Since the colonial era, western representations of Inner Asia, its past and traditions, have been dominated by images of fierce and free nomads, organized by the principles of pre-state kinship society into clans and tribes. Since this evolutionary model of political society has been largely rejected within anthropology, the discipline that developed it, I suggest a critical reevaluation of the application of the clan-tribe model to Inner Asian nomadic societies.^[1] This means looking again at dominant conceptions of political society more generally, since the current notion of tribe is entangled in the wider ideology of the nation-state. Dominant conceptions of identity reflect the influence of national populist thought, which conceived of volk (people), culture, and society as mass phenomena; autochthonous, grassroots entities, thought of in terms of commonality and solidarity. In this view, tribes and peoples were protonational units, sharing common cultural and social forms. Such notions can be contrasted with those generated by what Benedict Anderson^[2] calls the dynastic realm – in which political society is a product of rulership. This represents a very different view of the social units that became the subjects of empire, but it does, I think, serve as a better starting point for understanding the history of nomadic Inner Asia up until the twentieth century. It also helps us to suspend our commitment to the notion of “ethnic identity” as separate from power relations. In this article I try to summarize arguments proposed in my recent book The Headless State, and to suggest that categories commonly taken as “ethnic” represent historical processes of political designation rather than self-forming community.

THEORIES OF KINSHIP SOCIETY AND EUROPEAN COLONIALISM

By the late twentieth century, critiques by Godelier,^[3] Fried,^[4] and others had led anthropologists to shun the term “tribe.” Works such as Vail’s The Creation of Tribalism in Southern Africa^[5] had helped establish the view that “tribalism” was a product of colonial classification and administration. However, this critique has not seriously undermined the application of the model of tribal society to Inner Asia, or indeed much of the Middle East, where the institutions of “the state” date back to the beginning of recorded history. Here, scholars continue to apply parts or all of the tribal society model, although the analytical emptiness of the term becomes clear whenever it is seriously considered.
Notions of tribal society reflected the evolutionist social theory of the nineteenth century. Morgan, Maine, Marx, and McLennan all saw extended ties of kinship as forming the basis for pre-state society, later giving way to territory as the basis for social organization in civilizations. This theory of change, by which egalitarian kinship society preceded impersonal class society, was a powerful influence on the conceptions of tribe and clan developed within the social sciences. The most influential thinker in this regard, both for anthropology and Marxism, was Lewis Henry Morgan. For him “all forms of government are reducible to two general plans… The first, in the order of time, is founded upon persons, and upon relations purely personal, and may be distinguished as a society (societas). The gens [clan] is the unit of organization… The second is founded upon territory and upon property, and may be distinguished as a state (civitas).”^[6] Morgan’s work, as well as articulating the evolutionism of the age, was also intensely ideological. His commitment to the articles of faith of populist democratic politics led him to conclude that the most ancient forms of society – kinship society organized into gens (clans) – must embody the principles of liberty, equality and fraternity. He proposed a general scheme for kinship society in which clan members were thought to be united by common descent. These clans came together to form tribes, and these grouped themselves together to form “confederations.”^[7]
In fact there is nothing to suggest Morgan’s scheme was correct, even for ancient Greece. First, the “clan” does not seem to predate the state at all, but probably appeared after it, and like the later Roman gentes (clan), seems to be primarily an institution of the elite. Chester Starr remarks “The genos (clan) only became important when aristocrats began to play a central role in Greek political life.”^[8] Second, the phyle (tribe) does not seem to have been a kinship unit, but an administrative one, in which people were registered for the purposes of local civil and military government.^[9] Indeed, as Jack Goody notes, it is doubtful if even the classical Roman agnatic descent group, the gens, was ever the dominant organisational form that Morgan supposed.^[10] It was almost certainly not the pre-state social building block that the nineteenth-century social theory assumed. Tim Cornell for example notes that the gens spread with the process of urbanization and administration. “This feature of the evidence runs counter to the well-entrenched nineteenth-century theory that the gens originated as a ‘pre-political’ organisation, which was weakened and ultimately eclipsed by the rise of the state. In fact the evidence implies the contrary.”^[11]
Nevertheless, in the nineteenth century, Morgan’s theory was highly convincing and widely accepted. It was selectively adopted by Marx and Engels and so remained highly influential in the Soviet ethnography of Inner Asia. In western anthropology, Morganian theory was elaborated in the twentieth century with the structural functionalist model of segmentary kinship proposed by Fortes and Evans-Pritchard in “African Political Systems”, by which the units generated by kinship formed a nested series of descent groups. All these schemes shared a common vision of primitive society, and as Adam Kuper notes “The idea of primitive society served imperialists and nationalists, anarchists and Marxists.”^[12] Twentieth-century social science, both western and Soviet, represented nonstate society as made up of clans that united to form tribes and tribal confederations.
The era of European colonialism powerfully exported the term “tribe” as an administrative category and established it as a description of collective identity throughout much of the colonized world. Terence Ranger takes John Iliffe’s description of the creation of tribes in colonial Tanganyika (former German East Africa) as typical of the wider colonial process, and is worth quoting at length:

“The notion of the tribe lay at the heart of indirect rule in Tanganyika. Refining the racial thinking common in German times, administrators believed that every African belonged to a tribe, just as every European belonged to a nation. The idea doubtless owed much to the Old Testament, to Tacitus and Caesar, to academic distinctions between tribal societies based on status and modern societies based on contract, and to the post-war anthropologists who preferred “tribal” to the more pejorative word “savage.” Tribes were seen as cultural units “possessing common language, a single social system, and an established common law.” Their political and social systems rested on kinship. Tribal membership was hereditary. Different tribes were related genealogically.… As unusually well-informed officials knew, this stereotype bore little relation to Tanganyika’s kaleidoscopic history, but it was the shifting sand on which Cameron and his disciples erected indirect rule by “taking the tribal unit.” They had the power and they created the political geography.”^[13]

Anthropology inherited the terms “tribe” and “chief” from a colonial order in which civilizations were governed by monarchs and aristocrats, and more primitive “tribal” societies were ruled by chiefs and paramount chiefs. The colonial usage was later rationalized by structural functionalist modeling of many African political systems as smaller in scale, less centralized, and territorially looser than “the state.”^[14] But as the true grounds for the term had always been to distinguish civilized from barbaric societies, this terminology did not correspond to descriptions of any European forms of leadership since the Dark Ages. Elizabeth Colson notes the double standards of this “tribalizing” discourse.

“In terms of territory, population, wealth, bureaucratic development, social stratification, and the centralization of power, the Hausa state of Kano far surpassed many of the kingdoms of Medieval Europe. Yet most of those who referred to the Hausa as a tribe were not being facetious in the fashion of Weatherford when he wrote of the tribes of Washington.… Too many social scientists, as well as the general public, use [tribe] to maintain a false distinction between us and them, those people who used to be called primitive because they did not originate within the European tradition. Tribe, then, signals something about political domination but says nothing about the social complexity or political organization, now or formerly, of those to whom it is applied who may or may not have formed a polity in the past or present. In the 17th century when English-speaking explorers and settlers dealt with Native Americans as politically independent societies, they commonly referred to them as nations, placing them thus on a par with European nations.… As it became possible to ignore and inexpedient to recognize the full sovereignty of Native American rivals with whom the English settlements competed for land and political dominion, “nation” gave way to “tribe” which carried implications of lesser political status. Tribe thereafter became the term commonly used to distinguish among the populations being incorporated into colonial empires as these were created during the 19th century.”^[15]

QUESTIONING MORGANIAN KINSHIP SOCIETY

In western anthropology, despite some disquiet, the idea of tribal society as organized primarily by kinship rather than territory had been retained, although it was acknowledged that there was rather little ethnographic evidence that clearly illustrated this. However, in the second half of the twentieth century, the Morganian model of nested descent groups began to draw criticism. As early as 1961, Edmund Leach had suggested that the supposed structure of unilineal descent groups might be a “total fiction.”^[16] The ethnographic evidence for the segmentary kinship model had always been rather slight, and what there was faded away in the light of more critical later studies.^[17] As Adam Kuper^[18] pointed out, the actual local categories used to designate groups of people did not resemble those of descent theory at all.^[19] There is no Nuer word for “clan,”^[20] no Tallensi word for “lineage” or for the segments of a maximal lineage,^[21] and so on. As the dangers of foregrounding the analyst’s categories at the expense of the indigenous ones became more apparent, these descriptions simply lost their plausibility.
As Kuper writes “in the end even the Nuer, Tiv, and Talis cannot be said to have ‘true segmentary lineages.’” He concludes that “the lineage model, its predecessors and its analogs, have no value for anthropological analysis.… First, the model does not represent folk models which actors anywhere have of their own societies. Secondly, there do not appear to be any societies in which vital political or economic activities are organized by a repetitive series of descent groups.”^[22] Without its grounding in a distinctive form of kinship organization, the characteristics that are supposed to distinguish tribal society are reduced to the idea that certain societies represent a “pre-state” stage in political evolution. But the Weberian definition of the state – a political community possessing a monopoly of the legitimate use of force within a given territory^[23] – was actually devised to help distinguish the “modern” state. It was used by the structural-functionalist theoreticians to underpin the notion that the tribe was at least one evolutionary stage behind “the state,” just as Marxist theory used the notion of class relations to make the same distinction. In retrospect, this approach seems limited historically and culturally, and poorly suited for application beyond Weber’s original subject. Applied to the medieval world, for example, this definition of state would rule out almost every kingdom in Europe, as the Roman Church retained independent judicial jurisdiction over significant parts of most countries until well into the sixteenth century. That we recoil from the idea that France or England should be regarded as tribal chiefdoms until – say – the Bourbon or Elizabethan periods, reveals the residual ethnocentric evolutionism of the term. Europe had “passed that stage,” it had entered “historical time,” whereas polities in much of the rest of the world were viewed as examples of earlier stages, particularly in areas without systematic written histories to contradict the models projected onto them by the theorist. Marxist evolutionary schemes had the same tendency. Service provides a typical justification for keeping Europe out of the evolutionary scheme imposed on “the rest.” “European feudalism, then, was historically of a very complex, and perhaps unique sort. For this reason it cannot be considered a stage in evolution, or even a usual case of devolution.”^[24]
Although an earlier generation of scholars of Africa such as Rattray and Nadel had described precolonial African states in terms of feudalism, in western anthropology this terminology lost ground to the structural functionalist scheme, which presented itself as a more general and objective classification of political society.^[25] Similarly, Boris Vladimirtsov’s 1934 study of Mongolian society had described it as “nomadic feudalism,”^[26] but this approach gave way before the advance of the tribal discourse of Vasilii Radloff, Vasilii Barthold, Alice Fletcher, Anatoly Khazanov, and Lawrence Krader. The effect was to reproduce the old conceptual apartheid by which the colonized were subject to the primitivist language of tribe and chief, while the colonizers used such terms only for the most distant eras of their own historical narratives.

VISIONS OF THE TIMELESS TRADITIONAL NOMAD

Nineteenth-century evolutionist social science had pictured pastoralism as a sort of developmental dead end. In Engels’s scheme, for example, the evolutionary stages that preceded “civilization” began with “savagery” and progressed to “barbarism.” Each of these stages were characterized by a certain technological level: savagery by fire, hunting, and fishing, and barbarism by pottery and the domestication of plants and animals. Pastoral nomadism was thought of as essentially preagricultural, emerging in the “middle stage” of barbarism. Having branched away from the technological and social development that led to agriculture and successively higher forms of civilization, pastoralists had “gotten stuck” at this early stage. This may have been a plausible enough scenario in Engels’s time, but since then archaeology has shown it to be quite mistaken. It has become widely accepted that Eurasian pastoral nomadism developed long after the advent of agriculture, probably as a more specialized offshoot of the mixed farming that had developed in areas such as Southwestern Asia between the fourth and eighth millennium BCE.^[27]
But the notion of nomadic society as essentially tribal, traditional, and timeless remained influential. In his monumental study of world history, Arnold Toynbee describes it as one of several “arrested civilizations.” Along with the Eskimos and Polynesians, “nomads” represented “cases in which civilizations have remained static.”^[28] This was an essentialized vision of nomadism as a social form that had remained unchanged for countless generations, like the societies of ants and bees. Toynbee’s treatment helped set the scene for ecologically determinist accounts of nomadic pastoralism.
In western anthropology an explicit model of pastoral nomadic society emerged as a Weberian ideal-type. Central to this characterization was the persistent notion of the egalitarian nomad. When William Irons stated “among pastoral nomadic societies, hierarchical political institutions are generated only by external political relations with state societies, and never develop purely as a result of the internal dynamics of such societies”^[29] he was simply making explicit the implicit characteristics that both structural-functionalist and Soviet anthropology had already attached to the category, and which reflected the long-standing stereotypes of the “fierce and free” pastoralist. Many arguments were put forward to explain why nomads ought to be egalitarian. Some declared that the ecological constraints of pastoralism did not allow for the accumulation of wealth needed for hierarchical strata to develop.^[30] Others emphasized the political, rather than economic, constraints generated by mobility.^[31]
But by the 1970s this ideal-type of segmentary, egalitarian pastoral nomadic society began to draw criticism on both theoretical and empirical grounds. Telal Asad’s 1970 study of the Kababish Arabs^[32] showed they had clear hierarchy and hereditary rulers, and he formulated a Marxian critique of both nomadic pastoralism as an ideal-type, and the notion that pastoralists tended to be egalitarian. The more anthropologists studied societies they designated “pastoral nomadic” the clearer it became that they were diverse and different. That they all happened to employ mobility in raising livestock simply reflected the grounds for their inclusion in the category, and implied very little about their wider social orders. As Spooner acknowledges, “there are no features of cultural or social organisation that are common to all nomads or even that are found exclusively among nomads.”^[33] By the 1970s the notion of “pastoral nomadism” as an ideal-type had been all but abandoned by anthropological specialists, and in their review of the topic, R. Dyson-Hudson and N. Dyson-Hudson predicted that by 1990 “[t]he typology ‘nomadic pastoralism’ which locks together livestock herding and mobility into a single arbitrary category will then cease to define a subject suitable for review.”^[34]

TRIBAL SOCIETY IN INNER ASIA

Nevertheless, the idea that nomadic society was essentially tribal had a profound impact on the scholarship of “exotic” societies such as those of Inner Asia. The translation of historical texts is revealing in this regard. The observant Franciscan Friar John de Plano Carpini, for example, traveled to the court of Güyük Khan (Chinggis Khan’s grandson) in 1246 and left the first detailed eyewitness account by a European of the Mongol polity. He describes a hierarchical and aristocratic society:

“The dukes (duces) have like dominion over their men in all matters, for all Tartars [Mongols] are divided into groups under dukes.... The dukes as well as the others are obliged to give mares to the Emperor as rent... and the men under the dukes are bound to do the same for their lords, for not a man of them is free. In short, whatever the emperor and the dukes desire, and however much they desire, they receive from their subjects’ property.”^[35]

Carpini used the Latin word dux for senior Mongol and European nobles alike, and early translators such as Hakluyt (1598) translated these as “duke” (as I have done in the passage above).^[36] However, nineteenth- and twentieth-century translators such as William Rockhill^[37] and Christopher Dawson^[38] introduced an astonishing dual system whereby, when dux referred to a European noble such as the Russian nobleman named Vassilko, it was translated as “duke,” but where it applied to a Mongol it was translated as “chief,” thus confirming the tribal model of Mongolian society.^[39]
A more recent example is the work of Paul Geiss, who sets out to study the “tribal people of Central Asia,” that is, the Kazakh, Kyrgyz, and Turkmen in his book Pre-Tsarist and Tsarist Central Asia: Communal Commitment and Political Order in Change.^[40] Social structure is presented in terms of “kinship society” organized by descent.

“Genealogies played an important role in Central Asian tribal societies. They represented the backbone of the society, built a societal web.… Tribesmen formed a body of agnatic kin and traced their origin from common ancestors. In this way descent group names were inherited through the male line. Mutual relations were established according to the closeness and distance of shared ancestors.… Genealogies based on primogeniture could inform orders of seniority between groupings.”^[41]

Another familiar approach, represented by Barfield, was to incorporate into descriptions of “typical” Inner Asian political forms the notion of “confederation” derived from Morgan’s theory.^[42] Similarly, Svat Soucek describes Mongolia up to the advent of communist control in the 1920s and 1930s as “a confederation of tribal groups governed by a two-pronged aristocracy of lay tribal and Buddhist church leaders.”^[43] This is rather like describing the Habsburgs as “tribal leaders,” since the Borjigin aristocracy of Mongolia were quite as well established and socially elevated as the nobilities of Europe. And, as Anatoly Khazanov has noted, the term “confederation” is entirely inappropriate for the description of Inner Asian political formations, since most of these appear to have been formed by conquest.^[44]
But dominated by the notion of the pastoral nomadic society as an ideal-type, the tendency has been to graft received wisdom regarding segmentary tribes onto what historical evidence there was, so as to fill the gaps left in the historical record. The result was a timeless, essentialized model of steppe nomadic society:

“Throughout Inner Asia historically known pastoralists shared similar principles of organisation alien to sedentary societies.... Patrilineal relatives shared common pasture and camped together when possible.… Tribal political and social organisation was based on a model of nested kinship groups, the conical clan… an extensive patrilineal kinship organisation in which members of a common descent group were ranked and segmented along genealogical lines.... Political leadership was restricted to members of senior clans in many groups, but from the lowest to the highest, all members of the tribe claimed common descent.…”^[45]

However, since ethnographic accounts confirming this model were almost entirely absent, the lack of evidence had to be explained away in the following way: “When nomads lost their autonomy to sedentary governments, the political importance of this extensive genealogical system disappeared and kinship links remain important only at the local level.”^[46] In any historical description, the argument went, one would only expect to find elements, remnants of the complete kinship system of the past.
In the Mongol case, however, it is clear that the lack of a segmentary lineage structure dates at least as far back as the fourteenth century. Slawoj Szynkiewicz, for example, notes that

“towards the end of the Yuan dynasty... the lineage organisation, composed of a network of maximal lineages going back to the forefather of the clan and subdivided into segments, can be assumed to have ceased to exist. The few surviving maximal lineages may have comprised the aristocracy.”^[47]

So the essentialized tribal society is envisaged as existing before the Chinggisid state, at a period for which we have almost no historical material. Barfield writes: “At the time of Temüjin’s [Chinggis Khan’s] birth the steppe was in anarchy. Segmentary opposition was the basic form of political organisation.”^[48] But what, exactly, is the evidence for this sort of segmentary kinship structure in that period?

THE CASE OF THE MEDIEVAL MONGOLS

At the time of Chinggis Khan’s birth, the steppes of what is now Mongolia were divided between a number of political entities (such as the Kereid, Merkid, Tatar, Jürkin, and Tayici’ud). In the dominant interpretation of the Monggol-un ni’uca tobciyan (The Secret History of the Mongols), these are described as tribes or clans. In fact we know very little about the internal organization of these entities. What is clear, however, is that these polities were aristocracies with common subjects, and that the distinction between commoner and nobility long predated Chinggis Khan’s rise to power.^[49] The term for groups defined largely (but probably not exclusively) through descent was yasu (“bone”); nobles were referred to as of caγan yasu – “white bone” – this color contrasting with black – qara – used for commoners.
There is plenty of evidence, if one chooses to look for it, that these ruling houses or lineages were not related by descent to the people they ruled. The people known as the Jürkin, for instance, were one of several named groups living in Mongolia at the time. They have been described as a tribe^[50] or a clan,^[51] and from this it would appear they formed a homogeneous kin-based unit composed of clansmen. In fact, it is clear from section 139 of The Secret History of the Mongols that this political unit had originally been formed by a ruler (Qabul Qagan) and that the term “Jürkin” more properly referred to the ruling lineage/house only: “Chinggis Qahan made such arrogant people submit (to him) and he destroyed the Jürkin oboγ. Chinggis Qahan made their people, and the subjects (they ruled), his own subjects.”^[52] The conquests of other polities, such as the Tatar and Tayici’ud, for example, are described in very similar terms, with the name of the polity attached most specifically to the nobility who were clearly distinguished from their subjects.^[53]
There is no single term that corresponds to use of the word “tribe” in translations of The Secret History of the Mongols. Instead a series of different words – irgen (people, subjects), ulus (polity, realm, patrimony, apanage), ayimag [aimag] (division, section, group, province), and qari(n) (foreign, subject) – have been translated as “tribe” in places in the text when the unit concerned is believed to be tribal. Similarly, there are two terms, oboγ [oboq, ovog] (family, lineage, line) and yasu (bone, descent, lineage), commonly translated as “clan” depending on the context; often the term has simply been inserted next to any group noun that the translator believes to be a clan.^[54]
The term most commonly translated as clan, oboγ, appears most frequently in sections 9-49, which describe nobles founding various oboγ. These had usually acquired names derived from the founder’s own personal names or characteristics, and some had very recently been brought into being. The key point to be made here is that constitution of oboγ appears to have been more political than genealogical. Most of the ancestors and their descendants do not give rise to named clans or lineages; recognition of descent from a common ancestor alone did not generate a sociopolitical grouping in the way that the classic segmentary lineage model would suggest. At least one oboγ, that of the Jürkin, was created by a king as an apanage for his son. This had happened in the relatively recent past. The founder of the Jürkin oboγ was Chinggis Khan’s great-uncle, and his descendants alone cannot have been very numerous before the oboγ was destroyed and its subjects added to Chinggis Khan’s own. However, the Jürkin, and the subjects they ruled, were clearly a force to be reckoned with at the time of Chinggis Khan’s struggle for power. And it seems that recruitment may have been at least partially by choice; we know of at least one noble member of the oboγ who was not a descendant of the founder.^[55]
Some scholars, particularly those working with the original texts and aware of difficulties with the clan model, reconciled the historical evidence with Marxist theory by assuming that – once again – the kinship system was in breakdown by the time of historical records. Zhou Qingshu, for example, a Chinese historian of the Yuan, notes than in pre–thirteenth-century Mongolian society descent groups (xing shi 姓氏) included many who were not related by blood.^[56] Rincin, the Inner Mongolian historian, wrote that sometime before the twelfth century “oboγ institutions also lost their blood descent (cisun udum) character, and gradually became a unit of local relationships (gajar nutug-un qaricagan-u nigece).”^[57]
Paul Ratchnevsky has a similar tendency. Knowing that The Secret History of the Mongols describes a clearly stratified society, he pushes the mythic nomadic kin-based era back to the turmoil before Chinggis is born. By the time of Chinggis, he writes, “the kinship group lost its homogeneous character.”^[58] As with accounts of other Inner Asian societies where there are historical records, we have to conclude that the fully elaborated kinship structure has “broken down” because we never quite find it.
In a recent study of the appearance of the term oboγ [oboq] in The Secret History of the Mongols, Pavel Rykin concludes that there are no grounds to conclude that it was a kinship group and that the primary evidence suggests that it was used to indicate both the rulers of a group of people and the group itself.^[59] This need not contradict the central importance of tracing descent for the noble families of that time. Deuchler notes that, in both Tang China and Koryŏ Korea, aristocratic descent groups were “loosely structured, but nevertheless placed great value on descent.”^[60]
It might be safer, then, to use a term other than “clan” to describe aristocratic descent groups, since the term both evokes the primitivist discourse of kinship society and has been given an unhelpful technical definition in anthropology. I found the “house” to be useful, since it carries no such theoretical baggage and is routinely used for descent groups among the European nobility.^[61] In retrospect, the account of the origins of the various oboγ in The Secret History of the Mongols appears more like a description of the foundation of ruling houses or petty dynasties than of clans and tribes as we usually understand these terms.^[62] The standard translation of terms, however, gives the impression that it describes a pre-state “kinship society” of clans and tribes.

THE CASE OF THE KAZAKH KHANATE

The treatment of Kazakh “traditional society” provides another example of this tendency. Paul Geiss describes them in terms of segmentary structures of agnatic descent groups that could have been read directly from the African political systems put forth by Evans-Pritchard and Fortes.^[63] Of course, a proper examination of all the relevant historical evidence would require far more space, and indeed specialist knowledge, than are available to me here. However, even a brief and inexpert examination of the evidence suggests that this model rests on distinctly sandy foundations.
The Kazakh Khanate^[64] dates from the mid-fifteenth century, when two princes named Janibeg and Qarai established an independent polity.^[65] By the end of the sixteenth century, the khanate had conquered Tashkent, which became the seat of the Kazakh dynasty until 1723. Eugene Schuyler notes that at the height of their power the Kazakh Khans commanded over a million men and mustered more than 300,000 troops.^[66] However, in the eighteenth century, Russian colonial power extended into the region and Abilay Khan, ruler of one of the three great political divisions (“hordes”) of the Kazakh, swore fealty to the tsar around 1730.
There is no doubt that this was another aristocracy, but ethnohistorical accounts have nevertheless represented the society as composed of segmentary clans. Krader, for example, describes the polity as a “clan-confederation,” even though it was recognized that the units termed “clans” were not actually consanguineous at all.^[67] Geoffrey Wheeler, for example, tells us that “[t]he so-called clan system, which was originally based on the union of a number of families related to one another and sharing a communal economy, had begun to change by the seventeenth century.”^[68] Again, the fact that the historical accounts did not support the classical kinship society model was explained away by the notion that it had already begun to break down by the time that the records were written.
The notion that “tribal” society must somehow form itself through bottom-up processes of aggregation continues to dominate historical interpretation. Thus, Soucek writes that “[b]y 1730 the Kazakhs… tribes had coalesced into three confederations, the aforementioned Greater, Middle and Lesser Hordes.”^[69] In fact, as Khazanov and the historical record shows, quite the opposite process had been taking place: the fragmentation of a previously more centralized polity. The 1840 account of the Russian official Aleksei Levchine makes it clear that the “tribes” were actually administrative divisions or apanages created by a top-down process of political demarcation. These subdivisions “took the names of their chefs.”^[70]
However, Levchine also describes the Kazakh political system as “anarchy,” and this has been taken as evidence of the fierce and free egalitarian nomads that seemed so plausible to evolutionist social theorists. But a closer reading of Levchine shows that this was not the egalitarian ordered anarchy of Evans-Pritchard’s segmentary kinship society. This was an early nineteenth-century concept of anarchy, much closer to Hobbes’s notion of the state of war, a lawless and savage insecurity, from which the ordinary Kazakhs might heartily hope to be saved by orderly Russian colonial administration.^[71] Indeed, the instability of Kazakh society in Levchine’s time was the result of the disruption of the earlier, more centralized, political system of the Khanate. According to Levchine the Kazakh looked back on the “golden age” of Tauke (Tiavka) Khan who ruled from 1680 to 1718. Although a law code is now known to have been introduced by Qasim Khan in the early sixteenth century, Levchine credits Tauke Khan with the introduction of a law code that he described as “customary.”^[72] Tauke Khan had been militarily defeated by the Kalmyk in 1698, and his successor Pulad Khan (1718–1730) was defeated by them again in 1723, causing severe disruption of the Khanate.^[73] The lack of law and order that Levchine laments was the result of the collapse of central authority, a situation frozen by Russian imperial expansion.
The colonial administration set about undermining and, where possible, destroying the power of the Kazakh Khans and the realms they ruled. In the case of the Middle Horde, when its ruler Vali Khan died in 1818, the tsarist authorities did not recognize the new khan. At the beginning of the 1820s, the tsarist authorities introduced new administrative structures. In 1822 Spiransky’s “Rules for the Siberian Kyrgyz” were applied to the Middle Horde administered from Omsk, and in 1824 the “Rules for the Orenburg Kyrgyz” were applied to the remains of the Small Horde. Both statutes abolished the status of khan and hordes as political unities. The 1822 statutes divided the Horde territory into 4 okrug “county” districts, and 87 volosti subdistricts, governed by hereditary sultans, who had to be confirmed in office by the Russian governor-general of Omsk.^[74]
Levchine’s account has been read for evidence of a lack of “true” feudalism, usually conceived of in rather narrow Marxist terms, in which the appropriation of surplus value through land rent was regarded as a defining feature of feudal relations. Some of Levchine’s remarks were taken as evidence of the lack of regular tributary relations between vassals and lords.^[75]
Taxation, like full control of their subjects, had clearly become difficult for the Kazakh nobility faced with incorporation into the tsarist state. But Levchine makes clear that regular taxation had existed in the precolonial period “by the most ancient law of the Kirghiz.” Later in the text he describes the annual tax on all those of military age in the laws attributed to Tauke Khan. “All men able to bear arms must pay each year to the Khan and chefs of the nation a tribute of one twentieth of their goods.”^[76] So contradictory was this to the tribal model that the passage was somehow twisted to mean the exact opposite. Martha Brill Olcott states that evidence that the Kazakh khans did not have feudal powers and regular taxes can be found in Tauke Khan’s law code, which “records the levying of an extraordinary tax of one-twentieth of an individual’s wealth to pay for the provisioning of men to bear arms.”^[77]
Rather than being the basis of a bottom-up process of political self-organization, the tracing of distant descent seems to have been an elite practice. Khazanov makes the point that only the aristocracy kept genealogies, noting “Grodekov^[78] wrote of the Kazakhs and Kyrgyz that, in contrast to aristocracy, ‘… the poor do not know, apart from the names of their direct ancestors and their clan [rod] and tribe [koleno], anything about the distant branches of kinship.’ Levchine (1832, III: 11) also made this point some fifty years earlier.”^[79] As usual these discrepancies are explained away with the notion that the “tribal structure” was in decline at the time it was observed.^[80] Geiss writes “most of the reports were done during tsarist rule, when Russia has already introduced its administrative order.… Consequently, tribal affiliations and genealogies lost their political significance.”^[81]
It should be added that knowing one’s “clan” and “tribe” meant knowing the administrative units one belonged to.^[82] Although the tsarist empire governed its colonial subjects using terms such as rod and uprava – “administrative clans,” in reality, these did not require any particular kinship relations between members and “state officials transferred indigenous families from one clan to another by issuing resigning and reinstatement certificates.”^[83] However, Speransky’s 1822 statute enshrined the position of the indigenous elite using the paternalistic language of kinship. The statute declares that a native knyaz’ (prince) is “the governor of a clan” and “is accepted as an elder as if this clan is a single family” [emphasis added].^[84] As Ssorin-Chaikov argues, in the “customary laws” enshrined in the 1822 Speransky Statute, “a historically informed ethnographer could discern not an organic society but the legal code of a defeated empire.”^[85] This is exactly what Levchine’s description of the Kyrgyz-Kazakh suggests; the administrative code of the decapitated Kyrgyz-Kazakh state treated as the “customary law” of a traditional tribal people by the colonial state.

TRIBE AND NATION

The plausibility of the tribal model of kinship society reflected its place in the wider ideology of the nation-state. National populist thought pictured tribal and ethnic entities in ways that were analogous to the ideological construction of the nation. As in popular imaginings of the Völkerwanderung (Wandering of Peoples) period of Dark Age Europe, tribes and peoples were conceived of as protonational units, sharing common cultural and social forms.
As Martin Thom points out, the concept of the nation that emerged in the nineteenth century was powerfully influenced by debates surrounding the representation of Germanic tribes that invaded the western Roman empire from the fifth and sixth centuries. In his famous 1882 lecture What Is a Nation? Renan declared that “It was in fact the Germanic invasions which introduced into the world the principle which, later, was to serve as a basis for the existence of nationalities.”^[86] In Renan’s time, the conservative position of those such as the Comte de Montlosier, the eighteenth-century French publicist, was to identify the nation with the project of rulership, in particular, with the noble houses that had emerged from the Germanic invasions. Inspired by the populist vision of the nation that emerged with the French Revolution, Romantic historians such as Thierry rejected this approach. “For Montlosier, as for other ultra-royalists, the French ‘nation’ had consisted of the free, warrior nobilities that had invaded Roman Gaul, for Thierry it was the original inhabitants, the Gallo-Romans.”^[87]
Renan’s seminal theory of nationhood was part of a reaction by secular intellectuals, determined to reject the claims of the monarchists and promote the populist vision of the nation, constituted by its citizenry, not its rulers. That tradition lived on in the work of Fustel de Coulanges and his pupil Emile Durkheim, the “father of sociology.” The notion of both nation and society that developed in nineteenth-century social sciences reflected the victory of this populist politics over the “ancien régime.”^[88] Durkheimian thought bares the imprint of Renan’s nationalism, particularly his concern with solidarity and the collective conscience. One can almost see Durkheim’s work as elaborating Renan’s conception of the nation for use as a general theory of human aggregation.^[89] Like the nation, “society” stood for “the people” or volk as a whole, with its own generalized culture, traditions, and form. In this populist imagination, human aggregates were not simply the subjects of a ruler, but must form social and cultural collectivities, and this was reflected in the emergence of various forms of folk studies. Sociology and particularly ethnology,^[90] took as their objects of study the cultural and the social as mass or at least collective phenomena, conceived of in the populist mode very different from the royalist historians of an earlier era.
But, as Hobsbawm shows, in the age of populist national politics the notion of shared kinship was an important element in the new ideologies of mass mobilization.^[91] The idiom of familial and fraternal relations, projected onto the “family writ large” of the nation, became a dominant theme. Those engaged in the intellectual and political project of nation-construction made claims of national unity based on the idea of common descent. In the socially heterogeneous and divided region that was to become Albania, for example, Albanian nationalists such as Naim Frasheri (1846–1900), claimed “All of us are only a single tribe, a single family; we are of one blood and one language.”^[92] In this historical imagination tribes were the protonational groups, the natural units or subunits of a given volk (people). This generated a sort of “billiard ball” model of history, in which prenational, tribal peoples moved as discrete integral units across Eurasia to collide with and displace each other.^[93]
This vision of history has survived the numerous studies showing that “tribes” such as the Franks or Saxons were political projects including a wide and usually heterogeneous assembly of large and small noble houses and their subjects.^[94] This has been shown so frequently that Norman Davies, a popular historian of Europe, for example, notes: “Chroniclers and historians were tempted to write in terms of discrete, permanent, and self-conscious tribes where no such entities had necessarily existed.… All have suffered, too, from the attentions of nationalist historians in our own day, who think nothing of projecting modern identities into prehistory.”^[95] But lacking a replacement for the old model, Davies feels bound to reproduce it. He writes “In the absence of alternatives, it is difficult to know how one can describe the migrations except in terms of the traditional tribes.”^[96] The key problem in describing “peoples” is, as Wolfram put it “we have no way of devising a terminology that is not derived from the concept of nation created during the French Revolution.”^[97]
When these Germanic peoples migrated into or invaded parts of the Roman Empire, the historical record shows the process to have been very different from the vision of discrete population units displacing each other. The result of the Germanic invasions on Europe was not the replacement, en masse, of one human population with another, but the insertion of new aristocracies and their military followings into the existing order of the Roman provinces.
The Germanic “tribes” were not discrete bodies of kinsmen tracing common descent. They were the political entourages, and frequently the conquest projects, of noble families.^[98] They displaced, or intermarried with, Roman elites, and the membership of their political formations was recruited from all sorts of sources, often from Roman subjects who preferred the new masters to their old ones. As Herwig Wolfram notes, “From the first appearance of the Gothic hordes on Roman soil, they attracted people from the native lower classes.… Roman lower classes had been willing… since the third century: ‘to become Goths.’”^[99]
The early Germanic societies were a product of power relations between subjects and ruling families. They had a political, rather than a cultural or “ethnic” definition.^[100] Similarly, sociological terms developed in the age of popular nationalism are a poor guide to understanding what Anderson calls the “dynastic realm.”

“These days it is perhaps difficult to put oneself emphatically into a world in which the dynastic realm appeared for most men as the only imaginable “political” system. For in fundamental ways “serious” monarchy lies transverse to all modern conceptions of political life. Kingship organises everything around a high centre. Its legitimacy derives from divinity, not from population, who, after all, are subjects, not citizens. In the modern conception, state sovereignty is fully, flatly, and evenly operated over each square centimetre of a legally demarcated territory. But in the older imagining, where states were defined by centres, borders were porous and indistinct, and sovereignties faded imperceptibly into one another. Hence, paradoxically enough, the ease with which pre-modern empires and kingdoms were able to sustain their rule over immensely heterogenous, and often not even contiguous, populations for long periods of time.”^[101]

Societies are, as Leach notes, political units, in practice, and they are more clearly a product of common rulers than common cultures.^[102] As Ernest Gellner points out, the notion that people are bound to live in units defined by shared culture is a relatively recent one. “Culturally plural societies often worked well in the past: so well, in fact, that cultural plurality was sometimes invented where it was previously lacking.”^[103] But in anthropology the political construction of the social has often been backgrounded by the powerful heritage of classical ethnology, which assumed some more or less homogeneous cultural and social entity as its object of study (the ethnos) and tended to regard “peoples” as cultural wholes. This is reflected in theories of ethnogenesis developed by writers such as Lev Gumilev^[104] and Yulian Bromley^[105] that continue to be highly influential in post-Soviet Eurasia.

CONCLUSION: NATIONAL POPULIST HISTORY AND THE DISCOURSE OF TRIBE IN INNER ASIA

Models of tribal, nomadic, and kin-organized society were applied to the indigenous societies of Inner Asia, and led to the misinterpretation of historical and ethnographic materials to support the vision of tribal societies organized by principles of kinship. The subjects of “nomadic” polities were generally assumed to be tribes or clans. The genealogies of noble houses were taken as evidence of the general organization of society on kinship lines, rather than elite techniques linked to the project of rulership.
Rather than one of the “earliest acts of human intelligence” as Morgan supposed, I have argued that the organization of people into named unilineal descent groups with political functions was an act of state administration in much of Inner Asia.^[106] Comprehensive kinship organization, where it appeared, seems to have been a product of the state, not a precursor to it, even in the Greek and Roman world that had inspired Morgan. In the medieval Mongolian case the units described as clans (obog) did not resemble the model of pre-state kinship society at all, but seem to have represented the projects of rulership of the aristocracy. It was far later, in the twentieth century, that the elite tradition of aristocratic Borjigin descent from a common ancestor was extended to all Mongols as part of the ideological project of nationalism.^[107] In the tsarist Russian state the rod “administrative clan” was also used as a unit for the government of Inner Asian subjects. However, this should not be seen simply as a case of “secondary tribalism” in which the tribe becomes a product of the colonial state. These units did not operate in the way that the theories of kinship society supposed, but they provided an idiom of administration that could be used by later scholarship as evidence of autochthonous kin organization.
In the age of national populism, distinct “peoples” were imagined as constituting themselves and could therefore give rise to national leaderships that represented each volk and their interests. So it seemed natural that in pre-state society people should also form popular political collectivities – groups of kin with leaders that represented them. But until the age of national populism the putative content of the ethnonational notion of a people was not necessarily important for processes of political identification.
The extent to which something resembling a nation existed before the advent of the recognizable “modern” form differs in the work of primordealist scholars^[108] and constructivist ones.^[109] Gellner and Hobsbawm, for example, sought to counter dominant trends in history that projected modern national concepts onto past epochs, and tried to show how nationalist ideology creates nations where they had not previously existed.^[110] But the idea that nations are rooted in “ethnic” units of common descent has remained influential. Although Hobsbawm points out that nations are not generally formed by preexisting ethnic groups, for example, he still considers the ethnic group to be a form of proto-nation.^[111] Anthony Smith argues for the “vital role of… ethnies, in providing the basis for the emergence of nations.”^[112]
Unsurprisingly perhaps, theories of the ethnos as a protonational entity was elaborated most in Soviet scholarship, firmly rooted as it was in Marxist evolutionism. Ethnic studies in the USSR were based on the notion that ethnic groups were stable entities that transmitted their social structures from generation to generation.^[113] In the post-Soviet era, as Valery Tishkov noted, ethnos and ethnogenesis remained powerful and almost sacred themes in anthropological and public discourse.^[114] The theories of Lev Gumilev (1912–1992) became extraordinarily influential in late Soviet and post-Soviet Eurasia. Gumilev argued that the ethnic unit – ethnos – “naturally developed on the basis of an original stereotype of behaviour of a collective of people, existing as a power system (structure), opposing itself to all other similar collectives, proceeding from a sensation of complimentarity (komplimentarnost’).”^[115]
Gumilev’s work was highly appealing to the emergent leadership of the newly independent states of post-Soviet Central Asia, perhaps because it seemed to celebrate successful political figures as catalysts of social and historical development. Politicians such as President Akaev of Kyrgyzstan, made explicit use of Gumilev’s theories in their own nation-building ideologies.^[116] In Kazakhstan the new state named one of the country’s universities after him, and official histories reproduce his theory of the distinctive ethnos.^[117] In many ways, Gumilev and other theorists of ethnogenesis such as Bromley can be seen to have elaborated the well-established themes of national-populist social science.^[118] Having pictured populations in terms of distinctive volk (Fichte) or large-scale solidarities (Renan, Durkheim), scholars could elaborate theories as to how these complex mass phenomena came about and produced the putative unity and national or protonational leaderships attributed to them.
Entangled as it clearly is with nineteenth-century theories of common descent as the universal basis of political units, the concept of the ethnos seems of limited utility when it comes to rethinking these models. Now that modernism has come to be seen as a historically located ideology, we are bound to treat notions of the premodern, be that “traditional,” “ethnic,” or “tribal” society, as part of modernism’s own creation myth, an element of the movement’s self-description. The conceptions of tribe and ethnos that were inherited from the nineteenth century continue to be bound up with the project of rulership, in this case the nation-building efforts of the newly independent states of Central Asia. But it presents students of the region with the challenge of discerning evidence from supposition, escaping the articles of faith of the contemporary political order, and doing justice to the realities of a bygone age.