Bryant Evidence and Explanation in History and Sociology

8/3/2019 Bryant Evidence and Explanation in History and Sociology

1/18

Evidence and Explanation in History and Sociology: Critical Reflections on Goldthorpe'sCritique of Historical SociologyAuthor(s): Joseph M. BryantReviewed work(s):Source: The British Journal of Sociology, Vol. 45, No. 1 (Mar., 1994), pp. 3-19Published by: Blackwell Publishing on behalf of The London School of Economics and Political ScienceStable URL: http://www.jstor.org/stable/591521 .Accessed: 14/01/2012 09:53

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .http://www.jstor.org/page/info/about/policies/terms.jsp

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of

content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms

of scholarship. For more information about JSTOR, please contact [email protected].

Blackwell Publishing and The London School of Economics and Political Science are collaborating with

JSTOR to digitize, preserve and extend access to The British Journal of Sociology.

http://www.jstor.org
http://www.jstor.org/action/showPublisher?publisherCode=blackhttp://www.jstor.org/action/showPublisher?publisherCode=lonschoolhttp://www.jstor.org/stable/591521?origin=JSTOR-pdfhttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/stable/591521?origin=JSTOR-pdfhttp://www.jstor.org/action/showPublisher?publisherCode=lonschoolhttp://www.jstor.org/action/showPublisher?publisherCode=black


2/18

Joseph M. Bryant

Evidenceand explanation n historyandsociology:criticalreflectionsonGoldthorpe's ritiqueof historical ociology*

In 'The uses of history n sociology',John Goldthorpe 1991) raisesseveralseriousobjections o the currentrevivalof historicalmodes ofanalysis n our discipline.This critiquecomes at a most opportunetime, for advocatesof historical ociologyhave hithertoconcentratedmore on exposing the inadequaciesof mainstreampractices andassumptions - an understandableemphasis given their marginalstatus- and hence have not offered a systematicor comprehensivejustification of their own procedures. Goldthorpe's direct andinformed challenge should, accordingly,occasion a more advancedexchange on matters of theory and method, one which holds thepromiseof greateranalytical ophistication or all partiesconcerned.What ollows s a preliminary nd partial esponse o certainaspectsofGoldthorpe's ssessment,my objectives eing to (a)redraw he linesofcontention in accordancewith actual problems of investigation,asopposed to traditionaldisciplinarypreserves;and (b) reaffirm thenecessary nterdependenceof historical nd sociological nalysis.I. ON EVIDENCE, HISTORICALAND SOCIOLOGICALWhile acknowledging hat a continuingdialogue between historiansand sociologists s essential for both disciplines,Goldthorpe is dis-turbed by recent interdisciplinaryrends and by more extreme callsfor an amalgamation f explanatory ogics.These efforts - rooted inthe 'reaction gainstpositivism' nd a 'failureof nerve'over the idea of'social cience'- have gone too far in his view, llegitimately verridingaspects of the traditional diographic-nomothetic istinctionwhichshould be maintained Goldthorpe1991:212 - from here on all pagereferences o this paperwillappear n squarebrackets.). n brief com-pass, Goldthorpe reassertsthat historiansmust contextualize,withdatesand places,whereas ociologists hould strive o widen the scopeof their explanatoryarguments. That proposed division of laboursuggestsone obvious unctionfor history n sociological nalysisBJS Volumeno. 45 *ssucno. l March l 994


3/18

JosephM. BryantHistorymayserveas,so to speak,a 'residual ategory'orsociology,marking the point at which sociologists,in invoking 'history',therebycurb their impulse to generaliseor, in other words, toexplainsociologically, nd accepttherole of the specificandof thecontingentas framing- that s,as providingboth thesettingarmlhelimits of theirown analyses p. 212].

Goldthorpe's bjectivewould thusseemto entaila returnto thestatusquo nte, pointwhere sociologistsandhistorianswouldrestrainandcorrecteachother'snaturalexcessesor limitations,heone impartingthegift of theory,of generalization,heotheryieldingconcretedetailand bounded specificities.Given their respective research foci,Goldthorpeendorses the view that the two disciplineswill employdistinctive xplanatoryogics,even if'generalizing'and 'particulariz-ing'are mattersmore of emphasis hanstrictprinciple.Differentexplicanda,owever,are neitherthe sole nor the majorrationale or upholdingthe traditionalacademicdivide.Muchmoreimportant n Goldthorpe's iewis the fact thatthe twodisciplinesarefoundedupon distinctivempirzcalases,eaturingdifferencesbothinthe nature of the evidence utilized and in the manner in whichevidencecomesintobeing.The basiccontrasts presentedasfollows:whereasmodern sociologicalresearchtypically generates'or 'pro-duces' its own data or evidence, primarily hrough interviewsandsurveys,historical videnceconsistsof chancesurvivals r 'relics'romthe past. The immensedisadvantagesand problems posed by thelattercircumstance,Goldthorpecontends,havebeen largely gnoredby proponentsof historical ociology,resultingnot only in mistakenviewson the relationshipbetweenhistoryand sociology,but also inmisapplicationsf historicalmaterialsn sociologicalpractice.

Assessing he validity f Goldthorpe'sritiquewillobviouslyurnonthe adequacy of his distinction between 'sociologicaldata' and'historicalelics'.His startingpointis uncontroversial:a historicalactis an inference from the relics'[p.213]. Nor can one challengetheobservationthat historicalrelics - essentiallymaterialartifactsofvariouskindsanda wide rangeof literarydocuments-are both finiteand incomplete.Equally ound is the suggestionthat, for any givenperiod or setting,the survivingrelicsare likelyto constituteonly asmalland unrepresentative election.Giventhese limitingcircum-stances,historicalknowledgewillinvariablyprove to be both partialandlacunary, estricted o the inferences hatcan be drawnfrom thefragmentary emainsof the past.A reminder that historiography s a demanding enterprise issalutary n itself;but Goldthorpe'sdiscussion mpliesrathermore,that the evidentialbase upon which historicalresearchdepends isdistinctivelyproblematic,withthe consequencethat'historicalacts'are inordinately nferential.Is this characterizationustainableas a


4/18

Evidence ndexplanationn historyndsociology 5category distinction, marking off disciplines, or should we ratherdecide such matterson the basisof substantive oncernsand particu-lar questionspursued?Finitudeand incompleteness, fter all, seem tobe characteristic f most forms of scientific nquiry.The exhaustiverepresentation r recountingof 'reality's not the objective r practiceof any empirical science, and all confront, in varying degrees,limitations n sourcematerials, ither as actualdeficiencies n 'data'oras cognitiveor technical imits n accessing'all hat there is'.2Moretothe point, these limitationsvary in accordancewith the problemsinvestigated:a biologist studying cell division in garden vegetablesconfronts an empiricalsituation markedlydifferent from that of abiologist investigatingthe transmissionof the AIDS virus or themolecular mutations that promote cancer malignancy;a physicistconcerned with the properties of subatomicelements confronts anempirical situtation different from that of a physicist studyingchemicalkineticsor interstellarmagnetism; he historianwho investi-gates the causesof WorldWar I confrontsa researchchallengequitedistinct from that of a historian whose object is to identify thediplomatic maneuvers that resulted in the Treaty of Versailles;asociologist concerned with fertility patterns among developingnations confronts an empirical situation different from that of asociologist nvestigating ies betweencorporateand political lites orfor that matter,quitedifferent from a demographer tudying ertilityin developednations!If the finitudeand incompleteness f data sets are to be assessedonthe basisof substantive roblemsrather han disciplinary oundaries,the same principleshould hold with regard to the issue of inference.Science s of coursean inferential nterpriseat its verycore, in thatthephenomenal realm provides an experientialor empiricalbase fromwhichwe 'deduce'or 'infer' ertainproperties,patterns,and relations.In actual practice,this quest for intelligibility ypically nvolves anongoing dialectic between observation and conceptualization, n-duction and deduction: heories,concepts,and facts are all mutuallyimplicated.The qualityof scientific nferenceswill thus varywith thequality of the availableevidence, the existing technology of obser-vation and measurement,and the analytical ogency of the conceptsand theories employed; 'strong' and 'weak'areas of research willaccordinglybe found both acrossand withindisciplines.The conven-tionalviewthatthe sciences tand n a definitehierarchy wherebyallthat is carried out within the domains of physics or chemistry isdeemed more accurateand established han anythingthat has beenproduced in sociology or history - that view is quite misleading.Indeed, there are some forms of knowledge in the social-historicalsciences which are more secure than select areas within the naturalsciences.3Nor should it be overlooked that whatever the field ofinvestigation,the logic of inferential procedure remains basically


5/18

6 JosephM. Bryantuniform. The sociologistwho correlatescrime statisticsand sundrysocio-economic ndicators n order to postulatesome form of causalconnection s engaged in basically he same analyticalprocessas thearchaeologistwho infers widening stratification n the basis of thechanging composition of burial remains and signs of increasedagricultural roductivity nd monumentalarchitecture.Our remaining concern over the nature of historical evidencewould seem to be the issue of 'representativeness',nd this is indeedthe main focus of Goldthorpe'scritique. Owing to the distinctivemanner in which historical evidence 'comes into being', i.e., assurvivals r 'relics' rom the past, the historian ypically ncountersatwo-fold imitation:not only is the supply of extant materialspartialand fragmentary, ut design and sampling ontrolsare usuallyabsentgiven the haphazardand uneven circumstances y which 'relics'arepreserved. Determiningwhether a particular nscription,poem, orpaintingconveysrepresentative r aberrant nformation s, to say theleast, anything but straightforward.And even if we allow thathistorianshave worked out methods and inferentialrules for suchproblems usually ubsumedunder the broadrubricof Quellenkrztikr'source-criticism'),here is no denyingthe difficulties nvolved.

For Goldthorpe, his is the difference that divides:for unlike thehistorian, the sociologist is in a position to 'generate' or'invent'evidence,withattendingcontrolsover samplingand selection.Depen-dent upon what the past has bequeathed, he historian s constrainedto follow in pre-determinedpaths,a prisonerof the vestiges ime haspreserved.The sociologist,by wayof contrast,can chartan indepen-dent courseof research, ramingquestionsand then actively nterven-ing to create materials thatdid not existbefore' p.214]. And whileGoldthorpeoffers passingacknowledgmenthat generated vidence'poses ts own specialproblemsas regardscompleteness, eliability ndvalidity,he is confidentof the analytical dvantages hatobtain

where the nature and extent of available vidence is not restrictedby the mere accidentsof physicalsurvival;where, moreover, thecollectionof evidence can be 'designed'so as to meet the specificrequirementsof the inquiry n hand; and where questionsof thequality of evidence can always be addressed, as they arise, bygeneratingyet furtherevidence hroughwhich o checkand testtheoriginal pp. 21X15].The distinctionbetween research that investigates the past' andresearchconcernedwith 'the present' s a familiarone, and obviouslysubsumesmore than the disciplinesof historiography nd sociology.It is, moreover, a distinction ill-suited to serve as a criterion fordisciplinary emarcation, or withinevery empirical cience,physicalas well as social,one willfind a vastarrayof questionsand problemstheoretical s wellas substantive-that are intrinsicallyhistorical',.e.,


6/18

Evidence ndexplanationnhistoryndsociology 7concerned with processes and events that are both sequentially-ordered and time-dependent.4That said, it is undoubtedly rue thataccessing the past and accessing the present pose distinctivechal-lenges, and that,accordingly, ertainpeculiarities f procedurewillmarkthe two orientations.The question is, are we to presume aninherentsuperiorityof one to the other, or is it not the case that eachwill feature advantagesand disadvantages,withthese in turn beingassessed on the basis of specific research concerns? Proceedingcategoricallyrather than substantively,Goldthorpe'sanswer takesthe form of a decidedlyuneven 'balance heet',with all methodologi-calcreditsaccruingon the sociological ide of the ledger, principallyon the ground that surveytechniquesgenerate'data' hat are morerepresentativeand complete than the 'relics'which constitute thelimitedempirical und of historiography.5At first glance, Goldthorpe'saccountingseems plausible, almostaxiomatic.After all, whathistorianwould notjump at the chance toreturnto the past armedwith questionnaires or mass-mailing ndaudio-videorecordersof unlimited ape?The obviousanswer o thathypothetical s not conclusive,however, for the analogous situationalso holds: what sociologistwould not be keen to view the presentfrom the vantageof the future, withretrospectiveknowledgeof thetrendsthat emerged triumphantand those that fell by the wayside(often contraryto the subjectiveexpectationsof the participants!),and with the opportunity o peruse official and privaterecords thatweresealed fromcontemporaries?Goldthorpe'snvidiousdistinctionbetween 'sociologicaldata'and 'historical elics' s likewisemislead-ing,not so much for what tsays,but for what tleavesout of account.Admittedly, here can be few who wouldquestionthat investigatingthe present provides distinct advantages in matters of researchdesign and sampling. But 'representativeness'as a dual meaning:'typification'n a samplingor probabalisticense is certainlyencom-passed,but there is also a 'constitutive' imension, .e., the notion ofcorresponding to or embodying what is essential, what is 'real'.Goldthorpe's ontrastbetween evidence'discovered' historical) ndevidence 'invented'(sociological)attends only to the issue of sam-plingrepresentativeness, ypassing ubstantive ontentaltogether.Areassessment n that scoreis therefore urgentlyrequired.6To begin with the mostfundamentalpoint,we need to appreciatethat historical 'relics' are, for the most part, naturalor authenticelementsf pastsocialworlds.That is to say,human beings orient theirlives amidst the architectural orms that frame and define theirprivate and public spaces;they fashion their motives, ideals, andinspirations n and fromthe songs, poems, and prose that constitutetheirprimarymeans of communicativediscourse; hey projecttheirheartfelt hopes and anxieties onto the funerarymotifs which gracethe graves of the departed; they objectifyand preserve significant


7/18

8 JosephM. Bryantcollectiveexperiences n the socialritualswhichregulate he rhythmsof community ife; they administer heir publicaffairson the basisofcreatedpolitical nstitutions nd codified egal decrees; heir relation-ships with other communities and with nature are mediated byinvented nstruments f warand economy, .e., the panoplyof armourand weaponrywieldedby the warriorand the inventoryof tools andmachines mployedby the worker.Regarding he otherbasic ormsofhistorical evidence, much the same can of course be said: frompotsherds and pollen deposits to trash heaps and bureaucraticinventories all such 'relics' epresenteitherthe meaningfulcreationsor by-productsof social activity,or the bio-physicalrealitieswhichcircumscribedhuman experience. As residual 'traces'or 'objectifi-cations' f pasthumanactionsand existential onditions, he empiricalmaterialsof historiography ring us into rather mmediateor directcontactwith our subjects, theirworldsas experienced',and that kindof authenticity the 'hardness'of such data, one might say- mustsurelyrankhigh in the desiderataof any scientific nterprise.7Against the 'naturalness' r 'social authenticity'of historicalevi-dence, it is necessary o consider the 'artificiality'f much standardsociologicaldata. Goldthorpe'scharacterization f survey or inter-view-generated videnceas 'invented' s unintentionally pposite, orit is clear that research of that sort typically entails significantmodifications r deformationsof the subject nvestigated.Not only isthat kind of'invented' information fashioned and extracted in asomewhat stilted and artificialmode (e.g., responses mechanicallyscaled and transcribed n quantitative orm), and in an unnaturalsetting (i.e., the promptedresponseact is not partof the normal lowof social life), but the intrusion - whether as questionnaireor asinterviewer is recognizedas suchby the subject, husconditioningor'contaminating' he response (the so-called 'reactivemeasurement'problem).What must be confronted here are obvious parallelswiththe UncertaintyPrinciple n particlephysics,whereby nterventionistefforts to measure reality invariablyoccasion significantdisplace-ments or distortions n that reality.The frequentlyheard lamentthat'real'human beings and 'real'socialprocessesare obscuredor lost inthe trappings of modern research techniques and methods (orperhaps never found!) attests to the very real problem f trying oapprehendhe elusiveand highlyreactive ealities f social life throughprocedureshatareboth nnatural ndobtrusive.8Even apartfrom the problemsof artificialitywith invented'data, tneeds to be stressed hat surveysand fieldwork annotbe consideredfully representativeeven in the sense Goldthorpeemphasizes,forthese formats are severely limited in their capacityto lay hold ofcertain features of social totality. As regards survey methods, thecoded and statistically rocessed elf-declared ttitudesand actionsofindividualsyield a rather inadequateprism through which to view


8/18

Evidence ndexplanationn historyndsociology 9developmentsand transformationson the macro-structuralcale;whileparticipant bserversare for the most partphysically estrictedto workingin 'micro'settingsor other circumscribedmilieux.9Theinformationor datagatheredby suchmeans willaccordingly onsti-tute onlya limitedand partialrepresentation f social ife- useful tobe surefor various pecificresearchquestions,buthardlyadequateasthe empiricalbasisfor socialscience n general.10Anothercharacteristicf historical videncerelevant o the issue ofsubstantiverepresentativenesss what might be called its socially'loaded'or 'multivalent' uality, .e., the fact that,as an objectificationof humanactionor experience,the 'relic'or historical bject ypicallyembodiesor is stampedby variousaspectsof thatcontext.A particularpoem or text of legal discourse,for example,will articulatecertainvaluesand ideals,but in additioncarrya descriptionof dietaryhabits,residentialpatterns,and fashionsof dress; a particularpaintingwilldisclosesomethingof the aesthetictastesof artistsand patrons,butalso provide a window onto sundryroutinesof everydaylife; thematerials nd methodsusedin the construction f wallsandbuildingswill revealmuchregarding he economicand technical ircumstancesunder which the inhabitants ived, but also yield informationongeo-politicalconcerns and realities; unerary nscriptionsand gravedepositswill bear testimonyto the statusclaimsof the interred,butalso expresssomethingof thereligiousand cultural ensibilities f theperiod; a bronze cuirass and helmet will enable us to recreateconditionson the field of battle (includingthe physiologyof thecombatants),but in additiontestify to levels of craftsmanshipandwiderlinkagesof commerceand trade.In short,by subjecting uchdisparatematerialso analytical olligation,historians re oftenabletodrawout from a singlesourcea wide rangeof insightsandclues aboutimplicated ocialprocessesand relations preciselyecausehat videnceis 'natural',.e., anintegral rcomponentlementf theactions, outines,ndexistentialxperiencesf human eingsntheir wn ocialworlds.Here again a comparisonwith 'invented'sociologicaldata wouldseem to run counter to Goldthorpe'sone-sided assessment.Thecommonplacecharge that much of the information supplied bysurveyresearch s 'thin'or trivial(especiallyn regard to behavioralaspects)reflectsthe fact that,given the problemsof artificiality ndunnaturalnessdescribedabove,thereare groundsfor doubtingboththe validityand the significanceof the abstractedresponses.As atemplate of the social world, the research-inducedpsychologicalreactionsof individuals o a batteryof formalquestionsmustbe seenas quite limited,all the more so when methodologicalpressuresforstandardizationnd mass-processing'flattenut'or 'pre-package'hequalityand rangeof response.ll In contrast o historical elicswhichare typicallymultivalent, .e., 'saturated'n the social, there existsavery real possibility hat 'invented'sociologicaldata captureslittle


9/18


10/18

Evidence ndexplanationnhisto andsociology llimmersed, and the field within which they become intelligible'(1953:27-28). For a political party that repeatedly loses generalelections, each new campaign and subsequent oss is not a homo-geneous datum,but areality hat is overladenor 'surcharged' y whattranspiredbefore. For an individualcontractinga second or thirdmarital(or adulterous)relationship, he preceding experiences willbear significantlyon the present and future. The act of protestingagainstan authoritarian egime- its verymeaningand significancewill be defined by its occurrenceon the sociological ime-curve, tscontextual embeddednessin the precedinghistory of state-civilianrelations.The attitudeof consumers owardthe productsof certaincompanies s not simplyafunctionof currentcost-benefit alculations,butalso encompassesperceptionsof pastserviceand reliability.Thecourseof revolutionary pheavalortransformations conditionednotonly by the present constellationof social forces, but by the pastexperiencesof the participants, heir sense of the 'historicalmoment'and their appreciationof the collective destiny'of the nation. Thevariable hythmsof experiential ime thusenterinto all facetsof sociallife as a determinant eature, extending from the personalor microlevelon up to that of institutionsand macrostructures.l3Geneticorhistorical modes of explanation are accordingly indispensable tosociologicalanalysis,giventhat temporality- not as a homogeneousmetricbut as a culturallydefinedapperception- providesone of theessential ramesof meaningfor socialaction.l4The time-as-measurement iew, with historyproviding boundaryconditions'for sociologicalgeneralization, s likewise nadequatetocapture anotherfundamentalaspect of the ontology of social phen-omena: the fact thatpresent rrangementsinstitutions,oles,culturalforms are theproducts fpasthuman ctions.The 'past' s thus neverreally 'past', but continuously constitutive of the 'present', as acumulativelyand selectivelyreproducedensemble of practicesandideas that 'channel' and impart directionality o ongoing humanagency.The present, nother words, s whatthepast as receivedandcreatively nterpretedby the present- has made it, with the conse-quence that social scienceinvariably onfrontssituationsof 'layered'or 'ramified causality', .e., chains of dependence that repeatedlyrecede into precedingconstellationsof social factors.l5An investi-gationinto the 'causes'of gender inequalities n the workplace, orexample, cannot simplydraw up an inventoryof current attributes,interests,and powerrelations; t will also be necessary o show howpresent circumstances re the resultant moments'of a long chainoflinkedand contingentantecedents,a 'trajectory' r 'path' hat, in thisparticular ase, begins no later than the dawnof complexcivilization.History,as a phenomenon, s 'the cumulative ffect of past events oneventsof the present'(Tilly 1981:12). A sociology that ignores thepath-dependentnature or sequentialorderingof social phenomena


11/18

12 JosephM. Bryantis, in truth, no sociologyat all, for it fails to appreciate hat 'presentforms have their particularnature by virtueof their past' (Manicas1987:274).l6The dual temporalityof social phenomena- experiential ime asone of the dimensional ramesfor intentionality, nd the ontologicalrootednessof presentarrangements n pastpractices wouldseem tonecessitate historical ocialscience',and therewitha transcending fthe old idiographic-nomotheticntinomy. nsteadof artificiallyepar-ating the particular rom the general,the unique from the recurrent,the conjunctural rom the structured, hese qualitiesmust be expli-cated in terms of their ontological nterconnection, heir relationalimmanence. To assign one discipline the task of understandinghumanityby way of spatio-temporal pecifications,while enjoininganother to do so via generalizations nd abstractions, learlyviolatesthe essential unities and interrelationsof social life. Historicalso-ciology, in contrast,proceeds n conformitywith its object,by fusingidiographicand nomothetic modalitiesthrough a contextualogzc nwhich henomenareexplicatedndunderstoody racing oth heir enesisandtheir ntrtnsicelationso othermediatingtructuresndprocesses. hiscontextual logic is at once sociological, n that it attends to roles,institutions,and structures,and historical, n that it comprehendshuman agency in all its various forms as temporallyordered andconditioned. Those who would challenge either the logic or thenecessity of historical-sociological xplanation accordingly face aformidable ask, or rather han nvokedated methodologicalmaxims- canons that were in part formulated as self-legitimizinganddomain-preservingmechanisms- they are obligated o demonstratethat human actions are neither inherently social nor inherentlyhistorical. aditquaestio.III. A NOTE ON 'GRAND'HISTORICALSOCIOLOGYAND THE USEAND ABUSE OF SECONDARYSOURCESThe closingsectionof Goldthorpe's ritique s in my opinion the mostcogently reasoned and significant. His basic point is that 'grandhistorical ociology', .e., that genre of research eaturing the tracingout of long-termdevelopmentalprocessesor patternsor the makingof comparisonsacross a wide range of historicalsocieties or evencivilizations',s typically onstructed n the basisof extremely tenuousand arbitrary' inks between evidence and inference [pp. 220-22].Given their expansive analyticalrange, grand historical ociologistsare heavilydependentupon derivative r secondary ccounts or theirbasic data, a position which Goldthorpecontends reduces them tooffering 'interpretationsof interpretations'.More strikingly andironically,Goldthorpemaintains hat this sociological eprocessing f


12/18

Evidence ndexpkznationn histo andsociology 13historical rawmaterials's virtually denticalwith HerbertSpencer'spositivistic onception,whereinhistory s to provide stonesand bricks'for the vaultingarchitectural esignsof sociologyGrand historicalsociologists have to treat the facts, or indeedconcatenationsof facts or entire 'accounts', that they find insecondary ourcesas f theywereelativelydiscreteand stableentitiesthat can be 'excerpted'and then brought together in order thatsome largerdesignmaybe realized p. 221].Goldthorpegoes on to point out thatin contrast o secondaryanalysisof survey-based esearch,wherein primarydata and originalques-tionnairesare reanalyzed,grand historical ociologistsdo not makecontact with primarysource materials,but simply offer reinterpre-tationsof the interpretations fferedby historians.The charge thatgrand historical ociologistsare engaged in 'scis-sors-and-paste' istoriography s obviouslya serious one, and it fo-cuses much-neededattentionon a methodologicalproblemthat hasnot been adequately ddressed.An annotatedbibliography, uitesim-ply, is not enoughto sustainhistorical-comparativenquiry,and it is tobe hoped that Goldthorpe'spointedchallengewillelicit fullerdisclos-ure of analyticalprocedures rom practicinghistorical ociologistsnthe future.l7That said, a few observationswill be offered here in aneffort to show that the situation s not as 'impossibly oose'as Gold-thorpemakesout,and thatsociologistswho employsecondary ourcesare not nearlyso far 'removed rom theempirical' s is suggested.What Goldthorpe'saccount overlooks s that all worksof histori-ography are wovenfrom two distinguishabletrands:whatmight becalled reportagen the one hand, and interpretationn the other.Reportageconsistsof information hatpertains o basicquestionsofwhat,where,when,who, how many,etc.As an exampleonemightcitean episcopal etterfrom the mid-thirdcenturywhich relatesthat thechurch in Rome was then maintaining46 prebyters,7 deacons, 7sub-deacons,42 acolytes,readers, and doorkeepers,exorciststo thenumber of 52, and more than 1500 widows and poor people.'8Interpretation nvolvesestablishing he meaningand the significanceof these historicalfacts', .e., the materials hat constitutereportage.As to the famous letter just mentioned,historianshave long beenwaging various'interpretive'wars over, for example, what can beinferred regardingthe ecclesiastical tructureof the third-centurychurch, the socialcompositionof believers, he probablepercentageof Christians is-a-vishe total populationof the RomanEmpire,andso on. Now while t is true thatthe intricatenarrativeweavingsof mostworksof historiographyntegrateandcollate the 'facts'of reportagewithin interpretative tylings, t is usuallypossible for the informedreader to distinguishbetween the two- particularlywheneverothersecondary ourcesare consulted or comparative urposes.Historical


13/18

JosephM. Bryant4sociologists,grandor otherwise, husencounterprimarymaterialsnthereportageof historians, ndthoughthisformof mediatedaccess slimited - most seriously,one is dependent upon someone else'sselectionof the'facts' itis incorrect oclaim hathistoricalociologysa second-orderconstructionwithoutempiricalgrounding.l9Sociolo-gists should of course examine primarysourcesdirectlywheneverpossible;butgiventhatmanysignificantociological uestions equireextensiveknowledgeof different times and places,a cautiousandcritical elianceuponthereportageandinterpretationsf specialistssobviously ssential oradvancesn suchareas.20Goldthorpe'shargethattheofferingsof grandhistoricalociologyareinherentlyarbitrarys baseduponhisperception hatsuchworksaretenuouslyconnectedwithprimary vidence a viewwecorrectedin light of the distinctionbetween historiographic eportageandinterpretation.There is, however,anotherpoint to consider.WhenGoldthorpedoubtingly nquiresof the criteriaby whichsociologistscan possiblyreinterpretor adjudicatebetweenconflicting econdaryaccounts, heobviousanswer s,bymeansof theoryandcomparativeevidence.The disparate actsof reportagedo not 'speakfor them-selves',butarerendered ntelligible,meaningful,onlywhensetwithinsome kind of interpretiveor analyticat ramework.How well these'facts'are accountedfor, the scope and internalconsistencyof theanalysis, he compatibility f an interpretationwiththatestablishedfor comparable ases,all thiswillproverelevant n the evaluationofany cientific xplanation.Goldthorpehimself nvokessuchcriterianhiscriticisms f severalprominenthistorical ociologists, lleging hatsundry heoretical ndideologicalbiaseshavedistorted heiranalyses,or that aterrevisionist fferingshaveundermined heevidenceuponwhichthey indirectlyrelied.And for a definitivedisconfirmation fthe arbitrariness harge, one need only examine the many criticalexchangesamongsthistoricalociologistshemselves,wherequestionsof empiricaland theoreticaladequacyare continuouslyraised.2lGrandhistoricalsociologymay not have established fixedrules'toGoldthorpe's iking governingthe utilizationof secondarysourcesbut generalscientificstandardsor criteriado currentlyfunctiontodetermine the respectivestrengthsand inadequaciesof publishedresearch. ndeed,a casecouldbe madethatinternalcriticism ndthepossibility f knowledgecumulationarebothmoreeasiblen historicalsociology, given the stabilityof the 'facts'of reportage and theavoidanceof reactive-measurementnddesign-variabilityroblems.IV. CONCLUSIONThis paperhassoughtto establish hreefundamentalpoints:(1)thathistoricalevidence is characterizedby certainintrinsicstrengthsor


14/18

Evidence ndexpkznationn histo andsociology 15advantages - most notably the qualities of social authenticity or'naturalness' nd informational multivalence';2) that the temporalnature of social phenomena mandates a fusing of historicalandsociologicalmodes of analysis, .e., the adoptionof a contextual ogicof explanation;and (3) that the employmentof secondaryhistoricalsources - while entailing certain limitations does allow for non-arbitraryinkagesbetweenevidenceand interpretation. n addition, thas been suggestedthat researchproceduresmust be adaptedto thetypes of problems investigated; neither traditional disciplinaryboundariesnor peremptorymethodological ormats should delimit

, . . # . .t ze range ot sclentlhc nqulry.The case against historical ociology, I offer, has not been made.Goldthorpe's ariousstricturesmischaracterizehe natureof histori-cal evidenceand misconstrue he logic of sociological xplanation.Asa call to greatermethodological elf-awareness, is cautionary s to bewelcomedand endorsed;as a call to halt or reverse he trend towardsinterdisciplinary ynthesisand the consolidationof historicalsocialscience, t must, I submit,be rejected.The caseforhistorical ociology s still n the making.But it canclaimone manifest advantage: its analyticalprocedures have not been

modelled after the idealizations nd elisions characteristic f formaltreatiseson theory and method, nor have they been imported fromdisciplines that investigate fundamentallyalien ontological orders.Historical ocialscience s a 'grounded' cience, n that it proceedsbycomprehending he distinctiveand essentialpropertiesof its object:human agency as mediated by the constitutivecontextualframes ofhistorical ime and culturalmilieu.All methodological onsiderationsshouldbe similarly rounded.(Dateaccepted:July 992) JosephM.BryantDepartment f Sociology,University f New Brunswick

NOTES* I would ike to thankIrvingZeitlin,RandallCollins,Rod Nelson, Noel Iver-son, Jim Richardson,Allan Macdonnell,Marc Milner, Steven Turner, and JackGoldstone or theirhelpfulcomments nearlierversionsof this paper.Fellowshipsupport from the Social Sciences andHumanitiesResearchCouncilof Canadais alsogratefully cknowledged.1. But see the annotated bibli-ography n Theda Skocpol, ed.), V"ionand Methodn Historical ociology1984);

and of course the methodologically x-plicit offerings of Charles Tilly, mostnotablyAs SociologfMeetsHistory 1981).2. Jennifer Platt's 'Evidence andProof in DocumentaryResearch' 1981),provides an excellent overview of thechallenges involved, some unique tohistoricalresearch,but most shared byother standard sociological methods.Goldthorpe'spaper would have gained'balance'by engagement with this dis-cussion,and from consideration f some


15/18

16 JosephM. Bryantof the commonphenomenological ilem-masraised n PaulRock'sSomeproblemsof interpretive istoriography'1976).3. I think t safe to say, for example,that we know more about the originsofcivilizationhanwe do about heoriginsofthe universe,more aboutthe causesandprocessesof socialrevolution han aboutthe formation f galaxies r the transitionfrom norganic o organic,moreabout hedialecticof social structureand humanagency han about the paradoxof wave-particleduality.For a general analyticaldiscussion, eethe mportantessayyFritzMachlup, Are the SocialScienceReallyInferior?' 1963).4. One is reminded here of Marx'sfamousand - in lightof current rendsprescient emark n The German deology:'We know only a single science, thescienceof history.History anbe contem-plated from two sides, it can be dividedinto the historyof natureand the historyof mankind.However, he two sides arenot to be dividedoff; as long as menexist,the historyof natureand the historyofmen are mutually conditioned'. Thetrend I am referring o of course is therecent surge of interest in non-lineardynamics,system turbulence,hysteresiseffects, and contingenciesthat is sub-sumed under the general heading of'chaos heory',and extendsfrom physicsto biology and beyond. This broadintellectual urrentcan perhapsbest becharacterized s 'bringingHistorybackin' to the natural ciences.5. The conventional mage that his-torians confront a chronic dearth ofmaterials s, let it be noted, typically arfrom the actualities f problem-centredresearch. To be sure, extensive 'darkages'anddimly llumined spects f sociallife precludeany comprehensive econ-struction, but the sheer volume andvarietyof materialsbequeathedby thepast is not only unassimilableby thehistorical profession, but even mostspecialistsind it difficult o mastermorethanselectedareas n their ields.Not 'toolittle data', but 'too few fellow-workersand too little time' is a more realisticcharacterizationf the historian's light.Goldthorpe ppears o havebeen undulyinfluenced here by the partisan dero-

gation of mainstream historiographyofferedby quantitativecliometricians'.6. Whilemodernsociologyundoubt-edlyenjoysadvantagesn research esignandsampling n theory,practices usuallyless impressive.The followingdatum isparticularlyworth noting, for it bearsdirectlyon many of the most basicandimportant of sociological questions.Whereas historians are typically con-strained o workwith artifacts nd liter-arydocuments hatderive romthe com-mands and reflectionsof the privileged

and powerful, .e., thosewhoseresourcesenabledthem to dominate he mediaofcultural xpression andso the 'historicalrecord'); urvey researchers re usuallyhamperedby the oppositeproblem, hatof gaining access to the commandingheightsof their societies,with the conse-quence hatrandom amples f 'ordinary'respondents loom disproportionatelylarge in most sociologicalprojects. Ifpower is at all relevant o social reality,then 'selective ccess'wouldseem to be arathermoreconsequential roblem han'selectivedeposit'. The broad push torewrite history rombelow'by attendingto the marginalized xpressions nd dis-courseof variousdisprivileged roups-the lower classes, women, religious orethnicminorities, he conquered nd thecolonized is of course ntended o comp-lement and correct the perspectival/ideologicalbiases of more conventionalsources.7. For those who wouldquestionmyuse of the term immediate' ere- on theground that the analystmust interpretthe meaningand significance f the his-torical bject, hereby reating mediated'information-the point I am trying tomake is that 'relics'are aspects of thenatural environment of past socialworlds,unlike he artificially-inducede-sponses of survey research (which ofcoursealso require nterpretation!). is-torical interpretations, n other words,are interpretationsof objectifications;sociologicalnterpretationsf surveydataare one-step removed, being interpre-tationsof the respondent's ubjectiven-terpretations commonly forced' ntopre-selectedcategories of his or heractionsand orientations.


16/18

Evidence ndexplanationn historyndsociology 178. As documented n the classic tudyby Eugene Webb, et al., Unobtrusive

Measures:Interviews nd questionnairesintrude as a foreign element into thesocial setting they would describe,theycreate as well as measureattitudes, heyelicit atypicalroles and responses,theyare limited to those who are accessibleand will cooperate, and the responsesobtainedare produced n partby dimen-sionsof individualdifferences rrelevantto the topic at hand' (1966:1). Athorough review of these and relatedproblems nvolvedcan also be found inAaron Cicourel,Methodand Measurementin Sociology 1964). See also PierreBour-dieu, Jean-Claude Chamboredon,andJean-ClaudePasserson,The Craft of So-ciology: Epistemological Preliminartes(1991).9. See, most famously, C. WrightMills'critiqueof 'abstracted mpiricism'and the attendantdangersof 'psycholo-gism', .e., 'the tendency o reducesocio-logicalrealities o psychological ariables,chapter 3 in The Sociological magination( 1959).TheodorAdorno's Sociology ndEmpiricalResearch' 1976), likewise m-phasizes the 'epiphenomenal subjec-tivism' f the surveyand its inattention odeterminant tructural elations.To thisone should add Fernand Braudel'sre-peated nsistenceon the limitedpurviewof contemporaries, .e., the fact thatbecause actors are bound up in theimmediate, n the flux of the short-term,'livingworlds'are to some extent 'blindworlds', unaware of deeper historicalcurrents and enveloping structuralre-alities.See especially he essayscollectedin On Htstoo (1980). For a discussionofthe exigenciesof field research,consultHowardS. Becker 1970).10. And thatqualification oldseven fone assumes that problemsconcerningthe stability nd univocity f meaning nbothquestions nd repliescan be reason-ably managed - a large assumptionindeed! As Cicourelnotes, the problemsof interaction ffects alone are such thatthe survey instrument might best belimited to 'providing imple descriptivematerial f a non-threateningypefromalarge sample of individuals for somepractical purpose' (1964: 115). That

survey-based esearch has not to dateyielded the 'science' its proponentsprophesied, see Christopher G. A.Bryant (1985); and Peter Manicas(1987).11. Someone checking off response-boxes on a questionnaire s hardly 'en-gaged' in the way an individual s, say,when composinga poem, a diary, or atheological tract. To assume that the'documentary'value of the survey issuperior o the latterentailsquite a leapof faith, and certainly onfirms he sus-picion that far too much sociology isdrivenby methodological xpediency.12. Mills 1959:70). Cicourel'sMethodand Measurementrovidesa detailed ac-count of these intrusive contamination'effects and the attending problemsofartificiality.t would be difficult o chal-lenge his summary observation: Thecorrespondence between the hypo-thetical world inferred from question-naire items and actual behaviorof theactor remainsan open empiricalprob-lem' (1964:113). Or with Bourdieu, etal., that the 'classic echniques f empiri-cal sociology are condemned by theirvery nature to create situationsof fic-titious experimentationessentiallydif-ferent from the socialexperimentationsthat are constantlyproduced n the un-folding of social life' ( 1991:43). Allthings considered, extrapolatingfrom'survey-invented'materials to the realworld of situationalcontingenciesandemergentprocesses s perhapsmore 'in-ferential'than extrapolating rom 'dis-covered relics' hat do not entail corre-sponding researcher-respondent. .nteractlon.13. These points, though obvious,need continual restating, because so-ciologyas a disciplinehas long cultivateda systematicdisregard for temporality,largely as a consequence of the mis-guided effort to imitate nomothetic ci-ences like physicsand so gain distancefrom allegedly diographic nferiors ikehistoriography. examine he disastrousfalloutof this enduring egacy n 'Positiv-ism Redivivt4s?'1992a),and reply to at-tending commentsby Steve Fuller andJonathanTurner in 'Towards respect-able, reflexive, scientific sociology: A


17/18

18 JosephM. Bryantnote on the reformation required'( 1992b).14. Theoreticalattention o the 'tem-poralstructure f everydayife'isone ofthe virtuesof PeterBergerandThomasLuckmann'sThe Social ConstructionofReality.As they note:'Temporalitys anintrinsicpropertyof consciousness.Thestreamof consciousnesssalways rderedtemporally . . Allof myexistence nthisworld scontinuously rderedbyitstime,is indeed envelopedby it' (1967:26-7).Humanagency, notherwords, sfunda-mentallyime-referential, ith heconse-quence that social actions are to beunderstoodnot as episodic events ordiscreteoccurrences,but as complexdurationalprocesses,constitutively n-formed by past, present, and futuretemporalonsiderations.15. The law of entropy, in otherwords,s non-operativen thedomainofsocial-historicalhenomena:ertain ventsanddevelopmentsre 'pivotal' n that theydeczsivelylterthearcof histortcal ossibility,therebyontinuingo'condition'llsubsequentdevelopments.he distinctivepatternsofinterfacebetween 'religion'and 'civilsociety'n Christianand Islamicciviliz-ations,orexample,reflectsheenduringorreverberative'actthatthefounderofthearlier aithwascrucifiedbya hostilestateower,whereas he prophetof thelatteronqueredallbeforehim,success-fullyusingspiritual ndtemporalpowerfromheoutset.16. AndrewAbbott's eriesof articles(1988; 990; 1992) which expose thelimitationsf the 'variableparadigm'nmainstreamuantitativeresearch ad-dresseshese issuesin a highlyoriginalmanner, ith the aim of developinggeneralizable,ormalnarrativemethods.Particularlyelevant n thisregard s hisdemonstrationhat time-series andevent-historyethodsdo notadequatelysolvehe problemsof contextand tem-porality,s they are rendered compu-tationallynworkable nce theyattempttodvancebeyondontologicallysimpli-fied'ases, i.e., rudimentarynarrative'stages'ocusing on only one or twovariables.17. Let it be noted, however, thatrecourseo 'sanitizingbridgements'nd

'strategic oncealments's characteristicof all 'publicscience'.The informationgap betweenresearch-as-practicedndresearch-as-presenteds of course par-. . . . .tlCUry pernlclous n SOCIO ogy, glvenour fundamental disagreementsovermethodological rocedure.18. ThisepistlewasauthoredbyCor-nelius, hebeleagured ishopof Rome, nA.D.251, and is preservedn Eusebius'sEcclesiastical istoo, VI.43.5-11.19. The content-ratio f reportage ointerpretationvaries considerably,inworksof historiography s well as inworks of grand historical sociology.Wheneverhatratiois low,groundsforuneaserise appreciably.One expedientthatreadily uggest tselfis for historicalsociologistso include detailed appen-diceson sourcematerials, iscussinghesecondaryiteratureand the kinds ofprimaryatacontainedherein.20. More fundamentally,works ofscholarlyynthesisormanindispensable

componentwithinevery branchof sci-ence,and to insist upon contactwithprimaryata orallresearchstoendorsetheallacies f crudeempiricism. cienceis, fterall,acollaborativenterprise.21. Skocpol"s isionandMethod 1984)isa convenient place to start, for inadditiono the informativecriticalre-viewsof several distinguished prac-titioners, ibliographicreferences areextensivend include citationsof themajorebates.BIBLIOGRAPHYAbbott,Andrew 1988 'TranscendingGeneralLinear Reality', SociologicalTheoty(Fall):169- 186.Abbott,ndrew 1990 'ConceptionsofTimeand Events in Social ScienceMethods',istoricalMethods 3:14>50.Abbott,ndrew 1992 'FromCausestoEvents:otes on NarrativePositivism',SociologicalMethods and Research20(4):28-55.Adorno,heodore1976 'Sociology ndEmpiricalResearch', pp.6v86 inAdorno,t al. The PositivistDispute inGermanociology, ranslatedby G. Adeyand.Frisby,NewYork:Harper&Row.


18/18

Evidencendexplanationn historyndsociology 19Becker, Howard 1970 SociologzcalWork:Method rulSubstance,Chicago: Aldine.Berger, Peter and T. Luckmann 1967 TheSocial Construction f Reality,New York:Anchor Books.Bloch, Marc 1953 The Historian'sCraft,translated by P. Putnam, New York:Vintage Books.Bourdieu, Pierre et al. l 991 The CraftofSociology: Epistemological Preliminaries,translated by R. Nice, New York: Walterde Gruyter.Braudel, Fernand 1980 OnHzstoo,trans-lated by S. Matthews,Chicago: Universityof Chicago Press.Bryant, Christopher G. A. (1985) Positiv-ism n SocialTheoo andResearch,London:Macmillan.Bryant, Joseph M. 1992a. 'PositivismRedivivus?',Canaduznournal of Sociology17(1): 29-53.Bryant, Joseph M. 1992b 'Towards arespectable,reflexive, scientificsociology:A note on the reformation required',CanadianJournal of Sociology17(3):322-31.Cicourel, Aaron 1964 Method andMeasurementn Sociology,New York: TheFree Press.

Goldthorpe, John H. 1991 'The uses ofhistory in sociology: reflections on somerecent tendencies', Britishournalof So-ciology2(2): 211-30.Machlup, Frits 1963 'Are the SocialSciences Really Inferior?', pp.15v80, inM. Natanson (ed.) Philosophyf theSocialSciences,ew York:Random House.Manicas, Peter 1987 A HistooandPhilos-ophyof theSocialSciences,Oxford: BasilBlackwell.Mills, C. Wright 1959 TheSociologicalImagination,ew York: Oxford Univer-sityPress.Platt, Jennifer 1981 'Evidence and Proofin Documentary Research', SociologicalReview9(1): 3146.Rock, Paul 1976 'Some problems ofinterpretive historiography',Britishour-nalofSociology7(3): 35349.Skocpol, Theda (ed.) 1984 VisionandMethodnHistoricalociology,ambridge:CambridgeUniversity Press.Tilly, Charles 1981 As SociologyMeetsHistoty, ew York:Academic Press.Webb, Eugene, et al. 1966 UnobtrusiveMeasures, hicago: Rand McNally andCo.

Bryant Evidence and Explanation in History and Sociology

Documents