On the Study of Social Change*
FREDRIK BARTH
Traditional anthropological description in terms of pattern and custom, convenient as it is for certain
purposes, results essentially in accounts that do not adequately portray change. Change is more easily
handled if one looks at social behavior as allocations of time and resources. Analyses of ongoing process
that the latter view makes possible seem more productive of insight into the nature of social change than has
been the case with typological and comparative approaches.
T HE analytical contribution of modern
anthropology to the understanding of
social change has been limited, despite the fact
that our material is becoming increasingly rich
with most dramatic cases of change. I shall
use the opportunity that a brief and general
discussion of the wide theme of social change
offers to make a preliminary diagnosis of why
this should be so, and to suggest certain re-
quirements and reorientations that I feel are
necessary if we wish to remedy this situation.
I shall argue in favor of (a) a greater attention
to the empirical study of the events of change,
and a need for concepts that facilitate this;
(b) the necessity for specification of the nature
of the continuity in a sequence of change, and
the processual analyses that this entails; and
(c) the importance of the study of institu-
tionalization as an ongoing process.
We should not underestimate the effects
on our discipline that giving first priority to the
understanding of change may have. There
has been a comfortable convention in social
anthropology till now of treating "social
change" as if it were a topic of anthropological
investigation like "religion" or "domestic
organization," something that may be dis-
cussed in addition to, and preferably subse-
quent to, other substantive fields in the de-
scription of social systems. But if we couch our
description of these aspects of society as if we
were dealing with forms that do not entail
and reflect processes, we cannot expect that
the terms and concepts we develop in this de-
scription will serve us with equal facility in the
description of changing forms.
To understand social change, what we need
to do as social anthropologists is to describe all
• Plenary address to the American Anthropological Associa-
tion, 1966. Publication of this address is made possible by a con-
tribution from the Wenner-Gres foundation for Anthropological
Research. of society in such terms that we see
how it persists, maintains itself, and changes
through time. This may mean recasting many
661
University of Bergen
of our terms for the description of social
systems, not merely adding a chapter of
additional data. To do the job of analyzing
change adequately may mean that we will do
some of the old jobs less adequately, or at least
less simply, than we have been doing. To
someone who does not share this priority, the
efforts may look unnecessarily complicated and
relatively fruitless. But for those who give the
understanding of change high priority, it is
wishful thinking to expect that we can build
indiscriminately on all the concepts that our
discipline has developed for other purposes.
Because of our general unwillingness to
abandon well-established routines, studies
explicitly addressed to the investigation of
change have been prone to contain descrip-
tions of a social system at two points in
time—or even at one point in time?—and then
to rely on extrapolation between these two
states, or from the one state, to indicate the
course of change. I feel that if we want to
understand social change, we need concepts
that allow us to observe and describe the
events of change. Our contribution as social
anthropologists must lie in providing such
primary materials for understanding the
processes; it lies in our powers of observation
out there where change is happening today,
and not in producing secondary data by deduc-
tion and extrapolation. If this means that we
must recast our very description of social
systems in order to accommodate these data
about the events of change, that makes our
task more difficult but also more interesting.
The reason for the social anthropologist's
impasse when he tries to add change to his
662 American Anthropologist
traditional description of social systems is
found in the basic characteristics of the de-
scriptive concepts we habitually use. We wish
to characterize groups, societies, or cultures,
and to do this we have to aggregate individual
observations. We generally think of the pro-
cedure as one where we aggregate individual
cases of behavior to patterns of behavior, speci-
fying the common features of the individual
cases. Such patterns we think of as customs:
stereotyped forms of behavior that are re-
quired and correct. Some of us may choose to
emphasize the moral character of customs
(and thus the possibility of eliciting them
directly from informants) rather than their
stereotyped character, but in either case we feel
that the two are connected. We then construct
a system composed of such formal features,
and characterize the whole system as one
"with" dowry, or "with" cross-cousin mar-
riage, or "with" ambilocal residence.
This kind of morphological concept of cus-
tom as the minimal element of form has been
fundamental to our thinking because it serves
such a useful purpose. It allows us to aggregate
individual cases into a macrosystem and to
maintain the connection between the two
levels. We avoid the difficulties of some of the
other social sciences of using different kinds of
concepts for the description of the microunit
and the macroaggregate: a man "gives" a
dowry and a society "has" dowry. A custom
has morphological characteristics that are
like those of an individual item of behavior,
and on both levels we can use the same de-
scriptive and characterizing terms. And so we
can observe people practicing the very cul-
ture that we abstract, whereas nobody prac-
tices socioeconomic class or gross national
income.
But such a concept of custom makes the
pattern as a whole unobservable, except as
exemplified in the stereotyped aspects of each
individual case—the aggregate pattern can
never be observed by measurement. A
custom is revealed only in a series of more
or less representative exemplifications. And
change in a pattern, or change from one
pattern to another, is even less observable:
there is no way to observe and describe an
event of change, except perhaps in the field of
legislation.
A statistical view of the practice of customs
does not provide a way out. We may observe
breaches of custom—but is a breach of custom
an event of change? We may even summarize
a frequency, a rate of breaches of a custom; we
[69, 1967
will still know nothing about the probability
or imminence of a change in the custom, or
about the direction of change that frequent
breaches signal.
I feel that we need rather to use concepts
that enable us to depict the pattern itself as a
statistical thing, as a set of frequencies of
alternatives. If we, for example, look at social
behavior as an allocation of time and resources,
we can depict the pattern whereby people
allocate their time and resources. Changes in
the proportions of these allocations are ob-
servable, in the sense that they are measur-
able. New allocations are observable as con-
crete events that may have systematic effects
and thus generate important change. And this
view does not entail that we limit ourselves to
the description of an economic sector of activi-
ties only; it can be applied to the whole field
of social organization, to describe how people
in fact manage to arrange their lives.
Sharp's classic description of the introduc-
tion of the steel axe among the Yir Yoront of
Australia (Sharp 1952) stands out as an illumi-
nating case-study of social change precisely
because it adopts this perspective. It provides
an understanding of change by explaining the
changing bases from which people make their
allocations. We see how Yir Yoront women no
longer need to offer as much submission to their
husbands because they no longer need to go to
them to obtain an axe; we understand why
people no longer allocate time and resources
to intertribal festivals because they are no
longer dependent on them to obtain their tools.
This way of isolating the underlying deter-
minants of social forms, so as to see how
changes in them generate changing social
systems, implies a view of behavior and
society that is rather different from what has
frequently been adopted in anthropology.
What we see as a social form is, concretely, a
pattern of distribution of behavior by different
persons and on different occasions. I would
argue that it is not useful to assume that
this empirical pattern is a sought-for con-
dition, which all members of the community
equally value and willfully maintain. Rather,
it must be regarded as an epiphenomenon of a
great variety of processes in combination, and
our problem as social anthropologists is to
show how it is generated. The determinants of
the form must be of a variety of kinds. On the
BARTH] On the Study of Social Change
one hand, what persons wish to achieve, the
multifarious ends they are pursuing, will
channel their behavior. On the other hand,
technical and ecologic restrictions doom some
kinds of behavior to failure and reward others,
while the presence of other actors imposes
strategic constraints and opportunities that
modify the allocations people can make and
will benefit from making.
I would therefore argue that it is unfruitful
to explain a social form, a pattern, directly
by hypothesizing a purpose for it. Individual
actors and individual management units have
purposes and make allocations accordingly;
but a social form, in the sense of an over-all
pattern of statistical behavior, is the aggregate
pattern produced by the process of social life
through which ecologic and strategic con-
straints channel, defeat, and reward various
activities on the part of such management
units.
This analytic perspective stands in marked
contrast to the anthropological predilection
for going from a generalized type construct of a
social form to a list of "prerequisites" for this
general type. Though these two exercises are
so dose in many formal respects, their objeo-
tives are strikingly different. In one case, a
social form, or a whole society, is seen as a
morphological creature with certain require-
ments that need to be ascertained, in the func-
tionalist tradition, the better to understand
how it is put together. In the other case, a
social form is seen as the epiphenomenon of a
number of processes, and the analysis concen-
trates on showing how the form is generated.
Only the latter view- develops concepts that
directly promote the understanding of change.
I have been concerned recently to analyze
the institution of the beer party in the society
of the Fur, a village-dwelling population in
Darfur province of Sudan that subsists mainly
by the hoe cultivation of millet (Barth 1967).
One may describe the norms or customs
governing this institution and show how it
organizes a group of persons around a joint
task. Beer is supplied by a host, and guests
arrive to drink, sing, and work for the host.
Some of the guests are there by invitation;
many arrive unasked and unannounced, to
share in the work and the beer and the com-
pany. In all these respects, one beer party is
like another beer party, and this brief de-
scription summarizes the gross customary
features of the institution. As far as changes
in the institution go, all that the informants,
663
or an anthropologist with longer field work
in the area than myself, might be able to say is
that beer parties are becoming fewer or more
rowdy.
If one wishes to describe an institution as a
pattern of the allocation of time and resources,
one needs to specify the set of alternatives. In
a beer party you can be guest or host, or you
may choose to allocate your own labor directly
to your own millet field. Different frequencies
of these allocations entail different kinds of
community life: although they may be looked
at as behavioral outputs, their frequencies
have structural implications for the society.
Thus, where there is a predominance of
allocation of own labor to own fields, this
entails a limited circulation of labor services in
the community as a whole and a low level of
neighborliness and community life. Differ-
ences in wealth are constrained by the range of
the labor capacity of each cultivator-house-
holder.
Where on the other hand there is much beer-
party activity and reciprocity in the host-
guest relationship, this maintains an egali-
tarian, communal peasant community through
the constant circulation and redistribution of
labor services and rewards.
But the actual extent of reciprocity also
needs to be measured. If some consistently act
more as hosts than as guests, they are trans-
forming some millet into labor. An increased
rate of nonreciprocal allocations of this kind
leads to an increased social differentiation,
where some simultaneously obtain both
wealth and leisure; that is, it leads to change
in the direction of increased social stratifica-
tion.
One may therefore argue that these be-
havioral outputs feed back on the structure of
the community itself. The ubiquitous beer-
party guest, who is exchanging labor directly
for beer, does not ask himself : How will this
allocation affect our system of social stratifi-
cation? Yet his allocations, made on the basis
of limited considerations, do in fact create
directions and constraints on possible change.
It is only through attention to the frequencies
of allocations, by describing the pattern itself
as a certain set of frequencies, that it becomes
possible to observe and describe such quite
simple events of social change.
Because of an interest in observing events
of change, a group of us in Bergen decided to
664 American Anthropologist
turn our attention to the study of entrepre-
neurs (cf. Barth 1963). The choice was rather
obvious in that entrepreneurs are clearly
agents of change: they make innovations that
affect the community in which they are active.
Entrepreneurs are also much more common
and active in some communities and societies
than in others, and the dynamic character of
some societies has sometimes been explained
by the prevalence of entrepreneurs in them.
The anthropological study of entrepreneurs
and entrepreneurship has characteristically
sought to show the common characteristics of
entrepreneurs that differentiate them from
nonentrepreneurs, and thus the prerequisites
for the emergence of entrepreneurship. What
we did was to ask, not what makes the en-
trepreneur, but what does the entrepreneur
make: what can one say about his enterprise,
is it possible to characterize it as an event of
change?
Now in retrospect, one might see several
alternative ways of pursuing this question
and simpler ways of handling it than the ones
we adopted in that particular study. But what
proved stimulating to us then and later was
the way this question directed us to look for
ways of characterizing and describing change
itself, rather than the prerequisites for change.
We attempted to characterize particular cases
of entrepreneurial activity as new kinds of
allocation. But since our major interest was
not in an individual or a category of indi-
viduals, but in a social system, we had to go
on to characterize this social system and show
how the entrepreneurial activity in question
was changing it. We therefore had to try to
show the system of allocations in the entrepre-
neur's community and to place his new alloca-
tions in relation to these others. In this mate-
rial and elsewhere (Barth 1967) one finds that
entrepreneurs effect new conversions between
forms of goods that were previously not
directly convertible. They thereby create new
paths for the circulation of goods, often cross-
ing barriers between formerly discrete spheres
of circulation.
This activity cannot be without effect on the
culture of the members of an entrepreneur's
community. If we look for the bases on which
people make their allocations in primary
cultural facts such as people's categorization
of different kinds of goods and their preference
criteria for evaluating different outcomes of
their allocations, then we are relating their
choices to the cultural values or value orien-
[69, 1967
tations to which they subscribe. The entre-
preneurial coup, where one makes one's big
profits, is where one discovers a path by which
something of little value can be transformed
into something of great value. But looked at
this way, entrepreneurial successes produce
new information on the interrelations of
different categories of valued goods. The infor-
mation produced by such activity will render
false the idea that people have held till then
about the relative value of goods, and can
reasonably be expected to precipitate re-
evaluations and modifications both of cate-
gorizations and of value orientations. In other
words, it changes the cultural bases that deter-
mine people's behavior, and in this way
entrepreneurial activity becomes a major well-
spring of cultural and social change (cf. Barth
1966, esp. pp. 16-20).
However, the main point in the preceding
discussion is the most general one: I feel that
it is important for social anthropologists to
realize that we further our understanding of
social change best by using concepts that make
the concrete events of change available to
observation and systematic description.
There is also a requirement of another order
that needs to be observed in such studies. To
speak about change, one needs to be able to
specify the nature of the continuity between
the situations discussed under the rubric of
change. Change implies a difference of a very
particular kind: one that results from an
alteration through time and is determined by
the constraints of what has been, or continues,
in a situation. Let me use a very simple illus-
tration: Imagine a situation where you stand
looking into an aquarium, and you observe a
fish. A moment later you find yourself looking
at a crab in the same place where the fish was.
If you ask yourself how it got claws instead of
fins, you are implying a certain kind of con-
tinuity: this is the same body, and it has
changed its shape. If, on the other hand, you
say to yourself that this is the same aquarium,
you are specifying another kind of continuity,
implying a set of constraints that leads you to
formulate other hypotheses about the dy-
namics of change in this instance. Different
specifications of the nature of the continuity
that ties two situations together in a sequence
of change give rise to very different hypotheses
about the mechanisms and processes of change.
BARTH] On the Study of Social Change
For every analysis, it is therefore necessary for
us to make explicit our assertions about the
nature of the continuity.
In physical anthropology, the principle of
noninheritance of acquired characteristics
represents a step toward such a specification of
the nature of continuity. And the increasingly
rigorous study of change has only been made
possible through the explicit assertion that
what continues through time may be described
as a gene pool, and that changes in form re-
flect changes in the frequencies of genes in the
gene pool of the population.
In archeology, a hand-axe does not breed a
hand-axe, and the typological vocabulary
that seemed to imply this kind of continuity
has largely been dropped in favor of an ex-
plicit recognition that the continuity is found
in (a) the cultural tradition of the tool-makers.
However, the constraints on the processes of
change implied by this are very poorly under-
stood. Perhaps for that reason, archeology
seems so far to have been more successful when
specifying other kinds of continuity, such as
(b) the constancy of materials, implying con-
straints that help us understand courses of
change in techniques and art styles, or (c)
the continuity or slow change of environment,
enabling archeologists to see successive cul-
tures as changing adaptations to the environ-
ment.
In social anthropology, the specification of
continuity is highly problematical. To formu-
late hypotheses about change, we must be
able to specify the connection, that is, the
processes that maintain a social form, an
institution, or an organization. An item of
behavior does not breed an item of behavior.
What then is it that creates continuity of
society from one day to the next?
Obviously, one can say that society is in the
minds of men—as experiences and expecta-
tions. If forms of behavior can be described as
allocations with reference to evaluated ends,
then what persists in the minds of men can be
understood as items of credit and debt, as
prestations outstanding that make the actors
pick up where they last left off. In more general
terms, one can see a continuity of agreement
between people about the distribution of
assets—that is, about the location of rights in
statuses distributed in the population. Under-
lying these one might expect to find shared cul-
tural schemes of classification and evaluation.
But the aggregate pattern of behavior, the
665
structure of society, is not determined by this
alone, so this does not exhaust the factors of
continuity. What people do is also significantly
constrained by circumstance: a whole range of
facts of life, mainly ecological, enters as com-
ponents because people's allocations are
adjusted and adapted in terms of what they
experience as the observed outcomes of their
behavior. The strategic constraints of social
life also enter and affect behavior: people's
activities are canalized by the fact of competi-
tion and cooperation for valued goods with
other persons and thus by th e problems of
adapting one's behavior to that of others,
themselves predictive and adaptable.
I would argue that since these various com-
ponents are all involved as determinants of
the forms of aggregate social behavior, conse-
quently they must all enter into our specifica-
tions of the continuity connecting situations in
a sequence of change; and any hypothesis
about social change is inadequate unless it
takes all these constraints of continuity into
account. It may be a convenient shorthand for
structural comparison to say that a matrilineal
kinship system changes into a bilateral one, or
that a lineage organization develops into a seg-
mentary state. But such a formulation is not
a convenient shorthand for the series of events
of change that have taken place, since it begs
the whole analysis by implying a naïve and
mechanical kind of continuity between the
two forms, like that between the fish and the
crab in the aquarium.
Let me illustrate what I mean by a simple
example, again based on material from the
Fur.' Fur household organization is one where
each adult individual is an economic unit for
himself: each man or woman produces essen-
tially what he or she needs for food and cash,
and has a separate purse. Husband and wife
have certain customary obligations toward
each other: among other services, a wife must
cook and brew for her husband, and he must
provide her with clothes for herself and their
children. But each of the two cultivates sep-
arate fields and keeps provisions in separate
grain stores.
This arrangement can be depicted as a sys-
tem of allocations (Figure 1). A woman must
allocate a considerable amount of her time,
varying with the season, to agricultural pro-
duction. By virtue of the marriage contract,
she is also constrained to allocate time to
cooking and to brewing beer for her husband.
666 American Anthropologist
other activities
agricultural
production
agricultural
production
other activities
FIGURE
The husband, on his side, owes it to his wife to
allocate some of his cash to consumption goods
for her. Such patterns of allocation are thus
one way of describing the structure of Fur
family and household.
Some of these Fur couples change their mode
of life and become nomadic pastoralists like the
surrounding Baggara Arabs (cf. HAland 1967).
Together with this change in subsistence pat-
terns one finds a change in family and house-
hold form, in that such couples establish a
joint household. Their allocations change, as
compared to those of normal Fur villagers
(Figure 2). The husband specializes in the
activities that have to do with herding and
husbandry, while the woman cultivates some
millet, churns butter and markets it, and cooks
food. They have a joint grain store and a joint
purse and make up a unit for consumption.
In the anthropological tradition, one might
reasonably formulate the hypothesis that
what we observe here is a case of accultura-
tion: as part of the change to a Baggara Arab
way of life they also adopt the Arab household
form. This manner of describing the course of
change implies a very concrete view of house-
other activities
pastoral
production
agricultural
production
other activities
FIGURE 2
[69, 1967
saving
food and cash•
- -11
\0.7 , consumption
\to.
consumption
food and cash
saving
hold organization as one of the parts of Arab
culture, a set of customs that people can take
over.
Fortunately, the ethnographic material
provides us with a test case for the accultura-
tion hypothesis: some Fur cultivators, in
villages where they have no contact with Arab
horticultural populations, have recently taken
up fruit-growing in irrigated orchards as a
specialized form of cash-crop production.
Among such Fur too, one finds joint house-
holds, but with a slightly different pattern of
allocation (Figure 3). Here the conjugal pair
make up a unit both for production and con-
sumption, jointly cultivating the orchard and
sharing the returns.
To maintain the force of the acculturation
explanation of the form of the nomad house-
holds, one would have to look for similar fac-
tors in the case of the orchard cultivators and
hypothesize a change in values and accultura-
tion to modern life among them. But it is diffi-
cult to see the sources of influence for such ac-
culturation; more importantly, a restatement
of the nature of the continuity provides oppor-
tunities for other kinds of hypotheses. If we
__ i, consumption
food and cash
saving
BARTH] On the Study of Social Change
other activities
production —em- food and cash
other activities
FIGURE 3
agree that behavior in households is deter-
mined by several kinds of constraints, that all
behavior is "new" in that it constitutes alloca-
tions of time and resources made or renewed
in the moment of action, and that households
persist because their forms are recreated by
behavior each day, then we need to ask what
the other determinants of these allocations
are. To explain a changing pattern of activities,
we need not hypothesize changed categoriza-
tions and values: we can also look at the
changed circumstances that may well make
other allocations optimal when evaluated by
the same standards.
Indeed, the traditional range of behavior
and allocations in a Fur village indicates that
the Fur do not subscribe to any kind of pro-
hibition in joint conjugal households—such
arrangements are just not very convenient. A
fair autonomy of husband and wife is regarded
as a good thing, and joint economic pursuits
are a potential field for conflict. Moreover, the
techniques of millet cultivation are such that
persons work individually in any case; and
where a person desires help during peak
seasons, he or she can mobilize labor in bulk
through a beer work party. In the case of irri-
gated cash crops, on the other hand, the horti-
cultural techniques are such that it may be
convenient to cooperate. Persons with neigh-
boring plots often do so; occasionally, a hus-
band and wife will also decide to cultivate a
joint field—because they "like" to work to-
gether and because they can partly take turns
at irrigation, etc., partly cooperate.
The advantages of this jointness in cultiva-
tion are rather limited, only slightly reducing
the labor input required for the same result,
and few spouses choose to work jointly. But
in a situation where one of the spouses can
specialize in herding, the other in cultivation
and dairying, cooperation offers great ad-
vantages. Similarly, where a pooling of labor
667
consumption
'ob. saving
in specialized arboriculture and fruit-picking
gives far greater returns than millet cultiva-
tion, it is also clearly to the advantage of both
spouses to go together over production and
share the product jointly.
One may hypothesize a persistence of values
in all these different situations: (a) a preference
for husband-wife autonomy, and (b) a prefer-
ence for the minimization of effort in produc-
tion. How can spouses further these interests
in different situations where environmental
constraints change? Where effective produc-
tion can be pursued individually, persons will
be able simultaneously to maximize both
interests. Where pooling of labor in orchards
gives great returns with limited effort, this
allocation on the balance gives the greatest
advantage to both spouses. Where they thus
have a joint share in the product, it is difficult
and meaningless to divide it up when the
mutual obligations of cooking and clothing
tie the spouses together anyway for certain
aspects of consumption—so joint households
are generated. Finally, where complemen-
tarity and cooperation are not only advanta-
geous but necessary, as in a nomadic setting,
the necessary allocations will similarly create a
joint household, organized On a slightly
different pattern from that of the orchard
owners. It is by considering all the factors of
continuity in the situation of change—in
this case both valuational and technical-
economic—that we are in a position to formu-
late, and choose among, the full range of rele-
vant hypotheses.
In this example, then, we find that change in
household form is generated by changes in one
variable: the relative advantage of joint pro-
duction over separate production. This is
668 American Anthropologist
hardly a surprising conclusion. But if we attack
the problem in terms of a typology of house-
hold forms, we might be led to classify house-
hold type I (individual households for each
person) and household type II (joint conjugal
households) as very different forms and to
worry about how type I changes into type II,
which is like worrying about how the fish
changes into the crab. Yet the situation is
clearly not one where one household body
changes into another household body: it is
one where husband-wife sets, under different
circumstances, choose to arrange their life
differently. By being forced to specify the
nature of the continuity we are forced to
specify the processes that generate a house-
hold form. We see the same two people making
allocations and judging results in two different
situations, or we see a population of spouses
performing allocations in a pattern that gener-
ates predominantly individual households in
one opportunity situation, joint households in
another. We are led to seek the explanations
for change in the determinants of form, and the
mechanisms of change in the processes that
generate form.
In our efforts to understand social change,
this general viewpoint shifts our attention
from innovation to institutionalization as the
critical phase of change. People make alloca-
tions in terms of the pay-offs that they hope
to obtain, and their most adequate bases for
predicting these pay-offs are found in their
previous experience or in that of others in their
community. The kinds of new ideas that occur
can no more determine the direction of social
change than mutation rates can determine the
direction of physical change. Whatever ideas
people may have, only those that constitute
a practicable allocation in a concrete situation
will be effected. And if you have a system of
allocations going—as you always must where
you can speak of change—it will be the rates
and kinds of pay-offs of alternative allocations
within that system that determine whether they
will be adopted, that is, institutionalized. The
main constraints on change will thus be found
in the system, not in the range of ideas for in-
novation, and these constraints are effective
in the phase of institutionalization.
The comparative rates of pay-off of alterna-
tive allocations, which determine the course of
institutionalization, must be seen from the
point of view of actors or of other concrete
units of management that dispose over re-
sources and make allocations. Individual
[69, 1967
actors will naturally make frequent mis -
judgments of what the pay-offs of their alloca-
tions will be; but as the outcomes become ap-
parent through experience, they can be realisti-
cally evaluated. If the pay-offs are great, one
can expect the behavior to be emulated by
others; if, on the other hand, the results are
not desirable for the actor, he will not be
emulated, and he will also himself attempt to
revert to older allocations.
But the process of institutionalization is not
simply one of duplication; the allocations of
one unit can also have direct implications for
other units. They may find their opportunity
situation changed, not only through the possi-
bility of emulation, but also through a new
need for countermeasures or through new
opportunities for activity. The aggregate pat-
terns that can emerge in the population will
thus be shaped by the fact of competition and
the constraints of strategy. To depict these
constraints on actors and the way they will
determine the aggregate pattern of choices in
a population, we need models in the tradition
of game theory.
I do not wish to minimize the complexity of
the dynamics of such change and adjustment.
My main point is that most of the salient con-
straints on the course of change will be found
to be social and interactional, and not simply
cognitive. They will derive from the existing
social and ecological system within which
change is taking place. And finally, they can
most usefully be analyzed with reference to
the opportunity situation of social persons or
other units of management capable of deci-
sion-making and action: the mechanisms of
change must be found in the world of efficient
causes. It should follow from this that though
it may be a convenient and illuminating short-
hand of culture history to differentiate between
"emergent" and "recurrent" change, the
mechanisms involved seem to be essentially
the same: we must use the same tools to under-
stand the continuities that constitute society
in each case.
In summary, I should like to submit that
this general line of analysis—which is being
pursued in various ways by numerous col-
leagues in the United States and elsewhere—
makes it possible for us to improve our analyt-
ic and predictive understanding of social
 BARTH] On the Study of Social Change
change. I have had to harness it in this presen-
tation to specific, incomplete, and doubtless
in many ways inadequate exemplifications.
But its essentials are a concentration on the
observation of events of change and a specifi-
cation of the nature of continuity: the con-
straints of the whole system that is changing.
Conversely, I would suggest that approaches
that rely on typologies of overt social forms,
or seek to characterize and compare different
courses of change, will not provide as ready
insights into the nature of social change.
NOTES
This material derives from Gunnar land (1967) as well
as my own field material. REFERENCES CITED
BARTH, F.
669
1963 The role of the entrepreneur in social change in
Northern Norway. Bergen-Oslo, Norwegian Universities
Press.
1966 Models of social organization. Royal Anthropological
Institute Occasional Paper No. 23, London.
1967 Economic spheres in Darfur. In Themes in economic
anthropology. R. Firth, ed. London, Tavistock Publications.
HILAND, GUNNAR
1967 Ervervaform og etnisk tilhorighet. En studie av
nomadiseringsprosesser blant fastboende hakkebrukere i
det vestlige Darfur. University of Bergen. Unpublished
thesis.
SHARP, LAURISTON
1952 Steel axes for stone-age Australians. In Human prob-
lems in technological change. E. H. Spicer, ed. New York,
Russell Sage Foundation.