You are on page 1of 15

Analytical Dimensions in the Comparison of Social Organizations

Author(s): Fredrik Barth


Source: American Anthropologist, New Series, Vol. 74, No. 1/2 (Feb. - Apr., 1972), pp. 207-
220
Published by: Wiley on behalf of the American Anthropological Association
Stable URL: http://www.jstor.org/stable/672142
Accessed: 17-02-2017 14:33 UTC

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted
digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about
JSTOR, please contact support@jstor.org.

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
http://about.jstor.org/terms

American Anthropological Association, Wiley are collaborating with JSTOR to digitize, preserve and
extend access to American Anthropologist

This content downloaded from 193.48.45.204 on Fri, 17 Feb 2017 14:33:34 UTC
All use subject to http://about.jstor.org/terms
Analytical Dimensions in the Comparison of Social Organizations

FREDRIK BARTH
University of Bergen

The paper focuses on differences in the way social situations are defined in
different cultures and societies. It develops and illustrates some dimensions for
comparing social systems in this respect, viz: their inventory of statuses, the
repertoires of persons, and the cultural ways in which occasions are distinguished.
Some contrasting types of society are compared on these dimensions; and the
problem is raised as to the epistemological character of such types, and the
importance of the variable of scale. The dual nature of social organization as an
organization both of people and of tasks is emphasized.

A MAJOR AIM of social anthropology is organizational apparatus. If we seek to mak


to provide a comparative perspective on typologies of the ways of constructing
social organization, particularly one that societies, rather than the overt structures of
embraces all varieties of society from thesocieties, we escape at least provisionally the
simplest to the most complex. Such com- vexing problems of delimitation and scale in
parisons are fruitful because they lead to the comparison; we no longer compare cases
empirical generalizations, they expose of society but kinds of social systems. In the
analytical problems, and they allow for thelatter part of this essay I shall return to the
falsification of hypotheses. To perform suchconnection between membership size and
comparisons, however, one needs to developthe forms of organization, but the focus on
explicit typologies of social systems. Currentencounters makes it possible to postpone
trends in the analysis of society and culturethis discussion. We also escape at least some
would indicate the need for a typology of the pitfalls of comparative method ex-
posed by Leach in his contrast between
which takes account of differences in the
organization of encounters and allowscomparison
for and generalization (Leach
1961:2ff.), since we will mainly be attempt-
the exploration of interconnections between
ing to generalize some dimensions for com-
aggregate social forms and the construction
of social persons on the micro-level. paring social processes, not to construct a
In the following I shall discuss a few
taxonomy of institutional forms. To con-
major dimensions of social organization struct
cap- my typology, in other words, I seek
able of generating a fairly comprehensive
generalizable features-the kind of abstrac-
typology of social systems. This I shall tions that led to T6nnies' distinction be-
illustrate with some examples and sub- tween Gemeinschaft/Gesellschaft or led
stantive discussions; but the focus of the anthropologists to develop concepts like
paper is on the clarification of analytical substitutability (Fortes 1953) or multi-
dimensions, not on the elaboration of a plexity (Gluckman 1955) to characterize a
taxonomy. certain type of society.
The comparative analysis developed here A typology capable of being applied to
thus does not proceed from a macro-view of anthropological material clearly must be
societies through a comparison of sub- compatible with basic analytical concepts in
stantive institutional structures. Rather, it the discipline, particularly the concepts of
seeks to penetrate down to modes of inter- status and group. The analytical discussion
action in encounters, groups, and com- in this paper will expose certain basic dif-
munities and to see if we can characterize ficulties in their use; but provisionally we
how social systems differ in their basic may regard the concepts as unproblematical.

207

This content downloaded from 193.48.45.204 on Fri, 17 Feb 2017 14:33:34 UTC
All use subject to http://about.jstor.org/terms
208 AMERICAN ANTHROPOLOGIST [74,1972

Their still unworked potential are peculiar has


to eachrecently
case and cross-culturally
been demonstrated (Goodenough 1965; highly variable. The events through which
Mayer 1966). That this potential can be statuses, relations, and groups are made
increased for typological purposes by a manifest also have their form determined by
consistent emphasis on their system-oriented the actors' codifications of tasks and
rather than ego-oriented application was occasions-or, as Nadel puts it: social stru
exemplified by Nadel (1957 Ch. IV) in his tures have jobs to do. The variety of
exploration of "The Coherence of Role circumstances and purposes, of "jobs" can-
Systems." Nadel seems to have articulated nota profitably be left out of the formal
general perspective, and been followed by theory and then smuggled back in through
many, in seeing social systems as basicallyconcrete typologies, as, for example, by
constituted of a great number of abstrac- classifying corporate groups into subtypes
table elementary units of ascription which according to the kind of activity they serve
to organize. We should rather from the
we, pace Nadel, may call statuses.1 It has
outset give equal primacy to each of the
long been recognized (cf. Radcliffe-Brown
major axes of organization: the structures
1958) that such statuses always cohere in
structured contexts which may be seen asof statuses, and the delimitation of jobs.
relations or persons or corporate groups To encompass these twin aspects of social
according to the aspect of social reality on organization in one frame of reference, we
which one choses to focus. Indeed, I should can utilize the perspective so vividly pre-
like to add a fourth "natural" clustering of sented by Goffman (1959, 1961). For our
statuses in careers, to which we may have gross typological purposes, it is sufficient to
devoted too little attention apart from their emphasize the concept of "definition of the
standardized "life cycle" form. situation." In arriving at a definition of the
Most attempts at developing typologies situation, actors reach certain agreements
on this basis seem to have concentrated on about the rules of relevance in a particular
corporate groups, and then quickly to have encounter, both with regard to what are the
foundered on the great empirical variety of
relevant capacities of the participants and
such groups (see, e.g., International Ency- what is the "job" on hand, i.e., the occasion.
clopedia of Social Sciences: Political Anthro-This perspective opens very attractive pos-
pology: Political Organization). Constructing sibilities for developing analytical concepts
a typology of groups seems to turn into a as "scene" and "setting," and typo-
such
race with your own imagination, an im- logical concepts such as "forum" and
provisation of subtypes and additional types
"arena" (e.g., Bailey 1969); but these have
based on criteria fetched from outside a still been very little exploited.
theory of social systems and descriptiveThe of important thing in the present con-
the hopeless variety of habitat and circum-text, however, is not to develop a sub-
stance documented by the ethnographic stantive classification of activities into tasks,
record. institutions, or types of occasions, but to use
I would submit that these attempts fail the concept to make a first step at character-
because they fail to abstract one of the izing the way in which members of a society
major dimensions of social organization: organize their activity through the definition
that of tasks and occasions. To subdivide of situations. Thereby, we capture both the
and classify the flow of social eventscrucial
it is aspects of social organization, viz.
always necessary to specify not onlyboth
per-the social structures and the jobs they
sons, relations, and groups, but also thedo.rules
of relevance which actors impose on the I shall attempt to substantiate this claim
situation. Though these often arise from in the following paragraph. Very simply, we
necessities of a self-evident and common can visualize any society of which we are
members as follows. Each of us is a com-
human kind, they are codified in ways which

This content downloaded from 193.48.45.204 on Fri, 17 Feb 2017 14:33:34 UTC
All use subject to http://about.jstor.org/terms
Barth ] COMPARISON AND ANALYSIS 209

pound person, the encumbent choice are not satisfactorilyof many


explained by the
demonstration of the presence of some
statuses. When we come into each other's
presence we do so in a physical objective circumstances alone, but require an
environment-one which we perceive account of how these circumstances are
perceived and evaluated by actors. Their
selectively and classify culturally as a pot-
specific effects on social organization
ential scene for certain, and only certain,
depend on the way in which participants in
kinds of activities. We add to these con-
straints, or modify them, by communicatingencounters accommodate themselves to such
with each other as to who we are and what circumstances by taking them into account;
we intend to do, and thereby we arrive at thus
an they are present as part of the
"micro "-events of an encounter, as con-
agreed definition of the situation, which
straints on behavior, modified through the
implies which status out of our total reper-
actors' definition of situations and con-
toire we shall regard as relevant and to what
use we shall put it. The agreement will besiderations of choice. Only by showing how
workable only if all participants have a these codifications and evaluations are
status in their repertoire which articulates stereotyped and shared do I feel we ha
with those of the others and are willing to explained regularities in social behavior, i.
act in this capacity. A definition of the aggregate features of society.
situation thus implies the mobilization, as The feedback on societal macro-features
relevant and acceptable, of a set of articu- which arises from the objective con-
lating statuses. Through such under- sequences of behavior, e.g., on the dis-
standings, social statuses are mobilized andtribution of assets or facilities for control, is
activity ordered in the manner we can likewise neither explained by a denial of the
describe as social organization. Behind this objective consequences of acts nor by
creation of organized encounters, we can imputing purpose to all the unsought con-
identify the interests and goals that set social sequences of acts. An analysis requires the
life in motion: we can recognize social demonstration, for each connection claimed,
statuses as assets, and situations as occasions of the particular mechanisms whereby such
for realizing them by enactment. In part, consequences follow.
people will seek out the partners and places I thus see encounters, constrained by
that provide occasions for achieving this; circumstances
in and structured by common
part, they will merely find themselves in the understandings between the participants, as
proximity of others who call on them to the stuff of society. If you grant me that this
take social cognizance of each other. is one way of looking at social life, my
This implies a particular view of the question will be: Are the people in different
relation between the micro-level of separate
societies equipped with structurally similar
social encounters and the macro-level of arrays of statuses, alters, scenes, and oc-
societal form, one which it might be well to through which they can define
casions
make clear. I am not propounding a sub- situations and structure interaction
jectivist viewpoint which denies the sequences, thereby generating the re
objective consequences of social acts or the
ities in their encounters? Comparative
existence of objective social and ecological
material clearly provides no as an answer.
macro-features which operate as constraints
Societies differ, as I shall try to show, in
on behavior. What I am denying is the fundamental ways with respect to the
mechanical determination of behavior by elements of organization which they provide
these constraints. Human acts are pre- their members and by which these members
dominantly shaped by cognition and pur-create order in their social lives. And they
pose, asserted through awareness and differ not only in their status inventories,
voluntary behavior, i.e., through decision but in ways these statuses can be combined
and choice. Regularities in multiple cases of in persons and elicited by alters and the

This content downloaded from 193.48.45.204 on Fri, 17 Feb 2017 14:33:34 UTC
All use subject to http://about.jstor.org/terms
210 AMERICAN ANTHROPOLOGIST [74,1972

ways distinctive social occasions can be so that being mother to a child you are also
created. These differences would seem to child to another mother, whereas being
provide a significant and fundamental doctor
basis to a nurse you are not also nurse to
another doctor.
for classifying and understanding different
social systems.
To construct more exact dimensions for
In other words, the concept of "status
sets" expounded here is not sufficient to
such a typology we need to look more
describe the structure of social systems.
closely at what I have referred to as status
Such systems also exhibit fundamental dif-
sets. The concept is an expansion of the
ferences in the ways of constructing
dyadic social relationship to encompass any
"persons," or relevant sectors of persons. In
natural cluster of two or more reciprocally
the sphere of kinship each person is so
relevant statuses, exemplified by mother-
composed as to produce, in conjunction
child, or doctor-nurse-patient. It is meaning-
with others, a particular kind of larger
ful, indeed possible, to differentiate status
system which we may call "replicating": if a
sets only if actors themselves distinguish
person is A to B, he is frequently also B to
between social situations that systematically
another A, so most persons are indeed A +
elicit them as different sets, i.e., if alter-
B. Kinship sets internally also show another
native capacities exist for some of the
special feature: any status can only be made
persons which are elicited in other situa-
relevant toward a particular kind of other-
tions. Otherwise, all of the society would
they can be depicted as dyads, and, given
consist of only one set, making the term
any particular interaction partner, there is
redundant.
no freedom to manipulate this component
Let us compare the two examples of
of the definition of the situation, only the
status sets mentioned to note some di-
task at hand remains open to negotiation
mensions of structural contrast. Mother is
and agreement. The hospital set, on the
mother to (her own) children only; she is
other hand, does not presume any other
called so only by them and acts as mother
component of the definition of the situa-
only toward them. Nurse is nurse both to tion: both the occasion and the relevant
doctor and to patient, and must act very
alters must be specified before predictabl
differently toward the two. Hers is a status
behavior will be generated.
with several roles, as first clarified by Merton
(1957).' Which of these sectors of her rights These two systems imply very differen
and duties as nurse will be activated mainly organizations of activity. In the kinshi
depends upon which alter in the total set her system, most persons will perform mo
behavior is directed. tasks, but toward different alters. The ad
Yet, being mother also in fact implies ditional situational components determini
relations to others, just as being nurse does;what mother will do are in a sense seconda
by virture of motherhood a person should and are connected both with the scene and
also be a party to relationships to husband with what is instrumental in an environment.
and father-in-law and is liable to become Nor can mother repudiate the relationship:
mother-in-law and grandmother. Do all the these
rights and obligations of the status set
statuses together constitute the set in are thealways entailed in the interaction be-
analytical sense? Are we just duped by a tween these persons. Given the appropriate
different convention of naming, where in alter, there is thus really no alternative status
our conventions of kinship terminology eachthan can be made relevant, and consequently
role sector is given a separate term? Not no social interaction is necessary for the
quite, because, having indicated that the purpose of defining gross features of the
statuses involved are kinship statuses, we situation. The fact that the main situation
have implied a special kind of linkage variable--the available alter-is incorporated
whereby a number of statuses are connected in the status definition leads to the sub-

This content downloaded from 193.48.45.204 on Fri, 17 Feb 2017 14:33:34 UTC
All use subject to http://about.jstor.org/terms
Barth ] COMPARISON AND ANALYSIS 211

stitutability and multiplicity agreements are noted


made about inrelevant
such
systems. status sets, occasions, and tasks.
The structure of the hospital set, on the A short comparative series of thumbnail
other hand, implies the existence of alter- sketches of empirical types may demonstrate
native sets and the possibility of interaction the features brought out by these
in terms of such sets; consequently, activity dimensions. I appeal to the reader's know-
does not become organized until a gross ledge of the relevant ethnographies to
definition of the situation, indicating the elaborate and criticize the characterizations:
relevance of a particular (hospital) status set, they are included here merely to provide
has been reached. In systems of this kind, concrete referents for the typological terms.
activity is organized so that any one person I. Shoshonean or Bushman bands: (1)
performs relatively few tasks, but does so small inventory of statuses, (2) very small
toward many. The limiting case is the person individual repertoires, (3) environment-
who performs a unique combination of tasks oriented task organization. In these
and does so for all members of the society, elementary social systems, each person
being singularly unsubstitutable. This of would seem to occupy a relatively indivisible
course is the aspect of organization position. There is thus no clear separation
emphasized by the concept of organic solid- between a man's capacity as dominant adult
arity. A wide range of implications of the male/husband and father/hunter/etc., and in
contrast between these two kinds of system all these respects he contrasts with a child or
has been explicated by Levi-Strauss (1963), a married woman. Though such specially
suggesting that we are indeed dealing with codified capacities change in a regular career
two basically alternative and different pattern through a life cycle, they are at any
societal types. one time unambiguously distributed on the
Three crucial dimensions which depict local population, and interaction is pre-
the organizational bases for such differences dominantly organized by one dominant
between social systems can be formulated as status set. Situations differ from each other
follows. A person, as a combination of mainly with reference to the definition of
statuses, can be said to have a certain the task at hand, and this derives promi-
"repertoire" of statuses; different situational
nently from the opportunities offered by the
definitions bring out different components immediate environment.
of this repertoire. Further, a total societyII. Australian societies: (1) larger in-
may be said to be composed of a total ventory of statuses, (2) larger but highly
"inventory" of statuses--partly intercon- repetitive individual repertoires, (3) partner-
nected in sets, partly co-existing as alter-
specific elicitation of capacities. These
native sets. Finally, "social situations" arereplicating systems show considerably
conceived in different societies as differently
greater complexity than Type I, without,
constituted and contrasted. The dimensions incidentally, being associated with more
I propose for classification-not quantifiablecomplex productive technology. Character-
in numbers but clearly measurable as con- istically a person is the encumbent of status
trasting orders of magnitude-are thus: A vis-a-vis person one, of status B vis-a-vis
(1) The "inventory" of statuses in a
person two, etc.; those persons may turn to
social system: the total number of him dif-whenever they want prestations of kinds
ferentiated statuses known to the actors. appropriate to statuses A and B respectively.
(2) The "repertoires" of persons: the With classificatory ascription of kin statuses,
sizes and structures of the standardized marriage-class organization, etc., the life
clusters of statuses that make up personscycle does not involve drastic changes in the
in the society. constitution of the persons. Situations differ
(3) The components used in the pre-eminently in terms of which personnel is
actors' "definition of situations": what present; therefore activities are organized

This content downloaded from 193.48.45.204 on Fri, 17 Feb 2017 14:33:34 UTC
All use subject to http://about.jstor.org/terms
212 AMERICAN ANTHROPOLOGIST [74,1972

through the process ofduties seeking


of father out specific
will take place: father is one
partners, as well as by the
status. Butdefinition ofone such
it is possible to elicit
tasks. sector of "toddy-tapper," e.g., liquor seller,
Let me contrast these and twoallow types of simple
only that status be relevant to
societies with two types of complex interaction without becoming involved with
societies, using the same dimensions: the other statuses of toddy-tapper. In other
III. Indian villages and regions: (1) large
words, toddy-tapper is a standard repertoire
inventory of statuses, (2) large repertoires,of statuses which may be activated severally
which tend to be found only in a limited and singly in distinct social situations and
range of standardized constellations, (3) relationships. The fact that these statuses in
elaborate structures of occasions and tasks, Indian society tend to cluster in stereotyped
signaled by complex idioms widely mani- repertoires has important implications for
pulated to define situations. These involute the form which interaction may take: it
systems have occasioned considerable becomes possible to treat whole persons as
theoretical discussion, where also an attemptstandardized transactional partners; indeed,
of mine to account for them in general terms some parts of Indian social organization,
related to those exercised here (Barth such in as, for example, the jajmani systems,
Leach 1960) has been severely criticized are based on this fact. But the recognitionl
(Dumont, in Reuck 1967). The issue, as I thatsee this is not necessary, that the Indian
it, lies in what has been referred to as the system is one based on large repertoires and
"summation of statuses." I find it necessary not on a few wide and undifferentiated
statuses, is important because it leads one
to distinguish the versatility of a person that
comes about by his having a large, diverse pose two central questions that would
repertoire of statuses, from the versatility otherwise be meaningless: What maintains
that comes about by his occupying a status the stereotyped repertoires? What are the
with a large diversity of rights and obliga- aggregate consequences of those processes,
tions. My argument, now and previously, andis of the presence of non-stereotyped
deviance?
that these should be distinguished because
they come about in very different ways and We know that the stereotyped repertoires
exist in societies with very different con- are maintained by persons' own efforts and
stitutions and properties, although, from thea sanctions of others and that they depend
certain perspective, they may appear on a hierarchical concept of congruence
phenotypically alike. within repertoires. A single polluting status
No one seems to disagree with the gross in a cluster has a contagion effect on the
generalization that subcaste membership person as a whole, while a few more highly
tends to imply standard constellations of ranked statuses and capabilities give limited
capacities, i.e., standardized social persons. benefits when associated with the others.
These component capacities of a person can These premises have enormous implications
be distinguished as separate statuses only if for the social process, and thereby organize
they can be shown to belong in different activity in a characteristically Indian way:
status sets. This requires that actors have persons both seek, and avoid, occasions with
ways of defining a social situation so that elaborate care; they pursue a sensitive
only one sector of a person is made relevant, husbandry of social assets that channels
that this sector articulates with a limited community life and sustains complex
sector/status in an alter, and that the two cultural
or codifications for defining social
more parties need not involve themselves in
situations.
interaction in other sectors. It is not possible The fact that people are capable of
in this way to elicit one such sector of entering into limited interaction in a single
"father" and define situations so that no status set, on the other hand, leaves a way
other interaction involving other rightsopen and for persons to try to construct alter-

This content downloaded from 193.48.45.204 on Fri, 17 Feb 2017 14:33:34 UTC
All use subject to http://about.jstor.org/terms
Barth ] COMPARISON AND ANALYSIS 213

native repertoires, and organization


through has been characterized
them to as an
open network and been associated with
achieve cumulative results for themselves
tendencies to subjective isolation and
while at the same time creating new patterns
in the society. Thus while a career for most
alienation. (b) Perhaps more characteristic of
the life situation of many members of such
persons involves a rather limited and predict-
societies is the emergence of cliques and
able progression through successive life cycle
classes
phases, for some persons it constitutes a through a process whereby persons
tactically ordered, innovative pattern withwith partly similar repertoires seek each
potential consequences for community other out as interaction partners. Such
organization. It is thus possible for the partners will find that they potentially arti-
goldsmith to be a moneylender, the toddy- culate in terms of a number of different
tapper to be a liquor salesman, the toddy- status sets, and they seem to create op-
tapper (to his own and his customers' portunities for active switching between
mutual advantage) to double as a money- situation definitions rather than distinct fora
lender until, one day, the toddy-tapper is a for each separate set. Yet, these sets and
landowner. The local diversity and regional statuses retain some definitional distinctive-
complexity of Indian society is usefully seen ness, since two persons' repertoires never
in this perspective: not as an ultimate completely replicate each other, circles only
"explanation" of the creativity of the Indian partly overlap, and formal occasions with
tradition, but as a picture of the social heavy constraints on switching are also
maintained. The network pattern that results
organization through which this tradition is
consummated in behavior and thereby is more highly connected than in (a), but in
sustained and elaborated. my view the more significant aspect is the
activity of swift and transient definition and
IV. Modern Western societies: (1) very
large inventory of statuses, (2) very large redefinition of situations in encounters with
individual repertoires (but each much the same personnel creating the many-
smaller than the total inventory), highly stranded informalized relationships of
diverse in their constellations, (3) swift andfriendship. These again, as in the discussion
often transient definition of situation, greatof caste, need to be categorically dis-
variation in degree of formalization between tinguished from the many-stranded relations
situations. A label for such complex though between members of simple societies, be-
partially familiar societies is difficult to find;cause they are a part of very different total
but, in view of the great range of op- systems and are maintained by very different
portunities they offer for diverse and social processes, though showing clear
voluntary articulation with a very large range features of overt similarity.
of potential alters, we might characterize With respect to career forms, contract
them on the basis of their most explicit and societies are characterized by variable and
formalized kind of individual agreements as complex career structures in which
contract societies. sequences of statuses are assumed and shed
As to the social organization that by persons in patterns highly significant to
characterizes them, different potentialitiesthe
in organization of society.
these organizational elements are empha-
sized, in fact or description, in different
accounts: (a) In an impersonal, urban en-
vironment the parties to different kinds of The thrust of this way of depicting
interaction are segregated, i.e., each status society is generative, seeking to show how
set involves a different personnel and the more complex features are aggregated from
person moves in social space, articulating simple elements. It further assumes that this
only in one limited capacity with any aggregation is, in real life, effected by the
particular alter. The result of this kind of process of social interaction. The transfor-

This content downloaded from 193.48.45.204 on Fri, 17 Feb 2017 14:33:34 UTC
All use subject to http://about.jstor.org/terms
214 AMERICAN ANTHROPOLOGIST [74,1972

mation rules of the model,


The poverty leading from
of components in the
elements to aggregate social
definition forms,
of situations must
in societies of types I
therefore be such as rest onconnect
and II I would concepts like
with the totemistic
strategy and resource management, choice
use of taboo as a major idiom for defining
and opportunity. These
socialparts of the
persons. While argu-
taboo is an apt
ment I have allowed myself to treat mechanism for using distinctions in nature
sketchily and implicitly, since they have to create and define social distinctions, it
been expounded at greater length elsewhere simultaneously implies a denial that some of
(Barth 1966). I have rather concentrated on a person's identities can be latent and
clarifying the elements on which the process situationally segregated: taboos apply to
acts. It should be noted that our empirical whole, indivisible persons. During recent
task is to identify and describe these fieldwork in a marginal, newly contacted
elements, not to invent them. They are not New Guinea community I was very struck,
heuristic devices or intermediate variables coming from work in more complex
societies, by the constraints on status dif-
constructed to explain the recorded features
ferentiation and the simplicity of repertoires
of aggregate social form; on the contrary,
which characterize a system of social dif-
once they have been specified they should
also be available to observation as concrete ferentiation based on taboo. In such
features of actors and encounters on thesystems, different personal capacities tend
coalesce; situational restrictions of relevan
micro-level. This opens the way for empirical
arethe
testing of each main sector of the model: difficult to conceptualize for actors, a
elements, the transformation rules, and the facility to switch between conceptua
the
distinct
aggregate forms. The use of such a model as definitions of the situation is
a framework for comparative study should
lacking.
thereby become particularly fruitful. The most "primitive" distinction in the
definition
This perspective, then, opens the way for of situations is that between the
identifying, in the constitution of persons
sacred and the profane. By means of taboos
and encounters, the correlates of major
and other cumbersome and contrived idioms
morphological features of the macro-system. a supreme effort is made to conceptualize
A crucial concept in this connection isand thedistinguish the sacred situation from
concept of the definition of situation, profanewhich life. Perhaps some of our bewilder-
depicts how cultural factors of a variety ment ofwith much primitive ritual stems from
kinds are transformed by the actors into the fact that social anthropologists have not
constraints which operate on the process been ofsufficiently aware of this absence of
interaction. The comparison of some dif-
sophistication in the manipulation of
ferent types of society should have served definitions
to of the situation within many
emphasize the great differences in the native social systems. The organizational
apparatus for defining situations which message of the idioms seems clear enough to
actors in different cultures have at their us, and we are tempted to dismiss as "re-
command. The means whereby actors can dundancy" an elaboration of message which
create social organization is by their con- participants experience as highly necessary
ceptually differentiating kinds of social for so problematical a distinction. We
persons and kinds of occasions; it is their generally meet members of such societies
success in reaching agreement on some such after a recent history of coercive contact has
distinctions that makes a degree of order taught them to enter into interaction with
possible. The conceptual tools at the actors' strangers on the premises of the dominant
disposal for these purposes severely affect Westerner and they have as individuals
the kinds of social situations they can define achieved some sophistication in switching
and, thereby, the patterns of organization between whole social systems. We must
they can establish. recognize the cultural limitations found in

This content downloaded from 193.48.45.204 on Fri, 17 Feb 2017 14:33:34 UTC
All use subject to http://about.jstor.org/terms
Barth ] COMPARISON AND ANAL YSIS 215

the traditional apparatus of socialfor


institutions and the micro-level
defining of
social
situations and analyze interaction. each system in terms
only of those mechanisms To provide
on an adequate basis
which it foris com-
in
fact based. parative analysis, however, the framework
that has been developed above needs to be
In addition to the active manipulation of
relieved of at least two naive assumptions
situations characteristically practiced in
which have been implicit in my argument:
many societies, some complex systems are
(1) the empirical character of "pure types,"
also prominently organized by means of the
and (2) the irrelevance of scale. I shall try to
special kind of situational distinction that
show that these provisional simplifications
separate bureaucratic and personal
can be eliminated without seriously com-
capacities. Complexity results in part from
plicating the analysis.
the addition of this organizational mech-
anism to the others; and different domains
(1) I have already emphasized that the
typological dimensions developed in this
of social interaction within one population
paper do not apply to the description of
can become differently structured. Western
whole societies as abstract entities but to the
society is thus simultaneously built on
different ways in which social organizations
certain basic identities maintained by taboos
are constituted. The evidence seems over-
(such as sex, and often confessional member-
whelming that communities around the
ship in universal churches), on a wide variety
world differ in the way social persons are
of statuses which are very freely mani-
composed and social encounters are con-
pulated situationally, as well as on offices of
innumerable kinds which are delegated to summated. Yet it is not equally obvious how
encumbents who may act in these capacities one could go about classifying the social life
of any particular place with reference to the
only for certain purposes and may have
"pure types" generated by a few abstract
prestige reflected on them for occupying the
offices but do not command them as an dimensions. As was perhaps first shown
clearly in the "Analysis of a Social Situation
integral resource. The social person in such
systems is at best constituted as a highly in Modern Zululand" (Gluckman 1940),
social life at any one place encompasses
stratified repertoire of statuses including
elements of very diverse origins based on
imperative or inalienable identities,
very different sanctions. This has partly
situation-specific capacities, and offices.
come about through the colonialism and
The incomplete and provisional typology
other spectacular expansionism of a parti-
of social systems sketched in this paper
cular, and many places recent, phase of
history;
highlights the need for further analysis, in a descriptive accounts have often
common basic vocabulary, of the organ-depicted this as the clash of two "worlds"
ization of encounters in a wide range while
of concentrating on one or another of
different social systems. The concept of the
these worlds: the traditional or the
modernized.
definition of the situation may be utilized to But the life situation of any
describe the limitations which different one person in the community is compound
cultural traditions impose on the actors' and confusing and is rarely composed of two
control and ability to manipulate such such tidy sectors; also brokers and entre-
definitions. One can thereby depict the preneurs by their activities create social
essential duality of social organization systems
as a that are compound in their very
simultaneous ordering of activity both in constitution. On a wider scale, such activity
ties essentially all local communities
relations and tasks, and the great differences
together in a kind of complex global plural
in how such ordering is achieved in different
social systems. Empirical comparisons and society, and every basis for isolating and
analyses in this framework should contribute
classifying a part of this network seems to
have been lost.
both to our understanding of the macro-level

This content downloaded from 193.48.45.204 on Fri, 17 Feb 2017 14:33:34 UTC
All use subject to http://about.jstor.org/terms
216 AMERICAN ANTHROPOLOGIST [74,1972
What then is the utility of the
effect typological
boundary maintenance and brokerage.
This should
distinctions I have suggested reveal the
between character of the
dif-
ferently constituted systems? The fact
discontinuities of
that separate embracing an
local variation remains and encapsulated shows no systemssign of in terms of the soci
being ephemeral. Members processes that maintain, and change, the
of a Melanesian
and an Indian and a Norwegian institutional macro-systems. Eidheim (1971)
community
may interact in ways that are similarly presents one of the few case studies with
constrained by the rules and agents of such intentions (see also Barth, Ed. 1969).
modern bureaucracy or international busi- (2) By focusing the comparison on the
ness, but the ideas and capacities that they way in which social interaction is con-
themselves bring to their life situation, and stituted and channeled in different systems
that they have confirmed and maintained in rather than on the institutional features of
it, are radically different and variously gener- different societies, it is possible to ignore the
ate social systems with elementary, question of scale in membership when con-
replicating, involute or other characteristics. structing the dimensions for comparison.
Our problem, in other words, lies in defining But it is of interest to consider the question
the kind of abstraction to which the typo- of the organizational capacity of these dif-
logical properties of inventories, repertoires, ferent social systems with respect to scale.
and components for defining situationsOne cantends to take for granted that simple
be ascribed. Given the empirical continuum societies cannot organize large populations-
of one global society, these properties are it is difficult to demonstrate why
though
most reasonably ascribed to systems this of should be so. Conversely, there seem to
organization that have such separate be certain functional difficulties in maintain-
identities as the actors themselves give them, ing the viability of small and peripheral
and are often encapsulated one withincommunities the in industrial societies. Building
other (Bailey 1957, 1960, 1969). Thereby on the analytical framework developed here,
we avoid the need to dichotomize modem such questions can best be approached
and traditional institutional complexes, or to a discussion of information. I have
through
commit the simplification of distinguishing spoken as if persons in social systems share
between total societies as distinct bodies of codes and evaluations but have no significant
people. Rather, we can utilize the relativeprevious information about each other--i.e.,
discontinuities in the networks and premises
as if they have to arrive at a definition of the
of interaction to delimit social systems situation by signalling whenever they meet.
within their larger environment. Such Firstly, codes may be imperfectly shared
encapsulated systems can be characterized and only superficially understood, generat-
with respect to their internal constitutioning both unpredictability and dynamism as
well as surface agreement. With respect to
and workings, without ignoring the processes
whereby they articulate with their information in general much social inter-
environment (see especially Bailey action is in fact routinized in that previous
1969:146ff.) information makes signaling unnecessary,
The delimitation of the social systemand to even may constrain or invalidate efforts
which particular characteristics are ascribed to signal other definitions of the situation.
thus needs to be analytically validated in The amount of such previous information
each particular case. This requires a that is needed, or tolerated, in any particular
description of social organization in inter- organization will depend on the constitution
actional terms. Such a description must ofrest
that organization. On the other hand, the
on a demonstration on the one hand of the amount of such information that is actually
structure of encounters that provide the available will depend in part on the size of
vessels for internal social activity and, on the the interacting population. The capacity of
other hand, the encounters and roles that different social systems to organize societies

This content downloaded from 193.48.45.204 on Fri, 17 Feb 2017 14:33:34 UTC
All use subject to http://about.jstor.org/terms
Barth ] COMPARISON AND ANALYSIS 217

of different scale, i.e., cleartothrough


order studies
the in India
life (cf. of
Marriott
different sized total 1959populations,
and Vidyarthi 1961). is thus
closely connected with the amount of For a variety of interactions in industrial
previous information about persons that is Western society, no previous information is
required, and the mere signalling of the
required for interaction. This again depends
prominently on the techniques utilized by presence of an asset as an object of trans-
the actors to define situations. The storyactions
of is sufficient to elicit a status set and
Australian aborigines killing a strangerorganize
if an encounter. The social organiza-
they cannot establish a kinship connection tion also contains a great number of
with him (Radcliffe-Brown 1913) indicates functionally specific status sets in which one
of the positions is open to near universal
the necessity of certain vital items of infor-
mation for interaction in such a society. encumbency, as in the case of the doctor-
Without information on moiety and class, nurse-patient
or set. Interaction is prominently
perhaps certain indicative totems, there is organized
no by the fleeting mobilization of
way of ascribing any particular status and such sets by a multitude of otherwise
thus no basis for an ordered social inter- anonymous persons who merely demon-
action. A few such limited items of infor- strate their command of the minimal
mation, on the other hand, suffice to qualifications to be passenger, audienc
identify a whole social person and specify citizen, etc. Such an organization has th
his relations to all available alters in a capacity to structure immensely wide net
replicating type social system. works. The realization of this potential
By contrast, the involute systems require
further enhanced by the remarkable freed
more careful information management
of of
each individual person to accumulat
members. Indian villagers may interactinformation
for and act upon it by diversifyi
certain purposes across caste lines without
social relations and involving himself in
any previous information and merely deeper commitments with a particular alt
through a preliminary signalling of based
casteon this information. Whereas most
identities. Such interaction, then, maysocieties
take require such information to be
place with a very great number of potential
widely shared and legitimized, Western
partners and can organize relations in a
society largely allows it to be private and
large-scale network. In other spheres of only to the parties directly con-
known
activity, more information is required before
cerned, while remaining unknown to their
the suitability of the person as an interaction
many other alters. The contrast to other
partner is established, as, for example, if the
moderately large scale societies of involute
possibility of marriage is at issue. The widest
type is striking, where an initial lack of
potential scale of the kinship network is
information about a person gives him access
consequently much more limited than only thatto a provisional "stranger/guest" status,
based on some of the other status sets. The
which, though limited in various ways, is
capacity of such a social system to organize
multiplex and involves complex reciprocities
large populations thus varies, in an of a problematical kind and also shared
immediate way, with the status sets andobligations in a wider community.
kinds of assets involved in the encounters. The effects of imposing a Western form
The limits of the more narrow range net-of contract and situation-switching organiza-
works are set by the availability of infor- tion on a small scale population are however
mation, while the wider scale networks are familiar and disastrous. The mere fact of
limited only by the limits of shared codes limited personnel creates difficulty in
and evaluations. That the establishment and maintaining an organization based on
maintenance of such shared premises plays a specialization and division of labor: the
prominent part in the life of these limitation of the total inventory of statuses
traditional civilizations has become very that can be filled by the available members

This content downloaded from 193.48.45.204 on Fri, 17 Feb 2017 14:33:34 UTC
All use subject to http://about.jstor.org/terms
218 AMERICAN ANTHROPOLOGIST [74,1972

of the community here to the analysis


leads toof anetworks,
failure in place
ofof
vital functions. Small scale also has an structuralist macro-analysis.
immediate effect on the previous knowledge This discussion of interaction under vary-
actors bring to their encounters. Theing conditions of population size and
overlap
of personnel in different status setsorganization
leads to also brings out another
a surfeit of information and the coalescence important implication of this comparative
of separate occasions: persons have dif- attempt: that the very concept of "status"
ficulties segregating social situations and in these different social systems refers to
particularly separating out their bureau- rather different kinds of things. In the
cratically defined capacities. Existing simpler societies status refers to a sum of
relations are made more constraining andmultiplex capacities vis-a-vis alters with com-
given additional sanctions by the person's prehensive previous information about a
loss of freedom to withdraw from particular person. In involute systems it refers to
alters, while the safeguards to privacy whicha-perhaps compromising-component of a
multiple potential partners and formal stereotyped cluster of capacities. In modern
organizations provide in large scale com-contract society it may refer merely to the
munities are lost without the compensationability to demonstrate vis-d-vis strangers the
of adjustments and tolerance which multi- command of a very limited and specific
plex small scale organizations may provide. asset. In other words it varies between being
It has been suggested that a function of a total social identity, a compelling strait-
ritual in the social systems of small scale jacket, and an incidental option. The dif-
societies is precisely to overcome some of ference may be highlighted by the realiza-
these difficulties (Gluckman 1962:26). Most tion that a concept like that of role distance,
certainly the manipulation of the organiza- based on the distinction between subjective
tion by situational redefinition and switch- self and objective status (cf. Goffman 1961),
ing presupposes open networks and limited which seems very useful and fundamental to
information, and is defeated by small scale an understanding of status in our society,
(cf., e.g., Paine 1957 and 1965 for an becomes totally inapplicable in a social
empirical description of these effects). system of elementary type, based on only
very few status sets. Comparability in terms
It should perhaps be emphasized that thisof a concept of status is also seriously
perspective on scale does not attempt to impaired by the presence or absence of
derive structural principles by abstracting modes of communication as different as
from network form. Rather, it seeks to face-to-face interaction, writing, and tele-
identify the kinds of network that can be, communication, and fora as differently
and will tend to be, generated by actors constituted as domestic units, lecture halls,
interacting by means of certain organization- meetings of parliament, and art exhibits.
al aids. I would argue that it is not necessary, With the tools so far developed, any
or indeed possible, to reconstruct these attempt at all-embracing comparative ex-
organizational elements from a mere recordercises are consequently predestined to
of network form in a community. To failure in a number of gross ways, but may
provide such data in a systematic way is yet be of some value in the anthropological
most difficult and in itself insufficient, debate. I would particularly emphasize the
whereas the organizational elements are need, in any comparative reasoning, to cast
more readily identified by a close micro- one's analytical net somewhat wider than
analysis of encounters where the has been the practice and give attention
opportunity situation of each actor can be
simultaneously both to the wider organiza-
observed. The analytical perspective ex- tion of persons and of tasks and also to the
pounded here should provide an alternative structure of the encounters through which
and more feasible simulation-type approachany interaction must be consummated.

This content downloaded from 193.48.45.204 on Fri, 17 Feb 2017 14:33:34 UTC
All use subject to http://about.jstor.org/terms
Barth ] COMPARISON AND ANALYSIS 219

NOTES 1962 Essays in the Ritual of Social


Relations. Manchester: Manchester
University Press.
11I have elsewhere (Barth 1966) Goodenough,
pleaded W. H.
the utility of retaining the distinction
1965 Rethinking "Status" and "Role":
between status and role and will here apply Toward a General Model of the
Nadel's arguments to the status level of
Cultural Organization of Social Re-
analysis.
lationships. In The Relevance of
2 Merton speaks also of "role sectors" for Models for Social Anthropology.
these different faces, and introduced the Michael Banton, Ed. ASA Monograph
concept of "role set" for the whole inter- No. 1. London: Tavistock.
locking system seen from the point of viewGoffman, E.
of one ego. For the purpose of comparing 1959 The Presentation of Self in Every-
whole systems it is better to avoid the day Life. New York: Doubleday
egocentric perspective and focus on the total Anchor Books.
set of interrelated roles. Merton further uses 1961 Role Distance. In Encounters,
"status set" for the array of statuses Two Studies in the Sociology of Inter-
occupied by one individual, i.e., what action. Minneapolis and New York:
Radcliffe-Brown had already given the more Bobbs-Merrill.
useful name (social) "person." It seems to Leach, E. R.
me that the term "status set" can much 1960 (Ed.) Aspects of Caste in South
more usefully be reserved for the whole I ndia, Ceylon, and North-West
interlocking system of positions, as I do Pakistan. Cambridge Papers in Social
here, and "role set" be used to refer to the Anthropology No. 2. Cambridge:
interlocking system of behavior or activities Cambridge: Cambridge University
(cf. Barth 1966). Press.
1961 Rethinking Anthropology.
REFERENCES CITED London School of Economics Mono-
graphs in Social Anthropology No. 22
Bailey, F. G. London: Athlone Press.
1957 Caste and the Economic Frontier. IAvi-Strauss, C.
Manchester: Manchester University 1963 The Bear and the Barber. The
Press. Henry Myers Memorial Lecture, 1
1960 Tribe, Caste and Nation. Man- The Journal of the Royal Anthro-
chester: Manchester University Press. pological Institute 93:1-11.
1969 Strategems and Spoils. Oxford: Marriott, McKim
Blackwell. 1959 Changing Channels of Cultural
Barth, F. Transmission in Indian Civilization.
1966 Models of Social Organization. Chicago: The University of Chicago
London: Occasional Papers of the Committee on South Asian Studies
Royal Anthropological Institute, No. Reprint Series No. 7.
23.
Mayer, A. C.
1969 (Ed.) Ethnic Groups and Bound- 1966 The Significance of Quasi-Groups
aries. Oslo: Universitetsforlaget. in the Study of Complex Societies. In
Eidheim, H. The Social Anthropology of Complex
1971 Aspects of the Lappish Minority Socieites. Michael Banton, Ed. ASA
Situation. Oslo: Universitetsforlaget. Monograph No. 4. London: Tavistock.
Fortes, M. Merton, R. K.
1953 The Structure of Unilineal Descent 1957 Social Theory and Social Struc-
Groups. American Anthropologist ture. Revised and Enlarged Edition.
55:17-40. New York: Free Press.
Gluckman, M. Nadel, S. F.
1955 The Judicial Process Among the 1957 The Theory of Social Structure.
Barotse. Manchester: Manchester London: Cohen and West.
University Press. Paine, R.
1958 Analysis of a Social Situation in 1957 Coast Lapp Society, I. A Study of
Modern Zululand. Rhodes-Livingstone Neighbourhood in Revsbotn Fjord.
Papers 28. Manchester: Manchester Troms6: Troms5 Museums Skrifter,
University Press. (Papers originally Vol. IV.
published in Bantu Studies 1940, and 1965 Coast Lapp Society, II. Oslo: Uni-
in African Studies 1942). versitetsforlaget.

This content downloaded from 193.48.45.204 on Fri, 17 Feb 2017 14:33:34 UTC
All use subject to http://about.jstor.org/terms
220 AMERICAN ANTHROPOLOGIST [74,1972

Radcliffe-Brown, A. R.deReuck, A., and J. Knight, eds.


1913 Three Tribes of Western Australia. 1967 Caste and Race: Comparative
Journal of the Royal Anthropological Approaches. A CIBA Foundation
Institute 43:143-194. volume. London: Churchill.
1958 Method in Social Anthropology. Vidyarthi, L. P.
M. N. Srinivas, Ed. Chicago: The 1961 The Sacred Complex in Hindu
University of Chicago Press. Gaya. Bombay: Asia.

This content downloaded from 193.48.45.204 on Fri, 17 Feb 2017 14:33:34 UTC
All use subject to http://about.jstor.org/terms

You might also like