Peter Golden, in his review of The Headless State, kindly describes it as thought-provoking and rich, but attacks its central argument—that dominant notions of kinship society led historians of pastoral nomadic societies to assume they were composed of segmentary clans and tribes defined by descent from common ancestors. As this evolutionary model of political society has been largely rejected within anthropology, the discipline that developed it, I suggest a reinterpretation of groups termed “clans” and “tribes” in translations of texts describing nomadic societies such as the medieval Mongols. For Golden, this amounts to rewriting “the dictionaries of the languages of Inner Asia” (Reference Golden2009, 294). I would like to make three points.
First, the ambiguities surrounding each term allow for alternative interpretations without rewriting any dictionaries. In the cases discussed, “tribe” or “clan” are only two of many possible English terms given. Golden argues that Rashid ad Din's Jami’ at-Tawarikh describes Turko-Mongol peoples (aqwam) as “divided into qaba'il (‘tribes,’ …). Over time these aqwâm split into “branches” or “tribal subdivisions” (shu'ba) … [Wheeler] 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” (2009, 294). However, Thackston wrote, “Had he so chosen, Rashiduddin [Rashid ad Din] could have availed himself of Arabic vocabulary and been very precise in the definition of groups of various sizes. Instead he opted to use one word, qawm, for all groupings. He does not, therefore, distinguish among nations, tribes, or clans, although he occasionally uses the word shu'ba ‘branch’ for a branch of a family tree … The reader should bear in mind that Rashiduddin does not distinguish these groupings in any way” (1998, xvi–xvii). Thackston, then, translates shu'ba as “branch,” not as “tribal subdivision,” as Golden implies. As usual, the “tribal” is an interpretive addition.
Francis J. Steingass's dictionary, which Golden cites, gives many terms for qawm, including “people, nation” (not kinship terms), “a sect, some, a certain person, a company of men … The followers of King Abrahah.” Earlier translators, such as Étienne Quatremère (Reference Quatremère1836, 97) used “people” or “nation” for qawm where Thackston uses “tribe” (1998, 93). For qabila, which is not mentioned by Thackston in this regard, Steingass gives “class, kind; species, category” (Reference Steingass1892, 954) as well as “house”—the alternative to “clan” that I suggest for descent groups such as the Mongol obog. Al-Jabarti used qabila to describe political factions of Mamluks (Ayalon Reference Ayalon1960, 310), and Henry Munson (Reference Munson1989, 388) showed that in Morocco, qabila was a vestige of old administrative divisions, not of common descent. As Richard Tapper remarks, “Middle Eastern indigenous categories (of which perhaps the commonest to have been translated as tribe are qabil, ta'ifa, quam [sic], and il) are no more specific than are English terms such as ‘family’ or ‘group’” (Reference Tapper, Khoury and Kostiner1990, 56). Dictionaries are of limited help, then, in knowing what the social groups referred to were actually like. I have tried to rethink the composition of medieval Mongol society—where terms for aristocratic descent groups could stand for the larger entities they ruled. When it comes to the early Türks, this is not a job for someone like me, but for specialists with open and enquiring minds. However, even an inexpert examination of the terms suggests the old tribal model rests on distinctly sandy foundations. Gerald Clauson's dictionary gives a wide range of possible terms for bod/bodun, (clan/tribe), many with no tribal or kin connotation. In almost all the examples he gives, Clauson translates bodun as “people,” not as “tribe” or “clans,” for instance, “subject people,” “the people of the country,” “77,000 people” (Reference Clauson1972, 297).
Second, despite the impression that Golden gives, partly by misquoting me (2009, 293), I do not deny the significance of kinship for steppe society; however, I suggest that it was largely aristocratic kinship that was politically important (the genealogy of the ruling house, not its subject people). Mongol groups described as qawm did not resemble the Morganian model of clan and tribe. The insider's account, The Secret History of the Mongols, and the eyewitness accounts of Rubruck and Carpini indicate that they were divided into commoners and nobility, included large numbers of persons who were not kin of the rulers, and were not fully exogamous.
Third, far from “dodging” the question of what “the state” is, as Golden claims, I explicitly address it twice (2007, 5–11, 186–89) and offer a definition—“political authorities, systems of rules or arbitration, and the distinction between governing and governed persons” (10). Ignoring this point, Golden writes, “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” (2009, 295). He may think so, but I find it highly plausible given that the earliest accounts of steppe society describe aristocracies, and that comparable processes have been documented in the Korean, Manchu, and Chinese cases reviewed in the book.
