Maurice Mandelbaum The Anatomy of Historical Knowledge 1977

《Maurice Mandelbaum The Anatomy of Historical Knowledge 1977》由会员上传分享，免费在线阅读，更多相关内容在学术论文-天天文库。

THEANATOMYOFHISTORICALKNOWLEDGE BLANKPAGE MAURICEMANDELBAUMTHEANATOMYOFHISTORICALKNOWLEDGE,�W�.�THEJOHNSHOPKINSUNIVERSITYPRESSBALTIMOREANDLONDON ThisbookhasbeenbroughttopublicationwiththegenerousassistanceoftheAndrewW.MellonFoundation.Copyright©1977byTheJohnsHopkinsUniversityPressAllrightsreserved.Nopartofthisbookmaybereproducedortransmittedinanyformorbyanymeans,electronicormechanical,includingphotocopying,recording,xerography,oranyinformationstorageandretrievalsystem,withoutpermissioninwritingfromthepublisher.ManufacturedintheUnitedStatesofAmericaTheJohnsHopkinsUniversityPress,Baltimore,Maryland21218TheJohnsHopkinsPressLtd.,LondonLibraryofCongressCatalogCardNumber76-46945ISBN0-8018-1929-6LibraryofCongressCataloginginPublicationdatawillbefoundonthelastprintedpageofthisbook. CONTENTS..I··,,PREFACE..VIIPARTONE:HISTORYANDITSMODESCHAPTERONE:UNITYANDDIVERSITYINHISTORICALSTUDIES3CHAPTERTWO:VARIETIESOFSTRUCTUREINHISTORICALACCOUNTS24v VICONTENTSPARTTwo:CAUSATIONCHAPTERTHREE:CAUSALBELIEFSINEVERYDAYLIFE49CHAPTERFOUR:CAUSES,NECESSITY,ANDLAWS80CHAPTERFIVE:ONWHATANDWHYINHISTORY109PARTTHREE:OBJECTIVITYCHAPTERSIX:OBJECTIVITYANDITSLIMITS145CHAPTERSEVEN:OBJECTIVITY,CAUSATION,ANDLAWS169APPENDIXA:HUME195APPENDIXB:HARTANDHONORE199NOTES205INDEX227 PREFACE�W.·�.*Priorto1938,whenmyfirstbook,TheProblemofHistoricalKnowledge,appeared,therehadbeenrelativelylittleinterestamongBritishandAmericanphilosophersinthetopicswithwhichitdealt.Thatsituationhaschanged.Beginningin1942withCarlG.Hempel'sarticleon"TheFunctionofGeneralLawsinHistory,"andreinforcedbytheslowlyaccumulatinginterestinCollingwood'sposthumouslypublishedIdeaofHistory,philosophersintheUl1itedStatesandBritainhavebeenincreasinglyconcernedwithvariousaspectsofthesameoralliedproblems.Inthelastdecades,thenumberofseriousworksinthisareahasgrownsogreatthatitwouldbedifficulttodiscusseventhen10stimportantinacareful,critical,andfair-mindedway.Therefore,IhavemadereferencetosubsequentdiscussionsonlywhenIhavehappenedtofindthemespeciallyapposite.Itrustthatthiswillnotbetakenasamarkofdisrespect.Vll VillPREFACEInthepresentworkIhavereturnedtothemainissueswithwhichmyearlierworkwasconcerned,butitisonlynaturalthatIshouldnowtreattheseissuesinadifferentway.Insomecasesthishasbeenduetochangesthathaveoccurredinphilosophyitself,particularlyinthisfield.Forexample,theproblemof"fact"and"value"loomslesslargeatthemomentthanitoncedid.Takingitsplaceintheforefrontofattentionisthequestionofwhatkindsofexplanationaretobesought,orcanbefound,inhistory.Myownconceptionoftheproblemshasinsomecasesalsochanged.Forexample,Ihavebecomeincreasinglyawareofthevarietypresentindifferentkindsofhistoricalinquiry.Inwhatfollows,Ihaveattemptedtofocusattentiononthisfact,andtorelateittosomeoftheprobleITIswithwhichphilosophersdealingwithhistoryareconcerned.Itisinthisarea,andinInyattempttoofferamoredetailedandadequateexpositionofmyviewsconcerningcausation,thatthenoveltyofthepresentworkmaychieflylie.Ihavehadthehonortopresentvariousportionsofthisbook,invaryingforms,whenIdeliveredtheRabbiIrvingLeveyLectureatPrincetonUniversity,aFranklinJ.MatchetteFoundationLectureattheUniversityofCaliforniaatIrvine,theArthurO.LovejoyLectureatJohnsHopkins,andtheAlfredSchutzMemorialLectureattheUniversityofCaliforniaatSanDiego.Ithankmyhostsattheseinstitutionsfortheirinvitations,andtheseandotheraudiencesfortheirfriendlvcriticism../IalsowishtothankW.H.DrayandC.BehanMcCullaghfortheircarefulreadingofmymanuscript.Eachsavedmefrommanyinfelicitiesinexpositionandfromsomeimportanterrors;Iammostgratefultothenl. PARTONEHISTORYANDITSMODES..••••I BLANKPAGE ChapterOneUNITYANDDIVERSITYINHISTORICALSTUDIESInthisbookIshallattempttoclarifysomeproblemsthatareofconcerntophilosophersandtohistorianswhoreflectonthenatureofhistoryasadiscipline,onwhatconstitutesexplanationinhistory,andwhetherhistoricalknowledgeisasreliableasotherformsofknowl�·edgemaybe.Althoughtheseproblemshavelongbeendiscussedinoneformoranother,mostoftheissuesthatremaincurrentbegantobeintensivelydiscussedinthelastdecadesofthenineteenthcentury.Sincethen,thephilosophicIiteratureconcernedwithhistoricalknowledgehasgrownenormously,andtheinterestofhistoriansappearsnottohaveslackened.WhileIhopenottohaveoverlookedanyofthemostsignificantcontributionstovariousaspectsoftheargumentsthathaveflourished,itismyaimtoofferaconsistentviewofmyownratherthantoattemptanassessmentofalltheadmittedlyimportant3 HISTORYANDITSMODES4positionsthathavebeenadvanced,criticized,anddefendedbyothers.Asafoundationforwhatistofollow,Ishallfirstattempttoshowinwhatsenseonemayclaimthatthereisafundamentalunityinhistoricalstudies;Ishallalsoattempttoindicatesomeoftheimportantwaysinwhichthesestudiesdifferfromoneanother.Whenoneconsiderstheproliferationofformsofhistoricalwritingsincethebeginningoftheeighteenthcenturyandthedifferencesintheinterestandthepracticesofhistoriansofdifferentbackgrounds,itmayseemfoolishtopropose-asIshalldo-thatthereisunityaswellasdiversityinhistoricalstudies.Practicinghistoriansarethoroughlyawarenotonlyoftheindividualdifferencesamonghistorians,butofthelong-termchangesandtheshort-runfashionsinthekindsofsubjectmatterwithwhichhistorianshavebeenconcerned.Thisdiversityhasbeensoapparentthatonefindsveryfewhistoriansmakingaseriousattempttocharacterizeinacarefulorp;recisewaywhatisdistinctiveabouttheirdiscipline.Ontheotherhand,whenphilosophershavebeenconcernedwiththeproblemsofhistoricalknowledgetheyhavegenerallyfailedtorecognizethediversitythatexistsinthefieldofhistoricalstudies:Iknowthatmyownworkhassufferedfromthatdefect.InwhatfollowsIwishtolayequalstressontheunityandonthediversitythatarepresentwhenonetakesintoaccountthewholerangeofhistoricalstudies.ItistoaconsiderationofthequestionofunitythatIshallfirsttum.IItisacommonplaceintheIiteratureofoursubjectthathistoriansareconcernedwithparticulareventsthatoccurredatspecifictimesandplaces,andnotwiththemonlyinsofarastheyrepresenteventsofagiventype.Tobesure,historiansoccasionallyembarkonanalysesofsimilarevents,comparingandcontrastingtheiroriginsandnatures,asR.R.PalmerinTheAgeofDemocraticRevolutionstudiedonefacetofthepoliticalhistoryoftheeighteenthcentury.Theyalsofrequentlychoosesomeverylimitedtopicforstudybecausetheytakeittobetypicalofotherphenomenainwhichtheyareinterested,asamedievalFrenchhistorianmaystudyaparticularvillagetofindoutaboutFrenchvillagelifeatthetime.Similarly,ahistorianmaystudythelivesofpersonsbelongingtoaparticularsocialclassinVictorianEnglandinordertogainabetterunderstandingoftheclassstructureandclassattitudesprevalentthereatthetime.Suchstudiesaresometimessaidtoinvolve"generalizations"becausetheylllove UNITYANDDIVERSITYINHISTORICALSTUDIES5fromwhatwastrueinspecificcasestowhatgenerallyheldtrueatthetime.Itisimportanttonote,however,thatsuchgeneralizationsaresinlplymeansofdiscoveringanddescribingwhatwascharacteristicallytrueofsomeparticularplaceoversomeparticularspanoftime.Oneshouldnotsupposethatbecausehistorians"generalize"inthiswaythattheyareattemptingtoformulateorconfirmanygeneralizationastowhatalways,orusually,occursinsituationsofagiventype.Inthisrespecttheiraimsaredifferentfromthosecharacteristicofsociologistsorsocialpsychologistswhomightbedealingwiththesamenlaterials.Thus,thefamiliarthesisthathistoriansareconcernedwiththeparticular,ratherthanwithestablishingexplanatorygeneralizations,appearstomesound.TheclassicformulationofthisdistinctionisattributabletoWindelband,whodistinguishedbetweentheidiographicaimsofthehistorianandthenomotheticobjectivesofthesciences.1Unfortunately,thosewhoinitiallyintroducedthiscontrastintodiscussionsofhistoricalknowledgeformulateditinawaythatinvolvedmanyotherissues.Forexample,thecontrastbetweenidiographicandnomotheticinterestsbecameentangledwiththeclaimthateachhistoricaleventisuniqueandunrepeatable,andthatinthisrespecthistoricaleventsaredifferentfromtheeventswithwhichthenaturalsciencesdeal.Fromthisclaimitwasthoughttofollowthatwhilenaturecouldbeadequatelydescribedintermsoflaws,lawswerenotinprincipleapplicableinthedomainofhistory.Unfortunately,thisinturnwasthoughttoimplythathistoriansmustusemethodstotallydifferentfromthoseusedinthenaturalsciences.ItisagainstthistangledbackgroundofissuesthatCarlHempel'sfamouspaper"TheFunctionofGeneralLawsinHistory"shouldberead.Unfortunately,theseissueswerenotdisentangledwithinthatpaperitself.ForexampIe,thequestionofthefunctionofgenerallawsinhistoricalexplanationisnotequivalenttothequestionofwhatitisthathistoriansareattemptingtodo,yetHempelfailedtodrawthisdistinction:Mostofhisargumentwasinfactdirectedtowardshowingthathistoricalexplanationsinvolvetheuseofgenerallaws,butfromthishedrewtheunwarrantedconclusionthathistoricalstudiesarenotprimarilyconcernedwiththedescriptionofparticularevents.2SincethisconfoundingoftwodistinctissuesandothersimilarconfusionscontinuetobepresentinmuchoftheliteraturethatstemsfromHempel'sextremelyinfluentialarticle,onlyareturntotheoriginalidiographic-nomotheticdistinction,disentangledfromotherissues,willpermitustomakeafreshstart. 6HISTORYANDITSMODESItismyclaimthatanyworkwetaketobehistoricalinnaturepurportstoestablishwhatactuallyoccurredataparticulartimeandplace,orisconcernedwithtracingandexplainingsomeparticularseriesofrelatedoccurrences.However,thisdoesnotentailthatinfulfillingsuchataskthehistorianmaynot,atcertainpoints,havetorelyongeneralizationsinordertoofferacoherentaccountofsomeoftheoccurrenceswithwhichhedeals.Forexample,inattemptingtogiveanaccountofaparticularrevolution,ahistorianoftenhastomakeuseofcertaingeneralassumptionsconcerninghowindividualsgenerallybehaveinparticularsortsofsituations,suchasthosethataroseinthecourseofthatrevolution.Thisdoesnotinvolveanyabdicationofhisprimaryenterpriseasahistorian,sincebeforehecanmakeuseofanygeneralizationsconcerningtheirbehaviorhe.mustestablishthenatureofthesituationsinwhichtheseindividualswereplaced,andhemustalsohaveestablishedhowtheydidinfactbehave.Thefunctionofhisassumptionsconcerninghumanbehavioristhatoflinkingthebehaviorofindividualsandofgroupstothesituationsinwhichtheyfoundthemselves,renderingitintelligiblewhytheyactedastheydid.Unlikeasocialpsychologistwhomaybeconcernedwithrevolutionarymovements,thehistorianisnotattemptingtoshowhowcertainfactorsthatmaybepresentinallhumanbehaviorcanbeusedtoaccountfortheoccurrenceofrevolutions.Also,unlikesociologistswho'attempttoestablishtheoriesofrevolutions,historian,s,:areconcernedwithwhatactuallyoccurredin'specific'cases,�.ratherthanwithdiscoveringthegeneralsociologicalconditionsthatservetoexplainthosetypesofsocialandpoliticalchangedesignatedasrevolutions.Theforegoingexampleshouldmakeitclearthatthetaskofthehistorianisquitedifferentfromthatofthepsychologistorsociologist,whoseapproachtothesamesubjectmatterisnomotheticratherthanidiographic.However,itwouldbemisleadingtoassumethatthehistorian'sapproachisinallcasesasfreefromgeneraltheoreticcomponentsastheprecedingexamplemayhavesuggested.Insomecasesatleast,hemusthaveabroadacquaintancewithsocialtheoryinordertorecognizevariousalternativetypesofexplanationsthatcouldbeusedtoaccountforthefactswithwhichheistodeal.Asoneillustrationofwherethisisneeded,considertheproblemofahistorianwhowishestotracethehistoryofaparticularsetofbeliefsandanassociatedpatternofactions,suchasbeliefsinwitchcraftandtheprosecutionofpersonsaswitches.Ahistorianofwitchcraftmustseektoestablishwhetheritismoreplausibletoholdthatthespread UNITYANDDIVERSITYINHISTORICALSTUDIES7ofthebeliefsandofthepatternsofactionwithwhichheisconcernedisbestaccountedforintermsofdirectinfluenceandimitation,orinsomeotherway.Forexample,theymighthavearisenindependentlyofeachotherbecauseofthepresenceofsimilarpsychologicalorsociologicalconditionsindifferentplacesatthesametime.Inshort,thehistorianwilloftenbeenmeshedinpreciselythesamesortsofproblemsthatanyanthropologistmustfacewhenhedealswiththequestionofwhetherthedistributionofaparticularsetofculturetraitsistobeaccountedforbydiffusionorwhetheritisanexampleofindependentorigins.Thiswillalmostalwaysinvolvegeneralknowledgebasedonacomparisonofinstances;itisnotusuallyaquestionthatcanbesatisfactorilysolvedmerelybyexamininganyonecase.3However,thefactthatahistorianmustmakeuseofgeneralknowledgeofthissort,drawnfromhisfamiliaritywithothercases,doesnotsignifythathistaskinanyparticularcasedivergesfromaconcernwithunderstandinganddescribingwhatactuallyoccurredinthatonecase:Hisinterestremainsrootedinthatcase,andisidiographic.Iturnnowtoasecondgenericcharacteristicofhistoricalstudies,nomatterinwhatfieldtheyarepursued:Theydependuponinquiry,thepurposeofwhichistoestablishthetruthconcerningparticulareventsthatdidactuallyoccur.Inthis,historicalstudiesaretobedistinguishedfromstories,myths,andmemories,witheachofwhichtheyhaveonoccasionbeenclaimedtobeintimatelyrelated.Letusfirstconsiderthefactthattheelementofinquirydistinguishesthemfrommemory.Whileourownmemories,ortherecordsoftheobservationsofothers(whichinanextendedsensecanbecharacterizedasaformof"memory"),maygiveusinformationusefulforconductinganinquiryintowhathasoccurred,suchinformationdoesnotconstitutemorethanastartingpointforhistoricalstudies.Theaccuracyofourmemory,oroftherecords,mustbetestedagainstothermemoriesandrecords,andrelationshipsmustbeestablishedamongahostofsuchrememberedandrecordedevents.Itisforthisreason,andnotforthereasonsassignedbyBenedettoCroce,thatonemustdistinguishbetweenhistoryandthetypesofrecordsCrocedesignatedasmerechronicle:Chroniclespurporttobearecordofasetoffactsthatdidoccur,butachronicleneithersupportsitsstatelnentsbyauthenticatingthem,nordoesitnecessarilyprovideanaccountoftherelationshipsinwhichtheystandtooneanother.Itisthetaskofthehistoriantotakechroniclesandallotherrelatedrecordsofthepast,andtoestablish,throughinquiry,thevariousrelationshipsthatob- 8HISTORYANDITSMODEStainedamongtheserecordedfacts.OneofthemeritsofCollingwood'sworkonthenatureofhistorywasthatheinsistedupontheessentialroleinquiryplaysinanythingwedenominateasahistoricalwork.Thisthesisis,however,independentofotherdoctrineswithwhichCollingwoodassociatedit,suchashiscontentionthathistoryisthereenactmentofpastthought.Tomarkthisdifference,andalsotoseparatemyuseofthedistinctionbetweenhistoryandchroniclefromtheviewofhistorytowhichCrocesubscribed,Imightsaythattheroleofinquiryinhistoryisnotinstrumentaltoreliving,reenacting,orinanyotherwayexperiencingorbringingalivewhatispast.Itsroleistoallowustoknowwhatoccurred,andtoknowitasfullyandaccuratelyaswecan.ToputthematterinasstronganopposinglightasIcan:Historyaims,andoughttoaim,atbeingwissenschaftlich,whichistosaythatinlayingclaimtotruthitmustbeabletoadvanceexternalevidencethatvouchesforitstruth;indefaultofthis,itisnottobeconsideredahistoricalstudy.Inquiryinhistorynotonlyservestodistinguishitfrommemoryandfrommerechronicles,butalsodistinguishesitfromstoryandmyth.Neitherthestorytellernorthemythmakerneedseektoestablishthattheelementsenteringhisaccountactuallyoccurred;evenifheaimsatconveyingatruththroughhisstoryormyth,thattruthwillnotdependuponwhethertheelementsthroughwhichitiselucidateddoinfactrefertoactualevents.Althoughthereareworksofart,especiallyinpaintingandinfiction,thatstrivetodepictspecificevents,ortodepictveryconcretelythekindsofeventsthatdidoccurataparticulartimeandplace,suchattemptsarebynomeanscharacteristicofallworksofart;andifoneclaims-followingAristotlethatpoetryistruerandofgreaterimportthanhistory,itisnottosuchfeaturesoftheworkthatonewouldbeinclinedtorefer.Rather,thetruthonewouldusuallywishtoassigntoaworkofartortoamyth,ifoneweretoassignitsuchanattributeatall,wouldlieinitsrevelatorycharacterwhentakenasawhole.AlloftheeventsinastoryInaybefictitious,andinsomenovels,suchasKafka'sTheCastle,theymaybearonlyambiguousresemblancestoeventsthatordinarilyoccurineverydaylife;yetthestorycan,likeamyth,carryameaningthatrelatestohumanexperienceinsuchawaythatweareapttoregarditastrue,andasprofoundlytrue,orperhapsasfalse.Fiction,however,isnothistory.Onewouldbeextremelysillytoaskofanovelistwhethertherewasreallysomepersonexactlylikeacharacterhehaddepictedinoneofhisnovels,ortoaskwhetheranyonewhohadinsomerespectsservedasamodelforoneofhis UNITYANDDIVERSITYINHISTORICALSTUDIES9charactershad,inreallife,stoodinthesamerelationstootherpersonsashadbeendepictedinthenovel.Itisnotsilly,however,butabsolutelyessentialthatwedemandofahistorianthatheincludewithinhisaccountonlysuchpersonsasreallyexisted,andonlysucheventsandrelationshipsamongthesepersonsasdidreallyoccur.Puttingthemattermoregenerally,weimposeuponhistoricalstudiesatruthconditionthatisnotonlydifferentfromanyappliedtoartormyth,butonethatmaybemoreseverethanthetruthconditionplacedonthetheoreticalstructuresofthenaturalsciencesbysomephilosophersofscience.Weaskthatahistorian'saccountofaseriesofeventsbetruenotonlyofthatserieswhenviewedasawhole,butthatitsaccountofallofthecomponentelementsincludedwithintheseriesalsobetrue.Ifthisdemandwereabandoned,historywouldnotbeadescriptivediscipline,dependentforitstruthontheaccuracywithwhichitcouldinferwhathadactuallyoccurred.Thisclaim,thathistoricalstudiescharacteristicallyaimatdiscoveringanddescribingthenatureofparticulareventsandseriesofevents,andthatinquiryenablesthenltodoso,isaclaimthathasoftenbeenchallenged.Suchchallengeshaveoccasionallyarisenbecausehistoricalaccountsrefertothepast,which(beingpast)cannot,ofcourse,bedirectlyobserved.BothCharlesA.BeardandCarlL.Beckersonletilnesappealedtoargumentsofthistype.Takenbyitself,however,thisformofargumentisnotstrong.Onenlaynote,forexample,thatthehistoricalaspectsofvariousnaturalsciences,suchas·geology,arenotgenerallychallengedsimplybecausetheyconcernthepast.Norisitusual-orplausible-tochallengeallhistoricalreconstructionswhichconcernpastactions;forexample,wedonotbelieveitintrinsicallyimpossibletoestablishinacourtoflawwhatoccurredinaspecificcase.Whateverdifficultiesmaybesetsuchinquiries,theyarenotusuallysubjectedtowholesalechallenge.Itisthereforenotsurprisingthatskepticismwithrespecttoourhistoricalknowledgeisonlyoccasionallybasedonthefactthathistoriansdealwiththepast;itsmostusualsourcesaretobefoundinotheraspectsofthesubjectmatterwithwhichhistoriansareconcerned.Forexample,itisoftenassumedthattheparticularinterests,antipathies,andsympathiesofhistorianscannotbedisentangledfromthematerialswithwhichtheydeal.Furthermore,differenthistoricalaccountsareoftentakentobeSOdiverseanddisparatethatnogeneralstandardsareapplicabletothosethatwerewrittenatdifferenttimes,aroseoutofdifferentinterests,andwereguidedbydifferentassumptions.However,wereitpossibletoshowthatdifferinghistorical 10HISTORYANDITSMODESaccountsarenotintrinsicallyunrelated,butthattheytendtodovetailinspiteofdifferencesintheprejudgmentsofdifferenthistorians,andthattheyexploredifferentaspectsorfacetsofasinglebodyofconnecteddata,thenitmightbeclaimedthattheseaccountssupplement,corroborate,andservetocorrectoneanother.Werethisthecase,itwouldnotbenecessarytoabandontheassumptionthatthereisacommonstandardforjudginghistoricalinquiries,howeverdiversetheymaybeincontentorintheirapproachestothatcontent.Ishalllaterexaminethisprobleminsomedetail.AtthispointIshallconfinemyattentiontoshowingthatthedatausedindifferingsortsofhistoricalaccountsdoinfacthaveimportantfeaturesincommon,thuslayingthegroundworkformyclaimthattheyconstituteaconnectedsetofdata.ThenatureofthesedataconstitutesthethirdandfinalgenericcharacteristicofhistoricalaccountswithwhichIshallbeconcerned.Ithasoftenbeensupposedthatthedataofhistoryincludeallthatisaccessible,orbecomesaccessible,withrespecttowhatanyhumanbeinghasdone,said,felt,orthoughtoverthewholeofthepast.4Suchcharacterizationsofwhatbelongstohistoryarenotunusual,andtheyappeartobeattractivebecausetheydonotplaceantecedentrestrictionsonhistorians,exceptforthestipulationthathistoryisconcernedwiththehumanratherthanthenonhumanpast.Yet,characterizationsofthistypearetooinclusive,unlesstheirscopeistacitlyrestrictedtomakethemconformtotheactualpracticeofhistorians.Thisbecomesevidentwhenoneconsidersthefactthatthethoughts,feelings,andactionsofanyindividualmaybeseeninrelationtovariouscontexts,ofwhichtheirhistoricalcontextisonlyone.Forexample,apsychoanalystisinterestedinthethoughts,feelings,andactionsofhispatient;ajudgeisinterestedinthethoughts,feelings,andactionsofpersonsbroughtbeforehiminatrial;anditisnormalforanyparentorchildtoattendtowhatothermembersofhisfamilyfeelandthinkaboutmanymatters-otherwiseneitherparentnorchildwouldknowwhattoexpectwhenbehavinginonewayratherthananother.Thehistorian,however,viewsthethoughts,feelings,andactionsofindividualsinaspecialcontext:Heviewsthemintheirsocietalsetting,thatis,intermsofthevariouswaysinwhichtheyaffected,orwereaffectedby,thesocietyinwhichtheytookplace.Itisonlyinsofarasindividualsareviewedwithreferencetothenatureandchangesofasocietyexistingataparticulartimeandplacethattheyareofinteresttohistorians;itisnotthethoughtoractionofanyindividualviewedmerelyasthisspecificindividualwithwhich UNITYANDDTVER>SITYINHISTORICALSTUDIES11thehistorianisconcerned.ForthisreasonIelsewherecharacterizedthedomainofhistoriansas"thestudyofhumanactivitiesintheirsocietalcontextandwiththeirsocietalimplications."5ThatcharacterizationsufferedfromonedefectInowwishtocorrect:Ittended;tolayunduestressontheinstitutionalstructuresofsocietiesandlefttoo1ittleroomfortheconsiderationofthoseaspectsofhumanculture,suchasthearts,technology,orphilosophy,thatundoubtedlyhavesomerelationtotheorganizationofthesocietiesinwhichtheyarefound,butwhichcanalsobetheobjectsofindependent.;historicalinquiries.Inwhatfollows,muchmorewillbesaidconcerningthedifferencebetweeninstitutionallyorientedhistories,whichIhaveelsewherereferredtoas"generalhistory,"andhistories;'ofspecificaspectsofculture,whichItermed"specialhistories";6hereitisnecessaryonlytodiscriminatethewayinwhichIshallbeusingtheterms"society"(or"societal")and"culture"(or"cultural").Socialscientists,andsocialanthropologistsinparticular(e.g.,A.L.Kroeber,M.J.Herskovits,S.F.Nadel,andClydeKluckhohn,amongothers),drawadistinctionbetween"society"and"culture,"butnostandarddefinitionsofthesetermshavebecomeestablished.Ishallbeusingthem,andtheircognates,inthefollowingway.Asociety,Ishallhold,consistsofindividualslivinginanorganizedCOIDlllunitythatcontroIsaparticularterritory;theorganizationofsuchacOlllmunityisprovidedbyinstitutionsthatservetodefinethestatusoccupiedbydifferentindividualsandascribetothemtherolestheyareexpectedtoplayinperpetuatingthecontinuingexistenceofthecommunity.Iwishtodrawadistinctionbetweenasociety,asthusdefined,andmyuseoftheterm"culture."Indoingso,Ishallnotconformtoeitheroftwousagescurrentlyfoundinthewritingsofmostanthropologists.Inonesuchusagethetermisnotprimarilyemployedasagenericterm,butisusedtodesignatetheparticularwayoflife"characteristicofasociety,aswhenonemightsay"thecultureoftheNavahodiffersfromthatoftheKwakiutl."7Anotherusagetendstorestricttheterm"culture"toasystemofideasandvaluesthatshapesthebehaviorofindividualsinasociety.8Formypurposesatleast,neitherofthesewaysofdefiningcultureseemsadequate.Instead,Ishallusetheterminagenericsense,asE.B.Tylororiginallydid,andIshallincludeaselementsofcultureartifactsandthewaysinwhichtheyareused,asdidhe.9Thechiefreasonfordefiningitinthiswayisthatvariousculturalelementsneednotbetiedtoapartic- 12HISTORYANDITSMODESularsociety,butcanmigratefromsocietytosociety.Sotoocanartifacts,asmanyfamiliarstudiesofthediffusionofculturetraitsandculturecomplexeshaveshown.Also,nocomplexsocietyislikelytobeasmonolithicinitscultureasspeakingof"thewayoflife"or"thevaluesandideas"ofasocietywouldseemtosuggest.Althoughpersonsoccupyingdifferentinstitutionalrolesinasocietymayshareacommoncultureinmanyrespects,theymayalsobelongtodifferentculturalsubgroupswithinthatsociety,beingdistinguishedfromoneanotherbyspeech,possessions,anddress,aswellasbymanners,tastes,andmorals.Ishall,then,usetheterm"culture"asagenerictermdesignatingwhateverobjectsarecreatedandusedbyindividualsandwhateverskills,beliefs,andformsofbehaviortheyhaveacquiredthroughtheirsocialinheritance.Definedinthisway,"culture"doesnotincludeinstitutions,suchaskinshipsystemsorrulesgoverningthedistributionofpropertyandthedivisionoflabor,whichdefinethestatusandrolesofindividualswithinasocietyandregulatetheorganizationofitslife.Rather,Iamusing"culture"asagenerictermcoveringlanguage,technology,thearts,religiousandphilosophicattitudesandbeliefs,andwhateverotherobjects,skills,habits,customs,explanatorysystems,andthelikeareincludedinthesocialinheritanceofvariousindividualslivinginaparticularsociety.Usingthesedefinitionsof"asociety"andof"culture,"specialhistoriestracevariousaspectsofcultureastheyariseandchangeinasociety,orastheycrosstheboundariesseparatingsocieties,whereasgeneralhistoryisconcernedwiththenatureofandthechangesinparticularsocieties.Althoughthe'contrastIoriginallydrewbetweengeneralhistoryandspecialhistorieshasnothadanymarkedimpactondiscussionsofhistoriography,ithasonoccasionbeenreferredto;whatisofmoreimportanceisthatothershavequiteindependentlyusedtheterm"generalhistory"inessentiallythesamesenseasthatinwhichIusedit.MostnotableamongtheparallelsisthedistinctiondrawnbyOttoBrunner,themedievalhistorian,between"historyasadiscipline"andotherhistoricalstudies,adistinctionhedevelopedinalectureentitled"DasFach'Geschichte'unddiehistorischenWissenschaften."Heheldthathistoryasadiscipline-"History,toutcourt,"asitissometimesput-concernstheactionsofhumanbeings,bothindividuallyandingroups,inthecontextofaparticularorganizedsociety;thestructuresofsuchasocietydistributerelationshipsofpoweramongtheindividuals,anditisintermsofthesepowerrelationshipsthattheindividualsact.Incontrasttogeneralhistory,takeninthis UNITYANDDIVERSITYINHISTORICALSTUDIESsense,whichhecalled"GeschichteimengerenSinn,"Brunnercharacterizedthespecialformsofhistoricalstudy("diehistorischenFachwissenschaften")asdealingnotwithindividualsandgroups,butwiththeirculturalproducts,suchasphilosophicandreligiousviews,worksofliteratureandart,whichthehistorianinvestigates,interprets,anddepicts.loInasomewhatdifferentmanner,W.H.B.Court,theeconomichistorian,raisedthequestionoftherelationbetweeneconomichis.:.toryandgeneralhistory.llInthisconnectionhedistinguishedbetweenthehistoryofsinglenations,whichhereferredtoas"generalorintegralhistory,"andwhathereferredto(notaltogetherhappily,Ithink)as"universalhistory."Whilegeneralhistoryattemptstodealwithallaspectsofaparticularsociety,andmaythusbecalled"integral,"aneconomichistorianneedstotranscendtheconfinesofsinglenations,tracingthespreadofeconomicinstitutionsanddealingwithwholetradingareas,orevenwithallnations,fromaneconomicpointofview.Economichistory,inthissense,isaspecialdiscipline,aFachwissenschaftinBrunner'sterminology,concernedwiththenatureofandchangesinoneaspectofhumancultureratherthanattemptingtodepictthenatureofaparticularsocietyandthechangesinit.Tobesure,whetheronedealswitheconomicsorartorreligion,orwithanyotheraspectofhumanculture,innovationsandchangetakeplaceinparticularsocietiesandformpartofthelifeofthatsociety;theyarethereforeofconcerntothegeneralhistorianwhoattenlptstounderstandanddepictthenatureofagivensocietyanditschanges.Yet,theseaspectsofculturemayalsospreadbeyondthepointsoftheiroriginandpresentthespecialhistorianwithadistinctsubjectmatterofhisown.Thus,asbothBrunnerandCourtsuggest,andasIwishtoemphasize,whilethetasksofthegeneralhistorianandofthosedealingwithspecialhistoriesaredifferent,theyarenonethelesscomplementary,lendingeachothermutualsupport.Inthelightofwhathasbeensaidconcerninggeneralhistory(whichdealswithsocieties)andspecialhistories(whichdealwithculture),itmightbethoughtthatIwishtominimizethehistorian'sconcernwithindividualhumanbeings.Thisisnotthecase.ThatonwhichIwishtoinsistismerelythefactthatinorderforanindividualtobeofconcerntoahistorianhischaracterandactionsmustbeviewedinrelationtotheplacethatheoccupiedandtherolethatheplayedinthelifeofasocietyorinrelationtosomefacetofculture;andthisholds,aswell,forgroupsofindividuals.Suchrelationships HISTORYANDITS,MODESmayofcoursebeofdifferentkinds.Forexample,ahistorianmaybeconcernedwiththeactionsofanindividualbecausehehadacertaininstitutionalstatuswithinaparticularsociety,orbecauseofsomeimportantinfluencehehadinchangingthatsociety,orsimplybecausehemayserveasrepresentativeofsomeaspectofthesocialorculturallifeofthatsociety.Whatisimportanttobearinmindisthattheplaceindividualsoccupyinanyhistoricalaccountisrelativetothesocialorganizationandthecultureofthesocietythehistorianseekstounderstandanddepict.12Inordertoavoidmisunderstanding,itisalsoimportanttonotethatahistorian'sidiographicinterestinsomeparticularsocietydoesnotsignifythathistoriansmaynotundertakecomparativehistoricalstudies.Thesestudies,whicharebecomingofincreasinginteresttothehistoricalprofession,oftenhelptodirectattentiontoaspectsofsocietalorganizationorculturalactivitiesthatmightbeoverlookedbyanyhistorianswhosefamiliaritywithdifferentformsoforganizationandofculturetendstoberestrictedtowhatwasmostcharacteristicofthenatureandchangestobefoundinthehistoryoftheirownsociety.Furthermore,thestrikingpresenceofquitedifferentfeaturesinothersocietiesmaycallattentiontothepresenceofanalogousfeaturesinone'sown.Thus,comparativehistoricalstudieshelptopreventmisinterpretationsofothersocietiesandtheirculture,andmisinterpretationsofourown.However,suchstudiesneednotinvolveanabandonmentofthehistorian'sidiographicapproach,turninghistoriansintopsychologistsorsociologists;theysimplyattesttothefactthatadequatehistoricalunderstanding,inanyfield,demandssophisticatedinquiry:Whatispresentandneedstobedescribeddoesnotalwayslieonthesurfaceofthoseformsofsocialorganizationandofculturallifewithwhichhistoriansseektocometogrips.Insum,ithasbeenmycontentionthathistoricalaccountsareconcernedtoestablishthroughinquiry,andtovalidatethroughevidence,occurrencesthatrelatetothenatureofandchangesinaparticularsociety,or-usingthesamemethods-totracecontinuityandchangeinthosehumanactivitiesthatwemaydesignateformsofculture.IIItwillberecalledthatinintroducingmycontentionthathistoryasadisciplineisconcernedwithsocietalfactsandwiththeelementsof UNITYANDDIVERSITYINHISTORICALSTUDIEScultureproducedbymenindifferentsocieties,IindicatedwhyIattachimportancetoshowingthatthereisunityaswellasdiversityinhistoricalaccounts.Iftheeventswithwhichtheseaccountsdealareintrinsicallyrelated,andifthereareinterlockingdatathathistorianscanuseinestablishingtheirrelationships,thenvarioushistoricalaccountscanservetosupplement,corroborate,andcorrectoneanother.Underthesecircumstancesonewouldnotneedtoabandontheassumptionthatthereissomecommonstandardforjudginghistoricalinquiriesinspiteofthediversityintheircontentandintheirapproaches.OneimportantwayofestablishingthatthereareinterlockingrelationshipsamonghistoricalaccountsistotakenoteofthephenomenonIhaveelsewheretermedthescaleofsuchaccounts.13Somehistorianssurveylongertimeintervalsthandoothers,becauseofthesubjectmatterwithwhichtheychoosetodeal,andsomeareconcernedwithmoreinclusiveentitiesthanareothers,whochoosesubjectsofsmallerscopeandenterintothemingreaterdetail;butthesedifferencesinscaledonotmaketheiraccountsirrelevanttoeachother.Onewayofillustratingthisfactistosaythatinthefieldofgeneralhistory(althoughnot,asweshallsee,inallspecialhistories)theseriesofeventswithwhichthehistoriandealsisanindefinitelydenseseries,justasisthegeographicalterritorywithwhichacartographerdeals.Letusexaminewhat,precisely,thismeans.Amapmakeralwaysoperatesonsomescale;letusinitiallysupposethatheisdrawingamapofthecontinentalUnitedStatesonascaleoftwohundredmilestotheinch.Hemaythendrawamapofanyoneofthestatesonsomeotherscale:forexample,Connecticutattenmilestotheinch,Wisconsinfortymilestotheinch,andTexaseightymilestotheinch.TheConnecticutmapcanthusincludemoreinformationconcerningroadsandtownsthancantheTexasInap,buteithermapmaybeblownupfurther,toincludemoredetails.Suchmapsmayalsobesupplementedbymapsofthemainarteriesofthemajorcitiesthatliewithinthestate,andthesemapsmaybesupplementedbystreetmapsthatshownotonlythemainarteriesbuteachcitystreet.Further,theDepartmentofPublicWorksinacitycanusuallysupplymapsidentifyingbuildingsandutilitiesoneachblockofeachstreet,and,intheend,architecturalblueprintsmightbefoundshowingthefloorplansofthesebuildingsandeventhedetailsofwiringandcarpentryineachoftheroomsinthebuildings.ItisinthissensethatIwouldcharacterizethepossiblemapsofaterritoryasbeingindefinitelydense.Similarly,historicalaccountsofaseriesofeventsn1ay,inmanycases,formadenseseries,withthevarious 16HISTORYANDITSMODEStemporalandgeographicalsegmentsofanation'shistorybeingcapableofbeingexploredinfurtherdetail,insofarastherequisitedataareavailable.Theapplicabilityofthissimiletohistoricalstudiesshouldbeobvious.Somehistoriesareconcernedwithlongerstretchesoftimethanareothers;also,theregionalaswellasthetemporalscopeofdifferenthistoriesvaries,somehistoriansbeingconcernedwiththepoliticallifeofanationasawhole,whereasothersfocusonwhatoccurredinthepoliticsofoneregion,orwithpoliticalchangeswithinasinglecity.Tobesure,mapsdrawnondifferentscalesdonotpresentuswithrepresentationsoftheverysanleitems:Thedetailsshownonacitymaparenotrepresentedonanationalmap,nor-ontheotherhand-canacitymapgiveindicationsastowhatroadoneshouldchooseinleavingthatcityinordertofindthebestrouteacrossthecountry.Nevertheless,mapsdrawnondifferentscales,iftheyareaccurate,willfittogether,aIIowingustomovetowardanincreasinglysmalltarget,ortheywillallowustofindourwayfromanyparticularstartingpointtootherpointsthatareextremelyremote.Intherealmofhistoricalstudies,onecanalsomoveineitherdirection:frombroadhorizonstomoreaccurateanddetailedvision,orfromanacquaintancewithwhatismostfamiliartoentirelyunfamiliarterritory.Ineachcase,thehistorianwilleitherfollowconnectionsthathisowninvestigationsuncover,orhewillbehelpedbyconnectionsthatotherhistorians,workingonadifferentscale,mayhavelaidoutforhim.Itisinthissensethathistoricalstudiesusingdifferentscalesreinforceoneanother,eventhoughthedetailstheydepictwillnecessarilybedifferent-andwillbedifferentpreciselybecausedifferentscalesareused.TheinterlockingofhistoricaldataisalsotobefoundinthefactthatdifferenthistoriansdealwithwhatIshalltermdifferentfacetsofthesameevents.Toillustratethisaspectofhistoriographicalpractice,oneneedlnerelybearinlnindthatanystretchinthelifeofamodernsocietycanbeviewedintermsofpoliticalchanges,economicchanges,theinternationalrelationsofthatsociety,orthelike.Tobesure,ahistoriandealingwithanyoneofthesefacetswillalmostcertainlybecaIIedupontomakereferencetoeventsthathistoriansapproachingthesocietyfromanotherpointofviewwillalsobeforcedtomentionorinvestigate.Nonetheless,havingapproachedasocietyfromonepointofviewratherthananother,ahistorianwillnotbeobligedtotracealloftheconnectionsthatareofimportancetothehistorianwhoapproachesthesamesocietyfromanother UNITYANDDIVERSITYINHISTORICALSTUDIESangle.Onceagainacomparisonwithmapmakingmaybehelpful.Thesamegeographicalterritory-say,thatoftheUnitedStatesmaybemappedintermsofitspoliticalsubdivisionsor,alternatively,intermsofitsphysiographicfeatures;inthelattercasethernapneednotrevealstateboundaries,butitwillshowdifferencesinaltitudesinvariousareasofthecountry,whichapoliticalmapwillnotdo.OthermapsoftheUnitedStateswillindicateoursystemsoftransportation,butwillnotindicatethealtitudesofvariousregionsandneednotindicatestateboundaries"Inshort,mapsdrawnonthesamescaleandhavingreferencetothesameregionmaybeconcernedwithverydifferentaspectsoftheterritorytheymap.Yetthesemapsfittogether.Oneachofthemaps,theGreatLakesandtheGreatSaltLakemustbeshowninthesamerelationshiptooneanother,andmustbearthesamerelationstootherpoints,suchasCapeCodandSanFranciscoBay.Ifthesegeographicalfeaturesofthesameterritoryarenotdepictedinaconsistentmanner,wedemandthatoneoranotherofthemapsberevised.Andso,too,inhistory.Tobesure,whenhistoriansdepictdifferentfacetsofasocietycantroversiesfrequentlyariseastowhichoftheinte�lockingfactorswithwhichtheydealistobeconsideredthemoreimportant:forexample,whetherinthecaseofconcurrentchangesintheeconolnicandthepoliticalaspectsoflifeinasocietyoneofthesecanbeheldtoberesponsiblefortheother,orwhetherbothr.esultfromstrainsduetoSaInelong-standingdysfunctionintheorganizationofthesociety'slife.Thus,itmustnotbeassumedthatinhistorytherelationshipsbetweendifferentfacetsofasocietyareinertandstatic,asisthecaseingeography.Weshalllaterbeforcedtodealwiththeseandsilnilarproblenlsofcausationinhistory.Weshallalsohavetoraisethequestionofwhethertheadoptionofsomegeneraltheory,suchastheMarxisttheoryofsocialstructureandsocialchange,makesitimpossibletofindafitbetweenhistoricalaccountsthatdealwithdifferentfacetsofsocietallife.Forourpresentpurposes,however,whereweareconcernedwithunderstandingtheunityanddiversityinhistoricalstudies,itisnecessaryonlytotakenoteofthefactthathistoriansdealingwithoneandthesamesocietyoftendealwithdifferentfacetsofthatsociety,aswellasdealwithitondifferentscales,andforthatreasonthereisgreatdiversity-andtherewillalwaysbegreatdiversity-inthewaysinwhichthesamesocietywillbedescribed.Thishasbecomeespeciallynoticeablesincethehegemonyofpoliticalhistoryhasbeenbroken.Tobesure,asOttoBrunnerpointedout,history-as-a-discipline-thatis,generalhistory-alwayshasthe HISTORYANDITSMODESorganizationofsociety,andthereforerelationshipsofpoweramongthemembersofasociety,intheforefrontofattention.This,however,doesnotentailthatitisprimarilyconcernedwithpoliticallifeinanynarrowsense:Therelationshipsofpowerinmanysocietiesdonotturnonsuchinstitutionsaskingship,orontherelationsbetweenchurchandstate,oronrepresentativeformsofgovernment,aspoliticalhistorianswereinclinedtohold.Ithasbecomeincreasinglyclearthatinordertounderstandmanyofthemajorchangesinasocietyforexample,thoseintheUnitedStatesaftertheCivilWar-itiswhollyinadequatetofocusattentiononthecourseofpoliticaleventsalone.Inaddition,knowledgeofthesuccessivewavesofimmigrationisneeded,andso,too,isknowledgeofthedevelopmentoftechnologyanditsapplicationtothemanufactureanddistributionofgoods.Thesefacetsofourhistoryhave,ofcourse,beencloselylinked.TherealsoaremanyotheraspectsofAmericanlifethathistoriansrecognizeasimportantandfinditnecessarytotakeintoaccount.Forthesakeofincreasingaccuracy,checkingwhatmightotherwisebetoofacileoverallgeneralizations,allofthesestudiesmustbepursuedinconsiderabledetail.Thus,withinthescopeofwhatIhavetermed"generalhistory"wefindhistoricalaccounts(writtenonverydifferentscales)concerningourpolitics,economicorganization,foreignaffairs,changesintheformsofourfamilylifeandinthenatureofoureducationalandreligiousinstitutions,andtherolesplayedatvarioustimesbydistinctionsofwealth,ethnicorigin,andrace.Consequentlythereisboundlessvarietyinwhatisinvestigatedinanattempttounderstandthenatureofasinglesocietyandthechangesthatoccurredinthatsocietywithinalimitedperiodoftime.Andhistoriansarenot,ofcourse,interestedonlyinsomeparticularsociety,butdealwiththenatureandtheinterrelationshipsofmanydifferentsocieties.GiventhisvarietyinthematterswithwhichthosewhomIhavetermedgeneralhistoriansareconcerned,whatfurtherareasofinvestigationsareopenforthosewhopursuespecialhistoricalstudies,thatis,whoareconcernedwiththespecialFachwissenschaftenasOttoBrunnerdesignatedthem?Itiswiththesestudies,andwiththeirdifferencesfromgeneralhistory,thatthefollowingsectionisprimarilyconcerned.IIIItisnotusualtoregardhistoriansofliterature,ofscience,ofpainting,orofphilosophyas"historians,"eventhoughitcannotbe UNITYANDDIVERSITYINHISTORICALSTUDIESdeniedthattheyareconcernedwithestablishinganddelineatinghistoricalconnectionsinthefieldswithinwhichtheywork.Tobesure,whattheyestablishoftenthrowslightonvariousaspectsofwhatIhavetermedgeneralhistory;itisalsooftenthecasethatanawarenessoftherelationshipswithinsocietieswithwhichgeneralhistoriansareconcernedisofcrucialimportancetothesespecialhistories.Weshallshortlyhavemoretosayconcerningtheseinterlockingrelationships,butitisfirstnecessarytoturnourattentiontothedifferencesbetweenthekindsofinquiryinvolvedingeneralhistoryandtheproceduresfollowedbythosewhowritespecialhistories,suchashistoriesofanationalliterature,astyleofarchitecture,oraperiodinthehistoryofscienceorphilosophy.Onesuchdifferencerelatestothefactthatinwritinganaccountofsomeperiodinthelifeofasociety-say,ofFranceinthereignofLouisXIV-oneisdealingwithwhatcontinuouslyexistedinaparticularregionoverastretchoftime;insofarastheirmaterialspermitthemtodoso,thereisinprinciplenolimittothedetailintowhichhistorianscanenterinexploringthenatureofasocietyanditschanges.Differenthistoriansmayassigndifferentdegreesofimportancetodifferentaspectsofaparticularsociety,butevenasitsinstitutionschangetherearecontinuousongoingconnectionsamongthem.Forexample,thoughFrancenotonlychangedduringthereignofLouisXIVbuthaschangedevenmoreradicallysince,duetomajorfluctuationinitspoliticalstructuresandcontinuingchangesinitseconomicorganization,therehasbeen,throughtime,apeopleoccupyingagiventerritory,withpeoplesofotherlanguages,traditions;andformsof.socialorganizationaroundthem;howevermuchthesocialorganizationofFrenchlifehaschanged,Francehasremainedthroughoutthelastcenturiesasocietydifferentfromthosebywhichithasbeensurrounded,possessingadegreeofcontinuityandaunityofitsown.14IncontrasttothecontinuingexistenceofasocietysuchasFrance,withwhichgeneralhistoryisconcerned,letusconsidertheobjectofsomespecialhistory,suchasahistoryofFrenchliterature.Thehistorianofanationalliterature,oroftheliteratureofaperiod,isdealingwithacollectionofworksthatmayberelatedtooneanotherinavarietyofways,butindealingwiththiscollectionofseparateworks,heisnotdealingwithanythingthatconstitutesasinglefunctioningwhole.Thereare,sotospeak,gapsbetween·theseworks:Unliketheelementsinthelifeofasociety,theyarenotcontinuous,forminganindefinitelydenseserieswhich.canbeexploredineverincreasingdetail.Onecan,forexample,imaginethemainoutlinesof 20HISTORYANDITSMODESFrenchliteratureashavingremainedthesameevenhadsomeindividualplaysorpoemsornovelsnotbeenwritten,orevenifthereexistedothersofwhichalltraceshavenowbeenlost.Tobesure,influencesexist,andthespecialhistorianoftensearchesforaconnectinglinkbetweenworksthatseemsimilarandyetpuzzlinglydifferent,butthereisnoguaranteethatanysuchlinkeverexisted:Genuineinnovation,aswellasinfluences,mustbetakenintoaccount.Whiletherearehigh-watermarksandmajorturningpointsinanation'sliterature,thatliteraturedoesnotconstituteasinglecontinuingevent,asdoesthenationitself.Thisistruenotonlyofliteratureandthearts,butoftechnology.andofcustoms.ItisthesamedifferencetowhichIhavealreadycalledattention,betweenhabitsandartifactsthatindividualslearnandcreateinthecourseoftheirlives,andwhichIdesignatetheelementsofaculture,andtheinstitutionalpatternsinwhichtheseactivitiesarecarriedon.AsIpointedout,culturalhabitsmaymigratefromsocietytosociety;theyarenotrootedinanyoneformoflife,anddifferentindividualsinthesamesocietymaynotpossessacommonculture.Inunderstandingtheelementsinaculture,howtheyarise,change,andspread,wemustthereforenotlooktotheirinstitutionalbasisonly,butmusttraceparticularinfluences,allowforinnovation,modification,andchangesinuse.Itisinthiswaythathistoriansofthespecialdisciplines,andhistoriansofcustonlSandtechnologywhoarenotdominatedbyaninstitutionalbias,willproceed.ThedifferencebetweenthosewhohavewhatIwouldtermaninstitutionalbias,attemptingtoexplaintheelementsinaculturesolelyintermsofthesocietyinwhichtheseelementsarefound,andthosewhotreattheseelementsinpartialindependenceofthatsocietycanbestbesuggestedbyconsideringtherelationshipbetweenanauthorandhisworks.Onecannotunderstandaliteraryworksimplyintermsofitsauthor'scharacterandlife;onelllustalsotakeintoaccountthetraditionsofhiscraftthatheabsorbed-oragainstwhichherebelled-thereceptionaccordedhisownearlierwork,andtherelationshipsinwhichhestoodtootherauthorsandtotheaudienceshesought.Thus,thereisahistoryofanauthor'sworkwhichalthoughanintimatepartofhispersonalhistory-demandstreatmentinitsownrightifwearetounderstandthecharacteristicsofwhatheproduced.Similarly,tounderstanddevelopmentsinthescienceofaperiod,orchangesinarchitecturalstyle,weneedspecialhistoriesandcannotviewtheseformsofactivitysimplyintermsofchangesoccurringinthesocietiesinwhichtheywereproduced.Itiswiththe UNITYANDDIVERSITYINHISTORICALSTUDIES21semiautononloushistoriesofthevariousaspectsofculturallifethatspecialhistoricaldiscipIines,suchasIiteraryhistories,historiesofscience,andhistoriesoftheology,aswellashistoriesoflanguageandofcustoms,areconcerned.TosaythatthesehistoriesaresemiautonomousshouldcallattentiontothefactthateveniftheyarenottoberegardedmerelyasfacetsoftheinstitutionalIifeofasociety,theyfrequentlyaredeeplyinfluencedbythenatureofthatlifeandthechangestakingplaceinit.Onecannot,forexample,understandsomechangesthattakeplaceintheartswithoutfindinginthemreflectionsofhowpeoplereactedtoeventsthatoccurredintheirlives;norcanonealwaysunderstandchangesinthepositionofartistsorofscientistsintheirsocietieswithoutunderstandingchangesintheinstitutionalstructureofthosesocieties.Thus,thespecialhistorianoftenneedsbasichelpfromthegeneralhistorianifheistounderstandthematerialswithwhichitishistasktodeal;andthegeneralhistorianmayderivehelpfromthespecialhistorianincomingtounderstandhowpeopleviewedtheeventsoftheirtime,andhow,forexample,thedevelopmentofscienceandtechnologypavedthewayformajorformsofeconomicchange.Thisinterlockingofdifferentformsofhistoricalinquiryneedbenomoresurprisingthanthemannerinwhichinquiriesintodifferentfacetsoftheinstitutionallifeofasocietyoffereachothercorroborationandlllutualsupport.Whiletheseconnectionsbetweengeneralhistoryandspecialhistoriesshouldnotbeunderestimated,itremainstrue-asIhaveremarked-thathistoriansofliterature,ofpainting,ofscience,orofphilosophyarenotusuallyregardedas"historians,"andtheiracademicpostsmoreoftenthannotareoutsidedepartmentsofhistory.Itisnotdifficulttounderstandwhythisshouldbethecase.Ahistorianofliterature,forexample,willalmostalwayshavetofunctionasacriticaswellasahistorian,andhemaythereforefindamorenaturalplaceforhimselfamongthosewhoareconcernedwiththepracticeandtheoryofcriticism,withhermeneutics,stylistics,andlinguisticsthanamongthosewhoseinterestsaremorepurelyhistorical,focusingprimarilyonthenatureandchangeswithinasociety.Similarly,thehistorianofphilosophymusthaveaconcernwithphilosophicproblemsfortheirownsake,andtobeahistorianofartdemandsaestheticsensitivityandsomedegreeofconnoisseurship.Furthermore,asIhavepointedout,agreatdealoftheworkofthosedealinginspecialhistorieshastodowithcomparativestudies,withstudiesofinfluence,andwiththemigrationofculturaltraditionsover 22HISTORYANDITSMODEStime;thus,thefocusoftheirattentionwillnotbelikethatofgeneralhistorians,whoseaimitistounderstandthenatureandthechangesofparticularsocieties.Tobesure,therehasbeenatendencyamongmanyculturalhistorians,andamongsomeinstitutionalhistorians,toobliterateanyultimatedistinctionbetweengeneralhistoriesandspecialhistories;theydosobecausetheyassumethatwithinanysociety,orinanyage,thereisanoverridingunitythatembracesallaspectsofsociallife.Suchaunityissometimesheldtobebasedonthedominanceofsomeformofinstitutionalstructure,ashassometimesbeenheldbyMarxists;sometimesithasbeenidentifiedwithapervasiveZeitgeist;atothertimesitismoremodestlyclaimedthattherearecommonintellectualpresuppositionsandcommonformsofsensibilitythatunderlietraitscommontothesocialinstitutionsandtotheculturalproductsofanage.Anyoneofthesemonistictendencieswilldomuchtoobliteratethedegreeofindependencethat,Ibelieve,mustbepreservedinvariousfieldsofhistoricalinquiry.Thiscanmostclearlybeshownbyconsideringwhatisinevitablyinvolvedinthewaysinwhich,forvariouspurposes,weperiodizehistory.15Anyperiodizationofhistorydemandsthatweselectsomeaspectofthelifeofasociety,orsomeaspectofculturallifethatweregardasimportant,asthebasisformarkingthebeginning,middle,andendoftheperiodwithwhichweareconcerned.Onecannotassume,however,thatwhatmarksthebeginningofaperiodwhenseenfromapoliticalpointofviewalsomarksthebeginningofaneweconomicperiod,northataperiodizationintermsofsomeelementofculture,suchaspaintingorliterature,willbesynchronouswithaperiodizationbasedonscientificdiscoveriesoronphilosophicinnovation.Thisisnottosaythatitismisleadingtoofferperiodizationsofhistory,iftheprincipleofperiodizationusedismadeexplicit.Whatistoberejectedistheall-or-noneapproachofmonisticviewsofhistory,inwhichperiodizationsaretakentobeequallyapplicabletoallaspectsofasocietyanditsculture.Howgreatadistortionthiscaninvolveismoststrikinglyillustratedbythewayinwhichtheconceptof"theRenaissance"hassometimesbeenused,evenwhenthatconcepthasbeenappliedonlywithinItaly.Inliterature,twomajorItalianauthorsoftheRenaissance,PetrarchandBoccaccio,diedin1374and1375;buttheRenaissancepainterRaphaeldied146yearslater.Thus,theperioddesignatedas"theRenaissance"mustbedifferentlydatedifoneisconcernedwithpaintingratherthanwithliterature.Similarly,wereonetoviewthenewscienceofthesixteenthand UNITYANDDIVERSITYINHISTORICALSTUDIESseventeenthcenturiesasaRenaissancephenomenon,asissometimesdone,thelimitsoftheperiodwouldhavetobegreatlyextended,sinceGalileodied122yearsafterRaphael.Infact,thestretchoftimefromPetrarch'sbirthtothedeathofGalileoisapproximatelyequaltothatfromGalileo'sdeathtothepresentday.Itiscertainlyimplausibletoassumethattherewasoneandthesamespiritsuccessivelydevelopingindifferentareasofculturallifeoverthatlengthoftime.Infact,noinvestigationsoftheactualworksproducedindifferentfieldsduringthatperiodwouldsupportthenotionthatitislegitimatetoregard"theRenaissance"monistically,asifitwereaunitaryphenomenonpervadingallaspectsofItalianculturallifefromthefourteenthtotheseventeenthcentury.Norisitonlywithrespecttoculturalperiodizationsthatsuchasituationobtains.Ifonechoosesdifferentinstitutionalfacetsofasociety,onefindsthateventhoughtheyareintimatelyconnected,oneormoreofthemmayhavefailedtochangeatthesameacceleratingrateasothers.Insuchcases,ofcourse,majordysfunctionsmayhavearisen;butevenifitisassumedthatthesewillintilllecorrectthemselves(associologicalandanthropologicalfunctionaliststendtoassume),thehistorianwhowishestolookatwhatactuallyoccurredwithinaparticularperiodwillhavetotakesuchdysfunctionsintoaccount.Intaking"thelongview,"thelllonisticallyinclinedhistorianconcealsfromhimselfandhisreaderswhatdidactuallyoccur.Thisholdsaswellforthosewhotake"thebroaderview,"lookingonlyatatotalresult.InaccountingfortheoutcomeofapresidentialelectionintheUnitedStates,itisassuredlynecessarytotakeintoaccountavarietyofinterests,dissatisfactions,antipathies,andenthusiasmswhichcometogethertogivethefinalresult.Yet,thisresultdoesnotnecessarilygiveanaccuratepictureofthebasisforthevotesindifferentregionsofthecountry,noramongdifferentclassesofvoterswithintheseregions;theresultsoftheelectionasawholemaythereforeberelativelyunintelligibleuntilthesefragmentarydataareanalyzedanditisshowninwhatwaystheycontributedtothefinalresult.Whenahistoriantakesthebroadview,orwhenhetakesthelongview,heisinterestedonlyinfinalresults,notintracingthenatureofasocietyandthechangesinitastheseactuallyoccurredovertime.Forahistorian,Isubmit,thisisacontradictioinadjecto,forregardlessofthediversitytobefoundinthematterswithwhichdifferenthistoriansdeal,itistheirCOlnmitmenttoidiographicconcernsthatleadsustoregardthemashistorians. ChapterTwoVARIETIESOFSTRUCTUREINHISTORICALACCOUNTS..•••)Traditionally,thegeneralformofhistoricalaccountshasbeenthatknownas"narrativehistory."Eventhoughotherformshaveincreasinglyreplacednarrativehistories,anumberofphilosophershaverecentlytakenthenarrativetobethebestmodelforunderstandingthelogicofhistoricalexplanation.AmongexamplesofthisviewonemayciteW.B.Gallie,PhilosophyandtheHistoricalUnderstanding,ArthurC.Danto,AnalyticalPhilosophyofHistory,andMortonWhite,FoundationsofHistoricalKnowledge;earlierformsofasimilartendencyaretobefoundinW.H.Dray'Scontinuousseriesmodelofexplanation,inW.H.Walsh'saccountofhistoryas"significantnarrative,"andinsomeaspectsofthetheoriesofCroceandCollingwood.Ifindthistendencyunfortunate,asIhaveelsewhereattemptedtomakeclear.1Whatisofconcerntomehere,however,isnot VARIETIESOFSTRUCTUREINHISTORICALACCOUNTScriticism;instead,Iwishtoofferapositiveanalysisofalternativeformsofstructuretobefoundwithinhistoricalworks.Ineedmerelymaketwopreliminaryremarksconcerningtheconceptofanarrativeasappliedtohistoricalstudies.Inthefirstplace,describinghistoryasnarrativesuggests-andIassumeismeanttosuggest-thathistoriographyistobecomparedwithtellingataleorstory.Thisismisleadingevenwhenappliedtothemosttraditionalhistories.Ahistoriandealingwithanysubjectmattermustfirstattempttodiscoverwhatoccurredinsomesegmentofthepast,andestablishhowtheseoccurrenceswererelatedtooneanother.Oncethisresearchhasbeencarriedforwardtoapartial·conclusion,hemust,ofcourse,thinkabouthowhewillbestpresenthisfindings,andthis,andwhatfollowsuponit,mayberegardedas"constructinganarrative."Suchanarrative,however,isnotindependentofhisantecedentresearch,noristhatresearchmerelyincidentaltoit;thehistorian's"story"-ifonechoosestoviewitmerelyasastory-mustemergefromhisresearchandmustbeassumedtobeateverypointdependentonit.Itisthereforemisleadingtodescribewhathistoriansdoasifthiswerecomparabletowhatismostcharacteristicofthestoryteller'sart:Thebasicstructureofastoryortaleisofthestoryteller'sownchoosing,andwhatevermaybepreliminarytohistellingthatstorydoesnotservetocontroltheactofnarration.2Inthesecondplace,asweshallsee,thedemandsplacedonahistorianbyhissubjectmatterrarelypermithimtofollowanysimplestoryline.Toexplicatethechainofoccurrenceswithwhichhedeals,hemust,inmostcases,provideagreatdealofmaterialconcerningtheantecedentbackgroundoftheseoccurrences,andmustalsopaycloseattentiontomanycontemporaneousevents.Thestoryteller,ontheotherhand,isnotunderthesamenecessity:hestartsinmediaresandheneedmerelyintroducewhateverearlierbackgroundorwhateverreferencestoongoingeventsheregardsasusefulinhighlightinghisstory.Therefore,narrativestendtohaveasimpler,morelinear,andmoreself-containedstructurethandohistoricalaccounts,evenwhenthehistorianissequentiallytracingthechangesthatoccurredovertimewithrespecttothecentralsubjectwithwhichhisaccountisconcerned.Ishallthereforespeakofthestructureofthemoretraditionalhistoricalaccounts'assequentialincharacter,andavoidreferringtosuchworksasnarratives.AsIhaveremarked,thistraditionalstructurehasbeenincreasinglyreplacedbyotherwaysoforganizinghistoricalworks,butthesequentialformremainsonebasicpatternofwhichwemusttakecog- 26HISTORYANDITSMODESnizance.TherearetwoothersuchformswithwhichIshalldealinthepresentchapter:theexplanatoryandtheinterpretive.Whatmustbeemphasizedattheoutsetisthatthesearenot"pure"forms:Nohistoricalworkusesonetotheexclusionoftheothers.Asweshallsee,therewillbepointsatwhicheachofthethreemodescomesintoplayinanyhistoricalwork;however,theoverallstructureofanyhistoricalstudyislikelytoconformmorecloselytooneoftheseformsthantoeitheroftheothers.ThefactthatIshallnotattempttoisolateanddealwithanyotherformsdoesnotindicatethatIbelievemyclassificationtobeexhaustive.Ido,however,regardthesethreetypesofstructureasmorepervasivethananyothersonemightdistinguish.IForpurposesofexposition,Ishallfirstbrieflyconsidertheexplanatorystructure,whichdominatessomehistoricalaccountsandwhichcanbeofimportanceatalmostanypointinanyformofhistoricalinquiry.VhatIshalltermanexplanatorystructureispresentonlywhenaperson-inthiscaseahistorian-alreadyknows(orbelievesthatheknows)whathasinfacthappened,andseeksanexplanationofwhyithappened.Insuchacasehestartsfroillafacttakenaspresentandseekstotracebackitscauses-thatis,toestablishwhatwasresponsibleforitshavinghappened.Howwearetoconceiveofthecause,orthecauses,ofanoccurrenceisaproblemthatwilloccupyusinthesecondpartofthisbook.Whatishereofimportanceiswhatmaybecalledthedirectionofinquiryinanexplanatoryaccount.Speakinggenerally,inquirystartsfromagivenoutcomeandproceedsinadirectionthatisthereverseofthedirectioninwhichtheeventsresponsibleforthatoutcome,actuallyoccurred;inothervvords,anexplanatoryaccountinvolvesatracingbackofeventsfromthepresenttowardthepast.3Itisherethatthereemergesafirstandfundamentalpointofdifferencebetweenanexplanatorystructureandasequentialstructure,forthelatter-asweshallsee-followseventsintheorderoftheiroccurrence,andthoughtheoutcomeoftheseriesmayinfactalreadybeknownbythehistorian,itneednothavebeenknown.Ontheotherhand,inexplanatoryinquiries,theinquiryitself(althoughnottheaccountthatemergesfromit)movesbackfromwhatisknowntohaveoccurredandseeksanexplanationofitthroughtracingitsantecedents. VARIETIESOFSTRUCTUREINHISTORICALACCOUNTSFurthermore,inanexplanatoryaccountofaparticularoccurrence,theeventswithwhichthehistoriandealsmaybeextremelydiverse,notbelongingtogetherexceptinsofaraseachhappenedtocontributetotheparticularoutcolnethehistorianisinvestigating.Forexample,ifheistoaccountforasuddendeclineinanation'sforeigntrade,heisnotnecessarilyconfinedtoeventsthatoccurredwithinthatnation:Theeruptionofawarinvolvingoneofthenation'smajortradingpartnersandablockadeofthatpartnerbyitsneighborsInightwellbeimportantfactorstobetakenintoaccount.Inaddition,adroughtmayhavecausedadeclineinthenation'sagriculturalexportstoothernations;so,too,mayhaveastrikeofdockworkers.Allsuchfactorsarerelevantforthehistorianwhowishestoexplainwhatoccurred,butitisobviousthattheyareoftenwhollyindependentofoneanother.Thus,anexplanatoryaccountofaparticulareffectdrawsuponfactorseachofwhichhasitsownseparatehistory:Thehistorianwillnotbefollowinganyonecontinuousseriesofevents.Inthisrespect,suchaccountsdifferfromthecharacteristicstructureofsequentialhistories.Inasequentialhistory,thehistorianchoosesasubjectthathasadegreeofcontinuityinitshistory,andheseekstotracethestrandofeventsmakingupthathistory.Consequently,asequentialhistoryseemstopossessasingledominantstoryline-asnarrativesgenerallydo-ratherthanbeingananalysisofindependentfactorsthat,together,bringaboutaparticularresult.However,thecomparisonbetweensequentialhistoricalaccountsandnonhistoricalnarrativesmustnotbeoveremphasized,notonlybecauseofthedifferencestowhichIhavealreadyalludedbutbecause'whatweregardasastoryusuallyleadstoaspecificconclusion,anditiswithrespecttothatconclusionthattheepisodeswereselectedbythepersontellingthatstory.Ahistorian,however,maywriteahistoryofsomestillincompleteseriesofevents,asThucydidesdid:HecommencedhisaccountofthewarbetweentheAtheniansandthePeloponnesianswhentheyfirsttookuparms.Inthatcasethehistorianfollowsoccurrencesastheyoccur,notknowingwheretheywilleventuate,butkeepinginmindthatheistoincludeonlythoseeventsthatseemtohaveadirectbearingontheparticularsubjectwithwhichhishistoryisconcerned.Hissubjectmaybeapersonoranation,achanginginstitutionoraculturalmovement;inanyofthesecasesasequentialhistoryhasaunifyingtheme.Whenhistorianslookbackuponacompletedlifeortheendofanerainanation'shistory,oruponanearlierculturalmovementorasetofinstitutionalchanges,itisofcourseeasiertosee HISTORYANDITSMODESwhatwasrelevantthanwhentheyaredealingwithwhatisstillanongoingprocess;yet,eveninthesecases,thestructureofasequentialaccountmaybequiteunliketheusualstructureofanarrative-thereignofakingmayhavecometoaclosewithoutcomingtoaclimax,oraninstitutionmayhavechangeditsformsanditsfunctionswithoutourviewingeachofitsearlierchangesaspreparatorytowhatitlaterbecame.Inshort,inmuchofhistory,ifitbenarrative,itisnarrativesorelylackinginpoint.Thisfact,however,isapttobeconcealedbycasesinwhichthehistorianhaschosentodepictsomesequenceofeventsinwhichwearefascinatedbythemannerinwhich-asinastory-thepartsformthemselvesintoasingleunifiedwhole.Wearenowinapositiontoseetheessentialdifferencesbetweenanexplanatoryandasequentialapproachtothematerialswithwhichhistoriansareconcerned.Forthesequentialhistorianthereisaparticularseriesofeventshewishestofollow,seeingonegrowoutofanotherandobservinghowothereventsalteredtheirprogress,buthowintheendtheycametoformasinglehistorywhosecourseheisabletorenderintelligible.Inthis,thedirectionoftheflowofeventsandthedirectionofhisexplanationcoincide.Tobesure,atsomepointsinthisflowhemaybepuzzled,andhewillthenhavetopausetogiveanexplanatoryaccountastowhy,atthispoint,aneventofaparticularsortunexpectedlyoccurred;ingeneral,however,asequentialaccountwillfollowtheformofDray'scontinuousseriesmodelofexplanation-oneeventledtoanother,andittothenext,andsoontotheendoftheseries.Ontheotherhand,anexplanatoryaccountdoesnotsetouttogivethereaderasequentialviewofwhatoccurred,butseekstoansweradefinitequestioFl:Grantedthatthiseventdidoccur,whatfactorswereresponsibleforitsoccurrence?Sometimestracingacantinuousseriesofeventsprovidesanexplanationwithwhichwearesatisfied,buthistoricalanalysesoftentakeanotherform,tracingbackaconfluenceofotherwiseunrelatedeventsandindicatinghow,atsuccessivemoments,theyinteracted.ThedifferencebetweentheseapproachesissometimesasgreatasthedifferencebetweenexplainingthesizeofthepopulationoftheUnitedStatesin1976bytracingacurveofgrowthandaccountingfortherateofgrowthinanyparticulardecade.Wemustnowtumtothethirdformofstructureinhistoricalworks,theinterpretive.Perhapsthesimplestwayofindicatinghowinterpretiveworksdifferfromthosewhosestructureisprimarilysequentialorprimarilyexplanatoryistoconsiderthetaskfacedby VARIETIESOFSTRUCTUREINHISTORICALACCOUNTSalmostanyhistorianasheintroduceshisreadertothesubjectwithwhichheistobeconcerned.Inthatintroductionhemustportraythestateofaffairsexistingatthetimeandplaceatwhichhisaccountbegins,insofarasitsaspectsarerelevanttothatwithwhichheistodeal.Otheraspectsofwhatthenexisted,evenifknown,arenotofconcern:Whatisneededistopresentabackgroundagainstwhichfuturedevelopmentscanbeunderstood.Theportrayalofthesematerialscannotconsistinsimplylistingthem,sincetherelationsexistingamongthemwillbeimportanttorelationshipsamongtheeventsthatlateremergedfromthem.Thus,attheoutsetofanyhistoricalworkanattemptismadetodepictaparticularstateofaffairs,notinalldetailbutintermsofwhatwasmostsignificantinitsstructurewithrespecttothelatereventswithwhichthehistorianintendstodeal.Now,itispossibleforahistoriantoenteruponasimilartasknotsimplyforthesakeofpresentingbackgroundforunderstandingfurtherevents,butforthesakeofdepictingthatstateofaffairsitself.This,Itakeit,ispreciselywhathistoricalaccountsthatareprimarilyinterpretiveinstructureareintendedtodo.Tobesure,theyareunlikelytoberestrictedtodepictingthestructureofastateofaffairsmerelyatsomemOlnentintime;theywillbeinclinedinsteadtotreatofsuchstructuresasenduringinacontinuingformoversomedefinitespan.Forexample,G.M.YounginhisessayVictorianEn,gland:PortraitofanAgedealtwithEnglandfrom1831to1865;Burckhardt'sCultureoftheRenaissanceinItalyis,ofcourse,aninterpretivehistoryofevengreatersweep.Sincehistoricalstructurestosomeextentchangeovertime,theinterpretivehistorianwillgenerallyalsobeconcernedwithunderstandinganddepictingthesechanges.Thus,inportionsofhisworkhewilladoptsequentialmodesoftreatinghissubjectmatter.Further,hewillatvariouspointsattempttoexplainsuchchanges,andthusfromtimetotimehisaccountwillproceedmuchasexplanatoryaccountsproceed.Nonetheless,ahistorianmayenvisionhismaintaskasthatofrevealingthecharacteristicfeatureofsomeformoflife,ratherthansequentiallytracingorexplainingthevariousoccurrencesthatenterintohisaccount.ThismotivationisevidentintheworksofYoungandofBurckhardttowhichIhavealluded.4Itiseasytoidentifysuchworksasexamplesofinterpretivehistories,butthegenreisfarmoreinclusivethantheseillustrationssuggest.Forexalnple,ifonetakesRichardPares'swell-knownFordLectures,KingGeorgeIIIandthePoliticians,onefindsanattemptto HISTORYANDITSMODESportray"therulinginterestsandmotivesofBritishpoliticsinGeorgeIll'sreign;toexploretheking'suncertainandundefinedreIationswiththeHouseofCommons;andtoillustratetheconflictswhichthisuncertaintyandwantofdefinitionproducedfromtimetotime."5WhilePares'sdescriptioninvolvedelementsofexplanatoryanalysisandalsosometracingofsequentialconnections,hisemphasisthroughouthislectureslayontheinterpretationofrelationshipsamongindividualsandgroupswhoheldandsharedpower.Noristhisanisolatedexample:Almostanycross-sectionalstudyofpoliticalorsociallifeisboundtohaveastrongbiastowardaninterpretivestructure,evenif,later,explanationsareofferedastohowtherelationshipsamongthevarioussegmentsofthecommunitycametobeastheywere.Thus,interpretivestudiescannotbeconfinedtoastudyoftheinterplayofelementsinthecultureofapeople;theyoftendealwithinstitutionalrelationshipsaswell.Ofthisweshallhavemoretosaylater.IIItisnowtimetoturnfrom·thisfirstgeneralcharacterizationofthethreetypesofhistoricalstructuretoacloserconsiderationoftheirinterplay.Ishallfirstaddressmyselftotherelationsthatarelikelytoexistbetweensequentialandexplanatoryaccounts,andinthisconnectionweshallseeinfurtherdetailwhythecurrenttendencytoidentifyhistorywithnarrativeisbasicallymisleading,especiallyinsofarasgeneralhistoryisconcerned.Ingeneralhistories,thenatureandchangesofasocialorderarewhatserveasthefocusofahistorian'sinterest,andanysuchorderinvolvesacOInplexnetworkofrelationshipsevenwhereahistorianmaywishtoconfinehimselftoonlyonefacetofthesociety-say,toitspoliticalaspects.Inordertodepictthepoliticallifeofasocietyonecannotfollowanysimplenarrativesequence.Thisisevidentifoneconsidersevenasharplydelimitedsegmentofpoliticallife,suchasasinglepresidentialelectioncampaign.ItwouldbeagrossdistortionofthesubjectInatterifthehistorianweretoviewtheeventsconstitutingthecampaignasasinglelinearseriesinwhicheachstepiscausallyrelatedtoaparticularantecedentanditselfleadstoaspecificconsequent,as(forexample)MortonWhite'sanalysisofthebasicskeletalstructureofhistoricalnarrationwouldhaveusassume.6Usuallythereisanoverallstrategyinapoliticalcampaign, VARIETIESOFSTRUCTUREINHISTORICALACCOUNTSandthatstrategyoriginallydependsuponananalysisoftheestablishedvotinghabitsofdifferentsegmentsofthepopulation,anduponarecognitionofthecurrentinterests,disaffections,andneedsofvariousgeographic,economic,and·ethnicgroups.Thus,tounderstandthestratagemseachpoliticalpartyemploys,onemustgrasptheirrelationshipstolonger-enduringfactorsthatarenotthemselveslinksinthesequentialchainofeventsconstitutingthe"story"ofthecampaign.Moreover,muchthathappensinanationalelectionhappensatdifferenttimesandindifferentpartsofthenation,andtheultimateoutcomeofacampaign,eveninanageofrapid,widespreadcommunication,Inaydependnotuponwhathappensdaybyday,butuponwhereithappened,andbywhomitisknowntohavehappened,andhowitreIatestowhattheoppositionhadalreadyclainled.Inotherwords,anelection,unlikeachessmatch,isnotwonorlostbyaseriesofneatlyarrangedsequentialmovesandcountermoves;anymerelysequentialnarrative,or"campaignstory,"isthereforenotlikelytogiveitsreadersmuchinsightintowhytheelementsinthatstoryoccurredorwhytheyhadtheeffectstheydid.Ratherthanviewinganelectionasalinearsequenceofevents,itismoreaccuratetoviewitasawholemadeupofparts,withanunderstandingofthatwholedependinguponanunderstandingofitsvariouspartsandoftheirreIationstooneanother.Suchawholeis,ofcourse,tenlporal;itisnotawholethatispresentallatonce.SOlneofitspartswillprecedeothers,butmanymaybepresentatthesametime;sonlewilldependuponothers,butmanymaybeindependentofmostoftheothers,nonethelesscontributingsomethingtothewholeofwhicheachisapart.Whatistrueofasingleprocess,suchasanelectioncampaign,ortheenactmentofasinglemeasurebytheCongress,isobviouslytrueofanymajorongoingchangesinthepoliticallifeofanation.Theinterplayofcentersofpoliticalpower,andthedecisionsreached,thesetbacks,therecoupingoflosses,andthefinalresultsachievedcannotbechartedinalinearpatterninwhicheachrelevanteventistobeviewedasadiscretelinkinasinglecontinuouschain.ThereIationsbetweentheeventswithwhichpoliticalhistoriansdealarealwayssetagainstaninstitutionalbackground,withoutwhichtherelationsbetweenthevariouseventscannotbeunderstood.Furthermore,itisrarelythecasethatthepoliticallifeofanationexistsinisolationfromitseconomicandsocialchangesorisunaffectedbyitsrelationstoothernations.Underthesecircumstances,themodelofhistoricalaccountsthatnarrativistsproposeissooversimplifiedastoberadicallymisleading. HISTORYANDITSMODESNevertheless,thereisoneimportantaspectoftheviewofthosewhoadoptthismodelthatcanberetainedevenafteronedropsallotheranalogiesbetweenwritinghistoryandtellingastory:Thisaspectistheclaimthat,indescribingwhatactuallyoccurredandintracingtherelationshipsbetweentheseoccurrences,oneisinfactofferingwhatahistorianwilltaketobeanexplanation.Inotherwords,contrarytowhathasbeenheldbyHempelandothers,thereiswhatFredD.Newmanhastermed"explanationbydescription":Inordertoexplainactualevents,oneneednotinallcasesshowthattheyfollowedfromasetofantecedentconditionsaccordingtosomegenerallaw.7Thisisaconsequenceofwhathasalreadybeensaidintheprecedingparagraph:Anunderstandingofawholemaycomethroughunderstandingitspartsandtheirrelations;andwhiletheserelationsamongthepartsmaysometimeshavetobeexplainedwithreferencetogenerallaws,asweshallseeinchapter5,thisisnotinallinstancesnecessary.Considerwhathasbeensaidofanelectioncampaign:Toexplainitsoutcomewedonotinfactrelatethatoutcometosomeantecedentsetofconditionsbymeansofagenerallaw;weanalyzethecampaignasanongoingprocessinwhichthetacticsemployedbyeachcontendingpartyprovedtobesuccessfulorunsuccessfulwithparticulargroupsofpeopleinparticularlocalities,andwetakeintoaccounthoweventsoverwhichneitherpartymayhavehadcontrolalsoaffectedtheoutcome.Understandingwhatconstitutestheelectoralprocessinaparticularnation(oritssubdivisions)permitsonetoanalyzethatprocessintoitscomponentsandtoofferadescriptionofthesecomponentsandofhowtheyaffectedoneanother;thisistoofferwhathistoriansgenerallywouldtaketobeanexplanationofthecampaign'soutcome.Somehistoriansorpoliticalscientistsorsociologistsmaywishtogeneralizefromsuchanalyses,formulatingagenerallawconcerningrelationshipsbetweenparticulareconomicconditionsorwarsorpsychologicalfactorsandwhathavebeentheresultsofpresidentialelectionsintheUnitedStates;however,asweshallsee,itisdoubtfulwhethertherecanbeanysatisfactorylawsofsequenceofthattype,eveninthephysicalsciences.8Whilethereareindeedgeneralizationsthathistoriansmustoftenuse,thepointtobenotedisthatoneisgivingwhatahistorianwouldcountasanexplanationwhen,withinasequentialhistory,oneisabletofollowthetrainofeventsthatmakeuptheseriesasawhole.Unlesstherelationsbetweentheparticularepisodesremainopaque-whichtheyoftenarenot-tracingthecomponentswithinaprocessdoesyieldan VARIETIESOFSTRUCTUREINHISTORICALACCOUNTS33explanationofwhy,whenthesituationwasasitwasatagiventime,itlaterbecameradicallyaltered.Inshort,asourcontemporarynarrativistsinsist,tocomplete"thestory"istogivetheexplanation;butmypointisthatinahistoricalaccount"thestory"isbyriomeansasimplenarrativestory.Furthermore,Iwishtoinsistthattherelationsbetweentheparticularepisodeswithinahistoricalaccountmayremainopaque:Itmaynotbeintheleastclearhowonewell-authenticatedpartofaprocesswasrelatedtoanother,evenwhenonesuspectsthattheremustbesomerelationbetweenthem.Insuchcasesthesequentialstructureofahistoricalaccountmustbesupplementedbyexplanatoryorinterpretiveanalyses.Consider,forexample,thecaseinwhichitisobviousthatahistorianmustmakementionofadeepeconomicrecessioninordertoexplaintheoutcomeofaparticularelection.Campaignspeeches,newspapereditorials,andvotingpatternsInayyieldabundantevidencethattherewassuchaconnection.Yet,itInaynotbeatallclearfromthedocumentsavailabletothehistorianhowsucharecessionwasitselfconnectedtoanyforegoingevents.Toestablishsuchaconnection,heInayhavetocalluponeconomistswho,bytheuseofwell-establishedgeneralizationsconcerningeconomicprocesses,canexplaintheconnectionbetweentherecessionand·earlierpoliticallymotivatedpolicies.Inordertoapplysuchgeneralizationstotheeventswithwhichheisconcerned,thehistorianmayalsohavetoemploywhatIhavetermedtheinterpretivemethod,whichforceshinltolookforinstitutionalrelationshipsofwhichheInaynotpreviouslyhavebeenaware.Thus,anopaquerelationshipinwhatotherwiseseemstobeastraightforwardsequentialaccountmayleadhistorianstqseeknewfactorsinthesituation,andthebackgroundofearliersequentialaccountswilltherebyhavebecolneInorecomplex.This,ofcourse,isonewayinwhichadisciplinesuchaseconomicshasledtofundamentalchangesinthetraditionalforInsofhistoriography.9Analogousrelationshipsbetweensequentialaccountsandexplanatoryorinterpretiveinquiriesaretobefoundwhenweturnfromgeneralhistoriestothefieldofspecialhistories,suchashistoriesofscienceorliteratureorarchitectureorphilosophy.Inthesefields,however,thesequentialstructureraisesproblemsnotpresentinthecaseofgeneralhistory.Theseproblemsarisebecausethesubjectmatterofaspecialhistory-aswehavenotedinchapterI-issomephaseofculture,notaparticularsocietyoritsinstitutions.Therefore,thespecialhistoriandealswithacollectionofactivitiesandworks, 34HISTORYANDITSMODESratherthanwithanyentitythathasacontinuousexistence.Nevertheless,whenonereadsahistoryofFrenchliterature,ofGothicarchitecture,orofchemistry,theliteraryworks,thebuildings,orthediscoveriesthatmakeupthehistoryformarelatedseries;theyarenotmerelyrandomcollections,sequentiallyarranged.Obviously,thisisbecausethehistorianhasarrangedthemashehas;butwhy,onemayask,shouldthatarrangementhavebeenchosen?Theanswerthatmayfirstcometomindisthatonecanseehoweachoftheseworksmayhaveinfluencedsomeofthesucceedingones;thus,theseriesisformedthroughtheskeinsofinfluencethehistorianfollows.This,however,leavesoutofconsiderationamorebasicfactor:thedefinitionofhissubjectmatterthatledthehistoriantohisselectionofthosematerialstowhichprimaryattentionwastobedevoted.Asweshallsee,onlyafterhissubjectmatterhasthusbeendelimiteddoesthetracingofinfluencesbecomeimportantinestablishingthecontinuitytobefoundinanyspecialhistory.Onecanappreciatetheimportanceofadefinitionofthenatureofaculturalelementwhenonenotes,forexample,thatdifferentliteraryhistoriansdefine"literature"indifferentways.Forsome,itincludesonlyfiction,drama,poetry,essays,andjournals,whereasotherswouldalsoincludesermonsandcorrespondence,andperhapshistorical,scientific,orphilosophicalwritingsviewedfromthepointofviewoftheirrelationstootherliteraryforms.1°Furthermore,whetheroneusesabroadoranarrowdefinitionof"literature,"somewouldwishtoincludeonlyworksconsideredtobeofhighliterarymeritduringtheirauthor'sowntimeorsubsequently,whereasotherswouldincludepopularfiction,popularverse,widelydistributedpoliticaltracts,andsoforth,takingintoaccountbreadthofdistributionandgeneralinterest,aswellasliterarymerit.Similarconsiderationsapply,mutatismutandis,tohistoriesofarchitecture,ofscienceandtechnology,philosophy,oranyotherspecialhistory.Giventhissituation,itisnotstrandsofinfluence,butthehistorian'sconceptionofhissubjectmatterthat,inthefirstinstance,dictatestheprinciplesofinclusionofcertainmaterialsinhisworkandgovernstheexclusionofothers.Wefind,ofcourse,thatliteraryhistorians(forexample)donotdealwiththewholeofpastliterature,butsetthemselvesmorespecifictasks:Theymaydealwiththeworksofasingleauthororofaschool,withastyleoragenre,withtheliteratureofaperiodorofanation.Yet,whatisincludedwithintheirmorespecifictopicdependsuponwhattheyarewillingtoregardasliterature,andalsouponhow,forexample,theydelimittheperiodorthegenretheywishtoinves- VARIETIESOFSTRUCTUREINHISTORICALACCOUNTS35tigate.Thisprimarydelimitationofthematerialswith,whichaliteraryhistoriandealsisanalogoustothefactthatthematerialsthatfallwithinthescopeofageneralhistorian'sinterestare,inthefirstinstance,determinedbythesubjecthehaschosen,bytheaspectoraspectsofitwithwhichheisconcerned,andbythescaleonwhichhewishestopursuehisinquiry.Thereis,however,apointatwhichthisanalogybreaksdown.Aswehavenoted"thegeneralhistoriandealswithsomecontinuingsocietalstructure,analyzingitspartsandfollowingitschanges,whereasaliteraryhistorian,orthehistorianofanyotherbranchofculture,dealswithaseriesofworksthat,althoughtheyareseenasbelongingtogether,arenotpartsofanyactualsinglewhole.Itthenbecomesnecessarytoraisethequestionofhowaliteraryhistoriancanjustifythefactthatheincludescertainworksandexcludesotherswhenhecomestowritehissequentialaccountof,say,anation'sliterature,orofagenresuchasthenovel.Whilehisdefinitionofliteratureorofthenovelprovidesaninitialdelimitationofhismaterial,therewillundoubtedlybemanyexamplesthatfithisdefinitionwithwhichhewillnotdeal.Unlikeageneralhistorian,hecannotthensay,"ThesedonotbelongtothesubjectwithwhichIanldealing,"forthatwithwhichheisdealingisaclassofobjects,notacantinuingwholeanditsparts.Atthispointtheliteraryhistorian,oranyotherspecialhistorian,islikelytoinvoketheconceptof"importance":Hewillattempttoincludeallimporta'ntexamplesoftheclassofworkswithwhichheisdealing,butwillbewillingtoexcludethosehebelieveslackinginimportance.11Theterm"importance,"however,concealswithinitselfanumberofmeanings.Inthepresentcontext,forexample,animportantworkmaybeoneofoutstandingliteraryvalue;oritmaybeonethathadasignificantinfluenceonotherliteraryworksbyvirtueofitsthemeoritstechnique;oritmayberegardedasimportantbecauseitinfluencedsocialandpoliticalthought,ormoralandreligiousbelief,andwouldthereforebeincludedbythoseliteraryhistorianswhoadoptoneofthebroaderdefinitionsofliterature.Ofthesealternatives-foreachofwhichmanyexamplescanreadilybecited-itisonlythefirstmeaningof"importance"thatdoesnotincludeanobviousandexplicitreferencetothefactorofinfluence.�veninthatcase,thewayinwhichtheconceptof"outstandingliteraryvalue"isfrequentlyusedmaysuggestatacitreferencetothefactorofinfluence,forworkssodesignatedareusuallythosethatareconsideredclassics-thatis,workswhosereadershipandinfluencehavepersisted.Shouldthisbechallenged,andweretheretobeno HISTORYANDITSMODESreferencetoawork'sinfluencecontainedinvaluejudgmentsregardingliterarymerit,theroleofinfluencewouldneverthelessbedeeplyingrainedinanyliteraryhistory,sincethehistoryofaliterature,orofagenre,doesnotconfinei�selftocriticismandtothecomparisonofindividualmasterworks.Therefore,inaliteraryhistory,orinanyotherformofspecialhistory,thequestionofinfluenceandthespreadofinfluencestandsattheheartoftheproblemofcontinuity.Intracinginfluencesinliterature-toremainwithitasanexample-thehistoriansometimeshasexplicitevidencetoguidehim:Letters,diaries,theauthor'sconversationswithothers,andthelikeprovidethehistorianvithsuggestionsastotheworksandeventsthatmayhavemostinfluencedanauthor.Suchmaterialsare,ofcourse,ofgreatbiographicalvalue,butonecannottakeanauthor'sexplicitcommentsonotherauthorsorevents(norcanonetakeanyabsenceofsuchcomnlent)asanaccuratereflectionofthedegreetowhichhewasinfactinfluencedbythem.Inliteraryhistory,asinanthropologicalinvestigationsofthespreadofculturetraitsandofculturecomplexes,onemustbeguidedbyresemblancesthatarefound,aswellasbytheprobabilitythattherehasbeenanopportunityforthedisseminationofinfluence.12Letters,diaries,and·thelikecandirectlyattesttotheexistenceofcontact,andthustotheopportunityforinfluence;however,theliteraryhistorianhasother,lessdirectevidencehecanuseconcerningopportunitiesforinfluence:theavailabilityoftherelevantworksintheiroriginalformorintranslation,discussionoftheseworksinperiodicalsknowntohavebeenreadbytheauthorinquestion,andthelike.Yet,itisthroughresemblancesthatmostcluesastoinfluenceareoriginallydiscovered;andafterthepossibilityofcontacthasbeenestablished,itisonthestrengthoftheseresemblances,andtheunlikelihoodoftheirhavingbeenaccidental,thatthehistorian'sargumentforinfluencemustfinallyrest.Innotingresemblancesandtracinginfluencesamongliteraryworks,thehistorianofliteratureisledtoweaveacomplicatedpatternofrelationships:Eveninthesimplestcases,therewillnotbeanysinglelineofinfluencetofollow,inwhichainfluencedb,andbinfluencedc,butinwhichcwasinfluencedbynothingbutb.Anywork,c,willhavebeenwrittenbysomeperson,andtheinfluenceofbonhiswork-nomatterhowstrong-willnotexplainc,sincetheauthor'sownexperienceandstyle,aswellastheneednotmerelytorepeatwhathasalreadybeendone,willhaveaffectedhiswork.Norisitlikelythattheworkofanauthorwillbeaffectedbyonemodel VARIETIESOFSTRUCTUREINHISTORICALACCOUNTS37only:Whatevertheinfluencesfromthepastorfromhiscontemporariesthataffecthilll,theyarenotlikelytoremaindiscrete,butwillblendandaffecthisvariousworksindifferentways.Therefore,evenifoneweretowritethehistoryofsomelimitedliteraryform,inmovingfromoneauthortoanothertheliteraryhistorianmusttakemuchintoaccountbeyondthespecificformheistotrace.Hemustbepreparedtotreatthemostimportantworkswritteninthisformintheirowncontexts,andnotmerelyasillustrationsofthatform;furthermore,insomecaseshemayalsobeexpectedtoaccountforthefactthattheformwithwhichheisconcernedaroseandflourishedwhenitdid.Forexample,itwouldbeapoorhistoryoftheepistolarynovelintheeighteenthcenturyifonewereledbysimilaritiesintheirformtotreatPamela,LaNouvelleHeloIse,andTheSorrowsofWertherwithoutconsiderationofRichardson,Rousseau,andGoethe,asifthesenovelswerereIatedonlytoeachotherandnottothelivesandotherworksofthosewhowrotethem.Itwouldalsobeanimpoverishedhistoryoftheepistolarynovelifnoattemptweremadetoshowwhattheseworkshadincommonbeyondtheirepistolaryform,andhowsuchcommonelementswererelatedtootheraspectsoftheliteratureandthoughtofthetime.ItisforthisreasonthatIhavecharacterizedthetaskofaliteraryhistorianasoneofweavingacOlllplicatedpatternofresemblances(and,ofcourse,alsocontrasts)amongtheworkswithwhichhisdefinitionofliteratureandhischosenaspectofthatliteraturehasledhimtodeal.Andsinceheisahistorian,andisnotmerelyclassifyingworksaccordingtoresemblancesinabstractionfromquestionsastowhentheywerewritten,orbywhom,hewillseektoaccountfortheseresemblancesintermsofinfluences:Inotherwords,hewillseeknotmerelytoanalyzetheresemblancesanddifferenceshefinds,buttoexplainthem.13Notallsuchexplanationswillbeofthesametype,sincemanysortsofinfluencescanaffectanyliterarywork.Forourpresentpurposesitwillbesufficienttoclassifythemunderfourgeneralheads,whichmayberoughlydescribedasfollows:(1)influencescomingfromotherliteraryworks;(2)influencescomingfromtheotherarts,orfromreligion,thesciences,orphilosophy;(3)factorsinanauthor'sownpersonalityandinhisexperiencesthatcanberelatedtoindividualworksortohiscreativeworkasawhole;and(4)political.andsocialfactorsinthelifeofhistimebywhichhisworkcanbeshowntohavebeendirectlyinfluenced.Thisclassificationofthesortsofinfluencesonemayexpecttofindinliteratureisnotintendedto HISTORYANDITSMODESapplytoallkindsofspecialhistories,butitdoeshaveapplicabilityinfieldsotherthaninliterature-forexample,tothehistoryofthepictorialartsorofphilosophy.Ishallnotattempttodealwiththesedifferentformsofinfluenceindividually,butshallmerelyusethemtoillustratehow,inthesequentialstructureofaliteraryhistory,orinotherformsofspecialhistory,explanatoryelementsareintroduced.Forexample,whenthereisasuddenbreakintheIiterarytraditionsofanation,theIiteraryhistorianwillattempttoexplainthatchange.Insomecaseshecanappealtoinfluencescomingfromsourcesotherthanliterature,whetherthesearoseinotherartsorweretobefoundinintellectualorsocialchanges;inothercaseshemayfindtestimonythattheoldformshavewornthinandnolongerattract,andamajorinnovatorhasappeared.Insuchcasestheinnovationsmayhavebeendirectlyrelatedtothelifeexperiencesofoneorafewmajorinnovators;inothercasestheymayhaveresultedfromarebellionagainstcurrentfashionsandasearchforrenewalinearlierindigenoustraditionsorinexoticforms.Thepowerofnoveltyisnottobeunderestimated,andwhenitisfoundinamajorcreativefigure,hisworkmaysetapatternforhisgenerationand,foratime,forotherswhocomeafterhim.Inothercases,whereoneisnotdealingwithswift,revolutionarychanges,butwithnewemphasesandanewtone,theinfluenceofindividualsmayalsobestrong,buttheliteraryhistorianwillbemoreapttolookfortheexplanationofsuchchangesintheintellectualandsocial1ifeoftheperiodwithwhichheisconcerned.Thispointneednotbelabored:Theliteraryhistoriandoesnotmerelydepictthechangeshefinds,butseekstoaccountforthemintermsofthedifferentinfluencesthathaveeffectivelyshapedtheworksandhavechanneledtheinfluencesoftheauthorswithwhomhehaschosentodeal.Itisatthispointthatonecanmostclearlyseetheilnportanceofinterpretiveaccountsfortheliteraryhistorian,orforanyothertypeofspecialhistorian.Unlessabackgroundhasalreadybeenpresentedinwhichtherelationshipsbetweensocial,political,religious,philosophical,scientific,andotherimportantfactorsinthelifeoftheperiodhavebeendepicted,therewillbevirtuallynoInaterialsuponwhichtodrawwhentheliteraryhistorianwishestoexplainthechangeshefindsandtheeffectsthesechangeshavecarriedintheirtrain.Therefore,hemust,atvariouspoints,presentacross-sectionaldepictionoftheconditionofthesocietywithwhichhedealsandof VARIETIESOFSTRUCTUREINHISTORICALACCOUNTS39thevariouselementsenteringintoitsculture,orhemustpresupposethathisreaderswillhaveknowledgeathandregardingthesefactorswhenheintroducestheminexplainingthepersistenceorthechangesoftheformsandthemeswithwhichhedeals.Inpresentingsuchinterpretiveportraits,ahistorianneednotconfinehimselftodiscussingworksofrnajorimportance.Infact,heisunlikelytodoso.ManyworksthatarenotofoutstandingmeritnorofsignificantinfluenceInayfindtheirwayintoaliteraryhistory,orintoanyotherformofspecialhistory.Theseareworkstakentobemoreorless"typical,"thatis,worksthattoamarkeddegreeareparadigmaticofastyleoraformofsensibility,orofreactionstothepolitical,social,orintellectualaspectsofthetimes.Forexample,whileitisunlikelythatanyexamplesoftheeighteenth-centuryGothicnovelwouldbeincludedinhistoriesofEnglishliteratureiftheliteraryhistorianweretobeconcernedonlywithoutstandingliteraryvalueorwithsubsequentinfluence,suchnovelsdorepresentafacetofEnglishsensibility,andareusuallyincludedinhistoriesofEnglishliteratureinthatguise,beingtakenasharbingersofsomeoftheattitudesassociatedwith"Romanticism."Sincetheliteraryhistorianwill,aswehaveseen,introduceexplanatoryelementsintohissequentialaccountofalmostanyaspectofliteraryhistory,andsincesuchexplanatoryinterpretationspresupposeinterpretiveanalysesofthesocietyanditsculture,itisobviousthatinthefieldofspecialhistories,nolessthaningeneralhistories,asequentialaccountwillnottaketheformofasimplenarrativeexposition;itwillinvolvecrisscrossingrelationships,inwhichthehistorianpicksupfirstonethreadandthenanotherandweavesthemintoanintricatepatternthatfollowsthecomplexofrelationshipswithwhichheisconcerned.IIIAsimilarintricatepatternofrelationshipsistobefoundinhistoricalaccountsthatareinterpretiveintheirbasicaimandstructure.Inthem,however,thesequentialpatterningthatarisesthroughtracinginfluencesislessinevidence.Whileaninterpretiveaccountisnotusuallyconfinedtoasinglecrosssectionoftimebutspansaperiod,asIhavenotedinthecaseofYoung'sVictorianEnglandandBurckhardt'sCultureoftheRenaissanceinItaly,theemphasisinsuchworksisonthemannerinwhichaspectsofsocietyoroftheculture HISTORYANDITSMODESoftheperiod,orboth,fittogetherinapattern,definingaforInoflifedifferentfromthatwhichonefindsatothertimesorinotherplaces.Itisofinteresttoconsiderwhatlendsstructuretointerpretiveaccountsasthusconceived.InthefirstplaceitistobenotedthatsincethehistorianisseekingtodepictthenatureandreIationshipsofvariousaspectsoflifeatoneparticularplaceandtime,hisaccountwillhavetofitintoachronological,sequentialframeworkofgreaterdurationthanthatofthecontinuingstateofaffairswithwhichheisdirectlyconcerned.AsweseeinYoungandinBurckhardt,hewillinitiallyneedtoprovideabackgroundforthespecificperiodheistointerpret.Inprovidingsuchabackground,hispositionistheconverseofthatofsequentialhistorians:Theymustinformtheirreadersofthestateofaffairsobtainingattheoutsetoftheirsequentialaccounts,whereasthebackgroundtheinterpretivehistorianmustsupplyisasequentialbackgroundthatsetsthestageforthepatternsoflifewithwhichheistobeconcerned.Thus,theongoingprocessesinwhichaparticularstateofaffairsisembeddedcontributeoneelementofstructuretointerpretivehistoricalaccounts,andthechronologyofeventswithintheperiodcannotbeneglected.Tobesure,theinterpretivehistorianneednotbeconcernedwithalloftheknownfactsleadinguptoaninitialstateofaffairs,norwithalltheeventsthatlateroccurredbecauseofthem;noteverychangethatmayhaveoccurredintheprecedingperiodwillberelevanttothelaterstateofaffairswithwhichtheinterpretivehistorianisconcerned.Norwilleveryeventthathasalegitimateplacewithinasequentialaccountofaperiodalsoberelevanttoaninterpretivehistory,althoughmanyeventswillhaveaplaceinboth-.Whilethepresencewithinanyinterpretivehistoryofelementsderivedfromvarioussequentialhistories-suchasthoseconcernedwithpolitical,economic,literary,orreligiouslife-providesastructuralframeworkorprincipleoforganizationthatcannotbewhollyneglected,asecondandevenmoresignificantfactorinstructuringanyinterpretivehistoricalaccountisthehistorian'sviewofwhataspectsinthetimesweremostcharacteristic,pervasive,andfundamentalforthepatternoflifeheisattemptingtoportray.Sometimessuchconvictionsarederivedfromgeneralsociologicaltheories,orfromsomeparticularphilosophicorhistoricalbias,buttheyneednotbe:Manyhistoriansinfactdenythatthesameaspectsoflifearealways,andinallsocieties,equallyfundamenta1.Whateverthesourceofhisinterpretationmaybe,theinterpretivehistorianwillattempttoshowhow VARIETIESOFSTRUCTUREINHISTORICALACCOUNTS41theperiodwithwhichheisconcernedmaybeviewed-ormaybestbeviewed-intermsofsomebasicthemeorthemes.Thistypeofselectiveorganizationdiffersatleastindegree(and,asIshalllaterargue,alsoinkind)fromthestructuretobefoundinotherformsofhistoricalaccount.Whileitistruethattheworkofanyhistorianpresupposesthatcertaineventsoraspectsoflifearetakentobemoreimportantthanothers,inhistoricalworksthataimtobeprimarilyinterpretiveitistheinterpretivethemeitself,andnotaparticularseriesofintrinsicallyrelatedoccurrences,thatservestoexplainwhytheelementsthehistoriandiscusseshavebeenbroughttogether.Undertheseconditions,it'mayseemthatIacknowledgethataninterpretivehistorianhasfreerein,andcanineffectofferalmostanyinterpretationofaperiodthatfitsthethemehehaschosen.This,however,isnotthecase.Differenthistoriansmay,ofcourse,adoptdifferentpositionsastowhichbasicthemesprovidethegreatestinsightintoaperiod,andthusinwhatlighttheperiodmaybestbeviewed,butthereisastandardagainstwhichsuchclaimsaremeasured.Thatstandardishowwellagiveninterpretivethemeissupportedbyawiderangeofevidence.Thus,regardlessofwhatmayhaveledahistoriantoofferoneratherthananotherinterpretationofaperiod,itisonthebasisofevidencethatinterpretivehistories-nolessthanotherhistories-aretobejudged.Tounderstandthelimitsthatevidenceplacesonthehistorian'sinterpretationofaperiod,onemustfirstnotethatinterpretivehistoryisnotanindependententerprisethatcanproceedwithoutusingtheresultsofotherhistoricalinquiriesnotprimarilyinterpretiveintheiraimsandtheirstructure.Thisismerelytosaythatinorderforaninterpretivehistoriantogettoknowtheformoflifeinagivenperiodhemustdrawuponotherstudies,en1bracingthepolitical,economic,Iiterary,religious,artistic,andotherdevelopmentsthatoccurredwithinthatperiod;otherwisehewillnothavematerialstointerpret.Tobesure,theinterpretivehistorianshouldbemorethanasynthesizerofalreadyfamiliarmaterials;theremaybeagreatdealofresearchhehimselfmustdoafterhehasfoundtheframeworkthatforhimbestcharacterizestheformoflifeintheperiodwithwhichheisconcerned.Yet,eventhisfurtherresearchmustcoherewithwhathasalreadybeenknownconcerningtheperiod,anditisusuallyhisownpriorknowledgeofmanyaspectsoftheperiod-ratherthanpriorsociologicalorphilosophicalcommitments-thatsuggesttotheinterpretivehistorianinwhatlightthosephasesoftheperiodinwhichheisinterestedcanbestbeviewed.Atthispointhisinterpretation HISTORYANDITSMODESmustbeabletowithstandcriticism:Otherswhoareknowledgeableinthesamefieldswilljudgewhethertheinterpretationissupportedbytheevidenceadduced,andwillalsowanttoconsiderwhetherinlportantcontradictoryevidenceexists.Furth.ermore,interpretiveaccountsarepraisedorcondemnedonthebasisofwhetherthefitbetweentheinterpretationanditsevidenceleadsonetoseeotherpointsatwhichtheinterpretationisapplicabletotheperiodinquestion,orwhethertheinterpretationisplausibleonlybecauseitsauthorarbitrarilyrestrictedhisattentiontothoseaspectsoftheperiodthatillustratedthethemehehadchosentopush.Historians-nolessthanscientistsapplygeneralprinciplessuchastheseinestimatingthefitbetweenaninterpretationandtheevidenceneededtosupportit,andtheyapplysuchprinciplesconcretelyandindetailwithrespectbothtowhathasbeensaidandtowhathasbeenleftunsaidandshouldhavebeensaidinanyparticularwork.Interpretivehistoricalaccountsarenotonlyjudgedaswholes,inwhichcasetheprimarybasisofjudgmentwouldseemtobeaquestionofwhatevidencewasadduced,orwhatshouldhavebeentakenintoaccount;theyarealsojudgedinamorepiecemealfashion.Asisthecasewithrespecttoothertypesofhistoricalworks,oneexpectsthemnotonlytobeenlighteningwhentakenasawhole,butalsotobesoundintheirparts.Therefore,aninterpretiveaccountofaperiodthatisnotalsosoundinitsinterpretationofthespecificelementswithwhichitdealswithinthatperiodwillbeliabletocriticism.Atthispointitiswelltorecallthatinterpretiveaccountsalwayspresuppose,andatvariouspointsincludewithinthemselves,materialsdrawnfromavastnumberofrelevantsequentialhistories.Similarly,aswehavenoted,withinanyinterpretiveaccounttherewillbesuggestionsastohowoneistoexplainparticularchangesthattookplacewithintheperiodunderconsideration.Whereonepartofaninterpretiveaccountcomesintoconflictwithspecificfactsorrelationshipsthatmayhavebeengenerallyagreeduponbycompetenthistorians,itwillpresumablybetheinterpretiveaccount,notthesequentialorexplanatoryones,thatwillbedamaged.Suchdamagemaynotbeirreparabletotheinterpretationasawhole.Theconflictmaysimplyleadonetosubstitutesomeelementotherthantheoneinquestionasevidencefortheinterpretationoffered.Inotherwords,whenampleevidenceexists,emendationandnotreinterpretationmaybeallthatiscalledfor.However,incasesofbasicconflictitwillbetheinterpretiveaccount,notthesequentialorexplanatoryaccounts,thatwillcallforalteration. VARIETIESOFSTRUCTUREINHISTORICALACCOUNTS43Thisfactdoesnotsignifythatinterpretiveaccountsareinferiortootherformsofhistoriography;itfollowsfromthefactthatweareherespeakingoftheelementswithininterpretiveaccounts,andnotoftheinterpretationasawhole.Suchelementsserveasevidencefortheinterpretation,andwhateveraccountisgivenoftheseelementsmustholdupagainstthesamesortofcriticalscrutinyoneusesintestingthecontinuitiesdepictedinsequentialaccountsorthelinkagesamongeventsthataretakenasexplanationsofanyparticularoutcome.Inotherwords,theoverallinterpretationofaperiodthatisofferedbyahistorianwillnotserveasajustifyinggroundforhisviewofthevariouselementsthatareofconcerntohiminthatperiod.Itistheseelementsthatserveasevidenceforhisinterpretation,sincetheinterpretationasawholecannotserveasevidencefortheaccuracyoftheelementsuponwhichitissupposedlybased.Thus,withrespecttowhateverisincludedwithinanyinterpretivehistoricalaccount,thestandardsofcriticismtobeusedarethesameasthoseemployedinexaminingthereliabilityoftheelementspresentinanyothertypeofhistoricalaccount.IVInconcludingthisdiscussionofthetypesofhistoricalaccountsthatIhavedifferentiated,Iwishtoemphasizeonceagainthatalthoughthestructureofanyparticularhistoricalworkwillbepredominantlyofoneofthesetypes,ineverysuchworktherewillbepassages-andfrequentlylongpassages-inwhichtheotherformsarepresent.Thus,aswehavenoted,explanatoryaccountswillbeintroducedintoworksthatareprimarilysequentialorinterpretiveinnature,andasequentialframeworkmakesitsappearanceinbothexplanatoryandinterpretiveaccounts.Aswehavealsonoted,atleastarudimentaryformofinterpretivehistoriographyisalsotobefoundinsequentialandexplanatoryhistoriesbecauseofthenecessityfordepictingtheinitialstateofaffairsthatanysequentialorexplanatoryaccounttakesasitspointofdeparture.Thereare,however,eventightersetsofbondsthattietheseformsofhistoriographytogether.Thechiefoftheseisthatthesameeventsaresusceptibleoftreatmentinallthreetypesofaccounts:Whenevertheseaccountsdealwiththesamesocietyorsocietiesoverroughlythesamespanoftime(iftheyuseroughlysimilarscales),eachwillincludediscussionsofmanyofthesameevents.Theelementsthat 44HISTORYANDITSMODESthusmaketheirappearanceindifferenttypesofhistoricalaccountsmustbedescribedandinterpretedinnoncontradictorywaysiftheaccountscontainingthemaretobeconsideredreliablewithrespecttothem.Thus,solongastwoormorehistoriansareconcernedwiththesameevents,theirworkswillnotbeirrelevanttoeachother,nomatterhowdifferenttheirbasicapproachesmaybe.Itthereforefollowsthathistorianswhoworkinthesamefields,orinfieldsthatcommitthemtobeingconcernedwiththesameevents,arenottoberegardedasworkingatcrosspurposes,evenifdifferencesintheirassumptionsandtheirmethodsmightotherwiseleadonetoregardthemasopponentsratherthanasco-workers.Idonotwishtosuggestthatalldisputesbetweenhistoriansvillberesolvedandthatallhistorianswillultimatelyreachconsensusintheirviewsregardingtheeventswithwhichtheyaremutuallyconcerned.Imerelywishtoinsistthatinsofarastheantagonismsonefindsarebasedonradicaldifferencesinmethods-asdistinctfromantagonismsbasedonpersonalhostilities,prideofplace,orthelike-oneshouldnotassumethatitisinprincipleimpossibletoreachasubstantialmeasureofagreementastowhatoccurredingivensocietiesandhowithappenedthattheseeventsdidoccur.Shouldthiscontentionappeartobehopelesslyoutoftouchwiththerealitiesofthedisagreementsthatexistamonghistorians,thereareseveralpointstobenoted,onlytwoofwhichIshallmentionhere.First,itistobenotedthatIhavenotclaimed,andwouldnotclaim,thatdifferentinterpretivehistories,whentakenaswholes,aremutuallycOlnpatible;Ihavearguedonlythattheadequacyoftheirtreatmentsoftheelementstheyintroducecanbeassessed,andthattheiradequacywithrespecttotheevidencetheyadduce,orfailtoadduce,canalsobeassessed.Itremainspossiblethatafterthishasbeendonetheweightingoftheelementsintheinterpretationasawholemaydifferfromhistoriantohistorian,andtheremaynotbeanywayinwhichsuchdifferencescanbedecidedindependentlyofmoregeneralsociologicalorphilosophicalcommitments,orindependentlyofpurelypersonalpreferences.14Second,Iwishtostressonceagainthatdifferenthistoriansworkondifferentscalesandareconcernedwithdifferentaspectsofthelifeofaspecificsocietyatagiventime.Therefore,theproliferationofworksonwhatsuperficiallyappearstobeHthesamesubject"neednotbetakenasindicativeofaneedforcontinualattemptstostartafreshandtorewriteallthathasbeenwrittenconcerningourhumanpast.If,asweshallseeinsubsequentchapters,itisnormaltofindthatworkswrittenondifferentscalesandconcerned VARIETIESOFSTRUCTUREINHISTORICALACCOUNTS45withdifferentaspectsofasocietyprovetobeinterlocking,theproliferationofhistoricalaccounts,eachofwhichmakesapointofitsown,doesnotentailthatweshouldcommitourselvestoanyformofhistoricalrelativism.InthefinalchaptersIshallonceagaintakeupthisquestion,whichwastheissuewithwhichphilosopherswhowroteabouttheproblemsofhistoriographywereatonetimemostconcerned.Ingeneral,however,thisinterestsubsequentlyshifted,andproblemsconcerningthenatureofhistoricalexplanationcametodominatethefield.Suchproblemsare,ofcourse,relevanttoquestionsconcerningtheobjectivityofhistoricalknowledge,buttheydeservediscussionintheirownright.InthefollowingchaptersIshallbeconcernedwithsomeofthem;onlyinchapter5,however,willIdealspecificallywithhistory.ThisisbecauseIbelievethatquestionsconcerninghistoricalexplanationcannotbeadequatelydiscussedwithoutraisingissuesthatinvolveageneraltheoryofwhatconstitutescausalexplanations.Thattopic,inturn,involvesacarefulconsiderationoftherelationbetweencausesandlaws.SincetheviewsIholdonthesemattersareundoubtedlyheterodox,Ishalldeveloptheminthenexttwochaptersatconsiderablelength.Then,inchapter5,IshallsuggestsomeofthewaysinwhichtheseanalysesareapplicabletothevarioustypesofhistoricalstudyIhavealreadydelineated.Icanthenaddressthequestionofobjectivityonceagain. BLANKPAGE PARTTWOCAUSATION BlANKPAGE ChapterThreeCAUSALBELIEFSINEVERYDAYLIFE.,•'")Amongthemanychangesinphilosophicopinionthathavetakenplaceduringthepastthirtyyearsoneisofspecialconcerntothefollowingdiscussions.Thishasbeenachangeinthedominantviewofwhatconstitutesaproperanalysisoftheconceptofcausation.Theviewthatprevaileduntilrecentlymaybedesignatedindifferentways,dependinguponwhichofitsaspectsonewishestoemphasize�ItisoftencharacterizedastheHumeanview,whichcallsattentiontowhatwasundoubtedlyitsmostinfluentialsource.Ithasalsobeenreferredtoastheregularityviewbecauseittakesthecause-effectrelationtobeequivalenttoanempiricallyestablishedregularitybetweentwotypesofevents,oneinvariablyfollowingtheother.Itmightalsohavebeendesignatedinmoretechnicalterminologyasthenomotheticview,sinceitholdsthattheconceptofcausationisotiose49 CAUSATIONandshouldbeabolishedfromourphilosophicandscientificvocabulariesexceptwhenreferringtosomeparticularinstanceofanempiricallyestablishedlaw.Thesevariantformulationsarenottoberegardedasstrictlyequivalent.Forexample,Hume'spsychologicalaccountoftheroleoftheimaginationincausalattributionsdifferentiateshispositionfrommostsubsequentformulationsoftheregularityview;furthermore,hisemphasisontheimportanceofcausalbeliefsinallaspectsofourlivesisincompatiblewiththepositionofthosewho,holdingthenomotheticview,havewishedtoridusofanyrelianceuponcommonsensenotionsofcausation.Nevertheless,thesethreevariantformulations(towhichotherscouldbeadded)constituteatypeofphilosophicpositionregardingcausationthat,inspiteofsomecritics,tendedtodominateallothersforatleastoneortwogenerations.1Thefactthatthisdominationhascometoanendmaybestbesignalizedbymentionofonehighlyinfluentialbook,CausationintheLaw,byH.L.A.HartandA.M.Honore.2Itisnoexaggerationtosaythatsinceitsappearancein1959thewholetenorofdiscussionsofcausationinAnglo-Americanphilosophyhaschanged.Intheirpreliminaryanalysisofcausalconcepts,HartandHonoredrewasharpcontrastbetweentheplainman'snotionofcausation,whichtheyfoundtobedominantinhistoryandthelaw,andtheregularityview,whichtheyacceptedasbeing,onthewhole,applicableinthesciences.Inshort,theydidnotchallengetheprevailingviewwithinonedomain,butarguedthatitwasinapplicableinanother.Awillingnesstofollowthisgeneralstrategy-especiallyawillingnesstoholdthatscientificconceptsandproceduresarenotnecessarilyapplicabletotheaffairsofeverydaylife-hasbeenonecharacteristicfeatureofphilosophyinthelastdecades.Thereissomereasontobelievethatthisstrategyislosingitsappeal.Insteadofseglnentingphilosophicissuesinordertodealwiththempiecemeal,thereisnowatendencytotryandconnectthedistinguishablepartsoftheselargerproblemsthathadinitiallybeenseparatedthroughlinguisticanalysis.Withrespecttocausation,J.L.MackiehasrecentlyattemptedtobringvariousaspectsoftheproblemintoasinglefocusinhisdistinguishedbookTheCementoftheUniverse:AStudyofCausation.3InwhatfollowsIshallnotattempttodealwithalloftheproblemswithwhichheandothershavedealt,fortherearemanylogicalissuesthatarenot,Ithink,especiallyrelevanttohistoricalcausation.Ontheotherhand,Ishalldevotemoreattentiontothephenomenologicalfeaturesofthosesituationsinwhichwemakecausaljudgmentsthan CAUSALBE"LIEFSINEVERYDAYLIFEonefindsinMackie,orinanyotherdiscussionofcausationwithwhichIamacquainted.Idosoinordertofocusattentiononproblemsthatspecificallyconcerncausalattributions,ratherthanonquestionsconcerningthosegeneralizationsthat,inthenaturalsciencesorelsewhere,wetaketobelaws.Solongastheregularityviewwasdominant,suchadistinctioncouldnotbetolerated.Now,however,philosophersofdifferentschoolsareoftenwillingtodistinguishbetweencausallyexplainingwhyaparticulareventoccurredandformulatingageneralizationthatstatesthatwheneveraneventofaspecifictypeoccursitwillbefollowedbyaneventofsomeotherdesignatedtype.Aswehavenoted,HartandHonoreassumedthatthisdistinctioncoincidedwithadifferencebetweenourcausalattributionsinthecontextsofeverydaylifeandthe"characteristicstructureofscientificexplanation.Ibelievethattheirassumptionwasmistaken.Ishallthereforeattempttoshowinthepresentandfollowingchaptersthatalthoughthecausalexplanationof-aparticulareventdoesnotrestonknowledgeofwhatoccursinallcasesofaparticulartype-andthusthattheregularityviewismistaken-oneshouldnotconcludethatthisprovidesabasisfordistinguishingbetweencommonsense,history,andthelawontheonehand,andthesciencesontheother.4Myconcerninthepresentchapterwillbewithourcausalbeliefsineverydaylife.WhatIshallattempttoprovideisabroadrangeofexamplesofdifferenttypes,uninterruptedbydiscussionsofanyalternativeinterpretationsoftheseexamples.However,sincemyinterpretationdifferssomarkedlyfromtheviewsheldbyHume,andalsofromthoseheldbyHartandHonore,IfeelobligedtoexplainonwhatgroundsIfindtheirpositionsunacceptable.ThisIshalldointwobriefappendices.I.THEPERCEPTIONOFCAUSATIONItisnotuncommontofindthatsometermsfrequentlyusedbyphilosophersarerarelyusedineverydaylife."Cause,"withitscognates,isonesuchterm.Thisshouldnotsuggestthattheconceptofcausationisremotefromeverydayconcerns;onthecontrary,itmaysignifythatthecausalrelationissopervasiveinexperiencethatourordinarylanguagehasavarietyofexpressionsthatserve,inaroughmanner,todistinguishamongthemanyformsittakes.Forexample,ifweusetheexpression"xproducesy,"amoreimmediate,less CAUSATIONremoterelationshipissuggestedthanifwesay"ywasaconsequenceofx,"yetitiswhollylegitimatetotreateachexpressionasinvolvingtherelationofcauseandeffect.Again,whenweuseverbssuchas"kill,"wehaveinmindacausalrelationbetweensomeaction,orsomestateofaffairs,andthedeathitbroughtabout,yetneithertheword"cause"noranyofitsderivativesisusedwhenwesaythat"xkilledy."Similarly,wedistinguishbetweenpushinganobjectandpullingit,butbothpushingandpullingmaylegitimatelybetakenasinstancesinwhichacausalagencybringsaboutaneffect.Inallofthesecasesonemayrephrasetheoriginalstatementusingtheterm"cause"withoutintroducingasignificantchangeincognitivemeaning;itisthereforefalsetoassumethatsimplybecausephilosophershaveusedthemoregeneraltermtheyhaveinventedacoinagethathasnocorrespondinguseineverydaylife.Theiruseof"cause"wouldbedangerousonlyifitledthemtoidentifythattermwithsomeoneorsomefewtypesofcausalrelationship,neglectingallothers.ThisisadangerIwishtoavoid.Itisespeciallyimportantnottoconfineourdiscussionofeverydaycausalattributionstotoonarrowarangeofexamples,sinceallcausalnotions,whetherinscienceorelsewhere,andnomatterhowsophisticatedandalteredtheygrowtobe,havetheiroriginalrootsineverydayexperience.Onlyifwedirectourattentiontoawidevarietyofapparentlydisparateexampleswillweguardagainstoversimplificationandavoidadistortionofoureverydaycausalnotions.Inmyopinion,theprilnarysourceofmanyoversimplificationsanddistortionshasbeentheconventionalviewthatwhenwespeakofthecauseeffectrelationshipwealwayshaveinmindatemporalsequentialrelationshipinwhichsomespecificprioreventisthecauseofasubsequentevent.ThepervasivenessofthisviewisprobablyduetotheinfluenceofHume'sanalysis,foritisnotaviewtobefoundinAristotle,norinWesternmediaevalorRenaissancephilosophy,noramongtheRationalists.Yet,evenHartandHonoretakethisconventionalviewforgrantedandtreatitasifitwereclearlyadequateinallcausalsituations,whetherinscienceorineverydaylife.Tobesure,therearemanycasesinwhichwhatistakentobethecauseofaneventissomespecificeventthatwasprecedenttoitseffect.Ontheotherhand,therealsoaremanycasesinwhichtheconventionalviewdoesnotconformtoourordinarycausaldescriptions;wedonotalwaysregardthecauseanditseffectastwodistinctanddiscontinuousevents.5Insuchcases,acauseanditseffectappearasrelatedaspectsofasingleeventwhosecontinuitywedirectlyperceive,rather CAUSALBELIEFSINEVERYDAYLIFE53thanasdifferent,separableevents.Ishallfirstdealwiththesecases.Itwillthenbepossibletoshowthatevenwhentheconventionalviewisapparentlycorrectitdoesnotprovideasatisfactorymodeofanalysisforthecausalrelationship.Tospeak-asIshall-ofdirectlyexperiencingaconnectionbetweenacauseanditseffectis,ofcourse,tochallengeHume,andforthatreasonAppendixAisdesignedtoshowthathisargumentagainstthepossibilityofperceivingtheconnectionbetweencauseandeffectisbynomeanscompelling.Now,however,Iwishtopresentcasesinwhichitislegitimatetospeakofperceivingcausalconnections.6Ifweconsultourdirectexperience,uninhibitedbyHumeanobjections,andifweconsideronlythosecasesinwhichwebelievethatwedirectlyseeacausalconnection,wewillnot,Ibelieve,describewhatisseenjustasHumeandothershavedescribedit.Althoughithasbecomeusualtospeakofthecausalconnectionasexistingbetweentwoevents,inmanycasesinwhichwethinkweareseeingcausalpower,itiswithindifferentphasesofacontinuoushappening,andnotbetweentwodistincteventsthatweexperienceconnection.Insuchcaseswhatweseeas"theeffect"istheendpointofaseriesofchanges;whatweseeas"thecause"ofthiseffectiswhatleduptoitinthatcontinuousseries.Tobesure,thisisnotalwaysthewayinwhichwedescribeacause-and-effectrelationship:ItiscertainlynotthewayinwhichIwoulddescribethecausalconnectionwhenIswitchonalight.Ishalldealwithcasesofthelattertypeinthenextsectionofthischapter,andweshallthenfindthatsuchcausalattributionsdependuponregularityofsequence.Thus,mybeliefintheconnectionbetweenmyflickingtheswitchandhavingalightcomeonisabeliefthatmustbeacquiredinthewaythatHumesaidallcausalbeliefsareacquired.Now,however,Iamdealingnotwithallcausalbeliefs,butonlywiththoseinwhichwewoulddescribeourselvesasdirectlyseeingwhatHumedeniedwecaneversee:directconnectionortransferenceofpower.AsexamplesofsuchcasesIshallfirstcitetworeadilyfamiliarinstancesinwhichaspectatorwatchingagalnemaybesaidtoseetheconnectionbetweencauseandeffect.Whenweseeafootballplayerrunningwiththeballandthenbroughtdownbyahardtackle,wehavewitnessedaneventthatisseenashavingacertainunity.ThatithasthisunityforusInayinpartbeduetothefactthatweknowtherulesoffootball,butitisverydoubtfulwhetherthisisthetotalexplanationofitsexperiencedunity.7Thatquestionisnot,however,myimmediateconcern.Myconcernistolocatewhereintheeventweseeacausalconnection, 54CAUSATIONandwhatitisthatwetaketobethecauseandwhattheeffect.Inthiscase,asinothers,onemuststartfromtheeffect,forinthecauseeffectrelationshipnothingisdenominatedasacauseindependentlyofouractuallyrecognizingitasthecauseofaparticulareffect.Now,theeffectinthisparticularcasecanbedescribedinseveraldifferentbutentirelycompatibleways:forexample,asthefactthattheplayerwhocarriedtheballwashaltedbybeingbroughtdownbythetackler,orthattheplaywasstoppedonthefive-yardlinebynumber99,thetackler.(Thefactthattherearealternativecompatibledescriptions,differinginscale,isapointthatwilllaterbeoffundamentalimportancetous.)Ineachofthesedescriptionsthecauseoftheeffectwastheeffectivetackle,butthetacklewasnotanothereventprecedingtheeffect,asonewouldexpectifoneweretothinkofcauseandeffectintermsoftheregularitymodel.Thetacklewasthebringingdownoftheballcarrier,notaprecedentevent;or,onthealternativedescription,number99stoppedtheplayonthefive-yardline,shortofatouchdown,ineffectivelytacklingtheballcarrier.Thosewhowishtoholdtothenotionthatthecausemustalwaysbesomeeventprecedenttoanddistinctfromtheeffectmayseektoescapefromtheforegoingdescriptionsbyregardingthetacklingasaseriesofsequentialeventsofexceedinglyshortduration,eachofwhichcauseditsimmediatesuccessor,withthelastiteminthesuccessionbeingtheballcarrier'sbeingbroughttothegroundinthegraspofthetackler.However,suchananalysiswouldnotonlyfailasadescriptionofwhatitisthatwereallyseeinsuchacase(forwecannotdiscriminatewhatisactuallyoccurringateachoftheseassumedsuccessivestates),butitwouldalsofailtoestablishthat,properlyanalyzed,whatcausedtheeffectwasalinearsequenceofsuccessiveevents.Itwouldfailinthisrespectbecausewerecognize,eitherdirectlyorfromourownpastexperience,thatthereispresentasimultaneousconflictofforceswhenatacklertacklesarunner:Itisnotaquestionofaseriesofsuccessiveindependentmovementstakingplace,eachofwhichisfollowedbyadistinctresponse;forcesareateveryinstantexertedbytacklerandrunner,bygravityandbyinertia,withtheinteractionoftheseforcesendinginthiseffect.ItmightbeobjectedthatinmydescriptionIhavegivenapec�liarandtruncatedcausalanalysissinceIhaveomittedrelevantprecedentactions,suchasthefactthatthetacklerhadtocatchtheballcarrierinordertotacklehim;thus,itmightbeargued,wemustcarrythecausefurtherbackintime,towhatprecededtheeffect.Thereis,ofcourse,noreasonnottoincludethespeedwithwhichthetackler CAUSALBELIEFSINEVERYDAYLIFE55pursuedtheballcarrieraspartofthecauseoftheeffect;however,ifwedosowearenotconsideringas"theeffect"thattheballcarrierwasbroughtdownbytheforceofthetackle,butthattheplayendedatthefive-yardline,andnotinatouchdown.Inthisdescriptionoftheevent,thespeedofthetacklerinovertakingtheballcarrierispartofthecauseoftheeffect,butitobviouslyisnotacausethatprecededtheevent(theplaythatendedonthefive-yardline);rather,itisincludedwithinit.Letmenowusemysecondillustration,thatofathree-cushionbilliardshot,inordertomakethesamepointonceagain.8Inthiscase,too,wemustdistinguishbetweenalternativedescriptionsofwhatwetaketobetheeventthatendsinaparticulareffect.Whenourattentionisfocusedontheactualcollisionofthecueballwiththeredball,weseetheredballmovewhenitishit;wedonotactuallyseeitbeinghitandthenmoving.Thus,thecauseofitsmovingisitsbeinghit.If,afterthecollision,weareaskedwhyitmovedwithsuchforce,wearelikelytoanswerthatitwasbecausethecueballhadstruckitwithgreatforce,butevenhere,whenweareforcedtousethepasttense,wearereferringtowhatwasseenasasinglecontinuousprocess,ratherthanasbeingtwosuccessive,independentevents:Inthecollision,wehaveseentheforceofthefirstballapparentlytransferredtothesecond.Now,tosaythattheredballmovedbecauseitwashit,ormovedwithgreatforcebecauseitwashitwithgreatforce,maynotbeconsideredveryenlightening,buttheyareanswersthatwegiveandwegetineverydaylife.Whatisrevealingaboutthemisthateventhoughweusethepasttensewhenwelookbackuponwhatwesaw,whatwastakentobethecausewasnotanoccurrenceprecedenttoitseffect:Itwasthebeingstruckthatwasthecauseoftheball'smoving,anditwastheforcewithwhichitwasatthatinstantstruckthatcausedittohavetheforcewithwhichitstartedtomove.(If,astheballthenslowsdown,weshouldbeaskedwhatisslowingitdown,wewouldhavetoadmitthatwedon'tseewhatisdoingso,butwemaysay"itislosingforce";ifweareaskedwhatthatmeans,wecanexplainthatfrictionwiththetable'ssurfaceisslowingtheballdown.Ineithercase,onceagainwewouldbeusingthepresenttensetodescribetherelationofcauseandeffect.)Tobesure,wecantakea.longerviewoftheeventinquestioninthesearchforaprecedentcause,justasitwaspossibletodointheexampledrawnfromfootball;thoughhere,too,thestratagemwillbeunavailing.Wemay,forexample,saythattheredballmovedwithgreatforce,inthatparticulardirection,becausetheplayerwasmak- CAUSATIONingathree-cushionshotandthereforestruckthecueballinamannerthatwouldmakeitstriketheredballjustasitdid,sendingtheredballinthatparticulardirection;andwecansaythatthiswasn,ecessaryifthecueballwastotravelwithsufficientforcetowardthefirstofthethreecushionsitwastostrikebeforetouchingtheopponent'scueball.Onceagain,however,suchadescriptionincludesthecause-whattheplayerdid-withintheeventitself,theeventbeingathreecushionshotthatwassuccessfullycomplet�d.ThisviewofwhatweinmanyinstanceStaketobethecause-effectrelationshipmaybeillustratedbybrieflycitingotherexamplesdrawnfromeverydaylife.9Forexample,ateacherisattheblackboardwritingasentence,andthestudentswatchashewrites.Theviewofsomemaybeobstructed,butotherswillseethatashishandmoves,holdingthechalk,theindividuallettersandwordsareformeduntilthesentenceiscomplete.Thecompletedsentenceaswrittenontheboardistheeffect:Itisthestateofaffairsinwhichtheeventended.Butwasthecausesomeseparateanddistinctprecedentevent?Notifthecauseistakentobethewritingofthesentence,forthiswasconstitutedbyformingthelettersthatformedthewordsthatmadeupthesentence,andthiswaspartandparceloftheeventitself.Norwillthisconclusionbeescapedbysayingthatitwasthechalkmovingontheboard,orthatitwasthemovementofthehandthatwasthecause,fortheseoccurred,seriatim,atpreciselythetimethattheletterswerebeingformed.Noristhereescapeinadoptingthelongerrangepointofview,seekingthecauseintheteacher'sintentiontowritejustthissentenceandnoother,fortheeffect-thewrittensentence-wouldnothaveoccurredunlesstheteachernotonlyhadintendedtowritethesentencebuthadactuallywrittenit.Thus,theeventhassimplybeenassignedalongertimespan,includingtheteacher'sintention;however,thatintention(whichisnowseenascausallyrelatedtotheeffect)isnottobetakenasprecedingtheevent,butisviewedasapartofit.Casesofacause-effectrelationthataresimilarinstructuretothatofwritingasentencecanbemultipliedalmostindefinitelyifoneconsidersthosecasesinwhichordinaryactsofpullingorofpushingobjectsoccur..Forexample,ifIamdraggingaheavyobjectattheendofarope,orpushingapieceoffurniturefromonepartofaroomtoanother,howevershortadistanceorhoweverfarImoveit-andwhetherIevergetittothespottowhichIintendtomoveit-thecauseisnotsomeseparateprioreventbutconsistsinallthatisinvolvedintheactofhaulingorpushing:Itisincludedwithinthe CAUSALBELIEFSINEVERYDAYLIFE57boundariesoftheeventitself.Furthermore,justasinthecaseofwriting,suchaneventmaybeviewedeitherasapurelyphysicalprocesswhereanindividual'smotionsbringaboutaspecificeffect,orasanintentionalact;ineithercase,however,whatbringsabouttheeffectisapartoftheprocessthateventuatesinthateffect.10Thesignificanceofthesecasesinthepresentcontextisthattheyillustratethefactthatinmanyofoureverydayusesofcausalnotionstheseapplytowhatoccurswithinwhatisseenasacontinuous,ongoingprocess.Statingthisconclusionwithsomewhatgreaterprecisionthanitwasconvenienttoemployindiscussingtheprecedingexamples,ineachcasetheeffecttobeexplainedissomeparticularstateofaffairsthathasbeensingledoutforattention.Inthesecausalattributions,suchastateofaffairsisviewednotasanisolatedevent,butastheendpointofaprocess;whatwetaketobethecauseofthatstateofaffairsistheprocessleadingtoit,outofwhichiteventuated.Wemay,ofcourse,distinguishvariousphasesorvariouscomponentswithinthatprocess,butinourexamplesthesewerenotseenasaseriesofseparateanddiscontinuousevents.Tobesure,shouldwewishtodoso,wecanimaginetheprocess,onceithasoccurred,asifithadbeenlnadeupofaseriesofseparate,successivestates,butthisformofcinelnatographicalanalysis(asBergsoncalledit),inwhichcontinuousprocessesaretreatedasiftheywerecomposedofaseriesofdiscontinuous,independenthappenings,doesnotreproducewhatweperceivewhenweperceivecausalconnections.Whatisseenastheprocessleadinguptotheparticularstateofaffairstobeexplainediswhatistakentobethecauseofthatspecificstateofaffairs.llInshort,inthosecasesinwhichwemaybesaidtoperceivecausationanditisonlywithsuchcasesthatIamnowdealing-theconnectionbetweencauseandeffectliesinthefactthatbothareseenasaspectsofasingleongoingprocess,ofwhichtheeffectisviewedas.itsendpointorresult;thecauseofthisresultistheprocessitself.MichaelScrivenmayperhapshavehadthisinmindwhen,inthearticleIhavealreadycited,heheldthatcauseandeffectarenot(insuchcases)separate,butareonlyconceptuallydistinct.Wherevertheforegoinganalysisapplies,thesupposedll1ysteryofhowweperceivetheconnectionbetweencauseandeffectsimplydoesnotariseinanythinglikeitstraditionalform.Tobesure,thequestionofwhy,forexample,acomplexsetofmotionsisseenasacontinuousprocess,andnotasarandomsetofmotions,willstilldemandpsychologicalexplanation,butthisisnotaproblempeculiartothecauseeffectrelationship,asthosewhofollowHumehavegenerally CAUSATIONsupposed.Onthecontrary,itwillformpartofthemoregeneralproblemofthewaysinwhichstimulithatcomestouseithersuccessivelyorsimultaneouslyareorganizedintowholes.Hume'spuzzleastohowwecouldpossiblybesaidtoseeanyconnection'betweencauseandeffectaroseonlybecausehestartedfromtheassumptionthatinthecause-effectreIationshipwearealwaysdealingwithwhatwereoriginallyseenastwoseparatewholes.Becauseofthisassumptionhesoughtsomespecialrelationshipthatcouldaccountforthefactthatweseethemastiedtogether..When,however,onedoesnotassumethattwoseparateeventsareoriginallygiven,theproblemofexplainingourperceptionofacausalrelationtakesaquitedifferentform:WeIllUstseektoexplainwhatleadsustoseeaneventasasingleprocess,possessingaunityofitsown.Withrespecttomanyaspectsofthislargerproblemagreatdealofexperimentalevidenceisavailable,butnoneofittendstosupportHume'satomisticsensationalism.AmongthefactorswemaywithSOineassurancesaygiveunitytoanobjectortoaneventarethosethatweresingledoutforattentionbyGestaltpsychologistsininvestigatingsuchprinciplesoforganizationas"goodcontinuation"and"closure."Onemayseektoexplainsuchfactorsandtheirmodesofoperationinavarietyofdifferentways,dependinguponone'stheoreticalframework,buttheirprominenceintheperceptionoftheunityandstabilityofobjects,andoftheconnectionswithinwhatareperceivedascontinuousevents,cansurelynolongerbequestioned.Itisnotpartofmypurposetoattempttosortoutthevariousfactorsthatmayberesponsible,nortosayhowtheyoperateundervariantconditions:AstheexperimentsofMichotteindicate,evenslightchangesinthepatterningofthestimulileadtoashiftinwhethervisuallyapprehendedmovementsareseenastheuninterruptedmotionofasingleobject(asin"thetunneleffect")orwhethertheyareseenastwomotions,oneofwhichcausestheother.Nevertheless,thereisonetypeoffactortowhichIshallcallattention,sinceitisrelevanttoanothersetofexamplesofcausalperceptionwithwhichIshouldliketodeal.Thisfactor,whichispresentinmanycasesinwhichwehavetheconvictionthatweareseeingacausebringaboutaneffect,hasbeendesignatedbyothersasacorrespondencebetweenthecauseandeffect.12Toindicatewhatismeantbyacorrespondencebetweencauseandeffect,onecanpointtoinstancesofthefollowingsort:Whenahotobjectisappliedtoanotherobject,wearenotsurprisedthatthehotobjectshouldheatandnotcooltheother.Italsoseemsentirely CAUSALBELIEFSINEVERYDAYLIFE59naturalthatmarksleftbysomethingblack,suchascharcoal,shouldbeblack,orthatimpressionsinthesandshouldhavecontoursstructurallysimilartotheobjectsthatmadethem.Similarly,whenoneobjectstrikesanotheritdoesnotappearpuzzlingthatthemotionofthesecondshouldcontinueinthesameline,andthattheharderitisstruckthegreaterthespeedexpected.Ofcourse,inanygiveninstancetheseexpectationsmaynotbefulfilled;infact,wemaybeextremelysurprisedbytheunexpectedwaysinwhichoneobjectwillbeaffectedbyanother.Allthatthefactorofqualitativesimilarity(whichhasbeencalled"correspondence")helpstoexplainiswhyinsomecasesitiseasierthaninotherstoseeasequenceofeventsashavingcontinuityandunity.Thatfactitdoeshelptoexplain.13IcomenowtothenewsetofexamplesforthesakeofwhichIintroducedthisbriefdiscussionofcorrespondence.IntheseexamplesIaminterestedinthefactthatineverydaylifeweoftenthinkofacomplexsetofeventsasacausallyrelatedserieswhentheypossesssimilaritiesinstructureeventhoughwedonotnecessarilyrelateeachtotheothersinasimplelinearpattern.Insuchcasestheirsimilaritiesservetounifyandconnectthemwithoutourbeingawareofdirectconnectionsamongthespecificpartsofthewhole.14Inthesecasesitisasifthequalitativesimilaritiesamongtheeventswereexpressiveofsomeunderlyingcausalfactorthatmanifestsitselfineachofthem.ForasimpleillustrationofthisphenomenonwemayconsiderDuncker'sexampleoftemporalcorrespondenceswithrespecttoform:"Therhythmofsoundsofknockingcorrespondstotherhythmofthemotionsofknocking."15Inthisconnection,considerasimplerhythmicpatternofthreeknocksonadoor,thefirsttwobeingcloselyspaced,withaslightintervalbeforethethird.Regardlessofwhetheritiswewhoareknockingorwhetherwearewatchingsomeoneelseknock,wedonotrelateeachknockingmotiontoitsownsound,butthethreeknocksaregroupedinapatternthatalsocharacterizesthesound.Itisthissimilarityinpatternthatconnectswhatisfeltorisseenwithwhatisheard.ThatapatterncanplaythistypeofrolewillbeimportantinthosemorecomplexcasesofcausalconnectionInowwishtodiscuss.Whatitisimportanttonoteinthesecases-whichItaketobefrequentinoureverydaylife-isthatacomplexpatterningofphenomenaisoftenexperiencedasexpressiveofsomecausalpowerthatwedonotdirectlyexperiencebutcanonlyname.Take,forexample,theexperiienceofwatchingasummerthunderstormgatherandbreakonahot,stillday.Thereisoftenabriefmomentwhenonefeelsabreezespringup,andtheaircoolsandthe 60CAUSATIONleavesrustle,andsoonbranchesbend,thentossinthewind;thereisthunder,andtheraincomes.Thereishereaclearsequentialorderofwhichweareaware.However,wedonotexperiencethisorderasbeingasinglelinearprogressionfromatobtoctod.Therisingstormisfeltthroughalltheseeffectstogether:Thesuddenbreezethatcoolstheairalsorustlestheleaves,buttheirconnectionisfeltandseeninthechangeallaboutus,notthroughtheknowledgethatthebreezethatIfeelalsorustlestheleaves.Whatgivesunitytowhatisexperiencedinwatchingastormgrowisacrescendoofchange:changeintheairandinthelightandthesky,butaboveallintheincreaseofmoven1entandsoundasthestormgathersandbreaks.Thispatterningofeventsseemsexpressiveofcausalforceswithoutourbeingclearlyawareofwhattheforcesare:Itis"thestormitself"thatbringsthewindandtherainandthethunder,andwedonottracetheindividualrelationshipsthatexistamongtheseaspectsofit.Thisisquitedifferentfromwhatwefoundinourearlierexamples,whichmaybesaidtohavebeenseriallyordered.Inthoseearliercasesitwaspossibletotracehoweachsuccessivephaseofanongoingprocesswasapartinthatprocess.Forexample,inwatchingasuccessfulthree-cushionshot,·wefollowthelinethatthecueballfollows:SetinIllotionbythestroke,ithitstheredballandthenacushion,caromsofftohitasecondandthenathirdcushion,andtouchestheopponent'scueballbeforecomingtorest.Whateverphaseofthisprocessonemightwishtoanalyzefurther,onewouldstillremainwithinthissingle,causallyconnectedseries.Inthecaseofthestorn1,however,,vhatisexperiencedascommontotheindividualelelnentsthatenterintoitsstagesisthateachisincludedwithinadevelopingpattern,thateachispassingthroughatransformationsimilartothatwhichisalsocharacteristicoftheothers.Insuchcaseswhatisexperiencedmaybetterbedescribedas"orchestrated"relationsratherthancausalrelations.Nevertheless,insuchcasesweareawareofvhatappearstobesomecausalpower.Notexperiencingitinthespecificelements,butonlyinthepatternthatiscreatedbythem,weareledtoattributethatpowertoamere"something"underlyingthewholeseries;butwhatthat"something"Illaybewecannottrulysay.Insofarasdirectexperienceisconcerned,wecanonlysaythatitwasthestormitselfthatcausedthewind,thethunder,andtherain.16Inowturntoanothersortofcaseofthesamegeneraltype,inwhichwefindourselvesinthepresenceofcontinuitiesthatweexperienceasexpressiveofcausalforces,thoughwearenotdirectly CAUSALBELIEFSINEVERYDAYLIFEawareofthefactorsonwhichthatcontinuitydepends.AmongsuchcasesarethoseinwhichapatternofIllotionsappearsasdetermined,yetasdeterminedbyasourcethatremainshiddenfromus.Forexample,inwatchingadogeagerlyfollowingascentweseeapatterninhisaction-hisnosetotheground,hisbodytaut,rushingalongatrailweourselvescannotsee,losingit,sniffingabout,findingit,followingit.Weareawareofpursuit,butwhatgivesunitytothedog'sactionisnotanypartofwhatwedirectlyperceive.Inthiscase,however,weknowthatdogsdofollowscents,andthisknowledgehelpsustointerpretwhatwesee.However,evenwithoutthatknovledgewewouldseethedog'sbehaviorasgoal-directed.Thisisevidentinotherinstanceswherewhatisseenasgoal-directedbehaviorleadsustolooktowardthegoalinordertodiscoverthecauseofthebehavior,aswhenweseepedestriansonacrowdedstreetsuddenlyturningtostareinaparticulardirection.Similarly,incasesinwhichthegoal�directednessappearsasbeingnottowardbutawayfromsonlething,thatbehaviorisusedasawayofascertainingitsunknowncause,aswhenweattempttolocatethesourceofananimal'spainordiscoverwhatachildisafraidof.Theuniversalityandimportanceofthisfactorininterpretinghumanbehaviorcanscarcelybeoverestimated,butthatisnotmypresentconcern.Iusetheseillustrationsmerelytoshowthatwecanbeawareofthepresenceofunseenforceswithoutdirectlyexperiencingthem.Ourawarenessofthemrestsonthefactthatthecomplexmotionswedoseehaveacommonfactor:Theyareseenasbelongingwithinapatternthatrunsthroughthemall.Intheimmediatelyprecedingillustrationsthepatterningdependeduponthegoal-directednessofthemotions,butthisneednotalvvaysbethecase.Forexample,twofiguresorshadovvsapproachingandrecedinginrhythmicfashion,asifdancing,giveusastrongimpressionofbeingconnected,eventhoughweseenoconnectionsbetweenthem.17Simi�larly,inhumansituationssuchasexhibitionsofballroomdancingintheAstairefilms,whenthepartnersgivenosignof"leading"or"following,"itsometimesappearsasifthedancingweredonethroughthem,ratherthanbythem:Thepatternseemstohaveaforceofitsown.Thisisoftenevenmorestrikinglydisplayedbylargechorusesofdancersfollowingintricatelypatternedroutines,orbyexpertdrillteams.StillanotherexaInpleofhowpatternedmotioncangiveusastrongimpressionofsomeunderlyingforcewedonotdirectlyperceiveistobefoundinsomeofthewaysinwhichtheoceanappearstous.Inthe CAUSATIONrisingandswellingandsinkingofwaves,therhythmicpatternseemstoexpressanunderlyingcausalforce.Similarly,iffromthebeachwewatchasinglewaverisingandswelling,crestingandbreaking,itsrhythmgivesusasenseofitsunderlyingpower,evenbeforeitcrashesonthebeach.Wereittobesupposedthatthissenseofitspoweristobeattributedtowhatweknowandnottowhatwesee,Imightpointoutthatweknowagreatdealaboutthepowerofapiledriverbuttherepeatedfallofitshammer,thesoundthismakes,thevibrationswefeel,aremerelyrepetitiveandarenotexperiencedasrhythmicallyunified;thus,theygiveusnosenseofanunderlyingpowersuchaswefeelwhenwewatchthecrashofawave.Theimmediatelyforegoingsetsofphenomenologicalobservationsweredesignedtoshowthatwedosometimesexperiencewhatmaybedesignatedascausalpowerwithinacomplexeventevenwhenwedonotdirectlyexperienceanythingwewoulddenominateasthecauseofthatevent.Ihavesuggestedthatinsuchcasesitistherhythmicpatterningexpressedthroughthevariousaspectsoftheeventthatgivesustheimpressionofaforceunderlyingwhatispresentedtous.Iftheseremarksareaccepted,wehaveseenanotherreasontodoubtaHumeantypeofanalysisofourcausalbeliefs.Inthesecases,nolessthanincasessuchasthoseinwhichweseeonefootballplayerbringdownanotherwithahardtackle,thecausalrelationisnotonethatexistsbetweentwoseparateevents,onesucceedingtheother;rather,intheparticularcaseswithwhichIhavejustbeendealing,itisthepatterninginaseriesofeventsthatleadsustosupposethatbehindthoseeventsacausalpowerisbeingexerted.ThisdoesnotsignifythataHumeantypeofanalysismaynotbeapplicableinmanyothercases;itistoanexaminationofsuchcasesthatInowturn.II.ONREGULARSEQUENCESIhavealreadymentionedonecase-thatofswitchingonalightinwhichHume'saccountwouldseemtobeadequate.Inthatcase,thecauseanditseffectappearasdistinct,withoneoccurringjustaftertheother,andwecanreadilypictureeitherthecauseortheeffectasoccurringwithouttheother.Furthermore,insuchacaseHumeisentirelycorrectinholdingthatwecannotindicate,onthebasisofdirectinspection,whatitisthatconnectstheflickingoftheswitchandthelight'scomingon.Theawarenessofaconnectionbetweentheseeventswouldthereforeseemtobewhollydependent CAUSALBELIEFSINEVERYDAYLIFEuponpastexperience.Humeisalsocorrectinholdingthatafterrepeatedexperiencesofasin1ilarsortsomeintimateconnectionseemstodevelopbetweentheprioreventandwhatfollows.Itmustbeadmittedthatthereareindeedagreatmanycasesinwhichpast,experiencethusengendersconvictionsof"necessaryconnection."Ontheotherhand,therearemanyinstancesinwhichthesamefeelingofnecessaryconnectionneverarises.Forexample,experienceestablishesthatbreadnourishesandthatplantsneedlightandmoisturetogrow.Suchgeneralizationsarederivedfrompastexperience,butinthesecaseswedonotdevelopanyinclinationtoholdthattheconnectionbetweenthecauseanditseffecthasbeendirectlyexperienced.Thus,thesecasesdifferfromthoseinwhichtheeffectsofhabituationaresostrongthatthecausalrelationseemstobedirectlyexperienced.Infact,thereisawholespectrumofcasesrangingfromthoseinwhichpastexperienceengendersthefeelingthattheconnectionisexperienced,toonesinwhichwewouldnotInakesuchaclaim,eventhoughwebelievein,andrelyon,acausalconnectionbetweenwhathappenedatonetimeandwhathappenedmuchlater.Theexistenceofthesevariationsdoesnot,Ibelieve,overthrowwhatthepreviousargumentattemptedtoestablish:thatcauseandeffectarenottobeconstruedasdistinctevents,butaretoberegardedascomponentswithinsomesingleongoingprocess.Toshowthatthisisthecase,IshallconsideranumberofdifferentexampIes,and1shallstartwiththesortofinstancethatmightbethoughttobemostadvantageoustotheHumeanformofanalysis:thatofswitchingonalight.Insuchacase,asIhavenoted,itisclearlypastexperiencethatleadsustoconnectthetwosuccessiveevents,sincewecannotdiscernanydirectbondofconnectionbetweenthem;nevertheless,ourrecognitionofacausalrelationbetweentheseevents,thoughitwasacquiredonthebasisofpastexperience,seemsnolessimmediatethanthatwhichispresentwhenthereisadirectperceptionofcausality.Ishallattempttoexplainonwhatbasissuchinstancestakeonthisimmediacyofrecognition,whereasitislackinginotherregularsequencesthatserveasabasisforcausalattributions.18Inordertopavethewayforananswertothatquestion,Ishallfirstaskwhyinsuchcases-unlikethosecasesinwhichIhaveclaimedthatthereisadirectperceptionofcausality-thecauseanditseffectareseenastwodistinctevents.Thereare,Ibelieve,threeperceptualcharacteristicsthat,eitheraloneorincombination,leadusinthesecasestoviewthecauseasdiscontinuouswiththeeffect.Thefirstof CAUSATIONthesecharacteristicsisthattheyarenotspatiallycontiguous.Forexample,Iflickawallswitchandaceilinglightcomesall.Unlikewhatispresentedwhentwobilliardballscollide,orwhenachildtracesalineinthesandwithastick,spatialcantiguityislacking.Thisisonefactorthatseparatestheevents,leadingustoseethenlasdistinct.Nevertheless,thisfactorisnotinitselfdecisive,sincethereareotherinstancesinwhichchangesoccurintwospatiallyseparatedobjectsandyettheyappearasbelongingtogethertoformasingleevent:Forexample,twopersonswalkingorrunningmaybeseenasconnectedintheiractions,withoneleadingandtheotherfollowing,orvithonepursuingtheother.Asecondfactorthatseparatesflickingtheswitchandhavingthelightscomeon,andwhichmakesthemappearaseventsthataredistinctandseparable,isanabsenceofwhat,followingDuncker,Ihavetermeda"correspondence"betweenthecauseandtheeffect.Amanual,mechanicalmovement,suchaspushingabuttonorflickingaswitch,bearsnodirectresenlblancetothesuddenappearanceofabrightlight.This,too,enhancestheirdistinctness.Nevertheless,itcannotbedoubtedthatgiventherequisitepriorexperience(forIamheredealingvithcausaljudgmentsthatdependuponregularsequence),wedolookupontheflickoftheswitchasimmediatelyconnectedwiththelight'scomingon.19Thus,theremustinthiscasebesonlefactorthatcompensatesforthelackofspatialcantinguityandlackofcorrespondence.Whatseemsofcrucialsignificanceinestablishingthisfeltconnectionisnotthesheernumberofoccasionsonwhichweourselveshaveswitchedonlightsorhaveseenthemswitchedon;rather,itiswhatIshallcalltheir"instantaneoussuccession."Solongasthecontextinwhichtheseeventsoccurpermitsustonoticetheimmediacyoftheirsuccession(asopposedtowhateverotherchangesaresimultaneouslyoccurringaroundthem),thetwoeventswillcometobeseenaslinked.Thatthisisthecaseshouldnotbesurprising:Asconditioningexperimentsshow,rapidtemporalsequenceisanextremelyeffectivebondinlinkingtwoevents.However,inorderthatweshouldseethetwoeventsasinstantaneouslysuccessive,eachmustitselfbeofshortduration,seemingtobealmost(ifnotquite)instantaneous.2oThismaybeillustratedinthefollowingway:IfwewerealwaystoturnonalightbymeansofSainecontrivancesuchasacrankorapulleythattookanappreciabletimetooperate,andifwhileoperatingitthelightcameon,wewouldlearnthroughexperiencetoexpectthelightwhenweworkedthecrankorthepulley.However,iftherewerenodistinc- CAUSALBELIEFSINEVERYDAYLIFEtive,quasi-instantaneouseventimmediatelyprecedingtheappearanceofthelight,suchasaclickofthecrankorahesitancyinthepulley,theexperienceofadirectlinkbetweenwhatwearedoingandthelight'sconlingonwouldbelacking.Ontheotherhand,weretheretobesuchaclickorhesitancy,wewouldregardit-andnotthewholeoperationofcrankingorpulling-asthecauseoftheeffect.Thus,fortheretobe"instantaneoussuccession,"boththecauseandtheeffectmustthenlselvesappearasquasi-instantaneous.Incasessuchasthatofturningonalightbyflickingaswitch,eachoftheforegoingcharacteristicsispresentandcontributestoourseeingthetwoeventsasdistinct,eventhoughtheyareexperiencedascausallyrelated:Theeventsareseparatedinspace,theyarequalitativelyverydifferent,andeachappearsasquasi-instantaneous.Thus,whatprovideslinkagebetweenthelllistheirexceedinglyrapidsuccession.Consequently,insuchcaseswecaninfactdowhatHumethoughtitpossibletodoinallcases:Wecan,withoutdistortingouroriginalexperience,considereacheventseparately,asisolatedfronltheother.Tobesure,instantaneoussuccessionwithoutpastexperiencewouldnotleadustolinktwononcontiguous,dissimilarevents.Nevertheless,afterrepeatedexperience,eventsofthistypeappeartobelinkedinawaynotfoundaillongeventsinwhichinstantaneoussuccessionislacking.Wehavetolearnfronlexperience,throughacomparisonofinstances,whichfoodsdonotagreewithus,howmuchwaterisneededbydifferenttypesofplants,andthelike.Allsuchrelationshipsarecausal,butthereisnodefinitelllonlentatwhichaconnectionbetweenthemappears.Thus,theirconnectionisnotexperienced,butisaconnectionwerecognizeashavingobtainedinthepast.Itisinsuchcases,wherenolinkageisfeltinanyoneparticularinstance,thatattentionlllOStreadilyswitchesfromwhatistrueinonecasetowhatistrueincasesofaparticulartype.21ContrarytotheassumptionsofHartandHonore,suchcasesarebynomeansconfinedtoscientificexplanation,nortotheexplanationofnonhumanevents:1'heyarisewithgreatfrequencyinallaspectsofeverydaylife.Tobesure,notancausalattribution�areofthissort,yetinourdailyexperienceweoftenexplainaparticulareventthroughcitingwhatgenerallyhappensinothercasesofthesametype.22Consider,forexample,howwelllayexplaintoachildthatthetoycarhedroppedinthewatersankandwaslost,buthissister>srubberballwasnot.Wewillassuredlyappealtosomegeneralrubricconcerningthingsthatfloatandonesthatdonot:Thingsthataremadeofrubberfloat,orthingsthatarelightandnotheavyfloat,etc. 66CAUSATIONInshort,weappealtogeneralizationsastohowthingsofagiventypebehaveinsituationsofacertaintype,andthisisourexplanationoftheparticularcase.Noristhistrueonlyofexplanationsweusetofendoffthequestionsofchildren:Imayexplaintoyouwhythecropsaregoodthisyearbycitingtheweather,whichisofasortthatalwaysresultsingoodcrops,orImayexplainwhyonebedofrosebushesishealthierthananotherbycitingthefactthatitwasfertilizedwithAandsprayedwithX,AandXbeingparticularlyeffectivewithrosesofthistype.Sinlilarly,withrespecttohumanconduct,amothermayexplaintoherchildwhyherhusbandbecameangrybysaying,"Youknowthatwhenyouactthatwayyourfatherloseshistemper."Explanationsthatrestonsubsumingaparticularcaseunderpreviouslyfamiliartypesofcasesaretobefoundinallaspectsoflife.Itwillreadilybeseenthatsuchexplanationsserveusefulpurposesandareoftensatisfactoryasanswerstoparticularquestions.However,eventhoughtheymayhaveadequatepragmaticjustification,theyoughtnottobetreatedasiftheyweresufficientasexplanationsoftheoccurrencesthat,inashorthandmanner,theyareusedtoexplain.Thisbecomesevidentassoonasonenoteshowmuchadditionalknowledgeispresupposedintheiruse.Forexample,onemustknowsomethingabouttheconditionsnecessaryforthegrowthofplantsandaboutthefactorsthatinhibitgrowthbeforeanexplanationinvolvinghotandhumidweather,orfertilizersandsprays,canprovideananswertowhyoneyear'scrop,oronegardenplot,wasmoresuccessfulthananother.ThesetruncatedexplanationsservetosupplymissingpiecesofinformationthatbelongwithinamorecOlnprehensiveexplanationandhelptocompleteit.Thisistheiruse,butalsotheirlimitation:Unlesstheotherpiecesofthepuzzlewerealreadyinplace,thetruncatedexplanationwouldfailasanexplanation.Becausethislimitationisoftennotnotedbycontemporaryphilosophers,orbecauseitisnottakenseriously,theclaimisfrequentlymadethatexplanationsarealwayscontext-determined.Thisclaimwouldbecorrectifoneweretoequatesupplyingamissing.pieceofinformation,orcitingarubricthatgenerallyappliesinacertaintypeofcase,withprovidinganexplanationofsomeparticularoccurrence.However,aswehavejustnoted,theseshorthandresponsesserveasexplanationsonlywhenonealreadypossessesaframeworkofrelevantfactualandtheoreticalknowledgeintowhichtheyfit,andwhichtheyhelptocomplete.Therefore,whatiscontext-dependentisnottheactualexplanationoftheoccurrence,butonlywhatparticularsortofinformationmay,underthecircumstances,beneededtocom- CAUSALBELIEFSINEVERYDAYLIFEpleteit.Theinformationweseek,orthatwesupplywhenwerespondtorequestsforexplanations,willdependuponwhatsortofpuzzlementapparentlyliesbehindthequestionposed.Bearingthisinnlind,onecanseethattruncatedexplanationswillalwaysbecontextdependent,withdifferentanswersbeingproperlygiventodifferentpersons,orupondifferentoccasions.However,thereisnoreasontosupposethatthesedifferingexplanationsareeitherincompatibleormutuallyindependent.Onthecontrary,whatprovidesthebasicstructureintowhichthevaryingshorthandexplanationsmustfitwillnotitselfdifferfromcasetocase,butwillbethesame.Therefore,itwillbelessmisleadingifinsteadofsaying"explanationsarecontextdeterlnined,"wesaythat"answerstoquestionsthatareaskedorimpliedarecontext-determined."Weneednotthenspeakasifthesameoccurrence(or,nlorestrictlyspeaking,thesaIneoccurrenceviewedonthesamescaleandwithrespecttothesamefacets)lnayhavetwocorrectbutquitedifferentexplanations;wecansay,lnoreaccurately,thatthetwoinquirerswerepuzzledbythesameoccurrencefordifferentreasons.Inotherwords,neitherhadgraspedthecorrectexplanation,butthefailureofeachwasduetoadifferentreason:Eachlackedadifferentpieceofthecorrectexplanation.Whatthecharacteristicsofagenuinelyadequateexplanationnlaybeisaproblemthatliesinthebackgroundofallthatistofollow.Atthispoint,however,Ishalllimitmyargument,seekingtoshowonlywhywecannotbesatisfiedbyshorthand,context-determinedexplanations,evenintheaffairsofeverydaylife.Onemajorsourceofdifficultyinsuchexplanationsisthattheyrestontheassumptionthatthecauseofaneventissomesimplepriorevent.Oncethisassumptionismade,itbecomesnecessarytodrawadistinctionbetweenthecauseofaneventandwhataremerelyconditionsthatmakefor,orpermit,itsoccurrence.Asthenextchapterwillshow,thatdistinctionbristleswithdifficultiesandshouldbeabandoned.Here,Imerelywishtqexposethefactthatevenwhenweareexplainingordinaryoccurrencesineverydaylife,thedistinctionbetweencausesandconditionsisartificialandfailstoestablishtheadequacyofourshorthandexplanations.III.ONCAUSALEXPLANATIONInexaminingtheproblemofwhetheronecaninfactdistinguishbetweenacauseandtheaccompanyingconditions,letusfirsttakeas 68CAUSATIONanexamplethefactthatonesometimesexplainsanexceptionallygoodharvestintern1Softhatyear'sprevailingweather.Although"thecause"istakentobetheweather,itisnonethelesstruethatawholesetofotherconditionsmusthavebeenfulfilledfortheharvesttohavebeengood.Sinceweassun1ethatfarmerswillhavetheskillandopportunityneededtoplantandtotendtheircrops,andwealsoassumethatthecropswillgrowwhenproperlyplantedandtended,weregardtheweatherasthemostvariableofthefactorsrelevanttothesuccessofacropinaparticularyear.Therefore,itistheweathervesingleoutasbeingresponsiblefortheharvest.(WhetherpeasantsduringtheThirtyYears'Warwouldhavedonesoisanotherquestion.)However,itshouldgowithoutsayingthatthequalityoftheseedthatissownandthejudgmentoffarmersinknowingwhenitisbestsown,theirknowledgeofhowtotendtheircropsandtheirabilitytodosounhamperedbysocialupheavalornaturalcatastrophes,areallequallyessentialtohavingasuccessfulharvestattheendoftheseason.Itisthereforemyclaim,whichIshalllaterdefendatgreaterlength,thatitisnotintheendpossibletodistinguish"causes"from"conditions,"thoughinsomeparticularcontextsofinquiry,whereagreatdealcanbetakenforgranted,itisoftenconvenienttodoso.Whatleadstotheeasyacceptanceofthesepragmaticoversimplificationsisthatthereare,aswehaveseen,agreatmanycasesofregularsequencesinwhichweviewacauseanditseffectastwodistinctandseparableevents,ratherthanaselementswithinasingleongoingprocess.Solongastheanticipatedsequenceoccurs,anyhiatusbetveenacauseanditseffectisnotdisturbing.However,inthoseinstancesinwhichwhathasusuallyoccurredfailstooccur,weareforcedtogobeyondthesesimplisticexplanationsandfindsomeconditionsthatareusuallypresentbutareinthiscaseabsent,orsomeconditionsthatareusuallyabsentbutareinthiscasepresent.Thus,shorthandexplanationsthattakeintoaccount"thecause"butnotaccompanying"conditions"explainwhatoccursonlyin"normal"cases;inso-calledabnormalcaseswelllustrefernotonlyto"thecause"butto"conditions."Thisshouldmakethedistinctionsuspect.Furtherlllore,whileshorthandexplanationsoftenserveourpurposes,evenineverydaylifeanyprudentmanwhowishestotesttheadequacyofageneralizationwilllooknottopositiveinstancesonly,butalsotowhatconstituteapparentexceptions.Astheoldadageholds,itistheexceptionthatproves-thatis,tests-therule.Considerthefollowingcases.Weflickaswitchandalightcomes CAUSALBELIEFSINEVERYDAYLIFEon,andweregardthefirstasthecauseofthesecond.Similarly,inallsocietiespeoplehavelearnedthatifsomethingsareeatentheywillcausesicknessordeath,thatotherthingshelpallaypainorinducevisions,andinallsocieties-.orinalmostall-itisrecognizedthatsexualintercoursecausespregnancy.Insuchcasesitwouldseemasifnothingwereinvolvedbeyondregularsuccession,thatthereneedbenoinherentconnectionbetweenwhatisregardedascauseandwhatisregardedasitseffect.Yet,suchcasesadmitofexceptions.Thelightwillnotcomeonifthepowerhasbeencutoff,oriftheelementswithintheswitchdonotmakecantact,orifthefilamentinthebulbisbroken.Onecanpackintothegapawholeseriesof"conditions"thatmustaccompany��thecause"inordertobringabouttheeffect,butwhatoneisthendoingissubstitutingtheseconditionsforthecantinuousprocessthatoccurswheneveralightisactuallyswitchedon.Similarly,whatiseatenandusuallycausesillnessordeathdoesnotalwaysdoso,nordoessexualintercoursealwaysresultinpregnancy,evenapartfromcontraception.Toexplainthedifferencesbetweencasesinwhichtheexpectedeffectfollowsandthoseinwhichitdoesnot,wemusttracetheprocessesinterveningbetweentheeventthatiscalledthecauseandthatwhichweregardasitseffect.Intracingtheseprocesseswequicklydiscoverthatwhatappearedastwodistincteventsfromourordinarycommon-sensepointofviewwereinfactsimplypartsofonecontinuousprocess:Forexample,thelightcomingonismerelytheterminalstateofwhatoccurredwithinthatelectricalsystemwhenweswitchedonthecurrent.Similarly,theingestionofapoisonsetsupprocessesinthedigestivetractthat,throughtheirconnectionwithotherorgans,insomecasesleadsdirectlyandwithoutinterruptiontodeath,whereasinothersitmayleadtovomitingandthentorecovery.Thedifferencebetweentheresultsinthelattertypeofcasesneednotbeadifferenceinthepoisoningested,butinthereactionsoftheperson'ssystellltothatparticularpoisonwhen,ataparticulartime,itwasadministeredinaparticularamount.Inthisexample,asintheprecedingone,itistobenotedthattheactualoutconle,or"effect"-whetheritwasdeathorrecovery-isnottobeviewedasaneventdistinctandseparablefromthatwhichbroughtitabout:Itissimplytheterminalstateofaprocesswhoseinitialphasewastheingestionofthepoisonandwhichcontinuedasthebodyabsorbedthepoison,withaconsequentseriesofdisruptionsoforganicfunction,endingeitherinvomitingandrecovery,orindeath.Whenwerefineourshorthandexplanationsinordertoaccountforthosecasesinwhichapparentlysimilarcauses CAUSATIONproducevarianteffects,thecausecannolongerbeseenasaprioreventdistinctfromthecontinuousprocessthatterminatesintheeffect,nordoesitInakesensetoseparate"thecause"from"theconditions"withoutwhichtheeffectwouldnothaveoccurred.Asecondandquitedifferentsortofdifficultyarisesifweattempttotreatthecauseasaspecificeventdistinctfromitseffect,ratherthanviewingbothaselementswithinasingleongoingprocess.Thisdifficultyconsistsinthefactthatwewould,asaconsequence,sometimesnotbeabletodistinguishbetweencasesinwhichitiswarrantedtosaythatthefirsteventisitselfcausallyrelatedtotheeffect,andthosecasesinwhichaprioreventismerelyasignaccompanyinganothereventthatisresponsibleforwhatoccurs.Ahackneyedexampleofthisproblemisprovidedbytherelationshipbetweenchangesregisteredbyabarometerandchangesintheweather.Forexample,arapiddeclineinbarometricreadingsistakenaspresagingastorm,butnotascausingit.Onemightperhapsclaimthatthisisbecausethereisnoinvariantexperiencedsequencebetweeneventsofthesetypes:Stormsareexperiencedbythosewhoknownothingofbarometers.Furthermore,othersignsofanoncomingstormarefamiliartous,whetherornotwehappentoobservechangesregisteredonbarometers.Thus,eventsofthesetypesarenotexperiencedasconstantlyconjoined.However,ontheregularityviewthiswouldnotbeanadequatereasonfordenyingthatchangesinbarometricreadingscausestorlllS:Thosewhoanalyzecausationintermsofregularsequencesdonotinothercasesholdthatafailuretonoteparticularregularitiesinsonlecasesshouldcountagainstabeliefthatthereisacausalrelationbetweentwotypesofeventsifwheneverwelook··forthissequenceweareabletofindit.Itisbeyondquestionthateventhoughwedonotalwayscheckabarometerwhenwenoticeweatherchanges,wecanfindregularityinthechangesofbarometricreadingsandchangesintheweatherwheneverwedomakeacarefulcheck.Itisthereforenecessarytoseeksomeotherreason·';.yvhythisregularityisnotacceptedasindicatingadirectcausalreiationshipbetweenchangesinbarori1etricreadingsandtheonsetofstorll1S.Thereasonlies,Isubmit,inthefactthatnothinginourpreviousexperienceofsequentialrelationshipshasprovideduswithabasisforseeingchangesinthesereadingsasaninitialstateinacantinuousprocessthatterminatesinastorm.Spatialcontiguity,correspondence,andinstantaneoussuccessionarealllacking.Yet,thereareothercasesi�whichwelearnthroughexperiencethattwostatesofaffairsarecausallylinked,evenwhenthesesignsoflinkagearenot CAUSALBELIEFSINEVERYDAYLIFE'present.Thedifferenceinthiscaseisthatthereisnothinginourexperiencetosuggestthatanobjectsuchasanindicatoronabarometerdirectlyaffectswhatoccursintheenvironment.Apartfromthehumanpurposesitisdesignedtoserve,anyslightchangeinthepositionoftheindicatorwillappearasbeingInerelyaneffectofsomethingelse,andnotasthecauseofanyfurthereffects.Infact,itisonlyonthebasisofcomplexscientifictheoriesthatweunderstandtheconnectionbetweenreadingsonthissimplehouseholdinstrulnentandtheonsetofastorm.Theconnectinglinksarenotdirectlyobserved,butinvolveacomplexsetofprocessesofwhichwelearnonlywhenwegobeyondoureverydayexperience.Inordertogiveacausalexplanationoftheirlinkagewewouldhavetosaysonlethinglikethefollowing:ChangesinpressureresultinchangesinteInperature,andundertheseconditionsamovementofairsetsin,thecoolingofn10istairleadstocondensation,andthereforetoadownpour.Inallofthis,throughscientifictheories,wearetracingacontinuousprocessandarenotsimplyspecifyinganumberofdifferentconditionsthatarepresentinoneplaceatagiventilne.Itistheexistenceoftheprocessthatpermitsustosayhowthechangesinbarometricreadingsareconnectedwiththestorm:TheyareconnectedinsofarastheythemselvesareeffectsofthesaInechangesinatmosphericpressurethatresultinthestorm.23Intheforegoinginstanceitiseasytoseewhynooneislikelytobemisledintothinkingthatchangesinbarometricreadingscausestorms.Inthefirstplace,asIhavepointedout,spatialcontiguity,correspondence,andinstantaneoussuccession-eachofwhichcancontributetoaperceptionofcausality-arealllacking.Inthesecondplace,wewouldnotevenknowhowtointerpretubarometricreadings"withoutthesortofknowledgethatmakesitevidentthatbarometersareconstructedtoserveasindicatorsofcertainspecificprocesses,andnottocauseotherphysicalprocessestooccur.Never�theless,therearesomeinstancesinwhicharegularsequencebetweentwotypesofeventsmayleadonetoregardthefirstashavingacausalreIationtothesecond,eventhoughthefirstisinfactonlyanindi�catorthatonemayexpectthesecondtooccur.Suchcasesaresubsumedunderthefamiliarfallacyofposthocergopropterhoc.Thatthereisarecognizedfallacyofthistypearguesstronglyagainsttheregularityview,sincethefallacyispresentwhenonetakesasignforacause,eventhoughthesignregularlyaccompaniestheeffectandservesasanindicatorthattheeffectwillensue.Itwould,forexample,beaninstanceofthisfallacyifaphysicianweretofail CAUSATIONtodistinguishbetweenthesymptomsofadiseaseandthefactorsresponsibleforthecoursethediseasefollows,tryinginsteadtoestablishcausalconnectionsamongthevarioussymptoms.Eventhelaymandrawsthisdistinction.IfIsuddenlyfeelfeverish,andthenextdaynlyjointsache,mybreathingisconstricted,andIdevelopacough,IdonotattributethesetoInyfever,nortoeachother.Unlikeaphysician,Ifilaynotunderstandtheconditionsonwhichtheseeffectsdepend;nevertheless,thefactthatIhavesooftenexperiencedthemtogether,inthisparticularconstellation,doesnotmakenleregardthemascausallyrelatedtoeachother;instead,theyaretakenbynletobesynlptomsofprocessesgoingonwithinme,thenatureofwhichIdonotdirectlyexperienceorunderstand.24Inthiscase,asinothers,vemustdistinguishsymptomsfromcauses;inotherwords,sonleparticulareffect,orsomeseriesofeffects,mayserveasareliablesignofanongoingprocesswithoutitsbeingtruethatthisparticulareffectcausesthatprocessorthattheseriesofeffectscauseoneanother.Itisinlportanttonotethatindrawingthisdistinctionbetweenthecauseofaneventandwhatmerelyservesasasignindicatingthatsomeeventhasoccurred,isoccurring,orwilloccur,wedonotremainonthelevelofthoseeverydayshorthandexplanationsinwhichcausationpurportedlydependsuponregularityofsequencebetweentwotypesofevents:Evenwithinoureverydaypersonalexperience,whenwearenotrelyingonanyadvancedtheoreticalanalyses,wepresupposeconnectionsthatwedonotactuallyobserveintheinstanceimmediatelyathand.Thus,itisnotinthesciencesonly,butineverydaylife,thatexplanationsofeventsreachdeeperintothenatureofongoingprocessesthantheregularity-of-sequenceviewwouldleadonetosuspect.Wehavenowseenthistobetrueintwosortsofcases.Inthefirsttypeitwastrueofcasesinwhichaparticularexpectedeffectdidnotoccur,aswhenweflickedalightswitchandthelightfailedtogoon;wehavenowfoundthatitisalsotruewhenwedistinguishbetweencausesandindicators,orsigns.Incasesofbothtypes,thenewlevelofcausalexplanationthatmustbereachedisnotoneinwhichweareinvolvedinconnectingaseriesofdistinctandseparableevents,butoneinvhichwearetracingacontinuous,ongOIngprocess.Inadditiontotheprecedingsortsofdifficulties,whichwouldariseifoneattemptedtoremainonthelevelofexplanationinwhichthecauseofsomeoccurrenceisviewedasadistinctandantecedentevent,wemaynotethatinanyadvancedscienceanexplanation CAUSALBELIEFSINEVERYDAYLIFE73consistsnotincorrelatingtheoccurrenceoftwotypesofevents,butinfindingsystematicconnectionsamongvariousfactors-suchaspressureandvolume,ormassanddistance-thatareapplicabletoeventsofthemostdiversetypes.(SeeAppendixAformycriticismofHumewithrespecttothispoint.)Inotherwords,advancedscientificexplanationsconsistinstatingfunctionalrelationsthat,whentakeninconjunctionwiththeinitialandboundaryconditions,servetoexplicatewhathashappenedintheparticularcaseathand.Explanationsthatmerelycorrelateonetypeofoccurrencewithsomeoccurrenceofanothertypelacktheexplanatoryandpredictivepowerthatanyadvancedsciencepossesses:Theyfailtomakeclearhowthesamefunctionallawsapplywithequalforcetoverydiversetypesofoccurrences,thusofferinguniforlnexplanationsforeventsthat,onthesurface,appeartobeutterlydisparate.Inshort,thesciencesarenotconfinedtoexplainingeventsaccordingtorubricsthatconnectparticulartypesofeventsonthebasisofaseriesofobservedcorrelations:Thetheoreticalcomponentofanyreasonablyadvancedsciencedemandsthatthismodelofexplanationbeabandoned,andwithitsabandonmenttheHumeanviewofcausationnecessarilydisappears.Itmustnotbeassulned,however,thatthemodeofexplanationcharacteristicofanadvancedstageofsciencedoesnothaveanaloguesinsomeoftheexplanationsgivenineverydaylife.25Alnongtheinstancesonemightcitearethoseinwhichweexplainthewayanobjectbehaveswhenplacedinaparticularsetofcircumstancesbyappealingtooneofitsqualities,ratherthanbycitingsomeantecedentevent.Ihavealreadymentionedonesuchcaseinpassing:Weexplaintoachildwhyatoycarsinksbutarubberballfloatsbyappealingtothefactthat"thecarisheavier."Tobesure,exceptionsandcounterexamplestosuchanexplanationquicklypushustomoreadequategeneralizations,andultimatelytotheArchimedeanlawofdisplacement;however,onacommon-senselevelweoftenuseshorthandexplanationsofthistype.WhatisofinterestabouttheminthepresentcontextisthefactthatthecauseisnotSOlneantecedenteventdistinctfromtheeffect:Itisintermsofthepropertiesoftheobjectitselfthatweseektoexplainhowitbehaves.Thisisquiteobviouslytrueininnumerableexplanationsineverydaylife.Weexplainsomeaspectsofthebehaviorofaparticularanimalthroughreferencetothekindofanimalitis;or,tochoosemerelyoneotherexalnple,weexplainwhatoccurswhenweputsugarinourcoffeebyreferringtothesolubilityandthesweetnessofthesugar.Tobesure,someantecedenteventmayhavebeennecessarytoevoketheparticularin- 74CAUSATIONstanceofanimalbehaviorinwhichweareinterested,justasitisnecessaryformetodropthelumpofsugarinmycoffeetosweetenit;nevertheless,thisdoesnotmeanthatcitingtheantecedentservestoexplaintheparticularaspectoftheeventinwhichweareinterested.Rather,weareinterestedinwhy,underthesecircumstances,thisanimalbehavesinawayinwhichanimalsofotherspeciesdonot,justasweareinterest.edinwhylumpsofsugarsweetencoffeebutpebblesdonot.Crudeassuchexplanationsmaybe,thattowhichweappealishowobjectsofthistype-thatis,thosehavingcertainproperties-behaveundercertaintypesofcircumstances.Ingoingbeyondthesecrudeshorthandexplanations,whichresemblethosesatirizedbyMoliere,wemustofferfurtheranalysesofboththerelevantpropertiesoftheobjectsandtherelevantconditionsinthecircumstances.Suchanalysesdonot,however,leadusbackwardintimetosomepriorevents:Theyareanalysesofhowthepresentpropertiesoftheobjectarefunctionallyrelatedtootheraspectsoftheconditionsthatarealsopresentwhentheeventoccurs.ThispointcanbeillustratedbycitingHume'sownexampleofthecollisionoftwobilliardballs.Ifourconcernisnotconfinedtothegeneralcharacteristicsofoccurrencesofthistype,butifwewishtoexplainwhatactuallyoccurredinoneparticularinstance,Hume'sanalysisofthecausalrelationquicklyprovestobeinadequate.ThewayinwhichtwobilliardballsbehaveuponimpactisnotalwaysthesaIne:Itvaries,forexample,viththeaIII0untandthedirectionofthespinimpartedtothecueball,asisevidentinthedifferencebetweenshotsthat"draw"andthosethat"follow."Letusalsonoticethatdifferentbilliardballs,eventhoughtheybehaveinroughlysimilarways,donotbehaveinanidenticalmanner,forsomearemadeofivoryandothersoflesselasticmaterial.Experienceisundoubtedlyessentialifwearetolearnhowobjectsofdifferenttypesbehave,andalsohowtheeffectsoftheilnpactofonebilliardballonanothercanbemodifiedbythespinthathasbeenplaceduponit:IalllnotherearguingagainstHun1e'sappealtopastexperienceasthesourceofourknowledgeoftheseconnections.WhatIwishtopointoutisthathisassumptionthatthecausalrelationisalinear,sequentialrelationbetweentwodistincteventsismistaken:Ananalysisofthecauseofaparticularoccurrenceinvolvestracingthevariousfactorsthatarejointlyresponsiblefortheoccurrencebeingwhatitwas,andnotbeingdifferent.Suchananalysisisinfactananalysisoftheoccurrenceitself,anddoesnotleadustosearchforantecedent,independentcauses.Infact,asourpreviousargumentsweredesignedto CAUSALBELIEFSINEVERYDAYLIFE75show,solongasweremainonthelevelonwhichwesaythatacauseanditseffectappearastwodistinctandseparateevents,wecannotformulategeneralizationsthatwilladequatelyexplain(a)whyinsomecasesaregularsequence,whichisexpected,failstoobtain;nor(b)howwearetodistinguishbetweenwhatcausesaneffectandwhatislnerelyanaccompanyingsignofacausalrelation;nor(c)whyitisthatdifferentobjectsthatareinmanyrespectsverysinlilardonotalwaysreactinthesamewaywhenplacedinsimilarcircumstances.Thesereasonsshould,Ibelieve,besufficienttoleadonetoabandonHume'sviewofthecausalrelation,onceonehasbrokenwithhispsychologicalandepistemologicalassumptions.InthenextchapterIshallfurtherdevelopthealternativeviewIhaveherebeensuggesting,byshowinghowthatviewappliesinthecaseofscientificexplanations.First,however,itwillbewelltodrawtogetherwhathasbeensaiduptothispointconcerningourcausalbeliefsineverydaylife.IV.CONCLUSIONSInsunllnary,itshouldbeobviousthatourcomlnon-sensecausalexplanationsareofverydifferenttypes.Inthefirstplace,aswehaveseen,therearecasesinwhichthereisdirectperceptionofwhatappearsasacausalrelation;ontheotherhand,theconvictionthatthereissucharelationbetweenoneeventandanotheroftendependsnotupondirectperceptionbutuponourhavingdiscoveredthatthereareregularitiesinsequencesofthistype.Inthesecondplace,therearedifferencesbetweenthosecasesinwhichacausalrelationappearstoobtainbetweentwoeventsthatareseenasdistinctandinprincipleseparable,andthoseinwhichcauseandeffectappearascomprisingdifferentaspectsofonesingleevent.Furthermore,wehaveseenthattherearecasesofothertypes.Forexample,somecomplexoccurrencesappearasdependentuponacausewedonotdirectlyexperience,butwhosepresenceisadumbratedthroughapatterninginwhatweexperience.Inaddition,wehavenotedthattherearecasesinwhichaneffectisregardedasbeingdependentuponSOineparticularpropertyofanobject,ratherthanasbeingaresultattributabletosomeantecedentoccurrence.Thesedifferencesinthevarioustypesofcausalattributionsineverydaylifewillfirstoccupyus.Weshallthenturnourattentiontotheconclusionsthatfollowfromourargumentthatinallcaseswhenthesecasesarepressed-causalexplanationsleadustoviewa CAUSATIONcauseanditseffectaslinkedtogetherinsuchawaythattheymaybesaidtoconstituteaspectsofasingleongoingprocess,ratherthanbeingdistinctevents.A.ConcerningtheTypesofCausalExplanationsinEverydayLifeBecauseoftheinfluenceofHume'sanalysisofthecausalrelation,andbecauseofthepersistenceofpsychologicalassulnptionssimilartothoseheInadeconcerningwhatwecantakeas"given"insenseperception(asdistinctfromwhatmustbeassumedtobecontributedbyactivitiesofthelnind),itisstilldifficulttoconvincemostphilosophersthattherearecasesinwhichacausalrelationshipisdirectlyperceived.Perhapsthefollowingclarificationwillhelptodispelthatreluctance.Inmanyprin1itivesocietiesthroughoutthevorld,onefindsbeliefsidenticalwith,orsin1ilarto,thebeliefinmana,asecretorspiritualpowerpresentinSaIneobjectsorperson,butnotinothers.26Manacanflowbetweenobjectsandpersons;itcanbeacquiredbytouchingoreatingthatwhichpossessesit;itspresencebringsstrength,anditslossmeanslossofpower.Thoughweviewthisasaninstanceofsuperstitiousbelief,vestigesofthatbeliefarefoundinawidevarietyofphrases,actions,andbitsoffolklorestillevidentinoursociety.If,forexample,SOlneonehasachievedsomenotablesuccess,wemayjokinglysay,"Letmetouchyou."Furthermore,insomecasesthisbeliefisconnectedwiththenotionof"correspondence,"wherecauseandeffectarelinkedbyaqualitativesimilarity.Forexample,childrenaretoldthatcarrotsvillgivethemrosycheeksandcurlyhair,justasinprimitivesocietiesthemeatofabearhasbeensupposedtogivestrength,andmeatofdeerhasbeensupposedtomakehuntersfleet.Beliefinsuchdirecttransferencesofqualitiesis,ofcourse,superstitious;andifoneweretoexplainourcausalbeliefsinaHumeanwaysuchsuperstitiousbeliefsshouldneverhavearisen,preciselybecausetheyaresuperstitious:Beingfalse,theycouldnotpossiblyhavearisenthroughanyobservationofregularlyrepeatedsequences.Theirsourcemustthereforelieelsewhere,anditliesatexactlythepointthatwehaveseentobemostessentialintheperceptionofcausality:aqualitativecorrespondencebetweenwhatistakenascauseandaseffect.TIle'factthatsuchqualitativecorrespondencessometimesleadUStobelieveinconnectionsthatdonotexistshowsthatweshouldnotinallcasesrelyonwhatwedirectlyperceive.This,however,doesnot CAUSALBELIEFSINEVERYDAYLIFE77signifythatwedonotperceiveit,northatitmaynotbeareliableformofperceptioninothercases.Forexample,undercertaincircumstanceswethinkthatobjectsinourenvironmentaremovingwhenitiswewhoaremoving,andunderothercircumstancesthereverseisthecase;however,suchcasesofinducedmovement,thoughdeceptive,donotprovethatourperceptionofmovementisunderallcircumstancesunreliable.So,too,inthecasesofourperceptionofcausality.AsMichottehasshown,ourperceptionofcausalrelationshipsbetweenlllovableobjectssometimesaccuratelydepictsrelationshipsthatholdinnature,andsometimestheseperceivedcausalrelationshipslookequallyconvincingbutgoagainstallphysicallawsconcerningwhatoccursuponimpact.Inbothcasesweperceivearelationshipthatweregardasacausalrelationship,butinonecasetheperceptionisconsonantwiththelawsofmotionandintheothernot.Itishere,ofcourse,thatregularityofsequencebecolllesiUlportant:Itisatestthatisusedineverydayaffairswhenwecheckonthereliabilityofthecausalattributionswemake,andwhichothersalsomake.Givenaconstantconjunctionofonetypeofeventandanother,weexpectthesecondwheneverthefirsthasappeared,andnlanyofourshorthandexplanationsofcausalconnectionsineverydaylifedependuponourhavingexperiencedsuchconstantconjunctions.Nevertheless,aswehavehadoccasiontonote,thesatisfactorinessoftheseshorthandexplanationsisoftendependentuponthefactthattheyfitintoabackgroundoftheorythatallowsustoconnectthecausewiththeeffect:Theconnectionisnot,asHumeclaimed,27simplyamatterofhavingobservedaconstantconjunctionbetweenthetwotypesofevents.Furthermore,aswehaveseen,suchshorthandexplanations,thoughtheyareoftenserviceableineverydayaffairs,.tendtobreakdownanddemandsupplementationwheneverwehavetoexplaincasesinwhichtheconjunctiondoesnothold,orcasesinwhichwedonotregardconstantconjunctionasindicatingadirectcausalrelationbutonlyarelationshipinwhich,becauseofsomegenuinelycausalfactor,oneeventservesasasignoftheother.Finally,inthosecasesinwhichweattributethebehaviorofanobjectundercertaincircumstancestothepropertiesitpossesses,ratherthantowhateverantecedentoccurrenceledtoitsbeingplacedinthesecircumstances,aHluneananalysisofwhatistobetakenasthecauseofagiveneffectisnotrelevant.28Yet,indailylifewedoatten1pttoexplainmanyeffectsintermsofthepropertiesinherentinobjectsthatactinaparticularmanner,andthescientificsearchforinvariant CAUSATIONfunctionalrelationshipsisanextensionofsuchcommon-senseattempts.BearingthesepointsinInind,twogeneralconclusionselnergeconcerningthetypesofcausalexplanationsthatarecharacteristicofeverydaylife:First,itshouldbeobviousthattheseexplanationsarediverse,andshouldnotinallcasesberegardedasrepresentingadirectperceptionofcausality;nor,ontheotherhand,shouldtheyinallcasesbethoughttobedependentonobservationsofsequentialregularities.Second,theseillustrationsshowthat,exceptincasesinvolvingthedirectperceptionofcausalityinaparticularinstance,astrongadmixtureoftheoryisapttoenterintooureverydaycausalexplanations,andthiselen1entInaywellbecontinuouswiththekindsoftheoreticalframeworkfoundinthesciences.Therefore,itshouldnotbeassumedthattherearefundamentalandirreconcilabledifferencesbetweenscientificgeneralizationsandcausal'explanationsineverydaylife.B.TheUnderlyingCOlnmonFormofCausalExplanationsTheillustrationsthroughwhichIhaveattemptedtoleadthereadertotheforegoingconclusionswerealsointendedtoshowthattheconventionalforlnofdescriptionofthecause-effectrelationshipisinadequate:Onlyinshorthandexplanationsofaspecificeffectisthecauseofthateffecttakentobeanantecedenteventthatisnotitselfaphaseinthesingle,ongoingprocessthattern1inatesintheeffect.And,aswehaveseen,oncethetruncatednatureoftheseshorthandexplanationsisunderstood,andtheexplanatoryschemasonwhichtheyrelyarebroughttolight,theytoocanbeseentoshareacommonform.Todrawtogetherwhathasbeensaidaboutthatforn1,itwillbehelpfultorecallonepragmaticfeatureofanycausalexplanation:thateachsuchexplanationhasasitspointofdeparturetheobservationofsomethingconsideredasaneffect,theexplanationrequiredbeingonethatanswersaquestionastowhythatparticulareffectoccurred.Thepragnlaticaspectofthissituationresidesinthefactthatwhenweaskwhyaparticulareffectoccurred,wearetreatingagivenoccurrenceundersomeaspect,andnotwithrespecttoallofthewaysinwhichitmightbedescribed.Thus,whenweinquireintothecauseofaneffect,wearealwaysaskingforitscauseundersomeparticulardescription.Sinceitisundoubtedlytruethatpersonswithdifferentbackgrounds,knowledge,andinterestswillviewthesame CAUSALBELIEFSINEVERYDAYLIFE79concreteoccurrenceindifferentways,theywillchoosedifferentaspectsofthatoccurrenceasstandinginneedofexplanation.TIlus,whatquestionsareasked,andwhateffectsneedtobeexplained,willdependuponwhoitiswhoasksforcausalexplanations.Itmightatfirstglanceseemthattheintroductionofthispragmaticdimensionwouldleadtoachaosofconflictingexplanations.Infact,justthereverseisthecase.Inthefirstplace,itispreciselythispragmaticdimensionthatkeepsusfromanendlessquestinoursearchforcauses:Whileeachanswertoanyspecificcausalquestionmayraisefurthercausalquestions,thisdoesnotmeanthattheinitialquestionwasnotadequatelyanswered.Furtherlnore,therewouldbenoreasontoexpectconflictbetweentheanswerstothevariouscausalquestionsthatcouldberaisedwithrespecttoSOlneconcreteoccurrence:Solongaseachquestionreferredtosomeactualaspectoftheoccurrence,andsolongasthesequestionswerenotconfusedwithoneanother,thecausalexplanationsshouldbenolesscOlnpatiblethanarethedifferingaspectsthemselves.Thispointhasbeenillustratedinseveralofourearlierexamples,butitcanequallywellbeappliedinall.Itisinfactaprinciplethatcanbeexpectedtoobtainineverycase.Itmaybestatedasfollows:Whilenosingleexplanationwillanswerallcausalquestionsthatcanbeaskedconcerninganyconcreteoccurrence,anywell-formedquestionthatiscorrectlyansweredwillfitintoaconsistentpatternofexplanation.Weshallhavefurtheroccasiontonotethisfeatureofcausalexplanationsinourlaterchapters.Asweshallsee,itwillbeofspecialimportanceinunderstandinghowhistoricalexplanationscohere.Beforeturningtothatquestion,however,Ishallexaminesometraditionalproblen1sconcerningcausation,necessity,andlaws.Indoingsoitwillbecomeclearthatmyanalysisofthecausalrelationhaswiderapplicabilitythanhasyetbeensuggested,andisbynomeansconfinedtotheconcernsofeverydaylife. ChapterFourCAUSES,NECESSITY,ANDLAWSOfthemanyphilosophicproblemsthathavetraditionallybeendiscussedwithrespecttocauses,necessity,andlaws,Ishallconfinemyattentiontothree:first,problemsthatariseifoneattemptstoseparate"causes"from"conditions,"aswellassomethatariseifonerejectsthatdistinction;second,howtheconceptofcausationisrelatedtotheconceptofexplanatorylaws;and,third,inwhatsenseorsensesof"necessity"wemustassumethatnecessityispresentincausalrelationships.TheviewsIholdwithrespecttoeachofthesequestionshavealreadybeensuggestedintheprecedingchapterandtheappendices,butfurtherdiscussionisneededsothatwemaylaterseeinwhatways,andtowhatdegree,historicalstudiesinvolvecausalexplanationsandexplanatorylaws,andwhatrole,ifany,theconceptofnecessityplaysininterpretingthepast.80 CAUSES,NECESSITY,ANDLAWSI.ONCAUSESANDCO'NDITIONSOneusualfeatureofmodernphilosophicanalysesofcausationisthatadistinctionisdrawnbetweenthecauseofaneventandwhateverconditionswereresponsible-inadditiontothatcause-foritsoccurrence.SofarasIknow,noanalyticalhistoryofthisdistinctionhasbeenwritten.Whatisofinteresttome,inthepresentcontext,isthefactthatatleasttwoimportantmodernphilosophers,HobbesandMill,dealtwiththeprobleminawaythatlargelyundercutsthevalidityofthedistinction.Mill'sdiscussioniscomplexandvacillating,asHartandHonorehaveshown;IshallthereforetakewhatHobbessaysasmypointofdeparture.HisviewismostclearlysummarizedinonestatementinwhichhecontrastswhathetermstheHcausasinequanon"with"theentirecause"ofanoccurrence.Hesays:"Thataccidenteitheroftheagentorpatient,withoutwhichtheeffectcannotbeproduced,iscalledcausasinequanon,orcausenecessarybysupposition,asalsothecauserequisitefortheproductionoftheeffect.ButaCAUSEsimply,oranentirecause,istheaggregateofalltheaccidentsbothoftheagentshowmanysoevertheybe,andofthepatient,puttogether;whichwhentheyareallsupposedtobepresent,itcannotbeunderstoodbutthattheeffectisproducedatthesameinstant;andifanyoneofthembewanting,itcannotbeunderstoodbuttheeffectisnotproduced."lItisnotentirelyeasytosaywhatHobbes,inthispassage,wishedtodesignatebythetermcausasinequanon,exceptthatitmustrefertosomesinglefactorthatisnecessarytotheoccurrenceofagiveneffect,butwhich,alone,isnotsufficient.Yet,HobbesclearlyassertsthatboththecausasinequanonandtheentirecausearenecessaryinsomesenseorsensesofHnecessary"-iftheeffectistooccur.Toclarifyhismeaning,Isuggestthatthecausasinequanonbetakenasreferringtowhateverconditionisnecessaryforaneventofagiventypetooccur.Ontheotherhand,toexplaintheoccurrenceofsomeparticulareventonemustdiscoveritsentirecause.Thus,thecausasinequanonofaneventwouldbesomefactorthatmustbepresentintheentirecauseofaspecificeventinorderthatanyeventofthattypeshouldoccur.Onthisinterpretation,thecausasinequanonshouldnotberegardedasHthemostimportant"ofthetotalsetofoperativefactorsthatbroughtaboutaspecificeffect;instead,itwouldbedistinguishedfromotheroperativefactorsnotintermsofitsefficacy,butonlybecauseitisalwayspresentwhencasesofthesame CAUSATIONtypeoccur.UnlessHobbes'sdistinctionisinterpretedinthisway,Idonotseethatitwouldhavebeenworthmaking,sinceeveryfactorintheentirecauseisnecessaryfortheproductionofsomeparticularevent,andanyofthesefactorscould,therefore,equallywellbedesignatedasacausasinequanon.Tobesure,hadHobbesbeenoneofourcontemporaries,onemightassumethatbyacausasinequanonhemerelywishedtodesignatewhicheverfactormightmostusefullybesingledoutinordertoexplaintosomeonewhowaspuzzledbytheoccurrenceofsomeeventwhythateventhadoccurred.PerhapsHobbes'sphrase.:tcausenecessarybysupposition"mightbetakenassuggestingthisinterpretation.Ontheotherhand,itwouldnotbeeasytosquaresuchaninterpretationwiththefactthatHobbes'swholediscussionofcausationiscouchedinontologicalterms,focusingonthewaysinwhichthepropertiesofbodiesexplainevents.ItthereforeseemsessentialtointerpretHobbesasholdingthatacausasinequanonmaybeanecessaryconditionforexplaininganeventinsofarasitisaneventofagiventype,butthatonlythewholesetofconditions,makinguptheentirecause,isbothnecessaryandsufficienttoexplaintheoccurrenceofthateventitself,whenitoccurredandasitoccurred.Whileitisassuredlyusefultotryanddiscoverwhatconstitutesanecessaryconditionfortheoccurrenceofanyeventofaparticulartype,weareoftenjustifiablyinterestedindiscoveringpreciselywhysomeveryspecific,uniqueeventoccurredasitdid.Todoso,wemustdiscoverwhatconstituted"theentirecause":Itisnotsufficienttosingleoutsomeonefactoranddenominateitas"thecause,"distinguishingitfromwhataremerelyaccompanying"conditions."Thisfactmaybeillustratedinthefollowingway.2Ifoneistoexplainhowitcameaboutthatsomespecificconflagrationdestroyedalmostanentirehousewithinlessthananhour,butthatonewingofthathousesufferedonlynegligibledamage,onemusttakeintoaccountacomplexsetofconditions,suchasthefactthatalightedmatchwascarelesslydroppedintoawastebasketfullofpapers,thatthebasketwasnearinflammablecurtains,thatthewallswereeasilycombustible,thatsomewindowshadbeenleftopenandprovidedadraft,thattheconstructionofthewholehouse,withtheexceptionofonewing,wasnotbuilttoresistfire,butthatafireproofwallseparatedthatwingfromtherestofthehouse.Theseconditionsareallnecessarytoexplaintheseriesofeventsinthisparticularcase,buttheyarenot,ofcourse,relevantinallcasesinwhichtherearedevastatingfires.Inanothercaseafiremayhavebeendeliberatelyset CAUSES,NECESSITY,ANDLAWSbyanarsonistwholaidatrailofeasilyignitedchemicalsandthensetfiretoit.The,endresultmightineachcasehavebeenthesame,butacausalaccountofwhathappenedwouldinvolvetracingout,inasmuchdetailaspossible,howthefirehadstartedandspread,fornoteveryactofcarelessnessnorofarsonendsinthealmosttotaldestructionofahouse.Thewholeprocess,andnotmerelyitsorigin,wouldbeneededtoexplainthedamagethatactuallyoccurred.Thisisnottosaythatonecannotgeneralizeandstateconditionsnecessaryfortheretobeanyfirewhatsoever,whetheritbeahousethatburns,orafirethatburnsinacoalfurnace,oramatchoraforestthatburns.Knowledgeofthatsortisoftenextremelyimportantinhelpingusunderstandwhathasoccurredinaspecificcase.Nevertheless,itisnosubstituteforacausalanalysis.Becauseanysuchconditionwillapplyequallytoalloccurrencesofthesametype,itwillnot-byitself-beadequatetoexplainthetimeandnatureofanyparticularoccurrence.Infact,oneshouldnotevenassumethatwemustinallcasesrefertosomeuniversallynecessarycondition(thatis,tosomecausasinequanon)whenweareseekingtoexplainwhatoccurredinaparticularcase.Forexalnple,althoughweknowthatthepresenceofoxygenisanecessaryconditionfortheoccurrenceofafire,whenweattributeonefiretoarsonandanothertoamatchbeingdroppedinawastebasket,wearedealingwitheventsonacommon-senselevel,andarenotanalyzingthemonamicrolevel.Onthatlevelofexplanationwedonot-andneednot-includethefactthatoxygenwaspresent.Inthiscase,asinmanyothers,theonlynecessarycondition,orcausasinequanon,isnotdiscernibleinoureverydayexperience;itwillbesomefeatureofthemicrostructurethatispresentinallcasesofthesametype.Similarly,whenweexplainhowafirespread,citingthepresenceofconlbustiblernaterials,weneednotexplain(andonacommon-senselevelwecannotexplain)whythesematerialswereinfactcombustible.Aswenotedinanothercantext,differentlevelsofexplanationmustberecognized:OnemaysupplyanadequateanswertoaparticularcausalquestionwithoutansweringallfurthercausalquestionsthatanswermayraIse.Thispoint,withwhichwearealreadyfanliliar,followsfrolnthefactthatinourcausalexplanationswealwaysstartfromsomeeffectthatistakenasgiven,andweseektoaccountforthateffect,thatis,tofinditscause.Inanysuchanalysis,newcausalquestionswillarise,sinceeachoftheeventsthatispartofthecauseofaneffectispresumablyitselfaneffectofothercauses.However,toanswerour CAUSATIONfirstquestionweneednotalsoanswerthosefurtherquestionsitmayengender:Thequestforexplanatoryclosure,asHeITIpelhastermedit,isawhollyfruitlessquest.Thus,itshouldnotbeassumedthatinofferingacausalexplanationofaparticularoccurrencewearecommittedtoanendlessandthereforesenselesstask.Therewillalwaysbesatisfactorystoppingpointssincethespecificquestionsthatwereactuallyaskedwillhavefoundtheiranswers.TherearetwoadditionalpointstowhichIherewishtocallattentioninordertoavoidthepossibilitythattheymightentangleusinneedlessdifficulties.Thefirstisthefactthatalthoughweuseasingularterm,"thecause,"incausalexplanations,ouruseofthesingulardoesnotentailthatthecauseisasingle,simpleevent.Inspeakingof"theeffect"wealsousethesingular,butindescribinganyspecificeventorstateofaffairswemustalwaysrefertoalTIultiplicityoffactorsthat,together,constituteit.Therefore,eventhoughweuseasingularterminspeakingofthecauseofanevent,itdoesnotfollowthatatenabledistinctioncanbedrawnbetweenanyonecausalfactorandwhateverfurther"conditions"wereresponsiblefortheoccurrenceoftheeventweareattemptingtoexplain.Inthesecondplace,Iwishtopointoutthatmyargumentconcerningcausesandconditionsisnotintendedasananalysisofordinarylanguage.Aswehavealreadyseen,therearemanyoccasionsinordinarylifeinwhichwegiveandaccepttruncatedcausalexplanations;insuchcasesweareindeedapttospeakasiftherewereafundamentaldifferencebetweenthecauseofaneffectandwhateverattendantcircumstanceswerealsonecessaryforitsoccurrence.3However,whenwegobeyondsuchshorthand,truncatedexplanations,thedistinctionbetweencausesandconditionsbreaksdown,andthecauseofaneffectmustbetakenasincludingalloftheso-calledconditionsthatwereessentialtotheoccurrenceoftheeffecttobeexplained.Inordertopreparethewayforthisargument,letusreturntoHobbes'sconceptionofthecausasinequ.anon,sinceitmightbethoughtthathiscontrastbetweenitand"theentirecause"perITIitsustodrawadistinctionbetweencausesandconditions.This,however,isnotthecase.Aswehaveseen,Hobbes'scausasinequanonistobetakenasreferringtowhateverfactorsmustalwaysbepresentinorderforacertaintypeofeffecttooccur,asoxygenisnecessaryinallinstancesofburning.Thisbeingso,thecausasinequanonwillnotservetodistinguish"thecause"froITItheaccompanyingconditions,since(inmostcases)thepresenceofoxygenwouldnotbeidentified CAUSES,NECESSITY,ANDLAWSasthecauseofaparticularfire.Itwouldsimplycountasonealnonganumberof"conditions"withoutwhichthefirewouldnothaveoccurred.Thisfollowsfromthefactthatthosephilosopherswhodistinguishbetweencausesandconditionsusetheterm"cause"toapplytoSOlne"firingcause,"asthematterisoftenput:Whatevermayhavebeencontinuouslypresentpriortotheeffect-as·oxygenispresentbeforean1atchisstruck-isnotregardedbythemasthecauseoftheeffect,butonlyasanaccompanyingconditionofitsoccurrence.4ThispointofviewwasdevelopedinanoriginalwaybyC.J.Ducasseincharacterizingwhathetooktobethedifferencebetweenthecauseofaneventandaconditionofthatevent'soccurrence:"ThecauseofaneventBwasaneventAwhich,inthethenexistingcircumstances,wassufficienttotheoccurrenceofB...."AconditionofaneventBwasaneventAwhich,inthethenexistingcircumstances,wasnecessarytotheoccurrenceofB...."5Inintroducingthenotionof"thethenexistingcircumstances,"DucassewasguardingagainstholdingthatanygiventypeofeventAisalwayssufficienttocausetheoccurrenceofB,orthatsomeconditionAwill,underallcircumstances,beanecessaryconditionofB.Ishallnotarguethemeritsordemeritsofthisspecificformulation,butshalltakeitasmoreorlesstypicalofthosecasesinwhichthereisanattempttoholdthatthecauseofaneffectistobedistinguishedfromtheaccompanyingconditionsnecessaryforthateffect.Onthisview,thecauseissomeonespecificoccurrencethat,giventherequisiteconditions,issufficienttobringabouttheeffect.Ifsuchaconceptionofacauseistoberegardedasadequate,itmustholdinallcases,andnotinsomeonly.Incriticizingit,Ishallfirstshowthattherearecasesinwhichonecannotdrawalineofthissortbetweencausesandconditions.Ishallthenextendmyargumentbyrelatingittowhathasbeenshowninthelastchapter:Thatanyaccountofcausalrelationsinwhichcauseandeffectappeartobedistinctandseparateeventspresentsatruncatedandinadequateviewoftheirrelations.Amongthemostobviouscasesinwhichonecannotdrawaclearanddecisivelinebetweencausesandconditionsaresomefoundinphysiologyandmedicine.Othersatleastequallyclearariseinconnectionwiththeanalysisofhumanmotivation;and,asweshalllatersee,thecasesthatmaybeclearestofalloccurinhistoryandthesocialsciences.Tochooseoneobviousinstancefromthephysiologicalsphere,considerthediseaseoftuberculosis.IntheEncyclopaedia 86CAUSATIONBritannicawefindthatitissaidtobecausedbythetuberclebacillus,butthediseaseisthenmorecorrectlydescribedasfollows:Fundamentally,tuberculosisconsistsinaninflammatoryreactionofanyparticulartissuetotheinvadingbacilli,andsincetuberclebacilliarerelativelylittlevirulent,thistissuereactionissubacuteincharacter.Itconsistsintheformationroundthebacilliofamicroscopicagglomerationofcellsconstitutingtheso-called"tubercle."...Thoughintheearlieststagethetubercleismicroscopic,whenseveralareformedclosetogethertheybecomevisibletothenakedeye.Thepassagecontinuesbydescribingwhatoccursifresistancetotheinfectionislow,andwhatoccursifresistanceishigh;itconcludes:Tuberculosis,whenithasreachedthestageatwhichitisclinicallyrecognizable,mayberegardedastheend-resultofaslowprogressiveunrecognizedbacterialinvasion.6Whatistobenotedisthattheeffect,acaseoftuberculosis,isregardedastheendresultofaprocessbeginningwiththebacterialinvasion.However,whatisimportantisnotmerelytheinvasionbutalsothetissuereactiontoit;theactualendresultwillbedependentuponthosefactorsthat,together,constitutethebody'sreactiontotheinvasionandthepresenceorabsenceofsecondaryinfection.Inspeakingofanyparticularcaseoftuberculosis,wemustthereforetakeintoaccountnotonlytheinitialinvasion,butthewholecomplexprocessthatterminatesinthespreadoftheinfectionoritscontainment.Solongasweareattelnptingtoexplainwhatoccurredinaparticularcase,itwouldbefalsetoregardthebacilliasresponsiblefortheeffect.Theactualoutcomewillhavedependedonaninterplayofthevariousfactorspresentinthisspecificcase;nooneamongthemisentitIedtobeisolatedfronltheothersanddesignatedas"thecause"thatbroughtabouttheeffect.Itisonlyifoneconsiderstuberculosisasatypeofdiseasethatonespeaksofthetuberclebacillias�'thecause"ofthatdisease;butoneisnotthenusingtheterm"cause"torefertosomeeventthatwassufficient,underthecircumstances,toproduceagiveneffect,buttosomefactorthatisanecessaryfeaturepresentinallinstancesofthedisease.Inshort,insteadofgivingacausalexplanationofaspecificevent,oneisformulatingalawconcerningalleventsofagiventype.Bydrawingupontheliteratureofthephysiologicalsciencesonecanmultiplyalmostindefinitelycasesinwhichitisobviousthatwhatisresponsibleforaparticulareffectcannotbeidentifiedwithany CAUSES,NECESSITY,ANDLAWSsinglecausalfactor;thatonemusttakeintoaccount,asequallyefficacious,theongoingprocesseswithintheorganismthatarealsodeterminativeofthateffect.Thus,solongaswearespeakingofspecificeventsandnotoftypesofevents,thereisnojustificationforsinglingoutsomeoneimlnediatelyantecedentevent,andidentifyingitas"thecause"ofwhatoccurred,whenthepresenceofotherfactorswasnolessessentialtotheoccurrenceofthateffect.Thus,itismyclaimthatthelinebetweenwhatistakentobe"thecause"ofaneffectandwhataremerely"conditions"ofthateffectdisappears.Thisistruenotonlywithrespecttophysiologicalexplanationsbasedonscientificinquirybut,aswesawintheprecedingchapter,italsoobtainsinourcommon-senseexplanationsofwhatfollowswhenapersonswallowsapoison.Weshalllaterseethatsuchcasesarenotconfinedtophysiologicalprocesses,butariseinthephysicalsciencesaswell.Itmightbethoughtthateventhoughthedistinctionbetweencausesandconditionsbreaksdownwhenoneatten1ptstoexplainprocessesintheworldofnature,thedistinctioncanbemaintainedwhenweseektounderstandhumanactions.This,however,provesnottobethecase.Asafirstexample,takeasuddenoutburstofangerwhenapersonresentsaremarkandlashesbackwithathreatofviolentreprisal.Canwesay,quitesimply,thatthecauseoftheangerwasonlytheremarkthatwasInade?Ifwedo,itislikelytobeinexplicablewhyinthiscasesucharemarkevokedsuddenangerwhereasifithadbeenmadeatanothertilnetothesameperson,ortoanotherperson,itmighthavebeenshruggedoff;orthepersonagainstwhomithadbeendirectedmightevenhavebeenmildlyamused.Toexplainthereaction-whateveritmaybe-onemusttakeintoaccountmorethantheremarkitself:Onemustviewtheremarkinthecontextofwhomadeit,andofthecircumstancesunderwhichitwaslnade.Onemustalsounderstandwhetheratthattimethepersontowhomtheremarkwasmadewasirritableandfrustrated,orwhetherhewasrelaxed.Onemustalsounderstandhislonger-rundispositionaltendenciestoreactinonewayratherthananotherwheneverhisegoisinvolvedinsituationsofagiventype.Inshort,boththe"conditions"characterizingthecircumstancesand"conditions"obtainingwithrespecttothepersonhimselfformpartofthecauseofhisreactinginonewayratherthananothertoanyremarkdirectedagainsthim.Itisonlyinshorthandexplanationsthatweidentifythecauseofhisangerwithnothingbuttheremark.Yet,eventhen,wemustassumethe 88CAUSATIO,Npresenceoftheotherfactorsifwearenottobepuzzledastowhythatpersonreactedashedid,sinceweknowthatsuchremarksdonotuniversallyleadtoviolentdisplaysofanger.Anotherexampledrawnfromoureverydayunderstandingofhumanbehavioristobefoundinthewaysinwhichweoftenexplainfearreactions.Ifachildoranadultappearsterrifiedofsomesituationwearenotinmostcasescontenttoexplainthatterrorsimplyintermsofsomeaspectofthesituationheconfronts.Tobesure,therearecasesinwhichsomesingle,simpleevent,suchasasuddenloudnoise,seemssufficienttoaccountforthereactionoffear.Whenthisisthecase,itseemslegitimatetodrawadistinctionbetweenthecauseofanevent-thesuddenloudnoise-andwhatareconstruedtobemerelyconditionsrequisitefortheeffect-'forexample,thatthepersondidnotanticipatesuchanoiseanddidnotrecognizeitssourcewhenitoccurred.Inothercases,however,whenwearedealingwithfears,ratherthanwithwhatmightbecalledstartlereactions,thecauseisclearlynotregardedasasingle,simple,antecedentevent.Itisnot,forexample,thepresenceofacatalonethatterrifiesachild;weseektoexplaintheterrorthroughsomefactsuchasthechild'shavingbeenbadlyscratchedbyanothercat.Toseektodistinguishbetweencauseandconditioninsuchasituationishopeless,andcanberecognizedtobehopelesswhenwenotethatmuchofthechild'sterroroftendependsuponthecircumstancesunderwhichhefirstnoticesthecatintheroom-whetheritisstalkingsoftlyorisbeingcarriedbyapersonthechildknows.Similarly,toexplainaperson'ssuddenterrorifhefindshimselflostinunfamiliarsurroundings,wecannotisolateanyonefactorasthecauseofhisfear.Tobesure,wemaysaythatitisthethreateningnessofthesurroundings,orhisrecognitionofaperil,thatcauseshisterror,butthesecharacteristicsofthesituationinwhichhefindshimselfarenotisolablefromtheunfamiliarityofthesurroundings,norfromtheapprehensionlinkedtothisunfamiliarity.Toseektoisolateanyoneofthese-factors,makingitdistinctfromtheothersintheinterestoffindingsomesinglecause,istodistortsuchexperiences.Onecan,forexample,beawareofpossibledangersandyetnotfeelterrorifone'isin'familiarsurroundingsandknowswhatpathsofescapemaybeopen;oronemaybeinunfamiliarsurroundingsandenjoythemasexciting,preciselybecausetheyareunfamiliar,notfindingthemintheleastterrifying.Muchwillofcoursedependuponthecharacteroftheperson,forinthecaseofthetimidweexpectadifferentreactionfromthatexpectedofthosewhothroughouttheirlivesareassertiveandself- CAUSES,NECESSITY,ANDLAWSconfident.Aresuchcharactertraitstobecountedas"causes"oras"conditions"ofourreactions?Toattempttodrawanysuchlineis,Isubmit,tobemisledintothinkingthatwhatevermayoccurinsomecases-asinthosecasesinwhichfearisatypeofstartlereactionmustoccurinall.ItshouldnowbeclearfroBleachoftheprecedingexamplesthattherearemanycasesinwhichwecannotisolateanyfactorasthecauseofaspecificeffect,relegatingotherfactorstothestatusofbeinglnerelytheconditionsunderwhichthatcauseoperates.Nevertheless,therearecases-suchasthatinwhichastartlereactionoccurs-inwhichitseemsplausibletospeakinthisway.Usuallythesearecasesinwhichthereisaninstantaneoussuccessionofwhatareseenastwodistinctevents.However,aswenotedintheprecedingchapter,suchcasesconstituteshorthandexplanations,andarequicklyseentobeinadequatewhenanyexceptionoccurs.Toexplain,forexample,whyapersonwhogenerallydisplaysastartledreactiononhearingaloudnoisedoesnotappearstartledononeparticularoccasion,welnayhavetotakeintoaccountthefactthatinthiscasehewasexpectingthenoise.Thus,itisnotmerelyaloudnoisethatexplainshisnormalstartlereaction,butthepresenceofanunexpectedloudnoise.Inintroducingthefactorofunexpectednessintoourexplanationwearetakingintoaccountnotlnerelyanevent,butastateoftheperson;andonecannotexplainaperson'sreactioninanyoneinstancewithouttakingitintoaccount.Thus,thelinebetweentheeventthatappearsas"thecause"and"theconditions"thatarerelevanttotheeffectisalinethatdisappears.Thecaseis,then,preciselylikethatinwhichaflickofaswitchisusuallyconsideredtobetheadequate,sufficientcauseofalight'scomingon,butwhenanexceptionoccurs-aswhenthebulbisburnedout,orthepowerhasbeencutoff-welnustgobeyondourusualshorthandexplanation,inwhichsomesingleeventistakenasthecauseoftheeffect.Twodifferenttypesofobjectionswillquitepossiblyberaisedagainsttheprecedingargulnent.Thefirstwouldseektoshowthatifoneabandonsthedistinctionbetweencausesandconditions,treatingthelnasbeingofequalilnportance,itwillbecomeimpossibletoofferanadequateexplanationofanyoccurrence,sincethelistofconditionsthatcanbeshowntobenecessaryforthatoccurrencewillbeinexhaustible.Asecondobjectionmightbethat,inbreakingdownthedistinctionbetweencausesandconditions,abandoningthenotionthatthecauseofaneffectissomespecificantecedentoccurrence,onewouldbesurrenderinganessentialaspectofthecausalrelationship: CAUSATIONitstemporalnature.Ishallfirstbrieflyshowthatthefirstobjectionisunwarranted,andthenshowhowthesecondistobemet.Inordertoanswerthefirstobjection,itisnecessarytoshowthatthereisnotaninexhaustiblelistofrelevantconditionsthatmustbetakenintoaccountinordertoofferanadequateexplanationofanyoccurrence.Onestepinshowingthatsuchanobjectionisunwarrantedhasalreadybeentaken.AsIhaverepeatedlyargued,inexplaininganeventwecometoanaturalstoppingpointwhenwespecifytheelementsintheprocessthatterminatedinthatparticularevent;wearenotcommittedtoofferingfurtherexplanationsastowhyeachoftheseelementsitselfoccurred.ThisisnolesstrueontheviewIhavebeendefendingthanitisontheconventionalview.Thosewhobelievethatanoccurrenceistobeexplainedbycitingsomesingleantecedenteventneedfeelnoobligationtoofferafurtherexplanationoftheoccurrenceofthatantecedent,norofitsantecedentandallpriorantecedents.Mutatismutandis,thiswillalsobetrueoftheviewIallllnaintaining.Thereis,however,asecondreasonwhyitmaybethoughtthatabolishinganyultimatedistinctionbetweencausesandconditionsleavesoneopentothechargethatitwillthenbecomeimpossibletoofferanadequateexplanationofanyevent.ThisreasonconsistsinthefactthatwhatistakentobethecauseofaneventisalwaysSOlneactualoccurrence,whereasthesumoftheconditionsrelevanttoanyoccurrencecouldincludeanindefinitelylargenumberof"negativeconditions."Byanegativeconditionwouldbemeantanyeventthat,haditoccurred,wouldhavepreventedtheoccurrenceoftheeventwewishtoexplain.Thus,tochooseanexamplemadefamiliarbyMill,thesuccessofasurpriseattackonamilitaryencampmentmayhavebeendependentuponthefactthatthesentinelwasasleepandthereforefailedtogivethealarm.InwhatIhavesaid,noroomseemstohavebeenleftforsuchnegativeconditions,sinceithasbeenmyclaimthataneffectistobeexplainedbytracingtheactualcourseofeventsthatterminatedinit.Furthermore,itwouldseemthatsuchnegativeconditionscouldbemultipliedatwill,forwewouldalsohavetotakeintoaccountthefactthatnoothersoldierhappenedtobeawakeandtohaveseentheenemyapproachingtheperimeter;thatnoenemysapperhadhappenedtostumbleonamine;thatnospecialsensorshadbeeninstalled,thoughtheyweretohavebeensupplied,etcetera,etcetera.Toescapethiscriticismweneedonlynotethatnotallso-callednegativeconditionshavethesamestatus.Thesentinelwasasleep,no CAUSES,NECESSITY,ANDLAWSoneawakened,andthesappersweresuccessfulinavoidingthemines.Theseareallactualoccurrences.Theyarenomoretobedescribedas"negativeconditions"thanstandingstillratherthanwalkingisanegativestateofaffairs.Inthecasebeforeus,theonlytrulynegativeconditionswouldrelatetowhatdidnotactuallyhappen,butmighthavehappenedifthesentinelhadinfactawakened,orifasapperhadtriggeredamine.Inraisingthesehypotheticalquestionsoneisnotofferingacausalexplanationofwhatactuallydidoccur:Oneisnotidentifyinganythingthatcanrightlybetermedacausalconditionoftheactualeventoneisattemptingtoexplain.7Insupportofthisclaim,letmeciteanotherexamplethathassometimesbeenusedinanefforttoshowthatnegativeconditionsareascausallyeffectiveasarepositiveoccurrences.Apersondiesbecausehehasingestedapoisonandnoantidotewasgiven.Itispointedoutthathadtheantidotebeenpromptlyadministeredthepersonwouldpresumablynothavedied;therefore,thefailuretoadministertheantidotewasonecausalfactorresponsibleforhisdeath.Now,insuchacasetherecouldbemanyreasonswhyanantidotewasnotadministered.Perhapsnooneelsewaspresent,ornonewhowerepresentrealizedthatapoisonhadbeeningested,ornoantidotetothatpoisonwasknown.Anyoneofthesestatesofaffairswouldservetoexplainwhysomethingdidnotoccurthatmight,underothercircumstances,haveoccurred.Yet,tosaywhatmighthavehappenedifanynumberofthingshadhappenedotherthanwhatactuallyhappenedissurelynottogiveacausalexplanationoftheparticularcaseathand.InpointingoutwhatmighthavehappenedhadtheCirCUlTIstancesbeendifferent,weareonlycallingattentiontothefactthatdeathisnotanecessaryconsequenceofhavingtakenthispoison;wearenotgivinganaccountofwhatinthiscaseactuallycausedtheperson'sdeath.Itis,ofcourse,importanttoknowwhetherthereisorisnotanecessaryconnectionbetweenanoccurrenceofaparticulartypeandsomeparticulartypeofeffect;asweshallsee,thisbackgroundofknowledgeisoftenhelpfulwhenweseektoestablishtheactualcauseofsomeparticulareffect.Nevertheless,asIhavealreadyindicated,andasweshallhavefurtheroccasiontosee,thedeterminationofwhatisanecessaryconditionofatypeofeventisnottobeidentifiedwiththediscoveringthecauseofaparticulareventofthattype.Itisnotmerelyaconfusionbetweenthesetwodifferent,althoughrelated,questionsthathasledtoaninterestintheroleplayedbysocallednegativeconditionsincausalanalyses;inaddition,thereisour CAUSATIONinterestinfixingresponsibilitywhentheeventsanalyzedconcerntheactionsofmen.Forpurposesofassigningpraiseand-especiallyblame,whatapersonfailstodoisoftennolessinlportantthanwhatheactuallydoes.Thus,toomitdoingsomethingthatoneisexpectedtodoisparticularlyrelevanttowhatHartandHonorecalled"attributivecausaljudgments."However,asIargueinAppendixB,whenhistoriansmakeattributivejudgmentsconcerningmatterssuchas"thecausesoftheCivilVar,"theirjudglnentspresupposeexplanationsofwhatdidinfactoccur.IshouldnotexpectHartandHonoretotakeissuewiththatclaim;nonetheless,theyintroducednegativeconditionsintotheircausalexplanations-generallyintheformofomissions,thatis,intermsofwhatsomepersonfailedtodo.Forexample,inonediscussionofnegativeconditionsascausallyexplanatory,theycitedanexampleanalogoustoMill'sexampleofthesleepingsentinel:Agardenerfailstowateragardenandtheplantsdie.Tosubstantiatemydiscussionofthesentinelcase,Ishouldliketoexaminethiscaseaswell,since(unlikeHartandHonore)Idonotbelievethatitestablishesthatnegativeconditionshavethesamestatusaspositiveoccurrencesinournonattributivecausalexplanationsofwhatdidinfactoccur.8Letusinitiallyputasideallquestionsofblame,andacceptitasasimplematteroffactthattheplantsdiddiebecausethegardenerdidnotwaterthegarden.Thequestionthatthenarises,andwithwhichwemustdeal,concernsthesenseinwhichhisnotwateringthegardenisanegativeconditionsimilartoallothernegativeconditionsonemightciteinacasesuchasthis.Forexample,asHartandHonorepointout,hadanyoneelsewateredthegarden,evenhadthegardenernot,theplantswouldnothavedied.Nevertheless,HartandHonorerightlyholdthatinsuchacasewedonotsaythatthefailureofanyneighbortowaterthegardenwasacauseoftheplants'dying;yetwedoinfactsaythatthegardener'sfailurewasthecause.Theyholdthatthisissobecauseoftheroleplayedinourcausalexplanationsofwhatweregardasnormalandwhatweregardasabnormal.Theywouldhaveusbelievethatwedistinguishbetweenthegardener'snotwateringandaneighbor'Snot-wateringbecauseoneexpectsthatagardenerwillwater;hisfailuretodosoisthereforeabnormal,andwewill,asaconsequence,pickoutthatfailureasthecauseofwhatoccurred.Thereis,however,anotherandquitedifferentreasonwhywewouldsaythatthegardener'Sfailuretowateriswhatcausedtheplantstodie.Agardener,havingbeenhiredtotendagarden,iscountedontodoso.Seeninthislight,hisabsenceisnotproperly CAUSES,NECESSITY,ANDLA'VS93viewedas"anegativecondition,"asweviewthefactthatitdidnothappentorain,orthatnoneighborsawthattheplantsneededwatering.Rather,thegardener'snotcomingbreachedanagreementhehadundertaken.Hedidsomethingotherthanthatwhichhehadagreedtodo.Inthisrespecthisactionissimilartothatofasentine1'ssleepingathispost;eachisaderelictionofaduty,whateverextenuatingcircumstancestheremayhavebeen.Gardenersandsentriesbearresponsibilitiesthatoff-dutysoldiersandneighborsdonotbear;iftheyfailtoperformthedutiestheyarecountedontoperform,wedonottreatthisasmerehappenstance,butassomethingactuallydone,andweviewwhatwasdoneintermsofitsconsequences.Itisforthisreason,Isubmit,thatweattribute"vhathappenedtothegardentothegardener'sfailuretocome,andnottoanyothercircumstancesthatmighthavesavedthegarden,hadtheyindeedhappenedtooccur.Letmenowturnfromadefenseofmyviewagainstthosewhowouldattackitbecauseitfailstotakeaccountoftheroleof"negativeconditions"incausalexplanations,tomeetanotherchallenge:thattheviewIholdeitheromitsordistortsthetemporalaspectofcausalrelations.Actually,IbelievethattheaccountIhavegivenhassomedistinctadvantagesinthisconnection.Forexample,oneproblemthathassometimesappearedworrisomewithrespecttothetemporalaspectofthecausalrelationarisesoutoftheassumptionthatacauseanditseffectaretwodistinctandseparateevents,foronthisassumptiononeisforcedtoaskatwhatinstantthecausebringsaboutitseffect.9If,asIhaveclaimed,theeffectisnottoberegardedasaneventseparablefromitscause,butissimplytheterminalstateofasingleongoingprocess,thisproblemdoesnotarise.Itmightbeargued,however,thatmypositionescapesoneworrisomeproblelllonlytofallintoothersthatmaybemoreserious.Thefirstandmostbasicoftheseproblemsishowoneistoconstrue"thecause"ofaspecificeffect,ifcauseandeffectarenotdistinctand,inprinciple,separable.Asshouldalreadybeclear,whatIwishtodesignateasthecauseofaneffectissimplytheprocessthatterminatesintheeffect:Thecauseisthewholesetofactualongoingoccurrencesoreventsthatresultedinthis,andnoother,particulareffect.AsIhavenoted,thefactthatwemostoftenusethesingular,andspeakof"thecause"ofaneffect,shouldnotbetakentomeanthattheterrncannotrefertoaconjunctionoffactors,butmustrefertosomesingle,simpleevent.Furthermore,asIhavealsopointedout,whenweaskforacausalexplanationofsomeparticulareffect,wearenotaskingalimitless 94CAUSATIONquestion.Inthefirstplace,wearenotaskingforanexplanationofallaspectsofsomeoneconcreteoccurrence,butonlyofoneofitsaspects,undersomeonedescription.Inthesecondplace,weareaskingwhatwasresponsibleforit,notwhatwerethecausesofitscause.Finally,asIhavealsopointedout,wemustnotallowtheshorthandcausalexplanationsweoftenuseindailylifetoserveasparadigmaticcasesofcausalanalyses,sincethesealwayspresupposethatwecanfleshthemoutthroughanappealtomoregeneralschemasofexplanation;infact,weoftenfindourselvesforcedtodosowhenapparentexceptionsarise.ItisthereforetofullercausalexplanationsthatIhavelookedforparadigmaticexamples,nottotheirtruncatedversions.Insuchcases,asIhavetriedtoshow,wemustincludewithinthecauseofaneffectalloftheelementsthatenteredintotheactualprocessthatterminatedintheeffect.ItisatthispointthatwhatistobeconsideredasrelativelynovelinmyaccountofthecausalrelationIllOStclearlyappears.Aswehavenoted,MillandHobbesrecognizedthattherealorentirecauseofaneffectconsistedinacomplexsetofconditionsandnotinsomeoneeventonly.Intheirdiscussions,however,eachoftheelementsenteringintotherealorentirecausewasconsideredasaseparateorindependentconditionpresentintheagent,orinthepatient,orinthecircumstancesobtainingataparticulartimeandplace.Incontrasttothatconceptionofthecauseasacompoundofseparateelementsthatproduceanewanddistinctoccurrencewhentheycometogether,theconceptionIadvocatetakestheseelementsascomponentswithinasingleprocess,terminatinginthatparticularstateofaffairsweregardastheeffectoftheprocess.Tobesure,thevariouselementsinthisprocesshavetheirownhistories,whichmaybeentirelyindependentofoneanother;however,whenseenasconstitutingelementsthat,together,areresponsibleforaparticulareffect,theyformpartoftheprocessthatterminatedinthatgiveneffect.Forexample,inourearlierillustrationofho"vanexceptionallygoodharvestwastobeexplained,wenotedthatinadditiontothegoodgrowingweatheritwasnecessarytotakeintoaccountthelaborofthefarmersandtheirskills.Althoughthemeteorologicalfactorsresponsibleforthefavorableweatherandthefactorsexplainingtheskillsofthosewhoraisedthecropsconstitutetwowhollyindependentsetsoffactorswhichhadindependentancestries,onthisoccasiontheycombinedtoformoneprocess:Theweatherandtheskillstogether,duringthegrowingseason,ledtothisexceptionallyfinecrop.Intheforegoingexample,itisobviousthatwhatisseenascon- CAUSES,NECESSITY,ANDLAWS95stitutingasingleoccurrencemayberelativetotheeffecttobeexplained.Forexample,themeteorologicalfactorsthatwerepartlyresponsibleforthefinecropformpartofnumerousotherhistories,thatis,theymaybeseenascontributingnotjusttothisoneeffect,buttoothereffectsaswell.Theverysameconditionsfavorabletothecropmight,forexample,havebeenextremelyunfavorableforattractingtourists,andmightthereforehavebeenpartofthestoryofwhyincomewaspoorforonesegmentofthepopulationofaregionwhileatthesametimeitwasextremelygoodforanother.Theactualweathercanthusbeacausalfactorindifferentseriesofevents,formingpartofdifferentoccurrences;itsrole.inanycausalexplanationwilldependuponthespecificnatureoftheeffectonesetsouttoexplain.Whateverhasledtothisparticulareffect(whichisnotitselfadiscreteeventbutsilnplytheterminalstateofaprocess)isthecauseofthateffect,andthiscauseisnothingotherthanthewholesetoffactorsthatenteredintothatprocess.10Ifwethenaskwhattemporalrelationsobtainbetweencauseandeffect,itisclearthattheeffectis'theendpointofaprocess,andthecausalfactorsleadingtotheeffectmustprecedeit.However,thecauseasawhole-thatis,thewholesetoffactorsthattogethermakeupthisprocess-isnotaneventseparatefromtheeffect,andcannot,asawhole,beregardedasexistingpriortoit.When,forexample,Imoveaheavysofafromonepartoftheroomtoanother,eachphaseoftheprocessofpushingandhaulingandtuggingprecedesthefinalcoming-to-restofthesofainitsnewlocation,butuntilthesofaactuallyisinthatlocationtheprocesshasnotbeenconcluded:Thecauseisnotcomplete.Thus,aneffectdoesnotoccursubsequenttoitscause,butonlysubsequenttosomeoftheoccurrencesthatareinvolvedinthatcause;thecauseiscompleteonlywhentheeffectoccurs.Thisview,Isuggest,escapesthetemporalparadoxessometimesthoughttoinhereinthecasualrelationship.Notonlydoesitescapethedifficultyofhowacausecanbringaboutaneffectifitisprecedenttothateffect(andthereforeisnotoccurringwhentheeffectoccurs);italsoescapestheparadoxthatconsistsinholdingthatiftherelationbetweentwoevents,AandB,isanecessaryrelation,thenwemayaswellsaythatthelaterevent,B,wasthecauseofAasthatAcausedB,sincetheexistenceofeitherwouldentailtheexistenceoftheother.Thisparadox,ofwhichmuchhasrecentlybeenmade,11neednotarisewhenonerecognizesthatacauseanditseffectarenottwodistinctevents,butaretobedifferentiatedasbeing CAUSATIONaspectsofoneandthesameprocess.Sinceweexplaintheeffectbytracingouttheconcreteoccurrencesthatleduptoit,thetemporaldirectionofcausalattributionsisineluctable:Onecannotseparatecausefromeffectandspeakofthemasiftherelationbetweenthemweresomeabstractlogicalrelation.Thus,thisparadoxaswellasthefirsttemporalparadox-howacausecanbringaboutaneffectifitisprecedenttothateffect-,villdisappearwhencauseandeffectarenottreatedasiftheyconstitutedtwoeventsdistinctandseparatefromeachother.Notonlydoesthevie,vofcausationthatIamdefendingenableonetoescapetheseapparentparadoxes,but,moreimportant,itallowsustomakesenseofthewayinwhichcausalexplanationoccursinanyadvancedscience.IthasofcoursebeenarguedbyRussell,amongothers,thattheconceptofcausationisotioseinthesciences;however,theseargumentsrestedonidentifyingcausalexplanationswiththetypeofshorthandexplanationofanygivenoccurrencethatiscommonlyusedineverydaylife,inwhichonepicksoutsomeantecedentoccurrenceandidentifiesitasthatwhich,becauseofitsownnature,hasthepowertoproducethistypeofeffect.Whenonerefusestoviewsuchshorthandcausalexplanationsasparadigmsofwhatisinvolvedwhenseekingthecauseofsomeparticularoccurrence,itcanbeseenthatwheneverascientistiscalledupontoexplainwhathasoccurredinanygiveninstance,hetoousesacausalmodeofexplanationinwhichaparticulareffectisviewedastheoutcomeofacomplexprocessofchange.12Itisespeciallyapparentinscientificexplanationsthatthecausalfactorstowhichanyeffectisattributablearenotseparate,antecedentevents:Thephysicistinspeakingasaphysicistwillnot,forexample,explainthemotionofabilliardballsimplyintermsofitshavingbeenstruckbyanotherball;theimpact,whichtothenakedeyeappearstotalandinstantaneous,willberegardedasaprocessinwhichaseriesofchangesoccurswithinfieldsofforce;whatistakenastheeffect-themotionoftheballthathasbeenstruck-istheendproductofthesedistortions.Similarly,aswehaveseen,aphysiologicalaccountofthestateofatubercularpatientistobeexplainedastheendresultofaprocess,notasastatethatfollowedtheoccurrenceofsomeearlier,quitedifferent,isolableevent.Thus,inthesciencesaswellasinourmorecarefulcausalexplanationsineverydaylife,what.istobeexplainedisnotexplainedintermsofregularityofconnectionbetweendistincttypesofevents,butisaccountedforbytracingtheelementswithintheparticularprocessthatendedinthiseffect.Thisasweshalllatersee,makesitpossibletobridgethegapbetweenthe CAUSES,NECESSITY,ANDLAWS97causalexplanationsgivenineverydaylifeandthesophisticatedexplanationsofparticulareffectscharacteristicoftheadvancedsciences.Beforeshowingthatsuchaharmonydoesindeedexist,itwillbenecessarytodiscussthewayinwhichtheconceptofalawofnatureisrelatedtoouranalysisofthecausalrelationship.II.CAUSALEXPLANATIONSANDNATURALLAWSCausalexplanations,aswehaveseen,aredirectedtowardansweringthequestionofwhatwasresponsibleforsomeparticularoccurrence.Anaturallaw,ontheotherhand,consistsintheformulationofsomeinvariantconnectionbetweenpropertiesoreventsofspecifiedtypes.Aswehavealreadynoted,inordertodiscoverorconfirmanysuchlawonemustdealwithparticularoccurrences;itisalsotrue,asweshallsee,thatestablishedlawsareimportantinourcausalexplanationsofwhathasoccurredinaparticularcase.Thus,causalanalysesandtheformulationoflawsareconnectedenterprises,buthowtheyarerelated,aswellashowtheydiffer,remainstobeseen.Unfortunately,thedifferencesbetweentheproceduresrequiredingivingcausalexplanationsandindiscoveringandconfirminglawshavebeenconcealedbythefactthatparadigmaticexamplesofbothprocesseshavetoooftenbeendrawnfromshorthandcausalexplanationsandfromthecrudelawlikegeneralizationsthatarecharacteristicnotofthesciencesbutofeverydaylife.Ifonetakesasthemodelofacausalexplanationthefactthatcropsaregoodbecauseoftheweather,orthatanelectriclightcomesonbecauseaswitchhasbeenflicked,thereseemstobelittledifferencebetweensuchanexplanationandmerelycitingageneralruleorlaw.Thetrouble,however,isthatsuchcausalaccountsnotonlyareveryincompletebutpermitofexceptions,aswehaveseen;therefore,thelawthattheyaretakenasexemplifyingwillnotinfactstateaninvariantconnection,butwillbeacrudegeneralizationconcerningwhatgenerallyhappens,orhappens"undernormalcircumstances,"anditsapplicationwilldependuponintroducingavagueceterisparibuscIause.If,instead,afull-fledgedcausalanalysisischosenasanexample,itimmediatelybecomesclearthatthereisaradicaldifferencebetweenwhatsuchananalysisincludesandwhatisinvolvedinthefornlulationofalaw.Togiveacausalanalysisistotraceanongoingprocessthatterminatedinthespecificeffectwewishtoexplain;thisinvolvesdescribingaparticularsetofinterconnect�doccurrences.Informulat- CAUSATIONingalaw,ontheotherhand,oneisconcernednotwithaparticulareffect,butwithaneffectofaspecifiedtype;theobjectistoshowonwhatfactororfactorsaneffectofthistypealwaysdepends.Onemight,ofcourse,assumethatwheneveraneffectofagiventypeispresent,somespecifictypeofantecedentoccurrencewillalsohavebeenpresent,sothatalawsimplyexpressestheexistenceofacorrelationbetweentwotypesofoccurrences.Thisismoreorlesscharacteristicofoneofthewaysinwhichweareinclinedtoformulatelawlikegeneralizationsineverydaylife:Ifanoccurrenceofsomeparticulartypeisfrequentlyobservedtobeassociatedwithsometypeofeffect,andifeffectsofthattypehaveneverbeennoticedintheabsenceofsuchoccurrences,weareinclinedtoassumethatthereisanecessaryconnectionbetweenthem,andwetakethestatementofthatconnectionastheformulationofalaw.However,generalizationsofthissortarenotreallycomparabletothekindsoflawsfoundinanyadvancedscience.Inthefirstplace,suchgeneralizationswouldbebothtoocrudeandtoounreliabletocountaslaws.Theyarecrudebecausetheydonotidentifythekindsanddegreesofresemblancethatvariousoccurrencesmusthaveiftheyare·tobringaboutagiventypeofeffect.Theyareunreliablebecause,aswehaveseen,inanyactualcaseotherconditionsmayalsohavetobepresentifaspecificeffectistooccur;asaconsequence,aceterisparibusclausemustbeintroducedtobuttressthesecommon-sensegeneralizations.Inanyadvancedscience,however,nosuchclauseisincludedwhenalawisstated:Alawappliestoallcasesandnot"undernormalconditions"only.Inthesecondplace,andofmoreimportancefromageneraltheoreticalstandpoint,isthefactthatthesecommon-sensegeneralizationsareformulatedintermsofareIationshipbetweenspecifictypesofoccurrences.Inthesciences,however,alawusedinexplaininganyactualoccurrencealsoappliestowhatwouldhaveoccurredunderothercircumstances,hadcertainconditionsbeenfulfilled.Thus,suchlawsdealwithrelationshipsthatmaybepresentwithoutbeingmanifest,andwithfactorsthatbecomemanifestonlyunderotherconditions.Scientificlawshavethisrangeofapplicabilitybecausetheydonotattempttostateconnectionsbetweenactualoccurrences,butbetweenpropertiescharacteristicofoccurrencesofgiventypes.Forexample,inBoyle'slawanattemptwasmadetorelatethepressureandthevolumeofagas,underanyconditions,sothatgivenachangeineitherfactoracovariantchangeintheotherwouldoccur.Similarly,inGali1eo'slawsoffallingbodies,orinNewton'sgravitationallaw,thelawaims CAUSES,NECESSITY,ANDLAWS99tostatearelationshipnotbetweenactualoccurrences,butbetweenspecificcharacteristicsofcertaintypesofobjectsorevents.Inotherwords,insteadofseekingtofindsomeonetypeofoccurrencethatisinvariantlyconnectedwithaparticulartypeofeffect,ascientistwillfractionatetheseoccurrencesintofactors,seekingtoestablishwhatfactorwithinsuchoccurrencesisconnectedwithafactorinvariantlypresentintheeffect.ItisatthispointthatonecanseewhyHume'saccountofourgeneralizationsfromexperienceaccordssopoorlywiththemethodsandtheresultsofscientificinquiry,andwhysomeformofanalysissuchasGalileo'smethodofresolutionisnecessaryfortheformulationofscientificlaws.Theselawsexpressrelationshipsamongfactorsthatmaybepresentinoccurrencesofmanydifferenttypes;theyarenotconfinedtostatementsconcerninginvariantsequences.Consequently,onecannotmaintainthedistinctionthathassometimesbeendrawnbetween"causallaws"andlawsthatservetocorrelatetwosetsofproperties,suchasthelengthofapendulumanditsperiodicity,orthechemicalcompositionofametalanditsmelbngpoint.13ToattempttomaintainsuchadistinctionwouldalmostsurelyinvolvedenyingthatlawssuchasthoseofBoyleorofGalileoorofNewtonwere"causallaws."Inthissameconnection,itbecomesevidentthatoncetheHumeanregularityviewisabandoned,onecanalsoabandonthenotionthatthespecificfactorsresponsibleforagiveneffectmustbeoccurrencesantecedenttothateffect.AsMaxBlackhasobservedindiscussingthecausalrelation,theassumptionoftemporalpriorityfailstosquarewithexamplessuchasthefactthat"themoon'sgravitationalpulllastsaslongasthetideitproduces;differenceoftemperatureregistersthroughouttheperiodthatthermometricexpansionoccurs;acatalystcontinuestoactduringthechemicalreactionitisinfluencing;andsoon,foranynumberofsimilarcases."14Finally,wemayagainobservethatifonespellsoutthelawsinvolvedincasessuchasthoseBlackmentions,theselawsdonotconcernconcretetypesofoccurrences,asHumesupposed.Rather,theyexpressfunctionalrelationshipsbetweenfactorsincludedwithinthesituation.Forexample,anexplanationofthecausalrelationbetweenthemoonandthetideswillbeformulatedintermsofgravitationalattraction,whichisdependentuponrelationsofmassandofdistance,ratherthanintermsofsomeantecedentevent.Itisworthwhiletonoteonceagainthatanappealofthissorttospecificpropertiesorfactorsinasituation,asameansofexplainingaspecifictypeofeffect,isnotconfinedtotheadvancedsciences;itis 100CAUSATIONalsocharacteristicofexplanationsfrequentlygivenineverydaylife.Toexplainwhysomeobjectsfloatwhendroppedinapondandotherssink,wemayappealtowhattheyaremadeof,ortotheirweight.Suchexplanatorygeneralizations,aswehavealreadynoted,areinsufficientlyaccurate,sincewhilewecaninmostcasescountonwoodtofloat,waterloggedwoodwillsink;andweightisnotthedecisivefactor,asshipsoflargetonnageconclusivelyprove.Thus,thesestatementsarenolessinadequatewhenconsideredasstatementsoflawsthanareanycasesinwhichweseektoestablishaninvariantsequencebetweensomespecifictypeofantecedentoccurrenceandaspecifictypeofeffect.Nevertheless,thereisanimportantdifferencebetweenthesetypesofexplanations,bofhofwhichareoftengivenineverydaylife,anditistothisdifferencethatIshouldhereliketocallattention.Onecannotreadilymovefromanattempttoestablishcorrelationsbetweenactualoccurrencestoaformulationofstrictscientificla\:s,sinceinordertohaveastrictlyinvariantconnectionbetweensometypeofantecedentoccurrenceandsomespecifictypeofeffect,wemustdefinethenatureofeachmorenarrowly,andthelawtherebybecomesincreasinglylessgeneralinitsapplicability.Ontheotherhand,inmoreclearlyspecifyingthepropertiesonwhichsometypeofeffectdepends,inordertomakesurethatnoexceptionsarise,wemovefromrestrictedproperties,suchasthatanobjectwillfloatbecauseitismadeofwood,tomoregeneralones,suchasthatitmustnotbe"tooheavy,"andthentomoreaccuratestatementsoffunctionaldependenceinvolvingtheconceptofspecificgravity;suchastatementisevenmoregeneral,aswellasbeingmoreprecise,sinceitisapplicableunderallconditions,andwithoutrespecttowhateverotherpropertiesanobjectmaypossess.Thus,startingfromatypeofcrudegeneralizationthatisnolesscharacteristicofourexplanatorygeneralizationsineverydaylifethanaregeneralizationsconcerningsequencesofevents,wecanmovetowardtheformulationoflawsthathold,withoutexception,inawidevarietyofinstances.Unfortunately,nosuchtransitiontakesplacesolongaswesearchforlawsthatattempttostateinvariantconnectionsbetweenonetypeofantecedentoccurrenceandsometypeofeffect.Havingnowseenthatthelawsonwhichreliancemaybeplaceddonotconformtothemodelofregularity-of-sequence,butstatefunctionalrelationshipsbetweentypesoffactorsratherthanbetweentypesofactualevents,wemustexaminehowlawsconcerningthesefunctionalrelationshipsenterintoconcretecausalexplanationsofspecificevents. CAUSES,NECESSITY,ANDLAWS101Letusrecallthatwhenweseektotracethecauseofagiveneffectweareseekingtoestablishtheelementsthatenteredintotheprocessthatterminatedintheeffectwewishtoexplain.Theseelementsarethemselvesoccurrencesthat,whenconsideredasleadingtotheeffect,formasingleunifiedstrand.Thequestionarises,however,astowhatgivesunitytoanysuchstrand:Whyarejusttheseoccurrences,andnotothers,toberegardedasleadingtotheeffectthatistobeexplained?Onepartoftheanswertothisquestionisobvious:Insonlecasestheseoccurrencesaremerelyphaseswithintheprocesstakenasawhole,asthepositionsandthemotionsofeachbilliardballinathree-cushionshotarebutpartsoftheoccurrencewetrace.However,intracingtheoccurrencesthattogetherformthewhole,wefindthattherearepointsatwhichtheyintersect,andatwhichtheyappeartoaffectoneanother.Toexplainhoweachisaffectedbytheother,andthereforetoexplainwhathasoccurred,wecannotmerelydescribeeachofthephasesincludedwithintheprocess,butmustlooktothewaysinwhichthepartsoftheprocessmaybeconnected,thusformingonecausalstrand.Forexample,inthesuccessfulthree-cushionshot,thephasesoftheprocessconsistedinaseriesofoccurrenceswithoutwhichtheshotwouldnothavebeensuccessful,buttoexplainwhatlinkedtheseoccurrencestooneanotherinawaythatmadethemphaseswithinonecausalstrand,appealwouldhavetobemadetofactorssuchasthemOlnentumandtheangleofincidenceofimpactandthelike,which,ateachsuccessivejuncture,determinedthepathofthecueball.Inanaccountsuchasthis,themortarthatservestomakeasingleprocessoutofwhatmightotherwisebeseenasonlyasuccessiveseriesofoccurrencesistherelationshipoffunctionaldependencebetweensomeofthefactorsembeddedinthoseoccurrences.Thus,inexplaininganeffectastheoutcomeofaparticularprocesswemustnotonlydescribetheseriesofconcreteoccurrencesthatenteredintotheprocess,butmustalsotakeintoaccounthoweachoccurrencewithinthatserieswasrelatedtoatleastoneofitsothercomponents.Todoso,wemustappealtogeneralizationsconcerninghowsomeaspectoftheonewasdependentuponsomeaspectoftheother.Thisisnotonlytrue"vithrespecttoanexamplesuchasthatofathree-cushionshotinbilliards,wherewehavetheadvantageofbeingabletoappealtowell-establishedlawsconcerningmotion,italsoappliestocasesinwhichourgeneralizationsarelooseandrelativelyunreliable.(Totheextentthattheyhavethesecharacteristics"thecausalexplanationsthatdependuponthemmustalsoberegardedas 102CAUSATIONopentochallenge,andtosubsequentcorrection.)Forexample,inthefieldofhumanrelationships,ifweseektogiveacausalaccountofwhytwofriendsbecameestrangedandsuddenlyturnedintoenemies,weshallhavetotraceaseriesofepisodesduringwhichtherewasgrowingtensionbetweenthem,andweshallhavetotakeintoaccountthefinalclimacticepisodeinwhichtheirentirerelationshipchanged.Thestagesinthisprocesswillconstituteourcausalaccount,butatalmosteveryjunctureweshallhavetoexplainthedynamicsoftheseattitudinalchangesbyappealingtofunctionalconnectionsbetweenspecificfactorspresentwithintheepisodes,suchasthefactthatfailureleadstoenvy,andthathatredarisesoutofinjuredpride.AsoneseesinSpinoza's'accountofthepassions,itispossibletogeneralizewithrespecttofunctionalrelationshipsamongourattitudeswithoutreferringtothespecificoccurrencesinwhichtheseattitudesmanifestthemselves.Suchgeneralizationspermitustotraceconnectionsbetweendifferentstates,andshowinwhatwaystheseriesofoccurrencesbelongtogetheraspartsofasingleprocess.Thus,suchgeneralizationsprovidemortarforourcausalaccountseventhoughtheydonotprovidethematerialsoutofwhichsuchaccountsareformed.Thegeneralizationsthatareusedtoformulateconnectionsbetweentwooccurrencesoftenreferdirectlytosomefeaturesofthoseoccurrencesbymeansofwhichwewouldnaturallydescribethem.Such,forexample,wasthecaseintheimmediatelyprecedingillustration,inwhichtheconnectiverelationshipswerecharacterizedthroughusingtermssuchas"envy"and"pride."InothercasesweInayformulateourgeneralizationsinmoreprecise,technicalterms.Whilestillreferringtofeaturesthataredirectlyidentifiableinthoseoccurrences,suchtermsoftenindicatefactorsintheoccurrencethatwedonotordinarilyconsiderapartfromoneanother.Forexample,indescribingwhathappenswhenonebilliardballstrikesanotherwemaydistinguishbetweenmomentum,theangleofincidence,thespinoftheball,etc.,allofwhichrefertofactorsthataredirectlyrecognizableinwhatisobserved,butwhichweconsiderseparatelyonlywhencalledupontogiveacloser,analyticaccountofwhatoccurred.Insuchaccounts,generalizationsconcerningtheseisolablefactorsareusedinexplainingtheconnectionbetweenthemotionofonebilliardballandthatoftheother,thusservingasconnectivelinksinourcausalexplanation.Finally,therearegeneralizationsthatdonotapplydirectlytowhatisobservedinthecaseathand,butexplaintheconnectionsbetweenobservedfeaturesoftheoccurrencethrough CAUSES,NECESSITY,ANDLAWSreferencetolawsconcerningwhatmightbecalled"micro-occurrences,"thatis,tooccurrencesunderlyingwhatwedescribeasoccurring.Forexample,inexplainingwhatoccursinthecollisionofbilliardballs,weneedtounderstandthelawsgoverningtheirelasticity,buttheselawsrefertotheirinternalphysicalstructure,nottoanyfactorsdirectlyobservableintheoccurrencewedescribe.Similarly,thelawsthatservetoexplaintheconnectionbetweenbeingshotthroughtheheartanddyingarephysiologicallawsthatrelatethecirculationofthebloodandothervitalprocesses,buttheselawsapplytomicro-occurrences,nottotheeventswespecifywhenwesaythathewaskilledbyabulletthatpassedthroughhisheart.15Itisimportanttonotethatwhenweusegeneralizationsofanyofthesekindsasanleansofexplainingthelinkagebetweenspecificoccurrences,suchgeneralizationsdonotinanysenseserveassubstitutesforthecausalexplanationweseektogive.Forexample,thephysiologicalgeneralizationsthatservetoexplaintheconnectionbetweenaperson'sbeingshotthroughtheheartandhisdyingaregeneralizationsthatconnectaspectsofwhatoccurswhenabulletpassesthroughtheheartandthemicro-occurrencesthatconstitutedying;inordertomakeuseofsuchgeneralizationswemustdescribewhat,inthisinstance,occurred.ItisforthisreasonthatIhavesaidthatwhilethelawsweuseinourcausalexplanationsprovidemortarthatmayservetoconnectfactorsthatarepresentintheseoccurrences,theydonotprovidethematerialsoutofwhichsuchaccountsareformed.Thisrelationshipmaybefurtherclarifiedinthefollowingway.Aswehaveseen,anylawthatistakenasformulatinganecessaryrelationshipwillbephrasednotintermsofrelationshipsbetweenspecifictypesofactualoccurreQces,butintermsofrelationshipsbetweenfactorsthatcanbeconsideredinabstractionfromthespecificoccurrencesinwhichtheymaybepresent.Thisbeingthecase,onecouldnot(sotospeak)"reconstitute"anyactualoccurrencemerelybyknowingwhatlawswereapplicabletoit:Onewould,onthecontrary,havetobeinapositiontoofferaconcretedescriptionofthatparticularoccurrencebeforebeingabletoapplytheselawstoit.Thisisofcourseacknowledgedevenbythosewhoholdthatadequateexplanationalwaystakestheformprescribedbythedeductive-nomotheticmodel:Theytooinsistthatonemustpossessaccurateknowledgeoftherelevantinitialconditionsinordertousealawinexplaining(orinpredicting)thespecificeffectwearetoexplain(orwhichwewishtopredict).Wearethereforeforcedtostartourexplanationsfrom CAUSATIONourdescriptionsofwhatispresentinanygivencase.However,inexplainingwhatoccurs,itwillnotbesufficientmerelytodescribetheoriginalinitialconditions;onemustinsteadtracethesuccessiveoccurrencesbroughtaboutastheseconditionschange.Forexample,ifachangeinonebringsaboutachangeintheotherbecauseofsomeinvariantconnectionbetweentwoaspectsoftheoriginalconditions,thenwemusttracetheconsequencesthischangewillhaveforotheraspectsoftheconditionsthatwereoriginallypresent:Inmostcasesonedoesnotmovesmoothlyandwithoutanyinterveningprocessesfromachangeinsomeonepairofconditionsdirectlytothefinaleffect.Forexample,inCarlHempel'swell-knownillustrationofthecrackingofaradiatoronacoldnight,16onecouldnotdeducethefinaleffectfromthefreezingofthewateralone:AsHempelrecognized,thefreezingofthewaterhadtobeconnectedwithanexpansionofitsvolume,andthecoefficientofexpansionhadtobeconnectedwiththedegreeofpressureatwhichtheradiatorwouldcrack.Itisthereforemisleadingtospeakasifaparticularsetoflawswouldbesufficienttoenableustoexplain(orpredict)agivenoccurrence,oncetheoriginalinitialconditionswereknown.Instead,thelawsmustbeappliedseriatimtoasequenceofchangingconditions;andthisistosaythatacausalaccountmustbegivenoftheseriesofoccurrencesleadingtothefinaleffect.Tobesure,atleastsomeoftheconnectionsamongtheseoccurrencesmaybeforlTIulableintermsofsuchgenerallawsasthefreezingofwaterataspecifictemperature,theexpansionofwaterwhenitfreezes,andthelawsdeterminingtheburstingpointofacontainerhavingthestructuralpropertiesofthisparticularradiator,buttheselawsarenotappliedallatoncetotheinitialconditions,asHempel'saccountmightseemtoimply:Theyapplytoaseriesofoccurrences,andarethusinstrumentsusedincausalexplanationsandnotsurrogatesforsuchexplanations.Nothingcouldmakethisclearerthanthefactthatinexplaining(orinpredicting)aspecificeffectintermsofthedeductive-noluotheticmodelwemustnotonlytakeintoaccounttheboundaryconditionsthatexistwhentheinitialstateobtains,butwemustexcludechangesintheseboundaryconditionsastheseriesofeventsistakingplace.17ItwillbeseenfromallthathasgonebeforethatIholdourcausalanalysestobeprinlarilydescriptive:Theyanalyzeanongoingprocessintoaseriesofoccurrencesterminatinginaspecificeffect.Giventheseoccurrences,andnoothers,wemaysaythatitwasnecessarythattheprocessshouldhaveterminatedinthiseffect,butthisisnottosaythattheprocessasitactuallyoccurredwasitselfanecessary CAUSES,NECESSITY,ANDLAWSprocess.Itisnottosaythatnothingcouldhaveinterfered,afteritsinitialstate,topreventtheoccurrenceofwhatdidinfactoccur.Itisaviewofthelattersortthatisbestcharacterizedas"determinism."WhatIwishtoshowisthatdeterminism,iftakeninthissense,isnotanecessaryconsequenceofholdingthatallprocessescanbeanalyzedincausalterms,norofholdingthattherearenecessaryconnectionsamongthefactorsuponwhichsuchprocessesdepend.III.ON"DETERMINISM"Idonotdenythatsomeprocessesareinfactdetern1ined,inthesenseinwhichIamusingthatterm.Anyprocessthattakesplaceunderboundaryconditionsthatexcludetheinfluenceofoutsidefactors,andinwhichwhatoccurscanbesaidtorepresenttheselftransformationofasystemaccordingtosomesetoflaws,isaprocessthat·canbecalleddetermined:Giventheinitialstateofthesystem,nothingbutthisparticularendresultcouldOCCUr.18Chemicalreactionstakingplaceunderexperimentalconditionsluightserveasexamplesofdeterministicprocesseswhendeterminismisconstruedinthissense,anditisinthissamesensethatonecanformulatethepositionofmetaphysicaldeterminism.Thatposition,asonecanrecognizeinSpinozaorLaplace,takestheuniverseasawholetobeasinglesystemthatundergoesaseriesoftransformationsaccordingtoonesetoflaws;whateveroccursinsuchaclosedsystemhadtooccur,giventheearlierstateofthesystemanditslaws.However,unlessoneiswillingtoassumefromtheoutsetthattheuniversecanbesaidtobeasingle,unitaryprocess,noparticularinstanceofdeterminismshouldbeconstruedasifititselfhadbeendetern1inedtooccur.Forexample,whileachemicalreactionthattakesplaceunderexperimentalconditionsmaybeaninstanceofacOlnpletelydetermilJedprocess,thereisnosetofinitialboundaryconditions,andnoonesetoflaws,fromwhichitcanbededucedthattheexperimentitselfwouldbeperformed.Nevertheless,onceweknowthatitactuallywasperformed-andperformedwhenandwhereandinthewayinwhichitwas-thereisnoreasontodenythattheeffectwasdetermined,eventhoughtheeventasawholewasnot.Thereisnoparadoxinthisposition,sinceitisonlywhatoccurredaftertheexperimentstartedthatrepresentsatransfor�ationwithinaclosedsystem;theplanningoftheexperimentandthetaskofcarryingitoutmighthavebeeninterferedwithbyanynumberofextraneousfactors, 106CAUSATIONandtheeventasawholewouldnotthenhavetakenplacewhenorwhereorhowitactuallydid.Thus,iftheoccurrencesinwhichoneisinteresteddonottakeplaceundercircumstancesinwhichtheyareinsulatedagainsttheeffectsofotheroccurrencesintheirenvironments,whatoccurscannotbesaidtohavebeendetermined,solongaswearespeakingfromanempiricalpointofview,ratherthanassumingthetruthofacomprehensivemetaphysicaldeterminism.Thisdoesnotofcoursemeanthattheseeffectsmaynothavebeennecessary(insomesenseof"necessary"),giventhefactsofwhatdidactuallyoccur.Theproblen1istocharacterizethenecessitythatistobeascribedincasesinwhichwedonotinvokeac.ompletemetclphysicaldeterminism,andinwhichthereisnotthatformofempiricaldeterminismthatcharacterizestheself-transformationsthatoccurwithinaclosedsystem.Thefirstthingtobesaidisthatthisformofnecessityisnottoberegardedasanexampleoflogicalnecessity.Humeofferedreasonsadequatetoshowthatanycausalnecessityweascribetoeventsisnotequivalenttologicalnecessity.Inaddition,however,wemaynotethatwheneverwetracethecauseofaspecificeffectwemusttakeintoaccountamultiplicityofotheroccurrences,andnooneoftheseoccurrencescan,byitself,betakentobeasufficientconditionfortheoccurrenceoftheeffect.19Therefore,nooneoftheseoccurrencescanbesaidlogicallytoentailtheeffect;rather,theeffectfollowsfromthecontingentfactthateachoftheseoccurrencestookplacewhenandwhereitdid.Norcanweshowthatthisconcatenationofoccurrenceswasitselflogicallynecessary.Therefore,itwouldbealnistaketoviewthenecessitythatlinkscauseandeffectasaninstanceoflogicalnecessity.NorcanweaSSUlnethatthereis,inallcases,anecessaryrelationbetweenanyoneconditionandtheeffectweseektoexplain.Thiscanbeillustratedwithreferencetowhathasbeencalledoverdetermination.Forexample,ifamanisexecutedbyafiringsquad,hisdea�hisattributabletothebulletspenetratinghisbody;butifweassumethateachofthebulletsstrucksomevitalorgan,canwesaythatanyoneoftheshotswasanecessaryconditionofhisdeath?Underthesecircumstances,hisdeathwouldhaveensuedfroinanyofthebulletsthatstruckhilu,and,asaconsequence,nooneofthemwasanecessaryconditionofhisdeath.Asweshallnowsee,thisissueisnotrestrictedtocasesofoverdetermination,butisrelevanttomanyotherinstancesaswell.Anexamplechosenfroinamongthosealreadydiscussedluaybeusedtoillustratethispoint.Considerthecaseofafirethatdestroyed CAUSES,NECESSITY,ANDLAWSanentirehousewiththeexceptionofonewing,whichhadbeenconstructedtobeproofagainstfire.InsuchacaseSOlneofthespecificoccurrencesandconditionsthatenteredintotherealandentirecauseoftheeffectlnaybesaidtohavebeennecessaryforthatparticularresulttohaveoccurred,butothers,whentakenindividually,shouldnotberegardedashavingbeennecessaryeventhoughtheyformedpartofthesetofconditionsthat,whentakentogether,werebothnecessaryandsufficienttoproducethateffect.20Forexample,givenalloftheeventsconstitutingthisfire,somemaybesolinkedtothefinaleffectthatiftheyhadnothadthecharacteristicstheyactuallyhadtheeffectwouldnothavebeenjustasitwas:Noothereventwould,underthesamecircumstances,haveledtopreciselythiseffect.Ontheotherhand,someothereventslnayhavebeenessentialtotheoccurrenceofthiseffectonlybecauseofoneparticularcharacteristictheypossessed,andanyeventsharingthisspecialcharacteristicmightthereforebesaidtohavebeensubstitutableforitwithoutalteringtheeffectthatactually·occurred.Inourexample,forinstance,itisdoubtfulwhetherthefinalresultwouldhavebeenidenticalincharacterifthewingofthehousehadbeensavedbecausefiremenputoutthespreadingfire,ratherthanbecauseitwasbuilttoresistthissortoffire;anexalninationoftheremainsofthewingwouldalmostsurelygiveevidenceastowhichofthesetwoalternativeshadoccurred.Ontheotherhand,ifafirethatisknowntohavestartedinanatticconsumedanentirehouse,thesaIneresult.couldhavebeenbroughtaboutwhetherthefirestartedduetofaultywiringorduetothespontaneouscombustionofapileofoilyrags.Solongaseithersourceofheatwasintenseenoughtoignitewhatwasinimmediateproxilnitytoit,andsolongasthere,veresimilarInaterialsathandtospreadtheBaIne,thefirewouldhavespreadrapidlyfrointheatticthroughtherestofthehouseinanidenticalmanner,regardlessofitsoriginalsource.Inthatcaseweshouldnotbelikelyto,usetheconceptofnecessityindescribingtherelationbetweentheoriginalsourceandthefinaleffect,eventhoughthesourcewasoneofthatsetofconditionswithoutwhichthisspecificeffectwouldnothaveoccurred.Wewouldnotdosobecausewerecognizethatsomeoccurrenceslnightbesubstitutableforotherswithinanygivensetofcircumstances.Thisdoesnot,however,alterthefactthatsincewearedealingwithwhatdidactuallyeventuate-andnotwithwhatmighthavebeentrueunderanothersetofcircumstances-itislegitimatetoclaimthattheeffectwasanecessaryconsequenceofwhatdidactuallyoccur. 108CAUSATIONOnthebasisofsuchexampleswemaysaythatthesenseof"necessary"thatisapplicabletothecausalrelationhereunderconsiderationisarestrictedorconditionalone:Giventheentiresetofcausalconditionsthatdidoccur(andnoothers),itwasnecessarythattheeffectshouldhaveoccurred.ThisobviouslydoesnotentailthattheeventwasdeterminedeitherinthemetaphysicalsenseorintheempiricalsenseinwhichIhaveusedtheconceptofdeterminism.Wearenot,forexample,committedtosayingthat,giventhenatureoftheuniverse,itwasdeterminedthateachoftheconditionsthatoccurredmusthaveoccurredjustasitdid,andwhenitdid;norareweclaimingthatwhatoccurredinthisinstancewasuninfluencedbyexternalfactors,andrepresentedmerelyanecessarycourseofself-transformationoccurringwithinaclosedsystem.Inotherwords,contingencymaybeacknowledgedaroleinallcasesinwhichweexplainaneffectthroughtracingbacktheconditionsuponwhichitdepended,andyetwemaynonethelesssaythat,whatevermayhavebeentheexplanationofwhyjusttheseconditions,andnoothers,occurredwhentheydid,itwasnecessary-giventheiroccurrence-thattheeffectthatdidinfactoccurshouldhaveoccurred. ChapterFiveONWHATANDWHYINHISTORY�w�*.Itmayatthispointseemasifthetwoprecedingchaptersinvolvedanunnecessarybreakintheargumentofabookthatisostensiblyconcernedwithhistoricalknowledge.Yet,itwinrapidlybecomeclearthatonecannotunderstandthewaysinwhichhistoriansexplainthenatureandtheconnectionsofeventswithoutfirstdestroyingtheassumptionthatthecausalrelationshipisarelationshipbetweentwoseparateanddistinctevents,oneofwhich,byitself,mayberegardedashavingcausedtheother.Norwoulditbepossibletounderstandtheimportantbutauxiliaryrolethatlawlikegeneralizationsplayinmany(althoughnotall)historicalexplanations,unlessonehasfirstseparatedtheconceptofcausationfromtheconceptofanexplanatorylaw.Furthermore,itisimportanttodistinguishbetweenthenotionof"necessity,"whichhistoriansaswenaslaymenoftenuse, 110CAUSATIONandabeliefindeterminism.Thus,theargumentofthetwoprecedingchaptersisdirectlyrelevanttoquestionsthatariseconcerningthenatureandimplicationsofhistoricalexplanation.Itisthisthatthepresentchapterisdesignedtoshow.IshallfirstconsiderthenatureofexplanationinwhatIhavetermed"generalhistory"-ratherthan"specialhistories"-foritistherethatwhathasbeensaidreceivesitsclearestexemplifications.IInawell-knowndictunl-lnelnorablebecauseitissoextremeMichaelOakeshottsaid,UChangeinhistorycarrieswithititsownexplanations."lBythishemeanttoindicate,ashemakescleartwopageslater,that"theonlyexplanationofchangerelevantorpossibleinhistoryissimplyacompleteaccountofchange.Historyaccountsforchangebymeansofafullaccountofchange.Therelationbetweeneventsisalwaysotherevents,anditisestablishedinhistorybyafullrelationoftheevents."VariousreasonsledOakeshotttoadoptthisposition,andIshallnotattempttodisentanglethemandlayeachofthembare.2WhatisofinterestisthatOakeshotttookforgrantedthathisviewentailedthattheconceptofcausationhadnoapplicabilitytotheprocessesofhistory.Thisfollowedfromthefactthatheequatedtheconceptofcausationwithasetofmininlalconditionsalwaysassociatedwiththeoccurr,enceofaneventofaspecifictype;3thatis,hewaswillingtoaccepttheconventionalviewofcausation,inwhichthecause-andeffectrelationshipissimplyanindividualinstanceofsomeobservedregularitybetweentwotypesofevents.However,aswehaveseen,thisviewfitsonlysomeshorthandexplanationsineverydaylifeandsomerudimentaryscientificgeneralizations;itisinadequatewithrespecttothosecasesinwhichcausesandeffectsarenotperceivedasdistinctandseparateevents,anditisalsoinadequatewithrespecttotheexplanationscharacteristicoftheadvancedsciences,forinthenlananalysisofwhathasoccurredinaspecificcaseinvolvesfarmorethananappealtodirectlyobservedregularities.Oncethisisrecognized,thecauseofanyparticulareffectisnottobeequatedwithsomespecificpriorevent,butistobeconstruedasthatsetofconditionswithoutwhichtheeffectwouldnothaveoccurredasitdid,whenitdid.Thus,thegapbetweenhistoricaldescriptionandcausalexplanationisnottheunbridgeablegapOakeshotttookittobe. ONWHATANDWHYINHISTORY111ApositioninsomewayssimilartothatofOakeshottbutdevelopedindependentlyofanidealistnletaphysicsisdefendedbyLouisO.Mink.Indiscussinghistoricalexplanation,Minkholdsthat"theminimaldescriptionofhistoricalpracticeisthatthehistoriandealswithcomplexeventsintermsoftheinterrelationshipoftheirconstituentevents."4Thus,unlikenarrativistssuchasGallie,White,andDanto,whohaveemphasizednarrativestructure,MinkviewshistoricalexplanationnotinterIllSoflinearsequences,butinterIllSofdelineatinghowaprocessisformedoutoftherelationsamongitsnonsimultaneousparts.5AlthoughthereareaspectsofMink'spositionwithwhichIamnotinsympathy,Ifindthathisviewonthisparticularissueiswhollysound.Itispreciselytheviewthatseemsmostadequatewhenahistorianseekstoofferacausalexplanationofanyparticulareffect.Consider,forexample,anillustrationusedbyGardiner,byDray,andbyDanto:theproblemofexplainingthedeclineinpopularityofLouisXIV.6Toexplainsuchachangeintheattitudeofthepeopletowardtheirking,onedoesnotlooktoasetofantecedentconditions,buttotheking'sdecisionsandtothepeople'sreactionstothenl;thatistosay,theexplanationinthiscaseresidesintracingouttheelementsthatenteredintotheprocessthatledfromapointatwhichthekingwashighlyesteemedtotheviewheldofhiIllathisdeath.Silnilarly,totakeanexalnplefroIllMichaelScriven,whenthehistorianseekstoexplaintheriseoftheCityofLondonasafinancialcenter,heappealstochangesincludedwithinthatrise,whichtogetheraccountforit.7Sincenooneofthesechanges,takenbyitself,wasidenticalwiththeriseoftheCity'spower,thatrise(whichoneseekstoexplain)istoberegardedastheeffectofaconcatenationofindividualchangesthat,together,accountforit.Andthis,ofcourse,preciselyfitstheanalysisofcausationIofferedintheprecedingchapters.Itisnownecessary,however,todevelopingreaterdetailthewaysinwhichthisanalysisappliestotheworkofanyhistorianwhoisdealingwithsomephaseofgeneralhistory.Inanygeneralhistory,itwillberecalled,thefactswithwhichthehistoriandealsarealwaysseenasconnectedwiththenatureofaparticularsocietyandchangeswithinit.Evenwhenahistorianis,say,writingthebiographyofapoliticalfigure,orofsomerepresentativemanofhisage,itisinrelationtothesocietyinwhichheplayedarolethatthehistorianviewsthepersonwhoselifeisofinteresttohim.8ThisdoesnotentailthatthebiographerIllUStrelatehissubject'slifetoeveryaspectofthesocietyinwhichthatsubjectlived:NohistorianneedbecomIllittedtoaconcernwithallaspectsof 112CAUSATIONsocietallife.Whathefindstobeofinterestandworthyofstudymaysimplybesomeoneaspectofsocietallife,orsomeparticulargroupofpersonswhooccupyaspecialpositionwithinthatsociety.Whilemanygeneralhistoriansareinterestedinthelong-runchangesinasociety,othersmaybeprimarilyinterestedinstudyingvarioussegmentsoraspectsofthehistoricalprocessincloserdetail.Atpresent,itisprobablytruethatprofessionalhistoriansareapttofocusattentiononspecialtopicsandissues,ratherthantodealwiththenatureofasocietyconsideredasawholeandthechangesthathaveoccurredinit.This,however,doesnotinvolveanabandonmentofinterestinthenatureofasocietyandthechangesithasundergone:Anysocietywillincludeamultiplicityofinstitutionalstructures,andinanycomplexsocietythatextendsoveranappreciableterritorytherearelikelytoberegionaldifferentiationsandsomedegreeofregionalautonomyinitsinstitutions.Sinceanyoftheinstitutionsinasociety,oranyofitsregionalcomponents,mayserveasthesubjectmatterforhistoricalstudy,itisnotincompatiblewithwhatIhavetermedgeneralhistoryforhistorianstorestricttheirstudiestolimitedtopics.Thesespecializedstudiesbelongwithingeneralhistorybecausetheyconcernthenatureandchangesofspecificinstitutionsorofspecificfacetsoftheinstitutionalizedlifeofaparticularsociety.9Thisclarificationofwhatisheremeantbygeneralhistoryleadstoasecondpointitisimportanttorecallfromearlierchapters:Anyhistoricalstudy,ofwhatevertype,hasaparticularsubjectunderinvestigation,anditiswithreferencetothatsubjectthatthehistorian,inhisaccount,includessomeeventsandexcludesothers.Thesubjectsthathistorianschooseforinvestigationvarywidelyinscopeaswellasinkind,andanyhistorianwill,ingeneral,beworkingonaparticularscale,dependingonwhatsubjecthehaschosenasthecentraltopicwithwhichheistodeal.Therewillbemanyeventsthatnaturallybelongwithinhisaccountofthetopichehaschosen,andwhich,evenwithoutinvestigation,heknowscannotbeexcludedfromit.ApoliticalhistorianoftheUnitedStatescannot,forexatllple,failtoincludetheresultsofnationalelectionsduringtheperiodhehasunderconsideration.Aneconomichistoriandealingwithtllodernsocietiescannotexcludefrotllconsiderationtheindustrial,agricultural,anddetllographicfactorsintheeconomyheistoconsider,norcanheneglectitsfinancialstructures,itsmethodsofexchange,itslevelsofconsumption,anditsforeigntrade.Thus,inordertobeapoliticalhistorianoraneconomichistorian,apersoncannotbewhollynaIve,withoutanypriorconceptionofwhatfeaturesofsoci- ONWHATANDWHYINHISTORYetallifearelikelytobeofconcerntohiln.History,likescience,neverstartscompletelydenovo�Theactivityofbeingahistorian,dealingwithaparticularsubject,presupposesthatonealreadyhassomeknowledgeof,andsomeinterestin,thelnaterialswithwhichoneistodeal.Nevertheless,ahistorianc,annotstopwiththefamiliar;inordertounderstandwhytheeventsofwhichhehasknowledgeoccurredastheydid,hemusttracetheirconnectionswithotherevents.Someoftheseeventswillbeincludedwithintheongoingprocesshewishestoexplain,andthushewillsimplybeexaminingtheeventingreaterdetailsothathenlaydiscernitsconstituentelements;inothercaseshewillshowhoweventsthatwerenotthemselvespartofthatprocessenteredintoitsnaturebytheimpacttheyhadononeormoreofitsconstituentparts.Inaccountingforchange,historiansproceedinbothoftheseways.Yet,thesewaysarenotultimatelydifferent,sinceineithercasewhatisofconcerntothehistorianisaccountingforaprocessintermsofchangeswithinitsparts,whetherthosechangeswerebroughtaboutbyexternalforcesorwhethertheydevelopedwithoutsuchinfluences,duetofactorsofadjustlnentandreadjustmentwithintheprocessitself.WereonetoaSSUlnethathistoriansstarttostudyandwritehistoryknowingnothingvhatsoeveraboutthepastwithwhichtheyaretobeconcerned-·weretheytostartasLangloisandSeignobosasslllnedthattheyshouldstart,withnothingbutascatteringofisolatedfactsthatfitintonocontextuntilhistorianssynthesizethem10-itwouldbenexttoilnpossibletoexplainhoworderisbroughtintothen1aterialswithwhichanyhistorianInustdeal.However,ahistorianknowssomethingaboutthenatureofhisownsocietythroughhavinggrownupinit,andhewillhavelearnedthroughitscultures0111ethingaboutitspast;furthermore,inanysocietyinwhichthereisinquiryintothepast,ahistorianwillalsoknowsomethingaboutsocietiesotherthanhisown,andabouttheirpasts.AlthoughsuchinformationInayinitiallybelilnitedandoften111isleading,thedevelopmentofhistoricalinquirieswillsupplen1entandcorrectn1uchthatpassedforhistoricalknowledgeatanearliertilne.(Andthereisnoreasontosupposethatthereeverwillbeatilnewhenhistoricalknowledgecannotbefurtherextendedbothinrangeandindepth.)Whatisimportanttobearinmindisthatsuchknowledgeisneverchaotic.Fromthebeginning,ineventheearliesthistorians,itisboundtohavestructure,sincewhatevereventsappeartobenoteworthywithinthelifeofasociety,orintherelationsbetweensocieties,willpunctuatethe·flowof'tilne,andthesuccessionoftheseeventswillhelpdefinesuccessiveperiodsor CAUSATIONphasesinthehistoryofthesocietywithwhichahistorianisconcerned.Suchaframework,onceadopted,providesacontextintowhichhistoriansattempttoplacefurtherfacts,andashistoricalinterestsbroadenandchange,theframeworkitselfwillinsomerespectschangebyvirtueoftheknowledgeaddedtoit.Lookedatinthisway,whateverentityprovidesthesubjectluatterofahistorian'saccountisseenbyhimasbelongingwithinsomebroaderframework.Atthesametime,itisalsoseenashavingsomenatureofitsown,sinceanyhistorianwhoelectstodealwithacertainsubjectmattermusthavesomeknowledgeconcerningit,forhewouldotherwisehavenointeresteitherinfindingoutmoreaboutit,orinrelatingittoothereventsconcerningwhichhealreadypossessessomepreliminaryknowledge.This,Isuppose,isobviousenough.Whatmustnotbeoverlookedisthatthispreliminaryknowledgeisitselfhistoricalknowledge,howeverprimitiveitslevelmaybe.Itwinofcoursebesubjecttocorrectionaswellastoexpansion,buttheknowledgeahistorianseeksconcerningtheeventswithwhichhedealsisnotwhollydifferentincharacterfromhispreliminaryknowledgeconcerningthem:Atanystageinhiswork,itisonthebasisofsomeantecedenthistoricalknowledgethathemustbuild.Andso,asIhavesuggested,historicalinquiryneverstartsabsolutelydenovo.Now,theeventsthatahistorianknowsinapreliminarywaypertainingtothesubjectmatterinwhichheisinterestedformpartofitshistory.Sucheventsandtheothersofwhichhemaylearnandtheconnectionsamongthemthatheseekstoestablisharethehistoryinwhichheisinterested:Torecounttheseeventsastheyoccurredandintheirrelationshipstooneanotheris-quitesimply-towritethehistoryhesethimselftowrite.Evenifoneweretoattributetoanation,ortosomeotherhistoricalentity,arealitythatisnotexhaustedbywhatoccurredinitshistory-aswethinkofapersonassomethingmorethanthesumofhisactionsandofwhatbefellhilllnevertheless,inpresentingthehistoryofanation,thatnationexistsintheeventsthen1selves,andthey-andnothingelse-constituteitshistory.Similarly,whenoneisconcernednotwiththehistoryofanationtakenasawhole,butwiththenatureandchangesofsomeoneofitsfacets,itiswhathappenedwithinthisfacetofsocietallifethatisthehistorywithwhichthehistorianisconcerned.Thisisnottosay,ofcourse,thateitherasocietyoranyfacetofsocietallifedevelopsinaccordancewithsomeinnerdialecticofitsown,withoutbeinginfluencedatvariouspointsbytheimpingementofeventsthatotherwiseformnopartofitshistory.However,in ONWHATANDWHYINHISTORY115impingingonittheyenteritshistory.Theireffectuponitbecomesapartofthathistoryeventhoughothereventsofthattypeneveragaincomeintocontactwiththatparticularstrandofhistory.Thehistoryofsociety,then,consistsofacontinuingstrandofrelatedevents,theabsenceofanyoneofwhichwouldmakeadifferenceinthecharacteristicsofwhatdidinfactoccurintheparticularsubjectunderinvestigation.11Oneimmediateobjectionmaybeanticipated.Surely,itmightbesaid,ahistorianisnottobeexpectedtoincludewithinhisaccounteverythingthatactuallyoccurredwithinasocietyoroccurredwithrespecttosomefacetofthatsociety.Nopoliticalhistorianneedmentioneverytown,norneedabusinesshistorianincludeinhisaccountaconsiderationofeverystockflotation,everybankruptcy,oreveryadjustmentintarriffs.Werethisnecessary,itisobviousthatnosuchhistoriescouldeverbewritten.Theobjection,however,overlooksthefactthathistoriansworktoscale,andthescaleonwhichtheyworkisdeterminedbyhowtheydefinethesubjectmatterwithwhichtheyaretobeconcerned.AsIhavenotedinusingtheanalogyofmapmaking,ahistoriancandealwithhissubjectmatterinthelarge,showingwhatseriesofeventswentintoitsformation,buthecanthenshifthisattentiontoanyoneoftheseeventsandanalyzetheother,quitedifferentseriesofeventsonwhichitdepended:Heneednotknowtheminutiaecontainedwithinthecomplexprocessbeforeheisabletopickoutthemacroscopicfeaturesofthatprocess.Tobesure,itisalwayspossiblethathemayfind,asheproceedstoanalyzethesubeventsmorecarefully,thatsomeoftherelationshipsheoriginallybelievedtobepresentareopentoseriousquestion.However,whenthisdoesnotturnouttobethecase,hemaybesaidtohaveexplainedtheeventswithwhichhewasoriginallyconcernedbyshowingthattheyresultedfromthesubeventsincludedwithinthem,asoneexplainsotherwholesbyanalyzingtherelationshipsamongthepartsofwhichtheyarecomposed.Inhistoricalexplanations,ofcourse,oneisdealingwithatemporalwhole,notwithawhole,suchasawatch,thatispresentallatonce;butjustastheexplanationofhowawatchrunsdependsuponshowingtherelationshipsamongitsparts,soonewayofexplainingaparticularprocessistobringtolighttheparticularseriesofeventsoutofwhichitwascomposed.(Cf.whathasalreadybeensaidconcerninghowahistorianaccountsfortheoutcon1eofaparticularelection,chapter2,section2.)Thereishowever,anotherwayofdealingwiththerelationbetweenwholeandpartsingeneralhistory.Thelifeofasocietyneednotbe 116CAUSATIONviewedprimarilyasasuccessionofphasesorevents,butintermsofthecomponentstructuresthatconstitutethebasicfeaturesofthatsociety'sorganization.AsIsaidinchapter2,thesetwomodesofviewingtherelationofwholeandpartarenotmutuallyexclusive,andintheendtheymustbeusedtosupplementeachother.Yet,forpurposesofexpositionInowturntothesecondwayofviewingthewhole-partrelationship:thatwhichgivesrisetothehistorian'sinterestinspecificfacetsofsocietalorganizationratherthaninthecontinuitypresentinanytemporalsegmentofthesociety'slife.Onceagainwemustnotethatthehistoriandoesnotstartfromtotalignorance,withoutknovledgeoftheinstitutionalstructurespresentinhisownsociety:Hehasgrownupinthem,andhaslearnedtoreactwithrespecttothem,andisinfactbroughtupwithanentirenomenclaturerelatingtotheorganizationofthesocietyinwhichhelives.Gradually,ofcourse,one'sviewofthenatureandtheinterrelationshipsofinstitutionalstructureswillchange,andaccumulatedhistoricalknowledgeconcerningthetypesofinstitutionsfoundinothersocietiesandthedifferencesintheinterrelationshipsamongthemwillleadahistoriantoafarmoresophisticatedviewofinstitutionalpossibilitiesthanhewouldhavepossessedhadhelearnedaboutsocietalstructuresonlythroughhispersonalexperience.Anthropology,sociology,economics,studiesofcomparativegovemment,andthelikeplacehistoriansonthealertastothediversityofthestructuresandthedifferencesintheirinterrelationshipsthataretobefoundindifferentsocieties.Itmightbethoughtthatpsychology,too,hasanindispensableroletoplayinthisregard,pointingouttheneedsofmenthatmustbesatisfiedinandthroughsocieties.Yet,apsychologicalapproachtosocialinstitutionshasinthemainprovedmisleading,sinceanyparticularneedmaybefulfilledthroughdifferentinstitutionsandanyoneinstitutionmayhelptosatisfydifferentpsychologicalneeds.Itwasinfactoneofthemostdisastrousalthoughoneofthemostnatural-errorsinthesocialsciencesofanearliergenerationtosupposethatthestructuralelementsinsocietalorganizationcouldbestbeunderstoodthroughrelatingthemdirectlytospecificpsychologicalneeds.12Theoriginalbasisforourunderstandingofsocietalstructuresis,then,theexperienceofanindividualingrowingupinhissociety,andtheenlargementofhorizonsthatcomesthroughaknowledgeofothersocieties.OnceagainImustpointoutthatthisgrowthinsophisticationcomestothehistorianthroughdatathatheandhisfellowhistorians,andothers,havegatheredconcerningvarioussocieties.In ONWHATANDWHYINHISTORYattemptingtofitthesedataintotheiraccounts,historianswilloftenfindthemselvescommittedtofurtherinvestigationsasameansofresolvingcontradictionsorforthesakeoffillingingaps.Theywilldosobysearchinginmaterialstheyhavepreviouslyoverlookedforfurtherrelevantdata.Thisdoesnotcommithistorianstoestablishinggeneralizationsconcerningsocietalstructures;whattheyareendeavoringtofindareconnectionsthatmayexistamongthesestructuresinwhateverparticularsocietyistheobjectoftheirconcern.Whileitistruethatgeneralizationsdrawnfromthesocialsciencesaresometimesusefulinthisrespect-andthisisafacttowhichweshallshortlyreturn-itisimportanttonotethatahistorian'sknowledgeofconnectionsamonginstitutionalfactorsneednotrestonanygeneralizationswhatsoever:SuchconnectionsareoftendirectlyevidentwithinthesourcematerialswithwhichhistoriansdeaLForexample,ifapoliticalhistorianisdealingwithchangesinthepoliticalcomplexionofaparticularregion,theimpactofeconomicchangesforthatregionwillbeevidentinmanyspeechesinthelegislature,ineditorialsinlocalnewspapers,incampaignappeals,andinothermaterialswithwhichoneunavoidablydealsinwritingpoliticalhistory.Similarly,thesamesortsofmaterials,whicharereadilyavailable,usuallypermithistorianstotracetheimmediateimpactofsonlepoliticaldecisionontheeconomy,eventhoughtheymaylaterhavetorelyoneconomictheoryinordertotracethefurtherrepercussionsthatthedecisionhadfortheeconomyasawhole.Thepointisthatthehistorian'saccountisstructuredbyrelationshipsthatareclearlyindicatedinthematerialswithwhichhedeals.Thesematerials,independentlyofantecedenttheories,oftenservetodirecthisattentiontotheimpactofotherinstitutionsontheparticularfacetofthesocietywithwhichhehaschosentodeaLItisatthispointthatonecanbestseethecomplementarycharacterofthetwowaysofviewingtherelationsofpartandwholeinunderstandingthehistoryofanysociety.Ontheonehand,thehistoryofasocietyismadeupofacontinuousseriesofeventsandcanbestudiedlongitudinallyovertime;ontheotherhand,asocietyhasmanyfacets,orspecificinstitutionalstructures,whosechangesareofinteresttohistorians.Thesechangesbringaboutchangesinotheraspectsofthesociety,andsoareimportantfortracingthebasiclongitudinalchangesthesocietymayhaveundergone.Italsofollows,however,thatthehistoryofanyinstitutionwithinasocietycannotbeisolatedfromwhatoccursinatleastsomeofitsotherinstitutions,andthusthehistoryofaninstitutionmustbeseeninitscontextas CAUSATIONpartofthecontinuinglifeofthesocietyasawhole.Thecomplementarityoftheseapproachesunifiesthefieldofhistoricalstudiesinsofarasdifferenthistoriansdealwiththesamesocietyindifferentways,anditallowsafullerreconstructionofthenatureandchangesofparticularsocietiesthanwasthecasewhen-inanearliererahistorianstendedtobemoreexclusivelyconcernedwithviewingsocietiesintermsofpoliticalinstitutions,seekingthebasisofbothcontinuityandchangeinthenatureoftheseinstitutions.Whenviewedinthatway,theproblemoftheexplanationofcontinuingstabilityinspiteofmajorpoliticalchanges,andtheproblemofchangeitself,raiseddifficultiesthatalessrestrictedviewofhistoryavoids.Itiswiththeproblemofexplanationandwiththerelationof"what"and"why"ingeneralhistorythatthenextsectionofthischapterwillbeconcerned.IIAswehaveseen,Oakeshottbelievedthatinordertoexplainwhatoccurredinhistory,nothingisdemandedbeyondamorecompleteaccountofthenatureofthechangesthatdidoccur;ashesaid,"Historyaccountsforchangebymeansofafullaccountofchange."Thissuggeststhatthereisnoultimatedistinctiontobedrawnbetweenwhathappenedandwhyithappened,andthisviewissharedthoughonothergrounds-byCollingwood.SofarasOakeshott'spositionisconcerned,itislessparadoxicalthanitinitiallyappears.Theparadoxpersistsonlysolongasoneassumesthatwhatistobeexplainedissomesingle,isolatedevent,andnotsomecontinuousseriesofevents.Inordertoexplainthecharacteristicsofsuchaseries,onemustrefertotheelementsthatformedit,foritistheythatmadeitpreciselywhatitwas.Tobesure,asOakeshottsaw,onecanalwayspursuethematterfurther,askingwhytheseelementsthemselveswereastheywere.Ananswertothesefurtherquestionswouldinvolveashiftinscale,butnotashiftinthemethodofanalysis:Inattemptingtoexplaintheoccurrenceofeacheventthatwentintotheseries,thehistorianwouldregarditastheterminalpointinanotherprocess,tracingtheelementsthattogethermadeupthatprocess.Onceagain,then,hisexplanationwouldhavetakentheformoftracingoutaseriesofeventsthathadactuallyoccurred.Itwillbemycontentionthattheforegoinggeneralremarksareequallyapplicabletohistoricalaccountsthatareprimarilysequential ONWHATANDWHYINHISTORYinstructureandtothosewhoseprimaryaimistoexplainsomespecificoccurrenceorstateofaffairs.Insofarasexplanatoryaccountsareconcerned,wehave,forexample,seenthatifonewereaskedtoexplainsomespecificstateofaffairssuchasthedeclineofpopularityofLouisXIV,ortheriseofthefinancialinfluenceoftheCityofLondon,onewouldtracetheseriesofeventsthat,together,constitutedthatdeclineontheonehand,orthatriseontheother.Thesamesituationobtainsinanyotherformofexplanatoryhistory,inwhichitisaskedwhyaseriesofeventshadpreciselythisoutcome,andnotanother:Onestartsfromaterminalstateofaffairs,andanalyzesthespecificseriesofeventsthat,infact,leduptoit.Similarly,inasequentialhistory,theunderstandingofhowonestateofaffairsledtoanotherinvolvesanattempttotracetheinterveningsequenceofstepsfromtheoriginalpositiontowhateverlaterconditionsfollowedit.Thus,forbothsequentialandexplanatoryhistories,Oakeshottwascorrectinholdingthattheusualdistinctionbetweenwhathappenedandwhyithappenedtendstocollapse;however,itisbynomeansclearinhisaccounthowtheeventsinanyparticularseriesofoccurrencesarerelatedtoeachother,sothatitmaybesaidofthemthattheyformacontinuousseries.IbelievethatthisdeficiencyinhisanalysiscanberectifiedbyintroducingtwoconsiderationsthatOakeshottfailedtointroduce,largelybecauseofhismetaphysicalpresuppositions.ThefirstoftheseconsiderationsconsistsinacontentionIattemptedtoestablishintheprecedingsection,thattheprimary,orinitial,relationshipbetweentheeventswithwhichanyhistoricalaccountdealsisarelationshipofwholeandpart.ThehistorianisconcernedwithSOlnespecificsubjectmatter,notwith"historyingeneral";giventhissubjectInatter,thedatawithwhichheworksguidehiminseeingwhatpartsbelongtogetherandwhatpartsinfluenceoneanother,andthuswhichspecificeventsaretobetreatedascomponentswithinsomeparticularwhole.Inthesecondplace,asIhavealreadysuggested,therealsoarepointsatwhichtheoreticalgeneralizationsmaybeessentialtothehistorian'sunderstandingoftherelationshipsan10ngevents.Forexample,asInoted,theinitialimpactofapoliticaldecisionontheeconomyofacountrymaybeevidentinthematerialswithwhichaparticularhistoriannecessarilydeals,butanunderstandingofthefurthereconomic.repercussionsofthatdecisionmaypresupposeanacquaintancewithsophisticatedeconomictheories.Ahistorianwholackssuchknowledgemayonlybeinapositiontodealwithimmediateconsequences,andnotbeable 120CAUSATIONtoexplainthelonger-runchangesthattheimmediateconsequencesengendered,eventhoughwhathappenedatalatertimemaybeimportantintheseriesofeventswithwhichheisconcerned.Vhatholdsinsuchacaseholdsinmanyothers,andmuchofthemortarinhistoricalaccountsis,asIshallnowshow,afunctionnotofdirectdescriptionbutofgeneralizationsconcerninghowvariousfactorsthatmaybepresentinmanydifferentoccurrencesaffectthesituationswithwhichhistoriansdeal.Themostfamiliarofthesegeneralizationsconcernthewaysinwhichhumanbeingsmaybeexpectedtobehaveinvarioustypesofsituations,andsuchgeneralizationsareapttobeadmittedasimportantforhistoricalinquiriesevenbythosewhoarelTIOStopposedtotheintroductionofgeneralizationsofanyothersort.Thereasonforthisexceptionisthatitisuniversallyadmittedthathistoriansmustrelyontheir"knowledgeofhumannature"inordertoexplainwhyindividualsactastheydo.Theknowledgepresupposedisnot,however,aknowledgeofhowindividualsalwaysbehavewhentheyfacesituationsofcertainspecifictypes.Itis,instead,aknowledgeoffactorssuchaspride,loyalty,envy,courage,fear,asenseofrectitude,orgreed,whichmaybeevidencedinmanydifferentsortsofactions;also,itincludesanawarenessthatemotionsmayaffectaperson'sjudgment,thatpersonsmaybeinfluencedbywhatothersdoandsay,andofothergeneralfactorsthataffectthebehaviorofindividuals.Whilethisknowledgeofhumannatureisgeneral,beingapplicabletomanydifferentindividualsindifferentsituations,itdoesnotconsistinasetofgeneralizationsastohowindividualswillbehavewhenevertheyareconfrontedbyparticularsortsofsituations.Theinappropriatenessofattemptingtoformulatesuchgeneralizationsdoesnotstemfromanyconsiderationsconcerninghumanfreedom.Itfollowsfromthefactthathumanactionsvaryinaccordancewiththenatureandstrengthofanyspecificindividual'spsychologicaltraits,thesituationinwhichheisplaced,andhow,atthetime,heviewsthatsituation.Itisthereforenotpossibletoformulatesimplegeneralizations,exceptofaverycrudetruncatedtype,thatservetoconnectsituationsofagiventypewithspecificmodesofindividualaction.Notonlyisthisthecasewithrespecttohowdifferentpersonsbehaveindifferenttypesofsituation,itisalsotrueofhowthesamepersonbehavesunderdifferentcircumstances.Forexample,apersonwhoseactionsareusuallydon1inatedbygreedmayononeoccasionbemovedbypity,andhewillthenbehavecharitablytowardapersonwhomhe ONWHATANDWHYINHISTORY121wouldordinarilycheat.Hismotivationcannot,however,beknownatleastbyothers-untilafterhehasacted;thenonecandrawinferencesfromhisactionastohowthesituationprobablyappearedtohimatthetime.This,ofcourse,is.preciselywhathistoriansdowhentheyinterprettheactionsofothers;itiswhattheymustdoeveniftheagenthasexplainedhisownmotivationinjournals,letters,interviews,andbooks,forthatevidencemustalsobeinterpretedinterms.ofthehistorian's"knowledgeofhumannature"ifheistoassessitsreliability.One'suseofthisgeneralknowledgemaysometimesbemisleading,andisoftenespeciallymisleadingwhenoneattemptstointerprettheactionsofpersonswholivedinanotherageorinanunfamiliarculturalmilieu.Nevertheless,thepossibilityofbeingnlisledshouldnotbetakenasdiscreditingtheclaimthatthroughourownself-knowledgeandthroughtheobservationofotherswearriveatsomemeasureofunderstandingofthecapacities,theformsofreaction,thewaysinwhichdecisionsarereached,andmanyoftheotherdispositionaltraitsofpersonsthataregeneralinbothoftwosenses:lneyarewidespread,andtheyareevincedinmanydifferenttypesofsituations.Inthesetwosensestheirgeneralityresemblesthegeneralityoftheconceptsandrelationshipsinanadvancedsciencesuchasphysics,forthesearenotonlyapplicabletomanyinstancesbuttheirapplicabilityextendstoinstancesthat,fromadescriptivepointofview,arenotallofonetype.Atthispointweareinapositiontoseehowthisgeneral"knowledgeofhumannature"canserveasmortarinhoidingtogethersomeoftheelementsthatenterintoastrandofhistoricalevents.Giventhefactthattheactionsofaspecificpersonformedoneelementinaseriesofevents;"wemaybepuzzledastowhyheactedashedid:forexample,thathesuddenlychangedfrombeingastrongsupporterofsomemeasuretovoting.inoppositiontoit.Insuchcasesalegislatorgenerallyoffersreasonstoexplainthechangeinhisview,andthehistorianwiilhavetoseektodeterminewhetherthesewereinfacttherealreasonsforthedecision,orwhetherotherreasons,suchaspromisesorthreats,mayhavebeendecisive.Todecidebetweensuchalternativesahistorianmustofcourseknowagreatdealaboutthepoliticalsituationatthetime,buthemustalsousehisgeneralknowledgeofhumannaturetointerpretwhetherthereasonsgivenforthisparticulardecisionareworthyofcredence.Furthermore,ifthislegislator'schangeofmindoccasionedchangesinthevotingbehaviorexpectedofothers,·thehistorianwouldagainhavetousehisknowl- 122CAUSATIONedgeofthepoliticalsituationatthetimeandhisgeneralknowledgeofhumannaturetoestimatewhatfactorswerepresentthatinducedtheotherstochangetheirmindsaswell.Suchinterpretationsarenecessaryifahistorianistoexplainwhy,onaparticularoccasion,ameasurethatwasexpectedtopasswasdefeated;andwhatistrueinarelativelysimplecase,suchasthis,isalsotruewheneverhistoriansmustinterprettheactionsofthevariousprotagonistsinanyhistoricdecision.SeveralimportantaspectsoftheforegoingclailTIslTIUStbeexplicitlymentioned.Inthefirstplace,aknowledgeofhumannature,takenbyitself,isaninsufficientbasisforanyhistoricalexplanation.Asisalsothecaseinthesciences,inordertousegenerallawsinexplaininganyoccurrence,onemustpossessknowledgeoftheinitialconditionsthatwerepresent.Intheprecedingillustrativeexample,theseinitialconditionswouldincludebothpriorknowledgeofthenatureandfunctioningofpoliticalinstitutionsatthetime,andknowledgeofthespecificpoliticalsituationinwhichtheproposedmeasurewasbroughtupforconsideration:Itisonlybecausehepossessessuchknowledgethatahistorianisinapositiontousehisgeneralknowledgeofmen'smotivationtoexplainwhatoccurred.Inthesecondplace,itnlustbeadmittedthatthegeneral"knowledgeofhumannature"involvedinhistoricalexplanationsdoesnotconsistinclearlyformulablelawscomparabletolawsformulatedinanyadvancedscience.Thisisnotprimarilybecausethefactorsinvolvedarenotdealtwithquantitatively,ashasoftenbeenassumed;amorebasicreasonisthatthefactorsintermsofwhichweformulategeneralizationsconcerninghumannaturehavenotbeenreachedthroughequallyrigorousabstractiveanalysis,andtheyarethereforelesssimpleandlessgeneralthanthefundamentalconceptsusedinanyadvancedscience.Nevertheless,becausetheyremaincloselytiedtotypesofactionswithwhichweareacquaintedineverydaylifehistorianscanutilizethem.Althoughithasbeenpossibletodeveloparelativelyadvancedscienceofexperimentalpsychologyinsomeareasofhumanbehavior,theresultsthathavebeenobtainedintheseareascannotbedirectlyappliedtoexplainthesortsofactionswithwhichhistoriansareconcerned:Therelevantinitialconditionsfortheirapplicationaresimplynotknowninsuchcases.13Inthethirdplace,however,itistobenotedthatthelackofrigorintroducedintohistoricalexplanationsbytheabsenceofpreciselyformulatedlawscanbecompensatedforinwaysthathavenotrueparallelintheexactsciences.Althoughthescientistisnecessarily ONWHATANDWHYINHISTORYconcernedwithparticularcaseswhenheformulatesandtestshisgeneralizations,theanalysisofwhathappensinanysinglecaseofagiventypeisonlyincidentaltohismainpurpose,whichistoexplainallcasesofthattype.Thehistorian,ontheotherhand,attemptstoexplaintheparticularcase,andheusesgeneralizationsonlyincidentallyforthispurpose.Thefactthathisgeneralizationsareloosedoesnotpreventhimfromofferingincreasinglyaccurateandconvincingaccountsofwhathappenedinanyparticularcase,sinceoncehehasobtainedguidancefromageneralizationhecanlookforfurtherevidenceeithertosupportortomodifyhisuseofitinthecaseathand.Hecandosobecausetheoccurrenceswithwhichhe,asageneralhistorian,dealsformanindefinitelydenseseries:Havingofferedanexplanationofthelinkagebetweeneventsatonelevel,hecanpursuehisinquiriesinfurtherdetail,inordertoverifytheapplicabilityofthatgeneralizationconcerninghumanbehaviortotheparticularcaseheseekstoexplain.Suchgeneralizations,then,donothavetheexplanatorypowerthattheircounterparts·possessintheexactsciences,buttheyserveasusefultoolsinhistoricalexplanations,andareoftenseentobewhollyappropriateasawayofexplainingthelinkagebetweenspecificeventswithinaparticularhistoricalprocess.14Whathasherebeensaidaboutgeneralizationsthatenterintothehistorian's"knowledgeofhumannature"appliesalsotothegeneralizationsconcerninginstitutionalstructuresthatenterintohisaccount.Tobesure,therearerelativelyprecisegeneralizationsineconomics,andperhapsinothersocialsciences,whichhelptoexplainthelinkageamongdifferentoccurrencesinanongoingsocialprocess.However,inmostcasesthegeneralizationsactuallyusedbyhistoriansinseekingtoexplainthenatureofandthechangesinsocialorganizationarebasedontheiracquaintancewiththeirownandothersocieties,andevenafterstudyinganthropology,sociology,comparativepoliticalscience,andagreatdealofhistory,theirgeneralizationsremainlooseinformulation;theythereforehavegreaterheuristicthanspecificallyexplanatoryvalue.Itakethispointtobein1portant,andIshallthereforedealwithitinS0111edetail.15Onecan,aswehaveseen,viewthelifeofasocietynotintermsofasuccessionofphasesorevents,butinterlllSofthecOlnponentstructuresthatarethebasicfeaturesofitsorganization.Amongsuchcomponentsareitseconomicinstitutionsfortheproductionanddistributionofgoods;itssystemsofrulesconcerningkinship,fan1ilyorganization,anddifferentiationoffunctionaccordingtoageandtosex;itslegalandpoliticalstructures;itsreligiousinstitutions;andthe CAUSATIONlike.Itwouldseemthatinanygivensituationanyoneofthesecomponentsmayactononeormoreoftheothers,asinoursocietypoliticaldecisionsaffecteconomicaffairs,andecononlicchangesnotonlyleadtotheenactnlentofnewlawsbutmay,forexample,alsohaveadirectilllpactontheorganizationoffalllilylife.Ofcourse,anyoneofthesecomponentsisitselfcomplexanditsvariousaspectsaffectoneanother,sothatchangesinsomesectorofpoliticallifewillhaverepercussionsonotheraspectsofthepoliticalstructure,andanydrasticchangesinonepartofaneconolllYarelikelytoaffectthewhole.Itiswithsuchinterrelationshipsthatstudiesofgovernmentandthedisciplineofeconomicsarelargelyconcerned.Whilethegeneralizationstobefoundineconomicsarebothabstractandrelativelyprecise,thisisnotthecasewithrespecttomostgeneralizationsinthesocialsciences.Thelessabstractandlessprecisegeneralizationsthathistorianscanborrowfromtheothersocialsciences,orwhichtheyformulateinterlllSoftheirownknowledgeofothersocieties,mayneverthelessperforlllausefulheuristicfunction,justasdogeneralizationsconcerninghumannature.Furtherinvestigationmayprovethesegeneralizationstobecorrectinasufficientnumberofcasessothattheycanberegardedasofferingreasonablehypothesesastowhatmaybetrueinfurthercases,eachofwhichcanthenbeexaminedindetail.Forexample,indiscussionsofalternativeformsofgovernlllentitisacOlllmonplacethatdirectdelllocracycanexistonlywherethepopulationandtheareatobegovernedarebothoflimitedsize;andthatwhenthesizeofthepopulationandoftheterritorygrow,andthetasksofgovernmentbecomemorecomplex,directdemocracylllUStbesupplementedbysomerepresentativeformofgovernment,ifdemocraticproceduresaretosurvive.16Ageneralizationofthistype,linkingsizeandformsofgovernlnent,isnotprecise,nordoesitinvolveahighdegreeofabstractioninthetypesoffactorsitseekstorelatetooneanother;nonetheless,itisacommon-sensegeneralizationthatseemstofitin-manycasesinwhichonedealswithvariousformsofgovernance:notonlystategovernancebutgovernanceinreligiousbodies,inunionactivities,andinmanyotherinstitutions,suchasuniversities.Nevertheless,suchageneralizationcannotbeheldtobealawofsocietalorganization,notonlybecauseofitslackofprecision,butbecauseitcannotsupportacounterfactualconditionalstatementassertingthatassizeandareatobegovernedincreaseitbecomesimpossibletomakedecisionsbyvotingexceptthroughsome,formofrepresentativesystem.Asacounterexample,onecanimagine-thoughperhapsnotwithouta ONWHATANDWHYINHISTORYshudder-avotingsystembasedonmoderntechnologyinwhichawholepopulationcanbepolledonanyissuebyinformingthecitizensviatelevisionatannouncedtimesonwhatissuestheyaretovoteandhavinganelectronicvoting-machingterminalavailableineachresidencethatallowseacheligiblepersontovoteonce,butonlyonce,oneachissue.Thus,ageneralizationthatseekstolinkpoliticalfornlswithsizecannotbetakentobealawofsocietalorganization;itisonlyageneralizationthatseemstobeaccurateinmanydifferenttypesofcases,undertheconditionsthatusuallyobtainincasesofthesetypes.Whilenotclaimingnecessityforsuchageneralization,itisreasonabletoapplyittopreviouslyunexaminedcasesinordertoseewhetherornotitleadstoinsightintochangesinthesystenlofgovernancethathaveoccurredinwhateversocietyorwhateverparticularinstitutionsareunderscrutiny.Suchinsightcomesifinvestigationrevealsthatarepresentativesystemwasinfactintroducedtosupplementorsupplantdirectgovernanceinordertocopewithnewtasks,whetherthesearosethroughgrowthofpop'ulationandanexpansionoftheterritorytobegoverned,orwhetherchangesinotherinstitutionsmadeitnecessarytointroducegovernlnentalregulationintonewareasofsocietallife.Thereare,Ibelieve,manygeneralizationsofthissortthatcanbeusedinhistoricalinquiries,suchasthosegeneralizationsthatseektolinkformsofmarriage-'forexample,polyandryandpolygyny-withthewaysinwhichapopulationgainsitsmeansofsubsistence;orthosethatseektointerrelatevariousaspectsofmarriageindifferentsocieties,asTylorattemptedtolinkrulesofresidenceandrulesofdescent,17orasHomansandSchneiderintheirdiscussionofthetwotypesofunilateralcross-cousinmarriagerelatedthesetypestopatrilinealandmatrilinealkinshiprules.18Iwouldnotwishtosuggestthatallsuchgeneralizationshavebeenhelpfulinhistoricalinquiries.InthisconnectiononeneedmerelythinkofgeneraIizations"thatattemptedtolinkformsofgovernmentwithgeographyorclimatetounderstandhowmisleadingtheycansometimesbe.Ontheotherhand,Ialsodonotwishtosuggestthatallsuchgeneralizationsfallintotheclassofthosethatareonlyofheuristicsignificance:Somemayindeedcometobesopreciselystatedandwellauthenticatedthattheycanbeconsideredassociologicallaws,andwhenconjoinedwithadequateknowledgeofinitial"conditionscanbeusedtoexplainthedirectioninwhichchangehasproceeded.19Atpresent,however,itisatleastdoubtfulthatonecanpointtoinstancesofsuchlaws.Itis,forexample,extremelyplausibletotreatthatmostfamousofallmodern CAUSATIONsociologicalgeneralizations,theMarxiandoctrineoftherelationsbetweensubstructureandsuperstructure,asbeingaheuristicgeneralizationonly,ratherthanasbeing,inanystrictsense,alawconcerningsocietalchanges.Infact,inMarx'sprefacetohisIntroductiontotheCritiqueofPoliticalEconomyhespeaksofhisdoctrineashavingservedasaguidingthread(Leitfaden)inhisstudies,whichisnot,ofcourse,thesameasholdingittobeasociologicallaw.IdonotwishtosuggestthatMarxdidinfactregarditasaheuristichypothesisonly;onthecontrary,Ibelievethatheregardeditasastatementofasetofnecessaryrelationships,eventhoughheleftitunclearastohow,ifatall,theserelationshipscanservetoexplainallofthosespecificchangesinthetotalorganizedlifeofasocietywithwhichhistoriansmustdeal.Inmyopinion,oneshouldnotinfactregard"theMarxianhypothesis"asalawinthesensethatit,togetherwithaknowledgeoftheinitialconditionsthatwerepresent,wouldpermitonetoexplainconcreteevents.Nonetheless,itcanbeextremelyusefulinawidevarietyofinstancesofonetreatsitasaheuristichypothesisonly,usingitinasearchfornewevidenceconcerningsomeofthefactorsinvolvedinsocialchange,withoutassumingittobetrueofotherinstancesinwhichnoconvincingevidenceofitsapplicabilityhasbeenfound.Wearenowinapositiontounderstandhowaseriesofeventscanbesaidtobelongtogetherastherelatedpartsofasingleprocess,andtodosobecauseoftheirconnectionswithoneanother.Fromwhathasbeensaiditcanbeseenthattheeventswithwhichahistoriandealsintracingaprocessmaybelongtogethereitherbecausetheyare,quitesimply,constitutivepartswithinthatprocess,orbecausetheyhaveentereditthroughinfluencingoneormoreoftheseparts.Inspeakingoftheconstitutivepartsofaseriesofevents,Irefertothefactthatwhenahistorianseekstounderstandthenatureofandchangesinasociety,orinsomeaspectofthatsociety,heisdealingwithacomplexwhole,someofwhosepartshealreadyknows.Itistheseparts-andanyotherswhoseexistenceheuncovers-thatarethepartsofthewhole,astheseriesofplaysinafootballgame,orthefourquartersintowhichthegameisdivided,constituteitsparts.Inasociety,aswehaveseen,onecanalsoregarditspartsasbeingtheorganizationalstructuresthat,together,composeit,andthisstructuralwayofviewingthepartsofasocietyiswhollycompatiblewithviewingitasthesequenceofeventsthattogetherformitstemporalhistory.Infact,aswehaveseen,thesetwowaysofviewinganyhistoricalsubjectmatterarenotonlycompatible,butfuseinalmost ONWHATANDWHYINHISTORYanyhistoricalinquiry.Thus,onecanseethatwheneverahistoriancorrectlyanalyzesthestructurespresentinasociety,orwheneverhegivescorrectinformationastothesequenceofchangesthatitoranyofitsaspectshasundergone,hehasdealtwitheventsthatbelongtogetherbecausetheyarethepartsofonecontinuingwhole.Suchawholeisnotformedmerelybecausethehistorianhasdefinedhissubjectmatterinacertainwayandhasconfinedthescopeofhisinquirytowhatoccurredwithrespecttothatparticularsubjectmatterinacertainplaceandoversomerestrictedperiodoftime.Rather,theeventsthatheincludesasbelongingwithintheseriesofoccurrenceswithwhichheistodealarethosebetweenwhichhefindsinherentconnectionsbecausetheyhaveinfluencedoneanother.Inthefirstinstance,hisawarenessofsuchinfluencesarisesbecausehefindsthemexplicitlyindicatedinthematerialswithwhichhemustdealinhisinquiry,aswehaveseentobethecasewhenanyoneattemptstosay,concretely,whatactuallyoccurredinthecourseofanelectioncampaign.However,aswehavenowseen,whatisevidentinthefirstinstancemaybesupplementedbytheuseofhypothesesdrawnfromthehistorian'sknowledgeofhumannatureandfromhisacquaintancewithothercases.Suchhypotheses-especiallywhenheispuzzledbyanyturnofevents-willleadhimtolookforevidenceofconnectionsthatwerenotobviousintheoriginalmaterialswithwhichhedealt.20Inthisway,Inanyhypothesesthatareneitherprecisenorwellconfirmedmayservetosuggestwhereahistorianistolookforrelationsthatmayexistbetweenvariousfactorseffectingthecourseofevents.Ifhefindsevidenceofsuchrelations,hewillhaveestablishedinmoredetailanexplanationofhowitwasthataparticularseriesofeventscametotheconclusionitdid.And,ofcourse,shouldtherebeanycasestowhichanypreciseandwellconfirmedpsychological,economic,orsociologicallawcouldbeapplied,suchalawwouldservetocementtherelationshipbetweentheeventsthatwereinvolved.Atpresent,itdoesseemthatatleastthoserelationshipsthatthedisciplineofeconomicsisabletotracecanbeimportantforhistoriansinpreciselythisway.ChangesmaybebroughtaboutinasocietynotonlybecauseSOlneofitsstructuresundergochangeandaffectitsotherparts,butalsobecauseofexternalevents.Forexample,adroughtoranearthquakemayinfluencelifeinaparticularsociety,andthefactofitsoccurrencewillthenenterintothesociety'shistory,buttheeventitselfcannotbesaidtobelongwithinthathistoryinthesenseofbeingoneofitsconstituentparts.Entirelyparallelexamples,thoughonesthat CAUSATIONhavesometimesledtoconfusion,arecasesinwhichwhatoccursinonesocietyinfluenceswhatoccursinanother,asawaroradepressioninonecountrymayaffecttheeconomyofanother.Inthesecases,too,thefactthattherehasbeensuchawarordepressionwillproperlyenteranyhistoricalaccountoftheaffectedcountry;yetthewarorthedepressioncannotthereforebesaidtohavebeenapartofthatcountry'sownhistorysinceititselfwasnotatwar,nordidititselfsufferaneconomicdepressionatthattime.21Theinfluenceofanysuchexternaleventsonthehistoryofaspecificsociety,oronanyofitsinstitutions,canbetracedbyhistoriansinexactlythesamewaysastheytraceinternalchanges:forexample,byexaminingevidencethatrelatestohowpeopleatthattimeregardedtheseevents,andbymeansofevidenceastowhethertheirperceptionsoftheconsequencesoftheseeventswereorwerenotsound,and,inaddition,onthebasisofevidenceastotheshort-runeffectsthattheseperceptionsthemselvesmayhavehadonthesituation.Also,ofcourse,externaleventssuchasadroughtmayactdirectlyonthesizeofapopulation,justasaninvasionmayoverthrowanautonomousstate.Thefurtherconsequencesofanysucheventswillnotdependprimarilyuponhowthesituationwasperceivedbythepeopleaffected;thesocietyanditsinstitutionswillnothavechangedduetointernalfactors(paceHegelandToynbee),butthroughexternal'factorsthatintervenedinitsongoinglife.Thecollectionofallsuchevidence,themarshalingofitinaconvincingfashion,andawillingnesstosearchforfurtherevidencethatmayalterthepicturethatearlierevidenceseemedtosuggestcomprisethetaskofanyindividualhistorian.Hisknowledgeoftheviewsheldbyhispredecessorsonthebasisoftheevidencethatwasavailabletothemwillbothhelpandchallengehim,aswillanygeneralization.sconcerningeitherhumannatureorsocietalorganizationthatarenewinhisgeI).erationorthathavebeennewlyrevivedfromthepast.Allofthisraisesquestionsastohowmuchtherewritingofhistorypermitsonetosaythathistoryisadisciplineinwhichreliableknowledgeaccumulates,orwhetheritisafieldinwhichparticularsetsofbeliefsaboutthepastaremerelysupplantedbyothers,withnoneproving,inthelongrun,tohavebeenmoreworthyoftrustthanthosethathadgonebefore.Weshalldealwiththatquestion,whichisthequestionofobjectivityinhistoricalknowledge,insucceedingchapters;atthispointwemustconcludeourdiscussionofhistoricalunderstandingbyexamininghow"what"and"why"arerelatedtoeachotherinspecialhis- ONWHATANDWHYINHISTORYtories,andalsointhosehistoricalstudieswhosestructureisneithersequentialnorexplanatory,butisbasicallyinterpretiveinnature.IIIUnlikesequentialandexplanatoryformsofgeneralhistory,specialhistoriesdonothavecontinuingentities,suchasasocietyorsomefacetofitsstructure,asobjectsofinvestigation.Instead,theydealwithspecificaspectsofhun1anculture,suchasformsoftechnologyandliteraryorartisticstyles,whichoftenspreadacrosstheboundariesofsocieties,followingacourseoftheirown.Also,thesubjectswithwhichspecialhistoriesareconcernedmayhaveaninterlnittentexistence,asonecanseeinfollowingtheinfluenceexertedbyaliteraryorphilosophicfigure,ortheinfluenceofapoliticalorsocialideal.Inallspecialhistoriesthehistorian'sconceptionoftheculturalelementwithwhichheistodealisoffundamentalimportance:Whatconstitutes"literature"orwhatcharacterizes"Gothicarchitecture"orwhatisincludedunderUtechnology"willdeterminewhatistobetakenintoaccountinthehistoryheproposestowrite.Thisisnotthecaseingeneralhistoriesthatareeithersequentialorexplanatory.22Evenwhenhistoriansconfinethemselvestoexalniningparticularaspectsofsocietallife,choosingtoinvestigatepoliticalchangesorchangesintheeconomicinstitutionsorintheeducatiom�alsystemofaspecificsocietyoveracertainperiodoftime,theywillb;edealingwithcomplexstructuresinwhichtherelationshipsamongthepartsarenotdependentuponthehistorian'sinitialconceptionofthataspectofsocietallife.Whateverantecedentassumptionsahistorianmaybringtohismaterials,hewillfindthatifheistodescribethefunctioningofsomeparticularstructureinsociallife,bemustrelateittovariousotherstructures,andindoingsohemaybeforcedtorevisehisconceptionofthenatureoftheinstitutionswithwhichhewasattempting,todeal.Inspecialhistories�ontheotherhand,ifaparticularseriesofworksfallsoutsidethedefinitionofliteratureorofGotbicarchitectureoroftechnologythatahistorianhasadopted,hewillnotbe'c0111pelledtoincludetheminhishistory,eventhoughothers-workingwithotherdefinitions-willdoso.Thus,thelnannerinwhichahistorianusesgeneralconceptstobringorderintothematerialswithwllichheistodealplaysaprimaryroleinspecialhistories.WhewelI,inanotherconnection,referredtothisuseof 13°CAUSATIONconceptsas"colligation,"andanumberofrecentwritershaveappliedthattermtowhattheytaketobethefundamentalcharacteristicofallhistoricalwriting.Thatthesisisnot,however,mypresentconcern.(Foracriticismofthenotionofcolligationasappliedtogeneralhistory,seechapter6,section3.)VhatIherewishtopointoutistheverymarkeddegreetowhichgeneralconceptionsortheoriesconcerningsometypeofsubjectmatterorson1ehumanactivitydefineandlimitthetasksthatspecialhistorianssetthemselves.Take�forexample,thecaseofliterature.Notallprintedworkswillbeclassedasliterature,butonlythosethatconformtoaparticularviewofwhatseparatesliteraturefrolnotherformsofexpressionorcommunication.Similarly,"technology"isusuallynottakenasatermthatincludesalloftheinstrumentsthatlnenn1akeanduse,butreferstothewaysinwhichtoolsareusedtosatisfythebasicneedsofmenlivinginaparticularsociety.Consequently,whilefishhooksandcanoes,andthewaysinwhichtheyareused,areconsideredtobeaspectsofthetechnologyofsomesocieties,theyarenotsoconsideredinours.Ofcourse,thereismuchroomfordisagreementwithrespecttowhetherornotcertainobjectsorcertainactivitiesresembleoneanotherintheirfunctionsoraims.Suchdisagreementswillobviouslylimittheextenttowhichdifferentspecialhistoriansdealingwithwhatisostensiblythesamesubjectmatterwillpresentcomplementaryorconflictingaccounts.Withthatproblemweshalllaterhavetodeal.Atthispoint,whatmustbeexaminedaretherespectiverolesofdescriptionandexplanationinspecialhistories,andweshallfindthatinsuchhistories-incontrasttogeneralhistories-thereisnotendencyfordescriptionandexplanationtocoalesce.Inspecialhistories,whethertheybehistoriesofson1eforlnofartorartmovement,ofphilosophicaldoctrinesorschools,ofscienceoroftechnologicalinnovations,onefundamentaltaskforthehistorianistodescribethenatureoftheobjectsthatenterintohishistory.Hemustunderstandandbeabletopresenttohisreadersthesalientcharacteristicsofthoseworksofart,doctrines,ordiscoverieswithwhichthehistorydeals.Inaddition,ofcourse,hewillattelnpttoshowhowtheseworksmayhavebeenrelatedtooneanother,andtocertaintraditionsinwhichtheystand.Accountsofthelattersortplayalargeroleinanyformofspecialhistory,andIdonotwishtominimizethem.WhatIwishtoemphasize,however,istheextenttowhichdescriptiveanalysesofindividualworksenterintoeveryspecialhistory.Suchdescriptionsalmostinevitablyincludeanimportantelementofevaluation,andtheevaluationofanyworkwillinpart ONWHATANDWHYINHISTORYdependuponthehistorian'sconceptionoftheaimsandthefunctionsofsuchworks,andofthegenretowhichtheybelong.23Thisisnottosaythattheseevaluationsarenecessarilyboundtotheconventionsdominantinthehistorian'sownculturalmilieu:Historiansareinfactabletorecognizethatthearts,forexample,havedifferentaimsandperformdifferentfunctionsindifferentperiods;consequently,theirevaluationswillnotbetiedtowhatartmeansintheirownculture,butwilltakeintoaccounttherelativesuccessorfailureofdifferentworkswithrespecttotheperiodinwhichtheywereproduced.Supervenientuponsuchevaluations,aworkmayalsobeviewedbythehistorianintermsofitsabilitytoevokeadmirationasaworkofartonthepartofthosewhoseviewsoftheaimsandfunctionsofartarenotlimitedtothoseoftheartist'sowntime.Theabilitytotranscendoneperiod,evokingadmirationatothertimes(althoughnotnecessarilyinalltimes),iswhatmakes·aworkaclassic,whetheritbeinthearts,inphilosophy,orinscience.Suchclassics,ofcourse,playanimportantroleinspecialhistories,buttheydonotusurpthestagesolongasthehistorianisengagedinhistoricalinquiry,ratherthaninaspecificallycriticalenterprise,forevenaclassicmustbeunderstoodinrelationtootherworksofthetimeandcannotbetotallyliftedoutofitscontextandunderstoodsubspecieaeternitatis.Thisisnottosaythatanyworkofart,ofphilosophy,orofscienceiswhollyembeddedinthesocietyinwhichitisproduced,forthecomponentsofhumanculturemaypursueanitinerantexistenceand,ingeneral,theyhaveasporadicratherthananuninterruptedinfluencewithinthecultureofthosesocietiesinwhichtheysuccessivelyappear.24Thus,tounderstandaworkinrelationtoitstime,onemustbealerttotheculturalmilieuinwhichitmadeitsappearance,andthisincludesnotmerelythecultureindigenoustothatsocietybutwhateverculturaltraditionsareavailabletothosewhocreatedtheworksthehistorianseekstounderstand.Naturally,noworkcanbeunderstoodmerelyinterlllSofthatwhichhasbeenpassedalongtoitscreator,orcreators,bywhatothershavealreadydone.Worksofart,philosophicargumentsandsystems,theologicaldoctrines,scientificdiscoveries,andtechnologicalinnovationsaretheworkofindividuals,andinadditiontotheavailableculturalheritage-whichindividualsabsorbdifferently,andtowhichtheymayreactindifferentways-thepersonality,intelligence,andotherpredispositionsofanindividual,aswellaswhatheexperiences,willshapetheaimsandthenatureofhiswork.Itisherethatwecometotheproblemofexplanation.inspecialhistories. CAUSATIONInordertoofferanexplanationofthecharacteristicsofaparticularworkorsetofworks,thehistorianmustgooutsidetheworkitself.Hemustinthefirstplacerelateittotheculturaltraditioninwhichitstands,tracingwhateverinfluencesn1ayhavebeenbroughttobearonit.Yet,aswehavenoted,almostanysingleworkwillalsobearsOlneimpressofitsmaker,andthehistorianwhoisconcernedwithofferinganexplanationofwhatmaybemoreorlessuniqueinsuchaworkwillhavetoattempttorelateit,insofarashecan,tothetalents:tlimitations,interests,andexperienceofthepersonwhomadeit.Inaddition,ofcourse,anyculturalproductisapttoreflectsomecharacteristicsofthesocietyinwhichitwascreated,andboththesin1ilaritiestobefoundinaseriesofworksproducedinagivensocietyandthediscontinuitiesinworksproducedbydifferentgenerationswithinthatsocietywilloftenbeexplicableonlyintermsofchangesinvariousaspectsofthesocietyitself.Suchchangesoftenaffecttheideologicalcontentofliterature,philosophy,thearts,orreligion,andevenchangesinstylemaybeexplicableonlyintermsofchangesintechnologyorthroughchangesin.socialstratificationandsocialideals.Therefore,thehistorianconcernedwithsomeaspectofculturemustnotonlyappealtoinfluencesspreadingthroughaculturaltraditionwhenheattemptstotracechangesthathissubjectmatterhasundergone;hemustalsoviewthatsubjectmatterintermsofsocietalchange.Inphilosophy,religion,thearts,andalsointhesciences,thistaskisusuallyfarmoredifficult,andalsoInorecontroversial,thanattemptstotracethespreadofspecificculturalinfluenceswithinagivenfield.AsC.S.LewissaidindiscussinginfluencesonEnglishliteratureinthesixteenthcentury:"Thusfarwehavebeenconcernedwithideas,andideashaveaneffectonliteraturewhichcanbetraced,oftenwithgreatprobability,andsometimeswithcertainty.Whenweturntosocial,political,andeconomicconditions,we'lreinaverydifferentsituation.No�onedoubtsthatthesethingsaffectaInan'swritingatleastasmuchashisideasdo:buttheinfluenceisverymuchhardertoidentify."25Nonetheless,nohistoriancanneglectsuchinfluences,andinsomefields,suchastechnologicalinnovationsandtechnologicaldecline,theyoftenareofprimarysignificance.26Consequently,asweearlierhadoccasiontostress,specialhistoricalinquiriescannotatallpointsbewhollyautonon10USbutneedtodr�wupontheresultsofgeneralhistories.Wearenowinapositiontodrawasharpercontrastbetweenspecialandgeneralhistoriesthanwehavepreviouslydone,forthedifferencebetweenthemisnotonlyonebetweenahistorydealing ONWHATANDWHYINHISTORYwithasociety,orwithsomeaspectofitsstructure,andahistorydealingwithsomephaseofhumanculture;thereisalsoadifferenc�betweenthetypesofexplanationstobefoundinthem.Inspecialhistories,aswehavejustseen,theexplanationofthecharacteristicstobefoundinoneormoreworksisthroughrecoursetosomethinglyingoutsidetheseworksthemselves,forexample,throughanappealtocertainculturaltraditions,ortothetalentandtemperamentanduniqueexperienceofthepersonwhomadethem,ortotheimpactofsocietalchange.Ingeneralhistories,ontheotherhand,theexplanationofwhatitwasthathappenedisgiventhroughdeeperpenetrationintowhatdidactuallyhappen,justasOakeshottandothershavemaintained.Evenwhenanearthquakeoraninvasiondrasticallyaltersthenatureofasociety,itisbecausethesociety'seCOn0111Yhasbeenalteredorthesociety'spoliticalautonomyhasbeenoverthrownthattheeffecthascomeabout.Whenoneseekstoexplainhowitseconomywasalteredby,say,anearthquake,orhowitwasdeprivedofitspoliticalautonomythroughaninvasion,onegivesfurtherdescriptionsoftheseevents:Onetracesthedifferentwaysinwhichtheearthquakedisruptedtheeconomy,oronetracestheoutcomesofthebattlesthatfollowedtheinvasionandthepeacetreatythatbroughtthewartoaclose.Inallofthisonehasremainedwithintheframeworkoftheongoingprocessesthatmakeupthelifeofasociety.Ontheotherhand,theelementsthatmakeupsomestrandofculturalhistoryareindividualworks,oraspectsofthesework.Theyarenotcomponentswithinanongoingprocess;theyconstituteaseriesonlybecauseoftheirresemblancestooneanother,andthesereselnblancesareduepartlytothefactthattheyhaveinfluencedoneanother.Therefore,whensomeoutsideinfluenceaffectsthecharacteristicsofsuchaseriesofworks-·asachangeineconomicneedsaffectstechnology,orpoliticalchangeaffectsliterature,.ornewscientificconceptionsaffectphilosophy-·theexplanationofhowsuchachangebasbeenbroughtaboutisnotgivenbyprobingdeeperintothematerialsofthespecialhistorywithwhichoneisconcerned:Onemustexplainitbyrelatingittoth,ewaysinwhich,atthetime,theindividualswhoworkedwithintheseareaswerethemselvesaffectedbychangesaroundthemtowhichtheyresponded.Thus,itisonthebasisofspecializedbiographicalinvestigations,through,aknowledgeofotherformsofspecialhistory,andalsothroughaknowledgeofgeneralhistorythatthespecialhistorianisbestabletoofferconcreteexpla..nationsofthechangesintheculturalmaterialswithwhichhedeals.27 CAUSATIONFromwhathasbeensaiditfollowsthattheroleofexplanationinspecialhistoriesdoesnotconsistinfurther,closerdescriptionofwhathasinfactoccurredinanongoingprocess,butisoneofaccountingfornewinfluencesandforchangesofdirectioninthatseriesofworksthespecialhistorianhasunderconsideration.Becausetheexplanatoryaspectofspecialhistoriesdiffersfromtheirdescriptiveaspects,itisnotsurprisingtofindmanycasesinwhichahistorianhasadequatelyandsensitivelytracedthenatureofandchangesin,say,aseriesofworksofart,butinwhichonewouldbeforcedtorejecttheexplanationheoffersforthefeaturesthathefindstheypossessincommon,orforthetransformationsinthosefeaturesinthecourseofthemovementhetraced.Forexample,thefactthatonehistorianmayaccountforsimilaritiesandchangesinstylethroughappealingto"aspiritoftheage"shouldnotbetakenasjustifiedbytheadequacyofhisdescriptions,norshouldtheiradequacybeimpugnedbythosewho-quiteproperly,asIbelieve-wouldrejectthattypeofexplanationofthesefeatures.Yet,inthefieldofgeneralhistoriestheeventsthatexplainwhatoccurredarethemselvespartoftheseriesofoccurrences,andafalseexplanationissimplyonethathasmisrepresentedwhatdidoccur.Turningnowtohistoriesthatareprimarilyinterpretiveinstructure,ratherthanprimarilysequentialorexplanatory,itmaybesaidthatthehistorianaimstopresentadescriptionofwhathefindstobethedominantphysiognomiccharacteristicsofaperiod,andthesamewouldholdtruehadhechosentopresentaninterpretationnotofaperiodbutofsonleperson,orofsomeaspectofculturewithinagivenperiod,suchasanartisticstyle.However,inwhatfollows,Ishallbedirectlyconcernedonlywithinterpretiveaccountsthatdealwithasociety,ratherthanwithanindividualorwithsomeaspectofculturewithinaperiod,althoughwhatwillbesaidwillbesophrasedasalsotoapplytothem,n1utatis1nutandis.Thepossibilityofwritinggeneralinterpretivehistorypresupposes,ofcourse,averysubstantialstoreofpriorknowledgeconcerningthesocietywithwhichthehistorianistodeal,andthatstoreofknowledgecannotbeconfinedtoanyonestrandinthesociety'shistory,butmustincludeknowledgeofvariousfacetsofsocietalandculturallifeatthetime.Givensuchpriorknowledge,theroleofcolligation-thatis,ofbringingvariousaspectsoflifetogetherundersomeconceptorgeneraltheme-becolueshighlyimportant.Unlikespecialhistories,wherethehistorianworkswithsomepriorconceptofwhatconstitutes"literature"orHGothicarchitecture"or"technology"or"phi- ONWHATANDWHYINHISTORYlosophy,"aninterpretiveaccountofasociefypurportstoextractitsunifyingconceptDrthemefromanexaminationofthecharacteristicsthatweremostfundamental,andthereforemostpervasive,intheparticularsocietywithwhichitdeals.Therefore,aswehavepreviouslynotedwithrespecttointerpretivehistories,theadequacyofaninterpretiveaccountcaninfactbemeasuredagainsttheevidenceitmarshals,andagainstthepointsatwhichitsbasiccolligatorythemehasledthehistoriantooverlookotherevidence.Furthermore,itisusuallythecasethatburiedwithinaninterpretiveaccounttherearesomeexplanatoryassuInptionsthatthehistoriantakestobeindicativeofwhichelelnentsinthesocietyaccountedfortheoriginandspreadofthethemethatpervadeditsparts.Thisisnottosaythatitistheprimarypurposeofinterpretivehistoriestoofferexplanatoryaccountsoftherelationshipsamongthevariousaspectsofthesocietywithwhosenaturetheyareconcerned:Theirprimarytaskisdescriptive,portrayingthedistinctivefeaturescommontothevariousaspectsofasocietywithintheparticularperiodwithwhichtheydeal.Nevertheless,suchConl1TIOnfeaturesdonotarisesimultaneouslyinallaspectsofsocietalandculturallife,andthelocusofwhatistakentobetheirprimarysource,orsources,willrevealtheexplanatoryassumptionsonthebasisofwhichtheinterpretivehistorianhasinthiscaseseenfittoproceed.SuchassumptionsareopentocriticisIll:Theymustbesquaredwiththeresultsofexplanatoryinvestigationsthatothersconcernedwiththesamesocietyhaveproduced,ortheymustprovetobemoreconvincingbecausetheyincludeawealthofevidencethattheexplanatoryaccountsdidinfactoverlook.Thelatterisnotoftenlikelytobethecaseandtherefore,asweearliernoted,interpretiveaccountsarenotusuallygrantedprecedenceofauthorityovertheotherformsofgeneralhistory.Thereis,however,onewayinwhichaninterpretivehistorian'sthematicapproachnlayperInithinltofendoffsomeforlllsofcriticism.Thisdefenseliesinthenlannerinwhichhecharacterizestheperiodwithwhichhehaschosentodeal.Aswehaveseen,everyperiodizationofhistoryinvolvesachoiceastowhichaspectsofsocietalorculturallifearetobetakenasdefiningtheperiodunderconsideration;whatareregardedasmajorchangesintheseaspectsarethentakenassignalizingthebeginningandtheendoftheperiod.Sincedifferentaspectsofsocietalandculturallifedonotnecessarilychangeinasynchronousfashion,thethemesthatseempervasiveinsomeaspectsofsocietyorofculturallifeduringagivenspanofyearsmaynotbeepresentinotheraspectsduringthatsamespanoftime. CAUSATIONThus,itispossibletowriteverydifferentbutnonconflictinginterpretivehistoriesdealingwiththesamesocietyduringaparticularperiodoftime,solongasitismadecleartowhichareasoflifetheinterpretationisbelievedtoapply.Misunderstandingsonthisissueareapttoariseonlyifaninterpretivehistorianassumesthatthereissomesingleoverridingandall-embracingunityintheagewithwhichhedeals,butthatassumptionisnotplausiblewhenitisrecognizedthateveryperiodizationofhistorydependsuponthehistorian'sselectionofsomeaspectoflife,ratherthananyother,asbeingofprimaryconcerntohim.Theunityofaperiodmayinthisrespectbeverystriking,butoneshouldnotassumethatallotheraspectsofsocietalandculturallifewillshareinthatunity.28IVItrustthatitisnowclearthathistoricalstudiesarefarmorediversethanisusuallyassumed,andthatthisdiversityistosomeextentreflectedindifferencesbetweenthemodesofexplanationtobefoundinthem.Asaconsequence,oneneednotassumethatthemethodsemployedbyafirst-rateeconomichistorian,withhisabilitytouseabstractgeneralizationsdrawnfromthedisciplineofeconomics,haveanyexactcounterpartsinpoliticalhistory,northatthemethodsofgatheringandvalidatinginformationthatapoliticalhistorianmayusearelikelytohaveexactcounterpartsinliteraryhistory.Aswehaveseen,thereisevenadifferenceinthemannerinwhichtheconceptofcausationisappliedinthosecasesinwhichemphasisislaidonexplainingchangesinsocietallifeandinthoseothercasesinwhichoneattemptstoaccountforchangeinsomestrandofculture.Allthisisapttobeoverlookedbythosewhofailtoanatomizethetasksthatdifferenthistorianssetthemselvesinconductingtheirinvestigations,orwhoneglectthedifferencesamongthematerialswithwhichdifferenthistoriansdeal.Tochoosemerelyoneexample,itisnowgenerally-assumedthattheprimarysubjectmatterofhistoriansistobefoundin.humanactions,thoughthisassumptionwouldnotbeso-widely'heldwereitnotcoupledwith"methodologicalindividualism"-theviewthatwhenahistorianorsocialscientistrefersto"aninstitution"heisreferringinaquickandeasywayonlytosometypeofcustomaryorrule-governedbehavioronthepartofmostindividualswholivein ONWHATANDWHYINHISTORYcontactwithoneanotheratagiventimeandinagivenplace.29Thisview,asIhaveelsewhereargued,isfallacious;itspresentcurrencyrestsinlargepartontheassumptionthatitprovidedtheonlyalternativetotheviewsofthosenineteenth-centurysociologistsandphilosophersofhistorywhoheldthatsocietieschangeinaccordancewithanecessarylawofdevelopment,andthattheactionsofindividualsareeffectiveinhisto,ryonlywhentheyareinaccordwiththoselarger,impersonalforcesthatbringaboutsocialchange.3oThatassumptionisunwarranted.Asocietyneednotbemerelyanetworkofinterpersonalrelationsinorderforustobelievethatthechoicesofindividualsmayprofoundlyaffecthistoricalchange.Nevertheless,noonewouldholdthatallchoicesofallindividualsdoso,andthisbasicfactisoverlookedbythosewhoassumethattheprimarysubjectmatterofhistoriansistobefoundinhumanaction.Differentindividualsplaydifferentrolesinasociety,andtherolesthatanyoneindividualcanplaywilldependupontheinstitutionalstructureofthatsociety.Theextenttowhichthechoicesofanyindividualcarryoverintoaction,andtheextenttowhichtheythenhaveaneffectonthecourseofaffairs,willdependupontheinstitutionalsettingwithinwhichthatindividualactsanduponhisplaceinthatsetting.Untilthehistorianunderstandstheseinstitutionsandthepossibilitiesforactionthattheyallowatanyonetime,hewillnotbeinapositiontoexplainsocietalchange..Norisitpossibletounderstandmosttypesofculturalch�ngesolelyintermsofthechoicesofindividuals.ThearchitecturalorIiteraryhistorianwhotraceschangesinstyleandeventhehistorianoftechnologycanoftenappealtotheendsmenseekasformingsomepartofanexplanationofchangesthathavetakenplace.However,whenoneattemptstoexplainbothcontinuityandchange,suchchoicescannotbeisolatedfrom,norberegardedasmoreimportantthantheculturalandsocietalmatrixinwhichtheyaremade.Theintroductionofnewmodelsofcompositionthroughnewcontactswithothercultures,theavailabilityofnewbuildingmaterials,theimpactofnewscientificdiscoveries,orchangesinclassstructureareasimportanttotheriseofnewformsofexpressionorthecreationofradicallynewinventionsasisanythingthatisdirectlyattributabletotheendsmenseek.Thus,intherealmofspecialhistories,noless.thaningeneralhistory,thecausalexplanationofaseriesofeventsinescapablyinvolvesanappealtoamultiplicityoffactors,andthedistinctionbetweenwhatis"thecause"andwhatare"thecondi- CAUSATIONtions"breaksdown.Withitmustgotheattempttoholdthatexplanationsofsocietalorculturalchangemust"inthelastanalysis"restonanunderstandingofthemotivationsofmen.Onereasonwhythereisgreatreluctancetoabandontheviewthathistoricalexplanationsareultimatelytobecouchedintermsofthespecificchoicesandunderlyingpurposesofindividualmenisthatitseemseasiertoofferexplanationsofhistoricaleventsin·thesetermsthaninanyothers.Weknow,forexample,thattraitsofcharactersuchasprideandenvyarewidespread,weknowhowtheyareapttodominatethelivesandwarpthejudgmentofsomemen,andwerecognizetheconsequencesthatoftenfollowwhenmenareplacedincircumstancesofuninhibitedpower,ofdeprivation,orofpsychologicalstress.Thus,itwouldseemthatifhistoricalcausationwereessentiallyamatterofactionspurposefullyundertaken,itwouldposenoproblemswithwhichwearenotalreadyquitefamiliaronthelevelofcommonsense.Yet,historicalcausationdoesposeproblems,andtheexperiencedhistorianwillnotgivethesamesortsofexplanationofhistoricalchangeasaregivenbymostjournalisticinterpretersofcurrenteventsandasarecommonlyacceptedbythosewhohavehadnotraininginthecloseanalysisofthecircumstancesofsocialandculturallife.Themostfundamentaldifferencebetweentheexplanationsofcontinuityandchangethataregivenbyhistoriansandthosewithwhichothersmayremaincontentinvolvesthebreakdownofthesingle-factorviewofcausationwithwhichweoftenoperateonthelevelofcommonsense.Thisbreakdowninvolvesabandoninganyfirllldistinctionbetween"thecause"ofaneventandthoseaccompanying"conditions"withoutwhichtheeventwouldnothaveoccurredasitdidandwhenitdid.SolongasoneseekstodrawsuchadistinctionwhendealingwithanyofthefactsofsocietalorculturalchangeonewillnotbeabletoofferanadequateexplanatiQIlofwhathasoccurred.Therewillbenospecifictypeofeventth!atalwaysprecedesapoliticalrevolutionorachangeinarchitecturalstyle,astheHumeanlllodelofcausalexplanationwouldapparentlyhaveusassume.NorwilltheanalysisofHartandHonoreservebetter,sinceinthecomplexoccurrenceswithwhichhistoriansdealweshallnotbeabletoisolateanyfactoras':':thecause,"astheyusethatterm:Therewillnotbeanyonesinglefactorwithrespecttowhichthecaseinhanddiffersfromwhatusually,ornormally,occurs.NorisitsatisfactorytodrawadistinctionbetweencauseandconditionsinthewayinwhichOlafsonattemptedtodo.RelyingheavilyonHartandHonore,hesoughtto ONWHATANDWHYINHISTORYisolatethedecisiveelementinexplainingwhathasoccurredinhistory,andfoundthatelementalwaystobesomehumandecision.31However,aswehaveindicated,theeffectivenessofhumandecisionsinbringingaboutsocietalstabilityorchangedependsupontheinstitutionalrolesoftheindividualsmakingthesedecisionsandtheextenttowhichotherinstitutionalfactorsaffecttheimplementationofthesedecisions.Thus,historicalexplanationdemandsthetypeof,conceptionofthecausalrelationshipthatIhaveattemptedtoestablishasvalidinallcasesinwhichwewishtoofferaconcreteexplanationofaparticulareventorseriesofevents.Tohold,asIhaveheld,thatthecauseofaneffectistheactualseriesofeventsthatterminatedinthatspecificeffectisnotparadoxical.Itappearsparadoxicalonlyifonethinksof"anevent"assomequasi-instantaneoushappening,separatedintimeandspacefromthoseothereventstowhichitmaybesaidtobecausallyrelated.This,ofcourse,isanaturalenough,wayoflookingateventsifweidentifyaneventsolelywithrespecttothesortofeventitis,andarenotconcernedwithitasaconcreteoccurrence.However,whenlookedatasaspecificoccurrenceandnotasaneventofaparticulartype,anyeventisseenaspartofsomecontinuingprocess,ratherthanasanisolatedunit.Eveninthesciences,aswehaveseen,whenascientistisconcernedtoanalyzeaparticularcase,suchasthedeathofaperson,whatistobeexplainedisthatwhichhasbeenbroughtaboutinthecourseofaprocess.Inordertofillinthegapsinanysuchprocess,tracingitscontinuity,thescientistcannotremainonthelevelofexplanationwithwhichhestarted,butmustseektheconnectionswithintheprocessintermsofananalysisofthemicro-eventsonwhichitwasbased.Andthis,too,aswehaveseen,ischaracteristicoftheproceduresofhistorianswhoexplainconnectionsbylayingbarethefurthersubeventsthat,together,broughtaboutwhatoccurred.Yet,aswehaveseen,thistypeofcausalexplanationappliesonlywhenoneisdealingwithchangesoccurringwithinsomeongoingprocess,asisthecasewhen,ingeneralhistory,thehistorianisconcernedwithchangesinthenatureofasocietyorinsomeaspectofitssocietalstructure.Ontheotherhand,whenspecialhistoriansdealwiththeinfluencesresponsibleforchangesinsomestrandofhumanculture,theynlustlooktowhatmayhaveaffectedthosepersonswhoweretheinnovatorsofchange.Theseinfluencesmayhavecornefromcontactswithinnovationsinothersocieties,fromacross-fertilizationbetweendifferentculturalstrands,fromtheimpactofchangesinthestructureofthesocietyinwhichtheinnovationsarose,orfromallof CAUSATIONtheseand,inaddition,whateverfactorsintheinnovator'sownpersonalitythrowlightonhisinnovativetendenciesandonwhytheytooktheshapetheydid.Inallofthis,onecannotseparate"cause"from"mereconditions"andgiveasingle-factorexplanationofculturalchange..Everyculturalinnovationtakesplaceagainstabackgroundofwhatisalreadyfamiliar,andthecharacteristicsofthisbackgroundwillaffecteventhemostinnovativepersonalitiesinonewayoranother;itisthetaskofhistoriansofculture-inwhateverfieldtheymaywork-touncoverthisbackgroundsothattheymayunderstandbothcontinuingtraditionsandchange.Itisonlythelaymanwholacksthatbackgroundwhowillseeasingleinnovativepersonality,orviewasinglediscovery,asHthecause"ofmovementsinthearts,orinthesciencesortechnology,orinphilosophy.Historians,ontheotherhand,whentheydealwithproblemsoftraditionandchangetakeintoaccountnotonlydominantcreativepersonalitiesbutthewholecomplexpatternofinfluencesthatformculturalstands.32Thetemptationtoadoptasimplisticviewofhistoricalcausationis,ofcourse,strongwhenonesuddenchangeisseenfollowingimmediatelyuponanother,justastherapidsuccessionoftwoquasiinstantaneouseventsineverydaylifeleadsustothinkofoneasthesoleortruecausethatbroughtabouttheother.Yet,aswehavenotedinthecaseofflickingaswitchandhavingalightcomeon,experienceshowsthattherelationshipisbynomeanssimple:Awholesetofconditionsmustbepresentfortheflickingoftheswitchtoinitiatetheprocesswhoseterminusisthelight'scomingon.Inhistory,too,onemaybemisledbyapairingoftwochangesinrapidsuccession,butthehistorianwhohaslearnedtoanalyzesocietalchangeandisawareofthecomplexitiesinvolvedinculturalchangeswillnotbethusmisled.Butarethere,onemayask,anyfurtherparallelsbetweencausalattributionsinhistoryandineverydaylife?Tworen1aintobementioned.Inthefirstplace,aswehaveseen,inourdirectexperienceweoftendirectlysee-orbelievethatwesee-causalconnections,suchasthetransferenceofmotionintheMichotteexperiments,ortherelationshipbetweenthemovementsofaperson'shandwhenheiswritingandthechalklnarksontheblackboardthatseemtoflowfromthesemovements.Theclearestofsuchcases,asIhavesuggested,involvenotonlyspatialandtemporalcontinuitiesbutthequalitativesimilaritybetweencauseandeffectthatDunckerdesignated"correspondence.."Withinthematerialswithwhichbothgeneralhistori- ONWHATANDWHYINHISTORYansandculturalhistoriansdeal,someparallelstothesefactorsaretobefound;butinthesecases,asinthecaseofperception,thestrongimpressionoftheexistenceofacausalrelationshipthatoneimnlediatelyunderstandsmayprovetobemisleading.Tochoosemerelyasingletypeofexample,inthecaseofspecialhistoriesonemayhavethestrongimpressionofanuninterruptedstylisticcontinuityinasequenceofliteraryorartisticworks,butinvestigationmayprovethistobemerelyanartifactofselection:Theremayhavebeennopossibilityofinfluencebetweenoneworkandtheothers.Ontheotherhand,ofcourse,stylisticsimilaritiesoftendosuggestinfluence,andthespecialhistorianmaybeabletotracethefactthattherewereeitherdirectorindirectinfluencestoaccountforthesesimilaritiesinstyle.Casesofthissortclearlyparallelinstances,suchasthecollisionbetweentwobilliardballs,wheretheinitialperceptionofcausalrelationsisnotonlycapableofwithstandingobjections,butinwhichanalysisshowthattherewasacontinuousseriesoffiner-scaleeventsthattogetherconstitutedanongoingprocess.Similarly,inthecaseofgeneralhistory,thehistorianconcernedwithwhatappearstobeamountingcrisis,leadingtoarevolutionorawar,maynotbedeceivedbythesenseofmountingapprehensionandalarminthedocumentswithwhichhedeals,butwillbeabletotracetheinterplayofthosefactorsthatwereresponsibleforthealarlllandalsofortherevolutionorthewarthatactuallytookplace.Ontheotherhand,wedidnotewithrespecttooureverydaycausalbeliefsthatitisoftenthecasethatsimilarityinthepatterningofdiverseelements,suchasonefindsintheawarenessofagatheringstorm,sometimesleadstothebeliefthatwhileeachoftheseelementsisnotitselfacausalagency,itisanexpressionofsomesinglebasicunderlyingforce-thegatheringstormitself.Thisisafurtherpointatwhichitisusefultodrawananalogybetweencausalattributionsineverydayexperienceandcausalattributionsineithergeneralhistories.orhistoriesofculturalchange.Whenonefindsapatternofresemblingqualitiesinthevariousaspectsofcultureinanyperiod,itiseasytointerpretthemasexpressionsofthespiritoftheage.TheresemblancesmayalsocarlYoverintoatleastsomeaspectsofsocietalorganization,aswhenculturalrevolutionsinvariousarts,inscience,andinreligionseemtobeparalleledbyrevolutionarybreaksinpoliticaltraditionsordrasticrealignmentsinclassstructureinthatage.Itistheneasytoappealtosomebasicspiritualchangethatmanifestsitselfinalloftheseformsandisresponsibleforthem.Manygeneralhistoriesthatareprimarilyinterpretiveinstructure CAUSATIONhavedonepreciselythat.Yet,aswehaverepeatedlynoted,itisunlikelythatallaspectsofsocietallifeandallphasesofculturewillchangeinasynchronousfashion;oncethisisrecognized,thedegreeofunitytobefoundinanyagebecomesnotanexplanatoryprinciplebutsomethingthatisitselftobeexplained.Thus,theproblemshifts-asitmustalsoshiftinsimilarcasesineverydaylife-fromanattempttoexplainthepartsthroughanappealtothatwholetowhichtheyostensiblybelong;instead,itwillbeseenthattheexplanationofthewholewilldependuponunderstandingtheconnectionsthatexistinthepatterningofitsparts. PARTTHREEOBJECTIVITY�w.·�.* BIANKPAGE Cha.pterSixOBJECTIVITYANDITSLIMITS•....),.Thereshouldbythistimebenodoubtthatinthefieldofhistoricalstudiesthereisgreatvarietyinthematerialsstudiedandthatdifferenthistoriansoftensetthemselvesquitedifferenttaskswhendealingwiththesematerials.Therefore,whenoneraisesthequestionoftheextenttowhichhistoricalknowledgecanbeobjective,andwhatthelimitsofsuchobjectivitymaybe,onecannotexpectasingleanswerthatwillbeequallyapplicabletoalltypesofhistoricalinquiry.Instead,weshallhavetoconsidertheanswersthatarelikelytobemostadequateinthevarioussortsofinquirywithwhichhistorianshavebeenchieflyconcerned.First,however,itwillbeusefultoconsidersomeofthedifferentwaysinwhichtheconceptofobjectivityhasbeenused. OBJECTIVITYIWhenthequestionoftheobjectivityofhistoricalknowledgeisraised,theissueisoneconcerningtheaccuracyorreIiab:ilityofthatknowledge;butnotallusesoftheconceptofobj:ectivityareequallyconcernedwiththisproblem,whichhastodowiththetruthofwhatis-actuallyaffirmedordeniedin�hejudgmentswemake.Instead,theconceptisoftenusedinwaysthatdonotreferdirectlytothecontentofaparticularjudgment,butcallattentioninsteadtotheconditionsunderwhichthatjudgmentwasmadeandbywhichitmayhavebeeninfluenced.Itisinthissensethatwesaythatapersonhasbeenobjectiveifhehastriednottoletself-interestorfearorangerinfluencehisjudgment.Similarly,apersonmaybesaidtobeobjectiveifheisnotprejudicedfororagainstspecificindividualsbecauseoftheirclass,theirnationality,theirreligion,ortheirrace.Aperson'sjudgmentsmayalsoberegardedasobjectiveifhedoesnotexaggeratethevirtuesofthosetowhomheisattached,norexaggeratethefailingsofthosewhonlayhaveinjuredhim.Inallsuchcases,"objectivity"hastodowithkeepingpersonalconsiderations,sentiments,andemotionsfromwarpingone'sjudgment,whatevermaybetheobjectoneisjudging.Whiletheseforlnsofobjectivityoftenhaveabearingonthetruthorfalsityofaperson'sbeliefs,itmayturnoutthatapersonhasjudgedtrulyevenwhenhehasnot,inthissense,beenobjective;andhemayhavejudgedfalselyeventhoughhehas.Therefore,whentheconceptofobjectivityisusedinthissense,itshouldnotbetiedtoocloselytothequestionofthetruthorfalsityofthatwhichisbelieved.Similarly,thecriterionofobjectivitymaybeappliedinassessingmoraljudgments:Wechallengethevalidityofamoraljudgmentwhenitseemstospringfromself-interest,bias,orspecialemotionaltiestothosewhoarejudged.Thisattempttopurgemoraljudgmentsofsubjectivityischaracteristicofwhathasbeencalled"themoralpointofview,"anditisarecognizedelementinallmoraltheories,whethertheyareclassifiedascognitiveorasnoncognitive.Since,accordingtononcognitivists,truthandfalsityarenotapplicabletomoraljudgments,itisonceagainclearthattheconceptsofobjectivityandsubjectivityaresometimesappliedwithreferencetotheconditionsunderwhichajudgmentismadeandarenotnecessarilytiedtothequestionofwhetherwhatisassertedistakenastrueorasfalse.1Furthernlore,weInaynotethattheconceptofobjectivityisoftenusedwithreferencetothewayinwhichapersonconductshimselfinaparticularsituation,'suppressingprejudices,sentiments,orpersonal OBJECTIVITYANDITSLIMITS147inclinations,andinthesecases,too,wearenotprimarilyconcernedwiththetruthorfalsityofhisbeliefs.Forexample,wemayholdthatajudgedidordidnotdisplayobjectivityinthewayinwhichheconductedhimselfincourt,orinhischargetothejury;similarly,ateachermayormaynotbeobjectiveinhisassignmentofgrades.InsuchcasesHobjectivity"isclearlynotusedwithreferencetowhatonemayjustifiablyclailllthatoneknows.Thesefactshavenot,however,beenrecognizedinmostdiscussionsoftheproblemofhistoricalknowledge.Manycontroversieshavearisen,andhavebeenneedlesslyprolonged,becauseithasbeenassumedthatthebasicissueisoneconcerning"objectivity,"takeninthissense.Thus,ithasbeensupposedthatthetruthclaimsofanyhistoricalaccountmustbeassessedintermsoftheextenttowhichthehistorian'sworkhasbeeninsulatedfromallpersonalconsiderations,sentiments,oremotionsconnectedwiththeeventsheattemptstounderstandanddepict.Thishasbeenunfortunate,since"objectivity"-wheninterpretedinthissense-doesnotprovideanytestofwhetherornotastatementorsetofstatementsistrueorfalse,eitherinhistoryorelsewhere.Thereisafurther,quitedifferentwayinwhichtheconceptofobjectivityisfrequentlyused,butit,too,lacksanynecessaryconnection"viththequestionofthereliabilityofourknowledge.Whatisreferredtoasobjective,inthissecondsense,doesnothavetodowithwhetherourbeliefsarefreefromtheintluenceofourlikesanddislikes,ofself-interest,orofouremotions;instead,itinvolvesacontrastbetweenwhatisattributedtotheknowerandwhatexistswhetherornotitisknown.Takingthedistinctionbetweenthatwhichis'�objective"andthatwhichis"subjective"inthissense,philosophershaveoftenincludedwithinthesubjectiveallthatfallswithintherealmofhumanexperience,asdistinctfromwhatever-ifanything-existsindependentlyofbeingexperienced.Notinfrequentlythishasledthemtoidentifythesubjectivewith"themental."Theseparticularusesofthesubjective-objectivedistinctionaremoreoftenfoundinphilosophicdiscussionsthaninothercontexts,butitisalsothecasethatineverydaylifewecommonlydistinguishbetweenwhatissubjectiveinthesenseofbeing"ours,"andwhatisobjectiveinthesensethatitisindependentofus.Forexample,thetickleinournostrilsbeforewesneeze,thesorenessofourmusclesafterunaccustomedexercise,thethrobbingofaninjuredfinger,theshootingpainofaheadacheareallexperiencedassubjective,asbelongingpeculiarlytous.Inthesecaseswelocalizewhatisexperiencedasbeing OBJECTIVITYwithinourownbodies.These,then,are"subjective"states,andwedonotregardthemashavingexistenceindependentlyofourexperiencingthem.Similarly,weregardourthoughts,dreams,andmemoryimagesassubjective,ratherthanasexistingindependentlyofus.Withrespecttothelatterinstances,wearenotevenawareofanybodilyconditionsuponwhichtheypresumablydepend;thusweareapttotakethemastheclearestexamplesofthatwhichissubjective,andonlysubjective.Ontheotherhand,whatweseeortouchwetaketobeindependentofus.Inphilosophy,however,therehasbeenalong-standingtraditionholdingthatallthatisavailabletousasafoundationforknowledgeisthedataofconsciousness,takenassubjectivestates.Mostphenomenalistsandmanyidealistsbelongwithinthistradition.Thosewhorejectphenomenalismandrejectsubjectiveformsofargulnentinfavorofidealismtaketheopposedview,whichisalsoadoptedbymostlaymenwithrespecttothefoundationsofhumanknowledge:Theyassignprioritytowhatweimmediatelyidentifywiththeobjectivepoleofourexperience.Fortunately,thisepistemologicalquestionneednotbedebatedhere.Whatisimportanttonoteisthatdiscussionsofhistoricalknowledgehavesometimesbecomeentangledbyit.Thisisevident,forexample,inCroce.Hisinsistencethatallhistoryiscontemporaryhistoryrestedonhisviewthatartifactsanddocuments,consideredasobjectivefacts,arewithoutsignificanceuntilthehistorianwhoexaminesthembringsthemtolifethroughhisimaginativere-creationofthem.Thus,forhim,noneofthedatawithwhichhistoriansdealareultimatelyindependentofthesubjectsbywhomtheyareknown.Asimilaremphasisonsubjectivity,inthissense,istobefoundinCollingwood'streatmentofwhatconstituteshistoricalfacts.Unlikefactsconcerningnature,itisthe"innerside"ofevents,nottheirouter,objectiveformsofexpression,thatisofprimaryimportancetothehistorian,who-accordingtoCollingwood-mustgraspthisinnercorethroughthinkingthethoughtsthatwereresponsibleforwhatoccurred.Ontheotherhand,manywhohavediscussedtheproblemofhistoricalknowledgewouldnotacceptanysuchemphasisonthesubjective,insofarasthebasicmaterialsforhistoricalconstructionareconcerned.Forexample,althoughCharlesA.BeardfrequentlycitedCroceasasourceandanallyofhistheory,Beardregardedtheartifacts,documents,andmanyatomicfactsconcerningthepastasobjectivelygiven.Itwasonlywhenhediscussedhowthehistorianmakesuseofsuchfactsinconstructingahistoricalaccountthat OBJECTIVITYANDITSLIMITSBeardapproachesaCroceanposition;itwashisclaimthatthewayinwhichahistoriansynthesizesthesefactsisareflectionofhisexperienceandisnottobeconstruedasareflectionofindependent,objectiverelationshipsamongthefactsthemselves.Thus,evenatheorythatassumesthattheultimatedataforanyhistoricalaccountareobjectivelygivenlnaynonethelessclaimthatanyhistorian'swayofrelatingthesefactstooneanotherwillbe"subjective,"ratherthandependingonthenatureofthefactsthemselves.Itmayseemasifthequestionofwhatis"subjective"andwhatis"objective,"takeninthesenseofwhatiscontributedbythesubjectthroughhisownexperience,andwhatisindependentofthatexperience,wouldbethebasicissueconcerningtheobjectivityofhistoricalknowledge.This,however,isnotthecase.Ineveryfieldofknowledgethebackgroundandexperienceoftheinvestigatorwillaffecthisinvestigations.Forexalnple,itisonlybecauseofhisbackgroundandexperiencethatascientistdiscernstheproblemswithwhichheistodeal;furthermore,whateversolutionsheproposeswillhavebeensuggestedtohilnthroughwhathealreadyknows,orbelievesthatheknows.Inspiteofthis,thescientist'sexperience,interpretedas·asubjectivefactconcerninghim,isnevertakenasadequatetestimonyforthetruthorfalsityofhisinterpretationoftheeventsandrelationshipsheclaimstohaveestablished.Inshort,wheneverweclaimknowledge·ofanythingotherthanofourownimn1ediateexperiencethereisanappealtothatwhichisregardedasbeingindependentofthatexperience:.Thisistruenotonlywithrespecttoourknowledgeofnature,asSOlnehavebeeninclinedtohold;itistruealsoofourknowledgeofothers,andofhowtheyexperiencetheworld.Inclaimingthatweknowthebeliefsorintentionsoridealsofotherpersons,weareclain1ingtoknowsOlnethingotherthanwhatweourselvesareexperiencing.Thus,whenevertheobjectivityofsomeformofknowledgeisbeingdiscussed,thecontrastbetweenthe"subjective"andthe"objective"-inthepresentsenseoftheseterms-isirrelevant.EventhoughsubjectivefactsconcerningtheexperienceofanyinvestigatormayhelptoexplainhowSOlneofhisjudgmentscaInetobemade,theywillfailtosettleanyquestionsconcerningtheaccuracyand.reliabilityofwhatthesejudgmentsaffirlnordeny.Inadditiontothetwosensesoftheconceptofobjectivitywithwhichwehavesofarbeenconcerned,thereisathirdwayinwhichitisoftenused.Itisthissensethatisdirectlyrelevanttoourproblem.Ajudgmentcanbesaidtobeobjectivenotmerelybecauseitwasnotduetoself-interest,prejudice,orthelike,andnotmerelybecause OBJECTIVITYitreferstoeventsandrelationshipsthatexistedindependentlyoftheexperienceofthepersonjudging,butbecauseweregarditstruthasexcludingthepossibilitythatitsdenialcanalsobetrue.Theobjectivityofajudgment,takeninthissense,constitutesabasicprinciplethatispresupposedwheneverweseektoestablishthereliabilityofourjudgmentsconcerningmattersoffact,includingfactsconcerningthoseformsofdirectexperiencethatareinterpretedasbeingsubjectiveratherthanobjective.Thisprincipleisobviouslyrelatedtothesocalledlawsofthought,andwhenappliedtotheproblemofknowledgeitmaybestatedquitesimplyasfollows:Ourknowledgeisobjectiveif,andonlyif,itisthecasethatwhentwopersonsmakecontradictorystatementsconcerningthesamesubjectmatter,atleastoneofthemmustbemistaken.Itthenbecomesnecessarytosay,inanyparticularareaofdiscourse,howoneistoestablishwhichofthecontradictorystatementsismistaken,ortoadducereasonsforholdingthatbotharetoberejected.Aswehaveseen,thiscannotbedecidedonthebasisoftheattitudes,emotions,predispositions,orprejudicesofthepersonorpersonsresponsibleforoneoranotherofthecontradictoryjudgments.Norcanadecisionbereachedbycitingthoseelementsinagivenhistorian'sbackgroundandexperiencethatinfluencedwhateverjudgmentshemayhavemade,sinceallhistorians,aswellasallscientists,judges,orotherpersons,areinfluencedbytheirbackgroundsandexperience.Vhatevertestmustbeusedmustinallcasesbedirectlyappliedtowhatisbeingaffirmedordenied,nottowhateverrealorsupposedinfluencesmayhaveledtothataffirmationordenial.Vhenthisisrecognized,atleastsomeoftheconventionalargumentsforhistoricalrelativism,andagainsttheobjectivityofhistoricalknowledge,losemuchoftheirforce.InwhatfollowsIshallattempttoshowthatwhenoneclearsawaytheprec"edingmisunderstandings,theinterlockingconnectionsamongthedatawithwhichhistoriansareconcernedpermitustoholdthatthecumulativeresultsachievedthroughtheirindividualinquiriescan,inmostcases,beregardedasestablishingknowledgethatisobjectiveinthisthirdsenseofthatterm.Tobesure,thereareinnumerableindividualcasesinwhichthiscontentionappearsopentochallenge,butinmanysuchcases,asweshallsee,conflictsarisebecausethereferentsofthetwosetsofjudgmentshavenotbeenspelledoutwithsufficientcare.Insuchcases,thetwoconflictingjudgmentsmaynotinfactbecontradictory,andassoonasthedefecthasbeenremediedbothjudgmentscanbeacceptedwithoutviolatingtheprincipleofobjectivity.Nevertheless,asweshallalsosee,there 9BJECTIVITYANDITSLIMITSarecasesinwhichtheopposingjudgmentsdocontradicteachother,andadecisionconcerningthetruthorfalsityofoneortheothercannotbereachedunlessonecanappealtosomewell-authenticatedgeneraltheorythatlendsitssupporttooneratherthantotheother.IIInwhatfollows,Ishallnotinitiallybeconcernedwithanydifferencesthatmayexistbetweengeneralandspecialhistoriesinsofarastheproblemofobjectivityisconcerned;instead,Ishallfirstexaminequestionsthatariseinany'formofhistoricalinquiry,whetheritsdominantstructuralformissequential,explanatory,orinterpretive,andregardlessofthesubjectmatterwithwhichitdeals.Themostobviousofthesegeneralquestions,andtheonewithwhichitissimplesttodeal,isthequestionofhow,ifatall,thefactthatdifferenthistoricalaccountsdealwitheventsofdifferentdimensions,ratherthanexaminealleventsonthesamescale,isrelatedtotheissueofobjectivity.Asecond,analogousquestionarisesbecausedifferenthistoricalaccountsoftendealwithdifferentfacetsofthesameevents,andthislnayseemtoraisequestionsconcerningtheirobjectivity.Ouranswerstothesequestionswillhelplaythegroundvorkforaconsiderationofthemoredifficultquestionswithwhich,insubsequentsections,weshallbeconcerned.Totakeupfirsttheissuesarisingwithrespecttodifferencesinscale,itwillberecalledthatIusedthatnotiontorefernotonlytodifferencesinthetimespancoveredindifferenthistoricalaccounts,butalsotodifferencesinhowrestrictedorhowextensivethesubjectunderinvestigationmayhavebeen.Thus;InotonlycontrastedthescaleofahistoryoftheUnitedStateswithahistoryconfinedtotheeventsintheCivillVarperiod,butIalsocontrastedahistoryoftheUnitedStateswithahistoryofoneofitsstatesoroneofitsmunicipalities.Atfirstglance,theremayseemtobeafundamentaldifferencebetweentheproblemsthatarisewhenoneusestheconceptofscaleinthesetwoways,oneofwhichistemporalwhiletheotheris,sotospeak,geographical.ItwouldseemthateveryeventofhistoricalinterestfallingwithintheCivilWarperiodisalsoaneventbelongingtothehistoryoftheUnitedStates,althoughnoteveryeventthathappenedinGeorgia,orinAtlanta,canequallywellbesaidtobelongwithinUnitedStateshistory-howeverimportantitmayhavebeeninthepoliticallifeofGeorgia,orfortheeconomicdevelopment OBJECTIVITYofAtlanta.Nevertheless,noultimatedistinctionofthiskindcanbedrawn.AsIhaveindicatedinfirstintroducingtheconceptofdifferencesinscale,onedoesnotexpectmapsdrawnondifferentscalestoconveythesameinformation:Iftheydid,therewouldbenoreasonforashiftinscale.Andthisholdsnolesswhenwearedealingwithshiftsinthetimespancoveredinahistoricalaccountthanitdoeswhenweshiftattentionfromanyothermoreextensivesubjecttosomesubjectthatis,inasense,includedwithinit.Considerthelattersortofcasefirst.Thatwhichapoliticalhistoriandealingwithaparticularstatewishestotracewillbechangesinitspoliticallife,notinthepoliticallifeoftheUnitedStates.Changesinthelattermaydrasticallyaffectchangesinthelifeofanyorallstategovernments,aswhenvariouspowersbecomeconcentratedinthefederalgovernment;andtherealsoaretimesatwhichwhatoccursinthepoliticallifeofaparticularstatemaygreatlyinfluencewhatoccursonthelevelofthefederalgovernment,as:isthecasewhenanissuearisinginoneparti�ularstatehasanimpactontheissuesortheoutcomeofanationalelection.Nevertheless,ahistoricalstudythathasasitssubjectmatterthepoliticallifeofastateisnotastudyofthehistoryoftheUnitedStates,andastudyofchangesinthepoliticsandeconolni.cconditionsinonelllunicipalitywithinastateisnottobeconfusedwithahistoryofthatstate.This,Ishouldsuppose,wouldbeentirelyclear.Itisequallytrue,however,thatahistoryofeachofthestatesdoesnotserveasahistoryoftheUnitedStates,andaseriesoflocalhistoriesdoesnotconstituteanystate'shistory.Thefactthathistoriesofthesetypesinterlockatmanypointsandthatstudiesoftheonemustoftenrelyonstudiesoftheothershouldnotleadustoconfusethem:Ineachcasethespecificsubjectmatterisdifferent,andthisholdstruewhetherthehistoriesareprimarilysequential,explanatory,orinterpretive.Preciselythesamesituationholdswithrespecttohistoriesthatdifferintimescale.AhistoriandealingwiththeCivilWarperiodmusttracethepolitical,military,andeconomiceventsoccurringwithinthatperiod;hemustunderstandandfollowtheirconsequencesduringtheCivilWaritself.Naturally,someoftheseeventswillhavehadfurther,long-runconsequences;otherswillnot,andinthatcaseonlytheirimmediateimpactonthecourse.ofthewarwillbeofimportance.When,-however,aspecificeventsuchasLincoln'sEmancipationProclamationhasrelevancetowhathappenedinsubsequentperiodsofAmericanhistory,thateventwillfindaplaceinhistoriesthatuseadifferenttimescale.Theseotherhistories,which OBJECTIVITYANDITSLIMITSarenothistoriesoftheCivilWaritself,neednotgobackandtracethevariouscausesoftheeventwhosesubsequentinfluencetheyaretracing,norneedtheydescribeitsoriginalcontextindetail,bothofwhichaCivilWarhistorianwouldbeexpectedtodo:Itwillbesufficientforthemtoconsiderit,anditsconsequences,intheframeworkofthelongertimespanwithwhichtheyareconcerned.Puttingthen1attermoregenerally,thefactswithwhichhistoriansareconcernedwhentheyworkondifferentscalesarenot"thesamefacts,"eventhoughtheyrelatetothesameactualoccurrences.Thereisnothingoddaboutthis.Take,forexample,almostanyimportantepisodeinaperson'slife.Onemayviewsuchanepisodeineitheroftwoways:Onemaydescribeitandanalyzeit,treatingitasaparticularlymemorable,self-containedepisode,oronecanviewthatsameepisodeinalargercontext,asaturningpointinthatperson'slife.Whenoneviewssuchanepisodeinthesedifferentways,whichfeaturesappearasmostsignificantmaybequitedifferent,sincethesameepisodeisbeingviewedindifferentcontexts.Relativistsareapttoseizeonthisfactasestablishingthecontentionthatanyhistoricalaccountisdominatedbythehistorian'sowninterests,whichleadhimtoviewaneventinonecontextratherthaninanother.Theexistenceoftheinfluenceofone'sinterestsonthecontextinwhichonehappens,orchooses,toviewanoccurrenceisindisputable.Whatmustnotbeoverlooked,however,isthefactthatthesedifferentapproachesarenotintheleastcontradictory,sincethetruthofeachiscompatiblewiththetruthoftheother.Tobesure,ifanyhistorianweretoassumethathisaccountcouldcaptureeverythingthatoccurredwithrespecttohissubject-ifheweretoassumethathiswrittenworkcouldreplicateinalldetailtheactualoccurrenceitself,makingahistoricalworkequivalenttowhatBeardtermed"history-asactuality"-thentheexistenceofmultiplehistoriesdealingwiththesameoccurrenceswouldentailtheirbeingcontradictory.Yet,Iknowofnohistorianwhocanbesaidtohavebeenguiltyofsuchafoolhardyassumption.Itwouldinvolveconfusingawrittenworkwiththoseeventstowhichtheworkrefers.Tobesure,somehistorianshavebeenmisledbysomephilosophers,andhaveassumedthatwhenadocumentreferstoafactconcerninganoccurrenceitcanbetakentobetrue,andnotaviciousabstraction,onlyifitrefersatthesametimetoallaspectsofthatoccurrence.This,however,issimplytoconfusewhatisafactconcerninganoccurrencewiththatoccurrenceitself.Whileitistrulyafact-basedontestimonythereisnoreasontodoubt-thatCharlesA.BearddiedonSeptember2,1948,one 154OBJECTIVITYneedknowneitherthecausesnorthecircumstancesofhisdeathforthattobeafact,eventhoughitisthecasethatwhenanydeathoccurstherearecausesthatbroughtitabout,andbroughtitaboutatoneplaceratherthananother,andwithorwithoutothersinattendance.Whileitissometimesimportanttoinvestigatesuchcausesandcircumstances,thefactthatdeathdidoccuronthatdatewillnotbealtered,whatevernlaybetheoutcomeofthesefurtherinvestigations.Asimilarbutsomewhatmorecomplexsituationobtainswhenweturnourattentionfromdifferencesinthescaleoftheeventswithwhichdifferenthistoriansdealtodifferencesinthefacetsoftheeventsthatmayoccupytheirattention.Insomecasesitisfairlvobviousthattwodifferenthistoricalac-oJcountsthatdealwithdifferentfacetsofthesameoccurrencesaremorelikelytosupplementoneanotherthantoclash.Forexample,amilitaryhistorian'saccountofawarandapoliticalhistorian'saccountofthewarperiodcanbeexpectedtomesh,sincepoliticaldecisionsarefrequentlylinkedwithmilitarysuccessesorfailures,andmilitarysuccessesorfailuresmaydependuponpoliticaldecisions.Tobesure,thesehistorians-likeothers-maydisagreewithrespecttosomepointsthateachdiscusses,butthiswillnotbebecausetheyaredealingwithdifferentfacetsofthesameoccurrences;norshouldweassumethatthejudgmentsofonearemorereliableandmoreinconformitywithalloftheevidencesimplybecauseheisconcernedwithoneofthesefacetsratherthanwiththeother.Inthiscase,asinothers,theevidencecited,alongwithotherevidencethatmightbecitedorthatmightbediscovered,isthebasisonwhichthecontroversyistoberesolved.Whatisimportanttonoteisnottheexistenceofanysuchdisagreementsbutthefactthatneitherthepoliticalnorthemilitaryhistoriancanadequatelycultivatehisownspecialtywithoutrelyingondataandinterpretationswithwhichtheotherisprimarilyconcerned.Whatholdsinthisobviouscaseholdsalso,ingreaterorlesserdegree,inothercases-even,aswehaveseen,withrespecttothewaysinwhichaspecialhistoryof,say,DutchpaintingrelatestothepoliticalandsocialhistoryoftheNetherlandsduringthatperiod.Ifonethinksofthedifferentfacetsofasocietyandtheelementsenteringintoitscultureasdifferentperspectivalviewsofoneverycon1plexobject,theadvantageofmultiplyingtheperspectivesfromwhichoneviewsthatobjectbecomesobvious.Thisisanalogoustothefactthatifoneistolearnthetrueshapeofamountainonemustbeabletoseeitfrommanyangles,sincenoone OBJECTIVITYANDITSLIMITSperspectivalviewwill,byitself,revealitsdimensionsanditscontours.Soitiswithmultifacetedmaterialswithwhichgeneralhistorians,atleast,arerequiredtodeal.Troubledevelops,ofcourse,whenahistoriancontendsthatoneperspectivalviewismoreimportantthananyoftheothersbecausehebelievesthatfacetofsocietallifetobemorebasic,ormorerevealing,thananyother.Suchclaimsmaybemadewithrespecttospecificsocieties:thatinthissociety,inthisperiod,certaineventseconomicorpoliticalorreligious-hadapreponderatingroleinshapingthesocietyandthechangesitunderwent.Itisinfactalmostinescapableforthoseconcernedwithgeneralhistorytomakeassumptionsofthissortastowhichofthefacetsofaparticularsocietywereofgreatestsignificanceinensuringitsstabilityorinbringingaboutchangeinit.Itneednot,however,beassumedthattherewasonlyonesuchfacet,ratherthanseveral,thatplayedacrucialroleintheserespects.Whiledifferenthistorianswilloftendisagreeonsuchquestions,thereexiststhepossibilityofgoingfartowardtheresolutionoftheirdisputesthroughtheexaminationofevidenceastohoweventsthatwereprimarilyofonetypeinfluencedeventsofothertypes,andtheextenttowhichthelatterwereinfluencedbytheformer.2Whilesuchdisputesmaynotbereadilyresolvedtothesatisfactionofthedisputants,thefactthatunresolveddifferencesofopinionaretobefoundinthisareashouldnotleadonetoacceptarelativisticconclusion,anymore,say,thanunresolveddisputesbetweenmedicalpractitionersconcerningthecausesofanillnessanddisputesastohowbesttotreatitshouldleadonetoholdthatinmedicinetherearenocriteriaonthebasisofwhichsuchdisputescaninprinciplebesettled.Insuchcaseswedonotholdthateachpractitioner,givenhisbackground,interests,andpreconceptions,isequallyentitledtohisownviewastotherelativeimportanceofthefactorsthatareadmittedlypresentinthesituationconcerningwhichtheydisagree;instead,we.assunlethattherearewaysinwhichfurtherknowledgecanleadtoaresolutionoftheissue.WhileadequateinHlanysimplercases,thisanswermaybreakdownwhentheissuedoesnotconcernwhichofseveralfacetsisofprimaryimportanceinaspecificsocietyataparticulartime,butinvolvesatheorythatinallsocieties,atalltimes,certainfacetsarebasicwhileallothersaredependentuponthem.Theoriesofthistypearepredicateduponaviewofthenatureofasocietyanduponbeliefsconcerningthefactorsresponsibleforsocietalandculturalchange.Inthehistoryofsociologicaltheorytherearemanyexamplesofsuch OBJECTIVITYtheories,andmanyphilosophersofhistorywhowouldnotconsiderthemselvessociologistsalsorepresentthetendencytogeneralizeinthisfashion.Ineithercaseitisimportanttoexaminehowgeneraltheoriesinfluenceattemptstounderstandhistoricaleventsandwhatrelevancetheyhavefortheproblemofobjectivity.IIIItmaysafelybesaidthatthereisnouniversallyagreeduponmeaningoftheterm"theory"suchthatallphilosophersofsciencewilldistinguishinthesamewaybetweenwhattheyregardasalawandwhattheyregardasatheory.InsofarasmyownuseoftheterlTI"law"isconcerned,thereaderwillrecallthatIheldthatalawformulatesaninvariantfunctionalrelationshipbetweenfactorspresentinavarietyofconcreteoccurrences(chapter5,section2).Noone,Itakeit,wouldregardsuchacharacterizationasacorrectformulationofwhatwedesignateasatheory.Norisatheorysimplyageneralizationfromasetofexperimentallaws,sincetheacceptanceofatheorymayprecedethediscoveryofthoseobservationsandlawsthatserveaspartialconfirmatoryevidenceforit.Furthermore,itistobenotedthatdirectobservationsandexperimentalfindingslTIayinmanycasesbeexplainedthroughappealingtodifferenttheories;therefore,notheoryisrigidlyentailedbysomeparticularsetofdata.Speakinggenerally,thoughnontechnically,atheory(asIshallusethatterm)isawidelyapplicablehypothesisthatservesasanexplanatoryframeworkthroughwhichavarietyofobservationsandideally-avarietyoflawscanbeconnectedwithoneanother.Theunifyingfunctionofatheorydependsuponthetheorist'sabilitytoshowthatthebasicconceptsandassumptionsofthattheorycanbeappliedtoawidevarietyofphenomenaandcanusefullyservetoconnectadiversesetofapparentlyindependentlaws.Inthephysicalsciences,theparticulatetheoryofmatterisoneexampleofatheory,takeninthissense,foritisapplicabletoahostofobservationsandservestoconnectawidevarietyofexperimentallawsthatarenotknowntobededuciblefromanyonemoregenerallaw.Inthebiologicalsciences,ontheotherhand,therewerenoalreadywell-establishedlawsthatDarwin'stheorywascalledupontoconnect,butthereweremanyapparentlyindependentphenomenathathewasabletobringtogetherinasingleexplanatorysystem.Heaccomplishedthisbyfirstassumingthatinanynewgenerationsome OBJECTIVITYANDITSLIMITS157individualswillpossesscharacteristicsnotpossessedbytheirancestors,and·byassumingthatthereisaselectiveprocessoperativeinnatureaswellasunderdomestication.Onthebasisoftheseassun1ptions,andbyusingtheconceptsofa"struggleforexistence"and"thesurvivalofthefittest,"histheoryservedtoconnectmanyobservedfactsconcerningthedistributionofplantsandofanimals;itwasalsoabletoaccountfortheexistenceofsimilaritiesandofdifferencesbetweenapparentlydifferentspecies;italsoofferedaco�sistentinterpretationofthefossilrecord.Similarly,historiansandsocialscientistshaverepeatedlysoughttoestablishconnectionsamongbroadrangesofhistoricalfactsbymeansoftheoriesconcerningsocietalorganizationandthefactorsresponsibleforsocietalchange.3Marxisn1isoneexampleofsuchatheory;sotooweretheevolutionarydoctrinesofComteandofSpencer,variousformsoffunctionalisttheoryinanthropology,andthephilosophiesofhistoryofVico,Spengler,andToynbee.Infacttherearesomanysuchtheories,lllOStofwhicharemutuallyinconsistent,andsofewofwhichhaveseriouslysoughtconfirmationthroughanexaminationofasufficientlybroadsurveyofavailabledata,thathistoriansarelikelytoholdthatnogeneraltheoriesofsocietalorganizationandchangehaveaproperplaceinhistoricalinquiries.This,however,isamisguidedclaim.Historicalinquiriesdonoteverproceedwithoutatleastanimplicitacceptancebythehistorianofoneoranothersetoftheoreticalcommitments-asWernerSombartremarked,"Notheory,nohistory."4Amongsuchcommitmentswillbethosethatcharacterizethehistorian'sviewofthenatureofsocietiesandofthefactorsaffectingsocialstabilityandchange.Forexample,somehistoriansenvisionsocialinstitutionsasrefle·ctionsoftheconcreteaimsandidealgoalsofthosewhoshareintheongoinglifeofasociety,andtheythereforeseekexplanationsofstabilityandchangeinthevaluesandchoicesofindividualpersons.Othersusuallytermed"holists"-denythattheaimsofindividualsdeterminetheinstitutionsunderwhichtheylive;instead,theyregardinstitutionsasdevelopingandchangingtomeettheneedsofthesocietyinwhichtheyareembedded,regardlessofthegoalsthatindividualsmaywishtoattain.Suchdifferencesinthebasicconcep:..tionsofasociety(towhich,ofcourse,therearealternatives)willdeeplyaffectthesortsofinquiriesthatdifferenthistoriansarelikelytoundertake,andalsothewaysinwhichtheyexplainwhathastakenplaceinthepast.So,too,willdifferencesintheirviewsregardingthedegreetowhichvariousaspectsoflifeinasocietyformasingle OBJECTIVITYintegratedwhole.Somehistoriansexpectaveryhighdegreeoffreeplayintheinstitutionalandculturallifeofanysociety,whereasothersexpectthatchangesinsomeoneinstitution,orinsomeaspectofculture,willinvariablybecorrelatedwithchangesinallothers.Somehistorianswillnot,ofcourse,holdeitheroftheseextremeviews.Theywillworkwithinadifferenttheoreticalframework,believingthatdifferentsocietiesdisplaydifferentdegreesofunityintheirorganizationandthatnoonesocietyneedatalltimesdisplaythesamedegreeofunity.Notalloftherelevanttheoriesonemightciteareasgeneralasthese.Ihaveusedthemasillustrationstoshowthateventhoughhistoriansusuallylookwithgravesuspicionontheutilityofthemorespecifictheoriesthathavebeenformulatedbysocialscientiststoexplainsocietalorganizationandchange,theycannotthemselvesescapetakingastandwithrespecttosomegeneraltheoreticalissuesconcerningthenatureofsocietiesandthefactorsinvolvedinsocietalchange.Tothisextentatleast,thereisjustificationforSombart'sdictum"Notheory,nohistory."Thequestionnowarisesastowhetherthepresenceofsuchtheoriesnecessarilylimitsthedegreeofobjectivitytobefoundinanyhistoricalwork.Theanswerwilldependupontheextenttowhichthetheoryinformingahistoricalworkisitselfcapableofbeingtestedforitstruth.Ifthegeneraltheoryoneacceptedweretodependuponone'sattitudetowardtheworld,astheeconomistJ.C.R.Dowclaimedinreviewingabookconcerned,viththeworkofJ.M.Keynes,5itwouldnotbeplausibleton1aintainthatobjectivityisattainable.Ontheotherhand,evenifageneraltheorycannotbeconfirmedthroughshowingthataparticularsetoffactsisdeduciblefromitandnotfromanyofitsrivals,itmaystillbepossibletoshowthatthetotalweightoftheevidencefavorsonegeneraltheoryratherthananother;andthisisallthatonecanexpectbywayofconfirmationforageneraltheory(asdistinctfromanexperimentallaw),eveninthenaturalsciences.Inmostformsofgeneralhistory-towhichIshallfortheInomentcontinuetoconfinemyself-itisindeedpossibletomarshalconvincingevidenceinfavorof,oragainst,somegeneraltheories.Whenthetheoriesareasbroadasthoseconcerningtherolesofindividualsandofinstitutionsinfosteringstabilityandchange,acomparisonofdifferentsituationsinawidevarietyofdifferentsocietieswouldsurelyshowthatneitherextremeviewcanberenderedplausiblebytheevidence.Inthiscase,asinthequestionoftheunityofasociety,practicinghistorianswouldbeinclinedtotakethemiddleroad.Here, OBJECTIVITYANDITSLIMITStoo,theyareapttothinkthattheyarenotespousinganytheoryatall,simplybecausetheirtheorydoesnotproposethatwhateveristrueofonesocietyatonetimemustalsobetrueofallsocietiesatalltimes.Whatitdoespropose-andwhatpermitsonetoconsideritasatheory-istheviewthatinallsocietiesatalltimesonemustbepreparedtotakeintoaccountbothinstitutionalfactorsandtheactionsofindividualsifoneistounderstandstabilityorchange.Silllilarly,theviewthatthenatureandchangesofoneinstitutionoftenhaverepercussionsonotherinstitutionsisatheorythatoffershistoriansguidanceintheirattemptstoexplainmanyoftheeventstheyarecalledupontoexplain.Ifahistoriandoubtsthatthisconstitutesatheory-holdingthatitismerelycomillonsense-heneedonlyconsidertheextenttowhichearliergeneralhistoriesofnationswerewrittenprimarily-andsometimesalmostexclusively-intermsofpoliticallife.Implicitinthosehistorieswasatheoryofsocietalorganization,whetheritwasarticulatedornot,accordingtowhichtheinstitutionsofgovernmentwerethelllOStbasicfeatureinthesociety,andsocietalchangecouldbeunderstoodalmostwithoutreferencetochangesinanyotherinstitutions,exceptfortheroleplayedbyreligiousdifferencesinpoliticallife.Thefactthatthispoliticallyorientedtheoryhasnowbrokendowndoesnotmeanthatanolderformofhistoryhasmerelybeensupplantedbyanewanddifferentform.Therehasbeenanadvance.Politicalhistoryisstillwritten,andisverywellwritten,butsinceitisnowrecognizedtobemerelyoneformofhistoricalwriting,dealingwithonlyonefacetofsocietallife,historiansaremorealerttothepossibleeffectsofotherinstitutionsongovernment,andtheexplanationsofpoliticalchangethattheyareinapositiontoofferhaveprovedtobericheranddeeperthanthosepreviouslyofferedwhen"society"wasequatedwith"thestate."Forthoseacquaintedwiththerelevantdatafromanthropology,orforthoseinterestedintheMiddleAges,itcanscarcelybedoubtedthatitisamistaketoassumethatatalltimesandinallplacespoliticaleventsshouldbeallowedtooccupycenterstagewheneverahistorianseekstoexplaineitherstabilityorchange.Thus,theoriesthatwereonceverymuchtakenforgrantedaregivenupwhenhistoriansextendtherangeoftheirintereststoincludeevidencepreviouslyunknownorneglected.Tobesure,sometheoriesmaybedogmaticallyheldandbeveryresistanttochange,especiallyiftheyareconnectedwithmetaphysical.commitmentsorwithbasicpolitical-ethicalconvictions.ItisthenthattheyreflectwhatJ.C.R.Dowlabeled"rivalattitudestowardthe OBJECTIVITYworld."Yeteventhemostbasicattitudestowardtheworldaresometimesdrasticallyalteredbecauseofanaccumulationofnewdatathattheyareunabletoabsorb.Unfortunately,manygeneraltheoriesofsocietyandsocialchangehavedefendedthemselvesagainstthispossibilitybydefiningasrelevantonlysuchdataasconformtotheirinitialinterpretivescheme.Amongphilosophersofhistory,forexample,ithascommonlybeenthecasethattheonlyfactsorperiodsorevengeographicareasconsideredtoberelevanttothetruthofthetheoryarethosethatillustratewhatevergranddesignthephilosopherofhistorytakestobeparamountinimportance;allelseisdismissedasinsignificant"inthelongrun."ThistendencyisclearlyevidentinHegel,butitisnolesstrueofComteandofSpencer,whoheldthedisciplineofhistory,aspracticedbyhistorians,incontempt.6Marxisttheoryhasoften(andsometimeswithjustice)beencondemnedforthesamefault,althoughsomeformulationsofthattheoryinvolveanattempttobringitintolinewithawiderrangeoffactsthanMarxhimselfwasabletotakeintoaccount.Eveninthecaseofaninterpretivehistory,wherethethemeofthestudydictateswhatmaterialsaretobeincludedandatwhatpointsthemainemphasisistofall,itispossible(asIhavealreadysuggestedinchapter2,section3)toassessitsadequacybycheckingitsinterpretationagainstfurtherrangesoffact.rThus,whileIacknowledgethatthereoftenarelimitstoobjectivityintheory-dominatedworks,itisnottruethattheimpactofageneraltheory,orpointofview,makesitimpossible,inprinciple,tojustifyortorefuteaninterpretationofthepast.When,however,weturnfromanyformofgeneralhistorytospecialhistories,andconsiderthequestionofobjectivitywithrespecttothem,theproblembecomesmorecomplicated.Notonlywillthespecialhistorian,likethegeneralhistorian,beforcedtomakecertaintheoreticalassumptionsconcerningthefactorsthatinducesocietalandculturalchange,buthisworkwillalsopresupposesomecharacterizationoftheparticularsubjectmatterwithwhichhisspecialhistoryistobeconcerned.Forexample,anarthistorian'sworkpresupposesatleastanimplicittheoryofwhatseparatesartfromnonart.Similarly,underlyinganyhistoryofphilosophytherewillbeatleastatacitdefinitionofphilosophythatservestojustifylabelingsomepersonsasphilosophers,butwithholdingthatdesignationfromotherswhomaytosomeextentsharetheirintellectual,moral,andreligiousconcerns.Wehavealreadynotedthatthesamesituationobtainswithrespecttotheproblemofwhatconstitutes"literature." OBJECTIVITYANDITSLIMITSSometimesthedefinitionofaconceptsuchas"literature"derivesfromageneraltheoryofthearts;in.othercasesitrepresentsaprogramoratraditiontowhichtheparticularliteraryhistorian(knowinglyornot)tendstoadhere.Inneithercasecansuchdisputesbereadilyresolved,eveninprinciple.Inthisfield,therefore,thereissomereasontodoubtthatobjectivitycanbeattained.Thispointisworthyoffurthercomment.Insofarasthedefinitionofwhat,forexample,constitutes"literature"restsonageneraltheoryofthearts,themajordifficultyliesinthefactthattheparticularinstancesusuallyusedasevidencefororagainstsuchatheoryarenotneutralfactswhoserelevanceisbeyonddispute.Whenthereisadisputethatisfundamental,involvingtwodifferentgeneraltheoriesofthearts,eachdisputantwillbeinclinedtorejectthecounterexamplesthathisopponentwillbemostinclinedtocite,preciselybecausehisdefinitionofhissubjectmatterdiffersfromtheonehisopponentaccepts.Thus,iftheevidenceuponwhichonedrawsinsupportofageneraltheoryoftheartsisconfinedtotheparticularinstancesthatthetheoryattemptstointerpret,the.quarrelbetweenrivaltheoriescannotberesolved:Eachwillintheendbearguingcircularlyforhisowntheory.7Whatholdswithrespecttothedifficultyofestablishingobjectivityinthefieldofliteraryhistory,becauseofitsdependenceonageneraltheory,alsoholds,mutatismutandis,inallotherformsofspecialhistory.Thus,wehavehereapproachedthepossiblelimitsofobjectivityinoneformofhistoricalInqUIry.Inthosecasesinwhichdifferentdefinitionsofliteraturearenotprimarilydependentupontheacceptanceofoneoranothergeneraltheory,butrestuponfamiliaritywithaninterestindifferenttraditions,oruponcommitmentstodifferentprogramsastowhatisworthyofencouragementandwhatisnot,thelimitsofobjectivityarealsoreached.Insuchcases,solongaseachoftheopposedhistoriansmaintainshisposition,eachwillbewritingahistorythatisbiasedbythenatureoftheworkswithwhichheisalreadybestacquaintedandofthosethatheprefers.Whenonetakesintoaccountthefactthatliteraryformsandmodesofexpressionhaveundergoneradicalchangesovertime,onecanappreciatehowpervasivesuchbiasesamongliteraryhistoriansarelikelytobe:Aliteraryhistorywrittenatonetime,inthelightofthethenknownpast,willalmostcertainlydifferinorientationfromonewrittenwithfullawarenessofthechangesingenreandinstylethatsubsequentlyoccurred.Whatisin OBJECTIVITYthisrespecttrueofliteraryhistoriansisnolesstrueofthoseconcernedwithotherformsofspecialhistory,suchasthehistoryofphilosophy,ofscience,orofart.ItmayseemthatthefactorsIhavementionedaslimitingobjectivityinspecialhistorieswouldnecessarilylimittheobjectivityofanygeneralhistoryaswell.Suchanargumentmightbeconstructedalongthefollowinglines.JustasaIiteraryhistorian'sdefinitionofthespecialsubjectmatterwithwhichheisconcerneddeterminesthecontentofhiswork,sothegeneralhistorianselectscertaintypesofeventwithwhichhechoosestodealandconstructshisaccounttoincludeeventsofthistype,whileexcludingothers.Andjustastheliteraryhistorianislimitedbythetraditioninwhichhestands,orbyprogrammaticaimsastowhatisandwhatisnotofimportanceinliterature,sothegeneralhistoriantendstobeconfinedwithinoneoranotherhistoricaltradition,ortobeinfluencedbyoneoranotherprogrammaticaim.Finally,itmightbeargued,thefieldofgeneralhistoryhasundergonemanyradicalchangesinitsstyleandinitsconcerns,andmuchofitnowdiffersprofoundlyfromhistoricalworksthatwereproducedinGreeceorRome,intheRenaissance,orevenintheEnlightenlnent;toexpectthesedifferentwaysofwritinghistorytoyieldcompatibleresultsissimplyquixotic.However,itismistakentodrawthisparallelbetweengeneralhistoriesandspecialhistories.ItwillberecalledthatIemphasizedthefactthatgeneralhistorieshaveastheirsubjectmatterparticularsocieties,whicharecontinuingentitiesexistinginaregionoveraparticularperiodoftime,whereasthesubjectmatterofanyspecialhistoryisaclassofresemblingculturalproducts,manyofwhicharerelatedthroughstrandsofinfluencebutwhichdonotcompriseaunitaryongoingentity,asdoesasociety(cf.chapter1,section3).Asacansequenceofthisdifference,ageneralhistoriancannotexercisethesamefreedomindelimitingwhatwillandwillnotenterintohisaccount.Hemusttakeasocietyasitis,analyzingitintermsofconnectionsthatexistamongitscomponentparts.Thus,eventhoughageneralhistorianmaydefinehistaskasoneinwhichhewilldealdirectlywithonlyonefacetofasocietyandnotwiththesocietyasawhole,heisnottherebysetfreeofconstraintsastowhatmustbeincludedwithinhisaccount:Hewillhavetorecognizewhateverotherfactorsmayhavedirectlyaffectedthechangeswithwhichheseekstodeal.When,ontheotherhand,aspecialhistoriandefineshissubjectmatter,hedoesnotnecessarilyplacehimselfunderthesametypesof OBJECTIVITYANDITSLIMITSconstraints.Hemay,forexample,decidetodealwithacertainbodyofworksproducedbysomewriters,butnotothers,settinghimselfthetaskoftracingthesimilaritiesanddifferencesintheirinterestsandstyles,andcomparingthemwithrespecttotheesteeminwhichtheywereheld.Hemayormaynotchoosetoaccountforwhateversimilaritiesordifferenceshefindsamongthem,orintheirpopularity;onecannotsaythatgiventhetaskhehassethimself,heisobligatedtodoso.Thus,thecriteriaweusewhenestimatingtheworkofaspecialhistorianlllayshiftaccordingtothesortoftaskhehassethimself;wecandemandthathedowellwhateverhehassethimselftodo,butwecannotcriticizehimfornothavingdonesomethingelse.Inthecaseofageneralhistorian,however,thesituationisdifferent.Hepurportstounderstandanddepictwhatwasinfacttrueofsomesociety,ortrueofsomeaspectofit;thescaleonwhichhehaschosentoworkwillofitselfdeterminewhatheshouldinclude-aswellaswhatheneednotinclude-inhisaccount.If,forexample,hisaccountpurportstodealwithsomesegmentofthepoliticalhistoryofanation,andifhedidnottakeintoaccounttheimpactofsomereligiousoreconomicchangesonthechangesthatoccurredinthepoliticallifeoftheperiod,wedonotsaythatheneednothavedoneso;weholdthathisaccountstandsinneedofcorrection,eventhoughwemaystilladmirehisabilityonothergrounds.Inshort,eventhoughwerecognizethatdifferentgeneralhistoriesreflectdifferentassumptions,wedemandareconciliationoftheirdifferences,ratherthanacceptingboth.Ontheotherhand,inthefieldofspecialhistories,wedemandareconciliationonlyofdifferencesamongaccountsthatproceedonthebasisofthesameassumptions,oronthebasisofassumptionsthatarecompatiblewitheachother.Unfortunately,itisoftenthecasethatdifferenthistoriansofliterature,ordifferentculturalhistoriansgenerally,proceedonthebasisofincompatibleassumptions,andwhenthisoccursalimitofobjectivityhasbeenreached:Incomparingtwosuchaccounts,wecannotsaythatatleastoneofthewaysinwhichthepastwasdepictedmustberejected.Itwillberecalledthatitispreciselyinthissense-andnotinanyother-thattheconceptofobjectivityisbeingusedinthisdiscussion.Thecontrastjustdrawnbetweentheobjectivitytobeexpectedingeneralhistoriesandwhatmaybesaidoftheunresolveddifferencesbetweendifferentspecialhistorieshasabearingonrecentdiscussionsoftheroleofcolligationinhistory.Theterm"colligation"wasapparentlyfirstusedinaspecificallyphilosophiccontextbyWilliam OBJECTIVITYWhewellinhisdiscussionsofmethodinthenaturalsciences;itwasW.H.Walshwhofirstintroduceditintodiscussionsofhistoriography.8Whewellhadcharacterizedcolligationinsaying,"FactsareboundtogetherbytheaidofsuitableConceptions.ThispartoftheformationofourknowledgeIhavecalledtheColligationofFacts:andwemayapplythistermtoeverycaseinwhich,byanactoftheintellect,.weestablishapreciseconnexionamongthephenomenawhicharepresentedtooursenses."9Inhisuseoftheterm,Whewellwasreferringtothestepinscientificmethodthatheregardedasinterveningbetweentheestablishmentoffactsandourinferencestohypotheses:Colligationinvolvedbringingappropriateconceptstobearonwhathadbeenobserved.Thus,itwasasteppreliminarytotheformationofthehypothesesthatweretoexplainwhathadbeenobserved.VValsh,ontheotherhand,doesnotusetheconceptofcolligationtoapplytoasteppreliminarytoexplanation;heviewsitasthewayinwhichhistorianstransformtheirdatainto"significantnarratives"thatdonotstandinneedoffurtherexplanation.Hesays:"Differenthistoricaleventscanberegardedasgoingtogethertoconstituteasinglehistoricalprocess,awholeinwhichtheyareallpartsandinwhichtheybelongtogetherinaspeciallyintimateway.Andthefirstaimofthehistorian,whenheisaskedtoexplainsomeeventorother,istoseeitaspartofsuchaprocess,tolocateitinitscontextbymentioningothereventswithwhichitisboundUp."10AccordingtoWalsh,itisthroughtheintroductionof"dominantconceptsandleadingideas"thatthehistorianmovesfromamerely"plainnarrative"ofwhathappenedtoa"significantnarrative"inwhichweareabletoseewhyithappened,thatis,toseeitasapartwithinalargerintelligiblewhole.11Whetherthispositionleadstoasubjectivisticortoanobjectivisticpositiondepends,ofcourse,onthefactors,thatdeterminethenatureoftheseintelligiblewholes.OnWalsh'sview,theconceptsbymeansofwhichthehistoriancolligateshisfactsareconceptssuchas"theIndustrialRevolution"or"theEnlightenment,"whichheclaimsare"arbitraryandnotnaturalunits."12Thus,accordingtoWalsh,itisthehistorian'sownchoiceofconcepts,ratherthanthedatawithwhichheworks,thatunderliesthekindsofexplanationsheoffersoftheeventswithwhichhedeals.Thisinitselfissufficienttoundermineanyclaimtotheobjectivityofhistoricalknowledge.13Onthebasisofwhathasalreadybeensaidconcerningtherolethatdefinitionsplayindelimitingthematerialswithwhichhistoriansof"art"or"literature"or"philosophy"deal,therewouldbemuchtobe OBJECTIVITYANDITSLIMITS165saidforWalsh'senlphasisontheimportanceofcolligation-andthelackofobjectivityitintroduces-wereheconcernedonlywithwhatoccurswithintherealmofspecialhistories.This,however,isnothisconcern:Heisattemptingtodealwithallformsofhistoricalexplanation.Evenwerewetogrant-asheassumes14-thatinhistorygenerallywearealwaysconcernedwithhumanactionsandthathumanactionsarealwaystobeinterpretedteleologically,historiansarenotfreetousewhateverconceptstheychooseinordertoarrangethesedataintosignificantwholes.AsIhaverepeatedlyattemptedtoshow,onceahistorianhaschosenasubjectmatterandaworkingscale,itisinthefirstinstancethedatathatinquiryreveals,ratherthanthehistorian'sowninitialconcepts,thatservetocontrolthestructuredconnectionsofthefactswithinhisaccount.Whereverhisselectionoffactsappearstobedependentuponhisinitialconceptionofwhatthewholemustbelike,ratherthanbeingconfirmedthroughinquiry,hisreconstructionofthepastmaybeconsideredinterestingasrevealinghisownmindandasmirroringhisowntimes;butthatdoesnotmeanthatitwillbeacceptedbythosewhoareprimarilyinterestednotinhim,butintheeventshisworksoughttodepictandexplain.Asimilarcontrastcanbedrawnbetweentheobjectivitytobeexpectedingeneralhistoriesandthelimitationsonobjectivityinspecialhistoriesifoneconsidersthewaysinwhichperiodizationaffectseachformofhistoricalinquiry.Inbothcasesthehistorianperiodizesthepastintermsofeventsheregardsasmarkingthebeginningandtheendofaparticularlysignificantdevelopment.Thegeneralhistorianmaymarkoffperiodsintermsofthereignofaruleroradynasty,orintermsofwhatheregardsasthebeginningandtheendofsomesignificanteconomicdevelopmentwithinasocietyorgroupofsocieties,orhemaydosointermsoftheriseanddeclineofanationoranallianceofnations.Similarly,inthefieldofspecialhistoriesperiodsaremarkedoffintermsof,say,adominantstyleinliteratureorastylethatisheldtocharacterize.variousformsofartatthetime;oraperiodmaybemarkedoffintermsoftheacceptanceandsubsequentrejectionofasetofpresuppositionsinphilosophy,orinthebackground,development,andfinalacceptanceofaseriesofepoch-makingscientificdiscoveries.Forexample,onethinksofRomanticismoroftheBaroque,ofRationalism,ofseventeenth-andeighteenth-centuryEmpiricism,oroftheScientificRevolutionastermssometimesusedtocharacterizeperiodsinliterature,inthearts,inphilosophy,andinthesciences.Wheregeneralhistoriesandspe- OBJECTIVITYcialhistoriesdifferwithrespecttoperiodizationisnotintheirdelimitationofaperiodbymeansofsignificanteventsinthefieldwithwhichtheyareconcerned;thisiscommontoboth.Wheretheydifferisintheimpactoftheirperiodizationonwhattheiraccountswillincludeandexclude.Inageneralhistorytheperiodizationmayrestonpoliticalevents,oneconomicchanges,oronaviewofchangesintherelationsamongnationsoveraparticularspanoftime,butwhateverdictatesthechoiceoftheseeventsasmarkingoffaperiodwillnotjustifythehistorianinexcludingothertypesofeventsthatbroughtaboutchangesinsocietyduringthattime.Ontheotherhand,whenahistorianwritingaspecialhistoryhascharacterizedaperiodintermsofthedevelopmentofastyleorintermsofacommonsetofphilosophicpresuppositionsorintermsofthedevelopmentandacceptanceofanewsetofscientificconcepts,methods,andparadigms,whatheisobligedtoincludewithinhissurveyofthatperiodareonlythoseotherworksthatshareacommondenominatorwiththeworksthathavegivenrisetohisperiodization.However,theworksproducedwithinaparticularperiod,evenwithinasuitablydefinedgeographicalarea,arenotapttobecharacterizedbyanysimplehomogeneityinconceptionorexecution,evenwhenoneconfinesone'sattentiontothosethatareclearlycomparableinintentandinfunction.Disparitiesintasteandtheinfluenceofregionalandclassinterestsareevidentwheneveronelookscloselyatthetotalrangeoftheseworks,evenwhenthereisoneoverridingstylethatgenerallydominatestheperiod.15Thisisnotonlytruewithrespecttoliteratureandthearts,butalsoclearlyappliestophilosophyaswell.Forexample,itisunmistakablethattherewasacontinuingscholastictraditionbothinEnglandandontheContinentthroughouttheperiodinwhichDescartes,Spinoza,andLeibniz,aswellasBoyleandLocke,providedtheinnovationsthatmarkthenewRationalismandthebeginningsofmodernEmpiricism.Thedifferencebetweengeneralhistoriesandspecialhistorieswithrespecttoproblemsofperiodizationissymptomaticofthebasicdifferencebetweenthem:Generalhistorieshaveastheirsubjectmatterentitieshavingacontinuousexistence,andspecialhistoriesdonot.Therefore,aswehavealreadynotedwithrespecttohistoriesofliterature,andasnowalsoappearswithrespecttoperiodizationsinspecialhistories,unityisintroducedbyaprincipleofexclusionthatpermitsahistoriantoconsideronlycertainworks,andnotothers,onthebasisofhisevaluationoftheirimportancewhentheyareconsideredasrepresentativeofthetypeofculturalactivitywithwhichheiscon- OBJECTIVITYANDITSLIMITScerned.Nomatterhowintimatetheconnectionsamongtheseworksmaybe,itisnottheirrelationshipsthataredecisiveinsinglingthemoutforattention;ifitwere,thehistorianwouldalsohavetotracealloftheirotherinfluences,andnotbecontentwithhowtheyinfluencedworksthat,onhisview,werealsoimportant.Sincedifferenthistorianswilladoptdifferentviews,basedondifferenttheoriesorevaluativecriteria,onecannotexpectaresolutionofthedifferencesbetweenalternativespecialhistories,eachofwhichmaybeexcellentsolongasoneadoptsitspointofview,buteachofwhichwillproveunsatisfactoryifonedoesnot.Thus,inthisfield,onecannotexpectobjectivityinhistoricalknowledge.Onefinalcontrastmaybedrawnbetweengeneralhistoriesandspecialhistoriesinsofarastheproblemofobjectivityisconcerned.WhiledifferentgeneralhistoriesmaydealwithdifferentfacetsofasocietyandneednotInakeuseofidenticalscales,anyonesuchhistorywilldovetailwithothers,andacumulative,consistentrecordofpastsocietiescanbebuiltup.Similarly,differentbiographiesofthesameperson-thoughstartingwithvaryinginterests,andstressingdifferentaspectsofthatperson'scharacterandcareer-will,whentakentogether,yieldamoretrustworthyinterpretationthanwillanysinglebiographythatseekstointerprethisachievementsandfailuressolelywithreferencetooneofmanyalternativepointsofview.Whatholdsofbiographyalsoholdsofinterpretationsoftheliterary,philosophic,orscientificworkofanyspecificpersonwithwhomaspecialhistorianofliterature,philosophy,orsciencemaybeconcerned.Whilemanyinterpretationsofclassicfiguresinthesefieldsarepossible-andsomeprovetobeextremelystimulatingevenwhentheyareobviouslyone-sided-intheendanattemptmustbemadetoachieveaninterpretationofthatperson'sworkintermsofitsauthor,itscontemporarycontext,andtheculturaltraditionstowhichitbelongs.Whilepreferencesandpersonalbackgroundwilltendtodictatewhatanyoneinterpreterwillseeandwillstress,itisnecessarywhendealingwiththelifeandworkofaparticularindividualtotakediversepointsofviewintoaccountandtoofferanonidiosyncraticinterpretationofthepersonwithwhomoneisconcerned.Noristhisimpossible,evenwhenoneisdealingwithwriters,painters,scientists,orphilosophers.Therefore,whatprecludesobjectivityinsomeformsofculturalstudiesisnotthecharacterofthematerialswithwhichtheydeal,howevervalue-ladenthesematerialsmaybe.Rather,itisthefactthatifahistorianofcultureisnotdealingwiththelifeandworkofoneperson,orwithsomelimitedgroupofpersons,butis OBJECTIVITYseekingtotraceacontinuoushistoryofsomeformofculturallife,theprinciplebymeansofwhichhechoosesandorganizesthespecificmaterialswithwhichheistodealrepresentsmerelyoneamongmanypossibleprinciples.Inthefieldofgeneralhistory,thesamefactorofrelativelyfreeselectionisnottobefound.Itistoestablishthispoint,andtoconcludethisdiscussionofobjectivity,thatwenextturn. ChapterSevenOBJECTIVITY,CAUSATION,ANDLAWSTheproblemofobjectivityinhistoricalknowledgeturnsonthequestionofwhatcontrolstheworkofhistoriansoncetheyhavesetthemselvessomespecifictopictoinvestigate.Anynumberofdifferentreasonsmaybecitedtoexplainwhyparticularhistorianschosetoinvestigatethetopicstheydid.Suchreasonsaretobesoughtinbiographicaldataconcerningtheirbackgrounds,training,andpredilections,aswellasinthetopicsthatwereofinterestinthemilieustowhichtheybelonged.Furthermore,boththenatureoftheirchoicesandthewaysinwhichtheycarriedthemoutareboundtohavebeendependent,tosomeextent,onthehistoricalknowledgeavailableatthetime.Questionsconcerningthesefactors,bothinparticularcasesandingeneral,areofimportancetohistoriansofhistoriography,buttheyarenotdirectlyrelevanttothequestionofobjectivitywithwhich OBJECTIVITYwearehereconcerned.Ourproblemiswhetheritisthecasethatwhentwohistoriansdisagreeintheirinterpretationsorexplanationsoftheverysameevents,atleastonemustbeheldtobemistaken.Thefactthatdifferenthistoriansaremotivatedindifferentwaystoundertakewhateverinvestigationstheypursuewouldberelevanttothequestionofobjectivityonlyiftheconnectionstheytraceamongtheeventswithwhichtheyareconcernedwereprimarilyartifactsofwhatoriginallydeterminedtheirchoiceoftheirsubjectmatter.Disagreementswouldthenbeexpected,andwouldbeineradicableinprinciple:Anyhistorianseekingtojustifyoneaccountagainstotherscouldbeaccusedofbeingpredisposedtofavorthataccountbecauseofhisowninterestsandbackground.Ontheotherhand,ifhistoriansareconstrainedbythenatureofthematerialswithwhichtheymustwork,andiftheconnectionstheytracedidinfactexistintheeventstowhichthesematerialsrefer,thenitistobeexpectedthatdifferenthistoricalaccountscanbeusedtosupplementandcorrectoneanother;abeliefintheobjectivityofhistoricalknowledgecouldthereforebemaintained.Norwouldthisconclusionbeunderminedbythefactthateveryhistoricalaccountispartiallyconditionedbywhatmayhavebeenthestateofhistoricalknowledgeinitsowntime.Thatnewknowledgedevelopsandnewmodesoftreatingthepastarisedoesnotofitselfprovethatthereisnocompatibilitybetweentheoldandthenew.Onceagainitremainspossiblethatthesedifferingaccountscouldbeusedtosupplementortocorrectoneanother,andabeliefintheobjectivityofhistoricalknowledgecouldbemaintained.Thatthisistruewithrespecttogeneralhistories,whathasalreadybeensaidandwhatfollowsshouldservetomakeclear.Itwillberecalledthatgeneralhistoriesareconcernedwiththenatureandchangesofparticularsocieties,orwiththenatureandchangesofspecificaspectsoftheirstructure.Thedominantapproachinsomegeneralhistories,aswehaveseen,issequential;othersseektoaccountforaparticularstateofaffairsbyexaminingtheeventsonwhichitdepended;stillothersseektointerpretandportraythenatureofsocietythroughexaminingitsvariousfacetsandnotingtheirinterplay.Whiletheserepresentthreebasicallydifferentformsoforganizationinhistoricalaccounts,everyaccountwillatonepointoranotherhavetoincludeelementsthattypifytheotherapproaches.Wehavealreadyseenthatsomedegreeofobjectivitycanbeascribedtointerpretivehistories,basedontheextenttowhichthebasicinterpretivethemeemployedseemstoilluminateawiderangeofmaterialsinadditiontothosetheinterpretivehistoriancitedinfavorofhis OBJECTIVITY,CAUSATION,ANDLAWSinterpretation.1Furthermore,thatdiscussionshouldhavemadeitclearthattheadequacyofinterpretivehistorieswasdependentupontheaccuracyofthesequentialandexplanatoryaccountsonwhichtheyweredependent.Ishallthereforenowturntoexaminethedegreeofobjectivitythatcanbecreditedtosuchaccounts,firstconsideringthosethataresequentialratherthanexplanatoryintheirbasic,overallstructure.Anysequentialaccounthasasitsfocusofinterestsomechainofevents;thehistorianseekstodepictthenatureofthesuccessivelinksinthatchainandtheirconnectionswithoneanother.Thisisastrueofsequentialaccountsthatfallwithintherealmofspecialhistoriesasitisofsequentialgeneralhistories;however,itisonlywiththelatterthatIamnowconcerned.Inageneralhistorytheseriesofeventswithwhichthehistoriandealsareeventsthatoccurredwithrespecttosomeentitythathadacontinuingexistence,andthehistorian(aswehavenotedinChapter6)mustalreadypossesssomeknowledgeofthatentityandofsomeofitschangesinordertohavebecomeinterested-forwhateverreason-ininvestigatingthecourseofitshistory.Whatleadsthehistorian,givensuchaninterestandsometentativeandpreliminaryknowledge,toincludecertaineventswithinhisaccountandtoexcludeothers?Ashasjustbeenremarked,theanswertothisquestioniscrucialtoanydiscussionoftheproblemofobjectivity.Inthefirstplace,whenthehistorianselectsthesubjectwithwhichhisinquiryistobeconcerned,hewillalreadybeinpossessionofvariousdocumentsandreportsreferringtowhatoccurredwithrespecttothatsubjectatvarioustimes.Thissuppliesachronologicalframework.Otherdocumentsandreportsthathegathersreferringtohischosensubjectwillgenerallyfiteasilyintothissameframework,thoughoccasionallytheywillsuggestaneedforadjustmentinit.Inanycase,themerechronologicalsequenceofwhatoccurredwithrespecttoagivensubjectprovidesabasicskeletalstructureforthehistorian'saccount.Tobesure,thematerialshegathers,allofwhichtreatofthesamesubject,willdifferwithrespecttoscale,someofthemtreatinglargersegmentsofthatsubject'shistorywhereasothersbearevidenceastosomebriefepisodeonly.Thesedocumentsandreportsfallnaturallyintoplacewhen,forexample,sometellofthecourseofawarandothersprovideeyewitnessaccountsofoneofitsbattles.Ingeneral,anysuchsortingofdocumentswillinitiallyproceedeasilyandwithouthitch,andmayproceedwithoutthehistorian'sbeingacutelyawareofwhatheisdoing.Thus,inthe OBJECTIVITYaccumulationofmaterialswithwhichthehistorianissubsequentlytoworktherealreadyisanorder,andthatorderisbothtemporalandanalytic:ThematerialsrefertoasequenceofeventsthatarenotonlylinearlyorderedbutsomeofwhichpossessastructuredreIationshiptoothers,sincesomeareknowntobecomponentsubeventsofothers.Asthehistorianproceedsinhisattempttogatherfurtherrnaterialsconcerninghissubject,thisstructurebecomesmoreandmoreclearlyordered,ratherthanbecomingmoreamorphousbecauseofthewealthofaccumulateddata.Tobesure,amongthedocumentsandreportsthatthehistorianaccumulatestherewiIlbecontradictionsastowhatactuallyoccurred.Whensuchcontradictionsappeartheydonotusuallyunderminethebasicstructureoftheaccountthathasbeenbuilding;theyoftencontradictonlyspecificitemsthatahistorianhaspreviouslyaccepted,anddonotforceachangeinthegeneraloutlineoftheaccount.Thisoccurswhentheconflictinhismaterialsleadstothesubstitutionofoneelementforanother.AlthoughthiswiIlaffecttheaccount-perhapsevenalteringitsinterpretationofsomeothereventsinaradicalmanner-itwillnotinvolveatotaldismantlingofthechronologicalandanalyticstructureintheseriesofeventstakenasawhole.This.istruenotonlywithrespecttotheeffortsofanindividualhistorian,itistrueofthecollectiveenterpriseofhistoriansgeneraIly.Thosedealingwiththesameeventsmayfindthemselvesforcedtoofferquitedifferentinterpretationsofthoseevents,butthesedifferencesdonottotallyalterthebasicstructurethatprevioushistorianshavefoundtobepresentinthesameevents.2Nordoesthefactthatdifferenthistoriansareprimarilyinterestedindifferentfacetsofasocietynecessarilycastdoubtonthedegreeofobjectivityattainableinsequentialgeneralhistories.Anadequateunderstandingofthenatureandchangesofasocietycouldnotbeachievedifallhistoriansfocusedtheirattentionononlyoneaspectofthelifeofthatsociety.Thiswillbecomemoreevidentwhenweconsidertheexplanatoryformofgeneralhistories,butitisrelevanttomentionitnowsincewehavealreadynotedthatallsequentialaccountswillalsoincludesegmentsthatareprimarilyexplanatoryinnature.Thefactthatdifferenthistoriansfocusattentionondifferentaspectsofasocietyandconstructdifferenthistoriesindoingsowouldunderminetheobjectivityofhistoricalknowledgeonlyiftheseaccountswerecontradictory.Yettheyneednotbe.Anysocietyincludeswithinitselfmanydifferentinstitutions,andanyadequateaccountofthenatureofthatsocietymustthereforebebasedonaknowledgeof OBJECTIVITY,CAUSATION,ANDLAWStheseinstitutionsandthechangesthathavetakenplaceinthem.Sincenoonehistoriancanbeexpectedtoinvestigatethehistoryofeachoftheseinstitutionalfactorsindetail,aproliferationofhistoriesofthesamesociety,seenfromdifferentpointsofview,istobeexpected.Conflictamongthemisengenderedonlyinsofarasgeneraltheoriesoftherelativeimportanceofdifferentinstitutionsintervene,withdifferenthistorianstendingtoexplainallotherchangesinasocietyintermsofsomeoneorsomefewinstitutionalfactors.Whenthisoccurs,claimstoobjectivitymustberelinquishedunlesstherearewaystoassessthetruthofthealternativetheoriesthemselves.Thisisnotalwaysimpossible,astheargumentofthelastchapterattemptedtoshow.Thereis,however,anotherdifficultythatarisesbecausedifferentsequentialhistoriesofthesamesubjectconcentrateondifferentaspectsofit.Thisdifficultyconsistsinthefactthatsuchhistoriesarenotlikelytoperiodizetheiraccountsinthesameway:Whatappearsasthebeginningorendofasignificantepochinonemaybeseenasmerelyincidentalinanother.Suchdisparitiesinperiodizationwerealreadynotedinchapter1,buttheirrelevancetothequestionofobjectivitymustnowbediscussed.Werethereonlyonewayinwhichthepastcouldbecorrectlydividedintosuccessiveperiods,thentheexistenceofthemanydifferentperiodizafionsthatoneinfactfindsintheworksofdifferenthistorianswould,ofcourse,entailalackofobjectivityinhistoricalknowledge.However,themannerinwhichahistorianperiodizeshistorydependsuponthefacetofsociallifewithwhichheisconcerned.Inthecourseofeventsthereareindeedpointsatwhichnewdevelopmentstakeholdinasocietyandpointsatwhichtheirdominantrolecomestoanend.Eventhoughonecanalwaysfindrelatedantecedentsforthesedevelopments,andeventhoughonecan'expectalingeringinfluencetoremain,thereoftenaremajorturningpointsthatcanbesaidton1arkoffaperiodinsomeparticularaspectofthelifeofasocietyoraconnectedgroupofsocieties.Itisthereforenaturalthathistorieswrittenwithaparticularaspectofasocietyintheforefrontofattentionshouldtreatofthepastasifitwerepunctuatedbytheseevents.Yet,whenonetakesaviewofthesocietyasawhole,inallofitsseveralaspects,anyperiodshorterthanthatmarkedbyitsoriginanditsultimatecollapse(ifsuchpointsaretobefound)willappearasrelativetothehistorian'sfocusofinterestsince,asawhole,thesocietyfunctionsasacontinuingentity.Theproblemissimilarto"theproblemofgenerations,"whichisin OBJECTIVITYfactoneformthattheproblemofperiodizationsometimesassumes.Strictlyspeaking,therecannotbe"generations"inasociety,thoughthereareinfamilies.Everydaynewindividualsarebornandothersdie,andthestreamoflifedoesnotstartandstopattheboundariesofwhatahistorian-forexample,aliteraryorsocialhistorian-regardsasageneration.Yetthereisgoodreasonforhistorianstospeakofgenerations,markingthemoffintermsofasuddenchangeindominantinterestsorinstyle,sincesuchchangeslinkagroupofpersonsandseparatethemofffromthosewhoprecededthem.However,theselinesofdemarcationapplyonlytooneorafewcommoncharacteristics:Manyotheraspectsinthelifeofagroupofpersonsareapttobecontinuouswithwhatwentbefore.Wereahistoriantobeconcernedwithotherfacetsoflifeinaparticularsociety-say,withitsscienceinsteadofitsliterature,oritsentrepreneurialsysteminsteadofitspolitics-periodsandgenerationswouldbedifferentlydelimited;itwouldthenbeonlyonthebasisofsometheoryastowhichaspectofsocietallifewastobeconsideredprimarythatoneofthesewaysofdividinghistorywouldbegivenpriorityovertheothers.Intheabsenceofanestablishedtheoryofthissort,eachperiodizationmustbeseenasrelativetoaparticularpointofview.Thiswouldseemtounderminetheobjectivityofperiodizations,butonceitisrecognizedonwhattheyarebased,thevarietyofwaysinwhichhistoriesareperiodizedshouldnotberegardedascontradictory:Theirfocusesofinterestmaybedifferent,buttherelationshipsamongtheeventsthattheydescribemaybethesame.Thesituationwithrespecttoperiodizationiscomparabletothatwhichariseswheneverthepastisregardedfromanethnocentricpointofview.AsequentialhistoryconstructedtodepictthepastofwesternEuropeansocietalandculturallifewillleaveoutmuchthatwouldberelevanttothehistoryof,say,Indonesiansociety.Thoughsomecommonantecedentswillappearinboth,thereisnosinglesequentialstreamofhistoricalchangetowhichbothbelong.Thus,itisillegitimatetospeakasiftherewereasinglesequential"worldhistory,"asmanyphilosophersofhistoryandsomesocialevolutionists'havedone.Thosewhohaveviewedthewholehumanpastasifitwereasinglesequentialprocesshavedonesobecausetheyhavebeeninterestedprimarilyintracingouttheantecedentsoftheirpresenthistoricalposition,ratherthanconcernedwithotherstrandsofdevelopment.3Abandoninganethnocentricapproachofthissortdoesnotentailthatinquiriesintothepastthatareconductedfromtwodisparatepointsofviewlackobjectivity.Eachmaybewhollycom- OBJECTIVITY,CAUSATION,ANDLAWSpatiblewiththeothersinceeachwillbeconcernedwithmuchthatisofnoconcerntotheother,andevenwhenthesamesetofevents,suchascolonialexpansion,appearsinboth,itwilloftenbeinrelationtodifferentantecedentsanddifferentconsequences.Ishallnowtakeitforgrantedthattheobjectivityofhistoricalknowledgeisnottobeimpugnedsimplybecausedifferenthistorianswhoareconcernedwiththenatureofaparticularsocietynotonlyworkonverydifferentscales,butmaychoosetodealwithdifferentaspectsofthatsociety.Asaconsequenceofsuchchoices,theirperiodizationofchangesinthatsocietycanbeexpectedtodiffer;oncethereasonforthesedifferencesisunderstood,onecanholdthattheperiodizationofhistoryisrelativetoapointofvieweventhoughaknowledgeoftherelationshipsthatobtainedwithinanyperiodcanbeobjective.Itremainstoshow,however,thatobjectiveknowledgecaninfactbeobtainedinsequentialgeneralhistories.Abasiccluetothesolutionofthisproblemhasalreadybeensuggestedinearlierchapterswhen,contrarytocommonbelief,itwasheldthattheconnectionswithwhichhistoriansareconcernedarenotprimarilylinear,butarepart-wholerelationshipswherethewholesareprocessesandthepartsareeventsincludedwithintheseprocesses(see,forexample,chapter2,section2,andchapter5,section1).Whicheventsaretobeincludedisnotamatterlefttothehistorian'sfreechoice.Werehetopossessaseriesofdocumentsthatstatedonlybarefacts,suchasuWashingtoncrossedtheDelawareatMcKonkey'sFerry,nearTrenton,onDecember25,1776,"hewouldnotknowwhattodowiththem;withoutfurtherknowledgehewouldnotbeabletoconstructahistoricalaccountthatincludedthem.Tointerpretanysuchstatementasreferringtoafactthatcanbeofconcerntoahistorian,itmustbebroughtintocontactwithothermaterialsthatservetoplaceitinsomesocietalcontext,andinthiscasewhatwouldbeinvolvedisrelatingittotheRevolutionaryWar,ofwhichitwaspart.Thisisnot,however,merelyamatterofincludingitinaclassofmaneuvers:Thedateatwhichitoccurredwillbringitintoasequenceofbattles,sieges,advances,andretreats,andtheseareconnectionsthatmustbeestablishedbyfurtherdocumentation.Suchconnectionsarenotinanyrespectrelativetothehistorian'sownpointofview.IamnothereraisingthequestionofwhyWashingtoncrossedtheDelawareatthattime;todosowoulddemandanexplanatoryaccountofhischoiceamongvariousstrategies,anditisnotwiththeexplanatorysegmentsofsequentialhistoriesthatI'amnowcon- OBJECTIVITYcerned.WhatIwishtopointoutissimplythatifahistorianweretostartfromdocumentsthatmerelystatedsingleatomicfacts,hecouldproceedonlythroughasearchforotherdocumentsthatalsoreferredtothesefacts,andwhichplacedthemincontextaspartsofsomeongoingwhole.However,historiansalwayshaveotherprioraccountsthathelptointerpretwhatevernewdocumentstheyfind,andsothey.proceedbyenlargingandcorrectingprioraccountsthroughshowingingreaterdetailthenatureofthoseeventsuponwhichtheprocessofconcerntothemdepended.Thesearchforsucheventsisbynomeansblind.Thematerialswithwhichhistoriansworkarenotrestrictedtodocumentsthatmerelyreportisolatedatomicfacts:Suchmaterialsareapttocontainexplicitreferencestoatleastsomeothereventsuponwhichthatprocessdepended.Tbesecluesmustbesupplementedbyasearchforfurtherdocumentationastowhatoccurredwithrespecttotheeventsalreadyreported,andinthissearchthehistorianwillbeguidedbyhisknowledgeofthestructurebothofhisownsocietyandofothersocieties-inparticularbytheknowledgehehasalreadyobtainedconcerningthesocietywithwhichheisspecificallyconcerned.Theresultisthat,overthelongrun,aseriesofhistoricalinquiriesbuildsupaneverdensernetworkofaccountsreferringtotheeventsthatoccurredinasocietyduringagivenspanoftime.Whilethefocusofinterestoftheseindividualaccountswillvary,andwhilehistoriansmaydisagreeastowhichoftheseeventswereofgreatestimportance,knowledgewillgrowconcerningtheactualstructureofthesocietiesthathavebeenstudied,andchangesthattheyhaveundergonewillhavebeentraced.Whenasequentialhistoricalaccountconcernedwithgeneralhistorytracesaseriesofchangesthatendedinaparticularstateofaffairs,itprovidesalegitimateanswertothequestion,"Whatbroughtaboutorcausedthatstateofaffairs?"Suchananswerdiffers,ofcourse,fromexplanationsthatcitesomesingleeventasthecauseofwhateverstateofaffairsonewishestoaccountfor.Whilethelattermaybeaconvenientenoughshorthandanswerifenoughotherfactorsaretakenforgranted,itshouldbeevidentfromourpreviousdiscussionsthatsuchanswersinvolvegrossoversimplifications.Awar,forexample,canbesaidtohavebeenwonorlostinasinglebattleonlyifonehasalreadytakenintoaccounttheothereventsthatmadeitimpossibleforthelosertofighton.Thus,itisnotonebattleonly,butaseriesofchangesintherelativestrengthoftheopposingforcesthatisneededtoaccountforthefactthatonesidewasforced OBJECTIVITY,CAUSATION,ANDLAWStosurrendertotheother.Similarly,aswesawinchapter5,section1,thelossofpopularityofakingortheriseofaneweconomicpowerisastateofaffairstobeexplainedthroughrecountingtheseriesofsuccessivechangesthatledtoit.Whatgivesobjectivitytosuchcausalaccountsisthefactthat,inhistory,documentationisessentialtotheacceptanceofanyaccount,andwhateverconnectionsaretracedinaseriesofeventsmustbeshowntobesupportedbyallrelevantevidence.Wherelacunaearepresentintheevidencehistoriansmustsearchforfurtherevidence;andwhereverthereisevidencecontrarytoagivenaccount,thatevidencemusteitherbediscreditedorbeprovedcompatiblewithalltheotherevidenceathand.Thequestionofobjectivitybecomesmoreacutewhenweturnfromsequentialtoexplanatoryaccounts.Thereis,aswehaveseen,asenseinwhichasequentialhistoricalaccountoffersanexplanationofwhathasoccured,forittracesaseriesofeventsthatledtosomeparticularterminalstateofaffairs.Suchexplanationsappearadequatesolongastheeventsformingtheseriesareregardedaspartsofasingleongoingprocess,asthespeechesandtelevisioncommercialsinanelectoralcampaignarepartsofthatcampaign.Questionsarise,however,whenitisnecessarytoexplainthesepartsthemselves,sincewhatleduptothemwillnotinallcasesbeincludedwithintheoriginalseries.Inanelectoralcampaign,forexample,aparticularspeechorsetofcommercialsmayhavetobeexplainedbytracingtheimpactofsomepreviouslyunassociatedevent,suchasarevolutioninaneighboringcountry,uponthecourseofthecampaign.Whateverchangesinthecampaignwerebroughtaboutbythateventwillsubsequentlybeseenaspartofthehistoryofthecampaign,andbelongtoit.Nevertheless,toexplainthechangethatoccurredatthatpointinvolvesgoingoutsidethecourseofeventsthatwereincludedinthecampaign.Forthisreason,althoughahistorymaybeprimarilysequentialinstructure,therewillbemanypointsatwhichitwillincludeinvestigationsthatbreakthepatternofsequentialexposition.Consequently,questionsconcerningtheobjectivityofaccountsthathaveanexplanatorystructurearenotonlyimportantintheirownright,butaffectthereliancethatcanbeplacedonmostsequentialaccountsinthefieldofgeneralhistory.Itwillberecalledthatthedifferencebetweenthestructureofasequentialformofhistoricalaccountandanaccountthatisprimarilyexplanatoryinstructureisthatasequentialaccountisconcernedwithasingleongoingprocess,whereasanexplanatoryaccountstartsfromanystateofaffairs,andnotfromwhatisregardedastheendpointof OBJECTIVITYacontinuousprocess;itthenseekstodiscoverwhatotherevents,takentogether,accountforthisoccurrence.Suchaccountsmaytracethebackgroundofsomeoftheseeventsinasequentialmanner,buttheirprimaryfocusofinterestwillbeonhoweventswithdifferentantecedenthistoriesyieldtheresulttheydowhentheycometogether.Givensuchaproblem,thehistorianwillinthefirstplaceexaminewhateverdirectevidencethereistoshowthatsomeoftheseindependentlyoriginatingeventshadanimpactonothers.Suchevidencecansometimesbefoundinthewaysparticipantsthemselvesconnectsomeeventswithothers,asacandidateinacampaignmayconnectarevolutioninaneighboringcountrywiththeneedforachangeinthepoliciesofhisowncountry.Thehistorianweighingtheimpactoftherevolutionontheactualcourseofthecampaignwillthenhavetoestimatetheeffectivenessofthecandidate'sclaim,andthishewilldobygatheringevidencefromeditorials,opinionpolls,andthelike,andbyestimatinginthesamewayhoweffectivetheresponsesoftheoppositionwere.Whiletherewillalwaysberoomfordisagreementinweighingtheevidence,thereisnothingtosuggestthatanysuchdisagreementswarrantaskepticismmoreradicalthanwouldleadustosayitisoftenverydifficulttobecertainhowimportantoneoranotherfactorwasinbringingaboutaresult.Thismeasureof·doubtdoesnotprovideagoodreasonfordistinguishingbetweenhistoryandotherformsofinquiryinsofarastheirabilitytoattainobjectiveknowledgeisconcerned.Asisalsotrueinotherfields,therewillbegreatvariationsfromcasetocaseastotheamountofdirectevidenceonecangatherregardingtheimpactofoneeventonanother;insomecasesthereislittleroomfordebate,thoughinotherstheremaybernajordifferencesofopinion.Inadditiontoanydirectevidencethatcanbefoundregardingtheimpactofoneeventonotherevents,historiansrelyongeneralizationsthatarelawlikeincharacter.ThishasbeendeniedbymanywhoweredissatisfiedwithCarlHempel'sclassicpaperon"TheFunctionofGeneralLawsinHistory,"butthatpaper,asIindicated,wentfarbeyondtheissuewithwhichwearehereconcerned.4Hempelnotonlyarguedthatgeneralizationshaveaplaceinhistoricalinvestigations-whichiswhatIshallargue-butheclaimedthatallhistoricalexplanationsrequiredeductiveinferencesbasedonthesegeneralizations.Oneneednotgosofarinordertoshowthatgeneralizationsarepresupposedinmanyhistoricalaccounts.Inordertoindicatethatthisisthecase,IshallbrieflyexamineW.H.Dray'swell- OBJECTIVITY,CAUSATION,ANDLAWSknownexampleofanengineseizure,whichheusedincriticizingHempel'scovering-lawmodelofexplanation.5Inhisillustration,Drayhasanautomobilemechanicexplaintohimwhatcausedanengineseizurethroughtracingaseriesofstatesofaffairsinwhichaleakontheundersideoftheoilreservoirpermittedalloiltoescape,andthereforetheoilpumpcouldnotdeliveroiltothecylinders,andthemotionofthepistonsagainstthedrycylinderscausedthemtoexpandandthereforetolock;theenginecouldnotthencontinuetorun.Draycantendsthatthis"troubletracing,"whichenablesone"toenvisageacontinuousseriesofhappenings,"doesnotleaveoutanyessentialfactors,andprovidesaparalleltohistoricalexplanations.Unfortunately,however,thisillustrationpresu'pposesanacquaintancewithgeneralizationsatalmosteveryconnectinglinkinthechain.Itpresupposesthatoneknowsthatasubstancesuchasoilwillleakout,becauseofgravity,ifthereisaholeontheundersideofthereservoir,thatfrictioncausesmetalstobecomehot,thatmetalsexpandwhenheated,andsoforth.Withoutanacceptanceofthesefacts,whicharelawlikegeneralizationsandnotstatementsconcerningtheparticularsequenceofeventsinvolvedintheseizure,theseizurewouldnothavebeenexplained.Thus,thiscasediffersfromothers,suchasmyexamplesofelectioncampaigns,inwhichtherelationshipsofpartandwholeareexplicitlycontainedinthedocumentswithwhichthehistorianworks.IdonotuseDray'sexampletodisparagehiscontinuousseriesmodelofexplanationinthosecasesinwhichitdoesapply,butonlytoillustratethattheremaywellbeothercasesinwhichlawlikegeneralizationshaveanimportantroletoplayinhistoricalexplanations.6Economicgeneralizationsprobablyprovidetheleastcontroversialcasesinwhichhistorianscanuselawlikegeneralizationsinordertoshowtherelationshipbetweenoneeventandanother.(Ontheroleofeconomicgeneralizations,seechapter2,section2.)However,somegeneralizationssuggestedbyanthropologists,politicalscientists,andsociologistsmayatleastdirecttheattentionofhistorianstofactorstheymightotherwiseoverlookintheirexplanationsofwhathasoccurred(seechapter5,section2,forexamples).Whileitistruethatmostofthesegeneralizationsareformulatedwithlessprecisionthangeneralizationsineconomics,andwhileeventhelatterareprobablymorerestrictedwithrespecttotheirapplicabilityatdifferenttimesandplacesthanaregeneralizationsinthenaturalsciences,Idonotbelievethesetobetheonlyreasonstheyarelooked 180OBJECTIVITYuponwithsuspicion.Therealsoisatendencyalllongmosthumanisticallyinclinedhistorianstodenythatlawlike:generalizationscan,orshould,beappliedtohumanevents.Widespreadassuchatendencyhasbeen,itrestsonassumingthattobelievethatlawsareapplicableinhumanaffairsistantamounttobelievingthatdeterminismistrue.Aswehaveseen,thisisawhollyerroneousassumption.Infact,itrestsonamisunderstandingoftheroleoflawsinanyformofscientificexplanation,andthismisunderstandingappearstometobethebasicreasonwhymosthistoriansarereluctanttoadmitthatlawlikegeneralizationsmaybeimportantintheirdiscipline.Alaw,itwillberecalled,isastatementconcerninganinvariantrelationbetweentwoormorefactorsthatmaybepresentinmanydifferentoccurrencesofdiversetypes.Inordertousesuchalawtoexplainoccurrencesproducedbytheelementspresentinaparticularstateofaffairs,onemustpossessaccurateknowledgeofthenatureoftheinitialconditionspresentinthatstateofaffairs;onemustalsoknowtheboundaryconditions-thatis,what,ifanything,thatisnotapartoftheinitialconditionswillhaveanimpactontheensuingprocessbeforetheeventinquestionoccurs.Thus,lawsbythemselvesdonotenableonetopredicteventsthatwilllateroccur,eitherinhistoryorinthenaturalorsocialsciences.Furthermore,becausetheselawsconcerntherelationsbetweencertaintypesoffactors,theyareapplicableonlytosituationsinwhich'factorsofthesetypesarepresent.Therefore,iftherearelawlikegeneralizationsinthesocialsciences,ahistoriancanmakeuseofthemonlywhenthesituationswithwhichheisconcernedcontainfactorsofthatspecificsort,justasonecanuseBoyle'slawonlywhenoneisdealingwiththebehaviorofgases.Forexample,anygeneralizationsconcerningrulesofdescentascorrelatedwithmatrilocalorpatrilocalresidencehavenoapplicabilityinoursociety,wheretherearenosuchresidencerules;norwouldanygeneralizationsconcerningtheconsequencesofcrosscousinmarriagesbeofusetohistoriansoftheUnitedStates,sincecross-cousinmarriagesdonotrepresentanynormgoverningacceptablesocialbehavior.Ontheotherhand,notonlyeconomicgeneralizationsbutgeneralizationssuchas"theironlawofoligarchy"andgeneralizationsconcerningthetendencyofAmericansto"votetheirpocketbooks"mayhelptoexplainmanyoftheeventswithwhichhistoriansoftheUnitedStatesareconcerned.Tobesure,notallsituationsinwhichonemightexpecttofind"theironlawofoligarchy"exemplifiedwillconformtoone'sexpectations;nordoestheoutcomeofeveryelectionprovethatthevoters OBJECTIVITY,CAUSATION,ANDLAWShave"votedtheirpocketbooks."Thisshouldnotoccasionsurprise,anditdoesnotofitselfdisprovethatthesearelegitimatelawlikegeneralizations.ObjectsdonotfalltotheearthasonemightexpectonthebasisofGalileo'slawsoffallingbodiesalone,sincewindcurrentsandtheresistanceoftheairmaysignificantlyaffectthefallofsomeobjects,althoughnotthefallofallothers.Similarly,intimesofwarorimpendingwarpeoplemaynot"votetheirpocketbooks,"andtheremayalsobemanyindividualsituationsthatmitigatetheeffectsof"theironlawofoligarchy."Thesegeneralizations-ifauthenticated-wouldnomoreberefutedbysuchapparentexceptionsthantheeffectsofaircurrentsrefuteGalileo'slaws.Assumingforthemomentthatthereissufficientevidencetomake"theironlawofoligarchy"plausibleasalawconcerningthedistributionofpower,itwillpresentthehistorianwithatwofoldopportunity:first,totracethedegreetowhichsomeevidenceisexplainedbyit;andsecond,toseekevidencethatexplainswhy,inaparticularcase,theresultswerenotwhatonewouldhavepredicted,giventheinitialconditionsandthelaw.Silnilarly,whenahistorianacceptsthegeneralizationthatpeopletendto"votetheirpocketbooks"hewilllookforevidenceamongvariousclassesofvotersthattheydidsovote,buthewillalsosearchfor.evidenceastowhatfactors,inthiscase,servedtooffsetthattendencyandledtoaresultotherthanthatwhichhewouldhavepredicted,giventheeconon1icsituationandhisacceptanceofthelaw.Itismycontentionthatevenwhenhistoriansdealingwithgeneralhistorydenytheapplicabilityofgeneralizationsintheirinquiries,theymaketacituseofthemjustthesame.Thisisobviouslytrueinsofarastheypresupposegeneralknowledgeofthewaysinwhichdifferenttypesofindividualsmaybeexpectedtorespondtodifferenttypesofsituations;withoutpresupposingsuchknowledgeahistoriancouldneverinterpretmanyofthematerialsuponwhichheattemptstobuildareconstructionofwhatoccurredin,say,GreecearRame.Tobesure,thereisdangerinusinggeneralizationsofthistypesinceoneisinclinedtoassun1ethatwhateverclassificationsofindividualsandofsituationsoneregardsasnaturalwithinone'sownsocietywillapplyequallytoothersocieties,andthisneednotbetrue.Theprocessofcorrectingerrorsproceedsinthesamewayhereasinanyotherfield.Whentherearecontinuingdiscrepanciesbetweenwhatthegeneralizationwouldleadonetoexpectandwhatactuallyoccurred,onemusteitherchangeone'sanalysisoftheinitialconditions'presentinthesecases,ormustalterthegeneralizationitself.Inthe OBJECTIVITYfieldofsocialtheory,bothhaverepeatedlyoccurred.Insomecasesthesystemofclassifyingindividualsandtypesofsituationshashadtobealtered,asourconceptionsofwhatistocountas"religion"or"afamily"havealteredasacquaintancewithnonliteratesocietieshasgrown;inothercases,whathasbeencorrectedwerethegeneralizationsthemselves,aswhenknowledgeofothersocietieshasmadeitclearthatonecannotassumeanyinherentconnectionbetweenmaterialrewardsandwillingnessonthepartofapopulationtoengageproductivelyinarduouswork.Asinotherfields,itisonlythroughcontinuousattentiontoevidencethatthegeneralizationshistoriansusearediscarded,refined,orconfirmed.Inexplanatoryhistoricalaccounts-andalsoinexplanatorysegmentsofsequentialorinterpretivehistories-thefunctionofgeneralizationsistohelpexplainwhytwoormoreindependentseriesofeventsthatintersectataparticularplaceandtimeproducetheresultstheydo.ToillustratethisfunctionIshallfirstchooseanobviousexample.SupposethatasteepriseoccurredingrainpricesontheChicagocommoditiesexchange,andoneseeksanexplanationofthatfact.Thegeneralizationthatshortagessendpricesupsuggestsitself;andsincewhatistradedaregrainfutures,ananticipationofshortagescanbeheldresponsiblefortheriseinprices.Thus,aneconomicgeneralization,plusacomprehensionofwhattheexchangeisdealingin,offersanexplanation,butonetoolackinginspecificdetailtobesatisfactory.Foramoresatisfactoryexplanationonemustdiscoverwhataccountedfortheanticipationofshortages,andheregeneralizationsarenotlikelytohelp.Rumorsofbadcropsinoneormoremajorproducingareascouldbetheexplanation,butsocouldthelikelihoodofawar,whichwouldinterferewiththeproductionofgrain.Althoughsuchcausalaccountsofwhyashortagewasexpectedmaynotrelyongeneralizations,atleastonegeneralizationwasessentialintheexplanation:Ithadtobeassumedthatthereisadirectrelationshipbetweenshortages,whatevertheircauses,andariseinprices.Withoutthatknowledge,theexplanationoftheriseinpriceswouldcollapse.Thus,inahistoricalinvestigationthatisexplanatoryinform,ratherthanmerelysequential,onecanexpecttofindsonlegeneralization,butwhatisalsoneededisknowledgeoftheconcretenatureofthesituation,andofhowoneistoaccountforthepresenceinthesituationofthosespecificfactorsthatthegeneralizationemploys.Inthiscasethegeneralizationwasconcernedwiththeconnectionbetweenpricesandsupply.Knowledgeoftherelevantfactorsinthesituationincludedknowledgethattradingisbasedonanticipa- OBJECTIVITY,CAUSATION,ANDLAWStionofthenextharvest,andknowledgeofthecurrentpriceofgrainwhichhadbeengovernedbyprevioussupplies.Whatwasinthiscaseneededforthegeneralizationtoyieldaconcreteexplanationoftheriseinpriceswastoaccountfortheanticipationofafutureshortage;thispartoftheexplanationdidnotdirectlyrelyonanyfurthergeneralization,butonananalysisofwhatledtothebeliefsthatdrovepricesup.Becauseitisnecessaryforahistoriantoknowanddescribetheconditionsinitiallypresent,andtoaccountforthepresenceofwhatevernewfactorhasbeenintroducedintothatsituation,andbecausethesearemattersthatchieflyoccupyone'sattentioninreadinghisaccount,theimportanceofgeneralizationforhisanalysisistooeasilyoverlooked.Yet,astheprevioussimpleexampleshows,theuseofageneralizationisoftencrucial.Alessobviouscasethatcanservetoillustratethesamepointwouldbethefollowing.Ahistorianknowsfromcensusreportsthatoveraperiodoftwodecadestherehasbeenamarkeddeclineinthepopulationofruralareasandanincreaseinthepopulationofurbanareas,andhewishestoexplainwhythisshifthasoccurred.Migrationfromtheruralareastothecitieswouldprovideoneobviousexplanation,butothersarealsopossible;onemightbeamarkeddifferenceinthebirthratesoftheareas,andanotherwouldbeamarkeddifferenceintheirdeathrates.Thesepossibilitiescouldpresumablyberuledoutbychoosingasampleofruralandurbancountiesandexaminingtheirbirthanddeathstatisticstoseewhethertherehavebeensignificantdifferentialchangesinthenumberofbirthsanddeathsthatwererecorded.However,ahistorianwouldnotbelikelyinthefirstinstancetotakethisapproach,notbecauseitmightnotberelevant,butbecausehisbackgroundknowledgeofthefactorsthatcausemajorshiftsinthenumberofbirthsandofdeathsinagivenpopulationwouldnotseemtofitwithanydifferencesinconditionsinruralareasandincitiesthatwouldbesufficienttoaccountfortheshiftthathasoccurred.Thus,lawlikegeneralizationsconcerningthefactorscapableofinfluencingratesofbirthandratesofdeathwithinanysocietywillbeusedbyhistorianstoruleoutsometypesofexplanationaslikelywaysofaccountingforapatternofhistoricalchange.Inaddition,thehistorian'spriorknowledgeofthefrequencywithwhichshiftsinpopulationhaveoccurredbecausepeoplehavemovedinlargenumbersfromoneplacetoanotherwouldnaturallyleadhimtoinvestigatemigrationastheexplanationofthispopulationchange.Wherewouldhebeabletogatherevidencetoshowthatthisdidinfactprovidetheexplanationforwhichheislooking?Inalarge-scale OBJECTIVITYandnationwideshiftinpopulationfromruraltourbanareas,anyattempttorelysolelyuponaninterviewingtechniquewouldbeboundtofail.Thus,adirectcollectionofdatawouldnotinthiscasebeabletoprovideanexplanationofwhatoccurred.Tobesure,asamplingofpersonswhoareknowntohavemigratedfromoneregiontoanother-eitherinotherinstancesorinthecaseunderconsideration-couldrevealthekindsoffactorsthatleadtomigrations;forexample,thatpeoplemigrateforeconomicreasons,andthattheyalsosometimesdosobecausetheyfeelthemselvestobepoliticallyorsociallyoppressedandexpectabetterlifeelsewhere.Thus,atleasttwolawlikegeneralizationscouldbeformulated,oneconcerningthecorrelationbetweeneconomicfactorsandmigrationandtheotherbetweenoppressionandmigration,andthehistorianmightlooktoseewhich,ifeither,appliestothesituationhewishestoexplain.7Tbiswouldinvolveanalyzingtheconditionsactuallyprevalentintheruralareasandintheurbanareaswithwhichheisconcerned,inordertoseewhethereitherofthesegeneralizationsisapplicable,orwhetherbothare.Havingreachedatentativedecision,sampleinterviewscouldbeusedtoconfirmhishypothesis,orcouldleadhimtoreviseitinonewayoranother.(Forexample,interviewsmightshowthateconomicmotivesweredominant,eventhoughanexaminationofactualeconomicconditionsmightshowthatnoimportantdiscrepanciesexistedatthattimebetweenopportunitiesinruralandurbanareas.)Whatisimportant,however,istheroleofgeneralizationsinsuchaccounts.Theevidenceobtainedbyinterviewsandbycomparativestudiesofconditionsintheareasconcernedisofcourseimportantforthehistorianinauthenticating,throughfactualdetail,themannerinwhichhehasaccountedforthechange;withouthavingusedgeneralizations,however,hewouldnotinsuchcaseshavebeenableeventomakeastarttowardofferinganaccountofwhathewantedtoexplain.Itistobenotedthatthelawlikegeneralizationsthathavebeenmentionedaredifferentfromthesortoflawlikestatementsusuallybroughtunderattackwhentheproblemisdiscussedbythosewhorejectthepossibilitythatthereareexplanatorylawsapplicabletohistory.ThetypesoflawsthatIhaveciteddonotattempttoformulateinvariantconnectionsbetweenactualoccurrencesofspecificsorts,suchthatgivenanyeventoftypea,anothereventoftypebwillalwaysfollow.8Instead,theyconcernrelationshipsbetweenfactorsthatmaybepresentinanynumberofsituationsofdifferenttypes;theinvariantconnectionstheyattempttoformulateconcernthefunc- OBJECTIVITY,CAUSATION,ANDLAWStionalrelationshipsbetweenthesefactors,notconnectionsbetweenparticularsortsofoccurrences.9Whatoccurscannotthenbesaidtooccursimplybecauseanoccurrenceofanothertypeprecededit.Instead,whatfollowswilldependupontheprecisenatureofthesituationandthefactthatwhenoneoftwocorrelatedfactorschanges,theotherwillchangeaswell.Sincebotharepresentinthesamesituation,thesituationwillhavechangedfromwhatitoriginallywas.Theforegoinganalysisshouldsufficetoshowthatlawlikegeneralizationsplayanimportantroleinhistoricalaccountswhentheseareexplanatoryintheiroverallstructure.Itshouldalsosufficetolnakeclearthattheuseofsuchgeneralizationsdoesnotsupplanttheneedfortracingasequenceofeventsifaconcreteexplanationistobegivenofthechangesthathaveoccurred.Thisisevidentinthefactthatwhenahistorianreliesonageneralizationtoaccountforachange,hemustalsoshowwhatprioreventsleduptoachangeinoneofthefactorswithwhichhisgeneralizationisconcerned.Thus,theuseoflawlikegeneralizationsinhistorycannotbeheldtosupplantsequentialhistoricalaccounts,asHempelbelievedtheyshould.Althoughanyexplanatoryaccountwillinvolveasequentialmodeofexplanationaswellastheuseofgeneralizations,andmanyaccountsthataresequentialinstructurewill,atplaces,makeuseoflawlikegeneralizationstoexplainwhatoccurred,thereneverthelessremainfundamentaldifferencesbetweenhistoricalaccountsthatareprimarily'explanatoryandthosethatareprimarilysequentialinstructure.AsIpointedoutinchapter2,sequentialaccountsareconcernedwithsomecentralsubjectandtheirstructureisdirectedbyanattempttotraceitscourseovertime.Whiletheymayincludementionofmanyeventsotherthanthosethatformpartofthehistoryoftheircentralsubject,theydosoonlybecausetheseeventsimpingedonit,alteringitscourse.Theprimaryconcernofahistorianwillremaintheonecontinuingstrandofhistorythathesetouttotrace.Inspecialhistories,aswehaveseen,suchcentralsubjectswillnotbecontinuingentities,butwilldependuponthehistorian'scharacterizationoftheclassofobjectswithwhichhewishestodeal.Ingeneralhistories,ontheotherhand,thecentralsubjectwillbeasocietyorsonlefacetofasociety,oritmaybeanindividual'slifeorsomeaspectofhislife.Thisaccountsforthedifferenceintheobjectivitythatcanbeattributedtogeneralhistoriesasopposedtospecialhistorieswhenbotharesequentialinstructure.Whenahistoriansetsouttodealwitha 186OBJECTIVITYparticularsocietyorwithsomeindividualorwithsomefacetofthehistoryofeither,hecannotpickandchoosewhatevidenceisrelevanttohistask:Neitherthenatureofasocietynorthelifeofanindividualissomethingthatisalteredbyone'scharacterizationofit.Tobesure,everyhistorianworksonsomescale,andwhenheisdealingwitheventsononescaleheneednotincludeinhisaccountallthatwouldhavetobeincludedwereheworkingonanother.Thisdoesnotsignify,however,thathehasbeenfreetochoosewhatisrelevanttohisaccount;hisfreedolTIhasbeenexercisedinchoosingthesubjecthehaschosenandchoosingthescaleonwhichheistowork,butoncethesechoiceshavebeenmadehisfreedomtorejectevidencedisappearsbecausetheeventswithwhichhedealspossessastructureoftheirown.Itisbecausesocietieshaveastructureandarenotmerelyanagglomerationofeventsthathistoricalaccountsconcernedwithgeneralhistorycanprovidesatisfactoryexplanationsevenwhentheyonlytraceacontinuoussequenceofevents.Thisispossiblebecausethoseeventsarerelatedtooneanothernotmerelywithreferencetothetemporalorderinwhichtheyoccurred,butalsoasphasesorascomponentpartsofasinglecontinuouswhole.10Inanexplanatoryaccount,ontheotherhand,thehistorian'sattentionisdirectedtoasinglestateofaffairs,nottoacontinuousprocess.Whatheseekstoestablishiswhatprioreventswereresponsibleforthatstateofaffairs;thus,hestartsbackwardintimeandthenfollowstheresultsoftheseprioreventsuntiltheyhaveterminatedintheparticularstateofaffairsthathehassetouttoexplain.Tobesure,hisbackwardsearchisnotendless,foroncehavingfoundtheparticularsetofeventsthatbroughtaboutthestateofaffairsheistoexplain,heneednotaskwhatwasresponsiblefortheseeventsthemselves.Toaskthosequestionswouldleadtofurtherexplanatoryaccounts,buttheseaccountswouldnotbepartoftheexplanationwithwhichhewasoriginallyconcerned.Inadditiontothedifferenceinthedirectioninwhichhistoriansproceedinsequentialandexplanatoryaccounts,thereisanotherdifferencewehavealsonoted:thatinexplanatoryaccountstheremaybeincludedmanydifferenttypesofeventthathavenorelationtooneanothersaveforthefactthat,havingcometogetherataparticularplaceandtime,theywerejointlyresponsiblefortheparticularstateofaffairsthatthehistorianwishestoexplain.Theeventsthatservetoexplainthisresultarenotinsuchcasespartsofonecontinuousprocess;thereforehistoricalchangemusttothatextentbe OBJECTIVITY,CAUSATION,ANDLAWSregardedassubjecttomanycontingencies,ratherthanasdevelopinginadeterminedmannerthroughself-transformationswithinaclosedsystem.Neitherofthesefeaturescastsdoubtontheobjectivityofhistoricalinquiriesthatareexplanatoryinstructure.Tobesure,ifitweretruethatwhenahistorianattemptedtoexplainaparticularstateofaffairshehadtostartoutfromapositionoftotalignorance,hewouldnotknowwheretolookforthelikelyantecedentsofthatstateofaffairs,andhewouldhavenotentativehypothesesconcerninghowanytwofactorsinasocialsituationarelikelytoberelated.Inhistory,however,asinscience,oneneverstartsdenovo.Inmostcasesthehistorianhasathisdisposalpriorsequentialaccountsthatincludereferencetothestateofaffairsinwhichheisinterested;hethusknowssomethingaboutthestreamofeventsinwhichitwasembedded,andthereforeabouttheeventsprecedingit.When,however,itappearstohimthattheseprecedingeventsdonotadequatelyexplainthestateofaffairsinwhichheisinterested,hewillbeledtoofferanexplanatoryaccountofit.Indoingso,hewillbelookingforothercontributingfactors,andherehewillbeguidedbythesortsofhypotheseshistorianshaveusedinothercases,andbyhisownconjecturesastothekindsoffactorsthatcanbeexpectedtobringaboutaresultofthiskind.Thisrelianceontheaccumulatedexperienceofhistorianswithrespecttowhatsortsoffactorsarelikelytoaccountforaparticularkindofresultparallelswhatonefindsinthesciences.Sotoodoesthefactthatindividualhistoriansmustalsorelyontheirowninsightwithrespecttowhatmaybeimportantinaparticularcase,andtheymaytherebydiscoverrelationshipsbetweenfactorsofwhichtheirpredecessorsweregenerallyunaware.Inadditiontoaccumulatedexperienceandnovelinsights,generaltheoriesprovidestillanothersourceforthehypotheseshistoriansuseinexplainingparticularstatesofa�airs.Aswehavenoted,historiansoftendenythattheymakeuseofanygeneraltheoryintheiractual.practice,butthisisusuallybecausetheytaketoorestrictedaviewofwhatcomprisesatheory,identifyingitwithsomegeneralpropositionstatingthatallhistoricaleventsaretobeexplainedintermsofsomesinglefactor,suchasthemodesofproduction,ortheviewthatthebasisofallhumanactionisadriveforpower.Supposedly,suchtheoriescanbedirectlyappliedtotheeventswithwhichhistoriansareconcerned,providingexplanationsforthem.Ontheotherhand,aswehaveseen,therealsoaremoregeneraltheoriesconcerningthewaysinwhichoneistoconceiveofasocietyandofsocietalchange, 188OBJECTIVITYandthesearenotintendedtoexplainparticularhistoricaleventsanymorethantheparticulatetheoryofmattercandirectlyexplainspecificphysicalphenomena.Theoriesofthissortleaveroomforagreatvarietyofconcretehypothesesthathistoriansmayuseintheirexplanatoryinquiries.Nonetheless,theycontrolthedirectioninwhichahistorianlookswhenheattemptstoformulatethesemorespecifichypotheses.Thus,ageneraltheory,plusaknowledgeoftheinitialconditionsandthetypeofeventthatoccurred,willchannelahistorian'ssearchforalawlikegeneralization,leadinghimtocastasidesometypesofhypothesesandtotestothers.Thisneednothaveadistortingeffectonexplanatoryinquiries,underminingtheirobjectivity,solongasthevalidityofthehypothesisistestedintermsofallavailablerelevantevidence,notonlyinthiscasebutinothersinwhichsimilarfactorsaretobefound.Thispointcanbeillustratedwithreferencetothoseexplanatoryaccountsthatappearwithinspecialhistories.Specialhistories,itwillberecalled,haveasequentialstructure;however,likeanyothersequentialhistoriestheyarelikelytoincludemanyexplanatorysegments.Inahistoryofliterature,forexample,anattemptmaybemadetoaccountforaspecificchangeinstyle,andindoingsoaliteraryhistorianmaycallattentiontopoliticalandsocialchanges,toliteraryinfluencescomingfromabroad,ortotheimpactofchangesinreligionorphilosophyorinoneofthearts.Whiletheacceptanceofsomespecifictheory,suchassomeformofMarxisttheoryofthearts,hassometimesledliteraryhistorianstoexplainthesechangesintermsofasingle,dominantfactor,theexplanationsofaliteraryhistorianmayderivefromamoreflexibletheory,whichholdsthatculturalchangemustinallcasesbeassumedtoberelatedtoanumberofdifferentfactors,allofwhicharetobefoundineachandeverysociety.Forexample,hemayhold-asamatterofgeneraltheorythateveryaspectoftheculturepresentinanysocietyisalwaysrelatedtoitsowntraditionsaswellastoatleastsomeoftheotheraspectsoftheculturefoundinthatsociety,andthatitisalsoalwaysinfluencedbythestructureofthesocietyasawhole.Usuallysuchtheoriesareonlytactitlyheld,buttheyexertanimportantinfluenceonhowaspecialhistoryofsomeaspectofculturewillproceed.Tobesure,solongasoneisconstructingonlyasequentialaccountofwhatoccurred,generaltheoriesofthistypeareofonlylimiteduse;theysimplysuggestsomeofthequartersinwhichahistorianistolookinordertogathermaterialsrelevanttohissequentialaccount.However,assoonastheevidencesuggeststhatwhathasoccurredinthese- OBJECTIVITY,CAUSATION,ANDLAWSquenceapparentlycannotbeexplainedintermsofwhathasgonebefore,hemustlooktoeventsoutsidethatsequence.Herehisgeneraltheoryofthetypesoffactorsthatcanexplainculturalchangewillsuggesthypothesesastowhatotheroccurrencesmighthelptoexplainthechangethattookplace.Thereisnoreasontothinkthatthesamehypotheseswillbeadequatetodealwithalldifferentformsofculturalchange,butthetypeofgeneraltheorywithwhichwearehereconcerned(incontradistinction,say,toMarxisttheory)wouldnotdemandthatsuchbethecase.Itwouldmerelysuggesttypesoffactorsthatmightbecorrelatedwithoneanotherinexplanationsofculturalchange,leavingittoinvestigationsindifferentfields-suchashistoriesofliteratureinthemodernperiod,orhistoriesofphilosophyoroftechnology-toformulateconcretehypothesesthathelptoexplainthevariousspecificchangeswithwhichtheyareconcerned.Theforegoingremarksnotonlyshowthatgeneraltheoryisofuseintheexplanatorysegmentsofspecialhistoriesthatarepredominantlysequentialinstructure,buttheyalsoservetosuggestthatonecanclaimobjectivityfortheseparticularsegmentsofspecialhistories.Objectivitymaybeclaimedforthemsincethehypothesesonwhichtheyproceed,whichweresuggestedbygeneraltheory,aresubjecttothesamecheckagainstevidenceasisthecaseinanyotherexplanatoryaccount.Aswehaveseen,thisobjectivitycannotbeclaimedforanyspecialhistorytakenasawhole,sincedifferentculturalhistoriansdefinethesubjectmatterwithwhichtheyareconcernedindifferentways,andwhatisincludedineachwillthereforebesomewhatdifferentfromwhatisincludedinothers.However,assoonasaneffortismadetoexplainwhateverchangeshavecomeaboutwithinaseries-howeverthatserieshasbeendefined-theneedforevidencewillprovideacheckontheexplanatoryhypothesesthatwereused,andthepossibilityofobjectiveknowledgereemerges.Inshort,aclaimfortheobjectivityofhistoricalknowledgerests,inthefirstinstance,onthefactthathistoriansmustsupplyevidenceforthestatementstheymake.Thisfactwouldnotofitselfestablishobjectivitywereitnotthattheeventswithwhichhistoriansusuallydealfitsnuglytogether,sothatonehistoricalaccountcanlendsupporttoanother,ortheirfailuretofitwillbeobviousand,asaconsequence,atleastonewillhavetobemodifiedorabandoned.Inthecaseofspecialhistories,however,wherehistorianscanstructuretheirmaterialstoconformwiththeirowndefinitionsofwhatistobelong,ornotbelong,intheiraccounts,objectivitywillnotbeattainableinthehistoryasawhole,butonlyinitsexplanatoryportions.In OBJECTIVITYtheseportions,whenahistorianis,forexample,attemptingtotraceindividualinfluencesoraccountforbasicchangesinthecharacteroftheworkswithwhichhehaschosentodeal,whatconstitutesrelevantevidenceisnolongersubjecttoredefinitionbyhim:Havingoncedefinedthematerialswithwhichheistodeal,hewillhavetofollowwheretheevidenceleads.Thefactthatinapproachingtheproblemofexplainingwhatoccurredhemaybringtobearspecifichypothesesormoregeneraltheoriesthatheisalreadyinclinedtoacceptdoesnotvitiatetheclaimthatsuchexplanationscanbecharacterizedasobjective,asIamusingthatterm.Thehypothesesortheoriesmustthemselvesbesupportedbypriorevidence,andmustbeshowntobeapplicableinthisinstancebyevidencethatcanbegatheredinsupportoftheexplanationthatfollowsfromthem.Wheninsuchcaseshistoriansdisagree,thesedisagreementsinweighingevidencedonotentailthathistoricalknowledgeisnotobjective,anymorethantheparallelsituationin,say,psychologyorbiologywouldleadtothatconclusion.Arelativistmightobjectthatallthisiswellandgood,butthatitismerelyaconsequenceofthewayinwhichobjectivitywasdefined.Whatpurposedoesitserve,hemightask,ifonecanclaimthathistoricalknowledgeisinprincipleobjectiveandthatiftwohistoriansdisagreeatleastonemustbewrong,whenallthetimeweknowthatinpracticehistoriansdisagree,anddisagreebecausetheiraccountsarenotobjectiveineitheroftheothertwosensesthattheterm"objective"mayhave(chapter6,sectionI)?Tothiswemustnow,inconclusion,reply.Withrespecttotheclaimthathistoricalaccountsarenotobjective,inthefirstofthesensesof"objective"thatIdistinguished,itmustberecalledthatthereisagreatdifferencebetweentwoquestionsthathavesometimesbeenconfused:(a)thereasonswhyahistorianmayhaveelectedtowriteaboutsomeparticularsubjectmatter,and(b)thereasonswhyheexplainedthatsubjectmatterashedid.Itisnecessarytodrawthisdistinctionnotonlywithrespecttotheproblemofhowahistorian'sinterestsandvaluesaffecthiswork,butalsobecauseconfusionhassometimesarisenregardingthesenseinwhichexplanationsarecontext-determined.AsIarguedinchapter3,section2,andchapter4,section1,thequestionsindividualsaskandthekindsofanswerstheyexpecttoreceiveareinfactcontext-determined.Nevertheless,whenoneiscalledupontoassesstheadequacyofanexplanation,onemustdosowithrespecttoitsrelationtotheevidenceuponwhichitrestsandonthebasisofitsrelationtoother OBJECTIVITY,CAUSATIOX,AXDLAWSevidencesuppliedbyindependentaccounts.Theadequacyofamaterialexplanationisnotassessedvithrespecttowhatevermayhaveledthisparticularhistoriantooffertheexplanationhegave.Thisobviouspointiseasilyoverlookediftheworksofoneortwohistoriansareconsideredapartfromtheaccountsofthesamesubjectmatterthathavebeenbuiltupoverseveralgenerations.Throughacontinualaccumulationandsiftingofevidence,generalagreementisactuallyreachedconcerningmuchofwhatoccurredinthepast;however,whenspecificworksareindividuallyconsidered,theirdifferences,ratherthantheirsimilarities,arelikelvtostrikeoneastheirmostinterestingcharacteristics.Thisfixation"ofattentionondifferenceswillleadonetooverlooktheexistenceofthemoregeneralframeworkofhistoricalknowledgeintowhicheachofthesedifferingaccountsmayactuallyfit.Tobesure,theadvancementofhistoricalknowledgeisnotsmoothlycontinuous;therearemanyunexpectednewstartsasnewissuesariseoutofchancediscoveriesandoutofnewinterests,butthesesuddenshiftscanoccuronlybecausetherealreadyisanacceptedbackgroundofknowledgethattheynlaychallengeatpoints,butwithoutwhichwhateverisnewinthemwouldlackmeaning.11Itisalsopossibletoindicate,inanequallybriefform,agoodreasontorejectthebeliefthathistoricalaccountscannotbeobjectiveinthesecondofthesensesof"objective"thatIhavedistinguished.Thatbeliefrestsontheassumptionthattheorderandconnectionoftheeventsinahistoricalaccountisnotacharacteristicoftheeventsthemselves,butthattheydependuponthewayinwhichthehistorianorganizeshisfacts.Thisassumptionrests,however,onaconfusionbetweentheeventsthemselvesandtheevidenceforthoseeventswithwhichahistorianmustdeal.Suchevidencecancometohiminanyorder.Furthermore,thekindofevidenceavailableconcerningtwoevents,orconcerningtwoaspectsofthesameevent,maybequitedifferent,andneitherbitofevidencemaycontainanyreferencetotheother.Thisdoesnotmean,however,thatwhatisreferredtoinonebitofevidencewasinfactindependentofthatwhichwasreferredtointheother.Thereisnomysteryinhowsuchrelationshipscanbeestablishedevenwheneachbitofevidence,iftakenbyitself,failstorevealaconnectionbetweenoneeventandtheother.Eachmayrefertoaneventthatoccurredataparticularplaceandtime,andthehistorianwhohasaccesstobothwillthenfindthatheisdealingwithsuccessiveoccurrences,orwithwhataredifferentfacetsofthesameoccurrence.Thesearenotrelationshipsthathemakes;theyarerela- OBJECTIVITYtionshipshehasfoundonthebasisofevidence.Topushbeyondthismerelyskeletalknowledgeandfindoutjusthowtwosuccessiveoccurrencesortwofacetsofthesameoccurrencemayhavebeep.relatedposesafurtherproblem,butit,too,isto·besolve.dbymeansofevidence.Ahistorianmay,forexample,lookforsuchfurtherevidenceinstatementsmadebypersonswhoparticipatedintheseevents,orinthereportsofothers.Hemayalsoformahypothesis,basedonothercases,astohowtheeventswererelated,buttheapplicabilityofsuchahypothesistothecaseathandwillthenhavetobecheckedbystillfurtherevidence,orwillatleasthavetobemadeplausiblebyafailuretofindanydisconfirmingevidence.Thathistoricalaccountsgrowinthiswayisnotimplausiblewhenonenotesthattheevidencewithwhichhistoriansmostfrequentlydealisofasortthatrefersnottooneeventonly,buttoaneventinitscontext:toabattleinawar,aspeechinacampaign,theformationofanalliancebetweentwonationsagainstthethreatofacommonenemy.Thus,itisonthebasisoftheevidencewithwhichheworksthatahistorianisledtoconsidereventsasbelongingtogethertoformaseriesofrelatedoccurrences-awholeseriesofbattlesorofspeechesorofpoliticaldecisionsleadingtoanalliance.Thus,theonlybasisonwhichit,mightbeclaimedthathistoricalaccountsaresubjectiveinthesensethattherelationstheytracedependuponthehistorianhimselfisundermined.Onceevidenceisbeinggathered,itisthisevidenceitselfthatimposesstructureonanyhistorian'saccountsolongasheholdsclosetothatevidence.Giventhesefacts,wemayalsodismisstheargumentoftenusedbyCharlesA.Beard,amongothers,thathistorycannotbeobjectivebecausethehistorianmustselectfromamongawelteroffacts,decidingwhattoincludeandwhattoexclude.Theresultingaccount,itisclaimed,reflectswhateversubjectivefactorscontrolthehistorian'schoices,ratherthanwhateverobjectiveconnectionsmayhaveoriginallyexistedamongtheeventsthemselves.Thisargumentis,however,convincingonlyifoneoverlooksthefactthatwithinthernaterialswithwhichhistoriansworktherearedistinctionsofscaleandofthefacetsoftheparticulareventswithwhichtheydeal.Thus,itisnottruethatthesortalprincipleshistoriansuseinclassifyingevidencearemerelysubjective.Bearingthesedistinctionsinmind,onemayholdthatabasicstructureisimposedonahistoricalaccountbytheevidenceonwhichitrests;theexistenceoflacunaeinthatevidence,andthenewquestionsthatarepresentinit,directthehistorian's,attentiontotheneedforfurtherevidenceofaspecific OBJECTIVITY,CAUSATION,ANDLAWSkind.Thatthisshouldbethecaseisnotmysterious.WhateveristakenasevidencepointsbeyonditseltJustascombinationsofcolorsmaybesaidtocallattentiontothemselves,orlinesleadustofollowadesign,sothepreviouslyavailableevidencewilloftenbeseenashavinggapswithinit,orasbeingsuggestiveofanewdirectioninwhichotherevidencemaybefound.Thus,whateverevidenceisoriginallyavailabletoahistorianwillnotbeaninchoatemass,andthemoreevidencethereis,thelesschoicehewillhaveastothealternativewaysinwhichhemayreasonablystructurehisaccount.ItshouldnowbeapparentthatthereisacloseconnectionbetweentheviewofcausationIearlierdefendedandthedefenseofobjectivityIhavenowproposed.Iarguedthatanadequatecausalexplanationdoesnotconsistinsomerubricassertingthataneventofonetypeisregularlyfollowedbyaneventofanothertype.Historiansdonottrytoexplainwhatoccursinthismanner;theyareinterestedinparticularevents,notinalleventsofagiventype.Infact,theyareoftenespeciallyinterestedineventsthatdonotconformtowhatmighthavebeenexpectedonthebasisofwh·atoccurredinothercases.Whentheywishtoexplainanyevent,whetheritmightormightnothavebeenexpected,theyattempttodetermineonwhatconditionsitsoccurrenceactuallydepended,andthisinvolvesanalyzingtheconcreterelationsbetweenit,itscontext,andaseriesofprioreventswithwhichitwasdirectlyconnected.Itisonthebasisoftheconnectionsinherentintheevidencewithwhichhistoriansworkthattheycanproposeconcretecausalanalysesoftheeventswithwhichtheydeal.Consequently,themoreevidenceahistorianpossessesconcerningwhatoccurredinasocietyataparticulartime,thelessarbitraryandperfunctoryhisassertionsofcausalconnectionscanbe.Itisnot,.ticthen,atedon;ittheisbasisontheofbasisgeneraloflawevisdenthatcecaasusaltocowhnnatectacionstuallyareoccurrauthen-ed.Lawlikegeneralizationsareatsomepointsusefulinestablishingthesortofcovertconnectionthatexistedbetweentwoevents,butinmostinstancestheconnectionamongeventsisanopenrelationship,directlyattestedtobytheevidenceitself.Tobesure,historiansmustexercisecriticalreserveinacceptingtheassertionsofparticipantsastothetruecausalconnectionsamongevents.However,whatservestocorrecttheassertionsofparticipantsisnotageneralskepticismbasedontheclaimthatsuchassertionsalwaysrepresentsomeformofbias:Theycanbecorrectedonlybymoreevidenceshowingthatitwasindeedbiasthatledtotheassertionsmade.Itshouldnotoccasionsurprisetohavesomeonearguethathistori- OBJECTIVITYcalaccountsshouldbejudgedonthebasisofevidence,yetthisclaimwillundoubtedlybeviewedasanextremeexampleofsimplemindedness.Onehasheardsomuchconcerningthesubjectivefactorsthatenterintohistoricalaccountsthatonenowalmostforgetsthatthecreativityofahistorianisshowninhowhehandlesevidenceandwherehehashadtheinsighttolookfornewevidence;itisnotthathehastoldanoldstoryinanewwayandthathistellingofithasmadeitseemofgreaterinterestthanitpreviouslywas.Historyasadisciplineisnotaformofart,andwhateachhistorianaccomplishes.hedoesnotaccomplishalone,asanindividual:Thesignificanceofanyhistoricalinquiry,liketheresearchofanyscientist,dependsdirectlyuponwhatothershavealreadydoneorwillbeenabledtodobecauseofhiswork.Thisiseventrueinthecaseofspecialhistories,sinceeventhoughtheycannotlayclaimtothesamedegreeofobjectivityasistobefoundinanyformofgeneralhistory,wehaveseenthatatmanypointstheydependupontheworkofthoseconcernedwiththenatureofandchangeswithinsocieties,andthattheirinquiriescancontributetotheknowledgethatgeneralhistoriansseek.Whenonetakesintoaccounttheinterplayofgeneralandspecialhistories,andthefactthatalthoughsequential,explanatory,andinterpretiveinquiriesarediverseinformtheycannotexistincompleteisolationfromoneanother,theanatomyofhistoricalknowledgerevealsaunityofpurposeandaunityofmethod:tounderstandtheconcretenatureofsocieties,thechangestheyhaveundergone,andtheculturalproductstheyhaveproduced. -AppendixAHUMETherehavebeenmanyattacksontheatomisticsensationalismwhichwascharacteristicofBritishEmpiricistsingeneralandwhichprovidesanindispensableelementinHume'sanalysisofcausation.WhileIsharetheviewsofsomewhohavecriticizedHumeonthesegrounds,Idonotbelievethatoneneedmountawholesaleattackonhispsychologicalandepistemologicalassumptionsinordertoshowthathisargumentcontainsaflawthatvitiatestheattempttoprovethatwedonot-andcannot-everdirectlyperceiveaconnectionbetweencausesandtheireffects.ItisnecessarytoexposethisflawsothatmyphenomenologicalaccountofthosecasesinwhichwebelieveweperceivesuchaconnectioncannotbedismissedsimplybecauseitfailstoconformwithHume'spositionregardingthenatureofsenseexperience.1 APPENDIXA:HUMEIndiscussingHume'sanalysisofcausation,itisimportanttodistinguishbetweentwodifferentproblemswithwhichhewasconcerned:(a)whetherwecanjustifiablyclaimthattherelationbetweenacauseandits·effectisanecessaryreIation,and(b)onwhatbasisweareledtoidentifyoneeventasthecauseofanother.Theseproblemsclearlyarenotthesame,buttheywerenotstrictlyseparatedinhisargument,sincehisanswertothesecondprovidedoneimportantargumentinfavorofhisanswertothefirst.ItisonlywiththesecondthatIshallherebeconcerned.Withrespecttothefirst,Iamentirelywillingtoacknowledgethatifweconstruetheterm"necessary"inthestrictsenseinwhichHumeconstruedit,wehavenogroundsonwhichtoestablishtheclaimthattheconnectionbetweenacauseanditseffectisanecessaryconnection:Weretheworlddifferentfromwhatitis,eventhemostreliableofourpresentcausalattributionsmightturnouttobeincorrect.Thisdoesnot6fcourseentailthatiftheterm"necessary"weretakeninsomeothersense,causalrelationsmightnotbe"necessary"relations.Itisnot,however,withHume'spositionregardingcausalnecessitythatIamhereconcerned;rather,Iaddressmyselftothequestionofhow,accordingtoHume,wecometoidentifyoneeventasthecauseofanother.WhatIwishtoshowisthatthepositionheadoptedrestsonanassumptionthatisfactuallyfalse,andcanbethoughttobetrueonlyifonedisregardshisowndistinctionbetween"impressions"and"ideas."Abasicaxiomonwhichhisanalysisofcausationrestsishisoftenrepeatedstatementthatwhatisdistinguishableisseparable.Throughouthisepistemologicalanalysesheappliedthisaxiomtosimpleimpressions,whichhetooktobetheultimatebuildingblocksofallknowledgeofmattersoffact;thus,heheldthatanyimpressionsthataredistinguishableareinprincipleseparable.Hethenusedthisaxiomtoshowthatwecanneverdirectlyexperiencetheconnectionbetweencauseandeffect:Sinceourimpressionsofwhatconstitutethecausearedistinguishablefromourimpressionsofwhatconstitutetheeffect,eachisthereforeseparablefromtheotherandanyconnectivitybetweenthemdisappears.Humethereforeofferedapsychologicalaccountofourconvictionthatwedoexperienceaconnectionbetweenacauseanditseffect.Hedidsobyappealingtotheinfluencethatconstantconjunctionsexerciseuponourminds.Whilehispositiveaccountofourbeliefincausalrelationsisinmanycasesplausible,hisbasicaxiomconcerningthedistinguishableandtheseparableledhimtosupposethatthesameaccountwouldbeapplicableinall APPENDIXA:HUMEcaseswhatsoever.ItisthisthatIwishtochallenge,andIshalldosobycriticizingthemannerinwhichHumeputhisaxiomtowork.Inexamininghisdoctrinethatallsimpleimpressionsaredistinctandseparable,wemustkeepinmindthatwhatisatissueishowthisdoctrineappliestoimpressions(thatis,todirectsenseexperience),andnothowitappliestoideas.WhatevermaybethedifficultiesinHume'saccountofthedifferencebetweenimpressionsandideas(andIbelievethemtobemany),2whenheisanalyzingourbeliefsinthereIationofcauseandeffectheisprimarilyconcernedwithourbeliefsastohowobjectsthatareexperiencedascontiguousinspaceandtimeareconnected.Tobesure,onsomeoccasions-aswhenwereceivealetterfromafriend-wetakeitthattherewasacausalconnectionbetweensomedirectlyexperiencedobject(theletter)andanotherobjectnotnowimmediatelyexperienced(myfriend,whoisabroad);butweformsuchaconnectionbetweenimpressionsandideasonlybecauseofpriorexperiencesinwhichwhatwenowrecallwasactuallypresenttousalongwithanobjectofthesametypeastheonewenowholdinourhand.3Thus,anyconceptionofacausalconnectionmustinallcasesbeultimatelybasedonthespatialandtemporalrelationshipsofdirectlyexperiencedimpressions,notonarelationshipamongourideas.Attentioncanbedrawntothispointinanotherway.Thereaderneedmerelyrecallthat,accordingtoHume,spatialcontiguityaswellastemporalsuccessionisinvolvedintherelationofcauseandeffect,butwhileitisclearthatweknowwhatitmeanstosayoftwoobjectsthattheyarecontiguousinspace,itmakesnosensetosayoftwosuccessiveideas(say,oftwomemoryimages)thattheyarespatiallycontiguous,exceptinthederivativesensethattheobjectsofwhichtheyaretheimageswerethemselvesoriginallyexperiencedascontiguous.Therefore,Hume'sanalysisofthecausalrelationmustbeinterpretedasapplyingtoimpressionsastheyaredirectlygiven,andnottothosesimulacraofouroriginalimpressionsthatareourideas.ThispointisoffundamentalimportanceinevaluatingHume'suseoftheaxiomthat"whateverisdistinguishableisseparable"asitrelatestocausation,sinceifheistoshowthatwecan,havenoimpressionofanycausalconnectionhemustapplythisaxiomtoimpressions,andnottoideas.Yet,canitbesoapplied?If,forexample,Iaminspectingacream-coloredrectangle-say,asmallsampleofwallpaper-Icandistinguishtheshapefromthecolor,butasimpressionstheshapeandthecolorofasmallpatcharenotseparable.InordertoseparatethemImustpictureeachofthemin APPENDIXA:HUMEmymind'seye-thiscolormerelyascolor,thisshapemerelyasshape;butindoingso,Iamtransformingthemfromdirectlyexperiencedimpressionsintoideas.Thisistruenotonlywithrespecttothesimultaneouslygivenqualitiesofsimpleobjects,butalsowheresuccessionisinvolved.Forexample,inMichotte'sexperimentsontheperceptionofcausation,4ifIseeoneformmovingslowlytowardanother,andIseethemincontactandthencontinuinginmotiontogether,Icanimaginethemotionofthefirstwithoutthatofthesecond,andIcanimaginethemotionofthesecondwithoutthat'Ofthefirst,butIdonotseethis.ThemovementsIsawhaddistinguishableaspects,forIdidseetwoobjects,butIdidnotatthetimeseetheseobjectsashavingseparateandindependentmotions.Inotherwords,itisfalsetoassumewithrespecttoperceptionthatwhatcanbedistinguishedisseparable,solongaswearespeakingofwhatisgiven,andarenotspeakingofhowwhathasbeengivenmightbeanalyzedafterhavingbeentransformedfromasetofimpressionsintoasetofideas.Infact,Hume'satomisticsensationalismdoesnotrestonaconsiderationofthegiven,butonananalysisofhowwemaydecomposewhatisgivenwhenwereflectivelyconsideritnotasasetofimpressionsbutasasetofideas.Forthisreason,hisanalysisofourcausalbeliefsholdsonlyinthosecasesinwhichthecasualconnectionswefindinexperiencedependuponobservedregularitiesofsequence,andhasnothingtosayconcerningothercasessuchasthosediscussedinsection1ofchapter3. AppendixBHARTANDHONORETherearethreecountsonwhichIshallexpressmydisagreementwiththepositiondevelopedbyHartandHonoreintheopeningchaptersofCausationintheLaw.1First,Ifindthatthewayinwhichtheyhavesetupacontrastbetweenaninterestintheparticularandaninterestinthegeneralisseriouslymisleadingwithrespecttobothhistoricalknowledgeandscientificexplanation.Second,theirviewswithrespecttocommon-sensecausalexplanationsarealsomisleading.Third,theyaremistakeninassumingthatthenotionofcausationtheyattributetolawyersisalsoapplicabletotheexplanationsgivenbyhistorians.MyargumentwithrespecttoeachofthesepointswillraiseanumberofseparateissueswithrespecttowhichIfindtheviewsofHartandHonoreunsatisfactory.ItrustthatinattemptingtocompressmydiscussionoftheirviewsintothisbriefappendixIshallnotbeguiltyofmisinterpretingthem. ,200APPENDIXB:HARTANDHONORE(1)AsImadeclearattheoutset,Iacceptthewell-knowndistinctionbetweenthenomotheticinterestsofscientistsandtheidiographicinterestsofhistorians,butthisisnottosaythatIaccept(a)thecontrastHartandHonoredrawbetweenthehistorian'sinterestintheparticularandthescientist's'interestinthegeneral,or(b)theirassunlptionsconcerningthetypeofexplanationthatischaracteristicofthesciences.(a)HartandHonoreopentheirdiscussionoftheparticularandthegeneralinsaying:"Thelawyerandthehistorianarebothpril1larilyconcernedtomakecausalstatementsaboutparticulars,toestablishthatonsomeparticularoccasionsonleparticularocc·urrencewastheeffectorconsequenceofsomeotherparticularoccurrence.Thecausalstatementscharacteristicofthesedisciplinesareoftheform'Thisman'sdeathonthisdatewascausedbythisblow.'''2This,however,isanimpoverishedviewofthehistorian'sinterestintheparticular:Hisaimisnotmerelytoconnecttwoeventsinlinear,causalsequence,buttodiscoverandtodepict,inconcretedetail,awholesetofeventsconcerningagivensubjectmatter,andtotraceavarietyofconnectionsamongthem.Thus,intheexampleusedbyHartandHonore,thehistorianwouldbeinterestedinhowitwasthatthesemencametomeetwhentheydid,whattheirpreviousrelationshadbeen,andwhat,inthisparticularsituation,hadledonetostriketheother.Anidiographicinterestisnotmerelyaninterestinsomeparticularevent,butinunderstandinganddepictingthatparticulareventinitscontext.NorareHartandHonoremorefortunateincharacterizingtheinterestsofscientists.Inthefirstplace,theyfailtopointoutthatifscientistswere�otoriginallyinterestedinexplainingparticulareventstheywouldnotbeledtoformulategeneralizationsinordertoexplaineventsofthattyp�.Inthesecondplace,itisonlythroughanalyzingthespecificconditionsunderwhichparticulareventsofagiventypeoccurthattheycanconfirmtheirgeneralizations.Forthesereasons,thecontrastdrawnbyHartandHonoreismisleading:Itisnotthatscientistsfailtobeinterestedinparticulareven�sthatsetsthemapartfromhistorians;rather,itisaquestionofhowhistoriansareconcernedwiththeparticular,andinwhatwaysparticulareventsareofinteresttoscientists.Thisdifferenceisbroughtoutbythecontrastbetweenidiographicandnomotheticinterests;itisnot,however,madeclearbythesimpledichotomyofaninterestintheparticularversusaninterestinthegeneral.(b)Anevenmorefundamentaldifficultyarisesinconnectionwith ,APPENDIXB:HARTANDHONORE201theviewofHartandHonoreconcerningthenatureofscientificexplanation.Itistheirviewthatinthesciencesacausalexplanationtakestheformofexemplifying"somegeneralizationassertingthatkindsorclassesofeventsareinvariablyconnected."3Inshort,insofarasthesciencesareconcerned,HartandHonorearecontenttoacceptsomeformoftheHumeanregularityview.Yet,ifwetakeanyadvancedscienceasamodel,wefindthatthelawsthatareIllostcharacteristicofscientificexplanationsareformulatedintermsofspecificfactors,suchaspressureandvolume,ormassanddistance,whichserveasvariablesinawidevarietyofcasesofverydifferenttypes;theyarenotstatementstotheeffectthattheoccurrenceofsomespecifictypeofeventisinvariablyconnectedwiththeoccurrenceofsomeotherspecifictypeofevent,asHartandHonore,followingHume,wouldhaveusbelieve.Tobesure,somegeneralizationsorlawsillustrateHume'sviewofcausationinthattheyexplicitlyrefertoinstancesofagiventype,andtotheconditionsunderwhichsuchinstancesoccur.Forexample,"litmuspaperturnsredwhenimmersedinacid,"or"atsealevel,waterwillboilat2120Fahrenheit."Nevertheless,thegeneralizationsorlawsthathavethisformareusuallyregardedascallingforfurtherexplanation,andsuchexplanationsareusuallycouchedintermsofthespecificfactorsintowhichtheseoccurrencescanbeanalyzed.Thefactorsthataretakenasprovidingthemoreadequateexplanationsarenotconfinedtoobjectsoreventsofaparticulartype,butarepresentinavarietyofinstancesthatdiffermarkedlyintheirothercharacteristics.Infact,generalizationssuchasthoseaboutlitmuspaperturningredorwaterboilingwhenheatedtoacertaindegreearenotatalltypicaloftheadvancedsciences;theyareinsomerespectsmoresimilartocommonsenseexplanations,suchas"icewillmeltatordinaryroomtemperature,"or"porcelaincupsbreakiftheyaredropped."Theydo,however,differfromthesecommon-sensegeneralizationsinoneimportantrespect:Theyareassumednottoadmitofexceptions,whereasmanycommon-senseexplanationsareregardedasadequateeventhoughtheydonotholdinabsolutelyallcases,butonlystatewhatusually,ornormally,occurs.HartandHonoreplacegreatemphasisonthisparticulardifferencebetweencommon-senseexplanationsandscientificgeneralizations,andthatisasitshouldbe.Theyfail,however,tochallengeHume'sviewonotherpoints,exceptinsofarastheyareconcerned(astheyare)withcasesofinterpersonaltransactions.Inshort,theyarewillingtoacceptHume'sviewofscientificexplanationasbeingcon- ,202APPENDIXB:HARTANDHONOREcernedwiththerelationsbetweenspecifictypesofevents;itisthisthatledthemtoholdthatthesciencesareinterestedonlyinthegeneral,andnotintheparticular.Furthermore,sincetheydonotchallengeHume'sclaimthatonecannotinanycaseperceiveaconnectionbetweencauseandeffect,theyhavetoholdthatcausalattributionsinallcasespresupposesomebackgroundofobservedregularitiesastowhatfollowswhat.Therefore,ontheirview,themaindifferencebetweencommon-senseexplanationsandscientificexplanationsresidesinthefactthatinourcommon-senseexplanationswefocusonwhatwasoddaboutaparticularcasethatdidnotbehaveasonewouldhaveexpectedittobehave;wethensettleonwhateverabnormalfeaturewaspresentinthissituation,designatingitasthecauseofwhatoccurred.Whilemanycausalattributionsineverydaylifedoconformtothismodel,notalldoso.ThisleadsmetomysecondlineofcriticismoftheviewsofHartandHonore.(2)Therearetworeasonswhytheiranalysisofourcommon-senseviewofcausationisnotinallcasesadequate.First,asIattemptedtoshowinchapter3,therearemanycasesinwh�chwedodirectlyexperienceacausalrelationshipthatisnottobeaccountedforintermsofpastexperience,yetHartandHonore'suseofthedistinctionbetweenwhatisnormalandwhatisabnormalpresupposesthatallcausalattributionsrestonourknowledgeofwhathasoccurredinothercases.Inthesecondplace,theiranalysisdoesnottakeintoaccountthosecasesinwhichwedosaythatoneeventcausedanotherwhenthereiscompleteregularityofsuccessionbetweentheseevents,withoutanydeviationfromwhatwetakeasnormalandwhatweexpected.Inthethirdplace,throughouttheirdiscussionofcausation,HartandHonorepresupposethatitisinallcasesvalidtodistinguishbetweenthecauseofaneventandwhateverotherconditionswerenecessaryforthatcausetohavehadtheeffectthatitdid.Whilewesometimesdodrawadistinctionbetween"cause"and"conditions/'andwhilesuchadistinctionmayperhapsbeoftheutmostimportanceforaffixingresponsibilityinthelaw,oneshouldnotlightlyassumethatthesamedistinctioncanbedrawninallothercontexts.AsIrepeatedlytrytoshowinchapter4,itisactuallyfalsetoholdthatwedrawthisdistinctioninallcasesofcausalexplanationthatariseeitherinscienceorineverydaylife.Thisisespeciallytrueinhistory.(3)IdonotbelieveitunfairtosaythatalthoughHartandHonorerepeatedlylinkthehistorian'sconceptionofcausationwithcausationinthelaw,anycarefulexaminationofwhatinterestshistorians,and ,APPENDIXB:HARTANDHONORE2°3ofwhattheyactuallydo,issingularlylacking.Intheonlypassageinwhichthistopicisexplicitlydiscussed(pp.58-59)oneisledtobelievethatwhenhistoriansusecausalnotionsinanexplanatorywaytheirpurposeisconfinedtogivinganexplanationof"somepuzzlingorunusualoccurrence."Yet,historicalexplanationissurelynotthusconfined.WhatisofgreaterinteresttoHartandHonore'thancausalexplanationinhistoryorthelawistheattributiveuseofcausalnotions,thatis,thefixingofresponsibilityforwhatoccurred.Itiswiththisuseofthecausalconceptthatthepassageinquestionisprimarilyconcerned.Now,itisassuredlytruethathistorians,likepractitionersinthelaw,domaketheseattributivecausaljudgments,assigningresponsibilitytoindividuals,ortoparticularcircumstances,forhavingbroughtabout(thatis,"caused")certainevents.Forexample,inAmericansInterpretTheirCivilWar,4ThomasJ.Presslysurveyedoneparticularsetofjudgments,namelythosethatshowedchangingopinionsastowhatfactor,ortypeoffactor,wasmostbasicinleadingtotheCivilWar.However,severalpointsshouldbenotedwithrespecttothisexample,fortheyareofwiderapplicability.5Inthefirstplace,Pressly'saccountisnotanaccountoftheCivilWaritself,norisitadetailedstudyofthedifferingaccountsofthebackgroundandtheoutbreakofthewarthateachoftheauthorswhomhestudiedactuallygave.Rather,itdiscussestheseaccountsfromasinglepointofview,abstractingfromeachwhatitsauthortooktobethemostbasicfactorleadingtothewar.Thus,quitelegitimately,Presslywasconcernedonlywiththemostgeneralattributivecausaljudgmentstobefoundinanumberofrepresentativewriters;itwasnothisaimtoassessthedetailedtreatmentsthesewritershadgivenofalloftheotherfactorsthatmighthavebeeninvolvedintheoutbreakofthewar.However,ifahistorianistobeinapositiontomakegeneralattributivejudgmentsconcerningwhatwas"mostbasic"withrespecttotheCivilWar,hemustfirstknow-orbelievethatheknows-thenatureofagreatmanyfactsconcerningslavery,theabolitionmovement,theaimsandactionsofvariouspoliticalfigures,thestateoffeelingindifferentpartsofthecountryatdifferenttimes,andthelike.Ifheistobeconsideredareputablehistorian,itwillbeonthebasisofhisreadingofallsuchfactsthathewillhaveputforwardhisattributivecausaljudgment.Ingatheringthesefactsandintracingtheirconcatenations,thehistorianisnotengagedinafurtherseriesofattributivecausaljudgments,norisheengagedinexplanatorycausaljudgmentsifbythisismeantthatwhatheisseekingtoaccountforarepuzzlingorunusualoccurrences:Heisengagedinattemptingto ,APPENDIXB:HARTANDHONOREdiscoverwhathappened.Similarly,beforeacasegoestotrial,inquiryintomattersoffactmustbemadebythelawyers;anditisonlyaftersuchinquirieshavebeenmadethatonecanargueincourtwhereresponsibilitylies.6Onthisanalogy,itmaybesaidthatalthoughitisnotuncon1monforhistorianstobeconcernedwithattributivecausaljudgmentssimilartothoseHartandI-Ionorediscusswithreferencetothelaw,thiscannotbethefirstorprimarytaskofthehistorian.Furthermore,itisataskthatcanbeavoidedbyhistorians,eventhoughitisunavoidableforjudges.Historiansneednotsitasjudges.Itisthereforemycontentionthat,contrarytotheviewofHartandHonore�thereisnocloseanalogybetweenwhathistoriansdo,andwhatisdonebylawyersarguingincourtorbyjudgeswhentheydecideacase.NonethelessIadmitthatitisunfortunatelytruethatlaymen,beingneitherhistoriansnorlawyers,oftenthinkitwithintheirpowerstomakeattributivecausaljudgmentsconcerningwhathasoccurredinhistorywhentheyhavenotfirstinvestigatedthefacts.ForallthesereasonsIfinditnecessarytodisagreewiththeenormouslyinfluentialanalysisofcausationweowetoHartandHonore. NOTES�w��CHAPTERONE1.WilhelmWindelband,Praludien(5thed.;'fiibingen:J.C.B.Mohr�1915),2:144-45.2.TheconflationofthesedistinctquestionsisapparentinthefirsttwosentencesofHempel'spaper,whichreadasfollows:"Itisaratherwidelyheldopinionthathistory,incontradistinctiontotheso-calledphysicalsciences,isconcernedwiththedescriptionofparticulareventsofthepastratherthanwiththesearchforgenerallawswhichmightgoverntheseevents.Asacharacterizationofthetypeofprobleminwhichsomehistoriansaremainlyinterested,thisviewprobablycannotbedenied;asastatementofthetheoreticalfunctionofgenerallawsinscientifichistoricalresearch,itiscertainlyunacceptable."ForHempel'scriticismofmyclaimthattheprimaryaimofthehistorianlies"notintheformulationoflawsofwhichtheparticularcaseisaninstance,butinthedescriptionoftheeventsintheiractualdeterminingrelationshipstoeachother,"seethefootnotetosection7.4ofhispaper,inwhichthetwoissuesareagainidentified.205 206NOTESTOPAGES7-11Hempel'spaperwasoriginallypublishedintheJournalofPhilosophyin1942;itisreprintedinhisAspectsofScientificExplanation(NewYork:FreePress,1965).3.Forageneraldiscussionoftheproblemofexplainingculturaltraitsintermsofdiffusionorintermsofindependentorigins,aswellasforadiscussionofthephenomenonofconvergenceamongculturaltraits,thereadercanconsultA.A.Goldenweiser,Anthropology(NewYork:Crofts,1937),chaps.28,29.Ontheotherhand,thosewhowishaconcreteexampleinwhichrelativelysophisticatedgeneralizationsplayapartintheactualworkofahistoriancanconsultH.R.Trevor-Roper'sessay"TheEuropeanWitch-crazeintheSixteenthandSeventeenthCenturies,"whichwasoriginallypublishedinhisReligion,theReformationandSocialChange(London:Macmillan,1967).Whetherornotoneagreeswithhisinterpretation,Trevor-Roper'sinterestwasthatofahistorianbuthismethodincludedtheundisguiseduseofanumberofsociologicalandpsychologicalhypotheses.4.IhereparaphrasethepositionofCharlesA.Beardin"WrittenHistoryasanActofFaith";asimilarpositionwasadoptedbyCarlL.Beckerin"WhatAreHistoricalFacts,"andelsewhere.TheseessaysarereadilyavailableinHansMeyerhoff,ed.,ThePhilosophyofHistoryinOurTime(GardenCity,N.Y.:DoubledayAnchor,1959).(Fortherelevantpassages,seepp.140and124,respectively.)ThesamepointofviewhadearlierbeenespousedbyJamesHarveyRobinsoninhisessay,"TheNewHistory"(1912),wherehesaid:"InitsamplestmeaningHistoryincludeseverytraceandvestigeofeverythingthatmanhasdoneorthoughtsincefirstheappearedonearth"(apudFritzStern,TheVarietiesofHistory[NewYork:Meridian,1956],p.258).R.G.Collingwoodadoptedaparallelposition,definingthatwhichservesastheobjectofhistoryas"resgestae:actionsofhumanbeingsthathavebeendoneinthepast"(TheIdeaofHistory[Oxford:ClarendonPress,1948],p.9).Solongasanyaction,nomatterhowtrivial,canbetheobjectofthekindofinquirycarriedonbyahistorian,Collingwoodheldthatitisnotbeexcludedfromthedomainofhistory.ThisisclearinhisPhilosophyofHistory(HistoricalAssociationLeafletno.79,1930),wherehesaidthatthequestionofwhoplayedcenterforwardonavillagesoccerteamisasmuchahistoricalquestionaswhowonthebattleofCannae.5.TheProblemofHistoricalKnowledge(NewYork:Liveright,1938),p.9·6."TheHistoryofIdeas,IntellectualHistory,andtheHistoryofPhiloso-phy,"HistoryandTheory,Beiheft5(1965):42-47.Inasubsequentstudy,Ihavediscussedsomeoftheproblemsconcerningonetypeofspecialhistory,thehistoryofphilosophy,ingreaterdetail("OntheHistoriographyofPhilosophy,"PhilosophyResearchArchives[PhilosophyDocumentationCenter,BowlingGreenStateUniversity,BowlingGreen,Ohio],July1976).Crocealsousestheterms"specialhistories"and"generalhistory"(History,ItsTheoryandPractice[NewYork:Harcourt,Brace,1923],pt.1,chap.8,andapp.2).Whiletherearepointsofcontactbetweenhisusageandmine,hisconclusionsareentirelydifferentfromthoseIdraw.7.Forexample,Herskovitssays,"Acultureisthewayoflifeofapeople;whileasocietyistheorganizedaggregateofindividualswhofonowagivenwayoflife.Instillsimplerterms,asocietyiscomposedofpeople;thewaytheybehaveistheirculture"(ManandHisWorks[NewYork:Knopf,1948],p.29).Cf.Kluckhohn,"TheConceptofCulture"(1945),reprintedinhisessaysCultureandBehavior[NewYork:FreePressofGlencoe,1964],inwhichnoteespeciallyp.21.8.Forexample,CliffordGeertzdistinguishescultureasasystemofideasfromtheeconomic,political,andsocialrelationsinasociety,allofwhichareto NOTESTOPAGES11-14somedegreeinformedbythatsystemofideas(TheInterpretationofCultures[NewYork:BasicBooks,1973],p.362.)Suchasystemofideasischaracterizedbyhimas"anhistoricallytransmittedpatternofmeaningsembodiedinsymbols,asystemofinheritedconceptionsexpressedinsymbolicformsbymeansofwhichmencommunicate,perpetuate,anddeveloptheirknowledgeaboutandattitudestowardlife"(ibid.,p.89).Ifindthistypeofdefinitiontoorestrictive,sinceIfinditessential,asdomanyanthropologists,toviewlanguage,tools,andcraftsasaspectsofculture,nolessthanare,say,religiousbeliefs.KroeberandParsons,inanattempttoclarifytherelationshipbetweentheterms"culture"and"society"(or"socialsystenl"),alsotendedtoidentifyculturewithvaluesandideas,thoughtheyaddedtotheirdefinitionareferencetoartifacts("TheConceptsofCultureandofSocialSystem,"AmericanSociologicalReview28(1958):582-83).Itistobenoted,however,thatthisrepresentsamajordeparturefromKroeber'searlierviewasstatedinsections6and117ofhisAnthropology,newed.,rev.(NewYork:Harcourt,Brace,1948).9.IntheopeningparagraphofhisPrimitiveCulture(London:J.Murray,1871),Tylorusedtheterm"culture"toreferto"allcapabilitiesandhabitsacquiredbymanasamemberofsociety"andheincludedinhisdiscussionallartifactsresultingfromthosecapabilitiesandhabits.(Cf.chap.7,"GrowthandDeclineofCulture"inhisResearchesintotheEarlyHistoryofMankind[London:J.Murray,1865].)Partlyasaresultofhisevolutionaryviewsregardingsocieties,hespoke·of"culture"andnotofdiverse,particular"cultures."Whilenotsharinghisviewsonsocialevolution,Itooshallusetheterminagenericsenseandshalldefineitasincludingartifactsaswellasideasandvalues.Ineachoftheserespects,myviewissimilartothattobefoundinKroeber'sAnthropology.10.Iquotethereleventpassagesfrom"DasFach'Geschichte'unddiehistorischenWissenschaften,"HamburgerUniversitatsredenno.25(1959):GeschichteimengerenSinn,allgemeineGeschichtehateswedermitderPolitik,denpolitischenAblaufenurnihrerselbstwillen,nochmitderKulturasezutun,sondernmitderenTrager,demMenschen,sowohlmitdemeinzelnenMenschen,derunsstetsingesellschaftlicherTerbundenheitentgegentritt,wiemitmenschlichenGruppen.EineGeschichteimengerenSinnhabendahernurMenschenundmenschlicheVerbande,Familien,Dorfer,Stadte,Stande,Klassen,Staaten,Volker,Stammeusw.MenschenundmenschlicheVerbanderingenurnihreExistenz,behauptensichselbst;siehandelnindiesemSinne"politisch."Es·sindSozialgebilde,indenenHerrschaftsverhaltnissebestehen,rechtlichgeordneteMachtverhaltnisse.[po23]Incontrast,Brunnercharacterizedthespecialformsofhistoricalstudyinthefollowingway:Dasprimare,zentraleObjektderhistorischenFachwissenschaftenistebennichtderMenschunddiemenschlichenGruppen,sonderndessenWerke.HierwerdenInstitutionen,Rechts-undWirtschaftsordnungen,religioseundphilosophischeLehrmeinungen,WerkederKunstundderLiteraturen,dieSprachenundvielesandere,zuersteinmalabgehobenvonihrenTragern,alsSinngebildeuntersucht,interpretiertunddargestellt.[pp.25-26]11.InH.P.Finberg,ed.ApproachestoHistory(London:Routledge&KeganPaul,1962),pp.41-46.12.Biographiesconstituteaspecialform·ofhistoricalaccount,inwhichtheinterestisfocusedonthepersonwhoisthesubjectofthebiography.Bio- 208NOTESTOPAGES15-25graphicalstudiesdonot,inmostinstances,provideanexceptiontowhatIhaveemphasized,sincethepersonschosenforbiographicalstudyare,ingeneral,personsinwhomthereisinterestbecauseoftheirrolesinaparticularsocietyorbecauseoftheirrelationtosomeaspectofitsculture.Insomecases,however,apersonmaybethesubjectofabiographicalstudysimplybecauseofhischaracter,ratherthanbecauseofhisrelationtohissocietyortothecultureofhistime.Suchstudiescancontributeindirectlytoanunderstandingofthepast,buttheirprimaryfunctionislikelytobeeitherpsychologicalor(takingtheterminitsliteralandnoninvidioussense)hagiographic.ApointofvievdiametricallyopposedtothatadoptedhereistobefoundinFrederickA.Olafson'sarticle"HumanActionandHistoricalExplanation,"inNewEssaysinPhenomenology(ed.JamesEdie[Chicago:QuadrangleBooks,1969]).Olafsonregardsthesubjectmatterofallhistoricalaccountsasbeinghumanactions,andfindsthestructureoftheseaccountstobedictatedbythemannerinwhichweanalyzesuchactions.AsimilarpositionisimplicitinG.H.vonWright'sExplanationandUnderstanding(Ithaca:CornellUniversityPress,1971),chap.1and4(especiallypp.137-39).InassessingvonWright'sposition,itistobenotedthatheconfinescausalexplanationtodiscoveringconnectionsbetweengenericcharacteristicsofeventsorstatesofaffairs,ratherthaninananalysisofthesufficientconditionsofindividualoccurrences(cf.pp.38,43,and74).13.TheProblemofHistoricalKnowledge,pp.255-57and266-69.AsItherepointedout,IhadborrowedthisconceptfromaBerlindoctoraldissertationwrittenbyK.Milanov.IhaverecentlyfoundadiscussionofthesameconceptinSiegfriedKracauer,History:TheLastThingsbeforetheLast(NewYork:OxfordUniversityPress,1969),chap.5.14.In"CentralSubjectsandHistoricalNarratives,"HistoryandTheory14(1975):253-74,DavidL.Hullholdsthatoneshouldnotregardhistoricalnarrativesasbeingconcernedwithaseriesofconnectedeventsbutwithhistoricalentitieswhichhaveunityandcontinuityovertime.Insofarasgeneralhistoryisconcerned,Ifindmyselfinagreementwithmuchthathesays.However,Idonotfindhimconvincingwhenheholdsthatthisisalsotrueofallhistories.Infact,hisbasiccontentionthatbiologicalspeciesaretobeconsideredasunitarycontinuingentitiesseemstomeopentodoubt.15.Forastimulatingdiscussionofperiodizationinrelationtochronologicaltime,seeKracauer,Historychap.6.CHiPTERTWO1."ANoteonHistoryofNarrative,"HistoryandTheory6(1967):413-19.ThatarticlewascriticallycommenteduponbyRichardG.Ely,byRolfGruner,andbyW.H.DrayinHistoryandTheory8(1969):275-94.PerhapsthemostsympatheticassessmentofthenarrativistviewisDray's"OntheNatureandRoleofNarrativeinHistoriography,"inHistoryandTheory10(1971):153-71.NeitherthesearticlesnorotherformulationsofanarrativistpositionbyA.R.LouchinHistoryandTheory8(1969):54-70,andbyHaskellFaininBetweenPhilosophyandHistory(Princeton:PrincetonUniversityPress,1970),haveledmetoaltertheargumentsIused.IfindsupportforsomeofmyargumentsinanarticlebyC.B.McCullaghinMind78(1969):256-61.Inthebalanceofthisdiscussion,mycriticismsofthenarrativistpositionareretained,butthewiderfran1eworkwithinwhichtheyareplacedmayperhapsmakethemsomewhatmoreacceptabletomycritics.2.Tobesure,ifanauthorwritesahistoricalnovel,orseekshistorical NOTESTOPAGES26-34materialasbackgroundforhisfiction,hewillhavetoengageinpreliminaryresearch,andwhathewriteswillhavetobecompatiblewithwhathehaslearned.However,hisresearchwillnotdictatethestructureofthestoryhehaschosentotell;thereinliesanessentialdifferencebetweenhistoryandhistoricalfiction.3.ThispointisalsomadebyCordonLeffinHistoryandSocialTheory(University:UniversityofAlabamaPress,1969),whenhesays:"Thehistorianmethodologicallyworksbackfromtheproblemwhichhehasalreadyidentifiedtothecircumstanceswhichleduptoit"(p.96)and,again,"Historicalexplanationsbeginfromwhatistohowitbecamesuch"(p.100).ThepointisalsomadebyPaulK.Conkinin"CausationRevisited,"HistoryandTheory13(1974):4.Hesaysofthestrengthoftheimplicativerelationshipincausalattributions:"Allthestrengthisretrospective,butsoisthevisionofthehistorian.Inhiscausalanalysis,ifnotinhisnarration,heusuallymovesbackfromasignificantbutsomehowpuzzlingevent(theeffect)tosearchoutsomeofitscauses,oftenguidedbyadetaileddescriptionoftheevent."4.ThatYoung'saimwastointerpret,andnottorecountorexplain,canbeclearlyseeninaremarkhemakesintheintroductiontothe1953edition·ofVictorianEngland:PortraitofanAge(London:OxfordUniversityPress,1953):"thereal,centralthemeofHistoryisnotwhathappened,butwhatpeoplefeltaboutitwhenitwashappening."Burckhardt'sinterestintreatingtheRenaissanceinItalyasasingle"geistigesKontinuum,"whichitwashispurposetorevealthroughtreatingitsvariousfacets,isclearintheveryfirstparagraphofhisbook.5.KingGeorgeIIIandthePoliticians(Oxford:ClarendonPress,1954),p.182.6.TheFoundationofHistoricalKnowledge(NewYork:Harper&Row,1966),pp.221-25·7.SeeExplanationbyDescription:AnEssayonHistoricalMethodology(TheHague:Mouton,1968).Newmansays:"Thethesisof[this]essaywillbethathistoriansfrequentlyexplainanactionbyofferingadescriptionofthataction"(p.12).In"TheAutonomyofHistoricalUnderstanding,"inHistoryandTheory5(1965):24-47,LouisO.Minkespousedasomewhatsimilarposition.HisessayhasbeenreprintedinW.H.Dray,ed.,PhilosophicalAnalysisandHistory(NewYork:Harper&Row,1966);pp.178-79and180-81ofthateditionareespeciallyrelevant.Itisalsotobenotedthatintheconcludingsectionofapaperentitled"OntheNatureandRoleofNarrativeinHistory,"DraydiscussesthefactthatsuchaviewisimplicitintheclaimsofCallieandDanto,andisexplicitinOakeshott;heexpresseshisownstrongsympathywithit(HistoryandTheory10(1971):169ff.).8.Ifirstarguedthispointin"HistoricalExplanation:TheProblemofCoveringLaws,"inHistoryandTheory1(1961):229--42,andtheninmybookHistory,Man,andReason(JohnsHopkinsPress,1971),pp.114-20..9.Forsomeexamplesofhoweconomictheorymaybeimportantinexplainingparticularhistoricalchanges,seeA.C.L.Shaw,"EconomicsandHistory,"HistoricalStudies.AustraliaandNewZealand3(1949):277-86.Foramoreradicalapproachtotheproblem,theexampleofRobertW.Fogelandhisfellow"cliometricians"istobenoted.10.Asuggestiveexampleofhowone'sdefinitionofliteratureinfluencestheorganizationofaliteraryhistoryistobefoundinchapter1,"Prospective,"ofOliverElton'sSurveyofEnglishLiterature,1730-1780(NewYork:Macmillan,1928).Particularlyrelevantistherolethatactualspeechplaysinhisdefinitionofliterature(pp.2-3). 210NOTESTOPAGES35-50AbriefandverygeneraldiscussionofthepointistobefoundinReneWellekandAustinWarren,TheoryofLiterature(NewYork:Harcourt,Brace,1949),pp.30-31.11.Inwhatfollows,Iamconfiningmyattentiontotheproblemofselectioninsofarasitconcernsformsofspecialhistory;Ishallnotconsiderthequestionofimportanceasitarisesinothercontexts.Forexample,thetermissometimesconnectedwiththeproblemofweighingcausalfactors;sometimesitisidentifiedwithwhatevercriteriaahistorianmayuseinselectingasubjectmatter,andwhateverprinciplesofinclusionorexclusionheappliesindealingwiththatsubject.Foradiscussionoftheformerproblem,seetheconcludingsectionofErnestNagel,"SomeIssuesintheLogicofHistoricalAnalysis,"whichappearedinScientificMonthlyin1952,andisreprintedinPatrickGardiner,ed.,TheoriesofHistory(Glencoe,Ill.:FreePress,1959);forasuggestivetreatmentofthesecondissue,seeW.H.Dray;"OnImportanceinHistory,"inMind,Science,andHistory,whichisvolume2ofContemporaryPhilosophicThought:TheInternationalPhilosophyYearConferencesatBrockport(Albany:StateUniversityofNewYorkPress,1970).12.Inchapter1,Icalledattentiontotheparallelbetween.thetaskofahistorianintracinginfluencesandthatoftheanthropologistinassessingthediffusionofculturalelements,andIshallnotlaborthepoint.Theanthropologist,ofcourse,suffersunderthedisadvantageofnothavingwrittenhistoryonwhichtorely,anditisthereforeespeciallydifficultforhimtodeterminewhatmayhavebeenthepointoforiginofaparticularculturalelementandinwhatdirectionitspread.Goldenweiser'sdiscussionoftheproblem,whichIcited,affordsausefulcontrasttothemannerin",ohichaliteraryhistorianisgenerallyabletoproceed.13.Thesamesituationobtainswithrespecttotheworksofanyindividualauthor:Theyarenottobeviewedasformingasinglelinearseriesinwhicheachworkgrewonlyoutoftheworkthatprecededit.Fromworktoworkanauthor'sexperiencewillhavechanged,hisinterestmayhavebeencaughtbywhatothersweredoing,orhemayhavefeltthathehadexhaustedtheveininwhichhehadbeenvorking.Theseandmanyotherfactorsmayhavetobetakenintoaccountinordertointerprettherelationshipbetweenanyindividualworkandtheauthor'slifeorhisworkasawhole.14.Whatobtainsgenerallywithrespecttointerpretivedisagreementsappliesequallytodebatesconcerningsuchissuesas"thecausesoftheCivilWar,"whicharenotinfactconflictsbetweendifferingexplanatoryhistoricalaccounts,butbetweenrivalinterpretiveaccounts(seeAppendixB).CHAPTERTHREE1.WhileBertrandRussellwasamongthemostinfluentialfiguresinstandardizingtheprevailingview,itwasalsoaviewsharedbyallpositivists.Inaddition,itwasacceptedbyanumberofantipositivistAmericanphilosophers,suchasJ.B.PrattandMorrisR.Cohen.Amongthosewhorejectedthisview,EmileMeyersonwasforatimethemostinfluential.Theviewwasalsorepeatedlychallengedbyidealists,someofVhOnl(e.g.,A.C.EwingandBrandBlanshard)claimedthatlogicalnecessityisinvolvedinthecausalrelationship.Amongotherformsofchallenge,mentionshouldbemadeofVhitehead'sattackontheHumeananalysisandC.J.Ducasse'ssystematicattempttodefendanalternativepositioninCausationandOtherTypesofNecessity,UniversityofWashingtonPublicationsintheSocialSciences,vol.1,no.2(Seattle,1924).However,noneofthesesucceededinleadingtoanybasicrevisionofthedominantview,thoughDucasse'spositionhasrecentlybecomebetterknownandmorewidelydiscussedsinceitsreformu- NOTESTOPAGES50-53211lationinhisNature,Mind,andDeath(LaSalle,Ill.:OpenCourt,1951)andinhisTruth,Knowledge,andCausation(London:Routledge&KeganPaul,1968).2.CausationintheLaw(Oxford:ClarendonPress,1959).3.TheCementoftheUniverse:AStudyofCausation(Oxford:ClarendonPress,1974).4.ItistobenotedthatbothJ.L.MackieandK.Marc-Wogaualsoinsistthatoneshoulddistinguishbetweenthecausalexplanationofaparticularcaseandgeneralizationsconcerningcasesofaparticulartype,butdonotfindthatthedifferencesbetweenexplanationsinthesciencesandelsewhereinvolvedifferentconceptionsofthecausalrelation.SeeMackie,CementoftheUniverse,pp.121and270-71forsummarystatements,andMarc-Wogau,"HistoricalExplanation,"Theoria28(1962):214and215-16.Itisalsotobenotedthatattemptstodrawasharpcontrastbetweenhistoryandthesciencesismistakeninstillanotherrespect:Insomenaturalsciences,suchasgeology,inquiriesintotemporalsequencesinthepastareofgreattheoreticalimportance,providinganindispensablecomplementtothosenontemporal,functionalgeneralizationswithwhichscientistsareprimarilyconcerned.T.A.Goudgehasmadethispointclearinanarticlethathasimportantconsequencesforunderstandinghistoricalexplanation:"CausalExplanationsinNaturalHistory,"BritishJournalforthePhilosophyofScience9(1958):194-402.5.Intherecentliteratureofoursubject,one-ofthefirsttopointthisoutwasMichaelScriven.Hemakesthepointespeciallyclearlyin"Causes,ConnectionsandConditionsinHistory,"firstpresentedataconferenceattheUniversityofCincinnati,andlaterpublishedinW.H.Dray,ed.,PhilosophicalAnalysisandHistory(NewYork:Harper&Row,1966).Seeespeciallypp.242-43.6.Toavoidmisunderstanding,IwishattheveryoutsettosaythatwhenIspeakofsomeone's"perceiving"acausalconnectionitispossiblethatthepersonismistaken,andthatsuchaconnectiondoesnotexist.Tofindexamplesofthisoneneednlerelyrecallone'sexperienceinwatchingaskillfulnlagician.7.Forexample,Iassumethatevenanuninstructedspectator,whodidnotknowtherulesoffootball,wouldseeadifferencebetweenthebehavioroftheplayersbetweendownsandtheirbehaviorduringaplay,recognizingthelatterasasingleandunifiedeventinawayinwhichtheintervalsbetweendowns,orinatime-outperiod,arenot.OneaspectofanexperimentperformedbyHeiderandSimmelmayberelevanthere.Theydesignedam-oviethathasasits"characters"twotriangles,differinginsize,andacircle.Thesethreefiguresmovedinandaroundarectanglethroughanaperturethatopenedandshut.Ingeneral,allsubjectsinterpretedthefiguresinhumanterms,andinterpretedtherectangleasahouse.Inthepresentcontext,whatisrelevantaboutthisexperimentisthat,withinthetemporallycontinuousfilm,particularsetsofmotionsofthefiguresbecamesegregatedasbelongingtogether,withthetotalflow-ofmotionsbeingseenneitherasadisconnectedsequencenorasasingleevent;rather,itwasseenasaseriesofepisodespunctuatingaflow-ofaction.ThisisnotapointbroughtoutbyHeiderandSimmel,butasanobserverofthefilmIwasparticularlystruckbyit,andmypresentdescription-ofthephenomenonconformstotheiraccountofwhattheirfirst(uninstructed)groupofsubjectsreported(FritzHeiderandMarianneSimmel,"AnExperimentalStudyofApparentBehavior,"AmericanJournalofPsychology57[1944]:243-59).InthisparticularpaperHeiderandSimmeldiscusstheperceptionofcausationinthefilm,givinganaccountverysimilartothatwhichhasbeenmorefullydevelopedthroughtheMichotteexperimentstowhichIshalllatercallattention.) 212NOTESTOPAGES55-618.InthiscaseImustremindthereaderthatwearehereconcernedwithcausalbeliefsineverydaylife;wearenotdealingwiththeexplanationsofthemotionsofthebilliardballsthataphysicistmightgive.Weshallcometosuchexplanationslater.Iamhereconfiningmyselftowhatisseenwhenwesaythatwesawthatabilliardballmovedbecauseitwashitbyanother.9.ImustrepeatthatIamnotclaimingthatallcasesinwhichthecauseeffectrelationshipisappliedtoeventsineverydaylifeareofthiskind.Asweshallsee,manyconformtotheregularitymodel.Iamnowconcernedonlywiththoseinwhich'webelievethatwedoactuallyseetheconnectionbetweencauseandeffect.Furthermore,ImustonceagainsaythatIamnotnowdiscussinghow,ifatall,necessitymaybeinvolvedincausalconnections.10.IneachoftheprecedingillustrationsIhavebeenconcernedwithcasesinwhichtheeffectistakenastheendpointinaprocess,butwecanaskcausalquestionsconcerningcontinuingprocesses:Whydoesthetopcontinuetospin,ortheenginecontinuetorun?Inthesecases,too,ourcausalattributionsrefertowhatisoccurringwithintheprocessdesignatedasthetop'sspinningortheengine'srunning;itisnotsomepriorevent.11.ThefollowinggeneralstatementmadebyA.Michotte,theBelgianpsychologist,concerninghisexperimentsontheperceptionofcausation,isrelevanthere:"ItistruethatbyananalyticalandabstractapproachsuchasIhaveadoptedabove,itispossible,theoretically,todistinguishtwosuccessiveevents,themovementandthecontact.Butactuallytherearenottwoevents;thereisonlyoneeventwhichdevelopsprogressively.Asweshallseelater,theimpactisnotreallylimitedtothecomingintocontactofthetwoobjects;itconstitutesawholeprocess,ofwhichthemovementandthecontactarebothconstitutiveparts....Thewholeisonegradualdevelopment"(ThePerceptionofCausality[London:Methuen,1963],pp.24-25).Giventhisstatementandtheforegoinganalysisofcasesinwhichwemaybesaidtoperceivecausalconnections,itshouldbeclearthatthesecasesbearagreaterresemblancetothewaysinwhichhistoriansareapttousetheconceptofcausationthanaHumeananalysisprovides.Thus,thepresentdiscussionpreparesthe"vayforwhatistobesaidinchapter5.12.SeeKarlF.Duncker,"OnProblem-Solving,"PsychologicalMonographs58,no.270(1945):67-68,orSolomonE.Asch,SocialPsychology(NewYork:Prentice-Hall,1952),pp.101-2.13.ItistobenotedthatHumehimselfusedtheresemblanceofsuccessiveimpressionstoexplainourbeliefinthecontinuingidentity·ofobjects.Iamhereusingasimilarfactortoexplaincontinuityinevents.14.Inaquitedifferentconnection,BertrandRussellspeaksofsimilarityofstructureasgivinguswarranttoinfercausalrelations.Althoughhisdefenseofthisinductivepostulatedependeduponconsiderationsofprobabilityandnotuponperceptualbelief,inbothcasestheinfluenceofourrecognitionofsimilarityinstructureisafactorthatleadsustoregarddifferenteventsasbelongingtogetherinaconnectedseries.(SeeRussell,HumanKnowledge,ItsScopeandLimits(NewYork:SimonandSchuster,1948),pt.6,chap.6;pp.460-64areparticularlyrelevant.)15."OnProblem-Solving,"p.67.16.Analoguestothese"explanations"aretobefoundinmanyhistoricalworks.Ifindthemunsatisfactory,bothinhistoryandineverydaylife,sincetheexplanationisnotcouchedintermsofactualevents,butintermsofwhatisonlyapatternrunningthroughthem.wpensuchapatternisparticularlystrikingitisapttobereified.Insteadofbeingviewedaswhatresultsfromtheseeventsintheirinterrelationships,itistakenasifitexplainedthem.Seechapter5forafurtherdiscussionofthis.17.Ifthereisnorhythmicpattern,andtheybearafixedspatialrelationshipto eachother,theywillbeseenmovingasapair,asiftheywererigidlyconnected.(Forexample,seeMichotte'sdiscussionofthedifferencebetweendisplacementandwhatweexperienceasmovement,inPerceptionofCausality,pp.315-16.)18.Thisquestionis,ofcourse,apsychologicalquestion.Inattemptingtosuggestapsychologicalexplanationoftheoriginsofourcausalbeliefs,IammerelyfollowingthepracticeofHume.19.Inhisexperimentson"qualitativecausation"(asdistinctfrom"mechanicalcausation")Michottereachedanessentiallynegativeconclusion:Therewasnoperceptionofcausationunlessaconnectionwasseenbetweenthemovementofoneobjectandthemovement(orchangeofshape)ofanother.Iamnotdenyingthisnegativeconclusion.Whenweswitchonalightwecannotclaimtoseethelinkagebetweenwhatoccursattheswitchandthefactthatthelightcomeson;itisonlypastexperiencethatbindsthesetwoeventsdirectlytogether.Thus,thisisnotthesortofexamplewithwhichMichotte'sexperinlentswereconcerned:Heattemptedtofindcasesinwhich,independentofpastexperience,therewasadirectperceptionofsomefactorlinkingcausetoeffect.20.Evenfromaphenomenologicalpointofview,thetwoeventsdonotappearasinstantaneous.Themotioninvolvedwhenweswitchonalightdoesnotappeartousasinstantaneous.Eventhelight'scomingondoesnot,underscrutiny,alwaysappearinstantaneous,asthosefamiliarwith"gammamovement"willrecognize.21.Anotherconditionthatleadsusfromonespecificcasetocasesofaparticulartypeiswhensomeoneasksustovalidateourbeliefthatacausalconnectiondidinfactobtaininsomeparticularinstance.IaiiliIotheredealingwiththeproblemofvalidation,butonlywithwhatisinitiallyinvolvedinourascriptionsofcausalrelationships.InthisconnectionImightpointoutthatintwopassagesinwhichC.G.Hempelcriticizedmyviewsregardingcausation,holdingthatthecausalrelationreferstotypesofcasesandnotindividualinstances,hisargumenthingedontheissueofvalidatingouroriginalcausalascriptions.(See"TheFunctionofGeneralLawsinHistory,"asreprintedinHempel,AspectsofScientificExplanation[NewYork:FreePress,1965],pp.233n.and241n.)22.Thisisnottosaythatlawyersandhistoriansordinarilyusethisformofexplanationindealingwiththeinstancesofcausationwithwhichtheyareconcerned.AsIshallnoteinAppendixB,HartandHonoremayhavetooreadilyassimilatedcommon-senseviewswiththemannerinwhichcausationisusedamonglawyers.Ishallalsoarguethattheyhaveerredinassumingthattheproceduresofhistoriansaresimilartoboth.23.Conversely,ofcourse,theonsetofastormcanserveasanindicatorthatchangesinbarometricrecordingshaveoccurred,butforpragmaticreasonswearerarelyinterestedintracingthisparticularrelationship.Barometersaregenerallyofinterestonlybecausetheyforewarnusofchangesintheweatherthatwedonotyetdirectlydetect,orbecausetheyhelpusevaluateothersignsthattheweathermaychange.24.Suchcasesbearacloseresemblancetothecasesinwhich,asIpointedoutintheprecedingsection,whatisgivenindirectexperienceappearsashavingadefiniterhythmicpattern,andwefeelthepresenceofsomeunderlyingbutnotdirectlyexperiencedcause.25.Ontheotherhand,asIhavepointedout,therearealsomanyinstancesinwhichaHumeanmodelofexplanationseemstoserveourpurposesadequatelyinordinarylife.26.Theterm"mana"enteredanthropologicalliteraturethroughR.H.Codrington'sTheMelanesians(Oxford:ClarendonPress,1891),andbecameextendedinitsapplicationbyR.R.Marrett'sarticleon"Mana"inHastings' NOTESTOPAGES77-85EncyclopediaofReligionandEthics(Edinburgh:T.&T.Clark,1908-26),andbythewritingsofDurkheim,Lowie,andothers.Alternativeinterpretationsoftheconceptandofthephenomenatowhichitwasusedtoreferwerecloselytiedtoalternativetheoriesofmagicandreligioninprimitivesocieties.Atpresentasomewhatrestrictedinterpretationappearsdominant(seeE.E.EvansPritchard,TheoriesofPrimitiveReligion[Oxford:ClarendonPress,1965],p.110).Formypurposes,however,wheretheissueisoneofcausalefficacyandnotatheoryofreligion,andwheretheoreticalissuesconcerning.theroleofmagicinprimitivecommunitiesdonotarise,itispresumablylegitimatetousetheterminitsmoreextendedsense.Foronestandardworkonmana,seeF.R.Lehmann,Mana:DerBegriffdes"ausserordentlichWirkungsvollen"beiSudseevolkern(Leipzig:O.Spamer,1922).MyuseofthemanaillustrationinthepresentconnectionderivesfromaninterestingarticlebyWolfgangKohler,"PsychologicalRemarksonSomeQuestionsofAnthropology,"AmericanJournalofPsychology50(1937):271-88.AsimilarpointconcerningsuperstitionhasbeenmadebyWilliamRuddickin"CausalConnections,"Synthese18(1968):49.27.Forexample,HumesaysinAnInquiryConcerningHumanUnderstanding,sec.7,pt.2:"One.eventfollowsanother,butwenevercanobserveanytiebetweenthem.Theyseemconjoined,butneverconnected"(Selby-Biggeedition[Oxford:ClarendonPress,1966],p.74).28.ItisinsuchcasesthatHumespeaksof"secretpowers."SeeInquiryConcerningHumanUnderstanding,sec.4,pt.2,pp.32-34.CHAPTERFOUR1.EnglishWorksofThomasHobbes,ed.W.Molesworth(London:Bohn,1839-45),vol.1,chap.9,pp.121-22.2.Inhiswell-knownarticle"CausalRelations"(JournalofPhilosophy64[1967]:691-703),DonaldDavidsonwasalsoconcernedwithcausalexplanationsofparticularevents,andinthisconnectionhecitedasimilarexample.Hewas,however,criticalofMill'sinclusionofallrelevantconditionsaselementsinthetruecauseofaparticularevent.However,thatpointwasnotessentialtothemainthrustofhisargument(seep.692).3.InthisconnectionitisappositetociteJ.L.Mackie'sdefenseofMillagainstthecriticismsmadebyHartandHonore.SpeakingofwhatMintermed"thephilosophicalview"ofcausation,inwhichanydistinctionbetween"cause"and"mereconditions"disappears,Mackiesays:"Sincewhatwerecognizeasacause,ratherthanamerecondition,commonlydependsonwhatweknow-orwhatweknewfirst-orwhatiscloselyrelatedtoourinterests,thereismuchtobesaidforMill'srefusaltodistinguish'philosophicallyspeaking'betweencausesandconditions.Asananalysisofordinarylanguage,thiswouldbewrong;butfromatheoreticalpointofview,asanaccountofcausalprocessesthemselves,itwouldberight"(TheCementoftheUniverse:AStudyofCausation[Oxford:ClarendonPress,1974],p.120).4.Milloffersanaccountofwhywesingleoutsomeeventasthecauseofaneffect,andwhy,incontrast,wedesignatecontinuingstatesasbeingmerelyattendantconditions.Hesaysofthelatterthatthey"mightthereforehaveprecededtheeffectbyanindefinitelengthofduration,forwantoftheeventwhichwasrequisitetocompletetherequiredconcurrenceofconditions:whileassoonasthatevent...occurs,noothercauseiswaitedfor,buttheeffectbeginsimmediatelytotakeplace:andhencetheappearanceispresentedofamoreimmediateandcloseconnectionbetweentheeffectandthatoneantece- NOTESTOPAGES85-95215dent,thanbetweentheeffectandtheremainingconditions.Butthoughwemaythinkpropertogivethenameofcausetothatonecondition,thefulfillmentofwhichcompletesthetale,andbringsabouttheeffectwithoutfurtherdelay;thisconditionhasreallynocloserrelationtotheeffectthananyoftheotherconditionshas.Theproductionoftheconsequentrequiredthattheyshouldallexistimmediatelyprevious,thoughnotthattheyshouldallbegintoexistimmediatelyprevious"(J.S.Mill,ASystemofLogic,bk.3,chap.5,sec.3).(Forthisvariantofthepassage,seeCollectedWorksof1.S.Mill[Toronto:UniversityPress,1973],7:328.)5.C.J.Ducasse,Nature,Mind,andDeath(LaSalle,Ill.:OpenCourt,1951),p.108.ThisvolumecontainsDucasse'sfullestexplanationofhistheoryofcausation,but(asIhavepointedout)itcanbesupplementedbyanearliermonograph,CausationandOtherTypesofNecessity,UniversityofWashingtonPublicationsintheSocialSciences,vol.1,no.2(Seattle,1924),andbyhislater.collectionofstudies,Truth,Knowledge,andCausation(London:Routledge&KeganPaul,1968).6.Thispassagecomesfromthe14thedition,s.v.lttuberculosis."Theaccountofthepathologyoftuberculosisgiveninthe15theditionismoredetailed,butisentirelycompatiblewiththepointherebeingmade.EquallyclearistheaccountIntheAmericanEncyclopedia,wheretheetiologyisdiscussedintermsofthehost-parasiterelationship.7.Inadifferentcontext(namely,inattackingthenotionofthe"accidental"inhistory),MichaelOakeshottmakesasimilarpoint(cf.ExperienceandItsModes[Cambridge:AttheUniversityPress,1933],p.140;also,pp.129and142).AlexanderGerschenkron,theeconomichistorian,distinguishesbetweenwhatheterms"nonfacts"and"counterfacts"(ContinuityinHistoryandOtherEssays[Cambridge,Mass.:HarvardUniversityPress,BelknapPress,1968],pp.53-54).IfIunderstandhimcorrectly,thisparallelsthedistinctionIherewishtodraw.8.Cf.H.L.A.HartandA.M.Honore,CausationintheLaw(Oxford:ClarendonPress,1959),pp.35-36and37fortheiruseofthisexample.9.ThemostinfluentialstatementofthesupposeddifficultyhasprobablybeenthatofBertrandRussellinhisessay"OntheNotionofCause,"inMysticismandLogic(NewYork:W.W.Norton,1929),pp.184-85.10.Itisobviousthatinthisdiscussion,andelsewhere,Ihavebeenassumingthatthenotionofwhatconstitutes"aprocess"isnothopelesslyobscure.Idonotbelievethatitis,andIseenoreasontothinkthatitraisesspecialmetaphysicaldifficultiesthatarenotparalleledinthosecasesinwhichweclaimtoknowwhatwemeanby"anobject,""anevent,""astateofaffairs,"etc.AtthispointIshouldalsoindicatethatmypresentaccountofthecauseeffectrelationdiffersfromthemannerinwhichIstatedmypositioninTheProblemofHistoricalKnowledge(NewYork:Liveright,1938).ThereIspokeof"events"and"sub-events"ratherthanspeakingoftheendpointinaprocessandwhatleduptoit.Incorrespondenceandconversations(longago),HughMiller,formerlyoftheUniversityofCaliforniaatLosAngeles,pointedouttomesomedifficultiesinherentinmyearlierformulation.Ishouldliketothinkthatmypresentwayofformulatingmyviewsoftherelationbetweencauseandeffecthasovercomethosedifficultieswithoutgivingupwhatwasbasicinn1yearlierposition,andwithouthavingengenderedseriousnewdifficulties.11.Forexample,seeMichaelDummett,"CananEffectPrecedeItsCause?"inProceedingsoftheAristotelianSocietysuppl.vol.28(1954):27-44,and"BringingAboutthePast,"PhilosophicalRevie1v73(1964):338-59.Also,seeRichardTaylor'sdiscussionoftheissueinhistwoarticlesentitled"Causation," 216NOTESTOPAGES96-107oneofwhichappearedinTheMonist47(1963):287-313,andtheotherintheEncyclopediaofPhilosophy,s.v."causation."Forfurtherreferences,seeG.H.vonWright,ExplanationandUnderstanding(Ithaca:CornellUniversityPress,1971),p.185,n.18.12.IfIunderstandhimcorrectly,thispointwouldbeendorsedbyWesleySalmon,who,indiscussinghisownandReichenbach'sviewofcausalrelevance,says:"Oneverybasicandimportantprincipleconcerningcausalrelevance...[is]thatitseemstobeembeddedincontinuousprocesses"("TheoreticaExplanation,"inStephanKorner,ed.,Explanation[Oxford:Blackwell,1975],P.132).13.Foranattempttodrawadistinctionbetweencasuallawsandlawsthatconnectproperties,seeDucasse,Nature,Mind,andDeath,pp.127-28.14."MakingSomethingHappen,"inSidneyHook,ed.,DeterminismandFreedomintheAgeofModernScience(NewYork:NewYorkUniversityPress,1958),pp.23-24.15.Onthispoint,seeHartandHonore,CausationintheLaw,pp.36-37.16.In"TheFunctionofGeneralLawsinHistory,"asreprintedinAspectsofScientificExplanation(NevYork:FreePress,1965),p.232.17.Forexample,inHempel'sillustrationwemustexcludethepossibilitythattheownervillstarthiscarinordertowarmitsengine,orthathewilladdanantifreezesolutionasthetemperaturedropstothefreezingpoint.ThesequentialaspectoftheseriesofoccurrenceswouldsurelynotbedeniedbyHempel,eventhoughthisaspectdoesnotbecomeapparentinhisschematicaccountofhowinitialconditionsandlawsareusedinexplainingagiveneffect.18.Cf.ErnestNagel,"DeterminisminHistory,"PhilosophyandPhenomenologicalResearch20(1960):293-94;also,hisStructureofScience(NewYork:Harcourt,Brace,1961),pp.594-95.InbothplacesheusesanexampleofdeterminismthatwasformulatedbyL.J.HendersoninPareto'sGeneralSociology(Cambridge,Mass.:HarvardUniversityPress,1937),chap.3.19.Torelatethisandwhatfollowstomoretechnicaldiscussionsoftheissueinvolved,IfindmyselfinagreementwithJ.L.Mackiewhenhecharacterizesthevariousfactorsenteringintothecauseofaneffectasbeingwhatheterms"inusconditions,"andalsowithwhatItaketobeMichaelScriven'smeaningwhen,inreviewingErnestNagel'sStructureofScience,hesaysthatcausesare"contingentlysufficient."Ishallquotethesummarystatementsgivenbyeach.Mackiesays:"Inthecasedescribedabovethecomplexformula'(ABCorDGHorJKL)'representsaconditionwhichisbothnecessaryandsufficientforP:eachconjunction,suchas'ABC,'representsaconditionwhichissufficientbutnotnecessaryforP.Besides,ABCisaminimalsufficientcondition:noneofitsconjunctsisredundant;nopartofit,suchasAB,isitselfsufficientforP.Buteachsinglefactor,suchasA,isneitheranecessarynorasufficientconditionforP.YetitisclearlyrelatedtoPinanimportantway:itisaninsufficientbutnonredundantpartofanunnecessarybutsufficientcondition:itwillbeconvenienttocallthis(usingthefirstlettersoftheitalicizedwords)aninuscondition"(CenlentoftheUniverse,p.62).Inholdingthatcauses(bywhichhemeansspecificconditions)are"contingentlysufficient,"Scrivensays:"Theyarepartofasetofconditionsthatdoesnotguaranteetheoutcome,andtheyarenon-redundantinthattherestofthisset(whichdoesnotincludeallotherconditionspresent)isnotalonesufficientfortheoutcon1e"(ReviewofMetaphysics17[1964]:408).Scriventhenintroducesfiverefinementsofthisaccount(pp.408-12).20.In"Causation"(Monist47[1963]:287-313),RichardTaylortakesadifferentposition.Hearguesthateachoftheconditionsthat,together,arenecessaryfortheproductionofagiveneffectmustalso,whentakenindividually,beregardedasnecessary. NOTESTOPAGES110-16CHAPTERFIVE1.ExperienceandItsModes(Cambridge:AttheUniversityPress,1933),p.141.2.BothW.H.WalshandW.H.DraycontributeddiscussionsofOakeshott'sviewsregardinghistoryinavolumeofessaysentitledPoliticsandExperience(London:CambridgeUniversityPress,1968),editedbyPrestonKingandB.C.Parekh,andpresentedtohimonhisretiren1ent.3.ExperienceandItsModes,p.127.4.From"TheAutonomyofHistoricalUnderstanding,"originallypublishedinHistoryandTheory5(1965):24-47;reprintedwithminorrevisionsinW.H.Dray,ed.,PhilosophicalAnalysisandI-listory(NewYork:Harper&Row,1966),wherethispassageappearsonp.178.5.Drayconcludesanarticleentitled"OntheNatureandRoleofNarrativeinHistoriography,"inHistoryandTheory10(1971):153-71,withasectioncontrastingMinkwithOakeshott,andwithothernarrativists;he,too,findsMink'spositiononthispointespeciallycongenial.IntheReviewofMetaphysics21(1967-68):667-98,MinkreviewedthebooksofGa11ie,White,andDanto,butinthatessayhedidnotdevelophisownviewsbeyondwhatwascontainedinhisearlierarticle.6.Cf.PatrickGardiner,TheNatureofHistoricalExplanation,(London:OxfordUniversityPress,1952),pp.89-90and96-97;W.H.Dray,LawsandExplanationinHistory(London:OxfordUniversityPress,1957),pp.32-37;ArthurDanto,AnalyticalPhilosophyofHistory(Can1bridge:AttheUniversityPress,1965),p.234and,foraparallelexample,pp.234-35.7.See"Causes,ConnectionsandConditionsinHistory,"inDray,PhilosophicalAnalysisandHistory,pp.242-43.Scriven,however,doesnotspeakofthepart-wholerelationshipbetweenanyoneelementinthecausalconditionsandthetotaleffect;rather,hespeaksofthecauseas"physicallyidenticalandonlyconceptuallydistinct"fromtheeffect.8.Thereare,ofcourse,otherbiographicalstudies,suchasliteraryorscientificbiographies,inwhichsomefacetofculture,ratherthantheinstitutionalaspectsofaparticularsociety,providetheprimarycontextintermsofwhichthelifeandworkofanindividualareviewed.Somebiographies,dependinguponthecareeroftheirsubject,mayfusetheseinterests;othersmayfocusonthesubjectinrelationtoaninterpretivethesisregardingthecharacteristicsoftheperiodinwhichhelived.9.Incontrasttothis,asIpointedoutinchapter2,aspecialhistoryinvolvesahistoricalstudyofsometypeofculturalproduct,tracingtheconnectionsandthechangesinexan1plesofit;itisnottheprimarytaskofspecialhistoriestodealwiththenatureandchangesofthesocietyorsocietiesthateitherproducedorhavepreservedtheseproducts.10.C.V.LangloisandC.Seignobos,IntroductiontotheStudyofHistory(London:Duckworth,1925),p.214.11.Imightherepointoutthatanyeventspriortotheparticularseriesofeventsunderinvestigationarenottobeviewedasdeterminingitscharacteristics,exceptindirectly.Thus,whiletheymaybepartofthecauseofitscause,theyarenottobeincludedan10ngthecausalfeaturesresponsibleforit.12.Thiswasarelativelycommonpointofviewamongsocialpsychologists,ofwhomWillialTIMcDougallwasone.AmongsocialanthropologistsonefindsitrepresentedinB.Malinowski,AScientificTheoryofCulture(ChapelHill:UniversityofNorthCarolinaPress,1944),thoughnotinhisearlierworks.Inthreearticlestobecitedinn.30,below,Ihavearguedagainstarelatedview,ttmethodologicalindividualism,"whichholdsthatsocietalfactsare 218NOTESTOPAGES122-25tobeunderstoodandaretobeexplainedintermsofthebehaviorofindividuals.13.Thisremarkalsoappliestomostattemptstousepsychoanalytictheoriesinhistoricalexplanations.Whilesomehistoriansholdthatsuchtheoriesmaybefruitfullyappliedbothinbiographicalstudiesandinexplainingthepersistentpatternsofactionofvariouspoliticalfigures,manyoftheactionswithwhichhistoriansareconcernedinvolvethedecisionsofindividualsconcerningwhomthereisinsufficientknowledgetowarrantaconcreteapplicationofpsychoanalytictheory.14.AlanDonagan,in"HistoricalExplanation,"Mind66(1957):163,usesasimilarargumentasthatgivenabovetoexplaintheheuristicvalueofgeneralizations.WhathasbeensaidmayalsoremindthereaderofMichaelScriven'sthesisthatthegeneralizationshistoriansuseare"truisms."(Cf."TruismsastheGroundsforHistoricalGeneralizations"inPatrickGardiner,ed.,/'I'heoriesofHistory[Glencoe,Ill.:FreePress,1959],pp.443-75.)H·owever,Scriveninterpretsthesetruismsasrelatingtowhatindividualsnormallydo,andhethereforecallsthem"normicgeneralizations."Idonotbelievethat,asarule,thegeneralizationshistoriansusearemeanttorefertowhatisnormalbehavior,ortowhatnormallyhappens,ineitherthestatisticalorthequasi-normativesenseoftheterm"normal."Therefore,inspiteofasuperficialresemblancebetweenwhatIheresayconcerningtheloosenessofthegeneralizationsusedbyhistoriansandtheviewsbroughtforwardbyScriven,Iwishtoseparatemypositionfromhis.15.Inpart,itsimportancefordiscussionsofhistoricalmethodologyliesinitsabilitytoprovideanalternativetoCarlHempel'sviewthatbecauseoftheloosenessoftheirgeneralizationshistoriansofferonly"explanationsketches,"notexplanations.(Cf."TheFunctionofGeneralLawsinHistory,"sec.5.4,reprintedinAspectsofScientificExplanation[NewYork:FreePress,1965],_p.238.)16.TheonlyworkIknowthatmakesaseriousattempttodefinetheparametersofthistypeofproblemandtoelicitconclusionsisSizeandDemocracybyRobertA.DahlandEdwardR.Tufte(Stanford:StanfordUniversityPress,1973).17.E.B.Tylor,"OnaMethodofInvestigatingtheDevelopmentofInstitutions,"JournaloftheRoyalAnthropologicalInstitute18(1889):245-69.18.GeorgeC.HomansandDavidM.Schneider,Marriage,Authority,andFinalCauses(Glencoe,Ill.:FreePress,1955).Forthesummarystatementoftheirhypothesis,seep.28.19.OneofthemoststimulatingattemptstoprovidesuchalawseemstometohavebeenRobertMichels'developmentofMosca'stheories.AccordingtoMichels,thereisabasicprincipleofpoliticalorganizationthat-inamannerreminiscentofMarx-hecalled"theironlawofoligarchy"(PoliticalParties[NewYork:FreePress,1966],pt.6,chap.2;cf.FirstLecturesonPoliticalSociology[Minneapolis:UniversityofMinnesotaPress,1949],pp.141-42).Itwashiscontentionthattherewasinherentinthenecessityforpoliticalorganizationinsocietyaninescapabletendencyforcontroltobeexercisedbyaminoritygroup.Hephrasedthismostconciselyasfollows:"Organizationimpliesthetendencytooligarchy.Ineveryorganization,whetheritbeapoliticalparty,aprofessionalunion,oranyotherassociationofthekind,thearistocratictendencymanifestsitselfveryclearly.Themechanismoftheorganization,whileconferringasolidityofstructure,inducesseriouschangesintheorganizedmass,completelyinvertingtherespectivepositionoftheleadersandtheled.Asaresultoforganization,everypartyorprofessionalunionbecomesdividedintoaminorityofdirectorsandamajorityofdirected"(PoliticalParties,p.32).InUnionDemocracy(Glencoe,Ill.:FreePress,1956),S.M.Lipset,M.A.Trow, NOTESTOPAGES127-31andJ.S.ColemanexaminedtheInternationalTypographicalUnioninanefforttoexplainwhyit,aloneamonglaborunions,providedaconspicuousexceptiontoMichels'ironlawofoligarchy.(Seetheirfirstchapter,"DemocracyandOligarchyinTradeUnions.")Theiranswerinvolvedcertainofthestructuralfeaturesofthisunion,but'theirexplanationofthedevelopmentandpersistenceofthesefeaturesdemandedanappealtovarioushistoricaloccurrences:Thecumulativeeffectsoftheseoccurrenceschangedtheinitialconditionspresentforeachsubsequentstageinthedevelopmentoftheunion,thusnegatingtheapplicabilityofMichels'law.Fortheirbriefdiscussionofthisbasictheoreticalpoint,seepp.393-94and402-3.WhileMichels'formulationanddefenseofhispositionwasoverburdenedbyaconcernwithproblemsofsocialismanddemocracyinmodernpoliticallife,andwasthereforenotformulatedintermsthatmakeit,readilyapplicabletoallfonnsofsociety,onecanconceiveofamoregeneralizedstatementofitwhichcouldbeappliedtoallformsoforganizationandnotto.modernformsofpoliticallifealone.ImightaddthatoneconventionalgeneralizationIhavecited,concerningsizeofpopulationanddirectdemocracy,canitselfbeconsideredonlyaspecialcaseofMichels'"law,"ashehimselfattemptedtoshowinthechapterofPoliticalPartiesentitled"TheMechanical·andTechnicalImpossibilityofDirectGovernmentbytheMasses,"fromwhichtheabovequotationwastaken.(Seealsopt.;,chap.1ofthatwork,on"TheReferendum.")However,inhisexplanationofwhytheironlawofoligarchyholds,Michelsofferedtwobasicprinciples,onlyoneofwhichdependsupon"tacticalandtechnicalnecessities";theotherrestsonassumptionsastopsychologicalchangesthatindividualsundergowhentheyassumerolesofleadership(PoliticalParties,pp.400-401).Onlythefirstofthese,andnotthepsychologicalassumption,seemstomelikelytohaveuniversal.applicability.20.Cf.W.B.Gallie,PhilosophyandtheHistoricalUnderstanding(NewYork:SchockenBooks,1964),PP.107-8.21.Oncethisdistinctionis>drawn,sothatnotalloccurrencesthatmayaffectvariousaspectsoflifeinaparticularsocietyareregardedasbelongingwithinitsownhistory,thetemptationtoviewthewholehumanpastasconstitutingasinglehistoryshoulddisappear.However,idealistssuchasOakeshottwill,onmetaphysicalgrounds,ofcourserejectanysuchdistinctionbetweenexternalandinternalrelations.22.Asweshallsee,generalhistoriesthatareprimarilyinterpretiveinstructurefallbetweenspecialhistoriesandotherformsofgeneralhistoryinthisrespect.23.Whiletheterm"genre"isusuallyusedonlyinreferringtostylistictypesinthearts,Ibelievethat·itsusecanbeextendedtootherfields.Onecan,forexample,saythatdifferenttypesofphilosophicproblemsanddifferentmethodsofapproachingtheseproblemsresemblethediversitytobefoundindifferentliterarygenresorindifferentgenresintheplasticarts.So,too,inthesciencestherearemanysortsofproblemstobeinvestigated,andthereisalsovarietyinthestylesofinvestigationthatdifferentscientistsfollow.Indifferentperiodsson1etypesofproblemsandsomestylesofinvestigationmaybemoredominantthanothers,justasistrueinthearts.24.Itissaid-andprobablycorrectlysaid-thatinapritnitivesociety(thatis,inanonliteratesociety)traditionismorerigorouslyfollowedandinnovationismorerestrainedthaninothersocieties;asaconsequence,agreaterdegreeofcontinuityinthecultureofthatsocietyistobeexpectedthanwouldotherwisebethecase.Inaddition,thegreaterthecontactofasocietywithothersocieties,themoreopportunitythereisforculturalinterchangeand,therefore,themoreareinnovationsanddiscontinuitieslikelytooccur. 220NOTESTOPAGES132-4°25.EnglishLiteratureintheSixteenthCenturyExcludingDrama(Oxford:ClarendonPress,1954),p.56.26.Inthisconnectiononemaynotethatthereissomeevidencethateconomicfactors,ratherthantheexploitationofnewscientificdiscoveries,playadominantroleindevelopingmajortechnologicalinnovations.Foradefenseofthisthesis,seeJacobSchmookler,InventionandEconomicGrowth(Cambridge,Mass.:HarvardUniversityPress,1966),especiallychap.3.Inanysuchstudythereis,ofcourse,theproblemofdefining"majortechnologicalinnovations"inawaythatdoesnotprejudgetheissue,andSchmookler'scasestudiesmightbechallengedonthisground.Nevertheless,hisstudycallsattentiontotheroleplayedbyeconomicfactors-asdistinctfromscientificadvances-infosteringtechnologicalchange.27.Isay"concreteexplanations"sinceIthinkitentirelypossiblethatonemightbeabletosetupgeneralprinciplesrelatingtostylisticchangeinthearts,andperhapsinotherfieldsofculturalhistoryaswell.Suchprinciplesmightrefertochangesinfashionduetofactorssuchasthesatiationoftasteforastylewhenthatstylehasbeendominantforatime,orbecauseoffactorssuchasatendencyforstylestobevulgarizedastheirinfluencespreads.However,suchgeneralprinciples-ifanyaretobefound-wouldonlyexplainwhythereischangeandwouldnotconcretelyexplainthedirectioninwhichthechangetookplace.28.InEnglishLiteratureintheSixteenthCentury,C.S.Lewismadesomeveryappositeremarksonhowtheword"Renaissance"-whichwasoriginallyusedbyhumanistsconsciousoflivinginarenascentiainwhichGreekwasrecoveredandtherewasarevivalofclassicalLatin-becamedebasedbytheinclusionofotherelementswhichwerequiteindependentofthenewclassicallearning.Ashesaid,"Wherewehaveanounwetendtoimagineathing.ThewordRenaissancehelpstoimposeafactitiousunityonalltheuntidyandheterogeneouseventswhichweregoingoninthosecenturiesasinanyothers.Thusthe'imaginaryentity'creepsin.Renaissancebecomesthenameforsomecharacterorqualitysupposedtobeimmanentinalltheevents,andcollectsveryseriousemotionalovertonesintheprocess....NoonecannowusethewordRenaissancetomeantherecoveryofGreekandtheclassizingofLatinwithanyassurancethathishearerswillunderstandhim.Badmoneydrivesoutgood"(pp.55-56).29.Forexample,thisistheviewofMelvilleHerskovitsinhiswell-knowntextinsocialanthropologyManandHisWorks(NewYork:Knopf,1948),pp.21-28.Foraphilosophicdiscussionof"methodologicalindividualism"andsomeofitsalternatives,see"HolismandIndividualisminHistoryandSocialScience,"byW.H.Dray,intheEncyclopediaofPhilosophy.30.Iarguedagainstmethodologicalindividualismin"SocietalFacts,"BritishJournalofSociology6(1955):305-17,andin"PsychologyandSocietalFacts,"inLogic,Laws,andLife,ed.R-obertG.Colodny(Pittsburgh:UniversityofPittsburghPress,1977).In"SocietalLaws,"BritishJournalforthePhilosophyofScience8(1957):211-24,Iattemptedtoshowthatarejectionofmethodologicalindividualismdoesnotentailanacceptanceofthetypeofs-ocietallawthatPopper,IsaiahBerlin,andothermethodologicalindividualistshaveassumedthatitdoes.31.FrederickA.Olafson,"HumanActionandHistoricalExplanation,"inNewEssaysinPhenomenology,ed.JamesEdie(Chicago:QuadrangleBooks,1969),seeespeciallypp.366-67.32.ItisinstructivetonoteadiscussionoftheelementofnoveltyinearlyEnglishdramainF.P.Wilson'sTheEnglishDrama1485-1585(Oxford:ClarendonPress,1969).InspeakingofHenryMedwall'sworkshesays:"AstheauthorofNatureMedwallwouldbarelymeritamentioninthehistory-of NOTESTOPAGES146-58221ourdrama,butastheauthorofFulgensandLucrece,thefirstpurelysecularEnglishplaythathassurvived,heisasignificantfigure....Thediscoveryofthisplayin1919causedalmostasmuchsurpriseastherecentdiscoveryofanhistoricalplaydatingfromthegreatperiodofGreekdrama"(p.7).ButWilsonimmediatelyadds:"NodoubtMedwall'splayseemstousmoreoriginalthaninfactitwas,"andhetracesawholeseriesofworksofdifferenttypesthatalsowerepopularincharacterandwerebothsecularandquasi-dramatic.Tothis·headdstheconcludingcomment,drawnfromaremarkofSirEdmundChambers:"Sowebuildupthepast."Itisthisthatanexperiencedculturalhistorianisinapositiontodo,butthenoviceisnot.CHAPTERSIX1.Forbothcognitivistsandnoncognitivistsmoraljudgmentsarevalidonlyiftheyareobjective.ForthedistinctionbetweenthevalidityandthetruthofInoraljudgments,seechapter6ofmybookThePhenomenologyofMoralExperience(Glencoe,Ill.:FreePress,1955).2.Intheconcludingsectionof"SomeIssuesintheLogicofHistoricalAnalysis,"ErnestNagelsuggestedvariouswaysinwhichhistoriansattempttoestimatetherelativeimportanceoftwocausalfactorsinasituation.IwouldbewillingtoaccepteachofthefivewaysNageldifferentiated,butIwishtocallattentiontothefactthathefailedtoincludethetypeofassessmentIhaveindicated.Werehetohaveincludedthistype-buthisgeneralconceptionofcausalexplanationperhapsmadeitimpossibleforhimtodoso-hisfinalpositionmighthavebeenless·discouragingthanitwas.(HisarticlewasoriginallypublishedinScientificAmericanin1952;itisreprintedinPatrickGardiner,ed.,TheoriesofHistory[Glencoe,Ill.:FreePress,1959],pp.373-85.)3.Inthefirstthreesectionsofchapter25ofTheOpenSocietyandItsEnemies(Princeton:PrincetonUniversityPress,1963),KarlPopperarguesthattheorieshavenoplaceinhistory:Thereareatbestgeneralinterpretations,whichhealsocallsquasi-theories.Thiscontentionisbasedontwopremises,eachofwhichItaketobefalse.First,Popperassumesthatagenuinetheorycanbeformulatedonlyifonehasalreadyestablishedempiricallaws,but,asIhaveindicated,thiswasnotthecasewithDarwin'sevolutionarytheory.Second,Popperassumesthatthelawsuponwhichtheoriesregardinghistorywouldnecessarilydependwilltaketheformofsayingthat,givenanoccurrenceoftypea,anoccurrenceoftypebwillalwaysfollow.However,asIhaveelsewheretriedtoshow,sequentiallawsofthistypeconstituteonlyoneofthekindsoflawsthatmightbeusedinunderstandingsocietalchange;yetitistheonlytypethatPopperconsiderswhenherejectsthepossibilityofestablishinglawsconcerninghistory.(Seemyarticle"SocietalLaws,"BritishJournalforthePhilosophyofScience8(1957):211-24;reprintedinW.H.Dray,PhilosophicalAnalysisandHistory[NewYork:Harper&Row,1966].)4.CitedbyEileenPower,"OnMediaevalHistoryasaSocialStudy,"inTheStudyofEconomicHistory,ed.N.B.Harte(London:Cass,1971),p.115.5.Thepassagereadsasfollows:Itisalwaysdifficult,andfrequentlyimpossible,tobringtothepointofinductivetestingasystemofexplanationofasystemofinterrelatedfacts."IfKeyneswasreallytobesuccessful"(remarksSirRoyHarrodinhisLifeofKeynes)"heshouldhavebeenable,itisargued,torefute,say,Mr.D.H.RobertsonbyshowingasetoffactswhichtheKeynesiandoctrinewouldfit,whiletheotherwouldnot.Unhappily,thestateofeconomicsisnotsoadvanced."Bothoftwoalternativesystemsofexplanationmayfitmostofthefacts(andthoseitdidnot 222NOTESTOPAGES160-71-anadherentmighthope-couldbeexplainedawaywithfurtherresearchorreflection).Incomplicateddisputes,itisnotdetailedtheories,butrivalattitudestotheworldthatareincollision.[TimesLiterarySupplement(London),2May1975,p.471]6.Tobesure,ifoneweretointerpretHegel'sphilosophyofhistoryasbeingconcernedonlywiththedevelopmentofthenotionofpoliticalfreedom,andnotwiththegrowthanddeclineofthosegreatsocietieswhosedestinieshetraced,thisstricturewouldnothold;however,suchaninterpretationofwhathisphilosophyofhistorywasaboutwouldseemtomeuntenable.AstotheviewsofComteandSpencerregardinghistoricalmethodology,seemybookHistory,Man,andReason(Baltimore:JohnsHopkinsPress,1971),particularlypp.88-89.7.Inmyopinion,therecanberelevant"outside"evidencefororagainstanyformofgeneralaesthetictheory:Appealcanbemadetohistoricalandpsychologicalfacts,aswellastophenomenologicalinvestigations,allofwhichcangofartowardresolvingsuchdisputes.However,Ishallnotpressthepointhere.8.Firstin"TheIntelligibilityofHistory,"Philosophy17(1942):128-43;theninIntroductiontothePhilosophyofHistory(London:Hutchinson,1951),pp.23-24and59-64.9.PhilosophyoftheInductiveSciences(London:J.W.Parker,1847),2:36.ThispartofthatworkwasrepublishedasNovumOrganumRenovatum,andthepassageoccursonp.60oftheeditionpublishedbyJ.W.ParkerinLondonin1858.10.IntroductiontothePhilosophyofHistory,p.23.11.Ibid.,pp.31-33.Cf."TheIntelligibilityofHistory,"p.130.12.IntroductiontothePhilosophyofHistory,p.62,and"TheInteIligibilityofHistory,"p.133.13.MortonWhite'streatmentofcolligationinTheFoundationofHistoricalKnowledge(NewYork:Harper&Row1965),pp.252-54and257-64,isevenmoreexplicitlyopposedtothepossibilityofobjectivehistoricalknowledge,asIhavedefinedthatterm.14.SeeIntroductiontothePhilosophyofHistory,pp.59-64.15.AlthoughErwinPanofskywasbynomeansskepticalofthevalidityofnlostart-historicalperiodizations,oneofhisessays,"TheFirstPageofGiorgioVasari's'Libro':AStudyontheGothicStyleintheJudgmentoftheItalianRenaissance,"admirablyillustratesthepointIammaking.SeehisMeaningintheVisualArts(GardenCity,N.Y.:DoubledayAnchor,1955),pp.169-225,andespeciallysection5ofthatessay.CHAPTERSEVEN1.Cf.chap.2,sec.3.Whatwassaidthereholdsnotonlyofinterpretiveaccountsofparticularperiods,myprimaryconcerninthatdiscussion,butholdsalsoofinterpretivebiographiesandinterpretivestudiesof,forexample,theworksofanindividualauthor.InthisconnectionImightciteapassagefromJ.A.Passmore,"TheObjectivityofHistory":Sonlelittletimeago,IwroteabookwhichpurportedtobeaninterpretationofHume'sphilosophy.Onerevieweraddressedmesomewhatasfollows:"apossibleinterpretation,butotherinterpretationsareequallypossible."Howisonetoreply?...Whathappensissomethinglikethis:aninterpretationissuggestedbycertainpassagesinHume;thatinterpretationisthenconfirmedbypassagesIhad NOTESTOPAGES172-84notpreviouslysomuchasnoticed,whichtheproposedinterpretationservestoilluminate.OrIdiscoverthatpassagesIpreviouslycouldnotunderstandnowmakegoodsense....Ifareaderisconvincedofmyinterpretation,thiswillbebecausehehashimselfbeenpuzzledbypassagesinHume,andmyinterpretationsolveshispuzzleforhim.[.Philosophy33(1958):106-7]E.D.Hirsh'sboldyetcarefulstudyValidityinInterpretation(NewHaven:YaleUniversityPress,1967)shouldbeconsultedbyanyoneconcernedwithproblemsintheauthenticationofinterpretations.2.Toreplyinadvancetoanapparent·exceptionthatmightspringtomind,IremindthereaderofwhatIsaidinAppendixBconcerningtheproblemof"thecauseoftheAmericanCivilWar":Affixingofpraiseorblamepresupposesdetailed,structuredhistoricalknowledgeandisnotitselfaninvestigationofcauses,asIusethatterm.3.IhavedealtwiththisatgreaterlengthinanarticleappearinginaSpanishsymposiumonthephilosophyofhistory.Intranslationitwasentitled"HistoriayUniversalidad"anditappearedinRevistadelaUniversidaddeMadrid12(1963),no.45.4.Seechapter1,section1.Hempel'spaper,originallypublishedintheJournalofPhilosophyin1942,isreprintedinAspectsofScientificExplanation(NewYork:FreePress,1965).5.InLawsandExplanationinHistory(London:OxfordUniversityPress,1957),pp.67-68.FormyowncriticismofHempelconcerningthesamepoint,see"HistoricalExplanation:TheProblemof'CoveringLaws,'"HistoryandTheory1(1961):233-38.6.In"HistoricalExplanation,"pp.239-41,IcriticizedDrayinthesameway,anddevelopedthepointatsomewhatgreaterlength.Atthattime,however,Ifailedtosee-asInowsee-thattherearecasesinwhichtracingaseriesofevents,suchasthoseconstitutingacampaign,mayprovideanadequateexplanationofaparticularstateofaffairs.However,Iwishtoemphasizethatinsuchcasesonemustnotassume(asDraytendstodo)thattherelevanteventsformasinglelinearseries.Onthispoint,seechapter2,section1,andalsoanarticle,"ANoteonHistoryasNarrative"(HistoryandTheory6[1967]:413-19),inwhichIdevelopedthepointinamorepreciseway.DrayrespondedtothelatterarticleinHistoryandTheory8(1969):287-94.7.Thisisadmittedlyacrudestatementofthefactorsthatmightbeinvolved.Forexample,inspeakingof"economicfactors"Imightbereferringtotheconditionsobtainingeitheratthesourceofemigrationoratitsgoal.Foracarefulstudyofonecaseofimmigration,seeSimonKuznets,ImmigrationofRussianJewstotheUnitedStates:BackgroundandStructure,PerspectivesinAmericanHistory,vol.9,1975(Cambridge,Mass.:HarvardUniversity,CharlesWarrenCenterforStudiesinAmericanHistory,1976).Inthatstudy(pp.86-89),Kuznetscitescomparablecasesthatsuggestthathisfindingsaregenerallyapplicableandarenotconfinedtothecaseathand.8.Whenlawsaremistakenlyassumedtostateaninvariantconnectionbetweenspecifictypesofoccurrences,itisplausibletoholdnotonlythathistoriansdonotuselawstoexplainevents,butthatnoonecanciteanyexampleofawell-formulatedlawthatahistoriancouldpossiblyuse.When,however,itisunderstoodthatinthenaturalsciences,too,eventsarenotexplainedbydeducingthemdirectlyfromalaw-thatananalysisoftheinitialconditionsisessentialbeforealawcanbesuccessfullyapplied-itwillbecomeapparentthattheremaybeagreatmanygeneralizationsuponwhichhistorianscananddorelyinexplainingevents. NOTESTOPAGES185-2039.SeeAppendixBformyargumentthatthelawlikeconnectionswithwhichHartandHonore,andalsoHume,wereconcernedarefundamentallydifferentfromanybasiclawscharacteristicofanadvancedscience.10.Onphasesandcomponentpartsastwodifferentbutcompatiblewaysoflookingatthestructureofcontinuingevents,seechapter5,section1.11.As-omewhatanalogoussituationistobefoundinthedevelopmentofthemuchyoungerdisciplineofculturalanthropology,thoughitstheorieshavetendedtoshiftmorerapidlythanhavethoseinhistory,whereevidenceconcerningparticularsocietieshasalwaysbeenmorereadilyavailable.Whenthisdifferenceistakenintoaccount,whatCliffordGeertzhassaidconcerninghisdisciplinecanbeusedinsubstantiationofmyclaimregardinghistory.Hesaid:"Ratherthanfollowingarisingcurveofcumulativefindings,culturalanalysisbreaksupintoadisconnectedyetcoherentsequenceof·bolderandboldersorties.Studiesdobuildonotherstudies,notinthesensethattheytakeupwheretheothersleaveoff,butinthesensethat,betterinformedandbetterconceptualized,theyplungemoredeeplyintothesamethings.Previouslydiscoveredfactsaremobilized,previouslydevelopedconceptsused,previouslyformulatedhypothesestriedout;butthemovementisnotfromalreadyproventheoremstonewlyprovedones,itisfromanawkwardfumblingforthemostelementaryunderstandingtoasupportedclaimthatonehasachievedthatandsurpassedit"(TheInterpretationofCulture[NewYork:BasicBooks,1973],p.25).Inhistories,thenewsortiesthataremostapttochangepreviousinterpretationsinvolveexaminingthesameeventsonadifferentscale,orexaminingotherfacetsofthesameevents.Whenthisisso,shiftsininterpretationdonotwarrantadenialofobjectivity.APPENDIXA1.Ihavedevelopedthisargumentatgreaterlength,andwithmoreattentiontodetail,in"TheDistinguishableandtheSeparable:ANoteonHumeandCausation,"JournaloftheHistoryofPhilosophy12(1974):242-47.2.IhaveexaminedatleastsomeoftheminthethirdchapterofPhilosophy,Science,andSense-Perception(Baltimore:JohnsHopkinsPress,1964).3.Seetheopeningparagraphofbk.1,pt.3,sec.4oftheTreatiseofHumanNature.4.A.Michotte,ThePerceptionofCausality(London:Methuen,1963).Inthisconnection.Ihaveparticularlyinmindhis"entraining"experiments-forexample,p.21,experiment2.APPENDIXB1.H.L.A.HartandA.M.Honore,CausationintheLa'w(Oxford:Claren-donPress,1959).2.CausationintheLaw,pp.8-<).3.CausationintheLaw,p.9.4.AmericansInterpretTheirCivilWar(Princeton:PrincetonUniversityPress,1954).5.Inmyopinion,whatfollowscastsseriousdoubtonthethesissupportedbyW.H.Drayinanarticleentitled"SomeCausalAccountsoftheAmericanCivilWar,"Daedalus91(1962):578-98.Thethesisofthatarticle,andmuchofthearticleitself,arealsotobefoundinchapter4ofDray'sPhilosophyofHistory(EnglewoodCliffs,N.J.:Prentice-Hall,1964). NOTESTOPAGE2046.Tobefair,Imustpointoutthatindistinguishingbetweentheexplanatoryandtheattributivecontextsinwhichcausaljudgmentsaremade,HartandHonoreacknowledgethatitissometimesnecessaryinthelawtoofferexplanatoryaccountsbeforeattributivejudgmentscanbemade(CausationintheLaw,pp.22-23).Further,inthesamepassagetheyadmitthatsuchpreliminaryinquiriesmaysometimesbedifficult,buttheyaddthat"suchsearchesforexplanationarenotthesourceofthelawyer'smainperplexities."Thissuggeststhatwhentheyspeakof"thelawyer"theyhaveinmindprimarilythejudge,andnottheinvestigativeofficersofthecourt.ThisshouldsuggesthowmisleadingitmaybetoviewhistoriansasiftheirinterestswereessentiallylikethosethatHartandHonorefindtobecharacteristicof"thelawyer." BLANKPAGE INDEX.'�'�.•....Asch,Solomon,E.,212n.12science,73-74,77-78,96-97;Hobbeson,81-82,84;andinBeard,CharlesA.,9,148,153,192,stantaneoussuccession,64-65;and206n.4objectivity,193;ofparticularBecker,CarlL.,9,206n.4eventsvs.necessaryconditions,Bergson,Henri,5782-83,86;perceptionof,53-55,Berlin,Isaiah,220n.3057-58,76-77(seealsoAppendixBlack,Max,99A);andprocesses,52�53,57-58,Blanshard,Brand,210n.169,78,93,94-95,139;andshortBrunner,Otto,12-13,17,18handexplanation,66-67,68,72,Burkhardt,Jacob,29,39,4077,78,84,94,96,97;andtemporalsequence,52,54-57,62,73-Causalbeliefs,validationof,77,21374,78,93-96n.21Causes:vs.accompanyingsigns,70-72;Causation:andeffectsofexperience,vs.conditions,67-70,81-97,137-62-63;ineverydaylifeandhistory,38;andeffects;asdistinct,53,62-140-42;ineverydaylifeand65;and·laws,distinguished,51, INDEXCauses(continued)ofindividuals,136-39;andwhole97-98,103,193;astypesofpartrelations,31events,65-66Explanatorystructure,26-27,28Chambers,Edmund,221n.32Externalinfluences,,114-15,127-28,Codrington,R.H.,213n.26177Cohen,MorrisR.,210n.1Coleman,J.S.,219n.19Facets:ofhistoricalfacts,16-18;andColligation,130,134,163-65objectivity,154-55;andrelativism,Collingwood,R.G.,8,24,118,148,44-45206n.4Fain,Haskell,208n.1Comparativehistory,14Fiction,distinguishedfromhistory,8-9Comte,Auguste,157,160Fogel,RobertW.,209n.9Conkin,PaulK.,209n.3Contingency,108Galileo,methodofresolution,99Correspondence:betweencauseandGallie,W.B.,24,111,209n.7,217effect,58,64;andmana,76;andn.5,219n.20underlyingcausalpower,59-62Gardiner,Patrick,111Court,W.H.B.,13Geertz,Clifford,206n.8,224n.11Croce,Benedetto,7,8,24,148,206Generalhistories:andspecialhistories,n.6distinguished,11,12-13,19;strucCulture,11-12tureof,30-33,111-14Generalizations:andexplanation,6-7,Dahl,RobertA.,218n.1632,33,119-26,178-85,193;asDanto,ArthurC.,24,111,209n.7,heuristic,123-36217n.5Generations,inhistory,173-74Davidson,Donald,214n.2Gerschenkron,Alexander,215n.7Definition,roleofinspecialhistories,Gestaltpsychology,5834-35,129-30Goldenweiser,A.A.,206n.3,210n.Deternlinism:inhistory,180;meaning12of,105-6Goudge,T.A.,211n.4Donagan,Alan,218n.14Gruner,Rolf,208n.1Dow,J.C.R.,158,159Dray,W.H.,24,28,111,178-79,208Hart,H.L.A.,oncausation,50,51,n.1,209n.7,2Ion.11,21752,65,138,199-204,216n.15,nn.2and5,224n.5224n.9;oncausesandconditions,Ducasse,CurtJ.,85,210n.1,216n.81,92,202,214n.313Hegel,G.W.F.,128,160Dummet,Michael,215n.11Heider,Fritz,211n.7Duncker,KarlF.,59,64,140,212n.Hempel,CarlG.,5,84,104,178,185,12213n.21,218n.15Durkheinl,Emile,214n.26Henderson,L.J.,216n.18Herskovits,MelvilleJ.,11,220n.29Elton,Oliver,209n.10Hirsh,E.D.,223n.1Ely,RichardG.,208n.1Historicalrelativism,44-45;argumentsEvans-Pritchard,E.E.,214n.26for,147,148-49,150,153,155,E�Ning,A.C.,210n.1161-62,170,178,190-93Explanation:inadvancedsciences,72-History:dataof,10-11;andfiction,75,98-100;ascontextdetermined,8-9,25,27-28;asidiographic,66-67,78-79;throughdescrip4-7;andtheindividual,13-14;tion,32;ingeneralhistory,118-andinquiry,7-9;andinstitutional28;initialconditionsin,103-4;instructures,115-17,126-27;andsequentialaccounts,177;shortmap-making,15-17,115;andhand,inhistory,176;inspecialmemory,7;asnarrative,24-25,histories,37-39,188-89;interms30-31;andscience,5-7,122-23; INDEXstructuralformsin,25-27,28-29,Malinowski,B.,217n.1230-33,43-44Mana,andcorrespondence,76Hobbes,Thomas,oncause,81-82,84,Map-makingandhistory,15-17,11594.Marc-Wogau,K.,211n.4Homans,GeorgeC.,125Marrett,R.R.,213n.26Honore,A.M.SeeHart,H.L.A.Marxiantheory,126,157,160,188,Hull,DavidL.,208n.14189Humannature,knowledgeof,andhis-Memory,nothistory,7tory,120-22.Methodologicalindividualism,136-37,Hume,David,oncausation,58,62-63,217n.1274,75,76,77,99,138,224n·9Meyerson,Emile,210n.1(seealsoAppendixA);mentioned,Michels,Robert,218n.1949,53,73,106,201,202,212n.Michotte,A.,58,77,140,198,211n.13,213n.187,212n.11,213nn.17and19Milanov,K.,208n.13Idiographicvs.nomotheticaims,5-7,Mill,JohnStuart,oncausesandcondi122-23tions,81,90,94,214nn.2,3,andImportance,concept·of:inhistory,35-436;andfacets,155;inspecialhisMiller,Hugh,215n.10tories,166-67Mink,Louis0.,111,209n.7,217n.Institutionalstructure,1865Interpretivestructure,28-30;andcolliMosca,Gaetano,218n.19gation,134;descriptive,notexplanatory,135;andperiodization,Nadel,S.F.,11135-36;relationof,tosequentialNagel,Ernest,210n.l1,216n.18,andexplanatorystructures,39-41;221n.2andspecialhistories,38-39;valNarrativistviewofhistory,24-25,30-idationof,41-43,4431,217n.5"Ironlawofoligarchy,"180-181,218Necessity,andcausalexplanation,105-n.198"Negativeconditions,"andcausalatKeynes,J.M.,158tributions,90-93Kluckhohn,Clyde,11Newman,FredD.,32Kohler,Wolfgang,214n.26Nomotheticaims,5-7,122-23Kracauer,Siegfried,208nn.13and15Kroeber,A.L.,11,201n.8Kuznets,SimonS.,223n.7Oakeshott,Michael,110,111,118,133,209n·7,215n.7,217n.5,Langlois,C.V.,113219n.21Laplace,PierreSimonde,105Objectivity:andcausalanalysis,193;Laws:inadvancedscience,201;applicandexplanatoryaccounts,177-78,abilityof,180-85;incausalex187-89;andfacets,172-74;inplanation,100-103generalhistoriesandspecialhisLeff,Gordon,209n.3tories,162-68,189-90;limitsof,Lehmann,F.R.,214n.26inspecialhistories,160-63;meanLewis,C.S.,132,220n.28ingsof,146-50;andpart-wholereLipset,S.M.,18n.19lations,175-76;andsequentialLouch,A.R.,208n.1histories,171-77;summarized,Lowie,R.H.,214n.26189-93;andtheoryinhistory,187-89McCullagh,C.Behan,208n.1Olafson,F.A.,138-39,208n.12McDougall,William,217n.12Order,amonghistoricaldata,112-14,Mackie,J.L.,50,51,211n.4,214n.116-17,165,170,171-72,175-3,216n.1976,186,189,191-93 INDEXPalmer,R.R.,4Specialhistories:descriptionandevaluPanofsky,Erwin,222n.15ationin,130-31;andexplanationPares,Richard,29of,36-39,131-32,139-40;andParsons,Talcott,207n.8generalhistories,connectionsbeParticularevents,vs.typeofevent,139tween,21;andgeneralhistories,Part-wholerelationship.SeeWhole-partcontrasted,33-35,132-34;andrelationshipgeneralhistories,distinguished,11,Pastness,andhistory,912-13,18-26;andinfluences,34-Periodization,22-23,135-36;ingen37eralandspecialhistories,165-66;Spencer,Herbert,157,160andobjectivity,173-75Spengler,Oswald,157Popper,Karl,220n.30,221n.3Spinoza,Benedict,102,105Power,Eileen,221n.4Spiritofanage,134,141-42Pratt,J.B.,210n.1Pressly,ThomasJ.,203Taylor,Richard,215n.11,216n.20Psychoanalytictheory,andhistory,218Theory:inhistory,6-7;meaningof,n.13156;andobjectivity,155-56,158-Psychology,andsocialinstitutions,11660,161,173;roleof,ingeneralhistory,155-60;roleof,inspecialReichenbach,Hans,216n.12histories,160-61Relativism,historical.SeeHistoricalToynbee,Arnold,128,157relativismTrevor-Roper,H.R.,206n.3Robinson,JamesHarvey,206n.4Trow,M.A.,218n.19Ruddick,William,214n.26Tufte,EdwardR.,218n.16Russell,Bertrand,96,210n.1,212n.Tylor,E.B.,11,12514,215n.9Vico,Giambattista,157Salmon,Wesley,216n.12vonWright,G.H.,208n.12,216n.Scale:inhistory,15,115;andobjectiv-11ity,151-54,andrelativism,44-45Schmookler,Jacob,220n.26Walsh,W.H.,24,164-65,217n.2Schneider,DavidM.,125Warren,Austin,210n.10Science:andhistory,5-7,122-23;andWelIek,Rene,210n.10particulars,139,200Whewell,William,129-30,163-64Scriven,Michael,57,111,211n.5,216White,MortonG.,24,30,111,217n.n.19,218n.145,222n.13Seignobos,Charles,113Whitehead,A.N.,210n.1Sequentialstructure,25,26,27-28;andWhole-partrelationship,31,186;inexplanation,118-19;andexplanageneralhistory,114-18,119,126-torystructure,185-8727;andobjectivity,175-76Shaw,A.G.L.,209n.9Wilson,F.P.,220n.32Simmel,Marianne,211n.7Windelband,Wilhelm,5Societyandculture,distinguished,11-"Worldhistory,"174-7512Sombart,Werner,157,158Young,G.M.,29,.39,40 AbouttheAuthorMauriceMandelbaumisAndrewW.MellonProfessorofPhilosophyattheJohnsHopkinsUniversity.HeistheauthorofTHEPROBLEMOFHISTORICALKNOVLEDGE:ANANSWERTORELATIVISM;PHILOSOPHY,SCIENCE,ANDSENSEPERCEPTION:HISTORICALANDCRITICALSTUDIES;THEPHENOMENOLOGYOFMORALEXPERIENCE;andHISTORY,MAN,ANDREASON:ASTUDYINNINETEENTH-CENTURYTHOUGHT. LibraryofCongressCataloginginPublicationDataMandelbaum,MauriceH1908-Theanatomyofhistoricalknowledge.1.History-Philosophy.2.Causation.1.Title.D16.9.M2690176-46945ISBN0-8018-1929-6