This book is an invitation to reconsider the whole history of “nomadic Inner Asia”. The question is: how can we understand the emergence of these famous “empires of the steppes” which appear to plunge into the void so suddenly when they fell? Sneath's answer can be put that way: rather than assuming an opposition between two political forms (i.e. state vs. tribe), it is more productive to think of a continuum of “ruling houses” (also called “nobility” or “aristocratic orders”) developing more or less centralized forms of political organization. The best example of a strongly centralized form is the state established by Genghis Khan on the eve of the thirteenth century in Mongolia. The other extreme (a non-centralized form of state power) could be found in seventeenth to early twentieth-century Mongolia – a period of which the author, who is Director of the Inner Mongolia and Asia Studies Unit at Cambridge University, is a specialist. In 1640, Sneath explains, the Mongol and Oirat nobility put up a State (törö), with its laws and judges, serving a social system inherently unequal, but without capital and sovereign – in other words, a “headless state”. In Sneath's view, the engine of Central Asian history is no longer the environment (ecological constraints would prevent social stratification inside nomadic society and/or explain their moving) or the external influences (the political organization of the nomadic society would be closely linked to that of their sedentary neighbours), but instead the political strategy of these basic units that were the noble houses. In the end, Sneath highlights an bottom-up process: the (nomadic) society is produced by the centralized or “headless” states (for example through the creation of genealogies), and not the reverse.

To conduct this demonstration, Sneath analyzes in great details the links between history and anthropology. Indeed, since the nomadic societies have left almost no written trace, historians have been inclined to fill in with the help of anthropological studies. The problem, says Sneath, is that most of the anthropological scholarship was wrong and/or outdated. The enemy, because there is one, is the “colonialist” misconception that has obscured the aristocratic reality of Inner Asia, namely the notion of an egalitarian society, according to which the “chiefs” would have been primus inter pares, and the group members would have been united by a solidarity based on a common origin. Armed with Weber, Foucault and Lévi-Strauss, Sneath conducts a critical review of a large body of literature: anthropology (criticism of the kinship society theory), sociology (criticism of the structural functionalism and political evolutionism) and history (criticism of the essentialist view of the nomads). Drawing on numerous examples, from the Scythians to the Bakhtiaris, from the Türkmens to the Mongols, Sneath reassesses most of the studies which have been produced on the nomadic societies of Asia and proposes his own understanding.

This is a stimulating essay, whose ambitious project will be of great interest to scholars from diverse backgrounds. The Headless State is an invitation to think about the epistemological links between history and the other branches of social sciences, and the way they feed each other. While historians are invited to be careful not to use such outdated “anthropological” concepts (for example the concept of tribe), anthropologists are encouraged on their side to re-historicize the facts. Sneath sets an example by rereading Irons's field work and conclusion on the Yomut Türkmens from a perspective of what we know of the history of the Oghuz (pp. 142–56). He concludes that Yomut were not “a nation”, but a “ruling elite”. It is an interesting application of the constructivist theories developed by Hobsbawm and Gellner for European history to Central Asia. Beyond the questions of method, this book captures the attention by its thesis (the preeminence of the noble houses) and the effort, praiseworthy in these days of over-specialization, to propose a new model of analysis for an area which played such a pivotal role in world history. For the author of this review, a specialist of a sedentary country – Iran – when it was ruled by Turks of nomadic origin – the Saljuqs, this type of questioning is naturally welcome.

The question is to know whether the demonstration holds. The book poses a number of problems, which will prove prohibitive to many readers. The first reservation is about the manner in which Sneath presents his material. There is a Da Vinci Code side to this book: the signs (of the aristocratic reality) were everywhere, but they were masked by more or less dishonest scholars at the service of the “dominant view” (p. 59, or “dominant vision” p. 71, “received wisdom” p. 53, “typical justification” p. 49). Even “the Scythians were not allowed to challenge the notion of the essential nomad” (p. 124)! This demonstration is done at all costs, and sometimes with amazing shortcuts. To reassess the concept of nation (and hence of solidarity among the nomadic society), Sneath convenes in the same passage the French reactionary historian Montlosier and the Marxist Hobsbawm (p. 159). Besides, Sneath has a definitive way to contrast the elected scholars (those who go in the direction of his argument) from the misguided. Throughout the book he awards good and the bad points in an extreme binary way. Hurrah for Vladimirstov, who highlighted the “Mongolian nomadic feudalism” as early as 1934; down with Tolybekov, who thereafter tried to apply the model of “kinship society” to Central Asian societies, and who later brought in his wake other scholars (like Geiss).Footnote ¹ Hurrah for the travelers and explorers of the nineteenth century, whose perception was not veiled by the (false) kinship theory; down with the functionalist anthropologists of the twentieth century. Hurrah for Digard who insisted on class divisions among Bakhtiaris of Zagros; down with Garthwaite who stressed the importance of lineage divisions.Footnote ² Hurrah for Dresch who has shown that segmental genealogy is the product of a political project and down with the (numerous) others who took for granted these genealogies.Footnote ³ Not surprisingly the palm of “infamy” in this binary presentation belongs to the scholars who share with Sneath his field of specialization: Mongolia. KraderFootnote ⁴ and BarfieldFootnote ⁵ hold here the role of arch-villains, and there is no worse criticism than to say, for example, that “Barfield offers a description … that could have been written by Krader and Sahlins” (p. 60). Sneath seems to have a vital need to attack others, and when no researcher can serve as a victim, the journalists will do, as at the beginning of Chapter 4 (pp. 93–94). In many places the tone of some comments, the constant imprecations (speaking of Krader, he regrets “another of his entirely misleading flights of fancy”, p. 70) seems to me not only out of place, but detrimental to the project itself.

A second reservation is about the content. The detailed description of the anthropological basis which had allegedly “corrupted” the understanding of Inner Asia (esp. in Chap. 2, 3 and 5) goes back as far as Morgan's Ancient Society (1877). This is interesting but not really new. The same is true of the criticisms of the segmentary model (its limitations have been highlighted for a long time), and his use of the “house society”. Theorized in 1979 by Lévi-Strauss from the study of a Native-American tribe, a “house society” (société à maison) combines apparently contradictory principles (agnatic/uterine, kinship/alliance, etc.), and as such it enables the anthropologist to analyse “a field of power relations expressed in the idiom of descent” (p. 112). The concept provides Sneath's “headless state” with the theoretical basis he needed, and he uses it as a magical formula to understand the history of the Xiongnu, the Liao and the Mongols (pp. 112–19). However, in a major article published as early as 1987 in L'Homme, Digard had already referred to Lévi-Strauss's “société à maison” to explain the apparent contradiction he had noted within the society of the nomads of Bakhtyari.Footnote ⁶ The fact that Sneath does not even mention this article on a topic so close to his own concern, but instead refers to a collected volume in English on Lévi-Strauss is symptomatic of either the lack of visibility of the French scholarship or, worse, Sneath's limitations in the issue he deals with (as far as the anthropological approach is concerned, readers would be well advised to read the production of the research group Production Pastorale et Société which was active in France between 1976 and 1987).

From the point of the view of the historian, Sneath's stance presents many weaknesses also. His “aristocratic” interpretation of the terms noyan or obogs in the Secret History of the Mongol is interesting, as is his dismissing of a “tribal” interpretation. But given the state of our documentation, it seems that it can be at best only an hypothesis and not a certitude. Interestingly Peter Golden, whose work on pre-Chinggisid Inner Asia is often cited by Sneath in reinforcement of his theory, has been particularly critical of the way he uses historical sources on Mongol history (see Golden's review in the Journal of Asian Studies, 68:1, pp. 293–96, and the subsequent rejoinders in ibid. 69:2, pp. 658–63). The same is true of other texts mobilized to support the thesis of “aristocratic orders”. For example, while dealing with Türkmen society, Sneath refers extensively to the nineteenth-century French version of Levshin's account (first published in Russian in 1832) of de Levchine's account, but without addressing the two major issues of any historical reading: the issues of sincerity (the possibility of deliberate lies) and of accuracy (the possibility of error). Such casualness is problematic when one is attempting to engage in a re-historicization of the whole history of Inner Asia. Beyond these problems of interpretation, many factual errors are equally irksome. The Saljuq “established a power base in Persia” a century later than Sneath thinks (p. 25). Mistaking the identity of Ira Lapidus (p. 51), who might be legitimately considered as the greatest scholar of medieval Muslim societies, is not reassuring as to Sneath's level of historical knowledge. This is the kind of mistake that undermines the confidence of the reader in the rest of the author's case. And his wish not to use what he probably considers as suspect historical works results in surprising holes in the references (for example on the Aq-qoyunlu, Sneath quotes Minorsky but not WoodsFootnote ⁷).

Another defect is, I believe, the very structure of the book. Sneath starts with the general (a rereading of all the literature on nomadic societies) and finishes with the particular (seventeenth- to nineteenth-century Mongolia). It is a risky approach since he rejects at the end of the volume what he knows best and where he is least attackable. Before getting to that period, the reader is invited to consider elements of proof, on the solidity of which the author takes care not to be too assertive (for example about the main protagonists of the national epic of the Kyrgyzs, Sneath describes them as nobles who led groups of retainers “who do not seem to be their kin”, p. 89, emphasis added). No doubt he would have made his point more convincing by analysing first the case of Mongolia during the Qing period, and offering only subsequently a rereading of the history of Inner Asia in the light of his starting hypothesis.

Lastly, we can deplore the constant and tiring cross-references (for example p. 71) and equally incessant repetitions (the inje, or ‘human dowry’, is presented in the introduction, described on page 110 and again on p. 175; same for the refutation of geographical determinism, p. 16, p. 53 and pp. 135–36). Many references are not included in the references section (for example Khazanov 1984 quoted on p. 73 and p. 176; Khazanov 1990 quoted on p. 176; Hobsbawm 1990 quoted on p. 159; Fried 1967 quoted on p. 7; Irons 1975 quoted on p. 142). And there are also an unusual number of misprints (for example “Qara-Kdhitan” on p. 31, “Syr Dariya” on p. 85, “significan” on p. 109, “qahgans” on p. 169, “D. Senor”, on p. 255). All of this reflects precipitation and a lack of care.

In conclusion, this is a book that I have found deeply challenging (because of its argument and its audacity), but also deeply irritating (because of the way he has made a mess of it). The best thing is probably to take this book for what it is: a polemical pamphlet. It has all the features: an interesting thesis but few hard facts and much over-simplification. As for the hard words and the attacks, they are part of the project. With regard to Deleuze and Guattari's “nomadologie”, Sneath writes: “If the study of mobile pastoral society is to usefully inform wider social theory, it must do so on a firmer basis” (p. 202). Indeed.