This website will be closed down during 2014.

Thomas Hylland Eriksen's new site is now open at

You are now being redirected to the new Engaging with the world site.
























Ethnicity and Nationalism

Anthropological Perspectives

Thomas Hylland Eriksen

Pluto Press 1993 (second, expanded edition 2002, third edition 2010)

Go directly to Chapter 1

Detailed table of contents


1. What is ethnicity?

The term itself
Ethnicity, race and nation
Ethnicity and class
The current concern with ethnicity
From tribe to ethnic group
What is ethnicity?
"Kinds" of ethnic relations?
Analytical concepts and "native" concepts

2. Social classification: Us and Them

The ecology of the city
The melting-pot metaphor
Communicating cultural difference
Folk taxonomies, social distance
Dichotomisation and complementarisation
Ethnic stigma
Negotiating identity
Ethnicity for the individual
What are the criteria for ethnicity?

3. Groups and boundaries

Ascription as decisive feature of ethnicity
Boundary maintenance
Boundary transcendence
Degrees of ethnic incorporation
Ethnicity as resource competition
Levels of ethnicity
The theory of plural societies
Ethnicity and rank
The interrelationship between criteria
Some controversies

4. Ethnic identity and ideology

Order in the social universe
Analogic and digital; we and us
The emergence of ethnic identites
The creation of an ancestral identity
History and ideology
Social factors in identity processes
Does European identity exist?
What do identities do?

5. Ethnicity in history

The historical development of ethnic relations
Expansions of system boundaries
"Black" ethnogenesis
Indians in new worlds
"Ethnic revitalisation": From people to a people
Colonialism and migration
The semantics of ethnic labelling
Modern education and ethnic identity
Ethnicity, history and culture
Cultural history in ethnic identity

6. Nationalism

The race to nation
What is nationalism?
The nation as a cultural community
The political use of cultural symbols
Nationalism and industrial society
Communication technology and nationhood
Nationalism as religion and metaphoric kinship
The nation-state
Nationalism against the state
Nationalism and the Other
The problem of identity boundaries
A non-ethnic nationalism?
Nationalism and ethnicity reconsidered

7. Minorities and the state

Minorities and majorities
Minorities and the state
The creation of minorities in the modern world
Indigenous peoples
Territorial conflict
Stages in ethnogenesis
Factors in indigenous ethnogenesis
Urban minorities
Boundary processes
Does culture matter among urban migrants?
Identities and "cultures"
Ethnicity in the USA: Race, class and language
The paradox of multiculturalism
Public controversies over culture
Minorities and modernity

8. That which is not ethnic

Changes in social theory
Changes in the social world
Globalisation and localisation
Identities and loyalties
Gender and ethnicity
Beyond ethnicity?
The end of ethnicity?
The eye of the beholder


1. What is ethnicity?

It takes at least two somethings to create a difference. (...) Clearly each alone is - for the mind and perception - a non-entity, a non-being. Not different from being, and not different from non-being. An unknowable, a Ding an sich, a sound from one hand clapping.

Gregory Bateson (1979: 78)

Words like "ethnic groups", "ethnicity" and "ethnic conflict" have become quite common terms in the English language, and they keep cropping up in the press, in TV news, in political programmes and in casual conversations. The same can be said for "nation" and "nationalism", and many of us have to admit that the meaning of these terms frequently seems ambiguous and vague.

There has been a parallel development in the social sciences. During the 1980s and early 1990s, we have witnessed an explosion in the growth of scholarly publications on ethnicity and nationalism, particularly in the fields of political science, history, sociology and social anthropology.

In the case of social anthropology, ethnicity has been a main preoccupation since the late 1960s, and it remains a central focus for research in the 1990s. In this book, the importance of anthropological approaches to the study of ethnicity will be emphasised. Through its dependence on long-term fieldwork, anthropology has the advantage of generating first-hand knowledge of social life at the level of everyday interaction. To a great extent, this is the locus where ethnicity is created and re-created. Ethnicity emerges and is made relevant through ongoing social situations and encounters, and through people's ways of coping with the demands and challenges of life. From its vantage-point right at the centre of local life, social anthropology is in a unique position to investigate these processes. Anthropological approaches also enable us to explore the ways in which ethnic relations are being defined and perceived by people; how they talk and think about their own group as well as other groups, and how particular world-views are being maintained or contested. The significance of ethnic membership to people can best be investigated through that detailed on-the-ground research which is the hallmark of anthropology. Finally, social anthropology, being a comparative discipline, studies both differences and similarities between ethnic phenomena. It thereby provides a nuanced and complex vision of ethnicity in the contemporary world.

An important reason for the current academic interest in ethnicity and nationalism is the fact that such phenomena have become so visible in many societies that it has become impossible to ignore them. In the early twentieth century, many social theorists held that ethnicity and nationalism would decrease in importance and eventually vanish as a result of modernisation, industrialisation and individualism. This never came about. On the contrary, ethnicity and nationalism have grown in political importance in the world, particularly since the Second World War.

Thirty-five of the thirty-seven major armed conflicts in the world in 1991 were internal conflicts, and most of them - from Sri Lanka to Northern Ireland - could plausibly be described as ethnic conflicts. In addition to violent ethnic movements, there are also many important non-violent ethnic movements, such as the Québecois independence movement in Canada. In many parts of the world, further, nation-building - the creation of political cohesion and national identity in former colonies - is high on the political agenda. Ethnic and national identities also become strongly pertinent following the continuous influx of labour migrants and refugees to Europe and North America, which has led to the establishment of new, permanent ethnic minorities in these areas. During the same period, indigenous populations such as Inuits ("Eskimos") and Sami ("Lapps") have organised themselves politically, and demand that their ethnic identities and territorial entitlements should be recognised by the State. Finally, the political turbulence in Europe has moved issues of ethnic and national identities to the forefront of political life. At one extreme of the continent, the erstwhile Soviet Union has split into over a dozen ethnically based states. With the disappearance of the strong Socialist state in the countries of Central and Eastern Europe, issues of nationhood and minority problems are emerging with unprecedented force. On the other extreme of the continent, the situation seems to be the opposite, as the nation-states of Western Europe are moving towards a closer economic, political and possibly cultural integration. But here, too, national and ethnic identities have become important issues in recent years. Many people fear the loss of their national or ethnic identity as a result of a tight European integration, whereas others consider the possibilities for a pan-European identity to replace the ethnic and national ones. During the electoral campaign preceding the Danish referendum on European Union in June 1992, a main anti-EU slogan was: "I want a country to be European in". This slogan suggests that personal identities are intimately linked with political processes and that social identities, e.g. as Danes or Europeans, are not given once and for all, but are negotiated over. Both of these insights are crucial to the study of ethnicity.
This book will show how social anthropology can shed light on concrete issues of ethnicity; which questions social anthropologists ask in relation to ethnic phenomena, and how they proceed to answer them. In this way, the book will offer a set of conceptual tools which go far beyond the immediate interpretation of day-to-day politics in their applicability. Some of the questions which will be discussed are:

· How do ethnic groups remain distinctive under different social conditions?
· Under which circumstances does ethnicity become important?
· What is the relationship between ethnic identity and ethnic political organisation?
· Is nationalism always a form of ethnicity?
· What is the relationship between ethnicity and other types of identity, social classification and political organisation, such as class and gender?
· What happens to ethnic relations when societies are industrialised?
· In which ways can history be important in the creation of ethnicity?
· What is the relationship between ethnicity and culture?

This introductory chapter will present the main concepts to be used throughout the book. It also explores their ambiguities and in this way introduces some main theoretical issues.

The term itself

"Ethnicity seems to be a new term", state Nathan Glazer and Daniel Moynihan (1975: 1), who point to the fact that the term's earliest dictionary appearance is in the Oxford English Dictionary in 1972. Its first usage is attributed to the American sociologist David Riesman in 1953. The word "ethnic", however, is much older. The word is derived from the Greek ethnos (which in turn derived from the word ethnikos), which originally meant heathen or pagan (R. Williams, 1976: 119). It was used in this sense in English from the mid-14th century until the mid-19th century, when it gradually began to refer to "racial" characteristics. In the United States, "ethnics" came to be used around the Second World War as a polite term referring to Jews, Italians, Irish and other people considered inferior to the dominant group of largely British descent. None of the founding fathers of sociology and social anthropology - with the partial exception of Max Weber - granted ethnicity much attention.

Since the 1960s, ethnic groups and ethnicity have become household words in Anglophone social anthropology, although, as Ronald Cohen (1978) has remarked, few of those who use the terms bother to define them. In the course of this book, I shall examine a number of approaches to ethnicity. Most of them are closely related, although they may serve different analytical purposes. All of the approaches agree that ethnicity has something to do with the classification of people and group relationships.

In everyday language, the word ethnicity still has a ring of "minority issues" and "race relations", but in social anthropology, it refers to aspects of relationships between groups which consider themselves, and are regarded by others, as being culturally distinctive. Although it is true that "the discourse concerning ethnicity tends to concern itself with subnational units, or minorities of some kind or another" (Chapman et al., 1989: 17), majorities and dominant peoples are no less "ethnic" than minorities. This will be particularly evident in Chapters 6 and 7, which discuss nationalism and minority-majority relationships.

Ethnicity, race and nation

A few words must be said initially about the relationship between ethnicity and "race". The term race has deliberately been placed within inverted commas in order to stress that it has dubious descriptive value. Whereas it was for some time common to divide humanity into four main races, modern genetics tends not to speak of races, and this has two main reasons. First, there has always been so much interbreeding between human populations that it would be meaningless to talk of fixed boundaries between races. Secondly, the distribution of hereditary physical traits does not follow clear boundaries. In other words, there is often greater variation within a "racial" group than there is systematic variation between two groups.

Concepts of race can nevertheless be important to the extent that they inform people's actions; at this level, race exists as a cultural construct, whether it has a "biological" reality or not. Racism, obviously, builds on the assumption that personality is somehow linked with hereditary characteristics which differ systematically between "races", and in this way race may assume sociological importance even if it has no "objective" existence. Social scientists who study race relations in Great Britain and the United States need not themselves believe in the existence of race, since their object of study is the social and cultural relevance of the notion that race exists. If influential people in a society had developed a similar theory about the hereditary personality traits of redhaired people, and if that theory gained social and cultural significance, "redhead studies" would for similar reasons have become a field of academic research, even if the researchers themselves did not agree that redheads were different from others in a relevant way. In societies where they are important, ideas of race may therefore be studied as part of local discourses on ethnicity.

Should the study of race relations, in this meaning of the word, be distinguished from the study of ethnicity or ethnic relations? Pierre van den Berghe (1983) does not think so, but would rather regard "race" relations as a special case of ethnicity. Others, among them Michael Banton (1967), have argued the need to distinguish between race and ethnicity. In Banton's view, race refers to the categorisation of people, while ethnicity has to do with group identification. He argues that "ethnicity is generally more concerned with the identification of 'us', while racism is more oriented to the categorisation of 'them'" (cf. Jenkins, 1986: 177). However, ethnicity can assume many forms, and since ethnic ideologies tend to stress common descent among their members, the distinction between race and ethnicity is a problematic one, even if Banton's distinction between groups and categories can be useful (cf. Chapter 3). I shall not, therefore, distinguish between race relations and ethnicity. Ideas of "race" may or may not form part of ethnic ideologies, and their presence or absence does not seem a decisive factor in interethnic relations.

Discrimination on ethnic grounds is spoken of as "racism" in Trinidad and as "communalism" in Mauritius (Eriksen, 1992a), but the forms of imputed discrimination referred to can be nearly identical. On the other hand, it is doubtless true that groups who look different from majorities or dominating groups may be less liable to become assimilated into the majority than others, and that it can be difficult for them to escape from their ethnic identity if they wish to. However, this may also hold good for minority groups with, say, an inadequate command of the dominant language. In both cases, their ethnic identity becomes an imperative status, an ascribed aspect of their personhood from which they cannot escape entirely. Race or skin colour as such is not the decisive variable in every society.

The relationship between the terms ethnicity and nationality is nearly as complex as that between ethnicity and race. Like the words ethnic and race, the word nation has a long history (R. Williams, 1976: 213-214), and has been used in a variety of different meanings in English. We shall refrain from discussing these meanings here, and will concentrate on the sense in which nation and nationalism are used analytically in academic discourse. Like ethnic ideologies, nationalism stresses the cultural similarity of its adherents, and by implication, it draws boundaries vis-a-vis others, who thereby become outsiders. The distinguishing mark of nationalism is by definition its relationship to the state. A nationalist holds that political boundaries should be coterminous with cultural boundaries, whereas many ethnic groups do not demand command over a state. When the political leaders of an ethnic movement place demands to this effect, the ethnic movement therefore by definition becomes a nationalist movement. Although nationalisms tend to be ethnic in character, this is not necessarily the case, and we shall look more carefully into the relationship between ethnicity and nationalism in Chapters 6-7.

Ethnicity and class

The term ethnicity refers to relationships between groups whose members consider themselves distinctive, and these groups may be ranked hierarchically within a society. It is therefore necessary to distinguish clearly between ethnicity and social class.

In the literature of social science, there are two main definitions of classes. One derives from Karl Marx, the other from Max Weber. Sometimes elements from the two definitions are combined.

The Marxist view of social classes emphasises economic aspects. A social class is defined according to its relationship to the productive process in society. In capitalist societies, according to Marx, there are three main classes. First, there is the capitalist class or bourgeoisie, whose members own the means of production (factories, tools and machinery, etc.) and buy other people's labour-power (i.e. employ them). Secondly, there is the petit-bourgeoisie, whose members own means of production but do not employ others. Owners of small shops are typical examples. The third, and most numerous class, is the proletariat or working class, whose members depend upon selling their labour-power to a capitalist for their livelihood. There are also other classes, notably the aristocracy, whose members live by land interest, and the lumpenproletariat, which consists of unemployed and underemployed people - vagrants and the like.

Since Marx' time in the mid-nineteenth century, the theory of classes has been developed in several directions. Its adherents nevertheless still stress the relationship to property in their delineation of classes. A further central feature of this theory is the notion of class struggle. Marx and his followers held that oppressed classes would eventually rise against the oppressors, overthrow them through a revolution, and alter the political order and the social organisation of labour. This, in Marx' view, was the main way in which societies evolved.

The Weberian view of social classes, which has partly developed into theories of social stratification, combines several criteria in delineating classes, including income, education and political influence. Unlike Marx, Weber did not regard classes as potential corporate groups; he did not believe that members of social classes necessarily would have shared political interests. Weber preferred to speak of status groups rather than classes.

Theories of social class always refer to systems of social ranking and distribution of power. Ethnicity, on the contrary, does not necessarily refer to rank; ethnic relations may well be egalitarian in this regard. Still, many poly-ethnic societies are ranked according to ethnic membership. The criteria for such ranking are nevertheless different from class ranking: they refer to imputed cultural differences or "races", not to property or achieved statuses.

There may be a high correlation between ethnicity and class, which means that there is a high likelihood that persons belonging to specific ethnic groups also belong to specific social classes. There can be a significant interrelationship between class and ethnicity, both class and ethnicity can be criteria for rank, and ethnic membership can be an important factor for class membership. Both class differences and ethnic differences can be pervasive features of societies, but they are not one and the same thing and must be distinguished from one another analytically.

The current concern with ethnicity

If one runs a word search programme through a representative sample of English-language anthropological publications since 1950, one will note significant changes in the frequency of a number of keywords. Words like "structure" and "function", for example, have gradually grown unfashionable, whereas Marxist terms like "base and superstructure", "means of production" and "class struggle" were popular from around 1965 until the early 1980s. Terms like "ethnicity", "ethnic" and "ethnic group", for their part, have steadily grown in currency since the mid- to late 1960s. There may be two main causes for this. One of them is change in the social world, while the other concerns changes in the dominant way of thinking in social anthropology.

Whereas classical social anthropology, as exemplified in the works of Malinowski, Boas, Radcliffe-Brown, Lévi-Strauss, Evans-Pritchard and others, would characteristically focus on single "tribal" societies, changes in the world after the Second World War have brought many of these societies into increased contact with each other, with the state and with global society. Many of the peoples studied by social anthropologists have become involved in national liberation movements or ethnic conflicts in post-colonial states. Many of them, formerly regarded as "tribes" or "aboriginals", have become "ethnic minorities". Many former members of tribal or traditional groups have also migrated to Europe or North America, where their relationships with the host societies have been studied extensively by sociologists, social psychologists and social anthropologists.

Some ethnic groups have moved to towns or regional centres where they are brought into contact with people with other customs, languages and identities, and where they frequently enter into competitive relationships in politics and the labour market. Frequently, people who migrate try to maintain their old kinship and neighbourhood social networks in the new urban context, and both ethnic quarters and ethnic policical groupings often emerge in such urban settings. Although the speed of social and cultural change can be high, people tend to retain their ethnic identity despite having moved to a new environment. This kind of social change has been investigated in a series of pioneering studies in North American cities from the 1920s and in Southern Africa from the early 1940s, and we will return to these studies in the next chapter.

In an influential study of ethnic identity in the United States, Glazer and Moynihan (1963) stated that the most important point to be made about the "American melting-pot" is that it never occurred. They argue that rather than eradicating ethnic differences, modern American society has actually created a new awareness in people, a concern about roots and origins. Moreover, many Americans continue to use their ethnic networks actively when looking for jobs or a spouse. Many prefer to live in neighbourhoods dominated by people with the same origins as themselves, and they continue to regard themselves as "Italians", "Poles" etc., in addition to being Americans - two generations or more after their ancestors left the country of origin.

A main insight from anthropological research has been that ethnic organisation and identity, rather than being "primordial" phenomena radically opposed to modernity and the modern state, are frequently reactions to processes of modernisation. As Jonathan Friedman has put it, "[e]thnic and cultural fragmentation and modernist homogenization are not two arguments, two opposing views of what is happening in the world today, but two constitutive trends of global reality" (Friedman, 1990: 311).

Does this mean that ethnicity is chiefly a modern phenomenon? This is a tricky, but highly relevant question. The contemporary ethnic processes referred to above can be described as modern in character. In an influential statement on political ethnicity, Abner Cohen (1974a) has argued that the concept is perhaps most useful in the study of the development of new political cultures in situations of social change in the "Third World". However, important studies of ethnicity have been carried out in non-modern, non-Western societies as well.

The contemporary concern with ethnicity and ethnic processes is partly related to historical changes such as the ones mentioned above. It could nevertheless also be argued that the growing interest in ethnicity reflects changes in the dominant anthropological mode of thought. Instead of viewing "societies" or even "cultures" as more or less isolated, static and homogeneous units as the early structural-functionalists would tend to, many anthropologists now try to depict flux and process, ambiguity and complexity in their analyses of social worlds. In this context, ethnicity has proven a highly useful concept, since it suggests a dynamic situation of variable contact and mutual accomodation between groups.

From tribe to ethnic group

As mentioned, there has been a shift in Anglophone social anthropological terminology concerning the nature of the social units we study. While one formerly spoke of "tribes", the term "ethnic group" is nowadays much more common. Ronald Cohen remarks: "Quite suddenly, with little comment or ceremony, ethnicity is an ubiquitous presence" (R. Cohen, 1978: 379). This switch in terminology implies more than a mere replacement of a word with another. Notably, the use of the term "ethnic group" suggests contact and interrelationship. To speak of an ethnic group in total isolation is as absurd as to speak of the sound from one hand clapping (cf. Bateson, 1979: 78). By definition, ethnic groups remain more or less discrete from each other, but they are aware of - and in contact with - members of other ethnic groups. Moreover, these groups or categories are in a sense created through that very contact. Group identities must always be defined in relation to that which they are not - in other words, in relation to non-members of the group.

The terminological switch from "tribe" to "ethnic group" may also mitigate or even transcend an ethnocentric or Eurocentric bias which anthropologists have often been accused of promoting covertly. When we talk of tribes, we implicitly introduce a sharp, qualitative distinction between ourselves and the people we study; the distinction generally corresponds to the distinction between modern and traditional or "primitive" societies. If we instead talk of ethnic groups or categories, such a sharp distinction becomes difficult to maintain. Virtually every human being belongs to an ethnic group, whether he or she lives in Europe, Melanesia or Central America. There are ethnic groups in English cities, in the Bolivian countryside and in the New Guinea highlands.

Anthropologists themselves belong to ethnic groups or nations. Moreover, the concepts and models used in the study of ethnicity can often be applied to modern as well as non-modern contexts, to Western as well as non-Western societies. In this sense, the concept of ethnicity can be said to bridge two important gaps in social anthropology: it entails a focus on dynamics rather than statics, and it relativises the boundaries between "us" and "them", between moderns and tribals.

What is ethnicity?

When we talk of ethnicity, we indicate that groups and identities have developed in mutual contact rather than in isolation. But what is the nature of such groups?

When A. L. Kroeber and Clyde Kluckhohn investigated the various meanings of culture in the early 1950s (Kroeber & Kluckhohn, 1952), they found about three hundred different definitions. Although Ronald Cohen is correct in stating that most of whose who write on ethnicity do not bother to define the term, the extant number of definitions is already high - and it is growing (B. Williams, 1989a). Instead of going through the various definitions of ethnicity here, I will point out significant differences between theoretical perspectives as we go along. As a starting-point, let us examine the recent development of the term as it is being used by social anthropologists.

The word "ethnic group" has come to mean something like "a people". But what is a people? Does the population of Britain constitute a people, does it comprise several peoples (as Nairn, 1977, tends to argue), or does it rather form part of a Germanic, or an English-speaking, or a European people? All of these positions may have their defenders, and this very ambiguity in the designation of peoples has been taken on as a challenge by anthropologists. In a study of ethnic relations in Thailand, Michael Moerman (1965) asks himself: "Who are the Lue?" The Lue were the ethnic group his research focused on, but when he tried to describe who they were - in which ways they were distinctive from other ethnic groups - he quickly ran into trouble. His problem, a very common one in contemporary social anthropology, concerned the boundaries of the group. After listing a number of criteria commonly used by anthropologists to demarcate cultural groups, such as language, political organisation and territorial contiguity, he states: "Since language, culture, political organization, etc., do not correlate completely, the units delimited by one criterion do not coincide with the units delimited by another" (Moerman, 1965: 1215). When he asked individual Lue what were their typical characteristics, they would mention cultural traits which they in fact shared with other, neighbouring groups. They lived in close interaction with other groups in the area; they had no exclusive livelihood, no exclusive language, no exclusive customs, no exclusive religion. Why was it appropriate to describe them as an ethnic group? After posing these problems, Moerman was forced to conclude that "[s]omeone is Lue by virtue of believing and calling himself Lue and of acting in ways that validate his Lueness" (Moerman, 1965: 1219). Being unable to argue that this "Lueness" can be defined with reference to objective cultural features or clear-cut boundaries, Moerman defines it as an emic category of ascription. This way of delineating ethnic groups has become very influential in social anthropology (cf. Chapter 3).

Does this imply that ethnic groups do not necessarily have a distinctive culture? Can two groups be culturally identical and yet constitute two different ethnic groups? This is a complicated question which will be dealt with at length in later chapters. At this point, we should note that contrary to a widespread commonsense view, cultural difference between two groups is not the decisive feature of ethnicity. Two distinctive, endogamous groups, say, somewhere in New Guinea, may well have widely different languages, religious beliefs and even technologies, but that does not entail that there is an ethnic relationship between them. For ethnicity to come about, the groups must have a minimum of contact between them, and they must entertain ideas of each other as being culturally different from themselves. If these conditions are not fulfilled, there is no ethnicity, for ethnicity is essentially an aspect of a relationship, not a property of a group. Conversely, some groups may seem culturally similar, yet there can be a socially highly relevant (and even volatile) inter-ethnic relationship between them. This would be the case of the relationship between Serbs and Croats following the break-up of Yugoslavia, or of the tension between coastal Sami and Norwegians. There may also be considerable cultural variation within a group without ethnicity (Blom, 1969). Only in so far as cultural differences are perceived as being important, and are made socially relevant, do social relationships have an ethnic element.

Ethnicity is an aspect of social relationship between agents who consider themselves as being culturally distinctive from members of other groups with whom they have a minimum of regular interaction. It can thus also be defined as a social identity (based on a contrast vis-a-vis others) characterised by metaphoric or fictive kinship (Yelvington, 1991: 168). When cultural differences regularly make a difference in interaction between members of groups, the social relationship has an ethnic element. Ethnicity refers both to aspects of gain and loss in interaction, and to aspects of meaning in the creation of identity. In this way, it has a political, organisational aspect as well as a symbolic one.

Ethnic groups tend to have myths of common origin, and they nearly always have ideologies encouraging endogamy, which may nevertheless be of highly varying practical importance.

"Kinds" of ethnic relations?

This very general and tentative definition of ethnicity lumps together a great number of very different social phenomena. My relationship with my Pakistani greengrocer has an ethnic aspect; so, it could be argued, do the war in former Yugoslavia and "race riots" in American cities. Do these phenomena have anything interesting in common, justifying that we compare them within a single conceptual framework? The answer is both yes and no. One of the contentions from anthropological studies of ethnicity is that there may be mechanisms of ethnic processes which are relatively uniform in every inter-ethnic situation: to this effect, we can identify certain shared formal properties in all ethnic phenomena.

On the other hand, there can be no doubt that the substantial social contexts of ethnicity differ enormously, and indeed that ethnic identities and ethnic organisations themselves may have highly variable importance in different societies, for different persons and in different situations. We should nevertheless keep in mind that the point of anthropological comparison is not necessarily to establish similarities between societies, but it can also be to reveal important differences. In order to discover such differences, we must initially possess some kind of measuring-stick, a constant or a conceptual bridgehead, which can be used as a basis of comparison. If we first know what we mean by ethnicity, we can then use the concept as a common denominator for societies and social contexts which are otherwise very different. The concept of ethnicity can in this way not only teach us something about similarity, but also about differences.

Although the concept of ethnicity should always have the same meaning lest it ceases to be useful in comparison, it is inevitable that we distinguish between the social contexts under scrutiny. Some interethnic contexts in different societies are very similar and may seem easily comparable, whereas others differ profoundly. In order to give an idea of the variation, I shall briefly describe some typical empirical foci of ethnic studies, some "kinds of ethnic groups", so to speak. This list is not exhaustive.

(a) Urban ethnic minorities. This category would include, among others, non-European immigrants in European cities and Hispanics in the United States, as well as migrants to industrial towns in Africa and elsewhere. Research on immigrants has focused on problems of adaptation, on ethnic discrimination from the host society, racism, and issues relating to identity management and cultural change (cf. Chapters 4 and 7). Anthropologists who have investigated urbanisation in Africa have focused on change and continuity in political organisation and social identity following migration to totally new settings (cf. Chapter 2). Although they have political interests, these ethnic groups rarely demand political independence or statehood, and they are as a rule integrated into a capitalist system of production and consumption.

(b) Indigenous peoples. This word is a blanket term for aboriginal inhabitants of a territory, who are politically relatively powerless and who are only partially integrated into the dominant nation-state. Indigenous peoples are associated with a non-industrial mode of production and a stateless political system (Minority Rights Group, 1990). The Basques of the Bay of Biscay and the Welsh of Great Britain are not considered indigenous populations, although they are certainly as indigenous, technically speaking, as the Sami of northern Scandinavia or the Jívaro of the Amazon basin. The concept "indigenous people" is thus not an accurate analytical one, but rather one drawing on broad family resemblances and contemporary political issues (cf. Chapters 4 and 7).

(c) Proto-nations ("ethnonationalist" movements). These groups, the most famous of ethnic groups in the news media of the 1990s, include Kurds, Sikhs, Palestinians and Sri Lankan Tamils, and their number is growing. By definition, these groups have political leaders who claim that they are entitled to their own nation-state and should not be "ruled by others". These groups, short of having a nation-state, may be said to have more substantial characteristics in common with nations (cf. Chapter 6) than with either urban minorities or indigenous peoples. They are always territorially based; they are differentiated according to class and educational achievement, and they are large groups. In accordance with common terminology, these groups may be described as "nations without a state". Anthropologists have studied such movements in a number of societies, including Euzkadi or Basque Country (Heiberg, 1989), Brittany (McDonald, 1989) and Québec (Handler, 1988).

(d) Ethnic groups in "plural societies". The term "plural society" usually designates colonially created states with culturally heterogeneous populations (Furnivall, 1948; M. G. Smith, 1965). Typical plural societies would be Kenya, Indonesia and Jamaica. The groups that make up the plural society, although they are compelled to participate in uniform political and economic systems, are regarded as (and regard themselves as) highly distinctive in other matters. In plural societies, secessionism is usually not an option, and ethnicity tends to be articulated as group competition. As Richard Jenkins (1986) has remarked, most contemporary states could plausibly be considered plural ones.

The definition of ethnicity proposed earlier would include all of these "kinds" of groups, no matter how different they are in other respects. Surely, there are aspects of politics (gain and loss in interaction) as well as meaning (social identity and belonging) in the ethnic relations reproduced by urban minorities, indigenous peoples, proto-nations and the component groups of "plural societies" alike. Despite the great variations between the problems and substantial characteristics represented by the respective kinds of groups, the word ethnicity may, in other words, meaningfully be used as a common denominator for them. In later chapters, it will be shown how anthropological approaches to ethnicity may shed light on both similarities and differences between different social contexts and historical circumstances.

Analytical concepts and "native" concepts

The final problem to be discussed in this chapter concerns the relationship between anthropological concepts and their subject-matter. This is a problem with complicated ramifications, and it concerns the relationships between (i) anthropological theory and "native theory", (ii) anthropological theory and social organisation, and (iii) "native theory" and social organisation.

It can be argued that the terminological shift from "tribe" to "ethnic group" mitigated the formerly strong distinction between "moderns" and "primitives". The growing anthropological interest in nationalism entails a further step towards "studying ourselves". For if ethnicity can be non-modern as well as modern, nationalism must be identified with the modern age, with the French Enlightenment and German Romanticism as parallel starting-points. Nationalist slogans, movements and symbols have later penetrated into the heartlands of anthropological research. Nationalism, being a modern state ideology, is present in the social worlds in which the anthropologists themselves live. Although there are interesting differences between particular nationalisms, nationalism as such is a modern ideology. When studying nationalism in a foreign country, it is therefore difficult to use one's own society as an implicit contrast as anthropologists have frequently done when studying "exotic" societies. In fact, as Richard Handler (1988) has observed, nationalism and social science, including anthropology, grew out of the same historical circumstances of modernisation, industrialisation and the growth of individualism in the 19th century. For this reason, Handler argues, it has been difficult for anthropologists to attain sufficient analytical distance vis-a-vis nationalisms; the respective concepts and ways of thinking are too closely related (cf. also Herzfeld, 1987; Just, 1989).

Handler's point is also valid in relation to modern ethnopolitical movements. Their spokesmen tend to invoke a concept of culture which is in fact often directly inspired by anthropological concepts of culture, and in some cases they self-consciously present themselves as "tribes" reminiscent of the "tribes" depicted in classical anthropological monographs (Roosens, 1989). In these cases, there is an intrinsic relationship between anthropological theorising and "native theory". Additionally, when anthropologists study contested issues in their own societies, there is a real risk that the scholarly conceptual apparatus will be contaminated by the inaccurate and perhaps ideologically loaded everyday meanings of the words. For this reason, we should be particularly cautious in our choice of analytical terms and interpretations when we study phenomena such as ethnicity and nationalism.

The points made by Handler and others in relation to the study of nationalism and modern ethnopolitics, can nevertheless be seen as general problems of social anthropology. The main problem concerns how to articulate the relationship between anthropological theory, "native theory" and social organisation (Mitchell, 1974). In a sense, ethnicity is created by the analyst when he or she goes out into the world and poses questions about ethnicity. Had one instead been concerned with gender, one would doubtless have found aspects of gender instead of ethnicity. On the other hand, persons or informants who live in the societies in question may themselves be concerned with issues relating to ethnicity, and as such the phenomenon clearly does exist outside of the mind of the observer. But since our concepts, for example ethnicity and nationalism, are our own inventions, we must not assume that the actors themselves have the same ideas about the ways in which the world is constituted - even if they are using the very same words as ourselves! History and social identity are constructed socially, sometimes with a very tenuous relationship with established, or at least official, facts (cf. Chapter 4).

There are often discrepancies between what people say and what they do, and there will nearly always be discrepancies between informants' descriptions of their society and the anthropologist's description of the same society. Indeed, many anthropologists (e.g. Holy & Stuchlik, 1983) hold that it is a main goal of our discipline to investigate and clarify the relationship between notions and actions, or between what people say and what they do. One may disagree with their "rationalist" perspective, which seems to assume that a simple, "economic" means-end rationality underlies all social action, but the general problem remains important: why is it that people say one thing and then proceed to doing something entirely different, and how can this be investigated?

This discrepancy is relevant for ethnic studies, and it requires that we are clear about the distinctions between our own concepts and models, "native" concepts and models, and social process. In some societies, people will perhaps deny that there is systematic differential treatment between members of different groups, although the anthropologist will discover that such discrimination exists. Conversely, I have met many Christians in Mauritius who have sworn, in conversations, that they would (for ostensibly sound reasons) have nothing to do with Muslims; later on, it has turned out that they in fact entertain quite strong and sometimes confidential relationships with Muslims. It is, indeed, frequently contradictions of this kind that lead to anthropological insights.

©Thomas Hylland Eriksen 1993