1Research on nationalism, ethnicity and fundamentalism has become ‑ it seems ‑ an easy task. There are authors we all cite, and there are the objects of our study, which delineate themselves quite clearly. This selfdelineation poses however, as this paper shall argue, problems for social science, because it induces us to overlook 1. processes of switching between different frames of reference, 2. processes of (physical) reproduction (namely the link to families), which does not go without saying, 3. the fuzziness of the limits, which may fulfil specific social functions. The first issue ‑ the process of switching ‑ is rarely used as a reference point in the study of we‑groups, since the actors themselves rather insist on stability of their group. It seems to me however that this odd phenomenon can bring us new insights. We may use it as a plough to work through the garden of ethnicity theory.
2Some of authors in the field, namely those representing the essentialist tendency, have become victims of the discourse of the groups studied, insofar as they believe in the primordiality of emotions and traditions as the basis for the creation and maintenance of we‑groups.
3So‑called ethnic violence in Bosnia, Somalia and Turkey is explained this way. This sounds plausible to us, because the suffering of the victims is expressed in emotions and creates strong emotions in us. But this is only a link of connotations, not a logical one. Violence of the type reported from these countries, social (=organized) violence, with weapons, transport of actors to the field, the premeditation of ambushes etc. requires logistics and presence of mind, not spontaneous emotions (or the vagueness of historical references to traditions). There are central actors like warlords or politic al entrepreneurs to be studied, and we must look at the role of conflict‑chanelling institutions, of economic solidarity and of communication in order to understand the stabilization of we‑groups. Switching becomes an important focus in this field of study not only because understanding cases of failed stabilization helps us to better comprehend conditions of stability, but also because it openly contradicts assumptions about "deep‑rooted traditions" which might condition a given we‑group.
4First, we shall try to understand what might be the usefulness of referring to the complicated concept of "we‑groups" instead of limiting ourselves to nationalism or to fundamentalism.
5It is a common assumption that ethnicity is a case for anthropologists and nationalism one for political scientists. That's wrong ! Both phenomena can be classified as we‑group processes. They share with other we‑group processes, namely class movements and religious movements, the same driving forces and systemic patterns. Thus, there are no separate tools for their analysis, be it the hermeneutics of fieldwork or distanced systemic analysis.
6Examples for motives underlying nationalism and other we‑group processes :
Mega‑identities : migrants losing their frame of reference which provided prestige for them, are attracted by mega‑identities, which compensate for the lost individual one, like Pan‑Germany (GroBdeutschland), world revolution etc.
Clientelism : in competition over new resources it may be of advantage to create a c1ientelist network designed as a "we‑group".
Moral ethnicity : if behaviour seen as antisocial is felt as a personal threat, one option is to create a social form imagined as a community but framed as an organisation which excludes the immoral ones and unites the moral ones 1.
7Two elements underlie, as basic assumptions, the argumentation :
Neither ethnic groups nor nations constitute a "natural" order. They compete in human history with other types of social organisation for the place of the central organizing structure. We can even find social structures where there are no we‑groups of the ethnic type (Elwert 1989).
Nations and ethnic groups are social structures which have to be reproduced. That means, they have to be recreated by each successive generation. They may have to fulfil new functions and may use old plausibilities. But the transmission of plausibilities works only if the functions are satisfied.
8One remarkable feature of we‑groups is the process we call switching. That means a rapid change from one frame of reference to the other. A class movement may become a nationalist one, a nationalist movement transforms itself into a religious mobilization, or a religious network redefines itself as a class movement. Switching processes easily escape scientific treatment because they "change the department" ‑ from religious to social, to political studies and vice versa.
10Sri Lanka's Trotskytes turned their class movement into an ultra‑nationalist militant organisation.
11Iraq's secularist, nationalist Baath party redefined itself under Sadam Hussein as Islamic ‑ and won international support.
12Hamas is about to gain political supremacy by restating the (previously secular nationalist) Palestinian cause as an Islamic one.
13Patrilineal descent groups in the tension field of Afghanistan reacted with identification or by stating difference with the religiously defined powers and thus turned into religious sects (Canfield).
14We (Kristina Kehl & Georg Elwert) studied Alevites in Turkey 2. The first thing we know with some historical certainly, is that there were endogamous groups who wanted to differentiate themselves from the Sunnite Ottoman rule. The Shiite façade and the rule of secrecy covered in fact a wide range of beliefs. Under Atatürk's Republic they became fervent republicans, were over‑represented in the left of center republican and in socialist organizations. The installation of the holy men ("dedeler") who kept the oral traditions was discontinued. The holy books ("buyruk") were given away. But since the Alevites' identity was ascribed to them also from the outside and was maintained by their social environment, they were forced to continue endogamy.
15Now, the new conjuncture of religious prestige symbols and boundary definitions, the self declared moral rearmament of fundamentalism, induces them to re‑create a religious "identity". They are very active in retrieving ‑ with the help of foreign anthropologists ‑ old religious folklore, copying religious books ; they confront the painful task of harmonizing the incompatible and some move forward directly to sheer invention of "forgotten" beliefs (with some Iranian help). None of these transformations could alleviate violence against this group.
16The readiness to strike ‑ wrongly called "hatred" ‑ is conditioned by the internal dynamism of the dominant "community", which, in order to regain its self‑image of community, longs for the symbolic affirmation of difference. Because of the heterogeneity of the dominant ones the easiest option for the definition of the "we" is the creation of foes represented as a negative print of the morality, which one would like to have as a self‑description. The Alevites are easy targets for this, just because they happen to be there, not because of this or that feature 3.
17The concept of switching refers a) to alternations between reference frames and b) to moves between different more‑or‑less inclusive conceptions of the group's boundary. The switching may imply a redefinition of the group's boundaries, or these boundaries may be kept and reinterpreted.
18Redefinition of boundaries : A strategically important minority may be included or an "annoying" group of actors excluded (Thus Nazi Germany forged a concept of "Volksdeutsche"/ethnic Germans, which included the Austrian origin "Sudeten‑Deutsche" and the Dutch origin Mennonites of Russia and Ukrania and excluded the Jewish population of East Europe, which had been the main organizer of German schools and other cultural organizations in these countries and whose German ethnicity was hailed and instrumentalized by the German army's high command (Oberkommando der Wehrmacht) in the First World War). A segment of a larger group may discover their "real" identity or several groups may "re"‑unite under one umbrella.
19Redefinition of boundaries kept : In the case that the boundaries .are kept, the change of reference level brings other actors or resources into play or pushes them to the background where they risk being forgotten. Sensible (or shrewd) political actors preempt changes, takins the lead.
20Another example :
21Among the groups marginalized by dominating Christian and Islamic groups in the Horn of Africa emerged a label "Galla" or "Oromo" which was linked to a specific mode of cohesion 4. A generation and age class system allowed for the integration of neighbours. The expansion of the very militant Oromo groups along with and against the Ethiopian empire was a constant undercurrant of history at the Horn of Africa ‑ till they became, in the 20th century, the most extended ethnic label of Abessinia. Modern civil war put them under a different stress. Though one can still observe "Oromization" (Thomas Zitelmann) in refugee camps, in other situations the "identity" (belongingness) is redefined by using some (patrilineal) kinship links, (Kushitic) multilinguality, Islam or Christian tainted laicism as a reference in order to become part of a Tigre kindred, a Somali clan, spearhead of an Islamic movement or Ethiopian nationalist.
22The dynamics of maintaining or reshaping boundaries, of defining values and institutions as central or as marginal, is the product of tensions within the socio‑political context. This is valid whether the "identity" sought is the ideologized sentiment of value ascription or the definable characteristics of real or invented cultural traits. Similar contexts of tension will reproduce similar configurations of we‑groups (see also Wimmer). When the crucial political categories are religious, the marginal populations tend to become "heretics" (cf. Canfield). When the crucial political categories are national or ethnic, those marginal populations which can only escape the dynamics of peripherization by marking their difference, show ethnical distinction.
23In order to create difference or to reach incorporation switching may be necessary. This follows a pattern which might be called the "clarity imperative" : there is a need for both : an accepted code and clear markers.
24The code may be religion, ethnicity, local group, kinship. The code may even narrow down to a sub‑code or may widen. So "religion" may be narrowed down to book religion, or book religion to Islam. So the Pashtoo nomads who till the last century were Jews then had to mark their difference within the Islamic code, when only this became acceptable 5. Inversely the Christian Bogumil heresy in the Balkans increased the clarity of difference in reference to their Serb and Croat neighbours by switching to Islam, when conquered by the Ottomans ‑ which created the Bosnians. The codes of "locality" and "kinship", which now are presented as remnants from a past when human beings lived in accordance with nature, are not natural at all. Like age, generation, gender and physical features, locality and kinship are ways to give social order a natural appearance, "to naturalize" it 6. But there is no natural law that made any of these references a strictly observed organizing principle to be found in every human group. "Kinship" may be an idiom expressing also ties of neighbourship, "age" may actually mean generation etc : The code of locality in the definition of we‑groups may be a means of defining land rights ‑ among the Mapuche in Chile as among the Gur speaking groups of northern Ghana, Togo and Bénin "people of the earth" or "people from here" are common self‑given names. In other social environments marriage rights and inheritance are the crucial issues ; there kinship (especia1ly unilateral kinship) is the reference code to define the dominant we‑group structure (not excluding locality etc. as a secondary reference, to which one may switch).
25Changes of shape or parameters of reference can occur within we‑groups with amazing rapidity (which is conditioned by the volatility of power) (Wallerstein). This does not exclude the continuity of "ethnic" names, which is instrumental to any claims of political inheritance rights.
26Ethnic names, a specific type of music (see During on Uzbekistan), traditional norms and so forth constitute the inventory upon which a given group may draw in order to design its limits and its "identity". This inventory may be limited or it may be broad. It will be broad in situations of enhanced interregional communication. This inventory should be distinguished from a closet where one hangs a selection of one's coats, because these symbols are not owned. Some salient marker of distinction seen somewhere else may well be adopted, as a sign familiar from childhood on. Inventories are shared with other collective actors. Thus symbol use may overlap or may even alternate between enemy groups. The acquisition of symbols is, however, not random. It is conditioned by a) plausibility respect with of other markers of order and b) by the communicative need to create salience and difference. An example of the first is the importance of all references which suggest collective rights of past generations in an environment where inheritance is a strong argument for claims. Examples of the second need are too obvious to be detailed here (national dress, orthographic reforms, racial ideologies etc.)
27Toughts on the recurrence of essentialism :
28The phenomenon of switching is uncomfortable for any essentialist theory. Because those "primordial values", which presumably constitute by their very "nature" the boundaries of ethnic groups, suddenly lose relevance. At this point a discussion of the recurrence of essentialism might be appropriate. Has been Max Weber's time it shown that those groups which call themselves "ethnic" or "national" use boundaries in respect to criteria of functionality of the social structure they aimed at, on the one hand, and criteria of plausibility, which "goes without saying" in their respective context, on the other hand. It is society dressed as community, which produces the emotional cohesion of these we‑groups.
29Those who argued against Weber (or later against R. Thumwald, W. Mühlmann, F. Barth etc., who took up his ideas) that there were "natural" forces binding all ethnic groups together ‑ namely language, religion, culture ‑ could be confronted by empirical examples of ethnic groups, recognised as such by all their neighbours, which lacked at least one of these criteria (see the latest demonstration by Thomas Höllmann on Southeast Asian examples, and for the history of ideas Ernst W. Müller). The reason why, spite of all evidences for the formalist perspective, some authors come back to an essentialist view (or try to "harmonize" these logically incompatible concepts), may be due to the fact that we all are working in the environment of common popular ideas. Most ethnic groups try to present themselves as "natural" ; this induces a continuous stream of "immigration of protoscientific ideas" into social science.
30Groups and individuals may belong to different reference groups simultaneously ; according to the opportunity they may stress one or an other of these affiliations as their "real" one (see Geertz, Goody, Schlee, Zitelmann).
31The switching process uses a characteristic of human society, which animal social structures, so popular in theories of in‑group behavior, do not share : the capacity to preserve, in latency, different orders. We are all multilingual, at least in respect to language registers, we master different roles which we put into practice according the situation and we make consecutive use of several affiliations (several modes of belonging, or identities in the strict systemic sense). We may call this phenomenon polytaxis or polytactic potential (from the greek polus many and taxis order).
32Thought on the concept of identity :
33At this point we need to discuss the concept of identity. Currently there is a boom in essays on this concept. And it's not an accident that it is almost never operationalized.
34Most commonly identity means, the guise of a scientific concept, an ideologized sentiment. A closer look, however, reveals three different meanings which are sometimes brought together :
Belongingness (membership. The binding together of different people, this stems from the old philosophical meaning ; I refer to identity in this sense, if not mentioned otherwise.). There is logically no problem to be part of several categories.
Self information (what do people think that characterizes themselves).
Self‑valuation / attribution of prestige (what are parameters of esteem and where does one place one's own group in relation to others ; a central field of research ‑ as the foregoing one ‑ which should not be confused with identity in the strict sense which should however be seen in correlation to it).
35There is, however, a difference between the function of polytactic identities within closed and within open structures (in most cases identical with the opposition of ethnic groups and nations). For African ethnic structures it is the commonly case that individuals can claim several affiliations. There we find situational switching (or : ethnic conversion) as an individual process. It is quite rare that there is only one "ethnic identity". (In a famine a Fulani herder in the West African Borgu might opt for the sedentary way of life of a Gando, if he were poor, or might become a Dendi trader, if he had some wealth).
36This remarkable freedom of option, we can observe should however not induce us to anthropological romantism. The multiple opportunities of polytaxis alson create involuntary situational switching. A person may be excluded. e. g., if it is "discovered" by oracle that he or she belongs "in reality" not to our we‑group, but to the secret order of witches.
37There are some configurations where individual switching is a common answer to conjunctural ups and downs (Horn of Africa, the interior delta of the Niger) there are many others where this is uncommon and may be ultimately achieved only at burial after living as the "foreign" married wife or husband for a long period in the new community. The difference between these configurations seems to me to be created by the different (especially more or less elaborated) structure of the boundary zone and its "impact procedures", the liminality of social systems (see below).
38Once open (ethnic) groups are transforri1ed into "para‑national" bodies with established rights with respect to the state, there starts a painful process of opting between these ‑ quite convenient ‑ multipul options and to negate ail but one of them by pseudo‑historical text production. With nationalities it is rather the exception (and tolerated only for a transition) that there are multiple affiliations. Within the nation‑state ethnic options are rigidly welded on to the state structure and its formal procedures. Nationality and subnational entities with only singular options of belonging are such a case. Some countries even insist in a symbolically very prominent way on this singular subnationality by printing it on identity cards (in Rwanda these were instrumental for the organized genocide of Tutsis).
39This rigid coupling constitutes a major difference between identities in an ethnic or an national field of reference. Individual switching in the national field tends to become a form of "cheating". Potential multiple belongings (identities) become something dangerous, if agitators label them as potentially disloyal. Ethnic cleansing is then some nation states' final answer to polytaxis.
40I said above that switching is a rather common, although rarely noticed process (leaders prefer continuity and "historical roots" and anthropologists are the natural prey of ideological leaders). This does not imply that there is any theory for collective switching. Individual situational switching (opportunistic switching in the descriptive sense) is one thing, but collective ; coordinated : and synchronized switching is an other. Motives have to be redefined, new labels become the markers for connection, old labels have to be declared invalid, boundaries have to be redefined. Therefore a minimal consensus has to be sought.
41However the social structure provides in general no routines or institutions for such negotiation processes. This requires a very specific type of social actor : political entrepreneurs or central persons.
42My hypothesis is that switching processes are correlated with a process whereby the number of subjectively active actors in a given field (politics) is increased, whereas paradoxically in the same process the number of actors providing interpretation ‑ especially of new phenomena ‑ and implied in structurally relevant decisions declines. The subjective impression is that "we are all getting politicized" ‑ more people are entering the political arena.
43But after an initial ("chaotic") phase, the number of those who indicate (virtual) directions, leaders in the strict sense, decreases. If, subjectively, the chaos of future becomes a problem, if the "creation of a future", the selection of paths for future events became the desideratum, then the moment for a specific type of actor has come : These are people who can produce a restriction of future options under the appearance of opening ways in a thicket. In other words, they can create power "from nothing" ; it is a power which emerges from beyond the established decision‑making procedures. The ones who can produce this "miracle" we will refer to as central persons. They are big men, warlords or adopted leaders (from the outside).
44Big men and warlords both can convert prestige into power (= big men) or wealth into power (= warlords). They act in the resource triangle of power, wealth and prestige.
45Warlords turn the wheel clockwise. They create power with money for weapons and mercenaries, win prestige from power, which gives them credit for the acquisition of new wealth. Big men turn it the other way round. They transform prestige into power. Power may create wealth, but wealth and labour power has to be "devoted" to the people in order to create prestige.
46Power can also have its origin outside of the we‑group. We shall call these central persons "adopted leaders". Kemal Atatürk of the Alevi‑community seems, to me, to be an example of this case. Another more recent example is the role chancellor Kohl played in the East‑German transformation process (after the Protestant big men of the civil rights movement reached the limits of their mobilization capacity).
47But the formal mechanism of importing power from the outside or of creating power from prestige or wealth is not sufficient. Central persons need, for their success, a very specific personal quality : an interpretation capacity (Deutungsmacht) which links them in a feed‑back loop to the polytactic reference schemes of those they want to lead. Interpretation and orientation relieve one of the strain of complexity. The capacity to produce interpretation with seeming clarity, to reduce the complexity of the future with intellectual tools is the personal attribute which is required.
48Political entrepreneurs may also switch individually from separatism to nationalism or vice verso. (Thus the former separatist Konrad Adenauer became later a symbol of moderate national feelings in Germany. Silvio Magnano left the Italian fascist student movement to become a successful leader of the (separatist) South‑Tyrolians). One is tempted to say that in the environment of nation states one has to be a political entrepreneur in order to switch, since individual switching there is usually prohibited.
49Switching is not a random process. It is not even "normal". For social scientists the phenomenon of switching should of course be normal, but for the individual in most situations routine and continuity is the easier option. Switching has to refer, as formulated above, to established symbolic inventories and is limited by these ; if for example religious we‑groups have no historic precedence and if they do not constitute networks of communication or exchange nor provide institutions for conflict arbitrations, then there is no switching possible in the direction of this religious reference.
50Central persons may have to add to the inventory and they must alter the structure of relevances 7 (new goals have to become more important than those stabilizing the routine) before they can activate switching. To alter the structure of relevance new fears and new hopes have to be instilled.
51Changes in the .field of symbols which concretise and order perception have a peculiar algorithm. They should be close to the accustomed ways of thinking and yet at the same time appear to be innovations. 99% accustomed concepts + 99% innovations would be an appropriate formula. The best solution is travesty ! The innovations are presented as traditions (see Papstein and Hobsbawm & Ranger on the invention of tradition) and/or conceptual limits are retained and only their appearance is altered 8.
52If deeper changes in the social structure are sought or if marginal persons want to occupy center stage, then another algorithm might be more appropriate or be used concurrently : the creation of a co‑ordinate system. Something is defined as a core norm and every behaviour is judged in terms of distance this zero point. Belonging to a specific "race", belief in some value or practice of some rituals are defined as "the fundamentals" from which any other condition of well‑being can be deduced. It seems not convincing to me to accept the self made definition of the protagonists that the "fundamentals" were the doctrinal core of their religion. Whether the 'shador' or trinitarianism, these beliefs are seldom those found in the holy books. Symbols are chosen which react to other symbols in the time of present action.
53Even a non‑doctrinal religion such as Hinduism is attributed some "fundamentals" (see Randeria). The leaders define their own place as "zero". That is the irony of fundamentalism.
54It should not be forgotten that besides ideologies used for active switching, people may also be driven into an other self definition. Most of the Jews expatriated by Nazi‑Germany gave up the self definition "German" and kept the new American, Israeli etc. identity' even after the war, when they could opt back.
55A paradigmatic case is that of the Alsatians (and Lothringians 9), the inhabitants of the former German province ElsaB‑Lothringen, 1872 ‑ 1918, on the left bank of the river Rhine (following Jahr 1995). In 1914, when the First World War started, they were as engaged for the nationalist case as any other German "tribe" ("Stamm"). But since their "Germanness" dated only from 1872 (when Germany annexed this part of France without asking its inhabitants, based only upon the linguistic fact that they ‑spoke German) the military hierarchy was mistrustful. Thus they treated this territory, once the war started, as occupied land. When German troops accidentally fired on each other, it was suspected, that these were shots from a guerilla, though there were never any. Censorship was imposed upon newspapers and letters. If an Alsatian soldier did not return from the battlefield, it was immediately suspected that he was a deserter, though with others it was assumed that they were killed or taken prisoner. In spite of no statistical support for this prejudice, the Alsatian recrutes got a special treatment. Civilians who called each other names by making reference to the others' home region faced imprisonment, if the verbal aggressor was Alsatian and the other not – but not vice verso 10. People felt deserted by Germany. Without any underground pro‑French movement the opinion changed completely till 1918. When the French entered the territory after the peace agreement, they were hailed and everyone declared her‑ or himself to be French, although the majority of the population had given up bilingualism during the German period and for most people "French" was only their parents or grandparents identity. The same procedure seems to be used by the Turkish army "in order to" drive any Turkish identity out of the Kurdish‑speaking populations of its East. The Russian government probably "succeeded" this year driving out any remnant of Russian identity among the population of its province Chechnya.
56Populations deprived of equal access to the social and material goods relevant to them tend to opt for an other identity. The German proletariat of the 19th century who opted for the internationalist we‑group of "The working class" is, to my mind, such a case. This phenomenon is strengthened, if these people get ascribed a different identity by the majority -under this condition labelling and insult can become a material force. Some movements of the American blacks seem to me to be such a case (cf. Irek). A higher prestige status and the hope ; for a better material status associated with it make switching the best and easiest option.
57Switching is more commonly linked to daily behaviour than our anthropological or political research mentions. The definition of belonging to some group is directly linked to the field of kinship and alliance structures. It depends upon the kinship structure taken into account whether a person belongs to her/his father's, mother's, mother's brother's or spouse's ethnic group. Once ethnic groups or nation get established, the production of kinship laws is one of their first occupations.
58This brings us back to the classic definition of an ethnic group. Frederik Barth's definition, refering to ascription and selfascription of boundaries, is too broad. This remained unnoticed for some time, since every reader had some preconception of what was meant. The definition implies also, if we read it strictly, political milieus and other subcultures. What was implicit as a defining characteristic was that these groups are a sub‑set of those groups which transcend families and which integrate families as the central reproductive structures.
59Ethnic groups are family‑inclusive social categories, which ascribe themselves (or get ascribed by others) an (exclusive or multiple) identity.
60To be born into a family identified with a given group produces the right to belong to this group according to the membership definition attached to this family. Switching implies the negotiation of identities, legal conflicts over the question to whom someone belongs and the "underlife" ("cheating") present in all human behaviour towards norms. All this is exposed in individual strategies dealing with ethnicity via kinship and alliance. People may be included by an initiation ritual which is required for marriage. They may opt for matrilineality instead of patrilineality. They may, in a virilocal environment, use uxorilocality as a means of integration. Endogamy as a group norm is no accident. In fact, it may be a strategically necessary element.
61States are conscious of this. It is remarkable that even countries which claim ‑ in simplification ‑ that their citizenship is defined according locality ("ius soli") and not according to descendance ("ius sanguinis") like France and the US, devote much attention to the difference between marriage and "sham marriage" or to the marital status of their citizens' children born abroad.
62This link between the dynamics of we‑groups and reproduction has impacts upon demography, if we break down statistics to the level of groups differentiated by social structure or culture. The mode of inclusion affects demographic growth or shrinking on different levels. That migration is linked to individual switching is obvious. Whether migrants who marry in are counted as locals or not depends upon the specific mode of inclusion. Neither the modes of inclusion nor the groups' limits are stable. The limits may change within a generation.
63In some groups (e.g. the Berba of northern Bénin) women have no ethnic status at all. Their social integration is reached through sons, who undergo the initiation ritual. To bear children is a primary social necessity for women in these or similar social conditions.
64Once we see switching as a normal process which reveals to us the universal polytactic potential, then we can have a fresh look on those groups which do not switch or at least keep their boundaries.
65What stabilizes ethnic groups and other we‑groups ? Current theory is mostly concerned with processes of splitting, of re‑drawing boundaries and in the creation (invention) of legitimating ideology. Some hints may be found in the long listings of how nations are built in modernization theory. (There however every conceivable factor is touched in such a way that the theories can never be wrong.)
66It seems to me that two processes are central :
moral economy/practical solidarity : a distribution process of goods and services which, by man y actors, is not seen as (market‑ )economy. The principles of distribution are rather those which economic anthropology calls generalized reciprocity or redistribution. This extends from help for anonymous persons belonging to the same we‑group to redistribution through welfare and pension schemes (see Elwert 1987, Kohli 1987).
institutions of conflict resolution : the increase of the conflict processing capacities (Konfliktfähigkeit). This may be obtained by the creation of institutions for arbitration or by the implementation of a normative system equally relevant for all the members of the we‑group (see Albert Hirschman 1994). It was not "common values" but the new "B.G.B." (common law) which made, at the end of last century, national Germans out of the citizens of the German States. It is often stated that common values ‑ at least a minimal core ‑are needed to create a we‑group. Empirically we have difficulty finding cases to illustrate this point ; what we find are institutions and mutually accepted procedures which forego the creation of common values.
67As a German I am supposed to bring language into the discussion (following Herder, who "naturalized" nations by reference to language 11). There are, however, sufficient we‑groups of ethnic form without a common language, to contradict this idea. We should rather draw attention to the fact that communities of discourse can also be interpreted or seen in relation to arbitration of conflict and economics of reciprocity. Arbitration requires a code and this code has to have a linguistic form (a given language or two corresponding isomorphic registers of two languages). Norms about sharing knowledge or keeping it in secrecy ‑ and so making it an economic good ‑ constitute codes of communication, which make a language a relevant or an irrelevant tool for social action.
68Not one language directly, but one network of communication combined with shared social norms seem to be at the basis of all processes of we‑group formation. Even in situation of seeming powerlessness this combination allows for the attribution of honour and shame (the "reputation sanction") as means of social control.
69Those features of community which strike us most ‑ a common habitus (similarities in style and routines) ‑ have been studied by Ralph Bohnsack and his group in their analysis of group formation among German youth. It turned out that habitual community is a later development. It starts with a mutual articulation of individual plans for activities, creating an interdependence, which one may call community of fate (Schicksalsgemeinschaft). Norm‑breaking or violence which provokes revenge are powerful but risky means to create a community of fate in the emphatic sense ; the weaker the groups' capacity to create a common code or habitus, the stronger is the inclination to produce collective risk or foes by means of violence if necessary.
70Social systems do not only exist in the self‑limiting status of ethnic groups, although the mode of information storage in anthropology (arranged according to ethnic groups) may suggest this. There are two modes : one is characterized by opening up and networking with the imperative to inc1ude an optimal number of persons, who beforehand were strangers to me, into my network of reciprocal relations. The other is the restrictive closing up of we‑group formation, which this paper dealt with. But it is both forms, the bimodality of networking and group closure, which make for the dynamism of mankind. Though we may conceive some historical processes as an alternation of both modes, I want, rather, to draw attention to social arrangements which combine both ‑ the definition of people as strangers and their inclusion.
71The self description of those features of we‑groups which make for their stability rarely mentions the fact that all these arrangements seldom produce unambiguous borderlines. Yet the multiplicity of social subsystems produces rather complicated legal arrangements : differentiation between political and economic citizenship, civic rights for strangers as social or economic actors e.g. Each of these boundaries can be subject to a tidal movement whereby some actors are for some time in or out. This causes the social boundary ‑mostly unnoticed by the majority of population ‑ to become a broad "liminal zone" 12. This liminal zone fulfills important social functions. Here is a field where new ideas and arrangements are tested out before they become adopted as "ours". This very adoption as "ours", the "nostrification", requires for some items social arrangements, institutions.
72In order to find out what can pass and what needs to be scrutinized, this liminal zone acts as ship locks. Because of their precarious positions persons in this field are highly sensitized. They may perceive earlier the usefulness of a new arrangement (because they saw earlier its functioning outside the system or in the liminal zone) and they may react more allergically to something as "strange" (because they want to be "in").
73The strength of the human race in respect to other animals is our great flexibility. Multiple identities and switching processes contribute considerably to this. The creation of nation states (which was an "evolutionary success”, or, to be more precise, which proved to be instrumental for pattern expansion in other fields) produced, however, a set‑back with respect to flexibility for individuals' polytactic social behaviour.
74Rather complicated legal arrangements try to make up for with this draw‑back (differentiation between political and economic citizenship, civic rights for strangers as social or economic actors e. g.). They create social boundaries that become broadened liminal zones serving as surrogates for polytactic ethnic flexibility.
75From the perspective of an individual actor who makes opportunistic use of his multiple affiliations in an ethnically defined environment, a national structure still looks "underdeveloped". It is still a long way till we will reach the flexibility of we‑group identity, which is the strength of ethnic systems.
76Nationalism and other we‑group mobilizations are, however, a product of "modernization" in a restricted sense if we conceive as "formal modernization" the concomitance of expanding market economies with expanding communications systems linked by a positive feed‑back, then nationalism and fundamentalism are clearly aspects of this. We should, however, keep in mind that this definition of modernization does not imply any reference to (civic) values or to social differentiation through institution‑building.
77Warlord systems and genocides are part of modernization in this perspective. Insofar as modernization increasingly creates links to anomymous people and, as these links increasingly dominate the life world as "societal" (gesellschaftliche) relations, there emerge undercurrents which reinterprete and reconstruct social links as something built upon emotions or face to face relations (c.f. Frühwald). If, for example, marriage is less an affair to be regulated by lineages, and if selection of mating partners is opened to persons unknown to the heads of families, while the formal guarantees and rituals are handled more and more by state authorities, then "love" becomes an attractive code for the interpretation of decisions leading to marriage (cf. Luhmann). I put this ex ample into the foreground in order to underline the fact that nationalism is not the only "undercurrent". But it is a rather obvious case. Societal links are reinterpreted as community (Max Weber). The real process may be one of building institutions which link persons, who are, in most cases, anonymous to each other. The facade of self descriptions dwells upon images of community, using metaphors of face‑to‑face groups and invoking emotions.
78Nationalism and its concept of "Nation" contributed also of course to our image of a modem world. It contributed many ideological elements which today go without saying. It may be worth to discuss them.
79Modern politics is in one important aspect a symbolic process, motivated and limited through sociohistorical dynamics. That modem nationalism is so intimately linked with war is not an obvious or even natural combination. For the masses, who had to be integrated as soldiers, war was a costly and risky enterprise. A reconstruction of the prestige system was necessary in order to alter this attitude. The "Hero", and especially the dead one, had to be constructed as peak element of the prestige pyramid (taking off from the soldiers the ridiculous image of Plautus' "miles gloriosus") and even as an object of secular quasi‑religious devotion. This ideologem was then artfully intertwined with oscillating metaphors of "Freedom". Freedom was seen on the one hand as a controlled realm conceived as "own" ‑ one's own farm or workshop or family ‑ and on the other hand as accessible realm ‑ freedom of movement for migrants, traders and access to commodities. The oscillation between the two aspects comprised the strength of the concept.
80The more politics were associated with communication, the more it was plausible to link we‑groups to languages. Thus the idea that one nation should have one language and, later, that one language makes a nation, is incrementally gaining acceptance. Though at the period when nationalism was formulated, the social reality was still that of multilingualism. .
81In the beginning nationalism was far from being a plausible concept. Now the combination of an "obvious reality" ‑ language ‑ with strong sentiments of prestige ‑ heroes interwoven with equally strong sentiments of fear and hope ‑ freedom ‑ , seems almost unchallenged and contributes to our image of the modem world. This combination, however, was arrived at a two‑century long process of variation and selection. But this is not the end of we‑group ‑history ; competing concepts may yet be in their evolutive process...
82I would like to thank Artur Bogner, Ayse Çaglar, Chris Gregory, Krisztina Kehl, Ute Luig, Rainer Münz, Günter Schlee, Andreas Wimmer and Thomas Zitelmann for their useful hunts. And I would also like to thank, in advance, all those kind readers of this paper who would be willing to communicate to me their criticisms and suggestions.