亲属-1959Nonunilinear Descent

NonunilinearDescentandDescentGroups'WILLIAMDAVENPORTYaleUniversityFORmanyyears,interestinsocialstructurehasbeendirectedtowardsuni-linearsystems,whilebilateralstructureshavebeenallowedtoremainanunstudiedresidualcategoryforeverythingthatisnotunilinear.Somewriters(e.g.Fortes1953:33)havestressedtheuniversalityofcertainbilateralfea-tures,whileothers(e.g.RadcliffeBrown1935:302)haveconsideredtheab-senceofaunilinearprincipleofsomekindtobeunusualorrare.Eventhemostcursoryglanceatthenowvoluminousliteratureonsocietieswithbilateralstructuresrevealsthegreatvariationamongthemandsuggeststheneed,notonlyforabetterunderstandingofwhatthesevariationsare,butalsoofwhatstructuralprinciplesarebeingvaried.Murdock(1949:56-58),Pehrson(1954),andSpoehr(1950:3)havestressedtheneedforstudiesofbilateralsocialstructure,andinhisstudyoftheKon-kamaLapps,Pehrson(1957)setforthanddefinedtheconceptof"bilateral-ism."Alsoawareoftheinadequaciesinthecurrentbilateralconcept,Freeman(1955:7)hasofferedtheterm"utrolateralfiliation"todescribesomefeaturesofIbansocialstructure.Othershaverunintospecialproblemsindescribingandclassifyingsocietieswheretheordinarybilateral-unilineardistinctionsareeitherinappropriateorinadequate.MostnotableofthesehavebeenBoas(1920)withtheKwakiutl,Firth(1929:98)withtheMaori,Gifford(1929:29-33)withtheTongans,Mcllwraith(19481:117-162)withtheBellaCoola,Gluckman(1951:71-76)withtheLozi,andGoodenough(1955:73-75)withtheGilbertese.TheplanofthispaperisfirsttosurveyanumberofselectedexamplesofwhatGoodenough(1955:72)hascallednonunilineardescentgroups,andthentocomparethesewiththewell-knownunilineartypesandwithvariationsintheso-calledkindredtypeofstructure.Incomparingthesesystems,however,onedoesnotgetfarbeforeitalsobecomesclearthatanewlookattheconceptofdescentitself,aswellastherelatedconceptsofinheritanceandsuccession,mightberewarding.Thepurposeofthispaperisnotjusttosetupaseriesoftypesortostudybilateralityasaspecialtype,buttoofferatleasttentativelythreestructuralfeatureswhichseemtooperateinallsystems,andwhichintheirvariouscombinationsproducetherichvarietyofkingroupingswefinddescribed.Letuslookbrieflyattheconceptofdescent.Thishasbeenvariouslyde-fined,dependingupontheinterestsoftheinvestigatorandtheunitofanalysiswithwhichhestarts.ThemostcommonAmericanusageseemstobeginwiththeexistenceofakingroup;descentisthendefinedasthemannerinwhichpersonsaffiliatewiththisgroupbybirth(seeMurdock1949:15).Asimilar557Thiscontentdownloadedfrom182.242.229.207onWed,13May201510:35:37UTCAllusesubjecttoJSTORTermsandConditions558AmericanAnthropologist[61,1959procedureistostartwithsocietyastheunitandtodefinedescentasameansofgroupingorofdividingthesocietyintosegmentsofkin(seeLowie1948:7-9,57-59).Anotherapproachistoregarddescentasawayofascribingjuralrightsbymeansofkinshiprelationships(RadcliffeBrown1935).Theseseveralusagesdonotreallydifferessentiallyfromeachother,andeachhascertainadvantageswhenitcomestodescribingorcomparingsystems.Itshouldbepointedoutthat,whateverelseakingroupmaybe,itsmembershipischaracterizedbysomecommonalityofstatuswithrespecttoothergroupsofkin,andthatde-scentisasmuchawayoflimitingasocialunitandexcludingsomepersonsasitisofincludingothers.Asaminimaldefinition,therefore,letusconsiderde-scentasawayinwhichstatusesareascribedand/orwithheldonthebasisofkinrelationships.Statusescanbedefinedeitherbytheirassociatedbehavioralrolesorinlegalorquasi-legalterms.2Aruleofdescentwillbeconsideredunilinearwhenascriptionthroughaspecifiedkinrelationshipapproachesafrequencyof100percentandwhenthenormsofthesocietydonotundernormalconditionsprovidealternatives.Emphasisisplacedonnormalconditions,forfewifanysocietiesaresorigidasnottoprovidealternativesforspecialsituations(e.g.adoption,ambilanak).Thematrilinealandpatrilinealsubtypesarewellknown,butitiswelltokeepinmindthatthesearenottheonlyunilinearvariationsthathavebeende-scribed.3Nonunilineardescent,then,isascriptionorexclusionthroughspecifiedkinrelationships,butwheresocietalnormsprovidemorethanonepossibilityorwherenosinglealternativeruleapproachesafrequencyof100percent.How-ever,nonunilinearityisnottobeconfusedwithdoubledescent,wherediffer-entstatusesareascribedbydifferentunilinearrules(e.g.,Ranon,Arunta;seeLowie1948:58-59,260;Murdock1949:50-52).AmajorsteptowardbetterunderstandingofbilateralsocialstructurewastakenbyGoodenough(1955:71-72)whenhedifferentiatedbetweentwotypesofgroupsonOnotoaintheGilbertsandatthesametimeuntangledthetermi-nologicalconfusionintheuseoftheterm"kindred."FollowingAmericanusageandaprecedentwhichseemstohavebeenestablishedbyPhillpotts(1913:5)inherstudyofTeutonickingroupings,GoodenoughretainedthetermkindredforthekindofnonperpetuatingkingroupingwehaveintheUnitedStates,namely,asetofkinsmenwhoarerelatedthroughmalesandfemaleswithinaspecifieddegreeofcollateralitytoaparticularcommonrela-tive.Differentfromthekindredisagroupinwhichthememberstracetheirrelationshiplineallythrougheitherorbothsexestoacommonknownorun-knownancestor;Goodenoughaptlycalledthisthe"nonunilineardescentgroup."Thesenonunilineardescentgroupsresembleunilinearonesinseveralways.Forinstance,themembersbelievetheyaredescendedfromacommonancestor,thoughfrequentlytheycannotstatetheprecisegenealogicalconnections;thegroupsarecommonlynamedandhaveidentifyingemblems;landandotherproductivepropertyaretobefoundundertheircollectivecontrol;andoftenThiscontentdownloadedfrom182.242.229.207onWed,13May201510:35:37UTCAllusesubjecttoJSTORTermsandConditionsDAVENPORT]NonunilinearDescent559myth,ritual,andreligiousbeliefsarecloselyassociatedwiththem.Marriagemayalsoberegulatedbygroupmembership,butmoreoftenthannotexog-amydoesnotprevail.Tobesure,theMangaiankopuisexogamous,withmaximalextensiontobothmother'sandfather'sgroups,4butontheotherhand,theKwakiutlnumaymisactuallyagamous,eventhoughmarriagesarepreferentiallyexogamous.5TheBellaCoolaminmintisalsoagamous,butwithendogamicmarriagespreferred.6Thefactthatthesedescentgroupsarenotallexogamousdiveststhemofafeaturethathasoftenbeenusedtocharacterizetheclanandsib(seeLowie1948:9;Murdock1949:47),butitmustberememberedthattherearemanyinstancesofagamousorevenpreferentiallyendogamousunilinearkingroups.TheTikopiankainanga(Firth1936)andtheArablineagesarebuttwowell-knowninstances.Admittedly,anendogamouskingroupisquitedifferentfromanexogamousone,andtheconsequencesofthisdifferencebecomeespe-ciallysignificantinthecaseofnonunilinearkingroups.Nevertheless,weshouldbearinmindthatallornearlyallsocietiesprescribebothexogamyandendog-amyinaccordancewithvariousprinciples,includingsuchdiverseonesasage,race,socialclass,religion,andkinship.Nonunilineardescentgroupsresembleunilinearonesinstillotherstruc-turalcharacteristics.Onefindsrestrictedsegmentsoflargergroupsthatseemtocorrespondtothelineagedivisionsofunilineardescentgroups.Forexample,priorto1746theHighlandScottishclan,anendogamousnonunilineardescentgroupandnotaunilinearclan,wasfrequentlysegmentedintosmallerlineage-likedivisionscalledseptsonthebasisofdescentfromparticularclanmembers(Adam1952).Theexogamous'aigasaofSamoahassimilarsegmentsbasedonthesameprinciple,7andinboththesecasesthesmallerdivisionsorganizeandcontrolactivitiesthatthelargerdescentgroupdoesnot.Onealsofindsnonuni-lineardescentgroupscompoundedintophratriesormoietiesforspecialpur-poses.TheNewZealandMaoriiwiismadeupofconstituenthapuonthebasisoflegendaryassociations,8whiletheMangaiankopuaresimilarlylinkedinto"right"and"left"moiety-likedivisions.Finally,nonunilineardescentgroups,incombinationwithrulesofpost-maritalresidence,occurinlocalizedandindispersedorpurelyconsanguinealforms.ThisisofcoursethedistinctionwhichMurdock(1949:65-78)hasmadebetweensibandclanforunilinearkingroups.Itispertinenttonoteherethatitseemstobethelocalizedformsofnonunilineardescentgroupsthatarethemorefrequentlydescribed,sincethekindsofactivities,rights,andobligationsmostoftenregulatedbythemarethoserequiringresidenceandparticipationwithothermembersofthegroup.Rulesofpostmaritalresidence,too,likethoseofinheritanceandsuccession,canbethoughtofassystemsofascriptionthatpertaintoparticularstatusesoractivities.Letusnowlookatsomeofthevarietiesofnonunilineardescentitself.InMangaia,affiliationwiththeexogamouskopuoccursthrougheitherthefatherorthemother,butthechoiceisnotlefttotheindividual.Childrenareassignedalternatelytothefather'sthenthemother'skopu,accordingtoprearrange-Thiscontentdownloadedfrom182.242.229.207onWed,13May201510:35:37UTCAllusesubjecttoJSTORTermsandConditions560AmericanAnthropologist[61,1959mentbetweenrepresentativesofthetwogroups.InNukuoroasimilarsystemprevails.Thedescentgroups,calledte-haka-sa-aluna,donotregulatemarriage,butincaseofintergroupmarriagesthechildrenareassignedfirsttoonepar-ent'sgroupthentotheother,beginningwiththefather's.9WiththeIbanoftheBalehregionofSarawak,affiliationwiththeinde-pendent,corporate,exogamousthree-generationapartmentgroup,calledbilek,iswiththatofeitherthehusbandorwife,dependinguponwithwhichthemarriedcouplechoosestoreside.Thismeansthatrightsinland,ritualobli-gations,andotheractivitieswhichcomeunderbilekcontrol,aswellasmem-bershipinthecorporatelonghousevillage,aresubjecttosomedegreeofchoiceatmarriage.Acouplewhichaffiliateswithonetherebyrelinquishesitspoten-tialclaimontheother,sothatthereisnooverlappingmembership.Theirchildrencanbelongonlytothebilekwithwhichtheirparentshaveaffiliated,andlikethem,whentheymarrytheymustdecidewithwhichbilek(andlong-housegroup)theywillliveandaffiliate(Freeman1957).OnOnotoaoneofthedescentgroupswhichcontrolsland,feuding,andothereconomicactivities,butdoesnotregulatemarriageinanyway,iscalledthekainga.AsinthecaseoftheIbanbilek,affiliationiswitheitherspouse'sgroup,andonlytheonewhichischosenbythembecomesapossiblechoicefortheirchildren(Goodenough1955).Inbothsocieties,economicfactorsexertsomeinfluenceonthemarriedcouple'schoice.AmongtheKwakiutl(southernbranch)thedescentgroupsarecallednumaym,andaffiliationwithoneisdependentuponpossessionofasummerpotlatchnamethattraditionallybelongstothegroup.Residencewithanduseofnumaymlandsarealsocontingentuponpossessionofthename.Amanac-quireshisnamefromhisfather,fromacloseagnatesuchasabrotherorafather'sbrother,orfromhismother'sfather.Onlyoneinstanceisrecordedwhereamanactuallytookapotlatchnamefrom,andparticipatedwith,hiswife'sfather'sgroupinsteadofholdingitandpassingitontohisson.Marriageisnotcompletelyregulatedbynumaymaffiliation,butexogamyispreferred.InBoas'time,individualscouldbesimultaneouslyaffiliatedwithmorethanonenumaymthroughthepossessionofseveralnames.TheBellaCoolahaveasimilardescentsystem,withgroupscalledmin-mint.AswiththeKwakiutl,affiliationdependsuponhavinganancestralnamebelongingtoagroup.Amanmayreceivenamesfromanumberofhispaternalandmaternalrelatives,andmaythuspossessmorethanone.However,havinganameobligeshimtovalidateitbypotlatch,anditisdifficulttofulfillthisobligationformorethanonename.Somemendosoandthusbelongtomorethanoneminmint,butmostreturnunvalidatednamestotheirrespectivegroupsandtherebyrelinquishtheirclaimsofaffiliation.Hereagainachoiceofdescentismadeamongalternatives,andrelativesocialandeconomicadvan-tagesseemtoinfluencethechoice.Sincemarriageispreferentiallywithintheminmint,endogamytendstoreducethenumberofdifferentgroupsamongwhichamancanexercisehischoiceofaffiliation.WiththeLoziofBarotseland,eachpersonisaffiliatedwitha"descentThiscontentdownloadedfrom182.242.229.207onWed,13May201510:35:37UTCAllusesubjecttoJSTORTermsandConditionsDAVENPORT]NonunilinearDescent561name"groupcalledmishiku.EachofthesehasanancestralvillagesitelocatedonasmallmoundthatrisesabovethefloodleveloftheZambeziRiver.Partici-pationwithoneofthesegroupsinvolveslivinginitsancestralvillage,andonemaychooseanyofthevillageswherehisorhisspouse'sparentsorgrandpar-entshavelived.Sincethemishikudoesnotregulatemarriageinanyway,in-tra-mishikumarriagesfrequentlyreducethenumberofpossiblechoicestolessthanthemaximum.Sincelandisverylimitedonthesemounds,amarriedcouple'schoiceispartlyinfluencedbytheavailabilityofspaceforahouse.2?TheSamoan'aigasaisadescentgroupcontrollinggardenland,housesites,andcertainceremonials.Itisexogamous,sincetheincesttabooisextendedtoallknownrelatives.Associatedwitheach'aigasaisoneormoretitles,fromtheseniorofwhichthegrouptakesitsname.Titleholdersareselectedfromamongtheeligiblemalesparticipatingwiththegroup.Affiliationwithade-scentgroupisthrougheitherfatherormother,andamarriedcouplemayaffili-atewiththegroupsofeitherspouse.Apersonismainlyassociatedwithonegroup-thatwithwhichhelives-buthemayparticipatetosomeextentinseveral.Livingwithandusinglandsbelongingtoaparticular'aigasa,aswellastherighttospeakatmeetingsofthegroup,aredependentupontheconsentofitsmembers,andthismaybedeniedifthereisdoubtaboutaperson'sgenea-logicalrelationshiptothegrouporifhehasfailedtofulfillhisobligationstoit.Here,too,indefinitelyextendedrelationshipsthroughconsanguinitybecomelimitedbytheobligationsofparticipation.Eventhoughthereisconsiderableoverlappingmembership,thenumberofpotentialaffiliationsisappreciablyreducedbybothremissnessinobligationsandfailuretoremembergenealogicalconnections.ThiskindofnonunilineardescentisalsoexemplifiedbytheMaorihapu.Apersonusuallyaffiliateswitheitherofhisparents'orhisspouse'sparents'hapu,butitisalsopossibletoobtainmembershipinthehapuofagrandparentandevenagreat-grandparent,providedthegenealogicalconnectionisrecog-nizedbythehapuinquestion.Suchrecognitiondependsupontheextenttowhichtheclaimanthasparticipatedinormaintainedreciprocaltieswiththegroup.Ifhehasallowedthesetiesto"growcold"theylapse,anditisdifficulttomaintainthemwithalargenumberofhapu.Hereagainpotentialaffilia-tionsarereducedbyobligationstoparticipate.Nonunilinearaffiliationisofcoursenotlimitedtosedentarysocietieslikethosejustcited.AmongtheKonkamaLapps,bandmembershipisdeterminedthrougheitherspouse'skinrelationshipstothepairofmalesiblingswhoformthenucleusofthebandandbywhosenamethebandiscalled(Pehrson1957).Acomparablesituationprobablyoccursinnonunilocalbandselsewhere,in-cludingthosewithhuntingeconomies,butwefrequentlylacksufficientdatatodeterminewhattherulesofaffiliationare.Severaldistincttrendsaretobenotedintheaboveexamples.Oneisto-wardaneither/orsortofaffiliation,wherebyanindividualisassignedto,orchoosesamong,limitedalternativesandrelinquishesallotherpotentialaffilia-tionsforhimselfandhischildren.Likeclassicunilineardescent,thisisbestThiscontentdownloadedfrom182.242.229.207onWed,13May201510:35:37UTCAllusesubjecttoJSTORTermsandConditionsAmericanAnthropologist[61,1959562understoodasestablishingonerightofmembership,albeitthroughtheexer-ciseofchoiceratherthanautomaticascription,andatthesametimegivingupallothersuchrights,i.e.establishingano-right,inothergroups.Contrastedwiththisistheoverlappingkindofmembershipwhereapersonmaycontinuetoparticipate,eventhoughpartially,inmorethanonegroup.Residenceandsocialparticipationwithagroupfrequentlyserveaseithercontingentorfinalfactorsinfixingpresentaffiliationandthatofdescendants.Nonunilinearaffilia-tionalsodiffersastowhetherchoiceforamarriedcoupleincludesalternativesthroughoneorbothspouses.Possiblemodesofaffiliationmayalsobenar-rowedbygroupagamyandendogamy.Variableastheyare,theseexamplesdemonstratetheneedforclassifyingsuchkingroupsinamannerthatmakesthemcoordinateinmoststructuralrespectswithunilineardescentgroups.Theterm"descentgroup"servesequallywellforbothunilinearandnon-unilineartypes,andinaccordancewithcurrentusage,whentheyareunilineartheyarecalledsibsorlineages,dependingoninternalsegmentationandgenea-logicaldepth.Fornonunilineardescentgroupstheterm"sept"isproposed.Thoughemployedbyothersinvarioussenses,thistermisusedbyBoas(1920:114-115)fortheKwakiutlnumayminanattempttodistinguishitfromuni-linearkingroupselsewhereontheNorthwestCoast.Theterm"clan'm1maybeusedforthelocalizedcoreofsibandseptplustheirinmarryingspousesandotherdependents.Thefactthatseptsarecomparabletosibscanbedemonstratedbysomearealdistributionsandtheevolutionarytrendstobeinferredfromthem.TheseptcanbeconsideredoneofthedistinguishingtraitsofsocialorganizationofPolynesia,fortheyarefoundinTonga,Samoa,theSocieties,theCookIslands,NewZealand,theMarquesas,EasterIsland,andHawaii.Inalltheseculturesverysimilaractivitiesarefoundtobeorganizedbythissystem.Amongtheso-calledoutliersofPolynesia,however,wefindcertainofthesesameactivitiesorganizedbyunilineardescent.Tikopia(Firth1936)organizessomeofthembypatrilinealdescent,Kapingamarangi(Emory1956)regulatescertainoftheminpartthroughmatrilinealdescent,whileOntongJava(Hogbin1934)organizesbyboth,i.e.bydoubledescent.Pukapuka(Beaglehole1938),thoughnotstrictlyanoutlier,resemblesOntongJava,andinadditionorganizesathirdsetofactivitiesbynonunilineardescent.AsimilararealvariationwasnotedmanyyearsagobyBoasontheNorth-westCoastofNorthAmerica,butunlikePolynesia,itcannotbeassumedthatthisdifferentiationoccurredwithinonegeneticgroup,speakingcloselyrelatedlanguagesofasinglelinguisticstock.Letusnowturntotheothermajorformofbilateralorganization,thekin-dred.Herethemeansofascribingandrestrictingstatusesdiffersradicallyfromthatwhichwehavetermedthesept.Eachpersonhashisownkindred,thepersonnelandboundariesofwhichcoincideonlywiththoseofhissiblings,andsiblingsaremembersofasmanydifferentkindredsastherearekintypesintheirown(seeMurdock1949:56-58).InmakingthisdistinctionbetweenThiscontentdownloadedfrom182.242.229.207onWed,13May201510:35:37UTCAllusesubjecttoJSTORTermsandConditionsDAVENPORT]NonunilinearDescent563kingroupswhichareidenticalforsiblingsonlyanddescentgroupswhoseper-sonnelisthesameforallmembers,Phillpotts(1913:268,275)usedfortheformertheterms"shiftingkindred"and"fluctuatingkindred,"whilemorerecentlyLeach(1950:61-62)hassuggested"personalkindred."Thesetermsmeritseriousconsiderationinthisareawhereterminologicalconfusionhasbeentheruleandnottheexception.Thetermpersonalkindredwillbeadoptedhere,sinceitmostclearlydescribesthemostsalientfeatureofthiskindoforganization.OneofthemostcompletedescriptionsofpersonalkindredscomesfromtheIfugao.Herethekingroup,or"family"asBartonT2hascalledit,consistsofallthedescendantsofanindividual'seightpairsofgreat-grandparents.Itmayincludeasmanyastwothousandpersons,manymorethaninmostoftheseptsalreadydescribed.Indiscussinglegalrelationships,Bartonclearlyshowsthattheonlykintypeswiththesame"family"compositionaresiblings.Theex-tentofone'sinvolvementinlegaldisputesisapositivefunctionofone'sgenea-logicalclosenesstothelitigants.Inreferencetoproperty,arichmanwhoownsmuchlandhasadefiniteobligationtoprovidelandforhislesswell-situatedkindredmembers,anobligationwhichdecreasesasafunctionofcollateraldistance.TheKalingahaveasimilarorganization,althoughsomeformofseptmayalsobepresentinconnectionwiththeendogamoustribaldistricts.13Leach(1950)presentssomeinterestingcomparativedataonpersonalkin-dreds,asdefinedbymarriageregulations,amongseveralpeoplesofSarawak.TheSemambuMurutprefermarriagewithkinwhoaremoredistantthanfourthcousins,orjustwithinthisrange,i.e.,withthirdcousins,bypayingafineforincest,butmarriagewithfirstcousinsisabsolutelyprohibited.TheIban,however,prescribemarriagewithsecondtofifthcousinsandprohibitmarriagewithbothfirstcousinsandkinbeyondfifthcousins.ThegeneralizedversionofEskimoculturepresentedbyHoebel(1954:67-99)revealsapersonalkindredrangeforextensionoftheincesttaboo,obliga-toryactioninbloodrevenge,andotherkin-organizedstatuseswhichextendsonlytolinealsandsiblings.WiththecontemporaryNunamiutEskimoofAlaska,Pospisil(1957)reportsthattheincesttabooisextendedonlytopri-maryrelatives,allotherkinbeingeligibleasspouses.Strictobligationstoat-tendlife-crisisritesandtoextendspecialhospitalityareextendedtoawiderrangeofkinsmen,includinglinealstogreat-grandparentsandgreat-grand-children,firstcollateralstosiblings'great-grandchildren,secondcollateralstochildrenoffirstcousins,thirdcollateralstosecondcousins,andfourthcollat-eralstosiblingsofgreat-grandparents,butneverbeyondtheselimits.Allrela-tivesoutsidethisscope,eventhoughtheirrelationshipisknown,aresociologi-callynon-kin.Clearly,themostvariablefeatureoftheper