Mongol Manpower and Persian Population Author(s): John Masson Smith, Jr.

Reviewed work(s): Source: Journal of the Economic and Social History of the Orient, Vol. 18, No. 3 (Oct., 1975), pp. 271-299 Published by: BRILL Stable URL: http://www.jstor.org/stable/3632138 . Accessed: 02/02/2012 19:48
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.

BRILL is collaborating with JSTOR to digitize, preserve and extend access to Journal of the Economic and Social History of the Orient.

http://www.jstor.org

Journalof the Economicand Social History of the Orient,Vol. XVIII, Part III

MONGOL MANPOWER AND PERSIAN POPULATION

BY

JOHN MASSON SMITH, Jr.

(University of California, Berkeley) During the space of twenty eight years, as I have mentioned, the Scyths continued lords of the whole of Upper Asia. They entered Asia in pursuit of the Cimmerians,and overthrew the empire of the Medes, who till they came possessed the sovereignty. On their return to their homes after the long absence of twenty-eight years, a task awaited them little less troublesome than their struggle with the Medes. They found an army of no small size prepared to oppose their entrance.For the Scythian women, when they saw that time went on, and their husbands did not come back, had intermarriedwith their slaves. Herodotus, Persian Wars, IV: i What armyin the whole world can equal the Mongol army?In time of action, when attacking and assaulting, they are like trained wild beasts out after game, and in the days of peace and security they are like sheep, yielding milk, and wool, and many other useful things. In misfortune and adversity they are free from dissention and opposition. It is an army afterthe fashion of a peasantry, being liable to all mannerof contributions .... It is also a peasantryin the guise of an army, all of them, great and small, noble and base, in time of battle becoming swordsmen, archers and lancers and advancing in whatever manner the occasion requires. Juwaini (Boyle trans.), I, p. 30

I Modern scholars consider the Mongol conquests as triumphs of quality ratherthan quantity.They attributethe Mongols' extraordinary military achievements, the winning of an unequalled empire almost without the loss of a battle, much less a war, to their remarkable strategic and tactical skills, and to their good organization, great discipline and matchless leadership.These interpretationsare correct, as far as they go; the Mongols had these qualities, but some further qualificationis needed. Most of the methods employed by the Mongols in war were not new. The mounted archer,able to loose the "Parthian shot" (and a variety of others), had been riding all across the Inner

272

J. M. SMITH

Asian steppe and in parts of the Middle East for almost two millenia (since before the Parthians,in fact) and the styles of fighting and campaigning appropriateto him had long been worked out. The essential methods of evasion and encirclement had been used strategicallyby the Scythiansagainst Darius, and tacticallyby the Parthiansat Carrhae and the Turks at Manzikert,to cite only a few examples.And nomads before the Mongols had enjoyed the advantageover sedentarypeoples of cheaper horses and socially common cavalrymen.Mongol warfare was distinguishednot so much by its skill and aptitudesas by its scale and persistence. The size of the Mongol armieshas not been appreciated. The sources of Mongol history know the quality of Mongol troops, but they remarkas well the great size of their forces. Marco Polo claimed that the Mongol army numberedbetween six hundredand six hundredand fifty thousand men in Russia and the Middle East together 1) and Rashiduddin2) and Haython3) reported six hundred thousands in Russia alone. Modern scholars have disregardedthese figures. Some make perfunctory efforts, without reaching agreement or attempting precision, to estimatethe numbers of Mongol and non-Mongol troops in the imperialarmy, but most agree implicitly with Barthold that the sources give "fantastic figures" that "deserve no credence whatsoever" 4). They are too skeptical.The sources give us manpower data that the Mongols themselves compiled and relied upon in conducting their masterful warfare, and these data show that the Mongol armies were very large indeed. The Mongol conquests were the product of the irresistible combination of skill and numbers. The story of the Mongol conquests may seem, and is often made, one of Davids facing Goliaths, a handful of nomads taking on China
I) Marco Polo, The Travels(Penguin, 1958), PP. 310-311. of GenghisKhan, J. A. Boyle trans. (N.Y. and 2) Rashiduddin, The Successors London, I97I), p. I28. desCroisades, desHistoriens 3) Haython, Flos HistoriarumTerreOrientis,in Recueil Documents Arminiens,II (Paris, I906), p. 215. 4) W. Barthold, TurkestanDown to the Mongol Invasion(London: 2nd revised edition, 1958), p. 404.

MONGOL MANPOWER

273

and Russia, "a hundred thousand conquering a hundred million" 5). There is some truth in this. The population of China may have been one hundred millions in the twelfth century6), and the Mongols even today are only some three millions7). But the Mongols had reinforcements who tend to be overlooked, and, as we shall see later, resources that have not been correctly assessed. The Mongol empire expanded rapidly, and by the I240's included the whole Inner Asian steppe and all of its nomads and mounted archers. The "Mongols", owing to the elasticity, comprehensiveness,and linguistic and ethnic unconcern of tribalism,were not only Mongolians, but Turks and Tunguses and Tibetans. To some extent they were even Georgians, Russians and Chinese, to name only a few of the non-nomad peoples who participated in the Mongol conquests, although the differentiationbetween nomad and non-nomad meant more to the Mongols. The Mongol forces were thus larger and more heterogeneous than first impressions and many studies might suggest 8). The figures usually cited for the size of the Mongol forces derive from the Secret History of the Mongols,or from sources, such as Rashiduddin, indirectly dependent on it. The Secret History gives several enumerations of units from which, owing to their decimal organization,troops totals can be derived. In i 20o6,following Chingis' unification of the Mongols, 95 or 96 Thousands are listed, plus a timen (Ten Thousand) of guards,for an implied total of Io1 to Io6,ooo men 9). To these are added, following Jdchi's conquest of various
5) N. Riasanovsky, A History of Russia (N.Y., 1963), p. 75. 6) C. P. Fitzgerald, China: A Short CulturalHistory (N.Y.: 3rd edition, 1961),
P- 387.

7) D. Sinor, Inner Asia: A Syllabus(The Hague, 1969), pp. 37, 49. 8) Some historians have understood this, but they have not reached agreement or precision in their estimates of this larger force. Barthold (loc.cit.) estimates
150-200,000

C. K6priilti trans. [Ankara, p. 13) gives one million. B. Spuler (in Iran Mogollar:, see P. Pelliot, 9) Section zoz. For the text of the SecretHistory of the Mongols, Histoire secretedes Mongols(Paris, 1949). For a translation, see E, Haenisch, Die Geschichte der Mongolen Geheime (Leipzig, 1948).
1957], p. 439) has 1.4 millions. Most historians concentrate on Rashiduddin's figure of I29,000 men.

men. H. D. Martin (in The Rise of Chingis Khan [Baltimore, 1950],

274

J. M. SMITH

"forest peoples" in 1207, two tiimensof Oyirad and Kirghiz, and

other forces from other tribes 10). Subsequently, but before the beginning of the North China campaigns in i z i i, judging by the placement

of the report in the Secret History,there is an apportionmentof peoples and troops among the leading members of Chingis' family. Some 16 of the commandersof Thousands named in the zo6 list, presumably along with their troops, are assigned to membersof the Family, along
with 28,500 other tents (probably with the implication of one soldier per tent), for a total of 44,500 men 11). This would have left about 80

of the regular Thousands enumeratedin the 1206 list, and the Guards Tiimen.The Regular, Guard and Family forces together thus would

in I22712).

have totalled 134,500 men. Rashiduddin, using similar information, says that Chingis commanded 129,000 men at the time of his death

But these figures apply only to Outer Mongolia, as we shall see, and to the period before the beginning of imperial expansion. These are the "original" Mongols 13), perhaps, but the Mongol enterprise soon involved many other peoples who are not counted in the Secret History. Enumerationsof these exist; figures from Polo, Rashiduddin and Haython have been cited above, and Jiizjdni estimates a total force of from six to seven hundred thousand men 14). But because of their inconsistencies and especially because of the very large numbers involved, they are not much used. Canthey be reconciledand credited? One way of appraisingthe greater Mongol army is provided by data from the Middle East. Hiilegii was assigned "two persons out of every ten in the Eastern and Western armies" with whom to conquer the

io) Section 239. i i) The apportionment is in Section 242. The relation of the North Chinacampaign of 211 begins with Section 247. i2) Martin,op. cit., pp. 13-14 and p. 14, n. 6, citing Rashiduddinvia M.D'Ohsson, II (Amsterdam, 1835), pp. 3-5. Histoire des Mongols, 13) These "Mongols" already include some Turks, as for instance the Kirghiz. 14) Cited by Barthold, loc.cit., from Jaizjini, Tabaqit-i Nisiri, partial edition by W. N. Lees, et al. (Calcutta, 1864), pp. 273, 968,

MONGOL MANPOWER

275

Middle East 15). If we can establish the size of Hiilegii's army, we can

then also calculatethe numbers of the whole Mongol force, and of its regional subdivisions as well, since Mdngke Qaan, who gave not only
Hiilegii but also Qubilai a fifth of the Mongol troops
16),

was effecting

a reorganization of the Mongol territories and a redistribution of Mongol forces. He was revising the old, four-way division of the empire that had followed Chingis' death, adjustingfor the elimination of most of the descendantsof Ogedei by reassignmentof their assets; and equalizing (by reducing) the assets (and power) of the houses of J6chi and Chaghataithat had been disproportionatelyenlargedby the western conquests after 1227; and creating new portions for his brothers, Qubilai and Hiilegii. Mdngke, Qubilai, Hiilegii, the heirs of Chaghatai,and the descendantsof Jdchi were each to hold an equally valuablepart of the Mongol empire and an equal share of the Mongol nomad forces. The description of Hiilegti's forces in Juwaini17) shows that the "Eastern and Western armies" providing these portions for Hiilegii and Qubilai were the nomad, largely Mongolian and Turkish forces that constituted the main armies of the empire, and that the nonnomad auxiliaryforces of Iranians,Georgians, Russians, Chinese (and so on) were not also being reapportioned.Sedentarymanpower,when identified,and therewere only presentin Hiilegii's army,was separately one thousandhouseholds of Chineseartillerists -far from a fifthshareof
the non-nomad manpower available to the Mongols by the I25 o's.

Several sources give figures for the size of the Mongol force in the Middle East, and the figures are, again, large, round and perhaps questionable, although they are consistent. Marco Polo claimed that Hiilegii had three hundred thousand troops with which to withstand the attack of the Golden Horde in I26i 18). Rashiduddinhas Ketbuqa,
15) Juwaini, The History of the World-Conqueror, J. A. Boyle trans. (Cambridge [Mass.], 1958), II, p. 607; Ta'rikh-iJahdn-Gushi,Mirzi Muhammad Qazwini ed. i6) Ibid.
17) Ibid., trans., II, pp. 607-608; i8) Polo, op. cit., p. 3Io. 6, 1937), III, p. 90. text, III, pp. 91-93.

(3 vols.: Leyden and London, 1912

276

J. M. SMITH

after his defeat in I26o, taunt his Mamluk captors, saying (in effect), 'you've got me, but there are three hundredthousand more like me' 19). Haython, later, also counted three hundredthousand men in the army
of Oljeitii 20).

The number of Mongol troops is also reflected in their "order of battle". Occasionally,as in the Georgian account of Ghiz~n's Cbronicle's
Syrian campaign of I299-1300, the number of tiimens will be given in

the sources--thirteen in this case 21). More commonly, the number of units is omitted, but the principalcommandersare listed. I believe that these named commanders,except as otherwise specifiedin the sources, were commandersof tiimens. for Rashiduddinlists thirteencommanders
the 1299-1300 campaign: Ghazan, Chiiban, Sultan, Tagharilja,

Ilbasmish, Chichek, Qurumshi and Qurbagha in the Center; and Mulay, Satilmish, Qutlughshah, Yemin and Murtad in the Right Wing 22). In describing the Right Wing, Rashiduddin even specifies that each commanderled a ti'men,and we would not expect the commandersof the Centerto be of lesser rank. The correspondenceof the information from the two sources is exact. To assess Hiilegii's forces in similarmanner,let us look at his order of battle for the campaign of I257-58 against 'Irdq-i 'Arab, Baghdad and the Caliph,as given by Rashiduddin.The extensive strategyof the campaign itself suggests a very large force. The Center, marching direct on Baghdad from Hamadin via Kermanshah and Hulwun, included Hiilegii, Kuka Ilka, Arqatu and Arghun Aqa, as well as Suntai, who came in from some detached operation en route. In the Left Wing, advancing from Luristan via Khiizistdn, were Ketbuqa, Qadsunand Nerk Ilka. In the Right Wing, proceedingfrom Azerbdyjdn via Irbil, were the J6chid princes, Bulgha, Tutar and Quli, and the generals Buqa Timur and Sunchaq. Chormaghun and Baiju led a
de la Perse, E. Quatrem"reed. and trans. 19) Rashiddudin, Histoire des Mongols (Paris, I836), p. 352. zo) Haython, loc.cit. 21) M. F. Brosset, Histoire de la Giorgie,I (St. Petersburg, 1849), pp. 63o-632. 22) Rashiddudin,History of GhdZLdn Khan,K. Jahn ed., I (London, 1940), p. 127.

MONGOL MANPOWER

277

separate corps into Mesopotamia from Rim 23). Thus fifteen commandersare named, and since Chormaghunhad formerly led three or his force may have included more than two. Thus Hiilegii four tlimens, would seem to have commandedfrom fifteen to seventeen tiimens. has four tfmens Other sources bear this out. The Georgian Chronicle that in the force under Chormaghun originally occupied Azerbdyjin,
ca. 1230, and that was later made subordinate to Hiilegii (and moved by

him to Rilm, since he wanted Azerbdyjin for himself), and has six in Hiilegii's personal forces, stationed in Azerbdyjan 24). It goes tfimens on to mention the seconding to Hiilegii's command of other troops from the realms, and led by princes, of the Jachids and Chaghataids; it lists, however, only threeof these princes25). The numberof seconded princes-and with them, probably, tiimens-appears larger in other
sources. Juwaini gives six 26), and Bar Hebraeus, seven 27). If we take

for nearby Aaerbdyjdn, and those the figures of the Georgian Chronicle of Juwaini and Bar Hebraeusfor the princely forces (which may have been stationed in regions remote from Georgia and its chronicler)we have again, from different sources, the convergent reckoning of in Hiilegii's command. sixteen to seventeen tllmens How is this figure to be reconciled with the other total of three hundred thousand men commonly given for Hiilegii's forces? This largernumber can also be reachedby counting not only the "Mongol"

or Inner Asian nomad troops, but the "Tdjik"or local Iranian, Georgian, Armenian,and perhapsMiddle Eastern Turkish forces under Hiilegii's commandas well. These troops do not appearas of their disdainful clearlyin the sources.The Mongolswereprobably probablyfound them a capacities,and Muslimslike Rashiduddin
23) Rashiduddin, Jamic al-tawarikh, B. Karimi ed., II (Tehran, 1957), p. 707. 24) Brosset, op. cit., pp. 5II, 539. 25) Ibid., p. 541.

26) Juwaini, op. cit., trans., II, pp. 607-60o8;text, III, pp. 91-92. Grigor of Akanc', Historyof theNationof theArchers,R. P. Blakeand R. N. Frye ed. and trans.(Harvard Journal of Asiatic Studies, i2 [i949], PP. 269-399), p. 327, also has six princes, although as a unit Grigor is not an altogether reliable source: he has, for instance, the tiimen of thirty thousand men. E. Budge trans. (London, I932), I, p. 27) Bar Hebraeus, Chronography, 419.

278

J. M. SMITH

distastefulsubject,since many-perhaps most-of them were Christians. But they were there, if only noticed summarily, as by Rashiduddin when describing Hiilegii's Baghdad campaign, under the command of "the sul.tdns,mdliks and kdtibs of Iran" 28); these included, among others, the Georgian king and the "great men" of Georgia 29). These troops were also counted by the Mongols and organized in tells us that Hiilegii sent Arghun decimal units. The Georgian Chronicle Aqa to conduct a census in Georgia, and that he established nine
30). Elsewhere in Iran, Mongol administrators had been Georgian tfimens

in five Isfahin ('Irdq-i'Ajem), and of Tabriz-Azerbiyjdn-of tiimens arenot likelyto havebeenunitsof Mongols,since all31). Thesetfimens
they are located by cities, which would not have been a convenient way of specifying the mobile Mongol camps, and which would not have applied in particularto Qumm, Kdshin and Isfahdn,which were not in regions occupied by Mongols. Juwaini has in mind, surely, troops raised by local, non-Mongol commanders, vassal rulers and

at work on similarcensuses from the time of the conquest of Khurdsdn after the Khwdirezmian campaigns, and the occupation of Azerbayjan by Chormaghun.Their results are given in passing by Juwaini, who of Nishipuir-Tiis(Khurdsin), of Qumm, Kdshin and speaks of tiimens

in thesecitiesfromamong andkdtibs") administrators mdliks ("sul.tdns,

the local, non-Mongol (although perhaps nomadic) population. The Georgians, of course, were neither Mongols nor nomads. Taken or perhaps together, these Georgiansand others give us fourteentfimens, only thirteen, since the Georgian chronicler and Juwaini may both have been including the Azerbdyjin tiimenin their counts. Counted

up a force approximatingthe customary large total of three hundred thousand men.

theyhelpto make Mongoltiimens togetherwith the fifteento seventeen

28) Rashiduddin/ Quatremere, p. 264.

29)

Brosset, op. cit., pp. 548-549.

30) Ibid., p. 55T1.

31) Juwaini, op. cit., trans., II, pp. SII, ~i8; text, II, 248, 255.

MONGOL MANPOWER

279

II
From the preceding discussion can be seen, I believe, the numbers and kind of units the Mongol armies of Outer Mongolia, and later, of the Middle East, contained. But it is not clear how reliably a count of army units indicates actual manpower. The Mongol units were decimally organized, and named as Tens, Hundred, Thousands and Ten Thousands,but were these numbersrealistic,or only conventions? Ordinarily and rightly, estimates of numbers of troops, or indeed large quantitative statements of all sorts in pre-modern sources are treated with suspicion. The artistic tendency to exaggerate numbers of warriors so as to enhance the importance of the story, or of the protagonists, is well-known, and is not absent from Mongol storytelling. The SecretHistory has the Ong Khan and Jamuqa lead forty thousand men against the Merkits in retaliation for their raid on
Temtijin's camp and rape of Temiijin's wife, Bdrte 32). Implicitly, the

Merkits have forces of comparablesize, so that the episode as reported involves almost as many men as Chingis was later to mobilize from the whole of Outer Mongolia-an implausible suggestion, but one that follows the same artistic impulses as the Homeric story of the Trojan War. There are other reasons for distortion of estimates even by careful reporters.The pay and provisioning of armiesis very generallysubject to peculation, and one of the commonest devices for the diversion of funds is the padding of muster-rolls, which inflates the number of troops reported, enlarges the payroll, and enables the commander to pocket the surpluspay. As one of the Mongols' administrators put it:
Whenthey draw their pay and allowances the soldiers'numbersincreaseby but on the day of combattheirranksareeverywhere andthousands, hundreds and none presents himselfon the battlefield. A shepherd vagueand uncertain, wasonce calledto renderan accountof his office.Saidthe accountant: 'How the remain?' asked 'Where?' 'In the 'That', manysheep shepherd. register'. 'is why I asked:thereare none in the flock.' This is a repliedthe shepherd, to be appliedto their armies;whereineach commander, in orderto parable
32) Sections 106-107.

z80

J. M. SMITH increase the appropriationfor his men's pay, declares, 'I have so and so many men', and at the time of inspection they impersonate one another in order to

make up their full strength 33).

But Juwaini was not describing Mongol armies. He was citing these common corrupt and ineffectual practices to contrast them with Mongol honesty and efficiency. There are reasons why we may trust the estimates of the size of Mongol forces more than we do others, and ways in which the discrepancies in these estimates may be explained. The Mongols were notably effective campaigners,and the extent of their conquests and

in military of their victoriesarguefor their efficiency the regularity demanded as well. Moreover, their style of campaigning management and The Mongol field armieswere cavalryforces, cavalry efficiency. condition the numbers of horses since with unusually attached, large animalscould only be kept up to the high standards of the pastured the burdenof by Mongol strategyand tacticsby alternating required each warriorand his gear among a number of horses during the EachMongoltrooperthus took a stringof horsesto war; campaign. eachsoldierwas to report for Ghizdn's Syrian of I299-13oo00, campaign with five mounts34). These horses were normally supported by 35), so the Mongols had to be carefuland preciseabout the grazing becauseof numbersof men in their armiesand on their campaigns to findenough this high ratioof horsesto men, whichmadeit difficult the Mongol pasturefor a powerfularmy.Also, on some campaigns troops brough along not only their horses, but their familiesand unitswereconsidered as well. The Mongolmilitary animals subsistence to includeboth soldiersand their supportingestablishment: timens, their with are to the women, Chronicle, "myriads, Georgian according the size of a force would animals" and 36). Overestimating baggage
der 34) D'Ohsson, op. cit., IV, p. zz8; and J. von Hammer-Purgstall, Geschichte Ilchane,II (Darmstadt, 1843), pp. 85-86. The inadequate citations in these works make the source of this information uncertain; it is probably Wassif. 35) Polo, op. cit., p. 69. 36) Brosset, op. cit., p. 539.
33) Juwaini, op. cit., trans., I, p. 32; text, I, pp. 23-24.

MONGOL MANPOWER

28

cause it better to fit the pasture available,but might provide too few troops for the task at hand; underestimating could put too many horses and sheep on too little pasture. The fact that the Mongols so seldom got themselves into militaryor logistical difficulties--not even when campaigning in Russia in winter-suggests that they were usually well informed about the logistic potential of target regions, and able effectivelyto relatethis informationto the consumptiveneeds of their forces. To managethis they must have had a good idea of the size of these forces. Furthermore,the Mongol practice of re-equipping their troops by levies upon the animals and goods occasional (later, annual) qupchur of the adequately-stockedsoldiers for the benefit of those who had sufferedlosses during campaigns or from other hazards, assured that a close scrutiny of the numbers and equipment of the forces would be maintained 37). Some commanders might hope to gain extrabenefits from qupchur by over-counting their men, and some soldiers might over-representtheir losses, but other commandersand other soldiers -the ones who would have to pay-would do their best to ascertain that this paymentwas justified.

Finally,the Mongolshad less causethan most to be troubledwith of pay. The Mongol to enableembezzlement paddingof muster-rolls soldiers were not paid38). All in all, the Mongols had important reasonsfor trying to obtain accurate information about the size of their forces, and fewer causesthan most for attempting to miscount them. have not only to be supported The Mongol enumerations against out of but difficulties. Many helped general skepticism, particular armyof I227, but then have difficulty relatingthis force to a credible Mongol population.They have taken as the Mongols'generalconpracticeof levying as tax "onescriptionmethod their well-attested
see my "Mongol and 37) Rashiduddin/Jahn,p. 300. For a discussion of qupchur, Nomadic Taxation", Harvard Journalof Asiatic Studies(1970). 38) Juwaini, op. cit., trans., I, pp. 29-31; text, I, pp. 21-22.
18

scholars accept Rashiduddin's figure of i29,ooo men for the Mongol

282z

J. M. SMITH

tenth of everything",and have reconstructed from this a Mongol of about one million as thebasefor theforceof 129,000 39). population Thereare severalproblemswith this method.To begin with, when the Mongols appliedthe tithe as a conscription rate, they appliedit not to the generalpopulation,but to that segment of it that was liableto conscription--the adultmales40). Hadsucha method properly been appliedin Mongolia,the male populationby itself would have had to be over one million to provide 129,000 men, and the total would have had to be betweenfive and six millions.Such population a figureis out of the question:the modernMongolsonly amountto some threemillions at most41). Secondly,althougha Mongolianpopulationof one million in the early thirteenthcenturyis easily conceivable,a populationof one millionis OuterMongoliaalone is not. And it was Outer Mongolia that produced129,000 men, or even somewhatmore. The lists of forcesin the Secret History displaya Mongolstrengthof some 134,500 men priorto 1211 and the beginningof the North Chinacampaigns. Most of these troopscamefrom OuterMongolia,as is seen from the fact that the list of Thousandsof 1206 includesonly five thousand men of the Inner MongolianOnggiits, and no Tangqutsat all42). to these were madein the 1207 list, althougha dynastic No additions with the Onggiits is mentioned 43). Clearlythe population marriage and manpowerof Inner Mongolia,which probablyequalledor surpassed that of Outer Mongolia,then as now, was not accessibleto of 134,500 men. ChingisKhanat the time of his mobilization This placesa largeburdenon OuterMongolia.In recenttimes the populationof Outer (the PeoplesRepublicof) Mongolia-minus, to TannuTuva and some otherregionsfromwhich be sure,the Buryats, Chingisderivedhis earlyrecruits-has only just passedone million,
39) G. Vernadsky, The Mongols and Russia (New Haven and London, 1953), pp. 126-I27, 2I5-z26. Martin, op. cit., p. 14. 40) Brosset, op. cit., p. 55 . 41) Sinor, loc. cit. 42) Section o202. 43) Section 239-

MONGOL MANPOWER

283

and almost a quarter of these live in Ulan Bator. This population is probably double what it was in 1921 4"), and is surely rather larger also than that of the disturbedperiod leading up to 1207. How, then,
did Chingis raise 134,500 men from perhaps half a million people?

The difficulty appears again in even more exaggeratedform when such calculationsare appliedto the datafrom Mongol Georgia reported in the Georgian Chronicle. Hiilegfi's census enumerated nine timens, and the conscriptionapplied in Georgia took one man in ten 45).Nine tiimens-90o,ooo men-levied at a rate of one in ten would require an
adult male population of 900,000, and a total population of perhaps (multiplying by five) 4,500,000 persons. But, as the translator of the

Chronicle noted, in the early nineteenth century the Georgian Georgian was probably only around 225,000 46). In 1897 it was only population
some 1.3 millions 47). Just after the Mongol conquest it would scarcely

have been so large. Obviously, something is wrong, either with the force figures (which I would like not to believe), or with the understanding of the conscription technique through which the troops are related to the general population. The troops counted in the sources can be fitted to plausible populations, however, on the assumption that the Mongols conscripted all adult males into the army48). Working with this assumption,
134,500 troops, taken as one-fifth of the people of Outer Mongolia,

would imply a total population of some 650,000, only slightly more than the estimated population of 1921, and thus a believable figure. The results for Georgia would also be credible. The nine tiimensof adult males (of whomone-tenth were taken for service) would require
a total population of about 450,000.

This interpretationof the Mongols' census and conscription prac44) C. R. Bawden, The Modern History of Mongolia (N.Y. and Washington, 1968), pp. 404-405, 408. 45) Brosset, loc. cit. 46) Ibid., n. 2. 47) R. Pipes, The Formation of the Soviet Union (Cambridge [Mass.], 1954), p. 289. 48) For the Mongols, "adult males" were from 15 to 60 years of age: see Grigor of Akanc', op. cit., p. 325.

284

J. M. SMITH

tices makes sense in other ways as well. Adult male labor is not heavily involved in the subsistence routines of nomadism, as we can see from the accounts of both medieval and modern observers. Vreeland's analysis of the division of labor in an early twentieth century Mongol community has the women generally assigned to milking, collecting fuel, building fires, preparing and cooking food, and caring for the small children and for the sheep at night. The men carry on trading and caravanoperations, cope with shearing, slaughtering, butchering and tanning, engage in carpentry, ropemaking, metal-working and agriculture. Both sexes are involved in herding sheep; only men normally tend horses and camels. Vreeland notes that women often assist in loading animalsand in coping with the ger, that childrenbegin acting as shepherdsfrom age six or seven, and that boys begin shearing sheep as soon as they are strong enough to do so 49). Things have not changed much in Mongol camps in the past seven centuries.As William of Rubruck describedit:
It is the duty of the women to drive the carts, to load the houses on to them and to unload them, to milk the cows, to make the butter and grut [dried curd], to dress the skins and to sew them .... They also sew shoes and socks and other garments.... The women also make the felt and cover the houses. The men make bows and arrows, manufacture stirrups and bits and make saddles; they build the houses and carts, they look after the horses and milk the mares, churn the [kumis]that is the mares' milk, and make the skins in which it is kept, and they also look after the camels and load them. Both sexes look after the sheep and goats, and sometimes the men, sometimes the women, milk them50).

And things do not differ much among the various pastoral peoples. Among Barth's Persian Basseri, for instance, the women and girls usually perform most of the domestic chores, though men repair equipment and tents and make rope; all cooperate in making and breaking camp; and herding is usually (though not invariably) done by males, especially by unmarried men and boys down to age six, while milking is done by both sexes, but mostly by women. Barth
and KinshipStructure(New Haven: 3rd 49) H. H. Vreeland, MongolCommunity edition, 1962), pp. 48-51. Mission(London and in C. Dawson, TheMongol William of Rubruck,Journey, 5o) N.Y., I95 5), P. 103.

MONGOL MANPOWER

285

remarksthat the division of labor between the sexes is "highly elastic" and mostly determinedby pragmaticconsideration51). Thus women are usually responsible for the regular subsistence labor of nomadism,while men generallyattend to the more specialized and more occasional tasks, and even in these are often helped or replaced by women. The regular chore of herding normally involves male labor as well as female, but often in special forms and ways: using male childrenas much as adults, and with the men concentrating on the care of the great militaryand logistic animals.Consideringthis, and rememberingthe pragmatic "elasticity" of this division of labor by sexes, it is easy to understand and believe in the situation that Marco Polo reports:
And I assure you that the womenfolk [in Mongol society] buy and sell and do all that is needful for their husbands and households. For the men do not bother themselves about anythingbut hunting and warfareand falconry .... The wives aretrue and loyal to their husbandsand very good at their household tasks52).

Marco Polo was observing the Mongols in peacetime. But the implication of his descriptionis that, if the protection of the camps and herds could be provided for in some way not requiring the presence of the men-by the Pax Mongolica, for instance-then all the adult male Mongols could serve in the armies of Chingis Khan and his house. It would appear, furthermore, that the whole Mongol manpower was not only availablefor, but was used in war:
What armyin the whole world can equal the Mongol army?In time of action, when attackingand assaulting, they are like trained wild beasts out after game, and in days of peace and security they are like sheep, yielding milk, and wool, and many other useful things ... It is also a peasantryin the guise of an army, all of them,great and small, noble and base, in time of battle becoming swordsmen, archers and lancers and advancing in whatever manner the occasion requires.... The reviewing and musteringof the army has been so arrangedthat they have abolished the registry of inspection and dismissed the officials and clerks. For they have divided all thepeopleinto companies of ten, appointing one of the F. Nomadsof SouthPersia (London, 1961), pp. 51) Barth, 14-6. Sz) Polo, op. cit., p. 67.

286

J. M. SMITH
ten to be the commander of the nine others; while from among each ten commanders one has been given the title of 'commander of the hundred', all the hundredhaving been placed under his command.And so it is with eachthousand men and so also with each ten thousand... 53) [emphases added]

Here we see the Mongols as a people in arms, a regimented nomadic society, with all of their men counted, organized and used for war. Although the Mongols best exemplify this capacity of nomadic society for total mobilization, there are illustrations of it available from other periods as well, even going back'to the very beginnings of Inner Asian history. Herodotus begins his description of the Scythianswith an anecdote concerning a legendary Scythiancampaign into the Middle East 54). The Scythianmen went away to war, leaving their wives behind, and stayed away in their conquests for twentyeight years. Meanwhile their wives interbred with their slaves, which made for trouble when the Scythian husbands finally returned. The story is fabulous, but it had its roots in the real capacity of nomads to send all of their men to war. The story would not have survived the skepticism of either Herodotus or his audience except for this essential truth. But it survived not only among the Greeks, with their curiosity about, and knowledge of Scythian affairs, but remained persistently in circulation in the region as a valid illustration of the peculiaritiesof nomadic life: the same story appears in the fifteenth century Ottoman Siltfiqndme, told about the Golden Horde 55). The Mongols also counted and organized their sedentary subjects in the same way as they did the nomads in their empire, and for the same reason: they plannedto use all the adultmales in war, if necessary. Ordinarily,because the sedentarypeoples produced fewer cavalrymen, and thus a poorer sort of army,the Mongols only drew upon a fraction of this manpower, normally a tenth, and even that perhapsonly when a campaignwas undertaken.But on occasion, and in particularduring the period of the conquests, the entire sedentarymanpowerwas used:
53) Juwaini, op. cit., trans., I, pp. 30-31; text, I, pp. 22-23. 54) Herodotus, Persian Wars, IV: 1-4. 55) Ms. in the Topkapi Library,Istanbul, Hazine No. 1612, fols. 133-IS 2.

MONGOL MANPOWER

287

The Mongols then advanced on Khojend. When they arrived before the town, the citizens took refuge in the citadel .... When the Mongol armyarrived they found it impossible to capture the place immediately since it could be reachedneither by bowshot nor by mangonel. They therefore drove the young men of Khojend thither in a forced levy (hashar)and also fetched reinforcements from Otrar, Bokhara, Samarqandand the other towns and villages, so that fifty thousand levies and twenty thousand Mongols were assembledin that place. These were all formed into detachments of tens and hundreds. Over every ten detachmentsof ten of the Taziks there was set a Mongol officer.. .56)

And also:
When the town [of Bukhdri] and the citadel had been purged of rebels and the walls and outworks levelled with the dust, all the inhabitants of the town, men and women, ugly and beautiful, were driven out on to the field of the musalla.Chingiz-Khan spared their lives; but the youths and full-grown men thatwere fit for such service were pressed into a levy (hashar) for the attack on Samarqandand Dabusiya .... 7)

The Mongol enumerationsof troops raisedfrom the sedentarypeoples are thus also censuses of adult males, so that the lists of Mongol tiimens of whatever sort, nomad or sedentary,are at once counts of soldiers, estimates of manpower and military potential, and indicators of the size of the general population.

III
The Mongol population of the Middle East around I26o would have been about 85b,000 persons, if there were seventeen tlmens in Hiilegii's command (a figure that I prefer for reasons that will appear below). And since Hiilegti led one-fifth of the Mongol forces (and people), then the nomadic population of Inner Asia, all of it included in the
Mongol empire by this time, would have been about 4,250,000. Two

fifths of these would have been locatedin Mdngke'sOuter and Qubilai's Inner Mongolia, 850,000 people in each region, 1.7 million in Mongolia as a whole. One fifth, 85o,ooo people, would have been found in the Chaghataidrealm of Transoxiana,Semirechiyeand parts of Jungaria and the Tarim Basin. One fifth was in the J6chids' domainsin northern
56) Juwaini, op. cit., trans., I, p. 92; text, I, p. 71. 57) Ibid., trans., I, p. io7; text, I, p. 83.

288

J. M. SMITH

CentralAsia and the North Caucasianand South Russian steppe, and the remainingfifth was in the Middle East with Htilegii. These numbers are within plausible population limits for these regions, and well within their capacitiesfor the support of nomadism58). The sedentary populations of the western regions of the Mongol empire can also be estimated from data of this sort. For Russia, Verlocated between nadsky59) has reconstructeda list of forty-threetfimens Nizhni-Novgorod in the east and Galicia in the west, and north of the steppe zone occupied by the Mongols (Tatars) and the tiimensof the "regular" Mongol-Turkish nomad forces. The Grand Duchy of fifteento seventeen Vladimir tlimens, comprised according (respectively)

to the RogozhskyChronicle andthe Chronograph of 15 I2. The Grand Duchy of Nizhni-Novgorodcounted as five tlmens, and Tver, says
are listed in Vernadsky,could have been no smaller.The other tiimens a and a Crimean Polish letter. these documents are of yarliq Although the sixteenth century, they refer (explicitly in the yarliq) to, and incorporate information from the early fifteenth century, and thus they

of the lateperiodof the Golden probablystill recordthe arrangements under The list includes Kiev, Vladimir-in-Volynia, Horde, Toqtamish. Lutsk, Sokal,Podolia,Kamenets, Kursk,"EgolBraslav,Chernigov, and Pronsk-fifteen more Smolensk, Liubitsk, Polotsk, Riazan, day", To these Vernadsky would add threemore in Galicia(Galich, ftimens.
58) The Kazakhs in the Steppe Kray of Tsarist Russia's empire were some 1.8 to i . 9 millions in 1897, and most of them were nomads. This region, which was the equivalent of only the easternmost part of the Jdchids' domain, could thus have harbored the whole J6chid Mongol force of 850,000 people. See L. Krader,

Peoples of Central Asia (Bloomington and The Hague, 1963), pp. I8o, 199.

The Chaghataidrealm was the approximateequivalent of the Tsarist Turkestan Kray, including Semirechiye, which had perhaps i.5 million nomads in 1897, together with Chinese Turkestan, which had about half a million largely nomadic
Kazakhs and Kirghiz in the early 195o's (it contained some Mongols too, whom Asia: A Century of Russian Rule (N.Y. and London, 1967), p. Io4; G. Moseley, A Sino-Soviet Cultural Frontier: The Ili KaZakh Autonomous Chou (Cambridge [Mass.], 1966), pp. 17, z i; and Chang Chih-i, The Party and the National Question in China,

I have not been able to count). See Krader,op. cit., p. 199; E. Allworth (ed.), Central

G. Moseley trans. (Cambridge [Mass.], and London, 1966), p. i61.

59) Vernadsky, op. cit., pp. 217-219.

MONGOL MANPOWER

289

Lvov and Sanok), which had been held by the Mongols until 1349. These forty-threeRussian timensenrolled the adult male population of Russia: some430,000 men. Of them, the Mongols probablyordinarily employed about one-tenth in the armed forces; they may well have in been those Russians in China who were consolidated into a taimen 1330 60). Ten thousand Russians would have been close to one-fifth of the Russian force normally maintained (hypothetically) by the
Mongols, 43,000 men; they were perhaps sent to Qubilai as part of

the fifth portion underM6ngke's redistribution, althoughas mentioned, I have seen no mention of Russian contingents in the other regions of the empire, nor of substantial Chinese forces outside of the East. These forty-threeRussiantimenswould also imply a generalpopulation
in Mongol Russia of 2,1 5o,ooo persons.

There is a pleasing coincidence between this information and that of certainother sources that deal with Russia under the Golden Horde. For the Jdchids' realmas for the Ilkhans'there is a conventionallarge, round figure for military strength: 6oo,ooo men. The figure is found in Rashiduddin (although in an unconvincing story) 61), and in the
authoritative work of the Armenian Haython 62). Marco Polo has

another figure, or figures: three hundred to three hundred and fifty thousands63); he was uncertain of the exact strength of the Golden Horde, but wanted to indicate that it was about equal to that of the Mongol in the Middle East. The figure of six hundred thousands accords very well with Vernadsky'sforty-three Russian tiimenswhen these are taken in conjunction with the Mongol forces implied in the observations of the fourteenth centurytravellerIbn Bataita. Ibn Batpita visited the royal camp of the Golden Horde during the reign of Ozbek,
ca. 1332, and was told that there were seventeen commanders of

timensthen present at the camp for the festival at the end of Ramadan,
and that the army of the Horde was even larger than the 170,000 men
60o) Ibid., pp. 87-88. 63) Polo, op. cit., pp. 310-31.

p. 128. 6I) Rashiduddin/Boyle, 62) Haython,op.cit., p. 215.

290

J. M.

SMITH

that they led 64). These seventeen commanders would be the seventeen Muslim Mongol generals of the seventeen "regular"Mongol-Turkish nomad tinens assigned to the Golden Horde in accordancewith the reapportionmentof forces by Mdngke Qaan. This force, by the terms of the reapportionment,would have matched that of the Ilkhans in the Middle East (and those of the other three regions), so we may believe that Htilegii's army numbered seventeen tifmens also. Marco Polo was right, although imprecise, about the approximatelyequal strength of the two powers-in Mongol "regulars".The other forces alluded to in Ibn Batdita's account in addition to the seventeen tfimens whose commanderswere present, would have been the Russian troops, non-nomadicand probablyonly occasionallymustered,and commanded by Christianprinces; one would not expect them to be on hand in
the steppe for an Islamic holiday. The 170,000 nomad troops and the 430,000 Russians make up ,the required six hundred thousand men.

The conventional count of the forces of ChaghataidCentral Asia (which excluded most of Kazakhstan,but included parts of Jungaria and the Tarim Basin) was four hundred thousand troops, implying a total populationof two millions. This numberis given by Haython65).
men, or seventeen tiimens, as elsewhere, would have been "Mongols", leaving 230,000 men and I, 50,000 persons in the non170,000

"Mongol", probably largely sedentarypopulation. We do not have a detailedenumerationof the sedentaryunits such as we have for Russia (and, as we shall see, for the Middle East), although Ibn 'Arabshdh, in Timiir's time, mentions seven tiimensof Samarqandand environs, and its districts66). and nine tfimensof Andekan/Feraghdn The non-Mongol population of Iran in the Mongol period is more clearlyand fully exhibited. In the time of Hiilegti, the Mongols directly
Ibn Batata, Travels, H. A. R. Gibb trans., II (Cambridge, 1962), pp. 492-493. Haython, op. cit., p. 214. Ibn cArabshah,Tamerlane, J. H. Sanders trans. (London, 1936), p. 17. Emel in "Resimli bir Han Silsilenamesi", Islim Tetkikleri Enstittisi Dergisi, V as saying that (i973), p. 176, quotes the Tdrikh-iRashidiof Mirza Haydar Dughlat the Chaghataid Tughluk Timur (1347-1363) converted to Islam along with his Mongol armyof 6o,ooo men. 64) 65) 66) Esin,

MONGOL MANPOWER

291

controlled Azerbdyjdn,'Irdq-i 'Ajem and Khurdsin in Iran. These regions, or their modern equivalents, contained about half the population of Iran in 1956 67). If this proportion obtained in earliertimes, the same regions might have been expected ordinarilyto hold at least some two and a half to three million people; the population of Iran ca. i8oo, after two decades of Zand and Qajdrcomings and going, is estimatedat around five to six millions68). Hiilegti's agents, however, in those regions: one in Azerblyjdn,three could count only five tiimens in 'Irdq-i 'Ajem (Qumm, Kdshan and Isfahdn), and one in Khurisdn
(Nishipur-Tais). These five units, or 50,000 men, imply a population of only 250,000 people where, even by the undemanding early nine-

teenth centurystandard,there should have been two and a half millions. This is a very small figure-not ridiculously, but pitifully small. It shows, not mistaken arithmetic nor an incompetent count, since Persian eagerness to evade enumerationwas probably well balanced by Mongol willingness to use drasticmethods and plenty of manpower in taking the census, but rather the impact of the Mongols' total conscription and total warfare upon a sedentary society. The consequences of the Mongol invasions, always known to have been catastrophic,areherequantified. ManyMiddleEasternchroniclersproduced exaggeratedestimatesof the death-tollof the onslaught69); the Mongols actually counted the survivors, potential tax-payersand soldiers, and found only about one-tenth of what we might consider the normal minimum population. The Mongols had dispersed or destroyed the rest, some two millions and more. These losses are immense, but not disproportionate. Convergent evidence is supplied by a comparison of late Sassanian(early seventh century A.D.) and Mongol revenues obtained from the regions held by the Mongols. The Sassanianshad gathered the equivalent of 139
67) National and Province Statistics of the First Census of Iran: November 1956, I (Tehran, 1961). 68) C. Issawi, The Economic History of Iran, 18oo-z914 (Chicago, 1971), p. 20, has the population of Iran ca. 18oo as around five or six millions. 69) Summarized by I. P. Petrushevsky in the Cambridge History of Iran, V (Cambridge, 1968), pp. 484-488.

292

J. M. SMITH

million of the Mongols' silver "dinars",whereas the Mongols themselves, ca. I295, received only 17 millions, or only 12 % /oof the Sassanians'income. Thus both revenue and population figures suggest a
decline from "normal" of about 90 % 70).

The recovery of the population, at least in Azerbdyjdnand 'Iraq-i 'Ajem, is as phenomenal as its earlier losses. Around 1335, Qazwini in Iran that were not part of the Mongol enumeratedtwenty-five tiimens regularforces. Qazwini made specialmention of districtsthat supported Mongols with iqtd' (five Mongol tlimens in the Pishkin district of Azerbayjdn)and did not include these tiimensin the twenty-five 71). Nine of the twenty-five were in Azerbayjan,nine in 'Irdq-i'Ajem, and seven in Mdzandarin 72). Unfortunately,the military establishmentof Khurdsan,which was administrativelyautonomous, was not counted
by Qazwini. But the manpower and population of 'Iraq-i 'Ajem, which had been 30,000 men,and 150,000 people in ca. i26o, had tripled in about two generations, to 90,000 men and 450,000 people. That of
Mustawfi Qazwini in TheGeographi70) The revenue dataare given by ed. and trans., I, text (Leyden and cal Part of the Nuzhat al-Qulfb, G. Le Strange .Hamdullah London, 1915), P. 27; and II, trans. (Leyden and London, 1919), pp. 33 and n. i. The Sassaniansobtained 420 million mithqils of silver, and the Mongols, 17 million dindr-i rd'ij, or silver dinars. The mithqil used to count the Sassanians'revenue by Ibn KhurdHdbih,Qazwini's source, was 4-25 grams; see G. C. Miles, "On the Vatieties and Accuracy of Eighth Century Arab Coin Weights", Eretz-Israel, VII (1963). The Mongols' silver dinar weighed I2.96 grams; see J. M. Smith, Jr., "The and Social History of the Silver Currency of Mongol Iran", Journalof the Economic Orient,XII (I969). Thus the Sassaniansobtained 1,785 million grams of silver, and the Mongols only 220 millions. The population data are not as firm as those of the revenue. I have used the low to guess at the population of Iran estimates of the Iranian population of ca. 800oo in the period just before the Mongols' arrival. The estimate may be too low, since the period was one of a certain efflorescencein Seljuq Rum, CaliphalMesopotamia, KhwarezmianCentral Asia, and in Georgia and Armenia. For the late Sassanian period the estimate may be much too low. R. M. Adams, in "Agricultureand Urban Life in Early Southwestern Iran", Science, 136 # 3511 (13 April 1962) and TheLand that the population of 'Iraq-i 'Arab BehindBaghdad 15, argues (Chicago, 1965), p. I (in part) and Khaizistdnin Sassaniantimes was a great as, or greater than it is now. If the population of SassanianIran was similarly large, then its decline by Mongol times would have been on the order, not of 90 %, but 99 %. 71) Qazwini, op. cit., text, pp. 82-83; trans., p. 85. 72) Ibid., text, pp. 47, 75, 159; trans., pp. 54, 78, 156.

MONGOL MANPOWER

293

Azerbdyjin, rising from Io,ooo men and 50,000 people to 90,000 and 450,000, had grown nine-fold! Population growth was paralleled, and

is to a degree confirmedby the growth in the revenues of the Persian Mongol government. Qazwini recorded 17 million silver dinars in
the revenue of ca. two decades 74).
1295,

rising to

21

million in 1304

73).

By 1312 the

revenue had reached 30 million dinars, having nearly doubled in only Such a growth rate is hard to march in ordinaryexperience. Iran's population perhaps doubled between 18oo and 1914, and perhaps nearly doubled again by 195675). Egypt, with the most rapid growth of population in the modern Middle East, approximatelytrebled its numbers in the 73 years between 1883 and 195676). It may seem difficult to believe that medieval 'Iraq-i 'Ajem could rival modern growing three times faster still. Egypt, to say nothing of Azerbdyjdn Nevertheless, such growth can be explained. The Mongols under Hiilegii established a huge nomadic presence in Azerbdyjdn,with six tiimens or 300,000 persons, and 'Iraq-i 'Ajem was chronicallyexposed to the adjacentKurd and Lur nomads. These may well have replenished the populations of these districts. Barth'sstudy of the Basserisuggests that a very high rateof populationgrowth can be found among nomads:
The figures on present fertility seem consistent with those of the previous generation; and in the period 1908-38 in which that generation was born, none of the effects of modern medicine could yet have been felt, even indirectly, in the nomad camps of Fars. One is forced to assume that a consistently high rate of growth has been a characteristic of the tribal population in previous times as well as today. The evidence from the living generations in the Basseri

per tent, and 7.2 personsper adultsiblinggroupcamps today-4.25 children i.e. a treblingof the suggesta net growthfactorof at least 3 per generation, nomadpopulationevery 30-40 years. This generalpictureis, furthermore, not unique for the Basseri;superficial with neighboringArab acquaintance natural and Qashqai suggestcomparable growthrates"7).
73) Qazwini, op. cit., text, p. 27; trans., p. 33. 74) D'Ohsson, op. cit., IV, p. 54375) Issawi, op. cit., p. 20 and n. 2. 76) Ibid. and C. Issawi, The EconomicHistory of the Middle East, 1800oo-14 (Chicago and London, 1966), p. 373. 77) Barth, op. cit., pp. ii-16.

294

J. M. SMITH

Other cases show similar results. Joseph Birdsell, investigating Australianpopulation questions, has provided evidence from studies of isolated Tristan da Cunha, Pitcairn and Bass Strait islander populations, showing that, afterimplantationin these uninhabitedterritories, they have approximatelydoubled in each generation,despite their very simple economic and social circumstances,until the limit of the local resourceshas been reached. He presents also the case of an Australian aborigine family of one man and two women, who multiplied to 28 in thirtyyears,almost triplingtheir numbersin each of two generations, while "wandering from place to place in search of food, and living principallyupon black scrub kangaroos, which they sneak upon and spear", and obtaining water entirely from the roots of local plants 78). By comparisonwith growth rates of this sort, the suggested increase in the Mongol population looks rathermodest. In two generations of
35 years each, the original 30o,000 Mongols could easily have produced a surplus of 400,000 persons that would make up the numbers in Azerbayj~n from 5o,ooo to 450,000.

Since the Mongols had doubtless calculatedthe nomadic capacity of fairly closely before assigning the six timensto it, increases Azerbdyjdn in the nomadic population beyond the original 300,000 would have begun to produce imbalancesin the pastoral ecology, and would have had to be removed from the pastoral sector, probably by a process of sedentarizationakin to that described for the Basseri so clearly by of Mongols Barth. We can probably see parts of it. The five tf#mens mentioned by Qazwini as being supported by iqta' might be some of these over-numerousnomads, now maintainedby the sedentarysector. Others were probablyunable to avoid sedentarization,and were heard from when the government attempted to collect the taxes (qaldn) levied upon the non-Mongol sedentarypopulation:
Repeatedly, he (i.e., Ghizin) reprimanded the amirs, the judges (yarghzchis), and the (saying) that every time people come with complaints against wazirs and the mutasarrifs, they accept their words with haste; he indicated the hdkims, 78) J. B. Birdsell, "Some Population Problems Involving Pleistocene Man", Cold Spring Harbor Symposia on Quantitative Biology,XXII, Population Studies: Animal Ecology and Demography (1957), pp. 47-69.

MONGOL MANPOWER

295

that it was possible that those people had not given qaldnbefore then and had placed their burden on others, that the hdkim brought them into the qaldn, and that such people naturallycomplained79).

Because the Mongols had demolished the previous population of Azerbdyjin,therewould have been few obstaclesto this sedentarization, and even advantages in it. SedentarizingMongols would not have found themselves in the straits common to most failed nomads, who enter an establishedand often alien society and economy at the bottom, as agriculturallaborers, without land or other assets, and who suffer there an unusually high rate of attrition. The Mongols could avail themselves of the lands emptied by their earlier campaigning, and from the time of Ghdzin could do so with the encouragementof the Mongol government, which was trying to redevelop agriculture. In Azerbdyjin, moreover, they would have benefitted both from the sponsorship of Nature, which enables much of the region to be dryfarmed, and from the clement climate of the nearby Mongol government, which, as seen from the taxpayers'complaints and the grants of iqtd', tried to support its constituents. Conditions thus favored a rapid replenishmentof the sedentarypopulation. The bulk of this new, ex-nomad population would have been Turkish. The Mongol regular forces in the Middle East and Russia contained substantialnumbers of Turks, so that the lingua franca, and soon the language itself in these forces was Turkish. The Mongol court in Iran used Turkish by the time of Oljeitii. The evidence that shows the ruin of an older population and its replacementfrom among the conquerors thus also helps explain the transformationof Azerbayjdn'spopulation from Iranian to Turkish. This process is usually representedas startingin the eleventh century,with the Seljukinvasion. But the conditions that would have favored Turkification,almost ideal in the Mongol period, do not seem so ripe in Seljuktimes. The numbers of Turks coming into the Middle East with the Seljuks do not seem very large, perhaps tens of thousands rather than hundreds of thouTributary Practices of the Thirteenth Century", Harvard Journal of Asiatic Studies, 19 (1956), P. 333 and n. 61.

79) Rashiduddin/Jahn, p. I8o. Translation by F. Schurmann in "Mongolian

296

J. M. SMITH

sandsso) (as in the Mongol period), and not all of them went to Amerbiyjin. Those who did go there probably could not establish themselves as comfortably and productively as did the Mongols later, and probably could not clear the way for advantageoussedentarizationin the drastic Mongol manner. The sedentarysociety of Azerbdyjdnand the adjacent regions of Eastern Anatolia and Transcaucasiasurvived the arrival of the Turks, making arrangementswith the Seljuks81) and paying taxes (and very substantialsums by Mongol standards!) 82), or fighting persistently for independence8s). Because of this the pastoral potential of the Mughin steppe and the Aras and Kura valley lowlands, and the Qaribigh, Aldtdgh and Savildn highlands probably could not be fully realized, and the Turkish nomads had to rely more on predation in Ruim than on pastoralismin Iran. Their population surplus was probably too small, and their sedentaryopportunities too limited for them to effect much Turkificationin Azerbdyj5n.
Turkey(London, 1968), p. 33; and A. K. S. Lambton 80) C. Cahen,Pre-Ottoman znd ed. in "Iran", Encyclopaedia Islam, of the local dynasties of Azerbdyjin, Rawwidids of of the relations of details For 81) and of Shirwdnshdhs,with the Seljuks, see C. Bosworth Tabriz, Shaddddids Ganja, in the Cambridge Iran, V, pp. 44, 62, 94-95 and passim. History of trans., p. 78, claims that the revenue of Azettext, cit., 75; p. 82) Qazwini, op. bayjdnin the Seljukperiod was the equivalent of nearly 20omillion of the Mongol silver dinars. Since he calculated that one Seljuk gold dinar (which he doubtless dinars equated with the current mithq l of 4-32 gm.) was worth 2-1/3 Mongol silver a revenue be to he seems text figure Seljuk trans., reporting 27; 34), p. p. (ibid., lion silver dinars from Azerbdyjdn;of these, some 8.7 million came from the Tdbriz and z.z millions from land taxes, etc.: ibid., section 3, passim. tamghbd, 83) The Turkomans made headway into Armenia and Georgia only slowly and only with Seljuk support before 1071. During the reign of Malikshah they finally seem to have had full access to the pastures of these regions. Subsequently, the Georgians made a graduate reconquest, which, by the late twelfth century, had been extended beyond the frontiers of Georgia to Lake Van, Kars, and even Ardabil, thus threatening and perhaps controlling the summer pastures of the Azerbayjin Hellenismin Asia Minorand the region. See S. Vryonis, Jr., The Declineof Medieval the the Eleventh Processof IslamiZation Through FifteenthCentury(Berkeley and from
Los Angeles, I971), pp. 283-285; and Bosworth, op. cit., p. 179. of about 8.5 million Seljuk gold dinars. Ca. 133 5 the Mongols derived some 10.9 mil-

C. Cahen, in "L'Iran du Nord-Ouest face a l'expansion Seldjukide", Milanges Henri Massi (Tehran, 1963), pp. 68-69, gives information on resistance to the Seljuks by the Muslims of Khoy.

MONGOL MANPOWER

297

IV
In conclusion, let us note some further implications of the results arrived at above. We have been observing, first of all, the operation of a very simple and effective Mongol administrativemethod in this system of counting and organizing troops. The decimal organization of the army enabled Mongol commandersto calculate their military closely, since they knew the number potentialand logistic requirements of their men and could assume that each soldier had a certainnumber of mounts and a supportingfamily and herds of a certainaverage size. The rules of thumb here were probably ten horses, a five-person family, and ioo animals (in "sheep-equivalent" units) for each soldier. was basic herd the establishingtax liability, and this, or its ioo sheep in other was animals, probably required to maintain the equivalent desired level of military effectiveness84). A tiimencommanderwould thus know not only that he had ten thousand men to fight with, but that he had to administerfifty thousand people and find pasture for the equivalent of one million sheep. The system was also convenient because the exhaustive enumerationsof adult males, and their organization into decimal units, taken together with these conventional assumptionsabout their supporting families and herds, meant that the Mongol rulersknew the numbers,not just of their armies,but of their people, and could estimate their needs and assign them pastures and other assets accordingly. Counting the troops in effect accounted for everyone, and taking care of the troops resulted in caring for the whole people. And the calculations were simple. The Mongols clearly thought very highly of their system of decimal organizationof the army, since they impressed upon nearly every observer their sense of its importance, and caused it to be specially noted in most of our sources85). At first glance it seems only a sensible, scarcely remarkableway of organizing troops. But now we can see what else it accomplished,
85) Juwaini, op. cit., trans., I, p. 31 ; text, I, p. 23. Bar Hebraeus, op. cit., p. 354. Polo, op. cit., p. 69. Plano Carpini, in Dawson, op. cit., pp. 32-33. 19

84) J. M. Smith, Jr., "Taxation", pp. 68-69. SecretHistory, section 279.

298

J. M. SMITH

how easily it worked, and why the Mongols were so pleased with it. The whole manpower, and implicitly the whole population, human and animal, was comprehendedin these tens (or at worst, fives) and their multiples, and the Mongols could manage anything from a squadronto a nation without needing accountantsto do their figuring
-without even needing to be literate 86).

We see here also an index of the attachment of the Mongols to Chingis Khan and his enterprise. The Mongols could all be counted as soldiers because they could be counted upon to devote themselves wholly to Chingis' wars. This dedication, which is seen also in the Mongols' willingness to bear an exigent and incessant taxation, contrasts sharplywith the normalnomadic (not to say human)reservations about conscription. Most governments have only been able to use fractions of the nomads' manpower. The Japanese tried to recruit one-sixth of the Inner Mongols' men during World War II, but the
Mongols held out for a mere one-tenth
87).

Even the Ottomans, with

an attractiveenterpriseof their own, seem only to have obtained onesixth to one-fifth of the men in their nomads' camps 88). Full, or even heavy conscription, can only be applied appropriately in nomadic societies under special conditions. Ordinarily, nomadic peoples live in a state of greatinsecurity,of chronicinter-tribalrivalry,and, perhaps more important, of suspicion and fear even between camps of the same tribe and clan. Under such circumstances,most of the manpower of the camps, clans and tribes is immobilized by the requirementsof local defense. This manpower can only be drawn upon effectively
was prefigured by, andperhapsderivedby continuing 86) Mongolorganization from thatof the Hsiung-nu,described tradition of the by Ssu-maCh'ienin Records The B. Watson Historian Grand trans., of China, II (N.Y., 1961), pp.I 55-I192. Hsiungandtenthoua thousand, in units of ten, a hundred, nu has 3oo00,000troopsarrayed Shuo,taughtthem"how to sand(pp. 163-164). A Chineserenegade,Chung-hsing makean itemizedaccountingof the numberof personsand domesticanimalsin the country" (p. 170). And in peacetimetheirmen "had nothing to do" (p. 171)! from SechinJagchidof NationalChengchiUnicommunication Personal 87) Young University. versity,Taipei,and Brigham
88) H. A. R. Gibb and H. Bowen, Islamic Societyand the West I:1 (London,
5 5.
1950), P.

MONGOL MANPOWER

299

when unity of leadershipand purpose enables the suspension of these doubts and defenses. When it does, it gives the united community a great advantageover its disunitedneighbors. But unificationhas been achieved only rarely,and on a large scale only once-by Chingis Khan. Mongol successes were founded upon the military differentialthat favored nomadic societies over sedentary,not only in the raising and riding of horses, but in the capacity to bear conscription and pay taxes. Chingis Khan had only to discern these capacities, learn how to administer the hordes they provided-and interest the nomads, all the nomads of Inner Asia, in exercising them. He had only to transmuteconscription, like taxation, from liability to opportunity.