You are on page 1of 402

S Q O a O Q S Q 0

Classic Reprints in Anthropology

Bruce Kapferer and Clyde Mitchell
This series reprints classic texts in ethnography and methodology
which have exerted a major influence on anthropological thought.
The series w ill span a variety o f traditions and w ill also include
translations. Each volume w ill be introduced by a preface placing it
in the context o f current anthropological debate.. The overall aim is
to re-sensitize students and researchers to key issues which have
shaped the discipline and continue to have relevance today.
ISSN: 1354-3601

Schism and Continuity in an

African Society
A Stud y o f N dem bu V illage Life

V . W . T am er1,

Oxford W ashington, D .C .

First published in 1957 by the University o f M anchester on behalf o f

T he Institute for African Studies, University o f Zambia.
Reprinted in 1964, 1968 and 1972 by the University o f M anchester and
in 1996 by:
B erg
Editorial offices:
150 Cowley Road, Oxford, 0 X 4 1JJ, U K
13950 Park Center Road, Herndon, VA 22071, U S A

Edith Turner, executor for Victor Turner, 1996

. A ll rights reserved.
N o part o f this publication may be reproduced in any form
or by any m eans without the written perm ission o f Rerg.
Berg is an imprint o f Oxford International Publishers Ltd.

lib ra ry o f C ongress Cataloging-in-Publication D ata

A catalogue record for this book is available from the Library o f

B ritish lib r a r y Cataloguing-in-PubUcation D ata

A catalogue record for this book is available from the British Library.

ISB N 1 85973 110 4 (Cloth)

0 85496 282 4 (Paper)

Printed in the U nited Kingdom by W B C Book M anufacturers,

Bridgend, M id Glamorgan.


P reface to


1996 Edition


B ruce Ka p f e r e r .......... v ii

P reface to


1972 E dition


M ax G l u c k m a n ..........

Preface t o


1968 E dition ..................................................xxi

Preface to


1957 Edition .............................................



A cknow ledgements ............................................................ xxxiii


H istorical and Ecological Ba c k g r o u n d ........................ 1

T he V illage : T opography and D emography ................. 34
T he S o c ial C omposition o f the V il l a g e ......................... 61
M atrilineal D escent : T he B asic P rinciple of
V illage O rganization .......................................
M atrilineal S uccession and the D ynamics of
V illage Intrigue ..................................
V illage Fission, S lavery an d S ocial C hange ............... 169
V arieties o f V illage Fis s io n ............................................. 204
T he S tr u ctu r al I mplications of V irilocal
M arriage within the V il l a g e ..................................... 234
Political A spects of K inship an d A f f in it y .................. 258
T he Politically Integrative Function of R itual . . . 288
T he C hieftainship............................................................... 3 18
Po s t s c r ip t ........................................................................... 328
B ib lio g r a ph y ....................................................................... 332
I n d e x ...................................................................................... 337
A ppen d ices ...........................................................(end of book )

Map 3 and the Appendices have been omitted from this reprint.



IC T O R Turners Schism and Continuity is among the more
outstanding monographs that came out o f the work o f M ax
Gluckm ans Manchester School o f anthropology. T his group
developed from a research tradition established by Gluckm an at the
then Rhodes-Livingstone Institute for Social Research (now Institute
for African Research and part o f the University o f Zambia). Gluckman
succeeded the former Director (Godfrey W ilson) in the closing stages
o f W orld W ar II and set up a programme for the thorough
ethnographic study o f the region which included what are now
Malawi, Zambia, and Zimbabwe. W hen Gluckman was appointed
to the Chair o f Social Anthropology at the University o f Manchester
in 1947, the Rhodes-Livingstone Institute continued as the research
base for the school o f anthropology that Gluckman was to develop.
Those who gathered at the RLI and who largely built the fieldwork
and theoretical perspectives that made Gluckm ans M anchester
School an important development within British social anthropology
included Elizabeth Colson, Clyde Mitchell, B ill Epstein, John Barnes,
M ax M arwick, and Victor Turner. These...scholarsJucLtheir. own
distinctive ways pursued Gluckmans_insistence_that.anytheQretical
and!cQnceptuaL.undetstanding.Qf social forms and their definingideas
must be. the studyof social. practice. This .approachwhich
Gluckman him self elaborated from the example, of. Evans-Pritchard
became knownJbrQadly as method. The line that.Gluckman.encour.aged was also-informed
by a m ixture o f Durkheimianism (largely through the interpretations
elaborated by Radcliffe-Brown) and the ideas o f M arx and Engels.
O verall,.. the theoretical and m ethodological halim.ark o f .the
Manchester traditional concentrated on practices which were viewed
as revealing the inner logic o f
or more generally, these days, as culture.
It must be stressed that the Manchester position was innovative.
Many in British anthropology at the time regarded the Manchester
anthropologists to be m arginal to the mainstream and the


Preface to the 1996 Edition

Manchester/RLI group valued themselves in this way. They were

In Central A frica they became sharp critics,.Qf..CnlQniaLilule^and
werecfiticized in the press. Gluckman him selfwas officially refused
entry into the region (and other areas o f colonial rule such as
Australia-controlled New Guinea). Others in the Manchester group
were similarly prevented from continuing their research. Many o f
the Manchester/RLI group were directly active
England and elsewhere (although; Gluckman was .more suhduedLin
/ such affiliation, a fact that exposed him, to ,friendly jibesrom Jiis
colleagues and students). The Manchester/RLI group, in th e fifties
and^early sixties carried their generally-radical attitudes-into their
anthropological, practice. It was a distinctive feat.ur.e_ o f ..the
Manchester. School. They..piQneered perspectives that-only became
widely accepted much later in the seventies and eighties. The-.ev.ents
o f 1968 in Paris were arr -important dntellectuaL-watershed for
anthropology as. for o.thet_academic. disciplines.
The. Manchester emphasis on practice was sometimes compared
with .Firth,.s. contrast between..organization and-structure. They were
very_different. Firths observation largely.,th.e fact
(injmy view fairly trivial) that the representations that anthropologists
ancLthe subjects o f their observation make o f their forms, o f life
diverge from, what they actually do...He-stresaed-the.-impos'tan.ce o f
exploring the.actual- luid-.oxganizations_oactivit.y...Leacb_pmbably
carried this notion further.(a Malinowski)
as. did Barth. The developments within.the. Manches.tet.Scho.ol.were
more profound. They were not contradictions_within
cultural, systems (Leach)_ca^in the-dynamic&-QindividuaLraional
choice (Barth)aspects that_they.didnotdgnoresrhut.were,jdirected
more fundamentally to the.confcmdictionsjvithinhistcadcaL.,political
and economic forces o f a-globaI.nature .which_were^part-othe=wider
circumstance within which particular^humariupQpulations. struggled
to recreate or reproduce their social and. cultural ways o f life.
(Jonathan Friedmans critique o f Leach for overlooking the historical
forces should be noted.) The^Manchester. orientation,.was more
developed than many later and.more fashionable posfc-1968 ,attempts
at, a Marxist perspective in other major centres..of.anthropology in

Preface to the 1 996 Edition


England (for example, that at LSE where first then. Bloch
functionalist economism fact had n o n eed o fM arxo f Engels).
The originality and possibility o f the Manchester anthropology
perspective reached a milestone in Turners Schism and Continuity.
The argument extends from Gluckm ans lead, focussing on the
dynamics o f social conflict and its foundation within contradictions
at the heart o f systems o f relatedness through kinship and marriage.
This is the central problem atic that guides the analysis (the
contradiction between virilocality at marriage and a pattern o f
matrilineal descent and inheritance). The everyday conflict and the
course o f this conf lict was rooted in this contradiction which was
further complicated by the fact that the Ndembu o f the villages
studied by Turner were embroiled in larger historical changes effected
through Colonial Rule. Turner shaws.hQw..the-Contradictions.a the
heart o f Ndembu life became, in effect, more and more.irresolvable
in the co.ntext._of. larger, economic and. political developments. He
centres his account around the ambitions o f a key and in many ways
a tragic figure, Sandombu. Through the struggles ofSandom bu the
reader is lead into a remarkably vivid account o f the everyday life o f
Ndembu villagers..and,. in.Marxistvein,_how_they_come_tQ_partLcipate
in,the transformations o f their own world even as they_are caught
in structural processes that are ultimately beyond their..control.
Turher*s_metho.dological. innovations in Schism .and._Qo.ntijmity
overcame ..contradictions. ,in. social anthropology...betw een ..actororiented and.. structure-oriented .. perspectives.,... whereby, a
concentration on one aspect obv_iated_the other._The overcaming o f
this contradiction was. one .concern.oGluckmans-methodological
development o f situational analysis which Turner elaborated.and
transform ed. Broadly,, situational analysis expanded, beyond the
sociological use o f cases or instances from life to illustrateJarger
systems o f structure orinstitutions ofth e social order.that theanalyst
discernedthe case as illustrative. T he aim was-to..demonstrate the
intricacies o f larger processes oper.ating._within_the. dy.namics o f
particular events. Thus Gluckmart explored how a bridge-opening
ceremony in Zululand in Natal revealed the complexities o f a then
emergent system o f apartheid. He applied the method to the process

Preface to the 1996 Edition

of judicial decision in Lozi courts in Barotseland in Zambia. However,

it was Turner in Schism and Continuity who realized the full
possibilities o f the approach. N ot cjnly.dicl he demonstratexhe logic*
in the event, Turner (by taking a number o f events in_serlesJnvoiv ing
the same actors) was able to dem onstratehowparticipants.changed
and transformed the very structural. circumstance5_oLtheir_own
The general significance here o f Turners innovation should not
be missed. GLuckmans perspective opened out to...the.importance
o f considering how people themselves constructed their contexts o f
actionp.that. is,, they.had a ..role, to..pky=,in the and cultural realities.Tt_wasL^jmove.away-from.the,objectivist
position, .of .th e .anthropological, .observer, a in
Geettzjsense.well before the latter had formalized-the. idea. However,
Gluckman^and .his colleagues.
praetiee^c.auld.,not,dR fact,, escape .the-case,_ox .the^ an
illustration o f .sociaLpxinciples. that they devised.independently o f
participants.. Furthermore, .although,they stressed,an.attention to
the_change O F . systems of-social action, and-not their-tim eless
repetition, they could not break out o f a form o f analysis that was
more about how. systems. remained:the.same,rather_than. how they
changed or. transform ed. I note, that, this was also the .marked
difficulty o f other perspectives developing . in...British .social
anthropology at the timethose o f Leach and. o f Barth. Turners
analysis in Schism and Continuity broke out o f them ould. He did so
in a marked Hegelian manner.
Thus Turner analysed each crisis in the flux
o f Ndembu village life as a dynamic o f rupture-teachingJtowards its
overcoming in a new synthesis {resoiutiQn)_Qrorgani?ation-afjs_aciai
relations. T he wider .env ironment or political economic field o f an<i.intemal-to-Ndembu. village ,life (the
colonialpoUticaLorder,.capitalist economies-manifestin-urbanization,
labour,., migration, market farm*ng),^generated-forces ,that .were
expressed in the conflicts..that_Tumer_des.crib.e&_and.-which. the
villagers could not resolve. Howeyer,theireffQrts toJcontrolbsuch
forces involved villagers,..nonetheless,-.in_co.nstituting-dimensions o f
their changing universe and developing .and..elaborating original

Preface to the 1 9 9 6 Edition


cultural conception andpractice.

Coining, back tp the Hegelianism in Turners approach, he does
have a notion o f an ultimate higher unify. T his .is. estiablishejd in
ritual and the growth in importance .of rites,. such_as_the.Chihamba
curing rite, that cross-cut the social boundaries ofkin.gtQ d
village. The Chiham ba and its. political., importanceJs. the
unresolvable divisions and ruptures emergent n ot .jus.tL-fmm_.the
. contradictions underlying the traditional order. (thejcontradictipn
o f the m atrilineal andviriloeal principles) .butLinjemhracing.glob.aliy
connected political and economic changes. In an important.way the
Chihamba rite is generated-.and..reinvented-in_sudxjcircumsances.
It does not maintain the system, as a more. functianalistanaLysi&.might
have'it (e.g. Gluckm ans analysis o f the rites-.of-SwazLand-Zulu
kingship that were influential on Turner but from whose implications
Turner broke). Rather it is integral w ithin a. dynam ic.of Ndembu
social and political reconstitution.
The discussion o f the Chiham ba and other rites in Schism and
Continuity is preliminary to what became Turners key focusthe
symbolic processes o f rite. His later work, Chiham ba the "White
Spirit, is a wonderful elaboration on themes indicated in Schism and
Continuity. In.ihisiaterjgtudy
flight tP his. own creative
(literary, studies, Jungian and Freudian psychQ-analysis)_tQ_an
understanding, o f Mdembu symbolic processes anxLhow.they^extend
a comprehension o f the dynamics o f human far
beyond a particular cultural/histQrical cQntext...B.ritish anthropology
up to and beyond the work o f Turner is often characterized as static
British structural functionalism . T his common enough stereotype
can only be so if his work is ignored and attention is focussed on
what turned out to be the far more conservative and stultifying efforts
o f those centres o f anthropology located in London and Oxbridge.
I emphasize that it is by reading Schism and Continuity that a full
grasping can be achieved o f the radical direction that Turner was to
lead the anthropology o f ritual and symbolism.
Turner (and his_Manchester colleagues) are what might be generally
refexre.d, tO-as.dp.ractice theorists. Turner should be compared w ith
more recent developments in anthropology, for example, those o f


Preface to the 1 9 9 6 Edition

Sahlins and especially Bourdieu. T h e latter explicitly sets his

approach to practice in contrast to that developed from Manchester.
He chooses to distinguish his line from that o f Van Velsen (The
Politics of Kinship). Van Velsens study was conceived o f as a
development from that o f Turner. It took an individualistic strategic
choice direction. This was a possibility o f the Manchester orientation
but probably a retrograde step (I state this seif-critically because it is
also a direction in my own early work written in the Manchester
context). Bourdieus line on practice is also potentially individualistic
as evidenced in his free borrowing from the Am erican pragmatist
traditions o f symbolic interactionism and ethnom ethodology.
Turners analyses never hinges on a simple politics o f self-interest or
o f strategy and, in my view, extends beyond many o f the otherwise
positive and fruitful aspects o f Bourdieus work.
Bourdieu attempts a synthesis o f Husserls (also Heideggers)
phenomenology w ith a Levi-Straussian structuralism. His approach
is extraordinarily illuminating but his synthesis is not thoroughly
successful and the way human beings can shift their doxa or radically
alter the circumstances o f their habitus or habituated activities is
never clear. Bourdieu is committed to dynamics o f the reproduction
o f the same and is not oriented to the production o f difference and
originality. In this, I consider, Turner in Schism and Continuity and
in later work is probably more successful. He does not attempt to
force together tw o forms o f analysis (structuralism and
phenomenology) that in Bourdieus treatment at least appear to be
incompatible. Turners is an approach to practicean approach that
never relents on the density o f the cultural processes ingrained and
developing out o f practicethat indicates how new modes o f symbolic
comprehension and structures o f daily activity can be generated.
There is much else I could say about Schism and Continuity by way
o f introduction. For instance, here is an orientation that does not
seecultural/socialfarm sas^coherem ly.boundecL.systerns^Turner
concept oTsocial field^horrowecLfrQro the ^ c M pjsychologist. Kurt
Lewin (see Turners .later.The.Drjuxos_Q/_Aj^cti.Qn)==whic.h-CQncentrated on the dynamics o f social structuration w ith an open field o f
forces. Turner in his development was him self a generative centre

Preface to the 1996 Edition


for perspectives developed by others within the Manchester tradition

(e.g. Epstein, Bailey)* A lthough intellectually unrelated, Turners
unbounded field notion was reinvented* w ithin the contemporary
suggestions o f a postmodern anthropology. In Turner too we have
a sense o f the many voices o f Ndembu villagers as they participate
in the fashioning and refashioning o f their own existential realities.
Schism and Continuity is a central work in the history o f
anthropology. It is critical for a thorough understanding o f Turners
later writing. But it has far more than a historical significance. It is
exem plary o f the creative possibilities o f anthropological
ethnography, the centrality o f rigorous ethnography in the
anthropological contribution towards more general human
understanding. T he work should be read and reread as one example
o f how anthropologists might go about their research and develop
analytical understanding. T his is particularly so at this historical
moment in a general discourse w ithin anthropology concerning new
analytical directions and the role o f ethnography in demonstrating







O C IA L anthropology in its m odem form is o n ly about forty
years old, and during its short history it has changed ana
developed fast. For fo>m.theJdbmemhen..,anthrQpologists.hegan
to carry out intensive ileldrstudies,..they, have experim ented w ith
manylmethods: o f presenting their increasingly jrich .data. about
socisd life. D r. Turner attempts a n ew m ode o f presentation in
this book, but it is a m ode w h ich grow s out o f the history o f
anthropology since the W ar. H e has h e re co m h in ed .general
analysis w ith the individual, whatJLconsider_a..most
fruitful and iHuminating w ay.
W hen social anthropologists began to cope.w ith the. problem
o T handling their very, detailed data on the tribal societies o f
Africa, Oceania, Asia and the. Am ericas, they _seem_ first. tnJbaye
sought to establish that some kind o f system existed. in_these
societies. M alinow ski found the system in the interdependence
o f culture,.RadclifFe-Brow n found it in social structure. T he

among w hom I my se lf fell, continued this search for system. A s
I see it, our analyses abstracted a set o f regular interconnections
between various social relationships, ecological relationships,
modes o f belief, etc. From the tim e w hen Evans-Pritchard
analysed the Azande beliefs and practices relating to witchcraft,
oracles and m agic, to show that these constituted a rational
philosophy o f causation and philosophy o f morals, w hich w ere
related to a particular m ode o f social organization, we.Jhave
e3cbihi.ted_ th at, there, is ,a , system atic.structure in.-one. Jield o f
tribal life after another. T he evidence for the existence o f these
structures was m ainly qualitative, and consisted largely in illus
trations from different situations o f action in a num ber o f families,
villages, or political groups. These iUustrative,data-were-rarely
relatedtO-one_another. _In_r.esult,,when one,readsjeven the best
books o f this period, it is. n ot easy .to. put th.e. systeminto.. w orking
operation ,in actual life . In .order to. demonstratethe.existence
o f system at all, w e . discarded m uch o f the Hying, reality about
I consider, that this _was,a .necessary. stage_itrJhe_jdeyelcpment


Preface to the 1972 Edition

of_ our discipline. .. Faced.

form s o f social rektionsMps JvJiich_occ.urJii_t3ie,_tdbaL,sp.cieliesf
w eJ ia d first to produce a m orphology o f tKeir,,_.and
general analyses o f h ow their, systems- w orked,, to . exhibit, .some
stability w ithin constant , change .-ofpersonneL andrelations
between personnel. B ut I think all o f us have felt that m ore
could be done w ith the detail o f our data. It seems to me that
our systematic know ledge has now developed sufficiently fo r us
to achieve the next step. African Political Systems (1940), a col
lection o f eight studies, marked an im portant step in the systematic
study o f one field, and it was follow ed b y a series o f comparable,
but fuller, monographs. Y e t it is significant that political
studies since the W ar have, so to speak, put their systems back
into a particular history, w hile still aim ing at the demonstration
o f systematic interconnections : this is m arked in, fo r example,
Evans-Pritchards The Sanusi o f Cyrenaica and J. A . Barness
Politics in a Changing Society, an historical analysis o f the Fort
Jameson N goni. Some anthropologists, notably Schapera, have
even m oved fu lly into a m ethod o f w o rk m ore like orthodox
history, but still seeking for a systematic structure.
T he same change is likely to occur in the fields w hich social
anthropology has made peculiarly its ow n : for the problems
w e m eet in the field o f tribal politics and law , w e share w ith
political science, history, and jurisprudence. T he other fields are
the study o f domestic and village life, o f fam ilial and kinship
systems, and o f ritual religion, m agic, w itchcraft, etc. And
here Turner has made a significant contribution, follow ing leads
set, fo r Central Africa, b y M itchell, Colson and Barnes. Great
classics on domestic and village life, like Fortess books on the
Tallensi and Firths on the Tikopia, analyse fo r us the regularities
they find in a variety o f actual situations and relationships, and
the interdependence between those regularities. Theirs are mag
nificent analyses. B.UJL.they.use._ihe._m.ethod_o-appEopriate,
illum inating illustration, and hence, in a w a y , o f chance illus
tration. W e cannot from their analyses,w o rk o u th aw iaiiy..o n e
group o f people lived, at. a.-particular. time-.and in a_particular
place, w ithin this social structure and using_thes.eL_xusioms.
Turner has used a different (node o f iilustration, w hich I consider
deepens the understanding w hich he givesL_us .o f Mdembu. tribal
life. Indeed, it w ould be more_acc.uratej:o.sayLthathe^abandons

Preface to the 1972 Edition


x v ii

illustration. altogether,, to d evelo p w hat .is,. for . Africanstadies, a

new .m ode ofan alysis. H e first gives us a system atic outline o f
the principles on w hich N dem bu villages are constructed, and
measures their' relative im portance w ith unusually adequate
numerical data. Then he takes the history o f one village through
tw enty years to show how these abstract principles have operated
through that history, w ithin the chance occurrences o f illness,
death and other misfortune, o f go od luck, o f individual tem
perament and am bition, and finally o f the m ajor changes w hich
have resulted from British overlordship. H e thus shows us h o w
certain prindples_.a_orgamzation_and.xem im.daminmt~values
operate through.both.schism s andjceconciliations^ the
individuals ...and. gronps -varied
princ^es-andjalues_tCL-their.own eiids. Turner him self w ould
not claim any particular points he makes as original, for m ost
o f them he has taken individually from the w orks o f his predeces
sors. B ut this use o f the detailed case-study, in the background
o f general system atic analysis, com bines w hat he has taken over
into a pioneer study. The late B uell Q uain in his Fijian Village
attem pted this task, but his death perhaps prevented the final
pulling together o f his analysis, and it did not quite com e off.
O ne thing I can do, in this forew ord, is to answer a possible
objection to this m eth o d : h o w is the reader to assess that
Turner has selected a * typical * village ? This question is partly
answered in the careful num erical analyses o f m any villages
w hich precede the 4village-study \ Secondly, I am in the
privileged position o f having heard Turner present analyses o f
other village histories; and hence I can vouch that fo r all its
uniqueness, the village o f 4M ukanza is * typical * o f N dem bu
villages. Turner hopes to publish separately studies o f some o f
these other villages ; and collateral validation fo r this present
analysis w ill also com e out o f the study o f N dem bu ritual w hich
he is n ow w riting. For he plans to use the same method o f
analysis in studying ritu a l: and again I consider this w ill m ark a
significant advance in this field o f research.
heicalls 4the social, .drama %-r-one o_a series jofocriseswoccurring
in (the history o f the village, when, either a quarreLhetween some
o f ih e inhabitants, or a m isfortune ascribed by._foe_pejoplje.and b y
divination to ancestral spirits or.iorcery,.precipitates..threats to


xv iii


Preface to the 1972 Edition

the unity o f the village. T he village._asajvvhole, and itsjie ig h hours, as w ell as com|?Qnent groups^_witHn.Jthe-j^Ulage,^tEy. to
useLdifferent_forms..ofLredxess^to_meet_these_j;hreats. Turner
argues that when the confoc-emexges._frani...the .opposed-interests
and claims o f protagonists acting under a single social-principle,
jud icial institutions invoked to meet the crisis, fo x ajratiDnal
attem pt can be made to adjust, claims, w hich are similarly-.basedB u t w hen claims are advanced under different..sociaLpm iciples,
w hich "are inconsistent w ith one .another..eventothe-poin t o f
contradicting one another., there can be no rational settlement.
Here recourse is had to divination of.sorcery or ancestral wrath.
causing...misfortune..; and. ..ultimatelyto...a-.ritual reconciliation
whih can reassert a ll.the values-heldJby. decent..Ndemb.u,.Junder
the pretence that harm ony is restored withm . .those, yalues. H e
shows that after .this situation_has..occurredr there~is-a..teraporary
resp ite; but the ...deep, conflicts-between-groups-and-m dividuals
in- the. village continue_Thro.ugh...the..intervening,com parad ve ly peaceful .period, struggles-.continue-tilI-th.eyprecipitate a
new crisis. Each crisis..m arks.jhe culrrdna.riojX-o.anerioxd o f
altering, alignments o f pow er and shifts ,of. aUegiance-withinjthe
village though matrmneal attachment, is alw ays-dom inant in
the end. . . The crisis itself, through four stages which, hexarefiilly
delineates,. is... tem porarily . solved ; .. but. again .. thisinvolves a
definite shift in the villages internal balance-of pow er.
I have sketched som ething o f Turners m ethod : he can speak
better for him self.. B u t I m ake a few points to support him .
First, I hope no one w ill turn aw ay from his analysis in dislike
o f the phrase * social drama *. Several o f us have tried, w ith
Turner, to find another phrase w hich is less likely to meet objec
tions : w e have failed to, and he w ould be grateful for sugges
tions. Secondly, I w arn readers that it is not easy to fo llo w the
com plicated story o f this village, w ith its com plicated internal
genealogy and its im portant links in other villages. Turner has
done w hat he can to help the reader, w ith a main genealogy and
periodically w ith subsidiary genealogies o f those involved in each
drama. B ut the labour o f learning w ho the characters are, and
h o w they are related to one another, is w ell w orth w hile. The
reward is equal to the labour. I m yself found that I soon got
the main characters clear, and the rest follow ed easily. Then I
became absorbed in the story itself, and particularly in the tragic

Preface to the 1972 Edition


story o f Sandombu m oved n ot o n ly b y his personal am bition,

strong though that was, but also m oved by the pride o f his
lineage, to struggle for the village headmanship w hen everything
was against him . W ith his generosity and capacity for affection,
he fought w ith his quick tem per and the curse o f his sterility, to
achieve a headmanship ; and Turner shows clearly out o f his
general analysis, that Sandombu was doom ed to lose, and on
the w ay to defeat to incur a reputation fo r sorcery. H e is a
tragic figure, indeed. B ut w e learn too that his rivals could not
wholeheartedly take advantage o f his weaknesses ; fo r when they
had overcom e him , they in turn w ere m oved by guilt, for,
under the dom inant value o f N dem bu, was he not bom from
the same w om b as they ?
Into this story,Turner..has_woven. a.striking generaLanalysis,
o f jschism o f groxms . and .relationships,^.and...o....c.onrinmty o f
society .md_principle-and_Js^ue_ahov^e..jthe.-Schism=^a.xoniinuity

o f a * com m unity o f suffering , fo r it. is..misfox.tuneLJwliicli

assembles the cult-groups w hose rituals dram atize_foatxontinuity.
Thu^incidentally. as the. story .unfolds _we leam wiiaLareJ^Ideinbu
beliefe and customs.^c^kmg.~i^
social life,.jh ere, they both control people, and are exploited
b y people. .Custom s, and.. heKefeate.jreal,system atizedjthxough
social relationships : they are n ot mere adjuncts._Incw orking
all this, out, Turner, also-t-and this is an achievem ent brings in
the changes o f values m d. prindiples ,iQ fJ.Q rganim ttonjdbich are
em erging out o f British overlordsnip^the. developm ent.o f w ageearnjngand cash-cropping,Jthe estabUshmentopeace,_the^3kiIlin^
out of-gam e w h ich has struck at the dom inant m ale v alue o f
hunting. In all this I see a contribution to our know ledge o f
A frica, and to anthropological theory and m ethod. It is achieved
in a book w hich w ill fascinate anyone.
M a x G luckman
U niversity of M anchester,
June 1956



P R E F A C E T O T H E 1968 E D I T I O N
T is n ow m ore than a decade since Schism and Continuity was
first published. Seen from this peopectiye it^emeEgesjclearly
as ja transitional book between the prevalent structural-function
alism o f British anthropology in the 1940s the period in.w hich
I received m y training and the processual analysis ofltheuisxSos.
W hatever influence it. may. have had seems to have lainJnu the
rapidly advancing sub-disciplines o f diachronic m icrorsociology
and . .m icrorpolitics, Y e t it has-som etim es..heen._fbrgo.tten b y
those., caught , up in the first...enthusiasm, for processualism *
that.* process tim ately bound up _with A structured ancLthat
ati^de_q.uate_.analysisosaci_aLli6b.nece^sitates-a-ngQrousL-consideration o f the relation betw een them ._H istorical-hindsight
reveals a diachronic profile, . a.. * tem poral structure. -in. events,
and this_structure_ .understood in isolation. ftom -th c
series o f synchronic profiles w h ich make up the . structure o f a
social field at every significant point o f arrest in the tim e flow .
PrOcessuaLstudies-can... nexerJbe.. the.jnegatiotL-of. structuralism;
rather do they put the discoveries of. structuralism, to-new.use, in
In^HsL.intEoduction .tothe Jihstedition, o f this . hook..P.rofessor
G luckm anJiascalledaltentiontothe^ w ayT n-w hichxasem aterial
is(used to. facilitate, w hat he h aslater described asd_the-intensive
study-.o-the-processes-.o.CQntrol in..a.limited_.are.a..ojocial life
view edlover_ a. periQd.of_tim e (Introduction to The Craft o f
Social Anthropology, 1967 b y A . L . Epstein, London: Tavistock
Publications, x vi). H e contrasts this extended case m ethod *
w ith the m ethod o f apt illustration. . (of. abstract.. structural
principles), and suggests that its systematic.use. woulcLdeepemour
understanding o f law an d m orality. B ut it was not solely w ith
eyen .at _the_time_of w ritin g the book,, but .w ith. a_difEerent kind
o f analysis. In form ulating, the notion o f <social drama. .I.had in
mi_nd the explicit com parison.ofxhe temporaL&tructurejofcertain
types o f social processes w ith that o f dramas..on .the ..stage,with
their acts and scenes, each w ith its peculiar qualities^.and all
cum ulating towards a clim ax., In other w ords, I was groping














Preface to the 1968 Edition

towards the nQdQn..Qstiidyiiig.JdbL&..s0:ucturs pC.successive events

im social processes o f varying scope and depth. Rut. while .the
synchronic-structures. investigated by .many^anthropologists-could
be shown to rest upon custom and.habit,.aud. very -ofteu-toxemain
beneath .the level o f conscious awareness,.jhe. .d^chronic_ structure
o r . processional form s that I was interested,exposing-.and
analysing developed out o f clashes and.. alignments__of .human
volitions and purposes, inspired by private and public interests
and__ideals. , .Yet the social drama itse lf represented a. com plex
interaction bet ween .normative ,pat.terns.-laidsrx3o3ivn_in-.the,caurse
o f deep regularities, o f candid.Qnmg,and.^

immediate aspirations, ambitions and other conscious goals and

strivings o f individuals, and groups in. the here. A t the
tim e, how ever, I was only able to raise some o f the problems o f
processual analysis, not to provide any o f the answers. Thus, I
was able to suggest that to. the.differentphasesofthe-saciaLcham a,
breach,...crisis,^_redressive_ ..action.,...and... immediate-...result,....there
corresponded .a particular .style .of. social in teraction particu lar
patterns o f norms and values, specific .types. ofLgoals-^ndLgaah
orientated. behaviour and specific form s o f conjunctive^ancLdisjun ctive behaviour. B ut I was unable to probe these differences
m inutely or especially cogently. For one thing, there was little
at that tim e in the w ay o f com parative data o f this type. Since
then there has been a considerable accum ulation o f extended case
material, for exam ple, in the works o f van Velsen (The Politics o f
Kinship, 1964, Manchester U niversity Press); G ulliver (Social
Control in an African Society, 1963, London: Routledge and Kegan
Paul); and Abner Cohen (Arab Border Villages in Israel, 1965,
Manchester U niversity Press) to name but a few , w hile new
theoretical perspectives on the study o f social processes (particu
larly political processes) have been opened up by such scholars as
Gluckm an (Politics, Law and Ritual, 1965, Chicago: Aldine); B ailey
(Politics andSocial Change, 1963, B erkeley: U niversity o f California
Press); Adrian M ayer (The Significance o f Q uasi-groups in the Study
o f C om plex Societies, 1966, in the Social Anthropology o f Complex
Societies, A .S .A . M onograph N o. 4, London: T avistock Publica
tions) notably his concept o f the action set \ a concept further
elaborated b y G ulliver (Dispute Settlem ent w ithout Courts :
the N dendeuli o f Southern Tanzania, paper given in W ennerGren Sym posium N o. 34, 1966, Ethnography o f Law ) ; Sw artz,

Preface to the 1 9 6 8 Edition


Turner and Tuden (Introduction to Political Anthropology, 1966,

Chicago : Aldihe) ; Swartz (Introduction to Local Level Politics,
1968, Chicago : Aldine) ; Nicholas (Factions : a Com parative
Analysis in Political Systems and the Distribution o f Power, 1965,
A .S .A . M onograph N o. 2, London : T avistock Publications)
and the articles on case m ethod by A . L. Epstein and van Velsen
in The Craft o f Social Anthropology (op. cit.). O ther scholars are
making contributions to a steadily broadening stream o f studies
in processual analysis, but clearly a great deal remains to be done
i f serious and rigorous comparisons are to be made in crosscultural terms between diachronic structures.
T o end on an ethnographic note : I have but recently learnt
that headman M ukanza died at a ripe old age in 1967 and was
succeeded b y Kasonda. I predicted in this book that Kasonda
w ould found a new farm and that Sakazao w ould remain in
Mukanza V illage w ith most o f his lineage kin. This is one
illustration o f the difficulties besetting prediction in anthropology,
for in 1954 when I left the field t did not anticipate that Sakazao
w ould die w ithin a few months and that M ukanza w ould live
for a further thirteen years. B ut since m y m ain doubt as to
Kasondas succeeding to the M ukanza headmanship was based
upon his relative youth in 1954, and not upon his capability o r
the extent o f his potential political support, m y error was not
perhaps so grave. In 1967 Kasonda was sufficiently mature to
meet the tacit age requirem ent for a headman and a headman
he became !


General Forms have their vitality in Particulars, Sc every Particular is a Man.
W iix iam B lake ,


N this book I attem pt to isolate the cardinal factors under

lying N dem bu residential structure. I^ ^ u sjd ieJ n ^ sttg atio n
ujxon-tibte village, a significant locaLunit,-an<ianaLyse.itLsuccessisceIy
asjift.independent social system andas a lanit w thinseyeraLw dder
sets o f social relations included in _the__totaldelchof.Ndem bu
society. Interw oven w ith the analysis o f siu ctu raL fo rm I
present detailed studies o f situations o f crisis, w h ich arise periodically in village life. These crises m ake visible botr_contradictions
beiwe.en crucial p rin cip lesgo verm n gd lkgestra^ ^
flicts between persons^andLgroups^kL-.sets o f-so cia lrelations
governed b y a single principle.. From repeated observation, o f
suck situations I have evolved the concept o f the * social .drama *,
w hich I regard as m y principal unit o f description and analysis
in_the study o f social process. O n pages 91-3 I divide the
social drama into four phases its * processional form and
present reasons fo r doing so. Through..the,.social, .drama-one
m a m etim es.lQ o k b en ea h th esu rfa ce.Q fso cia lT eg u la rities
intoIthe hidden contradicrions. and conflicts,in.tho,sociaLaystem .
The .kinds o fred ressive mechanism, conflict,
the pattern o f factional struggle^ and m e sources^of initiarive.fco
end crisis, w h ich are all clearly m anifest .in d ie c ia l
provide va lu a b led u es to the character .o, system.
H iXtdie. sm dy .o f social dramas m ust ,be based on num erical
analysis o f. village census data and the. critical exam ination o f
genealogies.. B efore one can study, breach one m ust be aware
o f regularity. I have therefore tried to discuss in quantitative
terms such factors as the m agnitude and m obility o f villages,
individual m obility, and the social com position o f villages, before
undertaking the analysis o f social dramas.
HmceT.hay.eap.pXQache_dmxy;m ajo rfieM x> fsm d yititw o ways.
Firs&,_! have, com pared, a number. oE.villages, w ithjceference to
such^measurable .criteria a?^e...md_geneafo^alj:omposition.
Secondly,.! have analysed.a. sequence. osociaLdram as.involvm g
the membership o f a single village,- and-the m em bers..of other











Preface to the 1957 Edition

viU ag _ esJjiik ed to It^

iajthe first instance I have examined j:eg]dadries o^rm,ox:curring
throughout many v illa g e s . in . the,secQnd,.,Ldiscuss_regularities
o f process in the social maturation, o f. a. single,.,village. The tw o
approaches complement one another.

B y num erical analysis o f genealogical and census data I was

able to infer the effective principles determining village structure.
These on the w hole w ere in conform ity w ith the ideal pattern
o f residential relations presented by informants. W ithin villages
the dominant principles influencing residence w ere maternal
descent and virilocaH ty.1 U nder-hldem hu ~conditions, conflict
between them was xneradicable_and accomited..iEL_considerable
measure for. the unstable and.fissile~characteno~villageL-organization and fo r ..the ...high, degree o f .individualjonobility. It is
possible that hunting, a purely masculine pursuit, and virilocal
marriage, w hich binds together male kin in local descent groups,
are parallel expressions o f structural opposition between m en and
wom en in this m atrilineal society. Hunting, the m ens sphere
in the basic econom y, is socially valued above its objective con
tribution to the food supply, and is highly ritualized. Cassava
cultivation, the w om ens sphere, is correspondingly undervalued,
and is ritualized to a lesser extent. Y e t w om ens w ork ensures
the physical survival o f the group, since hunting is fitfu lly pur
sued and success in it is uncertain. H unting im plies a rather
greater stress on econom ic co-operation than cassava grow ing,
although in all sectors production is m ainly individualistic.
Throughout the W est-Central Bantu, hunting is linked w ith
eminence or aristocratic ranking. Am ong Ndem bu, professional
hunters are highly honoured, and all men hunt to some extent.
Hunting is equated w ith virility and reinforces in some ways
the structural opposition between m en and w om en. For mar
riage, as noted above, is v irilo ca l; wom en, on w hom the social
continuity o f villages depends, reside at their husbands villages
after m arriage. Nevertheless, maternal descent governs prior
rights to residence, succession to office and inheritance o f property,
1 * Virilocal * in this book refers to die post-marital residence o f a woman
in the village to which her husband takes her. Uxorilocal * refers to the postmarital residence o f a man in his wifes village. Patrilocal refers to residence
in ones fathers village. * Avunculocal refers to residence with ones mothers
brother. M atriJocal refers to residence with ones mother.

Preface to the 1957 Edition


even o f guns, the professional hunters m ost cherished items o f

equipment. T h e nuclear residential group consists o f m ale m atrilineally related kin. T o rem ain together this set o f kinsmen
must im port their w ives from other village lineages and export
their sisters. B ut w ith maternal descent as the basis o f village
continuity' a contradiction arises between the role o f men. as
fathers w h o w ish to retain their w ives and children w ith them ,
and their role as uterine brothers and uncles w h o w ish to recover
the allegiances o f their sisters and sisters children. W ith ou t that
allegiance m en cannot found enduring villages n or can they
effectively press their claims fo r headmanship w ithin their villages.
Thus both marriages and villages are inherently unstable and
in-laws struggle continually fo r control over w om en and their
Another consequence o f virilocal m arriage m aking for in
stability in residential structure is the great measure o f auto
nom y it confers on the m atricentric fam ily. Frequent divorce
weakens the link between father and children but strengthens the
tie between m other and children. B u t virilocal m arriage separ
ates a m atricentric fam ily from its maternal kin-group during
the m inority o f the children w h o are reared in their paternal
village. I f they return to their maternal kin after the divorce
or w idow hood o f the m other, die narrow er lo yalty o f the
members o f the m atricentric fam ily to one another tends to com e
into conflict w ith their w ider allegiance to the village as a whole.
A matricentric fam ily matures into a uterine sibling group w hich
is the m ost frequent unit o f secession to form n ew villages.
Radical incom patibility, then, betw een m aternal descent and
virilocality gives a keen edge to conflicts between uterine kinsmen
and husbands o f w om en, w h ich result in quarrels between vil
lages ; to m arital conflicts, producing a h igh divorce rate ; and to
conflicts o f loyalties between narrow - and wide-span maternal
descent groups, w h ich inhibit (he developm ent o f deep localized
Another effect o f virilocal m arriage is a high rate o f patrilocality for children o f m ale village kin. Children adhere closely
to mothers, and w h ile a m arriage lasts the children stay at their
fathers village. Since villages are sm all (mean size 10*6 huts),
they tend to be at once shallow lineages w ith a fringe o f seminal
children and cognatic kin, and bilateral extended families, i.e.

x x v iii

Preface to the 1957 Edition

often the headmans uterine sibling group, brothers outnum bering

sisters, and their children. There is a tendency to m erge kin
on the m others and on the fathers side as members o f a genea
logical generation. W ithin the village, membership o f a genea
logical generation cuts across affiliation b y m atricentric fam ily
and unites cross-cousins w ith parallel cousins. T h e senior o f
tw o adjacent genealogical generations exerts authority over and
levies respect from the jun ior. O n the other hand, alternate
generations jo k e w ith one another and behave m ore or less as
equals. Adjacent generations tend to build huts in separate arcs
or the village circle, alternate generations in the same arc.
Generation oppositions and alliances tend to reduce tensions in
the relationsmps between matricentric families and lineage seg
ments, and between children and sisters children o f headmen.
M arriages between classificatory grandparents and grand
children w ho frequently belong to the same maternal descent
group, and marriages between cross-cousins, also tend to bind
the various com ponents o f the village together.
Nevertheless, aespite-these-centripetaLand..accrerive-tendencies
N dem bu_villages remain, inherently unstable. N dem bu- have
the ideal aim o f building up large
ceasingly _rebutted by. -rea lity.-. For the.-XaliensL described - b y
Professor Fortes the deep localized lineage is the-skeleton o f the
social structure. For N dem bu, on the other hand, a deep, lineage
is a seldom-realized goal, the end-product o f the sagacious manip
ulation b y headmen o f a number, o f ..organizational, principles
w hich confiict w ith-one -another-in-various_situations.
N dem bu villages have a w ide range o f spatial m obility and
each maternal descent-group in the course o f tim e is scattered
throughout different vicinages (discrete clusters o f adjoining
villages) over the w hole region. Thus vicinages are hetero
geneous in com position, for adjacent villages are seldom linked
b y maternal descent. T h e m obility and instability o f villages
partly determines and partly is determined b y the absence o f
political centralization. T he Lunda ancestors o f the Ndem bu
came from the great pyram idal state o f M w antiyanvw as empire
in the C on go, but in course o f tim e succumbed to the decentraliz
ing influences o f their w a y o f life, accelerated b y the slave
trading and -raiding o f the nineteenth century. B u t a vestige
o f the Lunda state pow er remains in the ritual role o f the Ndem bu


Preface to the 1957 Edition


chief, Kanongesha. T h e chieftainship sym bolizes the unity o f

Ndem bu and their ownership o f a com m on territory.
T h e m stabiiity o f the secular social. structure can be .palliated
but.-not controlled b y secular means. B itu a l...^ otated -w ith
corporate groups...such _as lineages ancLjlage& is^meagreand
canuonly act as_^.Jm poraiy.b ra k e, against, fission^_ButL.ritual
periform edby cu lt associations, that cut. across^.inUages^jyidnages
and even adjacent .chiefdoin&.of Lunda. o rigin , acts tokeep L th e
com m on valu esof.N d em bu socd ety.con stan tlybeforeth erovin g
individualists o f w h ich com posed These,. values include
historical renow n, hunting, and virinty,. fertility .and m otherhood,
and health and strength. T he m obility o f the-society. isreflected
in the-contingent and occasional character o f _the~rituaL T h e
m k fa ttu n e s.o flifo .in clu d in ^ .b ^
wox3iensreproductive.disorders,andsevereillness jfor.both.sexes,
are.attributed to the punitive action o f ancestor spirits, w ho are
exorcised and placated in rituals specific to each m ode o f affiiction. /."The ..curt association is made up_,o .doctors aixd_adepts
who w ere themselves once patients m d can(hdates ,in.that cult.
The w idest com m unity o f N dem bu is therefore a com m unity o f
suffering. iu c th e c o n te x to f
the com m on .values o f
the w hole society are stressed in symbol,..mime... and .precept.
T he associations are transient-grQUps.caUed.into-existericeJby the
unpredUctablernisfortuiies o f m obile .individuals. H ence the cult
groups are too fleeting and shifting in com position to develop
internal stresses and divisions.,,. I suggest, because the
organisational princips. w h ih govern, the seadar .stm cture are
contradictoqLand_produce percnniaI_conflicts--between-persons
an d groups, that rituals, are constm dy being_perform ed.. b y
unitary .though transitotyassociarions^ a n d thaL-these-rituals
stress com m on values_.over..and above. th e.d ash of..section al
interests. N dem bu ^ritual does n ot reflect.or express,_as.does
Tallensi_ox-S-wazi. ritual, the structure o f a stable,,society,, w ith
ritual role. _corresponding, as it w ere, to secuiarrole-i.-rather, it
compensates fo r the integradonal deficiencies.ofla.poHticaUy-unstable-society. Toverty_Qf_seculaxstams is confrQrLted_with rich
developm ent o f ritual roles in m any c ult associations. T h e rang e
o f effective political or econom ic co-operation is .small _;.aUome
perform ances o f ritual m ore than a thousand people m ay attend.
In Chapter O ne I briefly describe the historical and ecological






Preface to the 1957 Edition

background to this study. Chapter T w o presents the relevant

dem ographic information w hile in Chapter Three I attem pt to
isolate structural principles governing the social com position o f
villages on the basis o f genealogical and census inform ation.
T h e next four chapters are centred on the them e o f maternal
descent as the main principle underlying village continuity. In
Chapter Four the concept o f social drama is first introduced,
and it is further refined in the course o f this and the follow in g
chapter w hich deal m ainly w ith struggles to Succeed to headmanship between the matrmneal kinsmen in a single village. In
these social dramas the pattern o f factional intrigue is revealed
and its consistency w ith, o r degree o f departure from , the basic
social structure under varying circumstances is discussed. Chap
ters Six and Seven are concerned w ith village fission and w ith the
pattern o f secession. Chapters Eight and N ine deal w ith the
second m ajor determinant o f N dem bu village structure, virilocal
marriage, and h o w it operates w ithin, and then between, villages.
In. Chapter T en the integrative role o f the cult associations is
analysed. In Chapter E leven the chieftainship, in its ritual
importance and secular weakness, is briefly discussed.
T h at the pervasive^them eof the book is conflict^andjhe_resolu ria m o f conflict .arises..-from, m y preddection_fon_the_yiews,
fastJbecoming a theory, o f .that school o f British sociaLanthropologists w ho are com ing to regard a- social. system,as - a^field
of.tension, fu ll o f ambivalence,^ofxcteQperatiomand..contra&ring
straggle -1 F or _ihese..antbrQpQlogb^
static< m odel, a harm onious p a ttern ,.n o e.^


o f a. monistic outlook... A,sociaL.system. is ~a-Jield~~ofLforces

in which, to quote TorteSj^ ^centrifogal tendencies^and centripetal tendencies pull against one anotherJ, and^whosepower
to persist is generated -by its ow n socially transmuted,.conflicts.


U nderlying-the. whole-_.smdy-is,,jlie jconcept^^most jrecently

reform ulated .by .G lu clq a ^ and Cpkon, .that ,gQUpsJiave.J an

1 Giuckman, M ., R ituals o f R ebellion in South-E ast A frica ( 1954) p- 2 1 .

Giuckman has recently developed and expanded this theory in his Custom and
C o n flict in A frica ( 1955)*
8 Fortes, M ., T he Dynam ics o f Clanship among the T allensi ( 1945) P* 244.
8 Giuckman, Political Institutions in T he Institutions o f Prim itive Sodety
(1954), pp. <56-8o; Colson, E., * Social C o n tr o l and Vengeance in Plateau
Tonga Society , A frica , xxiii, 3 (July 1953).

Preface to the 1957 Edition


inherenttendency'to segment and then to beemne .bound-together!elaionsJhips are
absorbed My analysis
o f village structure is influenced by this conception. But it must be
pointed out that among the Ndembu, conflicts in secular, non-ritual,
relations speedily sharpen to the point o f irreconcilability in terms
o f the maintenance o f local cohesion. The high rates both o f divorce
and o f village fission attest to this. But conflicts which split sub
systems tend to be absorbed by the widest social system and even to
assist its cohesion by a wide geographical spreading o f ties o f kinship
and affinity. CenmfugaL.te.nde.ncies... prevaiL.on the. whole...over
centripetal, tendencies .at the .level. o fc o r poratekinship and ..local
groupings, but.centrifugality is confined .within. the Jb.Q.unds..o,.the
totabsQcio-geagraphicaLsy&tem-.o.Tthe.Ndembumation. A n.overall
ritual unity .the.fissile.nature.,ofseculanlife^ecular
life shows unceasing attempts to build. up_coherentgrQUps,.hutthese
attempts are as,centrifugaUancL.fissile
tendencies. T h e concept o f Ndembu unity, transcending .all the
divisions o f the secular system, is the productofcinnumerable^fitfuliy
performed occasions o f ritual, .each couched in the idiom- o f unity
through common misfortune.



HIS account o f the Lunda-N dem bu is based on tw o periods
o f field research carried out betw een D ecem ber 1950 and
February 1952, and between M ay 1953 and June 1954, after
m y appointm ent as a Research O fficer o f the Rhodes-Livingstone
I have no skill adequately to express m y debt to the Ndem bu
people o f M w inilunga D istrict fo r their contribution to this book.
M any m ore than I can name instructed m e patiently and pains
takingly in their w ay o f life. Villagers w ere never offended b y
m y presence at m any o f their sacred or intim ate occasions.
From them I learnt not only some fascinating facts, but also that
human frailty m ust be forgiven in oneself as in others i f human
social life is to be em iching. M y especial thanks are due to
Musona, m y shrewd assistant, to Samutamba, to Sakazao, and
to Headman Kajim a. Others w ho gave m e unstinting help
were W indson Kashinakaji, M uchona and C h ie f Ikelenge. I am
grateful to M r. R . C . D ening, the D istrict Com m issioner, for
access to his excellent maps and to the D istrict N otebook.
Indeed, m any persons have contributed, directly and in
directly, to the shaping o f this book. M any separate skills and
much collective know ledge and w isdom w ere put at m y dis
posal as field-report became thesis and thesis was fashioned into
From Professor Forde I received m y first training in social
anthropology. It was in his lectures and seminars that m y
enthusiasm for the subject received form and direction.
D uring m y field-w ork and w hile w riting up, tw o sets o f
colleagues, several o f w hom belonged to both sets, gave m e
invaluable assistance. I refer to the Research Staff o f the
R hodes-Livingstone Institute and to the Departm ent o f Social
A nthropology at the U niversity o f Manchester. Professor M ax
Gluckm an, a form er D irector o f the Institute and present Head
o f the D epartm ent, has given m e training, guidance and criticism
o f the highest order throughout this period. It is w idely
acknowledged, am ong those w ho have w orked under his
direction, that he has the g ift o f arousing in his students a zest






r iv ^ ^ o

O (T' O O O O 'H



for anthropological theory w hich enables them to keep on

w orking w ith a w ill in spite o f the m any and often form idable
difficulties and discomforts o f field-w ork. I am deeply grateful
to Professor Gluckman for his inspiring teaching, keen criticism
and generous friendship.
Several members o f the Departm ent and the Institute have
greatly helped me by reading the manuscript o f this book w ith
critical attention. Professor Elizabeth Colson, a form er D irector
o f the Institute, made m any constructive comments during the
crucial period between spells o f field-w ork. Professor John
Barnes read the manuscript at a late stage o f preparation w ith
incisive thoroughness. His advice on the layout o f tables, dia
grams and genealogies was o f exceptional assistance. I w ould
also like to thank D r. A . L . Epstein for m any valuable observa
tions on m y general argument.
I ow e a special debt o f gratitude to Professor C lyd e M itchell,
w ho was D irector o f the Institute w hile I was in the field, fo r
his unfailing help and encouragement. H e gave me m y first
practical training in field-w ork methods during a happy fort
night o f research in Lambaland. His studies o f Y ao village life
opened up several fruitful lines o f thought to me.
M r. C . M . N . W hite, M .B .E ., for a tim e A cting D irector o f
the Rhodes-Livjmgstone Institute, gave me the benefit o f his w ide
know ledge o f the western tribes o f N orthern Rhodesia.
The influence o f Professor Fortes is w rit large in this book.
M any o f his ideas, first encountered during m y student days,
had already become part and parcel o f m y thinking w hen I
entered the field.
M y w ife collaborated actively in all aspects o f m y field-w ork.
Her contribution included most o f the photography and much
o f the measuring o f gardens and m apping o f villages. In addi
tion, she drew the maps for the book. Her comments on the
argum ent have helped to shape its form .
In short, this book is in a very real sense the product o f
collective authorship.
V. W . T.





HIS book is prim arily w hat Professor Fortes w ould call a
* histological * study, an attem pt to analyse in close detail
the form and functioning o f a sub-system, the village, w ithin a
w ider system, the totality o f N dem bu society. B ut som ething
must be said, how ever compressed, about the w ider historical
and socio-geographical background o f the village. O ur village
m icrocosm is a com ponent in a system o f social relations existing
at a specific tim e and place. B u t it also has a specific history.
B y European notions, this history is b rie f and blended w ith
m ythology. Nevertheless, fo r N dem bu it is a history as a
record o f successive, interdependent events, m aking an irrever
sible process, w hich in their view explains w hat they are and
where they are.
The M w inilunga Lunda-N dem bu, hereinafter called N dem bu , 1
num bering about seventeen thousand, today inhabit the western
portion o f M w inilunga D istrict in the N orth-W estern Province
o f N orthern Rhodesia, approxim ately between n and i z
south latitude. B efore European occupation the N dem bu chiefdom o f Kanongesha com prised m ost o f die territory between the
W est Lunga, Zam bezi and L u fw iji rivers. N o w the tribal land
is crossed b y die international boundaries that demarcate N orthern
Rhodesia, A ngola and the B elgian C on go. Kanongeshas cbiefdom is virtually bisected b y the Angola-R hodesia boundary, and
on either side o f the boundary there is a ch ief w h o calls m m self

1 A num ber o f Southern Lunda groups call themselves Ndem bu , including

the inhabitants o f the chiefHoms o f Kanongesha, Ishinde and Kazem be Mutanda
(or Xzaizai). M y inform ants say that N dem bu was originally the name o f a
river at which these chiefs and their follow ers stayed together for several
years after their departure from M wantiyanvwas capital (see p. z). I
em ploy the term Ndem bu throughout this book to distinguish Kanonge$ha*s
Lunda from the Lunda-K osa (or A kosa) o f C h ief M usokantanda. This ch ief
lives in the Belgian C ongo near M usonoi but has nom inated Sailunga as his
representative in Northern Rhodesia. The Kosa occupy M winilunga D is
trict to the east on the W est Lunga R iver. Government refer to C h ief
Kanongesha*s Lunda as * N dem bu * and to the Kosa group as Lunda * in
official publications.



( )
f. ;






Schism and Continuity in an A frican Society

Kanongesha and is recognized as such b y the colonial pow er.

This book is based on research am ong the Rhodesian Ndem bu
and m ay not apply to Ndem bu living in Angola or the C ongo.
History and Traditions1
T he Ndem bu, like their Kosa neighbours to the east, claim to
have com e as invaders from the N orthern Lunda 2 empire o f
M w antiyanvw a between the Kasai and upper Bushimaie rivers,
and to have conquered or received the submission o f small
scattered groups o f indigenous M bw ela or Lukolw e. The inva
sion appears to have taken place before the beginning o f the
eighteenth century. According to Dias de Carvalho 3 the de
parture o f Kanongesha and his follow ers from Luunda took
place during the reign o f the fifth M w antiyanvw a, Ianvo N oeji,
whose reign Verhulpen places between 1640 and 1660.4 Accord
in g to the traditions o f both N dem bu and Kosa, Kanongesha*
m igration occurred at about the same tim e as the migrations
o f other Lunda leaders such as : (1) Kazem be Mutanda, w ho
established a chiefdom ju st to the north o f Kanongesha*s;
(2) Musokantanda, w ho founded a chiefHom to the east o f the
N dem bu in w hat is now M usonoi D istrict in the Katanga, and
in eastern M w inilunga ; and (3) Ishinde, whose chiefdom , like
that o f Kanongeshas, was later divided between Angola and
N orthern Rhodesia where it occupies the eastern part o f Balovale
D istrict. For m any years these chiefdoms, like that o f Kazem be
on the Luapula, described b y Lacerda, Gam itto and others, w ere
tributary to M w antiyanvw a, and even today visits are made by
1 A considerable literature exists on Lunda history, much o f it represented
in the bibliography attached to Miss McCullochs T he Southern Lunda and
R elated Peoples, Ethnographic Survey o f Africa, ed. Daryll Forde (1951). In
addition to works cited below in the text, Lewis Gann { The End o f the Slave
Tirade in British Central Africa : 18891912. T he Rhades-Livingstone Journal,
xvi (1954), p. 36) makes some useful comments on the Angolan slave trade.
See also my annotated translation o f selected passages from Dias de Carvalho,
A Lunda Love Story and its Consequences
in T he Rhodes-Livingstone
Journal, xix (1955), for the Northern Lunda traditions.
8 The land o f Mwantiyanvwa is called Luunda or Luwunda , and his people
A luunda , by Ndembu.
3 Dias de Carvalho, H. A., ExpediSo Portuguesa ao Muatianvua (1890),
p- 5 4 *.
4 Verhulpen, E., Baluha et Balubaiss de Katanga (1936).

H istorical and Ecological Background

Southern Lunda chiefs or their representatives to the potentates

capital on the Lulua river w hen a new M w antiyanvw a succeeds.1
W hen a Southern Lunda ch ief succeeds to office, confirm ation
is sought from M w antiyanvw a. "When the British Govern
ment sought to abolish the N dem bu chieftainship o f Nyakaseya
in 1947, die incum bent hastened to M w antiyanvw a w ho w rote
a letter to the W estern Provincial Com m issioner on his behalf.
As a result he was reinstated.
According to W h ite,2 the Southern Lunda
largely retained their own language and to a greater degree their social
system. One reason for this seems to have been that they migrated
into a broken thinly-populated country where they had less chance
o f being absorbed by the people whom they found already there.
The latter were o f Lukolwe stock as evidenced by the names o f rivers
which they have le ft.. . . Indeed the Lukolwe, whose remnants today
are a very primitive people, may have been one factor which was
responsible for the failure o f the Lunda to show the same degree o f
enterprise as their neighbours. . . . Lunda also suffered continuously
from the incursions o f their Lwena and Chokwe neighbours which
may indicate that they were a later migration, at a disadvantage with
their earlier established neighbours, which prevented their expansion
and development. Hence the Chokwe and the Lwena were able to live
in more open, flatter country* whilst the Lunda took refuge in broken
hilly country between, the Zambezi and Kabompo headwaters where
they lived in small scattered communities.
There are today in the M w inilunga N dem bu region some
tw enty-six villages, the residential cores o f w hich claim to be o f
M bw ela * origin. T he term M bw ela * appears to have been
w idely applied b y in-com ing groups o f Lunda stock to the m ore
prim itive Bantu peoples, such as L ukolw e, w hom they encoun
tered to the south and w est o f their hom eland. Ndem bu today
use the term kabeta Kambwela to denote * the south \ Sometimes
they group M bw ela and K aonde together as having a cultural
affinity. T h e M w inilunga group o f putative M bw ela origin are
* As in 1951, when Mbaka succeeded. Both Kanongeshas sent represen
tatives, Musokantanda and Kazembe Mutanda went in person, and several
Native Authority Sub-Chiefs such as Ikdtenge, Nyakaseya and Mwininyilamba
visited the new Mwantiyanvwa, each with his entourage who were given
food by the potentate.
2 "White,C. M . N., The Balovale Peoples and their Historical Background*,
T he Rhodes-Livingstone Journal, viii (1949), pp. 35-6.

Schism and Continuity in an A frican Society

generally know n as *K aw iku *, after the name o f a plain to the

north-east o f M w inilunga Bom a, w here they are first said to
have been encountered by Kanongeshas invaders. According to
Ndem bu and K aw iku traditions they resisted the invaders for
m any years before they submitted. In A ngola, the Ndem bu apply
the name * H um bu * to a population o f M bw ela origin in the
south, and the senior Hum bu headman Kafwana plays an im
portant role in the Angolan Kanongeshas installation ritual.
T oday K aw iku and Hum bu are culturally and linguistically
almost indistinguishable from N dem bu, but on occasion stm
maintain that they are * the owners o f the country \
Ndem bu did not establish themselves as a higher caste or class
above the K aw iku. B oth groups interm arry freely and live at
the same econom ic level. Ndem bu, how ever, consider K aw iku
to be turbulent, clownish and discourteous people and jib e at their
selfish and quarrelsome behaviour. K aw iku take delight in
fostering these beliefs b y jo k in g about their ow n uncouthness.
It is doubtful whether Ndem bu have retained the social system
practised in their hom eland to the extent indicated by W hite.
Several w ell-inform ed accounts o f M w antiyanvw as state o f
Luunda exist w hich give the picture o f a h ighly centralized
political system.1 In the royal capital a host o f nobles and officials
w ere perm anently stationed, supported b y the tribute that Sow ed
in, at the height o f Lunda pow er, from chiefdoms as w idely
separated from one another as Kapenda-ka-M ulem ba to the west
and Kazem be on the Luapula to the east, a distance o f m ore than
seven hundred miles. N o t only tribute but trade contributed
to the surplus w hich maintained the Lunda nobility. M w antiyanvw a had trade agreements w ith a num ber o f Ovim bundu
chiefs w ho despatched caravans into the interior bearing trade
goods in exchange for slaves, ivo ry and, later, rubber and bees
w ax.2 Capello and Ivens 3 w rote in 1882, when the empire was
already on the wane, at a short distance from the mu-sumba
[Musumba, state capital ] are established vast markets, true
1 Dias de Carvalho ; Buchner, M., Das Reich des Mwata Yamvo und
seine Nachbarlnder , Deutsche Geographische Bltter, 1 (1883) ; Pogge, P.,
Im R eiche des M wata Yamvo (1880).
2 Childs, G. M., Umbundu K inship and Character (1949), p. 205.
3 Capello, H., and Ivens, R ., From Benguella to the T erritory fo Yacca, trans
lated by A. Elwes, (1882), vol. 1, p. 3 8 9 .

H istorical and Ecological Background

bazaars containing straight lanes or streets w here flour o f various

kinds, ginguba, palm -oil, fresh and dried m eat, massambaia,
salt, tobacco, m aluvo (palm w ine), mabellas, and other articles
are displayed, and are bartered for merchandise, such as blue and
red baize, cottons, printed calico, large w h ite and sm all red beads,
pow der, arms ana bracelets \ Verhulpen m entions h o w the
razzias o f the Lunda and Luba states sw ept die C entral African
area fo r slaves to sell to the Bangala and O vim bundu m iddle-m en
o f the Portuguese in exchange for guns and trade goods.
T o maintain this system o f tribute, trade and pillage, M w antiyanvw a had at his disposal a pow erful arm y quartered in the
capital and divided into advance, flank and rear guards, each
arm y corps having a reserve o f liaison personnel and scouts.
Capello and Ivens 12 m ention 4 wars o f exterm ination * w aged
by M w antiyanvw a against * recalcitrant tributaries * and dynastic
T h e N dem bu, on the other hand, seem never to have had any
thing like the same degree o f political centralization as their
Lunda ancestors. It is said that Kanongesha, a son o f M w an
tiyanvw a *, was accompanied, o r follow ed shortly afterwards,
by tw elve senior headmen drawn from various Lunda noble
families, w hen he first entered the present M w inilunga area.3
Each o f these was accom panied b y relatives, m ale and fem ale,
and a retinue o f warriors. It took at least a generation o f inter
m ittent raiding to subdue the indigenous peoples, and the
invaders w ere scattered far and w ide over the country in pursuit
o f them . Eventually, so the tradition goes, the Southern Lunda
chiefs, together w ith some neighbouring Lw ena chiefs such as
1 Capello and. Ivens, vol. i, pp. 390-2.
2 For instance, Livingstone writes in his journal for 1854 that in the Lunda
area o f (immediately north o f Kanongeshas Ndembu), * the
frequent instances which occur o f people changing from one part o f the
country to another show that the great chiefs possess only a lim ited pow er *
(my italics) M issionary Travels and Researches in Southern jdfrica (1857),

P* 305.

s Campbell, D. {W andering in C entral A frica (1929), p. 52 et seq.) is, I think,

substantially correct when he attributes the * colonial expansion * o f the
Lunda o f which the Ndembu migration was a part to * wars o f jealousy
and internecine strifes that scattered leaders, who, with ambitious followings
went farther and farther afield . . . nibbling into the weak frontiers o f other
tribes \







Schism and Continuity in an A frican Society

Chinyam a and N yakatolu, whose forbears had com e from M w atiriyanvw a in an earlier w ave o f m igration, m et together and
divided up the conquered M bw ela and Lukolw e land between
them , using the larger rivers as boundaries. Kanongesha re
ceived as his share the land between the Zam bezi, W est Lunga
and L ufw iji, w ith Ishinde as his southern neighbour, M usokantanda to the east, Kazem be Mutanda to the north, and the Lwena
dhieftainess, N yakatolu, to the west. Kanongesha recognized
the claims o f his tw elve headmen to certain tracts o f land w hich
they had conquered, and later allocated further areas to certain
o f his sons and dose matrilineal kin. H e retained the largest
share, including m ost o f the M w inilunga N dem bu region south
o f w hat is n ow the north-western pedicle o f M w inilunga Dis
trict, and the Angolan region around the L ovw a river, to be
under his direct control. T h e headmen and relatives to w hom
he had given areas w ere supposed to send him tribute o f various
kinds ; unrelated headmen had to send him tributary w ives from
their ow n matriUneages. H e him self had to send a portion o f
the tribute thus obtained to M w antiyanvw a, a custom w hich was
discontinued several years before die C h okw e invasion o f that
potentate's territories in 1885.
A fter the division o f the land, N dem bu tradition reports a
long period o f peace, during which there w ere no wars against
neighbouring tribes. In this period, how ever, it w ould appear
that the m inor and m inim al political units, the senior headmen's
areas and the villages w ithin them, became increasingly autono
mous and independent o f Kanongeshas control. Traditions
exist o f feuds during this period between senior headmen and
even between individual vmages and vicinages. Several factors
are responsible for this regression towards local autonom y.
A m ong these m ay be reckoned isolation from Luunda w ith its
relatively high level o f culture and social organization ; die
small num ber o f N dem bu invaders and M bw ela invaded alike,
scattered over a w ide region o f forested terrain ; and the lo w
level and rudim entary character o f econom ic production
associated w ith small shifting settlements. Villages were not perm anendy anchored to specific tracts o f land but m oved across the
face o f the Ndem bu region w ith relative freedom . H igh spatial
m obility contributed pow erfully to the considerable political
autonom y o f the village.

H istorical and E cological Background

T he decentralization and m obility o f Kanongeshas N dem bu

were accelerated b y the slave-trading and slave-raiding w hich
reached their height in this region in the second h a lf o f the nine
teenth century. O vim bundu slave-traders m ade trade pacts
w ith Kanongesha and im portant senior headm en such as Ikelenge, in north-w est M w inilunga,1 and these pow erful men
raided their ow n tribesmen and sold them into slavery in return
for guns and d oth . Later, C h okw e and Lw ena, heavily armed
w ith Portuguese muskets, attacked the N dem bu, w h o at first
presented no united front to the slave-raiders and w ere defeated
village b y village, the unfortunate inhabitants being marched o ff
to jo in the long lines o f slave caravans that converged on N yakatolus Lwena capital from die interior. T h e N dem bu w ere
rallied, not b y a ch ie f but b y a com m oner, Chipenge, w ho went
round the country urging diem to sink their private feuds and
jo in d y hurl out the invaders. H e inflicted several defeats on a
large w ar-party o f C h okw e and Lw ena w h o w ere under the
leadership o f the Lwena ch ief K angom be (mentioned b y Dan
C raw ford).2 A fter this, the raiders w ith drew into Angola.
B ut the N dem bu w ere n ot left lo n g in peace, for in 1906,
follow ing die announcement o f the K in g o f Italys arbitration
o f 1905, w hich settled a long-standing dispute between Great
Britain and Portugal over the exact position o f the western
boundary o f N . Rhodesia, officers o f the British South Africa
Com pany, accompanied by native police, commenced the
administration o f the new M w inilunga D istrict. As a result o f
the harshness o f an early administrator, there was a mass m igration
o f Ndem bu out o f the Com panys territory into the Belgian
Congo and A ngola. T he officer in question was discharged,
but hardly had the people returned than there was a new exodus
in 1913, w hen taxation was first introduced. The senior head
man Ikelenge, later to becom e a S u b -C h ief under the N ative
Authority, was the o n ly im portant N dem bu to remain, arguing
that die taxes w ould produce benefits fo r the people in the form
o f roads, services and peaceful governm ent. Chipenge, now an
1 Gann, L. (p. 33), points out that the Angolan slave dealers, unlike the
Arabs, never established political rule o f their own, but preferred to work
through native chiefs.
2 Crawford, E>., Thinking B lack (1912), p. 116; Tilsley, G. E., D an Craw fo rd , M issionary and Pioneer in Central A frica (1929), pp. 104-5.

Schism and Continuity in an A frican Society

old man, encouraged the flight, realizing that the Europeans

w ere too strong to eject by force. Gradually the population
trickled back and w ere regrouped into villages b y die N ative
Com m issioner, the leaders o f the malcontents being com pelled
to build near the Bom a (a w idely applied term for the head
quarters o f the British D istrict Adm inistration) where an eye
could be kept on them.
In 1906 Kalene M ission was founded b y D r. W alter Fisher
o f the Christian M ission in M any Lands at Kalene H ill in the
northern pedicle, and a hospital and school for Africans w ere built
soon afterwards. This event was to have im portant effects on
the distribution o f N dem bu villages and ultim ately on their
internal structure. For a time N dem bu regarded Kalene H ill
as an alternative administrative centre and headmen took tribute
to D r. Fisher as a mark o f vassalage. A fter each exodus from
the Bom a area, m any headmen returned to M w inilunga to build
near the M ission, and others came from A ngola w here illicit
slave-trading continued into the tw entieth century. A trading
store was early established at Kalene ; and in 1921, D r. Fishers
son, fFolliot Fisher, started H iliw ood Farm, near Ikelenges v il
lage, tw elve miles from Kalene, and a store was eventually added
to this.
Som e years later Sakeji School, a boarding school for the
children o f European missionaries o f various Protestant denom
inations, was founded close to H iliw ood Farm. In 1952 nearly
sixty children w ere in attendance at the school. Altogether, in
1952, there w ere about a hundred Europeans, including teachers
o f European and African school children, doctors, state registered
nurses, missionaries, and farm managers at H iliw ood, in the
>edicle area. Each o f the three European centres em ployed a
arge A frican staff, including hospital attendants, farm hands,
road labourers, school teachers, clerks and dom estic servants.
In addition to these, others w ere em ployed as store capitaos
(foremen) and maintenance men by European com m ercial firm s,
whose headquarters w ere on the N orthern Rhodesian Copperbelt. Each African in regular w age em ploym ent, whether he
lived w ithin the precincts o f the European settlements or in an
adjacent N dem bu village, tended to becom e surrounded by
kinsmen w h o sought his patronage.
Three senior headmen in the pedicle became N ative Authority


H istorical and Ecological Background

Sub-Chiefs under Governm ent, and each o f them is surrounded

by an entourage o f paid petty officials : clerks, kapasus (N ative
A uthority Messengers), assessors, and the like. A t each SubC h ief *s capital village there is a L ow er School under the super
vision o f the Christian M ission in M any Lands at Kalene H ill.
A t one, that o f C h ie f Ikelenge, there is a dispensary, a PostOffice, and a N ative A uthority M arket.
T he tendency to cluster round a European settlem ent in v il
lages w hich began in D r. Fishers tim e has continued, and today
each European centre, and indeed each N ative A uth ority head
quarters, has becom e the focus o f a large A frican population.
N o t only those already em ployed, but those w h o seek w age
em ploym ent, com e to settle in villages and other local units in
the pedicle, either b y attaching themselves to already existing
villages w here they nave kin, or b y com ing in kinship groups
and founding n ew residential units. T h e fact that the soil in the
pedicle is considered by N dem bu to be especially suitable fo r
cassava cultivation is a further strong inducem ent to settle there.
Paym ent o f domestic servants and labourers and the creation o f
opportunities for buying trade goods and selling native produce
in the northern pedicle led to further increases in its population.
Game was virtually exterm inated in the pedicle b y Africans w ith
m uzzle-loaders and Europeans w ith rifles.' Local trade in dried
meat and fish developed w ith Angolan and C on go Africans w h o
exchanged these com m odities fo r cash and store goods. This
area was the first in M w inilunga to grow crops fo r cash as w ell
as for subsistence. T oday this priority in developm ent is re
flected in the existence o f a num ber o f A frican traders, especially
in Ikelenges area near M illw ood Farm, w h o bu y cassava m eal
from the villagers fo r re-sale to the B om a and to B elgian traders
in the C on go. T h e pedicle area exhibits the highest degree o f
disruption o f the traditional social system seen in the Ndem bu
M any im m igrants, including some Lw ena and O vim bundu,
have entered the area since 1920, often to seek asylum from the
Portuguese w ho. had put dow n several rebellions o f the interior
tribes. T h e social com position o f the pedicle has thus become
extrem ely heterogeneous and its position astride labour routes
from the adjacent colonial territories to the Rhodesian lineof-rail have m ade it subject to continual stim uli from alien







r- \

Schism and Continuity in an A frican Society

cultures. In die last few years profound changes have occurred

in die residential structure in this area : the m ost noteworthy
has been the breakdown o f traditional villages into small units
headed by younger men w ho participate in the encroaching
cash econom y. These units know n as mafwami (from the
English * farms ) increased enorm ously in num ber between m y
first and second periods o f field-w ork. B u t since change is not
proceeding evenly either in tempo o r direction in all areas alike,
I found it possible to observe m any villages w hich retained the
indigenous structure, even in the pedicle, and w hose inhabitants
still tended to conform to traditional norm s o f behaviour.
T h e traditional village was a circle o f pole-and-m ud huts
typically containing a core o f m atrilineally related kin under the
leadership o f a member o f the senior genealogical generation
chosen b y the villagers. The ifwami or * farm * consists o f one
or m ore Kim berley-brick houses bordered b y a few mud huts and
it is occupied b y the form head, his elementary fam ily and a small
fringe o f kin and unrelated persons. In the pedicle near European
settlements and N ative A uthority Sub-Chiefs capital villages a
ribbon-developm ent o f forms has begun along the Governm ent
m otor roads. In this book I lay m ajor stress on the problems
posed b y the structure o f traditional villages and leave over for
a further study the social organization o f the form. B u t this
sudden m ultiplication o f forms in a few years was only the out
w ard manifestation o f changes in values and o f conflicts between
old and new principles o f social organization, w hich had been
goin g on w ith gathering intensity w ithin the fram ew ork o f the
traditional village for some time. I have tried to indicate some
o f the effects o f social and cultural change on the rate and type
o f village fission at different periods in this century in Chapters
Four to Seven. B ut until about 1950 the m ajor effect o f economic
and political changes was to increase the rate and not to alter
significantly the type o f fission. The basic unit o f fission was
a maternal descent group. T oday there are signs that the
elem entary fam ily is becoming the typical unit o f fission and
that the links binding together uterine siblings and matrilineal
parallel cousins are wearing thin. I m ention these tendencies
but do not discuss them in detail. In the group o f villages I
selected fo r particular study (most o f them in the area o f the
deposed N ative Authority Sub-C hief M ukangala), the effects

H istorical and Ecological Background


o f change on the social structure w ere visible, b u t they w ere far

less pronounced than in die pedicle.
W hen Indirect R u le was introduced in 1930, the N orthern
Rhodesian claim ant to the Kanongesha chieftainship was given
the status o f Senior C hief, and six senior headmen w ere made
Sub-Chiefs.1 These w ere : C hibw ika, Ikelenge (or Ikelengi),
Nyakaseya, M w ininyilam ba, M ukangala, and N tam bu Lukonkesha. C hibw ika like Kanongesha reduplicated in A ngola
is the title given to the nom inated successor o f Kanongesha. His
area lies to the south o f Kanongesha^ w ith w h ich it is contiguous.
Ikelenge is descended from an official called Kalula w ho accom
panied the first Kanongesha. A t the installation ritual o f a
Kanongesha his duty is to rem ove the lion and leopard skins
from the chiefly chair. T oday his area lies in the south-east o f
the pedicle.2 T o his west is the area o f M w ininyilam ba whose
ancestor was the Ifwota or * scout \ the leader o f a w ar-party w hich
went in advance o f Kanongeshas forces during the invasion.
Nyakaseya C she w ho pours out i) is a tide m ade for the first
Kanongeshas senior w ife (mwadyi) whose duty was to pour out
beer when the ch ie f offered it to his guests. T o d ay Nyakaseya
is a male Su b -C h ief in the north o f the pedicle. Mukang*ala*s
tide was created b y the second Kanongesha, Kabanda, for his
son w ho subdued the K aw iku. H is village was situated fairly
near the K aw iku Plain in order that he m ight prevent rebellion.
The senior K aw iku headman N sanganyi paid him tribute and
1 Actually, Government Had given Je fa cto recognition to these headmen
as * Sub-Chiefs * long before 1930. C . M . N . "White (personal communica
tion) prefers to call mem ' Sub-Chiefs * rather than * senior headmen in the
pre-European organization. White writes that * they have always had royal
insignia, their titles go back to the early days o f the Lunda chiefs' coming,
and the Lunda themselves in any interpretation, o f what the vernacular means,
regard them as Sub-Chief; *. I use different terms for their traditional and
modem positions because I believe, as I bave argued in Chapter Eleven,
pp. 323-7, that their status rested on different criteria before British rule.
Jh the past they enjoyed moral and ritual pre-eminence, but did not neces
sarily possess effective political authority. Today, since they are supported
by the Administration, they have greater powers o f coercion and are organized
into a hierarchy o f authority, but some or them, whose status rests on insecure
or inadequate traditional foundations, do not command the respect senior
headmen formerly possessed.
2 See Map 2.


Schism and Continuity in an A frican Society

acted, as his Chivwikankanu, a ritual office w hich gave him an

im portant role at the installation ritual o f a M ukangala. Kafwana,
the senior headman o f the Hirnibu in A ngola, was Kanongeshas
Chivwikankanu. Kafwanas office, w hich means literally * he
w ho puts on the lukanu *, involved the investing o f a new
Kanongesha w ith the lukanu bracelet o f human sinews and
portions o f genitalia, the supreme sym bol ofLrnida chieftainship,
and thereafter the periodical renewal o f its ritual pow ers by
washing it w ith special medicines. The other Sub-C h ief recog
nized b y Governm ent in 1930 was N tam bu Lukonkesha.
N tam bu was not tributary to Kanongesha in the past but paid
tribute directly to M w antiyanvw a. Although his territory was
m uch smaller than Kanongeshas his status was ju st as high. The
father o f the present Angolan Kanongesha was C h ie f N tam bu
and the father o f N tam bu was a form er Kanongesha. Tradition
ally, N tam bu had the right to sit on a chair in the presence o f
Kanongesha, but Kanongeshas senior headmen, som e o f w hom
are n o w Sub-Chiefs, had to sit on the ground.
In 1936 N tam bu, and in 1947 M ukangala, had recognition
w ithdraw n from them . M cC ulloch, on the basis o f inform ation
from W hite, w rites : 1 * Form erly these Sub-Chiefdom s seem
to have been quite firm ly established, but today political bound
aries have been greatly altered and in the interests o f centralization
recognition has been w ithdraw n from redundant Chiefs and
several areas have been amalgamated.* T he areas o f both
4redundant * chiefs have been incorporated in Kanongeshas ow n
Few o f the m odem Sub-Chief! w ere in fa c t4firm ly established
O f the tw elve headmen w ho came originally w ith Kanongesha,
o r follow ed him shortly afterwards, some are n o w ordinary
headmen. These include Chibw akata, the present incum bent o f
the office o f Kamhanji (* war-leader and * executioner ), whose
village is a few m iles from M w inilunga Bom a, Kabtm gu and
N yachilesya in A ngola, and M wanta waLuunda, Nyachikanda,
K abuya and C hota in M w inilunga District. B y an odd quirk
o f fortune, K akom a, the 4war-leader * o f M usokantanda (one
o f the senior m igrant c h ie f from Luunda), has been appointed
a S u b -C h ie f; and Musokantandas * em blem -purifier (Chiu1 McCulloch, p. 26.

H istorical and Ecological Background


wikankanu) Sailim ga, lias been appointed Senior C h ie f o f the

Akosa. Su b -C h ief M w ininyilam bas original area was on the
east o f the Lunga R ive r near the Kasanjiku Stream, in what is
now * Sailungas A rea*. Tow ards the end o f the nineteenth
century the incum bent o f the tim e fled after a quarrel from this
area and settled in w hat is n o w the pedicle. T h e senior headman
there was a perpetual * son * 1 o f Kanongesha, called K afw eku,
w ho collected tribute from village headmen fo r Kanongesha.
W hen Authorities w ere appointed in 1930, M w ininyilam ba,
although a refugee from another area, was appointed Sub-C h ief
o f K afw ekus area. N dem bu say this was done because the
M wm invilam ba o f the tim e was a forceful character w ith a
large village, whereas the K afw eku was a tim id m an w ho avoided
meeting Europeans.
It is doubtful in fact whether the headmen recognized as SubChiefs b y Governm ent w ere ever regarded as ch iefs*2 b y
N dem bu w ith the exception o f N tam bu Lukoukesha. T h ey
were sim ply village headmen w h o possessed im portant historical
titles. Certain o f them claim to have received their insignia from
M w antiyanvw a. I f they w ere strong m en w ith large villages
Kanongesha m ight have entrusted them w ith certain tasks.
Finally, i f he w ere n ot related to them b y perpetual kinship *,
1 The son o f a Lunda chief who has been granted his own area and at the
same time excluded from the succession to his fathers title is called M wana
awuta or M wanawuta. Mufcmgala is also a M wanawuta o f Kanongesha.
I. G. Cunnison (A Socia l Study o f a Bantu People , unpublished thesis, Oxford
University, p. 116), defines perpetual relationship as a relationship between the
titles o f headmen and chiefs winch always remains constant irrespective o f the
actual genealogical relationship o f the individuals holding titles, and which
expresses die political relationships between the different chiefs and headmen.
Cunnison has recently published a paper on Perpetual Kinship: a political insti
tution o f the Luapula peoples in T h e Rhodes-LivingstoneJournal, xx, pp. 28-48.
2 All were entitled to the respect due to their historical status as founders
o f the Ndembu people and all had certain insignia such as wooden slit-gongs,
xylophones, bangles, short swords, etc. Some utilized the initial advantage
given them by this status to dominate small areas. Harding (In Rem otest
Barotseland (1904), p. 146) describes Ikelenge (* jEakaling *) as a * very important
chief under Kanongesha \ His capital village was then at Kalene Hill, which
was an important centre in the slave-trade with the Ovimbundu. It is likely
that the Ikelenge headmen built up wealth and influence by trading in slaves.
Harding published his book in 1895. He visited Ikelenge in the x88o*s when
Ovimbundu slave-traders still dealt with Ndembu headmen.


Schism and Continuity in an A frican Society

he m ight m arry into their families. B ut although the boundaries

o f their areas have been rigid ly demarcated b y the Adm inistration
it appears as i f m any o f them ranged w id ely over the N dem bu
region in the past. M ukangala, for exam ple, during part o f
the nineteenth century made his capital village w ell to die east
o f the Lunga. I m ention below the spatial m obility o f other
im portant villages, including Chibwakata, and in the last chapter
discuss in some detail the role o f the senior headmen in the
indigenous social system.
T he Kanongesha in A ngola is regarded b y m ost N dem bu in
M w inilunga D istrict as their authentic chief, since he possesses
the lukanu given b y M w antiyanvw a to Kanongesha Kabanda (see
p. 320). T he first Kanongesha to enter M w inilunga D istrict,
M ulum bi, took this lukanu w ith him w hen he left A ngola in the
early i92os. M eanw hile a new Kanongesha was elected b y the
senior headman o f the Angolan Ndem bu w ho blamed M ulum bi
for leaving the Isenga R iver Island area w here the graves o f
form er Kanongeshas w ere located, ha 1927 the A ngolan
Kanongesha, Sakayoli C hifiiw u, conspired w ith Kafwana, the
Angolan H um bu senior headman, w ho then obtained the lukanu
from M ulum bi on the pretext that it had been defiled and had
lost its virtue. Kafwana told M ulum bi that he w ould wash it
w ith m edicine at the cross-roads to the west o f die capital village
at m idnight to cleanse it. H e took the lukanu and w ent straight
to A ngola w here he gave it to Sakayoli. M ulum bi tried to
obtain a new lukanu from M w antiyanvw a, but according to m ost
N dem bu was not given an audience b y him . H e returned indeed
w ith a new lukanu, but this is said to have been made b y a Luunda
court official,1 n ot b y M wantiyanvwa.
T he form al structure o f the Ndem bu political system under the
N ative Authorities Ordinance and the N ative Courts Ordinance
presents a rigid hierarchical appearance quite foreign to the
indigenous structure. In the past there was o n ly a single N dem bu
4ch ief (mwanta wampata) confirm ed in his office and tributary
to M w antiyanvw a, and bis effective authority was confined to
his ow n im m ediate area. The first senior headmen (ayilolu) 2
w ere initially allocated tracts o f bush in w hich they had rights
1 Chipawu.
2 For a discussion o f the meaning o f this term see p. 323.

H istorical and Ecological Background


to hunt and cultivate and to exact tribute from incom ers w h o

wished to hunt and cultivate there. In return for these rights
they had to g ive Kanongesha tribute w hen he asked fo r it and
tributary w ives from their o w n matrilmeages. B u t w ith the
continual m ovem ent o f villages from site to site, and w ith the
frequency o f village fission, it became extrem ely difficult fo r
senior headmen to exert political authority over the inhabitants
o f their areas. T h e basic econom y o f the N dem bu, cassavagrowing and hunting, did not com pel any individual or group
to remain in one specific area. In the period o f slave-trading and
-raiding, ties between groups and localities must have been further
loosened both b y internecine raiding and b y the instant necessity
to be ready to flee before an external enem y. In such an unstable
and highly m obile society Kanongesha could not enforce his rule
on senior headmen and the latter could not exact obedience from
the villagers in their areas. Villages resorted directly to force in
the form o f self-help in defence o f the rights o f individuals and
groups. I f a ch ie f or a senior headman managed to establish some
degree o f centralization in his area this was due m ore to personal
qualities than to his social position. U nlike such chiefs as
Kazembe o f the Luapula V alley or Chitim ukulu o f the Bem ba,
the Kanongeshas had no m ilitary forces perm anently stationed
at their capitals. T h ey had to rely on possessing rather larger
followings o f kin and slaves than other headmen, even to enjoy
local political pre-em inence. In Chapter Eleven I argue that
the position o f the Kanongesha was ritually rather than politically
important and that he was respected rather than obeyed.
The position o f the C h ie f and o f those senior headmen w h o
have been appointed as Sub-Chiefs has been considerably
strengthened b y Governm ent. M cC ulloch, follow ing W hite,
writes th a t1
today all civil cases and lesser criminal cases can be tried by the court
o f the Native Authority (the Chief's court) as constituted by the
Native Courts Ordinance. The C h ief is officially supposed to preside
at trials, but in practice cases are often heard only by assessors. From
this court, appeal lies in the first instance to a special Appeal Court
o f the Native Authority, then to the District Commissioner, and
1 McCulloch, pp. 2 6 -7 , see also p. 71, based on MSS. and personal com


Schism and Continuity in an African Society

finally to the High Court o f Northern Rhodesia. More serious crimes

are taken directly to the District Commissioner, who may remit to
the Native Authority any cases which he considers to lie under his
jurisdiction. Imprisonment can be given for certain criminal offences
by courts o f the Native Authorities, but this happens comparatively
rarely, and the usual penal sanction, is a fine. Both fines and court
fees are paid to the Native Treasury. In fact many minor cases tend
to be setded out o f court, but the Lunda Chiefs discourage this. . . .
Cases concerning land rarely come into the Native Authority courts,
but are usually settled by village heads.
In the past there do not appear to have been such centralized
judicial institutions. Kanongesha and his im portant headmen
and their im portance it m ust be remem bered varied w ith personal
and other adventitious factors m ight on application test a
diviners attribution o f w itchcraft or sorcery to a person b y
recourse to their ow n poison oracles. O ccasionally they m ight
officiate in the ritual o f Musolu to bring on belated rains and
confer fertility on crops, animals and men. B ut in general m ost
disputes w ere settled at the level o f the vicinage (chitungili).
M en skilled in the advocacy o f cases, know n as mahaku, o r in
judgm ent, know n as akwakusompesha, w ere jo in tly invited by
a plaintiff or his representative and a defendant or his represent
ative to discuss the case and give their verdict. Such m en were
often, but need not necessarily have been, headmen. I f there
w ere a senior headman (an im portant chilolu) in the vicinage
most cases in. that vicinage w ould be settled in his village. Cases
w ere seldom brought to Kanongeshas court on appeal. Kanon
gesha in fact was politically hardly m ore than a senior headman
w rit large and his effective authority was confined to the few
vicinages o f his im m ediate area.
T he N ative A uth ority and Sub-Chiefs are allow ed to legislate
under the guidance o f the Adm inistration. B ut there is m uch
passive resistance against the implementation o f N ative A uthority
Rules and Orders by villagers w ho resent the political powers
vested in the N ative Authority.
Apart from those areas in the pedicle nearest to European
setdements and Sub-Chiefs capital villages, m any features o f the
indigenous social system have tended to persist in much o f western
M w inilunga, especially at the village and vicinage levels. This
is probably due to a series o f factors, including the thinness o f

H istorical and Ecological Background


European settlem ent in the D istrict, the distance from urban

centres, the inadequacy o f roads until recently, and the high cost
o f transport. little stimulus has been, given to agriculture and
consequently the traditional system o f subsistence has not been
greatly m odified. Since this m ode o f subsistence was congruent
with the pre-European social organization the effects o f m odem
changes on the M w inilunga N dem bu have, until the post-w ar
period, been m uch less radical and far-reaching than in other parts
o f N orthern Rhodesia, such as the N orthern Province. In the
last few years, how ever, changes brought about b y the grow in g
participation o f N dem bu in the Rhodesian cash econom y and
an increased rate o f labour m igration, have in som e areas, notably
the pedicle, drastically reshaped some institutions and destroyed
others. In this book I shall consider in greatest detail the m ore
conservative area o f ex-S u b-C h ief M ukangala w here the effects
o f change have been hitherto less prom inent.
Social change o f the type w e have been considering is to some
extent reflected in T able I w here population densities in the pedicle
chiefdoms are compared w ith those in other N dem bu chiefdom s.
Population figures w ere kindly made available to m e b y the
T otai . P opuxation in N dembu A rea
N ative authority

Area (sq. iw.)


D ensity
(per sq. m .)


3 0 12

Pedicle chiefdoms :
Ik e le n g e ...............................
N y a k a s e y a ..........................


Total pedicle chiefdoms .


* 1,5 * 9


Other Ndembu chiefdoms :



3,3 5 *

3 *r7

Total other chiefdoms




Total Ndembu _ .....................

2 ,9 9 4







O '

) 1

t j '


Schism and Continuity in an African Society

D istrict Com m issioner, M r. R.. C . D ening, and. are based on taxregisters. The areas o f chiefdoms were plotted on a map o f
M w inilunga D istrict made b y M r. D ening. Table I shows
the m uch higher density o f population in the pedicle, especially
in C h ie f Ikelenges area,1 some reasons for w hich have been
previously considered. M ukangalas area, where m ost o f the
field, research for this study was conducted, is included in K anongeshas area b y Governm ent and must have roughly the same
population density as the latter. It is an area o f lo w density, and
in M w inilunga D istrict by and large lo w density is associated w ith
cultural persistence and higher density w ith culture change. In
M ukang alas area, and particularly am ong the K aw iku, the
traditional econom y is still dominant, including a certain am ount
o f hunting and collecting. In the next section o f this chapter m y
account o f the N dem bu econom ic system is founded on observa
tions I made in that area.
M cC ulloch 2 has summarized the available literature on the
physical environm ent and the main features o f the econom y o f the
M w inilunga N dem bu. I give below an outline o f her account
o f the physical environm ent.
The Ndem bu region is part o f the Plateau region o f the north
western area o f N orthern Rhodesia,3 a very ancient land surface
form ed upon the basement com plex o f Archaean rocks (in
tensely folded and metamorphosed gneisses or schists, quartzites,
limestones and shales, w ith intrusive granites). A ltitude at
M w inilunga station is 4,500 to 5,000 feet. M ost o f the D istrict
is fat w oodland w ith outcrops o f hills at Kalene.
The rains usually begin in O ctober but sometimes in September.
T h e heaviest rain falls during N ovem ber and Decem ber, the main
planting months, and the w et season ends in A pril. D uring the
period between 1926 and 1937 annual rainfall at M w inilunga
ranged from 47*88 in. to 74*83 in. w ith a mean o f 54*75 inT he mean m axim um temperature is 84o, and the m axim um
m onthly temperature remains very constant except in September
1 See also Map 2 , showing high concentration o f villages in the pedicle.
2 McCulloch, pp. 2-4, 14-17.
3 Trapnell, C. G . and Clothier, j . . Ecological Survey o f N orth-W est Rhodesia
(i9 3 7 )*

H istorical and Ecological Background


and O ctober, w hen it rises to 89*6. T he m onthly minima show

a greater range o f fluctuation, from 44*6 in June and July, to
6o*8 in January and February. T he year thus falls Into three
regimes o f temperature : N ovem ber to A pril, an equable period,
during w hich the m axim a and m inim a show a small range ;
M ay to August, characterized b y lo w m inim a ; and September
and O ctober, w hen high m axim a are registered. In the cold
season night frosts sometimes occur.
The soils o f M w in ilu n ga1 are m ostly o f the Kalahari contact
type, ranging from pale-coloured sandy loams and clay-sand
soils to better-class fight-reddish and brow n soils o f similar
texture, and representing denuded o r m odified remains o f a
greater extension o f the sands. W ith the exception o f redder
samples in M w inilunga D istrict they are o f lo w productivity.
M ost o f the D istrict is covered w ith Brachystegia woodlands.
Thick Cryptosepalum forest occurs in the south o f the Ndem bu
region between the K abom po, Lunga and M anyinga rivers. D ue
to its situation close to the Z am beri-C ongo D ivid e, M w inilunga
District is w ell w atered and in fact is a source-area o f m any rivers,
most o f them perennial. T h e Lunga is the o n ly fa irly large river
in the region o f the M w inilunga N dem bu. Grassy plains some
times surround rivers for some distance from their sources. D ry
patches o f grassland are not infrequent.
Game was form erly abundant, especially on river plains, and in
grassy tracts w here w oodland is stunted : but the gam e has been
almost exterm inated in the pedicle o f land in the north-w est o f
the D istrict that projects betw een A n gola and the B elgian C ongo.
In the rest o f the D istrict w oodland, duiker and bush-buck, and
grassland dik-dik, are still quite plentiful. B u t hunting, form erly
the dominant m asculine activity, has been considerably restricted
by Governm ent and N ative A u th ority rules and orders. M en
must obtain a licence to purchase a gun and the num ber o f licences
issued every year is small. Perm its m ust be obtained to buy
gunpow der or cartridges, and these are granted on ly for lim ited
quantities. C ertain types o f traps such as pitfalls and snares are
forbidden. There is a system o f Controlled Areas b y w hich the
D istrict is divided into First and Second-Class Controlled Areas
1 Trapnell and. Clothier. May be consulted for details and for an excellent
map showing the distribution o f soil and vegetation types.


Schism and Continuity in an A frican Society

and U ncontrolled Areas. In a First-Class Controlled Area,

where gam e is plentiful, even residents need a perm it to hunt.
These are given o n ly to know n hunters in the area, and the
D istrict Com m issioner has arranged w ith the Gam e Departm ent
for the Governm ent C h ie f to have a m ajor hand in control. In
Second-Class Controlled Areas, those w ho do not reside there
must have perm its to hunt.
T he N dem bu region is classified as N ative Trust Land w ith the
exception o f some C row n Land near the jun ction o f the Lw akela
and W est Lunga rivers, set aside as Forest R eserve in 1952, and
the territory occupied b y M ission Stations and b y H illw ood and
Mafconchi Farms.

N dem bu are shifting hoe cultivators and hunters.1 T heir

staple crop is cassava, o f both the sweet (manihot palmata) and the
bitter (imanihot utilissima) varieties. Finger-m illet is grow n,
m ostly for the m aking o f beer. M aize is cultivated in streamside
gardens (matempa) in black alluvial soil and also around villages,
in addition a w ide variety o f subsidiary crops are cultivated,
including beans, groundnuts, sweet potatoes, yam s, cucurbits,
tomatoes, onions, potatoes, cabbages and a num ber o f small
relish plants. B u t w ith the exception o f beans and sweet potatoes
these are grow n on ly in small quantities.
T h e m ain gardens (maha) are those containing cassava and, in
the first year o f cultivation, some finger-m illet. In the area w ith
w hich I am m ost fam iliar, that o f the deposed Su b-C h ief
M ukangala im m ediately to the north o f the D istrict Head
quarters, the usual practice is to clear high Marquesia or Brachystegia woodland grow ing on red o r orange loams and to heap
up the felled branches in irregular piles, occupying anything from
an. eleventh to an eighth o f the cleared area. These are then
burnt and the resulting ash-patches (mashita) are broadcast w ith
finger-m illet. Som e maize, pumpkins and other cucurbits are
planted on the ash-patches, but m any mashita m ay not be planted
at all i f seed is not available. T he finger-m illet is generally sown
in D ecem ber and harvested in M ay and June, w hen the rains are
1 McCulloch, p. 14, writes that * the Southern Lunda were in the past
primarily hunters and subsistence cultivators *.

H istorical and Ecological Background


over. A t the onset o f the next rains, in O ctober, the garden is

hoed uj> into mounds and planted w ith cassava cuttings. A t the
end o f the follow in g rains, i.e. about eighteen months after
planting, some cassava is taken from the garden and thenceforth
continuously fo r about tw o-and-a-half years until the crop has
been used up. T h e land is abandoned after one o r sometimes
tw o plantings o f cassava on it, and the bush is then allow ed to
regenerate com pletely, resulting in a fallow period o f about
thirty years.
W e studied the pattern o f land-holding o f
a vicinage in Septem ber 1953, in the neighbourhood o f M w inilongs airfield between four and seven miles from the Bom a 1
along the Bom a-K alene m otor road. T he gardens w ere plotted
b y compass and tape b y m y w ife and an African assistant trained
in surveying m ethods b y the A gricultural Departm ent. It was
found that fo r a total population o f 461 the area under fingerm illet and cassava cultivation totalled 441*1 acres, or *96 o f an
acre per person. Streamside gardens under m aize cultivation
totalled 52*5 acres, or *11 o f an acre per person. T aking both
bush and streamside gardens into account the total land-holding
o f the vicinage am ounted to 493*6 acres, or 1*07 acres per person.
I showed these figures to M r. Allan, form er D eputy D irector
o f Agriculture in N orthern Rhodesia. H e considered it reason
able to assume that cassava gardens are cultivated continuously
for either four o r five years and then left fo r thirty years to re
generate, and that the cultivable percentage o f land suitable fo r
the fertility requirements o f cassava under the vegetation-soil
conditions prevailing in M w inilunga lay approxim ately between
20-40 per cent o f the land surface. O n this basis he calculated
that the critical population o r Im d-carrying capacity for the
Ndem bu system o f land Usage probably lay betw een 17-38
persons per square m ile. These figures represent the approxim ate
population lim it w hich cannot be exceeded w ithout setting in
m otion the process o f land degradation. T h ey w ill be exam ined
later in relation to village m agnitude and distribution.
The Social Organization o f Gardening
In general it m ay be said that tree-foiling and clearing and
burning the undergrow th are the w o rk o f men, w hile planting,
1 Headquarters o f the District Administration.



' '


e .

Schism and Continuity in an A frican Society

weeding and barvesting, save for cassava w hich is planted by

both sexes, are wom ens w ork.
Gardens are owned and w orked m dividually and m ay be made
wherever suitable land is available. U sually the individual
members o f a fam ily m ake adjacent gardens. A polygynist
husbands cassava garden is in the centre o f a clearing and those
o f his w ives at either side. A com m on pattern is for a man, his
w ife or wives and a favourite sister to w ork contiguous gardens.
Brothers and male inatxilineal parallel cousins tend to cultivate
in separate blocks o f gardens. A s m ay be seen in M ap 3, the
holdings o f a single village m ay be scattered in discrete blocks
over a w ide area.
T he cutting and clearing o f bush, and later the initial hoeing
up o f the cleared ground into mounds, m ay involve a collective
w ork-party (chenda) o f kin and neighbours. W hen a man wishes
to fell trees and dear bush fo r a fm ger-m illet garden he lets it be
know n that he has brewed beer or killed gam e for those par
ticipating in a chenda to be held in a few days. Such a chenda is
restricted to men and older boys. T he beer is displayed at the
beginning o f the w ork w hich commences at dawn and continues
until the sun becomes very hot. I f m uch beer has been brewed
m uch bush w ill be cleared ; i f little, a m uch smaller area. The
productive individualism o f N dem bu finds expression in the
grum bling and mutual recriminations over the am ount o f w ork
to be done, w hich take place between the sponsor o f the chenda
and the volunteers w ho attend. Since m ost adult m en sponsor
w orking parties, it means in effect that the m ajority o f men in a
vicinage w ork collectively for each other in turn. Each man
sponsors a single chenda only, so that when it is over he has to
com plete the clearing b y himself.
Hoeing parties for first-year cassava gardens m ay be sponsored
b y men or wom en and contain members o f both sexes. T he
m en proceed in advance o f the w om en hoeing up the rough
ground, and the wom en follow , breaking up the soil m ore finely
and m aking neat mounds. M en and w om en usually w ork as
married couples, although a brother and sister or father and
daughter m ay w ork together. W h en the w orking party is over
the sponsor subsequently w orks alone or in partnership w ith his
or her spouse.
Streamside gardens are ow ned and w orked individually b y

H istorical and E cological Background


wom en only. Gardens around the village peripheries or in old

village sites are ow ned and w orked individually b y members o f
both sexes. V illage gardens, in contrast to bush and streamside
gardens, m ay be ow ned b y youn g unmarried individuals o f
either sex.
In the past m arriage was uxorilocal for the first year, in the
course o f w hich the husband had to build a hut for his m other-inlaw and clear and hoe up a garden fo r her w ith his w ife's help.
Then he took his w ife to his ow n village and cut, cleared and
hoed up the rough ground into a garden for her. In subsequent
years a m uch larger share o f cultivation fell to the w ifes lo t and
she was also expected to help her husband in his ow n gardens.
Nowadays, the w ork for the m other-in-law is often com m uted
into a cash paym ent, ranging from 10s. to 1 or m ore, depending
on the wealth and status o f the parties involved.
Agricultural production, then, is pronouncedly individualistic
in character. C ollective w orking-parties do take place but they
are brief and sporadic and perform ed for individuals. There is
no concept o f a jo in t estate, w orked collectively b y a village or
lineage-segment, the produce o f w hich is ow ned in com m on.
Marital and fam ily teams frequently w o rk together but b y no
means necessarily or invariably. M en and w om en ow n their
own gardens and spouses have no rights in one anothers gardens.
O n divorce a w om an m ay continue to w o rk in, and feed her
children from , her ow n garden at her ex-husbands village i f she
lives nearby. I f a man dies his w id o w s standing crop is sold on
her behalf b y her nearest m atrilineal kinsman when he comes
to take her back to his village after the period o f m ourning is over.
Today, when the surpluses o f subsistence crops are sold for cash,
and when in some areas new ly introduced crops such as rice are
grow n solely for the cash they bring in, husband and w ife retain
for their separate use the m oney obtained from the sale o f all
their respective produce. W hen m y w ife and I attempted to
collect fam ily budgets w e found that spouses never pooled
their incomes and that few o f them knew h ow much cash the
partners had obtained.
A lthough agricultural production is dom inantly individualistic,
consumption is m ainly com m unal. In villages the men eat in


Schism and Continuity in an African Society

the central thatched shelter (chota or njangu) and the w om en take

turns in cooking fo r the w hole group. In die village I knew best
there was a clockw ise rotation in w h ich each adult wom an in the
circle o f huts cooked successively fo r all the men. Cassava mush,
the invariable staple o f every meal, was prepared b y a wom an
and her daughters. T he dried cassava roots had been previously
pounded into flour b y one wom an, or m ore rarely a team o f
w om en, ranging from tw o to four. Such pounding teams con
sisted o f a pair o f co-w ives, or a group o f sisters, or a m other and
her daughters. W om en, girls and wicircum cised boys ate their
food, cooked on a separate fire from that on w hich the m ens
food had been prepared, in the kitchens, either in fam ily groups
o r w ith friends. Sometimes an elem entary fam ily, including the
husband, w ould eat together in the w ifes kitchen. In the
northern pedicle this arrangement is becom ing m ore com m on
today than the collective meal o f the m ens group. This is yet
another aspect o f the general breakdown o f the traditional social
organization in this area w here subsistence cultivation is steadily
givin g w ay to petty com m odity cultivation. It w ill give w ay
m ore rapidly w hen access to urban markets becomes easier and
when new local markets com e into existence.
Generally speaking, the bulk o f the cassava required fo r the
m ens food is taken from the mens gardens and the m ajor
quantity for the fam ily meals from the w om ens. V isiting kin
and friends are fed from the gardens o f their hosts, not those o f
the spouses o f their hosts. B ut w ives are required to process the
food and cook for their husbands* guests, and failure to do so is
placed high am ong the causes o f early divorce.
In the total agricultural system, from the clearing o f gardens
to the cooking o f food, w om en play a m uch m ore important role
than men. M en w o rk in short spectacular bouts o f energy ;
the regular patient labour o f the w om en in hoeing up mounds,
w eeding, digging up roots, soaking, carrying, drying, pounding,
and sifting, and finally in cooking the cassava mush provides the
steady sustenance o f the group. N dem bu consider a balanced
m eal to consist o f cassava mush and relish. E ven when meat or
fish is unavailable as relish wom en can substitute beans, cucurbits
and small relish plants from their streamside gardens.
M ale activity, generally in the co-operative form o f the chenda,
sets in train, as it w ere, an agricultural process which is then

H istorical and Ecological Background


maintained throughout the year b y w om en, each w orking alone

in her garden o r fitfu lly assisted b y her husband. In agriculture,
as in other aspects o f N dem bu life, fem ininity m ay be equated
w ith continuity, m asculinity w ith discontinuity.
Nevertheless, the overall productive role o f m en is not regarded
by Ndem bu as being m erely ancillary or peripheral to that o f
the wom en. T he men have a role in the productive system w hich
;V is autonomous and from w hich w om en are debarred not m erely
by its inherent dangers and difficulties but in addition by a
number o f ritual taboos. This role is that o f hunter.

Hunting: A Masculine Occupation

It m ay almost be said that the N dem bu social system is pivoted
> on the im portance o f hunting.1*3 This im portance does n ot derive
from the objective contribution to the food supply made b y the
chase. H unting owes its high valuation, on the one hand, to an
association consistently made am ong m any Central and W estern
Bantu between hunting and high social status, and on the other,
to an identification made am ong these peoples o f hunting w ith
There are num erous traditions ascribing the foundation o f
Central and W est-C entral A frican states to hunters. In other
societies in this area hunters appear as an aristocratic class. T h e
Lunda em pire in the C o n go w as said to have been founded b y a
Luban hunter called Chibinda (m eaning * hunter *) Ilunga, whose
son became the first M w antiyanvw a. T h e Luba state itself w as
established b y * a tribe o f hunters com ing from the great lakes
to the north*.8 T h e O vim brnidu chiefdom o f V iye (Bihe)
was founded b y a hunter o f Lunda origin. A ccording to Capello
and Ivens * the * real aristocracy am ong the Jinga o f western
Angola was com posed o f hunters and warriors \ Baumann
points out that the muri class o f nobles am ong the M bala w ere
hunters. T orday 4* writes that * the ch ie f character in the


1 In this they appear to resemble the N orthern Chokw e o f whom M cCulloch,

p. 35 writes : * N orthern Cholcwe society is oriented around the importance
o f hunting in its econom y.*
* Dias de Carvalho, p. 54.
3 Capello and Ivens, vol. ir, p. 69.

4 Torday, E ., * O n the Ethnology o f the South-W est Congo Free State*,

p . 152.

Journal o f the R o y a l A nthropological Institute (19 0 7 ),


Schism and Continuity in an African Society

peculiar politico-religious revolution which resulted in the institu

tion o f the riamba cult am ong the Bashi-lange was a hunter, and
the pakassero revolutionary society [sic.] o f Ladislau M agyar
(p. 2 6 6 ) was a society o f hunters \ Baumann * writes o f the
C hokw e that there is a hunting class w ith special privileges
and a high social status, and Chokw e religious and m agical rites
emphasize the importance o f hunting.2 T he Bangala tribe o f
W estern A ngola w ere founded b y a Lunda hunter, Chinguli.
These examples could be greatly m ultiplied from the literature,
and are cited here m erely to demonstrate the high value set on
hunting throughout the Central Bantu culture area, especially
am ong groups o f Lunda and Luba origin or affinities. M any o f
these tribes have traditions o f protracted m igration from one
region to another, during w hich they raided settled peoples for
their crops and killed game for their meat supplies. E ven when,
like the Bangala and Shinje o f western A ngola founded b y Lunda
leaders, they finally established themselves in a single area, their
villages still practised a semi-nomadic w ay o f life, livin g in m ore
or less tem porary villages and m oving to new sites w hen game
was scarce. It has been pointed out b y de Prville s that m any o f
these hunting tribes inhabiting the forest country to the south o f
the C ongo Bend grow cassava. He endorses D r. Livingstones
observation * that cassava-growing involves very little w ork
and that it is m ainly carried on b y the hunters w ives. This is
certainly true w hen one considers the extent o f masculine partici
pation in the agricultural w ork o f a m illet or sorghum econom y,
carried on b y chitemene ash-planting techniques. I do not mean
to suggest that grain cultivation is incom patible w ith an emphasis
on hunting but m erely wish to indicate that the extrem ely im
portant status o f hunting in the N dem bu hierarchy o f values is
congruent w ith a type o f cultivation w hich does not make heavy
demands on mens time and energy.4 Since cassava was intro1 Baumann, H. von, Lunda : bei Bauern und Jgern in Inner-Angola, Ergeb
nisse der Angola-Expedition des Museums fr Vlkerkunde (1935), p. 11.
2 Torday writes (Cam p and Tramp in African W ilds (1913), p. 208) : * the
most striking feature o f the Kioko in later times is that hey were essentially
a nation o f hunters and. ironworkers, but principally the former.*
3 Prville, A. de, L es Socits africaines (1894), p. 195 et seq.
4 Cassava certainly anchois a group to the ground for at least four years
after planting, but during this period men are less heavily involved in agri
culture than in a grain economy.

H istorical and Ecological Background


duced into Central A frica on ly in the sixteenth century and most

authorities;, are in agreement that the first m igrations from the
Lacustrine R egio n took place at an earlier date, it m ust be assumed
that the original hunting invaders did n ot possess cassava. B ut
its rapid and ready acceptance b y tribes w ith a hunting tradition
attests to its com patibility w ith hunting in the total system o f
The im portance o f hunting, then, am ong the N dem bu, a
migrant group from Luunda, has its roots in tradition. B ut its
value is also accentuated b y its role in the conflict between the
sexes.1 In the idiom o f N dem bu ritual, hunting and m asculinity
or virility are sym bolically equivalent, and the sym bols and gear
ofhuntsmanship are reckoned to be m ystically dangerous to female
fertility and reproductive processes. Thus, when a w om an is in
labour, her husband must rem ove his hunting equipm ent and
medicines from her hut and kitchen, otherw ise his lu c k and skill
at the chase w ill fail and she m ay die in labour o r have a mis
carriage. Conversely, a w om an m ust n ot approach too closely
a hunters village shrine or she w ill have a protracted and painful
menstruation. The same supernatural penalties w ill also be im
posed on her i f she approaches too near to the circum cision
medicine o f a senior circum ciser (tnbimbi), w hich, like m any
hunting-medicines, is placed in the cleft o f a forked stick (chishitiga
or muchanka). This m edicine (njutzda) is the supreme sym bol o f
masculinity as opposed to fem ininity, and the close parallel between
the effects o f its supernatural sanctions against w om en and those
o f hunting m edicine and apparatus supports the association w ith
huntsmanship. * For the man, huntsmanship ; fo r the wom an,
procreation * (neyala wubinda, namumbanda tusemu), is the Ndem bu
saying. This is illustrated b y a folk-tale in w hich a man, w hen
his village is raided, snatches up his gun and hunting gear, w hile
his w ifes first thought is for her baby. She runs into the bush
w ith her child, w hile he flees w ith w hat he m ost values, his
hunting things. Later she goes to her relatives. W hen her
husband follow s afterwards she refuses to return w ith him . H e
1 Conflict between the sexes is ritually expressed among a number o f related
tribes in western Northern Rhodesia and eastern Angola, including Iamda,
Lwena, Chokwe, Mbunda and jLuchazL Cf. M . Gluckman's The Role o f
the Sexes in W iko Circumcision Ceremonies *, in Social Structure : Essays
Presented to A . R. Radcliffe-Brow n , ed. Fortes (1949).


Schism and Continuity in an A frican Society

demands his child. H er relatives say, * N o , the child is hers, for

she has saved its life. The gun and axe are yours, for your only
w ish was to save them / Ndem hu, both men and wom en, put
forw ard this story as an argument in favour o f matrilineal descent.
Cassava, maize, sweet potatoes and other crops cultivated and
tended b y w om en frequently appear in rituals as emblems o f
fem ale fertility. T he conflict between m ale and fem ale is ritually
expressed in terms o f each sexs dom inant contribution to the
m ode o f subsistence. H unting and cultivation are com plem en
tary activities but they are also opposed and com petitive. In the
past, w hen hunters w ent into the bush they left their w om enfolk
undefended. I f a hunter was killed his w ife and children w ere
fo r a tim e left unprotected. A t certain times o f the year, a mans
econom ic roles as hunter and gardener m ight com e into com peti
tion. T h e first rains w hich softened the earth and made it
pliable fo r the hoe also stimulated the flush o f fresh leaves and
grass w hich attracted the antelope and made them easy prey to
the gun and bow . H unting brought m en together and divided
them from w om en. A s w e shall see, this probably had im portant
effects on the social com position o f residential groups, and the
form o f post-m arital residence.
The Social Organization o f Hunting
H unting, like agriculture, m ay be carried on either collectively
o r individually. N dem bu classify hunting techniques into tw o
m ajor categories and to some extent these determine the social
organization o f the hunt. Hunters are divided into hunters w ith
gu n s1 (ayiyanga ; sing, chiyangd) and hunters w ith bows and
traps {ayibinda ; sing, chihinda). T he chiyang*a tends to make a
profession o f hunting, devoting as litde tim e as he can to gardening
activities and as m uch as he can to roam ing about in the bush1
1 Gann, p. 36, states that4in. 1897 the Western Lunda had still been reported
to be without guns, but by the time the officials o f the British South Africa
Company began to enter their country [1904], they were well supplied*.
The Ndembu themselves say that they first obtained gum from Ovimbundu
traders and already possessed a limited number o f muzzle-loaders before the
Chokwe raid (c. 1885). One old man o f about eighty years o f age told me
that when he was a boy (c. 1880), the cult for hunters with, forearms, called
W uyang'a (see Chapter Ten) had only recently been introduced from tribes
to die west (Lwena and Luchazi). It is likely that Ndembu have had guns
since about 2870.

1 3


T h e V ailue

set o n

H u n t sm a n sh ip

Four hunter-adepts play wooden stridulators before a shrine for hunterspirits during a performance of the W uyang'a ritual. The winnowing basket
on the antelope horns contains herbal medicines to make the hunters invisible
to their prey and to attract animals within range of their ancient and unreliable

H istorical and Ecological Background


in search o f gam e. Frequently, he obtains vegetable produce in

exchange fo r m eat. Ayiyang*a participate in a cu lt in w hich
there are four degrees o f initiation. M em bership in this cu lt
enables them to range freely over Nclem bu ana indeed over
Lunda territory, for w herever they g o they are sure to fold a
cult-fellow w h o w ill give them tem porary hospitality in return
for a share in their kills. T he chiyang*a, the * big hunter *, w ho
is also a * gun-hunter tends to seek out the bigger game, such
as buffalo, eland, gnu, and roan and sable antelope, although most
If! o f his kills consist o f w oodland buck such as duiker and yello w backed duiker and small plains animals such as dik-dik. He is
infrequently a solitary, going w hither he w ill, and return%
ing to his ow n village w hen he chooses. H ow ever, ayiyang*a
occasionally form tem porary partnerships w ith other hunters,
especially i f dangerous animals such as buffalo are being pursued.
Sometimes, too, they take w ith them young m en w ho have
shown an early flair fo r hunting and w h o have been apprenticed
to them b y their fathers or mothers* brothers. M ost ayiyang a
endeavour to train their sons in hunting and w oodcraft, and the
father-son link am ong hunters is ritualized in a num ber o f w ays.
A hunter father is also entitled, to bestow one o f a lim ited num ber
o f names norm ally reserved fo r hunters on his son, even although
the latter m ay n ot be a hunter. In his life-tim e a hunter m ay
train a num ber o f apprentices : sons, m atrilineal kin, and even
the jun ior relatives o f his friends and neighbours.
A chiyang a m ay lead a sm all band consisting o f men and boys
o f his ow n village o r o f contiguous villages into the bush on quite
long expeditions lasting a w eek or m ore. O n such an expedition
the party build grass or le a f huts near an area know n to be fre
quented w ith gam e and continue to hunt until they have killed
several animals. T he m eat is cut up in to joints w h ich are sm oked
on a rough fram e o f branches to preserve them . T h e size o f the
band varies from about three to ten. T he boys, often in their
early teens, are taken into the bush fo r training and to carry the
meat back. I f the expedition is a lo n g one the boys carry supplies
o f vegetable food such as cassava roots and. green maize-cobs
for the party.
In the past there w ere annual organized gam e drives in the socalled ikuna areas, w here the trees are short and stunted and the
grass long and rank. These drives no longer occur am ong the


Schism and Continuity in an A frican Society

M w inilunga Ndem bu, but Singleton-Fisher1 has described a

gam e drive am ong the related Lunda o f south-western Katanga.
T he ikuna areas provide shelter fo r gam e when, the surrounding
bush has already been burnt, and the rem aining ikuna is burnt
in an organized w ay. T he operation is directed b y the village
headman \ T he ikuna is fired, and about thirty m en and boys
armed w ith guns and bows and arrows encircle it and k ill the
animals as they run from the flames. According to SingletonFisher, the village headman, w ith the headman from a neigh
bouring village, presides over the distribution o f the gam e
obtained. Each hunter w ho has killed an animal is given a
special ritual portion, including head, neck, heart, lungs and
intestines. This portion, called the * head * [mutu) or yinjila
(literally forbidden or * sacred * things), m ay be eaten only
by the hunter and is thought to maintain his m ystical powers
o f huntsmanship. H e is also given a leg. The remainder is
shared out am ong all those w ho have taken part in the burning.
Such drives are apparently organized b y any headman w ho
takes the initiative, and not necessarily b y a ch ief or senior
Hunters w ith bows and spears possessed much the same type
and range o f organization as hunters w ith guns. B ut whereas
almost every m an hunted w ith bow and trap, hunting w ith guns
was confined to a select few , and o f those few o n ly hunters w ith
great skill and ritual prestige belonged to the higher grades w ithin
the Wuyanga cult. N early all m en made traps and snares, o f
a w ide variety o f kinds, and each trap and the animals caught by
it belonged to the individual w ho made the trap. I have re
corded several disputes in the village courts between trap-owners
and men detected in the act o f stealing game from their traps.
It m ay be said that although in m any respects hunting conform s
to the pattern o f productive individualism 2 found in agricul
ture, nevertheless, it involves on the w hole a greater degree
and intensity o f co-operation than does crop-grow ing. M en
hunt in bands, men and w om en perform m uch o f their gardening
x Singleton-Fisher, W ., * Burning the Bush for Game , African Studies,
v, i ( 1948), pp. 3<5-8.
2 C f. Torday, op. cit., p. 15a : * the profession o f hunting naturally induces
an adventurous and self-reliant character, and encourages a roving disposition/

H istorical and Ecological Background


alone. It should be noted that the m ajority o f productive

activities in volvin g collective w o rk are carried on b y men
clearing the bush, hunting, and also hut-building, w hich w ill be
referred to below .
Ndem bu m ight hunt anyw here w ithin Lunda territory and
were not confined w ithin a restricted area.1 It was custom ary
for a hunter to give the chest or a thigh o f any large animal he
had killed to the senior headman o f the area in w hich he hap
pened to be hunting. He could then carry the rest o f his lo ll
back to his village fo r distribution. Certain plains and thicket
forests o f Cryptosepalum trees {mavunda) had a w ide reputation
for containing m uch gam e, and hunters w ould com e to them from
far and w ide. T he hunters cult and the great lateral range o f
remembered kinship ties w ere utilized b y hunters to obtain
hospitality and shelter in areas other than their ow n .
Meat Distribution and Consumption
The pattern o f distribution o f m eat varies w ith the size and
com position o f the residential unit. I f this unit is small, contain
ing a sm all bilateral extended fam ily (ntang*a), in addition to the
; portions reserved for the hunter and those given to the senior
headman or chief, a back leg w ill be given to the hunters brother
or m others brother or is divided betw een several brothers ;
a back leg or a front leg w ill be given to his m other ; a front
leg w ill be divided am ong his sisters ; the saddle w ill g o to his
w iv e s; the breast w ill go to his father ; and any small pieces
that remain w ill be distributed am ong boys too old to live in
their parents huts. In the larger villages containing tw o or m ore
minimal matrilineages (see page 8on.), a leg m ay be allocated
to the senior man o f each lineage, and it w ill be further sub
divided by the latter am ong the m arried m en o f his lineage.
Consum ption o f gam e, like consum ption o f cassava mush, tends
to be collective. It is considered to be good manners for each
man o r w om an w ho has received m eat to have a portion o f it
cooked for the mens eating-group in the village shelter, each
1 M cCulloch, p . 16, writes that * m ost hunters confine their activities to
within twenty miles o f their village, and to their tribal areas \ B u t I have
known hunters who have gone forty to sixty miles from their village to hunt.
This m ay have been due to the disappearance o f gam e from much o f north
western Mwxnilunga, com pelling hunters to rove further afield.


: D










Schism and Continuity in an A frican Society

retaining a share for his or her own. elem entary fam ily. A good
deal o f grum bling goes on both over the precise division o f the
m eat and over the amounts cooked for the mens group. Again,
hunters are frequently accused o f having made a kill and eaten
the meat themselves in the bush, feigning on their return that
they w ere unlucky in the chase. Throughout the econom ic
system a tension is set up between the individual producer or
killer o f food and the group w ho b y custom have claims in it.
A greedy person *, one w ho persistently retains w hat he pro
duces for himself, ultim ately m ay be expelled from the village.
Conversely, a professional hunter w ho resents the claims o f
rem ote classificatory kin on the product o f his hunting m ay lead
aw ay from a village a small group consisting o f his w ives, and
children, and his sisters children, to found a new village.
Successful gun-hunters are regarded as sorcerers,1 w ho acquire
their pow er in hunting from killing people b y means o f their
familiars. That is w h y great hunters seldom becom e successful
headmen, in the opinion o f Ndem bu. T heir nom adic inclination,
their tendency to favour prim ary rather than classificatory kin in
their ow n villages, and their association w ith sorcery, disqualify
them from perform ing a role w hich requires tact, generosity to
classificatory kin and strangers, and constant participation in the
group life, for its successful functioning. A bow hunter m ay
often make a good headman, a gun-hunter but seldom does so.
Summary o f the Main Features o f the Mode o f Subsistence
N dem bu practise a form o f subsistence cultivation in w hich
cassava-growing is associated w ith hunting. In addition to
1 The rancour and jealousy aroused in the situation o f meat distribution are
not only expressed in a number o f folk-tales (tuheka or ayiskitnu) but also
appear to influence strongly many beliefs about witchcraft and sorcery. A
witch or sorcerer kills his victim * for meat \ whatever his other motives for
killing are thought to be. Witches gather in covens to participate in necro
phagous feasts on the bodies o f their victims. In Social Drama I (p. 95),
and IV (p. 148), quarrels over meat distribution provide grounds for subse
quent accusations o f sorcery and witchcraft.
Professor Monica Wilson discusses a similar connection between witchcraft
and the lust for meat among the Nyakyusa o f Tanganyika ( Witch Beliefs
and Social Structure \ T he Am erican Journal o f Sociology , Ivi, 4, Jan. 1951).
She writes that * witches lust for meat and milk the prized foods o f the group
and it is this which drives them to commit witchcraft. They delight in
eating human flesh and gnaw men inside, causing death to their victims .

H istorical and Ecological Background


cassava, finger-m illet is grow n b y small circle ash-planting

(chitemene) methods m ainly fo r beer-m aking, and maize is
cultivated in streamside gardens fo r food and beer. O ther
crops are grow n in bush, village and streamside gardens m ostly
to provide relishes. In agricultural production and in the pre
paration o f food w om en have the dom inant role. H unting is
a male m onopoly. A lthough in both agriculture and hunting
production tends to be individualistic in character, where con-K operation is w ider than that between members o f a fam ily it
rends to be betw een men, w hile w om en w o rk on their ow n.
J =v|4 ; But members o f an elem entary fam ily often w ork together,
though co-w ives seldom assist one another in gardening. In
-.the sphere o f consum ption, on the other hand, the emphasis is
I V V ' on sharing the product o f gardening o r hunting am ong kinsfolk
, living in jo in t residence. Here again the larger collective units
0f consumption tend to be m ale village kin, w hile the wom en
eat separately w ith their unmarried daughters and uncircumcised



T H E V IL L A G E : T O P O G R A P H Y A N D D E M O G R A P H Y
H E N dem bu are settled in discrete circles o f huts w hich
are n ot scattered evenly over the w hole region but tend to
be loosely grouped into spatially distinct vicinages. These
circles o f huts are traditionally know n as nyikala (sing, mukala).
This term m ay be translated as * village \ T oday, N dem bu
distinguish villages entered on the Governm ent T a x R egister
(and still called nyikala)t from other residential units w hich have
not yet been so entered. The latter are know n as * farms *
(majwami), an English term said to have been borrow ed from
African tribes livin g near the Copperbelt w h o possess similar
units. Few farms are older than eight years. T he term * farm *
conceals tw o clearly distinguishable types o f residential grouping.
T he first o f these represents the first stage in the life-cycle o f the
traditional N dem bu village ; form erly such incipient villages
w ere know n as nyana ya nyikala, * the children o f villages V the
original village being regarded as the parent o f the new one. The
second is a new kind o f settlement, possessing certain special
features o f hut arrangement and social com position unknown to
the pre-European N dem bu social organization. These features
w ill be discussed in subsequent sections.
I propose to use the term registered villages * for the villages
appearing in the tax register, unregistered villages * for new ly
established villages organized on the basis o f traditional principles
o f social structure, and farms * for the new type o f settlement.
It m ust be noted, how ever, that at the present tim e there exist
a num ber o f transitional types between these kinds o f residential
unit, and that it is often difficult to distinguish unregistered
villages from * farms * unless one has personal know ledge
o f the econom ic interests and social ambitions o f the head
1 Sometimes the new village was known as mukala watwansi, * village o f the
juniors \


0 @

0 Q Q



^ cP manwsa




V IL laJ
r s

a n d



MARIO T f t - J J


D a u a V U t-} f

ELeW lfc**".

- j|p

\ I i MM vp QIH
tA A 'O rA

h ^ ssss? *



7fcAS0Nl>A a d



Y A W A .J -,

g S G
< /

mukaa^ ta / r e z Z M jF
i*: S i * G
y/UAAG /
/%"-------- moHma^KSlo.\
' '
J ^
g e R , . _ y f ' H A H A L > S A U > Y ,,r *
t U C A N A n r f
/ ' i{
II* M U K A N A / l ^ f f
* ** . & * n >
* ), I
I g8S8C
[^ * ^ '
H Y A M W A H A .F T
\ \
y /
y ft-l*
_ PfeSB'
ilIf I JctrA
^oMbA.cir Z o L lfflN A 'S V t > .
'"-. i 2?XS^li&OW&uji^MiiKRM^AS trfS ^ t y '/
(1|l| |k' 1
Il I I
>&A *
X U r Z U U I M M / ' S A K H flO
' "yivqm&#
u.;:/*;/ IH
n niTHJI
'SMaifijE l t / ,
r c y w
^ Cf
rn ) _ >/
. .
W A M VM NCW *'.- /
I y
J W W V r t lV A M I * S o i '
(f a r K N S < M iS E I-t/ lS S o o j ___ ..
'*---------------------- ---------------- ----------------------.ftfH n w .M (o'w 3
iw ic w j

ja m r o ;

k w k m

kabk o a





jN VA .vw e frtA ,G 4



o S e v^ V ' Wy
ij .|

C S AJK A ^ A O , ?
r f |im

p =si<n ?




A A T iK I^ ^




"J^ X J

K |A



f a - *.V t_VAMf*>



.1 AftAM

A A A M (J7


/ " 7 W * T IO ta , H l


V -.j


<1 m ............
fc V T r ' K lA 4fll9
t 19

t o

_CHAMSALQT, ft 2.0


CHAM W w t




\ X
1 M h in ^ A W V K G L .

... iiiillii!!!>!li!nih,.

i ^y a k a l u ^ a




C M w u to N e '


The V illage : Topography and D em ography


i Village Topography
The registered village consists o f a roughly circular or oval
arrangement o f huts around a cleared space in the centre o f w hich
is a round unw alled shelter (chotaor njangu), used as a men's
palaver hut and mess-room. M ap 4 shows a typical registered
village, Mtxkanza.1 Conspicuous features include the con
centration o f ancestor shrines before the huts o f the headman
and the senior representative o f another village lineage, the
maize patches around the village periphery manured by kitchen
rubbish, the m eal m ortar sites shared b y specific groups o f wom en,
the cassava-drying platform s, and the absence o f grain granaries.
The ancient hut type was the round or square grass hut, and
'. it was replaced in the present century b y the pole-and-m ud hut,
' which was introduced, say inform ants, from A ngola b y Lwena
;V immigrants. T he grass huts survive as seclusion huts for girls
undergoing the puberty ritual and as hunters* tem porary bush
camps. I am told that villages in the rem ote past w ere m erely
hunting camps o f grass huts, easily made and easily abandoned.
H owever, during the slave-raiding epoch they w ere usually
palisaded w ith strong stakes and sometimes w ere encircled w ith
high earthworks.
T he diameters o f village hut circles are fairly sm all and range
between about thirty and seventy yards.
T he most com m on type o f hut w hen I was first in the field
was the rectangular or square pole-and-m ud hut. H ut measure
ments lie betw een 9 ft. b y 9 ft. to about 15 ft. by 25 ft. for this
type. W hen I left, the pole-and-m ud hut was everyw here
giving w ay before the house o f sun-dried K im berley brick.
Some o f these exceeded 60 ft. in length, although the average was
about 30 ft. Pole-and-m ud huts usually contain o n ly a single
room , although som e m ay have tw o. M ost K im berley-brick
houses have at least tw o room s, and a few , the houses o f chiefs,
have as m any as six. A t one tim e there was a certain correlation
between size and type o f hut, and social status ; but the recent
1 As it was in 1953. Frequent reference will be made to this village in the
text. I have concealed its identity under a pseudonym and altered the names
o f its inhabitants. I have adopted this procedure elsewhere in the book whereever necessary.


Schism and Continuity in an A frican Society

Governm ent p olicy o f inducing all men to build brick bouses is

obliterating this distinction.
T he forest is cleared all round Ndem bu villages in a wide
ring to prevent damage from trees blow n over b y the d ry season
winds or felled b y die lightning strokes o f the rainy season.
Since the first cassava-gardens are made close to a new village i f
soil conditions perm it, m ost villages are clearly set o ff from the
environing deciduous forest o f the D istrict.
W hen a village moves to a new site in the past this took
place every four or five years at the end o f the rains circular
grass huts are built for the first few months or sometimes even
fo r the first year o f residence. Then the pole-and-m ud huts o f
the headman and senior m en o f his genealogical generation are
built for them b y the younger men, in an approxim ately semi
circular form ation, leaving fairly w ide intervals between huts.
W hen these are com pleted the m en o f the jun ior, proxim ate
generation build pole-and-m ud huts in a sem i-circle opposite
their seniors, assisted b y their sons and sisters sons. M ature
m en o f the ju n io r alternate generation to the headman m ay
then build huts between those o f their grandparents.
Unregistered villages are often built today, when village
m obility is less, w ith w ide intervals between the huts to leave
room for future members. In the past this was less com m on
since the diameter o f the hut-circle could soon be altered at the
next village m ove.
Farms present a quite different spatial arrangement from
villages. T h ey are no longer a circular grouping o f huts but
typically consist o f one or more K im berley-brick buildings
flanked on either side b y a few small pole-and-m ud huts and
kitchens.1 Occasionally, a hedge o f flow ering shrubs m ay be
planted in front o f a farm to screen it from the m otor road. M ost
true farms are situated beside the m otor roads, fo r the typical
farm -head is a m an w ho has earned m oney, often on the line-ofe
rail, and w h o intends to earn more locally. H e m ay be a petty
trader, a tailor w ith his ow n sewing-m achine, a 4tea-room *
proprietor, a * beer-hall * owner, o r a peasant producer raising
cash crops. For all these purposes easy access to m otor roads
is necessary and propinquity to administrative and trading centres
1 See inset, Map 4, Sandombu Farm.

The V illage : Topography and D em ography


advantageous. For these reasons m ost farms are found in the

northern pedicle, or near the B om a, or close to chiefs capitals,
or in the vicin ity o f M ission Stations i.e. w herever there are
large concentrations o f population and good com m unications
w ith urban areas.
Village Magnitude
Ndem bu villages have a sm aller average m agnitude than those
o f m ost other Central A frican peoples about w hich data are
available. I mapped, and collected genealogies in, fifty regis
tered villages, eighteen unregistered villages, and nine firm s. In
addition, I obtained census inform ation in sixteen o f the registered
villages, eleven o f the unregistered villages, and tw o o f the farms,
in w hich I collected genealogies. These genealogical and census
studies w ere examples, rather than a random sample. T h ey
include a high proportion o f settlements near to m y first and
second tour camps in Ikelenge area and Kanongesha area and
along or w ithin easy reach o f the m otor roads. T he proportion
o f registered to unregistered villages and farms is definitely
incorrect, since in Ikelenge A rea alone, in 1953, there w ere 15a
unregistered villages and farms as against 101 registered villages.
Nevertheless, I believe that the inform ation regarding the
relative average m agnitude o f registered and unregistered villages
and farms is n ot far from being correct. Estimates b y eye o f
the num ber o f huts in m any settlements 1 w here I did not collect
genealogies suggest that these settlements had closely similar
means and ranges to those in m y sample.
There was a total o f 657 huts in the 50 registered villages w here
I collected genealogies, the smallest having four huts and the
largest fifty-fou r. The mean num ber o f huts per village was
131 and the median was 12.
For unregistered villages there w as a total o f 135 huts in 18
villages, the smallest having one hut and the largest 16. T he
mean num ber o f huts per village was 7*5 and the median was 8.
For farms there was a total o f 23 huts in 9 forms, the smallest
having one hut and the largest 5. T he mean num ber o f huts
per form was 2*6, and the median was 3.

1 1 employ the term * settlement * as a generic name for registered and

unregistered villages and forms.


Schism and C ontinuity in an A frican Society

The distribution o f huts over the three types o f settlement is

shown in Table II.
For all types o f settlement it was found that in 77 settlements
there was a total o f 815 huts, or an average o f 10*6 huts per


um ber o f

H uts


S ettlem en ts



t )

settlement. T he median was 9 huts per settlement. B ut it must

be remembered that m y sample was not stratified in accordance
w ith the proportions existing between these three categories o f
residential unit in the total Ndem bu area. In the pedicle, for
example, farms and unregistered villages outnum ber registered
villages. O n the other hand, in the rest o f the N dem bu area,
registered villages probably outnumber the other tw o categories.
From the census sample it was found that there was a total
number o f 467 inhabitants living in 198 huts in the sixteen
registered villages, o r an average o f 2*35 persons per hut. In
the eleven unregistered villages 158 persons occupied 72 huts,
or an average o f 2*2 persons per hut, w hile there w ere 15 persons
livin g in the six huts o f the tw o farms, or an average o f 2*5
persons per hut. In the total sample 640 people occupied 276
huts, or an average o f 2*32 inhabitants per hut.
Thus the average registered village o f about 13 huts contains
about 30 persons, the average unregistered village o f 7-8 huts
about 16 -17 persons, and the average farm o f 2-3 huts about
58 persons.
In all villages w ith more than tw enty huts, it was found that
the main village circle contained in no case m ore than tw enty

T h e V illage : Topography and D em ography


huts. W hen this num ber was reached, either a n ew hut circle
was started a score or so yards aw ay, or new huts w ere built
outside and along the periphery o f the main ring.
I summarize below some inform ation on the size o f N dem bu
villages as com pared w ith the size o f villages in a num ber
o f Central A frican peoples.




Ngoni of Fort Jameson 2 .

Bemba : 3
Com moners..........................
Nyakyusa4 .*..............................
Plateau Tonga 6 .....................
Unga of Lake Bangweulu * .
Lamba, Mushiri reserve 7 .
Lamba: 89
Old A r e a ................................
New A r e a ..........................
Intermediate Area
. . . .
Kaonde8 ...............................
Lunda N d e m b u .....................

N o . o f huts
per village

N o. o f
per village

per -village

Up to 200-300
(divided into
sections o f
20 huts each)


17 (mean)


n*r (mean)
io*6 (mean)


1 McCulloch, M., pp. 66-7.

2 Barnes, J. A., * The Fort Jameson Ngoni Seven Tribes o f British Central
Africa, ed. Colson and Gluckman (1951), p. 206.
3 Richards, A. I., * The Bemba o f North-Eastern Rhodesia \ Seven Tribes,
4 Wilson, Godfrey, * The Nyakyusa of South-Western Tanganyika *, Seven
Tribes, p. 270.
* Colson, E., The Plateau Tonga of Northern Rhodesia*, Seven Tribes,
p. h i .
* Brelsford, W . V., Fishermen o f the Bangweulu Swamps, Rhodes-Livingstone
Paper No. 12 (1946). P- 3 *
7 Allan, W., Studies erf African Land Usage in Northern Rhodesia, RhodesLivingstone Paper No. 15 (i9 4 9 ). P* 27.
* Mitchell, J. C. and Bames, J. A., T he Lamba Village, Communications from
the School o f African Studies, New Series No. 24 (1950)* pp- 21, 29.
9 Watson, W ., * The Kaonde Village \ T he Rhodes-Livingstone Journal,
xv (July, 1954), PP* <5- 7 -


Schism and Continuity in an A frican Society

These figures are presented m erely to give a rough impression

o f com parative magnitudes, and not to make precise distinctions.
T he material available is in some cases not strictly comparable.
For instance, the data cited from Richards and W ilson w ere
collected before the Second W orld W ar, and it is highly probable
that there has been a marked reduction in the average size o f
villages in both areas since 1946. O n ly in the cases o f the Unga,
Lamba, Kaonde and N dem bu were the total num ber o f villages
and their range o f size stated. B ut in spite o f the obvious
deficiencies in the data it w ould seem that N dem bu villages have
a m arkedly smaller average size than those o f the other peoples
shown, w ith the exception o f the Kaonde, eastern neighbours
o f the M w inilunga Lunda. Today Ndem bu villages probably
have a sm aller average size than the figure given (io*6 huts)
since there has been a big increase in the num ber o f forms since
In The Lamba Village (p. 2) the term * O ld Area refers to
Reserve N o. 1, N dola D istrict, w hich was considered to be
overpopulated in relation to the requirements o f the agricultural
system. * Intermediate Area refers to n area o f C row n Land
reopened to Lam ba settlement where the people w ere bound
to observe certain agricultural rules, and where their villages
were laid out for them. In the * N ew Area * the boundaries
o f the area w ithin w hich a village could make gardens w ere
mapped b y agricultural officers according to the estimate o f its
needs o f arable land.
For the M w inilunga Lunda, W hite 1 gives a historical reason
for the small size o f villages. T he Lunda w ere never strong
enough to repel C hokw e and Lwena invaders, even i f concen
trated, and therefore preferred to seek safety in dispersion.
W hite suggests that the continued preference o f the Lunda for
living in sm all villages is due to their * general conservatism * :
* kinship ties remain stronger, and the village headman is m ore
respected, than am ong the Lwena or Luchazi. H ow ever, m y
older informants have repeatedly declared to me that when
they w ere you n g (some fifty years or so ago, w hen the slaveraids had hardly ceased), some villages were m uch larger than
those o f today. Some o f them have even told m e, from m em ory,
1 Cited by McCulloch, p. 24.

T he V illage : Topography and D em ography


how m any huts there w ere in the village in w h ich they grew up,
and some villages today w hich have o n ly a dozen huts once
contained about thirty. E arly travellers am ong the Southern
Lunda,1 adm ittedly for the m ost part am ong the Lunda o f C h ie f
Ishinde to the south o f Kanongeshas chiefdom , m ention * large,
stockaded villages * in the Rabom po area. H arding, for instance,
writes o f a Lunda induna w h o resided in a fortress o f no mean
dimensions, enclosed w ith a high earthen w all \ O n the other
hand, older N dem bu have told m e o f a tradition that long,
long ago \ before the slave trade, their ancestors used to live in
small encampments o f grass huts, o f similar shape to those
described and illustrated b y sketches fo r the Lunda o f M w antiyanvwa b y Dias de C arvalho.2 These settlements w ere appar
ently little m ore than hunting camps ; after destroying the game
in a given area the village w ould m ove on. I f there is any
historical truth in this; tradition then, the present sm all size o f
villages represents partly a return to an ancient type, and partly,
in the case o f * farms *, a response to changing socio-econom ic
conditions in Central A frica. This response has been assisted,
no doubt, b y the fact that Governm ent in spite o f its * ten tax
payers ru le did not m ake determ ined attempts to im pose on the
M w inilunga Lunda the rule that a village should contain a
minimum o f ten taxpayers. Since 1950, in fact, the p olicy o f
the D istrict Com m issioner has been to allow the unchecked
grow th o f sm all forms, so that in accordance w ith central govern
ment p o licy the * parish * rather than the village should in tim e
become the basic unit o f N ative Adm inistration. It w ould
seem, how ever, that some D istrict Officers made sporadic
attempts to com bine several sm all villages into single residential
units (see, fo r exam ple, p. 182).
Further evidence against the view that dispersion in the foce
o f slave-raiders was the on ly cause o f the small size o f villages is
found! in. the foct that the m ajor slave-raiders o f the epoch in
W est Central A frica, the C h okw e and Bangala, themselves lived
in small villages. Capello and Ivens,3 when passing through
1 Livingstone, op. dt., p. 283 ; Harding, p. 137 ; Cameron, V. L., Across
Africa (1885), p. 405 (of Lunda between sources o f Lulua and Zambezi ).
2 Dias de Carvalho, p. 220.
3 Capello and Ivens, voi. i, pp. 236-8.



1 ::








Schism and C ontinuity in an A frican Society

C hokw e territory in 1877, m ention m any sm all * hamlets \

* miserable little hamlets (vol. 1, p. 216), and they insert
illustrations o f villages ( senzalas ) o f eighteen huts and less.
In Bangala territory (vol. 1, p. 362) they m ention * a tibata or
village [which] is a favourable specimen o f its kind. The
stockade by w hich it is surrounded encloses an ample space
containing a couple o f dozen habitations/
Baumanns descriptions o f C hokw e villages, summarized by
M cC ulloch,1 show that they had a remarkable sim ilarity to
N dem bu villages. In the northern savanna country, where
hunting is the main activity, straw houses are m ore frequent,
palisades are rare, and settlements are more or less tem porary \
W ithin the villages houses are arranged round a cleared space, in
the centre o f w hich is the tsota or meeting-house. The two
village plans w hich M cC ulloch reproduced from Baumann
show that in one village, from the northern, m ainly hunting,
C hokw e, there are 14 huts ; and in the other, from the southern,
m ainly agricultural, C h okw e, there are 31 huts.
Capello and Ivens (vol. 1, pp. 217-23) describe tw o hunting
camps o f C h okw e th ey found in the bush, * o f grass-thatched
huts in an irregular circle \ containing * a score or so o f men
w hich resemble those described by older Ndem bu.
It is indeed quite probable that an ecological factor, i.e. a
com m on emphasis on hunting, played a prom inent role in
restricting village size am ong both C h okw e and Ndem bu.
Later die effects o f various sociological factors w ill be discussed
for N dem bu in this connection, but these too w ill be shown to
be directly and indirectly related to the high value attributed
to hunting. O n the basis o f their agricultural system alone there
is no reason w h y N dem bu should not live in larger villages.2
Indeed, today, in the north-western pedicle where hunting is
obsolescent, high population densities actually do subsist on
agriculture. B ut even in this area individual villages tend to be
small and are in fact becom ing smaller w ith the rapid grow th o f
* farms \ W hat one finds here is ribbon developm ent along the
road o f small settlements, a few hundred yards aw ay from one
another, each w ith its ow n headman and each containing a small
1 McCulloch, pp. 40-3, gives Baumanns hut diagrams.
2 See p. 21 for Allans estimate o f carrying capacity o f land under Ndembu
system o f cultivation.

The V illage : Topography and D em ography


corporate grouping o f kin. H ere the factor o f cultural persistence

or inertia seems to be im portant. B ut even in areas where
game is still plentiful, and where European contact and influence
is less marked, villages tend to average about a dozen huts. It
is possible that village size is here regulated rather b y the game
resources o f the surrounding bush and neighbouring plains than
by the fertility requirements o f cassava, the staple crop. In the
case o f the few larger villages fo r w hich I have records I shall
show that special factors are involved w hich maintain the
integrity o f uie settlement under a single headman.
It should, o f course, he m entioned here that the current grow th
o f small villages is a phenom enon found throughout N orthern
Rhodesia. It is partly due to G overnm ent p o licy w hich n ow
encourages the parish system at the expense o f die traditional
village ; to die pax britarmica w h ich has abolished slave-raiding,
blood-feuds and inter-tribal w ar ; and to the rapid exterm ination
o f lions. Everyw here the need to rely on ones kin and neigh
bours is being reduced b y these factors. Q uite as im portandy
perhaps, labour m igration to the urban industrial areas is posi
tively em ancipating the individual from his obligations to his
kinship group. Again, i f a man wishes to accum ulate capital to
set up as a petty trader or tailor, or to acquire a higher standard
o f Hving for him self and his elem entary fam ily, he must break
away from his circle o f village kin towards w hom he has tradi
tional obligations. E veryw here, w e see the spectacle o f corpor
ate groups o f kin disintegrating and the emergence o f smaller
residential units based on the elem entary fam ily.1
In most cases w here fission has occurred I shall show that the
parent village seldom contained at the tim e when it split m ore
than tw enty huts. In ail extant villages aanuuning around tw enty
huts pow erful social tensions exist and the lines or imminent
fission are clearly marked.
M y ow n hypothesis, necessarily tentative, about the historical
developm ent o f N dem bu villages is that before the period o f
the slave-raids m ost villages w ere about the same size as those
found today in areas rem ote from European setdements, and that
during those raids, some villages tended to unite into larger
1 For a more detailed examination o f changes brought about by the modem
cash economy see my article (with E. L. B. Turner) on * Money Economy
among the Mwinilunga Ndembu \ T h e R kodes-Livingstone Journal, xviii (1955).


Schism and Continuity in an African Society

concentrations w hile others dispersed into still smaller units.1

Several large com posite villages have been mentioned to m e by
older men. O ne such village was under the leadership o f the
* national * hero Chipenge, w ho organized his people against
the C hokw e, and it contained several groups o f N dem bu and
K aw iku w ho form erly lived in independent villages. Senior
headmen such as M ukangala Ibala controlled sim ilar composite
villages. O ther headmen, such as N yachiu, form ed com posite
villages from w hich they raided the traditional sm all villages o f
their fellow-tribesm en to sell the inhabitants as slaves to O vim bxrndu traders. Later, under the pax hritannica, these composite
villages broke up again into their com ponent groups o f closely
related matrilineal kin w ho founded separate villages.
I have discussed this question o f village size at some length
since, given the fact that N dem bu villages are inhabited by
kinsmen, lim itations on their m agnitude impose lim its on the
num ber o f kin w ho can d w ell together. I f maternal descent is
an im portant principle governing residential affiliation, and i f in
small villages lineages are stunted, it is probable that the social
mechanisms fo r prom oting cohesion and reducing conflict in
the residential unit w ill diner from those in societies practising
maternal descent w here the large village is typical. I w ill return
to this question later, m erely pointing out at this stage o f the
enquiry that there is probably a functional relationship o f some
im portance between the size and the structure o f the residential
Village M obility: The Changing Composition o f Vicinages
In the past (and in some areas today) N dem bu villages changed
their sites very frequently. O ne village, Nyaluhana, the core o f
w hich is com posed o f matrilineal kin o f the Senior C h ief
Kanongesha, is said to have m oved its site 17 times since it was
founded about 75 years ago. Its average duration o f occupation
o f a residential site is, therefore, about 4*4 years.2 M ost o f these
1 See p. 220 for a further discussion, on this point. Whether village members
dug themselves in and built stockades or dispersed and hid in the bush depended
on the fortunes o f war.
8 This agrees with Trapnells and Clothiers estimate that Lunda villages
generally move * after some four to six years * (Trapnell and Clothier,
p. 3 8 ).

The Village : Topography and Demography


moves have taken place in a narrow orbit w ithin a few miles from
the present village, but its fifth m ove took it to the vicin ity o f the
traditional capital o f Kanongesha in A ngola, and in 1908, after
returning to its present area, it w ent to A ngola again for about
ten years to escape from the British South A frica Com panys
administration. Since its return it has rem ained m ore or less
in its present vicinage. A nother village, M ukanza, o f K aw iku
origin, is said to have m oved tw enty-six times since its foundation
about a hundred years ago. This village again has tended to
remain w ithin about seven miles o f its present site, although,
when taxation was introduced in 1913, its members fled from the
vicinity o f the B om a to Ikelenges area in the pedicle about 40
miles aw ay, returning in 1919. B u t other villages have ranged
more w id ely across the face o f the country. T he village o f
Chibw akata, fo r exam ple, founded b y the w ar-leader or Kambanji
o f the first Kanongesha, at various times has occupied sites in
the extrem e south-west o f M w inilunga and in the Kosa chiefdom
o f C h ie f Saihmga, sixty m iles to die south-east o f its present
location w hich is five miles from M w inilunga Bom a. O ther
villages have m igrated into M w inilunga, before the Europeans
came, from Lunda chiefdom s in the C on go and Angola. In
general the K aw iku villages, containing descendants o f the auto
chthonous population, have tended to remain m ore firm ly fixed
in a particular locality than villages descended from the Lunda
invaders. In the past tw o or three decades m any villages have
m igrated to the pedicle area w here European settlem ent has held
out hopes o f em ploym ent fo r the m en and markets fo r vegetable
produce fo r the w om en. T od ay, a tendency towards permanent
settlement in one site, noted above, is becom ing clearly discernible.
A fter villages have m oved near a mission station, the Bom a, a
N ative A uth ority centre or a European farm , they tend to becom e
stabilized. A t the present m om ent a process o f fragm entation
o f traditional villages into sm all forms, containing elem entary
families w ith a sm all fringe o f m atrilineal kin o f the headmen, is
going briskly ahead. Since at m ost forms K im berley-brick
houses are being bu ilt it seems lik ely that these n ew settlements
w ill rem ain in their present sites fo r a long tim e.
I have made m any enquiries about w h y villages m oved in the
jast and in no case w as it stated that they m oved because accessible
and suitable for cassava cultivation had becom e exhausted. T he







t )





rt; '
K ;
t ;

I ,;

Schism and C ontinuity in an African Society

most frequent explanation for change o f site given me was that

in the past older people w ere buried under the floors o f their
huts w hich w ere then burnt dow n, follow ed by the m ovem ent
o f a village to a new site.1 Fear o f the spirits o f the dead led
to change o f site. For m any years, how ever, it is said, Ndem bu
have follow ed the European practice o f burying everyone in
graveyards. These are situated in the bush a few hundred yards
from the village. A s a result frequent change o f location has
becom e unnecessary. Another reason for m ovem ent was alleged
to be feuds between villages in the same or in adjacent vicinages.
Feuds were sometimes prolonged and the weaker villages fled to
new areas. Again, I have been told that when a village has
remained on the same site fo r a num ber o f years, larger game
and small mammals get destroyed in its immediate neighbour
hood, and the village m oves nearer to new hunting resources.
Thus the conscious m otives for change o f site are given as ritual,
political and econom ic, in that order. The econom ic m otives
are couched in terms o f the exigencies o f hunting although there
is little doubt that the proxim ity o f suitable soil (Ndem bu are
soil-selectors, choosing land fo r cassava cultivation b y the
dominant species o f trees on it) helps to determine their choice
o f one site rather than another. B ut it is unlikely that villages
w ere com pelled to m ove every four or five years b y the ex
haustion o f local soil. G iven a cultivable percentage o f about
20 per cent, a village o f thirty inhabitants, using the five-year
garden system described above, w ould require about i,ooo
acres to maintain it in perpetuity. Theoretically, therefore, i f
the population remained stationary and the forest regularly
regenerated, there is no reason w h y a village should m ove its
site at all. Factors other than those inherent in the agricultural
system were probably at w ork. A n attempt w ill be made later
to isolate and analyse the most im portant o f these factors.
M eanwhile it can be stated that N dem bu villages w ere small,
changed their sites often, and frequently m oved considerable
distances from one site to the next. W hen the new site was
w ithin a mile or so o f the old one, the old cassava gardens con-

1 It is said that after the death o f a headman, or after the consecutive deaths
o f two mature villagers, a village always moved.


The V illage : Topography and D em ography


tinued to be utilized from the n ew site until they w ere used up.
In the past, w hen a village m oved to a com pletely new area, its
inhabitants are said to have made large finger-m illet gardens in
the first year, for food as w ell as beer. Finger-m illet is a quickmatuiing crop (sown in D ecem ber, harvested in June), w hile
cassava roots take eighteen months to mature sufficiently to be
used as food. A gain, when a village m oved into a new area its
members used to borrow portions o f the standing crop in cassava
gardens from villages established there, cultivating extra land
for the donors the follow ing year. O r they m ight exchange
meat, honey or craft products for cassava until their first crops had
Villages arc rarely built in com plete isolation from neighbours.
Map 2 shows h ow villages tend to be grouped in discrete clusters,
o f varying numbers and form ation. Each such cluster is called
a chitung ili (from ku-tung a, * to build *). This term m ay be
translated as a * vicinage * or * neighbourhood cluster *. T he
vicinage has certain jural, econom ic and ritual functions w hich
w ill be discussed in subsequent chapters. Here I w ish to examine
it in relation to village m obility. Vicinages are not spatially
defined areas w ith permanent boundaries. The vicinage is not
an objectively based enduring unit, in this respect it differs
from a traditional senior headmans area or a m odem subchiefHom. A vicinage is a cluster o f villages, o f changeable
territorial span, and fluid and unstable in social com position. It
has no recognized internal organization w hich endures beyond
the changes in the identity o f the villages m aking it up. B ut
it is not just a neighbourhood round any village. As I w ill show
in Chapter Eleven, the vicinage becomes visible as a discrete
social entity in several situations, and a particular headman
within it usually exercises m oral and ritual leadership. Villages
in a vicinage, how ever, do not m ove as a unit, but each village
in its ow n tim e and to its o w n site either w ithin the same, or
another, vicinage. The frequency w ith w hich villages change
their sites means that the com position o f vicinages is constantly
changing, new villages com ing in from other areas and old
villages m oving out. It means that each vicinage is sociologically
heterogeneous, few o f its villages having mutual ties o f lineal
kinship or even originating in the same chiefHom. For example,
in the small vicinage in w hich Nyaluhana is the most im portant


Schism and Continuity in an A frican Society

headman 1 there are eleven villages and farms w ithin an area o f

about tw elve square miles. Nyaluhana was founded b y the
sister o f a Kanongesha w ho reigned about 1880. O ne un
registered village, W ukengi, has recently split from Nyaluhana
and has built about a quarter o f a m ile aw ay. A m ile aw ay is
W adyangamafu, w hich cam e in the nineteenth century from
the Lunda-Kosa chiefdom o f Nyam wana, in the Belgian C on go.
T w o miles aw ay is the K aw iku village o f Kafum bu, w hich broke
aw ay from M ukanza V illage in another vicinage. A m ile and
a h a lf on the other side is Sampasa w hich came from C h ie f
Chibw ikas area fifty years ago. Four miles aw ay in another
direction is Machamba, w hich recently broke aw ay from M w enilunga V illage, founded b y a son o f the first Kanongesha m ore
than tw o centuries ago. N ear Kafum bu are Sawiyem bi and
M ukom a Villages w hich cam e from senior headman Chikezas
area in A ngola thirty-five years ago and separated from one
another recently, the Sawiyem bi lineage having originally been
slaves o f the M ukom a lineage. N ear them again is Nyampasa,
recently founded b y a wom an-headm an m atrilineally related to
the K aw iku senior headman N sanganyi, whose village is in
another vicinage. A few small forms have detached themselves
from some o f these villages and built at some distance from their
parent villages. Num erous ties o f affinity and kinship n ow inter
link the earlier established o f these villages but few belong to the
same maternal descent groups. Villages w hich form erly in
habited this vicinage have now settled elsewhere. O ne village
in the vicinage is already m aking plans to m ove for to the south
o f the D istrict. In the rem oter, less stabilized areas I have
visited, I have seen m arked changes in the com position o f
vicinages between 1950 and 1954.
Fission o f Villages
Ndem bu thus live characteristically in loose and tem porary
collocations o f small neighbouring villages. B oth size and
m obility are influenced b y the fission o f villages w hich divides
larger villages into small units. I f fission has been accompanied
b y violent quarrels the seceding group often builds for from the
1 See also Chapter N ine fo r a detailed analysis o f the social com position o f
another vicinage.

The V illage : Topography and D em ography


parent village. T he sociology o f the situation o f fission w ill be

discussed at length below : here it is proposed to examine its
demographic effects only, the w ay in w hich settlements w ith
a com m on origin have been dispersed after fission across the
Ndem bu region.
Chibw akata V illage has already been m entioned in connection
with village m obility. This village in the course o f its history
o f perhaps tw o hundred years has ranged w id ely across N dem bu,
and even into Kosa, territory. D uring this tim e it has given
rise to no less than tw enty further villages b y fission, only one o f
which is n ot still in existence today. Som e o f them have given
rise in their turn to further villages. In 1953, 3 o f these offshoots
were in Kanongesha Area, close to their parent village, 3 w ere
in Chibw ika Area to the south, 2 w ere in N yakaseya and 1 in
Ikelenge Area in the pedicle, a total o f 9 villages in the N dem bu
region.1 In the K osa R egio n , w here Sailunga is the Senior
C h ief under the N ative A uth ority, 5 are in Sailunga Area proper
and one village has broken up there, 2 are in the form er C h ie f
M pulum bas Area, and one is in K akom a Area. O n e village
is in N yam w ana A rea and another in Kazem be Mutandas Area
in the Belgian C o n go , a total o f 11 in Lunda chiefdom s outside
the N dem bu R egion . O ther long-established N dem bu villages
have similar histories o f fission.
There is plenty o f evidence to show that the rate o f fission has
increased in the last three decades. A ll the K aw iku villages,
containing descendants o f the aboriginal population, claim to have
split o ff from N sanganyi V illage either b y direct fission or b y
fission from one o f its offshoots. I have collected inform ation in
all the extant K aw iku villages about the period at w hich they
broke o ff from N sanganyi or one o f its derivatives. This
inform ation has been checked against genealogical data and cor
relations have been made w ith historical events in the D istrict
the dates o f w hich are know n. W hereas during the w hole o f
the nineteenth century there w ere o n ly five remem bered instances
o f the form ation o f new K aw iku villages, there w ere no less than
five instances in the 1920s, three in the 1930s, and eleven, since
1940, m aking nineteen in the first h a lf o f the tw entieth century.
O nly four o f these have been * forms *, all founded since 1947.
1 See Map 2 for these chiefdoms, p. xxtv.




C ')




Schism and Continuity in an African Society

Thus in the w hole o f the nineteenth century five villages hived

o ff from one village w hile in h a lf o f the twentieth century nine
teen villages have hived o ff from five villages. I f this rate con
tinued it w ould mean that at the end o f the century nearly eight
villages w ould come from each village, an increase o f 37*5 per
cent. It is, o f course, possible that other K aw iku villages came
into existence, only to break up again, during the nineteenth
century, but careful enquiry did not reveal that this had been the
case. In the nineteenth century m y informants say that several
previously independent groups o f K aw iku came together in
quite large stockaded villages (which also contained some
Ndem bu) for self-defence. The K aw iku, in contrast to the
im m igrant N dem bu, have tended to remain in the vicin ity o f the
K aw iku Plain where they w ere first encountered b y the latter.
O f the 25 villages 14, or 56 per cent, still reside w ithin a few miles
o f the Plain, although their local m obility has been high. As
the result o f a violent quarrel in one village, M ukanza, many
villages seceded and later form ed three villages far from the
K aw iku Plain. Several younger men have established farms
in Ikelenge Area in the pedicle where they carry on trade. B ut
b y and large the K aw iku have tended to remain in approxim ately
the same area over long periods o f tim e. Each K aw iku village
has m oved its site m any times but always w ithin a m uch narrower
orbit than the m ajority o f true N dem bu villages. K aw iku and
N dem bu villages are interspersed in several vicinages.
Fission m ay occur w ithout the seceding groups departure
from the vicinage o f its parent village. This type o f fission is
becom ing m ore com m on today w hen one sometimes finds a
village surrounded by a num ber o f farms built b y enterprising
younger men w ho desire a measure o f econom ic and political
independence from the old conservative headmen. There has
been no open breach in these cases and the older m en are beginning
fatalistically, i f querulously, to accept the new order o f things
w hich is rapidly com ing about w ith the spread o f cash econom y.
U ndoubtedly, the accelerated rate o f fission in recent years is
related to a considerable increase in population in the District.
This increase is the product o f several factors. Immigration
from A ngola, and, to a lesser extent, from the Belgian Congo
into northern and western M w inilunga is perhaps the most
im portant o f these. The ending o f the slave-trade and

T h e V illage : Topography and D em ography


4 -:


prohibition b y the Governm ent o f internal blood-feuds are other

causes o f die increase. It is also possible that the introduction
by missions and Governm ent o f hospitals and dispensaries has
led to a reduction in the death-rate. O n the other hand, the
virtual exterm ination o f gam e in som e areas, a source o f meat not
yet replaced b y cattle w hich could be kept in the fly-free north
west o f the D istrict, must have led to a decrease in the consum ption
per capita o f anim al protein. This lack o f meat m ay have en
feebled resistance to disease and so prevented a m arked reduction
in the death-rate.
B u t the im m igration from neighbouring territories o f small
discrete groups o f kin w ho settle in every part o f the Ndem bu
region has undoubtedly provided a stimulus to m ore rapid fission
among villages already established there. Cash econom y tends
to destroy ties o f corporate kinship within villages. M any small
im m igrant groups, each living separately, provide for established
villages external models for a new type o f residential group. B ut
it must be stressed that these m odern trends do no m ore than
accentuate tendencies inherent in the indigenous social system .
The local kin group appears never to have been extensive, the
spatial m obility o f villages was alw ays considerable,1 fission was
relatively frequent and individual m obility was high. M odem
changes have not so far struck at the basis o f the indigenous
system nor radically disrupted it. Because o f the m obile fissile
nature o f its traditional residential units, the system has been able
to absorb changes that prom oted m obility and increased the rate
o f fission.
Individual Mobility
In addition to the high rates o f m obility and fission o f villages
themselves, individuals also tend to circulate rapidly through
villages in the course o f their lifetim es. T h e sociological factors
behind this m obility o f individuals w ill be exam ined in detail in
later chapters, but they must be m entioned at this point.
M en and w om en have different patterns o f m obility. M ost
men are bom in their fathers* m atrilm eal villages and some o f
1 See, for example, Livingstones remark that * people change from one
part o f the country to another * with frequency, quoted in a footnote in
Chapter One, p. 5.


Schism and Continuity in an A frican Society

them go to their ow n mothers* brothers* villages in their teens.

B ut if, as frequently happens, their parents are divorced, they go
w ith their mothers in early childhood to their mothers brothers
or mothers mothers brothers* villages. Some m en remain
w ith their fathers until the death o f the latter. I f a village splits,
men often have a choice o f m atrilineal villages in w hich to live.
T he m ajority o f adult men, how ever, reside avunculocally,
unless they found new villages o f their ow n. W om en m ove
m ore often than men. T hey are bom in their fathers villages ;
they m ay return to their uterine uncles villages for a w hile ; and
they g o to their husbands villages after m arriage. Since marriage
is brittle am ong N dem bu,12
*they m ay spend their reproductive
period alternating between their fathers and mothers brothers
villages, and successive marital villages. M any w om en, if
w idow ed or divorced after they have passed the menopause,
return to the villages o f their nearest m atrilineal kin to remain
there for the rest o f their lives. Thus the com position o f a
single village varies greatly from year to year as men, wom en
and children enter and leave it. This natural fluctuation is
supplemented b y other movements dependent on the personality
o f the headman. I f he is notorious as a sorcerer, or is unskilful
as a ju d ge, or mean and selfish, m any people w ill leave him .
I f he has the reputation o f being a w ise and generous man he w ill
build up a large village.
In Tables III and IV , I attem pt to express certain aspects o f
individual mobility. Table HI shows minimum m obility. It
is constructed from m y village census figures and offers an indica
tion o f w hat proportion o f adult m en and wom en live in their
natal villages.
These figures suggest that on ly one-seventh o f adult men and
w om en live in the villages where they are bom , w ith no sig
nificant difference between the m obility o f men and wom en.
H ow ever, since N dem bu are often born in one village and reared
in another w e can obtain a further index o f m obility b y setting
out a table w hich indicates the proportion o f adult men and
1 See Table VII for divorce ratios.
2 The layout o f these tables is based on Table I in E. Colsons 4Residence
and Village Stability among the Plateau Tonga, T h e Rkodes-Livingstotte Journal,
xii (i 9 5 i), P- 4 Z*

The V illage : Topography and D em ography


f women w ho have been reared in their natal villages. B y * village

o f rearing * I denote for w om en the village in w h ich they have
spent the longest period o f residence from birth to the puberty
ritual (Nkanga) w hich marks the entrance into w om anhood at
bout the age o f fifteen or sixteen. For m en I have arbitrarily
selected a sim ilar age-range, birth to sixteen years o f age, since
I n d iv id u a i . M

o b il it y

: N

atal an d

P r e s e n t V il la g e s ,

M ales

20-39 years
30-39 y ears

40-49 years

in natal

in other


(7 7 -i%)
(7 8 -4 % )

60-69 years


Total .

Fem ales
in natal

(lOO'O) %

(i 7 -o%)

(8 3 -0 % )


(9 0 -9 % )

(100-0% )

(9 *1% )

(l0 0 - 0 % )

(18 -7 % )

(0-0%) (100-0%)
(*<S-7 % ) (83 ' 3% )


(84 *9 % )



(89 -3 % )

(is -j% )

L ivin g
in other



50-59 years


b y age





(9 -1% )

(8 1 -3 % )
(9 0 -9 % )


(I3 7 % )

(86 -3 % )


(lO0O% )


(100-0% )

(l0 0 "0 % )

the boys* circum cision ritual (Mukanda) is often perform ed fo r

boys o f ten and under and is not com parable fo r our purpose
to the g irls puberty ritual w hich is usually follow ed im m ediately
by m arriage. T able IV , therefore, shows the com parative pro
portions o f those w ho have been reared in their natal villages
and those w ho have been reared in other villages. Table IV
shows that less than tw o-fifths o f adult m en and rather under
a h a lf o f adult w om en w ere brought up in their natal villages.
The slightly greater m obility o f boys than girls is probably due
to the fact that boys, after circum cision, m ay leave their fathers
villages for their maternal uncles* villages, o r vice versa, more or
less at w ill. Girls, on the other hand, tend to remain w ith their


Schism and Continuity in an African Society

mothers, or to m ove when they m ove. A t circum cision a boy

is sym bolically severed from his m other ; henceforth he is a full
member o f a male m oral com m unity, free on the w hole to choose
his site o f residence. B ut a girl does not lose her dependent status
until she has borne several children. She passes from mothers
to husbands village when she leaves m other fo r husband. Thus

M ales

in natal

20-29 years


30-30 years


40-49 years

(40 *5%)
(35 *7 % )

50-59 years
60-69 years


Total .

in other




<77 *i%)
(59 *5% )
(64 *3% )




(4 8 *3% )
(53 *0 % )

(51 *7 % )






in natal

in other

(39 *6 % )

(6 0 *4 % )







(5 9 *1% )

(4 0 *9 % )


(4 6 *9 % )

(53 *1 % )




(53 *x%)

(4 6 *9 % )


(lOOO% )



w om en have greater individual m obility than men, but boys

have greater m obility than girls.
T he natal village is in several respects an unsatisfactory point
o f departure for a study o f m obility. A person m ay be bom
either in his o r her fathers or m others village. I f the fathers
and m others villages are fairly close neighbours the expectant
m other is taken to her ow n village, usually the village o f her
close m atriiineal kin, a few days before the confinem ent, and
brought hack to her husbands village after the birth o f the child.
O r a w om an m ay go to her ow n village for the ku-tena ivumu
( m entioning the w om b *) cerem ony, perform ed at a wom ans

The V illage : Topography and D em ography


first pregnancy, and com e back to her husbands village to bear

her child. These customs underline the basically matrilineal
character o f the social structure. B ut they have the effect o f
confusing inform ants as to w hat was really their natal village
the village w here the m other actually gave birth, or the village
where she was residing w ith her husband. A better measure
I n d iv id u a l M o b il it y : V il l a g e

R e a r in g





Living in
village o f

20-29 years
3 0 -3 9


40-49 years

50-59 years
60-69 years


Total .


(2 4 *3 %)

in other


Living in
village o f



<68*8%) (100*0%) <S*7 %)

(75 *7 %) (100*0%) (18 *9 %)
(82*1%) (1000%)
(9 3 *1%) (100*0%) (20*0%)

(I7 9 %)
(6 -9 %)
(24-0%) (76-0%) (roo*o%)
(00%) (100*0%) (100*0%)


P r e s e n t V il l a g e ,

in other






(9 4 *3%) (100-0%)
(9 3 *8 %) (100*0%)
(80*0%) (100*0%)
(9 5 *5 %) (100*0%)

(4 *5 %)
(0*0%) (100*0%) (100*0%)

(79 *3%) (100*0%) (10 *7 %)



o f m obility m ay perhaps be obtained b y showing in a table w hat

percentage o f adult men and w om en continue to reside in their
villages o f rearing, whether or not these w ere their natal villages.
Table V shows this.
Table V shows that adult men have a stronger tendency to remain
in their villages o f rearing than adult w om en. This is probably
due to the fact that m arriage is virilocal. W om en are often
brought up patrilocally,1 m arry vixilocally, and after divorce reside
1 Patrilocal * refers to residence in ones fathers village. Matrilocal *
refers to residence in ones mothers village. It should be stressed that my
use o f these terms refers to residence with a parent, and not with a spouse.


Schism and Continuity in an African Society

avunculocally until remarriage. It should be noted that a much

higher proportion o f men under than over fo rty years o f age
reside as adults in their villages o f rearing : 25 per cent under
fo rty and 14 per cent over forty. This is probably due to two
main causes. T he first is that the m ore able and w ealthy men
w ho have not become headmen often establish villages or forms
o f their ow n, and the second is the transfer o f men from their
fathers to their uterine uncles villages w hen their fathers have
In m y sample I found that quite a high proportion o f the popu
lation had been born, reared, and dw elt at the tim e o f the enquiry,
in different villages. 46 out o f 179 adult men, or 25*7 per cent, and
57 out o f 197 wom en, or 28*9 per cent, had been born in one
village, bred in another, and lived at the tim e o f the census in
yet another.
T o sum up : Tables III, IV , and V show that about one-seventh
o f the census population lived in their natal villages and sixsevenths in other villages ; that about three-fifths o f adult men
w ere reared in villages other than their natal villages, as were
just more than h alf the adult wom en in the sample ; four-fifths
o f adult men currently resided in villages other than those in
w hich they had been brought up, and nine-tenths o f adult
w om en did likewise. A quarter o f the adult male population
and nearly three-tenths o f the adult wom en had been born,
reared, and dw elt at the tim e o f the census, in three different
villages. These figures give little indication o f the extent o f
spatial m obility o f N dem bu but m erely show transfers o f resi
dential affiliation from one headmans group to another. The
few individuals w ho have remained in the same village all their
lives m ay have travelled considerable distances across the region
as members o f that village. Those w ho have changed their
village affiliation m ay have m oved only a few hundred yards.
Some idea o f the extent o f individual spatial m obility may be
afforded by Table V I, w hich groups the census population into
chiefdom s 1 o f birth, rearing, and affiliation at the time o f the
1 The term chiefdom * here refers to a Sub-Chiefdom under the Native
Authority. Mukangala, however, ceased to be a recognized Sub-Chief in
1947. His area has been incorporated in that o f Chief Kanongesha.

The V illage : Topography and D em ography


Table V I shows that im m igration into each chiefdom in w hich

the census was taken, from other areas, has been considerable.
I n d iv id u a l M

o b il it y t h r o u g h

C h ie fd o m s

chiefdoms o f
in census


P/R/N %

Mukangala .
Ikelenge .

p/ r

27 40*3
2 57*8
15 45*4
14 42*4






















p/ h

Mukangala .
Ikelenge .




i i -3














*9 5


Males and Females

Mukangala .
Ikelenge .
p/ r / n
f/ r
p/ n


16 16-3
iS 23*4
XS 29*5
X2 I


Chiefdom o f
Chiefdom o f
Chiefdom o f
Persons bom
















birth, rearing and present residence the same.

rearing and present residence the same.
birth and present residence the same,
and reared outside present chiefdom.

The num ber o f im m igrants in the sample is nearly 50 per cent

o f the total population. I w ould like to repeat that the boundaries
o f chiefdom s have n ot been significantly altered during the


Schism and Continuity in an African Society

period under consideration, although m inor changes have been

made b y Governm ent. In the past the areas o f chiefs and senior
headmen w ere demarcated by topographical features such as
rivers and hills. Thus, in considering personal m obility through
chiefHoms, I am no longer analysing allegiance to a spatially
m obile unit (village), but am exam ining allegiance to a spatially
fixed one (chiefdom). Indeed a person m ay meander through
m any chiefdom s w hile rem aining the w hole w hile in the same
village. T he percentage o f persons bom and/or reared in
M ukangala area is m arkedly higher than that found in the other
three chiefdom s, all o f w hich are in the north-western pedicle
where there is m ost European settlement. M ukangala is a
deposed Sub-C h ief o f Kanongesha, his area is a * conservative *
and traditional one, and m any o f his people are K aw iku, des
cended from the autochthonous inhabitants o f the region.
M ost o f the im m igration into the pedicle has com e from Angola
and the Belgian Congo and has taken place since the beginning
o f the present century.
N dem bu society m ay be characterized as m obile rather than
stable w ith regard to residence. Villages m ove through space
and split through tim e. Individuals continually circulate through
villages. There are tw o main types o f village m obility : some
villages tend to m ove in a relatively narrow orbit, fo r instance,
K aw iku villages circulate around the m argin o f the K aw iku
Plain ; others, especially those o f Lunda origin, sometimes
m ove considerable distances between successive sites. Individuals
frequently change their village affiliations and few indeed o f the
adult members o f m y census sample w ere livin g in their natal
villages and villages o f rearing at the tim e o f m y enquiry. Since
villages are seldom large, and there is nowhere to get aw ay from
ones neighbours widiiix the village i f one quarrels w ith them,
people are constantly on the m ove. W h en villages reach a
certain * critical size they tend to split and the resultant immature
villages often m ove some distance from the parent villages. I
believe the critical sizes to have been, until ve ry recently, in
the neighbourhood o f about 20 huts and 50 people ; since 1950
they are probably less. Just as the membership o f villages is con
stantly changing so the membership o f vicinages is also changing,

The V illage : Topography and D em ography


although less rapidly. T he rate o f residential m obility con

siderably increased in the period between the w o rld wars as the
result o f im m igration into the Rhodesian N dem bu region from
Angola and the Belgian C on go, and recently attempts b y Govern
ment to stabilize the population have slow ed d ow n the rate o f
village m obility. B u t m y evidence suggests that high village
and individual m obility have always been features o f the N dem bu
social system. T h e accounts o f ethnographers and travellers
show that other groups o f Lunda origin and affiliations possessed
a similar unstable type o f social system .1 A ll these groups have
traditions o f m igration ; and although they have settled in more
or less defined territories in the last tw o or three centuries, w ithin
these territories villages and individuals have continued to range
w idely, w hile villages are ceaselessly com ing into being. M y
genealogical data suggest that m any sm all settlements have failed
to becom e established and have died out.
This high m obility in the social system m ay be associated w ith
a productive econom y in w hich men hunt fo r m eat and wom en
grow crops, and in w hich the mens role is valued m ore highly
than the w om ens. W om en also tend to accept the m ens
valuation o f hunting. T he dom inant characteristics o f the social
system at any given point in tim e m ay result from a m aledominated hunting econom y in a bush-habitat, the characteristics
being a high m obility, a lim ited degree o f co-operation, and,
where there is co-operation, this is m ainly between m en. O ther
characteristics, consistent w ith hunting, include personal in
dependence, resentment o f political control, and residential dis
continuity o f groups and individuals. O n the other hand, there
are regularities in behaviour w hich are consistent and constant
through these changes. R egularity, cohesion and persistence,
in the N dem bu system through tim e, are related to the roles o f
w om en both in production o f crops and in reproduction.
U ndervalued though it is w ith regard to its actual contribution
to the food supplies, the productive labour o f w om en provides
the regular staple fo r the physical survival o f N dem bu society.
W om en bear and nourish the children, and w here the mother
goes, w ith o r w ithout the father, the children go also. The
1 See, for example, Carvalhos {op. cit.) comments on the Chokwe, Bangala
and Songo.


Schism and Continuity in an A frican Society

present belongs to the men, and die social order at any given
m om ent bears the stamp o f male activity ; the past and future
o f the society are dominated b y die m others. T he group o f
m ale kinsm en form s the core o f the village today : tom orrow its
continuity w ill be found to have depended on their scattered
m atrilineal kinswom en. The m en are present and together
because o f dead w o m e n ; the w om en are absent because they
are m arried to other men, but their sons and daughters sons wifi
replenish the village. T he conflicts and cohesion in the kinship
and residential systems, associated w ith this dichotom y o f pro
ductive and reproductive roles between men and w om en, w ill
be die subject o f subsequent analysis.

C H A P T E R Ili
have already noted h ow , w ith the weakening o f chiefly
W authority, the village em erged as the significant local unit
among the N dem bu. B u t w e have also seen that N dem bu
viHages, like the T onga villages described b y C olson,1 * are not
necessarily enduring units w ith stable populations tied to par
ticular localities . . . villages m ay shift from spot to spot, w hile
through them stream a succession o f inhabitants \ Nevertheless,
: ft
at*any given m om ent it w ill be found that the inhabitants o f a
I li particular village are not m erely a random grouping o f unrelated
individuals but that the m ajority o f village members are linked
to the headman b y varying ties o f kinship and affinity. A n
Ndem bu village, like the Lam ba villages described b y M itchell
and Barnes 2 and the Kaonde village described b y W atson, m ay
i be defined in fact as * a local residential group w hose members
claim a genealogical relationship w ith one m an, usually the
In this the N dem bu, o f course, resemble m any
other Central A frican peoples.
In theory an individual, provided that he or she is n ot a slave,
a m arried w om an, an uncircum cised boy, or an uninitiated girl,
m ay reside w herever he or she m ay please, but in practice his or
.. - her
choice o f residence is usually circum scribed by a lim ited
o f modes o f residential affiliation, based on kinship,
affinity and slavery. T he m ost im portant o f these is matrilinea!
relationship to the village headman. H ow ever, this type o f
affiliation is drastically m odified b y the virilocal residence o f
married w om en. I f matrilinea! attachment to headmen and
virilocal m arriage w ere the sole determinants o f village affiliation,
villages w ould consist entirely o f male matrilinea! kinsmen, their
w ives and dependent children. A dult men w ould reside avunculocaBy, adult w edded w om en virilocally, and their young



1 C olson, op . cit., p . 41.

2 Mitchell and Barnes, passim.
3 Watson, p. 7.




Schism and Continuity in an African Society

children patrilocally. B ut structural rigidity o f this order is

never found in the m obile Ndem bu society. In practice, there
is a relatively high percentage o f patrilocally resident, adult men.
A number o f adult wedded wom en m ay also be found living
patrilocally. In the past, and in the more traditional areas even
today, marriage was uxorilocal fo r the first year and daughters
remained during this period in their fathers* villages w ith their
husbands. Again, young divorced wom en m ay return to their
fathers* villages i f their mothers are livin g there. Indeed, divorce
pow erfully affects the residential affiliation o f adult w om en o f
D iv o r c e R

a t io s

(T o t a l M

a r riage

E x p e r ie n c e


A. No. of marriages ended in divorce expressed

as a percentage o f all marriages . . . .
B. No. of marriages ended in divorce expressed
as a percentage of all completed marriages
C. No. of marriages ended in divorce expressed
as a percentage of all marriages except those
ended by death............................... ..... .


L iv in g In f o r m a n t s )

N o. in
census sample


347 /d58




3 4 7 /5 7 6


all ages. M arriage is extrem ely brittle (see Table VII i ), and
older divorced w om en usually go to live at the village o f their
nearest male matrilinea! kin until remarriage. A fter the death
o f their husbands, w idow s also tend to live w ith their near male
matrilinea! kinsmen until remarriage. W hen they have passed
the menopause w idow s and divorcees usually reside permanently
w ith their male matrilinea! kin. There is a strong tendency for
mothers and children to gravitate together as the mothers grow
older. The m other-child bond is perhaps the m ost pow erful
kinship link in N dem bu society. As minors both sons and
daughters tend to fo llo w their mothers wherever the latter m ay
be Hving. M ost children go w ith the m other on divorce. Boys
in their teens m ay sometimes leave their mothers and remain
1 See pp. 263, 265 for discussion on small percentage o f w idow s.

The So cial Com position o f the V illage


with their mothers brothers, but girls usually stay w ith their
mothers until the end o f the puberty ritual w hich is custom arily
follow ed b y m arriage. M others rejoin their sons, daughters
rejoin their mothers after w idow hood o r divorce. In their
m inority children live w ith their m other ; in her old age the
mother lives w ith her children again. In the intervening years
mother, son and daughter m ay spend m ost o f their tim e in
different villages, although they m ay visit one another frequently.
Patrilocal residence and frequent divorce, then, have further
modified the picture originally presented o f a village occupied b y
a core o f m ale m atrilineal kin, their w ives and children. W hat
we now find is a group o f siblings, brothers outnum bering
sisters, and som e o f their children o f both sexes. In addition
we m ay find spouses o f the above persons, w ith w ives o u t
numbering husbands, and some grandchildren.
In long-established villages a num ber o f classificatory as w ell
as prim ary kin w ill be found. T h e senior generation w ill con
tain tw o o r m ore groups o f m atrilineally related siblings and the
junior adjacent generation w ill contain classificatory as w ell as
full cross-cousins.
It w ill be rem em bered that the mean and median sizes o f
Ndem bu settlements in m y genealogical sample are small, io-6
huts and 9 huts respectively. N o w , given the tendency o f full
siblings and their m other to gravitate together, and given a
tendency for a certain proportion o f adult children o f headmen
and o f other im portant elders to reside patrilocally during the
lifetim e o f their fathers and w hile their parents* marriages remain
unbroken, this means that in m any long-established villages and
in m ost recently established villages prim ary kin o f the headmen
outnumber their classificatory kin. Thus in a recently estab
lished village containing, say, 8 huts, the headman m ay be accom
panied b y bis brother and w idow ed sister. I f he has tw o w ives
each o f them w ill have a separate hut, so that the sibling group
in the senior generation w ill jo in d y possess 4 huts. The head
man m ay have a married son livin g w ith him , and his sister a
married son and divorced daughter, m aking three huts for d ie
jun ior adjacent generation, and seven in all. T he other hut
m ight be occupied b y a male or fem ale parallel cousin o f the
headman, b y the son or daughter o f another sister residing else
where w ith her husband, o r by a sisters daughters child. O r


Schism and C ontinuity in an A frican Society

it m ight be occupied b y a member o f a maternal descent group

o f slaves 1 ow ned b y the headmans lineage, or even b y a stranger.
Ndem bu give different w eighting to each o f the factors affecting
village affiliation. M atrilineal relationship w ith the founder o f a
village gives one the righ t to reside there perm anently, other
things being equal. A person m ay stay in his or her fathers
village until the latters death, but must then leave. A free
fathers slave children are inherited b y his heir and thus remain.
This means that ones affiliation through ones m other is to a
group, whereas ones tie w ith ones father is personal. A man
is the end o f a m atrilineage, to paraphrase the Latin m axim , and
his son does not belong to that lineage. A wom ans real hom e
is where her close m ale maternal kin (her m others brothers, her
brothers and her sons) happen to be living, usually together.
These male members o f her lineage transfer her econom ic and
sexual services to a m ale member o f another lineage, but never
renounce their claim on her reproductive capacity. Eventually
her sons and daughters sons must go to them, and sooner or
later she w ill probably fo llo w them. B ut she spends m ost o f
her reproductive cycle in the village o f her husband or in the
villages o f her successive husbands. In Table VIII an attempt
is made on the basis o f data supplied in the village census sample
to discover whether m inors, under the age o f tw enty, do in fact
tend to reside w ith their mothers wherever the latter m ay be living.
Table VIII shows that about two-thirds o f the population
under tw enty years o f age, other than married girls living viri
locally, w ere residing w ith both their parents, and about a
quarter w ith their mothers only. A negligible percentage were
living w ith their fathers only, and hardly m ore w ere livin g in
their matrilineal villages w ith neither parent. M ore than ninetenths were residing in the same villages as their mothers, and
m ore than h a lf o f those w h o were not, w ere residing in villages
whose nuclei w ere their ow n matrilineal kin.
N o w let us try to find out at w hat period in the lives o f sons
this strong attachment to the mother is severed. O n ly a small
1 Slavery was formally abolished by the British South Africa Company
in the early years o f the present century. But it lingered on as a clandestine
institution in the villages for many years afterwards. See p. 187 if.

The Social Composition o f the Village


T: sample could be obtained b y com paring genealogical and census

data, but the results are suggestive.

e s id e n c e o f


M in o r s

nw edded

A ge groups

x viai


W ith whom residing


0 -4
N o.



N o.


N o.


N o.

N o.





* 3*5







3 3 '3








T o ta l.....................






5 **3



Females :
Mother and father .
Mother only
Father only .
Neither ; in village
o f matrikin .
T o ta l.....................
Both Sexes:
Mother and father .
Mother only
Father only .
Neither; in village
of matrikin .

*5 -*9

: Males:
Mother and father .
Mother only
Father on ly.
Neither : in village
o f matrikin .

14O 4 9 *<S




52 505

7 7 .7

23 22-3










3d 48*7




44 59*4 23







T o ta l ..................................................... j 103 100*0






















4 *i



47*9 *93
37*5 7 2<S*s
7 2*4






A ccording to Table IX , 21 out o f 37 men in the 2029 age

group whose m others w ere still alive w ere livin g in the same
villages as their mothers. In the age group 30-39, 5 out o f
11 m en whose m others w ere alive w ere livin g in the same v il
lages as their m others. In the w hole sample rather m ore than



Schism and Continuity in an A frican Society

h a lf o f the men whose mothers were alive were living with

them in the same village.
According to m y observations the mother-son link o f common
residence is on ly definitely snapped by the m others death. A
divorced or w idow ed m other o f advanced years w ill usually
spend her last years in her sons village. I have recorded a case
in w hich an elderly wom an actually divorced her husband in

e sid e n c e o f

A d u lt M

a les w h o se

M oth ers

a r e st il l a liv e

A ge groups

Men in same
villages as
own mothers





Unmarried .













i <5


Unmarried .









Men in different Married

villages from
Unmarried .
own mothers



order to take up residence w ith her eldest son, w ho had just

founded a village o f his ow n. This mother-son tie o f propin
quity is all the m ore remarkable when the countervailing ten
dencies set up by virilocal marriage o f w om en are considered.
The m other-child, and especially the m other-son bond, emerges
from these enquiries as the crucial link o f Jdnship and determinant
o f village affiliation. W here adult children reside patrilocally
this seems to be on account o f the childs attachment to the
m other rather than to the father. In Table VIII only 2*4 per
cent o f the 0-19 age group in the sample resided w ith their

The So cial Com position o f the V illage


fathers only, w hile 26-5 per cent resided w ith their mothers only,
follow ing their divorce or w idow hood. O f the fou r m en w ho
resided w ith their fathers only, I knew tw o very w ell, and in
both cases special circumstances w ere involved. The mothers
o f both w ere dead and their fathers w ere headmen. These
factors are, o f course, infrequently com bined, h i one case the son
feared w itchcraft in the village o f his near m atrilineal kin. T h e
matrilineal village o f the other was in A ngola and as a progressive
young man w h o wanted to earn m oney he preferred to remain
in M w inilunga in the w ealthier pedicle area.
The strength o f the sibling bond and its im portance as a deter
minant o f residence also derive from this m other-child tie. In
matters o f inheritance and succession it is m atrilineal descent that
counts ; the children o f one m other b y several fathers regard
each other as fu ll (chikupu) siblings in every respect in the social
context o f their maternal village, although they m ay visit in
their different paternal villages, and receive assistance from and
assist their father's m atrilineal groups independently o f one
another. A s the analysis proceeds, it w ill be shown h ow a
number o f social mechanisms are brought into play to counteract
this tendency for the m inim al unit o f m other and children to
cohere too closely and intensely and prevent the grow th o f
wider social units. Im portant am ong these to anticipate is an
emphasis on the social and spatial separation o f adjacent genea
logical generations and on unity w ithin each genealogical
generation. This emphasis opposes m other to child in a relation
ship o f authority on one side and respect on the other, and unites
full siblings w ith parallel and cross-cousins against the senior
generation as a w hole. T he persistence o f the m other-child link
as a determinant o f residence attests to its strength in the face
j e eration separation and o f other com petitive modes o f
W e are n o w in a better position to rate the various principles
determining village affiliation according to their degree o f em
phasis in N dem bu culture. First, there is, as w e have seen, an
extrem ely pow erful bond betw een a m other and her ow n children, acting against the dispersal o f this m atricentric fim ily and
towards its maintenance as a residential unit. B u t the principle o f
virilocal m arriage disrupts the m atricentric fim ily b y separating


Schism and C ontinuity in an A frican Society

daughter from mother and sister from brother. Henceforth

it is possible to regard the subsequent history o f the original
matricentric fam ily as an effort to reconstitute itself as a local
unit. M any sons remain w ith their mothers whether the latter
live virilocally or avunculocally. I f sons and mothers are
tem porarily separated, divorce or w idow hood o f the mother
sooner or later brings the old w om an back into the village o f
her son w ho b y this tim e m ay be fo lly mature, and w ho may
indeed have founded a new village o f his ow n. Sisters return
to brothers after their marriages have terminated and m ay often
fold their mothers there. T he m atricentric fam ily, united in its
early grow th but divided as the children grow up and themselves
reproduce, seldom comes together again in its full membership,
but it always strives to do so against virilocal marriage and the
m any other centrifugal tendencies o f N dem bu society.
Another important factor influencing residence is, as w e have
seen, the tendency o f males to com e together in kin groups to
form the persistent core o f villages. T heir collaboration is
couched in the dominant matrilineal idiom o f Ndem bu kinship
so that the village is regarded at its inception as a group o f
uterine brothers. B ut a w ider group than a pair or handful o f
brothers is required .for effective collaboration in economic
activities and in legal matters, and for offence and defence in
feuds and raids. The concept o f uterine brotherhood is extended
to include in the residential unit, m atrilineal parallel cousins.
V irilocal marriage, w hich had broken up the matricentric
fam ily, n ow appears as the ve ry means b y w hich the * adelphic
group * o f male uterine parallel cousins is enabled to remain
together. U xorilocal m arriage w ou ld disperse this group,1
1 Among the Nayar castes o f Malabar in India, sisters remained in their
brothers house. The Nayar kinship grouping was a matrilineal lineage living
on its own property. All Nayar women were married before puberty to
men o f their own caste and divorced after four days, so that children they
bore subsequently to lovers o f their own caste or higher castes would have the
correct caste on the whole for father \ Thus uterine siblings and matri
lineal parallel cousins remained together throughout life, for though each
brother might have several * mistresses * in other houses, he resided with his
sisters. In a sense, the frequent divorce among Ndem bu women parallels
the Nayar situation, since there is hardly a village in which, at a given moment,
one or more sisters are not found spending the intervals between successive
marriages with their uterine brothers anti matrilineal parallel cousins.

T h e So cial Com position o f the V illage


primary uterine kinship as the basis o f local association w ould

prevent its classificatory extension.
T he unity o f the m atricentric fam ily, virilocal m arriage, and
classificatory adelphic co-residence, are principles o f local group
ings w hich give rise to continual conflicts w ith in the village.
W e have seen h o w the ecological system o f the N dem bu, w ith
its com plem entary and opposing poles o f hunting and cassavacultivation, m ay itself be a direct source o f residential instability.
This instability is considerably aggravated b y conflicts w hich
arise out o f the social structure. In subsequent chapters I shall
examine in turn each o f the principles determ ining residence and
show h o w each interlinks persons and groups divided from one
another b y other principles o f social affiliation. In the course
o f the analysis it w ill be shown, fo r exam ple, h o w conflict
between the principles o f the unity o f the m atricentric fam ily
and virilocal m arriage tends to break up the m arital group, especi
ally in the early years o f m arriage, and on the other hand, h ow
stable m arriage tends to retard the grow th o f villages b y keeping
men apart from their uterine sisters w ith their children. It is
to a considerable extent b y divorce and w idow h ood that a village
is enabled to persist through tim e. W ith a w om an com e her
children, and i f she is divorced or w idow ed after she has passed
her menopause she and her sons com e back perm anently to the
village o f their m atrilineal kin and replace the w ives and children
o f their m ale kin w ho have gone out o f it. Thus divorce and
w idow hood act sim ultaneously as principles o f village recruit
ment and attrition.
W hat are the visible effects o f these principles o f local organ
ization on the social com position o f villages that I have actually
observed ? Is it possible to infer regularities in residential com
position from field data ? Is it then possible to construct models
o f different types o f village structure on the basis o f such observed
regularities ? A n attem pt to answer these questions is provided
in Tables X and X L T h e universe on w hich these tables are based
consists o f hut-ow ners in villages w here I have collected full
genealogies. O f the 77 villages in w h ich I collected genealogies
(see p. 37), I have used m aterial from <58 villages fo r the purpose
o f m aking these tables. T he nine discarded villages are those in
w hich the genealogical data are too scanty and fragm entary to




Schism and Continuity in an A frican Society

be suitable for use. Since in the great m ajority o f villages most .5

members are kin o f the headmen I have classified hut-owners
according to the kind and degree o f kinship or affinal links they
have to headmen. W here a person has m ore than one tie o f
consanguinity or affinity to the headman o f his village I classify
him b y the relationship w hich N dem bu hold to be the dominant
one in the context o f village membership. In establishing which
m em ber o f a m arital pair or o f a group o f im m ature boys or
girls inhabiting a single hut is the hut-ow ner I em ploy the
criteria used b y Ndem bu themselves. I f a w om an is livin g in a
village where she has a prim ary or classificatory m atrilineal rela
tionship w ith the headman, and her husband has not, she is
considered to be the hut-ow ner. I f both are matrilineaJly related,
as in the case o f a marriage between m atrilineal kin o f alternate
genealogical generations, the husband and not the w ife is taken
to be the hut-ow ner. I f a w om an is livin g in her fathers village
and her husband is not related or is a distant cognatic relative o f
the headman she is the hut-ow ner. In a cross-cousin marriage,
the spouse m atrilineally linked w ith the headman, regardless o f
sex, is die hut-ow ner. In a boys o r girls hut the oldest, linked
b y w hatever tie o f kinship to the headman, is classified as the
hut-ow ner. I f no link o f consanguinity or affinity can be traced
between a hut-ow ner and his or her headman, he or she is classi
fied as a * stranger . Strangers fall into tw o categories. They
m ay be first-generation slaves (sing, ndungu, pi. andungu) or
maternal descendants o f slaves, w h o even today m ay still be
regarded as belonging to free members ; or they m ay be im m i
grants from other areas in Rhodesia, A ngola or the Belgian
C on go w ho have attached themselves or have been attached by
a Governm ent Sub-C hief to a particular headman. I have not
been able to determine in every case w ho was or was not a slave
and have consequently been com pelled to classify all hut-owners
fo r w hom it was impossible to trace genealogical connection
w ith the headmen o f their village as 4 strangers \ I have called
a fem ale stranger a * hut-ow ner * i f she is so described b y the
headman o f the village in w hich she is living.
B y em ploying the descriptive system o f kinship nomenclature
used b y antnropologists I was able to distinguish no less than
7z different categories o f kin, in relation to headmen, living
in these 68 villages. These are subsumed in the Ndem bu


e l a t io n s h ip o f

H u t- O w ners t o H ea d m en
(male and female )


Ego (headmen themselves)



older G
younger G

of male

. . .


of male

older BC
younger BC
sibling rank)


. . .

older FC
younger FC

IS -

Children cf

older FSC
younger FSC

. . .

T o ta l............................
m Mother
G Sibling













71 176

83 58












M. F.

16 7
28 9
26 22
2 X




68 S et t l e m e n t s















188 X29






















255 102

C Child


50 132
15 31


























102 65


3 S7

l6 7



d Daughter
S Son'


L egend


B Brother
F Father


Schism and Continuity in an A frican Society

classificatory system under 14 kinship categories, or 18 i f cer

tain terms are qualified b y adjectives denoting sex. In Table X
I discriminate between matrilineal and patrilateral prim ary kin
according to descriptive criteria, and subsume all other kin
under N dem bu classificatory categories.
A number o f inferences m ay be drawn from the material
presented in Table X I. The first is that headmen and their kin
outnum ber strangers in villages by m ore than four to one. The
second is that headmen and their matrilineal kin account for nearly
T h e S o c ia l C o m p o sit io n


68 S et t l e m e n t s

Category o f hut-owners

Headmen and primary matrilinea!

k i n .........................................
Classificatory matrilineal kin
Own descendants o f male headmen
Children o f male matrilineal kin .
Patrilateral siblings and their children
Unspecified kin and affines



N o.








S 7


Total kin and affines

. . . .
Total hut-owners



4 '4




















n *4


668 OO-O

three-fifths o f all hut-owners in the sample. Children o f head

m en and o f their m ale matrilineal kin account fo r ju st under onesixth o f all hut-owners and one-fifth o f all related hut-owners.
A nother feature is the numerical preponderance o f prim ary
over classificatory matrilineal kin in a ratio o f alm ost tw o to one.
130 out o f 385 m atrilineally related hut-ow ners, or 33*7 per cent,
are w om en, an indication in a society practising virilocal mar
riage o f the strong tendency o f w om en to rejoin their male
m atrilineal kin. Since nearly all w om en are married ju st after
their puberty ritual this means that w om en found in their matri
lineal villages must have been divorced, w idow ed, or recently
separated from their husbands, or w ith husbands w ho live

The S o cial Com position o f the V illage


uxorilocally. T he evidence o f m y village census suggests that

divorce is easily the m ost com m on reason w h y w om en reside
w ith their m atrilineal kin. W om en live in uxorilocal marriage
only for special reasons, except that a few men still live uxori
locally during the first year o f marriage. T h e reasons behind
the m atrilocal and avunculocal residence o f adult w om en w ill
emerge in the course o f the analysis.
Headmen and their prim ary and classificatory brothers in both
lines o f descent account for 161 out o f 542 related hut-owners,
or nearly three out o f ten. W ith them are 49 w om en w hom
they call * sisters o f w hom 32 are their ow n sisters, m aking
up a sibling and classificatory sibling category w hich accounts
for tw o out o f every five related hut-ow ners. In this category
w e find that tw o out o f every three are prim ary siblings. This
tends to confirm the previous statement that uterine siblings
rejoin one another in the same village as they gro w older. The
table also demonstrates that whereas the ratio o f m ale parallel
cousins to brothers is 53 :4 7 , the ratio o f fem ale parallel cousins
to sisters is as lo w as 35 : 65, a fact that indicates the importance
o f the principle o f classificatory adelphic co-residence.
T he bilateral character o f the ju n io r adjacent genealogical
generation o f hut-ow ners em erges from the analysis o f the
data. Children o f headmen and o f those w hom headmen call
brothers * total 94 persons, w h ile those w hom headmen call
* sisters* children * total 134, a ratio o f about 2 : 3. For males
the ratio o f * sons * to 4sisters* sons * is about 7 : 8, and fo r ow n
sons to o w n sisters* sons approaches one to one.
I found in the villages from w h ich the figures w ere com piled,
that there was an approxim ate num erical balance between hutowners in the opposed generation categories. Thus i f hut-owners
in the headm ens generation are added to those in the second
descendant and second ascendant generations from the headmen,
members o f this linked generation category total 269 hut-owners.
I f members o f the first and third descendant generations and first
ascendant generations from the headmen are added together,
this opposed linked-generation segm ent possesses 273 hut-owners.
This num erical balance between the linked-generation segments
is to be expected, since in a given population, a hut-ow ner is
equally likely to belong to either generation-segm ent on a
random basis. Each linked-generation segm ent in Ndem bu





't "



Schism and Continuity in an A frican Society

society occupies its own. section o f the village, and, since villages
are generally circular in shape, inhabits a sem icircle o f huts
facing the semicircle o f the opposed generation segment. In
each village the number o f members in each segm ent is seldom
equal, but i f a sufficiently large sample o f villages is taken it
appears that the linked-generation categories approxim ately
balance one another num erically.
In order to bring out significant differences in social structure
between long and recently established villages I com piled a
further table illustrating such differences. For the purposes o f
this analysis I defined * a long-established village * as a village in
w hich there had been three or m ore successive headmen, including
the present incumbent, and a recently established village * as
one in w hich there had been less than three successive headmen.
This distinction does not precisely correspond w ith the distinc
tion previously made between registered and unregistered vil
lages and farms. According to the earlier distinction both
unregistered villages and farms are o f com paratively recent
establishment (cf. p. 34 and p. 36) since they consist o f settlements
not y e t recorded on the Governm ent T a x R egister. Thus by
this earlier reckoning I included as * registered villages settle
ments that had been founded m ore than about fiv e years before
the enquiry and w hich had been granted official recognition as
tax-paying units. Several o f such villages indeed had been
founded m ore than thirty years previously, but each village had
been under the authority o f a single headman throughout this
period. A long-established village w ith three or m ore successive
headmen is, in every case I collected, a village that was in existence
before the European governm ent began. There w ere 21 such
villages in m y sample and I compare the social com position o f
these villages w ith that o f 47 * recently established villages in
Table X II.
Table X II shows that in long-established villages the ratio o f
prim ary to classificatory hut-ow ning m atrilineal kin is about
1 : i *i , whereas in recently established villages the ratio is about
3 : 1 . In other words, recently established villages tend to con
sist dom inantly o f uterine siblings and their fam ilies, w hile in
long-established villages uterine kin are ju st outnum bered by their
classificatory matrilineal kin. In long-established villages also
w e find a significantly higher percentage o f cognatic hut-owning

The So cial Com position o f the V illage


kin than in recently established villages. This is because such

villages are regarded as som ething m ore than a group o f uterine
siblings. T heir membership is looked upon as a group w ith a
persistent m atrilineal core to w h ich a h in ge o f other cognates


S o c ia l C o m p o sit io n

o p L o n g - E s t a b l is h e d
w it h t h a t o f R e c e n t l y E s t a b l is h e d

M ates

V il l a g e s C
V il l a g e s

o m p a r ed


Fem ales

Category o f hut-owners

.2 'S
T3 ^
H *rt

N o.

N o.

N o.


l 8*6






H.M. and primary matrikin.

Classificatory matrikin
Descendants o f H.M.
Cognaric, patrilateral and un
specified kin, and affines
S tra n g ers..........................





3 *2
2 *4





T o t a l ...............................


71 *6





H.M. and primary matrikin.

Classificatory matrikin
Descendants o f H.M.
Cognatic, patrilateral and unspecified kin, and affines .
S tra n g ers..........................

X2 0







7 *i







T o t a l ...............................


6 9 *s





H.M. and primary matrikin.

Classificatory matrikin
Descendants o f H.M.
Cognaric, patrilateral and unspecified kin, and affines .
S tra n g ers..........................













2 *5
4 *2



T o t a l ...............................

















attach themselves. In these villages the m atricentric fam ily has

matured into the matrilmeage. T h e principal factors influencing
the persistence o f settlements beyond their initial phase as uterine
sibling residential groups is discussed on pp. 198-203. In a
matrilineal society w hich practises viriiocal m arriage, i f there is


Schism and Continuity in an A frican Society

virtually unlim ited access to resources so that settlements are

under n o constraint to occupy specific tracts o f land in perpetuity,
the transition from m atricentricity to m atriliny as the basis for
local groupings m ust always be hazardous and uncertain. It is
dependent on such factors as the ability o f the headman to keep
his follow in g together, the maintenance o f reasonably good rela
tions between the m en o f the m atrilineal core and their brothersin-law , and the biological accidents o f fertility and freedom
from disease.
The core o f the N dem bu settlement consists o f a group o f
uterine or m atrilineal kin, o f w hom the m ost senior is the head
man. V iiilo cal m arriage separates fem ale and m ale matrilineal
kin and at the same tim e enables men often to retain their
children, especially their sons, in their ow n villages after the
latter have matured. Nevertheless the dominant attachment o f
children is not to their fathers but to their mothers. Thus a man
can o n ly retain the residential affiliation o f his children w hile his
m arriage to their m other endures. O n divorce a w om an usually
goes to live w ith her nearest m ale matrilineal kin, and her children
tend to fo llo w her. T he crucial bond o f kinship is between a
m other and her children, w ho form a m atricentric fam ily. This
fam ily m ay be attached for longer and shorter periods to a
wom ans husband o r brother. As the children m ature, daughters
are separated from mothers, and brothers from sisters, b y virilocal m arriage. B u t there is a constant tendency, though it is
seldom com pletely realized, fo r the m atricentric fam ily to recon
stitute itse lf as a local unit. This is an im portant factor behind
the m obility and instability o f the individuals residence. It is
also a cause o f conflict between husbands and brothers-in-law
and between sons-in-law and m others-in-law. V irilocal mar
riage is related to the tendency for uterine brothers and male
matrilineal parallel cousins to reside together. This is part o f
a general tendency am ong Ndem bu to build up villages around
a fram ew ork o f close male kin : brothers, sons, sistere sons, and
sisters daughters sons. Evidence from matrilineal societies
practising uxorilocal marriage such as the Bem ba and Y a o 1
1 Richards, A. L, * Mother-right among the Central Bantu \ Essays Pre
sented to C . G. Seligman, ed. E. E. Evans-Pritchard, R . Firth, B. Malinowski,

The So cial Com position o f the V illage


suggests that w here the link between male village residents and
the headman is one o f affinity, and these men are linked to one
another not b y kinship but m erely b y propinquity, they co
operate very little am ong themselves. This is probably because in
most prim itive societies social control at the local level is associated
with position in the kinship structure. Senior kin exert authority
over and com m and respect and obedience from their juniors,
and betw een each category o f kinsmen custom has prescribed
an intricate and specific pattern o f behavioural expectations
which facilitates co-operation and inhibits dispute. A m ong
Ndem bu, m ale villagers co-operate in hunting and bush-clearing,
and form erly assisted one another in offensive and defensive war
fare. There was no strong central political authority and no
national arm y, and a mans first lo ya lty was to his village. For
all these purposes d ose and ready co-operation betw een men was
necessary. It was therefore appropriate that m en liv in g together
should be linked b y consanguineal kinship, the m ajor vehicle
o f social control. B u t the local attachm ent o f m ale kin probably
owed som ething o f its intensity, as w e have suggested in previous
chapters, not on ly to such em pirical factors, but masculine
resentment o f m atrilineal descent form ally expressed, as w e shall
see, in a num ber o f ritual contexts, and the decisive econom ic
im portance o f w om en. It is as though there w ere a general
though unconscious male conspiracy to exclude their female kin
from their local units. Nevertheless, men needed w om en for their
personal and corporate survival. T h e individual man ow ed his
care and nurture to a wom an, his ow n m other. Sim ilarly the
group o f male villagers needed their sisters to ensure the survival
o f their village as a social entity. T o bring them back into
the village entailed the male groups com ing into conflict w ith
other men. So that the v e ry exclusion o f fem ale kin from
villages in order that m ale kin m ight live together provided the
principal source o f conflict and unrest between N dem bu men
as a w hole. A gain , since in the econom y w om ens w ork
provided the staple o f subsistence, w om en w ere needed in the
and I. Schapera (1933), p. 267, and Bemba Marriage and Modern Economic
Conditions, Rhodes-Livingstone Paper N o. 4 (1940), pp. 33 fF. ; Mitchell,
J. C ., The Yao o f Southern Nyasaland \ Seven Tribes o f British Centra/
Africa, ed. Colson and Gluckman (195*). pp. 328 fF.


Schism and Continuity in an African Society

village. M en ow ned gardens, it is true, and w orked in them

sporadically, yet w om en had not only w orked m ore regularly
in their ow n as w ell as in their husbands* gardens but also dug
up, carried, processed and cooked the daily, never-failing supply
o f cassava. I f sisters w ere excluded, w ives had to be included
in the village personnel. In all their conscious statements about
the relative im portance o f wives and sisters, N dem bu have
continually stressed to m e that w ives outrank sisters, and that
indeed the m arital tie is the strongest and deepest they know.
In ritual, too, husband and w ife are regularly initiated together
into curative and fertility cults. In a decisive aspect the girls
puberty ritual is a drama o f marriage, emphasizing die importance
o f the conjugal tie. In spite o f this verbalization and ritualization o f the im portance o f the marital relationship, in fact Ndem bu
marriage is extrem ely brittle. This is due to the very real
im portance o f the brother-sister tie w hich is the socially, although
not biologically, procreative link in matrilineal society. M ale
kin can o n ly live together through die exclusion o f their sisters
and sisters* daughters and the im portation o f w ives linked to
other corporate groups o f male kin.1 B u t having thrust their
female kin out, the m en must reclaim these w om en and their
families i f the village is to persist through time, and is n ot to die
out for lack o f replenishment. Unfortunately fo r the men,
their ow n w ives, closely bound to their children, are subject to
the same * gravitational * compulsion from their male matri
lineal kin. E very adult Ndem bu man, under the operation o f
the same set o f social forces, is in the ambiguous position o f
striving to retain his ow n w ife and children by his side and
o f simultaneously endeavouring to w in back the sister whose
absence is the price o f his privilege o f livin g w ith his brothers and
sons in a matrilineal society. Put succinctly, to live w ith the
male kin he loses a friend to m arry an enemy. It is all part
o f the N dem bu male fligh t from the inexorable reality o f matri
lineal descent and fem ale control o f the econom ic basis o f sur
vival. Perhaps this is one factor in the Ndem bu preoccupation
1 In this discussion I have been considering some o f the factors relating to
the question o f why Ndembu are virilocal. I have taken the fact that they are
matrilineal for granted. I am also assuming here that there is a brothersister incest rule.

The So cial Com position o f the V illage


v with ritual, one aspect o f w hich is escape, and another, compen

sation, although these constitute only a m inute part o f the w hole
There is no guarantee, therefore, that a settlement w ill be per
petuated and w ill survive. I f the sisters o f the adelphic group
have successful marriages, and providing that their husbands
have long lives, it m ay be that their sons w ill live m any years
w ith fathers ; and that w hen these fathers die the sons w ill found
sibling villages o f their ow n rather than return to their uterine
uncles. It m ay be that conflicts am ong the brothers themselves,
arising from jealousies over succession and inheritance, w ill break
up a village. M uch depends in this society o f individualists,
only too ready to take offence, on the tact and diplom acy o f
individual headmen, whose best hope is not to browbeat or
domineer but to persuade and reconcile. B ut once the essen
tial link-up between first and third generations after the founding
o f a village has taken place, new sources o f conflict w ithin the
setdement com e into being. W hen a village contains only tw o
generations o f hut-ow ners it m ay still be regarded as a bilateral
extended fam ily. B u t w hen three generations o f adult hutowners form its membership, incipient cleavage along lineage
lines becomes detectable. I f a headman has tw o fertile sisters
each has becom e the founder o f a lineage. The principle o f the
unity o f the m atricentric fam ily com es into conflict w ith the
unity o f the m atrilineage, and each m atricentric fam ily or alli
ance o f m atricentric fam ilies w ith a com m on grandm other is a
potential source o f village cleavage and the potential startingpoint o f a new residential unit. H ere the unity o f siblings and
the unity o f classificatory adelphic m ale kin m ay be insufficient
to hold the village together b y cutting across lineage affiliation.
It m ay be that the original sibling-group w hich founded the
village has been reduced in numbers b y death. In any case its
authority is continually threatened w ith in the village b y the
maturation o f the ju n io r adjacent generation, the leading men o f
w hich are eager to obtain the headmanship. T o maintain its
authority w ithin the village on the one hand, and to prevent the
disruption o f the village as the result o f struggles between lineages
on the other, an alliance develops between the senior and the
second descendant generations, an alliance sometimes cemented
b y m arriage between persons w h o stand in the classificatory


Schism and C ontinuity in an A frican Society

relationship o f grandfather (nkaka) and 4grandchild (mwijikuiu)

to one another.1 W here such marriages occur they often link
a * grandfather * o f one lineage to a * granddaughter o f another.
Thus the grandparent and his siblings secure the allegiance o f the
granddaughters siblings against the thrustfui m iddle generation.
The m iddle generation, o f course, exercises authority over the
generation o f grandchildren and also threatens the authority o f
the senior generation. Grandparent-grandchild marriages also
bind together potentially conflicting m inim al 23matrilineages and
uterine sibling groups. T h ey are on ly made between classificatory kin and never between prim ary grandparents and
grandchildren. Thus they tend to occur m ainly w ithin longestablished villages. Nevertheless, in all villages containing at least
three generations o f hut-ow ners the alliance between alternate
generations, whether interm arried or not, finds physical and
spatial expression in the fact that both generations build their
huts in the same village semicircle o f huts facing and opposed
to the semicircle o f huts built by the intervening genealogical
generation. use the term 4alliance * and not 4unity or
4equivalence in describing the relations between alternate
generations, since their m ode o f association approaches m ore
closely that obtaining between cross-cousins (asonyi, sing.
musonyiY w ho interm arry, rather than that between siblings
w ho do not. Alternate generations, like cross-cousins, jo k e w ith
one another and interm arry, although the jokin g is less ribald
and m ore gently affectionate than the pleasantries o f cross-cousin
intercourse. Joking, according to R adclifle-B row n, implies a
fairly even balance between hostility and friendliness, and in fact
tension exists in grandparent-grandchild relations. In the first
place grandparents and grandchildren belong to different, and
therefore potentially opposed, matricentric families, the basic
1 It should be mentioned here that grandparent-grandchild marriages take
place not only within villages but between villages o f common origin (see
p. i7<5). Most, but not all, o f such marriages are between a man and his
granddaughter * (see p. 246).
2 In this book I regard a minimal matrilineage as a group consisting o f the
descendants through women o f a common grandmother, a minor matrilineage
as the matrilinea! descendants o f a common great-grandmother, and a major

lineage as those o f a common great-great-grandmother.

3 See Chapter Eight for analysis o f village kinship.

The Social Composition o f the Village


corporate kin units o f N dem bu society and the starting-point o f

new residential units- In the second place, grandparents, especi
ally the senior m ale grandparent, belong to the generation
w hich exerts grontocratie political and ju ral authority in die
matrilineal village or in other villages in w hich the grandchildren
m ay reside. A uth ority is ultim ately in their hands. A t the
same tim e, the position o f these tw o generations is essentially
similar vis-a-vis the m iddle generation as w e have seen above.
Again, the older needs the younger generation to replace as
allies the siblings w ho have died, and the younger needs the
older to m itigate the severity o f the intervening generation o f
uterine uncles and senior affines. T heir m utual inter-dependence,
strengthened b y m arriage and brother- and sister-in-law ties o f
fam iliarity and friendship, acts to bind together a residential
group constandy liable to disruption from lineage and adjacent
generation conflicts.
In the subsequent analysis I shall isolate each dom inant principle
governing village residence, and exam ine Its operation, illustrated
b y case m aterial, in prom oting cohesion or generating conflict
w ithin the local unit. This exam ination w ill concern itself not
only w ith secular behaviour but also w ith ritual. Afterw ards,
the w orking o f each principle in the field o f inter-village relations
w ill be analysed. Finally, there w ill be a discussion o f modes o f
social integration, other than those derived from kinship, and
these w ill be exam ined in terms o f their congruence w ith and
opposition to the kinship principles already considered.


H ERE is sense in regarding N dem bu villages as colloca
tions o f matricentric families interlinked by varying ties o f
Idnship and affinity. In addition, som e measure o f continuity
is provided by maternal descent w hich determines succession and
inheritance. In long-established villages w ith respected head
m en something like hierarchically organized lineages o f the type
made familiar in British anthropological literature b y the re
searches o f Evans-Pritchard and Fortes begin to em erge, although
such lineages are shallow b y comparison w ith those found among
such matrilineal and uxorilocal people as the H opi and Ashanti.
'1 able XIII shows lineage depth in 64 village genealogies. Farms
have been excluded from the sample and also several villages for
w hich m y data are unsatisfactory. Table X V appears to indi
cate that there is a significant increase in the m agnitude o f a village
i f six generations o f matrilineal kin o f its headman can be recorded
on a genealogy. Thereafter village m agnitude seems to remain
fairly constant regardless o f recorded lineage depth. T he differ
ence in genealogical recall between the older villagers o f ninehut and thirteen-hut villages respectively tends to correspond
w ith the difference previously made between * long and recently
established villages * (see Table X II). It tends to correspond also
w ith differences in social structure. R ecently established villages,
villages w ith less than three successive headmen, consist mainly
o f uterine siblings, their adult children and a few mature grand
children. Long-established villages, on the other hand, m ay con
tain m any adult members o f tw o or m ore segments o f equal
status w ithin a minor or m ajor village m atrilineage (see footnote
on p. 80). B u t few long-established villages, whatever the
generation depth o f their nuclear lineage, attain to a great
population size. For fission, occurring m ost frequently between
uterine sibling groups and minimal lineages, and virilocal mar
riage, and various kinds o f individual m obility, all prevent their

M atrilineal D escent


A lthough there is no absolute correspondence between the

lim it o f recall and the span o f the residential segm ent, one tends
L ineage D e p t h


N o. o f
g en era tio n s in
v illa g e g en e a lo g ie s


64 G en ea l o g ies

G e n e r a t io n

N o. o f
v illa g es

G en era tion
o f headm an




Total . . . .



H eadm an

N o. o f
v illa g e
h eadm en




Mean lineage depth . .

Median .
. . .


Mean generation level

Median .
. .
. . . . .




V il l a g e S iz e


N o. o f
g en e ra tio n s in
v illa g e g e n e a lo g y

L in e a g e D e p t h

M ea n n o. o f
h u ts in v illa g e


I 3 -I


8 & over

Mean 6*r


I z '9

to find a greater depth o frecall b y the headmen o f long-established

villages than b y those o f recently established villages. This is
because it is o f im m ediate interest to distantly connected kin in


Schism and Continuity in an A frican Society

long-established villages, to trace precise genealogical relation

ship to one another as members o f a single local com m unity.
O f course, one m ay find that headmen o f recently established
villages m ay be able to trace back their matrilineal ancestry to
quite a considerable depth. This is especially the case i f they
belong to the maternal descent group o f a ch ief or senior head
man, and therefore are possible successors to an im portant office.
B ut it is usual to find that headmen tend to trace back their
descent from that ancestress w hom the matrilineal core o f the
village have in com m on, and to forget her predecessors. Struc
tural amnesia typically begins above the ancestress whose genea
logical position is structurally significant for the living members
o f the village matrilineage. I f rem oter ancestresses are recalled,
one finds oneself in the field o f supra-village relations.
The lineage am ong the Ndem bu is not, as am ong the Tallensi,1
* the skeleton o f their social structure, so m uch as the endproduct o f a number o f social tendencies w hich under specific
circumstances place a curb on the m obility and instability o f
residential groupings. I discuss these tendencies in the subsequent
analysis. A m ong them m ay be mentioned a time-honoured
village title, continued proxim ity to an area where gam e is
abundant, and the good reputation o f its headman at the crucial
transition period from sibling group to structured lineage.
A m ong the Tallensi the lineage is ab initio a basic premiss o f social
organization. A m ong the N dem bu the local matrilineage is a
goal, and an ideal to w hich am bitious headmen aspire. O nce a
village has becom e established and consolidated, great efforts are
made not on ly b y its headman, but also b y those w ho hope to
succeed him , to prevent the disruption o f its widest span lineage.
T he form and direction o f such efforts w ill be considered in a
series o f detailed studies in this and the follow in g three chapters.
In m ost cases these efforts are ultim ately unavailing since the
ecological and structural pressures m aking for fission are too
strong, but the efforts are nearly always made. The basic unit
in Ndem bu society is not the lineage, but the matricentric fam ily
w hich in its life-cycle becomes, after the death o f the mother,
the uterine sibling group. This and not the elem entary fam ily is
the basic unit, since frequent and easy divorce, often consequent
1 Fortes, op. cit., p. 31.

M atrilin eal D escent


on the conflict between a w om ans husband and her brothers for

custody o f her and her children, renders the fam ily unstable and
impermanent. T he uterine sibling group is a constant source o f
danger to the continuity o f the village lineage ; the narrower unit
is the foe o f the w ider. It w ill be show n below that the most
frequent unit o f secession from a village is the uterine sibling
and n ot a lineage o f w ider span. Bach xnatricentric
/ is given an early autonom y from other m aternally-linked
matricentric fam ilies b y the institution o f virilocal marriage.
Hence w hen the m atricentric fam ily fin ally returns to the village
o f its nearest m atrilineal kin, it is seldom assimilated fu lly into
the latter, and tim e and again its narrow er loyalties prevail against
the value set on village unity. W e have seen h o w vinlocal
m arriage disperses the members o f the uterine sibling group for a
time, and h o w they tend to divest them selves o f other attach
ments in order to becom e reunited as a local autonom ous group.
The success o f the narrower against the w id er unit as a residential
nucleus is attested b y the small average m agnitude o f N dem bu
I cannot say w ith any certainty w hether this structural w eight
ing in favour o f the narrower maternal descent group is directly
associated w ith an ecological emphasis on hunting, w ith its con
com itant m obility and productive individualism , or whether it
is linked to hunting on ly through virilocal m arriage w hich
prim arily consolidates adelphit relations. It m ay w ell be that a
whole com plex o f ecological and historical factors is involved.
B ut it is clear that the tension betw een the w idest village lineage
and its com ponent uterine sibling groups, overlaid though it m ay
be by sundry other sources o f tension, nearly always asserts itself
in situations o f crisis and determines the dom inant m ode o f

N dem bu do n ot now possess structurally significant clans

(nyinyachi o r nyichidi) although they had them in the past, and a
few older people rem em ber their names. M ost o f the peoples
w ho came from the Luanda hom eland o f JMwantiyanvwa possess
a set o f clans in com m on, several o f them named after some
legendary hero (such as Saluseki or Sachingongu). A m ong
Lw ena today individual members o f these clans still have an
obligation to offer hospitality to fello w clansmen, even those









f "*;
( )



* t ;..J




Schism and Continuity in an A frican Society

o f other tribes, w ho travel across Lwena territory. B u t for

m any years N dem bu have had no clan in the sense o f a group
consisting o f a num ber o f m en and wom en, bearing a common
name w hich passes m atrilineally, dispersed w id ely over the
country, forbidden to interm arry, and claim ing vague kinship
w ith one another, such as exist am ong other Central African
peoples like the Bem ba, Kaonde, M azabuka T onga, Lamba,
Lunda o f Luapula, and m any other tribes. N dem bu m ay marry
a prim ary (though rarely) or classificatory sisters daughters
daughter, a m arriage w hich w ould have been forbidden b y clan
exogam y. N dem bu also have a fairly narrow range o f incest
prohibitions. Exam ples o f marriages w hich occurred between
reputedly forbidden categories o f kin and w hich have been
cleansed and made publicly acceptable by ritual, are given below
in case histories. In one case a wom an m arried her matrilineage
brother descended from a com m on great-grandm other and the
m arriage, w hich served as an index o f village fission, speedily
achieved general recognition (see p. 208). T he disappearance
o f clan organization is a further example o f the persistent ten
dency towards the consolidation o f narrower groups at the ex
pense o f w ider groups in all sectors o f the N dem bu social system,
except in the field o f ritual where, as w e shall see in Chapter Ten,
the reverse holds good.
Indeed, even villages whose nuclei claim com m on maternal
descent cannot trace precise genealogical connection w ith one
another, unless one has split o ff very recently from the other.
As w e have seen above, villages said to be inter-related b y mater
nal descent m ay be scattered far and w ide across the N dem bu
region in different Governm ent chiefdoms and in different
Nevertheless, N dem bu possess a lineage group w ider than the
village lineage w ith in w hich genealogical linking is possible.
This group extends beyond the village m atrilineage and includes
all, o r nearly all, o f the livin g matrilineal kin o f the village
m atrilineal core. T h e w om an w ho has married out o f a village,
and her children, are not forgotten by her kin, even i f she lives
m ore than fifty m iles aw ay in her husbands village. T he man
w ho has quarrelled w ith his matrilineal kin and has gone to live
w ith a paternal o r other cognatic relative is sim ilarly remembered.
Sisters children w ho as adults are living w ith their fathers are kept

M atrilin eal Descent


in mind. T h e continual flo w o f visits between m atrilineal kin,

how ever far apart in space, serves to maintain their connection.
These dispersed m atrilineal kin constitute a pool w h ich m ay at
some tim e be draw n upon to supplement and replenish the village
o f the nuclear maternal descent group. A gain , i f a man wishes
to start a village or firm o f his ow n he m ay attem pt to attract
to his new settlem ent such scattered kin. A man I kn ew w ho
was thinking o f starting a village told m e o f 17 m en and boys
and 10 w om en and girls o f his lineage livin g in his village, and o f
19 males and 19 females o f his lineage livin g outside it. H e had
in the period during w hich I knew him paid a series o f visits to
m any o f these m atrilineal kin in other villages to sound them
about com ing to jo in him in his contem plated n ew village.
This exam ple illustrates the point that N dem bu regard matrilineal
kin as the prim ary and essential elem ent in the social com position
o f a village. It is true to say that both a man and a divorced
wom an have a choice in the m atter o f residence in that they m ay
decide to live either w ith their m atrilineal kin or w ith their
father, or, indeed, in other villages under special conditions ;
but they have a life-lon g and autom atic right to reside in the
village o f their close m atrilineal kin, unless they have lost this
right through some sin or crim e they have com m itted such as
sorcery, w itchcraft or m urder. A man has the right to remain
in his fathers village w hile the latter is alive, but after his death
he must go to his m atrilineal village, unless his m other was a
slave, in w hich case he w ould belong to his fathers uterine heir.
Succession to Office
M aternal descent determines succession and inheritance. In
succession to headmanship uterine brother succeeds uterine
brother, and i f the village is a large one w ith a deep matrilineage
it sometimes happens that m atrilineal parallel cousins succeed
one another. W h en the ro w o f brothers and classificatory
brothers has died out, o r i f there are no suitable candidates am ong
them , the right to succeed passes dow n to the next senior genealo
gical generation o f sisters sons. D ue to this tendency to confine
succession w ithin the membership o f the senior genealogical
generation, it n ot infrequently happens that sisters sons becom e
im patient fo r office from w h ich they are barred b y adelphic
succession until * hope deferred m aketh the heart sick \ They


Schism and C ontinuity in an A frican Society

hive o ff from the village w ith their ow n uterine sibling gtoup

and the children o f that group and found new settlements.
Table X V I, based on data I have collected about succession to
village headmanship, supports this statement.

e l a t io n sh ip s

B e t w e e n H e a d m en

a n d t h e ir

S u c cesso rs 1

Relationship o f successor to previous incumbent

No. o f instances





mBS mB 1 z




B Brother
zS Sisters son
zdS Sisters daughters son
S Son
mBS Mothers brothers son
mB Mothers brother
z Sister
d Sisters daughter
1 Excluding the instance o f succession in Chikanga Village discussed on
p. 211.

From T able X V II, p. 205, it w ould appear that sisters* sons and
not brothers o f headmen are the most frequent founders o f new
villages. This tendency w ould be less marked i f there existed,
as in m any other matrilinea! societies, a clear-cut norm eryoining
nepotic succession to office. I f w e exam ine the pattern o f suc
cession to headmanship, Table X V I shows that brothers succeed
m ore often than sisters sons. W hen it is realized that the number
o f m en in the senior genealogical generation is far less than in the
jun ior proxim al generation, w hich contains men o f a low er
average age, it is clear that the chance o f succeeding to office is
m uch greater in the form er. Adelphic succession ensures that a
number o f aged office-holders w ill succeed one another, few o f
w hom w ill live long. Even i f a sisters son does succeed, he
probably succeeds much later in life than he w ould have done i f
he had been his m others brothers first heir. A value is attached
in N dem bu society to the wisdom w hich comes from age, and

M atrilin ea D esce


llS lf

Si 't u



age-seniority is an im portant qualification for headmanship.

The older an heir is when he succeeds, the shorter the tim e he is
likely to hold office. Consequently, there tends to be a high
turnover o f adelpfaic successors, w h o manage to obtain office
before they die. I f heirs manage to succeed young, they remain
in office longer, w hich means that potential successors have to
wait for m any years, and m ay eventually be passed over. For
these reasons, it is often a headmans sisters son. w h o leads a
dissident group from a village. I f he wishes to enjoy a long
period o f leadership, he m ay w ell prefer to give up his chance
o f succeeding to office in a long-established village, despite the
greater prestige o f such an office, than to w ait until he is old.
In this and subsequent chapters I present a num ber o f casehistories (social dramas) w hich throw light on the problem s o f
succession to headmanship. It is indeed distinctly unusual for
sisters sons to succeed w hile a m em ber o f the mothers* brothers*
generation remains alive. T able X I shows that out o f 385
matrilineal hut-ow ners on ly 7 w ere prim ary or classificatory
mothers brodiers o f the headmen, and o f these o n ly one was a
headmans ow n m others brother. This man, Biscuit (III, E 7),
lived in K am aw u V illage near M w inilunga Bom a ; he had spent
most o f his life in Southern Rhodesia, and. had returned as an old
man to live w ith his sister. She was a w id o w livin g in a village
which her sisters son had founded. W hen her brother returned
he had no w ife to cook for him and no other kin except her to
look: after him . H e did not contest the headmanship o f the
village w ith his sisters son, as he was feeble, forgotten b y m any
people, and an inveterate drinker. B y and large, how ever, it
m ay be stated that in practice as w ell as in principle brothers
succeed brothers and cousins succeed cousins before nephews
obtain the right to succeed.
Conflict over Succession within the Matrilineage
Thi&-bDokis_dom inantl^a..sjtu.dy_Qf^
so^d^al.iiiechaiihm s.hroughtm tcr play..toreduc_e_}_e2ccludeQr resolve
that conflict.__Beneath all other conflicts in Ndem bu society is

th e concealed opposition between men and w om en over .descent

and in the econom ic .tMsBasic apposition,
but-possessing their o w n autonom y, sets o f struggles an sew ith in
the sQciaL.strueture..:_ ..conflicts... between, .persons and. .between




T"1 /




Schism and Continuity in an A frican Society

group&jgho invoke different principles. o. residentiaL afifiliation

to support _and justify their ow n specific interests,., political, jural
in_tems_Qfia_cQmmon-J3orm..whicli each party claims the other
has_broken..; and_.CQnfUcts.betwe.en.-pjsrsQns,.wnitetiKy:..asingle
principle o f descent.. and_xesidence,fbr_pQsitions_o-auihority
determined by that principle. Struggles-aro.und_s.uccession to
village headmanship are instances. o.the.last.typeL-ox:onflict-and
it Js w ith these that I wish to. commence die analysis-of w hat I
propose to call * social dramas*. Eormal._analysisL-.of-a social
system enables us-to locate.and.isolate_criticaLpoints_and_areas in
itsL_structure where one. m ight expeot, on a priorgrounds, to
fmcLconilicts between the...occupants of-social-positions-carried
in jthe structure. Im the examination . o.the_,Mdemhmsystem I
have_inthis..xhapter_isQlated_the.matrilineal_descent_group and
shown .how the office ovillage-headman.-is. .vested-in_this_group.
I have looked at different categories-of.matrilineal.kimand_sh.own
howi^struggles fo r succession m ay be expected ,to take__place
between adjacent generations- and. b.etweenspecific-. kinship
positions, notably .between m other' and- .sistecs son.
Xl<:remains to., test out these., hypotheses._in..a-number_o cases,
regarded as typical, and to see whether struggles, do.in_fact_take
place. But the task does not end at this point. I f conflicts occur
w e want to see in what, w ay they are. handled .by the. memhers o f
the society. ImNdemhu..society..cQndu.ct.has..b.een regulated over
what_w.et_can-assume_tO-.have been a,_very .long period o f tim e by
norms, values,-beliefs., and sentiments associated, with.., kinship.
Conflicts o f interests arising_o.ut.of-.the .sociaLstructure are per
petuated by the observance of. thes.e.norms.... Hence, the.conflicts
must_also .followr a. regular course ..dic.tae.d.par.dy_.byjhese..norms,
and take a shape grow n fem iliar to..the.peapleJhraugh_repetition.
W e can expect to. find, in fact,...a mimber,..o social..mechanisms,
o f institutionalized ways o f behaviour, whidbu-havearisen in
response to .an almost endless reduplication o f such-conflicts, and
which have been designed b y group._.experience_tQ-mitigate,
diminish or repair them. C o rn e t .and..the.xesoiution-0.xoiiflict
hafce_ejfects.wm chareohservable in statisticaLand -genealogical
data But~the hints, and indications, afforded by,such.. data must
be .follow ed up by a close study o f social dramas. There we
observe the interlinked and successive events which fo llo w breach,

M atrilin eal D escen t


andLxnake visible, the. sources o conflicts^ _Xhis im tum -leads to

action'which may. restore theearlierset.orelatians,.or.recanstitute
them in a different pattern, o r even ,recognize- an-irreparable
breaking o f relationships betw een particular persons-or~ groups.
These last, nevertheless,! into the w ider pattern ofthfi-N dem bu

The Concept o f the Social Drama

O n a num ber o f occasions during m y field-w ork
aware o f m arked disturbance in the social life o f the
group I happened to be studying at the tim e. X h e jw h o j^ ^ ^ ^ l
m ight be radically cloven .in to. tw o conflicting - fa c t io n s t h e
quarrdling parties might-.compriseL.someb u t-n o t...a ll-o f its
members ; or disputes m ight be-m erely interpersonal iru.character,.... Disturbance in . short....had. a~ variable -_range_o_sodal
inclusiveness. A fter a w h ile ! began tO -detec^a-pattem .in these
eruptions o f c o n flict: I noticed phases_m.theirdevelo.pment
which seemed to follow-oneanother-in_a-m ore-or,-less-regular
sequence__Xhese~em ptions,._whkhIxalLlsacial-dram as./, have
* proeessionaL.ormj>._XJiav.e_-provisionally^.dmded .-the- social
process^which.constitutes_the-..sociaLdram a in.to__four major
phases :
(1) B reach o f regular- norm -govem ed social relations occurs
between persons or groups w ith in thesam e systemo. social
relations. Such a -breach is. signalized-hyi. the_.pnbKcbreach_ or
nonfulfillm ent o f som e crucial, norm xeguIating..the- intercourse
(2) F o llo w in g-b reach -o f . regular .social relations, a phase o f
m otm ting crisis supervenes, durm g w hich, ^
can-be sealed o ff quickly w ithin a lim ited area o f social.interacrion,
there is a tendency.for...the.breach to
a^ -extend.-r'Mntil it
becomes co~extensive with-som e dorninant.cleavageinth.e widest
set -oX felevant social relations to w hich the conflicting parties
belong. T h e phase o f crisis exposes _the_ pattern_ o..cvirrent
factional- stxu gglew ith in .th e relevmt_socdal grou p, be it village,
neighbourhood o r chiefdom ; and beneath, .ittherebecomes
visible the less plastic, m ore durable, but nevertheless ,gradually
fhangin g basic social structure, m ade up o f relations w h ich are
(3) In-order to lim it the spread o f breach certain adjustive and


Schism and Continuity in an A frican Society

regressive mechamsms,.inGrmaLoroxmd,are-speedily.jbrought
into-operation b y leading members of. the rele_vaxit_s.o.ciaLgroup.
These mechanisms vary in characterjw th su c h ..fa ^
and social significance o f the. breach,.the. sociaLinclusiveness .o f
the_crisis, the nature o f the .sooal^gro.up-vdthimwiuch-the.. breach
took place and the degree o f its autonom y .with referencero. wider
systems o f social relations. T h ey m ay range. from .personalad vice
andinform alarbitration, to form al juridical and legal machinery,
and^to resolve certain kinds o f crisis, to o f public
The finaLphase. I have distingmshed. consists either in t
reintegration o f die disturbed social group, o r in- the social recogJ|
nirion.ofirreparableL_breach_betw:een_the.cQn.testing. parties.
in ' short,. the processional form o the-social-dram a-m ay be
form ulated as (1) breach ; (2) crisis ; (3 ). redressive.. action ;
(4) re-integration or. recogniti.on.-of schism.
lL_must. be. recognized,... o f course,-..that- in- differentkinds o f
group, in different.societies,, and under..varying circumstances in
theisame kinds o f group in die.sam e _society, the prQcess m ay not
run- sm oothly or inevitably from , phase to.-phase.... Failure, for
exam ple, in the operation o f xedressive machinery, m ay result
in regression to crisis. In recently-form ed groups, institutional
ized ieg a l or ritual means o f handling, social, disturbance -may be
lacking, and breach m ay be succeeded -im m ediately- b y the
irreversible fission or ficagmentetion-.of-.the group.
In hidem busociety,although.villages- arise and..penshrthe ideal
form-othe village persists. M eanw hile, in order that any village
life ^should be possible, it is necessary, that members~o-a-village
should observe certain com m on. ..values,and-tha.tL_the._norms
governing behaviour between village..memhers^jcnasto.whom
arejmterlinked by ties o f kinship-and affinity, .shouldLbeupheld.
W here custom ary, norms .and..values .are,deeply.entrenched it is
usual to find institutionalized. m achinery. o f redress. Each
instance_o-breach Jn-sociaL-relations-is-made-the-OGcasion o f a
restatem ent-oftheh-xegulative. norms. The.iiature_oxedressive situations
is discussed later.
I have found the social drama a useful descriptive and .analytical
tool W-hen taken in conjunction w ith rn o re orthodox techniques
o f analysis. such as the genealogy,-the-census.and.theh.ut diagram.

M atrilineal D escent


niquesLj:eveals^ regularities in social relations., that _.we...may call
structural. A m ong N d e m b u ,fo r example^..we^fmdLby-. these
group^^-tha..tbere is a
ffighfrequency ochvQ:rce5.thataltemate_genealagicaLgeneEations
tend to build adiacently.and. adjacent ^enealogicaL-eenerations in
opposite sections o f a.:village, .and,, so xm l W e....find..that.jh^re is
a tendency..towards.addtahic..succession.and that sisters* sons tend
th e .m ost frequent founders o^ new ..villages. This leads
us to..suspect tension in the relationship and
sister s-son. T he social drama shows social
tendencies operate in practice ; how , in a givem situation^som e
m ay .support and others oppose one..ano.then^.andJhow_coiiflict
between persons or groups, in. terms~of_a_common_norm,or in
term s-of contradictory xesoLved-in.a-particular set
o f circumstances. In the social drama latent conflicts-of.interest
become^manifest, and kinship..ties,...whose..significance...-is-not
obvious in genealogies, em erge irito_key^ im portancc.
I f w e exam ine ft sequence o f social dramas arising w ithin the
same .social unit, each one glim pse, w ere^ of the
contem porary stage o f m aturation o r decay o f the .socialstructure
o f that unit. Xhope. ta..demonstrate ..thisin_pxesenting. a set o f five
consecutive social, dramas., a single long-stablished village.
T he social drama is a lim ited areaoftranspareiicy on the otherwise
opaque surface o f regular, uneventful so cial life. .. .Through it w e
are enabled to observe the crucial principles.of.the social structure
in their- operation., and_jtheirrelative dom inance at successive
points i n tim e.
O f'th e five social dramas based on material collected at
M ukanza Village,-thelastLthreeLcame.undeJLmyLdirect.ohsetvation.
T he ..a.large.num her.of interviews/and conversations w ith livin g . persons w h o .actively par
ticipated in them .
Social Dram a I illustrates the conflict that m ay arise between
m others brother and sisters son, and between m ale parallel
cousins, w hen on ly a few men rem ain in the senior, office
holding generation in a village, and there are several m iddleaged men ripe for office in the ju n io r generation. O ther kinds
o f conflicts becom e overt w ithin the fram ew ork o f these crucial







Schism and Continuity in an African Society

conflicts ; but the form er w ill not be analysed here, since they
involve other principles o f village organization than m atriliny.
This social, drama is one o f a series, .each o f j^ H ch .contains the
same principal characters, and. each o f w hich, reflects,,.different
aspects o f the same structural conflicts. It m ay be-objected that
such-factors as innate psycho-biological constitution and. personflity_ vananons..determ,the-early
years of,childhood take precedence,over sociological-factors in
shaping_the.. events, .to. be. described. .- B u t it is d e a r - t h a t , the
d i f ^ e n L _ p e r s Q n a h d e s . i n y o . l y e d . . occupy. s.adalp-Qsition&..tha..must
inevitably com e into conflict, and. each, occupant .of.,a.-position
must presentJhis_case in terms ..of.generally-acceptc<Lnorms. A
person can avoid disputesjover successionjonly-b^renouncing the
claim to office vested in .his. position._In-a.&ociety eovem ed by
rules/of.kinship,., he cannot abrogatehis-.position, iiito -which he
is horn and b y virtu e o f w hich h .a.niem ber o f -tha village
com m unity,... Personality may. influence^the.-forni.andJntensity
o f .he dispute, it cannot aboHsh ffie sitm tion. m -wlnGh-conflict
muit; arise. A- .person., w h o -.endeavours., -to. . avoidpressing his
claim to office w hen the position o f headm an-falls vacant is
subject to intense pressure , from his-uterine-kin and-.from his
children to. p u .forward. I f h e .fa ils to d o . so, there occurs a
displacement...o f the locus o f conflict,, not..areso lu tion or bypassing o f conflict. Instead o f leading a group o f kin against the
representatives o f other pressure groups, he becomes the target
o f criticism from members o f his ow n group. At_.some_p.02nt
in jh e _socid proces.s_.arismg^ com pelled to
turn_ad_defend .himself, .whateyer.-his..temperament-OE.cnaracter.
Thc/situation in an N dem bu viflage.xlosely-parallels-that-found
in Greek drama where one witnesses^the helplessness.oftheh.uman
individual before th e . Fates :. but.-in. this- case the-Fates-are the
necessities o f the social process.
Since the struggles to be described below are determ ined b y the
m atrilineai structure o f a long-established and m ature village, it
is necessary to give an outline o f this structure. Appendix I
shows the genealogy o f M ukanza V illage. M ukanza is a
K aw iku V illage, but the conflicts w ithin it are typical o f those
found in all N dem bu villages in similar situations o f crisis.1
1 1 am preparing for publication a paper which deals with similar conflicts
in Ndembu villages o f Lutufa origin.

M atrilineal D escent


The Bewitching o f Kahali Chandenda by his Nephew Sandombu
(compiled from informants)
One day, in 1947, Sandombu (I,Gio) trapped a duiker and divided
its meat between his village kin. His own mothers brother Kahali
Chandenda (LF5) was headman o f the village and should have received
by custom a back leg or the breast o f the animal as his share. Sandombu,
however, gave him part o f a front leg only. Kahali refused to receive
it, saying that Sandombu had shown that he despised his uncle. A
few days afterwards Kahali went to a village in C h ief Sailungas area
about eight miles from Mukanza Village and snared a bush-buck.
He sent his daughter back to Mukanza with the meat. Sandombu
took the meat and proceeded to divide it, retaining the breast, liver,
a front leg and the head, for himself and his wives. N ext day Kahali
returned, and finding Sandombu away, asked Sandombus wife
Malona, a Lunda from Angola, for some food, for he had no wife
o f his own at that time. She was insolently slow in preparing the
food, and in the end he went to his classificatory sister Nyamwaha
(I,F7), who gave him beer as w ell as food. A t night Sandombus
sister and Kahalis own niece, Mangalita (I,G n ), came to him in
private and told him, in anger and shame, how the meat had been
N ext morning Sandombu set o ff early to go to Hiliwood Farm,
twenty-five miles away, where he had seasonal employment as a
capitao in charge o f road maintenance. Meanwhile Kahali spoke
bitterly in the village forum (chota) about Sandombus action, and
the latters wife Malona wept tears at this public shaming o f her
After a week Sandombu returned to Mukanza Village and Malona
reported Kahalis remarks to him. A fierce dispute arose between
unde and nephew, in the course o f which each threatened the other
with *medicine (yitumbu, a euphemism for sorcery, w uloji). Sandombu
ended b y saying, I am going to Sailunga Area. The people o f this
village are worthless. Some people must look out. B y this people
took him to mean that he was going to seek out the services o f a
notorious sorcerer, thought to be Sakasumpa o f Shika Village, to
kill Kahali. It was believed by Mukanza people that Kahali kept his
familiar, an ilomba, or water-snake which possesses the face o f its
owner, in a stream in Sailunga A rea; and that Sandombu had gone
to pay Sakasumpa a fee to shoot the ilomba with his wuta wawufuku
or *night-gun \ a piece o f human tibia carved in the form o f a muzzleloader and primed with graveyard earth and decomposing pieces o f


2 ~ w 2

52 I

3 3




< 9





s *<


L fJj j
K > i.Hw. N Y A M U K O L A *



U I.D. MLABU------ r,Ej.---------------- ---------------------- [,Gi6_________ f * ^

L LU U U...I,J.,.





m a la b w

M AP 5. H ut P l a n o f M u k a n z a V h x a g e m 1947-4
Illustrating Social Dramas I and II
(Line down centre shows the separation o f adjacent
genealogical generations)



Schism and Continuity in an African Society

human bodies. When an ilomba has been killed its owner also dies.
After a few days Kahali fell ill, and. died shortly afterwards. A
rumour came to Mukanza Village that Sandombu had boasted in
Shika that he would kill his uncle by sorcery. N o divination was
made into the death since the people feared prosecution by the Govern
ment for making accusations o f witchcraft. Besides some, including
Mukanza Kabinda (1,FS) and Kanyombu (I.Fp), classificatory matriIineal brothers o f Kahali, said that Sandombu had condemned himself
out o f his own mouth both in Mukanza and in Shika Village.
Sandombu returned to his place o f employment and Mukanza people
said that Sandombu must not succeed Kahali, for he had shown himself
to be a man with a black liver * (muchima bwi), a selfish person and
a sorcerer. The question o f succession was left over for a time.
Sandombu was not expelled from the village because there was no
positive proof o f his guilt, such as might be obtained from the diviners
basket. He had only spoken in heat as had Kahali himself, although
good men did not speak in this way.
Mukanza Kabinda was made headman with the approval o f all,
and confirmed in his office by the Government Sub-Chief Mukangala.
T he account presented is a digest o f inform ation given me,
four years later, by most o f the senior residents o f Mukanza
V illage including Sandombu himself. Sandombu, how ever,
denied that he threatened to bewitch Kahali, although he admitted
that he w ent to Shika V illage after the quarrel. H e said that he
did this in order that peace m ight be restored in the village.
Kahali, he said, was an old man, and old men always die some
tim e or other. This rationalistic attitude was not, how ever,
typical o f Sandombu in other situations, as w e shall see presently.
A ll the other members o f the village concurred in the view that
Sandombu had bewitched Kahali and asserted that he had
publicly avowed that he w ould do so.
It is impossible to get any closer to the actual facts o f the case
fo r the events are no longer susceptible to enquiry and the
account has acquired a m ythical character. W h y has it b y the
consensus o f all save one acquired this character ? T he answer
lies, I think, in the genealogy o f M ukanza V illage itself, and in the
operation o f the principles o f residential affiliation discussed in
Chapter Three. The m yth o f the bew itchingo f. Kahali
in a. short period o tim e becom e the

M atrilin eal D escent


, m ythological, .charter .to.use M alinow skis expressive phrase,

the-sodal. justification, for the. exclusion...of. Sandom btuftom the
succession. W h y had. he been so excluded ? W as it sim ply
because he was regarded as a selfish and a quarrelsome fellow
and a sorcerer to b o o t? I do not think that this explanation
altogether fits the facts. Sandom bu was a m ost generous host
to m e as w ell as to others. H e was a diligent agriculturalist, w ho
grew m uch finger-m illet and gave aw ay beef brew ed from it free
o f charge. A lthough he m ight easily have follow ed the grow ing
practice o f selling beer, he preferred to give it to all and sundry.
B ut there is no doubt that w ithin the lim its o f his little w orld he
was a h igh ly am bitious man, eager fo r headmanship. and the
prestige that even today attaches to., that, .office- in.-the m ore
conservative areas. This is w h y he gave aw ay beer and also
food, for he wanted to put m any people under an obligation to
him, and in tim e to build up a follow ing w ho w ould com e and
live under his leadership in his village. This, am bition was a
little^too obtrusive fo r his relatives to. stomach,., since, he w ould
boast,, even w hen I knew him , that one day he w ould becom e
headman o f M ukanza V illage. This meant .only one thing to
Ndetitbu, that he w as im p a tie n tfo r office... and at
nothing, n ot even sorcery, too-btain bis ends.. B m Ms hospitality,
his obtrusive, am bition,, .and even, his w ild boastingin hi? cups,
suicidal though it was at the tim e. to. hisL.hopes,- ..all stemmed
from his position in the. sod ai structure.
It m ay be noted from A ppendix I that M ukanza is a lon gestablished village in w hich Kahali Chandenda (I,F j) was the
sixth successive headman, and in w hich there w ere tw o w elldefined lineage segments w hose com m on ancestress was in the
third ascendant generation from the present headman. Succestitinhad^beemadelphic-between-the. secondhand, third headman
and im:the_next-.genealogical.generation-thre.e,matrilineal parallel
cousins.had follow ed, one the.headmanship*-,In other
w ords, p_ositionof.theviEage.m usfJ^vesuccessfidly
undergone the_qualitative change from bilateraLextended-fem iiy
(i.e. .uterine siblings and their children) to matrilineage. and had
developed a .precise, and. extensive, system^of-complementary and
opposed genealogical generations.. B u t w ith the onw ard flo w
o f strutturai tim e and th e ..passage. ..of-successive-genealogical
generations the danger o f fission, becam e..progressively greater.

io o

Schism and Continuity in an African Society

T he lineages o f w hich nyachjntang a 1 (I,D i) and malabu

(I,D a) w ere the respective ancestresses w ere structurally ready
to divide from one another. A few years previously (1928)
another m atrilineage had split o ff after actual fighting in the
village between the senior m en o f M ukanza ; and it was touch
and go, so to speak, w hether the lineage (ivumu) o f malabu ,
the senior man o f w hich was Sakazao (I,H9), w ould not follow
suit. Sakazao, although in the second descendant generation
from Kahali Chandenda (I,F5) and M ukanza Kabinda (I,F8),
and in the first descendant generation from Sandombu (I,G io),
was older than the latter, and not m any years younger than
M ukanza Kabinda. M ukanza Kabinda was in fact the person
w ho held the tw o lineages together, since his w ife, b y w hom he
had m any children, was o w n sister o f Sakazao, head o f the
malabu lineage. H e was thus a matrilineal relative o f his w ife
by grandparent-grandchild m arriage. His w ife N yam ukola
(I,H io) was the link that united the tw o halves o f the village
and stood in an intercalary relationship between them. H er
children are children o f the tw o lineages : outside M ukanza
V illage they had now here to live. A t the same tim e, Mukanza
Kabinda (I,F8), his sister N yam w aha (I,F7), and his brother
Kanyom bu (I.Fp) w ere the last surviving members o f the senior
genealogical generation after the death o f Kahali. N o male o f
the m alabu lineage, to w hom the headmanship m ight pass,
remained alive in that generation. O ld Kajata (I,Fio) was senile,
and at that tim e seemed lik ely to die in K afw eku, his w ifes
village. M ost o f the senior members o f malabu lineage w ere
in the second descendant genealogical generation to M ukanza
Kabinda and his siblings, and w ere m oreover allied to him by
affinity and sibling-in-law ship. O ne man and three wom en
o f malabu lineage and o f G generation (to w hich Sandombu and
Kasonda also belonged), w ere alive at that time. T h e man,
Chinem a (I,G ai), was w orking on the Copperbelt as a road
labourer, and it was then considered that he w ould never return
to M ukanza V illage. H e had been headman o f Kahali Chandendas village in Ikelenge Area fo r a number o f years after
Kahali had returned to M ukanza. B u t since his village there had
been com posite in character, containing m any members o f
1 The

names o f lineages are in capitals.


another lineage and few o f his ow n m atrilineal kin, he had been

^involved in m any disputes w ith the other lineage. H e returned
v to M ukanza Village about 1945, perhaps w ith the hope o f suc
ceeding Kahali Chandenda. B u t he had quarrelled w ith
N yam uw anga, the notorious w itch o f Social D ram a FV, and it
was said that his fear o f her led to his flight to the Copperbelt.
N yam uw anga herself was barely tolerated in M ukanza Village,
and, as w e shall see, was subsequently expelled from it. Her
sister Shim ili (I,G i8) was at the tim e o f Social D ram a I living
'hi, virilocally w ith her husband in Chim bila V illage in Ikdenge
Area. O f her other sisters, Kalusa (I,G i9 ) was livin g v in locally at Shika V illage, w hile N yam alita (I,G zo) alternated
between her virilocal village Chim bila and K alene Hospital
where she w orked fo r lo n g periods as a ward-m aid. N one o f
the livin g members o f G generation in maiabxj lineage entered
the reckoning at the tim e o f Social D ram a I as candidates for
office or supporters o f a candidate. B u t Sakazao and Ins uterine
sisters and their children could be regarded as m ore or less
permanent residents in the village, and Sakazaos sisters w ere
married to M ukanza and his brother K anyom bu. M ukanza
K abm da.byanum beroreckonihgsQ ccupied-a_pivQ taI-position
with-Jtegarri-_to. JtbcLjtilUgitJK^..a..whole.^.,most_Q.its^stmctural
relations converged on his, position ,in the.kinsbip,system.
B efore considering die position o f Sandom bu in this situation,
another potential source o f danger to the continued existence o f
the m ajor m atrilineage should be m entioned. It has been noted
that fission m ay, and frequently does, occur as the result o f the
secession o f a fam ily o f mature uterine siblings and their children.
The possibility o f such fission existed in M ukanza V illage at die
tim e o f KahaH Chandendas death. T w o w om en, C haw utongi
(I,G i2) and M anyosa (I,G i3 ), both sisters daughters o f M ukanza,
had large families, one w ith six and die other w ith five children.
B oth these w om en w ere livin g avunculocally w ith w eak com
pliant husbands. T he husband o f one, nam ed N deleki, in fact
provides an illustration o f the fate o f a man w ho has n ot pressed
Ids claims to the headmanship o f a village. H e was a quiet and
selfecontained person, a devoted husband and father, but w ithout
that stubborn individualism , barely concealed b y the veneer o f
politeness and sociability, w h ich is typical o f N dem bu o f both
sexes. W hen the headman o f his m atrilineal village had died,


Matrilineal Descent



( )



4 )

Schism and C ontinuity in an A frican Society

he had n ot put him self forw ard as the successor. As a result

people tended to despise him or at best to ch ivy him in a goodhum oured w ay in his ow n village. H e phrased his social re
jection there in terms o f w itchcraft accusations against his kin
and w ent to live at his w ifes village. His w ife Chaw utongi
(I,G i2 ), on the other hand, was an aggressive and ambitious
wom an, high ly conscious o f her im portance as a m other o f many
children. H er m others sisters son, her ow n parallel cousin,
Kasonda (I,G i5) b y name, was an im portant man in the litde
village com m unity, shrewd to observe the social norm s o f
N dem bu life but also am bitious and seeking b y devious means
either to succeed to the headmanship, or better, to start a village
o f his ow n. The other m other o f m any children was his ow n
sister M anyosa (I,G i3), also a dom inating personality. A ll three,
Kasonda, M anyosa and Chaw utongi, w ere fu ll sisters* children
o f M ukanza Kabinda (I,F8), w ho in form er times w ould have
had the right, as their m others brother, to sell them into slavery.
He still exerted considerable authority over them . B u t one
factor in particular at that tim e m ilitated against Kasondas claim
to succeed his relative youth. A lthough he had lived fo r more
than ten years in urban areas, five o f them in B ulaw ayo, and
although he had acquired a little education in M ission schools,
traditional N dem bu notions rated him too young, at the age o f
thirty-five, to becom e a headman. In frank conversations I have
had w ith Kasonda about the question o f succession, he has told
me that he saw his best hope in exerting his influence over his
pow erful sisters to support the claim to succession o f his ow n
uterine uncle, M ukanza Kabinda. T h e latter was b y that time
an old man o f about 66 and could n ot be expected to live verylong. I f he died in a short tim e Kasonda w ould support
also in his sixties, fo r the headmanship. I f
M ukanza Kabindas life lasted a further ten years Kasonda w ould
press his ow n claims after his uncles death. Kasonda, although
he was careful to maintain the outw ard show o f friendliness
towards Sandombu, was privately jealous o f him , hated him as
a bar to his ow n advancement, and feared him on account o f his
widespread notoriety as a sorcerer and friend o f sorcerers. I f
Sandombu w ere to succeed at the age o f about fifty, it m ight
be m any years before Kasonda w ould get the chance to follow
him into office, and therefore Kasonda was determined to keep

M atrilineal D escent


Sandombu out o f the running at all costs, b y the secret and devious
ways o f whispered slander. H e, m ore than anyone else, was
responsible fo r the story that Sandom bu had bew itched Kahali
Chandenda, and he never allow ed the episode o f the quarrel to
becom e forgotten. For various reasons it suited him w ell that
M ukanza Kabinda, and not Sandombu, should succeed. M ukanza
Kabinda had a retiring and unaggressive character and at that
period at any rate did not possess Kasondas skill in the advocacy
and judgm ent o f cases. Judicial skill (wukaku) is rated high among
the accomplishments o f N dem bu men, and Kasonda hoped that
through his ow n gifts o f eloquence and deliberation he would
becom e the de facto headman o f M ukanza V illage. Kasonda
could also speak w ith facility and assurance to such Europeans
as D istrict Officers and missionaries w h o came sporadically to
trouble the people * (nakukabisha antu), as N dem bu put it. His
know ledge o f English, and his fam iliarity w ith documents,
w ould prove useful in this respect, and provide a sharp contrast
w ith M ukanza s illiteracy and ineptitude w ith mapaperi ( papers ,
a general term fo r tax receipts, gun and d og licences, bicycle
tax, permits to buy gunpow der, etc.). Kasonda saw himself^
in short, as the G rey Eminence o f M ukanza V illage, recognized
b y N dem bu as the effective village head, but delightfully free
from official responsibility for the delicts o f its members.
T h e links w hich bound M ukanza Kabinda, his brother-in-law
Sakazao (I,Hp), and his ow n sisters son Kasonda closely to
gether, w ere extended to include the close lineal kin and de
pendants o f these im portant men. Three linked pressure-groups
w ere thus form ed, all o f w hich supported M ukanza Kabindas
claim to succeed K ahali Chandenda (I,F5). Sandom bu (i,G io ),
although anxious to succeed, was the odd man out. W h y then
was he unable to obtain support from the other members o f the
village ? T he principal reason fo r his exclusion from considera
tion la y in his structural position. K ahali Chandenda was Ms
ow n m others brother and i f Sandom bu had succeeded, members
o f one lineage segm ent o f different generations w ould have
succeeded to headmanship, w hile a senior o f the other lineage
segm ent was available. A principle o f great im portance in
N dem bu political structure w ould thus have been broken,
nam ely that no single lineage segm ent should obtain a m onopoly
over a long-established office, for such positions are few , highly


Schism and Continuity in an A frican Society

valued, and jealously contested. For instance, the chieftainship

o f Kanongesha is vested in a dispersed maternal descent group
w hich is scattered over the N dem bu region to form the nuclei
o f diffrent villages. There is a rule o f succession w hich states
that no tw o successive incumbents should com e from the same
village. This m ode o f succession ensures that the honours should
g o round and not becom e the m onopoly o f a single village
m atrilineage belonging to the chiefly maternal descent group.
In a sim ilar w ay in long-established com m oner villages, classificatory adelphic succession, often cutting across affiliation by
lineage segment, is a means o f interlinking discrete uterine
sibling groups. I f the leading men in each consider that they
stand a reasonable chance o f succeeding to office they m ay w ell
hesitate before seceding from the village. T he basic member
ship o f long-established Ndem bu villages consists o f a rather loose
association o f free and independent elders w h o are not really
constrained b y econom ic exigency or political directive from
above to remain together. Nevertheless, i f the village is relatively
ancient, pow erful ties o f historical pride m ay buttress those o f
kinship to retain the elders jo in t allegiance. Perhaps their unity
tends to be further stiffened by the possibility o f succeeding to
the headmanship. T h e prestige attached to livin g in a longestablished village puts them under a certain am ount o f moral
pressure not to secede from it and * k ill the village as Ndem bu
put it. Since the headman, am ong the w atchfully egalitarian
Ndem bu, is hardly m ore than a primus inter pares, he cannot
coerce village members to remain in the village. T he unity and
cohesion o f villages depends on the fraternal association o f
generation mates, not on the dominance o f a single lineage
A nother reason fo r the prem editated exclusion o f Sandombu
lay in the history o f M ukanza V illage. In m ost Central African
tribes the nam e o f a village is a hereditary title through w hich
flow s a succession o f individual incumbents, w ho take the name
o f the founder o f the village. In the case o f senior headmanships
this is also the case am ong Ndem bu. Here the names o f the
founders becom e inherited tides. B u t com m oner villages are
constantly com ing into being as the result o f fission, and also
villages frequently becom e extinct ow in g to the secession o f
the m ajority o f their inhabitants. There is often, am ong

M atrilin eal D escent


N dem bu, a period in the developm ent o f a village during w hich

die village is generally know n b y the personal name o f each
successive headman* B ut w hen a village has becom e unquestion
ably established as a persistent social unit, the personal name o f an
outstanding headman tends to petrify into a title w hich is in
herited b y his successors.
Thus in the village n ow know n as M ukanza, the founder was
his first successor was N g unji (I,E z), w hile the
next three headm en (IE3 ; 1,^4 ; I, F5) bore the tide o f Kahali*
T he establishment o f this tide m arked as it w ere the social consoli
dation and m aturity o f the village, and it is lik ely that the village
w ould have continued to bear this tide had not the British Govern
m ent intruded upon the continuity o f N dem bu life. W hen
taxation was introduced b y the B om a in 1913 d ie village o f
Kahali Saluyi (I,F4), as it was then know n, fled to w hat is n ow
ikelenge A rea to seek the protection o f the first C .M .M X .
m issionary, D r. W alter Fisher. There a great hunter from the
village, M ukanza Kandulu (IF6), killed a roan antelope and
quarrelled over its division w ith Senior Headm an Ikelenge, a
descendant o f one o f the tw elve headmen w h o had accompanied
the first Kanongesha from M w antiyanvw a. Ikelenge demanded
a back leg o f M ukanzas kill* as his right, since he was a senior
headman whose ancestor had been appointed b y Kanongesha,
ch ie f o f the N dem bu. M ukanza refused to give him the meat
and thus refused to recognize Ikelenges authority in the area.
B u t Ikelenge had m ore follow ers in his ow n area than M ukanza
could m uster ; and M ukanza, having heard that the British
South A frica C om panys adm inistration was less harsh than had
at first been feared, suggested to Kahali Chandenda (I,F5),
uterine u n d e o f Sandombu, w h o had ju st succeeded Kahali
Saluyi (I,F4), that the village group should return to their old site
fiv e m iles from the Bom a. Kahali refused to g o and the village
divided, some returning w ith M ukanza and some rem aining w ith
K ahali. In 1919 M ukanza K andulu died, and some years later
K ahali Chandenda, w h o had him self quarrelled w ith Ikelenges
people, cam e back to the village and resumed his position as
headman. B u t in his absence, the village had generally becom e
kn ow n as M ukanza, and although K ahali refused to take the
title o f M ukanza, the village continued to be called b y that name
both b y its inhabitants and b y other N dem bu. It was argued, I

f '- 'H

V - /


l .=

i ... -



\ .



Schism and Continuity in an A frican Society

understand b y the siblings and sisters children o f M ukanza

Kandulu, that the name o f M ukanza was n ow established as the
title o f the village. In fact, Kabinda (I,F8), w ho eventually suc
ceeded Kahali Chandenda as headman and was still headman w hen
I was in the field, w ent so far as to succeed in 192S to the name
M ukanza. B y that act h e secured fo r him self, in N dem bu belief,
the tutelage o f the spirit o f M ukanza Kandulu (I,F6), in anticipa
tion o f his later accession to office. In a similar w ay, Sandombu,
w hile I was residing at M ukanza, form ally succeeded to the name
o f Kahali. Thus, as early as the twenties o f this century, a
cleavage had already developed between the follow ers o f
M ukanza Kandulu and those w ho remained in Ikelenge Area
w ith Kahali Chandenda (I,F5). Am ong the form er w ere
Sakazao (I,H9) and his sister N yam ukola (I,H io), w hom M ukanza
Kabinda m arried, thus linking the matrilineages o f nyachln tang a (I,D i ) and m alabu (I,Da). From this union m any
children w ere bom both o f whose parents belonged to the m ajor
matrilineage o f M ukanza Village, and w ho could be counted
upon b y their father to support him when he claim ed the headmanship. Sandombu (I,G ro), on the other hand, remained
w ith his uterine uncle K ahali Chandenda, and stayed for m any
years in Ikelenge Area, w here indeed he secured paid em ploym ent
as a road capitao at H illw ood Farm. H e visited M ukanza
V illage often, how ever, and when he saw that m ost o f Kahalis
original follow in g w ere trickling back to M ukanza, decided to
build his ow n huts there, and eventually to try to succeed there.
B ut by the tim e he had settled there a netw ork o f ties already
united the members o f the tw o lineages w ho had follow ed
M ukanza Kandulu (I,F6). N o t only affinity, but also the alliance
o f alternate generations (c p. 80 above), held together the
senior members o f these tw o groups. In the course o f tim e
genealogical generation and biological age had ceased to coincide
as between members o f the tw o lineages. This gave rise to a
situation in w hich several members o f the grandchild generation
o f malabu lineage w ere chronological contemporaries o f several
members o f the grandparent generation in nyachintang a
lineage, and w ere in fact older than most o f the senior members
o f the intervening generation in the latter. D ue to the friendship
that existed between these alternate generations it was quite
possible that Sandombu, and indeed Kasonda (I,G i5), w ould

S u c c e s s io n

to a


Headman Mukanza pours out white maize beer at the base of a newly-planted
muyombu sapling (see Social Drama II) to the spirit to whose name Yana (see
Chapter Ten) is succeeding. Yana is on the left. She will later wear the
white head-cloth draped over the muyombu. She will also wear a white bangle.
Manyosa, to the right, is anointed with powdered white clay (mpemba), and
holds some in her hand with which she will later anoint the whole ritual assembly.
White stands for health, good luck, strength, fertility, good will between persons,
atonement with the ancestors. The muyombu tree has white wood and exudes
white gum.

M atrilin eal D escent


be passed over in the succession, and that Sakazao (IjHp), the

* grandchild * and brother-in-law o f M ukanza K abinda (I,F8),
w ould fo llo w the latter i f he succeeded to K ahali Chandenda
(IJF5). This situation gave to Sandombus am bitions an edge
o f desperation that caused bis desire fo r office to becom e a little
too obtrusive fo r N dem bu tastes : he made statements publicly
w hich proclaim ed this desire and brought upon him the suspicion
that he w ould n o t hesitate to use sorcery to realize it.
A n im portant factor operating to defeat Sandom bus ambitions
in the period im m ediately p rio r to Social D ram a I, w as the absence
from M ukanza V illage o f close maternal kin descended from his
m others m other N yakapakata (I,E i). His sister M angalita
(I,G i i ) appeared to be sterile and his m others sisters daughters
daughter Bibiana (I,H4) was residing o n the C opperbelt w ith
her husband. H e could thus invoke no local support from
close m aternal k in to counter the pow erful local pressure-group
mustered b y M ukanza Kabinda and Sakazao (IH9).
Other.factors .arism gfrom biological and psychological accident
conspired to..hardm suspicion agaim t_Sm aom biiJhto_what for
N d en ffiu .w as.vk tu d ..^
Param ount amQag..ffig^.was the
fact that: Sandombu- had .no. children, . save fo r o n e. daughter
w hom scandal hinted ..was n o t his . child. Indeed, it is certain
that w hen he had gone to an urban area to w o rk in 1927 he
contracted gonorrhoea w hich seems to have rendered him
perm anently sterile. N o w fo r N dem bu .a. sterile m an is often
regarded as a sorcerer. It was said that his semen (matekela) was
not * w h ite {atoka), a colour sym bolically associated in m any
ritual and cerem onial contexts w ith purity, health, strength, piety
towards the ancestors, and good w ill towards ones follows.
Instead, his semen had becom e * red (achinana), a colour often
linked w ith w itchcraft, aggressiveness, and evil pow er obtained
through a w ilfu l breach o f the social norm s. Sandombu, ow ing
to his infection, had yello w semen (called * red * b y N dem bu) ;
and this was com m on know ledge am ong his follow -villagers,
perhaps because he had married successively tw o o f Mukanza
Kabindas daughters, the first o f w hom died. H e had often
beaten both o f them , and caused them to seek the protection o f
their parents, w hom th ey had doubtless inform ed, in the circum
locutions (iku-didyika) em ployed betw een members o f adjacent
generations, o f their husbands raisfortune.


Schism and Continuity in an A frican Society

.Sterility is also a misfortune, for an. ambitious-man-in -other

respects.. W hen-a man. wishes to..succeeci .tO -office. oE-to found
a village o fh iso w n , he looks fo r foe baching ohis-o-VNm childxeri
in,as w ell astohis.uterine.kin. A mans major
unit o f political support is the circle o f his closest kin, called by
Ndem bu ntanga, or b y the plural form antang a. In addition to
his ow n and his sisters children, these kin include his brothers
and their children. Such a group contains the nucleus o f a new
generation, the ju n ior adjacent genealogical generation over
which he andhis siblings exercise authority and control. Sandombu
was doubly unfortunate_in...that his_oxdy .full~sister_JMangalita
(I^Qi i ) . was barren, so that. he., had .neither...children.jrior own
sisters children to. support him.
fam ily, the fact, that Sandombu . was..a...di3igent_and .tireless
gardener, set the seal on the v iew that, he was a.sorcerer^a i/cyi
or mukwa kulowa....It.was .said that he. hacLa familiar, a wooden
figurine ,o f the type called katotajiy vAnch was activatedby. the
blood o f his previous victims, .and whicb- w orked-beside him ,
invisible to others, in his cassava gardens, heaping up countless
mounds. Sandombu from the fruits .o f his labours, washable
to...dispense to many people, most o f them strangers, a-lavish
hospitality, and to earn golden opinions fronx them. Thus he
aimed to lay the basis for the mclusion .in his futur& village o f
strangers w ho m ight be attracted thither b y diis- reputarion for
open-handedness. This again constituted a threat, tojhe_closelyknit and partly endogamous traditional com munity o f Mukanza
In summary, Sandombu was from m any points.of view an out
sider, an atypical, m arginal ,man.-in_Mukanza.. Village. H e
belonged to the Kahali faction w ho had remained in Ikelehge
Area when Mukanza Kandulu (I,F6) and his close kin and the
sibling group o f Sakazao (I,Ht>) had returned to the traditional
area where Kabonzu founded die village. H e was outside die
mesh o f interlocking reladonships that united die m ajority o f
members o f n y a c h in ta n g a ana malabxj lineages. He~was a
m em ber^ofthe intervening genealogical.generation,-which at the
time., was out o f pow er in die village.. He.was, sterile ,,.and in
dustrious, both indications .o f sorcery. lhe_succeede<
a single m inor lineage w ould gain control o f an office that should
have been sh ared dirtu m by all. the.lineages._of.the..village.

M atrilineal D escent


Finally, he w a s.a ch a n n elth ro u g h jw h ich .s

flo w
intO-.the tightly, orgaiuzedLand jflghly^xonservative. com niiinity
o f a jy p ic a l K aw iku. village. W h o better than, he could, serve as
a scapegoat fo r m isfortiines befalling the village ? W h o h etter
than h e'co u ld serve to umte- the-potentially-flssile~com ponents
o f -thfe remaining-jmembersMp,.. in_.teans_QL-Opp.osiJdotLJteiL. his
am bitionsandto. hisJmmoderate_way-of life ? H e tried to attach
him self to the chineng a sub-lineage o f n yach in tan g a lineage
and to malabu lineage b y m arrying daughters o f M ukanza, one
b y N yatungeji (II,D 3), and one b y N yam ukola. B y this last
m arriage hem ade him self the tw o lineages *son-in-law * (Ndembu
tnuku, properly speaking an affine belonging to an adjacent
generation to ego ) . B u t this. very reladonship-is one-involving
avoidance .o f fem alem -daw s, _and~unwilling-subm ission- to male
in-law s, o f the senior adjacent generation^dtism otLa.xelationship
o f fnendship a n d fm u n a rjty. ,-.^et ..stifl^jhis, odd m an, out was
audacious enough to aspire to be the mandat the helm .
W h y was it, then, that (he m arriage o f M ukanza to N yam ukola
proved to be a structurally effective alliance between lineage
segments, whereas that between Sandom bu and Zuliyana was
fraught w ith conflict ? Part o f the answer lies undoubtedly in
the fact that the form er m arriage w as fruitful and the latter barren.
It was a source o f chagrin to Sandom bu that Zuliyana had no
children and he was w on t to blam e her infertility and not his
sterility for this m isfortune. In fact, he paid several ritual
specialists to treat her w ith medicines to render her fertile. In
Chapter T en I discuss one such ritual in w h ich Zuliyana under
w ent treatm ent. B u t other factors w ere, in m y opinion, no
less im portant in rendering the alliance structurally ineffective.
T he marriage between M ukanza and N yam ukola did not
m arkedly increase the social distance betw een the spouses and
members o f the intervening genealogical generation, since the
parents o f M ukanza and o f N yam ukola w ere dead, and neither
had close in-laws in the senior adjacent genealogical generation.
B ut Sandombu, and fo r that m atter Kasonda, ranked as classifi
es to ry affines o f an adjacent generation to both spouses, and this
relationship b y m arriage, added to the tense relationship between
adjacent generations, did nothing to bring Sandombu closer to
M ukanza and his w ife. O n the other hand, Mukanza*s marriage
to N yam ukola strengthened the antecedent tie o f alliance between


Schism and Continuity in an A frican Society

alternate genealogical generations that connected M ukanza and

Sakazao, head o f malabu lineage. T he children o f their union
w ere Sakazaos sisters children and ciassificatory * children * o f
N yam ukolas sister N yatioli (I,H n ).
B iological and structural factors, interacting in the process o f
social life, lim ited the efiectiveness o f Sandom bus cross-cousin
marriage and increased that o f M ukanzas grandparent-grandchild
marriage. T he mifniitfulness o f the form er led to frequent
quarrels between the spouses and Zuliyana was usually able to
run to her parents fo r solace and protection against her husband.
The fhritfulness o f the latter gave M ukanza and d ie senior
members o f malabu lineage a jo in t interest in the children bom
in it.
T he relative duration of. the respective m arriag^ at_the_time o f
Social ..Dram a .I m ust taken into _accoxmt. M ukanzas
m arriage had. .stood the exacting test. o f .tim e....Nearly...all the
children o f M ukanza and.Nyam ukola w ere fu lh grow n adults, and
some had children o f their ow n. T he m arriage Jwas.everywhere
held up as a shining exam ple o f m atrim onial happiness {umluwt) .
O n th e other h a n d ,...Sandombu.. had. o n ly recen tly.. married
Zuliyana, a pretty and rather_dighty...girl, and., the first years
o f m arriage am ong N dem bu. always. ..possess., an.-experimental
character, frequently term inating rin divorce. E ven in 1953,
Sandombu was still almost pathologically jealous o f Zuliyana.
O nce he showed m e a spear that he kept in a com er o f his hut.
He said that he w ould stab anyone w ith it w hom he found
m aking advances to Zuliyana. H e did not trust her for a
m om ent out o f his sight. B ut he was determined to keep his
marriage in being, fo r although it was unsatisfactory in m any
respects, it still gave him a footing in the village com m unity.
I have. set .out. the sociological and historicaLbackgrouncLto this
first-social drama, a t.some, length, and deJineated-the-personalities
o f its m ain protagonists,, because ,w ithout- this. preMminary- w ork
it w ou ld be impossible., to make., a...satisfactory-analysis- o f the
subsequent social dramas, each o f. which--represents -a- further
stage, in the conflict over succession, im plicit in the matxilineal
structure o f M ukanza Village.
Itris evident from the events., described, at. theL.b.egfrming o f
SoriaLJDrama I, that tension had existed in. .the., relationship

M atrilineal D escent


between K a h a h C h a n ^
for sQmejdm e hefo
the,.episodes in vo lvin g meat disiribution had.,,precipitated a
situation o f open hostility. M y inform ants told m e that Sandoxnbu had neglected fo r several years before this to supply his
w idow er uncle w ith cassava meal from his ow n gardens. N yam waha (I,F7), M anyosa (I,G i3) and Kasonda (I,G i 5) had com bined
to feed th old man. Kahali Chandenda, even b y Ndem bu
standards, was a very poor man. Part o f his leg had been
; amputated, and w hen he travelled he had to enlist the help o f a
vounger man to push him on a bicycle. O n the credit side o f the
i- balance, he had. been a famous chiyanga, a hunter w ith a gun, and
was in addition a great 4law -m an (ihaku), ju d gin g and advo
cating cases w ith forensic skill and know ledge o f custom . B ut
he w as old-fashioned and could n ot cope w ith the m odem duties
o f a headman. H e was also feared as a sorcerer. Hunters are
thought to possess exceptionally pow erful fam iliars ; their
strength and skill in killing animals is acquired initially from
killing their ju n io r relatives b y m edicine. Thus KahalTs accom
plishments belonged to a rapidly passing social order, and
physically he was a burden to the village and a reproach to
Sandombu. H is reputation as a man w ith bad medicine was all
dnat restrained the hostility, felt b y several persons against him,
from breaking out.
Sandom bu, w ho had lived and w orked in Ikelenge Area most
o f his life, had also gained a certain notoriety as a sorcerer there.
I have been told b y members o f a village in Ikelenge Area in
w hich he lived fo r a lon g tim e, that w hen he left it finally, a large
hole w as found under the door o f his abandoned hut, in w hich it
was supposed he kept his ilomba, a serpent fam iliar w ith a human
face. I f there really was a hole, it was probably used fo r urination
at night. I have seen similar unhygienic devices in huts else
where in M w ind unga, and N dem bu consider that people w ho
have these holes are disgusting. But, w hatever the * circum
stantial evidence \ it is certain that Sandom bu had a bad reputation
as a practitioner o f the black art am ong the local inhabitants o f
Ikelenge Area. I m yself have often discussed sorcery and
w itchcraft w ith him , and although he always spoke indignantly
against it, he had undeniably a w ide know ledge o f its putative
techniques and materia medica. LLjs_passibIe, that. Sandombu,
fm ding . him self dubbed a sorcerer, on .account_o his .sterility,










Schism and Continuity in an A frican Society

.deliberately, fostered the belief,, by innuendo. when.sober,.vand by

the, violence ofhisbeha'viour when in lus cups^uLoxderL_to inspire
feat, in people and so to get his ow n w ay in a number o f situations.
B ut-it was even m ore in the interests o f his rivals. bt.,thc,village
succession, - particularly MuJkanza. Kabinda and Kasonda, to
harden suspicion o f ms sorcery into popular-certainty_and to
establish it as a local-dogm a.
N dem bu headmen pray to their ancestors before boys" .and
girls* initiation rituals : * G ive us in this village o n ly the meat
o f animals, not the meat o f m en/ This means, * Exclude sorcerers
and witches from our m idst/ i f the headman him self is a sorcerer,
in N dem bu belief, the village is in terrible danger, fo r it is thought
that headmanship adds additional pow er both for good and evil
to w hat a man already has in the w ay o f m ystical acquirements.
I f Sandombu w ere branded w ith this evil reputation, his chance
o f succeeding to office w ould be eliminated, except under special
circumstances w hich I w ill mention in the analysis o f a subsequent
B y -g iv in g Kahali an inferior .share o f his ow n .meat and by
eating a m ajor share o f Kahali*s meat, Sandombu made it clear
that.Jb.e no longer respected Kahali. as an uncle and a-headman,
and...would be glad to see the last o f him . .T h e fact-that in the
subsequent quarrel M ukanza people alleged...tha.t-Kahalk.too. used
threats o f sorcery against Sandombu^indicates that they- con
sidered that uncle and nephew were-eq.ualiy .evil.-..From-this the
people argued that their lineage..should no lon ger enter.into the
It m ay be asked w h y, i f Sandombu had the reputation o f being
a sorcerer, people thought it was necessary for him to go to the
black-hearted old Sakasumpa to enlist his aid against Kahali.
The answer lies in the belief that w hen a man wishes to k ill an
office-holder, such as a headman or a chief, he does n ot w ish to
take the fu ll responsibility him self fo r the sorcery. It m ay be
also that the diviner, to w hom the kin o f the dead headman may
repair, w ill be deceived b y this device. There are innumerable
Ndem bu tales, told round the chota fire or in the privacy o f huts
and kitchens, o f claimants for office w ho have sought fo r sorcerers
to slay b y m edicine the present incumbents. It is told h ow pay
ment o f guns, cloth and m oney, and form erly slaves, w ere made
for these services. It is also told how a great diviner w ill detect

M atrilin eal D escent


this device and h o w he w ill lay the blam e for the sorcery on the
claimant and not on the sorcerer w h o actually * shoots the
medicine . He was only doing a jo b for w hich he had been
hired, the people say, a w icked jo b it is true. B u t the real guilt
lay * in the liver *, as the seat o f the w ill and source o f pow er, o f
the relative w ho wanted to succeed. In such divinations the
name o f the hired specialist is never m entioned, perhaps because
o f the m ode o f questioning em ployed b y diviners w hich makes
the initial assumption that kin bew itch kin. O n ly i f they receive
negative answers, on consulting their oracles, to questions em
bracing the w hole gam ut o f kinship, w ill they seek fo r a sorcerer
or w itch outside the victim s kin-group.
This point is relevant to the analysis for later w e shall find
Sandombu accused b y the headman o f a neighbouring village
as the sorcerer called in b y that headmans relative to bew itch
In jfcjcta rra tive .o f. thebcw itching,.Q K ahali,.coloured,:though
it is by>the inclusionjQf.m ythical. m aterial, .we.detect...the .presence
o f those elements considered- ty p ic a lo f aULSQciaLdramas._Th.ere
had keen ,a breach in the custom ary xegularit^ social relations
between Kahali .and_.Sandombu,.and betweerL the-m em bers o f
M ukanza village as a w hole. This breach- took on. decirive~and
dramatic form as th e . breach-of_a cruciaLnorm ..-governing., the
relations o f m others brother and sisters, som in -bidembu society.
A sister^.on.shauid^give.hia-Uterine_ unclea spLe_cific_j.oiiitof meat
w heh-heJias m ade.aJrill.-... This, is o n e o f the. m ajor. obligations
inherent in the relationship. I f i t is n ot .carried, out,, w ithout the
excuse ..of exception al. circumstances, this is tantamount to a
refusal on the part , o f the nephew ..to_ accept.. the_continued
authority .of..his. .uncle... B ut it. goes, further, than that. ...It is a
challenge to one o f the., m ost. deeply , entrenched.principles o f
N dem bu social organization, that b y w h ich the .senior adjacent
genealogical generation is given authority, over..the.._.junior
generation, and the latter m ust-respect and . obey the .former.
W here generation authority, is linked w ith m atriliny,. as. in the
uncle-nephew relationship,.the.authority o f the .senior.generation
is exceptionally strong. It is n ot sim ply a breach in the relation
ship-betw een tw o persons, but a m atter involving, the w hole
group ; for it is a challenge to the authority o f the. senior genera
tion as-a w hole and ultim ately to the system itself.


Schism and Continuity in an A frican Society

H ie challenge is also one w hich not only jeopardizes the-general

authority structure, but w hich involves succession-to-ofce in
the group. It is argued by N dem bu that no~one woulcLmake
such a challenge against grontocratie authority, and assert his
claim to o ffic e , unless., he .possessed-special-powers,...beyond the
pow er inherent in his kinship position alone. T o m ake such a
breach in the social order, an order validated h y N dem bu ani
m istic religion, a man m ust have powers o f an anti-social kind,
w hich in this society are hostile to kinship. T o counteract these
powers, the forces o f the constituted order must be m obilized.
This is done b y a conscious, and public statement of-official-norms
o f correct behaviour and b y a public condemnation o f the-actions
o f the delinquent.
B u t w h y, in this case, was no recourse made to a diviner,, w ho
w ould alm ost certainly have given the M ukanza.people the
answer they desired, that Sandombu was a sorcerer and-should
be_expelled from the com m unity? T h e answer given b y the
com m unity that Governm ent w ould get to hear o f it and take
action-against the, accusers, is im portant, b u t it. is .not .the. w hole
tru th .,. A n, and., aggressivem an. like Sandom bu. m ight
havereported to the Borna an accusation o f sorcery ,made .against
him. I say he * m ight have reported the matter because this
betrayal w ould have led to his banishment from the. village,
w hich he did not w ant. B ut the people feared he m ight so
report .them, in w hich case the other senior men o f the village,
w ho made the charge, w ould undoubtedly have been sent to
prison. A nd it is fo r this reason that N dem bu rebuke. Govern
m ent in private discussion for * protecting witches and allow ing
them to m ultiply*.
B u t over and above this cogent argum ent there was the im por
tant consideration that, objectivelyspeaking,- Sandonibu^was an
asset_to.. foecom m unity. JBe was. in. regular em ploym ent and
hisr position as capitao enabled him to obtain labourers*, jobs at
H illw ood Farm fo r other men in the village. He was
the ear-of M r. Fisher, the em ployer, and-could flnd-places-for his
ow n kin.. Again, there w ere several-boys -in the village.w ho had
passed Standard II at the local mission out-school.and required
m oney fo r their boarding foes at the Middle. School , at. Kalene
Mission. These boys w ere partly financed by Sandombu...... From
Kalene they m ight go on to Mutanda School at.Solw ezi where,

M atrilineal D escent


ifth eyp assed StandardV I,. they_^otddjobtam_a^<juaIiicaticm for

fee-lucrative jo b .o fe le rk . T ^
assistjfeeirelders w ith m oney. O n the w h ole,. it w as. better to
retain Sandom bu as a_village m em ber than to incur hisdetennined
hostility b y expelling him . B u t he must at the sam e-tim e be
prevented from succeeding , to the. headmanship. T he upshot
o f the case was that M ukanza Kabinda becam e headman, an
appointm ent confirm ed by the Governm ent C h ie f M ukangala,
although he did n ot have a form al installation ritual. This
appointment was engineered b y im puting K ahalis death to
Sandombus sorcery b y w ord o f m outh, but not b y invoking the
crucial test o f divination. Sandombu could not argue against
this course, since the m ajority o f villagers gave M ukanza Kabinda
their support w hen C h ie f M ukangala came to the village to
ascertain their view s about w ho should succeed N o r could he
demand a divination to clear him self, fo r it w ould alm ost certainly
have gone against him . AH he could do was to prepare the
ground fo r an attem pt to succeed to M ukanza Kabinda.
This drama thus began w ith the breach o f kinship, norms
between"1Sandombu and .K ahali. This breach led - to -a -public
quarrel betw een them , in w hich the w hole villageto_ak. an
an interest. W h en Kahali died the quarrel gave rise to accusa
t i o n s o f sorcery against Sandom bu-behind, his...back. T he
accusations elim inated his prospect o f succeeding to headmanship.
T he drama culm inated in the appointm ent o f M ukan ^ . Kabinda.
This appointm ent led to the restoration o f a . social, order in
w hich Kahalis and M ukanzas generation contimied....tO-. hold
office and the possibility o f fission on lineage and...onsibling->
fam ily lines had been at-least tem porarily-averted. JSandombu
haiLno.t-been- expelled, from . the. wllage^.otit-had-beeti -excluded
from headmanship. T h e village had ..maintained its unity and
even strengthened it b y finding a scapegoat fo r-its . internal
B u t at this stage Sandom bu was still sm arting from his defeat
and further outbreaks o f hostility on his part occurred against
those w h o had ousted him . T h e m ost serious o f these led to
Social D ram a II, w h ich brings out clearly the. lines o f tension in
thejvillage structure o f M ukanza.



Schism and C ontinuity in an African Society

The Expulsion and Return o f Sandombu
(compiled from informants)



Sandombu bad married Mukanza Kabiizdas daughter Zuliyana

(IJ2). It may be supposed that he hoped by this means to attach
himself securely to the tw o village matrihneages o f CHIKENga
(r,E4), and malabu (I,Da), since she was connected to both through
her mother and father. He also hoped that she would give him
children, for although he had married eight times, no wife, except
his first, had borne him a child, and this was thought not to be his.
He had divorced Malona (see Social Drama I) before marrying
Zuliyana, because she was not acceptable to Mukanza people, being
reluctant to give them food and hospitality. .One _ojfthe__psychological drives behind Sandombus persistent . and openly , expressed
desire for headmanship probably consisted in the-act..that_the_names
o f .sterile persons are not given to children, -or rather, to~ put it in
Ndembu idiom, do not * come a name *.
Their spirits are called ayikodjikodji (sing, chikodnkodii) and in ritual
concerned with female fertility are given special offerings o f beer and
food by way o f placaidon at one side o f the temporary shrine, to keep
them away from the afflicted woman patient. Sterile people * die
fojr ever \ in that their names are not remembered
shrines o f quickset saplings are planted to honour them... Sandombu
probably feared that his name, i.e. his social personality, would be
forgotten, unless he became a headman, in which case, ho-would be
remembered as the occupant o f a poUticaL position. After a year
Zuliyana had borne no child although she came from a fertile family,
and Sandombu, while he was living at Hillwood Farm, used to vent
his spleen on her by frequently beating her.
Towards the end o f 1948 an Nkula ritual was performed at Mukanza
Village for Nyatioli (I,H n), sister-in-law o f Mukanza (I,F8), own
sister o f Sakazao (I,H9), and classificatory mother o f Zuliyana (IJz).
Nkula is most often performed for women with menstrual troubles,
and it appears that Nyatioli was experiencing a difficult menopause,
with much menstrual bleeding.
Sandombu, who was by this time working as a Road Capitao for
the Public Works Department in a camp about a mile from Mukanza
Village, was under the impression that an Nkang a girls puberty ritual
was to be held that weekend, and not Nkula. He ordered Zuliyana
to brew many calabashes o f beer for the occasion, for at Nkang*a nearly
all the people in a vicinage attend. A t the concluding phase o f Nkula
many people also usually attend, but this was only a * small Nkula *


Schism and Continuity in an African Society

confined to M ukan za V illage and a few external kin an d neighbours.

Z uliyana, w h o w as living in M ukan za V illage itse lf a t the time,
brew ed on ly tw o calabashes fo r this sm all ritual. W h en Sandom bu
cam e o f f d u ty m o st N d em b u rituals to d ay are h eld a t weekends
to accom m odate those at w o rk he fo u n d on ly tw o calabashes o f
beer. W ith ou t w aitin g to hear the explanation he began to beat
Z uliyana, w h o called o u t to her kinsfolk fo r help. K aso n d a ( I ,G i 5)
entered San d o m b u s village hut, and accordin g to his o w n account,
rem onstrated in a broth erly w ay w ith San dom b u. T h e latters only
response w as to take a b am b o o cane and thrash Z u liyana w ith it.
T h en a y o u n g m an called Benson, a m em b er o f the neighbouring
village o f N g o m b i, ju st back fro m the C o pperb elt, cam e n , snatched
the cane fro m San d o m b u and broke it across his knee. Sandom bu
lost his tem per an d bis head com pletely, accordin g to B en so n and
K ason da, and accused B en son o f bein g Z uliyanas lover. Others,
b y this tim e assem bled in the d oo rw ay, heard him g o o n to say that
he w o u ld m ake the m em bers o f M ukanza w h o w ere interfering with
his m arried life p a y fo r their intervention.
N e x t day, San d o m b u w as observed b y K a n y o m b u (I,Fp), M ukanza
K abin d as y o u n g er brother, and by their sister N y a m w a h a (I,F 7 ),
as h e w as stoo p in g ov er the footpath betw een M ukanza and B en son s
village o f N g om b i. T h en he circled the huts o f M ukan za and
N y am w ah a. H e w as heard, it w as alleged, to have in v ok ed the spirit
o f his deceased parallel cousin Lu pin da (I,G y), a hunter an d pow erful
personality in his life-tim e, to punish his enem ies. H e is then alleged
to h ave said to those sitting in the chota, I a m n o w returning to
Sak eji [i.e. to Ikelenge A rea]. T o m o rro w som eone w ill die in this
village. M ukan za peop le h ave n o sense.
E a rly n ext m orn in g w hile the aged N y am w ah a, sister o f M ukanza
and m other o f K aso n d a, w as p ou n din g cassava, she felt a spasm o f pain
and fell d ow n in a helpless condition. K ason da, w h o h ad been
w o rk in g as a gard en b o y fo r the D .C ., w as in form ed o f this and
obtained perm ission to return to the village, five m iles fro m the
B o m a . N y am w ah a w as gravely ill. H er last w ords before she died
w ere San d om b u has killed m e.
K aso n d a w rote an. an gry letter to San dom b u, retailing his m others
last w ord s an d d em an d in g th at he return to M ukan za and g iv e a foil
explanation o f his suspicious behaviour. San d o m b u cam e back and
all the senior m en o f M ukanza assem bled in the chota to hear w hat he
had to say. It w as p u t to him b y M ukan za that he h ad cursed the
village b y in v okin g. L u pin d as ghost. T h is fo rm o f curse is called
hushing*and b y N d e m b u and is classed w ith sorcery ( w uloji ). San
d o m b u denied that he had cursed the village in this w ay , b u t he ad
m itted that he had been an gry, and expostulated that he w o u ld never,

M atrilineai D escent


under an y circum stances, have killed his m other \ w h o m he loved.

H e to o k o u t a 10s. n ote and said : * H ere is m y m on ey. G o yourselves
to a divin er in A n go la, and find o u t the truth.* (T h e P o rtu g uese,
accordin g to N d e m b u , d o n o t regard w itchcraft accusations as a
crim inal offence, so that divination is le g a l inutheic territory.) K an y o m b u (I.Fcj), M u k an zas brother, to o k the m on ey a n d dem anded
. 3. xos. in addition, sayin g that the com pen sation he sh ould p a y fo r
causing N y a m w a h a to die m ust b e set at 4. It w as n o use, he w ent
on, den y in g that he h ad threatened the village w ith sorcery. M an y
people h ad heard h im , so that there w as n o need fo r a divination.
B u t San d o m b u insisted w ith tears in his eyes that under n o circum
stances w o u ld he k ill his m other, o r in deed, an y o f his kin . K an y o m b u
then said that after the m ou rn in g ritual (Mudyileji) San d o m b u m ust
leave the village, and stay aw ay until he sh ow ed b y his beh aviour that
he could liv e p rop erly w ith his kin. T h e others agreed, an d San dom b u
left the village fo r a b o u t a year. I 11 the end he h u m bly asked M ukanza
fo r perm ission to return.
In the m eanw hile, in view o f San d o m b u *s evident sorrow at
N y a m w a h a s death, public opin ion w av ered as to the cause o f death,
and secret divination into it w as m ad e b y m eans o f a m edicated
po u n d in g -p o le {ngotnbu yamwishi), and blam e w as fin ally ascribed
to the husband o f N y a m w a h a s d au gh ters d augh ter fro m C h ibw akata
V illage in the sam e vicinage. It w as alleged that N y am w ah a h ad
tried to keep her gran ddaugh ter in M ukan za w hile her husband
w an ted to take her to his ow n village. In an ger he h ad * sh o t m edicine
at her and killed her.
S an d o m b u w as absolved fro m b lam e and allow ed to return. B u t
in order to sh o w finally that his liver w as w hite tow ards the people
o f M u kan za, d ead an d alive, he p aid M u k an za a g o at and to o k part
in a village ritual perform ed to placate the restless spirit o f N y am w ah a
w h o had caused m an y p eop le to d ream o f her and caused a m inor
epidem ic o f illness. It w as th o u g h t th at she w as disturbed by the
* troubles in the v illag e \ S an d o m b u plan ted a tmtyombu 1 tree to
his m o t h e r in fro n t o f M u k an zas hut, and M an yo sa (I,G i 3)
inherited N y a m w a h a s nam e. B o th S an d o m b u and M ukan za prayed
to N y a m w a h a, and m entioned that th ey w ere n o w reconciled.
T h ro u g h the plan tin g o f the tree, th ey said, she w ou ld be rem em bered
b y her relatives. Finally, three lines o f p ow d ered w h ite clay w ere
sprin kled o n the g ro u n d fro m the base o f the tree b y M ukan za, and
all the m em b ers o f M u kan za V illage w h o w ere present, m atrilineai,







1 Trapneli and Clothier, p. So, give the botanical term for this tree, which
they call mwyombo, as Lannea antiscorbutua. C h eck L ist, issued by die Forestry
Office, Northern Rhodesia, identifies mwiombo [sic] as K irltia acuminata.


Schism and Continuity in an African Society

patrilineal and affinal, were anointed by Mukahza with white clay,

symbolizing the basic values o f Ndembu society good health,
fertility, respect for elders, observance o f kinship dues, honesty, and
the like.
The prodigal nephew had returned.
In,, this drama the breach o f relations w hich ..inaugurated the
series o f events was a quarrel between a married ,couple w ho not
only stood to one another in the relation o f husband and-wife
but_were also cross-cousins. Furthermore, since. Zuliyana (I,j2)
was jaxhild o f both the village matrilineages whose -members had
effectively prevented him from succeeding, the cleavage between
Sandombu and the rest o f the village entered., into...his. m arriage
relationship w ith Zuliyana. As usual, the .breach .of relations
received visible expression in the violation o f. a cruciaL norm
governing those relations. In this case the norm , governed
marital behaviour and can be stated as follow s.;. * N o husband
should beat his w ife violently w ithout good reason.or .without
a..fault on her part/ Znliyana.-had.brew&cLa^sinal 1enam ount
o f heexJthaa her. husband-had.otdered^not-because-sho.was lazy
or,Li,wilfully, disobedient, - butbecause-Jhen. husband- had mis
understood. w hat cerem ony was to be perform ed. Sandombu
probably.thought, that she shouldhave.brew ed ..aJarge amount
of-beer in any case, whether the ritual w as a m inor one. or not,
sim ply because he. had ordered her to do so. T he fact, that she
had not done so w ould almost certainly have been interpreted by
him as agreem ent w ith her .village relatives .and-jdisohedience to
him, her husband. B u t to the other members o fth e . village, his
beating o f his w ife must have seemecLa- breach o f the norm that
a m an must n ot beat his w ife -without-good. reason. T h ey felt
m orally justified in intervening^ although actual physical inter
vention was left to an outsider from another village. Their dis
approval o f his action was im m ediately interpreted b y Sandombu,
and probably not w ithout reason, as a recrudescence o f their
hostile attitude to his claims for headmanship. He lost his
temper, as I m yself have know n him to do on m any occasions,
and reviled everyone w ithin range. It is impossible now to find
out whether he m erely reviled (ku-tukana) the villagers or cursed

M atrilineal "Descent


(ku-shingana) them . It is allow able, though frow ned upon, to

revile in N dem bu society. O utlets fo r suppressed feelings are
sometimes venial in a small-scale society w here personal relations
are intense. B u t cursing, since it involves the raising o f a
m alevolent ghost (musalu) to kill, is illicit, a form o f sorcery.
Sandombus rivals say he cursed the v illa g e ; he him self admits
only to having reviled them .
U nluckily fo r Sandombu, the death o f Nyam waha, the
deatbuof Kahali ( fF j) on a previous occasion, occurred ju st w hen
it was m ost lik ely to confirm die w orst suspicions.ja.his_.ellow
villagers. H e was publicly denounced as.,.a-i/q/L.(sorcerer), and
in spite o f his genuine and obvious, g rie f wJhich-madeLa-profbund
impression, he was tem porarily banished.. fr.Qm Jb.e_.village.
Unlike_Dram a I,., the., mechanisms .used, to seal. .up. the. breach
between Sandombu and ..the rest o f .the village w ere of. a., m ore
form al-and institutionalized nature. In..Drama..I,. it-w ilL b e re~
callecLthat although.the village, inform ally. discussedJSandombus
behaviour in the chota, he him self was not present and na..action
was. taken against him. JustL. enough m obilization..o f ..public
opinion was made to bar effectively his w a y to the headmanship
and.-ensure that M ukanza Kabinda should succeed. .. But. in
Dram a II Sandom bu was form ally excluded from the. village fo r
a probationary period by a general assembly o f the village. Later.
doub.tarosem them in dso village..m em beE sw hether-tbe. grave
aLUsation m ade against him was.just^and-by. one-m ode o f divina
tion- he was adjudged innocent. Therefore he was. received back
into the fold b y means o f a ritual w hich reasserted the. solidarity
o f the lineage of. jstyachintang a (I,D i ), to w hich .not only
Sandomb.tLb.ut_.alsQ..headman.. M u kan za. Kabinda,^.Kanyombu,
Ka.soncla, and. M anyosa, among.. the,. aiid exheadm an.Kahali Chandenda and. the victim Nyam^vaha. am ong
It-w U lhe n o te d th a t ..although, action. .was.taken. against San
dom bu, the m atrilineal lineages o f M ukanza V illage were_cateful
to appear to be m orally justified in their actions and.t.o_avoid..o.pen
violence, against him F.or.example, as B.enson, an ojutsider,
w hobroke -Sandom bus. .cane. Kasonda. S^ d om b u s classificatory_bro.ther,. confined h im self,to.., mere. remon&tration...Kanyom btbL M and n ot the h eadman M ukanza
him selfo. pronounced the sentence o f. exile ; . and._.this__was a


Schism and Continuity in an A frican Society

tem porary, and not., a permanent,,..exile._T h en.^ decision. was

obtained from d^\dnatiQn_whicK allocatecl blam eTorM yam w ahas
death to an outsider, and not to Sandombu.....Mo. attem pt was
made to..bring,, a divorce suit, on, Zidiyana!s.b.ehal.agamst Saxidombu. b y her. relatives .o f. m a ia b u .lineage. T h ed aor-w as left
open for. the reconciliation that fin ally occurred....-At-the same
tim e it . was. m ade clear.. to..alL-who_.might have-supported San
dom bu in the village that he was com pletely unsuitable for the
position o f headman.
B y ju ra l and. ritual ,means the. crucial norms.of.kinship-and
affinity had been reaffirmed,.and Sandombu had.been, shown not
to have conform ed w ith them. Sandombu had been com pelled
to accept his ju n ior status in the new order o f things in which
M ukanza Kabinda was village headman. Sandombu had been
reintegrated into the./village com m unity, but .the. condition o f
his -re-admission .was that he had to return ,as a-penitent., and to
renounce his claims to office. Conflict.had .been,...temporarily
m itigated but. not,...of co.urse,..eHiiiiciated__ItsJnevitafale revival
snake, w ith. the..human. .face...which, is .used, by- Ndembu-sorcerers
as a.familiar,.^vividly_symb-olizes..the.-secret.malignity.and_s.udden
em ergence o f .masculine jealousy, o ver, succession-to office.
pQcess o f both jhese, dramas has been shown to. run from
| : '/ :br<^cft@arough crisis and redress-to. reintegration._Inform al and
i^ S^ ^ ^ B ijiiral.-m echam sm s, and ritual, w ere -brought--into play
to.restore the..equilibrium of:the. disturbed. so.ciaI..gro.u.p.s.._Is there
and-functions o.these..difietentmodes-oredEessing-and.adjusting
| conflict,-ju d ritual ? It-would-seem ^at-the-present stage
o.anthrop.ological -enquiry, - that .jural... m achinery_isem ployed
when conflict between persons, and.groups. is-couchedJln-terms o f
an appeal b y both m on-norm ,1 or,
when norms conflict, to a com m on fram e o f values.which_o.rganize
a society*SJiorms into a hierarchy. Thus, when Sandombu spoke
angrily to his relatives, they all accepted a com m on norm govern
ing social intercourse nam ely, that no person should curse
1 C Professor Gluckmans discussion o f how Lozi judges work with
norms in cross-examination in T he Judicial Process among the Barotse o f Northern
Rhodesia (1955), pp. 49~SO.

M ain lin e d D escent


another b y raising a pow erful ghost. In other w ords, human

beings w h o w ere members o f a single society had an interest in
one anothers w elfare and should not w ish one another dead. In
N dem bu society the w ish is thought often to be father to the
d eed ; in fact, w hen it is publicly expressed, it is stated b y
custom to have the m ystical p ow er to do the deed. Thus,
Sand0mbu..argued...dbathe. had
w hile his reladves held that he.had-not reviied^butJha<Lcursed.
B oth sought., to ju s tify . their actions in term s. of-the-com m on.
norm' that cursing is. anti-social and im m oral. T he corresponding
jural situation w here tw o norms com e into conflict but are
hierarchically organized is exem plified in the history o f Sandom bus
m arriage. A lth ou gh Sandom bu treated Zuliyana abom inably
in N dem bu eyes, beating her w ith out cause, no d ivorce action
was undertaken b y her relatives. A norm o f N dem bu society
is that relatives should protect a m arried w om an from an unjust
husband. B u t the unjust treatm ent m ust continue for a long
time and be exceptionally severe before they intervene to end the
m arriage. T h e norm that marriages should rem ain unbroken
supervenes over the norm that relatives m ay at tim es offer
sanctuary to an ill-treated w ife. In this case Sandom bus beating
o f Zuliyana did n o t bring ju ral mechanisms into action although
it m ight w ell have done so i f Zuliyana had refused to go on livin g
w ith her husband.
O n the other hand,, ritual mechanisms are. irivok.ed. w hen i t
is felt- that, the fundam ental .norm s o f society_ themselves__are
threatened, or. challenged, n ot b y a single individual but by_.the
social, group w hich, operates, b y means o f adherence to them . It
is not that persons o r groups phrase theindisagreem ents in terms
o f a .com m on norm to w hich all parties ,basically-conform , but
that hostility .comes. to. be .feltL.byall_concemed_against ,the. con
straints im posed by- the norms-tnemselves. R eb ellio n d evelo p s
against the very w a y the social system is ordered,-and..a-.challenge
is made- to the establislieclm oraLorder, withits^norm s-aiid.,their
evaJuatLve. fram ew.ork I -ventmeL_to_.suggest foat._every jtiine a
norm Js'broken b y o n e individual,.a b y
every other individual in the group, to .do. likew ise._Breaches
represent, .constant., tem p ta tio n sto ._the_members_o..the_group
to rebel'against norm s, critically connected-w ith-the-unity and
persistence o f the group Thesetendencies to-com e in to conflict


Schism and Continuity in an African Society

w ith the norms must be purged o f their socially disruptive-quality

ifth e group is to remain integrated. B itual is the saciai.mechan
ism/by which a group is purged of. the.. anarchic_and.disruptive
impulses which threaten its crucial norms and values.- These
impulses are present in the m ajority o f .its members.-.and.~come
dangerously near to overt expression i f there, has-been.a_.!ong
series o f quarrels between its members.
Hut this temptation to rebel and even revolt, may
constant disputes between individuals and factions,-which in fact
manifest deeper conflicts between different In
Ndem bu society, we.have..discemed_.thatn.cer.tain~-principles o f
organization, which appear to give rise to consistent-norms, in
practice originate processes which., w ork against _.one_another.
O ver-a-period o f time,. conflict-oprocess_Jeads_tc^breaking o f
particular relationships and fissionL .o f particular-grQups^-but new
relationships-and--gro.up&..on.the old-patteoi-areuestablished The
contrary processes- another, within. the-.s.tnucture
o f the changing,. but-, repetitive, .field.
processes have worked out their mutual._compensation.^_the
conflict between . them w ill- continue_to ._produce_naggingIy
persistent quarrels, which are ineradicable- untilne-w-relationships
are_formed. In a society, which is. not.changjng,..the_sa.ai6_pro
cesses w ill then produce similar quarrels, in the-new_relationships.

Q m this view o f society, norms and. their., supporting values

can only appear to be consistent, since they the
presence o f contradictions w ithin .the., structure, itself. The
situations must arise where the. norms..whichdetermine the
course, o f action, to be taken ..cannot be. clearly ..and ..consciously
afiirmed for the acceptance o f all parties, since, each _can_ claim
some'support from customary values.- It is here-that intrigue
may? become rife and. disruptive. In Ndem bu .society,....based
as_it is on close interpersonal ties, it is also at. this stage-tiiat illfeeling becomes charged w ith the malevolent, mysticaI_power
o f witchcraft and sorcery. Ill-feeling is not merely, im m oral:
it-is charged w ith the danger o f disease, -death-and-.other mis
fortunes to ones fellows.
Here judicial ..decisiomcan .condemn_6ne. -OX,^.. dis
putants, but it.cannot always, relieve the quarrels so .as. .to preserve
the threatened relationship. Accusations o f sorcery_and witch-

M atriineal Descent


craft m ay tem porarily emphasize for. tke, parries_them ystical

danger, o f their ill-feelings, as happened to Sandomhu- w henJiis
* m other * died. Ritual,, like beliefs in w itchcraft, is a custom ary
device w hich lifts the emotions o f peapieand-the. vahies..o society
to a. m ystical plane where they have power-beyond-their-secular
effects. It m ay he invoked to affirm th e uniryoTa-relationship
or group over and above the quarrels which-are rooted-in-confficts
between their organizing social principles. It is significant-that
as accusations o f w itchcraft and sorcery follow . orL._what. we: call
natural misfortunes,which._ate.__as.crib.ed_ to_. illdfeeling.. in the
group, so .similarL.situations,-Eather..than.-open. dispute,JeacLto-the
perform ance..of- rituals.1 -H ence, .it_seems.to_ me,_.afterL_amajor
crisis^roduced b y these conflicts, o f process,. rituak ace-employed
to affirm that reconciliation has been achieved. _These__rituals
may-'even be em ployed after a ju d icial decision appears -to have
settled rights and wrongs,, fact the cause_o_.disp.ute is
beyond settlement. -And finally,., w hen all-attem pts to preserve
existing relationships have failed,, final breaches, -often-provoked
b y w itchcraft charges, are confirm ed b y rituals whicburestate the
norms as consistent and enduring, even ,though ,n ew .relationships
have been established. Those w h o disputed bitterly fo r headmanship, w ithin a village, m ay-becom e helpful relatives-when, they
reside in tw o diffrent villages w hich each appear...to_confbmi to
the JSfdembu ideal.
In the tw o dramas set in M ukanza V illage, w.e-hav.e_ seen a
series o f these canflicts-and_a-series o f attempts-by. members, o f the
village m atrilineage to resolve them , , first .by.inform al,, then b y
form al, jural measures. W ith the expulsion of-Sandom bu a new
leuel o f .conflict had been reached. D oubts as to the..justice o f
their action in banishing him seem to have beset th e. leading
members o f the. village. T h ese.found_.expressian.Jn..the...result
o f .the divination and in the b elief that the spirit -o f N yam waha
was -haunting the village. T h ey appeared-to..represent the
uneasy stirrings o f a sense o f collective guilt. T he people o f the
village had exiled their kinsman, in w hom -flo w ed-the blood o f
1 C f. the view s expressed b y M . G . M rw ick in * T h e Social C on text o f
C e w a W itch Beliefs \ Africa, x x ii, 2 and 3 (1952) and b y J. C . M itchell
in * A N ote on the African Conception o f Causality \ The NyasahndJournal,
v , 2 (July, 1952), pp. 51-8.


Schism and C ontinuity in an A frican Society

collective life, a man w ho had in m an y w ays.treatedthem gener

ously ; and perhaps they felt that theyhad sirmed agam stjhe.deep
value set on kinship itself as an organizing principle o fso cia l life.
I go t the impression from the w ay they talked, that...they felt
gu ilty lest, in order to further their sectionafaiidpersonal interests,
they had excluded a brother and^a .benefactorsJroixL_participation their jo in t existence as a com m unity o f hlo.o_d.and place.
W hen Sandombu expressed his sense o f penitence and asked to
return, there follow ed not m erely an expression o f his ow n con
trition at the tnuyombu tree o f his * m other but prayers to her
b y her brother the headman, and b y her very ow n daughter that
a o f them should be forgiven. Xhe life .o f the. com m unity
should begin anew w ith Sandom idst-once more.
T h e -tnuyombu tree,, plan ted ...because- there, . had- been-conflict
between people and w ithin people themselves, _now.stands, and
w ilkgrow . and .branch-out .as. a sym bol, that peace, is. restored and
the values-.ofJkinship are once m ore renewed w ith fervour.
Judicial mechanisms., tend be. Jn.vo.ked_ tou.rcdr.ess...conflict,
where the conflict is overt,, and these judicial mechanisms-involve
rational investigation into the m otives and . behaviour o f the
contending parties. R itu a li mechanisms tend to. b e utilized
where conflict is at a d eeperlevel. Here, conflict-expresses itself
through projection that is ,. in the collective association o f
misfortunes w ith ill-feeling and the w orking o f m ystical beings
and. forces, w ith dreams, and w ith answers to divination,. T he
conflict is between norm and impulse in each individual.m em ber
o f the group, but since in a tightly-knit group similar impulses
assail com m on norms in situations w hich em brace.the, w hole o f
its membership, this conflict attains social recognition. ..through
repetition dow n the, years, and ultim ately cultural techniques
are devised to handle it. It appears to the. members o f the.sQ.ciety
as though mysterious forces are attacking the. v.ery_foundations
o f the .moral and social order, not from w ithin,
from w ithout, in the form o f witches, spirits and mystical, powers
w hich penetrate individual members or some representative
individual in the group in the form o f dreams, illness, infertility,
madness, etc. V iew ed in this w ay, the tnuyombu tree planted
to the dead * m other * N yam w aha m ay be regarded as a means
by which the forces which are potentially disruptive o f the village

M atrilinea D escent


are drawn o f f ; and die conflicting members o f die group unite,

purged o f anger, in am ity o f com m on worship, around it. The
very.tensionsan daggressionsjhat this social system thus
become means to give its norms a new charge.of^energy so that
nat_anly._are.they, rein stated h u t.alsoth ey.areth ejn orefferven tiy
accepted. Thus, Sandombu m ay w ell have w ept because he
believed, w ith other Ndem bu, that his norm -breaking behaviour
and disloyal self-seeking impulses m ight have caused Nyam wahas
death. Pressure is^exerted b y such beliefs ..on the individual to
conform to. the norm s o f N d em b u so ciety, .because,- i f h e per
sistently breaks them , people w hom . he. lo ves-w ill-d ie. The
overt expression o f ill-feeling, especially during periodsjafxrisis
in the a group,, is thought to. render, the. ohjectsjof.anger
vulnerable to m ystical danger. It is thought,..for.example^Jthat
the fomiliars o f w itchcraft and sorcery are emb.oldenedLby._the
anger/of their owners to dem and the * m eat f tliose_against
w hom their owners are. incensed If.hasty_..anger_iipens into
a ptatracted grudge (chtela) the fam iliar cannot be. gainsaidin its
demand to kill. V iolen t persons are thus constrained not on ly by
the-public opinion that they are witches, but also -b y them ow n
fears that inadvertently their anger m ight k ill ..som eone, they
A_further...diferen.Ge_betwc.c.n. the. kinds. ,_o.situatiQn. ^
jud^cial..md_ritual machinei^ respectively: em^
ployed, m ay here be considered. .R ecourse is often mad^To law
when the livin g quarrel. T h e nature o f the conflict brought to
lig h t is usually specific findLsL.remedyr and
i f possible, to bring about a reconcihationbetw een the opposing
parties. T h e conflict is man-made and im m ediately intelligible.
B ut- w hen a breach in social regularity is made b y _some
naturalm isfortune such as the death .o f a. m em ber,of the .group,
or a .fam ine, or a plague, and if. the naturaLorder-is.thought-to,be
sensitively responsive ,to the m oral condition ,o f society, .then .the
calam ity allow s o f a num ber o f alternative interpretations. A
w ide-range o f conflicts between persons and. factions... in -th e
disturbed group is brought to light. I take up this point in
greater detail in the follow in g chapter, in w hich I attem pt to
analyse the pattern o f intrigue in M ukanza V illage. B u t I w ould
like here to emphasize w hat seems to m e to be a significant
difference. W h en rules are broken b y livin g -persons, judicial


Schism and Continuity in an A frican Society

action can fo llo w and. this, action can speed ily seal o f f conflict
w ithin the orbit o f a single relationship or w ithin a small sector
o f th e'social system. B ut when., a., severe.. naturaLrais&rtune
precipitates crisis, practically every, .latent source_Q..conflict in
the.system is made manifest. These m ay be disclosed in accusa
tions o f w itchcraft/sorcery, or in confessions o f guilt b y those
w ho feel that they have broken some crucial norm governing
the intercourse o f the living w ith the living or w ith the dead.
Since nO-.specific..norm has.been .patently..broken-,it_~is im
possible to introduce judicial machinery.oredress.~.--Eitherxitual
o r the ascription ..o.witchcraft/sorcery .to .one,oits- menihers is
resorted the stricken, group. InJboth eYents.the.fundamental
norms_and ..values o f .the .community.are-.emphasizedJn-symbol,
m im e by.the .p.urging,,.o.a...witch., the group
reaffirm s.its solidarity. . ..In. such situations irremediable__c.onflict
has been felt .to exist between the m ajor principles b y w hich the
group. .is organized., There is n oth in gjforjt but to_.lay stress on
values to w hich all men subscribe,..regardless-of.their.particular
loyalties and interests.
It should be pointed out that w hile judicial m achinery is often
deliberately- .and consciously brought- in to . play w hen men
quarrel, ritual seems, .to. emerge, as .the spontaneous response to
the .moral discom fort experienced ,b y a group-disturbed by-som e
natural...disaster. L aw is thus directly related to man-made
breach in social, .regularity,...while. ritual_isohliquely..related to
natural breach in it.
Finally, when natural disaster.strikes^, .n o.legaLrem edy
for it. O ne cannot ,sue m ystical beings, and powers in .any court
and obtain - compensation. Y e t . the disaster -h as-to .heum ade
intelligible in terms o f a m oral .order . o_the.._universe^^This
can. be done, as I have mentioned^ b y x confession .o f ones ow n
g u ilt in.a ritual context, or b y a projection_o-the-g.uilt-outwards
onto an external enem y, a w itch or a sorcerer.1
It-has been shown how iiLa.long-estabKshedwiHage,xnatrilineal
affiliation constitutes the -cruciaLm eans-of.providing-village inte1 1 am grateful to Professor John Barnes, of the University o f Sydney, for
making this final point.

M atrilineal Descent


gration_B oth dramas pow erful .matrilineal des can.

be,-andJhow they resist-disruptive forces- put in to a ctio n -by con
flicts _of interest .w hich arise .withm jh e m^
m atriliny itse lf People live together, because...they-,,are.matrilineally rdlated, but ju st because-they are- m atrilineally related they
come into conflict over officeandover the inheritance o f property.1
Since the dogm a o f kinship assemtha.t-m atrilineal-.kin,participate
in. one anothers existence, and since the norm s _of.kinshipstate
that--m atrilineal .kin., must at all times help. one.another, open
physical violence between them seldom takes, .place. . Their
struggles are phrased in the idiom o f sorcery/w itchcraft and
animistic beliefs.- - spite-of. the..fear. arou&edJby-witchcraft
and sorcery, m atrilineal kin m ay go to strenuous.lengthsJto avoid
finally ostracizing one o f their num ber w h o is believed -to be
a sorcerer or w itch. ConfHct is endentic m.the-SO.cial ,structure
but a set o f mechanisms exists w hereby conSict-itself-is -pressed
into the service o f affirm ing group unity. . In the n ext chapter I
go on to show h o w the conflicts over succession to headmanship 2
inherent in the m atrilineal structure o f M ukanza V illag e broke
1 When a person dies, his or her most senior matrilineal kinsman, usually
an own mothers brother or uterine brother, becomes executor (nyamufu,
literally mother o f the dead o n e '). Traditionally, the standing crops in
the dead persons gardens were consumed by the funeral gathering (rhipenji)
that came nightly to mourn for a month or more. The funeral camp con
sisted o f kin and neighbours o f the deceased. At the end o f the chipertji (or
ckibimbt) the executor divides the dead persons movable property, including
such items as guns, tools, utensils, cash, small stock, clothes, bows, spears,
beds, stools, etc., among the matrilineal kin o f the deceased. Uterine kin
o f the executors own genealogical generation, i.e. uterine siblings, receive
the largest share ; next, his o r her sisters children, then sisters daughters
children. I f any property still remains it is allocated to more distant matri
lineal kin in order o f seniority. It depends on die good will o f the executor
and the agreement o f the matrilineal kin whether or not the wives and children
o f a dead man receive any o f his property. The deceaseds hut is burnt down
and he is buried in his clothes and blanket, while many o f his cups, plates,
and other utensils are nailed on trees in the graveyard as a memento mofi.
2 Headmen in the past enjoyed a few privileges, such as gifts from their
villagers o f first fruits and the first calabash brewed o f each kind o f beer.
At ritual gatherings and other public occasions when beer and food were
distributed headmen were entitled to the first share. But the primary value
o f the headmanship o f a long-established village, such as Mukanza, lay in its
prestigeit is a great achievement to build up an enduring residential unit
from Ndemhu individualists.


Schism and Continuity in an A frican Society

out afresh in a new series o f social dramas. Throughout my

second tour o f fieldw ork it became clear to m e that the village
was rapidly approaching the major crisis o f fission, w hich will
probably take place w hen M ukanza Kabinda dies. I then go
on in subsequent chapters to discuss a num ber o f situations in
w hich fission actually occurred in M ukanza and in other villages
and analyse the tensions that led up to it in different villages.
In .spite o f all efforts to maintain the unity of. village .members,
fissiah frequently^The_.circumstances.-in_which it
occurs and the different form s it assum ed.w illbe described and
analysed, and the relationship between, the- subsystem o f the
village and..the inclusive system, form ed b y N dem bu-society
w ill. be exam ined in a variety o f aspects.

N the last chapter I exam ined tw o social dramas which, made
visible and explicit the underlying social structure o f M ukanza
V illage, and .exhibited in. action the lines o f tension and* struggle
w ithin the m ajor m atrilineage. . wish .to. .analyse a further
set of-dramas w h ich arose out o f the same fundam ental structural
situation after a passage , o f three. years in, M ukanza V illage, as
that-village becam e the scene o f new conflicts, for-pow er. B ut
first I m ust outline the events that preceded the next o vert rupture
o f relations between the leading persons in the village,
social, dram a has passed a clim ax, and the group in v o lv p ^ ^ ^ p
to have-been reintegrated, there often follow s, in N d e^ h u h^ei:
a period o f apparent equilibrium during w hich, in the mami
interpersonal behaviour is in consonance w ith generally accepted
norms o f conduct. B ut beneath the surface conflicts, o f interests
go on and private intrigues provide m eans-whereby-individuals
seek, to realign the social structure in pursuit o f their ,own-advan
tage. Beneath the manifest pattern, o f daily, interactions, a re
shaping. ,o f the social groupis taking...placet-transferences o f
persons from one faction to another, loss an d .replenishment o f
personnel in the follow ings o f leading..m en^so-that-in- effect a
new and at first hidden set o f pow er relations gradually comes
into being. There com es a tim e w hen this reaUgnment-becomes
visible in.afresh, social.drama. A t the culm ination o f this drama,
i f unification is -re-established,...the^gro.up_may_exhibit^a.marked
shift-in the balance o f power.-between.. variau$__components.
In Jhse intrigues., the diffrent, principles. of-N d
organization are m anipulated b y the m ajor intriguers to. further
their o w n ends. - These interim periods
struggles...are really the., con tin u ation .of struggles-irt-adifferent
and-indirect guise....In th em .w e.fh id attem p ts Jbeingrnade_by
parties-who have secured social, gains, in preceding, social-dramas
to consolidate their position,, and. .attempts b yi defeated..parties
to repair their dam aged prestige, attach to. themselves- fresh



Schism and C ontinuity in an African Society

supporters, and w in over to their, side, members o f the - oppos

ing.- groups.
A fter the reconciliation between Sandombu and the other
members o f the major lineage o f M ukanza, several years elapsed
during which there prevailed on the w hole an appearance o f
outward harmony between M ukanza Kabinda (I,F8), Sandombu
(I,G io), Sakazao (I,H9) and Kasonda (I,G i5), the leading men
o f the village. It is true that, on occasion, bickering took, place
between Sandombu and the others over m inor matters. One
such dispute arose over the allocation o f a goat paid to Sandombu
b y the husband o f his classificatory sisters daughter as a customary
gift to the local head o f the nearest maternal descent group when
a w ife bears her first child. M ukanza Kabinda, as headman o f
the village, said that the goat should be killed and divided im m e
diately between all the village members for * there was meathunger in the village . B ut Sandombu pointed out that the
goat was in kid and that i f they were patient they w ould have
several goats instead o f on ly one to eat. M ukanza became angry,
but Sandombu, claim ing that he was the senior livin g m others
brother o f the wom an and thus had the right to do as he pleased
w ith the animal, sent it to a village some miles aw ay to run with
the flock o f his blood-brother. V illage opinion was divided on
this issue, some siding w ith M ukanza, others w ith Sandombu.
D uring this period o f peace, Sandombu was careful on the
w hole to avoid public dispute w ith M ukanza, and, in fact,
sought to ingratiate him self w ith the other village members.
From 1948 to 1951 he did not live in Mukanza V illage but in a
camp o f road labourers a m ile aw ay, where he was capitao or
H e managed to obtain the em ploym ent by the
Public W orks Departm ent o f a number o f M ukanza m en and
youths, thereby putting them under an obligation to him . In
his camp o f mud huts he occupied a position analogous to that
o f a village headman and at w ork he enjoyed greater authority
than N dem bu headmen norm ally possess. This situation was
brought to an end in m id-1951 when an African R oad Inspector
recommended to the D istrict Comm issioner that Sandombu
should be dismissed on account o f the poor condition o f the
stretch o f road for w hich he was responsible. Dismissed he
was, and it is interesting to note that in this situation all the
Mukanza people, including the headman, took Sandombus part

M atn lin eal Succession


in attributing his m isfortune to the repaym ent o f an old grudge

by the R o ad Inspector, w h o had quarrelled w ith Sandombu
when both w ere livin g in C h ie f Ikelenges area.
Sandombu n o w decided to build near M ukanza V illage. He
did not build his new house in the village itself since there was
little space in the hut-circle fo r such an edifice as he intended to
erect. H e selected a site about a hundred yards from the main
village, next to the main Bom aKalene road, where he w ould
be in constant contact w ith the daily stream o f people passing;
along it, could offer hospitality to the wealthier and m ore im
portant, and he aufa it w ith any D istrict news he m ight turn to his
advantage, such as advance inform ation about paid em ploym ent.
W ith his savings he em ployed an O vim bundu bricklayer to
make fo r him a large three-room ed house o f K im berley brick
and a carpenter to m ake him a handsome door, w indow-fram es
and furniture. H e then applied his indefatigable energy to the
task o f m aking large gardens o f m illet and cassava, and urged on
his tw o hard-w orking w ives to do likew ise. His senior w ife at
this tim e, K atiki, was the daughter o f a Lw ena w om an slave o f
Governm ent C h ie f N yakaseya, and had a w ell-grow n fam ily,
one o f w hom , a son b y a form er husband, built a pole-andmud hut beside Sandombus. m ediate aim was
to build up b y any means a follow ing, to settle, w ith him , and
whenJMukanza died to make, a renewed bid fo r the headrnanship
o f the- village. It is at this point that w e m ust pose the.question:
w h y did Sandombu still desire to becom e a traditional village
headman w hen through his participation in the^modem casheconom y he m ight have expected to prosper as a petty .trader-and
cash-crop farm er ?
A ll over M w inilunga, and especially in the northern pedicle
where Sandombu had spent a large part o f his w orking life, the
traditional village was givin g grou n d before. jfo e sm a ll ^farm *
asjthfc typical, form o f settlem ent..... JEarmJheads. w ere disencum
bering themselves o f m any o f the.obligations. o fk in sh ip , .and
retaining for their o w n use and for. the. use o fth eir. elem entary
families m oney they earned as .wages and jb y the_ sale_ of_.cashcrops o r surplus subsistence-crops. B.utJS.a.ndo.mbu .still hankered
after, the h ead m an sh ip o f a _tm ditionaL village. W hy.?.-. The
answer to .this question lies, ! think, in the -fact^tliat-the-kind o f
response made to socio-econom ic.


Schism and C ontinuity in an A frican Society

iage.ly_.a.. function .q, relative . age. Q ld men. lik eM u k an za

remained obdurately conservative,.anddeplored., w ays,
although they w ere indulgent towards the young people, who
practised them . Y ou n g m en lik e . Sakazao's uterine nephew,
Pearson (IJ5), capitao o f a European-owned, store at the Boma,
and.JVianyosas oldest son Daudson (fH y) w ho had. w orked in
Chfogola, had accepted the new order o f things., andJWOOS. smart
European clothes, ow ned bicycles and. gram ophones,.played
guitars, used Copperbelt slang, and attended-traditional rituals
only to jo in rings o f young, people, w ho danced the latest,..dances
im ported from the Rhodesian or Congo urban areas an.d sang.the
latest local or urban song-hits \ Re.tween .these_.tWQ. extremes,
Sandombu, ..and .to...a. lesser degree.Kasonda,.though., th ey had
grown, up .in .the .era....o f. m oney-earning, .. still_ belonged ..o a
generation_.which saw. success in life as.measured_by.the,numher o f
followers , a man could acquire, and. not b y the irisignia.of .conspicuous w ealth that could, be. purchasedb y m oney._.True,
Sandombu, and also Kasonda, had large ..houses. well..equipped
with' furniture. T hey had mosquito-nets and o il lamps. . B ut
these._signs o f wealth .were rather indices o f success in the traditional order than signs o f an altered w a y o f life, in vo lvin g the
acceptance o f entirely new modes o f . behaviour, and .o.a .new
scale o f values. R oth these m en continued- to -w their
gardens, to gossip and discuss cases. in .the yillage, chota, tgjp&rticipate in ritual as cult-m em bers and patients, to exercise their
traditional rights and fu lfil obligations as kin, and to interact
w ith the older generation in terms o f traditional .norms. T hey
were not ashamed, as m any o f the younger people w ere, o f the
ancient w ay o f life ; they lam ented that, so much, o f it. had already
passed aw ay, and m ore was passing beneath their v e ry eyes._ B ut
th ey felt that the royal road to eminence .within. .theJm lage w ay
o f life now la y through the acquisition o f cash._ Possession o f
cash gave them large houses,-bride=wealth for-several-w ives w ho
m ight g ive them children, and .them .to -offer-hospitality,
and the means o f retaining their chddren.and..givmg. .thesea g o o d
education. T h ey wanted m oney to-better- their-position w ithin
the .traditional system, not as a means .o f loosening-thek-ties w ith
it. For instance, for these men o f m iddle.years,_a.gunwas better
than a bicycle, and to obtain, .the wher.ewithal..to_buy one they
w ould save up fo r m any years, keeping...the money_in_a..bag or

M atr lineal Succession


boxJburied secrectly in the ground, could,

not-find it.
I woidd_like_ta.-emphasi2e a$ this_p.oint_ the..main^diiierences
between, tw o m odes o f incorporation in ^the.. .m odern. cash
economy. O n .the. one. hand, a m an. .cm ..acquire...wealth hy
w o rk in g in th e .W hite economy_as^a.wage labourer,-eitherJocally
or by- m igrating to the urban centres, on the R hodesian line^ofrail. O n the other hand, a m an can. obtain m oney Jocally by
selling surplus subsistence crops,, b y,_such as
rice onground-nuts,, for. European or African, traders,.and
by setting, up .in business as a petty trader, tailox. or. tearrxoom *
proprietor_-(i,.e. b y .selling.ibod .along a
labour, route). It seems often, ...even today,, to b e t h e a i m o f
returned, labour-m igrants or ...ofLpaid ^workers. in .local .employ
ment to obtain influence, and. subsequently. .office,. in .traditional
villages. M any o f them see the village as their, ultimate, hom e,
andjregard their w age-labour as. a_means._ofaeqmring.the.. wealth
that w ill give them prestige in the. village.sphere.._..Bju.ijhey
havejsubstantial savings, it m ay happen thanthey wish^to. invest
these in. capital equipm ent, such, as a sewing-m achine,--with the
aid o f w hich they m ake up store cloths in togarm ents for pay
ment, .a bicycle, on w hich they can travel , to buy. goods, cheap
and sell them elsewhere, at a profit,_or. a, storehouse-or.tea-r00m
madefirOm sun-dried bricks, to Once
they have taken this step and have invested m oney to .m akem ore
m oney, th ey find themselves increasingly embarrassed b y the
demands o f their k iir fojipresents in cash and kind. If they w ish
to become..-p.etty:-capitalists,.- they -must-separate--themselv:es~firom
the.village-Sphereuandthe v illa g e .w a y o f life JXhatris_.why it is
usuak to fin d traders and tailors hving w ith_theh iam ilies, in
different vicinages, and even.chiefdom s, from the. viUages.^ their
close m atrilineal kin. . T o 'a lesser extent, the same ..set,, o f con
ditions holds, good. for. .pettyxom m odity. c.ultiva.tors^.)dxeincipient
* kulaks-, w h o. grow crops specifically fbr sale If.the_.making o f
m oney tends to supplant, as a._major. aim. in life.._the._acqirisition
o f a-follow in g, people ,try .to.,,...t.Q.he, turned
into capital and the visible signs o f a. higher .status_in.. .terms o f
European values. A large fo llow in g then tends,.to_.become an
embarrassment rather than an asset. In essence, the..wage-eamer,
w ho acquires w ealth at the expense o f the W hites to .strengthen his


Schism and Continuity in an African Society

position in the traditional village, must be distinguished, fmm

the local ,m oney-m aker w h o grow s rich at the expense o f his
feUowrNdem bu.
B l4t.-$andomt>u, and f o r m a n y years Kasonda
sought to
usejheir..wages gained in w orking.for Europeans.. (Kasonda since
his__return from the Im e-of-raii had. w orked as-adomestic-servant
for a succession oD istrict O ificers).iii acqmringjHends^and sup
porters am ong their kin and neighbours to prepare the w ay for
succeeding to headmanship. Later on, towards jthe_.endLoF my
second period o f fieldw ork, both menjwere_cQming_to_realize
that the old order ..was doom ed,, and .that-tOb.ecom.e_eminent
they/ m ust. ...commit, themselves w hole-heartedlyto .the cash
econom y.... Hut. at the_time .of SociaLDxama^IIIthey- .were still
caught up in the contradiction odem
w ay o f getting a living. o f
values. The cash econom y. to..which .theyL..w.ere..c.ommitted was
breaking down. the structure .o f the. village_^-aiid_ye.t.itjwas their
am bition to obtain authority oyer a-disintegratmg_s.acial unit o f
this kind.
asystem w h ich , was.gradually-breakmg-do-wii-as-the-result o f the
introduction o f m oney. .He..was..not satisfied rto..remain_alone in
his h ig house, or even to attract Qthers,tofosL.settlement,_ He
wanted eventuaUy to becom e headman o f M ukanza and to spend
the evening o f his life as a Tespected eldet.o.the_n.eigh.bourhood
com m unity, supported.b y emQluments_fr.Qm the_salaried,_young
merL Jto .whose educatioix he had con.tributed- w lien .they were
M eanw hile bis prospective rivals for the office had been by
no means inactive. A member o f malabu lineage, w ho had
spent m any years on the Copperbelt, Line (I,H i 7), returned in
1951 w ith ms w ives. Sandombu prom ptly found him a jo b
as a road labourer, but Line soon quarrelled w ith him . Sakazao
(LH9) then persuaded Line to bring his uterine nephew Aram
(IJ7) w ith him and found a * farm * ju st behind Sakazaos own
huts. N yam uw anga (LG 17), the suspected w itch o f Social
Dram a IV , also had a hut built for her at Lines * form .
Sakazao, it w ill be recalled, was the brother-in-law and classificatory * grandson o f M ukanza Kabinda. H e was rather older
than Sandombu. Sakazao was much poorer both in cash and in

M u k a n z a V il l a g e


S ep t e m b e r 1951

The chota (mens forum) is in the centre of the village. Kasondas Kimberleybrick house is under construction (see p. 138).

!* iiwiiuttrii |"*T r~ *

Tm in-**~nriiir


kind than Sandom bu and Kasonda. His health was uncertain.

In his you n g m anhood he had been considered som ething o f a
* m ischief-m aker * (chipapokt). His favourite practice was to
travel to a part o f the D istrict w here he was little know n, offer
to sell a gun to a villager at a very cheap price, receive partpaym ent fo r it in advance, and set o ff hom e, prom ising that he
w ould return soon w ith the gun. H e never returned. H e had
been cured o f this practice b y his village kin w h o g rew tired o f
paying damages to Sakazaos victim s, m ost o f w hom trackedhim
dow n in the end. His personality was a curious m ixture o f
navet and shrewdness, but it was universally regarded as
lovable. Sakazao was not respected as M ukanza Kabinda was
respected, but he was loved. W hen his youthful exploits were
recounted in the chota, people w ould grin and shake their heads,
clucking their tongues and repeating his name. W h en he made
his pretendedly grum py sallies in response to teasing everyone
w ould be delighted. It was thought b y m ost m em bers o f both
village lineages that Sakazao w ould make an excellent successor
to M ukanza, since he was unaggressive, hated no one, and,
though som ew hat inactive, was hospitable to strangers. M ost
im portant o f all, no one had ever accused him o f practising
sorcery. B u t i f he became headman it was generally recognized
that the real m anagem ent o f affairs w ould devolve upon a younger
and m ore capable man. Sakazao w ould m ake an admirable
figurehead, but a poor representative o f the interests, jo in t and
several, o f his villagers. Y e t in his o w n w ay, and during the
struggle for pow er described in this chapter, Sakazao was astute.
Sakazao played on the suspicions o f bus classificatory * father \
N tololu (II,D z) o f Shika V illage, that his relatives w ere bewitch
ing him , in order to get N tololu to com e and live uxorilocally
in Line's farm . N tololu, perhaps influenced b y Sakazao, built
a hut there. N tololu s w ife Kalusa (I,G i9) was the sister o f
N yam uw anga (I,G i7 ), and both these w om en, although about
the same age as Sakazao (LH9), w ere his classificatory mothers o f
m alabu lineage. N yam uw anga had a hut built fo r her beside
Kalusas. Just previously, Kalusas and N tololu s daughter Ikubi
(I,H i4) had undergone her puberty ritual at M ukanza Village.
N tololu and some boys from M ukanza built a hut for Ikubi at
Lines farm , w here she was to aw ait her husbands return from
the Belgian C on go. Also at Lines farm w ere a wom an o f

M atrilin eal Succession

P '






Schism and C ontinuity in an African Society

m a l a b x j lineage and her husband, a road labourer in Sandombus

Thus, at the end o f the dry season in September 1951, it was
already possible to observe in the spatial structure o f Mukanza
Village visible signs o f gradually approaching conflict for author
ity. Behind Sakazaos hut w ere five huts belonging to members
o f malabu matrilineage, form ing, as it were, a nucleus o f social
pow er. A hundred yards from the village behind the circle
occupied by his generation mates o f nyachintang a lineage was
Sandombus great house, occupied by himself, tw o w ives and a
grandson o f his senior w ife, and her sons small hut. In Mukanza
Village itself Kasonda had built a large Kim berley-brick house
and a smaller one for his recently married third w ife.



Kasonda is Accused of Bewitching his Uncle Kanyombu

(my own observations)

/ o



Before proceeding any further with this account I must mention

my own impact and influence on the life o f the village. A t die
beginning o f 1951 1employed Kasonda as a cook and general henchman.
After a time I made a permanent camp in Chief Ikelenges area,
returning to Mukanza Village on intermittent visits to watch the
progress o f events. In June o f that year Kanyombu (I,Fp), Mukanzas
younger brother and a leading doctor in several cults, died after a
protracted illness. Kasonda told me that the cause o f his death was
probably Sandombus sorcery. He had received a letter from
Sakazao asking him to return for the funeral ritual. Since I was about
to pay a protracted visit to some bush villages at that time I was
unable to release Kasonda whose services I needed on tour.
In July Kasonda received a verbal message that his classificatory
mothers brother, Kajata (I,Fio), an extremely old man o f malabu
lineage who had recently returned to his matrilineal kin after about
fifty years absence, had also died; and the message urged him to come
to the funeral ritual. It happened that I had planned to pass by
Mukanza Village on m y way to the capital village o f a Kosa chief
and I took Kasonda with me. When we reached Mukanza we found
the entire village in an uproar. The day following Kajatas death
Mukanza himself had been taken seriously ill. Sakazao and Ikubi
(I,Hi4), his young female relative in Lines farm, had been ill for
about a week. A social drama was, in fact, in progress.
This time the breach in regular relations between villagers had not

t____________ _ J


Schism and C ontinuity in an A frican Society

been caused by a dispute but by the natural calamity o f what I took

to be an epidemic o f malaria. For among Ndembu -uiduckyevents
occurring in the natural order .are thought... to..originate. in_J>reaches
of-tbe moral order. More explicitly, in -the. co n textofvillage life,
these misfortunes onginate either-in_..tlie .maligmty o f kin. w ith evil
powers against other kin or in .the, p_uiutive ,actioii_oancestorL.spirits
against their living kin who have fbrgotten..thermor_transgressedJdnship "norms. The epidemic just mentioned was thought to have been
caused by one or the other o f these agencies.
I arrived at the phase o f crisis while the people were still arguing
about the origin o f their troubles and about the best means to banish
them. Kasonda was received w ith hostile looks, and it soon became
clear that he was regarded as the sorcerer who had bewitched
Kanyombu and Kajata, and who was now threatening Mukanzas
Presently, Nyamukola (I,Hxo), Mukanzas senior wife, and Nyakalusa, Sakazaos junior wife and slave, began openly to accuse Kasonda
o f having gone to Kalene Mission, where the missionaries have
created a sanctuary for people banished from their homes in Angola
as witches and sorcerers, in order to procure medicine to kill his three
senior relatives o f the first ascendant generation and his classificatory
sisters son Sakazao (I,H9).X Manyosa (I,Gi3), Kasonda*s uterine
sister, and Chawutongi (I,G iz), his matrilineal parallel-cousin,
strenuously defended him, as did his senior wife, Mangaleshi (II,Ez),
but they were in the minority. Kasonda had already begun, with the
help o f his cross-cousin Gideon (II,El), Mukanzas son by a deceased
wire, to build his Kimberley-brick house, and Nyamukola shouted
that this was clear proof that he was planning to succeed Mukanza.
W e went in to see Mukanza 12 who was lying on his bed, obviously
in a condition o f great fear and distress. Mukanzas two wives
followed. An extraordinary scene took place beside what looked like
becoming a death-bed. Kasonda became very angry and said:
* Alas, Africans are very, very stupid people. Kajata was an old man.
God (Nzambi) has taken him. But i f God has taken Kanyombu, it is
obvious that Kanyombus spirit {mukiski) is now troubling the village.
Women are m ighty fools and w ill tattle about anything. How could
I have gone to Kalene Hill when I was working for Mr. Turner all
1 Sakazao was grouped with the senior in the linking o f alternate generations.
2 Under normal circumstances Kasonda would not have entered Mukanzas
and Nyamukolas hut; since Mukanza was his mother s brother and Nyamu
kola rated as his mother-in-law, to be avoided, since she was the mother
o f his divorced wife, Koniya (I,Jr). B u t as m y servant, the dominant role he
chose to play, he ignored these customary rules.

M atrilin eal Succession


the tme ? W hy should I kill Kajata ? For meat ? I had plenty

o f beef, goats flesh and duiker meat at Mr. Turners place. I do
not eat human beings even i f other people in this village do.
A t this point I intervened and offered to take Mukanza without
delay to Kalene Hospital. The old man accepted gratefully, not, I
suspect, because he had much faith in European medicine, but because
he longed w ith all his heart to get away from the situation in which
sorcery was thought to be active. Once away from his relatives he
would be safer.
Others demurred and we all went into the chota to discuss the matter
fully. Sandombu was away at work at the time. Mukanzas two
adult sons, Line (LH iy), Kasondas tw o grown nephews, and the
headman o f a nearby village whose father had come from Mukanza
Village, were present. The women sat outside the chota and took
a lively share in the discussion.
But first I dosed the sick people with quinine and offered to
take them all to hospital.
Manyosa (I,Gl3) said that her mother* brother, Kanyombu, had
in his lifetime drunk mwiyanawu medicine, (that is, medicine
which would enable him to come back as a ghost (musalu) and kill
members o f the lineage o f the sorcerer who had bewitched him).
She implied by this that since so many members o f malabu lineage
were sick the sorcerer or witch had belonged to that lineage.
Nyamukola, w ife o f Mukanza, had now calmed down somewhat
with the prospect o f accompanying her husband to Kalene Hospital
before her, and interjected that it might well be that Kanyombu was
haunting the village, not because o f revenge medicine, but because he
had given Mukanza detailed instructions about how his funeral
stretcher should be made and medicated, and Mukanza had omitted
to carry them out. Perhaps he was angry at this neglect. If Mukanza
were taken away from the village and given European medicine he
might recover. Yes, he should go.
Sakazaos wife then interposed that Sakazao had been dreaming
about Kanyombu all the time he had been ill. Yes, Kanyombu*s
ghost was to blame for the trouble, not Kasonda.
Kasonda spoke up at this point, saying that M r. Turnea: had once
given Mukanza some money to pay Fishers lorry driver to transport
Kanyombu to Kalene Hospital when he first became ill, but that
Mukanza had divided the money among his wives and children and
not sent Kanyombu at all. This was another reason w hy Kanyombu
was angry. It was not Kanyombus ghost (musalu)* his evil power,
that was aedve ; Kanyombu, in spite o f what Manyosa had said, had
never drunk mwiyanawu for he had never observed the taboos o f
mwiyanawu, such as eating only one side o f an animal or chicken.



Schism and Continuity in an African Society

But Kanyombu's spirit (mukishi)yhis moral power, was angry because

his relatives had scorned his wishes and neglected to look after him
properly. Kanyombu had died because he had not been taken to
hospital. If, on the other hand, Mr. Turner took Mukanza to hospital
he would recover.
As Kasonda made these points, most o f the men clapped their
assent, and cried Eyo-o \ that's it
The upshot was that it was
agreed that Mukanza should go to Kalene Hospital, but Sakazao and
Ikubi (I,Hi4) refused with the excuse that m y medicine would soon
make them recover.
Mukanza and Sakazao both recovered, but the young woman
Ikubi died.




e !
o !|



In this situation misfortune was interpreted .hy._variaus_p.eople

in .terms .ofla...bxeach o f several-different, moral, norms,and thus
it hrQUght..tCL..the_.s.urfaee..a~number.x)..difierent latent conflicts
within the village. This tim e the lines QCconillct.were not clearcut.. I f there, is .a.real .dispu.te...b.etweenHving_vjhagejnianbers,
created b ya breach ..of rules, j.udiciaL.action_can_follo-w. W hen
serous illness occurs it is believed ,to be. .caused_by_one._of a
number o f mystical beings o r .forces.. The_operatioruoaIL.those
beings and forces is associated w ith disturbances...o- the_m oral
order. Therefore__every^laio_wi]LitenxQ_the.paients.b.ehayiour
in- the recent past is brought under review ,.jbi....o.rder_.ta..survey
possible sources o f disturbance, ...and., to._locate_that-which is
even_before.resort is made_tQ^ <livination. Various theories -are
advanced about the precise nature and .name-o-the- supernatural
cause o f affliction, on the. basis. o f .the.-patienth. kn.own_cond.uct.
Since it frequently happens that contradictory., theojries_are put
forward by parties whose.interesfcs. in. ffle_parietit.are_diametrically
opposed, representatives .o f the. major interestr-graups- go-join tly
to aniaccredfted-divmer,.whoJsusuallyirL.a..distant-aEeawhere the
vdIagers have no^ kin, in.Qrder.tO-0-b-tain.-a..decision--Tmtinged by
parrisan.bias.and. validated_byLjtriysLtical beliefs.
In. situations_a.g^ .misfortune, moral. misdemeanour. is in
volved, in. tw o .ways. . Either the .misfbrturLe. is. theLxesnlt.of the
victim 's ow n misconduct,. or. else it is -the.wickedness o f
others, in the first event, the cause of_misfQrtuneJs3.elieved to

M atrlineal Succession


he-the.p.unitive action,.o f
guardians o fkin sh ip m orality ; in. the second ewent. the-m alevole3PLce..Qwitches,._sorcerers _or .ghQs.tshas_heen_at w o rk.
T h ere-atecertain . exceptions ..ta this rule__It is . thought that
sometimes, in. extreme_cases,._theL.killing_by-sorcery. o.a_notorious
sorcerer, or w itch is justifiable.....Butnwhen a^sorcerer-is-believed
toJh-ave beenJkiHed_bysorcery, ^-.cetiain^ be
attaghftil m *b<- layer, ?1thougfr it* *h case
invoked. and his.identification. isdeft togossip.-and xumour^ . O n
unjustifiable m alevolence. th eir- living.....kirn. ....When a
spirit is th ough troh avecausecL unjustifiedseriousillnessor grave
misfortune,, it is not execrated but placated,-althougbuelements
o fe x o re ism m a y h e.foun d .in -the ritual.
W hether misfortune. _or illness .is..ascribed.. to...the..unrighteouS
ness..of the. victim^ or to the..unrighte.ousness-o..a-Corpareal or
spiritual enem y, depends on a num ber..ofactors :_. the-degree or
kind, o f m isfortim e_; or status of the victim in varying
social-situations. ;. and. the_-Current_leveJ_Q..morale._of_his. local
g ro u p ro fld n .... h i m ost cases, hath theories in vari.QUsforrns are
put_forward. Som etim es. sufficient consensus.. is.ohtrinecLahout
the particular cause to avoid recourse ,to -a.-d ivin er,-w h ich is
costly... and today risky as w ell. For instance, w hen a certain
pregnant w om an becam e seriously ill, it was agreed by her kin
and husband that the Wubwangu or tw in ritual should be per
form ed for her, since both her m other and m others m other had
been mothers o f twins, and it was considered likely that her grand
m others spirit was afflicting her reproductive processes. W hen
the w om an adm itted that she had forgotten to make an offering
o f first fruits to her grandm other, her kin regarded it as estab
lished that the latters spirit was offended and had afflicted her
in this w ay. Hut it is m ajor ...theories to
develop and the diviner in effect choose.betw een them.
This tendency for tw o .theories to .develop sometimes .produces
unexpected-consequences. W hen I was in M w inilim ga, a song
was current throughout C h ie f M w ininyilam bas area, the gist
o f w h ich was that w hen a headman died in one village, the senior
m en o f tw o village matrilineages each w ent to a diffrent diviner
and each obtained from his ow n diviner the name o f his rival
as the sorcerer w h o had bew itched the headman. T he song


Schism and Continuity in an A frican Society

related how they returned to the village at the same tim e and
each prom ptly accused the other o f the headmans death. A dis
pute arose and each led his kin group aw ay to form a new village,
so that the name o f the old village * died \ There is a droll irony
in the flat account o f the events w hich is irresistible. N dem bu,
in ia ct, often express scepucism.abnut.divinationsLmadeL.fbii-others,
but__se]dom abQut_thos_e _saught._o_ut_hy. themseives.
The situation in M ukanza V illage brought out into the open
not o n ly an im portant cleavage in the social structure, but also
interpersonal grudges. W hen N yam ukola accused Kasonda o f
bewitching her husband she was not o n ly attacking him out o f
jealousy because his im proved financial position as m y cook gave
him additional qualifications for headmanship, but also because
Kasonda had once been the husband o f her favourite daughter
K oniya and had treated her so badly that she had left him after
a m onth o f married life. A nd when M anyosa (I,G i3) said that
Kanyom bus ghost was attacking members o f m alabu lineage
because K anyom bu had possessed vengeance m edicine, she was
not m erely defending her brother from N yam ukola but also
revealing the hidden conflict that existed between nyachin tang a (I,D i ) and malabu (I,D z ) lineages ; fo r i f Kanyom bus
vengeance medicine w ere w orking, this could only mean that
someone o f that lineage had killed K anyom bu b y m ystical means.
Later in fact the m yth became established am ong members o f
nyachintang a lineage that N yatioli (I,H n ), sister o f N yam u
kola, and divorced w ife o f Kanyom bu, had bewitched the old
man. T h ey had lived a married life o f continual domestic
bickering and N yatioli had developed a grudge against her
form er husband. W om en are believed to possess familiars o f a
kind different from , and m ore dangerous than, those ow ned by
men. T h ey are called tuyebela, tushipa or andumba, and have the
variable form s o f little men w ith reversed feet, hyenas, jackals,
owls or small rodents. T hey are believed to be inherited m atrilinealiy and to make demands on their owners to kill junior
relatives. T h ey m ay be refused three times but after that, how
ever hard the ow ner (nkaka) m ay plead, they k ill their chosen
victim . W hen the owner o f tuyebela dies, they seek out one o f
her close kin, it m ay be a daughter or uterine sister, and attach
themselves to her whether she desires them or not. Their
victim s are the husbands or jun ior matrilineal kin o f their owners.

M atrilineal Succession


[f a husband is killed his ghost becomes w hat is know n as a

kahwehwi, and leads the band o f tuyebela. He is m ore pow erful
than they, and m ay on ly be gainsaid once. Tuyebela are excep
tionally active i f their ow ner has a grudge (chitela) against some
one, w hen they are liable to take instant action against him w ith
out asking their ow ner. N yatioli had been suspected for some
time o f having tuyebela, as indeed had her classincatory m other,
the notorious w itch N yam uw anga (I,G i7), also o f malabu
lineage, o f w hom m ore anon. N o doubt, one o f N yam ukolas
(I,H io) reasons for pinning the accusation o f w itchcraft on
Kasonda w as her fear that she herself, com ing from a lineage
the w om en o f w hich w ere suspected o f w itchcraft, m ight be
accused i f M ukanza w ere to die. His niece M anyosas reference
to K anyom bus vengeance m edicine was a fairly direct hint to
that effect.
Kasondas later handling o f the situation, after his first outburst
o f anger, was extrem ely diplom atic. First o f all he cleared him
self b y attributing the sickness in the village to K anyom bus
spirit. N ext, he cleared K anyom bu from the im putation that
he had used vengeance m edicine, a practice bordering on sor
cery. I f this argum ent w ere pressed too hard he realized that it
w ould be tantam ount to a public recognition o f the cleavage
existing betw een the tw o village m inor m atrilineages. B u t i f he
could gain general acceptance o f the vie w that K anyom bus m oral
spirit and not his vengeful ghost was troubling the village, he
w ould have produced a form ula that w ould unite the village
again. A ll members, including Sakazao, had erred in neglecting
the interests and ignoring the last wishes o f the old m an. That
was w h y he had afflicted so m any people w ith illness, as a rebuke.
Furtherm ore, since I had given them m oney to spend on Kan
yom bu, hut M ukanza had spent it on his ow n fam ily, Kasonda
subtly im plied that M ukanza was under an obligation to make
restitution to m e, and through m e, to Kasonda him self, w h o had
urged on them that K anyom bu be sent to hospital before he died.
This M ukanza could do in tw o w ays, first b y absolving Kasonda
from blam e, and secondly b y goin g to Kalene Hospital him self
for treatm ent. I f M ukanza did this he w ould show publicly
that he accepted Kasondas point o f v ie w and did not believe that
he was bew itching him . In other w ords, Kasonda com pletely
reversed the case against him b y posing as the w ould-be










O '


Schism and Continuity in an A frican Society

benefactor o f K anyom bu and by im plying that M ukanza and the

other elders o f the village had w ronged the old man. He omitted
to m ention the death o f the other old man, Kajata (I,F io ; see
p. 138), for ghosts (nyisalu) are believed to have pow er to kill,
though spirits (akishi) are rarely thought to do so. Later, he
told m e that it had been accepted, in village gossip, that N yam uw an ga (I,G iy) had killed Kajata and I have no doubt that
Kasonda did not strive officiously to kill this im putation. Later,
N yam uw anga was accused publicly by Sakazao and N tololu
(II,D2) o f having killed Ikubi (I,H i4), so that it became natural
to believe that she had killed Kajata as w ell.
Kasondas m ajor political triumph was to have created a
general opinion that everyone in the village had wronged
Kanyom bu and that the main burden o f guilt was M ukanzas.
H e had kept the tw o lineages together, and strengthened the
view that he and not M ukanza was the real peacemaker, the most
im portant o f an N dem bu headman's roles.
B ut w h y was Sandombu at no tim e accused o f sorcery in this
situation, except privately, to me, by Kasonda w h o was intensely
jealous o f him ? As always, where human beings are concerned,
there is no simple and all-sufficient answer. O ne reason m ay
have been that for several years now Sandombu had lived quietly
in his Public W orks Departm ent camp and had secured w o rk for
village members, thus enabling them to pay their tax and clothe
their families and themselves w ithout having to leave their
homes. T he scars o f the old conflict between Sandombu and
the rest had healed. B ut a m ore cogent reason, I think, was the
sudden increm ent o f members o f malabu lineage to the village
population, creating a num erical imbalance in favour o f malabu
people over nyach intang a people, whose lineage had always
held the leadership. This led to tension between the tw o groups,
tension w hich was openly expressed in M anyosas reference to
Kanyom bus vengeance medicine, w hich form ed a com ponent
o f N yam ukolas accusation against Kasonda, and w hich later
gave rise to the n yach intang a m yth that N yatioli (I,H n ),
a malabu wom an, had killed Kanyom bu b y w itchcraft. The
fact that M ukanza him self was so ready to lend an ear to Kasondas
arguments, w hich im plied that nyachintang a people should
perform a ritual to placate Kanyom bus spirit and thus consolidate
the lineage, suggests that he him self was reluctant to believe

M atrilineal Succession


that Kasonda, his uterine nephew, o f nyachintang a lineage,

was the sorcerer. Sakazao, too, M ukanzas grandson, brotherin-law and friend, was in agreement w ith Kasonda*s view , for
he also was partially responsible for not carrying ou t Kanyom bu-s
last wishes, and the b elief that Kanyom bus spirit was punishing
the whole village probably seemed to him a means o f reuniting a
settlement o f w h ich he hoped at some tim e in the future to be
headman. Kasondas remark that w om en are m ighty fools *
refers to the fact that N yam ukola, Nyakalusa and M anyosa
attributed the com m on m isfortune to personal animosities ; i f
their view s had been accepted, the village w ould alm ost certainly
have split into its com ponent lineages.
Another im portant factor w as m y role in..the..situation_as a
stranger....Q nceM ukanzajw a^ in.m y.vanettehe,w asm y.responsibility a n d n o lo n ger th a to f thevillagejm em hers Frankenberg 1
has_poin.ted.out, in an,analy$is o_a .Yillage_comimniityin N orth
W ales, how the, ahout-matters
on..which_ the villagejm em bers. interlinked b y lon g propinquity.
kinship, re lig io n ,.. etc.,...were...divided^am ong themselves^. was
often thrust, upon. strangers-or. village.m em bers-w ho. w ere struc
turally outside., the. particular ..situation .of controversy ,_Sirnilarly Jn_this...Mdemb.u_viUage..o.Mukanza, had
died-arK alene .Hospital, N dem bunotions-his death
w ould have lain at m y door, I.. w as ^JEum.pean no
action x o u ld . he., taken.against m e. _Sjnce I had taken the .first
step in asking M com e ..with me.,._ICas.Qnda_conld_.not
be held responsible for his fate. As. it happened_Mukanza made
a speedy recovery and bo th<j-uence
acquired enhanced.prestige_in.the. village.
In this situation the structural cleavage between n y a c h i n tang a and m alabu lineages came into the open in the form o f
accusation and counter-accusation phrased in the idiom o f mys
tical beliefs. A lthough these tw o lineages w ere linked together
by m any ties o f m arriage and alternate generation alliance, at the
same tim e the distance from their com m on ancestress was begin
ning to be reflected in social and spatial segregation. M ost
malabu lineage members w ere n ow in Sakazao Js section o f the
1 K inship and Com m unity in a W elsh Border V illage (1954), unpublished
thesis presented for the degree o f Ph.D. at University o f Manchester.


Schism and C ontinuity in an A frican Society

village. W hen meat was divided M ukanza gave a large share

o f the carcase to Sakazao w ho took it to his ow n hut and sub
divided it am ong his lineage kin. O n marriage and divorce
malabu and n yach intang a people dealt w ith the bride-wealth
matters o f their respective kin separately. W ith the progress o f
structural tame the village cell was developing tw o independent
T he next social drama is concerned w ith divisions w ithin
malabu lineage itself and shows h ow such divisions w ere ex
ploited by Sandom bu to build up his ow n follow ing.
The Expulsion of the Witch Myamuwang a
(my own observations)
Shordy after the above events, the young bride Ikubi (I,Hi4) died,
I believe o f malaria, in Line's farm.
Ntololu (II,Dz), her father, prompdy accused Nyamuwanga
(I,Gi7), her classificatory mother, o f having killed her by means o f
witchcraft-familiars (tuyehela). The usual post hoc ergo propter hoc
arguments were invoked. It was said that Nyamuwanga had asked
Ikubi for some meat that she had cooked shortly before her illness.
Ikubi had said that it was for her parents and that she had not enough
to give away. Nyamuwanga had become very angry with her and
had threatened her in a roundabout way.
Nyamuwanga defended herself by saying that Ikubis death was
probably due to Kanyombus vengeance medicine, not an argument
likely to commend itself to Sakazao (I,H9) and his lineage kin, since
it implied that Kanyombus killer had come from their ranks.
Gideon (II,Ei), son o f Mukanza by a dead wife from Shika Village,1
and uterine sisters son o f Ntololu, father o f Ikubi, then lost his temper
and started to beat up Nyamuwanga. Sakazao stopped him, but
told Nyamuwanga to leave the village, and stay with her sister in
Chimbila Village in Chief Ikelenges area ; not because she was a witch
but because she was a trouble-maker. Line (I,Hi7) endorsed this,
saying that he would have no witches in his farm. A ll formed a
united front against the old woman. She collected her belongings and
fled to Chimbila Village.
Later Sandombu said she had been accused without proof and
offered her a hut site in his farm. Nyamuwanga, who had met her
1 See Appendix II for skeleton genealogy o f Shika Village lineage.

O O O O D O D O 9 O O -3 O '3 Q


o "'1
D ;'y

^ O ^
-T\ O
O ^


Schism and Continuity in an A frican Society

daughter Ikayi (I,Hi2) in Chimbila, came back with Ikayi in 1952

and Sandombu built them a small hut beside bis house.
Nyam uwangVs hut in Lines farm had been burnt and pulled
down while she was in Chimbila. Ntololu returned, with his wife
Kalusa (I,Gi9) and their rexnaining children, to his own village o f
Shika, saying that there was less witchcraft there than in Mukanza.
This was not the first tim e b y any means that N yam uw anga
had been accused o f w itchcraft, hi the summer o f 1950 she had
been accused b y M ukanza V illage o f killin g her uterine sister,
Shim ili (I,G i8), at Chim bila V illage, although Chim bila people
blamed Shim ilis other sister N yam alita (I,G2o) for her death.
It was also said that N yam uw anga Bad killed her ow n children
by means o f tuyebela. Kasonda denied this, saying that N yam u
w anga first acquiredher tuyebela from her m other-in-law , a slave
o f Shika V illage. W hen I put it to Kasonda that the generally
accepted belief was that tuyebela w ere m atrilineally inherited, he
adm itted that that was usually the case, but said it had not hap
pened thus w ith N yam uw anga. This inconsistency is under
standable in terms o f the social structure. Kasonda, for instance,
had no hesitation in telling me privately at other times that
w om en o f maxabu lineage possessed inherited tuyebela, when
conflict arose between the tw o lineages w ithin M ukanza Village.
B ut since Kasonda had acted as representative o f M ukanza
V illage in the discussion o f death-paym ent (tnpepi) w hich fol
low ed Shim ilis death at Chim bila V illage, he had obviously
tried to whitewash his ow n village m atrilineal kin in a dispute
w ith members o f another village, by claim ing that N yam uw anga
tuyebela, i f she had any, w ere acquired, not inherited. H e had
either com e to believe this him self o r else wished to clear the
name o f M ukanza V illage in m y eyes.
H ow , then, had N yam uw anga com e to acquire her unenviable
reputation ? In the first place, she was o n ly h a lf Ndem bu, or
rather h a lf K aw iku, since her father had been a Lw ena ; and she
had spent m any years in A ngola at her fathers village. Persons
ofLm ixed tribal origin, or w ho have been reared, am on goth er
Headman Chibw akata (cf. p. loff.) w h o was captured as a child
by C h okw e and reared in a C hokw e village is a case in point.

M atrilineal Succession


T h e y d o n o t * quite^b.elotig to^tKe locaL society^ and -a& such

make useful scapegoats for m isfortune. Jbx th e secQnd _place_she,
like/Sandom buT.had been a prodigiously hard w orker in her
gardens, knew where. to_plant.her crops, so that.she-would. obtain
a higher yield than other people, jand often-w entjon,w orking in
the h eat o f the day. w hen others had retiredto_gossip..iti. the
kitchens. ,.And so. the meyitahle.questiQn w as.posed...fooiruyhat
som ebsdid she .draw h erxn ergy. a n d w h y w ereh en xrops-better
than, anyone elses ? O bviously ..she.had supem aturaLpowers
w h iclrgave her outstanding stxengthLandluck^.. .^ h en .sh e. was a
young w om an she also had the.,,.something
o f a_joymphomanaac. ...Mdembu J n .conversation strongly, condem njsxcessive sexual desire and associate, it .with._prop.ensities
to sorcery and w itchcraft...L ike .w itchcraft,-desitc-^tuvumbi) is
hot..!, and in the boys* circum cision .ritual,jm ysticaLprotection.
in jthe form o f medicines is used against both w itchcraft and for
nication. A gain, like w itchcraft, sexual desire causes social
disturbance-rrmrital inhdeHty,^crimes o f. passion, jm d.argum ents
over jh e repaym ent ofh rid e-w ealth . In the w ords o f the most
often sung N dem bu hunting song, * the buffalo in the bush, the
w om an in the village, are the death o f man \ T he-tw o beliefs
coincide in the notion o f the kahwehwi,. the g h o s to f a w itc h s
husband,, .k ille d . b y . her.sorcery,_X hc kahwehwi -is~thought to
be an incubus w ho continues to have intercourse w ith his w id ow ,
and tO cgive..her., inordinate, pleasure. -On. thesethree counts,
thexi7-h er _foreignness, her. diligence,, and her forrner_nyniphomania, N yam uw anga came to be reckoned a w itch as she grew
older. A series o f deaths in her lineage w ere attributed to her
w itchcraft : those o f Shim ili (I,G i8 ), Kajata (l,F io) and Ikubi
(I,H i 4).

O ld w oraen are .oten..considered .to. be.w itches. LfLthey are

widowsjor-divorceesH keiSfyam uwanga^wkhjiO-Sonsjorhrothers,
they_have no one to obtain m eat, fo r. them, and have to xely on
the^iimcertaiii - generosity.of. rem oter. kin__T h ey..are .constandy
grambling._ab.out^lack-.Qf itieat, and. in. .a..society_with. a lively
belief. in_the necrophagous_practices-.ofwitches,_this- is highly
suspicious, behaviour. . It is. beHe-Ved that wit.ches.coliect. together
in_the bush_with their_fomiUars^.and. divide_andLeat-.the..bodies o f
their, wictim s. T h ey substitute in the m eanwhile a simulacrum
o f the deceased, fabricated from his faeces and bod y dirt, in his


Schism and Continuity in an A frican Society

place in the village ; and this is mourned over and buried b y his
relatives. It is, I think, significant in terms o f this belief that the
j alleged cause o f N yam uw angV s grudge against Ikubi was the
I latters refusal to give her m eat.1
P itc h e s usually strike,, at...their... close. jnatemaL.kin,_on..whom
they; depend, and.w ho therefbre are .those, w ho neglect.-them by
notJuM lling their obligations. Itis usuahy_Qld.wiomen.who are
negletted because they tend to. be an economic_burden_to the
com m unity. Thus allowances are seldom _mad.eTar_old_wonien
detected as witches by any meansL_of-divination__In...the_past,
it is said, they w ere burnt to death. Today,...they_are often
driven, from their, villages. Accordjuag_to_my observations, old
they. m ay be suspected of.witchcraft,.are...never_treated_ in this
way... It is .the class x>f_unfortunateddLwomeii_who are w ithout
im m ediate, male kin, but .who. are. classificatoxy_ m atrilineal kin
o f the village headman, that acquire .the. role o f scapegoats for
misfortune. T hey consolidate the rest o f the .village .or lineage
membership against them. H ow ever, in the case o f N yam uWanga, it w ould appear that Sakazao m ade a mistake in banish
ing her. For unknow n to him at the tim e, N yam uw angas
daughter Ikayi (I,H i2), w ith her ow n small daughter, had just
returned from the C ongo, having been aw ay from M w inilunga
for five years. She was staying at Chim bila Village. Her
husband had recently received a long sentence o f imprisonment
for manslaughter after a figh t w ith Ikayis current lover. I f
Sakazao and Line had retained N yam uw anga they w ould have
acquired tw o further wom en o f their m atrilineage w ho w ould
have borne children for that group.
B ut it often happens .that after a death has occurred,...intense
animosity, is atoused, a n d if this is. focused-on_a_putative_witch,
a reasoned consideration.o f the resultmgsociaLsituarLan.Js_swept
a w a y b y anger and hatred against her. T he prim e movers in
this m atter appear to have been Ikubis father N toloiu (II,D2)
and her cross-cousin Gideon (II,E i), both from Shika Village
lineage. T here is a long tradition o f interm arriage between
Mukanza and Shika villages.2 Kasondas senior w ife (II,E2)
1 Cf. how quarrels in preceding episodes arose over the division o f meat.
8 See Appendix II.

" fS


came from Shika, Mukaixzas first w ife (II,D3) w as a Shika

wom an, Kasondas father (II,D i) cam e from Shika, and so on.
For this reason, since marriage is liable to in vo lve disputes
between the kin o f spouses, considerable tension as w e ll as am ity
existed between the tw o villages. N yam uw an gV s form er hus
band, W illiam M atorokoshi (11,D 7), cam e from Shika, and she
had stayed w ith him there fo r several years. W illiam had
recently divorced her because he said that he used to dream every
night o f * animals * (tunyama), m eaning that he feared that he
was being bew itched b y tuyebela. H e said that these belonged to
his w ife, and sent her back to M ukanza V illage. In addition, she
had quarrelled continually w ith members o f the Shika lineage,
refusing to cook fo r the men in the ckota, and squabbling w ith
the w om en.
W h en Ikubi (I,H i 4) died all the accum ulated dislike o f
N yam uw anga felt b y Shika fo lk, and their suspicions o f her
w itchcraft, broke loose and received expression in the beating
given her b y G ideon (II,E i). Sakazao, furious at the disruption
o f the fo llo w in g he was beginning to build up, ordered her out
o f Line Farm. She had already left her h u t near the main village
circle to build next to Ikubis m other, Kalusa (I.G ip ), her uterine
sister, because she had quarrelled w ith Gideon and M anyosa
(I,G i3 ), her neighbours. G ideon was n o w vio len tly opposed
to her return. In any case Sakazao had ordered her to g o and
live in Chim bila V illage i f the headman w ould allow her to
do so.
W h en Sandom bu heard that N yam uw an gV s daughter had
joined her m other in Chim bila he im m ediately saw this oppor
tunity both to add to the strength o f his ow n farm , and to ensure
its continuity through tim e b y inviting them to stay w ith him.
A t last he w ou ld be in a position to acquire the nucleus o f a
m atrilineal group. Three generations o f m atrilineal kinswom en
w ould be added to his follow in g, even i f one o f them was a
w itch, another a prostitute, and the third a child in arms. In
allow ing N yam uw anga to reside at his farm Sandombu was
openly defying the people o f M ukanza V illage, and behaving
like the headman o f a rival village. H e was acting as though he
intended to split M ukanza V illage.
N yam uw anga eagerly accepted Sandombu* s offer, for she was
persona non grata at Chim bila on account o f her reputation, and

M atrilineal Succession

0 Q















Schism and Continuity in an African Society

besides she had left her large gardens behind at M ukanza. She
and Ikayi (I,H i2) came and lived at Sandombus farm where, at
the tim e I left M ukanza V illage, Ikayi was m aking m oney by
selling her favours to passing lorry-drivers, African, traders and
other w ell-do-do Africans, although not to indigent villagers.
She was reputed to give N yam uw anga,and Sandombu a share
o f her earnings, and became a lively source o f scandal am ong the
respectable older w om en o f M ukanza Village. A bou t this time
too Sandombu added to his follow ing a young lad named
Chikim hu (I,J6), daughters son o f Ikubis m other Kaiusa (fC ip ).
Chikim bus m other had gone w ith her husband to the latters
place o f em ploym ent in the Congo urban area. Chikim bu was
too w ild and w ayw ard to be taken w ith her to her new husbands
place, she said, and N yam uw anga offered to look after him.
Soon he too acquired a reputation for sorcery in spite o f his
tender years : m y w ood and water carrier, a tall youth from a
nearby village, gave notice because Chikim bu, h a lf his size, had
threatened to kill him as the result o f a quarrel over a game o f
hopscotch. Chikim bu was the problem child o f the local
Mission out-school and his teachers could do nothing to discipline
I have given this thum bnail sketch o f Chikim bu to show w hat
kind o f a follow in g Sandombu had collected around him to
further his unsatisfied ambitions. A t this tim e his farm con
sisted o f his senior w ifes children b y other men and a number
o f rem ote m atriiineal kin o f malabu lineage w ho w ere social
outcasts and misfits. B u t for this very reason they could be
guaranteed to remain lo yal to him and w ork hard for him accord
ing to their varied abilities. Chikim bu, for exam ple, was always
running errands and doing odd jobs for Sandombu, although ne
w ould do nothing for anyone else. N o t unnaturally, the people
o f M ukanza V illage referred to Sandombus form w ith a selfrighteous shudder as that village o f w itchcraft * (mukala wawuloji). B u t Sandombu continued to visit the chota every day, to
take part in discussions on such important matters as bride-wealth,
the holding o f rituals, death-payments, cases o f petty theft,
slander, etc., and to state his opinions forcefully and authori
tatively. It was clear that at some time or other he w ould make
another bid fo r the headmanship.
B y the m iddle o f 1953 it became obvious that the headman

M atrilin eal Succession


Mukanza. Kabinda was becom ing very old and infirm . H e

spent m uch o f his tim e dozing in his kitchen, and although he still
w ent into the bush to hunt, he seldom killed anything. The
question o f succession to headmanship was once m ore beginning
to disturb the village. B y this tim e I had made m y permanent
camp ju st on the outskirts o f the village, n o t far from M ukanzas
ow n huts. I was, therefore, in a position to observe the trend
o f events far m ore closely than before. Kasonda, still m y cook,
was n ow m aking plans to found a form o f his ow n . H e said
that w hen M ukanza died the village w ould break up, som e going
w ith him , others goin g to Sandombu, and others staying w ith
Sakazao. Shortly after I arrived M ukanza publicly announced
in the chota that he had appointed Sakazao as his mulopu o r suc
cessor. This was done in the presence o f N sanganyi, headman
o f the oldest K aw iku village, and o f N swanakudya, headman o f
an im portant neighbouring village belonging to the lineage o f
C h ie f Kanongesha. Sandombu was aw ay at the tim e on a visit
to Ekelenge Area. W h en he returned and heard the news, his
first action was form ally to succeed to the name o f K ahali by
asking N sanganyis permission to use it, a name w hich, it w ill
be remem bered, was the ancient title o f the M ukanza headmanship. N sanganyi, after dem urring at first, agreed in the end to
give permission. M ukanza publicly approved Sandombus step,
and said sm ilingly in the chotat w hen both Sandombu and I
were present, * N o w there are two villages, M ukanza and N swanakahali.1 Sakazao w ill succeed in M ukanza V illage, and perhaps 2
someone or other w ill succeed in Nswanakahalid Sandombu
said nothing but shortly rose and returned to his village. Trouble
was patently brew ing.
A w eek later Sandombu invited m e to drink w ith him in his
house. H e was broodingly drunk w hen I arrived. Soon he
asked abruptly, * W h at is the nam e o f this village in the Govern
ment tax-register ? * I said that it was w ritten down, as K ahaliM ukanza \ * Yes,* he said, * in 1947 Bw ana Heath [the D istrict
Com m issioner] cam e to the village to collect tax and Kahali
* Literally, * Heir-of-Kahali \
* The use o f this * perhaps * (kw iji) was a powerfully compressed taunt
which contained the notion that Sandombu's followers were ail witches or
sorcerers who might bewitch one another and was an oblique reference to
Sandombu*s sterility.


Schism and C ontinuity in an A frican Society

Chandenda, m y uncle, told him to change the name in his booh

from M ukanza to Kahali, fo r the old name from lon g ago is
Kahali ; but M ukanza Kabinda and Sakazao told him that the
village had been founded b y M ukanza fCandulu (I,F6) and was
know n b y all as M ukanza. So Bw ana Heath was very clever
and w rote dow n Kahali-M ukanza. B ut he was w rong. Y o u
have w ritten our history and you kn ow that the true nam e o f
the village is Kahali. Please g o to the Bom a and tell the D .C .
to change it to Kahali.* I replied that although it was true that
the village used to be Kahali, the m ajority o f people today
thought that the village should be called M ukanza. A t this
point Kasonda cam e in, although he had not been specifically
invited. H e quickly picked up the thread o f the discussion and
sa id : * A long tim e ago the Lenje people had a village in the
place where B roken H ill n o w stands. W h o remembers the
name o f that village n o w ? People on ly m ention B roken
Hill**, Broken H uT *, B roken H ill* even Africans. N o w
a long tim e ago there was a village called Kahali. That
village divided and a new village was started b y M ukanza
Kandulu. M any people came and visited M ukanza V illage. For
m any years people have o n ly been speaking o f M ukanza. It is
like Broken H ill/ Sandom bu angrily retorted that he had ju st
visited C h ie f Ikelenges area and that there people still thought
o f M ukanza V illage as Kahali. Kasonda replied, * Ikelenge
Area is another place, not this place/ T he conversation turned
to other matters, and after a w h ile I left.
Sandombu.went-om drinking..heavilyand-iruthe...evening
another social drama began w h icK revealed that th e jealousy .over
succession had n ot been eradicated but had beenJfestering.beneath
the. show o f outw ard harm ony. Sandombu was w ell aware
that M ukanza, Kasonda and Sakazao w ere attem pting to seal him
o ff from any possibility o f succeeding, b y asserting that his farm
n ow constituted a separate village, and that M ukanza V illage was
not continuous w ith Kahali V illage, w hich had, in their reckon
ing, becom e extinct. Perhaps he felt that he had made a mistake
in taking the name o f Nswanakahali, and that M ukanza and
Kasonda w ere chuckling over their exploitation o f that mistake.
W hatever the cause the fact was that b y nightfall he was in a
prodigious temper, and threw over the restraints he had imposed
upon him self fo r years.

M atrilineal Succession


Sandombu Slanders and is Slandered
(my own observations)
W hile Sandombu was drinking great excitement was aroused in
the village by the arrival o f a masked ikishi dancer from a Boys*
Circumcision Camp some miles away. A t a certain point in die
seclusion period o f the boys, the dancers emerge from the camp and
visit ndgnbouring villages, accompanied b y ritual officials from the
camp, to obtain gifts in cash or kind from the villagers or to extract
various kinds o f forfeits from them. The whole village was in a
condition o f social excitement, with drumming, dancing, singing,
and drinking. This reached a high pitch when m y kitchen accidentally
caught fire.
Zuliyana (IJ2), Sandombus junior w ife and daughter o f Mukanza
and Nyamukola, went to join the other women who had assembled to
sing and dap around the ikishi dancer. She came back in a temper
because she did not possess a new dress as some o f the other women
did, and began to harangue Sandombu on this theme. She was
assisted by Katiki, the senior wife, and both accused Sandombu o f
giving away large quantities o f millet beer to ingratiate himself with
people instead o f selling some to buy them new cloths to be made up
into dresses by Kasonda, who had by this time bought a sewingmachine. They said that apart from organizing the ckenda collective
tree-felling party, Sandombu had done little in the w ay o f cultivat
ing the millet, while they had weeded it, harvested it, thrashed and
winnowed it, ground it and made it into beer. Certainly he was owner
o f the seed, but they were entitled to some reward for their labours.
Zuliyana then said that Sandombu just wanted to make himself a big
man ; that was w hy he had taken the name o f Kahali. He wanted
to become headman o f the whole village. In view o f the previous
arguments o f the day this was an unfortunate remark and Sandombu
staggered towards Zuliyana to beat her. But he was extremely drunk
and she easily evaded mm and ran o ff to her mother and father for
protection. Sandombu did not beat Katiki, who seemed to possess
some sort o f influence over him, and she persuaded him to sleep o ff
his rage.
But when he awoke later that night, his wrath had not abated and
he rushed to Mukanzas hut. Meanwhile Zuliyana had told her
father that Sandombu had been boasting to many people that he was
die elder (mukutumpi) o f the village while Mukanza was his junior
(kansi, literally, * child *), and that the name * Kahali * was a * heavier *.
name than 4Mukanza . Mukanza had also been drinking and a












C )



M atrilineal Succession


major dispute developed between them. Sandombu accused Zuliyana

o f trying to bewitch him with tuyebela which she had been given by
her mother Nyamukola. Zuliyana began to deny this energetically
from the kitchen where she was sitting with her mother, but both
men told her to keep quiet. Mukanza said : * Sandombu, you have
slandered (ku-tamukita) m y wife, and I w ill give you a big case. Also
you have been heard by others to have threatened me. W hy do
you want to kill me ? Have I not given you m y daughter ? You
have a very bad liver to speak as you do to me. You must take your
wife back again, treat her well, and buy her a new cloth. I f you
want to speak in this w ay you must stay away from the chota o f
Mukanza. Sandombu refused to take back Zuliyana, and repeated
that she and her mother wanted to bewitch him. He went back to
his house and for a long time his voice could be heard roaring out
abuse against the people o f Mukanza, while those in the chota who
were not immediately involved shook with suppressed laughter,
especially the little boys. Zuliyana slept that night in the hut o f
Lines wife, Maria (I,Hi9), the latters husband being away.
N ext day Sandombu came to the chota and demanded die return
o f his wife. She came back with him but Sandombu was told to
return the day after to explain his conduct at a meeting o f village
eople. He did so and persisted in accusing his wife and mother-iniw o f witchcraft. In such intra-village meetings in Mukanza the
summing up (ku-sompa) was usually left to a blind member o f malabu
lineage, Chayangoma (I,H i5), who had lost his sight as the result o f
an industrial accident in a sawmill in Bulawayo. He had neither wife
nor child, and used to stress that he felt free to say what he liked as
he was a dead person *, a mufu. Since he was already dead no
one had any motive for bewitching him, and he could speak impartially.
In his summary o f this case he said that grave charges had been made
both by Sandombu and Mukanza, involving witchcraft. In the
eyes o f the Boma witchcraft accusations should be severely punished.
Therefore he thought it wise i f the affair were kept within the village
and not discussed by the local * law-men *, the mahakuv or vicinage
elders with a high reputation for advocating and judging cases, since
news might leak out to the Boma and their relatives would be arrested.
Sak&zao said that Sandombu was drunk at the time and could not be
held responsible for his words. I f he apologized and paid 1os. to
Nyamukola and a new cloth to Zuliyana nothing more need be said.
But Sandombu refused to pay. He said that he had been wronged
by his wife, who had first caused trouble in the house by starting an
argument and had then borne tales to her parents against her husband
to cause trouble in the village. He had accepted her back and yet
they still wanted him to pay. As for the accusations o f witchcraft,

16 0

Schism and Continuity in an African Society

both parties had made them in anger/ He would pay nothing to

Nyamukola, for he had been greatly provoked by both women.
It so happened that Sandombu himselfat this very time was bringing
an action for slander against the headman o f a neighbouring village,
Chibwakata (cf. p. 150). A t a beer-drink at which I was present the
day following the Zuliyana affair, headman. Chibwakata had accused
Sandombu o f having accepted 10s. from Kasamba, Chibwakatas
classificatory son by a slave mother, to kill the headman by sorcery.
Sandombu at first thought that Chibwakata, who was his own classi
ficatory grandfather, was joking with him, and laughed heartily.
But Chibwakata, who was drunk, became very angry and said that
Anderson Mulumbi, another classificatory son and Kasambas mother*s
sisters son, had been witness o f the transaction. Sandombu promptly
charged Chibwakata and Anderson Mulumbi with slander (mutamu)
and asked Chibwakata to settle with him on the names o f mahaku
who should discuss the case.
Thus Sandombu was in the odd position o f being defendant in one
slander case and plaintiff in another at the same time, both cases
involving accusations o f sorcery and witchcraft.
The Kasamba case was settled first. The mahaku were Nswanakudya,
headman o f a village o f the Kanongesha lineage, headman N g ombi,
and a Kawiku elder. Mukanza was invited, but he refused to discuss
any case in which Sandombu was involved. Gideon (II,E i), son o f
Mukanza, acted as Sandombus witness (chinsahu). Sandombu
described how on the previous Sunday a quarrel had broken out at
a beer drink between Kasamba and Anderson because the former had
obtained the job o f D .O .s cook for which both had been compet
ing, and Anderson had pushed Kasamba into the fire. Kasamba had
attacked Anderson and chased him away. N ow Anderson was trying
to get his revenge by causing bad feeling between Chibwakata and
Mukanzas people. I f he said he had seen Kasamba and Sandombu
conspiring to bewitch Chibwakata he had lied. Chibwakata, who
had made the accusation, and Anderson should pay damages to
Kasamba and himself. He demanded 5 from them. After much
argument the court said that Chibwakata ought to pay Sandombu
30s., and Kasamba the same amount. Anderson had to pay each o f
the plaintiffs 10s. In this case Sandombu identified himself with
Mukanza Village and with the exception o f Mukanza and Nyamukola
all the Mukanza people were pleased with, the decision. They had
old scores to pay o ff against both Anderson and Chibwakata. A t a
deeper level, too, the level o f the Ndembu-Kawiku cleavage, they
were pleased, for Chibwakata was the descendant o f the Ndembu warleader who had first attacked the autochthonous Kawiku. In addition,
as w e have said, he had been reared by Chokwe, the hated slave-

M atrilin eal Succession


raiders o f the nineteenth century, and in certain circumstances was

identified with the Chokwe.
N ext day the Zuliyana case was settled, still at the intra-village
level. Chayangoma (I,Hi5) was again asked to arbitrate. Sakazao
was suggested, but turned down by all parties who claimed that he
had no inherent faculty (chisemwa) for settling disputes. Besides,
he was own brother o f the co-plaintiff. Chayangoma said that
Sandombu should pay Nyamukola 10s. for defaming her character,
since the previous day he had obtained damages from Chibwakata for
slander against himself. In both cases there had been witnesses and
the fact o f slander had been clearly established. Sandombu paid up,
and people joked with him afterwards, saying that after all he had made
a profit from his two cases. Peace was restored for a time in the village.
It-w ill be recalled that I divided the. social -drama ..into four
m ainphases.: _breach,..crisis,operation o f xedressive-mechanisms,
and-either reintegration ofodto-SO cial-gm up-or.sooaLrec^n|tie|i^^e
o f irreparable schism. -. .Im plicit .in. .the- n o tio n -o f xeintcgradc^ f^ ^ ^
the__CQncept._o.^oqial. eauiH hrium ... This, concept
view that a social system is made up o f interrelated
persons and groups, whose interests are som ehow ^.maintained
in balance ; ..and further, ..that when_disturbance.occurs^ read
justm ents. are. m ade w hich, have the effectofrestorin g. the balance.
B ut it is necessary to rem em ber that after disturbance. has_occurred
and readjustments have, been made, there m ay have, taken place
profound m odifications in the internal relations.
T h e new equilibrium is seldom a replica o f the old. T h e interests
o f certain persons and groups m ay .ha.ve_..gained__at., the. expense
ofithose o f others. Certain relations betw een persons and groups
mayJ 3ave-mcreasedinintensity^hileo.thers.m ayihaye_.dimuushed.
Others.again m ay have, been com pletely, rupm red w h fieiiew rela
tionships have com e into being. A social .system_is.in_dynamic
m ovem ent-through space and. tim e,-in som e-w ay. analogous to
an organic system in that it exhibits grow th and^decay,. m fact
the .process o f metabolism. In one aspect,., the..spcid drama is
a process w h ich reveals reaHgnments .o f social relations .at critical
points o f structural m aturation,_it_m ay be
regarded as a trial o f .strength betw een . conflicting-interests in
which_persons or groups try_.to_. manipulate., -to.. their.._own


Schism and C ontinuity in an A frican Society

advantage the actually existing network o f social, relations,-both

structural and contingent, within the_system...Thus the social
drama, m ay represent either the natural, inherent .developm ent
o f a given, social system thro-Ugh -space-tiine-at a distinct phase,
at a_critical point o f maturation, .or. the.
some .o f its m em bers -to._accelerate.-or. retard. that_dexelapment.
It m a y be . either an index or a vehicIeL o f -change. In most
case^Jboth aspects are.-present. Thus in Social Dram a III, the
crisis w hich follow ed the death o f Kanyom bu (IJFp) indicated
that there had been a change in the internal residential structure
o f M ukanza V illage, in that malabu lineage had increased in
membership and strength. A t the same time, through N yam ukola (I,H io) and Nyakaiusa, slave-w ife o f the senior elder o f
MAtABU lineage, an attem pt was made b y that lineage to im pair
the claims o f Kasonda, then the most prom inent m em ber o f
nyachintang a lineage, to succeed, b y fastening on him a charge
o f sorcery. Kasondas new role as a man o f relative w ealth and
influence residing outside the village w ith Europeans was also
expressed in the charge ; in the course o f the drama he used this
new role to rebut the accusations, to m obilize n yach in tan g a
lineage, and to further his ow n interest in the succession.
la this social drama it may also be noted that the^socio-spatial
system form ed b y M ukanza Village, was .b y n.o_means_the-same
systein as that revealed in Social Dram a I at-the-tim e-of the death
o f K ahali Chandenda. Form er members h ad go n e, n ew mem
bers had com e into it. The conflict between Sandombu and
the. rest was in abeyance ; a new cleavage existed between the
form erly allied lineages o f malabu and. n yach intang a . O ld
members had form ed new.and.broken o ff old. attachments_T h e
basic principles governing residential affiliation remained, m atriIiny,jthe unity o f m ale matrilineal kin, virilocal marriage,, the
relariom Jb-Ctween genealogical genfirari_ons.,_the.jamty.ojLiterine
siblings, etc.__ .But the .relations.. into
daily, contact b y these principles_had_altered. A nd in. .the course
o fjh n e the status o f genealogic.afi^.lh^ed-..p.ers.ons h^d changed
asym m etrically. M ukanza was now._a_ headman.. ancLN yam ukola a headmans w ife. Sandom ,rival
clairiiant for office but M ukanzas jun ior, .under-_his._authority.
Kasonda w as n ow old enough to be apo.werfulclaimmtLin.uture,
and as the result o f external factors, had increased his internal

M atrilineal Succession


influence, in _the_ system . Sakazao w as older and had a large

foUowing w ithin. the villa g e.,.. As b<^ye<m. groups M A i^ y
age had-gained and nyachintang *a lineage had lost .member
ship,-but n ot office and influence, within the .village_.Social
Drama III gave expression and ratification to the n ew situation.
Social Dram a V brings to ligh t further. changes._within the
system, w hich_ h ad _ heeiitm obtxusive^
number^of years. It was a repetition in m any respects'of Social
Drama II, the bew itching o f N yam w aha. L ike D ram a V , that
had speedily resolved itself into a trial o f strength between
Sandombu and headman M ukanza Kabinda. B ut this tim e, in
Drama V , a clearly discernible im provem ent in the standing o f
Sandombu w ithin the system becom es visible. T his m ay in
part be due to the less serious nature o f the charge laid against
him b y M ukanza slander instead o f sorcery. B ut the case was
undoubtedly brought in order to emphasize the fact that Sakazao
and n ot Sandombu had been nom inated b y M ukanza as his
mulopu and successor. It was brought to signalize Sandombus
exclusion from the succession and to show that K ahali Farm was
subordinate to M ukanza V illage, and n ot the reverse. It was
perhaps unfortunate from M ukanzas point o f new that San
dom bu should have been involved at that precise m om ent in a
case in w hich he stood as the representative o f the M ukanza
group against their inveterate enem y, Anderson M ulum bi, and
against the successor o f the ancient foe o f the K aw iku, Headman
Chibw akata. Anderson M ulum bi, a farm head, a man o f wealth,
trader and tailor, had lon g been engaged in disputes w ith M ukanza
people, and indeed w ith other K aw iku villages. R ecen tly, he
had accused some boys from M ukanza V illage o f having wounded
his hunting dog and killed one o f his goats. H e had taken this
case to C h ie f Kanongeshas court and judgm ent had been given
against him . Kasonda was his principal foe, and econom ic
rivalry was in volved in their relationship, since both w ere tailors.
Kasonda was delighted to see him in trouble because C h ie f
Kanongesha had said that i f Anderson was m ixed up in any more
local disputes he w o u ld order him o u t o f the area. In this
situation, therefore, Sandom bu had the tacit support o f the
m ajority o f M ukanza people, and it w ould have been difficult
to penalize him severely in one situation and applaud him in
another on the fo llo w in g day.


Schism and Continuity in an A frican Society

B ut apart from these considerations, Sandombu had in other

respects im proved his position in the years that intervened be
tw een Social Dramas II and V . H e was now a farm head with
a follow ing. H e had obtained w ork for m any M ukanza men.
H e had been liberal in offering beer to the villagers and to many
others in the vicinage. There w ere even some w ho, like
M anyosas husband Chikasa,1 and his m others sisters daughters
daughter Bibiana (I,H4) in Ikelenge Area (just on the point o f
divorce and m other o f four children), w ould support his future
claims to village headmanship. Even i f he did not succeed in
M ukanza V illage, he m ight yet be headman o f a large farm and
resuscitate the title o f Kahali.
It is this heightened importance o f Sandombu that was re
sponsible for the w a y the case was played dow n and kept within
the village, for the small am ount o f the fine im posed, and for the
fact that Kasonda kept in the background instead o f openly
givin g M ukanza his support. A conciliatory tone was adopted
towards Sandombu throughout, no attem pt was m ade to retain
his w ife, and after the case was over it was jo k ed about.
A lthough Sandombu paid damages to N yam ukola he did not lose
face thereby. In the altered climate o f opinion, as he had just
scored a victo ry over Chibw akata and Anderson M ulum bi, it
was in Sandombus interest to make a gesture o f goodw ill
towards the village. O n balance he had considerably im proved
his position after these tw o cases, and M ukanzas refusal to have
anything to do w ith a case involving Sandombu had not added
to the old mans popularity w hen Sandombu had been disputing
w ith a m ajor enem y o f M ukanza V illage.
Nevertheless, b y die end o f Social Dram a V , there is reason
to predict that M ukanza V illage w ill split up along certain lines
after the headmans death. Kasonda told m e shortly before I
left that whereas he had once thought o f supporting Sakazaos
claim to headmanship and remaining w ith the latter, he, Kasonda,
had becom e unpopular w ith the people o f maxabu lineage and
they w ould continually quarrel w ith him . H e w ould start a
farm near a form er site o f M ukanza V illage, about a m ile from
1 Manyosa supported her brother Kasonda, but her husband, Chikasa, was
very friendly with Sandombu. These two, Manyosa and Chikasa, had married
late in life, after each had reared families with previous spouses ; and they
quarrelled incessantly over almost every possible issue.

A Sy m p t o m o f S o c i a l C o n f l i c t
Headman Mukanza chats with his wife Nyamukola during an interval in an
Ihamba ritual. An apprentice-doctor from Kafumbu Village (see Social Drama VI)
is applying a cupping-horn to her back by suction. This ritual, to remove the tooth
of a punitive dead hunter from the patients body by 'catching it in the cuppinghorn, immediately preceded the Cftihamba ritual described in Chapter Ten.,
Nyamukola, principal patient in both rituals, occupied a social position under severe
strain at this time. Note the forked branch with a piece o f ant-hill at the base set
up as a shrine to the hunters spirit. Nyamukola sits on a duiker skin. Traces of
pounded herbal medicine adhere to her skin.






M atrilin eal Succession


the m otor road, w ith M anyosas (I,G i3) children, w ith Chaw utongi (I,G i2) and some o f her children, and w ith G ideon (II,E i),
at once his m atrilateral and patrilateral cross-cousin, and his
children. H e him self had tw o w ives and five children, so that
he w ould have a fairly large fo llow in g. M anyosas husband
m ight g o and live w ith Sandom bu but there was a strong possi
bility that M anyosa w ould divorce him and com e and stay w ith
Kasonda. A gain , the husband m ight accom pany his w ife to
Kasondas form .
Sakazao w ould be left w ith m ost o f his lineage kin. Some
members o f both lineages m ight build at Sandombus form.
Sandombu, how ever, told m e that i f Sakazao succeeded, many
w ould leave him and live at Sandom bus form , for Sakazao had
a poor head fo r * speaking cases \ was idle and w ithout w ealth,
and no-one really respected him . Sakazao said to m e that i f
he succeeded, the village w ou ld not break up, because unlike
both Sandombu and Kasonda, he had never been accused o f
sorcery and no-one feared him in this respect, and because some
o f M ukanzas children belonged to m alabu lineage and others
to the lineage o f his kinsman Lines w ife (I,H i9). H e thought
that Kasonda, M anyosa and C haw utongi w ould rem ain w ith
him, and i f they stayed their children w ou ld stay also.
O f these appraisals that o f Kasondas is perhaps the most real
istic. Sakazao, for instance, cherished the illusion that Kasonda
w ould remain w ith him , whereas I was aw are o f Kasondas plans
to m ove, perhaps even before the death o f M ukanza. T he main
link between nyach in tan g a and m alabu lineages is the mar
riage between M ukanza and N yam ukola. Since they have
already becom e genealogically so distant, once that link is snapped
there w ill be little to hold them together. W hen M ukanza dies,
Kasonda w ill becom e the senior m ale o f chinbng a (I,JS4) lineage,
the num erically predom inant segment o f n yach intang a lineage.
I f he founds a farm m ost o f its members w ill fo llo w him and
not Sandombu. Basically then the division w ill probably follow
the lines o f lineage segm entation. Sandom bus lineage o f
NYAKAPAKATA (I,E i), w ith Bibiana as its forem ost fem ale, w ill
live separately from the other sub-lineage o f nyachintang a
lineage, that o f chineng a . m alabu lineage w ill probably remain
in the site o f the present M ukanza V illage, w hich w ill becom e.
Sakazao Village.

M atrilin eal Succession


From beneath the m anifold and transitory shifts and realign

ments o f social relations from year to year in die village, w e can
see the fundamental cleavages between maternal descent groups
emerge in the lon g run. M arriage and alternate generation alli
ance are means o f delaying the final residential split between
lineages, but as the w edge o f tim e drives deeper into their division,
so these lateral and supplem entary linkages wear thin and break.
B u t at the same tim e this division into com ponent lineages,
in another aspect, represents a division betw een genealogical
enerations. M ost o f the leading members o f n yach intang a
neage belong to the senior adjacent generation (G) to the
leading members o f m alabu lineage. N yam uw axiga o f malabu
lineage belongs to the senior generation (G) but she lives w ith
Sandombu o f her ow n generation, but o f a different lineagesegment. Chinem a (I,G zi), Sakazaos classificatory m others
brother, also o f m alabu lineage, w ho is w orkin g atK itw e (a m ining
township on the Rhodesian Copperbelt), has reported that
when he returns he w ill not live at Sakazaos village, fo r Sakazao
is his ju n ior. Sandombu claims that Chinem a w ill stay w ith
him. M ukanzas eldest son b y N yam ukola, Zachariah (IJ3),
a m em ber o f m alabu lineage and sisters son o f Sakazao, has
said that he w ill go w ith Kasonda, his cross-cousin, i.e. in his
ow n genealogical generation, reckoning through his fathers side,
and in the alternate generation, through his m others side.
C onflict between lineages, and. hetwcen_adjacent^gexiealogical
generations throw s into high relief the role of-aUiance between
alternate genealogical generations _as_a_m em sofm tegratm g the
residential unit. In a sense__the._village of^ M ukanza was held
and N yam ukola. Tension between adjacent_generatiom_and
alliancerJbetween alternate generations w ere related ta .th e junity
and continuity o f the local grpup. O nce the .m arriage-between
the headman .and. his_.wife, the senior w om an o f the opposite
lineage* cam e to an end,, lineage cleavage, - w hich in this case
coincided w ith generation cleavage, w ou ld assert itse lf and irre~
parable breach between lineages w ould takeplace. In Chapter
T en I w ill describe and analyse a m ajor ritual w hich, to a socio
logist, looked like a concerted attem pt to overcom e the widening
cleavages that w ere threatening to disrupt the unity o f M ukanza
V illage. In this ritual N yam ukola was the principal patient,









'\ y



Schism and C ontinuity in an A frican Society

afflicted by the spirit o f N yam ukanga (I,E5), a daughter o f

M alabu ; Sandombu was the principal male organizer ; and
Sakazaos senior w ife was the principal female organizer. The
ritual adepts w ere drawn from all the villages o f the vicinage.
B ut since this ritual appeared to have the additional social func
tion o f reconciling conflicts that had arisen between Mukanza
V illage and other villages w ith which it was connected b y ties
o f kinship, affinity and spatial contiguity, the ritual cannot be
discussed until a careful consideration o f the sources o f inter
village disputes has been made. In the next tw o chapters I
present material indicating under w hat conditions fission o f
villages actually takes place, and what kinds o f social grouping
are involved.

fission I mean a situation in w hich a group or
B Y_.lvillage
section w ith in . the .total membership o f a village_deaches
itselfand builds a separate settlement. JT_a^umber_ofYadiyiduals
leave_a village and severally affiliate themselves to. .diiferentJieaclm erLthisisnot fission__Sii33iladLy,if.aften.thedeatfi.oLa^headman
or as_the result o f some disaster,.aviUage >hceaks. up.,com pleteiy
and its., members scatter in attachment to.many^cHfferentv,headmen, this is fragm entation,.not-fission. L d efin e.fission am ong
the N d em buas the division, of. a tilla g e .com m unity,.along lines
o f structural, cleavage so th at one section. .maintains ^continuity,
usually sym bolized b y the retention o fits name. withJthe_ojiginal
undivided village ; and.the other section .or sections ^tiamed, after
theirdeading elders, seek to establish themselves as independent
villages. The lines o f im m inent fracture in M ukanza V illage
demarcated, as w e have seen, three segments o f a matrilineage,
nyachintang a (I,D i ) and m atabu (I,D2) lineages o f equal span,
and within, n yach intang a lineage, n yakapakata (I,E i ) and
chineng a lineages (I,E4), also o f equal span. In the case o f
nyachintang a and malabu lineages, genealogical distance from
a com m on ancestress had led to their social differentiation as
corporate units w ith certain independent social, econom ic and
ritual functions. Breach o f spatial relations o f interdependence
had been postponed b y m arital and alternate generation ties.
Betw een n yakapakata and chineng a lineages, the possibility
o f social severance had been accelerated b y the ambitions o f
Sandombu, the leading elder o f the form er, and b y social and
econom ic changes brought about b y the introduction o f cash
econom y w hich facilitated the em ergence o f increasingly sm aller
settlements inhabited b y ever narrow er ranges o f kin. B u t in
both instances fission seemed about to take place dom inantly
along the lines o f lineage cleavage. In m any other villages fo r
w hich I have records fission took place between segments o f a
local matrilineage. Nevertheless, fission does not_always._oxeven.
dom inantly take place _eq.ual_arder o f
segmentation. V ery frequently the dissident ,group., .is. a uterine

ly o

Schism and C ontinuity in an A frican Society

sibling group,, aiicl their ,mother,_i she .is. stilLalive.

The seceding section m ay not consist o f m atrilineal .kin .o f ..the
headman at all, but m ay be descendants o f a female slave, {ndungu)
o f ...a village member. These., under the. pax. bdtannica,Jtizvc
experienced an im provem ent in status and wish to-em ancipate
themselves in fact and. in .sp ace from , their._former_.owiiers.
Again, it may. happen -that a .village, is composite. in_memhership,
some o f its ...inhabitants being, matrilmeafly. descended-.from a
sister and some from a w ifb o f the first headman. T he m atrilineal descendants o f th e w ife m ay-split. o if from_those o f the
sister, and build, separately. O rthe sistersdescendantsmay
m ove. B ut in nearly ,all cases o f fission foe...prmciple-.Qfjnaternd
de&cmt_is ..crucial,..and..maternalj^ ,a lm o s t alw ayilform -jhe.core
o f a n e w . village.
Ndernbu and Tallensi Social Organization
In order to bring out the essential character o f N dem bu village
fission it is instructive to com pare briefly the main features o f
Ndem bu and Tallensi social organization.1 Fortes w rites :
Tale society is built up round the lineage system. It is no exag
geration to say that every sociological problem presented by the
Tallensi hinges on the lineage system. It is the skeleton o f their social
structure, the bony framework which shapes their body p o litic;
it guides their economic life and moulds their ritual ideas and values.
[The Tale lineage is] a striedy unilineal, agnatic descent group,
[and] eight to ten ascendant generations are usually reckoned between
contemporary minimal lineages and the founding ancestor o f the
maximal lineage o f which they are part (p. 31).
[All Tale lineages] are hierarchically organized between the limits
o f the minimum span, i.e. the minimal lineage, on the one hand, and
the maximum span, i.e. the maximal lineage, on the other. Thus
every minimal lineage is a segment o f a more inclusive lineage defined
by reference to a common grandfather, and. this, in turn, is a segment
o f a still wider lineage defined by reference to a common great
grandfather ; and so on, until the limit is reached the maximal
lineage, defined by reference to the remotest agnatic ancestor o f the
group.. . . W ithin a lineage o f whatever span each grade o f segmenta1 Fortes, M., T h e Dynam ics o f C lanship among the T allensi (1945), p. 30.

Village Fission, Slavery and Social Change


don is foncdonaDy significant. Bach segment has its focus o f unity,

: and an index o f its corporate identity, in the ancestor by reference to
i whom it is differentiated from other segments o f the same order in
the hierarchically organized set o f lineages. Sacrifices to the shrine
: o f this ancestor require the presence o f representatives o f every
segment o f the next lower order ; and this rule applies to all corporate
acdon o f a ceremonial or jural kind o f any lineage---A lineage o f
any span emerges in any o f its activities as a system o f aliquot parts,
not as a mere collection o f individuals o f common ancestry (p. 31).
Tallensi do n ot live in discrete com pact villages but in hom e
steads, containing on average some 12 to 14 people, scattered
apparently indiscrim inately, som e d ose together, others farther
apart, over the country, in irregular blocks w h ich m ake up
* settlements
There is a distinct though sometimes narrow
territorial separation betw een the outerm ost homesteads o f
one settlement and those o f another. A ccording to Professor
stability and continuity are essential characteristics o f a settlement.
They are implied in the native concept o f a teng (settlement) as both
a definite locality and a fixed community. The ancient settlements
have attained a very high degree o f stability and continuity. . . .
Precise local orientation is essential in the economic activities and soda!
relationships o f the Tallensi (p. 157).
* A settlem ent is a m iniature o f the w h ole society, and reveals
all the basic principles o f the social structure9 (p. 154). A
settlement is a locality defined in relation to a corporate unit o f
social structure, a lineage, d an , o r inter-connected group o f dans.
T ale society, then, is built up round the hierarchically organized
lineage system , and lineages o f considerable depth are anchored
to specific localities. Each segm ent o f a single lineage o f w ide
tends to form a coherent residential cluster, situated on land which
has been owned by its members for several generations, and in close
proximity to the graves o f their ancestors. The lineage as a whole
tends to form a local aggregate, though a somewhat looser one than
any o f its segments (p. 197).
A ccording to Professor Fortes, * neighbourhood ties are ipso





Schism and Continuity in an A frican Society

facto lineage ties, and therefore econom ic, religious, jural and
moral ties 7 (p. 211). Lineage proxim ity is correlated w ith local
proxim ity.
This intense local fixation o f the lineage is no doubt congruent
w ith 4the rigorous precision w ith w hich its members fall into
a set o f hierarchically articulated segments, o f hom ologous form
but all exactly differentiated both genealogically and functionally
(p. 19s).

andJocal. ties.... Local, stability.and. continuity is-typical. o_Tale,
and_sp.atial_mobility is typical o f Ndembu-residentiaJLoxganization.
Lineages are anchored to specific tracts-.of.-land-, among-, the
Tallensi, and have deep genealogies ; and -the. ..fission., o f one
homestead results in no m ore .than the estabMshment-ofLan adja
cent homestead. A m ong Ndem bu an articulated lineage has little
depth (see Tables XIII, X IV . and X V , p.. 8.3), and-its.^fFective
structure is confined within_a_-smgle_village. Fission is often
accompanied b y violent internecine conflict o f the kind des
cribed in Chapters Four and Five, and results in the seceding
group putting m any miles between itself and the parent village.
A n N dem bu vicinage, w hich m ay perhaps be com pared w ith a
Tale 4 settlement \ is not a locality defined in relation to a cor
porate unit o f social structure, a lineage, clan, or interconnected
group o f clans \ but is extrem ely heterogeneous in character, so
that in certain vicinages none o f the com ponent villages m ay
be m atrilineally interlinked. Cer.tain__v.Ulages^ .often. ..dispersed
throughout., the entire Ndembu..region,^may. claim relationship
through putative maternal descent, but they do_not.trace..descent
to ajnamed com m on. ancestress and caim in a
hierarchically organized maximal. lineage_with_functionally signiIc4 n t .grades _of segmentation. Such..a.. .*.lineage.!.,.is_not 4a
system o f aliquot parts \ but * a mere collection o f .[small groups
o f] individuals o f com m on ancestry \
The contrast between Tale and Ndem bu lineage organization
is strikingly displayed in the ancestor cult. A m ong the Tallensi
ancestors have permanent shrines and graves, w hich are attached
to specific localities, w hich are usually surrounded b y the residen
tial and farm land held b y a maximal lineage, and w hich therefore
serve as foci o f the unity o f the lineage as a w hole and o f its
segments : hence, as Professor Fortes succinctly expresses it, 4 the

V illage F issio n , Slavery and So cial C h an ge


ancestor cult is the calculus o f the lineage system. B u t am ong

the Ndem bu, ancestor shrines consist o f quickset muyombu
saplings, planted to placate the spirits o f deceased village
members, w ho are still remembered as livin g persons b y their
kin. These spirits m ust be placated because they have returned
to afflict individual relatives w ith misfortunes, as the result o f
neglect * to m ention their names * and pour them out beer when
praying to other spirits, or because conflict has arisen between
village members. Such shrines are abandoned w hen the village
moves to a new site and it is very lik ely that n ew muyombu
saplings w ill never be planted for the same spirits again. R em ote
ancestors are seldom worshipped except in a few im portant and
infrequent rituals, such as the boys circum cision ritual {Mukcmda)
or the great ritual o f Chihamba ; and in both these rituals the
unity o f all N dem bu, even o f all Lunda, is emphasized rather than
the unity o f the particular village w hich sponsors the ritual.
Graves are never used as shrines b y N dem bu and are feared and
avoided rather than tended and used as sites o f sacrifice. For the
Tallensi the ancestor cult is related to the land, to agriculture, and
to permanent residence on the land o f w ell-defined corporate
lineages. Fi5r..N dem bu, the ancestor. cult_is_ass.acia.ted.with_the
busludts-dangers-and _hlessings._with. the transience o f settlement,
withjthe. hazards o f life , and .w ith the m obile human- group itself
rather.J th a n .itssp e cific habitation. It hasl been_.suggeste.d in
earlier.xhapters that the...high, m obility of-Ndemb.ur aticLthe small
size o f settlements and their frangible character,...are. congruent
w ith the high value set on hunting^ eans o f
providing an appetizing and nourishing...faad,_hut._also_ as an
index_of m asculinity and as a historically, validated.m eans o f
acquiring prestige. T he size and m obility of-settlem ents seem
to be m ore closely related to the local availability o f gam e and
sylvan .resources than to the prod uctivity. o f xhe soil.
N dem bu supernatural.beHefk regarding thed ead w arkjtow ard s
the prom otion o f m obility rather.than .tpwards the creation o f
deep attachments to the sites w here dead ancestors He buried.
In die .past, when a person died, he o r she. was. buried. under the
floor o f Jthe hut, and the village changed its site for fear o f the
ghost. Beliefs regarding the dead am ong. Ndem bu, then, accelerated~the already h igh rate .o f m obility. O n the other hand, the
ancestor cult o f the Tallensi tends to emphasize and validate the


Schism and Continuity in an African Society

intense attachment o f lineages to localities. W orship o f the

patrilineal ancestors and worship o f the Earth are the deepest values
o f Tale religion. A lthough in some respects they are opposed,
since the ancestor cult tends to emphasize the divisions between
lineage and clan segments w hile the. Earth cult tends to stress the
com m on interests o f the widest Tale com m unity, the tw o cults
are at the same tim e com plem entary, since the cultivations o f
individuals and o f particular lineages are, as Professor Fortes puts
it, * hut parcels o f land cut from the limitless earth *. Tallensi
are attached to the soil and, given the existence o f patriliny as
their dominant principle o f social organization, local and lineal
groupings tend to coincide, and to be interlinked w ithin a com
m on fram ew ork o f overlapping clanship and lineage ties that
embrace the entire
Ndem bu m ove over t
game, and given m atriliny as their dominant m ode o f social
organization, lineal kin tend to scatter across the region, local
lineages tend to be sm all and unstable, and no deep interlocking
local lineages and clans develop. T alc ritual expresses the unity
or corporate groups from the unity o f minor matrilineages to the
unity o f all the Tallensi, w hile in the great all-inclusive rituals
the basic units in the com position o f ritual assemblies are patrilineages. In N dem bu ritual, as w e shall see, the basic ritual units
are cult associations that cross-cut the relatively transient local
and lineal affiliations. The dominant links between, participants
in Tale ritual are links o f commonness and corporate identity ;
in N dem bu ritual they are links o f likeness and tem porary
association, lateral rather than lineal.
W ith this general background o f comparison in mind, let us
now consider the different modes o f fission in the residential
structure o f these tw o societies. Am ong Tale fission does not
represent a com plete break between the groups involved.
Rather, it represents a process o f regular branching-offby means
o f w hich the form ation o f new local units advances and projects
the social structure through time. T he new group does not
emancipate itself from the lineage ; it becomes a branch o f that
lineage. A cleavage occurs between tw o hitherto nascent seg
ments o f a m inor lineage inhabiting one homestead ; and w hile
this m ay give rise to the establishment o f a new homestead,
ties o f ju ral and ritual co-operation between the tw o groups

Village Fission, Slavery and Social Change


prevent radical breach and contain the cleavage w ithin the

fram ework o f the m axim al lineage.
But-am ong N dem bu, the cleavage_of a . village^-.W-hether or
not along lineage lines, results in the. form ation o f a com pletely
newjunit. A hiatus is made in the genealogical ,structure,of the
original m atrilineage in_such a w a y that, after the^passage o f a
single generation, no member o f either village distinctly, recalls
the precise.-point.of fission. A fter tw o or three generationsall
that is left is a vague sense o f com m on maternal...descent_-Yet
it often happens that, w hile the name o f the com m on ancestress
and the relationship o f the founding ancestor o f the new village
to members o f the old village are forgotten, members...of both
groups continue to apply to one another kinship ,terms .according
to-the.nom enclature, w hich expresses separation or. equivalence
in their'genealogical generations. Thus the present Rhodesian
C h ief Kanongesha N dem bi calls headman N yaluhana his older
brother (yaya), headman N swanakudya ms * older brother *,
headman Shika his younger brother (;mwanyika), and headman
Kajinga his 4 m others brother * (mandumi), although the exact
lineal relationship between these m en can no longer be traced and
all inhabit diffrent villages w id ely separated in space. T he facts
o f matrilineal kinship and the relevant genealogical generation
category are recognized, but a m ore precise identification cannot
be made. This system o f nom enclature m ust be distinguished
from perpetual kinship since the terms are applied between indi
viduals and do not express the relationship between political
positions in the social structure.
The generation principle governs the authority and marital
structure o f a settlement. It is im portant for members o f a
society w hich possesses a high rate o f individual m obility, in
w hich persons are frequently changing their village affiliation,
to kn ow in w hat relation they stand vis--vis the headman o f a
new village in terms o f genealogical generation. O nce this
relationship has been established a new m em ber o f a village
know s im m ediately w hom to respect and from w hom to le vy
respect, he know s w hich w om en he can m arry and w hich he
m ay not, he know s in w hat arc o f the village circle he m ay build
his n ew hut, and so forth. T h e stress laid b y N dem bu on genea
logical generation as a principle o f local organization and as a
means o f facilitating inter-village m obility, m ay be regarded as


( '}








( '

( :

.( "




! . .



Schism and Continuity in an African Society

an index o f the brittleness o f lineage structure, it m ay also be

regarded as a means o f organizing local units whose membership
contains a high proportion o f seminal children o f the male m atrilineal core, as w ell as a number o f cognatic kin. and slaves. In
Chapter E ight I w ill discuss conflicts that arise in villages on the
basis o f their m arkedly bilateral com position, and the role o f
genealogical generation organization as a mechanism for con
taining such conflicts w ithin bounds.
T he point I w ish to stress in this chapter is that. although at
theJvel o f lineage organization structural amnesia, w ith ..regard
to precise genealogical links , between villages, separated by
fissionlof a .localJiiieage speedily .occjuxs^_m.emor.y,is .retained o f
common, m aternal descent and .o f the .affiliation b y .-genealogical
generation between, the detached, groups. Separation has been
effected between tw o sub-groups.._within _one-s.ys.tem_oX_social
relations, that o f the village, but several .sets o f ties w hich form erly
interlinked these groups as members o f a single-com m unity con
tinue to operate between them. Each group. is_now a unit
w ithin a w ider system o f social relations,., and.w ithin-that w ider
system each maintains w ith the other-tieso f. a._moce. durable
nature than m erely those o f jo in t .membership._Q.a-_chiefdom,
spatial propinquity, or affinity. After, the feelingso.anim osity
associated w ith the. initial
has a
special claim on the hospitality. o the.,0 ther,,_ the. .members o f
both exchange long visits, and each.. may..setve_.in_tum as the
base.othe others hunting expeditions. It is thought appropriate
i f cross-cousin and grandparent-grandchild (cf. pages 80, 246)
marriage takes place between them .1 It-is_an...advantage to
merhbers o f a . highly.-nio.bile society. .tO-.he- ahle_to_jab.tain, hos
pitality, from distant matrilineal..kin in ..remote..areas,through
w hich they m ay wish to pass-.on.theiii-travels. Eission .extends
the geographical range o f each village!s ties_o.durable kinship,
ancLeach component, village in .a vicinage, -througbu.the frequent
fission, o f,. is. .interlinkedby. ..tiesof-com m om maternal
descent..with-_many . viUages._in_othervicinages. A m ong the
Ndernbu, therefore, the mesh to
gether in a single system distant and.discretejvullag-es-and vicinages
1 See, for example, Appendix III for marriages
Village and its offshoots.



V illage Fission , Slavery and Social C hange


and-to.p.CEmit.of..awidaraiig.a,Qiiijcliyidual^iabiIity: Am ong
the Taliensi, the lincagc-principlc. operates.. to...flx atidxonsolidatc
lineaLkirt in specific compact blocks o f land.
But it Tmust not be thought that Ndem bu consciously and
purposefully utilize the fission and dispersal o f villages to prom ote
the closcErjcohesioii o f the w ider social-system,.,the...tribe, or. tribal
section. - Independently o f their wishes. in..the..matter-it_may w ell
be that:.centrifugal tendencies dependent on .their .ecological
system-Constantly break up villages and change...the.. composition
o f vicinages. It is m ore likely that .the cohesion. .of,the. wider
system is maintained despite these tendencies, and.that Ndembu, in
fact, make a virtue o f necessity. .Fission is usually...regarded by
the people themselves as something deplorable and .they, take
every possible measure to avert it. InJtheir conscious behaviour
and judgments Ndem bu act and. spea.k_as _though_thc breakrup
o f a village or its departure from a vicinage..was a.calamity, and
the initiators o f fission are always condemned, n o t only by
memb.ers^of the village from which they led away a-section but
by members o f other vilIag.es_ inL_thc .vicinage. Again, if one
asks the headman o f a dissident, settlement, w h y he-broke. away
from the original village, he w ill most likely answer, that it was
an unfortunate necessity. It was unfortunate ,that, the village
should have split, he w ill say, but the headman .or on e.o f his close
relatives was a quarrelsome person or-a sorcerer,-or-theheadman
foiled to.Jbe impartial in his judgments, favouring his ow nprim ary
kin in disputes w ith the seceding group, and so ..forth... ..Both
groups, those w ho go and those w h o remain, w ill, in fact, stress
the value o f village stability and continuity, the original group
emphasizing the wickedness or fo lly o f the secessionists, and the
latter the exceptional circumstances which compelled_them to
But after a. lapse o f time, and especially after the deaths o f the
leading _<hsputants, _the tw o _viUages
relations, in the. w ay described .above. T h e conflict has -been
absorbedL/hy the w ider .social, system,..and .what.was-origm ally a
division between sections o f its membership is gradually filled
up by new sets o f ties w hich reunite, them at..a_.different level
o f organization. U ltim ately, a realistic ..attitude -towardsfission
may be taken,, and the members o f the original group m ay admit


Schism and Continuity in an A frican Society

tha_trthe_headman o f the seceding, group.was an eldec .with many

close-kin and had little chance o f succeeding to office.in_an estab
lished village. H o w this process illustrated
by. a social, drama. The first o f these is based on accounts given
to m e by members o f all three o f the groups involved. These
groups are Mukanza V illage and its offshoots Kafumbu and
Yim hwendi Villages. I use material from Mukanza once more
on account o f its plentiful detail, because it provides an interesting
comparison w ith the situations analysed in the tw o previous
chapters, and because it illustrates tw o m ajor types o f village
fission. I hope also that readers have begun to be acquainted
with some o f its inhabitants.
Yimbwendi and Kafum bu Secede from M u kan za
(compiled fro m informants)

In the period 1924-28 Mukanza Village passed through a crisis

which almost resulted in its disintegration. A t that time there were
three major factions in the village. The cleavage between
nyachintanga and malabu lineages was then only incipient, and
both these lineages were united as nyachipendi (I,C2) lineage and
as such were opposed to NYACHULA lineage (LD3, see Appendix I), o f
the same genealogical depth and o f equivalent span The common
ancestress o f all three lineages was Chipendipendi (I, A i), mothers
mother o f Kabonzu(I,Ci), the founder o f the village, but two inter
vening ancestresses between Chipendipendi and Nyachula had become
forgotten, nyachula lineage had returned to Mukanza Village .after
many years residence in Chibwika area. In the village also was a
large group o f hereditary slaves o f Kahali Webala (LE3), descended
from a woman, Mpeza (I,F i), who had many daughters by two
brothers o f nyachintang a lineage. The eldest daughter was
married to her cross-cousin Kafumbu (I,G i), also a slave belonging
to nyachintanga lineage. Another daughter was married to
Yimbwendi (I,F i i ), senior elder o f nyachula lineage.
The village was a large one but full o f constant bickering. One
o f the main reasons for this state o f tension as given me by the present
inhabitants o f Mukanza, was that at this time there was no recognized
headman. Kahali Chandenda (I,F5) had not yet settled permanently
in Mukanza but spent his time alternating between his kin in Ikelenge
Area and Mukanza Village. Mukanza Kabinda (I,F8) had inherited
the name Mukanza from his deceased elder brother Mukanza











18 0



i :



S chism and C o n tin u ity in an A frican S ociety

Kandulu (I,F6) as an indication to all that he hoped to succeed to the

headmanship. Yimbwendi (I,F i i ) and Kafumbu (I,G i) formed an
alliance with their respective followings, and Yimbwendi strongly
urged his own claims to office. When Kandulu died in 1919, Kafumbu
had been suspected o f killing him by shooting his ilomba, in the same
way as Sandombu had later been suspected o f shooting Kahali
Chandendas ilomba (see Social Drama I, p. 95). Kasonda told me
that Kafumbu had called in a sorcerer from Angola, named Mapachi,
to assist him in this deed, and that Mukanza Kabinda had consulted
a great diviner with a divining-basket who had named Kafumbu
as the sorcerer. I suspect that this tale is a myth retrospectively
attributing evil qualities to Kafumbu in view o f his subsequent de
fection from Mukanza Village. Although he was a slave (ndungu),
Kafumbu was a powerful man in Mukanza Village. He was a famous
hunter, a skilled orator, and had a large following in the children o f
Mpeza (I,F i). He claimed that slavery (wudungu) had been abolished
by the European Administration and that Kahali Chandenda and
Mukanza Kabinda no longer had any rights over the services and
allegiance o f their former slaves. He was constantly urging Yimb
wendi (I,F i i ), whose two sisters had many adult children, to succeed
to the headmanship. According to Mukanza Kabinda he knew that
if Yimbwendi formally succeeded Mukanza Kandulu, the people o f
nyachintanga and malabu lineages would leave the village and
found a new settlement, and the group o f slaves would be rid o f their
In n y a c h in t a n g a lineage at this period there were a number o f
adult men who, on the whole, supported the claims o f Kahali
Chandenda (I,F5) and desired his permanent return to the village.
These included the famous hunter Lupinda (I,G7), whose spirit
Sandombu was later alleged to have invoked to bewitch Nyamwaha ;
Tapalu (I,G8), later to become a wealthy African trader in Livingstone ;
and Kachambili (I,G9), Sandombus brother who fled to Broken Hill
in 1928 to escape from the sorcery o f Sandombu, according to some
accounts, after a quarrel, and who has remained there ever since.
Lupinda did not want to succeed, for he preferred the wild unanchored life o f a hunter. In addition, he was feared as a sorcerer
by the villagers.
In 1928 the simmering tensions in the village were brought to a
head when Kahali Chandenda decided to take up permanent residence
in Mukanza Village. This decision coincided with the decision o f
a special gathering o f village elders to change the site o f the village
to the southern edge o f the Kawiku Plain where the hunters o f the
village would be nearer to their major hunting ground. When
Kahali announced his intention, Yimbwendi said that he could no

V illage F ission , Slav ery and So cial C hange


longer remain in Mukanza Village as it was 4full o f troubles \ * people

were not living well together*, and so on. (In fact the change o f
site gave him a good opportunity to start a village o f his own.) He
was now, he said, a man with many children and sisters children.
He was an important elder (mukulumpi). He had been the headman
o f a settlement before coming to Mukanza. The time had come to
divide off (ku-baluka). He pointed to the fact that Kanyombu,
younger brother o f Mukanza, had married his (Yimbwendis) own
sisters daughter Nyamalata (I,G23), who was Kanyombu's classificatory sisters daughter, without performing the ritual to cleanse
an incestuous union (kudisola), as an indication that the lineages o f
nyachintanga and nyachula were no longer closely related.
He then asked Kafumbu to come with him as they had both married
into the one lineage, that o f mpeza (I,Fx). Kafumbu agreed and said
that he was weary o f the quarrelling in Mukanza Village. Kanyombu
and Lupinda became very angry at this and said that, according to
Lunda custom, Kafumbu and Mpezas descendants belonged to them
and had no right to dispose o f their own future. Kafumbu said
that a new law introduced by the British South Africa Company had
*killed * the old custom and slavery had ceased to be. After some
hot exchanges, fighting broke out in the village in the course o f winch
Kachambili (i,Gs>) and Yimbwendis uterine nephew Samunuma
(I,Ga2) were badly beaten up. Yimbwendi and Kafumbu declared
that the fight had clearly demonstrated that there could be no peace
between the followers o f Kahali and Mukanza, and the rest, and led
their followers away to the far side o f the Kakula river about seven
miles from Mukanza Village, and nearer to the Boma. Possibly
Kafumbu, who had chosen this site, thought he would be safer near the
Boma i f the Mukanza people tried to begin a feud.
Not ail the followers o f Yimbwendi and Kafumbu accompanied
them. Nyawunyumbi (I,G25), uterine niece o f Yimbwendi, remained
at Mukanza Village, for the reason, as she told me, that Yimbwendi
was a sorcerer, with a bad liver , and she had three daughters and
a son whom she wished to protect from sorcery. Mpezas son,
Kayineha (I,G<5), also remained. He was married to Kasondas
sister Mwendiana (I,Gi4), daughter o f Nyamwaha (I,F7), was
personally friendly with Mukanza Kabinda, and in fact played a major
part in bringing the fighting to an end. Nyamalata (I,G23) remained
with her husband Kanyombu (I,F9). Line (I,H i 7), on the other
hand, went with Kafumbu, although he belonged to Malabo, because
he was the cross-cousin o f Yimbwendi, and one o f Kafumbus sons
was his personal friend. Later he married Nyawunyumbis daughter
(I,Hi9) and returned with her to Mukanza. Thus, although three
major groups were clearly involved in the cleavage, several individuals


Schism and C ontinuity in an African Society

deserted from their own kinship factions and changed their allegiance.
Such individuals, although acting apparently in accordance with their
personal desires and interests, have an important role in reuniting the
initially hostile factions at a later date. Each assumes an intercalary
role in a wider social system, the units o f which are villages and
After the departure o f Yimbwendi and Kafumbu the District
Commissioner o f that period wanted to combine Mukanza Village
with Chibwakata and Nswanakudya Villages since the village had
lost more than half its population. But an appeal was sent by Kahali
(I,F5) to Ikelenge Area, to Sandombu and others living there to return
and save the independence o f the village. A number o f them, in
cluding Sandombu, and their wives returned and the dreaded amal
gamation was avoided.
In 1929, after one year, Kafumbu and Yimbwendi fell out over the
headmanship o f the new; village ; there was a fight between their fol
lowers, and Yimbwendi left Kafumbu with his children and sisters'
children to settle in Chief Chibwikas area, far to the south o f the
District on the edge o f Mbwela country, where Kabonzu had once
made a village about a century before, and where Yimbwendi had
once had his own village, although in a different vicinage. When
he died a few years ago his village split up once more after a quarrel
between Samunuma (I,G22), son o f his oldest sister, and Makumela
(I,G24), son o f a younger sister. Makumela, who had been nominated
mulopu or heir by Yimbwendi, was a younger man than Samunuma,
who hived off and made a farm o f his own with his children and
sisters children. When Kafumbu died, his deceased sisters son
(I,Hi) claimed the headmanship but he was rejected by the over
whelming majority o f villagers who chose Kafumbus son, Samlozanga (I,H2), as headman. Samlozanga was the senior elder of
m peza lineage to which most o f the villagers belonged ; he had
another claim in that he had 13 children, some o f them adults.
Samlozanga was supported by Mukanza himself, by this time reconciled
with the village through nis friend Kayineha (I,G6). Kayineha,
although mothers uterine brother o f Samlozanga (I,H2), did not
succeed because he was working at the Copperbelt when Kafumbu
died. Kafumbus uterine nephew (I,Hi) made a farm about half a
mile from Kafumbu Village where he lives alone with his wives and
children. His farm is called derisively by Mukanza and Kafumbu
people * Kafumbu Kamu , meaning, * the solitary Kafumbu , while
Samlozangas village is called Kafumbu Kevulu , * Kafumbu o f
many people *.

Village Fission, Slavery and Social Change


Commentary and Analysis

W hen a long-established village splxts, it often, happens that
the... seceding sections are .highly, unstable. T h e y ...lack-...longestablished bonds o f discrete .unity.-_and._they. often.break.. into
small setdements- o f uterine siblings and their families,..each under
the leadership., o f the., oldest , brother._.Gnce__the_strang bonds
which hold together the members _of a traditional, village .have
been broken,. the powerful .individualism o f the. JbJdemhu ethos,
nurtured-by. the. selfrreliant.liie.of. the. hunter,, asserts,, itself for a
time and-the.dissident group -breaks.-up-into. its. basic-units, the
remnants , o f what were once .matricentric-Jfamilies... But the
original village ...slowly .and painfully., .buildsup. _its_jnembership again. N dem bu take, great pride, in .belonging_tOL.a. village
with a famous historical name, and although- such -villages are,
like others, liable to fission, it is extremely__rare_.for__them to
perish utterly. Villages founded b y . the. twelve .headmen who,
according to tradition, originally accompanied-Kanongesha, are
still inexistence,.most..o_them large.and-thri-ving,-as-indeed are
villages., founded b y the .early chiefs-for-their-close relatives.
The ancient K aw iku villages too, such as Nsanganyi, Mukanza,
Kasai and Nyachiu, have persisted for many years in spite o f
frequent fission. The.-headmen o f. such ..villages, enjoy greater
respect than the headmen, o f newly-form ed villages and farms,
ana i f there is a gathering o f local headmen at a. ritual or at the
hearing o f a case they are always given beer or food before the
newer headmen.
A few words must be said about the w ay in. which this irre
parable breach o f intra-village relations has been recognized b y
the parties concerned and how new social relations have been
established between the villages. The people o f Mukanza
Village say that both Yim bw endi and Kafumbu were sorcerers,
and every death that subsequently occurred in their villages was
laid at their door as p ro o f o f this statement. Kafum bus slave
status, and that o f the mpeza lineage is also mentioned, and
blame is attributed to the Europeans w ho caused skvery to be
abolished, thus allow ing upstarts like Kafumbu to break away
from their rightful owners (ankaka). T h ey also allege that
Yim bw endi was a quarrelsome man, not the kind to inspire
respect or confidence in his judgment. Yim bwendi and


Schism and Continuity in an A frican Society

Kafumbu people, on the other hand, assert that Mukanza Village

contained many sorcerers and witches, mentioning Kahali, Lupinda, N yam uw anga and Sandombuin this connection. Mukanza
Kabinda was never considered a devotee o f the black art, even
by his enemies, but the dissident group say that at the time o f the
split he was a poor speaker in village lawsuits and was not suffici
ently firm in his decisions to give a village good guidance. Kahali
Chandenda was not acceptable to them as headman on account
o f his roaming propensities (ku-kimboka, * to move about from
one place to another), his irritability, and his suspected sorcery.
B ut although there appears to have taken place a complete
rupture o f social relations between the leading actors in this
social drama, Mukanza Kabinda, Kanyom bu, Kafumbu and
Yim bwendi, this did not occur between their several groups o f
followers, many o f whom were interlinked b y complex ties o f
kinship and affinity. Thus Sakazaos senior wife, Nyaluwema
(I,G4), was the youngest daughter o f Mpeza (I,Fi) by Chisela
M alwa (I,Fa), brother o f the fourth headman, and she was also
sister o f Kafumbus wife Katendi (I,Gz), so that Sakazao, accord
ing to Ndembu kinship nomenclature, called Kafumbu older
brother (yayrt). Chinema (I,Gai), at that time headman o f
the village in Ikelenge Area abandoned by Kahali Chandenda,
was the husband o f Nyaluwemas and Katendis other sister by
the same father, Kawila (LG3), and, it w ill be remembered,
Yim bw endi was the husband o f their other sister, Lusenga
Mulusa (I,G5), by a different father.
Eventually die children o f all these men, w ho called one
another older and younger brother would form the residential
core o f the new unit founded by Kafumbu. This is probably the
reason w h y Sakazao took no part in the fight which led to the
secession o f Kafumbu and Yim bwendi from Mukanza. Kayineha
(I,G6) o f the mpeza lineage was, it w ill be recalled, married to
M ukanzas sisters daughter, and was Mukanzas close friend, in
spite o f being his classificatory son-in-law (muku, a term em
ployed not only b y a womans father, but also by her mothers
brother, towards her husband). Ifciis._.nat. frequently.^enough
stressed by anthropologists that village kin,., whatever, their
genealogical relationship, .tend to develop, newuelationships o f
friendship or hostility based on their constant interaction and on
temperamental affinity or..animositywhich.jn..jthe_rcgular flow

f.Gi. K A F U M B U [,G2. KATE N D I

.G - K A W IL A ---I,C4- N Y A L U W E M A -

l t.Fi. M PEZA



I.F?. N Y A M W A H A O I.G14 - M W E N D IA N A -----

----------- - &i ,G 6. N y a m

............... -=A I.F11 . YIM B W E N D I-



n c a

I.H. S A K A Z A O -


Schism and Continuity in an A frican Society

o f dh^ .which cancels out in

many^situations the institutionalized^^
kinships rules. T ims' Mukanza and K.ayiiieha._shQuId .have par a relationship which
respect on the latter in his dealings w ith .the .fbrrner^
gave the
former peremptory authority, over. the. lat.ter,,..Butin practice the
two men. were, friends, visiting Angola together, where they
obtained a new variety o f cassava which gave rich yields and
which they introduced into Mwinilunga. They- slept-side by
sideJh-temporary camps when custom decreed that they should
never .:oeeupy.-die_.same._dwelling. In this case the fact that
Mukanza and Kayineha were more or less contemporaries, and
were brought up as adolescents in the same village, probably
helped to mitigate the rigours o f genealogical and affinal regula
tions. Kayineha did not leave Mukanza with the others but
remained until the death o f his wife several years later. He
finally departed from Mukanza on the best o f terms, and his
friendship w ith that village probably had much to do with the
rc-animation o f tics between the two settlements. Line (I,Hi7),
on the other hand, went with Kafumbu and Yim bwendi and
eventually married the daughter o f N yawunyum bi (I,G25),
Yinibw endis sister, a woman called Maria (I,Hi9) with whom
he eventually returned to Mukanza. N ya w u n yu m b i1 has re
mained in Mukanza Village, and after a few years married the
headman o f a nearby village, Mbimbi, son o f a male member o f
malabu lineage in Mukanza and her classificatory cross-cousin.
The children o f N yawunyum bi played an important role in inter
linking the various sections within Mukanza Village, as well as
in the rc-cstabJishmcnt o f relations w ith Yim bw endis new
village. In addition to die daughter Maria (I,Hi9) w ho married
Line o f malabu lineage, her oldest daughter, Seliya (I,H i 8),
married Mukanza himself, o f nyachintang a lineage ; and her
son Kenson (I,Hao) was brought up in Mukanza Village where
he shared a bachelor hut with Pearson (I,Js), son o f Nyatioli and
sisters son o f Sakazao o f malabu lineage, and with Daudson
(I,H7), son o f Manyosa o f nyachintang a lineage. Kenson is
now affianced to his classificatory cross-cousin Mwendiana
(II,Fi), daughter o f Kasonda o f malabu lineage. In 1953 he

1 Cf.

p . 18 1.

T 5!

V illage F ission Slavery and S o cial C hange




paid several visits to Yim bwendi village to buy meat from the
hunters there for resale in the vicinage o f M ukanza.
Mukanza and Kafumbu people regularly attended the others*
rituals; but Yim bw endi people, although sometimes notified o f
such events as the funerals o f important members o f these two
villages, never attended, and vice versa. I have heard Mukanza
wom en wailing in the traditional fashion w hen /news was
brought to them o f the death o f an elder in Yim bw endi Village,
but no representative was sent to the funeral ritual. O n the
other hand, when Pearson o f M ukanza V illage was killed in a
motor accident in 1954, many people came from Kafumbu
Village for the interment and mourning rituals. Again, I have
known large parties o f Mukanza people to g o to Kafumbu, in
the vicinage o f w hich Headman Nyaluhana is the principal
village headman (mwenimbu), when Kafum bu boys were initiated
at the boys circumcision ritual at Nyaluhana, and when a
Kafumbu girls puberty ritual was celebrated. I also attended
the Chihamba ritual in 1954 at w hich a final reconciliation o f the
two villages tacitly took place, and at w h ich I, an honorary
member o f Mukanza Village, was given the adept-name o f
Samlozang a Ndumba, also possessed b y the present headman o f
Kafumbu. T he possession b y tw o persons o f die same name
makes them majinda> * namesakes a mutual relationship o f mild
jo kin g and mutual assistance w ith goods and services.
In conclusion, jh erefo re,.it ..might. be.,saidjdiat. aftexL.die. fission
o f Mukanza Village, new relations offco rd iality,. expressed in
the maintenance and resum priom of. connubial-and, ritual ties,
were formed between Mukanza and Kafijmbu,^each o f which
became points o f contact between two. discrete vicinages. But,
on the other hand, w hile there has been a. marked, reduction, o f
open animosity between M ukanza Village andYim bwendi, the
great spatial distance set. between them, fu lly reflects- the con
tinuing., tension in their reciprocal relationship, and-is perhaps a
precaution against actual conflict.


t . )




Slavery and Social Change

For the purposes o f this analysis, Social Drama V I m ay be
said to have begun w ith the decision to change the site o f the
village, an event w hich m ay precipitate breaches o f regular soda!
relations which have been formed in a particular place. This



Schism and Continuity in an A frican Society

decision coincided w ith the permanent return o f Kahali Chandenda to the village as headman, an event which profoundly
modified the internal relations o f its members and precipitated a
crisis. Yim bw endi and Kafumbu refused to accept Kahali as
headman. For Yim bw endi had come to Mukanza Village with
a large follow ing about ten years before these events, and he had
hoped perhaps to succeed him self; while Kafum bu wished to
become independent o f his traditional masters, the members o f
n yach in tan g a lineage. Redressive machinery was brought
into action, in the form o f a special meeting o f the villages
leading men. B ut the aims o f the leaders o f the different groups
were divergent. Yim bwendi and Kafumbu wished to resolve
the crisis b y breaking away from Mukanza Village, i.e. b y irre
parable breach o f intra-village relations. Kahali, Mukanza
Kabinda, Kanyom bu, and Lupinda, the senior men o f nya
chintang a lineage, wanted to retain the lineages o f nyachula
and mpeza in the village. I f Yim bwendi had confined himself
to arguing a case for the secession o f his ow n follow ing there
m ight have been no quarrel, at least no overt violence, especially
since he and his kin had only been in Mukanza a short time.
B u t when he openly urged Kafumbu to go w ith him, and when
Kafum bu agreed to go, mentioning the discordant state o f
Mukanza Village, they touched o ff anger which culminated in
fighting. For at this point one o f the basic norms o f the tradi
tional social system was threatened, that w hich governed the
relations between master and slave. It was threatened not so
much by the demands and pretensions o f individuals, but by
major currents o f social change set in train b y British occupation
and legislation. Kafumbu, Tike people everywhere and at all
times, was merely utilizing new laws and norms to his personal
advantage. T w enty years earlier he could not have made such
a hid for independence. In the past runaway slaves were
hunted b y all free Ndembu, and i f captured were returned to
their owners w ho punished them severely. It was in the interest
o f all N dem bu to catch escaped slaves, for i f they did not do so
the system o f slave-ownership would break down. B ut slavery
had been speedily abolished under British Government and noone dared to take, action against slaves w ho had emancipated
themselves b y flight.
W h at were the main features o f slavery am ong Ndembu

V illage F ission , Slavery and So cial C hange


before British occupation ? Village slavery was distinguished

from commercial slavery. Village slaves were paid as fines to
terminate blood feuds, to settle debts, to compensate for homicide,
to pay a chief or senior headman whose poison-oracle had cleared
a person accused o f sorcery, and to discharge fines for a number
o f other offences. The principal mechanisms through which
payments o f slaves were made were : a ch iefs c o u r t; a council
o f elders or mahaku ; direct negotiation between the parties
involved in a dispute ; or the invocation b y the wronged party
o f some famous warrior such as Chipenge w h o defeated the
Chokwe and Lwena slave raiders, to press a claim against a
transgressor. T he persons paid as slaves were usually young
children. O ften they were already o f slave status; but i f a
man had to pay a relative, his choice usually fell on his ow n sisters
child. O n ly as a last resort w ould a man pay a junior sibling
and in custom he had no rights over the disposal o f his ow n
children in this manner. B ut in practice a man w ould resist the
demands o f a brother-in-law w h o wanted to pay his children,
the latters sisters children, as compensation for debt or homicide.
For example, the mothers brother o f one o f m y informants went
to m y informants father about thirty years ago in Angola and
tried to take his daughter to p ay her as compensation for a large
debt o f long-standing. B ut m y informants father became very
angry and told his brother-in-law to rem ove him self from the
village forthwith, saying, * I f yo u take one o f m y children, one
day, perhaps soon, w e shall have a fight. His brother-in-law
replied, Y o u have no cause to fight me. A re not these m y
sisters children ? I can do as I like w ith them.* The father
retorted, * N o , they are m y children. I f yo u have got into debt
because y o u love adultery and have had to pay many fines,
perhaps yo u can pay your ow n children. His brother-in-law
then said : Y o u are a bad brother-in-law (ishaku). Y o u have
no right to keep your children from m e. But he was afraid to
take the girl, and went away.
In an egalitarian society like that o f the Ndem bu, slaves were
not markedly exploited, and in fact were regarded more as rela
tives than chattels. I f the slave was young and his or her master
an elderly person, slave and owner initially stood to one another
in the fictitious relationship o f grandchild-grandfather. T he
term ttkaka in Lunda means both * grandparent and * owner *.


Schism and C ontinuity in an A frican Society

This relationship provided a point o f departure from, which the

slave could be assimilated into the kinship structure o f the village.
-T h e owners sisters son became his ^mothers brother , the
owners son his * father
and so forth. Female slaves were
often taken as concubines b y their owners and a new pattern
o f kinship terms with corresponding rights and obligations
developed from this relationship. T he children o f female slaves
by free Ndem bu remained in their slave status, the children o f
male slaves b y free mothers in the case o f such marriages the
slave husband had to pay bride-wealth were regarded as free
Ndembu and could go where they pleased. In other words,
status was inherited matrilineally. Slaves were inherited in the
same w a y as chattels, passing to the nearest matrilineal kinsman
o f the deceased. Slave widows, but not the w idow s o f free
Ndembu, could be inherited b y the brothers o f their dead hus
bands. Sisters sons also might inherit the slave widows o f their
uterine uncles. In the course o f time, as in M alcanza Village,
one finds in effect a large slave-lineage, like that o f mpeza , side
by side w ith and owned b y a free lineage. O ften, indeed, the
slave-lineage m ay be the larger, since most o f its members may
marry within the village while the virilocal marriage o f female
members o f the free lineage m ay reduce its effective membership
within the village. I have been told o f powerful slave-lineages
which, like cuckoos in the nest, ultimately ousted free lineages
from the headmanship o f villages. Perhaps Kafum bu had
cherished hopes o f doing likewise in Mukanza Village.
Slaves had to w ork in the gardens o f their owners, even, in
the case o f male slaves, after they had married, an act that normally betokens among free Ndembu the assumption o f adult
status and economic independence. I f they became hunters they
had to give the whole o f their kill, w ith the exception o f the
sacred portions reserved for the hunter, to their owners. I f the
owner o f a female slave did not wish to take her as his concubine
he received bride-wealth for her from any man w h o wished to
marry her, and her children belonged to him and to his heirs.
A slave could, however, purchase his independence and that
o f his relatives. I have recorded several examples o f this, which
suggest that slaves were entitled to accumulate property o f their
own. I give one case here b y w ay o f illustration.
T he father o f Kasonda o f Mukanza Village, a man named

V illage F issio n , Slavery and Social C hange



Kahumpu (II,D i), was an elder o f Shika Village, and his wife
v? Nyamwaha (I,Fy) o f Mtikanza Village was living w ith him virilocally. O ne day, he caught his w ife in delicto flagrante w ith a
' . man called Kalubinji, classificatory brother o f the headman o f
Kasai Village, one o f the ancient K aw iku settlements. Kahumpu
attacked Kalubinji and killed him w ith an axe. H e then ran
,; off and hid him self in the bush. N ew s was brought to Headman
Kasai o f the murder and he gathered a force o f armed men from
V his ow n and neighbouring K aw iku villages and came to the
s< periphery o f Shika Village. Headman Nsanganyi, senior head
man o f the K aw iku and son o f a former headman o f Shika
Village, came also to arbitrate between the groups. Kasai
shouted from the bush outside the village on tw o successive days
and demanded and obtained from Headman Shika Ikubi (II,C2),
:i: a promise to pay tw o slaves, tw o guns, 10 yards o f cloth and a
conus mollusc shell (itnba), as compensation for his brothers
death. Kahumpu (II,Dx) returned from the bush o n the second
day, and after pleading great provocation paid tw o slaves towards
the total compensation, a young man o f about fifteen called
Sakutoha and his sister, together w ith a gun. Kasai received
these. According to Kasonda, the reason w h y a large amount o f
compensation was made over to Kasai was that Shika and
Kahumpu stood in great fear o f the recently established European,
government (the murder took place in 1913) o f the British
South Africa Com pany which had made murder a capital offence.
Kasai had taken advantage o f the new situation to pitch his
demands very high. It was tacitly agreed between the parties
involved that no report w ould be made to the B om a i f the
Shika people agreed to Kasais terms, which were, in fact, a
form o f blackmail. But after compensation had been paid,
Kahumpu as a precautionary measure fled to Balovale District
for m any years.
The compensatory slaves, Sakutoha and his sister, were allo
cated b y Kasai to his classificatory mothers brother, Kahona.
They had com e originally from Nsanganyi Village and had
been paid as slaves to Kahumpus mothers brother, the late
headman o f Shika. B ut although they had thus been returned
to a K aw iku village sprung from Nsanganyi, they still retained
their status as slaves. Every little Ndem bu village, as long as it
endures, m ay almost be likened to a sovereign and independent


Schism and Continuity in an A frican Society

state. It was immaterial whether the slaves were Kawiku

or not, or whether they came from Nsanganyi Village or
not. This did not give them freedom. Nsanganyi and Kasai
were independent villages, each had its own business to
In the late twenties Sakutoha purchased his freedom from
Kahona with a muzzle-loading gun. A few years later he tried
to obtain his sister, w ho had borne Kahona several children, by
the offer o f another gun. Kahona did not wish to lose his con
cubine and children, the major part o f his follow ing in Kasai
Village ; he took the gun, feigning that he would release them
to Sakutoha in a short while. Sakutoha, w ho had returned to
Shika Village, was planning to break away and found a settlement
o f his ow n w ith as many o f his siblings as he could muster. But
Kahona then claimed that although Sakutoha had purchased the
enfranchisement o f his sister she still wished to remain w ith him,
and he offered Sakutoha bride-wealth to legalize the marriage.
He knew that Sakutoha could not prefer a charge against him in
a chiefs court, for the Boma might get to hear o f it, and the
British authorities might realize that slavery was still an active
institution among Ndembu. This would mean much inter
ference by Government in village affairs and possibly imprison
ment for both disputants in the case. Sakutoha, w hom I knew
well, was a simple and straightforward man, traditional in the
extreme, and he accepted the bride-wealth, although condemn
ing in round, terms Kahonas double-dealing. But his grudge
against Kahona and his family did not die out, for when Set,
Kahonas scapegrace son by Sakutohas sister, w ho had been
expelled from the Copperbelt for persistent thieving, was re
fused admittance to Kasai Village by the headman in 1953
Sakutoha w ould not allow him to build in his farm near Mukanza
Tim... case-history ..shows, the. tenacity w ith which Ndembu
clung to the institution o f slavery long after . British .rule had
overthrown or greatly modified major-aspects-of-the- traditional
political system. It also illustrates how. Ndembu,-both.slave and
free, played o ff the new laws against the old .customs.-to further
their, ow n interests.
It also reveals that slaves, although they might be inherited or
transferred like chattels and were compelled to remain in their

V illage F issio n , Slavery and So cial C hange


owners villages, nevertheless enjoyed certain rights. Sakutoha,

for instance, had been able to obtain a muzzle-loading gun o f his
own w ith which he had bought his freedom. The status o f slave
could not have been very onerous, for Kahonas slave woman
refused to leave him to jo in her brother, even though Sakutoha
had obtained her liberty and guaranteed her an important position
as the potential founder o f a village matrilineage.
tant institution in the thirties and. Lookedaskance.Qmihose who
had. resort to European-supported laws,, feeling.,against
those-who refused to conform to its time-honour.ed_norms.must
indeed .have been strong in. the twenties,. -when-Kafrimbu. (I,Gi)
made it clear that..he..\ new
ordinance. W h en Kafumbu and his followers joined, the fight
between -Yimbwendis and Kahalis parties, this act-represented a
violation, of. the. norms governing relations and
their_owners._ I n .the pasJLslaves. were.expected,.to..frght,.for their
masters 4/in this squabble the slaves fought against them. For
merly,, the sanctions that w ould have
such recalcitrant slaves might have included die .execution o f
their,, ringleader and the_suspension. o..the, rest b y their necks
from .thedforks o f tall trees. I f they had imposed these penalties,
the people o f Mukanza would have been supportecL b y the other
villages in their vicinage ; and C h ie f Kanongesha himself, i f
called upon, would have dispatched a_punitive force from.his own
vicinage, i f Kafum bu and. his group hadofferedresistance. But
the alien European authority was .known, to possess overwhelm ing force in the form o f armed police to. back np. its__laws and
render_nugatory attempts to uphold the norms o f the.traditional
political system. In short, the mechanisms which formerly
maintained the norms governing the relations o f slave-owners
and slaves could no longer operate, although m any Ndembu o f
both categories still adhered to those, norms. M oreover new
mechanisms had been introduced b y the British w hich applied
sanctions against Ndem bu w h o were c a u g h t.attempting to
maintain the institution o f slavery, mechanisms o f a legal charac
ter designed to enforce conform ity to norms whichjwere .opposed
to those o f the indigenous-society. Since there was no. way,
therefore, o f redressing a breach o f traditional relations between
slave-owners and slaves by traditional . machinery,, schism


Schism and Continuity in an African Society

beCwem. mcmbers o-die. tw.o categories within a. single vi Wage

resulted in^-the-fission of-that village.
Yim bwendi, on the other hand, was not condemned by the
people o f M ukanza br withdrawing his own lineage from the
village so much as for conniving w ith Kafumbu at the secession
o f the slave lineage o f mpeza (I,Fi ) from their hereditary owners,
the people o f nyachintang a lineage. Yim bw endis group,
after all, had previously split o ff from Kahali Village and could
no longer trace precise genealogical relationship w ith either
nyachintang a or malabu lineages, although they still remem
bered to what genealogical generation they belonged with
reference to the founding ancestor Kabonzu (I,C i). The lineage
o f nyach ula (I,D3) was sufficiently distinct from the combina
tion o f nyachxntang a (I,D i ) and malabu (I,D z ) lineages to
constitute a virtually autonomous group. B ut it seems that
Yim bwendi had at one time cherished ambitions to become
headman o f Mukanza and that he was encouraged in them by
Kafumbu, even i f this would have entailed the secession fronvthe
village o f most o f the members o f the other tw o lineages. But
he had not been successful in winning the support o f any o f die
leading men o f the established lineages without which he could
not have become headman o f Mukanza Village. He had to
content him self w ith securing the temporary allegiance o f
Kafumbu. This was an uncertain asset, for Kafumbu was also
an ambitious man, w ho wanted to use his relationship with
Yim bwendi as a lever w ith which to secure his ow n emancipa
tion and that o f mpeza s lineage, his personal follow ing, from
nyachintang a lineage.
Later, when the tw o men founded
their setdement, Kafumbu, whose follow ing at that time out
numbered Yim bw endis, claimed the headmanship o f the new
village and forced Yim bwendi, w ho naturally did not agree to
this, to leave the area.
W h y were ritual mechanisms not brought into operation to
resolve the conflicts In Mukanza Village, as in the case o f Social
Drama II in which Sandombu was reintegrated into the Mukanza
community ? The answer probably lies in the nature o f the
social relationships involved in the respective dramas. San
dombu was a close kinsman o f the principal actors in the earlier
social drama and an almost life-long member o f the residential
group. Kafum bu and his faction were slaves, and breaches o f

Village Fission, Slavery and Social Change



owner-slave relations w ere traditionally settled b y legal means
supported b y the sanction o f organized force. Y im bw en di and
his lineage kin w ere indeed matrilinea! relatives o f the M ilitanza
; -group, but on ly distant relatives, so distant in fact that tw o
marriages had taken place between members o f the respective
groups which w o u ld have been considered incestuous had kin
ship nomenclature alone been the criterion. B efore the split,
Kanyombu (I,F 9 ), o f n y a c h in ta n g a lineage, had married
Nyamalata (1,623), f n y a c h u ia lineage, w hom he called
sisters daughter ; and after the split Line (I,H i7 ), o f malabu
; lineage, had m arried M aria (I,H i9 ), o f n y a c h u la lineage, w hom
k; he called * sister * b y one means o f reckoning kinship. Th e sole
if common ancestress o f these m en and w om en was Chipendipendi
(I,A i), mother s m other o f Kabonzu (I,C i), the founder o f the
village that was subsequently known as Kahali, then as Mukanza.
Chipendipendi was in the fifth genealogical generation before
Kanyombu, the sixth before Nyamalata, and the seventh before
Line and M aria, so that the relationship between these spouses
was extrem ely distant. Yim bw endi and his group had lived
separately from the rest o f the village for about tw enty years
before they returned after Mukanza Kandulus death, so that
ties o f spatial propinquity between the tw o sections had lapsed
and had not been strongly revived in the short interval o f nine
years before the final split.
It w auld seerxis, therefore, , and this view by

eviden<^-foom.QfoeiL..vill^gesr...that.. _ritual.. mechanism sare _only

linealJdn w ith a. genealogical depth, o f about three-generations
at m ostbetween founding-ancestorand the. oldest, living-members
o f the lineage, w hat ..I call a 4minor-lineage -__RituaL mechan
isms are only brought .into operation,-moreover,_where_ties o f
lineage.Jdnship-.are .supported b y . a-long--tradition -o-joint resi
dence beweeELthepersons-.aiid-Subrliiieage&.concerned. More
over, as a rule only a single type o f ritual is utilized primarily
for this purpose, the ku-sw anika ijina ritual, in the course o f which
a muyombu tree is planted to a matrilinea! ancestor or ancestress
o f a village member w ho inherits on that occasion his or her
name (ku-swanika , to cause to inherit or succeed to, ijina, a
name). M iastjother xituals,. aithough-.they. m ay,Jiave the subordinate^function _-of reintegrating, the -ritual subject's.-village,











O .





Schism and C o n tin u ity in an A frica n S o ciety

dominantly emphasize the unity oalLNdem bu,-or- even- o f all

Lunda. But the lineage group which is reunited b y the ku~
swanika ijina ritual is the effective local lineage. Marriages
between members o f this lineage other than between classificatory
grandparents and grandchildren are reckoned to be incestuous
unions (chiinahvamalwa or chipikapika), and a ritual (ku-sola, or
ku-disola) is performed to cleanse the couple i f they have a
common great-grandmother. W ithin closer degrees o f matrilineal kinship than this, in pre-European times the couple might
have been permanently ostracized or put to death.
In Mukanza Village in 1954, when I left the field, the groups
predominantly concerned in ku-su/anika ijina rituals were tend
ing to become, not the mayor lineage o f nyachipendi (I,C2),
the mother o f Malabu (I,D2) and Nyachintanga (I,D i) ; but
the separate minor lineages descended from these daughters (see
M ap 4). The children o f Manyosas, Chaw utongYs and
Kasondas generation, and the grandchildren o f N yam ukolas,
N yatiolis and Sakazao s generation, when they g ro w up, will no
longer constitute together an effective local lineage, and ritual
mechanisms w ill no longer be able to reintegrate the whole
village i f there is a breach o f regular social relations between the
tw o sub-lineages. In any case, new social forces which are now
rapidly developing w ill almost certainly accelerate fission o f the
larger villages into ever smaller residential units.
to occu_iii_the_first -pexiad_Qf..sociai_change Traditional ties
became more brittle ; traditionaLsocial-machinery-for-reintegrating/ a . disturbed _group....became. .replaced., in.. scv.eralsituations
b y the'' legal machinery o f the. ..superoxdinate_aliem_authority ;
new -norms regulating behaviour were .introduced from above ;
or new norms began to develop on the basis_offnascent social
relationships o f a new type and ..came. .intoconfhct^with old
noons. But thdTirsL.efiect,othese.imiovations^ the
loos.emng..jQftraditionaL-ties^-rathertham-theirreplacemenc. by
newones. The <firsja..these_ties-to^snap^was-the. owner-slave
link, and all over Mwinilunga_DistrictJmtheGL92o!&_we-fmd the
growth, o f . new settlements.JnhahjtecLb.y_for.rner slaves. This
territorial .enfranchisement, was everywhere.._accompanied_. by
conflict between slaves and their former owners But-itw as not
long before the new situation .was acceptecLby...the., majority o f

V illage Fissio n , Slav ery and So cial C hange


the-people ;... and. although, in private conversation... a. certain

stigma-" is attached to membership o f such- .a-viUage._by free
in and
free l^embM-inteinnarry. freely; andtfie^ ch^
w om en, are. regarded- asJxee.
The role o f n yach u la lineage in the situation illustrates not
- only one type o f fission, but also accretion as it operates among
, the Ndembu. W h en Yim bw endi returned to M ukanza Village
after tw enty years he was utilizing ties o f matrilineal kinship
which had n o t yet fallen into complete abeyance. His ow n
uterine uncle Saluvaji (I,E6), the father o f Line, moreover, had
remained in M ukanza Village after (he first secession o f the
n yach u ia lineage from Kahali W ebalas village, and acted as a
living link b y means o f w hich the lineage could be attached to
the village once more. In most N dem bu villages one finds a
few scattered classificatory matrilineal kinsmen o f the headmen,
members o f sub-lineages that have split of hut w h o have re
mained perhaps on account o f marital ties, as in Saluvajis case,
with the principal village lineages, perhaps because they have
fallen out w ith their o w n primary matrilineal kin. These per
sons, for example, N yaw unyum bi (LG25) and her children in
Mukanza Village, in practice seem to assume the function o f
maintaining connection between the dissident and remaining
sections o f a divided village and to provide points o f re-entry for
members o f the seceding group. In other words they have
important roles in the system o f inter-village and inter-vicinage
relations, or from the point o f vie w o f members o f a village in
their system o f external relations. Fission is seldom along abso
lutely clear-cut lines. Ties other than those o f lineage affiliation
may temporarily or even permanently prove stronger than the
latter so that each core o f primary matrilineal kin in a village
may have a number o f classificatory encrustations. Such per
sons must not be regarded simply as social isolates withm a
v illa g e ; they have functionally significant roles in the under
system. W h en Kayineha (I,G6), for instance, was living at
Mukanza Village he held the door open, as it were, for a
reconciliation w ith Kafumbu.


Schism and C ontinuity in an A frican Society

Lineage or Uterine Sibling Group ?

In effect, the groups w hich split o ff from Mukanza Village
were groups o f uterine siblings. Kafumbu managed to gain
control over tw o o f the daughters o f Mpeza (I,Fi), and these,
together w ith their children, formed the core o f his new village.
Kayineha should have been the leader o f this group but his
attempts to persuade his sisters to remain in Mukanza Village,
where he had married, were resisted b y them. Kafum bu had
many children b y Katendi (I,G2), and Katendi herself had con
siderable influence as the oldest sister and a strong personality
in her ow n right over K aw ila (I,G3) w ho also had m any children.
Here the powerful bonds between uterine siblings and between
members o f a matricentric fam ily were exploited b y the husband
o f the oldest sister in his o w n interests. The success achieved
by Kafumbu (I,G i) against Kayineha (I,G6) in the struggle to
control the all-important fruitful wom en is one case in which a
husband defeated a brother. M ore often in such a struggle
between brothers-in-law the brother wins ; but the outcome
in Ndem bu society is never a foregone conclusion. Undoubtedly
the fact that Kafum bu was a cross-cousin o f his w ife and her
siblings was o f great assistance to him.
Yim bw endis follow ing consisted o f his ow n group o f uterine
sisters, and their and his ow n children. B ut his youngest sister
died ; and her daughter Nyawunyum bi (I,G25) suspecting
Yim hwendi o f having killed her mother by sorcery1 to enhance
his hunting powers, refused to leave Mukanza w ith him. His
uterine uncle Saluvaji (I,E6), w ho also remained behind, was a
feeble old man w h o had strong bonds o f marriage and friendship
with Mukanza Village and who, besides, would not accept the
authority o f his own. sisters son. B ut Yim bw endis other tw o
sisters and their children remained loyal to him.
Later when Samuntima (I,G22) split o ff from Yim bw endis
village after Yim bw endis death, he led aw ay w ith him his own
uterine sibling group, leaving Makumela (I,G24), his maternal
parallel-cousin, w ith Makumela*s ow n sibling group. The
grpup ^ f uterine siblmgs ..under foe ^leadership o f the oldest or
most_capable_Jbxo.foeii-is_the.basic_Ndembuceridenrial_unit and
1 Cf. pp. i8i, l 8<5.

T 3

V illage F ission , Slav ery and So cial (Change


the most com m on nucleus o f a new village... -Happy. is the

ambitious man .who. has .many .5is.ters and..unambitioiisLy.ounger
brothers w ith children o f their ow n. . Sandom.bu.-is. anuexample
o f a m am who is not favoured w ith fertile siblings^_ond,who_iias
no children o f his ow n. H e must resort tp discreditable-means to
scrape, together w hat follow ing he can,, and. is. always regarded
w itLsuspicion. This .nuclear..sibling., group ..has always, been
important as the pioneer element i n ..foundmg_a_jxew._.village,
according to m y genealogical material and
mation. 'collected from m y older informants.. B.ut_in._the..past,
say the latter, fission was much less..frequent,. his
own ,< a. m a n . w auid-hiveoffi-with- other
members^-of a--threer?generation .o r. ...V. .minimal-- - lineage, the
matrilineal descendants o f a common, grandmother...Today the
ties o f classificatory kmship .are wearmg. ever. thinner and each
uterine jsihling group_is..a. potential...unit o f -secession.
T o .achieve. .the qualitativetransformation, from , a-uterine
sibling village to a three- or four-generation matrilineal .village
represented a substantial social achievement in.-tbis unstable-and
volatile society. It required the proliferation o f a num ber-of
supplementary ties within the village interlinking... members o f
potentially opposed lineage segments. That is w h y emphasis
was placed in N dem bu culture on unity within the genealogical
generation and on the alliance o f alternate generations irrespective
o f the lineage affiliations o f their members, and w h y succession
and inheritance as far as possible ran laterally, even to cousins,
rather than lineally. T hat is w h y so much intra-village classificatory cross-cousin marriage took place between the children
o f matrilineal parallel cousins o f different sexes and lineages, and
w h y grandparent-grandchild marriage is found between lineal
kin belonging to different sub-lineages o f a single matrilineage.
That is w h y, also, members o f a slave lineage in a given lineage
were married b y members o f different village sub-lineages. For
the slaves acted, as it were, as warp to the lineage w eft. JB.ut.the
and Hxihvid^uahstic^. X^Memhupers.omlity....pxoxed_.a--St.umbling
block totheir. .development. Jealousy between classificatory
mothers brothers and sisters sons and also between classifica
tory matrilineal brothers over headmanship, w hen the last
headman o f the senior generation died, and..-the...extreme





l @


i. /

Schism and Continuity in an A frican Society

r e lu c t a n c e o f t h e . d e f e a t e d c la im a n t . to _ s .u b in it_ t< X - th n a u th Q r ity o f

the-successful one,, were frequent causes .offission .The.ahsen.ce o f
a strong, centralized political authority capable, .of. compelling
people to remain in a village after .conflict .had occurred..also
favoured village schisms. B ut once a village. had.-b.ecame_established .great efforts were usually made,as in JMukanzaJ^ji ilage, to
p r e v e n t its L .d is m p tio n . a n d _ t.Q _ Jc e e p . i t a l i v e .


V :






<r. /.

*vi ...";


Personal Factors
Much...depended .onthe._.personality.._Q_the.._headman, the
supreme agent o f the transition from_uterine...sibling_jallage to
matrilineal village, i f the group . remain._undiminished
through the period o f instability.which_super.venecLwhen the
proximal generation .to. the village .founders. str_uggle
among Ndembu _were .firni.-hutmnahtrusive_.pers.onaIijtie.s,.jjiiaggressive, and. ready to . share ,whatjwealth...they jnight._acquire
withjtheir relatives. Scrupulous, fairness is reckoned -to...begone
of_their most essential characteristics. B ut .notion o f
whatJSfdembu consider the ideal personality type, fbr..headmen
is afforded b y the follow ing-text I collected-about-Headman
N g arnbi during a discussion w ith somemembers-of-the vicinage
to which both N g ombi and Mukanza -Villages belong :
Headman N gombi, even before he succeeded to headmanship,
used to laugh with everyone, elders and children alike. He liked
both men and women. If Ms younger brother o f Ms own lineage
or o f a senior or junior lineage, or Ms sister*s son, or his father were
hungry he would remember to offer them food. If they are very old
(adinawevu, literally, *those who have a beard ) he helps them
(wayikwashang a, literally *he helps them all the time : ku-kwasha
means * to help in every way with goods and services, to be solicitous
to their every need ) ; i f they are sick he helps them ; i f they have
come from another village he helps them with food, with beer, and
gives them a place in wMch to sleep. And even i f the Europeans
send their messenger he will help Mm with one thing or another that
he may require. If CMef Kanongesha is coming to call a meeting o f
headmen he will bring a calabash o f beer or a sheep or cassava meal
as tribute. CMef Kanongesha is very pleased and says * You have
brought me tribute as a village headman; that is a good thing to do.*
He tells him The one who helps me in our village is a son in. every
respect. Headman N gombi has a good liver, he helps many people,







V illage Fission , S lav ery and So cial C hange


he makes gardens for many people (i.e. to give them hospitality), he

digs these gardens, he has strength, he has ability to argue cases,
{waheta wuhaku), he does not steal nor tell lies, he does not slander
people, he is not a sorcerer.
B y w ay o f contrast here is a text describing the personality
o f a headman w h o was foared and disliked :
He is a bad man, selfish (uraheta chifwa), proud, quarrelsome, given
to reviling people, lazy, a liar, without skill in legal argument, a thief,
lustful, a slanderer and a sorcerer. He is wqfwaha walu/a, a person who
scrounges beer from everyone without making a return. Such a
erson who takes without returning is like a mad dog. A man who
as self-respect (kavumbi)1 does not do so. He also respects other
people, without regarding them as useless.

Ndembu_do not admire overweening^_domineering^hea.dbtiien ;

men w ith such temperamentS-are condemnecLaiictaithough they
not^eldGm..acquirer Jarge.ibllow ing. B om Yim bw endi and Kafum bu appear to have been men
o f this type. M ukanza possessed many o f the qualities attributed
to N g ombi, although he was not such a good arguer o f cases.
Kasonda once told me that m any headmen eat privately in their
own kitchens, sending w hat is left o f their food to the chota
to be eaten b y the dkwachota (the group o f male villagers, regard
less o f lineage or generation affiliations). B u t Mukanza,
N g ombi, Kam aw u and M ulila (other respected headmen in the
vicinage) preferred to eat in the chota w ith * their juniors and
children \ T h ey were ridiculed for their commensaKty by such
headmen as Nswanakudya, Chibwakata and Nsanganyi, but
the people said that this behaviour proved that they were not
sorcerers (aloft), since it is believed that sorcerers share their food
with their creatures o f sorcery, such as malomba and tutotoji.
O nly chiefs are expected to eat alone or with their children,
apart from their people. O n account o f Chibwakatas pnde
(iwitiyi), argumentativeness and reputed sorcery, four groups,
led respectively b y N g om bis uterine uncle the founder o f the
village, N yakapwipu, Sayifon and M akayi, divided o ff (kubaluka, literally * to split a piece o f w ood *) from Ciiibwakata.
1 C . M . N . "White (personal com m unication) writes that * havumbi means
respect fo r others, n ot self-respect ** \ M y o w n informants, how ever, gave
m e the m eaning self-respect \ care to b thou ght a good man \


Schism and Continuity in an A frican Society

The_good headman.is_the..go.od.eUo_w,_the man. .who.. * laughs

w ith .everyone , w ho is hospitable, ..selferespecring,...helpful and
democratic. Perhaps the most significant section o f the com
mentary on Headman N g ombi is that in which he is described
as a man w ho would give food not only to * a younger brother *
o f his ow n lineage but also to members o f other lineages in his
village. His field o f friendly co-activity is not circumscribed by
narrow minimal lineage relations ; it extends outwards to include
everyone o f the village, regardless o f their precise degree o f
relationship to him. The headman in his person should typify
and exemplify the most general norms governing social inter
action within the village. His generosity and range o f interests
should not be confined to his ow n village ; he should offer
hospitality to visitors from other villages and should be able
to argue cases in other villages w ith freedom from prejudice.
Towards the agents and representatives o f external authority,
indigenous and alien, he should be courteous and respectful so
that his village should be w ell thought o f and he and his people
should not incur the rancour o f capricious officialdom.
Am ong Ndembu, however, not. ..fiad.m en who
measure up to these exacting standards o f generosity.,, impartiality,
and-unselfishness. For txadhrion. and_training-uphold .and pro
duce-among the majority of.m en. a many
w ays opposed to the ideal ty p e .o f.headman. This..i$ the ideal
personality o f the hunter,_The__professional hunter., (chiyang'a)
is a_man o f the.bush who., pitting
his-wits, against the.fieetness or-erocityL.oaxiimals^-fierce himself,
boastful,, b y .character
unable-to-stomach...the. .authority .o f another,. patriarchaLxn his
family, life in a matrilineai society,. a. wanderer.who. m ay travel
from M wantiyanvwas country to Ishindes chiefdom in Balovale
District in quest o f game, and given to long and frequent drinking
bouts round the calabashes, o f beer..madedfeom .the. honey which
he fearlessly collects from .the nests.Qf.w ild bees, in.tbe. woodland.
A prayer made by hunters at the beginning o f their Wuyanga
ritual defines the type o f man that is meant b y * a professional
hunter (chiyang a). * W e w ant a man w ho can sleep w ith ten
wom en in one day, a great thief o f a hunter. * T h ie f actually
refers to the theft * o f game from the bush, not to the stealing
o f property, but the bias o f adulation for what in ordinary life

V illage F ission , Slav ery and So cial C hange


would be illegal acquisition is plainly evident. .Again, a head- man should not be licentious ; in a hunter sexual..potencyr-^with
a fine.-.disregard o f its lawful directionis-reckoned a -strong
.. qualification. It is clear that such a personality type is diametric; aily o p p o sed .to that, thought..proper_for_a.headm an. Few
Ndembu..succeea-in..achiev3ng..a workable compromise between
these ideals. M any professional, hunters are headmen, .but few
o f them, are successful headmen ; many successful headmen.w ith
, large followings are not professional hunters. B o th Kafumbu
and Y im bw endi were hunters, and acted as hunters, but not
headmen, were supposed to act. O ne o f the reasons w h y
Yim bwendi w ent south into C hibw ika Area was because there
. was much gam e there. Although Mukanza was him self a
-hunter he was not a great and dedicated hunter, and spent most
o f his time in and around his village.
1 Because fe w men possess or develop the personality ideally
'required for headmen new settlements o fte n .fail to b e c o m e
established. T he spine o f continuity w hich maintains. aJlongestablished village is provided .by its. famous or by
the mere, fact that it has persisted fromJong.-aga~and~has.-thus
acquired-the prestige o f longevity..- Where...these-footers are
absent, w hen the founder o a_village dies, fissionjojften..occurs
between uterine sibling groups sprung from different, mothers.
-Thus Yim bw endi Village split into the groups led b y Makumela
and b y Samunuma, and Kafum bu into those o f Samlozanga and
Kafumbu K am u. T he o f the founder becomes extinct
in many .cases, and new discontinuous settlements., come into


. .

( .:
( .)



4' :

I ..



N order to demonstrate that the tendency for groups o f uterine

siblings to split ofF under the oldest or ablest brother is no
unique or exceptional phenomenon, I present Tables XVII, 4
X V III and X 3X compiled from data obtained from village
genealogies collected in four Government chiefdoms and from
discussion w ith the informants w ho supplied the genealogical
material. Table X V II shows the relationship o f the leader o f
the seceding group to the headman, whether Hassificatory or
descriptive, and Table X VIII shows the span o f the maternal
descent group w hich formed the nucleus o f his new settlement.
I restate m y use o f terms : * uterine sibling grouj^ * refers to the
children o f a single mother ; * minimal lineage^ to the matri
linea! descendants o f a single grandmother ; * minor lineage * to
the matrilinea! descendants o f a single great-grandmother ; and
major lineage * to the matrilinea! descendants o f a single greatgreat-grandmother. I found no lineages o f greater span than
the * minor lineage * as nuclei o f new settlements. Lastly, I give
in Table X IX the relationship o f the leader o f the seceding group
to the senior wom an o f his new settlement, in order to bring
out the feet that there is often a struggle between brothers-in-law
for the control over a wom an and her children, and sometimes
over her younger sisters as well, as in the case o f the rivalry
between K afem bu (I,Gi) and Kayineha (I,G6) described in
the previous chapter.
Table XVXI shows that 18 out o f 33, or roughly h a lf o f the
leaders o f seceding groups were sisters* sons, primary or classificatory, o f the headm en o f the original villages. O f the 11 classifica
tory sisters sons, 8 belonged to the same minor lineage as the
headmen but to different minimal lineages within it, 2 belonged
to the same major lineage as the headman but to different minor
lineages within it, while one, Kafumbu, was more distantly related
to the headman, KahaJi Chandenda. It has already been pointed
out that the secession o f sisters sons, both primary and classificatory, is often associated w ith the tendency towards adelphic suc
cession in Ndem bu villages (Chapter Three, p. 87). O nly one



V arieties o f V illage Fission


iS fe- uterine brother seceded from his older brothers village, and in
SMhl this case he made a small farm w ith his tw o sisters and his and their
R ela t io n sh ip s B

etw een

L e a d er s

o f S e c e d in g G r o u p s
H eadm en




z S .....................
. . . .
S .....................
Cross-cousin .
mB . . . .
Not known .






P r ev io u s



Total number o f seceding groups

' t.'ty '

a n d t h e ir


V ' ' j --


R e l a t io n s h ip o f L e a d e r o f
S e c e d in g G r o u p t o S en io r
W o m a n o f S e c e d in g G ro up

L in e a g e S p a n o f N u c l e a r
S e c e d in g G r o u p s
Category o f
nuclear group


S .V \

Number of

Uterine sibling
group . .
Minimal lineage
Minor lineage .


Class. B


Total .









'/ :

^ c


O B Older brother
Y B Younger brother
S Son
H Husband

Class. B

Classificatory brother
Sisters son
Mothers brother

children about a hundred yards from his original village. O f the

4 classificatory brothers w h o hived off, one belonged to the same
major, but to a different m inor lineage as the headman ; one


Schism and Continuity in an A frican Society

to the same minor, but to a different minimal lineage ; and two

to different uterine families in the same minimal lineage. In
both cases where classificatory mothers brothers split off* they
belonged to the same major lineage but to different minor lineages
within it.
Abundant confirmation is offered b y Table X VIH for the view
that the dominant nuclear unit o f a new settlement is the uterine
sibling group. In 21 out o f 33 new settlements the nuclear group
consisted o f uterine siblings ; in 11, o f a minimal lineage ; and
in only one case o f a minor lineage. Secession o f a minor lineage
occurred in one o f the oldest villages in the N dem bu region,
M wanta waLuunda in Nyakaseya area, the headman o f which
used to have the office o f supervising the caravan which took the
tribute that was irregularly despatched from C h ie f Kanongesha
to M wantiyanvwa. The classificatory sisters son o f the headman
built a new settlement quite close to that o f M wanta waLuunda
Village and continued to pay tax through the latter. The uterine
uncle o f this leader o f the seceding group had previously been
headman.1 In die majority o f cases in w hich a primary sisters
son founded a new settlement this was built at a considerable
distance from the original village. W h y this should be so I do
not know , unless it is because conflicts between close kin o f
proximal generations when they occur are liable to be more
violent than between distant kin, ow ing to the greater intensity
o f their previous interaction.
This theme o f the spatial dispersion o f close lineal kin after
fission in a village pervades Ndembu society just as the theme o f
lineal ramification in a single locality pervades Tallensi society.
I f w e examine the spatial arrangement o f almost any vicinage
in the Ndem bu region w e shall find few adjacent villages indeed
that are linked b y close lineal ties. Close lineal kin do not seem
to make good neighbours i f their leaders belong to adjacent
genealogical generations. O n the other hand, when uterine
brothers or parallel cousins o f the same minimal lineage divide,
the new settlement m ay often be built adjoining the old. Thus,
when Makayi, younger brother o f Headman Chibwakata,
founded a farm in 1952, he built about a hundred yards from
his brothers village ; when Sakanya, parallel cousin o f Headman
1 Cf. the case o f Sandombu in Chapter Four.

Varieties o f V illage Fission


Machamba, made a new settlement he built about fifty yards from

Machamba Village, and so on. W h en ties o f genealogical
generation affiliation are combined w ith lineal connections they
seem to act against the breaking o f links o f spatial propinquity,
r W hen, on the other hand, conflict breaks out between close
: lineal kin o f adjacent, and hence competitive, genealogical
generations, ties o f spatial propinquity are abruptly and irremedi
ably severed ; and the seceding group moves out o f the vicinage.
Indeed, frequently it moves out o f the senior headmans area to
v which the original village belongs, an area which m ay include
several vicinages. In seven out o f eight cases where fission took
place between groups headed respectively b y primary mothers
brodiers and sisters sons, I found that the seceding group had
settled more than six miles from the original village. O n the
other hand, where classificatory mothers brothers and sisters
^ sons had separated from one another, I found that in eleven cases
: the new group had settled within six miles o f the original village,
and in only one case more than six miles from it although in
five instances the new settlement was in a different vicinage from
the old.
In the m ajority o f cases (22-33), the leader o f the seceding
group was the oldest brother or classificatory brother in a group
o f uterine siblings or in a minimal lineage. In six cases he was
the son o f the senior wom an in his following. This again
illustrates the tenacity o f the mother-son bond in Ndembu
society. I f the mother o f a founder o f such a settlement is alive
he w ill take his mother w ith him. She w ill become the apical
ancestress o f the village lineage i f the village becomes firmly
established. In a sense, she is a living index o f the matrilineal
character o f the N dem bu village, and she is her sons embodied
hope that his sibling village w ill become a m inor local matrilineage.
In order to bring out more fully the structural implications o f
fission in Ndem bu society I give some examples o f different types
o f village cleavage below.
(a) Fission in Shika Village 1
Shika Village, about twelve miles from M ukanza Village today,
is a village whose nuclear lineage belongs to the chiefly matrilineal
1 Appendix II should be consulted for genealogical information.


Schism and C ontinuity in an A frican Society

group o f Kanongesha. In the latter half o f the nineteenth century:

its founder Kanjimana (II,Bi), uterine nephew o f Kanongesha
Kajinga Mpata Yibam ba (II,Ai), left the area o f Kanongesha
near the Lovw a R iver in Angola, on the occasion o f his uncles
death. He moved because he feared the jealousy o f the new
chief, the son o f Kanongesha Kajinga ; 1 and he settled near the
present site o f Shika Village b y the Lunga R iver w ith his brother
and three sisters, W ayanda (II,Bz), Nyam bunji (II,B3) and
Chilum bu (II,B4), his wives, and the children o f the uterine
sibling group. After his death he was succeeded by his uterine
nephew Ikubi Chingongu (II,C2). The older brother o f Ikubi,
Chanza (II,Cx), resentful that his younger brother had been
preferred to him as headman, founded a new village with his
children, one o f his sisters and her children, and his classificatory
sister N yakaw onga (II,C3), daughter o f Nyam bunji (II,B3),
and her children. N yakaw ongas son Nswanakudya (II,Dio)
became one o f the first African Messengers o f the British South
Africa Com panys Administration. Nswanakudya returned to
the new Boma near the Lunga, but after a short while left the
service o f the Administration and rejoined his kin at Shika Village.
B y 1914, five years after his return, he had become the leader o f
a strong uterine sibling group and had hopes o f succeeding Ikubi
(II,C2), now an old man. He was older than Chipwepu (II,D4),
uterine nephew o f Ikubi and his designated mulopu (deputy) ;
and had a larger follow ing than Chipwepu. B ut in this year
his sister N yawatwa (II,D ii) committed incest w ith her parallel
second cousin Samuheha (II,D6), uterine nephew o f Ikubi. The
old headman drove the pair violently from his village, which at
that time was quite near the present site o f Mukanza Village. The
incestuous couple lived for several years in the bush, shunning
the local villages. Eventually Ikubi forgave them and the ritual
o f ku-disola was performed to cleanse them, for they belonged
to the same minor lineage w ith a com m on great-grandmother.
B ut he said that they must live in another village lest his people
should feel ashamed by their presence. Nswanakudya (II,Dio)
m ay well have seen this expulsion o f his sister as a blow to his
hopes o f succeeding, for she had four children already by two
1 Succession to the Kanongesha chieftainship was patrilineal until 1890.
Cf. p. 321.

Varieties o f V illage Fission


'-.previous husbands, a pow erful pressure group to support his

claim for headmanship. In the early I920*s he broke away from
Shika V illage and invited his sister and her husband, to build w ith
him. He started his new v illa g e 1 w ith his m other and his
own uterine sibling group, including his older brother M atoka,
two younger brothers, and tw o sisters, one o f w h om had been
married to the late headman o f the nearby village o f Chibw akata.
Nswanakudyas ow n father had been headman o f Chibw akata
Village, and m any ties united and still unite the tw o villages,
both descended from the first Lunda invaders. Later Nswana- kudya was joined b y his m others m atrilineal parallel cousin
7 Nyakashilishi (II,C4), daughter o f Chilum hu (II,B4), daughter
t o f the apical ancestress o f Shika V illage lineage, nyachibamba
;.(II,A2). Nyakashilishi had recently been divorced b y her
husband and had elected to g o to Nswanakudya rather than to
Shika, her village o f origin. Perhaps this was because her
mother was closer to N swanakudyas grandm other in sibling
order than, to Headman Shikas grandm other from w hom all the
contemporary members o f Shika w ere descended. O ne often
finds on consulting N dem bu V illage genealogies that the matri
lineal descendants o f adjacent siblings o f similar age tend to live
together, rather than w ith the descendants o f a m uch younger
or older sister.
In this case w e find a m arriage, reckoned incestuous, between
members o f a village lineage, as the prelude and catalyst o f fission.
Had there been any able senior m en am ong the children o f
Nyam bunji (II,B3) it is possible that one o f them m ight have
succeeded Ikubi, and the division o f Shika V illage m ight have
been averted. B u t the senior lineage o f w a y a n d a retained the
headmanship and rendered virtu ally certain the secession o f
Nswanakudya (II,D io) w h o belonged to a ju n io r lineage. It
is difficult to say whether the group w hich follow ed Nswanakudya
was a m inim al lineage or a uterine sibling group to w hich w ere
attached Nyakashilishi (II,C4) and her daughter. Certainly w ith
reference to the com plem entary and opposed lineage o f w a y a n d a ,
the split occurred between that lineage and those o f nyambunji
(1133) and chilumbu (II3 4 ), her sisters ; but in practice, the
1 Nswanakudya told me that when he left Shika Village there were about
thirty huts there.

2 io

Schism and C ontinuity in an A frican Society

seceding group consisted o f Nswanakudya, his m other, and his

Finally, it should be noted that the breach o f residential con
tinuity w hich occurred w hen Nswanakudya and his uterine
sibling group left Shika V illage to go w ith Chanza (II,C l) prob
ab ly helped to disqualify him for headmanship in the eyes o f
those w ho remained, the close lineal kin o f headman Ikubi (II,C2)
and his uterine nephew Chipw epu (II,D4). M embers o f groups
w hich have once broken aw ay from villages, like those o f
N swanakudya in Shika and Yim bw endi in M ukanza, i f they
return are less likely to succeed to office than are persons who
remain. Residential continuity, difficult to maintain, is prized
in proportion to that difficulty.
(b) Fission in Nsang anyi Village andmits Offshoots 1
N sanganyi is the senior K aw iku village in M winilunga
D istrict, and was already in existence when Lunda invaders came
from M w antiyanvw a. T he headman claims also to have come
from Luunda the homeland o f all the peoples o f Lunda stock,
at a period long before the arrival o f Kanongeshas Ndem bu.
Informants say that all the K aw iku villages and farms in the
D istrict, tw enty-six altogether, hived o ff originally from
N sanganyi. In spite o f periodical loss o f membership Nsanganyi
V illage did not becom e extinct, but after each secession slow ly
but surely expanded once m ore. The follow in g account attempts
to trace briefly the history o f fission in N sanganyi and its recent
offshoots in the tw enty-five years between 1929 and 1954.
The present headman (III,D4) succeeded to office in 1924.
His predecessor had belonged to a different m atrilineage within
the village, and they knew o f no com m on ancestress from whom
both could trace matrilineal descent. The present headman called
his predecessor * older brother w hich im plies that they were
m atrilineal kin. N either lineage was reckoned b y the people o f
N sanganyi village to have been o f slave origin. Perhaps the
lineage o f the previous headman, that o f n yan k u k a (III,B i ), may
be regarded as a group analogous to w hat Professor Fortes calls
among the Tallensi an attached lineage , linked to the authentic
lineage, that o f nyachim a (III,A i ), by actual or putative cognatic
1 See Appendix HI,

V arieties o f V illage Fission


ties, and in the course o f tim e treated for ail practical purposes
as matrilineaUy connected. I have collected another village
' genealogy, that o f Chikanga, w hich shows that the founders
son succeeded him , and was succeeded in his turn b y the founders
sisters son o r sisters daughters son. T he uterine siblings o f the
founders son rem ained in the village and his sisters children
regarded it as their prim ary local group. Thus in a single village
. were found tw o distinct matrilineages, one descended from the
senior w ife and the other from the m other or sister o f the founding
ancestor. In the second or third descending generation, members
1 o f the tw o groups began to call one another * brother and
sister * instead o f * cross-cousin *, since they w ere genealogical
generation mates prim arily affiliated to the same village. Ties
o f co-residence and com m on genealogical generation affiliation
v took precedence over lineal tics w h ich m ight have differentiated
i; the tw o groups.1 T he same thing m ay have happened in
; Nsanganyi V illage, w hich is o f considerable antiquity and enjoys
great prestige. These factors m ay have induced both lineages
to remain in the same village.
W hen the present N sanganyi succeeded, tw o m en in particular
were disappointed in their claims fo r office. O ne was his
classificatory brother M atem pa (III,D3), w h o was son o f the
new headmans m others older uterine sister, and w ho was an
able jud ge and advocate and one o f Senior Headman M ukangalas
councillors. T he other was N sanganyis classificatory sisters
son, M w anaw uta Ikatu (III,E3), older than the headman, and
leader o f a large, uterine sibling group w hich included five sisters.
M wanawuta Ikatu was the son o f the older brother o f Senior
Headman M ukangala. In 1929 the D istrict Com m issioner made
a road along the C on go B order connecting the Bom a w ith
Solw ezi B om a. This road passed beside N sanganyi V illage
and, nearer the Bom a, came w ithin a m ile o f Senior Head
man M ukangalas capital village. M w anaw uta Ikatu (III,E3)
1 Yet, as we will see, Kamawus brother, a member o f n y a c h im a lineage,
who called Makumela (III,F5), o f k y a n k u k a lineage, his older brother \
married Makumelas sister. This marriage was not regarded as incestuous.
Thus in certain social situations, such as succession to headmanship, the two
village lineages regarded one another as lineally related ; in others, such as
marriage, distinctive matrilinea! origin was stressed. See Chapter Fight for
a full discussion o f the dual character o f Ndembu village structure.





z iz

Schism and Continuity in an fA rican Society

inform ed N sanganyi (III,D4) that he wanted to split off* and

make a village o f his ow n near his father * M nkang ala and closer
to the Bom a.1 He said that he had m any siblings w ith their
children, and that i f he lived nearer the B om a and Hose to
Senior Headman M ukangala, w ho, in 1930, had been elevated
to the status o f a Governm ent Sub-Chief, there m ight be an
opportunity fo r his jun ior kin to find paid w ork. N sanganyi
tried to persuade him to remain, hut he left the village and made
a new settlem ent near the Lwakela R iver bridge on the Govern
m ent road, a m ile from M ukangalas village and fifteen miles
from M w inilunga Boma.
In 1940 M atem pa (HI,D3) also broke aw ay from N sanganyi
and settled on the northern m argin o f the K aw iku Plain. W ith
him w ent his o w n uterine sibling group and all the members
o f n yan ku ka lineage. T he last tw o headmen o f the village
had been members o f that lineage. T h ey could not expect to
supply a third successor, but they objected to the appointm ent o f
N sanganyi on personal grounds.
In 1930, w hen the policy o f Indirect R u le was first introduced,
Nsanganyi had applied to the Bom a fo r recognition as a * Govern
m ent C h ie f. T he D istrict Com m issioner had called in C h ief
Kanongesha and Senior Headman M ukangala for consultation ;
and N sangianyi had adm itted in their presence that form erly
he had paid tribute to Kanongesha through M ukangala, son
b y perpetual kinship, and representative, o f the Senior Chief.
Kanongesha claim ed that although an early N sanganyi had
resisted the efforts o f the war-leader (Kambanji) o f the first
Kanongesha, founder o f Chibwakata V illage, to subdue his
* M bw ela * follow ing, the second Kanongesha had sent his son
M ukangala Kabanda against Nsanganyi. M ukangala had
defeated N sanganyi and com pelled him to pay tribute, half
o f w hich was taken by M ukangala for his ow n use and h a lf o f
w hich w as sent to Kanongesha. In addition N sanganyi had to
give M ukangala a tributary w ife ('ntombu). In compensation
N sanganyi was allotted the ritual office o f Chivwxhanharm or
* em blem -purifier * to M ukangala. T h e incum bent o f this
office was entrusted w ith the medicines o f M ukangalas senior
1 Before Mwanawuta left Nsanganyi there were more than thirty huts in
Nsanganyi Village.

V arieties o f V illage Fission


headmanship and perform ed an im portant role in his installation

ritual. B ut the acceptance o f this office excluded N sanganyi
from political authority over any village except his ow n, in the
indigenous political structure. O ther K aw iku villages recognized
Mukangala, n o t N sanganyi, as their senior headm an. O n
hearing this evidence, the D istrict Com m issioner turned down
Nsanganyis claim for Governm ent recognition and appointed
Mukangala as Su b-C h ief o f the area in w h ich m ost o f the
K aw iku villages w ere situated. It is lik e ly that the possibility
o f Nsanganyis appointm ent as a S u b-C h ief had been instrumental
in retarding M w anaw utas secession from his village, since he
m ight have hoped fo r an appointm ent in the N ative A uthority
bureaucracy i f he had remained. T h e rejection o f N sanganyis
claim was probably an additional factor in his decision to build
near M ukangalas village w here some o f his follow ers w ere later
in fact appointed as coim ciliors in the N ative A u th ority Court.
Matempa (IH,D3) also, w h o became one o f M ukangalas
councillors before he broke aw ay, probably le ft N sanganyi
because he regarded his status as councillor as incom patible w ith
his junior status w ithin the village. For in the village he was
subordinate to a headman w hose claim for an im portant position
in die new political order had been perem ptorily dismissed b y its
T he nuclear group o f M w anaw utas village w as a minimal
lineage descended from his m others m other, Nyachintang*a
(m ,C x). B u t in his village w ere three groups o f uterine siblings,
potential nuclei o f new settlements. There was his o w n uterine
sibling group, the children o f his m other, N yalubenji (IH ,D i).
There was also the sibling group led b y his parallel cousin Chipoya
(III,E5), the children o f his m others younger sister, N yam ahandu (HI,Da). Finally, there w ere the five children o f his w ife
Nyam pupa (HI,Ea), b y a previous husband. N yam pupa and
her brother Kakunda (IH,Ex) had fled originally from Angola
as the result o f a quarrel the nature o f w h ich I was unable to
ascertain, and had been given sanctuary b y N sanganyi. Kakunda
w ent w ith his sister w hen M w anaw uta (III,E3) left N sanganyi,
and neither he nor N yam pupa had any other hom e than M wana
wutas village.
M wanawuta, already quite an old man when he left N san ganyi, had becom e m ore or less senile b y 1947. In that year

2 X4

Schism and Continuity in an A frican Society

John. Kam banji (III,E6), his m other's sisters son, returned from
the Copperbelt w ith some cash savings. H e began to throw his
w eight about in the village and quarrelled w ith Kanem a (III,F2)
and Six (iU,F3), trwo m iddle-aged uterine nephews o f Mwanawuta, w ho had for the last ten years or so undertaken the practical
running o f village affairs. H e also quarrelled w ith Sondash
(m ,F i), the oldest son o f Nyam pupa (III, Ea) w ho b y this time
had becom e the leader o f a thriving uterine sibling group.
Kakunda (III,E i), the brother-in-law o f M w anawuta, was an
old man now , alm ost as decrepit as M w anaw uta (III,3) himself.
In fact, in the village the management o f affairs had fallen into the
hands o f the jun ior adjacent generation to the village founders.
In 1947, Sondash (III,Fi) told M wanawuta and Kanem a (III,Fa)
that he was now an elder (mukuiumpi) w ith m any siblings,
children and sisters children, that he found it im possible to live
in the same village as the obstreperous John Kam banji (III,E6),
and that he proposed to found a new settlement about h a lf a mile
away. Little resistance was offered to his departure fo r he was
not their matrilixieal kinsman, nor was he M w anaw utas ow n
child. Shortly afterwards, John Kambanji, whose older brother
Chipoya (IH,Es) had died several years before, left M wanawuta
Village w ith his three w ives, six children, and tw o sisters
children, and made w hat he called a farm * quite close to
M wanawuta. M w anawutas ow n younger brother Katoka
(III,E4), an old man w ho had quarrelled w ith Kanem a (III,F2),
went w ith John Kam banji. Later, a Lwena Public W orks
Departm ent road-labourer from a neighbouring cam p married
one o f Johns daughters and built a hut at his farm . B u t by
1954 John had lost nearly all his follow ing except his w ives and
children. His sisters daughter married o u t; his sisters son left
him for fear o f his irascibility ; his m atrilineal parallel cousin
Katoka died ; and the P .W .D . labourer also died. It was
hinted in. the villages o f the vicinage that John (III,E6) or one o f
his w ives practised sorcery or witchcraft. T he fact that John
was a blood-brother o f Sandombu o f M ukanza V illage, the
notorious sorcerer (see Chapters Four and Five), probably aid not
w eigh in his favour either.
In 1954 Six (lII,F3) also left M wanawuta and founded a small
form, consisting o f a single K im berley-brick house occupied by
his elem entary fam ily. This last type o f fission is now greatly

1 Cf. p. 36.



on the increase. In the pedicle area a rash o f such farms has

appeared in the last few years. Six has becom e a cash-crop
grow er.1 H e cultivates rice as w ell as taking in land fo r cassava
over and above his subsistence needs. H e sells to the Bom a
' cassava m eal w hich is m ainly used to feed road-maintenance
labourers in a large camp in the vicinity. Such petty com m odity
agriculturalists are tending to break aw ay from their matrilineal
kin in order to avoid the latters claims on earnings. John, too,
although his new settlement was originally an incipient village
by traditional structural criteria, is n ow the head o f a * farm * and
relies on the cultivation o f cash crops and sale o f surplus subsistence crops fo r his incom e.
Since 1947, then, three groups have seceded from M wanawuta
Village. L ittle conflict accom panied these w ithdraw als, for it
was generally recognized that the headman was extrem ely old
and infirm and unable to perform his role effectively. Before
the secession o f Sondash (III,Fi) M w anaw uta V illage contained
about seventy inhabitants occupying some thirty huts. It has
; been previously pointed out that today, w hen the population o f
a long-established village has reached about fifty inhabitants, it
tends to exhibit signs o f im m inent fission. W e have seen that
the social unit w hich m ost frequently form s the nucleus o f a new
settlement tends to be a uterine sibling fam ily, ranging from about
ten to fifteen persons. A village like M w anaw uta, which
; contains three or m ore sibling fam ilies, some o f w h ich m ay be
linked b y a clever leader into a single m inim al lineage, is more
conspicuous for its divisions than for its unity. O n ly the joint
operation o f a num ber o f factors m aking for cohesion (such as the
capability o f the headman, the historical fam e o f the village,
or the existence o f a num ber o f supplem entary ties o f kinship,
affinity, generation affiliation and friendship, w hich cut across
lineal attachm ent) can hold together such a congeries o f virtually
independent groupings. M w anawuta V illage possessed few o f
these countervailing tendencies : the headman had little personal
authority, the village was but recently established, it was situated
about sixteen miles from the nearest European centre, and its
com ponent sections w ere interlinked b y few o f the ties mentioned.
Sondash (III,Fi) and his siblings w ere step-children o f the

3 G G O G G


@ @ @ O O 8 O O O G S 0

Varieties o f Village Fission


Schism and Continuity in an African Society

headman and unrelated by consanguinity to the other village

members ; one marriage only had taken place between members
o f his and the headmans matrilineal kin. John Kam banji (III,E6)
had spent m any years on the Copperbclt and his experience and
interests w ere different and opposed to those o f the high ly con
servative old headman. He was in the senior genealogical
generation to Sondash and as the result o f some early quarrel
bitterly disliked him . Six (III,F3), although he was the uterine
nephew o f the old man, had also been a labour m igrant and saw
his future in terms o f personal advancement in the new cash
econom y. In fact Kanem a (IIX,F2) was the only senior kinsman
o f M wanawuta (III,E3) w ho had remained w ith him at the time
I left the field, and he had hopes o f succeeding to the headmanship
o f the depleted village, according to informants w ho knew the
situation. T he story o f the foundation, rise and decline o f
M wanawuta V illage is typical o f the fate o f m any Ndem bu
villages in the past tw enty years as the cash econom y has in
creasingly penetrated the traditional social system and corroded
first the links o f classificatory, and then o f prim ary, matrilineal
T he subsequent career o f M atempa V illage parallels in many
respects that o f M w anaw uta. A fter Matempas death his uterine
nephew Nswanam atem pa (III,E8) succeeded him and for a time
the village maintained its unity and increased in membership.
B ut in 1952, w hen Governm ent delim ited a new Forest Reserve,
w hich included the K aw iku Plain and most o f the hush adjoining
it, Nswanamatempas people, along w ith N sanganyi, C hibwakata, and other villages, w ere m oved into the vicinage to which
M ukanza, Nswanakudya and N g om bi belonged. T he leaders
o f tw o sections w ithin the village took this opportunity o f splitting
o ff from Nswanamatempa V illage w ith their follow ings. The
senior elder o f nyanktjka lineage was M akum ela (III,F5), a man
w ith eight livin g children, and he founded a farm , in reality an
incipient traditional village in social structure, w ith tw o younger
brothers, his m others sisters son, their children, and a number o f
his sisters children. His sister was the w ife o f the younger
brother o f K am awu (III,F4), leader o f the other seceding group,
and she lived virilocally w ith her husband. K am aw u was the
sisters daughters son o f Matempa (III,D3). His follow ing
basically consisted o f the sister o f his deceased m other and his

V arieties o f V illage Fission



mothers brother, his older and younger uterine brothers, his

sister, and his and their children. K am aw u had becom e headman
o f this group, in spite o f his lo w seniority, on account o f his
w idely recognized qualities as a man o f legal skill, as a hospitable
and conciliatory person, as one possessed, in fact, o f all the
virtues regarded as desirable in a headman (see p . 200). His
mothers brother Biscuit (III,E7) had returned as an old .man
after an absence o f m any years at B ulaw ayo, and had taken
up residence w ith his sister at K am aw us settlement. A s a
w idow er, he found it advantageous that the old w id ow was
w illing to co ok fo r him . H e was a sad-eyed hum orist w ithout
ambition, his on ly requirement a steady supply o f beer. M akumela built his farm facing Nswanam atem pa on the other side o f
the m otor road, and K am aw u established him self a hundred yards
from M akum ela on the same side o f the road. A ll three headmen,
Nswanamatempa (III,B8), M akum ela (IH.Fs) and Kam awu
(m ,F4), had asked M ukanza Kabindas perm ission to build in
that locality, fo r M ukanza was generally reckoned to be
mwenimbu, the headman longest established in that part o f the
vicinage. T here w ere several abandoned village and garden
sites and three separate graveyards o f the Kahali-M ukanza village
lineage near the three new settlements, and the ownership o f these
gave M ukanza his righ t to be reckoned mwenimbu. O ver the
w hole vicinage N sw anakudya (II,D io) was recognized as pos
sessing m oral authority since he belonged to the chiefly lineage,
but Nswanakudya had no say in the allocation o f land that had
once been occupied b y M ukanza V illage.
Nswanam atem pa (IH,E8) had the nam e o f being one o f the four
most notorious sorcerers in the vicinage. H e was thought to
have caused the deaths o f his tw o brothers, o f a sister, and o f
Kam awus m others brother and m others sister, b y his medi
cine *. This, alleged Kasonda m y henchm an, w as the * real
reason w h y M akum ela (HI,F5) and K am aw u (HX,F4) split o ff
from his village. K am aw u and M akum ela both told me,
how ever, that all three groups rem ained friendly and recognized
Nswanamatempa (III,E8) as their superior. In favour o f their
autonom y they argued that * the big village was finished n o w V
1 M eaning that the type o f large village containing a large group o f m atrilirteal kin had been supplanted by the sm all farm inhabited by close kin.


Schism and Continuity in an A frican Society

that they w ere elders (akulumpi), and that each had a large
follow ing o f his ow n close relatives. It is probable that both
rationalizations for secession were actually advanced b y them , the
form er privately, the latter in public ; but that really the prime
cause o f their secession was am bition to becom e leaders o f local
groups, am bition w hich had been given its opportunity to succeed
b y m odem developments.
Thus in the course o f tw enty-five years (1929-54) N sanganyi
V illage has given rise to tw o villages b y prim ary fission, M wanaw uta and M atem pa ; w hich in their turn have throw n o ff five
further settlements, Sondash, John Kam banji, Six, M akum ela and
Kam awu.
T he history o f this little group illustrates a num ber o f points
previously made. T he decline in political im portance o f Nsanganyi V illage under the British Adm inistration seems to have
weakened the bonds that held this form erly large and hetero
geneous village together. B ut m any o f the headmans character
traits w ere not such as to commend him to Ndem bu. The
leading elders o f his village blamed him for not obtaining the good
graces o f the Bom a, pointing out that another senior headman o f
M bw ela origin, Sailunga, had been persuasive enough to secure
his ow n appointm ent b y the D istrict Com m issioner as * Para
m ount C h ief* o f the Lunda-Kosa to the east o f the Lunga R iver.
T hey said that i f Nsang*anyi had been firm er and m ore eloquent
he m ight have becom e a Sub-Chief, and all his relatives w ould
have benefited. H e was garrulous but incom petent in discussing
village cases. H e was also m uch addicted to drinking. W hen
I knew him he was regarded as a com ic, but rather pitiable, old
man. He used to have his xylophone and slit gong, emblems
o f a departed prestige, played at night in order to rem ind his
departed kin that * he was hungry and thirsty * and that it was
their duty to help him . Although he was often neglected in
secular contexts, Nsanganyi was still, however, accorded con
ventional respect at rituals in K aw iku villages, w here he was given
beer and food before other K aw iku headman. H e continued to
preside over the installation rituals o f incumbents o f the
Mukang*ala senior headmanship, and to receive presents for it.
B ut his glo ry had clearly gone, leaving only decrepitude.
The cum ulative pressures o f social and cultural change are



1 J


Schism and C ontinuity in an African Society

im portant factors behind these successive cleavages. W e have

already seen h ow in M ukanza V illage m uch o f Kasondas and
Sandombu s behaviour is explicable in terms o f response to new
incentives, and h o w the abolition o f slavery b y the British gave
Kafiim bu the opportunity to secede from M ukanza Village.
Here again w e find that on the one hand direct political factors
such as the refusal o f the Governm ent to recognize N sanganyi
as a Sub-Chief, and, on the other hand, the pervasive influence
o f the cash econom y, have led to the gradual reduction o f a large
village into smaller and ever smaller components, m ost o f which
w ere hardly m ore than groups o f uterine siblings, and in one
case, a single elem entary fam ily. In. Chapter T w o it was sug
gested that in the latter part o f the nineteenth century some
N dem bu villages w ere quite large and others small,1 both types
representing different responses to the slave trade, and, later, to
slave raids. The large villages w ere often com posite, containing
several unrelated matrilineages o f relatively w ide span w ho had
settled together both for m utual protection against slave-raiders and
in order to m ake successful raids on other villages. T h e small
villages m ay have represented another w a y o f escaping the raiders
b y breaking up ana scattering before them into the deep bush.
In the present century, under British rule, w ith the abolition
o f the slave trade and o f internal slavery, w ith the slow addition
o f a cash to a subsistence econom y, and w ith a rapid post-war
increase in the rate o f labour m igration, the large villages
have tended to break up b y successive stages o f fission into the
smallest possible matrilineal units, uterine sibling families. B ut
change in the structure and stability o f residence has b y no means
everyw here and invariably taken place at the same rate and in the
same w ay. In a single area one m ay find traditional matrilineal
villages side b y side w ith m odem fam ily forms. N ext to diem
one m ay find settlements w hich m ay he regarded as transitional
types, consisting o f depleted uterine sibling groups. Tip-service
is still paid b y the m ajority o f N dem bu (despite the opinions,
quoted above,8 o f M akum ela and K am aw u), to the value o f
livin g in a large village o f matrilineal kin. T he members o f some
villages, for instance o f M ukanza V illage, until very recently have
deeply respected this value. B u t w ith the breakdown o f the
1 Cf. pp. 43 - 4 -

3 See p. 217.

V arieties o f V illage Fission


traditional authority o f the old m en o f the senior genealogical

generation in a village, and the em ancipation o f the younger men
from this authority, the norm s governing behaviour between
generations have becom e enfeebled : younger m en o f w ealth
and am bition no longer feel constrained to obey the w ill o f the
headman and his generation mates, but assert their financial and
political independence b y seceding and founding farms. T o
some extent the values inherent m m atxiliny as a principle govern
ing the com position o f residential groups still p ersist: uterine
ablings, that is, members o f the same genealogical generation,
between w hom relations exist o f approxim ate equality, still tend
to build together. B u t today, especially in the northern pedicle,
w ith its grow in g com m itm ent to a cash econom y, even the links
between siblings are becom ing m ore tenuous, and individuals
are founding farm s w ith their ow n elem entary fam ilies.
Nevertheless, although m odem developm ents have undoubtedly
accelerated the rate o f fission and perhaps reduced the span o f
the seceding m atrilineal group, these developm ents have only
recently, and in particular w ell-defined areas, produced the
spectacular atom istic effects described above. In the past, as to
day, the N dem bu lineage seems to have been shallow b y com
parison w ith the lineage am ong such peoples as the Tallensi
and the A sh an ti; and fission seems to nave taken place w ithin
the m inor or m inim al lineage rather than dichotom ously between
structurally balanced and opposed m ajor segments o f a village
lineage. V illages appear to have been small and spatially m obile
in the past also, as Livingstone observed. I f older informants
are to be believed, N dem bu villages, before the epoch o f slave
trading and slave-raiding, w ere hardly m ore than small hunting
camps ; so that w ith regard to m agnitude, i f not to social struc
ture, m odem * forms * resem ble ancient villages. B u t the social
core o f the * form * is, as w e have seen, an elem entary fam ily,
w hile the core o f an ancient village was probably a group o f
uterine siblings.
In the exam ples o f village fission given above and from the
tables constructed on the basis o f genealogical data, the uterine
sibling group appears as the basic unit o f secession. B ut the
seceding groups are not necessarily segments o f a local m atrilineage, since die group m ay be variously attached to its leader.


Schism and C ontinuity in an A frican Society

KLafumbus (I,G i) follow ers w ere a group o f uterine siblings and

their children, related to him b y ties o f m arriage and patrilateral
cross-cousinship. T he group led aw ay b y Sondash (III,Fi)
consisted o f stepchildren o f the village headman. Makumelas
(m ,F5) follow in g w ere not m atrilineal kin o f Nswanamatempa
(HI,E8), although M akum ela and Kam awu (IU,F4) called one
another respectively * younger * and older brother \ Appendix
in, in fact, shows that K a m a W s uterine brother had married
M akum elas sister. This marriage was not regarded as incestuous,
since the people o f K am aw u said, * W e have a different matrilineal
ancestress (nkakuluta) from M akum ela. Sometimes, it is true,
segments o f equal span broke aw ay from one another. If, for
instance, in M ukanza V illage, malabu lineage and nyachintang a
lineage build in separate settlements, this w ill represent fission
between tw o m inor lineages. A t a low er level o f segmentation
the secession o f M atem pa (IH,D3) from N sanganyi V illage and
o f John Kambarqi (UI.E6) from M wanawuta, w ere divisions
between tw o uterine sibling groups belonging to the same
m inim al lineage, divisions between the children o f tw o sisters.
T he most com m on unit o f secession is indeed the uterine sibling
group led b y a prim ary o r classificatory sisters son o f the village
headman. M w anawutas secession from N sanganyi V illage is
an exam ple o f this kind o f fission, as is Kam awus from Nswanamatempas V illage.
Ndem bu society, unlike T ale society, is not, in fact, * built up
round the lineage system
A fu lly developed ram ifying lineage
system requires as a prim e condition for its grow th an ecological
system in w hich lim ited access to resources is characteristic.
Seen in structural terms, it also requires that the m ode o f postm arital residence should be consistent w ith the m ode o f descent.
A system o f patrilineages, i f it is to be the skeleton o f the social
structure , requires virilocal marriage ; a system o f m atrilineagcs requires uxorilocal m arriage. T he patricentric fam ily
in the form er is w hat Professor Fortes calls * the grow ing tip *
o f the lineage system, and like the lineage is attached to a locality.
In m atrilineal uxorilocal societies the local m atricentric fam ily
is such a grow ing tip. I f fem ale members o f a matrilineage
remain in one settlem ent and im port their spouses, each wom an
becomes the source and nucleus o f a local lineage w hich m ay
attain considerable depth in a society where settlement is anchored

V arieties o f V illage Fission


to lim ited resources. In the next chapter I propose to discuss

the structural im plications o f virilocal m arriage w ithin and
between settlements in N dem bu society in considerable detail.
But some o f the consequences o f this antithesis between mode
. o f descent and m ode o f post-m arital residence have already
: become apparent in the last three chapters. In m any o f the
: simpler societies the tie between m other and children seems to be
the closest bond o f kinship. In societies w ith extended kinship
systems, there is a tendency fo r members o f the m atricentric
fam ily to gravitate together in co-residence how ever far apart
they m ay have been throw n b y other social tendencies, such as
virilocal m arriage am ong m any W est-Central Bantu peoples.
It is perhaps a tacit recognition o f the strength o f this bond that
has given rise in m any strongly patrilineal societies to institutions
directed against divorce,1 since the patrilineage o f the father has
a pow erful interest in retaining his children to replenish its local
membership, and there is a danger that w here the m other goes
her children w ill g o also. In N dem bu society, virilocal mar
riage, coupled w ith and opposed to m atrilineal descent, has the
effect o f weakening the local developm ent o f deep lineages ; and
at the same tim e it makes every fertile w om an w ho is livin g w ith
her husband the potential starting-point o f a totally new settlem ent
o f prim ary m atrilineal kin. I f for some reason a w om an is
unable to return to her ow n m atrilineal kin, it is likely that when
they mature her children w ill start a settlement o f their ow n.
Thus w hen slaves became emancipated, the children o f slave
wom en am ong N dem bu tended to form their ow n villages.
The m other o f Sondash could n ot return to A ngola, and her many
adult children, whose leader had been unable to succeed to headmanship in his stepfathers village, split o ff from M wanawuta
Village and made their ow n settlem ent. A gain, due to virilocal
marriage, a w om ans children are frequently brought up in their
fathers village, am ong his m atrilineal kin, and n ot in close
everyday contact w ith their o w n m atrilineal kin. Since they
do not interact intensively w ith the latter, they do not develop
strong sentiments o f m utual interdependence w ith them. Each
1 See G luckm ans argum ent on this p o in t in * K in ship and M arriage am ong
the L o zi o f N orth ern R hodesia and d ie Z u lu o f N atal \ in African Systems
ofKinship and Marriage, ed. A . R . R a d cliffe-B ro w n and C . D . Forde (1950).








Schism and C ontinuity in an A frican Society

uterine sibling group tends to regard itself as autonomous. In

their fathers village uterine siblings are outside the matrilineal
descent group w hich holds office. T h ey do not interact in the
affairs o f daily life w ith their ow n matrilineal kin from whom
they m ay be geographically separated by a considerable distance.
Even when they return to the matrilineal kin o f their mother,
either in a com pact block after the divorce or w idow hood o f the
latter, or singly and severally in later adult life, the ties inter
linking them am ong themselves are o f a m ore durable and
pow erful character than those attaching them to m ore distant
matrilineal k in -e v e n to the children o f a m others uterine sister.1
A number o f com pensatory structural devices exist in the social
system to prevent the secession o f uterine sibling groups. Local
exogam y, w hich is largely responsible for the initial cohesion and
autonom y o f such groups, later weakens their cohesion b y dis
persing their fem ale members, as they grow up, am ong the
villages o f their respective husbands. T he principle o f classificatory adelphic co-residence, w hich associates male matrilineal
kin o f the same genealogical generation, establishes bonds o f
co-operation between men belonging to different uterine sibling
groups and opposes them to members o f adjacent genealogical
generations. T he children, and sisters and brothers children,
o f the head o f a group o f uterine siblings are classed as members
o f a generation opposed to that o f their parents. T he uterine
sibling group begets the means o f its ow n potential dissolution,
since wom en m ay leave their brothers to accom pany seceding
sons. I have stressed the importance o f the headmans reputa
tion for justice and generosity as a means o f holding together
a settlement full o f potentially disruptive tensions between its
com ponent uterine sibling groups and adjacent generations. The
astute arrangement o f marriages between his ow n and his sisters
children, and between members o f his ow n and the ju n ior alter
nate generation in a village, is another means b y w hich a headman
endeavours to maintain the continuity and integration o f his
follow ing.


g ':

1 Unless they have been reared in the same village as the children o f a
mothers sister. If members o f a minimal matrilineage have been reared
together they tend to form a potential unit o f secession, e.g. the children o f
Manyosa and Chawutong'i in Mukanza Village.

V arieties o f V illage Fission


Thus the unity o f a village at any one tim e tends to depend, on

a precarious balance between different categories o f conflicting
alignments. Persons united as members o f a single uterine
sibling group are divided b y virilocal m arriage, and united w ith
members o f other uterine sibling groups b y membership in a
com m on genealogical generation. M ale members o f such a
sibling group are opposed to their o w n children w h o belong to
a different sibling group, belong potentially to a different village,
and belong actually to a different genealogical generation which
1 1 *
* ary and classificatory siblings and
Fem ale members o f a uterine
sibling group are opposed through m arriage to their brothers.
N orm ally, i f their marriages are to rem ain in being they must
live virilocally. Thus they set a lim it to the ambitions oF their
brothers, w ho m ay wish to form a n ew village w ith their help,
and in any case require the support o f their children to further
intra-village am bitions. T h ey are opposed to their brothers*
children if they live in the same village ; for their brothers tend
to favour ow n children w ith food and attention rather than sisters*
children, w ho w ill ultim ately oust brothers* children from the
village. Sisters are united w ith their brothers as members o f
the same genealogical generation, and w ith their other male and
female classificatory m atrilineal kin, against all members o f the
adjacent generations. In certain situations mothers are opposed
to their ow n adult children, w h o, in seceding from the villages
o f their husbands or brothers, disrupt and weaken those villages.
W hen, for exam ple, Sondash m ade his ow n 4farm *, his m other
remained w ith her husband, the o ld headman M w anaw uta ; and
she scolded her son fo r * spoiling the village
She also scolded
her brother, w ho w ent w ith Sondash because he was given the
office o f tnulopu or second-in-authority to Sondash, although he
was too old to perform the duties o f a headman effectively
Thus persons are interlinked in one set o f relations w ith persons
to w hom th ey are opposed in others. A lthough these cross
cutting loyalties restrain and m itigate the pow erful tendencies
towards fission in the residential unit, the uterine sibling bond in
its closeness and exclusiveness still tends to assert itself at the
expense o f all countervailing influences. T he nuclear group in
alm ost every instance o f fission cited is clearly the group o f


Schism and Continuity in an A frican Society

uterine siblings and their children, although in m any cases not

all the living siblings o f the leader o f the seceding group have been
w ith him at the point o f fission. Sometimes one or more o f
the leaders sisters have been livin g w ith their husbands in other
villages ; sometimes a brother has been elsewhere. T w o o f
Matempas (III,D3) sisters remained at N sanganyi and would
not com e w ith him . Kafum bu, w ho founded a village w ith his
w ifes siblings, could not get his w ife s younger sister to leave
Sakazao (fH p) her husband, o f M ukanza V illage, nor his w ifes
brother Kayineha (I,G6), w ho preferred to live uxorilocally at
M ukanza Village. Another o f Kafum bus w ifes sisters accom
panied her husband Yim bw endi (I,Fi i) to the village Yim bw endi
founded after his quarrel w ith Kafum bu (I,G i). B ut in time
most o f the siblings o f such a leader w ill com e to his settlement.
W hen his sisters are divorced or w idow ed they w ill jo in him
w ith their children. The strong tie between uterine brothers
also asserts itself. Gradually the uterine sibling group drifts
together. O ften the children o f sisters w ho did not secede with
their brother, jo in him when they grow up boys w hen they
m arry and girls after divorce. Thus the children o f Matempas
sisters w ho did not accom pany him later built huts in his village.
T h e structural principles, then, w hich govern residential
affiliation are manifold, com plem entary, and also conflicting.
M atriliny is in the final analysis dom in an t; but a peculiar cast
is given to m atriliny by viriiocal m arriage, w hich reduces the
span o f the effective matrilineal unit to the m atricentric fam ily.
That the system is able to persist depends upon a number o f
com pensatory principles : affiliation b y genealogical generation ;
opposition o f adjacent, and alliance o f opposed, genealogical
generations; and cross-cousin and grandparent-grandchild
marriages. These are all means o f interlinking matricentric
families in a wider system o f local ties and o f reducing the
strength o f uterine siblmgship. It is im portant to note, however,
that the effectiveness o f any one o f these principles varies from
settlement to settlement and depends on such local factors as the
length o f establishment o f the settlement, the fertility o f wom en
and men, the strength o f marital ties, the reputation and astuteness
o f the headman and o f candidates for headmanship, the age and
experience o f these candidates, the local numerical strength o f

V arieties o f V illage Fission


sibling groups and m inim al lineages, and so on. Social change

J produces alterations in die relative w eightings o f these principles
S and gives rise to n ew types o f interpersonal and inter-group
relationships. For instance, ties between members o f an elemen
tary fam ily are strengthened and ties between m em bers o f a local
maternal descaat group are weakened. Kinds o f struggle arise
within the village against w hich the traditional m achinery for
reintegrating a disturbed group m ay prove ineffective.
I am aware that w hen I w rote that * virilocal m arriage reduces
the span o f the effective matrilineal unit \ I was g u ilty o f over
sim plification. In practice, the com plex interaction o f a number
o f ecological and political factors, and n ot virilocality alone, is
responsible fo r the reduced span o f the m atrilineal unit. The
same set o f factors, w hich weakens the effectiveness o f m atriliny
as a bond o f local affiliation, assists, i f it cannot be shown to
determine, the em ergence o f virilocality as a com peting principle
o f residential attachment.
In societies governed by matrilineal descent, wide-span. local
matrilineages tend to be found w here access to land or other
valuable natural resources is lim ited. Such extended matri
lineages tend to be associated w ith the possession o f fixed estates.
In the natural region inhabited b y N dem bu, cultivable land is
relatively abundant, w hile population density is lo w . Social
groups are not constrained b y natural exigency to remain
jerm anently in particular areas, in pockets and tracts o f fertile
and. Indeed, since hunting is a h igh ly valued activity, and
game tend to m ove aw ay from hunters, there are positive
econom ic inducements to change the residential site periodically.
These inducements are all the m ore effective for a population
that does not invest in large, permanent houses, in long-grow ing
and long-bearing trees, and in stumped land. N o r are Ndembu
faced w ith the necessity o f settling near lim ited water-resources,
for their territory adjoins the C ongo-Zam bezi watershed and
m any streams and rivers take their rise in it.
Nevertheless, given abundance o f land and lack o f investment
in im m ovable property, it w ould still be theoretically possible
for a large m atrilineal descent-group to m ove about together,
and n o t to split up into smaller units. W h y does this seldom
happen ? Part o f the answer lies undoubtedly in the individual
istic productive system. The lim its o f econom ic co-operation,








Schism and Continuity in an A frican Society

as w e have seen, are extrem ely narrow. M en hunt alone or in

small bands. Com m unal working-parties to clear bush and
hoe up mounds are infrequently mustered, and m any Ndem bu
do not sponsor them. T he elem entary or polygynous fam ily
can, and today often does, satisfy m ost o f its needs b y its own
labour. N o t o n ly is access to resources virtually unrestricted,
but there is no concept o f a jo in t estate w orked collectively by
a unilineal kin group. N o one has a perennial interest in any
one tract o f land, owned jo in tly b y his lineage.
Furtherm ore, Ndem bu do not possess large herds o f cattle,
the jo in t ownership o f w hich m ight under certain circumstances
hold together a large nom adic group o f unilineal kin. The
m ost valuable item o f m ovable property they possess is the
m uzzle-loading gun, and o n ly a m inority o f m en ow n efficient
guns. Since such guns are scarce they are h igh ly esteemed and
are h o tly com peted fo r in situations in volvin g inheritance. They
tend to divide rather than unite matrilineal kin.
T he true problem , indeed, for Ndem bu society is not w h y local
matrilineages are small, but rather w h y they exist at all. W h y
do N dem bu live in villages, die cores o f w hich are male matri
lineal kin, and not in fam ily homesteads ? In the past, at any
rate, as w e have seen, one o f the reasons was provided b y the
necessity to unite in strong defensive groups against slaveraiders. O ne response to slave-raiding was to build palisades,
throw up earth-ramparts, and dig ditches around large villages.
Such villages w ere, in effect, fortresses, in a state o f perpetual
vigilance against siege. T h ey seem to have been attached to die
same tracts o f land for long periods. T heir male members must
often have co-operated to repel raiders, and to defend their
wom en, children and gardens.
O n the other hand, an alternative response to slave-raiding
was to scatter in small groups into the hush, each group holding
itself in readiness for instant flight. Y e t each such group,
according to tradition, had a nucleus o f uterine or matrilineal kin.
W h y w ere such groups w ider than the fam ily ?
N dem bu have pointed out to me the practical advantages o f
livin g in villages. T hey say that w hen one fam ily-head is short
o f food, especially o f meat or fish, supplies o f w hich are. sporadic
and uncertain, he can obtain it from a village kinsman or neigh
bour. Later, when he * finds meat * he can m ake a return. In

V arieties o f V illage Fission


other w ords, livin g in villages under a subsistence econom y raises

the general level o f consum ption and safeguards against individual
N dem bu have also told m e that before the Europeans came
N dem bu used to raid each other to obtain captives w hom they
could sell to O vim bundu traders as slaves in return for guns,
pow der and d oth . A large village was a m ore effective raiding
and defensive unit than a sm all one in those troubled times.
Then as n ow the local tie between uterine brothers was a very
dose one. A brother w ould help one in raids, and i f a man was
taken captive, it was his brothers duty to recover him , b y force
or paym ent, i f he possibly could. T he Icdandic m axim , which
Dasent made the m otto or his translation o f N jals saga, w ould be
equally appropriate to N dem bu circumstances : bare is back
w ithout brother behind it. A gain, i f a man w ere killed, it was
his brothers duty to avenge him o r to exact blood-com pensation.
As w e have noted on m any occasions, i f uterine brothers are
to remain together, they m ust m arry virilocally.
W e m ust regard matrilinea! descent itse lf as given in this
sodety. M atrilineal succession and inheritance are regarded as
axiom atic b y Ndem bu. M atriliny is sym bolized in m any kinds
o f ritual, and is supported b y the dogm a o f descent ; and its
im portance is emphasized in several folk-tales. I f male kin
resided together fo r defence, offence, and m utual econom ic aid,
the m ajority o f them w ould have had to be uterine and matri
lineal kin. It is on ly the small span o f the effective local matri
lineal unit that w e have to consider. It is probable that the
external provision o f law and order b y the British authorities
has contributed to the present small size o f local matrilineages.
Large units fo r defence and offence in raids are no longer neces
sary. E ven before the introduction o f cash, that solvent o f
corporate kinship groupings, the individualistic tendencies in
production m ust have resulted in the breakdow n o f large local
groups. These, as w e have noted, had no permanent jo in t
estates o r com m unal m ovable property to keep them together.
I f a m an wished to secede from a village w ith his fam ily and kin
w ho w ould fo llo w him , he could alw ays obtain residential and
agricultural land elsewhere.
These factors, then abundance o f land, individualism in
production, lack o f investm ent in valuable fixed resources, and


Schism and Continuity in an A frican Society

external provision o f law and order do not provide favourable

condition^ fo r the form ation o f wicle-span localized matrilineages. W hen, in addition, Ndem bu observe the custom o f
virilocal marriage, a positive tendency m aking fo r the reduction
o f the span o f the matrilineal unit m ay be said to be at w ork.
It is o n ly b y describing and analysing a num ber o f cases o f
village fission in terms o f the social drama, or, w here there are
insufficient data, in terms o f the case history, that the precise
effects o f these principles and o f social change in any given
situation can be accurately assessed. For exam ple, w hen one
examines the com position o f the seceding and rem aining sections
involved in village fission, one seldom finds that any one principle
o f social organization has clearly prevailed over the rest. In
Nswanamatempas V illage one m ight have expected that Biscuit
(UX,E7), uterine uncle o f Kam awu (M ,F4), and not Kam awu,
his nephew, w ould have becom e the headman o f the n ew settle
ment, since uterine uncles have authority over their nephews.
B ut w hen the particular circumstances o f the case are considered,
Biscuit is found to be a gentle, irresponsible drunkard, whereas
Kam awu is a respected ihaku (law-man) w ith m any siblings and
children. These factors outw eigh the values set on age and kin
ship position as criteria influencing selection for political office.
A m bition and ability becom e significant in any given instance,
and have repercussions on the social structure o f the local unit.
In the social drama w e can observe h o w particular individuals
manipulate the organizing principles o f social affiliation to serve
their ow n purposes. Sandombu, although childless and w ithout
fertile siblings, exploited cleavages in other sections o f M ukanza
Village and invoked the principles o f generation affiliation and
the m other-child tie, to build up a follow ing. B y his hospitality
to distant kin and strangers, he made apparent to all that he w ould
give protection and assistance to persons w ho, fo r whatever
good reason in N dem bu opinion, had been excluded from their
ow n villages. Social prestige and authority in societies based
on kinship principles are largely a function o f the relative fertility
o f individuals and families. Sandombu was unlucky in this
respect, and i f he had been a less forceful character, he m ight w ell
have resigned him self to be a man w ithout a follow in g, a social
nobody. B u t he was resolved to be a headman, and consequently
had to utilize other means o f obtaining a follow in g than those

Varieties o f V illage Fission


which lay to the hand o f any fertile man w ith a fertile sibling
group. In doing so he made m any enemies since he was com
pelled to surround him self w ith strangers, social undesirables, or
to attem pt to w in over b y hospitality those w hom m ore fortunate
men w ould have regarded as their * natural ' supporters-close
N um erical analysis tends to ignore as irrelevant the unique
features o f each instance o f fission and to stress regularities, the
statistically norm ative pattern. Y e t it provides the background
against w h ich each social drama assumes a significance. I f one
finds that the m ost com m on unit t>f fission, fo r instance, is
the uterine sibling group, and then, in a given case history or
social drama one discovers that the unit o f fission is the minimal
lineage o r the elem entary fam ily, one is led to inquire into the
reasons fo r this apparent anom aly. I f the data relevant to the
analysis are collected in terms o f the social drama, the lines o f
alliance and cleavage in the specific village becom e visible ; and
attention is focused on the particular norm s w hich are observed
and broken, on the actual m otives w hich guide the behaviour
o f the participants, and on the detailed econom ic, political and
other interests w hich unite and divide them. Apparent excep
tions to structural regularities discerned in the num erical data
are som etim es found to be the product o f a com bination o f
factors w h ich taken alone, or in their unhampered expression
in different situations, also possess regularity. Thus a series o f
marriages w ith in perm itted categories o f kinship betw een tw o
uterine sibling fam ilies or between tw o m inim al lineages m ay
decide w hether these groups secede o r remain together in a
particular drama o f fission. Indeed the dynam ic interaction o f
specific persons and groups in the process w hich I have called
the social drama falls w ithin the province o f the sociologist no
less than the analysis o f the statistical and ideal norm s o f social
structure. In the social drama w e see social structure in action.
O ur attention is draw n to the anomalous isolated instance, the
apparent exception to statistical regularity, ju st as m uch as it
is directed towards the manifestation o f that regularity. In
consequence, w e are led to inaugurate a series o f enquiries into
the nature o f the social mechanisms w hereby regularities are
maintained, and into w hat happens w hen regularities are broken.
W e are led to detect at ju st w hat points the fabric o f the social










Schism and Continuity in an A frican Society

structure is weakest, and w hat means are taken to solve problems

o f social integration posed by that weakness.
As an exam ple o f this social handling o f structural vulnerability,
let us take the recurrent situation o f the exceptional persons or
groups w h o transfer loyalty to the faction that opposes their own
prim ary kinship allegiance in a conflict w hich terminates in
fission. In M ukanza Village, Kayineha (I,G6), brother o f
Kafum bus w ife, was a case in point. W e w ere led to inquire
w h y he did not secede w ith his other siblings, and biographical
data w ere presented w hich showed that he was a personal friend
o f M ukanza Kabinda, and that he had married N yam w ahas
daughter. W e also found that in the fight between the Kafum bu
and M ukanza factions, Kayineha acted as peacemaker. Later
he assumed an intercalary role between the tw o settlements o f
M ukanza and Kafum bu. W e found that Sondashs (III,Fi)
m other, w ho did not secede w ith her son, acted in a similar capacity
as a lin k between M wanawuta and Sondash villages. M atempas
(ni,D 3) sisters w ho remained behind at N sanganyi but whose
children later w ent to their uncle, also occupied such an intercalary
position, as did M wanawutas full brother w ho seceded to Johns
In fact, persons w ho do not conform to the rule, m ay thereby
acquire structural importance. W ithin the system o f internal
village relations these persons appear to he the exceptions to the
rule that uterine siblings secede together. B ut a study o f their
behaviour at the point o f fission leads to a retrospective inquiry
into their m otives for action, and an exam ination o f their subrole as intermediaries between groups to w hich they are con
nected by ties o f close kinship, and groups w here they actually
reside. W e learn that they perform an essential function in the
w ider system o f inter-village relations, preventing the total
e s t r a n g e m e n t o f groups divided initially in anger.
Thus apparent
exceptions to statistical regularities obtained from genealogical
data on village fission prove to be themselves regularities within
a w ider system o f social relations.
V illage fission w hich divides the basic unit o f setdement, the
village, tends on the whole to contribute to the integration o f
the w ider system, the N dem bu socio-geographical region, by
the provision o f links o f maternal kinship w hich jo in vicinage
to vicinage, and senior headmans area to senior headmans area.

V arieties o f V illage Fission


The w ider system o f social relations gains at the expense o f the

M ale kin form the residential core o f villages : in a matrilineal
society such m ale kin must im port their spouses or separate from
one another : virilocal marriage in a spatially m obile society
prevents the form ation o f deep lineages and gives a high degree
o f cohesion and autonom y to the m atricentric fam ily w hich later
becomes the principal unit o f fission. T h e looseness o f ties
between m atricentric families leads to a high frequency o f fission,
and their small size determines the small size and large num ber o f
settlements. Within the village other principles o f social organ
ization than the lineage become im portant, notably the opposi
tions and alliances o f genealogical generations : between villages
relations o f spatial propinquity and affinity becom e m ore heavily
weighted than relations o f lineal kinship as factors m aking for
cohesion in a vicinage. Ties o f m atrilineal kinship, on die other
hand, are utilized to interlink separate vicinages. E cological
and structural factors, interlocking in a com plex fashion, jo in d y
determine the form o f setdement, the m ode o f its fission, and the
w ay in w hich setdements are interlinked w ithin the w ider social
systems o f vicinage and tribe.






HE first consequence...o f. jririlocal m ariiage_in Ndem bu
residential structure is that the m ajority o f village .members belong, to. a singlebilateralextended.faiiiily, g o f
the Jheadman, h is. w ives and cHldren^ and ofJiis-siblings,, their
es and childxen. There is no w ord in Lunda for * elem entary
y * but the term ntanga connotes a group o f bilateral kin
and their in-laws w h o are dependent on som e particular person.
Ntanga, the singular, or antang a, the com m only used plural
form , is always used w ith a possessive pronoun, as in ntang*ayami
or antang*ajami, * m y dependent kin group * o r m y dependent
kin (including in-laws) \ Its use can be extended to include all
ones kin w ith w hom one can trace genealogical connection,
but in practice the term usually refers to ones local kin-group
and is restricted to kin and in-laws w ith a jun ior status to oneself.
Thus in M ukanza V illage, all the villagers belong to M ukanza
Kabindas ntanga. Kasondas ntanga includes his w ife, his
minimal lineage sisters, and his and their children and grand
children. It does n ot include his older sisters husband w ho is
an older man than Kasonda and tends to m ove his residence
between his ow n farm in Kanongesha Area and M ukanza V illage.
Sakazaos ntanga comprises the w hole o f malabu lineage, and
the spouses and children o f lineage members. Consequently his
moral authority over ju n ior kin overlaps w ith that o f M ukanza
Kabinda, since bis sister N yam ukola is M ukanza Kabindas w ife,
and M ukanzas children b y her are his sisters children. The
ntanga is n ot th erefo rean _independent,_clearly demarcated
corporate unit, but varies-feom.senior..memberuto.senior-member,
as each, is regarded as its point o f reference.
Ntanga or antang*a must be distinguished from aumsoku (the
dural form o f kawusoku, * a kinsman or in-law *), in that the
atter term refers in practice to all ones kin and affines whereas
antang a is usually restricted to ones jun ior kin w h o reside w ith
one. O nes awusoku are, as it were, a pool from w hich antang a


V itilocal M arriage w ithin the V illage

m ay be drawn. B u t the term aumsohu m ay be broken dow n into
form er subdivisions according to the context o f situation. I f
one wishes to contrast ones kin from ones afones, one m ay
refer to the form er as aumsoku or tvusoku wakupwalika (* kinship
b y birth ), and to the latter as aka o r m atm ku, and mashaku (singu
lars, mukti and ishahu). A k u o r aumku refer to in-law s o f an
adjacent genealogical generation; mashaku describes in-laws o f
ones ow n genealogical generation. B oth categories m ay be
further subsumed as wusoku wakusumhwang ana (kinship by
marriage). B u t i f one wishes to distinguish the category o fo n e s
kin w ho are respected (ku-lemesha) and feared (ku-tiya woma)9or
w ho respect or fear one, from other categories o f kin, one also
applies to them the terms aku and awuku . A ll ones relatives,
including parents and children, m ay be so described w h o belong
to an adjacent genealogical generation. Brother and sister may
also sometimes refer to one another as aku.
T o m arry ones muku is incest (chimaltvamalwa, chipikapika or
hu-shim una). B u t the concept w uku admits o f a considerable
range o f behaviour patterns between com plete avoidance, and
respect tempered w ith affection and fam iliarity. In-laws o f
opposite sex in adjacent genealogical generations should avoid
one another ( bi-dichina ) ; a m em ber o f the ju n io r o f tw o adjacent
generations should 4feel fear for (ku-tiya tuoma kudi) his o r her
m others brother o r fathers sister, his potential parents-in-law,
w ho are often enough his or her actual disciplinarians; and a
person should * respect * (ku-lem esha) his or her parents and
siblings o f opposite sex. B u t one also * loves (ku -ken ga ) ones
parents and siblings o f opposite sex. In general the verb ku lemesha, 4 to respect , is used to summarize the totality o f respectfear-avoidance attitudes regarded as appropriate betw een aku.
B u t the best w a y to learn the idiom o f N dem bu kinship is to
exam ine local system s o f kinship, beginning w ith the ntanga or
bilateral extended fam ily o f the village headman.
I must-repeat that, in its simplest and m ost fondam entdfom i,
the basic m embership o f an Ndembu. viUage.consistsofa.]b^(hnan,
his uterine siblings,, and his. and their children and grandchildren.
In addition, it m aycontain spomes.ofsomejai^alfeofethes.epersons
and perhaps some great-grandchildren of the headman. This
group is the ntang*a o f the headman. T he spatial arrangement
o f the huts ow ned b y its members gives the first clue to the


ya BB









S ch ism an d C o n tin u ity in an A frica n S o cie ty

character o f their mutual interrelations. As I described in

Chanter Three, members o f. the_same_genealQgical_.gen.eratioii
tencLto live in the same sendchcie,^-nhileutneinbers o f
the adjacent generations live in the .opposite -semicircle^.. This
means that members o f altemate_genealQgical_generations build
together,, and their.hutsare.often, interspersed__The..arrangement
o f huts in fact suggestsjhatThe_village-is. coru:eived-not. as a set
o f hierarchically organized lineages. but _as ..a,. single. .extended
fam ily ; fo r as a rule members o f different lineages-of variable
span in weE-established villages build adjacendy, .so ,that extended
fam ilial unity takes precedence over. Hneage-diftbcentiation in the
spatial arrangement. Jelj*. sm allribjing.. vdkge.joihe-type-W ^
are considering, m atriliiieal kin, atid tlie-children and- patriliaeal
kin o f the headman, are interspersed. W ithin the ntang a, ego
refers to his matrilineal kin either as akwamama or mwivumu detu
(literally * in our w om b * or * in our lineage \ referring to matri
lineal co-descendants o f a com m on ancestress), and to patrilineal
kin as akwatata, w hich signifies persons descended through
agnatic links from a com m on ancestor, alm ost invariably a
common, grandfather. These are not patxilineages but i f m en
always bring their w ives hom e, a man and his. son and.his sons
children m ay all be living in one village.
T h eisntanga, thus, has lineally tw o m ajor components. Its
vertebral members belong to the matriiineage o e headman.
Female members o f the matriiineage maintain its continuity
through-tim e, ..w hile its._male_meird)rs-fo.rm.the_.residential core
o f the village, at. any. .given.. moinent.._...B_u.t..through..virilocal
marriage, the headman,- who-_is_freiju.enriy.. p-Qlygarnous, his
brothers and..bis sons, have, children who., together..constitute
quitea .high proportion, .o....the-..po.p.ulariQn_Q. a_viOUiage (see
Tables X and X I). In Chapter T w o (p. 44) I suggested that
there is a functional relationship o f some im portance between the
size and structure o f the local unit. N dem bu villages have a
mean size o f about ten huts, containing about tw enty persons.
Ent this small settlement, male village-kin outnum ber fem ale
village-kin b y m ore than tw o to one (see Tables X and X I).
Since the children o f brothers call one another siblings, the
village membership has quite as m uch the appearance o f an
extended fam ily as o f a local matriiineage.
This virtual equality , between.farnily .atid lineage as. .principles



V irilocal M arriage within the V illage


o f local organization is at least p a rd ^ y xesponsible__or_the

m erging o f patrilineal and m atrilineal kin as jo in t members o f a
single genealogical generation. Thus in m any situations^a mans
father and his m others brother have equivalent functions^and. are
treated alike. B oth should, according .. to custom ,_contdhute
the same sum o r the same am ount o f goods to a mans, bride
wealth. -W h e n a husband gives .bride^wealth- for...a-joroman, he
should give an equal portion to_her,

A womand^ph^es jhe..samejtexmsXatk,..mfewesi*M
to her husbands father and his mo.theris. brother.. ..and. must
equally avoid (ku-china) both. Conversely, a man should avoid
not only~his w ifes m other but also her fathers sister, to both o f
w hom he-refers, and both o f w hom he addresses,-as-muku or
mawenu. A man calls his o w n m other and also his m other's
brothers w ife mama (literally * m other ) and greets them in the
same w ay. H e greets his father and his m others brother in the
same w ay also. T h e equivalence o f father and m others, brother,
and of_m other and fathers sister, is expressed in a .number o f
other situations asw ell. -Ib is .-merging..o.^e^.Uy__iJistinct
members o f the senior adjacent and authority-holding, generations
is a further index o f the failure o f the m atrilineaL.principle to
becom e.indisputably dom inant am ong N dem bu.... Intends to
strengthen the bonds betweem m em bers.-ofihe-sam e-genealogical
generation, and. at the same tim e, b y the creation o f a divided
authority in the senior generation, to enable members o f the
ju n ior generation, to exploit the division between father and
m others brother, to their o w n advantage....Ih is_ d ivisio n o f
authority^nd.. consequently.. limitation._of^unilareral ^onU:ol b y
either father. orLmothers brQther imdQubtedly contributes__tp the
independent character orientation o f N dem bu, w hich,, in its. turn,
helps to account fo r the high frequency o f fission.
T h e Ndem bu_village, then^in its .s.Qcial.cQinpQsition-Eepj:esen||^
at once^a veiled -struggle, between- the_two_.powerfiiLpii||&pl^P
o familiaLan.d lineal organization, a n d -an -attem p t-to -i^ ^ ^ fi^
these b y a.jset..of com prom ise, form ations between..thei^il^rdey
o f social control._In the. .course o f the .stm ggle these_priniples
m utuallyinhibit .one another,_s.o..tha.t.Hneages xemam-shaEow. and
extended amilies..hteak.up. o r.lo se .their.membership, through the
pull o f m atrilineal affiliation....This, struggle. is itself a. manifesta
tion o f the deeper opposition between male.. .andL..female. in

23 S

Schism and Continuity in an A frican Society

Ndem bu society. Each village represents an attempt to establish

a patriarchal settlement in the teeth o f basic matrinneaL descent.
B iit the male kin w ho reside together w ith theirw ives and families
are themselves interrelated, b y matrstinea1.ties.which..persist,_while
ties jvith their children. which_interlink_themin .cof=residence are
m ore, tenuous and friable, snapping, w ith the divorce, o f their
wives5with_the mamageoftheix^daughters,and w ith th e frequent

m ajority., .On..a, viU age.,geni^ogy jh e,Iin e,o descent,, .the spine
o village conrinmty, is through jvon ien .whose_br_o_thers are
marginal to it. ,. But in any given_village,at.any,specific.point o f
tim ejthe line is broken or. in,rare,cases_even_imrisible.,.. For the
residential core o f a viHage consists^omen linked. to_,Qne,another
through wom en, most ow h om may, be,ab&ent, time,
or w h o are already .dead. In the course ,o f events^.however,
nearlyjevery. village contams some female,mLem.hers .o its nuclear
m atrilineage,whoJiave,retum,divorce-O E-w idow*
hood;. or-reside,diere_in.-cross=cousin-or-grandparent=grandchild

marriage, or whose husbands liv e uxorilocally These, w om en

betray b y their presence A e -fm A m m u tty matrilmeaLcharacter
or the* local unit.
Thus,the kinship basis, o f. a villagers .as m uch, bilateral as uni
lateral and represents a compr.omise .between,familial_ and. lineal
principles o f organizations....-In. m y ..opinion,, .the..socio-spatial
grouping..o f viliage kin by__genealo_gic_al.,ge_nerarion,_ as noted
above,,'tends. to .neutraH2e sQme,o .the .tensiQns,arisin.g.rom the
co-existence, o f ..disparate principles ina..single, lo it...C om
m on membership_oa genealogical.generation aligns persons w ho
belong to potentially opposed...descent .categories,. m others
brother and father, w ife and sister,.husband,,.son and
sister*s son, daughter and sister!s,daughters_memherSL.of,s.eparate
^grb^^^sf uterine siblings, or., members .o.separate--miniinal or
^Maaqfe?3ipfeages__In m ost N dem bu vihages the headm the
'H ^ t^ oj^ ll three m ajor modes o f organising, kinship-relations,
y lineage, b y extended bilateral fam ily, and. b y genealogical
generation. IheJs,.an_astute.nianJhe_endeavours,ta-reduce the
tendencies to disruption Q.the_viUage,,which, contra
dictions.between, these principles, by.encouraging-and taking part
in theqjrow th o f a netw ork o f afSnal ties These ties are.form ed
between persons belonging to opposed ...categories ...based on

V irilocal M arriage w ithin the V illage


differentprinciples^ of^ sacialorgam zatiori. Thus in M ukanza

Village, the headman (I,F8) b y his m arriage w ith N yam ukola
(I,H io) united the lineages o f n yach intang a and malabu .
B y the same m arriage he strengthened the alliance between
genealogical generations F and H (see A ppendix I). B y arranging
cross-cousin marriages between his daughters b y N yam ukola and
his sisters sons, Kasonda (I,G i5) and Sandom bu (I,Gxo), he
sim ultaneously sought to increase the internal cohesion o f
nyachjntang a lineage and o f the village lineage o f widest
span, that o f nyachipendi, sister o f the founder o f the village.
A t the same tim e, both b y means o f his ow n m arriage w ith
N yam ukola and b y means o f the marriages o f his daughters
w ith his sisters sons, he tried to overcom e the dichotom y
between his seminal children and his ju n ior lineal kin. Since
N yam ukola belonged to the village lineage his children by her
w ere fu ll village-m em bers. A n y children that Kasonda and
Sandombu m ight have begotten on his daughters w ould also
have been fu ll village-m em bers b y m atrilinea! descent, and the
rivalry betw een Kasonda and Sandombu m ight have been re
duced b y the fact that they had married sisters. Kasonda and
Sandombu belonged to different m inim al lineages, each o f w hich
represented a potential unit o f fission. B u t their children b y
tw/Ml\A h* ve belonged to the same
N yam ukola s oaugnters
m inim al lineage and w ould have been aligned w ith their grand
father M ukanza in the same linked-generatton segment.
M ukanzas other m arriage, w ith Seliya, a rem ote matrilineal
kinswom an, also ensured that his children b y her w ould belong
to the village m atrilineage and at the same tim e strengthened
his ties w ith Line (I,H i 7), a senior man o f m alabu lineage w ho
had m arried Seliyas sister. In this case too nyachintang a
and m alabu lineages ,w ould have a jo in t interest in the children
o f these m en. Kasondas divorce o f M ukanzas daughter and
Sandom bu s sterility thw arted some o f these intentions. B u t
M ukanza, like other headmen, clearly understood the necessity
ion forestalling b y m arital ties the potential disruption o f h js v
village. Sodrces__o.f disruption .w ere .divisiom....arism^^i^hjav^
m atriliny, conflict between_sem inalchildren ^ancLsisters j^ pf
and struggles betw een adjacent_genealogical^generarionll^
T he contradictions between these three principles o f village


Schism and Continuity in an A frican Society

organization lineage, fam ily, and-generatioifc

term inology.. of__the Ndem bu.. and influence. jdxeJhehavioural
noxms governing the interrelations .ofJdn. It is w orth repeating
that the simplest w ay to approach the analysis o f Ndembu
kinship is through the exam ination o f the spatial structure o f
the typical village. The village hut-circle consists o f tw o semi
circles, occupied respectively b y members o f adjacent genealogical
generations, and composed respectively o f members o f alternate
genealogical generations. Bach linked-generation segment calls
all the members o f the opposite segment collectively aku. Aku
o f the senior generation le v y respect from aku o f the junior
generation and exert authority over them. Betw een aku who
are in-laws, and, w ithin this category, between aku o f opposite
sex, the relationship o f respect is at its most form al and rigorous.
Aku o f opposite sex avoid one another, and aku o f the same
sex exhibit extrem e constraint in one anothers presence. But
affinal aku do not, in practice, enter into a relationship which
confers a high degree o f authority on the senior over the junior
generation. H ostility in the relationsjhip.pxedQminates.joyer the
factor, o f social control, resulting in .av.oidance....and....c.Qnst:raint
rather than in authority. N o t infrequently such aku physically
attack one another. For exam ple, one o f the prisoners at the
Bom a in 1950 was a man w ho had murdered his son-in-law after
a fierce quarrel. Benson o f N g om bi V illage (mentioned in
connection w ith Social Dram a II, p. 118) once beat his m otherin-law severely after she had upbraided him for taking a second
w ife and neglecting her daughter.
These examples indicate, h o w near to-xhe-surface-hostility m ay
comein_this. relation ship,jan d ^ hy _the..noxms^ stress
the .of. m utual.interaction. ...a,s,far.. asp.QSsible..... B ut a
relationship o f this.kind .can hardly., he. a. su itab leh n .a chain
o f authority. ... A -relationship.. o. superordination-subordination
must_also involve xo^operation.,..There is. a chain .of-comm and.
B ut .there...can .be little . co-operation..betweeru Jcm^who avoid
one .another or. experience-.extrem e..constraintJuGL-oneuanothers
presence. Am ong. Ndembu,.^authorityrelations_ar.e...carried in
the Jdnship..structure. Thus that category o f aku, over w hom
one effectively exerts authority and w hom one in practice obeys,
is made up o f consanguineal relatives o f adjacent village genea
logical generations. W rfhin._thiscategory,. a

V irilocal M arriage within the V illage


range^of behaviour patterns. O n the w hole, m atrilineal kin o f

the senior adjacent .generation exert greater^controLover- their
junior relatives than do paternal.kin.
B ut the; sharp differendatioii be.tween the jXLOthec s hrotlier as
a severely authoritarian_iig.ure and the. .ather as..a..herievolent
protector^ found in m any matrihneal . sQcieties,_. .sjui:h_._as_ the
Trobriand Islanders, is not m ade.. by.N dbm bu. D ue to the
partial assimilation o f functions associated w ith th e tw o positions,
the m others brother has m uch less control and...the. father rather
more than am ong the Trobrianders. I have, for exam ple, seen
fathers administer beatings to their youn g children, and I have
know n sons w ho left the villages o f harsh fathers to stay w ith their
mothers brothers. B ut the m ajority o f fathers assist their ow n
children financially ; help them to m arry ; represent them in
co u rt; teach them the skills o f hunting, housebuilding, and the
blacksmiths c r a ft; and instruct them in custom and law . The
father is head o f the domestic gardening team in bush-clearing
operations, and m ay call on his children to assist him in hoeing
his ow n garden. H e takes his sons w ith him to carry his p ro
visions and his kills when he goes on hunting trips.
There are a num ber o f folk-tales w hich reveal the existence o f
tensions in the father-son relationship. O ne tale describes how
a father killed his son in the bush w hen the pair failed to find game,
and h ow the father was detected in trying to pass o ff his sons
body as the dried carcase o f an antelope. A nother relates h ow
a father cursed his son fo r w alking beside him like an equal
instead o f behind him like an inferior w hen they w ent honey
collecting in the bush. As the result o f that curse the son could
not find a single hive. E ventually the lads m other persuaded
him to treat his father w ith respect and thus induce the father to
revoke his curse. A son, other than an infant, m ay not sleep in
his fathers h u t; nor, by extension, m ay C h ie f Kanongesha
spend the night in the village o f any o f ms senior headmen to
w hom he stands in the perpetual relationship o f father
the other hand, it is said that the father-son relationship is free
from mutual sorcery accusations, although the case o f Chibwakata
and Kasamba (p. 160) where a classificatory son was alleged by
his * father to be trying to bew itch him , seems to be an exception
to this rule. I have never heard that a father had been accused
o f bewitching his own son or vice versa, although in divination


Schism and Continuity in an A frican Society

paternal kin o f the victim s are occasionally named as sorcerers

or witches.
I have often witnessed acts o f indulgence and kindness per
form ed b y men on behalf o f their sisters sons. W hen, for
instance, the uterine nephew o f one man I knew carelessly
wrecked his uncles bicycle, the latter freely forgave him , telling
me that i f he w ere too severe on the lad, when the tim e came for
him to m ake a village o f his ow n his nephew w ould not come
w ith him . 4People are m ore im portant than things, he added.
I have know n men to pay their nephews fines in adultery cases.
B ut there is no doubt that uncles often punish nephews quite
harshly. For exam ple, Sakazao, usually the gentlest o f men,
grew exasperated one day in M ukanza V illage b y the persistent
naughtiness o f Chikim bu,1 his sisters son, and twisted his arm
out o f join t. Sisters children are usually at the beck and call
o f their uncles in the collection o f firew ood, the carrying o f food
from kitchen to chota, and so on. B u t the father is in the final
issue his ow n childs protector. In Ikelenge Area I once saw a
man angrily threaten his sisters son w ith a beating after the latter
had knocked down the form ers son. T he case cited on page 189,
in w hich a father refused to allow his w ifes brother to sell his
children into slavery, w ell expresses the difference between
avuncular and paternal attitudes. The marked hostility between
uncle and nephew emerges in m any sorcery accusations I have
recorded between these categories most o f them in connection
w ith succession disputes.
T he relationship o f father (tata) to daughter {mwana wamumbanda) is often one o f great affection ; indeed, I have only once
seen a father beat his young daughter. In fact, such an action is
roundly condemned by N dem bu w ho say that it is the task o f
the m other to punish the daughter and o f the father or m others
brother to beat the son. M others brothers m ay soundly rebuke
nieces w ho have been involved in adultery cases, as m ay mothers,
but fathers seldom do so.
M others frequently beat young children, but once a b oy has
been circum cised he comes under the general control o f the men
1 Chikim bu had transferred his allegiance from Sakazaos group to Sandombus group. The relationship between uncle and nephew was strongly
influenced by die political struggle between village factions. But the political
factor did no m ore than exacerbate a relationship that was already tense.

V irilocal M arriage w ithin the V illage


o f the senior adjacent generation and the m other loses her punitive
role. B u t m others exert quite a considerable authority over
daughters, levyin g their help in gardening, the preparation o f
food and the carrying o f w ater supplies. It is the m other w ho
receives the greater share o f gifts, as distinct from cash, bestowed
for her daughter b y a future son-in-law , and in the past it was for
her that her son-in-law used to build a hut and m ake a gardens
Jh the context o f m arital relations the m other is sometime,
1 g ? .'
a dken o f as the real * ow ner o f the child (mwenimwma ). This
Ose and authoritarian bond betw een m other and daughter is
*?-' 5 - dos
probably strengthened b y the fact that lineal descent passes
through w om en. T his bond is n o t broken b y death, since the
ritual histories I have collected show that the m others spirit is
<, C-- thought frequently to afflict her daughter w ith reproductive
troubles o r illness i f the daughter neglects to m ention her name
praying o r to m ake her an offering o f food or beer. Again
a m others spirit is thought to afflict her daughter i f she quarrels
w ith relatives. In fact, the m other-daughter tie is the second
frequent category o f relationship between afflicting spirit
l l ' ;and patient w hich recorded.
T he fathers sister (1tata wamumbanda or taiankaji) is in certain
contexts equated w ith the m other, hut she is treated w ith greater
: "'' constraint, since, w ith preferred bilateral cross-cousin marriage,
she is a potential m other-in-law . For that m atter the m others
brother is a potential father-in-law , and this tends to diminish
his authority over his sisters children i f they m arry his children.
A fter m arriage o f this kind, his nephews have little to do w ith
him directly, and his nieces avoid him .
So for w e have been considering relationships between primary
kin w h o call one another aku. TKe^dassifcatGr/, extension o f
kinship terms. tends_to_dififuseLOver_the. w h olegrou p-ofsen ior aku,
the control exerted b y prim aiy km o v e r foeiiL jmiiQrs in the
adjacent ^eneration, thus m itigatinn . its. narrowi sev.erity in
particular relationships. This co n tro l .by ,a .who
generation o f village kin over.another . w hole.-generation is
especially niarked in the case o f m en w h o form .in som e respect
a separate, m oral com m unity o f males.1* O ne o f the most

t^tS *


1 W om en do n ot form a coherent m oral com m unity since with virilocal

m arriage m any arc unrelated to one another and do not belong to the m atriiineal group o f the village.

r r




Schism and Continuity in an A frican Society

im portant functions o f the boys* circum cision ritual o f Mukanda is,

in fact, to rem ove boys from the authority o f their ow n individual
mothers, in the sphere o f the Idtchen, and to bring them under
the collective control o f the m en o f their village, in the sphere
o f the chota, the m ens court and forum* A g irl remains largely
under the authority o f her ow n m other until her husband has
made the kazundu paym ent, at a variable period after marriage,
w hich enables him to rem ove his w ife * from her m others
knee \ This..is__becanse...thebasic.format.jpf.a._village....consists
essentially, o f a set o f structured relationships. betw een,m dekin
w hojconlxol m d obey, one another in-econom ic.andjuralm atters.
Another aspect o f this .collective control is-connected_with the
conception o f a village., as._a -unit which-transcends- its-narrower
com ponents, the uterine siblm g groups .and elem eritary Families
discussed above.



W ithin.eachJhiked-gm erado.n_segm ent.inaviUage..them ost

general behavioural^ he sh d .tp. be^fajiiiliarity
and-equality. B ut the precise...w:dghting,Qf..such characteristics
varies in each segment. Im thegrandparent-grandchild segm ent,1
the generation .components are.geneaipgicaliyjunequal. and often
differ_considerably in_the mean age_of^jnember
ships; This fact tempers, the fam iliarity o f their, intercourse w ith
respect, .since. it_is that
the _eldest-bom. .is entitled , to the greates.t_respect._B ut these
components have. in.-coxnmon the fact-that .both are_.o.pp_osed to
the^mtervemng_.genealogieal. generad<ni_J^Mch..^ndeavours to
w rest _CQiatrolfromthegrandparentgeneration-._^With_r.eference
to the_medial generation they . stand. allied. . .Q nthe.other hand,
since the grandparent, generationJOLeazly__always holds .the. headmanship in a village, and since, ultim atelyall.,the_.authority
reladons .are.jvested_in...this.jQffice^a ..certain.J:ensi.on_exists in "the
rdaridnship between aJtemate generarions.^ A gain ^ if .we regard
the-village as a set o junked, corporate., groups.,..each a potential
unifcof fission, w e find that members.of_altemat_e_generations have
different uterine sibling group .afl5Hations_andJb.elongjta_djfferent
diementary families. This fact-again, in view . o f .the basic and
1 B y * grandparent is meant any member o f the senior alternate generation
w ith w hom a person can trace genealogical relationship through any line.

T h e A l l ia n c e


A l t e r n a t e G en e r a t io n s

A ll the children, w ith one exception, playing around Sakazao, the nominated
successor o f Headman Mukanza, are daughters children o f his sisters ISTyamukola
and N yatioli. The exception is the little girl standing behind him and clutching a
maize-cob. She is Kasondas daughter and is Sakazaos cross-cousin and joking
partner. A ll the children present in fact have a jok in g relationship (wusensi)
with Sakazao, their grandfather .

V irilocal M arriage within the V illage


~' ineradicable narrowness ofJSldemb.n_corp.otate. group, relations, is

a latent source o f tension between them.
That tension exists is again revealed b y certain sorcery beliefs.
I have been told that at night a sorcerer sends o u t his ilomba
: snake-familiar to listen invisibly to w hat his grandchildren
l (ejtkulu) are saying about him in private and to report hack to
him any expressions o f anim osity. In ritual perform ed to placate
female spirits w ho have returned to plague their livin g female
relatives w ith infertility or illness, grandm others form the largest
category o f punitive spirits.
Q n th e other hand, in custom the grandparents, are the genial
i advisefSL.and. instructors
m any matters.
Especially, is this.the case w here ed. Aku
and siblings o f opposite sex m ay n ot speak openly to one another
on sexual topics but must em ploy circum locutory speech (kudidyika). B u t this prohibition does n ot extend to grandparents
(cmkaka). In feet m any instructresses o f girls undergoing the
puberty ritual are * grandmothers \ These w om en (ankongu)
instruct their novices inter alia in the techniques o f sexual inter
course and in other matters connected w ith sex and reproduction.
Grandparents and grandchildren m ay sleep in the same hut and
witness each others sexual activities. Y o u n g children w ho are
considered too old to remain in their parents huts at night for this
reason are sent to sleep w ith their grandparents, sometimes in
the same bed.
Thejrelationship. between, grandparents, and. grandchildren*, like
that between cross-cousins, is summed up in the. iSTdemb.u term
w usensV' Wusensi m ay be translated as jo k in g .relationship *.
Joking partners m ay revile one another, excitin g amusement but
not anger. T h ey m ay claim any articles o f one anothers property
they m ay fancy or call on one anothers assistance ,in.,work^with
out paym ent., or. return.Joking -partners -of-opposite-sex- may
indulge_in, even-in. public. B ut. the jo k in g relationshipjw hich exists between grandparent.and-grandchilcLis-of a mild
and~tesfrained type, and tends to. b e asym m etrical. T he grand
parent tends to initiate jo k in g activities such as reviling, o r seizing
food carried b y the grandchild. A gain, although a young
child is often allow ed to take considerable liberties w ith his or
her grandparents, w hen the child reaches adolescence he o r she
must behave w ith greater respect towards them. The relationship







Schism and C ontinuity in an African Society

he!3een.grandfether ,.aiid. grm 4<dbugbter .tends,, to . bp more

ribald and egalitarian than that between grandmother jancL grand
son,. _/approxim ating in fact to the behaviour- between, crosscousins o f opposite see. B yfa tlh g l.tL p ro p o xu o n .o fg ra n d parent-grandchild mamages-are. between, m en ..oil the. senior and
w om en jof .the.junior genealogical generation.
hutnany. o thesim p lersoaeties. term inQ ^
made between grandparents and grandc3b4drensp_diat...the two
categories are regarded as siblings. B u t am ong^Ndem bu the
twojcategories do n ot apply to one term
but/each category is distinguished. A grandchild calls his or
her grandparent nkdkay and a grandparent calls his or her grand
child mwijikulu. Furthermore, w hile grandparents and grand
children o f opposite sex m ay sleep in the same hut and indulge
in sexual play together, such behaviour is rigorously forbidden
between brothers and sisters (as in other societies w here there is
this term inological identification). Grandparents and._grandcbildren behave tow ards one another m ore as though they w ere
cross^ Y e t in the case
o f grandfathers and grandsons kinship term inology and behaviour
patterns determined b y it do suggest that an approxim ation to
the equivalence o f brothers is involved in the relationship. For
exam ple, a man m ay call his m others m others brothers child
and also his fathers m others brothers child, a person o f parental
generation, 4 m y child * ; thus im plying that the speaker is
regarded as being in some sense the m others brother o f his ow n
parent. T h e fact that a wom an m ay also call her m others or
fathers m others brothers child * m y child *, suggests that she
is regarded as the cross-cousin or potential w ife o f her classificatory
grandfather. Thus, although all members m ale and female o f the
senior alternate genealogical generation are classified together as
ankaka b y all members o f the ju n ior alternate genealogical
generation, w h o are collectively called ejikulu b y them , crosssexual relations betw een these categories tend to be m odelled on
cross-cousin relations, and relationships between members o f the
same sex on those obtaining between brothers or sisters. Betw een
the categories as a w hole, therefore, there is a range in the
character o f their interaction between equivalence and alliance,
and equivalence and alliance are m odified b y the perceptible
degree o f authority exerted by the senior over the jun ior genera-

V irilocal M arriage w ithin the V illage


tion. Contradiction exists in die relationship in that a grandson

w ho is regarded in one sense as a younger brother is at the same
tim e a potential brother-in-law . Sim ilarly a granddaughter w ho,
as the sister o f a kind o f younger brother, m ust be regarded as a
sister, is also a potential w ife. B ehavionrassociated w ith these
respective ro contradictory. ^ T h e cornequent confusion is
one o f the. elements in the jo H n g relationship .w h ich is nearly
alw aysa coincidentia oppositorum.
This^confusion and contradictoriness-seems...a_m-.ta_brmg
out the essential nature o f the principle o f organization by
genealogical generation as a com prom ise form ation.betw een the
lineal, and fam ilial modes o f organization. I f .paternal and
maternal .kin are partially, merged, as. m em bers o f a sin g leg en ea logical generation it .is possible to preserve-their distinctive kinshipeharacteristics in that gen eration ; but in the. follow ing
generation, given the continuing -strength ~o -thegeneration
principle, such distinctions becom e-blutred. In a society w ith
strong . lineage organization, one aspect .of., w h ich_ is., lineage
exogam y,-m em bers^of-tliegenerationsjuniorto. ego. are.divided
into tw o categories :
(1) Lineage fe llo w s.with.,.whom..copulation., is- incest.
(2) Others, certain o f whom , are -potential-spouses.
B ut am ong _Ndemb.u,_members.ogeneratfohs-junior- to ego are
divideddnto tw o other categories :
(1) Thos in adjacent generations, w ith whom -there is hostility,
fear,-form ality, authority, etc.
(2) Those^in...alternate generations,, w ith .w hom ..there,is in
form ality, trust,. equality, .. and.the... like, and. w ho are
potential spouses except fo r ow n childrens .own children.
M arriage w ith ones ow n sisters daughters daughter is frow ned
upon but sometimes occurs, as in the case o f Nswanamatempa,
w ho married his ow n sisters daughters daughter N jonka (III,G l ;
see Appendix HI).
B u r the generation principle, trium phs.over the lineagepsrinciple
to the extent that a person may. marry-a..q.uite close JmeaLkmsman
o r kinswoman, belonging to.-an. alternate-generation. O n the
other hand, one m ay not m arry a com paratively ,distant paternal
or cognatic kinsman or kinswom an o f an adjacent generation.



Schism and Continuity in an A frican Society

The jo k in g relationship between alternate, generation kin is

made: up. o f ..two m ajor components One. such com ponent is
that in-effect a person m ay m arry someone or_her
classificatory sibling ; fo r true siblings this is tabooy.yet hereLit is
perfectly legitim ate, to do so. Embarrassment is- resolved by
laughter. T he neighbouring Lwena tribe regard this Ndembu
custom w ith as m uch abhorrence as Ndem bu display towards
the reputed Lwena custom o f wife-sharing. T he ..other com ponfcnt. consists in the. fact that two. categories, o f .genealogically
asymmetrical kin are regarded a- equal and allied, in contradiction
to the general prin cip le. that seniority -by birth or genealogical
order_confers authority. Thus ones equal ,is. a t the same time
one!s.inferior, a paradox again resolved b y .mutuaL-joking.
In both linked-generation segments in a village one o f the tw o
com ponent categories consists o f prim ary and classificatory sib
lings. In m any contexts siblings are equated, in others they play
opposed roles. T h ey are equated as members o f a uterine sibling
group, the basic unit o f fission and nuclear unit o f a new settlement. T he term mwivumu detu, * in our w om b is most often
applied to such a group, the children o f a single m other, although
it m ay be extended to all w ho claim descent through real or
putative matrilineal links from a com m on ancestress or even to
those w ho w ithout remembering the name o f that ancestress
claim maternal relationship w ith one another. B ut basically
mwivumu detu refers to ones ow n sibling group. There is no
general term for * brothers \ only terms for older brother
(yaya) and 4 younger brother * (mwanyika). A man does not
discriminate between an older or younger sister, but calls both
by the term muhela. O lder and younger sisters distinguish one
another by the terms yaya and mwanyika respectively. A wom an
calls both older and younger brothers by the term manakwetu.
Kwetu is the locative, not the possessive, form , and means 4at
our place . This is interesting, since it denotes that a wom an
regards her hom e as being where her brother is residing. The
plural form * kwetu \ 4at our place,* suggests that the sibling
group is regarded as a spatial unit. Sometimes a wom an w ill
refer to her brother, whether older or younger, b y the term yaya
w hich is usually applied to an older sibling o f the same sex as
ego . This implies that she recognizes his authority, since w ithin
the genealogical generation seniority goes w ith prim acy by


V iriiocal M arriage w ithin the V illage


birth. In the past I am told that a wom an used to call her

brother iyala, literally * man , a term now sometimes used as a
synonym for husband * (mfurnu).
W ith in a genealogical generation seniority in age and not in
genealogical position determines whether parallel cousins call one
another ' old er, or * younger sibling \ O n the other hand, sublineages o f a village matrilineage are rated as senior (ivumu
Jamukulumpi) or ju n io r {ivumu dakanst) according to the sibling
order o f their apical ancestresses. Thus one m ight find a man in
a senior lineage addressing his parallel cousin in a ju n io r lineage as
older brother since the latter was the earlier born o f the tw o.
Hence in each linked-generation segm ent there is a contradiction
among lineal kin o f the same sex between the principles o f
genealogical and chronological seniority. B oth principles are
invoked b y candidates for village headmanship.
Nevertheless,..despite the.. close -and,persistent_ ties~..that unite
members, o f .a single.... uterine, sibling, group,..pow erful..tensions
between its m em bers exist....The. closer ,the .tie, .the_greater the
am bivalence o f feeling. Betw een older and yo u n g eru terin e, hostility revealed., b y an... analysis...o.soxcery
accusations. B y adelphic succession a m ay
succeed to headmanship on the death o f his.Qlder .brother so that
he is a .natural target fo r suspicions w hich , themselves , m ay tend
to influence the result o f a ..divination. A... further source, o f
hostility resides in. the fact that an .old er .brother is .regarded b y
N dem bna' He
m ay be delegated b y his father w ith the task o f punishing a
younger brother or o f admonishing him severely for bad be
haviour. In the boys circum cision ritual, this role o f the older
brother is dram atized ; either a father or an older brother is
considered to be the m ost suitable person to act as * shepherd *
(chilombola) to a novice. Part o f a shepherds duties consist in
administering punishm ent to his charge fo r breaches o f lodge
customs, on the instructions o f the senior official o f the seclusion
lodge (mfumu watubwiku), w h o is him self usually the father o f the
greatest num ber o f novices. A n older brother can command
the services o f his younger brother in, for exam ple, the carrying
o f loads, the collection o f firew ood, hut-building, and many
other w ays. B u t ju st as the shepherd at circum cision protects
and attends to the wants o f his charge, so also does the older











S ch ism and C o n tin u ity in an A frica n S o ciety

brother give protection and assistance to his younger brother in

the affairs o f life. O nce when Daudson M w evulus (I,H7)
younger brother S ig n 1 was insulted at ah1Nkula ritual as being
too young to take part in the popular m odem chikinta dance which
today accompanies most rituals, and in w hich the younger element
dom inantly participate, M w evulu im m ediately started a fight
w ith Signs detractor and then knocked dow n the latters father
w ho had intervened on his behalf.
In m ore form al matters older brothers w ill assist their younger
brothers w ith bride-wealth, w ith court fines, and w ith deathpayments, and w ill share their food and possessions w ith them
w ithout question. As brothers gro w older their relationship
becomes m ore egalitarian and reciprocal and often som ething like
a division o f skills emerges between them. Thus Mukanza
Kabinda became a headman w hile his younger brother Kanyom bu
acquired a reputation as a ritual specialist and herbalist. In
another village M ulila, the younger brother, acted as headman,
w hile Itota, the older brother, devoted m ost o f his tim e to hrmting.
In Ikelenge Area Nswanandonga was at once headman o f a
village and a hunter w hile his younger brother M akw ayanga
participated fu lly in. the cash econom y at various times as a casual
labourer, kapasu or N ative A uthority policem an, and cash-crop
U terine brothers usually spend the w hole o f their lives to
gether. M others and sons m ay be separated b y the divorce and
rem arriage o f the form er. Sisters are divided from brothers by
virilocal marriage. Children m ay be separated from fathers by
divorce, sex, matrilineal pressures and vim ocal m arriage. Uterine
brothers are united b y descent and generation. Thus the link
between brothers is deep and abiding, despite the possibilities o f
conflict inherent in it. W hen a man founds a new settlement his
m ost faithful supporter is his brother, for their sisters have another
lo yalty to their husbands.
In die relationship between parallel cousins the ties are less
close and the cleavages m ore clearly evident. Parallel cousins
belong to different matricentric fam ilies and m ay have been
reared in different paternal villages. W hen different groups o f
uterine siblings build together and becom e co-m embers o f a
1 Manyosas sons in Mukanza Villag.

V irilocal M arriage w ithin the V illage


village genealogical generation, the divisions initially existing

, between them are n ot im m ediately or easily overcom e. Even
although it is custom ary w hen m eat is divided to giv e an equal
share to all kin w ho bear die same kinship term, in practice people
tend to favour their closer at the expense o f their m ore distant
kin and this becomes a source o f jealousy and back-biting,
especially between sibling groups. Sim ilarly, although it is
expected that close kin should contribute and receive the bulk o f
marriage paym ents, fines, etc. and that distant kin should contri
bute and receive lesser amounts, m uch w rangling actually takes
place between tw o o r m ore uterine sibling groups w ithin a minimal
or m inor m atrilineage about the precise sums given and received.
B ut it is w hen succession to office drops from the last living
member o f the senior village generation to the ju n io r adjacent
generation that relations between its com ponent uterine sibling
groups tend to becom e m ost strained, it w ill be remembered
that in M ukanza V illage the relationship m ost fraught w ith
tension was that between Sandom bu and Kasonda, parallel
cousins belonging to different sibling groups. It was pointed
out in the analysis o f Social D ram a I that according to N dem bu
opinion succession to office in a chieftainship o r in a w ellestablished village should not be confined w ithin a single narrow span lineage, so that a man was succeeded b y his uterine nephew.
In point o f fact, how ever, a headmans ow n sisters son often
strongly urges his claim and this brings him into conflict w ith his
male parallel cousins and also w ith their siblings w h o support
them. This kind o f conflict probably influenced the secession o f
Nswanakudya (II,D io) from Shika V illage w hen his parallel
cousin seemed lik ely to succeed, and that o f M atem pa (III,D3)
from N sanganyi V illage, w hen the present headman did in fact
succeed to headmanship (see Appendices II and III; also p. 208 and
p . 2 1 2 ).

The tie between brother and sister is also a very close one, and
like the preceding tie, contains several possibilities o f tension.
Brothers and sisters are frequently reared together in their fathers
village, but the tie o f spatial propinquity is usually broken b y die
virilocal m arriage o f sisters after puberty. Brother and sister are
linked b y descent and generation, but are divided b y sex. In
certain respects the relationship o f brother and sister resembles
that between aku (in-laws) belonging to adjacent generations.


Schism and C ontinuity in an A frican Society

Strong taboos against incest divide them and introduce constraint

into their behaviour towards one another. N either m ay speak
overtly o f sexual matters to the other. O n ly as very small
children m ay they sleep in the same hut. A man was once
strongly suspected o f sorcery w hen on a visit to his m arried sister
he sat on her bed and it broke beneath his w eight. T o obtain
a peculiarly potent hunting m edicine a man must com m it incest
w ith his sister, and the suspected possession o f such m edicine by
a hunter causes him to be regarded w ith abhorrence. One
N dem bu oath believed to be pow erfully binding in a village
court has the form , * I f w hat i say is untrue I have had sexual
relations w ith m y sister/ This is considered to be a statement
o f the impossible. Y e t in spite o f this incest barrier sisters are
equated w ith w ives in several contexts. A t a chief's installation
ritual, fo r exam ple, the chief-to-be is secluded during the night
before his accession either w ith his senior w ife or, i f she is in a
state o f pollution due to her pregnancy or menstruation, w ith his
sister. Again, in the Nkula ritual, perform ed for w om en w ho
are thought to be afflicted w ith menstrual disorders b y the spirit
o f a deceased relative, the patient is escorted from the small spirithut built behind her dw elling-hut to the ritual fire b y a male
helper * (chaka chaNkula), w ho m ay be either her husband, her
brother or her son. A t the boys* circum cision ritual either the
m other or a father's sister o f a novice m ay cook for him at the
sacred fire (ijiku daMukanda).
In their ow n generation the principal conflicts o f interest
between brother and sister spring from marriage and the fam ily.
A brother desires his sister's children to sw ell his follow ing, but,
ow ing to the strong ties that interlink members o f a matricentric
fam ily it is difficult to fu lfil this wish unless his virilocally resident
sister is divorced from her husband. I have recorded several
cases in w hich a husband demanded compensation from his w ife's
brother in. a N ative C ou rt for retaining her in his village when
she had paid him a visit. O n the other hand I have heard
complaints from brothers that their sisters have remained w ith
their children in their husbands' villages so that the brothers
have been unable, though senior men, to found their ow n settle
W hen brothers and sisters inhabit the same village, disputes
sometimes arise between them over their respective groups o f

V irilocal M arriage w ithin the V illage


children. Sisters accuse brothers o f neglecting their nephews and

nieces in favour o f their ow n children. Quarrels arise between
brothers and sisters over the distribution o f bride-w ealth received
for daughters o f the latter. I have often heard sisters bitterly
scold their brothers fo r failing to assist them w ith m oney, and
for givin g cloth to their w ives but not a rag to their sisters. The
sister has considerable pow er in this relationship, and her threat
to leave her brothers village often has the effect o f bringing him
to heel. For exam ple, C haw utongi o f Mukarxza V illage, in
censed that Kasonda, her first parallel cousin, had n o t bought
clothing for her younger children, le ft the village fo r a w hile to
stay at M bim bi V illage, a short distance aw ay. She w ould not
return until he had made up a dress fo r her daughter on his
Despite the possibilities o f discord inherent in the relationship
between siblings, w ithin each linked-generation segm ent prim ary
and classificatory siblings, m atrilineally related to the headman,
confront the other members o f their generation as a unitary
group. Thus in the senior segm ent the headman and his matriHneal generation mates confront the adult members o f the
* grandchild category, w h o are usually num erically inferior and
m ay be divided am ong themselves b y varying maternal and
paternal affiliations. Some m ay be m atrilineal members o f the
village, others seminal children and childrens children o f male
village members. In the ju n io r segm ent prim ary and classifica
tory m atrilineal siblings confront their cross-cousins, the seminal
children o f male village members o f the senior segment. Since
the m atrilineage is to some extent a persisting corporate group
w hile the tie to the father is personal and ephemeral, in this
segment also the m atrilineal sibling group tends to have greater
internal cohesion. This group is linked to the co-members o f
its segment b y cross-cpusinship.
Cross-cousmship, like the grandparent-grandchild relationship,
is a jo k in g partnership. Perhaps on account o f the egalitarian
nature o f the tie betw een cross-cousins in N dem bu society, the
jo k in g is o f a m ore boisterous typ e than that betw een kin belong
ing to alternate generations. It is sym m etrical, sexual in content,
derogatory in manner, and often contains reference to the sorcery
or w itchcraft o f the jo k in g partner. Cross-cousins m ake free
w ith one anothers possessions, those o f opposite sex indulge in


Schism and Continuity in an African Society

sexual play, and cross-cousin marriage is the preferred form o f

marriage. N dem bu m en appear to m arry their patrilateral and
matrilateral cross-cousins w ith equal frequency.
B ut a distinction must be made between the kinds o f behaviour
that typically occur between close and distant cross-cousins.
Cross-cousins w h o inhabit the same village behave to one
another m ore as though they w ere brothers and sisters than
jo k in g partners. Their relationship exem plifies the statement
previously made that in the village there is a convergence between
lineal and fam ilial principles o f organization, and that the organ
ization b y genealogical generation represents a compromise
between them . As members o f a com m on generation crosscousins are united in opposition to the adjacent generations. As
children they are jo in tly under the authority o f the senior adjacent
generation, as adults they jo in tly exert authority oyer the jun ior
adjacent generation. Together they co-operate m communal
working-parties at the beginning o f the gardening season. Male
cross-cousins m ay bunt together, female cross-cousins m ay pound
in the same m ortar. Their fam iliarity in daily tasks and the
sim ilarity o f their structural position in the organization by
generation tend to reduce the element o f unlikeness derived from
their distinct lineal affiliations. I f they lived in different villages
the feeling o f dissimilarity w ould be m uch stronger.
U nfam iliarity and hostility are components, o e r a ille r y
between jo k in g partners. Cross-cousins m a y . be regarded as
unlike and equal. Relations, between siblings m ay be described
as parallel relations. Siblin gsare regarded one
another. T h ey are alike and equal. V illage cross-cousins also
tend^to interact in parallel rather than in com plem entary and
opposed relations. T hey tend, .in fact, to. behave...toQne. another
like siblings. This m ay result in a reduction_of_joking and
perhaps in a dampening dow n .o f the._sexual_attraction between
opposites. Cross-cousin m arriagebetweem .closeJdum isrelatively
infrequent. O n the other hand marriage between persons who
call one another cross-cousms..-.(dLyoytl,_but--who^cannot trace
precise relationship between one. another, is. quite-com m on. This
kind of. m arriage supports m y jprevious stetement tha^villages
w hich claim com m on precise
m am lineal link but can still class. one._another!s-members b y
genealogical, generation. Thus. the. m e m o f .one- village know

Virilocal Marriage within the Village

2 $$

roughly w hich w om en in the other village fall into the marriage

able category and. w hich w om en fall into the forbidden categories.
Am ong the Lam ba o f the N dola D istrict o f N orthern Rhodesia
whom M itchell and Barnes found to possess a h igh frequency
o f virilocal m arriage (38-1 per cent o f extant m arriages), and
among w hom , therefore, cross-cousinship m ust constitute an
im portant ingredient o f village kinship, prim ary and traceable
cross-cousin marriages form ed an extrem ely lo w proportion o f
the total marriages recorded (7 prim ary, 2 classificatory crosscousin marriages out o f 360 extant and dissolved marriages,
or 2'5 per cent) -1 It m ay w ell be that am ong m atrilineal peoples
with shallow local lineages and virilocal m arriage an identification
tends to arise betw een the children o f siblings o f opposite sex
on the basis o f co-m em bership o f a genealogical generation w hich
inhibits frequent intra-village cross-cousin m arriage. A m ong
Ndem bu, at any rate, there is a m arkedly greater tendency to
m arry distant than close cross-cousins. A nd it is a matter o f
observation that jo k in g behaviour reaches its fullest expression
between distant cross-cousins, and betw een cross-cousins,
whatever their genealogical distance, w ho norm ally reside in ^
w idely separated villages. JSach. village. contjdns..g_s?pa^e arSEf^**
autonomous system o f kinship relations,.jcon&txuedjcm ..^^^j t^^!
tional principles w h ich m um aU y.m odif^ one_ anQ tbei:^ Sl^ % ^ l&t^
principlethen, achieves.....its..m axim um - expression. Concrete
behaviour is their mean w o rk in g .. In specific_siimtLQns,, how
ever, one or other m ay become, dom inant Marriages_do. .in fact
take place between village cross-cousins.... -When, children build
leaf-huts in the h u s b a n d -p la y athusbands_andw iv e s-9xrosscousinsdbut n ot brothers ^ a sisters jxume__foe_marital_roles, In
other kinship categQries_we..have-seen^examples-o.die,.kinds o f
situation in w hich, fo r instance, fatherss tancLopposecLta .mothers
brothers; and children _to. nephews. JBnt-imorder that the village
m ay be a-going, concern, and a viable..:unit,_.it_is. .necessary that
confficting nm tiilineal. and patrilateral loyalties_shauldLb.e recon
ciled. O rganization b y genealogical... generation., is...Jthe__major
means by.w hich. this reconciliation is effected.
1 Mitchell and Barnes, The Lamba Village, p. 41.


S ch ism and C o n tin u ity in an A fr ica n S o ciety

Virilocal-..mamage..imites a . nunib_er o narrow-spaii .maternal
descent groups within the bilateral extended fam ily, to form the
locaLunit o f residence. Each such descent.gro.up. is. a. matricentric
fam ily w hich in the course o f the social, process strives-to dis
engage itself. from all .other ties of--local-affiliation- and-achieve
residential autonomy. It ispossible, to..regard most other
categories o f local kinship ties as. co.unterbalancing..the .centrifugal
tendency o f the matricentric -family. I n .its earlystages the
matricentric fam ily is typically linked through.its_apicaLmember,
the mother, to a village lineage to which, h er husband-belongs.
Its members then have a personal tie o f loyalty to their father,
to tbe-bilateral extended family. (Kto^,ii.) .ofwliich-he-is-a. member,
and to. his matrilineal .village itself, .regarde<ijas_somethmg more
than the sum .o f its parts, as. a. unit osociaLspace-which m ay bear
an ancient, and respected name. I f the father is a headman this
tie to the village and to its dominant kin-group, his bilatral
extended fam ily, is exceptionally strong. I f he is not a headman,
but the senior brother in a large uterine sibling group, his
childrens dominant attachment may w ell be through him to his
ow n, narrow bilateral extended kin group which comprises his
ow n siblings and their children. I f this group splits o ff from the
village his w ife and her children w ill go w ith them. BujLeven
withim.-tfais-.villag&.-tO-which ,tfa.ey~amJinke<l..primarily through
their J&ther,_.the., children ...are.- -divided..-.from- -their, -parents by
generation.and united w ith their.patrilatemLctQSSO.usin&^who,
like_their .father,., are...meinb.ers-.of.the. -village-lineage^-- O n the
othei^Kand,..even..while-they live, in-dteir, fathers-village,-they pay
regular, visits, to their mothers .village,.in..vdrichjthey.iiaye ties o f
corporate ,group membership. It is in this village that some o f
them m ay one day succeed to office and inherit property. O n
the divorce or w idow in g o f their m other they m ay return to it.
O n -thir. return,-either-in_a compact-gijaupL.or-severally in stages,
they, fo rm a- tight-groupw hich-with- the maturation
o f its membershipbecomes-a-potentiaL-unit-. o f fission. Their
emancipation, as a_separate_residen.tiaL.unitris_pEevented, or at
least retarded, b y the cross-cuttmg^aHegiances-mentkmed above,
and b y a division o f their loyalries between the-narrow gxoup o f
uterine siblings and the-wider. groupings, of. minimal and minor

V irilocal M arriage within the V illage


matrilineages and o f the village as a w hole. V iriloral marriage

w h ich w as a tth e root o f their in tcn scco h esio n n o w operates to
separate-the. siblings b y sex. Nevertheless, asrwe have seen, the
tendency to jo in t secession evinced b y the g ro u p -o f- mature
uterine jsiblings frequently asserts itself-to retard the grow th of*
large villages.



"E have already discussed the w ay in w hich ties o f maternal
descent are utilized to establish links between villages in
separate vicinages. Such ties have a political function. Because
ties o f kinship, w hich are often coterminous w ith ties o f friend
ship, and o f econom ic, ritual and jural co-operation, enjoin
coflaboration and are almost always means to establish residential
affiliation, they have a high political value in a decentralized,
m obile and unstable society. Even m ore im portant in this
respect are affinal ties w hich interlink exogam ous and potentially
conflicting groups. In the preceding analysis o f internal village
relations, w e have seen h ow intra-village marriage in a longestablished village allies uterine sibling groups, belonging to the
same generation, by means o f cross-cousin marriage, and inter
links uterine sibling groups belonging to alternate generations
b y grandparent-grandchild, marriage. In this case kinship ties,
both lineal and cognatic, are supplemented b y affinal ties w ithin
the settlement. In this chapter links o f in-lawship between
villages w ill be examined as a mechanism for establishing political
relations between them.
T w o m ajor aspects w ill be considered : (i) the effects on inter
village linkages brought about by the particular form o f marriage
am ong Ndem bu, and (2) the range o f such ties, whether they
interlink villages w ithin a vicinage or whether they associate
separate vicinages w ithin a connuhium.
Table X X , based on a sample o f 165 extant marriages in m y
V illage Census about w hich I was able to obtain inform ation
from both m arital partners, shows to w hat extent virilocal
marriage is the prevalent form . N early seven out o f ten mar
riages w ere virilocal as com pared w ith seven out o f a hundred
w hich w ere uxorilocal. Som e uxorilocal marriages represented
the first stage o f a first m arriage in w hich the husband was still
conform ing to the tradition o f initial uxorilocality. Others, like
the marriage o f N deleki in M ukanza V illage (see pp. 101-2),
w ere special cases in. w hich the husband had abandoned his matri258

P olitical A spects o f K insh ip and A ffinity


lineal village on account o f fears o f w itchcraft or because o f serious

disputes w ith his kin.
T h e structural effects o f virilocal m arriage w ithin the village
have already been discussed in Chapter Bight w here it was found
that a relationship existed between this forn i o f marriage the
close cohesion o f the uterine sibling group, the shallowness o f
the m atrilineage, the organization b y genealogical generation
within the village, the tendency towards the m erging o f fam ilial

jeseoence o f

H u sba n d


if e

Sample of 165 -extant marriages :


N o. in sample


Virilocal . . . .
Intra-village .
Uxorilocal .
. . . .


i 64




7 3


and lineal principles o f intra-village organization, and the high

rate o f individual m obility. B etw een villages its principal
consequence is to mesh together b y ties o f affinity, w inch ulti
m ately giv e rise to links o f consanguinity, a num ber o f spatially
distinct settlements, vicinages and even senior headmens areas.
M atriJiny is the foundation o f each villages continuity in tim e ;
matrilineal descent is therefore the dom inant m ode o f attachment
b y kinship. V irilocal m arriage scatters the nuclear w om en o f
the m atrilineal kin-group o f each village through m any other
villages, so that the dom inant bond o f kinship is utilized not
m erely to consolidate the local unit, but to strengthen the ties
between separate villages. Thus each village contains not on ly
its ow n nuclear group o f m atrilineal kin but also a num ber
o f m atricentric fam ilies belonging to the nuclear matrilineages
.o f other villages. Conversely, through the virilocal m arriage o f
its ow n w om en, the nuclear m atrilineage o f each village at any
one tim e has a num ber o f its cells, consisting o f m atricentric


Schism and C ontinuity in an A frican Society

families lineally related to it, dispersed through other villages. It

is as though each village entrusts other villages w ith the task o f
fostering its ow n future membership and simultaneously brings
up the potential personnel o f other villages. B ut there is no
precise correspondence between the villages to w hich each village
exports w om en and from w hom it receives w om en. In other
words, there is no m arked tendency towards the form ation o f
connubial clusters o f villages w hich constitute closed inter
m arrying groups. Interlocking ties o f affinity connect separate
chiefdoms. Kanongeshas Ndem bu interm arry freely with
Musokantandas Kosa, and w ith Kazem bes C ongo Lunda. And
w ithin Kanongeshas chiefdom , senior headmans area is inter
w oven w ith senior headmans area b y the busy shuttle o f virilocai marriage. I f w e consider each remembered matrilineage
separately w e find that it has a local nucleus, consisting typically
o f a m inor lineage in w hich m ale members outnum ber females
by about tw o to one, and a w id ely scattered peripheral member
ship in w hich the proportions tend to be reversed (if absentees,
both m ale and fem ale, at urban centres are excluded from the
In certain matrilineal societies that practise viriloca! marriage,
such as that o f the Trobriand Islanders described by M alinowski,
it is custom ary fo r the sisters son to return to his m others
brothers village on attaining puberty or social m aturity as
defined by other criteria. N o such custom exists among
Ndem bu. A son w ill remain w ith his father after attaining man
hood i f he obtains any m aterial advantage from so doing, or i f
he has a strong tie o f affection, w ith him . For exam ple, Gideon
o f M ukanza V illage, son o f the headman, prefers to remain w ith
his father, even although his m other, from Shika V illage, has
been dead for several years. Gideon is in his m iddle thirties and
says that he w ill stay in M ukanza until his father dies. Daughters
w ill often return to their fathers villages after divorce, especially
i f their mothers also reside there, instead o f taking up residence
w ith their mothers brothers. This duality o f residential choice
results in a state o f endem ic and barely concealed conflict between
a man and his w ifes village matrilineal group, w h o w ant his w ife
and children, and between him and his sisters husbands village
lineage, w ho seek to retain his potential local follow in g and heirs.
This conflict o n ly becomes acute as the children and sisters

P olitical A spects o f K in sh ip and A ffin ity


children gro w up, but it exists at every stage o f a marriage.

There is really no w ay finally to resolve the situation except b y
divorce o r death. In die final issue village continuity depends
on m arital discontinuity.
In practice, o f course, since each village is in effect a set o f
variously linked uterine sibling groups w ith attached individuals,
who m ay be kin o r strangers, and since the units o f affinal conflict
are uterine sibling groups, losses in som e groups are compensated
by gains in others. For a tim e indeed m arital continuity is an
indispensable condition o f village continuity, since children
replace the lost sisters children, both in the uterine sibling group
and in the village. B u t in the end, i f the sisters children do not
return, the village w ill perish as a m atrilineal descent group. The
village o f M bim bi, fo r instance, d ose neighbour o f M ukanza,
was, in 1954, in a fair w ay towards becom ing extinct, since
M bim bis sisters* children w ere either livin g w ith their fathers
or had gone to other villages, his o n ly livin g daughter was hving
virilocally, and his son had left him to reside w ith his m other s
brother fo r fear o f the w itchcraft o f his fathers senior w ife,
w ho was thought to have killed several children in the
Each village is linked to m any other villages b y single marital
ties and in this w a y it extends the geographical range o f its
political interconnections. O n the other hand, one finds the
not infrequent establishment o f a relationship o f reciprocal marital
exchanges betw een tw o villages. T his arrangem ent aims at
reducing conflict since each o f the tw o in term arrying groupings
rears the future membership o f the other. It is in the interest
o f both villages to m aintain in being the marriages that interlink
them , since the breakdow n o f a m arriage in one village m ay lead
to the exertion o f pressure b y the m atrilineage o f the other on
their virilocally resident kinsw om an to leave her husband and
bring her children to her o w n village.
Nevertheless, as w e shall see, m arital ties, far from establishing
enduring friendship betw een villages, m ay g ive rise to frequent
disputes between them .
fii order to Bring out the political function o f inter-village
m arital linkages, the field o f affinal relations o f a single village
w ith other villages w ill first be considered, follow ed b y an
analysis o f the ties o f affinity and o f kinship affiliation, inaugurated




b y inter-village marriage, existing between the headmen o f

com ponent settlements in a single vicinage.






Schism and Continuity in an A frican Society

Since I have analysed the internal structure o f M ukanza Village

in some detail, I shall discuss the external relations o f the same
village constituted b y affinity in similar detail. In m ost respects
this village is typical o f the consolidated and traditional Nderabu
village. I regard all livin g members o f nyach intang a and
m alabu lineages, together w ith N yaw unyum bi and her children,
o f n yach ula lineage, w ho have made their hom e at Mukanza,
as * village members \ B oth extant and fruitful completed
marriages, whether com pleted b y death o r divorce, between
village members and members o f other villages w ill be con
sidered.1 Persons w ho w ere absent in urban areas in N ovem ber
1951 w ill be excluded from the investigation.
R eciprocal m arital exchanges have taken place between
M ukanza and Shika Villages for several generations. Shika
V illage xTxatxilineage belongs to the w ide m atrilineal descent
group from w hich the incum bent o f die Kanongesha stool is
elected. There is a distinct tendency for villages w hich par
ticipate in the Kanongesha chieftainship to interm arry w ith
K aw iku villages. Thus Shika V illage and its offshoot Nswanakudya V illage tend to interm arry w ith N sanganyi, w ith Kasai,
and w ith M ukanza Villages, all K aw iku settlements. Sim ilarly
there is a tendency for Chibw akata and its offshoot N g ombi
V illage, both o f Lunda origin, to interm arry w ith K aw iku
villages. B oth Shika and Chibw akata refer to themselves as
* the husband (1mfunrn) o f N sanganyi. Nsanganyi, on the
other hand, calls him self * the husband * o f Chibw akata. In
actuality, although both parties, the one descended from the
original invaders, the other from the autochthonous people,
claim superiority over one another b y the use o f the term
* husband \ both are able to m ake this claim b y virtue o f the
fact that the chain o f reciprocal interm arriage began w ith an
exchange o f sisters between the heads o f the respective groups.
1 one sense, therefore, the reciprocal m term arrying between
M ukanza and Shika V illages is an aspect o f the political rdation1 See M ap 6, M ukanza Village M arriages, for spatial range o f extant mar
riages and fruitful completed marriages o f m atrilineal membership.

P olitical A spects o f K in sh ip and A ffinity


ship between Luncfa and K aw iku. B ut, m ore im portantly it,

is a relationship between tw o N dem bu villages. A t one tim e
Shika V illage belonged to the same vicinage as M ukanza V illage,
and it was at this tim e that M ukanza Kabinda (I,F8), then, a young
man, married N yatungeji (II,D3), sisters daughter o f Headman
Shika Ikubi (II, C z ). In the same year N yattm gejis brother
Kahum pu (II,D i ; see pp. 190-1) m arried M ukanza Kabindas
sister N yam w aha (I,F7 ; see p. 118 et seq.), thus interlinking n ot
only the tw o villages but also the uterine sibling groups o f
M ukanza Kabinda and Kahum pu b y reciprocal interm arriage.
Although I have not been able to trace the exact genealogical
connection, M ukanza Kabinda called both N yatungeji and
Kahum pu his * grandchildren * (ejikulu), w hich suggests that
previous interm arriages had taken place between the tw o villages.
M ukanza Kabindas m arriage was term inated b y the death o f
N yatungeji in 1943, and the circumstances associated w ith her
death became a source o f perennial discord in the relations between
the tw o villages. W hen a death occurs am ong N dem bu the
lineal kin o f the surviving spouse have to pay the lineage o f the
deceased w hat is fo r them a substantial sum, called mpepi, o r
* death-paym ent , in com pensation fo r the loss they incur o f his
or her labour potential, or in the case o f a w om an, o f her fertility
in addition. Mpepi also includes paym ent for various ritual
services perform ed b y the village lineage o f the deceased for the
surviving spouse, the m ost im portant o f w h ich is to rid the latter
o f the spirit o f the dead spouse {ku-fumisha mukishi windi) w hich
w ill otherw ise d in g to the survivor and afflict him or her w ith
bad dreams and illness. Such a spirit m ay also jealously attack
the survivors next w ife or husband w ith disease. N o w in order
to avoid paying m pepi1 it frequently happens that a man o r
wom an m il divorce a spouse w h o is fatally ill. T he lineage o f
the deceased cannot then claim that the death occurred w hile the
m arriage was extant, and that therefore according to N dem bu
notions responsibility for the w elfare o f the deceased was vested
in the surviving spouse and his o r her village lineage. T he
practice o f divorcing an ailing spouse has undoubtedly distorted
' r-fr*
;il V

1 It should be m entioned that N dem bu villagers believe that the payment

o f mpepi is forbidden under a N ative A uthority regulation, but they tacitly
agree to maintain the custom clandestinely. B u t i f a sick spouse is divorced
there is no remedy at law .

t _____


Uns*. ncuBi



P o litic a l A sp e cts o f K in s h ip an d A ffin ity


the divorce ratios in Table V II, since a significant proportion

o f the divorces recorded must be in reality m asked w idow
W h en N yatungeji seemed to be on the point o f death, M ukanza
Kabinda divorced her, and carried her, w ith the aid o f kinsmen,
back to Shika V illage on a stretcher. H e alleged that someone
am ong her m atrilineal kin in Shika V illage had bew itched her * in
order to eat the mpepi hinting that Headman Shika him self was
to blam e. H e m en affected to take um brage because the
m arriage-paym ent (nsewu) he had made for her in cloth was not
returned. In the long process through w hich an N dem bu
m arriage becom es consolidated a num ber o f paym ents in cash
and kind are m ade. T h ey include the small chijika muchisu
4w hich closes the door * (today about 2s. 6d.) w h ich signifies
b etroth al; (he kazundu, 4taking a w ife from her m other *
(today about 5s.), w hich gives the husband his righ t to take his
w ife to his o w n v illa g e ; and the m ost im portant, the nsewu,
4the arrow , the marriage paym ent proper, form erly paid in
cloth, livestock o r a gun, but n o w in cash. Nsewu is generally
regarded b y N dem bu as a precaution against a d u ltery; in
formants have com pared it w ith a fence round a cassava garden
to keep out w ild pigs. I f a man has paid nsewu fo r a wom an he
is entitled to com pensation from his w ifes lo ver both in a N ative
A uthority C o u rt and in a village court, i f adultery is proven.
M oreover, a man is entitled to the return o f his nsewu on divorce,
w hichever party is adjudged culpable. I f he wishes to retain
w ith him im m ediately one o r m ore o f his children he w ill not
seek the return o f his nsewu from his w ifes kin. Again, i f he
wishes to m aintain friendly relations w ith his w ifes village
kin w h o w ill help to lo o k after his children, he w ill leave the
nsewu w ith them \ In the case o f N yatungeji no death-paym ent
was m ade and no m arriage paym ent was returned. M ukanzas
tw o surviving children, including Gideon (see Social Dram a IV ,
p. 148), b y N yatungeji, rem ained in their fathers village.
A lthough no o vert breach o f social relations took place between
the tw o villages, an undercurrent o f hostility remained which
manifested itself in subsequent situations.
Kasonda (I,G l5 ), son o f Kahum pu (II,D i) and N yam w aha
(I,F7), m arried his fathers classificatory sisters daughter
M angaleshi (II,Ea), o f Shika V illage lineage. She is know n as









Schism and Continuity in an A frican Society

his cross-cousin on both Others and m others sides,1 has borne

him five children, and their marriage is a stabilizing factor in the
relations between the tw o villages. Kasondas older daughters
have both acted in the capacity o f ritual handmaids {tunsonselelu)
to girls undergoing the puberty ritual at Shika V illage, and w ill
themselves be secluded there. T h ey pay frequent visits to Shika
unaccompanied b y either o f their parents. M angaleshi herself
goes to Shika to take part in any m ajor ritual that is perform ed
m ere. H er classificatory brother Gideon and Kasonda have
built adjacent huts in M ukanza V illage, and Kasondas con
nections w ith Shika have stood him in good stead both in his
internal intrigues and in his intercalary role betw een the tw o
villages. It is im portant for the M ukanza people to maintain
friendly relations w ith Shika V illage, for it is possible that a
m em ber o f that village w ill some day succeed to the Kanongesha
chieftainship.2 T he husband o f Kasondas classificatory sister
C haw utongi (I,G i2 ), N deleki (II,D8), is also the son o f a Shika
man, and he and his w ife have also built adjacently to Kasonda
on the other side to Gideon. These persons, linked to one
another through kinship ties o f varying character w ith Shika
V illage, form a sm all pressure group under the leadership o f
Kasonda w ithin M ukanza V illage. W ith in M ukanza too they
support the interests o f nyachintang a lineage to w h ich they
are affinally or filially attached: A t the same tim e they provide
a num ber o f linkages w ith Shika V illage w hich b y their m ulti
plicity and strength prevented the conflict betw een M ukanza
Kabinda and the elders o f Shika V illage from developing into
a permanent breach o f friendly relations between the tw o
T w o members o f malabu lineage w ere also affinally linked w ith
Shika V illage in 1951, and conflict arose from both relationships.
N yam uw anga ( I ,G i7 ; see Social Dram a IV , p. 148) was
m arried to W illiam M atorokoshi, (II,D7), son o f the late Shika
Ikubi. In that year he divorced her on the grounds that she was
a w itch and wanted to kill him b y her w itchcraft. Since
N yam uw anga was feared for the same reason in M ukanza
1 Kasonda stresses this double tie to support a politically valuable marriage.
But Mangaleshi is not his true matrilateral cross-cousin. She is classificatory
sister to Gideon (II,El), who is Kasondas matrilateral cross-cousin.
2 See pp. 207-8, 322.

P olitical A spects o f K insh ip and A ffin ity


V illage itsel and since die pair had no children, little was said
about the m atter. B ut Sakazao (I,H9), senior elder o f malabu
lineage, resented having to return W illiam s nseunt, since he was
a poor man. H e was said to have upbraided N yam u w aag a for
her reputed occult activities and to have demanded that she
contribute to the repaym ent o f nsewu. I did n o t hear whether
she did so, but it is probable that she did, since she w as living in
Sakazaos quarter o f the village in the dry season o f 1951 when
Social Dram a IV took place.
N yam u w an gV s sister Kalusa (I,G i9) was m arried to Kasondas
fathers uterine brother N tololu (I,D2), and this pair w ere the
parents o f Ikubi w h o , as described in the account o f Social Dram a
IV, was supposed to have been killed b y the w itchcraft o f
N yam uw ang a. N to lolu at that tim e was livin g iixorilocally in
M ukanza V illage, having quarrelled w ith his classificatory
brother, the headman o f Shika, Shamupaku (II,D5.) A fter
Ikubis death N to lolu returned to Shika V illage w ith Kalusa,
bitterly angry w ith N yam uw anga and in fact at odds w ith the
people o f M ukanza V illage as a w hole.
The death o f Kalusa, sister o f N yam uw anga, a t Shika V illage
hi 1953 touched o ff a social drama w hich made o vert m any o f
the hidden tensions both w ithin M ukanza V illage and between
M ukanza and Shika Villages.
Two Headmen Dispute over a Death Payment
(my own observations)
O n July 6th, 1953, a messenger arrived one somnolent afternoon
at Mukanza Village with news from Shika that Kalusa (LG19) was
dead. Instantly w ailing broke out from all the Mukanza women who
rushed into the central clearing between the huts and sat or stood
each separate from one another. In a short while the plaints became
organized into songs o f grief, led by Manyosa (I,Gi3), who sat with
her hands crossed on her shoulders, shaking and nodding her head
while tears ran down her free. Am ong her rhythmic cries could
be heard brief phrases, 4 She died all alone \ * m y father and mother
and m y unde Kanyombu are dead*, a death by andumba [a term for
witchcraft frmiliars or tuychela] . A t first all the women had been
freing in the direction off Shika Village. When the last phrase was
uttered all turned in the direction o f Sandombu Farm with anger

P olitical A spects o f K insh ip and A ffin ity


on their faces. Men had joined the bewailing group, although in

keeping with custom they did not weep. Sandombu (I,Gio) came
slowly towards them from his ow n place. He came forward into
the middle o f the dispersed group and a harsh argument developed
between him and Nyakalusa, the slave wife o f Sakazao (LH9),
privileged as we have noted already by virtue o f her role as an *out
sider * to speak openly what was in the minds o f all. Chayangoma
(i,H i5), classificatory sisters son o f Kalusa and Nyamuwanga
(f,G i7), and classificatory brother o f Sakazao, all in malabu lineage,
also joined issue with Sandombu. Chayangoma since he was, as we
have described, a blind man, as well as a close lineal kinsman o f
Kalusa, was also free to speak his views. They attacked Sandombu
for harbouring a witch in his farm, meaning Nyamuwanga. Nyamu
wanga had recently visited her sister at Shika Village and they recalled
how Nyamuwangas sister Shimili (I,Gi8) had died at Chimbili
Village after Nyamuwanga had visited her, and how Xkubi (I,Hi4),
Kalusas daughter, whom in the words o f Mangaleshi (II,E2),
Kasondas wife, * Kalusa had followed , had died after receiving a
malediction from the old witch. Sandombu shouted back at them
that there had been no proof that Nyamuwanga was a witch. She
had loved her sister, he said. He returned to his house in animated
conversation with Chayangoma (I,H i5) who was led along beside
him by Kasondas small son who was incidentally his fathers confidant.
Later the men assembled in the chefta. It was decided that they
should send round to the other villages to buy beer for the funeral
gathering (Chipenji) o f Kalusa (LG19) which would be held in Mukanza
Village. It was also decided to send a small delegation to Shika to
ascertain more fully the facts o f the situation and to make arrangements
about the funeral gathering. Feelings had calmed somewhat, following
a rumour that Headman Shika (HJDs) was contemplating a witchcraft
action in a chiefs private court against unspecified Mukanza. people
It was suggested that in this w ay he wished to evade the payment o f
mpepi to Mukanza Village, since i f he could get his charge upheld the
latter would be culpable and his own group could not in any way be
held responsible. Mpepi is paid by the kin o f the relict, not in com
pensation for their witchcraft, but inter alia for failing to protect the
deceased spouse from real or supernatural harm. But i f her own kin
had bewitched her, Kalusas arnnes need not pay them mpepi. This
rumour, which may have been inspired by someone in Mukanza, had
the effect o f uniting the village against Shika. They discussed the
question o f mpepi, which is divided into two consecutive payments, a
small sum called kutotafu (* to take away the dead ) for ritual services
rendered to the w idow or widower, and the mpepi proper, and
decided to ask only 2s. for the former since they could not obtain




( ;


( y



Schism and C ontinuity in an A frican Society

enough beer or provide enough food at that time for a long funeral
gathering (in the past this sometimes lasted three or more months
accompanied by elaborate ritual). The mpepi proper was to be left
over for negotiation between the tw o villages after the all-night
Mudyileji dance which concludes the funerary ritual. The emissaries
to Shika were as follow s: Mukanzas sons, Gideon (H,Ei) whose
mother came from Shika, and Zachariah (XJ3), uterine nephew o f
Sakazao (I,Hp) ; Short (I,Hi6), younger uterine brother ofChayangoma (I,Hi5) ; Zuliyana (IJz), Mukanzas daughter and Zachariahs
sister ; and Nyakalusa. The women were included only because the
deceased was a woman. Village unity had been restored to such an
extent that Nyamuwanga (I,Gi7) was nominated as executrix
(1nyamuju, * mother o f the dead ). She would receive and distribute
Kalusa s property, sell her Shika gardens, and take charge o f the
mpepi paid by Headman Shika. This nomination was a demonstration
o f the solidarity o f the Mukanza group in the face o f the outside world.
Normally this office would devolve upon the uterine brother or
mothers brother o f a dead woman but Kalusa had no brothers and
Nyamuwanga was her closest uterine relative. Sakazao, as head
o f malabu lineage, might have been appointed but since he had
recendy been fined for not returning the marriage payment to the
divorced husband o f Kalusa*s daughter Evinaki (I,Hi3) and was
looking round desperately for cash from whatever source, it was
thought inadvisable by the majority o f village elders to entrust him
with the office o f executor.
B y the mechanism, therefore, o f the chota debate a crisis which
threatened to activate the hostilities latent in the village was brought
under control and the ranks closed against Shika Village.
The next stage o f this social drama followed the month o f the
Chipenji gathering, during which Ntololu (II,Dz) kept the customs o f
widowhood at Mukanza Village where a grass hut (nkunka) was
built for him. O n the night o f Mudyileji (the final dance), Headman
Shika and other elders o f his village, together with many women,
came to Mukanza. Nearly every village in the vicinage o f Mukanza
Village sent a contingent. The general atmosphere was euphoric and
no hint was afforded o f the acrimonious debate that was to take place
on the following day. Both Mukanza and Shika danced, and much
apparent good-humour characterized their relationship round the
beer pots.
Next day the real business began, the controversy over mpepi.
Three mahaku were called in to judge between the contending parties,
Headman N gombi (see p. 200), Headman Mulila who had come
some years ago from Angola into the vicinage, and Headman
Nswanamatempa. None o f these men had close consanguineal ties

P olitical A spects o f K inship and A ffinity


witli cither Shika or Mukanza, although N gombi had married

Mukanzas daughter Koniya (I,Ji). But N gombi*s reputation for
scrupulous fairness made him immune from the charge o f favouritism.
Shika called in Headman Nswanakudya (IIJD10), whom, it w ill be
remembered, he called * older brother * (p. 208 et seq.), to assist him
in putting forward his point o f view. He was also supported by
the widower Ntololu (D^Dz) and by a classifrcaeory sister s son from
Shika Village. O n the Mukanza side, Mukanza Kabinda himself
spoke, assisted by Sandombu, Chayangoma, and Sakazao. Nyamuwanga refused to discuss the question o f mpepi, saying that she had
been despised * (ku-sawula) by the Shika people on many occasions
and would not hear* herself insulted again. Her place was taken,
curiously enough, by Nyakalusa, who had originally accused her o f
witchcraft. Once more Mukanza people had closed their ranks.
The discussion took place outside the chota since it was not about
internal village affairs. Sandombu opened the debate by demanding
a sum o f ^ 5 as mpepi. Ntololu, he alleged, had made it clear during
the Chipenji funeral gathering that he intended to retain with him in
Shika Village his tw o daughters by Kalusa, so that their services would
be lost to Mukanza. Furthermore, * bad words * had been spoken
by Shika people about witchcraft in Mukanza and it was only just
that they should make amends by paying a big mpepi. He went on
to say that X5 was not such a large sum after all b y modem standards
and mentioned cases in which 7 and even io had been paid.
Shika, a big, Blustering man, who, h is said, had once tried to bribe
Kanongesha to appoint him to the Chibwikaship (see Chapter Eleven),
hotly disputed the sum named, pointing out that the two villages had
for many years been in the habit o f exchanging wives, and had come
to he interrelated by many ties. It was an unfriendly act to make
heavy demands on ones kin. Besides, he recalled that when Nyattmgeji (H,D3) was on the point o f death Mukanza had divorced her,
only 2$. 6d. to Shika. Again, at the beginning o f Chipenji
sent 1 0 5 . to Mukanza by Ntololu to contribute to the cost
o f divination into the death. W hat did the chicken say ? he went
on, referring to the poison test (ng'ombu yamwaji) administered to
fowls in one mode o f divination. It would appear that Mukanza
Kabinda had not gone to a diviner but had * eaten * the money. D id
Mukanza think that he was a muloji (sorcerer) to ask him for such a
large mpepi ? Mukanza him self then spoke, saying that he had
no quarrel w ith Shika but that the latter must know that it was the
custom nowadays to pay a large mpepi since * every kind o f food
and beer now cost money, and cloths, pots and pans were very dear .
He had not gone to a diviner, for, i f such a matter came to the car
o f the Boma, all o f them might * receive a very heavy case *. But


Schism and Continuity in an A frican Society

he had taken the money already paid into account when he had
decided in consultation with his fellow-villagers on the amount of
mpepi he would ask for.
Other elders now spoke on both sides, presenting variants on the
main arguments. After a time the judges were consulted and they
suggested that Shika should pay 2. 10s. Headman N gombi said
that bitter things had been said on both sides and it had to be taken
into account that Mukanza had divorced Nyatungeji in order to avoid
the paymcnj o f mpepi. But since many people in both villages were
kin, it would be a bad thing if one group started to nurse a grievance
against the other. He thought that Mukanza should be satisfied
with half the amount demanded, in view o f the Nyatungeji affair,
o f kinship relations between the villages, and o f what Shika had
already paid as a contribution towards a divination which had never
taken place. Sandombu, speaking on behalf o f Mukanza Village,
refused to accept this, saying that while they would no longer press
for f s they would be content with nothing less than .3 10s. After
more argument Shika agreed to pay this sum. But he told the
Mukanza people that they were mean and selfish to drive such a hard
bargain. He said that he could only pay j i that day, as he was a
poor man (kazwengi) but that he would pay the rest later in instal
ments. He then left the village * in a huff* (ku-fundamoka) with his
followers. Kalusas children remained at Shika Village, and for
more than six months there was a marked reduction in visiting between
the villages, and much mutual backbiting. In the last chapter I shall
describe how one o f the functions o f the Chikamba ritual I witnessed
was to effect a reconciliation between the villages.
Social Drama VII illustrates a number o f points already made.
Am ong these m ay be noted : (1) the w ay in w hich inter-village
disputes help to prom ote intra-village unity strikingly brought
out by the change o f front shown by Chayangom a and N yakalusa, w ho began by accusing Sandombu and N yam uw anga o f
witchcraft and abetting w itchcraft and ended by jo in tly arguing
the case for M ukanza V illage, and even, in Nyakalusas case, by
acting as representative for the one initially stigm atized ; (2) the
interest shown by representatives o f the w ider social system o f
the vicinage in w hat was prim arily a m atter concerning tw o inter
m arrying villages ; and (3) the role o f m atriliny in providing
links between villages in different vicinages, exem plified by

P olitical A spects o f K insh ip and A ffinity


N sw anakudyas advocacy o f his classificatory brothers case.

But its m ain interest for us in this chapter is the ligh t it throws
on the kinds o f cleavages and ties between interm arrying villages.
E very breach in a m arriage, whether made by divorce or death,
is accompanied by conflict between the villages, but the existence
o f other ties o f affinity and kinship prevents the conflict from
rupturing all ties between them. Cross-cutting affiliations ally
those w ho are in conflict. Kasonda and M anyosa, for exam ple,
w ould not argue on behalf o f M ukanza V illage. T h ey said that
their father Kahum pu came from Shika V illage and that he was
the full-brother o f N tololu. Kasonda said that in addition his
ow n children belonged to Shika V illage through their m other.
H ow then could he speak against men w hom he called 4father ,
and w ho w ere grandfathers o f his ow n children, i.e. members
o f the same linked-generation segm ent in Shika V illage ? Again,
M ukanza him self spoke w ith great m oderation, fo r his three
children b y N yatungeji belonged b y m atrilineal descent to
Shikas follow in g. B y another reckoning, through M ukanzas
marriage w ith M angaleshi, he was a * co-parent-in-law (nktilu
nami) w ith Headman Shika, since he was muhu to M angaleshi,
and Shika was, in the context o f affinity, muku to Kasonda.
Persons w h o call one another nkulu nami, m ay sleep in one
anothers huts, receive hospitality from one another, and in
general behave w ith considerable fam iliarity, although there
is no jo k in g between them. T o m y m ind this is another ex
pression o f the solidarity o f genealogical generations. Sandombu,
on the other hand, had no close kinship or affinal tie w ith Shika
Village to prevent his speaking the M ukanza case unreservedly.
Shika recognized in him the authentic voice o f M ukanza Village,
when he asked fo r ^ 3 10s., and agreed to pay up to preserve
peace between the villages because he knew Sandom bus demand
to be a final offer.
B ut although the villages did not break o ff relations it was felt
by the Shika people that M ukanza V illage had had the better o f
the bargain. Resentm ent lingered ; Shika people complained
in m any villages o f the meanness o f the M ukanza group, coupling
their reprobation o f the village w ith disparagement o f the
K aw iku as a w hole. T he latent conflict between Lunda and
K aw iku, the invaders and the indigenous people, never entirely
damped dow n, began to sm oulder. M oreover, the people o f










Schism and Continuity in an A frican Society

M ukanza felt rather uneasy about their ow n conduct in the

affair. M embers o f other villages hinted that they had not
behaved w ell in demanding such a large sum as ntpepi. W hen
it was decided to hold the Chihamba ritual for N yam ukola 1
a special invitation was sent to Shika V illage to attend. N tololus
children came and a number o f children linked to both villages
w ere chosen as candidates for initiation. A t the ritual level
both villages w ere tem porarily reconciled.
B ut outside this relationship o f m arital exchange w ith Shika
V illage the people o f M ukanza had a w ide and extensive range o f
marital connections w ith other villages. Even divorce and death
do not break such ties, i f the union has been productive o f
children, for children act as a focus o f com m on interest between
the maternal and paternal groups. Children w ho live w ith their
m other often exchange visits w ith their father. In custom, a
man must show no animosity towards his ex-w ifes present
husband w hom he calls nkulu nami, the same term as he applies
to his childs spouses parents or uncles and aunts. This rather
tense relationship o f outward friendliness prescribed b y custom
enables a man to visit his ow n children at his divorced w ifes
hom e, w hether she be livin g in her ow n or in her new husbands
village. The customs are directed against the outbreak o f physical
violence due to jealousy. B ut i f a couple have had no children,
the relations between their respective villages w ill nearly always
cease on their divorce. Affinal ties are strengthened enormously
w ith the advent o f children, in whose w elfare the groups to which
each spouse belongs by various principles o f attachment have an
I exam ine the extant inter-village marriages o f M ukanza people
and the fruitful marriages o f present members w hich ended in
divorce and death. Starting w ith members o f n yach intang a
lineage, in order o f seniority, M anyosa (I,G i3) was married
to Chikasa, o f M pawu V illage in the deposed Governm ent SubC h ie f N tam bus area. She had previously been married to
Chisem pi, w hom she divorced and b y w hom she had five
children ; he was a member o f a village in the vicinage in which
Nyaluhana (see p. 48) is principal headman, in C h ie f Kanongeshas


1 See p. 310.



Schism and Continuity in an A frican Society

area. Her oldest son lives w ith his father, although he sometimes
visits Kafum bu V illage, and she goes to see him w hen she visits
Kafum bu V illage w hich is near Sawiyem bi V illage, w here her
form er husband and her son reside. A ll her other children,
except R.osina (I,H8) w h o lives w ith her husband in Kafum bu,
reside in M ukanza V illage. Chaw utongi (I,G i2 ), whose
husband N deleki (II,D8) lives uxorilocally, provides a link with
N g om bi V illage through N delekis matrilineal connection there
and w ith Shika through N delekis hither. Kasonda (I,G i5),
by his divorced w ife M ika, w h o had a baby b y a lover, main
tained connections w ith N tam bus area. These connections,
though snapped b y divorce, between Kasonda and M ika, were
renewed b y Kasondas m arriage w ith ex-C h ief N tam bus sisters
daughters daughter Luciana. nyachintang a lineage had
maintained relations b y affinity w ith the ntambu chiefly lineage
for several generations. T he Ntam bu chieftainship was reckoned
equivalent in status to that o f Kanongesha in Lunda tradition :
C h ie f N tam bu sent tribute directly to M w antiyanvw a as Kanon
gesha did in the past, and the tw o chiefly lineages interchanged
spouses ju st as Shika and M ukanza V illage lineages did. The
father o f M ukanza Kabinda and N yam w aha, Santapu, was the
younger uterine brother o f an incum bent o f the N tam bu chief
tainship. T he marriages o f N yam w ahas son and daughter
maintained the connection. Kasonda had in fact m arried quite
astutely. His children b y Mangaleshi m ight succeed or bear
successors to the chieftainship o f Kanongesha. His children by
Luciana m ight succeed to the stool o f N tam bu. Kasonda
cherished hopes that Governm ent w ould eventually restore that
chieftainship to its form er position in the N ative Authority.
His son by Luciana m ight then perhaps becom e a Governm ent
C h ie f in that case. I f Jerry, his son b y M angaleshi, became
Kanongesha, Kasonda m ight be the father o f tw o chiefs.
Sandombu kept up Intermittent contact w ith the daughter or
his first w ife N yalundaw u. This daughter, regarded b y Sandombu
alone as his, had m arried into a village in Ikelenge Area and had
a small daughter. W hen Sandombu visited Ikelenge Area on
business he always stayed w ith his * daughter *. Sandombus
sister M angalita (I,G i i) had form erly been married to a man from
nearby Nswanakudya V illage, but in 1951 she was living at
K afw eku V illage w ith her new husband. K afw eku, form erly a

P olitical A spects o f K in sh ip and A ffin ity


senior headmans village, had previously been linked w ith

M ukanza b y the uxorilocal m arriage o f Kajata (I,Fro), o f malabu
lineage. It is in M w ininyilam ba Area, far from M ukanza, on
the border o f A ngola.
Bibiana (I,H4), m others sisters daughters daughter o f
Sandombu, also lived w ith her husband, a crippled ex-askari, in
Chim bila V illage near Ikelenges capital village. She had three
sons and tw o daughters. T he sons regularly visited M ukanza
in the school holidays, and w hen I left, Bibiana, divorced b y her
husband in 1954, was talking o f settling in Sandom bu Farm.
Bibianas first parallel cousin, N yam pesi, was m arried to a store
capitao w orkin g for a European trader, near the Bom a. She
made frequent visits to Sandom bu Farm. H er husband had a
useful role from the point o f view o f M ukanza V illage in placing
the younger m en in paid jobsM anyosas oldest daughter Kandeieya was m arried to M akanjila, son o f Headm an M bim bi, w hose decrepit village im m ediately
adjoined M ukanza. M anyosas son Daudson (fH y ) had tw o
w ives. O ne was the daughter o f an O vim bundu im m igrant
w ho had m ade a farm near M ukanza. T he other war. Sakutohas
daughter (see p. 191 et seq.) from her fathers adjacent settlement,
w hich paid tax through M ukanza. H er other daughter Rosina
(I,H8) was m arried to Kantam oya (I,H 3), the younger brother
o f the headm an o f K afom bu V illage.
In malabu lineage Sakazao (I,Hp) was connected to Kafum bu
by m arriage through N yaluw em a (IC4). H is slave w ife
N yakalusa no longer seemed to m aintain contact w ith her
Angolan hom e village. N yatio li (I,H n ), his sister, had lived
in M w ininyilam ba area fo r m any years w ith a husband from
Nyansengi V illage. H er daughter Yana (IJ4) had married a
man from the neighbouring village to N yansengi, Kam bim hi
V illage. In 1951, w hen Y ana was still livin g virilocally, I
visited K am bim hi V illage, far from the m otor road along a path
w inding over troublesom e spurs o f rock, w ith a number o f
M ukanza people ; and w e w ere all given hospitality by Yana
and her husband, in the form o f fo o d and beer. Yanas brother
Pearson (I,j5) was m arried to a g irl from N g om bi V illage in
the vicinage. T h eir younger sister Serita had previously been
m arried to a man from Ibulu V illage in C h ib w ik a A rea, b y w hom
she had had a daughter w ho was livin g w ith her father. R ecently


Schism and Continuity in an A frican Society

she married a man from N swanakudya V illage in the vicinage

o f M ukanza. N yam uw anga (I,G i7) o f M ukanza had form erly
been married to a man from Kafum bu V illage, by w hom she
had borne a daughter. This man subsequently died, and his
death was popularly supposed to have been caused b y N yam uw an gV s w itchcraft. B ut her daughter Ikayi (I,H i2) still visited
her fathers village from tim e to tim e. N yam uw angas sub
sequent marriage to a man from Shika V illage has already been
mentioned. N yam alita (I,G zo), the surviving uterine sister o f
N yam uwanga, lived viriiocally for m any years w ith her husband,
a farm headman in Ikelenge Area, and later obtained w o rk as a
W ard M aid at the Kalene M ission Hospital. H er younger
children, like those o f Bibiana (see p. 164), frequently visited
Mukanza V illage in school holidays, and it was expected by
M ukanza people that they w ould look on it as their hom e when
they grew up.
Chmema (I,G 2l) and his uterine siblings remained in Ikelenge
Area after Kahali Chandenda (I,F5 ; see p. 180 et seq.) had
returned to M ukanza V illage. Together w ith some disgruntled
claimants for the M ukangala chieftainship he established a new
composite village, called after him. H e and his siblings married
in Ikelenge Area and I shall not enumerate the marital and consanguineal ties that n ow interlink the matrilineage o f M ukanza
through Chinem a and his siblings w ith villages in Ikelenge Area.
Some o f Chinem as uterine nephews have n ow returned to
M ukanza, including Chayangom a (I,H i5) and Short (I,H i6 ;
see Social Dram a V II, p. 269 et seq.), neither o f w hom w ere
married at the tim e o f w hich I am w riting. Lines senior w ife
came from Sawiyem bi V illage, near Kafum bu V illage, and he
frequently accompanied her on visits to her village. His jun ior
w ife, M aria, as noted on page 186, belonged to the lineage o f
n y a c h u l a , led o ff from M ukanza V illage b y Yim bw endi, as
described in Social Dram a V I (p. 181). O f the nyachujla people
w ho remained in M ukanza V illage, N yaw unyum bi (I,G2s)
through previous marriages w hich w ere fertile provided links
w ith villages in C hibw ika and Kanongesha areas. H er son
Kenson (I,H 2i), w ho resided at M ukanza Village, was married
to a wom an from nearby Chibw akata V illage. She herself, as
already mentioned, was m arried to Headman M bim bi, and her
youngest daughter (I,H2o) was married to a man from Kasai

P olitical A spects o f K in sh ip and A ffinity


Village, a K aw iku village in another vicinage (see p . 191 et seq.).

Aram (IJ 7), Lines sisters son, was m arried to a w om an from
Chikanga V illage, in the same neighbourhood as Kasai Village.
O f N yam ukolas children, her daughter K oniya (I,Jr} was
married to Headman N g om bi, w hile her son Zachariah (IJ3)
was m arried to the sisters daughter o f Farm Headman Sakutoha
(see pp. 192-3), close to M ukanza V illage.
Affinal ties, then, linked M ukanza V illage to other villages in
the same vicinage, to villages in adjoining vicinages, and to
villages in distant vicinages and even in different Governm ent
Chiefdom s. B y m arriage the village was interconnected w ith
villages in C h ibw ika Area fo rty o r fifty miles to the south, and
with M w ininyilam ba Area about the same distance to the north
west. Genealogies suggest that similar w ide ranges o f marital
ties interlinked villages in pre-European times. Num erous close
ties o f kinship and affinity w ith Shika V illage can be placed at
one end o f a scale, on the other end o f w hich w e find single
threads o f connection radiating in all directions to settlements
near and far.
Table X X I, based on village census m aterial, gives some idea
o f the spatial range and dispersion o f m arital ties. I was un
fortunately unable to w ork out in m any cases w hat marriages
had taken place between members o f villages belonging to die
same vicinage. M ukangala is classed as a * chiefilom *, although
the incum bent was deposed b y the Governm ent in 1947. The
high proportion o f extra-chiefHom marriages in the Ikelenge
sample is probably due in large measure to the high im m igration
into the area from A ngola, the B elgian C on go and from other
parts o f M w inilunga D istrict (see p. 9 and p. 59). The table,
despite its deficiencies, does show how w id ely the netw ork o f
affinity, w hich interlinks villages, vicinages and chiefdom s, is
Kinship and Affinal Connections between Headmen in a Vicinage
Before discussing the netw ork o f affinal and kinship ties
interlinking the headmen in the vicinage o f w hich M ukanza
Village form s part, a few w ords should be said about polygyn y
and about the polygynous marriages o f headmen in particular.
Table X X II gives the age and conjugal status o f all members o f


Ike Ikdenge

Muk Mukangala
Nya Nyakaseya
Mwi Mwininyilamba


Spatial R angb

P olitical A spects o f K in sh ip and A ffinity


24 villages and 6 farms, w hile T able X X III compares the age and
conjugal status o f 28 headmen w ith those o f 107 commoners o f
thirty years old and over.1 M ore than h a lf the headmen had tw o
or m ore w ives w h ile on ly about a sixth o f the com m oners had
tw o or m ore w ives. Headmen have greater obligations than
commoners to offer hospitality, and the m ore w ives they have the
better are they able to do this. B u t m arital ties betw een headmen
and members o f other villages are politically stronger links than
those between com m oners, since the headman represents a group
and sometimes inherits a tim e-honoured title. M en often,
like Kasonda, anticipate headmanship b y judicious polygynous
marriage w ith im portant maternal descent groups ; others
find that they m ust m arry m ore than one w ife to fulfil their
duties after they have succeeded. Since also the children o f head
men often remain w ith them , the polygynous headman w ith
fertile w ives is enabled to beget a follow in g w ithin the social
system o f the village o f w hose lo ya lty he can be almost sure.
C onflict and tension between co-w ives is, how ever, a frequent
cause or excuse fo r divorce so that p olygyn y, w hile it offers
advantages to the am bitious, m ore than doubles the possibilities
o f conflict, never absent from the N dem bu m arital relationship.
Before 1952 o n ly five villages com prised the vicinage (chi
tting*Hi or itung a) to w hich M ukanza V illage belonged. These
were N s wanakudya, Chibw akata (in a different site, n ow in a
Governm ent Forest R eserve), N g om bi, M ukanza, and M bim bi.
N o w N sw anakudya and Chibw akata w ere interlinked by a
number o f m arital and consanguineal ties. T he headman o f
N swanakudya was the son o f a Chibw akata man, K aluw a, w ho
was the m other's brother o f the headman o f Chibw akata. Thus
the tw o headm en w ere cross-cousins. Headman Nswanakudya
married N yachipoya, classificatory sister o f Headman C h ib
wakata. W h en N g om bi split o ff from Chibw akata, N yachipoyas uterine siblings w ent w ith N g om bi. Headman N swanakudyas oldest son b y N yachipoya later settled in N g om bi
V illage w here he was mulopu o r deputy o f the headman. Head
man N g om bis m other was older uterine sister o f N yachipoya.
Several other marriages interconnected the three villages but
1 The lay-out o f these tables is based on Table X IX in Mitchell and Barnes,
T he Lam ba Village , p. 4 6 .

;. ,'*r* *\>



A ge

Sample of 24 villages and 6 farms:


C onjugal Status

Number in sample



Males :

Percentage in each age group



Total .


Females :


Total .




Total Unw.

































i *7























52*8 100*0


2-0 100-0
60*8 100*0
86-0 100-0
94*5 100-0
84*4 100*0
70-4 100*0
40-0 100*0


4 (5*7



iS -5

n *l



3 *7 ,

54 *6


Past i . A ge



I li
. !

PIS pf!

C onjugal Status





Sample of 107commoners:

2* A ge











ip P lS ;







C onjugal Status






Total Unw.

Percentage in each age group





iM^h p i i l l

Number in sample



4 (5-7



70*6 100*0
87*2 IOOO
__ 90-3 100*0
ioo-o 100*0
2-9 85*7 100*0


Sample of 28 headmen

Total .


PPPp V ;



Vj X ? P

Total .



30-39 .











68*8 25*0

100*0 100*Q






100*0 100*0
100*0 100*0
100*0 100*0



100*0 100*0


100*0 100*0
73*7 100*0


87-8 100*0


C ommoners (male)









Unw. Unwedded
D Divorced, and not remarried
R Relict; widower or widow, not remarried
! j 2, 3, 4 1, 2, 3, 4 Spouses
Pu Puberty

86*8 100*0






Schism and Continuity in an African Society

those mentioned w ere politically the m ost outstanding. N g ombi

Village, it w ill be remembered, split o ff from Chibw akata Village
in the 1920s.
M ukanza V illage was linked w ith M bim bi V illage in a number
o f ways. T he headman o f M bim bi was the son o f a m em ber o f
malabu lineage in M ukanza V illage, and his son w as married
to Kandeleya, daughter o f M any osa o f nyach intang a lineage,
the favourite * granddaughter o f M ukanza Kabinda. M bim bis
second w ife was N yaw unyum bi, m other o f M ukanza Kabindas
second w ife and Lines second w ife. The prim ary relationship
between M ukanza Kabinda and M bim bi was that o f classificatory
* father to * son , since M bim bis father Ibeleka was classificatory
older brother to M ukanza. Through subsequent marital
ties M ukanza could be considered the classificatory son-in-law
o f M bim bi, since M bim bi had married M ukanzas m other-in-law
N yaw unyum bi. B y matrilineal affiliation M bim bi belonged to
the descent group from w hich incumbents o f the M ukangala
senior headmanship w ere chosen.
Betw een these tw o groups o f closely related and intermar
ried villages, N sw anakudya-Chibw akata-N gom bi on the one
hand, and M ukanza-M bim bi on the other, stretched threads o f
genealogical connection. Headman N g om bi was the husband
o f M ukanzas and N yam ukolas daughter K oniya, form er wife
o f Kasonda, M ukanzas uterine nephew. Few m arital ties
directly interlinked Nswanakudya and M ukanza, but the close
relationship o f marital exchange w hich existed between Shika,
from w hich N swanakudya had seceded, and M ukanza, b y classi
ficatory extensions o f kinship terms and behaviour, made it
possible for M ukanza Kabinda and Headman Nswanakudya to
call one another ishaku, * brother-in-law *. Kasonda called
Nswanakudya father since N swanakudya was the classificatory
* older brother o f his father. Kahum pu. T hrough the marriage
o f his sister N yam w aha to the previous headman o f Chibw akata,
M ulow a, M ukanza could also call the present headman o f
Chibw akata, Chonku, classificatory younger brother o f the late
headman M ulow a, his brother-in-law (ishaku). Kasondas
fathers father came from Chibw akata. A few marriages between
younger people interlinked the tw o villages. B u t on the whole
few ties o f kinship and affinity connected M ukanza V illage with
these three villages. M bim bi once m arried a w om an from

P olitical Aspects o j K insh ip and A ffin ity


Shika V illage w h o died m any years ago, but o n the strength

o f this form er m arital tie, he could call Headman N swanakudya
* brother-in-law * (ishaku). I could not trace any genealogical
connection between M bim bi and Chibw akata.
In 1952 Governm ent m oved a num ber o f villages out o f the
vicinage in w h ich N sanganyi was the principal village w hen
their form er territory was declared a Forest R eserve. These
villages w ere told to build in the vicinage o f M ukanza V illage.
M ap 3 (p. 21) shows the distribution o f these villages and their
gardens in 1953. It is instructive to note the w ay in w hich kin
ship, affinal and historical connections influenced the choice o f
their n ew sites. Nsang*anyi w ould not build near its offshoot
Nswanam atem pa, in view o f the tension that still existed after
the secession o f the latter (see p. 212) as recently as 1940. N or
w ould N sanganyi build dose to M ukanza, w h o claim ed to be
mwenimbu, principal headman, o f that part o f the vicinage
aborning the Governm ent m otor road. N sanganyi was the
senior and m ost ancient K aw iku village and could n o t recognize
the local pre-em inence o f any o f its offshoots. Nswanamatempa
broke up as described on page 216 into three settlements, all o f
w hich built near M ukanza, a fellow -K aw iku but distantlyrelated village. A nother group o f villages and farms, settled on
the ve ry verge o f the Forest Reserve, w ere interrelated w ith each
other and w ith villages already in the vicinage. Itota, the head
man o f one, was m arried to the sister o f the headman o f another,
named Lambakasa. Lambakasa and his sister w ere in their turn
siblings o f Headm an M bim bi. T heir father cam e from M ukanza
Village, from malajbu lineage. Lambakasa was a fu ll brother o f
M bim bi ; Ifcotas w ife was their half-sister. H er full-brother
Nsanji built a small farm opposite the road from Lambakasa, his
half-brother b y the same father, and on the same side o f die road
as Itota. In this group also was a form founded b y a wom an
headman N yam um w em a w hose daughter was Lambakasas
junior w ife.
N ext to N sanganyi V illage on the Itota side was Kanyahu
Village. Headm an Kanyahu had m arried N sanganyis uterine
niece. K anyabus ow n uterine niece had once been m arried to
Sakazao o f M ukanza Village, where; indeed, she had died. Her
daughter b y a previous husband was the w ife o f Gideon, the son
o f M ukanza Kabinda.


Schism and C ontinuity in an A frican Society

These villages then, w ere intricately interlaced by affinal and

kinship ties. N o ties o f either type existed between Mukanza
and Itota. In this case, how ever, the headmen made the contract
o f blood-brotherhood (wulunda). M ukanza used this relation
ship to secure for nothing the services o f Itota as senior ritual
specialist in the Chihamba ritual, to be described in Chapter Ten.
A fter 1952 also, a few small farms grew up in the vicinage.
Anderson M ulum bi (see Social Dram a V , p. 160 et seq.), son o f
a form er female slave o f Chibw akata, made a farm between
M ukanza and the Nswanamatempa group, and set up in business
as a sewing-machine tailor. His w ife was sisters daughters
daughter o f Nsanganyi, and m any disputes arose between the
K aw iku and Anderson over his cruel treatm ent o f her. Andersons
sister, married to an O vim bundu ritual specialist w ho had intro
duced the new ritual o f Tukuka into the vicinage, occupied the
next farm. N ext to her again was another O vim bundu, a petty
trader whose daughter was married to Daudson, M anyosas son,
o f M ukanza V illage. Beside him was the small farm o f Muchona
whose w ife belonged to the Nswanakudya village lineage. He
was a close friend o f the O vim bundu w itch-doctor and they
jo in tly officiated at several kinds o f ritual. N ear Chibwakata,
the younger uterine brother o f the headman, M akayi, made a
fam ily farm w ith his tw o w ives and children.
In the absence o f strong overall political authority, such ties
o f kinship and affinity becom e o f the utmost im portance in giving
some degree o f cohesion to a fluid and volatile society. B ut since
strong possibilities o f conflict exist in these very relationships
themselves, kinship and affinity, unsupported by other kinds o f
linkage, do not constitute reliable political bonds between
villages. Even where, as in the case o f M ukanza and Shika,
conflicts w hich arose between tw o villages over a broken mar
riage could to some extent be contained b y the countervailing
influence o f other ties o f kinship and affinity w hich interconnected
their members, tensions w ere set up w hich seriously reduced the
frequency o f interactions between them . Indeed, such tensions
m ight, w ithin each village, give rise to incessant bickering
between its ow n nuclear lineage and affines and kin belonging to
the lineage o f the other. In other words, the instability in the
marriage relationship, arising from the conflict between matrilineal descent and virilocal marriage, was at the same time a

P olitical A spects o f K in sh ip and A ffin ity


source o f instability in the political relationships between groups

allied by marriage Kinship and affinity in such a society divide
as often as they bind. T h ey provide brittle foundations for a
social p o lity. T im e and again one notices the tenuity o f the
kinship and m arital ties that hold together the com ponent elements
o f local groupings o f w hatever kind. Villages constantly shed
uterine sibling groups w hich becom e potential nuclei o f new
settlements. Sometimes they divide at a higher order o f lineage
segmentation. Individuals m ove from village to village in
pursuit o f their im m ediate interests. M ost w om en m arry ou t
at least once. Betw een villages, strife is endem ic at the level
o f consanguineal and affinal relations. T he best neighbours are
not ones closest kin- Since m any kin participate in marriage and
death paym ents, and since these are nearly alw ays subjects o f
dispute in w hatever context o f social relations they m ay be
considered, conflicts between villages over them are continually
taking place. Such conflicts seldom reach a solution satisfactory
to both parties. The contradiction between principles o f social
grouping within villages inhibits the grow th o f deep lineages, tends
to rip the m atricentric fam ily prem aturely from the m atrilineage,
sets adjacent generations at odds w ith one another, divides sisters
from brothers, and enfeebles the bonds o f m arriage. Betw een
villages, the same contradiction engenders disputes, makes o f the
vicinage an unstable grouping, and disperses kin throughout the
w hole N dem bu region. U nlike the societies described in
African Political Systems as * segm entary V in w hich lineage
organization form s the permanent fram ew ork o f social relations,
the cohesion o f N dem bu society depends on the strength o f its
mechanisms fo r resolving conflict w inch arises from the weakness
and instability o f lineage organization and from the ambiguous
character o f kinship ana affinal ties.
1 Fortes, M. and Evans-Pritchard, E. E., * Introduction * to African Political
System (1940), p. 10.

w ifoout-govem m entaL-m stitations ,-.~sociaL. cohesion-appears
to dep<md~on._[foe]._divirion of-society-intO-.a-series-ofLopposed
groups,..w ith -crosscutting membership
Again, he says, that
social ties w ere - . . established to link together people who
in other contexts w ere enemies \
Nd^mbw. village,...cohesion.^tsoiLTO.^isanedJhy the
many*?sided. social affiliations o its ..camponent,.graups and
persons.. In .the. .same w ay,.villages, are,, interlinkexl .in. vicinages,
in senior headmens ,areas, a n d in the. N.dembu-.tribaL_r.egion, by
varying sets o f ties,: knislnp,,affiiiity, spatial _pro.pmqmty,. bloodbrotherhood, and so on. Analysis. _o.the_s_QC.iaLdranias..has shown
how^persons and.. g ro u p sd iv id e d -iii.o iie .se to fso o ia l relations
are allied in other .sets. But. the. analysis o f social -dramas has
also .shown that cohesiomis.,hymo-mean&,necessarily ,.or-invariably
maintained in., a., group outbreak ,o,con&ct,.between
persons, or .groups, w ith in .it.- -Eoninstance,-w.ehav.e^ seen that a
village consists_o.a number o f .persons .bound^together b y several
principles o f social.affiliation,^som eofwhichhecom e-dom inant in
certain .situations,, in spite ^ the countervailing -pressures exerted
by the. others, and w hich destroy -the-unity o f the village.
Thus,in m ost cases_ o.fission...examined,jth^..pdnciple o f the
un ityjof the uterine sibling group prevailed against, such principles
as the,umty- o-thn minimal, lineage,,the-um ty,o$he genealogical
generation,- and. the, unity.,o f the. village. , . Again^ the, principle
o f the- unity o f the: uterin^ribIing.,group-,cam ^into continual
confkctiW ith th e .principle ,of.the..unity, o f. theTam ily : and the
success of-the form er can be measured in terms of-th e high fre
quency of divorce. B u t w e hav.e also .noticed .how . the unity
o f the w idest political .unit, the Ndem bu people, gains at the1

1 * Political Institutions *, in T he Institutions o f Prim itive Society, p. 67.

also Fortes, T h e Dynam ics o f Clanship Am ong the Tallensi, pp. 241 et seq.


The Politically Integrative Function o f R itu al


expense ,o f its significant local unit,.the village_Frequentfission,

associated w ith w ide-ranging spatial_m obility o -theu-secedmg
group, tends to create through tim e, numerous., .sets* o f .local
kinship nuclei, each set consisting o f villages w hich have arisen
from a "com m on source or from one another, by:.fission. T he
villages in each set, though ,.widely., scattered , in., space,.,.may be
interlinked b y b elief in com m on m atrilineal..descent,,, or b y
know ledge o f actual kinship connections. T h e separate sets are
interlinked b y affinal, vicinage and other ties. Thus .fission and
m obility, w hile they break up villages, interlock the nation, w hich
has n o effective overall -political-authority..- But-, when w e
examined m ore closely the political ro le -o f ues.of-km ship and
m arriage enjoining co-operation w e found that theseties-too. w ere
o f an unstable and friable nature. .
It w ould seem then that links o f kinship and- o f political and
econom ic co-operation per se are insufficiently strong to hold
together village and nation. A nd yet N dem bu are conscious and
proud o f their .unity as a discrete group, and as a sub-group o f the
great Lunda nation, the unity o f w h ich is expressed in -the kingship o f M w antiyanvw a.
Such a consciousness o f national unity .obviously depends on
the jo in t recognition b y all N dem bu o f a. com m on system o f
values and a com m on set o f norm s regulating behaviour. Their
unity_is a m oral, rather than a political, unity.... By_.what .social
mechanisms are these com m on norms and values inculcated and
periodically re-anim ated ? In this book I shall say nothing about
the educative process w hereby youn g children are indoctrinated
in the basic values o f their group. N o r, except in passing, have
I been able to deal w ith the jurid ical mechanisms. M ore im
portant perhaps than either as a means o f keeping continually
before the people their com m on norm s and values, in terms o f
sym bol, m im e and precept, is the com plex system o f ritual.1
The.profusicm of_types_and .fre.quency-.of performance, o f ritual
in N dem bu.society are, in. a w ay,, confessions, o f failure in the
pow er o f seculat mechanisms to redress and absorb, conflicts that
arise in.and between local and kinship have.suggested 1
1 1 do not intend here to make a cultural analysis o f Ndembu ritual but
simply to isolate from the ritual complex those sociological features which
are relevant in this book.




Schism and Continuity in an A frican Society

that the contradiction between virilocal marriage..and...matrilineal

descent is a crucial .determinant, o f structural instabiiity^at-all levels
o f organization. I w o u ld postulate that where the m ode o.f postmarital residence, is inconsistent., w ith ..the.mQde_ojo5ckonmg
descent in unilineally organized .sQcieties, local groupings in such
societies w ill be inherendy unstable, unless other ..factors intervene,
such a-s lim itations on access to resources. I. have tried to show
how the tenacity and tendency to autonom y o f the uterine
sibling group are related to virilocal marriage, and h ow the
tenacity o f tins group tends, in situations o f conflict, to jeopardize
the cohesion and continuity o f villages. I have also tried to trace
a relationship between virilocal marriage and the high divorce
rate, w hich, in its turn, impairs political relations between inter
married villages. So radically disruptive o f local cohesion is this
contradiction that not infrequently the com bined operation o f
all countervailing relations is powerless to prevent absolute schism
in villages.
M ost Ndem bu rituals tacitly recognize the. instability o f villages
and .o f the relations betweerL villages,, but .positL-the .ultimate unity
of-all-Ndem bu in.a.snigle.m oraLcom m unity. The.instability o f
kinship- and political relations is. recognized. in the fact, that the
dominant social elem ent in the-.composition, of-ritual assemblies
is not .a kinship group but an association o f adepts..who belong
to m any, kinship groups... Also,..the.. .dominant-symb-Qls in the
cluster -.of-sym bolic objects, and activities, .associated-with each
ritual-do not reflect or express major aspects o f structure,
but rather the values w hich all Ndem bu possess m on, such
as, for example, the fertility o f women,, crops .and. animals,
huntsmanship, health, and the pow er o f the ancestors-to .bestow
or w ithhold such benefits. The ultimate unity o f aU-Ndembu
is expressed in the com position o f ritual assemblies. Thus, all
Ndem bu, all Lunda even, have the right to attend any perform
ance "of ritual o f whatsoever typ e.... Indeed,-i theyJaave. been
initiated into its special techniques and esoteric, know ledge, its
* mysteries * {mpangu), they have the right Jparticip.ate as
officiants (ayimbuki},1* whether-they- are- related-to-the- -principal
subjects o f the ritual or not. Itris also the recurrence
1 Singular cbim bukt. This term has the dual significance o f * doctor or
leech * in so far as the ritual is regarded as therapeutic, and adept * (perhaps
even 4priest * for a senior adept) in so far as it is a national cult.

The P olitically Integrative Function o f R itu a l


throughout the ritual system, o f sym bols represeating.theJbistorical

origin jo f the N dem bu in M w antiyanvw as em pireandLofsym bois
representing grontocratie authority in general and o f m atriliny
in general rather than specific political positio n so fiau th o rity or
specific matrilineages.
T ke.u n ity _o_f_aUNdembu is. n ot maintained, by. pohtical control
fro m a strong centre o f authority. N o r is their polity^ o n eco n sisting o f hom ologous unilineal descent , groups, formed., b y a
process o f fission into structurally eq u ivalen tan d juxtaposed
segments. M oreover, there is no considerable overlapping o f
territorial and kinship groupings such as one finds in segm entary
societies. In such societies force tends to be distributed according
to the like but com petitive interests o f hom ologous segments.
In both the pyram idal and segm entary types o f society, ritual
relatiom .tend .both to .express and toJbuttressthe_major political
relations,.and ritual sym bolism tends, to express.parricular political
offices .or to stand, fo r the unity o f specific corporate g roups o f kin.
In states and .segm entary societies,..pohtical,. kinship, ..territoriaLand ritual relations_tend to some
In these respects, both types, differ markedly..&QmJhs&idfeinhin.
A m ong Ndem bu, relations established b y ritual .jc.ut_.sharply
across kinship and territorial ties, and. even across tribal .affiliation,..
smce_members^_of_other Lunda.graups, ...the. Kosa o f
Musok^htanda and.the N dem bu oXlshinde (and .evsnXvfena w ho
practise, analogous rituals), can participate in a n y cult into which
they have been initiated. B y establishing ties o f co-participation
in cults w h ich operate independently o f kinship and locallinkages,
the rituaLsystem compensatesjto.some extent for the lim ited range
o f effective pohtical. controLand fo r the instability, o f kinship and
affinal, ties to w hich, political. value.is_attached. It w ould trem
almost t a be the consistent id io m o f N dem bu social structure that
certaim basic sets o f ties w h ich tendto. be interwoven_so_M to
reinforce one another in m any .small~scale societieSy.shauld here
be separated from one another ; further,_thexohesiom othe total
society tin som e measure depends, not on the. juxtaposition and
correlative operation o f the ties, but on heir disjunction, and
even o n . their opposition. Thus m aternal descent. .groups, have
1 Fortes and. Evans-Pritchard, * Introduction * to African Political Systems,
pp. 16 et seq.











Schism and Continuity in an African Society

no local centres^juid Yicmages^are..made. up.of.yfflages^m o f

w hich are unrelated b y .lineal descent.^. .Descent..and..pQSt-marital
residence are not congruent... ..< throw
territorial,, kinship. and..pohtical.afiiliations..o.ut--of-alignment. A
fewdfixed. points exist in the. social, structureto-pro.vide.a-jneasure
o f stability w ithin the
the politico-ritual roles o f the chie and,ocertain_seniQr.headmen
in tnis_ cO nnection.B ut, Jby and. large, .the. function o f maintain
ing the unity o f the w idest social imit, the M dembu people,
dexolves.maiiily.upQn. the ^rituai .system,. although.Jthe ritual
system spreads evenwider- than this.
Mdemhu .haye. tw o major categories p f public rituals. These
are Jifencrisis. rituals, and. cults o f affliction.
Life=crisis rituals are perform ed irL.the^first place to signalize
im portant points in... the social or. b iologicald evelop n ien t o f
individuals (whether.these.coincide or not), asthese developments
are^ interpreted b y N dem bu culture__In.-the second^place, life
crisis rituals handle disturbances in the^social structure set up by
the.._change in social status o f the. principal .subject,jor_subjects o f
the,.: ritual. In the third., place, ffley.provide..occasions for a
demonstratiorL-of the..unity-oall Ndemb.u In..the.ibuEth place,
they.frequently act to reaffirm .ties.ofJ.ocal. propinquity between
villages in a given vicinage or . som etim es-between adjacent
vicinages. The three most im portant lifercrisis-iituals are the
boys circum cision ritual (Mukanda) , _the_ girls!_pixbertry ritual
(Nkan^-d), andthe., funeral ritual.
A, detailed, and in .this book , I. shffl. ..consider.. .only.. cultsL_o.affliction.
W hat Idiave called cults o f affliction are.perfbrmed.for.indiyiduals,
w ho are said -by- Ndembu-to. -have- been --caughxJ-fku^kwata) by
the spiritf o f deceased .relatives. whom-.they_haKe^-fbrgotten to
hononr_with .smaU.gifK.Qf..cxops.andB.eer,ox^who.m-fbey_have
offended, b y onrittingto..mentiQn,,their..namesL,.wheruprayers are
made^at the yM^&^yiyQmhu^(sing^rtiuyornhu)-4see-p. -X73) treeshrines. People m ay also be * caught .for. quarrellingcwith close
kin-or as representatives o f kin groups tom.-by-.conflicts.. Being
* caught means to be afflicted w ith bad-luck _at_h.unting in the
case o f .men, w ith reproductive disorders in .th e x a se o fw o m e n ,
an<LwithJllnessJnJboth cases.

The P olitically Integrative Function o f R itu al


which, afflicts,..and its. m ad^afeafflictian._T h esp iiit ..that afflicts
is a Jm'own ..and. named Jieceascd_i^attx:e.j>jdie_aiflicted person
or p atien t (muyej.i)__XKe.p^^^^|d3i:e
for.initiation into, the cu lt;... the., doctor- is. im adep.Lxti that
cult^^.-^hfcjoJe..<^^irhoreers,.to_ceitain.characteidsdcs_of this
spirit, .whidht are.conrelate.d.wifo.Q^^
sufferers m isfortune-.or illness. Thus a spirit w h ich afflicts a
wom an in the m ode o f Nkula (or as N dem bu pu t it, w hich
* comes out in N kula \ as in the phrase mukishi warni wedikilili
muNkula, * m y spirit em erged in Nkula ), 4 sits in her back (a
euphemism fo r her reproductive organs), and afflicts her w ith
menstrual disorders. These disorders include m enorrhagia, dysmenorrhoea, and amenorrhoea. T he w om an m ay dream that
her dead relative has appeared to her equipped like a hunter,
wearing the red w ing-feather o f a lourie in her hair, as only those
w ho have shed blood, such as hunters, manslayers and circum cisers, are entitled to do. Sim ilarly the m ain sym bols o f the
ritual fall into the class o f red objects : red d a y , the bark scrapings
o f a tree that secretes red gum , the blood and entrails o f a red
cock, lourie feathers, and the like. T he nam e o f the m ode o f
affliction in this ritual is called Nkula. The nam e o fjth e ritual
perform ed to treat the afflicted person is also railed Nkula. N o w
the carriedjouthy^a.nunffl.exLaf^doctors_QiL^adepts *
(ayimbuki or ayimbanda), both m ale and fem aleT w h o have them selves been closely associated w ith .previous perbrmances_oLthe
Nkula ritual. T h e w om en doctors, m ust at one .time^havejbeen
patients..theimselves and m ay not become, practising- doctors unless
they are .generally considered to h ave been.cmed_byjhej:reatirient.
The mate-doctors. m ust have ^
ckaka-JzhaNkula,. that is. to. say , m ale .ritual. asshtant..oimsi..Nkula
patient, a rolew hich may. bep.eife:medh^hechmb.aad,Jbjrother
or son, aiid.iixvolves foe.perform anceofafew ^taaLtasks,~ such
as guiding the patient out backwards fro m jh e spirit-rhut .to the
ritual .fire. T h e doctors or adepts are arranged in a loose hier
archy o f ritual prestige and each perform s a different ritual task.
Betw een doctors_.and patients^there. m ties-ofokinship
or affinity at all iCties. interlink them , ,w hich in secular situations
w ould -be associated w ith m utual avoidance, -these_.ties are
reckoned inoperative in the. context o f the ritual,_so_.that a man,


Schism and C ontinuity in an A frican Society

in his role o f doctor, m ay splash w ith medicine the almost

naked body o f a wom an w hom he calls 4 m other-in-law * or
daughter-in-law and w hom he norm ally avoids. I_a. patient
has been treated successfully,. sheJs.jentidcd.^to_take^paxt-iix future
performances o f the N ku la ritual.., as,.j^}w uot-.6octCH ^(chim bttki
wanyanya, * Hide doctor ). H er ritual status becomes higher
the m ore she participates in rituals o f this type, and the m ore she
learns about the medicines and techniques. T o whatever part
o f the N dem bu region she m ay m ove afterwards she w ill still
be entitled to participate as a chim huki (adept) in N kula . The
spirit w hich had once afflicted her is n ow thought to give her
healing pow er {ngovu yakuuka ) as a doctor, as w ell as to make
her forme. It is her spirit * (m ukishi windi). ThisLjspirit is
thought ,to have heen Nfe/a patieiit aii(fjdoctox, during
its lifetim e.
T he ritual itself has exoteric and esoteric phases. D uring the
exoteric phases, w hich typically take place in the village clearing,
a big secular dance often accompanies the curative process
proper, and as m any people as are able to leave their other
tasks in the vicinage, and, indeed, in villages in other vicinages
w ho have received tidings o f the ritual, com e to dance, sing,
drink, gossip and lo o k on. T h e headman o f the village and the
close kin o f the patient bu y and brew as m any calabashes o f beer
as they can to satisfy the clamorous demands o f the throng o f
dancers. I f the secular dance is a success, the prestige o f the
headman w ill tem porarily be high, but i f there is insufficient or
unfairly distributed beer and food for im portant attenders, it
w ill sharply decline for a tim e, whether the patient recovers or
not. The esoteric phases o f the ritual are attended b y doctors
and adepts only. T h ey take place either in the secrecy o f the
patients hut, or in the bush, w here medicines {yitumbu) are
collected w ith song and sym bolic action, and w here towards
the end o f several phases o f ritual, a chicken o r goat is sacrificed.
In the past doctors drove the uninitiated aw ay, b y threatening
them w ith bows and arrows, from areas in the bush w hich they
had sacralized for ritual purposes.
Bach culturally-defined mode^Taiflii^ojn^hy_a^spirit is associated-with a cult, whose adepts coflaborate-to-placate-the offended
spirit. Since the m ost com m on canse.,o,afflictk>nis_the neglect
in g or forgetting * (ku^vulamena) ,of.-the-spirit^the_m ost im -

The Politically Integrative Function o f R itu al


p cu tan tasp ect o f the_ process afip lacation i& .to back
to m em ory, * to m ake it kn ow n before m any peopleban d * to
m en ricm its name in th ek hearing \ in 'so c io lo g ic a l terms,
thismieans that in. a .m obile .and fissile society. th.ere_is.. a., strong
tendency towards structural am nesia,^countered b y
ritualslw M cli^ ontm ually. reviv.e.-tha.memos^-jo-deacL-persons
through whom , the Uving ^e_significandy._jmerconnected.
W hen the senior doctor addresses the .afBictmg^spint. at various
points in a ritual he mentions a num ber-of-significant kinship
connectionsof.thedeceased.. -Sometimes heLgives_a b rief outline
o f the kind. o f beha\aaur w hich. clivinatioix.has .suggested as the
cause ,o f the spirits anger__(Thiso.utline, incidentally, gives the
mthre^pologist-Jvaluable.-mfbrmation_about. the.Jkind&_of,conflict
diatL often arise in d ie. kinship system A .. In_the..context, o f the
ritual, situation, dus ritm g-ofLgenealagical. ties m ddiscussion o f
contem porary conflicts, has .the dual_effect-on-l>idembjU..of reviv
ing kinship a n d a ffin a l lhdcages.-which-.are_ be.gmuiug .to w ear
thin, and .o f invoking the participation.-of tbe wider.-S.oeiety,
represented b y the adepts draw n fro m . all .over the.. .N dem bu
region, in the conflicts o f the. narrower-unit^of-the village. B y
means, o f ritual,, the. im m ediate past and the__poHticorrkinship
relations Originating in it are kept .alive,-in. .spite, .o f the social
forces w o rk in g .in_the _app.Qsite__airection.
The other cults have die same general structure and functions
as Nkuia.1 In each, there is a band o f doctors and adepts, w h o
share out am ong themselves the various ritual tasks. In each,
there is a patient o r subject o f ritual, w h o is at the same tim e a
candidate fo r initiation into the cult. In each, a named, deceased
relative o f the patient is divined as the cause o f the affliction.
This spirit afflicts in a specific m ode. T h e mode is com m on to all
members o f the society : the spirit belongs to particular persons
and narrow kinship groups. T he name o f the m ode is the
name o f the ritual. T he same spirit m ay afflict the same livin g
relative in several modes, or m ay attack several relatives in the
same m ode. For exam ple, N yaluw em a o f M ukanza V illage
was afflicted b y her m other's m others spirit w ith reproductive
1 For a preliminary classification o f Ndembu rituals see my * Lunda Rites
and Ceremonies
O ccasional Papers o f the R hodes-Livingstone M useum , New
Series, No. 10 (1953) A detailed, analysis o f Ndembu ritual is in preparation.


Schism and C ontinuity in an African Society

troubles in the mode o f Wubwangu, and w ith illness in the mode

o f Chihamba. She is now an adept in both cults, and as w e shall
see an im portant doctor o f Chihamba. O n the other hand,
N yam ukanga o f m alabu lineage in M ukanza afflicted her
daughters daughter N yam asunga, her daughters daughters
daughter N yam ukola, and her daughters daughters daughters
daughters Zuliyana and Yana, w ith illness in the m ode o f
Chihamba. T h e spirit is m ost frequently a m other or a mothers
m other in w om ens cults. Thus the cults act as a m nem onic o f
direct matrilineal kinship links. V irilocal marriage, w hich may
scatter wom en far from their m atrilineal kin, is to some extent
counteracted by cults that stress m atrilineal ties between dead
and living wom en. B ut sometimes a wom an m ay be afflicted
b y a spirit from a different village sub-lineage from her ow n.
N yatioli, N yam ukolas sister, for exam ple, o f malabu lineage,
had tw ice been afflicted b y N yachintanga, apical ancestress o f
the other village lineage, in the modes o f Isoma and Wubwangu.
Thus the cause o f affliction is thought to be connected w ith
a breach o f kinship norms in some specific kinship group. But
the treatment o f the afflicted person is carried out b y an association
o f ritual specialists w ho need not necessarily be related to him
or her.
Sinee-most.. adult. members.o,any_.village-.are,.adep.ts,in_.atJeast
one cult, and. since .members., o f ail...c.ults_miay_.he_.found_evjs;qrwhere in the N dem bu and Kosa regions, it YoHows .that the-total
ritual system provides a set o f. interconnectionswhich~in effect
perform a . pobticaLfonction. This w ill be clear i f I present a
highly schematized picture o f the w ay ritual interlinks persons
and groups in a vicinage, w ith reference to cults concerned with
female reproduction only. There are four main cults w hich are
today perform ed to placate spirits interfering w ith female repro-i
duction : their names are Mkula, Chihamba, Wubwang*u and
Isoma.1 Let us for convenience call these cults A , B , C , and D .
Let us suppose that there are four villages in the vicinage, each
containing eight adult w om en belonging to the village descent
group. N o w let us suppose that in each village tw o wom en
belong to each cult association. This means that the wom en
1 There is another cult Kalem ba which used to be frequently performed
but is now extremely rare.

T he P olitically Integrative Function o f R itu a l


belonging to a single kinship group in each village are divided

by membership in four a its , and that four cu lt associations
interlink the four villages w hich, let us suppose, belong to four
unrelated descent groups. N o w let us suppose that in each
village there are eight w ives o f male village members, each w ife
from a diffrent village in a different vicinage. I f tw o w ives
in each village belong to each cult association, each pair o f w ives
is linked to tw o fem ale village members b y cult affiliation, and
to fem ale members and w ives o f m ale members in all the other
villages in the vicinage. T h ey themselves, since they originate
in other villages and vicinages, act in effect as representatives o f
the N dem bu as a whole in the context o f the ritual. N o w let us
suppose that in each village one m an is a doctor or adept in
each cult. In his ow n village he is linked through cult member
ship w ith tw o kinswom en and tw o fem ale affines, none o f w hom ,
perhaps, is his ow n w ife. H e is cross-linked w ith one man
and four w om en in each o f the other villages in his vicinage,
and divided from all the other m en, and from three-quarters
o f the w om en, in his ow n village, b y membership o f a cultassociation. Siblings and parents and children m ay be divided
from one another b y cult affiliation, w hile the status o f members
o f adjacent genealogical generations m ay be reversed during the
period o f perform ance o f a ritual. W h en ritual is being per
form ed in connection w ith cult A , members o f cults B , C , and
D are excluded1 from the esoteric phases o f the ritual, although
senior doctors o f these cults m ay be given sacralized beer or
food in token o f their high ritual status in other situations.
Sim ilarly w ith the other cults. Thus in a society characterized
by the poverty o f its status system in secular social relations, the
status differentiation w ithin each cult and the exclusion o f unin
itiated persons, w hatever their rank or standing in other contexts,
from its central mysteries, provides some compensation for the
frustrations o f ambitious urges, or for the occupation o f an
inferior status in secular life.
I have o n ly m entioned one section o f the ritual system, in
dealing w ith cults relating to fem ale reproduction. In addition,
1 Gluckman has shown in * The Role o f the Sexes in Wiko Circumcision
Ceremonies \ that exclusion from the arcana o f a ritual is sociologically as
significant as participation in them.



Schism and C ontinuity in an A frican Society

there are tw o distinct but interlocking categories o f cults relatm g to huntsmanship ( Wubinda and Wuyanga), one o f which
(Wubinda) contains five separate cults. Social bonds o f mem
bership in the hunters cult likewise interlink individuals in
different localities, and even cross tribal boundaries. Then there
are curative cults o f various kinds, a cult w hich concerns itself
w ith the initiation o f diviners, and a number o f anti-witchcraft
cults. T oday, new cults, often organized b y O vim bundu and
Lw ena im m igrants, are directed against disease and misfortune
caused b y the m alevolent influence o f living and dead Europeans,
and, in some areas, these are beginning to oust the traditional
cults. T o the m anifold ties interconnecting people in these cults
m ust be added the links created by the boys circum cision ritual,
w hich jo in m en in opposition to w om en,1 and by. the girls*
puberty ritual, w hich jo in w om en in opposition to men.
Furthermore, there is the M ungongi association w hich carries
out im portant duties during the funeral rituals o f its members
and w hich carries out an elaborate initiation ritual o f its ow n in
connection w ith the funeral rituals.
These_cidfclm is
frequently performed..^ I .do_not_know_j^hether_there has been
any increase in the frequency o f performance o f rituals in recent
years ; but I have personally witnessed thirty-one performances
o f no less than fifteen different categories o f cults o f affliction
and anti-w itchcraft cults, as w ell as m any life-crisis rituals, and
have been notified o f the performance, in localities w here I was
staying at the tim e, o f m ore than tw ice that num ber mostly
in m y second period in the field. It is likely that rituals were
perform ed at least as frequently in the past, for a num ber o f
rituals in the bow-hunters cult (Wubinda) have becom e extinct,
as w ell as ritual perform ed b y the w om ens Chiwila funerary
association, and a cult called Musolu, perform ed b y a ch ief or
senior headman, to bring rain i f there had been a drought at
the beginning o f the rainy season. M any cults in vo lve the per
form ance o f tw o successive rituals, separated b y a period during
w hich the patient undergoes partial seclusion from secular life.
T he first o f these rituals is in the w om ens cults called ku-lembeka
1 C f. G luckm an, M ., * T h e R o le o f the Sexes in W ik o Circumcision
Cerem onies *.


The P olitically Integrative Function o f R itu a l


or ilemhi, and the second, and m ore elaborate and im portant,

ku-tumbuka. A candidate w h o has undergone ku-lembeka m ay
play a m inor role in the cult, but cannot becom e an officiant
until she has passed ku-tumbuka. O ften a person, is considered
cured, that is, freed from the ban o f her fertility im posed b y the
spirit, after ku-lembeka, and m ay never undergo ku-tumbuka.
T h e 'fa ct that m ost cults in vo lve tw o separate perform ances o f
ritualJtnultipHes the num ber o f occasions o n w h ich each kind
o f ritual m ay be perform ed__Again,. rituaharesQme_times con
sidered tohavef^edim theirm irpose__fQ ra_num ber^Q fjte3soiis.
T he diiriner J w h o p xj^ q o b ed .jfe..^
deceived b y -a . w itch._O r he m ay have divined fo e.rig h t. spirit
but o f affliction,_O r severaLspirits_ m ay have
been afflicting .the .patient^simuitaneously, in,..diSm nt_m odes o f
affliction,.sodiatrituahm m tbe_successively_perform eclby.seyeral
cult associations before a cure is efiected_Thus^.personsuentitled
to participate .as adepts in cults, are comtmtly._being_mobilized, candidates fo r initiation, as._patients, .o r in .fo e guise o f
patimts,- are constantly., undergoing..treatm ent.
Som e w om en, such as Y ana (I,j4), daughter o f N yatioli
(I,H zi) In M ukanza V illage, have been initiated into all the
w om ens cults. Yana, w h o was o n ly thirty-four in 1954, was
a patient in the ku-lembeka phase o f Nkula w hen she was about
seventeen years old . Nkula in this case had been perform ed after
the m iscarriage o f her first child. H er husband h ad sought out
a diviner w h o had named her mothers m other N yam asunga
as the afflicting spirit. Yana considered the treatm ent successful,
for her next child was bom successfully. A fte r about ten years
she had menstrual troubles and her husband w ent w ith Sakazao
her uterine u n d e to consult a diviner. H e diagnosed the renewed
anger o f N yam asungV s spirit as the cause o f her trouble and
recommended that the ku-tumbuka phase o f Nkula should be
perform ed. A fter this, her m other N yatioli, w ho had been the
senior fem ale doctor in the ritual, taught her the fu ll * m ystery *
(mpang*u) o f Nkula, and its fu ll curative aspect (wumbuki). The
follow ing year she became ill during pregnancy, and divination
indicated that her m others m others m other had afflicted her
in the m ode o f Isoma. T h e Isoma manifestation o f a spirit can
cause abortions and miscarriage. T o prevent these she under
w ent the fu ll ritual, in this case called ku-tumbuka, o f Isoma,


Schism and C ontinuity in an A frican Society

perform ed b y Kanyom bu, M ukanza Kabindas brother, and at

that tim e her mother*s husband* There is no ku-lembeka at Isoma,
B ut w hen "she bore her child w ithout mishap, the kwidisha, or
causing to com e out * (i.e. from the grass seclusion hut), was
perform ed, to celebrate the success o f the treatment.
Yana also underwent as patient a perform ance o f IWubwangu,
k.U'-tumbuka on ly. Wubwangu is most frequently perform ed for
a w om an w ho has had twins, fo r her and their w elfare, or for a
wom an w h o is expected to bear twins because she herself or her
m other is a tw in. Yana passed through ku~tumbuka only, be
cause ku-lembeka is perform ed during pregnancy and not after
parturition. O nce the tw ins have been bom , ku-tumbuka is
enacted. Yana's twins both died and the treatm ent was reckoned
unsuccessful. Nevertheless, Yana claim ed that she w as entitled
to act as a ju n io r adept (<chimbuki) in Wubwangu and had in fact
attended tw o performances since her ow n, in both o f w hich7the
patients w are cured. She said that she had paid the senior wom an
doctor, w ife o f the headman o f her husband's village, w here the
ritual took place, ten shillings fo r full instruction in the collection
and preparation o f the herbal * medicines * (yitumbu) o f Wuhwangu. Y ana was also a candidate for initiation into the
Ckihamba ritual at M ukanza w hich I am about to describe. In
a relatively short period then, between the ages o f about seventeen
and thirty-four, Yana had undergone six performances o f ritual
in four w om en's cults. She was entitled to participate as an
officiant, w ith varying status in each, in four cults. In Nkula
she was quite an im portant offician t; in Isoma she played a
prom inent role ; in Wubwang u she was entitled to help in the
collection o f m edicine in the bush w ith other adepts but did not
prepare and apply them ; and in Ckihamba she had acquired the
right to enter the sacred enclosure from w hich the uninitiated
are excluded. H er m other N yatioli was an im portant officiant
in Isoma ; perform ed ritual tasks in Nkula ; was an adept, but
not an im portant one, in Wubwangu ; and had never been
initiated into Ckihamba. In Ckihamba, then, Yana rated higher
than her m other, about equally an Wubwangu, and low er in
Nkula and Isoma. Yana, how ever, had been initiated, but not
N yatioli, into the new fashionable Tukuka and Mxvana Ipana
rituals introduced from A ngola and the Belgian C o n go respec
tively. Yana displayed rather more than the usual interest in

T he P olitically Integrative Function o f R itu a l


ritual com m on in Ndem bu wom en but it is no exaggeration to

say that m ost N dem bu w om en are initiated into at least tw o cult
concerned w ith fem ale reproduction between puberty and the
D oes the frequency w ith w hich such ritual is perform ed
indicate a h igh rate o f reproductive disorders ? I obtained the
impression that it was indeed high, but not exceptionally high
for A frica. I w ould like to postulate that the h igh frequency
w ith w hich curative and gynaecological * ritual is perform ed
by N dem bu is ju st as m uch socially as biologically determined.
M isfortune, .illness and reproductive. troubles. am on g ,Ndembu,
i f severe enough, are associated ..with the activitieso f spirits,
witches and sorcerers. W h en these are causes
o f affliction, one does not have to lo o k very fa r .finds
confflcfc in social .relations. E very one o f the social dramas I. have
usecLto illustrate and carry forw ard-the analysis, contains .at least
one reerence i:cLa supematural agency, incon nection w ith a social
dispute._N dem busociety, e-have-show n,is-to m w ith

peeimi.aLdUspur_es>-therei s ^
between its dom inant p rin cip les-o f sociaL_organization, anci in
secularJife there is little to bind-together..m ore-than a small
number o f people in habitual co^operation. U n ity in . such a
societyisinaintained by m aking each-outstanding case, o f p er sonal
m isfortune the occasion on w hich the .m oral norms and _values
shared b y all N dem bu are prom inently displayed in a number o f
waysrrrin prayer, precept, sym bol, m im etic action, .and in the
rituahassociation o f those w h o have. suffered_regardlesso. their
kinship or other interlinkages, T he com m on m isfortunes o f
m ankind..are. m ade occasians_for .restating_the_.common. values
o f N dem bu.
A m on g the .Ndem bu,.as w e have seen, each little m obile local
unit_is_tom w ith continuous dissension. Nevertheless put, o f
cojnfflct its e lf u n ity is engendered._Each, severe case of illness
bringSrintQ^ the_ppen ^raiikling dispute w ith in a
group^-between-specifrc.persons w ith opposed
such disputes .are irresoluble sincejhey. spring^from cont
in the soeial structureitself. Since they are ultim ately irresoluble
the m ost that can be-done is to. brm g about.a.tem porary palliation,
to produce, as it w ere, the illusion -o f harmony^ within- the dis
turbed-group. This is. done b y stressing .in. each.little. localized


Schism and Continuity in an African Society

performance o f ..ritual the co m m o n in tere^ .o fth e whole, sodety^

and,, what.. Fortes., and Ey^-Eritcharid,.calJl^<LfhcL.basic. relations
that_make up the social order
such a& m atrijiny,.gerontocracy,
the historical unity o f the Ndernbu, ancLsp .on^ ,,-Agam, the ritual
is/ carried out b y an W JW assadation.of-cidfcim tiates,,.w ho re
present n ot this or that particular interest-group-but -Ndembu
society as a w hole. Such a groupJs jtao_ephemeralto develop
internal conflicts o f a serious kind.
T hus, each individuaTs m isfortune brings to ligh t some specific
and localized conflict- in inter-personal-on inter^group relations.
R itu a l is. then_inyoked,- under_ the-jpretext_of cu n n gjd ie patient
or rem ovinghism isfortune, to -settle. the-conjflict^ ^-But-the same
ritual has far wider. im portance in_that^it,afl&rmsL. the. com m on
interests o f all lsidembu, .often expressing -this-xom m unity o f
interests in . sym bols... representing pivotal-aspects... o a. culture
shared b y all, how ever distant.from .oneLmather jhey_m ay be in
space^-however opposed to.-one another in .political action.
I have suggested that there is a relationship between w hat was
in the past a marked emphasis on hunting and the institution o f
virilocal marriage ; ana also that the contradiction between
virilocal marriage and maternal descent in a context o f seminom adic hunting and cassava-cultivation gives rise to many
conflicts between sections and factions varyin gly affiliated in the
social system. JLhave argued., that ..therefore^
o f ^manifold and unceasing, struggles, of-schisiru and.,fission. I
believe further that the society exhibiting
could not maintain .a n yso rt ofcoh eren cefor_ l.on g.w e r e jt not
fo r its plastic and adaptable.system o f ritual._Tbis ritrial, .although
it operates in a contingent .and fitful .mamier-in .particular.-situations,-and lacks great national, seasonal onregular perfoxmances,
still contrives to. stim ulate in .its. members ..sentiments, -ofjcribal
unity, ^ o f a. ..general belonging to g eih e rjw h ich tr^sce|ici the
irreparable divisions and cpiiflicts^ofinterejs.tsin
Theiafflictionofeachis-theconcern.ofaU -;.Jhkenes.sof unhappy
lo t is the .ultimate bond. o f rituaL solidarity. The., adepts have
themselves know n the suffering, the candidates..are experiencing.
T he spirits w ho cause them to suffer themselves-suf&red in the
1 African Political System s p. 18.

The P olitically Integrative Function o f R itu al


same. w ay.. J3u t afHiction is potentially...a baoii>^.not_merely a

pre&ent'pain, fo r it is th e ra y a fro a d ta ritual em m encein a cult.
Affliction b y a . spirit follow s a breach o f regular custom ary
relations betw een livin g and^dead,_i.e_relations^ o f worship
invob^gjthejaei^tioningjojrfie. .names, o f,the dead in. prayer, o r
a breach between, livin g, and.Jiving,_vdh.emklmquarid,._which is
fflen^puhished b y the dead. "Thus the unity or. a ll Ndem bu is
only- perceived in situations.. arising_aut-ofothe-&tttf&_of specific
relations, usually couched in terms o f. .. kinship. A society
continually threatened w ith disintegration is continually per
form ing reintegrative ritual.. R itu a l among_Ndemhu_does n ot
express the kinship and political structure as in a firm ly organized
society ; rather it. compensates forffleird efIcien a labile
In order to bring out m ore fu lly the im plications o f this analysis
I propose to exam ine in som e detail the social background o f an
im portant ritual w h ich took place near the end o f m y second
tour o f fieldw ork at the village o f M ukanza.
* Chihamba9 Ritual at Mukanza Village
In this analysis I try to isolate social relationships and processes
from their cultural integum ent as for as possible. I hope to make
a detailed exam ination o f the cultural structure o f N dem bu
ritual in a separate study. B u t som e account o f the main cultural
features o f die Chihamba ritual is necessary, i f w e are to grasp
d early its sociological im plications.
T h e Chihamba ritual is said b y N dem bu to be a very im portant
one {chidyika chalema nankashi, * a very h eavy ritual *). Som e
claim that it is m ore im portant than the hoys circum cision ritual,
others deny this. B u t all agree that its ku~tumbuka phase is
usually attended b^r m any m ore people than com e to any other
ritual except boys* circum cision. M any calabashes o f beer are
brewed b y the village sponsoring it, and a large am ount in cash
and kind is paid to the ritual officiants afterwards.
Chihamba is a specific m anifestation o f an ancestor spirit. It
is an exceptionally dangerous one, fo r w hile most manifestations
o f spirits cause m isfortune, in fertility or disease only, a spirit that
has * com e out in Chihamba 9 can kill the person it afflicts.
T h e Chihamba manifestation o f a spirit afflicts its victim in
several ways. M ost com m only it causes pains in the w hole body,


Schism and Continuity in an A frican Society

especially in the neck, and a feeling o f extrem e cold. Sometimes

it is said to induce decay in crops planted by the object o f its
wrath. Again, a man w ho is caught * by Chihamba may
experience bad luck in hunting. It m ay also afflict wom en with
reproductive troubles. T he Chihamba manifestation is, in fact,
a sort o f compendium o f all the misfortunes that can happen to
a person. B ut Chihamba is not sim ply the manifestation o f an
angry ancestor. Another category o f supernatural being is also
involved. A t several points in the ritual a strange supernatural
being is represented in various ways. It speaks indecencies in a
throaty vbice and brusquely questions the candidates. This
being, called Kavula (a name know n in theory only to initiates),
is not the spirit o f a dead person, but has his ow n independent
existence. Kavula is an archaic term for the lightning, and is
perhaps etym ologically connected w ith nvula, * rain \ A t one
phase in the ritual he is said to enter the hut o f the principal patient,
as the Ughming does, through the grass * top-knot * {ntungu).
O n the evening o f the first day o f the ku-tumbuka phase, a
doctor impersonating Kavula, screened from the view o f adepts
and candidates, addresses them in the principal patients hut in
a throaty voice. A fter enquiring o f the candidates w h y they
have com e to visit him , and after reviling them, he confers on
them ritual names w hich o n ly Chihamba adepts are entitled to
use. N ext day a contraption is made sym bolizing Kavula in a
secret place called isoli w ithin a sacred enclosure in the bush about
a quarter-of-a-m ile from the village. The contraption consists
o f an inverted m eal-m ortar containing sym bolic objects and a
fram ework o f sticks to w hich rattles are attached. O ver this is
spread a blanket or skin whitened b y cassava meal. A concealed
string attached to the fram ew ork can be manipulated by a male
adept to make Kavula dance \ Just before nightfall the candi
dates, each in order o f ritual status, but wom en before men, are
brought to greet Kavula their * grandfather , the one who
know s everything and must be praised \ O ne by one they crawl
on their stomachs to the leaf-hedge surrounding the white image
o f Kavula, banging their heads on the ground as they advance,
first on one side and then on the other. W om en squeeze their
breasts * to give m ilk to the ch ief as they m ove forward. Kavula
is compared w ith M w antiyanvw a at this moment. W hen he
reaches Kavula, each candidate is told *to kill Kavula \ by striking

The P olitically Integrative Function o f R itu al


him on the * head * (really the inverted m eal m ortar) w ith a

Chikamha rattle-each candidate has such a rattle, o f special form
elaborately carved for him or her before the perform ance. The
candidates are then led aw ay b y adepts and brought bacfc to find
the blanket gone and the m ortar running w ith the blood o f a
sacrificed chicken. T h ey are told b y the adepts that they have
killed * the c h ie f (mwantci) and that they have seen Kavula*s
T he w hole o f that day they have been driven, w earing only a
waist-cloth, between the sacred enclosure and the principal
candidates hut b y the adepts. T h ey have been called the
slaves * (andungu) o f Kavula and have been forced to wear
sym bolic slave-yokes (mpanda). Nkaka means * ow ner (o f
slaves and dom estic animals) as w ell as * grandfather *, It seems
reasonable to infer that one o f the m any meanings o f Kavula is
gerontocratic authority in general. Kavula jokes (ku-senseka)
w ith adepts and candidates, as a grandfather does w ith his
grandchildren. H e reviles them sexually in a thick, m irthrevoking, yet nevertheless, rather terrifying, voice. Some o f
is esoteric names, Samasenga for exam ple, said b y informants
to be derived from ku-senguka, 4 to m ultiply (o f fruit)* suggest
that he presides over fertility and that his authority has its benevo
lent side. His associations w ith the rains, w ith cassava, and
w ith other cultivated crops, support this view . H e can rem ove
sickness and m isfortune and bestow health and fertility after his
m ystical death. B u t it is diear that there is am bivalence in the
relationship between Kavula and candidates. H e it is w ho
strikes like lightning w ith disease, w h o treats the candidates like
slaves, w h o forces them to grovel before him , w h o taunts them
w ith teasing questions and reviles them , and w h o finally is
bloodily killed b y them . This am bivalence is n ot surprising in
view o f the strains and conflicts in the kinship system, and
between slaves and their masters.
Kavula is said to be a male. T he spirit o f the afflicting ancestor,
on the other hand, is always fem ale, and its name is found by
divination. O ne inform ant told m e that * Kavula is the husband
and X (the aifiicting spirits nam e) is the w ife. T h ey jo in
together to catch the patient.* In the Chikamha ritual there is
norm ally a principal patient or candidate (the term muyeji has,
o f course, both meanings), always a wom an, w ho has been ill or


Schism and Continuity in an A frican Society

unlucky in the ways mentioned above. A divination has been

m ade for her and the diviner has said that such-and-such a spirit
has com e out to her in Chihamba \ O ther sick w om en in the
village m ay then also be thought to have been afflicted b y the
same spirit and they are also nom inated as candidates. O n the
m ost im portant day o f the ritual, that on w hich the candidates
are chased back and forth like slaves from the sacred enclosure
to the hut o f the principal candidate, the pursuing doctors may
capture any persons o f any age o r sex w ho happen to cross their
path and add them to the group o f candidates. People often
get caught deliberately in order to be initiated, in spite o f the
hum iliations to w hich they are subjected. Parents urge their
young children to get caught, for this renders them im m une from
capture at subsequent performances o f Chihamba and gives them
the right to enter the sacred enclosure, i f not to approach closely
to the central shrine. The chasing goes on all day until sunset.
Each tim e the candidates are driven back into the sacred enclosure
from the village, they are made to sit dow n w ith their backs to
the central shrine in a long row , w hile adepts crouch beside them
on either side asking each in turn a num ber o f riddles all concerned
w ith the identity o f Kavula. M ost candidates m ake ridiculous
mistakes in their answers at w hich the adepts laugh derisively.
Each tim e they return they advance a little nearer to the
central shrine where the sym bolic representation o f Kavula is
being m ade. T hey are told that they w ill see * som ething terri
fyin g*. U ntil the final obsequious greeting o f Kavula no
candidate ever mentions his name. There is a fiction that only
the initiated kn ow it, although I have obtained evidence that
uninitiated people have heard o f it. Eventually th ey are told by
adepts w h o are their friends to m ention the name o f Kavula when
asked b y the senior doctors, w ho are always male. T he wom en
claim that Chihamba was once their ritual but that the men * stole
it from them *. M ale doctors and adepts concern themselves
principally w ith the m ore esoteric aspect o f the ritual such as the
construction o f Kavula's representation, w hile the w om en adepts
act as a sort o f police force guarding the sacred enclosure from
desecration b y the uninitiated, chasing the candidates and seizing
those w h o cross the path between village and sacred enclosure.
Thus m en are the m ore important adepts, w hile w om en are the
principal candidates.



I n t e r r o g a t io n

of C h ih a m b a

C a n d id a t e s

The candidates sit modestly in a line with their backs to the small sacred
enclosure (isoli) where Kavttla (see Chapter Ten) is represented. Nyamukola
is the candidate nearest to the camera. Manyosa stands on the right, arms
akimbo, holding her adepts rattle while other adepts ask the patients, riddles.
The child-candidates can be seen top-left, harangued by Headman Itota, who
enacted Kavttla the previous night. Nyamuwanga, the suspected witch o f
Social Dramas IV and VII, is the bare-headed female adept near the centre o f
the photograph.

T he P olitically Integrative Function o f R itu al


Kavula is impersonated in the patients h u t behind b y a screen

b y the m ost senior m ale adept or an experienced adept w ith a
suitable vo ice fo r reviling in the style required. Another senior
adept makes the blanket-covered fram ew ork that represents
Kavula in the bush. O ther m ale adepts take turns to manipulate
the string that makes the structure shake and rattle. N o w om an
m ay pass behind the screen o f leaves w here Kavula thus * dances *,
but the leading fem ale adept m ay sit d ose to the front o f it.
Both, m en and w om en adepts m ay chase the candidates, although
m ore w om en than men do this. B oth m en and w om en m ay
question the candidates, although men ask the m ost esoteric
A fter Chihamba is over, a special form o f friendship similar to
blood-brotherhood and blood-sisterhood (wulunda) is made
between adepts and those w h o have been candidates in the ritual,
called tmbwambu or wulunda waChihamba. This umbwambu pro
vides yet another means o f interlinking members o f different
villages, fo r members o f the same villages are seldom united in
this w ay. T h e adept w ith w hom a candidate is join ed in ritual
friendship is the one w ho has helped him or her to answer
correctly the final riddles put b y the senior m ale adepts, before
the killin g o f Kavula.
This necessarily abbreviated account o f the m ain features o f
the hu-iumbuka phase o f Chihamba m ust suffice as an introduction
to the sociological analysis o f a specific perform ance o f this
ritual. It w ill be remem bered (see Chapter fiv e , p. 164) that
the village o f M ukanza was rapidly approaching the crisis o f
fission in 1954, near the end o f m y second tour o f fieldw ork. It
seemed lik e ly that the village w ould undergo a three-w ay split,
w ith Sandom bu revivin g the nam e o f Kahali V illage for bis farm
and obtaining bis follow in g from Bibiana and her children ;
Kasonda d ividing o ff w ith M anyosa and C haw utongi and
founding a fa rm ; w hile Sakazao, w ith o r w ith out Chinem a,
and supported b y m ost o f m alabu lineage, w ould succeed to the
name and position o f M ukanza. It w ill also be recalled that the
relationship under m ost strain in this situation was die marriage
between M ukanza Kabinda and N yam ukola w hich interlinked
tw o village lineages and also tw o genealogical generations.
M ukanza and N yam ukola w ere a m ost devoted couple, w ere
the parents o f m any children, and w ould never divorce one









Schism and C ontinuity in an A frican Society

another. H ostility between village lineages and generations did

n ot express itself in marital quarrels between them but in criticism
directed by other villagers against N yam ukola. M ukanza him
self was rarely attacked ; in some w ays he seemed to be above
the battle, as a personification o f the values o f village unity and
persistence. B u t N yam ukola became involved in several
quarrels, some o f them quite sharp. M ost o f her disputes were
w ith members o f the intervening genealogical generation. Social
Dram a V (pp. 157-61) includes a quarrel between Sandombu
and N yam ukola. She also had a dispute w ith Chaw utongi w hich
led to the latters tem porary defection from M ukanza Village.
N yam ukola had com plained publicly that w hen she had been
to Kalene Hospital for treatm ent o f her leprosy, Chaw utongis
sons, Kasenzi (!,H 5), w ho w orked fo r (he European headmaster
o f Sakeji School, w ith K aseloki (I,H6), w ho was a schoolboy at
Kalene M iddle School, had o n ly given her a small piece o f beef
between them. Chaw utongi, w ho had recently visited Kalene
Hospital, herself accused N yam ukola o f lyin g, and claim ed that
her sons had given N yam ukola a lo t o f meat. As (he quarrel
progressed m utual recriminations became m ore and m ore savage.
Ghaw utongi accused N yam ukola o f w itchcraft, claim ing that
the latter had developed a grudge (chitela), the activating principle
in witchcraft, against Kasenzi, because he had once, on receiving
the present o f a goat from his father- and m other-in-law after the
first pregnancy o f his w ife, charged N yam ukola and M ukanza
some m oney fo r a portion o f its meat instead o f givin g it as an
obligation o f kinship. A fter saying this she demanded com
pensation from M ukanza for the slander his w ife had uttered