pol ant

8/3/2019 pol ant

1/22

Political Anthropology: The Analysis of the Symbolism of Power Relations

Author(s): Abner CohenReviewed work(s):Source: Man, New Series, Vol. 4, No. 2 (Jun., 1969), pp. 215-235Published by: Royal Anthropological Institute of Great Britain and IrelandStable URL: http://www.jstor.org/stable/2799569 .

Accessed: 14/11/2011 04:26

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

Royal Anthropological Institute of Great Britain and Irelandis collaborating with JSTOR to digitize, preserve

and extend access toMan.

http://www.jstor.org

8/3/2019 pol ant

2/22

POLITICAL ANTHROPOLOGY: THE ANALYSIS OF THESYMBOLISM OF POWER RELATIONSABNER COHEN

School fOrientalndAfricantudies,niversityf ondonA decade go,an eminentoliticalcientistarrieduta surveyfpoliticalstudies ithinocial nthropologyo assessheir ontributiono the tudy fpoliticsenerallyEaston959). Hisconclusionasbriefnd onfident:Politicalanthropologyoes otyet xist'. ocial nthropologists,e rgued, erenterestedinpoliticsnlyndirectly,nd nlyn ofar s t ffectedthernstitutionalariablesin ociety.Purelyoliticalonsiderationsrencidentalothe mphasisnkinshipstructurend itsgeneral ocial effects'i959: 2I2).Apartromome rief,austicemarksyBaileyI968: 28 ) againstt, aston'sverdict as remainednchallenged.et I canthinkf no other ommentnsocialnthropologyhich asbeen sdamagingnd sacademicallyrresponsibleas this ne. t sdamaging,irstly,ecauseaston soneofthe eading iguresn

politicalcienceodayndhis iews ountwidelynthe ocial ciences.econdly,his rticlenpoliticalnthropologyI959) andhis reoccupationithhe onceptofpoliticalystem'avemade im opular ithocialnthropologistsGluckman& Eggan 965), some fwhom eem oattachmuchweightohisudgement.Even nanthropologistf he alibre fSouthalleems ohave cceptedaston'sviewand,with pologeticemarkso forestallossibleeach-typeharges fbutterflyollecting,asurged hatwe should vercomeurweaknessy em-barkingnnew,more ophisticated,unidimensional'lassificationsfpoliticalsystemsSouthall965).Easton'somments rresponsibleecauset sbased nwhat eems ohave een hurriedeadingf fewmonographshat appenedohavebeenpublishedhortlyefore emade t.AsI indicateelow, eneglectswholetreamsf houghtithinnthropologyhoseontributionothe tudyfpoliticsave eenmmense. orst f ll,hecompletelyisunderstandshe atureof he entralheoreticalroblemsithwhich ocialnthropologyeals.Eastonwritess fthere ere consensusstowhat oliticalnthropologysabout. ut uch consensusoes otyet xist.ndeed,oliticalcientistshemselvesarenot greedbout he xact omainfpoliticalcience.na recenturveyfhisowndiscipline,astonI968) states:Politicalciences tillookingortsdentity'and nhis ookAframeworkforpoliticalnalysis,ublishedsrecentlys 965,he sstill oncernedith he uestionfwhat ariableshould e ncluded ithinis'politicalystem'.f thisstodayhe asewith oliticalcience,boutwhich eknows omuch, ecertainlyould othave een osure enyearsgo aboutwhatvariablesere he oncernfpoliticalnthropology.This snota matterfplaying ithwords nddefinitions,rofquibblingbetweenival isciplines.he ssues refarmore undamental;or hey ose hewhole opicaluestionf he ole f ocialnthropologynthe tudy,ot nly f

8/3/2019 pol ant

3/22

2I6 ABNER COHENthe changingocietiesf thenewly ndependenttatesf Africand Asia,but alsoofthe omplexocietiesfthedeveloped ountries.ur traditionalubjectmatter-the solated,mall-scaleociety-hasundergoneundamentalolitical hange ndcanno longerbe studiedas if' itwerestill elativelyutonomous, utmustbeconsidered ithin he nstitutionaltructuref the new state.This brings ocialanthropologyace o facewithpolitical ciencewhich, ince he econdworldwar,has rapidly evelopedntoa verypowerful iscipline,oth n terms ffinancialand manpower esourcesnd nterms f ts nfluencen governmentsndon otherdisciplines.What ismore, n recent ears largenumber fcompetentoliticalscientistsave nvadedour territory'nd have undertakenesearchn thepoliticsofthenew states f the thirdworld'.'These developments ighlighthe stateof transitionhroughwhich socialanthropologys now passing nd call fora re-examinationf aims,methodsand theoretical rientation.hey also raise the questionof the natureof therelationshipetween oliticalcience ndsocial nthropology. hatcan theyearnfrom ne another?Whatkindof division f abour nd ofco-operationan bedeveloped etween hem?Underlyinghese uestionssthemajor uestion:whatis political nthropology?Thecentralheoreticalroblemnsocial nthropologyA disciplines definedn termisf themajorproblemswithwhich tdeals.A'problem' n this ontext eferso theneedfor he nalysisf nteractionetweenmajorvariables. he advanceof a disciplines as much n the dentificationndisolation f these ariabless inthe nalysis ftheirnterdependence.sHomansoncepointed ut, neofthe essons e learn rom he lder ciencesstocutdown,as far swe dare, henumber fvariableswithwhichwe deal.The firstmajortheoreticalndmethodologicalreakthroughn thedevelop-ment f ocial nthropologyccurred henDurkheim,nd ater adcliffe-Brown,advocated he nalyticalsolation f socialfrom istoricalndfrom sychologicalfacts.t is truethat n recent ears hishasbeensubjected o criticismysomeanthropologists,n a varietyfgrounds. utthis riticismas o far eendirectedagainstherigidity,ot the heoreticalrinciples,nderlyinghis eparation.venthose nthropologistshoregardocial nthropologys a kindofhistoriographyand callfor he nalysisfhistoricalataby nthropologistsgree, evertheless,hatsocial nstitutionsannot esociologicallyxplainedn terms fpast ventsEvans-Pritchard956: 6o). Similarly,ven in the study f symbols nd of symbolicbehaviour, hoseoperationscloselynvolvednpsychic rocesses,he eparationof the social from hepsychichas beensystematicallyaintainedLeach I958;Gluckman963; I968; Turner964).Followingthe theoreticaleads by Durkheimand Radcliffe-Brown,ocialanthropologistseveloped he so-calledholistic'approach o society nd con-centratednthe tudy fwhat ameto be known s social structure'.ut, ettingasideexplicitheoreticalndmethodologicalormulations,hequestionhould easked:whathavesocial nthropologistsctually one norder o study he ocialstructureolistically?he answer an befoundnthemonographictudies hichtheyhaveproduced.Broadly peaking,ocialanthropologistsave interpretedhe holism' of the

8/3/2019 pol ant

4/22

THE ANALYSIS OF THE SYMBOLISM OF POWER RELATIONS 2I7social tructurenterms f a limited umber f specificnstitutionsBeattie959).A survey f themonographs ill show that heyhave generally oncentratednthe tudy f four road nstitutionalields: conomic, olitical, inshipnd ritual.On a higherevelof abstraction,hese ournstitutionalields omprise wo majorvariables: hepolitical nd the ymbolic.The separation etween he conomic nd thepoliticalnsocio-anthropologicalstudiess often ery rbitrary. hat hascome to be known s economic nthro-pology'is nfact n admixturefdescriptionsf economic rocessnd economicrelationships.hese two aspects f economic ctivity elong to two differentconceptualchemes hat avebeendeveloped y wodifferentisciplines.conomicprocessrefers o the interaction etweenman and relatively carce resources.Economicrelations,n theother and, efero interactionetweenmen nvolvedin the economicprocess. ocial anthropologistsave been interested ainly neconomic elations,.e. nrelationsetweenndividualsndgroupsn theprocessesof production, xchange nd distribution,nd most ocialanthropologistshohave studied rocess ave doneso mainlyn so far s process ffectedconomicrelations.2 ut these economic relations re relations f power and are thusessentiallyolitical, orming majorpartof thepolitical rder nany society.These wo types fpower, he conomic nd thepolitical,recertainlyifferentin a number frespectsndthey reassociatedwithdifferentypes f sanctions.But theyare intimatelynterconnectednd are in many contexts nseparable.Inbothwe are nfact ealingwith elationshipsfpowerbetweenndividualsndgroups,when hese elationshipsreconsideredtructurallyhroughouthe xtentof a polity. n both institutions,elationshipsre manipulative,echnical ndinstrumental,smen ndifferentituationsse one another s means o ends ndnot as ends n themselves.Similarly, inshipndritual, hough istinctn form, avea great eal ncom-mon, ndthe eparationetween hemsoftenrbitrary,nd ometimes isleading.Theyare bothnormative, epending n categoricalmperativeshat re rootedin the psychic tructuref men n society hrough ontinual ocialisation. othconsist f ymbolsndof ymbolic omplexes. hese ymbolsrecognitive,nthatthey irecthe ttentionf men electivelyo certainmeanings. hey reaffective,inthat hey renever motionallyeutral; hey lways gitate eelingsnd senti-ments. heyareconative,nthat heympelmen to action.These characteristicsdeterminehepotencyfsymbols, hich an be ranged na continuum,rom heleast otent, meresign',to themost otent, 'dominant ymbol' Turner964;I968).Symbolsre ystematisedogether ithin heframeworkfdynamicdeologies,orworld-views,nwhichthesymbols fthepolitical rder re integrated iththose ealingwith heperennial roblemsf human xistence: hemeaning f ifeanddeath,llness ndhealth,miseryndhappiness,ortunendmisfortune,oodand evil. Thesetwo symbolic omplexes upport ne anotherwithin unifiedsymbolic ystem.Bothcategoriesf ymbols,hose fkinshipndofritual, avebeenused lmostinterchangeablynthe rticulationfpolitical roupingsndofpowerrelationshipsbetweenndividualsndgroups. itual ymbolsorm art fmost inshipystems,andkinship ymbols orm art f most itual ystems. inship ymbols resaidto

8/3/2019 pol ant

5/22

2I8 ABNER COHENbe particularlyuited to articulate hanging nterpersonalelationships, hileritual ymbols resaid to be particularlyuited o express olitical elations f ahigher evel.But there remany aseswhere kinshipdeology smade to arti-culate thepolitical rganisationf large populationsn bothuncentralisedndcentralisedocieties. he Bedouin of Cyrenaica Peters960; I967) andtheTal-lensi Fortes 945; I949), forexample, xpress heir olitical rganisationn theidiomofkinship. hesame anbe saidofthe rganisationfkingdoms.hewholepolitical tructuref theSwazi is expressedna lineagepatternhat ervades hewhole kingdomfrom hehighest o the lowest evels Kuper I947). In othercentralisedocieties inship ymbols rticulate oliticalgroupingsnd politicalrelations n only some levels. Among the Mambwe (Watson I958) and theLunda of theLuapulavalley Cunnison959) the tabilityf thepolitical tructureatthetop ssymbolisednterms f perpetual inship' elationships.mong theAshanti, n the ther and, nly he owerpart f he oliticaltructuresorganisedona kinship asis Fortes 948). But evenwhenwe considerhe ymbolismf nter-personal elationshipsn large-scale,ontemporaryndustrialociety,we canseethat hese ymbolsrticulaten endlessrray f nformalolitical roupings hoseoperations a fundamentalartofthetotalpolitical tructurefthe ociety.Similarly,itualymbols eednot be exclusivelynvolvednthe rticulationftherelativelyigh evel, arge-scale,olitical roupings,nd can be seen o expressvarious ypes f nterpersonalelationships.hus, s GluckmanI962) points ut,in most tribal ocieties,nterpersonalelationshipsrehighly ritualised'.Also,inmanyMediterraneannd LatinAmerican ountriesxtensive se smadeof theritualkinship elationshipsreated y the nstitutionf god-parenthood',om-padrazgo,n theorganisationf various ypes f nterpersonalelationshipsnd ofgroupings,nsomecasesbetween he ocially qual, n others etween he ociallyunequalMintz WOlf 950; I956; Pitt-Rivers958;Deshon963; Osborn968).Kinship symbols nd ritual ymbols re highly nterdependentnd neithercategory an operatewithout he other.The distinctionetween hem s oftenbased,not on objective ociological nalysis,utonnative sages nd deologies.The same can be said of the broaderdistinctionetweensacredsymbols' nd'profane ymbols',or betweenritual nd ceremonial enerallyLeach I954;Martin965; Douglas966).This s not to saythat here reno significantifferencesetween ymbols,rthat ymbolshouldnotbe categorised.ut symbols rehighly omplex ocio-cultural henomenand can be classifiedccording o a variety f criteria,e-pending n thepurpose ftheclassification.n otherwords, uch classificationdepends nthenature ftheproblem fthe nalysis hich,nturn, epends n thevariableshat re considerednthe tudy. amarguing ere hat nsocialanthro-pologythecentral heoreticalnterestn thestudy fsymbolss theanalysis ftheirnvolvementntherelationshipsfpower, nd that hiswill call for type fclassificationhichmayoften e at variancewiththatprovided y theculturaltraditionsf which he ymbolsrepart.FormndfunctionnsymbolismIt sessentialhatwe distinguishetweenymbolicformsndsymbolicfunctions.The same ymbolic unction, ithin particularolitical ontext,anbe achieved

8/3/2019 pol ant

6/22

THE ANALYSIS OF THE SYMBOLISM OF POWER RELATIONS 2I9by a variety f symbolic orms. or example, verypolitical roupmusthavesymbols f distinctiveness,.e. of identitynd exclusiveness.ut this can beachievedn differentymbolic orms: mblems, acialmarkings, yths forigin,customs f endogamy r of exogamy, eliefs ndpracticesssociatedwiththeancestors,enealogies,pecificeremonials,pecial tylesf ife, hrines,otions fpurityndpollution,ndso on (Cohen 969: 20I-I4). Thus,ritual ymbolsndkinshipymbols iffernform ut notnecessarilyn function.It is also importanto rememberhat thesetwo forms f symbolsdo notexhaust etween hem hewholesymbolic niversena society. here remanyother ormsf symbolshat renotordinarilyubsumednder itherhe ategoryof kinship r ritual.This is suchan obviouspointthat t seemsunnecessaryomentiont.Yet it s surprisingow oftenwe tend o forgettand thus eadour-selves straynourobservationnd analysis. his s particularlyhe ase whenwestudy hanging re-industrialocietiesrmoredeveloped ocieties. ftennsuchcases raditionalymbolsfkinshipndofritualosetheir ignificancendwe thenbegin o talkof socialdisintegration'r,whenwe refero ritual articularly,f'secularisation'.t thenbecomes asyto slip ntothe theoreticalosition hat hehold ofsymbols n socialrelationshipssweakenings the ociety ecomesmoresocially ifferentiatednd moreformallynd rationallyrganised. ut,asDuncan(I962) pointsout, there an be no socialorderwithout he mystification'fsymbolism.his strue, otonly fcapitalistocietiess Marxmaintained,ut lsoof ocialistocieties, here mblems,logans, anners, ass arades,itles, atrioticmusic nd songs, nd,inevitably,he worldview' of dialecticalmaterialism-these nda host f all sorts f other ymbols laytheir art nthemaintenancefthepolitical rder. Secularisation', ritesMartin i965: I69), 'is less scientificconcept han toolofcounter-religiousdeologies'.Thus,although inship ymbols nd ritual ymbolsmaybecomeobsolete nmodem ociety,ther ymbolsake heir lace narticulatingld,as wellasnew,symbolic unctions. changeofsymbolic ormdoesnotautomaticallyntailchange fsymbolic unction,ecause he amefunctionanbe achieved ynewforms. imilarly, continuityfsymbolic ormneednotautomaticallyntailcontinuityfsymbolic unction,orthesameform an fulfil ew functions.nsomesituationsld symbols re revived o perform ew functionsGluckmanI942; Cohen965). As argue lsewhereCohen969: 2II-I4), the hallengeosocial nthropologyodaysthe nalysisfthis ynamicnvolvementf ymbolism,or of custom,n thechanging elationshipsfpowerbetween ndividualsndgroups.Societies ften doptdifferentymbolic orms o achievethe sametypesofsymbolicunctions.his swhatwemeanbycultural ifferences.hesedifferencesarise s a result fdifferentombinationsfcircumstances,ome of which an behistorical,ulturalndecological. ome symbolic ormsreadoptedfrom therpeoplesthroughnteraction iththem t differentistoricaleriods;others reconditionedy special cological actors.orexample, people iving n a forestareawillmakeuseof treesncarvingymbols,r ngeneralymbolic epresenta-tion,whilea people iving n thedesertwill make use of othermedia and ex-periencesn constructingheir ymbolic orms. gain,because slam s categori-cally pposedto the mploymentfpainting,arving, ancing nd music n its

8/3/2019 pol ant

7/22

220 ABNER COHENsymbolism,xtensive se smade northodoxslamic ountriesfa widerange flinguisticorms-rhetoric,roverbs nd the ike.

Symbolic orms re theproductsfcreative ork.Their nternaltructures adramatictructurendtheir tudyspartly studynthe ociology fart.Manysymbols re thecreation f anonymous rtists.t is only n moreadvanced ndsophisticatediterate ocieties hat special,named,artistsre commissionedocreate ymbols or pecific unctions-to esign flag,write nanthem,omposemusicfor hymn, aint picture f a saint, tage ceremonial. ut we areallpotentialreators fsymbols. hrough urdreams,llusions,pontaneousctivi-ties,moments freflectionndin the general lowof ourconsciousness e con-tinually roliferateymbolsndmanipulatehem.Manymenkeeptheirymboliccreativityo themselves.thers xternalisetandtry o sharetwithothermen.Thissymbolic roliferationithin ach one ofusis notentirelyurautonomouscreation, ut s theproduct fa dialecticalnteractionetween urselvesndoursocialreality. t times f change, ome men's ymbolic orms an provide ettersolutions o thecurrentroblems fa groupthanother ymbolsndthosemenwho create,mobilise, r articulate hemmay become leaders nd have theirsymbols dopted ythegroup.There sthus great ealofthe reative rtistnthepoliticaleaderwho, through isrhetoric,logans ndtacticsmanipulatesxistingsymbols r creates ew ones.Whenthis reativitysparticularlyriginal, hen thelps to articulate r to objectify ew groupings nd new relationships,edescribe im s charismatic'.

Social nthropologistsnalyse ymbolic ormsnorder odiscover heirymbolicfunctions. ne ofthemost mportantf these unctionss theobjectificationfrelationshipsetween ndividuals nd groups.We can observe ndividuals b-jectivelyn concrete eality, ut therelationshipsetween hem reabstractionsthat anbe observed nlythrough ymbols. ocialrelationshipsevelop hroughandare maintainedy symbols.We 'see' groups nly throughheir ymbolism.Values, norms, ules, nd abstractonceptsikehonour, restige, ank,ustice,good andevilare madetangible hroughymbolism,ndmen nsocietyre thushelped o be aware of theirxistence,ocomprehendhem nd to relate hem otheir aily ife.Symbols lsoobjectifyoles ndgivethem realitywhich s separate rom heindividualpersonalitiesf their ncumbents.Men are trained or theirroles,installedn them, ndhelped o perform heir utiesn the course fa series fstylised ymbolic ctivities.y objectifyingelationsnd roles, ymbols elptodifferentiateetweenthem,a function articularlymportantn multiplexrelationshipsGluckman962).By objectifyingoles ndrelations,ymbolismchieves measure fstabilityandcontinuity ithoutwhich ocial ife annot xist. ower s an erraticrocess.A vengeance roupmayhave towaitforyears eforet findstselfnvolvedn acase ofhomicide hatwillrequire ction n thepartofall of tsmembers. ut itmust e ready or ction llthe ime;for uch n event anoccur tanymoment.Its members annotafford o disband n themeantime, ut mustkeep theirgrouping live.This continuityf thegroupcan be achievedmainly hroughgroupsymbolism, ot through heirregularxercise f power. Similarly,l-though regimemaycometo officendmaintaintself or ome timepurely y

8/3/2019 pol ant

8/22

THE ANALYSIS OF THE SYMBOLISM OF POWER RELATIONS 2 Iforce, ts tabilityndcontinuityreachievedmainly hroughhe ymbolism fauthority hich tmanipulates.ubjects o notstartheirives verymorning yexamininghedispositionsfpower ntheirociety o seewhether heregimesstillbackedbythe ame amount fpoweras before, rwhether hat ower hasdiminishednd theregime an thereforee overthrown. he stabilitynd con-tinuity f theregime re madepossible hrough complex ystem f symbolismthatgives t legitimacy y representingt ultimatelys a 'natural'part of thecelestial rder.Through he mystification'hichtcreates,ymbolism akes tpossible or hesocialorder osurvive hedisruptiverocessesreatedwithin tbythe nevitableareas of conflictingaluesand principles.t does thisby creating ommunionbetween otentialnemies. proverbmongArabpeasantstates: I againstmybrother; andmybrothergainst urcousin; andmybrother ndmy cousinagainst he utsider'. mandiscoversis dentityhroughnteractionith thers.To co-operate ithhisbrothergainst heir ousinhemust econcile ishostilitytohisbrother ith heneedtoidentify ithhim nthefight gainstheir ousin.He, hisbrother,nd their ousinmust chieve ommuniono contain heir n-mitiesfthey re to co-operate gainst he ommon nemy.As Smithpoints ut I956), all politics, ll struggleorpower, s segmentary.This means hat nemies tonelevel mustbe allies t a higherevel.Thusa manmust eanenemynd anallywith he ame et fpeople, nd t smainly hroughthe mystification'enerated y symbolismhat hese ontradictionsrerepeti-tively aced ndtemporarilyesolved.ndeed,Gluckman oes ofarnelaboratingonthis unctionfsymbolisms to state hat itual ndceremonialo notsimplyexpress ohesion nd impress hevalueof society nd itssocialsentimentsnpeople,as in Durkheim's nd Radcliffe-Brown'sheories, ut exaggerate ealconflictsf social rulesand affirmhat there s unitydespite heseconflicts(Gluckman963: i8).The degree f mystification' ounts s the ocial nequalitiesetween eoplewho should dentifyn communionncrease. his s a point tressedndgreatlyilluminatedy Marx in hisexposure fthemysteriesf capitalistymbols nd'ideologies'. t is furtherlaboratednddiscussedyDuncan I962) who pointsout that ll socialorder nvolveshierarchy,hat ll hierarchynvolves elationsbetween uperiors,nferiorsnd equals, nd thatrelationshipsetween hese redeveloped ndmaintainedhroughhe mystification'fthe ymbolismfcom-munion.It isnotmy ntentiono attemptogivehere survey fthevarious ymbolicfunctionshathave beendiscovered y socialanthropologists.anysuchfunc-tions avebeen dentifiedndanalysed; utthe ystematicearch or hem nd theanalysis ftheways nwhich hese unctionso theirob is still t itsbeginning.What want ostresss that ocial nthropologistsavebeen ollectivelyoncernedwith he tudy f nterdependenceetween womajorvariables: owerrelationsandsymbolicction.3Power nd ymbolismnanthropologicalnalysisI must asten osay hat heresnothing heoreticallyew nthis. eading ocialanthropologistsaveexpressedhe ameview, though ometimessingdifferent

8/3/2019 pol ant

9/22

222 ABNER COHENterms. husmorethan ifteen ears go,Leachmaintainedhat hemain task fsocialanthropologyas to interpretymbolictatementsnd actionsn terms fsocialrelationsI954). Similarly, luckmanI942: I965) has for longtimeheldtheviewthat ocial nthropologyiffersrom heother ocial ciencesn thatt sconcernedwithcustoms,which reessentiallyhat amcalling ymbols. ocialanthropology,e states,s concerned ith he nalysisfcustomn the ontext fsocial relationsI965). Again,Evans-Pritchardtates hatsocial anthropology'studies.. socialbehaviour enerallyn nstitutionalisedorms,uch sthe amily,kinship ystems,olitical rganisation,egal procedures,eligiousults, nd thelike, nd therelationsetween uch nstitutions'I956: 5).Thisdoes notmeanthat ll social nthropologistsre nagreementhat hey reprincipallyoncerned ith he tudy fthe ymbolismf powerrelations. sweshall ee below,a fewofthem rebarelynterestedn the tudy fsymbolsndconcentraten thestudyof powerrelationsnd powerstruggles etween n-dividuals nd groups.Othersocialanthropologists,n the otherhand,are notinterestedn the tudy frelationshipsfpowerand concentraten the tudy fsymbols s such. But theoverwhelmingmajority f socialanthropologistsallon thecontinuumetween hese wo extremesnthat heirwork consistsn theanalysisfvarious ypes f ymbols ithin ssentiallyolitical ontexts. ften heyalternaten their nalysis etween hese wovariables,hough omedo so moreconsciously,xplicitlyndsystematicallyhan thers.The twovariablesre n fact wobroad spects fnearlyllsocialbehaviour. sNadelandGoffmanave hown, ll socialbehaviourscouchednsymbolicorms(Nadel ig5i: 28-9; Goffman959). On the otherhand, s many ocialanthro-pologists ointout,relationshipsfpowerareaspects fnearly ll socialrelation-ships. n thewordsof Leach: 'Techniqueandritual, rofanendsacred, o notdenote ypes faction utaspectsf almost nykindofaction' I954: I3).There s no assumptionere hat hese woaspects ccount xhaustivelyor llconcreteocialbehaviour; or his sa highly omplexprocesswhich cannotbereduced o theoperation f a fewvariables. ower relationshipsnd symbolicbehaviourreonly nalyticallysolated rom oncreteocialbehaviour,n order ostudy he ociological elationsetween hem.t is also mportanto notethat hetwo variables re notreducible ne to the other. ach is qualitativelyifferentfrom heother. ach has its own special haracteristics,tsown typeofprocessthat s governedby its own laws. Symbols re not mechanical eflections,rrepresentations,fpolitical ealities.heyhavean existenceftheirwn, n theirown right,nd canaffectowerrelationsn a variety fways.Similarly, owerrelationshipsave realityftheirwnand can n nowaybe said obedeterminedbysymbolic ategories.f theonevariablewerean exactreflectionftheother,then he tudy ftheirnterdependenceouldbe of ittle ociological alue. t isonlybecausethey redifferent,et nterdependent,hat heir solation nd thestudy ftherelationsetween hem anbe fruitfulnd lluminating.It snotrelevantoaskwhetherhe solationfthesewovariabless validornot.One can isolatefor nalysis nyvariables rom oncrete ehaviour, or t is anaxiomatic ssumptionhat ll the variablesnvolved n thatbehaviour re,to alesser rgreaterxtent, irectlyr ndirectlynterdependent.he questionsonlywhether hevariables solatedforanalysis re significantlynterconnectednd

8/3/2019 pol ant

10/22

THE ANALYSIS OF THE SYMBOLISM OF POWER RELATIONS 223whether he study f this nterconnexions likely o develop systematicypo-theses nd to lead to furthernalysis. he work and achievements f socialanthropology aveso fardemonstratedhe value and the analytical ossibilitiesofthe tudy f thetwovariables iscussed ere.Analysisnsocial nthropologyasconsistednthe tudy f nterdependence,rof dialecticalnteraction,etween hetwo variables ather han n thestudyofeitherfthe ariableseparately. concentrationnonly ne, otheneglect ftheother, illresultmainlyndescriptionshose heoreticalaluewillbe imited. hisis of course baldstatement,or ach ofthe wo variables ontainssub-variables'whoseoperationndinterdependenceustbe analysed o makeour descriptionof themajor variablemore refinedndmoreaccurate. he differenceetweenanalysisnddescriptionsa matter fdegree.

There are at presentwo experimentalrendsn social anthropology,ach ofwhich sconcernedrincipally ithonlyoneofthe womajorvariables.The ctionheorists.ne trendsa reactiongainsthe mphasis lacedbyearlieranthropologicaltudies n collective epresentations'n the lassical urkheimiantradition.his chool fthoughtends osweep he heoreticalendulumowardsanorientationmanatingromWeberianaction heory'. his heoreticalpproach(see Bailey I968; Barth 966; I967; Boissevain 968; Mayer I966; NicholasI965) distrustsnalysisnterms fgroups nd of group ymbols,nd concentrateson the ctivitiesf politicalman' who is ever mpelled o thepursuit fpower.Mayer tates his na cautiousway: It maywell be that, s social nthropologistsbecomemore nterestedncomplex ocietiesnd as the implerocietieshemselvesbecomemorecomplex, n increasingmountof workwill be basedon ego-centred ntitiesuch saction ets ndquasi-groups,atherhan n groups ndsub-groups' I966: i 9). In a recent rticle oissevain ushes his osition o its imit:'The accentmust hift rom hegroup owards he ndividual. . Individuals,ndthe oose coalitionshey orm rethusogically rior ogroupsnd ociety. viewwhichpostulateshereverses illogical' I968: 544-5).Anthropologistsf this chool ofthought resent picture f political ife ntermsf continuinggame', n which veryman sseekingomaximise ispowerby perpetuallycheming,truggling,ndmaking ecisions.very ctionhe con-templatess theoutcome fa transactionnwhich hereturnsre at east qual to,ifnot nexcess ftheoutlay.Action heorynthropologistsavedeepened urunderstandingf thedynamicprocessesnvolved nthe truggleor ower hat oeson,notonlywithin hangingsocieties, utalso within raditionalocieties. hey haveuseda 'microscope'toshowuspoliticstthegrass-rootsevel, nd have ntroducedntoourvocabularynumber f valuable terms o label 'non-group' collectivities:faction', ego-centric etwork', action et. Ina recentook,Bailey I969) presentsnd discussesa bodyofconceptsnd terms esignedodeal, n a very erceptivendpenetratingway,withthesubtletiesfpolitical ehaviour t this evel. Theseconcepts ndterms irect ur attentionotypes fgroupingsndtoprocessesfpolitical nter-action hathave so far scaped urattention,nd thus rovide swith mportanttools,notonlyfor nalysis ut also for hecollection f field ata.But when his rientationspushed o ts xtremend spresented,sBoissevain

8/3/2019 pol ant

11/22

224 ABNER COHEN(I968) does, s a substituteor he old methods', t becomes ne-sidednd thusgives distortedicture f socialreality. o put t metaphorically,hemicroscopethat his choolholds sso powerfulndisclosinghedetails fface-to-faceoliticalinteractionhat t is powerless, r out of focus, o reflect hewider structuralfeatures f society.Boissevain s certainlyight n stating hat he ndividuals prior o thegroup,butonly fhe sreferringo thebiologicalndividual.nsociety, owever, e do notdealwith iologicalndividualsutwith ocial ersonalities.hegreater art four'human nature' s acquiredfrom ociety hrough ocialisation. s Mead (I934)shows, elf-identity,hevery oncept f I', is acquiredby manthrough nter-actionwith thermen,withwhomhecommunicateshroughymbols. man sborn ntoa societywitha culture nd a structureywhichhe is shaped.Thissocioculturalealitys an objective actwhich confronts im from he outside.To that xtenthegroup sprior o the ndividual. hisdoesnotmean hatman sdwarfed ythat ealityndthathisnaturend hiswill are determinedy t. Manalso develops n autonomy fhisown,his self', by whichhe reacts n society.Therelationshipetweenman nd ocietys thus dialecticalne Radcliffe-Brown1952: I93-4; Berger& Luckman 967). Butwe mustnotexaggerateheextentto which man sfree rom hegroups o whichhebelongs. orexample,noursocietywe believe hatwe are free o chooseourpartnern marriage, hatwemarry or ove. We certainlyo so toa large xtent. ut,asmany tudiesncon-temporaryndustrialocieties ave shown,most of us marry ur socialequals.Anthropologistsallthis lass ndogamy.ndogamy,swe allknow, s a mechan-ism formaintainingheboundariesfgroups nd forkeeping heirmembershipexclusive o prevent he encroachmentf undesirableutsidersnto them. npre-industrialociety, ndogamy s formallynstitutionalised,s in traditionalIndiansociety.n our societyt is notformallynstitutionalisedutis, instead,enforcedn a subtle,mostly nconscious ay throughheoperationf a bodyofsymbols hatwe acquirethroughocialisation.he status roups o whichwebelong mplant pecialagents', pecial ymbols,n ourpersonalitiesndmakeusrespond o somecategoriesfmembers f the other ex rather hanto others.Whenwe acquire he ymbolic ehaviourmplicitn the pecialstyle f ife' ofastatus roup,we are n fact herebyutomaticallycquiringherestraints,he ol-lective epresentations,f that roup.This means hat venwhenwe feelthatweare cting s freendividualsnfollowingurown motiveswe can n fact eactingas members fgroups.Groups ctthroughhe ctions ftheirmembers. uringan election ampaign, andidates, rokers,mediatorsnd votersmanipulateneanother, ollowingheir wnprivatenterests.heyform actions,ction ets,ndloose alliances. ut they t thesametime,knowingly r unknowingly,ct asmembers f arger olitical roups.4Some actiontheoristsake the rulesof thegame, .e. thesymbols overningsocialbehaviour,sgiven ndasbeing utside he arena' nwhich he truggleorpowertakesplace,when n fact hese ymbols redramaticallynvolved n thewholeprocess t every ne ofitsstages.n otherwords, his pproach ssumesstabilitys itstudieshange. or an ambitiousndcleverman tobe ableto mani-pulateothermen,he mustbe able to manipulate ymbols y interpretingndre-interpretinghem.Thesesymbols re the collective epresentationsfgroups

8/3/2019 pol ant

12/22

THE ANALYSIS OF THE SYMBOLISM OF POWER RELATIONS 225and onlywhen a man himself articipatesn suchgroups nd accepts he con-straintsf their ymbols an he succeed n his endeavour.

If we concentratexclusivelyn the tudy f politicalman' we shall nevitablydealonlywithhis conscious nd private ndeavour. ut factions,ction ets ndother non-groups' re not entities', ut partial ections bstractedrom widerand more nclusive ocialfield.No amount f study f ego-centric etworks illreveal o our view the political tructuref society. he ego-centric etwork smeaningfulnlywhen t s seenwithin he ontext f the totalnetwork' BarnesI968).The thoughttructuralists.he otherextreme rend n social anthropologytpresentoncentratesn the tudy f ymbols, r ofcollective epresentations,ftenquiteout of thecontext,f powerrelationships.ts orientations neatly escribedby Douglas I968: 36I): 'Anthropology as moved from he simple nalysis fsocial tructuresurrentn the1940'sto the tructuralnalysis f thoughtystems'.Anthropologistsf this chool-among themT. 0. Beidelman,R. Needhamand P. Rigby-are greatly nfluenced y the 'structuralism'f Levi-Strauss.AsJacobson nd Schoepf, hetranslatorsf his Structuralnthropology,tateI968:ix): 'His approach s holistic nd integrative.. He conceives f anthropologyin thebroadest ense, s the study fman, pastandpresent,n all hisaspects-physical,inguistic,ultural, onscious nd unconscious.. He is concerned ithrelatinghe ynchronico thediachronic,he ndividualo thecultural,hephy-siological o thepsychological,he bjective nalysisf nstitutionso the ubjectiveexperiencef individuals'. evi-Straussakes n his stride, mong manyothervariables, othsymbolismnd power relationshipsn his analysis. hus, n hisstudy f mythhe takes t forgranted hat n any particularontextmyth s a'charter or ocial ction'.But, sLeach I967) points ut,Levi-Strausss nterestedin furtherroblems. e aimsat the discovery f the language of myth'.He isindeedultimatelyoncernedwith discovering othing ess thanthe language',the thoughttructure',ehind ll culture.The thoughttructuralistselieve hatwe see objectivereality', oth naturaland social,not as it 'really s', but as 'structured'n terms f logically elatedthought ategorieshat rebuilt nto urpsyche.Whatever rder heres nnatureand n societys largely heoutcome f the ctivitiesf manunder heguidanceofhis programmed'mind.The keyto understandinghestructurefsocietysthus,not theanalysis f thedynamic n-goingpatternsf interactionetweenmen,butessentiallyhe code', or the ogic,thegrammarhat s implicitn thethought ategoriesnd nthe ystemsf relationsetween hem.Thought truc-turalistsre thereforeent n breakinghe ode', for ll time ndfor llculture.To do this hey oncentraten thestudy f symbolic orms nd symbolic e-haviour.Thus,whilethe actiontheoristsoncentraten thestudy f politicalman', thethoughttructuralistsoncentraten thestudy f ritualman'.

Thought tructuralistsavegreatlyefined urunderstandingf thenaturendworking f symbolism. heyhavere-emphasisedheview-recentlyweakenedby thedeparturef many anthropologistsrom ome of thetenets f classicalDurkheimianociology-thathe ymbolicrdersnotjust mechanicaleflection,oranepiphenomenon,f thepolitical rder, ut s a fact aving nexistence f tsown, in its own right.Theyhave drawn attentiono thesystematicelations

8/3/2019 pol ant

13/22

226 ABNER COHENexisting etween hedifferentarts f that rder. ikethe ction heoristsn thefield of powerrelations, heyhave provided nthropology ith a number fimportant oncepts nd terms hatcan be used as tools for both analysis nddescriptionn thefield f symbolism.It iswhentheyose direct eferenceo social nteractionhat hey ecomeone-sided nd stray rom he main stream f social anthropology. ost of them refully wareofthisdanger nd almostnvariably egin heir ifferentissertationswitha declaration f faithn social structuralism'nd a promise o bring heiranalysis f houghttructureo bearupon hedynamicntricaciesf ocial rganisa-tion e.g.Willis 967). But, as the xpositionroceeds,hepromised nalysiss putoff ntil he nd,when t becomesargely nconsequential.This s inno way an indicationf analytical eakness ut s rather matter forientationnd nterest.he problemshat his pproach oses renot ociologicalproblems, utprincipally eal withthe relations etween ymbols. hus,Need-ham's earned rticle n Nyoro symbolic lassificationI967) dealswith cultural'puzzle'-that among heBunyoro,while ll that s good and propitiouss assoc-iated withtheright and,5 he helpful iviner ses his eft and n throwinghecowrieshells,which he uses as a diviningmechanism. he problem husdealsessentially ithrelations etween ymbols,withoutmuch referenceo socialinteraction.roblems f a similar ature re also raised or xample y Beidelman(I968a), Douglas (I968) and Rigby I968). These are of coursevery mportantproblems or socialanthropology,ut only f they re systematicallynalysedwithin hecontext fpowerrelationships.6This sbecause here an be no general cience fsymbolic ehaviour s such.Symbolic henomenarehighly omplex henomena hich an be studied romdifferentngles, ependingn thenature f theother ariableshat re ncludedin the nalysis.n social nthropologye are nterestednsymbolsmainlynsofaras they ffectnd are affectedypower relations.n otherwordswe study hesymbolic s it is structured,r systeniatised,ot by a special ogic nherentn it,butbythedynamicsf nteractionetweenmen n society seeEvans-Pritchard1937). At every tage nthestudy, eferenceas to be madeto bothvariables.study fsymbolicystemsn their wn will nevitably e 'undisciplined',nthesense hattwill have no specificim or frame freference,nd sthereforeikelyto wanderndifferentirections, ixingmetaphysicsith ogic, art,psychology,theology,r inguistics.his sindeed hereasonwhyscholarsikeLanger I964:55)andGeertzI964) complain fhow ittle asbeen chievedn thedevelopmentofa 'science fsymbolic ehaviour'.All this swell known o the houghttructuralists,ut heir ilemmas that oomuchnotice f the nvolvementfsymbolismnpowerrelations ill nevitablylead toa departurerom heneat ogicofthought ategories.believe hat hissthe ource f Beidelman's omplaint,wice xpressedecentlyi968b; I969), thatV. W. Turner lacksappreciationfthose ogicaland formal ualitieswhichallsymbolicystems. . possess'. eidelmanimselfi968b: 483) points isfingeron thereal ssuewhenhe states hatTurner mphasizesymbolss expressionsfforces;Levi-Straussmphasizesheirnominal ualities. .'. The thoughttruc-turalistsertainlylluminateheformal ropertiesfsymbols, ut, nthewords fFortesI967: 9) 'at the ostofneutralisinghe ctor'.

8/3/2019 pol ant

14/22

THE ANALYSIS OF THE SYMBOLISM OF POWER RELATIONS 227All ofthepractitionersn either f these pposing amps, he action heoristsandthethoughttructuralists,reaccomplishednthropologists,ith great eal

ofworkbehind hem n theholistic' tudy fthe nterdependenceetween owerrelationsndsymbolicction.Fully ware ofthemethodologicalndtheoreticalimplicationsfwhat hey redoing, hey ancertainlyffordoconcentratenthestudy fonevariable,whilebracketing,rholding onstant,heother ariable.But t stheir isciples hoare ikely o become ne-sidedndthus osetrack fthecentralroblem f thediscipline. his snoticeablen somepost-graduateorkofrecent earswhich ends oconcentraten onevariable otheneglect fthe ther.The main reasonswhy this ne-sidednessppeals o beginnerss that trequireslittle nalyticalffort.t solvesfor hem he rksome roblem fhaving ofind'problem' for he nalysisfethnographicalata.To concentraten the tudy feither owerrelationshipsrof symbolism oesnot nvolve great eal ofanaly-tical ffort;tposesmainly roblemsfunidimensionalescription.n account fhow individualstruggleorpower,or of howpeople behave ymbolically,s acategoricalescriptionffacts hich anbeeither rue rfalse.t sonlyby posingproblemsnvolving he nvestigationf sociological elations,r of dialecticalinteraction,etween ifferentets ffacts, rvariables,hat ignificantnalysisanbeundertaken.Theprincipalontributionf ocial nthropologyopoliticalcienceInmyview,thegreatestndmost aluable ontributionfsocial nthropologyto the tudy fpoliticssnot omuch he imple ypologiesfpoliticalystemshathavebeendeveloped,sthe nalysis fthe ymbolismfpowerrelationsenerally.The mostpenetratingndenduring artofthe Introduction' oAfricanoliticalsystemss thatdealingwiththe mystical alues' associatedwithpolitical ffice(Fortes& Evans-Pritchard940: I6-22). Eastonis right n stating hat socialanthropologistsre interested ainly n non-politicalnstitutionsike kinship,religion ndforms ffriendship. hathefails osee,however,sthat he pecialisa-tionof social nthropologys in thepoliticalnterpretationfthese ormally on-politicalnstitutions.ur major nterests not n theone-sided ffect fpoliticson thesenstitutions,she maintainsEaston 959). On the ontrary, e generallyseek o explain hese on-politicalnstitutionsnterms fpolitical elations. hustheanalysis fgreatpublic ymbolic ramasikethoseofthe TallensibyFortes(1936; I945), of theShillukby Evans-PritchardI948), of theSwazi by Kuper(I947) andGluckmani954), f nArabShi'ite illage yPetersI963)-to mentiononly few-is analysisnpolitical erms. o are tudiesffictitiousenealogies yBohannani952) and Petersi959; I967) orof okingrelationshipsyColson(I962). Even studies f suchapparentlydomestic'relationshipss marriage yLeach I96I), Peters I963) and Cohen (i965), and manyothers,re essentiallypolitical tudies.Again, Gluckman's xplanation f thestabilityf marriagesformulatednpurely olitical ermsIg50).This ineofanalysiss ofcrucialmportanceorpoliticalcience. irstlyecause,as MannheimI936) pointed ut ong ago,thepolitical cientisttudyingis ownor a similarocietyshimselfaught pinthe ame ystem fsymbols hichhe stryingo decode.Symbols re argely ootedntheunconscious ind ndarethusdifficulto identifynd to discuss y peoplewho live under hem.The central

8/3/2019 pol ant

15/22

228 ABNER COHENconcernfpoliticalcience s the tudy f the ffectf nformalolitical roupingson thefunctioningftheformal tructuref governmentnd of other arge-scaleorganisations.ll behaviour, hethern formal r in informal roups, s largelycouched n symbolic orms. he very onceptsnd categories f thoughtwhichpolitical cientistsmployntheirnalysis rethemselvesart f thevery oliticalideologywhich hey ry o understand.t is true hatMannheim's aradox anbeto some extent vercomeby slow, accumulative,mpirical nd comparativeresearch.What is more, ome great hinkersike Marx, Durkheim nd othershave developed heanalysis f thesymbolismf advanced ndustrialocieties.Butall this snot enough.Despitemanydecades fthemost ntensiveesearchon theseines, heres asyetvery ittle nalysisfwhatMackenzieI967: 280) hascalledpolitical itual' ncontemporaryolities. ery ittlesyetknown bout heway the British abinetworks SSRC I968: 25), about how decisions f vitalimportanceo theBritish conomy re taken seeLupton & Wilson I959), andthere s little greementmong political cientistsbout the nature f politicalideology in contemporaryndustrial ocieties.Mannheinm'shallengeremainsvalid.Secondly, hesymbolic rderof a society an be understood nly when t isstudiedwithin hetotal cultural radition fwhich t is a part. This traditionincludes osmology,heology,rt nd iterature.ecauseof their elativesolationand of their imple echnology,hesmall-scale,re-industrialocietieshathavebeen studied y anthropologistsave little ccupationalnd institutionaliffer-entiation. heir ulturesrethereforeotvery ophisticatednd tend ouse imitedthemes nd experiencesnthe onstructionf their ymbolic ystemsseeDouglasI968: I7). Industrialocietynthe ther and sveryhighly omplex,with greatdealof division f abour, multiplicityfgroupings,nda highdegree fsocialand cultural eterogeneity.his complexity,ogetherwithhighlydevelopedliteracynd channels orthespeedy ommunicationf culturaltems,make tssymbolic ystem eryhighly omplex ndthereforeerydifficulto analyse. hisdoes notof coursemean hat ostudies f the ymbolismfcontemporaryocietyhavebeensuccessfullyarried ut. Some important ork n thisfieldhasbeendonebya larger umber fthinkers,ncluding arx,Carlyle,Weber,DurkheimandK. Burke.Butthisworkhas lways een hwartedythe roblemsf deology,scale, omplexity,onceptsndtechniques hich havejustmentioned. ne ofthemostpenetratingnalyses f symbolismn contemporaryociety s Duncan's(I962) Communicationnd ocial rder. et one can see after tudyingt that t isessentiallyased on sheer ntuitive ork without nymethodicalndsystematicstudy fempiricalituations.What want oemphasiseere s that he ocial nthropologist,y analysinghesymbols fpowerrelationshipsnsmall-scale,re-industrialocieties,asgainedgreat eal of nsightnto he ymbolismfpowerrelationshipsenerally. nthro-pological nalysis f the ymbolsnvolvedn thedevelopment,rganisationndmaintenancef varioustypesof kinship elationships,f marriage, riendship,patron-clientelationships,orporate olitical roupings, itual, nd of differentsystems f stratification,an give thepolitical cientists orking n industrialsociety ignificantonceptsndhypothesesor nalysingwvvholeange f nformalpoliticalgroupings nd informal elationships.hese informal roupings nd

8/3/2019 pol ant

16/22

THE ANALYSIS OF THE SYMBOLISM OP POWER RELATIONS 229relationshipservade the whole formal tructuref contemporaryndustrialsociety. heyare ndeed he veryfabric fwhichthestructuref all societiessmade, nd their nalysiss thecentral roblem fpolitical cience.Apart rom hese nalyticalindings,ocial nthropologyasdeveloped echni-ques and methods hat an be ofgreathelpto at least ome of the branches fpolitical cience. n recent ears nthropologistsavebeenadaptinghesemethodsand techniqueso the tudy fcommunitiesndgroups n complex ocietiesn bothdevelopedand developing ountries. rapidly ncreasing umber f anthro-pologists re now applying heirmicro-sociological'echniqueso studies nurban areas,where the struggle or powerbetween ndividualsnd groups sintense. s Firth oints ut i95i: i8), althoughnthropologicalechniques remicro-sociological,nthropologicalormulationsan be macro-sociologicalndcan thus e adapted or he tudy f state-levelolitics.The esson f oliticalcienceThe small reas fsocial ife,nwhose tudy ocial nthropologypecialises,renow everywhereecomingntegralarts f arge-scaleocial ystems. icro-socio-logical echniquesannot nthemselveseal with hehigherevels fthese ystems.Social nthropologistsavebeenwell ware fthis roblemnd todealwith thavedeveloped uchconcepts s 'social field' and 'plural society'.These arepurelydescriptiveoncepts nd thequestion s notwvhetherhey re valid or not butwhetherhey rehelpful n analysis. heyarecertainly elpfuln directing urattentionocertain haracteristicsfthenew societiesut, nmyview, hey onotface hecentral roblem quarely. he greatestolitical evolutionfour time sthe mergencef henew tates f he thirdworld'. n both hedeveloping nd thedeveloped ocieties,he tates today hegreatestolder ndarbiterf economicandpolitical ower.Social anthropologistsave done a greatdeal of work on relativelymall-scale primitive tates.7 ut, apart from a few exceptions see for example,Lloyd i955; Bailey I960; I963; Mayer 962; Cohen I965: I46-73) theyhaveignored he mportance f themodern tate n thestudy f thepolitics f smallcommunities,ortwo mainreasons. he firsts thatwhen theynitially ecameawareof this roblem,many fthe ommunitieshich hey tudiedwere n andsstillunder olonial rule. This was particularlyhecase in Africa,where nter-national oundaries ad been argely he creation fcolonialpowers. n formerBritisherritories,ndirectulehelped operpetuatehe xclusivenessnd utonomyoftherelativelymall ribalommunities. nder hose ircumstancesherewas no'state'toconsidernd themost hat nanthropologistoulddo wastotry ostudythe olonial dministration.utalthough omeanthropologistseganover thirtyyears go to advocate hat heEuropean dministratorndmissionaryhould estudied longwiththenative hief nd witchdoctors partof the amepolitical

system see Schapera 938), no serious ttempts ere made to probeinto thedomain fthe olonial dministration.ne reasonwas that n many ases twasthecolonialgovernment hich nitiatednd financedhe research.The second easonwhyanthropologistsavenottaken hemodern tate s thecontextwithinwhich heanalysis f small ommunitieshouldbe made, s theirearlier bjections o the study fpolitical hilosophy hichhad dominated he

8/3/2019 pol ant

17/22

230 ABNER COHENstudy f the tate ntil boutthetime f the econdworldwar. The tonewas setby the ditors fAfricanoliticalystemshenthey tatedhat hey ad found hetheories f political hilosopherso be of ittle cientificalue because heir on-clusionswere not formulatedn terms f observed ehaviourFortes& Evans-Pritchard940: 4).This ast bjectionsnolonger elevant ecause he tates now being mpiricallystudied y political ciencewhichhas growntremendouslyn staturen the asttwo decades. n theU.S.A. and Britain lone, normous inancialnd manpowerresources ave been allocated o theempirical tudy fstate-levelolitics n boththedeveloped nd developingountries. herehasbeen spectacularroliferationofdepartmentsf political cience n theuniversities,ith orrespondingacilitiesforresearch,ravel nd publicationsee Wiseman 967; Mackenzie 967; SSRCI968). Some excellentmonographsnd articles n thepolitics f countriesfthe'third world' have been published nd are being used in courses n politicalscience n theuniversities.While t strue hat olitical ciencesstilllookingfor ts dentity'nd that t sstill xploring arious pproacheshat ave become he pecialisationsf differentschools f thoughtwithint, there s, nevertheless,n underlyingnterestn thestudy f state-levelhenomena,nd t s nthis especthat ocial nthropologistscan earn great eal.Someanthropologistsay dismiss hefindingsf political cience or his eryreason, .e. forpolitical cience eing macro-political'.hey would argue hat ttakes n anthropologistver year ffieldwork,nd manyyears fprocessingndanalysing isdata, o make study f the ocial ystem fa simple ommunityf afewhundred eople; and thatt s thereforebsurd o attachny cientificalue tothefindingsf political cientistsho make generalisationsboutwholesocietieswithmanymillions f inhabitants.ut this rgumentgnores wofundamentalissues.The firsts thatbecausethe state xists nd playssuch a crucialrole inchanging he tructurend theculture f our small ommunities,omeonemuststudy t. Such a study s essential ot only academicallyutalso for variety fpractical,mainly dministrative,onsiderations.t is absurd o saythat hestudyofthe tate, s a whole, hould waitthedevelopmentf micro-sociology';hismaybe a long-termevelopmentnd, n themeantime,hepolitical cientistsmeeting hechallenge. he second s thatpolitical ciencehas developednewconceptsndnew techniquesordealingwith tate-levelolitical henomenananeffective ay. Therehasbeen revolutionnmethodsf ndexing ast mountsfinformation,rocessinghem nd employing hemn futurenalysisseeMack-enzie 967: 66-74; Deutsch 966).Political cience oday pproacheshestudy fsmall ommunitiesndgroupswithreferenceo the state. n thepolitical cientists'onceptual ramework,hetribes, ands nd solated ommunities, hichhave beenthemajorobjectofourstudies, re now eithern the process f integration ithinnew socioculturalentitiesr, ffor ny political easons hey till ling o their raditionaldentity,the most that can be said about theirdistinctivenesss thattheyare 'interestgroups exerting ressuren the tate r ongroupswithinhe tate. hus, s showelsewhereCohen 969), thephenomenonalled tribalism' r retribalisation'ncontemporaryfrican ocietiess theresult, ot of ethnicgroupsdisengaging

8/3/2019 pol ant

18/22

THE ANALYSIS OF THE SYMBOLISM OF POWER RELATIONS 23 Ithemselves romone another fterndependence,ut of increasingnteractionbetween hem,within he ontextfnew politicalituations.t s the utcome, otofconservatism,utofa dynamicocioculturalhangewhich sbroughtboutbynew cleavages nd new alignmentsfpowerwithin heframeworkf the newstate.A great eal ofprogressn the tudy f such interestroups'hasbeenmade nrecent ears y political cientists.ndeed,many oliticalcientistsee thepoliticalstructuref the state s being pluralistic'-using his erm n a differentensefrom hat f ocial nthropologists-thats, sconsistingf nnumerableroupingsofvarious ortswhichmediate etween he ndividual nd the state see BentleyI949; Finer 958; Eckstein960). The developmentf interestroups, nd thenature f therelationshipsetween hem nd the tate, epends n the tructurefthe state. ome states llow a greatdeal of group pluralism';other tates is-courageor even prevent hedevelopmentf suchgroupings y conductingnendless truggle gainst hem.Thesedifferencesetween tates ave been studiedby political cientistsmpiricallynd comparativelyseeEhrmann964; CastlesI967). The term politicalculture'has been sometimes sed to describe hesestructuralifferencesetween tates. he anthropologisthostudies mallgroupswithin hecontemporarytate annot fford o ignore uch tudies.ndeed, gofurthernd saythat heanthropologist ustdeliberatelyormulate isproblemsin such wayastomakereferenceo the tate necessaryart fhis nalysis.ConclusionPolitical nthropologyiffersrom olitical ciencentworespects:heoryndscale. Political ciencesessentiallynidimensional,eingmainly oncerned iththe tudy fpower: tsdistribution,rganisation,xercise,ndthe truggleor t.As itdealswithonlyonevariable, olitical ciences descriptive.n thewordsofone of itspractitioners,tseffortsmainlyto delineate elevant henomena,ogenerate seful lassificationsnd breakdowns,nd to pinpoint heimportantcharacteristicsfpoliticalctivities'Young 968: 5). Itsuniversefreferencesthemodern tate.Political nthropologyn the otherhanddeals with much smaller reas ofpolitical ife,but compensatesorthis imitation f scaleby greater epthofanalysis.t is,as I havesuggested,oncernedwith theanalysis f thedialecticalinteractionetween wo majorvariables: elationshipsf powerand symbolism.This sessentiallycollectiveoncern,houghndividuallynthropologistsifferntheir mphasis n the one variable ather hanon theother.A great ealof workhasbeenaccomplished ysocial nthropologistsn theselines.A survey f thiswork,with n analysisf thevariousnterestsndschoolsofthought ithint,wouldrequire full-length onograph.thasnotbeen myintentionn this rticle oattemptvento outline uch survey.What s more, norder ohighlightomepoints havehad to oversimplify any ssues.8Inmyview, t spossible ow for olitical nthropology,n thebasis ftheworkalready one,toproceed o investigateuestionsuch s these:How do symbolsarticulatehedifferentrganisationalunctionsfpolitical roups?What is therange fvariationn the ymbolic orms hat erform he amesymbolic unctioninpolitical ontexts nder ifferentultural raditions? hat scommon ndwhat

8/3/2019 pol ant

19/22

232 ABNER COHENis differentetween hese ymbolic orms? o these ifferentorms ifferntheirefficiencynd effectivenessnthedevelopmentndmaintenancef specificowerrelationships? hat are thepoliticalpotentialitiesf thepatternsf symbolicbehaviour ssociatedwithvariouskinds f interpersonalelationships?ow dopolitical ndsymbolic ctivitiesnteract ithone anotherwithin heorganisationofthe ndividual iography?Whatarethemainprocessesnvolved nthedevelop-ment ndmaintenancef a political deology?What are the differentypes fsymbolictechniques, ound under differentulturaltraditions,or keepingideologies live?How do symbolic ndpolitical rocessesffect ne another nsituationsfrapid hange?How is art ffectedy, ndhow does taffect,oliticalrelationships?Manysocial nthropologistsave n fact eendealingwith uestionsikethese,though otalwaysdirectlyndsystematically.hat sneedednowis a synthesisof ourfindingsofar, ndamore ystematicrientationowards he nalysisftheinvolvement f symbolic ction n political ontexts. olitical nthropologysindeednothing ther han ocialanthropologyrought o a high evel of ab-straction,hroughmorerigorousnd more ystematicnalysis.

NOTESI want to thank ProfessorMax Gluckman, Mrs A. Hayley and Mr S. Feuchtwang forvaluable comment and criticism; heyare in no way responsible or any of my errors.I See for xampleColeman I958; Post I963; Sklar 963; Mackintosh 966 on Nigeria alone.See also the papers ncluded n Geertz 963; Apter 965; Almond & Coleman I960.2 Analysis of economic process by anthropologistss a contributionmainly to economics(see Dalton I969). On thesepoints n general ee also Firth 967.3 I want to emphasise hatthis s not a reflection f my own personal nterest ut is, in myview, the unfoldingof the full mplications f our concepts and techniques. Social anthro-pologistsstilluphold the view that their pproach is 'holistic' and that even when theyareinterestedn the study f one social nstitution,ike aw or marriage, heyhave to analyse t inrelation o the othermajor institutional ariables n the society.This means thatwhether heychooseto concentraten specific ymbolicphenomenaor on specific ower relations heyhaveto carry n their nalysiswithin he context f both variables. also believe thatmost, f notall,social anthropologiststilluphold the Durkheimianmethodological rule' that a social factshouldbe consideredn tsmanifestationhroughouthe extent fa society.When,for xample,the father-son elation s studied n thisway, itwill be seen as a relation xisting etween two

socialcategorieswhich ncludemostof the males of the ociety. n somesocieties athersxercisea great deal of control over their ons, while in other societies hey exercise ess.This makesthe father-son elation significant eature fthepoliticalorder n any society.Thus in someArab villages which I studied I965) local governors ffectivelyontained youth power' byenlisting heco-operationofthe elderswho exercised greatdeal of authorityver their ons.In British ocietyon theotherhand, where fathers xercisemuch less power over their ons,the authorities ave to deal directlywith youth by mobilising greaterpolice force. Socialanthropologistslso say thattheir pproach s comparative, nd this nevitably eads to a highdegreeofabstraction.f llthese oints re assumed, henmy contention hat ocial anthropologyiscollectively oncernedwith the analysis f the ymbolism f power relationswillfollow.Eachmonographic tudy s in fact n experimentn the analysis f thesetwo variables.As in othersciences, he greaterpart of the work of the social anthropologist onsists n 'preparing theexperiment'.This consists n analysing nd sifting he datain order to isolate the two variablesfrom hoseothervariableswhichthe anthropologist rackets s 'other thingsbeing equal'.4 'We are all, to a greateror lesserdegree, mean, selfish, ishonest, azy, inefficient,ndgrasping; nd yetwe have ideals of generosity, nselfishness,onesty, ndustry, fficiency,ndcharity.Althoughwe seldomlive up to these deals, our behaviour s affected y reaffirmingthem' (Devons I956).5 Beattie,who is theauthority n theBunyoro, questions he validityofthisgeneralisationbyNeedham (Beattie 968).

8/3/2019 pol ant

20/22

THE ANALYSIS OF THE SYMBOLISM OF POWER RELATIONS 23 36Jbelieve hat hemost timulatingarts fDouglas'soutstandingookPuritynddanger(I966) are hose ealing irectlyith he ymbolismfpower elations.7 See for xample adel 942; Smith960; Lloyd 965 and he tudiesontainedn Forde&

Kaberry967.8 A detailedmonographn the ubjects npreparation.REFERENCESAlmond,G. A. &J. S. Colemaneds). 960. Thepoliticsfdevelopingreas. rinceton: niv.Press.Apter, . P. I965. Thepoliticsfmodernisation.hicago:Univ.Press.Bailey, . G. I960. Tribe,aste ndnation. anchester:niv.Press.I963. Politicsnd ocial hange.erkeley: niv.of Califomia ress.I968. Parapoliticalystems.nLocal evel oliticsed.)M. Swartz. hicago:Aldine.I969. Stratagemsnd poils. xford:Blackwell.Barnes, . A. I968. Networksnd political rocess.n Local evel oliticsed.) M. Swartz.Chicago:Aldine.Barth, . I966. Models f ocial rganisationOcc. Pap. R. anthrop.nst. 3). London:RoyalAnthropologicalnstitute.I967. On the tudy f social hange. m.Anthrop.9, 66I-9.Beattie, .H. M. I959. Understandingnd explanationn social nthropology.rit. . Sociol.10, 45-60.- 968. Aspects fNyoro ymbolism.frica8, I3-42.Beidelman, . 0. I968a.Some Nuernotions f nakedness,uditynd sexuality. frica8,II3-32.I968b.ReviewofV. Turner heforestf ymbols.frica8, 483-4.I969. Reviewof V. TumerThedrumsf ffliction.frica9,9I-3.Bentley, . F. I949. The rocessf overnment.anAntonio: ricipia ress fTrinity.Berger, . L. & T. Luckman967. The ocial onstructionf eality.ondon:AllenLane,ThePenguinress.Bohannan, . I952. A genealogicalharter.frica2, 30I-I5.Boissevain, . 968. The placeofnon-groupsnthe ocial ciences. an N.S.) 3, 542-56.Castles, . G. I967. Pressureroups ndpolitical ulture.ondon:Routledge& KeganPaul.Cohen,A. I965. Arab order-villagesn srael: studyf ontinuitynd hangen ocial rganisation.Manchester:niv.Press.I969. Customnd oliticsn urban frica.ondon:Routledge& KeganPaul.Coleman, . S. I958. Nigeria: ackgroundonationalism.erkeley,osAngeles:Univ.ofCali-fomia ress.Colson, . I962. ThePlateau onga fNorthernhodesia:ocialnd eligioustudies. anchester:Univ. Press.Cunnison,. G. I959. TheLuapula eoples fNorthernhodesia:ustomnd historyntribalpolitics. anchester: niv.Press.Dalton,G. I969. Theoreticalssuesneconomicnthropology.urr.Anthrop.0, 63-I02.Deshon, . I963. Compadrazgon a henequenaciendanYucatan: structurale-evaluation.Am. Anthrop.5, 574-83.Deutsch, . W. I966. Thenervesf overnment.ew York:The FreePress; ondon:Collier-Macmillan.Devons, . I956. The roleof themythn politics. heListener,une I.Douglas,M. I966. Purity nddanger.ondon:Routledge& KeganPaul.I968. The social ontrolf cognition:omefactorsn oke perception. an N.S.) 3,36I-76.Duncan, . D. I962. Communicationnd ocial rder.ondon:OxfordUniv. Press.Easton,D. I959. Politicalnthropology.n Biennialeviewf nthropologyed.)B. J. Siegel.Stanford: niv. Press.I965. Aframeworkoroliticalnalysis.nglewoodCliffs: rentice-Hall.I968. Politicalcience.n The nternationalncyclopediaf he ocialciences.ew York:

Macmillan.Eckstein,. I960. Pressure roup olitics.ondon:Allen & Unwin.Ehrmann, . W. (ed.) 964. Interestroupsnfourontinents.ittsburgh:niv. Press.Evans-Pritchard,. E. I937. Witchcraft,raclesndmagicmonghe zande f he nglo-EgyptianSudan.Oxford:Clarendonress.I948. Thedivine ingshipf he hilluk f heAnglo-Egyptianudan. ambridge: niv.Press.

8/3/2019 pol ant

21/22

23 4 ABNER COHENEvans-Prichard,. E. I956. Social nthropology.ondon:Cohen & West.Finer, . I958. Anonymousmpire.ondon:Pall MallPress.Firth, . i95i. Elementsf ocial rganisation.ondon:Watts.

I967. Themes neconomicnthropology: general omment.n Themesn economicanthropologyed.)R. Firth. ondon:Tavistock ublications.Forde,D. & P. Kaberryeds) 967. WestAfricaningdomsn thenineteenthentury.ondon:OxfordUniv.Press.Fortes,M. I936. Ritualfestivalsnd social ohesionn thehinterlandf theGoldCoast.Am.Anthrop.8, 59o-604.I945. Thedynamicsf lanshipmongheTallensi.ondon:OxfordUniv.Press.I949. Theweb fkinshipmongheTallensi. ondon:OxfordUniv.Press.I967. Totem nd taboo.Proc. . anthrop.nst.966, 5-22.& E. E. Evans-Pritchardeds). ntroduction.nAfricanoliticalystems.ondon:OxfordUniv. Press.Geertz, . (ed.) 963. Old societiesndnew tates. lencoe,ll.: FreePress.I964. Ideologys a culturalystem.n deologynd iscontented.)D. Apter. ew York:FreePress.Gluckman, . I942. Analysisf social ituationnmodernululand. anchester: niv. Press.i950. Kinshipnd marriagemong heLozi ofNorthem hodesia ndtheZulu ofNatal. nAfricanystemsfkinshipndmarriageeds).A. R. Radcliffe-Brown D. Forde.London:OxfordUniv.Press.I954. Rituals frebellionn south-eastfrica. anchester:niv.Press.I962. Les rites e passage.n Essays n the itualf ocial elationsed.)M. Gluckman.Manchester: niv.Press.I963. Ordernd ebellionn tribal frica.ondon:Cohen & West.1965. Politics,aw nd itualn tribalociety.xford:Blackwell.I968. Psychological,ociologicalndanthropologicalxplanationsf witchcraftndgossip: clarification.an N.S.) 3,20-34.& F. Eggan 965. Introduction.n The relevancefmodelsor ocial nthropologyed.)

M. Banton Ass. ocialAnthrop. onogr. ). London:Tavistock ublications.Goffmann,. I969. Thepresentationf elfneverydayife. ondon:AllenLane,ThePenguinPress.Jacobson, laire & B. G. Schoepf968. Translators'reface.n Structuralnthropologyby)ClaudeLevi-Strauss.ondon:AllenLane,ThePenguin ress.Kuper,H. I947. AnAfricanristocracy:ank monghe wazi.London:OxfordUniv.Press.Langer, . I964. Philosophicalketches.ew York: The New American ibraryfWorldLiterature.Leach, . R. I954. Politicalystemsfhighlandurma. ondon:Bell.I958. Magicalhair.J.R. anthrop.nst. 8, 47-64.I96I. Rethinkingnthropology.ondon:Athlone ress.I967. Introduction.n The tructuraltudyfmythnd otemismed.) E. R. Leach Ass.socialAnthrop.Monogr.5). London:Tavistock ublications.Lloyd, . i955. Thedevelopmentfpolitical artiesn western igeria. m. ol.Sci.Rev.49,693-707.I965. The politicaltructuref African ingdoms:n exploratory odel. n Politicalsystemsndthedistributionfpower ed.)M. Banton Ass. socialAnthrop.Monogr. 2).London: TavistockPublications.Lupton, . & S. Wilson 959. Backgroundndconnectionsf opdecision-makers.anchesterSchool959, 30-51.Mackenzie,W. J.M. I967. Politicsnd ocialcience. armondsworth:enguin.Mackintosh,.P. I966. Nigerianovernmentnd olitics.ondon:Allen& Unwin.Mannheim,. I936. IdeologyndUtopia. ondon:Routledge& KeganPaul.Martin, . I965. Towards liminatingheconcept f secularisation.n Penguinurveyf hesocial ciences965 ed.)Julius ould.Harmondsworth:enguin.

Mayer,A. C. I962. System ndnetwork: n approacho the tudy f political rocessesnDewas. In Indian nthropologyeds) T. Madan & G. Sarana.Bombay:AsiaPublishingHouse.1I966.The significancefquasi-groupsn the tudy f complexocieties.n The ocialanthropologyf omplexocietiesed.) M. Banton Ass. ocialAnthrop. onogr. ). London:Tavistock ublications.Mead,G.H. I934. Mind,elfnd ociety.hicago:Univ.Press.

8/3/2019 pol ant

22/22

THE ANALYSIS OF THE SYMBOLISM OF POWER RELATIONS 23 5Mintz, . W. & E. Wolfi950. An analysisf ritual oparenthoodcompadrazgo).West.J.Anthrop. , 34I-68.- I956. Canamelar: he ubculturefa rural ugar lantationroletariat.n Thepeople

of uertoico ed.)J. Steward. rbana:Univ. of llinois ress.Nadel,S. F. I942. A black yzantium:he ingdomf he upe fNigeria. ondon:Oxford niv.Press.I95I. The oundationsf ocial nthropology.ondon:Cohen & West.Needham, . I967.Right nd eft nNyoro ymboliclassification.frica7,425-52.Nicholas, . W. I965. Factions: comparativenalysis.n Politicalystemsnd he istributionfpowered.) M. Banton Ass. ocialAnthrop. onogr. ). London:Tavistockublications.Osborn,A. I968. Compadrazgo ndpatronage: Colombian ase.Man N.S.) 3, 593-608.Peters, . L. I960. The proliferationf segmentsn the ineage f theBedouin fCyrenaica.J. R. anthrop.nst. 0, 29-53.I963. Aspects frank nd tatusmongMuslimsn a Lebanese illage.nMediterraneancountrymened.)J.Pitt-Rivers.aris:Mouton.I967. Some structuralspectsf thefeud mong he amel-herdingedouin fCyre-naica.Africa7, 26I-82.Pitt-Rivers,.A. I958. Ritualkinshipn Spain.Trans. .Y. Acad. ci.20, 424-3I.Post,K. W. J. 963. TheNigerianfederallectionsf 959. London:OxfordUniv.Press.Radcliffe-Brown,. R. I952. Structurendfunctionn primitiveociety.ondon:Cohen &West.Rigby, eter 968.Jokingelationships,in ategoriesndclanshipmong heGogo.Africa8,I33-5.Schapera, . 938. Essay n fieldmethodsnthe tudy f modern ulture ontacts.nMethodsof tudyf ultureontactsnAfricaed.) L. Mair Mem. nt. nst.Afr. ang. 5). London:InternationalnstituteorAfrican anguages nd Cultures.Sklar, . L. I963. Nigerianoliticalarties. rinceton: niv. Press.Smith,M. G. I956.On segmentaryineage ystems.. R. anthrop.nst. 6, 39-80.I960. Governmentn Zazzau. London:OxfordUniv. Press.

Southall, . I965. A critique f the ypology f states ndpolitical ystems.n Politicalystemsand he istributionf owered.)M. Banton Ass. ocialAnthrop.Monogr. ). London:Tavistock ublications.SSRC (The Social ScienceResearchCouncil) I968. Researchn politicalcience.ondon:Heinemann.Turner, . W. I964. SymbolsnNdemburitual. n Closed ystemsndopenmindsed.) M.Gluckman.dinburgh: liver& Boyd.I968. Thedrumsf ffliction.xford:Clarendon ress.Watson,W. I958. Tribal ohesionn moneyconomy.anchester: niv.Press.Willis, R. I967. The head and the oins: Levi-Strauss nd beyond. Man N.S.) 2, 5I9-34.Wiseman,H. V. I967. Introduction:overnment,olitics nd political cience.n Politicalscience.ondon:Routledge& KeganPaul.Young,0. R. I968. Systemsf oliticalcience.nglewood liffs: rentice-Hall.