.ON“TURNS”;ONNATURALISMANDREALISMTurnsingeneralmaybesaidtobedecisivephasesinculturaldevelopment;theyintroducetrends.Gonzalezstudiesthenaturalistturnandthesocialturnalongsidewithscientificrealism.Theseturnsaresaidtobepartof–orperhapseventocon-stitute–a“post-historicalturn”,bywhichhepresumablymeansthedevelopmentssinceT.S.Kuhn’sinfluentialtheoryoftheformationofscientificparadigmsandrevolutions.Gonzalezdistinguishesfivetypesofnaturalisminscience,andseventypesofrealism.Inbothcases,animplicationholds:naturalisminthesocialandculturalsciencespresupposesnaturalisminscienceingeneral–andrealisminthecul-turalandsocialsciencespresupposesrealisminscienceingeneral.Thetypesthatnaturalismshareswithrealismconcernsemantics,epistemology,methodology,ontologyandaxiology.Theadditionaltypesofrealismconcernlogicandethics.Naturalistsstressdescription,empiricaltesting,observationandpractice,andtheirstanceisanti-metaphysical(inthesubstantivesense).‘Reducibility’isthekeywordhere:ofmeaningtouse,ofvaluetofact,ofprescriptiontodescription.Real-ists,ontheotherhand,focusonpossiblerelationsbetweenlanguageandreality,andbetweenscienceandfacts.Thisleadsthemtoquestionassumptionsoftruthandobjectivityinnatural,culturalandsocialsciences.Therearetwokindsofalternativetonaturalism:anti-naturalismandamod-erateview,thelatterbeingtantamountto“aninterpretativeperspective”.Inthis4Bochenski,Loc.cit.,p.114.5Cf.Husserl,Ibid.,esp.sections3,10-14,34,61and68.\nStateoftheArt249respect,itseemsthatyetanother“turn”canbeposited:theinterpretiveone.ThisclaimmaybesustainedbymentioningtheachievementsofHans-GeorgGadamer6(historicalunderstanding),PeterWinch7(aninwardapproachtosocialsciences),andGeorgHenrikvonWright8(clarificationoftherelationsbetweenexplanationandunderstanding).Cf.alsotheworkentitledTheInterpretiveTurn.9Asapredecessortothesedevelopments,onemaymentionHeinrichGom-perz’sarticle“Interpretation”10,inwhichhedistinguishesthehistoricalsciencesfromsociologyandanalyzesinterpretationas„oneparticularmethodofhistoricalresearch“.11Thiswouldbeaddressedtoconduct,languageandtexts.Ofthese,heanalyzesconductandcharacterizesthemethodsofitsinterpretationaseithersubjectiveorobjective.ItwouldbefruitfultocontinueGomperz’sprogramme,analysinglanguageandtextsaccordingtohisguidelines.AmuchearlierattempttoclarifytheconceptofinterpretationisBernardBol-zano’sanalysisin§§385-387ofhismainwork.Here,hediscussesthefollowingprocedures:inquiryintojudgmentsbyotherpersons,discoveryofintentionsbe-hindgivenactions,andinterpretationofgivensigns.12Heusestheterm‘herme-neutics’inhisremarksinthetext.13Thisisanobjectivisticanalysistowhichatten-tionisdue.OnemayalsomentionFranzBrentano’sstudyofintentionalityinhis1924workonempiricalpsychology.Gonzalezappliestheconceptofinterpretationto‘naturalism’and‘realism’themselves,becausethevarioustypesofnaturalismandrealismcanbeseenassomanyinterpretations.Ashesays,thereis“aplethoraofversionsofscientificrealism”.Hementionstwooftheseversions:thestructuralrealismofJohnWorrallandthescientificcriticalrealismofIlkkaNiiniluoto.Onemaymentionanothervariant:theontologicalrealismbyHansReichenbachthatwasbasedonprobabi-listicconsiderations.Accordingtohim,theseindicate“theworldasitis,objec-tivelyspeaking.”146WahrheitundMethode.Tübingen:J.C.B.Mohr(PaulSiebeck)1965(originally1960).7TheIdeaofaSocialScienceanditsRelationtoPhilosophy.London:RoutledgeandKeganPaul,1958.8ExplanationandUnderstanding.Ithaca–NewYork:CornellUniversityPress,1971.9DavidR.Hiley/JamesF.Bohman/RichardShusterman(Eds.),TheInterpretiveTurn,Ithaca–London:CornellUniversityPress,1991.10Erkenntnis,7,1937/38,pp.225-232.11Gomperz,„Interpretation“,p.226.12Wissenschaftslehre.Band3.Aalen:ScientiaVerlag,1981(originally1837).13Ibid.,p.553.14ExperienceandPrediction.Chicago:TheUniversityofChicagoPress,1938,p.220.\n250ArtoSiitonen4.DESCRIBING,EXPLAINING,PREDICTING,ANDCONFIRMINGEmpiricalsciencesconcernfacts,eventsandprocesses.Thepointofviewfromwhichtheseareapproacheddeterminesthedifferencesbetweennatural,socialandculturalsciences,andbetweenthemanydisciplinestheyinclude.Eachdisci-plineappliesitsconceptstotheobservationsmadeinitsownsphere,andraisesquestionsinwaysthathavebeenestablishedinitsowntradition–whichinturnisreinvigoratedbynewinquiriesandchallengedbycreativeresearchers.Whatisimportantintheeventstobeclarified,areminutedetailsontheonehand,andpat-terns,regularitiesandtrendsontheother.Itisthetaskofascienceofanykindtoaccountforthese–inotherwords,tomakesenseofthem.Howistheaccounttobegiven?Eachdisciplinedoessoinitsownway;butthequestionalsoconcernstypicalfeaturesthatcharacterizescienceasscience.Everyinvestigationmustdescribefactsandputforwardsuppositionsinresponsetothequestionsraised.Descriptionsanswerthe‘what’and‘which’questions,hy-pothesesthe‘how’and‘why’questions.Insofarasthehypothesesareconfirmed,theyyieldanexplanation.Asuccessfulpredictionhelpstoconfirmahypothesis.Ataskofphilosophyofscienceistostudythecapacitiesofdescription,expla-nation,predictionandconfirmation.Therearetwokindsofquestionabouttheseprocedureswithinculturaldisciplinesandsocialsciences:(1)Howfardoculturalsciencesproceedbydescribingphenomena,andhowfarbyexplainingthem?(2)Whatarethelimitsofexplanationandpredictionwithinthesocialsciences?Con-cerning(1),onemayalsoaskaboutdifferencesamongtheculturalsciences.(Ac-cordingtoGonzalez,thedescription/explanationcontroversyaffectsespeciallytotheculturalsciences,amongthemtosocialanthropologyandhistoricalsciences).Moreover,insuchcasesasgrammar,itappearsthatnotonlydescriptionbutalsoprescriptionisrelevant.Grammarcanbeseenasapartofappliedlinguistics,notasascience.Asappliedlinguistics,itrequiresprescription,andasappliedscience,itrequiresresolutionofconcreteproblems.Asfor(2),onemaywonderifpredic-tionplaysanyroleinculturalsciences.Certainlywehaveexpectationsnotonlyaboutnaturalevents,butalsoaboutsocietyandculture.Sometimeshumanfactscanbepredictedevenbetterthaneventsinnature.Thus,werelymoreonthecon-stancyofourfriends’characterthanonnextweek’sweatherforecast.Withrespecttobothnatureandculture,certainobservedregularitiessustainouranticipations.Insocialsciences,descriptionpresumablyhasaminorrolecomparedwithexplanationandprediction.Thefollowingquestionsarise:howfardothesepro-ceedbyexplainingphenomena,howfarbypredictingthese–andhowfarbyprescribingthem?Aretheredifferencesinthisrespectamongthedisciplinesofso-cialscience?Howaretheresultsofsocialscientificresearchconfirmed?Howarevariouscausalhypothesestested?Whatinterpretationscanbegiventothenotionofcausalityinsocialaffairs?Howcanthegamutofactions,causesandreasonsinpsychologyandsocialpsychologybethoroughlyclarified?Furthercomplications\nStateoftheArt251areself-fulfillingprophesies.Howtoaccountforprescriptionsthatarepresumablybasedonpredictions?(Examplesaboundinsociologyandpoliticalscience).Onehasgoodreasontoexpectbothsocialandculturalsciencestoraise,andtogivewell-foundedanswersto,notonly‚what‘and‚which‘questions,butalsoquestionsthatrequireexplicationandexplanation–i.e.‘how’and‘why’.5.ONMATHEMATICSANDEXPERIMENTATIONFurtherproblemsareraisedbytherolesofmathematicsandexperimentationinsocialsciences,andinhumanities.Thesetwoissuesarecloselyconnected,becauseonemaysupposethatdevisingandorganizingexperimentsisenhancedwhenper-formedwithinamathematicalframework.Mathematicsisconcernedwithorders,structures,relations,classesandnumbers;or,fromapragmaticpointofview,withordering,classifying,deducingandcounting.Realitycanberepresentedbymathematicalmodels.Humanistsandsocialscientistsseemtohavedifferentattitudestowardsusingmathematicsandexperimentsinresearch.AccordingtoGonzalez,“thepredomi-nantview”amonghumanistsfavoursthequalitative,suchassymbolsandtrends,overmathematics.Thosewhoadoptthisviewalsodonotbelievethatexperimen-tationmightbeusefulinstudyingculturalfactors,whereassocialscientistsusu-allyhaveamorepositiveattitudetomathematicalmodelsandexperimentsintheclarificationofsocialissues.Intheresearchofsymbolsandcultures,asforinstanceinsocialanthropologyorinarthistory,mathematicsisnotusedbuttheapproachfocusesonqualitativeaspects.Onthecontrary,incontemporaryhistoryorinsociology,quantitativefac-torssuchaseconomicissuesareinkeyrole,withconstantuseofstatistics.Relevantproblemsincludethefollowing(cf.philosophyofmathematics):a)Theverystatusofmathematicalmodelling.Ismathematicstobeunderstoodasa“language”usedtoestablishknowledgebyproceduresofproof;orasa“heuristictool”connectedtodiscovery?b)Donaturalsciencesdifferfromsocialandcultur-alsciencesinrespecttomathematicalmodelling?c)Whatistherelationbetweenqualitativeandquantitativemodelswithinthesocialsciences?Concerningthealternativesgivenina),thequestionis:Isthis‘or’exclusiveorinclusive?Onemaysuggest,“bothand”.Mathematiciansfruitfullyinterpretmathematicsasalanguageofproof,whereasphysicistsandengineersconsideritasatoolfordevelopingnewideastobetested.Onemaypresumethatvariousheuristicproceduresareusefulalsoelsewhereinscientificresearch.Problemb)raisesanewdifficulty,becausenowthetermsofcomparisonarenaturalvs.social/culturalsciences.Thisproblemcanbetackledinrespecttohisto-ryofscience,orwithinasystematicdimension.Mathematicalmodelswereearlierdevelopedinnaturalsciencesandtechnology,ratherthaninthestudyofsocieties.\n252ArtoSiitonenHowever,statisticshasreceivedakeyroleinmodernsocialsciences.Indeed,thereisnoempiricalareathatcouldbeshownincapableofbeingtreatedmathemati-cally.15Problemc)concernstherelationbetweenquantitativeandqualitativemodelsinthesocialsciences.Itisreasonabletoadopta“bothand”approach.Accordingtotheproblemtobetackled,statisticsandmathematicalmodelsmaybeneeded,oranexplicationofqualitativedifferences,orboth.Gametheoryisafruitfulmath-ematicalbasisforanalysingsocialissues.Ininformationtheoryandcybernetics,thequalitativeandthequantitativeaspectsarebothrelevant.Thesameistrueofeconomics.Astohumanities,forinstancehistorycanbeenrichedbystatisticsofhuman,animalandplantpopulationsintheareaanderawithwhichtheresearchiscon-cerned.Inlinguisticsaswellasliteraryresearch,thestudyofwordfrequenciesisusefulforsomepurposes.Themostconvincingexampleismusicology.MusicandmathematicswerecloselyconnectedalreadybythePythagoreans,becauseinbothofthem,thenotionsofmeasure,numberandharmonyarecentral.Thequestionofexperimentationiscloselyconnectedtotheissueofmath-ematicalmodelling.Gonzalezmentionsasexamplesexperimentaleconomicsandexperimentalpsychology.Onemayaddthatinhumansciences,testsandexperi-mentsareused,forinstance,inlinguisticsandphonetics.Furthermore,therearenoobstaclestotheiruseinmusicologyorinthestudyofliteratureorart,inarchae-ology,etc.6.IGNOREDBRANCHESBasicscienceisoftendistinguishedfromappliedscience;as,forexample,puremathematicsisfromappliedmathematics.Gonzalezmaintainsthatphilosophyofsciencetraditionallyfocusedonbasicscience,whereasappliedsciencesingen-eralhavereceivedlessattentionthanbasicscience.“Amongthefieldsthathavebeenhithertoignoredbyphilosophyarethesciencesoftheartificialunderstoodas‘sciencesofdesign’.”Onemaymentioncommunicationresearch,librarysci-ence,computerscience,economics,medicine,psychiatry,pharmacology,andallbranchesofengineeringsciences,includinggeneticengineering.Adistinctioncanbedrawnbetweentheconcepts‘natural’and‘artificial’;andthusnaturalsciencesmaybedistinguishedfromsuchapplieddisciplinesas“sciencesoftheartificial”and“sciencesofdesign”.Thelatterdeserve,withgoodreason,tobecomenewob-jectsofaphilosophicalanalysiswhosetaskwouldbetoclarifytheirconnectionstoculturalandsocialsciences.15Cf.GerhardFrey,DieMathematisierungunsererWelt.Stuttgart:Kohlhammer,1967,p.120.\nStateoftheArt253Computertechnologyhascontributedtotheemergenceofthevirtualworld,whichhasanartificialcharacter.Thereisacorresponding“scienceofthevirtual”cultivatedinso-called“multi-media”laboratories.Arelevantphilosophicalques-tionis:whatisthedifferencebetweenthevirtualandthefictive?Bothcanbecontrastedtotherealworld.Logiciansmayalsowonderhowpossibleworldsarerelatedtothevirtualworld.Astopharmacology,onemaynotethatitsearlyphilo-sophicalanalysiswasaccomplishedbythecontroversialphysicianParacelsusinthe16thcentury.Hecriticizedthetraditional“iatrocentric”(physician-centred)conceptionofmedicamentsandthushelpedtomakepharmacologyascienceinitsownright.Allinall,inthesciencesofdesignandoftheartificial,practiceandtheoryproceedinafruitful,mutualrelation.7.NOVELTYANDTHEFUTURELastly,Gonzalezaddressesthequestionofemergingandfuturedirectionsinphi-losophyofscience;these,hesays,“maybenewapproachesinthesocialconcernonscience”.Aremarkablepropellingforcebehindthisprojecteddevelopmentisthetensionbetweenrealismandsocialconstructivism.Anotherstraininmethod-ologyconcernsindividualismandholism.Theseissuesarerelevanttoallbranchesofscientificresearchandtheirinter-disciplinaryconnections,whileaspecialfocuswillbegiventoculturalandsocialsciences.Inthisrespect,itisremarkablethatmedicinecanbeunderstood,becauseofitsnewapproaches,asaculturalandsocialscience.AttheendofSection7,Gonzalezsuggeststhattherevivalofscientificreal-ismisconnectedwiththepossibilityofobjectivityinsocialandculturalsciences.Onemaywonderwhereitsoppositeforce,constructivism,issupposedtolead.Tosubjectivismandrelativism,perhaps?However,theremaybesomeobjectiveconstraintsthatwouldenableustoidentifyviableconstructions,andtoproceedtowardsanobjectivewayofregardingsocialandculturalphenomena.8.REVIEWThefinalremarksbyGonzalezconcernthebackgroundfactorsofthephilosophyofsocialandculturalsciences.Theseincludethedevelopmentofsocietyandcul-tureand,correspondingly,socialphilosophyandthephilosophyofculture.Otherbranchesofsciencearealsorelevanthere;notably,mathematicsandnaturalsci-ences.Thisisbecausemathematicsplaysaroleinalldisciplines,andbecausesocialandculturalsciencesarecomparable,inissuessuchasexplanationandprediction,tonaturalsciences.\n254ArtoSiitonenHistoryenjoysaspecialroleamongthebackgroundfactorsofsciences,in-cluding‘history’asoneoftheculturaldisciplines.Inatimelessframe,itispossibletospeakof‘things’and‘facts’,butinusingsuchconceptsas‘events’,‘processes’or‘trends’,wearepresupposingatimeframeandwithit,thebasisforhistory.Thatfactsbecomescientificfactsanddevelopthroughtheprocessofinquiry,wasthebasicinsightofLudwikFleck.16ThisideawasfurtherdevelopedbyThomasS.Kuhnintothetheoryofparadigmsandscientificrevolutions.17Kuhnsaysthatheencounteredby“randomexploration”the“almostunknownmonograph”byFleck,andsawthatit“anticipatedmanyofmyownideas”18.Kuhn’swork,inturn,wascontinuedbyImreLakatos’stheoryofresearchprogrammes.Thisdevelopmentintroducestwomoreturns.Gonzalezreferstothepresentstageofphilosophyofscience“afterthedeclineofthe‘historicalturn’”.Currentlythereis“anactivecompetition”among(1)naturalists,(2)thosewhosubscribetothe“socialturn”,and(3)realists.Ofcourse,thisdoesnotmeanthathistoricalresearch(conceivedas‘history’)cannotbefurthercultivatedanditssubstantialfoundation(conceivedashistory)considered.Gonzalezmentionshistoryamongsocialsciencesanddistinguishesbetweenthemethodologiesof“NewHistory”and“narrativehistory”inSec.5ofhiscontribution.Theformerisanareainwhichmathematicalmodelsareused,whereasinnarration,suchmodelspresumablyarenotemployed.Thus,thenar-rativehistoryispredominantlyqualitativeandfocusesonindividualorsocialagents.TheNewHistoryis“impersonal”,focusingonquantitativefactorssuchasproductionandtransportation.9.THEPROGRAMMEAHEADThetopicsofdiscussionandtasksofresearchahead,onayear-by-yearbasis,arethefollowing:(1)Clarifytheculturalandsocialsciencesandthephilosophical-methodologicalapproachestothem.(2)Analyzethemethodologicalcontroversyondescription,explanationandpre-dictionandtheirlimitsinculturalandsocialsciences.(3)Studyofthedebateonmathematicalmodellinginsocialandculturalsciencesanditscontributiontosocialissues.(4)Analyzethedifferencesbetweentheculturalandsocialsciences,ontheonehand,andthesciencesoftheartificial(e.g.libraryscience,communicationre-search,pharmacology,economics)ontheother.16Cf.hisworkEntstehungundEntwicklungeinerwissenschaftlichenTatsache.Frank-furt/Main:Suhrkamp,1980(originally1935).17TheStructureofScientificRevolutions.Chicago:TheUniversityofChicagoPress,1970(originally1962).18Ibid.,pp.vi-vii.\nStateoftheArt255(5)Investigatethenovelviewsabouttheinfluenceofsocialconstructivismandrealismonsocialsciences.BIBLIOGRAPHYInnocentM.Bochenski,ContemporaryEuropeanPhilosophy.BerkeleyandLosAngeles:UniversityofCaliforniaPress1957.BernardBolzano,Wissenschaftslehre.Band3,Aalen:ScientiaVerlag1981(firstpublished1837).LudwikFleck,EntstehungundEntwicklungeinerwissenschaftlichenTatsache.FrankfurtamMain:Suhrkamp1980(firstpublished1935).GerhardFrey,DieMathematisierungunsererWelt.Stuttgart:Kohlhammer1967.HansGeorgGadamer,WahrheitundMethode.Tübingen:J.C.B.Mohr(PaulSie-beck)1965(firstpublished1960).HeinrichGomperz,“Interpretation”,Erkenntnis,7,1937/38,pp.225-232.DavidR.Hiley/JamesF.Bohman/RichardShusterman,TheInterpretiveTurn.IthacaandLondon:CornellUniversityPress1991.EdmundHusserl,DieKrisisdereuropäischenWissenschaftenunddietranszen-dentalePhänomenologie.Haag:MartinusNijhoff1962(HusserlianaBandVI).ThomasS.Kuhn,TheStructureofScientificRevolutions.2nded.,Chicago:TheUniversityofChicagoPress1970(firstpublished1962).HansReichenbach,ExperienceandPrediction.Chicago:TheUniversityofChi-cagoPress1938.BernandWilliams,PhilosophyasaHumanisticDiscipline.PrincetonandOxford:PrincetonUniversityPress2006.PeterWinch,TheIdeaofaSocialScienceanditsRelationtoPhilosophy.London:RoutledgeandKeganPaul1958.GeorgHenrikvonWright,ExplanationandUnderstanding.Ithaca–NewYork:CornellUniversityPress1971.UniversityofHelsinkiDepartmentofPhilosophySiltavuorenpenger20ASF-00014HelsinkiFinlandarto.siitonen@helsinki.fi\n\nMATTISINTONENSCIENTIFICREALISM,THENEWMECHANICALPHILOSOPHERS,ANDTHEFRIENDSOFMODELLING1.OFMECHANISTSANDMODELLERSAspectreishauntingphilosophyofscienceingeneralandphilosophyofthesocialsciencesinparticular.Inhighlysimplifiedterms,thereisaconsensusthatthesyn-tacticorstatementofscientifictheories(theReceivedView,RV),withproblemsidentifiedanddefinedintermsoftheirrepresentationsinsomeformallanguage,isbeyondpale.Itsmostseriousformalrival,thesemanticview(thestructuralistviewisitsEuropeanvariant)isclaimedtobeinsensitivetohowscientificinquiryisactuallyconducted(seethearticlesine.g.MorganandMorrison1999).HereNMPshavebeenjoinedinbyFriendsofModelsof(FoMs,inbrief)whomaintainthatthefocusshouldbemovedtomodelsandmodelbuildinginsteadoftheoriesinthesenseofRVs.HopesthereforerunhighthatthethealliancesofNMPsandFoMswillchangeourconceptionoftheresultsandtheprocessesofscientificinquiry.NeitherNMPsnorFoMssubscribetoanysingledoctrine.Rathertheyformsomewhatloosegroupsofphilosopherswhowishtoredirecttheeffortsfromthatofformulatinguniversallawsandincreasinglytruthlikeuniversaltheoriestothatofexplicatingmechanismsandbuildingmodelsofnarrowerscope.Suchagoalishighlycongenialtomanysocialsciences,sincetheyneitherpossessnorexpressinterestinpossessinguniversallawsandcomprehensivetheories.However,likethelifesciencesthesocialandculturalsciencesaboundwithmechanismsthatareusedforthepurposesofexplaining.Socialscientistsalsooftenthinkofthem-selvesasengagedinbuildingmodelsthatinturnaimsatdescribingunderlyingmechanisms(forthenotionofmodelingmechanisms,seeGlennan2005).Ifthereisanotionthatcollectsthelifesciencesandthesocialsciencesunderonemetathe-oreticalumbrellaitseemstobethis:theyareinthebusinessofbuildingmodelsofmechanismsthatgeneratemoreorlessregularorindeedsometimeslawlikeconnectionsbetweenphenomena.Nordoesthismetatheoryconfinetothesesci-ences.Onthecontrary,bydrawingattentiontotheweaknessesoftheresearchprogrammeinitiatedbyRVtheNMPsandFoMsopenthecasesofmanyothersciences:manyofthephysicalandtechnicalsciencesareinthesameconceptualboat.ThisisalsothereasonwhyIshallstartwithageneralexaminationofmecha-nismssincemechanismswasthepatternofintelligibilityfortheOldMechanicalPhilosophers(seesection3).F.Stadler(ed.),ThePresentSituationinthePhilosophyofScience,ThePhilosophyofScienceinaEuropeanPerspective,DOI10.1007/978-90-481-9115-4_20,©SpringerScience+BusinessMediaB.V.2010\n258MattiSintonenAdoptingNMP,andembracingthepointofviewofmodelbuilding,hascon-sequencesforoneofthemostpersistentissuesinphilosophyofscienceingeneralandinthesocialsciencesinparticular.TheallianceofNMPsandFoMscallsforamorerealistviewofwhatscientistsactuallydo.AtthesametimeitconductsamutinyagainstrealismasitcametobedefinedinRVanditsaftermath.Realiststraditionallymaintainthattheworldis“outthere”,independentlyoftheinquirer.Thedirectionoffitisfromwordandthoughttoworld,nottheotherwayaround,fortheinquirer´sepistemic(orconative)attitudestowardstheworlddonothavedirect(causal)impactonwhatthings,propertiesandrelationsthereare.ThispartofrealismcausesnoalarmamongstNMPsandFoMs,butthestatementortheoryvariantofrealismwhichmaintainsthatscienceaimsattheoriesthatcapturethecompleteliteraltruthinfielddoes.Thecriticismisnotthatempiricaladequacy,warrantedacceptability,orsomeothervegetariannotionshoulddislodgetruthasthecentralconceptthatmarksasuccessfultheory.Rather,theNMPsandFoMsclaimthatascribingscientiststhegoalgettingatthecompletetruthinthesensedescribed,ofgettingclosertothetruth,ismisguided.Butwhilstthestatementortheoryvariantofrealismmightmissthemarkvis-à-visthesocialsciencesinparticular,thecredentialsofcausalrealismarealtogetherdifferent(theabovedistinctionsdonotofcourseexhaustthevarietiesofrealism).Manysocialscientistsaswellasphilosophersofsocialscienceholdthatthesesciencesonlyenterthecompanyofsciencesproperiftheyareabletorefertocausaleventsandprocessesthatreallyexistand,perhapsalso,trulyarere-sponsiblefortheirsocialexplananda,singularandgeneral.Inthenaturalsciencestherealism/antirealismdebateofJohannesKeplerandUrsusmarkedthehistoricalbeginningsofphilosophyofscience,sinceitwasthefirstconsciousdebateoverthegoalsofscience:forrealiststhegoalisphysicaltruthofcausaeverae,andnotmeremathematicaleleganceinsavingthephenomena(orindeedofassumingmorethanisneeded).1Thesocialsciencesarelatecomersonthescene,butthesamestrategymightworkhere.1PhilosophyofscienceissaidtohavebornwhenJohannesKeplerdefendedTychoBrahe’sproposalforanewworldsystem.ThedebatebetweenKeplerandUrsusin-volvedaccusationsofplagiarism,hurtpridebutalsosubstantialdisagreementsandphilosophicallysophisticatedargumentsoverthegoalsofscience.Keplerarguedthataphysicalinterpretationandaphysicalexplanationwasalegitimateconcerninas-tronomy,andabilitytosavethephenomenabymathematicalmeanswasnotenough(Jardine1991,p.134-135).LaterIsaacNewtonformulatedamethodologicalviewthatanimatedscienceforcenturies:“wearetoadmitnomorecausesofnaturalthingsthansuchasarebothtrueandsufficienttoexplaintheirappearances.”(Newton1934,398).TherequirementofgettingholdofrealcausesisthecornerstoneofNewton’smethodologicallegacy,anditwashighlyinfluentialatleasttillCharlesDarwin´ssearchforthecausesofevolution.Weneedtodistinguishbetweenseverallogicallyspeakingdifferentkindsofclaims.First,onecanclaimthatatypeofcausalprocessissomethingthatreallyexistsinnature,somethingthatisaveracausaandnotjust\nScientificRealism259Thereisnowaythatallaspectsoftruth,realismandmechanismsinthesocialsciencecouldbetreatedadequatelywithinonepaper.Itisimpossibletocoverallofthesocialsciences(orthesciencesofman,orhumaniora)–althoughthereis,forinstance,aninterestingdebateinthephilosophyoflawovertherealityoflawsthatarenotsanctionedorenforced.Ishallnotdwellontheontologicalandepistemologicalconsequencesofthefactthatculturesandsocietiesonlyexistbycourtesyofhumanbeings,andhenceare,inanimportantsense,mind-dependent.Icannotpausetoconsiderthemethodologicalproblemthat,forethicalaswellaspracticalreasons,carryingoutcontrolledexperimentsinmanyoftheculturalandsocialsciencesisdifficultorimpossible(seee.g.Gonzalez2007).Ishallstartwiththeroleoflawsornomicgeneralizations(#2)andIthenmoveontoconsiderNMP(3#).Thefollowingsection(#4)dealswithanalyticsociologywhichnotjustdefineswhatthesocialsciencesneedtorisefrom(orlowerto)grandgeneralschemestosocialsciencesproper.Section#5addresses,verybriefly,sometypesofsocialmechanisms.HerethequestioniswhetherthesocialsciencesindeedmeettherequirementsofNMP,howeverloose.Acrucialingredientinthisisthenotionofamechanicalexplanation,although,followingJaakkoKuorikoski’srecentproposal,itneedstotakeintoaccountthatmechanicalexplanationsinthesocialsciencesareaspecialvariantofNMP(Kuorikoski2009)(#6).Ishallthentakeupactiontheorywhich,Isuggest,couldprofitfromstrategicadvicefromanalyticsociology(#7).IshallthenexpandthispatternofmechanicalexplanationtothecommunicativemechanismsproposedbyPaulGriceandDeir-dreWilsonandDanSperber(#8).Thethirdexampleiseconomics,atradewheremodelbuildingisthebreadandmechanismsthebutter(#9)–andafieldwheredebatesaboutrealismhaveplayedaprominentrole.IshallthentakealookattheissueofrealismfromthepointofviewofthepragmaticturnadvocatedbysomerecentFoMs(#10).ToconcludeIshalltransfertotheinterrogativeandinquisi-tivemode.MuchasIappreciatetheadvantagesofNMP,thereisnoonecoherentnotionofamechanismtocoverthevarioussciences.NorisitobviousthatNMPworksasanindependentmetatheoryforscience.ItmaywellbethatNMPsandFoMshavekickedawaythesyntacticandsemanticladdersabittooearly(#11).theinquiringmind´swishfulthinking.Secondly,onecanclaimthatatypeofcauseissuchthatitcan,inprinciple,haveaneffectofacertainsort.Andthird,aftermattersofprinciplehavebeensettled,onecanclaimthataparticulartypeofcauseinfactisre-sponsibleforaparticulartypeofeffect.JonathanHodge(1977,1987)callstheseexist-ence,adequacy,andresponsibilityclaimsrespectively.Althoughthistripartiteanalysismighthavemuchtoofferinthesocialsciences,allIwishtopointoutisthatthelegacyinsistsonthecausesbeingrealaswellstrongenoughtobringabouttheeffects,andnotsomethingthatarefeigned.\n260MattiSintonen2.DOWNWITHLAWS!Thereislittleinterestinlawsanduniversaltheoriesinthesocialandculturalsci-ences–andinacknowledgingthistheycanfindcomfortfromtherealizationthattheyareingoodcompany.Thelifesciencesclaimtomanagewithoutlaws.Theyaredifferentbutalmostequal.Noristhere,forthisreasonmostly,interestinme-tatheorythatbuildsonthatview.IndeedNMPjoinsthosewhomaintainthatRVwasanartificeofthetool-kitoflogicalempiricism.Thelogicalempiricistreliedontherigorofmodernsymboliclogic(“logistic”)intheexplicationoffundamen-talmetatheoreticconcepts.Hencetheorystructure,confirmation,andexplanationwerealldealtwithintermsoflogic,withlogicalinferenceasaunifyingtheme.InRVlawsornomicgeneralizations–theoriesessentiallyconsistedofsetsoflaws–wereneededtosystematizetheclaimsofafield,tomakepredictionandtoenablecontrol,and,mostimportantly,toguaranteeexplanation.Butashasbeenshownagainandagain,moreisneededthannomicness.ConfrontedwithablackbirdImightask“Whyisthisbirdblack?”andbegiventhefollowinglaw-basedexplanation:“Itisaravenandallravensareblack”.Thiscanbeturnedintoaperfectdeductiveargumentforthetargetexplanandum.Takeanotherexample,thistimefromHempel´sclassicpaper.Iobservethatsomethinginmycarleaksandturntoinspectandaskagain“What?”ItturnsoutthatthecoolerhascrackedandIask“Why?”Iamthenofferedanaccountintermsofthelawsthatgoverne.g.theexpansionofwaterandcontractionofmetalwhenthetemperaturedropsbelowzerocentigrade–againtheputativeexplanationcanbeturnedintoaperfectdeductiveargument.Thefirstexampleofanexplanationishighlyunsatisfactory.Merelyknowingthatallravensareblackdoesnotexplainwhythisbird,araven,isblack–atleastifwehaveinmindscientificexplanation.Wewouldliketoknowsomecausalstory,intermsofproximatecauses,astohowtheraven’sgeneticconstitutionandmechanismsofgeneexpression,yieldthisparticularcolour.Orwewouldliketoknowanothertypeofcausalstory,intermsofdistalcauses,ofwhatsortofselec-tivepressureorotherevolutionaryforceshavebroughtaboutthegeneticconsti-tution.Theofferedpotentialexplanationsoundscheatingbecauseitisnotclear,firstofall,thatitinfactisnomicoreventrue,andsecondly,itfailstospecifyanyintelligiblemechanismforthecolour.Ashasbeenpointedout,thesearenotthetypeofanswerssciencesareinterestedin–noraresuchquestionsinfactraisedinbiology.Theclaimthatlawsarenotneededinallexplanationisnonews.WilliamDray(1957)challengedtheorthodoxybyarguingthathistoriansdonotmentionlawsintheirexplanationsbecausethereare,inhistory,nolaws.Draymaintainedthathistoricalexplanationssucceed(whentheydo)inbringingunderstandingbecausetheyrelyonapeculiarviewoftheactionsofhistoricalagents.Anaction,hesaid,isunderstoodnotbysubsumingitunderoneormoregenerallawsbutbyseeing\nScientificRealism261thatitwas,underthecircumstances,therightorrationalthingtodo.TheoutcomewasparalleltothetraditionofthelogicoftheGeisteswissenchafteninthatitin-sistedondrivingawedgebetweennaturalscientificexplanationbasedon(causal)lawsandexplanationinthesciencesofmanthatisbasedonthesuigenerispatternofVerstehen(foranilluminatingdiscussionondifferentapproachestothisnotionseeGonzalez2003).BothCarlHempelandKarlPopper,thetwomainproponentsofthenomo-theticviewofexplanation,admittedthathistoriansseldomifeverrefertolawsintheirexplanations.Popper’sreasonsforthissorrystatewasthathistoryinfactneedslawsifitexplains,buttheselawsarenotpeculiarhistoricallawsbutrathercommonplacetrivialitiesorperhapssimplepsychologicallaws.Hempelinturninsistedthatthelawsrequiredforexplicithistoricalexplanationsaretoocomplextospellout.BothPopperandHempelthereforesubscribedtothe“Nolaws,noexplanation!”view,althoughfordifferentreasons.Myreasonsformakingthisbriefdetourisnotthewishtodrivehomeanydeepconclusionsabouttwotypesoflogicofexplanation.Rather,Ihopethatitservesasareminderofacoupleoffacts.The“Nolaws,noexplanation!”stancehasenjoyedthestatusofadogmaforlong,butnotwithoutreason.Wedohavetheintuitionthatareasonableamountofregularityandsomedegreeoflawlikenessisrequired.However,itisnotobviouswhatlawlikenessis,orwhatlawsare.2Thelatteraresometimesglossedasexceptionlessgeneralizationsthatmeetavarietyofconditions:–theyonlycontainqualitativepredicates,theyareprojectivleandsupportsubjunctiveandcounterfactualconditionals,theyareuniversaloratleastverywideinscope,theyenablepredictionsandhaveacrucialroleinthesystema-tizationofknowledge,andofcoursetheyunderlieexplanatorypractices(seee.g.Woodward2000,section6foradiscussion).Buttheissueisnotjustwithlackofdeterministiclaws,sincethereareormaybeprobabilisticlawsthatcarrysimilaroranalogousnomicweight.Yetthefirstimpressionisright:deterministicorprob-abilisticgeneralizationbotharesupposedtomaketheexplanandumphenomenonexpected,thoughtodifferentdegrees,anditisthisgeneralrequirementthatcallsfortrouble.Nomicnessassuchgivesnoclueto,andnointelligibleaccountof,whyageneralizationholds.Thetwoputativenomotheticexplanationsarenot,infact,onthesamefoot-ing.Wedohavethefeelingthattheexplanationofthecrackingofthecoolerisilluminating,thoughperhapstoomuchsoforsomeone´sconcern.Thereasonit2CertainlylawlikenessorGesetzmässigkeitwasadifferentmatter,amuchwidernotion,inclassicalGermanscienceandphilosophythanitisincontemporaryphilosophyofscience–andthisisaveryEuropeanthing!Insection5belowItakeupvonWright’sactiontheorywhichwasbasedontheviewthatintentionsarenotHumeancausesofaction.Theyarenottiedtogetherbycausallaws.YetvonWrightdidnotmaintainthatactionarenotlawfulconsequencesfromtheintentions,andheevenacceptedthat,assomeothersusedthenotions,actionscouldbecaused.ButforvonWrightthelawlike-nesswassuigeneris,neitherlogicalnorcausal.\n262MattiSintonenworksisthatitreferstowell-understoodmechanicalprinciples,andnotbecausethereareuniversallaws.ThisiswhattheNMPsclaim.Woodward’sownaccount,nottheonlyoneonthemarket,buildsontheideathatexplanatoryrelationships,unlikeunexplanatoryones,canbeusedtomanipulateandcontroltheexplanan-dumphenomenon.Explanatorygeneralizationsarenotexceptionlesslawsbutratherinvariances.Invariancesinturnarerobustandresilientundervarioustypesofinterventions–andtheythereforeadmitofdegrees.Woodward’sviewofexpla-nationcaneasilybeincorporatedintoNMP.3.LONGLIVEMECHANISMS!ToputtheclaimsofNMPsinperspectiveitisusefultonotethattheyaremostlyconcernedwithscientificexplanation:sciencesshouldaimatexplanationandnotatmeredescription.Furthermore,reachingthisaimisbestachievedbyreveal-ingthemechanismsthatgeneratethephenomenatoexplained,whethersingulareventsandfactsorgeneralizationsandlaws.Toseewhatthisamountsto,letusagreethatallexplanationsareanswerstoexplanationseekingquestions:weaskwhyacertaineventhappened,orwhyageneralizationobtains,orhowaphe-nomenonwasborn(seeSintonen1989,2005).NMPholdsthatthecrucialfeaturethatmakesforexplanatorinessisthattheanswertothequestion“WhyP”,viz.“PbecauseQ”referstoa(orthe)mechanismthatproducesP.Thisistobringthe“cause”backtothe“Because”,butwiththeextrarequirementthatitreferstoacausalmechanism.Thereisasenseinwhichwearewitnessingareturntooldideas(thisisonemeaningof“revolution”),althoughinasubstantiallymodifiedform.ThestandardofintelligibilityfavouredbytheOldMechanicalPhilosophers,OMPs,suchasDescartesandGassendiandRobertBoyle,wasamechanism.3OMPswerein-spiredbythemetaphysicalviewthatallcausesandinteractionsofnaturalbodiesweremechanical,andsomelikeThomasHobbesembracedtheextrememetaphys-icsthatallthereisismatterinmotion.4TheexemplarmechanismforOMPswasamachineorcomplicatedcontrivancesuchastheStrasburgclock.Toexplainitselaboratemovementsonedidnotneedoccultentitiesorforcesbutmechanicalforce,pulleysandgears.Andofcourseitdidnothousemysteriesbecauseitwas3JulianReiss(2007)usestheabbreviationNMPforthe“newmechanicalperspective”.Ihavewantedtomaketheconnection,howeverloose,tothe“oldmechanicalphiloso-phy”of17thCentury.4HereisaquotationfromHobbes’IntroductiontoLeviathan(Introduction):“Forseeinglifeisbutamotionoflimbs,thebeginningwhereofisinsomeprincipalpartwithin,whymaywenotsaythatallautomata(enginesthatmovethemselvesbyspringsandwheelsasdothawatch)haveanartificiallife?”\nScientificRealism263aman-madeartifact.Withitsinnerconstitutionexposed,onecouldseehowitproducedtheactions.5Thisontologynolongerhasgreatfollowing,andthestrictpatternofintelligi-bilityhasbeengivenupformoreencompassingnotions(seeAllen,2005,andtheotherpapersinCraverandDarden,2005,fortheusesof“mechanism”).Thereisnosinglecharacterizationsofwhatamechanismis,andwhatmechanicalexplana-tionamountsto.6Yetthereistheconsensusthat,tounderstandaphenomenonistoseehowregularitiesaboutthatphenomenonarisethroughtheworkingsofanunderlyingmechanism.Suchunderstandingrequiresanaccountoftheentitiesthatareinvolved,andthewaystheseentitiesareorchestrated,intimeandspace,toproducetheregularities.LetusstartwithBechtel’sandAbrahamson’s(2005,423)definition:“Amechanismisastructureperformingafunctioninvirtueofitscomponentparts,componentoperations,andtheirorganization.Theorchestratedfunctioningofthemechanismisresponsibleforoneormorephenomena”.BechtelandAbrahamsenproposethatananalysisproceedsviafunctionalandstructuraldecomposition,inthisorder:onefirstlooksintowhatasystemdoesandthenwhatpartsgointothisdoing.Thefirststagethereforeconsistsoffunctionaldecomposi-tion:onestartsfromoverallfunctionsandthenproceedstocomponentoperations.Thesecondstageconsistsofstructuraldecomposition:heretheinquireridentifiesthecomponentpartsthatdothework,payingduerespecttotheoperations.Amechanismscanalsobenestedwithinanotherone,thusleadingtoare-searchstrategyinwhichevernewfunctionallydescribedblackboxesaresub-jectedtothetreatment.HerewecanappreciatethevirtuesofNMPoverandabovetheadvanceitallowsinthetheoryofexplanation.PartoftheRVwasaparticularviewofascientifictheory:itwastoconsistofanuninterpretedsetofformulasthat5ThusBoyle(1688,1968),p.397-398:“Theworldislikearareclock,suchasmaybethatatStrasbourg,whereallthingsaresoskillfullycontrived,thattheenginebeingoncesetamoving,allthingsproceed,accordingtotheartificer’sfirstdesign,andthemotionsofthelittlestatues,thatatsuchhoursperformtheseorthosethings,donotrequire,likethoseofpuppets,thepeculiarinterposingoftheartificer,oranyintelligentagentemployedbyhim,butperformtheirfunctionsuponparticularoccasions,byvir-tueofthegeneralandprimitivecontrivanceofthewholeengine”.Itshouldbenoted,though,thatBoylecombinedteleologywithmechanismsinawaythattheNMPsdonot,byarguingthatwhenthecontrivanceviewisextendedtonaturalthingssuchasorganismsitisnecessarytoresorttoanintelligencethathaddesignedit.Strictlyspeak-ingthestandardofintelligibilityforCMNincludedadevinedesigner!6AusefulanthologyformechanismsinbiologyisCraverandDarden2005,forthebrainandtheneurosciencesseeCraver2007,forcognitiveneuroscienceseee.g.Bechthel2008inparticular.AclassicpleaformechanismsinthesocialsciencesisJonElster1989,wherehereferstocausalmechanismsasthebasicunitsofthesocialsciences,butseealsoElster1999wherehebringsthenotionofamechanismtobearonemo-tions.Someoftheincreasingliteratureformechanismsinthesocialsciencesaremen-tionedinthesequel,butseeHedströmandSwedberg1999inparticular,aswellastheHedström2005.Foracriticalaccountofthesufficiencyofmechanismsinthesocialsciences,seeReiss2007.\n264MattiSintonenweretheninterpretedbyhelpofobservationsentencesandbridgelaws(aswellassemanticrules)togiveatrue(offalse)theory(“picture”)ofapartoftheworld.Progresswasseenasincreaseingeneralityorscopeandwaseffectedviaconceptandtheoryreduction:conceptsofareducedtheorycouldbedefinedonthebasisofareducingone,andlawsoftheformerinferredfromthoseofamorefundamentalreducingtheory.Theoreticalscientificprogress,then,wouldconsistofaseriesofreductionsmovingtowardstheleveloffundamentalphysicsthatservedastheglimmeringrock-bottom.AlthoughdiscoverywasnotontheofficialagendaofRVitdidhaveanunofficialone:tryandfindtermsandbridgelawsetc.thatmakesuchreductionpossible.Thisviewofinquiryhasbeenuniversallyrejected,andtheattractionofNMPproposesapositivealternative:scientists,andworkinglifescientistsandsocialscientistsinparticular,arenotinthebusinessofformulatinglaworexception-lessregularities,norexplainingbyderivingexplanandafromtheselaws(andini-tialconditions),noraretheyformulatingformalderivationsbetweentermsandtheoriesatdifferentlevels.Theoriesarenotlinguisticentities.Rather,scientistsareinthebusinessoffindingandarticulatingmechanisms.AndwhenNMPsarejoinedbyFoMstheresultis:scientistsareinthebusinessofdesigningmodelsofmechanims.NMPalsoprovidesanaturalsettingforincorporatingpragmaticsintothepic-ture.Ifthestandardofintelligibilityisnotthatofderivationofsingularorgeneralfactsfrom(more)fundamentallawsbutratherthatofexposingmechanisms,sci-entistsandscholarlycommunitiesarenolongerbystanderswhowitnessthatthereisaparticularrelationshipbetweentwosetsofsentences,theexplanandumandtheexaplanans,butparticipantsinaction.Notonlydotheyconstructmechanismsbuttheyalsousethesemechanismswhentheyexplaintotheiraudienceswhyandhowtheexplanandumphenomenontakesplace.WhatthisamountstoisimportantforunderstandingofthecredentialsofNMP,andfortheissueofrealism.Thestand-ardofintelligibilitymightmeanthattheobjectsstudied(individualpeopleandsocieties)aremechanismsandthatitisthetaskofscientiststouncovertheirnutsandbolts.Oritmightmeanthatalthoughtheobjectsstudiedarenotmechanisms(indeed,machinesoratleastcontrivances,inthemostintelligibleofallpossibleworlds),partsoftheirwaysofworkingcanbeexplainedbyreferringtomechani-calprinciples.Theformerreadingwouldrequirespecifyingexactlywhatamecha-nismis–andthisseemsratherahopelesstask.However,itisthelatterclaimthataseriousNMPwantsorshouldwant:resortingtomechanisms,andbuildingmodelsofmechanisms,neednotmakeheftymetaphysicalclaims.Rather,scien-tistsandthosewhoutilizetheirresultstopracticalandeducationalpurposescanusemechanismsandmodelsofmechanismstoexplainhowthingswork.Exactlywhatresourcesareneededtoreachtherequiredresult,theepistemicstateinwhichtheexplaineeunderstandswhyaphenomenonarisesandhowitworks,neednotbespecifiedonceandforallforallpossiblecases.Allmanneroflinguisticandnon-linguisticdevices,graphs,pictures,videos,livespecimen,models,anything\nScientificRealism265thatworksisallowed.Forthequestionofrealismthismeans:thetoolsusedtorepresentaproblemandthesolutionneednotbetrueoftheworld.Whatcountsisthatclaimsmadebyhelpofthemaretrue,andthatthecausalmechanismsthattheystandforarethewaytheyareclaimedtobe.4.ANALYTICALSOCIOLOGYTheOldMechanicalPhilosophy,OMP,founditsshape(orshapes)duringtheScientificRevolution.Thefocuswasonnaturalphilosophy(i.e.science),buttheclaimwasmoreuniversalandalsoencompassedmanandsociety.However,upuntilthe19thCenturythisreflectiontooktheformofspeculationratherthansci-ence.Thereweresocialandpoliticalphilosophiesthatfocusedonwhatanindi-vidualshoulddotocarryoutagoodlife,orwhatformasocietyshouldtaketoen-ableindividualsandthesocietytoflourish.Ortheycouldbuildonsomeparticularphilosophicalanthropologyormetaphysicsofhumannatureandsociety,advocat-inge.g.organismsormachinesasexemplars.Descartesthoughtthatbrutesarejustcomplexmachines,andinsofarashumansaretreatedfromthepointofviewoftheircorporealbeing(asinthestudyofreflexaction)thesamemethodcouldbefollowed.AnextremeformwasHobbes’BodyPoliticinLeviathanwhereeventhesocietyorstatewasconceivedasasuperorganism.Buthowdidtheidealfinditswaytothesocialandculturalsciences?Ononeimportantanalysisnoteventheclassicsofsocialscience,suchasthefunctionalistsorMarxists,managedtoraisesocialsciencestotheleveloftheory.AccordingtoPeterHedström(2005)theystillweretoogeneraltoreachthelevelofexplanation(seealsoHedströmandYlikoski,2009).Asocialtheoryworththetitlemustbeabletodissectthecomplextotalitiesintotheircomponentconstituentsandactivitiesandtoshow,withprecisionandclarity,howthesocialexplanandaarisefromtheintentionalactionsofindividuals.HedströmandRi-chardSwedberg(1998)pointoutthatthisgoalcanbeachievedbyhelpofmid-dle-rangetheories,anotionintroducedbyRobertMerton.Amiddle-rangetheorysteersbetweengrandconceptualschemes(suchasfunctionalism,nottospeakofarmchairphilosophies)ononehand,andmeredescriptionsontheotherhand.Itisnotenoughtoclaimthatallinstitutionshaveafunction.Whatisneededaretheo-riesthatfocusonexplanandathatarebetweenthesocietyasawholeand“thickdescriptions”ofparticularphenomena.Middle-rangetheoriesalsoeschewoverlysimplisticexplanantia,suchasrationalchoicetheory,ifthesearesupposedtopro-videfull-blownexplanationsofallmannerofsocialphenomena.Duetobiasesandinadequacesaswellasdistractingmechanismspeople´schoicesandactionscanfailtherequirementsofrationalitybutstillbeamenabletoexplanationthroughtoolsinthekitoftheanalyticsociologist(seeHedström2005,p.61).\n266MattiSintonenThisiswhereNMPgetsintothepicture.Therearefewlawsinthesocialsci-ences,butplentyofgeneralizationsandmechanismsthatareinvokedtoexplainthem,fromearlycharacterizationsofmarketmechanismsandinvisiblehandstomorerecentgame-theoreticaccounts.Againthereisnosingleaccountofwhatasocialmechanismis,noronenotionofexplanationintermsofmechanisms.JonElster,apioneerofmechanismsinsocialscience,writesthatmechanimsare,roughlyspeaking,frequentlyoccurringandeasilyrecognizablelow-level“causalpatternsthataretriggeredundergenerallyunknownconditionsorwithindeter-minateconsequences.”(Elster1999,p.1).InwhatfollowsIshallfocusonHed-ström’sanalyticalsociologysinceitperhapsbetterthanothersreflectsthespiritofNMP:iturgessocialscientiststodissectmechanismsintoentitiesandtheirworkingsandhencetorevealtheirmannerofworking.Socialsciencesfrequentlyrefertoexplanationscouchedintermsofthevari-ablesthatrepresentfeaturesofanindividualortheenvironmentinwhichthein-dividualacts.HedströmandSwedbergadmitthatsurveyanalysesandstatisticaltechniquesthatarestandardlyusedinchasingandformulatinggeneralizationsthatcapturetheinfluenceofsocialconditionsonindividualsandgroupsarehighlyvaluable.However,the“because”oftheresultantexplanations,evenifcausalinthesenseofcausalmodeling,donotcarrytheforceofanintelligibleconnectionthatwould,orshould,satisfyarationalignoramuswhoislookingforunderstand-ing.Itrecentlyturnedout,inFinland,thatmostabusedgroupofyouthsareboysratherthangirls.Somethingisknownofthestatisticallyrelevantfactorsbutstillthereisthefurtherquestionofhowthegeneralizationsarisefromthebehaviourofindividuals.JonElsterputsthisconstraintonintelligibilityeloquently:tosat-isfyanexplanationmust,intheend,beanchoredinhypothesesaboutindividualbehaviour.Ifweaskwhyconsumersbuylessofagoodwhenitspricegoesupasocialscientistneedstoadoptandtestaspecificassumptionaboutthereactionstoindividualconsumertochangesinprices.(Elster2005,p.45).5.TYPESOFSOCIALMECHANISMSWhatarethesemechanisms,then?DotheyfittheanalysesofNMPs,suchasBechtelandAbrahamsen(orStuartGlennanorJamesWoodward)?7InHedström’sviewtheexplanatorypowerofsociologydoesnotrestondeterministic(orindeter-ministic)lawsbutonsocialmechanisms,understoodasconstellationsbof“enti-tiesandactivitiesthatarelinkedtooneanotherinsuchawaythattheyregularlybringaboutaparticulartypeofoutcome”(Hedström,2005).Bytheseconstella-7ForreasonsofspaceIhavenotbeenabletodealwithGlennan’sorWoodwards’s(2000,2003)importantaccounts,buthereisGlennan’sdefinition:“Amechanismforabehaviorisacomplexsystemthatproducesthatbehaviorbytheinteractionofanumberofparts,wheretheinteractionsbetweenpartscanbecharacterizedbydirect,invariant,change-relatinggeneralizations.”(Glennan2005,p.445)\nScientificRealism267tionsHedströmreferstogenerativemodelsthatarepreciseandexplicitenoughtotietogetherentitiesandactivities,suchasagents,actions,andinteractionofagents.Whatthesemodelsmanagetodoisshowtheplausiblemechanisms(theremaybemorethanone)thatyieldtheoutcome,thedescribedsocialexplanandumphenomenon.SwedbergandHedströmprovideatypologyofsocialmechanismsthatessen-tiallydrawsonthetypeofmechanismatwork.Asituationalmechanismexplainsthewayaspecificsocialsituationshapese.g.thebeliefsandopportunitiesofanindividual.Takeasanexamplethebeliefformingmechanisminworkwhenaninitiallyfalsebeliefturnsintoatruth.Thereareactualexamplesofthismechanismatworkinrecenteconomichistory:supposethatanindividualhearsarumor,orisgivenacluehoweverweak,thatherbankisindeeptroubleandfacespossiblein-solvency.Inordertosavehersavingsacustomermaydecidetotakeallhermoneyout.Someoneelsemayhearofthis,andwhenothers“inevitably”joinintheresultisamassivewithdrawalofmoneyfromthebankandindeedastateofinsolvency.Apartfromsituationalmechanismstherearealsoaction-formationmechanisms(tobediscussedinthenextsection)andtransformationalmechanismsinwhichindividualsinteractwitheachotherandinwhichtheseindividualactionsconspiretoproduceanunintendedconsequence.Swedberg’sandHedström’sexamplesofthelattertypeofmechanismincludestandardprisoner´sdilemmamodelssuchasthetragedyofthecommons.Elster,SwedbergandHedströmallalignthemselveswiththerealists.Ineco-nomics,forinstance,abstracttheoriesarehighlyvaluedbuttheyareoftentreatedwithasusefulinstrumentsratherthan,asinthetraditionofKeplerandNewton,asrepresentationsofrealcausalprocessesormechanisms(orrealisticrepresenta-tionsoftheseprocessesandmechanisms).Butthevirtuesofsimplicityandel-egance,thoughimportantinmodel-building,shouldnotbeguileaninquirerintofictionalism.Atheoryinthesocialsciencesthataimsatunderstandingmustthere-forecombinemethodologicalindividualismwithcausalrealism:itmustspecify“thesetofcausalmechanismsthatarelikelytohavebroughtaboutthechange”,andthesemechanismsmustbethosethatdotheactualwork,“notthosethatcouldhavebeenatworkinafictionalworldinventedbythetheorist”(Hedström,2005)8.Herethemechanism-basedviewdiffersfromrationalchoicemodelssinceinthelatterindividualactorsarerepresentedbyproxyandnotbyindividualswithreallifeprofiles(seeLehtinenandKuorikoski2007).8Hereweseetheveracausastrategy(seefootnote2)atworkagain.Inthesamespirit,Elster(1989)insiststhatthesocialsciencesshouldnotbehappywithstorytelling,withhowthingscouldhavehappened,butshouldtryandfindouthowtheytheyactu-allytookplace.Andherecausalmechanisms,orcausalchains,haveaplaceofpride.\n268MattiSintonen6.SOCIALMECHANISMSASABSTRACTFORMSOFINTERACTIONAgreedmechanismsaretheentitiesweliveby,thereareacoupleofquestionsthatwemustraise.First,howclosetotheexemplarsofOMPorNMP,mechani-calsystemsorevenacontrivances,doweget?Andsecondly,howshouldweconceivethedistinctionbetweenamechanismandmechanicalexplanation.AsJaakkoKuorikoski(2008)hasshown,theanswertothefirstquestionis:Notveryclose.Individualsarenotmechanisms,althoughtheymighthousebiologicalorpsychologicalmechanismsatsomelevel.Societiesarenotmechanismseither,noristheremuchsubstancetotheclaimthattheyhousemechanismsthatwouldcon-sistofcomponentpartsandtheiroperations,orchestratedinspaceandtimeintherequiredsense.What,then,becomesofthesocialsciences’comingofagethroughidentifyingmechanismsthatreallyarethere,generatingsocialregularities?Kuorikoski’sdiagnosisisthatthenotionofamechanismsuitablefortheso-cialsciencesisnotthesameasthatinsomequartersofbiology.Thecharacteri-zationofBechtelandAbelson(2005,seesection2above)fitstwomorespecificandhencedifferenttypesofmechanisms.Ine.g.molecularbiologycomplexsys-temsarestudiedthroughdecompositionandlocalization(seeBechtelandRich-ardson1993).Herethestudyobjectsare,asKuorikoskiputsit,systemsthatarenearlydecomposableintocomponentparts,andthesepartsperformtheirtasksinaccordancewiththeintrinsiccausalpowersoftheparts.Kuorikoskicallsthesemechanismscomponentialcausalsystems(CCSs),andsincetheirconditionsofidentityrefertothecausalpowersoftheparts.TheresearchstrategythatcomeswithCSSscallsfortheopeningofblackboxesatmorefundamentallevel(whichiswhystudyinmolecularbiologistseasilyleadstocooperationbetweenbiolo-gists,biochemistsandbiophysicists).Themechanismsthereforearereallythereinthesensethattheycontainorchestratedandlocalizablepartsthatproducetheregularities(notuniversallaws,sincetheworkingofamechanismiscontingentoninternalandexternalmatters).InKuorikoski’sviewsocialmechanismsarenotCCSsbutabstractformsofinteraction,AFIs.AnAFImechanism,suchasamechanismsthatregulatepriceformationonthemarket,areabstractinthattheyonlytakeintoaccountsomecausallyrelevantfactors.AFIsarealsodecomposableintopartsandtheiropera-tionsthatcontributetothesystem’soverallperformance,butthistheydonotdoinaccordancewiththeirintrinsiccausalpowers.ThecomponentoperationsofsocialmechanismscannotbepairedwithclearlylocalizablepartsasinCCSs,butAFIsareasrealaspiecesofthefurnitureoftheworldasCCSs.ThespectreofNMPishauntingtheworld,andthecallforallNMPs(andFoMs)oftheacademicworldtounitemakessense.Mechanisticexplanationistoruletheworld,reductioninthesenseofidentifyingmorefundamentalmechanismswiththeirowncomponentpartsandtheirorchestratedoperationsisthestrategyto\nScientificRealism269follow.Butalthoughthecalltouniteisnothollowitcomeswithadisclaimer:theunderlyingontologiesandmethodologiesaredifferent.7.EXPLANATIONANDUNDERSTANDINGOFACTIONUnderstandingindividualactions,wesaw,isacrucialconstraintonsocialtheory,andHedström’sandSwedberg’sactionformationmechanismsaredesignedtohelptoseehowspecificactionsarise.Situationalmechanismsinturnfocusonaspecificsocialsituationaffectstheagent.Herethefocusisonspecificallysocialaction,butonecanaskifthenotionofamechanismcouldthrowlightonindi-vidualintentionformationandhenceonhowanagent’sactions(notjustsocialones)aredetermined.G.H.vonWrightsetouttodoinhisactiontheory.Hechallenged,inanoriginalway,boththe“Nolaw,noexplanation”viewandtheideathatactionsarecausallyexplainabletobeginwith.Aleadinginterpretativistandintentionalistheemphasizedthatactionexplanationsaresuigenerisandcannotbereducedtocausalexplanationbyhelpofnomiclaws.Hereishowhedescribedhisview:Toexplainanactionis,broadlyspeaking,togiveatruthfulanswertothequestionofwhytheactionwasdone(performed,undertaken)…AsacommonnameforallthefactorswhichexplainactionIshallusethetechnicaltermdeterminantsofaction.(vonWright1980,p.27)VonWrightdistinguishedbetweenactionsthatareinternallydeterminedandthosethatareexternallydetermined.Supposewehavethequestion”WhydidagentAdop?”Onetypeofpotentialanswerstosuchquestionsareoftheform”Adidpbe-causesheintendedtoobtainq,andtookptobenecessaryfor(orsufficientfor,oratleasthelpfultowards)theobtainingofq”.Theintentionalistmodelofactionex-planationdiscussedindetailinvonWright’sclassicExplanationandUnderstand-ingwasgearedtosuchinternallydeterminedactions(vonWright1971).HereputativeexplanationsrefertoA´svolitionalattitudes(willing,wanting)andtohercognitiveorepistemicattitudes(believing)whichconcerntherequiredmeans(herep)ofobtainingthestateofaffairswanted(q).Togetherthesetwotypesofmentalattitudesformagroundorreasonfordoingp.VonWright’s(1971)wascriticizedfornotbringingintheactionsofothers,sincehispracticalsyllogismfocusedontherelationshipbetweentheintentionofanagentAtobringaboutq,theactionpathatAthinksisneededtobringaboutq,andA´ssettingouttop.Aninternaldeterminantisso-calledbecausethereasonhere,asaorthedeterminantofaction,“isacombinationoftwomentalattitudeswithanagent”(vonWright1980,p.28).Thenotionofanexternallydeterminedactionisapartialresponsetocriticismsofexcessiveinternalism.Externaldeter-minantsarecharacteristicallyresponsessymbolicchallenges.Suchactionsareof-\n270MattiSintonenten,butnotalways,verbal,andtheyresultfromparticipationininstitutionalprac-tices.Learningtorespondtosuchchallengesislearningthesepracticesthroughsocialization.Symbolicchallengescanbepersonal(likeanswerstoaquestion)oranony-mous(likereactingtothe”challenge”posedbyatraffic-light’sturningred).Pos-ingsuchachallengeisanexternaldeterminantofaction.Itistypicalforsuchactionsthattheagentdoesnotformapriorintention:shesimplyreactstothechallenge.Whenaskedforthereasonshecouldrespond:“Youaskedme!”or“Thelightturnedred!”Anexplanationofaperson´sactiontobrakebecausethetrafficlightturnedredcanthereforebeacompleteone:theexternalchallengeisacompellinggroundforperformingtheaction,althoughwewouldnotusuallysaythatitforcesonetoperformit.Anotherformofparticipationinaninstitutionalpracticeisbehaviourwhichfollowsarule.Rulescanbelaws,moralrulesorrulesgoverningmanners.Someoftheseareconstitutivewhileothersareprescriptive.Theformerdefineinstitutionalpracticesandenterintoactionexplanationsonlyindirectly.Thelatterare,invonWright’sview,alsodeterminantsofaction.Itiseasytoseewhy(thoughperhapsnothow)aninternaldeterminantleadstoaction,sinceonvonWright’sviewthereisasortofconceptualconnectionbetweenthedeterminantandtheaction.Butthisdoesnotextendtoexternallydeterminedactions:thatapersonunderstandsarule(oraninstitutionalpractice),andrecognizesthatthesituationcallsforcompliance,doesnotsufficeforaction.Theremustbefurthermeansofsecuringthis.AndherevonWrightfollowsaleadbyJürgenHabermas.Tosecurepublicintereststhesocietytriestomakeitsmembersparticipateinpatternsofcommunicativeaction.AndherevonWrightmakesexplicitreferencetoasocialmechanism:“Tothisendaspecial“motiva-tionalmechanism”hastobeinvented.ItsefficacyiswhatIherecallthenormativepressureinasociety.”(VonWright1980,p.45).IfAdoesnotrespondinanap-propriatemanner,shewillbesubjecttounpleasantconsequences,anythingfromlegalsanctionstoraisedeyebrows.Butpeopledonotinthemainrespondtochal-lengesinordertoavoidpunishment.Themembersofthecommunityinternalizeinstitutionalpracticesandsimplyrespondtothechallengeswithoutfeelingthemasrestraintsontheirfreedomtoact.Itisclear,bythelightsofanalyticsociology,thatweherehavecausalsocialmechanismsatwork.Notonlydowehavethedeterminantsofactions,buttherearemechanismsthatexplaintheformationofintentions,theentitiesthatoncausalaccountscauseactionsdirectly.Whydoindividualsformtheparticularintentionstheydo?Ultimately,vonWrightthought,theyarisefromtwosources,viz.,fromwantsandduties.Wantsmayhavenaturalobjects(suchashealthandhappiness)butalsocontingentones,aswhensomeoneistreatedwithsomethingandacquiresataste.Butpeoplefrequentlyperformactionsthattheydonotparticularlyfancydoingorwanttoperform,suchasreadingpilesoftermpapers.Whydotheytothis?Whatexplainsintentionstodothingsthatonedoesnotwant?Thereasonisthatpeoplealsohavedutiesthatareanimportantkindofdeterminantinintention\nScientificRealism271generation.Butwhencedoesthiscome?Theexplanation,again,isthesamemo-tivationalmechanismsthatmakespeopleparticipateincommunicativeaction:theauraofnormativepressure.Peoplefeelthepressurebuteventually,ifthingsrunsmoothly,internalizethevoiceofdutyandformtheintentionsthatsocietycallsforinparticularcircumstances.Butgiventhatwecoulddrawaflow-chartofdeterminantsofaction,within-ternalandexternaldeterminantsinplace,andgiventhatthesedeterminantscouldberepresentedbyboxesthatmightstillbeopenedtorevealyetdeeperbelief-andintention-formingmechanisms,whydidvonWrightwishtosteerawayfromcausalaccountsand,presumably,explanationsintermsofcausalmechanisms?Toseewhy,notehowquotationinthebeginningofthissectioncontinues:Incolloquiallanguagethesefactors[determinants]aresometimesreferredtoascauses,sometimesasgroundsorreasons,sometimesasmotives.…Theterm“cause”itselfisusedwithamultitudeofmeanings.Forreasonsofexpediency,Ishallreservethetermforwhatisoftencalled“Humean”ornomiccauses.Theyare,roughlyspeaking,causesrelatedtotheireffectsbyalawwhichisaninductivegeneralization(vonWright1980,loc.cit).ActionsforvonWrightarenotinthecausalrealm,andwantsdonotserveascausalspringsofactionbecausetheybothpresupposefreewill,freedomtoactinaccordancewiththeintention,andfreedomtomakechoicesinaccordancewithone´swants.Thisiswhyalsothedeterminantsofactionhaveaninstrumentalistflavour.Theyarenotpartofthefurnitureofthecausalworld.IntentionsandtheirdeterminantsarenotHumeancausesofactions;intentionsandtheirdeterminantsarenotinthehead,whichiswhytheycannotbedetectedindependentlyofandpriortoaction.Inasenseintentionexplanationarelikeexpostacturationaliza-tionsofaction:weseesomeonedosomethingandwelookforamotivationalbackgroundthatmightmaketheactionintelligibleorunderstandable.Butthereisnowaywecould,equippedbythepsychologists’,cognitiveneuroscientists’orsocialscientists’instruments,capturetheentities.However,thepatternsofexpla-nationthatpertainstotheoutwardaspectofintentionalaction,thebodilypart,wasaccordingtovonWrightamenabletoHumeanlaws.VonWright’sviewthereforewasadualisticwithrespecttotheexplananda:ononehadtherewerethebodilymovementsaswellasthebrainevents,ontheotherhandthereweretheactions.8.ANALYTICACTIONTHEORYThesuggestionthatcausationandmechanismsareincompatiblewithintentional-istaccountsofactionbuildsonametaphysicallyloadednotionofacause.Ifthistieissevered,actiontheorymightwellbeincorporatedintothecausal-mechani-calpicturecanvassedbyanalyticsociology.Consider,tostartwith,vonWright’sexperimentalistnotionofcausation:\n272MattiSintonen[T]othinkofarelationbetweeneventsascausalistothinkofitundertheaspectof(pos-sible)action.Itisthereforetrue,butatthesametimealittlemisleadingtosaythatifpisa(sufficient)causeofq,thenifIcouldproducepIcouldbringaboutq.Forthatpisthecauseofq,Ihaveendeavoredtosayhere,meansthatIcouldbringaboutq,ifIcoulddo(sothat).(vonWright1971,p.74)Thisviewofcausationisessentiallytheonewhiche.g.JamesWoodward(2003)startedwithinhiselaborationofthemanipulationistnotionofcausation:causesaresomethingthatcanbeusedtobringaboutsomestateofaffairs.Knowingthecauseofaphenomenonenablesustomanipulatetheworld.Nowignoring,forawhile,vonWright’sdistinctionbetweentheinnerandtheouter(bodily)aspectofactionaswellastheso-calledlogicalconnectionargu-ment,vonWright’sviewfitsNMPreasonablywell.Giventhattheabilitytobringabouteventsorstatesofaffairsisthetouchstoneofacause–causalrealityshallwesay,whatreasoncouldtherebefortheexclusionofthisnotionfromactiontheory?Surelythereisnoobjectionassuchtomaintainingthatwants(ordesires)andbeliefsbringaboutintentions,andintentionbringabout,inappropriatecir-cumstances,actions.Furthermore,thereisnoneedtothink,paceDavidson,thattherearestrictphysicallawsbetweenwants,beliefsandactions,whenproperlyconstrued.TheproposalIsuggestthoughcannotargueforhere,isthatactiontheory,andspeechacttheoryforthatmatter,couldbenefitfromthedistinctionbetweenCCSsandabstractformsofinteraction,AFIs.Recallthattheformer,butnotthelatter,areidentifiedthroughtheintrinsiccausalpowers(ordispositions)ofthepartsthatperformtheoperations.Theyarethereforestuff-dependentandtheirpartscouldbelocalizedinspaceandtime.Nowifthereis,inaCCSsystem,aneventandthateventcausesanotherone,theyareseparatelyidentifiabletokenoccurrences,andhenceHumean(IdonotwishtotakeastrongstandastoexactlywhatHumemaintained,though).Nowthemechanismsthatgenerateintentions,throughthecomponentsofbelief-andwant-generation,arenotCCSsbutAFIswhichiswhysomeoneequippedwithathicknotionofcausemightfinditobjectionabletocallthesecomponentscausallyefficatious.Theregularitiesgeneratedneednotbelaw-like.Andofcourseherewehavetheextradifficultyofnotbeingabletointervenewiththepartstoseehowthesystemoperates,althoughthenewimagingtech-niquesmakeusthinkdifferentlyinthefuture.Thecrucialquestionis:couldoneneverthelessbringaboutdeterminantsofaction,thatis,beliefs,wantsandintentionsinanotherperson?Certainly.Icaninduceinyouabeliefthatitisraining,whenyouarenotabletochecktheweatherforyourself,byuttering“Itisraining”.Icantherefore,andthereby,bringaboutintentionsandactionsinyou,inaratherreliableway.ForinstanceIcouldbringabout,inyou,theintentiontotakealongtheumbrellawhenyougoout.Inducingwantsanddesiresinotherpeoplemaybeamoredifficultmatter,butnothopeless.BydisplayingachocolateboxonthekitchentableIcanprettyreliablybringit\nScientificRealism273aboutthatmychildrencometohavethedesiretohavechocolate,andoftentheintentionanditsexecutionfollow.Thedesire,andthebelief,thereforeproducein-tentionandaction.Dittoforsymbolicchallenges.Whenyouareabouttowalkun-deramovingcar,Iyell“Watchout!”andsaveyoufrominjury.WhenIamasked,indinnertable,topassthesaltviatheverbalrequest“Couldyoupleasepassthesalt!”Icomply.Infactjustaboutallperformancesofspeechactsfallwithinacausaltheoryofactionsuitablyconstrued:anutterercanperformalocutionary,illocutionaryorperlocutionaryactbywayofutteringatokenutterance,andthesearesuccessfultotheextentthespeakermanagestoutteratokenexpression,ormanagestogetthehearerrecognizehercommunicativeintention,ormanagestogetthehearerrespondinthewayshewishes.Thereisnoreasontodenythatwecanbringaboutmentalstatesinotherpeo-ple,andindeedwedosoregularlyinhumanverbalandnonverbalcommunication.Itmayofcoursebeobjectedthatmanipulationofotherpeople’sactionsbypro-ducinginthemcertainmentalstatesistobemorallycondemned.Indoctrinationisacaseinpoint,and,arguably,usingsocialscienceinmarketingtogetpeoplebuyproductswhichtheydon’tneed,ordidnotevenknowbefore,mightbefoundobjectionable.Butmanipulationassuchismorallyneutralandastheinnocentexamplesaboveshowweemploymanipulationtechniquesallthetimetobringaboutmorallygoodorneutralresults.9.THESCIENCEOFCOMMUNICATIONTheso-calledGriceantheoryofcommunicationinfactisanaccountofthemech-anismthatenablesustoexplainthywayface-to-facecommunicationworks,whetherlinguisticornon-linguistic(Grice1989;seealsoforBachandHarnish1979foranearlyaccountofthedetails;again,newimagingtechniquesmightwellbringaboutfactsabouttheneuro-cognitivebasisofthesemechanisms–andhencehelptodecidebetweenalternativeones).Therehadbeenrivalproposalsforthemechanismbefore,andtherehavebeenimprovementsandnovelproposalssince.Moreover,theexplanatorystrategyfitsNMPalsointhatwecandistinguishsubmechanismsintheoverallmechanisms.Verybriefly,thestandardmechanismforcommunicationthatreignedpriortoGricecanbecapturedwiththecodemodel.Accordingtothismodelaspeakerisequippedwithadevicethatconsistsofseveralcomponentssuchastheseman-tic,syntacticandthephonologicalcomponent.Thisdevicecodesshespeaker’sthoughtintoaformthatcanbetransmittedviavibrationsofair.Thelistenerhasanidenticalcontrivancethatworksintheoppositeorder.Thevibrationsreachtheearandthenthebrainwherethemessageisdecoded,startingwiththeoperationofthephonologicalcomponent,andeventuallyreachingthesemanticone.99Howmuchofthecognitiveneuroscientificmechanismsaretobeincludedinthemodel\n274MattiSintonenPaulGricechallengedthismodelbyarguingthatcommunicationtakesplacethroughintentionsandtherecognitionoftheseintentions.InGriceancommunica-tionthespeakerandthelistenerhavesomemutualcontext-dependentinterpre-tationstrategiesormaxims;similarly,theyhavesomemutualandotherbeliefsconcerningreality,includingwhatotherpeoplewantandbelieve.Inanattempttocommunicatethespeakermodifiesthelistener’sacoustic(orvisual)environmentbyofferingdirectorindirectevidenceabouthercommunicativeintentions.Thelistenerreasons,onthebasisofverbalevidenceandnon-verbalclues,relyingonmutualbeliefs,whatthespeaker’scommunicativeintentionsatanyonetimeare.TheGriceanmodeldifferssubstantiallyfromthecodemodelofcommunica-tion,sincethemechanismsitpostulatesisbasedonreasoningratherthanexplicitcoding.Totheextentthereasoningandtheprinciplesonwhichthisreasoningfol-lowscanbemadeexplicit–andthishasbeenthetargetofspeechacttheorysinceGrices’s(andAustin’s)seminalpapers–thereisreasontothinkthatalsospeechacttheoryfitsNMP.TherehavealsobeenlaterproposalsthateitherimproveorreplacetheGriceanmechanismoritssubmechanismsbydifferentones.DanSperberandDeirdreWilson(1976)propose,intheirRelevanceTheory,thattheGriceanmaximscanbereplacedbyasimplerCognitivePrincipleofRelevance(CPR):Humancognitiontendstobegearedtothemaximisationofrelevance.Whensomeoneutterssomethingthehearer,guidedbyCPR,triestofindthein-terpretationthatismostrelevantinthecontext.Thisisnottheproperplacetogointothedetails,sufficeittopointoutthatthereisindeedpromisethatitleadstoempiricallycontrolledresults–andthatitindeedfitsNMP.TheRelevancetheoryclaimsconsiderablepredictiveandexplanatorypowerinawiderangeofapplica-tions.AndWilsonandSperber(2004)write:Theuniversalcognitivetendencytomaximizerelevancemakesitpossible,atleasttosomeextent,topredictandmanipulatethementalstatesofothers.Knowingofyourtendencytopickoutthemostrelevantstimuliinyourenvironmentandprocessthemsoastomaximizetheirrelevance,Imaybeabletoproduceastimuluswhichislikelytoattractyouratten-tion,toprompttheretrievalofcertaincontextualassumptionsandtopointyoutowardsanintendedconclusion.Forexample,Imayleavemyemptyglassinyourlineofvision,intendingyoutonoticeandconcludethatImightlikeanotherdrink.ofcommunicationisamatterofchoice.HereishowPeretzandZatorre(2005,90)sumupthepath,muchthesameforspeechandmusic:“Asoundreachingtheeardrumsetsintomotionacomplexcascadeofmechanical,chemical,andneuraleventsinthecoch-lea,brainstem,midbrainnuclei,andcortexthateventually–butrapidly–resultsinapercept.Thetaskofauditorycognitiveneuroscienceistofigureouthowthishappens.Musicalsoundsandallothersoundssharemostoftheprocessingstagesthroughouttheauditoryneuraxis.”\nScientificRealism27510.REALISMANDTHEPRAGMATICSOFMODELBUILDINGSomevoicesintheallianceofNMPsandFoMsclaim–soitseemsatleast–thatthephilosophyofscienceofyesteryearputtoomuchemphasisonclaimsandtheoriesasunitsofanalysis.ThecomplaintwasthatneitherRVnorthesemanticviewgivesreal-lifescientistsorscholarlycommunities,withtheiraimsandpossi-blydifferingexplanatoryidealsandinterests,theirdue.Andnotonlydotheyleaveoutthepragmaticdimensionwhichdescribescognitiveprocessesandsocialinter-actionwithinacommunity,thetheory-centeredandproposition-obsessedviewsdonotpayenoughattentiontothematerialaspectofinquiry.Giventhefocusonmodels(andmechanims),thegoalofgetting(closerto)theTruth–highlyimportantforcriticalscientificrealismforinstance–isabouttoloseitsprideofplaceintheagendaofphilosophyofscience(forthis,seeNi-iniluoto1999).10ThusRonaldGiere(1988)poposedamodel-basedaccountofscienceinwhichscientistsbuildidealizedsystemsor“theoreticalmodels”(or“models”,forshort)withoutdetoursviatheirlinguisticrepresentations.Totheex-tentthereis,forinstance,asuitablerelationshipbetween,say,equationsandtheircorrespondingmodelsonecan,Gierewrites,speakoftruth.Buttruthherehasnoepistemologicalbite:theequationsdogiveatrueaccountofthecorrespond-ingmodel,butonlybecausetheyaresodefinedastosatisfytheequations(Giere1988,p.79).Astotherelationshipbetweentherealsystem(lateroftencalledthetargetsystem)andthemodel,therelationshipisnotoneoftruth(sinceneitherisalinguisticentity)orisomorphismbutratheroneofsimilarity.GierelatersummeduphislonggoodbyetoRVbymaintainingthattruthisa“carryoverfromanolderpictureofscience.”(1999).Againthisishighlycongenialtophilosophersofsocialscience,andthepres-suretojettisonthenotionoftruthhasbeenvisibleineconomictheoryinparticular.Philosophyofeconomicsthereforecontainsoneofthebest-focusedrealismde-batesintheentirefieldofsocialsciences.Econometricmodelsareoftensaidnottorepresentrealityasitis,andhencetogiveafingertoinstrumentalism.Somehavethoughtthatthisamountstoabetrayalofthecause,whereasothersholdthatitisagoodthingandonlyshowsthateconomicsislikeanyscience:itaimsatdistillingthetruth,anddoingthisrequiresremovingirrelevantdetail.Econometricmodelsneednotbetrueofeconomicrealitybuttheyrathersimplify,idealize,etcbyomit-tingalotofdetail.10IshouldperhapsmakeitclearattheoutsetthatIdonotsubscribetoallthesecriticismsoftheoldphilosophy.Thisshouldbecomeobvioustowardstheendofthepaper.Butverybriefly,fromthelogicalpointofviewtheallegeddividebetweenthesyntacticandthesemanticviewsisnotallthatgreat,sincetheycanthetwotypesofformalizingtheoriesareinthesamelogicalboat.Atheorygiventhroughitsmodelscanbegivenalinguisticinterpretationandviceversa.Secondly,oneNiiniluoto’spointsispreciselytodevelopatheoryoftruthlikenessinwhichclosenesstotruthorverisimilitudeadmitsofdegrees.\n276MattiSintonenIcanhereonlypickupacoupleofdevelopmentsthatseempertinenttothetheme.InaforthcomingpaperUskaliMäkiconfessestobeingarealistinhisphilosophicaloutlookbutpointsthattheoverallpictureismorecomplicatedthanisusuallyadmitted.Economicmodels,hewrites,arenodifferentfromphysicaltheoriesinthattheyrefertoidealizations:wherephysiciststalkoffrictionlessplanesandperfectlyelasticmolecules,economistsfeelfreetomakehighlyideal-izedorstraightforwardlyfalseassumptionsconcerninge.g.completeinformationandzerotransactioncosts.Butinsteadofjustpraisingorcriticizingsuchanappar-entlyrecklessbentofmindMäkipointsoutthatfalseidealizationsareameanstoanoble(indeed,moretruthlike)end,viz.,thatof“theoreticallyisolatingcausallysignificantfragmentsofthecomplexreality”(Mäki,fortgcoming;seealsoMäki1994,1996aswellasLehtinenandKuorikoski2007).Onethereforeneednotdenyontologicalrealism–thattheeconomicworldhasacertainstructurethatisindependentoftheeconomictheorizer,norindeedthattheoriesandmodelsaretruetotheextenttheymanagetocapturethosestructures.WhatmatterstoMäkiarenotthefeaturesthatscientificrealistsnormallywishtocelebrate,suchasap-proximatetruthofactualtheoriesorpredictiveandreferentialsuccess,butratherthatthingsintheworldareinonewayratherthananother,andthatascienceworththetitleshouldtryandfindoutwhatthatwayis.Itisthisnoblegoal,then,thatjustifiesusingfalsehoods.Unrealisticassump-tionsaboundineconomicmodellingbutthisshouldnotcausealarmsincetheseidealizingassumptionsareexploitedforthepurposesoftheoreticalisolation.Mäki’spetexampleisJ.H.vonThünen’smodelofagriculturallandusethatas-sumescitieswithoutanyresptrictionsondimensions,orneighbouringtowns,aswellasuniformfertilityandclimateetc.(Mäki,forthcoming,p.79).Therearenosuchcitiesandevearyoneknowsthat.Butcananythingmorebesaidaboutthenoblegoaloftruth?InMäki’scaseitisrealismaboutthemechanismsthatareclaimedtooperatenotjustintheimaginedmodelbutintherealortargetsystem.Needlesstosay,thetargetsystemwasnotanyparticulartown(theoreticalsciencesseldomareconcernedaboutparticularsystems,onlyparticularsystemsthatareusedtomakemoregeneralpoints).Thisbeingthecase,thefalsitiesadoptedarejustameanstobuildingamodelthatisrealisticinthesensethatitcapturesthemechanismsinoperationinthetargetsystems.ThereisfeaturethatFoMsarefondofemphasizing,asofrecently,viz.thepragmaticaspectsofmodelbuilding.Ishallconcludewithitsinceithighlightsamoregeneralfeature.Ihaveassumed,above,thatscientificexplanationscanbeusefullyconceivedasanswerstowhy-andhow-questions.ThisindeedwasHempel’sgambitinhisclassicnomotheticaccount.ButHempelalsomadeitclearthathewantedtocapturethelogicalandnotthepsychologicaloranthropomor-phicalascpecsofunderstanding,andhenceofexplanation.NowHempel’slogicofexplanationwasintendedasanexplicationofthelogicalaspectofexplanation,anditwascarriedoutbyconstruingtherelationshipoftheexplanandumandtheexplanansasatwo-placedrelationshipbetween(setsof)sentences.Nowcritics\nScientificRealism277wereeagertopointoutthatwhatcountsasanexplanationdependsonwhotheexplaineeisandwhatshealreadyknows(andwishestoknow,perhaps).Later,thiscriticismleadtopragmaticmodelsofexplanationthatmadetheexplanationrelationshipa3-placedone(“SexplainstoHwhyQ”),orincorporatedevenmoreingredientsintothestructure.MattiSintonen(1989)arguedthattherecanbenotwo-placedformalrelationstocaptureexplanatoriness,butclaimedthatbringinginsubjectsandtheirknowledgeaswellastheirinterestsandthecontextdoesnotleadtopernicioussubjectivismorrelativism.Thesepragmaticfeaturesarenowbeingbroughtintomodel-building.Ron-aldGiere’scognitiveapproachtoscientifictheorizingalreadyleftthisdooropen.Thenotionofsimilarityiscrucialtothataccountbut,asGierenotes,thisraisestheproblemofrespectsanddegreesofsimilarity.Indeed,hethoughtthat“itisrespectsofsimilarity,notdegrees,thatprimarilyseparaterealistsfromanti-real-ists”(Giere1988,p.93).Butassessingrespectsinevitablybringsinthescientistsandtheircommunities,fortheyaretheagentsthatdotheassessmentworkhere.Gierelatermakesthispragmaticcommitmentexplicitbyallowingpurposestoenterthepictureinhisattempttounderstandrepresentationalpracticesinscience.Hewritesthat“wearelookingatarelationshipwithroughlythefollowingform:SusesXtorepresentWforpurposesP”.Asalreadysuggested,andasGierenotes,thisproposaldrawsanexactparallelinthetheoryofexplanation.GiereindeedallowsSstandfor“anindividualscientist,ascientificgroup,oralargerscientificcommunity”.Togetahandleonadmissibleorfruitfulwaysoffalsifyingreality(myterm)wemustnotethatmodelbuildingisnotjustatwo-placerelationsbetweenthetargetobjectorsystemandthemodel–whetheraphysicalmodelor,moreoften,systemofequationsofwhatnot.Rather,modelsaredesignedbypeoplewithpar-ticularpurposesandaspectsinmind.Similarly,thereisaparticularaudienceforwhichthemodel,ortheresultsofthemodeling,areaddressed,andperhapsalsofurthervariables.Ihaveargued,elsewhere,thattakingintoaccountpragmaticas-pectsofexplanationdoesnotleadtosubjectivismorantirealism.Similarly,bring-ingpragmaticsintothebusinessofbuildingmodels(ofmechanisms,often)isnotadenialofrealismoroftherequirementofreality.AsMäkiputsitinthecaseofeconomicmodelling,itisbutaconcessiontothefactthatmodelingaswellassimulationareconductedwithparticularpurposesinmind.HereMäkiwantstogodeeperonthepragmaticpathandbringinfurthervariablestotheeffectthathisac-countofrepresentationbecomes-placed:“AgentAusesobjectMasarepresenta-tiveofsometargetsystemRforpurposeP,addressingaudienceE,(potentially)promptingissuesofresemblancetoarise.”\n278MattiSintonen11.INTHEINTERROGATIVEORINQUISITIVEMOODWestartedthisbrieftourintorealisminthesocial(andcultural)sciencesfromtheclaimoftheNMPsandFoMstoprovideaphilosophyofscienceingeneral,andinthesocialsciencesinparticular,thatcouldbemorerealisticthanRVingivinganaccountofwhatscientistsactuallydo.Thecasesexaminedonlyconstitutedapossiblyone-sidedsample:socialtheoryorsociology,actiontheoryandthetheoryofcommunication,andeconomics.ItremainstoaskiftheNMPsandFoMshavebeensuccessfulintheirclaims.TheclaimoftheNMPsstandsstrongatleastinthissense:thestrategyoffindingandrefining(causal)mechanismscertainlyissuperiortopreviousaccountsofwhatsocialscientistsconceivethemselvesasengagedin.AstotheclaimoftheFoMsgoes,thequestionwhethermodelsshouldbetheunitofappraisalismoot.Clearlyalsotheirclaimofbeingmorefaithfultoactualpracticeiswelljustified.YettherearesomeconceptualcloudsintheNMP(cumFoM)horizon,anditconcernstheultimateconceptualautonomyorsufficiencyofmechanismsandmodellingvis-à-vissomespecificallyphilosophicaltasks.RVhadaspecificphilo-sophicalmission:acanonicalwayofspecifyingthecontentofatheory.Thisca-nonicaltoolextendedtothenotionofanexplanations(whatisanexplanation),toprogressthroughreduction,ortoinductivesupportprovidedbyobservationstatementstoatheory.Theseaccountsweregivenalinguisticandaformalguise,andonecouldthenaskaboutthetruthortruth-likenessofatheory,orcomparethecontentsofrivaltheoriesanddebatetheirmeritswithrespecttoeachother.ThemajormotivationforusingRVnotionsinadescriptionwasthatthesenotionscouldprovideanobjectivehandleonwhatcountsasascientificclaim(asagainstloosetalk),oranexplanation(asagainstsubjectivesenseofcomprehension)orconfirmation(whichwouldbeindependentofpragmaticfactors).Especiallynowthatpragmaticaspects,peoplewiththeirknowledgeandinterests,areexplicitlywelcomedintometatheoryonecouldraisesomenaggingquestions:first,doestheallianceofNMPsandFoMsgiveusanormativehandle,soimportantforphiloso-phyofscience,todiscussthemeritsofalternativeproposals.Ifitdoes,doesitdosowithoutresortingtoRVor,shallwesay,formalnotionsinanessentialway?REFERENCESGarlandAllen,“Mechanism,vitalismandorganicisminlatenineteenthandtwen-tieth-centurybiology:theimportanceofhistoricalcontext”,Stud.Hist.Phil.Biol.&Biomed.Sci.36,pp.261-283.KentBachandRobertM.Harnish,LinguisticCommunicationandSpeechActs.Cambridge,MassachusettsandLondon,England:TheMITPress1979.\nScientificRealism279WilliamBechtel,MentalMechanisms:PhilosophicalPerspectivesonCognitiveNeuroscience,Routledge,2008.WilliamBechtelandAdeleAbrahamsen,“Explanation:amechanistalternative”,Stud.Hist.Phil.Biol.&Biomed.Sci.36,pp.421-441,2005.CarlF.CraverandLindleyDarden,“Introduction”in:StudiesinHistoryandPhi-losophyofBiologicalandBiomedicalSciences.36,2005,pp.233-244.CarlF.Craver,ExplainingtheBrain:MechanismsandtheMosaicUnityofNeu-roscience.Oxford:ClarendonPress.2007.NewtonC.AdaCostaandStevenFrench,“TheModel-TheoreticApproachinthePhilosophyofScience”,PhilosophyofScience57,1990,pp.248-265.WilliamDray,LawsandExplanationinHistory.Oxford:OxfordUniversityPress.1957.JonElster,NutsandBoltsfortheSocialSciences.Cambridge:CambridgeUniver-sityPress1989b.JonElster,AlchemiesoftheMind:RationalityandtheEmotions.Cambridge:CambridgeUniversityPress1999.StevenFrenchandJamesLadyman,“ReinflatingtheSemanticApproach”in:In-ternationalStudiesinthePhilosophyofScience13,2,pp.103-121.1999.StuartGlennan,“ModelingMechanisms”in:StudiesinHistoryandPhilosophyofBiologicalandBiomedicalSciences36,2005,pp.443-464.RonaldN.Giere,ExplainingScience:ACognitiveApproach.ChicagoandLon-don:TheUniversityofChicagoPress.1988.RonaldN.Giere(ed.),CognitiveModelsofScience.MinnesotaStudiesinthePhilosophyofScience,Vol.XV.Minneapolis:UniversityofMinnesotaPress.1992.RonaldN.Giere,SciencewithoutLaws.ChicagoandLondon:TheUniversityofChicagoPress.1999.RonaldN.Giere,“HowModelsAreUsedtoRepresentReality”in:PhilosophyofScience71,pp.742-752.2004.WenceslaoJ.Gonzalez,“FromErklären-VerstehentoPrediction-Understanding:TheMethodologicalFrameworkinEconomics”,inSintonen,M.,Ylikoski,P.andMiller,K.(eds),RealisminAction:EssaysinthePhilosophyofSocialSciences,Kluwer,Dordrecht,2003,pp.33-50.WenceslaoJ.Gonzalez,“TheRoleofExperimentsintheSocialSciences:TheCaseofEconomics”,inKuipers,T.(ed),GeneralPhilosophyofScience:Fo-calIssues,Elsevier,Amsterdam,2007,pp.275-301.PaulGrice,StudiesintheWayofWords.Cambridge,Mass.,HarvardUniversityPress1989.RomHarré,“MaterialObjectsinSocialWorlds”in:Theory,Culture&Society19(5/6):2002,pp.23-33.PeterHedström,DissectingtheSocial.OnthePrinciplesofAnalyticalSociology.CambridgeUniversityPress.2005.\n280MattiSintonenPeterHedströmandRichardSwedberg,“SocialMechanisms:Anintroductoryes-say”,inPeterHedströmandRichardSwedberg(Eds.)SocialMechanisms.AnAnalyticalApproachtoSocialTheory.Cambridge:CambridgeUniversityPress.1998.PeterHedströmandLarsUdéhn,“Analyticalsociologyandtheoriesofthemiddlerange”,inPeterBearmanandPeterHedström(eds.)TheOxfordHandbookofAnalyticalSociology.(Forthcoming).PeterHedströmandPetriYlikoski,“AnalyticalSociology”in:SAGEHandbookofPhilosophyofScience(editedbyI.JarvieandJ.Zamora-Bonilla),SAGE,NewYork,forthcoming,2009.CarlHempel,AspectsofScienftificExplanationandOtherEssaysinthePhiloso-phyofScience.NewYork:TheFreePress1965.M.J.S.Hodge,“TheStructureandStretegyofDarwin’sArgument”.BritishJour-nalfortheHistoryofScience10.Pp.237-246.1977.M.J.S.Hodge,“NaturalSelectionasaCausal,EmpiricalandProbabilisticThe-ory”.InL.Krüger(ed.),TheProbabilisticRevolution,vol.2.Cambridge,Massachusetts:MITPress.Pp.233-270.1987.NicholasJardine,JardineTheScenesofInquiry:OntheRealityofQuestionsintheSciences,ClarendonPress,Oxford.1991.JaakkoKuorikoski,“TwoConceptsofMechanims:ComponentialCausalSystemandAbtractFormofInteraction”,InternationalStudiesinthePhilosophyofScience.(forthcoming).AkiLehtinen,andJaakkoKuorikoski,“UnrealisticAssumptionsinRationalChoiceTheory”,PhilosophyoftheSocialSciences,vol.37,no.2,pp.115-138.2007.UskaliMäki,“Isolation,idealizationandtruthineconomics”,IdealizationinEconomics,editedbyBertHammingaandNeildeMarchi,specialissueofPoznanStudiesinthePhilosophyoftheSciencesandtheHumanities,38,pp.147-168.1994.UskaliMäki,“ScientificRealismandSomePecularitiesofEconomics”,inR.S.Cohen,R.HilpinenandQiuRenzong(eds.),RealismandAnti-RealisminthePhilosophyofScience.Dordrecht:Kluwer,pp.427-447.1996.UskaliMäki,“Myphilosophyofeconomics:Realisticrealismaboutunrealisticmodels”,inHandbookofthePhilosophyofEconomics,ed.HaroldKincaidandDonRoss.OxfordUniversityPress.Forthcoming.MaryS.MorganandMargaretMorrison(eds.),ModelsasMediators.Perspec-tivesonNaturalandSocialScience.Cambridge:CambridgeUniversityPress.1999.IlkkaNiiniluoto,CriticalScientificRealism.Oxford:Clarendon.1999.IsabelPeretzandRoberJ.Zatorre,“BrainOrganizationForMusicProcessing”,AnnualReviewofPsychology.56:89–114.2005.JulianReiss,“DoWeNeedMechanismsintheSocialSciences?”PhilosophyoftheSocialSciences,Volume37,Number2,pp.163-184.2007.\nScientificRealism281MattiSintonen,“Explanation:InSearchoftheRationale”,in:PhilipKitcherandWesleyC.Salmon(eds.),ScientificExplanation.MinnesotaStudiesinthePhilosophyofScience,Vol.13.Minneapolis:UniversityofMinnesotaPress.pp.253-282.1989.MattiSintonen,“FromtheLogicofQuestionstotheLogicofInquiry”,in:Ran-dallE.AuxierandLewisE.Hahn(eds.),ThePhilosophyofJaakkoHintikka.LibraryofLivingPhilosophers.Illinois:Carbondale.2005.DanSperber,andDeirdreWilson,Relevance.Cambridge,Mass:HarvardUniver-sityPress.1986.MauricioSuárez,“ScientificRepresentation:AgainstSimilarityandIsomor-phism”,InternationalStudiesinthePhilosophyofScience17,pp.225-244.2003.FrederickSuppe,TheSemanticConceptionofTheoriesandScientificRealism.Urbana:UniversityofIllinoisPress1989.DeirdreWilsonandDanSperber,“RelevanceTheory”,inG.WardandL.Horn(eds.)HandbookOfPragmatics.Oxford:Blackwell,607-632,2004.JamesWoodward,“ExplanationandInvarianceintheSpecialSciences”in:Brit-ishJournalforthePhilosophyofScience51,2000,pp.197-254.JamesWoodward,MakingThingsHappen:ATheoryofCausalExplanation,Ox-ford:OxfordUniversityPress.2003.G.H.vonWright,ExplanationandUnderstanding,London:Routledge&KeganPaul.1971.G.H.vonWright,FreedomandDetermination,ActaPhilosophicaFennica,Vol.XXXI.1980.PetriYlikoski,“SocialMechanismsandExplanatoryRelevance”,inSocialMech-anismsandAnalyticalSociology(editedbyP.Demeulenaere),CambridgeUniversityPress.Forthcoming.PetriYlikoski,“IllusionsinScientificUnderstanding”,in:DeRegt,Leonelli&Eigner(eds.)ScientificUnderstanding:PhilosophicalPerspectives,Pitts-burghUniversityPress2009.Forthcoming.ProfessorofPhilosophyUniversityofHelsinkiP.O.Box2400014UniversityofHelsinkiFinlandmatti.sintonen@helsinki.fi\n\nDANIELANDLERISNATURALISMTHEUNSURPASSABLEPHILOSOPHYFORTHESTCSCIENCESOFMANINTHE21ENTURY?Jean-PaulSartrefamouslywrote,nearly50yearsago,thatMarxism“remainsthephilosophyofourtime.Wecannotgobeyondit.”InhiscriticRaymondAron’swords,MarxismwasforSartrethe“insurpassable[or,inothertranslations:unsu-perable]philosophyofourtime.1”Takenincontext,Sartre’spronouncementwasatoncedescriptiveandprescriptive:itwas,accordingtohim,neitherobjectivelypossibleforthephilosophertoleavetheconfinesofMarxism,norethicallyper-missibletoattempttodoso.This‘thick’orhybridmodalitywascharacteristicofdialecticalmaterialism:theeventualoverthrowofcapitalism,thedictatorshipoftheproletariat,thesubse-quentdisappearanceofthestate,thesewerestageswhichwereatonceinevitable,andtheproperaimsofpoliticalactionatsuccessivemomentsofthehistoricalprocess.Dialecticalmaterialismwasatonceatheoryofthehistoricalandso-cialprocess,anoverarchingperspective,amethodologyforarrivingatthetruthregardingthesematters,andfinallyapractical(ethicalandpolitical)norm.Theunderlyingnecessitywasmaterial,notmetaphysical:ideaswerethoughttobethenecessarybyproductofobjectiveeconomicconditions,inparticularoftheproductionrelationsandtheaccompanyingclassstruggle.Onemechanismwhichwassupposedtounderlayorimplementthisnecessitationwasaprincipleofideo-logicalreflection:2dialecticalmaterialism,asatheoryorperspective,atrainofthought,wasthoughttobesecretedbytheeconomicset-up,becomingtheideol-ogyoftheproletariat,thusmotivatingitsmemberstoundertaketherevolutionaryactivitieswhichwouldeventuallyleadtotheoverthrowofcapitalism,etc.Indeed,Sartre’squoteaboveistruncated:“[Marxism]remains[...]thephilosophyofourtimebecausewehavenotgonebeyondthecircumstanceswhichengenderedit.”31TheFrenchwordis“indépassable”.2ThelabelisonewhichIamcoiningforpresentpurposes.IamnotaMarxianscholarandaswillbeimmediatelyobviousmygoalinthispaperhasnothingtodowithpoliti-calphilosophyorhistory.Inparticular,althoughIamawareofthedistancebetweenMarxhimselfandlaterformsofMarxismsuchasdialecticalmaterialism,andofdif-ferencesbetweenvariousforms,‘vulgar’andotherwise,ofMarx-inspiredthought,Ihavenousehereforsuchdistinctions.Interestednon-specialistreadersmightliketoconsulthttp://marxmyths.org/.3Jean-PaulSartre,Critiquedelaraisondialectique,vol.1:Questionsdeméthode,Pa-ris:Gallimard1960,p.29.EnglishtranslationSearchforaMethod.NewYork:AlfredKnopf1963,p.30.InFrench:“Il[lemarxisme]restedonclaphilosophiedenotretemps:ilestindépassableparcequelescirconstancesquil’ontengendrénesontpasF.Stadler(ed.),ThePresentSituationinthePhilosophyofScience,ThePhilosophyofScienceinaEuropeanPerspective,DOI10.1007/978-90-481-9115-4_21,©SpringerScience+BusinessMediaB.V.2010\n284DanielAndlerThusSartre,faithfultomaterialism,recognizedatoncetheinevitabilityofacer-taintrainofthoughtatagivenmomentofhumanhistory,anditscontingentchar-acter:theprocessofwhichitwasapartwouldeventuallyleadtoanewsituation,inwhichadifferenttrainofthoughtwouldbecomeavailable(andwouldinfactinevitablybetakenup,thuspresumablybecomingtheinsurpassablephilosophyofthenewepoch).Thisremembranceofthingspastmotivatesthispaper’stitleanditsgeneraldirection,asIwilltrytomakeevidentpresently.ButfirstIneedtomakeperfectlyclearthatIdonotintendtobaseavaluejudgmentonnaturalismontheparallelIamdrawingwithSartre’sversionofMarxism.ItisperfectlyobviousthattherearecontinuitiesbetweenMarxismandcontemporarynaturalism,butIdonotin-tendtodrawthemoutinthispaper.Marxism,especiallyoftheSartriansort,isheldinlowesteeminmanyquartersnowadays,inparticularamongamajorityofcommittednaturalists.Iamemphaticallynotsuggestingthatwhat(atleastuntiltherecenteconomicevents)appearedtomostpeopleashistory’snegativejudg-mentonMarxismhasanybearingoncontemporarynaturalism.Infact,Iwillbedefendingapositionwhichfallsintheballparkof‘liberalized’naturalism.Oneofthedifferencesbetweenmypositionandstrongerorstricterformsofnaturalismconcernsthemodalstatusofthenaturalisticstance,andthisiswheretheparallelwithMarxismcomesin,merelyasaheuristicorexpositorydevice.1.NATURALISM:DESCRIPTIVEANDNORMATIVEWhatisvariouslyknownasscientificorphilosophicalnaturalisminthecontextofcontemporaryanalyticphilosophyappears,atleasttooureyeswhichdonotyethavethebenefitofhindsight,asoneofthosebicentennialgroundswellswhichsweeptheentirephilosophicalscene.Asmanyauthorshavestressed(ithasindeedbecomeanidéereçue,aready-mademorselofphilosophicalconversation),nearlyeveryone(intheEnglish-speakingworld)isanaturalistofsorts.Itturnsoutthattherearedifferentkindsofnaturalism.Somesaythatdifferentphilosophersmeandifferentthingsbynaturalism,4butIprefertothinkthatphilosophershavedif-ferentviewsaboutthenature,structureandscopeofnaturalism,conceivedasaverygeneralstancetowardshumanknowledgeandtheroleplayedbythenaturalsciences.Inthemostgeneralsense,Iseenaturalismastherecommendationthatencoredépassées.”IamindebtedtoThomasFlynnforlocatingtheexactpassagesintheFrenchoriginalandtheEnglishtranslation.Italicsinthetextaremine.4“Itisacommonplacethat‘Naturalismmeansmanydifferentthingstomanydifferentpeople’.”MarioDeCaro/DavidMacarthur(Eds.),Naturalisminquestion.Cambridge,MA:HarvardUniversityPress2004,Editors’introduction,p.3.Theembeddedquota-tionisfromLawrenceSklar,“Naturalismandtheinterpretationoftheories”,in:Pro-ceedingsandAddressesoftheAmericanPhilosophicalAssociation75,2,2001,pp.43-58.\nIsNaturalismtheUnsurpassablePhilosophy?285naturalsciencebetakenwithutmostseriousness.Fromthisasstartingpoint,theroutestomoreevolved,fleshed-outphilosophicalpositionsaremany.Thetwomainbranchesareontologicalormetaphysical,whosecentraltenetisthatnothingtrulyexistsbutwhatthenaturalsciencespurporttoprovideknowledgeabout;andepistemicormethodological,whichtakethenaturalsciencesasaparadigmofallknowledge-seekingactivity.Thesebranchesforkinturn,againandagain,andtheresultingpathscanmeetagainorremaindisjoint.Describingtheresultingandever-changinglandscapeisnotataskwhichIwillundertakehere,althoughIwillhavetoengageinawhileinabitofsimpletaxonomy.5Thepointhere,besidesfixingsometerminology,istostressthephilosophicalnatureofthevariousstrainsofnaturalismthusconstrued:theyareconjecturesregardinghowcertain(abstract)thingsactuallyare,researchprogramsaimingatshowingthemtobethisorthatway,argumentspurportingtoprovideevidencethattheyare,etc.Inotherwords,naturalismisseenasantheoreticalenterprise.Butseeninanotherlight,aswhennaturalistsareengagingtheirphilosophicalopponents,itsometimeslooksasifnaturalismregardsitselfasacause,whoseop-ponents,whetheroutrightantinaturalistsormereskepticsoragnostics,arewrong,notjusttheoreticallywrong,butethicallywrong.Thinkersonbothsidecansoundasiftheybelievedthatitistheirdutytodefend,asthecasemaybe,naturalismorantinaturalism:notmerelyanintellectualduty,butamoralandapoliticaloneaswell.Theothersideisoftenseenascommittingthesinofscientism,or,tothecon-trary,ofantiscientificobscurantism.Thereisasymmetricaldisagreementaboutwhichsideistrulydefendingthehumanityinmankind:Isitthenaturalistwhoisdenyinghumanstheirbasicdignitybyputtingthemonparwith(nonhuman)ani-malsor(biological,naturallyevolved)machines?Orisittheantinaturalistwho,whethershemeansitornot,“objectively”(astheMarxistwouldsay)sideswiththeopponentsofprogressandthedefendersoftheestablished,superstitiousorviolentorder,bypreventingthelightsofsciencefromshiningovertheillsofhu-mankindandbringaboutsomecures?Themilitantnaturalistisalsoacrusader,onewhobelievesthatitwillneces-sarilyturnout,inthelongrun,thattheworldfallsentirelywithinthescopeofthenaturalsciences,andthatthosenotinhiscamp,rangingfromdeclaredopponentstotheuncommittedandweak-kneed,arestandinginthewaythusretardingtheeventualtriumphatthecostofunnecessaryintellectualtoilandhumansuffering.Themilitantanti-naturalistinturnseesthenaturalistasdivestingphilosophicalenergyinadirectionwhichwillturnouttobeeventuallyfruitless,andinthemeantimeasslowingdownherattemptstobringoutthetheoretical,andeventhepracticalconditionsofamorehumaneworld.Itmightbethoughtthatthismilitantstreakisnothingoverandabovethenormalpassionwhichenlivensthetrulyimportantphilosophicaldebates.Granted,5See.e.g..,PhilipKitcher,TheAdvancementofScience.NewYork&Oxford:OxfordUniversityPress1993;DavidPapineau,PhilosophicalNaturalism.Oxford:Black-well1993;DeCaro/Macarthur,ibid.\n286DanielAndlerthedebatearoundnaturalismisnotauniqueoccurrenceofmilitantexcitementinthehistoryofphilosophy.Butnorisitalwaysthecasethatadisagreement,howeverdeep,inphilosophicalorientationisaccompaniedbysuchmilitancy.Anexampleinpoint,inthephilosophyofscience,istheissueofrealism:thedivideisdeep,positionsareentrenched,yetnobodyineithercampbelievesthat,shouldtheopponentprevail,somethingterriblewouldhappentoscience,tophilosophy,ortohumanity.While,asStevenHorstwritesinarecentbook,“somenaturalists[…]claim,inessence,thatthemindmustbenaturalized,orelsesomethingunseemlyfollows”.6Thisisthenormativeformofnaturalismwhichsometimes(notalwaysofcourse)shadowstheepistemic,philosophicallylegitimateformofnaturalism.Normativenaturalism(inthisverypeculiarsenseofthephrase,quitedistinctfromthevariousdoctrinesputforthandillustratedunderthatlabelbysuchauthorsasLarryLaudanorJosephRouse7)hasaprescriptiveandaneschatologicaldimen-sion,justlikedialecticalmaterialism:itwillnecessarilycomeabout,youmusthelpitalong,orelse…Wearenowinapositiontounderstandwhytheproblemofnaturalismcanbeposedontwolevels.First,wecanaskwhethernaturalism(inoneoranotheroftheavailablevarieties)istrue.Second,wecanaskwhetherthequestionmayreason-ablybeasked,orwhether,assomewouldhaveusbelieve,itisthe‘insurpassablephilosophyofourtime’.Thispaperdoesnotproposedirectanswerstoeitherques-tion.Rather,itdefendsastancewhichpermitstopursuethefirstquestionwhiletakingthesecondintoaccount.Iwillpleadneitherfororagainstnaturalisminitsstrongform.Inparticular,asIdon’tthinkthateveninitsstrongform,naturalismiscommittedtoreductionism,Idonotfeeltheneedtoraisetheanti-reductionistflag.NordoIbelievethatnaturalismisnecessarily‘normative’,althoughitisanimportantpartofthecurrentdialecticalsituationthatitsometimesis,andthusIcertainlydon’tthinkitshouldbecombatedassuch.Itshouldbediscussedonitsmeritsasaphilosophicalthesis,andasafruitfulmetaphysicalresearchprogram,inPopper’ssense.8Thereishoweveratensionwhichthereadercannothavefailedtonoticebe-tweentheclaimthatnearlyeveryonetodayisanaturalistandthedescriptionoftheepicstrugglewhichnaturalistswageagainsthardenedopponentsandmeekskep-6StevenHorst,BeyondReduction:PhilosophyofMindandPost-ReductionistPhiloso-phyofScience,Oxford:OxfordUniversityPress,2007,p.15;quotedbyD.GeneWitmer,inNotreDamePhilosophicalReviews,2008:http://ndpr.nd.edu/review.cfm?id=128.WitmerdisputestheaptnessofHorst’scharacterization,arguingthatthenorminquestionisnothingoverandabovetheperfectlyreasonablecommitmentofthehonestscholartobringherandothers’beliefsinlinewithwhatappearsasthebestavailableevidenceprovidedbyscience.7LarryLaudan,“NormativeNaturalism”,in:PhilosophyofScience57,1990,pp.44-59;JosephRouse,HowScientificPracticesMatter.Chicago:UniversityofChicagoPress2002.8KarlR.Popper,RealismandtheAimofScience.London:Hutchinson,1983,§23,pp.89-193.\nIsNaturalismtheUnsurpassablePhilosophy?287tics.Thetensioncanbepartlyrelievedbyrelativizingthenear-universalquanti-ficationtoasubpopulationofphilosophers,andconcomitantlynoticingthattheremainingpopulationisfarfromempty,comprisingatleastanalyticphilosophersofaWittgensteinianbent,andavastmajorityofContinentalphilosophersandsocialscientists.Thereisastruggle,butatthispointintimethenaturalistsseemtohavetheupperhand;whythisisthecaseisaquestionworthasking,andwewillgetbacktoitpresently.Someconsiderabletensionremainshowever,anditisbornebyanincreasinglyvisiblegroupof‘liberalized’naturalists,9philosopherswhosubscribetothemostfundamentalinspirationofnaturalism,yetseemajorob-jectionstosomeofthetenetsofthepuresortofscientificnaturalismwhichseemsprimafacie,andoftenclaimstobe,theonlynaturalisticgameintown.Thereisindeedagnawingsuspicionthatliberalizednaturalismisamoreorlesssubtlydisguisedformofantinaturalism.Thesuspiciondoesnotariseonlyinthecriticalmindofthe‘card-carrying’naturalist:itsitsattheheartoftheliberalizednaturalistherself.10Inwhatfollows,Iwillpresentmyownformofliberalizednaturalism,andtrytoshowthatdespiteitscombininginsightsfrombothsides,itisaprinci-pledposition.ButIdon’texpecttoconvincethereaderthatitisaperfectlystableposition,notbeingfullyconvincedmyself.2.CAUGHTINTHEMIDDLEHereinanutshellisthepredicament.Thefactisthatrightbeforeoureyesnatural-isminspires,andreceivesreciprocalsupportfromfruitfulscientificworkandinter-estingphilosophicalexplorations;moreover,thereissynergybetweenthescienceandthephilosophy,whichrathercomfortsthemetatheoreticalaspectofnatural-ism,theideathatphilosophyandscienceare,astheexpressiongoes,‘continuous’.Thisseemstorequiretheresponsiblephilosophertobenaturalisticenough.Ontheotherhand,generalargumentsinfavorofnaturalism,andspecificproposals,whetherbroadlyphilosophicalorbroadlyscientific,arefarfromentirelysuccess-ful;somegeneralantinaturalistargumentsdoseemtocutsomeice;andthereisatpresentnosignatallthatongoingnaturalisticprogramscanreachmuchbeyondtheareaforwhichtheyaretailored:theassumptionthattheywilleventuallygen-eralizeseemsatthepresentstagequiteimplausible.Ourresponsiblephilosophermustthereforenotbetoonaturalistic.Finally,heshouldnotmerelysteeramiddlecourse,unpalatabletobothsides:heshouldmaketheantinaturalistsideofhispositionacceptableoratleastintelligibletohisnaturalistfriends,andvice-versa,thenaturalistsidecomprehensibletohisantinaturalistfriends.9SuchisthelabelwhichDeCaroandMacarthurclaimforthediversegroupofauthorsintheircollection.10Seetheeditors’attempttojustifytheirattachmenttonaturalismwhilereaffirmingarejectionof“scientific”or“scientistic”naturalism:op.cit.,pp.13-14.\n288DanielAndlerThefair-mindedphilosophermustkeepfirmlyinmindthetwomainfeaturesofthedialecticalsituation.Oneisthat,grantingthattheassessmentjustproposediscorrect,thereisanuneasystand-offbetweennaturalistsandanti-naturalists,withstrongargumentsonbothsides.Thesecondisthateventhenon-naturalistcansenseadissymetryinthedebate:theinitiativeisonthesideofthenaturalists,andtosimplydiscardtheevidenceandtheongoingscientificandphilosophicalworkonwhichtheybasetheircasewouldbeirresponsibleand,yes,somehowwrong.Theantinaturalistscan’tclaimthatignoringtheirevidenceandtheworkinprogressontheirsidewouldbeunreasonableorintellectuallyunethicalinthesameway,becauseallormostofitisofanegativenature.Notonlyarenatural-istsdoingmostofthemovingonthefield,butwhentheylosecontroloftheball(asforexamplewhentheyaresubjectedtodevastatingcriticism,sometimesfromtheirownranks)theypromptlypickitup:thereisacertainself-sustainingrobust-nessintheirgamewhichmakestheirwayofplayingtherightwaytoplay.Thisisperhapsthesenseinwhichnaturalismisindeedthe‘unsurpassablephilosophyofourtime’.Thefair-mindedphilosopher’staskisthusfairlystraightforward:togiveduecredittotheongoingresearchprogramswhichareinspiredby,oraregristtothemillofscientificnaturalism,andyetredressthegamesoastoblocktheunsurpas-sibilitythesis.Itmightbeobjectedthattheresponsiblephilosopherhasamorepressingduty:toarbitrateasbesthecanbetweenthetwosides,andcomeupwithhisownconsideredjudgment,issuinginaverdict.Butthiswouldbepreciselyforcinghischoiceinawayreminiscentofafamiliarmilitanttechnique.Thereisalegitimatethirdwaybetweenthrowingone’sarmsintheairandchoosingsides,whichistoexaminethearguments,findtheminconclusiveandasknot,onceagain,Whoisright,allthingsconsidered?,but,Whatshouldmystancebe,allthingsconsidered?The‘minimalnaturalism’whichIwillproceedtodefendismyanswertothesecondquestion.Naturalism,Isuggestedabove,inthemostgeneralsense,takestheformofarecommendation:Takenaturalsciencewithutmostseriousness.InR.B.Perry’sterms,itis“thephilosophicalgeneralizationofthesciences”.Butwhatdoesthisentail?Nointerestingformofnaturalismstopsatsuchgeneraldeclarations.Whatmakesanavatarofnaturalisminterestingistheproblemsituationitproposes11.It11AsHuwPricehasrecentlybeenarguing(seehis“Naturalismwithoutrepresentational-ism”,in:DeCaro/Macarthur,op.cit.,pp.71-88),theveryfirstandmostbasicchoicemaywellconsistindecidingbetweentwopossibletargetsofthegeneralization:shoulditbethesubjectwhosepositionintheoverallschemeofthingsistobecharacterizedasthatofaninhabitantofnatureassciencerevealsit;ortheobjectswhichwetalkaboutandrepresentlinguisticallywhichmust,asheputsit,placedinthenaturalrealm?InPrice’sownwords,subjectnaturalismassertsthat“Wehumansarenaturalcreatures;humanknowledgeisitselfanaturalphenomenon”,whileobjectnaturalisminsiststhat“whateverexistsexistsinthenaturalrealms”.AlthoughPricemakesnomentionofit,itseemstomethatMcDowell’searlierproposalthatweconceiveof“thinkingastheexerciseofpowerspossessed[...]unmysteriouslybyathinkingbeingitself,ananimal\nIsNaturalismtheUnsurpassablePhilosophy?289startswithacertaincontrastbetweenaclassofprimafacie“natural”entities,andaclassofprimafacienon-“natural”entities.Thefirstclassfixesthereferenceof“nature”atthestartofthegame,thesecondspecifiestheproblemathand.Soforexample,thefirstclassmightbethesetofentitiespostulatedbyourbestcurrentphysicaltheories,andthesecondmightcontainmentalstatesandprocesses.Thecorrespondingnaturalisticstanceisexpressed,ontheontologicalplane,bysomekindofphysicalismaboutthemental,andontheepistemicplane,bythedemandthatpsychology,orcognitivescience,beincludedinthenaturalsciences.Orthefirstclassmightconsistintheontologyofthenaturalsciences,andthesecondmightbethesetofsocialprocesses:naturalisminthissituationmightamountontologicallytotherejectionofanautonomoussphereofsocialentities,andepis-temicallytotherejectionofabifurcationbetweenthenaturalandthesocialsci-ences.Andsoforth:thestructureoftheissueconsistsinthespecificationofClassIandClassII,andtheclaimthatClassIIisinfactincludedinClassI.Afinalcondi-tionforaformofnaturalismtobeworthinvestigatingisthatthisinclusionrelationbenon-trivial:itmustbemootandrequireseriousscientificand/orphilosophicalwork.Dogmaticassertionsoftheform“Everythingis(or:isatbottom)natural”trivializetheproblemanddeservenoconsideration,nomorethansuchcounter-partsas“Everythingis(or:isinthelastanalysis)sociallyconstructed.”Inowcometoadistinctionwhichisnotmadeexplicit,asfarasIknow,intheliterature,perhapsbecauseitistooobvious,perhapsbecausemostphilosophersfocusonjustonesideofthedistinction,and/ortaketheothersideforgranted.Someformsofnaturalism,regardlessoftheirchoiceofClassesIandII,in-cludeaproposedstrategyforestablishingtheinclusionofClassIIinClassI.Duetothenon-trivialitycondition,itisnotonebitobviousthatthestrategywillwork,butthenaturalistmeanstoarguethatinthefullnessoftime,everymemberofclassIIwillbeshown,bysomecleverapplicationoftheproposedstrategy,tobelongtoClassI.ProgramsofthissortareinstancesofwhatIcallanchorednaturalism.Free-floatingnaturalism,bycontrast,consistsinargumentsofageneralnaturepurportingtoestablish,inonefellswoop,theinclusionrelation,sothat,oncetheargumentisaccepted,thereisnoworklefttobedone.Althoughofcoursethereisnothingtopreventaphilosopher,orascientistforthatmatter,fromproposingbothspecificreductivestrategiesandgeneralarguments,sometimesindifferentthatlivesitslifeincognitiveandpracticalrelationstotheworld”isaformofsubjectnaturalism.(QuotetakenfromHilaryPutnam,WordsandLife.Cambridge,MA:Har-vardUniversityPress1994,pp.307-8;thereferenceprovidedistoMcDowell’sAu-gusteComtelecturetotheLSEof2February1993).SeealsoJohnMcDowell,MindandWorld.Cambridge,MA:HarvardUniversityPress,1994,inwhichMcDowelldistanceshimselfwithwhathecalls‘baldnaturalism’.Price’snaturalismisclearlyofthe‘liberalized’sort,andisdevelopedsoastoex-plicitlyreject(object,mainstream,contemporaryscientificorbald)naturalism.WhatIdiscussinthesequelisobjectnaturalisminPrice’ssense;myClassI/ClassIIproblemiswhathecallsa‘placementproblem’.Myownversionofliberalizednaturalism,tobesketchedpresently,is(Ibelieve)compatiblewithPrice’s.\n290DanielAndlerwritingsoratdifferentmomentsofheritinerary,infactphilosopherstendtofallononeortheothersideofthefence.ThusQuine,andthoseworkinginthenatural-izedepistemologytradition,tendtobefree-floaters,whilePutnam,whetherinhisformer,optimisticmoodregardingnaturalism,orinhispresent,pessimisticmood,isinterestedinanchorednaturalism.Andhebecomesimpatientwithfree-float-ingnaturalism,whichhefinds“verypuzzling”,oraformofoutright“Iknownotwhat”.Poincaré’sverdictonarelatedmatterisjustasfinal.Poincaréisconcernedwiththeunityofnature:“Thequestionwemustaskisnotwhethernatureisone,buthowitisone.”12.Inotherwords,theonlyquestionworthasking,if,forexam-ple,weareworriedaboutthe‘imponderablefluids’suchascaloric,phlogistonetc.,isnot,Canwereduceoreliminatethem?,butrather,Howdowedoit?Inotherwords,whatiscalledforishardscientificwork,genuinescientificimagi-nation,possiblyproddedandsupportedbyphilosophy,ratherthanoverarching,armchairconsiderationstotheeffectthatnecessarily,someresearchstrategyorotherwillsucceedinaccomplishingthecalled-fornaturalization.3.METHODOLOGICALNATURALISMThefollowing2-by-2logicaltablemaybethoughttoprovideasimple,naturalwayoflocatingvariousformsofnaturalism:EpistemologyÎYesNoÐOntologyYes1-1(full-bloodednaturalism)1-0(caution,typeB)No0-1(caution,typeA)0-0(full-bloodedanti-naturalism)Table1:AlogicalspaceofattitudestowardsnaturalismbasedonontologyandepistemologyPosition1-1isoccupiedbyfull-bloodednaturalists,suchasavastmajorityofcontemporaryphilosophersofmind,whobelieveboththateverythingbelongstothenaturalorderandthatnaturalscienceisthesolemeanstoacquireatruepictureoftheworld.Position0-0isoccupiedbyfull-bloodedanti-naturalists,whobelieveneither.Position1-0inparticularbythosewhoaredoublycautious,beingwillingneithertopositnon-naturalentitiesnortoruleoutepistemicresourcesoutsidenaturalscience(if,forexample,theydoubtthatnaturalsciencecanprovideafullandfaithfulpictureoftheworld,andtheybelieveorsuspectthatwedohaveagrasp,howeverimperfect,ofcertainaspectsofit,crystallizedincommonwisdom,practicalknow-how,non-naturalscience,literature,etc.).Finally,position0-1isdefendedbyanothertypeofcautiousphilosopher,whoseesnoreasontodeny12HenriPoincaré,Lascienceetl’hypothèse,Paris:Flammarion1902,chap.IX,p.161;myitalics.\nIsNaturalismtheUnsurpassablePhilosophy?291thepossibilityofnon-naturalentities,butconsidersnaturalsciencetobetheonlyvalidwaytoacquiregenuineknowledge.Oneclearlimitationofthistableisthatitdoesnotdistinguishbetweentwowaysofrulingoutthe‘No’answers:positivedenialoragnosticism.Todeclinetobecommittedtothenon-existenceofnon-naturalthings,orofnon-natural-scientificsourcesofknowledge,isonething;tobecommittedtotheexistenceofnon-naturalentities,ornon-natural-scientificsourcesofknowledge,quiteanother.A3-by-3tablewouldremedythisshortcoming.Butanotherlimitationisthatitleavesnoroomforpositionswhichdistinguishbetweencommitmenttoscienceandcommitmenttonaturalscience;a3-by-4tablewouldbeneededtotakecareofthatproblem.ThetableIproposeinsteadtouseinordertolocateintermediatepositions,includingtheoneIwanttodefend,isabitsimpler.Itisbasedonapairofdimen-sionswhichareorthogonaltotheonesusedinTableI.Onthehorizontalaxisfiguresthecontrastclass,viz.what‘natural’iscontrastedwith.Thebasicpos-sibilitiesare:non-naturalequalsnon-physicalornon-material;notaccessiblebyscientificmeans;andnotaccessiblebythemeansofthenaturalsciences.Notethatthecontrastcanbeconstruedeitherontologicallyorepistemicallyorboth.Ontheverticalaxisonefindsthetwobasicstances,orvaluesofcommitmenttowardsthenaturalistthesis:rejectionandacceptance;andinbetween,ahalf-way,cautiousstancewhichIlabel‘methodological’.Contrastclass:Non-physicalNotscientificallyNotnatural-naturalvsÎ(non-material)accessiblescientificallyÐCommitmentaccessibleRejectionSupernaturalismAtheisticpluralismScientificdualism(theologicaldua-(bifurcationism)lism,non-religiousspiritualism)MethodologicalMethodologicalMethodologicalna-Methodologi-stancenaturalism,standardturalism,mysensecalnaturalism,sense(MENA)Chomsky’ssense,oranti-bifurcatio-nismAcceptanceAtheologicalnatu-BroadscientificStrictscientificralism(anti-super-naturalismnaturalism(physi-naturalism)calismorpluralisticnaturalism)Table2:AlogicalspaceofattitudestowardsnaturalismbasedoncommitmentandcontrastclassLetuslookatthefirstline.Theanti-naturalistmayaffirmorrefusetoruleouteithernon-materialentities(column1);orentitiesundetectedorunregimentedbysciencetoutcourt(column2);or,morerestrictively,bynaturalscience(column\n292DanielAndler3).Thefirstcellistypicallydefendedbyreligiousbelieversandbynon-religiousdualists,whobelieveintheexistenceofaspiritualrealmseparatefromthema-terialrealm;thesecond,byphilosophers(andlayfolks)whobelievethereareprovincesofrealitywhichareoutofthereachofsciencealtogether;thethird,byphilosophersandsocialscientistswhobelievethereareprovincesofrealitywhichareoutofthereachofnaturalsciencebutnotinaccessibletoallformsofscience.Occupantsofcell0-0inTable1straddlethefirstandsecondcellsofthisline,whiletheoccupantsofcell1-0straddlethesecondandthird.Thethirdline’sthreecells,bycontrast,neednotbeoccupiedbytheoristswithdifferentbeliefs,butbythinkerswithdifferentopponentsinmind.Inthefirstposition,thenaturalistinsistsonrejectingthenotionofanon-materialrealm(eitherbydiscreditingtheveryidea,orbyarguingthatthepurportednon-materialrealmispartofnatureafterall).Inthesecondposition,thetargetisthepluralistwhobelievesthatscienceleavesoutentitiesorphenomenawhichbelongtotherealworld.Inthethirdposition,themaintargetistheviewofarealmconsistingofmeaning,norms,values,history,culturewhichisseparatedfromthenaturalrealmyetistheproperobjectofstudyofthesciencesofman.Thisthirdpositionisdefendedbythetypicalscientificnaturalist(theoccupantofcell1-1inTable1).ThesecondlineistheoneIwishtodrawattentionto.Ihavedubbedit“meth-odological”toindicateitsintermediatepositiononthecommitmentscale:anydoctrinewhichstopsshortofcommittingtonaturalismortoanti-naturalism,conceivedasclear-cutviewsbasedonpositive(ratherthanmerelyskeptical)ar-guments,belongs,onmytaxonomy,inthemethodologicalgroup.However,thephrase“methodologicalnaturalism”isusedintheliteratureinatleasttwodiffer-ent,morerestrictedsenses.WhatIcallthestandardsenseistheoneusedinthedebateconcerningthepracticeofnaturalsciencebyChristianbelievers.Itisthestancerecommendedtosomeonewhoisbothabelieverandascientist:quabe-liever,heisacommittedsupernaturalist,butquascientisthesuspendshisbeliefinthesupernaturalandconductshisscientificbusinessasifonlythematerialworldexisted.AsMichaelMartinputsit,“inthecontextof[scientific]inquiryonlynatu-ralprocessesandeventsexist”.13QuiteanothermeaningisgivenbyChomskyinhisrecentdiscussionsofnatu-ralism:14forhim,methodologicalnaturalismistherejectionof‘methodologicaldualism’,theviewthathumancognitiveand(inparticular)linguisticprocessesaresubjecttoadualdescription,oneprovidedbythenaturalsciences,theotherbyotherrationalsources.ToChomsky,theoreticalunderstandingofwhateveror-derofphenomenahasbutonesourceandtakesbutoneform,whetherthetopicbetheformationofwaterfalls,thecollisionofelectronsorthehumanmindand13“Justifyingmethodologicalnaturalism”(2002)(availableonlineatwww.infidels.org).14Chomsky,Newhorizonsinthestudyoflangageandmind.Cambridge:CambridgeUniversityPress,2002andsubsequentwritingsdiscussedinteraliabyPierreJacob,“Chomsky’snaturalism:itsscopeandlimits”,forthcoming(Frenchversionin:Chom-sky,specialissueofCahiersdel’Herne,Paris,2007,pp.202-214).\nIsNaturalismtheUnsurpassablePhilosophy?293itsvariouscomponents.Chomsky’sdoctrinehereisatbasenothingparticularlynew–itisinfactthepositionofthetypicaloccupantsofcolumn1inTable1:itconstitutesadetailedandupdatedversionofthemonistviewofthesciencesofman,defendedandattackedforatleast150years15.ButwhileChomskyhasfelttheneed,throughouthiscareer,toreasserthisstandinthefaceofwhatremains(atleastontheContinent)amajorityviewinthehumanandsocialsciences,hismainconcerninhisrecentwritingshasbeentorefutetheontologicalnaturalismwhichisespousedbytheoverwhelmingmajorityofhisfellowcognitivescientists:unlikethem,hedefendscell0-1inTable1.16TheChristianandChomskyancon-strualsofmethodologicalnaturalismoccupythefirstandthirdcellrespectivelyofline2inTable2.Whatwillserveasmystartingpointisthepositionwhichoccupiesthemiddlecell.Methodologicalnaturalism,inmysense(whichIwillabbreviateasMENA),isdistinctfromtheothertwovarietiesinasmuchasitwithholdsbothanegativeandapositivefinaljudgmentonontologicalnaturalism.Itisradicallynon-com-mittalinthatsense,andalsobecauseittakesnostandonthebifurcationthesis:themainopponenthereisnotthedefenderofanichefornon-naturalistichumansci-ence,butonewhodeniesthelegitimacyofascientificapproachtoallsortsof(pri-mafacienon-natural)things.Nogreatleapofphilosophicalimagination,MENAinoneoranotherformisconstantlyre-discoveredbyphilosopherswhoareeitheruncomfortableaboutthemajorityontologicaldoctrineinphilosophyofmind,oruneasyabouttheunderlyingphilosophyofscience,orunnervedbytheunending‘warofmethods’inthesciencesofman,or(likemyself)aboutallthree.IshouldsayrightoffthatIwillnotendupdefendingMENAasIamabouttopresentit,butItakeitasafirstapproximationofthepositionwhichIwillrecommend.MENAisexpressedinthefollowingmaxim:Engageinwhateverinquiry,atanygivenstageofthescientificproblemsitua-tion,isrecommendedbyscientificnaturalismwiththeaimofsecuringaposi-tiveresult,butrefrainfromanycommitment,explicitorimplicit,regardingtheoutcomeoftheinquiry.15Chomskyisemphaticallynotclaimingthatnaturalscienceistheonlysourceofknowl-edgeconcerningmankind:“Someonecommittedtoit[methodologicalnaturalism]canconsistentlybelieve(Ido)thatwelearnmuchmoreofhumaninterestabouthowpeo-plethinkandfeelandactbyreadingnovelsorstudyinghistorythanfromallofnatu-ralisticpsychology,andperhapsalwayswill.”NoamChomsky,“Chomsky,Noam”,in:SamuelGuttenplan(Ed.),ACompaniontothephilosophyofmind.Oxford:Blackwell,1994,p.153.Butnaturalsciencedoeshaveanexclusiveresponsibility,accordingtohim,indevelopingatheoreticalunderstandingofmindandlangage.Thereisnothirdway.16MyinterpretationofChomsky’sontologymightbequestioned,butnothingcentraltothepresentpaperhingesonit;thepointhereisterminological:methdologicalnatu-ralismisusedbyChristiansontheonehand,byChomskyontheother,indifferentthoughrelatedsenses,andthenotionIproposetodefendisyetsomethingelse,albeitinthesameballpark.\n294DanielAndlerBya‘positiveresult’,Imeanademonstrationofthefactthatsomememberoftheclassofprima-facienon-naturalentitiesactuallybelongstotheotherclass.Non-commitmentmeans:noassumptionmadethatthenaturalisticinquiryaboutoneparticularentityorprocessisboundtosucceed,letalonethatthereexistsonestrat-egy,whichweareboundtodiscovereventually,whichwillworkoneachmemberofthenon-naturalclass.Indeed,MENArejectsanycommitmenttoathesisoftheform:everythingisreally,atbottom,natural,ortoanyothermanifestationsoffree-floatingnaturalism.Indeed,inasecundarysense,‘methodological’alsoexpressesanimplicitdemandforamethodorfamilyofmethods.Twoobjectionsmightberaisedrightaway.Thefirstisthewell-knownprecar-iousnessof‘asif’positionsgenerally.While,astheexampleofinstrumentalisminthephilosophyofsciencewouldtendtoshow,thereisnologicalinconsistencyindefendingMENAwhilerejectingontologicalnaturalism,thereisathreatofpragmaticincoherencein(i)adoptingamaximorstrategyorheuristicwhosesuc-cessdependsontheexistenceofcertainentitiesorprocesseswhile(ii)invokingontologicalabstinencewithrespecttothoseentitiesandprocesses.Myansweristhis:MENAencouragespiecemealattempts,onhopefulcandidatesfromClassII,tobringthemintoClassI,butitdoessowithoutpostulatinguniversalin-principlesuccess.Itevensupportsastrategyofconsideringnot-hopefulcandidatesinordertozero-inonwhatgetsintheway.Thehistoryoflogicprovidesanexamplewhichshowsthatthiscanbeagoodstrategy:provingthatcertainnumber-theoreticfunc-tionsarecomputable(orfeasiblycomputable)isagoalwhichonecanrationallypursuewithoutbelievingthatallfunctionsarecomputable(orfeasiblycomput-able).Andfailingtoshowthatafunctioniscomputable(orfeasiblycomputable)isnotnecessarilyfailuretoutcourt;ithelpsonegettheknack,graspapattern,andguesswhetherthecaseathandfitsthepattern.Thesecondobjectionisthatdespiteitsapparentoppositiontofree-floatingnaturalism,MENAactuallycomesperilouslyclosetoit,duetoitsnon-commit-menttoanyparticularnaturalizationstrategy.Incombinationwithitsas-ifcharac-ter,thislackofconstructivecontentmakesittooblandtobeofmuchtheoreticalorpracticalhelp.MENAneedstobestrengthened.Iwilltryandshowthatanintermediatepositionwithalittlemorebitecanbereached,whichistheminimalnaturalismwhichIthinkweshouldcountenance.ButfirstIneedtosayafewwordsabouttheobstacleswhich,asIseeit,blockbothabrutalupgradingorabrutaldowngradingofMENA.3.ONREFORMINGRATHERTHANOVERTHROWINGMENAInthissectionIwilllimitmyselftothenaturalizationofthemind.Thereare,asImentionedattheoutset,otherentitieswhichonemightwanttonaturalize,andwhichinfactarebeingsubjectedtonaturalisticapproaches.Mostofthem\nIsNaturalismtheUnsurpassablePhilosophy?295howeverareconnectedto,ordependentonthenaturalizationofthemindandtheconceptsdeployedtothatend.Iwillbeginwithtwosetsofconsiderationsblockingfullontologicalnatural-ism,thefirstfromphilosophyofscience,thesecondfromphilosophyofmind.Iwillmoveontoargumentsagainstoutrightrejectionofnaturalism.TheupshotwillbethatMENAshouldbebereformed,ratherthanrejectedinfavorofoneortheotherpositionoccupyingthelogicalspaceoutlinedabove.Thediscussionwillperforcebehighlycondensed.a.Limitsfromabove,1:TherecordofcognitivescienceItissometimesthoughtthatcognitivescienceistheeatingproofofthenaturalisticpie.Itcontributesinnosmallmeasuretomakingnaturalism‘unsurpassable’.Itissaidtobe‘naturalizing’themindrightbeforeoureyes.Thankstothelong-awaitedconceptualandinstrumentaltoolswhich,unlikeitspredecessorsinpre-scientificandscientificpsychology,cognitivesciencefinallyhassecured,itshowseverysignofbeingsuccessful.Asweshallseeshortly,thereissomedefinitemerittothisappraisal,butfornowthefocusisonitsmoredubiousparts.Onefamiliarbutdeepphilosophicalquestionbearsonhownaturalisticcog-nitivesciencereallyis.Towhatextenthasitfreeditself,andcanitfreeitself,ofitshybridvocabulary,partintentionalpsychologyplusinformationprocessing,partneuroscienceplusmathematical-physicalmodeling?Tosomephilosophers,thelackofaclear-cutanswertothisquestionprettymuchclosesthecase:cogni-tivescienceisandisboundtoremainanon-naturalscience.Butperhapstheyarelaboringundertoonarrowanotionofthenatural,justasLocke,beforehegot“convincedbythejudiciousMrNewton’sincomparablebook”,17waslaboringundertoonarrowanotionofthephysical.Ratherthanpursuethisdifficultmatter,Ichooseamorepedestrianroute.First,thebareempiricalfactisthatcognitivescienceinitspresentstateofdevelopmentpresentsacharacteristically‘gappy’structure.Itismostclearlysuc-cessfulfor‘inputsystems’(Fodor’s1983expression18),andatleastuntilrecentlyprogresson‘higher’or‘central’processeswaswidelyregardedaslessthanim-pressive.Fodor’sparticularwayofdrawingthecontoursofthegapisoutdated,andtherehasbeen,partlyasaresponsetoFodor’sgrimassessment,awealthofinterestingworkinareas(suchassocialandmoralcognition,emotions,con-sciousness…)formerlyallbutclosedtocognitivescience.Yettheveryabundanceofnewconcepts,paradigms,results,originatingincognitiveneurosciencebutalsoattheinterfaceofdevelopmental,cognitiveandsocialpsychology,evolutionarybiology,anthropology,retroactivelyproveshowdeeplyignorantwewereallalong17Locke,ReplytoStillingfleet,1699,quotedinRobertoTorretti,ThePhilosophyofphys-ics.Cambridge:CambridgeUniversityPress,1999p.75(fulltitleofLocke’sopusculeinTorretti,p.478).18JerryL.Fodor,TheModularityofMind,Cambridge,MA:MITPress1983.\n296DanielAndleraboutcountlessfunctionsofthebrain/mind.Whatweareacquiringrightnowisnotonlynewknowledgeincognitiveandbrainscience,butreasonstosuspectthatwearestilltoday,asitturnsoutwewereyesterday,moreignorantthanknowl-edgeable.Second,thegeneralstrategyfollowedbycognitivesciencesinceitsinceptionistheoneatworkinbiologyingeneral.Itconsistsincombiningatop-downandabottom-upapproach:thefirstidentifiesasetoffunctionsandtheirinterrela-tions;thesecondidentifiesthecorrespondingsetof‘forms’(materialstructuresorsystems)andtheircausalinterconnections.Bottom-upnaturalizationofthemindisthegoalofneuroscience.Top-downnaturalizationisofferedbytwodistinctre-searchprograms:functionalism(theinformation-processingparadigm,alsoknownascomputationalism,classicalorotherwise),andevolutionarypsychology.Fullvindicationofontologicalnaturalismregardingthemindrequiresatriplesuccess:completionofthetop-downanalysis;completionofthebottom-upanalysis;and,importantly,articulationofthetwoapproaches.Inthepioneeringstageofcogni-tivescience,thisarticulationwasthoughttobeprovidedbyanexistenceproof.Thecomputerwasseenasaninformation-processing,mechanicallyrealizedcog-nitiveorgan:howeverdefectiveinitsdetails,themodeldidshowwhatsensecouldbemadeofanarticulationofthetop-downandbottom-upanalyses.Thatwasthewholeidea,thegrandideaoffunctionalistneo-orTuringmechanism,articulatedbysuchfoundersofthefieldasNewellandSimon,MarrandFodor.19Thehopewasthatsuitablecomplexificationofthisparadigmwouldyieldasatisfactorynotionofanaturalmind.Nowwhilesuccesshasbeennotable,thereareconceptualproblems,andreasonstodoubtthatcognitivescienceisonitswaytosolvingthem.Itwouldbefutiletotryandreviewheretheever-expandingcri-tiqueoftheclassicalapproachincognitivescience,butitisworthstressingthatthekeyideaofthearticulationbetweenthetop-downandbottom-upapproachesisunderattack.20Neuro-imagery,evolutionarytheorizing,psychology,mathemati-cal/informationalsimulation,nolongerseemtofittogetherinthewayproposedbyfunctionalism,noristhereatpresentanyclearalternativeframework.Further,19Newell,A.&Simon,H.A.(1976),”Computerscienceasempiricalenquiry:Symbolsandsearch”,in:Comm.Am.Ass.ComputingMachinery,19,pp.113-126;repr.inJohnHaugeland(Ed.),MindDesign.Cambridge,MA:MITPress1981.DavidMarr,Vi-sion:AComputationalInvestigationintotheHumanRepresentationandProcessingofVisualInformation,SanFrancisco:HenryHolt&Company1982.JerryL.Fodor,TheLanguageofThought.NewYork:Thos.Crowell1975;repr.Cambridge,Mass:HarvardU.P.;JerryA.Fodor,Representations:PhilosophicalEssaysontheFounda-tionsofCognitiveScience.Cambridge,Mass.:MITPress,1981.20Seee.g.JaegwonKim,MindinaPhysicalWorld:AnEssayontheMind-BodyProb-lemandMentalCausation.Cambridge,MA:MITPress,1998;LawrenceA.Shapiro,“Reductionism,Embodiment,andtheGeneralityofPsychology”in:H.LoorendeJong&M.Schouten(Eds.),TheMatterofMind.Malden,MA:BlackwellPublishing2006,pp.101-120;DenisMareschaletal.,Neuroconstructivism:HowtheBrainCon-structsCognition.NewYork:OxfordUniversityPress,2007.\nIsNaturalismtheUnsurpassablePhilosophy?297the‘deep’faculties(consciousness,intentionality,spontaneity)seemtoresistbothbottom-upandtop-downnaturalization,despitethenumerousattemptsmadetothisday,andongoing.Themoraltodraw,itwouldseem,isthattheknowledgewehaveacquiredaboutthemind,considerableasitis,hasnotreachedthelevelwherewecancon-fidentlypredictthevindicationofontologicalnaturalismaboutthemind.Andfol-lowing,interalia,Chomsky’srecommendation,thenaturalisticspirititselfrecom-mendsheedingthisconsideration.b.Limitsfromabove,2:TheargumentfromcontextRunningonroughlyparalleltrackswithseveralphilosophersofnote,suchasCharlesTravis,21butwithadifferentstartingpoint,Ihaveovertheyearssketchedacontextualist,orsituated,viewofcognitioningeneral,andofrationalinquiryinparticular.Initself,contextualismappearstoraisenoinsuperableobstacleagainstnaturalism:naturalisticmodelsofcontext-sensitivityareinfactquiteanactiveareaofinvestigation.However,thewayIproposetothinkaboutcontextualityisasanirreduciblynormativedimensionofthought,andthis,itwouldseem,doesgetinthewayofallbutthemosthybridnaturalizationprograms,programswhoseclaimtonaturalnessareextremelydubious.22Ifthereisanymerittomyarguments,theyseemtoimplyseverein-principlelimitationsontheveryideaofprescriptivenaturalmodelsofhighercognitiveprocesses,asopposedtopermissivemodels.Thiswouldbynomeansspelltheendofongoingeffortstospecifysuchpermissivemodels(whichdeterminetheenvelopeoffeasiblecognitiveacts),butitwouldverymuchdulltheontologicalandethicalteethofnaturalism:whoeverdoubtedthattherearenaturalconstraintsonwhatwecanthink?Ofcourse,aswegainempiricalknowledgeandconceptualsharpnessonthesenaturalconstraints,ourpictureofthoughtprocessescanunder-goprofoundrevisions.Buttheywillcontinue,onemightwellreckon,tocontainanon-naturaldimension–secondnature,culture,spontaneity,history,…c.Limitsfrombelow:therecordofcognitivescience,revisited,andthelivelinessofphilosophicalpsychologyDespiteitslimitedsuccess,anddespiteitsshakyfoundations,cognitivescienceisthriving.Itisleadingphilosophyofmindandepistemology(briefly,philosophi-calpsychology)incornerswhichtheyhadn’tvisitedbefore,despitecenturiesofhardwork.Asaresult,ourviewsaboutthemindanditsnaturalunderpinningsareundergoingprofoundchanges.21CharlesTravis,Occasion-Sensitivity:SelectedEssays.NewYork:OxfordUniversityPress,2008.22DanielAndler,“Thenormativityofcontext”,in:PhilosophicalStudies100,2000,pp.273-303;DanielAndler,“Context:thecaseforaprincipledepistemicparticularism”,in:JournalofPragmatics,35,3,2003,pp.349-371.\n298DanielAndlerArguinginfavorofsuchaviewiscertainlynotimpossible.Butfirstitishardtodoinafewsentences,andsecond,itisevenmoreamatterofjudgmentthanofargument.Familiaritywiththefield,notjustitsresultsbutitsinnerproc-esses,itsongoingdiscussions,itsspeculativeenergy,inducesastrongimpres-sionofathrivingresearchprogram.Ofcourse,thesameimpressionwascertainlyconveyedtooptimistic,nottosaygullible,witnessesorparticipantsinscientificprogramswhichhavesincereputedlygonebankrupt,suchasGestaltpsychology,behaviorismorclassicalartificialintelligence.Indeed,aswewillseeinaminute,thisconsiderationplaysacrucialroleinthestiffeningofmyrecommendationforanacceptableformofnaturalism.Perhapsweshouldbecontentatthisjunc-turetorecordasafactthataformofphilosophicalanti-naturalismwhichwouldpurporttoshowthatwhatisgoingonincognitivesciencetodayisessentiallyawasteoftimewouldmeetwithconsiderablescepticism,andthattheburdenoftheproofwouldrestontheanti-naturalist.23Hewouldhavetoshownotsimplythatsomeclaimsareexaggerated,somephenomenalikelytoremainuntouched,sometensionsorevencontradictionsexistbetweensubfieldsandschools,thisorthatstrategyisboundtofail,etc.:hewouldhavetomakeaconvincingcasethatama-jorityofresultsareeitherunsound,oruninformative,oragainrequireacompletereinterpretationinordertofitintosomekindofconceptuallyacceptablepicture.Atallorder.Meanwhile,therealisticspiritagaincommendsrejectingaviewwhichwouldallbutdenycoherenceandfruitfulnesstothisscientificandphilosophicalactivity.Butthereisanegativeargumentwhichtheanti-naturalistcoulduse,andinfactHilaryPutnamhasdevelopeditatlength.24Putnam,asIsaid,rightlytakesseriouslyonlyanchoredformsofnaturalism,and,regardingthemind,heknowsofoneproposal,basedonthefunctionalistschemewhichhehimselfpropounded,andwhichhehasfaultedintheoreticallydeepways.Putnamseemstoconclude,intheabsenceofalikelystand-in,that‘computationalpsychology’,alongwithassociatedattemptsto‘naturalize’themind,isallbuthopeless.Icertainlydon’tmeantochallengePutnam’sdiagnosisoffunctionalism’s‘troubles’.ButIdowanttoquestionwhatItaketobethelogicofhiscaseagainstcognitivescience.Letmestartwithabanalconsiderationfromthehistoryofscience.Weweredoingperfectlyrespectablechemistrybeforequantummechanicscameonthescenetoinformusofwhatthechemicalbondphysicallyconsistsin.Pre-quantum23Mentionshouldbemadeoftworecentcollectionsaimingatcastingdoubtonthevi-abilityofcognitivescience:DavidM.Johnson&CristinaE.Erneling(Eds.),TheFu-tureofthecognitiverevolution.NewYork:OxfordUniversityPress,1997;andCris-tinaE.Erneling.&DavidM.Johnson(Eds.),TheMindasascientificobject.Betweenbrainandculture.NewYork:OxfordUniversityPress,2005.Theybothcontainveryvaluablepapers,andIpersonallysympathizewithanumberofcriticalperspectivesdevelopedthere.Iwouldhowevernotgoasfarastakingthemtobemorethanjustthat:criticalperspectives,whichleavethetargetalive,thoughbruised.24HilaryPutnam,Representationandreality.Cambridge,MA:MITPress1988.\nIsNaturalismtheUnsurpassablePhilosophy?299chemistrywasanaturalisticinquiryintokindsandcombinationsofstuff,anin-quirylegitimateandprogressivedespitenotbeingendowedwith(whatwewouldlaterregardasproper)naturalisticfoundations.Thiskindofox-before-the-cartsituationissurelynotunique.Evolutionarybiologyisanotheroft-mentionedcase;infact,itseemsclosertobeingtherulethantheexceptioninthehistoricaldevel-opmentofscientificdisciplines.Now,whatreasonshavewetothinkthatcognitivescienceis,orshouldbe,anydifferent?Whatiswrongwiththeideathatthisflurryofactivityisprogressiveandlegitimate,despitehavingasyetnosolidfoundation?PerhapsPutnam’s,andothercritics’,reasonsforthinkingitiswrongarethefol-lowing.Cognitivepsychology(andthuscognitivescienceasawhole,insofarasitincludescognitivepsychologyasacorepart),theybelieve,ispredicatedontherepresentationalist-computationalistscheme.Ifthatschemeisincoherent,cogni-tivepsychologycollapses(orrequiresatleastacompletetheoreticaloverhaul).Well,thislineofargumentseemstomeperfectlysoundwhenappliedtothesortofcognitivepsychologyassociatedwithsuchfiguresasNewellandSimon,andwhichwasbarelydistinguishablefromartificialintelligence.(Alotmorewouldneedtobesaidatthispointtodojusticetoboththeprogram,itscritics,anditseventualbreakdown,butthiswouldtakeustoofarafield.)Butitdoesn’tapply,atleastdirectly,tocognitivepsychologyandcognitivescienceintheircontem-poraryform.Thepulltowardsbelievingotherwisemaybeduetoover-relianceonphilosophers’reconstructionsofthecognitiveenterprise.Inconsistenciesorfrailtiesinsuchreconstructionsmaybeduetothefaultymodeling,nottowhatismodeled,viz.theactualscience.Infact,Ihavelongarguedthatalargeproportionoftheworkdoneunderthelabelofcognitivesciencehappilylivesinano-man’slandwherenoduesareowedtofunctionalism,orconnectionism,ordynamicalsystems,ormethodologicalsolipsism,orexternalism,orlearning-theoreticorbi-osemanticnotionsofrepresentation,etc.Thisdoesn’tmean,ofcourse,thatthetheoreticalbasesoftheworkshouldn’tbeactivelysoughtbyscientistsandphi-losophers.Butthereisnogroundforimputingtoit,bydefault,aprovablyfaultyfoundation.Theworkshouldbecriticallyexaminedpiecemealanddirectly,notthroughthelensesofarationalizingphilosopher.d.UpgradingMENAtoMINAIamnowinapositiontostrengthenmyinitialproposalforaformofwell-tem-perednaturalismcompatiblewiththeconstraintsdiscusseduptothispoint.WhatMENAlacksis,Isuggested,amodicumofanchoring.Yetsomeonewhoremainsunconvincedbyanyofthecurrentproposalswillbeloathtoanchoritinaparticu-larnaturalisticstrategy.HowthencanIprovidemyfavoredformofnaturalismwithsomeanchorwithoutanchoringitentirely?Tosolvethislittleriddle,itsufficestogobacktoAI.WhatiswrongwithMENAisshownbythefactthat,atthetime,itwouldnothavehadthetools,northemission,tocriticallyanalyzeAI(bywhichImean‘goodold-fashionedarti-\n300DanielAndlerficialintelligence’orGOFAI,inJohnHaugeland’sterminology25):atthetime,itseemedthatscientificnaturalism’sbestguessastowhatlineofinquirytopursueregardingthemindwasAI,andsoMENAhadnochoicethenbuttofollowsuit.Atthesamemoment,avowedanti-naturalistssuchasHubertDreyfus,andlaterHaugeland,aswellasphilosopherssuchasRobertCumminsandPutnamhim-self26didhavethemeans,andthecouragetodoit,anddidproduceaconvincingandenlighteningcritiqueofAI.Nowofcoursethecurecannotconsistinamendingourdefinitiontoread:Engageinwhateverinquiry,atanygivenstageofthescientificproblemsitua-tion,isrecommendedbyscientificnaturalismwiththeaimofsecuringaposi-tiveresult,exceptifthisinquiryhappenstobeGOFAI,butrefrainfromanycommitment,explicitorimplicit,regardingtheoutcomeoftheinquiry.Notonlywoulditbeuselessandunprincipled,butitwouldalsonotpreventanoth-erpossibledisaster,theuncriticalacceptanceofthenextfadincognitivescience.Forexample,evenwiththeproposedamendment,MENAwouldhavenothingtosayabouttheexaggeratedclaimsmadeonbehalfoffMRI-inspiredresearch.OurresponsiblenaturalistcannotbecontentwithlettingcognitiveneurosciencebasedprimarilyonfMRIandotherbrain-imagingtechniquessimplytakeovercognitivescience.Minimalnaturalism,orMINA,incontrastwithmereMENA,isnotalais-sez(science)-faireatttitude.Perhapsonecouldcharacterizeminimalnaturalismascriticalmethodologicalnaturalism.Onelastconstraintmightputthisdemandinsharperfocus.Whatexperienceseemstosuggestisthatthemindtemptsthosethatstudyit,againandagain,whethertheybephilosophers,computerscientistsorneuroscientists,totreattheempiricalevidencewithlessthanthecareitdeserves,andaltogetherdisregardlargechunksofit.Soperhapsminimalnaturalismshouldbephrasedthus:Engageinwhateverinquiryscientificnaturalismrecommendswiththeaimofsecuringapositiveresult,withoutforegoingacriticalexaminationoftherec-ommendation,andwithdueregardtotheentireempiricalevidence,whetheravailablethroughcommonsense,phenomenology,non-naturalisticorpre-naturalisticscience,oragainscientificexperimentationinthestyleofnatural25JohnHaugeland,ArtificialIntelligence:Theveryidea.Cambridge,MA:MITPress1989.26HubertL.Dreyfus,WhatComputersCan’tDo,NewYork:Harper&Row1972;augm.edition:WhatComputersStillCan’tDo,Cambridge,MA:MITPress1993.JohnHaugeland,1978.“Thenatureandplausibilityofcognitivism”,in:BehavioralandBrainSciences,1,1978,pp.215-226;repr.inJohnHaugeland(Ed.),MindDesign.Cambridge,MA:MITPress1981.RobertCummins,Meaningandmentalrepresenta-tions.Cambridge,MA:MITPress1989.Putnam,op.cit.and“MuchAdoaboutNotVeryMuch.”Daedalus117.1(Winter1988):269-281.Repr.as“ArtificialIntelligence:MuchAdoaboutNotVeryMuch”inWordsandLife(1994),pp.391-402.Seealsotheeditor’sintroductoryandconcludingchaptersinDanielAndler(Ed.),Introductionauxsciencescognitives.2ndéd.,Paris:Gallimard2004.\nIsNaturalismtheUnsurpassablePhilosophy?301science.Andrefrainfromanycommitment,explicitorimplicit,regardingtheoutcomeoftheinquiry.Inanutshellthen,minimalnaturalism(whichisnotlimitedinitsapplicationtothestudyofthemind,thoughIhaveusedthatimportantspecialcaseasaguide)ismethodologicalnaturalismwithphilosophicallywideopeneyes,aworkingphilosophybasedonacloseinteractionwithscientists,onewhichprecludesnei-therthecollaborativeproductionofresults,norcriticismandthepossibilityofrenouncingorreorientingthecollaboration.4.MINIMALNATURALISMINTHESCIENCESOFMANThesciencesofmanincludepsychology,ofcourse,andcognitivesciencecanberegarded,veryroughly,aspsychologypursuedbynovelmeans.Tothatextent,theconsiderationsdevelopedaboveareipsofactoapplicabletothesciencesofman.Yeteveninthiscasesomeinterestingissuesarise.Onemightimaginethatscientificpsychologyispsychologynaturalized,abranchofcognitivescience.Butalthoughthisisnotquitefalse,norisitquitetrue.Notonlyaretherebranchesofpsychologywhichthinkofthemselvesasbothscientificanddistinctfromcognitivepsychology(socialandpersonalitypsychologycometomind),butsomefollowastrictlynaturalscientificmethodology,andsomeevensharetheirsubjectmatterwithcognitivepsychology.Why,forexample,isthereafieldcalled‘mathematicalpsychology’,whichdoesintersectwithcognitivepsychologybuthasnottothisdaybeenabsorbedbyit,andwhosefoundingfathersarenotamongtheheroesofcognitivescience?Thereasonscouldinpartbehistorical,butIbelievethereareother,deeper,conceptuallymoresignificantreasons,havingtodopreciselywithdivergencesconcerningthenaturalisticcharacterofthemethodologies.27Linguisticspresentsanothercase,inwhichnotonlycancognitivescienceclaim,asofnow,nomorethanafractionoftheleadingresearchprograms,buteventhosewerenotuntilfairlyrecentlyuniformly‘naturalistic’inthesenseinuseincognitivesciencetoday.(Todotthe-i-s,thesebranchesoflinguisticshadimpeccablenatural-scientificcredentialswithoutbeingfullynaturalisticatleastinsomesense:theywerestatedinarigorous,formallanguageallowingforcumula-tiveknowledgeandhypothetico-deductivereasoning,yetwerenotconcernedwithprovidingcausalexplanatoryaccountsoflanguage,whetherproximalordistal).Hereagainthereissomedegreeofarbitrarinessandhistoricalcontingencyinthecartographyofthefield,butsomeconceptualissuesarealsoinvolved.27Thehistoricalandconceptualoriginsofthesplitareintertwined.Infact,thehistoricalrecordneedstoberectifiedinorderfortheseconnectionstocometolight.SeeGaryHatfield’srevisionarystudiesinthephilosophicalhistoryofpsychology,e.g.GaryHatfield,“Remakingthescienceofthemind.Psychologyasnaturalscience”,in:C.Fox/R.Porter/R.Wokler(Eds.),InventingHumanScience.Eighteeth-CenturyDo-mains,Berkeley&LosAngeles:UniversityofCaliforniaPress1995.\n302DanielAndlerThesuggestionisthatnaturalismincludes,andsometimesconfuses,twoin-dependentdemands.Thefirstisformal:thetheoreticalstudyofanareashouldbeconductedinwayswhichareformallyinaccordwiththoseof(established)naturalsciencessuchasphysics,andleadtoabodyofknowledgewhoseformiscomparabletotheknowledgeproducedbythosedisciplines.Thesecondiscausalandgenetic:thetheoreticalstudyofanareashouldaimatbringingtolightthecausalstructureofthephenomenaand(inparticular)showhowtheyhavecomeintoexistence.Theseconddemandcanbemetwithoutitbeingclearthatthefirstisaswell;hencethedebate,nowessentiallyover,aboutthescientificstatusofbiology.Butmanyareaswithinthehumanandsocialsciences,includingcognitivescience,canboastahighgradeontheseconddemandandfallexceedinglyshortofsatisfyingthefirst.Thereverseisnotonlyalsotrue,butitraisestheoreticalissueswhichappearnowtobeofcentralimportance.Inthewordsofacontemporarydefender“ofathoroughandstrictnaturalism”,MarkBickhard,“[i]tisdistressinglyeasytoespousenaturalism,butneverthelesstofailinaprojectofnaturalism.[…]Manymodelsfailtobeconsistentwithnaturalisminspiteofthebestintentionsoftheirauthors.[…]Anessentialcharacteristicofanynaturalisticmodelofanyphenom-ena,therefore,isthatitbeconsistentwiththenaturalemergenceofthosephenom-ena.”28Onthatcount,itnowappearsasiffewamongthescientificbranchesofthehumanandsocialsciencesarenaturalistic.Indeed,regardingthesocialsciencesproper(sociology,economics,geogra-phy,anthropology,demography…),thequantitativeandformalsubdisciplineswhicharenotnaturalistic(inthesecondsense)vastlyoutnumberthenaturalisticprogramsandresults.ThisholdsnotonlyunderBickhard’sstrongdefinitionof‘naturalistic’,buteveninalessstringentsense,countingasnaturalisticanaccountwhichprovidescausalmechanisms,atleastinoutline,andforegoingthedemandforaphylogeneticstory(anaccountofhowthephenomenaunderscrutinycameintobeinginthefirstplace,aftertheBigBangandoutofconditionsthenprevail-ing).Howarewetomakesensethenoftheexistenceofnaturalsciencesofmanwhicharenot‘naturalistic’inasenseacceptabletotoday’sscientificnaturalist?Thisisadifficultquestionwhichcannotbefullyexploredhere.Still,wecanmakeuseofourgeneraldistinctions.Thereexistfullynaturalistic,partlynaturalistic,andanti-naturalisticanswers.Thecommitted,up-to-datenaturalistcangooneoftwoways.Themostin-clusivegoalhemightespouseisafullreconciliationoftheformal-quantitativeandcausal-geneticaccounts.Suchwouldbetheoutcomeoncewefindouthowcertainformalstructuresandquantitativerelationscancomeintobeingthrough28Bickhard,M.H.,“CriticalPrinciples:OntheNegativeSideofRationality”,in:W.Herfel/C.A.Hooker(Eds.),BeyondRulingReason:Non-formalApproachestoRatio-nality.Forthcoming.(Myitalics).\nIsNaturalismtheUnsurpassablePhilosophy?303thedevelopmentofincreasinglycomplexsystemsboundbygenerallawsgovern-ingtheircomponentmechanisms.Thisambitiousnaturalistwantseverything,theformalrelationsandthecausalstory,andheregardsclassicalcognitivescienceasanadmittedlyunfinishedyetpromisingquestforsuchunification,whichwillinduecourseprovidetheexpectedarticulation,modeledinaverygeneralsenseaftertheSimon-Marr-Fodorschemeofphysicalrealizationofformalprocesses,althoughpossiblyatvariancewithit.Thenaturalistmightalsobelesssanguineabouttheformaldimension,andbemorephysicalisticallyorbiologicallyinclined.Toher,whatreallycountsarethecausal-emergentfacts.Theformaldescriptionsmightnotbecomeavailable,andiftheydotheymightnotplaymorethananinstrumentalrole.Suchwouldbetheinclinationofquiteanumberofcognitiveneuroscientists,forexample,butmoregenerally,ascientificnaturalistofthatilkwouldholdontothecausal-emergentgoal,andrenouncetheformalgoalifitturnedoutthatitwereunreachablewithinthenatural-scientificperspective.Theanti-naturalistmightaccepttheformal-quantitativeaccounts,andevengrantthemelevatedstatus:hemightforexample(likemoststructuralistsinSar-tre’stime)regardthoseaccountsasexpressingsomeinvariantsproducedorneces-sitatedbyculturalsystemssetupbymankind(Levi-Strausswasuniqueamongthisgrouptoseetheanthropologicalinvariantsasadistantreflectionofthenaturalstructureofthehumanmind).Ontheotherhand,theanti-naturalistwillrefusetogiveprideofplacetothecausal-emergentaccounts,whichhewouldseeatbestasprovidinglimitingconditionsonthecreativepowersofhistoricalhumans.Sofinally,howshouldamethodologicalnaturalist,andmoreparticularly,aminimalnaturalist,viewthesituation?First,hewouldnottakeforgrantedtheeventualfusionofthedomainofformal-quantitativeandcausal-emergentac-counts.Second,hewouldnottakeforgrantedtheeventualregimentationoftheentirefieldofhumanandsocialscience(orofanyofitsbranches)underthefor-mal-quantitativebanner.Third,hewouldnomoretakeforgrantedtheeventualtriumphofthecausal-emergentapproach.Infact,anyofthethreepossibilitieswillappeartohimasunlikely,inthelightofpastexperience,presentachieve-ments,andprospects.LikeOttoNeurath,29hewouldcallforan‘orchestration’ofthesedifferentapproaches,andwiththatbest-possibleoutcomeinmind,hewouldsupport,examineandpossiblycontributetowhateverresearchprogramisrecom-mendedbyscientificnaturalism.Thiswouldbehiswayofrecognizingnaturalismastoday’sunsurpassablephilosophyforthesciencesofman.UniversitédeParis–SorbonneEcolenormalesupérieureInstitutuniversitairedeFrancedaniel.andler@paris-sorbonne.fr29OttoNeurath,PhilosophicalPapers1913–1946.Dordrecht:Reidel1983.\n\nANTÓNIOZILHÃOWHATDOESITMEANTOBEANATURALISTINTHEHUMANANDSOCIALSCIENCES?ACOMMENTONDANIELANDLER’S“ISNATURALISMTHEUNSURPASSABLEPHILOSOPHYFORTHESCIENCESOFMANINTHETWENTY-FIRSTCENTURY?”Intheversionofthepaper“IsNaturalismtheUnsurpassablePhilosophyfortheSciencesofManintheTwenty-firstCentury?”DanielAndlersentmeafewdaysago,heputsforthapositioninthesciencesofManhestartscallingliberalizednat-uralism.Inthecourseofthepaper’sdevelopment,however,Andler’sownbrandofliberalizednaturalismisfurtherclarifiedasminimalnaturalism.Furtheron,hecharacterizesminimalnaturalismasmethodologicalnaturalismwithphilosophi-callywideopeneyes.Thisisthecomplextermheendsupselectingasthedesigna-torofthepositionhewantstomarkout.Whatisthenmethodologicalnaturalismwithphilosophicallywideopeneyes?Andlerpresentsushispositionintermsofacontrastwiththreeotherposi-tions,namely,thepositionshecalls‘anti-naturalistic’,‘partlynaturalistic’,and‘fullynaturalistic’.HetellsusthenthatwhatdistinguishesthesepositionsinthesciencesofManisthekindofapproachtheyfavourtotheobjectthesesciencesstudy:theanti-naturalisticpositionascribesaprivilegedstatustoaformalap-proach,thepartlynaturalisticpositionascribesaprivilegedstatustoacausalap-proach,andthefullynaturalisticpositionwantstoreconciletheformalwiththecausalapproaches.MethodologicalnaturalismwithphilosophicallywideopeneyesisthencharacterizedbyAndlerasapositionthatdeemstheunilateralsuc-cessofanyofthesethreepositionstobehighlyunlikelyandthatthereforecallsforacombinationofalloftheminordertoobtainresultsthatmightcontributetostrengthentheresearchprogramofontologicalnaturalism.Thetermontologicalnaturalismis,inturn,definedasexpressingaformofcommitmenttowardsthenaturaliststance,namely,fullacceptance.Finally,thenaturaliststanceisdefinedbothintermsoftheinjunctionTakenaturalsciencewiththeutmostseriousnessandintermsofthecommitmenttosomeformofreductionoftherealmofthenon-naturalintotherealmofthenatural.Now,Isupposethat,inthecontextofthiscolloquium,thequestionIamre-quiredtoansweristhefollowing:doIagreewithAndler’sbrandof‘liberalizednaturalism’?Well,itisdifficulttogivea‘yes’or‘no’answerrightaway.Ifeelthatsomegroundneedstobeclarifiedfirst.Thus,Iwillpostponemyanswertotheclosingpartofmycomment.F.Stadler(ed.),ThePresentSituationinthePhilosophyofScience,ThePhilosophyofScienceinaEuropeanPerspective,DOI10.1007/978-90-481-9115-4_22,©SpringerScience+BusinessMediaB.V.2010\n306AntónioZilhãoLetmebeginbyidentifyingtwoaspectsinAndler’spaperwithwhichIexpe-riencemajordifficulties.Inthefirstplace,itisdifficultformetomeseehowblendingnaturalisticwithanti-naturalisticviewsonthesciencesofMan,asAndlerencouragesustodo,mightcontributetodefineaphilosophicallycoherentorstableposition.ThismightnotbeaproblemifAndlerwerejustputtingforthapragmaticapproachtothefieldofresearchand,consequently,refusingtotakesidesinthedisputebetweendifferentphilosophicalwaysofmakingsenseofthefirstorderknowledgeactuallyproducedbythesciencesofMan;buthisversionof‘liberalizednaturalism’,nomatterhowliberal,issupposedtobeaformofphilosophicalnaturalism,namely,aformwhichimpliesfullacceptanceofthenaturaliststance,andnotaformofsuspensionofphilosophicalbelief.Ifindthisperplexing.However,Iwillnotpur-suethisissuehere,sinceIfindtheuseofthelabels‘fullynaturalistic’,‘partlynaturalistic’and‘anti-naturalistic’inthiscontexthighlyconfusing.Secondly,itisnotatallcleartomethatthestrengtheningoftheresearchprogramofontologicalnaturalisminthesciencesofMan,asdefinedbyAndler,mightpossiblybeachieved;thus,itisdifficultformetoseehowcouldonepos-siblycontributetosuchastrengthening.Letmebelabourthispointalittlebit,asIthinkitisthemostrelevant.AsImentionedabove,liberalizednaturalistsoftheformAndlerspecifies,arerequiredtostrengthentheresearchprogramofontologicalnaturalism.Thisresearchprogramis,inturn,characterizedbyafullacceptanceofthenaturaliststance.Theacceptanceofthisstanceis,inturn,characterizedbythefulfilmentofthetwoabovementionedrequirements.Ihavetroubleswiththefulfilmentofanyoftheserequirements.HereiswhatIfindtroublingwiththefulfilmentofthefirstrequirement.TheformulationoftheinjunctionTakenaturalsciencewiththeutmostseriousnessseemstometoimplythatthereisorthatthereshouldbeasinglefieldofresearchcalled‘naturalscience’,themethodofwhichisorshouldbeunifiedandtranspar-ent.ButIamnotsurethatthereisorthatthereshouldbeamethodofnaturalsci-enceoverandabovethemethodsofthedifferentnaturalsciences.Thereare,ofcourse,somegeneralstandardsconcerning,forinstance,sever-ityoftesting,waysofmakingsurethattheevidenceisdealtwithimpartially,ordefinitionofconstraintsonwhatmaycountasgoodevidence,whicharecommontoallnaturalsciences.Buttheseseemtometobefairlygeneralstandards.Inpar-ticular,Idonotfinditatallself-evidentthatmanyanthropologists,historiansorlinguiststhatconsiderthemselvestobesidingwithinterpretivismratherthanwithscientificexplanationmightnotacceptthesegeneralmethodologicalstandardsastheirown.Asamatteroffact,lotsofthemdo.However,if,inordertoexcludethemfromthesetwewanttodefine,wetrytomakethemethodologicalcharac-terizationofgeneralnaturalsciencemorespecific,wewillbeboundtorealizethatnaturalsciencesarelessunifiedthanwetendedtothink.Wewillbeboundtorealize,forinstance,thatsomeperfectlyacceptablenaturalsciencesdonotlive\nWhatDoesitMeantoBeaNaturalist…?307uptothestandardsofwhatisusuallyconsideredtobetherolemodelofnaturalsciences,namely,physics.Oncewerealizeandacceptthatsuchastateofaffairsisindeedthecase,itwillbehighlyproblematictojustifyadiscriminationagainstsomeexceptionsbutnotagainstothers.Andifwedonotdiscriminateagainstanyofthenaturalsci-entificexceptions,wewillhaveahardtimejustifyingourdiscriminationagainstthosenon-naturalsciencesthatabidebytheverygeneralmethodologicalscientificstandardsmentionedabovebutdonotpartakeofsomeofthemorespecificmeth-odologicalprinciplesfollowedby,e.g.,physicists.This,Ithink,isapointJerryFodormadealongtimeago.Letmenowdealwiththesecondrequirement.Thecommitmenttoontologicalreductionseemstometoimplythatitalwaysmakessensetotry,atleast,toreducetheentitiesandpropertiesthesciencesofMantalkabouttoentitiesandpropertiesnaturalsciencestalkabout.However,asitstands,thisthesisseemstomenottobequitetrue.Letmeintroducemypointthroughtheconsiderationofthefollowingstandardexample.Considerapredictionprovidedbystandardeconomictheoryaccordingtowhich,inanormalmarketeconomy,loweringinterestrates,undersomerelevantconditions,appropriatelyspecifiable,boostsprivateinvestmentandthusfacilitateseconomicgrowth.Letusassume,forthesakeoftheargument,thatpredictionssuchasthisonetendtoturnouttruemoreoftenthannotandthatwefeelconfidenttoprovideanexplanationforthesurgeofeconomicgrowthintermsofthelower-ingoftheinterestrates(ceterisparibus,ofcourse).Howarewetocountenanceaseriousreductionoftheentitiesreferredtoinanexplanationbelongingtothismodeofdiscoursetonaturalscientificentitiesundersomesuitabledefinitionofwhattheseare?Theobviousrouteistobringbackthetermsoccurringintheabovementionedexplanationtoverycomplexdescriptionsoftheunderlyingnaturalfacts(Iwillnottackleherethemorespecificquestionofdeterminingwhatkindofreductiverelationistheterm‘bringingback’supposedtoindicateinthiscontext).Needlesstosay,underlyinganyeconomicfactstherearemillionsofhumanbeingsrelatingtoeachotherincertainways.Theserelationsareinturnsomehowmanagedbytheirbrains;thesebrainsareinturncontinuouslyinthebusinessofprocessingelectro-chemicallycomplexvisualorauditorystimuli;theoutcomeofthecer-ebralprocessingisinturnsomehowtransducedintoelectrochemicalimpulsesthatmakethebodiesassociatedwiththesebrainsbehaveinsomespecificways,andsoonandsoforth.Questionsofpracticalfeasibilityapart,Isupposethatitshould,atleastinprinciple,bepossibletodescribeeventssuchasthe‘loweringoftheinter-estrates’orthe‘upsurgeinprivateinvestment’intermsbelongingexclusivelytothelevelsofdiscoursewithinwhichwetheorizeabouttheseunderlyingnaturalphenomena.Now,ifthenaturalist’saimweresimplytorejecttheideathattheremightbesomespookysupernaturalentitiesgoverningthewholeeconomicprocess,high-\n308AntónioZilhãolightinginthiswaytheunderlyingpsychobiophysicalcomplexityhiddenbehindtheentitiesandpropertiestalkedaboutineconomicsmightindeedbeanusefulandilluminatingreductivestrategy.ButItakeitthat,understoodinthisway,thenaturalisticreductionwemightachieveisofanotparticularlyinformativekind.Actually,inhispaper,Andlerhimselfmentionsthissortofreductionasbeingtriv-ial.Thisthenseemstomeanthattheideaassociatedwiththeontologicalreductionaimedatinthestrengtheningoftheprogramofontologicalnaturalismshouldbeunderstoodinsomestrongerandlesstrivialsense.Whatmightthissensebe?Itakeitthat,inordertobephilosophicallyandepistemologicallymeaning-ful,anysortofreductionworthundertakinghastosatisfythedesideratumofbe-inginduciveofexplanatoryprogress.AndhereiswhereIthinktheproblemlies.Forevenifwesupposethatwemightbeable,perimpossibile,toachieveafulldescriptionoftheabovementionedprocesseswhichwouldreferonlytosomesortoffullynaturalizedentitiesandtheirproperties,suchadescriptionwouldpresumablyleaveuscompletelyinthedarkaboutthequestionofwhytheentireeconomicsystembehavesinthewayitdoes.Asamatteroffact,itseemstomethatmostoftheinformationcontainedinsuchanoverwhelminglycomplexde-scriptionwouldbeexplanatorilyuseless.Atthesametime,italsoseemstomethatalotofusefulandpotentiallyexplanatoryinformationprovidedbyhithertounre-duceddisciplinessuchas,e.g.,history,socialpsychologyoranthropologywouldbecompletelyleftoutofthepicture.IfIamrightonthisaccount,thenitseemstofollowthatitisnotthecasethatitalwaysmakessensetotry,atleast,toreducetheentitiesandpropertiesthesciencesofMantalkabouttoentitiesandpropertiesnaturalsciencestalkabout.Asamatteroffact,wecanidentifyapatternherewhichisactuallymoregeneralthantheproblemofreducingthesciencesofMantothenaturalsciences.Considerforinstancethecaseofevolutionarybiology.JohnMaynardSmithandRichardDawkins,twohard-nosednaturalistsinanydecentaccountoftheterm,havebothdefendedaviewoftheirsubjectaccordingtowhichtherightdirectiontofollowinbringingbacksomeofthecentralconceptsofevolutionaryandde-velopmentalbiologytosomeotherconceptssimultaneouslymorebasicandmoreexplanatoryisthedirectionofinformationandcommunicationtheory,notthedirectionofphysicsormicrophysics.Dawkinstalksofgenesbeing‘longstringsofpuredigitalinformation’andMaynardSmithtalksofgenesbeing‘symbols’andstressesthefactthattheuseofinformationaltermsinbiologyisnotmetaphoricalbutliteral,inthatitimpliesintentionality,thepropertynineteenthcenturyphiloso-phersidentifiedasbeingthemarkofthementalasopposedtothephysical.Ofcourse,weknowthatDawkins’sviewofthegenesasitemsofdigitalinfor-mationencapsulatedinwetwareorMaynardSmith’sviewofnaturalselectionasaproviderofbiologicalintentionalityintolivingstructuresinnowayconflictwiththethesisofthecausalclosureofthephysicalrealm.Butthisisnotmypoint.Mypointissimplythatjustasthedirectionofexplanatorilyilluminatingintra-naturaltheoreticalconnectionsisnotdeterminedbeforehandbyapreviouslygivenhierar-\nWhatDoesitMeantoBeaNaturalist…?309chyoflevelsofnaturalscientificitystartinginMicrophysics,anexplanatorilyillu-minatingreductionofahumanorsocialsciencedoesnothavetobeundertakeninthedirectionofsomenaturalscienceorother.Conversely,Iseenoapriorireasonwhyareductioninthedirectionofsomenaturalscienceorother,evenifpossible,wouldhavetobeexplanatorilyilluminating.Now,Andler’snon-trivialityconditionandhiscriticismoffree-floatingnatu-ralismexpressconcernssimilartothoseIhavejustmentioned.Andthesecon-cernsseemtounderlyalsotheinjunctionwithwhichAndler’sterminateshisfinalcharacterizationofminimalnaturalism,namely,theinjunction“refrainfromanycommitment,explicitorimplicit,regardingtheoutcomeoftheinquiry”.Butthenitishardtoseehowbeingsoradicallynon-committalregardingtheoutcomeoftheenterpriseofontologicalreductionisactuallycompatiblewiththeinjunctionthattheresultsofyourinquiryshouldcontributetostrengthentheresearchpro-gramofontologicalnaturalism.Howcanyoucontributetostrengthenaresearchprogrambyrefrainingfromcommittingyourselftooneofthetwobasictenetsintermsofwhichthisresearchprogramisdefined?Andlerhimselfdetectsathreatof‘pragmaticincoherence’here,butheclaimsthathisfinalformulationofminimalnaturalismsolvesit.Idonotseehow.GiventhecriticismsIvoicedabove,wheredoIstandtheninthedisputethatrevolvesaroundphilosophicalnaturalismandthesciencesofMan?FromwhatIhavealreadyargued,itseemstofollowthat,accordingtomystandpoint,ifthereissomethingspecificcharacterizinganaturalisticviewofthesciencesofMan,thatisneithertheideathatthesciencesofManshouldapethespecificmethodsofanyparticularnaturalsciencenortheideathattheentitiesandpropertiestheytalkaboutshouldbesomehowontologicallyreducibletoentitiesandpropertiessomenaturalscienceorothertalksabout.So,where,ifanywhere,liesthespecificityofanaturalisticapproachtothesesciences?Myanswertothisquestionisthefollowing.Iamconvincedthattherightwaytoconstruethenaturalismversusnon-naturalismdebateinthehumanandsocialsciencesisintermsofamethodologicaldisagreement.Andthisdisagreementcon-cernstheconceptofexplanationeachoftheperspectivesendorses.Accordingtomystandpoint,whatdistinguishesnaturalistsinthisdisputeisthefactthattheyconsidercausalexplanationtobetherightsortofexplanationhumanandsocialscientistsshouldstrivetoprovideintheirwork.Ontheotherhand,non-natural-ists,suchas,e.g.,interpretivists,associateexplanationwithrationalreconstruc-tion,understanding,roleidentification,ortheestablishmentofsyntheticaprioriprinciplesthataresupposedtobeconstitutiveofthedomainofthehuman.Thisstandpointis,ofcourse,notparticularlynew.Itgoesbackto,atleast,someoftheremarksmadebyC.G.Hempelinhisfamousessay“AspectsofScientificExpla-nation”andhasbeenendorsedbyanumberofdifferentauthorssincethen.Finally,letmenowreturntomyoriginalquestion.ItshouldbeclearbynowthatIsympathizewithAndler’sattempttoavoidcommittinghimselftoontologi-calnaturalism.ItisalsoclearthatIdisagreewiththewayhecharacterizesthe\n310AntónioZilhãogeneralfeaturesofthetheoreticallandscapeagainstthebackgroundofwhichthenaturalismdisputeinthesciencesofMantakesplace.Givenhischaracterizationofthislandscape,hisownversionofliberalizednaturalismsoundstooclosetobeingcontradictory.However,againstthebackgroundofmyowncharacterizationoftherelevanttheoreticallandscape,thepositionsummarizedinhisfinaldefini-tionlosesitsparadoxicalaspect.Asamatteroffact,itappearsquitesensibleandjustified.BIBLIOGRAPHYRobertAxelrod,TheEvolutionofCooperation.NewYork:BasicBooks1984.TylerBurge,“Mind-BodyCausationandExplanatoryPractice”,in:JohnHeil/AlfredMele(Eds.),MentalCausation.Oxford:ClarendonPress1993.PaulChurchland,“TheLogicalCharacterofAction-Explanations”,in:ThePhilo-sophicalReview79,1970,pp.214-36.PaulChurchland,“EliminativeMaterialismandthePropositionalAttitudes”,in:TheJournalofPhilosophy78,1981,pp.67-90.JamesColeman,FoundationsofSocialTheory.Cambridge(MA):TheBelknapPressofHarvardUniversityPress1990.DonaldDavidson,EssaysonActionsandEvents.Oxford:OxfordUniversityPress1980.DonaldDavidson,“ParadoxesofIrrationality”,in:RichardWohlheim/JamesHopkins(Eds.),PhilosophicalEssaysonFreud.Cambridge:CambridgeUni-versityPress1982.DonaldDavidson,InquiriesintoTruthandInterpretation.Oxford:ClarendonPress1984.DonaldDavidson,“CouldThereBeaScienceofRationality”,in:InternationalJournalofPhilosophicalStudies,3,1995,pp.1-16.RichardDawkins,RiverOutofEden:ADarwinianViewofLife.NewYork:BasicBooks1995.DanielDennett,TheIntentionalStance.Cambridge(MA):TheMITPress1988.WilhelmDilthey,EinleitungindieGeisteswissenschaften.VersucheinerGrund-legungfürdasStudiumderGesellschaftundderGeschichte–GesammelteSchriften,Band1.Göttingen:VandenhoekandRuprecht2008.WilliamDray,“ExplanationinHistory”,in:JamesH.Fetzer(Ed.),Science,Expla-nation,andRationality–ThePhilosophyofCarlG.Hempel.Oxford:OxfordUniversityPress2000.JonElster,ExplainingSocialBehaviour:MoreNutsandBoltsfortheSocialSci-ences.Cambridge:CambridgeUniversityPress2007.JerryFodor,“SpecialSciences(or:TheDisunityofScienceasaWorkingHypoth-esis)”,in:Synthese,28,1974,pp.97-115.\nWhatDoesitMeantoBeaNaturalist…?311JerryFodor,“MakingMindMatterMore”,in:J.Fodor,ATheoryofContentandOtherEssays.Cambridge(MA):TheMITPress1990.DagfinnFøllesdal,“TheStatusofRationalityAssumptionsinInterpretationandintheExplanationofAction”,in:Dialectica,36,4,1982,pp.301-316.PeterGardiner(Ed.),ThePhilosophyofHistory.Oxford:OxfordUniversityPress1974.CliffordGeertz,TheInterpretationofCultures.NewYork:TheBasicBooks1973.CarlGustavHempel,“AspectsofScientificExplanation”,in:CarlGustavHempel,AspectsofScientificExplanationandotherEssaysinthePhilosophyofSci-ence.NewYork:TheFreePress1965.MartinHollis,ThePhilosophyofSocialSciences:AnIntroduction.Cambridge:CambridgeUniversityPress1995.PhilippKitcher,“ReasonablePeople”,in:JamesH.Fetzer(Ed.),Science,Expla-nation,andRationality–ThePhilosophyofCarlG.Hempel.Oxford:OxfordUniversityPress2000.MichaelMartinandLeeMcIntyre(Eds.),ReadingsinthePhilosophyofSocialScience.Cambridge(MA):TheMITPress1994.JohnMaynardSmith,“TheConceptofInformationinBiology”,in:PhilosophyofScience,67,2000,pp.177-194,.DavidPapineau,ForScienceintheSocialSciences.London:MacMillan1978.DavidPapineau,PhilosophicalNaturalism.Oxford:Blackwell1993.WesleySalmon,CausalityandExplanation.Oxford:OxfordUniversityPress1998.WesleySalmon,FourDecadesofScientificExplanation.Pittsburgh(PA):TheUniversityofPittsburghPress2006.GeorgHenrikvonWright,ExplanationandUnderstanding.Ithaca(NY):CornellUniversityPress1971.PeterWinch,TheIdeaofaSocialScienceanditsRelationtoPhilosophy.London:Routledge,1958.DepartamentodeFilosofiaFaculdadedeLetrasdaUniversidadedeLisboaAlamedadaUniversidade1600-214LisboaPortugalAntonioZilhao@fl.ul.pt\n\nPartV(TeamD)(teamleaderDennisDieks)PhilosophicalFoundationsofthePhysicalSciences\n\nDENNISDIEKSREICHENBACHANDTHECONVENTIONALITYOFDISTANTSIMULTANEITYINPERSPECTIVEABSTRACTWetakeanotherlookatReichenbach’s1920conversiontoconventionalism,withaspecialeyetothebackgroundofhis‘conventionalityofdistantsimultaneity’thesis.WearguethatelementsofReichenbachearlierneo-Kantianismcanstillbediscernedinhislaterworkand,relatedtothis,thathisconventionalismshouldbeseenassituatedatthelevelofglobaltheorychoice.ThisiscontrarytomanyofReichenbach’sownstatements,inwhichhedeclaresthathisconventionalismisaconsequenceofthearbitrarinessofcoordinativedefinitions.1.INTRODUCTIONThehistoryofthephilosophyofphysicshasbeenshapedbyacomplicatedandfascinatinginterplaybetweenphysics,philosophicalideasandexternalfactors.Thishistoryisnotonlyaintriguingsubjectforstudyinitsownright:histori-calconsiderationscanalsoshedlightonthecontentofdoctrinesputforwardbyphilosophersandarerelevantfortheappraisalofsuchdoctrines.Theaimofthispaperistoillustratethisgeneralpointbyacasestudy,namelytheintroductionbyHansReichenbachofthenotoriousConventionalityThesisregardingsimultaneityinrelativitytheory.Thereisan,admittedlyold-fashioned,standardloreconcerningthehistoryoftheconventionalitythesisthatgoesmoreorlesslikethis.Attheendofthe19thcenturyempiricisminthephilosophyofphysicshadgotthewindinitssailsasaresultoftheworkofErnstMach,andthisempiricistMachianatmospheredecisivelyinfluencedEinstein’sthinking.Inhis1905paperthatestablishedspe-cialrelativityEinstein1accordinglyadoptedanempiricistandevenoperationaliststance.Inparticular,whenEinsteindiscussedspatiotemporalnotionshedeclaredthatinordertomakesuchconceptsphysicallymeaningfulwehavetoendowthemwithconcretephysicalcontentintermsofmeasuringprocedures.Forexample,inthecaseoftimeataparticularplacethesoughtdefinitionoftime(Einstein’sterm)1AlbertEinstein,“ZurElektrodynamikbewegterKörper”,in:AnnalenderPhysik17,1905,891-921.ThestandardEnglishtranslationisin:H.A.Lorentz,A.Einstein,H.MinkowskiandH.Weyl,ThePrincipleofRelativity.London:Methuen1923(repub-lishedasaDoveredition,NewYork:Dover1952).F.Stadler(ed.),ThePresentSituationinthePhilosophyofScience,ThePhilosophyofScienceinaEuropeanPerspective,DOI10.1007/978-90-481-9115-4_23,©SpringerScience+BusinessMediaB.V.2010\n316DennisDiekscanbegivenas“thepositionofthehandsofaclocksituatedatthesamespot”.Timethusdefinedisapurelylocalconcept,however,sothatweneedafurtherdefinitiontocomparetimesatdifferentplaces.ForthisreasonEinsteinfamouslyaskedhimselfhowtosynchronizeclocks.Heansweredthisquestionbyassert-ingthatsimultaneityisbydefinition(italicsusedbyEinstein)achievedwhenallclocksaresetsuchthatthevelocityoflight,measuredwiththeirhelp,isthesameinalldirections.Withthis,thecharacterizationoftimeinaframeofreferencebe-comescomplete:givenonestandardclock,itstimecanbepropagatedeverywherebymeansofthesimultaneityrelation.Einsteinemphasizedthewords‘bydefinition’inhisdescriptionofthesyn-chronizationprocedure.Indeed,thetemporalandspatialnotionsintroducedear-lierinhisarticledonotyetfixthesimultaneityrelation,duetothefactthatwecannotdeterminethespeedofanysignalifwearenotyetabletocomparetimesatdifferentlocations.Ifwedidknowthespeedofsomesignal,thatoflightforexam-ple,wecouldsimplysynchronizeclocksbysendingalightsignalfromoneclocktoanotherandbytakingintoaccountthatthissignaltakesatimeL/ctoreachitsdestination(withLthedistancebetweentheclocksandcthespeedoflight).Thesituationbeingwhatitis,however,itseemsthatweneedtostipulateasynchroni-zationprocedurethatfixesbothsimultaneityandthespeedoflight.Stipulationscannotbetrueorfalse,sosimultaneityandthevalueofthespeedoflightcomeoutofthisanalysisasnothavingafact-like,butratheraconventionalcharacter.Thislatterstatementshouldnotbeinterpretedinthetrivialsensethatwehavetochooseunitsfortimeandlengthbeforewecansayanythingaboutthevalueofthespeedoflight:evenafterwehavemadeachoiceforsuchunitsitisstillundecidedwhatthespeedoflightalonganygivendirectionis.Foralthoughitistruethatwecanmeasuretheround-tripvelocity,bydetermininghowmuchtimeittakesforthelighttotravelfromclockAtoclockBandbackagain,thiswillnottellushowmuchtimewasneededtogooneway,fromAtoB.Inparticular,itisimpossibletoestablishthatthetoandfrolightspeedsbetweenAandBareequal.Ifthesethingscanonlybestipulated,thenitshouldalsobepossibletomakeotherchoiceswithoutcomingintoconflictwiththefactsalreadyfixedbypriordefinitions.Thispointwasworkedoutinaphilosophicallyprecisemanner,thestandardstorycontinues,byHansReichenbach,especiallyinhisepoch-makingbookThePhilosophyofSpaceandTime(1928)2.Reichenbachtheresubsumedhisinvestigationofsimultaneityunderageneralanalysisofthestatusofphysicalnotions,accordingtowhichallindependentconceptsshouldbecoordinatedtoconcretephysicalthingsandproceduresbymeansof‘coordinativedefinitions’.Reichenbachemphasized,inlinewithgenerallogicalempiricistdoctrine,thatthis2HansReichenbach,ThePhilosophyofSpaceandTime.NewYork:Dover1957.Origi-nalGermanversion:PhilosophiederRaum-Zeit-Lehre.Berlin:WalterdeGruyter1928.\nReichenbachandtheConventionalityofDistantSimultaneity317coordinatizationisfundamentallyconventionalincharacter:likeallcoordinativedefinitions,thedefinitionofsimultaneityisarbitrary3.ThestandardstorythustellsusthatthereisaLeitmotivofempiricistandoper-ationalistconsiderationsbothinthedevelopmentofspecialrelativityitselfandinthephilosophyofspaceandtimelinkingupwithrelativity.Themostemblematicelementofthisstoryistheaccountitgivesoftheconventionalityofrelativisticsimultaneity.ThisThesisoftheConventionalityofDistantSimultaneitynotonlyencapsulatestheempiricistphilosophythatisintimatelyconnectedwithrelativitytheory,italsorelatesdirectlytothedrasticrevisionoftemporalnotionsthatises-sentialforthetheoryitself.Thisstandardaccounthasnotgoneunchallenged.Inparticularaftertheap-pearanceofMichaelFriedman’sReconsideringLogicalEmpiricism,4ithasbe-comeoutdatedtotreattheearlyworkofReichenbachandotherlogicalempiricistsasadirectcontinuationofMach-likeempiricism.ItisnowwelldocumentedthatatleastReichenbach’sownversionoflogicalempiricismoriginatedfromneo-KantianconsiderationsandthatitwasonlyundertheinfluenceofMoritzSchlickthatReichenbachafter1920cametospeakabouthiscoordinativedefinitions(firstproposedbyhimasneo-Kantiansyntheticaprioriprinciples)asarbitraryconven-tions.InthispaperIshallfollowFriedman’slead,withaspecialeyetothespecificcaseoftheconventionalityofdistantsimultaneity.ThequestionIshallattempttoansweriswhetheramoredetailedandmorehistoricallyinformedaccountofthedevelopmentofReichenbach’spositionthantheoneprovidedbythestand-ardstorycanshednewlightontheConventionalityThesis.ItisundeniablethatReichenbach,afterhisdiscussionwithSchlick,explicitlyandrepeatedlyclaimedthatdistantsimultaneityisconventional;butdidhemeanexactlythesamethingasotherconventionalistsandlatercommentators,andwashefullyconsistent?IshallarguethatoncloseranalysistracesofReichenbach’searlierneo-Kantianstancebecomevisible,andthatthesecreateanunresolvedtensioninReichenbach’sposi-tion.TostartwithIshallhavealookatEinstein’sownsupposedlyoperationalistandconventionalistpositionin1905–theplacewherethewholeconventional-itydebatehasitsorigin.Partlydrawingonanotherpaper,5Ishallarguethatone3Thereisabundanttextualsupportforthispartofthestandardaccount.See,e.g.,Re-ichenbach,ThePhilosophyofSpaceandTime,op.cit.,sections4and19;HansRe-ichenbach,“Lasignificationphilosophiquedelathéoriedelarelativité”,in:RevuePhilosophiquedelaFranceetdel’Étranger94,1922,pp.5-61(Englishtranslationin:StevenGimbel/AnkeWalz(Eds.),DefendingEinstein:HansReichenbach’sWritingsonSpace,TimeandMotion.Cambridge:CambridgeUniversityPress2006,chapter10);HansReichenbach,AxiomatizationoftheTheoryofRelativity.Berkeley:Uni-versityofCaliforniaPress1969,section2(Germanoriginal:Axiomatikderrelativis-tischenRaum-Zeit-Lehre.Braunschweig:Vieweg&Sohn1924).4MichaelFriedman,ReconsideringLogicalPositivism.Cambridge:CambridgeUniver-sityPress1999.5DennisDieks,“TheAdolescenceofRelativity”,in:VesselinPetkov(Ed.),Minkowski\n318DennisDieksshouldnotreadanoperationalism-basedconventionalismintoEinstein’sstate-mentsof1905andlater.2.EINSTEINANDTHEDEFINITIONOFSPACEANDTIMETheemphasislaidbyEinsteininhis1905paperontheneedtodefineourno-tionsofspaceandtimebeforewecanevenstarttodophysicscertainlysuggestsoperationalistsympathies.However,weshouldnotethatthe1905paperisnottheonlyplacewhereEinsteinexpresseshimselfinthisfashion:remarkably,hereferstotheneedfordefinitionsofphysicalconceptsevenincontextsinwhichheexplicitlyopposesoperationalistandlogicalempiricistideas.Forexample,inhisAutobiographicalNoteswefindEinsteinreminiscingaboutthediscoveryofspecialrelativitywiththefollowingwords6:Onehadtounderstandclearlywhatthespatialco-ordinatesandthetemporaldurationofeventsmeantinphysics.Thephysicalinterpretationofthespatialco-ordinatespresup-posedafixedbodyofreference,which,moreover,hadtobeinamoreorlessdefinitestateofmotion(inertialsystem).Inagiveninertialsystemtheco-ordinatesmeanttheresultsofcertainmeasurementswithrigid(stationary)rods.…If,then,onetriestointerpretthetimeofaneventanalogously,oneneedsameansforthemeasurementofthedifferenceintime…Aclockatrestrelativetothesystemofinertiadefinesalocaltime.Thelocaltimesofallspacepointstakentogetherarethe“time”whichbelongstotheselectedsystemofinertia,ifameansisgivento“set”theseclocksrelativetoeachother.Thisisanalmostverbatimrepetitionoftherelevantpassagesfromthe1905paper,includingtheuseoftheterm‘define’,andwiththeexplanationthatspaceandtimecoordinatesmeanwhatisindicatedbyrodsandclocks;andallthiswithoutanyaccompanyingcommentthatmightindicatethatEinsteininthenineteen-fortiesdeemedsomekindofqualificationofhis1905statementsnecessary.SowemaysafelyassumethatEinsteinishereexpressingthesameviewastheonehehadinmindinhisoriginalrelativitypaper.Thisisstrikingbecauseelsewhereinthesesameautobiographicalnotes,andalsoinEinstein’s‘RepliestoCriticism’inthesamevolume7,wefindanexplicitandstrongrejectionofBridgman’soperationalismandReichenbach’sempiricismasviablephilosophiesofscience.Forexample,aboutBridgman’soperationalismEinsteinprotests8:Spacetime:aHundredYearsLater.NewYork:Springer2009.6P.A.Schilpp(Ed.),AlbertEinstein:Philosopher-Scientist.LaSalle:OpenCourt1949,p.55.7Schilpp,ibid.,pp.665-688.8Schilpp,ibid.,p.679.\nReichenbachandtheConventionalityofDistantSimultaneity319Inordertobeabletoconsideralogicalsystemasphysicaltheoryitisnotnecessarytodemandthatallofitsassertionscanbeindependentlyinterpretedand“tested”“operation-ally”;defactothishasneverbeenachievedbyanytheoryandcannotatallbeachieved.Inordertobeabletoconsideratheoryasaphysicaltheoryitisonlynecessarythatitimpliesempiricallytestableassertionsingeneral.EinsteinmadethesamepointingreaterdetailinhisReplytoReichenbach.InhiscontributiontotheEinsteinVolume,Reichenbachhadstatedthatthephilosophicallessontobelearntfromrelativitytheorywasthatbasicphysicalconceptsmustbegivenmeaningbymeansof‘coordinativedefinitions’:itisonlythe‘coordination’ofaconcretephysicalobjectorprocesstotheconceptsinquestionthatbestowsphysicalsignificanceonthem.AsReichenbachwrote9:Forinstance,theconcept“equallength”isdefinedbyreferencetoaphysicalobject,asolidrod,whosetransportlaysdownequaldistances.Theconcept“simultaneous”isdefinedbytheuseoflight-rayswhichmoveoverequaldistances.Thedefinitionsofthetheoryofrelativityareallofthistype;theyarecoordinativedefinitions.Reichenbachcontinuedbyexplainingthatthisdefinitionalcharacterofbasicphysicalconceptsimpliesthattheyarearbitrary:Definitionsarearbitrary;anditisaconsequenceofthedefinitionalcharacteroffundamen-talconceptsthatwiththechangeofthedefinitionsvariousdescriptionalsystemsarise.…Thusthedefinitionalcharacterofthefundamentalconceptsleadstoapluralityofequiva-lentdescriptions.…Allthesedescriptionsrepresentdifferentlanguagessayingthesamething;equivalentdescriptions,therefore,expressthesamephysicalcontent.InhisresponseEinsteinobjectedthatanyconcretephysicalobjectissubjecttodeformingforces,andcanthereforenotbeusedtodefineconcepts.Weneedatheoryofthesedeforminginfluencesinordertobeabletocorrectforthem,andsuchatheoryalreadyusesanotionoflength.Therefore,wemusthaveanideaofwhat‘length’ispriortotheuseofanyactualmeasuringrod.FromthisEinsteinconcludesthataconceptlike‘equalityoflength’cannotbedefinedbyreferencetoconcreteobjectsatall;suchconcepts“areonlyindispensablewithintheframe-workofthelogicalstructureofthetheory,andthetheoryvalidatesitselfonlyinitsentirety”.10Theunitoflengthcanonlybesupposedtoberealizedbyanimaginaryidealrod,whichcanatitsbestbeapproximatedbyaconcreteobject–andthisonlyontheconditionthatwearethinkingoftheconceptoflengthincircum-stancesinwhichitmakessensetoassumetheexistenceofrodsatall!Actualrodshavethustobeadjustedonthebasisoftheory,andthismeansareversaloforder9HansReichenbach,“ThePhilosophicalSignificanceoftheTheoryofRelativity”,inSchilpp,ibid.,p.295.10Schilpp,ibid.,p.678.\n320DennisDiekscomparedtotheanalysisthatbeginswithoperationaldefinitionsandstartscon-structingatheoryonlyafterwards.AnotherthoughrelatedpointstressedbyEinsteinisthatmacroscopicdeviceslikerodsandclocksdonothaveafoundationalroletoplayintheinterpretationoffundamentalphysics.Thereferencetothemonlyservespracticalpurposes:itmakescontactwithfamiliareverydaycircumstancesandthusdirectsourthoughts.Makingthisuseofthemisonlyatentativemanoeuvre,“withtheobligation,how-ever,ofeliminatingitatalaterstageofthetheory.”11Sowhenwelookatthehistoricalevidenceinamoredetailedway,itbecomesveryplausiblethatbyhisuseofthetermdefinitionin1905Einsteindidnotwanttoimplythatwearedealingwitharbitrarymeaningstipulationsthatmustprecedetheoryconstruction.12Theverysamepointcanbemadewithrespecttothenotionofsimultaneity.ThereisinfactaremarkablecontinuityinEinstein’sutterancesfromtheearlytwentiesonwards,whenhefirstexplicitlyaddressesphilosophicalquestionsrelatingtospaceandtime.InthesephilosophicalwritingsEinsteincon-sistentlyrejectstheprojectofdefiningconceptsalongthelinesofoperationalismorlogicalempiricism.ThestrikingfactthatEinsteinusestheterm‘definition’torefertothecontentofspatiotemporalnotionseveninthiscontextillustratesthathedidnotrealizetheextenttowhichthistermisabletoexcitephilosophersandcangiverisetomisunderstandings.Initselfthisisquiteunderstandable:Ein-stein’spapersonspecialrelativityareevidentlyphysicspapers,addressedtoaphysicistaudience.Einsteinwasfacingthetaskofconvincinghisreadersthatthespatiotemporalconceptsofclassicalphysicswerenotbeyonddiscussionandthat,indeed,certainchangesintheseconceptswouldmakeitpossible,inasurprisingmanner,toconsistentlycombinethetwopostulatesofrelativitytheory.Heat-temptedtodemonstratethatactualmeasurements(oftheusualkind,traditionallyemployedtodeterminespatiotemporalrelations)didnotprovethesoleapplicabil-ityoftheclassicalnotions.Inparticular,itwasimportantforEinsteintomakeitclearthattherewasnoempiricalsupportfortheabsolutenessofsimultaneity,norforthepre-relativisticideathatthetoandfrovelocitiesoflighthavetodifferinalmostallinertialsystems(namelythosemovingwithrespecttoabsolutespace).TheinterdependencethatEinsteinnotesbetweensimultaneityandthevalueofthespeedoflightisemployedbyhimtoconsistentlyapplythesamesynchronisationprocedure,withthesamevalueofthespeedoflight,inallinertialsystems.Thisisaquitedifferentprojectthanarguingthatthesenotionsarearbitraryinanysingleframeofreference.Einstein’sspecialtheoryofrelativityhasservedasabeaconfortwentieth-centuryphilosophyofscience;butquiteafewcommentatorshavemisinterpretedthephilosophicalimplicationsofthetheory.AsHoward13correctlyconcludes,it11Schilpp,ibid.,p.59.12SeeDieks,“TheAdolescenceofRelativity”,op.cit.,formoreextensiveargumentationconcerningthispoint.13DonHoward,“EinsteinandtheDevelopmentofTwentieth-CenturyPhilosophyofSci-\nReichenbachandtheConventionalityofDistantSimultaneity321wasonlywiththedownfalloflogicalempiricism,andtheQuineancriticismoftheanalytic/syntheticdistinction,thatphilosophyofsciencecaughtupwithEin-stein’sthinkingaboutthestatusofphysicalconcepts.Reichenbach’sdiscussionsofrelativityhavecertainlycontributedtothemisunderstandings.Hisanalysisofthenotionofsimultaneityinparticularhasbeeninstrumentalinreinforcingtheideathatspecialrelativityshouldbeseenasboththefruitandvictoryofastrictlyempiricistphilosophyofscience.However,Reichenbach’sideasaresophisticatedandcomplex,andasweshallseetheyleaveroomforthesuppositionthatdifferent,conflictingconceptionswerecompetingforpriorityinhisthinking;notallofwhichfitinwiththestandardread-ingofhiswork.ThiscomplicatedcharacterofReichenbach’sideascanbebroughtoutbylookingattheirinterestinghistory.3.REICHENBACH,RELATIVITYTHEORYANDTHEAPRIORIHansReichenbachwasoneofthestudentsattendingEinstein’sfirstrelativitycourseattheUniversityofBerlinin1919;ayearlaterhisRelativityTheoryandAprioriKnowledge(RelativitätstheorieundErkenntnisAPriori)14appeared.Asthetitleindicates,theproblematicReichenbachwasdealingwithinthisworkwastherelationbetweenKantianphilosophyandrelativity.WhenonereadsthebookitverysoonbecomesevidentthatReichenbachisnotatallattackingKant’sepistemologyfromanempiricistpointofview,usingEinstein’stheoryasanally–asonemightexpectonthebasisofthelorethatseesadirectempiricistlinkgoingfromMachviaEinsteintoReichenbach.Quiteonthecontrary,Reichen-bachsetshimselfthetaskofsalvagingasmuchaspossibleofKantiandoctrine,giventheproblemsthetheoryofrelativityadmittedlycausesforit.15InhisfirstchapterReichenbachnotesthataccordingtospecialrelativitythetemporalorderbetweentwoeventsisnotuniqueinallcases:foreventswithspacelikeseparationthisorderdependsonthechoiceofaframeofreference.Indeed,thesimultaneityrelationassociatedwithagiveninertialframeofreferencedetermineswhichoneofthetwoeventsisearlier;goingfromoneframetoanothermeansadoptingdif-ferentjudgementsaboutwhicheventsaresimultaneous;inthecaseofspacelikeeventsthischangemayreversethetemporalorderoftheevents.ThisresultisincompletecontrasttoKant’sdoctrineofthereineAnschauung,accordingtowhichitisaprioricertainthatallpossibleeventsareembeddedinoneuniquetemporalseries.ButReichenbachstressesthattheexistenceofthisconflictdoesnotimplyence”,toappearin:TheCambridgeCompaniontoEinstein.Cambridge:CambridgeUniversityPress.14HansReichenbach,RelativitätstheorieundErkenntnisApriori.Berlin:Springer1920.Englishtranslation:HansReichenbach,TheTheoryofRelativityandAPrioriKnowl-edge.Berkeley:UniversityofCaliforniaPress1965.15Cf.Friedman,op.cit.,chapter3.\n322DennisDieksthedownfalloftheKantianapproach.ForaccordingtoReichenbachKantwascertainlyrightinpointingoutthatapriorielementsareabsolutelyindispensableinanyempiricalinvestigation:weneedtoavailourselvesofconceptsbeforewecanevenstartstudyingnature.WhatReichenbachisthusarguingin1920istheinevitabilityofa‘constitutiveapriori’,consistingofanetworkofconceptsandprinciplesthatmakeitfirstofallpossibletogetagriponanyfieldofresearchandthatinthisway‘constitute’thefield.Butthisconceptualframeworkcanandwillchangeinthecourseoftime:itispossibletodevelopandadaptourconceptsinresponsetounexpectedrelationsbetweenempiricaldataandtheemergenceofnewtheoreticalideas.Thecontributionofhumanreasonisthereforenotgivenonceandforall,asoriginallyclaimedbyKant,butconsistsinevolvingprinciplesbymeansofwhichweorderthedataofexperience.Kant’sdoctrineoftheapriorishouldaccordinglybesplitupintoaconstitutiveandanapodicticcomponent:theconstitutiveapriorimustberetainedwhereastheapodicticpart,whichsaysthattheconceptsfurnishedbyintuitionhavepermanentandabsolutevalidity,shouldberejected.Reichenbach’smanoeuvrehereistypicalofneo-Kantianism:RelativitätstheorieundErkenntnisAPrioricanbeconsideredaneo-Kantiandiscussionofrelativitytheory,closeinspirittotheMarburgschool(Cassireretal.).Itisnaturaltoaskhowthisevolutionofourconceptsshouldbethoughtofindetail,andReichenbachpaysexplicitattentiontothisquestion.Onemightimag-ine,giventherejectionoftheapodicticsignificanceoftheapriori,thatcompletelyfreeconceptualchangescanbemadeinthefaceoftensionsinthenetworkofourknowledge.ButthisisnotthewayReichenbachdealswiththeissuein1920.Instead,hearguesthatthereisanimportantprinciplegoverningconceptualde-velopment,namelytheprincipleofwhathecalls‘continuousextension’(stetigeErweiterung):ourconceptualnetworkisadaptedinapiecemealfashion,sothatcentral,well-embeddedelementsremainunchangedandonlythemost‘margin-al’onesundergorevision.Althoughinprincipleallourconceptsmayeventuallychangeundertheinfluenceofnewempiricalfindings,inpracticesomeofthemarevirtuallyimmunetosuchrevision.Think,inparticular,oftheconceptsusedfordescribingdailyexperience:thesewillnotchangeasaresultoftheevolutionofsciencesincetheyaresoutterlycentralinourexistingconceptualframeworkandindispensableformakingcontactwithalreadyexistingknowledge.Thisprincipleofcontinuousextensionmakesitpossibletorelatenewtheoriestoolderonesand,importantly,itjustifiesusinperformingobservationsinwhichweimplicitlyuseoldertheories.So,evengrantedthatthereisnoapriorifoundationofourconceptualframe-workinthesenseofarockbottomwitheternalvalidity,onlythosenotionswillactuallychangethatarenotlinkedupwithdirect‘un-theoretical’observation;thenotionsthatchangearethoseneededtodealwithsituationsdirectlyaffectedbythenewtheories.Thebetteraconceptisembeddedinthenetworkofconceptsused\nReichenbachandtheConventionalityofDistantSimultaneity323todescribesituationsthatarenotdirectlytouchedbytheoreticaldevelopment,thelesschangeableitis.Reichenbachgivesseveralexamples.16Forinstance,inordinarymeasure-mentsoflengthsandtimeswedonotneedtotakeintoaccountrelativisticcontrac-tionsanddilations,noteveniftheeventualgoalistotestpredictionsofrelativitytheoryitself.Andifwelookthroughatelescope,inordertotestthepredictionsofgeneralrelativity,wemayforgetaboutthefactthataccordingtogeneralrelativ-ityinthetelescopeitselflightdoesnotpropagatealongEuclideanstraightlines.Althoughgeneralrelativitytheorynecessitatesdrasticchangesinthearsenalofconceptsneededtodescribetheuniverseatlarge,andalthoughtheobservationsmadewithatelescopeareofundoubtedrelevanceforthiscosmologicaldescrip-tion,thetelescopeitselfandthefindingsarrivedatwithitshelpcanbedescribedwithnotionsfromclassicalpre-relativisticphysics,whichinturncoincidetoalargeextentwitheverydayconcepts.Withoutthiscontinuityofdescriptionandtheapproximatevalidityofoldertheorieswewouldbeatalossinconnectingournewtheorieswithearlierobservations.4.THEFLEXIBILITYOFTHENOTIONOFSIMULTANEITYWhenweapplytheseratherconservativeideasabouttheevolutionofconceptstotheparticularcaseofdistantsimultaneityweareledtoanaccountalongmoreorlessthefollowinglines.Firstofall,whatwemostdirectlymeasurearelocalquan-tities,sinceweourselvesarespatiotemporallylocalcreaturesthatrespondtolocalstimuli.Theobservationofthehandsofaclockhereandnowisanexampleofsuchanimmediatelocalmeasurement.Likeallobservations,alsosuchdirectonesmustemployconstitutiveprinciples;buttheseareofaveryrobustkind,centrallyembeddedinourlanguageandpractice.Forallpracticalpurposes,theconceptsrelatingtodirectlocalobservationsarethereforeimmunetorevisionundertheorychange.Whenitcomestothecomparisonofeventsthattakeplaceatadistancefromeachother,however,observationsbecomelessdirect.Itistruethatclassicaltheo-ryprovidesuswithstandardproceduresformakingsuchcomparisons,andinpar-ticularforestablishingsimultaneityatadistance,butitisnotimmediatelyobviousthattheseprocedurescanbeconsistentlyappliedalsowithinthecontextofthenewtheory–andtotheextentthattheycan,itisnotself-evidentwhatthepropertiesoftheresultingsimultaneityrelationwillbe.Hereweclearlyfindourselvesataplaceinthenetworkofconceptswhichislesscentral,lessdirectlylinkedtoobservation,andthereforeofferingmoreroomtoflexibilityandrevision.Thisisgoodnews,foritturnsoutthattheconceptofsimultaneityhastobecomedifferentfromitsclas-sicalcounterpartifthetwopostulatesofspecialrelativity(therelativitypostulate16Reichenbach,RelativitätstheorieundErkenntnisApriori,op.cit.,pp.66-67.\n324DennisDieksthatsaysthatallinertialframeshaveequalstatusandthelightpostulatethatsaysthatthevelocityoflightishasaconstantvaluethatisindependentofthevelocityoftheemittingsource)aretobeconsistent–thisfollowsdeductivelyfromthetwopostulates,asdemonstratedinEinstein’s1905paper.Inotherwords,intheinitialstageofReichenbach’sphilosophicaldiscussionofrelativitytheemphasisisonthepossibilityofconceptchangeinthetransitiontonewtheories.Thisprocessofconceptualevolutionisinterpretedasacontinu-ousprocessofadaptation,whichstaysascloseaspossibletothealreadyexist-ingframework.Theflexibilitythatisrequiredisjusttheroomneededtomakethenewtheorypossibleatall;inthecaseofsimultaneitythiswouldleadtothestandardrelativisticsimultaneityrelationusedbyEinstein(correspondingtothesynchronizationproceduredescribedinsection1),whichdeviatesfromtheclassi-calsimultaneityrelationinthatdifferentinertialobserverswillcometodifferentconclusionsaboutwhicheventsaresimultaneousbutisidentical,quasynchroni-zationprocedure,toclassicalsimultaneityintheetherframe(inrelativitytheorythereisnosuchpreferredframe,andinaccordancewiththerelativitypostulatethesameprocedureisappliedinallinertialframes).Onemayintroducetheterm‘conventionality’inthiscontext:theconceptsthatareopentorevisionarethosenotcompletelyfixedbythecentral,morerobustconceptsandthefactsformulatedwiththeiraid.Inthissensetheyarenon-factual,conventional.Wecandecidetoadjustthemwithoutcomingintoconflictwithdirect,localfindings.Butthistypeofconventionalityonlyservestocreateroomforthenew(frame-dependent)relativisticconceptofsimultaneityandhaslittletodowiththemuchstrongerconventionalitythesisthatsurfaceslessthantwoyearslaterinReichenbach’swork.5.THEREICHENBACH-SCHLICKEXCHANGEANDTHECONVENTIONALITYOFSIMULTANEITYAftertheappearanceofRelativityTheoryandAprioriKnowledge,stillin1920,ReichenbachsentacopyofhisbooktoMoritzSchlick.Schlickrespondedposi-tivelyinabriefletter,afterwhichamoreextensiveexchangedeveloped.Thiscor-respondence,whichhasreceivedmuchattentioninthephilosophicalliterature17,apparentlyhadagreatimpactonReichenbach:afteritReichenbachrephrasedhispositioncompletely.InhislettersSchlicktakesReichenbachtotaskforpayingtoomuchtributetoKant.AccordingtoSchlick,itisexactlythecombinationoftheapodicticandcon-stitutiveaspectsthatischaracteristicofKant;separatingthesetwoaspectsinthe17See,alsoforfurtherreferences:FlaviaPadovani,ProbabilityandCausalityintheEarlyWorksofHansReichenbach.Geneva:dissertationattheUniversityofGeneva,FacultyofLetters2008.Also:Friedman,op.cit.\nReichenbachandtheConventionalityofDistantSimultaneity325wayReichenbachhasdoneleadstoadistinctivelynon-Kantianstancethatratheraccordswithempiricism.SchlickadmonishesReichenbachtoavoidmisunder-standingsabouthisalliancesandtoeschewKantianterminology:oneshouldusetheterm‘convention’,àlaPoincaré,insteadofspeakingabout‘aprioriconstitu-tiveprinciples’.AsSchlickwritesinhisletterof26November1920:18ThecentralpointofmyletteristhatIcannotfindoutwhatthedifferencereallyisbetweenyouraprioriprinciplesandconventions,sothatweseemtoagreeontheessentialissue.WhathasamazedmemostinyourmanuscriptisthatyoudisposeofPoincaré’sconvention-alitydoctrineinonlysofewwords.19Inhisreply20Reichenbachobjectsthatheterm‘convention’maycreatethemis-understandingthatthereisnofactualcontentinscientificstatements.Hepointsoutthatthereisnostraightforwardarbitrarinessintheaprioriprinciples,becausetheycannotbeconventionallychangedindividually,onebyone.Itisonlythetotalsystemofsuchprinciplesthatcanbesaidtoadmitempiricallyequivalentalterna-tives,andthenstillthesetofalternativesisrestricted(comparethediscussionofgeometrycumphysics,G+F,below).AsReichenbachstresses,suchacombina-tionofprinciplesrepresentsanobjectivePROPERTY(Reichenbach’scapitals)ofreality.21Alittlebitearlierinthesameletterhehadexpressed‘aprofounddistrust’(einstarkesMißtrauen)aboutwhetherthechoicebetweensuchalternative(i.e.,empiricallyequivalent)systemscanbemadeonthebasisofconsiderationsofsimplicity(asSchlick,likePoincaré,wouldmaintain);andhehadalsoreported‘aninstinctiveaversion’(eineinstinktiveAbneigung)totheideathatsuchachoiceisonlyamatterofpragmatics–heproposedtosuspendjudgmentonthiscompli-catedissue.ButSchlickretortson11December22thatPoincaréwasofcourseawareofsuchandsimilarcomplications,andconcludesthatafar-reachingagreementbe-tweenhimselfandReichenbachhasnowbeenreached.23Indeed,Reichenbach18MoritzSchlicktoHansReichenbach,26November1920,HR-015-63-22,reproducedbypermissionoftheUniversityofPittsburgh,allrightsreserved.19„EsistderKernpunktmeinesBriefes,dassichnichtherauszufindenvermag,worinsichIhreSätzeapriorivondenKonventioneneigentlichunterscheidensodasswiralsoimwichtigstenPunkteeinerMeinungwaren.DassSieüberdiePoincarescheKonven-tionslehremitsowenigenWortenhinweggehen,hatmichanIhrerSchriftammeistengewundert.“20HansReichenbachtoMoritzSchlick,29November1920,HR-015-63-20,reproducedbypermissionoftheUniversityofPittsburgh,allrightsreserved.21JedesmöglicheSystembesagtinseinerMöglichkeiteineEIGENSCHAFTderWirk-lichkeit.22MoritzSchlicktoHansReichenbach,11December1920,HR-015-63-19,reproducedbypermissionoftheUniversityofPittsburgh,allrightsreserved.23Actually,Schlickhimselfoscillatedbetweenthe‘local’viewpointthatconceptsshouldbegivenmeaningbyindividualcoordinativedefinitionsandmoreholisticviews.E.g.,inhisletterof26November1920Schlickstatedthatthe‘arbitrariness’onlyentersat\n326DennisDieksfromthispointonstartsdeclaringthatscientificconceptsarefixedby‘coordina-tivedefinitions’,which‘likealldefinitions’arearbitrary.Intheverybeginningthisisstilloccasionallymixedwithreferencestotheconstitutiveroleofsuchdefini-tions,butafter1922thereisnomentionanymoreoftheconstitutiveapriori.24Therearerecurringwarnings,though,alsoinReichenbach’slaterwork,againstpossiblemisunderstandingscausedbytheterm‘conventionalism’.Conceptspos-sessthestatusofaconvention,becauseoftheirdefinitionalcharacter,butthisdoesnotmeanthatstatementsformulatedwiththeirhelplackfactualcontent;thedesignation‘conventionalism’isthereforeunfortunate,Reichenbachrepeatedlydeclares.25InThePhilosophyofSpaceandTimewefindReichenbach’sfullarticulationoftheideathatdistantsimultaneityinrelativitytheoryissuchanarbitraryconven-tion.AsReichenbachexplains,26todeterminethesimultaneityofdistantevents–bysynchronizingtwoclocksAandB–weneedtoknowthevelocityofasignal(forexamplealightsignal)connectingtheclocks,buttomeasuresuchavelocitywerequirepriorknowledgeofthesimultaneityofdistantevents,sothatwegetcaughtinaviciouscircle.Heconcludes:Theoccurrenceofthiscircularityprovesthatsimultaneityisnotamatterofknowledge,butofacoordinativedefinition,sincethelogicalcircleshowsthataknowledgeofsimultaneityisimpossibleinprinciple.Wealsonoticethatthesecondcharacteristicofacoordinativedefinition,namelyitsarbitrariness,issatisfied.ItisarbitrarywhichtimeweascribetothearrivalofthelightrayatB.Alittlebitfurtheronthesamepage(p.127)wefindthepassageinwhichReichen-bachexpressesthisconventionalitybymeansofhisfamousεformula.Discussingthejust-mentionedlogicalcircleinanyattempttosynchronizeclocksAandBbysendinglightfromAtoBandbackagain,hewrites:ItisthisconsiderationthatteachesushowtounderstandthedefinitionofsimultaneitygivenbyEinstein,t=t+½(t–t),whichdefinesthetimeofarrivalofthelightrayatBasthe2131mid-pointbetweenthetimethatthelightwassentfromAandthetimethatitreturnedtoA.Thisdefinitionisessentialforthespecialtheoryofrelativity,butitisnotepistemologicallynecessary.Einstein’sdefinition,too,isjustonepossibledefinition.Ifweweretofollowanarbitraryrulerestrictedonlytotheformt=t+ε(t–t),0<ε<1,itwouldlikewisebe2131adequateandcouldnotbecalledfalse.thelevelofthetotalsystemofprinciples.“EssindjanursolcheKonventionenges-tattet,diesichineingewissesSystemvonPrinzipieneinfügen,unddiesSystemalsGanzeswirddurchdieErfahrungbestimmt;dieWillkürkommterstbeiderArtseinesAufbauhineinundwirdgelenktdurchdasPrinzipderEinfachheit,derÖkonomie,oder,wieichliebergesagthabe,dasPrinzipdesMinimumsderBegriffe.”Seeforanilluminatingdiscussion:Friedman,op.cit.,chapter1.24Padovani,op.cit.,p.177.25E.g.,Reichenbach,ThePhilosophyofSpaceandTime,op.cit.,p.36.26Reichenbach,ThePhilosophyofSpaceandTime,op.cit.,pp.126-127.\nReichenbachandtheConventionalityofDistantSimultaneity327SoReichenbachappearstohavebeencompletelyconvincedbySchlickthattheconceptsoftheoreticalphysicsinthefinalanalysishavetobedefinedbyosten-tionandpossessthestatusoffreestipulations(atleast,thisishowReichenbachinterpretedSchlick;seenote23).Reichenbachwillrehearsethispointofviewexplicitlyandemphaticallyinseverallaterpublications,forexample–aswehavealreadyseen–inhiscontributiontothe1949Einsteinvolume.6.COORDINATIVEDEFINITIONSVERSUSCONSTITUTIVEPRINCIPLESTheanalysisintermsofconventionaldefinitionsasadoptedbyReichenbachafterhisexchangewithSchlickhasthegreatadvantagethatitcaneasilybepresentedandargued:itissimple,systematic,logicallyneatandclear.However,insection2wealreadynotedthatthisaccountdoesnotaccordwiththeactualpracticeofphys-ics:individualtheoreticalconceptsarenotstrictlydefinedbymeansofconcreteempiricalprocedures–therelationbetweentheoryandexperimentisofamoreholisticnature.Ofcourse,thisobservationhaslaterbecomeoneoftheessentialpointsofcriticismoflogicalempiricismingeneral(aswehavenoted,Einsteincanberegardedasaprecursorhere).Remarkably,however,alsoinReichenbach’sownworktherearesignsthattherewereotherconceptionsinthebackofhismindthatstruggledwithhis‘arbitrarycoordinativedefinitions’account.Insection2ofhis1924bookAxiomatizationoftheTheoryofRelativity,enti-tledTheLogicalStatusoftheDefinitions,Reichenbachdistinguishesbetweentwokindsofdefinitions,conceptualandcoordinative.Conceptualdefinitionsclarifythemeaningofaconceptbyestablishingrelationswithotherconcepts.Ifweonlypossessedsuchdefinitions,however,wecouldnotdophysicsbutwouldremaincaughtinanabstractconceptualnetworkwithoutempiricalcontent,asinmathe-matics.Wethereforeneed‘physical’definitionsaswell.Reichenbachformulatesitlikethis:27“Thephysicaldefinitiontakesthemeaningofaconceptforgrantedandcoordinatesittoaphysicalthing:itisacoordinativedefinition”.Therefore,co-ordinativedefinitionspresupposeconceptualdefinitions(asisexplicitlydeclaredbyReichenbachalittlebitlateronthesamepage).Butthiscanonlymeanthatphysicalconceptsmustalreadyhavemeaningandbepartofanetworkofinter-relatedconceptsandprinciplesbeforeany‘coordination’cantakeplace;andthisintroducesanelementoftheoreticalholismintheircontent.WecanseehowthisworksoutindetailwhenwelookatReichenbachdiscus-sionoftheconventionalityofgeometryinThePhilosophyofSpaceandTime.Henotoriouslydoessoviatheconsiderationofdeformingforces(alreadyintro-ducedinhis1922and1924publications),insteadofbyfocussingdirectlyonthearbitrarinessofthedefinitionsoflengthandcongruenceasonewouldexpect.Heexplainsthatactualmeasuringrodscannotbeusedforthedefinitionoftheunitof27Reichenbach,AxiomatizationoftheTheoryofRelativity,op.cit.,p.8.\n328DennisDiekslength:wehavetocountinthepossiblepresenceofforcesthataffecttheirlength.Apparently,wealreadyknowalotaboutthemeaningof‘length’,‘rod’and‘force’beforethecoordination,andthisalertsustothecomplicationoftheinterrelationsbetweentheseconcepts;inphysicalterms,thepossibilitythatdistortingforcesarepresent.Now,thereisasystematicprocedurefordetectingthepresenceofsomeoftheseforces,namelythe‘differential’ones.Theseaffectdifferentmateri-alsdifferently,sothatwecancomparerodsofdifferentchemicalconstitution,findregularitiesintheirvariousresponsestotheforcefield,andfinallymakecorrec-tionsforthedifferentialeffects.This,again,bringsoutthatReichenbachthinksofhiscoordinativedefinitionforlengthintermsofatheoreticallydetermined,idealmeasuringrodthatcanonlybeapproximatedbyanactualobject.Nowthereisafurthercomplication:28universalforces,whichactonobjectsinthesamewayre-gardlessoftheirchemicalcomposition,cannotbedetectedviathejust-mentionedprocedure.SinceonlythecombinationG+Fofgeometryandphysicsleadstotestablestatements,thismeansthatwecanhavedifferentgeometries,combinedwithdifferentphysicaltheories(differingfromeachotherbypositingdifferentuniversalforces),thatleadtoexactlythesameempiricalpredictions.This,then,isacaseofglobaltheoreticalunderdetermination,inwhichthetheoreticalframe-work(inthiscasethedivisionoflabourbetweenGandF)failstobecompletelydeterminedbyempiricaldata.ThedifferentgeometriesGsaydifferentthingsaboutwhichdistancesareequaltoeachother–buttheyagreeontheconceptualmeaningoflengthandcongruenceinthesensethattheyallsaythattheunitoflengthiseverywhererealizedbyanideal,undeformedmeasuringrod.Theimportantpointhereisthatinthisanalysisofthenon-uniquenessofthegeometricaldescriptionthebasicideaisnotthatwearefreetostipulatewhateverdefinitionofcongruencewehappentolike.Rather,itistakenforgrantedthatcon-gruenceisimplementedbythetransportofanidealandundeformedmeasuringrod;thenon-uniquenessofthegeometrynowderivesfromtheunderdetermina-tionbyempiricaldataofthephysicaltheorythatspecifiesthemagnitudeoftheuniversallengthdeformations.Consequently,ifthereisconventionalityhereitisnottheconventionalityassociatedwiththearbitrarycharacterofstipulative,ostensivedefinitions.Insteadwearedealingwithaglobalkindofconventional-ity,associatedwiththeunderdeterminationofthetotaltheoreticalsystem.InThePhilosophyofSpaceandTimeReichenbachproposestooptforthesystemthatsetstheuniversalforcesequaltozero(F=0),onthegroundsthatthischoiceleadstomore‘descriptivesimplicity’(notbecauseitwouldbe‘moretrue’).Headds:29“ThisconceptionoftheproblemofgeometryisessentiallytheresultoftheworkofRiemann,Helmholtz,andPoincaréandisknownasconventionalism.”28Reichenbach,ThePhilosophyofSpaceandTime,op.cit.,sections3-8.29Reichenbach,ThePhilosophyofSpaceandTime,op.cit.,p.35.\nReichenbachandtheConventionalityofDistantSimultaneity329OncloserinspectionReichenbach’sconventionalismthusturnsouttobemoresophisticatedthansuggestedbythemotto“alldefinitionsarearbitrary”.Infact,thattheunityoflengthisgivenbyanundeformedmeasuringrodishandledbyReichenbachassomethingthatis‘unconventionally’givenapriori,beforeanyempiricalinvestigationcanevenbegin.Aglobalnetworkofconceptsspecifyingtherelationbetweenlengthsanddeformingcausesisalreadyinplace,andthisnetworkplaysarolecomparabletothatofthecombinationofconstitutiveaprioriprinciplesinReichenbach’searlierwork.SomethingsimilarcanbesaidaboutReichenbach’sdiscussionofsimultaneity.Iftheconventionalityofsimultaneityreallywerebasedonasupposedcompletearbitrarinesstypicalofalldefinitions,wewouldexpectthatnorestrictionatallwouldhavetobeimposedonthetem-poralcoordinationofeventsatdifferentspatialpositions(withrespecttoagivenframeofreference).However,Reichenbachdoesrestricttheconventionalityofsimultaneitybytherequirementthattheεinhisformulaliebetween0and1.Thebackgroundofthisrequirementcanbefoundinsections21and22ofThePhilosophyofSpaceandTime,whereReichenbachintroduceswhathecallsthe‘topologicalcoordinativedefinitionoftimeorder’:“IfEistheeffectofE,then21EiscalledlaterthanE.”30Remarkably,hereReichenbachrefrainsfromany21mentionofthearbitrarinessofthisdefinition.Onthecontrary,heemphasizeshowclosethis‘criterion’(ashealternativelycallsthedefinition)istoeverydayexperi-enceandpractice:“Theprocedurewhichwehavedescribedisusedconstantlyineverydaylifetoestablishatimeorder.”31Reichenbachnowclarifieshisrestrictiononε:weshouldevidentlynottaketwoeventsassimultaneousifoneofthemislaterthantheother,andthisimpliesthatεmustbebetween0and1.Headds:“Simultaneitymeanstheexclusionofcausalconnection”and“Theconceptsimultaneousistobereducedtotheconceptindeterminateastotimeorder.”32Thereisthennocompletearbitrarinessinthemeaningofsimultaneityafterall!Apparentlythereareprinciplesthatalreadyfixthemeaningoftheconcept,asinthecaseoflength.ThissituationremindsusoftheprincipleofcontinuousextensionandtheconstitutiveaprioriinTheTheoryofRelativityandAprioriKnowledge.Asinthisearlierneo-Kantianwork,Reichen-bachassumesthatthereareglobalandgeneralprinciplesthatregulatetheuseofspatialandtemporalconcepts.Ofcourse,ifsimultaneityindeedpreciselymeans‘impossibilityofcausalconnection’,itfollowsfromrelativitytheorythatthereisnotoneuniqueeventatspatialpositionBthatissimultaneouswithagiveneventatpositionA.Alleventswithspacelikeseparationfromthegiveneventqualify,sothatafiniteintervalratherthanapoint-eventissingledoutassimultaneouswiththeeventatA.ThisconsequenceishailedbyReichenbach,33becauseitcreatesroomfortherelativity30Reichenbach,ThePhilosophyofSpaceandTime,op.cit.,p.136.31Reichenbach,ThePhilosophyofSpaceandTime,op.cit.,p.138.32Reichenbach,ThePhilosophyofSpaceandTime,op.cit.,p.145.33Reichenbach,ThePhilosophyofSpaceandTime,op.cit.,p.145.\n330DennisDieksofsimultaneity(i.e.,theresultofrelativitytheorythatwhetherornottwoeventsaresimultaneousdependsonheframeofreferencewithrespecttowhichthesi-multaneityrelationisconsidered).Ashewrites:Ourepistemologicalanalysisthusleadstothediscoverythattherelativityofsimultaneityiscompatiblewiththeintuitiveconceptionwhichweconnectwithsimultaneity.Itisnotthisconceptionwhichisincorrect,buttheconclusionderivedfromitthatsimultaneitymustbeuniquelydetermined.Thisstatementisstrikingfortworeasons.First,Reichenbachhereconcedesthattheconceptofsimultaneityisgivenmeaningviaantecedentnotions(‘theintuitiveconception’)thatcanbemaintainedtobecorrect.Second,itnowbecomesclearwhyitisimportantforReichenbachtorestrictthemeaningofsimultaneitytoexactly‘lackofcausalconnectibility’andwhyhedoesnotwanttoconsideranyadditionalmeaningingredients.Ashestates,ifitwouldfollowfromthealreadygivenmeaningofsimultaneitythatexactlyoneeventatB–insteadofawholein-terval–wouldbesimultaneouswithanygiveneventatA,itwouldbeimpossibletoaccommodatethespecialrelativisticrelativityofsimultaneity.Thisargument,however,ismistaken.Thereisnoharminaddingmoreintuitivenotionstothemeaningofsimultaneity,andtofixoneeventatBassimultaneouswiththeeventatA,ifthisisdoneinaframe-dependentwaysothatitisgoingtodependontheframeexactlywhicheventistheuniquelychosenone.Thisisinfactthestandardapproachinrelativitytheory,andinthenextsectionweshallseethatthisapproachcanbefittedintoideasthataccordnotonlywithactualphysicsbutalsowiththe‘holistic’strandinReichenbach’sownapproach.Thereisaclearparallelbetweentheconstitutiveaprioriprinciplesof1920andtheglobalelementsinReichenbach’slaterwork.Forexample,in1920ReichenbachcountsEuclideangeometryamongtheconstitutiveprinciplesthatinclassicalphysics(butnolongeringeneralrelativityandinacceleratedframesinspecialrelativity)governthemeaninganduseofthephysicalconceptofcongru-ence.Theprocedureofcontinuousextensionweakensthistoalocalprincipleinthecaseofgeneralrelativityandacceleratedframes;theglobalgeometrywillinthesecasesgenerallybecomenon-Euclidean.After1920–anearlyexampleisReichenbach’s1922paperThePhilosophicalSignificanceoftheTheoryofRela-tivity34–Reichenbachstillholdsthattheglobalgeometryimplicitlydefinesthemeaningofcongruence.Inotherwords,theanalysisisstilltop-down,goingfromthetheoreticalframeworktoindividualconcepts35.Itistruethatthereisnownolongeranapriorispecificationoftheglobalgeometry;Reichenbachnowstressesthatthegeometryisunderdetermined,becauseofthepossiblepresenceofun-detectableuniversalforces.AccordingtoReichenbachthisunderdetermination34StevenGimbel/AnkeWalz(Eds.),op.cit.,chapter10.35Cf.DennisDieks,“Gravitationasauniversalforce”,in:Synthese73,1987,pp.381-397.\nReichenbachandtheConventionalityofDistantSimultaneity331entailsthatthechoicebetweendifferentgeometriesisinprinciplearbitrary(headvisestoadopttheoptionF=0,butemphasizesthatthisisamerelypragmaticchoice).Conventionalityandarbitrarinessarethereforelocatedonthelevelofglo-baltheorychoice,linkedtothetheoreticalunderdeterminationofthetotaltheory.Thatunderdeterminationthusimpliesarbitrarinesswillprobablynotbeacceptedbymostmodernphilosophersofscience.Butwhateverpositiononetakesonthequestionwhethertheorychoiceinsuchcircumstancesisconventional,arbitrary,purelypragmaticornot,onceachoicehasbeenmadethetheoreticalframeworkfunctionsasconstitutiveofthefieldofexperienceinthesamewayasthe1920constitutiveprinciples.Therethereforestillisa‘relativizedapriori’,relativetoanevolvingconceptualframework.Possiblywehavetosaythatthistheoreticalframeworkcontainsimportantpragmaticelements–butwhetherthisentailsthattheconstitutiveprinciplesoftheearlier1920neo-Kantianpositionwereepistemi-callymorerobustandmore‘truth-like’isnoteasytoanswer.Afterall,questionsaboutthetruthofthecategoriesoraprioriprinciplesaredistinctlyun-Kantian.7.THEHOLISTICAPPROACHINPRACTICE:MINKOWSKI’SANALYSISOFSPACEANDTIMEAsitturnsout,thereactuallyexistsanapproachtorelativitytheorythatisexplic-itlyalongtheholisticlineswehaveoutlinedabove.Althoughthisseemsnottohavebeennoticedintheliterature,Minkowski’sfamous1908lectureRaumundZeit36incorporatesseveraloftheideasthatarecentraltoReichenbach’s1920book–andofwhichsome,aswehaveargued,canalsobefoundinhislaterwork.Minkowskistartshisconstructionofwhathecalls‘theobjectiveworld’bydescribingelementary,localphysicalphenomenaintermsofarbitrarilychosencoordinates;hethenperformstransformationsontheresultingequationstobringtheminasimpleandsymmetricalform;andfinallyheintroducesspatialandtem-poralconceptsasembodiedinthatsystemofreferenceinwhichtheequationstaketheirdesiredglobal‘standard’form.ThelawsandequationsMinkowskiactuallyconsidersarethoseofMaxwell;buthemakesitclearthatheassumesthatalllawsofnature,includingthoseyettobediscovered,e.g.thoseresponsibleforthestabil-ityofmatter,willexhibitthesameglobalspacetimesymmetrypropertiesasMax-well’sequations.InMinkowski’sapproachspaceandtimearethusdeterminedonthebasisofconsiderationsabouttheglobalformofthephysicallawsthatrepresenttheregularitiesandpatternsinthephysicalphenomena,startingfromadescriptionoflocalevents.Theconceptsneededtogivethislocaldescriptionareassumedtobegivenbeforehand,butfortheglobalspaceandtimedeterminations36HermannMinkowski,„RaumundZeit“,in:PhysikalischeZeitschrift10,1909,pp.104-111.EnglishtranslationinLorentzetal.,op.cit.\n332DennisDieksthisisnotthecaseandthetaskisaccordinglytofixthemviatheoreticalconsidera-tions.Insomemoredetail,Minkowskifirstintroducescompletelyarbitrarycoordi-natesinordertolabelevents:x,y,zandt.Theeventsinquestionaretobethoughtofasphysicaloccurrencesthathappenlocallytosmallmaterialsystems,specksofmatter.Thepatternsamongtheseeventsshouldnowbecapturedbymathematicalequations,laws,stillexpressedinthearbitrarycoordinatesx,y,zandt.Thetaskthenistotransformtheselawssuchthattheyconformtocertainformrequire-ments,desideratathatarespecifiedbeforehandbutthatclearlyareinspiredbyempiricalknowledgeasformalizedinearliertheories.Inparticular,theequationsshouldbebroughtintoastandardformthatdisplaysspatialhomogeneityandisot-ropy,theequivalenceofinertialframes,etc.Minkowskihimselfdescribeshispro-cedureasfollows:37Fromthetotalityofnaturalphenomenaitispossible,bysuccessivelyenhancedapproxi-mations,toderivemoreandmoreexactlyasystemofreferencex,y,z,t,spaceandtime,bymeansofwhichthesephenomenathenpresentthemselvesinagreementwithdefinitelaws.Thereferencetothetotalityofnaturalphenomenashouldbenotedhere.Wearenotdealingwithlocaldatathatarepastedtogetherbymeansofarbitraryconven-tions,butwearelookingatglobalpatternsintheworldanduse‘apriori’princi-ples(inarelativizedsense!)tosimplifytheirformulation.Conceptslikespatialcongruenceandsimultaneityaresubsequentlyderivedfromthisholisticanalysis:thecongruenceandsimultaneityrelationareimplicitlydefinedbythephysicallawsintheirstandardform.Minkowskifinds,ofcourse,whenhedeterminesthestandardinertialspa-tiotemporalcoordinatesthisway,thatthereisnouniqueness.Thereareinfinitelymanysolutionstotheproblem,correspondingtotheinfinitelymanypossiblein-ertialframesofreference.Inalltheseframesthelawstakethesamestandardform.Thismeans,inparticular,thatspatialcongruenceandsimultaneitybecomeframe-dependent.Sothereisthefamiliarrelativityofsimultaneityinthesenseofframe-dependence;butconventionalityinthesenseofarbitrarinesswithinanyoneframedoesnotemergeinanaturalmanner.8.CONCLUSION:HOLISMANDREICHENBACH’SCONVENTIONALISMInMinkowski’sapproachtherelativisticsimultaneityrelationthusemergesaspartofatotaltheoreticalpackagesatisfyingglobalsymmetryrequirements.AsemphasizedalsobyReichenbach,38theexistenceitselfofsuchanisotropicand37Lorentzetal.,op.cit.,p.79.38Reichenbach,ThePhilosophyofSpaceandTime,op.cit.,p.166.Seealsonote21.\nReichenbachandtheConventionalityofDistantSimultaneity333homogeneousdescriptionofnatureisacontingentmatteroffact.Althoughwithinthelimitsofapplicabilityofspecialrelativitywecangiveallphysicallawsaforminwhichtheyassignidenticalphysicalpropertiestoallpointsinspaceandtime(theirstandardε=½form),suchasymmetricaldescriptiondoesgenerallynotex-istinmorecomplicatedspacetimes.Infact,mostgeneral-relativisticspacetimesdonotallowaconsistentglobalε=½simultaneityrelation.Thestandardsimul-taneityrelationthereforereflects,andisadaptedto,anobjectiveglobalsymmetrypropertythatistypicalofspecialrelativisticspacetime.Itcanbearguednonethelessthateveninspecialrelativityitremainsamatterofourchoicewhetherornotwemakeuseofthisproperty,whetherwechooseatheoreticaldescriptionthatisadaptedtotheglobalsymmetry.Reichenbach’scon-ventionalismcanbedefendedthisway.Indeed,aswehaveseen,Reichenbach’sconventionalismisdifferentfromthe‘arbitrarinessofdefinitions’conventional-ismthatassignsanarbitrarymeaningtoeachindividualterm.Thelatterbrandofconventionalismisalocalaffair;theformerhasaglobal,holisticcharacter.AlthoughReichenbachinmanyplacesproclaimedthatthelocalapproachconsti-tutedthebackboneofhisconventionalism,infactheturnedouttobecommittedtoholisticideasthatinseveralrespectsremindusofhis1920positionandpermitonlyaconventionalismrelatingtothechoicebetweentheories.Reichenbach’sconventionalismboilsdowntoacombinationoftheideathatlocalempiricaldataunderdetermineglobaltheoriesandthethesisthatthechoicebetweenthevariouspossibletheoriesisamatterofourconventionaldecision.Thiskindofconventionalismisevidentlyacontroversialissueinthepresent-dayphilosophyofscience.Manywouldargue,forexample,thatthefactthatdifferenttheoreticalschemesarecompatiblewiththesameempiricaldatadoesnotentailthattheyareequallysupportedbythesedata.Relatedtothis,sayingthatmakinguseofanobjectivepropertyofnatureisamatterofconventionseemssimilartoholdingthatspeakingaboutourdailyenvironmentintermsoftables,chairs,etc.,ispurelyconventional;andthisseemsanunexcitingpositiontotake.Reichen-bach’searlierapproach,outlinedinsections3and4,seemsmoreinteresting;per-hapswiththeadditionthattheprinciplesgoverningtheorychoicearethemselvestheresultsofahistoricalprocessofadaptationtoempiricalfindings,inwhichthebestfittingprincipleshavebeenabletosurvive.Buttheseareissuesthattranscendthescopeofthepresentpaper.HistoryandFoundationsofScienceUtrechtUniversityP.O.Box80.010NL3508TAUtrechtTheNetherlandsD.G.B.J.Dieks@uu.nl\n\nMAURODORATOONVARIOUSSENSESOF“CONVENTIONAL”ANDTHEIRINTERRELATIONINTHEPHILOSOPHYOFPHYSICS:SIMULTANEITYASACASESTUDYMyaiminthisnote1istodisambiguatevarioussensesof‘conventional’thatinthephilosophyofphysicshavebeenfrequentlyconflated.Asacasestudy,Iwillrefertothewell-knownissueoftheconventionalityofsimultaneityinthespecialtheoryofrelativity,sinceitisparticularlyinthiscontextthattheabovementionedconfusionispresent.MyplanistostartbysketchingReichenbach’soriginaltreatmentoftheprob-lem(section1).Insection2,IwilltrytolocateReichenbach’sproblemwithinamuchmoregeneralphilosophicalframework,essentiallyproposedbytheAmeri-canphilosopherWilfridSellarsalmostfiftyyearsago.2Iregardthissecondsec-tionasparticularlyimportant,andnotonlyasageneralintroductiontoourtopic:contemporaryphilosophyofphysicsisaffectedbyadangeroustemptationofex-cessivespecialization,andbyanattitudethatconsiderstechnicalitiesasendsinthemselves.Quaphilosophers,weoughttounderstand,asSellarsputit,“howthings(inthewidestpossiblesenseoftheword)hangtogether(inthewidestpos-siblesenseoftheword)”.3Insection3,Iwillthendistinguishamongfivedifferentsensesof“conventional”,andwillthenstudytheirlogicalrelationshipvisàvistheproblemofestablishinginwhichofthesesensestherelationofsimultaneitycouldberegardedasconventional.NotonlywillIpressthepointthat,asnotedbyDieksinhispaper,muchofthecurrentphilosophicaldebateonconventionalitylackscontactwiththeissuesReichenbach’sanalysiswasmeanttoaddresstostartwith.4FollowingFriedman’s1ThisnoteoriginatedasacommenttoDennisDieks’spresentationThephilosophyofphysicsinperspective,heldinViennainDecember2008forthemeetingoftheEuro-peanScienceFoundation.IwouldliketothankfirstandforemostDennisDieks,andthentheaudience,forthecommentsIreceivedtomycomments.2WilfridSellars,“PhilosophyandtheScientificImageofMan”,inRobertColodny(Ed.),FrontiersofScienceandPhilosophy,Pittsburgh:UniversityofPittsburghPress,1962,pp.35-78.3Ibid.4Seethebeginningsentenceofhiscontributiontothisvolume:“Thehistoryofthephilosophyofphysicshasbeenshapedbyacomplicatedandfascinatinginterplaybe-tweenphysics,philosophicalideasandexternalfactors.Thishistoryisnotonlyanin-triguingsubjectforstudyinitsownright:historicalconsiderationscanalsoshedlightonthecontentofdoctrinesputforwardbyphilosophersofphysicsandarerelevantfortheappraisalofsuchdoctrines.”F.Stadler(ed.),ThePresentSituationinthePhilosophyofScience,ThePhilosophyofScienceinaEuropeanPerspective,DOI10.1007/978-90-481-9115-4_24,©SpringerScience+BusinessMediaB.V.2010\n336MauroDoratooutlooktowardthehistoryoflogicalpositivism,5Iwillalsoarguethat,withouttakingintodueaccounttheoriginalprobleminspiredbyKant’sphilosophy−andconsistingintryingtoseparatewhatinourscientifictheoriesis“duetous”fromwhatiscontributedbytheexternalworld−thedebateontheconventionalityofsimultaneitylosesmuchofitsinterest.Thisconcession,however,shouldnotmakeusblindtothefactthatthespecialtheoryofrelativityhasepistemicallydriven,verificationistfoundations,inspiredbytheworkofMachandespeciallyDavidHume,afactthathasbeenveryimportantforthesubsequenthistoryofbothphys-icsandphilosophy.Astothemorecircumscribedquestionoftheconventionalityofsimultaneityperse,wewillseehow,inanimportantsenseof‘conventional’,therelationalcharacterof“beingsimultaneouswith”impliesbyitselfthatsimul-taneityisconventional,despiteMalament’scelebratedresult,6whileinanothersense,simultaneityappearstobenon-conventional.1REICHENBACH’SORIGINALFORMULATIONOFTHEPROBLEMAfterReichenbach’sgroundbreakingThePhilosophyofSpaceandTime,7theproblemoftheconventionalityofsimultaneityistraditionallyregardedasthequestionofestablishingwhethersimultaneity,besidesbeingnon-controversiallyrelativelytoagiveninertialworldline,isalsoconventional,ornon-unique,evenafteranarbitraryinertialworldlinehasbeenpicked.Asiswell-known,theoriginalproblemthatEinsteinhadtosolvein1905wasthatofsynchronizingtwodistantclocksatrestinthesameinertialframe.Adoptingsometerminologythathasbeeninventedonlylater,hereisEinstein’ssolution.EmitalightsignalatpointeontheinertialworldlineOtowardanother,parallelinertialworldlineO’:parallelismindicates,ofcourse,thatOandO’areinthesameinertialframe.SupposethatthelightsignalisreflectedbyamirroratpointfonO’andisreceivedagainaftersometimeatpointronO.Theproblemofsynchronizingthetwoclocksthattraceoutthetwoinertialworldlinesisequiva-lenttothequestionofdeterminingwhicheventtalongOissimultaneoustothefepistemicallyinaccessiblereflectioneventfonO’.InordertogainsomegeneralityinEinstein’ssimplethought-experiment,Reichenbachintroducedarealnumberε,0<ε<1,suchthatt=ε(t-t)fre5M.Friedman,ReconsideringLogicalPositivism,Cambridge:CambridgeUniversityPress,1999.6DavidMalament,“TheConventionalityofSimultaneity”,inNoûs11,1977,pp.293-300.7HansReichenbach,ThePhilosophyofSpaceandTime,transl.byMariaReichenbachandJohnFreund,NewYork:Dover,1958,pp.123-129.\n„OnVariousSensesof“Conventional337NowReichenbach’squestionabouttheconventionalityofsimultaneitycanbeformulatedwithprecision.CanweassumethatthereflectioneventfonO’hasoccurredexactlyatt=(t-t)/2,freonO,asEinsteinhadoriginallyassumed,andthereforefixthevalueε=½?Or,onthecontrary,isthechoiceofoneamongthenon-denumerableinfinityofvaluesofε≠½comprisedbetween0and1absolutelyunconstrainedbyanyfact,andthere-foreconventional?Intheformerhypothesis,thespeedoflightonthetwolegsofthejourney(fromOtoO’andback)isassumedtobethesame(isotropyofthepropagationoflight).Inthelatterhypothesis,lighthastwodifferentspeedsonthetwolegsofthejourney,insuchaway,however,thattheonlymeasurablequantityonthepartofO,thetotaltwo-waytime,isalwaysthesame.Reichenbach’soriginalclaimisthatthechoiceofε=½isunconstrained,orconventional,duetotheimpossibilityofmeasuringtheone-wayspeedoflight.Inordertomeasuretheone-wayspeedoflight,infact,accordingtoReichenbachwewouldalreadyneedasynchronizedclockonO’,butthewhole“radarprocedure”thatEinsteinproposedwasmeanttosynchronizethetwoclockstobeginwith.Sotheconventionalityofsimultaneity,forReichenbach,isaconsequenceoftheimpossibilitytoattributeisotropytolightwithoutfallingintotheviciouscircleofusingacriterionofsimultaneitytodeterminetheone-wayspeedoflightandusingthisone-wayspeedinordertoestablishacriterionofsimultaneity.Reichenbach’sargumenthasspurredvariouscriticismsthathaveoftenreliedondifferentsensesof“conventional”:inordertolookbackatthewholedebateonemoretime,thesesensesneedtobedisentangledwithcare.Beforedoingso,however,somemoregeneralremarksareappropriate.2THEAMBIGUITYOF‘CONVENTIONAL’,ANDSELLAR’SCONFLICTBETWEENTHEMANIFESTANDTHESCIENTIFICIMAGETheterms“convention”and“conventional”areflagrantlyandintricatelyambiguous.Ontheonehand,theconventionalistheordinary,theusual,thetraditional,theorthodoxasagainstthenovel,thedeviant,theunexpected,theheterodox.Ontheotherhand,thecon-ventionalistheartificial,theinvented,theoptional,asagainstthenatural,thefundamental,themandatory.8Whilethisquotationobviouslyreferstothesocialandepistemicsenseofconven-tion,onethatdoesnotseemdirectlyconnectedtoourtopic,itneverthelesshelpsmetomakeexplicitthegeneralphilosophicalperspectivefromwhichIwanttotackleReichenbach’soriginalproblem.8NelsonGoodman,“JusttheFacts,Ma’am!”,inMichaelKrausz(Ed.),Relativism:In-terpretationandConfrontation.NotreDame:UniversityofNotreDamePress,p.80.\n338MauroDoratoInafirstsenseof‘convention’,Goodmanwrites,‘conventional’seemstorefertowhatistheresultofawidelyshared,butnotnecessarilyintentional,agreement,somethingthatcanberegardedaspartandparcelofcommonsense.Inthissense,weclaimforinstancethat,beforetherevolutionaryviewproposedbyCopernicus,thesharedbeliefinastaticEarthwaspartoforthodoxy,orsimplyofour‘conven-tionalknowledge’.Ontheotherhand,andfollowingtheancientsophists,Goodmancontrasts‘conventional’alsowithwhatis‘natural’;inthissecondsense,itwasalsonatural−i.e.non-conventionalandmandatory−forusalltobelieveintheimmobilityoftheEarth.Suchabeliefseemsinfacttobeentrenchedintheso-calledfolkornaïvephysics,whoseuniversalandinnateholdonourbrainssomanyexperimentsinthecognitivesciencehavenowconfirmed.Fromthisviewpoint,scientificknowledgeartificially,non-naturallyand‘conventionally’goesagainstsomeofour‘natural’,naïvephysicalbeliefs,aswhenitconvincesusthat,forinstance,theEarthmovesalongitsorbitattheaveragespeedofapproximately30km/sec.TheconflictbetweenGoodman’stwosensesof‘conventional’couldthenbereformulatedbyclaimingthatpartofwhatistraditionallyandconventionallybe-lievedbyhumanbeings(inGoodman’sfirstsenseof“convention”)issobelievedbecauseofnatural,non-conventional“forces”thathavebeenshapingourbrainsduringourbiologicalevolution.Consequently,thewholescientificenterprisecouldberegardedasaninventionofourculturebasedonournaturalcapacities,aninventionthat,however,hadtogoagainst(atleast)someofournaturalbeliefs,likethecentralityandspecialityofhumanbeingsinnature,theexistenceofacosmicnow,thefactthatmovedbodiesrequiremovers,ourpervasivetendencytoprojectthenotionofpurposeontothenaturalworld,andsoon.Itthenseemsnaturaltoclaim,alongwithSellars,thatoneofthemainaimsofthephilosophyofphysicsistoinquireintothecompatibilityofthephysicalim-ageoftheworld(thephysicalimageoftime,space,matter,identity,etc.)withthe“manifestimage”,whichistheworldofourexperience,asithasbeenexplicatedbyphilosophersandphenomenologistsofthepast,andinvestigatedempiricallybyneuropsychologistsandcognitivescientiststoday.Inaword,ifIamrightinclaim-ingthatconventionalbeliefs(inGoodman’sfirstsenseof‘convention’)arepartofthemanifestimage,andthattheyarenaturally,non-conventionallybelievedbycommonsense(inGoodman’ssecondsenseof‘convention’)becauseofevolu-tionaryreasons,thedifferencebetweenthesetwosensesof‘convention’isSellars’conflictbetweenthemanifestandthescientificimageoftheworld.TheimportanceofSellars’conflictinthispaperistwo-fold.Firstly,wewillseeoneofitsmanifestationsinthethicketofquestionssurroundingtheconven-tionalityofsimultaneity,inparticularaproposofournatural,naïvephysicalbeliefintheexistenceofacosmicpresent,whichfounditsexpressionintheNewtonian,absolutecharacteroftherelationofsimultaneity.Secondly,focussingonSellars’conflictisalsoparticularlyimportantinjudgingtheimportanceoftheaprioriinthegrowthofscientificknowledge,atleasttotheextentthatmanyaprioripresup-\n„OnVariousSensesof“Conventional339positionsofsciencecomefromthemanifestimageoftheworld.Howchangeableisthisapriorivisàvistheevolutionofscientifictheories?3VARIOUSSENSESOFCONVENTIONALINTHEPHILOSOPHYOFPHYSICSThefollowinglistisnottoberegardedasexhaustive,9butsimplyasafirstguidetofurther,moredetailedwork.1)‘Conventional’iswhateverisopposedtosyntheticorfactual,thatis,some-thingthatcanbesaidofapropositionthatiseitheranalyticallytrue,orsimplydevoidofanytruth-value;thisisasemanticsenseof‘conventional’;2)‘Conventional’canberegardedaswhatisconstitutiveofourtheory(Reichen-bach’sconstitutiveapriori,tobecontrastedwiththeaprioriregardedas‘uni-versallyvalid’,or‘apodictic’);‘conventional’herecorrespondstotheformalelementofknowledgeinthekantiansense,somethingthatintheconstructionofascientifictheoryisexclusively‘duetous’;3)‘Conventional’canberegardedaswhatisnon-reducible/definableintermsofaphysical-causalrelation:thissenseinvolvesthecausaltheoryoftime,andisthetargetofMalament’s1977muchdiscussedtheoremonthenon-conven-tionalityofsimultaneity;104)‘Conventionalasmorechangeable’,orlessentrenchedepistemically(opposedto‘analytic’inQuine’ssense)becausenotpresentindirectexperience.ThissenseisdiscussedbyReichenbachinhis1958famoustreatiseonspaceandtimeandbyDennisDieksinthepapercontainedinthisvolume(section3);itisdefinitelyanepistemicsenseofconventional;5)‘Conventional’canberegardedaswhatisreferredtoachoicethatissupposedtofixagauge.WhileIwillnotcommentinadetailedwayonthislastsense,itisimportanttolistittogetherwiththeothers,asinsomeapproachesthechoiceofasynchronizationfornon-inertialframesisequivalenttogaugefix-ing.119Afirstattemptofseparatingdifferentsensesofconventionalisduetomyformerteach-erRobertRynasiewicz:seehisabstractin“VarietiesofConventionality“,inJacekCachroandKatarzynaKijania-Placek(Eds.),VolumeofAbstractsofthe11thInterna-tionalCongressofLogic,MethodologyandPhilosophyofScience.Cracow:Kopio-rama.1999,p.329.Mylistdoesnotoverlapmuchwithhis,however,andafterhisoralpresentation,Rynasiewiczneverwroteapaper.10Seenote6.11DavidAlba,LucaLusanna,“GeneralizedRadar4-CoordinatesandEqual-TimeCauchySurfacesforArbitraryAcceleratedObservers”,inInternationalJournalforModernPhysics,D16,2007,pp.1149-1186.SeealsoDavidAlbaandLucaLusanna,“ChargedParticlesandtheElectro-MagneticFieldinNon-InertialFramesofMinkowskiSpace-time,”arXiv.\n340MauroDorato3.1Thesemanticsenseofconventional:conventionalasthenon-syntheticInoneclearsenseoftheword,possiblythemostcentralone,‘conventional’isop-posedtowhateverisfactualorsynthetic.Therearetwowaysforasentencetobenon-synthetic:itcanbedevoidofanytruthvalue,oritcanbeanalyticallytrue.Letmestartbyexemplifyingthesenseofconventionthatweareafterinthissectionbybeginningwiththeformeralternative.Afterthespecialtheoryofrelativity,andthediscoveryoftherelativityofsimultaneity,weknowthatthereisnofactofthematterthatcouldbeinvokedtoanswerthequestion:Q=whatishappeningrightnowinAndromeda?simplybecauseweknowthatinthespecialtheoryofrelativitythereisnocosmicpresent:inthistheory,thenowdoesnotextendinspaceatall(itispointlike),orextendsinspaceatmostlocally,inasenseoflocalthatisinanycasedifferentfromthatrulinginquantumphysics.12TheanswertoquestionQisconventionalinthisfirst,semanticsense,sinceitdependsonanarbitrarychoiceofareferenceframeorofaninertialworldlineand,assuch,itcorrespondstonofactwhatsoever.Consequently,thesentence“eventeonAndromedaissimultaneouswithaneventhere-now”,inthepost-Newtonianuniverseslacksadefinitetruthvalue;beinganincompletesentence(itlacksarelationalterm),itisneithertruenorfalseorevenmeaningless.Asaninstanceofasentencethatisnotsyntheticallytruebutstill‘conven-tionallytrue’becauseanalyticallytrue,thinkoftheconventionalistreadingoftheaxiomsofthegeometry.In1902,Poincaréthoughtofsuchaxiomsasdisguiseddefinitions:assuch,theaxiomsofgeometrycouldnotberegardedasareflectionofempiricalfactsnor,giventhemultiplicityofgeometries,assyntheticapriorijudgments.Theaxiomsofgeometrythereforeareneithersyntheticapriorijudgmentsnorexperimen-talfacts.Theyareconventions;ourchoiceamongallpossibleconventionsisguidedbyexperimentalfacts;butitremainsfreeandislimitedonlybythenecessityofavoidingallcontradiction.13AccordingtoPoincaré,axiomsare“true”atbestinthesenseinwhichdefinitionslike“bachelorsareunmarriedmen”are(necessarily)“true”.Inanothersense,however,quadefinitions,theyareneithertruenorfalse,exactlylikesentencesthathavenotruth-makercomingfromtheworldoffacts.Adefinitioncanbeuse-fulandaptforourgoalsornot,butnotreallytrueorfalse.CouldweclaimthatafterwefixaninertialworldlineOpassingthroughour“here-now”,theanswertotheabovequestionQisamatterofmeredefinition?Thisharderquestionwillbetackledintheremainderofthepaper.12Dennis,Dieks,“Becoming,relativityandlocality”,inDennisDieks(Ed.)Theontol-ogyofspacetime,Amsterdam:Elsevier,2006,pp.157-176.13Henri,Poincaré,ScienceandHypothesis,inTheFoundationsofScience.Trans.GeorgeHalsted.NewYork:TheSciencePress,1902/1905,p.65.\n„OnVariousSensesof“Conventional3413.2TheconventionalastheconstitutiveaprioriAccordingtoEinstein’soriginaltreatment,establishingwhether“eventeonAn-dromedaandaneventhere-nowaresimultaneous”necessarilyrequiresaFest-setzung(astipulation),asEinsteinputitin1905.14Thatis,giventheepistemicinaccessibilityofeventeonAndromedafromourhere-nowandconversely(seethefourthsenseofconventional)evenwithinasingleinertialframe,inordertoansweranyquestionaboutdistanteventsrelativelytoachoseninertialframeweneedsomeoperational/conceptualconvention,liketheradarmethodillustratedabove.TotheextentthatEinstein’sradarconventiontransformsmeaninglessques-tionsintoempiricalquestions,theconventionitselfisalso‘constitutive’ofthespecialtheoryofrelativity,inReichenbach’speculiarsenseoftheconstitutiveapriori.15Andthisexplainsthetransitionfromthesenseofconventionaldiscussedin3.1tothecurrentlydiscussedsense.Inanutshell,themainideaofidentifyingtheconventionalelementinasci-entifictheorywithitsconstitutiveelement(s)isthis:notallscientificconceptsinatheoryareepistemicallyonapar,sinceconventionaltruthsaboutsomeofthem(simultaneityinourcase)makeempiricalquestionspossibleinKant’ssense.Sincewithoutaconventionofsomesort(radarmethod),wecouldnotevenaskquestionslikeQabove,theradarmethodandtheresultingconceptofsimultaneityconstituteorgroundthewholetheory.NotethatthisalsocorrespondstoFriedman’sdoctrineoftherelativizedapri-ori:MichaelFriedman,followingtheearlyReichenbach,andvariousotherlogicalpositivists,separatesfromtheoriginalkantianmeaningoftheaprioriregardedassomethingpertainingtoajudgmentthatisuniversallyvalidandunrevisable,theconceptofanpriorithatisconstitutiveofascientifictheory.16AsReichenbachhadit,sucha“constitutiveapriori”maychange;itisthereforerevisableacrossscientificchanges,andisthereforenotuniversallyvalid.Itsflexibilityiscompat-iblewiththefactthatwhatisconstitutiveaprioriforonetheorycanbeabandonedinthelaterconceptualizationofanewtheoreticalframework.WhydoIrefertosuchconstitutiveapriorielementsofourscientificknowl-edgeasconventions?HereIfollowMoritzSchlick’soft-quotedlettertoReichen-bach,onethatwasveryimportanttoconvincethelatterthathehadtoabandonhispreviouskantianlanguage.Schlickwrites:itisthemainpointofthisletterthatIcannotseewhatistherealdifferencebetweenyouraprioristatementsandconventions…Thedecisiveplacewhereyoudescribethecharacter14AlbertEinstein,“OntheElectrodynamicsofMovingBodies”,inAlbertEinsteinetal.,ThePrincipleofRelativity,NewYork:Dover,1952,pp.37-71.15HansReichenbach,TheTheoryofRelativityandAPrioriKnowledge.Berkeley:Uni-versityofCaliforniaPress,1965.16MichaelFriedman,DynamicsofReason:The1999KantLecturesatStanfordUniver-sity,Stanford:CLSIPublication,2001.\n342MauroDoratoofyouraprioricorrespondenceprinciplesseemtomenothingshortofaccomplisheddefini-tionsoftheconceptofconvention.17EventhoughReichenbachinhislaterwritingsimplicitlykeptonbelievingintheimportanceofsomeconstitutiveapriorielementinthefoundationsofscientificknowledge,18itisinanycasehighlysignificantthatafterthiscrucialletterandhisexchangewithSchlick,hewillmakereferencetoKantmoretocriticizehimthantovindicateorreviseaspectsofhisthought.Asaconsequence,afterthislet-terReichenbachwillcoherentlyabandonanyformofkantian-soundinglanguage.Thishadapragmaticmotivation:theneopositivistshadtomakeacarrierintheGermanUniversityaftertheFirstWorldWar,inaculturalenvironmentthatwasdominatedbyneokantiants,byheirsofclassicalGermanidealism,andbyphe-nomenologistslikeHusserlandtheirstudentslikeHeidegger.Beforemovingtothethirdsenseof‘conventional’,Iwouldliketoaddfourremarks.Firstofall,whilenotallconstitutiveapriorielementsinourscientifictheorieshavetoberegardedasconventions,inourcasetheavailabilityofothermethods(operationalcriteria)tofixthemeaningofdistantsimultaneitymakestheradarmethodconventional,ornotdictatedbyfacts.Inthissense,Schlickwasright.Secondly,asReichenbachhimselfmakesclear,alsothechoiceofacoordi-natesysteminthetheoryofrelativityisunderdeterminedbyallpossiblefacts.Accordingly,inhisRelativityTheoryandKnowledgeapriori,heinsiststhattheinvariancewithrespectto(Lorentz)transformationsrepresentstheobjective,fac-tualcontentofreality,whilethestructureofwhatin1920hestillcalls‘reason’(thesourceofthewholeaprioristructureofatheory)isexpressedbythearbitrarinessoftheadmissiblecoordinatesystems…‘thesubjectiveformthatmakesourde-scriptionpossible’.Sothechoiceofacoordinatesystemisconventional(becausesubjectiveandthereforeapriori),andyetindispensibleforthedescriptionofthephysicalworld.Inthisrelatedbutslightlydistinctsenseofconstitutiveapriori,forReichenbachthechoiceofacoordinatesystemisconstitutiveofthetheory(‘itmakesourdescriptionoftheworldpossible’)becausewearespatiotemporallylocatedbeings,sothatwemustdescribetheworldfromsomewhereandsome-when.Theaprioricharacterdependsentirelyonthearbitrarinessofthechoiceofareferenceframe:Thattheconceptofobjecthasanorigininreasoncanberevealedonlybythefactthatinittherearecontainedelementsforwhichnochoiceisprescribedandthatareindependentofthenatureofreality…Thecontributionofreasonisnotexpressedinthefactthatinthe17QuotedfromAlbertoCoffa,TotheViennaStation,Cambridge:CambridgeUniversityPress,1993,pp.201-2.18Foranillustrationofthisclaim,seeMassimoFerrari,Categorieeapriori,Bologna:IlMulino,2003.\n„OnVariousSensesof“Conventional343coordinationsystemthereareinvariantelements,butratherinthefactthatinittherearearbitraryelements.19Thirdly,notethatwhatisconstitutiveofscienceneednotbealsoconstitutiveofourexperienceoftheworldinKant’ssense(anaspectthatneokantianstendtofor-geteventoday).Infact,thereisaprimafacieconflictbetweenthespecialtheoryofrelativityandourexperienceoftime(seeabove):thelatterstronglysuggestsanaturalbeliefinacosmic,absolutepresent,whichispartandparcelofthemani-festimageoftime,abeliefthattheformerexplicitlydeniesbyinsistingontherelativityofsimultaneity.ThisremarkcreatessometensioninKant’sphilosophy,totheextentthatinhisthoughttheconditionsofpossibilityofourexperienceareregardedalsotheconditionsofpossibilityofscientifictheories.Finally,notehowthefirstandthesecondsenseofconventionalaredeeplyrelated:thepresenceofconventionalelementsinscienceinthesecond,constitu-tivesenseentailstheviewthatscientifictheoriesare“afreecreationofthehumanmind”(asEinsteinoftenputit),andthattheyarenotsimply“deduciblefromfacts”,butpartiallydependonus.Inotherwords,scientificmodelsandtheoriesarehumanartefacts:thoseconventionsthatareconstitutivelyaprioriare,inanal-ogytotheformalelementofKant’stheoryofknowledge,duetous,andthereforenotextractablefromtheworldoffacts,whichistherealmofthoseinvarianttrans-formationspreservingthestructureofspacetime.Itisthereforequitecrucialtonotethatwhenwediscusstheproblemoftheconventionalityofthemetric,oroftherelationofsimultaneityinamere“tech-nicalsetting”,weoughtnottoforgetthestrugglesoftheearlyneopositiviststoconfrontthemselveswithKant’sthought,andinparticularwiththeroleoftheconstitutiveaprioriinscience.3.3Conventionalas“non-definableintermsofaphysical(causal)relation”Malament’sfamous1977resultconcernstheuniquedefinabilityofEinsteinε=1/2simultaneityrelationintermsofatime-symmetricrelationofcausalconnectibil-ity,andthereforeintermsoftheinvariantstructureofMinkowskispacetime.Sup-pose,alongwiththedefendersofthecausaltheoryoftime,thataspatiotemporalrelationxisconventionaliffxisnotdefinableorreducibleintermsofaphysical/causalrelation.Then,Malamentprovesthattotheextentthatthecausaltheoryoftimeoughttobeendorsed(in1977hedidnotexplicitlydefendit,butpresentedhisphilosophicalclaiminaconditionalform),hisuniquedefinabilityresult−alreadyimplicitlypresentinaworkbyAlfredRobb20−rulesouttheclaimthattherelationofsimultaneityisconventional.19HansReichenbach,TheTheoryofRelativityandaprioriKnowledge,ibid,mytransla-tion,p.138.20Alfred,Robb,ATheoryofTimeandSpace.Cambridge:CambridgeUniversityPress,1914.\n344MauroDoratoNotethatthisthirdsenseofconventionalseemtobetotallyalientoReichen-bach’soriginal(kantian)philosophicalworriesaboutthe“constitutiveapriori”.Inthishypothesis,suchathirdsense(Malament’sdefinability)wouldbetotallyirrelevantforthesecondsense(constitutivity),sothatweshouldnotconfusethemundertheheading“conventionalityofsimultaneity”.Ontheotherhand,wecouldtrytodefendtheviewthatcausationisaconsti-tutiveapriorielementofthespecialtheory,inthesensethattheobjectivityandinvarianceofthepartialtemporalorderavailableinthetheorydependsontheob-jectivityandinvarianceofthecausalorder.Ofcourse,totheextentthatcausationcanberegardedasarelationthatisimposedbyusontothephysicalphenomenainthesenseofakantiancategory,wecouldaswellconsideritasanelementthatis“duetous”.Butthisclaimwouldneedadditionalargumentsthatcannotbepre-sentedinthelimitedspaceofthispaper.AllIcannoteinsupportofthisclaimhereisthefollowing:sinceinRobb’saxiomatizationofthespecialtheoryofrelativity,therelationof‘beingafter’en-joysafoundationalrole,thereducibilityofthisrelationtocausationwouldgivealsothecausaltheoryoftimea“constitutive”role,sincethecausalrelationwouldbecomethetrueand“primitive”buildingblockofthetheory.Thishypothesis,however,ishighlycontroversial,sincewestilldonotagreeonwhatcausationis.Forinstance,ifweacceptedaHumean,reductionistac-countofcausationasmereregularity,regularitiescertainlydon’tdependonus,sothatcausation,whilepossiblyconstitutiveofthetheoryintheaxiomaticsense,couldnotberegardedasbeingapriori.Ananalogousconclusionwouldholdifweadoptedatheoryofcausationinvolvingarealistattitudetowardpotenciesorcausalpowers.21Therelevancebetweenthethirdsenseandthefirstisnotveryeasytoestablisheither.Ontheonehand,Malament’sresultseemscertainlyrelatedtothefirstsenseofconventional,atleasttotheextentthatwhatisnon-conventionalquadefinableintermsofacausal/physicalrelationalsoappearstobenon-conventionalquafactual.Afterall,aphysical/causalrelationbetweeneventsrepresentsaphysicalfact,albeitrelational.Ontheotherhand,ifoneinsistedinholdingthattheonly“facts”inSRaretheinvariant(worldline-independent)facts,thentherelativityofsimultaneitywouldautomaticallyimplyitsconventionality.Therefore,duetotheambiguityaboutthenotionoffactinthespecialtheoryofrelativity,andtherelateddifficultyofestablishingwhetherrelationalfactsalsocountasfactinthattheory,thethesisthatMalament’sresultimpliesthatsimul-taneityisnon-conventionalinthefirstsenseremainscontroversial,evenafteraninertialworldlineisgiven.Hereisanotherwayoflookingatthisquestion,andappreciateitscomplex-ity.ConsideraneventaonaninertialworldlineO:isitstillmeaninglesstoask21Foradefenceoftheimportanceofcausalpowersinmakinglawstrue,IrefertoM.DoratoTheSoftwareoftheUniverse,Ashgate,2005,andA.Bird,Nature’sMetaphys-ics,OxfordUniversityPress,Oxford,2007.\n„OnVariousSensesof“Conventional345whetheraneventeinAndromedaisoccurringsimultaneouslywithaonO,rela-tivetoO?Asiswell-known,thisquestionhasbeensubjecttovariousdiscussions.Ontheonehand,ReichenbachandGrünbauminsistedontheclaimthat,sincewecan-notmeasuretheone-wayvelocityoflight,wecannotdeemtheanswertotheques-tionasbeingbasedonmattersoffact.Westillneedtheconventionalassumptionthatlighthasthesamespeedinalldirections.So,evenafterweassignaworldlineO,relativelytoOsimultaneityisconventional.Ontheotherhand,manyphilosophersafter1977havetakenMalament’sresultashavingsolvedthisproblemonceandforall.Therelationof“beingorthogonaltoO”,saidofastraightintersectingpointa∈Oidentifiesuniquelyasimultane-ityrelation.Thisrelation,beingdefinableintermsoftheautomorphismsofthestructureofMinkowskispacetime,preservesthatstructure.Ifnoothersimultane-ityrelationscanpreservethestructure,givingupEinstein’sε=1/2choicewouldamounttogivingupthewholestructureofMinkowskispacetime.22Furthermore,arelationlike“issimultaneouswitha∈OrelativetoO”wouldseemtobeobjec-tiveandfactual,quainvariantforallpossibleobservers.DebsandRedheadhaverecentlyclaimedthat,inasense,bothoftheseposi-tionssketchedaboveturnouttobecorrect.23Pickaworldline,andthenaskwhethertwopointsselectedbyEinstein’sstandardsynchronyareobjectivelysimultaneousrelativetothatworldline:theanswermustbeinthepositive,forthereasonsjustgiven.However,thecorrectnessofthisanswercruciallydependsonapreviouschoice,thatofeliminatingLorentzboostsfromthefullgroupofautomorphismsofMinkowskispacetime.24DebsandRedheadconcludethatconventionalismisstillwithusforatleasttworeasons:1)thechoiceofadoptingarestrictedsetofsymmetries(ratherthanthefullset)asaninvariancecriterionfortheobjectivityofarelationis,insomesense,initselfconventional(notdictatedbyfacts);2)oncewedecidetoincludeLorentzboostsinthefullgroup,thenconven-tionalismseemsagaintobecorrect,becauseaboostwilltiltthehyperplaneofsimultaneityorthogonaltotheoriginalworldlineandwillnotpreserveit.Unfortunately,thisirenicclaimseemstoforgetthatthechoicebetweenleav-ingLorentzboostoutoftheautomorphismsgroupornotisnotso“free”.Ifthequestionistodecidewhethersimultaneityifconventionalevenafterhavingfixedaninertialworldline,asReichenbachhadoriginallyposedthequestion,thenDebsandRedheadshouldconcludethatMalamentisright,becauseLorentzboostmustbeleftoutofthefullgroupofautomorphisms.Ofcourse,“thechoiceofwhethertousethelineofsimultaneitydefinedbyO,O’,oranyoneofanyinfinitenumber22ThisargumentisdefendedinM.Friedman,FoundationsofSpacetimeTheories,Prin-cetonUniversityPress,1983.23TalalA.Debs,MichaelL.G.Redhead,Objectivity,Invariance,andConvention,Har-vard:HarvardUniversityPress,2007,p.95.24Ibid,p.97.\n346MauroDoratoofinertialworldlines”25isfullyconventional.However,thisisaconsequenceoftherelativityofsimultaneity,andunlessweareconvincedthatthereisnodiffer-encebetweentherelativityandtheconventionalityofsimultaneity,weshouldstickwithadistinctionbetweenthetwonotions.Inaword,ifthedebateontheconventionalityofsimultaneityisabouttheuniquenessofanε-value,onceagivenworldlinehasbeenconventionallyfixed,thefullPoincarégroupisnotaliveop-tion.3.4TheConventionalastheepistemically“morechangeable”ThisepistemicsenseofconventionalhasbeenclearlydefinedbyDieksinhispaper:26heseemstoimplythatdistantsimultaneityisanon-localconceptandthat,assuch,itislessfirmlyanchoredindirectexperience;thereforemoreopentochange,ormore“conventional”.Conceptsusedindirect,localobservationsareinpracticeunrevisable“althoughinprincipleallourconceptsmayeventuallychangeundertheinfluenceofnewempiricalfindings,inpracticesomeofthemarevirtu-allyimmunetosuchrevision”.27Hereonecouldraiseapointthatinvolvesthemeaningofdirect,localobser-vation.Distantsimultaneityisnotdirectlyobservableofcourse,if‘direct’means‘local’,butinordertodecidewhatcountsaslocallyanddirectlyobservablewealwaysneedatheory,namelythespecialtheoryofrelativityandclassicalelectro-magnetism.WhatDieksseemstoneglectisthatitisalwaysascientifictheorythatdecidesforuswhatisdirectly,locallyobservableandwhatisn’t.Consequently,ifforscientificreasonswecouldadmitaninstantaneoustransmissionoflight,thenour“directexperienceoftime”(inaslightlyenlargedsenseof“experience”)wouldincludeacosmicnow.Andnotethataspartofourmanifestimageoftime,wefirmlybelievethatthereexistsacosmicpresent,andthatsimultaneityisab-solute,andthisseemspartofourdirectexperienceoftheworld.However,whatseemsglobalisinsteadonlylocal,sincebylookingatastarinthenightsky,wewronglybelievethatwedirectlyobservethelightemittedbyitatpresent,butweobserveonlylightemittedlightyearsago.Thisremarkislinkedtothefactthat,againstDieks’opinion,Ithinkthatsciencemaydisposeevenofconceptsthatappeartheresultofourdirectandmostentrenchedexperienceoftheworld,forthesimplereasonthat“localobservation”isatheory-ladenconcept.Interest-ingly,conventionalinthisfourthsenseisfullysynonymouswithQuine’ssenseofsynthetic,giventhatthelattermeansmorerevisable,becauseimpingingontheperipheryofthewholenetworksofbeliefsinwhichourscientificknowledgeconsists.Inanycase,itisbecauseofthetheoryderived,non-directlyaccessiblechar-acterofthesimultaneityrelationbetweentwodistanteventsthatinordertojudge25Ibid.,p.87.26Seethepaperinthisvolume,section3.27SeeDieks’spaperinthisvolume.\n„OnVariousSensesof“Conventional347suchtwoeventsassimultaneousweneedsomeadditionaloperationalprocedure.Theprocedureinquestionmusttranslatetheunobservabilityofthedistantsi-multaneityoftwoeventsintoalocallydiscerniblecoincidenceofpoint-events(twodistantlightsignalsintersectingtwomirrorsposedinfrontofus).ThelackofdirectaccessibilityisthemajorreasonforassumingthatquestionslikeQlackfactualcontent.Therefore,conventionalinthisfourthsense,assumingpositivistictheoriesofmeaning,mightbetakentoexplainconventionalinthefirstsense:thereisnomatteroffactsmakingacertainassertionaboutthesimultaneityoftwolightlikeseparatedeventsastrueorfalse,simplybecausethetwoeventsinquestiondonot,andcannotinprinciple,fallwithinthelimitsofasingleperception.Ontheotherhand,itisthelackofdirectepistemicaccessbetweendistanteventsthatEinstein(andPoincarébeforehim)hasexploitedtointroducesomesortofaconstitutiveaprioriconvention(thesecondsenseoftheword)inthetheory.Aswehaveseen,thisconventioniscapableoftransformingameaninglessquestionlikeQintoanempiricalquestion.Analogously,thefactthatthewholeregionoutsidethelightconecenteredinapointpofMinkowskispacetimeiscausallynon-connectiblewithrespecttop(theelsewhereregionrelativetop)gavephilosophersadditionalmotivationstodefendthecausaltheoryoftime,alreadydefendedbyKant.Theepistemicnon-accessibilityofdistantsimultaneity(itsconventionalityinthefourthsense)findsanexplanationinthelackofaphysicalconnectionbetweendifferentregionsofspacetime(thirdsenseofconventionalgivenbythecausaltheoryoftime).Iftherecannotexistinprinciplephysicalsignalsconnectingwithppointsintheelsewhereregionrelativetopointp,thenanysortoftemporalrelationshipthatwemayfancytointroducebetweenspacelikerelatedevents(simultaneityincluded)isgoingtobeconventionalinboththethirdandfourthsenseoftheword.Theonlyinvarianttemporalorderisgivenbythecausalconnectibilityrelation,whichisleftinvariantbythefullgroupofautomorphismsoftheMinkowskispacetime.Sotheconnec-tionbetweensensethreeandfourofconventionaliscertainlynotunimportant.Insum,theimportanceofthisfourthsenseofconventioncanhardlybeexag-gerated,afactconfirmingthatspecialrelativityisanepistemicallybasedtheory.Alsothepoint-coincidenceargument,thatEinsteindefendedlaterinthecontextofthegeneraltheoryofrelativityinordertoavoidthedireconsequencesderivedfromtheholeargument,isbasedontheclaimthatthedirectlyobservablerelationsarethefoundationalelementsofanyspacetimetheory.ThismeansthatwhileIagreewithDieksthatthecentralityofMach’s(andHume’s)influenceuponEin-steinandtheneopositivistsneedstobere-examinedwithcare,Ithinkhewillagreewithmethatitiscertainlydifficulttodenythatthespecialtheoryofrelativityhasempiricist,epistemically-drivenfoundations.\n348MauroDorato3.5Theconventionalasderivingfromgauge-fixing,i.e.determiningsimultaneityfornoninertialobserversInarecentessay,LusannabeginsbyremarkingthatrealobserversareneverinertialandforthemEinstein’sconventionforthesynchronizationofclocksisnotabletoidentifygloballydefinedsimultaneity3-surfaces,whichcouldalsobeusedasCauchysurfacesforMaxwellequations.28AsLusannaclarifies,whatisrequiredinthiscaseisa3+1splittingofMinkowskispace-time,namelyafoliation…whoseleaves…[are]bothaCauchysurfaceforthedescriptionofphysicalsystemsandaninstantaneous(ingeneralRiemannian)3-spaceofsimultaneityimpliedbyaclocksynchronizationconventiondiffer-entfromEinstein’sone.29AfteratechnicaldiscussioninvolvingtheHamiltonianconstraintapproachtotheproblem,Lusannaclarifiesthatalltheadmissible3+1splittings,namelyallthead-missibleproceduresforclocksynchronization,andalltheadmissiblenon-inertialframescenteredontime-likeobservers,aregaugeequivalent.Thequestionthatisquiteinterestingandnovelforourproblemisthefollow-ing:inLusanna’sapproachtoestablishingsimultaneityfornoninertialframes,thegaugefixingislinkedtotheconventionalchoiceofanextendedphysicallabora-tory.Thespatio-temporalphenomenaastheyareviewedfromnon-inertialframesarethereforecoordinate-dependent,inthesamesenseinwhichtheyarecoordinatedependentwhenwechooseinertialframes.InthismoregeneralapproachinvokedbyLusanna,however,theinertialframescenteredoninertialobserversbecomeaspecialcaseofgaugefixing.Inparticular:Foreachconfigurationofanisolatedsystemthereisaspecial3+1splittingassociatedtoit:thefoliationwithspace-likehyper-planesorthogonaltotheconservedtime-like4-momen-tumoftheisolatedsystem.CONCLUSIONInsum,Ihavetriedtoshowthatwecannottackletheproblemoftheconventional-ityofsimultaneityasifitweresolvablemerelywithtechnicalmeans:theabove28Lusanna,Luca,“GeneralcovarianceanditsimplicationsforEinstein’sspace-times”,talkattheMeetingLaRelativita’dal1905al2005:passato,presenteefuturoorgan-izedbySIGRAVandSISM,DepartmentofMathematicsofUniversityofTorino,June1,2005,p.3.29Ibid,p.4.\n„OnVariousSensesof“Conventional349illustratedconflationofvarioussensesof‘conventional’canbeavoidedonlybygivingbothhistoricalandconceptualconsiderationstheirdue.Ontheonehand,asalreadypointedoutbyFriedman,weoughtnottoforgetthedeepinvolvementwithKant’sphilosophybothofReichenbachandoftheothermembersoftheViennacircle.ThisinvolvemententailsthatthequestionoftheconventionalityofsimultaneitywasforthemsimplyacasestudyusedtotestthevalidityoftheframeworkproposedbytheCritiqueofPureReasonafterthenewrevolutionaryresultsintroducedbythetwotheoriesofrelativity.Ontheotherhand,thehistoricalimportanceofempiricistmethodsforthefoundationsanddiscoveryofthespecialtheoryofrelativity−andtherelativityofsimultaneityinparticular−canhardlybedenied.Notonlyisthisillustratedbythefourthsenseofconventionalpresentedabove,butalsobyimportantevidenceprovidedbyEinsteinhimselfontheroleofMachandHume’sthoughtontheoriginthe1905theory.30InalettersentbyEinsteintoSchlickattheendof1915,weread:Yourexpositionisalsoquiterightthatpositivismsuggestedrelativitytheory,withoutre-quiringit.AlsoyouhavecorrectlyseenthatthislineofthoughtwasofgreatinfluenceonmyeffortsandindeedE.MachandstillmuchmoreHume,whosetreatiseonunderstand-ingIstudiedwitheagernessandadmirationshortlybeforefindingrelativitytheory.Verypossibly,Iwouldn’thavecometothesolutionwithoutthosephilosophicalstudies.31[myemphasis]However,Diekswillhardlydisagreewithmeonthispoint.ItisworthwhiletorecallthatwhileEinsteinlaterdisownedit–byclaimingthatagoodjokeshouldnotberepeated−anappealtotheverificationist/operationalistfoundationsofEinstein’scritiqueoftheconceptofsimultaneitywillbeofimmensehistoricalimportance.Thisappliesnotonlytothehistoryofscientificphilosophy(thinkofitsinfluenceonWittgenstein’sthought,ortheViennacircleetc.),but−consider-ingtheenormousinfluencethatithadinHeisenberg’sandBohr’sthought,andinthemanyphysiciststhatstillfollowthem−alsototheinterpretationofquantummechanics.DepartmentofPhilosophyUniversityofRome3VialeOstiense23400144,RomeItalydorato@uniroma3.it30SeeJ.Norton,„HowHumeandMachHelpedEinsteinFindSpecialRelativity,“pre-paredforM.DicksonandM.Domski,eds.,SynthesisandtheGrowthofKnowledge:EssaysattheIntersectionofHistory,Philosophy,Science,andMathematics.OpenCourt,forthcoming.31A.Einstein,LettertoSchlick,December14,1915,Papers,A,Vol.8A,Doc.165,A.\n\nROMANFRIGGANDCARLHOEFER1DETERMINISMANDCHANCEFROMAHUMEANPERSPECTIVE1.INTRODUCTIONOnthefaceofit‘deterministicchance’isanoxymoron:eitheraprocessischancyordeterministic,butnotboth.Nevertheless,theworldisrifewithprocessesthatseemtobeexactlythat:chancyanddeterministicatonce.Simplegamblingde-viceslikecoinsanddicearecasesinpoint.2Ontheonehandtheyaregovernedbydeterministiclaws–thelawsofclassicalmechanics–andhencegiventheinitialconditionof,say,acoinitisdeterminedwhetheritwilllandheadsortailswhentossed.3Ontheotherhand,wecommonlyassignprobabilitiestothedifferentout-comesofacointoss,anddoingsohasprovensuccessfulinguidingouractions.Thesamedilemmaalsoemergesinlessmundanecontexts.Classicalstatisticalmechanicsassignsprobabilitiestotheoccurrenceofcertainevents–forinstancetothespreadingofagasthatisoriginallyconfinedtothelefthalfofacontainer–butatthesametimeassumesthattherelevantsystemsaredeterministic.Howcanthisapparentconflictberesolved?Oneresponsetothisproblemwouldbetoadoptanepistemicinterpretationofprobabilityandregardtheprobabilitiesweattachtoeventssuchasgettingheadswhenflippingthecoinorthespreadingofthegaswhenopeningtheshutterasareflectionofourignoranceabouttheparticularsofthesituationratherthanthephysicalpropertiesofthesystemitself.Outcomesreallyaredetermined,butwedon’tknowwhichoutcometherewillbeandsoweuseprobabilitiestoquantifyouruncertainlyaboutwhatwillhappen.Thereisnocontradictionbetweendeter-minismandprobabilitiesthusunderstood.However,thisisunsatisfactory.Therearefixedprobabilitiesforcertaineventstooccur,whicharesubjectedtoexperimentaltestandwhich,inmanycases,are1Theauthorsarelistedalphabetically;thepaperisfullycollaborative.Tocontacttheauthorswritetor.p.frigg@lse.ac.ukandcarl.hoefer@uab.es.WewouldliketothankAlanHájekandAidanLyonforhelpfulcommentsonearlierdrafts,andtheaudiencesinVienna,Canberra,ParisandMarylandforstimulatingdiscussions.FriggwouldliketoacknowledgesupportfromtheSpanishresearchgrantFFI2008-01580.Hoefer‘sresearchforthispaperwassupportedbySpanishresearchgrantsHUM2005-07187-C03-02andFFI2008-06418-C03-03/FISO.2Orifoneinsiststhatatbottomtheworldisquantummechanical,thentheproblemisthatprobabilitiesliketheonesweattachtocoinflipsdon’treducetothemicroprob-abilitiessincequantummechanicsassignsvaluesclosetoeither1or0ratherthan1/2toeventslikegettingheadswhentossingacoin.3ForadiscussionofdeterminismseeEarman(1986,Ch.2).F.Stadler(ed.),ThePresentSituationinthePhilosophyofScience,ThePhilosophyofScienceinaEuropeanPerspective,DOI10.1007/978-90-481-9115-4_25,©SpringerScience+BusinessMediaB.V.2010\n352RomanFriggandCarlHoefergovernedbyprobabilisticlaws(suchasthelawsofstatisticalmechanics).Sotheseprobabilitiesseemtohavenothingtodowiththeknowledge,oreventheexist-ence,ofconsciouscreaturesstudyingthesesystems:thechanceofacointolandheadsis0.5andagasisoverwhelminglylikelytospreadwhentheboxisopened,nomatterwhatanybodybelievesabouttheseevents.Thevaluesoftheseprob-abilitiesaredeterminedbyhowthingsare,notbywhatwebelieveaboutthem.4Inotherwords,theseprobabilitiesarechances,notcredences.Thisisanunwelcomeconclusionbecausechanceanddeterminismseemtobeincompatible.InthispaperwearguethatatleastforaHumeanthisincompatibil-ityisonlyapparentandthattheproblemcanberesolvedsinceHumeanobjectivechancesarecompatiblewiththerebeingunderlyingdeterministiclaws–Lewis’ownproclamationtothecontrarynotwithstanding.5Inourdiscussionwefocusonasimpleexample,acointoss,thendevelopaHumeanaccountofchance,andthenshowthatonthisaccountthereisanon-trivialsenseinwhichcoinflipsarechanceeventswhileatthesametimebeinggovernedbydeterministiclaws.Inthelastsectionwebrieflyindicatethatchancesareintroducedintostatisticalmechanicsessentiallyinthesamewayasinthecaseofthecoinandsothebasicideaofdeterministicchancedevelopedherecanbecarriedovertostatisticalmechanicswithout(much)furtherado.2.FLIPPINGACOINCointossingisthemostwidelyusedexampleofarandomprocess,andwearefirmlyconvincedthatthechanceforgettingeitherheadsortailsis0.5.Atthesametimewearealsofirmlyconvincedthatcoinsobeythelawsofmechanicsandthatthereforetheirflightaswellastheirlandingheadsortailsaredeterminedbytheirinitialconditionsandtheforcesactinguponthem.Canweconsistentlyupholdbothconvictions?ThisquestionhasbeendiscussedfromaphysicspointofviewbyKeller(1986),andlater,buildingonKeller’swork,byDiaconis(1998)andDiaconis,HolmesandMontgomery(2007).Webelievethatthisapproachprovidesalltheingredientsneededtoexplainwhythechanceforheadsequals0.5,andwhythereisnoconflictbetweenthisandthefactthatcoinsaregovernedbythelawsofclassicalmechanics.However,theexplanationweofferdiffersfromKeller’sandDiaconis’.WenowreviewinsomedetailtheirargumentssincetheyserveasthespringboardforourowndiscussionofchanceincoinflipsinSection4.4Thispointisoftenmadeinthecontextofstatisticalmechanics;seeforinstanceAlbert(2000,p.64),Loewer(2001,p.611)andGoldstein(2001,p.48);seealsoHoefer(2007,p.557,pp.563-4)andMaudlin(2007,pp.281-2).5Loewer(2001;2004)haspresentedareconciliationofdeterminismandchancefromaHumeanperspective.However,webelievethisreconciliationtobeproblematicforthereasonsdiscussedinFrigg(2008b).\nDeterminismandChancefromaHumeanPerspective353Kellerintroducesthefollowingmechanicalmodelofthecoinflip.Consideracircularcoinofradiusr,negligiblethickness,andwithhomogeneousmassdis-tribution.Theonlyforceactingonthecoinafterbeingtossedislineargravity,andthesurfaceonwhichitlandsismushysothatthecoindoesnotbounce.Fur-thermorethecoinisflippedupwardsinverticaldirectionwithvelocityvatinitialheighth(abovethesurfaceonwhichiteventuallylands)sothatitrotateswithangularvelocityωaroundahorizontalaxisalongthediameterofthecoin(i.e.weruleoutprecession).SolvingNewton’sequationsforthissituation(andassumingthatthecoinisflippedinhorizontalposition)yieldsgt2x(t)=vt–+h(1)2φ(t)=ωt(2)wherex(t)isthecoin’sheightattimetandφ(t)thecoin’sanglerelativetotheplane.Usingthecoin’sradiusonecanthendeterminewhichpointofthecointouchesthesurfacefirst,andtogetherwiththeassumptionthatthecoindoesnotbouncethisdetermineswhetherthecoinlandsheadortails.Thesecalculationsthenallowustodeterminewhichinitialconditionsresultinthecoinlandingheadandtailsrespectively;thatis,theyallowustodetermineforeveryquadruple(x0,v,φ,ω)whetherthecoinhavingthisinitialconditionlandsheadortails.Wehave0assumedthatallcointossesstartatheighthandthatallcoinsleavethehandinhorizontalposition:x=handφ=0;hencedifferenttossesvaryintheirvertical00velocityvandtheirangularvelocityω.Assumingthatthecoinstartsheadsup,initialconditionslyingintheblackareasofthegraphshowninFigure1comeupheads,whilethoselyinginwhiteareascomeuptails.ForthisreasonKellercallsthehyperbolicblackandwhitestripesinFigure1the‘pre-images’ofheadandtails.Figure1.Thepre-imagesofheadsandtails(Diaconis1998,p.803).\n354RomanFriggandCarlHoeferWhatfollowsfromtheseconsiderationsaboutthechanceofgettingheads?Kellerpresentsanargumentintwosteps.Thefirstistoregardtheinitialconditionasarandomvariablewithacontinuousprobabilitydistributionρ(ν,ω)withsupportintheregionshowninFigure1(i.e.ω≥0andν≥0).Thentheprobabilityforheads,p(H),isgivenbyp(H)=∫ρ(ν,ω)dνdω(3)BwhereBdenotestheblackregionsinω≥0andν≥0.MutatismutandisEquation(3)alsogivestheprobabilityfortails,p(T).Thesecondstepconsistsinshow-ingthatp(H)=p(T)=0.5.TothisendKellerprovesalimitingtheorem,basicallyshowingthatiftheboundariesoftheregionoverwhichtheintegralinEquation(3)iscalculatedisshiftedtowardsinfinity(i.e.ifweintegrateoverB’={(ν,ω):ν≥k,ω≥k}andletktendtowardsinfinity),thenp(H)=0.5nomatterwhatdistributionρ(ν,ω)wechoose.Thisresultbecomesintuitivelyplausiblewhenwerealisethatthestripesgetthinnerasthevaluesofωandνincrease(seeFigure1),andsotheintegralbecomeslesssensitivetofluctuationsinρ(ν,ω).Hence,inthislimitthereisauniqueprobabilityforheads.6WenowturntoadiscussionofHumeanchanceandthenreturntotheques-tionofhowtojustifyp(H)=0.5inSection4.ThemaindifferencebetweenourandKeller’sapproachisthatwemakeessentialuseoffactsabouttheHumeanmosaic(i.e.,thetotalityofallactualoroccurrentevents–seesection3.3)andtherebyavoidappealtoalimitingresult.3.HUMEANOBJECTIVECHANCEInthissectionweintroducetheconceptofHumeanObjectiveChance(HOC),onwhichourreconciliationofdeterminismandchanceisbased.7TheviewsdiscussedhereareanextensionofthoseintroducedinHoefer(2007),butherepresentedinawaythatpaysparticularattentiontothosefeaturesofthetheorythatbearontheissueofthecompatibilityofdeterminismandchance.3.1DefiningHumeanObjectiveChanceThedefinitionofchancethatwepresentinthissectiondiffersfromLewis’ca-nonicaldefinition(1994,p.480).Inpartthisisamatterofpresentation;butinpartitalsoresultsfromcorrectingcertainomissionsandmodifyingafewcentral6Diaconisetal.(2007)generalisethisresultbyrelaxingsomeoftheabovemodellingassumptionandtherebytakingintoaccounttheprecessionofthecoin.Thisaddsinter-estingfeaturestothemodel,butsincethemainfeaturesremainthesamewekeepusingthesimplemodeldiscussedinthissection.7Onemightarguethat‘objectivechance’isapleonasmsincechancesareobjectivebydefinition.Trueenough,butthephrase‘objectivechance’hasbecomecustomaryintheliteratureandsowesticktoithere.\nDeterminismandChancefromaHumeanPerspective355features.Threechangesareparticularlycrucial.First,wecorrecttheomissionofanyreferencetothePrincipalPrinciple(PP)inLewis’definition.InourviewPPisessentialforanunderstandingofobjectivechanceandthereforehastoappearinonewayoranotherinitsdefinition.Second,ourdefinitionisofchancesorchancelawsalone,andisnotadefinitionoflawsofnaturemoregenerally.Andfinally,ofcourse,ourdefinitionwillallowfortheretobegenuinechancesinaworldthatisdeterministicatbottom.Wereturntothesepointsinduecourse.Letebeanevent,forinstanceacoincomingupheadsoradielandingsothatitshowsthreespots.8Wedefinechanceasfollows.Definition1(HOC):Thechanceofevente,ch(e),isarealnumberintheinterval[0,1]suchthat:(1)thefunctionchsatisfiestheaxiomsofprobability,(2)ch(e)isthecorrectplug-inforXinthePrincipalPrinciple,and(3)thefunctionchsupervenesontheHumeanMosaicintherightway.ChancesthusdefinedareHumeanObjectiveChances(HOC);forbrevitywerefertothemsimplyas‘chances’.Weuse‘THOC’torefertotheentiretheoryofchancepresentedinthissection.Theelementsofthisdefinitionareinneedofexplication,andprovidingtheneededexplicationsisthetaskforthissection.Letusbrieflyindicatewhatthistaskinvolves.Thefirstclauseisstraightforward,butneverthelessnotentirelytrivial.Lewisthoughtitamajorproblemtoprovethatobjectivechancessatisfytheaxiomsofprobability,andhearguedatlengththatchancesindeedhavethisproperty.9Inourviewthereisnothingtoprovehere.THOCdefineschance,andwearefreetobuildintoadefinitionwhateverseemsnecessary.Afunctionthatdoesnotsatisfytheaxiomsofprobabilitycannotbeachancefunctionandsowesimplyrequirethatchsatisfytheaxiomsofprobability.Thesecondclauseneedsunpackingintworespects:weneedtointroducePP,andweneedtoexplicatewhatmakesaplug-inforXacorrectplug-in.Muchhangsonthis,andacarefulexpositionisimperative.Forthisreasonwededicatesubsec-tionstoeachpoint(Subsections3.2and3.4).Thethirdclauseisalsoproblematic.WefirsthavetointroducetheHumeanMosaic,thensaywhatwemeanbyafunctionsuperveningontheHumeanMosaic,andwethenneedtoexplicatethenotionofsuperveningontheHumeanMosaicintherightway.Thesecondclauseofthedefinitionentersheretoo,becausean8Twodisclaimersareinorder.First,nothinginwhatfollowsdependsonamoreprecisecharacterisationofevents.Second,weattributechancestoeventsbecausethislooksmostnaturalinthecaseswediscuss.Butnothinghangsonthat;wecouldtakeproposi-tionsinstead.Infact,aswillbecomeclearfromthecontext,incertainformulaebelowletterssuchaseandXwillstandforpropositionsdescribingeventsratherthandirectlyforevents.Thisisinconsequentialforourviewsonchance.9ForadiscussionofLewis’argumentsseeHoefer(2007,pp.560-62).\n356RomanFriggandCarlHoeferimportantpartofwhat‘therightway’meanshereis:insuchawayastopermitasolidargumentjustifyingPPtobemade.WeturntotheseissuesinSubsection3.3.3.2IntroducingthePrincipalPrincipleChances,firstandforemost,areguidestoaction.Welooktochanceswhenmakingdecisions:ifthechanceforraintodayis0.95Itakemyumbrellawithme,butifitis0.05Idonot.AsLewisinsisted,themostcentralandimportantrequirementonatheoryofchanceisthatitmakeitpossibletoseehowchancescanplaythisaction-guidingrole.ThisaspectofchancesisenshrinedinPP,whichestablishesaconnectionbetweenchancesandthecredencesarationalagentshouldassigntocertainevents,whereby‘credence’wemeananagent’ssubjectiveprobabilityordegreeofbelief.TheintuitiveideainPPisthatarationalagent’scredenceforaneventetooccurshouldbesetequaltothechanceofe,aslongastheagenthasno‘inadmissible’knowledgerelatingtoe’soccurrence.Definition2(PrincipalPrinciple):Let‘cr’standforarationalagent’scredence.ThePrincipalPrinciple(PP)istherulethatcr(e|X&K)=x,(4)whereXisthepropositionthatthechanceofeisx(i.e.X=‘ch(e)=x’),andKis‘admissible’knowledge.Beforespellingoutwhatwemeanbyadmissibleknowledge,letusaddsomeclarificationsaboutthepurposeofK.AtfirstsightitseemsunclearwhyKshouldappearinEquation(4)atall,andmoreneedstobesaidthefunctionthatKismeanttobeperform.ThepresenceofKshouldnotbeinterpretedasarequesttogatheraparticularkindofknowledgebeforewecanusePP.Onthecontrary,wealwayshaveknowledgeaboutsituations,andKsimplystandsforthesumofwhatwedefactohappentoknow.DependingonwhatkindofpropositionsKcontains,weshouldorshouldnotuseEquation(4)tosetourcredences.Theprescriptionissimple:ifKcontainsnoinadmissibleknowledgethenuseEquation(4);ifKdoescontaininadmissibleknowledgethendon’t.InthelattercasePPissilentabouthowtosetourcredences.Thequestionnowiswhatcountsas‘admissible’knowledge.Lewis’originalcharacterisationis:Admissiblepropositionsarethesortofinformationwhoseimpactoncredenceaboutout-comescomesentirelybywayofcredenceaboutthechancesofthoseoutcomes.(Lewis1980,p.92)Thischaracterisationhasgivenrisetocontroversy.Infact,Lewishimselflaterregardeditastooimpreciseandreplaceditwithatime-indexedversion,inpartinordertobeabletosaythatallpasteventshavechance0or1.Foradiscussionof\nDeterminismandChancefromaHumeanPerspective357Lewis’reviseddefinitionandtheissueoftimeseeHoefer(2007,553-5and558-60).Weherebuildonthisdiscussionandassumethatthesecorrectionsarenotonlyunnecessary,butalsowrong.Chancesattachtocircumstances(the‘chanceset-up’)andnottoworlds-at-specific-times.Theoriginaldefinitionofadmissi-bilityLewisgavewasessentiallyright.Chanceisaguidetoactionwhenbetterinformationisnotavailable.Sotheessenceoftherequirementofadmissibilityistoexcludetheagent’spossessionofotherknowledgerelevanttotheoccurrenceofe,thekindofknowledgethepossessionofwhichmightmakeitnolongersen-sibleordesirabletosetcredenceequaltoobjectivechance.Tousetheusual(andsilly)example:ifyouhaveacrystalballthat(youbelieve)reliablyshowsyoufutureevents,youmayhaveinadmissibleknowledgeaboutachanceeventsuchasthecoinflipaweekfromnow.Ifyourcrystalballshowsyouthecointossland-ingtailsandyoutrusttheball’srevelations,youwouldnotbereasonabletosetyourcredenceintailsto0.5forthatflip;youhaveinadmissibleknowledge.Thisexamplehelpsmakethenotionofadmissibilityintuitivelyclear,andalsopointstowardaveryimportantfactinourworld:inadmissibleevidenceisnotsomethingwetypicallyhave–ifwedid,thenchanceswouldberatheruselesstohave.Still,itispossibletogiveaslightlymoreprecisedefinitionofadmissibility(Hoefer2008,Ch.2)Definition3(Admissibility):ApropositionPisadmissiblewithrespecttoanout-come-specifyingpropositionEforchanceset-upS(Esaysthateventeoccurs)iffPcontainsonlythesortofinformationwhoseimpactonreasonablecredenceaboutE,ifany,comesentirelybywayofimpactoncredenceaboutthechancesofthoseoutcomes.Thisdefinitionmakesclearthatadmissibilityisrelativetoachanceset-upanditsattendantpossibleoutcome-events.Itisalsorelativetotheagentwhosereason-ablecredencefunctionisinvokedinPP.Theagent-relativityofadmissibilitymaybemoreorlessextreme,dependingonhowhighlyconstrainedacredencefunctionmustbeinordertocountas‘reasonable’or‘rational’.Forourpurposesagent-rela-tivityisnotgermane,andwewillassumethatallreasonableagentsagreeaboutwhetherapropositionPshouldorshouldnothaveanimpactoncredenceinE,whenPisaddedtoafurtherstockofbackgroundknowledgeK.3.3HumeanSupervenienceTheHumeanMosaic(HM)isthecollectionofeverythingthatactuallyhappens;thatis,alloccurrentfactsatalltimes.Thereisaquestionaboutwhatcredentialssomethingmusthavetobepartofthemosaic.Nothinginwhatfollowsdependsonhowthedetailsofthisissueareresolved.Whatdoesmatteristhatirreduciblemodalities,powers,propensities,necessaryconnectionsandsofortharenotpartofHM.Thatisthe‘Humean’inHumeansupervenience.\n358RomanFriggandCarlHoeferThesuperveniencepartrequiresthatchancesareentailedbytheoverallpatternofeventsandprocessesinHM;inotherwords,chancesareentailedbywhatactu-allyhappens.Wecanmakeacomparisonwithactualfrequentism,whichsatisfiesHumeansupervenienceinaparticularlysimpleway:theoverallpatternofeventsuniquelydeterminestherelativefrequencyofanevent,andhenceitsprobability.Actualfrequentismhasnofrequencytolerance,andhencefrequentistprobabilitiessuperveneonactualevents.Thiscontrastswithpropensitytheories,whichhavemaximalfrequencytolerance.THOCstrikesabalancesbetweentheseextremesbyrequiringthatHOC’ssuperveneonHM,butnotsimply:THOCpostulatesthatchancesarethenumbersassignedtoeventsbyprobabilityrulesthatarepartofaBestSystemofsuchrules,where‘best’meansthatthesystemoffersasgoodacombinationastheactualeventswillallowofsimplicity,strengthandfit.TheideaofaHumeanBestSystemofchancescanbeillustratedwithathoughtexperiment.Tothisend,weintroduceafictitiouscreature,Lewis’De-mon.IncontrasttohumanbeingswhocanonlyknowasmallpartoftheHumeanmosaic,Lewis’Demonknowstheentiremosaic.ThedemonnowformulatesallpossiblesystemsofprobabilityrulesconcerningeventsinHM,i.e.rulesassign-ingprobabilitiestoevent-typessuchasgettingheadswhentossingacoin.Inthemereformulationofsuchrules,nointerpretationofprobabilityisassumed.Therulesinthesesystemsassignnumberstoevents.Thesenumbershavetosatisfytheaxiomsofprobability–thisiswhytheyare‘probabilityrules’–butnothingoverandabovethisisrequiredatthisstage.Thenthedemonisaskedtochoosethebestamongthesesystems,wheretheBestSystem(BS)istheonethatstrikesthebestbalancebetweensimplicity,strengthandfit.Theprobabilityrulesofthesystemthatcomesoutofthiscompetitionasthebestsystemthen,bydefinition,become‘chancerules’,andthechanceofaneventesimplyisthenumberthatthischanceruleassignstoit.Thatis,thechancesforcertaintypesofeventstooccur(giventheinstantiationofthesetupconditions)simplyarewhatprobabilisticlawsofthebestsystemsaytheyare.Definition4(HumeanBS-supervenience):AprobabilityruleisHumeanBS-su-pervenientonHM(‘HBS-supervenesonHM’,forshort)iffitispartoftheBestSystem,i.e.thesystemthatstrikesthebestbalancebetweensimplicity,strengthandfitgivenHM.Clause(3)inDefinition1cannowbemadeprecise:thefunctionchHBS-super-venesonHM.Needlesstosay,muchdependsonhowweunderstandsimplicity,strengthandfit.Beforediscussingtheseconceptsinmoredetail,letusillustratetheleadingideaofHBS-superveniencewithanexample.Thequestionwehavetoaskishowcertainaspectsofevent-patternsinHMmaybecapturedbyaddingachanceruleaboutcoinflips.Coinsarefairlyubiquitousandwehavethecustomofflippingthemto\nDeterminismandChancefromaHumeanPerspective359helpusmakechoices.Sotheevent-typewecall‘agoodflipofafaircoin’iswide-spreadinHMaroundhere.Furthermore,itisafact,first,thatinHMtherelativefrequencyofeachdiscernibleside-typelandingupwardisverycloseto0.5and,second,thattherearenoeasilydiscernedpatternstotheflipoutcomes(itisnotthecase,forinstance,thatalongsequenceofoutcomesconsistofalternatingheadsandtails).THOCnowasksustoconsiderallpossibleprobabilityrulesforagivenclassofeventsandthenchoosetheonethatstrikesthebestbalancebetweensim-plicity,strengthandfit.Thereareofcourseinfinitelymanyrules.One,forinstancehasitthatp(H)=0.1andp(T)=0.9;anotherrulepostulatesthatp(H)=p(T)=0.5;andyetanothersaysthatp(H)istheactualfrequencyofheadsandp(T)istheactualfrequencyoftails.Giventhatthefrequencyofheadsandtailsisroughly0.5,thefirstrulehasbadfit;atanyrateitsfitisworsethanthefitoftheothertwo.Buthowdoweadjudicatebetweenthesecondandthethirdrule?Atthispointconsiderationsofstrengthcomeintoplay.InfacttheremaybeanevenbetterchancerulethatcouldbepartoftheBestSystem,whichwouldembracecoinsanddiceandtetrahedraanddodecahedraandothersuchsymmetric,flippable/rollablesolids.Therulewouldsaythatwheresuch-and-suchsymmetryistobefoundinasolidobjectofmiddlingsizewithnpossiblefacesthatcanlandupward(ordownward,thinkingoftetrahedra),andwhensuchobjectsarethrown/rolled,thechanceofeachdistinctfacebeingtheonethatlandsup(ordown)isexactly1/n.Givenwhatweknowaboutdiceandtetrahedraandsoforth,itisquiteplausiblethatthisrulebelongsintheBestSystem;anditentailsthecoin-flipchances.Soitenhancesbothsimplicityandstrengthwithoutmuchlossinfit,andhenceonbalanceitisbetterthanthesystemwhichsetschancesequaltorelativefrequencies.Hence,thechanceofheadsonafairflipofacoinwouldseemcer-tainlytoexist,andbe0.5,inaBestSystemforourworld.Howarewetounderstandsimplicity,strengthandfit?Letusbeginwithsim-plicity.Thisisanotoriouslydifficultnotiontodefineprecisely,yetwethinkthatthereisenoughonecansayaboutittomakeTHOCtick.Asweunderstandit,simplicityhastwoaspects,simplicityinformulationandsimplicityinderivation.Theformeriswhatisusuallymeantwhensimplicityargumentsareputforward:alinearrelationbetweentwovariablesissimplerthanapolynomialoforder325,ahomogenousfirstorderdifferentialequationissimplerthananon-linearintegro-differentialequation,etc.Itisnoteasytopindownwhatgeneralruledrivesthesejudgments,butthisdoesnotrepresentaseriousobstacletousbecausenothinginwhatfollowsdependsonsimplicityjudgmentsofthiskind.Anothercomponentofsimplicityinformulationishowmanydistinctprobabilityrulesasystemcontains.Ceterisparibus,thefewerrulesasystemhasinit,thesimpleritis.Thesecondaspectofsimplicity,simplicityinderivation,focusesonthecomputationalcostsincurredinderivingadesiredresult.Thequestionis:howmanydeductivestepsdowehavetomakeinordertoderivethedesiredconclusions?Somesystemsallowforshorterderivationsthanothers.Itisimportantnottoconfusesimplicityinthissensewithasubjectivenotionofaderivationbeing‘easy’or‘difficult’.Theissue\n360RomanFriggandCarlHoeferatstakehereisthenumberofdeductivestepsneededtoderiveaconclusion,andthisisacompletelyobjectivequantity,whichcouldbequantified,forinstance,byusingameasuresuchasKolmogorov’scomputationalcomplexity(roughly,thelengthoftheshortestprogrammecapableofderivingtheresult).Simplicity(inthislattersense)couldalwaysbeimprovedbycuttingperfectlygoodchancerulesoutofthesystem.However,ingeneralimprovingsimplicityinthiswayisnotagoodstrategybecauseitcomesattoohighacostintermsofstrength.Thestrengthofthesystemismeasuredbyitsscope.Thewiderthescopeofthesystem,thestrongeritis.Inotherwords,thelargerthepartofHMthatthesystemisabletoaccountfor,thebetteritfaresintermsofstrength.Theaboveex-ampleillustratesthepoint:asystemthatcoversonlycoinsisweakerthanasystemthatalsocoversotherchancesetupssuchasroulettewheels,dice,etc.TheBestSystemshouldnotonlyascribechancestolotsofeventtypes,anddosoinassimpleawayaspossible;itshouldascribetherightchances!Butwhicharetherightchances?Everysystemassignsprobabilitiestopossiblecoursesofhistory,amongthemtheactualcourse.WithLewis,wenowpostulatethatthefitofthesystemismeasuredbytheprobabilitythatitassignstotheactualcourseofhistory,i.e.byhowlikelyitregardsthosethingstohappenthatactuallydohappen.Asanillustration,consideraHumeanmosaicthatconsistsofjusttenout-comesofacoinflip:HHTHTTHHTT.Itfollowsimmediatelythatthefirstsystemabove(p(H)=0.1andp(T)=0.9)hasworsefitthanthesecond(p(H)=p(T)=0.5)since0.150.95<0.510.Thisexamplealsoshowsthatasystemhasbetterfitwhenitstaysclosetoactualfrequencies,aswewouldintuitivelyexpect.10Sothewaysinwhichweevaluatesystemsisobjectiveandnoappealto‘prag-matic’orspecifically‘human’valuesorlimitationshasbeenmade.Nevertheless,weaccepttwo(notverycontroversial)assumptionsthatassurethattheBestSys-tem,whateveritsconcreteform,sharesatleastsomeessentialcharacteristicswithscienceaswe,humans,knowit.Thefirstassumptionisontologicalpluralism,whichdeniesthatonlybasic/microentitiesexist.Somehard-headedreductionistsdenythatanythingexceptthebasicmicroentitiesexist.Thus,chairs,rivers,cats,trees,etc.aresaidnottoexist.Wedenythis.Thatcoinsconsistofatomsdoesnotmakecoinsunreal.Coinsexist,nomatterwhatmicrophysicstellsusabouttheirultimateconstitution,andsodorivers,chairs,andcats.Hence,eveninaclassicalworld,HMconsistsofmuchmorethanelementaryparticlesandtheirtrajecto-ries.Thesecondassumptionislinguisticpluralism,thepositthatthelanguageinwhichtheDemonformulatesthesystemsthatsubsequentlyenterintothesim-plicity-strength-fitcompetitioncontainstermsformacroscopickinds.Thatis,thelanguagehasnotonlythevocabularyofmicrophysics,butalsocontainstermslike10Elga(2004)aguesthatthisdefinitionoffitrunsintoproblemsifthereareinfinitelymanychancyevents,andsuggestsasolutiontobasedonthenotionofatypicalse-quence.Thisconcernisorthogonaltotheproblemswediscussinthispaperandhencewewillnotpursuethematterfurther.\nDeterminismandChancefromaHumeanPerspective361‘coin’and‘river’.Sowenotonlybelievethatmacroobjectsexist,wealsoequipthedemonwithalanguageinwhichhecantalkabouttheseassuigenerisenti-ties.113.4JustifyingthePrincipalPrincipleThereiscontroversynotonlyoverthecorrectformulationofPP,butalsooveritsstatus.Strevens(1999)arguesthatitisnomorepossibletoofferasoundargumentjustifyingPPthanitistojustifyinduction,andthatwethereforehavetoacceptitassomethinglikeafirstprinciple.Butnoteverybodysharesthispessimism.Infact,webelievethattheuniquefeaturesofHOCpermitademonstrationthatitisirrationalnottoapplyPPwhenitsconditionsarefulfilled.Spaceprecludesafulldiscussionhere,sowewillsimplypresentabriefversionoftheargument;foranin-depthdiscussionsee(Hoefer2008,Ch.4).Aswehaveseeninthelastsubsection,itisaresultofacarefulanalysisofwhatitmeansforthefunctionchtosuperveneontheHumeanMosaicintherightwaythatwheneverthereisalargenumberofinstancesofachancesetup,thechanceofacertainoutcomeisclosetotherelativefrequencyofthatoutcome.Forthisreason,THOCcanbeunderstoodasa(major)sophisticationoffinitefre-quentism,andunderstandingwhyPPisjustifiedforHOCbeginsbyrecallingthisaffinity.12TherearetwowaysofjustifyingPPbasedonthisaffinity,an‘apriori’anda‘consequentialist’one.TheformerissimilartothejustificationofPPHowsonandUrbach(1993)giveforvonMisesfrequentism.Asubjectivedegreeofbeliefcorresponds(bydefinition)totheoddsatwhichanagentfeelsabetoneithersideofaquestion(Eversusnot-E)wouldbefair.AnagentwhohasnoinadmissibleinformationpertinenttotheoutcomesofalongseriesofinstancesofchancesetupSshouldhavethesamedegreeofbeliefintheE-outcomeineachtrial–havingareasontoassignahigherorlowerdegreeofbelieftoEonaspecifictrialautomati-callyandbydefinitionamountstopossessinginadmissibleinformation.Hence,ifanagentassignsdegreeofbeliefptooutcomeEinasingletrialofchancesetupS,heshouldassignthesamecredencetoanE-outcomeineachinstanceofa(nindefi-nitely)longseriesoftrialsofS;nottodosoistotakeoneselftohaveinformationrelevanttoanE-outcomethatdoesnotcomefromE’schance(whichbystipula-tionisthesameineachtrialofS),andhencetohaveinadmissibleinformation.HavinginadmissibleinformationmakesPPinapplicable,sowemayassumefor11ForfurtherdiscussionoftheissueofthelanguageusedinformulatinglawsseeLewis(1983).12Sinceclause2ofourdefinitionofHOCabovestipulatedthatch(e)isthecorrectplug-inforthePrincipalPrinciple,onemightexpectaquickandeasydemonstrationthatHOC’ssatisfyPP:itistruebydefinition!Clearly,thisisabittooeasy.Thetwojusti-ficationsofPPforHOCofferedinthissectionaresubstantial,departingfromconnec-tionsbetweenHOCandfrequenciesofevents,andareentirelynon-circular.\n362RomanFriggandCarlHoefertherestoftheargumentthattheagentdoesnotvaryhiscredenceinE-outcomesfromtrialtotrial.SotheagenttakesbettingonEineachtrialofanindefinitelylongseriesatoddsp:(1-p)tobefair.Assumethathebetsonthesamesideinalltrialsinthesequence,i.e.eitheronEinalltrialsornot-Einalltrials.Becausetheagentthinksthebetisfair,hemustthinkthatthereisnoadvantagetobettingonEratherthannot-E(orviceversa);thatis,hemustbeindifferenttowardswhichsideofthebethetakes.ByassumptionthereisachanceforgettingEonatrial,ch(E)=q.FromtheaccountofTHOCaboveweknowthattherelativefrequencyofE’sinanindefinitelylongsequenceisoftrialsmustbeequal(oratleastverycloseto)thechanceofE.Itisasimpleresultofprobabilitycalculusthatifagentsdon’tbetinaccordancewithrelativefrequencies,thenonesideofthebetisdoingbetter.Thiscannotbeiftheagentbelievesthebettobefair.Itisthenasimplearithmeticfactthatifqdiffersnon-triviallyfromp,andtheagentbetsonEatp:(1-p)oddsthroughoutthelongseries,thentheagentwillcertainlylose(orwin)inthelongrun.Theagent,understandingTHOCandthatch(E)=q,mustknowallthistoo;buthecannotbelievethisandyetbelievethelongseriesofbetstobefair.Soifp≠q,theagentholdscontradictorybeliefs,whichisirrational.Sotheonlyrationalassignmentofprobabilitiesisp=q,asPPprescribes.Theconsequentialistargumentismorestraightforward.Itasksus,inthespiritofHumeanism,tolookatHM,whichnotonlycontainsoutcomesoftrialsbutalsoagentsplacingbets.Ifwelookatallagentsplacingbetsacrosstheentiremosaicandcheckonhowtheyaredoing,wewillseethatthoseagentswhosettheircredencesequaltothechancesobtain–atmostplacesandtimes,atleast–betterresultsthanthosewhoadoptcredencessignificantlydifferentfromthechances.Inotherwords,inthewiderdomainjustasinLasVegas,ifonehastogambleonchancyevents,onedoesbestifoneknowstheobjectiveprobabilities.Forthisreasonitisrationaltosetone’scredencestoobjectivechances,asPPrequires.133.5TheEpistemologyofHOC’sLetusclosethissectionwithabriefremarkabouttheepistemologyofHOC’s.Atfirstsight,anapproachtoprobabilitywhosecentralconceptsaredefinedintermsofeverythingthatactuallyhappensatanypointintimeandatanyspatiallocation–theHM–andanomniscientcreature–Lewis’Demon–maystrikesomeasratherdisconnectedfromactualhumanendeavours.Thisimpressionismistaken.Needlesstosay,theappealtoHMandLewis’demonareidealisations,andonesthattakeusratherfarawayfromouractualepistemicsituation.ButthisdoesnotturnTHOCintoanepistemicpipe-dream.First,thelimitationsofactualhumanexperienceareafactorthateveryepistemologyhastocopewith.Inparticular,alsothosepositionswhobelieveinmetaphysically‘thick’lawsandprobabilities13CaveatsanddetailsoftheconsequentialistargumentarediscussedinHoefer(2007,sec.5)and(2008,Ch.4).\nDeterminismandChancefromaHumeanPerspective363(universals,causalpowers,etc.)havetobasetheirviewsonthenatureandchar-acteroftheseonactualexperienceandthereisthepossibilitythatfutureeventsmayprovethemwrong.Anyviewaboutprobabilisticlaws–Humeanornot–hastobaseclaimsabouttheseonourlimitedactualexperience,andthisinvolvesaninductiveleap.Howtohandlethisleapisofcourseatime-honouredphilosophicalpuzzleonwhichmuchinkhasbeenspilled,andthereisnoroyalroadtosuccess.ThepointtostresshereisthattheHumeanisnotalonewiththisproblem.Second,therequirementthatonlyoccurrentpropertiesbepartofHMisinharmonywithscientificpracticeasweknowit,sinceoccurrentpropertiesarewhatsciencecanobserve.InthisrespectTHOCisevenclosertoactualsciencethanapproachesthatpostulatemodalentitiesthatsciencecanneverobserve.Third,therulesthataregiventotheDemonhaveanobvious‘humanflavour’:theomniscientDemonhimselfwouldprobablynotcareaboutsimplicityandstrengthsinceheknowseverythinganyway.TheserequirementsaremetatheoreticalvirtueshumansvalueinscienceandhencewhattheDemonisaskedtodoisintheend‘humanstyle’scienceasbestasitcanbedone.HencetheDemon’sactivityisnotdifferentinkindfromtheendeavoursofhumanscientists;thedifferenceisthathecanperformwithperfectionwhatwecandoonlyinadequately.4.COINFLIPPINGFORHUMEANSLetusnowreturntoflippingcoins.AstrikingfeatureofthediscussionsofaristhealmostcompletemismatchbetweenhowprobabilitiesforthecoinflipweretreatedinSections2and3respectively.ThetreatmentinSection2startedwithadeterministicmechanicalmodelandsoughttoretrievethe50/50chancerulefrommechanicallawsplusaprobabilitydistributionoverinitialconditions.TheapproachtakeninSection3didnotmentionmechanicsatallandinsteadfocussedthepatternofoutcomesinHM.Atleastonthefaceofittheseapproacheshavelittleincommonandsothequestionariseswhethertheyarecompatibleatall,andifsohow.Inthissectionwearguethattheyarecompatible,and,whatismore,thattheyareactuallycomplementary.Inordertoreachthisconclusionweneedtoanalysethetwoaccountsandtheirstatusvis-a-viseachotheringreaterdetail.Tofacilitatethediscussion,wesetupatemporarydebatebetweentwoviewpoints:‘mechani-cism’versus‘macro-statistics’,theirproponentsbeing‘mechanicists’and‘macro-statisticians’.Mechanicistsarelikelytoarguethattheirpointofviewisprivilegedsincetheiraccountisbasedonfundamentallaws:byassumptionweliveinaclassicaluniverseandsoHMconsistsoftrajectoriesofobjects,amongthemthetrajecto-riesofcoins,andclassicalmechanicsisthefundamentaltheoryofthisuniverse.1414Thosewhoalsoupholdmicro-reductionism–theviewthatmatterconsistofatomsand\n364RomanFriggandCarlHoeferChancerules,ifthereareanyatall,havetobeformulatedintermsofthefunda-mentalentitiesofHMandinthelanguageofthefundamentaltheorydescribingthem.Equation(3),supplementedwiththespecificationofaparticulardistribu-tionρfitsthebill:itisarulethatassignsprobabilitiestogettingheadswhentoss-ingacoin,anditdoessosolelyintermsofbasicmechanicproperties.Theruleissimple,hasgoodfit,andsinceKeller(1986,pp.194-6)showsthatitcaneasilybegeneralisedtootherchancesetupssuchasroulettewheelsitalsohasstrength.SowehavegoodreasonstobelievethatwithintheclassofallprobabilityrulesEqua-tion(3)istheonethatstrikesthebestbalancebetweensimplicity,strengthandfit,andhencetheprobabilitiesitdefinesarechancesinthesenseofTHOC.Themacro-statisticiandisagreeswiththispointofviewfortworeasons.Thefirstobjectionisconceptual,thesecondtechnical.Theconceptualobjectiontakesissuewiththemechanicist’sreductionistoutlook.Eveniftheworldisclassicalatbottomandclassicalmechanicsisthefundamentaltheoryoftheuniverse,itdoesnotfollowthateverythingthatcanbesaidaboutHMhastobesaidinthelanguageofthefundamentaltheory.Morespecifically,themacro-statisticianadoptsameth-odologicalpluralism(MP),thepositionthatprobabilityrulescanbeformulatedinamacrolanguagepertainingtoacertainlevelofdiscourse,andthatprobabilitiesthusintroducedaresuigenerisHOC’siftheprobabilityrulesinquestionstrikethebestbalancebetweensimplicity,strengthandfitrelativetoallothersystems.Todothis,theyneednotprovelogicalindependencefrommicro-levelchancerules;theyneedonlywinoutincompetitionwithalternaterulesformulatedinthesamelanguage,thatofthemacro-level.Onthisview,then,the1/nruleforgamblingdevicesisasuigenerischancerulebecauseitstrikesabetterbalancebetweenthethreebasicmetatheoreticalvirtuesthananyotherprobabilityruleformulatedinthelanguageofcoins,wheels,throws,etc.(Wecomebacktothisprincipleatlengthbelow.)Themacro-statistician’stechnicalobjectiontomechanicismturnsonthesta-tusandmathematicalformofthedistributionρ(ν,ω)inEquation(3).Atageneralleveltheworryisthatthemechanicistis‘cheating’.Noprobabilitiescanevercomeoutofapurelydeterministicapproach(‘noprobabilitiesin,noprobabilitiesout’),andthemechanicistjustputstheminbyhandwhenheintroducesρ(ν,ω),whichisnotwarrantedby(orevenrelatedto)anypositofmechanics.Thereforetheintroductionofρ(ν,ω)isanadhocmanoeuvre,unmotivatedfromthepointofviewofmechanics.And,asisoftenthecasewithsuchmanoeuvres,itmaywellraisemorequestionthanitanswers.Thefirstproblemwithρ(ν,ω)isthatitisnotclearwhatitisadistributionfor.Themostbasicquestionwehavetoaskabouteveryprobabilitydistributionis:whataretheseprobabilitiesprobabilitiesfor?Anditisnotclearwhattheanswerinthecaseofthecoinis.Wemighttakeitthatthebehaviourofmacroscopicobjectslikecoinseventuallyhastobeexplainedintermsofthebehaviourofitsmicroconstituents–canreplacethetrajectoryofacoinbythebundleoftrajectoriespertainingtotheatomsmakingupthecoin.Mutatismutandistheargumentsremainthesame.\nDeterminismandChancefromaHumeanPerspective365tobegivingtheprobabilityofacoinflip’shavinginitialconditionswithinagivenrangeν+dν,ω+dω.Butnothinginmechanicscangroundsuchadistribution.Theproblemwiththemathematicalformofρ(ν,ω)isthefollowing.Keller’slimitingargumentshowsthatthemathematicalformofρ(ν,ω)isimmaterial,andhencethequestionofwhatρ(ν,ω)tochosebecomesobsolete.However,thislim-itingargumentisofnorelevancetoactualcointosses.Diaconishasshowninexperimentsthatfortypicalcointossestheinitialupwardsvelocityvisabout5mphandthefrequencyωliesbetween35and40revolutionspersecond(1998,p.802).Thisisveryfarawayfrominfinity!Theproblemisthatoncewerevoketheinfinitelimit,itisnolongerirrelevantwhatρ(ν,ω)onechooses.Sowhichρ(ν,ω)istherightonetoplugintoEquation(3)?Intuitivelyonewouldchooseauniformdistribution.Foroneitissimple;foranother,itwouldgive(roughly)a0.5probabilityforheadssince,asbecomesobviousfromFigure1,theblackandthewhitestripesoccupyapproximatelythesamearea.Butnothinginthemechanicalapproachjustifiesthisassumption.Letusnowstepback,evaluatetheargumentsoneitherside,andexplainhowthetwoviewseventuallycometogether.Takethemechanicist’sinsistenceonfun-damentallawsfirst.HewillobjecttoMPonthegroundsthatthechanceforheadsisnotindependentofthemicrophysicsoftheworld.Surely,sotheargumentgoes,theremustbesomedependencethere!Ifthephysicsofourworldwasvastlydif-ferentfromwhatitis,thenthechanceforheadsshouldbedifferenttoo!Thereisagrainoftruthinthis,butwemustnotbemisled.Thephysicsofourworldmightbevastlydifferent,andyet(forwhateverreason)thepatternofheads-andtails-outcomesinHMmightbeexactlythesame;inwhichcase,thechanceswouldbethesame.Inmostreasonablyimaginablecounterfactualscenarios,thephysicswillmattermuchlessthantheactualpatternofoutcomesinHM.Macrolevelfactsdependontologicallyonmicro-levelfacts–intheobviouscomposi-tionalsense–butinourworldtheydonotdependonthemconstitutively(i.e.,themacrochancefactsarenotentaileddirectlyorindirectlybyfundamentalphysics;theydependonthepatternofmacroeventsnomatterwhatthemicrophysicsis.).Oncethisisrealised,theotherproblemshaveelegantsolutionstoo.WecanchoptheGordianknotinfourcuts.First,withthemacro-statisticianweaffirmMP,fromwhichitfollowsthatchancesformacroeventslikecoinflipsdependontheoutcomepattern,notonthedetailsoftheunderlyingphysics.(WejustifyMPbelow.)Second,withthemechanicistwetakeontologicaldependenceseriously.Thequestionishowtotakethisintoaccount,andthisiswhereanewelemententers.WesharewiththemechaniciststheviewthatEquation(3)–andsimilarequations–matter,butinterpretthemdifferently.Thisequationdoesnotgiveusthechanceforheads.Wedon’tneedtobegivenanything–wehavethechance,andthechanceis(constitutively)independentofthemicrophysics.Rather,weseeEquation(3)bothasa‘consistencycheck’andanexplanation.Letustaketheseinturn.ThedifferentpartsofaBestSystemhavetobeconsistentwitheachother(whichisnottosaythat\n366RomanFriggandCarlHoeferonehastobederivablefromtheother).Forthisreason,wheneverthesetupcondi-tionsofamacro-levelchanceruleandamicro-levelchancerulearethesame(ex-tensionallyequivalent),thenthechancestheyascribemustagreeorbeverynearlyinagreement.This,ofcourse,doesnotruleoutthepossibilityofminoradjust-ment.Forexample,assumeweadoptedthe50/50ruleforheadsandtails.Nowweknowforsurethatwegetreductiverelationsrightandwehavethecorrectmicrotheory,andbasedonthesewefind49/51.Thisisnorealconflictbecausethereissomeflexibilityaboutthemacrochancesandifthereareverygoodoverallreasonsformakingadjustments,thentheHumeancanmakethese.Butthereisabreakingpoint:ifthemicrotheorypredicts80/20,wehavetogobacktothedrawingboard.Thesecondelementisexplanation.Wedon’twanttoplacetoomuchemphasisonthis,butthereisthepervasiveintuitionthatifamacroresultcanbederivedfromamorefundamentaltheory,thereisexplanation.Thosewhosharethisintui-tion–amongthemus–canseeEquation(3)asprovidinganexplanation.Thosewhodon’tcanrenounceexplanatorygoalsandrestcontentwiththeconsistencyrequirement.Thethirdcutisthatthemechanicisthastoadmitthattheintroductionofρ(ν,ω)isastepbeyondmechanicsandassuchρ(ν,ω)hastobejustifiedelsewhere.Butfarfrombeingaproblem,thisactuallyisanadvantage.Whenthinkingaboutρ(ν,ω)inthe‘THOCway’,weimmediatelyhaveanaturalinterpretationofρ(ν,ω):itistherelativefrequencyofcertaininitialconditions.Ofcourseallactualinitialconditionsareacollectionofpointsinthev−ωplane,andnotacontinuousdistribution.Butarguablyacontinuousdistributionismuchsimpler(inthesenseofsimplicityinformulation)thanahugecollectionofpoints,andsotheHumeancanargueconvincinglythatfittingasuitablecontinuousdistributionthroughthepointsmakesthesystemsimplerandstronger.ThisdistributionthenisjustanelegantsummaryoftheactualinitialconditionsofallcoinflipsinHM.Fourth,thecommonintuitionthatthereissomethingepistemicaboutthechanceofgettingHeadsonthisflip–afterallithasoneandonlyoneinitialcondi-tionandgiventhisinitialconditionitisdeterminedwhetheritcomesupheadsortails–isaddressedbypayingcloseattentiontoTHOC’sprescriptionaboutwhentousechancestoguideourcredence.Informationaboutthepreciseinitialcondi-tionofagivencoinflipiscertainlyinadmissible:suchinformationlogicallyim-pliesthecointossoutcomeandhenceprovidesknowledgeabouttheoutcomeofatossthatdoesnotcomebywayofinformationaboutchances.Thecrucialpointisthatintypicalsituationsinwhichwetossacoin,wejustdon’thaveinadmissibleinformation,andthatiswhyweusechancesandPPtosetourdegreesofbelieve.Soweusechanceswhenwelackbetterknowledge.Letusillustratetheadmissibilitypointinsomemoredetail.Considerthesce-nariodescribedinSection2,andanagentAwhohasonlytheusualsortofknowl-edgeinhisbackgroundKandwhoneedstodecidehowtobetonthecoinflip.AshouldapplyPP,clearly,andsethiscredencesforheadsandtailsoutcomesto0.5.ButnowconsideragentL,aLaplace-demon-in-training,whoalsomustdecide\nDeterminismandChancefromaHumeanPerspective367howtobet.LknowsallthatAknows,but–crucially–Lalsoknowstheexactmicro-stateoftheworld(orabigenoughlocalregionofit)justpriortotheflip,andknowsthelawsofNewtonianmechanics.ShouldLsethercredencesforheadsandtailsoutcomesequalto0.5?Evidentlynot!Shecancalculate,onthebasisofherbackgroundK,preciselywhatwillhappen.Let’sassumeshecalculatesthatthecoinwillinfactlandheads.Lhasinadmissibleknowledge.15Shehasinforma-tionrelevanttowhetherthecoinwilllandheads(maximallyrelevant!),andtheinformationisnotrelevantbywayoftellingherabouttheobjectivechances.SoLshouldnotapplyPP;andthisisintuitivelytherightverdict.Theconclusionisnotthattheobjectivechanceofheadsis1.Itisthat(giventhepaststateandthelaws),thecoinwilllandheads;andanyonewhoisawareofthesefactsshouldsettheircredenceinheadsto1(astherulesofsubjectiveprobabilityrequire),andnotto0.5.16Thetruthofdeterministiclawsentailsthat,givenacomplete-enoughstateofaffairsatamomentoftime(andperhapsboundaryconditions),futureeventsarefullydetermined.AndthisentailsthatifyoucangetsuchLaplace’s-demonstyleinformation,andifyoucanactuallycalculateanythingwithit,thenyoumayhavebetterinformationwithwhichtoguideyourcredencesaboutfutureeventsthantheinformationHOC’sgiveyou.Whatisentailed,however,isnotthatobjectivechancesdonotexist,butratherthatcertaingodlikebeingsmaynothaveanyuseforthem.Wehumans,alas,neverhavehadnorwillhaveeithersuchinformationaboutinitialconditions,orsuchdemoniccalculationalabilities.Forus,itisagoodthingthatobjectivechancesexist,andthatwecancometoknow(anduse)them.Withthesepointsinmind,nowwecanseehowdeterminismandnon-trivialob-jectivechancesarecompatible,andwealsoseethattheadmissibilityclauseinPPplaysacrucialroleinthat.WenowturntoadefenceofMP,whichwemerelystatedabove.Whyshouldwesubscribetothisprinciple?Whywouldabestsystemcontainanythinglikechance-rulesaboutcoinsandothermacroobjects?Letusdistinguishtwocases,aworldinwhichphysicalistreductionismaboutchanceistrue,andoneinwhichitisfalse.Physicalistreductionismaboutchanceistheclaimthatallchance-factsariseoutofthelawsofphysics.Physicalistreductionismquitegenerally(notmerelyaboutchance)ispopularinparticularwithelementaryparticlephysicists;seeforinstanceWeinberg(1994).Ifreductionismofthiskindisfalse,thenitisobviousthatthebestsystemwouldcontainrulesaboutmacroobjects:theserulesdonotfollowfrombasiclawsofphysicsandthereforeputtingthemintoasystemwillgreatlyincreaseits15AccordingtoLewis’officialdefinitionofadmissibility,informationaboutlawsofna-tureandaboutpaststatesoftheworldarefullyadmissible,henceLdoesnothaveinadmissibleinformation.ThisadjudicationmakesitimpossibleforLewistoretainnon-trivialchancesifthetruelawsofnaturearedeterministic.ForadiscussionofthispointseeHoefer(2007,pp.553-555).16Formally,cr(H|XK)=1isrequiredbytheprobabilityaxioms,sinceK⊃H.Weempha-size,itisnotcorrectbycontrasttosaythatK⊃[ch(H)=1.0].\n368RomanFriggandCarlHoeferstrength.Themoredifficultcaseisifphysicalistreductionismistrue.Iftherulesaboutcoinsandwheelsarebutcomplicatedapplicationsofthelawsofphysics,whywouldwehavesuchrulesinourbestsystem?17Thisseemstomakethesys-temlesssimplewithoutaddingstrength.Thereasontoputtheminneverthelessiswhatweabovecalledsimplicityinderivation:itishugelycostlytostartfromfirstprincipleseverytimeyouwanttomakeapredictionaboutthebehaviourofaroulettewheel.Sothesystembecomessimplerinthatsenseifwewriteinrulesaboutmacroobjects.Thereisalsoamoreintuitiveargumentwhythisindependenceofchancesfrommicrophysicsiscorrect.ItisthebasicpositofHumeanismthatthechanceofacertaineventHBS-superveneonthepatternofoccurrenceofeventsofthesamekindinHM,andassuchthischanceisindependentofhowtheseeventsrelatetootherfeaturesofHM.InourconcreteexamplethismeansthatthechanceofheadsonlydependsonthepatternofheadsinHM,orperhapsthepatternofoutcomesinrolls/flipsofn-sidedsolidswiththeappropriatesymmetriesandnotontherelationthat‘obtainingheads’bearstootherpartsofHM,inparticularthebasicmechani-calpropertiesofmatter.Asnotedabove,thesesortsofpatternsmayobtaineveninworldswithradicallydifferentmicro-laws.ImagineauniverseinwhichmatterisacontinuumandobeyssomethinglikethelawsofCartesianphysics;imaginethatcoinsexistinthisuniverseandaretossedrepeatedly.Despitethebasicphysicsbeingverydifferent,supposeitturnsoutthattheoverallpatternofoutcomesofrolls/tossesofsuchn-sidedobjectsinthecontinuumuniverse’sHMisverysimilartothepatterninouruniverse.Whatwouldthechanceofheadsbeinthecontinuumuniverse?Clearlyitwouldbegivenbythe1/nrule,sincethisisthebestrulerela-tivetothatHM,irrespectiveofthemicro-constitutionofmatter.5.ENVOYAswehaveindicatedintheintroduction,thispaperisaboutmorethancoins.Infactexactlythesameconsiderationscanbeusedtoexplainchanceinstatisticalmechanics(SM).Afullexpositionofthistheoryisbeyondthescopeofthispaper,butwewouldliketobringourdiscussiontoaclosebyverybrieflyindicatinghowtheinsightsgainedwiththeexampleofthecoincarryovertoSM.18Consider17Itmaybehardtoseehowprobability-factscouldfollowfromfundamentalphysicallaws,orlawsplusinitialconditionseven,ifthelawsarefullydeterministic.Wedobelievethat‘noprobabilitiesin,noprobabilitiesout’holdshere.Butonemightpositafundamental-physicsprobabilitylawasasupplementtothedeterministiclaws,pre-ciselyinordertoallowderivationofprobabilitiesforavarietyofphysicaleventtypes,includingperhapsmacroevents.Loewer’sversionofBestSystemHumeanismdoespreciselythis;seeLoewer(2001).18ForadetaileddiscussionofstatisticalmechanicsseeUffink(2006)andFrigg(2008a).\nDeterminismandChancefromaHumeanPerspective369atypicalSMsystem,forinstanceagasincontainer.Thegasconsistsofabout1023molecules.Thesemoleculesbouncearoundundertheinfluenceoftheforcesexertedontothemwhentheycrashintothewallsofthevesselandwhentheycol-lidewitheachother.Themotionofeachmoleculeundertheseforcesisgovernedbythelawsofmechanics.Hencethegasisalargemechanicalsystem:itsstateisfullyspecifiedbyapointinits(6×1023-dimensional)phasespace–inthiscontextreferredtoasits‘micro-state’–anditsevolutionovertimeisfullydeterminedbythelawsofmechanics.Atthesametimethesystemisalwaysinacertainmacro-state,whichischar-acterisedbythevaluesofmacroscopicvariables,inthecaseofagaspressure,tem-perature,andvolume.Itisoneofthefundamentalpositsof(Boltzmannian)SMthatasystem’smacro-statesupervenesonitsmicro-state,meaningthatachangeinthemacro-statemustbeaccompaniedbyachangeinthemicro-state.Forin-stance,itisnotpossibletochangethepressureofasystemandatthesametimekeepitsmicro-stateconstant.Hence,toeverygivenmicro-statetherecorrespondsexactlyonemacro-state.Thisdeterminationrelation,however,isnotone-to-one.Infactmanydifferentmicro-statescancorrespondtothesamemacro-state.Wenowgrouptogetherallmicro-statescorrespondingtothesamemacro-state,whichyieldsapartitioningofthephasespaceintonon-overlappingregions.Wecanthendefineanentropyfunction(theso-calledBoltzmannentropy)thatassignsapar-ticularentropyvaluetoeverymacro-state.Systemscharacteristicallystartoffinalowentropystateandthenevolveintoequilibrium,themacro-statewithmaximumentropy.TheSecondLawofthermo-dynamicstellsusthatthisiswhatinvariablymusthappen.OneofthecentralaimsofSMistoshowthattheSecondLaw–whichisapurelymacroscopiclaw–actu-allyisaconsequenceofthemechanicalmotionofthemoleculesofthegas,anditdoessobyshowingthattheapproachtoequilibriumisoverwhelminglylikely.Andthisiswherewemakecontactwiththecoinexample.Inordertojudgesomethingaslikely,trivially,wemustintroduceprobabilities.SMdoesthisbyputtingauni-formprobabilitymeasureovertheregionofphasespacewhichcorrespondstothesystem’sinitiallowentropystate,andthenaimstoshowthatmicroconditionsthatlieontrajectorieswhicheventuallymovetowardsequilibriumareoverwhelming-lylikely.Thelogicofthisislikeinthecaseofthecoin,theonlydifferencebeingthatwesortinitialconditionsintoonesthatbehaveastheSecondLawsrequiresandonesthatdon’t,ratherthanintoonesthatyieldheadsandonethatyieldtails.Letusthenmarktheonesthatbehaveasweexpectwhiteandtheotheronesblack.Wethenputameasureovertheseallinitialconditionsofthesamekindasρabove.Thedifferencejustliesinthevalues:wenowdon’texpecta50/50divisionbe-tweenwhiteandblack,butrathersomethinglike99.9999/0.00001(omittingmany9sand0shereforbrevity).Butthebasicideaisthesame:putadistributionoverinitialconditionsandshowthattheoutcomeprobabilitiesentailedfitwellwiththepatternsinactualevents.Andindeedtheydo,notonlythe(essentially)exception-lesspatternofSecondLawbehaviourformacroscopicfluids,butalsonon-trivial\n370RomanFriggandCarlHoeferprobabilitiesforsmallercollectionsofparticles.SowhatwehavelearnedfromthecoinalsosolvestheproblemofinterpretingprobabilitiesinSM!TheycanbeelegantlyaccommodatedinaHumeantheoryofobjectivechance.REFERENCESDavidAlbert,TimeandChance.Cambridge/MAandLondon:HarvardUniversityPress2000.PersiDiaconis,“APlaceforPhilosophy?TheRiseofModelinginStatisticalSci-ence”,in:QuarterlyofAppliedMathematics56,4,1998,pp.797-805.PersiDiaconis,SusanHolmes,andRichardMontgomery,“DynamicalBiasintheCoinToss”,in:SIAMReview49,2,2007,pp.211-235.JohnEarman,APrimeronDeterminsim.Dordrecht:Reidel1986.AdamElga,“InfinitesimalChancesandtheLawsofNature”,in:AustralasianJournalofPhilosophy82,2004,pp.67-76.RomanFrigg,“AFieldGuidetoRecentWorkontheFoundationsofStatisticalMechanics”,in:DeanRickles(Ed.),TheAshgateCompaniontoContempo-raryPhilosophyofPhysics.London:Ashgate2008a,pp.99-196.—“ChanceinBoltzmannianStatisticalMechanics,”PhilosophyofScience(Sup-plement)75,2008b,pp.670–681.SheldonGoldstein,“Boltzmann’sApproachtoStatisticalMechanics”,in:JeanBricmont,DetlefDürr,MariaCarlaGalavotti,GianCarloGhirardi,FedericoPetruccioneandNinoZanghì(Eds.),ChanceinPhysics:FoundationsandPerspectives.BerlinandNewYork:Springer2001,pp.39-54.CarlHoefer,“TheThirdWayonObjctiveProbability:ASceptic’sGuidetoObjec-tiveChance”,in:Mind116,463,2007,pp.549-596.—ChanceintheWorld:BookManuscript2008.ColinHowsonandPeterUrbach,ScientificReasoning:TheBayesianApproach.2nded.Chicago:OpenCourt1993.JosephB.Keller,“TheProbabilityofHeads”,in:AmericanMathematicalMonth-ly93,3,1986,pp.191-197.DavidLewis,“ASubjectivist’sGuidetoObjectiveChance”,in:RichardC.Jeffrey(Ed.),StudiesinInductiveLogicandProbability.Vol.2.Berkeley:UniversityofCaliforniaPress1980,reprintedinLewis1986,83-132,withpostscriptsadded.—“NewWorkforaTheoryofUniversals”,in:AustralasianJournalofPhiloso-phy61,4,1983,pp.343-377.—“HumeanSupervenienceDebugged”,in:Mind103,1994,pp.473-490.BarryLoewer,“DeterminismandChance”,in:StudiesinHistoryandPhilosophyofModernPhysics32,2001,pp.609-629.—“DavidLewis’HumeanTheoryofObjectiveChance”,in:PhilosophyofSci-ence71,2004,pp.1115-1125.\nDeterminismandChancefromaHumeanPerspective371TimMaudlin,“WhatCouldBeObjectiveAboutProbabilities?”in:StudiesinHis-toryandPhilosophyofModernPhysics38,2007,pp.275-291.MichaelStrevens,“ObjectiveProbabilityasaGuidetotheWorld”,in:Philosophi-calStudies95,1999,pp.243-275.JosUffink,“CompendiumoftheFoundationsofClassicalStatisticalPhysics”,in:JeremyButterfieldandJohnEarman(Eds.),PhilosophyofPhysics.Amster-dam:NorthHolland2006,pp.923-1047.StevenWeinberg,DreamsofaFinalTheory:TheScientist’sSearchfortheUlti-mateLawsofNature.NewYork:Vintage1994.RomanFriggDepartmentofPhilosophy,LogicandScientificMethodLondonSchoolofEconomicsandPoliticalScienceHoughtonStreetLondonWC2A2AEEnglandr.p.frigg@lse.ac.ukCarlHoeferDepartamentdeFilosofiaUniversitatAutònomadeBarcelonaEdificiB-708193BellaterraSpaincarl.hoefer@uab.es\n\nLÁSZLÓE.SZABÓWHATREMAINSOFPROBABILITY?IWhenIsay“probability”IdonotmeanaKolmogorovian‘probabilitymeasure’orsimilarmathematicalnotion.Ofcourse,adefinition–bothcoordinative1andlogical–playsaconstitutiveroleintheveryconcepttobedefined.Nevertheless,thereseemsnoreasontoincludeintothedefinitionofprobabilitythatit“satisfiestheKolmogorovianaxiomsofprobabilitytheory”.For,onceweknow–suppos-edlyfromtherestpartofthedefinition–what“probability”isinourworld,itbecomesacontingentfactoftheworldwhetheritsatisfiestheKolmogorovianaxiomsornot;whichcanbeknownbyaposteriorimeans.Inotherwords,theaimofthesocalled“interpretations”ofprobabilityisnottofindaninterpretationoftheKolmogorovianaxioms,buttogiveasoundmeaningtoscientificstatementscontainingtheterm“probability”.Forexample,considerthefollowingassertionsofquantummechanicsorsta-tisticalmechanics:p(a)=trPˆaWˆ(1)(∑iNi)!p{Ni}i=1,2,...=(2)∏iNi!Andcomparethemwithothersimilarscientificassertions,liketheCoulomblaw,r−rqE(r)=q

(3)
r−r
3qorjustasimplestatementaboutthelengthofarod:l=4m(4)Incase(3)and(4)itisclearwhattheformulasassert.Whenweassertthatther−rqstaticelectricfieldstrengthofapointchargeisequaltoq

r−r

3,wehaveaprevi-qouslydefinedphysicalquantity,electricfieldstrength,and(3)expressesacontin-gentfactaboutthisquantity.By“definition”Imeanempirical/operational/verificationistdefinition.Thisisnottheplacetoargueforverificationismoroperationalism.Ionlymentionthatmy1H.Reichenbach,TheTheoryofRelativityandAPrioriKnowledge.BerkeleyandLosAngeles:UniversityofCaliforniaPress1965.F.Stadler(ed.),ThePresentSituationinthePhilosophyofScience,ThePhilosophyofScienceinaEuropeanPerspective,DOI10.1007/978-90-481-9115-4_26,©SpringerScience+BusinessMediaB.V.2010\n374LászlóE.Szabóapproachisbasedonaveryweakoperationalist/verificationistpremise:scientifictermsassignedtoquantitiesliketheonesappearingin(1)–(4)musthaveempiricaldefinitions(exceptiftheequationsinquestionweredefinitions).Inotherwords,thosesentencesofascientifictheorywhicharesupposedtodescribeobjectivefactsoftheworldmustbeexpressibleinobservational/operationalterms.Withoutthisconditionascientifictheorycouldnotbeempiricallyconfirmableordiscon-firmable.Ibelieve,thisviewiswidelyacceptedamongphysicists;although,thepreciseoperationaldefinitionofaphysicalquantitycanbeanon-trivialissue,eveninthecaseofbasicspatiotemporalquantities.2However,initself,thispremiseisnotyetequivalenttooperationalismorverificationismingeneralphilosophicalsense.Itdoesnotgenerallyimplythatastatementisnecessarilymeaninglessifitisneitheranalyticnorempiricallyverifiable.Now,contraryto(3)and(4),itisfarfromobviouswhatformulas(1)and(2)actuallyassert.Whatisthedefinitionofthequantitiesonthelefthandsideoftheseformulas?Whatistheprobabilityofanevent?Thisisthebasicquestionofthephilosophyofprobability.Strangelyenough,inspiteofthefactthattheterm“probability”isusedintheeverydayscientificdiscourse,thereisaconsensusinthephilosophicalliterature3,4thatwehavenosatisfactoryanswertothisques-tion.IIThevariousapproachescanbedividedintotwomajorgroups.Accordingtotheobjectivistschool,theprobabilityofaneventissomethingwhichcharacterizesafeatureoftheexternalworld;roughlyspeaking,itisapropertyoftheeventandthecircumstances.Accordingtothesubjectivistapproach,onthecontrary,theprobabilityofaneventissomethingwhichcharacterizesafeatureoftheinternalworld;itisnotapropertyoftheeventandthecircumstances,butapropertyofaparticularintentionalstateofmindabouttheeventandthecircumstances;a“degree”ofbelief.Objectivists’probabilityisoftencalled“chance”(ch);subjec-tivists’probabilityiscalled“subjectiveprobability”or“credence”(cr).Thus,itmustbeemphasizedthatchanceandcredencearenotdifferentinterpretationsofthesamething,buttheyaretwodifferentthings,belongingtodifferenttypesof2L.E.Szabó,“EmpiricalFoundationofSpaceandTime”,inM.Suárez,M.DoratoandM.Rédei(eds.),EPSA07:LaunchoftheEuropeanPhilosophyofScienceAssociation.Dordrecht:Springer2009.3J.EarmanandW.Salmon,“TheConfirmationofScientificHypotheses”,in:M.H.Salmon,etal.(eds.),IntroductiontoPhilosophyofScience.EnglewoodCliffs,NewJersey:PrenticeHall1992.4A.Hájek,“InterpretationsofProbability”,inE.N.Zalta(ed.),TheStanfordEncy-clopediaofPhilosophy(Summer2003Edition),http://plato.stanford.edu/archives/sum2003/entries/probability-interpret.\nWhatRemainsofProbability?375phenomena.Iftheyexistinourworld,theydosoindependently;iftheyareatallconnected,theirconnectionmustbeacontingentfactoftheworld.Nevertheless,westilldon‘tknowexactlywhatchanceisandwhatcredenceis.IIINodoubt,thereissuchathingasaperson‘scredenceorbelief;and,nodoubt,itismeaningfultotalkaboutthedegreeofabelief,asabeliefcanbestrongerorweaker;andperhapsonecancharacterizeitwithanumberbetween0and1,justlikeanumericscalefrom0to10canbeaskedtocommunicatetheintensityofapatient’spain.Andonecaneasilyimaginesomerulesgoverningthecomplexmentalprocessesthatdeterminethisnumber;adynamicsbywhichthisnumberchangesintimeundervariousconditions.Allthesethingsbelongtothescopeofordinaryempiricalscienceslikepsy-chology,cognitivescience,humanethology,orsociology.Strangelyenough,how-ever,inthesubjectivistliterature,wecannotfindanyreferencetotheresultsoftheseempiricalsciences.Itworthmentioningthattheassertionsinquestionare,atthesametime,quiteambitious.Consideronlytwoexamples.Lewis’PrincipalPrincipleassertsthataperson‘scredenceisstrictlydeter-minedbysomeothermentalstates,namely:cr(A|‘ch(A)=x’&K)=x(5)where‘ch(A)=x’standsfortheperson’sknowledgethatch(A)=xandKstandsforsomefurther“admissible”(mis)information.Ifthisistrue,itisquiteastrictcausal/dynamicallawofmentalprocesses.Nottomentionthatthoseconditionsthatmakeproposition‘ch(A)=x’knowledge(thatis,true)andpropositionKad-missible,areconditionsintheexternalworld.So,initsstrongerunderstanding,theprincipleisastatementabouttherelationshipbetweentheexternalworldandone’smentalstates.MysecondexampleistheBayesianlawofconfirmation.Itdescribeshowthedegreeofaperson’sbeliefinthetruthofAchangesduetogettinginformationaboutanewevidenceE:cr(E|A)t1cr(A)=cr(A|E)=cr(A)(6)t2t1t1cr(E)t1wherechistheperson’spreviouscredencefunctionbasedonsomeearlierbodyt1ofbelieves,chisthenewcredencefunction,basedonthepreviousbelievesplust2E.This,too,isaverystrongclaimaboutthedynamicsofmentalprocesses.\n376LászlóE.SzabóIVHowisitthatthesubjectivistinterpretationofprobabilitycanclaimsoprecisequantitativelawsabouthumanmindwithoutanyreferencetotheresultsofem-piricalsciences?Therearetwostandardexplanationsofthisignorance,andbothraisefurtherproblems.Thefirstpossibilityisthatthewholesubjectivisttheoryisregardedasakindof“armchairmetaphysics”.Typically,inLewis’“Subjectivist’sguide”thePrinci-palPrincipleisbasedonsome“evidences”drawnfromhisownintuitiveanswerstohisQuestionnaire.Hewrites:Ihavegivenundefendedanswerstomyfourquestions.Ihopeyoufoundthemobviouslyright,sothatyouwillbewillingtotakethemasevidenceforwhatfollows.Ifnot,dopleasereconsider.Ifso,splendid–nowreadon.5Thatis,wemakeaprioriassertionsabouttherealworld,onthebasisofoureve-rydaypre-scientificandpre-philosophicalintuitions,withoutanyreferencetotheepistemicmeansbywhichtheassertedfactsoftheworldcanbeaccessed.However,ifastatementisobviouslytrueinourworld,thenthestatementmusthaveanobviousmeaning;theremustbeanobviouswayinwhichthestatementcanbeverified,whetheritistrueindeed,ornot.Inotherwords,weneedsomeempirical/coordinativedefinitionsofthebasicterms“credence”and“chance”.Thesecondpossibilityisthatthesubjectivisttheoryisregardednotasatheoryaboutrealpersons‘believes,butaboutthe“credencesofanabstractagent”.Termslike“chance”,“credence”or“agent”areonlymathematicaltermswithoutanyreferencetotherealworld;thewholetheoryisaformal/mathematicalconstruc-tion,justlikeKolmogorov’saxiomatictheoryofprobability,grouptheory,orge-ometry.Inthismathematicalconstruction,onecandefinea“gamble”inwhichtheagentismaking“bets”on“outcomeevents”;onecandefinethenotionof“rationalagent”asanagentwhosebetsandcredencesareinacertainrelation.Onemaymakeassumptionsaboutthe“behaviorofthegamblingdevice”;forexampleonecanassumethatthe“chancesoftheoutcomeevents”satisfytheaxiomsofKolmogorov’sprobabilitytheory;andfromthesepremises,forexample,onecanprovetheDutchbooktheorem.Thisis,ofcourse,possible,butIbelievethisisnotthefinalaimofsubjectivistinterpretationofprobability.Attheendofthedaywewouldliketoapplytheseab-stractconstructionsinthemetaphysical,epistemological,andscientificdiscourseabouttherealworld;thereforeweneedtoknowhowtoapplyourtheoreticaltermslike“credence”and“chance”.Wewouldneedtounderstandthemeaningofthesetermseveniftherulesofsubjectivisttheoryofprobabilityshouldnotbeseenasafactualdescriptionofactualhumanreasoning,butratherasa“normativestandard5D.Lewis,“ASubjectivist’sGuidetoObjectiveChance”,in:PhilosophicalPapersVolumeII.NewYork:OxfordUniversityPress1987,p.86.\nWhatRemainsofProbability?377ofrationality”;sinceweneedtounderstandtherulesinordertofollowthem.Thus,ineithercase,whatwearemissingistheempirical/operational/verifi-cationistdefinitionsofthebasicconcepts,firstofallof“credence”;thatistosay,asubjectivist’sguidetoobjectivechancewould,firstofall,requireaguidetothesubjectivist’scredence.VTherefore,itcouldonlyleadtocircularitiesifthedefinitionofchancech(A)in-cludedanyessentialreferenceto‘ch(A)=x’inthePrincipalPrinciple.Thereasonisthattheallegedrelation(5)holdsonlyifpropositionKis“admissible”.“Admis-sibility”is,however,aconceptthedefinitionofwhichrequirespriordefinitionsofcredenceandchance,independently:ApropositionPisadmissiblewithrespecttoanoutcome-specifyingpropositionEforchanceset-upS(Esaysthateventeoccurs)iffPcontainsonlythesortofinformationwhoseimpactonreasonablecredenceaboutE,ifany,comesentirelybywayofimpactoncredenceaboutthechancesofthoseoutcomes.[myitalics]6VILetusreturntotheoriginalproblemofwhat“probability”meansintheproba-bilisticassertionsofthesciences.First,itisworthpointingoutsomeconceptualconfusionwhichneedstobesortedout.Itisobviousthattheconceptofprobabil-ityinscience,especiallyinphysics,isobjectiveprobability.Whenthebehaviorofaphysicalsystemisdescribedbymeansofaprobabilisticmodel,probabilitiesaresupposedtodescribesomeobjectivefeaturesoftheexternalworld;nomatteriftheunderlyingphysicalprocessesaredeterministicorindeterministic.Classicalstatisticalmechanicsisatypicalexample.Webelievethatthedetailedprocessisgovernedbythedeterministiclawsofclassicalmechanics;but,becauseoflackingtheknowledgeofthedetails,wegivealessdetailedprobabilisticmodel.Theprob-abilitiesinsuchamodelaresometimescalled“epistemic”probabilities.Thisishoweveramisleadingterminologyofthephysicists,whichdiffersfromthetermi-nologyofthephilosophers.For,thereasonwhywegiveaprobabilisticdescriptionofthesystem,insteadofacompletelydetaileddeterministicone,isindeedrelatedwithalackofknowledge,yettheprobabilitiesinthemodelareobjectiveprob-abilitiesandhavenothingtodowith“knowledge”,“lackofinformation”,etc.Allstatementsoftheprobabilisticdescriptionofthesystemremainvalidevenifwegetcompleteinformationaboutthedetails;becausenoneoftheobjectivefeatures6R.FriggandC.Hoefer,“DeterminismandChancefromaHumeanPerspective”,inthisvolume.\n378LászlóE.Szabóofthesystemdescribedbytheprobabilitiesinquestionchangesbyknowingmoreinformationaboutthesystem.Letmegiveaneverydayexample.Youarewaitingforthenexttraininasubwaystation.Ifyouknewtheexacttimetable,youcouldmakepredictionslike“Thenexttrainwillarrivein3minutes.”Ifyoudon‘tknowthetimetablebutonlyknowthatthetrainscomeinevery5minutes,youcanmakelessambitiousclaims.Forexample,youcansaythat“Iwillwaitlessthan5minutes.”;oryoucanpredictthefollowingresultofalong-runexperiment:“ProvidingthatthemomentsatwhichIentertothestationareuniformlydistributedintime,thelong-runaverageofmywaitingtimeis2.5minutes.”Now,thevalidityoftheseclaimsdoesnotchangeifyougetknowthetimetable.VIIWhatweobservehereisnothingbutakindofHumeansupervenience.Objec-tiveprobabilitiessuperveneonthecollectionoftheparticularfactsoftheactualhistoryoftheworld,thatis,alloccurrentfactsinallregionsofspacetime;ontheHumeanMosaic.Andthisistrue,nomatteriftheworldisdeterministicorinde-terministic;eitherinthesensethatthedifferenttimeslicesoftheactualhistoryarenotfunctionallyrelated;eitherinthesensethatthereexistotherpossiblehistoriesoftheworldbesidestheactualone;eitherinthesenseofamoresophisticatedbranchingstructureofpossiblespacetime-histories.Ontheonehandthisistriviallytrue;ontheotherhand,onehastoputitinamorepreciseform:Thetruthorfalsityofallmeaningfulstatementsaboutobjec-tiveprobabilitiessuperveneontheHumeanMosaic,where“meaningful”ismeantinaverificationistsense;thatis,astatementismeaningfulifitisexpressibleintermsoftheHumeanMosaic.Note,however,thatthisisalsotrueforsubjectiveprobabilities:ThetruthorfalsityofallmeaningfulstatementsaboutsubjectiveprobabilitiessuperveneontheHumeanMosaic.Thequestionis,ofcourse:whatarethemeaningfulstate-mentsaboutsubjectiveprobabilities?Anothernote:althoughthetruthorfalsityofallmeaningfulstatementsaboutbothobjectiveandsubjectiveprobabilitiesaredeterminedbytheactualcontentofHumeanMosaic,theirtruthorfalsitycanbeknownonlybyaposteriorimeans.VIIINevertheless,whatischanceandwhatiscredence?Westilldonothaveatenabledefinitionofprobability,neitherobjectivenorsubjective.Andhowisitpossiblethatphysicsandotherempiricalsciencesdonotnoticeproblemsarisingfrom\nWhatRemainsofProbability?379theseunansweredfundamentalquestions?Astotheconceptofobjectiveprob-ability,inoneofmyearlierpapers7IproposedapossibleresolutionwhichIcall“No-probabilityInterpretationofProbability”.Thekeyideaofmyproposalisthis:thereisnosuchpropertyofaneventasits“probability”.Thatiswhythestandardinterpretationsfailtogiveasounddefi-nitionofprobability;andthatiswhyempiricalscienceslikephysicscanmanagewithoutsuchadefinition.Wheneverweusetheterm“probability”inscientificdiscourse,itsmeaningvariesfromcontexttocontext:itmeansdifferentdimen-sionless[0,1]-valuedphysicalquantities,ormoreprecisely,differentdimension-lessnormalizedmeasurescomposedbydifferentphysicalquantitiesinthevariousspecificsituations.Moreover,thesecontext-dependentmeaningsreducethecon-ceptof“probability”toordinaryphysicalquantitiesofempiricalmeanings,likerelativefrequencyonafinitesample,ratioofphase-spacevolumes,orthequanti-tiesontherighthandsideofformulas(1)–(2).ConsidermyexampleinpointVI.Onecancalculatetheaveragewaitingtimeonlyfromthefactthatthetrainscomeinevery5minutesandthatthemomentsoftimewhenIentertothestationareuniformlydistributed.Thesefactsareordinary,empiricallyverifiable,physicalfacts.Thecalculationrequiresonlykinematics,withoutevenmentioning“probabilities”.If,however,someonewouldliketoen-forceaprobabilisticlanguage,thecalculationoftheaveragewaitingtimecouldbepresentedinthefollowingway:Idon’tknowwhenIenteredtothestationrelativetothearrivalofthetrain.Sinceallmomentsoftimeofmyenteringareofequalprobability,Icalculatewithauniformprobabilitydistribution…Andtheresultwillbethesame.But,thereappearedaterm–probability–whichhasnodefinition.Thestatementscontainingthisterm,like“allmomentsoftimeofmyenteringtothestationareofequalprobability”,aremeaningless;itisimpossibletoverifywhethertheyaretrueornot.Astheexampleshows,however,theconceptof“probability”iscompletelyneedless.TheresearchwaspartlysupportedbytheOTKAFoundation,No.K68043.DepartmentofLogicInstituteofPhilosophyEötvösUniversityH-1518Budapest,Pf.32Hungaryleszabo@phil.elte.hu7L.E.Szabó,“Objectiveprobability-likethingswithandwithoutobjectiveindetermin-ism”,StudiesinHistoryandPhilosophyofModernPhysics38,2007,pp.626-634.\n\nHOLGERLYREHUMEANPERSPECTIVESONSTRUCTURALREALISMStructuralRealism(SR)isamoderatevariantofscientificrealismandcanroughlybecapturedbytheideathatweshouldbecommittedtothestructuralratherthanobject-likecontentofourbestcurrentscientifictheories.AquickviewonthelistofsomeofthemainproponentsshowsthatSRisbasicallyaEuropeanphilosophyofsciencemovement(andjustsuitsourESFProgramme):JohnWorrall,IoannisVotsis,StevenFrench,AngeloCei,JamesLadyman,SimonSaunders,MichaelEs-feld,VincentLam,KatherineBrading,MauroDorato,DeanRickles,FredMuller,and–exceptionsprovetherule–AnjanChakravarttyandJohnStachel.Thelistisofcoursenotexhaustive,moreover,thedebatehasabroadperiphery.AnotableexampleofthisisBasvanFraassen’sstructuralempiricism.Thepaperisakindofopinionatedreviewpaper.InwhatfollowsIwillpassthroughthemostprevailingtopicsinrecentdebatesoverSR.Mydiscussionwillbeorganised,perhapsabitunorthodoxly,inshortsections,hereandthenIwilloutlinemyownviews.1THENOTIONOFSTRUCTUREThenotionofstructureisnotoriouslyvague,andthisisalreadyoneofthemanyproblemsofSR.Thenotionisofcoursenotvagueasfarastheabstractmathemati-calconceptofstructureisconcerned.Compare,forinstance,Shapiro(2000):Defineasystemtobeacollectionofobjectswithcertainrelationsamongthem.[...]Defineapatternorstructuretobetheabstractformofthesystem,highlightingtheinterrelation-shipsamongtheobjects,andignoringanyfeaturesofthemthatdonotaffecthowtheyrelatetootherobjectsinthesystem.Themathematicaldefinitionsaysthatthereareentities,therelata,thatcomeequippedwithastructure,butthattherelataaredeterminedbystructuralorrela-tionalpropertiesonly.Hence,agoodworkingdefinitionforSRisthatstructuresaresetsofobjects,domains,withsetsofrelationsimposedonthem.Theproblemisthatdespitethemathematicaldefinitionthereexistsnopracti-cal,straightforwardmethodtoextractthestructuralcontentfromagivenscientifictheory.Theproblemisobviousasfarasnon-formalizedtheoriesinthehigherspe-cialsciencesareconcerned,butitprevailsevenregardingfundamentalphysicaltheories.InthispaperIdonotdelveintothisproblem,butIwillmostlytaketheF.Stadler(ed.),ThePresentSituationinthePhilosophyofScience,ThePhilosophyofScienceinaEuropeanPerspective,DOI10.1007/978-90-481-9115-4_27,©SpringerScience+BusinessMediaB.V.2010\n382HolgerLyresymmetrystructureastheprimary,genuinecandidatetocharacterizethestructuralcontentofmodernphysicaltheories.2TWOROUTESTOSTRUCTURALREALISMSRhasalongstandingtraditioninthe20thcenturyandevenearlier.Thereiscon-sensusthatthemoderndebatewasinitiatedbyJohnWorrall(1989).Thediscus-sionofthelasttwodecadeshasactuallytakentworoutestoSR,theWorrall-typeandFrench-Ladyman-typeroute,asIprefertocallthem.Worrall,Votsis(2003)andothersgaveargumentsinfavourofSRfromthephilosophyofscience–forinstancebyarguingthatSR’scommitmenttostructureandnotobject-likecontentcanbeusedasanantidoteagainstprominentanti-realisticargumentslikethepes-simisticmeta-inductionortheoryunderdetermination.French-Ladyman-typeau-thors,ontheotherhand,trytopresentargumentsfromthesciencesdirectly,morepreciselyfromthewaymodernscience,notablyphysics,informsusabouttheontologyoftheworld.Meanwhile,allmajorfieldsofmodernphysicshavebeenconsideredtostrengthenargumentsinfavourofastructuralontology:QuantumMechanics(FrenchandLadyman2003a,Esfeld2004),QuantumFieldTheory(Cao2003,Saunders2003),GeneralRelativity(Dorato2000,EsfeldandLam2008,Stachel2002),GaugeTheories(Lyre2004a,b),QuantumGravity(Ricklesetal.2006)orphysicsingeneral(Muller1998,Redhead2001).Notethatthedis-tinctionbetweenthetworoutesisnotthesameastheESR/OSRdistinction(seebelow).Cao(2003),forinstance,proposesFrench-Ladyman-typeESR.3ANTEREMVERSUSINRESTRUCTURALISMDebatesonstructuralisminmathematicsshowasimilaritytostructuralisminsci-ence,butmustultimatelybeseparatedfromthem.Shapiro(2000)isforinstanceknowntoupholdananteremstructuralistpositioninthephilosophyofmathemat-ics,i.e.aPlatonistconceptionoftheexistenceofstructurespriortoandindepend-entoftheirexemplificationinthephysicalworld.FrenchandLadyman(2003b)madeitsufficientlyclearthatSRshouldnotbeconfusedwithPlatonismbutisexplicitlyintendedasarealismaboutstructuresnotasabstractentitiesbutasinrestructuresinthephysicalworld.4EPISTEMIC,ONTICANDSEMANTICSRAsiswell-known,JamesLadyman(1998)firstcoinedthedistinctionbetweenEpistemicandOnticSR.WhileESRproponentsbelieveinthestructuralcontent\nHumeanPerspectivesonStructuralRealism383oftheoriesasanepistemicconstraintand,hence,upholdtheviewthatobjectsmayexist,butthatourepistemicaccessisrestrictedtostructuresonly,OSRpro-ponents,accordingtoLadyman,takestructuretobeprimitiveandontologicallysubsistent.Ithink,however,thedistinctionshouldbeabitmorerefined.Inlinewiththeusualthreefolddistinctionbetweenepistemic,onticandsemanticformsofscientificrealism,wemayaccordinglydistinguishbetweenthefollowingthreeoptions:•EpistemicSR:scienceconveystrueknowledgeaboutstructures,•SemanticSR:thecontentsandtermsofscientifictheoriesrefertostruc-tures,•OnticSR:structuresexistindependently(fromourepistemicandlinguis-ticcapacities).5ELIMINATIVEANDNON-ELIMINATIVESRWhat’sgenerallyunfortunatewiththeabovedistinctionsisthefactthateverythingstilldependsonourproperunderstandingoftheterm“structure”.Giventhemath-ematicaldefinitionofstructureassetsofobjectsorrelatawithsetsofrelationsimposedonthem,thereare,onthefaceofit,threepossibilities:•EpistemicSR:therearerelationsandrelata,butthatwehaveepistemicaccesstorelationsonly,•Non-eliminativeOSR:therearerelationsandrelata,butthatthereisnoth-ingmoretotherelatathantherelationsinwhichtheystand,•EliminativeOSR:thereareonlyrelationsandnorelata.Notethatunderthisclassificationthewidelydebatedquestionwhethertheslogan“structureisallthereis”leadstotheproblematicpositionof“relationswithoutrelata”doesnotdependonLadyman’sESR/OSRdistinction,butratheronthedis-tinctionbetweennon-eliminativeversuseliminativeversionsofSR.ItisperfectlypossibletoupholdSRasametaphysicalpositionabouttheworldwithoutbeingvulnerabletothe“relationswithoutrelata”-problem.Well-knownproponentsofeliminativeOSRare,oratleasthaveinitiallybeen,StevenFrenchandJamesLadyman(FrenchandLadyman2003a,French2006,LadymanandRoss2007),aproponentofnon-eliminative(ormoderate)OSRisMichaelEsfeld(2004).6STRUCTURALLYDERIVEDINTRINSICPROPERTIESIdoactuallybelievethattheabovethreefolddistinctionisstillnotexhaustive.Generalconsiderationsaboutsymmetrystructuresenforceustoassumetheexist-enceofnotonlyrelationalbut(inacertainsense)intrinsicpropertiesoftherelata.Technicallyspeaking,asymmetryofadomainDisasetofone-to-onemappings\n384HolgerLyreofDontoitself(a.k.a.symmetrytransformations),suchthatthestructureofDispreserved.Thesymmetrytransformationsformagroupandexemplifyequiva-lencerelations(i.e.apartitioningofDintoequivalenceclasses).Naturallyandnecessarily,wealwaysgetcertaininvariantsunderagivensymmetry.Inaphysi-calcontext,suchinvariantsprovidepropertiessharedbyallmembersofD.ThesepropertiesareintrinsicpropertiesinthesensethattheybelongtoanymemberofDirrespectivelyoftheexistenceofotherobject-likeentities.Ontheotherhand,theinvariantpropertiesdonotsufficetoindividuatethemembers,sinceallmemberssharethesameinvariantpropertiesinagivendomain.Structureinvariantsdonotleadtoindividualsbuttoobjectclassesonly.Thishighlightstheimportanceoftheinvariants:weusethemtoindividuatedomains,notindividuals.Nowacrucialpoint:insofarastheyarestructuralinvariants,theintrinsicproperties`depend’(inasensestilltobedetermined)onthestructure,weshouldaccordinglyandproperlyconsiderthemas“structurallyderivedintrinsicproper-ties”.Nevertheless,theyareintrinsicratherthanrelational,sincetheysubsistir-respectivelyoftheexistenceofotherobject-likeentities.7INTERMEDIATESRWearethusleftwithanevenmoremoderatenon-eliminativeversionofSRwhichIshalldub“IntermediateSR”(cf.Lyre2009).Itistheviewthattherearerelataandstructurallyderivedproperties,butthatthereisnothingmoretotherelatathanthestructurallyderivedproperties,wherethestructurallyderivedpropertiescompriserelationalpropertiesandinvariantsofstructureasstructurallyderivedintrinsicproperties.NotefurtherthatthisisstillaviableSRpositionanddoesnotcollapsetooldfashionedentityrealism,sinceneitherarewecommittedtoessentialpropertiesnorarewecommittedtoindividuals(seebelow).Structur-allyderivedpropertiesdonotindividuateobjectsbutobjectclassesordomainsofstructureonly.8ANILLUSTRATION:THELONEELECTRONThefollowingGedankenexperimentprovidesanillustrationoftheparticularna-tureofstructurallyderivedintrinsicproperties:Supposeapossibleworldwithoneelectrononly(andwithrelationalspacetime).Doestheloneelectronpossessanelementarycharge?Undertheclassicviewthatintrinsicpropertiesareprop-ertiesanobjecthasofitselfandindependentlyoftheexistenceofotherobjects,theloneelectronhascertainlyacharge.Itseems,however,thatforproponentsofbotheliminativeandmoderateOSR,whoacceptrelationalpropertiesonly,aloneelectroncannothaveacharge,sincetherearenootherobjectsleftinvirtueof\nHumeanPerspectivesonStructuralRealism385whichtheelectron’schargemightbeconsideredasrelational.FromthepointofviewofIntermediateSRasanothernon-eliminativeversionincludingstructurallyderivedintrinsicpropertiesthereisnoproblemtoapplychargestoloneobjects.ForeveninthetrivialcaseofonlyonememberinD,theobjectwillpossessthesaidsymmetry-invariantproperties.Theobjecthastheinvariancepropertiesinvirtueofthestructure,thestructurecomesequippedwithsuchproperties.Inmorephysicalterms:evenaloneelectronisaproperinstantiationoftheinreU(1)gaugestructure.Butcouldn’twejustsaythatthechargeisrelationaltothestructure?Theproblemisthatinthiscaseonecannotexcludethepossibilitythatthestructureasarelatumoftheexemplificationrelationcanexistforitself.Hence,oneopensthedoortounexemplifiedstructures–aclearrenunciationofinrestructuralismandadangerousflirtwithPlatonism.TheideahereisthatthestructurewearetalkingaboutintheloneelectronscenarioistheU(1)structuredisplayedintheMaxwellequationsandinstantiatedbythatveryelectron.Fromanoperationalistpointofview,ofcourse,suchstructurecanonlybeobservedfromthebehaviourofmorethanjustonetestcharge.Butstructuralismisnotpersecommittedtooperational-ism–bothviewsshouldlogicallybekeptdisentangled.9AFURTHERARGUMENT:GAUGEINVARIANTSTheimportanceofstructuralinvariants–structurallyderivedintrinsicproperties–canmostclearlybeseenfromthemostimportantcaseofsymmetrystructureinmodernphysics,thecaseofgaugetheoreticstructures.Onecrucialfeatureofgaugesymmetriesisthattheypossessnorealinstantiations.Notethatwemustcarefullydistinguishbetweensymmetrieswithrealinstantiationsasopposedtosymmetrieswithoutrealinstantiations.Examplesoftheformerareforinstancethepossiblespace-timetransformationsofaphysicalobject.Examplesofthelat-terarescaletransformations,coordinatetransformations,and,inparticular,gaugetransformations.Therefore,agaugetheoreticcharacterizationofaphysicaltheoryisafortioriallandonlyacharacterizationbymeansofthesymmetryinvariants,sinceonlythegaugesymmetryinvariantsallowforarealisticinterpretation.Gaugetransformationspossessnorealinstantiations(cf.Lyre2004a,b).Inthecaseofgaugetheories,theSRcommitmenttostructurecanonlybeacommitmenttothestructureinvariants.TheseinvariantsaregivenbytheeigenvaluesoftheCasimiroperatorsofthevariousgaugegroups,whichintheirphysicalinterpretationareconsideredtobemass,spinandthevariouscharges.Infact,mass,spin,andchargeprovideparadigmaticcasesofintrinsicpropertiesofelementaryparticles.Theyaretheattributesbywhichweclassifythefundamentalparticlezoo.Theyare,infact,themostfundamentalstructurallyderivedintrinsicproperties.\n386HolgerLyre10IDENTITY,HAECCEITISMANDMETAPHYSICALUNDERDETERMINATIONAnother“structuralattack”ontraditionalentityrealismhastodowithissuesaboutidentityandindividualityinmodernphysics,notablyquantummechanics.TheempiricalindistinguishabilityofquantumobjectshasoriginallybeenregardedasafailureofLeibniz’principleoftheidentityofindiscernibles(PII).French(1989,1998),however,arguesthatweareratherleftwithakindof„metaphysicalunder-determination“:eitherquantumobjectsviolatePIIandarenoindividuals,ortheyareindividualssincePIIappliesbyreferencetosomekindofprimitivethisness,bareparticularityorhaecceity(orhoweverwemaycallit).Thedeeperlessonisthatscienceleaveseventhemostprofoundmetaphysicalquestionaboutindividu-alityunderdeterminedandso,followingFrench,webettergiveupentityrealismaltogetherandstickwithastructuralontology.Obviously,thislineofreasoningpavesthewaytoeliminativeOSR.11WEAKDISCERNIBLESSaunders(2006)hasarguedthatalthoughfermionsarenotabsolutelydiscernible(intermsofintrinsicmonadicproperties),theyareneverthelessweaklydiscern-ible.Indeed,thisobservationcanbeseenassupportingstructuralnon-elimina-tivism(andtogiveuphaecceitism).TomakethisclaimplausibleconsiderfirstBlack’scaseoftwoequalspheresinrelationalspacewithadistanced.DosuchspheresviolatePII?CallobjectsthatviolatePIIabsolutelydiscernible,butobjectswhichallowforirreflexiverelationsweaklydiscernible(Quine1976).RecallthatarelationRisreflexivewhenforallxinthedomainR(x,x)holds.Inthecaseof¬R(x,x),Riscalledirreflexive.ForinstanceeachBlacksphereisadistancedapartfromtheotherbutnotfromitself.Sothedistancerelationisirreflexive.Thesameholdsinthecaseoffermionicparticlesinanentangledstatefortherelationofhavingoppositespin.FredMuller(inprint)hasrecentlyevenextendedthisresulttoparticlesirrespectiveoftheirspinbyconsideringtheHeisenberg“commuta-tionrelation”ofhavingcomplementarypositionandmomentum.Wemaysaythatquantumobjectsareinfactgenerallyweaklydiscernibleduetothepossibilityofcanonicallyconjugatevariablesbasedonthenon-commutativealgebrastructureofquantumtheory.Thecaseofweakdiscerniblesaccountsfortheexistenceofrelatathatareweaklyindividuatedbyirreflexiverelations.Itrunscountertorelata-eliminativ-ism,butdoesatthesametimenotendorsefullentityrealismofabsoluteindividu-als.Indeed,irreflexiverelationsarestructurallyderivedrelationsinthesensethattheyreflecttheallowedquantumstatesofthenon-commutativealgebrastructure.Asinthecaseofstructurallyderivedintrinsicproperties,theyareontologicallyonaparwiththestructurewithoutpresupposingtheindependentexistenceofeither\nHumeanPerspectivesonStructuralRealism387thestructureortheproperties(ortherelata).Rather,theyareinreexemplifica-tionsofthestructure.12THEPROBLEMOFUNINTENDEDDOMAINSLet’scometosomemoreintricateproblemsofSR.Reconsidertheideaofstructureinvariantsasderivedintrinsicproperties.Thecrucialquestioniswhetherandhowwewilleverknowaboutsuchpropertiesasintrinsicnaturesofobjects.Takenliter-ally,theideatoindividuatetheoriesbymeansoftheirpurestructuralcontent(inthesenseofpuremathematicalstructure)isfartooweak.Thereasonliesinwhatonemightcallthe“problemofunintendeddomains”.Thereareinfactlotsofcaseswheredistinctphysicaltheoriesshowbasicallythesamemathematicalstructure,hencewemustqualifythestructure’sdomain.Herearesomephysicsexamplesofsuch“structuralequivalents”:(i)classicalelectrodynamicsandhydrodynamicsarebasedonmoreorlessthesamemathematicalapparatusaboutunspecified`cur-rents’includingcontinuityequations,theoremsofGaussandStokesetc.;(ii)thegaugetheoriesofstrongandweakisospinarebothbasedonSU(2);(iii)thegroupU(1)figuresinquantumphysicsbothasthegroupoftemporalautomorphismsandasthegaugegroupofQED.Surelywe’vesaidthatthedomainsareindividuatedbythestructureinvariantsasderivedintrinsicproperties,butsofarwedidnotspelloutwhetherandhowtheyprovideanindependentwaytomakecontactwithsuchinvariantproperties.Let’sleavethisopenforthemomentanddiscusssomefurtherrelatedissuesfirst.13STRUCTURALUNDERDETERMINATIONWemayexercabatetheproblemofunintendeddomainstotheproblemofstructur-alunderdetermination.AccordingtotheWorrall-typeroutetoSR(asmentionedinsection2),SRcanbeseenasanantidoteagainsttheoryunderdetermination(TUD).TheideaisthatwhileTUDunderminesentitycontent,SRseemstoavoidthisbynotcommittingustothetheory’sentitycontentbuttostructuralcontentonly.However,asI’vearguedelsewhere(Lyre,inprint),thereis,onthefaceofit,nowaytomakesurethatthestructuralcontentoftheoriesisnotunderdeter-minedeither.Onthecontrary,thereseemtoexistcasesinourbestfundamentalscience,notablyintheoriesofgravity,wherewearedirectlyconfrontedwithcasesofstructuralTUD.Thismeansthatweareconfrontedwithstructurallyinequiva-lentbutempiricallyequivalenttheories.Insuchcasesthestructureofatheoryisunderdeterminedbyempiricalevidence.\n388HolgerLyre14THERAMSEY-CARNAP-LEWIS-ACCOUNTOFTHEORETICALTERMSWemayreiterateandgeneralizethetwoproblemsmentionedabove.InordertodosowemustreconsidertheRamsey-Carnap-Lewis-accountoftheoreticalterms(cf.Lewis1970).Asavariantofscientificrealism,SRisarealismabouttheunob-servable.Taketheclassicdistinctionbetweenobservationalandtheoreticaltermsoandt.TheRamseysentenceofatheoryTcanbeunderstoodasamachineryforiiexpressingthestructuralcontentofT.ItisobtainedbyreplacingthetheoreticaltermsofTwithboundvariables:T(t,…t,o,…,o)→∃x,…∃xT(x,…x;1n1m1n1no,…o).Undersuchanaccountthetheoreticaltermsarenoteliminatedbutare1mexpressedintermsofthestructuralrelationsbetweenthevariablesxinT.TheiRamseysentenceleavesuswithapurestructuraldescriptionofthetheoreticalknowledgeabouttheworld.TheearlyRussellandCarnaptookthisasamotiva-tiontoupholdanextremeepistemicstructuralism.15MULTIPLEREALIZABILITY,QUIDDITISMANDRAMSEYANHUMILITYMultiplerealizabilityisinfactanimmediateconsequenceoftheRamsey-Car-nap-Lewis-accountoftheoreticalterms.Ourknowledgeaboutthereferentsofthetheoreticaltermsisjustknowledgeabouttheoccupantsorplaceholdersofdescrip-tivecausalroles.Thequiddisticnatureoftheplaceholdersisindetermined,theyarethusmultiplyrealizable.ApossibleresponseistoadvocateRamseyanHumil-ityaboutquiddities.Recallthathaecceitismistheviewthatapermutationofindividuals(orto-kens)makesadifference.Itamountstoassumeprimitivethisness.We’vealreadyseenthatSR,clearlyinitsnon-eliminativistbranch,dismisseshaecceitism(section10).Quidditism,ontheotherhand,istheviewthatapermutationofproperties(ortypes)makesadifference.Itamountstoassumeprimitivesuchness.Sostructural-istsusuallyrejecthaecceitism,butshouldtheyrejectquidditismaswell?TheproblemnotonlyforSRbutinfactforanyvariantofscientificrealismwhichcommitsitselftotheRamsey-Carnap-Lewis-accountoftheoreticaltermsisthatquidditismamountstomakingadifferencewithoutadifference.Neverthe-less,DavidLewis(2009)subscribestoquidditism,butatthesametimeadvocatesRamseyanHumility,atermhehasborrowedfromRaeLangton’s(1998)KantianHumility.KantianHumility,inturn,shouldcaptureKant’sviewthatthingsasweknowthem,phenomena,consistentirelyofrelationsandthatwehavenoknowl-edgeoftheintrinsicpropertiesofthingsinthemselves.SofollowingLangtonKant’sattitudeisnoidealism,butratheranepistemichumility.Accordingly,Ram-seyanHumilityistheviewthat“noamountofknowledgeaboutwhatrolesareoccupiedwilltelluswhichpropertiesoccupywhichroles”(Lewis2009,p.204).\nHumeanPerspectivesonStructuralRealism389Asecondanswertotheproblemofquidditismisthatwemightneverthe-lessbeincontactwithquiddisticnatures,i.e.toadvocateamoredirectrealismthansuggestedbytheindirectcausalandnomologicalknowledgeprovidedbytheRamseysentence(seealsoSchaffer2005).Andtheremightevenbeathirdstanceasregardsquidditism,namelysimplytodismissitasanexaggeratedmetaphysicswhileatthesametimeclaimingthistobeaviablerealistanswerdespiteitsappar-entempiricistflavor.Iwillmakenofurtherattemptheretodecidewhichwaytogo(inpartalsosince,again,theproblemisnotspecialtoSRbutaffectsrealismintoto).16THENEWMANPROBLEMAsiswell-known,MaxNewman(1928)raisedaseriousobjectionagainstRus-sell’s(1927)earlyversionofSR(seeDemopoulosandFriedman(1989)foramodernresumption).Theideaisthatifabstractstructureisallwecanknowfromourtheoriesabouttheunobservableworld,thenonlycardinalityquestionsareopentoempiricaldiscovery.AsNewman(1928,140)putit:…givenany‘aggregate’A,asystemofrelationsbetweenitsmemberscanbefoundhavinganyassignedstructurecompatiblewiththecardinalnumberofA.Andfurther:…thedoctrinethatonlystructureisknowninvolvesthedoctrinethatnothingcanbeknownthatisnotlogicallydeduciblefromthemerefactofexistence,except(“theoretically”)thenumberofconstitutingobjects.Sostructuralismisnear-vacuous,ineffectitcollapsestoempiricism.Allwecanknowisjustcardinality.ThepointoftheNewmanproblemisnotonlythatrelationsdonotsufficetopickouttheintrinsicnatureoftheobjectsinthedomain,butthatalsothenatureoftherelationsthemselvesremainsindetermined!AccordingtotheearlyRus-sellonlyabstractmathematicalstructureisknown.Butwithoutfurtherempiricalqualification,anysuchabstractstructurecanbeimposedonagivenset(modulocardinalityconstraints).Inasense,theNewmanproblemistheinverseofmultiplerealizability.Whereasinthelattercasewehavemultipleinstantiations(collectionsofentities)thatfitthestructuraldescription,Newman’sproblemamountstosayingthatagivencollectionofentitiescanbeendowedwithanyarbitrarystructure,aslongasthecollectionhastherightcardinality.AsvanFraassen(2008)haspointedout,\n390HolgerLyreNewman’sproblemshowsaninterestingsimilaritytoPutnam’smodel-theoreticproblem,butweshallnotdelveintothedetailsofdisentanglingthemhere.17FOURPROBLEMSREVISITEDWe’vediscoveredfourproblemsinconnectionwithSR:unintendeddomains(sec-tion12),structuralunderdetermination(13),multirealization(15)andNewman’sproblem(16).Theyactuallycomeinpairs.Whilethefirstpairhastodowiththepracticalandvaguenotionofstructureinphysicaltheories(forinstancethesym-metrystructuregivenbythesymmetrygroupsinphysics),thelatterpairhastodowiththepreciselogico-mathematicalstructureofatheory(cf.section1).ThedifferencebetweenthetwopairsisthatthesymmetrystructureofTismostcer-tainlynotexhaustive,sincethecompletestructureofTisalmostcertainlymoreextensive.Bywayofcontrast,thelogico-mathematicalstructureoftheRamseysentenceisexhaustive,insofarastheRamseysentenceofatheoryprovidesacompletedescriptionofT.Despitethisdistinction,problems12and15aswellas13and16aremoreorlessvariationsofthesametheme–with12and13asspecialpracticalcasesofthemoregeneralizedabstractcases15and16.Itisnotatallimplausibletoassumethatallfourproblems(oratleastthree,structuralTUDisperhapsmorespecial)aresostronglyconnectedthattheyseekforacommonanswer.Andbasically,therearetworoutesfromhere,aHumeanandananti-Hu-meanroute,asIshalloutlineinthefinalsections.18MODALSTRUCTURESSeveralSRproponentsinrecentdebateshavearguedinfavourofmodalorcausalstructures(Chakravartty2004,2007;Esfeld(inprint);Ladyman&Ross2007).Thismeansthatstructuresareconceivedasdispositionalratherthancategorical.Thebasicidea,notablyinChakravartty(2004),istoendowstructuresthemselveswithcausalpowers.Esfeld(inprint)considersthisaninevitablestepinordertocopewiththeproblemofquidditism(section15)byassumingthatthemeta-physicalcausalitybehindtheobservableregularitieshasitsrootnotinepistemi-callyhiddenquidditiesbutinthecausalnatureofthestructuresthemselves.WhileLewisbelievesthatbecauseoftheRamseyaccountoftheoreticaltermswehavenoepistemicaccesstoquiddities(buttocausalroles,i.eobservableregularitiesonly),thecausalstructureassumptiondismissesquidditiesaltogether(andis,therefore,ratheradissolutiontotheproblem).OthersevenseecausalstructuresasapossiblewaytoovercomeNewman’sproblem.Russell’searlystructuralismwasaboutabstractstructures,notaboutconcreteinrestructures.Itwas,inotherwords,aboutsecondandnotfirstorder\nHumeanPerspectivesonStructuralRealism391relations.ToovercomeNewman’sproblemthestructuralistmustconsiderfirstorderrelationswithcausalpowersasinstantiationsofabstractstructures.Thecausalstructuresstrategyisperhapsawayoutoftheconundrumofprob-lems12and16inparticular.But,asusual,onehastopayaprice.Thestrategyincludesadouble-step:first,toinvokefirstorderrelationsand,second,toinvokecausalpowers.Andthesecondstepportraysadecisivenon-Humeanelement,theallegedlymodalordispositionalnatureofstructures.Therearewell-knowndiffi-cultiesconnectedwithmodalordispositionalontologies,notablyunclearidentityconditions,whichIshallnotexplorehere.Rather,myprojectwillbetooutlinetheperspectivesofSRfromastrictHumeanpointofview.19AHUMEANRESPONSETONEWMANConfrontedwithNewman’sobjection,Russellimmediatelyrealizedthathemustrefinehisposition.Inordertojustifyaparticular,intendedstructure,wemustsomehowbedirectlyacquaintedwithcertainstructuralrelations.Russellthusdemanded“spatiotemporalcopunctuality”betweensense-dataandphysicalob-jectsasabasicrelation.IcannotnotdiscussherewhetherRussell’sproposalofspatiotemporalcopunctualityisalreadythecorrectanswertothequestforbasicrelations,butIwanttoemphasizethathisideaofknowledgeaboutstructuresbyacquaintanceratherthanmeredescriptionis,inprinciple,aviablesolutiontothenotoriousproblems12and16,perhapseven15.Itis,infact,asolutionwhichisalsoopenformodernproponentsofSRpavingthewayforaHumeanconceptionofSR.Theessentialclueisthatwearenotboundtorelationalpropertiesonly.Foraswehavealreadyseen,SRmusttakestructurallyderivedintrinsicpropertiesintoaccount(sections6-9).Wemightthereforeenvisagedirectobservationalacquaint-ancewithstructurallyderivedintrinsicproperties.Whetherandwhichplacehold-ersofastructuraldescriptionexist,i.e.whetherandhowastructureisinstantiated,isanempiricalquestion.Andwhetheritis,forinstance,electromagneticorhydro-dynamiccurrenthastobedistinguishedonthelevelofobservationalphenomenaandcannotbeknownfromthepuretheoreticalandstructuralcontentalone(giventhestructuralequivalenceofthemathematicalaccounts).Inourexperimentalob-servationsweare“incontact”withthecategorical,structurallyderivedintrinsicnatureofthecurrents.Sotheideaisbasicallythis:Insofarastheyare(structurallyderived)intrinsicweneednotinvokeacquaintancewith(causal)structuresandinsofarastheyarecategoricalweneednotinvokecausalpropertiesatall(betheystructuralornot).ThispavesthewayforaHumeanresponse.Andfinally,insofarasweassume“direct”acquaintancewiththemwerediscoverRussell’soptiontocircumventNewman’sproblem.SowegetahybridofaHumeanandRusselianresponseto\n392HolgerLyreNewman.Note,moreover,thatweaklydiscerniblerelationsarealsoperfectlycat-egorical:theydonotinvolveanyquantumprobabilities.20HUMEANPERSPECTIVESONSTRUCTURALREALISMAproperHumeanperspectiveonSRistodemandcategoricalstructuresandtodismissmysteriousmodalities(cf.Sparber2009foranaccountsimilarinspirit).Humeanmetaphysics,asusuallyconstrued,isbasedonatleastthreeconditions:1.amicro-physicalistsuperveniencebaseoffundamentalintrinsicandcat-egoricalproperties,2.regularity(i.e.non-necessitarian)viewaboutlaws,and3.reductionismaboutlaws.InanattempttocombineHumeanmetaphysicswithSR,atleastoneofthethreeconditionsmustbechanged.Letusconsiderthemsubsequentlyinthefollowingsections.21SUBVENIENTHOLISTICSTRUCTURESThefirstconditionisbestcharacterizedinLewis’famousconceptionofHumeansupervenience,hisviewof“theworld[as]avastmosaicoflocalmattersofpar-ticularfact”with“nodifferencewithoutdifferenceinthearrangementofqualities.Allelsesupervenesonthat”(Lewis1986,ix-x).Meanwhilehowever,itiswidelyacceptedthatHumeansupervenienceisboundtofail.Itfailsaccordingtomodernscience–accordingtothecasesofquantumentanglementandgaugetheoreticholism(cf.Healey2007,chap.4.5;Lyre2004b;Maudlin2007,chap.2).Lewisevenacknowledgesthethreatofquantumentanglement:maybethelessonofBell’stheoremisexactlythattherearephysicalentitieswhichareunlocalized,andwhichmightthereforemakeadifferencebetweenworlds…thatmatchperfectlyintheirarrangementsoflocalqualities.Maybeso.I’mreadytobelieveit.ButIamnotreadytotakelessonsinontologyfromquantumphysicsasitnowis.FirstImustseehowitlookswhenitispurifiedofinstrumentalistfrivolity…and–mostofall–…ofsupernaturaltalesaboutthepoweroftheobservantmindtomakethingsjump.If,afterallthat,itstillteachesnonlocality,Ishallsubmitwillinglytothebestofauthority.Butwhetherthequantummeasurementproblemhastodowithfrivolityornot–sincethecaseofnonlocalitycanbemadeingaugetheoriesaswell(afactLewiswasobviouslynotawareof),itistimetorealizethatHumeansuperveniencemustdefinitelybegivenup.ForproponentsofHumeanSRthisisnobadnews,sinceitisexactlythisconditionabouttheLewiseanHumeanbasewhichmustberejected.Insteadofa\nHumeanPerspectivesonStructuralRealism393mosaicofintrinsic,categoricalproperties,HumeanSRconsiderswholestructuresinthesuperveniencebase.Thisisadismissalofnaïvemicro-physicalism,notaboutthecategoricalnatureofsuchstructures.Structuresareholisticandglobalratherthanlocalentities,physicallyexemplifiedandmanifestlycategorical.Thereisnoneedtoassumecausalstructures,aswealreadysawinthediscussionofNewman’sproblemandaswe’llseenowinthediscussionofthesecondHumeancondition.22STRUCTURALNON-NECESSITARIANISMABOUTLAWSHumeanSRisactuallyinaccordancewiththesecondconditionfromsection20.Structuresarenotarbitrary,butregularglobalsetsofrelations.Hence–andthisisaquiteimportantpoint–regularity,thecrucialingredientoflaws,isalreadyentailedbyinvokingstructures.Structuresarelaw-like.Take,forinstance,theMinkowskispacetimestructureofspecialrelativity.Itisaglobalgeodesicstruc-tureexemplifiedbythetrajectoriesoffreefallingbodies–aseeminglyregularbehaviour.Moreover,thebehaviourofafreeparticletofollowgeodesicsisnodispositionoftheparticle,norisitadispositionofthegeodesicstructure,itisanexemplificationofthemanifest,categoricalinrestructureofspacetime.Thesameholdsforotherfundamentalstructures,forinstance,theU(1)-structureoftheworldbeingexemplifiedbychargeconservation.Remarkably,suchastructuralistregularityviewaboutlawsofferstoavoidwell-knownproblemsoftheorthodoxregularityview.Oneproblemisthatnotallregularitiesarelaw-like.Indeed,notallregularitiesarelaws,onlystructuresare.UnderHumeanSR,structuresshouldbeconceivedas“world-built-inpatterns”orglobalregularities.Theholisticaspectofstructuresiscrucialhere:theparticlefol-lowingageodesicisnotasubsequenceofdisparateeventswhich,withoutfurtherexplanation,showaregularbehaviour.Itisanexemplificationofaglobalregular-ityitself–thegeodesicstructure.Thereis,again,noreasontoassumethatthereare“empty”laws.Inrestruc-turalismconsidersonlyexemplifiedstructures.Suchstructuresaren’tnecessarilyexemplifiedatany(world)time,buttheyareatleastgloballyexemplifiedonthewholespacetimeextension.ThisisperhapsthemoststraightforwardwaytothinkofexemplificationinHumeanstructuralworlds:consideraworldinwhichonlyoneparticleataninfinitesimallysmalltimeperiodhastravelledalikewisein-finitesimalspatialpath.Thisparticleisaproperinstantiationofthefullspacetimestructureofthatpossibleworld.HumeanSRhasfurthermoretheresourcestoexplaintheobviousuniversalityofstructureinvariantswithoutrecoursetoessentialism.Becauseoftheholisticorglobalnatureofstructures,thestructuralinvariantsbehaveasuniversallyvalid.Butsuchuniversalitydoesnotcomeequippedwithmysteriousnecessity.Itsu-\n394HolgerLyrepervenesontheHumeanbaseofstructures.Itisamereregularityitselfthatsomeparticularstructureisinstantiated.Nonecessitiesareinvolvedhere.ItfollowsfromthesamelogicthatHumeanSRcanaccountforexception-lesslaws.Anyinstantiationofastructurewillshowthesameregularbehaviourencodedinthestructure.Exceptionsmustnotbeexpected,unless,however,thewholestructureitselfchanges.Thislatterpossibilitycanofcoursenotberuledout.Afterall,structuresprovidetheHumeanbase,whetheraparticularstructuresubsistsornotisamatterofpureregularityitself.23NON-REALISMABOUTLAWSTheideathatstructuresprovidetheHumeanbaseguaranteesthatHumeanSRisinaccordancewiththethirdconditionfromsection20.Lawsarereducedtostruc-tures,lawssuperveneonthestructuralHumeanbase.SomemightthinkthatSRiscommittedtoarealismaboutlawsbecauseofthefollowingargument:accordingtoSRstructuresarerealandlawsarestructures,solawsmustobviouslyberealtoo.But,aswe’veseen,HumeanSRjustconsidersstructuresasglobalregularitiesanditemsoftheHumeanbase.Soagain:whetheraparticularstructuresubsistsornotisamatterofpureregularityitself.Lawsaren’tliterallystructures,andstruc-turesareonlylaw-likeinthesensethatlawscanbereducedtoglobalregularities(whichwecallstructures).24TRANSFERTHEORYOFCAUSATIONHowshouldHumeanstructuralrealistsfinallyconstruecausality?Theymightinfactwelcomeatransfertheoryofcausation(cf.Dowe2000).Theroughideaisthatacausalprocessisthetransmissionofconservedquantitieswithcausalinter-actionsasintersectionsofsuchprocessesprovidinganexchangeoftheconservedquantities.Accordingtofundamentalphysics,conservedquantitiesareidentifiedwithstructuralinvariants.ThisisduetoNoether’stheoremwhichstatesthattoeverycontinuoussymmetrygeneratedbylocalactionstherecorrespondsacon-servedquantity.Suchconservedquantitiesand,inturn,thecausalprocessesandinteractionsareexemplificationsofthefundamentalstructures.Structurescomeequippedwithconservedquantities.Somemightcomplainthatthetransfertheoryisnon-Humean.Butthisisatbestaproblemforamicro-physicalistHumeanbase(accordingtocondition1).IfweconsiderwholestructuresintheHumeanbasethencausalprocessesandtransferofconservedquantitiessuperveneonthatbase.AndthisisalltheHumeanneeds.\nHumeanPerspectivesonStructuralRealism39525PRELIMINARYCONCLUSIONSections18to24presentargumentsagainstcausalstructuresandprovideperspec-tivesforaHumeanSR.Thereisnoneedtoendowstructureswithcausalpowers.What’sstillmissinginthepictureis,perhaps,howdynamicscomesintotheworld.We’vebasicallyoutlinedastaticpicture.Andthisispresumablythebiggestne-glectsofar.Non-HumeanSRwithcausalstructures,however,doesn’tsolvethisproblemeither.Metaphysicalcausationandphysicaldynamicsaredistincttopics,proponentsofcausalstructureshavenobettergripondynamicsthanopponents.Here’scertainlymuchtobedoneinthefuture.Admittedly,thispaperwaslargelyprogrammatic.WecouldmerelytouchuponsomefewmotivesandperspectivesonHumeanSR.Buttheperspectivesarequitepromising,perhapspromisingenoughtopursuetheminmoreelaboratedexaminations.Acknowledgements:Manythankstoananonymousrefereeforvaluablesuggestions.REFERENCESTianYuCao(2003).Structuralrealismandtheinterpretationofquantumfieldtheory.Synthese136:3–24.AngeloCeiandStevenFrench(2006).Lookingforstructureinallthewrongplac-es:Ramseysentences,multiplerealizability,andstructure.StudiesinHistoryandPhilosophyofScience37:633–655.AnjanChakravartty(2004).Structuralismasaformofscientificrealism.Interna-tionalStudiesinthePhilosophyofScience18(2&3):151-171.AnjanChakravartty(2007).AMetaphysicsforScientificRealism:KnowingtheUnobservable.NewYork:CambridgeUniversityPress.WilliamDemopoulosandMichaelFriedman(1985).Criticalnotice:BertrandRussell’sTheAnalysisofMatter:Itshistoricalcontextandcontemporaryin-terest.PhilosophyofScience52:621–639.MauroDorato(2000).Substantivalism,relationismandstructuralspacetimereal-ism.FoundationsofPhysics30(10):1605-28.PhilDowe(2000).PhysicalCausation.NewYork:CambridgeUniversityPressMichaelEsfeld(2004).Quantumentanglementandametaphysicsofrelations.StudiesinHistoryandPhilosophyofModernPhysics35(4):601-617.MichaelEsfeld(inprint).Themodalnatureofstructuresinonticstructuralreal-ism.InternationalStudiesinthePhilosophyofScience.MichaelEsfeldandVincentLam(2008).Moderatestructuralrealismaboutspace-time.Synthese160:27-46.\n396HolgerLyreBasvanFraassen(2008).ScientificRepresentation:ParadoxesofPerspective.NewYork:OxfordUniversityPress.StevenFrench(1989).Identityandindividualityinclassicalandquantumphysics.AustralasianJournalofPhilosophy67:432–446.StevenFrench(1998).Onthewitheringawayofphysicalobjects.InElenaCas-tellani(ed.):InterpretingBodies:ClassicalandQuantumObjectsinModernPhysics,pp.93–113.Princeton:PrincetonUniversityPress.StevenFrenchandJamesLadyman(2003a).Remodellingstructuralrealism:Quantumphysicsandthemetaphysicsofstructure.Synthese136:31–56.StevenFrenchandJamesLadyman(2003b).Betweenplatonismandphenomenal-ism:ReplytoCao.Synthese136:73–78.StevenFrench(2006).Structureasaweaponoftherealist.ProceedingsoftheAristotelianSociety106:167–85.RichardHealey(2007).GaugingWhat’sReal:TheConceptualFoundationsofContemporaryGaugeTheories.NewYork:OxfordUniversityPress.JamesLadymanandDonRoss(withDavidSpurrettandJohnCollier)(2007).EveryThingMustGo:MetaphysicsNaturalised.Oxford:OxfordUniversityPress.RaeLangton(1998).KantianHumility:OurIgnoranceofThingsinThemselves.Oxford:OxfordUniversityPress.DavidLewis(1970).Howtodefinetheoreticalterms.JournalofPhilosophy67:427–446.DavidLewis(1986).PhilosophicalPapersVol.II.Oxford:OxfordUniversityPress.DavidLewis(2009).RamseyanHumility.InDavidBraddon-MitchellandRobertNola(eds.):ConceptualAnalysisandPhilosophicalNaturalism.Cambridge,MA:MITPress.HolgerLyre(2004a).LokaleSymmetrienundWirklichkeit.Paderborn:Mentis.HolgerLyre(2004b).HolismandstructuralisminU(1)gaugetheory.StudiesinHistoryandPhilosophyofModernPhysics35(4):643-670.HolgerLyre(2009).Structuralrealismandabductive-transcendentalarguments.InMichelBitbol,PierreKerszbergandJeanPetitot(eds.):ConstitutingOb-jectivity.TranscendentalPerspectivesonModernPhysics.Berlin:Springer.HolgerLyre(inprint).Isstructuralunderdeterminationpossible?Synthese.TimMaudlin(2007):TheMetaphysicsWithinPhysics.NewYork:OxfordUni-versityPress.FredMuller(1998).StructuresforEveryone.Amsterdam:Gerits&Son.FredMuller(inprint).Witheringaway,weakly.Synthese.MaxNewman(1928).Mr.Russell’s“CausalTheoryofPerception”.Mind37(146):137-148.WillardV.O.Quine(1976).Gradesofdiscriminability.JournalofPhilosophy73(5):113-116.\nHumeanPerspectivesonStructuralRealism397MichaelRedhead(2001).Theintelligibilityoftheuniverse.In:AnthonyO’Hear(ed.):PhilosophyattheNewMillennium.Cambridge:CambridgeUniversityPress.DeanRickles,StevenFrenchandJuhaSaatsi(2006).TheStructuralFoundationsofQuantumGravity.Oxford:OxfordUniversityPress.BertrandRussell(1927).TheAnalysisofMatter.London:GeorgeAllen&Un-win.SimonSaunders(2003).Structuralrealismagain.Synthese136:127–133.SimonSaunders(2006).Arequantumparticlesobjects?Analysis66:52–63.JonathanSchaffer(2005).Quiddisticknowledge.PhilosophicalStudies123:1–32.StewartShapiro(2000).ThinkingaboutMathematics.NewYork:OxfordUniver-sityPress.GeorgSparber(2009).UnorthodoxHumeanism.Frankfurta.M.:Ontos.JohnStachel(2002).Therelationsbetweenthingsversusthethingsbetweenrela-tions:Thedeepermeaningoftheholeargument.In:DavidMalament(ed.):ReadingNaturalPhilosophy:EssaysintheHistoryandPhilosophyofScienceandMathematics.LaSalle,IL:OpenCourt.IoannisVotsis(2003).Isstructurenotenough?PhilosophyofScience70:879–890.JohnWorrall(1989).Structuralrealism:Thebestofbothworlds?Dialectica43:99–124.PhilosophyDepartmentUniversityofBielefeldGermanylyre@uni-bielefeld.de\n\nF.A.MULLERTHECHARACTERISATIONOFSTRUCTURE:DEFINITIONVERSUSAXIOMATISATIONABSTRACTCrucialtostructuralrealismistheCentralClaimthatentityBisorhasstructureS.Wearguethatneithertheset-theoreticalnorthecategory-theoreticalconcep-tionsofstructureclarifytheClaiminawaythatservestheneedsofstructuralrealism.Oneoftheseneedsistohaveaviableaccountofreference,whichal-mostanyvarietyofrealismneeds.Thereisalsoaviewofstructurethatcanadoptbothset-theoreticalandcategory-theoreticalconceptionsofstructure;thisistheviewthatadoptsB.C.vanFraassen’sextensionofNelsonGoodman’sconceptofrepresentation-asfromarttoscience.Yettheensuingfountainofperspectivesisamoveawayfromrealism,structuralrealismincluded.Wethensuggestthatanewtheoryofstructureisneeded,onethattakestheword‘structure’toexpressaprim-itivefundamentalconcept;theconceptofstructureshouldbeaxiomatisedratherthandefinedintermsofotherconcepts.WesketchhowsuchatheorycanclarifytheCentralClaiminamannerthatservesadescriptivistaccountofreference,andtherebystructuralrealism.1.PREAMBLEAfterhavingdiscardedanumberofcharacterisationsofscientificrealism,i.e.real-isminthephilosophyofscience,inhisclassicTheScientificImage,VanFraassenprovidedthefollowingminimalcharacterisation:Scienceaimstogiveus,initstheories,aliterallytruestoryofwhattheworldislike;andacceptanceofascientifictheoryinvolvesthebeliefthatitistrue.Thisisthecorrect1statementofscientificrealism.ThefirstandforemostdistinctioninthevarietyofscientificrealismcalledStructuralRealism(StrR2)istheonebetweenepistemicStrR(allthatsciencepro-videsisknowledgeofthestructureofthephysicalworld)andonticStrR(allthere1Fraassen[1980],p.8.2Weshallambiguouslyuseabbreviation‘StrR’alsofor‘astructuralrealist’.F.Stadler(ed.),ThePresentSituationinthePhilosophyofScience,ThePhilosophyofScienceinaEuropeanPerspective,DOI10.1007/978-90-481-9115-4_28,©SpringerScience+BusinessMediaB.V.2010\n400F.A.Mullerisinthephysicalworldisstructure).3Yetwhenwetakeheedofthefactthatknow-ledgeimpliestruth,andtaketruthtoimplyontologicaladequacy,thegapbetweenepistemicandonticStrRnarrowsquickly.4Whetherepistemicorontic,StrRneedstheconceptofstructure.ForStR,aliterallytruestoryofwhattheworldislikewillinvolvestructuresonly,andthereforeweneedtoknowwhatstructuresare.WhatmakesStrRstandapartfromothervarietiesofrealismisthatitissup-posedtobemorecautious,moremodest,initsrealistclaims,inordernottofallpreytothepessimisticmeta-inductionoverthehistoryofscience.Thepremiseofthisinductiveargumentisasequenceofpastscientifictheoriesthatnowhaveallbeenrejected.Theyareplausiblyfalse.Soifyouthinkthatourcurrentlyacceptedscientifictheoriesarenotfalsebuttrue,thenwhatareyou,stupid?ButStrRmustremainsufficientlysubstantivetoprovideabasisfortheno-miracleargument:theonlyexplanationforthefactthatinsequencesofsuccessivelyacceptedtheorieseachtheoryisempiricallyandtechnologicallyatleastas,andgenericallymore,successfulthanitspredecessoristhattheylatchontothestructureofthephysicalworldbetterandbetter.Wetakeacloserlookattheset-theoretical(Section2)andthecategory-theore-tical(Section3)conceptionsofstructureandwefindtheminadequatetoservetheneedsofStrR,specificallytheneedtohavealiteraldescriptionoftherefer-entstowhichtermsinscientifictheoriesrefer.ThenweexplorethepossibilityofretainingbothconceptionsofstructurebyadoptingVanFraassen’sconceptofrepresentation-as–asopposedtorepresentation-of–,whichcanbemarshalledtoevadetheobjectionsleveledagainstset-theoreticalandcategory-theoreticalcon-ceptionsofstructurewheninterpretedliterally.Thepricetopayseemshowevertoohighforrealism,becausethisadoptionintroducesaperspective-dependencythatstandsopposedtotheveryideaofrealism(Section4).Sincebythenallavail-ableoptionsseemexhausted,weargueforthecasethatStrRneedsanewtheoryofstructure,thattakestheconceptofstructureasfundamental,thatis,asaprimitiveconceptthatoughttobeaxiomatisedratherthandefined(Section5).SuchatheorywillservetheneedsofStrR,orsoweargue.Throughoutthispaperwetake‘structure’tomeanmathematicalstructure,becauseinsciencethesestructuresareusedinsciencetomodel,torepresent,todescribe,toexplain,tounderstand,etc.theworld.2.SET-THEORETICALCHARACTERISATIONAlthoughtherigorousset-theoreticalcharacterisationofstructureiswell-knownandwidespread,easytoexplainandeasytoillustrate,itsrigorousdefinitionintheformallanguageL2ofpureset-theory(ZFC,say)israthercommanding.Bourbakiwasthefirsttoprovideadefinitionofstructureofextremegenerality,withinthe3DuetoJ.Ladyman[1998].H.Lyre’s[2009]reviewinthisvolumeandLadyman’sencyclopediaarticle[2009]drawmoredistinctionswithinStrR.4AsIhavearguedelsewhere;seeMuller[2009],Section1.\nTheCharacterisationofStructure:DefinitionversusAxiomatisation401frameworkofhisownset-theory,inChapterIVofhisTheoryofSets(1949).5WecallthesetsinthedomainofdiscourseVofZFC–whichharbourspuresetsandnothingbutpuresets–answeringtothisdefinitionset-structures.Bourbakipro-poundedtheviewthatmathematicsisthestudyofstructureandsetout,bymeansofset-structures,tocreateLaw&Orderintheexuberantproliferationandprogres-sivesplinteringof20th-centurymathematics,whichhadturnedthedisciplineintoaTowerofBabel.PatrickSuppesfamouslycametopromulgatetheuseofset-structuresinthephilosophyofscience,notablytocharacterisescientifictheories.Suppes’Slogan:toaxiomatiseatheoryistodefineaset-theoreticalpredicate.6Informally,aset-structureisapolytupleofthefollowingform:•®basesets,subsetfamilies,relations,functions,operations,constants;(1)ormoreprecisely:•®B1;:::;Bb;F1;:::;Fs;R1;:::;Rr;F1;:::;Ff;O1;:::;Oo;C;(2)whereb2N+(positivenaturalnumber)andr;s;f;o2N(naturalnumbers).Set-structure(2)has:bbasesets;ssubsetfamilies,eachoneofanyofthebasesets;rrelationsbetweenthemembersofbasesetsorthoseofthesubsetfamilies;ffunctionseachofwhosedomainandco-domainisoneofthebasesetsorsubsetfamilies,orissomeCartesianproduct-setofthesesets;ooperationseachofwhichhasasadomaininaCartesianproduct-setofgenerallyoneofthebasesetsandthesamebasesetasitsco-domain;andasetC(ofConstants)whichcontainsmembersofthebasesetsorsubsetfamiliesthatplayaspecialrole(aszero,one,ˆtop,bottom,,singularity,etc.).SetCishoweveroftenomittedforthesakeofbrevity.Letusconsideranexamplefromphysics,thatweshalltakeastheleadingexampleinthispaper:aHeliumatom(He)inauniformmagneticfield(B:R3!0R3,hx;y;zi7!h0;0;Bi).Aquantum-mechanicalstructureusedtodescribethis0compositephysicalsystem(whichweshallhenceforthabbreviatebyHeB)isofthefollowingtype:•®S(HeB)´L2(R3);H(B);Ã;Pr;(3)0twherethefouroccupantsareasfollows.7ThebasesetistheHilbert-spaceofsquare-integrablecomplexfunctionsofthreerealvariables(‘complexwavefunc-5Bourbaki[1968].ForanattemptatanaccessibleexpositionofBourbaki’sdefinition,aswellasabriefdescriptionofBourbaki’sprogramme,seeMuller[1998],pp.106–115;foramoresmoothandaccessibledefinitionofstructure,seeDaCosta&Chuaqui[1988],whospeakof‘Suppes-predicates’,inhonorofPatrickSuppes[1960],[1967].6SeeSuppes[1960],[2002];seeDaCosta&French[2000]forareviewofdevelopmentsinthisareaoverthepast30years.Lyre[2009]alsoemploysset-structures.237Theoccupantsofapolytuplelike(3)arethefouritemsinthere:L(R)isthefirstoccupant,etc.Theyarenotthemembersoftheset(3),butiteratedmembers.SeeMuller[1998],p.24,forarigorousdefinitionofoccupant.\n402F.A.Mullertions’).8LinearfunctionH(B):D!L2(R3)istheHamiltonian,theopera-0torthatrepresentsthephysicalmagnitudeofenergy;itsdomainDliesdenseinL2(R3).(ThemagneticfieldBandotherrelevantphysicalmagnitudes,such0asthelinearmomentumoftheHe-atom,arepresentinH(B0)butsuppressednotation-wise.)FunctionÃ:R!L2(R3);t7!Ã(t)(4)isthesolutionoftheSchrodinger-equation;itiscontinuousiffnomeasurements¨areperformed.Finally,functionPrt:¢7!Prt(¢)istheBornprobabilitymea-sure,oneforeveryt2R,thatgivestheprobabilityoffindingavaluein¢½RfortheenergyoftheHe-atomwhenmeasuredandwhenthestateoftheHe-atomisÃ(t):Pr(¢)=hÃ(t)jPH(¢)jÃ(t)i;(5)twherePH(¢)istherelevantmemberofthespectralfamilyofprojectorsofH(B).0Mathematicallythereismoregoingon,whichwehavesuppressedinthedecep-tivelysimplenotation(3).Firstofall,Hilbert-spaceL2(R3)isitselfastructure:•®L2(R3);+;¢;h¢j¢i;k¢k;0;(6)where,first,+:L2(R3)£L2(R3)!L2(R3)(7)istheoperationofadditiononthecomplexwavefunctions,leadingtoanAbelianadditivegroup;secondly,¢:C£L2(R3)!L2(R3)(8)isthescalarmultiplicationofwavefunctions,whichinteractsdistributivelywithaddition,leadingtoacomplexvectorspace;thirdly,mappingh¢j¢i:L2(R3)£L2(R3)!C(9)istheinner-product;fourthly,k¢k:L2(R3)!R+(10)isthenorm,generatedbytheinner-product,leadingbothtometricalandtopologi-calstructure;andsixthly,0isthezero-function,theneutralelementoftheadditivegroup.Inturn,thereals(R)alsoformsomealgebraicstructure:•®R;<;+;£;f0;1g;(11)238MorerigorouslyonehastoidentifymembersofL(R)whichareequalalmostev-erywhere,thusgivingrisetoasetofLebesgue-equivalenceclassesofcomplexwave23functions,denotedasL(R).\nTheCharacterisationofStructure:DefinitionversusAxiomatisation403andthecomplexnumbers(C)too.Thenaturalnumbers(N)arealwaysneededandtheyalsoformaparticularstructure:•®N;S;0;(12)whereS:N!Nisthesuccessor-function(allarithmeticaloperationscanbedefinedinductivelyintermsofS).StructureS(HeB)(3)alsoharboursaKolmogorovianprobabilitystructure:•®B(R);[0;1];Prt;(13)wheretheprobabilityfunctionPrt:B(R)isanormedmeasureontheBorelsetsB(R),whichinturnisalsoastructure,aBoolean¾-lattice:•®B(R);µ;[;\;n;?:(14)Thusthewave-mechanicalstructureS(HeB)isinfullsplendour(permutingtheorderoftheoccupants):DN;S;0;R;<;+;£;f0;1g;C;+;£;f0;1g;EL2(R3);+;¢;h¢j¢i;k¢k;0;H(B);Ã;B(R);µ;[;\;n;?;[0;1];Pr0(15)Sinceoperationsareaparticularkindoffunctionsandfunctionsareapartic-ularkindofrelations,andrelationsbetweenmembersoftwoarbitrarysets,DandRsay,aresubsetsoftheirCartesianproduct-setD£R,andthusmembersofthepower-setofD£R,andsincetheCartesianproduct-setD£Risamemberofthe3-timesiteratedpower-setoftheunion-setD[R:D£R2}3(D[R);(16)oneseesthatstartingfromtheinfinitenumbersetsN,RandC,thestructureS(HeB)(3)livesatalevelinthecumulativehierarchyofsetsVthatisacon-siderablenumberofapplicationsofthepower-setoperationhigherthanwereN,RandClive.WecalltomindCantor’sPowerTheorem,accordingtowhichthepower-set}(D)isstrictlylargerincardinalitythansetD,toseethatstructureS(HeB)(15)harboursvarioussetsmuchlargerthanthecardinalityofthecontin-uum(R).Thestandardrouteforset-theoreticiansistotakethefinitevonNeumannordi-nalsasthenaturalnumbers(N´!);thenthereisauniquesetofnaturalnumbers.Thestructuralistroute(Bourbaki’s)istodefinea‘naturalnumberstructure’bymeansofaset-theoreticalstructure-predicate(aSuppes-predicate),asa‘Peanostructure’(12),orasa‘Dedekindstructure’,orasa‘Fregestructure’;inallthesecasesthereisnolongeraunique‘naturalnumberstructure’butanabsoluteinfin-ityofsuchstructures(asmanyastherearesetsinthedomainofdiscourseVof\n404F.A.MullerZFC).9Thesametworoutesareavailablefortheothernumberstructures(inte-gers,Z;rationals,Q;reals,R;complexnumbers,C):theycanbeconstructedinVbyset-theoreticalmeansfromN=!soastoendupwithuniquenumberstruc-tures(rationalsasorderedpairsofintegers,realsasBolzano-CauchysequencesofrationalsorasDedekind-cuts,complexnumbersasorderedpairsofreals);ortheycanbedefinedbystructure-predicates(seefootnote9).Whenonefollowsthefirst,constructive-likeroute,thenZ2}4(N);Q2}7(N);R2}8(N);C2}10(N):(17)ThenforthesetofwavefunctionsfromS(HeB)(3)wehave¡¢L(R3)2}3}3(}8N[}9N)[}9N;(18)andfortheHamiltonian:¡¢H(B)2}6}3(}8N[}9N)[}9N;(19)0andthewavefunction:³¡¢´Ã2}3}9N[}3}3(}8N[}9N)[}9N;(20)andtheprobabilitymeasure:¡¢Pr2}3}8N[}9N:(21)tFortheorderedquadrupleS(HeB)(3)wethenobtain:³¡¢´S(HeB)2}3L2(R3)[}3H[}3(Ã[Pr):(22)tButproperlyconstrued,astheordered28-tuple(15),structureS(HeB)isamem-berofafarmoreinvolvedset-structure.With(18),(19),(20),(21)and(22),onecanworkoutexactlyhowmanyiterationsofthepower-setweare,withS(HeB),beyondthefirstinfiniteordinallevel(!)inthecumulativehierarchy,whichweleaveasanexerciseforthewillingreaders.Presentlyitwillbecomeclearwhywehavebotheredtopointthisallout.Now,whatdoesStrRclaimwithregardtostructureS(HeB)(3)?WhenwefollowPatrickSuppes10inconsideringtheclassofstructureslikeS(HeB),andsimilarones(withotherphysicalmagnitudes,mixedstates,etc.),toconstitutethetheoryofquantummechanics(QM),thenittriviallyfollowsthatallthatQMtellsusaboutphysicalreality,actuallyevenallthatQMcantellusaboutphysicalreality,suchasaboutelementofphysicalrealityHeB,isthatthisphysicalsystemisorhasstructureS(HeB).ThusJohnWorrall[1989]isrightwhenhesaysthatallscience9SeeMuller[1998],pp.56–64,wherethisisallspelledout.10Suppes[1960],[1967],[2002].\nTheCharacterisationofStructure:DefinitionversusAxiomatisation405providesuswithisknowledgeofthestructure(s)oftheworld,ratherthanofthenature(s)oftheworld.EpistemicStrRseemsinevitable.Whenknowledgeimpliestruth,andtruthimpliesontologicaladequacy,thenknowingthatstructureS(HeB)isthestructureHeBimpliesthatS(HeB)trulyisthestructureofHeB.OnticStrRisjustaroundthecorner!Perhapsweshouldlimitourclaimstoso-calledontologicalsubstructuresofS(HeB),butonticStrRremainsjustaroundthecorner.11Wegetaroundthecornerwhenweassumeinadditionthatsciencetellsus,oreventuallywilltellus,everythingthereistotellaboutthephysicalworldingeneral,andaboutHe-atomsinuniformmagneticfieldsinparticular(scientificoptimism).12Nothingwillbeleftunsaid.SincethephysicalworldisbuiltfromatomsandaccordingtoonticStrRtheyarestructures,thephysicalworldiscom-posedofstructures.Adfundumstructuresdetermineeverythingthereisinthephysicalworld.TheconclusionseemstobethatSuppes’structuralistviewonscientifictheo-riesconjoinedwitharealistattitudeyieldsepistemicStrRandoptimisticallyalsoonticStrR.AsWorrall[2009]hasrecentlyputit:“StructuralRealismistheonlygameintown.”Endofstory?Notyet.ForwhatdoesitmeanexactlytosaythatHeBisorhasstructureS(HeB)?ThisisaninstanceoftheCentralClaimofStrR.BeingBisorhasstructureS(abeingis(23)anything,anyentity,thatexists),independentlyofus,humanbeings,ofouractivities,attitudesandcapacities,ofourveryexistence.ThisCentralClaimstandsinneedofclarification,aswillemergebelow.The‘is’obviouslycannotmeantheidentity-relation,becauseS(HeB)(22)isanabstractmathematicalentity,towitacomplicatedset-theoreticalconstructionoutoftheemptyset,livinginthecumulativehierarchyofallandonlypuresets,inthedomainofdiscourseVofZFC,whileHeBisaconcretephysicalentity,‘outthere’inthephysicalworld.CertainlyaHe-atominauniformmagneticfield(HeB)itisnotaset.Perhaps,then,‘is’meanspredication,asdoes‘has’.HeBhasastructure,averyspecificstructure,namelyS(HeB),justasatomatohasacolour,averyspecificcolour,namelyred.Letusseewherethisleadsus.Weexpresspropertiesinourlanguagebymeansofpredicates.Thepropertyred–ifthereare‘properties’–isexpressedbythepredicate‘red’,andtheascrip-tionofthepropertyredtoatomatoisexpressedbysayingthat‘Thistomatoisred’istrue,orthatthistomatofallsunderthepredicate‘red’.Theobviouscan-didateforthepredicatethatascribesthewave-mechanicalstructuretoHeBisthe11Theideaofconsideringontologicalsubstructuresofstructuresforrealistclaimswassuggestedmorethantenyearsago,andiscloselyrelatedtoM.L.G.Redhead’sideaof‘surplusstructure’.SeeinMuller[1998],p.356ff.,andRedhead[1975],p.88.12SeeMuller[2009],Section1.\n406F.A.Mullerset-theoreticalonethatdefinesstructureS(HeB)(3),callit»(¢).Thegeneralformofthispredicateis:13¡¢»S(HeB)iff9X1;9X2;9X3;9X4:•®S(HeB)=X1;X2;X3;X4^(24)X=L2(R3)^X=H(B)^120X3=Ã^X4=Prt:Thiswillnotdoeither,because»(¢)(24)isanopensentenceinthelanguageL2ofZFCandthusonlyappliestoinhabitantsofV.OurHeBdoesnotinhabitVandthereforecanneverfallunder»(¢):formallyspeaking,‘»(HeB)’isnonsense.Nowwhat?Thewaytogowithoutleavingset-theoryseemstoenrichVwithphysicalsystems.14ThismakesLatwo-sortedlanguage,withset-variablesandphysical-2system-variables,sayaandbforthenewsort.Physicalsystemscanbecollectedinsets,sothat‘a2X’etc.becomewell-formedatomicsentencesoftheenrichedlanguage,callitL¤.Expressions‘X2a’,‘X=a’,‘a2b’,‘X62a’etc.are2forbiddeninL¤becausethephysicalsystemsarenotsupposedtobeentitiesthat2canhavemembers,theyarenotsets.TheaxiomsofZFChavetobereformulatedintheenrichedlanguageL¤,andoneaxiomhastobeaddeddeclaringtheexistence2ofphysicalsystems,butthat’sall.ThusoneobtainsZFCU.NoadditionalaxiomsarepresentinZFCUtogovernthephysicalsystems.(ItispossibletoenrichZFCwithmereologicalaxiomsthatgovernthephysicalsystems,bytakingthesubsystem-relationasaprimitivedyadicpredicateinthelanguageadditionaltothemembership-predicate;theorythusobtainedisaconservativeextensionoverZFCandthereforeconsistentrelativetoZFC,andthereforetoZF.15Weshallnotdothishere;wehavedoneitalreadysomewhereelse(seepreviousfootnote).)OurHeBwillnowhopefullybecomeavalueofthefreshvariables,because,asweallknow,tobeistobethevalueofavariable.TheconclusionthatHeBisasetwillthenhavebeenavoided.Butstill,structure-predicate»(¢)(24)issuchthatonlyaparticularkindofpolytuple,henceaset,fallsunderit,namelyapolytupleoftheformS(HeB)(3).Formally,from(24)weseeimmediatelythat‘»(a)’isnonsensebecause‘a=h:;:;:;:i’isnonsense.Thereforewehavetoadjust»(24)ofL2tosomeotherpredicateofL¤,say'(¢),suchthat‘'(a)’makessenseandstructure2S(HeB)issomehowinthere–asitmust,becausethatiswhatQMprovides.Thiscanbeachievedintwosteps.13Noticethattheright-hand-sidesoftheidentity-statementsinthedefiniens(24)areas-sumedtobeantecedentlydefinedsingulartermsinthelanguageofZFC;thisisdoneforbrevity,morestandardistowrite‘X1isaHilbert-space’.14Thetechnicaltermforobjectsthatarenotsetsisprimordialelements,orUr-elements,fromtheGermanUrelemente.SeeFraenkel[1973],pp.23–25.15SeeMuller[1998],pp.189–252,fordetailsandproofs.\nTheCharacterisationofStructure:DefinitionversusAxiomatisation407Thefirststepistoletaoccupythestructurepolytuple:¡¢»¤a;S¤(HeB)iff9X;9X;9X;9X:1234•®S¤(HeB)=a;X;X;X;X^1234(25)X=L2(R3)^X=H(B)^120X3=Ã^X4=Prt:Thedyadicpredicate»¤expressesarelationbetweenstructureS¤(HeB)andphys-¤icalsystema.Since»relatesaphysicalsystem,whichwehopetoidentifywithconcretephysicalobjectHeB,toanabstractobject,structureS¤(HeB),whichisattheendofthedaystillaset,justlikeS(HeB),thisisnotquitewhatStrRneeds.¤Thesecondstepistoturn»(25)intoamonadicpredicateofabyexistentiallyquantifyingS¤(HeB)away:¡¢'(a)iff9S¤(HeB):»¤a;S¤(HeB):(26)Formally,weseemtobegoingintherightdirection.Forletuscomparethingsagaintoredtomatoes.Supposethereisatomatoontheplateinfrontofus.Thesentence‘Red(this-tomato)’istrueandtheexpression‘this-tomato’triviallyreferstothetomatoontheplateinfrontofus.Similarlywewanttosaythat‘'(a)’istrueandthat‘a’referstoaHe-atominauniformmagneticfield.But‘a’isavariableandvariablesdonotrefer.WhatwewanttosayinsteadisthataisaHe-atominauniformmagneticfieldiff'(a),because'(¢)(26)istheset-theoreticaltranslationofthecharacterisationofHeBthatQMprovides.16Theextensionof'(¢)thenincludesallandonlyactual(andperhapspossible)He-atomsinauniformmagneticfield.Thesymbol‘HeB’canthenbeofficiallyinauguratedasavariablerunningoverthisextension,aso-calledHelium-atom-in-a-uniform-magnetic-field-B0-variable.Iftheofficiallyinauguratedvariable‘HeB’assumesavaluefromthisextension,wecansaythatHe-atomsinauniformmagneticfieldexist,orthatthevariableplurallyreferstothosephysicalsystems;orifwecanlocatebylasercoolingtechniquesasingleHe-atominthelaboratory,andgiveitaname,wecansaythatthisnamesingularlyreferstotheatom,justasinthecaseoftheredtomatoontheplateinfrontofus.Thisstoryhastobegroundedinsomeaccountofreference.Forsuchunob-servablephysicalsystemsasHe-atomsinmagneticfields,theonlyviableaccountofreferenceisadescriptivistone.17Therelevantdescriptionhereisthedescrip-tionofourHeB.Science,bymeansof(ourset-theoreticallyreconstructed)QM,16He-atomsareusuallycharacterisedbytheirconstitutiveparts(anucleusconsistingoftwoprotonsandtwoneutrons,andtwoelectrons)andtheirmass,chargeandspin.Usuallythiscanbereadof‘readoff’theHamiltonianH(B0)andthereforeisincludedbutis,unliketheuniformmagneticfieldB0,notation-wisesuppressed.17Forwhytheonlyavailablealternative,theKripke-Putnamcausaltheoryofreference,failstoprovideageneralaccountofreferenceforscience,seeGauker[2006],pp.130–132.\n408F.A.Mullerdeliversthisdescription:'(a)(26).Letusnexttakeacloserlookatthisdescrip-tion.Description'(a)(26)literallysays:thereissomeparticularpolytuple,i.e.set-structureS¤(HeB),thathasaasitsfirstoccupant(25).OnecaneasilyprovethatifS(HeB)(3)existsinV,thenS¤(HeB)(25)existsinV¤,foreveryaindis-criminately:ZFCU`9S:»(S)¡!8a:'(a):(27)Sincetheantecedentcanalsobeproved,socantheconsequent:ZFCU`8a:'(a):(28)Recallthattheideawastoobtainadescription–basedonQM–suchthatthoseafallingunderthedescriptioncanbesaidtobeHeB.Butif'(¢)(26)isthatdescription,thenasaconsequenceeverysinglephysicalsystemqualifiesasaHeB(28).Whichisabsurd.18Thesituationisactuallyworsethanabsurd,becausethisallgeneralises.Foreveryset-structureAinV,onecaneasilyprovethereareasmanystructuresastherearephysicalsystemsinthatforeveryphysicalsystemb,thereisastructureAbthathasbasitsfirstoccupantandthatsharesallitsoccupantswithA.Thuseveryphysicalsystemiseverything.Thedescriptionsarethereforevoid.Notonlyistheputativedescription'(¢)(26)void,inspiteofappearancestothecontrary,buteveryotherdescription,basedonanyotherstructureA,ratherthanS(HeB)orS¤(HeB),willalsobevoid.NodescriptivistaccountofreferencecantakeoffinthecontextofZFCU.Nowwearedone.Ourprovisionalconclusionisthattheset-theoreticalroadtophysicalrealityforStrRseemsaroadtonowhere.Realismwithoutreference,then?Hmmm.Smellslikerealismwithoutreality.Beforerealistsgetthatdesper-ate,theyshouldexploreallotheroptions.OneoptiontoclarifytheCentralClaimofStrR(23)istoreplaceset-theorywithcategory-theory.3.CATEGORY-THEORETICALCHARACTERISATIONWhenitcomestodealwithstructures,inparticularinabstractbranchesofmath-ematics–abstractincomparisontonumbertheory,analysisandthegeometryoffigures,curvesandplanes–,suchasalgebraictopology,homologyandhomo-topytheory,universalalgebra,andwhathaveyou,avastmajorityofmathemati-ciansconsidersCategory-Theory(CT)vastlysuperiortoset-theory.CTalsoistheonlyrivaltoZFCinprovidingageneraltheoryofmathematicalstructureandinfoundingthewholeofmathematics.ThelanguageofCTistwo-sorted:itcon-tainsobject-variablesandarrow-variables.Anarrowsendsobjectstoobjects;an18Whenweidentifythe‘objects’thatBrading&Landry[2006],p.572,taketobe‘pre-sented’byastructureasUr-elements,thentheorem(27)alsomakestroubleforthem:everythingcanbe‘presented’byeverystructure,soallthem‘present’everything,orconversely,everystructurecan‘present’anything.\nTheCharacterisationofStructure:DefinitionversusAxiomatisation409identity-arrowsendsanobjecttoitself.Simplyput,structuresarecategories,andacategoryissomethingthathasobjectsandarrows,suchthatthearrowscanbecomposedsoastoformacompositionmonoid,whichmeansthat:(i)everyobjecthasanidentity-arrow,and(ii)arrow-compositionisassociative.ThelanguagesofCT(L")andZFC(L2)areinter-translatable.InCTthereisthespecificcategorySet,whoseobjectscanbeidentifiedwithsetsandwhosearrowsaremaps.InZFConecanidentifyobjectswithsetsandarrowswithorderedpair-setsoftypehf;Ci,consistingofamappingfandaco-domainC.19Inspiteofthefactthatsomemathematicalphysicistshaveappliedcategoriestophysics,notasinglestructuralrealistonrecordhasadvocatedreplacingZFCwithCT.Oneoftheveryfewcriticsoftheuseofset-theoryforStrR(ifnottheonlycritic)isE.M.Landry[2007],whohasarguedthattheset-theoreticalframeworkdoesnotalwaysdotheworkithasbeensuggestedtodo;butevenshedoesnotopenlyadvocateCTasthesuperiorframeworkforStrR,althoughshedoesadvocateitformathematicalstructuralism.20TheobjectsofCTaremoregeneralthantheUr-elementsonecanintroduceinZFC,becausewhereasprimordialelementsarenotsets,theobjectsofCTcanbeanything,arrows,sets,functorsandcategoriesincluded.SimilartoZFCUisthatCTdoesnothaveaxiomsthatsomehowrestricttheinterpretationof‘object’.ACT-objectisanythingthatcanbesentaroundbyanarrow,similartothefactthataset-theoreticalUr-elementisanythingthatcanbeputinaset.CT-objectsobtainan‘identity’,a‘nature’,fromthecategorytheyarein:differentcategory,differentidentity.Outsidecategories,theseobjectslosewhateverpropertiesandrelationstheyhadinthecategorytheycamefromandtheybecomeessentiallyindiscernible.OnegreatadvantageofCTisthatstructures,i.e.categories,arenotaccom-paniedbyallthesesetsthatarisebyiteratedapplicationsofthepower-setandunion-setoperation,aswehaveseenin(6),(19),(20),(21)and(22).Neverthe-less,thegrimstorywehavebeentellingforStrRintheframeworkofZFC,canberepeatedintheframeworkofCT,ofcoursewithafewappropriateadjustments.ObjectsplaytherolethatUr-elementsplayedevenbetter:weendupwithsome-ˆthingverysimilarto(28),ontopofsayingthatHeBdefinitelyisnotacompositionmonoidofobjectsandarrows.Sincethereislittlepointinre-tellingtheentirestory,weleaveitasanexerciseforthescepticalreader.TheendofthestoryisthesameproblemaboutreferenceanddescriptionwelandedinwithZFCU.Ourconclusionisthatthecategory-theoreticalroadtophysicalrealityforStrRtowalkonalsoseemsaroadtonowhere.BeforewekissZFCandCTgoodbye,wewanttoexplorethepossibilityofretainingthemboth.ThisseeminglyimpossiblepossibilityariseswhenweputtheconceptofrepresentationcenterstageandseewhetheritcanhelpuswithclarifyingtheCentralClaimofStrR(23).19SeefurtherMuller[1998],pp.485–496.20WhenLandry[2007]arguesagainstSuppes,French,etc.thataset-theoreticalframe-workisnotnecessarytomakethingsrigorous,shetakes‘necessity’inasensethatisstrongerthanSuppes,French,etc.haveevermeantitwhenevertheyuseditorsiblingphrases.\n410F.A.Muller4.REPRESENTATIONRecentlytheconceptofrepresentationhasgainedmomentuminthephilosophyofscience.21Thesimplestconceptofrepresentationconceivableisexpressedbythefollowingdyadicpredicate:structureS(HeB)representsHeB.S.French[2003]defendedthattorepresentsomethinginscienceisthesameastohaveamodelforit,wheremodelsareset-structures;then‘representation’and‘model’becomesynonymsandsodo‘torepresent’and‘tomodel’(consideredasaverb).Never-theless,thissimplestconceptionwasquicklythrownoverboardastoosimplebyamongstothersR.N.Giere[2004],p.743,whoreplacedthisdyadicpredicatewithaquadraticpredicatetoexpressamoreinvolvedconceptofrepresentation:ScientistSusesmodelStorepresentbeingBforpurposeP;(29)where‘model’canherebeidentifiedwith‘structure’.AnotherstepwassetbyB.C.vanFraassen.Asearlyas1994,inhiscontributiontoJ.Hilgevoord’sPhysicsandourViewoftheWorld,VanFraassen[1994]broughtNelsonGoodman’sdis-tinctionbetweenrepresentation-ofandrepresentation-as–drawninhisseminalLanguagesofArt(1968)–tobearonscience;hewentontoarguethatallrepresen-tationinscienceisrepresentation-as.WerepresentaHeliumatominauniformmagneticfieldasaset-theoreticalwave-mechanicalstructureS(HeB)(3).InhisnewtomeScientificRepresentation[2008],VanFraassenhasmovedessentiallytoahexadicpredicatetoexpressthemostfundamentalandmostinvolvedconceptofrepresentationtodate:Repr(S;V;S;B;F;P);(30)whichreads:subjectorscientistSisV-ingartefactStorepresentBasanFforpurposeP.Example:Inthe1920ies,Heisenberg(S)constructed(V)amathemat-icalobject(S)torepresentaHeliumatom(B)asawave-mechanicalstructure(F)tocalculateitselectro-magneticspectrum(P).Weconcentrateonthefollowingtriadicpredicate,whichisderivedfromthefundamentalhexadicone(30):ReprAs(S;B;F)iff9S;9V;9P:Repr(S;V;A;B;F;P);(31)whichreads:abstractobjectSrepresentsbeingBasanF,sothatF(S).Briefhistoricalinterlude.Giere,VanFraassenandcontemporariesarenotthefirsttoincludemanifestationsofhumanagencyintheiranalysisofmodelsandrepresentationinscience.Almosthalfacenturyago,PeterAchinstein[1965],pp.104-105,expoundedthefollowingasacharacteristicofmodelsinscience:Atheoreticalmodelistreatedasanapproximationusefulforcertainpurposes.(:::)Thevalueofagivenmodel,therefore,canbejudgedfromdifferentthoughrelatedview-points:howwellitservesthepurposesforwhichitiseimployed,andthecompletenessandaccuracyoftherepresentationitproposes.(:::)ToproposesomethingasamodelofXistosuggestitaswayofrepresentingXwhich21Suarez[2003],Giere[2004],Frigg[2006],Fraassen[1994],[2008].´\nTheCharacterisationofStructure:DefinitionversusAxiomatisation411providesatleastsomeapproximationoftheactualsituation;moreover,itistoadmitthepossibilityofalternativerepresentationsusefulfordifferentpurposes.Oneyearlater,M.W.Wartofskyexplicitlyproposed,duringtheAnnualMeet-ingoftheAmericanPhilosophicalAssociation,WesternDivision,Philadelphia,1966,toconsideramodelasagenusofrepresentation,totakeinthatrepresen-tationinvolves“relevantrespectsforrelevantforpurposes”,andtoconsider“themodellingrelationtriadicallyinthisway:M(S;x;y),whereStakesxasamodelofy”.22Twoyearslater,in1968,Wartofskywroteinhisessay‘TelosandTech-nique:ModelsasModesofAction’thefollowing(ouremphasis):Inthissense,modelsareembodimentsofpurposeand,atthesametime,instrumentsforcarryingoutsuchpurposes.Letmeattempttoclarifythisidea.Noentityisamodelofanythingsimplybyvirtueoflookinglike,orbeinglike,thatthing.Anythingislikeanythingelseinaninfinitenumberofrespectsandcertainlyinsomespecifiablerespect;thus,ifIlike,Imaytakeanythingasamodelofanythingelse,aslongasIcanspecifytherespectinwhichItakeit.Thereisnorestrictiononthis.Thusanarrayofteacups,forexample,maybetakeasamodelfortheemploymentofinfantrybattalions,andmatchsticksasmodelsofmu-mesons,therebeingsomepropertiesthatanyofthesethingssharewiththeothers.Butwhenwechoosesomethingtobeamodel,wechooseitwithsomeendinview,evenwhenthatendinviewissimplytoaidtheimaginationortheunderstanding.Inthemosttrivialcases,then,themodelisalreadynormativeandtelic.Itisnormativeinthatischosentorepresentabstractlyonlycertainfeaturesofthethingwemodel,noteverythingallatonce,butthosefeatureswetaketobeimportantorsignificantorvaluable.Themodelistelicinthatsignificanceandvaluecanexistonlywithrespecttosomeendinviewor23purposethatthemodelserves.Further,duringthe1950iesand1960iestheroleofanalogies,besidesthatofmodels,wasmuchdiscussedamongphilosophersofscience(Hesse,Achinstein,Girill,Nagel,Braithwaite,Wartofsky).Wepredictthatseveralinsightsburiedintheensuingliteraturewillbere-discoveredbythecontemporarydivisionofrepresentationalists.Endofbriefhistoricalinterlude.Onthebasisofthegeneralconceptofrepresentation(30),wecanechoWartof-skybyassertingthatalmostanythingcanrepresenteverythingforsomeoneforsomepurpose.24Inscientificrepresentations,representansandrepresentandum(tointroduceanotherpairofLatinbarbarisms)willsharesomefeatures,butnotallfeatures,becausetorepresentisneithertomirrornortocopy.Realists,a-realistsandanti-realistswillallagreethatReprAs(S;B;F)istrueonlyifonthebasisofF(S)onecansaveallphenomenathatbeingBgivesriseto,i.e.onecancalculateoraccommodateallmeasurementresultsobtainedfromobservingBorexperi-mentingwithB.WhilstforstructuralempiricistslikeVanFraassenthisisalsosufficient,forStrRitisnot.StrRwillwanttoaddthatstructureSoftypeF‘is22CollectedinWartofsky[1979],quotationonp.6.23CollectedinWartofsky[1979],p.142.24Almostanything,noteverything:hasanyoneevertakentheuniverseasawholetorepresentsomething?\n412F.A.Mullerrealised’,thatSoftypeFtrulyisthestructureofbeingBorreferstoB,sothatalsoF(B).StrRwillwanttoordertherepresentationsofbeingBthatscientistshaveconstructedduringthecourseofhistoryasapproachingtheoneandonlytruestructureofB,itsstructureansich,theKantianregulativeidealofStrR.Butthistalkoftruthandreference,ofbeingsandstructuresansich,isindissonancewiththeconceptofrepresentation-as.SomebeingBcanberepresentedasmanyotherthingsandalltheensuingrepresentationsareallhunky-doryifeachoneservessomepurposeofsomesub-ject.Thatistheideaof(30).Whentheconceptofrepresentation-asistakenaspiv-otaltomakesenseofscience,thenthesortof‘perspectivalism’thatGiere[2004]advocatesismoreinconsonancewiththeensuingviewofsciencethanrealismis.Giere[2004]attemptstohammeraweakvarietyofrealismintohis‘perspectival-ism’:allperspectivesareperspectivesononeandthesamerealityandfromeveryperspectivesomethingissaidthatcanbeinterpretedrealistically:incertainre-spectstherepresentansresemblesitsrepresentandumtocertaindegrees.Asingleunifiedpictureoftheworldishowevernottobehad.NancyCartwright’sdappledworldseemsmoreneartoGiere’sresidenceofpatchworkrealism.Aunifiedpic-tureofthephysicalworldthatrealistsdreamofiscompletelyoutofthepicturehere.Withfriendslikethat,realismneedsnoenemies.Thereisprimafacieaway,however,forrealiststoexpressthemselvesintermsofrepresentation,asfollows.First,fixthepurposePtobe:todescribetheworldasitis.Whenthisfixedpurposeleavesavarietyofrepresentationsonthetable,thenchoosetherepresentationthatisempiricallysuperior,thatis,thatperformsbestintermsofdescribingthephenomena,becausethephenomenaarepartoftheworld.Thiscanbeestablishedobjectively.Whenthisstillleavesmorethanonerepresentationonthetable,whichthussavethephenomenaequallywell,choosetheonethatbestexplainsthephenomena.Inthiscontext,VanFraassen[1994]mentionsthemanyinterpretationsofQM:eachoneconstitutesadifferentrepresentationofthesamebeings,orofonlythesameobservablebeings(phe-nomena),theirsimilaritiesnotwithstanding.Doalltheseinterpretationsprovideequallygoodexplanations?Thiscanbeestablishedobjectivelytoo,buteveryjudgmentherewilldependonwhichviewofexplanationisemployed.SupposeweareleftwithasinglestructureA,oftypeG.Thenweassertthat‘G(B)’istrue.Whenthis‘G’predicatesstructuretoB,westillneedtoknowwhat‘struc-ture’literallymeansinordertoknowwhatitisthatweattributetoB,ofwhatAisthatBinstantiates,and,evenmoreimportant,weneedtoknowthisforourdescriptivistaccountofreference,whichrealistsneedinordertoberealists.Yes,wenowhavearrivedwherewewereattheendoftheprevioustwoSections.Weconcludethatthiswayforrealists,toexpressthemselvesintermsofrepresentation(asannouncedatthebeginningofthisparagraph),isadeadend.Theconceptofrepresentationisnotgoingtohelpthem.Weconcludethatapplaudingforavarietyofdifferentrepresentations-asofthebeingsdoesnotservetheaimofrealism,StrRincluded.Theneedforsubstantiveaccountsoftruthandreferencefadeawayassoonasoneadoptsaviewofscience\nTheCharacterisationofStructure:DefinitionversusAxiomatisation413thattakestheconceptofrepresentation-asasitspivotalconcept.Fundamentallydifferentkindsofmathematicalstructure,set-theoreticalandcategory-theoretical,cantheneasilybeaccommodated.Theyare‘onlyrepresentations’.Thatismov-ingawayfromrealism,StrRincluded,dissolvingratherthansolvingtheproblemforStrRofclarifyingitsCentralClaim(23)–‘dissolved’,because‘isorhas’isreplacedwith‘isrepresented-as’.RealismwantstoknowwhatBis,notonlyhowitcanberepresentedforsomeonewhowantstodosomethingforsomepur-pose.WhenwetakeitforgrantedthatStrRneedssubstantiveaccountsoftruthandreference,morespecificallyadescriptivistaccountofreferenceandthenanaccountoftruthbymeansofreference,thenacharacterisationofstructureasdi-rectlyaspossible,withoutcommittingonetoaprofusionofabstractobjects,ismandatory.ThisissueweaddressinthenextandfinalSection.5.DIRECTCHARACTERISATIONSupposewehaveazoologicaltheoryEaboutElephants.Theword‘elephant’islogicallyspeakingan‘elephant-variable’.Sinceelephantsareconcreteobservableanimatebeingseasytorecognise,itservesnoscientificpurposetothinkofpostu-latesforEthatwillcharacterisewhatanelephantis.(Thiswouldbedifferentifwewereconsideringatheoryofants,ofwhichthereareabout12;000species.)Thatiswhyonewilllookinvaininzoologyforsuchpostulates.SupposenextwehaveamathematicaltheoryNaboutNaturalNumbers.Un-likeelephants,naturalnumbersareabstractandthereforeunobservableobjects,butlikeelephantstheyareeasytorecognise.Childrenrecogniseelephantsandnaturalnumberseffortlessly.Sincenaturalnumbersarehowevercruciallyinvolvedinthemostrigorousintellectualpraxisthathumancivilisationsofarhasproduced,i.e.mathematics,whereintheoremsareprovedaboutnaturalnumbersandtheoremsareprovedaboutotherabstractobjectsthatemploynaturalnumbers,itservesamathematicalpurposetothinkofaxiomsforNthatwillcharacterisewhatsortofabstractobjectanaturalnumberis.Weneedtoknowexactlywhatholdsfortheminordertoknowwhatwecanuseandwhatwecannotuseinproofsoftheoremsthatareaboutthemorinvolvethem.Thatiswhyonewillnotlookinvaininmath-ematicsforsuchaxioms.GottlobFrege,RichardDedekind,GiuseppePeanoandWilliamLawverehaveprovidedsuchaxioms.25WhatStrRneeds,wesubmit,isatheorySaboutStructures.JustaswecansaywithaclearphilosophicalconsciencethatZFCimplicitlydefinestheset-concept,wewanttosaythatSimplicitlydefinesthestructure-concept.26HencejustasthelanguageofZFC,L2,takestheset-conceptasprimitivebyhavingset-variables,25SeeMuller[1998],pp.56–64.26Forhowtoobtainaclearphilosophicalconscience,seeMuller[2004],[2005].Thisconceptof‘implicitdefinability’shouldnotbeconfusedE.W.Beth’sconceptthatisexpressedbythesameexpression,whichbythewayisbetterexpressedby‘semanticdefinability’.\n414F.A.Mullerthelanguageofpurestructuretheory,callitLS,musthavestructure-variables.Theconceptofstructureoughtnottobereducedtootherconcepts,suchassets(Section2)orobjects-cum-arrows(Section3).TheprojecttoconstructLSandSwillhavetowaitforanotheroccasion.Fornow,letussupposethatwepossessS.WillthatbeofanyhelpinclarifyingtheCentralClaimofStrR(23)?WillitprovidealiteraldescriptionofBthatanysubstantiveaccountofreferencerequires?Tobeginwith,justasanelephant-realistwilltakehiselephant-variablesinErunningoveratleastallelephantsonplanetEarth,StrRwilltakethestructure-variablesofLStorangeoveratleastallstructuresinphysicalreality.ThedomainofdiscourseofS,callitS,willofcourseharbouraplenitudeofstructures,justasVofZFCharboursaplenitudeofsets.WhenwetakeZFCtoprovidethefoun-dationsofmathematicsthatisusedinscience,thenonlysetsinthelowertipofV,saybelowordinallevel!+!,willbeemployed.Similarly,onlysomeofthestructuresinSoftheorySwillbecandidatesofwhichStrRwillwanttosaythattheyare‘realised’,orinstantiated,inphysicalreality.Science,physicsinparticu-lar,willtelluswhichonesarethosecandidates.StrRwillthensubmitthatthosepredicatesinLSthatsingleoutthesecandidatesprovideliteraldescriptionsofthosestructures.Exactlyhere,withintheconfinesofS,asubstantivedescriptivistaccountofreferencewillfinditsArchimedeanpoint,andreferencewillleadtherealisttotruth.RecallthattheCentralClaimofStrR,beingBisorhasstructureS(23),stoodinneedofclarification.TheclarificationwehaveintheoffingwiththeorySruns,insummaryfashion,asfollows.First,weadviseStrRtosaythatbeingBisastructureSoftypeF,where‘F’isapredicateinLSsuchthatF(S),andwhere‘S’nowisastructure-variableofLS.ThenStrRshouldsimplystipulate,onthebasisofF(S),thatpredicateFalsosuppliesastructuraltype-descriptionofbeingB,inotherwords,StrRshouldalsoassertthatF(B).This‘F(B)’istheliteraldescriptionofBthatanydescriptivistaccountofreferencecantakeaboard.But,waitaminute,beforethispaperends:howabouttheinscrutabilityofreference?Whataboutit?Thatisaproblemforeveryonewhowantstoadoptasemantictheorythatunrestrictedlyaimstosaveallandonlyobservablebehaviouristicfactsofascentanddissentoflanguage-users.Assoonasmoreisrequiredofasemantictheory,inscrutabilityargumentsareblocked,whichisnottodenytheseriousnessoftheproblemofwhatpreciselytheseadditionalrequirementsare.27Sincein-scrutabilityisaproblemforrealistsandanti-realistsalike,ratherthanaproblemforscientificrealismletaloneStrRinparticular,theissuemakeslittledifferenceintherealismdebateandwehavethereforeignoredit.¥27SeeWilliams[2008]forthestrongestinscrutabilityresultsofar:apermutationargu-mentforD.K.Lewis’Montague-basedgeneralsemantics.\nTheCharacterisationofStructure:DefinitionversusAxiomatisation415AcknowledgmentsManythankstoIoannisVotsis(HeinrichHeineUniversitat,D¨usseldorf,Germany)¨andDennisDieks(UtrechtUniversity,TheNetherlands)forcomments.REFERENCESP.Achinstein,‘TheoreticalModels’,BritishJournalforthePhilosophyofScience16(1965)102–120.N.Bourbaki,TheTheoryofSets,Paris:Hermann,1968(translationofTheoriedes´Ensembles,Paris:Hermann,1949).K.Brading,E.M.Landry,‘ScientificStructuralism:PresentationandRepresen-tation’,PhilosophyofScience73(2006)571–581.N.C.A.daCosta,R.Chuaqui,‘OnSuppes’Set-TheoreticalPredicates’,Erkenntnis29(1988)95–112.N.C.A.daCosta,S.French‘Models,TheoriesandStructures:ThirtyYearsOn’,PhilosophyofScienceAssociationProceedings67(2000)S116–S127.B.C.vanFraassen,ScientificRepresentation:ParadoxesofPerspective,Oxford:ClarendonPress,2008.Fraassen,B.C.van,‘Interpretationofscience;scienceasinterpretation’,in:PhysicsandourViewoftheWorld,J.Hilgevoord(ed.),Cambridge:CambridgeUni-versityPress,1994,pp.169–187.B.C.vanFraassen,TheScientificImage,Oxford:ClarendonPress,1980.S.French,‘AModel-TheoreticAccountofRepresentation’,PhilosophyofScience(PSAProceedings)70(2003)1472–1483.R.Frigg,‘ScientificRepresentationandtheSemanticViewofTheories’,Theoria55(2006)49–65.G.N.Giere,‘HowModelsAreUsedtoRepresentReality’,PhilosophyofScience71(2004)742–752.C.Gauker,‘ScientificRealismasanIssueinSemantics’,in:P.Greenough,M.P.Lynch(eds.),TruthandRealism,Oxford:ClarendonPress,2006,pp.123–136.J.Ladyman,‘StructuralRealism’,TheStanfordEncyclopediaofPhilosophy(Sum-mer2009Edition),E.N.Zalta(ed.),URL=hhttp://plato.stanford.edui.J.Ladyman,‘WhatisStructuralRealism?’,StudiesintheHistoryandPhilosophyofScience29(1998)409–424.E.M.Landry,‘SharedStructureneednotbeSet-Structure’,Synthese58(2007)1–17.H.Lyre,‘HumeanPerspectivesonStructuralRealism’,inthisbook.A.A.Fraenkel,etal.,FoundationsofSetTheory(2ndEd.),Amsterdam:North-Holland,1973.F.A.Muller,‘WitheringAway,Weakly’,toappearinSynthese,2009.F.A.Muller,‘TheImplicitDefinitionoftheSet-Concept’,Synthese138(2004)417-451.\n416F.A.MullerF.A.Muller,‘DeflatingSkolem’,Synthese143(2005)223-253.F.A.Muller,StructuresforEveryone,Amsterdam:A.Gerits&Son,1998.M.L.G.Redhead,‘SymmetryinInter-theoryRelations’,Synthese32(1975)77–.M.Suarez,‘Scientificrepresentation:againstsimilarityandisomorphism’,´Inter-nationalStudiesinthePhilosophyofScience17.3(2003)225–244.P.Suppes,RepresentationandInvarianceofScientificStructures,Stanford:Cen-treforLogic,LanguageandComputation(distributedbyChicagoUniversityPress),2002.P.Suppes,‘WhatIsaScientificTheory?’,in:S.Morgenbesser(ed.),PhilosophyofScienceToday,NewYork:BasicBooks,1967:55–67.P.Suppes,‘AComparisonoftheMeaningandUseofModelsinMathematicsandtheEmpiricalSciences’,Synthese12(1960)287–301.M.W.Wartofsky,Models:RepresentationandScientificUnderstanding,Dordrecht,Holland:D.Reidel,1979.J.R.G.Williams,‘ThePriceofInscrutability’,Nous42.4(2008)600–641.J.Worrall,‘StructuralRealism:theBestofBothWorlds?’,Dialectica43(1989)99–124.J.Worrall,‘RealismsforSale’,talkatworkshop‘ScientificRealismRevisited’,London:LondonSchoolofEconomics,28April2009.FacultyofPhilosophyInstitutefortheHistoryandFoundationsofScienceErasmusUniversityRotterdamBurg.Oudlaan50,H5–163062PARotterdamf.a.muller@fwb.eur.nlDept.ofPhysics&AstronomyUtrechtUniversityBudapestlaan6,IGG–3.083584CDUtrechtTheNetherlandsf.a.muller@uu.nl\nINDEXOFNAMESNotincludedare:Footnotes,Tables,ReferencesandFiguresAbrahamsen,A.263,266Bolzano,B.249Achinstein,P.410,411Bonicalzi,F.51Adorno,T.W.29Bonini,V.91Agre,202Borges,J.L.30Allen,G.263Bourdieu,P.50,54Althusser,L.42,52Boutroux,É.59Altman,202Bovens,L.88,91-93,115,142Ampère,A.-M.58Box,G.150Andler,D.305,306,308,309Boyd,R.166,189Anscombe,F.146Boyle,R.262Apel,K.O.224Brading,K.381Aristotle30,166.175,178,112Braithwaite,R.B.411Aron,R.283Brams,S.148,149Arrow,142,143,145Braunstein,J.-F.50Austin,J.L.274Brenner,A.41,48-50Babich,B.71Brentano,F.249Bach,K.273Bridgman,P.W.318Bachelard,G.29,42-49,51,52,55,59,Brunschvicg,L.42-47,57-60,62-6460,62,63,71Buchner,202Badiou,A.54Buttasi,C.89Balibar,E.50Camerer,C.91Baltag,A.116Canguilhem,G.42-45,51,52,58,62,71Barberousse,A.73Cantor,G.403Barnes,B.230Cao,T.Y.382Batterman,R.115Carnap,R.16,23-26,30,32,35,37,38,Bechtel,W.263,266,26869-72,75,76,87,111,112,114,388Belnap,N.119,120Cartwright,N.412Benacerraf,P.103-105Carus,A.W.76Bensaude-Vincent,B.49,50Cassirer,E.17,18,22,29,31,32,76,322Bergson,H.72,77Castelão-Lawless,T.53Bernard,C.58,169,170Cavaillès,J.55Berr,H.57Cech,T.202Biagioli,M.49Cei,A.381Bickhard,M.302Cevolani,G.92Birkhoff,G.116Chakravartty,A.381,390Bishop,M.91Chater,N.88,90Blaug,M.224Chimisso,C.57-60,64Blay,M.49Chomsky,N.292,293,297Blondel,C.49Christie,J.49Bloor,D.230Ciechanover,A.202Bochenski,I.M.247,248Clifford,78Bogen,J.120Cohen,L.33,76Bohr,N.349Collins,H.230Boltzmann,L.71Comte,A.41,42,44,46,58,60-62F.Stadler(ed.),ThePresentSituationinthePhilosophyofScience,ThePhilosophyofScienceinaEuropeanPerspective,DOI10.1007/978-90-481-9115-4,©SpringerScience+BusinessMediaB.V.2010\n418IndexofNamesCondorcet,N.de42,44,143Febvre,L.47,48Copernicus,N.338Fechner,G.T.69Cournot,A.42,44,58Feigl,H.68Cousin,V.59,61Fenn,202Cox,M.201Festa,R.89,92Craver,C.207,263Feyerabend,P.15,52,69,87,163Crick,F.202Fichant,M.50Crupi,V.89,90,92Fine,K.118Cummins,R.163,164,300Fisher,202Curley,E.33Fitelson,B.89-92,115Dagognet,F.42,50Fleck,L.29,254Danto,A.17,20Fodor,J.178-183,186,192,193,196,Darden,L.263295,296,303,307Darwin,C.71Follette,W.91Davidson,D.272Forster,M.200Davis,D.91Foucault,M.42,43,54,63Dawkins,R.308Frank,P.23-26,30,35,38Debru,C.48,50Fraassen,B.van100,104,107,108,389,Debs,T.345399,400,410-412Dedekind,R.413Frege,G.15,16,71,105,111,413Delaporte,F.54French,S.381-383,386,410Demopoulos,W.389Freudenthal,G.49Descartes,R.30,42,79,262,265Friedman,M.70,76,317,335,341,389Dewey,J.37,38,82,248FruteaudeLaclos,F.49Diaconis,P.352,365Gadamer,H.G.224,249Dieks,D.7,335,339,346,347,349Galavotti,M.C.10Dietrich,F.115,125-127,132,133,135,Galilei,G.112,115136,142,145Gane,M.54Dilthey,W.75Gassendi,262Dobzhansky,T.159Gauss,C.F.387Dokow,E.126,132,135Gayon,J.44,50Dorato,M.381,382Ghisleni,M.214Dorn,A.75Giere,R.275,277,410,412Dowe,P.394Gil,D.50Dray,W.260Glennan,S.266Dreyfus,H.300Golinski,I.49Duhem,P.15,16,29,31,42,43,49,52,González,W.J.7,89,90,244-254,259,53,60-62,64,71,77,79261Dummett,M.16Goodman,N.88,338,399,410Earman,J.116Gomperz,H.249Eccles,J.172Granger,C.G.233Eigen,M.170,171Greco,M.53Einstein,A.68,71,315-321,324,326,Grice,P.259,273,274327,336,337,341,343,345,347-349Griffiths,P.163Elgin,M.205,216Griffiths,T.92Ellis,B.176,177Grossi,D.145Elster,J.266,267Grünbaum,A.345Esfeld,M.381-383,390Gutting,G.41,42,48\nIndexofNames419Habermas,J.270112,321,322,324,336,341-343,347,Hacking,I.53,54,63,88,174,175,183,349,388189,190,193,237Kaye,D.91Hayek,F.v.224Keller,J.B.352-354,364,365Hahn,H.23,25,26Kelly,K.102Hahn,U.93Kepler,J.258,267Hanson,N.R.15Kilgour,D.148,149Hardcastle,G.34,35Kim,J.156,157,165,167Harnish,R.M.273Kitcher,P.31,157Harris,A.93Koyré,A.42,43,47,52,57,59,62,63Hartmann,S.7,88,91-93,115Knobe,J.87Haugeland,J.300Koehler,J.91Hawthorne,J.115Kolmogorov,A.N.360,376Healey,R.392Konieczny,S.126Hedström,P.265-267,269Kornhauser,L.144,145,150Hees,M.van125,133Krebs,202Heidegger,M.342Kripke,S.117,176Heidelberger,M.49,69,75,77,83Kuhn,T.15,34,53,69,70,87,101,104,Heisenberg,W.349,386,410106,113,163,195,229,248,254Helmholtz,H.68,70,71,75,77,328Kuipers,T.A.F.111,114Hempel,C.G.88,261,276,309Kuorikoski,J.259,267,268,276Henderson,L.170Kusch,M.55Hendricks,V.F.105Ladyman,J.381-383Hershko,A.202Lagnado,D.92Hertz,H.71,75,76Lakatos,I.69,195,196,199,200-203,Hesse,M.411232,254Hilbert,D.15Lam,V.381,382Hilgevoord,J.410Landry,E.M.409Hintikka,J.92Lange,F.A.69-71,75-77Hobbes,T.262,265Langton,R.388Hoefer,C.354,357,361Largeault,J.49Holmes,S.352Latour,B.230Holton,G.53,77,78Laudan,L.286Holzman,R.126,132,135Laugier,S.71Horst,S.286Lawvere,W.413Horvitz,H.R.202LeBlanc,G.50Howard,D.35,320Lecourt,D.41-43,48,50,51,54Howson,C.88,361Lehninger,A.201Huber,F.89,115Lehrer,K.93Hume,D.46,272,336,347,349Lehtinen,A.267,276Husserl,E.16,247,248,342Leibniz,G.W.31,386Jacob,F.172LeRoy,E.16,49,61James,W.77-83,100,248Lesk,A.199Jerusalem,W.80,81,83Levi,I.91,92Joyce,J.89Levi-Strauss,C.303Kahneman,D.91,92Lévy-Bruhl,L.45-47,57-60,62,64Kaller,R.10Lewis,D.375,376,388,390,392Kant,I.55,69,70,71,75,76,78,81,105,Libet,B.70\n420IndexofNamesLipton.P.49Morgan,M.257List,C.125-127,132,133,135,142,144,Moro,R.91145Morris,C.30,37,38Lloyd,E.210Morrison,M.257Locke,J.175,295Muller,F.381,382,386Loewenstein,G.91Müller,T.114Lombrozo,T.93Mullis,202Lorentz,H.342Nagel,E.113,115,156,411Löwe,B.114Natorp,P.76Löwy,I.49Nehring,K.125-127,134,135,137Lewis,D.38,352,354-357,360Nelson,D.201Lukasiewicz,J.15Nelson,J.90Lusanna,L.348Nestroy,J.29Lyre,H.382,384,385,387,392Neumann,J.von116Mach,E.15,16,68,70,71,75,77-79,81,Neurath,O.23,25,26,29,30,35,38,61,112,169,315,317,321,336,347,34962,69,71,72,79,81,82,226,303Mackinnon,202Newell,A.296,299MacLeod,M.10Newman,M.389,390,392,393Mäki,U.224,276,277Newton,I.31,78,115,267,295,353Malament,D.336,339,343-345Nichols,S.87Manktelow,K.90Nickerson,R.91Marcum,J.209Nietzsche,F.72Marr,D.296,303Niiniluoto,I.231,232,236,249,275Martin,M.292Norris,C.52,71Martini,C.93Oaksford,M.88,90Mastropasqua,T.89Oderberg,D.176Maudlin,T.392Osborne,T.53Maxwell,J.331Osherson,D.91,92MaynardSmith,J.308Ostrogorski,M.143,146,147Mayr,E.159Ostwald,W.71McAllester,M.54Over,D.90McArthur,D.52Paoletti,C.10McKenzie,C.90Papini,G.78Meijs,W.92Paracelsus253Mendel,G.156Pauly,M.125,133Menger,C.224,226Peano,G.78,105,413Merton,R.265Peirce,C.S.37,77,78,82,100Metzger,H.43,44,47,49,53,57,64Perales,J.91Meyerson,E.42,49,53,60,62,64,71Perry,R.B.288Mikkelsen,L.90Perutz,M.202Milhaud,G.49,59-63Pettit,P.125-127,133,135,142,144,145Mill,J.S.175,185Petzold,J.15Millikan,R.178,185,191,192Pigozzi,G.115Milne,P.90Pino-Perez,R.126Minkowski,H.331,332Plato163,175,178,247Mises,L.v.226Poincaré,H.15,16,42,49,60-62,64,68,Montague,R.11671,75,77,79,232,290,298-300,325,Montgomery,R.352328,340,347\nIndexofNames421Polizzi,G.51Sellars,W.335,338Popper,K.15,87,172,224,232,261,286Selten,R.235Prior,A.N.118Sen,A.144,234Puppe,C.125-127,134,135Serres,M.42,55Putnam,H.99,157,158,176,196,390Shanks,D.91,92Pylyshyn,Z.174Shapiro,L.196,197,199Quine,W.V.O.61,71,72,79,339,346,Shapiro,S.381,382386Shogenji,T.92Rabin,M.91Sides,A.92Rabinow,P.54Simon,H.234,236,296,299,303Rabinowicz,W.142Sintonen,M.262,277Ramsey,F.117,388Smets,S.116Redhead,M.345,382Smith,V.202,235Redondi,P.49,51Sober,E.160,196Reichenbach,H.18,24,62,68,70,75,76,Socrates163249,315-319,321-333,335-337,339,Sonnenschein,C.211341,342,344,345,349Soto,A.211Reisch,G.70Sparber,G.392Renouvier,C.58Sperber,D.259,274Rey,A.16,42,43,47,48,57,62,63Spinoza,B.111Reydon,T.189-193Sprenger,J.93Rheinberger,H.-J.44Stachel,J.381,382Richardson,A.34,35,72,76,268Steel,D.89Rickert,H.16Stegmüller,W.15Rickles,D.381,382Stein,E.90Riehl,A.72,76Stich,S.90,91Riemann,B.328Stokes,387Robb,A.343,344Stotz,K.87Romizi,D.10Strevens,M.361Rose,N.53,202Summer,202Rosenberg,A.158-164,168,171Suppes,P.401,404,405Ross,D.383Swedberg,R.265,267,269Rouse,J.286Tanaka,202Russell,B.15,71,111,388,389,391Tannery,P.60,61Sager,L.144,145,150Tarski,A.15,103Salmon,W.31Tauber,A.210Samuels,R.91Teklès-Klein,E.49Sarton,G.43Tenenbaum,J.92Sartre,J.-P.283,284,303Tentori,K.89-92Saunders,S.381,382,386Thomason,R.H.118Schaffer,J.389Thompson,M.159,165-168Schiller,F.80Thomsen,A.78Schlick,M.9,18,23,31,75,317,324,Thünen,J.H.v.276325,327,341,342,349Tijattas,M.52Schopenhauer,A.72Tiles,M.52Schrödinger,E.170Tosi,L.49Schupbach,J.93Travis,C.297Schuster,P.170,171Tversky,A.91,92\n422IndexofNamesTwardowski,K.15Uebel,T.7,29-36,38,41Urbach,P.88,361Ursus258Vaihinger,H.70,72,76Vailati,G.78-80,83VanRegenmortel,M.208,210Vernant,D.52Vineis,P.214Vinti,C.51Votsis,I.381,382Wagner,C.93Waismann,F.23Wartofsky,M.W.411Wason,P.90Watson,J.202Webb,D.55Weber,M.7,169-171,206,224Wedell,D.91Weinberg,R.211Weinberg,S.367Weir,L.53Whitehead,A.N.37Wilson,D.259,274Wimsatt,W.208Winch,P.249Windelband,W.16,33Wittgenstein,L.20,23,68,71,117,228,349Woodward,J.120,165,261,262,266,272Woolgar,S.230Worrall,J.231,232,249,381,382,387,404,405Wright,G.H.v.229,249,269-272Wüthrich,K202Wyman,J.171Ylikoski,P.265Zahar,E.49Zwicker,W.148,149