Note: This is a classic statement of what social theorists now call "boundary maintenance." It was written for anthropologists, but if you look beyond the ethnographic details you may find some good organizing principles for your paper. Although optional, this reading should be very useful for many students.
This collection of essays addresses itself to the problems of ethnic groups and their persistence. This is a theme of great, but neglected, importance to social anthropology. Practically all anthropological reasoning rests on the premise that cultural variation is discontinuous: that there are aggregates of people who essentially share a common culture, and interconnected differences that distinguish each such discrete culture from all others. Since culture is nothing but a way to describe human behaviour, it would follow that there are discrete groups of people, i.e. ethnic units, to correspond to each culture. The differences between cultures, and their historic boundaries and connections, have been given much attention; the constitution of ethnic groups, and the nature of the boundaries between them, have not been correspondingly investigated. Social anthropologists have largely avoided these problems by using a highly abstracted concept of 'society' to represent the encompassing social system within which smaller, concrete groups and units may be analysed. But this leaves untouched the empirical characteristics and boundaries of ethnic groups, and the important theoretical issues which an investigation of them raises.
Though the naïve assumption that each tribe and people has maintained its culture through a bellicose ignorance of its neighbours is no longer entertained, the simplistic view that geographical and social isolation have been the critical factors in sustaining cultural diversity persists. An empirical investigation of the character of ethnic boundaries, as documented in the following essays, produces two discoveries which are hardly unexpected, but which demonstrate the inadequacy of this view. First, it is clear that boundaries persist despite a flow of personnel across them. In other words, categorical ethnic distinctions do not depend on an absence of mobility, contact and information,  but do entail social processes of exclusion and incorporation whereby discrete categories are maintained despite changing participation and membership in the course of individual life histories. Secondly, one finds that stable, persisting, and often vitally important social relations are maintained across such boundaries, and are frequently based precisely on the dichotomized ethnic statuses. In other words, ethnic distinctions do not depend on an absence of social interaction and acceptance, but are quite to the contrary often the very foundations on which embracing social systems are built. Interaction in such a social system does not lead to its liquidation through change and acculturation; cultural differences can persist despite inter-ethnic contact and interdependence.
There is clearly an important field here in need of rethinking. What is required is a combined theoretical and empirical attack: we need to investigate closely the empirical facts of a variety of cases, and fit our concepts to these empirical facts so that they elucidate them as simply and adequately as possible, and allow us to explore their implications. In the following essays, each author takes up a case with which he is intimately familiar from his own fieldwork, and tries to apply a common set of concepts to its analysis. The main theoretical departure consists of several interconnected parts. First, we give primary emphasis to the fact that ethnic groups are categories of ascription and identification by the actors themselves, and thus have the characteristic of organizing interaction between people. We attempt to relate other characteristics of ethnic groups to this primary feature. Second, the essays all apply a generative viewpoint to the analysis; rather than working through a typology of forms of ethnic groups and relations, we attempt to explore the different processes that seem to be involved in generating and maintaining ethnic groups. Third, to observe these processes we shift the focus of investigation from internal constitution and history of separate groups to ethnic boundaries and boundary maintenance. Each of these points needs some elaboration.
Ethnic group defined
The term ethnic group is generally understood in anthropological literature (cf. e.g. Narroll 1964) to designate a population which:
1. is largely biologically self-perpetuating 
2. shares fundamental cultural values, realized in overt unity in cultural
3. makes up a field of communication and interaction
4. has a membership which identifies itself, and is identified by others, as constituting a category distinguishable from other categories of the same order.
This ideal type definition is not so far removed in content from the traditional proposition that a race = a culture = a language and that a society = a unit which rejects or discriminates against others. Yet, in its modified form it is close enough to many empirical ethnographic situations, at least as they appear and have been reported, so that this meaning continues to serve the purposes of most anthropologists. My quarrel is not so much with the substance of these characteristics, though as I shall show we can profit from a certain change of emphasis; my main objection is that such a formulation prevents us from understanding the phenomenon of ethnic groups and their place in human society and culture. This is because it begs all the critical questions: while purporting to give an ideal type model of a recurring empirical form, it implies a preconceived view of what are the significant factors in the genesis, structure, and function of such groups.
Most critically, it allows us to assume that boundary maintenance is unproblematical and follows from the isolation which the itemized characteristics imply: racial difference, cultural difference, social separation and language barriers, spontaneous and organized enmity. This also limits the range of factors that we use to explain cultural diversity: we are led to imagine each group developing its cultural and social form in relative isolation, mainly in response to local ecologic factors, through a history of adaptation by invention and selective borrowing. This history has produced a world of separate peoples, each with their culture and each organized in a society which can legitimately be isolated for description as an island to itself.
Ethnic groups as culture-bearing units
Rather than discussing the adequacy of this version of culture history for other than pelagic islands, let us look at some of the logical flaws in the viewpoint. Among the characteristics listed above, the sharing of a common culture is generally given central importance. In my view, much can be gained by regarding this very important feature as an implication or result, rather than a primary and definitional characteristic of ethnic group organization. If one chooses to regard  the culture-bearing aspect of ethnic groups as their primary characteristic, this has far-reaching implications. One is led to identify and distinguish ethnic groups by the morphological characteristics of the cultures of which they are the bearers. This entails a prejudged viewpoint both on (1) the nature of continuity in time of such units, and (2) the locus of the factors which determine the form of the units.
1. Given the emphasis on the culture-bearing aspect, the classification of persons and local groups as members of an ethnic group must depend on their exhibiting the particular traits of the culture. This is something that can be judged objectively by the ethnographic observer, in the culture-area tradition, regardless of the categories and prejudices of the actors. Differences between groups become differences in trait inventories; the attention is drawn to the analysis of cultures, not of ethnic organization. The dynamic relationship between groups will then be depicted in acculturation studies of the kind that have been attracting decreasing interest in anthropology, though their theoretical inadequacies have never been seriously discussed. Since the historical provenance of any assemblage of culture traits is diverse, the viewpoint also gives scope for an 'ethnohistory' which chronicles cultural accretion and change, and seeks to explain why certain items were borrowed. However, what is the unit whose continuity in time is depicted in such studies? Paradoxically, it must include cultures in the past which would clearly be excluded in the present because of differences in form differences of precisely the kind that are diagnostic in synchronic differentiation of ethnic units. The interconnection between 'ethnic group' and 'culture' is certainly not clarified through this confusion.
2. The overt cultural forms which can be itemized as traits exhibit the effects of ecology. By this I do not mean to refer to the fact that they reflect a history of adaptation to environment; in a more immediate way they also reflect the external circumstances to which actors must accommodate themselves. The same group of people, with unchanged values and ideas, would surely pursue different patterns of life and institutionalize different forms of behaviour when faced with the different opportunities offered in different environments? Likewise, we must expect to find that one ethnic group, spread over a territory with varying ecologic circumstances, will exhibit regional diversities of overt institutionalized behaviour which do not reflect differences in cultural orientation. How should they then be classified if overt institutional forms are diagnostic? A case in point is the  distributions and diversity of Pathan local social systems, discussed below (pp. 117 ff.). By basic Pathan values, a Southern Pathan from the homogeneous, lineageorganized mountain areas, can only find the behaviour of Pathans in Swat so different from, and reprehensible in terms of, their own values that they declare their northern brothers 'no longer Pathan'. Indeed, by 'objective' criteria, their overt pattern of organization seems much closer to that of Panjabis. But I found it possible, by explaining the circumstances in the north, to make Southern Pathans agree that these were indeed Pathans too, and grudgingly to admit that under those circumstances they might indeed themselves act in the same way. It is thus inadequate to regard overt institutional forms as constituting the cultural features which at any time distinguish an ethnic group these overt forms are determined by ecology as well as by transmitted culture. Nor can it be claimed that every such diversification within a group represents a first step in the direction of subdivision and multiplication of units. We have well-known documented cases of one ethnic group, also at a relatively simple level of economic organization, occupying several different ecologic niches and yet retaining basic cultural and ethnic unity over long periods (cf., e.g., inland and coastal Chuckchee (Bogoras 1904-9) or reindeer, river, and coast Lapps (Gjessing, 1954).
In one of the following essays, Blom (pp. 74 ff.) argues cogently on this point with reference to central Norwegian mountain farmers. He shows how their participation and self-evaluation in terms of general Norwegian values secures them continued membership in the larger ethnic group, despite the highly characteristic and deviant patterns of activity which the local ecology imposes on them. To analyse such cases, we need a viewpoint that does not confuse the effects of ecologic circumstances on behaviour with those of cultural tradition, but which makes it possible to separate these factors and investigate the nonecological cultural and social components creating diversity.
Ethnic groups as an organizational type
By concentrating on what is socially effective, ethnic groups are seen as a form of social organization. The critical feature then becomes item (4) in the list on p. 11 the characteristic of self-ascription and ascription by others. A categorical ascription is an ethnic ascription when it classifies a person in terms of his basic, most general identity, presumptively determined by his origin and background. To the extent that actors use ethnic identities to categorize themselves and others for  purposes of interaction, they form ethnic groups in this organizational sense.
It is important to recognize that although ethnic categories take cultural differences into account, we can assume no simple one-to-one relationship between ethnic units and cultural similarities and differences. The features that are taken into account are not the sum of 'objective' differences, but only those which the actors themselves regard as significant. Not only do ecologic variations mark and exaggerate differences; some cultural features are used by the actors as signals and emblems of differences, others are ignored, and in some relationships radical differences are played down and denied. The cultural contents of ethnic dichotomies would seem analytically to be of two orders: (i) overt signals or signs the diacritical features that people look for and exhibit to show identity, often such features as dress, language, house-form, or general style of life, and (ii) basic value orientations: the standards of morality and excellence by which performance is judged. Since belonging to an ethnic category implies being a certain kind of person, having that basic identity, it also implies a claim to be judged, and to judge oneself, by those standards that are relevant to that identity. Neither of these kinds of cultural 'contents' follows from a descriptive list of cultural features or cultural differences; one cannot predict from first principles which features will be emphasized and made organizationally relevant by the actors. In other words, ethnic categories provide an organizational vessel that may be given varying amounts and forms of content in different socio-cultural systems. They may be of great relevance to behaviour, but they need not be; they may pervade all social life, or they may be relevant only in limited sectors of activity. There is thus an obvious scope for ethnographic and comparative descriptions of different forms of ethnic organization.
The emphasis on ascription as the critical feature of ethnic groups also solves the two conceptual difficulties that were discussed above.
1. When defined as an ascriptive and exclusive group, the nature of continuity of ethnic units is clear: it depends on the maintenance of a boundary. The cultural features that signal the boundary may change, and the cultural characteristics of the members may likewise be transformed, indeed, even the organizational form of the group may change yet the fact of continuing dichotomization between members and outsiders allows us to specify the nature of continuity, and investigate the changing cultural form and content. 
2. Socially relevant factors alone become diagnostic for membership, not the overt, 'objective' differences which are generated by other factors. It makes no difference how dissimilar members may be in their overt behaviour if they say they are A, in contrast to another cognate category B, they are willing to be treated and let their own behaviour be interpreted and judged as A's and not as B's; in other words, they declare their allegiance to the shared culture of A's. The effects of this, as compared to other factors influencing actual behaviour, can then be made the object of investigation.
The boundaries of ethnic groups
The critical focus of investigation from this point of view becomes the ethnic boundary that defines the group, not the cultural stuff that it encloses. The boundaries to which we must give our attention are of course social boundaries, though they may have territorial counterparts. If a group maintains its identity when members interact with others, this entails criteria for determining membership and ways of signalling membership and exclusion. Ethnic groups are not merely or necessarily based on the occupation of exclusive territories; and the different ways in which they are maintained, not only by a once-and-for-all recruitment but by continual expression and validation, need to be analysed.
What is more, the ethnic boundary canalizes social life it entails a frequently quite complex organization of behaviour and social relations. The identification of another person as a fellow member of an ethnic group implies a sharing of criteria for evaluation and judgement. It thus entails the assumption that the two are fundamentally 'playing the same game', and this means that there is between them a potential for diversification and expansion of their social relationship to cover eventually all different sectors and domains of activity. On the other hand, a dichotomization of others as strangers, as members of another ethnic group, implies a recognition of limitations on shared understandings, differences in criteria for judgement of value and performance, and a restriction of interaction to sectors of assumed common understanding and mutual interest.
This makes it possible to understand one final form of boundary maintenance whereby cultural units and boundaries persist. Entailed in ethnic boundary maintenance are also situations of social contact between persons of different cultures: ethnic groups only persist as significant units if they imply marked difference in behaviour, i.e.  persisting cultural differences. Yet where persons of different culture interact, one would expect these differences to he reduced, since interaction both requires and generates a congruence of codes and values in other words, a similarity or community of culture (cf. Barth 1966, for my argumentation on this point). Thus the persistence of ethnic groups in contact implies not only criteria and signals for identification, but also a structuring of interaction which allows the persistence of cultural differences. The organizational feature which, I would argue, must be general for all inter-ethnic relations is a systematic set of rules governing interethnic social encounters. In all organized social life, what can be made relevant to interaction in any particular social situation is prescribed (Goffman 1959). If people agree about these prescriptions, their agreement on codes and values need not extend beyond that which is relevant to the social situations in which they interact. Stable inter-ethnic relations presuppose such a structuring of interaction: a set of prescriptions governing situations of contact, and allowing for articulation in some sectors or domains of activity, and a set of proscriptions on social situations preventing inter-ethnic interaction in other sectors, and thus insulating parts of the cultures from confrontation and modification.
Poly-ethnic social systems
This of course is what Furnivall (1944) so clearly depicted in his analysis of plural society: a poly-ethnic society integrated in the market place, under the control of a state system dominated by one of the groups, but leaving large areas of cultural diversity in the religious and domestic sectors of activity.
What has not been adequately appreciated by later anthropologists is the possible variety of sectors of articulation and separation, and the variety of poly-ethnic systems which this entails. We know of some of the Melanesian trade systems in objects belonging to the highprestige sphere of the economy, and even some of the etiquette and prescriptions governing the exchange situation and insulating it from other activities. We have information on various traditional polycentric systems from S.E. Asia (discussed below, Izikowitz pp. 135 ff.) integrated both in the prestige trade sphere and in quasi-feudal political structures. Some regions of S.W. Asia show forms based on a more fully monetized market economy, while political integration is polycentric in character. There is also the ritual and productive cooperation and political integration of the Indian caste system to be considered,  where perhaps only kinship and domestic life remain as a proscribed sector and a wellspring for cultural diversity. Nothing can be gained by lumping these various systems under the increasingly vague label of 'plural' society, whereas an investigation of the varieties of structure can shed a great deal of light on social and cultural forms.
What can be referred to as articulation and separation on the macro-level corresponds to systematic sets of role constraints on the micro-level. Common to all these systems is the principle that ethnic identity implies a series of constraints on the kinds of roles an individual is allowed to play, and the partners he may choose for different kinds of transactions. In other words, regarded as a status, ethnic identity is superordinate to most other statuses, and defines the permissible constellations of statuses, or social personalities, which an individual with that identity may assume. In this respect ethnic identity is similar to sex and rank, in that it constrains the incumbent in all his activities, not only in some defined social situations. One might thus also say that it is imperative in that it cannot be disregarded and temporarily set aside by other definitions of the situation. The constraints on a person's behaviour which spring from his ethnic identity thus tend to be absolute and, in complex poly-ethnic societies, quite comprehensive; and the component moral and social conventions are made further resistant to change by being joined in stereotyped clusters as characteristics of one single identity.
The associations of identities and value standards
The analysis of interactional and organizational features of interethnic relations has suffered from a lack of attention to problems of boundary maintenance. This is perhaps because anthropologists have reasoned from a misleading idea of the prototype inter-ethnic situation. One has tended to think in terms of different peoples, with different histories and cultures, coming together and accommodating themselves to each other, generally in a colonial setting. To visualize the basic requirements for the coexistence of ethnic diversity, I would suggest that we rather ask ourselves what is needed to make ethnic distinctions emerge in an area. The organizational requirements are clearly, first, a categorization of population sectors in exclusive and imperative status categories, and second, an acceptance of the principle that standards applied to one such category can be different from that applied to another. Though this alone does not explain why cultural  differences emerge, it does allow us to see how they persist. Each category can then be associated with a separate range of value standards. The greater the differences between these value orientations are, the more constraints on inter-ethnic interaction do they entail: the statuses and situations in the total social system involving behaviour which is discrepant with a person's value orientations must be avoided, since such behaviour on his part will be negatively sanctioned. Moreover, because identities are signalled as well as embraced, new forms of behaviour will tend to be dichotomized: one would expect the role constraints to operate in such a way that persons would be reluctant to act in new ways from a fear that such behaviour might be inappropriate for a person of their identity, and swift to classify forms of activity as associated with one or another cluster of ethnic characteristics. Just as dichotomizations of male versus female work seem to proliferate in some societies, so also the existence of basic ethnic categories would seem to be a factor encouraging the proliferation of cultural differentiae.
In such systems, the sanctions producing adherence to group-specific values are not only exercised by those who share the identity. Again, other imperative statuses afford a parallel: just as both sexes ridicule the male who is feminine, and all classes punish the proletarian who puts on airs, so also can members of all ethnic groups in a poly-ethnic society act to maintain dichotomies and differences. Where social identities are organized and allocated by such principles, there will thus be a tendency towards canalization and standardization of interaction and the emergence of boundaries which maintain and generate ethnic diversity within larger, encompassing social systems.
Interdependence of ethnic groups
The positive bond that connects several ethnic groups in an encompassing social system depends on the complementarity of the groups with respect to some of their characteristic cultural features. Such complementarity can give rise to interdependence or symbiosis, and constitutes the areas of articulation referred to above; while in the fields where there is no complementarity there can be no basis for organization on ethnic lines there will either be no interaction, or interaction without reference to ethnic identity.
Social systems differ greatly in the extent to which ethnic identity, as an imperative status, constrains the person in the variety of statuses and roles he may assume. Where the distinguishing values connected  with ethnic identity are relevant only to a few kinds of activities, the social organization based on it will be similarly limited. Complex polyethnic systems, on the other hand, clearly entail the existence of extensively relevant value differences and multiple constraints on status combinations and social participation. In such systems, the boundary maintaining mechanisms must be highly effective, for the following reasons: (i) the complexity is based on the existence of important, complementary cultural differences; (ii) these differences must be generally standardized within the ethnic group i.e. the status cluster, or social person, of every member of a group must be highly stereotyped so that inter-ethnic interaction can be based on ethnic identities; and (iii) the cultural characteristics of each ethnic group must be stable, so that the complementary differences on which the systems rest can persist in the face of close inter-ethnic contact. Where these conditions obtain, ethnic groups can make stable and symbiotic adaptations to each other: other ethnic groups in the region become a part of the natural environment; the sectors of articulation provide areas that can be exploited, while the other sectors of activity of other groups are largely irrelevant from the point of view of members of any one group.
Such interdependence's can partly be analysed from the point of view of cultural ecology, and the sectors of activity where other populations with other cultures articulate may be thought of as niches to which the group is adapted. This ecologic interdependence may take several different forms, for which one may construct a rough typology. Where two or more ethnic groups are in contact, their adaptations may entail the following forms:
(1) They may occupy clearly distinct niches in the natural environment and be in minimal competition for resources. In this case their interdependence will be limited despite co-residence in the area, and the articulation will tend to be mainly through trade, and perhaps in a ceremonial-ritual sector.
(2) They may monopolize separate territories, in which case they are in competition for resources and their articulation will involve politics along the border, and possibly other sectors.
(3) They may provide important goods and services for each other, i.e. occupy reciprocal and therefore different niches but in close interdependence. If they do not articulate very closely in the political  sector, this entails a classical symbiotic situation and a variety of possible fields of articulation. If they also compete and accommodate through differential monopolization of the means of production, this entails a close political and economic articulation, with open possibilities for other forms of interdependence as well.
These alternatives refer to stable situations. But very commonly, one will also find a fourth main form: where two or more interspersed groups are in fact in at least partial competition within the same niche. With time one would expect one such group to displace the other, or an accommodation involving an increasing complementarity and interdependence to develop.
From the anthropological literature one can doubtless think of type cases for most of these situations. However, if one looks carefully at most empirical cases, one will find fairly mixed situations obtaining, and only quite gross simplifications can reduce them to simple types. I have tried elsewhere (Earth 19646) to illustrate this for an area of Baluchistan, and expect that it is generally true that an ethnic group, on the different boundaries of its distribution and in its different accommodations, exhibits several of these forms in its relations to other groups.
These variables, however, only go part of the way in describing the adaptation
of a group. While showing the qualitative, (and ideally quantitative) structure
of the niches occupied by a group, one cannot ignore the problems of number
and balance in its adaptation. Whenever a population is dependent on its exploitation
of a niche in nature, this implies an upper limit on the size it may attain
corresponding to the carrying capacity of that niche; and any stable adaptation
entails a control on population size. If, on the other hand, two populations
are ecologically interdependent, as two ethnic groups in a symbiotic relationship,
this means that any variation in the size of one must have important effects
on the other. In the analysis of any poly-ethnic system for which we assert
any degree of time depth, we must therefore be able to explain the processes
whereby the sizes of the interdependent ethnic groups are balanced. The demographic
balances involved are thus quite complex, since a group's adaptation to a niche
in nature is affected by its absolute size, while a group's adaptation
to a niche constituted by another ethnic group is affected by its relative
The demographic problems in an analysis of ethnic inter-relations in a region thus centre on the forms of recruitment to ethnic groups and the question of how, if at all, their rates are sensitive to pressures on the different niches which each group exploits. These factors are highly critical for the stability of any poly-ethnic system, and it might look as if any population change would prove destructive. This does not necessarily seem to follow, as documented e.g. in the essay by Siverts (pp. 101 ff.), but in most situations the poly-ethnic systems we observe do entail quite complex processes of population movement and adjustment. It becomes clear that a number of factors other than human fertility and mortality affect the balance of numbers. From the point of view of any one territory, there are the factors of individual and group movements: emigration that relieves pressure, immigration that maintains one or several co-resident groups as outpost settlements of larger population reservoirs elsewhere. Migration and conquest play an intermittent role in redistributing populations and changing their relations. But the most interesting and often critical role is played by another set of processes that effect changes of the identity of individuals and groups. After all, the human material that is organized in an ethnic group is not immutable, and though the social mechanisms discussed so far tend to maintain dichotomies and boundaries, they do not imply 'stasis' for the human material they organize: boundaries may persist despite what may figuratively be called the 'osmosis' of personnel through them.
This perspective leads to an important clarification of the conditions for
complex poly-ethnic systems. Though the emergence and persistence of such systems
would seem to depend on a relatively high stability in the cultural features
associated with ethnic groups i.e. a high degree or rigidity in the interactional
boundaries they do not imply a similar rigidity in the patterns of
recruitment or ascription to ethnic groups: on the contrary, the ethnic inter-relations
that we observe frequently entail a variety of processes which effect changes
in individual and group identity and modify the other demographic factors that
obtain in the situation. Examples of stable and persisting ethnic boundaries
that are crossed by a flow of personnel are clearly far more common than the
ethnographic literature would lead us to believe. Different processes of such
crossing are exemplified in these essays, and the conditions which cause them
are shown to be various. We may look briefly at some of them.. 
Factors in identity change
The Yao described by Kandre (19676) are one of the many hill peoples on the southern fringe of the Chinese area. The Yao are organized for productive purposes in extended family households, aligned in clans and in villages. Household leadership is very clear, while community and region are autochthonously acephalous, and variously tied to poly-ethnic political domains. Identity and distinctions are expressed in complex ritual idioms, prominently involving ancestor worship. Yet this group shows the drastic incorporation rate of 10 °/o non-Yao becoming Yao in each generation (Kandre 1967a: 594). Change of membership takes place individually, mostly with children, where it involves purchase of the person by a Yao houseleader, adoption to kinship status, and full ritual assimilation. Occasionally, change of ethnic membership i$ also achieved by men through uxorilocal marriage; Chinese men are the acceptable parties to such arrangements.
The conditions for this form of assimilation are clearly twofold: first, the presence of cultural mechanisms to implement the incorporation, including ideas of obligations to ancestors, compensation by payment, etc., and secondly, the incentive of obvious advantages to the assimilating household and leader. These have to do with the role of households as productive units and agro-managerial techniques that imply an optimal size of 6-8 working persons, and the pattern of intra-community competition between household leaders in the field of wealth and influence.
Movements across the southern and northern boundaries of the Pathan area (cf.
pp. 123 ff.) illustrate quite other forms and conditions. Southern Pathans become
Baluch and not vice versa; this transformation can take place with individuals
but more readily with whole households or small groups of households; it involves
loss of position in the rigid genealogical and territorial segmentary system
of Pathans and incorporation through clientage contract into the hierarchical,
centralized system of the Baluch. Acceptance in the receiving group is conditional
on the ambition and opportunism of Baluch political leaders. On the other hand,
Pathans in the north have, after an analogous loss of position in their native
system, settled in and often conquered new territories in Kohistan. The effect
in due course has been a reclassification of the settling communities among
the congeries of locally diverse Kohistani tribes and groups.
Perhaps the most striking case is that from Darfur provided by Haaland (pp. 58 ff.), which shows members of the hoe-agricultural Fur of the Sudan changing their identity to that of nomadic cattle Arabs. This process is conditional on a very specific economic circumstance: the absence of investment opportunities for capital in the village economy of the Fur in contrast to the possibilities among the nomads. Accumulated capital, and the opportunities for its management and increase, provide the incentive for Fur households to abandon their fields and villages and change to the life of the neighbouring Baggara, incidentally also joining one of the loose but nominally centralized Baggara political units if the change has been economically completely successful.
These processes that induce a flow of personnel across ethnic boundaries will of necessity affect the demographic balance between different ethnic groups. Whether they are such that they contribute to stability in this balance is an entirely different question. To do so, they would have to be sensitive to changes in the pressure on ecologic niches in a feed-back pattern. This does not regularly seem to be the case. The assimilation of non-Yao seems further to increase the rate of Yao growth and expansion at the expense of other groups, and can be recognized as one, albeit minor, factor furthering the progressive Sinization process whereby cultural and ethnic diversity has steadily been reduced over vast areas. The rate of assimilation of Pathans by Baluch tribes is no doubt sensitive to population pressure in Pathan areas, but simultaneously sustains an imbalance whereby Baluch tribes spread northward despite higher population pressures in the northern areas. Kohistani assimilation relieves population pressure in Pathan area while maintaining a geographically stable boundary. Nomadization of the Fur replenishes the Baggara, who are elsewhere becoming sedentarized. The rate, however, does not correlate with pressure on Fur lands since nomadization is conditional on accumulated wealth, its rate probably decreases as Fur population pressure increases. The Fur case also demonstrates the inherent instability of some of these processes, and how limited changes can have drastic results: with the agricultural innovation of orchards over the last ten years, new investment opportunities are provided which will probably greatly reduce, or perhaps for a while even reverse, the nomadization process.
Thus, though the processes that induce change of identity are important to the understanding of most cases of ethnic interdependence, they need not be conducive to population stability. In general, however,  one can argue that whenever ethnic relations are stable over long periods, and particularly where the interdependence is close, one can expect to find an approximate demographic balance. The analysis of the different factors involved in this balance is an important part of the analysis of the ethnic inter-relations in the area.
The persistence of cultural boundaries
In the preceding discussion of ethnic boundary maintenance and interchange of personnel there is one very important problem that I have left aside. We have seen various examples of how individuals and small groups, because of specific economic and political circumstances in their former position and among the assimilating group, may change their locality, their subsistence pattern, their political allegiance and form, or their household membership. This still does not fully explain why such changes lead to categorical changes of ethnic identity, leaving the dichotomized ethnic groups unaffected (other than in numbers) by the interchange of personnel. In the case of adoption and incorporation of mostly immature and in any case isolated single individuals into pre-established households, as among the Yao, such complete cultural assimilation is understandable: here every new person becomes totally immersed in a Yao pattern of relationships and expectations. In the other examples, it is less clear why this total change of identity takes place. One cannot argue that it follows from a universally imputable rule of cultural integration, so that the practice of the politics of one group or the assumption of its pattern of ecologic adaptation in subsistence and economy, entails the adoption also of its other parts and forms. Indeed, the Pathan case (Ferdinand 1967) directly falsifies this argument, in that the boundaries of the Pathan ethnic group crosscuts ecologic and political units. Using self-identification as the critical criterion of ethnic identity, it should thus be perfectly possible for a small group of Pathans to assume the political obligations of membership in a Baluch tribe, or the agricultural and husbandry practices of Kohistanis, and yet continue to call themselves Pathans. By the same token one might expect nomadization among the Fur to lead to the emergence of a nomadic section of the Fur, similar in subsistence to the Baggara but different from them in other cultural features, and in ethnic label.
Quite clearly, this is precisely what has happened in many historical situations. In cases where it does not happen we see the organizing and canalizing effects of ethnic distinctions. To explore the factors  responsible for the difference, let us first look at the specific explana-, lions for the changes of identity that have been advanced in the examples discussed above.
In the case of Pathan borderlands, influence and security in the segmentary and anarchic societies of this region derive from a man's previous actions, or rather from the respect that he obtains from these acts as judged by accepted standards of evaluation. The main fora for exhibiting Pathan virtues are the tribal council, and stages for the display of hospitality. But the villager in Kohistan has a standard of living where the hospitality he can provide can hardly compete with that of the conquered serfs of neighbouring Pathans, while the client of a Baluch leader cannot speak in any tribal council. To maintain Pathan identity in these situations, to declare oneself in the running as a competitor by Pathan value standards, is to condemn oneself in advance to utter failure in performance. By assuming Kohistani or Baluch identity, however, a man may, by the same performance, score quite high on the scales that then become relevant. The incentives to a change in identity are thus inherent in the change in circumstances.
Different circumstances obviously favour different performances. Since ethnic identity is associated with a culturally specific set of value standards, it follows that there are circumstances where such an identity can be moderately successfully realized, and limits beyond which such success is precluded. I will argue that ethnic identities will not be retained beyond these limits, because allegiance to basic value standards will not be sustained where one's own comparative performance is utterly inadequate. The two components in this relative measure of success are, first, the performance of others and, secondly, the alternatives open to oneself. I am not making an appeal to ecologic adaptation. Ecologic feasibility, and fitness in relation to the natural environment, matter only in so far as they set a limit in terms of sheer physical survival, which is very rarely approached by ethnic groups. What matters is how well the others, with whom one interacts and to whom one is compared, manage to perform, and what alternative identities and sets of standards are available to the individual.
Ethnic identity and tangible assets
The boundary-maintaining factors in the Fur are not immediately illuminated by this argument. Haaland (pp. 65 f.) discusses the evaluation of the nomad's life by Fur standards and finds the balance between advantages and disadvantages inconclusive. To ascertain the comparability  of this case, we need to look more generally at all the factors that affect the behaviour in question. The materials derive from grossly different ethnographic contexts and so a number of factors are varied simultaneously.
The individual's relation to productive resources stands out as the significant contrast between the two regions. In the Middle East, the means of production are conventionally held as private or corporate, defined and transferable property. A man can obtain them through a specific and restricted transaction, such as purchase or lease; even in conquest the rights that are obtained are standard, delimited rights. In Darfur, on the other hand, as in much of the Sudanic belt, the prevailing conventions are different. Land for cultivation is allocated, as needed, to members of a local community. The distinction between owner and cultivator, so important in the social structure of most Middle Eastern communities, cannot be made because ownership does not involve separable, absolute, and transferable rights. Access to the means of production in a Fur village is therefore conditional only on inclusion in the village community i.e. on Fur ethnic identity. Similarly, grazing rights are not allocated and monopolized, even as between Baggara tribes. Though groups and tribes tend to use the same routes and areas every year, and may at times try in an ad hoc way to keep out others from an area they wish to use, they normally intermix and have no defined and absolute prerogatives. Access to grazing is thus an automatic aspect of practising husbandry, and entails being a Baggara.
The gross mechanisms of boundary maintenance in Darfur are thus quite simple:
a man has access to the critical means of production by virtue of practising
a certain subsistence; this entails a whole style of life, and all these characteristics
are subsumed under the ethnic labels Fur and Baggara. In the Middle East, on
the other hand, men can obtain control over means of production through a transaction
that does not involve their other activities; ethnic identity is then not necessarily
affected and this opens the way for diversification. Thus nomad, peasant, and
city dweller can belong to the same ethnic group in the Middle East; where ethnic
boundaries persist they depend on more subtle and specific mechanisms, mainly
connected with the unfeasibility of certain status and role combinations. 
Ethnic groups and stratification
Where one ethnic group has control of the means of production utilized by another group, a relationship of inequality and stratification obtains. Thus Fur and Baggara do not make up a stratified system, since they utilize different niches and have access to them independently of each other, whereas in some parts of the Pathan area one finds stratification based on the control of land, Pathans being landowners, and other groups cultivating as serfs. In more general terms, one may say that stratified poly-ethnic systems exist where groups are characterized by differential control of assets that are valued by all groups in the system. The cultures of the component ethnic groups in such systems are thus integrated in a special way: they share certain general value orientations and scales, on the basis of which they can arrive at judgements of hierarchy.
Obversely, a system of stratification does not entail the existence of ethnic groups. Leach (1967) argues convincingly that social classes are distinguished by different sub-cultures, indeed, that this is a more basic characteristic than their hierarchical ordering. However, in many systems of stratification we are not dealing with bounded strata at all: the stratification is based simply on the notion of scales and the recognition of an ego-centered level of 'people who are just like us' versus those more select and those more vulgar. In such systems, cultural differences, whatever they are, grade into each other, and nothing like a social organization of ethnic groups emerges: Secondly, most systems of stratification allow, or indeed entail, mobility based on evaluation by the scales that define the hierarchy. Thus a moderate failure in the 'B' sector of the hierarchy makes you a 'C', etc. Ethnic groups are not open to this kind of penetration: the ascription of ethnic identity is based on other and more restrictive criteria. This is most clearly illustrated by Knutsson's analysis of the Galla in the context of Ethiopian society (pp. 86 ff.) a social system where whole ethnic groups are stratified with respect to their positions of privilege and disability within the state. Yet the attainment of a governorship does not make an Amhara of a Galla, nor does estrangement as an outlaw entail loss of Galla identity.
From this perspective, the Indian caste system would appear to be a special case of a stratified poly-ethnic system. The boundaries of castes are defined by ethnic criteria: thus individual failures in performance lead to out-casting and not to downcasting. The process  whereby the hierarchical system incorporates new ethnic groups is demonstrated in the sanscritization of tribals: their acceptance of the critical value scales defining their position in the hierarchy of ritual purity and pollution is the only change of values that is necessary for a people to become an Indian caste. An analysis of the different processes of boundary maintenance involved in different inter-caste relations and in different regional variants of the caste system would,. I believe, illuminate many features of this system.
The preceding discussion has brought out a somewhat anomalous general feature of ethnic identity as a status: ascription' is not conditional on the control of any specific assets, but rests on criteria of origin and commitment; whereas performance in the status, the adequate acting out of -the roles required to realize the identity, in many systems does require such assets. By contrast, in a bureaucratic office the incumbent is provided with those assets that are required for the performance of the role; while kinship positions, which are ascribed without reference to a person's assets, likewise are not conditional on performance you remain a father even if you fail to feed your child.
Thus where ethnic groups are interrelated in a stratified system, this requires
the presence of special processes that maintain differential control of assets.
To schematize: a basic premise of ethnic group organization is that every A
can act roles, 1, 2 and 3. If actors agree on this, the premise is self-fulfilling,
unless acting in these roles requires assets that are distributed in a discrepant
pattern. If these assets are obtained or lost in ways independent of being an
A, and sought and avoided without reference to one's identity as an A, the premise
will be falsified: some A's become unable to act in the expected roles. Most
systems of stratification are maintained by the solution that in such cases,
the person is no longer an A. In the case of ethnic identity, the solution on
the contrary is the recognition that every A no longer can or will act in roles
I and 2. The persistence of stratified poly-ethnic systems thus entails the
presence of factors that generate and maintain a categorically different distribution
of assets: state controls, as in some modern plural and racist systems; marked
differences in evaluation that canalize the efforts of actors in different directions,
as in systems with polluting occupations; or differences in culture that generate
marked differences in political organization, economic organization, or individual
The problem of variation
Despite such processes, however, the ethnic label subsumes a number of simultaneous characteristics which no doubt cluster statistically, but which are not absolutely interdependent and connected. Thus there will he variations between members, some showing many and some? showing few characteristics. Particularly where people change their, identity, this creates ambiguity since ethnic membership is at once a question of source of origin as well as of current identity. Indeed; Haaland was taken out to see 'Fur who live in nomad camps', and I have heard members of Baluch tribal sections explain that they are 'really Pathan'. What is then left of the boundary maintenance and the categorical dichotomy, when the actual distinctions are blurred in this way? Rather than despair at the failure of typological schematism, one can legitimately note that people do employ ethnic labels and that there are in many parts of the world most spectacular differences whereby forms of behaviour cluster so that whole actors tend to fall into such categories in terms of their objective behaviour. What is surprising is not the existence of some actors that fall between these categories, and of some regions in the world where whole persons do not tend to sort themselves out in this way, but the fact that variations tend to cluster at all. We can then be concerned not to perfect a typology, but to discover the processes that bring about such clustering.
An alternative mode of approach in anthropology has been to dichotomize the ethnographic material in terms of ideal versus actual or conceptual versus empirical, and then concentrate on the consistencies (the 'structure') of the ideal, conceptual part of the data, employing some vague notion of norms and individual deviance to account for the actual, statistical patterns. It is of course perfectly feasible to distinguish between a people's model of their social system and their aggregate pattern of pragmatic behaviour, and indeed quite necessary not to confuse the two. But the fertile problems in social anthropology are concerned with how the two are interconnected, and it does not follow that this is best elucidated by dichotomizing and confronting them as total systems. In these essays we have tried to build the analysis on a lower level of interconnection between status and behaviour. I would argue that people's categories are for acting, and are significantly affected by interaction rather than contemplation. In showing the connection between ethnic labels and the maintenance  of cultural diversity, I am therefore concerned primarily to show how, under varying circumstances, certain constellations of categorization and value orientation have a self-fulfilling character, how others will tend to be falsified by experience, while others again are incapable of consummation in interaction. Ethnic boundaries can emerge and persist only in the former situation, whereas they should dissolve or be absent in the latter situations. With such a feedback from people's experiences to the categories they employ, simple ethnic dichotomies can be retained, and their stereotyped behavioural differential reinforced, despite a considerable objective variation. This is so because actors struggle to maintain conventional definitions of the situation in social encounters through selective perception, tact, and sanctions, and because of difficulties in finding other, more adequate codifications of experience. Revision only takes place where the categorization is grossly inadequate not merely because it is untrue in any objective sense, but because it is consistently unrewarding to act upon, within the domain where the actor makes it relevant. So the dichotomy of Fur villagers and Baggara nomads is maintained despite the patent presence of a nomadic camp of Fur in the neighbourhood; the fact that those nomads speak Fur and have kinship connections with villagers somewhere does not change the social situation in which the villager interacts with them it simply makes the standard transactions of buying milk, allocating camp sites, or obtaining manure, which one would have with other Baggara, flow a bit more smoothly. But a dichotomy between Pathan landowners and nonPathan labourers can no longer be maintained where non-Pathans obtain land and embarrass Pathans by refusing to respond with the respect which their imputed position as menials would have sanctioned.
Minorities, pariahs, and organizational characteristics o f the periphery
In some social systems, ethnic groups co-reside though no major aspect of structure is based on ethnic inter-relations. These are generally referred to as societies with minorities, and the analysis of the minority situation involves a special variant of interethnic relations. I think in most cases, such situations have come about as a result of external historical events; the cultural differentiae have not sprung from the local organizational context rather, a pre-established cultural contrast is brought into conjunction with a pre-established social system, and is made relevant to life there in a diversity of ways.
An extreme form of minority position, illustrating some but not all  features
of minorities, is that of pariah groups. These are groups actively rejected
by the host population because of behaviour or characteristics positively condemned,
though often useful in some specific, practical way. European pariah groups
of recent centuries (executioners, dealers in horseflesh and -leather, collectors
of nightsoil, gypsies, etc.) exemplify most features; as breakers of basic taboos
they were rejected by the larger society. Their identity imposed a definition
on social situations which gave very little scope for interaction with persons
in the majority population, and simultaneously as an imperative status represented
an inescapable disability that prevented them from assuming the normal statuses
involved in other definitions of the situation of interaction. Despite these
formidable barriers, such groups do not seem to have developed the internal
complexity that would lead us to regard them as full-fledged ethnic groups;
only the culturally foreign gypsies clearly constitute such a group.
The boundaries of pariah groups are most strongly maintained by the excluding host population, and they are often forced to make use of easily noticeable diacritica to advertise their identity (though since this identity is often the basis for a highly insecure livelihood, such over-communication may sometimes also serve the pariah individual's competitive interests). Where pariahs attempt to pass into the larger society, the culture of the host population is generally well known; thus the problem is reduced to a question of escaping the stigmata of disability by dissociating with the pariah community and faking another origin.
Many minority situations have a trace of this active rejection by the host population. But the general feature of all minority situations lies in the organization of activities and interaction: In the total social system, all sectors of activity are organized by statuses open to members of the majority group, while the status system of the minority has only relevance to relations within the minority and only to some sectors of activity, and does not provide a basis for action in other sectors, equally valued in the minority culture. There is thus a disparity between values and organizational facilities: prized goals are outside the field organized by the minority's culture and categories. Though such systems contain several ethnic groups, interaction between members of the different groups of this kind does not spring from the complementarity of ethnic identities; it takes place entirely within the framework of the dominant, majority group's statuses and institutions, where identity as a minority member gives no basis for action,  though it may in varying degrees represent a disability in assuming the operative statuses. Eidheim's paper gives a very clear analysis of this situation, as it obtains among Coast Lapps.
But in a different way, one may say that in such a poly-ethnic system, the contrastive cultural characteristics of the component groups are located in the nonarticulating sectors of life. For the minority, these sectors constitute a 'backstage' where the characteristics that are stigmatic in terms of the dominant majority culture can covertly be made the objects of transaction.
The present-day minority situation of Lapps has been brought about by recent external circumstances. Formerly, the important context of interaction was the local situation, where two ethnic groups with sufficient knowledge of each other's culture maintained a relatively limited, partly symbiotic relationship based in their respective identities. With the fuller integration of Norwegian society, bringing the northern periphery into the nation-wide system, the rate of cultural change increased drastically. The population of Northern Norway became increasingly dependent on the institutional system of the larger society, and social life among Norwegians in Northern Norway was increasingly organized to pursue activities and obtain benefits within the wider system. This system has not, until very recently, taken ethnic identity into account in its structure, and until a decade ago there was practically no place in it where one could participate as a Lapp. Lapps as Norwegian citizens, on the other hand, are perfectly free to participate, though under the dual disability of peripheral location and inadequate command of Norwegian language and culture. This situation has elsewhere, in the inland regions of Finnmark, given scope for Lappish innovators with a political program based on the ideal of ethnic pluralism (cf. Eidheim 1967), but they have gained no following in the Coast Lapp area here discussed by Eidheim. For these Lapps, rather, the relevance of Lappish statuses and conventions decreases in sector after sector (cf. Eidheim 1966), while the relative inadequacy of performance in the widest system brings about frustrations and a crisis of identity.
Culture contact and change
This is a very widespread process under present conditions as dependence on the products and institutions of industrial societies spreads in all parts of the world. The important thing to recognize is that a drastic reduction of cultural differences between ethnic groups does  not correlate in any simple way with a reduction in the organizational relevance of ethnic identities, or a breakdown in boundary-maintaining processes. This is demonstrated in much of the case material.
We can best analyse the interconnection by looking at the agents of change: what strategies are open and attractive to them, and what are the organizational implications of different choices on their part? The agents in this case are the persons normally referred to somewhat ethno-centrically as the new elites: the persons in the less industrialized groups with greater contact and more dependence on the goods and organizations of industrialized societies. In their pursuit of participation in wider social systems to obtain new forms of value they can choose between the following basic strategies: (i) they may attempt to pass and become incorporated in the pre-established industrial society and cultural group; (ii) they may accept a 'minority' status, accommodate to and seek to reduce their minority disabilities by encapsulating all cultural differentiae in sectors of non-articulation, while participating in the larger system of the industrialized group in the other sectors of activity; (iii) they may choose to emphasize ethnic identity, using it to develop new positions and patterns to organize activities in those sectors formerly not found in their society, or inadequately developed for the new purposes. If the cultural innovators are successful in the first strategy, their ethnic group will be denuded of its source of internal diversification and will probably remain as a culturally conservative, low-articulating ethnic group with low rank in the larger social sytem. A general acceptance of the second strategy will prevent the emergence of a clearly dichotomizing polyethnic organization, and in view of the diversity of industrial society and consequent variation and multiplicity of fields of articulation probably lead to an eventual assimilation of the minority. The third strategy generates many of the interesting movements that can be observed today, from nativism to new states.
I am unable to review the variables that affect which basic strategy will be adopted, which concrete form it may take, and what its degree of success and cumulative implications may be. Such factors range from the number of ethnic groups in the system to features of the ecologic regime and details of the constituent cultures, and are illustrated in most of the concrete analyses of the following essays. It may be of interest to note some of the forms in which ethnic identity is made organizationally relevant to new sectors in the current situation.
Firstly, the innovators may choose to emphasize one level of identity  among the several provided by the traditional social organization. Tribe, caste, language group, region or state all have features that make them a potentially adequate primary ethnic identity for group reference, and the outcome will depend on the readiness with which others can be led to embrace these identities, and the cold tactical facts. Thus, though tribalism may rally the broadest support in many African areas, the resultant groups seem unable to stand up against the sanctioning apparatus even of a relatively rudimentary state organization.
Secondly, the mode of organization of the ethnic group varies, as does the interethnic articulation that is sought. The fact that contemporary forms are prominently political does not make them any less ethnic in character. Such political movements constitute new ways of making cultural differences organizationally relevant (Kleivan 1967), and new ways of articulating the dichotomized ethnic groups. The proliferation of ethnically based pressure groups, political parties, and visions of independent statehood, as well as the multitude of subpolitical advancement associations (Sommerfelt 1967) show the importance of these new forms. In other areas, cult-movements or mission-introduced sects are used to dichotomize and articulate groups in new ways. It is striking that these new patterns are so rarely concerned with the economic sector of activities, which is so major a factor in the culture contact situation, apart from the forms of state socialism adopted by some of the new nations. By contrast, the traditional complex poly-ethnic systems have been prominently based on articulation in this sector, through occupational differentiation and articulation at the market place in many regions of Asia and Middle America, or most elaborately, through agrarian production in South Asia. Today, contending ethnic groups not infrequently become differentiated with respect to educational level and attempt to control or monopolize educational facilities for this purpose (Sommerfelt l967), but this is not so much with a view to occupational differentiation as because of the obvious connection between bureaucratic competence and opportunities for political advancement. One may speculate that an articulation entailing complex differentiation of skills, and sanctioned by the constant dependence on livelihood, will have far greater strength and stability than one based on revocable political affiliation and sanctioned by the exercise of force and political fiat, and that these new forms of poly-ethnic systems are probably inherently more turbulent and unstable than the older forms. 
When political groups articulate their opposition in terms of ethnic criteria, the direction of cultural change is also affected. A political confrontation can only be implemented by making the groups similar and thereby comparable, and this will have effect on every new sector of activity which is made politically relevant. Opposed parties thus tend to become structurally similar, and differentiated only by a few clear diacritics. Where ethnic groups are organized in political confrontation in this way, the process of opposition will therefore lead to a reduction of the cultural differences between them.
For this reason, much of the activity of political innovators is concerned with the codification of idioms: the selection of signals for identity and the assertion of value for these cultural diacritics, and the suppression or denial of relevance for other differentiae. The issue as to which new cultural forms are compatible with the native ethnic identity is often hotly contended, but is generally settled in favour of syncretism for the reasons noted above. But a great amount of attention may be paid to the revival of select traditional culture traits, and to the establishment of historical traditions to justify and glorify the idioms and the identity.
The interconnection between the diacritics that arc chosen for emphasis, the boundaries that are defined, and the differentiating values that are espoused, constitute a fascinating field for study. Clearly, a number of factors are relevant. Idioms vary in their appropriateness for different kinds of units. They are unequally adequate for the innovator's purposes, both as means to mobilize support and as supports in the strategy of confrontation with other groups. Their stratificational implications both within and between groups are important: they entail different sources and distributions of influence within the group, and different claims to recognition from other groups through suppression or glorification of different forms of social stigmata. Clearly, there is no simple connection between the ideological basis of a movement and the idioms chosen; yet both have implications for subsequent boundary maintenance, and the course of further change.
Variations in the selling for ethnic relations
These modern variants for poly-ethnic organization emerge in a world of bureaucratic administration, developed communications, and progressive urbanization. Clearly, under radically different circumstances, the critical factors in the definition and maintenance of ethnic boundaries  would be different. In basing ourselves on limited and contemporary data, we are faced with difficulties in generalizing about ethnic processes, since major variables may be ignored because they are not exhibited in the cases at our disposal. There can be little doubt that social anthropologists have tended to regard the rather special situation of colonial peace and external administration, which has formed the backdrop of most of the influential monographs, as if this were representative of conditions at most times and places. This may have biased the interpretation both of pre-colonial systems and of contemporary, emergent forms. The attempt in these essays to cover regionally very diverse cases is not alone an adequate defence against such bias, and the issue needs to be faced directly.
Colonial regimes are quite extreme in the extent to which the administration and its rules are divorced from locally based social life. Under such a regime, individuals hold certain rights to protection uniformly through large population aggregates and regions, far beyond the reach of their own social relationships and institutions. This allows physical proximity and opportunities for contact between persons of different ethnic groups regardless of the absence of shared understandings between them, and thus clearly removes one of the constraints that normally operate on interethnic relations. In such situations, interaction can develop and proliferate indeed, only those forms of interaction that are directly inhibited by other factors will be absent and remain as sectors of non-articulation. Thus ethnic boundaries in such situations represent a positive organization of social relations around differentiated and complementary values, and cultural differences will tend to be reduced with time and approach the required minimum.
In most political regimes, however, where there is less security and people live under a greater threat of arbitrariness and violence outside their primary community, the insecurity itself acts as a constraint on inter-ethnic contacts. In this situation, many forms of interaction between members of different ethnic groups may fail to develop, even though a potential complementarity of interests obtains. Forms of interaction may be blocked because of a lack of trust or a lack of opportunities to consummate transactions. What is more, there are also internal sanctions in such communities which tend to enhance overt conformity within and cultural differences between communities. If a person is dependent for his security on the voluntary and spontaneous support of his own community, selfidentification as a member  of this community needs to be explicitly expressed and confirmed; and any behaviour which is deviant from the standard may be interpreted as a weakening of the identity, and thereby of the bases of security. In such situations, fortuitous historical differences in culture between different communities will tend to perpetuate themselves without any positive organizational basis; many of the observable cultural differentiae may thus be of very limited relevance to the ethnic organization.
The processes whereby ethnic units maintain themselves are thus clearly affected, but not fundamentally changed, by the variable of regional security. This can also be shown by an inspection of the cases analysed in these essays, which represent a fair range from the colonial to the poly-centric, up to relatively anarchic situations. It is important, however, to recognize that this background variable may change very rapidly with time, and in the projection of long-range processes this is a serious difficulty. Thus in the Fur case, we observe a situation of externally maintained peace and very small-scale local political activity, and can form a picture of interethnic processes and even rates in this setting. But we know that over the last few generations, the situation has varied from one of Baggara-Fur confrontation under an expansive Fur sultanate to a nearly total anarchy in Turkish and Mahdi times; and it is very difficult to estimate the effects of these variations on the processes of nomadization and assimilation, and arrive at any long-range projection of rates and trends.
Ethnic groups and cultural evolution
The perspective and analysis presented here have relevance to the theme of cultural evolution. No doubt human history is a story of the development of emergent forms, both of cultures and societies. The issue in anthropology has been how this history can best be depicted, and what kinds of analyses are adequate to discover general principles in the courses of change. Evolutionary analysis in the rigorous sense of the biological fields has based its method on the construction of phyletic lines. This method presumes the existence of units where the boundaries and the boundary-maintaining processes can be described, and thus where the continuity can be specified. Concretely, phyletic lines are meaningful because specific boundaries prevent the interchange of genetic material; and so one can insist that the reproductive isolate is the unit, and that it has maintained an identity undisturbed by the changes in the morphological characteristics of the species.
I have argued that boundaries are also maintained between ethnic units, and that consequently it is possible to specify the nature of continuity and persistence of such units. These essays try to show that ethnic boundaries are maintained in each case by a limited set of cultural features. The persistence of the unit then depends on the persistence of these cultural differentiae, while continuity can also be specified through the changes of the unit brought about by changes in the boundary-defining cultural differentiae.
However, most of the cultural matter that at any time is associated with a human population is not constrained by this boundary: it can vary, be learnt, and change without any critical relation to the boundary maintenance of the ethnic group. So when one traces the history of a ethnic group through time, one is not simultaneously, in the same sense, tracing the history of 'a culture': the elements of the present culture of that ethnic group have not sprung from the particular set that constituted the group's culture at a previous time, whereas the group has a continual organizational existence with boundaries (criteria of membership) that despite modifications have marked off a continuing unit.
Without being able to specify the boundaries of cultures. it is not possible to construct phyletic lines in the more rigorous evolutionary sense. But from the analysis that has been argued here, it should be possible to do so for ethnic groups, and thus in a sense for those aspects of culture which have this organizational anchoring.
1. The emphatic ideological denial of the primacy of ethnic identity (and rank) which characterises the universal religions that have arisen in the Middle East is understandable in this perspective, since practically any movement for social or ethical reform in the poly-ethnic societies of that region would clash with conventions and standards of ethnic character.
2. The difference between ethnic groups and social strata, which seems problematical at this stage of the argument, will be taken up below.
3. I am here concerned only with individual failure to maintain identity, where most members do so successfully, and not with the broader questions of cultural vitality and anomie.
4. As opposed to presumptive classification in passing social encounters I am thinking of the person in his normal social context where others have a considerable amount of previous information about him, not of the possibilities afforded occasionally for mispresenting one's identity towards strangers.
5. The condemned behaviour which gives pariah position to the gypsies is compound, but rests prominently on their wandering life, originally in contrast to the serf bondage of Europe, later in their flagrant violation of puritan ethics of responsibility, toil and morality.
6. To my knowledge, Mitchell's essay on the Kalela dance (Mitchell 1956) is the first and still the most penetrating study on this topic.