Skip to main content

Full text of "African systems of kinship and marriage"

See other formats




^JICEL£r^ ^^OFCALIFOfti^ ^OFCAllFOft^ 

^ommin^ "^(^Advyaiv^^^ 

>t m^f) A nv/-i 






i3 i tr^ ^ 





-/^ ^lOSAIICtUt. 

I s 






^ %a3AINn3WV^ %0JITV3JO'^ ^^OJIIVJ-JO"^ 

> = 

o _ 












^ ^OFCAUFOff^ ^^WEUNIVER% ^lOSANff'?' 

2 ^f 






fj^ ^lOSANCE^f^ 5^1UBRARYQ^ 

^ ^lOSANCElij^ 



■^/saaAlN(l3y^^ ^^ojitydjo'^ ^^mmyi^^^ 





ir -5^t-UBRARYGr^ 

^ ^. 




^JIIVOJOV' ^XiUQNV-soi^ '^/ia3MNn-3\^ 


1:^11(5^1 iV/ 

#=\^ ^ 





In July 194S, in response to an invitation from the United 
Nations Educational, Scientific, and Cultural Organiza- 
tion, the International African Institute submitted a 
vietnorandum setting forth particulars of a number of 
research projects lohich the Institute was prepared to 
undertake in collaboration with u.n.e.s.c.o. 

One of these projects related to the publication of a 
volume devoted to studies of kinship and marriage in a 
number of representative African societies. It zvas pro- 
posed that the volume should be published in English and 

In fulfilment of a resolution adopted by the General 
Conference of the United Nations Educational, Scientific, 
and Cultural Organization at its third Session, and in 
accordance with the decisions taken by the Executive 
Board of u.n.e.s.c.o. at its sixteenth Session, a grant 
zvas allocated towards the cost of preparation and publica- 
tion of this volume. 

The English version of the volume, in zohich a nwnber 
of distinguished British and South African anthropo- 
logists have collaborated under the joint editorship of 
Professor A. R. Radcliffe- Brown and Professor Daryll 
Forde, is now presented; the French version is in pre- 
paration and will be available shortly. 




Edited by 




Published for the 


by the 




Oxford University Press, Amen House, London E.C.4 


Geoffrey Ciimberlege, Publisher to the University 



KINSHIP looms large among peoples who obtain their livelihood 
in small groups with simple tools. Among such peoples dif- 
ferences of aptitude and special training and duties do not, as 
in more complex societies, overwhelm the bonds between 
those who are born together and intermarry. The local groups within 
which personal relations are developed in work, rite, and recreation are 
at the same time bodies of relatives who have ancestors in common and 
among whom a complex web of ties links every person with others 
throughout the community .iThe way in which comprehensive obliga- 
tions of kinship direct the activities and relations which, in our society, 
are segregated out as more specifically political, economic, and religious 
is a commonplace of social anthropology. But the detailed operation of 
particular factors, their relative weight in different circumstances, the 
principles under which these can be subsumed, and the real character 
and role of the many patterns of kinship organization that result are 
less clearly grasped. These are, however, fundamental questions in a 
scientific sociology^^hey are also of immediate practical urgency when 
peoples who have lived in a largely kinbound society are reacting to 
pressures and incentives from another social world. For those attempt- 
ing to achieve a smooth transition, and to elicit the energies and loyalties 
of such peoples, the reason for many intangible obstacles and discords 
lies in unintended and often avoidable disharmony between the indi- 
genous and the invading social values. 

The concomitant of culture contact is social strain. The International 
African Institute has sought to promote research in the social anthro- 
pology of Africa in the belief that more adequate resources, intellectual 
and financial, should be devoted to such studies if scientific knowledge 
is to advance on a scale commensurate with the social problems to which 
it should be applied. In an earlier volume on African Political Systems 
some of the results of researches into the political systems of African 
peoples by Fellows and others associated with the Institute were brought 
together. The wide and continuing demand for this study, which has 
afforded insight into the situations with which every student, educator, 
and administrator has to deal, has encouraged an attempt to provide a 
similar conspectus on an equally fundamental aspect of the indigenous 
social life of African peoples. The present volume consists of an Intro- 
duction, in which the general principles underlying African systems of 
kinship and marriage are reviewed in the light of our present knowledge, 
and a series of detailed studies by authors who have made intensive 
field investigations of the particular social systems they analyse. Each 
presents the essential characteristics of one or more varieties of African 
patterns of kinship. 



The purpose of the volume is to present a general view of the nature 
and impHcations of kinship in Africa. It cannot claim to be formally 
comprehensive, for our knowledge is patchy and there are many studies 
— one on the Galla-speaking peoples of north-east Africa might be 
instanced — that remain to be carried out. But much valuable work in 
this field has appeared and is appearing elsewhere and it will become 
easier to remedy the more serious omissions from other sources in the 
future. Meanwhile, save for the peoples of the smallest scale of social 
organization, the Bushmen and Negritos, and the kinship pattern among 
the Amhara of the Ethiopian Kingdom, for none of which studies were 
available, the chief varieties of kinship organization occurring in trans- 
Saharan Africa are illustrated and considered. Their significance is dis- 
cussed by the several authors to whom we express our thanks for their 
ready co-operation in the preparation of this volume. 

January 1950 A. R. RADCLIFFE-BROWN 



PREFACE ....... V 


INTRODUCTION. By a. r. radcliffe-brown, Professor^ ,V 
Emeritus of Social Anthropology, Oxford University . . i 

Importance of scientific study of kinship — An analytical as opposed 
to an historical approach — Definition of kinship — Kinship relations 
exhibited in social behaviour — Structural principles of kinship 
systems illustrated in the Teutonic and the Masai systems — The 
unilineal principle — Classificatory terminology — The generation 
principle — Rank in social relations — ^Omaha classificatory termino- 
logy — Principle of lineage unity expressed in terminology — Privileged 
behaviour — Clans and Jineages — Marriage customs — Early English 
marriage — Modern English and American conceptions of marriage — 
African marriage: marriage payments, marriage as a developing pro- 
cess — Modification of kinship relations by marriage — Rights of 
husband over wife and children — Marriage as an alliance between two 
groups — 'Mother-in-law avoidance' — 'Joking relationships' — Regula- 
tions concerning marriage between relatives — Prohibited and pre- 
ferential marriages — Exogamy — Relation between rules of marriage 
and kinship structure — Incest, parricide, and witchcraft — Mother- 
right and father-right — The social functions of kinship systems. 

KINSHIP AMONG THE SWAZI. By hilda kuper, Lecturer 
in Social Anthropology in the University of the Witzcatersrand . 86 
The clan — The family unit — The homestead — Basic behaviour 
patterns : husband and wife ; parent and child ; siblings — Maternal 
and paternal kin : paternal aunt and maternal uncle ; cousins ; grand- 
parents — Affines: parents-in-law; sisters- and brothers-in-law; the 
groups — Rank and kinship. 

NYAKYUSA KINSHIP. By monica wilson, Professor of Anthro- 
pology, Rhodes College, Grahamstozvn . . . . iii 
Introductory — Kinship groups — Kinship and locality — Bonds of 
kinship and affinity — Cattle and kinship relations — Cattle and mar- 
riage — Dissolution of marriage — Permanence of tie created by 
marriage — Religious interdependence of relatives — Father-in-law 
avoidance — Extensions of kinship behaviour — Kinship and political 
organization — General principles — Maintenance of kinship and 
marriage relationships — Relationship terms. 

I. schapera, Professor of Social Anthropology, University of Cape 
Town ........ 140 

The social setting — The kinship system — Kinship terminology — 
Prohibited and preferential mating — Marriage regulations — Secondary 
unions — Frequency of kinship marriages — Conclusions. 

By MAX GLUCKMAN, Professor of Social Anthropology, Victoria 
University, Manchester, formerly Director of the Rhodes-Livingstone 
Institute . . . . . .166 

Introduction — Kinship structure and the affiliation of children — 
Forms, conditions, and personal consequences of marriage — Property 
consequences of marriage — Marriage payment and the social structure. 




CENTRAL BANTU. By A. i. Richards, Director of the East 
African Institute of Social Science, Makerere, formerly Reader in 
Social Anthropology, London School of Economics . . . 207 

Characteristics of matrilineal kinship organizations in Central Africa — 
Varieties of family structure : a. Alayombe-Kongo : Economic deter- 
minants ; ideology and principles of descent and succession ; the mar- 
riage contract; authority; residential units — b. i. Bemba-Bisa-Lamba 
group: economic determinants; ideology and principles of descent; 
the marriage contract; domestic authority; residential units — b. ii. 
Yao-Cewa group — c. Ila type: principles of descent and succession; 
the marriage contract; domestic authority; residential unit — Varia- 
tions in lineage structure — Conclusions. 


M. FORTES, Professor of Social Anthropology, Cambridge University . 252 

Introduction — Matrilineal descent: political aspect; domestic 
aspect; mother and child — Paternity: concept of Ntoro; father 
and child — Relations between other kin: mother's brother; siblings; 
grandparents ; father's sister — Marriage — Conclusion. 

-;> DOUBLE DESCENT AMONG THE YAKO. By daryll forde, 
Professor of Anthropology, University of London, Director, Inter- 
national African Institute . . . . . .285 

Introduction — Household and family — Patrilineage and clan — Leader- 
ship in patrician and lineage — Adoption into a patrilineage — Matrikin 
and matriclan — Matrilineal rights and obligations — Ritual and politi- 
cal authority of matriclan priests — Yakpan — Terminology of kinship 
— Marriage and affinal relations — Conclusion: Dual sentiments of 

Reader in Anthropology, King's College, Durham University . 333 

Nyaro society: settlement; descent, patrilineal and matrilineal; clan 
rights and obligations ; gens rights and obligations ; family and 
kindred; political and religious institutions; dual principle — Tullishi 
society : settlement ; descent, patrilineal and matrilineal ; myth of 
origin ; sex antagonism — Conclusions. 


NUER. By e. e. evans-pritchard, Professor of Social Anthropology, 
Oxford University . . . . . .360 

Introduction — Villages and camps — Lineage and kinship — Village 
of Konye — ^^Village of Nyueny — Cattle camp at Yakwac — Village of 
Yakwac — Village of Mancom — Village of Kurmayom — Conclusions. 

INDEX 393 










Paternal and Maternal Kin (Swazi) . 

Affinal Kin (Swazi) . 

Zulu Man's Terms for Cognates 

Zulu Man's Terms for Affines 

Zulu Woman's Terms for Affines 

Lozi Man's Terms for Cognates 

Lozi Man's Terms for Affines 

Lozi Woman's Terms for Affines 

Sketch Map of the Territory and Settlements of the 
Yako and their Neighbours 

Woman's House; Man's House 

Composition and Affiliation of Compounds 

The Village of Umor 

Diagram to Show the Location and Relations 
clans in Umor 

Diagram of Matriclan Links in Umor 

Yako Terms for Kin 

Konye Village 

Konye Village — genealogy 

Nyueny Village — genealogy . 

Nyueny Village 

Yakwac Cattle Camp 

Yakwac Cattle Camp — genealogy 

Yakwac Village 

Mancom Village 

of Patri 


' Id 



facing p 




























facing p. 




page 372 



facing p. 


page 381 





'Avoir affaire aux nations sans les connaitre, sans les comprendre, c'est bon 
pour les conquerants; moins bon pour des allies et meme pour les pro- 
tecteurs ; et rien n'est plus detestable et plus insense pour des civilisateurs, 
ce que nous avons la pretention d'etre.' gobineau. 


FOR the understanding of any aspect of the social life of an African 
people — economic, political, or religious — it is essential to have t 
a thorough knowledge of their system of kinship and marriage. 
This is so obvious to any field anthropologist that it hardly needs 
to be stated. But it is often ignored by those who concern themselves 
with problems relating to economics, health, nutrition, law, or adminis- 
tration amongst the peoples of Africa, and it is hoped that this book will ) 
be read not only by anthropologists but by some of those who are respon- 
sible for formulating or carrying out policies of colonial government in 
the African continent. 

A book that would deal thoroughly and systematically with kinship 
organization throughout Africa cannot yet be written. This volume of 
essays is intended only to illustrate by a few samples the results that 
have been reached by social anthropologists at the present time in this 
branch of their studies. It has been thought desirable to add the present 
essay, which offers an introduction to the general comparative and 
theoretical study of kinship organization. 

The literature dealing with kinship is loaded with theories that can 
only be described as pseudo-historical. There are many varieties of such^ _ 
theories, but they all have one thing in common. Starting from some / 
known condition in the present or in the historically recorded past, an 
'explanation' of it is invented by imagining some condition or event in 
the unrecorded past and arguing on a priori grounds that the known 
condition might or must have had its origin in this way. The devotion to 
pseudo-history has had unfortunate results. It has led to the adoption 
of false ideas about the facts as they are, and has often influenced or 
vitiated observation and description. This legacy of erroneous ideas is 
only gradually being got rid of by field studies aiming at the analysis of 
social systems as they are without reference to their origin, where that 
origin is not known from history and can only be conjectured by a priori 

History, the authentic record of events and conditions of the past, is 
entirely different from pseudo-history. If we ask how it is that a society 
has the social institutions that it does have at a particular time the answer 
can only be supplied by history. Where we have records we can trace in 




greater or less detail the way in which the institutions have come to be 
what they are. For European countries we can thus trace the develop- 
ment of social institutions over several centuries. For most African 
societies the records from which we can obtain authentic history are 
extremely scanty or in some instances entirely lacking except for a very 
short period of the immediate past. We cannot have a history of African 

The method adopted here is neither that of history nor of pseudo- 
history but one combining comparison and analysis. Social systems are 
compared so that their differences may be defined and beneath their 
differences more fundamental and general resemblances may be dis- 
covered. One aim of comparison is to provide us with schemes of classi- 
fication. Without classification there can be no science. 

Analysis, as the term is here being used, is a procedure that can only 
be applied to something that is in itself a whole or synthesis. By it we 
separate out, in reality or in thought, the components of a complex whole 
and thereby discover the relation of these components to one another 
within the whole. To arrive at an understanding of kinship systems we 
must use comparison and analysis in combination by comparing many 
different systems with one another and by subjecting single systems to 
systematic analysis. 

A study of kinship systems all over the world by this method reveals 
that while there is a very wide range of variation in their superficial 
features there can be discovered a certain small number of general 
structural principles which are applied and combined in various ways. 
It is one of the first tasks of a theoretical study of kinship to discover 
these principles by a process of abstractive generalization based on 
analysis and comparison. 

There are two sides to science and to the activities of scientists. On the 
one side there is the task of creating and establishing a general theory of 
a certain class of phenomena. For the scientist engaged in this task, or 
while he is so engaged, any particular instance of the phenomena in 
question is only of interest as part of the material that he can use to 
formulate or test his hypotheses. This kind of scientific activity is fre- 
quently referred to as 'pure' or theoretical science. On the other side 
there is the task of applying whatever theoretical knowledge has been 
established to the explanation or understanding of particular phenomena. 
Such activities are often referred to as applied science, and in this sense 
medicine is an applied science based on such theoretical sciences as 
physiology and pathology. For the scientist engaged in a task of this 
kind the aim is not so much to make an addition to general theoretical 
conclusions as to arrive at an interpretation of a particular instance of the 
phenomena with which the science is concerned. The doctor uses his 
theoretical knowledge in order to understand a particular instance of 
disease, to interpret the symptoms, and make a diagnosis and prognosis. 


The two kinds of scientific activity are in close interdependence and 
may on occasion be carried on together. To understand their relation to 
one another, however, it is necessary to distinguish them. In any study 
it is a good thing to know what you are trying to do. 

In the study of social institutions such as the study of kinship, the 
theoretical social anthropologist regards any particular social system as 
supplying him with a body of factual material which he can use for 
formulating or testing theory. But on the other side theoretical knowledge 
can be used to give an understanding of the features of some particular 
system. In the light of the general theoretical knowledge that has re- 
sulted from the comparative study of societies of diverse types the 
scientist may undertake the analysis of a particular system so that 
any single feature is seen in its relation to other features of that system 
and in its place in the system as a whole. The value and validity of 
any such study of a particular system will obviously depend on the 
extent and soundness of the general theoretical concepts by which it 
is directed. 

A system of kinship and marriage can be looked at as an arrangement z*^* 
which enables persons to live together and co-operate with one another 
in an orderly social life. For any particular system as it exists at a certain 
time we can make a study of how it works. To do this we have to conr 
sider how it links persons together by convergence of interest and 
sentiment and how it controls and limits those conflicts that are always 
possible as the result of divergence of sentiment or interest. In reference 
to any feature of a system we can ask how it contributes to the working 
of the system. This is what is meant by speaking of its social functio n, y 
When we succeed in discovering the function of a particular custom, f.e. */ 
the part it plays in the working of the system to which it belongs, we 
reach an understanding or explanation of it which is different from and 
independent of any historical explanation of how it came into existence. 
This kind of understanding of a kinship system as a working system J^ 
linking human beings together in an orderly arrangement of interac-*^^ 
tions, by which particular customs are seen as functioning parts of the y 
social machin ery, is what is aimed at in a synchronic analytic study. In >/ 
such an analysis we are dealing with a system as it exists at a certain 
time, abstracting as far as possible from any changes that it may be 
undergoing. To understand a process of change we must make a dia- 
chronic study. But to do this we must first learn all that we possibly can 
about how the system functioned before the changes that we are in- 
vestigating occurred. Only then can we learn something of their possible 
causes and see something of their actual or probable effects. It is only 
when changes are seen as changes in or of a functioning system that they 
can be understood. 





2, We have first of all to try to get a clear idea of what is a kinship system 
or system of kinship and marriage. Two persons are kin when one is' - 
descended from the other, as, for example, a grandchild is descended 
from a grandparent, or when they are both descended from a common, 
ancestor. Persons are co gnatic kin or cognates when they are descendecTi 
from a common ancestor or ancestress counting descent through malesj 
and females. 

The term 'consanguinity' is sometimes used as an equivalent of 
'kinship' as above defined, but the word has certain dangerous implica- 
tions which must be avoided. Consanguinity refers properly to a physi-"~7 
cal relationship, but in kinship we have to deal with a specifically social J 
relationship. The difference is clear if we consider an illegitimate child" 
in our own society. Such a child has a 'genitor' (physical father) but has 
no 'pater' (social father). Our own word 'father' is ambiguous because it 
is assumed that normally the social relationship and the physical relation- 
ship will coincide. But it is not essential that they should. Social father- 
hood is usually determined by marriage. The dictum of Roman law was 
pater est quern nuptiae demonstrant. There is an Arab proverb, 'Children 
belong to the man to whom the bed belongs' . There was a crude early 
English saying, 'Whoso boleth my kyne, ewere calf is mine'. Social father- 
hood as distinct from physical fatherhood is emphasized in the Corsican 
proverb, Chiamu bahha a chi mi da pane. 

The complete social relationship between parent and child may be 
established not by birth but by adoption as it was practised in ancient 
Rome and is practised in many parts of the world to-day. 

In several regions of Africa there is a custom whereby a woman may 
go through a rite of marriage with another woman and thereby she 
stands in the place of a father (pater) to the offspring of the wife, whose 
hysical father (genitor) is an assigned lover. 

Kinship therefore results from the recognition of a social relationship 
between parents and children, which is not the siH«e thing as the 
physical relation, and may or may not coincide with it. "^'here the term ~: 
'descent' is used in this essay it will refer not to biol6^cal but to social j 
relations. Thus the son of an adopted person will be said to trace descent 
from the adopting grandparents. 

The closest of all cognatic relationships is that between children of the 
same father and mother. Anthropologists have adopted the term 'sib- 
ling' to refer to this relationship; a male sibling is a brother, a female ' 
sibling is a sister.' The group consisting of a father and a mother and 
their children is an important one for which it is desirable to have a 
name. The term 'elementary family' will be used in this sense in this 
essay. (The term 'biological family' refers to something different, 

' In Anglo-Saxon 'sibling' meant 'kinsman'. 


namely, to ^;enetic relationship such as that of a mated sire and dam and 
their offspring, and is the concern of the biologist making a study of 
heredity. But it seems inappropriate to use the word 'family' in this con- 
nexion.) We may regard the elementary family as the basic unit of kin 

ship structure. What is meant by this is that the relationships, of kinship 
or affinity, of any person are all connexions that are traced through his 
parents, his siblings, his spouse, or his children. 

We must also recognize what may conveniently be called 'compound ^ 
families'. Such a family results in our own society when a widower or 
widow with children by a first marriage enters into a second marriage 
into which children are born. This gives such relationships as those of 
half-siblings and of step-parent and stepchild. In societies in which 
polygynous marriages are permitted a compound family is formed when 
a man has two or more wives who bear him children. Families of this 
kind are, of course, common in Africa, and the difference between full 
siblings (children of the same father and mother) and half-siblings 
(children of one father by different mothers) is generally socially im- 

Where there are truly polyandrous marriages, as among the Todas of 
south India, a family may consist of a woman with two or more husbands 
and her children.' But we should distinguish from this an arrangement 
by which the oldest of two or more brothers takes a wife, of whose 
children he will be the father (pater), and access to the wife for sexual 
congress is permitted to the man's younger brother until such time as he 
in turn is married. For it is not sexual intercourse^at constitutes mar- v/ 
riage either in Eurofifb or amongst savage peopleL-^Marriage is a social i^^,,^ 
arrangement by which a child is given a legitiniate position in the society, v^ 
determined by parenthood in the social sense. \ 

The elementary family usually provides the^basis for the formation 

of domestic groups of persons living together in intimate daily life. — 

Of such groups there is a great variety. One common type is what may be 
called the 'parental' family in which the 'household' consists of the 
parents and their young or unmarried children. We are familiar with 
this type of family amongst ourselves, but it is also a characteristic feature 
of many primitive peoples. The group comes into existence with the 
birth of a first child in marriage; it continues to grow by the birth 
of other children ; it undergoes partial dissolution as the children leave it, 
and comes to an end wjth the death of the parents. In a polygynous 
parental family there are two or more mothers but only one father, and 
a mother with her children constitute a separate unit of the group. 

What is sometimes called a patrilineal extended family is formed by a 

custom whereby sons remain in their father's family group, bringing their 

.^ives to live with them, so that their children also belong to the group. 

Among the Bemba of Northern Rhodesia there is found a matrilineal 

' Rivers, The Todas, 1906, p. 515. 


extended family, a domestic group consisting of a man and his wife 
with their daughters and the husbands and children of the latter. The 
group breaks up, and new groups of the same kind are formed, when 
a man obtains permission to leave his parents-in-law, taking his wife 
and children with him. 

Most men who live to maturity belong to two elementary families, to 
one as son and brother, and to the other as husband and father. It is this 
simple fact that gives rise to a network of relations connecting any single 
person with many others. We can get a good idea of this by considering 
what may be called orders of relati onship by kin ship and marriage . . 
Relati onships of the first order a rgj hoae wit hjnjbheelementor^ 
viz. the relation of parent and child, that of husband and wife7and that 
between siblings. Relationships of the second order are those traced 
through one connecting person such as those with father's father, 
mother's brother, stepmother (father's wife), sister's husband, brother's 
son, wife's father, &c. Those of the third order have two connecting 
links, as mother's brother's son, father's sistePs husband, and so on. So 
we can go on to the fourth, fifth, or nth order. In each order the number 
of relationships is greater than that in the preceding order. This net- 
work of relationships includes both cognatic relationships and relation- 
ships resulting from marriage, a person's own marriage, and the 
marriages of his cognates. 

The first determining factor of a kinship system is provided by the 
range over which these relationships are effectively recognized for social 
purposes of all kinds. The differences between wide-range and narrow- 
range systems are so important that it would be well to take this matter 
of range as the basis for any attempt at a systematic classification of 
kinship systems. The English system of the present day is a narrow-range 
system, though a wider range~of relationship, to second, third, or more 
distant cousins, is recognized in rural districts than in towns. China, on 
the contrary, has a wide-range system. Some primitive societies have 
narrow-range systems, others have wide-range. In some of the latter 
a man may have several hundred recognized relatives by kinship and by 
marriage whom he must treat as relatives in his behaviour. In societies 
of a kind of which the Australian aborigines aff(^rd examples every 
person with whom a man has any social contact during the course of his 
life is a relative and is treated in the way appropriate to the relationship 
in which he or she stands. 

Within the recognized range there is some method of ordering the 
relationships, and it is the method adopted for this purpose that gives 
the system its character. Later in this essay we shall consider some of the 
prmciples which appear in the ordering of relatives in different systems. 

A part of any kinship system is some system of terms by which rela- 
tives of different kinds are spoken of or by which they are addressed as 
relatives. The first step in the study of a kinship system is to discover 



what terms are used and how they are used. But this is only a first step. j\ 
The terminology has to be considered in relation to the whole system of 
which it is part. 

.There is one type of termiaology that is usually referred to as 'descrij)- 
tive'. In systems of this type there are a few specific terms for relatives 
of the first or second order and other relatives are indicated by com- 
pounds of these specific terms in such a way as to show the intermediate 
steps in the relation. It is necessary in any scientific discussion of kinship 
to use a system of this kind. Instead of ambiguous terms such as 'uncle' 
or 'cousin' we have to use more exact compound terms such as 'mother's 
brother', 'father's sister's son', and so on. When we have to deal with 
a relationship of the fifth order, such as 'mother's mother's brother's 
daughter's daughter', and still more in dealing with more distant rela- 
tions, the system presents difficulties to those who are not accustomed 
to it. I have found it useful to invent a system of symbols to use instead 
of words. ^ Descriptive terminologies in this sense, i.e. those using 
specific terms and compounding them, are to be found in some African 
peoples, and illustrations are given in the section on the Yako in this 

In many systems of kinship terminology a single term is used for two 
or more kinds of relatives, who are thus included in a single terminological 
category. This may be illustrated by the English system of the present 
day. The word 'uncle' is used for both the father's brother and the 
mother's brother and also by extension for the husband of an 'aunt' 
(father's sister or mother's sister). Similarly with such terms as 'nephew', 
'niece', 'cousin', grandfather', &c. 

The categories used in the terminology often, indeed usually, have 
some social significance. In English we do something which is unusual 
in kinship systems when we apply the term uncle (from Latin avunculus, 
mother's brother, literally 'little grandfather') to both the mother's 
brother and the father's brother. But this corresponds with the fact that 
in our social life we do not make any distinction between these two kinds 
of relatives. The legal relationship in English law, except for entailed 
estates and titles of nobility, is the same for a nephew and either of his 
uncles; for example, the nephew has the same claim to inheritance in 
case of intestacy over the estate of either. In what may be called the 
socially standardized behaviour of England it is not possible to observe 
any regular distinction made between the paternal and the maternal 
uncle. Reciprocally the behaviour of a man to his different kinds of 
nephews is in general the same. By extension, no significant difference 

' 'A System of Notation for Relationships', Man, xxx, 1930, p. 93. 

^ See Herskovits, Dahomey, 1938, i, pp. 145 et seq. ; Northcote W. Thomas, 
Anthropological Report on the Edo-speaking Peoples of Nigeria, 1910, Part I, 
pp. 112 et seq. and Anthropological Report on the Ibo-speaking Peoples of Nigeria, 
1913, Part I, p. 72. 



is made between the son of one's mother's brother and the son of one's 
father's brother. In Montenegro, on the contrary, to take another 
European language, there is a different system. The father's brother is 
called stric and his wife is strina, while the mother's brother is ujak and 
his wife is ujna, and the social relations in which a man stands to his two 
kinds of uncles show marked differences. 

In the eighteenth century Lafitau' reported the existence amongst 
American Indians of a system of terminology very unlike our own. 

'Among the Iroquois and Hurons all the children of a cabin regard all 
their mother's sisters as their mothers, and all their mother's brothers as 
their uncles, and for the same reason they give the name of fathers to all 
their father's brothers, and aunts to all their father's sisters. AH the children 
on the side of the mother and her sisters, and of the father and his brothers, 
regard each other mutually as brothers and sisters, but as regards the 
children of their uncles and aunts, that is, of their mother's brothers and 
father's sisters, they only treat them on the footing of cousins.' 

In the nineteenth century Lewis Morgan, while living with the 
Iroquois, was impressed by this method of referring to kin and set to 
work to collect kinship terminologies from all over the world. These he 
published in 1871 in his Systems of Consanguinity and Affinity. He found 
systems of terminology similar to that of the Iroquois in many parts of 
the world, and such systems he called 'classificatory'. 

The distinguishing feature of a classificatory system of kinship 
terminology in Morgan's usage is that terms which apply to lineal 
relatives are also applied to certain collateral relatives. Thus a father's 
brother is 'father' and mother's sister is 'mother', while, as in the type 
described by Lafitau, there are separate terms for mother's brother and 
father's sister. Consequently in the next generation the children of 
father's brothers and mother's sisters are called 'brother' and 'sister' 
and there are separate terms for the children of mother's brothers and 
father's sisters. A distinction is thus made between two kinds of cousins, 
'parallel cousins' (children of father's brothers and mother's sisters), 
who although 'collateral' in our sense are classified as 'brothers' and 
'sisters', and 'cross-cousins' (children of mother's brothers and father's 
sisters). There is a similar distinction amongst nephews and nieces. 
A man classifies the children of his brothers with his own children, but 
uses a separate term for the children of his sisters. Inversely a woman 
classifies with her own children the children of her sisters but not those 
of her brothers. Classificatory terminologies of this kind are found in 
a great many African peoples. 

There are other types of classificatory system, found less frequently, 
in which the term 'father' is applied to the brothers of the mother as 
well as to those of the father, and both mother's sister and father's sister 

' Lafitau, Mreurs des Sauvages Ameriquains, Paris, 1724, vol. i, p. 552. 


are called 'mother' ; or cousins, both parallel and cross-cousins, may all 
be treated as 'brothers' and 'sisters'. 

In classificatory systems the principle of classification may be applied 
over a wide range of relationship. Thus a first cousin of the father, being 
his father's brother's son, whom he therefore calls 'brother', is classified 
with the father and the same term 'father' is applied to him. His son in 
turn, a second cousin, is called 'brother'. By this process of extension 
of the principle of classification nearer and more distant collateral 
relatives are arranged into a few categories and a person has many 
relatives to whom he applies the term 'father' or 'mother' or 'brother' 
or 'sister'. 

The most important feature of these classificatory terminologies was 
pointed out long ago by Sir Henry Maine. 'The effect of the system', he 
wrote, 'is in general to bring within your mental grasp a much greater 
niimber of your kindred than is possible under the system to which we 
are accustomed. '^ In other words, the classificatory terminology is' 
primarily a mechanism which facilitates the establishment of wide-range 
systems of kinship. 

There is more to it than this, however. Research in many parts of the 
world has shown that the classificatory terminology, like our own and 
other non-classificatory systems, is used as a method of dividing relatives 
into categories which determine or influence social relations as exhibited 
in conduct. The general rule is that the inclusion of two relatives in the 
same terminological category implies that there is some significant 
similarity in the customary behaviour due to both of them, or in the 
social relation in which one stands to each of them, while inversely the 
placing of two relatives in different categories implies some significant 
difference in customary behaviour or social relations. Some anthropolo- 
gists make a great point of real or supposed exceptions to this rule, but 
they seem to forget that there can only be an exception when there is 
a general rule to which it is an exception. 

There is a complication resulting from the fact that in classificatory 
systems there are necessarily distinctions between near and distant 

' relatives included in the same category. Thus amongst the men referred 
to as 'father' the nearest relative is, of course, the actual 'own' father. 

)r After him come his brothers and after them his parallel first cousins 
and perhaps in some systems the husbands of the mother's sisters. So 
on to more and more distant relatives of the same terminological cate- 
gory. The attitude and behaviour of a person towards a particular rela- 
tive is affected not only by the category to which he belongs but also by 
the degree of nearness or distance of the relationship. In classificatory 

y systems there are many women whom a particular man calls 'sister'. In 
some systems he will be prohibited from marrying any of these women. 

> In some others he may not marry any 'near' 'sister', i.e. any one of these 
' The Early History of histitutions, 1874, p. 214. 


women who is related to him within a certain degree of cognatic relation- 
ship, but may marry a more distant 'sister'. , 
— Morgan tried to classify all terminological systems into two classes 
as being either classificatory or descriptive. But the ordinary English 
system should not be called descriptive and there are many other non- 
classificatory systems that are also not descriptive.^ A people using a ' 
classificatory terminology may also make use of the descriptive principle 
or method in order to refer to the exact genealogical relation between 
two persons.^ The study of kinship terminologies is valuable because 
they frequently, or indeed usually, reveal the method of ordering 

The reality of a kinship system as a part of a social structure consists 
of the actual social relations of person to person as exhibited in their 
interactions and their behaviour in respect of one another. But the actual 
behaviour of two persons in a certain relationship (father and son, 
husband and wife, or mother's brother and sister's son) varies from one 
particular instance to another. What we have to seek in the study of a 
kinship system are the norms. From members of the society we can ( 
obtain statements as t o how two persons in a certai rLxelatioaship ought 
to-trchavrtowards one^another. A sufficient number of such statements 
will enable us to define the Tdeal or expected conduct. Actual observa- 
tions of the way persons do behave will enable us to discover the extent - 
to which they conform to the rules and the kinds and amount of devia- 
tion. Further, we can and should observe the reactions of other persons 
to the conduct of a particular person or their expressions of approval or 
disapproval. The reaction or judgement may be that of a person who is 
directly or personally afi'ected by the conduct in question or it may be 
the reaction or judgement of what may be called public opinion or public 
sentiment. The members of a community are all concerned with the .1 
observance of social usage or rules of conduct and judge with approval 
or disapproval the behaviour of a fellow member even when it does not 
''affect them personally. 

A kinship system thus presents to us a complex set of norms, of usages, — , 
of patterns of behaviour between kindred. Deviations from the norm 
have their importance. For one thing they provide a rough measure of 
the relative condition of equilibrium or disequilibrium in the system. 
Where there is a marked divergence between ideal or expected behaviour 
and the actual conduct of many individuals this is an indication of 
disequilibrium; for example, when the rule is that a son should obey his 

' For an example of a terminology that is neither descriptive nor classificatory , 
see N. W. Thomas, Anthropological Report on Sierra Leone, 1916, Part I, 'Law 
and Custom of the Timne'. 

^ For a kinship terminology making use of both the descriptive and the ' 
classificatory principles see 'Double Descent among the Yako', by Daryll Forde, 
in this volume. 



father but there are notably frequent instances of disobedience. But 
there may also be a lack of equilibrium when there is marked disagree- 
ment amongst members of the society in formulating the rules of conduct 
or in judgements passed on the behaviour of particular persons. 

In attempting to define the norms of behaviour for a particular kind of 
relation in a given system it is necessary to distinguish different elements 
or aspects. As one element in a relation we may recognize the existence 
of a personal sentiment, what may be called the affective element. Thus 
we may say that in most human societies a strong mutual affection is 
a normal feature of the relation of mother and child, or there may be in 
a particular society a typical or normal emotional attitude of a son to his 
father. It is very important to remember that this affective element in 
the relation betvveen relatives by kinship or marriage is different in 
different societies. 

We may distinguish also an element that it is convenient to refer to 
/by the term 'etiquette', if we may be permitted to give a wide extension 
of meaning to that word. It refers to conventional rules as to outw^ard 
behaviour. What these rules do is to define certain symbolic actions or 
avoidances which express some important aspect of the relation between 
two persons. Differences of rank are given recognition in this way. In 
some tribes of South Africa it would be an extreme, and in fact unheard 
of, breach of the rules of propriety for a woman to utter the name of her 
husband's father. 

An important element in the relations of kin is wh^t will here be called 
the jural element, meaning by that relationships that can be defined in 
^terms of rights and duties. Where there is a duty there is a rule that 
a person should behave in a certain way. A duty may be positive, 
-prescribing actions to be performed, or negative, imposing the avoidance 
of certain acts. We may speak of the 'performance' of a positive duty and 
.the 'observance' of a negative duty. The duties oi Ato B are frequently 
spoken of in terms of the 'rights' of B. Reference to duties or rights are 
simply different ways of referring to a social relation and the rules of 
behaviour connected therewith. 

- In speaking of the jural element in social relations we are referring to 
customary rights and duties. Some of these in some societies are subject 
to legal sanctions, that is, an infraction can be dealt with by a court of 
law. But for the most part the sanctions for these customary rules are 
what may be called moral sanctions sometimes supplemented by 
"Veligious sanctions. 

^ There are, first, what we may call personal rights and duties, rights 
in personam in legal terminology. A has a right in personam in relation 
^o B when A can claim from B the performance of a certain duty. The 
right and the duty are both determinate. Thus, in the relation between 
an African husband and his wife each of the partners has personal rights 
imposing duties upon the other. Those personal rights and duties that 


form a most important part of relations by kinship and marriage are 
, different from those estabhshed by contract or in a contractual relation- 
ship. In such a relationship a person accepts a certain definite obligation 
or certain obligations towards another. When the specific obligations on 
both sides have been fulfilled the contractual relation is terminated. But 
relations of kinship are not of this kind. They are not entered into 
voluntarily and they normally continue throughout life. It is true that 
a marital relationship is entered into ; but it is not a contractual relation- 
ship between husband and wife; it is best described as a union.' It may 
be terminated by death, but in some societies not even then (witness the 
custom of the levirate by which a woman continues to bear children for 
her husband after he is dead); where divorce is recognized it may be 
terminated by that means. But the rights and duties of husband and wife 
are not like the obligations defined in a contract; they are incident to the 
relationship in the same sort of way as the rights and duties of parents 
and children. 

We must distinguish from personal rights (jus in personam) what are 
designated rights in rem. Such a right is not a claim in relation to a cer- 
tain particular person but a right 'against all the world'. The most 
characteristic form of such rights is in relation to things. The right 
which I have over something which I possess is infringed if someone 
steals it or destroys it or damages it. The use of the legal term jus in 
rem implies that in certain circumstances a person may be treated as a 
thing (res). 

'Thus when a father or master brings an action for the detention of, or 
for injuries inflicted upon, his child or apprentice, or when a husband 
sues for injuries inflicted upon his wife, the child, apprentice, and wife are 
in fact held to be things. The action is not brought in pursuance of the 
legal rights of the child, apprentice, or wife.'^ 

I shall refer to these rights (in rem) over persons as 'possessive rights', 
but it must be remembered that the term is used in this special sense. 
In the formation of systems of kinship and marriage these possessive 
rights over persons are of great importance. Thus in most African 
systems a husband has possessive rights in relation to his wife. His 
rights are infringed, i.e. he suffers a wrong, if a man commits adultery 
with her or if someone kills her or abducts her. Later in this essay we 
shall have to consider the subject of possessive rights over children. It 
should be noted that possessive rights over persons can be shared by a 

' An agreement between two families, whereby one promises to give a 

daughter in marriage and the other undertakes to see that the marriage payments 

' are made, is a contract in the proper sense of the term. This is a preliminary 

to the marriage, just as the Roman sponsalia or betrothal was a preliminary 

promise or contract which was fulfilled in the nuptiae. 

^ Sheldon Amos, The Science of Lazv, 1888, p. 87. See also Sir Frederick 
Pollock, First Book of Jurisprudence, chap. iv. 


number of persons or may be held collectively by a definite group of 
persons, just as may possessive rights over land or other property. 

A kinship system is therefore a network of social relations which con- 
stitutes part of that total network of social relations which is the social 
structure. The rights and duties of relatives to one another are part of the 
system and so are the terms used in addressing or referring to relatives. ( ' ^ ' 

By using the word 'system' we make an assumption, for the word 
implies that whatever it is applied to is a complex unity, an organized 
whole. This hypothesis has already received a considerable measure of 
verification by anthropological studies. But we must distinguish between 
a stable system which has persisted with relatively little change for some 
period of time and the unstable condition of a society which is under- 
going rapid change. It is in the former, not in the latter, that we may 
expect to find some fair degree of consistency and congruence amongst 
the items that make up the whole. 


In this and following sections the more important structural principles 
^ which are found in kinship systems will be briefly indicated. 
"^ Two persons who are kin are related in one or other of two ways: 
either one is descended from the other, or they are both descended from 
a common ancestor. It is to be remembered that 'descent' here refers to~i 
the social relationship of parents and children, not to the physical rela- 
tion. Kinship is thus based on descent, and what first determines the 
character of a kinship system is* the way in which descent is recognized 
and reckoned. 

One principle that may be adopted is the simple cognatic principle. 
To define the kin of a given person his descent is traced back a certain 
number of generations, to his four grandparents, his eight great-grand- 
parents, or still farther, and all descendants of his r£Cognized ancestors, 
through both females and males, are his cognatgsJAt each generation 
that we go backwards the number of ancestors is double that of the 
preceding generation, so that in the eighth generation a person will have 
sixty-four pairs of ancestors (the great-grandparents of his grcat-great- 
grandparents). It is therefore obvious that there must be some limit to 
tracing kinship in this way. The limit may simply be a practical one 
depending on the inability to trace the genealogical connexions, or there 
may be a theoretically fixed limit beyond which the genealogical con- 
nexion does not count for social purposes. 

Another way of ordering the kindred may be illustrated by the system 
of ancient Rome. Within the body of a person's recognized cognates 
certain are distinguished as agnates. Cognates are agnates if they are 
descendants by male links from the same male ancestor. ^ In the Roman 

— *^^Sunt autem agnatl per vinlis sexus personas cognatione juncti' (Gaius) ; 
'Agnati sunt a patre cognati' (Ulpian). 



system there was the strongest possible emphasis on agnatic kinship 
i.e. on uniUneal descent through males. 

In some other societies there is a similar emphasis on unilineal descent- 
through females. With such a system a person distinguishes from the 
rest of his cognates those persons who are descended by female links 
only from the same female ancestress as himself. We can speak of these 
as his matrilineal kin. 

There are few, if any, societies in which there is not some recognition 
of unilineal descent, either patrilineal (agnatic) or matrilineal or both. 
Thus in the modern English system surnames descend in the male line. 
In many countries, in a mixed marriage, children acquire by birth the 
nationality of the father, not that ""of the mother. But what matters in the 
study of any society is the degree of emphasis on the unilineal principle 
and how it is used. 

One important way in which the unilineal principle may be used is 
in the formation of recognized lineage groups as part of the social struc-. 
V ffcture. An agnatic lineage consists of an original male ancestor and all his 
descendants through males of three, four, five, or n generations. The 
lineage group consists of all the members of a lineage alive at a given 
time. A woman belongs to the lineage of her father, but her children do 
not. With matrilineal reckoning the lineage consists of a progenetrix and 
all her descendants through females. A man belongs to his mother's 
lineage, but hiS children do not. Lineage groups, agnatic or matrilineal, 
are of great importance in the social organization of many African 

A lineage of several generations in depth, i.e. back to the founding 
ancestor or ancestress, will normally be divided into branches. In an 
agnatic lineage of which the founding ancestor has two sons each of them 
may become the founder of a branch consisting of his descendants in the 
male line. The two branches are united by the fact that their founders 
were brothers. As a lineage continues and increases the branching 
process continues, resulting in a large and complex organization. In 
some parts of China we find in one village a body of persons all having 
the same name and tracing their descent in the male line to a single 
ancestor who may have lived eight or nine hundred years ago. This is 
therefore a very large lineage which may number several hundred living 
persons. Genealogical records of the whole lineage are usually preserved 
and are sometimes printed for the use of the families of the lineage. 
There is a complex ramification of branches. An important feature of the 
Chinese system is the maintaining of the distinction of generations. 
A common method of giving names is one by which a man has three 
names; the first is the lineage or family name; the second indicates the 
generation to which he belongs; the third is his distinctive personal 
name. From the second part of the name any member of the lineage can : 
tell to which generation any particular individual belongs. 


What is here called a branch of a lineage is, of course, itself a lineage. 
A lineage of ten generations may include two or more branches of nine 
generations and one of these may contain two or more of eight, and so 
on. A lineage of several generations includes dead as well as living. We 
may conveniently use the term 'lineage group' to refer to a group formed 
of the members of a lineage who ajce alive at a particular time. Lineage 
groups as thus defined are important as components of the social struc- 
ture in many African societies. A lineage group that is socially important 
may itself consist of smaller groups (branch lineages) and it may itself 
be part of a more extended and recognized group formed of related 

It is desirable to illustrate by examples the differences in the ordering 
of kindred as the result of relative emphasis on the cognatic principle or 
on the unilineal principle. As an example of a cognatic system we may.^». 
take the kinship system of the Teutonic peoples as it was at the beginning 
of history. This was based on a widely extended recognition of kinship 
traced through females as well as males. The Anglo-Saxon word for 
kinsfolk was maeg (magas). A man owed loyalty to his 'kith and kin'. 
Kith were one's friends by vicinage, one's neighbours; kin were persons 
descended from a common ancestor. So, for 'kith and kin' Anglo-Saxon 
could say 'his magas and his frynd', which is translated in Latin as 
cognati atque amid. 

The arrangement of kin by degrees of nearness or distance was based ' 
on sib-ship (English sib, German Sippe). A man's sib were all his cognates 
within a certain degree. One method of arranging the sib was by refer- 
ence to the human body and its 'joints' (glied). The father and mother 
stand in the head, full brothers and sisters in the neck, first cousins at the 
shoulders, second cousins at the elbow, third cousins at the wrists, 
fourth, fifth, and sixth cousins at the joints of the fingers. Finally come 
the nails, at which would stand the seventh cousins. On one scheme 
these nail kinsmen (nagel magas) were not included in the sib, though 
they were recognized as kinsmen (magas) if known to be such. The sib 
therefore included all kinsfolk up to and including sixth cousins. They 
were the sibgemagas, the sib kinsfolk.' 

It is evident that no tsvo persons can have the same sib, though for 
two unmarried full brothers, A and B, every person who was sib to 
A was sib to B, and A and B were sib-kinsmen of one another. A person 
cannot be said to 'belong' to a sib or be a member of a sib in the sense — 

' This account of the 'joints' of the sib is the one given in article 3 of the 
first book of the Sachsenspiegel. I have used the Leipzig edition of 1545. The 
statement about the 'nail kinsmen' is as follows: 'in dem siebenden steht ein 
nagel und nicht ein glied darumb endet sich da die sip und heisst ein nagel 
freund. Die sip endet sich in dem siebenden glied erbe zu nemen'. This is 
explained in a Latin nota: 'gradus cognationis finitur in septimo gradu, necesse 
est ergo in petitione haereditatis, q. haeres et petitor articulet eum, vel se 
defuncto infra septem gradus attigisse'. 


in which he can be said to belong to a lineage or a clan or a village 

The innermost circle of the sib of an individual included his father 
and mother, his brother and sister, and his son and daughter — the 'six j 
hands of the sib'. Another circle that was recognized was that of rela- * 
tives'' within the knee'. This word (cneozv) seems to have referred to the 
elbow, so that kinsfolk 'within the knee' would include all the descen- 
dants of the eight great-grandparents. 
"The sib was thus an arrangement of kindred as it were in a series of 

V concentric circles, with the person whose sib it was at the centre. One 
circle included all those kin with whom marriage was forbidden. It is 
difficult to discover exactly where this was. In some Teutonic systems 
and in ancient Wales it is said that the prohibition against marriage" 
extended to 'the fifth degree'. This would seem to include all third 
cousins, but the matter is not quite clear. 

. . Another way of reckoning degrees of kinship was by 'stocks'. There , 
were first the four 'quarters'of the sib (kliifte, in Frisian), the four stocks 
of the grandparents, each consisting of the descendants of one of the 
four pairs of great-grandparents. The wider sib included the eight 
stocks (fechten or fange) of the great-grandparents, each consisting of the 
descendants of one of the eight pairs of great-great-grandparents. The 
eight stocks (Old High German ahta, Old Norse oett) therefore included 
all kinsfolk as far as third cousins. 

\ In the Middle Ages another method of reckoning was adopted, the 

parentelic system (Latin parentela). A person's sib was arranged in five 
parenteles: (i) his own descendants; (2) all descendants of his parents 
(excluding i); (3) all descendants of his two pairs, of grandparents ' 
(excluding i and 2) ; (4) all descendants of his four pairs of great-grand- 
parents (excluding i, 2, and 3); (5) all descendants of his eight pairs of 
great-great-grandparents (excluding i, 2, 3, and 4). This system seems " 
to have been used principally for regulating inheritance. Within a 
parentele the degree of relationship was fixed by the greater or smaller 
distance from the common ancestor. Thus the uncle is more nearly 
related than his son, the cousin. But nephews (in the second parentele) 
are more nearly related than uncles (in the third). Both the 'stock' and 
the parentelic method of reckoning seem to go only as far as the wristj 
'joint', i.e. as far as third cousins on both sides. They include, therefore, 
only part of the total sib. 

"^ In more recent times the present method of reckoning by cousinage 
was introduced. It has been suggested that the system of reckoning by 
first, second, third cousins originated in Spain and Portugal as a result 
of Teutonic invasions. In England this reckoning by cousins has j 
replaced the older system of the sib. ! 

I 'Sib' may be defined as meaning computable cognatic relationship for 
definite social purposes. W'e have seen that it was used for fixing the 


degrees within which marriage was forbidden. After the introduction 
of Christianity the relation between godparents and godchildren was 
included under sib. The godfather and godmother were 'god-sib' 
(modern 'gossip') to their godchildren and marriage between them was 
forbidden. Sib-ship also regulated the inheritance of property. Persons 
who were not related to a deceased person within a certain degree had 
no claim to inherit. 

Where the functioning of the Teutonic sib can be best studied is in the 
customs relating to wergild, which was the indemnity that was required 
when one person killed another., It was paid by the person who had killed 
and his sib, and was received by the sib of the deceased. When the 
system was in fall force the number of kinsmen who might be called on 
to contribute to the payment or who might receive a share of it was con- 
siderable. Theoretically it seems to have extended in some communities 
.as far as sixth cousins on both sides, i.e. to all the 'joints' but not to the 
nail. But practically, and in some instances in theory, duties and claims 
seem to have been effective only as far as fourth cousins. There were 
laws or rules fixing the total amount of the wergild and the amounts or 
shares to be contributed or shared by each class of kinsfolk. The nearest 
kin paid and received most, the most distant paid and received least.' 

The payment of wergild was an indemnity for homicide paid to those 
persons who had possessive rights^ (rights iti rem) over the person who 
was killed. In the Teutonic system these rights were held by the 
cognatic relatives of the slain man by what was essentially a system 
similar to partnership. Each relative held as it were a share in the 
possession, and the consequent claim for indemnity and the share of any 
relative depended on the nearness of the relation so that, for example, 
the share claimed by the second cousins was twice that belonging to 
third cousins. 

We have now to ask what use was made of the unilineal principle in 
the Teutonic systems. A man's kin were divided into those of the spear 
side (his paternal kin) and those of the spindle side (his maternal kin). 
In some of the Teutonic systems relatives on the father's side paid or 
received twice as much as those on the mother's side in a wergild trans- 
action. This was so in the England of King Alfred. Similarly in ancient 
Welsh (British) law the gala?ias (the Celtic term for wergild) was paid 
two-thirds by paternal kinsmen and one-third by maternal, up to the 
fifth cousin, but not including the son of the fifth cousin or the sixth 
cousin. So far as the Teutonic peoples are concerned this may have 
been a late development. But in any case this did not mean the recogni- 
tion of unilineal descent, but only that a father's sister's son was a 
nearer relative than a mother's sister's son. 

Amongst some at least of the Teutonic peoples there existed large 

' The most readily accessible account in English of wergild payments is 
Bertha S. Phillpotts, Khidred atid Clan, Cambridge, 1913. 


house-communities under the control of a house-father or house-lord, 
of the type of what it is usual to call the patriarchal family. Sons con- < 
itinued to live with their father under his rule and daughters usually 
■joined their husbands elsewhere. But although the patrilineal principle ) 
was general or usual, it was not always strictly adhered to. Thus in the 
Icelandic Nyal's Saga the house-community of old Nyal included not 
only his wife Bergthora and his three married sons but also a daughter's 
husband, and, with children, men-servants, and others, the household 
numbered some fifty persons. 

The patrilineal principle also appears in the preference given to sons 
over daughters in the inheritance of land. But in default of sons, 
daughters might inherit land in some Teutonic societies. It appears that 
the principle of unilineal descent was only used to a limited extent in th^ 
Teutonic system. 

About the Teutonic system two pseudo-historical theories have been 
advanced. One is that these peoples in prehistoric times had a system of 
matriarchy, which, whatever else it may or may not mean, implies an 
emphasis on unilineal descent through females. As evidence is quoted 
the statement of Tacitus that amongst the Germans the mother's brother 
was an important relative.' 

The other theory is that in prehistoric times the Teutonic peoples had , 
a system of patriarchy emphasizing agnatic descent. This is a deduction 
from the general theory that originally all the Indo-European peoples 
had a patriarchal system. For neither theory is there any historical 

" An example of an arrangement of relations of kinship on the basis of \ 
unilineal descent may be taken from the Masai of East Africa. Though , 
the kinship system of the Masai has not been adequately studied, the 
arrangement of the various kin can be seen in the terminology which has 
been recorded by Hollis.^ 

The terminology is classificatory. The father, the father's brother, and 
the father's father's brother's son are called by the same term, menye. 
The sons and daughters of these men are called 'brother' and 'sister' 
{ol-alashe and eng-anashe). To all the children of men he calls 'brother' 
a man applies the same terms as to his own son and daughter {ol-ayoni 
and en-dito). All these persons belong to a man's own agnatic lineage 
(descendants of the father's father's father). We do not know how far the 
recognition of the lineage connexion is extended, but for our present 
purpose this does not matter. ^ 

' Tacitus, Germania, c. 20. 4: 'Sororum filius idem apud avunculum qui ad 
patrem honor.' 

^ A. C. Hollis, 'A Note on the Masai System of Relationship and other Matters 
connected therewith', J. Roy. Anthrop. Itist. xl, 1910, pp. 473-82. There are 
two misprints in the table facing p. 482. The terms e-sindani e-anyit (wife's 
sister) and ol-le-' sotwa (her husband) have been transposed. 

^ The Masai are divided into five clans and each clan is subdivided into 


Own lineage 


F.F. A = O 
6l-akwi I okoi 

O = A 
okoi ol-le-sotwa 

F. A = O 
menye I ngoto 

O = A 
ngoto I ol-aputani 

Br. A = O A O = A A O 

ol-alashe I e-sindani EGO eng-anashe I ol-aputani ol-apu en-e-'ng-apu 

Son A = O 
ol-ayoni I en-gerai 

A = O 
ol-akwi en-gerai 

O = A A O 

en-dito I ol-aputani ol-apu en-e-'ng-apu 

O = A A O 

eng-akwi ol-aputani ol-akwi eng-akwi 



Mother^s lineage 

A = 


1 ^ 



1 - 

A = 


O = A 
ngoto 1 menye 

A = 



O = 



A = 










A feature of the Masai system is that these terms for 'father', 'brother', 
'sister', 'son', 'daughter', though used as classificatory terms, are not 
appUed except to a man's agnatic kindred. Thus while in some societies 
with a classificatory terminology^ the mother's sister's husband is called 
'father' and her children are 'brother' and 'sister', this is not so in Masai. 
Thus the Masai terminology emphasizes the distinction between 
agnatic and other kindred. 

It is of some significance that the distinction between agnatic and 
other kin is not carried back into the generations of the grandparents and 
great-grandparents. Male relatives of the second and third ascending 
generations on both the father's and the mother's side are ol-akwi 
and their wives and sisters are okoi. Similarly, and reciprocally, the 
terms ol-akwi (grandson) and eng-akwi (granddaughter) are applied to 
daughters' children as well as sons' children and to the grandchildren, 
through sons or daughters, of any 'brother' of the lineage. 

The emphasis on the agnatic lineage also appears in the fact that there 
are certain terms which apply only to women who have come into the 
lineage by marriage : e-sindani for the wife of any 'brother', and en-gerai 
for the wife of any 'son' or 'grandson' of the lineage. 

Let us consider the women of a man's own lineage of his own and 
the preceding generation. The sisters of the men he calls 'father' are 
referred to descriptively (eng-anashe-menye) or the term for 'mother' 
(ngoto) is applied to them. The husband of a 'father's sister' or a 'sister' 
is called ol-apiitani, which simply means 'relative by marriage', but their 
children are called ol-apu and en-e-'ng-apu. A man's ol-apu is his 
'sister's son' or 'father's sister's son' and is connected with the man's 
own lineage through the mother. The children of his ol-apu are again his 
kin, ol-le-'ng-apu and en-e-'ng-apu, and the son of ol-le-ng-apu is again 
ol-le-ng-apu. The cognatic relationship is continued from father to son 
in the male line. But the offspring of an en-e-ng-apu (sister's daughter, 
father's sister's daughter, sister's son's daughter, &c.) are ol-le-'sotwa and 
en-e-sotwa. It is evident from the account given by Hollis that these two 
terms do not refer to any definite relation of kinship but simply mean 

sub-clans. The sub-clan is the exogamous group, i.e. marriage is permitted within 
the clan but not within the sub-clan. A similar organization is found in tribes 
related to the Masai. The Kipsigis, for example, are divided into clans (oret) 
and the clan is subdivided into segments which Peristiany refers to by the native 
term kot op chi and calls a subdivision of the clan (Peristiany, The Social Institu- 
tions of the Kipsigis, 1939). For the Nandi, Hollis tells us that 'second cousins, 
like cousins, are called brothers ; more distant cousins are called piek-ap-oret 
(people of the family)'. This presumably refers to agnatic cousins only, and it 
would mean that the lineage proper (within which cousins are 'brothers') con- 
sists of the descendants of father's father's father. The Masai system may have 
been similar. 

' In the Kipsigis, for example, who belong to the same general area as 
the Masai and have a somewhat similar social organization. See Peristiany, 
op. cit. 


'relative'. The husband of an en-e-'ng-apu is not ol-aputani (relative by 
marriage) but simply ol-le-'sotzva. 

A man is connected through his mother with her agnatic lineage. He 
calls his mother's brother ol-apu. This is therefore a self-reciprocal term 
used between mother's brother and sister's son. The relationship, 
according to Hollis, is an important one. The mother's brother's son and 
daughter, the cross-cousins, are ol-le- ng-apii and en-e-ng-apii. Once 
again the relationship is continued within the lineage, i.e. through males 
but not through females. The s on and daughter of the mo ther's brother's 
son are ol-le- ng-apu andfen-e-'ng-apu, but the children of the mother's 
brother's daughter are simply sotzva relatives. Specific kinship is also 
not continued through the mother's sisters, whose children are sotzva 

A man also recognizes the mother's brother of his father as a kinsman 
and calls him ol-apu, as he does his own mother's brother. This man 
would call him ol-le-' ng-apu since he is his sister's son's son. Hollis does 
not inform us what a man calls his father's mother's brother's son, but 
it seems likely that he would be ol-le-' ng-apu. 

The emphasis on agnatic descent is also observable in the relation of 
man to his wife's lineage. All the male members of the lineage, including 
the wife's father's father and the wife's father's father's brother and all 
their male descendants in the male line, are ol-aputani. On the other 
hand, while the wife's father's sister is eng-aputani, the relationship does 
not continue and her children are only sotzva. In the wife's mother's 
lineage she is eng-aputani and her brothers and sisters are ol-aputani and 
eng-aputani, but the relationship continues no farther and the wife's 
mother's brother's children become sotzva. 

The Masai kinship terminology thus presents an interesting and 
illuminating example of a system in which the ernphasis is placed on 
agnatic lineage. The most important kinship relations of a man are 
evidently those with the members of his own agnatic lineage. But the 
children of his sister and father's sister (female members of his lineage) 
are persons with whom his kinship is important. The kinship connexion 
is maintained in the male line but not in the female line. The other side 
of this relationship is that of a man with his mother's lineage. Here again 
he is related to the members of the lineage, his mother's sister and his 
mother's brother, but the relationship only continues in the male line 
to the children and the son's children of his mother's brother, and not 
to the children of his mother's sister or to the children of his mother's 
brother's daughter. B y marriage he is related to his wife's agnatic lineage, 
to female members as well as male, and again the relationship is only 
continued in the male line. His relationship to his wife's mother is con- 
tinued to her brothers and sisters but not to the rest of her lineage. 

The emphasis on the agnatic line is shown clearly in the fact that the 
husband of the father's sister is ol-aputani, a relative by marriage, and 


her children are kinsfolk, while the husband of a mother's sister, far from 
being a 'father', as he is in some systems, is merely a relative (sotzva), as 
also are her children. The general principle is that relationships, includ- 
ing those established through females or by marriage, continue in the 
male line only. ^ 

An interesting and distinctive feature of the Masai system is the use 
of the terms ol-le-sotwa and en-e-sotwa. Hollis tells us \.\\?l\. sotwa means 
'peace' or 'relative'. It does not connote any definite kind or degree of 
relationship either by kinship or by marriage. Thus sotwa relatives may 
marry. Hollis states that the restrictions on marriage are two: (i) a man 
may not marry a woman of his own sub-clan, which is a patrilineal group ; 
(2) 'No man may marry a nearer relation than a third cousin, and then 
only if the terms of address used are ol-le-sotwa and en-e-sotwa.^ This 
statement as it stands implies that there are certain sotwa relatives who 
may not be married, for example, the mother's sister's daughter and the 
mother's mother's sister's daughter's daughter. On such a point as this 
it would be better if we had more precise information. The fact is clear, 
however, that the sotwa relationship is not itself a bar to marriage while 
other kinship relations are. 

The Masai system arranges a man's kin into a few large categories, 
(i) His agnatic relatives belonging to his own lineage (or sub-clan); 

(2) his apu kin, if we may call them so, the descendants of those women 
of his lineage who are his father's sisters or his sisters, and on the other 
hand, the members of the mother's lineage: the relationship is estab- 
lished through a female and is then continued in the male line only; 

(3) his aputani relatives, i.e. relatives by marriage, either the kin of his 
wife or persons who have married one of his kinswomen; (4) those 
relatives to whom he applies the classificatory terms 'grandfather', 
'grandmother', 'grandson', 'granddaughter', some of whom belong to 
his lineage while others do not; (5) his sotwa relatives, a sort of fringe of 
persons not belonging to the first four classes, with each of whom some 
indirect connexion can be traced: they are vaguely his 'relatives', 
persons with whom he should be at peace. 

The structural difference between organization by sib and by lineage 
can be readily seen by comparing a 'stock' with a 'lineage'. A 'stock'/ 
includes all the descendants of a man and his wife counting descent' 
through females as well as males. In the sib system a man belongs to each 
of the eight stocks of his eight pairs of great-great-grandparents and all 
descendants of these ancestors are his kin. With a patrilineal lineage 
system he belongs to a lineage which includes the descendants of a male 
ancestor counting through males only. The persons of his mother's 
lineage are his kin, but he does not belong to the lineage. 

We are here concerned with the ways in which different societies 
provide an ordering of the kin of an individual within a certain range, 
wide or narrow. One way is by tracing kinship equally and similarly 


through males and through females. There is a close approximation to 
this in modern European societies and in some primitive societies. In 
various societies we find some greater or less emphasis placed on uni- 
lineal descent, but there are many different ways in which this principle 
can be applied. In the Masai system there is marked emphasis on the 
male line, so that the most important of a man's social connexions are 
with the members of his own agnatic lineage and with those of his mother's 
agnatic lineage and those of the agnatic lineage of his wife. In other 
societies there may be a similar emphasis on the female line, so that 
a man's chief relations through his father are with the latter's matrilineal 
lineage. There are also systems in Africa and elsewhere in which uni- 
lineal descent through males is given recognition and also unilineal 
descent through females. 

There are a great variety of ways in which the unilineal principle may 
be used. It is therefore only misleading to talk about matrilineal and 
patrilineal societies as was formerly the custom of anthropologists. 
Some more complex and systematic classification is needed to represent 
the facts as they are. 


Reference has already been made to the classificatory systems of kin- 
ship terminology which are found in very many African peoples. The 
theory of pseudo-history is that this method of referring to kin is a 
'survival' from a time in trie past when the family as it now exists had 
not made its appearance. In those remote days it is imagined that there 
existed a system of 'group-marriage' in which a group of men cohabited 
with a group of women and when a child was born all the men were 
equally its fathers and all the women equally its mothers. This fantastic 
example of pseudo-history was put forA\^ard by Lewis Morgan in his 
Ancient Society (1878). 

The theory here proposed is not in any way concerned with the origins xf 
of classificatory terminologies, which ixt found widely distributed in 
Asia, Africa, America, and Australasia, but with their social functions. 
It starts from the simple and obvious postulate that in order to have a 
system of kinship it is necessary to have some way of distinguishing and 
classifying a person's kin, and that one very obvious and natural means 
of doing this is through the kinship nomenclature. We give the same 
name to a number of things when we think that in some important 
respect they are similar. We use in English the same name — uncle — for 
the mother's brother and the father's brother because we think of them 
as similar, as relatives of the same kind. In a classificatory system a man 
uses the same term for his father and his father's brother because he 
thinks of them as relatives of one general kind. 

The principle on which the classificatory terminology is based may be 
called the principle of the unity of the sibling group. This refers not to 


the internal unity of the group as shown in the relations of its members 
' to one another but to the fact that the^group jnay^onstitute a unity for 
a person outside it and connected with it by a specific relation to one^F 
its members. Thus a son may, in a particular system, be taught to regard 
his father's sibling group as a united body with whom he is related as 
their 'son'. 

The sibling group, i.e. the body of brothers and sisters of common 
parentage, has its own internal structure. In the first place there are the 
very important social distinctions between the sexes, which divide the 
brothers from the sisters. These distinctions are diflferently exhibited in 
usages in different societies, and the relation between brother and sister 
is therefore an important feature of any particular kinship system. 
Secondly there is the order of birth, which is translated into social terms 
in the distinction of senior and junior. The importance of this distinc- 
tion in African tribes is shown by the existence of separate terms for 
'senior brother' and 'junior brother'. In a polygynous family there is the 
further difference between full siblings and half-siblings, and in African 
peoples this is usually, if not always, important. 

Within an elementary family, i.e. amongst full brothers, the dis- 
crimination between older and younger siblings is made on the basis of 
order of birth. But in a polygynous compound family the wives may be 
unequal in rank, the great wife, or the first wife, having a higher rank 
than others. In the Kaffir tribes of the Transkei a son of a wife of 
inferior rank will apply to all the sons of the great wife (his half-brothers) 
the term used for 'elder brother' even when they are younger than 
himself. The two words are therefore better translated as 'senior brother' 
and 'junior brother'. 

In these same tribes (which have a classificatory terminology) the 
sons of an older brother of the father must be called 'senior brother' 
irrespective of actual age, and those of the father's younger brother will 
be called 'junior brother'. Amongst the Yao, who have matrilineal 
descent, not only the sons of the older brother of the father, but also 
those of the older sister of the mother, must be called 'senior brother' 

The main principle of the classificatory terminology is a simple one. 
xf A and B are two brothers and X stands in a certain relation to A, then 
he is regarded as standing in a somewhat similar relation to B. Similarly 
if A and B are two sisters. In any particular system the principle is 
applied over a certain range. The similarity of the relation is indicated 
by applying a single term of relationship to A and B. The father's brother 
is called 'father' and the mother's sister is called 'mother'. The father's 
father's brother is regarded as similar to the father's father and therefore 
his son is also called 'father'. Once the principle is adopted it can be 
applied and extended in different ways. However the principle is used, 
it makes possible the recognition of a large number of relatives and their 


classification into a relatively few categories. Within a single category 
relatives are distinguished as nearer or more distant. 

As a general rule, to which, of course, there may be exceptions, 
towards all persons to whom a given term of relationship is applied there is 

■ some element of attitude or behaviour by which the relationship is given 
recognition, even if it is only some feature of etiquette or an obligation 
to exhibit friendliness or respect. Rules of behaviour are more definite 
and more important for near relatives than for more distant ones. 

v/ Itis_a general characteristic of classificatory terminologies that the 
father's brother is called 'father' and the mother's sister 'mother'. When 
we come to the mother's brother and the father's sister, for these relatives 
there is a possible choice between two difterent structural principles. 
One may be called the 'generation' principle. There are a few examples 
in Africa of peoples with classificatory systems who call the father's 
sister and the mother's brother's wife 'mother'. The Masai provide one 
example, and there are others. This means that female cognates of the 
first ascending generation are placed in a single category. It is only 
rarely, and not, so far as I am aware, anywhere in Africa, that the 
mother's brother is called 'father'. Where this principle is used the 
emphasis is on generation and sex. 

/ The other principle is that of the unity of the sibling group. The 
father's sister belongs to the father's sibling group and therefore she is 
a relative of the same general kind as the father and the father's brother, 
and similarly the mother's brother is a relative of the same kind as the 
mother and her sister. This way of thinking about uncles and aunts is 
shown in the kinship terminology of many African societies. Writing 
about the Kongo, Father van Wing says : 'All the sisters and brothers 
of the mother are considered as the mothers [ngiidi) of the Mukongo 
{ngudi nsakela or ngudi nkasi) while all the brothers and sisters of the 
father (se) are considered by the Mukongo as fathers (mase).'^ The term 
for 'mother's brother' in many Bantu tribes is imtahmie, which is literally 
'male mother', uma being the term for 'mother' and linne meaning 'male'. 
Similarly, in a number of African peoples the father's sister is referred 
to by the term for 'father' or is called 'female father'. Examples are the 
Kitara, the Ndau, the Yao, the Huana, and there are others. 

To some Europeans this use of the terms 'female father', 'male mother' 
may seem the height of absurdity. The reason for this is simply a con- 
fusion of thought resulting from the ambiguity of our own words for 
father and mother. There is the purely physical relation between a child 
and a woman who gives birth to it or the man who begets it. The same 
relation exists between a colt and its dam and sire. But the colt does not 
have a father and a mother. For there is the social (and legal) relation 
between parents and children which is something other than the physical 
relation. In this sense an illegitimate child in England is a child without 
' J. Roy. Anthrop. Inst. Ixxi, 1942, p. 96. 




a father. In the African tribes with which we are dealing it is the social 
and legal relationship that is connoted by the words which we have to 
translate 'father' and 'mother'. To call the father's sister 'female father' 
indicates that a woman stands in a social relation to her brother's son 
that is similar in some significant way to that of a father with his son. 
It is more exact, however, to say that a father's sister is regarded as a 
relative of the same kind as a father's brother, with such necessary 
qualifications as result from the diflFerence of sex. 

The principle of social structure with which we are here concerned is 
therefore one by which the solidarity and unity of the family (elementary 
or compound) is utilized to order and define a more extended system of 
relationships. A relationship to a particular person becomes a relation 
to that person's sibling group as a social unit. This shows itself in two 
ways.-Eirat, in some similarity in behaviour, aswhen the kind of behaviour 
that is required towards a father's brother is in some respects similar to 
that towards a father. Sficond, in the provision that in certain circum- 
stances one relative may take the place of another, the two being siblings. 
Thus in some African societies the place of a father, a husband, or a 
grandfather may be taken by his brother. In the custom known as the 
sororate the place of a deceased wife is taken by her sister. In one form 
of the levirate the brother of a deceased man becomes the husband of 
the widow and the father of her children. Amongst the Hehe the 
grandmother plays an important part in the life of a child. This should 
be the child's own grandmother, but if she is dead her sister can take her 
place. ^ 

Miss Earthy^ says that amongst the Lenge the father's sister {hahane, 
female father) 'ranks as a feminine counterpart of the father, and some-, 
times acts as such, in conjunction with or in the absence of the father's 
brothers'. She may offer a sacrifice on behalf of her brother's child, in 
case of illness, in order that the child may recover. Such sacrifice would, 
of course, normally be made by the father or his elder brother. 

The purpose of this section has only been to indicate the existence 
of a structural principle which is of great importance in a very large 
number of kinship systems not only in Africa but also in many other 
parts of the world. Where the principle influences the terminology of 
kinship it may appear in the form of a classificatory terminology. But ' 
the absence of such a terminology does not mean that the principle of 
the unity of the sibling group is not effectively present in the social 
structure and in the organization of norms of behaviour. The classifi- 
catory terminology in its most characteristic form is the utilization of the 

' Elizabeth Fisher Brown, 'Hehe Grandmothers', ^ Roy. Anthrop. Inst. 
Ixv, 1935, pp. 83-96. ^^ 

^ E. Dora Earthy, 'The Role of the Father's Sister among the Valenge of 
Gazaland', South African Journal of Science, xxii, 1925, pp. 526-9. Also in 
Valenge Women, 1933, pp. 14 et seq. 


principle of the unity of the sibHng group to provide a means for ordering 
relatives in a system of wide-range recognition of relationships. 


Within the elementary family there is a division of generations ; the 
parents form one generation, the children another. As a result, all the. 
kin of a given person fall into generations in relation to him, and there 
are certain general principles that can be discovered .in his different 
behaviour towards persons of different generations. 

The normal relation between parents and childre/i <;an be described 
as one of superordination jyad subor-dination. This results from the fact 
that children, at least during the early part of life, are dependent on their 
parents, who provide and care for them and exercise control and 
authority over them. Any relationship of subordination, if it is to work, 
requires that the person in the subordinate position should maintain an 
attitude of respect towards the other. The rule that children should not 
only love but should honour and obey their parents is, if not universal, 
at least very general in human societies. 

" There is therefore a relation of social inequality between proximate 
generations, and this is commonly generalized so that a person is subor- 
dinate and owes respect to his relatives of the first ascending generation 
— that of his parents. To this rule there may be specific exceptions, for 
the mother's brother in some African societies, for example, or for the 
father's sister's husband in some societies in other parts of the world, 
whereby these relatives may be treated disrespectfully or with privileged 
familiarity. Such exceptions call for explanation. 

^' The relation between the two generations is usually generalized to 
extend beyond the range of kinship. Some measure of respect for persons 
of the generation or age of one's parents is required in most if not all 
societies. In some East African societies this relation is part of the 
organization of the society into age-sets. Thus among the Masai sqxual 
intercourse with the wife of a man belonging to one's father's age-set is 
regarded as a very serious offence amounting to something resembling 
incest. Inversely, so is sexual connexion with the daughter of a man of 
one's own age-set. 

The social function of this relation between persons of tvvo proximate 
"generations is easily seen. An essential of an orderly social life is some 
considerable measure of conformity to established usage, and conformity 
can only be maintained if the rulesjiave soipe sort and measure of 
authority behind them. llRe continuity of the social order depends upon 
the passing on of tradition, of knowledge and skill, of manners and 
morals, religion and taste, from one generation to the next. In simple 
societies the largest share in the control and education of the young falls 
to the parents and other relatives of the parents' generation. It is their 


authority that is or ought to be effective. All this is obvious, and it is 
unnecessary to dwell upon it. 

But the further effects of this in the organization of the relations of 
generations are not always so immediately obvious. We shall be con- 
cerned in what follows with the relations between persons and their 
^'relatives of the second ascending generation, that of their grandparents. 
If the exercise of ailthority on the one side and respect and obedience 
on the other were simply, or even primarily, a matter of relative age, we 
should expect to find these features markedly characteristic of the 
relations between grandparents and grandchildren. Actually we find 
most commonly something almost the opposite of this, a relation of 

1 friendly familiarity and almost of social equality. 

} In Africa generally there is a marked condition o£-ceatraint on the 
behaviour of children in the presence of their parents. They must not 
indulge in levity or speak of matters connected with sex. There is very 
much less restraint on the behaviour of grandchildren in the presence of 
their grandparents. In general also, in Africa as elsewhere, grandparents 
are much more indulgent towards their grandchildren than are parents 
to their children. A child who feels that he is being treated with severity 
by his father may appeal to his father's father. The grandparents are the 
persons above all others who can interfere in the relations between 
parents and children. The possibility of this interference has important 
social functions. Elizabeth Brown remarks that 'in Hehe society the 
presence of the grandmother minimizes possible friction between mother 
and daughter'.' In any relation of subordination and superordination 
conflict is always possible. This is true of the relation of fathers and sons 
and of mothers and daughters in a great many societies. The son is 
subordinate to his father, but the latter is subordinate to his father in 
turn, and similarly with a mother and mother's mother. Control of the 
behaviour of parents towards their children therefore falls in the first 
5 place to their parents. In South Africa a man who is appealing to his 
ancestors for help, as when he is offering them a sacrifice, frequently, 
perhaps usually, makes his first appeal to his deceased father's father, 
and asks him to pass on the request to the spirit of his father and to the 
other ancestors. 

J In the passage of persons through the social structure which they 
enter by birth and leave by death, and in which they occupy successive 
positions, it is not, properly speaking, children who replace their parents, 
but those of the grandparents' generation are replaced by those of the 
grandchildren's generation. As those of the younger generation are 
^ moving into their positions of social maturity those of the older genera- 
tion are passing out of the most active social life. This relation of the two 
generations is recognized in some African peoples. In East Africa, where 
age-sets are arranged in cycles, the cycles are such that a son's son may 
' Elizabeth Fisher Brown, op. cit. 


frequently belong to the same one as his father's father. This explains 
also some of the African customs as to the relation of a child to its great- 
grandparent. Rattray reports for the Ashanti that a great-grandchild is 
called 'grandchild don't touch my ear', and the touch on the ear of such 
a relative is said to cause speedy death. Remembering the way in which 
small infants frequently reach for the ear of anyone close to them, this 
is a way of indicating the existence of a social distance between men and 
the children of their grandchildren. In the normal organization of 
generations there is no place for any close definite relation between . 
these two relatives. For any man the birth of children to his grand- 
children is the sign that he is approaching the end of his life. ' 
^/^The relation betweengrandparents and grandchildren that has here 
been~5nefly^dicatecl ismstitutionalized in various ways in African and 
other societies. There"ls a'widesplread custom of privileged familiarity 
between grandchildren and grandparents. The grandchild may tease 
his relative and joke at his expense. This custom of permitted disrespect 
to grandparents is found in tribes of Australia and North America as well 
as in many African peoples. A good example is one from the Oraons of 
India reported by Sarat Chandra Roy.' 

, The replacement of grandp arpnt-:; by thpjr grandchildren is in a way 
recognized in the w idespread cuatom-of giving a child the name of a 
grandparent. Amongst theHenga, when a child is born the husband 
is greeted with the words 'A father has been born to you to-day', having 
reference to the fact that the child will be given the name of its grand- 
father. A further step is taken in some peoples by the formation of a 
belief that in some sense the grandparent is 'reincarnated' in a grand- 

One aspect of the structural principle with vyl uch we are here con- ' 
cefned IS that one generation is replaced in course of time by the genera- 
tion of their grandchildren. Another aspect of the same principle is that 
the two generations are regarded as being in a relation, not of super- 
ordination and subordination, but of simple friendliness and solidarity 
and something approaching social equality. This may sometimes result 
in what may be called the mergi ng of alternate geperations^ a structural 
principle of fundamentar importance in the native tribes of Australia 
and in some Melanesian peoples. A man with his 'father's fathers', his 
'son's sons', and his 'brothers' in the classificatory sense form a social 
division over against his 'fathers' and 'sons', who constitute another 

Where the principle makes itself apparent in some AfricarL^oples is', 
in a peciiliai"~feature"orthe kinship terminology, whereby the term that; 
primarily means^ Svife'_ is applied3y^ a man to his granddaughters ori 
his grandmothers, and a woman applies to her grandson the termi 
meaning^Tiusband'. The custom has been reported for the Ganda, thcj 
' Sarat Chandra Roy, The Oraons of Chota Nagpur, Ranchi, 1915, pp. 3 5 2-5- 



Pende, the Ila, the Yao, the Ngonde, and the Henga. Thus amongst 
the Ngonde the term nkasi which is appUed to the wife and to the 
brother's wife is also apphed to all a man's 'granddaughters' in the classi- 
ficatory sense (but not to the 'granddaughters' of his wife) and to the 
wives of his 'grandsons', own and classificatory. Inversely it is applied to 
the wives of the men who are called 'grandfather'. The real significance 
of this is that these women are thus merged with those of his own genera- 
tion. That this is the real meaning of ihe custom m^y^be seen by reference 
to the term mzcinangu which Sanderson translates 'compeer' and which 
is applied primarily to persons of one's own generation and implies 
equality. 'For a grandson mwisukulii-niwinangu is always used in 
preference to mwisukulii by itself, and indicates that a grandson is 
treated as a "brother", a younger "brother" but an equal.' Similarly the 
term zvamyitu, meaning 'kinsman', is used for relatives of one's own 
generation, and may also be used for any relative or connexion except 
the generation immediately above or below that of the speaker (e.g. 
father or son) as it implies a degree of intimacy not permitted with those 

/ The use of a terrnmea ning 'wife' for a granddaug hier-iaugt not be 
^ assumed to imply the existence of a custom of marriage with a grand- 
daughter either in the present or in the past. Amongst the Ngonde 
and Henga such marriages are not permitted. But once the grand- 
daughters and grandmothers have been included in one's own genera- 
tion by this merging of alternate generations, the possibility of marriage 
suggests itself. There are therefore African tribes in which such marriages 
are regarded as permissible. It is said to be legal, though rare, amongst 
the Yao for a man to marry the daughter of his own child (Sanderson). 
Amongst the Kaonde a man may marry the daughter of his brother's 
son or of his brother's daughter, but may not marry the daughter of his 
sister's daughter (Melland). In the reverse direction, in the Ngonde 
tribe a man may be required to marry a 'grandmother' on the death of 
his 'grandfather', apparently in order to provide for her if she would 
otherwise be destitute. 

There is therefore discoverable in African societies, as in many other 
societies in various parts of the world, a social structure based on rela- 
tions of generation. Between two proximate generations the relation is 
normally one of essential inequality, authority, and protective care on the 
one side, respect and dependence on the other. But between the two 
generations of grandparents and grandchildren the relation is a con- 
trasting one of friendly familiarity and near equality. The contrast 
between the two kinds of relation is itself an important part of the 
structural system and is emphasized in some of the accompanying 
institutions ; for example, the contrast between restraint in the presence 

' Meredith Sanderson, 'The Relationship Systems of the Wangonde and 
Wahenga Tribes, Nyasaland', J. Roy. Anthrop. Inst. Hii, 1923, pp. 448-59. 


o£ a father or his brother and the freedom of joking with a grandfather. 
The way in which this structural principle provides for a condition of 
equilibrium in social relations is one deserving of careful investigation. 
There is not room to deal with it here. 

In most African kinship terminologies no distinction is made between 
grandparents on the father's side and those on the mother's side. We 
have seen that even in a system such as that of the Masai, where there 
is great emphasis on the distinction between agnatic and other cognatic 
relatives, this distinction is not made in the second ascending generation. 
This seems to be associated with a single general pattern of behaviour 
towards all those one calls 'grandfather' or 'grandmother', which does 
not, of course, exclude the existence of certain special relations, such as 
that of a man with his own father's father. 

^ InAjdaat' might perhaps be called the normal use of the generation 
principle in kinship structure, the generations provide basic categories. 
Any category of relatives can be placed in one particular generation: 

, uncles in the first ascending, nephews in the first descending, and so on. 
But in some terminologies a single term may be used for relatives of two 
or more generations. A study of instances of this serves to throw light 
on the principles involved in the various methods of ordering relation- 

In order to deal with this subject we have to consider the question of \ 
interpersonal rank or status in relationships. In a social relation two 
persons may meet as equal or approximately equal in rank, or one may 

/be of superior rank to the other. Differences jn rank may show themselves 
in many different forms. They a re perhap s jnost easi ly, seen in the rules 
of etiquette, or in an attitude ofdeference that the inferior is expected 
to show to his superior. 

T here may be i ne^iialityLof rank within one generation. There are 
systems in Oceania and Africa in which a father's sister's son is superior 
in rank to his mother's brother's son. In a great number of systems an 
older brother, or a classificatory 'senior brother', is superior in rank to 
his junior. The difference in rank may be more emphasized in some 
systems than in others. 

-^^ There seems to be uniyersally some more or less marked inequality of 
rank be twe en certain,relatives belonging to two proximal generations, 
as between father and son, and there is some evidence of the existence 
of a tendency to extend this to all relationships between two such 
generations, the person who is senior by generation being superior to 
the one who is junior. But, as we shall see, this general tendency is 
sometimes overborne. We have already seen that between the alternate \ 
generations (of grandparents and grandchildren) there is a widespread \ 
tendency to make the relationship one of approximate equality. 


A relationship of unequal rank is necessarily asymmetrical. A^^rn^ 

metrical relationship is one in which each of the two persons observes 

the same, or approximately the same, pattern of behaviour towards the 

C/other. In an asymmetricaLrelationship there is one way of behaviour for 

one of the persons and a different, complementary behaviour for the 

other; as when a father exercises authority over a son, and the son is 

deferential and obedient. Terminology can also be symmetrical or 

J-^ asymmetrical. In the former case each applies the same term of relation- 

shipto the other; in the latter there are two different, reciprocal terms, 

as uncle — nephew. Where terminology in a relationship is symmetrical 

it is frequently an indication that the relationship is thought of as being 

approximately symmetrical in respect of behaviour. 

/ . We sometimes find terms of relationship that ha ve no gener ation 

.refgiieijee. An example is provided by the Masai term sotzca, which may 

be applied to relatives of any generation. It refers to persons who are 

'relatives' in a general sense but do not belong to any one of the specific 

categories into which nearer relatives are divided. It corresponds to 

some extent to the Old English term *sib', and it is interesting that both 

these terms have reference to 'peace' ; sib or sotzva are those with whom 

one should live at peace. 

The English 'cousin' can be used for kin of different generations, and 
generation position has to be indicated by terms such as 'once removed'. 
The word is derived from the Latin consobrinus (mother's sister's son), 
and therefore originally had a generation reference to one's own genera- 
tion, but in the Middle Ages it seems to have been regarded as equivalent >, 
to the Latin cotisanguineiis. 

The frequently observed absence of marked inequality between 
grandparents and grandchildren may be occasionally reflected in the 
terminology, either by the use of a single self-reciprocal term between 
these relatives or by applying to grandparents and to grandchildren a 
term that is normally used for relatives of one's own generation. This 
has been noted in the last section. 
I D . An inte resting example of a system in w hich terms having a primary 
y^ reference to one generation are applied to relatives of other generations 
46- provided by the Nandi of East Africa. The term kamet refers in the 
first instance to the mother and her sisters (first ascending generation). 
But the mother's brother's daughter is also called katnet and is addressed 
by the same term of address as the mother. Correspondingly the term 
imamet refers primarily to the mother's brother, but is also applied to 
the mother's brother's son. Further, the children of the mother's 
brother's son, who belong to the first descending generation, are also 
called 'mother's brother' and 'mother'. 

The same feature appears amongst the Bari and the Kitara. In the 
former a single term [mananye) is used for the mother's brother, his 
children, and his son's children, and we may note that the wife of any 


male mananye is called 'grandmother' (yakanye). Amongst the Kitara 
the mother's brother and his son are both called 'male mother' (nyina 
rumi) and the mother's brother's daughter is called 'little mother' {nyina 
ento) like the mother's sister. Reciprocally a man calls both his sister's 
son and his father's sister's son by the same term mwhiwha. There is 
a partial application of the same principle amongst the Masai, where the 
mother's brother is ol-apu and his children and his son's children are 
ol-le-'ng-apu and en-e-'ng-apu. 

This peculiar type of_terminology_has_been found in a number of 
soci^_tie^ri_different parts of the world, and is called by anthropologisl&- 
thejQniaha-type. Its widespread occurrence shows that it cannot be 
regarded as the product of some accident of history ; we should seek some 
^theoretical explanation. It can be regarded as a method of expressing 
and emphasizing the unity and solidarity of the patrilineal lineage group. 
A man belongs to a patrilineal lineage. He is closely connected with his 
mother's lineage, which plays an important part in his life, second only 
to that of his own. His connexion with that lineage, being through his 
own mother, is with the first ascending generation. By the terminology 
he treats all the members of that group, through three (or more) genera- 
tions, beginning with that of his mother, as belonging to a single cate- 
gory; the females are 'mothers' to him and the males are 'mother's 
brothers'. For all these persons, and for the group as a whole, he is a 
'sister's son'. Thus in its relation to this person the lineage of three 
generations is a unity ; we can therefore speak of the structural principle 
that is applied in these systems as the principle of the unity of the 

mother's father 

A O = A 

mother's brother mother \ father 

A O 

'mother's brother' 'mother' 


A O 

'mother's brother' 'mother' 

_^^xln systems of the Omaha type the principle is applied not only to the 
mother's lineage but also to other lineages with which a person is closely 
connected by some individual relationship. Amongst the Masai, for 
example, all the members of the lineage of a man's wife, beginning with 
his wife's father's father, are his aputani. 

In a previous section the classificatory terminology in general was 

-inferpreted as a way of recognizing the principle of the unity of the 
sibling group. The special Omaha form of the classificatory syste m is 


here interpreted as a way of recognizing^ t] w unhy of thf^tae age gr oup. 
Thus a single method of interpretation is apphed throughout, and this 
gives a simphfication or economy of theory. If the whole question were 
merely one of the use of terms of relationship the subject would not be 
of any importance; but it is here held that the terminology is used as a 
means of ordering relationships for social purposes, and of this there is 
already abundant evidence.' 

For the purpose of the analysis that is to follow we must note here 
that in some of the tribes of this region of Africa there is a single self- 
reciprocal term for mother's brother and sister's son. In Masai this term 
is ol-apu. In Nandi the two relatives may each address the other as mama. 
This symmetrical terminology suggests that the social relation may also 
by symmetrical, i.e. that there is a single pattern of behaviour towards 
an ol-apu or a male mama whether he is a mother's brother or a sister's 
son. The accounts we have of these tribes do not permit the assertion 
that this is really so, though there is some slight indication in Hollis that 
it may be so for the Nandi. 

The use of a symmetrical (self-reciprocal) terminology between 
mother's brother and sister's son is also found in some of the tribes of the 
Nuba Hills in Kordofan, namely, Heiban, Otoro, Tira, Mesakin, Koalib, 
sC^^fp^ and Nyima. All these tribes also use a self-reciprocal terminology for the 
J "" relationship of wife's father and daughter's son, as do the Masai. Also, 
Csk the first four of the Nuba tribes mentioned above use a self-reciprocal 

terminology between grandparent and grandchild. The use of self- 
reciprocal terms for these relationships puts relatives of different 
generations into one terminological category. 

There is a special variety of the Omaha type of terminology found in 
the Shona, Ndau, and Shangana-Tonga peoples of Southern Rhodesia 
and Portuguese East Africa.^ The most southerly tribe of the Shangana- 
Tonga group is the Ronga; according to Junod this tribe uses the 
common Southern Bantu term malume (literally 'male mother') for the 
mother's brother and his son, the reciprocal being miipsyana, which thus 
applies to the sister's child and the father's sister's child ; the mother's 
brother's daughter is called mamana ('mother'). Thus the Ronga 
terminology is similar to that of the Kitara. 

The other tribes of this cluster have a different system involving an 
extended use of the term for grandfather. The mother's brother is called 

' For an exposition of the theory see A. R. Radcliffe-Brown, 'The Study of 
Kinship Systems', J. Roy. Anthrop. Inst. Ixxi, 1941, pp. 1-18. 

^ For the Shangana-Tonga tribes see Henri Junod, The Life of a South African 
Tribe, 1913, and E. Dora Earthy, Valenge Women, 1935. For the Ndau, Franz 
Boas, 'Das Verwandtschaftsystem der Vandau', Zeitschrift fiir Ethnologic, 1923, 
pp. 41-51. For the Shona, B. H. Barnes, 'Relationships in Mashonaland', Man, 
xxi, 1 93 1, p. 210. Mr. J. F. Holleman has kindly permitted me to see the 
manuscript of his very thorough analysis of the system of the Hera tribe of 


'grandfather' — sekuru in Shona, tetekulu in Ndau, kokzvana in Lenge 
and other tribes of the Shangana-Tonga group ; both the Shona and the 
Ndau terms are derived from a stem meaning 'father' and kulu or kuru 
meaning 'great'. The sons and son's sons of the mother's brother are 
also called 'grandfather'. The reciprocal of 'grandfather' is, of course, 
'grandchild' — muzukuru in Shona, muzukidu in Ndau, ntukulu in 
Shangana-Tonga. This term, which is used by a man for the children of 
his son or daughter, is thus also applied to his father's sister's child and 
his sister's child, both of whom call him 'grandfather'. A man thus may 
have 'grandchildren' who are older than himself and 'grandfathers' who 
are younger. 

In these tribes, as in the Ronga, the daughter and the son's daughter P^^ j, 
of the mother's brother are called 'mother' — mayi or mayi nini (littleNf t 
mother) in Shona, mai in Ndau, manana in Shangana-Tonga. The4(^ JL. 
children of these women are therefore called 'siblings' {makwabu or ^ ^ 
makzveru in Shangana-Tonga). . C 

y This terminology expresses the unity of the mother's lineage. All the 
female members of a man's mother's lineage in her own and succeeding 
generations are his 'mothers'. All the men of the lineage through several 
generations fall into a single category, but instead of being called 
'mother's brother' they are called 'grandfather'. The male members of 
the lineage are placed in a category that refers primarily to the second 
ascending generation, while the females are placed in one that refers 
primarily to the first ascending generation. 

The principle of lineage unity is recognized in these tribes in other 
features of the terminology. Thus in the Shona tribes a man calls all the 
men and women of all generations of the lineages of his father's mother 
and his mother's mother 'grandfather' and 'grandmother', and they all 
call him 'grandchild'. In each of these lineages his individual genealogical 
relation is with a grandmother, so the whole lineage as a unity becomes 
a collection of 'grandfathers' and 'grandmothers'. When a man marries, 
all the women of his wife's lineage are his 'wife's sisters' ; the term in the 
Hera tribe is ?nuramu, and Holleman translates this as 'potential or even 
preferential wife or husband', the term being a self-reciprocal one. The 
term for wife's father is mukarabghwa or mukarahzve, reciprocal ?}mk- 
washa ; all the male members of the wife's lineage through all generations 
are called 'father-in-law'. There is inequality of rank between father-in- 
law and son-in-law, the former having the position of superiority, and 
this applies apparently to all relationships in which these terms are used. 
These systems of terminology are clearly based on the principle of the 
unity of the lineage. But that does not explain why the mother's brother 
is called 'grandfather'. To understand that we must refer to certain 
features of the social relationship between mother's brother and sister's 
son in these tribes. 

In the type of kinship system with which we are now dealing a person 


is under the authority and control of his agnatic kin, his father and 
father's brothers and the male, and sometimes the female, members of 
his own agnatic Hneage. It is with this groupan^jlSLXoeiohers that a 
person has his most iinpui'laiil jural rtd Mions V^'ps, i {;. relationshi ps 
(^,— — -deri lltiJ by duliti^ iiiid ligltlj. Th(. mollTrfSoes not belong to this group, 
though she is attached to it by marriage and is to some extent herself 
under its control, particularly under the control of her husband. A father, 
however affectionate he may be to his son, is a person to be respected 
and obeyed, and so are his brothers and, in most of the systems of this 
type, his sisters. The mother, though she must, of course, exercise some 
disciplin e over ji ^^ung ch ^\drf-n^ is prima rily _the peiiJUl'i' wlm Jifives 
affectioiiate care. Just as the relation to the father is extended to his 
sibling group, so the relation to the mother is extended to hers. The 
mother's sisters are also 'mothers' and the mother's brothers are -male 
motliers'. The mother's brother is not a person to exercise authority 
over his sister's son; that right is reserved to his 'fathers'. It is the 
mother's brother's part to show affectionate interest in his nephew and 
give him aid when he needs it. In the family of his mother's brother a 
man is a specially privileged person. 

To establish and maintain a fixed pattern of behaviour for a particular 

1 type of relationship it is useful, and in some instances necessary, to adopt 

I some conventionalized symbolic mode of behaviour which expresses in 

I some way the character of the relationship. In the type of kinship system 

^ we are here considering the kind of relationship that is appropriate for 

a sister's son is symbolically expressed in certain definite customs of 

privileged familiarity exercised by the nephew. He may, for example, 

be permitted, or indeed expected, to joke at his uncle's expense, as in the 

Winnebago and other tribes of North America. He may be permitted 

to take his uncle's property, as in the vasu custom of Fiji and Tonga and 

as described for South Africa by Junod.' 

This relation of privileged behaviour may be extended from the 
mother's brothers to the whole of her lineage, though it will be primarily 
exercised towards the own mother's brother. The extension sometimes 
takes in the dead members of the lineage, the ancestral spirits. In South- 
east Africa a man's own ancestors of his own lineage are believed to 
watch his conduct and punish him for any breach of duty. His mother's 
ancestors have no business to exercise authority over him in this way; 
on the contrary he may go to them for help, not approaching them 
directly, but through his mother's brother or mother's brother's son, 
who can sacrifice to his own ancestors to obtain their help for his nephew 
or cousin. Junod gives an account of how the sister's sons and daughter's 
sons of a dead man symbolically express their relation at the final funeral 

■ See A. R. Radcliffe-Brown, 'The Mother's Brother in South Africa', South 
African Journal of Science, xxi, 1924, pp. 542-55, in which this analysis was 
first offered. 


ceremony, interrupting the prayers and grabbing and running away 
with the sacrificial offering, which they then eat. 

There is an East African custom which has, I think, been misinter- 
preted. This is the custom of the mother's brother's curse. It is said that 
a man fears the curse of his mother's brother more than that of any other 
relative. This is sometimes interpreted as though it means that the 
mother's brother regularly exercises authority over his nephew and that 
his authority is greater than that even of a father. I suggest that the 
proper interpretation is that the mother's brother will be the last person 
to use his power of cursing and will only do so in exceptional and serious 
situations, and that it is for this reason that it is feared more than the 
curse of the father. A matter that ought to be inquired into is whether, 
where the relationship appears to be symmetrical, as in the Nandi, the 
uncle may also be cursed by his sister's son. 

^^Jrf these syste ms the behaviour tQ w^^'ds a mnthi^r'F; brithfr in in 
marked con trast to that towards a father . A man's relation to his mother's 
relatives is as important as that to his father's kin, but of a different and 
contrasting kind. In the matter of interpersonal rank the father is very 
definitely superior to his son. But the mother's brother is not, or not 
markedly, superior to his nephew. The relationship may be treated as 
one of approximate equality, or the nephew may even be treated as 
superior. Thus in Tonga and Fiji the sister's son is quite definitely 
superior in rank to his mother's brother and to his mother's brother's 
son. In Tonga the term eiki (chief) is used to indicate a person of superior 
rank, and a sister's son is said to be eiki to his mother's brother. Similarly, 
Junod reports from the Ronga that a nephew is described as being 
a 'chief to his mother's brother. 

If there is, indeed, as there seems to be, a general tendency to attribute 
superior rank in interpersonal relations to relatives of the parents' 
generation, the special relation to the mother's brother that exists in 
these societies is directly contrary to it. The 'normal' rank relation of 
proximal generations, if we may venture to call it so, is destroyed in the 
relation to a mother's brother. There are two ways in which this may be 
reflected in the terminology. One is the use of a self-reciprocal term 
between mother's brother and sister's son, as in the Masai and Nandi and 
some Nuba tribes, whereby the relationship is treated as one that is 
symmetrical and therefore of approximate equality. Another is to place 
the mother's brother in the second ascending generation (by a sort of 
fiction) and call him 'grandfather'. We have already seen that there is 
a widespread tendency to make the relation between the grandparent 
generation and the grandchild generation one in which there is an 
absence of marked inequality of rank and one in which the junior 
generation is privileged in its behaviour towards the senior. This does 
not prevent the recognition of some measure of superiority for some 
relatives of the second ascending generation, for example, the father's 


father and his sister in some patrilineal systems. The inclusion of the 
mother's brother in the category of 'grandfathers' removes him from 
that generation-category to which there is a tendency to attach superiority 
of interpersonal rank and places him, by a fiction, in one towards 
which the relation is one of easy familiarity, approximate equality, or 

/ In the system of the Shona-Ndau-Tonga cluster of tribes, therefore, 

we have first of all the application of the principle of unity of the lineage. 

/For any person his mother's lineage is a single united group. The women 

jfof the lineage, who are mostly dispersed amongst the families into which 

I they have married, are all relatives of one kind, his 'mothers'. Amongst 

I the men of the group his closest connexion is with his mother's father 

I and brother and his mother's brother's son. With them he stands in 

I a specially privileged position ; they are expected to show him affectionate 

I indulgence. They and all the other males of the lineage constitute a single 

unity of which the representative individual is the mother's father; they 

i are all 'grandfathers' and he is for them a 'grandchild'. For the lineage 

I as a unity the women born into it are its children, but their children are 

I not. In Africa the term 'child of our child' is sometimes used for the 

I child of a woman of the family, a sister's or daughter's child. So for the 

\ lineage treated as a unity the children borne by its female members into 

\other lineages are 'children of our children' — grandchildren. 

It may be noted, for the purpose of comparison, that our English word 
'uncle' is derived from the Latin avunculus, which was the term for 
mother's brother but not for father's brother, and which is literally 'little 
grandfather', from avus. Moreover, the philologists believe that the Old 
English term for mother's brother, eme or earn, was originally a modifi- 
cation of the word for grandfather. 
{^^ Qnej^oiEpose of this section has been to show how the division of kin 
into generations can be used as a means of formalizing relations of inter- 
personal rank. One further example may be added by referring to the 
arrangement of nominal generations in the Nkundo of the Belgian Congo. 
Although Father Hulstaert's statements are not very clear, it would 
seem that the important social group, which he speaks of as a clan or 
exogamous group, is a patrilineal lineage of seven generations or so. 
The brothers and sisters of the father and members of the lineage of that 
generation are all 'fathers' (baise). The children of a female 'father' 
(father's sister) are also 'fathers' male and female, and therefore the 
children of a mother's brother (male mother — nyangompane) are 'chil-r 
dren' (bana). There is thus established a difference of rank between 
cross-cousins, the father's sister's children ranking above their mother's 
brother's children in a way similar to that in which a father ranks above 
his child. That this is a matter not merely of terminology but also of 
behaviour is evident from Father Hulstaert's account. Consistently with 
this a man treats the children of his father's sister's son (who is 'father' 



to him) as 'brothers' and 'sisters', A diagram may help to make the 
matter clear. 

nkoko (grandfather) 



ise (father) 


ise (father's sister) nkoko (grandfather) 


A O A O 

ise ise ise ise 

A I 

'brother' ? ise 

Since my father calls his father's sister's children 'fathers' I call them 
'grandparents', and I call the children of my father's father's sister's 
son 'fathers' male and female. Father Hulstaert does not tell us the rela- 
tion of a man to the children of his father's sister's daughter (who is 
his female 'father'). It seems very probable that they also are male 
and female 'fathers'. If this be so, then for any person there is a series 
of female lines each stemming from his own lineage ; the descendants 
through females of a father's sister or a father's father's sister are 
'fathers' to him and rank above him, he being 'child' to them. For our 
present purpose the Nkundo system affords anot-h^r examp le nf the use 

^o f terms hav ing a generation reference to establish relatio ns, i)f rank, 
tog ether with the us e_ot .such terms, to establish categories containing 
relatives of different real generations. Relatives of one's own generation 
are given superior rank by being called 'father'. 


Every kinship system provides each person in a society with a set of 
dyadic (person to person) relationships, so that he stands, as it were, at 
the centre of a narrower or wider circle of relatives. During his life the 
body of his relatives is constantly changing by deaths and births and by 
marriages — his own marriage and the marriages of his relatives: 
___^ In many societies the kinship system also includes a different kind of 
structure by which the whole society is divided into a number of separate 
groups, each consisting of a body of persons who are or who regard them- 
selves as being a unilineal body of kindred. Such kinship groups are 
moieties, clans, and lineages. Moieties, by which the society is bisected, 
do not exist in Africa except amongst the Galla, though they are impor- 
tant in some parts of the world. The distinction b^veen clan and lineag^ 
is that in a lineage group each member can^ctually, or at least theo- \ 
reticatty7TraCe~^^ts-genealogical connexion ^^^th any other member by i 
descent from a known common ancestor, whereas in a clan, which is I 


usually a larger body, this is not possible. A moiety may be divided into 
clans and usually is so. Clans may be divided into sub-clans, and clans 
or sub-clans may be divided into lineages. A lineage of any considerable 
size is usually divided into branches, which are themselves smaller 
lineages, and these again may be subdivided. For structures having suc- 
cessive segmentations the term 'polysegmentary' has been suggested. 
Such systems have been excellently described by Evans-Pritchard for 
'the Nuer, and by Fortes for the Tallensi. 

\^ It is usual to apply the term 'clan' to both patrilineal and matrilineal 
groups, but some American ethnographers use the term 'clan' only for 
matrilineal groups and 'gens' for patrilineal, and Dr. Nadel has adopted 
this usage in describing in this volume the two sets of groups of the 
Nyaro. Some writers in the volume have made use of compounds-;- 
patri-clan, matri-clan, patri-lineage, and matri-lineage. If these seem to 
some readers somewhat barbarous it must be remembered that some ■ 
technical terms are needed for concise description, and have to be 

.^ -—The term 'clan' has often been used without any clear definition. 
There are, of course, many different kinds of clan systems, but the term } 
should be used only for a group having unilineal descent in which all \ 
the members regard one another as in some specific sense kinsfolk. One 
>eay of giving recognition to the kinship is by the extensive use of the 
/x"'^ classificatory terminology, so that in a system of patrilineal clans a man 
regards all the men of his clan aS being his classificatory 'fathers', 
'brothers', 'sons', 'grandfathers', or 'grandsons'. Frequently, but not 
universally, the recognition of the kinship bond uniting the members of 
the clan takes the form of a rule of exogamy which forbids marriage 
between two members of the same clan. Where clans are divided into 
sub-clans it may be only to the smaller group that the rule of exogamy 

.^^ Member^ip of a clan is normally determined by birth : where clans 

are matrilineal the children of a woman belong to her clan ; where they 

/ / are patrilineal the children belong to the father's clan. But in some tribes 

J ^^ there is a custom of adoption. Where a man is adopted into a patrilineal 

' ^ clan, thereby abandoning his membership of the clan into which he was 

born, his children belong to the clan of his adoption, not that of his 

birth. In some African tribes the position of a child in the social structure 

depends on the source of the marriage payment for his mother. Thus, 

among the Lango children belong to the clan that has provided the cattle 

for the marriage payment for their mother. The father might not be a 

clansman, but might be a war captive or the sister's son of a clansman 

who was provided with a wife through cattle belonging to the clan, or 

the children might have been born outside marriage and the mother 

later married into the clan.' 

' T. T. S. Hayley, The Anatomy of Lango Religion and Groups, 1947, p. 40. 


If we look at a structure of clans or lineages from the point of view of 
an individual it appears as a grouping of his relatives. In a patrilineal 
system the members of his own clan are his agnatic kinsfolk, and the 
nearest of these to him are the members of his own lineage. The 
members of his mother's clan or lineage are also his kin, through his 
mother. He may apply to them the appropriate classificatory terms, and 
in some systems he may be forbidden to marry any woman of his 
mother's patrilineal clan. The members of his father's mother's clan and 
his mother's mother's clan may also be recognized as relatives, and those 
of his wife's clan or lineage may all have to be treated as relatives by 

> ^.y ^ clan syst em, however, also provides a division of the tribe into a 
number of distinct separate groups, each having its own identity. The 
clans may then, as groups, play an important part in the social, political, 
or religious life of the tribe. The extent to which they do this depends 
on the degree to which they are corporate groups. A group may be 
spoken of as 'corporate' when it possesses any one of a certain number of 
characters : if its members, or its adult male members, or a considerable 
proportion of them, come together occasionally to carry out some col- 
lective action — for example, the performance of rites; if it has a chief or 
council who are regarded as acting as the representatives of the group 
as a whole; if it possesses or controls property which is collective, as 
when a clan or lineage is a land-owning group. In parts of Africa it is 
very common to find that land is held or owned by lineage groups, which 
are thus corporate groups. 

An example of a society in which there are both patrilineal and matri- 
lineal corporate kin groups is provided by the Yako, described in this 
volume by Professor Daryll Forde. There are corporate patrilineal 
lineages (yeponema) each having a leader and collectively owning landed 
property. There are also corporate patrilineal clans (yepun), each con- 
taining several lineages, and the ideal arrangement is one by which the 
clan has a ritual head, a shrine for clan rites, and a meeting-house for the 
men. The Yako have, in addition, a system of matrilineal clans [yejima) 
divided into lineages, and these also are corporate groups uniting for 
clan rituals. The patrilineal clans are 'compact', i.e. the male members 
with their families live together in one delimited area; whereas the 
matrilineal clans are 'dispersed', the various members being scattered 

. through the village settlement and living in the different areas of the 
patrilineal clans. 

It should be noted that as a rule it is the adult men who really consti- 
tute the corporate kin group, and this is so for those systems that have 
descent through females. A good example is provided by the tribes of the 
lower Congo, described by Dr. Richards in this volume. Villages or 
hamlets are formed of matrilineal lineages; all the men of a single 
lineage live together with their wives and young children, boys when 



they reach a certain age leaving their parents to join their mother's 
brother and his village. It is therefore the men of the lineage who form 
the corporate group, holding rights over land and acting collectively in 
.various ways. 

Professor Gluckman regards the absence of corporate kin groups (clans 
or lineages) as an important distinguishing characteristic of a number of 
tribes of Central Africa. The typical corporate group in that region is 
a village constituted by the persons who attach themselves to a headman. 
This group is an open, not a closed group ; that is, individuals or families 
may join or leave it, moving from one village to another. It is usual that 
a number of the inhabitants of a village at any time should be related, 
either by cognatic ties or through marriage with the headman or with 
one another, but they do not form a unilineal kin group, which is by its 
constitution a 'closed' group. 

Some of these tribes have clans, patrilineal in some instances, 
matrilineal in others, but the clans are dispersed and not corporate. 
Thus the Ila and Bemba and other tribes have dispersed matrilineal 
clans. The members of one clan are scattered through the tribe; they do 
not ever come together to take any kind of collective action, and have no 
single authority (headman or clan council). They have no p ogitive cl an 
rites ; the identity of the c1anj _gnd its unity as a separate groupof kin dred , 

is maintauied by negative ritual obse rvanc es common to all the memb ers, 
"sucTTas refrai ning from killing or eatmg a certain animal (the 'totem' of 
the clan). "A member of the clan does not know all the other members, 
but if two persons meet who know or discover that they belong to the 
same clan they are expected to behave towards one another as kinsfolk, 
and since all members are kin they may not intermarry. It does not seem 
that in these tribes matrilineal lineages are given social recognition 
except in royal families. 

One of these Central African tribes, the Lozi , described in this volume, ^ 
does not recognize either clans or lineages. The system, though it shows 
a slight preference in some respects for kinship in the male line, is 
characteristically one of cognatic kinship, tracing relationship through 
males and females. Theoretically the range of recognized kinship 
extends to the descendants of a common great-great-grandparent, i.e. 
to third cousins. 

The unilineal principle of reckoning rela tmngViip in nnn Unn (miln nr 
fe male) is utilized in a great vari ety of ways in different kinsb'p gyg«^^^g 
iVhere i t is used~to" create a system ot clans it facilitates th at^wide-range 
recognition of relations of kinship to whic h there is a tendency in many 
societies. A person will thereby-fiftd-kinwelf connected by specific social 
ties, subject to established institutional modes of behaviour, with a large- 
number of other persons. In the absence or weak development of politi- 
^cal structure this gives an effective system of social integration. It is not 
' See below, pp. 166 ff. 


possible to provide such very wide range in a system based on cognation, 
since that impHes the tracing of genealogical relationships through all 
lines. But eve n j nore important is that unilineal reckoning makes it 
possible to create corporate km^groups having cbntmuity in tim^ extend- . 
ing beyond the lite ot an mHividual or a family. There are innumerable 
social activities that can only be efficiently carried out by means of 
corporate groups, so that where, as in so many non-literate societies, the 
chief source of social cohesion is the recognition of kinship, corporate 
kin groups tend to become the m^ost important feature of social structure. 
Thus it is the corporate kin group, whether clan, sub-clan, or 
lineage, that controls the use of land, whether for hunting, for pastoral 
life, or for cultivation ; that exacts vengeance for the killing of a member, 
or demands and receives an indemnity. In the sphere of religion the kin 
group usually has its own cult, whether of its ancestors or connected with 
some sacred shrine. A continuing social structure requires the aggre- 
gation of individuals into distinct separated groups, each with its own 
solidarity, every person belonging to one group of any set. The obvious 
instance is the present division of the world into nations. In kinship 
systems cognatic kinship cannot provide this; it is only made possible 
by the use of the principle of unilineal descent. This is, indeed, obvious, 
but there have been writers who have used much misplaced ingenuity in 
trying to conjecture the origin of clans. 


* — In order to understand the African customs relating to marriage we 
have to bear in mind that a marriage is essentially a rearrangement of 
5iacial structure. What is meant "By" social structure is any arrangement 
of persons in institutionalized relationships. By a marriage certain 
' existing relationships, particularly, in most societies, those of the bride 
to her family, are changed. New social relations are created, not only 
between the husband and the wife, and between the husband and the 
wife's relatives on the one side and between the wife and the husband's 
relatives on the other, but also, in a great many societies, between the- 
relatives of the husband and those of the wife, who, on the two sides, arej 
interested in the marriage and in the children that are expected to result 
'from it. Marriages, like births, deaths, or initiations at puberty, are 
i'earrangements of structure that are constantly recurring in any society ;i 
they are moments of the continuing social process regulated by custom ; 
there are institutionalized ways of dealing with such events. 

We tend, unless we are anthropologists, to judge other people's 
customs by reference to our own. To understand African marriage we 
must remember that the modern English idea of marriage is recent and 
decidedly unusual, the product of a particular social development. We 
think of a marriage as an event that concerns primarily the man and 
woman who are forming a union and the State, which gives that union 




its legality and alone can dissolve it by divorce. The consent of parents 
is, strictly, only required for minors. Religion still plays some part, but 
a religious ceremony is not essential. 

We may compare English marriage with the following account of 
a 'wedding' in early England.^ 

'If people want to wed a maid or a wife and this is agreeable to her and 
to her kinsmen, then it is right that the bridegroom should first swear 
according to God's right and secular law and should wage (pledge himself) 
to those who are her forspeakers, that he wishes to have her in such a way 
as he should hold her by God's right as his wife — and his kinsmen will 
stand pledge for him. 

'Then it is to be settled to whom the price for upfostering her belongs, 
and for this the kinsmen should pledge themselves. 

'Then let the bridegroom declare what present he will make her for 
granting his desire, and what he will give if she lives longer than he does. 

'If it is settled in this way, then it is right that she should enjoy half the 
property, and all if they have a child, unless she marries another man. 

'All this the bridegroom must corroborate by giving a gage, and his 
kinsmen stand to pledge for him. 

'If they are agreed in all this, then let the kinsmen of the bride accept 
and wed their kinswoman to wife and to right life to him who desires her, 
and let him take the pledge who rules over the wedding. 

'If she is taken out of the land into another lord's land, then it is advis- 
able that her kinsmen get a promise that no violence will be done to her 
and that if she has to pay a fine they ought to be next to help her to pay, 
if she has not enough to pay herself.' 

The marriage here is not any concern of the State or political authori- 
ties ; it is a compact between two bodies of persons, the kin of the woman 
who agree to wed their daughter to the man, and his kinsmen who 
pledge themselves that the terms of the agreement will be carried out. 
The bridegroom and his kinsmen must promise to make a payment (the 
'marriage payment') to her father or other legal guardian. He must also 
state what present he will give to his bride for permitting the physical 
consummation of the marriage; this was the so-called 'morning-gift' to 
be paid after the bridal night. There was further an agreement as to the 
amount of the dowry, the portion of the husband's wealth of which the 
wife should have the use during her lifetime if her husband died before 
her. The agreement is concluded by the giving of the wed, the symbolic 
payment made by the bridegroom and his kin to the woman's kinsmen. 

In modern England the pledge or gage, in the form of a 'wedding' 
ring, is given, not to the bride's kinsmen when the marriage arrange- 

' Quoted by Vinogradoff, Outlines of Historical Jurisprudence, i, p. 252, from 
Liebermann, Gesetze der Angelsachsen, i, p. 442. The 'wedding' was the agree- 
ment or contract entered into by the kinsfolk of bride and bridegroom, equivalent 
to the Roman sponsalia, not the ceremony of handing over the bride (the Roman 
traditio puellae). 



ment is made, but to the bride herself at the wedding ceremony. The 
change in custom is highly significant. The 'giving away' of the bride 
is a survival of something which at one time was the most important 
feature of the ceremonial of marriage. 

Thus in Anglo-Saxon England a marriage, the legal union of man and 
wife, was a compact entered into by two bodies of kin. As the Church 
steadily increased in power and in control of social life, marriage became 
the concern of the Church and was regulated by canon law. There was 
a new conception that in marriage the man and woman entered into a 
compact with God (or with His Church) that they would remain united 
till parted by death. The marriage was under the control of the Church; 
matrimonial cases were dealt with in the ecclesiastical courts. 

At the end of the Middle Ages there came the struggle for power 
between Church and State in which the State was, in Protestant 
countries, victorious. Marriage then came under State control. At the 
present day to legalize a union of man and wife the marriage, whether 
there is or is not a religious ceremony, must be registered by someone 
licensed by the State and a fee must be paid. It is the State that decides 
on what conditions the marriage may be brought to an end by a divorce 
granted by a court which is an organ of the State. 

A most important factor in the development of the modern English 
(and American) conception of marriage was the idea of romantic love, 
a theme that was elaborated in the nineteenth century in novel and 
drama and has now become the mainstay of the cinema industry. In its 
early development romantic love was conceived as not within but outside 
marriage, witness the troubadours and their courts of love and Dante and 
Petrarch. In the eighteenth century Adam Smith could write: 'Love, 
which was formerly a ridiculous passion, became more grave and 
respectable. As a proof of this it is worth our observation that no ancient 
tragedy turned on love, whereas it is now more respectable and influences 
all the public entertainments.' The idea that marriage should be a union 
based on romantic love leads logically to the view that if the husband 
and wife find they do not love one another they should be permitted to 
dissolve the marriage. This is the Hollywood practice, but conflicts with 
the control of marriage by the Church or by the State. 

Another very important factor has been the change in the social and 
economic position of women during the nineteenth and twentieth 
centuries. A married woman may now hold property in her own right; 
she may take employment that has no connexion with her family life 
but takes her away from it. In the marriage ceremony many women now 
refuse to promise that they will obey their husbands. 

Not only are marriage and ideas about marriage in England and 
America the product of a recent, special, and complex development, 
but there is good evidence that they are still changing. The demand for 
greater freedom of divorce is one indication of this. Yet it is clear that 


despite all this some people take twentieth-century English marriage 
as a standard of 'civilized' marriage with which to compare African 

<^The Africa n does not think of marriage as a union based on romantic 
love although beauty as well as character and health are sought in the 
choice of a wife. The strong affection that normally exists after some 
years of successful marriage is the product of the marriage itself con- 
ceived as a process, resulting from living together and c«-operating in 
many activities and particularly in the rearing of children. 

An African marriage is in certain respects similar to the early English 
marriage described above. The dowry or dower does not exist in Africa, 
though writers who do not know, or do not care about, the meanings 
of words use the term 'dowry' quite inappropriately to refer to the 
'marriage payment'.' There is also in Africa nothing exactly correspond- 
ing to the English 'morning-gift' regarded as a payment for accepting 
sexual embraces, though it is usual for the bridegroom to give gifts to 
his bride. The two other features of the early English marriage are 
normally found in African marriages. Firstly, the marriage is not the 
concern of the political authorities but is" established by a compact 
between two bodies of persons, the kin of the man and the kin of the 
woman. The marriage is an alliance between the two bodies of kin based 
on their common interest in the marriage itself and its continuance, and 
in the offspring of the union, who will be, of course, kin of both the two 
kin-groups. The understanding of the nature of this alliance is essential 
to any understanding of African kinship systems. Se condl y, in Africa 
generally, as in early England, and in a great nunrofF'oPsocieties in 
ancient and modern times in all parts of the world, a marriage involves 
the making of a g3^ifia«ftH5y the bridegroom or his kin to the father or 
guardian of the bride. Africans distinguish, as we do, between a 'legal' 
marriage and an irregular union. In modern England a marriage is legal 
if it is registered by a person licensed by the State. Only children born 
of such a union are legitimate. But in Africa the State or political 
authority is not concerned with a marriage. How, then, are we to distin- 
guish a legal marriage ? The answer is that a legal marriage, by which the 
children who will be born are given definite 'legitimate' status in the 
society, requires a series of transactions and formalities in which the two 
bodies of kin, th ose of the husba ndjnd, those of the wife, are involved. 
InlnosrAfncanmarriages, as in the early English marriage, the making 
of a payment of goods or services by the bridegroom to the bride's kin 
is an essential part of the establishment of 'legality'. 

" _Some peopl e regard_ |Dayments of this kin d as being a 'purch ase' of 
a wifeinjthe sense injw Hiclrin EiigTand' "to^ ay^a man may_purchase 

' Belgian and some French writers make a similar misuse of the term 'dot', 
which is a woman's marriage portion of which the annual income is under her 
husband's control. 


a hoFse n z a m otor-car. In South Africa it was at one time held officially 
that a marriage by native custom with the payment of cattle (lobola) was 
'an immoral transaction' and not a valid marriage. The Supreme Court 
of Kenya in 1917 decided that 'a so-called marriage by the native custom 
of wife-purchase is not a marriage'. The idea that an African buys a wife 
in the way that an English farmer buys cattle is the result of ignorance, 
which may once have been excusable but is so no longer, or of blind 
prejudice, which is never excusable in those responsible for governing 
an African people. 

A marriage in many, perhaps most, African societies involves a whole 
series of prestations^ (payments, gifts, or services), and while the most 
important of these are from the husband and his kin to the wife's kin, 
there are frequently, one might say usually, some in the other direction. 
One of the best accounts of the whole procedure is that given by Father 
Hulstaert for the Nkundo of the Belgian Congo. ^ The procedure begins 
with the presentation, on the part of the future husband, of the ikula, 
at one time an arrow, now two copper rings. The acceptance of this by 
the woman and her kin constitutes a formal betrothal. The marriage, 
i.e. the 'tradition' of the bride, may take place before any further pay- 
ment. At the marriage, gifts are made to the bride by the parents of the 
bridegroom, by other of his relatives, and by the bridegroom himself. 
The next step is the formal prestation of the ndanga, formerly a knife, 
to the bride's father. It signifies that the husband thereafter becomes 
responsible for accidents that might befall his wife. In return there is 
a prestation from the bride's family to the husband and his family. This 
is part of the nkomi, the payment that is made to the bridegroom by the 
wife's family. The marriage is not fully established until the husband 
pays his father-in-law the zualo, a substantial payment consisting chiefly 
of objects of metal. After this the woman becomes fully the man's wife. 
When the zvalo is handed over the woman's family make a return pay- 
ment {nkomi) and give a present of food to the husband's family. The 
husband must also make a special payment to his wife's mother and 
must give a considerable number of presents to the father, mother, 
brothers, and other relatives of the bride. The relatives of the husband 
then demand and receive presents from the wife's family. The final 
payment to be made by the husband is the bosongo, formerly a slave, 
now a quantity of copper rings. 

There is, of course, an immense diversity in the particulars of presta- 
tions connected with betrothal and marriage in different societies and in 

' 'Prestation' is defined in the Oxford Dictionary as 'the act of paying, in 
money or service, what is due by law or custom'. The prestations with which 
we are here concerned are all those gifts and payments of goods or ser\'ices which 
are required by custom in the process of establishing a valid marriage. 

^ R. P. G. Hulstaert, Le Mariage des Nkundo, Inst. roy. colon, beige, 
M^moires, tome viii, 1938, chap. ii. 


each case they have to be studied, with regard to their meanings' and 
functions, in relation to the society in which they are found. For general 
theory, however, we have to look for general similarities. In the first^ 
place it is necessary to recognize that whatever economic importance ' 
some of these transactions may have, it is their symbolic aspect that we - 
chiefly have to consider. This may be made clear by the English customs 
of the engagement ring, the wedding ring, and the wedding presents. 
Though an engagement ring may have considerable value (more than 
many Africans 'pay' for their wives), the giving of it is not regarded as 
an economic or at least not as a business transaction. It is symbolic. 

In what follows the term 'marriage payment' will be used for thc) . 
major payment or payments made by the bridegroom to the wife's kin; < 
Where there is a payment from the wife's kin to the husband (as in the 
Nkundo) this will be called the 'counter-payment'. The rule in many.^ 
African societies is that if there is a divorce the marriage payment and 
the counter-payment must be returned. There are qualifications of this; 
for example, in some tribes where on divorce there are children and they 
belong to the father the marriage payment may be not returnable, or 
returnable only in part. Also, there are tribes in Africa in which, instead 
of a payment in goods, the bridegroom must serve for his wife by working 
for her kin, just as Jacob served his mother's brother Laban seven years 
for each of the two sisters, Leah and Rachel, his cousins, whom he 
married (Genesis xxix). This service, the equivalent of the marriage 
payment or of part thereof, is of course not returnable if there is 

Let us return to the early English marriage. In the formulary quoted 
above the marriage payment was called 'the price of upfostering' and 
was thus interpreted as a return to the father or guardian of the expense 
of rearing a daughter. But in somewhat earlier times the payment was 
diff"erently interpreted. It was a payment for the transfer of the woman's 
mund from the father or guardian to the husband, whereby the latter 
gained and the former lost certain rights. The term for a legitimately 
married wife in Old Norse law was mundi kjobt, meaning one whose 
mund has been purchased. In Sweden the transfer of mund was not by 
purchase but by gift, and the expression for marriage was giftarmal. In 
Roman law the marriage by coemptio, sometimes called 'marriage by 
purchase', was not the sale of a woman but the legal transfer of manus 
to her husband, and mund and manus are roughly equivalent terms. In. 
these Roman and Teutonic marriages the important point is that to 
legalize the union of a man and a woman, so that it is really a marriage, 
legal power over his daughter must be surrendered by the father and 
acquired by the husband, whether the transfer be by gift or by payment. 
yf^The early English marriage was of this type. '' 

\j /^^ InJ^irica an unmarried womanjs in a position of dependence. She 
\,y lives under the control and authority of her kin, and it is they who 


afford her protection. Commonly, if she is killed or injured her guardian 
or her kinsfolk can claim an indemnity. At marriage she passes to a 
greater or less extent, which is often very considerable, under the control 
of ber husband (and his kin), and it is he (and they) who undertake to 
afford her protection. (Note the ndanga payment amongst the Nkundo, 
by which the bridegroom accepts responsibility for accidents that may 
befall the bride.) The woman's kin, however, retain the right to protect 
her against ill treatment by her husband. If she is killed or injured by 
third parties it is now the husband and his kin who can claim an 
indemnity. It is this transfer of miind, to use the Old English term, that 
is the central feature of the marriage transaction. 

T o understand African m arriage \ve^must think of it not as an event 
Or a conditionjbut as a developing process. The first step is usually 
a formal betrothal, though this may have been preceded by a period of 

^courtship or, in some instances in some regions, by an elopement. The 

• betrothal is the contract or agreement between the two families. The 
marriage may proceed by stages, as in the instance of the Nkundo men- 
tioned above. A most important stage in the development of the marriage 
is tlie birth of the first child. It is through the children that the husband 
^d wife are united and the two families are also united by having 
descendants in common. 

We may consider African marriage in three of its most important 
aspects. Fir.^the marriage involves some modification or partial rupture 
of the relations between the bride and her immediate kin. This is least 
marked when the future husband comes to live with and work for his 
future parents-in-law while his betrothed is still a girl not old enough 
for marriage. It is most marked when, as in most African societies, the . 
woman when she marries leaves her family and goes to live with her 
husband and his family. Her own family suffers a loss. It would be a 
gross error to think of this as an economic loss.^ It is the loss of a person 
who has been a member of a group, a breach of the family solidarity. 
, This aspect of marriage is very frequently given symbolic expression in 
the simulated hostility between the two bodies of kin at the marriage 
ceremony, or by the pretence of taking the bride by force (the so-called 

' 'capture' of the bride). Either the bride herself or her kin, or both, are 
expected to make a show of resistance at her removal. 

Customs of this kind are extremely widespread not only in Africa but 
all over the world, and the only explanation that fits the various instances 
is that they are the ritual or symbolic expression of the recognition that 
rharriage entails the breaking of the solidarity that unites a woman to 
the family in which she has been born and grown up. Ethnographical 

- literature affords innumerable instances. One example may be given 

' This is the view of modern English law. If an unmarried woman is seduced 
her father can recover damages for the loss of her 'services' ; as though the only 
value attached to a daughter is as a servant. 



here.' In Basutoland, or at least in some parts of it, on the day fixed for 

the marriage the young men of the bridegroom's group drive the cattle 

that are to constitute the marriage payment to the home of the bride. 

When they draw near, the women of the bride's party gather in front of 

the entrance to the cattle kraal. As the bridegroom's party try to drive 

the cattle into the kraal the women, with sticks and shouts, drive them 

away so that they scatter over the veld and have to be collected together 

again and a new attempt made to drive them into the kraal. This goes on 

for some time until at last the cattle are successfully driven into the 

kraal. The women of the group make a show of resistance at the delivery 

of the cattle which will have as its consequence the loss of the bride.. 

• The proper interpretation of these customs is that they are symbolic 

:' expressions of the recognition of the structural change that is brought 

' about by the marriage. 

When this aspect of marriage is considered the marriage payment can 
be regarded as an indemnity or compensation given by the bridegroom 
to the bride's kin for the loss of their daughter. This is, however, only 
one side of a many sided institution and in some kinship systems is of 
minor importance. In societies in which the marriage payment is of •' ^ 
considerable value it is commonly used to replace the daughter by 
obtaining a wife for some other member of the family, usually a brother 
of the woman who has been lost. A daughter is replaced by a daughter- 
in-law, a sister by a wife or sister-in-law. The family is compensated for 
its loss. 

A second important aspect of legal marriage is that it gives the -^ 
husban^and his kin certain rights in relation to his wife and the children 
she bears. The rights so acquired are different in different systems. Some 
of these are rights of the husband to the performance of duties by the 
wife (rights in personam) and he accepts corresponding duties towards 
her. He has, for example, rights to the services of his wife in his house- 
hold. But the husband usually also acquires rights in rem over his wife. " 
If anyone kills or injures her, or commits adultery with her, he may claim 
to be indemnified for the injury to his rights. 

The husband acquires his rights through an action by the wife's kin 
in which they surrender certain of the rights they have previously had. 
The marriage payment may be regarded in this aspect as a kind of 'con- 
sideration' by means of which the transfer is formally and 'legally' made. 
It is the objective instrument of the 'legal' transaction of the transfer of 
rights. Once the payment, or some specific portion of it, has been made . 
the bride's family have no right to fetch their daughter back, and in most 
tribes, if the union is broken by divorce at the instance of the husband, 
the payment has to be returned and the woman's family recover the 
rights they surrendered. 

The rights obtained by a husband and his kin are different in some 
respects in different systems. The most important difference is in the 


matter of rights over the children the wife bears. An African marries 
because he wants children — liberornm quaerendorum gratia. The most 
important part of the 'value' of a woman is her child-bearing capacity. 
Therefore, if the woman proves to be barren, in many tribes her kin 
either return the marriage payment or provide another woman to 
bear children. 

In a system of father-right, such as the Roman patria potestas, the 
rights of the father and his kin over the children of a marriage are so 
preponderant as to be nearly absolute and exclude any rights on the part 
of the mother's kin. On the other hand, in a system of mother-right 
such as that formerly existing amongst the Nayars of southern India, the 
father has no legal rights at all : the children belong to the mother and 
her kin. This does not, of course, exclude a relationship of affection 
between father aYid child. Both father-right and mother-right are 
exceptional conditions ; most societies have systems which come between 
these .extremes and might be called systems of joint right or divided V 
right. The system of division varies and there may be an approximation 
either to father-right or to mother-right. 

Some societies in Sumatra and other parts of the Malay Archipelago 
have two kinds of marriage. If a full marriage payment is made the 
children belong to the father; we may call this a father-right marriage. 
But if no payment is made the children belong to the mother and her 
kin, the marriage being one of mother-right. 

The same sort of thing is reported from some parts of Africa, for 
example from Brass in Southern Nigeria.' The father-right marriage, 
with a substantial marriage payment, is the usual form, but if only a 
small payment is made the children belong to the mother's kin. The 
most definite example is from the Nyamwezi. In the kukwa form of 
marriage there is a payment (nsabo) made by the bridegroom to the 
father or guardian of the bride ; children of such a marriage fall into the 
possession of the husband and his agnatic kin. In the butende form of 
marriage there is no payment and the children belong to the mother and 
her kin. 

Therejs an other aspect of marriage that must be taken into account. 
In Atrica a marriage is not simply a union of a man and a woman ; it is ' 
an alliance between two families or bodies of kin. We must consider the 
marriage payments in this connexion also. 

In so-called primitive societies the exchange of valuables is a common 
method of establishing or maintaining a friendly relation between separ- 
ate groups or between individuals belonging to separate groups. Where 
material goods are exchanged it is common to speak of gift- exchange. 
But the exchange may be of services, particularly those of a ritual 
character. There are societies in which there is an exchange of women, 
each group (family, lineage, or clan) providing a wife for a man of the 
' P. Amaury Talbot, Southern Nigeria, 1926, vol. iii, pp. 437-40. 



Other. The rule governing transactions of this kind is that for whatever 
is received a return must be made. By such exchanges, even by a single 
act of exchange, two persons or two groups are linked together in a more 
or less lasting relation of alliance.' 

There are societies in some parts of the world in which the marriage 
payment and the counter-payment are equal or approximately equal in 
value. We may regard this as an exchange of gifts to establish friendship 
between two families, of which the son of one is to marry a daughter of 
the other. The kind, and t6 some extent the amount of the gifts is fixed 
by custom. But where the marriage payment is considerable in amount 
and there is a much smaller counter-payment, or none at all, we must 
interpret this as meaning that the bride's family is conferring a specific 
benefit on the bridegroom by giving him their daughter in marrfage, 
a benefit that is shared by his kin, and that the marriage payment is a 
return for this. The transaction can still be regarded as a form of 'gift- 
exchange' and as such establishes a relation (of alliance) betsveen the 

It is characteristic of a transaction of purchase and sale that once it 
has been completed it leaves behind no obligations on either the buyer 
or the seller. (This does not, of course, exclude claims based on warranty.) 
In an African marriage the position is very diff"erent. For one thing the 
marriage payment may in certain circumstances have to be repaid. In 
some tribes where the payment consists of cattle it is the same cattle with 
all their increase that should be returned. Further, in some African 
societies the family that has made the marriage payment continues to 
have an interest in the cattle or other goods of which it consists. The 
payment received for a woman's marriage may be used to obtain a wife 
for a member of her family, usually her brother. This sets up a number 
of important relations between the persons involved. 

A = b B = c C= d 

B and b are brother and sister, and so are C and c. A marries h and makes 
a marriage payment which is used to obtain a wife {c) for B. In various 
tribes the marriage payment establishes a series of special personal 
relations between b and B, between A and B, between b and c, and 
between A and c. These are defined diff"erently in different tribes. We 
may briefly consider three varieties. 

It is usual to speak of B and b as 'linked' brother and sister, and B is 
the 'linked' mother's brother of the children of A and 6, while b is the 
'linked' father's sister of B's children. In the Shangana-Tonga tribes 
there is a very special relation between A and his 'great mukonwana' c, 
the wife that B married with the payment provided by A. A can claim 

' See Marcel Mauss, Essai sur le Don. 



in marriage a daughter of c, particularl}' if his wife h dies and there is no 
younger sister to take her place.' 

In the Lovedu the relations between the families of A, B, and C 
ought to be continued in the next and succeeding generations. A son of 
A should marry a daughter of B and a son of B should marry a daughter 
of C. There is thus established a chain of connected families. The B 
family (or lineage) gives brides to and receives cattle from the A family 
and gives cattle to and receives brides from C. The linked sister h is 
said to have 'built the house for her brother' B, and she 'has a gate' by 
which she may enter the house. She has the right to demand a daughter 
of the house to come as her daughter-in-law, to marry her son and be her 
helper. Thus, in this tribe, cross-cousin marriage is systematized in terms . 
of marriage payments, and a complex set of relations between persons 
and betsveen families is created.^ In the Shangana-Tonga tribes b can 
demand a daughter of c as her co-wife or 'helper', the wife of her husband, 
not as her daughter-in-law.^ 

Amongst the Nkundo the relationships are given a different form. 
There is a special relation of b to c, the wife of her linked brother whose 
marriage was provided for by her marriage payment. The sister b is the 
nkolo of c, who is her tiMta. The ?ikoh [b) stands in a position of superi- 
ority to the nkita {c). This relation is continued in the succeeding genera- 
tions; the children of b (the nkolo) are in a position of superiority to the 
children of c. This is connected with a peculiar ordering of relations 
amongst the Nkundo by which the relation between cross-cousins is an 
asymmetrical one in which they are treated as if they belonged to differ- 
ent generations. The children of the father's sister are 'fathers', male 
and female (baise), to their cousins, the children of the mother's brother 
who are their 'children' (bana). The 'children' must show respect to 
their 'fathers' and help them. As a consequence a man regards the son 
of his father's sister's son as his 'brother' and uses that term for him.'^ 

It should now be evident that the marriage payment is a complex 
institution having many varieties in form and function. In any given 
society it has to be interpreted by reference to the whole system of which 
it is a part. Nevertheless, there are certain general statements that seem 
to be well grounded. In Africa the marriage payment, whether it be 
small or large, is the objective instrument by which a 'legal' marriage is 
established. In some instances it is a compensation or indemnity to the 
woman's family for the loss of a member. This is particularly so where 
the marriage payment is considerable and is used to obtain a wife for the 

' H. Junod, op. cit., pp. 231 et seq. 

^ E. J. and J. D. Krige, The Realm of a Rain Queen, 1943 ; J. D. Krige, 'The 
Significance of Cattle Exchanges in Lovedu Social Structure', Africa, xii. 
4 Oct. 1939, pp. 393-424- 

^ E. Dora Earthy, Valenge Women. 

* Hulstaert, Le Manage des Nkundo, p. 164 et seq. 


woman's brother. The payment may in some instances be regarded as 
part of an exchange of a kind that is used in many parts of the world to 
establish a friendly alliance between two groups. In some societies of 
South Africa and the Nilotic region it is the derivation of the cattle used 
in the marriage payment that fixes the social position of the children 
born of the union. Where the same cattle or other goods are used in two 
or more successive marriages this is in some tribes held to establish 
a special relation between the families thus formed. Where cattle are 
sacred in the sense that the cattle of a lineage are the material link 
between the living and their ancestors (having been received from those 
ancestors and being used for sacrifices to them), the use of cattle in 
marriage payments has a significance which a transfer of other goods 
would not have. This is not intended as a complete survey, which would 
be impossible within the limits of this essay. It is only an indication of 
how this institution, which is the procedure by which a husband ac- 
quires those rights which characterize a legal marriage (rights that vary 
in different societies), may be elaborated in different ways. 


It has been said above that an African marriage has to be regarded as 
a developing process. One aspect of this is the development of the rela- 
tion between the two allied families as children are born and grow up. 
We think of kinship only as a relation between two persons who have 
a common ancestor. But there is a kind of reve rse kinship between, 
persons who have a'common descendant, and it is relatiorislirps of this! 
fcTTrd"T:hat- m ' e c r ea te d by t 4^te-^^^afriage~ronceived as a process. When a 
chiTd^sborn the father-in-law of the child's father becomes the grand- 
father (mother's father) of the child, and the man's brother-in-law 
becomes his child's uncle. It is usual to speak of the relation of a man to 
his brother-in-law as an affinal relation and to trace it through the wifcr- 
But the real relation that is established as the marriage proceeds is~^ 
between the father and the mother's brother of a child or children. This 
is an elementary observation to make, but the failure to recognize 
clearly this simple fact is an obstacle to the understanding of a number 
of features of kinship systems. 

African systems differ as to the rules concerning marriage between 
kin. In many the general rule is that a man and woman who are kin, or at 
any rate closely related, may not marry, and thus no bonds of kinship; 
unite the two families before the marriage. On the other hand, there are 
many African societies in which it is thought very appropriate that a man 
should marry his cross-cousin, most usually the daughter of his mother's 
brother, more rarely the daughter of his father's sister. In such marriages 
the two families are already related before the marriage occurs. In 
marriage with the mother's brother's daughter a connexion between the 


families or lineages that has been formed in one generation is repeated 
in the next. There is also the very exceptional case of the Tswana, where 
a man may marry not only a mother's brother's daughter but such a near 
relative as the father's brother's daughter. We may expect that the 
social relations that result from a marriage alliance will differ in these 
different kinds of marriage, and a comparative study of the difference is 
desirable. It cannot be undertaken here/This section will deal only_v^h 
certai n feature s- th a t c haracte r ize the ici aT iiLnib uf a ill^U IcT Kiswife's 
rel atives in a great number of African j)ec^jes^andare also found among 
njaay other peoples in many parts of the world. 

For seventy years anthropologists have paid a good deal of attention 
to a custom found in many parts of the world and commonly referred to 
as 'mother-in-law avoidance'. This is a custom by which social contact 
between a husband and his wife's mother is limited in significant ways 
or in extreme cases entirely prohibited. 

A theory favoured by a number of anthropologists is that the purpose 
of this custom is to prevent incestuous intercourse with the wife's 
mother. It is not explained why such special and in some instances drastic 
measures are necessary, when incest with the mother or sister and other 
relatives is avoided without them. It would seem to be assumed that in 
some societies every man has a strong desire to have connexion with his 
wife's mother. It is an example of the kind of speculative theory that has 
been all too frequent in anthropology, made, in defiance of scientific 
method, without consideration of the relevant facts. 

What is really the same custom varies from complete or nearly com- 
plete avoidance to the maintenance of social distance by a reciprocal 
attitude of reserve and respect. Amongst the Ganda 'no man might see 
his mother-in-law or speak face to face with her'.' Amongst the Galla 
a man must not mention the name of his mother-in-law (actual or pros- 
pective), but he does not appear to be prohibited from speaking to her. 
But he may not drink milk from a cup she has used nor eat food of her 
cooking.^ Thus the custom has many varied forms. 

It is not confined to a man's own mother-in-law. In some societies 
a man must practise the same sort of avoidance towards the mother-in- 
law of his brother. In many there is a similar avoidance of the sisters of 
the mother-in-law, and occasionally of the wife's grandmother. But a 
man must also avoid, or maintain a respectful distance from, some of his 
wife's male relatives, particularly her father, sometimes her father's 
brothers, and in some societies her mother's brothers. It is said that 
amongst the Toro of Albert Nyanza the avoidance between son- and 
father-in-law is even more rigid than that between son- and mother-in- 
law; and amongst the Lendu, another tribe in Uganda, the father-in-law 
can never visit his son-in-law except in the event of the serious illness of 

' Roscoe, The Baganda, 191 1, p. 129. 

* Werner, in ^. Afr. Soc. xiii, 1914, p. 139. 


his daughter, whereas the mother-in-law may visit her son-in-law and 
his wife when two months have passed since the marriage.' 

With this custom of maintaining a respectful distance between a man 
and his wife's parents and other relatives of the same generation there is 
frequently associated a directly contrary relation between a man and his 
wife's brothers and sisters. This is the kind of relationship that is usually 
called the 'joking relationship'. It is fundamentally a relation expressed 
in disrespectful behaviour. Persons between whom such a relationship 
exists are not merely permitted but are expected to speak and behave 
to one another in ways that would be insulting and offensive between 
persons not so related.^ 

.^.^^^These customs of 'avoidance' and 'joking' are too frequently found 
together for us to treat the association as accidental, particularly as in 
them we find two directly contrary modes of behaviour used in a single 
social context, that created by a marriage. We cannot regard as worthy 
of serious consideration any theory or explanation that does not deal 
with both of them. 

i As a first step towards the formation of a theory we must bear in mind 

thaTm these relationships (both of 'avoidance' and of 'joking') behaviour 
is highly conventionalized. In any society the kinds of abusive speech or 
behaviour that joking relatives may use are defined by custom. The rules 
that must be observed towards the wife's parents are similarly defined 
in detail, such as the Galla rule that a man must not drink milk from 
a cup that his wife's mother has used, while this does not apply to dadi, 
the intoxicating drink made from honey or from the fruit of the Borassus 
palm, because this drink is 'a thing of great kindness'. A very widespread 
rule is that forbidding the uttering of the personal name of an avoidance 
relative. Thus much of the behaviour imposed in these relationships 
must be described as symbolic behaviour and the rules are essentially 
similar to rules of etiquette. The acts and abstentions imposed by such 
rules are the conventionalized symbolic expression of the relative posi- 
tion of persons in a particular social relation or situation. 

t,,^The3^iew taken here is that the customs of 'avoidance' and 'joking' 
have the same general social function. The differentiating principle 
between them is that by which, as a general rule (to which, as we have 
seen, there are exceptions which require special explanation), behaviour 
towards relatives of the parents' generation should be respectful while 
towards relatives of one's own generation there is a nearer approach to 
equality and familiarity. But within one's own generation there is often 
a differentiation of senior and junior with a rule that the junior person 
must show respect to the senior. So in some societies the rules of 
avoidance are applied to the wife's elder sister as well as to the wife's 

' J. F. Cunningham, Uganda and Its Peoples, 1905, pp. 54, 331. 
^ A. R. Radcliffe-Brown, 'On Joking Relationships', Africa, xiii. 3, July 1940, 
pp. 195-210. 


mother, even where there is a joking relationship with the wife's younger 

In the building of social structures means must be provided for avoid- 
ing, Hmiting, controUing, or settHng conflicts. In the new structural 
situation resulting from marriage there are possibilities of conflict. 
While there is a union of the husband and wife the two families (in the 
sense of bodies of kin) remain separated, only linked together by their 
separate connexion with the new family that is coming into existence. 
It is the separateness of the two groups, together with the need of main- 
taining friendly relations between them, that has to provide the basis for 
their personal relations. 

^ ^ The 'jokin£^ relationship in its reciprocal form can be regarded as 

a kmd of friendliness expressed by a show of hostility. The mutual 
abusive behaviour would be simple hostility in other connexions, but 
the joking relatives are required not to take offence but to respond in the 
same way. The social separation of the man and his wife's relatives is ■ 
symbolically represented in the sham hostility, ruled by convention, 
and the friendliness is exhibited in the readiness not to take offence. This 
interpretation applies to other instances of the reciprocal joking relation- 
ship that have nothing to do with marriage." 

: -^he jo king relationship is clearly only appropriate between persons 

who m the general social structure can treat each other as equals, and 
this generally means persons of the same generation or those related 
as 'grandparent' and 'grandchild'. For the wife's parents and other 
relatives of that generation, and sometimes for the wife's elder sister, an 
attitude of respect is required. But it must be a different kind of respect 
from that which a man shows in some African tribes to his father and in 
others to his mother's brother as the person who is entitled to exercise 
authority over him. This respect is totally incompatible with any open 
show of hostility. In a man's relations with his wife's parents the social 
separation is symbolically expressed in conventional rules such as the 
avoidance of the utterance of their personal names or the Galla prohibi- 
tion against eating food cooked by the wife's mother. In the most extreme 
form of the custom there is complete avoidance of social contact with the 
wife's mother, to whom a man may never speak and whom he may never 
meet face to face, and there is sometimes a similar complete avoidance 
of the wife's father. 

, There might be a temptation to regard this avoidance as a form of 
hostility, since we tend to avoid persons with whom we do not get on 
well. This would be a mistake. Amongst the Australian aborigines there 
is complete avoidance of the wife's mother. When I asked a blackfellow 
why he had to avoid his mother-in-law his reply was: 'She is my best 

' A. R. Radcliffe-Brown, op. cit. ; R. E. Moreau, 'The Joking Relationship 
(Utani) in Tanganyika', Tanganyika Notes and Records, xii, 1941, pp. i-io, and 
'Joking Relationships in Tanganyika', Africa, xiv. 3, July 1944, pp. 386-400. 


friend in the world: she has given me my wife.' Though this may seem 
strange to our way of thinking, I think his answer was logical and 
adequate. What disturbs or breaks a friendship is a quarrel. You cannot 
quarrel with a person with whom you have no social contact, or with 
whopi your contacts are strictly limited and regulated by convention. 
/,.,^<A marriage produces a temporary disequilibrium situation. In the . 
small and close-knit groups with which we are here concerned any 
removal of a member results in disequilibrium. The event that most 
markedly produces this result is a death. But on a smaller scale the 
removal of a daughter by marriage is also a disturbance of equilibrium 
in her family. Moreover, the intrusion of a stranger into a group of kin 
is similarly a disturbance. Among the Nguni of South Africa the bride 
during the early period of her marriage has to give presents to and per- 
form services for the women of her husband's group and only after a 
lapse of time is she accepted as one of themselves. Any reconstruction 
of a disturbed equilibrium inevitably takes time — longer or shorter as 
the case may be. 

The establishment of a new equilibrium after a marriage requires that 
in certain types of kinship or family structure there is a need felt for 
emphasizing the separateness of the two connected families. There are 
many customs in which this is shown, but a single example must suffice. 
In the Nguni tribes the personal name that a woman has in her own 
family, as a daughter, may not be used by her husband's family, who have 
to provide her with a new name, which again will not be used by her own 
relatives. She is a different person in the two groups. _ 

The principal points of tension in the situation created by a marriage 
are between the wife and the husband's parents and between the husband 
and his wife's parents. In order to condense and simplify the argument 
we are considering only the latter, but we must remember that the 
former is equally important. The point of maximum tension seems to be 
between the wife's mother, who is the person most closely and intimately 
connected with the wife before the marriage, and the son-in-law to 
whom have been transferred control and sexual rights over her daughter. 
This is, of course, what lies behind the vulgar English jokes about the 
mother-in-law." The conventionally maintained 'distance' between 
son-in-law and wife's mother does have the effect of avoiding conflict 
between them. 

In Bantu languages the customs of avoidance are referred to by a word 
of which some of the various forms are ntloni, nthoni, hloni. Some 
writers translate this as 'shame'. Thus Torday and Joyce report that 

' After a lecture given more than thirty years ago in which this theory was 
explained, a member of the audience asked: 'Would it not be a good thing to 
introduce this custom (the avoidance of the wife's mother) amongst ourselves?' 
His question aroused a roar of laughter from the audience, which, I imagine, 
was what he aimed at. 


amongst the Huana a man may never enter the house of his parents- 
in-law, and if he meets them on a road he must turn aside into the bush 
to avoid them. 'Repeated inquiries as to the reason of this avoidance on 
the part of a man of his parents-in-law elicited the invariable reply "that 
he was ashamed"; to a further inquiry of what he was ashamed, the 
answer would be "of marrying their daughter". No other reason could 
be obtained.'^ 

This 'shame' has been interpreted by Westermarck and other writers 
as being specifically sexual shame, but it is something much more 
fundamental and general than that, and indeed the English word 'shame' 
is not adequate as a description. What is really referred to is a felt con-\ 
straint of one person in the presence of another, which limits his 
behaviour and keeps him at a distance from that other. Shyness is a 
similar phenomenon, and both shyness and shame are commonly asso- 
ciated with blushing. But it is to be noted that this constraint on a person 
in his relations with his wife's parents is not the spontaneous product of 
his own feelings, but is imposed upon him by social custom and the — 
rules of etiquette that give expression to it, just as is the relative absence 
of constraint that is exhibited in the joking relationship with a wife's 
brothers or sisters. 

Westermarck makes much of the fact that in Morocco all indecent 
talk is particularly prohibited in the presence of a man and his parent- 
in-law. But in a great many African tribes it is not only the father-in-law 
but also the father in whose presence one may not utter or listen to such 
talk. Obscene expressions are used in many societies as expressions of 
hostility, as in swearing. Their use is therefore often considered appro- 
priate in joking relationships, and since these are the polar contrary 
of the relationship to parents-in-law, it is easy to see the meaning of 
avoiding obscenity in the presence of these relatives. Freedom to refer 
to sexual topics is in general characteristic of a certain kind of social 
intimacy, and is directly contrary to the distant reserve that has to be 
observed towards the wife's parents. 

The argument could be supported by the examination of other features 
of the etiquette relating to parents-in-law, but only one can be mentioned 
here. It has been noted above that amongst the Galla a man may 
not eat food cooked by his wife's mother nor drink from a cup that she 
has used. In many other African tribes a man may not eat food from his 
wife's family, and where there is not complete avoidance of all contact 
a man may not eat food in the presence of his parent-in-law. Dr. Nadel, 
in his book on the tribes of the Nuba Hills, suggests that such customs 
also express, symbolically, sexual shame; though he has himself indi- 
cated their true significance by referring to rules that forbid members of 

diflPerent clans to eat meat or drink milk together, stating that these 

' E. Torday and T. Joyce, 'Notes on the Ethnography of the BaHuana*, 
J. Roy. Anthrop. Inst, xxxvi, 1906, p. 285. 


eating avoidances are the medium 'through which these tribes express 
' social proximity and distance'. The sharing of food, and still more, eating 
and drinking together (commensality), are all over the world expressions 
of social solidarity. The symbolism of customs of etiquette in the matter 
of eating is obvious in the customs themselves, and there is no need to 
search for obscure Freudian symbols by which eating is a sexual activ- 
ity. Amongst the Nguni of South Africa a bride may not drink the milk 
of her husband's kraal until after a lapse of time, usually not till after 
she has borne a child, and after a ceremony has been performed. This is 
the symbolic expression of the fact that she is no longer a stranger or 
outsider but a member of the group. 

Customs about the avoidance of names can be interpreted in the same 
way as customs about eating, and it has not yet been suggested that they 
have sexual significance. All these various rules, it is here held, are con- 
ventional rules by which a man is obliged to maintain social distance 
between himself and his wife's parents. By maintaining this kind of 
relationship any tensions that exist or may arise are prevented from 
breaking, by open conflict, what should be a friendly relation. 

In some societies in Africa and elsewhere, the rules of avoidance are 
somewhat relaxed in the course of time, i.e. as the marriage develops 
through the birth of children. This is, for example, reported of the 
Kamba.' It is probably true of many other societies from which it has 
not been reported. This is easy to understand if we think of the marriage 
as a developing process. Whatever tensions or dangers of conflict there 
may be between the son-in-law and his wife's parents are at a maximum 
in the early period of the marriage. The imposed 'shyness' between 
them is functionally most significant immediately after the marriage. 


A part of evg ry kinship syste m is a set of regulations concerning 
marriage between persons related by kinship or through marriage. There 
aTe, in the first _^lace, rules which prohibit marriage between persons 
who stand m certain relationships. An example is afforded by the list of 
prohibited degrees in the English Book of Common Prayer. The rules 
vary greatly from one system to another, and in a given society may vary 
from one period of its history to another. In many societies there is what 
is called a rule of exogamy, by which a man is forbidden to take a wife 
from amongst the women of his own group (lineage, sub-clan, clan, or 
moiety). On |Vie nth^r hjj^H in some systems there are certain relatives 
between whom marriage is not merely permitted but is regarded as 
desirable. The term 'preferential marriage' is commonly applied to 
customs of this kind. The commonest examples are cross-cousin marriage 
(marriage with the daughter of the mother's brother or of the father's 

' Lindblom, The Akamba, which gives a good account of the relations 
between athoni (avoidance relatives). 


sister) and marriage with the wife's sister or the wife's brother's 

There are also rules relating to sexual intercourse outside m arriage. 
Incest is the sin or crime of sexual intercourse between persons related 
either by kinship or through marriage within degrees defined by law or 
religion. Marriage and sexual intercourse outside marriage are not the 
same thing, and the rules relating to them must be separately considered. 
Most of the discussions about these rules have been vitiated by the 
failure to distinguish two distinct, though obviously related, problems. 
Here we are concerned primarily with the problem of the rules relating 
to marriage. 

There is, in most societies, a tendency to condemn sexual intercourse 
between persons who are forbidden to marry. But there are many 
instances in which a man and woman who may not marry may carry on 
a temporary affair without this being considered the grave offence to 
which we give the name of incest, and without being subjected to any 
legal or religious sanction. Amongst the Tallensi of West Africa there 
are women whom a man is prohibited from marrying but with whom 
intercourse is not regarded as incestuous; the Tallensi themselves say 
'copulation and marriage are not the same thing'. Similarly, amongst the 
Nkundo of the Belgian Congo there is a special term (lonkana) for sexual 
intercourse with women whom a man may not marry but with whom 
such connexion does not constitute incest; they are women of a clan 
(or lineage?) which is related to his own and into which he may not 
marry for this reason.' 

Confining our attention to the regulation of marriage, we can see that 
there are certain requirements thafmust be met by a theory if it is to be 
worthy oT aiiy'consideration. It must ofTer a general theory of the varia- 
tions in these rules in different societies, and it must therefore deal not 
only with prohibited but also with preferred marriages. It must give 
some significant clue towards an understanding of why any given society 
has the rules it does. The test of a scientific theory is in its application 
to the explanation of particular instances. By this criterion many of the 
speculative hypotheses that have been put forward are entirely useless 
and it would be a waste of time to discuss them. To anyone propounding 
a theory we might put the following question : How does the theory give 
us a clue to the understanding of why, amongst the Nkundo, a woman 
is forbidden to marry (in a second marriage) the husband's father's 
brother's son of her first husband's mother's brother's daughter; or 
why, amongst the Hera of Mashonaland, a man may not marry a womarj 
of the lineage of his wife's brother's wife, although he may marry his 
wife's sister or a woman of the lineage of his mother's brother's wife ? 

' M. Fortes, 'Kinship, Incest and Exogamy of the Northern Territories of 
the Gold Coast', in Custom is King, ed. Dudley Buxton, 1936; Hulstaert, 
op. cit. 


The th eory here p roposed is sim ply a sp eciaLapptteatieR o^f the g^ eneral 
theoryTHat the raison d'etre oFaii institution or custom is to be found in 
its~social function. 'I'he theory is, therefore, that the rules or customs 
relating to prohibited or preferred marriages have for their social func- 
tion to preserve, maintain, or continue an existing kinship structure as 
a system of institutional relations. Where a marriage between relatives 
would threaten to disrupt or throw into disorder the established system 
it tends to be disapproved or forbidden, and the greater and more 
widespread the disturbance that would be caused by a marriage, the 
stronger tends to be the disapproval which it meets with. Inversely, 
preferential marriages are those which have for their effect to renew or 
reinforce the existing system. 

Whit \~i hirrn rnllrd i trndrnry in "nnirthinfr t hat can be disc overed by 

observation. In some instances the objection to a certain type of marriage 
takes a quite definite form, as in a system of law. But in other instances 
there may be, in a given society at a particular time, a disagreement 
between individuals as to the desirability of encouraging, permitting, 
or prohibiting a certain type of marriage. This can be illustrated from 
English history. Until the Reformation marriage with a deceased wife's 
sister was forbidden by the canon law of the Roman Church. But persons 
of influence who could pay for the privilege could obtain a special dis- 
pensation permitting the marriage to take place. By an Act of King 
Henry VIII dispensations were abolished and marriage with a deceased 
wife's sister or with a deceased husband's brother was made illegal. An 
Act of Queen Mary legalized marriage with a deceased husband's brother 
but not with a deceased wife's sister. In 1835 Lord Lyndhurst brought 
forward in the House of Lords a Bill to legalize marriage with the 
deceased wife's sister; from that date till 1907, when an Act was finally 
passed permitting such marriages, there was a continued and passionate 
controversy. This can be studied in the debates on the subject that 
occurred at intervals in the Lords and Commons and were reported in 
Hansard. An association called the Marriage Law Defence Union was 
formed to prevent the passing into law of the proposal. Articles and 
pamphlets were printed on both sides of the controversy and of these 
a bibliography by Huth has 257 entries between the years 1840 and 
1887. After the law permitting such marriages was passed some clergy- 
men refused to solemnize them in church ; though permitted by the law 
of the State they were judged by some to be contrary to religion. 

This episode from the history of the English kinship system has been 
mentioned for several reasons. It illustrates the fact that while in a par- 
ticular society there may sometimes be unanimity of opinion as to the 
desirability or otherwise of a particular kind of marriage, there may 
sometimes exist a marked divergence of opinion. In England there was, 
and still perhaps is among some people, a very strong sentiment against 
marriage with the sister of a deceased wife, while others feel that such 


marriage is permissible or even desirable. Similar divergence of opinion 
can sometimes be noted in primitive societies and its existence calls for 
theoretical explanation and affords a good means of testing any general 

An examination of the documents of the controversy that lasted in 
England for seventy years shows that the objection to this kind of 
marriage is based on sentiment rather than on any sort of reason. It is 
just felt to be wrong. 

Nevertheless, it is possible, by analysis of the English kinship system, 
to see how this divergence of opinion and sentiment was possible. The 
decline in the social importance of kinship which has been taking place 
in England for several centuries, and is still continuing, resulted in 
a situation in which there was no clearly defined and generally accepted 
pattern of behaviour as between a man and the brothers and sisters of 
his wife. In the absence of any institutional norm this kind of relation- 
ship could be, and inevitably was, diflFerently treated by different per- 
sons. Since thought and sentiment may be very strongly influenced 
by words, the English term 'sister-in-law', understood as 'sister' by 
marriage, served to provide some persons with a pattern for the rela- 
tionship; the sister-in-law is a sort of sister. Marriage with any sort of 
'sister', implying sexual intimacy, was emotionally felt to be a sort of 
symbolic incest. Other persons felt that sister-in-law and brother-in-law 
are not really relatives at all. If any man, on the death of his wife, wishes 
to enter another marriage, the wife's sister should be just as eligible as 
any unrelated person. There was a third view: the relation between two 
sisters is, or should be, one of affection and intimacy; for a widower with 
young children there is no one who could so suitably replace his wife as 
her unmarried sister; on this view marriage with the wife's sister was 
viewed as a marriage to be preferred. 

From this well-documented historical instance, and from the parallel 
instance of the discussions in South Africa over the proposal to legalize 
marriage with a deceased husband's brother, which had previously been 
forbidden by Roman-Dutch law, we can draw a generalization. Where 
the kinship system is one that has a structure with a set of clearly defined 
institutional relationships there is likely to be complete agreement as to 
whether a particular kind of marriage should be prohibited, permitted, 
or preferred. Divergence of opinion or sentiment indicates an absence 
of rigidity, and that, of course, is not a bad thing in itself. A political 
system, such as that of England, may depend essentially on the differ- 
ences of opinion and sentiment of political parties. But in a system of law 
some rigidity is necessary, even though from time to time the law may 
be modified. A law works best when it is backed by a nearly unanimous 
public opinion. 

Xh ^e are significant differences in human societies in this matter of 
prohibiting, permitting or preferring marriage of a woman with her 


husband's brother, or of a man with his wife's sister. Polyandry, the 
marriage of a woman to two or more husbands, is an institution that is 
only rarely found, but its most usual form is adelphic polyandry, in 
which the woman marries two or more brothers. Corresponding to this 
there is the much more widespread institution of sororal polygyny, in 
which a man marries two or more sisters. Amongst the Australian 
aborigines this is regarded as the ideal form of marriage, and still more 
ideal is the arrangement by which an elder brother marries the two eldest 
sisters of a family and his younger brother marries one or two of the 
younger sisters. 

There are very widespre ad marriage customs that are commonly 
referred to by the terms levlrate^ an d 'sororateVjJbiit~we have to distin- 
guish'different institutions to which these terms are applied. In the true 
levirate, exemplified by the customs of the Hebrews, and in Africa by 
the Nuer and Zulu and many other peoples, when a man dies and his 
wife has not passed the age of child-bearing it is the duty of the man's 
brother to cohabit with the widow in order to raise children, which will 
be counted, not as his, but as children of the deceased. The widow 
remains the wife of the dead man, for whom the brother is a surrogate 
and thus not strictly speaking her husband. A different institution is 
widow inheritance, in which a brother takes over the position of husband 
and father to the widow and her children. 

There is a similar d istinction with regard to the sororate. In some 
tribes of South Africa, ^uctTas-the-^ttki^-il-a-wGman proves to be barren 
her kin will provide a sister to bear children who will be counted as 
children of the barren wife. This is parallel to the true levirate, A differ- 
ent custom is that by which when a wife dies her kin may supply a sister 
to replace her. In sororal polygyny a man, having married an older sister, 
also marries her younger sister. 

All these customs of preferential marriage can be seen to be continua- 
tions or renewals of the existing structure of social relations. All of them 
are also examples of the principle of the unity of the sibling group, since 
brother replaces brother and sister replaces or supplements sister. 

Professor Gluckman's paper in this volume gives an illuminating 
comparison between the Lozi and the Zulu in this matter of the sororate 
and sororal polygyny. The Zulu have both ; they approve of marriage 
with the wife's younger sister which reinforces the relationships estab- 
lished by the first marriage, and which, if sisters behave in a 'sisterly' 
way, increases the solidarity of the family group. The Lozi object to mar- 
riage with even a classificatory 'sister' of the wife, and say that the com.- 
petition and rivalry between co-wives is likely to destroy the relationship 
that ought to exist between sisters. But there is really more in the matter 
than this, and it may be held that the basis of this difference between the 
Zulu and the Lozi lies in certain very important differences of social 
structure. The unity of the sibling group, with its implication of 

is replaced by a sister, unless s 
child-bearing age. In the very 


substitution of brother for brother and sister for sister, is a major prin- 
ciple of the Zulu system and a relatively minor feature in that of the 
Lozi. Following this the Zulu system emphasizes the solidarity of the 
lineage group, and this can hardly be said to exist in Lozi. In sororal 
polygyny one woman of a sibling group as part of a lineage supplements 
another; in the sororate she replaces her. Her duties are imposed on her 
by her lineage affiliation. To quarrel with her sister or neglect her sister's 
children is not simply a neglect of her marital duties ; it is contrary to her 
obligations to her own closest kin. In the Zulu system a marriage estab- 
lishes a relation between a man and his brothers and the family of his 
wife, which should be permanent. Divorce is objected to because it is 
destructive of this permanence. If the man dies his wife passes to a 
brother. If the woman dies the aelation can only be fully continued if she 

le has borne children or is beyond the 
different Lozi system the levirate and 
sororate could not possibly have the functions they have in the Zulu 

Marring'' wit^H^^*^-^**^'^ ci'utf^r l^^n^thp whole, more frequently found 
in associalign_mtJh die_patrilineal lineageand what may be called father-' 
right marriage, and it is precisely in such circumstances that it functions 
most effectively to maintain or strengthen the relationships set up by 
a marriage. In societies with matrilineal institutions there are variations. 
Thus the Ashanti do not permit marriages of this kind, and the paper by 
Dr. Fortes enables us to see why. On the contrary, amongst the Bemba 
such marriages are approved, and the reason again lies in the social 
structure. By his first marriage a man becomes attached to the family of 
his wife, with whom, for at least some time, he must take up residence. 
Marriage with the wife's sister would strengthen this bond and introduce 
no new factor ; whereas if, in a second marriage he unites himself with 
a different family this must complicate and is likely to disturb the exist- 
ing system of relations. Theoretically one would expect that the Bemba 
should have given, at any rate in former times, a definite preference to 
marriage with the wife's sister. 

In a number of African tribes there is a custom by which a man is 
given his wife's brother's daughter as a wife. This is in a sense a variant 
of marriage with the younger sister of the wife. It exists in tribes in which 
the patrilineal lineage is a predominant feature of the social structure, 
and in such tribes a marriage of this sort renews by repetition the 
relationship set up by a first marriage between a man and the patrilineal 
lineage of his wife; he takes a second wife from the same lineage group, 
just as in the sororate he receives a second wife from the sibling group 
of the first wife. The second wife supplements or replaces, not her elder 
sister, but her father's sister. The structural principle involved is that of 
the unity of the lineage group. 

In some African societies a man is not permitted to marry the daughter 



of his mother's brother, and this rule is extended to the whole lineage to 
which she belongs, and sometimes to the clan. Other societies permit 
such marriages of cross-cousins, and in some they are given preference. 
(It must be noted that in a system of lineages or clans, whether patri- 
lineal or matrilineal, cross-cousins belong to different groups.) There 
is no space available here for a general discussion of the reasons for this 
variation, as the subject is complex. Where the institutionalized rela- 
tions of a man to his mother's brother (and his wife) are in important 
respects incompatible with his relations to his wife's father and mother, 
marriage with the mother's brother's daughter tends to be forbidden. 
Where there is no reason of this kind, then marriage of this sort renews 
in one generation a relation between families that was established in the 
preceding generation, and thus tends to be approved or preferred. A man 
takes a wife from a certain family or lineage and establishes a relation 
with her kinsfolk. His wife's brother becomes the mother's brother of 
his children ; if his son marries the daughter of this man there is a repeti- 
tion of the previous connexion. 

The Hehe of East Africa do not forbid marriages of this kind. Gordon 
Brown has reported the existence of a difference of opinion in this tribe. 
Some people think that marriage with the mother's brother's daughter 
is desirable because it renews the already established relation between 
the families, but others think that the tensions that are likely to arise 
between the kin of a man and those of his wife may disrupt the existing 
friendly relations that have resulted from a former successful marriage, 
and therefore think it better to avoid a marriage of this kind. By way of 
contrast it is apparently the unanimous opinion of the Lobedu of the 
Transvaal that a relationship established in one generation by a marriage 
should if possible be renewed or repeated in the next generation through 
marriage with the mother's brother's daughter. 

The paper on the Ashanti by Professor Fortes gives us further insight. 
The Ashanti formerly gave definite preference to marriage with the 
mother's brother's daughter, and had their own particular rationaliza- 
tions for the custom. This kind of marriage is becoming less frequent as 
a result of social changes, and there is now a divergence of opinion as to 
the desirability of what was apparently formerly accepted as an estab- 
lished custom. Recognizing the existence of a number of factors in the 
minds of the people themselves, Professor Fortes sees the change as 
one involved in the gradual transformation that is taking place in the 
social structure of the Ashanti people. 

The re; is a very important general difference in the regulation of 
marriage between societies that build their kinship system on cognatic 
relations traced equally through males and through females and those 
that adopt the unilineal principle. In a purely cognatic system, such as 
that of Anglo-Saxon England or ancient Wales, the prohibition against 
marriage applies to all cognates within a certain degree of kinship; 



marriage is forbidden between persons who have a common ancestor or 
ancestress within a certain number of generations. For example, two 
persons who have a common great-great-grandparent may be forbidden 
to marry. In a unilin eal system th e primary rule is that two jiersons may 
not marry i f they both belong toasocially recognized unilineal descent 
group. This"~iliay be a lineage, orlt may be a clan. A rule of this kind is 
called 'exogamy'. Perhaps the most extreme example is the Chinese rule, 
not always, I believe, observed in these days, that two persons having 
the same surname may not marry, since such names are patrilineally 
inherited and therefore the two persons of one name may be supposed 
to have had an ancestor in common, though it may be three thousand 
years ago. There is no special problem about exogamy. The exogamy 
of a clan is the same thing las the exogamy of a lineage, with a wider 
recognition of kinship. The essence of the system of clans is that a man 
is required to recognize all the members of his own clan as his kin and 
to behave to them accordingly. The rule of exogamy, where it exists, 
is a way of giving institutional recognition to this bond of kinship. Like 
the classificatory system of terminology, which is frequently found 
associated with clans, exogamy is part of the machinery for establishing 
and maintaining a wide-range kinship system. 

Jn unilineaj ^syst^'-nQ mgnatir l^inghip nntgiHp tha unilineal 'group may 
also b? r^r^g^'^^di T^ns in ancient Indian law a man might not marry 
a sapinda, a person descended patrilineally from one of his patrilineal 
ancestors within seven generations. He might also not marry certain 
cognatic kin, but the connexion had to be a nearer one, within five, or in 
another system of law within three generations. There is also such a 
thing as lineal-cognatic kinship. In a system of patrilineal lineages or 
clans the rule of exogamy forbids marriage within the group ; but there 
may also be a prohibition against marriage with a person of the mother's 
group. Inversely, in a system of matrilineal groups a man may be for- 
bidden to marry a woman of his father's group ; the relationship is what 
is here called lineal-cognatic. 

It is evident from these_^rem_arks. that in societies that make use of the 
umimeal^fmcipTe there is an immense variety in the rules relating to 
marriage, THel;ontrast between a cognatic system and a unilineal system 
is brought out in the comparison of the Lozi with the Zulu . Amongst the 
Lozi, with a cognatic kinship system, the regulation of marriage takes 
the form that marriage is forbidden between any two persons who are 
cognatically related within a certain degree ; for this purpose genealogical 
relationships are not traced farther back than the fourth generation, and 
in fact marriages do take place within these limits. The rule, therefore, 
even if it is not always observed, is that third cousins, descendants of 
one great-great-grandparent, should not marry. This seems to have been 
the rule at one time in England and Wales. 

The Zulu have a system of agnatic lineages, as did the ancient Romans, 


and marriage between agnates recognized as such within a certain range 
is forbidden. The Zulu also recognize lineal-cognatic kinship and forbid 
the marriage of a man with a woman of his mother's lineage. The regula- 
tion of marriage is very different in the Lozi and the Zulu, and the 
difference corresponds to a fundamental difference in social structure. 
/<^. In the construction of a kinship system it is necessary to fix in some 
way the range over which relationships are to be institutionally recog- 
ntSed. In this the rules relating to marriage may have great importance. 
The system of the Australian aborigines is one in which every person in 
the society is related to every other person with whom he has any social 
contact whatever. The regulation of marriage therefore takes the form 
of a rule that a man may only take a wife from some one category, or 
some categories, of kin. Fndoj^nnwis^n nilf^ onft of thf fiinrtionq of which, 
at aj iy_jate in some parts ol InJia, is to circumscribe the range of rela- 
tionships, since a member of an endogamous group cannot possible be 
related, by kinship or through marriage, with any person outside the 
group. In a cognatic system the range of relationships depends on how 
far any person traces his genealogical connexions. In the Teutonic 
system of the sib, theoretically, any descendant of any of the sixty-four 
pairs of the grandparents of one's great-great-grandparents was a sib- 
kinsman, thus including all cousins up to sixth cousins. It would seem 
quite impossible that anyone would recognize as kinsmen all of these ; it 
was a theoretical construction of the lawyers, not something used in 
daily practice. It illustrates the fact that a wide-range cognatic system is 
not very practicable. In setting up rules for marriage, therefore, a cog- 
natic system, such as the Lozi, rarely, if ever, goes beyond third cousins, 
descendants of a common great-great-grandparent. 

Unilineal systems have to work in quite a different way. Unilineal kin 
groups normally tend to increase in size in successive generations. A clan 
may grow into a group numbering many hundreds ; but membership of 
the same clan in the case of dispersed clans may only be of significance 
amongst persons in regular and frequent social contact. Lineages also 
tend to expand in volume, and in lineage systems we find some procedure 
by which a lineage that has grown to a size in which the institutions that 
maintain its unity do not function well can be divided into two or more 
separate but still connected lineages. A method by which this is some- 
times brought about in the Nguni tribes of South Africa is interesting 
as illustrating the thesis of this section. It may happen that when 
a lineage group has grown to a considerable size a young man may 
decide that he wants to marry a girl of the lineage who is not closely 
related to him. The natives themselves say that a marriage within the 
lineage is disruptive of its unity, since it would create within the group 
relationships by marriage which are entirely incompatible with the 
established lineage relations. There will therefore be resistance to the 
proposed marriage. But if there is sufficient opinion in favour of it 


the lineage can be divided into two separate connected lineages between 
which marriage becomes possible. Even when a lineage has become too 
large to be functionally fully effective the Nguni are inclined to try to 
maintain its unity, and therefore tend to wait for such an occasion as 
the one here described. Marriage within the lineage group would be 
thoroughly disruptive of its structure, and therefore cannot be permitted ; 
but if the structure is changed by fission of the group the marriage is 
permissible between the two newly created groups. This illustrates 
the relation between rules as to marriage and the kinship structure. 

A brief reference must be made to the Tswana, whose system is 
described in this volume by Professor Schapera. The Tswana are 
decidedly exceptional in Africa, and might almost be regarded as an 
anomaly ; but in the comparative study of social systems exceptions and 
apparent anomalies are of great theoretical importance. There is a strong 
contrast between the Tswana and the Nguni tribes. The Tswana recog- 
nize patrilineal lineage ; they seem to have had a preference for marriage 
with mother's brother's daughter, which is characteristic of a number of 
tribes with whom they are ethnically related. But they also permit 
marriage within the lineage to the daughter of a father's brother. This 
would be impossible amongst the Nguni, who indeed refer to the Sotho, 
related to the Tswana, with expressions of disgust as 'those people who 
wear breeches and marry their sisters.' How the Tswana arrived at their 
present system is an historical question about which we can unfortunately 
only speculate. But the way the system works can be studied. Amongst 
the Nguni the way a man is required to behave towards the relatives of 
his wife is entirely incompatible with the way in which he behaves to 
persons of his own patrilineal lineage, his father's brother, with the wife 
and children of the latter. The absence of such incompatibility amongst 
the Tswana, while it makes possible the marriage with father's brother's 
daughter, also marks their system as being of an unusual type amongst 
indigenous African peoples. The Arabic peoples also practise marriage 
with the father's brother's daughter, but their kinship system is in many 
respects very different from that of the Tswana or those of African 
peoples in general. 

It is only possible he re to deal very b riefly with the subject of incest 

in itk jela'tion 16 t lieTegnlatinn-nf the marriagp nf kin Incest is properly 

speaking the sin or crime of sexual intimacy between immediate relatives 
within the family, father and daughter, mother and son, brother and 
sister. In human societies generally such conduct is regarded as unthink- 
able, something that could not possibly occur, and the idea of it arouses 
a strong emotional reaction of repugnance, disgust, or horror. It is 
characteristically conceived as an 'unnatural' action, contrary not so 
much to law and morals as to human nature itself. It is this emotional 
reaction that we have to explain if we are to have a theory of incest. 
Another example of a kind of action frequently regarded as 'unnatural' 


is parricide, the killing of a father or mother. The parallel between incest 
and parricide is illustrated in Greek drama. 

Almost everywhere in h unaaH-SOcieties as..^ e know them the first 
experience that any person has of society is in_ the-£argntal jamily^tbLe 
intimate domestic group of father, mother, and children .""^Certain 
emotional attitudes are developed in such a group with sufficient force 
to come to be thought of as 'natural' in the sense of being part of human 
nature itself. The kind of emotional attitude existing in sexual intimacy, 
and the kinds of emotional attitude developed in the family towards the 
nearest kin, are felt to be violently contrary, incapable of being combined 
or reconciled. This is a matter of the logic of sentiments, not the logic 
of reason, and this is what is really meant when writers say that the 
repugnance to incest is instinctive, for there is a certain logic of the 
emotions which is the same in all human beings and is therefore inborn, 
not acquired. Individuals who behave contrary to this logic of sentiment, 
as by the murder of a mother, are behaving 'unnaturally'. 

The study of what are regarded as 'unnatural' offences, incest, 
bestiality, in some societies homosexuality, patricide, and matricide, is 
a special branch of the comparative study of morals. One offence that is 
frequently thought of as 'unnatural' is witchcraft in the sense of working 
evil on members of one's own social group. (Black magic used against 
one's enemies in other groups, as amongst the Australian aborigines, is 
an altogether different matter.) In Africa incest and witchcraft are often 
thought of as connected. A South African native with whom I was dis- 
cussing the subject remarked about sexual intimacy with a sister, with 
horror in his voice, 'That would be witchcraft.' There is a widespread 
belief in Africa that a man can obtain the greatest possible power as a 
sorcerer by incestuous intercourse with his mother or sister. Intercourse 
with a more distant relative would be quite ineffective. 

In Europe in Christian times incest, bestiality, homosexuality, witch- 
craft, as 'unnatural' offences were quite logically regarded as offences 
against the Creator, and therefore the concern of the Church. In England 
it is only recently that incest has been treated as a crime to be dealt with 
by the secular courts. In many primitive societies it is thought that incest 
will be punished by supernatural sanctions. These points are alj signifi- 
cant for an understanding of the attitude towards incest. The_fomil^_is 
n ormally reg arded as__ SQm''<'^'''''g gRrrg d; incest, like patrigide-of-matfi- 
cice, is sacrilege. 

'l he attitude towards incest, in the narrow sense, may be extended to 
sexual intimacy between other relatives, but in different ways in differ- 
ent societies. In some primitive societies sexual intimacy with the wife's 
mother is felt to be not less evil than with one's own mother. But there 
are other societies in which a man may be married at one time to a woman 
and to her daughter by another husband. Sexual intercourse between 
husband and wife is an element in the institutional complex of marriage. 


Every society therefore makes a distinction between this and sexual 
intimacy outside marriage. There is a tendency to make a distinction 
between persons between whom marriage is prohibited and other 
persons, a tendency that is much more powerful in some societies or 
in some instances than in others. There is very great variation in such 

The categories of relationship recognized in classificatory terminolo- 
gies undoubtedly have considerable effect. If a woman is to be called 
'mother' or 'sister' or 'daughter' the relationship with her is thought of 
as being similar to that with one's own mother or sister or daughter, and 
it is obvious that this will very frequently be felt to forbid any sexual 
intimacy. It would be a sort of 'symbolic' incest and as such objection- 
able ; to have a sexual relation with a classificatory 'mother' is a symbolic 
offence against one's own mother. Such symbolic incest, except ritually 
on specific occasions, is strongly reprobated amongst the Australian 

So me w riters assume, or seem to assume, that prohibitions against 
rnarnagejre the~re5Tihrof feelings that sexual intercourse between the 
persons concerned would b e wrong. The J:ruth, in the majority of 
instances, is the reve ree_ot£_this; sexual intercourse is felt to be wrong 
betwe6ir^o_personsif by the rules of society they may not marry. It is 
extremely rarely, however, that the reaction to such conduct, or to the 
idea of it, is of the same kind and intensity as reaction to the idea of real 
incest with immediate relatives. This is exemplified in the laws of 
modem states where sexual intimacy between relatives who are by law 
forbidden to marry is not in all cases punishable as a crime. 

Theoretical discussions on the subject of incest and the regulation of 
marriage are often, one might even say usually, full of confusions. The 
theory outlined here attempts to get rid of these, (i) Incest is the sin or 
crime of sexual intercourse between members of a parental family. It is 
not a question of prohibition of marriage. There are societies in which 
intercourse between brother and sister is incestuous, but kings or chiefs 
may marry, or may even be expected to marry, their own sisters. The 
condemnation of incest is based on the emotionally or 'instinctively' felt 
violent incompatibility between sexual intimacy and family relations of 
affection and respect. This may be rationalized in different ways. (2) The 
rules relating to marriages between related persons in any society are an 
intrinsic part of the complex of institutions that make up a kinship 
system. In each instance the rules can only be understood by reference 
to the system to which they belong and the social structures that consti- 
tute the basis of the system. The general law which each instance exem- 
plifies is that the rules have for their function to maintain the continuity 
of the general system of institutional relationships, either by preventing 
marriages which would be disruptive or by encouraging marriages which 
reinforce the existing arrangement of persons. (3) The strong feelings 


about incest, by a readily recognizable psychological process, tend to 
spread, so that sexual intimacy between persons who may not marry 
tends to be regarded as being somewhat of the same kind as incest, 
though differing in the degree with which it meets with condemnation 
or reprobation. 

The whole theory is therefore one of social structure and of the neces- 
sary conditions of its stability and continuity. Incest, as here defined, is 
not merely disruptive of the social life of a single family, it is disruptive 
of the whole system of moral and religious sentiments on which the 
social order rests. Prohibited marriages are for the most part simply 
those which would prevent the continuance of normal relations between 
the few persons who would be immediately affected. There are, however, 
instances in which a marriage between kin, not necessarily closely related, 
is felt to be an attack on the whole social order ; this is so in the kinship 
systems of Australian aborigines. Such a marriage, if attempted, is a sort 
of crime against society and is likely to be treated as such. 


One of the most famous pseudo-historical speculations of the 
anthropology of the last century was the idea that the earliest form of 
society was one based on 'matriarchy' or 'mother-right'. One definition 
of this, given in the Encyclopaedia Britannica of 1910, is 'a term used to 
express a supposed earliest and lowest form of family life, typical of 
primitive societies, in which the promiscuous relations of the sexes 
result in the child's father being unknown'. An alternative definition, 
frequently used, was a social condition in which kinship is reckoned 
through females only, and in which there would be no recognition of any 
social relationship of fatherhood. We have no knowledge of any societies 
of this kind in the present or in the past; it is, as Robertson remarked in 
his History of America in the eighteenth century, a pure product of 

But early anthropologists also applied the term 'mother-right' to 
certain existing societies, McLennan to the Nayars of southern India, and 
Tylor to the Menangkabau Malays. We may take these two societies as 
providing us with a special type of system to which we may continue to 
apply the term 'mother-right', and we may add to them the Khasi of 
Assam. It is to be noted that the Nayars and the Malays, so far from 
being 'primitive', are advanced, literate, and cultivated peoples. The 
Nayars are a military, ruling, and land-owning aristocracy in a civilized 
community; they esteem learning and the arts and have produced an 
extensive, and according to accounts, admirable, literature in their own 
language, Malayalam. The Malays similarly have an extensive literature 
which is admired by those who know it. Though the Khasi are less 
advanced they are very far from being savages. Thus the typical instances 


of mother-right are found, not amongst the more primitive peoples, but 
in advanced or relatively advanced societies. 

The Nayars are sometimes referred to as a caste, but they are a numer- 
ous people divided into a number of subdivisions, 130 such being 
enumerated in the census of 190 1. There are undoubtedly differences of 
custom in different sections of this large community. Their system of 
marriage and kinship has been undergoing change during the past 
seventy years, and with regard to some of its features there is a lack of 
agreement in different accounts. I believe that what follows is a sub- 
stantially accurate generalized account of the system as it formerly 
existed. It is put in the past tense because it is not an accurate descrip- 
tion of the conditions of the present day.' 

Xhe i mportant unit oX .structure-iQ,the_Nayar kinship system was a 
matrilineal _ lineage gr oup called-farafarf. It consisted of all the descen- 
dants th rough fema le? of a gjrglp known an££stLPss^ and might number 
mor e tha n a hundred persons. It is spoken of in Indian law as a 'joint 
family', since the pfoperEyT^SToitrtlyTJwned by the group as a corporate 
unity; but this is not a family in the ordinary sense, since it includes 
women and their daughters and sons and brothers, but not their 
husbands. All sons of the taravad inherited from their mothers the right 
to share during their lives in the produce of the land or other property, 
but could transmit no rights to their children. The control of the 
property was in the hands of the karanavan (known in legal terminology 
as the 'manager'), who was normally the oldest male member. If a 
taravad became too large, or for other reasons, it might be divided. If 
the first ancestress had three daughters, all of whom had left descendants, 
three new lineage groups would be formed, the property being divided 
into three equal portions. The partition of property did not destroy the 
relationship, and kinship therefore extended beyond the taravad. 
A number of related taravad formed a group for which the name in 
north Malabar is kulani ; we may call this a clan, though it may have been 
a very large lineage. The clan was exogamous, and the various lineages 
of one clan shared pollution, a birth or death in one lineage group caus- 
ing pollution in the other groups of the clan. A number of clans consti- 
tuted a sub-caste, and in south India the sub-caste is normally the 
endogamous group. Within the sub-caste it would seem that a taravad 
had a special relationship with one or more other lineages, not belonging 
to the same clan, whose members were their enatigar (allies). This was 
an important feature of the kinship system. The Nayars had a cult of the 

' There have been many accounts of the Nayars from the early fifteenth 
century to the present day. The Report of the Malabar Marriage Commission of 
1894 is important in relation to Nayar marriage. More easily accessible is the 
article 'Nayars' in Hastings's Encyclopaedia of Religion and Ethics and an article 
by K. M. Panikkar on 'Some Aspects of Nayar Life' m J. Roy. Anthrop. Inst, 
xlviii, 1918, pp. 254-93. 


ancestors or deceased members of the lineage ; once a year an offering of 
food and drink was made to the ancestral spirits, including the mother's 
brothers and mother's mother's brothers of the living members.^ 

Before a Nayar girl reached puberty she had to undergo a ceremony of 
which the essential feature was the tying on of an ornament, usually of 
gold, called a tali. A number of girls of one taravad might pass through 
the ceremony together provided that they were all of the same genera- 
tion. The tali was tied by a man invited for the purpose, called the 
manavalan, selected from the enangar of the taravad. There is some 
evidence that, at least in former times, the ceremony included the ritual 
defloration of the girl. There are two interpretations of the ceremony. 
One is that it is a sort of religious marriage which was dissolved at the 
end of the ceremony. The rite of tying the tali is a marriage or betrothal 
rite amongst some of the peoples of south India. At the end of the Nayar 
ceremony, which lasted four days, a cloth was severed in two parts and 
one part given to the manavalan and the other to the girl ; this is a rite 
commonly used in south India as a rite of divorce. When the manavalan 
died the girl or woman on whom he had tied the tali had to observe for- 
malities of mourning. Dr. Aiyappan, on the other hand, holds that 
amongst the Nayars the tying of the tali is nothing more than a rite of 
initiation into womanhood.^ For present purposes it does not matter 
which view we accept, and indeed it is largely a matter of the choice of 
words. The manavalan does not become the real effective husband of the 
girl on whom he ties the tali. 

When a Nayar young woman reached a suitable age she formed a 
union which will here be called a marriage. The most usual union was 
with a man of her own sub-caste, who must be of the same generation 
as the woman and must be older. It seems that marriage with an 
enangan was regarded as specially suitable. While the evidence is imper- 
fect, what there is points to the members of the allied taravad of the 
girl's own generation being all her cross-cousins; so that the Nayar 
system, like so many others in Dravidian India, would be based on 
cross-cousin marriage; the woman would take a husband preferably 
from the taravad of her father's sister. These marriages were referred 
to by the Sanskrit term sambandham, and the Nayars did not apply to 
them the usual Malayalam term for marriage. The Sanskrit bandhu 
refers to friendship, and may sometimes be used for the relation between 
a woman and her lover. The rite of marriage was of a simple kind, the 
essential feature being a gift of clothing (pudaka) from the man to the 
woman, which is a rite of marriage observed in some other castes of 
south India. 

' See V. K. Raman Manon, 'Ancestor Worship among the Nayars', Man, 
XX, 1920, p. 25. 

^ Dr. A. Aiyappan, 'The Meaning of the TaH Rite', Bulletin of the Rama 
Varna Research Institute, ix, part ii, July 1941, pp. 68-83. 


An important feature of the Nayar system was that a woman might 
form a sambandham union with a man of the caste of Nambutiri Brah- 
mans. That caste had a patriHneal system in which only the oldest son 
of a family might marry a woman of his own caste and bring her home 
with him to raise up children. The children born of a union of a 
Nambutiri man and a Nayar woman would have no legal position in the 
family of the father. 

The sambandham husband did not take his wife to live with him, but 
visited her in her own home. He did not have any legal rights over her 
or her property or over any children that resulted from the union. It 
seems that the wife could divorce her husband by asking him to discon- 
tinue his visits. It would appear that he was expected to offer her gifts 
from time to time. T he^ayars hav e_be en often quoted as a people 
practising polyandry, but on this subject there is a differen ce of opinion 
amongst recent write rs.^ jt seems clear from the historical evide nce that 
in formfijil imes, at any rate amongst some of the Nayars, a woman might 
form a union with two or more men who might be simultaneously her 
sambandham', th"e~Nayar marriage did not give a man exclusive right to 
sexual relations with the woman. 

There has been discussion whether the Nayar sambandham union is or 
is not marriage; this obviously depends on how the word is defined. The 
Malabar Marriage Commission of 1894 decided that it did not consti- 
tute a legal marriage according to systems of law accepted in India. We 
may say that it was a marriage in the sense of a socially recognized union 
which always had some permanence and might continue through a life- 
time. It was, however, a union based not on legal bonds but on ties of 
personal affection. 

We can use the Nayars to define a certain type of kinship organization 
to which the term 'mother-right' may be suitably applied. The system 
of the Menangkabau Malays was similar in essentials, and so was that of 
one section of the Khasis. The domestic group (the joint family of Indian 
legal terminology) is a matrilineal lineage of fewer or more generations. 
The parental family, the domestic group of man and wife with their 
young children living together as a household, does not exist. Jural 
relationships are practically confined to persons connected in the female 
line; so also are those religious relationships that are constituted by 
ancestor worship. All important property passes down in the matrilineal 
lineage, and only members of the lineage have claims over it; inversely 
the most important duties of a person are towards his or her lineage 
group and its members. 

In the purest form of the system a man does not acquire by marriage 
any rights of possession over his wife and her children. His relationship 
with them contains no jural element, but is one of mutual affection. He 

' See, for example, the correspondence in Man, IVIarch, April, October, 
November, and December 1932 and August 1934, 


gives his wife gifts, but cannot transfer to her or to his children property 
which belongs to his lineage or over which the lineage members have 
a claim. There are some variations. Amongst the Khasi a man did have 
certain rights over his wife since adultery was severely punished, and in 
this system there was an elaborate religious ceremony for marriage. The 
Khasi father occupies a position of respect and is revered by his children 
after his death. A widow may keep her deceased husband's bones for 
a time (thus keeping his spirit with her), but sooner or later the bones 
must go back to the man's lineage and clan. But yet divorce is common 
and may occur for a variety of reasons. 

Mothe r^rigb^- '" ^ontra.ste d with 'father-right' and the peop le chosen 
by anthropologists as aff ording a typical example of the latter were the 
ancient Romans, or ratlier one form of family that existed in Rome, 
characterized by manus mariii and patria potestas . There was a structure 
of patrilineal clans (gentes) and lineages. The father-right type of marriage 
must be distinguished from what is sometimes called 'free marriage' or 
marriage sine manu, which also existed "in Rome, and in which a woman 
retained her connexion with her own family. In the father-right marriage 
possessive rights over a woman were ceremonially transferred from her 
father or guardian to her husband, and in the coemptio marriage a pay- 
ment was made to the woman's family in consideration of this transfer 
of rights. The wife thus passed under the power and authority of her 
husband; she transferred her allegiance from her own household deities 
and ancestral spirits to those of her husband. The father, the pater- 
familias, had exclusive possessive rights over his children; they were 
under his power, \ht patria potestas . Over his sons he had the power of 
life and death, and might sell a son — the jus vitae et necis et vendendi. 
With this right the law or the mother's relatives could not interfere, but 
in his exercise of his power the father had to obser\^e the duties of reli- 
gion, and for an abuse of power might receive censure from the members 
of his gens, or at one period perhaps from the Censors. 

While in one se nse mother^ight an d father-right a re opposite types 
of systg m, the re is another sense in which they are only contrasting 
varieties of a smgle type. What they have mcommon is the extreme 
emptiasis'gn the lineage, m atrilineal or patrilineal, and they both contrast 
strongly w^ith systems based mainly on cognatic kinship; in both, the 
jural relations in kinship are rigidly confined to one lineage and clan. 
Possessive rights over the children belong entirely to the lineage, and 
inversely it is within the lineage that the individual has his most impor- 
tant duties and also his most significant rights, such as the right of 
support and rights of inheritance over property. In religion also 
(remembering that the Romans had patrilineal ancestor worship) it is the 
lineage or clan ancestors to whom one owes religious duties, and from 
whom one may ask for succour. The institutional complex of which 
mother-right and father-right are contrasting forms is thus one that can 


hardly make its appearance except at a relatively high stage of social 
development, where property and its transmission have become impor- 
tant, and where social continuity has come to be based on lineage. 

We have to ask, therefore, what is the differentiating principle between 
thcTwo ^ontrastiYi g formo o f lIil Luiupkx. -i^-j iag gofpfti ^i^^J],^'"" sup- 
pos^ fKat mother- right is the resuft of emphasis on the social bond 
between rriother and^hild, but an examination of evidence gives no 
support whatever to this view. In primitive societies, whether they have 
matrilineal or patrilineal institutions, it is normally recognized that the 
closest of all kinship bonds is that between mother and child. Even in 
societies that incline to father-right it is felt that a child is of the same 
flesh and blood as its mother. So children born from one womb have 
the same flesh and blood, and in the patrilineal tribes of Africa children 
of the same mother are much more closely connected than children of 
the same father by different mothers. As Gluckman, following Evans- 
Pritchard, points out, in strongly patrilineal societies such as the Nuer 
and Zulu it is through the mother that the social position of a child is 
determined.' We have only to consider the position of the materfamilias 
in the Roman family to realize that under father-right the most intimate 
personal bond is with the mother. 

The contrast between father-right and mother-right is one of two 
types of marriage. A woman is by birth a member of a sibling group; 
strong social bonds unite her to her brothers and sisters. By marriage 
she enters into some sort of relation with her husband. To provide 
a stable structure there has to be some sort of institutional accommoda- 
tion of the possibly conflicting claims and loyalties, as between a 
woman's husband and her brothers and sisters. There are possible two 
extreme and opposite solutions, those of father-right and mother-right, 
and an indefinite number of compromises. 

In the solution provided by mother-right the sibling group is taken as 
the most important and permanent unit in social structure. Brothers and 
sisters remain united, sharing their property, and living together in one 
domestic group. In marriage the group retains complete possession of 
a woman; her husband acquires no legal rights at all or a bare minimum; 
but at the same time he has few duties towards her or her group. Rights 
of possession over children therefore rest with the mother and her 
brothers and sisters. It is these persons to whom the child must go for 
every kind of aid and comfort, and it is they who are entitled to exercise 
control or discipline over the child. 

It must be emphasized that this is a matter of jural (including legal) 
relations only, and these do not constitute the whole of social life. We 
may revert to the theory of Aristotle that the two chief factors on which 
social harmony depends are justice and philotes. Justice corresponds to 
what are here being called jural elements in social relationships. We may 

■ See p. 185. 


take the other term as applying to all personal relationships of attach- 
ment and affection. A system of mother-right, in which a father has no, 
or almost no, legal rights over or legal duties towards his children, does 
not debar, but possibly encourages, mutual affection; for affectionate 
attachment can perhaps flourish best where there is a minimum of the 
kind of constraint that may result from the obligations of a jural rela- 
tionship. Certainly there is evidence of frequent instances of strong and 
lasting affection between a man and his wife and children under a system 
of mother-right. The system simply separates out the jural relations, 
which are confined within the lineage, and the personal relations of 
affection, esteem, and attachment. 

The solution offered by father-right is opposite. Possession of a 
woman, and therefore of the children of her body, are surrendered by 
marriage to her husband and his kin. The Roman husband acquires 
manus over his wife and her children fall under his potestas. The 
mother's kin, her brothers and sisters, in this kind of marriage, have no 
rights over the children, who, in turn, have no rights over them. The 
jural bonds between a woman and her siblings are severed by her 
marriage. But this leaves open the possibility of relations of personal 
affection. It is characteristic of systems that approximate to father-right 
that the mother's brothers and sisters are expected to extend to her 
children affectionate care and friendly indulgence, and that the sister's 
child is expected to exhibit affection towards the maternal uncles and 
aunts. Once again there is a separation made between jural relations and 
relations of personal attachment. 

Mother-right- as repr c sc nted^y the Nayars, and father-right as repre- 
"SentedTByjoiie form of marriage in Rome, give us useful points of refer- 
ence for an attempt to establish a systematic typology of kinship systems. 
In both these types the structure is one in which legal and jural relation- 
ships are as nearly as may be possible limited to the lineage and its 
connexions. The contrary of this is to be found in what are here called 
cognatic systems, in which jural relations are based on cognatic kinship 
traced equally through males and females. Such systems, or close 
approximations, are found in some primitive societies such as the Anda- 
man Islands, and in advanced societies such as Anglo-Saxon and modern 

In Africa a system that approximates fairly closely to the ideal type 
of a cognatic system is that of the Lozi, in which there is a minimal 
emphasis on unilineal kinship, so that lineages can hardly be said to 
exist as features of social structure. Professor Gluckman has brought out 
the marked contrast there is between the Lozi system and that of the 
Zulus which is a very close approximation to father- right.' The Lozi 
system also contrasts with one of mother-right. 

Relations of kinship involving rights and duties may also be traced 
' See below, pp. i66 ff. 


through both male and female lines in a double lineage system in which 
the structure includes both patrilineal and matrilineal lineages or clans. 
An example of this type of system is provided in this volume by the 
paper on the Yako by Professor Daryll Forde. Every individual has 
a well-defined set of relationships within his patrilineal lineage and clan, 
and another within his matrilineal lineage and clan. 

Thii^j^^^^jTavp^FniTj typpg Q£_aystgr!2g_{idjp^l types based ^n empirical 
ex ampi eait^!gn?g us a framework within which to construct a typology : 
father -right, m other-right, purely cognatic systems, and double lineage 

In Africa the nearest approach to pure mother-right is the system of 
Ashanti, of which an analytical description is given in this volume by Pro- 
fessor Fortes.^ The system is undergoing modification under European 
influences of various kinds, but still retains some of its former features. 
What has in this discussion been held to be the basic structural feature 
of mother-right, the close and continued solidarity of the sibling group 
of brothers and sisters of one flesh and blood, is illustrated in Professor 
Fortes's paper. It is true that the parental family exists, but it is not the 
standardized form of the domestic group. The importance of lineage 
transmission of property is well illustrated, and also the religious bonds 
that unite persons of common descent in the female line. 

What can be called 'qualified' systems approximating not very closely 
to mother-right are characteristic of some parts of central Africa and 
have been compared in the paper by Dr. Audrey Richards. In the system 
of the lower Congo, the type A of Dr. Richards, there is an interesting 
compromise in the form of a division of rights. The male members of 
a matrilineal lineage continue to live together in a corporate group 
owning property and having their own religious cult. But the group 
surrenders certain rights over its female members to her husband when 
she miarries. She retains a significant connexion with the group of her 
brothers, but does not live with them. The domestic unit is the parental 
family, man, wife, and young children. Boys, when they reach a certain 
age, leave the parental family to join the group which consists of the 
male members of the lineage. The system is one of a division of rights; 
the father has rights over his wife and his young children within the 
parental family, and of course corresponding duties. But his sons 
ultimately fall under the power and authority of the brothers of their 

The system of certain tribes of Rhodesia and Nyasaland is a compro- 
mise formation in which the division of rights does not seem to be clearly 
defined, so that there are variations in practice not only from one tribe to 
another, but also within a tribe. This is connected with the local struc- 
ture of this region, with its marked mobility by which persons move from 
one village to another, or establish new villages, so that the personnel 
' See below, pp. 252 fF. 


of a particular village varies from time to time. The structural principle 
of mother-right appears in a contrary form to that of the Congo tribes, 
in the tendency for the group of sisters to continue living together, 
at any rate for some time. This is illustrated by the enlarged domestic 
group of the Bemba, consisting of a man and his wife and their married 
daughters with their husbands and children ; the group breaks up when 
a man with his daughters forms a new domestic group of the same kind. 

Both Professor Fortes and Dr. Richards draw attention to the existence 
of tensions and strains in the kinship systems of the Ashanti and the 
Bemba. But tensions and strains and possibilities of conflict exist in any 
system of rights and duties. The constraint of social obligations may often 
be felt as irksome. There is an unfortunate tendency for human beings in 
some circumstances to insist on their rights rather than to be punctilious 
in performance of duty. But it is obvious that a system based on com- 
promise or on successive compromises is more likely to reveal tensions 
or conflicts than one in which jural relations are clearly defined and 
socially accepted. There is no reason why a system of mother-right 
should present more difficulties for individual adjustment than a system 
of father-right. But a system like the Bemba, with its division of rights 
and its occasions of rearrangement of structure, must obviously depend 
on the way in which individuals make personal adjustments with each 

The Ila appear to have a mixed system which is farther removed from 
mother-right than it is from father-right, in spite of the existence of 
matrilineal clans and some other matrilineal features. 

The Nguni peoples of South Africa, represented in this volume by 
the Swazi and the Zulu, may be described as having father-right. 
Possession of children is determined by the marriage payment. 'Cattle 
beget children' and 'The children are where the cattle are not' is the way 
the people themselves express this. But during their infancy the children 
belong to their mother. This is symbolically expressed in the custom by 
which she may protect them from sickness by making for them necklaces 
of hairs from the tail of her ubulunga cow, which belongs to the ances- 
tral herd of her own lineage, and which she takes with her on marriage, 
so that during the first period of her marriage she can drink the milk of 
her own lineage cattle. The ubulunga beast, cow for a woman and bull 
for a man, is a link between the individual and the sacra of his or her 
lineage, the cattle, the kraal, and the ancestral spirits. A woman after her 
marriage is entitled to the protection of the gods of her own family, and 
so also are her infant children who are attached to her more than to their 
father. It is at adolescence that a boy or girl becomes fully incorporated 
in the father's lineage. There is something similar in the Ashanti system, 
and still more definitely in the Congo systems, in the way in which a 
boy's relation with his mother's brothers becomes the preponderant 
fact in his life after adolescence. 


This section illustrates the method of typological analysis applied to 
institutions of kinship. The procedure is to select certain types which can 
be used as standards with which to compare others. For the type of 
mother-right it was necessary to go outside Africa, since even the Ashanti 
have only a qualified system of mother-right and, moreover, the Nayar 
and Menangkabau systems had been selected by anthropologists of the 
last century as best representing actual mother-right or matriarchy. It 
has also long been customary to take the ancient Romans as an example 
of father-right or patriarchy, though if we want an African example it 
might be possible to take the Zulu. It has been argued that the major 
structural principle of both father-right and mother-right is the maxi- 
mum emphasis on the lineage as the source of jural and legal relations. 
The opposite type is therefore that of cognatic systems in which lineage 
has very little or no recognition. This gives three fixed points on what 
can be pictured as a chart on which systems could be given position by 
reference to these points. 

To understand certain features of kinship we have to recognize that 
in many systems the structural unit consists of a woman and her children. 
This is very clearly seen in the patrilineal tribes of South Africa in which 
a polygynous family consists of two or more such units, each with its 
separate dwelling and food-supply, united by the relation to the man 
who is husband and father. It is by the position of this structural unit 
in the total kinship structure that we can define the contrast between 
mother-right and father-right. In true mother-right the unit group of 
mother and children is completely incorporated, jurally or legally, in the 
group of the woman's brothers and sisters. In true father-right the unit 
group is incorporated for jural purposes in a group consisting of brothers 
with their wives and children. 

In trying to classify kinship systems a most important feature to con- 
sider is the way in which the relationship of a person to his mother's 
siblings and to his father's siblings is institutionally defined. I n any 
system that approximates to „ tke-cogn atitL-£y2g__there is a tendency to 
treat the lather's brother and Jhe, mother's brother as relatives of the 
same kind, and similarly with_father's sister and mother's sister; but the 
assimilation may be less completejwhere there is some recognition, even 
though rt~besIigEt7of unilineal relationships. The degree of assimilation 
may sometimes be indicated in the terminology, as in the English use of 
'uncle' and 'aunt'. In classificatory terminologies it may appear in the 
inclusion of the mother's brother under the term for 'father' and of 
father's sister under the term for 'mother'. In the Lozi system a man calls 
all the cognatic relatives of his mother in her generation (the children 
of her father's and mother's brothers and sisters) 'mothers', male and 
female, and classifies as 'fathers' all the cognatic relatives of his father. 
It is therefore not patrilineal or matrilineal lineage that is recognized in 
this terminology, but the Old English distinction amongst the cognates 



of a person between those on the 'spear' side (through the father) and 
those on the 'spindle' side (through the mother). It contrasts with the 
common Bantu custom of using 'father' for the father's brothers and 
sisters and other persons of the father's Hneage and generation, which is 
an appUcation of the uniHneal principle. 

Cognatic systems are rare, not only in Africa but in the world at large. 
The reasons have already been indicated : it is difficult to establish and 
maintain a wide-range system on a purely cognatic basis; it is only a 
unilineal system that will permit the division of a society into separate 
organized kin-groups. 

In a typological classification of unilineal systems an important place 
must be given to those systems, of which the Yako are a good example, 
which recognize and attach importance to both ijiatrilineal and patri- 
lineal lineage relationships. This provides a special way of organizing 
a system of divided right by a cross-segmentation of the society. 


The view advanced in this Introduction is that to understand any 
kinship system it is necessary to carry out an analysis in terms of social 
structure and social function. The components of social structures are. 
human beings, and a structure is an arrangement of persons in relation- 
ships institutionally defined and regulated. The social function of any 
feature of a system is its relation to the structure and its continuance 
and stability, not its relation to the biological needs of individuals. The 
analysis of any particular system cannot be effectively carried out except 
in the light of the knowledge that we obtain by the systematic comparison 
of diverse systems. 

All the kinship systems of the world are the product of social evolu- 
tion. An essential feature of evolution is diversification by divergent 
development, and therefore there is great diversity in the forms of kin- 
ship systems. Some idea of the diversity of African systems can be 
obtained from the sections of this volume. Comparison of diverse systems 
enables us to discover certain resemblances. Some of these are features 
which are confined to one ethnic region. An example in Africa is the 
important part played by cattle and their transfer in the system of 
marriage and kinship. This is a feature of the patrilineal cattle-keeping 
peoples of East and South Africa from the Sudan to the Transkei. It is 
illustrated in this volume in Professor Wilson's paper on the Nyakyusa. 
It would be valuable if some student could give us a systematic compara- 
tive study of the various customs of this kind. As an example of the kind 
of thing that is meant may be mentioned the custom in some parts of the 
Transkei that if a man drinks milk from the cattle of a lineage other than 
his own he may not thereafter marry a woman of that lineage ; he is their 
kinsman through milk and cattle. In the same region the ceremony that 


completes a marriage is that in which the wife, who has usually borne at 
least one child, drinks for the first time the milk of her husband's herd. 

There are other similarities of custom which, so far from being limited 
to one ethnic region, are widely distributed in parts of the world distant 
from one another. Customs of avoidance and joking between relatives 
by marriage are found in regions scattered all over the world. The custom 
of privileged familiarity of a sister's son towards his mother's brother is 
found in some peoples of Africa, Oceania, and North America. Then 
again the Swazi in South Africa and the Cherokee in North America 
both apply the term 'grandmother' {ugogo in Swazi) to all the women of 
the lineage or clan of certain grandparents (father's father and mother's 
father in the matrilineal Cherokee, father's mother and mother's mother 
in the patrilineal Swazi) and give some measure of preference to marriage 
with such a 'grandmother'. A general theory of kinship must be tested by 
the help it gives us in understanding or explaining these resemblances. 

For a kinship system to exist, or to continue in existence, it must 
'work' with at least some measure of effectiveness. It must provide an 
integration of persons in a set of relationships within which they can 
interact and co-operate without too many serious conflicts. Tensions and 
possibilities of conflict exist in all systems. Professor Fortes and Dr. 
Richards have pointed out the tensions that exist in societies that 
approximate to mother-right. In systems approximating to father-right 
the tensions are different but exist none the less, as may be seen in the 
account of the Swazi. For a system to work efficiently it must provide 
methods of limiting, controlling, or resolving such conflicts or tensions. 
Dr. Nadel, in his paper on the Nuba, contrasts the Nyaro system, which 
he regards as providing an effective social integration, with the Tullishi 
system, which functions less efficiently. The Tullishi, he thinks, have 
been less successful than the Nyaro in constructing a well-ordered 
system of social relations, and he looks for the reason. 

Whether a kinship system functions well or not so well as a mode of 
social integration depends on the way it is constructed. Just as an archi- 
tect in designing a building has to make a choice of structural principles 
which he will use, so, though in less deliberate fashion, in the construc- 
tion of a kinship system there are a certain limited number of structural 
principles which can be used and combined in various ways. It is on the 
selection, method of use, and combination of these principles that the 
character of the structure depends. A structural analysis of a kinship 
system must therefore be in terms of structural principles and their 

The unit of structure everywhere seems to be the group of full siblings 
-T-brothers and sisters. The group has its own internal structure by virtue 
of the distinction between the sexes and the order of birth. Its members, 
however, are of 'one flesh and blood', and every system makes some use 
of this solidarity between siblings. This means that everywhere it is felt 


that brothers and sisters ought to exhibit affection and ought to 
co-operate and interact without serious conflict. Some of the ways in 
which the sohdarity and unity of the sibling group is utihzed in building 
wider structures have been illustrated in this Introduction. 

Since in all societies the closest parental bond is that of children with 
their mother, the group of brothers and sisters with their mother con- 
stitutes a more extended unit of structure. This can be best seen in the 
polygynous families of patrilineal societies, and is illustrated in the 
accounts of the Zulu and Swazi. There the compound parental family 
consists of mother-children groups, each forming a separate 'hut' or 
'house' (indlu), united to the man who is the father and husband. 

In this Introdu ction ^^ ha.'s_hppn suggested that one of the most im- 
portant questions to ask about a system is, in^what way, if at alj^ it- makes 
use of unilineal kinship aS distinct from cognatic kinship. Unilineal 
kinship receives only a minimum of recognition, if even that, in the Lozi ; 
matrilineal kinship is emphasized in the Ashanti, and patrilineal kinship 
in the Zulu and Nuer : both matrilineal and patrilineal kinship are made 
use of in the construction of the system of the Yako. Between these four 
selected types there are many intermediate forms. 

A further important feature of the social structure of any people is 
the way in which the kinship system is connected with the territorial 
arrangement of persons. In the past this has been often overlooked. 
Careful studies of this are now being made in Central Africa by the 
Rhodes-Livingstone Institute and the paper in this volume by Professor 
Evans-Pritchard is directed to a study of this aspect of the social struc- 
ture of the Nuer. It is in the contact and co-operation of neighbours in 
a territorial group such as a village, or what Professor Schapera calls 
a ward, that relations of kinship have their most continuous influence on 
the social life. Professor Evans-Pritchard draws attention to the differ- 
ence between the dyadic, person to person, relationships that every 
kinship system includes, and the group relationships that are established 
by a system of lineages or clans. They are, of course, both included in 
what has been called here a kinship system, but the difference between 
them needs to be recognized. 

In this volume it has not been possible to deal systematically with the 
relation of kinship systems to other parts of the social system, to religion, 
to political organization, and to economic life. This can, however, be 
studied in monographic studies of particular peoples. Professor Evans- 
Pritchard in his book on the Nuer, for example, has dealt with the part 
played by the system of lineages in the political organization of that 
people, and Dr. Kuper in An African Aristocracy has shown how the 
kinship system is connected with the political organization of the Swazi. 

African societies are undergoing revolutionary changes, as the result 
of European administrations, missions, and economic factors. In the 
past the stability of social order in African societies has depended much 


more on the kinship system than on anything else. In the new conditions 
kinship systems cannot remain unaffected. The first changes are 
inevitably destructive of the existing system of obhgations. The 
anthropological observer is able to discover new strains and tensions, 
new kinds of conflict, as Professor Fortes has done for the Ashanti and 
Professor Daryll Forde shows for the Yako. How far the disruption of 
the existing social order will go, and in what direction reconstruction 
will be attempted or possible, it is at present impossible to judge. The 
sanctions provided by the kinship systems for the control of conduct are 
being weakened. For example, some of those sanctions were religious 
and cannot persist where missionary enterprise is successful. Judging 
by what is happening in some parts of Africa the new sanctions, of which 
the agents are the policeman and the priest or minister of the church, 
are proving much less effective than those of which the agents were 
kinsmen speaking with the authority of the ancestors behind them. 

The process of change is inevitable. To a very limited extent it can 
be controlled by the colonial administration, and it is obvious that the 
effectiveness of any action taken by an administration is dependent on 
the knowledge they have at their disposal about the native society, its 
structure and institutions, and what is happening to it at the present 
time. A wise anthropologist will not try to tell an administrator what he 
ought to do; it is his special task to provide the scientifically collected 
and analysed knowledge that the administrator can use if he likes. 



THE Swazi of south-eastern Africa are an amalgamation of 
Nguni and Sotho clans, welded into a political unit in the late 
eighteenth century by a conquering Nguni aristocracy. By this 
military conquest, allegiance to the king overrode clan loyalty 
and gave rise to a system of government in which the clans — the farthest 
unilateral extension of a kinship group — were subordinated to a central- 
ized military and political organization. Kinship patterns of the Sotho 
were blended, as we shall see, with those of the dominant Nguni. 

Fictional kinship extending throughout the nation was created through 
the royal family, with the king and his mother as symbolic parents of 
the people, and much of their authority is expressed in kinship terms 
and the associated rights and obligations. Thus in the field of religion 
the rulers appeal to their royal ancestors on behalf of the entire nation, 
on the same principle by which the head of each homestead appeals to 
his own ancestors on behalf of his related dependants. Swazi kinship 
operates on both the domestic and national level, but the latter is a 
derivative and I will concentrate in this essay on the domestic aspects.' 
The Swazi word for kinship — uBunitii — defines relationship in the 
general as well as the specific sense. The term isinini (kinsman or kins- 
woman) is extended to non-relatives as a title of affection and gratitude. 


Every Swazi belongs to a patrilineal clan, each of which has a dis- 
tinctive isiBongo or clan name. The clan names (of which I counted over 
seventy) are either eponymous or refer to episodes in the history of the 
founders. A number of clans which are now separate were once one, 
and this appears from the isinanatelo (from ukwenana, to borrow with the • 
intention of returning). The isinanatelo is really an extended clan name 
providing useful historical clues to clan movements and affiliations. If 
the same name appears in more than one isinanatelo, the groups can be 
regarded as related sub-clans and intermarriage between members is 
taboo for them as well as for clansmen. Only in the case of the royal 
clan — the Nkosi — is intermarriage between sub-clans permitted. 

Clan subdivision continued at the same time as the centralized 
political authority developed, and fission was most frequent in the 
ruling (Nkosi) clan. The Nkosi Ginindza, Mamba, Dvu, Ngwenya, 
and others are all offshoots of the original Nkosi Dlamini. When the 
tie with the royal parent became so distant that it carried no privileges, 

■ I have described the national aspect of kinship in An African Aristocracy, 
1947. PP- 54 ff. 


the clan name Nkosi was itself occasionally dropped. Clans subdivided 
through the rivalry of brothers for leadership; or to reward a clan 
member for services, thereby bestowing on him a following and an 
area; or (in the Nkosi clan) to sanction intermarriage. Subdivision is 
now only continued in the royal clan enabling the king to marry women 
otherwise prohibited by the law of exogamy. Indirectly this keeps the 
Dlamini nobility a numerical minority and at the same time allows 
the king to reaffirm his hold on all sections of the people through the 
ties of polygynous marriage. 

At one period of Swazi history every clan had a territorial centre, but 
at present only a limited number have a clearly defined locality. 
Evidence of the local base is provided by the changing meaning of the 
word isifunza : in reply to the question, what is your isifunza, informants 
usually give their isiBongo, but when asked 'Who is the chief of the 
isifunza V some give the name of the political head of the district though 
his clan name is different from their own, and others give the name of 
the man they consider the direct heir of the founder of their clan though 
he does not live in the same locality. A distinction has thus emerged 
between the isiBongo, which may be scattered over a number of political 
units, and the isifunza, which is a local political unit with or without a 
kinship nucleus. 

Segmentation of the Swazi clans has not given rise to any distinctive 
interlinked lineage structure, that is, to a structure in which the links 
are remembered. The senior lineage is not even always known, except 
in clans which have specific tasks in the national organization. As a rule 
the lineages are parallel, unable to trace exact genealogical connexion. 
There is no specific word for a lineage : the closest approximation is the 
word lusendvo, which was originally applied to the entire clan but is now 
restricted to the effective family council which coincides roughly with 
a lineage. The lineage depth of a commoner is rarely more than five 
generations; for the aristocrats it increases to eight or ten. 

Clan members follow similar taboos and rituals, but only a few clans 
still practise rituals involving the clan as such. There is a grading of 
clans into a rough hierarchy according to their position in the tribal 
structure: at the apex is the conquering Nkosi Dlamini in which the 
lineage of the king is pre-eminent. This is followed by clans that have 
provided queen mothers, then by clans with their own ritual of chief- 
tainship, then by clans holding positions of national importance, and 
finally by clans with no co-ordinating clan ceremonies, no national 
officials, and no local centre. The grading of clans does not depend on 
the customs, all of which are regarded as having equal merit for their 
members. Moreover, the grading is flexible, some clans having risen 
through diplomacy or loyalty and others having been demoted through 
conquest or treachery. The position of the royal clan sets the limit of 


It is clear that the effectiveness of clan membership depends on the 
national standing of the clan. To men and women of the more important 
clans clanship involves political and social privileges, but to isolated 
members of scattered, defeated, or immigrant clans, the clan is simply 
an exogamous group with a common name and a few common customs ; 
it has no defined organization or claim on active loyalties. 

Swazi exogamy extends beyond the sub-clan to the clan, but, as has 
been mentioned, this is not enforced in the case of the king. Moreover, 
in contrast with other Nguni a man may, as will be shown, marry into 
the clans of his mother, paternal grandmother, and maternal grand- 
mother, although certain relatives in each of these clans are forbidden 
him. The principles affecting the kinship side of the clan activities 
are developed from behaviour in the family. 


The starting-point of the Swazi kinship system — the 'elementary 
family' of father, mother, and child — depends as in all societies on a 
recognition of a social relationship which may or may not coincide with 
a physical tie. The Swazi say 'A child is one blood with its father and 
its mother', but if a man has refused to perform the recognized marriage 
ceremony the children will belong to the mother's kin. The indisputable, 
but by no means sole, evidence of legal marriage is the acceptance 
by the bride's people of a number of cattle known as emabeka under 
the custom of uhiloBola} UkuloBola is specifically connected with the 
woman's fertility and with rights over children. In cases where the 
woman is sterile and cattle have already been paid, a substitute must be 
sent 'to bear for her'. The rights conferred over the children appear when 
a subject obtains marriage-cattle from his chief; the first daughter of 
the union is then spoken of as the chief's child, and cattle received on 
her marriage go to his heir. If the king contributes even one beast, he 
receives the full emabeka of the first daughter. In both cases the woman 
is recognized as the legal wife of the subject and the payment of the 
cattle gives the other men no sexual rights. All children other than the 
first daughter are regarded as the husband's, even though he may not 
be the genitor. A man may loBola a woman long after she has borne him 
children, and thereby establish them as his legitimate offspring, but — 
and in this the Swazi differ markedly from neighbouring Bantu tribes — 
he may simply loBola the children, leaving the woman to marry elsewhere. 

The sociological relationship of father and child may conflict with 
the physiological bond, for while the genitor may be denied the right 
of social paternity, a child is said to be impelled by the 'call of blood' 
to seek him out in later life. Here we have one of the many instances in 
which a social norm is challenged by individual emotions ; viewed from 

' Only the king need not give emabeka for his queens. 



the other angle, the emotional interests of a man and woman for legiti- 
mate 'blood' offspring is a major sanction behind family stability. The 
importance which children attach to the respectability of their mother, 
and to their own rights in the paternal family, appears from the cases 
where sons, after the death of the father and sometimes even after the 
death of the mother, give emaheka for her to her people. 

Swazi consider the production of legitimate children a social obliga- 
tion of adult men and women. For a few years after puberty, boys and 
girls are expected to have tingani (sweethearts), but the girls may not 
fall pregnant without losing value on the marriage market and the boys 
are fined anything up to ten head of cattle depending on the rank of the 
girl and the ruling of the court. After marriage the emphasis is on child- 
bearing. 'The marriage-cattle are closed by the carrying sling' (in which 
the baby is tied on the mother's back) is a common saying. Should a 
grown man die without children, but leaving behind a lover, it is the 
duty of his father or his father's heir to give cattle for her and provide 
a 'thigh' from the kinship group to raise children for the deceased. This 
is an instance of what Evans-Pritchard has styled 'ghost marriage'. 
Should a woman be childless, her husband is entitled to a co-wife from 
her family to produce children, and they are spoken of as the children 
of the older woman as well as of the younger. The extension of maternity 
through kinship of co-wives is described as 'to put into the womb of 
another'. A husband may even take the child of any junior wife and 
'put it into the womb' of a senior unrelated co-wife as heir. This type 
of adoption is most frequently practised in the case of main wives who 
have no sons, and it is then always best to choose boys whose own 
mothers are dead or whose mothers are deeply attached to the senior 
woman. Adoption of non-related children is unknown, the nearest 
thing to this being the ancient custom of taking over orphans whose 
parents had been killed in war or for witchcraft. In many respects these 
children were identified with those of the family head, but they were 
described — in their absence — as tigcili (domestic serfs), and were not 
included in the full kinship circle. 

Wives and children are regarded as a man's greatest assets and poly- 
gyny is the ambition of the tribesmen, an ambition most frequently 
achieved by aristocrats and wealthy elderly commoners. The king is 
expected to take more wives than any of his subjects, and the present 
king, who was born in 1902, has more than forty wives. Not only do the 
queens enhance his prestige and provide him with labour, but they are 
diplomatically selected from a wide range of clans which are thereby 
drawn into in-law relationship with the royal family. Swazi say, 'poly- 
gamy (isithembu) is the nature of man, while the nature of woman is 
satisfied through children'. 



On marriage a Swazi girl leaves her family and goes to live in her 
husband's homestead (wnuti). If his parents are alive he may attach 
his huts to the paternal circle or build in the vicinity. Should a patriarch 
have a number of wives, there is a strong tendency, especially when the 
sons marry, for his homestead to subdivide into smaller units of mothers 
and sons. Women always move with their own sons, but as long as the 
patriarch is alive he is the main authority in all legal and economic 
issues. Full brothers are more likely to continue to share a homestead 
after marriage than sons by different mothers, but when they have 
many wives, full brothers often build separately, the mother staying 
with the older son and visiting the younger. The section of the homestead 
in which the mother lives is known as the capital (umpakatsi), and in 
relation to it any other section is merely the barracks (lilau). 

The homesteads, scattered irregularly over the country-side, vary 
considerably in size, with the wealth and status of the headman. In a 
sample area of 66 homesteads, the average number of occupants was 7-2 
for commoners' families and 22-5 in the homesteads of aristocrats. The 
largest umuti in the whole county is owned by the ruling family; the 
smallest that I came across belonged to a Christian widow living with 
two unmarried sons several miles from their nearest kin. The small 
isolated group without even a single complete family unit is a modern 
development and is still most unusual, though for various reasons — 
economic, religious, and political — homesteads are undoubtedly dwin- 
dling in size. The diminishing of the local kinship group is an indication 
of certain changes taking place in the traditional structure, but space 
does not permit an analysis of these changes. 

A conservative homestead is shaped in a semicircle facing a central 
circular cattle byre. Each married woman has her own quarters con- 
sisting of living- and store-huts and kitchen enclosed by a high reed 
fence. An enclosure is spoken of as 'the hut of so-and-so' and relation- 
ship is counted through the different 'huts'. There is no rigid placing 
of 'huts' in order of rank as amongst some southern Bantu, but the 'great 
hut' [indlimkuhi) is usually in the middle of the semicircle. It is occupied 
by the mother of the headman, or, if she is dead, by a classificatory 
mother or a special wife raised to the status of 'mother'. The indlunkulu, 
associated with the elders of the headman's lineage, is the shrine of the 
homestead. Wives are grouped on either side of it in an order largely 
determined by the wishes of the headman, though he frequently places 
the enclosure of the first wife on the right of the 'mother' and the second 
on her left, the third on the right, and so on. Subdivision of a homestead 
does not, however, follow any defined grouping of huts, and it breaks 
up the original order. In the harem of the king and leading princes the 
arrangement of the wives is somewhat different : there are a number of 


'big queens', each with a number of 'Httle queens' attached to their 

In every homestead the headman's Hne is dominant. All his children 
have the paternal isiSongo, which is the same as that of his brothers and 
their children and any sisters living in his home. Wives always keep 
their own clan names and this distinguishes them as the 'strangers', the 
out-group. Unrelated people may attach themselves to the headman 
and live in his umuti, but a sharp distinction following the difference 
in isiBongo is drawn between 'our folk' and 'cold people'. 

Like all southern Bantu, the Swazi live by cultivating crops and by 
animal husbandry, and the homestead is the main unit of tribal economy, 
with the division of labour regulated by sex and age. The women are 
primarily responsible for the routine work, which carries lower prestige 
value than the activities monopolized by men. 

In the homestead Swazi religion, like Swazi economics, gives greater 
power to the males than to the females. While both men and women 
become ancestral spirits after death, it is the men who officiate as family 
priests for the entire homestead. 

The group living in the homestead is always impermanent and fluctu- 
ating, affected by marriages, births, deaths, migration, and immigration. 
At the same time, initial behaviour is shaped in the elementary family 
with its local basis in the homestead. Here the individual acquires the 
stereotyped norms reflecting the economic, legal, and religious values 
of the society, essential for the stability and perpetuation of the total 
social structure. 


In the following section I will outline the three basic sets of behaviour 
patterns in the immediate family — between the spouses, parents and 
children, and siblings. Since the ideal family is polygynous, the position 
of co-wives will also be discussed. The behaviour patterns fit into a legal 
framework which curtails individual preferences and aversions. Inevi- 
tably the norms of kinship formulated by the people give little idea of the 
degree of latitude, or of flagrant violation, or of counter interests, 
which emerge from observed behaviour, and this can be but touched 
upon in this article. 

{a) Husband and wife 

The formally expressed relationship between husband and wife 
emphasizes the husband's legal rights and makes little mention of the 
wife's emotional influence. Marriages based on affection are common, 
but in polygynous homes bestow lower status on the wife and her 
offspring than marriages arranged by parents or than preferential 
marriages. A husband is described as the owner (umnikati) of his wife; 


he cannot, however, do with her as he wishes without being called to 
account not only by his kinsmen but by her own family. While he is the 
legal head of the home, she is entitled to definite considerations. She has 
the right of iiBufati (wifehood) to demand land to support herself and her 
children, cattle to provide the children with milk and an inheritance, 
help in various economic and social ventures, protection against insults 
and attacks from outsiders. While she may not dispose of her crops 
without her husband's consent, he may not use any of the property of 
her hut — even the animals he has allocated to her children — without her 
consent. Only over the property of the main hut has he, together with 
his 'mother', undisputed control. By payment of a beast to her parents 
he acquires rights to his wife's independent earnings as a herbalist or 
diviner, though he must consult with her as to their distribution. Swazi 
say, 'A wife is bound to her husband with a law case' ; at the same time 
a husband knows that if he wishes to have a harmonious home, he must 
temper mastery with consideration. Should a wife be lazy, neglect her 
children, refuse to cook for the man and his kin, or in other ways break 
her side of the marriage bargain, he is entitled to beat her, albeit not to 
excess — he must not 'break the skin' — and if he should, she returns 
home and her people will exact a fine before allowing her to return. 
Adultery by the wife is spoken of as 'stealing' from the husband, and in 
former times the wife was sorely beaten and the lover, if caught in the 
act, could be killed. To-day the lover is fined in the court, and the 
woman may be beaten — though if the husband has been away for some 
years her action is often condoned and she is said 'to hit birds' for her 
husband who has the right to the adulterine children. This recent 
tolerance of adultery under present conditions, when large numbers of 
men are away from their homes for long periods working for Europeans, 
appears to be a defence technique of the husband and his group 
anxious to benefit by the fertility of the woman for whom they gave 
the marriage-cattle. 

Divorce is practised by the Swazi, but is extremely rare. A husband 
can 'return his wife' if she continually misbehaves, and he will receive 
back his cattle minus two for a daughter and one for a son. It is very 
much more difficult for a woman to refuse to return to her husband if he 
asks her people to send her back and is prepared to pay a fine for any 
wrong he admits to having committed. In the only case that I came 
across of a woman succeeding in her divorce, her husband had beaten 
her so brutally that her people laid a charge against him at the Com- 
missioner's court and he was sentenced to prison for six months. His 
older brother agreed that the woman should not return, and said her 
people could keep the cattle. Commenting on this case, old councillors 
insisted that the husband still had the right to any children she might 
bear. Only if a husband — or if he is dead, his heir — agrees to 'break 
the stomach' or 'to break the loin' can a woman marry a second time. 


The second man pays a fine, usually ten head of cattle, of which half 
goes to the chief and half to the group of the first husband. Children 
born after this payment are the legitimate offspring of the second man, 
who gives a few emabeka to the girl's father, who does not return the 
animals given by the first husband. 

A rigid etiquette of hlonipha (respect or shame) is demanded from 
a wife in her husband's home (ekhakhakhe) . She is prohibited from 
using the names, or words similar to the principal syllable of the names, 
of her husband's nearest senior male relatives — his father's father, his 
father, his father's senior brothers, his own senior brothers — living or 
dead. If she forgets, she is threatened that her tongue will rot, and if 
she continues to speak the forbidden sounds, she is sent to fetch a fine 
from her family. A wife must also avoid certain places in the husband's 
home; she may not eat various foods, including milk, and she may not 
wear the style of clothing permitted to the girls of the family; finally 
she must behave with restraint and conspicuous humility on numerous 
occasions. The restrictions on her conduct ekhakhakhe are most marked 
during the first year of marriage, and, apart from the language taboos, 
decrease with the birth of children and years of service for the in-laws. 
When she is well established the husband, usually on the suggestion of 
the mother-in-law, gives her, if he can afford it, one or more animals 
known as lipakhelo (from kupakhela, to serve). The main purpose of 
these is to provide against destitution in her old age and to enable her 
to assist her children. Once she has the lipakhelo she is allowed to eat 
curdled milk ekhakhakhe. The laws of ukuhlonipha, apart from the 
language of respect, are considerably modified if a wife is already a 
kinswoman of her husband, or is attached as co-wife to one of her own 
relatives, in which case the first woman is said to 'have finished the 

The wife is subordinate to the husband in the family religion. Her 
ancestors may affect the health of herself and her children, but they 
can only be placated at the home of her own kinsmen and through her 
male line. The husband is the priest for his ancestors and it is his 
obligation to keep them well disposed towards his wife. Difficult birth 
is sometimes attributed to the anger of his ancestors because of 'bitter- 
ness' between himself and his wife. When the young couple have 
'cleansed the heart', the baby is permitted to come forth. 

Wives of a polygynist hold unequal status; during his lifetime they 
are graded primarily on the basis of seniority, the first taking precedence 
over the second and so on, but after his death the children's rights to 
inheritance and succession are determined by their mothers' rank and 
mode of marriage. The first wife of an important man is chosen tor him 
by his senior paternal relatives and by his mother when they consider 
that he is ripe for marriage. The first wife is his isisulamsiti (wiper away 
of darkness) and will not be chosen as mother of the heir except in 


unusual circumstances. Her son is the lisokancanti (first circumcised), 
the adviser and confidant of junior brothers, including the heir. 

Except in special circumstances, the successor to a polygynist is not 
known publicly during his father's lifetime. Only after the death of the 
husband does the family council choose the heir in accordance with the 
status of the wives. 

The public behaviour of spouses is also regulated by rank and mode 
of marriage. Princes and chiefs exact more respect and obedience than 
commoners, and the king's wives are constantly guarded. On the other 
hand, daughters of the aristocracy are treated with exceptional honour 
and given considerable freedom. 

Co-wives call each other dzadze or fetu (sister), but usually speak of 
each other by their different clan names or as mother of so-and-so. 
Co-wives participate in a number of joint activities in the daily routine; 
they each contribute to the main family meal and eat it in the yard of 
the 'great hut'; if the headman wants beer, all his wives must grind 
the grain, fetch the wood and water, contribute utensils ; if he wants his 
fields weeded all the wives work in them. Between a woman and an 
inhlanti — junior attached co-wife — there is particularly close co-opera- 
tion including the sharing of store- and cooking-huts. 

On the whole, co-wives are on friendly terms with each other and it 
is not unusual for a woman to ask her husband to take a junior wife — 
particularly a sister — to help with the chores. But underlying the friend- 
liness is always an undercurrent of suspicion; Bukwele is the special 
word applied to the jealousy of co-wives, and the majority of accusations 
of witchcraft are levelled against these women, particularly the headman's 
favourite. The keeping of peace in the harem depends largely on the tact 
and wisdom of the mother of the headman and his first wife. Close rela- 
tives are considered useful allies as co-wives, and the kinship structure 
enables women to have as tinhlanti a range of kinswomen — full sisters, 
half-sisters, and brothers' daughters. The reasons for this selection will 
appear later. 

{h) Parent and child 

From the Swazi point of view the strongest affective relationship is 
not that between husband and wife but between mother and child. 
This is created in the first place by the child's physical dependence on 
the mother. She fondles and tends the baby, carries it around on her 
back, and suckles it until it is weaned in the third year. The husband 
may not cohabit with his wife towards the end of her pregnancy and for 
a few moons after she has given birth, and he is not allowed to hold the 
child before it is three moons old. His masculine interests keep him from 
being with the mother to any great extent and prevent him from giving 
the infant intimate care. In the polygynous home, in particular, the 
children regard the father from some distance, and with more, respect 


than love. The mother is 'soft', indulgent, a buffer between the child 
and the father's 'hardness' and authority. (Throughout life, and 
symbolized in ritual, the feminine is associated with 'softness', the 
masculine with 'hardness'.) A mother has a strong influence on her 
children's marriage, and will stand out against her husband if he tries 
to force his daughter into a match which the girl finds objectionable; 
a mother even goes so far as to encourage the girl to run away to the 
maternal kin until the matter has been smoothed over. When a girl 
marries, the mother's care and affection are recognized in the special 
beast known as the 'wiper away of tears' (imsulaninyembeti) given her by 
the groom. This remains the mother's property and may not be taken 
by her husband ; it is inherited by her youngest son, the accepted darling, 
and not by the eldest who inherits the main property of her hut. Should 
a woman die leaving only daughters, the timsulamnyenibeti of the girls 
should be used by her husband or his main heir to loBola a wife who will 
be regarded as the 'mother' of the girls, though she may in fact be a 
sister-in-law. She will feed them, when they visit the home, from the 
late mother's gardens, and her son will be the heir of their hut. 

A Swazi woman reaches fullest social stature as 'mother' in the home 
of a married son, w^here she takes charge of his domestic affairs, and 
lives in the great hut. A Swazi riddle runs: 'If your mother and wife 
were drowning, which would you save first V and the correct answer is : 
'My mother; I can get another wife, but not another mother.' 

There is a recognized preference on the part of Swazi, particularly 
women, for male children. Herbalists tell me that pregnant women 
often come to have the sex of the unborn child 'turned into' a boy, but 
none of them could give instances of the desire for transformation into 
a girl. On the death of an only son, the mother's stereotyped wail is: 
*Who now will bury me ? Who will care for me ? I have lost my support. 
I am abandoned on the veld.' For while a woman holds a privileged 
position in the home of her married son, she cannot live at ease with a 
married daughter and her son-in-law, especially if he has more than 
one wife. 

The grown son's behaviour towards his mother combines deference 
with affection. For him to swear, uncover himself, or behave lewdly 
in her presence is believed to evoke direct ancestral punishment, and 
he will also be publicly rebuked and possibly fined by his family council. 
The mother is expected to scold him if he neglects his duties as son, 
husband, or father, and he must not answer back in anger. 

The emphasis is always on the own mother — 'the mother who bore 
me'. Her hut is ketfu — our home. Though a casual European visitor 
receives the impression that the children do not differentiate between 
the various wives of their father, each of whom they address as 'mother', 
and that the women treat all the children, whom they call 'my child', as 
their own, to a Swazi there is no confusion. No co-wife may suckle 


another's child; even if its mother is very ill, it will be sent to her own 
mother or to a full sister not in the same harem or to her father's sister. 
Children describe the 'mothers' by their status vis-a-vis their own 
mother. A senior wife will therefore be 'big mother', and a junior will 
be 'little mother'. A polygynist's grown children may be older than his 
youngest wives, but will always call them 'mother'. (Unlike the neigh- 
bouring Thonga, Swazi look with horror on the idea of a 'son' on the 
death of his father cohabiting with a young half-mother. In one case 
where this was known to have happened, the boy ran away to Johan- 
nesburg and refused to return. The woman fled to her own people.) 
Calling a woman 'mother' involves more than lip-service : the 'mothers' 
cook food which all the children share and when the real mother is ill 
or away her child still has its share. The element of personal affection 
is, however, important, and the real mother usually asks a special 
friend among her co-wives to look after her child so that it is not 
neglected — or bewitched. 

The mother's interest in her own children often comes out in disputes 
over property, and the intensity with which she guards the interests of 
her own hut. A woman is so unwilling for the property of her hut and 
the marriage-cattle of her daughters to be merged with the general estate, 
that if she has no son she will urge her husband to put a motherless son 
'in her stomach' or attach a junior co-wife to her for the specific purpose 
of providing a son. Women may not inherit property, but they are 
entitled to own it and pass it on to their youngest sons. 

Succession and inheritance go hand in hand, and the fundamental 
principle underlying the selection of the main heir of a polygynous 
estate is that property and power are inherited from men and acquired 
by them, but are transmitted through women whose rank, more than 
any other factor, determines the choice. 'A ruler is ruler by his mother.' 

Children are trained to regard the father as the legal and economic 
authority. It is hard to convey the extent of the subservience of the 
Swazi son to his father. He works for him, consults him in all his 
negotiations, refers to him as 'his head', takes legal oaths 'by father'. 
From infancy sons are taught to obey the father's word, and even 
married sons are never regarded as free from his control. As long as they 
live in his homestead they are expected to hand over to him whatever 
they may earn, and he may if he wishes give them back a portion. On 
the other hand, he must provide a son with cattle for his first wife, he 
must sacrifice on his behalf, and is responsible before the courts for any 
torts committed by his sons while in his homestead. 

The status of the wives affects the behaviour between the father and 
his various sons. The first-born son of the first wife is the father's 
confidant — he is told about the allocation of family property and who 
the father considers a suitable heir. The father makes special provision 
for the first-born during his lifetime, and may also 'cut portions' for 


some of his other children, but the main distribution is only clarified 
after the father's death and the appointment of the main heir. 

Daughters are removed from any legal title to the father's property, 
but a considerate father will provide them with a beast. The eldest 
daughter, the inkosatana (princess), usually receives special considera- 
tion. While close companionship between father and daughter is not 
expected, there is a recognized affection between them as a basis of 
reciprocal duties. A father keeps a watchful eye on his married daughter, 
anxious to secure her happiness and also to guard himself against any 
claim for return of cattle. On her marriage he provides her with a full 
bridal outfit and he periodically sends her gifts for herself, her children, 
and her in-laws. Should she be in need ekhakhakhe (at her in-laws) she 
seeks help from kuBo (her own family), and this can never be refused 
without incurring the anger of the ancestral spirits at her father's home 
who remain actively interested in 'their child'. 

(c) Siblings 

The different claims of the children, dependent on seniority and sex, 
to family property and position affect their behaviour to each other. 
While they are still very young they sleep in the same hut, play together, 
and eat from the same dish, but at about the age of eight, when they 
have the lobes of the ear slit, the boys are separated from the girls, 
given masculine duties, and in the game of 'house house' learn not to 
choose their sisters as their wives. Together with the emphasis on sex 
and age differences comes an awareness of the importance of seniority 
of birth, which even overshadows the question of age. They learn to 
call people by their kinship terms, and on the classificatory system may 
find themselves addressing as 'father' or 'mother' children of their 
own age or even younger. No special term of address exists for older 
brothers or sisters, though in describing their position in a family they 
use such terms as 'older', 'younger', 'first-born', 'the entrails (last born)', 
and such. 

The social identity of siblings is closest for children of the same sex 
by the same mother as well as by the same father. Swazi practise the 
levirate and the sororate. When a man dies, his younger full brother is 
expected to 'enter' the widows and bear children for the deceased. But 
should there be no younger full brother, the duty will fall on another 
junior kinsman rather than on a senior brother, the deceased's oldest 
brother in particular being treated with the respect and avoidance 
accorded to a father-in-law. Nor will the widows be 'entered' by the 
main heir, though he be junior in years to their own husband, for he 
assumes on the patriarch's death the role of 'father' in all practical 

A man regards his brothers as essential members of his family council 
{lusendvo), and would not decide such important matters as the marriage 


qS kinship among the swazi 

of a child, or the disposal of cattle or moving of a village, without 
consulting the senior members of the lusendvo. 

The tie between sons of the same father undoubtedly depends largely 
on the status of the wives, and it is over succession and inheritance that 
cleavages between half-brothers come out sharply and bitterly. But the 
possibility of friction even between blood brothers appears from the 
injunction in the royal family that a queen mother should have only 
one son. 'The king is not followed' (by blood brothers). Swazi clan 
histories are marked by tragic instances of fraternal disputes. 

While the main heir is supposed to assist his other brothers in the 
same way as a father, the unevenness in inheritance creates jealousies. 
The main heir receives far and away the major share — not only does he 
take the cattle of the main hut and of his mother's hut, but the emabeka 
of the first daughter of each independent wife as well as any unallotted 
property. Only if there is more than one daughter will the oldest son of 
an independent wife benefit directly by a sister's marriage. He will, 
however, receive the emabeka from the oldest daughter of any junior 
wife attached to his mother's hut. While the older brothers are thus 
usually provided for and the youngest son of each hut inherits the 
mother's property, the middle son is often left destitute unless the father 
makes special provision during his lifetime. Occasionally a father allots 
a particular daughter to such a son, and informs the family council 
that he will receive her emabeka, but this is rare : Swazi do not generally 
practise the custom of 'linking' brothers and sisters as do the Tswana 

The relationship of sisters is not affected to the same extent by 
primogeniture. While an older sister nearly always marries before her 
younger full sister, the daughters of different mothers do not necessarily 
I marry in order of age or status. In aristocratic families, moreover, the 
inkosatana (princess) has the privilege of retaining her girlhood freedom 
longer than other daughters of her father. 

Affection and mutual help are the norm of sisterly behaviour. We 
have already mentioned the desire of an older sister to have a younger 
as her co-wife. If married to different men, and living long distances 
apart, the girls still try to keep in touch with each other through 
travellers and it is not unusual for them, and even for half-sisters, to 
'borrow' each others' children, whom they call 'my children', as 
nursemaids and helps. 

While brothers and sisters of the same father, albeit by different 
mothers, are taught from an early age that sex relationship between them 
is prohibited, they do not by any means avoid each other. They often 
walk along hand in hand (a demonstrativeness tabooed a man and his 
wife), call each other to drink from the same bowl, give each other 
gifts, and render each other many informal services. A sister acts as 
intermediary in a brother's love afTairs and works for him the bead 


ornaments worn to make an impression on the girls. As usual, the 
degree of social distance varies with age and seniority: a boy shows 
more restraint towards his older than his younger sisters, and a girl 
knows that she will have to obey the main heir after her father's 

Despite the difference of sex, the 'unity of the sibling group' appears 
in the marriage ritual when the groom makes a symbolic bid for another 
girl from the bride's family.' Should the bride have no younger full 
sister, but should there be a brother by the same mother, this boy is 
pointed out as providing her ifihlanti. His daughter is a suitable junior 
wife for his sister — that is, for the girl's father's sister. While this 
arrangement is usually explained by the Swazi on the ground that the 
brother benefited by the sister's marriage-cattle and is therefore 
obliged to provide her with a co-wife if necessary, the underlying 
principle of the social equivalence of siblings is clear. It, and not the 
Swazi rationalization, accounts for the fact that a sister is the most 
suitable inhlanti, and only in the absence of a full sister does the obliga- 
tion fall on the full brother. In finding a substitute for an independent 
wife who has died childless, and without a full sister, the choice of a 
daughter by the full brother is largely determined by her status. The 
strength of this obligation appears from a case in the royal family. 
The late king Bunu had a full sister Tonga Tonga, who married an impor- 
tant Ndwandwe chief who gave for her 100 cattle. Tonga Tonga was 
accompanied, as inhlanti, by a half-sister by a junior wife of Bunu's 
father. Tonga Tonga died childless soon after the marriage, and it fell 
to Sobhuza, heir of Bunu, to provide a substitute. Tonga Tonga had no 
full sister, and the family council selected as her substitute Bunu's 
leading daughter, SencaBapi. Tonga Tonga's inhlanti then became 
SencaBapi's inhlanti, though she, SencaBapi, was her inhlanti's brother's 
child — an apparent reversal of the usual role, brought about through 
the emphasis on full status equivalence. 

The extent to which a brother benefits directly by the bride-wealth 
of a sister depends on his place in the family, but all the brothers are 
considered to share in the increased family prosperity. This point was 
frequently made by informants — and so brothers by the same father 
help sisters when in need. A legal obligation rests on the heir or on the 
brother who received her marriage-cattle, and a moral obligation on 
all the brothers. 

A sister enjoys a specially privileged position in the hom^e of married 
brothers, and can take from their wives (her sisters-in-law) clothing, 
utensils, and other odds and ends, which may never be refused. The wives 
should appear glad, and only if articles of too great a value are removed 
may they complain against 'the girl of the home'. The difference between 

' For a fuller description of marriage ritual see my article 'Marriage of a 
Swazi Princess', Africa, xv. 3, July 1945, pp. 145-55. 


the position of a woman as wife and sister is very marked. Whereas she 
is always an outsider ckhakhakhe, when she comes kiiBo she is addressed 
as 'princess', 'our girl', or by her first name; she does not use the 
hlonipha language of shame and respect and has the free run of the home. 
Her children, described affectionately as 'the calves', often spend many 
years kaBonina (at the mother's home) with their mother's brother 
{umalume, male mother). When the food is not cooked, or the fields 
are not hoed, the wife and not the sister is blamed, and when illness 
attacks a home the sister is seldom divined as a source of evil. Because 
of the trust that brothers have in their sisters being eager to promote 
their well-being, they are amongst the category of people to whom it is 
safe to send an heir till he reaches the strength of maturity. 


Beyond the immediate family, which I have discussed primarily from 
the angle of behaviour, extends a range of kinsmen on both the father's 
and the mother's side. The way in which these are recognized provides 
the kinship structure which, albeit it hinges at each level on an immediate 
family, has a permanency and consistency greater than that of any 
particular family. 

Though a Swazi retains links with the families of both his parents, 
the emphasis on the paternal kin [kaBoyise) and maternal kin {kaBonina) 
is not identical. The agnatic line has greater time depth, going back 
five to eight generations, and in the royal family even farther, while in 
the maternal group descent can be traced for only three or four genera- 
tions, except in those cases where the mother belongs to the royal 

The broad principles of the extended kinship structure can be seen 
from the chart of kinship terminology (Figs, i, 2). It shows the typical 
feature of the classificatory kinship system in which a limited number 
of terms applied to lineal relatives are also applied to collateral. Thus the 
term uBaBe is not only applied to my own father, but to my father's 
brother (by the same or different mothers) and to my father's father's 
brothers' sons ; and by a process of extension on the maternal side, it is 
also used for husbands of all the women I call mage — my own mother, 
my mother's co-wives, mother's sisters, mother's mother's sisters' 
daughters, and my mother's father's sisters' daughters as well as the 
wives of the men I call 'father'. 

Swazi kinship terminology covers a range of five generations in all — 
two ascending, the contemporary, and two descending. Beyond the 
second ascending — the grandparent generation — everyone is classified 
as Bogogomkulu, and beyond the second descending — the grandchild 
generation — as BantfwaBantfwaBami. If relationship with a person of 
the clan name of my mother or father cannot be traced, the generation 


GaSe'mkulu orgogo 


r6 k 

BaBefmkulu orjogo 


mntfwana- 6a6e. 6a6e iomsikati mntfwano'gogo 

•909° I 


malumc mage 



6aSe or 6a6e 





malume maqc ma- mooe 


.4.. ..A; 


6a6e mage 

Zl i.' 

\mzala\ mnaketfu dzodze- \mnaketfu dzadio- mnaketfu dzadzi 

I y^etfu EGO >fff" I r!f" 

,rt rH.; ...i^ rt ^....^....t3. f3,a- 

Jc ^ k k.' ••J-d..-b.t. 00.0,0-^-^?- -M^i-b-b.' ■d.A'^.h 




laketfu dzadze- I mialA 

mntjwana miaia 


mntfv/ana 'mnthnanami 


Each term is in the singular and is in the mcatiife case 
A= male o^ female 




,^•0 ^=0 


gogo 'mkuk 

"X = ,6 A=o 

6a6e'mkulu orc/ogv 


"5 = 


X - .0 ^r. 


SaSe'mMu or gogo 

6a6clomsikati j Sa6i 

.£ ST 


mags malume 

6a6eloms/ko(i Oat 




,A=0 = ^ 


2. = <5- 

(my husband] 

w<3<7« malume 

A=o 5Ti ^iO 6 = ^ 

malogatana mkwenyana mshana 


(mj mjej 

-- o ^^""^ h, 

mkwe- mzala 

nkwen- mahgatana mkwen- mlamu 

jana fana 



kinship terms are usually applied, any one of roughly the generation 
of my father being called 'father', and so on. 

Within each category of relations a distinction can always be drawn 
between close and distant kin. This is reflected in social obligations, the 
general rule being that the closer the tie the more precise and definite the 
obligations. The degree of distance will be elaborated in the descriptive 
term even though it does not appear in the term of address, which only 
indicates the broad category. Thus while a man will be able to describe 
accurately the 'huts' of his father's different wives, he will call them 
each mage, and while he describes his father's older brother as BaBeynkulu 
(big father), he may address him, as well as his father, as Bade. 

Different terms of address for the same person do not all have the 
same emotional meaning, some carrying more affection and others more 
deference. There is a special word kunambitsisa, derived from kiinambitsa 
(to have a pleasant after-taste from food), associated with the selection of 
the more demonstrative term. A woman will nambitsisa her brother's 
child if she calls him jmitfwanaketfu (child of our place), but if she were 
angry she would reprimand him as mnifwana^ mnaketfu (child of my 
brother), to imply his indebtedness to her, his father's sister. The word 
kunambitsisa is also used to describe the friendly familiarity permitted 
between various sets of relations, but the so-called 'joking relationship' 
is not developed among the Swazi. 

(a) Paternal Aunt and Maternal Uncle 

Swazi share with other Southern Bantu the custom of identifying 
each parent with his or her own siblings, with a slight differentiation 
according to sex. Not only are the father's brothers called BaBe, but his 
sister may be addressed as BaBe or by her descriptive term BaBelomsikati 
(female father). Similarly the mother's brother is mage or tnalume (male 
mother). The paternal aunt exercises the same type of authority over 
her brother's children as the father, and must receive respect and obedi- 
ence from them. I remember when Sobhuza was visited by his paternal 
aunt in his modern office that she sat on the floor and he thereupon 
got off his chair and squatted beside her, explaining to me afterwards 
that he could not be 'above her'. The father's full sister takes part 
in the choice of her nephew's main wife as a member of his family 
council, and, as we have already stated, the paternal niece is a potential 

The fact that the father's sister bestowed benefits on the father 
through her marriage cattle entitles her, by Swazi standards, to make 
various demands on him, on his wives, and on the children of the union 
which she made possible. On the other hand, the maternal uncle 
received advantages through the marriage of his sister, and she and her 
children may receive help from him. The umalume is always a great 
favourite; the bashana (sing, umshana, niece or nephew), may take beer 


from his hut without asking his permission, they bring him their 
troubles and seek his intervention in disputes with their own father. 
An umalume is under no legal obligation to assist even a full sister's son to 
collect bride-wealth, but he may do so from affection, and there have 
been a few cases where a man's own children resented the favours 
showered by him on a sister's children. 

Marriage between an umalume and a niece is not favoured since it 
introduces the tension of sexual interest into an easy protective relation- 
ship. Furthermore, by such a marriage the man's sister becomes his 
mother-in-law, and familiarity following on the sibling tie is distorted 
by the avoidance imposed on a son-in-law. Even if the umalume is the 
brother of the wife of a classificatory father, the marriage is not en- 
couraged. For this reason pressure was exerted, albeit in vain, against 
the marriage of Lamtana Dlamini, daughter of King Bunu's youngest 
brother, Lomvase Dlamini, to her umalume, Mzululeke Ndwandwe, 
brother of Bunu's main wife. 

If a woman has no child or if she has borne only daughters, she may 
ask her brother to provide an inhlanti. The case of the late queen- 
mother Lomawa is usually cited as an example of this arrangement. 
Her mother, laNdlela (of the Ndlela clan), married Ngolotsheni 
Mkwatshwa and had three daughters, of whom Lomawa Avas the eldest. 
Ngolotsheni spoke to his wife and suggested she should get an whlanfi 
so that the cattle for her daughters need not be eaten by the great hut. 
LaNdlela had no sister, but her mother's brother of the Madlamalala 
clan had a daughter, and she was given to laNdlela (her cross-cousin — 
umzala) as an inhlanti, and bore the required heir. 

Because of the father's sister's claim to her brother's daughter, her 
husband addresses her niece by the term of a potential spouse — mlamu 
— to which she responds with ynkwenyewctfu. And because of the 
brother-sister obligation, the nephew is also called umlamu. The differ- 
ence in generation is not important in husband-wife relationship in a 
polygynous society, where it is not unusual for a man's oldest sons to 
be senior to his youngest wives. Between umlamu and umktvenyezvetfu 
is an easy familiarity more fully described in the account of the behaviour 
of a man and his wife's sisters and brothers, the bulk of his Balamu 
(P- 108). _ 

A Swazi idiom states 'The seed is begged from the kin' {inhlanyelo 
iyacehva esininini). This is applied to cases such as those mentioned 
above in which a man marries more than one woman because of the 
women's relationship to each other. 

It is also applied to marriages with women to whom the man is 
related in his own right. It is in the attitude towards such marriages that 
we see the influence of the Sotho on the dominant Nguni pattern. 'The 
Nguni', Mrs. Hoernle writes, 'rigidly prohibit marriage, or sexual 
relations of any kind, with people related through any of the four 


grandparents' (p. 74). 'The Sotho tribes all allow certain types of kin 
to marry' (p. 86).' 

(b) Cousins 

Parallel cousins are classed as siblings and marriage between them 
is condemned as incest; cross-cousins are distinguished as Bonizala and 
marriage between them is permitted. In the case of close cross-cousins 
it is discouraged, but there is one instance in which marriage with a 
maternal uncle's daughter is approved by a section of the people. If a 
man's own mother is dead and he has no classificatory mother to put 
in her place, some Swazi consider it a sign of filial affection to marry 
a woman with the mother's clan name. Daughters of her full brothers 
and half-brothers are considered 'too close', but the daughter of a 'clan 
brother' is considered suitable 'to wake the hut of mother [kuvusa indhi 
yamagey . This gesture is usually made by a polygynist who puts the 
woman into the 'mother's hut' of his main homestead. A cross-cousin 
is never selected as main wife, but her link with the 'mother' gives her 
the privileges of 'mother' during her 'son's' lifetime. It is interesting to 
note that the Swazi are not agreed on the advisability of such marriages. 
Some regard them as admirable, others look on them with disfavour. 
The cleavage may originally have followed the division between Sotho 
and Nguni clans, but to-day it seems largely a matter of personal 
opinion. The discussion came out clearly when Timba Maseko, whose 
own mother, Tana Hlophe, died, married Nkaniso Hlophe, daughter of 
his mother's father's half-brother's son. Some informants considered 
it a good marriage, others were very much against it and argued that 
he should have put a 'cold' wife — a non-relative — into the hut of the 

The kinship marriage which is generally approved is that with 'ugogo' 
— a woman of the family or clan of the paternal or maternal grand- 
mother. This can be explained by the general behaviour between grand- 
parents, parents, and grandchildren. 

(c) Grandparents 

The social inequality between proximate generations was illustrated 
in the parent-child relationship, where it also became clear that the 
generation distinction could be overruled by primogeniture as in the 
promotion of the main heir to the status of 'father'. Marriage between 
relatives two generations apart perpetuates the family links without 
impinging on the parent-child behaviour. 

Since marriage is patrilocal, a child often grows up in the home of the 
paternal grandparents. The grandparents are said 'to teach the young 

' The Bantu- Speaking Tribes 0/ South Africa, edited by I. Schapera, Routledge, 


to respect their parents'. Grandparents 'scold by the mouth', parents 
more often 'with the stick'. The law of the past issues in the voice of the 
grandfather, and he must give his sanction and opinion on any important 
event, such as the marriage of a grandchild. The power exercised by 
apparently decrepit ancients over sons and grandsons is partly explained 
by the fact that they are 'the nearest to the ancestors' and essential 
officiators in all domestic rituals. The grandmother — ugogo — is the main 
teacher of the young. She sees that the correct ritual is performed to 
ensure the health and proper development of the grandchildren and she 
supervises the numerous ceremonies which punctuate their growth in 
status. In many districts a woman, after she has borne a child, is not 
allowed into the great hut until the baby can crawl, when it is led to the 
supporting poles in the hut and then into the courtyard where the granny 
is waiting ; she puts a little ash from the hearth on its forehead so that 
'it will be one with its forefathers'. When the grandchildren are weaned 
they are usually put in her charge, and they sleep with her till approach- 
ing puberty, when boys and girls move to separate quarters. 

So honoured and important is the paternal grandmother that any 
woman, irrespective of her generation, with the same clan name is 
addressed as gogo. Here the generation principle is violated and clan 
relationship triumphs over the family. Marriage with a woman of the 
paternal grandmother's clan gives her high status in a polygynous 
menage, for she is said 'to wake the hut of granny' and also 'to wake the 
hut of father'. Such a wife is freed from most of the early restrictions 
of hlonipha since they were 'finished' by the dead grandmother whose 
place she revives. 

The maternal grandparents show a proverbial affection and indul- 
gence: 'at the home of the mother, the grandchildren are adored.' 
When a daughter is pregnant for the first time her mother usually pays 
her a visit bringing special beer, and the son-in-law kills a beast for her 
in return. Maternal grandparents beg the father to send the child to 
them on visits, and when it arrives lavish on it admiration and tender- 
ness. 'It is the child of the calf.' If at all wealthy, the girl's father or 
brother gives the baby, whether boy or girl, an animal known as the 
'beast to kiss'. The maternal grandparents also slaughter a goat to 
provide the babe with a carrying sling. The maternal kin must be kept 
informed of any illness, and if it is divined as caused by lidloti lakiiBo 
(the ancestor of the maternal home) the girl's father or brother must 
provide the sacrifice and officiate at his own hearth. Illness is attributed 
more frequently to the paternal than the maternal kin, while very often 
a woman and her children are said to be 'saved' by the 'ancestors of the 
maternal home'. 

Marriage with a woman of the clan of the maternal grandmother 
{ugogo otalamage) is rated almost as high as marriage with a woman of 
the clan of the paternal grandmother {ugogo otala6a6e). Such a wife is 


said 'to wake the hut of the mother's people', and she too is absolved 
from some of the taboos of other wives. 

The position can be seen from the following diagram : 

Ai = bi Bi Ci = di Di 




B2 I C2 D2 

A3 b3 C3 d3 

Capital letters represent males, lower-case letters females. For A3 his 
mother's clan is C, his father's mother's clan is B, and his mother's mother's 

clan is D. 

The grandparent-grandchild relationship has various repercussions. 
Thus, by the marriage of A^ to b^ (marriage into the paternal grand- 
mother's clan), Az becomes the father-in-law of the woman b^ who is 
his mother's relation. The cattle that left Ai's group for the woman bi 
made A2 legitimate and created a bond with the B group, to give 62's 
son the greater part of the estate oi Ai's grandchild. A^, as the meeting- 
point of the four clans A, B, C, and D, may obtain his main wife from 
B or D, the groups to w'hich cattle w^ere given by A and C. The men of 
B and D have no reciprocal claim to the sisters of Az, though if they 
should intermarry many families rejoice that the 'cattle have returned 
home'. Other Swazi disapprove of such ties, saying that it is too like 
'buying from each other'. 

The order of preference in choosing a main wife of A^ from amongst 
63, C3, and d^ w'ould be b^yd^yC}. (Other modes of marriage, such as that 
mentioned in the final section on rank, are rated even higher.) Since 
no Swazi desires to create unnecessary conflict it very rarely happens 
that anyone marries both a paternal and a maternal ugogo, and in such 
cases the paternal line has preference unless the mother be alive, when 
she may persuade the family council to favour her own clanswoman. 
This happened in the case of that forceful character, the queen-mother 
Gwamile Mdluli. Her younger son Malunge Dlamini included amongst 
his wives a w'oman of the MaBuzo clan and another of the Nkambule. 
Nkambule was the clan of his paternal ugogo, mother of the King 
Mbandzeni himself, and MaBuzo was Gwamile's own mother's clan 
name. Malunge died and the family council wanted to choose as main 
wife the Nkambule woman, but Gwamile, though her own father was 
not of great importance, argued so strongly in favour of BakuBo that 
they agreed to choose the MaBuzo wife. 

In marriages with either a paternal or maternal ugogo the link may be 
through a common grandparent, as w^ell as with more distant members 
of these clans. 



In societies where marriages between kin are permitted the distinc- 
tion between relatives 'by blood' and 'by marriage' is not usually sharp. 
But amongst the Swazi marriages with kinsmen arc rare except in large 
polygynous homesteads. Of 35 monogamous marriages, i was with a 
maternal cross-cousin and i with a maternal uncle; on the other hand, 
5 out of 12 chiefs whose genealogies I traced were selected as main heirs 
because their mothers were of the ugogo clan — 2 paternal and 3 maternal 
(of the remaining 7, 3 were selected because their mothers were princes- 
ses of the ruling family; 2 were 'arranged' marriages (ukzvenzisa); 2 were 
women of exceptional character). 

Despite marriages between kinsfolk, the dominant pattern in affinal 
relations is set by the Nguni section of the population to which the 
royal clan belongs, and opposition between the two intermarrying 
groups is formally stereotyped. The husband's relatives, the Bekhakhakhe, 
are put into a different legal and emotional category from the BaktiBo, 
and the husband draws a parallel distinction with the kaBomfati (wife's 

{a) Parents-in-lazv 

The restrictions imposed on a wife have already been mentioned, and 
since marriage is normally patrilocal the position of wife and daughter- 
in-law overlap locally. The strongest avoidance is manifested between 
the bride (umkoti) and her father-in-law, stressing the sexual barrier 
between them. In a polygynous hom^e a father's young wife may be a 
contemporary of his son's wife, and a strong artificial sanction, based on 
maintaining peace between father and son, removes the son's wife from 
the range of potential mates of the old, dominant male. The young girl 
becomes the 'handmaiden' of the mother-in-law, whose own reproduc- 
tive sexual life must cease when her oldest child marries. 

The husband's behaviour towards his wife's parents is on the same 
pattern as hers to his group, but is eased by the fact that he does not live 
in their home. The difference of sex involves him in showing the 
strongest avoidance of his mother-in-law, his nmkzcegati. A man will 
be fined by his wife's family council if he makes lewd jokes or in any 
way misbehaves in her presence, and 'shame' prevents him eating (an 
act psychologically and verbally associated with sex) in her presence. 
She reciprocates with similar obser\'ances towards the son-in-law, whom 
she usually addresses as mkicenyana, or, if she wishes to nambitsisa him, 
as mntfanami (my child), but never by his personal name. Though 
some of the irksome taboos are absent in the case of kinswomen, and in 
any case decrease with the passing of the years, it is considered impos- 
sible for a man, particularly a polygynist, to live easily with his mother- 
in-law. 'She knows what I am doing to her child, who is a woman like 


herself.' One section of Swazi consider that the strongest prohibition 
against marriage, as strong as the law forbidding marriage with a person 
of one's own clan, is with a woman of the clan of the wife's mother, 
even though no relationship between the two women can be traced. 
We have seen that other sections, following the Sotho pattern, approve 
of this marriage in certain circumstances. 

A man treats his wife's father with respect and courtesy, knowing 
that he is the person to whom the wife will appeal if badly treated. 
Being of the same sex, the relationship with the father-in-law is less 
strained than with the mother-in-law, but the umkwenyana is always 
on his best behaviour on his occasional visits. The father-in-law helps 
the situation by inviting him to eat and drink with him, and by extending 
opportunities for co-operation in male occupations. A good son-in-law 
is recognized as an asset in economic ventures. 

The mother-in-law is the great umkwegati, but there are other women 
in a wife's home who are also tabooed and called Bakwegati — they are 
the wife's brothers' wives. Though the social distance between a man 
and these women is not as irksome as with his mother-in-law, they are 
removed from all sexual contact. They are, in fact, potential mothers-in- 
law, because they may be called upon to provide his wife with an 

(b) Sisters- and brothers-in-law 

In sharp contrast to the avoidance of Bakwegati is the familiarity 
between a man and his wife's sisters and brothers — the Balamu. A wife's 
sisters can be taken as independent or subordinate wives, and until 
they are married they indulge in rough and often suggestive horseplay 
with the married sister's husband {iimkwenyewetfu). Should he, however, 
make them pregnant without having negotiated marriage, he is fined 
and treated as an ordinary suitor. And once an umlamu is married, 
intimacy with her is regarded as adultery. The male umlamu will, if he 
provides a daughter as inhlanti to his sister, also be identified with a 
father-in-law, but the umlamu relationship is the stronger of the two. 

The popularity of the marriage of a man with more than one woman of 
the same family appears from these limited figures : out of 24 marriages 
of commoners, each with two wives, 7 married two sisters. In the harem 
of the king Sobhuza II we find the following marriage types for his 43 
wives : 14 women are unrelated to him and to each other; 4 women each 
have one full sister in the harem ; 2 women have clan sisters ; 8 women 
are of the Nkosi clan (one of them is linked with him through a common 
grandfather by different wives and this marriage evoked such criticism 
because of the 'closeness' that he set her up in a completely separate 
homestead from his other wives) ; 4 women are of the paternal grand- 
mother's clan; i woman of the maternal grandmother's clan. 

The husband's senior brothers receive some of the respect accorded 


to the father-in-law. Until his death all the husband's brothers and 
sisters are classed as BomkuluBakhe, and they reciprocate with the same 
term. After his death the main heir becomes 'father-in-law'. By the 
custom of the levirate, the younger brothers are a wife's potential 
mates, and they may call her jokingly 'my wife', to which she will reply 
'my husband'. But it is dangerous to show too much friendliness during 
the husband's lifetime lest they be accused of having caused the death. 
(The levirate relationship is fraught with difficulties and is in fact 
never gladly accepted by a brother, particularly if he has his own 
wives.) The term umkuluwakhe is also used of a woman and her brother's 
wives, and we have already described the right which she has to her 
husband's sister's daughter. Umkuluwakhe connotes mutual obligations 
and familiarities and a potential link through marriage. 

(c) The groups 

The two sets of parents-in-law address each other as Bakoti and are 
polite and hospitable to each other. The man's people slaughter a beast 
when the girl's father or mother pays them a visit and the girl's folk 
send beer and presents. Naturally a great deal depends on how close 
the homes are to each other, distance making regular effective contact 
difficult. However, certain courtesies, particularly notification of births, 
illnesses, and deaths, are ritually enforced by the belief that their 
omission would evoke 'the bitterness of the ancestors'. To meet a 
person and call him or her umakoti is to na?nbitsisa that person exceed- 
ingly. The term Bakoti is also extended to the grandparents of the 
young couple. 

Apart from the relatives-in-law whom I have specially mentioned and 
to whom man and wife react in different norms, there are a number of 
in-laws to whom man and wife behave in similar fashion. Swazi say: 
'My wife and I are one, through marriage, and her relations become 
mine', and vice versa. This explains the similar terms used towards 
each other's father's sister, mother's brother, and other categories evident 
on the chart (Fig. 2, p. loi). 


An essay on Swazi kinship would be incomplete without special 
reference to rank, for Swazi kinship, like all other Swazi modes of 
behaviour, is characterized by the importance of rank by birth. Even 
within the immediate family the relationships reflect the pedigree of 
the parents : the mother of the king, for example, is married with cattle 
provided by representative sections of the people, making the king 
'child of the country', and similarly chiefs of the district have the etnaheka 
of their main wives contributed by their subjects. 

The selection of a main heir is determined by the rank of the mothers, 
and we have indicated the preference given to women known as ugogo. 


Even more important are women of the ruling clan, particularly sisters ' 
and daughters of the king. These women are sent as wives to important 
chiefs, spreading the network of the dominant aristocracy. When the 
polygynist Kambi Maseko died, his wives included a daughter of the 
king Mswati (by the levirate) and one with the clan name of the paternal 
grandmother. The family council appointed the latter as main wife, 
and announced their choice to the ruling king. When he found out that 
a royal princess had been overlooked, he told the council to reconsider 
their decision, for 'it is not law that the king should eat on the ground 
while the dog eats from the plate'. 

To discuss fully the repercussions of rank on kinship would lead us 
into an analysis of the political structure, and we can in this essay merely 
indicate that the two are closely interlocked. 



THE Nyakyusa' are a Bantu-speaking group, numbering some 
150,000, who live in a portion of the Great Rift Valley at the 
north end of Lake Nyasa. Their country is fertile, the rainfall 
is heavy and well distributed, and they are skilled and diligent 
cultivators, growing a great variety of crops. They also keep cattle, and 
fish in the lake and streams. Hence the land can carry a heavy population, 
averaging over 80 to the square mile, and in contrast to many of 
their neighbours the Nyakyusa are well fed and vigorous. 

They are peculiar in that their villages consist not of kinsmen but of 
age-mates. Herd-boys of 10 or 11, who are the sons of neighbours, begin 
to build together at a little distance from their fathers. Gradually they 
are joined by younger boys, also from the village of their fathers, until 
the group covers an age-span of about five years ; then the next set of 
boys is told to start a new village of its own. Most of these age-mates 
remain together all their lives. As they marry they bring their wives to 
the village and the children they beget live with them until the daughters 
marry, and the sons, in their turn, build new villages. Though the site 
of a village changes the group survives until the death of its members. 
Once in each generation a great ceremony is held at which the chief- 
dom divides into two halves, each under a son of the chief, and authority 
is handed over to these sons and the leaders of the villages of young men. 
At this ceremony the old men within each half move aside, leaving the 
bulk of the land for their sons, who now rule the country. All the 
boundaries of the villages are redrawn, and the young men move from 
the site on which they built as boys to one which allows more room for 
their expanding families. Thus the Nyakyusa combine with chieftain- 
ship the formal handing of authority from generation to generation, so 
typical of the East African peoples organized in age-grades. 

Being surrounded on three sides by high mountains, and on the 
fourth by the stormy head-waters of the lake, the Nyakyusa long 
remained isolated from the outside world. They were scarcely touched 
by slave-raiders and successfully repelled marauding bands of Ngoni. 
Not until 1 89 1 did European missionaries settle in the country and the 
authority of the German Government was only established somewhat 

' The material on which this essay is based was collected between 1934 and 
1938 by my late husband, Godfrey Wilson, and myself. During that period 
he was a Fellow of the Rockefeller Foundation, and I a Fellow of the International 
African Institute. We are indebted to these bodies for makinc this research 


later. But by the middle of the nineteen-thirties (to which period this 
description refers) the Nyakyusa were doing a vigorous trade with the 
outside world in coffee, rice, and other local products, and many young 
men spent periods working on the nearby Lupa gold-diggings. Some 
1 6 per cent, professed Christianity, and there were demands for more 
schools and trained teachers. Such is the group with whose marriage 
and kinship relations we are concerned.' 


The basic unit of Nyakyusa society is the elementary family consisting 
of a man, his wife, and children. All families begin as elementary 
families and some remain so, but polygyny is permitted and honoured 
and the majority of men are able, as they grow older, to marry other 
wives, thus linking together a number of elementary families in a 
compotmd family. 

There is no evidence of any marked difference in the survival rate 
of males and females, but there is a difference of ten years or more in 
the average marriage-age of girls and men, and it is this differential 
marriage-age which makes polygyny possible. A legal marriage is effected 
by the transfer of cattle from the groom to the bride's father, and the 
system is linked with a late marriage-age for young men and with a 
privileged position for the older men. Many men have not cattle with 
which to marry until they are well over 25, while their fathers may be 
able to afford more than one wife. The tax registers show in part of the 
district, among 3,000 men presumed to be 18 years or over, 34 per cent, 
bachelors, 37 per cent, monogamists, and 29 per cent, polygynists, and 
investigation proves that, generally speaking, it is the young men 
who are bachelors, and the men over 45 who are polygynists. Few 
commoners have more than 7 or 8 wives, but a claief may have as many 
as 40. 

Within a compound family each wife and her children form a distinct 
segment. Each wife has either a hut to herself or else a separate alcove 
in a long hut with partitions, and there she has her fire for cooking. 
She fetches her own firewood and water, or sends an unmarried daughter 
for them. At meal times each mother eats with her own daughters and 
young sons apart; more rarely one or two women, either co-wives or 
neighbours, choose to eat together, but this group probably does not 
include all the wives of one man, for only those who are particularly 
friendly will join. Each mother provides food for her own sons, who, 
after 10 or ii, usually eat with friends of their own age, the group of 
age-mates visiting the mother of each member in turn. The husband 
eats alone or with neighbours of his own age, his wives each sending 
him a dish or taking turns to cook the whole meal. 

' For an outline of Nyakyusa society see Godfrey Wilson, 'An Introduction to 
Nyakyusa Society', Bantu Studies, x, 1936, pp. 253-91. 


The separate identity of each segment or house of the compound 
family is also reflected in the family names. Every man has three 
family names, that of his paternal grandfather, that of his father, and 
that of his mother, and a girl is referred to by her mother's family 
name as well as by that of her father. Thus the mother's name groups 
together the full brothers and sisters within a compound family, and if 
an individual in detailing his or her genealogy says, 'We were so many', 
the implication always is that there were so many children the offspring 
of one mother, sharing the mother's name. 

Among commoners the first wife betrothed to a man while she is 
yet unmarried' ranks as the great wife {unkasikulu), and her eldest 
son is her husband's ultimate heir, inheriting the bulk of the family 
property. She herself takes precedence in ritual and she is entitled to 
the respect of her co-wives, but she has no authority over them, and 
the other wives are not ranked in order. A wise husband, it is said, 
divides his favours equally, visiting his wives in turn, and dividing what 
milk or meat there may be fairly between them; but in practice every 
man is thought to have a favourite with whom he sleeps more often, 
and to whom he gives more fine food or ornaments, than to the others. 
Often the latest bride ousts a former favourite and there is bitter 
antagonism between them — accusations of witchcraft against co-wives 
are frequent. Polygyny is thus correlated with jealousy and competition 
which involve a subordinate position for women in relation to their 

In a homestead {akaja) there commonly live together only a man, his 
wives, and their young children ; sons, as we have seen, build for them- 
selves at a little distance before they reach puberty ; they work with their 
fathers and are cooked for by their mothers until they marry, but they 
live in a separate village. Girls also leave home early, for they are often 
betrothed and begin to visit their husbands from the age of 8 or 9, 
moving permanently to their husbands' homesteads at puberty. A man 
may build for an elderly mother in his homestead, but the obligation 
of a widow to go to the heir is insisted upon at least until she be past 
child-bearing; only in Christian families is it usual for a middle-aged 
widow and her son to live together. Occasionally a rich man with 
several wives may build two homesteads in different parts of the 
country, but this again is unusual except with chiefs. 

Inheritance and succession in Nyakyusa society are dominantly 
patrilineal; genealogies are commonly traced farther in the male than 
in the female line, and economic and ritual co-operation extend farther 
among agnates than among matrilineal kin. Agnatic lineages of a span 
of three or more generations are concerned with the control and exchange 
of cattle, with inheritance, and with certain rituals, but for other purposes 

' A divorcee ranks lower than such a girl even although marriage cattle have 
been given for the divorcee before an unmarried girl is betrothed. 



it is not strictly lineages, but rather groups of cognates with a patrilineal 
bias, which co-operate. The Nyakyusa have no word for a lineage. 
All a man's kindred — the group of his cognates — make up his ikikolo. 
He distinguishes between his father's kindred {ikikolo kya tata) and his 
mother's kindred {ikikolo kya juba or abipwa), but both groups include 
relatives in the male and female lines; a child may participate in a 
ritual in his mother's sister's or his father's sister's family, as well as 
in those of his father and his mother's brother. The importance of links 
through females is brought out by the fact that at every point full 
siblings are held to be more closely related than half-siblings. For 
example, marriages between the descendants of a common great- 
grandfather are much disliked, but they are considered less bad if only 
one great -grandparent, not two, are common, i.e. if the related grand- 
parents are children of the same man by different wives.' 

The span of lineages and groups of cognates co-operating varies 
somewhat with particular circumstances. Among commoners, lineages 
of three generations control cattle, inheritance, and prayers to the 
ancestors. Agnatic genealogies include six to nine generations, and 
genealogies of cognates for four generations are practically important, 
since marriage between the descendants of a common great-grandparent 
is disapproved, and when such marriages occur those concerned are 
made to drink protective medicines. Between the descendants of a 
common grandfather marriage is impossible,^ and this is the group 
which, as we shall see, regularly co-operates. 

Chiefs' genealogies occasionally include as many as nineteen genera- 
tions traced in the male line — an old chief Mwandesi, famous as an his- 
torian among the Nyakyusa, gave us such a genealogy — but more often 
they include no more than twelve or thirteen generations. Wider groups 
of kin do in fact co-operate among chiefs than among commoners, for 
the chiefs are reponsible for communal rituals in which the agnatic 
descendants of a common great-great-grandfather participate, and for 
the ordinary domestic rituals more distant kin are summoned than 
among commoners. Should a commoner's family claim kinship with a 
chief, that link is traced farther than a link with commoners; thus if 
the father is a commoner and the mother belongs to a chief's family 
there may be wider co-operation with her kin than with his. This is 
very noticeable when it comes to an analysis of the individuals avoided 
by a daughter-in-law in different kinship groups. 

' Since divorce was rare in the traditional Nyakyusa system and the children 
of an inherited widow counted as children of the deceased, the children of one 
woman were practically always classed as full siblings. Should a woman have 
children by unrelated men, however, their grandchildren might possibly marry. 

^ There is a tradition among some sections of the Nyakyusa that very long 
ago they practised one type of preferential cross-cousin marriage — that between 
a man and his mother's brother's daughter. 



As we have seen, the Nyakyusa build not with relatives but with age- 
mates, and the age-span of a village (that is, the difference in age of its 
members) averages about five years. Now the spacing of births in 
Nyakyusa families was wide, for it was thought wrong for a woman 
to become pregnant again until her previous child was able to run 
beside her should the war-cry be shouted. No woman, it was argued, 
could flee carrying two children, therefore the elder child must be 
able to run before she was burdened with another. Hence four or five 
years between births was normal. This meant that full brothers (except 
for twins) were seldom in the same age-village, though half-brothers 
were often together. Since peace has been established births are less 
widely spaced, three years between them now being considered proper, 
so a pair of full brothers may well join the same village; but still the 
very close bonds which link together all the sons of one mother and 
father rarely include common residence in one village. It is, however, 
usual for afHnal relations (abako) to inhabit the same village, for it is 
thought good to marry the sister or daughter of an age-mate. 

The members of one chiefdom are not a kinship group any more than 
are the members of one village, and there is no myth of their common 
descent, but two factors tend to keep relatives within a chiefdom. First, 
a man has higher status in the chiefdom in which his father and paternal 
grandfather lived than he has outside it. Where his ancestors lived he is 
umwilima — an owner of the soil — a member of an old family, and as such 
entitled to greater respect and support from his neighbours than he will 
receive anywhere else. Old men explain that before the coming of the 
Europeans an 'owner of the soil' had much greater security of life and 
property than a stranger, for the latter, not being defended by his 
neighbours as one of themselves, was always liable to persecution by 
the chief. His wife might be taken by the chief and he himself murdered, 
or his cattle seized on some slight pretext. Moving from one chiefdom 
to another always occurred, but in recent years it has greatly increased. 
This is partly because land is becoming scarce, and men move to seek 
better gardens, but mainly because the traditional methods of dealing 
with accusations of witchcraft have been prohibited, and when a man 
suspects that he or his family or cattle are being attacked by witchcraft 
he thinks his only safeguard is to move. Probably 80 per cent, of 
Nyakyusa men now move at least once during their lifetime, but of these 
many return later on to their home chiefdoms, for the prestige of being 
an 'owner of the soil' is still a strong inducement to do so. 

The second factor which tends to keep relatives within a chiefdom 
is that marriages within a chiefdom are preferred to those beyond. 
Marriage with a woman of a neighbouring chiefdom was permissible 
but beyond that it seldom occurred, since the danger of travelling made 


it impossible for a wife to visit her parents. To-day the range of marriage 
is increasing, but parents still object to their children taking partners 
from forty or fifty miles away, on the ground that such a distance 
prevents that easy exchange of visits and gifts between in-laws which is 
held to be essential to a successful marriage. 


Relations within and between the kinship groups we have defined 
are articulated in co-operation in cultivation and the sharing of food, 
in building, in the passage of cattle, and in participation in common 
rituals; these activities we must now consider. 

In an elementary family husband and wife co-operate in cultivation, 
the man hoeing, his wife sowing, weeding, and reaping. Each wife in 
a compound family has her own fields in which she works with her 
unmarried daughters. Unmarried sons hoe with their fathers, working 
in some households only in the fields of their own mothers, while in 
others they work as a group with their father, hoeing each woman's 
.plot in turn. Often a boy brings his friends — usually only two or three, 
but sometimes as many as ten — to work in his mother's fields, and he 
in turn will assist these friends. Those who have been w'orking together 
always eat food prepared by the woman whose field they have been 
hoeing, for the produce of a woman's fields is used to provide food for 
her husband, her own children, herself, and any others who have 
been working in her fields for the day on which they have worked. The 
relative rights of husband and wife to dispose of crops vary with the 
different crops,' but no co-wife has any rights over another's food. A 
woman may, and often does, ask her co-wives to help her in her fields, 
and sends them presents of cooked food from her own fire; she may 
help them and receive presents in return ; but they have no rights over 
her fields or her food. 

From the time he is betrothed a young man begins to work in the 
fields of his mother-in-law (for so she is called from the time he has 
handed over the betrothal gift), dividing his time between her fields 
and those of his own mother, and bringing friends to work in both 
sets of fields. Often also he hoes a field near his own house which his 
'mothers-in-law' (that is, his betrothed's mother, her co-wives, and 
possibly some of her sisters or neighbours) plant and weed and reap. 
The crop from this field belongs to his mother-in-law until his betrothed 
reaches puberty and comes to live with him permanently as his wife. 
Then the field is hers, and though her 'mothers' come to help her in the 
field the food belongs to the young people. 

After marriage the young husband continues to help both his father 

' For an analysis of these rights see Godfrey Wilson, The Land Rights of 
Individuals among the Nyakyusa, 1938, Rhodes-Livingstone Paper No. i, 
pp. 16-29. 


and father-in-law in hoeing, while both his own 'mothers' and his 
wife's 'mothers' help her with planting and weeding and reaping; 
gradually, however, the amount of such co-operation diminishes. A 
man spends less and less time in the fields of his mother and mother-in- 
law as he grows older and has an increasing family of his own to support, 
while the help from his mother and mother-in-law diminishes cor- 
respondingly. As one newly married informant explained: 'My wife's 
mothers still help her a lot in cultivation; they say, "She is a child, let 
her dance and enjoy herself." But when a girl gets older and has borne 
one or two children, then her mothers stop helping her so much. They 
say, "We helped you before because you were only a child. Now you 
are a woman, do your own work." ' 

In building there is co-operation of the same sort. A young man helps 
his father, bringing with him a group of age-mates, if a hea\y job is on 
hand, and similarly he works with his father-in-law. His mothers and 
mothers-in-law in their turn help him with the women's tasks of carrying 
bamboos and grass, and plastering. We once watched a party of twenty- 
one women and girls, 'mothers' of the wife of a chief, come to help their 
'daughter' build. 

To grasp the significance of all this co-operation it must be realized 
that the Nyakyusa get the great bulk of their food from their fields, and 
that skill and diligence in cultivation bring prestige to both men and 
women. House-building is also very important in this climate of tor- 
rential storms and cold mists. The rainfall (averaging 100 to 130 inches 
in the year) makes substantial houses a condition of survival. 

But not only do Nyakyusa relatives work together in field and in 
homestead, they are also continually handing on from one to another 
their dearest possession, cattle. Indeed,- the giving of cattle is involved 
in all the closest relationships, and their reciprocal transfers of cattle 
are what hold families together. Cattle are prized not only because they 
provide two favourite foods, milk and meat, a supply of which makes it 
possible for a man to feast his neighbours, but also because they are 
necessary to an honourable marriage and a pious funeral. 

Cattle are not the exclusive possession of one individual ; rather they 
are property in which all the male members of a lineage of a span of 
three^ generations have certain rights. A man who is the senior son of 
a senior son, and whose grandfather's and father's full brothers are all 
dead, controls not only the cattle inherited from his father, and those 
from the marriages of his own full sisters and daughters, but also those 
from the daughters of his full brothers and the daughters of his sons. 

' Since the average marriage-age of men is about 25, agnatic lineages of full 
brothers of more than three surviving generations of married men are rare. 
Agnatic lineages of half-brothers of four surviving generations of married men 
are common enough, but exchange of cattle is not maintained to the fourth 
generation between the descendants of half-brothers. 


At the same time he is responsible for providing marriage-cattle for his 
sons and sons' sons, and for his younger full brothers and their sons 
and sons' sons. He is also responsible for providing one or more cows 
to kill on the death of any dependant on whose account he provides or 
receives marriage-cattle, and on the death of the husband or wife of any 
of these. On the other hand, the rights of disposal of the 'father' of the 
lineage are limited, firstly, by the strong claim every group of full 
brothers has to the cattle from their own full sisters and daughters, and 
secondly, by the practice of giving cows to younger brothers and sons 
over which they have ultimate rights. Certain individual rights thus 
exist, but the unity of a group of full brothers and their interest in a 
common stock of cattle are maintained through three generations. 

Relations between half-brothers vary somewhat. Those whose 
mothers are sisters, or otherwise closely related (for, as we shall see, 
it is common for a man to marry kinswomen), share the same reciprocal 
rights over cattle as full brothers, and half-brothers who are not 
linked through their mothers may yet be linked by cattle. As we have 
noted, each group of full brothers has a special claim to the cattle that 
come from the marriages of their full sisters, but their father or grand- 
father disposes of them as he wishes. If he gives some of their cows to 
a half-brother he thereby creates between the eldest of the group and 
this man the special relationship of 'milking each other's cows' [uku- 
kamanila), and it is expected that the cows given will in time be returned 
by this half-brother from the marriage-cattle of a full sister or a daughter 
of his own. A similar link is established between half-brothers if a father 
uses the cattle from a daughter's marriage to marry another wife himself. 
The new house thus begun is linked to the house from which the cattle 
originally came, and at least one cow must be returned when the eldest 
daughter of the new house marries. 

The exchange once begun goes on; as their sisters and daughters 
marry, linked half-brothers continually give each other cows; between 
their respective sons also the exchange usually continues, but between 
their grandsons it lapses. Reciprocity is expected and can be legally 
enforced, but it is not insisted upon so long as the two men concerned 
are friends, and its legal enforcement at once breaks the relationship. 
If two such half-brothers (or their sons) quarrel one may take legal 
action for separation against the other. The cows that have been handed 
over by both sides since the beginning of the relationship (perhaps 
twenty or thirty years before) are counted up, and one or the other is 
ordered to pay over the balance to make the numbers equal. After this 
the two are 'no longer kinsmen', they have no mutual obligations, and 
do not attend each other's funerals. 

It is usual for a father to create this relationship between his senior 
son and the eldest son of each of his other houses, so that by this 
exchange of cows the family is held together. Its continuance depends 


very largely on personal friendship. Good relations between half- 
brothers and their sons are valued, but quarrels leading to legal separa- 
tion occur. By contrast, quarrels between full brothers seem rarely to 
be serious and never to go to the length of a lawsuit; the unity of a 
group of full brothers is in fact maintained by their distinction from, 
and in some situations by their opposition to, their half-brothers. 

The separation of each house and the importance of links between 
houses through 'milking one another's cows' is apparent in the rules of 
inheritance. A man's heir is his next living full brother, or failing a 
full brother, a half-brother linked through their mothers. Should such 
an heir be lacking, the eldest son of the dead man inherits, provided he 
is grown up. If he is still a child a half-brother linked by cattle inherits, 
failing that a half-brother with whom there is no such bond. 

The heir, if he be a brother, takes all the widows of the deceased; 
if he be a son he takes all save his own mother and her kinswomen who 
must go to the son of another house. Whether he be brother or son he 
inherits the cattle of the deceased, and has the right to move into his 
homestead and take his gardens if he chooses to do so. He succeeds to 
the position of 'father' (tata), and as such is not only entitled to the 
respect and obedience of his younger brothers and brothers' sons, but 
also has the same obligations to them as the dead man had. The heir is 
avoided by those affines who avoided the dead man; often he takes the 
dead man's name. Indeed so close is the identification that even the 
barrier between succeeding generations is leapt, and a son who is his 
father's heir may be welcomed to the village of his father as the equal of 
his father's age-mates. 

Now this form of inheritance operated in a society in which there was 
little opportunity for individuals to acquire wealth. Occasionally a poor 
man would hoe for a chief and, after many seasons' work, gain cattle 
with which to marry; young men also may sometimes have acquired 
cattle in war. But the great bulk of property was family property v.'hich 
every male, provided he lived long enough, enjoyed in his turn. The; 
eldest son of an eldest son married earlier than did his juniors, and if 
he lived he acquired more wives and begat more children than they 
did, but each man lived in the expectation of one day enjoying the 
possessions, in wives and cattle, of his seniors. 

Since the coming of the Europeans the situation has changed. Op- 
portunities for the individual to acquire wealth either in employment 
or by the sale of produce are considerable, and the property of an 
energetic and capable man is often very much greater than that which 
he has inherited. Correlated with the change in the economy there is in 
process a change in the traditional law of inheritance. A younger full 
brother still has the right to claim his elder brother's property on the 
latter's death, but not uncommonly he relinquishes this right in favour 
of the deceased's senior son (provided that son is grown up) on the 


ground that if he takes his elder brother's property, then, when he 
himself dies, that brother's senior son will claim all that he, the uncle, 
held, including property he earned himself as well as that which he 
inherited. It was explained that if a younger brother, who already has 
considerable property in his own right, accepts the inheritance of an 
elder brother, there will almost inevitably be quarrelling between their 
sons over the division of the property when the younger brother dies. 
This tendency towards the direct inheritance of sons is further em- 
phasized in Christian families, where there is no polygyny, and con- 
sequently no group of wives to be inherited, and where the right of the 
one widow to refuse to be inherited is recognized. 

We have shown that a man's heir is commonly either his own younger 
full brother or else his senior son; now a distinction must be drawn 
between the positions of these two possible heirs. A man's younger 
brother inherits all his powers and responsibilities, and exercises 
authority over all the dead man's children; but when the heir is the 
senior son his authority is more limited. He is without a rival in the 
position of 'father' to his younger full siblings, but in relation to his junior 
half-siblings that position is shared with the eldest of each group of full 

A separate inheritance for each house is not so clearly marked among 
the Nyakyusa as among the Nguni people of the south. Though gifts 
of cattle are sometimes made to individual sons, the estate is not 
apportioned to each house, nor are the cattle of each house milked for 
the benefit of that house, as is common in Pondoland. Neither are land 
rights inherited by houses, for all the land falls to the heir who moves 
into the dead man's homestead. Indeed inheritance of land rights was 
relatively unimportant, since village boundaries were redrawn and 
homestead sites and most arable lands redistributed in each generation, 
at the great 'coming out' ceremony.' Nevertheless the eldest son of each 
house inherits certain rights on the death of his father and of his father's 
full brothers. He then gains the right of disposing of the cattle of his 
own full sisters, of his daughters, and of his full brothers* daughters. 
As we have seen, a senior son who inherits from his father (or father's 
brother) may or may not exchange cattle with each group of his half- 
brothers; if he does so at all his dealings are with the eldest of each 
group, who receives the cattle for full sisters and daughters of the 
group, to whom his younger full brothers look for cattle to marry, and 
for whom they hoe. Thus in each generation there is some separation 
of property between the different houses of a compound family. 

The precedence of the senior son appears most markedly in ritual, 
for although in minor matters the eldest of each group of brothers may 
officiate as priest, whenever serious trouble comes the prayers of the 
senior brother must be sought on behalf of all his junior half-brothers, 

' Cf. Godfrey Wilson, The Land Rights of Individuals among the Nyakyusa. 


unless they have been legally separated from him. So much for the 
control of cattle within the lineage, and their part in binding together its 

The exchange of cattle is also one of the main bonds between 
lineages. A legal marriage is effected by the handing over of cattle — 
traditionally i to 3 head, now 5 to 20 — by the representatives of the 
bridegroom to the father of the bride. It is the transfer of cattle which 
makes it possible for a woman to bear legitimate children, and which 
gives the husband legal control of these children. If a father refuses to 
accept cattle on behalf of his daughter she cannot be married and is 
greatly disgraced, since it is considered very shocking for an unmarried 
girl to bear a child. Indeed, such an occurrence is still rare. If no cattle 
pass there is no relationship {ubukamu), the Nyakyusa say, between the 
husband's people and the wife's people. 'A wife for whom cattle have 
not been given is not my relative {unkarmi)\ and 'since you (Europeans) 
do not give cattle at marriage where does the relationship (ubukatnu) 
come from? With us relationship is cattle {Uswe ubukamu syo nombe).' 
Such were typical Nyakyusa comments on the passage of cattle at 
marriage. And should a woman make a runaway marriage she cannot 
visit her parents at all, not even to attend a funeral, until her husband 
takes a cow to her father to create relationship. 

In fact, a legal form of marriage without cattle did exist traditionally, 
but the status of those who married in this way was much lower than 
the status of those who married with cattle. Informants explained that 
since in the old days cattle were scarce and owned only by the richer 
families, a father might agree to the marriage of his daughter in return 
for labour. Tf a poor man had hoed for me, the father, and built a house 
for me, then I said, "I am satisfied; because he does not speak proudly, 
he was my servant, I'll give him my child".' The consent of the bride's 
father made the union a legal one, but some, at least, of the children, 
including the first-born daughter, belonged not to the husband but to 
his father-in-law, and the husband was referred to scornfully as 'a cock' 
since 'chickens are the hen's, the cock goes about alone'. Thus marriage 
by service was traditionally an alternative, albeit a less honourable one, 
to marriage with cattle. 

Even when cattle are given at a marriage the son-in-law is expected 
to do a certain amount of hoeing for his father-in-law, as we have seen, 
and a connexion is made between his diligence in hoeing and the de- 
mands for cattle. If he hoes satisfactorily he is not pressed for the cattle 
outstanding, but if he is lazy they are demanded very quickly. And 
however many cattle he gives, hoeing is spoken of as an essential part of 
his obligation to his affines. Indeed nowadays, when a young man in 
paid employment is unable to hoe, he often gives two or three shilHngs 
to his father-in-law in lieu of labour. 

A legal marriage was commonly (and now is always) created by the 


transfer of cattle, and it is dissolved by the return of those cattle. 
Traditionally a husband could claim back all his cattle together with 
their progeny if the wife leaving him had borne no children, but if she 
had borne even one child one cow was left, for, the Nyakyusa say, 'If we 
left no cows the children, when they went to visit their mother's people, 
would be in a difficulty. Their mother's people would say to them : "You 
are not our relatives. Where are your cows?" ' Nowadays, if a woman 
has borne several children, two cows are left, but there is no exact 
balancing of the number of children against the number of cows left. 
And since 1935 calves have no longer been returned; the difficulty of 
tracing all the calves, when an increasing number of cattle is being 
given at marriage, was such that the British Administration pressed 
the chiefs and councillors to agree to a change in the law regarding 

No matter who has initiated the divorce the children belong to the 
husband, and he can even claim the child yet unborn, if the divorce 
takes place while his wife is pregnant. Did he not give cattle for her? 

The importance of cattle in creating a marriage is shown by the fact 
that formerly a husband might be compelled to divorce his wife solely 
on the ground of his poverty, irrespective of their personal relations. 
Traditionally, when a fine was imposed for theft or manslaughter, a 
man might be compelled to divorce his wife in order to pay the fine, 
or if his sister with whose cattle he had married left her husband, he 
might have to divorce his own wife in order to return the cattle to his 
sister's husband. Old men explained that a father might refuse a young 
man reputed to be a thief, or quarrelsome, 'because the father did not 
want that man's cows in his homestead'. The father thinks to himself, 
'If this youth's cows are finished they (his creditors) will come and take 
those he gave me. They will say, "Fetch back your daughter and return 
the cows." ' Nowadays the courts have no authority to compel a divorce 
in order to enable a husband to pay a fine, but there is still the feeling 
that 'a man's property is with his father-in-law' and may be recovered 

Despite the possibility of divorce on account of poverty marriage 
was traditionally stable. Informants are unanimous on this point. The 
sanction against adultery was spearing or impaling the man. A husband 
finding his wife with another man speared him, or, if the couple had 
run away to another chiefdom, he followed them up, along with two or 
three of his own kinsmen, and sought to kill the adulterer and bring his 
wife back. Sometimes fleeing lovers got away, but often they were over- 
taken by the enraged husband and his kinsmen. An adulterer could hope 
for no support from his neighbours if his lover were a woman from his 
own chiefdom, and even if she came from another chiefdom he was not 
secure, for his fellow villagers might drive him out lest they should all 
be attacked by the injured village. 'Men were afraid; they said, "An 


adulterer brings war." ' Hence, the Nyakyusa say, men seldom ran off 
with married women. Divorce which did not begin with an elopement 
was equally rare. A woman could secure a divorce from her husband for 
consistent ill treatment, if her kinsmen agreed, but they were slow to 
do so ; more often her father and brothers would catch the husband and 
beat him for ill treating their daughter and sister; and if a wife behaved 
badly her husband would beat her rather than divorce her. 

To-day the situation is very different. More than half the pagan 
women have been divorced at least once (some of them three or four 
times) and these divorces are commonly initiated by running off with 
a lover. An attack on an adulterer (not caught in flagrante delicto) is 
treated as a criminal assault or homicide; instead adultery is punished 
by a fine of three cows, two being paid by the man found guilty of 
adultery to the husband of the woman concerned, and one by the father 
of the woman to her husband, since the father is regarded as being in 
some measure responsible for his daughter's conduct. 

But formerly divorce was relatively infrequent, and even if a husband 
or wife died, the affinal relationship between their families was not 
dissolved; instead a younger brother or sister took the place of the 
deceased. A widow w^as inherited by her husband's heir, and this right 
of inheritance was insisted upon even though the 'widow' was a child, 
not yet initiated though betrothed. For a widow to refuse the heir was 
comparable to deserting her husband; if she ran off the heir pursued 
her as her husband would have done, and indeed he was regarded as 
her rightful husband. Nowadays if a widow refuses the heir and returns 
to her father no damages can be claimed by the heir — he only gets back 
the cattle given for the woman — but still there is a very strong feeling 
that widows should accept the heir as husband. We heard one father, 
persuading his daughter to remain with the heir, say to her : 'See, he also 
is a husband who married you,' And in fact many pagan widows agree 
to be inherited because they wish to remain with their children ; if they 
refuse the heir they are parted from their children like divorced women. 
Only in Christian families may a young widow remain with her children 
without accepting the heir as her husband, and without causing her 
father to lose the cattle given for her. And even among Christians a 
widow who remarries must leave her children with their father's heir. 

If a wife dies she is very often replaced by a younger sister (unuguna) 
or by a daughter of her brother (umzvanasenga). Should she die childless 
such a substitute must be provided, or else all the marriage cattle have 
to be returned. If she has borne children a sister is, in Nyakyusa eyes, 
the best person to mother these children; indeed it is often the children 
who go to their maternal grandfather or uncle to ask for 'another mother 
to care for us'. In the old days if the wife who died was young no further 
cattle were asked for the sister given in her place, but now the father 
often claims marriage-cattle all over again. Fathers argue thus: 'These 


cattle have gone with her who died. Do you wish men to say of the 
younger sister, "They gave no cows for her, they just took her"? ' 
Nevertheless a smaller number of cattle is often accepted for the second 
sister than for the first. 

And here we come to the crux of Nyakyusa ideas of marriage : relations 
between affines (abako) are ideally permanent — a divorce should never 
occur; a dead husband should be replaced by his heir, a dead wife by 
her younger sister or brother's daughter; the individuals concerned may 
change but the relationship between the families remains. This idea of 
permanence is pushed to the point at which some say that whenever a 
woman dies she should be replaced, no matter what her age is; in fact 
women who have borne children are not replaced if they die after their 
own husband and his brothers have died. But marriages between two 
families already linked are strongly encouraged, provided that it does 
not mean marriage between the children of a common grandfather. 
Very frequently a man does not wait for the death of his v/ife to marry 
her younger sister or brother's daughter. Ukiisakula — to bring a younger 
relative to one's husband as wife — is the form of polygyny generally 
preferred and is the mark of successful marriage relations, for a man 
will not marry the younger kinswoman of a wife with whom he is on 
bad terms, nor will a wife consent to bring her little sister or niece to him 
if he is a harsh husband. Instead she will advise her father to refuse 
the offer, should the husband suggest it, for the initiative in such a 
marriage may come either from the husband or from the elder sister 
already his wife. 

Sometimes two brothers marry two sisters, or a young man may seek 
a wife in the family from which one of his stepmothers came. 'He 
marries from where his father married.' Now, as we have mentioned, 
there is a tradition among some of the Nyakyusa of a former custom of 
cross-cousin marriage (with the mother's brother's daughter), but one 
informant argued that it had never been real cross-cousin marriage, but 
only marriage with the half-brother of a father's sister's son. T call 
my mother's brother's sons my cousins (abatani), but his daughters are 
my sisters {abilumbu). It never was the custom to take them to wife, but 
it was, and still is, the custom to give them to my half-brothers. I am 
the owner of these women [ndi mzvene bakikiilu aba), I may take one of 
my brothers by a different mother to my mother's brother's place and 
say to my brother, "There is the girl for you." ' Whatever the traditional 
practice, cross-cousin marriage does not now occur, but marriage with 
the half-brother of a cross-cousin does, though less frequently than 
marriage with a wife's brother's daughter. It is to be noted that marriage 
with a wife's brother's daughter is incompatible with cross-cousin 
marriage, and that it operates in a system in which the older men have 
first right to women. The girl who, under a system of preferential cross- 
cousin marriage, would become the wife of a young man, often, under 


the present Nyakyusa system, becomes the wife of that young man's 

The suggestion that a man has some rights over his mother's brother's 
daughters fits in with other facts, for the passage of cattle creates a 
special relationship not only between a husband and his affines but also 
between his children and his wife's people, more particularly with the 
brother who benefited from her cattle. From this uncle (umzvipzca) a 
nephew, when he grows up, can claim a cow (ijabzvipzva), and, as one 
informant put it, 'This custom of getting the cow, ijahwipiva, is to show 
the relationship of the mother's people to the young man and his father, 
and also to the mother of the young man. Especially we think it is to 
honour their daughter, the mother of the youth; it is to say, "Thank 
you, we have received cattle on your behalf, you have borne a child 
for us." ' Mother's brothers who did not receive cattle on her account 
are abipzva too, but 'that which makes this relationship important is 
cattle. It is an insignificant relationship if the uncle is not milking the 
cows from the young man's mother.' Because of the cows a nephew 
can take food, especially the prized thick milk, in his uncle's homestead, 
without asking permission, and should he quarrel with his father it is 
to his linked mother's brother that he goes. He also has certain rights of 
inheritance, for should the uncle to whom he is linked die without a 
brother (own or classificatory) or a son to inherit, the nephew claims 
the cattle given for his mother. 

Between the sons of brother and sister (abatani) there is freedom and 
friendship as between equals, and a mother's brothers' daughters are 
almost as sisters; we heard no suggestion of unusual jocularity between 

We have explained the Nyakyusa view that relationship [ubukamu) 
between two lineages is initially created by the passage of cattle. It is 
maintained by co-operation in cultivation and building, by the occasional 
exchange of gifts of food, and above all by coming to mourn and bring- 
ing cows to kill at each other's burials. Unless he is utterly destitute 
a son-in-law must bring a cow to kill at the death of his wife's father or 
mother, and his wife's father (or this man's heir) must, on his part, bring 
a cow to the burial of his daughter or her husband. Thus we come to 
the religious interdependence of relatives which is so important to the 
Nyakyusa that relationship is often defined by them in terms of atten- 
dance at each other's rituals — to say of a man 'we do not go to each other's 
burials' is to imply that he is not counted as a relative. 

The rituals in which relatives join are those celebrated at puberty 
and marriage, at birth (especially at a twin-birth), at death, and in case 
of sickness or misfortune, particularly if blood has been shed or if one 
of the family has been bound with ropes. Here we can only indicate 
very cursorily the nature of these rituals, for the description of them 
and an analysis of their symbolism would make a book in itself. 


The overt purpose of the rites in each case is fertiUty and health; 
neglect of them is believed to result in sterility, sickness of one type 
or another, and 'madness' (ikigili), or, more accurately, specific neurotic 
symptoms. Different pathological conditions are associated with the 
neglect of different rituals. From the point of view of kinship and mar- 
riage the essential fact is that relatives are believed to be mystically 
affected by the very fact of their relationship. Should a son not take 
part in the death ritual for his father he is in danger of going mad; 
should a sister's son not be called to drink medicines at the birth of 
twins to his mother's brother the nephew's children may swell up and 
die; should the wife of a murderer eat with the widow of the victim 
before purification both will begin to cough blood. It is not the same 
set of relatives which is concerned each time, for the range varies with 
the ritual. The previous experience of individuals also affects attendance, 
for once a person has gone through the death or twin ritual as a principal 
he is not required to take part again. 

A large group of cognates is concerned in a twin ritual. At one we 
watched, all the descendants of the father of the twins, all his full 
siblings and their descendants, his half-brothers and their agnatic 
descendants (but not his half-sisters), his father's full siblings and half- 
brothers and their agnatic descendants (but not his father's half-sisters), 
as well as his wife's brother and sister and their children, were treated. 
The wives of all the men concerned took part, but not the husbands of 
the women, and it was noticeable that while links through both males 
and females counted with nearer relatives, more distant connexions 
were traced only through males. For example, the children of full 
siblings' daughters were treated, but not the children of half-siblings' 
daughters — only the children of half-brothers' sons. The only distant 
relative connected through females who was treated was the baby of the 
daughter of the principal's father's sister's son.' 

At a burial a similar large group of cognates and their wives is expected 
to attend, along with the sons-in-law, sisters' husbands, and fathers- 
in-law of the deceased, if he be a mature man ; but when it comes to 
the death ritual {ubunyago) for purifying the chief mourners and driving 
away dreams of the dead from their minds, only a small group of kin — 
siblings, own children, and grandchildren — together with the widows 
or widower, of the deceased, is concerned. Similarly, after the shedding 
of blood it is only these relatives of the slayer and slain who drink 

One further principle which is necessary to an understanding of 
kinship and marriage emerges from a study of the rituals : it is the separa- 
tion of sexual activities and the parent-child relationship. The Nyakyusa 
believe that the sexual fluids are extremely dangerous to children, hence 
(they say) the restrictions on the parents of a young child sleeping 
' Possibly because infants are identified with their mothers. 


together. The mother may not conceive again until the child is weaned, 
and she must always be in a position to wash scrupulously before going 
from her husband to the child. 

At the same time a newly married wife who is 'hot from her husband' 
is held to be dangerous to her own parents, and she must not go into 
the inner part of her mother's house where the bed is, though before 
puberty she was free of the house; likewise a widow who has been 
inherited cannot eat with her parents until they have drunk a medicated 
beer together, lest they fall ill on account of the 'dirt' of her new husband 
on her. For a son to sleep with his father's wives is believed to be 
mystically dangerous to both men, and if, as sometimes happens when 
the father is old, permission has been given to the son to take those 
wives who will be his inheritance, elaborate ritual precautions are taken. 
And when the father does die one of the overt purposes of the death 
ritual is to make it safe for the heir to sleep with the widows. 

So strong is the feeling that the sexual activities of succeeding 
generations must be separate, that it is forbidden for a woman to bear 
children after the marriage of her son, and she must not risk conceiving 
a child from the time of her daughter's puberty until her daughter has 
conceived. If the mother 'oversteps' her daughter in this way it is 
thought that the latter will be barren. 

The separation reaches its greatest elaboration in the avoidance of 
father-in-law and daughter-in-law (abakamwana). A woman may never 
look at her father-in-law nor enter his house, nor meet him on a path, 
nor mention his name, or words like it. She avoids even the cow which 
looked into his grave, and the fiesh of a cock which was in his 
homestead, and the banana grove where her husband prays to his 
dead father. This avoidance of the father-in-law (ukutila twkamzvana) 
is extended to his brothers, half-brothers, and male cousins, the sons of 
full siblings of his parents. For example, a woman avoids the sons of her 
father-in-law's mother's full sisters and brothers, but not those of her 
half-sisters and half-brothers. It is extended in a modified degree to the 
father-in-law's sisters, half-sisters, and female parallel cousins, for all 
are abakamzvana. She cannot look at these women or greet them directly, 
though she does not fly at their approach as she does at the approach 
of her father-in-law or one of his brothers, own or classificatory ; to veil 
her face before a female unkamzvana is sufficient. And when the father- 
in-law dies his heir is avoided in his stead, though less strictly than the 
dead man was, and the wives of the heir (if he be the son of the deceased) 
go through a special ritual to enable them to meet their husband, who 
is now also their 'father-in-law'. 

But it is not only the father-in-law's siblings and heir that a wife must 
fear; though she can talk freely with her mother-in-law, the latter's 
brothers and half-brothers, and the husbands of her full sisters (but 
not of her half-sisters) are treated as fathers-in-law, though again the 


avoidance is less strict than it is towards the husband's own father. And 
the extensions apply to those identified with the daughter-in-law as 
well as to those identified with the father-in-law — a wife avoids not only 
her own 'fathers-in-law' but those of her sisters and half-sisters also. 

These general rules are modified by particular circumstance, for the 
range of avoidance varies somewhat with proximity, friendship, and 
social status. A parallel cousin of the husband's father, who lives at a 
distance, may well not be known and so not avoided, whereas a more 
distant relative with whom the family is friendly, or who is a chief, 
may be treated as a father-in-law's brother. All who are reckoned as 
kinsmen of the father-in-law and of his generation are in fact avoided; 
indeed the Nyakyusa themselves take the range of avoidance as a 
criterion of kinship. Often a man will say: 'I am related to so and so, 
though distantly; tny wives avoid him.' But avoidance shows relationships 
in consecutive generations only, for a woman conspicuously does not 
avoid her husband's grandfather and his siblings.' 

When the Nyakyusa are pressed to give their reasons for insisting on 
so strict an avoidance between father-in-law and daughter-in-law they 
relate a myth concerning a chief who looked on his son's beautiful 
young wife, and coveted her and took her. 'People thought this so bad 
that from that day they forbade father-in-law and daughter-in-law to 
look on one another.' And any familiarity whatever between father-in- 
law and daughter-in-law savours of incest or indecency to the Nyakyusa. 
As one old woman put it: 'If a girl looks on her father-in-law she 
compares him with her husband; she sees the physical resemblance 
between them. It is as if she had looked on her parents-in-law sleeping 

The great elaboration of avoidance is correlated with the age-village 
organization; were father and son to be near neighbours such strict 
avoidance between father and daughter-in-law would not be possible, 
and indeed the necessity for avoidance between them is given as a reason 
for fathers and sons building apart. Only among those Christians who 
relax the rules of avoidance are the homesteads of father and son found 
near one another. It is to be noted also that this emphasis on avoidance 
occurs in a society in which polygyny is approved, and in which the 
cattle with which to marry are mostly controlled by middle-aged and 
elderly men who customarily marry young girls. 

Between mother-in-law and son-in-law (abako) there is reserve and 
a measure of avoidance — he may never go into the inner part of her 
hut where her bed is, and the food she cooks for him will be brought by 
someone else — but the restrictions on his behaviour are slight compared 
to those imposed on his wife. 

' Here the Nyakyusa system differs from that of the Pondo. 



The forms of behaviour primarily appropriate to kinsfolk are extended 
in some measure to neighbours who, owing to the age-village organiza- 
tion, are very often not related at all. All one's father's mates in his 
village are 'fathers in the village' {batata pa kipanga), their wives are 
all 'mothers in the village' (abanna pa kipanga), and their children of 
opposite sex are 'sisters (or brothers) in the village' {abilumbu pa 
kipanga). Neighbours co-operate in herding, those who live next door 
to one another sending out their cattle to pasture together, and they join 
also in cultivating and building. We have shown that a young man 
working for his father or father-in-law is very often accompanied by a 
group of friends, neighbours from the boys' village, and with this group 
he eats after work. So also with the older men and women : for any major 
undertaking, such as hoeing or reaping a big field, or building a house, 
at least one work-party is usually held to which close neighbours as 
well as relatives are expected to come. An investigation of any working 
group will show that some of the members are there because they are 
relatives of the organizer, others because they are village neighbours of 
his own or of his son or son-in-law. 

But such economic co-operation stops short at the exchange of 
cattle. Very occasionally an intimate friend, who is also rich, may 
bring a cow to a funeral, but there is none of that circulation of cattle 
which binds together kinsmen and afRnes. Nor is the assimilation of 
neighbours to kinsfolk pushed to the point of village exogamy. Marriage 
within the village is and always has been permitted, though traditionally, 
it is said, a man would not readily marry the daughter of a next-door 
neighbour, with whom he ate, and with whose cows his were herded. 
For the daughter of a next-door neighbour is 'like your own daughter'. 
But there could be nothing against marriage with the daughter of a 
fellow villager living farther off, and marriages between the children 
of fellow villagers are felt to be particularly appropriate. 'It is good to 
marry a sister of the village', the Nyakyusa say; 'our fathers were friends 
and built together'. 

As in the material, so in the religious aspect, there is a measure of 
co-operation between neighbours. At the great events in life — marriage, 
birth, and death — neighbours are expected to show sympathy and to 
rejoice or mourn with their friends as may be appropriate. Not to attend 
a burial in the village is to incur suspicion of having killed the dead man 
by witchcraft; not to dance at a wedding feast is likewise unneighbourly. 
At the initiation of a girl her age-mates — half-sisters and neighbours — 
keep her company, women neighbours share with her senior female 
relatives the task of instructing her, and both neighbours and relatives 
are included in the party which formally takes her to her husband. 
When food and firewood are required to provide for the guests at 



initiation and marriage feasts, and at burials, neighbours as well as 
relatives contribute, food at a burial being provided on the first day by 
the bereaved household, on the second by relatives, and on the third 
by neighbours. 

But when it comes to the domestic rituals — the treatment with 
medicines and the prayers to the ancestors — then, with one exception, 
it is relatives alone who are concerned. No neighbour is purified and 
protected from dreams of the dead at the ritual after death, nor washed 
from the contamination of blood and ropes ; no neighbour is concerned 
in the fertility rites of initiation and marriage. Only in the fearful event 
of a twin-birth do next-door neighbours come with their children and 
their cattle to drink or be sprinkled with medicines, to save the cattle 
from purging and the humans from monstrously swollen legs. Even in 
this case fellow villagers living at a little distance and herding their cows 
in another group are not aflfected, whereas all relatives, no matter where 
they may live, must be treated. Thus neighbours play a small part in 
the domestic rituals though, as we shall see in the following section, 
they are believed to exert a mystical power to enforce obligations between 
relatives, including the performance of the rituals. 

We have dealt with the extension to fellow villagers of the forms of 
behaviour primarily appropriate to kinsfolk; it remains to note the 
part which kinship plays in the political organization. 


As has been shown, there is no all-embracing kinship bond between 
the members of one village or one chiefdom, and, according to tradition, 
chiefs and people are of different stocks; but relationship is traced 
between the chiefs of different chiefdoms, and by virtue of their 
relationship various groups of chiefs share in common rituals. The 
welfare and fertility of a country is held to be bound up with the health 
of its chief, and, just as a living chief has mystical power over his land, 
so dead chiefs are thought to have power over the countries they ruled. 
Thus half-brothers who divide their father's chiefdom pray together at 
the grove in which he is buried on behalf of both their countries, and in 
certain circumstances a number of chiefs combine to pray at the grave 
of a reputed common ancestor. The religious obligations of kinship are 
then a bond linking chiefdoms ; nevertheless, participation in common 
rituals did not in practice exclude war. 

Despite the fact that chiefs and people are believed to be of different 
stocks there is no cleavage between the descendants of chiefs and 
commoners, the descendants of the junior sons of chiefs being quickly 
assimilated into the group of commoners. The headmen of the age- 
villages (amafumii) cannot be the sons of chiefs or of former village 
headmen ; they are commoners par excellence, and the Nyakyusa insist 
that were son to succeed father in the office it would become a chieftain- 


ship. Thus, although there are hereditary chiefs, kinship does not play a 
dominating part in the political organization. 


We are now in a position to summarize certain general principles 
embodied in the Nyakyusa organization. It is clear that the older 
generation, especially the men, have a privileged position. The older 
men control cattle and labour; the majority of them have more than 
one wife, and the possession of many wives, together with unmarried 
sons and young sons-in-law who hoe for him, means that a man has 
at his disposal plenty of food; and food makes lavish hospitality, the 
foundation of prestige, possible. Along with the control of wealth goes 
religious authority, for not only do men have mystical power over their 
descendants, they are also the mediators between them and their 

The privileges of seniority accrue not only to the senior generation 
but also in some measure to a senior son — the eldest son of a great wife 
— and in a lesser degree to the eldest son of each house. Though in- 
heritance passes from brother to brother there is an ultimate, albeit 
delayed, efficiency of the principle of primogeniture. The senior line 
of any family is likely to be the most wealthy and the most numerous, 
for the senior son of a senior son is likely to marry earlier and have 
more wives than his juniors. In ordinary everyday speech the differences 
between senior (unkulu) and junior (unuguna) siblings of the same sex 
are marked. 

Secondly, we can observ^e a certain identification of siblings, half- 
siblings, and father's brothers' children of the same sex, all of whom 
are abakulu or abanuguna} The identification appears further in the 
extension of the term for father {tata) to his brothers, half-brothers, and 
father's brothers' sons, and a similar extension of the term for mother 
(juba), and in their reciprocal use of 'my child' (mzvatiangu) both in 
address and reference. It appears in the replacement of a dead husband 
by his younger brother, half-brother, or father's brother's son, and the 
similar replacement of a dead wife ; and it appears in the extension of the 
rules of avoidance to include all those whom the husband calls father, 
and whom the daughter-in-law calls sister. 

Even between siblings of the opposite sex (abihwibu) there is a 
measure of identification. A man's wife must be deferential to his sister 
(ugwifi), for this sister is very close to her husband. As one informant 
put it, 'our sisters are as men, those who married our wives'. On the 
same principle a woman avoids her husband's father's sister. One day 
an old woman in a crowd drew our attention to this fact: 'See, I do not 

' Mother's sifeters' children are referred to as brothers and sisters, not as 
cousins (abatani), but when it comes to the control of cattle, inheritance, and 
rituals they are not treated as siblings. 


go to greet them, my brother's son's wives. I am Hke the father of their 
husband.' And a girl is very respectful to her father's sister (unnasenga), 
even though the latter is a mere child ; she fetches and carries for her, 
and never touches her back to smear her with ointment, as she would 
a mother, for 'father's sister is father'. 

A sister is felt to be more nearly the equal and intimate of her brother 
than a wife can ever be — indeed one day the Nyakyusa, seeing a European 
woman brush a caterpillar off her husband's sleeve, laughed and said : 
'Why, she behaves like a sister! A wife would never take such liberties.' 
A father's sister may even be asked to arbitrate between her brother and 
his son, in case of dispute, which the wife and mother could never do. 
And the assurance which a sister has in dealing with her brother is 
extended in some measure to her children, a man being free with his 
mother's brother (umwipwa) in a way in which he cannot be with his 

But the tendency to identify brother and sister is limited at every 
turn by the sex difference. Brother and sister are much less familiar 
than brother and brother, or sister and sister; much more freedom is 
possible with a father's sister than with the father himself, much less 
with a mother's brother than with a mother herself. The sex difference 
appears vividly in the rules of avoidance: a woman does not greet the 
sister of her husband's father, or look her in the face, but she may be 
in the same company; whereas should word come that her father-in- 
law or his brother is approaching she disappears immediately; and she 
avoids her mother-in-law's brother only a little less strictly than her 
husband's father, but with her mother-in-law she can talk and work. 

We have shown that siblings, half-siblings, and father^s brother's 
children are grouped together, but at the same time very clear distinc- 
tions are made between each group of full siblings. How these appear 
in family names, in the distribution of cattle, and in inheritance has 
already been discussed (vide supra, pp. 114, 1 17-18). It remains to note 
that precise terms exist for the distinctions. Unkulu, unuguna, and 
ulilumbu {vide supra) may all be qualified thus : 

unkulu munna elder full sibling of the same sex. 

,, mhannabo senior half ,, ,, (mothers related). 

,, mimumba ,, half ,, ,, 

,, mbisabo father's elder brother's son (or daughter, woman 

Along with the grouping together of those of the same generation 
goes the separation of those of succeeding generations, which is carried 
to the point of territorial segregation. Fathers and sons ynust live in 
different villages. They do not eat and drink together, sharing those 
urbane conversations over a well-cooked meal which to the Nyakyusa 
are the essence of the good life ; in a word they do not ukwangala, for 


that is only possible between equals; it is not possible between men 
and women or betsveen fathers and sons. The separation is especially 
marked in the field of sfex. One of the reasons given for sons building 
apart is that they should not hear lewd conversation and see sexual 
play between their parents; for children to learn anything about sex 
from their parents is shameful ; such instruction must come from their 
immediate seniors, older brothers and sisters, not parents; and, as we 
have seen, elaborate precautions are taken to prevent any action felt to 
symbolize an association of the sexual activities of parents and children. 

This separation of the generations is broken by inheritance, for a 
senior son may inherit most of his father's wives and, joining the village 
of his father, be treated almost as his father. The parallel among women 
is the custom of replacing a dead woman by her brother's daughter 
should no younger sister be available. Among women, indeed, the 
differentiation between the generations is much less marked than 
among men, for co-wives are commonly of different ages, and a brother's 
daughter may well join her aunt as a junior co-wife. Even mothers and 
daughters will eat together. 

Unlike parents and children, grandparents and grandchildren (abisti- 
kiihi) are expected to be familiar. As one man explained when asked if 
his wife avoided his grandfather: 'No, he is my comrade. I am 
Mwakalambo and he is Mwakalambo.' And significantly the words 
'husband' (undume) and 'wife' (unkasi) are used between grandparent 
and grandchild of opposite sex. 

Lastly, we repeat that the effective range of Nyakyusa kinship varies 
with personal friendship, proximity, and social status, and that there is 
some correlation between range in the contemporary and historical 
moments, those who recite the longest genealogies recognizing the 
widest connexions among the living. 


Some indication has already been given of the types of social pressure 
which maintain these relationships. Now we shall consider more closely 
the nature of that social pressure. In all the kinship and marriage 
relations of the Nyakyusa we find that there are mutual economic 
obligations, and some of these are enforced by reciprocity, that is, if 
one party neglects its obligations, the other does likewise. 

Both boys and girls owe their parents, the Nyakyusa think, a material 
recompense for the trouble and expense of feeding and bringing them 
up ; while a boy, in addition, must make some return for the cattle his 
father gives him to marry with. A girl fulfils her obligations mainly by 
getting and staying married. Since marriage involves a transfer of 
cattle to her father, and divorce involves their return to her husband, 
it is above all by behaving well in the relationship of marriage, and so 
avoiding divorce, that a woman discharges her obligation to her parents. 


'If a daughter ran home in the old days', we were told, 'and said that 
her husband had beaten her, then her father sent to inquire among the 
neighbours about the quarrel ; if they said that it was she who was in the 
wrong, then her father would beat and scold her, saying : "I thought you 
were a grown woman and were supporting me, but instead of this you 
spoil my wealth." ' The word which we translate 'were supporting me' 
is a form of ukuswila which usually means 'to feed a child', and it thus 
brings out very precisely the element of economic reciprocity in the 
relationship of father and daughter. We asked our informant what was 
meant by 'spoiling my wealth'. 'Would not her husband come to fetch 
his cattle back (i.e. divorce his wife)?' he replied. 

The obligation to the parents which a daughter thus fulfils by an 
unbroken marriage is discharged by a son in the fields; he earns the 
cattle for his marriage and repays his parents for his food and upbringing 
by hoeing, and nowadays by earning money for them as well. The 
labour of an unmarried son is one of the most important sources of 
wealth in Nyakyusa society, and it is directly secured to his parents by 
his dependence on his father for marriage-cattle. A boy's economic 
value to his parents is increased by every year he stays unmarried, but 
the quicker a girl gets married the better for them. And it is an essential 
condition of the economic value of children that the marriage-age of 
young men is, on an average, ten years older than that of their sisters. 
For a father pays out as many cows for a son's marriage as he receives 
for a daughter's ; it is because he has time, between receiving and paying 
out, to keep the cattle and let them increase in his house that the transfer 
of cattle benefits him, while this interval of time is also filled by the 
useful labour of his son. Now the length of this interval depends partly 
on the diligence of the son in hoeing. For a son who is lazy a father will 
be slow to give marriage-cattle,' whereas for a diligent youth the father 
should provide cattle as soon as he can. And should the father, having 
cattle, not provide for his son's marriage, that son may take his strength 
elsewhere, going to hoe for his mother's brother, or even for a kinsman 
of his father, in the expectation that in due course the man for whom he 
has worked will repay him, like a father, by finding the marriage-cattle. 
Or nowadays he goes off to work for Europeans and keeps his earnings 
for himself. 

This example brings us to the kernel of our argument. Between father 
and son there is direct reciprocity, a balancing of labour against early 
nurture and cattle, and should one neglect his obligations the other may 
do likewise. Between father and daughter the case is somewhat different : 
the economic obligations of the father are largely fulfilled before his 
daughter marries, and the return she owes him is enforced by religious 
and conventional rather than by economic sanctions. Everyone agrees 

' For an example see Godfrey Wilson, 'An Introduction to Nyakyusa Law', 
Africa, x. i, Jan. 1937, p. 33. 


that a girl ought to stay married so that her father may benefit from her 
cattle; she is urged to do so by the fear of losing the power to bear 
healthy children, owing to the anger of her parents, and by the fear 
of being shamed by the refusal of her father to receive cattle on her 
account. She does not fear much direct economic loss; only indirectly 
in the food given her by her father to present to her husband and in the 
food and cow given to her son by her brother, may she be said to get 
some return for the cattle given on her behalf. It is noticeable that while 
the divorce rate has risen sharply, sons, for the most part, still hoe 
diligently for their fathers. 

In all the relations in which cattle pass there is direct economic recipro- 
city between the men who give and those who receive them; brother 
gives to brother in the expectation of a return and, should no return 
be made when the occasion for it arises, a legal separation with a careful 
balancing of accounts — a counting of cattle given and received — follows. 
Between son-in-law and father-in-law there is a similar balancing of 
obligations : the son-in-law works and gives cattle in return for a wife 
and children; if his wife leaves him, or dies, or is sterile, she must be 
replaced, or the cattle returned. Cattle killed at in-laws' funerals are 
carefully reckoned in the tally of those owing from one family to the 
other, and even in the exchange of gifts of food a rough balance is 

Between husband and wife also there are reciprocal economic 
obligations. Traditionally a wife planted, weeded, reaped, cooked, and 
did the housework, while her husband saw to the stock and hoed. 
Nyakyusa women are aware that one road to a man's heart lies through 
the kitchen, and good cooking and cleanliness are recognized means of 
becoming the favourite wife; while the diligent man who hoes ample 
fields for his wives is honoured by them on that account. Should a man 
fail to hoe for his wife she has grounds for divorce, while should she be 
lazy he is considered justified in beating her. 

This illustrates a second type of pressure very important in the old 
Nyakyusa society, the right of private people to use force. A father or one 
of his village neighbours might beat a child for doing wrong ; a husband 
beat his erring wife ; a father-in-law and his sons beat a son-in-law who 
was ill-using his wife, their daughter and sister ; and a husband, aided 
by his brothers, tortured and killed the seducer of his wife. If cattle 
were owing from a man in another chiefdom the creditor might go 
armed with one or two kinsmen to collect his debt. Such a use of force 
was considered perfectly legitimate, though for a junior to strike a 
senior, or a wife her husband, or for one to seize another's cattle without 
good cause, provoked further punishment. 

This right of using force without reference to any court is becoming 
more and more circumscribed. The man who beats his neighbour's son 
is liable to be summoned to court by his neighbour, and a Christian 


may be reprimanded by his Church even for beating his own wife 
when she has done wrong. The man who goes armed to collect Kis debts 
or who pursues the seducer of his wife and kills him is punished, for the 
right to use force has become the prerogative of the courts. As one 
Nyakyusa put it: 'Of old if a boy insulted me I'd beat him; now if I do 
so his father summons me to court and the judge says, "Are we not 

But it must not be thought that courts did not play a great part in the 
traditional Nyakyusa society. Constantly the Nyakyusa take quarrels to 
older relatives or respected neighbours for arbitration; this was the 
traditional custom, and traditionally there was a chief's court with 
power to enforce its judgements within the chiefdom. 

Quarrels between relatives are most often taken to senior relatives 
to be settled; disputes between brothers or a brother and sister, or 
between a man's wives or his children, are taken to the father, and one 
between kinsmen or brothers whose father is dead is referred to the 
senior kinsman of the lineage; but the exact relationship of the arbi- 
trator to the quarrelling parties does not matter so much as that they 
should both respect him, and indeed relatives may take their cases to 
a respected neighbour as they would do if they were not related at all;' 
it is only thought more appropriate to go to a fellow kinsman. What is 
considered bad is that a dispute between kinsmen should go to the 
chief's court; they should always settle the case privately if possible; 
though, as we have seen, half-brothers or parallel cousins who are on 
bad terms may come before the chief's court for a formal separation 
and sorting out of the cattle they have exchanged. 

Quarrelling between relatives is not, however, merely a secular matter, 
for pagan Nyakyusa (and many Christians also) believe implicitly in the 
mystical power of senior relatives, living and dead, over their descen- 
dants, and the neglect of kinship obligations which angers the seniors 
is believed to bring sickness or misfortune.^ This power is most com- 
monly spoken of as operating within the agnatic lineage, a father having 
mystical power over his children and sons' children, and the senior son 
inheriting power over his sisters and half-sisters, his junior brothers and 
half-brothers, and their children and sons' children. The anger of a 
father-in-law with his son-in-law is thought to affect the latter's child- 
ren, i.e. a man has power over his daughter's children, and in case of 
sickness may pray on their behalf, but there is never any mention of 
power over more distant descendants in the female line. A woman 
remains under the mystical power and protection of her own kinsmen, 
living and dead, all her life, and she is also affected by the anger of her 
husband's kin. 

' For an example see Godfrey Wilson, 'An Introduction to Nyakyusa Law', 
Africa, X. i, Jan. 1937, p. 27. 

- Cf. Godfrey Wilson, 'An African Morality', Africa, ix. i, Jan. 1936, pp. 83-5. 


The sins commonly spoken of as bringing down the wrath of senior 
relatives are unfilial behaviour, such as a son speaking angrily to his 
father, or actually striking his father, or wasting his father's substance 
by committing adultery, so that his father is forced to pay the fine, or 
failing to give his father a share of his earnings. 'If a married man, 
married with his father's cattle, goes away to work and gains much 
wealth and does not send any home to his father, he will fall sick. The 
ancestors will be angry and say: "He is not wise, he has given his father 
nothing!" He will fall sick there at his work. But if he is an unmarried 
man it is different ; his father has given him nothing, he may keep the 
cows to get himself a wife and nothing will happen ; only the first cow 
he gains he should send home to his father. If he does not do this he 
will fall sick. The other cows he can keep.' And if a daughter leaves her 
husband without good cause she risks her father's wrath, which, it is 
believed, may cause her to be barren. In the same way the heir who 
inherits the position of 'father' is believed to suflFer if he neglects his 
dependants. If he refuses to provide marriage-cattle for a younger 
brother or the deceased's son, or if he refuses to give bananas to the 
daughter of the deceased when she comes 'home' to him so that she 
may have a gift to take back to her husband, it is said that the heir's 
children will fall ill. 

Quarrelling betw'^een husband and wife, quarrelling between kinsmen, 
adultery of women, incest, neglect of rituals (including failure to 
summon relatives to rituals), and breaches of the rules of avoidance, 
all these are thought to rouse the wrath of senior relatives and bring on 
the wrongdoer or his children some such evil as sterility, illness, thin 
and sickly children, a lingering death, or solitude in the world of the 
dead. For if an heir has neglected his wards it is said that his kinsmen 
will not come to meet him on the road when he makes the journey to 
the land of the dead. Painful periods and difficult labour are also directly 
related to sin, and if a brewing of beer— the pride of a good housewife — 
goes sour, immediately it is whispered that she who brewed has been 
quarrelling with her husband. Have not the ancestors manifested their 
displeasiure ? 

But it is not only relatives who are believed to bring sickness and 
misfortune for breach of kinship obligations ; whenever fellow villagers 
are shocked (ukuswigaY at some breach of customary behaviour their 
breath — 'the breath of men' {embepo syabandu) — is believed to fall on 
the culprit and cause him (or her) to fall ill. Villagers are shocked, not 
only at the neglect of obligations towards themselves, but also at 
breaches of the laws of kinship and affinity, such as incestuous unions, 
neglect of avoidance, unfilial behaviour, neglect of his wards by an heir, 
and so on. Here we see the importance of the 'fathers of the village', for 
if a son has insulted his father it is not only the father and the ancestors 
' Literally 'to be astonished', but often with the implication of disapproval. 


who must be appeased, but also the father's fellow villagers. An erring 
son who has quarrelled with his father and later falls ill will bring a bull 
to kill at his father's homestead for the father and his neighbours to eat, 
so that the curse of the village may be removed. The exact nature of the 
supposed mystical power of neighbours need not concern us here ; it is 
related by the Nyakyusa to the power of witchcraft rather than to the 
power of the ancestors,' but the 'astonishment' of villagers is considered 
to be a legitimate use of mystical power, and it operates as one of the 
main sanctions for the maintenance of obligations between relatives as 
well as between neighbours. 

While emphasizing the mystical danger, in Nyakyusa eyes, of angering 
relatives and neighbours, we must not overlook the fact that shame also 
operates. Informants spoke of actions which would cause one 'to die of 
shame' (ukufzva nesoni), and to tax someone with being 'shameless'^ is 
one of the commonest forms of reproof. The line between being 
ashamed and fearing that the shocked astonishment of one's neighbours 
will affect one's health is ill defined, but it is quite clear that the Nyakyusa 
dread being shamed by comments on their stinginess, quarrelsomeness, 
or ill manners. Generosity and urbanity in kinship relations carry 
prestige in the same way as they do in village relations. One nch man, 
whom we knew intimately, provided marriage-cattle for several sisters' 
sons; he had no obligation to do so, but he gained great honour among 
his relatives as well as among neighbours by his generosity. 

It is apparent that these various types of social pressure were effec- 
tive, in the traditional Nyakyusa society, in maintaining the form of 
kinship and marriage relations we have described. To-day they are not 
wholly effective and the form is changing. The most obvious break- 
down is in marriage. The high degree of polygyny among elderly men 
and the large number of young bachelors unable to marry for lack of 
cattle make the temptation to adultery, often followed by divorce, 
very strong. The Nyakyusa themselves say that the most common 
reason for a woman to leave her husband and elope with a lover is sexual 
neglect. Some speak of their husband's laziness in hoeing, or his unfair 
division of milk and other food, but most often the complaint is sexual 
neglect, and it is possible that the lower divorce rate among Christians 
(18 per cent, as against 59 per cent, in a small sample) is related to the 
fact that they are the one group of families which remain monogamous. 

The sexual privileges of the older men were maintained by the right 
of the aggrieved husband to pursue and kill an adulterer and by the 
denial to women of any right of personal choice. It is clear that formerly, 
unless a woman could prove by the witness of neighbours that she was 
being severely ill treated by her husband, her father would refuse to 
countenance a divorce, nor would he agree to her refusing the heir on 
the death of her husband. Nowadays, under the influence of the British 
' Cf. Godfrey Wilson, op. cit., pp. 85-93. 



Administration, the courts uphold a woman's right of choice in marriage 
or inheritance, and her right to a divorce in certain circumstances. 
This new freedom for women, and the denial of the right of an individual 
to wreak private vengeance on an adulterer, are incompatible with the 
traditional form of marriage. 


(t/) tata^ My father, father's brothers and male parallel cousins. 

Also mother's brother in address. 
(JJ) juba My mother, mother's sisters and female parallel cousins, 

father's wives. 
Also father's sister in address. 
Umwanangu My child, son or daughter; sister's son or daughter 

(woman speaking); brother's son or daughter (man 

Brother's son or daughter in address (woman speaking). 
Sister's son or daughter in address (man speaking). 
Undume Husband and all his brothers, sisters' husbands, grand- 

son, grandfather (woman speaking). 
Unkasi, Wife and all her sisters, brothers' wives, granddaughter, 

grandmother (man speaking). 
Ulilumbu Sibling or parallel cousin of opposite sex. 

Used in some contexts of a cross-cousin of opposite sex. 
Unkulu Senior sibling or parallel cousin of the same sex. 

Unuguna Junior sibling or parallel cousin of the same sex. 

Untani Cross-cousin. 

{U) jubasenga^ Father's sister or female parallel cousin. 
Umivanasenga Brother's child or male parallel cousin's child (woman 

Umwipwa Mother's brother or his male parallel cousin. 

Sister's child or female parallel cousin's child (man 

Utnwisukulu Grandparent, grandchild; sibling of grandparent, 

sibling's grandchild. 
Unko Father-in-law and his brothers, mother-in-law and her 

sisters (man speaking), son-in-law and his brothers. 
Unkamwana Father-in-law and his siblings and parallel cousins, 

brothers of mother-in-law and husbands of her sisters, 

fathers-in-law of sisters (woman speaking). 
Daughter-in-law and her sisters and parallel cousins 

(man speaking). 
Undamu Wife's brother, sister's husband (man speaking). 

Ugzvifi Husband's sister, brother's wife (woman speaking). 

Unjimba Father of son-in-law or daughter-in-law (man speaking). 

' The initial u is always dropped in the vocative case ; it is seldom heard at all 
with juba, jubasenga, and tata. 

^ Esenga is an avoidance word for cattle. The father's sister is sometimes called 
umuenesenga which might be translated 'the owner of the cattle'. 



THE Tswana tribes of the Bechuanaland Protectorate seem to 
have fewer marriage restrictions than any other group of 
Bantu-speaking peoples in southern Africa. They lack exo- 
gamous units such as are commonly found elsewhere (e.g. 
among the Nguni, Venda, Tsonga, and Shona), and they allow marriage 
between first cousins of all kinds and various other close relatives. In 
addition, they practise the levirate and the sororate. However, the rules 
of mating are not always identical, and a union permitted in some tribes 
is forbidden in others. In this paper I shall compare briefly the marriage 
regulations of nine different tribes that I have studied in the field." My 
main objects are to determine if those regulations conform to a single 
fundamental type, and to see how far they can be related to other 
aspects of Tswana social structure. I shall comment also upon the local 
variations that occur, and discuss the problems that they present for 


I have already described elsewhere the main features of Tswana 
social organization.^ For our present purpose it is enough to note that 
each of the tribes dealt with here has its own territory, and forms a 
separate political unit under the leadership and authority of a chief who 
is subordinate only to the British Administration. In size they vary 
greatly, the extremes being represented by the Tlokwa, with about 2,000 
people, and the Ngwato, with about 110,000. Their members have been 
recruited from many different stocks ; the proportion of 'aliens' is largest 
in the north, smaller but still fairly high in the south, and smallest in the 
east.^ This ethnic diversity is sometimes reflected in the occurrence 
within a single tribe of several different languages and other cultural 
variations. But the Tswana proper, who are the ruling community in 
each tribe, all speak the same language and have much the same culture ; 

' The tribes are the Kgafela-Kgatla, Malete, and Tlokwa (Eastern Tswana), 
Tshidi-Rolong, Ngwaketse, and Kwena (Southern division of Western Tswana), 
and Ngwato, Tawana, and Khurutshe (Northern division of Western Tswana). 
There are in the Union of South Africa many other Tswana tribes, but the 
relevant information about them is not available. 

^ See especially Handbook of Tszvana Law and Custom (Oxford University 
Press, 1938), chap, i; Married Life in an African Tribe (Faber, 1940); and 'Some 
Features in the Social Organization of the Tlokwa', Southwestern Journal of 
Anthropology, ii, 1946, pp. 16-47. 

^ See note i above for the geographical distribution of tribes. 


unlike most of their alien subjects, moreover, they tend to concentrate 
in large compact villages containing several hundreds or even thousands 
of inhabitants. In this paper I shall deal with them alone and ignore the 
other elements of the tribal populations. 

The smallest well-defined unit in the social system of the Tswana is 
the household, a group of people occupying the same enclosure of huts. 
It consists basically of a man with his wife or wives and their unmarried 
children, but often also includes one or more married sons, brothers, or 
even daughters, with their respective spouses and children. It may 
contain up to fifteen people, occasionally more, but the general average 
is from five to seven. 

Several different households, living together in the same part of a 
village or ward settlement and acknowledging a common 'elder' [mogol- 
zoane), constitute a family-group. This group consists basically of families 
whose men are all agnatic descendants of the same grandfather or great- 
grandfather ; the man senior to the rest by right of birth is their accepted 
leader. However, it may also contain married sisters or daughters of those 
men, with their husbands and children; possibly one or more uterine 
nephews who have come to live permanently with their mother's people ; 
and in rare instances even siblings of a woman married into the group. 
In effect, it is a form of extended family, dominantly but not exclusively 
patrilocal in character. It usually has from twenty to fifty members. 

A number of family-groups, living together in a distinctive portion of 
a village, or sometimes even in a separate village, make up a ward. The 
ward is the basic unit in the administrative system of the tribe, its 
members being subject to the authority of a hereditary headman with 
well-defined judicial and executive powers. The number of wards varies 
roughly with the size of the tribe; the Tlokwa, for instance, have only 
five, and the Ngwato approximately 300. The number of people in each 
also varies considerably, some containing less than 100 and others well 
over 1,000. 

Occasionally the men belonging to a ward all have a common agnatic 
ancestor, the headman being senior to the rest in line of descent. But 
the great majority of wards now also contain people of alien origin. Some 
may have been placed there by the chief to help the headman and 
strengthen the ward when it was founded ; others may have been allowed 
to transfer from their own ward because of internal dispute; or, if a 
group of new-comers to the tribe is considered too small to form a sepa- 
rate ward, the chief will attach it to one already existing. Such recruits 
are seldom related by birth to the nuclear lineage of the ward, but they 
may subsequently marry into it. Neither the ward nor even the family- 
group is exogamous, and marriages between members are in fact fairly 
common. However, should a wife be an outsider, she must normally be 
brought to live among her husband's people, and the children all belong 
to his group. 


It will have been gathered tliat most of a man's neighbours may be 
closely related to him, especially in the agnatic line. His immediate 
maternal relatives and/or relatives-in-law are sometimes also of his own 
ward, but they generally belong somewhere else. In addition, he has 
more distant relatives of all kinds widely scattered over the tribe and 
even in other tribes, for the Tswana carry recognition of relationship 
much farther than is common in our own society. Almost everybody with 
whom genealogical connexion can be established, no matter in how 
remote a degree, is considered a kinsman. Even if no direct link can be 
traced the mere fact that someone belongs to the same group as a close 
relative may lead to his inclusion within the body of kin; thus, all 
members of the ward from which one's mother comes are loosely classed 
as 'maternal relatives' {baga etsho mogolo), and all members of the ward 
into which one's sister or daughter is married are classed as 'sons-in- 
law' (bagzce). 

In general, relatives are expected to be friendly and to help one 
another : a man looks to his kinsmen of all kinds for hospitality, assistance 
in work, and support in times of trouble. But he relies above all upon 
the members of his own family and such other close relatives as his 
parents' siblings with their spouses and children or his wife's parents 
and siblings. These people advise and help him in all his problems and 
undertakings, a special 'family council' being summoned if necessary; 
according to their particular status they make him prescribed gifts of 
food and other commodities, and have certain duties to perform at the 
ceremonies that he organizes; and they also are the only ones directly 
affected by the marriage regulations. His more distant relatives, unless 
they are actual neighbours, seldom figure prominently in his life, and 
even on special occasions of festivity or mourning, when all the available 
kindred are assembled, his close relatives invariably take pride of place. 

The importance of close relatives is reflected particularly in the system 
of 'linking' {go rulaganya) that is so conspicuous a feature of Tswana 
kinship. In almost every family of any size the children are usually 
paired together — elder brother with younger brother, elder sister with 
younger, and brother with sister. Those of the same sex are paired 
together alternately (first and third, second and fourth, and so on), those 
of opposite sex go in relative order of birth (eldest brother with eldest 
sister, second brother with second sister, and so on).' Moreover, a man's 
linked sister is also the linked paternal aunt of his children, just as he is 
the linked maternal uncle of hers; and linked brothers (or sisters) are the 
linked senior and junior paternal uncles (or maternal aunts) respectively 

* The arrangement depends mainly upon the actual composition of the 
family and the decision of the parents, and many variations are found from the 
order given above. 


of each other's children. Among his immediate relatives, therefore, 
a man may have one of each kind to whom he is specially attached, and 
with whom he is said 'to work together for life'. It is with them that he is 
most closely associated in kinship obligations, and it is they who above 
all are expected to carry out the roles conventionally assigned to relatives 
of their class. Should they be living elsewhere, a sibling of the same sex 
who is close at hand will do as a substitute, but on all important occa- 
sions they are specially summoned. 

The Tswana themselves habitually group their closer relatives into 
three major categories : ba ga etsho, agnatic relatives ; ba ga etsho mogolo, 
maternal relatives; and bagzvagadi (m.s.) or bo-tnmatsale (w.s.),' the 
kinsmen of one's spouse. The first is the most directly important, for 
many social institutions are organized on a patrilineal basis. Member- 
ship of tribe, ward, and family-group is determined primarily by descent 
traced through the father; property and rank normally pass from father 
to son, or, failing a son, to the next male member of the same lineage ; 
a man's surname is usually the given name of his father or paternal 
grandfather, and wards or family-groups often bear the name of their 
leader's agnatic ancestor; if a marriage is dissolved the children 
(especially the sons) always remain with the father; and, in the old days, 
people worshipped the spirits of their deceased paternal ancestors. In 
genealogies, too, the father's relatives are invariably remembered 
farther back and in greater detail than the mother's (unless she is 
descended from a chief). 

Moreover, as already noted, a man's close agnatic kinsmen usually 
belong to his own family-group and ward. Consequently they are the 
relatives with whom he habitually associates and from whom he expects 
immediate support and protection. Their mutual dependence unites 
this localized category of kin into a co-operative body whose solidarity is 
generally recognized. Fa gare ga bana ba mpa ga go tsenwe, says the 
proverb, 'Outsiders should not intrude upon children of one womb.' 

But the relations between individual members of the group are also 
governed by well-defined principles of discipline and authority. The 
head of a household is responsible in tribal law for the conduct and 
liabilities of his children and other dependants, from whom he accord- 
ingly demands unfailing obedience and respect.^ A mother, although 
entitled to similar consideration, is more openly affectionate and lenient 
than the father, and is usually the medium through whom he is ap- 
proached. A younger brother, again, should defer to and serve his elders, 

' Throughout this paper I shall use the abbreviations m.s. for 'man speaking' 
and w.s. for 'woman speaking'. 

^ It should be understood that the summary descriptions given here of 
behaviour patterns embody the substance of what informants defined as the 
recognized norms ; there are many individual instances in which the norms are 
not observed. See Married Life in an African Tribe, chaps, ix and x. 


who in turn control but also advise, support, and protect him. A similar 
pattern of behaviour obtains between sisters. Siblings of opposite sex 
are not so rigidly differentiated according to relative age, nor are the 
rules of conduct between them well defined. In general, a brother should 
look after the welfare of his sisters and help them in their troubles, 
especially once the father is dead ; they in turn render him such domestic 
services as fall within their scope. All such obligations apply chiefly to 
linked brothers and sisters, but in a lesser degree they are observed by 
other siblings also. 

The behaviour patterns obtaining in the family are extended to the 
other close relatives, but with modifications for age, sex, and seniority. 
Among those living in the same group the paternal grandparents and 
especially the grandfather share in the obedience and respect due to 
one's seniors generally, but they tend to be kindly and tolerant rather 
than insistent on strict discipline. The father's elder brothers have greater 
authority than even he, and are if anything more esteemed ; his younger 
brothers, on the other hand, are the recognized auxiliaries of their 
nephews, to whom they are much more affable and friendly. Paternal 
aunts are also respected, but seldom figure as prominently in a man's 
life as their brothers, especially if, as often happens, they leave the ward 
after marriage. They certainly do not command anything like the great 
authority that they are said to possess among the Venda and some of the 
Northern Sotho. The children of paternal uncles, again, are regarded 
much as brothers and sisters, and are one's usual playmates in childhood ; 
in later life, however, they are differentiated according to the relative 
status of their father and the appropriate conduct is applied to them. 
More remote agnates are distinguished mainly according to line of 
descent : if senior to one's father by birth, they are entitled to obedience 
and respect; if junior, their services can be freely commanded. The say- 
ing that a man's 'elder brother' is his chief, and his 'younger brother' his 
subject, summarizes adequately the accepted relationship. 

In everyday life the hierarchy of age and seniority is seldom considered 
oppressive, and most people remain on good terms with their agnates, 
co-operating willingly and harmoniously with them as occasion arises. 
But disputes sometimes occur owing to arbitrary exercise of authority 
and rival claims to property or position, and it is not fortuitous that 
most accusations of sorcery are made against one's relatives in the same 
ward. The close proximity in which they live, and the rules of patri- 
lineal succession and inheritance, breed jealousies and conflicts that 
may prove stronger than the ties of mutual dependence. 

One's maternal kinsmen {baga etsho mogolo) are not as a rule involved 
in situations of the kind just described ; they cannot be rivals for property 
or position, and they generally (although by no means invariably) belong 
to some other ward. In consequence, perhaps, they are notoriously more 
aff'ectionate and devoted than agnates. Children when small are often 


sent to live for a while at their mother's parental home, which they are 
afterwards encouraged to visit frequently; there they are assured of a 
warm welcome and generous hospitality, and enjoy many privileges. 
Ngwana mogolo kwa gabo-mogolo, says the proverb, 'A child is important 
at the home of its mother's people.' 

A linked maternal uncle, in particular, must be consulted in all 
matters specially affecting his sister's children; his opinion is so impor- 
tant when their marriages are being arranged that his veto is sometimes 
decisive; he helps with food, clothes, and other gifts at all their rites de 
passage; special exchanges of property are common between him and 
his nephews ; and he (or his successor) also has the ritual duty of pre- 
paring their corpses for burial. It is to his maternal uncle perhaps more 
than anybody else that a man looks for disinterested advice and aid in 
times of difficulty ; and when disputes arise between father and son, or 
brother and brother, it is often also the uncle who reconciles them or 
with whom the oppressed child or younger brother goes to live if peace 
cannot be restored. The mother's parents and sisters, similarly, are 
commonly said to be more affable and indulgent than the father's. 

Cross-cousins are expected to associate together on terms of the 
greatest intimacy. They are regarded as the most suitable mates for 
each other ; but even if they do not marry, or are of the same sex, they 
are entitled and encouraged to be very familiar. In speaking to each 
other, for instance, they can use obscene or insulting language which, if 
addressed to any other person, would be a just ground for offence ; they 
can help themselves freely to each other's personal belongings, and 
behave without restraint in various other ways. Mmapa le ntsalae 
moakodi, says the proverb, 'Beside his cross-cousin a man is happy.' 
This joking relationship, known as go tlhagana, prevails symmetrically 
between cross-cousins of both kinds, no distinction being made accord- 
ing to line of descent. 

The behaviour patterns between relatives-in-Iaw are not as formalized 
as some of those already mentioned. When a man marries one of his 
relatives his original kinship ties with her family tend to persist. A 
father-in-law who is also an uncle, or a brother-in-law who is also a 
cross-cousin, continues to be treated along much the same lines as 
before, except perhaps when questions arise affecting the stability of the 
marriage. Most men and women, however, marry outside their kindred, 
and thus enter into special relationships with a new series of people. 
Those who concern them most are the parents and siblings of their 
spouses; other relatives, such as uncles, aunts, and cousins, come into 
prominence only on festive or other special occasions. 

The basic attitude in all instances is one of friendship and co-opera- 
tion. A woman often continues to live at the home of her parents for 
some time after marriage, her husband going there to sleep with her at 
night. As a result, he sees much of her parents and other close relatives, 



who, without being familiar, treat him cordially and with respect. After 
taking his wife to his own home he is expected to visit his relatives-in- 
law frequently, help them at work, and invite them to all his domestic 
celebrations. They reciprocate in the same manner. A woman's relations 
with her husband's people may at first be characterized by mutual 
politeness, accompanied perhaps by a show of authority on their part, 
but if she proves a good wife they gradually unbend and she is accepted 
and treated as a daughter. There is no form of taboo or other prescribed 
avoidance between relatives-in-law of any kind. 


The kinship terminology reflects many, but not all, of the social 
distinctions. The basic pattern is set by the terms used within the family. 
Husband and wife call each other mosadi (woman) and monna (man) 
respectively;^ there is also a term for spouse, mogatsa, but it is used 
mainly in descriptive formations. Parents are distinguished according to 
sex : father is rre (dial, ntate), and mother is mme. Children can also be 
distinguished according to sex, a son being morzva and a daughter 
morwadi, but the common term for either is simply ngwana, child. 
Siblings of the same sex are distinguished according to relative age: 
older brother (m.s.) or older sister (w.s.) is mogolole (dial, nkgonne), and 
younger brother (m.s.) or younger sister (w.s.) is nnake. Those of oppo- 
site sex, i.e. brother (w.s.) or sister (m.s.), call each other kgaitsadi. 

The terms just given are generally also used in compound families, 
but, if necessary for purposes of reference, distinctions can be made by 
means of descriptive terms. A stepfather, for instance, may be called 
mogatsa-mme, mother's husband ; a stepmother mogatsa-rre, father's wife; 
and a half-sibling ngwana-rre, father's child, or ngwana-mme, mother's 
child. Co-wives of a polygynist call each other mogadikane (said to be 
connected in origin with the word bogadi, bride-wealth) ; alternatively, 
they may use mogolole, older sister, or nnake, younger sister, according 
to their relative status. Similarly, a woman may call her senior co-wife's 
child ngivana-mogolole, older sister's child, and her junior co-wife's 
child ngzvana-nnake; the child reciprocates by calling her either mmang- 
ivane, mother's younger sister, or mme-mogolo, mother's older sister. 

The distinctions made between siblings are extended to parent's 
siblings and to siblings' children. The terms for father's siblings all have 
the root rre, and those for mother's siblings all have the root mme, but the 
appropriate distinctions are made for sex or relative seniority. Father's 
older brother is rre-mogolo ('great father') and father's younger brother 
rrangwane ('little father'), but any father's sister is rrakgadi ('female 
father'); similarly, mother's older sister is mme-?nogolo and mother's 
younger sister mmangwane, but any mother's brother is malome ('male 

' All Tswana terms are given here in their simplest form; various modifica- 
tions can be made to indicate person or number. 


mother'). Reciprocally, older brother's child (m.s.) or older sister's child 
(w.s.) is ngwana-mogolole, and younger brother's child (m.s.) or younger 
sister's child (w.s.) ngwana-nnake, but brother's child (w.s.) is ngwana- 
kgaitsadi and sister's child (m.s.) motlogolo. Here all the terms used are 
descriptive, except the last, and no sex distinctions are made. 

As already noted, father's siblings are distinguished from mother's. 
In the second ascending generation the two lines are merged ; father's 
father and mother's father are both rre-mogolo, and father's mother and 
mother's mother are both mme-mogolo. These are also the terms for 
father's older brother and mother's older sister respectively, so that here 
we have persons of different generations being classed together. Reci- 
procally, son's child and daughter's child are also merged together ; the 
usual term is the descriptive ngzoana-ngwana, although motlogolo (sister's 
child, m.s.), is occasionally heard as an alternative.' Here again, as with 
other relatives of descending generations, sex distinctions are lacking. 

Parallel cousins are sometimes classed with siblings. This usage is 
especially common for agnates (the children of paternal uncles), although 
the descriptive terms ngzvana-rre-mogolo or ngzvana-rrangwane also occur 
ever5rwhere. The children of maternal aunts, however, are habitually 
called by descriptive terms, sometimes abbreviated in address to 
mme-mogolo or mmangwane respectively ; their classification with siblings 
is a less common and apparently not even universal alternative.^ On the 
other hand, cross-cousins, regardless of sex or relative age, always call 
each other by the special term ntsala (dial, motswala). 

The children of agnatic cousins, again, are classed with the children 
of siblings ; those of materterine cousins are called by descriptive terms 
(e.g. ngzvana-ngwana-mfne-mogolo), sometimes abbreviated in address to 
simply mme-mogolo or mmangwane ; and those of cross-cousins are also 
called by a descriptive term, ngwana-ntsala. 

The terms for other remote relatives are of the same pattern. The 
siblings of grandparents are classed, not with the grandparents (as is 
usual in other Bantu systems), but with uncles and aunts, i.e. they are 
called by the same terms as are applied to them by one's father or mother 
respectively. In reference, line of relationship is generally distinguished 
by adding the words a rre, 'of my father', or a mme, 'of my mother'; 
thus, the younger brothers of father's father or of mother's father are 
both addressed as rra?igwane, but the former is referred to as rrangwana- 
rre and the latter as rra?igzvana-mme. The agnatic cousins of one's parents 
are also classed with uncles and aunts ; the materterine cousins, however, 
are simply called either mme-mogolo (mother's older sister) or mmang- 
wane (mother's younger sister), regardless of sex; and the cross-cousins 

' The use of this alternative has been noted only among the Rolong, Ng- 
waketse, and Ngwato. 

^ I did not find it among the Kwena, Tawana, and Tlokwa, although I would 
not say that its use is completely unknown to them. 


are addressed as ntsala, but referred to as ntsala-rre or ntsala-mmi 
according to line of relationship. The children and other descendants 
of all these people are referred to descriptively but addressed by the 
same terms as their parents; the only exception is that the terms for 
siblings and brothers' children (m.s.) are occasionally applied to agnates. 

The general effect of these linguistic usages is that one's remoter 
kinsmen tend to be grouped into three major categories : {a) agnates, for 
whom descriptive terms may be used but who are sometimes also classed, 
according to generation, with paternal uncles and aunts, siblings, or 
brother's children (m.s.); {b) materterine relatives (including those of 
one's father), all of whom, regardless of sex or generation, are normally 
addressed as mother's sister, and (c) 'cross' relatives (ntsala), including 
the descendants of anybody called father's sister or mother's brother.' 

The terms for relatives-in-law vary considerably and are difficult to 
summarize briefly, since alternative usages occur even within a single 
tribe. Distinctive terms are found everywhere for husband's father and 
mother (both called mmatsale), wife's father and mother (both called 
mogwagadi), daughter's husband {mogwe, dial, mokgonyana), and son's 
wife (ngwetsi). Mmatsale and mogwagadi (or the descriptives ngzvana- 
mmatsale and ngwana-mogwagadi) are often also applied to the siblings 
and other close relatives of one's husband or wife respectively, although 
it is equally common for these people to be called by the same terms as 
are used for them by one's spouse. The eastern tribes, however, have 
a special self-reciprocal term, ?nogadibd, for husband's sister and brother's 
wife (w.s.), and the Kgatla further use molamo for wife's brother and 
sister's husband (m.s.). Elsewhere sister's husband (m.s.) is generally 
classed with daughter's husband. The same usage occurs in the western 
tribes for the husbands of nieces and other junior relatives, but the more 
common tendency, both here and in the east, is to class them with their 
respective wives. The wives of brothers and nephews, again, are either 
classed with their respective husbands or called by descriptive terms. 

The terms for the spouses of uncles and aunts are still more varied. 
The eastern and southern tribes class father's sister's husband with his 
wife; the northern tribes prefer the descriptive mogatsa-rrakgadi. The 
eastern tribes, again, class mother's older sister's husband with father's 

' The basic pattern may be represented by the following diagram: 
Ascending Contemporary Descending 

{A rremogolo -^ mogolole — > (ngwana-)mogolole 

A rrang\\'ane -^ nnake — > (ngwana-)nnake 

O rrakgadi^ 
( A malome / ~^ "*^^'^ ^ (ngwana-)ntsala 

Maternal I ^ mmemogolo — > (ngwana-)mmemogolo -^ (ngwana-ngwana-) 
I mmemogolo 

V O mmangwane — > (ngwana-)mmangwane -> (ngwana-ngwana-) 

The words in brackets are often omitted in address. 


older brother and mother's younger sister's husband with father's 
younger brother; this usage occurs also in the northern tribes, but more 
commonly both they and the southern tribes call such a man either by 
the same term as his wife or by a descriptive term. The wives of paternal 
uncles are usually classed with maternal aunts, but in the north descrip- 
tive terms are the more common alternatives. The northern tribes also 
use a descriptive term for mother's brother's wife; the eastern tribes, 
surprisingly, class her with grandmother ; and the southern tribes have 
both usages. 

In general, one is left with the impression that the terminology for 
relatives-in-law, and even to some extent for blood relatives, is nowadays 
going through a process of change. Such special terms as mogadibo and 
molamo, for instance, seem to be known only to elderly people; the 
abundance of alternative usages indicates a lack of standardization ; and 
there is also a marked tendency for both siblings and siblings-in-law 
of all kinds to call each other mogolole or nnake, terms that correctly 
should be used only for persons of the same sex as the speaker, or, 
especially among the eastern tribes, to substitute terms of Afrikaans 
origin, such as sebara (brother-in-law) and ausi (elder sister). 


Apart from being expressed in the behaviour patterns already noted 
kinship affects sexual relations. The Tswana prohibit marriage or 
cohabitation between relatives of certain categories; they also approve 
or even encourage the mating of others. These regulations apply to both 
'marriages' and 'secondary unions'. By a 'marriage' I mean here a form 
of mating in which the man and woman are legally recognized as husband 
and wife, and are subject to all the rights and duties that the relationship 
entails. The woman's people, for instance, are entitled to receive bogadi 
(where the practice still prevails), and the children that she bears belong 
to the man. A 'secondary union', on the other hand, is merely an exten- 
sion of an existing marriage. Its essential character is that, for purposes 
of child-bearing, one of the original parties to that marriage is replaced 
by another person of the same sex, who is regarded as a bodily substitute, 
and not as an independent spouse. Thus, if a woman is childless, her 
younger sister (or some other relative) may be provided to bear children 
on her behalf, such children ranking according to the status of the barren 
wife; or, when a husband dies, his younger brother may cohabit with 
the widow in order to raise seed to the deceased. The original marriage, 
in such cases, still persists, even although it is no longer the same people 
who cohabit. The distinction between the two forms of mating will 
appear more fully in the course of our discussion. It is important in the 
present context because a man is sometimes allowed to form a secondary 
union with a woman whom he may not marry. 

Before I deal with the actual rules of mating something should be 


said about Tswana marriage generally. For a marriage to be valid, under 
old tribal law, two essential conditions must be satisfied : (a) a formal 
agreement, reflected in the betrothal ceremonies, must be made between 
the two family-groups concerned, and (b) the bridegroom's family must 
give cattle to the family of the bride. These cattle are known as bogadi. 
Their number, which in practice usually ranges from four to ten head, 
is decided by the boy's people alone, the girl's family having no say in 
the matter. In the old days, once the betrothal had been confirmed, the 
boy was allowed to cohabit with the girl at her parents' home, and it was 
only after a year or more, perhaps when she had already given birth to 
a child, that she would be taken formally to live among his own people. 
This custom of temporary matrilocal marriage, known as go ralala, has 
been abandoned among the Ngwato and Tawana, but is still found in 
most other tribes. The bogadi cattle should normally be given by the 
time that the girl leaves her own home to join her husband's people; in 
fact, they are seldom produced until much later. 

There have been several other changes in marriage law, owing mainly 
to contact with Western civilization and especially with Christianity. As 
we shall soon see, the rules governing the choice of a mate have altered 
in some tribes. Again, the adoption of European civil marriage introduced 
various consequences and implications not found in old Tswana life, 
but this form of marriage is practised by only a small proportion of men. 
More important, perhaps, was the abolition of bogadi among the Ngwato 
and Tawana towards the end of last century, and its abandonment by 
many Christians in other tribes.' The giving of bogadi is accordingly no 
longer a universal condition of marriage. Greater stress is laid nowadays 
upon the consent of both family-groups, which must still be obtained 
before the tribal courts will regard a marriage as valid. 

Marriage Regulations ^\ 

The rules stipulating whom one may, or may not, marry apply to both 
kinsmen and affines. The kin whom a man is specifically forbidden to 
marry include, firstly, all women to whom he is related in the direct line 
of descent (grandmother, mother, daughter, granddaughter, &c.). The 
same prohibition applies everywhere to his full sisters, and nowadays 
also to his half-sisters (both paternal and maternal). Marriage with 
a half-sister was formerly permitted among the Ngwaketse and Rolong, 
but apparently nowhere else. Among the Ngwaketse it is said to have 
been abolished by Chief Bathoeng I (1889-1910). I do not know how or 
when it ceased among the Rolong.^ 

' This topic is discussed more fully in my Tribal Legislation among the Tswana, 
pp. 44 flF. (London School of Economics Monographs in Social Anthropology, 
No. 9, 1943.) 

^ Z. K. Matthews writes as if marriage with a half-sister is still allowed 
('Marriage Customs among the Barolong', Africa, xiii. i, Jan. 1940, p. 11); my 
own informants said that it is nowadays considered wrong. 


The only other kin whom a man is generally forbidden to marry are 
his parents' sisters and his sisters' daughters. Instances are known among 
the Kwena of marriage with a father's half-sister, and among the 
Ngwaketse of marriage with a mother's half-sister, but both were con- 
sidered irregular. Such marriages have not been recorded in any other 
tribe. Among the Ngwaketse, too, and among them alone, a man was 
formerly allowed to marry his half-sister's daughter, but this also was 
forbidden by Bathoeng I. In the other tribes it has apparently always 
been considered incestuous. 

The rules about marriage with a brother's daughter vary considerably. 
The Kwena, Ngwato, Tawana, Malete, and Tlokwa permit marriage 
with any brother's daughter. The Ngwaketse and Khurutshe say that 
a man may marry his half-brother's daughter, but not his full brother's. 
The Rolong and Kgatla, finally, prohibit marriage with both varieties. 

Except for the women just mentioned a man may marry any of his 
kin. Such marriages are indeed strongly encouraged, especially with 
a cross-cousin. My informants in all tribes said that cross-cousin 
marriages are preferable to any other. Ntsala zva motho ke tnogatse, states 
the proverb, *A person's cross-cousin is his (rightful) spouse.' In prin- 
ciple this applies especially to the daughter of the linked maternal uncle. 
Failing her, the daughter of any other maternal uncle will do. Then 
come, in stated order of preference, the daughter of the linked paternal 
aunt, and, failing one, the daughter of any other paternal aunt. Marriage 
with an agnatic cousin is also approved of everywhere. Here again the 
union is sanctioned by a proverb : Ngzvana rrangzvane, nnyale, kgonio di 
boele sakeng, 'Child of my father's younger brother, marry me, so that 
the (bogadi) cattle may return to our kraal.' Despite the wording, one 
may also marry the daughter of a senior paternal uncle. On the other 
hand, my Kwena informants spoke disapprovingly of marriage with 
a materterine cousin ; they said, too, that it was extremely rare among 
them, and would probably not be allowed nowadays, 'because your 
mother's sister's daughter is very much like your own sister'. Some of 
the Ngwato, too, said of such a marriage that 'it is unbecoming' (go a 
rona). But the same objections were not voiced in any other tribe, and 
certainly none, with the possible exception of the Kwena, prohibit 
a marriage of this kind. 

In the old days, according to my informants, so much importance was 
attached to cousin marriages that a boy's parents almost invariably first 
sought him a wife among the daughters of their own brothers and sisters ; 
if they did not, they were said to have 'violated our law' [ha tlotse vwlad 
wa segarona). Priority was given to the daughters of maternal uncles, and 
it was mostly with them that child betrothals occurred. A girl's cousin, 
moreover, could generally obtain preference over any other suitor, and 
if no one at all seemed anxious to marry her, her maternal uncles or 
paternal aunts were expected to find her a husband among their sons. 


SegSlS se tsholwa ke ba-ho-sona is the proverb quoted in this connexion, 
'A cripple is looked after by his own people.' 

However, cousin marriages are not compulsory. The claims of 
relationship can be offset by the lack of other desirable qualities, and if 
either the girl or her parents fail to reach satisfactory standards of 
conduct and character a wife will be sought somewhere else. Her own 
people, similarly, may refuse to hand her over to some cousin of whom 
they disapprove or with whose parents they have quarrelled. And, of 
course, it happens often enough that no suitable cousin is available. In 
such cases it is held that a more remote relative should be married if 
possible, but there is no special category for whom preference is 

We come now to relatives by affinity. As regards marriage regulations 
we may group them into three classes : those whom a man is allowed to 
marry, those with whom he may neither marry nor cohabit, and those 
whom he may not marry but with whom he is allowed to contract a 
secondary union should the occasion arise. I shall for the moment 
confine myself to the first two, and will discuss the other below. 

Among all the Tswana a man is to-day not allowed to marry or cohabit 
with his mother-in-law or his stepdaughter. The Ngwaketse formerly 
permitted marriage with the latter, provided that the man himself had 
not begotten children by her mother ; however, this also was prohibited 
by Bathoeng I. Except for these two, a man may marry any of his wife's 
relatives, but there is no special preference, e.g. for her brother's 
daughter, such as is found among some other Southern Bantu. He may 
also marry the widows or divorced wives of any of his male relatives, 
excepting only his stepmother and daughter-in-law; but he may enter 
into a secondary union with the former, and in some tribes (formerly) 
also with the latter. Among the Kwena and Ngwaketse, finally, one is 
allowed to marry a step-sister (the daughter of the father's wife or 
mother's husband by another marriage). In all other tribes such a union 
is considered incestuous, although instances are on record among the 
Ngwato and the Rolong. 

Secondary Unions 

So far I have been dealing with unions where the man and woman 
become full husband and wife. However, marriage among the Tswana 
is not merely a relationship between husband and wife ; it also establishes 
certain rights and duties between their respective families. One obliga- 
tion, common to both, is to ensure that the marriage shall be fruitful. 
Hence, if the wife is barren or dies fairly young, her parents may have 
to provide another woman to bear children on her behalf. This substi- 
tute is kno\vn as seantlo. She does not rank as an independent wife, but 
is attached to the 'house' of the woman whose place she is to fill; and, in 
the absence of a direct heir, her eldest son succeeds to the status and 


property of that house. Again, should a husband die, his widow if young 
enough is expected to cohabit with one of his kinsmen in order to con- 
tinue bearing him children. This man is not regarded as a new husband ; 
the children he begets by the widow are legally those of the dead husband, 
and they call their genitor, not 'father', but by the relationship term 
appropriate to his original status in their circle of kin, e.g. rrangwane, 
father's younger brother. No hogadi is given either for a seantlo or for 
'entering the hut' {go tsena mo tlung) of a dead husband. 

The choice of a seed-raiser is governed partly by the nature of his 
relationship to the dead husband; only certain relatives are eligible, 
others being debarred by the rules of incest. The man who above all is 
expected to cohabit with a widow is her husband's younger brother, 
preferably the next in order of seniority.' Failing a younger brother, the 
seed-raiser can be a junior paternal uncle, the son of a paternal uncle, or 
some more remote relative of approximately the same standing, e.g. an 
agnatic second cousin. The eldest son of a polygynist is also allowed to 
raise seed to his father, but only by junior widows, not by his own mother. 
This form of levirate is specifically sanctioned by the proverb, Molala le 
mtnaagwe ga a bolawe, itsala monnazve, 'The man who lies with his 
"mother" is not killed, he is begetting his (own) younger brother.' The 
Ngwaketse say that Chief Bathoeng I prohibited its practice among 
them, but it is still recognized in all the other tribes. 

In the two varieties just mentioned the seed-raiser is junior in status 
to the dead husband. However, an elder brother may also take on the 
role. This form of levirate, although not prohibited by any of the 
Tswana, is regarded with disapproval by the Ngwato, Rolong, and 
Kgatla. Their objection is due mainly to the difficulties its practice 
raises about the status of children. The proverb says, Kgosiga e tsalele, 
'A chief does not beget children for others.' Hence, it is pointed out, 
if a man begets children by his younger brother's widow they may tend 
to claim a higher status than that to which they are entitled. In law, they 
are the children of the younger brother, but, owing to the senior status 
of their genitor, they will prefer to claim him as their father instead. It 
is because of such possibilities of conflict that this form of levirate is 
disliked. There is apparently no objection on grounds of incest to 
cohabitation between a man and his younger brother's widow. 

All the tribes to-day forbid and condemn a man's 'entering the hut' 
of his son's widow, i.e. of his daughter-in-law. But this w^as formerly an 
accepted usage among the Kwena and Kgatla, and instances are also 
known of its having occurred among the Ngwato and Rolong (as well 
as among the Mmanaana-Kgatla, a formerly independent tribe now 
subject to the Ngwaketse). I lack the relevant information for the 
Tawana, but in the four remaining tribes I was told that the practice 
had never been permitted (although my Tlokwa informants said that it 

' The Tawana say that he should be the husband's linked younger brother. 


was beginning to occur nowadays). Even among the tribes where it 
formerly existed it did not have the same recognized status as the 
fraternal levirate ; the husband's father was not formally chosen as seed- 
raiser, but entered his son's hut surreptitiously and of his own accord. 

The only other recognized form of levirate is with a maternal uncle's 
widow. This variety is allowed by the Khurutshe, Kgatla, Malete, and 
Tlokwa, but is said to be relatively infrequent and to be practised only 
when no one else is available. The Rolong stated that they formerly also 
had the custom, but that it was forbidden long ago by Chief Tawana 
(c. 1805-49). The Ngwato, Kwena, and Ngwaketse say that it has always 
been prohibited among them as incestuous. I lack the relevant informa- 
tion for the Tawana. 

The substitute (seantlo) for a dead or barren wife must also be one of 
her close relatives. The most favoured is her younger sister, especially 
the one linked to her. Failing a younger sister, an unmarried elder sister 
may be provided instead. If no sister is available some other close 
relative will do, who should be of the husband's own or a younger 
generation. Preference is given to a matrilineal relative (a sister's daughter 
or maternal aunt's daughter), but an agnatic niece or cousin is also con- 
sidered suitable. A cross-cousin is generally excluded; only among the 
Kgatla and Rolong was I told that a maternal uncle's daughter could be 
provided if the wife had no other available relative. 

However, although if need be some such relative may be substituted, 
the wife's sister is always considered by far the most suitable person to 
take her place. This is reflected in a custom that was formerly universal, 
but has now been abandoned everywhere (among the Ngwaketse it was 
specifically abolished by Bathoeng I) : if a dead wife had no unmarried 
sister her parents might give the husband a younger sister who was 
already married to someone else. This form of sororate was practised 
especially when a man had married his cross-cousin. The husband of the 
woman who was thus taken to become a seantlo could keep the children 
she had borne him, but he could not recover his bogadi nor claim any 
later children as his own. 

It may be added that the sororate in general was never obligatory. 
Where the two families were previously related, or if they were on 
friendly terms, they might agree to substitute another woman for a dead 
or barren wife; but the wife's people were not bound to provide such 
a substitute, nor the husband to seek one. Greater importance was 
attached to the levirate, and although the custom is admittedly decaying, 
especially among Christians, public opinion even to-day still tends to 
condemn a man who fails to raise seed to his dead brother. 


The description just given embodies the substance of statements made 
by informants. It remains for us to see how far the Tswana do actually 



marry relatives and what categories they then prefer. The material upon 
which the following discussion is based consists of genealogies collected 
in eight tribes.^ For three (Rolong, Tawana, and Khurutshe) I have the 
genealogies of the royal families only; in the others I also obtained 
genealogies of people who do not belong to the ruling line, but who are 
of true Tswana stock. 

As will be shown in a moment, the proportion of marriages with near 
kin is markedly higher among members of the royal family than among 
commoners. Accordingly, I shall deal first with the five tribes where 
I obtained genealogies of both groups, and where my figures are conse- 
quently more representative of the population generally. These genealo- 
gies embrace altogether 2,574 marriages. For preliminary discussion we 
may classify the marriages into two major categories: (a) those where 
husband and wife had a common grandparent or were more closely 
related, and {b) those where they were more distantly related or not at 
all. The following table gives the number and percentage proportion of 
marriages of type (a) found in each tribe. (The heading 'Nobles' covers 
marriages recorded in the royal genealogies, and 'Commoners' the 
others; 'S' is the size of the sample in each case, and 'No.' the number 
of marriages between close relatives.) 

Table I 
Proportions of Marriages zoith Near Kin 



































Ngwato . 




















Tlokwa . 










Total . 



1 1 -4 







The figures show that, in general, the proportion of marriages between 
near kin is just about one in thirteen. There is fairly close agreement 
among all tribes except the Ngwato, who are w'ell above the average. 
However, since the genealogies do not constitute a true random sample, 
the tribal variations need not be stressed. On the other hand, the class 
variations are significant. Among the Ng\vaketse, Kwena, and Ng\vato 
the proportion of marriages between near kin is roughly two to three 
times greater among nobles than it is among commoners, and even 
among the Kgatla and Tlokwa it is appreciably greater, although not to 

' I have omitted my Malete genealogies, since they are not detailed enough 
for the present analysis. 


the same extent. The general average for nobles is one in nine, and that 
for commoners one in eighteen. 

In the three tribes for which I have genealogies of the royal families 
only, the corresponding figures are: 






Khurutshe . 








Including the figures already given, we get a total of 1,546 marriages 
of nobles, of which 177 (ii"4 per cent.) were between close relatives. 

The two sets of genealogies together embrace 265 marriages between 
people who had a common grandparent or were more closely related. 
We must now analyse these marriages to see if they indicate any special 
form of preference. To facilitate discussion I shall combine the figures 
for all the tribes, but where necessary I shall refer to significant local 

Table II 

Types of Marriage between Near Kin 

Wife's relation to husband 



Sister .... 
Mother's sister 
Sister's daughter 
Brother's daughter . 
Mother's brother's daughter 
Father's sister's daughter . 
Father's brother's daughter 
Mother's sister's daughter 










Before we consider the details it may be noted that marriages of the 
first three types together constitute less than 0-3 per cent, of the grand 
total (3,145), those with a brother's daughter i-i per cent., and those 
with first cousins 7-1 per cent, (cross-cousins 3-7, parallel cousins 3-4). 

Marriages with a sister occurred only among the Ngwaketse (3) and 
the Rolong (i); the woman, in each case, was a paternal half-sister. As 
already noted, such marriages were formerly permitted in both tribes, 
although nowadays they are considered incestuous. The marriages with 
the mother's sister and the sister's daughter were confined to the 
Ngwaketse, and here again, in each case, the connecting link was a 
paternal half-sister and not a full sister. Marriage with the sister's 
daughter, although nowadays prohibited, was formerly permitted, but 
marriage with the mother's sister is said to have always been regarded 
as incestuous; I failed to inquire why the two in question had been 



recognized. It may be added that with one exception (a marriage with 
the sister's daughter) all the marriages in this group took place last 
century and the people concerned have long been dead; among the 
Ngwaketse, moreover, two instances of marriage with a sister and one of 
marriage with a mother's sister were found in a single collateral branch 
of the royal family. 

Marriages with a brother's daughter occurred in all tribes except the 
Kgatla, where, as already noted, they are prohibited. The Rolong, 
according to my informants, also prohibit them, but one early instance 
is found in the genealogy of the royal family; and among the Ngwaketse, 
where it was said that a man may marry his half-brother's daughter but 
not his full brother's, there were two instances (out of twelve) of marriage 
with the latter. 

The figures for cousin marriages show a marked difference between 
stated preferences and actual practice. The order of preference, as laid 
down by informants, is mother's brother's daughter, father's sister's 
daughter, father's brother's daughter, mother's sister's daughter. The 
actual order of frequency, as shown in Table II, is father's brother's 
daughter, mother's brother's daughter, father's sister's daughter, 
mother's sister's daughter. There is, however, very little difference 
between the first two, although they are far more common than the 
others. Moreover, they occurred in every tribe. No instances of marriage 
with a daughter of either father's sister or mother's sister were found in 
the Khurutshe genealogy, but this may be due partly to the smallness 
of the sample. The Rolong also lacked instances of mother's sister's 
daughter marriage. On the other hand, one was noted among the Kwena, 
who it will be remembered spoke disapprovingly of such a marriage ; but 
it should be added that the man in question, a member of the royal 
family, had settled and married among the Tlokwa, whose chief was his 
maternal uncle. 

Further analysis of the cousin marriages reveals another important 
feature. In the five tribes where I have genealogies of both nobles and 
commoners the various forms were distributed as follows among the two 
classes : 

Table III 

Class Variatiotis in Cousin Marriages 










Mother's brother's daughter 
Father's sister's daughter . 
Father's brother's daughter 
Mother's sister's daughter 

43 303 
20 14- 1 
68 479 
II 7-7 





142 loo-o 

80 1 loo-o 


Among nobles nearly half the marriages are with the father's brother's 
daughter; the mother's brother's daughter comes next, but is obviously 
less favoured ; and then, much lower in the scale, are the father's sister's 
daughter and, finally, the mother's sister's daughter. (In the three tribes 
for which I have genealogies of the royal families only, the corresponding 
proportions, out of 52 cousin marriages, are : father's brother's daughter, 
48-1 per cent.; mother's brother's daughter, 36-5; father's sister's 
daughter, 9-6; and mother's sister's daughter, 5-8.) Among commoners, 
on the other hand, there were as many marriages with the mother's 
brother's daughter alone as with all the other types of cousin combined ; 
only one-fifth were with the father's brother's daughter, and the mother's 
sister's daughter ranked equally with the father's sister's daughter, both 
nearly as common as the father's brother's daughter. 

The genealogies of the commoners are not sufficiently extensive for 
further detailed analysis, but those of the nobles allow us to determine 
the proportions and kinds of marriages with more remote relatives than 
those already discussed. Of the 1,546 marriages in this group (the 
figures are for the eight tribes combined), 177 (as already noted) were 
between people related through a common grandparent or more closely. 
There were another 200 (12-9 per cent.) in which husband and wife had 
a common great-grandparent, 408 (26-4 per cent.) in which they were 
still more remotely connected, and 761 (49*2 per cent.) in which there 
was no discoverable relationship. In all, that is, about half the noble 
marriages are between people genealogically connected. Among com- 
moners I was able to establish genealogical relationship in only 331 out 
of 1,599 niarriages (20-7 per cent.). 

The only figures calling for further comment are those for marriages 
among nobles having a common great-grandparent. Of the 200 marriages 
in this group, no fewer than 112 (56-0 per cent.) were between agnates 
(father's father's brother's daughter, 15 ; father's father's brother's son's 
daughter, 58; father's brother's son's daughter, 36; brother's son's 
daughter, 2). If we add the marriages of type (a), we find that of 377 
marriages in which husband and wife had a common great-grandparent 
or were more closely related, there were altogether 213 (56-5 per cent.) 
with an agnatic relative, 80 (21-2 per cent.) with a paternal 'cross' 
relative, 67 (17-8 per cent.) with a maternal 'cross' relative, and only 
17 (4*5 per cent.) with a matrilineal relative. It seems evident that men 
of the royal family, when they do marry fairly close relatives, prefer 
women of their own lineages. Some possible reasons will be suggested 

We may now turn to afiinal marriages or secondary unions. The inci- 
dence of the sororate is easy to determine from the genealogies. In all, 
347 men had married more than once, the total number of their wives 
being 889. Of the 542 additional wives 16 were explicitly stated to have 
been taken as substitutes for a childless or deceased relative. There were 



36 Other instances in which a man had married independently a close 
relative of an earlier wife. This gives a total of 52 marriages (9-6 per cent.) 
in which the additional wives of polygynists or widowers were closely 
related to predecessors. The actual relationships were as follows: 






Older sister .... 




Younger sister .... 





Mother's younger sister 




Sister's daughter 





Brother's daughter 




Mother's sister's daughter . 




Father's brother's daughter 





Mother's brother's daughter 





Father's sister's daughter . 








It will be seen that almost half the unions were with a wife's sister, 
and almost a fourth with her father's brother's daughter. The one 
instance of true sororate with a mother's brother's daughter occurred 
among the Kgatla. 

Information about the extent to which the levirate is practised is 
seldom available in the genealogies. Since the widow's children in such 
cases are held to belong to her late husband, the name of the actual 
father if (someone else) would not necessarily be mentioned ; nor, I must 
add, did I when compiling the genealogies think of inquiring into the 
point. However, the relevant information was occasionally volunteered, 
especially about the widows of chiefs. In all, I learned of 16 instances. 
The seed-raiser's relationship to the dead husband in these instances was 
as follows : son, 5 ; younger brother, 4 ; brother's son, 3 ; father's brother's 
son's son, 2 ; older brother, i ; mother's sister's son, i . The last, which 
occurred among the Tlokwa, is perhaps the only one calling for special 
comment ; here the husband had been adopted as a child by his maternal 
aunt, and the man who subsequently acted as his seed-raiser was there- 
fore also his foster-brother. 

There were also 129 women who had married again after being 
widowed or divorced. The total number of their husbands was 274. Of 
the 145 additional husbands, 20 (13-8 per cent.) were closely related to 
a predecessor. The exact relationships were as follows : father's brother's 
son, 8 ; younger brother, 5 ; father's younger brother, 2 ; brother's son, 
mother's brother, sister's son, father's sister's son, and mother's brother's 
son, I each. The one instance of marriage with the husband's sister's son 
(i.e. of a man with his mother's brother's widow) occurred among the 
Ngwaketse, where it is said to be forbidden ; the man was subsequently 
overwhelmed by a series of misfortunes, which were piously attributed 
to the breach of the taboo. 

l6o kinship and marriage among the tswana j 

Conclusions \ 

The material presented above shows that, in general, the Tswana 
prohibit marriage only between kin of the first and second orders, 
excluding a brother's daughter; there are also only a few affines whom 
a man may not marry (his wife's mother, wife's daughter, father's 
wife, son's wife, and stepsister). Marriage is allowed with any first 
cousin, but preferably with a cross-cousin; various forms of levirate are 
practised (junior fraternal, senior fraternal, and filial), and so is the 
sororate, especially with the wife's sister. 

However, there are several local variations of the general pattern. The 
Rolong and Kgatla prohibit marriage with any brother's daughter, the 
Ngwaketse and Khurutshe with a full brother's daughter; the Ngwaketse 
and Kwena allow marriage with a stepsister; the Ngwaketse formerly 
permitted marriage with a half-sister, half-sister's daughter, and step- 
daughter, and the Rolong with a half-sister ; and the Kwena and Kgatla 
formerly recognized the paternal levirate (cohabitation with a son's 
widow). The Khurutshe and the three eastern tribes (Kgatla, Malete, 
and Tlokwa) also allow cohabitation with a mother's brother's widow, 
which is prohibited elsewhere, although it was practised long ago by the 

Genealogical and other data show that the mating prohibitions have 
occasionally been violated, although on the whole instances are rare.' 
In addition, the cousin marriages that have actually occurred do not 
conform to the stated order of preference. Among commoners the 
mother's brother's daughter is by far the most frequent choice, but the 
father's brother's daughter comes next, ahead of the father's sister's 
daughter; among nobles the father's brother's daughter is preferred to 
even the mother's brother's daughter, and both are married more 
commonly than the two other types of first cousin. Nobles also show 
a definite bias towards marriage with agnates among their more 
distant kin. 

These facts pose various problems for discussion. Can we account for 
the tribal variations in rules of mating ? Why do cousin marriages differ 
in actual order of frequency from the order of preference laid down by 
informants, and why do nobles, in contrast with commoners, marry 
agnatic relatives more frequently than others? How are the rules of 
mating, in general, related both to other aspects of the kinship system 
and to the social structure as a whole, and, in particular, can we explain 
why the Tswana have such a limited range of restrictions ? 

It may be said at once that the Tswana material does not fully support 
Rivers's well-known contention that 'the terminology of relationship has 
been rigorously determined by social conditions', including especially 

' I have dealt with this topic more fully in a paper on 'The Tswana Concep- 
tion of Incest', in Social Structure (ed. M. Fortes, 1949), pp. 104-20. 


forms of marriage.' It is true that relatives to whom different patterns 
of behaviour apply are often also called by different terms, e.g. paternal 
and maternal uncles, elder and younger brothers, cross-cousins and 
parallel cousins. On the other hand, the same term of relationship is 
sometimes used both for people with whom mating is forbidden and for 
those who may be married or taken as secondary consorts. Thus, 
kgaitsadi = (a) sister (prohibited), (b) father's brother's daughter and 
mother's sister's daughter (permitted) ; rrakgadi = (a) father's sister 
(prohibited), (b) father's father's brother's daughter (permitted); 
mmangwane = {a) mother's younger sister (prohibited), [b) junior step- 
mother and father's younger brother's wife (both permitted); and 
mogwagadi = {a) wife's mother (prohibited), {b) wife's sister (permitted). 
In such cases there is generally an alternative term for the woman who 
may be married. A parallel cousin, for instance, may be called by a 
descriptive term, and a wife's sister by the term for 'sister (w.s.)'. But, 
with parallel cousins descriptive terms are far more commonly used for 
the mother's sister's daughter than for the father's brother's daughter, 
although marriage with her is much less frequent; on Rivers's argument, 
we should have expected the opposite tendency to prevail. Again we 
may find all tribes using the same term for a certain relative, but differing 
in the rules about mating with her. The best example is the brother's 
daughter, who may be married in some tribes but not in others, yet who 
is called by the same term throughout. 

The mating regulations also do not conform exactly to the other rules 
of behaviour between relatives. Parallel cousins are said to be treated 
'much the same as sisters', but, unlike them, may be married; and the 
practice of the levirate with the widows of one's father or paternal uncles 
does not accord with the respect that should be shown to such women 
as 'mothers'. It can of course be argued, and correctly, that in these 
instances the behaviour patterns are not really the same, the rules 
varying in intensity with the nearness of the relationship. But the same 
argument will not explain why some tribes allow and others prohibit 
marriage with the brother's daughter, although in all she is regarded 
'as a daughter'; nor does the formal attitude towards the mother's 
brother's wife in other respects differ according to whether or not she 
is a potential mate. Again, the symmetrical joking relationship between 
cross-cousins fits in well enough with the preference expressed for 
marriage between them, but this does not explain why the mother's 
brother's daughter is married far more frequently than the father's 
sister's daughter. 

The observations just made suggest that among the Tswana there is 

apparently little causal connexion between kinship terminology, mating 

regulations, and the other rules of kinship behaviour, i.e. we cannot say 

that one of them necessarily 'determines' the others. The kinship 

' W. H. R. Rivers, Kinship and Social Organization, 19 14, p. i. 



terminology is not derived from the rules of mating, nor are the latter 
always consistent with the other rules of behaviour between relatives. 
To explain why the Tswana marry as they do we must therefore look 
also to other features of their social system. 

We must note, firstly, that marriages are usually arranged, not by the 
young couple themselves, but by their parents and other close relatives; 
formerly, indeed, it was not uncommon for girls to be betrothed during 
infancy or even before they were born. Moreover, in choosing a bride 
people look for the qualities likely to ensure a stable marriage : the girl 
herself should be industrious, modest, chaste, obedient, and amiable, 
and her parents should be of respectable ancestry and good character, 
free from any suspicion of practising sorcery. 

It is largely because parents wish to find a good wife for their son that 
they prefer to marry him, if possible, to the daughter of some close 
relative, with whose conduct and reputation they are themselves well 
acquainted. Near relatives, informants also argue, are apt to be more 
tolerant of each other than strangers, and, because of the pre-existing 
ties, their kin will take greater interest in the welfare of the marriage and 
try to ensure its success. Such marriages, moreover, bind the two 
families together even more closely than before and make for increased 
harmony and co-operation — a factor of much importance in a society 
where people depend greatly upon their relatives for help in major 
household activities. 

These statements help us to understand why among the various types 
of cousin marriage those with the daughters of uncles predominate. 
A man's brothers normally live in his own family-group and ward, so 
that he sees them daily, whereas his sisters will generally have moved 
away after marriage. A married woman, similarly, maintains closer 
contact with her brothers, who as the male representatives of her own 
family continue to have some responsibility for her welfare, than she 
does with her sisters, whose husbands are not so directly concerned. In 
both instances a child's parents are apt to associate more intimately with 
his uncles than with his aunts, so that the former's daughters tend to be 
considered first, because they are better known, when a wife has to be 
chosen for him. 

Several other factors determine the preference shown by commoners 
for marriage with the mother's brother's daughter. A man whose own 
wife has proved satisfactory readily turns to her family for a daughter- 
in-law, because experience has shown him that they train their children 
well. The bogadi cattle received by a man when his sister is married are 
also held to give her sons a prior claim to his daughters, and he himself, 
since he has an important say in choosing a wife for his nephew, may 
succeed in arranging a marriage with his daughter. The joking relation- 
ship between cross-cousins is another reason often given : the familiarity 
that it encourages and even enjoins, some informants state, makes for 


tolerance and indulgence between husband and wife and thus helps to 
ensure a stable marriage.' 

Similarly, in explaining the relative popularity of marriage with 
a father's brother's daughter, the Tswana often add that it keeps the 
bogadi cattle in the same family circle and prevents them from passing 
into the hands of outsiders.^ In the old days, when polygynous 
marriages were much more common than now, this may have been an 
important factor, since bogadi would have to be given for several wives. 
But even then the cattle received when a girl married did not remain the 
sole property of her father or brother; they were divided among her 
close relatives generally, including especially her maternal uncle, who in 
some tribes had the first claim upon them. The economic argument must 
therefore not be stressed unduly; moreover, as we have seen, among 
commoners the father's brother's daughter is married far less commonly 
than the mother's brother's daughter, in whose case it is not relevant, 
since marriage with her again involves the alienation of cattle from the 

Among nobles, where marriage with the father's brother's daughter 
is actually the most common form of cousin marriage, the dominant 
factor is certainly not bogadi but status. It is considered highly desirable 
that the chief's heir and other senior children should marry persons of 
rank. The motive is not so much 'to keep the blood pure' as to secure the 
political advantages attached to union with powerful or influential 
families. Hence the choice of a mate is in effect restricted to the royal 
families of other tribes, the leading families of important subject com- 
munities within the same tribe, or collateral branches of the royal family 
itself. There is another reason why agnates are preferred. Since, in the 
Tswana social system, a man's maternal relatives are expected to be his 
main partisans, a chief by marrying into his own lineage can ensure that 
his sons, especially the heir, have a powerful backing close at hand. There 
are several instances in tribal history where a claimant to the chieftain- 
ship, whose maternal relatives were of the same tribe, succeeded against 
a rival whose mother was a foreigner. A man whose maternal uncles were 
also his agnates was thus in a particularly favourable position. The fact 
that he was their sister's son would probably also help to remove the 
sources of conflict which, as we have seen, sometimes enter into the 
relations of agnates. 

So far I have been trying to show why the Tswana, in general, tend 
to marry certain types of relatives more frequently than others. It is 
more difficult to explain why some tribes allow kinship marriages that 
others prohibit. As I have already indicated, except for the rules of 

' Other informants, however, maintained that the joking relationship has the 
opposite eflfect, since a wife may presume upon it to be cheeky and even imper- 
tinent to her husband instead of showing him due deference and submission. 

^ Cf. the proverb quoted above, p. 151. 


mating there appear to be hardly any significant tribal variations in either 
formal behaviour patterns or terminology for the relatives concerned ; 
we cannot, therefore, connect the differences in mating regulations with 
other differences in the kinship system. This applies not only to the 
sister, sister's daughter, and stepdaughter (with whom marriage is now 
prohibited in the tribes where it was formerly allowed), but also to the 
stepsister and brother's daughter, about whom usage still differs. The 
only relative for whom kinship terms vary is the mother's brother's wife. 
The eastern tribes call her 'grandmother', the northern tribes use a 
descriptive term, and the southern tribes have both usages. This may 
be associated with the fact that it is also only among the eastern tribes, 
and the Khurutshe,' that one may cohabit with her; but, here again, 
there is nothing to indicate that the conventional behaviour patterns 
differ in other respects. 

It may well be that I did not investigate these variations deeply enough 
while in the field; I certainly did not inquire, as I should have done, into 
the formal reasons for every individual type of marriage prohibition, 
a procedure that would possibly have thrown light upon some of the 
inter-tribal differences. At present I can only suggest that the variations 
must be attributed, at least partly, to independent local developments 
due to specific historical incidents. I cannot account in any other way 
for the former occurrence and subsequent abolition among the 
Ngwaketse of some types of marriage that were prohibited everywhere 
else, nor for the complete prohibition of marriage with the brother's 
daughter among both the Rolong and the Kgatla, who belong to different 
divisions of Tswana and have never been intimately connected. It is 
possible also that diffusion among neighbouring tribes may account for 
the former existence of sister marriage among the Rolong and Ngwaketse, 
and of the paternal levirate among the Kgatla and Kwena. On the other 
hand, it should be noted that the Rolong and Kgatla are both marginal 
tribes, who have closer affinities with the Tswana peoples of the Union 
than with those of the Protectorate. It is possible, therefore, that com- 
parison with the Union tribes may help to explain some of the differences 
noted. However, detailed information is not at present available about 
the mating regulations of those tribes. 

This brings me to the final point. I have shown that the Tswana 
tribes (of the Bechuanaland Protectorate) permit marriage with all but 
a few very close relatives ; but I have not tried to explain this feature of 
their kinship system. In other Sotho groups, such as the Pedi of the 
Transvaal, marriage is prohibited with the mother's sister's daughter, 
and in still others, such as the Lobedu, marriage is permitted only with 
cross-cousins, especially the mother's brother's daughter. There are 
other southern Bantu, such as the Nguni, who prohibit marriage with 

' I failed to get the kinship terminology of the Khurutshe, and am therefore 
unable to say what they call the mother's brother's wife. 


any first cousin at all, and who are also organized into exogamous 
patrilineal clans. It seems evident that for an adequate explanation of 
Tswana mating regulations we need a still wider comparison than I have 
suggested above. We must compare the Tswana with other groups of 
Sotho, and the Sotho generally with the Nguni and other groups of 
Southern Bantu. Only then, I think, will it be possible for us to state 
why the Tswana regulations differ from those of other groups. This I 
hope to show more fully in a later publication. 






IN this essay I contrast the marriage and kinship systems of the Lozi, 
the dominant tribe of Barotseland, with those of the Zulu.' I make 
this comparison because it brings out clearly the attributes of the 
Lozi system, and these attributes are characteristic of a large number 
of Central African systems which I consider have not yet been defined. 
On our present information I tentatively define these systems negatively, 
by the absence of a corporate lineage (i.e. of any organized kinship group 
of several generations depth reckoned in one line, which is internally 
segmented on a genealogical system, the whole and each segment having 
identity and unity against corresponding groups) .yThe Zulu system is of 
this latter type, wherein the corporate lineage and its segments endure 
in time irrespective of changes in personnel, and form the nuclei of 
villages and local groupings. 

In Central Africa villages are also corporate groups of kindred, but 

* I worked among the Lozi as a research officer of the Rhodes-Livingstone 
Institute of Social Studies, Northern Rhodesia (1940-2, 1947), and among the 
Zulu on a grant from the National Bureau of Educational and Social Research 
(Carnegie Fund), Union of South Africa Education Department (1936-8). I 
thank Dr. Elizabeth Colson and Professor A. R. Radcliffe-Brown for criticisms 
of an early draft of this essay. It was written at the Rhodes-Livingstone 
Institute in Livingstone and I was restricted in making comparative reference 
to the books available there and at libraries in Johannesburg which I visited. 

The argument advanced in this essay arose directly from E. E. Evans- 
Pritchard's Some Aspects of Marriage and the Fainily among the Nuer, Rhodes- 
Livingstone Paper No. 11, 1945 (reprinted from Zeitschrijt fur Vergleichende 
Rechtsivissenschaft, 1938). I acknowledge gratefully the stimulus I received from 
this analysis, though I have only referred to it occasionally in my text, and from 
discussions with Professor Evans-Pritchard himself. 

The best available descriptions of the Lozi are my 'The Lozi of Barotseland' 
in Seven Tribes of British Central Africa, London : Oxford University Press 
for the Rhodes-Livingstone Institute (forthcoming) ; Economy of the Central 
Barotse Plain, Rhodes-Livingstone Paper No. 7, 1941 ; and Essays on Lozi Land 
and Royal Property, Rhodes-Livingstone Paper No. 10, 1943. A general account 
of the Zulu, not based on field research, is E. J. Krige's The Social System of the 
Zulus, London: Longmans, 1936. See also my 'The Kingdom of the Zulu of 
South-East Africa' in M. Fortes and E. E. Evans-Pritchard (editors), African 
Political Systems, London: Oxford University Press, 1940. In general I describe 
Zulu law as practised by the Zulu, not as laid down by the Natal Native Code 
and Native Appeal Court decisions (see W. G. Stafford, Native Law as Practised 
in Natal, Johannesburg: Witwatersrand University Press, 1936). 


they are formed on a different framework. The villages move frequently 
in search of virgin woodland for cultivation and their membership 
changes at each move and between moves. Villagers constantly alter 
their residence. A village is constituted by the attachment of small 
groupings of kindred, of families, and of individuals, to a headman with 
whom they are linked in various ways. In many matrilineal and matrilocal 
tribes of Northern Rhodesia and Nyasaland a series of small groups of 
matrilineal kin are linked to the headman, who is a member of one, by 
various cognatic ties (e.g. Yao and Chewa, and perhaps Bemba and 
Lamba).' Other kinsmen and kinswomen also attach themselves to 
a popular headman. Bemba may move through four villages during their 
lives. Among the Ila and their neighbours, the Tonga, who are matri- 
lineal but dominantly patrilocal, the links of villagers to headman are 
more varied. With a headman will be some of his sons, some of his 
uterine nephews, and some of his sons- and brothers-in-law, as well as 
other relatives, with some of whom he cannot trace genealogical con- 
nexions. In both these matrilocal and patrilocal complexes the inhabi- 
tants of neighbouring villages are not related to one another by links 
which form them into a kinship grouping standing against other kinship 
groupings. Nor is it clear from anywhere that the relations of groups of 
k^n are determined by their positions relative to one another in a 
genealogical system. The only corporate group of kindred is the village. 

Associated with this type of village are kinship systems which are 
shallow (1-4 generations from adults) but ramify widely in all lines. 
Shallow kinship systems ramifying in all lines, short-lived villages of 
varied constitutions, and slash-and-burn shifting cultivation form a single 
coherent complex. However, I believe that Central African kinship 
systems may have a common pattern, independent of this environmental 
setting. It appears in an unusual form among the Lozi largely because 
in their environment the size to which their villages can grow is limited 
and kinship groupings cannot expand in one locality, though each 
village may endure through centuries. Initially, we may note that the 
Lozi and their related sub-tribes are dominantly patrilineal and patri- 
local (like their Lunda forebears), though the surrounding tribes are 
matrilineal and often matrilocal. Part of the Lozi kinship system is a 
marriage complex differing in almost every rule and consequence from 
that of the Southern Bantu. 

The Lozi live mainly in the great flood plain of the upper Zambezi 
river. Most Lozi build their villages on 'mounds', higher parts of the 
plain, which stand above the waters of the summer floods. Towards the 

' See on the constitution of these villages M. Gluckman, J. C. Mitchell, and 
J. A. Barnes, 'The Village Headman in British Central Africa', Africa, xix. 2, 
1949, pp. 89-106; and essays by J. A. Barnes on the Ngoni, E. Colson on the 
Tonga, M. Gluckman on the Lozi, and J. C. Mitchell on the Yao in Sezen Tribes 
of British Central Africa. 


height of the flood the Lozi move for a few months to other homes, 
built above the water-hne on the margins of the plain, where they remain 
until the fall of the waters in winter enables them to return to their 
permanent homes. Not all the members of one plain village go to the 
same margin village. Especially since the 1890's, many Lozi have made 
their permanent homes on the margin near markets created by missions 
and administrative posts. The flood, and the movements of people, 
cattle, and fish associated with it, dominate Lozi life. Gardens are 
covered and uncovered, watered and fertilized, by it; it fixes the pastur- 
ing of cattle; it conditions methods of fishing. 

The Lozi have a mixed economy. They are proficient gardeners and 
fishermen, planting and reaping crops and catching fish in various sites 
during almost every month of the year. Their gardens and fishing-sites 
are dispersed in the plain, from its centre at the river into the encircling 
woodlands. Often diflFerent activities have to be pursued at the same 
time in widely separated places. They have to send their cattle to flood- 
season grazing in the woodland and the small plains within it when they 
are gardening and fishing on the Zambezi plain; the cattle return to 
graze on pastures in the plain when the flood falls, before the people 
move back to their homes. 

The flood compels the Lozi to build their homes on the mounds in 
the plain and therefore these are the basis of their social organization. 
To each mound are attached certain gardens and fishing-sites which in 
general can only be used by the residents of the village built on the 
mound. The villages are of different types. Some villages are inhabited 
by groups of kinsmen, consisting usually of a small core of agnates and 
other cognates with their wives and children, and possibly an occasional 
male afiine or stranger as well as serfs. The limited area of the building- 
space on a mound, and of the resources centred in it, restricts the size 
to which villages can grow. Some have only a couple of men in them and 
the average village varies from six to ten men. The other type of village 
belongs to the royal family and to the titles of councillors at court. These 
are inhabited by members of diflFerent tribes, both Lozi and subject 
peoples, collected by kings. Their size runs from tens of people into a 
couple of hundred, and thousands at the capitals. Royal villages are 
aggregations of small groups of kindred which resemble the groups in 
family villages, so that in eflf"ect a royal village is like a number of family 
villages concentrated together. Inhabitants of a royal village have per- 
manent rights of holding in it. The Lozi imposed Lozi rules on members 
of subject tribes brought into their villages, and to-day it is impossible 
to distinguish their descendants by behaviour from true Lozi, of whom 
indeed there are admitted to be very few, because of intermarriage. 

The Lozi villages therefore have histories running into hundreds of 
years, for since they came to the plain they have built on the same 
mounds. Yet because mounds are small and scattered in the plain, 


members of each village as its population grows must seek new land 
at a distance; perhaps, therefore, they have not been able to develop 
extended unilineal kin-groups. In addition, I believe this pattern is 
generally foreign to Central Africa. 

The Lozi have not remained in undisturbed possession of their plain. 
In 1836 the Kololo, a people with a Basuto nucleus, conquered them. 
Some Lozi remained under the Kololo. Others retired to northern 
swamps whence, in 1866, they annihilated the Kololo and regained their 
homeland. I consider that the Kololo conquest did not affect the Lozi 
kinship system. 

The Zulu cultivate their gardens and graze their cattle on the hills and 
in the valleys of south-east Africa.vThey live in small villages inhabited 
by groups of agnates and their wivesyOften the members of a number of 
neighbouring villages are related to one another by patrilineal ties to 
form a group [iimdeni) within a lineage system (ulusendo), in which 
a genealogically senior headman, descended by primogeniture from 
chief wives of his ancestors, is recognized as superior by other headmen. 
j^The people practise shifting cultivation, but villages tend to be settled 
in one area for many decades.'^s a village grows it splits into patrilineal 
branches, which normally establish independent villages within the 
lineage land-area. 

Both Lozi and Zulu are organized in powerful kingdoms. 


\/The Zulu kinship system is similar to that of the Swazi, described by 
Dr. Kuper elsewhere in this booki/The Zulu are divided into a number 
of exogamous clans (there is a word, isiBongo, for clan name, but not for 
the group) each ofjwhich is an association of dispersed agnatic lineages 
which are /fcorporate groups, of kinspeople who to-day trace common 
descent in a putative patrilineal line over six to nine generations .vThere- 
after there is a genealogical gap before clan founders are referred to. 
/However, this genealogical gap dates from the establishing of the Zulu 
kingdom in 1818-28 ; before that the people of the Zululand region were 
organized in small tribes whose nuclei were lineages with a genealogical 
framework of some twelve generations depth, since these nuclei were 
then localized groups. 

\/rhe lineages within the clan are usually residential units.*' Their 
segments are cores of villages and a number of segments living in one 
neighbourhood form a recognized group (umdeni) against other similar 
groups, in their own and other clans. uThe segments are placed in a 
genealogical structure yMembers of a segment hold rights in each other's 
herds and lands, consult on personal questions, and arbitrate in quarrels 
between members. vThe umdeni exists even when its constituent villages 
are divided by political boundaries between provinces of the nation 
or between nations. ^There are also rights and obligations between 


fellow-clansmen. vThey have residual inheritance rights; they should be 
hospitable to and help one another ; they may drink milk with each other ; 
they may not marry each other's sisters ; they share a common clan-head 
and his ancestors and a sacred clan-song, and in some clans common 
ritual and taboos. 

Each exogamous clan obtains wives from other clans by giving cattle 
which transfer to the husbands and their agnates rights to all the children 
of the wives, whoever are their genitors.'«A child is absolutely a member 
of his mother's husband's lineage and all his rights of inheritance lie in 
it.vHe has also certain rights and duties in his mother's lineage : he cannot 
marry in it and he can drink its milk, but he does not inherit in it or 
normally come under its ancestors. An illegitimate child, unless redeemed 
by his genitor with cattle, or by his or another's subsequent marriage to 
the mother, belongs to his mother's agnatic lineage. 
I' In the kinship terminology relatives of the father's and mother's sides 
are clearly distinguished, and lineal and collateral kin on either side are 
identified with one another in the first ascending generation i.'The father's 
brothers (real and classificatory) are 'fathers' and his sisters are 'female 
fathers'; the mother's sisters are 'mothers' and her brothers are 'male 
mothers'. Ndn the grandparents' generation the same terms are used on 
both sides, but the behaviour pattern to the father's father is of submis- 
sive respect, against familiarity on the other side.t4Cinship farther back 
is traced in the patrilineal line only.iln ego's own generation there is no 
distinction on a lineage basis : all of the ortho-cousins are brothers and 
sisters and both sets of cross-cousins are called by one term (umzala). 
-v/ Marriage is not allowed with any member of one's father's or mother's 
clan, nor with anyone who shares a common grandparent. JPhe unity 
of the lineage does appear to some extent in the terminology i^esides 
using the terms set out in Fig. 6 for siblings, a man or woman calls 
all siblings and cousins of both sexes in his or her lineage 'brother' 
(mfowethu). 'H^his fits in with the identification of father and father's 
sisters who are also members of the lineage: Children of all siblings, half- 
siblings, and cousins, are called 'my child', and all their children are 
grandchildren. Descriptive phrases indicate sex difference and the exact 

i' The husband calls his and his brothers' and sisters' affines umlanda, 
except his wife's father and mother who are mamezulu.)'Th.e wife and her 
family call her husband's kin mkwenyana, or the wife may say ekhakaw.i. 
But, in addition, both spouses use specific and distinctive terms for 
certain of their affinal relatives j. In the grandparents' and senior genera- 
tions they use the same terms as for their own kin, and similarly in the 
grandchildren's and junior generations .M^he husband can courteously 
use the/term for parents to his wife's parents, but there are distinctive 
terms .'^ The wife also has distinctive terms for her parents-in-law, but 
she, resident in her husband's village, tends to adopt the terms he uses. 



boba bobo bsbo boba \ 

■j"i%'i,t LjiuAci.„i,.i, 

!>> i~i ,ii , ^du;j„, 

~1 -"*" 

I — 1 I 1 r 

mialo m^ola mfo dade mi 

fniota mzola 

'/tj/ji? mtana mtana ■ 

vith descnptive terms 

lilllill 11 1 1 11 1111 1111 

tvifh deicriptiye terms. ~^ / r; ^/'^"^^Z ami ^ nith dcicripfive terms . 

Fio. 6 



and lower generations ■ 'mtonomtariomii remembered trt potntin 
Alternate upwards ■ 


k h ijt i„i,f, i„, i„ rj 

r" T^rr^^ J:;^!; — 

iliiilll llji,^ii iiiiilil 

Meif generation - all are 'mivana' l ,„qII tines 

Next qenerafton - • - ' /r.mWt/V 

Allernate doivn>vardj. 


vThe influence of lineage categories is seen in that the husband calls his 
wife's brothers and sisters, and also his wife's brother's child, mlamu. 

VThey are members of his wife's lineagel'But the same term is also applied 
to wife's brother's wifeVThe wife adopts the term used by her husband 
for his siblings of both sexes, and all their children are her children, as he 
calls her sister's child 'my child' i In short, a wife calls all her husband's 
relatives by the same terms as he would use, allowing for sex differences ; 
thus the wives of her husband's brother are her fellow-wives, while her 
husband calls them 'my wives'. 
J^T\iG relationships of afiines are marked by restraint and avoidance. 

%l A wife avoids the milk and meat of her husband's home until a special 
ceremony is performed for her. She must not go to certain parts of the 
village, must cover her body before her senior male affines, and avoid 
using the radicals of their names. \As she bears children the relationship 
gradually eases, and when her sons grow up her position as 'mother 
(father) of the village' is powerful .vA man avoids his mother-in-law, but 
is freer than his wife.vlie cannot eat meat at her home unless a beast is 
especially killed for him. 

i/"A man calls his wife's sister's husband mnazvethu (my comrade). 
Their children by the sisters cannot marry, but their children by other 
wives may. 

v/Zulu chiefs in the past could marry in their clans, thus splitting them 
and producing new clans, but this belongs to the study of political 
organization, as does the rule that a Lozi paramount chief can marry his 

No corporate unilineal group of kinsmen exists among the Lozi. 
Every child, legitimate, illegitfmate, and adulterine, has a right to make 
its home in a village of either of its mother's parents and to inherit there. 
It also has these rights with the kin of its father, who in Lozi law is the 
genitor. The difference among the Lozi in a child's relations with its 
father's and mother's kin is one of emphasis. There is no dominant 
unilineal kin-group, either in father's patrilineal or mother's matri- 
lineal lines. A child's proper home is with its father and it should go to 
his village to live and inherit ; but this may be the village of his father's 
father or his father's mother. If a child does not get on well with his 
father, if his father fails to provide properly for him, or if he quarrels 
with his paternal relatives, he will go to live at his mother's home, which 
again may be his mother's father's or his mother's mother's. I have even 
recorded Lozi growing up and establishing themselves at their father's 
mother's mother's and their mother's mother's mother's homes. While 
in Zululand there are very many cases in which there are disputes over 
the guardianship of children, which hinge on the question of who paid 
marriage-cattle for the mother and whether it was delivered, similar 
cases are practically unknown among the Lozi. Lozi say 'the child 
belongs to both sides', and the courts will not constrain a young boy (or 


girl) to leave his mother's home to go to his father. In any case, when the 
child grows up, it will itself choose freely where it wants to live. 

Manner of residence and rules of inheritance are consistent with the 
environmental situation of Lozi settlements. The small size of mounds 
means that Lozi villages can be inhabited only by small groupings of 
kindred, though usually villagers are related to one another. In a small 
family village on a mound, or in a family's part of a royal mound, it is 
impossible for the headman to place all his own sons and brothers' sons 
on land. Therefore some of them seek new land from the king or beg 
land from other relatives. Owing to the dotting of mounds in the plain, 
and because they are all owned, this may be far away. Frequently the 
king moves people into his royal villages. It happens thus that a village 
may fall below its full strength. Other relatives then move in to the 
vacant places. They have rights to inherit. When an elder dies no suc- 
cessor is definitely indicated. Men and women in all branches of the 
kindred meet to select an heir on the basis of his character — his wisdom 
and generosity and above all his ability to keep the village contented. 
They prefer to select a son, but he may be born of any wife of the husband 
and be one of her younger sons. If there is no suitable son they will 
choose a brother, a brother's son, a son's son, a sister's son or a daughter's 
son. The heir then replaces the dead man : he takes over all his land and 
cattle, including land and cattle allotted to wives, and divides them 
among the dead man's sons, grandsons, and nephews, in all lines, while 
retaining the largest portion for himself. The wives are not inherited; 
they may remain with their children, but they cannot claim land from 
the new headman, or they may return to their homes and marry again. 
Only succession to the chieftainships is by selection from agnatic 
descendants of ruling chiefs. 

The large choice of homes which is open to a Lozi is reflected in the 
mode of tracing descent. Instead of having a single important kin-group, 
reckoned from father through all his male ancestors, as among the Nuer, 
Swazi, and Zulu, or from the mother through all her female ancestors, 
as among the Ovambo and Ashanti, or in two lines, matrilineal and patri- 
lineal, as among Herero and Yako, the Lozi reckon descent theoretically 
in all, including mixed, lines, and in practice in every line with whose 
members they maintain connexions. They have no clans, but they have 
mishiku (sing, mushikii), -which I translate as 'descent-names'. This defines 
their attributes clearly, for the people sharing a single descent-name are 
not considered to be a group, nor do they have any specific obligations 
to each other, nor are there any ritual beliefs or practices attached to any 
descent-names. Nevertheless, people who share a descent-name feel 
that they are kinsfolk. The descent-name indicates that somewhere in 
his ancestry a man has a male or female ancestor or ancestors who, again 
in any one or more of his or her or their lines, had that descent-name. 
Lozi say everyone has eight descent names, but admit that everyone 





kive mkwekazi 

^ext generation - os yvife 
Ubukhwe (or) uiiulanda . 

mk^ekaj/ mkive mk^^e. ( mame) mawme mkwekazt 


= A 4=0. dadc ^ = 

nka mfo mktvenyanc, 




Collecti\>eh .as_. , , , , 
fkakami \ [as- her husband does 

(or) bakivenjjana J 

mameza/a mamezala 
(babo). (mame) 



1 = 

mne £ao' ndoda Zaknirtu mfa ZaUetu dade mktventjana 

=A mloyenyano ~mal(oii 


mtana mfa. 

■na m 

fano mtana mfana mfana 


might ultimately be able to claim all the Lozi descent-names. A Lozi 
will remember the descent-names of those kinsmen with whom he lives 
and co-operates. Thus, where a man's mother comes from near by, when 
asked he will give spontaneously the descent-name of his mother as well 
as his father; if her mother, his maternal grandmother, lives so far away 
that he never sees her family, he readily forgets her descent-name. 
I have found in questioning Lozi about their descent-names that most 
remember two, the father's in the paternal line and the mother's in the 
paternal line, while next most frequently they remember the father's 
in his mother's paternal line. Some Lozi give five descent-names, but 
very many know only two and others only one. 

People who share descent-names certainly do not expect to trace 
genealogical relationship. The Lozi, except for the royal family, trace 
their descent genealogically for only three or four generations from 
adults. There are no broadly based unilineal groups associating in 
common rights of residence, ownership, inheritance, production, &c., 
and consequently the structural depth of the kinship system is shallow, 
while it ramifies widely in all lines. The Lozi is interested in tracing his 
genealogical relationship with particular kin through any of the many 
lines of his descent rather than to claim or trace a distant relationship 
with someone in one special line. The patrilineal bias shows, however, 
in a tendency to change matrilineal links into patrilineal links by treating 
women in genealogies as men, or maternal uncles as fathers, where they 
link people with a village. 

A Lozi may marry a woman with the same descent-name as himself. 
He should not marry a woman with whom he can trace genealogical 
relationship in any way, and we have seen that Lozi fail to trace genea- 
logical relationships beyond four generations. In fact intermarriages 
occur earlier, while relationship can still be traced. The Lozi say, sikowa 
sakukzvata okusatunda kasapula, 'a kindred gets so big that relationship is 
broken with time'. They apply this saying when a young man wants to 
marry a girl whom his elders consider is related to him. The elders 
protest, the young man counters with the maxim and marries her. 
It is significant for my later analysis that the couple will be cursed : 'You 
have chosen to marry. Only death may now divorce you; you may not 
separate from each other,' They establish a new relationship by affinity 
which ultimately produces new cognatic ties between the two kindreds. 
Lozi say that a man should not marry a woman with whom he is linked 
only by patrilineal ties, but such marriages occur. 

Most Northern Rhodesian tribes have matrilineal exogamous clans but 
also forbid intermarriage with ortho-cousins on the father's side. They 
allow, and often prefer, marriage of cross-cousins. Among the Lozi, as 
shown on the genealogical chart (Fig. 6), all cousins are classed as brothers 
and sisters and marriage with them is forbidden. Thus, if the relationship 
is remembered, a man will call his mother's father's mother's daughter's 


son's daughter 'my sister' (kaizeli). Here I have deliberately given an 
example in which there is a jump from maternal to paternal line at almost 
every step, which is the most complicated combination. He should not 
marry her, though he may force the marriage by claiming that the 
relationship has grown old. That is, while the older people still call each 
other 'brother' and 'sister', the young people say in effect, 'There is no 
longer relationship between us.' By marrying they establish a new set of 
relationships. They no longer call each other 'brother' and 'sister' but 
'husband' and 'wife'. The husband's and wife's siblings also stop calling 
each other, and the spouse of their sibling, 'brother' and 'sister', and they 
become 'brothers- and sisters-in-law' (balamu) to each other. Their 
fathers or their mothers, or their father and mother, depending on 
through whom the old relationship was traced, are no longer siblings 
(brothers and sisters), but call each other 'the other parent' (mukwange 
mushemi). The spouses no longer call the parents 'my father' or 'my 
mother', but 'my relative-in-law'. Thus the marriage breaks cognatic, 
and establishes affinal, relationships. The children of this couple will 
call the children of their parents' siblings 'my brother' and 'my sister', 
and though the marriage ban is broken for the one generation, it is 
renewed in a lower generation by the creation of new cognatic relations 
between their descendants. Since the spouses were related they will 
probably share at least two descent-names, and their children will have 
the same descent-names in more than one of their ancestral lines. I fre- 
quently found this in recording genealogies. In this dominantly patri- 
local and patrilineal society relationship is obviously more easily broken 
where there is a female at some point in the chain of genealogical con- 
nexions; and more easily broken where there is descent from one man 
through two wives than where it is through one wife. 

Even where there is no blood-relationship but there are certain 
marriage connexions marriage is banned, but here relationship becomes 
weak earlier. The Lozi say mwaboyu ta mwaboyii, 'the relatives of a man's 
relatives are his relatives' (see Fig. 7). Since divorce is frequent and 
a widow is rarely married again by one of her husband's kinsmen, 
a man's half-siblings through his mother or father are related through 
his mother's or father's other marriages to people who have not even 
putative blood-connexions with him, but through his half-siblings he 
claims them as relatives. Men and women related thus call each other 
'brother' and 'sister' and do not marry. Similarly, a man calls his wife's 
brother's wife 'my sister', and most of my informants agree that he 
should not marry her if his wife's brother divorces her or dies. A woman 
calls her co-wife's brother 'my brother' (see Fig. 8), and again most 
informants say that he should not marry her if she is divorced or 
widowed. I know of one marriage in the former set of relationships; the 
man's earlier wife, sister of his new wife's divorced husband, left him 
and he did not appeal to the court. 



/^exf qenerah'on ••- All are l^uku " 

Neoct veneration :- /III men ore "ndate' all ivomen ore'me ' 



• ndole „ I " mMitem, 

(no musoh} (ndaTe 


0=^-, , — 


"'"""^ mukweOLiani 

Fig. 8. LOZI 



Nexfqenerafinn -nil nr/= ' kuku' 

r-L i i 1 1 


TiuknenLjani ma fun 


jj'k koaeli^tf^^a muhalizo 


Men who marry related women, or have sexual relationships with the 
same woman, are regarded as related and address one another by the 
special term mufubalume ('fellow-husband)'. This denotes a man who 
has slept with the same woman as oneself, and it applies between two 
lovers of one woman, between an adulterer and his cuckold, and between 
a man who has married a divorcee and her former husband. Lozi explain 
it as 'a man who quarrels with you because of sexual intercourse'. It is 
extended, on the principle of the equivalence of siblings, to apply 
between a man who has married a widow and the brothers of her dead 
husband, and between men who have married classificatory sisters. In 
general, the descendants of these men should not marry, but though 
their marriages will be resisted, it will not be as strongly as where there 
is a real cognatic relationship. 

Thus in all lines except the purely patrilineal the Lozi allow marriages 
where a distant genealogical connexion can be traced, and ultimately 
welcome them, to renew the relationship. Perhaps, therefore, these 
marriages are cursed with a ban on divorce. /This process also occurs in 
some Zulu lines jYit is approved for a man to renew relationships by 
marrying a woman of his father's mother's or mother's mother's clan 
with whom he cannot trace links. He cannot marry a known cognate. 
Nor can he marry a woman of his father's or mother's clan, even though 
she belongs to a different lineage with which genealogical links are no 
longer remembered. 

The Lozi thus tend to trace and merge relationships as widely as 
possible in the present living generations of immediate cognates, and 
not to trace relationship between kinsmen in one line who may be widely 
dispersed over the country. Nor does the Lozi system generally differ- 
entiate between the lines. This appears clearly in Figs. 6, 7, 8. In the 
same generation siblings are distinguished by sex : a sibling of the same 
sex is miihulwani if older, munyani if younger, but kaizeli is a self-reci- 
procal term between siblings of opposite sex. This sex distinction is not 
maintained downwards, where there are only two terms used: child 
(mzvana) and grandchild [mwikulii). Descriptive terms, 'of boy' and 'of 
girl', are used to distinguish between them, and descriptive combinations 
are necessary to distinguish the exact linkage. Great-grandchildren are 
again 'children', great-great-grandchildren again 'grandchildren', and 
thereafter in theory the alternation proceeds regularly, though the 
relationship is rarely traced. The same process occurs in the senior 
ascendant generations: all grandparents and great-great-grandparents 
are kuku, irrespective of sex. In the parent's generation all the father's 
brothers and sisters are ndate (my father) like the father, though the 
father's sister can be called 'female father' {ndate wamusali). Similarly, 
the mother's sisters and brothers are all me (my mother), though the 
mother's brother may be called malume ('male mother'). The terms of 
the parents' generation are extended to the great-grandparents and 


great-great-great-grandparents, but the distinction between father's or 
mother's sides is overridden : all males are ndate (father) and all females 
are me (mother) on both sides. 

This merging also extends to the spouse's relatives in lower and 
higher generations. People whom your spouse calls 'child' or 'grandchild' 
or 'grandparent' or 'great-grandparent' are similarly related to you, and 
you may not marry them. You primarily distinguish your parents- and 
children-in-law by the self-reciprocal term mukwenya?ii, but you also 
call them respectively 'father' and 'mother' and 'child', and your 
spouse's more distant connexions, such as mother's brother, you address 
by the same terms as your spouse does. His or her siblings, irrespective 
of sex, are balamu (sing, malamu), or husband (w.s.) or wife (m.s.), again 
irrespective of their sex. Beyond that a man or woman takes over the 
terms of address which his or her spouse uses for kinsfolk. 

Thus among the Lozi relationships in all lines of descent tend to be 
merged by generations. The actual importance of the different lines 
varies from individual to individual, with a general bias to the paternal. 
But the mother's immediate home is more important than the home of 
someone distantly connected in the patrilineal line. Nearness of cognatic 
relationship, and not membership of a unilineal kin-group, is empha- 
sized. There are slight differences in the behaviour patterns to father and 
mother's brother, and to mother and father's sister, but in general they 
are similar. Father and mother's brother are both treated with affection- 
ate respect, but the relationship is easier with the latter. However, while 
among the Southern Bantu a man may claim something he wants from 
his mother's brother, among the Lozi a man could take anything of his 
father's or mother's brother's or of any other relative {kiifunda : this has 
now been declared theft). A man can claim help from his mother's 
brother as from his father, and father and mother's brother share 
marriage payment of girls equally. My impression is that generally 
a man is friendlier with, and takes more care of, his sister's children than 
his brother's children. The father's sister is treated with considerable 
respect, while people behave with more ease to mother's sister. 

There is one variation in behaviour to parents' siblings of the same 
and different sexes. A mother's brother also calls his sister's daughter 
'my wife' and his sister's son 'my fellow-husband' (mufubalume), and 
father's sister also calls her brother's son 'my husband' and her brother's 
daughter 'my co-wife' (muhalizo), as well as 'my child'. Grandparents 
and grandchildren of opposite sex can address each other similarly. 
Between persons of opposite sex this form of teasing ceases when the 
youngsters, who find it embarrassing, grow up : they use the terms grand- 
parent, father, mother, and child, out of respect. But mother's brother 
and sister's son still call each other 'fellow-husband' because the nephew 
can sleep in his uncle's hut, fondle his uncle's wife's breasts, and marry 
her if she wants to when his uncle dies. Lozi also say that these terms are 


used because these relationships are friendly and easy, and the young- 
sters can consult their elders on sexual matters (e.g. about the impotence 
of niece's husband or frigidity of nephew's wife) ; but clearly they fall 
into the pattern of joking relationships, since we have here a woman 
who is a father, and a man who is a mother. Similarly in the case of 
grandparents there is conflict because the terms of nomenclature iden- 
tify odd and even generations despite the great difference in age. 

Behaviour to affines varies markedly according to generation. Parents- 
in-law and children-in-law, who address each other by a self-reciprocal 
term (mukwenyani), are treated with great respect. Son-in-law and 
mother-in-law, and father-in-law and daughter-in-law, practise mutual 
avoidance and if they touch each other must pay damages 'to wash away 
the fault'. They must not sit alone in a hut or use bad language before 
each other. A son-in-law should not sit on the same mat as his mother-in- 
law, (^his respect is not, however, carried as far as among the Zulu, 
where the daughter-in-law must avoid using any word like the names of 
her senior male relatives-in-law, and must not walk through the cattle- 
kraal or the men's place. 

Directly related brothers- and sisters-in-law behave freely to each 
other. They are 'man and wife', even when of the same sex, though they 
more often call each other mulamu than 'wife' and 'husband'. But they 
should not marry. They joke and tease each other. A spouse takes over 
from his or her partner behaviour patterns to grandparents, nieces, and 

The corporate groups of kindred which in theory have lasted from the 
beginnings of Lozi history are the villages. Since people have constantly 
movedor been moved by the king between villages and neighbourhoods, 
the pattern of kinship links in local communities is constantly altering. 
No umlirieal framework exists to which they can be referred ; only royal 
people close to ruling kings can trace their descent from Mboo, the 
founder of the Lozi kingdom, son of God and His daughter. This 
agnatic line is the core of the whole nation's history. Nevertheless, for 
commoners, too, kinship links are all-important in defining membership 
of a village and rights in it, and denoting relationships with and claims 
in other villages. Descent-names provide the ultimate reference for 
common origin by blood. By whatever links cognates are united they 
express their kinship by citing their descent-names. Correspondingly, 
strangers who find they share a descent-name feel that they are kin. The 
descent-names have this high value because they originated with the first 
Lozi who lived with King Mboo. They symbolize the notion that kin- 
ship, relationship by blood as such, comes from the beginning of time, 
even though the Lozi can only remember their genealogical ties back for 
a very few generations. They give historic depth to the shallow genealogi- 
cal system. Every descent-name is related to some stage of Lozi history: 
the Lozi names are those of the first inhabitants of the plain, those of 



foreign tribes date to the absorption of these tribes in the nation. The 
villages which are the corporate groups of kindred are built on mounds 
which provide the only sites for homes in the plain. Certain of these 
mounds are known as those of the Lozi's distant forebears, and each 
Lozi descent-name is referred to an ancestral village on one of these 
mounds. The possession of a Lozi descent-name enables a man to point 
to his ancestral village, and claim to be a true Lozi, a true inhabitant of 
the plain, a descendant by blood of the original settlers related to the 
daughter of God, a kinsman of all Lozi. 

, The Lozi kinship is thus largely consistent with their modes of 
-->;■ residence and production. We have seen that it is physically impossible 
for large groups, tracing unilineal relationship, to live on the mounds 
in the plain or on neighbouring mounds. As the group on a mound 
increases some members must seek land at other mounds far away, while 
the people at each mound must depend on their neighbours for help in 
prosecuting their many dispersed and contemporaneous activities. With 
these they intermarry and become related in a variety of ways. Cognatic 
relationships are periodically broken by marriages between people who 
can still trace relationship. These marriages establish new cognatic ties, 
and, I repeat, it is these marriages alone which are cursed by a ban on 

Nevertheless, the Lozi pattern of kinship, despite its unique environ- 
mental setting for the region, has attributes which are characteristic of 
Central Africa. The system of nomenclature of the Bairu of Ankole, in 
distant Uganda, is more like the Lozi system of nomenclature than any 
other African system I know of,' but the free choice of homes for 
children, and the inability of either set of kin to claim children against 
the other, produce similar results in many of the tribes of Central 

I summarily contrast the Zulu system. /There a man's rights of 
inheritance reside only in his agnatic lineage.|/His mother's and his 
grandmothers' homes and people are very important in his life, but his 
major ties of kinship and his economic interests pull him to his agnatic 
lineage.vA Zulu inherits only from his agnates and he usually lives with 
and obtains land from them.\For purposes of inheritance and familial 
pride Zulu are more interested in their distant patrilineal relationships 
than in close relationships in other lines. 

' K. Oberg, 'Kinship Organization of Banyankole', Africa, xi. 2, April 1938. 

^ See e.g. A. T. and G. M. Culwick, Ubena of the Rivers, London, 1935; 
G. G. Brown and A. McD. B. Hutt, Anthropology in Action, Oxford, i93S. 
pp. 82 ff. ; A. I. Richards, Land Labour and Diet in Northern Rhodesia, London, 
1940, chap, viii; and W. Allan, M. Gluckman, D. U. Peters, and C. G. Trapnell, 
Land-holditig and Land-usage among the Plateau Tonga of Mazabuka District, 
Northern Rhodesia, Rhodes-Livingstone Paper No. 14, 1948, chap. iii. 



v-nThere are thus four striking differences between the Lozi and Zulu 
systems; the presence of the agnatic hneage among the Zulu and the 
Zulu distinction betvveeiv'^ortho-cousins and cross-cousins, neither of 
whom they can marry, against the absence of this lineage among the 
Lozi and their merging of all cousins as brothers and sisters;-1:he Lozi 
alternation of descending and ascending generations as children and 
grandchildren, and as parents and grandparents, respectively; and the 
Lozi merging of cognatic kin in numerous lines, and these kin with 
affines, to a greater extent than do the Zulu. 

I observed no important differences in the psychological relationships 
of husbands and wives, though their social relationships differ widely. 

^Zulu and Lozi marriages do not generally spring from romantic attach- 
ments, and their concept of love between the sexes is on the whole 
restrained, though sexual attraction between men and women may make 
them brave severe sanctions.t/In both societies there is only a limited field 
in which men and women can associate. tThey sleep together and produce 
children together, and their union founds an economic unit, but in 
general social and public life their contacts are limited.!^ man seeks 
companionship with his fellow men, a woman with her fellow women. 

'/Except in the intimacy of the marriage bed it is difficult for a man to 
enjoy the company of a woman ^ere a Zulu is freer than a Lozi : he can 
spend his time with his sister, whom he can call 'my brother'. A Lozi 
may not touch his sister or a female affine without committing sindoye 
(a breach of the sexual ban) and he must not be alone with her in a hut. 
If he is, he is accused of sorcery. At beer-drinks men and women sit 
apart, and at dances they stand on opposite sides of the circle. By Lozi 
standards if a man walks along a path with another's wife w^ho is not 
related to him, or gives her snuff or a drink of beer, or even speaks to her 
when no one is by, he commits 'adultery', even if he does not sleep with 
her. Men are often afraid to work gardens for, or help, the wives of their 
brothers who are labour migrants, lest their brothers charge them with 
being lovers of their sisters-in-law.i-In this background Lozi marry for 
se^al satisfaction, to secure an economic partner, to obtain children, 
but not to find a companion.LThis applies in general terms to the Zulu, 
save that, as marriage is far more stable among them, suspicions of 
adultery do not arise so easily .i^For example, men always help their 
absent brothers' wives. 

uHowever, the importance of the marital relationship must not be 
underestimated. Lozi trust and confide in their mothers and sisters, 
with whom their ties are enduring, rather than in their wives, with 
whom their ties are ephemeral. But a wife has a right to know all her 
husband's affairs and she is, with him, 'lord of the home'. If he neglects 


to treat her thus, and favours his kinswomen, she will be granted a 

>«<Both societies are polygynous, and jealousy between the co-wives of 
one man for his sexual and other favours is pronounced .vlt is recognized 
in proverb and saying and appears in charges of sorcery .\2n this similar 
background they evaluate sororal polygyny in opposite ways which reflect 
graphically the difference between the two marriage systems. 
^The Zulu approve of marriage to two sisters." Often a woman will urge 
her husband to marry her younger sister, who will become subordinate 
to her, so that she may have a 'companion' who will share her labours 
and her loving care for her children without the jealousy of a stranger 
co-wife. j.'The love of sisters overcomes the jealousy of polygyny. '»-The 
husband cannot marry his wife's senior sister, but this is because she 
cannot be a subordinate wife to her own junior. Among the Lozi it is 
considered very bad to marry even a classificatory sister of your wife 
while the latter is alive. Your wife and relatives will protest strongly, 
and I have records of wives who walked out on their husbands when 
they married thus. Lozi say, 'The jealousy of polygyny spoils the love 
of sisters. It will break up their family.' They must not be put in the 
position of competing for a man's favours or their children in the posi- 
tion of competing for their inheritances in both their lines. Among the 
Zulu their inheritances and relative status would be fixed. 

This variant approach to sororal polygyny is one of many differences 
in the family complexes of Zulu and Lozi. The Lozi family of parents 
and children is very unstable as a domestic grouping, and in addition 
divorce is frequent and easily obtained by both man and wife. A man 
who wishes to divorce his wife merely sends her home and need give no 
reason. The wife, unless she can persuade her husband to divorce her, 
must seek release in court, but she can gain it on many grounds. More- 
over, she will be released by the court if the husband does anything 
which can be construed as 'driving her away'. If he says 'get out' to her, 
if he pushes her, if he takes from her any article which he may have 
given her, if he favours another wife, if he does not sleep with her, she 
will get a divorce. A woman has very little security in marriage; a man 
has not much more, unless he is patient, controlled, and careful not 
to act unguardedly in a way the courts can interpret as dismissal of 
his wife. 

In Loziland to-day people discuss divorce all the time.'^n the other 
hand, divorce is almost unknown in Zululand.vlt can be obtained in 
a magistrate's court, but men do not want to divorce their wives and 
women rarely seek divorce.Un native opinion a woman should only be 
released if her husband is intolerably cruel. ^I never heard divorce dis- 
cussed as a social problem in Zululand: the problem was always 

Obviously it is difficult to eliminate the effect of modern conditions 



on the stability of marriage (We can say that marriage has always been 
stable in Zululand, and old records on the Zulu' indeed affirm that they 
were a 'moral' people among whom divorce, adultery, and illegitimacy 
were rareVChastity was highly valued, and was part of the ethical code 
enforced by the age-set regimental organization which was directly under 
the king.vRegiments could only begin to marry when the king gave them 
permission as a set to marry a younger age-set of girls, who by this time 
would be well in their twenties yA man in an older regiment could marry 
a girl whose set had not yet had its mates indicated. ''Intercourse intra 
crura by youngsters was allowed.vlf an unmarried girl became pregnant 
by a young warrior, both they and their families were liable to be killed. 
/^Therefore she would be hurriedly married to a man whose regiment had 
leave to marry. \fLu\u. themselves attribute the high and increasing 
illegitimacy rate to Government's banning of this sanction. Girls, be it 
noted, still do not marry till they are in their twenties:/ Adultery too was 
severely punished, with death, or flogging with thorny branches, or 
cacti were thrust into the woman's vagina.^ 

On the other hand, early travellers to Loziland, and every missionary 
and administrative record, describe men and women as promiscuous 
and marriage as unstable. Holub, one of the earliest visitors to the 
country, reported that marriage was lax and particularly that women, 
including the queens, looked upon the marriage tie as loose. Macintosh -^ S^ 
wrote that 'marriages are unmade at a moment's notice . . . the wife 
being deprived of everything, even her beads, blankets, and clothes, 
and told to go. . , , Should a man take a liking to someone else's 
wife he will have an interview with her and bring her home.' Texts, 
life-histories of old men, and genealogies, confirm this. The evidence 
contradicts Lozi assertions that girls, who are now usually married 
shortly after puberty, used in the past to be much older before they 
married, and grew up, unspoiled, with boys at cattle-posts where 
strict morals were inculcated. Seduction, adultery, and abduction of 
wives were undoubtedly rife. The quotation from Macintosh bears out 
what Lozi told me: that adulteries, divorces, and abductions did not 
come to court at the capital until 1918-20, when the paramount chief 
issued an edict saying that his court would hear these cases. To-day it 
does little else. Previously cases of adultery and abduction of wives were 
settled by the kindreds involved. The wrongdoer hid from the wrath of 
the cuckold and his kin, who came armed and had to be appeased with 
gifts. Only if blood were shed did the king intervene. 

Therefore it seems reasonable to assume that divorce was always 
relatively frequent in Loziland. Old men bemoan the looseness of modern 

' I must emphasize that I am writing here of the Zulu proper, and not of 
Natal or Cape Nguni. 

^ See M. Read, 'The Moral Code among the Ngoni' (of Zulu stock), Africa, 
xi. i, Jan. 1938. 



women in contrast to the fidelity of their mothers, and undoubtedly 
modern conditions have increased marital laxness.i/fiut modern condi- 
tions have borne more severely on the Zului^^evertheless, despite the 
abolition of their kingship, and the effects of labour migration, &c., Zulu 
marriage is firm and enduringivMowever long a labour migrant is away, 
his wife should not be divorced: she can bear children for him by his 
kinsmen or a lover. If a Lozi is away for three years, now, his wife is 
entitled to a divorce : the period has been steadily reduced from seven 
years. I argue therefore that modern conditions have given full rein to 
tendencies already present in Lozi society, and that when the paramount 
chief tried to control divorce and abduction he was not tackling a newly 
created problem. 

I begin my analysis from the dissolution, and not from the contract- 
ing, of marriage because its instability among the Lozi and its stability 
among the Zulu set the fundamental difference between the marriage 
complexes in the two societies. On this hinges a whole set of rules. 
'/Initially, we may see here why there is a different reaction to sororal 
polygyny.' Among the Zulu this is approved because the sisters help one 
another and it is unlikely that divorce of one sister will spoil the marriage 
of the other or that their competition for their husband will lead to 
a divorce.VlVlarriage to two sisters binds the lineages involved more 
firmly together. Among the Lozi, each marriage is potentially so unstable 
that a sister co-wife would very likely get involved in trouble with her 
husband arising from a quarrel with her sister, and their competition is 
likely to lead to the divorce of one or other. We have seen how widely 
the concept of 'sister' extends. The Lozi also disapprove of a marriage 
which gives rise to two sets of kinship terms for the same people, and 
which produces two sets of children with identical attachments in all 
their lines. They are interested in spreading the net of their cognatic and 
affinal connexions as widely as possible. Marriage to two women who are 
sisters would prevent this. However, a man can marry, though not claim, 
the sister of his dead wife. Here her interest in her dead sister's children 
is emphasized, and he renews his relationship to his affines. It is also 
proper for a woman to marry her sister's husband's brother, especially 
a half-brother. 
.,-t/ln brief, among the Zulu marriage moves a woman from her kin-group 
to her husband's ;( the 'love of sisters' is an additional tie between 
co-wives. Among the Lozi a woman's main tie remains with her kin, 
jyt" ' p." ''and is not with her husband's: 'the love of sisters' must not be endan- 
' , gered. 

I The Zulu in Natal have the same types of marriage-forming families 
^ as are reported by Evans-Pritchard from the Nuer in the far distant 

i6^' Sudan^ and I follow his definitions. 

^ i z' \ 'X Outside of marriage, there is the natural family, which is formed by 

' Op. cit., pp. 5 ff. 


a man living with his concubine and her children. This is rarely found 

V2. There is thesimple legal family formed by a man and a woman for 
whom he has given marriage-cattle, and her children >If a man has more 
than one wife his household constitutes a compound family /Among the 
Zulu these wives are gradedvOne, the chief wife of the great house, who 
may be married late in life, will produce the main heirKShe has placed 
under her a number of subordinate wives /Another wife is head of the 
left-hand house which also contains subordinate wives, and another 
group of wives in very big families may form the right-hand house. 
A separate wife is sometimes set apart in the chief house to bear an heii^ 
to it if it has no sons.' Finally, the first wife married (who may be called 
mame, my mother) becomes mother of the uyise wabantwajia (father of 
the children), a son who ritually replaces the father on his death./These 
are indigenous gradings. The Natal Code of Native Law hoped to 
reduce litigation by ruling that the first wife married is the great wife. 
(y(^3 . Where the husband dies and an approved relative of his lives with 
the widow and the children, he begets more children for the dead man ; 
this is the leviratic family. .The pro-husband does not pay marriage- 
cattle. tx^vans-Pritchard has emphasized the important difference 
httwttnleviratic marriage, where the dead man is still the husband of his 
widow and her future children are his and do not belong to the man 
living with her, and^tvidozv inheritance, where the widow is expected or 
even compelled to rharry a relative of the dead man. This is a new union 
and future children belong to the new husband, iin leviratic marriage 
future children address their genitor by the same term as do their 
mother's other children ; in widow inheritance he is their father, so that 
they and their mother's other children may address him by difi'erent 
terms. The Zulu have leviratic marriage K Sometimes a Zulu widow goes 
to live with a stranger; then, in Evans-Pritchard's terms, she is living in 
widow concubinage]^ But, as among the Nuer, the children still belong to 
the dead man. Until the reign of King Mpande (1840-72) a woman's 
person was also permanently transferred to her husband's agnatic 
lineage. A widow could never go to a man not related to her husband 
without his being prosecuted. Mpande passed a law allowing a widow to 
go to any man she liked, but the children were to remain the dead 
husband's. Her lover, their genitor, could claim one beast from the 
marriage payments for his daughters, 'the beast of the knees', since he 
had gone on his knees to beget them, but correspondingly he should 
contribute one beast to the marriage payments given by his sons for their 
wives. Latterly, under Government regulation which prevents women 
from being compelled to marry against their will, widows can marry 
other men, but the new husband must return the marriage payments to 
the heirs of the widow's dead husband, less one beast for each child born. 
The Zulu, like the Nuer, also have ghost marriage. |X"here are two 


forms of this : (a) if a man was betrothed and died, his fiancee should 
marry one of his kinsmen and bear children for the dead man, as if she 
were a widow ; and (b) a man may 'waken' a dead kinsman who was never 
betrothed by marrying a wife to his name and begetting children for him. 

U'5. As among the Nuer a rich and important Zulu woman can marry 
another woman by giving marriage-cattle for her, and she is the pater of 
her wife's children begotten by some male kinsman of the female 
husband. 'They belong to the latter's agnatic lineage as if she were a man. 
' If a man dies leaving only daughters and no son, the eldest daughter 
should take his cattle and marry wives for her father to produce sons for 
him.*' This and the preceding forms of marriage are weighty customs 
enforced by ancestral wrath, and they arise from the importance of 
continuing the agnatic line. 

'^ 6. An impotent Zulu man can marry wives with whom he asks a 
kinsman to have intercourse, so that he can get children. iThis is a simple 

(_^^7. Among the Zulu in addition it is considered proper that when 
a chief or important man marries his chief wife, the heads of all the 
subordinate social groups under him should contribute the marriage- 
cattle. I Her son is then a child of the group who gave the cattle, and 
therefore should rightly succeed to the heirship. 

i/' These forms of marriage all depend on the legal consequences' which 
flow from the cattle which a Zulu gives for his bride. i/The cattle make 
him pater of all her children, whether or not he is their genitor.^he 
Zulu says 'cattle beget children', which means that the marriage-cattle 
given for a woman indicate the giver (usually her husband) to be the 
pater of her children. Once a man has given marriage-cattle for a woman 
he is pater of all her children, even if she deserts him or if he dies. 
,/^herefore men who are dead or impotent, and women, all physically 
incapable of begetting, can be paters of children who have been begotten 
by other men, and a whole group can be pater to children whose mother's 
marriage-cattle it has contributed. A man secures his illegitimate children 
either by giving cattle for them separately or by securing all an unmarried 
woman's children by subsequent marriage (giving of cattle) ; otherwise 
they belong to the mother's agnatic lineage. /The purpose of the cattle 
is to procure children, and Zulu say that in the past they began to pay 
cattle only when a child had been born. 

i/The effect of this rule is to give great permanence to legal marriage, 
and to ensure that the members of the simple legal family live together. 
v/Marriage endures beyond the death of the husband and father, for he is 
still married to his widow now living either in leviratic marriage with 
a kinsman of his, or in widow concubinage, and he is still father of her 
future children. These children take his clan name and have rights to 
live and inherit in his agnatic lineage; they have not those rights else- 
where vThe death of the wife may break the marriage, but, though it is 


not compulsory, it is considered proper, if the husband has been a good 
spouse and son-in-law, that she should be replaced by a younger sister 
who steps into her place under the sororate. The sister is now the dead 
woman and the family is reconstituted. If a woman dies before bearing 
children she has not fulfilled the purpose for which the marriage-cattle 
were given, even though she has worked and given sexual services, and 
her family must replace her with a younger sister or else they must 
return the marriage payment.'/ A younger sister should also be sent to 
bear children for a barren woman, or her parents must return the cattle. 
^If they do the latter, a good Zulu does not necessarily send his barren 
wife home; he may use the returned marriage-cattle to obtain a wife 
whom he puts 'into the house' of his barren wife to bear children for her. 
Thus even maternity can be fictitious. I suggest that the low rate of 
divorce is another aspect of this general permanence of marriage, the 
effect of which is to bring in a woman from another clan to bear children 
for the agnatic lineage which, under exogamous patriliny, is unable to 
reproduce itself. 

u^ulu marriage thus constitutes a long enduring union between the 
spouses, which extends to their kin, above all their agnatic lineages. This 
is still the position in Zululand now when European statute has fixed 
the amount of the marriage payment, but in the past these consequences 
were even clearer. Then there was no fixed amount of marriage pay- 
ment: bridegrooms were proud to give what they could. tThe first beast 
was sent after the birth of a child, and thereafter the son-in-law, known 
as 'the handle-of-a-hoe', gave a beast to his father-in-law every time one 
of his wife's brothers got married I'He was also supposed to help his 
father-in-law with cattle whenever the latter was in need.^ulu say that 
cattle continued to pass as long as the relationship between the families 
was remembered, in both directions but chiefly to the woman's home, 
even after the original spouses had died. 

Evans-Pritchard's description of Nuer marriage and the family 
emphasizes that the legal bonds resulting from marriage are very strong, 
though often a man's wife may not live with him or his proxy. Nuer 
women do desert their husbands and many widows live in widow 
concubinage and not in leviratic marriage. Nevertheless, the cattle given 
at marriage always make the giver pater of a woman's children, and 
lineage ties and rights of inheritance draw the children back to their 
pater's home. The same rule determines the affiliation of children in 
Zululand. There, in addition, the members of a legal family almost 
always live together. Most widows go into leviratic marriage and few 
wives desert their husbands. Courts order deserting wives to return to 
their husbands. 

\/VVe may say then of the Zulu, as Evans-Pritchard did of the Nuer, 
that 'the social principle of agnatic descent is, by a kind of paradox, 
tracedthrough the mother, for the rule is that in virtue of payment of 


bride-wealth all who are born of the mother are children of her husband. 
Thus all the children of a woman may be children of a man whom their 
mother has never seen. Or a woman may bear children by many men 
/ yet they all have one father.' '(-The principle is expressed in Zulu, where 
/ the word isizalo means both the womb of a female and a man's origin, 
i.e. his tribal or clan name, which is derived from the pater, 
l^/ln ranked Zulu society the principle that 'agnatic descent is traced 
through the mother' is additionally important, because the status oi sons 
depends on the status of their mothers in the compound family .KHence 
the importance of the levirate which, as among the Nuer, places the 
children in a genealogy which reflects the segmentary lineage system. 

The Lozi do not have leviratic marriage, ghost marriage, woman- 
marriage-to-a-woman, and do not contribute to the marriage payments 
for the wives of their leaders. A woman is always released from marriage 
to an impotent man, who is 'another woman'. Wives are not graded in 
any special status. By order of marriage, irrespective of their relative 
ages and tribal status, they are referred to as the first, second, &c., wife, 
and senior wives are supposed to be treated respectfully by junior wives, 
and have greater rights and obligations. But no wife has instituted 
- authority over another, despite the fact that the Lozi is a ranked society, 
1 and no wife has status conferring special rights in inheritance on her son 
or sons. The main heir selected may be the son of a young wife, and 
among one wife's sons a younger may be preferred to an elder son. 

<A Lozi wife and her children have no status to persist after her divorce ; 
even if a Zulu wife is divorced her children retain their positions 
derived from her status in the household. 

Until the 1900's, when the average Lozi commoner married he gave 
the bride's parents a few hoes and mats. It is comparatively recently 
(1900-10) that their paramount chief Lewanika issued a law that two 
beasts or £2 should be given for a virgin and one beast or ^^i for 
a girl who is not a virgin (a widow, a divorcee, an unmarried girl who 
has had a child, or an unmarried woman who is so long past puberty 
she should not be a virgin). To constitute legal marriage now there must 
be consent of the parties and consent of the bride's parents or guardians, 
which is implicit in their accepting the marriage payment. The last is 
the most important indicator of a legal marriage, at which a Lozi court 
looks to see if it will award adultery damages if another man sleeps with 
the woman. If her parents have consented to the union but the man has 
not delivered the cattle or money, the court will not award him damages. 
It will say that she is his concubine, 'a wife of the country'. Thus the 
marriage payment gives a man the right to exclude other men from his 
wife, and the right to control her. Though divorce is frequent and 
a woman can get it easily, if her husband resists she must go to court. 
If she refuses point-blank to live with her husband, neither he nor the 

' Op. cit., p. 64. 


courts can compel her to do so. If she deserts him he can always in this 
polygynous society take other wives or concubines, while she can only 
he with or go to another man under threat of an adultery charge, when 
her lover will pay damages of two cattle to the husband, or three if she 
has come to him as a wife, and she will have to pay a fine of one head to 
the court. (These fines were raised by ^1 in 1947, after this essay was 
first written.) Thus ultimately a woman, if her lover is eager to have her, 
can force a divorce from her husband. When an abductor pays damages 
to the husband the latter loses his rights over his wife. If she agrees to 
return to her husband she remains his wife; but she can remain with 
her abductor. The third beast he gave to the husband counts as a 
marriage-beast. This usually happened in the past. To control these 
moves the Lozi have made it a criminal offence for a woman's abductor 
to give her parents the fine she must now pay the court. 

In Lozi law-^ delivery of the marriage payment is essential to pass these 
rights over the wife's person; in Zulu law agreement to deliver is 
sufficient. This does not affect the issue that the husband gets these 
rights in both societies, though it means that what is a proper marriage 
in Zulu eyes is not reckoned to be so by the Lozi. 

Payment of the marriage-cattle does not give the Lozi man the right 
which a Zulu has to all the woman's children. In general, under Lozi 
law the pater is the genitor, not the mother's husband. The difference 
between the two sets of rules appears clearly in rights to illegitimate 
children. A Zulu must redeem his illegitimate children with cattle and 
also pay seduction damages to the girl's father and mother as well as 
a fine to the chief, and he can never claim his adulterine children. If 
a Lozi gives an unmarried girl a child he pays her father a beast, to 
compensate for the one that will be lost from the marriage payment. 
Formerly he paid no damages for giving a child to a non-virgin, but these 
were imposed by a Native Authority law in 1947. In both cases the child 
is always his, even if he does not deliver the damages for seduction. The 
Lozi say one beast is paid at marriage for the bride's untouched fertility 
if she is a virgin, and one is 'the beast of shame because husband and wife 
know each other', i.e. it is paid for sexual rights. In theory, the marriage- 
cattle do not give rights to the woman's children who, if adulterine, may 
belong to the adulterer and not to the husband. However, the husband 
has in practice a limited right to all her children. He is 'the owner of the 
hut' and it is assumed that all the woman's children are his. He may 
suspect a child is adulterine and overlook this. If he claims damages for 
adultery on the grounds of birth of a child he loses the child, unless he 
too was sleeping with his wife when she conceived. An adulterer can 
claim that a married woman's child is his if he alone had access to her. 
If she admits it he takes the child, though he will have to pay damages. 
Before men went to labour centres or jail for long periods it is unlikely 
that this question arose, but the Lozi reaction in this situation is the 


opposite of the Zulu's. One man noted for his many children told me he 
had gone into the huts of his infertile friends, at their requests, to give 
children to their wives. He could claim the children. Usually, how^ever, 
sterile men ask this favour of kinsmen to avoid the claims. 

Adulterine children are related to their mother's husband and other 
kin under the rule that the relative of one's relative is one's own relative. 

Barrenness in a Lozi wife is not a breach of an essential condition of 
the marriage contract. The husband cannot claim a sister in marriage or 
a refund of the marriage payment. He can divorce her, and if she was 
a virgin at marriage get back half the payment if she has not conceived, 
but divorcing a woman for sterility is very different from having the 
right to demand a female relative to produce for her. 

When a Lozi dies his kinsmen have no rights to keep his wives. These 
are free to go and marry where they please. They will be approved of if, 
having children, they marry his brother, or resident uterine nephew, 
so as to prevent the children being drawn from their paternal home, but 
it is a new marriage for which a new marriage payment must be given 
and its children are of the new husband, not of the dead man. Though 
the Lozi lack the levirate, they do have a limited sororate. Until 1946-7 
if a man married a woman and she died shortly after marriage, leaving 
no children, he could not claim her younger sister or the return of the 
marriage-cattle as a Zulu can. He had no case in law; the Lozi held that 
God took the woman and she and her parents were not at fault. The 
parents might of their own good heart, if they approved of him, give him 
another sister to replace the dead wife, or he might beg for one. Then 
he paid no cattle. If his wife died leaving a child he had no right even 
to beg. If he wished to marry her sister he had to pay marriage-cattle 
again. However in 1946-7 the High Court held that if either wife or 
husband died and the wife had not conceived, according to Lozi law 
(on which it was misinformed) one beast must be returned, and Lozi 
courts now follow this decision. 

We have seen that if a Zulu wife dies childless, and is not replaced 
by a sister, the marriage-cattle return, because she has not fulfilled an 
essential part of the contract. Some cattle.return if the wife has only one 
or two children, none if she has three. n 'Similarly, in the rare event of 
divorce, no cattle are returned if there are three children, but some are 
returned if there are only one or two children. Under the European- 
imposed Natal Native Code an erring Zulu husband can be deprived of 
the cattle as punishment. If a divorcee or widow marries elsewhere, the 
new husband may pay cattle to her house within her former husband's 
compound family. Among the Lozi no cattle used to return on the 
death of a childless wife. On divorce, the beast given for a non-virgin 
does not return; one of the two beasts given for a virgin returns if 
she has not conceived, i.e. if she has miscarried the husband loses his 
rights to it. 


*^The long enduring rights conferred on the Zulu husband extend 
forward to betrothaUIf a girl breaks an engagement she must return the 
gifts given to her, and her family must return most goods transferred 
during the negotiations. The court would, were it not for Government, 
compel marriage X If the boy is in the wrong, he forfeits the gifts and may 
have to pay compensation for wasting the girl's time. Among the Lozi, 
if an engagement is broken by the girl any gifts given by the boy to the 
girl do not .return. iXozi courts do not therefore recognize infant 
betrothals .JThe Zulu do. 

-^ To sum up, in both societies marriage gives the husband the right to 
claim damages from an adulterer, and to some extent to control his 
wife's movements. Among the Zulu he also gets the right to all the 
woman's children, and this right vests for ever in him and his agnatic 
group, even after his death. If she is childless she must be replaced or 
the cattle returned, and the amount of cattle returnable on divorce 
depends on the number of children the wife has had. Once a woman is 
betrothed her fiance's kin have a right to claim her in ghost marriage. 
A man redeems his illegitimate children with cattleV^hus Zulu marriage 
transfers a woman's fertility absolutely to her husband's agnatic kin- 
group, and an essential element in the contract is that she have children. 

Associated with this situation we have rare divorce. Among the Lozi 
a child should go to its genitor.^ Children belong automatically to their 
mother's and their genitor's families. Therefore the Lozi do not have 
Zulu forms of marriage in which women and dead men can marry wives. 
An illegitimate child goes to its genitor independently of payment. When 
the husband dies his widow is free to marry elsewhere, though she would 
be approved if she entered into a new marriage with one of his cognatic 
kinsmen. The children of this new union belong to the new husband. 
A man's kin have no claim on his fiancee if he dies. That the woman 
should produce children is not an essential part of the contract. While 
she lives, if she is childless her husband should not be given a sister, and 
he cannot claim refund of the marriage-cattle unless he divorces her and 
she was a virgin. If she dies childless shortly after marriage her parents 
can of their goodwill replace her; the husband has no claim in law. With 
this is associated a high rate of divorce. 

' My reading so far reveals a similar rule only among Hehe, Bena, Kikuyu, 
and Ganda: Brown and Hutt, op. cit., p. 102; G. Gordon Brown, 'Legitimacy 
and Paternity amongst the Hehe', American Journal of Sociology, xxxviii, 2, 
Sept. 1932, pp. 184-93 ; A. T. and G. M. Culwick, op. cit., p. 372 ; J. Kenyatta, 
Facing Mount Kenya, London, 1938, p. 184; J. Roscoe, The Baganda, London, 
191 1, p. 61; L. Mair, A71 African People in the Tzventieth Century, London, 
I943> [on Ganda] p. 57. Among the Tallensi a child belongs to the woman's 
husband, but he may run to his genitor's home where he will be received : M. 
Fortes, Marriage Law among the Tallensi, Accra: Gold Coast Government, 1937, 
pp. 12-13; so too with the Shona, see C. Bullock, The Mashona, Cape Town, 
1927, p. 328. 



^ The fact that divorce hangs over every marriage in Loziland, but not 
in Zululand, appears in the marriage ceremonial. The Lozi ceremony, 
if it is performed, is simple and attended by few people. The bride's 
senior relatives do not come to the bridegroom's village, and the bride- 
groom's senior male relatives do not attend the ceremony if there is one. 
,,\' 'It is a play of children.'^ Zulu bride should be escorted by many 
senior relatives, though her own parents do not attend. l/The bride- 
groom's parents participate in the complicated ceremonial which spreads 
over many days. jThe ceremonial chiefly expresses the hostility and 
conciliation of the two lineages concerned.^ 

I have analysed two contrasting types of marriage complex as they 
appear among Zulu and Lozi. Among other things, we see that the 
divorce rate is only one index of the general durability of marriage which, 
I suggest, is a function of the kinship structure as a whole. Recent 
researches have shown that a high rate of divorce is not always an 
indication of general social instability. In many tribes the family is 
unstable; in some (e.g. Nuer) as a domestic unit, despite strong marriage 
ties, while in others the marriage tie also is fragile. Stability inheres 
in the extended kinship groupings and relationships which form the 
structure of the society.^ 

I have surveyed the literature on many African tribes^ and affirm 
tentatively that divorce is rare and difficult in those organized on a system 
of marked father-right, and frequent and easy to obtain in other types. 
The correlation has been stated, without explication, by Loeb for 
Indonesia : 

'Nowhere does the form of social organization show influence more than 
in the laws of divorce. In a strict patrilineal society divorce can be obtained 
at the will of the husband, since the husband would not wish to lose the 
bride-price. An exception is made in the case of adultery. In the sibless or 
bilateral families of Indonesia (Mentawei excepted) divorce is frequent, 
and where there is a bride-price the rule almost everywhere holds that the 
bride-price must be paid back if the fault lies with the woman and not 
paid back if the fault lies with the man. The village or family head checks 
too frequent divorces. Among people with matrilineal sibs, as the Minang- 
kabau, there is no bride-price and divorce is very frequent and can be 
obtained at will by either party. '^ 

' A. W. Hoernle, 'The Importance of the Sib in the Marriage Ceremonies of 
the South-eastern Bantu', South African Journal of Science, xxii, 1925. 

^ See e.g. Evans-Pritchard, op. cit., pp. 64-5 ; D. Forde, Marriage and the 
Family among the Yako in South-Eastern Nigeria, London, 1941, pp. 1 13-14; 
M. Fortes, 'The Significance of Descent in Tale Social Structure', Africa, xiv. 3, 
July 1944, at pp. 362, 373-4, 382. 

^ As stated earlier, my comparative survey was limited by library facilities. 
Here I quote only on tribes whose marriage rules indicate the lines which guided 
my hypothesis, or whose marriage rules are against my hypothesis. The litera- 
ture is not clear on most points. 

■* E. M. Loeb, Sumatra: Its History and People, Vienna, 1935, pp. 68 ff. 


Loeb here relates the relative frequency of divorce to the amount of the xj/^ 
marriage payment, and this is still commonly done in books on Africa.' r.iC^- 

Evans-Pritchard strongly criticized this correlation : the assumption that • '}^ 

the maintenance of marriage relations is due to economic motives ^-^ . 

'is at the basis of the assertion often made that the function of bride- ' 
wealth is to stabilize marriage. The v^^ord "function" carries no meaning 
in this context. Is it true, moreover, that the relations between husband 
and wife persist through what amounts to economic blackmail? No 
evidence is adduced to justify belief in a functional relationship between 
the amount of bride-wealth paid to the bride's group and the durability 
of her union with her husband. No one would deny that the difficulty of 
returning a very large amount of wealth may be a motive in the pressure 
which the parents of a girl bring to bear on her to remain with her husband, 
but it is a hopeless distortion of social realities to regard this as an explana- 
tion of bride-wealth. . . . This point of view also neglects the fact that the 
relations between husband and wife^re not of primary importance, and 
that the stability of marriage rests on the goodwill of a man's father-in-law 
or^fothefs-Tn-Iaw ... it is evident that the stability of the family is not s^,^' L 
really a function of economic motives, but of moral and legal norms from . \/',tr'V ^ 
present-day conditions in Zandeland, for, in spite of payments, divorce 
is rife. It is morals that censure divorce and law that refuses to recognize 
grounds for divorce which ensure the stability of the union of husband 
and wife. It derives its stability from the restraint imposed by law and 
morals^ and not from economic blackmail. In the past Azande regarded 
marriage as an indissoluble union between man and wife, and as unsever- 
able relations between their affines, and divorce was allowed only for 
a flagrant breach of obligation on the part of the husband towards his 
wife's father or brothers.'^ 

As Evans-Pritchard says, individuals when they contemplate divorce 
may consider difficulties about marriage payment. I suggest that this is 
the least important side of a complex relation /The frequency of divorce 
is an aspect of the durability of marriage as such, which in turn is a 
function of the kinship structure^''! have no space to cite all my support- 

■ See e.g. A. I. Richards, Bemha Marriage and Present Economic Conditions, 
Rhodes-Livingstone Paper No. 4, 1940, pp. 53-4; A. T. and G. M. Culvvick, 
'The Function of Bride-wealth in Ubena of the Rivers', Africa, vii. 2, April 
1934, p. 151 ; G. G. Brown, 'Bride-wealth among the Hehe', Africa, v. 2, April 
1932, p. 146. A Nyasaland Government inter-departmental memorandum states 
the correlation of high divorce rate with patriliny and low divorce rate with 
matriliny, and refers it to the difference in quantity of marriage payment: see 
summary in Nada, xvi. 1939, pp. 43-4. 

^ E. E. Evans-Pritchard, 'Social Character of Bride-wealth, with special 
reference to the Azande', Man, xxxiv, 1934, No. 194, pp. 172 flf. M. Fortes 
refers to this statement and says his Tallensi data support it : 'Kinship, Incest and 
Exogamy of the Northern Territories of the Gold Coast', in Custom is King: 
Essays presented to R. R. Marett, London, 1936, p. 250. It seems from C. R. 
Lagae, Les Azande, 1926, p. 155, that divorce is also now common among the 
Belgian Congo Azande. Schwernfurth (The Heart of Africa, 1873, ii, p. 28) 
praised the stability of marriage among the Azande. 


ing data from other tribes, but it seems that rare divorce goes with 
levirate, sororate, rights to claim a betrothed girl, &c., and with them it is 
found in father-right societies. Marriage is for a woman's lifetime even 
if her husband dies. The marriage payment transfers the woman's pro- 
creative power absolutely to her husband's agnatic lineage for her life, 
and therefore divorce is rare. This emerges most clearly among the Nuer, 
where with constant changes of households the legal bonds in the family 
resulting from marriage are very stable. If the society has many goods 
marriage payment can be high. Among the Lozi and other tribes, who 
recognize inheritance in all lines, the marriage payment gives the 
husband rights over his wife, and indicates generally the children's 
paternal attachment while the marriage lasts, though in some tribes 
children should go to their genitors. Divorce is frequent, for the woman 
produces for her own line as well as her husband's. In matrilineal 
societies the woman produces for her own matrilineal line and the 
marriage payments do not transfer her fertility. Divorce is frequent. 
Therefore the marriage paym^t tends to be low; there is none among 
the cattle-owning Ovamboi/On this hypothesis, which needs more 
testing, the amount of goods transferred and the divorce rate tend to be 
directly associated, but both are rooted in the kinship structure. It is 
rare divorce which allows high marriage payment, rather than high 
marriage payment which prevents divorce. 

The amount of goods transferred at marriage will obviously be 
influenced by factors extraneous to the kinship-marriage complexes. 
Thus the types and quantities of property available, and the proportions 
of the sexes and their relative marriage-ages, may by the laws of supply 
and demand affect the marriage payment and the divorce rate.' I give 
a few examples to support my hypothesis that the influence of property 
is secondary. The patrilineal-patrilocal Azande with no cattle had 
initially low payment in spears and in the past rare divorce. I think the 
Ganda may also be taken as a propertyless people with father-right and 
rare divorce,^ but data on this type of tribe in Africa are too confused 
to be cited. On the other hand, there are several African peoples who 
had large herds of cattle but did not use them for marriage payments, 
and they are all peoples with mother-right or an unspecified or double 
descent system of succession. The Lozi fall into this category, as do the 
Ovambo and Herero.^ If property transfers determined the stability of 
marriage, these cattle-owning tribes could have used their herds to effect 

' See e.g. Oberg, op. cit., p. 136. 

^ Roscoe, op. cit., p. 97. 

^ C. H. L. Hahn, 'The Ovambo', in Tribes of South-west Africa, Cape Town, 
1928, pp. 32-5; M. Rautenan, 'Die Ondonga', in Rechtsverhdltnisse von einge- 
borenen Vdlkern (ed. S. R. Steinmetz), Berlin, 1903, pp. 330 flF. On Herero see 
G. Viehe, 'Die Ovaherero', in ibid, at pp. 305-6; B. von Zastrow, 'Die Herero', 
in E. Schultz-Ewerth and L. A. Adam, Das Eingeborenenrecht, Stuttgart, ii. 
p. 254; E. Dannert (jun.), Zww Rechte der Herero, Berlin, 1906, p. 46. 


this, since they all approve of stable marriages. Indeed, the mother-right 
Ila, who are patrilocal, have high marriage payments from large herds, 
but among them divorce is frequent.^ They may have high payments 
because they are said to have few young women. I lack space to cite 
further examples, but those given clearly refute the explanation put 
forward by Loeb and others. Therefore I suggest that the divorce rate 
is a reflex of the kinship structure itself. The morals and laws which, as 
Evans-Pritchard says, sanction the stability of marriage, and the relations 
between affines which affect it, are aspects of that structure, even where, 
as among the Zulu, /chastity and fidelity are also enforced by the ethical 
code of the regiments in a military nation. 

However, it is important to stress that even where marriages are un- 
stable, they always exist. Theoretically there is no reason why, given the 
ban on incest, the men of a matrilineal group should not keep their 
sisters and daughters and allow strangers to impregnate them in casual 
encounters, as was in effect believed to be the position among the Nayar. 
Always, however, even among the Nayars as we now know, the associa- 
tion for begetting children is institutionalized and important. Marriage 
cannot be referred altogether to the kinship structure. 


V Lozi and Zulu families are economic units. 'The husband must give 
his wife land for gardens, feed her from his own gardens and herds (and 
fishing-sites among the Lozi), and provide her with a hut and clothing. 
■^/The wife must work her gardens and do household duties. 

In Lozi law husband and wife have equal rights in the crops she has 
worked — the land is his but the labour is hers. They and their children 
feed from these crops. On divorce they share the crops equally. If he 
dies and she leaves his home she takes half the crops. If she dies her 
heirs, who are her sons and daughters, her parents, and her siblings, can 
claim her share against her husband, his sons by other wives, and his 
brothers and parents. If the husband lived in one of her homes in 
affineship (bukwetunga) he has no rights in the crops : neither the land 
nor the labour were his. 

Similarly, father and mother each have a right to half the marriage 
payments for their daughters. 'They are both parents, they both gave 
birth to, the child', say the Lozi. If £2 or two beasts are given for a virgin 
daughter, each should get one. When ^^i is given for a girl who is not 
a virgin they share the money, and if a beast is given and it is killed the 
two families share the meat. However, usually the father keeps it and 
gives the first and all odd calves to the mother and her kindred. Each 
kindred should divide the cattle, or money, or meat, which it receives 
among its own members. The bride herself has rights in both portions, 

' E. W. Smith and A. Dale, The Ila-speaking Peoples of Northern Rhodesia, 
London, 1920, i, p. 298; ii, pp. 50 ff. 



which is inconceivable to a Zulu. Lozi say 'the bride brought the cattle'. 
If either kindred ever kills a descendant of a marriage-beast the bride 
who brought the beast should be given the tongue, and specific portions 
should be sent to the other kindred.' Frequently a father, or maternal 
uncle if the bride lives with him, does not share the marriage payment 
with the other kindred as he should, or share his portion with his own 
relatives. People can sue the other kindred for their rights in the 
marriage payment, but rarely do so 'for it will spoil their relationships'. 
They content themselves with the reflection that if there is divorce of 
a virgin who has not conceived or if the woman is fined for adultery, 
which are both very likely to occur, they will be under no obligation to 
contribute to the return of a beast or to her fine. 'We knew nothing of the 
marriage.' If they have shared in the payment they should contribute. 
Whichever kindred the man pays the woman is legally his wife. 

The mother's own cognates are heirs to her rights as against the 
father's relatives, i.e. failing her own sons, her parents, siblings, and their 
children inherit. The father, his sons by other wives, and his brothers 
have no rights in the maternal portion of a dead woman's daughter's 
marriage payment. If both parents are dead their sons inherit both sets 
of rights, but they should divide the payment in half and distribute one 
half among their paternal kindred and one half among their maternal 

If the chief or any other person gives a man the marriage payment for 
his wife he gets no rights in the payments for her daughters. 
'^^ Since divorce is rare the Zulu have no fixed rules about the division 
of the crops, but on my information I should say they would belong 
altogether to the husband. Since widows should remain in leviratic 
marriage there is no law about the division of crops on the death of one 
or other spouse, and^d incline to think that if the widow left her husband's 
kin they would not allow her to take any of her crops. I did not inquire 
carefully. V Certainly, the mother and her kin have no rights in her 
daughter's marriage-cattle, if marriage payment has been given for her. 
v^All go to the bride's father who may divide them among his agnates. 
A separate beast, which is not reckoned part of the marriage payment 
{ilobolo), is given by the groom to her mother for his bride's virginity 
and for her mother's care. The mother usually eats it with other women. 
The same beast is paid by a seducer of a virgin to his love's mother. The 
bride's maternal uncle can claim a beast if he reared the girl. 
\y The extent to which the marriage-cattle are owned in the agnatic 
lineage appears in a very complicated set of rules about their distribution, 
y/ In a simple legal family, if a man gave the cattle himself, he or his sons 
are entitled to all the cattle of his daughters. VBut if he borrowed the 
cattle and has not repaid the loan, the lender can claim the cattle of his 

' I give examples of actual divisions of the marriage payment in my Essays on 
Lozi Land and Royal Property, pp. 65-7. 


eldest daughter. If the lender is the father's chief, the father would not 
offer to return them, because it is an honour for his eldest daughter to 
'belong' to the chief. If the husband did not deliver the marriage -cattle 
to his affines they claim the eldest daughter's cattle; and a poor man can 
marry a wife by pledging his eldest daughter thus. If he has no daughter, 
there is no liability j^While the father is alive the cattle of other daughters 
are his and he can deal with them as he pleases .'\4^o we ver, if he uses the 
cattle to marry other wives, these become subordinates of 'the house' 
(i.e. the wife) from which the cattle came. The junior wife and her 
children are under the authority of the senior wife and her children. The 
eldest son in that house can claim the cattle of the subordinate wives' 
eldest daughters. The chief heir of the great house can claim the cattle 
of the eldest daughter of each constituent house. He can also claim the 
cattle of the eldest daughter of the left- and right-hand houses, to 
establish his seniority, even though the left- and right-hand houses are 
established with independent herds. -Their heirs have rights within their 
own houses similar to those of the main heir in his house : they take the 
cattle for eldest daughters of houses established from their herds, 
i/^ubject to these rights of senior sons, cattle of a woman go first to her 
father and own brothers, failing these to the main heir of her house, and, 
failing one in her house, to the main heir of the compound family. 
Beyond that inheritance is in the agnatic lineage. The mother's kin have 
no claim once marriage-cattle have been given for her, save for the beast 
of virginity. Their rights to the marriage-cattle of the eldest daughter if 
her mother has not had cattle delivered for her do not affect this principle. 
'^/Marriage-cattle among the Zulu are thus definitely property of 'the 
house of the woman', subject to certain contingent claims, but only 
within the agnatic compound family. \/vVhat we may call 'the house- 
property complex' has wider implications iA Zulu allots land and cattle' 
to each of his wives .\/Land and cattle thus allotted become irrevocably 
property of the house, even if the woman is divorced or demoted in 
status. This property can be used by the father, subject to the rules set 
out above, but it is inherited by the sons of that wife against their half- 
brothers by their father's other wives. As widows usually remain attached 
to their dead husband's agnatic lineage, even if they are 'entered' 
(ukungena) by a sister's son of the dead man, they retain the land allotted 
to them for gardens ; when a woman or widow dies her land goes to her 
own children, as do the cattle. If the husband while alive uses these 
cattle or this land for another wife, she becomes subordinate to. the first 4, ^^'' 
wife. My notes are not clear about a woman's own property .i/Zulu law 
says a wornan cannot own anything, and if she earns property as a doctor 
it is^in the care of her male guardian 1/lt is property of her house and on 
her death should go to her sons, and failing them to sons of a sister 
married in sororal polygyny. I am uncertain if after this it goes to her 
agnatic kin or her house in her husband's family. Where a woman has 



herself married a wife with cattle, she is treated as a man and her 
property passes to 'her' sons by her wife or wives on regular rules of 
agnatic succession. 

Any land or cattle which the father did not allocate go to the main heir, 
who must use them for the benefit of all his father's sons, just as the heir 
in each house must help his siblings. If an agnatic line dies out for lack 
of sons, an heir is raised by marrying a wife with the cattle left or they 
pass in the agnatic lineage. The chief enters into heirless estates before 
maternal kinsmen. 

That is, while a woman gets no rights of ownership in the cattle and 
fields allotted her by her husband, save during her marriage, she has 
rights of holding which vest through her in her sons. But they do not 
through her get any rights in the property of her family. /Thus though 
Zulu descent and succession to property and status are strongly patri- 
lineal, they are reckoned through an unrelated woman: 'the social 
principle of agnatic descent ... is traced through the mother' (Evans- 
Pritchard).'i'he sons' rights and positions in their father's home and in 
their agnatic lineage are determined by the positions of their mothers. 
Some of the main sources of litigation among the Zulu are disputes 
between half-brothers about their rights arising from the respective 
status of their mothers, and what cattle and fields and other goods were 
allotted to their mothers while the father was alive. (The positions of 
wives' huts in the village, their status in the tribe, the order of their 
marriage, their wedding ceremonial, the source of their marriage-cattle, 
are all considered in evidence. 

Among the Lozi the inheritance of property and status goes by very 
different rules. Here the mother does not fix agnatic descent; the child 
goes to its genitor. Correspondingly, her sons inherit no property rights 
from their father through her. The Lozi man, like a Zulu, allocates some 
of his cattle and gardens to each wife. He also gives cattle, gardens, and 
fishing-sites to his sons and daughters as they grow up, and they are 
entitled to these against his selected general heir. They transmit this 
property to their own heirs, agnatic or uterine. They are not entitled to 
claim the fields and cattle allotted to their mothers, against the general 
heir, who takes them all, and the marriage -cattle of his half-sisters, 
though he is under an obligation to give some to his siblings and half- 
siblings and cousins. Here a woman does not transmit property rights 
to her children in her husband's home, but she transmits to them pro- 
perty rights in her own homes, paternal and maternal. Not only are they 
heirs to property she may herself have acquired, as among the Zulu, but 
they may also succeed to her brother, their maternal uncle, or to her 
father, their maternal grandfather, or even to their matrilineal great- 
grandfather, their mother's mother's brother. The contrast emerges 
clearly in the Lozi institution of 'the beast of the fire'. If a woman has 
lived for some years with a man, cooking and gardening well, preferably 


bearing children, her husband will give her a 'beast of the fire' in thanks. 
This is hers, and is inherited by her children ; failing them it goes to her 
parents, her brothers or sisters, or even more distant relatives in her 
paternal or maternal family, as against her husband's sons by other wives 
or his brothers. The husband will warn her to send 'the beast of the fire' 
to her own home lest his general heir claim it, as he is entitled to claim 
cattle given for her nurture and use, against her sons. 

Further, among the Zulu, when the husband dies the widow is 
expected to remain in his home to bear more children in his name, and 
therefore is entitled to retain and work the gardens he gave her; but 
among the Lozi, as soon as the husband dies the widow loses rights to 
land at his home. She takes half the planted and reaped crops, but even 
if she has children in the village and elects to remain with them, the 
general heir is not bound to give her land to cultivate for herself. She 
must get land in the portion allotted to her children under their right to 
land at their father's home. Thus among the Lozi the place of a married 
woman is not fixed in her husband's village by her marriage; she does 
not pass with her fertility to bear children for them only. Her position 
and her rights to property remain in her own paternal and maternal 
homes, and therefore she transmits rights of succession and inheritance 
to her children there, while they get no rights through her in her 
husband's home. I must stress that I refer here under 'house-property 
complex' to the inheritance of rights through the mother. In practically 
every Bantu society each wife's house has its own property while the 
husband lives, as among the Lozi. The wives do not pass rights in this 
property to their sons. I suggest that this is why wives do not have fixed 
status in ranked Lozi society as they do among the Zulu. Even the Lozi 
Paramount's widows become nonentities save through their children. 
/The Zulu king's widows have status in their own right. 

These differences in Zulu and Lozi laws of property are consistent 
with the internal structure of their marriage complexes. However, we 
cannot assert that in African societies organized in corporate agnatic 
lineages a woman transmits to her sons specific rights in their patrimony, 
and no rights in her own family property. On present information, it can 
only be claimed that it is not recorded from a single tribe without the 
agnatic lineage, that a woman passes rights in the patrimony to her sons 
against their father's sons by other wives. Clearly in purely matrilineal 
tribes the rule would not be found, but neither does it exist in any account 
of tribes organized 'bilaterally'.' On the other hand, what I have called 
the 'house-property complex' in patrilineal societies seems to have 
a twofold geographical distribution. First, it is found in South Africa 
among all Nguni, Sotho, Tswana, and Venda peoples, roughly as far 
north as the Limpopo. Offshoots of these, like Ngoni, Ndebele, and 

' See e.g. Forde on the Yako, op. cit., pp. 45, 56, 70 ; H. Vedder, 'The Herero', 
in The Native Tribes of South-West Africa, p. 195. 


Shangana, have carried the complex with them. Generally the complex 
stops north of the Limpopo. Bullock says specifically that, in contrast 
with the Southern Bantu, the Shona (who have an agnatic lineage system) 
do not pass property through the wife. Jaques's description shows that 
Junod's Thonga do not have it. Among Shona and Thonga a general 
heir is appointed who divides the patrimony^ as in the patrilineal and 
bilaterally organized tribes in Central and Eastern Africa.^ This seems 
from Roscoe's account to be the position among the interlacustrine 
Bantu. ^ In north-eastern Africa, however, the complex is found again 
among the Chaga,"* probably the Kikuyu and Kamba,^ and to some 
extent the Bantu Kavirondo.^ Driberg reports it among all Nilo-Hamites 
and Nilotes, and individual records confirm this.' In West Africa the 
complex does not occur among the Tallensi, Ibo, or Dahomey, all of 
whom have agnatic lineages.^ 

The absence of the house-property complex appears to be associated 
among the Shona and Thonga, and the West African peoples cited, with 
the rule that inheritance of a deceased man's inherited estate passes 
through his brothers before it drops a generation to their sons and 
nephews, all of whom hold it before it drops again to the grandson 

' C. Bullock, The Mashona, p. 370, and A. A. Jaques, 'Terms of Kinship and 
Corresponding Patterns of Behaviour among the Thonga', Bantu Studies, i, 
1927-9, pp. 327-48. H. A. Junod, The Life of a South African Tribe, London, 
1927, is not clear, but as he does not describe 'the house-property complex' 
it is unlikely that it is present. However, A. Clerc implies that the kindred 
Ronga devise property through their wives ('The Marriage Laws of the Ronga 
Tribe', Bantu Studies, xii, 1938, at pp. 103-4). 

^ See e.g. Brown and Hutt, op. cit., pp. 82 ff. ; A. T. and G. M. Culwick, op. 
cit., pp. 275 ff. ; G. Wilson, 'An Introduction to Nyaklyusa Society', Bantu 
Studies, X. 3, Sept. 1936. 

^ J. Roscoe, The Baganda, The Northern Bantu, 1915 ; The Banyankole, 1923 ; 
The Bakitara, 1932; The Bagesii, 1924. 

* C. Dundas, Kilimandjaro and Its Peoples, London, 1924, pp. 304 ff. 

^ C. Dundas, 'The Organization and Laws of Some Bantu Tribes in East 
Africa', J. Roy. Anthrop. Inst., xlv, 1915, at p. 294; Kenyatta, op. cit., pp. 32 
and 177; and W. S. and K. Routledge, With a Prehistoric People, the Akikuyu, 
London, 1910, p. 47 and pp. 142 ff. But G. Lindblom, The Akamha, Upsala, 
1920, is not clear: see pp. 162 ff. 

* G. Wagner, The Changing Family among the Bantu Kavirondo, Supplement 
to Africa, xii. 2, 1939, pp. 19 flF: marriage payments go to the bride's full brothers, 
other property is given out as sons grow up. 

'' J. H. Driberg makes a general statement for the Nilotes and Nilo-Hamites: 
'The Status of Women among the Nilotics and Nilo-Hamitics', Africa, v. 4, 
Oct. 1932, at p. 419. See also J. H. Driberg, The Lango, 1923, pp. 174 fT. ; 
J. G. Peristiany, The Social Institutions of the Kipsigis, London, 1939, pp. 211-12; 
M. Merker, Die Masai, Berlin, 1910, p. 203 ; Evans-Pritchard on the Nuer, op. 
cit., p. 49; C. G. and B. Z. Seligman on the Dinka, Pagan Tribes of the Nilotic 
Sudan, London, 1932, pp. 172 ff". 

* M. Fortes, Marriage Lazv among the Tallensi, pp. 9 ff. ; C. K. Meek, Law 
and Authority in a Nigerian Tribe, London, 1937, pp. 319 ff. ; M. Herskovits, 
Dahomey, New York, 1938, pp. 87 flf. 


generation. Clearly in such a system the estate could not be divided into ] 
sections passing directly to groups of sons demarcated by their mothers. ; 
Nevertheless in these systems each man creates also his own estate. 
His inherited estate passes to his brothers before vesting in the genera- 
tion of sons; his earned estate vests immediately in his sons and passes _ 
from one to the other of them before vesting in his grandsons. Among ) 
the Southern Bantu the house-property complex is part of the graded 
organization of the component families within the compound poly- 
gynous family. In north-eastern Africa, too, it seems that inheritances j 
pass directly to sons and not through a generation of brothers first. 

I have described these two contrasting sets of inheritance rules 
because, as stated above, they fit logically into the various kinship and 
marriage systems respectively. I have also indicated the geographical 
distribution of the types of rules, and the occurrence of other associated 
rules. Though the comparative data do not yet allow us to generalize 
firmly about them, I draw attention to them because the significance of 
the difl^erence has been missed in regional surveys in the past,^ and 
because information on this point is not always clear in the literature. 


The evidence from the Lozi and Zulu indicates that we may have to 
distinguish different types of marriage payments. There are several 
aspects of marriage. It breaks or modifies certain existing social relation- 
ships and creates new social relationships : the union of a man and his 
wife, and an alliance between the kinsfolk of the two spouses. It unites 
men and women to produce children to occupy specific positions in the 
kinship system, since normally it is through the marriage of their 
parents that children acquire their kinsfolk on both sides. But clearly 
the structure of the kinship system primarily determines the conse- 

' e.g. in I. Schapera's (editor) The Bantu-speaking Tribes of South Africa, 
London, 1 937, especially essay on 'Work and Wealth' by L Schapera and A. J. H. 
Goodwin, at pp. 162 ff. The Seligmans in discussing what I call the house- 
property complex among the Dinka call it 'not an inheritance with female 
descent. ... It would appear to be a compromise between patrilineal and 
matrilineal descent, and to be associated with the Nilotic relationship system in 
which brothers and sisters, children of the father and mother, are distinguished 
from one another' (loc. cit., p. 174). E. Torday ('The Principles of Bantu 
Marriage', Africa, ii. 3, 1929, at p. 263) sees the legal position correctly: 'For 
inheritance of property, every house has its own rights as far as these fields and 
stock are concerned, and, even where the patriarchal system prevails, inheritance 
is by the mother.' But he wrongly applies this rule to all Bantu, though he cites 
only Dundas, Kilimandjaro, pp. 304-9, on the Chaga, and Routledge, op. cit., 
p. 47, on the Kikuyu, in support. He also refers to Van Wing, Etudes Bakongo, 
pp. 223 ff. (misprinted pp. 233 ff.), and Torday and Joyce, Les Bushongo, p. 272, 
for the allocation of property to wives. Van Wing and Torday and Joyce do not 
state that these allocations are inherited by wives' sons. 


quences of marriage in the affiliation of children and therefore the 
attributes of the marriage payment. 

Marriage in most societies transfers to the husband a certain common 
minimum of rights : almost always an exclusive right to the wife's sexual 
services, or to the lending of them, certain powers over her person, 
rights to the produce of her economic activity balanced by economic 
obligations to her, and a prima facie right to be pater to her children. 
The extent and duration of these rights varies greatly. We have seen that 
Lozi marriage payment is low, considering their wealth in cattle, and 
that even then it dates from a recent enactment. It does not transfer the 
woman's fertility altogether to her husband, let alone to her husband's 
kindredj/Comparative data show that in general marriage payments fall 
with the decreasing dominance of patrilineal descent until very little is 
r given in purely matrilineal societies. In Central Africa a son-in-law may 
"give some years of service, but it is in matrilocal marriage. The matri- 
lineal Ovambo are rich cattle-owners, but a husband gives for his wife 
only a present to her mother, and it is killed for the wedding feast. That 
is, when children are not dominantly joined to the husband's line he 
tends to give but little for his wife, whose economic and sexual services 
^ alone are transferred to him. In almost every African tribe he does give 
something to get these rights and prima facie rights in relation to her 
children. On the other hand, in patrilineal societies of the Zulu type 
the marriage payment permanently transfers the woman's procreative 
capacity to her husband's lineage. Therefore, relative to the society's 
wealth, the payment tends to be large. 

I suggest that it may no longer be wise to name the common institu- 
tion of transferring goods by a single term (marriage payment, bride- 
wealth, bride-price). This leads to disputes about whether the attributes 
ascribed to it, e.g. by Evans-Pritchard in East Africa, apply to the 
transfer of goods among the Yako or Lozi or some other tribe. The Lozi 
institution, on the surface a similar transfer of property, is not the Nuer 
or the Zulu institution when we come to examine their structural rela- 
tions. For the common element, the rights transferred in all societies, 
. we may follow Radcliffe-Brown's use of the term marriage payment, 
! but it may be necessary to distinguish at least marriage payment Type A, 
Type B, &c. I have analysed two types, and there are likely to be more. 
Certainly my crude categories will not cover every variation, and further 
analysis requires to be done in many tribes. 

I consider that the data we have indicate that the basis of the varia- 
tion in the complexes is probably the different affiliation of the children. 
'i/Zulu society is chiefly distinguished from Lozi by its structure of 
unilineal agnatic groups, which are exogamous, and which are associated 
with a marriage rule by which the giving of cattle to a woman's father 
transfers her fertility for all her lifetime to the agnatic group of her 
husband, who may indeed have been dead before she was married to his 


name, or who may be a woman or an impotent manHf the wife goes off in 
adultery, the children are her husband's ; if he dies, she continues to bear 
for him ; if she is barren, or dies before bearing, a younger sister should 
replace her. /The outstanding fact is the extreme endurance of the 
husband's rights and their passing on his death to his agnatic heirs. 
Legal marriage and the domestic unit it establishes are thus very stable, 
though there may be frequent adultery. The legal emphasis is the same 
among the Nuer, even though they have frequent changes in the consti- 
tution of households. Marriage is stable in that wherever the woman is 
her husband accompanies her, even after his death, to be pater of her 
children, for whoever their genitors are, they belong to the man or 
group which gave cattle for her. The husband or his heirs may let his 
wife go and not reclaim the marriage payment, since they get the children 
even though they lose her other services as wife. In these tribes marriage 
payment thus binds the woman's reproductive capacity to the perpetua- 
tion bf the extended agnatic lineage. Each such group loses its daughters, 
but insists on its rights to the fertility of its daughters-in-law. 

Moreover, rights of inheritance vest in children exclusively from the 
agnatic lineage of their pater. They have no rights in their mother's 
family unless they are unredeemed illegitimate children, when they rank 
as if they were members of her agnatic lineage. Therefore economic and 
other interests pull the children almost entirely to the home of their 
pater and his agnatic lineage. 

The Lozi have no extended unilineal kin-groups. Among them the 
corporate groups which endure in time are the cognates in family 
villages or in sections of a royal village, based on the mounds which dot 
the plain and which are centres of surrounding gardens, fishing-sites, 
and pastures. Lozi's relatives in the patrilineal and matrilineal lines may 
be scattered at many places far over the surface of the great flood-plain. 
Because their productive activities are varied, and many fall in distant 
places in the same month, they require co-operation and help from many 
people. Kinship provides the framework for getting this help, but since 
neighbours in the various parts where they have economic resources are 
linked to them by patrilineal, matrilineal, and mixed ties, they emphasize 
relationship in all of their lines of descent. For them it is not important 
to fix an individual's relationship to a single line, but to emphasize his 
links with many lines ; therefore marriage does not tie a woman's pro- 
creative power to one line. She produces for many lines .<> Among the 
Zulu economic and other interests coincide with the pull of agnatic 
lineage ties; among the Lozi, with their limited resources centred in 
restricted dwelling-mounds, economic interests may pull a man to settle 
with his mother's kindred, in either of her lines of descent. As a woman's 
productive capacity is not tied to a single lineage group, the child goes 
to its genitor, not to the legal husband of the woman. This removes the 
main buttress of marriage ; since the children can shift their allegiance 


and emphasize that relationship which most pleases or profits them, the 
family as a whole is an unstable association. I am not here referring only 
to the instability of households : this is also marked among the Nuer, but 
among them lineage ties always draw the family together about the 
woman's husband. Among the Lozi, too, a large proportion of children 
grow up elsewhere as foster-children and not in the villages where they 
will ultimately settle. There is no dominant pressure of interests and 
law to induce them to return to their father's home, though most do. 
They may go to the home of any of their near ancestors. Adulterers can 
claim their children, and adultery is incessant. Men divorce their wives 
! easily and at personal will; women are always straining to be released. 
Marriage, as well as domestic association, is unstable. I have suggested 
that the reason for this is to be sought in Lozi kinship structure, which is of 
a pattern common in Central Africa, but which here is directly related to 
their; modes of production and settlement in their physical environment. 

We see then that in these societies the types of marriage and the forms 
of the family, with inheritance of property, rights of children in father's 
and mother's lines and the rights of these lines to claim them, laws of 
betrothal, destination of widows, rates of divorce, status of wives in their 
husband's homes, are all consistent with certain types of kinship system. 
In some societies the household group which is usually designated as the 
family may be unstable. The structural stability of the society rests in 
the extended kinship lines, one of which may be the nuclear framework 
on which local communities are organized as corporate groups. These 
reach back into the history of the tribe through all changes of personnel, 
while the domestic family itself is always an ephemeral, and often an un- 
stable, association.' 

I have presented my argument more strongly than I myself feel is 
justified at present, in order to make clear the type of data and analysis 
I consider likely to be most fruitful. I should have preferred to do further 
comparative research before publishing it, had it not been for the 
opportunity of presenting it in this symposium. I am fully aware of the 
difficulties of establishing the validity of the hypothesis, but even if it is 
wrong it may be useful. Some of the difficulties are inherent in socio- 
logical analysis, since in this there are always complicating variables. 
Others arise from the vague and embracing use of categories and con- 
cepts (of which I too am guilty) such as patrilineal, lineage, marriage, 
divorce, &c. When is a marriage complete, and when can we class the 
separation of a cohabiting couple as divorce ?^ Are the Bena patrilineal 

' See references p. i66, n. i, especially Evans-Pritchard, Nuer Marriage, 
pp. 64-5. 

^ For example, on the difficulty of establishing when there is a marriage and 
when divorce, see Richards, Bemba Marriage and Modern Economic Conditions. 
It is difficult to tell what Fortes means by divorce (see his Marriage Law among 
the Tallensi, 1946). He speaks of a high divorce rate but divorce from a 'proper' 
marriage is obviously a most serious ritual and social step (see his The Dynamics 


and the Ila matrilineal, and have they Hneages ? The Hterature is generally 
confused and imprecise, and the posing of problems may help to clarify 

/I am myself uncertain whether it is the stability of people's attachment 
to specific areas, or patriliny or father-right itself, or the agnatic lineage, 
or all of these together, which, whatever the other variables are, tend I 
to be associated with a strong marriage tie. That any or all of these are j 
significant seems definite to me. Evans-Pritchard has stated that the J^ 
Azande, organized on father-right but without the agnatic lineage, had 
rare divorce in the past. He has not reported on the problem fully, but 
it might be that the State power prevented women from leaving their 
husbands if they wanted to, i.e. there was instituted authority to compel 
them to remain in marriage. When that authority was restricted by 
British occupation divorce became rife. On the other hand, despite 
European support for women against men in Zululand, something there, 
I suggest the agnatic lineage, has maintained the stability of marriage ; 
while in Loziland the marriage tie has become looser and looser, even 
though tribal ethics approve of firm unions, while not condemning 
divorce. However, as we have seen, the code of the Zulu age-set regi- 
ments, backed by the State, punished severely unchastity and infidelity: 
perhaps these codes are still the main sanctions on marriage despite all 
structural changes. Yet the ban on prenuptial conception does not seem 
to have retained similar force. Illegitimate births are frequent while 
divorces are rare. According to old records and modern field-workers 
this seems to be the position among all Southern Bantu tribes who had 
tribal initiation into age-regiments,' save for the Thonga of Portuguese 
East Africa. In his book Junod implies that divorce is frequent, though 
this seems to contradict an earlier article on another group of Thonga, 
and Clerc states explicitly of the related Ronga that divorce is impossible 
once the wife has borne a child : 'she is forever bound to her husband's 

I give these references in detail for, while all the Southern Bantu have 

of Clanship among the Tallensi, 1945). In Dahomey (Herskovits, op. cit.) the 
divorce rate varies according to which of the thirteen types of marriage is 
contracted, but it seems to vary according to my postulate. 

' H. Kuper, paper on the Swazi in this book, and An African Aristocracy, 
1947, p. 104; Read, 'Moral Code of the Ngoni', pp. 16-18; E. CasaHs, Les 
Bassoutos [written 1859], Paris, 1930, pp. 231-2; D. F. Ellenberger and 
A. Macgregor, History of the Basuto, 1912, p. 273; E. H. Ashton, unpubhshed 
thesis on the Southern Sotho, University of Cape Town, quoted with his per- 
mission; C. L. Harries, The Laws and Customs of the BaPedi and Cognate 
Tribes, 1929, p. 36; H. Sta^'t, The Bavenda, 1931, p. 152; I. Schapera, Married 
Life in an African Tribe [Tswana], 1940, p. 294. 

^ H. A. Junod, op. cit. i, pp. 198-9. Cf. his 'The Bathonga of the Transvaal', 
Addresses and Papers, British and South African Associations for the Adz-ancement 
of Science, 1905, iii, pp. 258-9, together with Harries, loc. cit., previous footnote; 
A. Clerc, 'The Marriage Laws of the Ronga Tribe', p. 89. 


agnatic lineages, it is of tribes without age-sets that a few records state 
there is frequent divorce. It seems clear that among the Natal Nguni 
and the Shona divorce was rare/ Hunter on the Pondo says 'dissolution 
was usual [though disapproved] .... Very many of the married women 
whom I knew had been married more than once, and the number of 
[unmarried divorcees] is considerable.' Cook implies that Bomvana 
divorces are frequent, but while divorced women are free to remarry, 
he had 'been unable to find a woman married again. There seems to be 
a strong aversion to this. Further, a woman who has returned from her 
husband's [village] is in the nature of a byword for female frailty.' Soga 
states that among the Xhosa^ 'divorce is not so common as might be 
supposed', and a deserting husband can return after many years to claim 
his wife and children." Most of the old WTiters on these Cape Nguni 
peoples who dealt with the subject stated explicitly that divorce was not 
frequent, was not easy to obtain, and required the sanction of the chief; 
but a few imply that a man merely sent his wife home and women could 
run away, without stating how frequently this happened.^ Generally 
writings on this part of Africa did not state the frequency of divorce, and 
were concerned to examine, as were the 1883 Cape Commissioners on 
Native Law and Custom, whether it was possible for a woman to escape 
from a marriage which they regarded as a sale into slavery. However, 
since a competent authority states that the Pondo have frequent divorce 
and I have been unable to find a specific reference to them in the early 
literature, it is perhaps worth recording that Sir Theophilus Shepstone 
told the 1883 Commission that the Pondo, unlike the Zulu, had become 
'far more immoral' than they were fifty years before. He ascribed this to 

' C. T. Nauhaus, 'Familienlieben, Heirathsgebrauche und Erbrecht der 
KafFern', Zeitschrift fiir Ethnologie, xiv, 1 882, at p. 2 1 o ; L. Marx, 'Die Amahlubi', 
in S. R. Steinmetz (ed.), Rechtsverhdltnisse von Eingeborenen Vdlkern, p. 353; 
and information from Dr. S. Kark who kept extensive records for a southern 
Natal grouping over seven years recently and noted one divorce. On the Shona, 
Bullock, op. cit., p. 367. 

^ M. Hunter [Prof. M. Wilson], Reaction to Conquest, 1936, p. 212; P. A. W. 
Cook, Social Organization and Ceremonial Institutions of the Bomvana, 1932, 
pp. 156-7; J. H. Soga, The Ama-Xosa, 1932, pp. 283 ff. 

^ Clear statements on the rarity of divorce are: H. Lichtenstein, Travels in 
Southern Africa in the Years 1803, 1804, 1805, and 1806, 1930, i, p. 322 and p. 
326, at pp. 322-3 quoting L. Alberti, De Kaffers aan die Zuidkust van Afrika, 
Amsterdam, 1810 — on the Xhosa; Warner, Dugmore, and Brownlee in J. 
Maclean (ed.), A Compendium of Kafir Laws and Customs, 1858, pp. 53-4, 70, 
and 163 — on Xhosa, Tembu, Fingoes; J. Macdonald, 'Manners, Customs, 
Superstitions and Religions of the South African Tribes', J. Roy. Anthrop. Inst., 
xix, 1889, pp. 271-2 — chiefly on Xhosa; A. Kropf, Das Volk der Xosa-Kaffern, 
1889, pp. 154-5. Note that none of these refer to the Pondo. The strongest 
indication that divorce was frequent is in the report of the Cape Government 
Commission on Native Law and Custo?n, 1883 (Parliamentary Paper G. 4-1883), 
Cape Town: W. A. Richards, p. 37, para. 95. D. Kidd did not specify the tribes 
to which he applied his contradictory statements: The Essential Kafir, 1904, 
2nd edit. 1925, at pp. 225 and 358. 


the removal of the threat of attack by the Zulu which led to a decline in 
military virtues.' 

I have referred to the Southern Bantu material in some detail partly 
to indicate the difficulty of assessing the data on these problems. It must 
be emphasized that records on the Southern Bantu, especially on marriage, 
are better than those on most regions of Africa. But I feel I must also, 
even in this compressed essay, cite the Pondo and Thonga as peoples 
who are said to have unstable marriages, since they not only iiave the 
agnatic lineage but are close to being neighbours of the Zulu.-l indicate 
that it is possibly the moral code of the age-sets which helps to stabilize 
Zulu marriage, to show that I do not discount the influence of ethical 
rules themselves. It is clear that special variables have to be examined in 
each tribe. ^ 

I have considered briefly some of the possible influences that the types 
and quantities of property available to particular peoples may have on 
the marriage payment and hence on marriage. We have seen that the 
presence of considerable movable property in the form of cattle among 
Lozi, Ovambo, and Herero is not itself sufficient to produce high 
marriage payments; and where it has done so among the Ila, marriage 
is not thereby stabilized. I suggested that the Azande and Ganda are 
father-right peoples who lacked property which could produce high 
marriage payments, but nevertheless had stable marriage. However, 
I have not had space to examine whether a kindred group becomes stable 
about a fixed piece of land, or herds of cattle, and whether this and the 
rights of the children to the property, are not what stabilize marriage, 
rather than agnatic descent and father-right themselves. It appears to 
me that property in this form is by itself insufficient, since we have 
many records of peoples with matrilineal succession to valuable lands 
or herds among whom marriage is very unstable.-' 

In this essay I have restricted my survey in general to Black Africa. 
I believe that for this region my argument is likely to be validated as 
a demonstrable tendency. So far I have been able to make only an even 
more cursory survey of the literature on other primitive groups. I have 

* Report of the Commission cited in previous footnote, p. 31, Q. 564. 

* e.g. why do the Xhosa not practise sororal polygyny, the levirate, divorce 
for barrenness, though their marriage rules are so similar to those of the other 
Nguni? {Opera cita, Lichtenstein, i, pp. 319-25; Ayliff in Maclean, p. 163; 
Macdonald (on Gaika and Tembu), p. 272 ; Soga, p. 1 39 ; 1 883 Cape Commission, 
pp. 1 14-15). 

^ See e.g. M. Fortes, 'The Ashanti Social Survey', Rhodes-Livingstone jfournal, 
vi, 1948, p. 19 on the attachment of the Ashanti matrilineage to village and lands, 
and J. B. Danquah, Akan Lazvs atid Customs, 1928, at p. 156: 'it cannot be 
exaggerated how easily and rapidly marriages may be dissolved with little 
trouble', e.g. also K. M. Panikker, 'Some Aspects of Nayar Life', Jf. Roy. 
Anthrop. Inst., xlviii, 1918, at pp. 266 ff. See also various works on the Pueblo 
Indians, as F. Eggan on 'The Hopi Lineage' in Studies in Social Structure: 
Essays presented to A. R. Radcliffe-Broivn, 1949. 


cited Loeb's statement of the one part of the correlation, the divorce 
rate with the kinship structure, for Indonesia,^ and the thesis seems to 
be borne out generally in North America and Oceania. The largest 
group of peoples which can by most criteria be classed as at all primitive, 
for whom the thesis appears to be incorrect, is the Bedouin. Under 
Islamic law divorce is simply effected and most records on the Bedouin 
state that it is frequent. There are contrary statements, and it is impos- 
sible for me to give a decision on this set of peoples who, having agnatic 
lineages, should on my theory have stable marriage. I record that the 
present bias of the data is against my argument.^ It is, of course, doubtful 
whether an Islamized people can be considered primitive. 

I do not even venture to express an opinion on the stability of marriage 
in more developed communities, in towns and complex rural economies, 
for here many complicated variables enter. I am aware that my analysis 
ultimately involves the question of what is the relation of the family to 
the total social structure, of how far the stability of the family is related 
to the stability of that structure, of what effects moral rules have as 
compared with structural determinants, &c. In defence of my attempt to 
penetrate these intricate problems I plead that we must isolate a few of 
them at a time. Among Lozi and Zulu the kinship system, and the 
marriage payment and its consequences, form complexes which are 
internally consistent and which differ in practically every relation. 
Therefore they offer ideal material for initial comparison. 

' I read this after I formulated my argument. 

^ H. Granqvist (Marriage Conditions in a Palestinian Village, Helsingfors: 
Societas Scientarium Fennica, 1931, pp. 268 ff.) made a quantitative analysis 
of divorces and concluded : 'as far as I can see there is little which goes to prove 
that divorce is so frequent among Muhammedans'. Otherwise all statements on 
the urban and peasant fellahin emphasize a very high divorce rate, though 
Blackman's description implies that divorce is a serious business (The Fellahin 
of Upper Egypt, p. 95). The typical statement on the desert Bedouin is that of 
G. W. Murray (Sons of Ishmael. A Study of the Egyptian Bedouin, London, 1935, 
at pp. 225-6) that 'divorce is very common among the Bedouins, for a variety 
of reasons'. A. Kennett (Bedouin Justice : Laws and Customs among the Egyptian 
Bedouin, 1925, p. 100) writes that 'divorce, although a very much simpler 
business than in civilized countries, is by no means as easy as is popularly 
believed ; and Bedouin public opinion is usually sane and healthy'. I dare affirm 
that modern studies will show Bedouin marriage to be stable. 




MOST of the Bantu peoples of Central Africa reckon descent 
in the matrilineal rather than the patrilineal line, and many 
of them practise some form of what is usually known as 
matrilocal marriage. In fact, it is the matrilineal character 
of their kinship organization which distinguishes them so clearly from 
the Bantu of East and South Africa, and for this reason the territory 
stretching from the west and central districts of the Belgian Congo to 
the north-eastern plateau of Northern Rhodesia and the highlands of 
Nyasaland is sometimes referred to as the 'matrilineal belt'.-' 

Within this group of 'matrilineal' tribes there is a remarkable degree 
of uniformity as to the principles governing descent and succession and 
the various ideologies by which people explain their adherence to the 
mother's rather than to the father's line, and stress their community 
of interests with their maternal relatives.^ Blood is believed to be passed 
through the woman and not through the man. The metaphors of kinship 
stress the ties between people 'born from the same womb' or 'suckled 
at the same breast', and in some tribes the physical role of the father 
is believed to be limited to the quickening of the foetus already formed 
in the uterus. The duty of the woman to produce children for her lineage 
is emphasized, and descent is traced from an original ancestress or a 
series of ancestresses known as 'mothers' of the lineage or clan, and also 
in some cases from the brothers of these founding ancestresses. The 
ancestral cult centres round the worship of matrilineal rather than 
patrilineal ancestors, although spirits of the father's line are sometimes 
the subject of subsidiary rites. 

A child belongs to his mother's clan or lineage, and succession to 
ofHce follows the common matrilineal rule, that is to say, authority 
passes to the dead man's brothers or to his sisters' sons, or to the sons 

' There are, of course, a number of patrilineal or mainly patrilineal tribes in 
the Belgian Congo, including most of the Luba (as distinguished from the Luba- 
Hemba), the Songe, and the Nkundo. 

^ Matrilineal is used here in inverted commas because it is generally recognized 
that no society is entirely matrilineal or patrilineal as regards descent, inheritance, 
succession, and authority, but that the family system provides a balance of 
interests and rights between the two sides of the family with a predominant 
emphasis on one side or the other, and it is in this sense that I shall use the term 
in this article. See my 'Mother Right in Central Africa' in Essays presented to 
C .G. Seligman, 1930, for a description of Bemba kinship from this point of view. 


of his maternal nieces. Among some of the Central Bantu women 
succeed to the titles of royal ancestresses, or hold positions as chieftain- 
esses, with special ritual functions. 

But with these principles of descent and succession the similarity 
ends. The Central African people differ in a rather striking way as to 
their family structure, and in particular as to the various forms of 
domestic and local grouping based on the family. 

These variations in the family structure depend largely on the nature 
of the marriage contract and the extent to which the husband is able to 
gain control over his wife, who belongs, by virtue of matrilineal descent, 
to the lineage and clan of her mother and of her brothers and sisters; 
and also on the extent to which he manages to achieve a position of 
authority over the children she bears. In matrilineal societies the man's 
control over his wife and her children can never be complete, except 
in the case of a union with a slave woman, but he can gain considerable 
power over his wife's labour, her property, and her child-bearing 
powers, as well as rights over his children's work and their marriages, 
by virtue of the service or payments he makes to his father- or brother- 
in-law. Moreover, the ways in which domestic authority is divided 
between a man and the head of his wife's kinship group are surprisingly 
varied. In some cases there is a formal allocation of rights and privileges 
between father and mother's brother in return for service and payments. 
In other cases the balance is less well defined, and every marriage 
produces what can only be described as a constant pull-father-puU- 
mother's-brother, in which the personality, wealth, and social status 
of the two individuals or their respective kinsmen give the advantage 
to one side or the other, and a number of alternative solutions are 
reached within the same tribe. 

In this balance of privileges and duties between the patrikin and the 
matrikin the crucial point is obviously the husband's right to determine 
the residence of the bride. If she and her children live in the same home- 
stead or village as his kinsmen, his domestic authority is likely to be 
greater than where they remain with the wife's relatives. 

Throughout this area, at any rate, the rule of residence at marriage 
seems to me to be the most important index of the husband's status. 
It also provides the most convenient basis for the classification of these 
different forms of matrilineal family. In Central Africa we find every 
gradation from the marriage in which the husband has the right to 
remove his bride immediately to his own village, to varieties of trial 
marriage, temporary unions, or customs by which the husband takes 
up more or less permanent residence in the wife's group. 

For this reason I have found the old terms 'matrilocal' and 'patrilocal' 
of little use for comparative purposes. Many writers have pointed out 
that the words are in themselves confusing, since there are two parties 
to a marriage and what is 'residence with the mother' for the one is 


'residence with the mother-in-law' for the other. If the terms are to be 
retained for purposes of classification it would be necessary to adopt 
a convention by which, for instance, they were always used with 
relation to one sex. Firth and Adam suggest as an alternative the use 
of the words 'viri-local' and 'uxori-local' to meet this difficulty,' and 
I shall use these terms from time to time. 

Another difficulty with regard to the Central African area is the variety 
of forms of marriage relationship which could reasonably be included 
under the title 'matrilocal'. In societies practising matrilineal descent 
a man may live with his wife's people because the marriage is on trial ; 
because he is fulfilling a marriage contract by service instead of the 
payment of goods; or because he means to settle permanently in his 
wife's group. ^ He may have sex access to his wife at night and work in 
her fields, but act as head of his sister's house and spend a large part of 
his day with the latter. This is the practice of the matrilineal Menang- 
kabau of the Padang highlands of Sumatra and the kindred peoples of 
Negri Sembilan.^ A similar position holds good among the Hopi of 
Arizona. Alternatively a husband may live at his wife's village but often be 
away visiting his own relatives, as amongst the Yao (see p. 233); or he 
may spend alternate years in his own and his wife's village, as in Dobu 
Island. In any of these cases there will be some years, some months, or 
some hours in which the marriage may reasonably be called matrilocal. 
The term gives no indication of the length of time a man spends in his 
wife's village, or the degree to which he is incorporated with her 
matrikin, or isolated from his own family."* For this reason I have 
found it better to use the phrases 'marriage with immediate right of 
bride removal' or 'marriage with delayed right of bride removal' 
or 'marriage without bride removal' in describing the family systems 
of the Central Bantu. 

The terms 'matrilocal' or 'patrilocal' are similarly lacking in precision 

' See R. Firth, We, the Tikopia, 1936, p. 596, and L. Adam, in Man, Jan. 
1948, p. 12. N. W. Thomas, who suggested the terms patrilocal and matrilocal 
in 1906, put them forward 'not as being specifically appropriate but as being 
parallel to patrilineal and matrilineal, denoting descent in the female or male 
line respectively' {Kinship and Marriage in Australia, 1906, p. 108), thus show- 
ing that he realized from the start the ambiguity of the terms. 

^ Rivers points out, in his article on 'Kinship', in Hastings's Encyclopaedia, 
that marriage by service 'passes insensibly into the matrilocal form of marriage' 
and trial marriages 'shade insensibly into trials before marriage on the one hand 
and into ease and frequency of divorce on the other'. 

^ Cf. Verkerk-Pistorius, Studien over de inlandische Houshoiidijig in de Padang- 
ische Bovenlander, 1871. Quoted by Taylor in 'Matriarchal Family System', 
The Nineteenth Century, xl, 1896, p. 96, where he describes the 'Chassez Croisez' 
which takes place at dusk when each man leaves his sister's house where he has 
been by day, to join his wife and children at night; cf. also F. C. Cole, The 
Peoples of Malaysia, 1945. 

* R. Linton stresses the importance of the propinquity of the two settlements 
in determining the character of the descent system in Study of Man, 1936, p. 168. 



when used as a means of classifying types of family and domestic unit. 
According to the terminology adopted in this book, a parental family is 
a household or homestead composed of a man, his wife or wives, and 
their children. Such a unit naturally grows in size in societies where it 
is customary for either the sons or the daughters to live with their 
parents after marriage. This wider group, usually based on the principle 
of unilineal descent and composed of members of three generations, 
is here described as an extended family.' Members of extended famiUes 
of this kind usually live in a separate kraal or homestead or they may 
form' the nucleus or core of a village or a section of a village. They tend 
to co-operate in economic affairs, to exercise some common rights over 
property, to accept their genealogical senior as a common authority, 
and to practise some joint ritual. Where it is a common practice for a 
man who has married sons to separate off from his father's homestead 
to start a new community, and where the land situation makes this 
possible, a three-generation extended family of this kind becomes the 
normal pattern of residence as it is, for instance, amongst the Zulu. 
Where, however, the villages are more permanent and the splitting 
off of new extended families is not so easy, the residential pattern 
is naturally much less uniform. The eldest brother of a group of 
siblings may succeed to the position of his father in a patrilineal society 
or to that of his mother's brother in a matrilineal society, and various 
other changes of this kind may take place. Thus in the more or less 
permanent villages there may be two or three extended families or 
remains of extended families. There may also be two or three alternative 
types of marriage in the same community, and a number of principles 
of residence. Such larger residential units, usually based on a nuclear 
extended family but with a number of additions, I have called local 
kin groups, specifying whether they are predominantly matrilineal or 
patrilineal in composition. 

In such extended families or local kin groups extension takes place 
on the basis of certain nuclear or pivotal relationships. For instance, 
according to the rules of residence at marriage, the children of one sex 
marry out of the homestead, whereas those of the other remain within 
it. Rules of succession and inheritance similarly determine the incidence 
of authority and economic privilege within the local community. 
Property rights and authority may go from father to son, from mother to 
daughter, from brother to sister's son, and so forth. I find it convenient 
to classify the different types of extended family by means of these 
nuclear relationships, even though the categories constantly overlap. 
Thus I use the term 'father-son extended family' for a Zulu homestead ; 

' I have found it useful to employ the term 'grand-family' suggested to me 
by R. Firth to indicate a three-generation family' descended in the direct line, 
patrilineal or matrilineal, when it is necessary to distinguish between this and 
other forms of extended family. See pp. 218, 227, 228. 


or the 'mother-son extended family' for a local unit which is a common 
pattern amongst the Swazi, where the woman leaves her husband's 
kraal at the time of her son's marriage and lives with him in a position 
of authority to the end of her life. Parent-daughter extended families 
of one kind or another exist where uxorilocal marriage is practised. 
These include the mother-daughter family of the Hopi, where property 
in land or houses is passed through the woman; the sororal family, 
where a group of married sisters and their daughters live together, 
usually under the care of an elder brother (see p. 232); or the sibling 
family found amongst the Nayar of Malabar, where a man lived with 
his sisters and the latter were visited by their husbands at night; or a 
father-daughter family, where married daughters live with their parents, 
but their father is very much the head of the group and may determine 
residence instead of the mother (see p. 227). Other types of extended 
family include the matrilineal fraternal family, where a group of brothers 
live with their wives and their sisters' sons. 

It is obvious that in the case of residential units such as these the 
nuclear relationships vary from one type of extended family to the 
other ; that is to say, one homestead is based on the close relationship 
of a group of the men of the family, as in the case of the patrilineal 
father-son family or the matrilineal fraternal family; while another is 
founded on the kinship of women, as in the mother-daughter family of 
the Hopi. The interests of brother and brother, brother and sister, or 
mother and daughter may be identified by one marriage system or the 
other, while the extended families so formed may attract to themselves 
additional households of kinsfolk or slaves, according to fixed rule or 
more casual association. 

I hope to distinguish four different types of family structure amongst 
the Central Bantu by means of the criteria indicated, that is to say, the 
type of marriage contract, the distribution of domestic authority, the 
residential units, and the primary kinship alinements. 

The literature on this area is as yet scanty. When further material 
is available I expect to find a range of variation between these types 
and even the appearance of new forms. It is my thesis that the balance 
of interest between the two sides of the family is bound to be an uneasy 
one in the case of matrilineal communities, and for this reason variation 
between neighbouring societies within the region is to be expected 
as well as different types of marriage and family within each tribe itself. 


{Type A) The Mayombe-Kongo Group 

I have selected the Mayombe and the Kongo as typical of a group of 
western Congo peoples which practise matrilineal descent, succession, 
and inheritance, and give high marriage payments with the right of 


immediate removal of the bride, but with the return of her children to 
her brothers' village at puberty. To the same group of people belong 
the tribes of the Kwilu basin such as the Huana, Yaka, Yanzi, and 
Sakata, as well as the Songo of the Luniungu and Gobari areas and the 
Ambundu to the extreme south-west of the district.' 

The Mayombe inhabit the lower Congo area between Shiloango and 
the Luakula basin, south-west of Brazzaville. The Mayombe were 
described by Van Overberg in 1907, but since that date our information 
has been supplemented by a series of articles by N. de Cleene, as well as 
an important study by F. P. Van Reeth.^ 

The Kongo inhabit the region of the cataracts south of the Congo 
between Matadi and Stanleypool. Van Wing's study of the Mangu 
tribe of the Kongo group was written in 1921 . Further material appeared 
in Torday's article in 1933.^ I shall base my description on the Mayombe 
material in the first instance."* 

Economic Determinants 

The peoples of the lower Congo area appear to have more economic 
resources at their disposal than the other selected tribes. Their material 
culture was more highly developed than that of most other Bantu 
peoples, and ironwork, copper, wood- and ivory-carving as well as 
weaving, reached a high level. This was particularly the case among the 
Kongo. Copper and iron objects were used for marriage and other 
payments. Trade in ivory and slaves seems to have been established for 
some hundreds of years. This area has also been longer in contact with 
Europe than have Northern Rhodesia and Nyasaland, since the Portu- 
guese occupation of the western coast district dates from the fifteenth 
century. Under present-day conditions these people are described as 
producing almost all the palm oil and coconut oil of the Congo and they 

' E. Torday and T. Joyce, 'Notes on the Ethnography of the Bahuana', J. Roy. 
Anthrop. Inst, xxxvi, 1906; P. J. Denis, 'L' Organisation d'un peuple primitif', 
Congo, i. 4, 1925, pp. 481-532; De Beaucorps, 'Le Manage chez les Basongo 
de la Luniungu et de la Gobari', Bull. Jur. Indig., xi. 6., 1943, pp. 109-26; 
G. Week, 'La Peuplade des Ambundu', Congo, ii, 192, 1937. 

^ N. de Cleene, 'La Famille au Mayombe', Africa, x. i, Jan. 1937, pp. 1-15; 
'L'Element religieux dans I'organisation sociale des Bayombe', Congo, i. 5, 1936, 
pp. 706-1 1 ; 'Individu et collectivite dans revolution economique du Mayombe', 
Inst. Roy. Colon. Beige Bull. xvi. 2, 1945, pp. 254-60; Van Overbergh, Les 
Mayombe, 1907 ; F. P. Van Reeth, De rol van dem rnoederlijken oom in die inlandsche 
faniilie, Inst. Roy. Colon. Beige Memoires, 1935 ; 'De bruidsprijs bij de huwelijken 
in Congo', Semaine de Missiologie, Louvain, 1934, pp. 192-215 ; M. Nauwelaert, 
'Note sur la Societe Yombe', Congo, i. 4, 1938, pp. 504-9; L. Bittremieux, 
Mayombsche idioticon, 1923; H. Deleval, Les Tribus Kavate du Mayombe, 1913. 

^ Van Wing, Etudes Bakongo, 1921 ; E. Torday, 'Dualism in Western Bantu 
Religion', _7- Roy. Anthrop. Inst. Iviii, 1933, pp. 225-45; Causeries Congolaises, 

"* N. de Cleene, 'Les Chefs indigenes au Mayombe', Africa, viii. i, Jan. 1935, 
PP- 63-75. 


seem to have considerable wealth in money. Agricultural land is fertile, 
although there appears to be considerable pressure of population in some 
districts, and rights over forests, garden and house sites, and palm-trees 
will be shown to be important determinants of kinship ties. Small stock 
such as goats and sheep are kept in some districts and are used for 
marriage payments, but the ritual attitude to stock which is characteristic 
of the Eastern and Southern Bantu does not exist. Differences in 
economic and social status were evidently marked at the time of the 
arrival of the Belgians. Slavery was such an important institution in 
pre-European days that the people are described as falling into three 
groups, chiefs, freemen, and slaves,' and this high incidence of slavery 
affected the balance of the kinship system, since rules of descent differed 
for slaves and for freemen. Chiefs and commoners were differentiated 
according to their wealth in material possessions. 

Ideology and Principles of Descent and Succession 

The Mayombe believe thdt blood passes through the woman and 
not through the man. The child belongs to a series of kinship groups 
based on matrilineal descent: the clan (mvila), which is an exogamous 
division tracing descent to a legendary founding ancestress and associated 
with a territory; a major matrilineage (dikanda), a subdivision of the 
clan, which is also exogamous and which reckons direct descent to a 
depth of about six generations from a known founding ancestress, and 
which is associated with a district; a minor matrilineage (mvtimti, pi. 
bivuniu), which is a local group forming a village or section of a village 
and is composed of a group of brothers and their sisters' sons, with 
sisters who marry elsewhere but contribute money obtained by their 
work to the mvumu funds. ^ 

The descent system of the Kongo is similar. Van Wing speaks of 
matriclans (dikanda), tracing descent from a legendary ancestress, and 
matrilineages (luvila), which he describes as 'lignees'. The matrilineage 
traces descent to an actual ancestress, acknowledges a common head, 
and shares rights over a strip of territory, but the institution of chieftain- 
ship is more pronounced in this tribe and De Jonghe speaks of 'crowned 

' Van Overbergh, op. cit., pp. 383, 405, 409. 

- I use the term 'clan' for the largest group which traces unilineal descent, 
usually putative, from a common ancestress, and 'lineage' for a named group 
tracing actual descent from a known ancestress. I use this term for a variety 
of kinship groups based on the tracing of direct descent in one line, and dis- 
tinguish different types by qualifying terms such as 'patrilineage' or 'matriline- 
age'. I find Evans-Pritchard's use of 'maximal', 'major', 'minor', or 'minimal' 
lineages a ven.^ convenient one where two or three levels of subdivisions of the 
clan exist and they remain structurally connected. I am using the term 'lineage' 
to cover a number of diflferent unilineal descent groups such as the Ramage 
described by Firth in Tikopia, the sub-clan as used by Malinowski in the 
Trobriands, and the 'section' employed by Nadel in accounts of the Nuba. 



chiefs', 'clan chiefs', 'lineage chiefs', and village headmen.' The clans 
are also arranged in something like an order of hierarchy according to 
traditions of first arrival in the area. The Kongo also recognize as a 
special category the father's people or kitata. There is no legal identifica- 
tion of a child with his father's matrilineage, but the ancestors of this 
line are remembered for four generations and are honoured 'by respect'. 

The descendants of the same father are described as having a fraternity 
from 'the grave of the father',^ that is to say they were brought up in 
their father's village and return to bury him there. They are also 
described as being 'brothers of the same village'. A similar recognition 
of paternal ancestors is clearly marked at the time of the marriage of 
a girl, since she then calls on the spirits of her father's matriclan, her 
paternal grandfather's matriclan, and then her maternal grandfather's 
matriclan.^ It is for this reason that Torday speaks of dual descent 
amongst the Kongo, but the term kitata seems merely to mean the 
father's people. It is the name of a category of relations who are honoured 
and remembered with affection. The Kongo peoples similarly use a 
collective term for their matrikin {kingadi), for the brothers and sisters 
of a man's wife (kinzadi), or for the father and mother of a son's wife or 
a daughter's husband.'* 

The rule of matrilineal descent was formerly associated with the 
developed cult of the maternal ancestors, De Cleene points out that 
the chief of the Mayombe clan formerly had religious functions and his 
village was the centre of an ancestral cult;^ the same kind of cult, 
associated with clan and lineage ancestors, was definitely marked 
amongst the Khimba.^ In the case of the Kongo, ceremonies performed 
for the illness or death of a member took place within the clan and 
occasionally the lineage, but this type of cult seems to have largely 
lapsed amongst the Mayombe and the Kongo at the present day. 
Mayombe chiefs are mentioned as possessing fetishes, and the heads of 
local lineage groups are said to have magic powers or kindoki, but it is 
expressly stated by both De Cleene and Van Reeth that they have no 
significant part to play in an ancestral cult. 

The principles of descent and affiliation are clear. All freeborn 
Mayombe belong to matrilineal descent groups, and it is only the 
children of slave wives who belong to their father's matrilineage, or 
descendants of immigrants who have attached themselves to a chief 
and intermarried with women of his matrilineage. Children may also 

' De Jonghe, Preface to Van Wing, op. cit., p. vii. 

^ Van Wing, op. cit., p. 134; also E. Torday, J. Roy. Anthrop. Inst. Iviii, 1933, 
p. 238. 

' Van Wing, op. cit., p. 201. 

* De Jonghe, Preface to Van Wing, op. cit., p. x. 

' De Cleene, 'Le Clan matrilineal dans la societe indigene', pp. 12 and 13; 
Van Wing, op. cit., pp. 122, 271, 272. 

* Bittremieux, La Societe secrete des Bakhimba, 1936. 


be obliged to belong to their father's clan if their matrikin have been 
unable to return the marriage payment when the father has asked for 
a divorce. The children in this case will be kept as hostages until the 
repayment has been made, but they will always be reckoned as socially 
inferior members of their father's clan.' 

Succession to office amongst the Mayombe follows the matrilineal 
rule, that is to say, it goes from brother to brother to uterine nephew and 
then to uterine grandson. The offices which are hereditary are those of 
chief, head of a village or village section, or senior brother (khazi) of a 
matrilineal fraternal family. De Cleene states that the chiefs, whose 
powers seem to have been rather exiguous, were always succeeded by 
their brothers, but with regard to the post of head of the kinship village 
or village section, there seems to have been considerable room for 
exercise of choice within the 'filiation uterine'.^ This post is so important 
that young men are trained for it. Van Reeth describes the careful 
selection of two or three uterine nephews by the head of the village, 
and the education of these boys for their future work.^ The succession 
is usually matrilineal, but a complete stranger may be appointed to 
the village headship for want of a sufficiently intelligent hereditary 

Rules of inheritance and ownership of property are exceedingly 
important in the maintenance of kinship ties. Property consists of 
houses and garden sites, palm-trees, household goods, and wealth in 
money. Each man has individual rights to cultivate land allocated to him 
by the head of the minor matrilineage (mvumu), but the property of this 
group of brothers and sisters is held jointly and administered by the 
senior brother. Women who leave their village to marry keep their 
possessions distinct, and money made from the sale of produce or in 
other ways is handed over to their senior brothers, and hence is finally 
inherited by their children. It is not clear from the evidence whether 
any forms of property could be passed from father to son or father to 
daughter, but this seems unlikely. 

The Marriage Contract 

Marriage amongst the Mayombe is described as a system by which 
a man acquires sex access to a woman, and certain clearly defined rights 
to her services and those of her adolescent children, in return for a 
substantial payment in money or goods. The Mayombe husband never 
acquires full authority over his wife or children, as we shall see, but the 
marriage payment enables him to remove his bride to his own village 
immediately on marriage; his sex rights over her are exclusive, and 

' De Cleene, Africa, x. i, Jan. 1937, p. 6; Deleval, op. cit., p. 29. 
^ De Cleene, Inst. Roy. Colon. Beige Bull, ix, 1938, p. 68 ; Van Reeth, op. cit., 
pp. ii, 20. 

^ Ibid., p. II. 


payments as damages for adultery are heavy. In the old days, in fact, 
adultery was a crime punishable by death. 

The marriage payments can be described as substantial in relation 
to the tokens and gifts passed at marriage in the other selected type 
societies. Van Reeth speaks of 50 lengths of cloth valued at 12 francs 
each in the old days ; 400 to 500 francs in money after the 1914-18 war and 
a rate in 1936 of 3,000 francs.' Weeks gives figures for the lower Congo 
area in general as follows: 55 beads (very scarce and dear) in 1885; 3° 
pieces of cloth to the value of ^9 in 1883 ; and beads or money to the 
value of ;^20 in 1914.^ Van Wing gives figures of 40 to 60 francs to 100 
francs amongst the Mpungu in 19 10 to 1915.^ Interesting material is 
given by Mertens for the Banbata group in 1948, in which he compares 
the marriage payments with the average wage for an unskilled worker, 
e.g. 1,500 to 2,000 francs for the total marriage payments for a man 
earning 50 francs a month."* 

The Mayombe woman, although she is potentially the founder of a 
lineage, is definitely in an inferior position when she marries into her 
husband's village. She is there a stranger. De Cleene describes her as a 
servant in her husband's village ('Elle n'est d'ailleurs qu'en service chez 
son mari').^ Whether she is ever finally incorporated in any sense of the 
word in her husband's group is not clear. When she first marries it is 
evident that she is definitely an outsider in her husband's nivumu. On 
the other hand, the services she must do for her husband are limited 
by custom. If she is asked to do anything beyond these prescribed tasks 
for her husband or his brothers, she must be paid something additional 
for her work ; in fact a Belgian ethnographer uses the word 'wages' when 
describing such payments. Moreover the wife, as has been pointed out, 
keeps her own property distinct. She has her own gardens from which 
she is only obliged to give half the produce to her husband, while she 
controls the rest herself, and she usually gives it to the members of her 
own mviimu. Both in terms of economic interests and principles of 
affiliation she is evidently very closely identified throughout her life 
with her matrikin. 

The marriage payment here performs different functions from those 
ascribed to the loBola of the patrilineal Bantu of East and South Africa, 
but it is an equally significant element of the contract. Van Reeth shows 
that where a girl marries without the permission of her own mvumu her 
husband will give an unusually high marriage payment to her khazi, and 
will thus gain complete possession over her and her children. She then 
abandons all property rights in her matrilineal inheritance.^ The bride- 

' Van Reeth, op. cit., p. 196. 

^ Weeks, op. cit., p. 142. ^ Van Wing, op. cit., pp. 205-9. 

■* V. Mertens, 'Le Mariage chez les Bambata (Bakongo) et ses lefons sociales', 
Zaire, ii. 10, dec. 1948, pp. 1099-126. 

* De Cleene, Africa, x. i, Jan. 1937, p. 4. ^ Van Reeth, op. cit., p. 27. 


w ealth is provided by the men of the bridegroom's matrilineage and 
this seems to give them certain residual rights over the bride. A 
husband's brothers may produce children by the wife if the husband is 
impotent, and may inherit her, if she becomes a widow, with only a 
slight addition to the original payment made. 

The father has certain rights to his children's services during their 
early youth. A young man pays his earnings to his father until his 
marriage when he returns to his own mviimii, but it is probable that these 
rights are considered as a return for the amount a man pays for the 
upkeep of his wife's children and the fact that he is responsible for 
paying their debts and providing magicians to cure them if they are ill. 
In the case of divorce the marriage payment is returned, either wholly 
or in part, and in that case the children are immediately removed by 
their mother's brother. In fact it is said that the khazi comes to fetch 
his sister's children as a sign that he wants a divorce to take place.' 

I Authority ■ 

The children remain under the authority of the father until an age 
described variously as puberty, or even earlier, or at marriage.^ While they 
are small the father has considerable powers over the children, and if the 
khazi wishes to take two or three of his sister's sons to train as heirs at 
an early age, he must do so with the father's consent.-' But De Cleene 
states that the father can get obedience by appealing to his children's 
goodwill or using 'artifices'. If he is dictatorial, the mother reminds him 
that the children do not belong to him, and that they will leave him at 
once for their maternal uncle if they are badly treated. ''^ Deleval com- 
ments on the affection felt by children for their father, but says the 
mother is usually the disciplinarian. ^ Van Reeth says that if a child 
runs away to his khazi the father will have to make a formal request to 
get him back again. In any case, whatever the temporary powers of the 
father, the mother's brother always has overriding rights. An intractable 
boy is sent by the father to his mother's people to be corrected. The 
khazi appears if one of his nephews or nieces is ill, especially if sorcery 
is suspected. When they are adolescent and have come to live in his 
village he has absolute rights over their services and he can sell them 
into slavery and determine their marriages. He provides for the marriage 
payments of the boys, with the aid of the men of his mvumu. He receives 
those given for the girls. Neither boys nor girls can leave their mvumu 

' Van Reeth, op. cit., p. 20. 

^ Van Wing gives a figure of 8-10 years for the Kongo, op. cit., p. 261 ; Van 
Reeth, op. cit., p. 20. 

^ Girls are also apparently selected for training in this way and are sent to a 
mother's sister. It is not clear what they are trained for and why they should be 
sent to a woman not living in the mvumu. Van Reeth, op. cit., p. 11. 

* De Cleene, Africa, x. i, Jan. 1937, p. 8. 

* Deleval, op. cit., p. 31. 


unless they have been enslaved or have refused to marry their selected 
spouse (see p. 216). Even when a man has chosen to live with his father 
during the whole of the latter's lifetime, which occasionally happens in 
the case of a father of high birth or great wealth, he returns to his 
mother's people once his male parent is dead. In these unusual cir- 
cumstances he is referred to as 'a stranger chief in his father's village 
{ftimu angatii) and is given particular respect.' 

De Cleene quotes Mertens's account of an even more extreme form 
of the avunculate amongst the Badzing.^ In this tribe a crowned chief 
could beat his nephews and nieces or sell his nieces for the benefit of 
the matrilineage or in order to liquidate his debts or accumulate wives 
for himself. In this area a father can buy back his son or daughter, but 
he can never secure for them clan membership, and when he dies they 
are scorned as slaves. 

Residential Units 

The basic domestic unit among the Mayombe is a parental polygynous 
family composed of a man and his wives and their young children. Such 
a homestead also included the huts of slaves in pre-European days. 
This polygynous family differs from those of the patrilineal Bantu of 
South and East Africa, however, since it can never become a grand- 
family. The sons and daughters of the head of the homestead always 
marry and settle elsewhere, and a Mayombe does not, therefore, 
normally live with any of his grandchildren, whether the children of his 
sons, as amongst the Ila, or those of his daughters, as amongst the Bemba 
(see p. 227).^ 

He has, however, young boys and married men dependent on him. 
These are his sisters' sons, who are described as his 'subjects'. Un- 
married nephews apparently form part of the domestic unit and his 
wives cook for them, but it is not clear whether married nephews form 
part of the homestead in the sense of living in a cluster of huts marked 
off from the rest of the village or fenced around in any way. Nor is it 
clear whether all the sons of several sisters attach themselves to one 
eldest brother. 

What is certain, however, is that the sons of one mother, together with 
their wives and young children, their sisters' grown sons, unmarried and 
married, and their sisters' unmarried but adolescent daughters form a 

' De Cleene, Africa, x. i, Jan. 1937, p. 8; also Deleval, op. cit., p. 32. 

^ J.Mertens, Les Baldzing de la Kamtsha, Inst. Roy. Colon. Beige Memoires, 
1935. See also De Cleene, op. cit., 1946, p, 36. 

^ I say 'normally' because it is quite clear that some sons remain with their 
fathers in some areas at any rate, both from Van Reeth's account of the selection 
of special maternal nephews as heirs, and De Cleene's statement that the sons 
of rich men may be tempted to stay with their fathers. This is obviously a point 
on which more detailed analysis of the composition of particular villages would 
prove useful. 


compound family which acknowledges the authority of the senior 
brother (the ngadi khazi) and owns land in common from which 
individual gardens and trees are allocated. This matrilineal fraternal 
family is in the first instance the nucleus of the village, but an extended 
family naturally grows with the passage of years. The polygynous 
households of the brothers are increased by those of their married 
nephews until there are two, if not three, groups of married brothers 
residing in one village. New communities occasionally split off as in the 
case of other Bantu societies, but apparently this does not occur often, 
and some bivumu reach the size of 300 to 400 inhabitants. 

It is possible to distinguish amongst the Mayombe a corporate 
matrilineage in the sense of a group of men and women who trace direct 
descent from a common ancestress, own land and property in common, 
which the men cultivate together and to which the women contribute 
money, and who all accept the authority of a senior brother. The men of 
this matrilineage and their wives and young children constitute the kin- 
ship settlement which is an extended fraternal family or several such 
extended families. Both the descent group and the residential group are 
described by the Mayombe as bivumu. 

The mvumu is a very interesting type of residential unit on account 
of its joint activities and possessions and the marked authority of its 
head, the khazi or 'protector'.^ This headman may be the eldest of 
a group of brothers, or he may be selected from amongst the most 
competent of the men who form the elders of the mvumu, or one of 
two or three boys chosen as potential headmen and educated by the 
existing khazi for the post (see p. 215). As has been seen, he may even 
be an outsider specially chosen for his gifts of leadership and brought 
into the mvumu for the purpose. 

T he khazi ha s suprem e authority over the members of the matriline- 
age_^ According to Van Reeth's account he allocates land, palm-trees, 
and garden sites from those held in common by the men of the mvumu .^ 
He is in charge of a joint purse composed of a levy on the earnings of 
the men and the contributions of the absent women who send him part 
of the money they earn by the sale of garden produce (see p. 216). This 
money is used to provide bride-wealth for the men and also dowries for 
the girls, and to pay for the expenses of the latter's weddings. It also 
pays for the services of magicians (feticheurs) or healers, if these are 
required by members of the group (see p. 234), and for funeral expenses. 
Van Reeth emphasizes the fact that the khazi is in charge of the labour 
of the members of his mvumu rather than their actual property.^ They 
are his 'subjects' (bilezi), and he could, if necessary, sell them into 

* De Cleene, Inst. Roy. Colon. Beige Bull, ix, 1936, p. 64. As distinct from 
the real eldest brother, the 7igadi khazi. 
^ Van Reeth, op. cit., pp. 18, 19. 
' Ibid., p. 14. 


slavery in the old days. The khasi acts as judge in mvumu disputes and 
it is his job to prevent quarrels that might break up the group, and to act 
as arbitrator in marriage negotiations and discussions as to the amount 
of the bride-wealth to be returned in case of divorce. He must use the 
common purse carefully in order to keep the members of the mvumu 
united and must specially avoid wastefulness. Lastly, he protects the 
group against sorcery and does this by means of his own strong magic 

The members of the mvumu usually constitute a village, but several 
bivumu may build together and in that case each retains its own centre 
and its own presiding khazi. The khazis of such a joint village or town 
together form a council.' 

It will be seen from the following diagram that a matrilineal fraternal 
family composed of two brothers A and Ai and the sons of their sisters 
a and ai, and the sons of the latter's daughters az and 03, will grow in 
two or three generations to form a much wider group of matrikin in 
which new extended matrilineal fraternal families appear. Az and A^ 
might well form the nucleus of such new extended families, and one or 
more of these may in time split off and form a new mvumu. Polygyny of 
course increases the number of residents in such a community but not 
the number of members of the matrilineage, since, however many 
wives a man has and however many children the latter may bear, they 
will not become members of the mvumu. 


I ^ I I I I I I 

X— 9- X-^ A2 32 -> X— >■ X-^ A3 33- 

X— > X A4 a4-> X— s- x-5- A5 35- 

X-» X A6 36^ X-^ X A7 37-> 

Diagram I. Composition of the Mayombe Mvumu 

The members of the mvumu 3re a (female) and A (male). X and x are sons 
and daughters of the male members of the mvumu who return to their own 
mothers' villages and are therefore not given special symbols. Succession passes 
from A to Ai and from Az to A2, to A4. and then As. a, ai, az, 03, 04, and ^5, 
a6, ai, leave the village to marry, as indicated by the arrows, but send back 
money contributions to the head of the family and sometimes return to live 

there in old age. 

The strength and permanence of the mvumu is probably emphasized 
by the shortage of land which would tend to make young men anxious 

' Corporate matrilineages with shared funds and a matrilineal fraternal 
family as the basic residential unit also exist amongst the Sakata and Songo 
(see ref. on p. 212, n. i). 


to return to their maternal uncles' villages to claim their rights to 
garden sites, and possibly also to the importance attached to the per- 
manent ownership of palm-trees, which yield profits over a long period 
of years. The common funds of the mvumu seem mainly to consist of 
money, and it seems probable that the use of a money currency also 
makes for the corporate nature of the matrilineage, curious as this may 
sound. It is a convenient way in which the absent sisters can send 
contributions to the mvumu and it makes it possible to provide for all 
the needs of the mvumu from one source and therefore from one hand 
— the khazi's. 

The nuclear relationships in such a residential unit are plainly those 
between brothers and between brothers and their sisters' sons. It is a 
group based on the ties between men and not on those between women . 
A Mayombe boy is closely associated with his brothers all his life. He 
lives with them in his father's homestead; he moves with them to his 
own mvumu at the time of puberty. He is bound to them by ties of 
economic co-operation and joint ownership throughout life. With his 
paternal half-brothers, on the other hand, he has few links after the days 
of early childhood, since half-brothers go when the time comes to 
settle in their several biviimu. 

{Type Bi) The Bemba-Bisa-Lamba Group 

A very different type of matrilineal organization is to be found among 
a group of kindred peoples who stretch from the Luapula basin across 
"the plateau of north-eastern Rhodesia and parts of north-western 
Rhodesia. These tribes follow rules of matrilineal descent and succession. 
They contract marriage by service and token payments and a series of 
HtuaTacts" which give the bridegroom ultimate, but not immediate, 
rights of removing his bride to his own village. There are thus variable 
rules of residence at marriage, and a divided authority between the 
father and the mother's brother. 

_The Bemba and Lala of the north-eastern plateau of Northern 
Rhodesia, the Bisa and other inhabitants of the Bangweulu swamps, and 
_the Lamba on the Kafue river seem to have a very similar family 
system. The Kaonde and the Lunda of the Kasempa Province are 
probably similar in type, but we have only scanty information on their 
kinship system from Melland's work published in 1923.' 

I have taken the Bemba as characteristic of the group, since I have 
the fullest material on this tribe, which I collected during two visits 
made in 1930 and 1933 respectively." The Lamba were described by 

' F. W. Melland, In Witch-bound Africa, 1923. 

^ I summarize here information contained in m^' 'Mother- right in Central 
Africa' in Essays presented to C. G. Seligman, 1936; Land, Labour and Diet i?i 
Northern Rhodesia, 1939; Bemba Marriage and Modern Economic Conditions, 
Rhodes-Livingstone Institute Paper No. 3, 1940; and an essay in African 



Doke in 1931, but there is no information on the Lala, to the best of my 

Bisa inhabit the Mpika district of north-eastern Rhodesia, and also 
some of the islands and adjacent mainland in the Lake Bangweulu 
area. These peoples have not been fully described. I visited the Bang- 
weulu swamps in 1934 for a period of six weeks. Brief as my visit was, 
it enabled me to establish the fact that the kinship system of these people 
was more matrilineal in emphasis than that of the Bemba, whom they 
resemble very closely in other ways. I did not visit the Bisa settled 
farther to the east of the plateau. 

Economic Determinants 

The peoples of this group are poor in economic resources compared 
to those of the lower Congo. Their material culture is less developed 
and they have been in contact with Europe for a shorter period. The 
soil on which most of them live is poorer and the crops which they 
grow, mainly finger-millet, sorghums, and maize, have not yet been 
developed on a commercial basis and provide little by way of money 
income. Most of these people are shifting cultivators, moving their 
villages every four to seven years. Bemba, Lala, and Lamba are sparsely 
distributed over bush land with little or no shortage of land for 
cultivating. The Bemba population per square mile was reckoned as 
3 "65 in 1934. Tsetse-fly prevents the keeping of stock, except for 
occasional goats. There is thus practically no inheritable wealth amongst 
these peoples, either in the form of land, stock, or money, and this fact 
naturally influences profoundly the nature of their residential and 
kinship groups. 

The Bisa, Ushi, and Unga, on the other hand, live in the Bangweulu 
swamps where cultivable land is scarce and where fishing-rights in the 
lake and the main river are valuable, and considerable trade in fish has 
been developed under modern conditions. 

I think more exhaustive researches will show that some of the varia- 
tions in the family structure of the swamp peoples are due to these 
diflferences in economic resources. 

Ideology and Principles of Descent 

According to Bemba dogma, blood passes through the woman and 
not through the man. The semen merely activates the foetus in the 
womb. Directly associated with this theory of procreation is the Bemba 
stress on the ties of a man with his sister 'born of the same womb' and 
her children, his uterine nephews and nieces. In the royal clan it is the 
duty of each woman to produce as many children as she can to succeed 

Political Systems, 1940. My data on the Bisa and other swamp peoples have not 
yet been published. 

' C. W. Doke, The Lamba, 1931. 


her brothers, irrespective of the status of their fathers, and she is 
allowed, arid even expected, to bear children by a series of lovers in 
order to carry on the line.' An ancestress is remembered by the number 
of children she bore, and even the consort of a princess, who is not a 
member of the royal clan at all, may appear as an important figure in 
tribal legend if he was the begetter of many children for the royal 

With this theory of unilateral descent is associated the dogma of 
matrilineal guardian spirits attached to land. But guardian spirits of 
either line are attached to individuals. Ancestors enter the womb of a 
pregnant woman and become the guardian spirit of the baby to be 
bom, and such ancestors may be of either sex. An interesting index of 
the differing patrilineal-matrilineal balance within the royal dynasty 
and outside it is the fact that in the case of commoners the guardian 
spirit may belong to the father's or mother's line, whereas in the royal 
clan it is the maternal ancestors and ancestresses only who return to 
act as guardians to the child. In the case of both commoners' and 
chiefs' matrilineages the injured spirits of the father's line may return 
to afflict the living with punishment, and the curse of the father or the 
father's sister is specially feared. 

The child belongs to his mother's totemic clan (mukoa) and with this 
clan membership go rights of hospitality, blood compensation, reciprocal 
funeral offices within paired clans, ^ forms of greeting and joking relation- 
ships, and exogamy rules. The child also acknowledges ties of some kind 
with his father's clan, of which he will readily give the name when asked. 
The children of men who belong to the royal clan naturally try to stress 
their links with their father rather than their mother and are given 
special titles as 'sons of chiefs'. Children take their father's personal 
name, not their mother's. 

The Bemba also reckon membership of a matrilineage or 'house' 
{yanda). This is a loosely organized group reckoning its descent from an 
ancestress 3 or 4 generations back in the case of a commoner, 13 or 
20 generations in the case of hereditary court officials, and 25 to 30 in 
the case of the paramount chief himself. This matrilineage is composed 
of a man, his brothers and sisters, his sisters' sons and daughters, and the 
sons of his mother's sisters, and his mother and her sisters and brothers. 
The 'house' is not a local unit nor is it corporate in the sense of owning 
property in common or accepting common authority, but it is a group 

' Note that among the Mayombe, with their corporate matrilineages, it is 
thought to be the duty of all women to produce as many children as possible for 
their brothers, whereas the insistence is only strong in the case of the royal 
women among the Bemba, and it is only the ruling dynasty that can be reckoned 
as a corporate matrilineage in any sense of the word. M. G. Mar\vick tells me 
that among the Cewa children were merely desired as providing future workers 
and as making for more stable marriages. 

^ Cf. my 'Reciprocal Clan Relations', Man, xxxvii, 1937, pp. 188-93. 


in which positional succession is reckoned and which provides wives 
in the case of a default in the marriage contract. The authority of the 
genealogical senior of a 'house' is now of a rather tenuous kind. His 
word carries the preponderating weight in matters concerning marriage 
and other family affairs. He is summoned on all ritual occasions. For- 
merly he could sell members of the matrilineage into slavery. 

The rules of succession follow the common matrilineal pattern. A 
man is succeeded by his brother, sister's son, or sister's daughter's 
son; and a woman theoretically by her sister, but actually by her 
uterine granddaughter.' I have not sufficient data to show clearly the 
span of the matrilineage which acts as the succession unit, but most 
genealogies of commoners are reckoned to a depth of three to four 
generations only, and the succession of brothers and classificatory 
brothers appears to take place within this unit.^ 

Succession to the chieftainship is only a special case of the charac- 
teristic 'positional succession' (ukupyamka) of the Bemba by which the 
social status and kinship position of each dead person, man or woman, is 
passed on to a selected heir or heiress. A successor must be found to 
acquire the guardian spirit of a person recently dead, and the bow of 
the man or the girdle of the woman, and to be addressed by the name 
of the dead and by the same kinship terms. In the case of an im- 
portant man, the name of the departed will be regularly used by the 
successor, but where the dead man is a nonentity the name lapses, 
although fellow villagers will remember, if asked, which name has been 

The inheritance of personal property is not an important determinant 
of kinship sentiment among the Bemba, since most possessions are 
perishable and the dead man leaves little more than his hereditary bow, 
the use of a house which will decay in a few years, or the produce of 
gardens that are soon to lapse into bush. Succession to office and 
positions of authority and the right of access to ancestral spirits are the 
significant links between one generation and another. Bemba succeed 
either to a position in a ruling dynasty or to the social status a commoner 
has acquired for himself by his individual efforts during the course of his 
lifetime. It is not the title of head of a clan or lineage that is inherited, 
but the particular individual status to which each man or woman has 
attained, and I found in fact that the names of insignificant people were 
quickly forgotten, whereas those of eminent men or mothers of large 
families were remembered for some generations. 

' A daughter cannot succeed her mother or she would become the social 
equivalent of her father's wife. 

^ Cf. W. V. Brelsford, The Successioyi of Bemba Chiefs, 1944, and my essay in 
African Political Systems, 1940. 

■* Such a form of 'positional' succession appears to exist also among the 
Lamba (see Dolce, op, cit.), as well as the Cewa and the Yao. 


The Marriage Contract 

^ By the Bemba marriage contract the man gives a series of small 
payments and a contribution to the initiation ceremony of his bride. 
He also works for his parents-in-law for an indeterminate period of 
years. The payments given by a bridegroom were, and are, small in 
relation to the high marriage payments for the Kongo area. They con- 
sisted of one or more barkcloths or some hoes in the old days and 
anything from 5^. to lo^. nowadays. £•}> would be considered a very 
large payment.' Under modern conditions a bridegroom who is away 
working at the mines sends back money specifically 'to cut the trees 
for his father-in-law' if he cannot do the years' service. 

Through his services the husband acquires various rights over his 
wife and children. In the first months of the marriage the girl merely 
sweeps his house and performs small services for him. Later she sleeps 
in Tiishut until the time of puberty when she is withdrawn until her 
initiation ceremony has been completed and the formal consummation 
of the marriage has taken place. The husband henceforth has complete 
sex rights over his wife, except in the case of the consort of a royal 
princess, and the penalties for adultery were very severe in the old days. 
A husband also has the right to the labour of his wife and she has very 
rarely a separate granary of her own, although she distributes the grain 
in the family granary. This is a male dominant society and, even though 
descent is reckoned through the mother, the wife is very much under the 
control of her husband even while he is an outsider in his wife's village. 

The wife's parents give the son-in-law a series of 'gifts of respect' 
in the form of food, and these acts of politeness are very much valued 
and mark the different stages of his acceptance as a permanent son-in- 
law. Finally the rite known as the kuingishya or 'the entering in of the 
son-in-law' is performed, and by this ritual act the last in-law taboos 
are removed and the young man is definitely accepted as a member of 
his wife's lupwa — her family or group of closely related kinsmen. 
Paradoxical though it may seem, once the young man is admitted to the 
family he is free to remove his wife and children to his own village 
should he desire to do so. It is for this reason that I have described 
the marriage contract as marriage with delayed right of bride removal. 
The right is not always exercised, as some young husbands may have 
settled happily in their wives' village and wish to remain there. I did 
not make any quantitative estimate of the conversion rate of such 
marriages among the Bemba, but Barnes and Mitchell give the figure of 
36 per cent, for the conversion of such marriages amongst the Lamba.^ 

' T. CuUen-Young in fact uses the term 'token-transfer marriages' of this 
type of union. See his 'Tribal Admixture in Nyasaland',^. Roy. Anthrop. Inst. 
liii, 1933, pp. 1-18 ; cf. also my Bemba Marriage and Present Economic Conditions. 

* J. Barnes and C. Mitchell, The Lamba Village: a Report oj a Social Survey, 



I should think it likely that the Bemba conversion rate is very much 
higher since the power of the father is stronger in this area. 

Once the husband has been admitted to the wife's family he acquires 
various subsidiary marriage rights. He can, for instance, ask for his 
wife's sister in marriage and he gives no service to his father-in-law in 
respect of this union. On the other hand, even when he has removed his 
wife to his own village she will not necessarily be a permanent member of 
that group. She will probably return to her own people on the death 
of her husband or even when she has ceased child-bearing. Her 
husband's brother can inherit her if that is what she wishes ; this is 
by no means considered a right, but rather something that may be a 
convenient arrangement for her support. 

It will be seen that the marriage relationship in this area is one that 
is progressively established. A man builds himself a hut in his wife's 
village and lives there for four or five years or even longer. During this 
time he is gradually being made free to join in the activities of his wife's 
relatives by a series of ritual acts which bring to an end the taboos that 
kept him apart from his in-laws ; but Bemba marriage in its early stages 
has many of the elements of a trial or temporary union. A bridegroom 
will leave his wife's village during the early stages of marriage if he does 
not think he is being treated with respect, and in this case no formal 
act of divorce seems to be necessary ; Bemba divorce in the old days did 
not, in any case, involve the return of token payments or damages for 
services given. A man's parents-in-law might also refuse him the right 
to remove his wife and her children to his own village after three or 
four years' residence, and the wife herself occasionally refuses to follow 
her husband because she does not want to leave her own family. It is 
only in the case of a successful marriage, where the man has proved 
himself a good husband and son-in-law and has produced children, that 
he gains the right to remove his bride to a community of his own 
choosing, either to that of his mother's brother or even to that of his 
father or, in the case of a middle-aged man, to set up a village of his 

Domestic Authority 

Authority over Bemba children is divided between the father and the 
mother's brother. The maternal uncle formerly had rights of life and 
death over his sister's children, and could sell them into slavery or use 
them in payment of a blood debt incurred by a member of the matriline- 
age. He can still intervene in matters of marriage and can claim part of 
the girl's marriage payment. He still sometimes removes his sister's 
daughter from her husband and marries her to someone else to whom he 
owes obligation. But the father has authority over the children during 
their youth and can maintain this authority if he is a man of personality, 
birth, and status. Thus a chief will be able to keep his sons as well as his 


maternal nephews with him and will give some of them headmanships. 
It might be said in fact that all members of the royal clan contrive to 
gain control over their children as against the authority of their mother's 

A father is always consulted about the marriage of his daughter even 
if he has been divorced from her mother some time previously, and he 
can claim part of any marriage payment made for her. 

The avunculate seemed to me to be much more strongly developed 
among the Bisa and other Bangweulu tribes, and this also seems to be 
the case among the Cewa according to Marwick. It occurred to me that 
this might be correlated with the absence of a strong centralized govern- 
ment and a ruling clan, and I was therefore interested to hear an edu- 
cated Bemba volunteer the statement : 'The father is stronger amongst 
the Bemba than the Bisa. It is a matter of chieftainship. Members of the 
crocodile clan always try to get hold of their own sons as well as their 
sisters' sons.' Is there a correlation between divided authority, as 
between the father and mother's brother, and the existence of social 
differentiation and a hierarchical political system.^ It is a suggestion 
which needs testing in other areas. 

Apart from the powers a man is able to gain over his wife and children 
as distinct from those of the mother's brother, every Bemba who 
establishes a permanent marriage builds up his authority in a variety 
of other ways. To begin with he gains, as a grandfather, what he loses 
as a father. The Bemba bridegroom starts his married life as a stranger 
in his wife's village, but since his daughters remain with him for some 
years after marriage he builds up a grand-family composed of his 
daughters, their husbands, and their children. In some cases he is able 
to persuade his sons-in-law to remain permanently with him, while in 
others he is even able to keep his sons in his own village. In time some 
of his sisters' sons, with their wives, will choose to come and settle 
with him, so that he will be head of a group of rather heterogeneous 

Residential Units 

It will be seen that the basic domestic unit of the Bemba is a matrilineal 
extended family of the father-daughter type, that is to say, a group 
composed of a man, his wife, his young married daughters, and their 
husbands and children. This extended family is composed of separate 
parental families housed in huts in the same village and not fenced off 
in any way from the rest of the community ; but it must be reckoned as 
forming one domestic unit, since the daughters' households are closely 
linked with those of their mother. When a young couple are married 
they have no garden or granary of their own for some years and the girl 
is not even allowed her own fireplace for a year or two. During the 
interval, cooking is done at her mother's house and dishes are supplied 


to her husband as he sits in the men's shelter. After the girl has been 
given her fireplace, much joint cooking will still be done and the gardens 
of mother and daughter will be made near to each other. 

From some points of view the Bemba extended family might be 
described as patriarchal, since the authority of the grandfather over the 
little group of dependent households is so very marked. It is true that, 
according to Bemba theory, the important tie is the one between mother 
and daughter. If the father is dead, or divorced, the mother will live 
with her married daughters and her sons-in-law and will form a domestic 
unit as regards cooking and gardening; but where the father is alive, 
and the marriage has been stable for many years, he is very obviously 
the head of the grand-family. The house is described as his and not 
hers. He selects garden sites and supervises the work of his sons-in-law. 
The latter are described sometimes as 'working for their father-in-law' 
and sometimes as 'cutting trees for their mother-in-law'. It seems to 
depend on the relative social status and personality of the man and 
the woman concerned. But it is certain that the position of a Bemba 
grandfather is one that gives great power and authority and is in some 
respects more enviable than that of the grandfather in a typical patrilineal 
Bantu community, since the head of a Bemba grand-family has estab- 
lished his rule by individual effort and not through the help of his 
brothers or his patrilineage. The status of head of the grand-family is 
reflected in the spirit world. The names of maternal grandfathers are 
remembered in prayers as well as those of maternal uncles, by men as 
well as women, even though, as will be realized, these two belong by 
rule of descent to different clans. A man prays to his own ancestors on 
occasions of family affliction and in some cases he addresses those of his 
wife as well.' 

The extended family is rendered an even more stable unit when 
cross-cousin marriage has taken place. Both types of cross-cousin 
marriage are practised in Bemba society, but I think the marriage of a 
man's daughter with his sister's son is the most common. Such a 
marriage brings into the village sons-in-law who are not strangers, but 
are closely identified by descent with the leading men of the group. 

Though the father-daughter family is the basic domestic unit, it may 
exist in a number of different residential units since, as is common in 
matrilineal societies, there are several alternative forms of marriage. 
Chiefs or members of the royal clan send payments to their wives' 
parents for the cutting of their gardens and do not give service. Their 
superior status enables them to remove their brides to their own villages 
immediately. It is common in such cases for the girl's mother to move 

' Ancestral spirits are of course associated with districts, and where a man is 
still living with his wife's family she or her brother tend to conduct the prayers. 
Where, however, a man has removed his family to his own village he may pray 
to his own and his wife's ancestors as I have described. 


with her to the chief's village so that the son-in-law mother-in-law 
pattern is set up, but in other cases the girl goes alone to marry, and 
this is what would be described as a typical viri-local marriage. Poly- 
gyny, however, raises difficulties in this as in other matrilineal societies. 
Among the Bemba the second marriage is usually viri-local. Only very 
occasionally do the two wives live at separate villages with their own 
people, as is the case in parts of Nyasaland and the Congo. A man who 
marries a widow also generally has the right of immediate bride removal 
and the payment in goods and service is usually very small in such cases; 
but whatever variation there may be in the form of marriage, by the 
time the third generation has appeared some form of the father-daughter 
family exists. 

The basic political unit of the Bemba is a village (Umushi) with a 
nuclear father-daughter grand-family and its numerous accretions. In 
some cases there are two or more of such nuclear households. A Bemba 
village comes into being when a man has sufficient dependants to warrant 
his asking his chief for permission to set up a new community; but the 
headmanship is hereditary and with the passage of years two or three 
grand-families, with their dependent households, become recognizable, 
and the village may have two or three cores. 

The principles of accretion to the nuclear household are numerous. 
The headman's sister will very generally live with him if she is divorced 
or widowed or tired of married life. In any of these cases she will usually 
bring to live with her a few of her married daughters and in some cases 
her sons. The headman may also support some of his wife's relatives, or 
even his brother's wife. 

From figures collected by Barnes and Mitchell in sixteen Lamba 
villages, 60-5 per cent, of the villagers were attached to the headman 
by matrilineal ties; 11 '8 per cent, were more distant connexions; 19*7 
per cent, were affinal relatives, and 8-2 per cent, connexions of different 
kinds." It would be roughly true to say that the more important and 
popular a man is, the more different relatives he is able to attract to live 
with him, although a large village of this type, composed of more distant 
relatives, is likely to split into separate villages with the death of the 
headman in question. It must be remembered too that Bemba villages 
are rebuilt every four to five years, and each shift of the village site 
provides an opportunity for a change in the composition of the com- 
munity. The more casually associated households drift away and join 
up with other senior relatives. The diagram on p. 230 makes some 
attempt to illustrate the different principles of association in a Bemba 

It will be seen that the basic ties in the Bemba family are those of a 
woman with her daughters. This group of women co-operate closely 
in domestic activities for many years. Whereas the Mayombe girl 

' Op. cit., p. 41. 


changes residence from her mother's household to that of her maternal 
uncle and finally to that of her husband, the Bemba girl tends to 
remain \vith her own mother and sisters for a large part of her life. 
The Bemba girl is also closely associated in interest with her own 
brother, who will be her natural supporter in the first years of her 
married life and if she becomes divorced or wishes to leave her husband. 
The brother also shares with her husband control over her children as 

A = b bi 

A2 = c <-A3 33 = R <-B ba = S bs = T 

C CI = D 34 A4 <-Bi b4 bs B2- 

C2 C3 

Diagram II, Composition of a Bemba Umushi 

^ is 3 youngish headman living with two married daughters bz and 63 and 
their husbands S and T and children and his wife's unmarried sister bi. His 
son B had left to marry elsewhere, ^'s widowed sister az lives with him with 
her married daughter 03 and the latter's husband R and their children. His 
eldest sister ai is dead, but the latter's son Az, and his married daughter ci 
and her husband D and children, have come to join A. Thus this village is com- 
posed of members of the clan of the headman A and that of his wife b. A2 is 
his probable successor and heir. 

has been described. The Bemba boy, in distinction, has few economic 
or residential links with his brothers, since the boys of one family 
tend to marry women in other villages and to go and live there. It is 
only when two brothers both decide to bring back their wives to the 
village of their mother's brother in later life that the two will be members 
of one village, and I do not remember a case in which this happened 
during my visit. The boy is linked with his mother's brother by the 
ties of matrilineal descent and the possibility of future succession. It is 
also possible that he may inherit his mother's brother's wife. He is afraid 
of his maternal uncle's authority and views him with much greater 
reverence than his father. However, there are few economic links 
between a man and his mother's brother and a boy is not attracted to 
the latter's village by the hope of inheriting cattle or land there, as is 
the case in the Congo and probably Nyasaland. 

(Type B2) The Yao-Cewa Group 

The kinship system of the matrilineal peoples of Nyasaland is very 
similar to that of their Northern Rhodesian neighbours, but the avun- 
culate seems to be much more strongly developed and there are some 


interesting differences in the composition of the local group. Hence my 
decision to classify the Yao and Cewa as a separate sub-type. 

Preliminary accounts of the social organization of the Yao were 
published by Stannus and Johnson in 1922.^ Mr. Clyde Mitchell, of the 
Rhodes-Livingstone Institute, is at present at work on the Yao and I 
quote from his preliminary reports.^ Mr. Max Marwick has also been 
working recently among the Cewa in the south-west corner of the Fort 
Jameson district and gives me permission to quote from an unpublished 

Both the Yao and the Cewa differ from the Bemba in their lack of 
hierarchical forms of chieftainship and ruling dynasties. The Yao were 
scattered by the Ngoni from 1864 to 1870 and again by punitive raids 
of the British Administration during the anti-slave campaign from 1891 
to 1895. The Cewa were even more dispersed. Both tribes seem to 
consist of a series of largely autonomous villages or groups of villages 
under the overriding rule of a chief. 

The Nyasaland peoples form quite a marked contrast from an eco- 
nomic point of view also. The Yao were formerly great traders and were 
in contact with Arabs for some 200 years before the advent of the British. 
Their territory, mainly in the Liwonde, Fort Johnston, and Zomba 
districts, is accessible to the railway, making it possible for them to 
develop the sale of cash crops, and Mitchell gives figures of the annual 
average income from the sale of tobacco in two districts in 1943-4 as 
£^. ly: lod.^ Wealthy individuals can make as much as £()0 a year. 
This figure would be thought high in Bemba villages, where the rough 
estimate of yearly cash income that I made in one village during 1934 
was about £2. Cattle are not numerous in the Liwonde district east of 
the Shire river where Mitchell has been working. Cattle have been kept 
by the Cewa described by Marwack for the last twenty to thirty years, 
although they are less wealthy than the Yao in other ways, and live too 
far from the railway to make it profitable for them to grow cash crops. 

Another striking difference is the much greater population density 
in the Nyasaland area. Mitchell quotes a figure of 53*4 per square mile 
for the Yao as against 3 7 amongst the Bemba and i -63 in some of the 
Bisa areas. In some of the Cewa districts there seems to be considerable 
land shortage. Marwick gives a figure of over 19 per square mile for the 
area with which he is concerned.'* Both the type of cultivation and the 
land situation may account for the greater permanence of the Yao and 

' H. Stannus, The Wa-Yao of Nyasaland, Harvard African Studies, 1922, 
and W. P. Johnson, Nyasa the Great Water, 1922. 

^ See J. C. Mitchell in 'The Village Headman', Africa, xix. 2, April 1949, and 
'An Outline of Social Organisation of the Yao of Southern Nyasaland', in Seven 
Tribes of Central Africa, to be published shortly. 

^ Admittedly only a small proportion of the population here grows tobacco. 

* R. H. Fraser, 'Land Settlement in the Eastern Province of N. Rhodesia', 
Rhodes-Livingstone Institute Journal, No. 3, June 1945. 


Cewa villages, which in its turn influences the type of kinship tie 
developed in the local group. It is true that new villages hive off from 
the old amongst the Nyasaland peoples, but Mitchell states that many 
settlements have a history of thirty years in one place and of five genera- 
tions of hereditary headmen. This must be compared with an average 
life of four or five years for the Bemba village with three generations of 
headmen at the most. It would seem, therefore, that the Yao have far 
more inheritable possessions than the Bemba, including valuable garden 
sites, house sites, houses, and money, while the Cewa have cattle to 
leave to their descendants as well. 

A Yao belongs to his mother's clan. These clans have a name (hikosyo). 
They were formerly exogamous, but are so no longer. They are now 
widely dispersed and seem to fulfil few important sociological functions 
in the present-day life of the people. Clan names are pronounced after 
sneezing, and there was evidently some recognition of kinship between 
immigrant Yao bearing the same clan names as resident Nyanja, since 
the lukosyo name was used by the Yao as a means of claiming help. 

Within the clan the Yao distinguish major matrilineages or houses 
known as mawele (sing, liwele) (breasts) or tnilango (doorways). These 
groups trace descent from a founding ancestress {lipata, likolo). They 
are similar to the 'houses' of the Bemba, but each house reckons seniority 
according to the position of its founding ancestor or ancestress and is 
described as being 'of the large breast' or 'of the small breast' respectively. 

The effective descent group is the mbimiba, a small unit consisting, 
in theory at least, of a group of sisters and their children under the 
leadership of their eldest brother.' The mbumba traces its descent to 
a common grandmother or great-grandmother, but apparently the 
matrilineage rarely has a greater depth than four generations. Succession 
to ofiice follows matrilineal rules. A man is succeeded generally by his 
eldest sister's son and not by his brother. The most important type of 
succession is to the semi-autonomous and more or less permanent 
headmanships, which are coveted positions giving a man the right to 
rule over a group of his married sisters and their descendants. The 
installation ceremonies for headmen in this area are much more 
elaborate than those found among the Bemba, where the most important 
ritual of this kind centres round the installation of chiefs of the ruling 
dynasty, and this fact reflects the greater importance and permanence 
of the Yao headmanships. The people also practise a form of positional 
succession similar to that of the Bemba. Property is inherited by sisters' 
sons apparently. 

Clan membership among the Cewa presents some differences since 
the vizcongo or clan names are now inherited patrilineally as the result 
of Ngoni influence, although lineage descent and succession are matrili- 

' Bemba use the word bumba for any group of kinsfolk who are under the 
leadership of a senior man. 


neal as among the Yao. Further details on the functions of the Cewa 
clan are not available. 

The marriage contract differs from that of the Bemba in that the 
ceremonial is less protracted and elaborate. Amongst the Yao the 
exchange of presents between the two families is insignificant and 
the bridegroom is not obliged to make gardens for his father-in-law. He 
merely makes a garden for his own wife, and these services do not 
entitle him to remove his bride to his own village in the majority of 
cases. Mitchell states that two types of marriage are recognized by 
different names : kulombela, or marriage without removal of the bride, 
and ktdowosya, or marriage with immediate removal; but he suggests 
that the latter form only takes place in the case of the marriage of a 
headman. Transfer marriages also occur when a man succeeds to a 
position of headmanship and is obliged to leave his wife's village. 

A significant feature of Yao marriage is the importance of the four 
'sureties' or witnesses of the contract — senior and junior representatives 
of the husband's and wife's matrilineage respectively. These sureties, 
who watch over the success of the marriage and adjudicate in cases of 
dispute or divorce, seem to be a sign of the much more corporate nature 
of the Yao matrilineage as compared to that of the Bemba. 

Among the Cewa, service is given to the parents-in-law for a year 
or so, as occurs in Bemba country. Native authorities have recently 
tried to enforce a marriage payment of 30^. to be given to the witnesses 
of a marriage, usually the mother's brother {tsibweni) or the brother of the 
girl, in order that claims may be made in the case of adultery ; but the 
practice is not often followed and this group of Cewa reject the Ngoni 
custom of giving bride-wealth, here described as cizvololo, since this 
would give the father control over the children of the marriage. In fact 
they refuse cizvololo in the case of a marriage between a Ngoni man 
and a Cewa woman. The Cewa headman is allowed to remove his bride 
immediately at marriage. The commoner can remove his wife to his 
own village at the end of a year's service and after the birth of one 
or t\vo children, as among the Bemba, but it is not clear how frequent 
these conversion marriages are. 

In both these tribes the degree of incorporation of the son-in-law 
is never so great as amongst the Bemba apparently. Mitchell says that 
unsuitable husbands are dismissed with compensation and sent away. 
He adds that the husbands of the women of a village are often away 
visiting and are definitely not reckoned as members of the community. 
Stannus states that a widower is usually given a present with the sugges- 
tion that he go elsewhere since he no longer has any standing in his 
wife's village. Marwick indicates that Cewa marriage ties sit loose and 
speaks of a man and his sister and 'her current husband'. 

The Cewa talk of the father as a stranger. 'He is as a beggar; he has 
simply followed his wife.' At divorce he leaves his wife's village with his 



hoe, his axe, and his sleeping-mat and has no right to any of the children 
of the marriage, even if he may have begotten as many as seven children. 
The authority of women seems to be higher in Nyasaland than in 
Northern Rhodesia. Johnson notes that in Yao villages 'the woman head 
of the family is the main outstanding person',' and Mitchell speaks of 
women acting as heads of hamlets composed of their daughters and the 
latter's children, or of two sisters with their children. This would be 
an unusual phenomenon in Bemba society and may possibly be corre- 
lated with the fact that men have always tended to be absent for long 
periods in the year, as Mitchell points out, formerly raiding or trading 
and now doing wage labour. Nevertheless a woman and her children, or 
a group of sisters and their children, are normally under the authority 
of their eldest brother, who is the asyene or owner of their mbiimba, which 
Mitchell translates as 'sorority group'. He is referred to by a special 
kinship term by his sisters and younger brothers and definitely treated 
as the head of a group of siblings. As a young man the eldest brother 
probably lives in his wife's village, but later in life he may succeed to the 
headmanship of his sisters' village and live with them and their married 
daughters. The head of the nibumba settles disputes between members of 
the group, represents his sisters in a court case, pays their fines, makes 
decisions as to the marriage of his sisters' sons and daughters, and pays 
for diviners in the case of illness. In the old days he presumably had the 
right to sell his sisters' children into slavery, as in other parts of Central 
Africa, and Johnson states that a man might choose one of his sisters' 
sons to live with him from the age of 8 or 9 if he wanted to do so. 
The Cewa tsibweni, or mother's brother, under present conditions has 
the duty of correcting and punishing his sisters' children, whom their 
father, in theory at any rate, should never beat. He is also responsible 
for sending them to school. He can get his maternal nephews to herd 
his cattle for him, and they are glad to do so because they hope one 
day to inherit these cattle. So far the duties of the maternal uncle of the 
Yao and the Cewa do not differ very much from those of the mother's 
brother in Bemba society, but primogeniture is evidently much more 
marked amongst the Nyasaland peoples. The eldest brother amongst 
the Bemba is not distinguished linguistically from his juniors and is not 
regarded specifically as head of a group of siblings. He may succeed to 
his mother's brother's position and so become the head of a house, and 
possibly also of a village, but if he does not so succeed I do not think 
he is regarded as having specific powers or rights over his brothers 
and sisters. 

The rights and duties of the Yao father have not yet been clearly 

described. Johnson says that a boy remembers his obligations to his 

father, 'but this is often from a good feeling of pity or of family regard* 

rather than from duty.^ It is clear, however, that some fathers become 

' Johnson, op. cit., p. 68. ^ Ibid., p. 87. 


heads of large households, and achieve considerable domestic authority 
over their wives' children, although not, of course, ultimate control 
over their destinies. In the old days the head of a polygynous family with 
incorporated slaves must have wielded authority over a domestic unit 
of considerable size. But even under present conditions it seems that 
whenever a marriage relationship has been maintained for a number of 
years and daughters have been born, have married and produced child- 
ren, some kind of father-daughter grand-family of the Bemba type is 
bound to appear, and a middle-aged Yao who is the father of married 
daughters probably achieves much the same position as that of a Bemba 
father and grandfather. The difference, apparently, is that this grand- 
family is less permanent amongst the Nyasaland peoples, since it falls 
to pieces immediately the father or the mother dies, and the children 
and grandchildren then return to their own matrikin. There is also a 
linguistic distinction which indicates the different emphasis: a Bemba 
son-in-law is said to be working for his father-in-law, whereas the Cewa 
is said to be working for, or living with, his mother-in-law. 

a = X XI = B x2 Xi^^ 

I Az^ A 

ai <-X2 x3 c = X5 x6 

<-A2 a2 <-X3 x4 X6 xy 

Diagram III. The formation of Yao Mhumha 

X lives with two married sisters, xi and X2, and their married daughters, 
X2 and x6. X, xi, and X2 represent one mbumba. As X2 grows up he assumes the 
guardianship of X3 and x6 thus making another mbumba. The residential group 
here consists of X, his sisters and their married daughters together with A''s 
wife a and her married daughter, ai. 

Cross-cousin marriage between a man Xz and his mother's brother's daughter 
ai will keep X2 in the group in which his sisters live. 

In the case of a senior brother who is selected for the position of head 
of a sororal family, he takes his wife with him and marriage therefore 
becomes viri-local. The head of a Yao village of this sort thus has 
authority over his wife's children and over those of his sisters, but the 
position of the wife's children seems to be clearly differentiated from 
that of the sisters' children, and Mitchell speaks of a traditional hostility 
between them. I never saw evidence of such hostility in Bemba villages, 
which also often contain a man's children and his sisters' children. 

Both Yao and Cewa villages can contain a number of extraneous 
units. Johnson speaks of settlements composed of kinsmen, incorporated 
refugees, traders, and slaves, and Mitchell shows the number of separate 
sororal families which are to be found in a village of long standing. 
Marwick describes a village as possibly combining several groups 


composed of the married daughters and one or two sons of an old 
woman, dead or alive, and additional extraneous simple families attracted 
to the village for economic reasons. 

The matrilineal system of Nyasaland therefore differs from those of 
north-eastern Rhodesia in the greater strength of the brother-sister 
group, the more pronounced avunculate exercised by the eldest brother 
in a society in which primogeniture is emphasized, and the growth of 
villages round a sororal extended family core rather than round a 
father-daughter grand-family. The ceremonial of marriage is less pro- 
tracted, the services demanded of a son-in-law less severe, and marriage 
ties sit looser. Conversion marriages are common among the Cewa, but 
rarer, apparently, among the Yao. It is probable that the ties uniting 
members of the matrilineage are strengthened in these areas by the 
greater shortage of land, the greater permanence of villages, which are 
usually based on matrilineal ties, and the presence of inheritable wealth 
in the form of money, houses, and cattle ; but full studies of land tenure 
and property inheritance have not yet appeared. 

( Type C) The Ila Type 

The Ila-speaking peoples of the Kafue river basin in Northern 
Rhodesia form the fourth distinguishable type of kinship organization 
in Central Africa. It is a system which is described as matrilineal since 
clan membership is reckoned through the mother, and the avunculate 
is strong, but there is some evidence that the individual belongs to 
double descent groups, one patrilineal and the other matrilineal; suc- 
cession and inheritance may take place from the father as well as the 
mother's brother and in the ancestral cult the patrilineal ancestors have 
precedence. The man gives a high marriage payment in the form of 
cattle, which entitles him to remove his bride to his own village 
immediately and to maintain considerable authority over his sons, who 
remain living with him after their marriage, and who invariably inherit 
at least some of his cattle. 

The Ila have linguistic and cultural similarities with the matrilineal 
peoples of north-eastern Rhodesia and Nyasaland, but they have tradi- 
tions of origin from the east, and they keep cattle and express many of 
the ritual attitudes associated with the cattle cult as it is commonly 
practised among the Eastern and Southern Bantu. They apparently 
conquered their present territory about 200 years ago. Their history is 
one of internal warfare and raiding, and they never developed a cen- 
tralized system of government. They are grouped to-day in a series of 
small semi-autonomous chieftainships. 

The Ila were first described by Smith and Dale in 1922,' but have 
recently been revisited by Dr. Smith, who outlined some further features 
of their social structure in a paper read at the Royal Anthropological 

' Smith and Dale, The Ila-speaking Peoples of Northern Rhodesia, igzz. 


Institute, which he has kindly lent me. The Sala described by Brelsford 
in 1935 seem to be similar but more markedly matrilineal.' The Tonga, 
on which Dr. E. Colson is working, and the Lenje probably belong to 
the same group. 

Principles of Descent and Succession 

A man belongs to his mother's clan (mtikoa). This is an exogamous 
unit with members who are dispersed throughout the various terri- 
torial divisions. The mukoa is associated with legends of origin and a 
totemic name and taboos. It is also named after its place of origin and 
prominent ancestors. Its members have certain common rights and 
duties, such as the payment of blood compensation and the giving of 
mutual aid, and they are described as taking decisions in the case of 
the marriages of members. The Ila also recognize allegiance to other 
descent groups known as kameko, amongst these the matriclan of the 
father and those of the grandfathers.^ 

Spirits from both sides of the family can be re-born in a new child, 
although one statement shows that a woman's ancestors have the right 
to be born first and another that a man prefers sons to daughters because 
he can only be re-born through his sons,^ while in a third case a child 
with guardian spirits from both sides of the family is described. Ancestors 
in both the father's and the mother's line are honoured, and it is expressly 
stated that 'All the deceased members of a man's family are his mizhimo' 
(family deities).'* But it is an interesting and unusual feature of Ila 
ancestral beliefs that though clan membership is reckoned through the 
mother the ancestral spirits of the father are mainly responsible for the 
welfare of the children and are appealed to by him. We are told that 
'The wife's (spirits) are her own and only to a certain degree of the 
children, in that they help the father's a little in shepherding and guard- 
ing the children. '5 A man continues to worship his ancestral spirits and 
a woman hers, and there are separate spots on each side of a hut door 
for prayers of this sort — a feature not recorded elsewhere in the area.^ 

According to Smith's new material a man also belongs to a local 
descent group (lunungu) which is composed of a man, his sons, and their 
children, and which therefore might be described as a patrilineage. 
Hence we get a three- or four-generation local patrilineage which 
practises the worship of the patrilineal ancestors, which gives rights to 
the inheritance of some cattle and which is also exogamous, as well as 
a matrilineal exogamous clan. Smith was also told that women can start 
their own lunungu composed of their daughters and the latter's children 

' W. V. Brelsford, X Roy. Anthrop. Inst. Ixv, 1935, pp. 205-15. 

^ Smith and Dale, op. cit., vol. i, p. 295. 

^ Ibid., vol. ii, p. 2. 

* Ibid., p. 165. ^ Ibid., p. 173. 

^ Presumably a woman worships her father's spirits or her mother's. 


which looks as though there may be local matrilineages as well as 
patrilineages. This is obviously a matter on which further research is 

Succession to chieftainship follows the matrilineal line amongst the 
Nanzela group/ and Smith gives an instance of a line of descent of 
twelve chiefs ; but there is considerable elasticity in the case of the suc- 
cession to chieftainships and nothing like a fixed order of succession, as 
is found amongst the Bemba for instance. Amongst the Bwila a brother's 
son or sister's son may succeed to the chieftainship, or even an unrelated 
man chosen by the elders as successor to a vacant post.^ There seems, 
however, to be some effort to get a chief from the same clan as his 
predecessor, and if this is not possible the new chief may take the clan 
of the old as a courtesy title. 

Amongst commoners there is evidently a form of positional succession 
similar to that amongst the Bemba. It is expressly stated that each man 
and each woman has an heir 'to eat the name'. The heir is selected after 
considerable discussion and it is not clear whether he is necessarily a 
brother or a sister's son. The heir inherits a large proportion of the 
property but not all, since a dead man's sons, his sisters' sons, and 
others have a right to a share in the dead man's goods {kukoma)? 

A characteristic of this type of kinship system is the wide group that 
shares in the inheritance and the fact that they are called from both 
sides of the family. Property consists of the dead man's wives, cattle, and 
formerly, slaves. In some of the divisions made the cattle of a dead man 
went to his heir, to his younger brothers, to his sons and daughters, and 
to his sisters' sons. In one case three widows went to the heir, one to the 
sister's son, and one to the deceased mother's family. 

The allocation of cattle as between the different wives of a polygynous 
household is not described and we do not know whether these herds 
remain distinct, as amongst some other Bantu peoples, so that the heirs 
of these herds form a separate subsidiary line of descent as amongst the 
Swazi for instance. 

It is important to note, however, that women can inherit property 
in their own right and can acquire wealth individually. Foodstuffs are 
divided between husband and wife on divorce, and a woman can make 
a private garden or can get cattle from her father, presents from her 
husband, or a tithe of the cattle obtained as compensation by the husband 
when she has committed adultery. A woman in fact can become so 
wealthy that she is chosen as a chief.'^ It may therefore be that a lunungu 
composed of a man's daughters and their children arises in the case of 
women who have acquired or inherited cattle which they wish to pass 

' Smith and Dale, vol. i, p. 304. 

^ Ibid., pp. 300, 304, 305 ; we are even told of a competition to select the best 

^ Ibid., p. 303. ■* Ibid., pp. 380-1. 


on to their own children. If this is the case the lunungu is in fact a cattle 
inheritance unit based on inheritance either through a man's sons or, in 
some instances, through his daughters, and it is distinct from the matri- 
lineal unit of descent, the clan. The inheritance of residence and the 
inheritance of property in cattle seem, in every case, to take precedence 
over the clan affiliation. 

The Marriage Contract 

A Mwila pays a substantial sum in bride-wealth. This was formerly 
composed of mixed goods (blankets, shells, hoes, and cattle). It is now 
normally paid in cattle and was reckoned as four to five head, or about 
j(^i2. 15^., in 1922 — a high payment, since wages must then have been 
not more than ^i to ^(^2 a month. By this payment he secures the right 
to remove his bride immediately to his own village and to possess the 
children, who remain his in case of divorce, even though by clan 
membership they belong to their mother's mukoa. Because of the bride- 
wealth, the woman remains the property of the husband's family, and is 
inherited in the event of her husband's death. Cattle are returned in the 
case of divorce. They are provided mainly by clan members with some 
contribution from the patrikin. The chiko is returned in the case of 
divorce, or where the woman proves barren, unless another woman can 
be provided in her place. The chiko cattle are divided amongst an 
unusually wide group of relatives, mainly the maternal kinsmen. The 
parents are expressly stated to get a very small share. 

Domestic Authority 

The question of authority over the children is a complex one. A man 
could apparently sell his sisters' children into slavery in the old days or 
use them for blood compensation or plan their marriages. Smith and 
Dale say: 'If the interests of the clan conflict with those of the family, 
the former prevail over the latter.'' On the other hand, as has been said, 
the children belong by right of marriage payment to the father and 
remain with him in the case of divorce. They are associated with him by 
ties of residence, economic co-operation, and hopes of inheritance. 

Residential Unit 

The basic domestic unit amongst the Ila is an extended family, 
composed of a man, his wives, his married sons, and the latter's children, 
his unmarried children, and his servants and slaves. This extended 
family has a title — the tnukwashi. It is separated from the rest of the 
village houses by a fence, and its members share in economic activities, 
in particular the care of the joint cattle of the establishment. It is 
not clear how often the mukwashi divides into new groups, but it is 

' Smith and Dale, vol. i, p. 284. 


expressly stated that it is 'not large' and a unit of twenty-five was 
described as unusually big.' There is thus an anomalous situation of a 
typical patriarchal extended family as a basic residential unit of a society 
with matrilineal descent, and the ties of joint residence and ownership 
of cattle are recognized to be as strong as, if not stronger than, those of 
matrilineal descent, common legends of origin, the keeping of taboos 
and rights of mutual help. 

Nevertheless, membership of the clan gives a man a special position 
in the village from which his mother came and where he has rights of 
succession, and he is here described as *a possessor of the land'^ rather 
than as 'treading the land of others'. The Ila recognize the difference 
between the domestic unit under the authority of the father and the 
descent unit of which the mother's brother is head. 'The clan is your 
mother's, the family is your father's.'' 

-Xi XI = Az A3 = b 

X2 x2 bi B 

Diagram IV. The Ila Mukwashi 

X lives with his married sons Az and At, and the latter's children. He does 

not live with a member of his own clan unless his son A2 marries the daughter 

of his sister x. The children of this union Xz and xz will then be members of his 

clan ; his daughter az marries out. 

A man's daughters of course leave the mukwashi to marry elsewhere. 
Men of the mukwashi marry women of different clans, and thus the 
children brought up in the village may belong to a variety of different 
clans and the residential unit is not based on descent at all. However, 
cross-cousin marriage with the father's sister's daughter is practised 
and this will reduce the number of alien clans within the mukwashi? It 
would appear in fact that here, as elsewhere in the area, a cross-cousin 
marriage obviates some of the difficulties of patriarchal rule combined 
with matrilineal descent. 

It will be seen that the nuclear relationships in this residential unit 
are those between father and son. A boy is linked with his brothers from 
his earliest years. He lives with them and shares in the care of the same 
herd of cattle, and hopes to inherit some of these cattle. He is, however, 

* Smith and Dale, vol. i, p. 284. 

^ Ibid., p. 286. 

^ Miss Mary Tew has pointed out to me that the Ila and the Yao practise 
opposite types of cross-cousin marriage in conformity with this residential 


dependent on his matrikin for his bride-wealth, and has rights of suc- 
cession in the village of his mother's brother, where he is 'at home'. He 
has a number of rights and obligations associated with his membership 
of a matriclan. Girls are identified with the members of their mukwashi 
by right of residence and are also 'at home' in the village of their mother; 
but they live in their husband's village and will usually be inherited by 
one of the latter's brothers if they become widows. Thus both boys and 
girls have a double allegiance as regards local affiliation and economic 
co-operation, and this is reflected closely in the ritual and ideology of 

I realize that many points in this analysis are of the nature of hypo- 
theses since further data on descent and residential grouping amongst 
the Ila-speaking peoples are required. I thought the material worth 
including, however, since it shows an extreme case of the variety of 
residential pattern which can exist in a society which practises matrilineal 
descent, and of the power the father can obtain in such societies. It also 
provides an example of a type of social structure in which the principles 
of unilineal descent are less important as a basis of grouping than 
residential kinship ties. 


The lineage structures correlated with these types of family system 
cannot be fully analysed in the compass of this article and with the 
present gaps in our knowledge, but it will be useful to summarize some 
of their salient characteristics. 

The lineage system of the Mayombe and Kongo could be described 
as a segmentary type, using the term in the way that Evans-Pritchard 
has employed it in his descriptions of Nuer kinship and political 
organization, although there are many irregularities due to the intrusion 
of foreign elements. 

The Mayombe are divided into nine exogamous matriclans (tnvila), 
of which the members trace descent from a putative ancestress and 
remember a common legend of origin. The fnvila are divided into 
major matrilineages (dikanda) the members of which trace actual 
descent from a common ancestress. Van Reeth shows one such divi- 
sion with a six-generation span.^ The dikanda are associated with 
named territorial districts and their members have ultimate rights over 
all the land in the district, although each minor matrilineage {mviwiu) 
actually exercises cultivating and hunting rights over its own particular 
area. The head of the senior minor matrilineage in each dikanda is 
recognized as chief [pfumu dikanda). Such a man would rule over his 
own mvumu and others attracted to him, such as immigrant bivumu or 

' Van Reeth, op. cit., p. 4; cf. also De Cleene, Inst. Roy. Colon. Beige Bull. 
ix, 1938, p. 68. 



those of slaves. He would take a leading part in councils of lineage elders 
and is described as forming a centre of political organization in the 
territory rather than as being its chief.' 

Disputes over authority and land have apparently led to a further 
process of division, and the split off of minor matrilineages, the bivumu, 
which trace direct descent from an ancestress, practise exogamy, and 
own land and property in common in the way that has been fully 

The Kongo system is similar except that the lineages are of a dual 
and not a triple order and there is a certain linguistic difference, since 
the kanda are here the exogamous matriclans and the mvila the matrili- 
neages. The Kongo clans are arranged in some kind of hierarchy of 
genealogical seniority as are the matrilineages. There are usually three 
such matrilineages, known respectively as the children of the eldest 
sister, those of the middle, and those of the youngest born, but Van Wing 
records cases of five, seven, and even nine such constituent matrilineages 
of a clan.^ The men of the luvila may form a village, or two or three 
may combine, in which case each will obey its own head and keep its 
own land rights distinct. 

In both these sets of tribes a group of brothers and their sisters form 
a corporate matrilineage accepting the rule of a senior brother. The men 
and their absent sisters are kept closely united by joint ownership of 
land and permanent trees yielding a cash income. Where a split occurs 
and one brother leads off his descendants to form a new mvumu, it 
remains structurally connected with its parent body and with the other 
bivumu in the dikanda in which it can claim residual rights to land; 
it acknowledges the precedence of descendants of genealogically senior 
ancestresses ruling over other bivumu, and accepts the political authority 
of the head of the senior lineage known as the dikanda head, the crowned 
chief or the senior clan elder. 

The Bemba-Bisa system differs widely. Here there is no corporate 
matrilineage owning property rights in common, and the residential 
unit is not based on the co-operation of a group of siblings of one sex. 
The Bemba are divided into forty exogamous totemic clans (mikoa) 
which have the name of an animal or plant, a legend of origin or 
separation from the parent stock, greetings and praise titles, and reci- 
procal burial obligations and joking relationships amongst paired clans."* 
Founding ancestresses and ancestors are remembered in the case of 
the royal crocodile clan and senior clans of which the members claim 

' Nauwelaart, op. cit., 'formant des noyaux de groupements politiques 
voisinants', p. 406. 

^ The 'sub-clan' of De Cleene and the 'schoot' (literally 'womb') of Van 

^ Van Wing, op. cit., p. 119. 

* Cf. my 'Reciprocal Clan Relationships', Man, Dec. 1937. 


descent from warriors and councillors who accompanied the first chief 
when he occupied Bemba territory some 200 years ago. Otherwise 
ancestresses of clans are not remembered with great interest, if they 
are called to mind at all. 

Clans are arranged in a certain order of precedence and are attached 
in people's minds to particular districts wliich their ancestors first 
occupied and where their descendants are still numerous, at any rate 
according to the prevailing belief." Clan members still acknowledge 
obligations of hospitality and mutual help. Senior clans provide here- 
ditary officials for the chiefs' courts. Where a woman is required to 
fulfil a marriage obligation or to perform a ritual act of intercourse, as in 
some of the ceremonies connected with chieftainship, she may be sought 
within the clan, if there is none suitable within the matrilineage. The 
mukoa has no recognized heads except the holders of hereditary offices 
at the paramount chief's court, who are heads of matrilineages which 
have become established in the vicinity of the Citimukulu's village. These 
are sometimes loosely described as 'heads of clans'.^ 

Within the clan there are a number of matrilineages or 'houses' 
{yanda). These are loosely organized groups reckoning descent from an 
ancestress 3 or 4 generations back in the case of a commoner, 13 to 20 
in the case of a hereditary court official, and 25 to 30 in the case of the 
paramount chief himself. The 'house' is not a corporate unit in the sense 
that it owns property in common, but it is the unit in which positional 
and office succession is reckoned and which provides wives in the case 
of a default in the marriage contract. There is no structural inter- 
connexion between these small three- to four-generation matrilineages, 
and they perform no political functions. It is only the royal dynasty 
that acknowledges a common authority and exercises common rights 
and privileges. 

The Yao clans seem to have survived only as distinguishing names 
associated with exogamous rules. They have no local attachment or clan 
head. The effective descent group is a corporate matrilineage consisting 
of the descendants of a common ancestress who acknowledge the 
authority of a senior mother's brother or brothers. Of this group the 
married sisters and the eldest brother form the core of a residential 
group. These small matrilineages arc not structurally inter-connected 
in any way. 

The Ila are grouped into ninety-three exogamous matriclans {mikoa). 

' I have collected no statistical data on this point, but the connexion is so 
clear that I found I was able to guess the probable district from which men came 
when I knew their clan. 

^ The Lamba are divided into thirty-two exogamous clans and Doke states 
that 'property must remain within the clan' (op. cit., p. 193). The Kaonde are 
divided into sixty-two exogamous clans, but their functions are described as 
insignificant except for the performance of reciprocal burial rights (Melland, 
op. cit.). 



These clans form the unit in which blood compensation used to 
be paid and which still entails obligations of mutual aid and the 
payment of the bride-wealth of its members. Clan members exercise 
common rights over fishing-pools and are 'associated in the minds 
of people with certain localities or places of origin'.' They could 
claim the right to parcels of land in such localities and an eminent 
member of a mukoa could be buried there. But clan members were 
dispersed throughout the various territorial divisions and exercised no 
political authority. They performed no joint ceremonial, although they 
occasionally gave competitive displays at feasts for the exhibition of 

The Ila also recognize another more restricted descent group, the 
lunungu, of doubtful structure which is the unit of succession and 
ancestor worship. 

These units are also exogamous. Dr. Smith's recent account suggests 
that a man and his sons and their children form such a unit, which 
would in this case be a patrilineage in a matrilineal society. He also 
suggests that women may start their own lunungu composed of their 
children. I have suggested elsewhere (p. 239) that these are in reality 
units of cattle inheritance which have formed themselves in a society 
in which clan membership is matrilineal, but marriage payments give 
the right to viri- local marriage, and cattle are left to a man's sons and 
daughters as well as to his maternal nephews. The lunungu are not 
apparently structurally inter-connected at all. 



Lineage structure 

Mechanism of 

Rights and obligations 

Type A. Lower Congo. Lineage system with triple segmentation 


1. Clan (mvila) 

2. Maj. lineage 

3. Min. lineage 

Name of ancestress. 
Legend of origin. 

Name of ancestress. 
Legend of split off. 

Name of ancestress. 


Named territory. 
General territorial 

Political authority by 

senior lineage head. 


Authority with senior 

Joint land and property 

Men form residential 


E. W. Smith (unpublished manuscript). 




Lineage structure 

Mechanism of 

Rights and obligations 

Lineage system with dual segmentation 


I. Clan (kanda) 

2. Lineage (luvila) 

Name of ancestress. 
Legend of split off. 
Praise-names in 
archaic language. 

Name of ancestress. 


Arranged in precedence. 

Authority over named 

territory with clan 

Cult of clan ancestors 

particularly at illness 

or death. 

Corporate rights as with 

mvumu above. 
Joint participation in 

burial rites with other 


Type B. N. Rhodesia and Nyasaland. Dispersed clan with local 3-4 generation 





I. Clan {mukoa) 

2. Min. lineage 

I. Clan (Jukosyd) 

2. Maj. lineage 

3. Min. lineage 

Clan (ciwengo) 

Totem symbol. 
Legend of origin. 
Food taboos. 

Name of ancestress 
or her brother. 

Totemic symbol and 
name pronounced 
when sneezing. 

Name of founding 

Name of senior 
brother or group of 

Ciwengo name inheri- 
ted patrilineally. 


Mutual help. 

Reciprocal burial 

Provide chiefs, council- 
lors, and priests. 


Unit of succession. 

Unit of wife replace- 
ment and blood 

Exogamy. Formerly 

Name used to claim 


Recognition of senior 
and junior lines ('large 
or small breast'). 

Brothers, sisters, latter's 

children under 

authority of senior 

Residential group — 

sisters with selected 

senior brother. 

Not knowTi. 




Lineage structure 

Mechanism of 

Rights and obligations 

Type C. N. Rhodesia 


Clan (tnukoa) 

Totemic animal. 


Name of centre. 

Mutual help. 

Name of prominent 



Rights of burial in clan 

Blood compensation. 
Displays at feasts. 

? Patrilineage 

Common house. 

Residential unit. 


? Cattle ownership. 


We have been able to distinguish at least three types of kinship 
organization among the Central Bantu. In all these tribes the matrilineal 
system makes for certain elements of conflict for which some kind of 
solution has to be found. The problem in all such matrilineal societies 
is similar. It is the difficulty of combining recognition of descent through 
the woman with the rule of exogamous marriage. Descent is reckoned 
through the mother, but by the rule of exogamy a woman who has to 
produce children for her matrikin must marry a man from another 
group. If she leaves her own group to join that of her husband her 
matrikin have to contrive in some way or other to keep control of the 
children, who are legally identified with them. The brothers must 
divide authority with the husband who is living elsewhere. If, on the 
other hand, the woman remains with her parents and her husband joins 
her there, she and her children remain under the control of her family, 
but her brothers are lost to the group since they marry brides elsewhere 
and they are separated from the village where they have rights of 

There is the further difficulty that in most societies authority over 
a household, or a group of households, is usually in the hands of men, 
not women, as are also the most important political offices. Thus any 
form of uxori-local marriage means that an individual of the dominant 
sex is, initially at any rate, in a position of subjection in his spouse's 
village, and this is a situation which he tends to find irksome and tries 
to escape from. 

There are, of course, a number of solutions to the matrilineal puzzle. 
The first of these may be described as the matriarchal solution, in that 
property, and particularly houses and lands, pass through the woman 
as well as the line of descent. The eldest brother usually acts as manager 
of the estate. This is achieved either by the institution of the visiting 
husband or by that of the visiting brother. The women of the taravad 


or matrilineal joint family of the Nayar of Malabar live with their 
brother or brothers and are visited by their husbands at night. In the 
case of the Menangkabau the men were members of their sisters' joint 
household (rumah) and spent much of the day there, but they returned 
to their wives' households at night.' Among the Hopi the situation is 
slightly different, in that a group of married women live together and 
own land and houses jointly, but each husband acts as manager of his 
wife's land and chief worker on it, while the brother acts as spokesman 
and ritual officiant for his sisters and as host in their houses.^ 

Such solutions; only seem to me to be possible in big settlements with 
permanent housing, where a man can walk easily from his own house- 
hold to that of his sisters and perform his two functions without clash. 
This is the case amongst the Nayar and something similar occurs in the 
large towns of Ashanti Province in the Gold Coast where a woman 
visits her husband at night. But Central Africa is much less closely 
settled. In Northern Rhodesia shifting cultivation is practised, and 
villages not only move, but are more sparsely distributed, often at a 
distance of 17 to 30 miles apart. In these circumstances either the wife 
or the husband may have to live some distance away from the group to 
which they are legally affiliated. The more densely occupied Yao area 
allows husbands to be away visiting their own relatives more easily 
and it appears that they often do so.^ 

The second solution is that of the fraternal extended family with 
sisters and children 'loaned away'. This is the solution adopted by the 
Trobriand islanders and the western Congo people. Here viri-local 
marriage makes it possible for a groupi of brothers to live together and 
to exercise full male authority over the community, while their sisters 
are loaned out to men in other communities and the children of the 
matrilineage are reclaimed at puberty. 

The third solution might be said to be that of the borrowed husband. 
These are the various forms of uxori-local marriage described in this 
area. The conflict of interests in these societies is probably the most 
extreme, since all the men of a community cannot at the same time act 
as mother's brothers with authority over their own local descent group 
and also as husbands living in their own wives' villages, as the rule of 
uxori-local marriage demands. Amongst the Cewa and Yao it seems 
that the majority of marriages are uxori-local but that marriages are 
easily broken. A man who cannot stand the situation in his wife's village 

' F. C. Cole, op. cit., pp. 253, 266. 

~ Verbal communication from Mrs. Aitken ; cf. also Daryll Forde, 'Hopi 
Agriculture and Land Ownership', jf. Roy. Antlirop. Inst. Ixi, 1931, pp. 357-99. 

^ Daryll Forde suggests to me that the necessary factor is not the size of the 
settlement but its compact and endogamous character. A man can act simul- 
taneously as husband and manager-brother where he has married within the 
same community. Both the size and the endogamous character of the local unit 
seem to me to be relevant factors. 


leaves and goes elsewhere. This might in fact be described as the solution 
of the detachable husband. Added to this certain senior men have a 
right from the start to practise viri-local marriage since they are selected 
as heads of villages or inherit such positions. They thus escape from the 
difficult position in which their younger brothers are placed. In other 
words, there is here a stress on primogeniture with viri-local marriage 
as one of the privileges of the first-born. The Bemba deal with this 
difficulty by the transfer or removal marriage, by which the young man 
is dependent on his father-in-law in youth but gains authority as the 
head of a father-daughter grand-family in his old age; this is a solu- 
tion which is possible in areas where land is plentiful and the setting up 
of new homesteads and villages is easy, and where the political system 
allows of the constant creation of new units. 

Still another way out of the difficulty is the solution of the selected 
heir, which allows the head of the matrilineage to choose one or more 
of his sisters' children to succeed him and to transfer these boys to his 
own village while the remaining children are allowed to stop with their 
father. This seems to be the practice among the Mayombe, according to 
Van Reeth (see p. 217). Kirchoff quotes the similar custom of the Tlingit 
by which one or two sons of a woman marry their cross-cousins and go 
to live with their mother's brother, whereas the rest of the children 
marry viri-locally. The Nuba often adopt one or two sisters' sons to be 
brought up as their own sons in the same way.' 

In every case there are constitutionally recognized alleviations for the 
socially childless father in these matrilineal societies. Slavery allows him 
to gain control over the children of slave wives who thus join his clan, 
and Torday speaks of Kongo fathers choosing slave wives for preference 
for this reason.^ Cross-cousin marriages may give the father control over 
some of his grandchildren that he cannot maintain over his children, 
or bring into his village members of his clan. By the mother's brother's 
daughter marriage of the Yao the son of one of the married sisters 
belonging to the sororal family marries the daughter of the manager- 
brother and therefore contracts what is virtually a viri-local marriage. 

Polygyny and the marriage of widows also allow men to contract at 
least some viri-local marriages in most of the Central African tribes, 
and, whatever the rule of descent, a man who has succeeded in becoming 
the head of a polygynous household is in fact a patriarch over his wives 
and young children. Lastly men of wealth and distinction are able to 
reverse the usual rules of residence. 

' P. Kirchoff, 'Kinship organisation', Africa, v. 2, April 1932, pp. 184-91; 
S. F. Nadel, The Nuba, 1948, p. 31. This list does not of course exhaust the 
logical possibilities. The solution of alternating residence among the Dobu has 
been mentioned already. There is also the separate men's and women's houses 
of the Ga of the Gold Coast. 

* E. Torday, Causeries Congolaises, p. 103. 


Authority over the children of a marriage consists of a series of rights 
and privileges, and many kinship systems allow for the exercise of 
different types of authority by different members of the patrikin or 
matrikin. In Central Africa there may be a fairly complete severance 
between domestic authority exerted over the children brought up in 
their father's homestead and rights over their persons and the marriages 
they contract. These latter may be exerted by a man who lives some 
distance away. Domestic authority is a treasured asset in an area where 
local communities grow up round the core of the extended family and 
the head of such a family is actually or potentially the head of a village, 
and hence of the unit that is the basis of the whole political structure. 
On the other hand, this is an area in which slavery was formerly a very 
prominent institution, and, therefore, whatever hold the father may 
have gained over his children, the power of the mother's brother to sell 
them into slavery must have remained as a potential threat to be con- 
stantly feared. Even amongst the Ila, where the father is a patriarch 
ruling over a large polygynous homestead very similar to that found 
amongst the patriarchal Bantu, Smith and Dale describe the woman's 
brother as having 'power of life and death' over her children. The right 
to sell into slavery is apparently the power described. 

A balance of rights and duties between the patrikin and the matrikin 
tends to produce secondary forms of descent. While it would be true 
to say that in every matrilineal society a man must know his father's clan 
and get some privileges from his patrikin, there are tribes in which both 
patrilineal and matrilineal relatives are organized in definite descent 
groups with names and corporate functions as amongst the Yako. In 
Central Africa the existence of such double descent groups does not 
seem to be proved, but there is a secondary reckoning of descent in those 
tribes in which the patrilineal-matrilineal balance is most even. Amongst 
the Kongo the father's matrikin are known by a separate name and a 
man has sentimental ties wath the men of his father's village in which he 
was brought up, and these informal ties seem to have led to the sug- 
gestion that the Kongo practise dual descent. 

Ancestor worship also reflects the duality of the reckoning of descent. 
The Bemba commoner can inherit a guardian spirit from his father's 
side as well as his mother's, and will pray to his maternal grand- 
father as well as to his maternal uncle, although this would not be the 
case in the royal clan. The Ila pray to the ancestors both of the father 
and of the mother, and Kongo ceremonies honouring the father's 
ancestors have also been mentioned (p. 214). 

In all these forms of family structure the crucial point, I have sug- 
gested, is the question of residence at marriage, and the determining 
factor in this regard is the marriage payment or the type of goods and 
services which the bridegroom gives. In Central Africa the tribes which 
give large amounts of goods or money in bride-wealth, as do the 


Mayombe and the Ila, seem invariably to practise viri-local marriage; 
whereas those who give service or token payments, Hke the Bemba or 
Bisa, do not. The importance of the marriage payment is shown by the 
fact that even amongst the western Congo peoples, where the avunculate 
is most pronounced, a man who gives an unusually large sum to his 
bride's family is able to gain permanent possession of his wife and to 
keep her children with him instead of returning them to their maternal 
uncle at puberty. The passage of cattle as chiko amongst the matrilineal 
Ila is an equally important determinant. The chiko gives the father 
permanent possession of his children and enables him to keep his 
married sons with him to form the basic residential unit, the miikwashi. 
The division of cattle among his sons and his daughters sets up what 
seem to be a number of three- or four-generation patrilineages as well 
as similar groups united through rights of inheritance of cattle from 
their mother. In Northern Rhodesia there are no high marriage pay- 
ments and a man cannot remove his wife and children to his village 
immediately. He may earn the permission to do so through service or 
joint residence, but he will never acquire complete possession of his 
wife and children. 

As regards the stability of marriage in these different types of family, 
we have hints and impressions of various kinds but very little accurate 
quantitative data. The payment of bride-wealth is assumed to encourage 
a stable marriage since the wife's family is pledged to return cattle or 
goods received in this way if she breaks her contract. If this assumption 
is correct, divorce should be rare in the western Congo, where the 
marriage payments are considerable, but I have no figures to show 
whether this actually is the case or not. The divorce rate should also 
be low amongst the Ila, but as a matter of fact Smith and Dale say that 
it is frequent. It seems clear to me that the size of the bride-wealth and 
the way in which it is contributed are determining factors but by no 
means the only ones. Where the avunculate is strong the mother's 
brother usually has the power to take his sister's daughter away from 
the husband to whom she has been 'loaned' and give her to another man, 
and this must be reckoned as at least one of the causes of instability in 
marriage. It occurs amongst the Mayombe, the Ila, and the Bemba, 
and probably elsewhere. Moreover, where the residential units are 
based on ties of kinship through the woman rather than the man, the 
link between mother and daughter, or sister and sister, may become 
so strong that it can threaten the marriage of one of these women. 
Divorces sometimes occur in Bemba society at the time of the conversion 
of a marriage, that is to say when the husband has won the right to 
take his wife back to his own village. The woman may refuse to go with 
him because she wants to stay with her mother and sisters. 

There are also positive inducements to permanent marriage. Bemba 
give token presents not bride-wealth, but a man stands to gain by a 


Stable union because only by this means is he able to build up the basic 
domestic unit which puts him in a position of authority over several 
households and enables him to set up house where he wishes and 
ultimately to build his own village if he pleases. A stable marriage for 
a Bemba is the first step on the road to a headmanship. It is interesting 
to correlate different forms of family structure with greater or less 
stability in marriage in this way, but conclusive answers will have to 
wait until we have even some rough approximation to the differential 
divorce rates in these and contiguous areas. 

Various economic and social determinants of family structure have 
been suggested in these pages, though fuller details of the economic 
organization of all the selected tribes would be required to test the 
hypotheses made. The presence or absence of inheritable possessions 
in the form of land, trees, cattle, or money have been correlated with the 
type of marriage, and the position of the father as against that of the 
mother's brother. It has also been associated with the corporate nature 
of a unit like the western Congo matrilineage, as against the dispersed 
descent groups of the peoples of Northern Rhodesia. It would no doubt 
be possible to start from the same set of facts and reach very different 
conclusions. Is it, for instance, the shortage of land and the value of the 
permanent palm-trees that make members of a matrilineage willing to 
live together in one closely organized group? Or is it the system of 
corporate matrilineages which makes them anxious to develop their 
land together instead of doing so alone, and encourages them to accept 
the rule of their genealogical head ? No doubt there is something to be 
said from both points of view. Where the type of marriage makes it 
possible for a group of brothers to live together, an economic situation 
such as that of the western Congo increases their economic inter- 
dependence and counteracts centrifugal tendencies. Where, on the 
other hand, the family structure provides no basis for a group of co- 
operating men, and where the land supply is so ample that it is possible 
for men to set up new settlements easily, they do so and the matrilineage 
loses its corporate function. Detailed studies of contiguous people living 
in rather similar economic conditions are needed before we can generalize 
usefully on such questions. At present the variables are too numerous 
to permit more than likely guesses. 

This article merely attempts to make some crude comparisons between 
four types of family structure associated with the rule of matrilineal 
descent in Central Africa, and to suggest some possible correlations 
between marriage type, residential grouping, economic and political 



THE Ashanti are the most numerous' of the Akan-speaking 
peoples who occupy most of the southern half of the Gold 
Coast. Though they all speak dialects of a common language^ 
and have a common culture, the Akan peoples were never 
politically united until they all came under the rule of the British 
colonial government in 1903.^ Indeed the Ashanti are famous for the 
wars they fought against other Akan groups and against the British 
throughout the nineteenth century. During the eighteenth century 
Ashanti had emerged as a powerful national state consisting of a con- 
federation of strong and semi-autonomous chiefdoms grouped round 
the kingship whose politico-ritual symbol was the Golden Stool.'^ The 
basis of this confederacy was war. It was first formed in order to throw 
off the yoke of Denkyera, an Akan state which held the Ashanti in 
subjection, and once established, it was strengthened and gradually 
expanded by wars of conquest against other Akan neighbours. 

As the extent and power of the Ashanti confederacy increased, trade 
with the Europeans on the coast became a vital necessity, especially for 
the procuring of flintlock guns and ammunition. In attempting to gain 
control of the trade routes to the coast the Ashanti came into conflict 
with the coastal Akan, and subsequently with the British who had 
assumed the protection of the latter. The Ashanti fought six wars 
against the British. The last was the Ashanti revolt of 190 1 which ended 
in defeat and annexation. 

Though it was able to wage external wars with success, the Ashanti 
State was not internally altogether stable. The great chiefs guarded 
their regional autonomy jealously against encroachment by the king 
and this sometimes led to armed rebellion. The king's forces, aided by 
those of the other great chiefs, always, however, succeeded in suppres- 
sing a rebellious chief. The necessity of unity was recognized by the 
great chiefs of the confederation and had the backing of powerful 
religious sanctions. 

' They now number approximately 750,000 people. 

^ See J. G. Christaller, Dictionary of the Asante and Fante Language, 2nd 
ed., Basel, 1933, Introduction. 

^ The standard history of the Gold Coast is W. W. Claridge's History of the 
Gold Coast and Ashanti, 191 5. See also W. E. F. Ward, A History of the Gold 
Coast, 1948. The extension of British rule over all the Akan peoples was a 
gradual process which took half a century. 

* Cf. R. S. Rattray, Ashanti, 1923, chap, xxiii. 


^ The political history and structure of the State decisively influenced 
the whole social order of Ashanti. The large degree of autonomy 
reserved by the component chiefdoms of the confederacy led to emphasis 
on the local allegiance of individuals and groups as opposed to citizen- 
ship of the confederacy. Captives and fugitives from defeated tribes, 
added to slaves obtained by purchase, introduced alien elements into 
every chiefdom. Sometimes also many of the subjects of a rebellious 
\ chief sought refuge in other chiefdoms. And in spite of the legal 
priority of local allegiance, individuals often emigrated from one 
chiefdom to another for economic reasons, such as trade, as the result 
of marriage, for religious reasons, such as the desire to become an 
adherent of a nationally famous god, and so forth. In these vi^ays kin 
groups often became dispersed. These influences are reflected in the 
kinship institutions of the Ashanti. 

The main features of Ashanti social and political organization are 
well known from the ethnographic works of the late Captain R. S. 
Rattray.' Indeed the outlines of the Akan system were known to scholars 
long before Rattray first described it.^ As regards Ashanti kinship 
institutions, in spite of rapid social and cultural change during the past 
forty years, the principles elucidated by Rattray still prevail. 

In this paper I shall give a broad outline of Ashanti kinship organiza- 
tion as it emerges in their social structure to-day. It will be useful to 
give special consideration to certain aspects which Rattray did not 
sufficiently stress, in order to correct misinterpretations of Ashanti 
kinship based on his work. This applies particularly to the view that 
Ashanti kinship is characterized by the equal recognition of both 
matrilineal and patrilineal descent.^ 

' R. S. Rattray, Ashanti Proverbs, 1916; Ashanti, 1923; Religion and Art in 
Ashanti, 1927; Ashanti Law and Constitution, 1929. The accuracy and insight 
of Rattray's writings, based as they were on intimate contact with the Ashanti 
for nearly twenty years and a deep knowledge of their language, are acknowledged 
by Ashanti chiefs and elders. Dr. K. A. Busia, himself an Ashanti, has paid 
tribute to the quality of Rattray's data in his (unpublished) thesis on 'The 
Position of the Chief in the Modern Political System of Ashanti' ; and my own 
inquiries in Ashanti during 1945-6 (vide my paper 'The Ashanti Social Survey: 
A Preliminary Report', in Rhodes-Livingstone Institute Journal, vi, 1948) fully 
confirmed his judgement. 

^ There is an excellent summary of what was then known of Akan kinship in 
Frazer's Totemis77i and ^xoga?ny, 1910, vol. ii, pp. 553 ff. It was well known, 
from the writings of early travellers, that the Akan peoples reckoned descent 
in the female line. When Frazer wrote the 7jtoro divisions had also already been 
discovered by Mr. (later Sir) C. H. Harper; cf. reference in Frazer, op. cit. 

^ Cf. e.g. G. P. Murdock, 'Double Descent', Amer. Anthropologist, n.s. xlii. 
4. I, pp. 555 ff., who writes, ititer alia, 'As is now well knovm, the Ashanti have 
both matri-sibs (abusua) and patri-sibs (ntoro).' 


The Political Aspect 

The rule of matrilineal descent, as Rattray made clear,' is the key to 
Ashanti social organization. This is due mainly to the fact that matrilineal 
descent is the basis of a localized lineage organization which is 
generalized through the social system as a whole by an organization of 
dispersed clans. 

Every person of free matrilineal descent (an odehye) is by birth a 
member of his mother's lineage (abusiia) and a citizen of the chiefdom 
in which this lineage is legally domiciled. A matrilineal descendant of 
an alien or slave woman is a member by right of birth of the lineage into 
which that woman was tacitly incorporated,^ though he is subject to 
disabilities which not even the lapse of generations can altogether wipe 
out. Through marriage or, formerly, through flight or falling captive 
in war, a w'oman might live and have offspring outside her natal chief- 
dom. In such a case her matrilineal descendants still retain their con- 
nexion with her natal lineage and their rights of citizenship in her natal 
home, though they can exercise them only by taking up residence there. 
Dispersal does not deprive members of a lineage of their status, of 
rights to succeed to office or property vested in the lineage, or of their 
relationship to the lineage ancestors and gods. Thus one of the most 
prized rights of citizenship to-day is the right to occupy and farm virgin 
forest land freely in the chiefdom to which one's lineage owes allegiance. 
Nowadays it often happens that descendants of women who were 
captured in war return to their ancestral homes to claim land on this 
basis. Again, when a stooP falls vacant it often happens that a member 
of a branch of the lineage which has for several generations been 
resident outside their natal chiefdom is invited to return to take up the 
office. To the Ashanti themselves a lineage consists of all the descendants 
of both sexes by a known genealogy of a single known ancestress in the 
unbroken female line. 

The localized matrilineage is found in its most precise and explicit 
form in those parts of Ashanti which have a tradition of permanent 
settlement over a long period. Every long-established Ashanti village 
or township was until recently divided into wards (brono), each of which 
was occupied by the majority of the members, both male and female, 

' Rattray, 1923, and passim in his other works. 

- As described by Rattray, 1929, chap. v. 

^ A chiefship, lineage headship, or other public office held by a man or a 
woman is always referred to as a stool (kotmua). Carved wooden stools play a 
big part in Akan social life as symbols of office and of status and, duly consecrated, 
also serve as ancestor shrines. The holder of any public office is installed by 
being ritually presented to the ancestors of the group in the room which houses 
the ancestral stools. See Rattray, 1927, igzg, passim. 


of a single lineage. In some cases this lineage would be a maximal 
lineage tracing common descent from an ancestress ten to twelve 
generations back ; in others this unit would be a major segment of such 
a lineage, the main body of which occupied what was regarded as the 
original ward of the whole lineage. The lineage has a segmentary struc- 
ture, each segment being defined in relation to other segments of a like 
order by reference to common and to differentiating ancestresses. This 
allows of both accretion to and differentiation within lineages. Thus non- 
local Ashanti women coming into a village as wives or captives auto- 
matically adhere to the lineage of their own clan. If their matrilineal 
descendants remain in the village they become an attached segment of 
the local lineage. For most social purposes they are treated as true 
members of the lineage ; but they are not eligible for any office held by 
the lineage. Foreign slave women gave rise to many attached branches of 
authentic lineages, for the lineage principle is so dominant that a person 
or group cannot be completely absorbed by an existing lineage but is 
always treated as an actual or potential lineage segment. The norms of 
residence are the chief institutional means through which the lineage 
principle thus asserts itself (see below, pp. 261-2). Members of an 
attached lineage are known to be of alien origin and are not entitled 
to call themselves odehye. But it is a gross insult, actionable in law, for 
anyone to refer to their origins in public. This is in keeping with the 
maxim which has the force of a moral injunction as well as of a legal 
rule that no one should reveal another's origins.^ 

Though the lineage is segmentary in form it is dominated by the 
rule of inclusive unity. There is no hierarchy of jural status or religious 
authority corresponding to the hierarchy of segments. The corporate 
unit recognized for political, legal, and ritual purposes is generally the 
most inclusive lineage of a particular clan in the community. The 
solidarity of such a lineage in relation to the rest of the community is 
impressive. Ashanti are reluctant to admit to outsiders that lineage 
kinship is a matter of degree. They insist on the identification with one 
another of all lineage kin. But in personal matters and in the relations of 
lineage members among themselves, degrees of matrilineal connexion 
are closely observed. 

Every such politically unitary lineage has a male head {abusiia panin) 
who is often ipso facto one of the chief's councillors. He is chosen from 
among all the living male members (adehye), irrespective of age or 
generation, by consensus of the whole body of members of both sexes.^ 

' Obi nkyere obi ase. Rattray rightly emphasizes this in many places. This is 
one of the main reasons why genealogical inquiries are difficult in Ashanti and 
explains the apparent confusions one meets with in the ideas Ashanti have about 
their kinship system. 

^ In practice, of course, the decision rests mainly with the older men and 
women, especially with the former, and only mature men are chosen for the 


Personal qualities of tact, leadership, intelligence, and knowledge of 
affairs determine the choice. 

It is the duty of the lineage head, with the aid and advice of tjie older 
men and women, to watch over the welfare of the whole groupJ He has 
the power and the duty to settle private disputes between any of his 
fellow members so that peace and solidarity can prevail in the group. 
He is its chief representative in its political and legal relations with other 
lineages and with the community. He must take the lead in organizing 
corporate obligations, such as the funeral of a member, and he has to 
1 I contribute the burial cloth. The crucial payment by the bridegroom by 
^;;;J which a daughter of the lineage is legally espoused goes to him and his 
V concurrence is essential for a divorce of any member of the lineage to 
x^*^ be valid. Most important of all, both as a sign of his status and as 
"^ sanction of his authority, is his custody of the male ancestral stools of the 
lineage. These ancestor shrines comprise the consecrated stools of his 
predecessors in office and they belong to the lineage as a whole. A sub- 
ordinate segment cannot have its own ancestor shrines. On Adae days' 
\ he and his elders meet to pour libations and make offerings to the 
ancestral stools on behalf of the lineage. Individual members of the 
lineage may bring offerings to him for the stools in cases of personal 
misfortune or in expiation of sin or sacrilege. The heir of a deceased 
member must be formally approved by the lineage head and his elders, 
and a widow of any member cannot remarry without their consent. 
Such a lineage also has its god or gods (obosom) whose shrines are vested 
in the lineage. 

The commonest occasion on which the unity and solidarity of the 

lineage receive public expression is the funeral of a member.^ It is 

presided over by the lineage head. All the members of the lineage gather 

to take part in it. The expenses, which may be heavy, are shared equally 

by all the adult members — not, be it noted, among the segments of the 

lineage. This is nowadays one of the most important criteria of lineage 

membership. All are expected to observe mourning taboos, though 

distant lineage kin of the deceased will not do so as strictly as near kin. 

A death is felt as a shock and a loss by all members of a lineage wherever 

they may be. 

•/ J (In all these matters there is a very high degree of equality between 

// male and female members of the lineag^. It is common for the head to be 

I assisted by a senior woman (obaa panin) informally chosen by him and 

his elders. This is extended to the State as a whole and to each chiefdom. 

The senior woman of the royal lineage is the Queen Mother of the 

j chiefdom. Her duty is to watch over the morals of the women and 

' For an account of the Adae ceremonies see Rattray, 1923, chaps, v-ix. 

^ Until quite recently every maximal lineage in a village or township had its 
own cemetery, which was sacred ground used for burial only by its own members, 
Christianity has changed this. 


*" girls, to supervise such feminine n^atters as girls' puberty ceremonies, 
which are occasions for joint activities by most of the women of the 
lineage, and to help to make peace in family quarrels. In particular she 
is often the best genealogical authority in the lineage and this gives 
her great influence. The female ancestors of a lineage which owns a 

\ chief ship are also commemorated and paid ritual homage. The con- 
secrated stools of the queen mothers are preserved in the custody of 
the ruling queen mother and she performs rites of worship similar to 
those of her brother the chief on Adae days. This does not apply to 

-commoner lineages. 

"In these situations, moreover, all. members recogniz ed as belon ging 

to the lineage are treated equally, irrespective of their true origin. It 

is only when it comes to the headship of the lineage that non-authentic 

members are excluded,' and this is because they have no direct access 

'to the lineage ancestors. 

Ashanti say 'one lineage is one blood' (abusua haako mogya baako); 
and again, that a lineage is 'one person' {ntpa koro). This is not a ques- 
tion of physiological theories but a way of stressing the corporate unity 
of the lineage. It is symbolized in the rule of lineage exogamy and in the 

^extension of the notion of incest {mogya fra — mixing the blood) to 
include the whole lineage. Incest is a crime and a sin. In the old days it 
was punished by death. ^ Nowadays the culprits are heavily fined and 
are obliged to provide sheep for expiatory sacrifice to the State ancestral 
stools. The public scorn they incur very often causes them to leave the 

^community. The rule of exogamy is strictly observed, the apparent 
exception being that marriage is allowed with a descendant of an alien 
woman adopted into the lineage not more than four or five generations 

; Though all political offices, including the kingship, are vested in 
maximal lineages, ownership of land and other economically useful 
property is commonly vested in a segment of such a lineage. Yet in 
theory any member of a maximal lineage is eligible for selection as heir 
of a deceased member if he or she belongs to the appropriate kinship 
category. This rule also applies to the right to marry the wddow of a 

1 lineage member. In practice inheritance and the levirate (which is 
thought of as a duty owed by the heir in return for the property he 
inherits) are restricted to a segment of the lineage. Mutual aid — as when 
a member gets into debt or funeral expenses have to be met — is extended 

' Except when no true descendants survive, or, rarely, if none of the true 
members is deemed fit for the office. Cf. the right of slaves to inherit in the 
absence of true matrilineal kin, Rattray, 1923, pp. 43-4. 

^ Cf. Rattray, 1929, p. 304. Mogya die, eating the blood, is an alternative term 
for incest. 

^ Data on the marriages of i ,000 women in one chief dom show only 4 per cent, 
of apparent intra-lineage marriages. In every case one partner was of alien 
maternal descent. 

.V V 


throughout the hneage, but the heaviest responsibility falls on this 

This segment, sometimes described as the children of one womb 
(yafunu), generally consists of the uterine descendants of an ancestress 
not more than four, or occasionally five, generations antecedent to its 
living adult members. That is, it consists of matrilineal kin whose 
female parents (and grandparents) usually grew up in a single matrilineal 
household as a group of siblings. 

It is primarily within this segment that the jural authority of the 
mother's brother is effective. The segment has no appointed head, but 
one of the experienced senior males is accepted as the leading personality. 
As in all departments of Ashanti social life, any matter of importance 
that concerns the segment or one of its members is always dealt with 
after consulting the senior men and women. But the leading man will 
have the decisive voice. If the segment owns land, or fishing-rights, or 
other corporate property he administers it, and his successor will be 
chosen by reference to his capacity to act as leader in the affairs of the 
segment. In jural or religious matters (such as marriage or the swearing 
of an oath) in which the assent of the maximal lineage head is necessary, 
the procedure is as follows : the principal party goes first to his (or her) 
legal guardian, his mother's brother, who takes him to the segment 
leader, who then goes with them to the lineage head. 
- The maximal lineage found in a community is thought of as the 
widest local extension of the yafunu segment, and the kinship terms 
used in the matrilineal household are applied to all members of the maxi- 
mal lineage. People of the same generation call one another 'sibling' {nua), 
irrespective of sex. Women of the mother's generation are called 'mother' 
{ena, ni, 7nafni) ; men of this generation are addressed and spoken of as 
'mother's brother' (zvofa), and the men and women of the mother's 
mother's generation are called 'grandparent' (nana). A woman calls all 
the members of her children's generation ba, child, and a man calls 
them sister's child, wofase. Both men and women call their fellow 
members of their grandchild's generation nana, grandchild.' The con- 
ventional norms of behaviour symbolized by these terms apply through- 
out the lineage, but again more strictly within the yafunu. Much 
conduct felt to be absolutely binding within the yafunu is felt to be a 
matter of goodwill and decent standards outside that range. It is of 
great significance, too, that witchcraft (bayi) is believed to be effective 
only within the lineage. It is only within the lineage that amity and 
solidarity are binding in virtue of absolute sanctions, and the inevitable 
inter-personal hostilities and jealousies thus aroused are expressed in 
this belief about witchcraft. 

' A ver>' full list of kinship terms, including their variant usages, is given in 
Rattray, 1923. These terms are still in use, though the English loan-words juami, 
mother, and ^a^a, father, are rapidly replacing the Akan terms in everyday speech. 


K The maximal lineages which constitute a township or village generally 
vary in numbers and in status. The lineage which provides the chief 
and queen mother ranks above the others and is usually the most 
numerous. My inquiries suggest that there are usually two or three 
maximal lineages which account for the bulk (half or more) of the 
population, while the remainder of the population is distributed among 
the other lineages. The significance of the lineage principle is clearly 
seen in the terms of address used towards the lineage kin of one's father 
and paternal grandparents. Any male member of one's father's lineage 
may be called or referred to as 'father' (agya), and a female member of 
his lineage may be called sezva (lit. female father). Similarly, any 
member of one's father's father's lineage may be called 'grandparent' 
(nana). These terms are used when the speaker wishes to show filial 
respect to members of these classes of paternal kin. Ashanti explain that 
this is because any member of one's father's lineage may become his 
heir, and the same rule applies to one's grandfather. Indeed, it applies 
to any kinsman or kinswoman. 

A maximal lineage is regarded as the local branch of a widespread 
matrilineal clan. The same term abusua is used for the clan as for the 
lineage. It is believed that all the lineages of a clan are the matrilineal 
descendants of a single remote ancestress for whom a mythological 
emergence is generally claimed.' The local lineages of the same clan 
cannot, however, show their precise genealogical connexion with one 
another. Members of the same clan but of different local lineages behave 
towards one another as if they were distant kin of the same lineage. The 
rule of exogamy applies. A person coming to a strange place will receive 
hospitality from his clansfolk there, and will be accepted as one of them 
if he wishes to stay. He will be given kinship status in accordance with 
his age. If no suitable candidate for a lineage ofiice can be found locally, 
the elders may seek out a clansman from elsewhere. But clanship does 
not, of course, automatically confer rights to property and office in a 
different lineage from one's own. 

There are only eight clans in Ashanti^ and every lineage belongs to 
one or other of these clans, and every clan is usually represented in 
every chiefdom. This is attributed to dispersion due to migration, 
wars, &c. 

Rattray asserts that each Ashanti clan has a totemic avoidance, an 
animal or bird which clansmen are forbidden to kill or eat. I was unable 

' Cf. Rattray, 1929, passiin, for records of the mythological ancestn,^ of 
Ashanti clans. As he (ibid., chap, viii) and other writers before him and since 
(cf. J. B. Danquah, Akan Laics and Customs, 1920) have pointed out, the same 
clan names and clan organization are found among all the Akan peoples. 

^ The correct list is given in Rattray, 1929, pp. 66-7. The 'dual organisation' 
he refers to appears nowadays only as an alternative naming usage for each clan. 
Cf. also M. J. Herskovits, 'The Ashanti Ntoro', J. Roy. Anthrop. Inst. Lxvii, 
1937, PP- 287 ff. 


to obtain confirmation of this. It may be that the belief is dying out 
among the majority of Ashanti or else that it is an esoteric ritual secret 
(as Rattray implies) associated with chiefship and not readily divulged 
to a foreigner. 

The clan system is important as a unifying force in the political 
organization and as an expression of the cultural unity of the whole 
people.' The fact that there are only eight clans and that one or more 
high-ranking chiefships are vested in each of them strengthens this 
V effect. Every child knows that every Ashanti belongs to one of these 
■ eight clans and knows the name of his own clan. So, also, every child 
knows that the kingship is vested in a lineage of the Oyoko clan, and 
all members of this clan shine in the reflected glory of this connexion. 
This gives Oyoko lineages a little extra prestige; but every clan has 
claims to prominence in one area or another, for one reason or another. 
Hence all are considered to be of equal rank. An Ashanti is proud of the 
chiefships vested in lineages of his own clan. 
^ Chiefs who belong to the same clan call one another 'brother' and this 
'•' is not a mere title of courtesy. Often this connexion has the support of 
a tradition that the founders of the chiefships were the sons of one 
mother. Chiefs thus fraternally related often consult together over 
urgent public issues irrespective of their immediate allegiances. When 
one of them is installed, his brother chiefs send him obligatory gifts 
and he, in turn, sends gifts to thank them. They may have special 
ceremonial duties at his installation; and again, when a chief dies his 
brother chiefs must attend the funeral with special gifts and may have 
ceremonial duties in connexion with it. These ties are thought of as 
holding between chiefdoms — the stools — not as being personal, and 
they often in the past formed the basis of concerted political action.^ 

This is not the only way in which kinship norms enter into the 
political organization at the level of the chiefdom and the State. Marriage 
with daughters of other chiefly lineages as well as with daughters of 
commoner lineages has always been deliberately contracted by all 
chiefs, and especially by the Ashanti kings, as a means of creating and 
cementing political alliances. The lineage principle ensures the per- 
petuation of such ties as bonds between chiefly lineages and they are 
often invoked for political ends. Chiefs of different clans or of different 
sections of the national army^ thus often had traditional ties of clas- 
sificatory kinship. These links arose also from the practice of reserving 
certain subordinate chiefships coming under the jurisdiction of the 

' The fact that the clan system is identical for all the Akan-speaking peoples 
is the most important symbol of their cultural unity. 

^ In a recent constitutional case collusion was suspected to have occurred 
between two chiefs on this score. 

^ See Rattray, 1929, chap. xv. The groupings followed in the national army 
still regulate procedure in big national gatherings and in the work of the 
Confederacy Council. 


king or of the leading divisional chiefs for sons {oheneba) and sons' sons 
(ohenenana) of the king or the chief. 

The Domestic Aspect 

The rule of matrilineal descent governs the legal and moral relations 
of individuals as well as the structure and relations of politically signifi- 
cant groups. But its action is subject to limitations arising out of the 
other genealogical ties rooted in Ashanti family organization. This 
comes out graphically in the structure of the domestic group,' that is, 
the individual household occupying an independent dwelling. The 
household may be under a male head or a female head. If it is under a 
female head, it is normally a segment of a matrilineage consisting of the 
head and her children, her sister and her children, and perhaps her 
own and her sisters' uterine grandchildren. The household of a male 
head, on the other hand, may be either a parental family, consisting of 
a man and his wife or wives and their children, or it may include the 
head's sister and her children as well as his wife and his own children, 
with, sometimes, the children of his children or of his nieces. These 
domestic arrangements represent different ways of reconciling the 
potentially conflicting claims and sentiments characteristic of Ashanti 
kinship. On the one side are the overwhelming bonds of matrilineal 
kinship which embrace those arising out of motherhood ; on the other, 
the ties of marriage and of paternity. Ashanti are much preoccupied 
with this problem and constantly discuss it ; and though it has obviously 
been aggravated by missionary and modern economic influences, there 
is much evidence to show that it existed in the old days. 

Missionary teaching lays special stress on the bonds of marriage and 
parenthood ; and modern opportunities for accumulating private means 
and holding fixed property, such as cocoa farms and buildings, work in 
favour of the ties between parents and children. As a result the solidarity 
of the maximal lineage has declined in matters of a personal or domestic 
nature ; but the strength of matrilineal kinship within the constellation 
of kinship ties crystallized in the domestic groups remains unimpaired. 
By Ashanti law the right of inheritance is confined to the matrilineage 
and men take precedence over women in the inheritance of a man's 
property .y Brothers should take precedence over sisters' sons and the 
latter come before sisters. Formerly a mother's sister's son (a brother — 
nua-barima — in Ashanti terminology) inherited in preference to own 
sister's son. Nowadays this is not so and if no male heir descended from 
the deceased's mother is eligible, his own sister will demand acknow- 
ledgement as heir rather than permit the mother's sister's son to inherit. 
In inter-personal relations the trend is to stress lineal connexions in 

' A detailed analysis is given in my paper 'Time and Social Structure — An 
Ashanti Case Study', in Social Structure — Studies presettted to A. R. Radcliffe- 
Brown, 1949. 


successive generations in preference to collateral connexions of higher 
genealogical status. 

The relative strength of the potentially divergent family ties we have 
mentioned can be roughly gauged from data given in the paper just 
cited. It is unusual for people to reside permanently v^^ith distant 
maternal or paternal kin. In a relatively stable Ashanti community 
between 40 and 50 per cent, of the population live in matrilineal house- 
holds under female heads, and only about a third of ail married women 
reside with their husbands, the remainder living chiefly with matrilineal 
kin. About half of the children under 15 live with their fathers, a large 
proportion because their parents are living together, and the other half 
live in households presided over by their mothers' brothers. In Ashanti 
there are no avoidances between relatives-in-law. It is significant, there- 
fore, that a man and his wife's brother {akonta reciprocally) are never 
found as members of a common household and that it is uncommon 
(though not unknown) for a woman and her husband's sister (akonta 
or kunjna, lit. female husband) to reside together. On the other hand, it 
is not uncommon for a man to have both his own children and his 
sisters' children residing with him, but the preference of most men is to 
have only one class of filial dependants living with him. These are the 
arrangements found in practice, though the ideal of Ashanti men is for 
marriage and the domestic family to be patrilocal. It appears that the 
norm is for the ties of matrilineal kinship to be more or less equally 
balanced by those of marriage and fatherhood in inter-personal social 
relations; and this is achieved either by segregating their respective 
fields of action or by assimilating one class of social ties with its opposed 
class. This comes out clearly in Ashanti kinship usages and norms. The 
balance, however, is precarious, and the conflict and discord arising 
out of failure to reconcile antagonistic kinship rights and duties are 
chronic symptoms of the instability of Ashanti family relationships. 
The most general sign of this state is the resentment constantly expressed 
by Ashanti of both sexes and all stations in life against the restraints and 
frustrations which they attribute to the observance of the rule of matri- 
lineal descent. Yet they continue to abide by this rule, acknowledging 
that its force derives from the political and legal system. 

The Ashanti regard the bond between mother and child as the key- 
stone of all social relations. Childlessness is felt by both men and 
women as the greatest of all personal tragedies and humiliations.' 
Prolific child-bearing is honoured. A mother of ten boasts of her 
achievement and is given a public ceremony of congratulation. As 

' 'A barren woman (obo?iin) is looked upon with pity not unmixed with scorn. 
She feels an outcast. And the lot of the childless man (okrawa) is equally hard. 
However rich he may be he feels that there is something seriously lacking about 
him if he is sterile.' So writes Mr. T. E. Kyei (Dip. Ed., Oxon.), who was Prin- 
cipal Research Assistant to the Ashanti Social Survey. 


intercourse between husband and wife is prohibited only during the 
seclusion and convalescent period of eighty days after delivery, children 
often follow rapidly after one another. By custom, still generally 
observed outside the big towns, birth must take place in the mother's 
natal home/ This not only ensures that the woman is under the care 
of her own close maternal kin, in particular her mother, at this time, 
but it fixes, in a very tangible way, the lineage affiliation and citizenship 
of the child. This custom greatly strengthens the inclination of women 
to reside with their mothers or close maternal kin rather than with their 
husbands during the early years of marriage. Moreover, the care of a 
young child falls almost wholly on the mother ; and as she generally has 
her own farm to cultivate or other income-earning work to do, it is a 
great advantage if she can leave her children in her mother's care during 
her daily absences.^ Maternal grandmothers (nana) play a great part in 
child-rearing. Indeed they can sometimes be very autocratic in this 
regard, arguing that a grandchild (nana) belongs more to the lineage 
(abusua) than to its parents and therefore comes most appropriately 
under its grandmother's care. 

Mother and Child 

But the critical feature, Ashanti say, is the bond between mother and 
child. They look upon it as an absolutely binding moral relationship. An 
Ashanti woman stints no labour or self-sacrifice for the good of her 
children. It is mainly to provide them with food, clothing, and nowadays 
schooling that she works so hard, importunes her husband, and jealously 
watches her brother to make sure that he discharges the duties of legal 
guardian faithfully. No demands upon her are too extreme for a mother. 
Though she is loath to punish, and never disowns a child, an Ashanti 
mother expects obedience and affectionate respect from her children. 
She is always spoken of and addressed as 'mother' (maami, ena, &c). 
To show disrespect towards one's mother is tantamount to sacrilege. 
Ashanti say that throughout her life a woman's foremost attachment is 
to her mother who will always protect and help her. A woman grows 
up in daily and unbroken intimacy with her mother, learns all feminine 
skills from her, and above all derives her character from her. This often 
leads to differences between husband and wife and so to divorce. For 
a man, his mother is his most trusted confidant, especially in intimate 
personal matters. A man's first ambition is to gain enough money to 
be able to build a house for his mother if she does not own one. To be 
mistress of her own home, with her children and daughters' children 
round her, is the highest dignity an ordinary woman aspires to. There 
are very many sayings and proverbs that express the attachment of 

' Cf. Rattray, 1927, pp. 56 ff. 

^ Until weaned at the age of about one year, an infant is almost never separated 
from its mother. She carries it to her work and it sleeps beside her. 


Ashanti to their mothers and indicate the importance of the mother in 
social hfe.' In the poHtical sphere the concept of 'mother' is appHed to 
the chief's female counterpart, the 'queen mother' (ohemaa), who is in 
fact usually his own or classificatory sister. Like a mother's control of 
her children, a queen mother's authority depends on moral rather than 
legal sanctions and her position is a symbol of the decisive function of 
motherhood in the social system. As Ashanti often point out, a person's 
status, rank, and fundamental rights stem from his mother and that is 
why she is the most important person in his life.^ In the individual's 
life-history his or her mother stands for unquestioning protection and 
support against the world at large. It is of special theoretical interest, 
therefore, that she also stands for the source of the most dangerous 
occult force that can impinge on his life, that of witchcraft {bayi). 

'Mother' to an Ashanti generally means his own mother. But the 
terms of reference and address used for own mother are widely applied, 
first of all to all women of his mother's generation in his lineage, and 
secondly to other women of her generation, as a term of courtesy. A 
person's own mother is, however, unique and his feelings and attitudes 
and modes of behaviour towards her are extended to no one else. 

Custom, upbringing, and mode of residence decree very close 
identification of full sisters, and public behaviour towards one's 
mother's sister is indistinguishable from the behaviour one shows to 
one's mother. Indeed it is considered disgraceful to distinguish between 
them in public. An orphan left motherless in early childhood will 
generally be brought up by his mother's sister and in such a case she 
wall be treated as if she were his own mother. But in normal cases it is 
accepted that people feel differently about their own and their proxy 


The Concept of Ntoro 

Ashanti beliefs about the physiology of conception reflect the social 
values attached to the parents. The traditional ideas recorded by Rattray^ 
are widely known and are still accepted among Ashanti who have had 
little contact with Western teaching. Even among the more sophisticated 
(but not necessarily literate) men and women, who maintain that a child 
has its father's blood {bogya) as well as its mother's, the belief that there 
is a closer physiological bond with the mother than with the father 

' A very common saying is Wo na azvu a, zvo ahusua asa—ii your mother dies 
you have no Hneage-kin left (cf. Rattray, 1916, p. 129). Other proverbs also 
stress the idea that one's mother is irreplaceable, that no one else can do for 
her child what she can do, that it is unthinkable for any person to disown his 

^ Cf. Rattray, 1923, p. 79. 

^ Rattray, 1923, ch. xi; 1927, p. 51. As he puts it, conception is believed to 
be due to the blood (bogya) of the woman mingling with the spiritual (ntoro) 
element of the male. 


prevails. Those who speak EngUsh say 'your mother is your family, 
your father is not'. But it is also generally accepted that there is a unique 
spiritual bond between father and child. This is what Rattray described' 
as the 'male transmitted ntoro (spirit)'. There is nowadays, as Herskovits 
has shown in a penetrating critical discussion of Rattray's statements,^ 
some confusion among Ashanti about this concept. 

According to my own observations (and these are confirmed by the 
independent inquiries of Dr. K. A. Busia and Mr. T. E. Kyei),^ the 
account given by Rattray of the traditional complex of beliefs on this 
matter is substantially correct.'* These beliefs are rapidly disappearing 
except among the older men and among lineages in which chiefships of 
siiperior rank are vested. A State's well-being is bound up with the 
health and well-being of its chief; and no matter what his private beliefs 
may be, a chief is obliged to ensure his and his State's good fortune by 
regular observance of all the religious acts connected with his position. 
The washing (i.e. ritual purification) of his kra is such an act. To know 
one's ntoro and perform its rituals is a hall-mark of free if not noble 
ancestry. It is a product, also, of the political importance attached to 
patrilateral ties among chiefs and councillors. 

Father and Child 

The crux of this aspect of Ashanti kinship lies in the great aflPective as 
well as jural weight attached to the recognition of paternity. The general 
rule that a child is considered by law to be the offspring of its pater 
prevails in Ashanti,^ but the ideal is that pater and genitor should be 
the same person. It is a sin and a crime, for which both parties are 
nowadays liable to a heavy fine and to public obloquy,^ for a girl to 
conceive before her puberty ceremony. But after this ceremony no 

' Ibid. 

^ M. J. Herskovits, loc. cit. 

^ In unpublished memoranda kindly placed at my disposal by them. My own 
inquiries were made in several villages, including Asokore, where Herskovits 

■* Herskovits states that 'according to the Asokore men then, Rattray's terms 
ntoro and abusua are to be replaced by 'Ara-washing group' and ntoro res- 
pectively'. This is certainly not standard Ashanti usage. The confusion arises 
to some extent because the ritual beliefs and practices associated with the ntoro 
concept are becoming obsolete. To most young people the term is a vaguely 
understood archaic word for kinship, hence they equate it with terms for the 
kinship groups which rule their life. But there is also deliberate misrepresenta- 
tion. Only Ashanti of freeborn male ancestry have ntoro, in the traditional sense. 
People of slave descent in the male line have none. It is such people who some- 
times insist, no doubt in self-defence, that ntoro refers to the clan. 

^ Thus an adulterine child belongs to its mother's legal husband, in accord- 
ance with the maxim Okromfo nni ba, a thief has no child. 

*■ Part of the fine is used to procure sheep for a purification sacrifice to the 
communit5''s ancestral stools. Formerly (cf. Rattray, 1927, p. 74; 1929, p. 306) 
both culprits were put to death or driven out of the village. 


disgrace follows premarital pregnancy, if the girl has not been promis- 
cuous. What is, however, reprobated and considered shameful, to the 
man as well as to the girl and her maternal kin, is refusal on his part to 
acknowledge paternity of the child. The latter is fully legitimate, as far 
as his status in his matrilineal lineage is concerned, but he carries a 
stigma which may be thrown at his head in later life in a quarrel. 
Paternity is acknowledged by the man's accepting the responsibility of 
maintaining his lover during her pregnancy and by his giving her and 
her child a number of customary gifts immediately after delivery.' On 
the eighth day after its birth (or nowadays, very soon after the eighth 
day) the child is named by its acknowledged father; and this is the 
critical assertion of fatherhood. If an unmarried girl's child and its 
physical parents suffer disgrace it is because the child was not named by 
its father.^ 

A man generally chooses a name for his child from among those of 
his forebears on either side of his parentage, but he is not bound to do 
so. Ashanti say that a man wants children so that he can pass on the 
names of his forebears. It is a very important filial duty as well as a 
source of pride. Hence the children of brothers and the children of a 
polygynist by different wives often bear the same names. The modern 
custom of using one's father's or one's own personal name as a surname 
(and not a name indicating one's lineage affiliation) goes back to the 
father's right of naming. 

Ashanti connect this with the father's sunsum (his personality con- 
ceptualized as a personal soul) or his kra (his spirit, the source of his 
life and destiny), or, if they are versed in traditional religion, his 
ntoro. The ideas connecting the three ritual concepts thus associated 
with fatherhood cannot be discussed here.^ It is enough to say that 
Ashanti believe that a child cannot thrive if its father's sunsum is 
alienated from it;'^ that its destiny and disposition are fixed by the kra 
which is transmitted by the father ; and that this puts every person into 
one of a limited number of named quasi-ritual categories, the ntoro 
divisions. Each ntoro appears to be under the aegis of a god [obosom), 
from whom it takes its name, and whose abode is a river or lake.^ The 
number of ntoro groupings recognized varies from seven to twelve in 
different areas, as the major nation-wide groupings have split up in 

' Rattray, 1927, p. 6. Nowadays the gifts consist of imported toilet articles, 
clothes, soap, trinkets, &c., which may cost two or three pounds. 

^ Ibid. This rule applies as strictly to Christian as to pagan Ashanti. If the 
father refuses to name a love-child its mother's brother will do so and will accept 
paternal responsibility for it. 

^ Cf. Rattray, 1923, pp. 77 ff. ; 1927, pp. 51 ff., 154. Cf. also Herskovits, 
loc. cit. * Cf. Rattray, 1929, p. 8. 

* This is seen from the names of the most widespread ntoro, e.g. Bosommuru 
— the god of the river Buru, Bosomtwe — the god of the sacred lake of the same 
name, &c. See Rattray, 1923, chap. xi. 


adaptation to local circumstances. Each local ntoro division has its 
sacred day for the ritual purification of the kra of its adherents, specific 
totemic taboos, a distinctive form of response to a greeting from any of 

- its members, and a number of names which are commonly though not 
exclusively found among its members. It is also believed that each ntoro 
goes whh. a particular type of character which its adherents are prone 
to have. The ntoro concept thus gives expression through ritual beliefs 
and sanctions, through etiquette and through names, to the value 
attached to paternity. Women as well as men are bound to honour 
paternity. A woman does so by observing her husband's ntoro taboos 
during pregnancy and nursing. If she does not, the child suffers injury 
through the anger and hostility of the father's ntoro. The ntoro concept 
emphasizes, in particular, the bonds of father and son which give 

1 continuity to the male side of family and kinship relations. Adherents 
of the same ntoro in any locality do not meet for ritual or social purposes 
and do not constitute an organized group analogous to the lineage. Nor, 
according to my information,^ is membership of the same ntoro division 
a bar to marriage, as Rattray asserts. What is prohibited is marriage 
between patrilateral kin connected through males only not farther back 
than a common great-grandfather. As Ashanti put it, marriage is pro- 
hibited between the descendants of a man in the male line up to the 
fourth generation. The possession of a common ntoro is incidental in 
this connexion. It is the kinship thus symbolized which is the bar to 
marriage, not the ritual bond. This rule is asserted not by a count of 

f generations but by reference to known ancestry ; and paternal ascendants 
farther back than the great-grandfather (or descendants of a man beyond 
the fourth generation) are rarely remembered.^ One of the main reasons 
for this is that there is no corporate organization based on the father's 
line, nor are jural or political rights (such as rights of inheritance) or 
duties derived from paternal descent.^ The relations of father and child"* 
are rooted wholly and solely in the fact of paternity. The bond is asW 
unique as that of mother and child, though with very different emphasis. 
One's father's brother (i.e. any man whom one's father calls brother {nua 
harima)) is addressed and spoken of as 'father' {agya, papa, &c.) and his 
sister (any woman whom he calls sister {nua baa)) is one's 'female father' 
{sewa) ; but the respect and the frequent affection and goodwill shown 

' I have records of cases of men marrying women of their own ntoro, and 
Dr. Busia has given me instances of this as well. Elders assured me that this is 
not due to modern changes. 

* See, in this connexion, footnote 2 on p. 279 below. 

' The exceptions are certain high palace functionaries serving the king whose 
titles are transmitted from father to son. This is explained either on the assump- 
tion that the original holder of a title was a slave or captive, or on the grounds that 
the duties involved can only be learnt by growing up with them, as a son grows 
up under his father's care and instruction. 

* Rattray, 1929, chap, xi, discusses this subject with profound insight. 


to them are said bluntly to be because of one's father. It arises from 
their mutual identification as siblings. 

An Ashanti father has no legal authority over his children. He cannot 
even compel them to live with him or, if he has divorced their mother, 
claim their custody as a right. Ashanti say that children should grow 
up in their father's house' but investigation shows that not more than 
50 per cent, of pre-adolescent children are found living with their 
fathers at a given time. Nevertheless it is regarded as the duty and the 
pride of a father to bring up his children, that is, to feed, clothe, and 
educate them, and, later, to set them up in life. Even after the divorce 
or the death of the mother, a conscientious father discharges these 
duties. This applies especially to sons. Their moral and civic training, 
in particular, is his responsibility and this gives him the right to punish 
them if necessary. If anything, Ashanti fathers (unlike mothers) tend 
to be over-strict in exacting obedience, deference, and good behaviour 
from their children. Nowadays it is strongly felt that a child's schooling, 
which is widely regarded as the most important preparation for gaining 
a living in the modern world, is his father's responsibility. Formerly he 
took his sons to farm with him or taught them his craft. Investigation 
shows that in practice approximately half of all children attending 
school are kept there by their fathers. Of the rest, half (25 per cent, of 
the total) are supported at school by their maternal uncles, and between 
10 and 15 per cent, by their mothers.^ An uncle, a mother, or other 
kinsman will only take on this responsibility if the father fails to do so. 

The spiritual tie believed to unite father and child is a ritual symbol 
of the moral aspect of their relationship. Ashanti say that a man has no 
hold over his children except through their love for him and their 
conscience. A father wins his children's affection by caring for them. 
They cannot inherit his property, but he can and often does provide 
for them by making them gifts of property, land, or money during his 
lifetime or on his death-bed^ (by written will nowadays). To insult, 
abuse, or assault one's father is an irreparable wrong, one which is 
bound to bring ill luck (mmtisuo). While there is no legal obligation on 
a son or daughter to support a father in old age, it would be regarded 
as a shame and an evil act if he or she did not do so. And, as with all 
kinship ties, the bond with the father is given tangible expression in 
funeral rites. It is the sons and daughters (mmaa mma), own and 
classificatory,'* of a man and the brother's sons and daughters of a 
woman who provide the coffin. 

' The proverb says, 'A child grows up in its father's house but does not remain 

^ Unpublished data based on a sample of nearly 700 children of all ages. 

' Cf. Rattray, 1929, chap. xi. 

■* Generally those descendants of the deceased's father's father whom he called 
sons and daughters. 


Parents have a very important part in the marriage of their children, 
for it is by marriage that a child is established as a socially responsible 
person. The consent of both parents, though not legally compulsory, is 
by usage deemed indispensable for any marriage to take place. A father 
should provide his son with a wife ; and even if he seeks out a bride for 
himself, no respectable young man would marry her without his father's 
approval. Nor would a respectable girl marry without her parents' 
approval. In recognition of the parents' labours in nurturing her and of 
their concern for her well-being, special gifts are due to them from the 
bridegroom on her marriage. In this respect the father receives as much 
honour and consideration as the girl's legal guardian, her mother's 
brother. Indeed the former carries greater weight than the latter, as can 
be seen from statistics collected in 1945. It was found that in a sample 
of 525 women of all ages, about 33 per cent, had been given in marriage 
by their fathers, as compared with 28 per cent, given in marriage by 
their mothers' brothers ; and though success in courtship depends to a 
very great extent on winning the girl's mother's favour, yet without 
the father's approval (if he is alive) the match would not be allowed. 
For it is the father's duty to make sure that the suitor is able to support 
his daughter. This appears from the sample just mentioned. Only 13 
per cent, of the women interviewed claimed that their mothers gave 
them in marriage to their husbands. 

The idea that a father is primarily responsible for his son's moral 
behaviour is seen in the rule that if a man commits adultery with another 
man's wife, the responsibility for paying the adultery damages falls on 
the father, not the mother's brother. Ashanti say that is why fathers take 
pains to get their sons married as soon as the boys are old enough. A 
girl's moral conduct is chiefly her mother's concern. 

Ashanti say that no man loves his sisters' children as much as his 
own children. They say that sons are the support of their fathers. 
Chiefs, in particular, stress this. The more sons and sons' sons a chief 
has, the more secure does he feel. As their social standing depends on 
him and as they have no rights to his office, they will support him in all 
circumstances. They are his most trusted followers, and important 
chiefs appoint their own and their brothers' sons and sons' sons 
(ohenemma and ohenenana) as titled councillors to attend closely on 
them. In the old days, men say, a youth preferred to follow his father 
to war rather than his mother's brother. 

Ashanti kinship, both from the point of view of the individual and 
from that of the system as a whole, is an arrangement of polar relation- 
ships. Where kinship is involved, a person is always looking two ways. 
The division between the sexes is a first consideration. Men have Nj( 
greater political power than women; but political status comes from 
lineage affiliation which is conferred by women, and this redresses the 
balance. The result, in practice, is that there is a high degree of equality 


between the sexes.' In terms of personal behaviour and attitudes, there 
is often no apparent difference between the relations of mother and 
children and those of father and children. The warmth, trust, and 
affection frequently found uniting parents and offspring go harmoniously 
with the respect shown to both. In terms of jural and moral relations, 
however, Ashanti contrast the parents who, as we have seen, very often 
live apart. Through one's father one is given access to social relationships 
in which voluntary choice can play a part. These affect one's personal 
life, not one's public character and roles. Men and women are usually 
very fond of their half-siblings by the same father; but there is no 
sense of inescapable obligation about this, as there is about all social 
ties through one's mother. And this is what Ashanti emphasize in regard 
to everything that concerns matrilineal kinship. 


The Mother's Brother 

Polar social roles in members of one sex give rise to sharper dicho- 
tomies. Ashanti bring out the characteristics of a father's role by con- 
trasting it with that of the mother's brother.^ A person's oldest living 
mother's brother (zvofa) — using this term in the generalized or clas- 
sificatory sense— within the lineage segment sprung from his mother's 
grandmother is the male of the parental generation in whom is vested 
sole legal authority over him. A wofa cannot discipline his sister's child 
(wofase) unless she or her husband asks him to ; but he can command,^ 
and in former times had the power to pawn, his wofase. Nowadays a 
man can demand financial assistance from his nephew but not from his 
son — though it is the latter who is most likely to help out of love and 
goodwill. A good uncle will help his nephew, e.g. by paying for his 
schooling or setting him up in business, in much the same way as a 
father often does with his son; but Ashanti look upon the former as 
creating an obligation to render an equivalent return either by doing the 
same in the next generation for their own nephews or by doing some- 
thing for the generous uncle's other matrilineal dependants. For a 
marriage to be legal it is essential that the girl's wofa should accept the 
nsa payment and hand it to the head of his lineage. In divorce the wife's 
mother's brother's acquiescence and willingness to return the nsa 
payment is essential to make it legal. 

The critical element in the relation of mother's brother and sister's 
son is the latter's status as the former's prospective heir. This comes out 
in the kinship terms used in addressing or referring to them by their 

' Rattray showed this in many places in his books. 

^ Rattray's analysis (1929, chap, iii) of the mother's brother's position is still 
valid, and shows his usual deep understanding of the Ashanti. 

^ It is very significant that a slave called his master wofa, mt father (agya). 


non-lineage kin. Thus a man's children address his sister's son as 
father (agya) if they wish to show respect to him, since he might well 
step into their father's position some day. An honourable and con- 
scientious man who inherits his mother's brother's property and position 
will treat the latter's children with the consideration due from a father. 
Again, a man can address his zvofa's wife as wife (yere) since he has the 
right to inherit her if his uncle dies. Cases of adultery with an uncle's 
wife occur now and then. The sister's son is fined and his action is a 
grave wrong against the uncle.' Though never condoned, it is explained 
by reference to the potential right of the levirate symbolized in the 
kinship terms used for a mother's brother's wife. 

In theory nephews inherit in accordance with the seniority by age of 
their mothers ; but as the heir has in every case to be approved by the 
lineage head and his elders, the claim of an incompetent or spendthrift 
man can be set aside in favour of a junior nephew of better character. 
For it is emphasized that the heir inherits not only property such as 
farms, houses, his uncle's gun,^ clothes, &c., but also his debts and 
responsibilities. One of his chief responsibilities as heir, according to 
traditional custom, is to care for his predecessor's young children and 
widow. The simplest way to fulfil this duty is to marry the widow {kuna 
aware). Young people, both men and women, are averse to this nowa- 
days. The women insist just as strongly as the men on freedom to go 
their own way after the death of a husband. But the formalities of 
inviting the widow to accept her husband's heir still form part of the 
final funeral ceremony on the eighth day after burial. Older men often 
honour the obligation,^ which they describe as a just return for the 
benefits derived from the property inherited. Those who still adhere 
to traditional beliefs say that the uncle's ghost (saman) will punish the 
nephew who 'eats' (di) his property but refuses to care for (fwe) his wife 
and children. 

Matrilineal inheritance is still the legal and customary norm, though 
it is denounced on all sides. The Confederacy Council has endeavoured 
to establish a rule first introduced by one of the Christian missions for 
its own members. According to this rule a man's estate is divided into 
three parts, one for the matrilineal heir, one for the children, and one 
for the widow's use during her lifetime, but reverting to the heir on 
her death. The Confederacy Council has, however, no means of enforc- 

' It is said that certain branches of lineages owning chiefships were dis- 
inherited because of such a crime committed by sons of the women who founded 
these branches. 

^ The heir is described as the one who takes over his uncle's gun, this being 
the chief symbol of a man's status as citizen. An inherited estate belongs to the 
whole lineage segment descended from the original owner's mother and it is 
the heir's duty to manage it for the common benefit. 

^ In the previously mentioned sample of marriages, only 3 per cent, of the 
women admitted to being inherited wives and all were over 40. 


ing such a drastic change, even though pubHc opinion appears to 
support it, and the measure has remained a dead letter outside the 
^ Christian community referred to. Criticism is especially severe of the 
increasing tendency of matrilineal heirs to eject their predecessors' 
widows and children altogether from the enjoyment of properties in 
the building up of which they have often assisted. This is said to under- 
mine the ties of marriage and of paternity which are, in any case, 
precarious by comparison with those of matrilineal kinship. Women 
_V- prefer to work for themselves as an insurance against the future, rather 
than to assist their husbands. It is felt, also, to be inconsistent with the 
ideas of justice and of reciprocity which, in Ashanti thought, underlie 
the institution of inheritance. Nevertheless, only a minority of men take 
care to provide for their wives and children in the customarily lawful 

; way by making gifts of property to them during their lifetime or by 
verbal (or, nowadays, written) testament. Inquiries show that three out 
of four cocoa farms pass from one generation to the next by matrilineal 

Ashanti say, as Rattray records,' that a sister's son is his mother's 
brother's enemy, waiting for him to die so that he may inherit.^ But they 
insist equally on their common interests and axiomatic mutual loyalty. 

/' In fact, both aspects are observable in the relations of uncle and 
nephew. No social relationship in Ashanti is so ambivalent. The ideal 
is to assimilate the relationship, on the personal and affective side, to 
that of father and child. This, however, is achieved only by a minority 
of exceptionally conscientious men. Even men and women who grew 
up in the care of an uncle speak of him with respect but often without 
real affection. Adult nephews are often found living and working with 
an uncle in the most amicable way ; but they are nevertheless very ready 
to find fault with the uncle for alleged selfishness in spending what 
should be preserved for the nephew to inherit, or for unfairness in the 
demands they make though they have a legal right to do so. If such 
criticism is levelled at a father the importunities of his sisters and their 
children are blamed. 

The easiest solution is, as has been pointed out, to segregate potentially 
conflicting ties within the field of kinship. The more complete the 
segregation is the more satisfactory is the solution. A divorced woman 
or the mother of a love-child feels entitled to lean more on her brother 
than one who is married. The brother feels the obligation to provide 
for his sister's children as fully as he does for his own. Though their 
father is held by custom to be obliged to support them, divorce often 
makes him less amenable to their claims on the grounds that a wife and 
children of an existing marriage require prior consideration. A man 
whose sister's children live with him can more easily be a kind of father 

' Rattray, 1929, p. 20. Wojase eye dom is the Ashanti saying. 

^ Hence it is a taboo for a man to sit on his living mother's brother's stool. 


to them, as well as their legal guardian, if his own children are living 
elsewhere. And the same principle comes out again in the common 
Ashanti view that a child stays with its father till adolescence and then 
goes to his uncle, or, if a girl, to her husband. Many individual cases 
conform to this generalization, but it is not universal; indeed not 
more than 10 per cent, of adults fit the rule. As we have seen, a con- 
siderable proportion of children live with their mothers' brothers during 
childhood. It applies strictly only in the sense that after adolescence the 
influence of the mother's brother becomes preponderant in a person's 
social activities owing to the increasing importance of legal relations 
such as those of marriage, land holding, &c. The affairs of adults are 
considered to be the concern of the lineage and hence of their immediate 
legal guardians, not of their fathers. This is compatible with the 
Ashanti ideal that every married man should set up an independent 
household as soon as he is able to do so. 

The obsolescence of widow remarriage is a symptom of the increasing '/a 
divergence between the ties of fatherhood and marriage on the one 
hand, and those of matrilineal kinship on the other, which is found in 
Ashanti to-day. This is associated with the narrowing down of the 
lineage range within which precedent rights of inheritance are accepted 
— that is, the range within which the bond of siblingship is accepted as 
automatically binding. 


Next to the bond between mother and child none is so strong as that 
between siblings by the same mother. Ashanti see that it is simply the tie 
between mother and child translated to the level of generation equality. 
The most important difference socially recognized between siblings is 
that of age. An older sibling is entitled to punish and reprimand a 
younger brother or sister and must be treated with deference. He is, 
conversely, obliged to help his juniors in trouble. This applies especially 
to the mother's first-born (piesie) who is regarded as the head of the 
sibling group and also receives special consideration from his or her 
parents.' But in all other respects it is the equality and solidarity of 
siblings which Ashanti specially stress. The theme is common in dis- 
cussions of the vicissitudes of marriage. 'Your brother or your sister, 
you can deny them nothing' is the way it is often put. Complete frankness 
and intimacy are possible only between siblings. Great as is the horror 
of incest, there are no avoidances between brother and sister. Ashanti 
are very particular about good manners and modesty of dress. It is 

' It is the piesie whose birth transforms the woman into a parent (obaatan) 
and his or her position in the sibling group is unique for that reason. It is of 
interest that it is taboo for a piesie to have any contact with the medicine set up 
for twins. This accentuates his unique social position, the antithesis to which is 
a place in the sibling group that is shared. 



only between siblings that easy informality of manners and dress is 
allowed. Brothers and sisters may joke with each other, tease each other, 
and even quarrel, as they are prone to do in childhood, without causing 
offence.' This accent on equality is not found in any other kinship 
relationship. It is felt to be immodest for adults of different generations 
to bath together ; siblings of the same sex may do so. The attachment 
and mutual identification of sisters is notorious. Two sisters (in this 
case all women of the same generation in the lineage) cannot be the 
wives of one husband at the same time or in succession. They would 
quarrel incessantly — and so involve their immediate maternal kin in 
quarrels — because they cannot be kept separate in their relations with 
their husband. Sisters try to live together all their lives. A woman 
treats her sister's children so much like her own that orphan children 
often do not know whether their apparent mother is their true mother 
or her sister. This holds, though to a lesser degree (because of the legal 
inferiority of paternal ties), of brothers. Siblings think of one another's 
possessions as joint property. Borrowing between siblings cannot create 
debts^ and, as we know, a sibling is a person's first heir. 

The social identification and solidarity of siblings are particularly 
important and explicit in the relations of opposite sex siblings. Ashanti 
invariably contrast them with the relationships of husband and wife 
and of parent and child. The essential nature of the sibling tie lies, to 
most Ashanti, in its antithesis to conjugal ties. This is a bitter subject 
to many, both among the educated and among the illiterate folk ; but it 
draws attention to the fact that the sibling tie shows its strength primarily 
in opposition to other ties, and in virtue of the constraint inherent in 
the rule of matrilineal descent. The pivot of the Ashanti kinship system 
in its function as a system of jural relations is the tie between brother 
and sister. A brother has legal power over his sister's children because 
he is her nearest male equivalent and legal power is vested in males. 
A sister has claims on her brother because she is his female equivalent 
and the only source of the continuity of his descent line. It is typical 
of the conflicts engendered by Ashanti kinship that men find it difficult 
to decide what is more important to them, to have children or for their 
sisters to have children. But after discussion most men conclude that 

' Conventionalized joking is not customary in any Ashanti kinship relation- 
ship. The nearest approach to a joking relationship is the privileged familiarity 
that obtains between Ashanti and some neighbouring Akan peoples like the 

^ This applies also to all lineage kin, by contrast with other kin ; but siblings 
(including children of full sisters) may borrow one another's personal goods 
without asking permission ; other kin, except parents, may not. The borrowing 
and lending of money is in a class of its own. It is a common complaint that 
irrecoverable 'loans' cannot be withheld from lineage kin, but this complaint 
is seldom made of siblings. A brother or sister must be helped even if the 
circumstances of his need are blameworthy. 


sad as it may be to die childless, a good citizen's first anxiety is for his 
lineage to survive. That is why cohabitation without the legal formalities 
of marriage is not stigmatized. 

Quoting their own experiences, men say that it is to his sister that 
a man entrusts weighty matters, never to his wife. He will discuss 
confidential matters, such as those that concern property, money, 
public office, legal suits, and even the future of his children or his 
matrimonial difficulties with his sister, secure in the knowledge that 
she will tell nobody else. He will give his valuables into her care, not 
into his wife's. He will use her as go-between with a secret lover, know- 
ing that she will never betray him to his wife. His sister is the appropriate 
person to fetch a man's bride home to him, and so a sister is the best 
watch-dog of a wife's fidelity. Women, again, agree that in a crisis they 
will side with their brothers against their husbands. There is often 
jealousy between a man's sister and his wife because each is thinking of 
what he can be made to do for her children. That is why they cannot 
easily live in the same house. Divorce after many years of marriage is 
common, and is said to be due very often to the conflict between 
loyalties towards spouse and towards sibling. 

But the basis of constraint in the sibling relationship is often felt to 
be irksome. There is much suppressed hostility between siblings. It is 
seen in the watchfulness of chiefs lest their sisters, the queen mothers, 
should plot to displace them in favour of their sons. It follows inevitably 
from the conflict of paternal love and avuncular duty or between a 
woman's loyalty as a sister and her dependence on her husband for 
sexual satisfaction and for the fulfilment of her purpose of motherhood. 
The high incidence of adultery among women is a consequence of this. 
It is no new thing, as the elaborate scale of adultery damages, which 
forms part of the traditional legal code, shows.' 

Ashanti custom provides no collectively approved means of giving 
expression to the underlying hostility in the sibling relationship. It is 
rightly felt to be an inevitable result of matrilineal descent. It is therefore 
merged in the other suppressed hostilities aroused by these obligations. 
The accepted expression for these is the belief that witchcraft acts only 
within the lineage. Accusations of witchcraft are everyday occurrences 
in Ashanti, and their volume is increasing as claims based on lineage 
ties come to be felt as more and more onerous. Side by side with this is 
found a rapidly growing addiction to cults purporting to give protection 
against and to detect witches. Illness, death, barrenness, economic loss, 
and other misfortunes are often ascribed to witchcraft, and those 
accused are most often close matrilineal kin of the sufferer, especially 
a mother or sister. 

Ashanti draw attention to the constraint inherent in the relations of 
maternal siblings by contrasting with them the relations of paternal 
* Cf. Rattray, 1927, chap. viii. 


half-siblings. Half-siblings are spoken of and addressed by the same 
kinship terms as full siblings." They rarely grow up in the same house- 
hold. But Ashanti say that one's father's child {agya ba) by another 
wife than one's mother is often one's most trusted and loved friend. It is 
a matter of one's own preferences. One has no binding obligations to 
him (or her) or to his (or her) children. It is purely a bond of sentiment 
based upon the feelings both parties have for their father;^ thus it is 
a trusted half-sibling to whom a person in need of honest and dis- 
interested counsel will go. 


The grandparents {nana, reciprocal nana) on both sides are the most 
honoured of all one's kinsfolk. Their position and status are of very 
great importance in the social system. In Ashanti it is the grandparents 
who are the prototypes of persons and institutions commanding 
reverence and submission to the norms of tradition. 

In terms of personal relations there is no distinction between paternal 
and maternal grandparents. In accordance with the principle of the 
identification of siblings, the terminology used and the correlated 
norms of behaviour are extended also to the siblings of the grandparents, 
and, by the lineage principle, to the lineage kin of the grandparents who 
are of different lineage from their grandchildren. 

The maternal grandmother holds a special position as she is often the 
female head of the domestic group and this gives her great influence in 
the bringing up of children. She is the guardian of morals and of 
harmony in the household. 

The young children are her special care ; but the paternal grandmother 
also has this role. Grandparents (and their siblings) lavish affection on 
their grandchildren, who are their greatest pride. Grandparents can 
reprimand and punish their grandchildren for minor acts of disobedience 
or impropriety. In cases of serious misbehaviour they must call on the 
parents to take disciplinary action. It is from the grandparents of both 
sexes that children learn family history, folk-lore, proverbs, and other 
traditional lore. The grandparents are felt to be the living links with the 
past. They are looked up to with reverence, not only as the repositories 
of ancient wisdom but also as symbols of the continuity of descent. 
This applies particularly to the mother's mother's brother who is 
commonly the head of the lineage segment.^ 

' Indeed, the terms for sibling (niia), brother (nua baruna), and sister {jiua 
bad) are freely used in speaking of or to any kinsman on the mother's side or the 
father's side of one's own generation. All Ashanti kinship terms are used in this 
wide way. 

^ This is the present-day version. Old people still point out that there is the 
bond of common kra washing and the taboos that go with it. 

^ A full discussion of the status of grandparents would take us beyond the 
scope of the present essay. A few points, however, deserve to be noted. It is 


The Father's Sister 

The father's sister (sewa, Ht. female father) receives the respect and 
often the affection due to a father, but there is not the deep attachment 
and trust felt for the father and his brothers. She describes her brother's 
child as 'my child {ba)' and she may scold but may not punish him. He 
feels at home in her house, and readily asks for and is given food or 
sleeping accommodation ; but there are no binding obligations on either 
side. A sewa has no authority over or legal claim upon her brother's 
child. There is often an element of strain in the relationship owing to 
the actual or supposed influence of the sewa on her brother in the 
interests of her own children. She stands for the claims of his lineage on 
him, especially those of his potential heirs which conflict with those of 
his children. The affective distance and the stress on deference rather 
than on filial trust are increased through the position of the father's 
sister as potential mother-in-law by cross-cousin marriage. Formerly 
a sister usually acted for her brother in the ritual of naming his child,' 
but this is falling into disuse with the obsolescence of the name-giving 
ritual and the increasing emphasis on lineal relationships at the expense 
of collateral relationships. 

As has been indicated before, all Ashanti kinship terms are used with 
a wide range of reference and there are often alternative terms for any 
one relationship.^ Individuals are connected in more than one way 
owing to the high incidence of marriage within a community. The 
general rule that where there is kinship there is a bond of amity and 
goodwill, which may be invoked if necessary, applies to all genealogical 
connexions. And where circumstances lead individuals to wish to show 

of great significance that the term for grandparent (nana) is also the title of 
respect used by a subject to a chief, and, indeed, is the general term of address 
or reference for a chief. It is used, similarly, to show respect to any person of 
rank, to old people, to the dead, to gods and fetishes, to priests and priestesses, 
&c. The reciprocal term (nana) is used by chiefs, elders, priests, &c., in addres- 
sing subjects or followers or suppliants to whom special consideration is being 
shown. The status of grandparents explains the importance so often attached 
to kinship within four generations. In brief, four generations is about the 
maximum extension of a descent line over which all the progeny of a man or 
a woman can remain united either under his or her authority (as sometimes 
happens in a matrilineal household) or through living contact with him or her. 
More usually four adult generations (the equivalent of five generations inclusive 
of children) embrace people of common descent strongly united by kinship 
sentiment focused on a progenitor or progenetrix who was the parent or grand- 
parent of its oldest members. The 'grandparents' are decisive in this chain 
of living contact as they form the bridge between their grandparents and 
'their grandchildren. By the sixth generation, except where the lineage principle 
is utilized to maintain the cohesion of the descent group, the chain is broken 
and kinship sentiment is no longer based on living contact. A great-grandparent 
is felt to be so near the ancestors that he or she is regarded with almost religious 

' See Rattray, 1927, pp. 62 ff. ^ Ibid., 1923, chap. i. 


special goodwill to or demand special consideration from each other, 
they find justification for this in kinship ties. As has been mentioned, 
on the principle of lineage unity any member of one's father's lineage 
may be addressed or described as father {agya, papa, se), or father's 
sister (sewa) if the filial relationship is stressed ; or if the person con- 
cerned is of one's own generation and this is stressed, the term for 
sibling (nua) can be used. This rule applies to the lineage and genera- 
tion alinement of all the kinsfolk of one's kin. But the jural and affective 
relationships associated with the use of these terms in their primary 
contexts are not automatically extended in this way. I may, half 
jestingly, address a daughter of a distant lineage brother of my mother 
as 'wife' (yere), but that does not mean I have a special claim to marry 
her. If, however, I court her successfully, he will probably explain his 
readiness to consent to the match on the grounds that I am his 'sister's 


The marriage customs of the Ashanti have been described in detail 
by Rattray.' In rural areas there has been little change in these customs, 
though greater latitude is nowadays allowed in their observance. The 
most important change has been the substitution of money payments 
for most of the gifts in kind customary in former days. The same rules 
apply to all ranks and classes, the only difference being that girls of 
chiefs' and high-ranking councillors' lineages are espoused with larger 
payments, graded according to their rank, than are commoners. Most 
educated men still marry in accordance with native custom.^ Respect- 
able citizens complain that illicit and casual unions are becoming a 
serious problem, and the older people blame disregard of parents' 
advice in the choice of a spouse for the notorious frequency of divorce. 
While there is evidence to support the view that shortlived illicit unions 
are increasing, especially in the towns, the data collected during the 
Ashanti Social Survey in 1945 indicate that a high divorce rate has 
been characteristic of the Ashanti for the past twenty to thirty years. ^ 

A girl may be betrothed in childhood (asiwa), but cannot be taken 
in marriage until after her puberty ceremony. Girls marry between 16 
and 18 years of age, youths between 20 and 25. 

Ashanti marriage is restricted in accordance with the Ashanti concept 
of incest. Marriage is prohibited between individuals who would be 
committing the sin of incest if they had sexual relations. Thus marriage 
is forbidden between members of the same matrilineage, and this 

' Rattray, 1923, 1927, ig2g passim, especially 1929, chap. iv. 

^ Church marriages and marriages in accordance with Colonial Civil Law 
are confined to a very small section of educated people. 

^ These data show that rather more than half of all men over 40 and about the 
same proportion of all women over 35 have been divorced at least once. The 
incidence is of course less with younger people. 


includes accessory lineages which have been attached to the authentic 
lineage for four or more generations. This taboo extends to the clan 
throughout Ashanti.' Marriage is also prohibited with any 'patrilineal' 
descendant of one's father's father's father — i.e. of one's father's 
paternal nana up to the fourth generation. This taboo is an extension 
of the taboo on sex relations with a father's brother's child (a sibling 
{nua) in Ashanti kinship terminology), which is regarded as a sin almost 
as heinous as incest with a full sibling. The basis of this is the identifica- 
tion of brothers. Thus the taboo applies to the children and sons' 
children of half-brothers by the same mother but different fathers. 
Lastly marriage with a lineal descendant to the fourth generation, or with 
a lineal ascendant (or the sibling of such an ascendant) to the fourth 
generation, is prohibited.- The ideal match is with a cross-cousin on 
either side, though there is a preference, for men, for marriage with a 
mother's brother's daughter {wofa bay — i.e. the daughter of a man of 
his mother's generation belonging to the lineage segment descended 
from her great-grandmother. Again, marriage with a member of one's 
own village or chiefdom is preferred to marriage with an 'outsider' 
(ahoho). The reasons given are that the character and family background 
of the parties are known to their respective parents and this enables 
them to make a careful choice. It is also said that such a marriage enables 
the parents and other kin of both spouses to take a personal interest in 
and help in the care of their children. But undoubtedly the most im- 
portant factor is the strength of matrilineal ties in domestic organization. 
Intra-village marriage enables husband and wife to reside with their 
maternal kin if they wish to or have to. My investigations in 1945 
showed that 75 per cent, and upwards of married people have spouses 
from the same village or adjacent villages. Chiefs still have two matri- 
monial prerogatives recorded by Rattray. Twin girls (ntaa) who are not 
kin to the chief of their place of birth within forbidden degrees are 
regarded from birth as his future wives and cannot marry anyone else 
unless he relinquishes his prior rights. Secondly, every ranking chief 
has a number of stool wives (ayete). These wives are obligatorily pro- 
vided by various lineages of the chief's subjects ; and if a stool wife dies 
she must be replaced by another girl of her lineage."* 

A marriage may begin with a period of cohabitation approved by 
the parents of the couple and accepted as a proper marriage for all 

' And, where clan affiliation is known, throughout the Akan-speaking peoples. 

^ The prohibited marriages listed by Rattray, 1923, p. 37, all fall under these 
rules. As has been pointed out, a common ntoro is not by itself a bar to marriage. 
Any attempt to defy these rules is nowadays treated as a crime subject to a fine. 
It should be noted that the count of four generations is always exclusive of the 
generation of Ego. 

^ Cf. Rattray, 1927, chap. xxix. 

* This is the only form of 'sororate' practised in Ashanti. Cf. Rattray, 1929, 
p. 27. 


practical purposes. The man, in such a case, has no right to demand 
adultery damages if his 'wife' has a lover and there is nothing to prevent 
the couple from separating by mutual agreement. 

The decisive formality for the establishment of a legal marriage is the 
giving of the tiri 7isa (lit. head wine), which Ashanti describe as a thank- 
ing gift (aseda). Most commonly this consists of two bottles of gin or an 
agreed equivalent in cash. It is handed over on behalf of the husband 
by the head of his lineage to the head of the bride's lineage through her 
legal guardian in the presence of representatives of both groups. Half 
of the gift is sent by the head of the bride's lineage to her father, and the 
remainder is distributed amongst the representatives of the two 

The payment of tiri nsa gives the husband exclusive sexual rights 
over his wife and the legal paternity of all children born to her while the 
marriage lasts. It gives him, also, the right to essential domestic and 
economic services from her. He in turn is obliged to provide her and 
their children with food, clothing, and housing if she has none. He must 
give her sexual satisfaction and take care of her in illness, is responsible 
for debts she contracts,' and last but not least, must obtain her consent 
if he wishes to take an additional wife. These material rights and 
obligations are specifically recognized in Ashanti customary law. Chronic 
failure on the part of a spouse to fulfil any of them can constitute 
grounds for divorce which husband and wife have equal rights to 
demand. But the crucial right conferred on both spouses by the giving 
and accepting of tiri nsa is that of sexual satisfaction for procreative 
purposes. The husband thus acquires exclusive sexual rights over his 
wife. If she commits adultery he can, as we have seen, sue for compensa- 
tion from the adulterer and claim the paternity of a child conceived in 
adultery. He has no such rights over a woman with whom he is merely 
cohabiting. A married woman can divorce a husband who neglects her 
sexually or who can be proved to be impotent or sterile. In such a case 
no efforts are made to reconcile the parties. This is generally attempted 
if divorce is demanded on other grounds. The rule that a man must 
get his wife's consent if he wishes to take a second wife is due to the fact 
that he is in effect asking her to share her sexual rights with another 
woman. The recognition of matrilineal descent is responsible for this 
high degree of equality between spouses in Ashanti marriage law 
and custom. It allows the conjugal relationship to be envisaged as 
a bundle of separable rights and bonds rather than as a unitary all-or- 
none tie. 

Tiri nsa may be paid before or at any time after the couple begin to 

cohabit. In addition, the bride's father, mother, lineage head, and 

brothers are entitled to certain customary gifts from the bridegroom, 

but these may be waived. Offerings may have to be presented to gods 

' But if she finds treasure trove her lineage takes possession of it. 


(obosom) or medicines (suman) under whose protection the bride's 
^ parents have Hved. The bride herself usually receives a gift of money 
from her husband when she first goes to live with him. This occasion 
is celebrated by her cooking a sumptuous meal for him and his kinsfolk 
and fellow villagers for which she herself and her mother supply the 

Except for the tiri nsa, none of the customary gifts is returnable on 
divorce. For a divorce to be legal, tiri nsa must be returned through 
the same channels by which it was given. 

A further payment may be demanded from the husband at any time 
during the marriage. If there is urgent need of money in the lineage 
segment to which the wife belongs, her mother's brother, as her legal 
guardian, can, with the consent of the other members of the segment, 
ask the lineage head to demand tiri sika, or head money, from her hus- 
band. Any amount may be asked for. It is in reality a loan for an 
indefinite period which serves as a pledge of the fidelity of the wife. It is 
returnable only on the termination of the marriage by divorce or by 
death of the wife. The husband may, however, make a free gift of the 
tiri sika to his wife and her lineage after some years of marriage, on 
divorce, or if she dies. 

Polygyny is permitted, and chiefs often have large harems. It is, 
however, rare, nowadays, for commoners to have more than three wives 
at the same time ; 80 per cent, of all married men have only one wife at 
a time. It is notorious that co-wives often show great jealousy of one 
ainoiEirer. They call each other kora, 'jealous one', and the usual practice 
is for a polygynist's wives to live separately. A polygynist must be 
scrupulously fair in sharing his time, his sexual attentions, and the 
material provision he makes for them, equally among his wives. 

As has been mentioned, a cross-cousin is regarded as the most 
satisfactory spouse. At least this is still the view of the older people. 
Many young men take a diff'erent view. They say that a cross-cousin is 
'too near', almost a sister, and so she is never as attractive as an unrelated 
girl. Some argue, also, that cross-cousin marriage reinforces the 
authority of the mother's brother, who is now also father-in-law, to an 
intolerable degree. But in rural areas there are many young men who 
approve of marriage with a mother's brother's daughter (wofa ba) on 
the grounds that it creates an additional bond with their maternal uncles 
and that it is a more secure marriage than a match with an unrelated 
woman. Women often argue in favour of the custom. They say it 
strengthens their claims on their husbands and their children's claims 
on both paternal and maternal kin. The older people — parents, mothers' 
brothers, and fathers' sisters — in whose interests and at whose insistence 
cross-cousin marriages are arranged^ defend the custom on various 
grounds. The commonest argument is on the grounds of property and 
' As Rattray points out, 1927, chap. xxix. 


wealth.' Cross-cousin marriage most often occurs between the children 
of full siblings or of uterine first cousins. Such a marriage, it is con- 
tended, ensures that a man's daughter and her children derive some 
benefit from the property he is obliged to leave to his sister's son; or 
(where a man's son marries his sister's daughter) that his son's children 
get much benefit from his property. Importance is still attached, also, 
to the transmission of names, though not so much as when Rattray 
wrote. ^ Furthermore, parents praise cross-cousin marriage because 
there can be no mistake about the character or family background of 
one's child's spouse. They maintain, also, that such marriages are less 
liable to break up on frivolous grounds than marriage between non- 
kinsfolk. The kinship between all the parties concerned gives all of them 
a strong interest in smoothing out difficulties and supporting the mar- 
riage. But undoubtedly the strongest motive in favour of cross-cousin 
marriage derives from the conflicts latent in Ashanti kinship relations. 
Cross-cousin marriage offers a possibility of reconciling the conflicting 
influences of conjugal, parental, and matrilineal kinship ties. 

The divergent opinions about cross-cousin marriage suggest, how- 
ever, that there are strong tendencies working against it. In any case 
only a small proportion of all marriages can, at a given time, be cross- 
cousin marriages. My inquiries showed that this proportion has been 
decreasing over the past thirty to forty years, ^ and that cross-cousin 
marriages are not more stable than other marriages. This is in keeping 
with the general trend towards greater individual liberty of movement, 
of choice of occupation, and of selection of spouses due to new economic 
opportunities and new ideas derived from mission and school teaching. 
Chiefs and elders say that it is also one of the symptoms of declining 
parental and avuncular authority. But the chief factor working against 
cross-cousin marriage appears, to the outside observer, to be the 
increasing tendency to seek a solution for the tensions inherent in the 
Ashanti kinship system by segregating the field of conjugal and parental 
kinship from that of matrilineal kinship. 

' Rattray, 1927, chap. xxix. 

^ Ibid. The belief that children sometimes reincarnate recent ancestors still 
prevails among pagan Ashanti. The point about the names, however, is that 
cross-cousin marriage ensures that a man's paternal antecedents may become the 
namesakes of his nearest lineage kin of the next succeeding generation. This is 
greatly desired. 

^ In a sample of 525 women from various parts of Ashanti who were or had 
been married, approximately 8 per cent, had been married to cross-cousins. 
A rough breakdown by age-groups gave the following approximate results: 

Percentage of cross-cousin marriages 
Among women under 30, approx. 2 per cent. 
Among women aged 30-40, approx. 5 per cent. 
Among women aged 40-50, approx. 9 per cent. 
Among women aged 50 and over, approx. 14 per cent. 



^ The status of parents-in-law (ase — the generic term for all in-laws) 
is not very precisely defined in Ashanti. There are no specific forms of 
etiquette or obligation between them and their son's wife or daughter's 
husband. There is only the general rule that they must be treated with 
special respect,' and must be given assistance and support whenever 
they need any. This is especially important in the early years of marriage. 
A man should go out of his way to help his mother-in-law or father-in- 
law with farm work, in house-building, or in other ways, and should 
not neglect opportunities of giving them small presents. These usages 
apply also, in an attenuated degree, to siblings-in-law. A woman should 
show deference and obedience to her parents-in-law. But it is accepted 
that there is constant danger of friction with parents-in-law. Both 
husband and wife are apt to resent the strong hold of a spouse's parents 
(in particular of the mother) and siblings over him or her. Ashanti, 
therefore, do not seek frequent contacts with their parents- or siblings- 
in-law. When a divorce occurs the spouse who feels wronged is apt to 
blame his or her parents-in-law or siblings-in-law. Men, in particular, 
are prone to lay a wife's failings at the door of her mother. 

As has been mentioned before, there is a high incidence of divorce 
among Ashanti. This is due primarily to the strength of matrilineal 
Kinship ties. Divorce usually makes little change in the domestic 
circumstances of a woman or in her economic situation, nor does it 
affect her jural status or that of her children. Though it may involve 
personal distress, it carries no moral stigma and no social penalties. 


Kinship plays a very important part in Ashanti social life. It is not 
only the source of the critical norms governing the jural and personal 
relations of individuals in many fields of social life, but it also deter- 
mines the structure of the corporate groups on which the political 
organization is based and influences political relations at all levels. 
The dominant principle of Ashanti kinship is the rule of matrilineal 
descent. In spite of nearly forty years of rapid social change under the 
influence of European economic, social, and cultural agencies, the 
Ashanti have tenaciously upheld this rule. The chief problem of kinship 
relations among them is to adjust the jural and moral claims and bonds 
arising out of marriage and fatherhood to those imposed by matrilineal 
kinship. Conflict between these rival claims and bonds is inherent in 
their kinship system. This problem is most acute for men and they 
seek a solution to it along two lines. Some attempt to assimilate the 
conflicting kinship ties to one another, and so to circumvent the dilemma 
of a choice of loyalties. Most, however, try to solve the problem by 

' The proverb says Ase fie, yenko no basabasa — 'The parent-in-law's house, 
we do not enter it carelessly'. 



segregating the respective fields of the conflicting kinship ties. This 
method of dealing with the problem seems to be gaining ground under 
modern conditions. Though it appears to be an adequate method, it 
is undoubtedly uneconomical, imposing a heavy emotional and moral 
as well as material strain on the parties concerned. This is the crucial 
problem of contemporary Ashanti society. 



THE relations arising from parenthood extend in all societies to 
form a wider system of kinship whereby, both for individuals 
and groups, rights of inheritance and succession, and ties of 
mutual obligation, are established on accepted principles of 
descent. But parenthood is dual, and if recognition were symmetrically 
accorded through both parents at each generation, kinship ties would 
proliferate indefinitely in ever widening aggregates. Moreover, the sex 
differences among both parents and siblings are a pervasive factor in 
differentiating cultural activity and attendant social status. In the process 
of establishing coherent and continuing social relations between kin, 
endless variety is possible in the stress and limitations set on the recogni- 
tion of particular ties. The sex distinction may in some contexts be so 
dominant that women are grouped and succeed through their mothers, 
while men are grouped and succeed through their fathers.' But in the 
maintenance of groups of wide social relevance the sex difference between 
individuals is normally subordinated to the principle of affiliation of both 
sons and daughters through one of the parents to give rise to mutually 
exclusive kin groups of both sexes. ^ 

Among many peoples one line of descent, either that through fathers 
or that through mothers, is stressed and a single system of unilineal 
groups is formed. But the ascription of status within groups based on 
one unilineal reckoning does not exclude the concurrent establishment of 
rights and obligations, if only between close relatives, with kinsmen in 
other lines of descent. The unilineal tendency itself contains the alter- 
natives of patrilineal and matrilineal reckoning and these are not, as was 
once assumed, mutually exclusive. Both may be operative as principles 
of affiliation and group organization in distinct social fields, so that rights 
of succession and inheritance fall into distinct categories. 

Systems of double descent or double unilineal kin-group organization 
have long been encountered in Africa, but they have often been mis- 
understood and few of them have been closely studied.^ A people of 

' This tendency, which gives rise to distinct descent groups among, for 
example, the Ge tribes of the Matto Grosso in South America, is exemplified 
over limited fields of inheritance in Africa; see, e.g., the paper by Dr. S. F. 
Nadel, in this volume. 

^ For a discussion of the range of conditions under which this tendency finds 
expression in the formation of a stable system of mutually exclusive unilineal 
kin groups, see my paper, 'The Anthropological Approach in Social Science', 
in The Advancement oj Science, London, 1947. 

^ On the Herero system, which still awaits full analysis, see I. Schapera, Notes 
on some Herero Genealogies, University of Cape Town, 1945. 


importance in this connexion are the Yako, one of a number of 'semi- 
Bantu' speaking peoples' Hving some seventy miles north of Calabar on 
the eastern side of the middle Cross river in the Obubra Division of 
Ogoja Province, South-Eastern Nigeria (see map in Fig. 9). Their sys- 
tem affords an instance of full and simultaneous development of both 
patrilineal and matrilineal kin groups. Until the present generation, 
among whom many features are, as will be seen, being modified under 
the impact of Western institutions, matrilineal descent has been as 
outstanding in its sphere as patrilineal, and both are corporately 

The Yako number about 20,000 persons grouped in five compact 
village settlements, one of them very large, situated at distances of from 
three to six miles from each other. The dwelling-area of the largest 
village, Umor, seven miles to the east of the Cross river, extends over 
a quarter of a square mile and had a population of nearly 11,000 in 1935. 
Six miles to the north lies Ekuri, smaller than Umor ; the main settle- 
ment is again away from the river, although one of its wards has a water- 
side hamlet. Five and eight miles away to the east are two still smaller 
and closely neighbouring settlements of Nko and Ngkpani, while little 
more than three miles south of Umor is the fifth and smallest village 
of Idomi. A very high degree of linguistic and cultural homogeneity 
is maintained between the villages by a continual interchange of 
visitors, temporary residents, and permanent migrants. But there is 
no centralized political organization and sporadic, shortlived fighting 
between the villages, particularly between Umor and Ekuri, occurred 
in the past. 

Yako settlement has been comparatively stable. In the larger and 
traditionally older villages of Umor and Ekuri the house sites of impor- 
tant ancestors four generations back are pointed out and there are in 
places mounds several feet high said to have been formed by super- 
position of successive house floors. Of the smaller villages, Idomi and 
Ngkpani are held to have been established by migrations from Umor, 
the latter after a bitter quarrel between two Umor wards. But despite 
a rapid increase of population over at least the past two generations and 
the growth of Ekuri and particularly Umor into very large and crowded 
settlements, there is a very strong tendency for men to remain with their 
patrikin. When, as is increasingly common, dwelling areas in the village 
become fully occupied, the pressure is relieved by building hamlets 
(sing, kowu) on the nearest tract of farming land belonging to the group 

' Others of these peoples which have a similar dual system of kin grouping are 
the Ekumuru, Abayong, Agwa'aguna, Enna, Abini, Agoi, and Asiga. The 
Agwa'aguna and Yako share a common tradition of overland migration from 
an area further up the Cross river in what is now Ikom Division and trace their 
origin to Okuni. Among them both 'friendship' and social distance are expressed 
in taboos on fighting and intermarriage and in joking relations. 

ADUM ^^ 





K^ i 



Idoini '' 




• Nko Settlements 

■■••«* Yako tTibat "boundary 

Yako vtlla5c territories 

_^__ Boundaries of Tietghl?ourin5 

""""" triljcs 

••»••» Provincial "boundary 

^caU I I I . I I I— 

I I 1 I Allies 


Fig. 9 



and within half a mile or less of the village. The occupants of these 
hamlets continue to participate as fully as they can in the life of the main 
village. With the increase in size of settlements the areas appropriated 
for cultivation and the collection of forest products have been extended, 
but the overall density of population has not so far risen high enough to 
cause deleterious pressure on land. Even in the territory of Umor, where 
the density is highest, at about 230 per square mile, there are still some 
tracts of unfarmed land. 

The Yako are predominantly subsistence cultivators ; they collect and 
prepare oil-palm products for external trade, but take little part in the 
river traffic with Calabar and the lower Cross river. A wide range of bush 
products is collected for local use and petty trade, but hunting is now of 
minor importance. Interests are concentrated on the growing of yams 
and secondary crops which are cultivated jointly by men and women on" 
household farm plots. These plots are held individually by men asjiouse- 
hold heads within the framework of the wider collective rights of kin 
groups to be described later. But a household only cultivates in any one 
year about a fifth of the plots to which it has established rights. The rest 
is left in bush fallow and cultivated in rotation over a period of years.' 


The elementary family is the nucleus of the Yako household, but the 
majority of the older men have more than one wife. In polygynous 
households every wife has a right to occupy a separate house provided 
by her husband and has personal property in food-supplies and domestic 
equipment (see Fig. 10, Woman's House). She rears her own children 
and in both productive and recreational activities is in principle free 
from the interference of any other wife. Although co-wives normally 
occupy adjacent houses there are, from the point of view of daily food- 
supply, as many households as there are adult able-bodied women. The 
single household head has equal rights and obHgations in each. Food 
should be prepared for the husband by each wife in turn, but the length 
and regularity of the periods appear to be very variable. But in farming, 
a man with his several wives normally constitutes a single farming unit, 
although each wife has her specific interests, rights, and duties which do 
not extend over the entire farm plot, and one wife does not control the 
farm labour of the others. The close proximity of the houses of plural 
wives and their common concern with the aff'airs of one husband involve^ 
personal relations between the wives, but these are informal in character. 
They depend on individual temperament and vary according to the 
particular circumstances. Attitudes are not determined in advance either 

' For an account of the settlement pattern, agricultural cycle, and land 
utilization, see my paper 'Land and Labour in a Cross River Village, Southern 
Nigeria', Geogr.J. xc. i, July 1937, pp. 24-57. 


by rules of domestic co-operation or by the relative status of wives. The 
first married or the senior existing wife has no formal superiority or 
prerogative. Yako say that each wife has an equal claim on the time, 
attention, and energy of the husband. Similarly each wife has an equal 
obligation to participate in household duties, including especially farm 
work and preparing food for the husband, and is responsible for the 
w^ell-being of her own children. In practice, owing to differences in the 
aptitudes and inclinations of wives and in the sentiments of the husband 
towards them, there may be considerable differences among them both 
in their participation in domestic activities and in their prestige in the 
compoiihd. But such differences are not subject to the control of the 
senior wife. On the farm it is the husband who tells each wife what he 
wishes her to do. Mutual helpfulness between wives in farming and 
domestic activities is by no means absent, but at the same time it is quite 
common to find an age-mate of one wife helping her in some task at 
home or on the farm while the other wife is working independently. The 
relations between wives of one man may range from real companionship 
to a minimum of contact punctuated with outbursts of hostility. The 
failure of a wife to live amicably with another in the compound is 
a recognized ground for dismissing the aggressor. 

The yams of a man and those of his wife or wives are planted and 
harvested separately. They are tied on distinct groups of uprights in the 
husband's harvest stack and the separate harvests of each are used in 
turn to provide household supplies according to w'ell-established rules. 
The lesser crops are the wives' responsibility. Each owns and controls 
her own harvests, but she is at the same time under obligation to provide 
her husband and children with adequate supplies during the year. A wife 
whg^ for any reason, is unable to carry out all the farm work that falls to 
her lot should find and recompense helpers from among her age-mates 
or her kin. If she lacks sufficient yams to supply her share for the needs 
of her husband and children she is expected to work on other people's 
farms at busy times, and especially at planting and harvesting, when she 
will be given yams for each day's work which will help to remedy her 
own deficit. But a husband with a large harvest of yams will himself give 
a wife extra yams in recognition of her farming services to him if her own 
crop is short. 

On the other hand, fruit-trees and oil-palms are not held by women. 
They are planted or tended individually by men, and rights to them are 
transmitted to male heirs. A husband is expected to provide his wife 
with supplies of palm-nuts for the preparation of kernels which she sells, 
and failure to do so is a very common source of friction between them. 
This Ke"commonly does in the course of his collection of palm-fruit for 
the preparation of palm-oil, the marketable surplus of which belongs to 
him. There is a division of labour between men and women in the suc- 
cessive processes, the wife normally receiving the nuts for cracking and 



disposal of the kernels as part of her share in the product and her reward 
for assisting in the production of the oil.' 

There is no limitation, apart from inclination and the resources he 
commands, on the number of wives that a man may have at any one time, 
and demographic conditions, at least in the recent past, have been 
favourable to a high incidence of polygyny/ Plural marriage is an 
undoubted advantage in maintaining a large farm every year, for 
weeding the farm, cultivating secondary crops and carrying the yam 
harvest, and also in the preparation of oil-palm products for sale. Men 
frequently explain their later marriages by saying that increase in their 
yam harvests or the death of one wife made it necessary for them to seek 
another without delay. But plural marriage is also valued as a means of 
rearing a larger number of children and so of increasing the strength of 
a man's patrilineage and of his own prestige within it. On the other hand, 
the increasing adoption of trading pursuits by younger men who, in 
consequence, can clear only small farms, is rendering polygyny difficult 
or expensive for them. Such men also tend to be less strongly concerned 
in the traditionally high value placed on numerous offspring. Moreover, 
there is little emphasis on the number of wives as a direct expression of 
an individual's status or prestige. The priest leaders of the clans, both 
patrilineal and matrilineal, for example, are not distinguished by a 
markedly higher degree of polygyny than the adult male population in 

A man takes further wives according to his needs and opportunities 
throughout his earlier life, but as old age approaches he does not marry 
again. Elderly widowers, although they maintain their ov/n farms, often 
do not remarry late in life, but are dependent on the aid of daughters or 
sons' wives for farm work. There is thus no tendency among the Yako 
for old men as a class to secure a disproportionate number of women as, 
wives at the expense of younger men. 

Young men usually marry girls of their own age in the first instance. 
On the other hand, in later marriages men often take women who differ 
widely from them in age, and the general tendency is to marry younger 
women. There is a particular inducement to take a second wife very 
shortly after the first if the latter has borne a child, for a child is normally 
suckled for two years during which intercourse between the parents is 
forbidden. In such a situation the second wife is often a previously 
unmarried girl. 

The sex ratios at birth and at marriageable age appear to be normal 

' Wives retain for their personal use any oil left over after filling the 4-galIon 
tins used as units in oil trading. Any surplus less than half a tinfull the wife 
claims for domestic use and for disposal in small occasional gifts to kin, affines, 
and neighbours with whom she is on good terms. 

^ See my Marriage and the Family among the Yako in South-Eastern Nigeria, 
Monographs on Social Anthropology, No. 5, London, 1941. 


among the Yako. It is therefore likely that the general practice of 
polygynous marriage has been made possible only through the purchase 
of foreign children, mainly from Ibo country, west of the Cross river, 
who are known as yafoli, and are adopted into the household and kin 
groups of the purchaser, and the great majority of whom have been girls. 
There has recently been increasing governmental check on this traffick- 
ing in children, but it has for a considerable period made available a 
surplus of girls especially for marriages to older men as later wives. 

Divorg e as well as polygyny is frequent among the Yako and both 
produce a situation in which a man's children fairly often consist of 
several groups which are half-siblings to one another. All will be 
members of the father's patrilineage, but their matrilineal affiliations and 
Qonsiderable rights and obligations associated with them will be diver- 
gent. A man is not usually concerned, for example, in his paternal half- 
sister's marriage ; her children are not among his heirs. But a full sister 
and a maternal half-sister are equally among his closest matrilineal kin. 

The household of an older man will often include not only his wives 
but also divorced or widowed kinswomen, among whom mother, sisters, 
^fenghters, and sisters' daughters are the most frequently encountered. 
A woman widowed late in life usually joins one of her sons in whose 
farm she will plant her crops. A younger woman who has been widowed 
or divorced may temporarily rejoin her parents or go to her mother's 
brother or to an older brother. Such women are given houses of their own, 
but, urileFs they are infirm, they provide their own food-supplies and 
assist the head of the household. Their rights and obligations are in 
many respects similar to those of the wives, and any young children with 
them will be treated like foster-children. Elderly widowers are also to be 
found as dependants in the households of sons or other patrikin. There 
are, too, injnany Yako households young children of previous marriages 
related to only one of the spouses. 

~ Thenfn^ofityof Kousehold heads provide themselves with a separate 
m an's ho use, usually a smaller and simpler dwelling than a woman's, in 
which they keep their personal belongings, meet their friends, and often 
sleep. Young men before and after marriage frequently share the house 
of a father, elder brother, or other male relative living in the same com- 
po und .__(See Fig. 10, Man's House, and Fig. 11.) 

Yako feel that physiological paternity constitutes a right to social 
fatherhood which, where they conflict, may be asserted against the rights 
derived through marriage. Although, as will be seen, social fatherhood 
may be later attenuated if not extinguished following adoption by a 
foster-father and his lineage, a man's physiological paternity of his wife's 
child is generally axiomatic, for there is not usually any knowledge or 
assumption to the contrary. But rights to social fatherhood based on 
physiological paternity outside marriage can in another context be 
successfully asserted over those derived from marriage to the mother. In 

K t Y 

/^///// mud and stick. waIL5 

E:;:::;:::;::a:::i firf fixs 
I.I 'i >.i|i> 

, t) P 2 J ■» i. 


A r: 

Section A-B.with Jistaaw.facinj C. superimposed in )>ToVjin lines 

'Letoma (verandan) ,1,1 M i,.^— --~:^- — - — ^ {firewouii stacM 
■ h'lll ',': ^1 III '^---^= ^^ 



(■pmuoof of matting; 
for storing carjy yams) 

froof frame 

Epilo 11 


£ . ^ H (short b^d ) 

^^ — " raissdS 


■ ,^ Kinakwo 
sQ': ( fireplace^ 

'"1^^—^ Pl S<?' t pilo TvLinakuwa 


[main Floor) 


positwn ofridjep^ 5 

Ke toKpoy um o 
( bed hn^lyin^) 
2'2" raised I'l" 

,, ,,,, i| I I Mil M ,f 


Hona bid) 
raised 3' 

^^i/'ii'j'ii' rii"M i''iij'i ''"' ''' ' '' '' '■^'■ 

^ LC -'-/i .u,''J^w-A^ I vi^-^ ' -^^-^ 

Fig. io 

g IViL 

( fiff pit 

for bed ) 


Other words, an adulterer can in some situations successfully claim the 
social fatherhood of his child. It is said that claims which would give 
rise to such a conflict are not frequently made, but they are held to be 
justified where a woman has left her husband and has a liaison with 
another man, before any divorce has been arranged by repayment of 
marriage money. The rights of the husband and his patrikin are then 
confined to compensation for adultery and a return of the marriage 
money, but do not include fatherhood of the child or its membership of 
his patrician. And this has been held to be so even when the woman 
eventually returns to her husband.' 


Yako dwellings are usually built round small four-sided compounds 
{akamsunga, sing, lekdtnsunga) on to which face from five to ten separate 
women's and men's houses. A compound usually contains the houses of 
several household heads, but the latter, as will be seen, are nearly always 
close patrilineal kin and despite the crowding of houses in the parts of 
the villages that have been long occupied, an effort is made to extend the 
compound and maintain the unity of the related households if they 
increase in number. Thus a Yako child brought up in the house of its 
mother is one of a considerable group of children of various ages. Most 
of them will be linked in patrilineal kinship both among themselves and 
to the men of the compound, although, as will appear later, foster- 
parentage, and the fact that young children usually accompany their 
mother if she leaves her husband's compound on divorce, introduce 
some who are connected by other ties. More important at this stage is 
the fact that a child grows up in the company of age-mates from different 
households and, as it grows up, takes an increasing part in the activities 
of the compound as a whole. A boy goes out with his father and with his 
father's brothers from an early age when they are clearing bush or 
collecting palm products. Unless relations between his own parents and 
others happen to be strained he is made to run messages or take food 
out to the farm during busy seasons for any grown-up of the compound. 
Later he becomes a regular helper in work on his father's farm and in the 

' In one case, which came for decision to the village head in Umor, a woman 
left her husband, went to live under the protection of a matrikinsman, and had 
a liaison with one of the latter's neighbouring patrikin, giving birth to a son. 
Her husband, wanting her to return, had not brought a charge of adultery or 
claimed return of marriage money and a year later she did return to him. The 
lover then claimed the child before the village head, who, with some of the 
council, held that the child belonged to him and his clan to which it should later 
return. The lover was told he should compensate the husband for the adultery, 
but more stress was laid on their living in peace, and no compensation was in 
fact made. Nevertheless at about 12 years of age the boy was brought back by 
the lover to his patrician and has grown up there as his son and a member of his 


co-operative working parties of kinsmen and others in which his father 
joins. A girl accompanies her mother to the farm or looks after younger 
children in the compound when the older women have to be away 
farming or at the market. When a young man reaches 17 or 1 8 years and 
payments have been made for his first marriage, he is given his first farm 
plot and help in building a house by his father and other patrilineal kin 
in the compound (see Fig. iia). 

But the compound in which a man is born and grows up is only one 
segment of a larger settlement group of patrilineal relatives and their 
wives. It is this larger group, normally a cluster of several adjacent 
compounds, which collectively maintains rights to a section of the village 
site and to tracts of farm-land in the surrounding country. In one house 
area examined in Umor, nineteen out of a total of thirty-two men, whose 
fathers were still living, lived in the same compound or house cluster as 
their fathers. Of the nine adjacent compounds in this dwelling-area, four 
were almost exclusively occupied by men who were close patrilineal 
relatives. Each of these sets of close patrikin neighbours, together with 
a few individuals living elsewhere, constitutes a self-conscious group 
known as an eponama (pi. yeponama), a term derived from the Yako 
word for the urethra and thus stressing the biological link through 
males. The yeponama are corporate patrilineages whose members trace 
descent through birth or adoption from a common ancestor three to five 
generations back, from whom they name themselves as a corporate 
group, e.g. Eponama Etung Enamuzo. They will vary in strength from 
half a dozen to thirty adult men. If a lineage becomes much larger than 
this, one or more branches tend to drift apart and their ancestors of three 
or four generations back become the point of reference for the segmen- 

The patrilineage is thus a small group of kinsmen with, as will be 
seen, their adopted adherents, the internal coherence of which lasts only 
so long as the group remains small and intimate; in conditions of 
increasing population it is likely to persist as a stable and undivided 
group for only three or four generations. The patrilineage rather than the 
single compound is the most important corporate group beyond the 
household of which a Yako is conscious. Its members claim and distri- 
bute among themselves succession to rights in house sites, farm plots, 
oil-palm clusters, and planted trees that are not made use of by actual 
sons. A senior man of standing, referred to as uwo womon (our father), 
arbitrates in disputes among them and is their leader in affairs with other 
groups. And they may also speak of themselves and be referred to as his 
people, e.g. Yanen bi Ikpi Esua. 

But the patrilineages are aggregated into larger groups or patricians 
which are also territorially compact. These larger groups are known as 
yepun (sing, kepun). Each has a recognized and even demarcated 
dwelling-area {lekdm, pi. aksm), in which live the great majority of its 



^ <5> 

. ^ 

























• un« 

^ i^ 

II II II »• 

(a) Diagram of four adjacent compounds in Egbisum Kepun, Umor. 

(b) Lineage of Ikpi Esua (Obot Kepun) to show affiliations of occupants. 

Numbers refer to houses in the diagram of compounds. All living adult males 
of the lineages are shown. Those without numbers live elsewhere. 


men with their wives and children. Each has a name which is used to 
refer both to the members and to the dwelHng-aren. These names do not 
refer to ancestors, actual or mythical. Some refer to natural and other 
features suggesting that conditions connected with the site led to the 
naming. Loseni, for instance, relates to the name of a tree. But several 
yepun in different wards have the same name.' 

The boundary between the dwelling-areas of two patricians is often 
marked by a narrow gap, a pathway, a shallow ditch, or even a row of 
posts, and the adjacent houses of the two groups are built back to back 
on either side of the boundary, emphasizing by their orientation the 
separateness of the adjacent groups (see Fig. 12). 

Within the dwelling-area there is a large open-sided meeting-house 
(lepema) built of massive timbers where kepun feasts are held. Here, 
rather than in their houses, men spend their leisure hours and to it they 
often bring some task they have in hand. Here, as well as in his own 
compound, the death of a man is ceremonially recognized when his 
name is drummed on the kepun gong. The patrician house is not ritually 
barred to women, but they have no concern with it. Near the meeting- 
house is the shrine of the kepun (epundet, pi. yepundet), a low mound of 
small boulders surmovmted by some chalk-stained pots, usually set in 
the shade of a tree. The epundet, and the recurrent rites performed there, 
symbolize for everyone from childhood the corporate group. 

But the coherence of the kepun is not conceived by the Yako in terms 
of contiguity alone. The kepun membership of every person is normally 
established at birth, for a child born in the village is automatically 
recognized as a member of its father's kepun, and continues to be so 
regarded unless special circumstances result in a de facto transfer of 
membership later in life. Although, as will be seen, individuals may be 
adopted into the group either as children or even when adult, the 
majority of the members of a kepun at any one time belong to it in virtue 
of patrilineal descent. 

On the other hand, the kepun is not a single lineage of wider span than 
the eponarna in the sense that all members are held to be descended from 
a single ancestor to whom genealogical connexion is traced or even 
ascribed. Yako may and do say of their kepun fellows, 'We are all 
brothers', and imply common descent in this way, but in other contexts, 
particularly in disputes over dwelling-sites or prior claims to the use of 
tracts of farm-land, the separateness of the lineages and their alleged 
distinct origins are emphasized. Moreover, while, in response to genea- 
logical inquiry, the named ancestors of two lineages are occasionally 
held to be siblings and their father's name is given, in the majority of 
cases ignorance or doubt is expressed concerning any patrilineal kinship 

' Many kepun names incorporate the Yako term for dwelling-area, e.g. letekam, 
place of smithing; lekpafigkdm, place of kekpan (= grouping). Others, however, 
e.g. ibenda, unebu, &c., omit this combination. 





5 z <: - 



go :2 1 







^ -g Z ^ g c 
a &, J 6. ^ E > 

•S.I'- "^^S^S^^ 

o - 

, / 




?■? - ^ -1 -§ ■= 3 

o _ 


\ '^^X 

§ E 




\ <p 

i2 Z 



111 i^ III 3 

o - 



-..^^^ i; 



i B < 1 1 i . 



>3 \ 



\ -^ \ 


o - 


\ -^ 




\ '^ 





* "-••'"■ \ 


o - 





-to „ i _.-\ 

/^ \ 

o — ' 



x- \ 
V ^ \ 










— -7 

^**\ ^ 











\ "^ 


< V 



y V r a vsn 







v- ■J. 

\ D / O 




w n n 

■M V \ 

/^ 1 ■ 

"^ /I'^^jrf *" 







/ +/ ^ 





/«\ h- 


z ^ 

/l < 

K. I 

■ o W "> / > 


/'3 \ 




2: % 



/ '^-''"iV 

/ N^ 









'1 •■■ -> \ 

^- z 


/ 1 



/ "^ ^ 


\ < 


/ i 




/ ,/ 


^ / 

=» /o 


/ 5 



U/ ^ \ 

"^ X ( '3 



/ •^ 


-" / 




/ °/ 


' to / 




„ 1 1 
>" -g c ^ E 

- c E 5 1= 

(fi o S S. ;>. •< UJ 
• + 6 o < > 


between the founders of lineages in a kepun, and in some a definite 
tradition of arrival as a stranger accounts for the ancestor's existence. 
The sociologically important point is that the separateness of the lineages 
within the kepun is in these ways frequently emphasized. 

The patricians are, nevertheless, strictly exogamous and the rule is 
enforced with no great difficulty, since post-marital residence is patri- 
local. At marriage a woman leaves the compound and kepun of her birth 
to join her husband, usually elsewhere in the same village. Thus, apart 
from returned widowed or divorced daughters, the women living in 
a patrician area {lekam) are wives and not members of the kepun kin. 
The women who are kepun kin are scattered, according to their marriages, 
all over the village and to a lesser extent in other villages. 

/ The Yako kepun may therefore be summarily described as an exo- 
gamous, corporate, and territorially compact patrician composed of 
a number of separate lineages with collective rights to a delimited 
dwelling-area in the village and to tracts of farm- land in the territory, 
possessing a shrine and an assembly house for group rites and social 
intercourse. Priority in rights to particular tracts of house and farm-land 
are claimed by the various component lineages. A Yako thinks of a kepun 
in terms of all these attributes, but social groups are plastic, liable to 
departures from the recognized norm and subject to growth, decay, and 
the impact of other social institutions. Increase in numbers, the develop- 
ment of friction within groups, the concomitant recognition of matri- 
lineal ties, opportunities for house-building on new sites, and for 
farming in new directions have all operated to complicate the formal 


The data summarized for the village of Umor in Table I and on the 
map (Fig. 12) and diagram (Fig. 13) show how the actual yepun often 
diverge from the ideal pattern. If the groups are defined in terms of the 
named dwelling-areas alone, we find twenty-six in all. But there are five 
areas in which there are two shrines (each with its own priest), and two 
assembly houses are usually maintained in these areas. In each there are 
two sets of lineages with independent kepun rituals forming separate 
exogamous groups, for there is no bar to marriage between members of 
the two groups ; in some cases they unite in farm activities and recognize 
a single leader for each of the paths to their farm tracts, others do not. 
On the other hand, there are eight named dwelling-areas the occupants 
of which have their own elders, farm-road leaders, and assembly house, 
but which contain no shrine. The members of such groups belong to 
a larger ritual group and the shrine with its priest is found elsewhere. 
In all such cases the two separately named groups form a single exo- 
gamous unit and have joint rights to tracts of land. 

In Yako thought and practice patriclanship and exogamy apply to all 
persons who supplicate at a single shrine (epundet), so that there are in 
Umor, on the one hand, nominally distinct residential groups, which 


Table I 
Patricians ( Yepun) of Umor 

Showing linked groups, sites of yepun shrines, and number of component 
lineages (yeponama). 



1 . ( Lebokem* 

2. \ Lekpangkem I 

• 3 






( -Lekpangkem* 
\ Lewangkem . 

3. f Lekpangkem II 

4. \ Ugom I . 




: I] 




5. Ugom II* 


Akugum If . 


6. Letangkem 



Akugiim Ilf . 

7. Lebulibulikom . 



Kikongkula I 


8. Aboni . 



Kikongkula II 

9. Kebung . 





ID. Otalosi . 

. ? 

II. Utongt • 

• 3 




12. {Lekpangkem* . 

13. |Ndai 

• 3 



• 4 

• 5 



• 3 

14. ( Egbisum . 

15. \Letekem 

• 3 



• 4 

• 3 


Ibenda I 

• M 

16. Usadja . 

• 3 


Ibenda II 


Distinct territorial groups sharing a single shrine bracketed on the left. 

Groups with separate shrines but a combined dwelling-area bracketed on the 

Groups with shrines in their dwelling-areas shown in italics. 

Groups of which one or more lineages have migrated to other dwelling-areas 
indicated by *. 

A group composed of, or including in its dwelling-areas, lineages from other 
clans indicated by t- 

Incipient clan composed of discrete migrant lineages indicated by J. 

constitute a single ritually united exogamous entity, and, on the other 
hand, single residential groups which in fact consist of two exogamous 
units each with its independent ritual. The position may be summarized 

Patrician dwelling-area names ........ 26 

Exogamous groups each with its own shrine . . . • .22 

Named groups with separate dwelling-areas but lacking exclusive shrine . 8 
Exogamous groups with named sub-groups occupying distinct dwelling- 
areas ........... 6 

Groups with a single named dwelling-area, consisting of two or more 

exogamous units with separate shrines ...... 5 

Exogamous groups, many members of which (all or part of one or more 

lineages) live in the dwelling-area of another group . . . .7 

These divergencies in territorial arrangement and ritual organization 
result from processes of fission and accretion whereby the structure and 



composition of many of the patricians has been modified during a period 
of marked and doubtless uneven growth of population. Thus the Utong 
people in Umor consists of three lineages, two of which are recognized 





1 D J U M 


Groups are arranged by wards on the four sides and are numbered in ac- 
cordance with Table I. Circles indicate groups with patrician shrines in their 
dwelling-area. Squares indicate separately named territorial groups recently 
formed by migrations of lineages. Broken lines indicate recorded migrations of 


as being descended from migrants from other villages but are now 
regarded as having combined with one native lineage. The group as 
a whole recognized the ritual authority of the shrine and priest of the 
patrician Anedja-Lekpangkem from which the original lineage itself 
migrated. The three lineages of Unebu are similarly recognized as 
descendants of migrants from three different patricians. Here fusion has 


not progressed so far. Two of the lineages still consider they belong to 
their parent yepun and resort to the shrines there, while a recent action 
of the third demonstrates a stage in the process of actual kepun fission. 
After a dispute which came to a head in connexion with a rite at the shrine 
of the parent clan, they determined to create their own shrine at Unebu. 
This creation of a new epundet with a material nucleus of pots and 
boulders obtained from an earlier foundation is a recognized if rare pro- 
cedure, and several are pointed to as having been established in the past 
generation,' If the rift between lineages in a clan is deep the dissident 
group may, as recently happened in Akugum in Umor, seek the nucleus 
of pots and boulders with which to establish their new shrine not from 
that of the clan they are leaving but from one in the Ugom kepun known 
as Elamalama which is held to be the first epundet to be established at 
the foundation of the village. This ritually expresses the complete 
severance of former clan ties and at the same time the continued 
adherence of the group to the wider village community. 

Fission and fusion in varying degrees imply that the wide range of 
social activities characteristically grouped within a coherent patrician 
may not, in practice, always be so tidily arranged. The customs of 
exogamy, of economic and ritual co-operation, and the underlying 
sentiment of kinship maintain a general stability, but all can be weakened 
by internal strains and can be reforged in a new grouping.^ 

The provision of food-supplies at funeral rites is a special concern of 
the dead man's lineage, and the inheritance of membership in men's 
associations falls to a lineage-fellow before it is oflFered to, or expected 
by, other members of the kepun. Moreover, the offices of kepun priest 
and his assistant are properly filled by members of certain lineages only. 
The choice is almost always restricted to one or two lineages ; when more 
than one lineage is concerned, these provide the priest in rotation. 

The estimate of the numerical strength of the patricians will of course 
depend on the criteria adopted to define them, but they may be said to 
range in size from about 50 to 150 grown men. They are therefore com- 
paratively small groups whose residential unity obviously strengthens 
their social coherence. Born and growing up in one dwelling-area, 
playing from childhood with others who will later be his neighbours in 
the village and his companions in farm work and palm tending, a man 

' Patricians which have thus recently come into existence by fission sometimes 
lack a name of kepun type. Members of such groups distinguish themselves 
by using the name of the priest, that is, by calHng themsehes 'the kepun of So- 
and-So', or by using the Hneage names in addition to a kepun name shared with 
another group, e.g. Ibenda comprises two clans distinguished as 'Ibenda yi 
Iwara Kekbiyen' and 'Ibenda yi Iwara Edo'. 

^ On the other hand, individuals may for particular reasons live outside their 
patrician dwelling-area without dissociating themselves from it. When, for 
example, a man succeeds to a matrilineal priesthood he should go to live in the 
lekama where the shrine is situated. 


comes to regard his kepun group as a community within the larger 
village. This corporate sense extends to outlying farm and forest lands, 
since the patrician provides the framework within which each family 
head secures personal rights to the tracts from which he obtains his 
food-supply and palm-oil. A farm-road elder {oponotam eta), often the 
head of one of the lineages, adjusts the claims of individuals in each tract 
of land claimed by the patrician. 

In the larger villages the patrician dwelling-areas are grouped in 
districts or wards known as yekpatu (sing, kekpatu). It is necessary here 
only to emphasize that the grouping together of certain patricians within 
a single ward and the organizations based on the ward, of which the age- 
sets and some men's associations are the most important, are purely 
territorial and are not based on any assumption of kinship ties between 


The Obot Kepiin, the priest who officiates at its shrine and is custodian 
of the sacred tusk trumpets and staves, is the ritual head of a patrician. 
He is treated with considerable deference. He receives gifts from the 
participants in rites and from all men of the clan at the annual New Yam 
ceremonies. It is also his duty to maintain harmony in the group, to 
compose disputes and announce decisions. But in this he is much depen- 
dent on the balance of opinion among the elders of the clan (yaponota?n, 
sing, oponotam) and particularly the heads ('fathers') of the component 
lineages. For he is primarily a priest and a peacemaker, and active leader- 
ship and the mobilization of opinion on public affairs are left to the play 
of personality among the elders. A clan may be dominated at a particular 
time by a leading man in one of the lineages. If so, the Obot Kepun will 
accept him as his agent and representative in arranging its affairs. At 
other times there may be serious rivalry between factions among the 
lineages and his task is more difficult. But the priest himself is expected 
to listen to all sides, to avoid partisanship and, unless it flagrantly disre- 
gards custom, to accept the prevailing opinion on any question at issue. 
He is not, therefore, the active manager, far less an autocrat, in the clan. 
On the other hand, no question of substance is considered properly 
settled until it has been thrashed out before him and his judgement has 
been given. Disputes between clansmen are heard by him together with 
the elders concerned and delinquents are ordered to pay small fines and, 
for ritual offences, to make an offering at the shrine which is shared 
among them. Each priest has a regular assistant — 'the pourer of palm 
wine' in rituals at the epundet — who should succeed to the priestship. 
His recognition as assistant in fact depends on his being the designated 
and accepted successor to the priest. 

The accepted leader of a lineage, its 'father', who is turned to for 
advice and who acts as its spokesman, takes the initiative in composing 


any internal disputes, and is expected to speak on behalf of any individual 
or group in the lineage accused of an offence or suffering a grievance. 
Both men and women make him small gifts when they seek his services. 
Although seniority in patrilineal descent is emphasized, succession 
depends largely on personality. A leader usually chooses, from among 
his male relatives a few years junior to himself, an acceptable lieutenant 
who acts as his messenger and general supporter. This man, if he gets 
on well with his fellows and shows qualities of moderation, succeeds 
without discussion or formal appointment. There is, however, a tendency 
for the leadership to revert to a descendant in the direct line of a former 
head. Alternate succession may be the rule when a lineage has two 
branches. But the head of a lineage should not merely be eligible by 
descent and seniority ; he should be notable and prosperous, impressive 
and conciliatory. 


When the genealogical structure of Yako patrilineages is examined 
in detail, it soon becomes apparent that many and sometimes the 
majority of the members do not trace actual descent from the founders, 
but are descendants of men adopted into the group. Thus in the Ndai- 
Lekpangkem kepun in Umor only seven of the sixteen living adult men 
of Etung Enamuzo lineage are lineal descendants of the founder, while 
there are four groups of descendants of individual adoptions. Similarly 
in Itewa lineage, only five of the present twenty-one adult members are 
lineal descendants, while the rest are members in virtue of four adoptions 
at various times. These adoptions, usually effected in childhood or 
youth, may bring into the group men from other yepun within a village, 
men from other villages, and even strangers from other language groups. 
Adoption is not marked by any rite whereby a man's status is formally 
and abruptly changed and an earlier kin allegiance annulled in favour 
of his new status. It may often be a matter for debate whether or not, 
at a particular moment, a man has become a member of a new kin group 
by adoption. 

Very young children nearly always accompany their mother if she 
leaves her husband, and if their mother dies they are as often given to 
the care of her relatives as to those of the father, and so go to live, at 
least temporarily, in another patrician. With older children the customary 
attitude differs as between boys and girls. Although the father may in 
a particular case agree that a son should go with the mother, this does 
not imply his consent to adoption by the mother's patrikin, and he is 
upheld by his own kin and by the village head if he refuses to let the 
child accompany her. A girl, however, almost always accompanies her 
mother, and, although her father can claim his customary portion, the 
matrilineal kin of the mother in such circumstances may take complete 
responsibility for her marriage and receive all the marriage payment. 


This is symptomatic of the weaker attachment of women to their 

A father and other patrikin usually maintain contact with an absent 
son during his childhood, and encourage his return at manhood to the 
patrician of his birth, and if a son is with a stepfather he does nearly 
always return. But when the child, on the death or divorce of his mother, 
is entrusted to a mother's brother, he is likely to associate himself more 
closely with the patrikin of his foster-father, and to be offered a house 
site and farm-land there when he grows up. Acceptance, which would 
normally imply an intention to attach himself permanently to the 
patrician of his foster-parent, will depend on whether his actual father 
is still living and is sympathetic to him, and on his relations with his 
brothers and half-brothers in the patrician of his birth. The crucial issue 
is often the provision of funds for marriage payments. If these are 
provided by his foster-parent, the youth will be expected henceforth to 
behave as a member of that patrician. On the other hand, provision of 
the greater part of the marriage payment by his own father or other close 
patrikin will be dependent on the return of the youth to his patrician, 
where he will find a house site and farming lands. 

Foster-children or wards, known as yawunen (sing, owonen), are quite 
frequently pressed to stay by their foster-fathers, who desire to increase 
the number of their households and descendants. They undertake the 
marriage payments for their adopted sons, and sponsor them in all 
claims to status in the new kin group. Love-affairs or marriage with 
women of the foster-father's patrician will then be forbidden as incestu- 
ous to the children of an adopted son, who will be free, on the other 
hand, to mate with any but close relatives within the patrician in which 
their father was born. The decision to adhere permanently to the patri- 
kin of a foster-father does not, however, entitle a man to seek a wife for 
himself in his own father's patrician. Many associations and obligations 
connected with his kinship status during infancy still cling to him. It 
should also be stressed that adoption is never irrevocable. A man is held 
to be free to return to the kin of his birth at any time, and should he do 
so the rights, obligations, and ritual restrictions, such as kin exogamy, 
which operated with regard to the kin into which he had been adopted 
are regarded as having lapsed. 

The adoptions so far considered involve transfer from one patrician 
to another within the village, but persons may also be adopted from other 
settlements. A stranger may have been invited to live with a clan because 
of his skill in a craft such as wood-carving, but the majority have come 
as the kinsmen of women who have come from other villages as wives. 
A widowed or divorced woman from another village, whether it be Yako 
or one in the neighbouring territories such as Ekumuru or Adim, may 
bring with her a young son who, by the time he is adult, will be accepted 
as a member of his stepfather's kepun. An adult kinsman of the foreign 


wife may also come as a guest, help her husband in the farm-lands for 
a season or two, and eventually be taken to the patrician head and farm- 
road elders 10 be given a plot of his own. The majority of adopted 
foreigners whose personal histories could be obtained were found to be 
such relatives of foreign wives. 

A second class of extra-village adoption arises from the recognition 
of matrilineal kinship. It consists essentially in the adoption in childhood 
of a boy or, less often, of a girl from among the foster-parent's matrilineal 
kin in another village. This form of adoption also commonly results 
from extra-village marriages. The sons and daughters of a woman who 
has married into another village will normally become members of that 
community. But a man often seeks to adopt a sister's son, especially if 
he is without sons of his own, while, on the other hand, such a child may 
be offered to his care in consequence of some dislocation in the sister's 
household. In addition to actual sisters' sons, more remote and putative 
matrilineal kinship may be the occasion for such extra-village adoption. 

But a considerable number of children of completely alien origin have 
also, as has been mentioned earlier (p. 291), been adopted into Yako 
households and so into the patricians. These children, known as yafoli 
(sing, ofoli), have been brought to the village by strangers and have been 
handed over in return for payments, usually made in money. This may 
have occurred more frequently than the genealogical data reveal, for 
both the inferior status of the descendants oi yafoli and the punishment, 
under slave-dealing ordinances, of the sale and purchase of children have 
probably led to the concealment of such actions in the past, and to the 
misrepresentation of yafoli as yawunen, or actual begotten children. 
During the period of increasing wealth, with the growth of the palm-oil 
trade over the past four or five decades, the adoption of yafoli has 
certainly contributed to the rapid growth of the population which is 
indicated by the expansion of the villages. To purchase an ofoli has been 
a meritorious display of wealth, and an achievement necessary for one 
who claims to be a man of substance {osu, pi. yasu) and to partake in the 
reciprocal funerary feasts of the rich. Yafoli have been acquired far 
more frequently by men than by women, but childless wives often urged 
their husbands to get a child in this way. Both boys and girls have been 
obtained, but girls more frequently than boys. The adoption of girls is 
expressly stated by the people themselves to reflect the desire to increase 
the numbers and prestige of the matrilineal kin. It is also valued as 
giving a woman a companion and helper in her farm work. Yafoli have 
not usually been married by their purchasers and the status and signi- 
ficance of a girl ofoli is more often that of a daughter. For a short period 
she may make possible an increase in the size and productivity of her 
foster-father's farm, and later he receives a marriage payment for her. 
After that she is likely to have little significance for the patrician unless 
serious matrimonial disputes arise, for her children are matrilineal, not 



patrilineal kin of her foster-father. The boy ofoli Ukewise ranks as 
a potential if inferior matrilineal heir, but he also grows up as an 
adopted member of the patrician of his purchaser. He can inherit status 
there from his foster-father, although he ranks bejow a begotten son, 
and his origin may be used against him in abuse in moments of anger 
both by kinsmen and by other villagers. 

An adopted youth, whether originally a kinsman {owunen) or an ofoli, 
can often be a direct economic asset. A man will farm and build, collect 
palm products, and engage in petty trading for at least fourteen years 
before any son of his own is old enough to act as a regular and effective 
male helper. A young married man, therefore, often welcomes an oppor- 
tunity to adopt a boy lo or 12 years of age. But even more influential 
than these economic benefits is the widespread desire to enlarge the 
household and increase the number of descendants in lineage and clan. 


The patrilineal kepun is not, however, the only group of kin within 
which a Yako has rights and obligations. It has already been seen that 
ties of matrilineal kinship underlie many of the adoptions of men into 
yepun. These ties not only involve rights and duties between persons 
closely related in this way ; they are also the foundation of matriclans, 
known as yajima (sing, lejima), complementary to the patricians just 
described. Apart from full siblings, the persons composing the patrikin 
and the matrikin of each individual are normally distinct, and it is of 
importance for the understanding of the social relations that are 
established through matrilineal kinship to bear in mind that, among 
close relatives, only full brothers and sisters will of necessity belong to 
both groups. Since residence is patrilocal, the members of a matriclan 
must be dispersed among the patrician territories of a village while 
a minority will be living in other villages. 

The rights and obligations which derive from matrilineal kinship do 
not formally conflict with those derived patrilineally. Over the greater 
part of the fields of economic activity, ritual observance, and succession 
to property, a clear distinction between their spheres of application has, 
until very recently, been maintained. Matrilineal kinship should take 
precedence over patrilineal in the inheritance of transferable wealth, 
especially livestock and currency, in the receipt of payments made to 
a woman's kin at her marriage, in the corresponding responsibility for 
the return of payments received for women who later unjustifiably leave 
their husbands, in responsibility for debts incurred by an individual, and 
also in obligations and rights in respect of recompense for bodily 
injuries. On the other hand, patrilineal rights and obligations, as has 
been seen, largely relate to the use of land and houses and to the pro- 
vision of co-operative labour, especially in the annual farm-clearings at 


the beginning of the agricultural season. Yako say that 'a man eats in his 
kepun and inherits in his lejima\ 

In the present generation, however, matrilineal ties are being under- 
mined, especially in the villages of Ekuri and, to a lesser degree, Umor, 
by successful claims of sons to their father's personal possessions and by 
the retention of marriage payments by fathers. Administrative confirma- 
tion of decisions in this sense in the Native Authority courts have 
strengthened this tendency. But in the thirties the majority of the Yako 
still adhered to earlier custom, and, although some are now determined 
to flout them, no Yako denies that matrilineal rights exist. 

Close matrilineal relatives will be in touch with all important events 
and circumstances in each other's households and will give mutual 
support within the framework of the accepted kinship obligations. 
Although in the larger villages many members of one's matriclan may 
be encountered only occasionally, all are made vividly aware of their 
fellowship two or three times a year in rituals which are the visible 
symbols of the unity and social reality of the matrilineal groups. 

The matriclans are similar in numerical strength to the patricians. 
The largest do not much exceed 100 adult males and some are much 
smaller. The prefix j'«, signifying 'the people of , . .', is incorporated in 
the names of the matrilineal kin groups which are quite distinct in form 
from those of the patrilineal groups.^ Some are held to have been inde- 
pendent groups at an indefinitely remote period. Such clans have priest- 
hoods. Others are ritually attached to one or other of these groups and 
are in some cases held to be off^shoots of the clan from which the priest 
of the spirit {yose, pi. ase) they recognize is chosen. Other secondary 
groups are held to have transferred their ritual dependence to another 
shrine when they separated from a parent group. Fig. 14 and Table II 
summarize the position in the village of Umor, where it will be seen that 
although 23 distinct matriclans are recognized, there are only 10 cult 
groups. Twelve of the former, although they are independent with 
respect to other rights, depend for their supernatural benefits and for 
the ritual validation of decisions by their elders on rites carried out by 
the priests of other clans. All but four of the priests officiate for more 
than one matriclan and these four are said to be priests of groups which 
were themselves dependent on others before they established shrines 
and priesthoods of their own.^ 

In every matriclan there is a group of elders known as yajijnanotam — 
old men and women of the lejima — from six to a dozen of the more 

' Thus Yabaye refers to one matriclan as a corporate entity as well as to a 
plurality of members. An individual member is an obaye. But the name of a 
kepun such as Egbisum is distinct from the term for members, who would be 
referred to individually as Ogbisum and collectively as Yagbisum. 

^ For details and documentation of these processes see my earlier paper on 
'Kinship in Umor'. Amer. Anthropologist, xli, 4, Oct.-Dec. 1939, pp. 523-53. 


senior and active members and leaders in component lineages who, 
although they have no ritual functions and no ceremonial duties and 
maintain their numbers by informal co-option from time to time, do in 

Ritual Groups 

9,10, ,(12713), 14 I ,(l6), 17, 18,19,20,21, ,(22), 23 


j Clan from which a Yose priest is chosen 

Recorded fission and transference 


Table II 
Matriclans ( Yajima) of Umor 












f Yabot II (from 





Yabot I) 


Yakpambot of 

(Obot Lo- 








Vangyo (from 

Obete Edet 





Yakoibot (from 



Yabot I) 





Yosenibot (from 




Yabot I) 



, , 




19. I 


, , 



Yakpolo (from 

. . 



Yabot I) 




\Yanali (from 


Yabaye (from 




22. ^ 

23. ^ 



Capitals indicate matriclans providing ase priests and not known to have 
separated from another group. 

Italics indicate matriclans said to have separated from others as indicated to 
create their own ase. 

Ase which may not be moved from the site of their alleged foundation indi- 
cated by *. 

Nine matriclans have an exclusive right to elect priests; two have rights in 
alternation ; twelve provide no priests ; one is in process of fission ; eight (seven 
and a section of an eighth) have transferred their allegiance and five of these have 
created new ase. 


fact exert considerable influence both on the members and on village 
affairs that concern the group, Recompense for offence by one member 
of the matriclan towards another can be enforced by requesting the 
priest of the yose to declare the offender excommunicated. By appeal to 
the village council they are able, as will be seen, to protect the interests 
of particular members of the group in the payments and responsibilities 
involved in the marriages of their matrilineal kin. On the death of a 
matriclan priest they nominate his successor. 

Within each village one of the matriclan spirits is held to be superior 
in power and in ritual status. Its cult is maintained, not only for the bene- 
fit of members of the matriclan of its priest and of any others second- 
arily attached to it, but also on behalf of the village as a whole. Its shrine 
has precedence in all village rituals and the priest, as ritual head of the 
local community, is known as Obot Lopon (Leader of the Village). But 
this does not confer any formal authority or superior status on the 
members or elders of the matriclan itself. Indeed, the matriclan organiza- 
tion, like the patrilineal groups already discussed, exhibits a characteris- 
tic absence of systematic ranking of groups. 

While all members of a matriclan within a village are not only ritually 
associated but regard one another as kin, they are not an undifferentiated 
group. The exercise and transmission of rights operate within smaller 
units of close relatives known as yajimofat (sing, lejimafat). These 
matrilineages are comparable in size and span to the patrilineages and 
it is within these smaller groups that most rights and obligations between 
matrilineal kin are exercised and transmitted. The collective term for 
matrilineal kin — yatamban — usually refers to the matrilineage, although 
it may apply also to the clan. 

It is only within the matrilineage that exogamy is now strictly enforced. 
Formerly marriage within the matriclan was strongly disapproved. Old 
people still hold that such a marriage is likely to be sterile and to weaken 
the beneficient power oi the yose. But during the thirties, starting first in 
Ekuri, both practice and attitude were changing in connexion with the 
conflict over inheritance rights between sisters' sons and sons. There 
has been a general tendency in recent years for sons to lay successful 
claim to at least a share in the movable goods left by their father which, 
by former custom, passed entirely to matrilineal heirs. Marriage with 
a girl of another lineage in one's own matriclan has come to be encouraged "^ 
as a device whereby this conflict can be resolved, the son of such a 
marriage being regarded as the matrilineal heir of his father. Such sons 
are successfully claiming sole inheritance of their father's movable goods'^- 
as members of his matriclan and virtual members of his matrilineage, - 
while young men are being encouraged by elders of their own matri- 
lineage to make their first marriage with a girl of another lineage in the 
matriclan, so that they will be able to transmit their wealth to their 
eldest sons. 


Unlike the patrician system, which does not extend beyond the village, 
the matriclans in one village are held to correspond with clans in others. 
There are no common rituals, but there is a considerable parallelism in 
names of clans and clan spirits, and even a vague belief in ultimate 
common descent. People coming from another village, including not 
only individuals bent on trade or visiting a close kinsman who has 
migrated but also parties of young girls who commonly make stays of 
several weeks in the rainy season, can expect hospitality from members 
of the corresponding matriclan. Women marrying into another village 
are, together with their children, automatically accepted as members of 
the appropriate matriclan there. Formerly sons of such women lost no 
rights as matrilineal heirs of their mother's brothers in her village of 


It has been seen that rights to dwelling-sites, farming land, and the 
more important forest resources are obtained by virtue of membership 
of a patrilineal group ; but that when attention is transferred from rights 
in economic resources to the transmission of accumulated wealth, 
matrilineal kinship comes into prominence. The funeral and the disposal 
of a dead person's property are supervised by a close matrikinsman, 
a mother's brother or a sister's son, who in the case of a man is chief 
heir. All currency, whether it be in brass rods or modern coinage, and 
all livestock, should by custom pass to matrilineal relatives who also 
receive the greater share of the implements, weapons, household goods, 
and any stores of food. The movable property of women, which is 
usually less considerable, passes mainly to their daughters, but sons 
should expect very little from their parents. A man may obtain a gun or 
cutlass that was his father's ; the rest will be claimed by his father's 
brother or sisters' sons. 

With these rights of inheritance are associated corresponding obliga- 
tions, and, in particular, responsibility for debts of matrikinsmen, a 
readiness to make reasonable loans at need, and the duty of providing 
a part, although a minor share, of the currency and goods transferred 
by a sister's son at marriage. If an outside creditor is unable to obtain 
satisfaction for a debt he will often seize a goat or cow belonging to 
a close matrilineal kinsman of the debtor who can obtain no recom- 
pense through either the Village Council or, in more recent times, the 
Native Authority court, if the existence of an unduly prolonged debt is 
satisfactorily established. Considerable payments are involved in mem- 
bership of men's associations among the Yako and debts are often con- 
tracted in this connexion both with individuals and the clubs themselves. 
If the debtor is dilatory, his matrilineal kinsmen are summoned and 

' Marriage with women of other villages appears to have been largely confined 
to Umor. 


requested to make arrangements for settlement. If they refuse or fail to 
carry out their agreement, the Village Council would formerly have 
approved the seizure of their livestock by one of the associations. Nowa- 
days most disputes over considerable debts are taken to the Native 
Authority court established by the Nigerian Government, but the 
responsibility of an offender's matrilineal kin to contribute if necessary 
has been recognized there. The court likewise summons the matrilineal 
kinsmen of a bankrupt debtor and orders them to undertake a settlement. 
Finally, if a man at his death leaves outstanding debts, his matrilineal 
kinsmen succeed to the obligations. 

When a man dies, his household usually continues to function as 
a productive unit with the aid of brothers and sons until the yam harvest 
of the farm is gathered in.' The dead man's crop, which will be con- 
siderably larger than those of his wives, is at the disposal of his matri- 
lineal kin. But since the chief claimant is often a full brother, who is also 
a patrikinsman and undertakes obligations to the children of the dead 
man if they remain in the kepun, a considerable part may in fact be used 
for the benefit of these children. The widows will remarry or join either 
their grown sons or their own parents, and unless they are still caring 
for young children of their dead husband, when they will be given a 
portion, they have no right to any share of their dead husband's yams. 
The rest of the surplus is taken by brothers and adult sisters' sons for 
planting in their farms or for sale. But the dead man's children will 
themselves go later to his brothers, as to a 'father', for a supply of 
yams to make their first farms. 

The position of women as wives and as owners of property in their 
husbands' farms is similarly the concern of their matrilineal kin. When 
a wife dies, both her rights and duties in her husband's farm pass for the 
rest of the farming-season to one or more matrikinswomen. If she dies 
after a harvest, her yams in the household yam store are claimed by her 
matrilineal kin and are usually taken by a sister or a grown daughter. 
If, on the other hand, a woman dies during the growing-season, the 
husband asks to be given the services of one or more of her 'sisters' to 
care for the farm and carry in the harvest. When the crop is dug, the 
yams of the deceased wife are carried away by the women who have 
taken her place, for distribution among themselves and other matri- 
lineal kin. 

Apart from the recent open challenging by sons of the inheritance 
rights of matrikin referred to earlier (p. 309), some evasion by men and 
their sons is said to be of long standing. Hoards of brass rods and later 
of Nigerian currency have not infrequently been hidden by secretly 
burying them in farm-land. Apart from the ease of concealment, the 
profound suspicion of magical malpractice which is aroused by trespass 

' For an account of relevant fanning practice see my 'Land and Labour in 
a Cross River Village'. 


on another's farm-land and the magical safeguards against it render such 
a cache in farm-land fairly secure. A man may confide in one or more of 
his sons the whereabouts of such a hoard while withholding knowledge 
of it from his matrikin so that the former may quietly take possession 
after his death. 

Matrilineal kinship confers no settled right to economic resources 
which are exploited within the patrician, that is, to farm plots as distinct 
from specific harvests, to oil-palm groves or to planted trees. But the 
categories are not entirely clear cut since matrikin should supply one 
another when there is special and occasional need. A man should never 
refuse a reasonable request from a sister's son or other matrikinsman for 
some of the produce of his planted trees and will seek approval from the 
head of his patrilineage for him to collect palm products on their land. 
Moreover, a considerable number of the valuable clumps of raphia 
palms, largely confined to swampy ground not used for farming, which 
provide much valued building material for houses and yam stacks are in 
fact the property of matrilineages. It is recognized that a man who plants 
a new clump may share it during his lifetime with his sisters' sons and 
leave it to one or more of them at his death. The right of control then 
passes matrilineally, although conversely owners' sons in each generation 
, are then allowed access to the clump for their own needs. This practice 
runs counter to the general principle of patrilineal succession to farm 
and forest resources and appears to be related to the very uneven distri- 
bution of the scarce raphia-bearing swamps among the lands of the 
patricians. Finally each matriclan claims the right to collect palm-wine 
over a tract of village territory for use in the funeral rites of matriclan 
elders, at the installation of their priests, and in certain other rituals. This 
temporarily overrides the rights of those who have prepared the trees in 
their patrician lands, and enables a matriclan to act corporately in 
assembling a large supply of palm-wine. 

Membership in most of the men's associations among the Yako is 
obtained by succession to a deceased member. This may be either 
patrilineal or matrilineal, since a man's right, which is also an obligation, 
to join a society may be inherited from either a matrilineal or a patri- 
lineal kinsman and this right should pass on by the same rule as that 
whereby the deceased succeeded. The close kin, both matrilineal and 
patrilineal, are expected to contribute to the fees required, but the more 
substantial payment is made by the group providing the successor, and 
in their case the obligation extends beyond the close relatives and con- 
cerns all the older members of the lineage and other notables of the clan. 

Homicide within a village also initiates a series of customary claims 
for compensation which are the concern of the kinsfolk of those directly 
involved. Formally no distinction is made between deliberate and 
accidental homicide, and while it is not admitted that murder of one 
villager by the deliberate violence of another is a danger to be guarded 


against, anger is little tempered by evidence, however obvious, that the 
injury was accidental. If a man is killed, the reactions and claims of the 
two sets of his kinsfolk, patrilineal and matrilineal, are clearly distinct. 
In both groups anger is aroused, but the immediate crisis and the danger 
of retaliation is m^uch greater among the patrikin. When the body is 
brought into the compound, there is a likelihood that the close patri- 
lineal relatives will urge the men of the patrician to attack the offender 
and his patrikin. Fights arising in this way have fairly frequently 
occurred. Indeed it is customary for the offender to seek shelter in the 
compound of the Village Head, andstay there while the Village Speaker is 
sent to the assembly house of the dead man's patrician bearing the village 
elephant-tusk trumpet and drum which, by supernatural sanction, impose 
restraint and an obligation on the bereaved patrician to keep the peace. 
The offender, or one of his close patrikin, has also on occasion been 
seized as a hostage by the kepun of the dead man and held until a pay- 
ment was made to them for his release. 

Nevertheless, if one of their number, man or woman, is killed the 
patrilineal kin have no customary right to compensation, material or 
ritual, from the patrician of the offender. This concerns only the matri- 
clans of the deceased and the person held responsible. A matriclan which 
suffers loss by the violent death of a grown member at the hands of an 
outsider has a recognized right to compensation which is ritually safe- 
guarded. It is required that the offender and also, where an adult man 
has been killed by another, one of the former's matrilineal kinswomen 
be transferred to the dead man's lejinia in a ceremony at its yose shrine. 
The transfer of the woman has, however, been commonly compounded 
by a payment sufficient to purchase a female ofoli. There is a keen sense 
of loss over those who have been transferred, but they must attend and 
contribute conspicuously at the funeral rites of members of their new 
matriclan and must avoid ostentatious participation in rituals of their 
former matriclan. The transfer of a woman, which is always permanent, 
results, if she bears children, in the numerical increase of the matriclan 
she has entered and also in the transfer, by inheritance, of property from 
the people of the offending lejima. It was fairly clear that restoration of 
the strength of the injured lejvna was the most vital aspect of these trans- 
fers and they reflected a sentiment, more consciously developed in the 
matrilineal than in the patrilineal groups, which is being continually 
enhanced by the fertility rituals at the ase shrines. 

This practice of compensating the matrilineal kinsfolk alone is not 
intelligible from a material point of view. Any children that a man who 
is killed might later have begotten would have been no concern of his 
matrilineal kin and, since his existing property should pass to them in 
the ordinary way, any loss to the matriclan is confined to wealth he 
might subsequently have accumulated. The patrician, on the other 
hand, suffers the loss of a breadwinner and a progenitor by the death of 


a grown man, yet it can legally claim no compensation from the offender's 
patrician either in services or supplies of food for the household, nor are 
any males transferred to membership of the bereaved kepun. The strong 
sentiments concerning fertility and peace within the lejima as described 
later appear to be responsible for the maintenance of these practices, but 
there is nevertheless a lack of symmetry. The matrikin are given the 
emotional satisfaction of restitution but the patrikin are not, while in 
economic terms the matrilineal kin are over-compensated and the 
patrilineal kin remain without recompense. It is held that these obliga- 
tions to the matrilineal kin of victims of homicide are not affected by 
close patrilineal kinship between the parties. The murder or killing of 
a half-brother, a son, or other kepun kinsman would involve restitution 
to the bereaved matrilineal kin in the ordinary way. On the other hand, 
no claim for compensation could arise from a killing that involved two 
men of the same matrilineal group. 

Rights to inheritance of property, to aid in the accumulation of 
their marriage payments, and sometimes to succession in associations 
inevitably strengthen the ties between youths and their senior matri- 
lineal relatives, and there is usually an intimate relation between a man 
and his sister's son — the classic relationship which cuts across parental 
ties in societies stressing matrilineal descent. And among the Yako this 
relationship is often converted when opportunity arises into foster- 
fatherhood with subsequent adoption of sisters' sons into the patrician 
of the mother's brother. 


Every matriclan is associated with a fertility spirit — yose (pi. use). This 
is embodied in a miscellaneous set of cult objects, including decorated 
skulls, figurines, helices, and penannular rings of brass and copper and 
various pots, which are kept on an altar in a miniature house in the 
compound of the priest and arranged on an adjacent open-air altar at 
public rituals. The successive invocation of the fertility spirits at these 
shrines constitutes the central act of the village rituals at various stages 
in the farming year, and the ase, not the yepundet, spirits of the 
patricians, are in native belief the spirits primarily active in maintaining 
the well-being of the village. Every girl is brought by a close matrilineal 
kinsman, usually an older brother or a mother's brother, to the shrine 
of the fertility spirit of her matriclan during her first pregnancy. Gifts 
and materials for an offering are brought on an appointed day when 
a rite is performed to safeguard her and her unborn child, a future 
member of the group. A corresponding ritual, in which the woman is 
accompanied by her father, is performed at her patrician shrine, but is 
regarded as less powerful. The priests of the ase, each known as the ina 
(pi. biHna) of the spirit in question, e.g. Ina Atalikumi, have, in conse- 
quence of this greater ritual power, a prestige and authority superior to 


those of the priests of the patrilineal kin groups. This power of the ase^ to 
which their priests alone have direct access, is the ritual sanction of 
authority in the wider field of social control in the village as a whole, 
which overrides authority within the several patricians. The ase priests 
are the nucleus and strength of a sacerdotal council in each village which 
has the power to invoke the destructive or beneficent actions of the 
spirits themselves, and thus commands the strongest supernatural 
sanctions in the village. 

Each is chosen from a certain matriclan and proposed by its elders 
but acts on behalf of every clan dependent on the spirit concerned. 
Although a close matrilineal relative of a former priest is considered 
most appropriate, succession is not rigidly prescribed. Choice usually 
falls not on one of the elders but on a man in early middle life and, 
although he is selected by the clan elders, he must be approved, and then 
instructed in his ritual duties, by the group of ase priests as a whole. 

A priest should reflect in his own physical well-being, in the serenity 
of his temperament, and in his peaceable behaviour the beneficent 
power of the spirit, and he will usually ask the priest of another spirit 
to deputize for him if he is ailing or has suffered distress at the time of 
a ritual. If two or more priests of a single yose die in succession after 
only a few years of office, a notion of restoring life-maintaining power 
by infusing new blood into the priesthood comes into play. It is believed 
that a weakness revealed by the chain of misfortune can be overcome by 
selecting as a successor a son of a man of the matriclan, an okpan, that 
is, by breaking the rule of succession within a group of matrilineal kin. 
This may be and occasionally is done apart from previous misfortunes, 
if a man, well fitted by physique and temperament on whom the clan 
elders and the ase priests are agreed, cannot be found within the group. 
A priest appointed in this way is ritually adopted into the lejima at his 

Although each priest carries out private rituals at his shrine for all 
persons of the matriclans concerned, the seasonal rites are conducted 
jointly by all the priests and for the village as a whole. In each village one 
of them is, as has been said, recognized as the Leader of the Village 
(Obot Lopon), in virtue of his control of the cult of the premier spirit 
which is regarded as the fertility spirit of the village and the most power- 
ful supernatural force within it. Selected from a single matrilineal kin 
group, he is as pre-eminent among the priests as is that spirit among the 
others and he is recognized as the religious head of the village. 

This corporation of priests, known collectively as the Leaders 
(Yabot),^ does not consist solely of the BVina but the other members 

' The term Obot (pi. Yabot) can best be translated as Leader or Head, but its 
connotation varies with its context. Used alone in the plural it is understood 
to refer to the indigenous sacerdotal council of the village, the kepun heads are 
distinguished as Yabot Yepun. The same term is also, however, the title of 


such as Okpebri (the village Speaker and prayer-leader in collective 
rituals) all belong to it in virtue of their appointment to ritual offices at 
the instance of the other priests. Its functions are not, however, restricted 
to ritual performances and the securing of supernatural benefits. It also 
guides village affairs, reaffirms customary law, and attempts to settle 
major disputes. In Umor the priest of Odjokohi is Obot Lopon and head 
of this council. His compound, known as Lebokem, is situated on one 
side of the village assembly square (Keblaponga), near the centre of the 
village and not in the territory of any patrician or ward. In his compound 
at the Odjokobi shrine all the public rituals of the village are initiated 
and reach their climax and in it, too, the council assembles to discuss 
village affairs and to hear disputes that are brought before them. 

The Yabot thus form a close-knit corporation whose members are 
linked to the rest of the village through the matriclan system. They are 
not representative of the patrilineal groups or of any territorial section. 
They are responsible for announcing the times at which many seasonal 
activities should begin, and for refusal to comply with their regulation 
of such activities they impose fines. They have authority both in civil 
disputes, where they serve as a court of appeal, and in public offences, 
for which they can impose fines and order expiation. ' Disputes between 
persons and groups which cannot be settled by the arbitration of the 
elders of the kin groups concerned, or by the ward heads in the larger 
villages, are taken to them by one or other of the contestants. They would 
also intervene on their own initiative to deal with public offences both 
secular and religious.^ The Yabot command powerful sanctions both 
physical and supernatural to enforce their decisions. The supernatural 
sanctions lie in their own hands. They can refuse an offender and his 
close matrikin all access to the shrine of his matriclan spirit. More 
drastic is a ceremonial declaration of the offence to tht yose and a request 
that the beneficence of the spirit be withdrawn from the offender and if 
need be from his lejima. By coming in procession and placing their staffs 
before the entrance to his compound or his house, they can forbid 

two matrilineal groups from one of which the Obot Lopon is selected (see 
Table II, p. 308). 

' The estabUshment first of a Warrant Chief's Court and more recently of 
Native Authority Councils and Courts in the Yako area have reduced the judicial 
and executive authority of the Yabot, but civil disputes, especially those between 
kin groups, are brought to them and they still intervene against ritual offences. 
See my paper, 'Government in Umor', Africa, xii. 2, April 1939, pp. 129-61. 

^ Their power of action in the public interest may be exemplified as follows : 
it is the duty of the Speaker acting on behalf of the Yabot to quell serious 
disorders arising from civil disputes and summon the offenders before them. 
Anyone alleged to have caused a fire in the village would be tried by the Yabot 
and if guilty pay them a fine of a cow and 10 brass rods. For abortion, as a grave 
ritual offence believed to impair fertility in the community as a whole through 
action of the ase spirits, they impose a fine of a cow which is sacrificed and eaten 
in a rite of expiation at the shrine of the offender's matriclan. 


a person's entering or leaving the dwelling until the fine imposed for 
an offence has been paid. If their decisions are flouted one of the men's 
associations is authorized by them to despoil the offender and his 
close kin. 


It will be appreciated that, with the tracing of distinct lines of descent 
and the formation of groups of kin through father and son on the one 
hand and through mother and daughter on the other, every Yako will, 
through parents and children, have kin to whose lineal groups he or she 
will not themselves belong. Such will be the matrilineal kin of one's 
father and those of a man's children and the patrilineal kin of one's 
mother and those of a woman's children. The children's kin of the other 
group will, of course, be those with whom affinal ties were established for 
the parent at marriage, but the relation is felt as distinct from, or at 
least as a special reinforcement of, affinal ties as such. Because they are 
kin groups of one's own closest kin, mutual consideration and helpful- 
ness second only to that expected between patrikin and matrikin is felt 
to be meet. A man can, for example, seek assistance from them or from 
the elders of their clans if he is in need of additional land to farm or of 
raphia supplies for house-building. More specifically both men and 
women are under obligation to visit and bring gifts at the funerals of 
members of these patri- and matriclans, occasions when many of them 
are assembled and sympathy and support can be expressed at a time of 
loss and grief. For this relation, as will be seen, the Yako employ the 
specific and reciprocal term okpan (pi. yakpan). It is, it should be 
emphasized, not a relation within and between groups, but between 
a person and groups of kin to which he or she is individually linked by 
the cross-ties arising from patrilineal and matrilineal kinship. 


The kinship terminology of the Yako clearly expresses the dichotomy 
between matrilineal and patrilineal affiliation as well as the unity of the 
kin groups so formed. But, save on formal (ritual or periodical) occasions 
and when there is emotional stress, kinsfolk below the grandparental 
generations are usually addressed by their personal names. The terms dis- 
cussed here are therefore most frequently used for reference (see Fig. 15). 

The terms used for parents {uwo = (my) father; jtmka = (my) 
mother)^ may also be employed in a classificatory sense when referring 

" The terms for parents in Yako are as follows : 

my, our father, uzco, — mother, 7nuka 

thy, your father, awo, „ aka 

his, their father, yate, ,, yaka 

my, our fathers, buwo, — mothers, bamuka 

thy, your fathers, bawo, ,, ba'aka 

his, their fathers, bsyate, ,, bay aka 


to collateral kin of their generation if they are of the same sex as well as 
the same lineage as the parent, i.e. to male patrikin and to female matri- 
kin of this generation. Thus father and father's brother and father's 
father's brother's son are all referred to and may be addressed as uwo, 
but a mother's brother will not. Similarly, mother's sister and mother's 
mother's daughter will be called muka, but father's sister will not. These 
terms can also be extended to all men immediately senior in generation 
in one's patrician and to all women senior in generation in one's 

Step-parents are not kin. Another wife of one's father is referred to as 
such, yanen wuwo, and a stepfather is known as uwo keplaku (father in 
daylight), a term which conveys the element of quasi-paternal authority 
and responsibility he may have as head of the domestic group. 

A mother's brother, as are all men immediately senior in generation 
in one's matriclan, is wenyaka wamuka (child of mother of mother). 
A father's sister, like all women immediately senior in generation in 
one's patrician, is wenyaka wuwo (child of mother of father). 

Among siblings the maternal link has precedence whereby full and 
maternal half-siblings are designated wenwomuka (child of my mother, 
pi. benbomuka), while only paternal half-siblings are called wenwuwo 
(child of my father, pi. benbowuwo). 

Patrikin and matrikin junior in generation are, however, merged and 
may all be referred to individually and collectively by the term wenomi 
(pi. benomi) which is more specifically used for one's own child. 

The paternal grandfather is called uwo otam (old father) and the 
maternal grandmother muka otam (old mother). The paternal grand- 
mother may also be called muka otam if the grandchild has been much 
with her, but she is more usually addressed and is always referred to as 
father's mother (yaka wuwo), while the maternal grandfather, of whom 
a child usually sees little, is never called or referred to as 'old father', but 
always descriptively as mother's father (yate womuka). These terms are 
now, however, applied to other kin of the grandparents' generation. 
Those who are male patrikin and female matrikin are merged with the 
father's and mother's generation as buwo and bamiika respectively; 
female patrikin and male patrikin are addressed and referred to by terms 
which stress lineage membership but ignore generation level, while only 
descriptive terms or in some cases a reciprocal term (see below) are 
applied to those belonging to neither of one's own lineages. Here, again, 
the patrilineal and matrilineal ties are emphasized throughout. 

One's own grandchild is called wetiwawen (omi) — child of (my) child — 
without distinction between children of sons and those of daughters, 
although they will be members of different patri- and matriclans. But 
this term is not extended to collaterals of this generation ; the latter as 
kin junior in generation may, as indicated above, be merged with the 
generation of one's children as wenomi. But they may also be referred to 


according to affiliation by the merging terms for patrikin, for matrikin, 
or for the members of one's father's matri- or mother's patrikin (see 

More inclusive classificatory terms which further ignore generation 
differences are more usually employed in referring to unilineal kin 
outside the elementary family. They are wenuwo (pi. henuwo), which 
refers to all members of the patrician, and wenamuka (pi. benamuka), 
which correspondingly refers to all matrikin. These terms are not, how- 
ever, applied individually to one's own father, paternal grandfather, or 
grandchild, or to one's own siblings, children, mother, or maternal 
grandmother, save when included in a wider group of unilineal kin. 
They combine the roots for child and father, and child and mother 
respectively, but are, it should be noted, distinct from those used 
specifically for siblings — wenwomuka and wenwuwo — given earlier. 

A single reciprocal term {okpan, pi. yakpan) is used to express the 
kinship ties between a person and the father's matrikin, the mother's 
patrikin, the matrikin of a man's children, and the patrikin of a woman's 
children, i.e. the members of the unilineal groups to which one's parents 
or children belong but to which one does not belong oneself. Thus, if 
ego's father is a member of the Yanyo matriclan, ego is an okpan of Yanyo, 
and they are yakpan to him. If his mother was born in Anedja patrician, 
he is also an okpan of Anedja and they are yakpan to him. 

The patrikin, matrikin, and the two sets of yakpan comprise all the 
persons who are grouped together under broad classificatory terms. 
There is no single word for kin as such and remote cognates who fall 
outside these classes can be and are designated when occasion arises 
only by compound descriptive terms which state the genealogical links 
(D in Fig. 15). They are, like other persons, addressed by their proper 
names. The terms applying to wide classes of kin are not customarily 
used in address, since, as mentioned above, both patrilineal and matri- 
lineal kin other than own parents and grandparents are addressed by 
personal names usually prefixed, in the case of those senior in generation, 
by the terms tata (for men) and ma (for women). These last terms like 
wen (child) are expressive of generation difference rather than of descent 
itself and may be used in addressing all elders. 

In addition to the terms for parents, children, siblings, grandparents, 
and the classificatory terms for the various groups of unilineal kin with 
which a Yako is related, descriptive terms are also employed to designate 
the kinship links between persons by combining the roots for child 
(wen), the father (yate), or the mother (yaka). To these may be added 
where required an indication of the mode of descent: borne by (a female) 
— zvomani — or begotten by (a male) — wdmponi — together with the sex 
of a person: male — wodo'oddm, or female — wodoyanen, compounded 
from the terms odam (male) and yanen (female) plus kin. In these fuller 
descriptive terms, precedence is again given to matrilineal descent in 


that the maternal link is used where, as among full siblings, there is both 
a matrilineal and a patrilineal relation. We have seen that children of one 
mother, whether full or half-siblings, are all zvenwomuka to one another, 
while only paternal half-brothers and sisters are wenwuwo. Similarly, 
where a relationship between patrilineal kin is to be more precisely 
indicated by the use of a descriptive term, the link is traced back to the 
mother, not to the father of the siblings from whom they are respectively 
descended. Thus a father's sister is designated wenyaka wiiwo zvodoyanen 
(child of mother of father: female). Grandchildren can be distinguished 
as wenzvawenzvamponi (child of child I begot) by a grandfather and 
wenwawenwomani (child of child I bore) by a grandmother. They too 
may be differentiated in sex as zvodo'odam (male) and woddyanen 


It will already be appreciated that marriage, with the establishment 
of a new household and family, is the concern of several distinct sets of 
relatives, of the matrikin and patrikin of the husband and those of the 
wife. In addition to the direct affinal relations of each spouse with kin 
of his or her partner, certain of these relatives also undertake obligations 
to one another. But the roles of the matrilineal and patrilineal kin of the 
husband and those of the wife are not symmetrical. The domiciliary 
factor, with patrilineal succession to house sites and land, is dominant 
on the husband's side and his patrikin are most concerned in the 
marriage. His children will be members of his patrician but not of his 
matriclan. On the woman's side, however, the dominant aspects both of 
fertility and of the responsibility accepted in receiving marriage pay- 
ments are felt to be preponderantly a concern of the matrilineal kin. Her 
children will be members of her matriclan, they will be the matrilineal 
successors and heirs of her brothers — her own mother's sons. The links 
of the children with her patrician will, like their links with their father's 
matriclan, be limited to occasional acts of kindness and the ceremonial 
expression of friendship and sympathy due from them as yakpan. 

During betrothal and the marriage ceremonies, indeed until she 
actually takes up residence in her husband's dwelling-area, which is 
generally subsequent to the birth of her first child, a wife has little 
contact with and no formal obligations to her husband's patrikin. Later, 
when resident with the patrician of her husband and children, she is 
likely to be a close neighbour of her father- and brothers-in-law with 
whom she will in favourable circumstances be on easy terms, being 
treated as a daughter and sister. Apart from her mother-in-law, she has 
much less to do with her husband's matrikin, although these will be 
yakpanen {yakpan people) of her children. 

For a man, on the other hand, the affinal relation of major significance 
' Affinal terms are discussed in a later section; see p. 328. 



is that with one or more of his wife's senior male matrikin. During the 
betrothal he is as a suitor dependent on the goodwill of his future wife's 
parents. But once the marriage payments have been accepted and, more 
particularly, when his wife has come to live with him, it is her mother's 
brother, or one of her own elder brothers acting in his place, who, as her 
closest senior matrikinsman, takes responsibility should serious disputes 
arise and the marriage be threatened or actually dissolved.' 
/ During the period of betrothal {keplo kesi) before the first marriage 
of a girl, which may last from one to two years, the suitor is obliged to 
perform specific services for his prospective bride and her parents and 
to make them gifts. He will be asked to help and to bring his age-mates 
to assist in any considerable tasks that the father-in-law undertakes, such 
as the lifting and tying of yams at harvest, building, and repairing houses. 
And it was formerly customary not only to bring the future father-in-law 
a daily calabash of palm-wine but also to carry water for the mother-in- 
law, a task which otherwise is woman's work. During the annual lehoku 
rites, at the time when parents give special food and ornaments to their 
children, a suitor should make gifts not only to his betrothed but also to 
her parents. 

In the ceremonials and exchanges of marriage itself there are no direct 
relations between the kinsfolk of the bride and those of the groom. Indeed, 
a most remarkable feature of Yako marriage ceremonial is the emphasis 
on the relations of the chief participants each to their own friends or 
age-set mates, who need not be and often are not kin, and the lack of 
emphasis on their relations to either kinship groups or particular circles 
of kinsfolk as such. 

Apart from later rituals at her matriclan and patrician shrines, when 
the bride is pregnant, the clans of the bride and the groom have no cor- 
porate relation to the marriage ceremonies. The patricians of the bride 
and groom have an interest in maintaining the exogamic rule which 
would be sustained by the intervention of the Obot Kepun if an infrac- 
tion was threatened, but they otherwise take no corporate action con- 
cerning the marriage. Nor does the matrilineage of the bride, although it 
has, as will be seen, an interest in the maintenance of the marriage once 
established, participate in its ceremonial recognition. It is only after the 
last wedding-feasts which the bride's parents severally give to their own 
personal friends, particularly their age-mates, that her father should go, 
or send a message to, his wife's brother as the closest senior male among 
the matrilineal relatives of the bride, informing him that the wedding 
feasts have been concluded, that the marriage payment has been made, 
and that he is ready to hand over the customary portion to him. 

In considering the provision of the marriage payment {libeman, 
marriage money), which formerly consisted mainly of brass rods, but is 

' For an account and fuller analysis of betrothal and marriage see my Marriage 
and the Family among the Yako in South-Eastern Nigeria. 


now, apart from a few token rods, handed over in Nigerian currency, 
a distinction must be made between the first marriage of a youth and the 
later marriages of older men. For the latter the responsibility for pay- 
ment is usually entirely the husband's, although he may often borrow 
from matrikin. But for his first marriage, while labour services are made 
by the youth himself, his father normally takes the main responsibility, 
although his mother's brothers and often more distant kin on both sides 
will make contributions varying according to the individual circum- 
stances. If the mother's brother is the foster-father of the groom he 
replaces the father as the main contributor. The marriage payment 
should properly be handed over by age-mates, i.e. age-set friends, of the 
groom to the father or foster-father of the bride when she returns to 
the village after the clitoridectomy rite which should form part of the 
marriage ceremonies. Frequently only a portion of the payment is 
transferred on this occasion and the balance is later handed over 
informally in a number of instalments. By the wife's father and her 
mother's brother the marriage money is regarded as provided by the 
husband. It is he, not his father or any other kinsman, who claims its 
return and the fact that it is the groom himself who offers it is empha- 
sized in the transfer. 

Transfers involved in marriage payments are not directly reciprocal 
between the households or kinsmen involved. A father receives for his 
daughter much less than he gives for his son. A mother's brother 
receiving for a bride does not give correspondingly for her brothers. 
There is no doubt that it is the transfer of the libeman which gives the 
husband legal rights to his wife's services and to the social fatherhood 
of the children born to her during the marriage. The retention of a 
portion of the marriage money by the father, like its initial transfer by 
the groom to him, is not regarded by the Yako as an indemnity for any 
economic value of a daughter. It is to be explained by the fact of paternal 
authority in the household. A father has the established right to control 
the domicile of his unmarried children. No daughter may go or be 
removed from his household and parental authority without his consent. 
The acceptance of the marriage money and the retention of a portion of 
it mark the curtailment of the father's parental authority and signify his 
consent to his daughter's departure from his household. His share of the 
marriage payment is a consideration for that consent. 

But the husband is securing at marriage not merely the curtailment 
of paternal authority over the bride. He is also placing the bride under 
a positive obligation to perform certain services. Should she fail, com- 
pensation cannot be secured from the bride herself and the obligation 
to return the marriage money to the husband devolves not on the father 
but on those who in Yako law have responsibility for ensuring the fulfil- 
ment or compensation of all personal obligations, namely, the matrilineal 
kin. It is on account of this potential obligation that a matrilineal 


kinsman is considered entitled to receive from the father the greater part 
of the payment. It should be observed that the obligation to return what 
was received relates to the whole of the marriage money, including the 
portion retained by the father. That the kinsman who undertakes it and 
in consequence receives the greater part of the marriage money should 
be the mother's brother is consonant with the Yako principle of matri- 
lineal transmission. As the closest matrihneal kinsman senior in genera- 
tion to the wife, compensation would be claimed from him if she should 
fail in her marital obligations, while her matrihneal brothers are suc- 
cessors both to this obligation and to the property of the mother's 

Thus the Yako marriage payment is variously related to the several 
interests involved. It expresses the value of a wife to her husband and 
confirms his intention to maintain a stable union. It rewards the bride's 
father for the renunciation of his parental authority while establishing 
his consent and that of the matrihneal kin to the assertion of marital 
rights by the husband. It reinforces the responsibility of matrihneal kin 
for ensuring that the wife shall fulfil her marital obligations by giving 
the closest senior matrihneal kinsman of the bride a pecuniary obligation 
with regard to her good conduct. - 

The relations of a widow to her deceased husband's kin depend on 
whether she has grown sons. There is neither leviratic marriage nor 
widow inheritance and a widow, if she is young, is free to remarry where 
she and her own kin wish. Only if she is widowed while suckling a child 
has she a right to remain for any length of time in her late husband's 

' The customary obligation of the bride's father to hand over the greater 
part of the marriage payment to her mother's brother or other close matrilineal 
relative has, as already mentioned, been increasingly disregarded in recent 
years, and in a third of the cases in a sample studied it was all retained by the 
father or someone acting in his place. This was particularly common where 
the father was a rich man (osu, pi. yasu), and was not a matter of cupidity but 
rather an assertion of paternal authority and also of wealth and independence. 
For a mother's brother can rely on his matrilineal kinsmen to assist him in 
providing the means of returning a marriage payment, but a father cannot, since 
the marriage is not the concern of the father's matrilineal kin ; nor do his daughter's 
children belong to his patrilineage. Thus a father who retains the marriage 
payment faces the possibility of having to refund a large part of it, perhaps 
many years later, entirely from his own resources : a risk a poor man would be 
more reluctant to assume. 

^ The role of mother's brother in relation to a girl's marriage and other 
juridical and ritual obligations should properly be assumed by the senior among 
the brothers (including maternal half-brothers) of the mother, and where there 
is no true brother the senior matrikinsman in the senior collateral line should 
act. But, while this is fairly strictly adhered to in ritual contexts, in situations 
requiring personal initiative and responsibility the rights and duties may be 
assumed by a younger brother or other matrikinsman who, through personality 
or possessions, has greater prestige. Thus marriage payments are not infre- 
quently nowadays handed over by the mother's eldest brother to a wealthier 
junior who is more ready to accept the contingent responsibility for repayment. 


compound and to be given farming facilities by one of his patrikin. She 
may or may not prefer to rejoin her parents or other relatives, but in any 
case she cannot remarry until she is ready to wean the child ; she is there- 
fore dependent on either the late husband's or her own relatives for 
farming facilities and food-supplies. On the other hand, a widow has no 
claim to the livestock and personal possessions of her late husband, 
although she may, at the distribution of his goods during the funeral 
rites, be given a cloth and a few other small objects by the matrilineal 
relative of her husband who is the chief heir. A young widow usually 
remarries within a year or two. She leaves her late husband's compound 
after the harvest following his death and joins the household of either 
her parents or a senior matrikinsman. On the other hand, an older widow 
who has adult sons living in the patrician area, and usually in the same 
compound as her late husband, will often remain there and join the 
household and farming unit of one of her sons, who is, of course, her 
closest matrilineal kinsman. 

It is rare for a Yako to take the initiative in divorcing a wife, but if he 
determines to do so he informs her parents and the matrikinsman who 
received the marriage money that she must go. As a last resort he will 
refuse to admit her to his farm in the coming year. But, unless the 
woman's conduct has been outrageous, the husband forgoes all claim 
to a return of the marriage payments and the woman is free to marry 
elsewhere. Divorces arise far more frequently from the voluntary depar- 
ture of the wife from her husband's compound, and the husband then 
becomes entitled to a return of his marriage payment. If she goes to live 
with another man, her new mate becomes liable, not only for the payment 
of damages to the husband on account of the adultery, but also for 
refunding to her matrilineal kinsman the amount of the marriage pay- 
ment which the latter is required to restore to the husband. When, 
and only when, this payment has been made, is the divorce accepted 
and the adulterer recognized as a second husband. In recent years, 
however, deserted husbands, finding it difficult to