The Nayars and the Definition of Marriage

Author(s): E. Kathleen Gough

Reviewed work(s):
Source: The Journal of the Royal Anthropological Institute of Great Britain and Ireland, Vol.
89, No. 1 (Jan. - Jun., 1959), pp. 23-34
Published by: Royal Anthropological Institute of Great Britain and Ireland
Stable URL: http://www.jstor.org/stable/2844434 .
Accessed: 23/03/2012 09:28

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 support@jstor.org.

Royal Anthropological Institute of Great Britain and Ireland is collaborating with JSTOR to digitize, preserve
and extend access to The Journal of the Royal Anthropological Institute of Great Britain and Ireland.

http://www.jstor.org
 The Nayars and the Definitionof Marriage
E. KATHLEEN GOUGH
THE PROBLEM OF A SATISFACTORY DEFINITION OF MARRIAGE has fordecades
vexed anthropologists
and has beenraised,butnotsolved,severaltimesin recentyears.'Overtimeit becameclear
thatcohabitation, ritualrecognition,definitionof sexualrightsor stipulation of domestic
services
eachhadtoolimited a distribution
toserveas a criterionforall theunionsanthropologists
intuitivelyfeltcompelledto call 'marriage'.For good reasontherefore theNotesandQueries
ofI95I makesno reference
definition toanyofthese:'Marriageis a unionbetween a mananda
womansuchthatchildren borntothewomanarerecognized legitimate ofbothparents.'
offspring
Admirably concisethoughit is,thisdefinition
tooraisesproblems in a numberofsocieties.
The Nuerinstitution of woman-marriage-to-a-woman wouldbe a case in point.Here,both
partiesto theunionarewomenyet,as Evans-Pritchard (I95I, pp. IO8-9)has shown,thelegal
provisionsoftheunionarestrictly comparable to thoseofsimplelegalmarriage betweena man
and a woman.Few therefore wouldquestionEvans-Pritchard's logicin callingthisuniona
marriage.
The NotesandQueries definitioncontainstwocriteria:thatmarriage is a unionbetween
one man and one woman,and thatit establishes the legitimacy of children.Nuerwoman-
marriage doesnotconform tothefirstcriterionbutitdoestothesecond.Atthispointtheproblem
therefore becomes:is a definition
feasiblewhichwouldinsistonlyon thesecondcriterion, that
oflegitimizing children?
In Europe,2 Dr EdmundLeachinitiated themostrecentchapterin thisdiscussion (Leach
I955), and ratherthanreviewitswholehistory it is pertinent forme to takeup theargument
wherehe andothers haveleftit.In effect,Dr Leachanswered 'no' tothequestion posedabove.
He arguednotonlyagainstthevagueness ofthephrase'legitimate offspring'butalso against
anyuse ofpotential legalpaternity as a universal criterionofmarriage. He concludedin fact
thatno definition couldbe foundwhichwouldapplyto all theinstitutions whichethnographers
commonly referto as marriage.Insteadhe namedtenclassesofrights3 whichfrequently occur
in connection withwhatwe looselytermmarriage, addedthat'onemight perhapsconsiderably
extendthislist',and seemedto concludethatsinceno singleone oftheserights is invariably
established by marriage in everyknownsociety, we oughtto feelfreeto call 'marriage'any
institutionwhichfulfilsanyone or moreoftheselectedcriteria.
Thereis,surely, a quitesimplelogicalflawin thisargument. Forit wouldmeanin effect
thateveryethnographer might extendat willDr Leach'slistofmarital andinshortdefine
rights,
marriage in anywayhe pleased.Thismaybe legitimate in describing a singlesociety. ButI
wouldarguethatforpurposes ofcross-cultural
comparison, we do needa single,parsimonious
definition,simplyin orderto isolatethephenomenon we wishto study.
In supportofhisargument againstusingthelegitimizing ofchildren asa universal criterion
ofmarriage, Dr LeachcitedtheNayarcase.On thebasisoftwoof mypaperson theNayars
(GoughI952, I955a), he statedthattheNayarstraditionally had 'no marriage in thestrict
(i.e. NotesandQueries)senseofthetermbutonlya "relationship ofperpetual affinity"between
linkedlineages(GoughI955a). The woman'schildren, however theymightbe begotten, were
simply recruitsto thewoman'sownmatrilineage.' He statedfurther, 'The notionoffatherhood
is lacking.The childusesa termofaddressmeaning "lord"or "leader"towards all itsmother's
lovers,buttheuseofthistermdoesnotcarrywithit anyconnotation ofpaternity, eitherlegal
orbiological. On theotherhandthenotionofaffinity is present,as evidenced bythefactthata
womanmustobserve pollutionat herritualhusband's death(GoughI955a).' LaterDr Leach
concludesthat 'among the matrilinealmatrilocalNayar,as we have seen,rightJ (to establish
24 E. KATHLEEN GOUGH

a socially significant"relationship of affinity"between the husband and his wife's brothers)

is the only marriage characteristicthat is present at all' (Leach I955, p. I83).
This paper has two objectives. It will begin by analyzing traditional Nayar marital institu-
tions and therebyshowing that in fact the notion of fatherhoodis not lacking and that marriage
does serve to establish the legitimacyof children. My analysis will, I hope, not only dispose of a
misinterpretation on Dr Leach's part, but will in general clarifywhat has always proved a crucial
but difficultborderlinecase fortheoristsofkinship.The paper will conclude with a new definition
of marriage which will again make the status of children born to various types of union critical
for decisions as to which of these unions constitutemarriage. The ultimate aim is not of course
to re-definemarriage in a dogmatic way to suit a particular case, for definitionsare tools of
classificationand not aims of research. The aim is to show that there is a common element not
only in the institutionsanthropologistshave confidentlylabelled 'marriage' by the Notes and
Queriesdefinition,but also in some unusual cases to which that definitiondoes not apply. Whether
we call the element 'marriage' does not much matterprovided it is made explicit,but it would
probably be convenient to do so.

NAYAR MARRIAGE IN CENTRAL KERALA

This account will referto Nayars in the formerkingdomsof Calicut, Walluvanad, and Cochin
in the centre of the Malabar Coast or Kerala. In the northernmostkingdoms (Kolattunad,
Kottayam) and probably also in the southernmostkingdom of Travancore, Nayar residence
appears to have been avunculocal even beforethe period of Britishrule, marriage was optionally
polygynousbut not polyandrous,and individual men appear to have had definiterightsin and
obligations to their children. Full informationis not available for these northernmostand
southernmostkingdoms in the pre-Britishperiod. But it seems probable that in the northern
kingdomsat least, even the Notesand Queriesdefinitionof marriage was applicable to the Nayars.
It was certainlyapplicable in the latterhalf of the nineteenthcenturyforwhich I have accounts
from informants.
My account of marriagein the centralkingdomsis a reconstructionof a state of affairswhich
appears to have been general before I792 when the Britishassumed governmentof the Coast.
As I have shown elsewhere (Gough I952) Nayar kinship was slowly modifiedin the nineteenth
century and more rapidly in the twentieth.But in remote villages the traditional institutions
persisted until towards the end of the nineteenthcentury and were remembered by a few of
my older informants.Their reports are not contradicted and are substantially corroborated
by writingsof Arab and European travellersof the fifteenthto eighteenthcenturies.
In this account I shall use the terms 'marriage', 'husband' and 'wife' without definition.
My reasons for doing so will appear later.
In each of the threecentral kingdomsthe Nayar caste was divided into a number of ranked
subdivisions characterized by differentpolitical functions. Chief of these were (a) the royal
lineage, (b) the lineages of chiefsof districts,(c) the lineages of Nayar village headmen and (d)
several sub-castes of commoner Nayars. Each of these last either served one of the categories
(a) to (c) or else served patrilineal landlord familiesof Nambudiri Brahmans. I shall deal first
with the commoner Nayars of category (d).
There were presentin each village some four to seven exogamous matrilineagesof a single
sub-caste of commoner Nayars. These owed allegiance to the familyof the head of the village,
which might be a patrilineal Nambudiri family, a Nayar village headman's matrilineage, a
branch of the lineage of the chiefof the district,or a branch of the royal lineage. The commoners
held land on a hereditaryfeudal-typetenure from the headman's lineage and, in turn, had
authority over the village's lower castes of cultivators,artisans, and agricultural serfs. Each
retainer lineage tended to comprise some four to eight property-owningunits which I call
property-groups.The property-groupformed a segment of the total lineage and was usually
composed of a group of brothersand sisterstogetherwith the children and daughters' children
of the sisters.The members owned or leased propertyin common, lived in one house, and were
 THE NAYARS AND THE DEFINITION OF MARRIAGE 25

underthelegal guardianship oftheoldestmale (karanavan) ofthegroup.Boththeproperty-group

and the lineagewere called taravdd.
Nayar men trainedas professional soldiersin villagegymnasia,and forpart of each year
theytendedto be absentfromthevillagein warsagainstneighbouring kingdomsor formilitary
exercisesat thecapitals.Only thekdranavan, the womenand the childrenof the property-group
remainedpermanently in theirancestralhomes.
The Nayarsofone villageor of two adjacentvillagesformeda neighbourhood group (kara
or tara)of some six to ten lineages.Each lineage was linkedby hereditary of
ties ceremonial
co-operation withtwoorthreeotherlineagesoftheneighbourhood. Theselinkageswerereciprocal
but not exclusive,so thata chain of relationships linkedall the lineagesof the neighbourhood.
The lineageslinkedto one's own werecalled enangar; thetotalneighbourhood group,theenangu.
Atleastone manand one womanofeach linkedlineagemustbe invitedto thehouseofa property-
groupforthelife-crisis ritesofitsmembers.Itslinkedlineageswerealso concernedifsomemember
of a lineagecommitteda breachof the religiouslaw of the caste. It was theirdutyat once to
breakoffrelationswiththeoffending lineageand to call a neighbourhood assemblytojudge and
punishthe offence.Its linkedlineagesthusrepresented the neighbourhood groupas a wholeto
theoffending lineageand werespecialguardiansofitsmorality.Sometimesin smallneighbour-
hoodsthecommonerNayarlineageswereall enangar to each other,but in largerneighbourhoods
thiswas not feasible,forthe heads of property-groups would have had too manyceremonial
obligationsto fulfil.
The linkedlineagesplayed theirmostimportantrole at the pre-puberty marriagerites
of girls(Gough I 955b). At a convenienttimeeveryfewyears,a lineageheld
(talikettukalydnam)
a grandceremony at whichall itsgirlswhohad notattainedpuberty, aged aboutsevento twelve,
wereon one day rituallymarriedby men drawnfromtheirlinkedlineages.The ritualbride-
groomswere selectedin advance on the advice of the village astrologerat a meetingof the
neighbourhood assembly.On theday fixedtheycame in processionto theoldestancestralhouse
of the hostlineage.There,aftervariousceremonies, each tied a gold ornament(tdli)roundthe
neckofhis ritualbride.The girlshad forthreedayspreviously been secludedin an innerroom
of the house and caused to observetaboos as if theyhad menstruated. Afterthe tJli-tyingeach
couplewas secludedin privateforthreedays.I was toldthattraditionally, ifthegirlwas nearing
puberty, sexualrelationsmighttakeplace. This custombeganto be omittedin thelate nineteenth
century,but fromsome of the literatureit appearsto have been essentialin the sixteenthand
seventeenth centuries.At the end of the periodof seclusioneach couple was purifiedfromthe
pollutionof cohabitationby a ritualbath. In Calicut and Walluvanad each couple in public
thentorein twotheloin-clothpreviously wornby thegirlduringthe 'cohabitation'period,as a
tokenof separation.This riteappears to have been omittedin Cochin. In all threekingdoms
however,theritualhusbandsleftthehouseafterthefourdaysofceremonies and had no further
obligationsto theirbrides.A bridein turnhad onlyone further obligationto herritualhusband:
at his death,she and all her children,by whateverbiologicalfather,mustobservedeath-pollu-
tionforhim. Death-pollution was otherwiseobservedonlyformatrilinealkin. In Cochin,even
iftheirmother'sritualhusbandnevervisitedhiswifeagain,herchildrenmustreferto himby the
kinshiptermappan.Childrenin thelower,patrilinealcastesofthisarea used thiswordto referto
thelegal father,who was presumedalso to be thebiologicalfather.In Walluvanadand Calicut
I did nothearofthisverbalusageand do notknowbywhatterm,ifany,Nayarchildrenreferred
to theirmother'sritualhusband.
The pre-puberty tali-ritewas essentialfora girl. If she menstruated beforeit had been
performed, she shouldin theorybe expelledfromher lineageand caste. In fact,however,my
informants told me thatin such a case the girl'sfamilywould conceal the factof her maturity
untilafterthe ritehad been performed. But it was a gravesin to do so and one whichwould
neverbe publiclyadmitted.
The tali-rite
markedvariouschangesin thesocial positionofa girl.First,it broughther to
socialmaturity. She was now thoughtto be at leastrituallyendowedwithsexualand procreative
26 E. KATHLEEN GOUGH

functionsand was thenceforward accordedthestatusofa woman.Aftertheritepeopleaddressed

her in public by the respectful titleammameaning'mother',and she mighttake part in the
ritesof adult women.Second, afterthe tali-ritea girl mustobserveall the rulesof etiquette
associatedwithincestprohibitions in relationto menofherlineage.She mightnot touchthem,
mightnot sitin theirpresence,mightnot speakfirstto themand mightnot be alone in a room
withone of them.Third,afterthe tdli-rite and as soon as she becomeold enough(i.e. shortly
beforeor afterpuberty),a girlreceivedas visitinghusbandsa numberof men of her sub-caste
fromoutsideher lineage,usuallybut not necessarily fromher neighbourhood. In additionshe
mightbe visitedby any Nayar of the highersub-castesof villageheadmen,chiefsor royalty,
or by a NambudiriBrahman.All of theserelationships were called sambandham.Amongcom-
monerNayar women,however,the greatmnajority of unionswerewithmen of commonersub-
caste.
Relationsbetweenany Nayarwomenand a man oflowerNayar sub-caste,or betweenany
Nayar womanand a man of one of the lower,non-Nayarcastes,were strictly prohibited.If a
woman was foundguiltyof such a relationship her lineage'senangarcarriedthe matterto the
neighbourhood assembly.This temporarily excommunicated the woman'sproperty-group until
justicehad been done. In thenineteenth centuryand earlythiscenturytheproperty-group was
re-acceptedintocasteonlyafteritskdranavan had dismissedthewomanfromherhouseholdand
caste,neverto return.In pre-British timesa woman so dismissedbecame the propertyof the
kingor chiefand mightbe soldintoslaverywithforeign however,themen
traders.Alternatively,
ofherproperty-group had theright,sometimes exercised,to killboththewomanand herlover
and thuspreservethegood name oftheirlineage.
Aftertheritualmarriagethebridegroom need have no further contactwithhis ritualwife.
If bothpartieswerewilling,however,he mightenterinto a sexual relationship withhis ritual
brideaboutthetimeofherpuberty.But he had no priority overothermenoftheneighbourhood
group.Thereis someuncertainty as to thenumberofvisitinghusbandsa womanmighthave at
one time.Writersofthesixteenth and seventeenth centuriesreportthata womanusuallyhad some
threeto eightregularhusbandsbut mightreceiveothermen of her own or a highercaste at
will.Hamiltonin I727 statedthata womanmighthave as husbands'twelvebut no moreat one
time' (Hamilton I727 I, p. 3Io). As late as I807 Buchanan reported that Nayar women vied
with each other as to the number of lovers they could obtain (Buchanan I807 I, p. 4I i). A
fewof my older informantscould rememberwomen who had had threeor fourcurrenthusbands,
although plural unions were being frownedupon and had almost died out by the end of the
last century.There appears to have been no limitto the number of wives of appropriate sub-caste
whom a Nayar might visit concurrently.It seems, therefore,that a woman customarilyhad a
small but not a fixed number of husbands from within her neighbourhood, that relationships
with these men might be of long standing, but that the woman was also free to receive casual
visitorsof appropriate sub-caste who passed throughher neighbourhoodin the course of military
operations.
A husband visited his wifeaftersupper at night and leftbeforebreakfastnext morning.He
placed his weapons at the door of his wife'sroom and if otherscame later theywere freeto sleep
on the verandah of the woman's house. Either party to a union might terminateit at any time
without formality.A passing guest recompensed a woman with a small cash giftat each visit.
But a more regular husband fromwithinthe neighbourhoodhad certain customaryobligations.
At the start of the union it was common although not essential for him to present the woman
with a cloth of the kind worn as a skirt.Later he was expected to make small personal giftsto her
at the three main festivalsof the year. These giftsincluded a loin-cloth,betel-leaves and areca-
nutsforchewing,hair-oiland bathing-oil,and certainvegetables. Failure on the part ofa husband
to make such a giftwas a tacit sign that he had ended the relationship.Most important,however,
when a woman became pregnant it was essentialfor one or more men of appropriate sub-caste
to acknowledgeprobable paternity.This theydid by providinga feeofa cloth and some vegetables
to the low caste midwifewho attended the woman in childbirth.If no man of suitable caste would
 THE NAYARS AND THE DEFINITION ON MARRIAGE 27

consent to make this gift,it was assumed that the woman had had relationswith a man of lower
caste or with a Christian or a Muslim. She must then be either expelled fromher lineage and
caste or killed by her matrilinealkinsmen.I am uncertainof the precise fate of the child in such
a case, but there is no doubt at all that he could not be accepted as a member of his lineage
and caste. I do not know whether he was killed or became a slave; almost certainly,he must
have shared the fate of his mother.Even as late as I949, over a hundred and fiftyyears afterthe
establishmentof Britishrule, a Nayar girl who became pregnant before the modern marriage
ceremonywas regarded as acting withinthe canons of traditionalreligiouslaw ifshe could simply
find a Nayar of suitable sub-caste to pay her deliveryexpenses. But if no Nayar would consent
to this she ran the danger of total ostracism,with her child, by the village community.I heard
of several cases in which such a girl was driven fromher home by her karanvanat the command
of the sub-caste assembly. Her natal kinsmenthen performedfuneralritesfor her as if she had
died. In each case the girl took refugein a town before or shortlyafterher child was born.
Althoughhe made regulargiftsto her at festivals,in no sense of the termdid a man maintain
his wife. Her food and regular clothingshe obtained fromher matrilinealgroup. The giftsof a
woman's husbands were personal luxuries which pertained to her role as a sexual partner-
extra clothing, articles of toilet, betel, and areca-nut the giving of which is associated with
courtship,and the expenses of the actual delivery,not, be it noted, of the maintenance of either
mother or child. The giftscontinued to be made at festivalsonly while the relationshiplasted.
No man had obligations to a wifeof the past.
In these circumstances the exact biological fatherhood of a child was often uncertain,
although, of course, paternitywas presumed to lie with the man or among the men who had
paid the delivery expenses. But even when biological paternity was known with reasonable
certainty,the genitor had no economic, social, legal, or ritual rightsin nor obligations to his
children afterhe had once paid the fees of their births.Their guardianship, care and discipline
were entirelythe concern of theirmatrilinealkinsfolkheaded by theirkdranavan. All the children
of a woman called all her currenthusbands by the Sanskritword acchanmeaning 'lord'. They did
not extend kinship terms at all to the matrilineal kin of these men. Neither the wife nor her
childrenobservedpollution at the death of a visitinghusband who was not also the ritualhusband
of the wife.
In most matrilineal systems with settled agriculture and localized matrilineal groups,
durable links are provided between these groups by the interpersonalrelationshipsof marriage,
affinityand fatherhood.The husbands, affines,fathers,and patrilateral kin of members of the
matrilineal group have customaryobligations to and rightsin them which over time serve to
mitigate conflictsbetween the separate matrilineal groups. The Nayars had no such durable
institutionalizedinterpersonal links. This does not mean that men did not sometimes form
strongemotional attachmentsto particular wives and their children. My informationindicates
that theydid. I know forexample that if a man showed particular fondnessfora wife,his wife's
matrilineal kin were likely to suspect the husband's matrilineal kin of hiring sorcerersagainst
them. For the husband's matrilinealkin would be likelyto fear that the husband mightsecretly
convey to his wifegiftsand cash which belonged rightfully to his matrilinealkin. This suspicion
was especially rifeif the husband was a karanavanwho controlled extensiveproperty.Informal
emotional attachmentsdid thereforeexist between individuals of differentlineages. But what I
wish to indicate is that among the Nayars, these interpersonalaffinaland patrilaterallinks were
not invested with customarylegal, economic, or ceremonial functionsof a kind which would
periodically bring members of differentlineages togetherin mandatory formsof co-operation.
Four special kinship termsdid apparently exist for use in relation to affinesacquired through
the sambandham relationship,although, as I have said, there were no patrilateral termsfor kin
other than the mother'shusbands. All men and women currentlyengaged in sambandham unions
with members of ego's property group, and all members of the property-groupsof these
individuals, were collectivelyreferredto as bandhukkal('joined ones'). A currentwife of ego's
mother's brotherwas addressed and referredto as ammayi,and a wife of the elder brother as
28 E. KATHLEEN GOUGH

jyeshtatiamma(lit. 'elder-sister-mother').Finally,the own brotherand the sambandham husband

ofa womanemployedthereciprocaltermaliyanto referto each otherbutusedno termofaddress.
All thecurrentbandhukkal ofa property-group wereinvitedto householdfeasts,but as individual
affinestheyhad no ceremonialor economicobligationsand were not obliged to attend.As
representatives ofenangar lineages,however,someof thesesame individualsmightbe obligedto
attendfeastsand to fulfilceremonialobligationsas enangar. But as particularaffines
theyhad no
obligations.In place therefore, of institutionalized
interpersonal patrilateraland affinallinks,
the Nayars had the hereditary institutionof linkedlineages.Whetheror not, at a particular
time,sexualrelationships existedbetweenindividualsoflinkedlineages,thelinkedlineagesmust
fulfiltheirobligationsat householdceremoniesand give neighbourly help in such emergencies
as birthand death. In the patrilinealand double unilinealcastesof Kerala preciselythe same
obligationsare fulfilled
by thematrilateral kinand affinesofindividualmembers ofthepatrilineal
group. The linkedlineagesof the Nayars must I
therefore, think,be regardedas having a
relationshipof 'perpetualaffinity', which carried the more formalfunctionsof affinity and
persistedthroughthemakingand breakingofindividualsexualties.
In view of thesefacts,it is convenientto mentionhere that Dr Leach's statementthat
Nayar marriageserved'to establisha sociallysignificant relationshipbetweenthe husbandand
his wife'sbrothers'is not,strictlyspeaking,correct.The sambandham uniondid not establish'a
sociallysignificant relationship'betweenbrothers-in-law, forin spiteof the reciprocalkinship
termthesepersonshad no institutionalized obligationsto one anotherby virtueoftheparticular
sambandham tie. Further,the tdli-rite a relationship
did not establish betweenthe ritualhusband
and the brothersof his ritualbride.The ceremonyset up no specialobligationsbetweenthese
persons;it was merelythattheirlineageswere,hereditarily, enangar,both beforeand afterany
particulartdli-rite.What theritedid establish was a ritualrelationshipbetweenthe tdli-tierand
his ritual bride, and, as I shall tryto show later, a relationshipof group-marriagebetween the
bride and all men ofher sub-casteoutside her lineage. But a particulartdli-rite in no way modified
the hereditaryrelationshipsbetween male enangar.It is for this reason that I call the enangar
relationshipone of 'perpetual affinity'between lineages,which, though it carried the ceremonial
functionsof affinity, persistedirrespectiveof particular sambandhams and tdli-rites.
The Nayars of this area were thus highlyunusual. For they had a kinshipsystemin which
the elementaryfamilyoffather,motherand childrenwas notinstitutionalizedas a legal, productive,
distributive,residential,socializing or consumption unit. Until recent years, some writershave
thought that at least as a unit for some degree of co-operation in economic production and
distribution,the elementaryfamilywas universal. This view has been put forwardmost forcibly
by Murdock (Murdock I949, chapter I). Radcliffe-Brown,however, was one of the earliest
anthropologiststo observethat ifthe writtenaccounts of the Nayars were accurate, the elementary
familywas not institutionalizedamong them.4My research corroborateshis findings.
I turn brieflyto marital institutionsamong the higherNayar sub-castesof village headmen,
districtchiefs,and royalty.At various times during the pre-Britishperiod these lineages were
accorded political officeand set themselvesup as ofhigherritualrank than the commonerNayars.
The ritual rankingbetween thesemajor aristocraticsub-divisionswas fairlystable, but the mutual
rankingof lineages withineach sub-divisionwas in dispute. Most village headmen acknowledged
the ritual superiorityof districtchiefs,and most chiefs,of the royal lineage. But some village
headmen disputed among themselvesfor ritual precedence and so did many chiefs.As a result,
each of these aristocraticlineages tended to set itselfup as a separate sub-caste, acknowledging
ritual superiors and inferiorsbut acknowledging no peers. In the course of time, moreover,
followingthe vicissitudesofpolitical fortune,such lineages could rise or fall in the ritual hierarchy.
It was in these lineages thereforethat hypergamousunions became most highlyinstitutionalized,
for most of these lineages refusedto exchange spouses on equal terms. Instead, most of them
married all their women upwards and all their men downwards. Women of village headman's
lineages entered sambandham unions with chiefly,royal, or Nambudiri Brahman men. Men of
these lineages had unions with commoner Nayar women. Chieflywomen had unions with royals
 THE NAYARS AND THE DEFINITION OF MARRIAGE 29

or Nambudiris; chieflymen, with the women of village headmen's or commonerNayar lineages.

Royal women for the most part had unions with Nambudiri Brahmans of the highestrank. A
few, especially in Calicut, however, had unions with men of older and rituallyhigher ranking
royal lineages which had throughconquest become politicallysubordinate to theirown. Among
Nambudiri Brahmans, only eldest sons were permittedto marry Nambudiri women and beget
children for their own families.Younger sons of Nambudiri households might have sambandham
unions with Nayar women of any sub-caste.
In all these hypergamous unions the visitinghusband owed the same periodic giftsto his
wifeas in the case of equal unions between personsofthe same commonersub-caste.The husband
in a hypergamousunion was also held responsibleforpayment of deliveryexpenses at the birth
of a child to his wife. Hypergamous unions differedfrom 'equal' unions in that in the former,
the husband, being of higher ritual rank, might not eat in the house of his wife. The husband
was also prohibitedfromtouchinghis wife,her children,or her otherkinsfolkduring the daytime
while he was in a state of ritual purity. Finally, although children called their mother's higher
caste husband by the term acchanplus the caste title,Nayars as a whole were not permittedto
use affiiial terms toward the Nambudiri husbands of their womenfolk,nor did Nambudiris
address or referto theirNayar wives' brothersas affines.Nayars insisthowever that a sambandham
union with a Nambudiri Brahman was of the same character as a sambandhamunion with a
Nayar of equal sub-caste. It seems that fromthe legal point of view we must also judge it to be so,
since the Brahman husband, like the Nayar, was responsibleforpaymentsat the birthof a child
to his Nayar wife. During my fieldwork,the three Nambudiri Brahmans whom I was able to
question closelyon thissubject told me thatfromtheirpoint ofview only marriage to a Nambudiri
woman with Vedic rites could be regarded as true marriage and that sambandham unions with
Nayar women were a kind ofconcubinage. There seemsto me no reason whywe should not regard
theselatterunions as concubinage fromthe point ofview of the Brahmans and (since theyfulfilled
the conditionsof Nayar marriage) marriage fromthe point of view of the Nayars. This seems to
me, in fact, the only possible interpretation,since the Brahmans are patrilineal and the child
of a Brahman-Nayar union is not legitimized into the Brahman caste. The contrastfrom the
Brahman point of view appears most sharplyin the case of an eldest son, who may marryone or
more Nambudiri women with Vedic rites and may also have liaisons with one or more Nayar
women. The Brahman wife's children are of course fully legitimized into the Brahman caste
frombirth.But the Nayar wifeand her childrentraditionallyhad no rightsof patrilineal descent
or inheritance whatsoever,might not enter the kitchen of the Brahman house and might not
touch its inhabitants.
Consistentlywith the differencein directionof sambandham unions, the enangarinstitutionin
these aristocraticNayar lineages differedsomewhat from that in the commoner sub-castes. In
general, an aristocraticNayar lineage had as enangartwo or more lineages of a sub-caste higher
than itselffromwhich its women were wont to draw husbands in the sambandham relationship.
The linked lineage relationshipwas in these cases not reciprocal. A chieflylineage might act
as enangarforthe lineages of one or two village headmen, but had as its own enangarone or two
chieflyor royal lineages of higherrank than itself.Nambudiri Brahman lineages acted as enangar
forthe highestranks of chiefsand royalty.In this case too the aristocraticNayar lineage had of
course no reciprocal ritual obligations towards the Brahman familieswith which it was linked.
The functionsof these aristocraticenangarwere, as far as I can detect, the same as in the case of
commoner Nayars. In particular, men of the higher ranking enangarlineages tied the tali at the
pre-pubertymarriage of aristocraticgirls-appropriately, for it was from these and other such
higher ranking lineages that the girls would later draw visitinghusbands. Plural unions were
customaryin these aristocraticlineages as among commoner Nayars. Obviously, however, the
choice of husbands became more and more restrictedas one ascended the scale of ranked sub-
castes, and at the top of the Nayar hierarchyit was restrictedto Nambudiri Brahmans.
I turnnow to myinterpretationofNayar maritalinstitutions.To accomplishthisitis necessary
to classifythe rightsand obligationsobtaining between 'spouses' and between 'fathers'and their
30 E. KATHLEEN GOUGH

union.
'children'. These fall into two categories: those of the tali-riteand those of the sambandham
In relationsbetweenspousesofthe tali-rite,the importantrightsare thoseofthe woman. The ritual
husband had, it is true, apparently at one time the rightto deflowerhis bride. But the accounts
of many writersindicate that this rightwas not eagerly sought, that in fact it was viewed with
repugnance and performedwith reluctance. The ritual husband also had the right that his
ritual wifeshould mourn his death. But we may assume that this righthad more significancefor
the wife than for the husband, forit was not attended by offeringsto the departed spirit.These
could be performedonly by matrilineal kin. The ritual bride's rightswere complementaryto
her husband's, but for her they were of supreme importance. She had, first,the rightto havea
ritual husband of her own or a superiorsub-caste beforeshe attained maturity.Her lifedepended
on this,forif she was not rituallymarried beforepubertyshe was liable to excommunicationand
mightpossiblybe put to death. She held this claim against her sub-caste as a whole exclusive of
her lineage, or (in the case of aristocraticlineages) against a higher sub-caste. This group must,
throughthe institutionof the linked lineages, provide her with a ritual husband of correctrank
and thus bring her to maturityin honour instead of in shame. It was the duty of her lineage
kinsmen to see to it that some representativefromtheir linked lineages fulfilledthis right. The
ritual wife's second rightwas that of observingpollution at the death of her ritual husband. I
interpretthis as a mark of proofthat she had once been marriedin the correctmanner and that
this ritual relationshiphad retained significancefor herthroughouther ritual husband's life.
The tdli-tierhad no rightsin his ritual wife's children except that they should observe
pollution at his death. From the child's point of view, however,his mother'sritual husband must
have been a figureof great symbolicsignificance.For a child whose motherhad no ritual husband
could notacquire membershipin his caste and lineage at all. The birthofa child beforehis mother's
tali-ritewas absolutelyforbiddenand, in the nature of the case, can scarcelyever have happened.
If it did occur, mother and child must certainlyhave been expelled and were most probably
killed. The child's observance of pollution for his mother's ritual husband-like the use of the
kinship term appan in Cochin-was a formal recognitionthat, for ritual purposes, he had been
'fathered'by a man of appropriate caste.
Turning to the sambandham union, it seems clear that the husband had no exclusive rights
in his wife. He had only, in common with other men, sexual privilegeswhich the wife might
withdraw at any time. Again it is the wife's rightswhich are important.The wife had the right
to giftsfromher husband at festivals,giftsof little economic value but of high prestigevalue, for
they established her as a woman well-favouredby men But most significantwas the woman's
rightto have her deliveryexpenses paid by one or more husbands of appropriate caste, that is,
to have it openly acknowledged that her child had as biological fathera man of required ritual
rank. Her matrilinealkinsmencould if necessarypress for the fulfilmentof this rightin a public
assemblyof the neighbourhood: in cases of doubtfulpaternityany man who had been currently
visitingthe woman could be forcedby the assembly to pay her deliveryexpenses. But if no man
of appropriate rank could be cited as potential father,woman and child were expelled fromtheir
lineage and caste.
The sambandham fatherhad no rightsin his wife's children. Here again, however, the child
had one rightin his possible biological fathers:that one or more of them should pay the expenses
associated with his birth, and thus entitle him to enter the world as a member of his lineage
and caste.
It is clear thereforethat although the elementaryfamilyof one father,one mother and their
children was not institutionalized as a legal, residential, or economic unit, and although
individual men had no significantrightsin their particular wives or children, the Nayars did
institutionalizethe concepts of marriage and of paternity,and gave ritual and legal recognition
to both. It is here that I must contradictDr Leach's interpretationof the situation,forit is not
true that 'the notion of fatherhoodis lacking' nor is it true that 'a woman's children, however
theymightbe begotten,were simplyrecruitsto the woman's matrilineage' (Leach I955, p. I83).
For unless his motherwas rituallymarriedby a man of appropriatecaste and, unless his biological
 THE NAYARS AND THE DEFINITION OF MARRIAGE 3I

paternitywas vouched for by one or more men5 of appropriate caste, a child could never enter
his caste or lineage at all. As I pointed out in both the papers quoted by Dr Leach, the Nayars
were aware of the physiological functionof the male in procreation and attached significance
to it, forthey expected a child to look like his genitor.Like all the higherHindu castes of India,
they based their belief in the moral rightnessof the caste systemin part upon a racist ideology
which involved the inheritanceof physical,intellectual,and moral qualities by a child fromboth
of its natural parents, and which held that the higher castes were, by virtue of their heredity,
superiorto the lower castes. It was ostensiblyforthis reason that the Nayars forbade with horror
sexual contacts between a Nayar woman and a man of lower caste, and that they expelled or
put to death women guilty of such contacts. This racist ideology also provided a motive for
hypergamous unions, for Nayars of aristocraticlineages boasted of the superior qualities they
derived fromroyal and Brahmanical fatherhood.
Moreover, although individual men had no significantcustomaryrightsin their wives and
children, marriage and paternitywere probably significantfactorsin political integration.For
hypergamousunions bound togetherthe highersub-castesofthe political and religioushierarchies.
Multiple sexual ties,as well as the enangarrelationship,linked office-bearing lineages to each other
and to theirretainersin a complicated manner. And Nayar men phrased theirloyaltyto higher
ranking militaryleaders, rulers,and Brahmans in termsof a debt owed to benevolent paternal
figureswhose forebears had collectivelyfathered them and whose blood they were proud to
share. The generalized concept of fatherhoodthus commanded the Nayar soldier's allegiance
to his wider caste unit, to the rulers of his village, chiefdom,and kingdom and to his religious
authorities.It was associated with tender loyaltyand with fortitudein war.
I cannot entirelyblame Dr Leach for underestimatingthe significanceof Nayar paternity
on the basis of his reading of my earlier papers. For in those papers I was concerned to emphasize
the lack of rightsof individual men in theirspouses and children. It is true that in I952 I wrote:
'Marriage . . . was the slenderestof ties, while as a social concept fatherhoodscarcely existed'
(Gough I952, p. 73). I had not then realized the fundamental necessityto a Nayar of having
both a ritual and a biological fatherof appropriate caste. Moreover I myselfconfusedthe issue
by referringto the sambandham partnersas 'husbands' and 'wives' in my firstpaper (Gough I952)
and as 'lovers' and 'mistresses'in my second (Gough I 955a). For it was not until some time after
I read Dr Leach's paper that I decided to classifyNayar unions unequivocally as marriage and
arrived at a definitionof marriage which would include the Nayar case. In my own defence I
must, however, note that in my paper of I955 I mentioned that children must observe death
pollution fortheirmother'sritual husband, and that in Cochin they used the kinshipterm appan
for this ritual father.In both papers quoted by Dr Leach, finally,I noted that sexual relations
were forbiddenbetween a Nayar woman and a man of'lower caste or sub-caste, and that the
currentsambandham husbands of a woman must pay her deliveryexpenses.
I regard Nayar unions as a formof marriage for two reasons. One is that although plural
unions were customary,mating was not promiscuous. Sexual relations were forbiddenbetween
members of the same lineage on pain of death. It was also forbiddenfor two men of the same
property-groupwittinglyto have relations with one woman, or for two women of the same
property-groupto have relations with one man. (This rule of course automatically excluded
relations between a man and his biological daughter.) Further, relations were absolutely
prohibited between a Nayar woman and a man of lower sub-caste or caste. These prohibitions
are directlyconnected with my second and more important reason for regarding these unions
as marriage,namelythattheconceptoflegallyestablishedpaternitywasoffundamentalsignificance
in establishinga child as a member of his lineage and caste.
Granted that Nayar unions constituteda formof marriage,we must I thinkclassifythem as
a clear case of group-marriage.This was the interpretationto which I inclined in I952 (Gough
I952, p. 73) and it is, I now think,the only interpretationwhich makes sense of the descriptive
material I have presented.The tali-rite,as I see it, initiatedforeach individual Nayar girl a state
of marriage to a collectivityof men of appropriate caste. First, the rite ceremonially endowed
32 E. KATHLEEN GOUGH

the girl with sexual and procreativefunctions.(The mock menstrualseclusion beforethe rite is
relevantto this,as is the actual defloration.)Second, the woman's natal kinsmensurrenderedthe
newly acquired rights in her sexuality, though not in her procreative functions,to a male
representativefromoutside her lineage. This appears in that rules of etiquette associated with
incest prohibitionscame into forcefromthis date. Third, rightsin the woman's sexuality were
received by her enanganas representativeof the men of his sub-caste as a whole. This appears in
that the individual enangan,as a special sexual partner,was dismissedat the end of the ceremonies
and might approach the woman again only as one among a series of equal husbands. In the
commoner sub-castes the enanganwas of the same sub-caste as the woman, and throughhim as
representativesexual rightsin the woman were conferredon all men of her sub-caste as a collec-
tivity.They were also in fact extended to any man of highersub-castewho mightfavourher with
his attentions.In aristocraticlineagestheritualhusband was ofa sub-castehigherthan thewoman's,
and throughhim, as representative,sexual rightsin the woman were conferredupon all men ot
higher sub-caste as a collectivity.Fourth, the tdli-rite,by providing the woman with a ritual
husband who (in my view) symbolizedall the men of his sub-caste with whom the woman might
later have relationships,also provided her children with a ritual father who symbolized the
correctnessof theirpaternity.The children acknowledged theirdebt to him by mourningat his
death.
The later sambandhamunions, by this interpretation,involved the claiming of sexual
privilegesby men all of whom were potential husbands by virtue of theirmembershipin a sub-
caste. The husbands had, however, no individually exclusive rightsand could be dismissed at
the woman's wish. Their duties as members of their caste were to provide the woman and her
lineage with children and to acknowledge their potential biological paternity through birth-
paymentswhich legitimizedthe woman's child.

THE DEFINITION OF MARRIAGE

I have called the Nayar unions marriage because they involved the concept of legal paternity.
It is clear however that such a formof group marriage will not fitthe Notesand Queriesdefinition
of 'a union between a man and a woman such that children born to the woman are recognized
legitimateoffspringof both parents' (my italics). For legitimacyin the case of the Nayar child
required both a ritual fatherand a 'legalized genitor' of appropriate rank, and indeed a child
might have more than one 'legal genitor' if two or more men had jointly paid the expenses of
his birth.
As a tentative move toward a new definitionwhich will have cross-culturalvalidity and
will fitthe Nayar and several otherunusual cases, I suggestthe following:'Marriage is a relation-
ship established between a woman and one or more other persons,which provides that a child
born to the woman under circumstancesnot prohibitedby the rules ofthe relationship,is accorded
full birth-statusrightscommon to normal membersof his societyor social stratum.'
A few footnotesto this definitionmay help to vindicate its inevitably clumsy phraseology.
'One or more persons' (in place of 'a man') will bring into the definitionboth group-marriage
of the Nayar type and also true fraternalpolyandry.6It also brings within the definitionsuch
unusual types as woman-marriage-to-a-woman.'Under circumstances not prohibited by the
rules of the relationship'would bring into the definitionvarious problematic cases. It is possible
forexample that there are patrilineal societiesin which a husband may legally repudiate a child
illicitlybegotten upon his wife by another man, without divorcing the wife herself.In this case
the previous establishmentof the marriage would notensure full birth-statusrightsto the child,
for the rules of the marriage relationship would have been broken through the circumstances
which led to his birth.'Full birth-statusrightscommon to all normal members. . .' is a compressed
referenceto all the social relationships,property-rights, etc. which a child acquires at birth by
virtueofhis legitimacy,whetherthroughthe fatheror throughthe mother.For patrilinealsocieties
the phrase 'full birth-statusrights'will include the rightswhich a child acquires in his pateras a
person and in hispater'sgroup. It will include, that is to say, the legitimizationof fatherhood,or
 THE NAYARS AND THE DEFINITION OF MARRIAGE 33
more precisely,of 'father-sonhood'.The phrase is, however,broader than any concept of specific
rightsin a particular father.It will thereforetake care of a case like the Nayar in which all rights
are acquired through the mother but in which a relationship must be established between the
motherand one or more other personsin order forthese matrilinealrightsto be ratified.Such a
process may be called the legitimizationof motherhood,or more preciselyof 'mother-sonhood'.
Moreover 'full birth-statusrights'is, I think,not only broader but more precise than 'recognized
legitimateoffspring',to the vagueness of which Dr Leach took exception. The inclusion of 'society
or social stratum' makes allowances for class or caste systemsin which birth-statusrightsvary
between strata.The case of the Nayars, who are a matrilinealcaste in a predominantlypatrilineal
society,is an obvious example of this.
It should also perhaps be pointed out that this definitiondoes not state that full birth-status
rightscannot be acquired by a child except throughthe marriage of its mother,but only that
marriage provides for the acquisition of these rights.The definitiondoes not thereforeexclude
societies like the Nuer in which a man may legitimize the child of an unmarried woman upon
payment of a legitimization fee, without becoming married to the mother (Evans-Pritchard
I95I, pp. 2I, 26).
Prince Peter has objected to the Notes and Queriesdefinitionand, by implication, to any
definitionwhich would make the legitimizationof children through the mother's relationship
to another party the distinctivecharacteristicof marriage (I956, 46). His reason for objecting
is that in some societies like the Toda, 'marriage and legitimacyof the children can be looked
upon as two differentand separate concepts, and it may be necessaryto go througha ceremony
of legitimizationof the offspring(the Toda pursiitpimi ceremony) in order to establish who is
the legal father,because marriage rites are insufficientin themselvesto do this.'
However, it seems fromRivers' account that preciselywhat distinguishesthe Toda institu-
tion which Prince Peter translates as 'marriage' (mokh-vatt) from that which he translates as
'concubinage' (mokhthoditi) (I957, 35), is that a 'husband' holds the rightto legitimizesome or all
of his 'wife's' children by the pursiitpimi ceremony, whereas a lover in the mokhthoditi union,
being of a differentendogamous group fromthe woman, does not hold this right (Rivers I909,
p. 526). A husband acquires the right to performthe pursuitpimiceremony,it seems, by virtue
of arranged marriage to an infantor throughpaymentof cattle to a formerhusband or to a group
of formerhusbands of the wife. The Toda marriage union at its inception does thereforeprovide
that a child born to the woman (under circumstancesnot prohibitedby the rules of the relation-
ship) mustbe legitimized before his birth; the pursiitpimi ceremony confirmshis legitimacy by
attaching him to a particular fatherand giving him rightsin the father'spatrilineal group. In
the Toda case again thereforethe concept of legal paternityis thedistinguishingcharacteristic
of marriage, even though the individual husband, because of polyandry,may be permittedto
legitimize only some and not all of the children born to his wife. The Toda case thereforefits
my definition,7whether we regard the pursdtpimi ceremony as the final one of a sequence of
marriage rites,or as a legitimizingact which, under circumstancesnot prohibited by the rules
of the relationship,one or another of the woman's husbands is legally obliged to fulfil.
I do not argue that all societiesmust necessarilybe foundto have marriageby my definition.
There may yet turn out to be whole societies-or more probably whole social strata-in which
children acquire no birth-statusrightsexcept throughtheirmother,by the simple fact of birth.
It is possible for example that some slave populations do not have marriage in this sense of the
term. What I do wish to suggesthowever is that for most if not all the societies for which we
now have information,including the Nayar, marriage as I have definedit is a significantrelation-
ship, distinguishedby the people themselvesfromall other kinds of relationships.My definition
should thereforeenable us to isolate marriage as a cross-culturalphenomenon, and fromthereto
proceed to the more exciting task: that of investigatingthe differentialcircumstancesunder
which marriage becomes invested with various other kinds of rightsand obligations. Some of
the most important of these Dr Leach has already listed for us.
34 E. KATHLEEN GOUGH
NOTES

'The fieldworkon which thispaper is based was carriedout in threevillagesof Kerala betweenSeptember
I947 and July I949 withthe aid of a WilliamWyse StudentshipfromTrinityCollege, Cambridge.Writingit has
formedpart of a projectfinancedby the AmericanSocial Science Research Council.
2In AmericaMiss Alisa S. Lourie,Douglass College,RutgersUniversity, has recentlyworkedon thisproblem,
and I have been stimulatedby correspondence withher and by readingan unpublishedpaper of hers,Concepts in
FamilySociology.In thispaper Miss Lourie formulatesa definitionof marriagewhichis narrowerthan mine, but
when her workis publishedreaderswill see that I was helped towardmy definitionby her analysis.I have also
profitedmuch fromdiscussionswith my husband,David F. Aberle.
3A. To establishthe legal fatherofa woman'schildren.
B. To establishthe legal motherofa man's children.
C. To give the husbanda monopolyin thewife'ssexuality.
D. To give the wifea monopolyin the husband'ssexuality.
E. To give the husband partial or monopolisticrightsto the wife'sdomesticand otherlabour services.
F. To give thewifepartialor monopolisticrightsto the husband'slabour services.
G. To give thehusbandpartialor totalrightsoverpropertybelongingor potentiallyaccruingto thewife.
H. To give the wifepartialor total rightsover propertybelongingor potentiallyaccruingto the husband.
L To establisha joint fund of property-a partnership-forthe benefitof the childrenof the marriage.
J. To establisha sociallysignificant'relationshipof affinity'
betweenthe husband and his wife'sbrothers.
(Leach I955, p. I83.)
4Radcliffe-Brown expressedthisview mostrecentlyand fullyin his Introductionto African Systems
ofKinship
andMarriage(I 950, pp. 73 seq.).
5I do not knowwhetherthe Nayars believedit possiblefortwo or more men to contributeto the formation
ofone embryo.I thinkit possiblethattheydid, forI foundthisbeliefamongvillagersoftheTamil country.Among
thesecastesit formedpart of a beliefthatseveralacts of intercourseare necessaryto 'feed' the embryoand assist
it to grow.
6I agree with Dr Leach that the Iravas of CentralKerala had truefraternalpolyandry.My own enquiries
producedevidencesupportingAiyappan'sview that the brotherssharedequally both sexual rightsin the woman
and also legal paternityofthechildren,in thesame mannerin whichtheywereco-ownersoftheancestralproperty.
The eldestlivingbrotherat any given timewas simplythe legal representative of thiscorporation.
7I agreewithDr FischerthatPrincePeter'sdefinition ofmarriageis a tautologyand so ofno assistance(1956,
92). All that PrincePeter'ssecond note shows(I957, 35) is thatseveralpeoples ofhis acquaintancehave different
termsfordifferent kindsofrelationshipsbetweenmen and women.But unlesswe approachthesewithsomeguiding
conceptsofour ownin mind,we cannotdecidewhichofthemto translateas 'marriage'and whichas 'concubinage'.

REFERENCES

BUCHANAN, FRANCIS (Hamilton) I807. A 7ourney

fromMadras through Canaraand Malabar. 3 vols. London.
Mysore,
EVANS-PRITCHARD, E. E. I95I. Kinshipand MarriageamongtheNuer.Oxford.
FIsCHER, H. TH. I956. For a New Definition of Marriage. MAN, I956, 92.
GOUGH, E. KATHLEEN i952. ChangingKinshipUsages in theSettingofPoliticaland EconomicChange amongthe
Nayars of Malabar. J. R. ANTHROP. INST. 82, pp. 7I-87.
GOUGH, E. KATHLEEN I955a. The TraditionalLineage and Kinship Systemof the Nayars. [Unpublishedmanu-
scriptin the Haddon Library,Cambridge.]
GOUGH, E. KATHLEEN I955b. Female InitiationRites on the Malabar Coast, J. R. ANTHROP. INST. 85, pp. 45-8o.
HAMILTON, ALEXANDER I727. A New Account of theEast Indies,2 vols. Edinburgh.
LEACH, E. R. I955. Polyandry,Inheritance,and the Definitionof Marriage. MAN, I955, 199.
LOURIE, ALISA S. I957. 'Conceptsin Family Sociology.' I957. [Unpublishedmanuscriptkindlymade available to
this author.]
MURDOCK, G. P. I949. SocialStructure.
New York.
Notesand Queriesin AnthropologyI95I. 6th ed. London.
H.R.H. PRINCE PETER OF GREECE AND DENMARK For a New Definition of Marriage. MAN, I956, 46; 1957, 35.
RADCLIFFE-BROWN, A. R. & Forde,D. (eds.) 1950. African ofKinshipandMarriage.Oxford.
Systems
RIVERS, W. H. R. I906. The Todas.London.