-
International Library of AnthropologyRoutledge & Kegan
Paul
Editor: Adam Kuper, University of Leiden
/\rhm Sdcntiae/\rhln Vilac
/\ cataloguc uf IIlhcr Sud'll Scicnce buoks published
byRoullcugc & Kcgan Paul will bc fuund at the end ur Ihis
volumc.
-
Process and form in social lifeSelected essays of Fredrik
Barth:
Volume I
Routledge & Kegall PaulLondon, BosI"n and HCllley
-
Process and fonn in social life
-
First published in 1981by RUUl/edge & Kegan Paul Ltd39 Store
Street,Lundun WCIE 7DD.9 Park Street,Buston, Mass. 02108. USA
andBroadway House,Newtuwn Road,Henley-on-Thames,OxonRG9IENSet in
lOon 12 ptIBM Press Roman byAcademic Typing Service. Ge"ards Cross,
Bucksand printed in Great Britain byBilling & Suns
LimitedGUi/dford, London, Oxford and Worcester Fredrik Barth 1981No
part uf this book may be reproduced inany form without permission
from thepublisher, except for the quotation ofbriefpassages in
criticism
British Library Cataloguing in Publication Data
Barth, FredrikScJected essays uf Fredrik Barth. -
(Internationallibrary ufanthrupulogy). Vol. I: Prucess and fonnin
social lifeI. t:thnulogyI. Series306 GN325 80-41283
ISBN 9710007205
-
Contents
Introduction II Anthropological models and social reality 142
Models of social organization I
Introduction 32The analytical importance of transaction 33
3 Models of social organization IIProcesses of integratiun in
cllioire 48
4 Models of social organization IIIThe prublem ofcomparison
61
5 'Models' reconsidered 766 On the study of social change 1057
Analytical dimensions in the comparison of
social organizations I 198 Descent and marriage reconsidered
1389 Economic spheres in Darfu r 157
10 Competition and symbiosis in orth East Baluchistan 17911 A
general perspective on nomad-sedentary relations in
the Middle East 18712 Ethnic groups and boundaries 198Notes
228Bibliography 232Index 239
-
Introduction
These essays were written during the period 1955-72 (,'xeept
furchapter 5. Volume I and chapter 7. Volume II. which were
especiallyprepared for this collection). as steps in my own
development as ananthropologist. They are panly my response to thc
thoug.hts of otheranthropologists: but mainly they arc my response
- aided and en-cumbered by such thoughts - to various glimpscs inlO
the realities ofother people's lives and social relations that I
have obtained throughfieldwork. As I understand my own work, in
other words. my in tdlce-lUal biography would focus mainly on the
various fieldwork I havedone, and not on the books I have read or
the SdlllOls where I havestudied and taught. The major results of
such fieldwork have heenpresented in a succession of monographs
(see Bibliography). whkhcontain my main anthropological
contributions. both substantive andanalytical. In the lectures and
anicles republished here. I sought 10extract some of the more
general positions and understanding.s atwhich I had arrived,
prescnting them together with smaller fragmentsof data which I
judged particularly enlightening for my analyticalargument.
Argument, polemical or otherwise, is central to moSI of the
essays;and they inevitably address issues as these appeared to
me:1I a particular time. One might tJlink that they therefore echo
old battles longforgotten. If they were to be written loday, they
would have beendesigned somewhat differently in responsc to
contemporary debatesand arenas. But I regard their essential thrust
and content as stillcurrent and vatid, since I find thai the
viewpoints Ihey contain areproductive for my continuing work and
have only been incompletelyutilized and accepted by otJlers, and
that the views I challenge are suchas persist or reappear in only
slightly varying fonns in the thinking ofanthropological
colleagues. For these reasons, though the Ihemes andissues that
engage me today are partly different (see Bibliography), itis also
true to say that I 'stand for' views as argued in these texts.
-
2 /nrrodUClion
TIle essays may be usefully juxtaposed, since they are related
asdevelopments of a few fundamen tal themes and so supplement
oneanother. But I do not imagine that they add up to a complete
andunified theory of culture and/or society. On the contrary, they
wereintended to enter into the wider corpus of social anthropology
toinnuence and modify its development on a few fronts and therebyto
contribute 10 the collective endeavour that is our discipline,
ratherthan erect a personal memorial or a sectarian Grand Theory.
In otherwords, although my work has sometimes been interpreted this
way, Ihave never wished to join or delineate any particular '-ism'
in anthropo-logical analysis.
The anthropology which J met in the 1950s was, however, in
myjudgment seriously crippled by an inattention to fundamental
aspectsof people's lives. Most particularly, I felt the need to
acknowledge theplace of the individual, and the discongruity
between varying interestsand various levels of collectivity. Much
of my writing in the ensuingyears has consequently focused on the
task of developing a perspectiveon the subjective and goal-pursuing
actor. This has entailed taking upthe questions of what place
considerations of value and utility have incanalizing the behaviour
of persons; the variation exhibited in be-haviour and the factors
generating this variation; and what it is thatpropels and
constrains individual actors and thus shapes their behaviourand
Iheir lives. I believe that a conceptual apparatus which can
treatthese topics and integrale Ihem wilh other major
anthropologicalconcepts is an indispensable component in any
theoretical systemfor social anthropology. All the essays collected
here relate, in oneway or another, 10 this theme.
There are several alternative vocabularies with which to discuss
thecircumstances and consequences of these aspects of
intentionality, goalorientation, and rationality. I do not think
that social anthropologyshould adopt any particular one of them and
thereby the custom-madeepistemology which it enlails. We who study
the different epistemo-logies and differing praxis of different
cultural traditions should ratherengage aclively to design our own
procedures and positions withrespect to such fundamental questions.
To do so we need to be awareof the formulations and debates that
are current in other disciplines,but not fight shy of assuming the
arrogance needed to reshape themaccording to our own needs. That we
have so far done this only littleand weakly is no reason nol to
embrace it as a programme.
My own position on these central issues of epistemology and
theory
-
Introduction 3
is only occasionally taken up in an explicit way in these
essays, butis given a brief presentation in chapter I, I hold that
we must a.:know-ledge that most of the phenomena we study are
shaped by human .:on-sciousness and purpose. Since social acts are
tllus not simply 'caused'but 'intended', we must consider these
intentions and understandingsof actors if we wish to capture the
esscntial contexts of acts. I seelittle possibility, and no
desirability, in defining our object of study soas to eliminate by
exclusion this subjectivity of human actors. WhereI seem 10 part
ways with most of the transcendental philosophies, andwith
anthropological colleagues who tend to follow them, is in
Clll-phasizing the need to understand behaviour Sil1lll/ranevlIs!v
in two,differently constituted, contexts. O!!e is the semiotic onc,
wherestrings of events are shaped by actors so as to embody
meanings andtransmit messages and thus rel1ect the rules and
constraints of codifica-tion. But the same events also enter into
the material world of causesand effects, both because acts have
consequences and because personsmust relate to others who also
cause things to happen. This latter con-text forces actors 10
consider the instrumentality of acts, in ways whichrel1ect both the
constraints of knowledge and value, and the pragmaticsof
cooperation and competition. Fi!lilly, I see a dialectic -
albeitbetween entities on conceptually distinct macro- and
micro-Ievels-between these codes values, and knowlcdge on tile one
hand andhuman acts on the other. Not only do the former provide
premisesIand constraints for particular acts, but acts also affect
codes, values
. and knowledge by increments and so can change and modify
theirown preconditions.
Within this comprehensive perspective, most of my discussion
Inthe period covered by these essays has focused on actors'
stratcgiesof instrumentality and the aggregate social consequences
of suchstrategies. In quite another connection, Bateson (1972:
490-3) hascharacterized as a fundamental epistemologica1 fallacy
the Westerntendency to th.ink and act in terms of the wrong units:
in terms ofseparate individua1s rather than systems of interacting
persons andobjects, or in terms of human groups and populations
rather than thesein interaction with their environment. Systems
theory has taught us(recently) instead to follow the pathways along
which information andeffects now, and see these whole circuits or
loops as the unit - notthe severed chunks which we produce by
cutting across these conncc-tions of interrelationship. I contend,
however. that this fallacy is notlimited to Western thought, but is
prevalent in many contexts in many
-
4 lnrroducrion
other cultures as well. In such cases the logic by which actors
operate -predicated by their subjective understanding of the
situation andtheir own interests in it - is often at odds with the
aggregate system ofinterdependencies within which they are acting.
In societies with lesspowerful technologies this is not as
ecologically suicidal as it may bein our case, but it is certainly
pervasively consequen tial in shapingsociety and culture. The focus
in the following essays on the strategiesof actors provides a
methodology for exploring the consequences ofsuch egocentric
epistemologies on the part of actors, inherent as I seeit in the
inescapable directed ness of consciousness towards
'something'distinguished from the subject who is exercising the
consciousness(Slagslad, 1976; Skjervheim, 1959).
My perspective, and my production over the years, have
beenshaped by the effort to introduce these considerations in
opposition,as an alternative or as a supplement, to conceptual
habits that aredeeply entrenched in anthropological thinking and
keep assertingthemselves in various guises and combinations. What I
see as mostdistinctive to my approach is perhaps most vividly
expressed in contrastto these other, in many ways more orthodox,
perspectives.
The most dominant and most indispensable of them is
sTructural-ism. ( believe it to be fundamental to the
conceptualization of anykind of complex reality. But the way it has
been adapted to anthropo-logical materials has entailed a
predominant focus on 'systems ofthought', even when avowedly
speaking about the connections ofsocial interaction and people's
relationship to their environment.It has also focussed strongly on
the macro-level of forms, institutionsand customs, ignoring the
micro-level of the distribution and inter-connections of concrete
acts and activities; or else it has confoundedthe two as if the
acts of individuals were a simple homologue or 'expres-sion', of
collective macro-structures. As a method, it has tended toachieve
clarity - indeed often brilliance - in the depiction of patternsby
a high degree of selectivity: by backgrounding and
eliminatingvariation and abstracting the norm, thereby ignoring
increasinglyrnore of what seems to me real and vital in people's
lives (on the opera-tion of backgrounding in the construction of
shared understandings,see Douglas, 1975: 3-4, and elsewhere). In
these ways, structuralismhas developed in social anthropology in
directions which ever reduceour universe of observation. The
culturally and humanly rich occasionswhen people cOllie together in
the creative act of shaping collectiveunderstandings and
cosmologies are reduced to tex ts, to be distilled
-
lnrroducrion 5
into a series of abstract binary oppos.itions: eating together
is depictedin the ciphers of conventional patterns of dishes: and
the multiplexphenomena of being and changing iden tity become
principles of recruil-ment. There are undeniable insights in these
abslractil'l1s. but theycapture only very few sides to life: what
is more. I mistrust them forbeing flawed even wilhin their
restricted focus. Thus. I expt'ct there isusually a deeply
systematic difference between people's reilectivegeneralizations
about macro-features of tlteir world and society, andtheir
conceptualization of their socbl and physical environment seenas an
opportunity s.iluation for action. To search for the reflectionof
tlle fonner in the cases and patterns of social behaviour is
there-fore unduly simpListic. no matter how sophistic'lted the
analysis ofthese collective representa tiuns may be. lllere is much
perhaps pro-saic but truly illuminating groundwork III he done with
a f:lr hroaderperspective and sense of the problem before one can
hop(' to rcpresentthe premises of meaning for acts with such rigour
and logic - if in-deed valid representations of tllese mailers can
ever achieve suchsimplicity.
The following essays try to do parts of tllis groundwork. In
theirway, they may also go too far in seeking to achieve a degree
of con-ceptual clarity by oversimplifying and obliterating parts of
reality.[ feel attracted bOtll to the 'thick description' of Geertz
(Geertz, 1973)whereby the layers of meaning and context arc
elaborately exploredand exhibited, and to a methodology of
'watching and wondering'in the gentler tradition of the naturalist
(Tinbergen, 1951). But I shouldnot let polemics against a narrow
structuralism tempt me 10 negale myown basic ideal: the analytic
virtues of seeking to construct tight andsimple models on limited
premises. I expect such models to be par-ticularly illuminating,
however, if they seek 10 identify and depicIempirical processes ra
ther than perfolln the arbit rary opera Iions ofsynthesizing and
backgrounding. Thereby they should also avoidthe illusory
timelessness of these other forms of structuralism, andarticulate
with the irreversibility of praxis (cL Bourdieu, 1977) andthus the
realities of people's lives. Furthermore, hy insisting on eventsand
interaction as central features of our object of study, we
arecompelled to confront a far broader and more diverse segment
ofreality. And finally, the virtue of developing models more
rigorouslyand ruthlessly lies not in believing them to be the only
and wholeIruth, but in the occasion it provides to discover their
implications,and what they can and cannot do.
-
6 Introduction
Anothet deeply entrenched mode of explanation in anthropologyis
historical, whether sweepingly evolutionary or more modestly
con-cerned with particular sequences and developments. I have
consistentlyrejected this as the major focus of my anthropology for
three majorreasons:
I It has a depressive effect on the anthropological endeavour
toadopt the view that social facts are the results of (previous)
criticalevents rather than being themselves critical: it is more
productive toassume until proved wrong that everything influencing
the shape ofan event must be there asserting itself at the moment
of the event.This realization was in my view the greatest
achievement of functional-ism, and we are ever in danger of losing
its benefits under the influ-ence of the linear, causal, and
development-oriented modality thatdominates our traditional
thinking.
2 A method that seeks to explain that which one can know
muchabout (the present) by means of that which one must be content
toknow less about (the past) is topsy-turvy and ever in danger of
pro-ducing spurious insight.
3 A historical perspective easily encourages the
misrepresentationinvolved in lightly transforming typological
series into developmentalsequences.
These considerations need not and should not lead one into
asynchronic dogmatism and the rejection of a time perspective.
Anyoneinterested in discovering processes must be alert to
sequences overtime - though not be prepared to confuse any and
every sequence ofchange in aggregate form with a process, as is
frequently done (cr. ch.5, pp. 77 -9). A certain amount of early
training in palaeontology andbiology made me conscious of how much
more productive is the searchfor the mechanisms of change rather
than composing descriptions ofits phylogenetic course, even in the
rigorous sense that such coursesare identified from the fossil
record (cr. ch. 6). I therefore judge it tobe particularly useful
to study variation with meticulous attention toco-variation, bOlh
regionally and through time, as I have done in con-nection with
each of my field studies (cL for example, Vol. II chs. 1,2and 6).
But the alternative to a structuralistic indifference to
particularevents of change is not to mystify change by vesting it
with moral ormetaphysical propcrties, as anthropological
evolutionists and marxiststcnd to do. We must struggle to ascertain
the dynamics of cultures andsocictics in timc as ongoing systems
and through time as emergentscqucnccs. But this we do best by
establishing the facts of the past
-
Introduction 7
where possible. and not by conjectural interpretations based on
preestablished schemas or by pursuing the craft of historiography.
nomatter how competently. Data from the past are analytically
usefulwhen they can surprise us and falsify our hypotheses:
otherwise. I seeno reason why they should receive privileged
interest.
The third dominant mode of thought in anthropology which weneed
constantly to challenge is filllcriotlalism. In
anthropologicaltheory this viewpoint has mainly nourished through
the inappropriateapplication of considerations of purpose and
utility to tile macro-level of aggregate systems, rather than the
micro-level of dccisionmaking units where it belongs. But it is
also nourished by the anthropo-logist's recurrent experience during
fieldwork of Ihe covert fitnessand felicity of customs which had
initially struck us as arbilrary andsenseless.
The essays that follow are concerned to pursue questions of
purposeand utility systematically where persons and groups making
dccisionscan be observed, and rigorously avoid them where they
provide onlyinappropriate metaphor or unsupportable teleology. I
think we cando better anthropology that way: and it should not be
difficult to main-tain a broad and meticulous attention 10 contexl
in thc fieldworksituation, as taught by Malinowski, without
embracing the holisticand functionalistic axioms by which he
justified it. In ils vulgar form.a functionalist stance seems 10
assert that the mere existence of a pattern in a culture other than
our own is proof of its functional fit, whichagain serves as the
explanation of its existence. At the same time thisimage of other,
functionally integrated cultures has served anthropologists as a
foil against which to condemn our own culture and lifesituation
where we experience a disparity between our own desiresand ideals,
and the prevailing (aggregate) patterns and circumstances.Surely,
these are frustrations which people in aU cultures experience,in
the disparity between desire and circumstance, what is and
whatmight have been; and only a gross insensitivity to Ihe actual
life situalion of real people in other cultures could lead one to
think otherwise?Nor can I find a priori justification, or empirical
validation, for a harmony model of culture and society, or culture
and environment. eithergenerally or in any particular case. Indeed,
even where an approximation of such functional harmony is found, it
does not provide in itselfany explanation for the existence of the
cultural patterns in queslionunless the processes whereby this
degree of harmony arises can be rigorously explicated.
-
8 Introduction
The inadequacies of each of the three main viewpoints I have
hereattacked so summarily arise in each case from the simplicity of
one oranother aspect of their implicit constituting premises. How
could onehope to develop an alternative viewpoint which does not
contain analo-gous naws of its own? 0 statement in the following
essays is intendedto claim that I know of such an alternative; and
indeed I have no trustin our ability ever to construct a
theoretical framework capable ofquestioning its own constituting
premises. It is in the way that we relatetheory to the reality
which it seeks to depict that we can hope to createthis possibility
of falsification and correction of assumptions about thatreality.
Perhaps it is in my ideals - and hopefully my praxis - on thispoint
that I feel I differ most from a number of my colleagues: I wantto
work in a discipline in which theory and empirical data are
confronted at a diversity of levels.
It is in the very nature of our object of study that
difficulties shouldarise in performing such confrontation with
reality with sufficientrigour to achieve anything approaching the
operations of testing and'proof' in the sciences (cL ch. I). The
anthropologist in studying humanbehaviour is faced with such a
diversity and near infinity of potentialdata, even for the most
modest encounter (cL e.g. the records ofkinesics, as in
Birdwhistle, 1970, or the problems of constructing ajournal of
observations, as in Pittenger et al., 1960). We are consequently
ever in danger of having our own selectivity become the mainsource
of patterning in our data. What is more, our methodology
ofcomparison, so fundamental to the theoretical constitution of
socialanthropology, suffers in comparison with the biological
sciences: we donot compare specimens and cadavers, only synthetic
descriptions of ourobjects of study. Above all, the symbolic and
social construction ofpeople's realities entail the necessity of
comprehending interpersonalevents by interpreting them, on many
simultaneous levels of meaningand significance, by means of the
codes and keys employed in theirown cullure as well as analysing
them by canons which we can acceptas objectively, materially
adequate. Nor can I, as noted above, see anyway of escaping these
difficulties by a methodology which restrictsour data to forms and
topics where a greater rigour can be achieved(whether by social
survey techniques or New Ethnographies), sincethese procedures only
force more of what may be essentials of humanlife into the
inaccessible void of the unrecorded, and thereby unknowable.
TIle essential problem of any anthropological endeavour is
perhaps
-
Introduction 9
most simply described as adopting procedures of thought and
observa-tion whereby the anthropologist can transcend his own
categlJries andpremises, and thus make discoveries of any degree of
profundity. andexperience surprises that will afTect his
preconceptions about cultures,societies and human life.
When the issue is put this way, it becomes apparent that we in
facthave a number of procedures at our disposal though these have
rarelybeen used in concert or given proper recognition as the
essential under-pinnings of a methodology for social anthropology.
(What followsassumes the practice of the cluster of techniques
summarized under thecaption 'participant observation', needed to
allow us to be in an ade-quate observer position at all.) Being
essentially ways of relating theo-retical constructions and
reality, these procedures describe perhapsmost succinctly my own
ideal of what anthropology should be about.
I We need to adopt the naturalist's stance of 'watching and
wonder-ing'. This involves the construction of extensive and
meticulous descrip-tions, but presented with an economy of words
and dimensions. There-fore, it entails model-building - but models
constructed for the pur-pose of more thorough, comprehensive, and
economical description.The emphasis on 'watching' entails a
recognition of the essentialseparation of rhree levels: (a)
theories, explanations, and other con-structions; (b) observations,
data; and (c) reality ou t there. 'Wa tching'means trying to
capture chunks of this reality by transforming them toobservations
by procedures that entail tlle least possible selection
orconstruction in terms of whatever theory is at issue, so tllat
falsifyingconfrontations of theory and observations become
possible.
2 We should construct gel1er'UiYe mo.dcJs (eL chapters 2 and
5)rather than attempt to distil re'presen~tjQDaI models from
complexdata. Thereby, we are led to search for the empirical
processes behindobservable regularities in the material, and to
depict these processesin the operations, or transformation rules
which we employ in ourgenerative models. What is more. such models
can be used inductively,and the patterns they produce can be
confronted wilh data collectedindependently of the model's
premises. Thereby the procedures forfalsification are strengthened
drastically.
3 We should capitalize on our unique advantage: that our 'object
ofstudy' can help us actively to transcend our categories by
leaching liStheir own. This means recognizing that the actors'
categories providea way to understand reality. as well as being
part of that reality. Inpractice, probably most of the productivity
of the anthropologist
-
10 lntroducrion
derives from this source, even though his arrogance as a
professionalacademic, and his defensiveness when his own reality is
being threat-ened by the enchanted world of another culture, both
militate againstsuch learning.
4 We should expose our scholarship to the demands of
practicalaction so that its incompleteness and unworkable ness can
be revealed.I see such experiences in applied anthropology as
immensely stimulat-ing and chaUenging episodes in any academic
career and have experienced them as such myself. Applied activities
raise other and difficultissues, however. The colonial and
imperialist involvements of someprevious and contemporary coUeagues
give cause for misgivings. like-wise, our exposure to the
insensitive complacency of internationalorganizatiuns, or
disillusionment with the independent governmentsof new nations, may
repel. One also becomes less confident of one'sown ideological
standards when one recognizes the confident idealismwith which
errors have previously been perpetrated by
thoughtfulanthropologists. Yet unless we are to commit ourselves to
completesocial passivity in this world as constituted, there is no
alternativeto acting while enquiring carefully into whose interests
we serve byaction, and whether the act to which we become party is
in itselfdefensible, in respect of alternatives. The intricacy and
agony of thesequestions are as inducive to realism in our
anthropology as the moreexternal ascertainment of the consequences
of practical action.
S We should focus our observations on real people in real life
situa-tions, with a curiosity as to what an investigation into
their situationmay bring - rather than pursue highly abstracted and
demarcated topicsof investigation according to predetermined
methods. The latter alter-native is more suitable for the
construction of good grant applicationsthan for the pursuit of
innovative scholarship. Both topic and methodmust be allowed to
develop in response to the concrete situation offieldwork and the
findings that accumulate.
6 We should espouse the ideal - though not the realism - of
cumula-tive mapping to produce an ethnographic Whole World
Catalogue. Adiscussion of the 'crisis of anthropology' in terms of
the disappearanceof its subject ma tier can only arise from an
academic practice of con-ceptual 'backgrounding' to the point of
subjective blindness. Realityis entirely different: social
anthropology has as its object an immenseand still, in many
essentials, truly uncharted world. The urgent tasksof anthropology
- the recording of swiftly disappearing cultural andsocial forms
which embody the results of millenia of cultural creativity
-
fnrroducrion II
in countless distinctive traditions - are far more numerous and
complexthan we can solve with available resources. What is more.
the publicizeJperfonnance of these tasks may in its direct and
indirect consequcnasperhaps also be the most powerful way of
resisting the tragic and s,'nse-less destruction of ways of life
cherished by other people. some clf themperhaps happier than
ourselves. But on the even hroader ,-anva, lklln,'lIby the life
situations of people. there are manifestly nHlIr pc,,'pk inthe
world today than ever. cullurally and sorially mor,' intric"ately
inter-dependent than ever before. II seems to me a shllCking
rellection "f thc'scholasticism of some of our colleagues that this
hould Iwt be IIniwr-sally recognized to constitute an inexhaustible
field f(l' anthropc'logic:tlinvestiga tion.
The essays republished below reflect this conception of our
objectof study and seek to achieve such a relationship between
concept andtheory and the empirical object. Thus, the theoretical
t1evelopmentsthey contain are motivated by - or perhaps more
correctly. their neces-sity has seemed to me to have arisen from -
the substantive events that( observed. Anyone who has ever
attempted serious fieldwork willrecognize that events have a sheer
mass and diversity far greater Ihanmy total published corpus; I
have never subscribed to the illusion thatit is fruitful to publish
materials outside of the areas that have been atleast provisionally
grasped by a conceptual framework. These essays.much more
singlemindedly than my main writings in monograph form,focus
primarily on such frameworks: on various ways of conceptualiz-ing
which may enable us to build models suitable for
fundamentallydescriptive purposes: models that enable us 10 provide
coherent,reflective accounts of the interconnections that obtain
between wideand varied sets of observations. I see our task in
developing theory asan unending struggle to increase the
comprehensiveness of materialwhich we can thereby cover, and the
elegance and completeness of ourcoverage.
The various issues and contributions with which I have
struggledinterweave, reappear and are restated and modified in
these variousessays. I should like only briefly to point to tJlOse
which I myself, atthe present time, see as the most important:
ri-An emphasis on The concrete life situation or opportunity
situa-tion of the actor as the essential and significant context
for his act. Itis not by moving directly to the more rarified
levels of norm and pat-tern, and derivative conceptualizations of
deviance or structure, thatwe can identify the actor's point of
view and thus the factors that
-
12 Inrroducn'on
impinge directly on - or more correctly, are an inherent part of
- hisact. First of all, our understanding of what he is doing must
be solidlygrounded in a knowledge of that particular actor's
resources andassets, his alters and significant others, his
manifold commitmentsand interests, and his environmental and
communicative options.
2 A persisten t effort to develop and clarify our main
conceptualvehicle for describing social systems and
institutionalized behaviourviz the concept of status. This theme
seems to me to receive entirelyinadequate attention by most
colleagues, even though the inadequaciesof textbook definitions are
widely recognized.
3 The dynamic interconnection between macro- and
micro-phe-nomena (particularly the concepts of aggregation and
consrraint) ,which may provide the most fruitful ways to articulate
the centralquestions of choice, freedom, history and the ontology
of society andculture.
4 The relationship between social and cultural phenomena and
theirenvironment, particularly their preconditions and consequences
in thatenvironment. For this purpose I have sought to utilize an
ecologicalperspective without submerging the anthropological task
in it. Con-sequently, despite my early and fairly comprehensive
efforts to intro-duce ecological thinking in to our discipline, I
have reservations againstsome of the major currents and trends in
this sub-field, as will beapparent from a reading of the relevant
articles.
5 An interest in developing concepts to analyse the process
ofexchange in in teraction - not so as to apply a market model to
sociallife but on the contrary to study the implications of
exchange as oneof the basic forms of constraints whereby the
behaviours of two ormore actors are constituted as systems, and
thereby to contribute toa general theory of social systems.
6 The rela tionship between codification and praxis, or how
formsof understanding arise from experience, and reciprocally how
behaviourand experience are predicated by collective
representations. This isonly partly articulated in the present
collection (see particularly chap-ter 8) but has been developed
further in a separate monograph (Barth,1975). I believe there is a
possibility of integrating much of my pre-vious theoretical' work
in a broad kind of sociology of knowledge treat-ing this cen tral
theme.
TIlere is a temptation in writing an introduction to one's
ownold articles to repeat, refonnulate, and construct ever more
cryptic
-
III rroduction 13
and speculative meta-statements over the same themes. R:llher
thancontinue. it would probably be wiser to let the various texts
speak forthemselves.
FreJrik BarthOslo. June \1)7
-
1 Anthropological modelsand social reality
In trying to show you the character of social an thropology as
anacademic discipline, I migh t try to sketch some substantive and
perhapsintriguing findings in the field, or the history of its
development, orsome of its major inte ect~~ problems tQSl~ . I have
chosen the last ofthese alternatives, because by showing the
general problems we aregrappling with I hope to reveal to you, in
part no doubt inadvertently,the ways that anthropologists think,
and also how our difficulties inpart arise from the character of
the social reality itself, which weconfront and try to
understand.
The fundamental questions which social anthropology asks
areabout the forms, the nature, and the extent of order in human
social /'life, as it can be observed in the different parts of the
world. Thereis no need to prejudge the extent of this order; as
members of onesociety we know how unpredictable social life can be.
But concretely,human life varies greatly around the world, and it
seems possible tocharacterize its forms to some extent. We seek
means systematicallyto discover, record and understand these
forms.
I wish to emphasize this fundamentally empirical view: we di
.;.... cover and record .~ do not comment and evaluate. The
fundamental
approach is thus that'at: jici
-
Anthropological models and social reali(\' 15models - models in
the broad sense that they are represelltationsof an interrelated
set of assumed faclOrs Illat determine or 'expbin'the phenomena we
observe.
But there is one circumstance that makes our discipline
differentfrom the natural sciences. From our own life, we feel that
it is un-deniable and true that human behaviour is prominently
shaped byconsciousness and purpose. Anthropologists are thereforc
prepared tospeak about things like beliefS, obliga~ons, and values,
not just L1ll ,mediate, overt behaviour. T11is also means that an
cxplanalmy Illodclf 'our can be different from the models used in
nalural scicnce.Human beh~~ia'uT) is 'explained' if we show (a) the
utility of its con-sequences in terms of values held by the aClOr.
and (b) the awarenesson the part of the actor of the conneclion
between an act and its speci-fic results. This is what has been
called the SJ,IbjeCliPe vjewpoinl. or byMax Weber
'Verstehend!:LSoziologie'. Indeed, most anthropologistsseem to feel
that unless the actor's point of view can be made under,standable
in such cognitive and valuational temls, his behaviour
cannotbe-explained. On the other hand, the psychological mechanism
behindsuch 'purposeful action' constitutes a different field of
inquiry whichdoes not concern us: our interest is not 10 refine our
underslanding ofit. Our concern is to understand the different
societies and cultures Ithat are based on Ihis mechanism by
relating social behaviour 10 theconceptual and valuational syslems
of the actors. I
This view, however, has methodological implications for how we
goabout finding the causes or determinants of behaviour. To take
afrequently used example: traffic stops at a red ligh!. Can we
point toa red stop-light as the 'cause' of auto-drivers stopping
Iheir cars, inview of the fact that occasionally a driver will not
stop at a red light?If one contrary case does not prove us wrong,
holV many cases wouldit take before this explanatory model was
falsified?
There is a sense in which this argument misses the poinl by
nulunderstanding what kind of claims to adequacy we make for
ourmodels. To find out if cases of not-stopping challenge our
model, wewould want 10 know (a) can they be understood - and
dismissed - ascases of inattentiveness? That is, was it the actor's
immmediate percep-tion that failed in this situation, though Ihe
cognitive rule that redmeans 'stop' still holds true? Or (b) can
these exceptions be dismissedas cases of atypical utilities?The
evaluation which makes people stop iscompounded of the danger of
accidents and the probabilily of sanc-tions by the police. If a man
has some uniquely urgenl reason to travel
-
16 Anthropological models and social reality
fast, the utility of arriving quickly might outweigh the risks
of notstopping. Only if both these interpretations fail can such an
event betruly regarded as a 'con trary case'. Sta tistics on the
overt behaviourof populations do not readily supply a test of the
explanation. Theycertainly do not challenge or confirm statements
such as 'red meansstop' - and such statements constitute some of
our major discoveries inthe analysis of exotic societies. One might
be tempted to say that afrequent discrepancy between the observed
and the predicted behaviourreduces the relevance, but not
necessarily the validity, of such models.Yet we do wish to falsify
our models by confronting them with empiri-cal facts, and this
involves methodological problems to which I shallreturn
shortly.
The empirical bent in anthropology takes a form which is
reminiscent)'--of natural history. Indeed, it has sprung directly
from that academic
tradition. In this anthropology differs from most of the
behaviouralsciences, which to greater extent have their roots in
philosophy andthe humanities. An thropologists tend to seek the
strange and theexotic, approaching such matters with the
naturalist's curiosity. Theeffect of such an approach is that one
exposes oneself to the greatestpossible variety and mass of
impressions. The immediate purpose ofthis is, of course, to collect
and cross-check information. But it alsocolours the
anthropologist's theoretical orientation by encouragingeclectic
model-building: th_e anth!opologisLis concerned abouLthedirect
relevance of theories to empirical facts, because of his con-stant
neeoto order a diversity of such facts. This does not mean
that,temperamentally, many social anthropologists may not wish to
protecttheir orthodoxy from uncongenial facts, or to work
deductively froma basic theory of society. But the situation that
they place themselvesin, surrounded by the strangeness of a foreign
society, is one thatmakes this maximally difficult. In this they
differ for example fromthe economist, who works deductively at his
desk and seeks to falsifyhis constructs against a limited range of
data. Among us, the term'armchair anthropologist' has been a
standard term of abuse. Butperhaps I am merely saying that to the
extent social anthropology isa science, it is a very primitive
science?
Let me practise this natural history concreteness by taking
youbriefly to one of the classical loci of British social
anthropology, toillustrate the confrontation of anthropological
models and socialreality. The place is one or another of the
communities in the Suddswamps of the Sudan, inhabited by the Nuer
people, made famous
-
Anthropological models and social reality 17
through the writings of Professor EvansPritchard.111ese people
live in shifting camps in periodically inundated ~ountry.
They practise agriculture. though they are more interested in
theirherds of cattle. In the wet season. the period of lloods. they
congregatein villages of 50 to 200 persons. Around ovember they
disperseth.rough the country in small cattle camps. until
increasing desiccationdrives them together around permanent sources
of water in large dryseason camps of 100 to 1000 members.
Of the various kinds of order which one may search for in such
communities, let us first concentrate our attention on the
relations of communities to each other, and to larger territories -
what might be calledlitical lif'e:) We find a situation where, in
disputes and battle. thepeople who are allied at one moment, over
one issue, are opposed toeach other the next moment. There are no
chiefs or. apparently, stableleaders of any kind. The problem migh
t seem to be one of finding anykind of order at all. It was one of
EvansPritchard's conlributions toshow how a political or(!er is
maintained in this society without theinstitution of chiefs and
stable authority, through the operation of aframework of kinship
relations of a scale quite unmatched ill our ownsociety.
Let me frrst try to depict this order aN system of jural rules
govern-ing member~h!1Lin gr.o!,!ps and the institutional relevance
of tllesegroups; Nuer men identify themselves as members of tribes
and theirlocal segments, down to the level of the local community.
111is identi-fication refers to residence, and to political duties
of a kind whichimplies or produces a political corporation. The
duties are relevant forbehaviour in daily life and at ritual
occasions, and they are especiallydrama tically manifested in feud,
when groups of warriors wreak collec-tive vengeance.
In all this activity there is a characteristic rela!"' of
groups. Thusif a tribe were divided into two local segmen _ and B,
each with sub-divisions I and 2, the people of B2 will haVe pnm
obligations toeach other and unite against the men of BI; but if a
man in BI isharmed by a man in AI or A2, they will all combine as
B's, holding allthe A's collectively responsible. There is thus a
pattern of situationalfusion and fission of groups, depending on
the relative positions ofthe original parties to the dispute within
a hierarchy of segments.
Furthermore, each such tribal territory or segment is
identifiedwith a clan or lineage within the clan. Clan and lineage
membershipis ascribed by atrilineal descent - namely, determined by
the father's
-
18 Anthropological models and social reality
position. The Nuer can thus!!resent an agnatic genealogy that
corres--ponds to the hlerarcQy of ern! r~~nts..:. the tribe is
identifiedwith a deceased ancestor, each of its primary
subdivisions with oneof his sons, etc. In that way, the duties
between co-members of a com-munity and those between agnatic
kinsmen become one
This whole pallern can b~ represented as a set o?jur~'
~~Therules are jural, although there is no formal system of courts
and lawenforcement, both in that they are. embraced b the Nuer as
moraland n t, andm -lilat they _are sanctioned -
compliancelSpra1SeCl
and supported, breaches are criticized and punished by public
opinion.A jural rule model of the system depicts the regularity of
behaviourin terms of group structure: a set of rules regulating the
mode ofrecruitment to groups, and the duties that group membership
entails.
Many of the primitive political systems which have been
analysedprove to utilize kinship-based criteria of recruitment to
political groups.Where the ascri tion .. aI rights and obligatioos
is rie.Ll.tD.-O.!le~ of descent either male or female -lInd enealo
'es are remembere
-
Anthropological models and social reality 19
are those of agnatic kinship. Political systems of this type
have therefore somelimes been characterized by social
antlHopologists as systemswhere politics is cast in the 'idiom' of
kinship. If the actors lhemselvesconceptualize political soda1~s
groups of brothers. th,')' use aconcept which implies situational
disregard l)f collateral distance. andwhich force the
representation of relalions betwcen grl'ups intl) asegmentary
mould. The characleristics uf siluatiunal fusi,'n and fissiunfollow
from this. and the anthropologist can thus und,'rstand regubri.
1ties in the political life by analysing lhe implications of hmv
lhc actursconceptualize their political groups - what might bc
called the rclevantcognitive structures in their culture. The fact
lhat all mcmbers of thecommunity are not brotllers. or even
agnales. is nut as problema I ical asit is for the jural rule
model: t~.!itorial ~gnlro!;llion is implicit in thi:basic
calegories in ~hich political .groups are conceptualized. Thusthe
detailed discrepancies become rather unimportant and uninteresting:
the reality with which such models should be confronted is
thecategory of discourse, of ritual affirmation, etc. (cr., for
example.Leach ,1961).
What have been called 'structural' models of societies have
tended tocombine features of these two ways of representation: both
as juralrules and cognitive categories. This has enabled
anthropologists toexhibit the major groups in a society, how they
arc interrelated interms of recruitment, and what is their major
institutional field ofrelevance.
Nor do such models need 10 ignore alternative pallerns of
behaviourand their implica tions for the system. Professor
Gluckman, in areanalysis of the Nuer, has specifically emphasized
the effects of theintermixture in villages of persons from
different lineages and clans.The man living in the territory of a
different lineage from his own willhave essentially similar
obligations (a) to his distant lineage mates, and(b) to his local
neighbours. This may create dilemmas uf cross,cullingloyalties in
the case of feud between groups: as a member of a lineagehe may be
on one side, as a member of a village identified with
anotherlineage he may be on the other side in a conflict. Such a
positionboth motivates and qualifies the man to act as gobetween
and peacemaker in the feud; so the consequence of the dispersal of
personneloutside the territories of their clans is to create both a
force and amechanism in the service of maintaining ~~~ over larger
regions,and thus paradoxically to maintain the formal pallern w UCh
they,individually, have not followed.
-
20 Anthropological models and social reality
This way of arguing exemplifies in its most explicit form
anothertraditional feature of anthropological thinking: the
society-organismanalogy as a basis for _functionalism, whereby one
shows how aspectsof form serve a purpose in maintaining the system
as a whole. As a wayof depicting the prerequisites of system
maintenance, this representsa different explanation of form from
that of rmding the actor's own'purpose' - a circumstance which has
occasioned some confusion inanthropology but which has largely been
overcome. Most anthropo-logists, however, have chosen not to pursue
this view of society furtherand have not adopted a Darwinian
perspective.
Structuralist models thus enable us to depict the gross
morphologyof societies, and have made it possible to describe and
compare agreat diversity of forms. But with an anthropologist's
field experience,it is difficult for long to avoid raising, in some
form, the questionof why specific actors behave the way they do. Th
uctural kind ofmodel emphasizes cognitive cate ories and the
ositive ans~ ;- ns 0 u r But living in our Nuer village, we cannot
but beimpressed by the 'case stories' around us: why did our
neighbourchoose to live with his mother's or his wife's kin? Some
cases areeasily explained. Among the Nuer there are a number of
small lineagesand clans that have no land of their own and must
live interspersedamong others. There are also individuals and lines
that have beenadopted from defeated neighbouring tribes. But many
others, thoughmembers of clans which have aristocra tic rank in
their home territories,yet choose to live elsewhere as foreigners.
Why?
To answer such questions, we must go back to the basic view
ofbehaviour and utility which served as our initial point of
departure.The simplest structural argument would be that there are
jural rules,implying sanctions, which support the established
patterns; so mostactors acknowledge these and act accordingly,
while some deviate fromthe rules in random fashion. Alternatively,
we may rmd that the rulesare themselves inconsistent and thus
produce variability in behaviour.But rather, we may look more
closely at all cases of behaviour and askwhat it is that makes
people act the way they do: what goods arepeople after, and how do
cognitive categories, rules and sanctions - aswell as the local
ecology - constrain their possible courses of action inpursuit of
such goods? Though he does not utilize it in constructing hisbasic
model of Nuer society, EvansPritchard, the fieldworker, does
notfail to recognize such factors. He notes that there are
variations inpopulation density throUgll Nuer country, and that it
therefore may be
-
Anthropological models and social reality 21
mOte advantageous for a man to live as a foreigner in an area of
3mpleland, water and grazing than as a member of lhe dominant .:1Jn
in alloverpopulated area. He also refers to some men's desire for
autonL1my:collateral lines are arranged in order of seniority. and
members ofjunior lines may sometimes achieve greater autonomy away
from lheircollaterals than close to them.
To depict such facts and discover and interrelate such
explanatoryfactors systematically, however. we need to build models
of a differentk.ind. Regularities in behaviour can be represented
as, relative fre::'>
uencle i.e-as....a...kirul.. f order that emerges from the
independentactivttie~ of n1J!!!iple
-
22 Anthropological models and social reality
land can be conquered only at someone else's expense; and there
is aclear tactical advantage in attackin ates - thereb
mobilizingthe smallest ossible 0 osed grOUP, since only those more
closelyrelated to the victims than yourself will rise in defence of
the owner's,and their own, prior rights to the lands in question.
This means thatlineages tend to break up in opposed splinters or
fractions. When suchsplinters seek alJies for support against their
close collateral rivals. theycan lind them in similar, distantly
related splinters of other lineagesegments, by promising mutual
support against their respective rivals.In this situation, then, Al
and BI tend to form a stable coalitionagainst A2 and B2. The
stronger coalition will be victorious in a con-frontation -: which
means tbat:l!ie.lC WI _...e a D.nsistenLpLeSSure.-towards the
systematization of larger and larger coalHions, culminatingin the
forma tion of a two-party syste_m of two large, dispersed
politicalblocs.
This system can be simula ted as a game of strategy and
coalition-forming by utilizing the Theory of Games (Neumann and
Morgenstern,1947). Thereby, a more dgOr.o.US...lllitd.d.can he
constructed. based onthe notion of value and utility and showing
the tactical constraints onchoice which generate such a political
fOJ]11 (Barth, 1959b). We canthus see that a kind of order, in the
sense of regular patterns of be-haviour, emerges that cannot be
adequately depicted in terms of ruleswith jural sanctions, nor as
the direct expression of a cognitive pattern.A jural rule model
would show us the structure of lineages with refer-ence to land
rights, but would not explain political alignments. Thejural rules
governing political relations merely state that a client may'join
any patron he likes, and a patron may join any coalition he likes
-rules that very imperfectly depict the regularities that may be
observed.Nor does the cognitive structure of a two-party system and
the partici-pants' awareness of the processes of balance between
the parties explainthe regularity of political opposition that
divides each lineage. Thegames model thus alJows us to depict some
further aspects of politicalorder Whkh the previous models could
not readily encompass.
The lineage-politics problems I have sketched above are of
courseonly one limited aspect of the social order among the Nuer,
as amongother peoples. Anthropologists are equally concerned to
study econo-mic or ritual behaviour. But such institutional
categories as politicS;or rilual are perhaps best regarded as
conveniences of communicationand provisional description: through
studying such fields of behaviour,!we seek to arrive at an
understanding of a more comprehensive orderl
-
Anthropological models and social reality 23which can be
represented as a social system: a complex system ofsocial posi
tions and in terrela lions in terms of which
behaviouralregularities in many institCitional fields may be
understood and theircongruence and interdependence discovered.
In the previous uffield lecture, Professor olow discussed the
problems of micro- and macro-analysis in economics. and pointed out
thedifficulties experienced in that discipline in connecting the
analyses ofdeterminants of individual behaviour with those of
aggregate behaviour.Any study of human behaviour will be concerned
Wilh this question;yet you may have noted that I have not treated
this as a particularlytroubling problem from the social
anthropologist's point of view.Indeed, I would say that it is one
of the strengths of amhropologicalanalysis that we are more
consistently concerned and able to inter-relate the individual and
the aggregate level than most of the behav-ioural sciences, though
this is probably achieved at the cost of someelegance in our
analyses of both.
In part, the greater unity of the micro- and macfl}level is
achievedby our emphasis on understanding the actor's subjective
point of viewas our own point of departure, This leads us to try to
c laracteriLe theaggregate system, not by operationally
sophisticated and well measuredindices, but b the shared cognitive
categories and ross values of t1~
1lMJ.i9.~ e s stem, Whether as an interest in a nalive
informant's description of 'custom'. a careful attention to native
words andconcepts, or an increasingly sophisticated participation
in discussions ofthe Realpolitik of a local area, the
anLhropologist seeks to form a pic,ture of the macro-system in
categories related to those which theparticipants 'hold, t ou 1
without accepting their account of the systemas empirically
correct. Also it should be recognized that must anthropo-logists
work among groups or are interested in problems for which weare
provided no aggregate data by statisticians, We must accumulalethem
ourselves through our observations and from informants. andwill
thus tend to accumulate them in categories that are those of
theindividual actor's cognitive schema and our own intended
analysis,producing thereby also an inadvertent congruence between
the microand macro-levels.
A common feature of anthropological models is that they
presumethe relative stability and maintenance of form and are
ncerned wilhexplaining pers~tence rather than chan e. lllis is not
to deny the factso c ge, u spnngs rather from a realization of the
precariousnature of any semblance of order and continuity in so
volatile a mailer
-
24 Anthropological models and social reality
as interpersonal behaviour in a human population. The
explanationof relative stability and order is sough t in the
permanence of determin-ing or limiting factors. The underlying
viewpoint has been expressed inthe axiom that 't.here are no
privileged moments' (cf. Gellner, 1958)-that patterns of social
form are not rela ted in a uniquely illuminating orerivative way 0
specific events in the past, and thus adequately under-
stood by a historical explanation. Though the axiom in this
generalform is probably untrue, it has proved very fruitful in
directing investigation towards verifiable constraints in the
present rather thanreconstructed events in the past, deduced often
from the very factswhich they are called upon to explain.
Naturally, this viewpoint isequally adequate as a basis for dynamic
models as for social statics-an assertion which is regarded as more
debatable among humaniststhan I expect it will be among a group of
scientists. Some attempthas. been made to express this viewpoint by
means of-a distinctionbetween 'historical' and 'structural' time;
mostly it is merely implicitin the kinds of explanations that are
sought and the models that arebuilt.
In the manner of a naturalist, then, the anthropologist builds
amodel, or a number of loosely connected, partial models, of the
societyhe observes around him as an ongoing system. But what are
the obser-vations whereby the adequacy or inadequacy of these' odel
areested, r documented, by the investigator?
n thropologists are here faced with a great methodological
problem:the distance between our concrete sensory impressions of
events 'outthere' in a society, and the kind of data one can use to
falsify anthropological models. The question at issue is,
fundamentally, the nature ofthe 'social reality' with which we can
hope to confront our models:what kind of order do we have reason to
think that we will fmd?According to the anthropological view there
are, in each society...andculture. 'ru!tllraL categories of events
- cgnsisting QC~ol;ial ~tatuses!!.!1d '!leanings - whi h--!fe
unigue for each system and must be dis-covered; ll!'d the extent of
order ,in social life becomes apparent onlywhen observation and
description takes place with reference to such
-
Anlhropologicalmodels and social rea!ily
't~nslate' the meanings of evelllS in a foreign .:ultur
-
26 Anthropological models and social reality
probably even knows that A is a common villager and B a
headman.One can therefore judge by context which alternative
meaning holds inthis case, and understand that A is expressing his
submission under Bby giving tribute. Bu t as observers who wish to
demonstrate or falsifythe presumed inequality between A, who is a
relatively undistinguishedviUager, and B, who has been identified
at ano~asion as a head-man, how does this event help us? We have to
translat to give the actits appropriate meaning in the culture, and
thisseems to require pre-vious knowledge both of the 'code' and of
the social relations.
3 Still one open question remains: were A and B acting in
theircapacities as commoner and village headman? These are small
andintimate local groups where we observe the same people in what
wemust distinguish as different capacities: we see the village
headman athome being a father and a mother's brother to various
people, we meethim in the fields being a cultivator on his own
land.
We also see the same equipment, a spear, used as a kitchen
utensilto cut meat, or as a weapon, or as an object of wealth for
barter andpayment, or as a sign of rank. How do we break the code
so we nolonger talk about naive categories like spears and men, but
aboutheadmen, fatheIs...s.igns of rank, items of tribute? Our
anthropologicalmodels are not about people and spears, and cannot
be falsified by eventhe most exhaustive statistics on the gross
activities of people or theuses of spears. Indeed, we can
legitimately doubt that such statisticswould reveal very marked
regularities at all - certainly most of theorder in social life
characterizes other units than these. Events involv-
1ing men and spears need to be translated in terms of what they
mean,in a social system and a cultural idiom, before they become
data to theanthropologist. '
But once we have performed this translation, have we not
'cooked'our evidence so that it must fit our models - is the
operation of trans-lation one that is independent enough of our
models so that it can pro-duce data that can falsify these
models?
Though some of my colleagues might argue otherwise, I would
saythat we have no..l!.~etlc'!!!i: founded methodology whereby we
canbreak into this closed circle and start trans a mg.--S-ome
-linguisticallyoriented anthropologists in the United Statesare
trying to develop sucha methodology (cf., for example, Frake,
1962): the British tradition ismuch weaker in its theory on this
point. Its practice, on the otherhand, has been bolll rigorous and
successful, involving a nearly totalsubmersion in the foreign
culture through the technique of participant
-
Anthropological models and social reality
observation. The purpose of this technique is 10 creale an
oplimallearning situation, and this involves control of Ihe native
language. wntinualpresence in the community, disconlinuation of
contact wilh membersof one's own sociery, and participation in as
wide a range of altivitiesin the native conununity as possible.
This is a situation that makesone maximally sensitive to the
reactions and conlrols of members inthe foreign society. Not only
does it increase one's awareness of what isgoing on in that
sociely; it also enhances ils inlportauce to oueself. asthe only
source of human companionship and conta,1. It is in Ihissituation,
by an imaginative leap into the unknown. tJut one breaksinto the
closed circle of the foreign culture and obtains the 'bridge-heads'
from which one can begin to work.
It is not too clear what are tJle clues which we use to obtain
thisinitial minimal understanding. I have had the experience of
Hying toanalyse and edit tJle rich, but unprocessed, field notes of
a colleaguewho died during fieldwork, and have experien,ed this
frustrating feel-ing of having so much information but no
understanding, somehowlacking some part of the key to the
significance of the recorded events.I was finally able to go to the
area in question, and see the peoplespring to life. Even after
that, I am not quite sure whal had been miss-ing and what it was
that I observed tJtat suddenly gave meaning to mycolleague's
record. But parI of it was the setting, the landscape - notin the
poetic sense, but as a place to live. I knew of course tJlat
thenotes dealt with a herding people in a mountain desert - but my
col-league, living in it, had taken tJle details of this setting
for granted;and to understand the significance of acts I needed 10
know, or see,what the feasible alternatives to any particular act
were, which aspectsof the fonn of behaviour were connected with the
pure necessities ofthe demanding task of living in this
environment, and which were arbitrary, reflecting the prejudices of
tJle culture and the spiril of the actor.
Some methodology can also be based on the fairly readily
observableand objective difference in a social relation between
parties who initialea sequence of interaction and those who cannot
do so, or those whocan terminate such a sequence of in teraction
and those who mustalways respond. Such criteria come closer than
anything else to provid-ing an observationally based, culture-free
definition of 'authority' (cLChapple and Coon, 1942).
But mostly, I think the clues we use are those that reveal
'1!liludesaDd Vaill es by emotional ~i81'18Y.We have assumed
provisionally, and sofar without being disproved, that basic
reflexes such as laughing, crying,
-
28 Anthropological models and social realityscreaming,
trembling, etc., are released by the same basic attitudes
andemotional states in all Homo sapiens. By observing such
emotionaldisplay, therefore, we can obtain clues about the
significance or meaning of cultural idioms. Thus Bateson (1958)
tells how he constructeda provisional interpretation of a
transvestite ceremony among a peoplein New Guinea on the basis of
second-hand information, and how hisunderstanding of its meaning
was suddenly transformed when he saw,during the first complete
ceremony he attended, how the males dressedas women were treated as
figures of fun and laughed at during the performance, while the
females personified arrogant male warriors andwere proud and
aggressive.
By thco;c means, the anthropologist in the field catches hold of
cluesas 10 Ihe meanings of acts, and progressively builds up a
participant'sImp of soda I conventions, of the different social
situations, and what isregarded as suitable for persons in
different positions in the differentsituations. lie thereby learns
10 differentiate the roles and statuses ofthe social system, and to
interpret activities in the contexts th~cuTturaJJy
appropri~.~,..!I1..d Ulat define their meaning. Fortunately,
allidioms do not, like the example of the transfer of the spear
discussed"bove, st"nd for radically opposed meanings in different
situations:some turn out to be very unequivoc,,1 and can be used as
signpostswherever they appeilr. By aggregating very many diverse
facts and,through the successes and embarrilssments of
participation, havingthem related and systematized, the fieldworker
progressively improvesand expilnds his understanding of the social
life around him. Participation also gives opportunities for what
has sometimes been called socialdnll11
-
Anthropological models and social reality
meaning and valuational altitudes exemplified in the various
idiomsand forms of behaviour. thereby simplifying the descriptions
and'making sense of whole syndromes of activities.
In other words. it is as field worker rather than as analyst
that theanthropologist produces data in this 'translated' fonn.
TI,e o(>
-
30 Anthropological models and social reality
What is more, the monograph has a classic unity of time and
place: itis typically a synchronic study of a small, delimited
community. [tthus provides a picture of unity and integration. Even
where our actualanalysis fails to show the interdependence, this
image of the culture andsociety as a 'whole' remains - a unity of
perhaps spurious, but none theless convincing, character.
By these means, we try to collect and systematize observations
thatenable us to depict the reality that confronts us: the
regularities of lifein human communities. We regard this reality as
the resultant of thebehaviour of many actors, separately shaping
their own acts accordingto their subjective view of the
opportunities offered by their world andtheir society. But what we
observe is overt behaviour and the objectiveconsequences of this
behaviour - whether sought or unsought by theactors.
Social an thropologists construct models which seek to take
thisinto account. We feel we cannot adequately represent, much less
under-stand, the forms cf regularity except with reference to their
under-lying determinants, and this entails the need to map out the
culture,the actor's own categories and meanings, to depict the
social system.
I have tried to sketch for you the three most common kinds
ofmodel which social anthropologists make use of, whereby they
depictthe social order:
J A system of ju ral rules, consisting of the basic ideas about
inter-personal obligations and rights which are embodied in a
culture. Someanthropologists would go further and claim a reality
of its own forthis, the 'moral community'.
2 A system of cognitive categories, especially those whereby
themembers of a society conceptualize and order their own social
statusesand social groups.
3 An interactional system - a model that depicts the constraints
onindividual behaviour that arises from the behaviour of others in
a socialsystem. Such models assume some fOnTI of 'economizing' or
competi-tion or strategy on the part of actors whereby they
mutually modifytheir behaviour in respect to the objects that are
valued in their culture.
By relying mainly on the first of these fOnTIS of
representation, onewill lend 10 construct rather static models of
each society as a set ofrules. and one can show a remarkable
consistency and integrationbetween various features of order. Less
interest centres on actualbehaviour in deviance from these rules,
and the timelessness of asynchronic sludy lends to lead to an
assertion of stability for such
-
Anthropological models and social rea/in' 31
systems. [n its naive and ori al fornl. the expbnati
-
2 Models of social organization I
Introduction
TIlesc essays are an attempt to consolidate and use generative
models insocial anthropology. Drawing on concepts and viewpoints
which arecurrent in anthropological literature, I delimit and
construct a fewsimple models of this type, discuss the way they can
be utilized andtheir relation to empirical processes in society,
and illustrate theirutility as applied to a few examples. The
essays are essentially the textof three lectures presented under
the heading 'The explanation ofsocial forms' at the London School
of Economics in the winter of1963.
Form in social life is constituted by a series of regularities
in a largebody of individual items of behaviour. Much effort in
social anthropo-logy has been concentrated on the necessary step of
constructingmodels or patterns descriptive of such forms, whereby
structural fea- rtures of the society are exhibited. The kind of
models which I discusshere are of a different kind. They are not
designed to be homologouswith observed social regularities; instead
they are designed so that they,by specified operations, can
Gellercl.l~ su.cll...ITgylarities or forms. Theyshould be
constituted of a limited number of clearly abstracted parts,the
magnitude or constellation of which can be varied, so that onemodel
can be made to produce a number of different forms. Thus by aseries
of logical operations, forms can be generated; these forms may
becOI1J.Pared to empirical forms of social systems, and where there
iscorrespondence in fonnal features between the two, the empirical
fornlmay then be characterized as a particular constellation of lhe
variablesin the model.
In these respects the models which I discuss are similar to
thosealready in usc in certain fields of anthropology, notably the
cornponential models of kinship systems (Goodenough, 1956) and
Leach's
First published by the Royal AnUHopological tnstitute of Gre'l
Britain andIreland, 1966.
32
-
Models of social organization I 33topological models (Leach.
1961). "Where I believe I differ fr"11\ these.but follow the
example of varillus other dis"plines (llleo), "I'
Gamt's.evolutionary genetics. et,.) is in adding a further. ;;ery
imp"rtantlimitation: the logical operations whereby fmms are
generated sh'Htidmirror actual, empirical processes whkh ,an be
identified in the realitywhich is being analysed. This limitation
is necessary to mah theoperations significant (cf. eumann and
Morgenstern 19-1-1). allli itmakes the modelbuilding activity
subservient to the objel'tives llf anempirical science.
Very brieny, I hold that such models can have three important
uses:I They provide a kind of understanding and explanati'Hl which
amodel of form. however meticulous and adequate. can newr give.
Tostudy form it may be sufficient to describe it. To expbin form
oneneeds to discover and describe the processes that gent'rale the
form.2 11ley provide the means 10 describe and study change in
sodal formsas changes in the basic variables that generate the
forms.3 Finally, they facilitate comparative analysis as a
meth"dol,'gicalequivalent of experiment. Models descriptive of form
merely I",nnitone to layout typological series and point to
differences and similari-ties, or to specify the logicaLly
unrestricting transformations whcrebyone fonn may be converted into
another. The adequacy of:1 gcnerativemodel. on the other hand. is
tested by its success or failure injtencGlt-ing the observed forms:
it contains implicit hypothese about 'possible'and 'i,;PoSsiblc'
syslems which may be falsified by comparative data.
11le level of complexity and sophistication reached in these
essays isvery low. However, I believe that the study of social
anthropology cannot today be advanced much by sophistication and
refincment of itscurrent lotal stock of concepts and ideas. Rather.
we should make acareful selection among them, and among concepts
available in relatedfields, to isolate a minimal set which is
logically necessary and empiri-cally defensible. The implications
of any such SCi should then be ex-plored and exhausted before
further complexity is added. The follow-ing is presented as a first
sketch of one such set of concepts, and apreliminary exploration of
some of its implications.
The analytical importance of transaction
The validity and utility of the following discussion depends on
a par-ticular view of the constitution of social phenomena. Though
thisview is neither unorthodox nor susceptible to serious
challenge, it is
-
34 Models ofsocial organization Iimportant to establish it
explicitly. For this purpose, one of Radcliffe-Brown's last and
perhaps most authoritative statements may be cited:
the concrete reality with which the social anthropologist is
con-cerned in observation, description, comparison and
classification, isnot any sort of entity, but a process, the
process of social life. Theunit of investigation is the social life
of some partieur"a-r-re-gJ7"on of theearth during a certain period
of time. The process itself consists ofan immense multitude of
actions and interrelations of human beings,acting as individuals or
in combinations or groups. Amidst thediversity of the particular
events there are discoverable regularities,so that it is possible
to give statements or descriptions of certaingeneral features of
the social life of a selected region. A statement ofsuch
significant general features of the process of social life
con-stitutes a description of what may be called a[orm
o[socigIJife. Myconception of social anthropology is as the
comparative theoreticalstudy of forms of social life amongst
primitive peoples. (Radcliffe-Brown, 1952: 3-4)
The' neral features', the regularities in social life, thus have
to do withthe repetitive nature of acts, which we observe in our
investigation of asocial system. Our first descriptive
characterization of our fmdings,then, must be one involving
frequencies. No matter how crude ourtechniques for registration and
counting are, our claim must be that we
1""1 .... t ('J t') l'have discovered some non-ran om frequency
distribution in actions.The patterns we report may most
realistically be viewed as frequencydistributions around a mode. My
argument in the following pages fol-lows from this, and is briefly
summed up in the simple statement thatour theoretical models should
be designed to explain how the...obsefY-able frequency patterns or
re laritie enerated.
TIle most simple and general model available to us is one of an
aggre-gate of people exercising choice while influenced by certain
constraintsand incentives. In sueh situations, statistical
regularities are produced,yet there is no'absolute compulsion or
mechanical necessity connectingthe determining factors with the
resultant patterns; the connectiondepends on human dispositions to
evaluate and anticipate. Nor can thebehaviour of anyone particular
person be firmly predicted - suchhuman conditions as
inattentiveness, stupidity or contrariness will,for the
anthropologist's purposes, be unpredictably distributed in
thepopulation. This is also how we subjectively seem to experience
our
-
Models ofsocial organization 1 35own social situation. Indeed,
once one admits lhal what we empiri-cally observe is nOI 'customs',
but 'cases' of human behaviour. it seems10 me that we cannot escape
lhe concept uf choia in ,'ur analysis: "urcentral problem becomes
what are the constraints and incentives thatcanalize choices.
The sttucturalist's view has been. as I have understood it. that
Iheseconstraints on choices are moral: society is a moral system.
nlis viewleads to a type of analysis where regularities in the
patlern of bdlaviourare related to a set of moral constraints and
incentives which stipulatethe critical features of lhat regularity.
Thus for example the regul3ri-ties summarized in a status position
are specified as a series of rightsand obligations which summarize
all lhe regular aspects of behaviourwhich are associated wilh lhat
status.
By lhis transformation, one form, in the sense of a set of
regularpatterns of behaviour, is translated into anolher. virtually
congruentform, made up of moral injunctions, which are made
logically prior tobehaviour. The model does not depict any
intervening social processbetween lhe moral injunction and the
patlern. nlere is indeed noscience of social life in lhis
procedure, no explanation of how actualforms, much less frequency
distributions in behaviour come about, be-yond lhe axiomatic: what
people do is influenced by moral injunctions.
If a concept of process is to be analytically useful, it must
refer tosomething Illat governs and affects activity, something
that restrictsand canalizes the possible course of events. nlese
restrictions shouldgo beyond what can be contained in static or
general kinds of limita-tions, Just as lhe description of a game of
cricket is more than a de-scription of its binding rules, so a
description of a process of inter-action should contain more lhan a
listing of reciprocal obligations.The study of process must be a
study of necess,ary or probable inter-dependencies which govern the
course of events. We have recently seenhow the variety of forms of
domestic unit in a society can be under-stood by a view of family
development cycles (Goody, 1958). Thegeneral lesson we may learn is
that by a simple analysis of a process wecan understand lhe variety
of complex forms which it produces.
This is the kind of understanding and explanation of form
whichs1lOuld be sought. Explanation is not achieved by a
description of thepatterns of regularity, no matter how meticulous
and adequate, nor byreplacing lhis description by olher
abstractions congruent with it, butby exhibiting what
~ke_!.Jh~...IDl11ern, i.e. certain processes. To studysocial forms,
it is certainly necessary but hardly sufficient to be able
-
36 Models ofsocial organization Ito describe them. To give an
explanation of social forms, it is suf-ficient to describe the
processes that generate the form.
In the following I wish to explore the extent to which patterns
ofsocial form can be explained if we assume that they are the
cumulativeresult of a number of separate choices and decisions made
by peopleacting vis-a-vis one another. In other words, that the
patterns are gener-ated through processes of interaction and in
their form renect the con-straints and incentives under which
people act. I hold that this trans-formation from constraints and
incentives to freqoentative patterns ofbehaviour in a population is
complex but has a structure of its own,and that by an understanding
of it we shall be able to explain numerousfeatures of ,ocial form.
Indeed, as I interpret the above citation fromRadcliffe-Brown, the
prucesses which effect that transformation areour main field of
study as social an th ropologists.
As the object uf systematic investigation, these processes have
beensorely neglected. Even the work which has been heralded as a
prototypeof prucessual analysis - that of the development cycle in
domesticgroups - focuses on processes which arise not from social
facts, butfrom the extraneous biological fact of creeping
senescence.
The consetjuences of having overlooked generative problems are
per-haps most strikingly revealed in- uOnost-used concept
of,Istatus, Intro-duced into the field in a pa' with role, the two
concepts togetheretjuip us to analyse the fundamental social
process whereby bindingrights and obligations are made relevant in
particular social situations.Instead anthropologists chose to
concentrate on the static concept ofstatus, and the analytic
poSsibilities of the paired concepts have onlyrecently been
explored by Goffman (1959). Goffman argues thatagreement on a
definition of the situation must -beestablished and-maintained to
distinguish which of the participants' many statusesshould form the
basis for their interaction. The process of maintainingthis
agreement is one of skewed communication: over-communicatingthat
whidl confirms the relevant status positions and relationships,and
IIl1der-eommunicating that which is discrepant. TIle effect is
togenera te stereotyped forms of behaviour in these situations -
behaviourwhich is not specified in the rights and obligations
comprising thestatus, but which emerges as regular features of the
role - becauseof tJlese situationally determined requirements of
over- and under-comll1unication which he aptly calls 'impression
management'.
We may distinguish two types of problems to which such viewsmay
be applied. One concerns the way in which a person completes/
- Models ofsocial organization I 37consummates a successful role
performance by sele.:ting frl)m his tL)t:l1repertoire those
gestures and idioms whkh will servL' his needs fl'r'impression
management'. TIl is is where Goffman L"lllKentr:ltes hisdiscussion.
TI,e other type of problem. with which WL' arc mainly .:"ncerned in
the present connection. is th:1t of institutilmali/ati"n: Iww
amultiplicity of individual decisions under thl' influenL"l' "I'
l":lnaIiLingfactors can have the cumulative effect of produdng
dl'ar p:ltlL'rnS :lndconventions. Goffman's arguments :Ire also
:lpplicable here. Thl' s:lmcproblems of impression management arise
for all inL"umbents "I' :1status. The punishments and rew:lrds of
varying degrees l,f snccess willmake a majority modify tJleir
perfl1nnanee in the dirl'eti"n Ill' theoptimum: and tJle n1l1re a
cenain Ivpe ,)f helwvil'lIT is statistie:lllyassociated with a
status. Ihe mOTe it WIll he rl'inf,'fL"l'd thfllugh s,'r\"ingas an
idiom of iuentific:ltion. We ma) Ihus cl'llStrUl"1 a mLldel
whl'Tebycomplex and comprehensive p:ltterns "f h
-
38 Models ofsocial organization Ihehaviour of the other in a
progresssional sequence.
The ge.!lera1 notion of reci rocity is of course old and
familiar inanthropology; indeed, it seems to be fundamental to our
view of socialrelationships. Though its meaning is not frequently
made explicit, !should think few will quarrel with one of Leach's
formulations: 'In anysuch system of reciprocities one must assume
that, overall, both parties... are satisfied with their bargain,
and therefore that the exchangeaccount "balances".' (Leach, 1952:
51.) Yet the analytical status of'reciprocity' in social theory is
far from clear, but is capable of beingdeveloped in several ways
(see esp. Homans, 1958; Stanner, 1959). Inthe present context, it
lies at the heart of an analytic concept of trans-action: one may
caIJ transactions those sequences of interaction whichare systema
tically governed by reciprocity.
Each and every case of interaction does not have these
characteris-tics. It is possible to define a relationship of
incorporation as an analyti-cal opposition to a transactional
relationship - one where a value opti-mum, probably for a
restricted range of values, is sought for the sum ofpartners, and
not for a single party. Such partnerships can engage asunits in
transactions vis-a-vis other persons or groups; internally,
theirinteractions will not be systematicalJy governed by
reciprocity. Yetthere are limits in most cases to the losses, or
inequalities of gains,which people are willing to bear through such
incorporation, and theserestrictions on the constitution of
relationships of incorporation willbe developed and utilized below
(pp. 63-75). FinaIJy, we have all of usseen cases of what we have
interpreted as altruism - a kind of funda-mental negation of the
transactional relationship. But transactions wehave also seen, and
a clear concept of transaction, leads us to a recognition of a very
fundamental social process: the process which resultswhere parties
in the course of their interactions systematicalJy try toassure
that the value gained for them is greater or equal to the
valuelost.
Put this way, one may see that transactions have a structure
whichpermits analysis by means of a strategic model, as a game of
strategy.They cunsist uf a sequence of reciprocal prestations,
which representsuccessive moves in the game. There must be a ledger
kept of valuegaineu and lost; and each successive action or move
affects that ledger.changes the strategic situation, and thus
canalizes subsequent choices.Many pussible cuurses of action are
ruled out because they are patentlyunsatisfactury. i.e. an actor
must expect that value lost be greater thanvalue gaineu. In such a
model the incentives and constraints on choice
-
Models ofsocial organization I 39are effective through the way
they detennine what ,'an he gained andlost; and each actor's social
adjustment to the other party in the trans-action is depicted in
tenns of alter's possible moves. and how they intum affect ego's
value gains. The structure depicted in this model isa successional
one over tin1e - in other words. it is a model of preess.
The in1portance of process and the analysis of choice has been
givenincreased recognition in anthropological writings.
particularly throughFirth's important formulation of the
distinction between structure andorganization (Firth, 1954). I have
elsewhere used some of the formalapparatus of the Theory of Games
to analyse political choice (Barth.1959b). The particular formalism
of the Theory of Games is not asin1portant for anthropological
purposes as is the theory's fundamentalcharacter as a generative
model. It can serve as a prototype for a proces-sual model of
interaction: and in concentrating on transact;o!l as theanalytic
isola Ie in the field of social organization, I am emulating whatI
regard as the most crucial aspect of the theory for our purposes.
Whatis useful in Ihis view of transactions is that it gives us a
logically consis-tent model of an observable social process. It is
a model whereby onemay generate forms according to the rules of
strategy. given the para'meters of value; and these fom1s generated
by the model may then becompared to the empiJical patlerns which
one has observed. The logicalanalysis can be rigorously separated
from the cumulative presentationof data: and the adequacy of such a
model can in each case be judgedby the degree of fit between the
patterns which are logically generatedand the patlerns which are
observed. We are not committed by anyprejudged 'view of society' -
the adequacy of the transactional modelfor any and every particular
relationship is continually on trial. Andsince the model c1ain1s to
depict actual empirical processes, all its partsand its operations
- its exchanges. its value parameters, etc. - may bequestioned and
checked.
A measure of the analytic importance of such a concept of
transac-tion is the fact that it is implicit in our whole .idea of
values. It ismeaningless to say thaI something has value unless
people in real lifeseek it, prefer it to something of less value,
in other words maximizevalue. This can only be true if they usually
act stralegically with respectto it, that is, make it the object of
transactions between themselves andothers.
I should emphasize that this does not imply a claim for
explanatoryadequacy for all aspects of all behaviour. 0 doubt man
has a psychological constitution which affects the way he hehaves.
But as a social
-
40 Models ofsocial organization Ianthropologist I am concerned
with exploring the effects of socialdeterminants on human
behaviour, and these vary with reference tosuch factors as value
and strategy. Furthermore, in real life, men alsoenjoy and consume
value, and act with indifference to it. What I amsuggesting is that
transactions are of particular analytical importancebecause (a)
where systems of evaluation (values) are maintained, trans-actions
must be a predominant form of interaction; (b) in lhem therelative
evaluations in a culture are revealed; and (c) they are a
basicsocial process by means of which we can explain how a variety
of socialforms are generated from a much simpler set and
distribution of basicvalues.
The first two of these themes will be explored further in the
nextchapter; the third I shall now scek to illustrate. The major
elements ofthe generative model called for in such an illustration
have been pre-sen ted. Transactional behaviour takes place with
reference to a set ofvalues which serve as generalized incentives
and constraints on choice; italso takes place with reference to a
pre-established matrix of statuses,seen as a distribution of values
on positions in the form of minimal clus-ters of jurally binding
rights. From this point, through the formation ofstatus scts, and
the implications and restrictions of transactional rela-tions and
impression management within