The Toluid Civil War was a war of succession fought between Kublai Khan and his younger brother, Ariq Böke, from 1260 to 1264. Möngke Khan died in 1259 with no declared successor, precipitating infighting between members of the Tolui family line for the title of Great Khan that escalated to a civil war. The Toluid Civil War, and the wars that followed it (such as the Berke–Hulagu war and the Kaidu–Kublai war), weakened the authority of the Great Khan over the Mongol Empire and split the empire into autonomous khanates.
The dominions of Qaidu and the Chaghadayids
The conflict in Central Asia between the supporters of Du’a and of Chapar in 705/1305–6 was a brutal affair. Usually the details in our sources relate to the plundering and slaughter of nomadic armies – actions that in themselves, of course, had grave consequences for the economic prosperity of a region. But occasionally we are given a glimpse of unbridled attacks upon a sedentary population, as when, following a victory over Chapar’s brother Shāh Oghul in 705/1305–6, Du’a’s supporters, headed by the Chaghadayid prince Yasa’ur, laid waste the districts of Talas, Yangī, Kenjek and Chigil, tormenting the inhabitants, driving off livestock and setting light to anything they could not take with them. Waṣṣāf wrote that the whole of Turkestan and Transoxiana had been rendered desolate by the warring of the princes, the concentration and movement of troops, and the destruction of dwellings. Merchants ceased to travel, while headmen and cultivators were ground down by the requisitioning of provisions and levies of grain, and many folk had emigrated.
With the outbreak of conflict between Esen Buqa and Köpek, on the one hand, and their kinsman Yasa’ur, on the other, Transoxiana was once again subjected to a harrying. In 713–14/1314, prior to Yasa’ur’s submission to the Ilkhan Öljeitü, his troops sacked Samarqand, Sāghraj, Kish, Nakhshab, Kūfīn and other towns; Bukhārā and Khujand avoided the same fate only through the good offices of Shaykh Badr al-Dīn Mandānī. The notables from the stricken cities contrived to escape to Khwārazm, but Yasa’ur’s forces drove a vast number of the populace across the Oxus (rendering it hard to share Qāshānī’s confidence that the devastation occurred without the prince’s knowledge or consent). In Khurāsān their sufferings were considerable. Continued hostilities with Köpek obliged Yasa’ur to move them southwards from an initial camping-ground in the Murghāb region into the territory of Herat, with the result that as many as 100,000 souls allegedly perished from the cold or from starvation. Unsurprisingly, Yasa’ur has been characterized as a champion of nomadic interests and one impervious to the needs of the sedentary population.
Khwārazm also suffered periodic devastation. At the time of Alughu’s assault on Jochid Transoxiana in the early 1260s, some of his forces carried the war into Khwārazm. And when Abagha’s army invaded Transoxiana in 671/1273, a number of his commanders were sent against Khwārazm, conducting a general massacre (qatl-i ‛āmm) in Gurgānj (Ürgench), Khiva and Qarāqash. In Jumādā I 715/August 1315, Baba Oghul (a descendant of Chinggis Khan’s brother Jochi Qasar), who had earlier abandoned Chapar for the Jochids, deserted them in turn for the Ilkhan Öljeitü. Having repelled an attack by the governor of Khwārazm, Qutlugh Temür, he subjected the entire province to a campaign of devastation that is graphically described by Qāshānī, plundering several major towns, notably Hazārasp, Kāt and Khiva, committing numerous atrocities and carrying off 50,000 captives. Although he was relieved of his prisoners in a surprise attack by Yasa’ur, they arrived back home in a wretched state. Öljeitü responded to Özbeg’s incensed protests by executing Baba in the presence of the Jochid envoy, but – understandably, given that Baba had acted without his authority – offered no other form of reparation.
The impact of internecine conflict
Instances of devastation, then, during the numerous wars within the Mongols’ ranks, are not far to seek. To assess the magnitude of the damage is a different matter; nor is it easy to determine how far these conflicts prolonged the economic instability inflicted by the earlier campaigns of conquest. We are too often confronted by generalization – and suspect generalization at that. Rashīd al-Dīn tells us that on account of Qaidu’s ‘rebellion’ many Mongols and Tājīks had been killed and flourishing regions laid waste; he further claims that through the assistance given to Qaidu and Du’a by the renegade Nawrūz much damage was done to Khurāsān and innocent Muslims were killed. But as an Ilkhanid minister he had every reason to stress the adverse consequences of Qaidu’s activities (or, following the downfall of Nawrūz, of that noyan’s earlier disloyalty) for the Toluid cause. And how many times did a town have to be sacked, or an entire region devastated, before recovery became an impossibility? It is difficult to repress a nagging scepticism.
The numbers involved in the attacks by Mongol potentates on their neighbours and rivals are an inadequate gauge of destructive capacity. Often suspiciously large, they doubtless represent the total strength notionally available to a ruler (bearing in mind that all males over fourteen, at the very least, were warriors) rather than an army on a particular campaign. Thus one version of Marco Polo’s book has Hülegü and Berke do battle in 1262 with 300,000 men each. The figure recalls Rashīd al-Dīn’s epic narrative concerning ‛Ayn Jālūt, in which Kedbuqa defiantly tells his Mamlūk captors that Hülegü has 300,000 horsemen and could therefore afford to lose him, and al-‛Umarī’s statement that in its heyday the Ilkhanate could field thirty tümens. The same reservation applies to the figures that have come down to us for the Golden Horde armies. Insofar as they have any value, it can only be as a reliable indication that the Jochid ulus could draw on larger reserves of manpower than its neighbours. Al-‛Umarī indeed confirms that the Golden Horde’s forces greatly outnumbered those of the Chaghadayids, while Hayton of Gorighos attributes 400,000 and 600,000 men to Chapar and Toqto’a respectively. All these totals pale alongside the incredible claim by an informant of al-‛Umarī’s that when Toqto’a mustered one man in ten for a campaign against Esen Buqa, this produced an army of 250,000. Whether the Jochid totals include the auxiliary forces of client princes, we cannot know; Toqto’a was certainly accompanied on campaign by Rus´ soldiers.
We seem to have more modest, and hence more likely, figures for the campaigns by the Central Asian Mongols. Baraq is said to have entered Khurāsān with 150,000 men in 668/1270, and Du’a with 100,000 men in 695/1295; Sarban and his confederates moved on Ṭūs with 50,000 men seven years later; the force with which Köpek and his Chaghadayid colleagues invaded Khurāsān in 713/1313 is variously set at 40,000, 50,000 or 60,000 horse (although at one point Sayfī gives a more precise figure of 56,000); and on the eve of his downfall Chapar allegedly planned to face the Yuan army at the head of twenty tümens (doubtless including the troops of his supposed ally Du’a). Conceivably all the figures cited above were originally expressed in tümens. We should therefore remember that, as noted earlier, the tümen only notionally comprised 10,000 men and such a unit at any given time would have been reduced by campaigning.
Too much uncertainty attaches to the circumstances in which towns were deserted. In the early seventeenth century it was believed that Balāsāghūn had been covered by sand since soon after Chinggis Khan’s time. The fact that Ḥamd-Allāh Mustawfī has nothing to say about the town and describes the region’s inhabitants as predominantly nomads in his own day157 may mean that it was among the places cleared in advance of Hülegü’s passage. Yet Rashīd al-Dīn’s statement that in his time there were a great many villages in the region between the Ili and the Chu rivers, where Qaidu was buried, suggests that not far away agriculture was still flourishing after 1300. Supposedly eyewitness Muslim travellers, too, afford us scant help. Ibn Baṭṭūṭa describes Samarqand and Bukhārā as largely in ruins. Travelling through the Chaghadayid-Ilkhanid borderlands, he notices evidence of considerable destruction, but attributes it to Chinggis Khan’s forces over a century earlier. One wonders whether it resulted, in reality, from more recent strife. In any case, the fact that Ibn Baṭṭūṭa describes Balkh as in ruins when, according to the Timurid author Sharaf al-Dīn ‛Alī Yazdī, the city had been restored by the Chaghadayid khan Köpek, aggravates the doubts surrounding his true itinerary.
Ibn Faḍl-Allāh al-‛Umarī, drawing, in all likelihood, on information obtained in the 1330s, sketches another picture of desolation:
I was told by someone who passed through its [sc. Turkestan’s] countryside and travelled through its villages, ‘There remain of its settlements nothing but traces of their location and dilapidated ruins. You see in the distance a settlement with impressive buildings and verdant surroundings, and are cheered by the expectation of finding there congenial inhabitants. But when you arrive, you find the said buildings empty of people and inhabitants other than tent-dwellers and herdsmen. There is no sowing or tillage. Its greenery is that of pastures which God created. The vegetation there is of the steppe: no sower has sown it nor cultivator planted it.’
Impressions of this kind defy evaluation. Which parts of Turkestan did al-‛Umarī’s informant see? Was this wilderness a direct consequence of the conflicts of Chapar’s era, or of the warfare between Qaidu and Qubilai? Or did it date back several decades, to a time when agricultural land was converted to pasturage? The simple answer is that we cannot know.
Nomadic khans, their military and their sedentary subjects
The foregoing section has been concerned with the way in which Mongol princes treated the subjects and territories of their enemies. Let us now consider the policies they adopted towards their own, which we might reasonably expect to have been more restrained. There was undeniably a marked variation in social and economic conditions between the two Mongol polities centred on the major regions of sedentary culture, namely Yuan China and Ilkhanid Iran, and those dominated by a pastoralist nomad economy. Yet two qualifications should be made. In the first place, throughout the Mongol world Chinggisid rulers and their military continued to follow a nomadic lifestyle well into the fourteenth century, and it was not merely invading armies that inflicted devastation on the agricultural regions. The Ilkhans themselves still adhered to the customary seasonal movements, following, as Charles Melville has demonstrated, a rhythm of migration between summer and winter quarters right down to Öljeitü’s reign. And secondly, in many cases Mongol rulers whose territories were predominantly steppelands were by no means hostile towards urban culture.
The needs and aspirations of nomadic cavalry forces certainly conflicted with agrarian interests.164 In the spring of 1264, Arigh Böke’s troops, quartered in the Almaligh region, were obliged to feed their mounts with wheat instead of barley; in consequence, many of the townsfolk of Almaligh starved to death.165 During Mubārak Shāh’s brief reign over Chaghadai’s ulus, Rashīd al-Dīn tells us, the military continued to pillage and harass the populace, though the khan – a Muslim – reportedly prevented them from oppressing the peasants. His successor Baraq twice hatched the design of plundering Transoxiana, but was prevented by Qaidu on the first occasion and dissuaded by Mas‛ūd Beg on the second. He nevertheless caused a serious dearth by requisitioning wheat as well as barley to fatten up his horses in preparation for his invasion of Khurāsān.
In Iran the Ilkhans’ own forces, perhaps on the move to repulse an enemy, damaged sedentary land. Anecdotes in the Ṣafwat al-ṣafā of Tawakkulī Ibn Bazzāz, an account of the life of Shaykh Ṣafī’ al-Dīn Ardabīlī (d. 735/1334), testify to the damage done to agrarian districts and the peasantry by the passage of troops. It seems to have been important to Rashīd al-Dīn that on the long journey from Azerbaijan to Khurāsān to do battle with Baraq, Abagha had forbidden his men to damage a single stalk, and that during their return march they refrained from injuring a single creature; but Egyptian authors (not necessarily a dispassionate source either, in this context) insist that his cavalry grazed their mounts in cultivated fields. So too, it was the Ilkhan’s troops who on occasions allowed their zeal to carry them away when conducting a punitive campaign against unruly subjects, like the Türkmen groups in the Konya region of Anatolia in the early 1280s, since which time, says an anonymous author writing over eight decades later, it had remained ‘virtually uninhabited (hamchunīn kharāb)’. During his operations as governor of Khurāsān against the rebel Nawrūz, according to Rashīd al-Dīn, Ghazan forbade his soldiery to allow their livestock into fields or orchards, or consume the grain, or treat the peasants with violence, which (whether true or not) suggests that these abuses were common. His decree for the allocation of iqṭā‛s (land-grants) to the Mongol soldiery referred to their habit of damaging the provinces when setting out on campaign.
There was also periodic conflict between the Ilkhan’s sedentary subjects and nomadic groups introduced by the Mongols at the time of the conquest. This category includes the tamma forces that had been quartered within the regions since Hülegü’s era, or even earlier, and had virtually acquired a title to their pasturelands through long usage. By the early fourteenth century some of these provincial Mongol armies had been transmuted into artificial tribal formations that took their name from their original commander (much as the residue of the Jochid troops had regrouped in eastern Iran at an earlier date to form the Negüderis): the Jurma’īs of Fārs and Kirmān, for example, whose nucleus was the force led by a certain Jurma, or the Ūghānīs, initially a thousand on the frontier of Kirmān commanded by Ughan of the Jalayir. Although Jurma’īs, having intermingled with the local village population, were caught up and enslaved in the Negüderi raid of 677/1278–9, and although they and the Ūghānīs for some days put up a spirited defence against a Chaghadayid invasion in 700/1301 (above, pp. 197–8), friction also readily arose between these nomadic troops and the sedentary population. A local chronicler speaks of the iniquities that the Ūghānīs perpetrated against ‘the warm regions and the cold’ in the 1280s, and they and the Jurma’īs were still plaguing Kirmān in the post-Ilkhanid period, when the new Muzaffarid rulers were called on to check their depredations. A fifteenth-century poet preserves the laments of the people of Abīward, in Khurāsān, regarding their oppression by the Mongols of the Ja’ūn-i Qurbān.
Responsibility for damage to agricultural land in Iran also lies with outside pastoralist groups technically subject to the Ilkhans’ own authority. Waṣṣāf describes the Qara’unas/Negüderi bands who submitted to Abagha in Khurāsān in 677–8/1278–9 and entered his service as ‘like demons rather than humans and the most brazen of the Mongols’, and says that plundering was their customary activity. These Ilkhanid Qara’unas wrought havoc in the civil war between Tegüder Aḥmad and Arghun, when they devastated the Dāmghān region, and again in the upheavals towards the end of Arghun’s reign. A band led by Dānishmand Bahādur laid waste the Juwayn district in 689/1290. In 698/1299 another Qara’unas thousand, under a certain Buqa, abandoned its quarters near Ṭārum and headed east to rejoin the Negüderis, ravaging the borders of Yazd and Kirmān.
Further nomadic groups accompanied fugitive Chinggisids seeking fresh pasturelands and offering their services as auxiliaries. Following Chapar’s defeat by Du’a, several princes abandoned Central Asia. Chapar himself, as we saw, sought asylum in the Yuan dominions. Others took refuge in the territories of the Ilkhan Öljeitü. First was a group numbering 30,000 and headed by Chapar’s brother Sarban, Mingqan, a grandson of Arigh Böke, and Temür, of Jochi Qasar’s line, who entered Khurāsān together early in 706/the summer of 1306. Later, in the summer of 708/1308, the Chaghadayid Dhū l-Qarnayn too entered Khurāsān to seek fresh camping-grounds from the Ilkhan. Supervising this influx of successive nomadic groups, and their absorption, cannot have been easy. Since Sarban, Temür and Dhū l-Qarnayn all died shortly after crossing the Oxus, the province was possibly spared a problem of the scale that it would experience in the next decade. This was the responsibility of Yasa’ur, who, accompanied by various Ögödeyid princes, had been settled in Khurāsān by Öljeitü but who then fomented trouble during Abū Sa‛īd’s minority, trying to subject to himself local dynasts, including the maliks of Herat and Sīstān. In the course of these struggles, Khurāsān and Māzandarān suffered considerable devastation. Yasa’ur was overthrown in 720/1320 by his old enemy Köpek, to whom the ruler of Herat had appealed for assistance.
The second qualification that must be made to the time-honoured view of the Mongol states is that even those khans who have been cast as proponents of the nomadic lifestyle and customs made efforts to remedy the damage inflicted by the campaigns of conquest or by more recent warfare. The quriltai that Qaidu summoned in 1269 voiced a concern to preserve the flourishing towns of Transoxiana, under threat from the rapacious Baraq; it was agreed that the princes would live in the mountains and plains, avoiding the urban centres, and would not pasture their livestock on cultivated land. After Transoxiana had lain desolate for seven years as a result of the depredations of the sons of Baraq and of Alughu, Mas‛ūd Beg, on Qaidu’s orders, reassembled the scattered populations of Samarqand and Bukhārā and took steps to revive economic activity (c. 681/1282). Waṣṣāf, writing in 699/1299–1300 (and thus before the conflicts that began in 1305), heard that it was once more a thriving region. According to an early fifteenth-century author, Du’a (probably with Qaidu’s backing) restored a number of towns in Turkestan and Farghāna, notably Andijān, where the development of a flourishing commercial entrepôt is implicit in the reference to separate quarters for each of the various ethnic groups. Some decades later, as we saw above, Du’a’s son Köpek would restore Balkh. Old Tirmidh, reduced to a ruin by Chinggis Khan, had evidently been refortified by 716/1316, when Köpek’s forces were able to hold out there against Yasa’ur’s troops. Khwārazm, where the irrigation network had suffered during the Mongol attack of 1221, was again a prosperous agricultural region by the early fourteenth century. Sighnāq and Sawrān, on the lower Sīr-daryā river, revived when the region became the centre of the ulus of Orda around the same time. The Jochids are known to have founded many small towns as craft and supply centres.
Much of this princely solicitude was directed towards towns and trade. We know that in the Ilkhanate, at least, taxes on commerce and crafts yielded larger sums than did those on the land. The monarchs, their families and the upper echelons of the bureaucracy and the military were maintained by these taxes; the costs of administering the empire were met chiefly from agricultural revenues. Philip Remler argues that this explains the relative neglect of the agrarian sector prior to Ghazan’s reign and the fact that the pastoralists were allowed to live off the countryside. Such information is lacking for the other two western khanates; but here too the rulers often made their seasonal quarters in the vicinity of towns, which served as centres of revenue collection and craft production and in some cases as mints. Qaidu’s principal residence, as we saw, was near Talas. From 696/1296–7 the town of Ṣāqchī (Isaccea, in present-day Rumania), on the lower Danube, was an important centre in Noghai’s steppe domain; coins were minted here in his name, and it is described by the Mamlūk author Baybars al-Manṣūrī as one of his halting-places (manāzil). Some of these walled centres had even been constructed by Mongol qaghans or other princes: Emil, on the river of that name, by Ögödei, for instance. The Chaghadayid khan Köpek built Qarshī, a few miles from the town of Nakhshab in Transoxiana, though whether as Esen Buqa’s viceroy (when the Nakhshab-Kish region is known to have been his summer quarters) or during his own reign, we are not told. These were tent cities. Qarshī was still a royal residence over a decade later, Ibn Faḍl-Allāh al-‛Umarī tells us, confirming, however, that the khans otherwise eschewed living behind walls. The most celebrated example is the city-encampment of Sarai (Pers. sarāī, ‘palace’) that Batu had constructed by c. 1250 in the Volga delta, not far from the site of the one-time Khazar capital of Itil. At some subsequent date another city may have been built upstream, possibly by Özbeg, and known as Sarāī-yi jadīd (‘New Sarai’), though it has been argued that this was merely another name for Batu’s foundation. Even if far from implying the abandonment of the nomadic life, activities of this kind must have fostered (and even, at times, expressed) a certain sensitivity to the needs of town-dwellers.
On the other hand, the Ilkhans could no more afford to be unmindful of the exigencies of ruling over nomads than were their counterparts in the Pontic-Caspian steppes or in Central Asia. Rashīd al-Dīn makes Ghazan assure the Mongol military of his readiness to fleece the agriculturalists of his realm, should that prove the most profitable means of governance in the longer term:
I am not on the side of the Tāzīk [Tājīk, i.e. Persian] peasants (ra‘iyyat). If there is a purpose in pillaging them all, there is no one with more power to do this than I: let us rob them together …
But Ghazan goes on to advocate restraint. His opening words (no doubt inspired by Rashīd al-Dīn) are admittedly designed to ‘sell’ his reforms by pointing to a community of interests; but otherwise the sentiments are almost worthy of a Baraq. Yet the crucial distinction, for our purposes, is not a geographical or ecological one. It separates, rather, rulers who failed to rise above the short-term view, extracting from their sedentary subjects the maximum possible, from others who perceived their own long-term advantage in nurturing the same subjects’ prosperity and hence their potential as a source of revenue. This exercise in delayed gratification was more pressing, perhaps, for Mongol khans who embraced Islam – or at least for Muslim authors who wrote about them.
* * *
Just like the sources that describe the campaigns of Chinggis Khan’s armies against the Khwārazmshāh’s empire and lesser Muslim kingdoms, those for the inter-Mongol conflicts of the later thirteenth and early fourteenth centuries are prone to exaggerate and to avoid the specific; but there is one marked difference. Only rarely, as on the occasion of the Ilkhanid assault on Khwārazm in 671/1273 (the work, we might note, of troops from a regime often credited with a greater sympathy for sedentary culture), do they give the impression of the wholesale massacre of an urban population. In general the damage and its consequences were of the kind normally associated with medieval warfare. We have no way of knowing whether the harmful impact on agrarian or urban societies exceeded that of the late Saljuq and Khwarazmian periods. Nor is it possible to judge whether the influx of pastoral nomads created greater dislocation than that caused by, say, the Ghuzz during and for some decades after the collapse of Sanjar’s empire in the mid-to-late twelfth century. The most we can say with some degree of certainty is that these conflicts impeded or reversed the work of reconstruction which various Mongol rulers and their ministers had undertaken in the aftermath of the initial conquest.
