Who Were the Mongols?

 My new site: www.idiocentrism.com

At gmail dot com I am known as emersonj

"Instead of ethnic designations we shall have to deal with terms of nomenclature reflecting essentially the form of the constitutional organization of nomadic groups. We believe that the dark mists obscuring the history of the steppe would be dispelled sooner if emphasis were laid on the study of the migration of political symbols rather than that of hypothetical migrations of ethnic units, on the 'alarums and excursions' of political groupings, rather than on the mythical meanderings of self-conscious ethnoses, each bent on proagating its particular linguistic or ethnic self." (Boodberg, p. 305).

What is a Mongol? In this section I will briefly discuss different ways of answering this question, as put forth respectively by modern scholars, by traditional civilized historians, and by the Mongols themselves as seen in the Secret History and Rashid ad-din.

In modern scholarship the names by which barbarian peoples are designated are usually thought to refer to racially, linguistically, and culturally defined ethnic groups (with "culture" often being just a label for the way of life of a linguistically-defined group -- e.g. "Polynesian culture"). For a variety of reasons I think that this practice is unhelpful when discussing the steppe peoples. The large steppe confederations such as the Huns or the Mongols were never mono-lingual or racially uniform, but were coalitions of diverse origin. Even the component tribes were mixed: practices such as kidnapping, slave-trading, exogamy, fostering, alliance, adoption and coerced military recruitment meant that bilingualism and intermarriage were highly prevalent. Furthermore, the mobility of the steppe contrasted sharply with the rootedness of peasant families who lived generation after generation in a single village. The steppe world was fluid and frequently-stirred, and by and large the steppe peoples shared a common steppe culture defined by pastoralism and war rather than by ethnicity, and the peoples spoken of in the civilized histories are normally armies: the steppe enemies of the civilized world.

For the record, the earliest steppe barbarians were, by and large, Caucasians speaking Iranian languages (the Scythians, et al); but later on East Asian peoples, speaking first Turkish and later Mongol languages, replaced them. But for the reasons given above, to speak of separate Scythian, Turkish, and Mongol cultures is misleading. At any given time there might be difference between the eastern steppe and the western steppe, and over time there were historical changes, but there was no significant systematic contrast between Mongol-ness, Turkish-ness, and Scythian-ness. (To spare the reader, further discussion of this point has been relegated to Appendix One).

In the light of this, the famous sloppiness of Chinese, Persians, and  Europeans in their use of ethnonyms seems much less problematic. European writers of the later period used "Scythian", "Turk", "Hun", "Tatar", and "Mongol"almost interchangeably based on the literary effect desired.  The Chinese used "Hu", and sometimes "Ti" as generic terms, and often identified other peoples as "Hsiung-nu" even though they were clearly a different people. The Chinese called the Mongols (and some of the Turks) of Chinggis' time "Tatars", which was the name of just one of the Mongol tribes.  The word "Hun", which was used both near China and in Europe, may have been a generic Persian or Turkish word. All this makes sense if you understand that these words were functionally and politically, rather than linguistically or racially defined. The Scythians/Huns/Turks were "whichever cavalry people north of the Black Sea it is that keeps raiding us" -- and similiarly for China and Persia.   In fact, I believe that for a historian this functional definition is the best one; modern ethnographic definitions add relatively little of value and can cause more trouble than they're worth.

A careful examination of the sources can help us understand how Mongol groups were defined and how the Mongols themselves named human groups. Below I will discuss Rashid ad-Din first, and then The Secret History.

Mongols in Rashid ad-din

Rashid ad-Din's "Book of the Tribes" (written more than a half century after Chinggis Qan's death by the Persian chief minister of Persia's Mongol rulers) treats Mongols as a subclassification of Turks. (This is not really a bold stroke, since the Mongol and Turkish languages are closely related and the ways of life of many of the Mongols and Turks were almost identical. The majority of Chinggis Qan's horde were Turks). Rashid divides the tribes into four groups (with some overlap): first, six Turkish tribes from west of Mongolia descended from the legendary Oguz, all of whom joined the Mongol confederation rather late. Second, a group of fourteen Mongol tribes which had members among the Mongol hordes, but which did not exist as political units and were commoners or unfree. Third, nine important peoples who had their own leaders, including among them Turkish and Mongol groups together with the Tanguts. And finally, twenty-four Mongol tribes more independent than the tribes in the second group, of whom the first eight (the Terligen), were less honored than the next sixteen, the Nilun, who were the most closely related to Chinggis Qan. Rashid's lists begin with the tribes furthest from the Qan and ascend to those closest (without, however, mentioning the royal clan itself.)

By and large these tribes are treated consistently with what we see in the earlier Secret History. One exception is the Dorben, which should not be treated as Nilun because they were descended from Alan Qo'a's brother-in-law Dobun rather than from Alan-qo'a herself. The fact that Bolad, one of Rashid's most important informants, was a Dorben almost certainly had something to do with this upgrade. Not only in Rashid's work but also in the Secret History, the chroniclers' reports on the past can be presumed to have been distorted by their awareness of the relative contemporary statusses of the descendants of historical figures described, and particular the chroniclers' own claims based on descent from historical figures. Mary Douglas has shown how even in literate China geneologies were mostly fictitious further back than the sixth generation; the levels older than that really existed only to ground politically-useful kinship relationships. This is only the more true among the Mongols, whose culture was almost entirely oral before about 1206 A.D.

In sum, the tribes in Rashid are defined two different ways: by kinship, and politically. Politically the groups ranged from dispersed commoner Mongol groups, through dispersed commoner Turkish groups, through miscellaneous semi-autonomous groups, and finally to honored Mongol groups with varying degrees of claims to kinship to Chinggis Qan. As a rule of thumb it is safest to assume that the political definitions are dominant and that the kinship claims are secondary or fictitious.

Tribal Identification in The Secret History

The Secret History, the Mongol version of which was compiled as early as 1228 AD (immediately following the death of Chinggis Qan) is the most authentically-Mongol source for the life of Chinggis Qan. In this work we see three forms of group designation. First, a present- and future- oriented clan defined by kinship; second, a past-oriented clan, also defined by kinship; and finally, group membership based, not on kinship, but on allegiance to a war chief. As a complicating factor, the kinship-defined groups could be variously defined depending on which ancestor was chosen as the cut-off, giving each individual potential membership in several different groups of varying sizes, with the smaller groups nested in descending order of size within the larger groups.

For example, Temujin was simultaneously a Mongol, a Kiyan, and a Borjigin. The Borjigin were limited to Temujin, his brothers, and their descendants; this group was defined by descent from Temujin's father Yesugei. The Kiyan were defined by descent from Temujin's grandfather Bartan-ba'atur, and consisted of the Borjigin plus Temujin's uncle Daritai, his cousins Quchar and Onggur, and their descendants. (The Mongols did not limit inheritance to sons, but left it open to other male kin not directly descended from the departed Qan). The Mongols were most probably defined by descent from the legendary Qaidu, Temujin's great-great-great-great grandfather, but drew its prestige from Temujin's great-grandfather Qabul Qan, who had been a important figure during the first part of the twelfth century (which was really the last time that the Mongols had had much success until Temujin came along).

This narrow definition of clan defined eligibility for inheritance -- thus, if Temujin had been defined as a Borjigin, only his brothers and his uncle Daritai and their descendants could have inherited Chinggis Qan's position, whereas if he had been defined as a Kiyan several more distant uncles and cousins would have been eligible. However, there is also a second definition of clan which defined the scope of a leader's claims. Once Temujin had defined himself as a Kiyan, he could claim leadership for his core Kiyat group he led over all descendants of any male ancestor, back past Qabul and Qaidu all the way to the legendary Bodonchar and even further. Furthermore, he could claim for himself all the prestige and magical power of any of his male ancestors, which in this case especially meant Qabul and Qaidu. This is the backward definition of clan, which mostly lays claims in the present on the basis of kinship relationships with the dead, including legendary and mythical figures. The names Kiyan, Borjigin, and Mongol all identify clans with only two or three generations of living members with ancestors ten or more generations in the past (Mongqolin-qo'a, Borjidai-mergen -- both of whom are mentioned in the Secret History -- and Qiyan the brother of Neguz. This rather resembles the Chinese practice of naming usurper dynasties after great past dynasties, or the American practice of naming new towns after major European cities.) This does not mean that all descendants of Bodonchar, for example, could claim to be the Borjigin leader, but that the Borjigin here on earth, defined as Yesugei's descendants, were laying claim to the allegiance of (and making an offer of leadership to) all the descendants of Bodonchar.

The third form of group membership is simply military allegiance to a war chief, without any regard to kinship relations. In context, anyone following a Kereit, Mongol, or Tatar leader into battle might be called a Kereit, a Mongol, or a Tatar. In practice, folowers could be of many types, including sworn brothers, volunteers, subject tribes, tribes and individuals which had recently been pressed into service, and even actual kin. The steppe armies of any size were always defined primarily in this manner, and within them kinship really defined only the inheritance of leadership, including the leadership of a few trusted and privileged subgroups. Most members of Temujin's horde had joined, voluntarily or involuntarily, as individuals, without regard for kinship bonds, and if they had joined as part of a kin group in most cases their group was dispersed. (This is one of the senses in which it might be said that the Mongols were a "modern" rather than a "traditional" society.)

Temujin was a direct descendant of Qabul Qan, Qaidu Qan, and the legendary founder Bodanchar. He inherited some of their nobility and prestige, and was able to make special claims on any of the other descendants of these men. He also was the son of Yisugei, who it seems clear had been ready to make a claim, based on his successes against the Tatar enemy, to be the Mongol Qan against his distant Tayyichi'ut cousins who had had little success in that position. But all the claims Temujin was able to make made him enemies. Besides the Tayyichi'ut, his Jurgin and Kiyat cousins also had their claims, and when he was defeated in his first try at leadership, most of them separated from him. Ultimately he was to kill all of the Kiyat and Jurgin kin of his generation except for most of his brothers and one of his Kiyat cousins (Onggur). When he defined himself as a Borjigin through Yesugei, in theory all of his father's descendants were in line to succeed him, but in fact only his own sons were considered. The definition of the Borjigin as the descendants of Bodonchar was merely an ambitious claim.


The primary organizing principle of Mongol society was a functional one -- military leadership. Military groups were united by an absolute allegiance to a leader. Noble descent was a major resource for leaders putting together a following, as were a reputation for generosity and fairness, as well as omens and prophecies of various sorts. Kinship was also one of the resources a leader had in gathering his following, though it was a treacherous one since near kin were also competitors for power. A leader could also gain followers from his tribe's hereditary dependents and his mother's and his wife's peoples, but many of his followers were unrelated "nokor" who came as volunteers on an individual basis. Once recruited, in any case, followers were part of the leader's military machine and functioned according to its rules.

Purer forms of clan organization probably did exist at times, especially during the periods of disunity. But these groups were not the ones known to history as "The Mongols", and even if defined by kinship, their primary function was still military. It's interesting to ask what the function of clan membership, and genealogy, was for dispersed, leaderless, powerless tribes such as the Jalair. It would seem to be limited to incest rules about intermarriage, and the faint, distant hope of someday reclaiming past glory.

This squares with most of what we know about the barbarian peoples of history. Whether on the borders of China or Rome, these defined themselves as political and military units organized in obedience to a war chief, rather than racially, linguistically or culturally (except insofar as they shared a common military culture). The language of the group and its nominal ethnicity was normally that of the war chief. When the Huns disappeared, this does not mean that all the Huns were killed. It means that the political unit ruled by Attila disintegrated, and its members dispersed into various smaller groups, most of which were not led by Huns. Barbarian "nations" existed for the military purpose of gaining concessions of various sorts from the civilized world, and the groups were defined by their military forms of organization.

Boodberg , Peter, “The Altaic Word for ‘Horn’ in the Political Nomenclature of the Steppe”, Selected Works, California, 1979, pp. 296—305.
Lindner, Rudi Paul, "What was a Nomadic Tribe?", Comparative Studies in Society and History, 1982.
Wolfram, Herwig, tr. Dunlap, History of the Goths, California, 1988.
Wolfram, Herwig, tr. Dunlap , The Roman Empire and its Germanic Peoples, California, 1990.
Heather, Peter, The Goths, Blackwell, 1996.

Dunnell: Xi-Xia.

Wittfogel: Qara-qitai

Shu Ching: Wen/ Wu Wang.


During the nineteenth century and before, many attempts (which we would call reductionist) were made to put the study of mankind on a scientific basis. Enormous effort went into classifying the peoples of the world according to race, language, and (toward the end of this period) "culture" -- the last of which often seemed simply to mean the way of life of the speakers of a given language or language group (e.g. "Malayo-Polynesian culture".) Without going too far in a postmodernist direction, I think that it is fair to say that these attempts, especially when they attempted to find a describable "nature" in racial, linguistic, and cultural groups, were guilty of "essentialism" in their tendency to posit essentially-fictional "pure" races, languages and cultures at some indefinite point in the past. The mere fact of sexual reproduction tends to ensure, given the vicssitudes of history, that racial groups will all be mixed, and cultural and linguistic borrowing are also powerful influences toward impurity.

European political events during the 1940's tended to put an end to attempts describe humanity according to racial categories, but there still exist strong tendencies to describe peoples according to their language and culture. Thus, Mongols are of Mongol language and culture (and also, as it happens, also "Mongoloid" in race). Turks and Manchus are of Turkish and Manchu language and culture respectively, and so on for Scythians, Tokharians, Yeniseians, Hungarians, and so on.

It is one of the definitions of science that it finds unexpected, relatively simple ideas which succeed in explaining a large amount of confusing data. These simple ideas are often called "causes". Thus, diabetes is caused by the failure to produce insulin, smallpox is caused by the smallpox virus, and so one. The nineteenth-century thinking about race, language, and culture tended to regard them as causes. My argument here is that race and language are not causes at all, but only markers. Peoples who live pretty much the same way really are are pretty much the same, even though their languages and races (e.g. Scythian, Turkish, Hungarian, and Mongol) are widely different. Likewise, peoples whose languages are more or less closely related to each other but not to most other languages (e.g., the Hungarians, the Finns, the Mordvins, and the Vogul) are not necessarily similiar at all. (For the idea to that language is causal to have any power, the Hungarians should differ from the Austrians in a way analogous to the way that the Finns differ from the Swedes -- I case I find hard to make). And finally, peoples whose languages are unique and unrelated to any other (e.g. the Basques, the Burushaski, and the the Nivkh) are not themselves unique.

I have taken the long way around to getting to my main point. When you are talking about the Mongols who founded the Mongol empire, rather than analyzing "Mongol culture", you should analyze the whole history of the steppe/sedentary confrontation. The political and military relationships that anthropologists tend to bracket out are the most important things to look at. On the steppe, at least, race, language, and "culture" are useful in the reconstruction of the past history of a people, but they don't tell you what you need to know about what that people is really like. The Mongols were most like Turks and Scythians, and the early Hungarians were like all three of them -- regardless of language or race. They all were defined by the steppe way of life and its military relationship to civilization.

The Mongols, Tangut, Toba, Khitai, Karakhitai, Jurchen, Ottomans, Hsiung-nu, Huns, Kushans, and so on which you read of in history were all large, complex, politico-military groups of mixed origin. Their tribal names were derived from the tribal designations (some of them rather nominal) of the clans from which their leaders were chosen. But the Khitai who founded and ruled the Liao dynasty, for example, were not really the same as the "tribal" Khitai who existed before the Liao dynasty and who also survived it. The pre-Liao tribal Khitai were merely one of the steppe peoples. The founding Khitai were most the peoples of the steppe -- as organized and led by the Khitai. The Khitai Liao were the rulers of North China and their civil and military supporters of Khitai origin (including some "tribal" Khitai). The post-Liao tribal Khitai were once again just one of the many steppe peoples. (This analysis can be repeated, with variations, for the fugitive Khitai founders of the Karakhitai state).

So back to the original question: Most broadly, the Mongols were the steppe army, constituted of troops of widely varying ethnicities, which conquered much of the world in the thirteenth century. This army united almost all the forces of the steppe and represented the steppe barbarians versus civilization.

Most narrowly, the Mongols were the members of the kin group, descended from Chinggis Qan, which led the Mongol army and governed the Mongol Empire. In between, the Mongols were those who could establish a patrilineal kinship link with this ruling kin group through Yisugei, Qabul Qan, Qaidu Qan, Bodonchar. Probably those with a link through the female line (e.g. the Onggirat) were also Mongols, simply because they were often closer to the leadership group than many patrilineal kin. Finally, all Mongol-speakers were also Mongols, regardless of how distant the relationahip to the ruling clan was. But this anthropological definition of the term was not the one the Mongols used, nor is it a useful definition for historians.



Buell on Mongol kinship

Also oboq etc. Gather Buell / Rachewiltz

Collar, smokehole, head -- Golden lion -- a "people" was a led unit united by someone. Two contrasting lists from SH. Composite political entity.

Class, kinship are options. Like language, culture, race, exaggerated.

Dissect kinship (intermarriage; solidarity; leadership; inheritance). Otchigin and mother (Bekter). Matrilineal aspect.


The idea that the specifics of languages have a causal effect on culture and behavior is identified with Whorf, but the following example from Levi-Strauss, (Structural Anthropology, Basic Books, 1963: "Language and the Analysis of Social Laws", especially pp. 63-5) conjectures that various language groups are each correlated to a specific kinship system, and thus to a specific kind of society. For example: "Sino-Tibetan kinship systems exhibit quite a different type of complexity.... Translated into more general terms applicable to language that would conform to the following linguistic patterns, we may say that the structure is complex, while the elements are few, a feature which may be related to the tonal structure of these languages.....The widely recognized feature of the Oceanic kinship system seem to lead to the following formulation of the basic characteristics of the linguistic pattern: simple structure and few elements.....The linguistic patter to correlation of this situation [American Indian kinship] is that certain of the American Indian languages offer a relatively high number of element which succeed in becoming organized into relatively simple structure by the structures' asymmetrical forms."

It has even been proposed (Eberhart, Wolfram, Conquerors and Rulers, Brill, 1952, pp. 69-72) that the Mongol, Turkic, Tibetan, and Tungusic peoples each had their own distinct political forms, ways of life, and even their own types of pastoralism varying according to the combined effects of language group and the type of livestock raised (Tibetans raising yaks and nomadizing over a rather small area, for example). Since groups speaking languages belonging to each of these language groups lived in various physical environments suitable for various sorts of livestock, it's hard to imagine how this idea could be more than a very rough rule of thumb.

Jackson tr. Rubruck, p. 120: 1253 Mongol self-reference as "Mo'ol" rather than Christian.

In Yuri Tambovtsev's study of the phonological relationships between the various Turkic languages, Mongol seems to be closer to Uzbek and Anatolian Turkish, the most central Turkic languages, than many languages which are classified as Turkic (Mongolian Studies, XXIV, 2001, pp.41-84.)

Mary Douglas, "Passive Voice Theories in religious sociology," In an Active Voice, Routledge Kegan Paul, 1982, argues against the idea that languages have causal effects..

There has been a tremendous amount of speculation about the languages of the Huns and the Hsiung-nu ("Were they Turks or something else?"), and also about the possible identity of these two peoples (since a Sogdian letter exists calling the Hsiung-nu "Huns".) As I have said, I think that the former question is of relatively little importance. The problem with the second is that the Hsiung-nu disappeared in the east about two centuries before the Huns appeared in the west, and no one can say what happened in between. It seems likely to me that "Hun" was a generic Iranian or Turkish word, and that the West learned it from a Iranian- or Turkish-speaking people. And even if the "Hun" name in the West was meant as a reference back to the Hsiung-nu in the east, it was a semi-fictitious attempt to appropriate the past glory of an rather distantly related ancient nation, the way Tamerlane claimed the heritage of Chinggis Qan, the Tangut claimed the mantle of the Toba, and various ephemeral Chinese dynasties took on the names of such ancient Chinese states as the Ch'in, Chin, Wei, Yen, etc.

A look at Dutch, Low German, Catalan, Gallego, Provencal, Scots, Sardinian, Romansch, Danish, Norwegian, and Swedish will illustrate one of my points. Some of these are very distinct languages which have always been called dialects, some once were languages but now are dialects, and some are independent languages which are only very weakly distinguishable from other independent languages. The reasons are to be found not in linguistics but in politics.

Very roughly speaking, the first steppe peoples were probably Caucasians speaking a Northern Iranian language: the Scythians who raided Persia and Greece, and who seem later to have spread far to the east. As time went on, Turkish and then closely-related Mongol peoples (both probably originally forest peoples) of East Asian race occupied Mongolia and began to drift south. The Hsiung-nu who fought China and the much later Khirgiz were probably Turkish speaking and of East Asian race, but some believe that they spoke paleo-Siberian or Yeniseean languages belonging to a pre-Turkish group which is now almost extinct (and very little known); these peoples were also of East Asian race. In Europe, the Hungarians (ca 800-1000) were Caucasians speaking a Finno-Ugric language, and presumably had earlier been northern forest peoples. However none of these groups was ethnically uniform.

The ascription of causal effects to language or race was a widespread nineteenth- century error, and this error still has not entirely disappeared in the case of language. (I have mercifully put my arguments for this point in an appendix at the end). The value of language and race for steppe historians is purely as markers. Genetic studies have confirmed that the Hazara in Afghanistan are indeed closely related to the Mongols (even though they are now Muslims speaking an Iranian language). The atypical language of the Chuvash Turks lends support to the idea that they are the descendants of the first wave of Turkish migration to the west, and perhaps the descendants of the Volga Bulgars. The Indo-European -- but not Iranian -- language of the Tokharians does lend some credence to the theory that this Caucasian people came from the west at a very early date. And so on.