Gellner and Geertz in Morocco: a Segmentary Debate

ABSTRACT

Segmentary lineage theory dominated (especially British) anthropology for about 25 years after the Second World War. In its heyday it was hailed as one of the greatest theoretical achievements of the discipline (Fortes 1953). At that time the theory was applied to describe social structure of dozens of ‘primitive’ societies around the world. Since the beginning of the 1960s the model has repeatedly came under strong criticism and nowadays, in 2003, it commands little attention.

In the meantime, starting in the late 1960s several scholars who had conducted their fieldwork in Morocco published monographs that put the country in the center of anthropological debates on the nature of fieldwork (Dwyer 1982; Rabinow 1977), ethnographic writing (Crapanzano 1980; Munson 1984), and Islam (Eickelman 1976; Geertz 1968; Gellner 1981a). One of the monographs, Saints of the Atlas (1969), made its author Ernest Gellner the main advocate of the segmentary model for the years to come. His compelling argument re-invigorated the segmentary debate and re-focused it on political life of pre-modern Arab and Berber tribes of the Middle East and North Africa. Since then several leading anthropologists have presented widely differing views on the issue and their contributions touched upon important theoretical and epistemological issues of the discipline.

The purpose of this paper is to analyze and assess that part of the debate whose participants conducted anthropological fieldwork in Morocco.

ORIGINS OF SEGMENTARY LINEAGE THEORY

Building concepts of the segmentary lineage theory come from the work of social thinkers of the 19^th century. The idea of society composed of mutually resembling and economically independent segments appears in Durkheim's De la division du travail social (1893). It is a basis of his concept of ‘mechanical solidarity'. Another influence was an evolutionist theory of ‘primitive society’ elaborated, among others, by Morgan (1877) and Maine (1861). Attention of these scholars concentrated on kinship seen as the main force integrating society. The underlying issue was the constitution of primitive polity and the implications for a civilised political order (Holy 1996: 72; Kuper 1982: 73). Two sets of principles were considered important. First, an interplay between ‘blood’ (kinship) and ‘soil’ (territory). Second, an interplay between family (conceived as a bilaterally traced web of kin members) and clan (conceived as a group of unilaterally traced kin members, later renamed to lineage by Gifford [Kuper 1982: 79]).

These and other influences intersected in the 1940s in several classical ethnographies, namely African Political Systems (Fortes and Evans-Pritchard 1940a), The Nuer (Evans-Pritchard 1940b), and The Dynamics of Kinship among the Tallensi (Fortes 1945). In them a distinction was made between domestic and public domain. In the public domain emphasis was put on jural aspects of kinship as opposed to interpersonal relationships that were considered important in the domestic domain.

The main issue raised by Fortes and Evans-Pritchard was the regulation of social order and political life in so called ‘stateless societies’ in colonial Africa. Their analysis departed from the then dominant topical (religion, magic) and geographical (Pacific islands) preoccupations and combined functionalism of Malinowski with a novel concept of social structure. Radcliffe-Brown understood social structure as relationships between persons while his students Fortes and Evans-Pritchard defined it as relationships between groups (Evans-Pritchard 1940b: 262; Kuper 1996: 82; Radcliffe-Brown 1965: 9).

The central innovation of Fortes' and Evans-Pritchard's theory was the idea that in the ‘stateless societies’ public sphere is regulated through complementary opposition between fusing and splitting segments whose memberships was defined by descent. It is now accepted that the society that served as a blueprint for the segmentary theory were not so much the Nuer but Arab tribes of the Middle East (Eickelman 1981: 100; Kraus 1998: 19). They were described by Robertson-Smith in his Kinship and Marriage in Early Arabia (1885), a study that drew on the 19_th century speculations about family, kinship, territory, and the nature of society mentioned above. The book influenced Evans-Pritchard who, in the 1930s, conducted fieldwork not only in British-administered Sudan and Kenya but also in what is today Libya. A resulting monograph, The Sanusi of Cyrenaica appeared in 1949 but the fieldwork among the pastoral Bedouin was conducted concurrently with the Nuer.

Segmentary theorizing had a large impact on the discipline. It stimulated a number of detailed ethnographic studies, brought conceptual innovations to the then dominant anthropology of Malinowski and Radcliffe-Brown, and provoked a long-running anthropological debate. It also offered a model of the maintenance of social order in egalitarian societies without centralized government. Nevertheless, the weaknesses of the theory were substantial and the criticism of it from the 1960s onwards devastating (for overview see Holy 1979a; Kuper 1982). Yet there is one field where segmentary lineage theory keeps provoking passion and curiosity: the analysis of tribal societies of the arid zone of Islamic world stretching from Morocco to Pakistan. The story begins and ends ‘where brothers unite with brothers against their cousins and then unite with the cousins to fight a common enemy ...’

In retrospect, the way the segmentary lineage theory achieved preeminence seems to be a coincidence rather than a well-thought theoretical shift. The cultural and social importance of patrilineality among the Arabs elaborated by Robertson-Smith and a common interest of the two editors of African Political Systems in the ‘stateless’ societies gave inflated importance to theorizing about the role of unilateral kin groups in political life.

INITIAL FORMULATIONS

The earliest and clearest formulation of the segmentary lineage theory can be found in Evans-Pritchard's writings on the Nuer of southern Sudan (Evans-Pritchard 1940a). Using his account as a guide one can characterize a segmentary lineage society in the following way:

(1) It has no centralized government and is largely egalitarian. Appointed chiefs have limited powers (if any) to enforce decisions (Evans-Pritchard called the Nuer society an ‘anarchic state’ [1940a: 272] or ‘ordered anarchy’ [idem, p. 296]).

(2) It consists of political units that exist on several levels. In the Nuer the smallest political unit was the village, the middle range unit was the district, and the largest unit was the tribe. The larger political units are always composed of (and subsume) the smaller ones. These political units form, to a larger or lesser extent, territorial entities as well.

(3) It divides into lineages (descent groups) with exclusive membership traced unilineally. A tribe may comprise a few major lineages. Each lineage divides into lineages of smaller size which again divide into smaller ones etc. Thus a genealogical tree of related lineages can be drawn with important ancestors at cleavage points all ultimately descending from the founder of the tribe. The principle of lineage organization is the same as the principle of political organization. In the Nuer the lineages could be differentiated ‘only in reference to rules of exogamy and certain ritual activities’ (idem, p. 286).

(4) The relationship of political (residential, territorial) segments and lineage affiliation is one of the main sources of confusion in the theory since they do not necessarily coincide: ‘Every Nuer village is associated with a lineage, and though the members of it often comprise a small proportion of the community, it is identified with them ...’ (ibid.). At the same, the lineage idiom is the unifying force of the overall political organization that determines the ordering (so called fission and fusion) of the political units.

(5) In case of a conflict between two or more parties the situation is defined in terms of lineage (or residential) affiliation of the people involved. The lower level political units then may unite along their ‘genealogical’ proximity to form two opposing groups each backing one of the disputing parties. According to the theory, a balanced opposition emerges since the political units combine into groupings of the same order that tend to have similar size and strength. The conflict is then resolved without recourse to open large-scale violence and social order is maintained.

(6) Disputes within the smallest political units are usually quickly resolved to avoid disintegration of the group.

ANTHROPOLOGY OF MOROCCO

Morocco comprises three main environmental zones that correspond roughly to three different settlement patterns and life-styles. About three-fourths of the population live in coastal plains and plateaus with Mediterranean climate. They are urban dwellers or farmers. Most are Arabic speakers. One fifth of Moroccans live in the highland areas of the Rif and Atlas mountains. They are mostly Berber speakers who abandoned nomadism under pressure from French colonial administration. About 5 percent of the population lives in oases in dry pre-Saharan and Saharan areas to the south of the Atlas. They are mostly of mixed African descent, a majority of them speaks Berber.

In pre-modern period, the Berber tribes of the High Atlas were not ruled by the Sultan. They lived in what Gellner calls ‘institutionalized dissidence’ (1969: 1). They remained untouched by modernization and inaccessible to foreigners well into the 20_th century. The French proclaimed Morocco their protectorate in 1912 but the last tribes succumbed to them only in 1933. Until 1956 when Morocco gained independence the social change in the High Atlas was relatively slow thanks to the French policy of upholding Berber traditional customs and institutions.

The first detailed ethnographic information about the tribes comes from French colonial officers who administered the region from 1933 till 1956. However, their reports did not lead to systematic research and remained largely unpublished (Kraus 1998: 1). It was only in 1950s and 1960s when anthropological research of the tribes began in earnest.

An American David Hart who spent most of his life in Spain and Morocco started his research of Berber tribes of the Rif in 1952. His detailed ethnographies are good examples of the more traditional use of segmentary lineage theory. He was followed in 1954 by a British anthropologist Ernest Gellner who conducted his fieldwork among the Berbers of the High Atlas (Davis 1991; Gellner 1969: 303). He subsequently formulated the most sophisticated version of the segmentary theory to date. In his Muslim Society (1981) he integrated the segmentary theory into a general model of the classical core of the Islamic civilization.

A group of anthropologists of completely different intellectual background descended upon Morocco between 1965 and the early 1970s. They were all Americans. Those who later contributed to the segmentary debate were mainly from the team led by Clifford Geertz: Lawrence Rosen, Paul Rabinow, and Hildred Geertz.

The research of Hart and Gellner on the one hand, and of the group around Geertz on the other, can be contrasted in two important aspects:

(1) Gellner and Hart were trying to formulate structural-functionalist sociological explanations while the Americans studied culture conceived as a system of meanings.

(2) The Americans stayed in and around Arabic speaking towns in the lowlands and were interested in the present. Hart and Gellner worked mainly the Berbers in the mountains and tried to reconstruct the tribes in their pre-modern condition.

During the 1970s and 1980s the spectrum of anthropologists conducting fieldwork in Morocco further diversified bringing in other epistemological, theoretical, and topical interests and approaches. However, with regard to the segmentary lineage debate the tone of the discussion had already been set. More recent contributions can easily be understood as elaborations, commentaries, criticisms, or refutations of earlier arguments (for detailed intellectual history of segmentary theory in the context of anthropology of Morocco see Gellner 1985; Hall f. c.).

SEGMENTARY THEORY IN ANTHROPOLOGY OF MOROCCO

Four anthropologists found segmentary theory indispensable to explain their Moroccan material: Hart, Gellner, Combs-Schilling, and Kraus. Ernest Gellner and David Hart started their research without realizing the relevance of the theory for their findings. Initially, they did not know each other and employed the segmentary theory independently. Gellner came to Morocco influenced by Robert Montagne and only later turned his attention to Evans-Pritchard (Kraus 1998: 19). Two younger scholars, Elaine Combs-Schilling and Wolfgang Kraus, already knew Gellner's and Hart's work when they began their fieldwork.

Between 1954 and 1968 Gellner spent more than one year in total among the Berber tribes conducting research for his PhD dissertation published in book form as Saints of the Atlas (1969). His approach to ethnographic writing was unorthodox and it testifies to Gellner's deep interest in philosophy and social sciences in general rather than simply anthropology.

The first part of the book, the one that later provoked so much controversy, is an attempt to make a fundamental contribution to Western political philosophy. It tackles a long-standing question of political life in a society without centralized government (Davis 1991: 71; see also Eickelman 1981: 99). However, this elaboration of segmentary theory drawing upon the ideas of Durkheim and Evans-Pritchard concerns the lay Berber tribes rather than the saints that gave name to the book. The saints are the subject matter of the second, ethnographic part of the monograph. Here, Gellner examines political life of a cluster of saintly villages interspersed among transhumant and sedentary lay tribes. Saintly communities of the High Atlas were in a minority compared to the lay population. Among the total population of the High Atlas of a few hundred thousand people they numbered less than 5 % (Hart 1981: 1, 62; Gellner 1972b: 61). This plan makes the book quite bizarre. The highly abstract, logically ‘pure’ functionalist argument about the majority population is followed by an ethnographic analysis of an anomalous minority. Gellner himself later admitted ‘[t]hey [the lay tribes] were at the edge of my field of vision or attention, whilst I was actually focusing on the saints’ (Gellner 1995: 821).

There are four aspects of Gellner's version of the segmentary model the Berber tribes of the High Atlas that are worth mentioning here:

(1) Gellner tied the theory to ecological conditions in the tribal territory. Resources were limited and chances of violent confrontations high. In the absence of government or centralized authority some other mechanism of regulating access to the resources had to be in place.

(2) It was beneficial for every segment to have clan members planted in all ecological zones to be able to balance climatic variation by migration to one's kin in more favourable place in case of necessity. This ensured loyalty of the segments (Gellner 1995: 822).

(3) The fact that each key sub-segment is present in all ecological zones and thus along different sections of tribal frontier ensures cohesion of the tribe in case of an attack from outside. In other words, this distribution explains the loyalty of the segments to the total confederation (ibid.).

(4) For segmentary regulation to work among the tribes it had to complemented by permanent saintly settlements located in the interstices in between tribal segments. The most powerful and charismatic saints (igurramen) served as arbitrators in tribal disputes. Their saintly, non-tribal descent, moral authority and formal pacifism put them above the anarchic, egalitarian, and feud-addicted tribes.

Hart's copious writings on Morocco are, in many regards, the opposite of Gellner's. His are voluminous descriptive ethnographies full of details with little emphasis on theory. Hart and Gellner came to know each well, shared a common interest in the question of social order (Hart 1994: 234) and were exchanging both ideas and data. Hart employed segmentary lineage terminology and ascribed segmentary lineage organization to the tribes, though, compared to Gellner, he did so in a very conventional way. By the late 1980s his work became an easy target. In 1989 Munson turned Hart's own data against him so convincingly that Hart subsequently abandoned the theory. Towards the end of his career Hart expressed an ‘extreme dissatisfaction’ with anthropology as a discipline (ibid.) and turned subscribed ‘more frequently to social history than to social anthropology’ (Hart 2000: 1). He also endorsed a sort of social constructionism (Hart 1994: 235) that squares uneasily with his earlier structural-functionalist aspirations.

Combs-Schilling and Kraus belong to a younger generation of ethnographers who conducted their fieldwork in 1970s and 1980s respectively. Combs-Schilling, an American, researched commercial activities in a Moroccan boom town of about 8,000 people located on a major highway in the foothills of the High Atlas. She turned her attention to segmentary theory after incidentally discovering that 69 % of the town's merchants ‘made their initial transition to town and to commerce by means of patrilineal ties and resources’ (1985: 666). This led her to argue that segmentary social structures may well co-exist with alternative behavioral strategies such as pragmatic make-up-as-you-go alliances observed during periods of rapid modernization.

An Austrian anthropologist Wolfgang Kraus has raised the segmentary debate to a new level of sophistication. His research focuses on the Ayt Hdiddu of the High Atlas who live to the east of the Ait Sokhman and Ait Atta observed by Gellner. As Gellner and Hart, Kraus is interested in pre-modern tribal social organization. In his most important article to date he inquires into collective identities of the Ayt Hdiddu (Kraus 1998). The identities show a segmentary pattern but Kraus analyses them in neither ideal (as Gellner) or superficial terms (as Hart). He distinguishes two interconnected levels of individual affiliation: a genealogical one (‘original’) and a political one (‘actual’). He also pays close attention to important factors that spoil any simple picture of egalitarianism and exclusive membership ‑ inequality and doubt. In doing so Kraus provides an analysis of concrete processes of how collective identities form. He is less specific about the ways these identities put constraint on collective action. In contrast to Gellner and Hart, his account is based on logical separation of segmentary mechanism from agnatic descent. Kraus contends that his research lends support to two key features of Gellner's ideal model: There is a tree-like structure of nested segments which balance each other in terms of their political force (1998: 16).

CULTURALIST CRITICISM

The segmentary debate in anthropology of Morocco began as a series of high-profile parallel monologues that only gradually, over many years, changed into a real debate. As one would expect the main target of most criticism was Gellner's logically ‘pure’ functionalist model of the tribes of the High Atlas. Hart's work came under scrutiny much later. The whole polemic began by a dismissal of Gellner's model by Clifford Geertz and his colleagues.

Gellner defended his PhD thesis in 1961, eight years before Saints of the Atlas were published. Hall notices (2000: 1) that the text was known to Clifford Geertz who mentions it in Islam Observed as ‘unfortunately still unpublished Ph.D. dissertation’, ‘a fine study of a saint cult’ (Geertz 1968: 124), and employs its ideas in chapter two of his book. Three year later in his brief review of Saints of the Atlas Geertz does not engage Gellner in any serious argument and dismisses him as an ‘old believer’ who thinks ‘there is still an object ‘out there,’ like Everest ...’ and a person who ‘descended upon it [Morocco] with a finished theory looking for an instance’ (1971: 20, 21). This a priori rejection is complemented by a more empirically sounding disagreement about the importance of social structure in what Geertz calls ‘Maghrebi’ or ‘North African’ society (1971: 20, 21). His position is unequivocal: There is not such thing in this society. There are only ‘ad hoc constellations of miniature systems of power which compete, ally, gather strength, and, very soon overextended, fragment again’ (1971: 20).

According to Geertz, Moroccan and North African society is best viewed as consisting of unstable and temporary small social networks that are constantly created and re-created by pragmatic individuals (Geertz, C. 1971, 1979: 235, 264). Similar ideas were expressed by other American anthropologists (Crapanzano 1980; Eickelman 1976, 1981; Geertz, H. 1979: 315-317, 351–356, 377; Rabinow 1975; Rosen 1984: 73–76; for a critique of this model see Abu-Lughod 1985). In contrast to Geertz these scholars express their opposition to Gellner in empirical terms, as if it was a matter of different interpretation of the same data, or perhaps an interpretation based on different sets of data.

CRITICISM BY HAMMOUDI AND MUNSON

Two other criticisms of the segmentary theory in Morocco deserve attention. They were made by Abdellah Hammoudi, a Moroccan anthropologist living in the United States and Henry Munson jr., an American. Unlike the culturalist dismissal they provoked a series of exchanges with Hart and Gellner.

In 1974 Hammoudi published an important article where he reflects on Gellner's segmentary model. His methodology is mostly ethnohistorical combined with his own fieldwork in the Ait Atta territory and secondary sources. He comments critically on several cornerstones of Gellner's model. First, he questions egalitarianism of the tribes and points out that immigrant and conquered groups, though genealogically assimilated, remain in an inferior position. Relationships of patronage and elements of stratification are common. This observation resembles ‘incomplete integration’ described later by Kraus (1998: 10). Second, he claims that Gellner underestimates the importance of chiefs who always come from a limited number of powerful extended families. Third, using historical sources, he shows that, on a macro-social level, the saint, far from simply playing a role of inter-segmentary arbitrator is better seen as an ambitious tactician who ‘preaches to the crowds, raises an army and inserts himself into the dynastic process’ (Hammoudi 1996: 278). Combined with the other observations, understanding the powerful saints not as an anomaly facilitating the functioning of otherwise egalitarian segmentary structure (Gellner), but as an integral part of the structure, allows Hammoudi to redefine the Berber tribes of the High Atlas as a stratified society ruled by religious specialists. Gellner himself did not view Hammoudi's critical commentary as necessarily disproving his model though he did not answer his objections adequately. He acknowledged validity of some of Hammoudi's observations, disputed others but in general he welcomed the piece as a ‘valuable re-historicisation of the [my] ethnography’ (Gellner 1996: 656; cf. Hall 2000: 30).

Henry Munson jr. conducted fieldwork for his PhD dissertation in northwestern Morocco in 1976–1977 and made three additional summer visits in 1987, 1988, and 1990. In his articles he attempts to show irrelevance of the segmentary theory both in the Rif and the High Atlas. His argument is strictly empirical and his method straightforward. He re-read available ethnographies and tried to extract from them descriptions of real events that were presumed to follow the segmentary model. His own research in Morocco played only a marginal role in this effort. The opening sentence of his article from 1989 is characteristic: ‘It is not unusual for scholars to be so mesmerized by a model that they fail to see that their own data demonstrate its inadequacy’ (p. 386). Here, Munson's target is Hart's depiction of the Rifian tribes as a segmentary lineage society. To disprove it Munson drew mainly on Hart's opus magnum The Aith Waryaghar of the Moroccan Rif (1976). Using individual ethnographic examples given by Hart he showed that what Hart called ‘agnatic lineages’ never acted as corporate groups. To the contrary, the most important source of hostilities was disputes between agnatic relatives over individually inherited land. These conflicts typically involved brothers and sons of brothers who were backed by factions whose membership ran across genealogical ties. In Munson's view, these factions, called liffs, did not operate according to segmentary principles either. As indicated above, Munson's critique was so successful that it forced Hart, without really giving a fight, to retract segmentary theory in his Moroccan research altogether and to express doubt about Gellner's conclusions as well (Hart 1994, 1996).

Four years later and using the same method Munson attacked Gellner's work on the High Atlas (1993). Apart from re-reading available ethnographies he also interviewed a Berber man who was Gellner's and Hart's research assistant for many years. Munson's conclusions were similarly categorical: ‘...no one has ever described a single actual conflict in which this mechanism [fission and fusion of groups and their balanced opposition] was operative’ (1995: 831). ‘… Gellner's … model of a polity in which order is maintained primarily by the balanced and complementary opposition of groups generated by an unambiguous tree does not correspond to anything in the minds or the behaviour of the precolonial Ait Atta’ (1993: 278, original emphasis). In spite of the strong words, the result was less convincing than previously. It was so for several reasons. First, Gellner never published any detailed ethnographic descriptions of the lay tribes either in Saints of the Atlas or anywhere else. In addition, he accompanied his theory of the lay tribes by idealized or very general data rather than descriptions of concrete events that could be used for later re-analysis (Kraus 1998: 5). Therefore, Munson had to rely mainly on Hart's relatively brief (at least in comparison to his work on the Aith Waryaghar) work on the Ait Atta (1981), material assembled by French colonial administrators (mainly Spillmann), and statements from a several hours long interview with the former research assistant. Second, Gellner never claimed that daily events invariably correspond to his model. Instead, he focused on developing a logically ‘pure’ model and acknowledged ‘impurity’ of real events. Kraus even suggested that Saints of the Atlas could be read as an account of the main divergence from the ideal model – the saints (1998: 19).

Gellner seemed almost delighted by Munson's critique as if he was glad that someone sharing his general methodology finally challenges him. His reply to Munson was doubly paradoxical. He admitted that he never conducted any systematic research of the lay tribes (Gellner 1995: 821). They were, as remarked earlier, at the edge of his ‘field of vision’. In addition, ‘... at the edge [of ethnographer's field of vision], one must note discrepancies, but they need to be at least moderately visible. One does not dig up the ground so as to find them ...’ (Gellner 1995: 821). At the same time, he turned Munson's re-analysis against him and used it to support the segmentary theory: ‘...Munson's admirable map only reinforces my feeling that my intuition was correct’ (Gellner 1996: 644, my emphasis). As usual, his arguments were abstract as well as empirical and concentrated on the large-scale, ‘systemic’ parameters. They hinged upon Gellner's claim already mentioned above, that there was a particular (segmentary) logic to the distribution of tribal segments along the frontiers of the tribal territory and across the ecological zones. He also re-iterated the other key features of the situation: lack of resources, high chances of violent conflict, egalitarianism, no government, but no chaos and relatively stable social order. Gellner thus managed to withstand Munson's critique in general, but was unable to supply concrete ethnographic examples (see also Hall f. c.; Kraus 1998: 4–5). There is only one spot in Gellner's reply where he yields a few millimeters to Munson: ‘[T]he evidence marshalled by Munson ... may indeed show that under the pressure of circumstance, men often forge alliances contrary to the segmentary map, perhaps more often than I had supposed’ (1995: 827).

Only Hammoudi and Munson challenged the segmentary theory in Morocco on empirical grounds. Their criticism resulted in Hart changing his mind and Gellner entrenching himself. He continued to hold his views until he died in 1995.

EPISTEMOLOGICAL ISSUES

Are there any important epistemological issues at stake or are the differences between the contributors to the debate primarily ‘empirical’? This question has to be asked mainly about the culturalists around Clifford Geertz. It is clear from Clifford Geertz's review cited above that his target was not so much segmentary social structure as the fundamental principles of the kind of social science Gellner practices. Segmentary theory happened to be a mere example of it. Upon closer reading the same applies to all culturalist criticism. Besides general statements these authors published elsewhere, two kinds of evidence point in this direction: methodological statements in culturalist ethnographies on Morocco (Geertz et al., 1979) and the way the culturalists go about their criticism of Gellner.

The reasons (or rather their absence) that lead the culturalists to reject Gellner's model are unclear in the first place. Gellner describes pre-modern political organization of Berber-speaking tribes of the High Atlas. Sefrou region, the research site of Geertz's team, differs from the High Atlas not only in language (Arabic versus Berber), but also in ecology (hilly Sefrou region versus tribal territories stretching from the Sahara to snow-capped mountains), settlement pattern (town versus villages or camps), economy (complex economy of a modern town versus subsistence agriculture and transhumance), history and political life (Sefrou as part of a centralized state versus ungoverned tribes), and length and extent of exposure to European influences (remote tribes were subdued by the colonial power 21 years after the establishment of the French protectorate and modernized much more slowly).

Given the different research location and focus the culturalists could not engage Gellner in any serious argument. They could not question his data or use their own against him. Why did they pay attention to Gellner at all? There are some suggestions as to why this happened in the introduction to Meaning and Order in Moroccan Society (1979), the major publication by Clifford and Hildred Geertz and Lawrence Rosen on Morocco. They see their book as part of ‘the search for more adequate ways to render the special contribution of nook-and-cranny anthropological work to the wider, multidisciplinary effort to comprehend ... Morocco ... the Maghreb ... the Middle East ... the Third World .. the Modern World Order ...’, they present it as an ‘attempt ... to find a form in which particular facts can be made to speak to general concerns’ (idem, p. 1–2). And they fulfil their intention in a very straightforward manner: They extend their findings from in and around the town of Sefrou (30 000 inhabitants) situated ‘at the point where the foothills of the Middle Atlas Mountains meet the western plain’ (idem, p. 7) to the whole of Morocco and further. Similarly, Clifford Geertz in his review article In Search of North Africa claims that ‘[t]he social order is a field of small, pragmatical cliques … The cliques are somewhat more stable in the Moroccan High Atlas or the Tunisian steppe than they are in Algiers, but the difference is only relative’ (Geertz 1971: 20). Authority of this statement implicitly derives from Geertz's research in Sefrou. In brief, the culturalists were prone to easy generalizations.

There are other reasons to believe that the culturalist rejection of Gellner's model had to do with epistemology rather than interpretation of one common ethnographic reality. For example, why does Crapanzano bother to mention segmentary theory in his subjectivist biography of the urban outcast Tuhami? Also, as noted by Combs-Schilling (1985) the culturalist position is incoherent because it relies on authors whose rejection of the segmentary theory is strictly empirical. The most important of them, Emrys Peters claims that segmentary processes do not manifest in behavior but he admits their existence as folk ideology (1967). At the same time folk ideology, local meanings, categories, concepts, metaphors, and figures of speech are precisely the material making up culture, the research focus of Geertz and his colleagues.

Commenting on the culturalist critique Kraus writes: ‘segmentation ... was compromised by the empiricist-positivist background with which its study was associated’ (1998: 3). The culturalists a priori rejected ‘objective’ social science. In this sense the attempts at showing that the views of Gellner and the culturalists were not mutually exclusive (Combs-Schilling 1985; Hart 1994: 235) mask the essence of the problem. Gellner never denied importance of pragmatic one-to-one relationships in towns (1981: 215). The disagreement simply runs deeper. Kuper's account of the culturalist school in American anthropology (1999) supports this conclusion. The debate between Geertz's team and Gellner can be interpreted as an encounter of two irreconcilable positions: relativism and empiricism.

Gellner himself responded to American culturalist anthropology of Geertz and others in his own way. He wrote that Americans are culture-blind by virtue of being born into an exceptional civilization. When an American anthropologist specializing in the study of culture is exposed to a radically different society he or she is prone to ‘reification and fetishism of idiosyncratic and varied systems of meaning’ and ‘hermeneutic intoxication’ (1995b: 21–22). This citation comes from a short provocative piece, written in seemingly light-hearted and joking style, but undoubtedly meant seriously. Underlying Gellner's sarcasms is a radical but extremely sketchy argument about the influence of American geography, history, and social life on the practice of some cultural anthropologists doing participant observation abroad. To assess this assertion in any depth would require a separate paper.

In sum, the culturalists on one hand, and other contributors to the debate on the other hand, base their arguments on differing (one can even say opposing) epistemologies (see for example Hart 1994; Munson 1995: 831). To examine these social scientific assumptions per se is not my intention here. In the rest of the essay I will focus on the segmentary theory itself.

GELLNER'S MODEL

Gellner developed an abstract, logically coherent functionalist model of political behavior. He did not demonstrate convincingly that the model worked among the tribes of the High Atlas. Nor he has ever provided descriptions of real events that were channeled by segmentary logic or given a systematic account of the folk notions of political action (what Evans-Pritchard called ‘values’ in The Nuer). The strength of the model derives from the elegant way it ties together general observations about patrilineal kinship, conflict between individuals and groups, tribal economy and ecology of tribal territory.

Can such a model be refuted by showing, as Munson attempted to do, that social action does not neatly follow it? Kraus disagrees and calls such a view ‘naive empiricism’ (Kraus 1998: 19). Individual events can never conform to such an abstract model. Segmentary processes are always accompanied by other phenomena such as inequality, stratification, chiefs, marriage ties, treason, ambiguity of social identities, informal ties, personalities of main actors, etc. and their primary role is impossible to document. But then, how to assess validity of the model? Kraus suggests that ‘... we should concentrate on the particular quality of his [Gellner's] account ‑ the clarity of the ideas organizing his model ‑ and try to apply these ideas to our data’ (idem, p. 5). He finds Gellner's model useful in that it may ‘assist us in grasping local idioms’ (idem, p. 16) and it ‘provides us with the key to understanding the ambiguities that he [Gellner] himself prefers to ignore’ (idem, p. 18). Similarly, Hall, in his assessment of Gellner’s work concludes: ‘Gellner's model captures the spirit of the key institutions and of the self-understanding of the actors themselves’ (Hall 2000: 33). None of these citations suggests a possibility that the value of the model lies in predicting the course of individual events. Rather, the model seems to serve as a useful organizational idiom that helps the researcher to make sense of the mass of data, to grasp quickly nuances of both folk ideology and practice.

In terms of ‘hard’ sociological data and predictive value the model may do better on the macro-sociological level. It can account for limited intra-tribal violence, regulation of the use of resources, activation of men for defense against outside threat. This is an empirical question that would require further research. The argument would go along the lines of Gellner's response to Munson's critique. It would delve into distribution of clans across ecological zones and along tribal frontiers, handling of resources, comparisons of levels of violence under changing environmental conditions and historical pressures, comparisons with other tribal groups, etc. The assembled evidence would have to cover the whole tribal territory and be both qualitative and quantitative. At present there is not enough data, perhaps there will never be. So far, only Gellner, Hart, and Kraus conducted fieldwork in the High Atlas proper, Hammoudi in the Dra valley. Almost nobody who may still remember the pre-modern tribal organization is alive and the research into pre-modern condition has become purely ethnohistorical. Gellner himself did not make any explicit distinction between the micro and macro-social level but the way he initially presented and later defended his model points in this direction. In Saints of the Atlas he hardly provided any ethnographic descriptions of real events and his defense against Munson was coined in purely systemic, i. e. macro-sociological terms.

I will end this commentary by one final remark. Gellner's way of thinking reflects in the kind of models he formulates. He openly admitted it in his interview with Davis: ‘... I very much like neat, crisp models and I try to pursue them, and I would be very uncomfortable if I didn’t have one’ (Davis 1991: 71). Gellner aims at models that are not only logically coherent and highly abstract (they have almost mathematical qualities) but also parsimonious – they use the least number of concept and logical connections to explain the largest possible part of reality.

LAW AND ORDER

There is a deeper level on which the segmentary theory needs to be assessed. It has to do with the initial question that stimulated Gellner to formulate his model in the first place. The whole debate does not make sense if the initial question about the maintenance of social order is misplaced. In other words, is the Hobbesian framework justified? This is one of the basic themes of political philosophy that is beyond both my abilities and the scope of the present essay (cf. Caton 1987: 78, Eickelman 1981: 99). Here, I will limit myself to a few basic observations.

A strong case for the importance of specifically tribal political arrangements can certainly be made. A good example is an overview of reasoning about the tribe as opposed to the state by Crone where she defines the tribe as a ‘descent group which constitutes a political community’ (Crone 1986: 51). According to this classical view ‘...the resolution of conflict rests on self-help, and one of the prime objectives of the tribal organization ... is to regulate and mitigate the disruptive effects of its use’ (idem, p. 50). Where do the contributors to the segmentary debate stand with regard to this school of thought? Gellner's position it more than clear and it supersedes the only earlier theory of political life of the Berber tribes developed by a French scholar Robert Montagne (1973). Among contemporary scholars Munson did not offer any such theory (Gellner wasted no time to point out this deficiency). The culturalists did not deal with tribal politics which is hardly surprising given their different research focus. This leaves us with Hammoudi who did offer an explanation.

In contrast to Gellner, Hammoudi emphasized stratification and the peculiar role the tribes played in wider Moroccan society. For Gellner, the society whose law and order has to be explained are the lay tribes. In his model the saints are external to the lay tribes though at the same time their role as mediators makes the segmentary system work. The result is a strictly synchronic argument. Hammoudi, on the other hand, take a more historical approach. He views both the lay tribes and the saints as parts of the wider Moroccan society with the saints being a special kind of ruling stratum. His broader focus, the different unit of analysis, bypasses the need for a separate political theory of the lay tribes. On the other hand, a historical dimension does not necessarily disqualify the segmentary model. Again, further research would be needed to ascertain what phenomena are better amenable to Hammoudi's historical interpretation and what can be explained in purely functionalist terms.

SEGMENTARY THEORY IN A WIDER PERSPECTIVE

In ethnographic writing on the Middle East and North Africa the segmentary theory has taken various directions. In what way can they illuminate our present discussion? Three questions are important here. Does the author address the issue of social order? What status is ascribed to folk (segmentary) lineage ideology? What status is ascribed to segmentary model of social action?

The most skeptical view is that of Peters (1967) who researched the Bedouin of Cyrenaica (the same ethnic group that had been studied by Evans-Pritchard in the late 1940s). He claims that the segmentary lineage model is simply a folk model the Bedouins hold about their society and anthropologists such as Evans-Pritchard simply mistook it for a sociological model of how the Bedouin society really functions. For Peters the folk segmentary ideology has only a cognitive function, it ‘enables them [the Bedouin] ... to understand their field of social relationships, and to give particular relationships their raison d’être’ (1967: 270, original emphasis). In Peters' view social action can be explained in non-segmentary, non-lineage terms, ultimately in terms of group competition for scarce resources (the reasoning behind the argument is basically adaptationist). With regard to our questions Peters does not distinguish between lineage system and segmentary political system (cf. Salzman 1978: 64) and he dismisses segmentary lineage model as one thing. In doing so he does not have much to say about social order besides stating that many factors influence relationships between groups.

Peters' account was challenged by Salzman using Peters' own ethnography plus material about Iranian and Somali tribes (1978). He acknowledges importance of non-segmentary factors such as affinal, matrilineal, and residential ties etc. emphasized by Peters but he qualifies his conclusion on several points. The most important of them is Salzman's argument that segmentary folk ideology represents a ‘social structure in reserve’, a political framework that may be disregarded at times of territorial stability but is activated for social mobilization under specific ecological and historical circumstances, namely increased nomadic movements, population mixing, and related uncertainty. A similar point was later made by Kraus who argues that patrilineal ties come to the fore when the tribe is expanding into a new territory or is threatened by another tribe's expansion, and recede into background if it has successfully stabilized in a new area (Kraus 1998). Salzman is aware of the difference between lineage system and segmentary political system but he primarily focuses on the relationship between lineage ideology and social action rather than social order itself.

Another group of authors did not dismiss the segmentary/lineage model altogether but tried to move it to a symbolic level. Caton (1987) re-analyzed segmentary ethnographies of Evans-Pritchard, Peters, Gellner, Barth (Iran), and Jamous (Morocco) and followed Jamous in claiming that segmentary character of Berber tribes derives from symbolic behavior concerning the concept of honor. This is in sharp contrast to Gellner for whom groups form because of shared interests (1969: 50). More importantly, Caton proposed a new solution to the issue of social order. According to him, in seemingly egalitarian tribes power resides in the ability of chiefs to persuade through the use of socially valued political rhetoric. Tribal leaders, powerless in terms of physical coercion, enforce their decisions through the use of symbolic power.

A purely structuralist account of segmentary theory was offered by Dresch in his case study of a Yemeni tribe (1986). He separates lineage theory from (what he calls) ‘segmentation’ but along different lines than other authors. In his understanding lineage theory describes how ‘solidary groups form, and then combine or conflict, in predictable ways within a system sustained by a balance of power’ (idem, p. 309). Segmentation, on the other hand, deals with formal relations and is characterized by balanced opposition of concepts (such as honour) on multiple levels. The resultant deep structure contains implicit assumptions people hold about the nature of social order, exists prior to any action and gives meaning to it (cf. Galaty 1981 for the Maasai). Dresch compares this ‘segmentary operator’ to ‘matrilateral operator’ of Lévi-Strauss (Dresch 1986: 320, 323). The segmentary principle is supposed to be the defining feature of Middle Eastern tribalism.

A cognitive version of segmentary theory was proposed by Maynard (1988). He compares allegiance to Protestant churches in urban Ecuador with segmentary organization of the Nuer as described by Evans-Pritchard. Maynard's main argument is that social identity relies on segmentary hierarchical taxonomies of cognition. These cognitive taxonomies supply the segmentary form of social identities while phenomena such as unilineal kinship supply their content (cf. Gellner 1996: 639, Kraus 1998: 17). Human agency is the third key concept in Maynard's equation: ‘...social identity lends to sociocultural organisation the propensity to be segmentary, yet ultimately the actual shape of social practice is relative to human agency and the unintended consequences of action’ (Maynard 1988: 114).

In what way do these accounts contribute to our understanding of the Moroccan debate? The magic seems to be that there is something inherently logical about segmentary processes. For anybody who takes the segmentary model seriously a question immediately arises: Where is the logic coming from? Where is it located? It is only processualists such as Holy (1979b) who are not faced with this dilemma. Others are forced to take one of the three following positions:

(1) The segmentary logic is a property of an explicit folk segmentary (lineage) ideology. Authors who want to base their argument on the ideal normative rules (Evans-Pritchard) have problems to explain social action which is always more complex. Others prefer to emphasize ‘impurities’ and other factor that shape events (for example Salzman who combines a concept of segmentary ideology as a cultural resource with adaptationist, functionalist, and instrumental elements). In both cases the segmentary ideology is understood as a factor constraining behaviour. And there are those such as Peters for whom the segmentary ideology simply makes behaviour meaningful while not having any constraining effect upon it.

(2) Authors such as Maynard place the segmentary model in human cognition and relate it to the way humans classify phenomena. Gellner may fall into this group as well (Kraus 1998: 2). In this regard it is worth noticing that in one of his early articles Gellner suggests that there may be a relation between mathematical logic and kinship (Gellner 1957; cf. Hall f. c.).

(3) For others (Caton, Dresch, Galaty) segmentary model has implicitly semiotic structure. Its units are defined by relationships to other units. Language is the ultimate template and we are studying a behavioral grammar.

CONCLUSION

Why did the segmentary debate in anthropology of Morocco evolve the way it did? Kraus touches upon this question:

I would generally agree with this assessment with two additional comments. First, the more interesting part of the debate, the one between Hart, Munson, Gellner, Hammoudi, Kraus, and Combs-Schilling did have an empirical character and lead to significant progress of our knowledge about the Berber tribal organization. Secondly, I would not say that the debate concentrated on epistemological issues ‘openly’, rather the opposite. It was only Gellner and, on one occasion, Geertz who made unambiguous epistemological statements in the context of the segmentary debate. Geertz (1971) made one laconic remark in his brief review of Saints of the Atlas published in the New York Review of Books. Gellner put forward a sustained critique of both the epistemological and ethnographic position of the culturalists but nobody reacted to it. Thus the actual reasons behind the culturalist rejection of the segmentary model remained largely unspoken which undoubtedly contributed to the confusion surrounding the debate.

Generally, the segmentary debate reflects the dynamics of the discipline in the 1970s and 1980s. Some called this period in the history of social sciences linguistic or relativistic turn. In anthropology it saw the culturalist school trying to win over already established anthropological approaches. Recently, there was a small-scale replay of the debate this time concerning the use of segmentary theory in Iranian nomadic tribes. It strongly echoed the Moroccan example and involved a full spectrum of perspectives from structural-functionalist to postmodernist (Barth 1992; Ganzer 1994; Salzman 1995; Street 1992, 1995; Wright 1992, 1994, 1995).

However, these historical considerations should not distract our attention from a series of coincidences that shaped the segmentary debate in anthropology of Morocco. The fact that two very prominent scholars – Geertz and Gellner – conducted long-term fieldwork so close to each other is one of them. So it happened that culturalist school in its purest form (Geertz) encountered one of the staunchest proponents of causal explanation in anthropology (Gellner). And from Hall's recent account (Hall f. c.) it is clear that in Gellner the culturalists encountered not only a scholar holding opposite philosophical assumptions but someone whose conviction about the correctness of these assumptions stemmed from a very particular life history. The second noticeable coincidence are the circumstances informing the position of most American fieldworkers. The result is known. The commanding theoretical influence of Clifford Geertz must have been caused by a mix of factors: Geertz's scholarly reputation, his literary qualities, the fact that he led a team of graduate students and supervised some of them, perhaps even his personality.

NOTE

_* I wrote an earlier version of this essay to conclude my MA studies in anthropology at McGill University in Montreal. The text has its limitations, the most obvious being the fact that it bypasses the francophone contributions to the debate by Jacques Berque and other scholars. I thank to John Hall for useful comments, Peter Skalník for editorial help, and Petr Paroulek for technical support. Contact: radim.tobolka@elf.mcgill.ca.

REFERENCES

Abu-Lughod, J.

1985. Bargaining for Reality: The Construction of Social Relations in a Muslim Community, by Lawrence Rosen. Chicago: University of Chicago Press, 1984 [book review]. Contemporary Sociology 14: 47–48.

Amselle, J.-L.

1998 [1990]. Mestizo Logics: Anthropology of Identity in Africa and Elsewhere. Stanford: Stanford University Press.

Barth, F.

1992. Method in our critique of anthropology. Man (N. S.) 27: 175–177.

Bigelow, J. C.

1998. Functionalism in social science. In Craig, E. (ed.), Routledge Encyclopedia of Philosophy (vol. 3, pp. 813–819). London: Routledge.

Caton, S. C.

1987. Power, persuasion and language: a critique of the segmentary model in the Middle East. International Journal of Middle East Studies 19: 77–102.

Combs-Schilling, M. E.

1985. Family and friend in a Moroccan boom town: the segmentary debate reconsidered. American Ethnologist 12: 659–675.

Crapanzano, V.

1980. Tuhami: Portrait of a Moroccan. Chicago: University of Chicago Press.

Crone, P.

1986. The tribe and the state. In Hall, J. A. (ed.), States in History (pp. 48–77). Oxford: Basil Blackwell.

Davis, J.

1991. An interview with Ernest Gellner. Current Anthropology 32: 63–72.

Dresch, P.

1986. The significance of the course events take in segmentary systems. American Ethnologist 13: 309–324.

Durkheim, É.

1893. De la division du travail social. Paris: Alcan.

Dwyer, K.

1982. Moroccan Dialogues: Anthropology in Question. Baltimore: The Johns Hopkins University Press.

Eickelman, D. F.

1976. Moroccan Islam: Tradition and Society in a Pilgrimage Center. Austin: University of Texas Press.

1981. The Middle East: An Anthropological Approach. Englewood Cliffs, NJ: Prentice-Hall.

1989. The Middle East: An Anthropological Approach, 2^nd edition. Englewood Cliffs, NJ: Prentice-Hall.

Evans-Pritchard, E. E.

1940a. The Nuer of the southern Sudan. In Fortes, M., and Evans-Pritchard, E. E. (eds.), African Political Systems (pp. 272–296). London: Oxford University Press.

1940b. The Nuer: A Description of the Modes of Livelihood and Political Institutions of a Nilotic People. London: Oxford University Press.

1949. The Sanusi of Cyrenaica. London: Oxford University Press.

Fortes, M., and Evans-Pritchard, E. E. (eds.)

1940a. African Political Systems. London: Oxford University Press.

Fortes, M., and Evans-Pritchard

1940b. Introduction. In Fortes, M., and Evans-Pritchard, E. E. (eds.), African Political Systems (pp.1-23) London: Oxford University Press.

Fortes, M.

1945. The Dynamics of Clanship among the Tallensi. London: Oxford University Press.

1953. The structure of unilineal descent groups. American Anthropologist 55: 17–41.

1969. Kinship and the Social Order: The Legacy of Lewis Henry Morgan. Chicago: Aldine Publishing Company.

Galaty, J. G.

1981. Models and metaphors: on the semiotic explanation of segmentary systems. In Holy, L. (ed.), The Structure of Folk Models (pp. 63–92). London: Academic Press.

Ganzer, B.

1994. Constructing tribal identity in Iran. Man (N. S.) 29: 182–183.

Geertz, C.

1968. Islam Observed: Religious Development in Morocco and Indonesia. New Haven: Yale University Press.

1971. In Search of North Africa. New York Review of Books, 22 April, pp. 20–24.

1979. Suq: the Bazaar Economy in Sefrou. In Geertz, C., Gertz, H., and Rosen, L., Meaning and Order in Moroccan Society: Three Essays in Cultural Analysis (pp. 123–313). Cambridge: Cambridge University Press.

1982. Conjuring with Islam. New York Review of Books, 27 May, pp. 25–28.

Geertz, H.

1971. Saints of the Atlas. By Ernest Gellner. Chicago: University of Chicago Press, 1969 [book review]. American Journal of Sociology 76: 763–766.

1979. The Meaning of Family Ties. In Geertz, C., Geertz, H., and Rosen, L., Meaning and Order in Moroccan Society: Three Essays in Cultural Analysis (pp. 315–391). Cambridge: Cambridge University Press.

Geertz, C., Geertz, H., and Rosen, L.

1979. Meaning and Order in Moroccan Society: Three Essays in Cultural Analysis. Cambridge: Cambridge University Press.

Gellner, E.

1957. Ideal Language and Kinship Structure. Philosophy of Science 24: 235–242.

1969. Saints of the Atlas. London: Weidenfeld and Nicolson.

1972a. Introduction. In Gellner, E., and Micaud, C. (eds.), Arabs and Berbers: From Tribe to Nation in North Africa (pp. 11–21). London: Duckworth.

1972b. Political and Religious Organization of the Berbers of the Central High Atlas. In Gellner, E., and Micaud, C. (eds.), Arabs and Berbers: From Tribe to Nation in North Africa, (pp. 59–66). London: Duckworth.

1981a. Muslim Society. Cambridge: Cambridge University Press.

1981b. Flux and Reflux in the Faith of Men. In Gellner, E. Muslim Society (pp. 1–85). Cambridge: Cambridge University Press.

1981c. Saints and their Descendants. In Gellner, E. Muslim Society (pp. 207–213). Cambridge: Cambridge University Press.

1983. The Tribal Society and its Enemies. In Tapper, R. (ed.), The Conflict of Tribe and State in Iran and Afghanistan (pp. 436–448). London: Croom Helm.

1985. The Roots of Cohesion [review article]. Man (N. S.) 20: 142–155.

1995a. Segmentation: Reality of Myth? [Comment on Munson 1993]. Journal of the Royal Anthropological Institute (N. S.) 1: 821–829.

1995b. The Politics of Anthropology. In Gellner, E., Anthropology and Politics: Revolutions in the Sacred Grove (pp. 11–26). Oxford: Blackwell.

1996. Reply to Critics. In Hall, J. A., and Jarvie, I. (eds.), The Social Philosophy of Ernest Gellner (pp. 623–686). Amsterdam: Rodopi.

Hall, J.

2000. The Sociology of Islam. McGill University: Manuscript. In Hall, J. (f. c.), Ernest Gellner: An Intellectual Biography. London: Verso.

2003. Ernest Andre Gellner (1925–1995). McGill University: Manuscript. In Hall, J. (f. c.), Ernest Gellner: An Intellectual Biography. London: Verso.

2004 (f. c.). Ernest Gellner: An Intellectual Biography. London: Verso.

Hammoudi, A.

1996. Segmentarity, Social Stratification, Political Power and Sainthood: Reflections on Gellner's Theses. In Hall, J. A., and Jarvie, I. (eds.), The Social Philosophy of Ernest Gellner (pp. 265–289). Amsterdam: Rodopi.

Hart, D. M.

1976. The Aith Waryaghar of the Moroccan Rif: An Ethnography and History. Tucson: University of Arizona Press.

1981. Dadda `Atta and His Forty Grandsons: Socio-political Organisation of the Ai` Atta of Southern Morocco. Wisbech: Middle East and North African Studies Press.

1984. The Ai` Atta of Southern Morocco: Daily Life and Recent History. Wisbech: Middle East and North African Studies Press.

1989. Rejoinder to Henry Munson Jr. American Anthropologist 91: 765–769.

1994. Faulty Models of North African and Middle Eastern Tribal Structures. Revue du Monde Musulman Mediteranéen 68–69: 225–238.

1996. Segmentary Models in Morocco. Journal of the Royal Anthropological Institute (N. S.) 2: 721–722.

2000. Tribe and Society in Rural Morocco. London: Frank Cass Publishers.

Holy, L. (ed.)

1979a. Segmentary Lineage Systems Reconsidered. Belfast: Queen's University Press.

1979b. Segmentary Lineage Structure and its Existential Status. In Holy, L. (ed.), Segmentary Lineage Systems Reconsidered (pp. 1–22). Belfast: Queen's University Press.

1996. Anthropological Perspectives on Kinship. London: Pluto Press.

Kincaid, H.

1998. Positivism in the Social Sciences. In Craig, E. (ed.), Routledge Encyclopedia of Philosophy (vol. 7, pp. 558–561). London: Routledge.

Kraus, W.

1998. Contestable Identities: Tribal Structures in the Moroccan High Atlas. Journal of the Royal Anthropological Institute (N. S.) 4: 1–22.

Kuper, A.

1982. Lineage Theory: a Critical Retrospect. Annual Review of Anthropology 11: 71–95.

1996. Anthropology and Anthropologists: The Modern British School. London: Routledge.

1999. Culture: The Anthropologists' Account. Cambridge, MA: Harvard University Press.

Maine, H. S.

1861. Ancient Law. London: Murray.

Manning, R. N.

1998. Functional Explanation. In Craig, E. (ed.), Routledge Encyclopedia of Philosophy (vol. 3, pp. 802–805). London: Routledge.

Maynard, K.

1988. On Protestants and Pastoralists: the Segmentary Nature of Socio-cultural Organization. Man 23: 101–117.

Montagne, R.

1973. The Berbers: Their Social and Political Organization. London: Frank Cass.

Morgan, L. H.

1877. Ancient Society: Researches in the Lines of Human Progress from Savagery through Barbarism to Civilization. New York: Holt.

Munson, H. J.

1989. On the Irrelevance of the Segmentary Lineage Model in the Moroccan Rif. American Anthropologist 91: 386–400.

1993. Rethinking Gellner's Segmentary Analysis of Morocco's Ayt Atta. Man 28: 267–280.

1995. [Reply to Gellner 1995]. Journal of the Royal Anthropological Institute (N. S.) 1: 829–832.

Munson Jr., H.

1984. The House of Si Abd Allah: The Oral History of a Moroccan Family. New Haven: Yale University Press.

Peters, E.

1967. Some Structural Aspects of the Feud among the Camel-herding Bedouin of Cyrenaica. Africa 37: 261–282.

Rabinow, P.

1975. Symbolic Domination: Cultural Forms and Historical Change in Morocco. Chicago: University of Chicago Press.

1977. Reflections on Fieldwork in Morocco. Berkeley: University of California Press.

Radcliffe-Brown, A. R.

1965. Structure and Function in Primitive Society. New York: Free Press.

Robertson-Smith, W.

1885. Kinship and Marriage in Early Arabia. Cambridge: Cambridge University Press.

Rosen, L.

1984. Bargaining for Reality: the Construction of Social Relations in a Muslim Community. Chicago: University of Chicago Press.

Salzman, P. C.

1978. Does Complementary Opposition Exist? American Anthropologist 80: 53–70.

1995. Understanding Tribes in Iran and Beyond. Journal of the Royal Anthropological Institute (N. S.) 1: 399–403.

Street, B. V.

1992. [Reply to Barth 1992]. Man (N. S.) 27: 177–179.

1995. [Reply to Salzman 1995]. Journal of the Royal Anthropological Institute (N. S.) 1: 403–404.

Waterbury, J.

1991. Bargaining for Segmentarity. In Joffé, G., and Pennell, C. R. (eds.), Tribe and State: Essays in Honour of David Montgomery Hart (pp. 4–13). Wisbech: Middle East and North African Studies Press.

Wright, S.

1992. Method in our Critique of Anthropology: a Further Comment. Man (N. S.) 27: 642–644.

1994. [Reply to Ganzer 1994]. Man (N. S.) 29: 183–186.

1995. [Reply to Salzman 1995]. Journal of the Royal Anthropological Institute (N. S.) 1: 404