This is a polemical and thought-provoking work that seeks to recalibrate what the author presents as misguided assumptions regarding the internal structure and dynamics of Inner Asian society. Based on his fieldwork in Mongolia, David Sneath concludes that “nothing like the popular image of kinship had existed in Mongolia” (p. 1). Moreover, the conceptualizations of “tribe” and “clan” that prevailed in the nineteenth- and twentieth-century scholarship, with its emphasis on egalitarian “segmentary kinship” groups (p. 156), were premised on “colonial-era notions of tribalism.” Weighted with the “pejorative, colonial baggage” associated with the word “tribe” (p. 52), they produced a skewed vision of the history of the peoples of the steppe, requiring outside catalysts (neighboring sedentary states) to bring kinship-based “timeless … nomadic, tribal society” (p. 3) to statehood. Kinship, clan, and tribe (disputed conceptual categories in modern anthropological literature and not without problems for historians; see p. 64) are set aside. Rather, Sneath suggests, it was “aristocratic power and statelike processes of administration” that proved to be “the more significant features of the wider organization of life on the steppe” (p. 1). These “aristocratic orders,” through “a configuration of statelike power formed by the horizontal relation between power holders” (p. 2), produced a “headless state,” that is, a polity without a central ruler, capital city, and so on (chapter 7), but functioning as a state. Sneath has restored the role of the aristocracy, so often, in his view, swept aside in an imagined unstratified, egalitarian nomadic society (pp. 73–74).

There are problems with Sneath's sweeping revision of Inner Asian history. In the process, he has to rewrite the dictionaries of the languages of Inner Asia and of the sedentary societies whose sources provide much of our information on the nomads, medieval and modern (cf. his discussion of Chin. xing, “surname,” “clan,” “people,” pp. 116–17). Among the transgressors is Igor de Rachewiltz's recent translation of the Secret History of the Mongols (Leiden: Brill, 2004), which, Sneath asserts, inserts or assumes “clan” where “principality” could have been used (pp. 106–7).

Similarly, Wheeler Thackston's 1998 translation of Rashîd ad-Dîn (d. 1318), whose world history, Jâmi’ at-Tawarîkh (Compendium of Histories) contains a lengthy taxonomy of the Turko-Mongolian steppe dwellers, misleadingly “supports a tribal model,” Sneath contends, as the author only uses “a single term, qawm, for all groups of people, which can mean a group of any size and contains no inherent kinship content” (p. 107). However, qawm (plural aqwâm) in Persian means “people,” “nation,” “tribe,” “family,” or “kindred” (see F. Steingass, A Comprehensive Persian-English Dictionary [London: Routledge and Kegan Paul, 1892; repr., Beirut: Librarie du Liban], 995) and, pace Sneath, implies kinship. It is a loanword from Arabic, in which it has an identical range of meanings (“people,” “tribe,” “kinsmen”—see Vl. V. Polosin, Slovar poétov plemeni abs: VI–VII vv. [Moscow, 1995], 408; peuplade, tribu, peuple—see A. de Biberstein Kazimirski, Dictionnaire Arabe-Français [Paris: Maisonneuve, 1860], 840). Rashîd ad-Dîn was exceptionally well informed about such matters through Bolad Aqa, a Mongol who had come from the Yuan court to serve in Iran, and his presentation of the Turko-Mongolian world, reflecting indigenous traditions, has to be taken seriously.

Moreover, the Persian text has a more extensive terminology than Sneath claims. Rashîd ad-Dîn reports that the aqwâm-i atrâk (“nations” or “peoples of the Turks”) were divided into qabâ'il (“tribes,” singular qabîla, also meaning “family,” “kindred,” “parentage”). Over time these aqwâm split into “branches” or “tribal subdivisions” (shu'ba), each with its own laqab (“agnomen,” “surname,” “family name”—see Rashîd ad-Dîn, Jâmi' at-Tawârîkh, ed. M. Rawshan and M. Mûsawî [Tehran, 1373/1994], 1:39–41). Thackston, one of the best (and most prolific) translators of medieval and early modern Persian sources, did not read in notions of kinship and tribe; these are in the texts.

The terminology used by the pre-Chinggisid nomads—our examples largely come from Turkic speakers—was kinship based. Maḥmûd al-Kâshgharî, in his Dîwân Lughât al-Turk (Compendium of the Turkic Dialects, c. 1077, trans. R. Dankoff and J. Kelly [Cambridge, Mass., 1983–85]), reports that when “two men who do not know each other” meet, they say, boy kimdir? or “Who are your kinsmen, your clan, your people?” (2:219). Kâshgharî (whose work Sneath ignores), a scion of the Qarakhanid ruling house, had intimate knowledge of the Turkic steppe world. He uses four Arabic terms to convey the sense of boy (Old Turk. bod “clan,” see Gerard Clauson, An Etymological Dictionary of Pre-Thirteenth Century Turkish [Oxford: Oxford University Press, 1972], 296–97, perhaps originally denoting “body,” see also Sergei Starostin, Anna Dybo, and Oleg Murdak, Etymological Dictionary of the Altaic Languages [Leiden: Brill, 2003], 1:365): qabîla and qawm, already explained, and rahtṭ (“a man's people and tribe,” see E. W. Lane, An Arabic-English Lexicon [Edinburgh, 1863–93], 1169), and ‘ashira (“a man's kinfolk or his nearer or nearest relations or next of kin … a small subtribe,” see Lane 1863–93, 2053, “clan” in Modern Arabic). Kâshgharî was no colonialist ethnographer. For him the Turkic groupings (tṭabaqât, “categories, generations”), which he knew as an “insider,” were divided into qabâ’il (“tribes”) based on kinship. Sneath says that the Dede Qorqut tales make “no mention of descent groups at all” (p. 150), but a look at the Turkic text reveals boy, qabile, oba (Clauson 1972, 5–6, early on a term for “clan” or perhaps an even smaller social unit, but defined by Kâshgharî as qabîla) and the well-known Oghuz tribes (the Qayï, Bayat, Döger, also mentioned by Kâshgharî). Sneath is certainly correct in his critique of the simplistic construction of kinship-based clan > tribe > state, but in his concern to expunge kinship and tribes, he creates an alternate universe of “aristocratic orders” as the source of political identity.

Much of the modern scholarship on the Eurasian nomads long ago recognized that “tribe” and “clan” were complex phenomena, involving the political integration, often unstable, of heterogeneous elements. (Cf. L. P. Lashuk and Anatoly Khazanov, among many others—the essential ethnographic studies by N. A. Aristov, Usuni i kyrgyzy ili kara-kyrgyzy [Bishkek: Ilim, 2001] and Trudy po istorii i etnicheskomu sostavu tiurkskikh plemen [Bishkek: Ilim, 2003]; R. G. Kuzeev, Proiskhozhdenie bashkirskogo naroda [Moscow, Nauka, 1974]; V. V. Vostrov and M. S. Mukannov, Rodoplemennoi sostav i rasselenie kazakhov [Alma-Ata: Nauka, 1968]; Faruk Sümer, Oğuzlar [Istanbul: Ana, 1980]; and others are not noted in Sneath's bibliography), and Sneath is aware of this (p. 108, cf. his discussion of the Mongol obugh/obog). Sneath's contention that the Russians and Iranians used oba as an “administrative and tax unit” for the Türkmen (p. 146) ignores its long earlier history in Turkic, where it is first clearly associated with kinship-based groupings. As with Mongol aimag (“tribe, clan” and “administrative unit,” see Ferdinand Lessing et al., Mongolian-English Dictionary, 3rd ed. [Bloomington: Mongolia Society, 1995], 21), Sneath projects the latter meaning alone into the early history of the steppe (p. 67). Are we to posit “administrative units” as developing first (implying the existence of a state) and then their transformation into tribes or aristocracy-led named groups? This seems more than unlikely. When examining the medieval sources, it is hard to avoid, as Sneath would have us do, kinship, fictive, politically expedient, or based in some distant or not so distant reality, clan and tribe as the vocabulary of political affiliations among the Eurasian nomads. Consequently, at some earlier point, they must have been the building blocks on which later, expanded and no longer stricto senso kinship-based political structures were constructed.

Another fundamental problem is that Sneath nowhere tells us how his “aristocratic orders” arose in the pre-Chinggisid steppe and what was the source of their local power. Indeed, in his concluding remarks, he comments that the “concept of aristocracy as an analytical and comparative term deserves to be developed more fully than I have been able to in this work.” If we are to accept his proposed tribeless order and “headless state,” in which nameless servitors, slaves, and other dependents cluster around power-wielding “aristocratic orders,” we need to know much more about the latter and the mechanisms by which they were able to attract and retain followers.

Finally, the work contains some striking errors, misstatements, or understatements of historical fact. The fall of the First Türk Qaghanate to the Tang (630 in the East and 657–59 in the West) did not set off the migrations of the Pechenegs and Oghuz (p. 25). We do not know when precisely the Pechenegs came to region around the Syr Darya. We do know that they were present when the Oghuz first came there in the 770s and drove them into the Volga-Ural Mesopotamia and thence into the Pontic steppes in the ninth century. The movement of significant Turkic groupings from the Eastern Türk state (e.g., the Qarluqs in 745) began after the fall of the Ashina Türks in 742 and the brief Basmïl Qaghanate, which was quickly toppled by the Uighurs (744). The Khazars did not “apparently” convert to Judaism. They did convert, as Khazar and non-Khazar Hebrew sources, Muslim Arabic and Persian accounts, and Western Christian Latin sources attest. The Khazars did not fall “victim” to the Rus' (assisted by the Oghuz) in the “late ninth century” but in the late 960s. The Seljuks did not “establish a power base in Persia in the tenth century” but in the mid-eleventh century. Sneath refers to the Uighurs as the Türks (p. 117), a confusion that neither would have appreciated. And Ira Lapidus, the well-known historian of the Middle East, is not a “she,” but a “he” (p. 50).

Acephalous polities among the nomads (cf. the Pechenegs, pre-Seljukid Oghuz, and Cuman-Qïpchaqs) are hardly unknown. Long ago, W. Barthold suggested that khans usurped power by force and that statelessness was the “natural” condition for the nomads, who did not need a strong central authority for the effective functioning of their stock-herding economy. Were these acephalous polities states? This requires a definition of the state, and Sneath dodges this, burying this key question with the comment in a note that “there is not space to properly review” (p. 210 n. 9) the issue. Indeed, one could argue that Sneath's “headless state” is a retro-projection to the whole of steppe history of the 1640 compact (the Mongol-Oirat Code) made between the fractious later Chinggisids of Mongolia and the Oirats, which produced a kind of central khan-less pan-Mongol legal order—for a time.

While I think that Sneath's thesis may be feasible for the late Chinggisid Mongol world, with its contested, fragmented loci of power and long history of khan-led states, I do not find it a useful tool for assessing the pre-Chinggisid steppe polities. Nonetheless, this is a rich book, and a brief review cannot do justice to its many reinterpretations of steppe society or to the rejoinders that such viewpoints will most certainly engender.