“The Impact of the Mongol invasions on Iran, Iraq and Central Asia; A Revaluation” by Professor Charles P. Melville, Cambridge University, U.K.

Seige Underway by Mongol Forces - from a 14th century manuscript.


“The Impact of the Mongol invasions on Iran, Iraq and Central Asia; A Revaluation” by Professor Charles P. Melville, Cambridge University, U.K.

Please note that this lecture paper is a work-in-progress and is not to be copied, quoted or excerpted from without expressed written permission from Professor Charles P. Melville, Cambridge University, United Kingdom.


“The Impact of the Mongol invasions on Iran, Iraq and Central Asia; A Revaluation” by Professor Charles P. Melville, Cambridge University, U.K.

The Indo-Mongolian Society Lecture at New York University,
March 12th, 1997.


This lecture paper titled “The Impact of the Mongol invasions on Iran, Iraq and Central Asia; A Revaluation”, was transcribed from a verbal presentation by Professor Charles P. Melville on March 12th, 1997 at New York University for the Indo-Mongolian Society of New York. Professor Melville addresses one of the most important issues in Mongolian history in this lecture, which is the consistent distortion and large-scale exaggeration by many Persian, Arab and Central Asian historians about the scale of destruction wrought by the Mongol campaigns and the Mongol rule over these dominions.

The oft-repeated view of massive destruction carried out by Chinggis Khan and his armies in several campaigns has helped form the standard view of Chinggis Khan marauding murderous hordes perpetuated through literature in most parts of the world. This extraordinary re-examination of the primary historical sources and period accounts about the Mongol military campaigns provides scholars and the lay public the opportunity to have a more balanced view of Chinggis Khan, the Mongols and this critically important aspect of world history.

Professor Melville’s methodical reexamination of historical sources from the Mongol Ilkhanid period and those prior to the Mongol conquests illuminates the closer-to-actual population densities of the Ilkhanid regions for a more precise examination of statistical assessments of regional population figures of areas in the path of the Mongol invasions.

Since most of the historical accounts about Chinggis Khan and the Mongol conquests were written by historians whose countries had been conquered by the Mongols, there has been a natural tendency for historical distortion and statistical exaggeration in their writings about the Mongol empire and its military campaigns. Professor Melville’s comparative analysis of the reports of local historians, travelers accounts, and contemporary sources helps provide much needed scrutiny of questionable statements in the writings of native historians who suddenly found themselves to be subjects of Mongol rule.



“The Impact of the Mongol invasions on Iran, Iraq and Central Asia; A Revaluation” by Professor Charles P. Melville, Cambridge University, U.K.



“I’ d like to dedicate this lecture to the memory of the great French scholar the late Jean Aubin, who died recently. Apart from the general debt that everyone working in this field owes to his amazing research over a period of forty years, in this particular presentation, I am building on arguments that I first heard him articulate at a conference in France in 1992 and which I believe remain unfortunately unpublished. Any residual traces of sarcasm and wit that may be detectable in this talk will readably be recognized as the hallmark of Aubin’s refreshingly sardonic style. Views have differed dramatically over the impact of the Mongol invasions on the Islamic world, the debate has ebbed and flowed like all historical arguments depending on the particular time and circumstances of the historians concerned.

As for E.G. Browne writing in Cambridge in the early years of this century, and as Bernard Lewis implies, maybe rather jealously, in a haven of shelter of civilization, “the Mongol assault was a catastrophe, which changed the face of the world and inflicted more suffering on the human race than any event in world history.”

Things didn’t seem quite so bad to Barthold, the first scholar to make an objective analysis of the invasions in his extraordinary doctoral thesis defended in 1900 and still by far the best work on the subject, that is “ Turkestan Down to the Mongol Invasion”, and particularly so after the First World War.

The work of revisionism was taken a stage further by the post-Second World War and post-Holocaust historian Bernard Lewis, who remarks that “ The immediate blows of the Mongols though thought no doubt trivial by modern standards were terrible and overwhelming, yet they were limited in extent and duration.” For Lewis, only in Iraq did the Mongol conquest leave permanent injuries, the decline of the elaborate irrigation works on which the prosperity, even the life of the country depended. But whereas in Persia there was partial recovery once the regime was firmly in control, in Iraq there was hardly any.

Even further down this path, when the interesting thesis of Pai-nan Rashid Wu in a dissertation that was published or done at the University of Utah in 1974 - unlikely to become another Barthold - whose work on the fall of Baghdad and the Mongol rule in Al-Iraq concluded that “Most or all the accusations against the Mongols are rendered dubious or without foundation. The Mongol invasion of Baghdad and the elimination of the Abbasid Caliphate created no more than ripples in a pool which soon returned to normal.” No respecter of persons, he interestingly concentrates particularly on the question of irrigation as we shall see. More mature in his judgment in the fin-de-siecle gloom of Thatcherite Britain when it was all too easy to imagine the horrors of the past. David Morgan concluded that Iraq became a neglected frontier province and for Persia the Mongol period was a disaster on a grand and unparalleled scale. While at the same time bringing a welcome breath of common sense into the analysis of the forces that the Mongols actually brought to bear. His views are largely shared by Professor Lambton in her recent book touching on the subject.

It may seem unlikely that there is anything more to be said about the Mongol invasions though excellent articles continue to be written on the subject, notably recently by Jirgen Paul. Certainly it is a vast subject which cannot be tackled in great detail this evening, despite the optimistic impression given by the title of my talk. To bring the topic down to size we shall not be concerned with the effects of Mongol rule as such, though some before and after comparisons will be useful. But more with the short term impact of the conquests themselves, as you know these occurred in two main waves, the first in Transoxania and Khurasan in the years 1219 to 1223, and the second through Iran and Iraq in 1256 and 1258 that’s thirty years later.

I should emphasize for the only time that much of southern Iran was not affected directly by the Mongol invasions, though this must be part of the argument in any overall assessment. The situation in the south is more a matter of Mongol rule, which is not itself claimed to have destroyed cities or decimated populations. Nor can we be concerned with the longer term impact of the Mongols on Persian history although this is of course an interesting subject. As you are all aware, Persian culture reached extraordinary heights under the Ilkhans and later Mongol rulers, but that is a different subject. As David Morgan with his customary wit put it, “We may justly have our doubts, over how impressed the Persian peasants - as they did their best to avoid the Mongol tax collectors - would have been by developments in miniature paintings.”

Finally the Mongols motives for the invasion and for their destructiveness also only feature very briefly in my argument. So to assess the short-term impact and the immediate casualties and destruction of the invasions we need some perspective that will allow us to compare the situation before and afterwards. If over a million people were killed in Herat for example, we need to be sure that there were a million people there beforehand. One of the main problems as this suggests is the question of numbers and how we can get around them.

Our sources are no more immune from the often unconscious influence of their own milieu then are modern historians, but unlike most of us unfortunately, they also made some effort to write literature consistent with their education. Also they were writing under some important constraints, nevertheless they speak with an impressively unanimous voice of a great trauma. Its not my intention or desire to minimize or belittle this trauma, nor to play down the terror that the Mongols created, its not an amusing story. Nevertheless, we are in the business of explanations and not emotions, and it is useful to attempt some more objective measure of our subject. Furthermore the 14th century Persian satirist Ubajdi Zakani managed to extract considerable humor from the situation, and I hope that only traces of levity in my own presentation will be seen as an attempt not to spoil your evening with too many mountains of corpses.

I realize so far this has been all talk and no action unlike the Mongols themselves. To keep within bounds I propose to examine some individual episodes rather than the whole course of the invasions to bring out the salient points of my approach. I will then try to draw some conclusions together from these and other cases. The slides are purely illustrative rather then crucial to the argument, though once or twice they do provide some compelling visual proofs. I regret that I have not yet managed to produce a satisfactory map as you’ll see from my first slide. (Slide shown) The idea of the map is just to show in the most general sense a sketch of Genghis Khan’s and other Mongol invasions. Transoxania, Iran, especially northern Iran, and Baghdad is here somewhere (pointing) There’s also a map that I’ve handed out. This shows Alexander the Great plucking up the people of Gog and Magog behind the mountains for the protection of the civilized world. In Christian and Islamic mythology their emergence is promised at the end of the world. This explains the Muslim rationalization I think of the origins and nature of the Mongol attack. The Mongols point of view also of course have an idea of breaking out from the mountains valleys from which they were encased to form a nation. These are just some pictures of Mongol warriors breaking out, looking somewhat fierce. This shows the main routes of the invasions progressing from Samarkand and Bukhara down to Herat, and Nishapur. As you know Genghis Khan’s invasion was launched against the territories of the Khwarizm Shah, the ruler of the area before their arrival, in retaliation for the murder of the Mongol Muslim trade mission at Utrar followed by the execution of his ambassadors who were sent to protest. In fact I am not going to refer very much to the conquest of Transoxania largely because very few figures are given and its the figures particularly that I’m interested in. Leaving his son Chaghatay to prosecute the siege of Utrar that was the scene of the massacre, Genghis Khan himself moved to Bukhara in February to March of 1220. Various figures are given for the size of the garrison, 12,000 in one source, 20,000 in another, and 30,000 in another. As it happened however most of the army decided to abandon the city and they fought their way out, an operation from which very few survived. The city then very sensibly surrendered and the population left the town which was plundered only the last defenders of the citadel were massacred, we are now told that they are only 400 of them. The city was not leveled to the ground, nor was there a general massacre. Though some fires broke out and caused damage. The mosques were pillaged however and the Mongols horses are said to have use of Koran stands for fodder troughs. An equally famous and similarly symbolic story is told of Genghis Khan’s addressing the inhabitants of Bukhara from the minbar and informing them that he was sent as a scourge from God. This of course is the only rational explanation available. Now we have a picture of him demanding that the place be dug up so he can find the treasure I think. (Showing slide) There doesn’t look like there has been much carnage. The important thing to notice for the moment is the discrepancy in the sources over the numbers. This becomes more acute at Genghis Khan’s next port of call Samarkand. Here the garrison is given as 110,000 by Juvaini, 60,000 by Juzjani, 50,000 by Ibn al-Athir, 40,000 by Nasavi. In one sortie in their first flush of their enthusiasm, the besieged lose either 70,000 men according to Juvaini or 50,000 by Juzjani, this is just in one attack. As in Bukhara the inhabitants themselves decided to surrender, and the Qadis with 50,000 people under their protection were spared being plundered. The rest of the inhabitants were driven out of the town, which was sacked. But the garrison in the citadel, 1000 Qatlugs perished in the mosque in the fire, and 30,000 were massacred when the citadel falls. However the city was not razed to the ground and again there was no general massacre. Although 30,000 artisans were given to the sons of Genghis Khan. However the city does seem to have been subject to further tribulations on later occasions and when the Chinese monk Chang Chung stayed there in December 1221 he reckoned the population had dropped to a quarter of its previous level which he puts at a fantastic figure of 100,000 families, so that’s presumably round about 500,000 people. Nevertheless life went on, he noticed there was much merchandise in the bazaars; this is December 1221 about a year and a half after the sack of the place and a flourishing and productive gardens stretched to an estimated 30 miles around the city with which not even Chinese gardens could compare.

Massacres did take place at some of the other cities of Transoxania, notably Utrar the scene of the original massacres of Mongols, Termiz and Organj or Gorganj where the besieging army was said to be over 100,000 strong, even before Juchi’s forces arrived. Despite a prolonged and bloody siege, 100,000 artisans were carried away to the East, and women and children were being enslaved and the rest were massacred. Each Mongol soldier, of whom there is now said to be only 50,000, that is half as many as are said to be taking part in the first place were given 24 men each to butcher. Which gives a total figure of 1.2 million dead. Juvaini had heard an even higher figure but for some extraordinary reason he couldn’t bring himself to quote it. Nevertheless the situation in Transoxania region in Khwarizm seems to have stabilized rather quickly and since the whole area came under direct undisputed Mongol rule, the work of reconstruction could begin immediately. The result was that the cities of the region recovered far more rapidly than those in Khurasan and Iraq as everybody agrees. (Showing slide map of Merv, Ray Nishapur and Herat) This of course is the main route along which Genghis Khans’s generals persued the fleeing Khwarazim Shah. Passing quickly into Khurasan and on to Balkh the situation there is confused. The city seems originally to have surrendered voluntarily and to have escaped a massacre, but then to have rebelled and suffered the fate of other cities in Khurassan. This slide shows the walls of Balkh as they remain today. The next one shows, there is nothing inside them except gardens, in other words this is an abandoned site. The Taoist monk I’ve already referred to Chang Chung passing about in September of 1222 noted that “there was a very large city which had recently revolted, the inhabitants had fled and we heard the barking of dogs in the city.”

It was still in ruins in the time of the famous Muslim traveler Ibn Battuta in the 1330’s and I suppose recovered some time after that, but not in the immediate sight. The survival of the empty shell of the walls is a strong visual witness to the abandonment of the city. Here we have no details of the figures. Its really with the effects in the other major cities of Khurasan that I’m mainly concerned. Genghis’s son Tolui was sent to subdue the province with the following reported results, in Merv between 700,000 and 1.3 million casualties, Nishapur an estimated 1.7 million casualties, excluding women and children. Here it took twelve days to count the dead. In Herat the first siege left 12,000 of the Sultan’s forces killed here. The townspeople were spared. After a revolt however after in a second siege the city was sacked. In June of 1222, with an estimated 2.4 million people butchered in one day. 200,000 survived, another figure suggests that with 1.6 million, 40 survived. So these are the reported facts.

We should first note the literary methods of the Persian bureaucrats to denounce the Mongol carnage. Where there were 100,000 people there remained 100 says Juvaini. An approximate survey of the provinces shows that only one in ten were prosperous and the rest were in ruins says Rashid ud Din. The use of such rhetorical figures was typical of the literary divanians, that is the men of the divan, who are used to bandying around large numbers, but who also fail to comment on the enormous accounting errors that cost so many of them their necks.

Well, the reports of events in these three cities are instructive. All three, but let us focus on one and notice in particular the fantastic figures. In Merv, each trooper was given 300 to 400 people to kill, some sources say 200. The process of killing them took five days. A Sayyid who went around counting the dead found more than 1.3 million. This was just the people who were laying out in the open not the ones who were in ditches and everything. As it took thirteen days, he must have counted a 100,000 people a day. Its much slower than it took to kill them in other words. Ibn al-Athir however who was unaware of the Sayyid’s efforts says that only 700,000 perished, thats’ half as many. 400 artisans were spared, you may note here the regular use of multiples of four and seven, 400 troops were left behind to complete the executions in Nishapur, where again 400 artisans were spared, 70,000 people were killed in Sabzivar the same number in Nisa.

If 1.3 million bodies, or before they were bodies were divided up into lots of between two or four hundred, this suggest that the army was only either three and a half or six and a half thousand strong. And in fact, by some astonishing coincidence, the figure given for the army at the siege of Merv is 7,000. Barthold makes a rare slip here and says it was 70,000. The figures for the Mongol forces suddenly become quite small, even realistic. The buildings and defenses are supposed to have been razed too. Yet in Merv in November 1221 following the massacre of Balkh, the population rose in revolt but it didn’t dare leave the city which was therefore presumably still standing, its new rulers repaired the fortifications and the walls, as well as agriculture, irrigation works, and so on. The people gathered there from all around attracted by the abundance of its wealth, when 5,000 Mongols arrived in the summer to crush the uprising, a further 100,000 people are left dead according to some reports with only 4 survivors. A few months later however the town was repopulated by those who had hidden in the desert or remained in their villages, and the walls were rebuilt. A local commander came and took charge and rallied a force of 10,000 men, a Mongol general returned with 100,000 men. This is the third visit by the Mongols, and carried out widespread torture for forty days, only 100 souls survived this. At the end of it all there are only 10 or a dozen Indians left residing in the city, I don’t know how they managed to get away with it.

These obviously contrary and fantastic figures which deserve no credit whatsoever, though they have often been regarded with suspicion, they have never been dismissed out of hand. Is there any way to achieve a more exact measurement? Demographic data of course are totally lacking for both the period before and after the invasions and unfortunately they are not available either for Mongke’s census, which was carried out in the 1250’s. In favor of the figures quoted above, as orders of magnitude, there is the demographic question mark that hangs over Yuan China, where incidentally we also get a sense of the tiny Mongol population compared with the native Chinese. One million Mongols for seventy million Chinese according to the census of the 1290’s carried out in China. This also reveals a drop in the population from the previous Chin and Sung periods where the population is a 100 million. This decline continued throughout the Mongol period; the census of 1293 showing a population of 60 million. Its difficult to account for such drops purely on account of plagues, disease, etc., whatever the underlying uncertainty in the figures themselves. What I mean here is that there is an unaccounted and dramatic drop in population in China as well, so that would obviously support the general view that there was also a similar drop in population in Persia. I might just mention something about the size of the army here quickly. I haven’t read out all the figures as they go along of the size of the forces involved but you will have noticed that quite often that they are supposed to be a 100,000 in a troop even with people going back to crack what should have been a tiny nut in Merv was a troop of a 100,000. Barthold estimates the total Mongol forces at between 150,000 and 200,000. The army at the time of Genghis Khan’s death was 129,000, reasonably well set out what it consisted of. And certainly there were a few additions added on since then, and of course as we all know a lot of people were scooped up on the way. Turkish tribes particularly or forcibly joined the Mongol forces on their way through. Nevertheless Barthold’s estimate of the upper end of 200,000 is not unreasonable. For the second invasion, thats the one under Hulegu against Baghdad, an army of around 170,000 is proposed out of a total Mongol population according to John Masson Smith of about 850,000. A higher figure nearly double that much, 300,000, technically at Hulegu’s disposal, whether it includes the Turco-Mongols or the whole army including all the extra various units were certainly never mobilized in a single campaign. The point about a lot of these other figures are too, it is very unusual for the whole army to be in action at the same place at the same time.

As we know Genghis Khan was with part of the army himself going down towards the Indian border at the same time other sons were knocking out Organj and then someone else was chasing the Khwarizmshah across Iran. So the units involved were probably fairly small, even if in sum they were quite big by the standards of today; I mean 200,000 is a pretty significant army. Well, what about some methods to try to quantify the populations involved. I may say that I approach this with extreme hesitation, and I noticed unfortunately although I shouldn’t say so that Jean Aubin had got himself in a complete muddle when he tried to do this so my figures are probably a little better than his, but they just show how dodgy it is to bandy figures around at all.

Nevertheless, this is just to try to establish an order of magnitude really. Are we talking about millions, hundred of thousands, tens of thousands or what? No one is ever going to pretend we can know an exact figure, but it certainly would be useful to have a rough order of magnitude. At the moment the only way it seems possible to do this is through using models of urban population density in areas within the walls in the rural areas. There is of course a complete lack of archives, but we may note for sake of comparison the Ottoman census of the rich merchant cities of Aleppo and Damascus which were provincial capitols in large oases each had 10,000 taxable hearths which may lead to a population of somewhere around 50,000 just multiplying five by hearths. As I say this is very imprecise, but it gives you a method of comparison. Yet Herat on the eve of the Mongol invasions is supposed to have 444,000 hearths. That’s 40 times as many, and implies using the same multiple roughly a population of 2,000,000. Petrushevsky the famous Russian scholar noticed that the population of 2,000,000 would more or less allow the massacre of 1.6 million in the siege, so he thought that this was a credible figure because it was entirely consistent with the other evidence available. But having worked out to his own satisfaction that the sums did add up, he then said that it didn’t seem really quite realistic, so he arbitrarily slashed the number in half and he said there was a population of 1,000,000. And of the Aqquuyunlu - who as you know are a late 15th century dynasty - Shiraz had a population of around about 20,000 people or 3,500 hearths in about 1515. Sticking with Shiraz at the end of the Qajar period, the census of people within the medieval walls was 6,200 houses inhabited by around 53,000 people. That’s the end of the 19th century. If you like to look at your handouts, medieval Damghan, in which the walls are still standing, encloses an area of approximately 470 hectares. Sheradil estimated with a population at its peak should have been round about 25,000 people. Not on a very scientific basis I admit but its roughly on the basis that in 1930 when that picture was drawn the size of the population was concentrated in roughly a fifth of the area available within the medieval walls and therefore as a maximum multiplied by five might have been 25,000 at its peak. This yields a figure of roughly 53 inhabitants per hectare which is quite low. In Nishapur where the medieval site was abandoned rather like Balkh. I’ve walked over the site at Nishapur - its rather dramatic. Its clear to see the old city walls with abandoned ruined fields with pottery shards all over them which have been excavated rather imperfectly unfortunately, but a lot of the material is in the Metropolitan Museum, it was very nice to see it there. Nishapur, from the extent of the ruined fields has been estimated by Bulliet at 1,680 hectares. He suggests a multiplier of between 100 to 200 people per hectare as embracing the highest and the lowest generally agreed figures from studies in other parts of the Islamic world as sort of population density. He applies this to two-thirds of the whole area which therefore allows for public spaces, gardens, mosques, all the rest of it, so its not absolutely cram packed the whole area.

Taking a range then of the minimum from Damghan of maybe 55 people per hectare and a maximum at the top end of the range round about 200 people for two thirds of a hectare. We can apply this range to the sizes of the circuits of the medieval cities. Unfortunately for the ones that we haven’t investigated thoroughly yet we have to rely on the figures given by Mustaufi, a contemporary geographer and historian of the circuits of the walls. This process itself is not itself without objections because there seems to be some argument over the length of the pace. But if we take it as being roughly a meter, which seems to be in the middle of the various options, then it also makes it much easier to do the sums. I am going to come up with some extremely approximate approximations of a range from the smallest one on the density population of 55 people per hectare up to 200 people for a reduced area two-thirds of a hectare. Nishapur, these are using the size of the cities as they were contemporaraly within their walls, has a population by these calculations of between 75,000 and 180,000. Ray, between 48,000 and 120,000. Tabriz, at the time between 12,000 and 30,000. Sarakhs between 8,000 and 21,000. Qum between 35,000 to 89,000. Qazvin 35,000 to 88,000. Shiraz 52,000 to 130,000. In other words most of these places are somewhere in the region of 30,000 population, and at the most in the case of Nishapur, 180,000. Even Nishapur, therefore the biggest of these cities can hardly have boasted a population of more than 200,000 people at its peak, 200,000. Its a place where 1.7 million are said to have been killed. For comparison, just to show these figures aren’t totally ludicrous, the new city of Nishapur which is built on the ruined fields to the west of the old site which has an area within its walls - which were still standing when these figures were taken of 105 hectares is very small - had a population who were still living within the medieval walls in 1910, of between 10,000 and 15,000. That’s about a fifth the size of Nisahpur roughly on the eve of the Mongol invasions at the smallest or a tenth of it at the biggest.

In the first census of 1956 by which time the population had spread outside the 14th century walls it had a population of 25,000. Now we know that on the eve of the Mongol conquest the city of Nishapur was greatly reduced from its former splendour thanks to the devastation of two earthquakes, devastating raids by the Turkish Ghuzz tribes about 60 years before the Mongol invasions followed by violent internecine strife within the city between the Shafi’i and Hanafi factions. Despite Yakut’s reports of the flourishing state of the city we know that the bulk of the inhabitants had moved to the suburb of Shadjiakh to the southwest of the city an area which contained only roughly only 165 hectares, the walls are still there - I have a photograph of them at home - suggesting a population of between 16,000 and 40,000 people, so that would be a perhaps more accurate estimate of the population of Nishapur at the moment of the Mongol invasions rather than at its peak of prosperity which had long been past.

In Ray, we may notice, talking of internecine strife, that before the Mongols arrived a city of 30,000 mosques was there in which 100,000 people had been killed in one incident of fighting between the Hanafi and Shafi factions, or according to another source 600,000 people. And in fact by the time Yakut visited it shortly before the invasions only one quarter of the city was left. It was basically a fight between the Hanafis, the Shafiis and the Shiites. The Shiites were the largest group. They were eliminated by the Shafis and the Hanafis, the Shafii and the Hanafis then fought it out amongst themselves. The Shafii who were actually the smallest group and whose quarter of the city was the smallest, won, and so when the Mongols got there it was actually only the Shafii quarter that was still there. Yakut specifically says that the city was deserted and an empty shell. The Shafiis submitted to the Mongols but their quarter was sacked anyway due to the presence in the locality of Khwarazmian forces. Its interesting concerning Ray that Ghazan Khan the later Mongol tried to revive the city but actually failed to do so, and the population and prosperity, as it were, moved to a neighboring town of Veramin. Also in Isfahan in the period between the original invasion and the second invasion there’s also a report of factional fighting between the Shafis and the Hanafis, which having run its course and caused a lot of damage eventually led to the Mongols capturing the city because one of them let them in. The result was that they all were killed. So these were an example of the sort of problems that were affecting some of these large cities before the Mongol even arrived.

Current excavation at Samarkand and at Merv should help provide similar opportunities long term. Going back to my number crunching wandering around to evaluate the size of the built up areas of these major cities to evaluate at different periods and ideally before the Mongol invasions. I’ve really just used Nishapur, which at the moment is the only one that has been excavated as an example.

So as in the previous situation as I mentioned we should not seek the economic causes for the collapse of Iran too casually in the destruction of the great cities of Merv and Nishapur, the most brilliant period of their civilization was already over. The Mongols invasions were preceded by long decades of disorganization and local difficulties. The Mongols themselves cannot be held responsible for the decline. This had already started.

You may notice Fars had already been said to have been ruined by the 11th century by the invading Shabansiqariar tribesmen, and in the 12th century Khuraasan and Kirman were ruined by the Ghuzz. Ruined; what does that mean? Anyway, not in the peak of their condition. Following this we have the ravages of the Khwarazimshahs and various Turks and Turkmen tribes and groups following their own warlords ever since the collapse of the Seljuqs.

So the figures I quoted above and the list of places are probably maximal. I think it is rather unlikely that they are underestimates. Elsewhere in northern Iran outside these large cities which I suppose are the most eye catching where resistance was offered, for instance in Ardabil, Nakhjawan and Marazbeh, there was trouble of course, but otherwise there was some pillage and some deaths but no wholesale sacking. Its interesting to compare perhaps some of it a selective way some of the other evidence given by Yakut who traveled through the area both on his way east and then on his way west fleeing from the Mongols shortly before their arrival. By comparing his evidence with that provided by later authorities. First of all in Azerbiajan, Yakut noticed that discord was endemic there almost all the town are falling into ruins and the villages are deserted. Ardabil, which I mentioned where the Mongols sacked it, despite two Mongol assaults, he says it now, maybe more flourishing than it was before the invasion. At Urjan Yakut noticed a walled town with a market, but mainly in ruins. Mustowfi writing in the Mongol period noticed that Ghazan rebuilt it with mortared stone walls and dedicated all the income from its agricultural harvest to charitable trusts.

Qarghazkunan, a place not all that far away, Yakut calls small with a nice bazaar but half-ruined. Mustowfi also says it was ruined in the invasions, I suppose there is a difference between ruined and half-ruined and it became a nomadic settlement.
Mar and also in Azerbijan - obviously I am focusing on Azerbijan because thats where the Mongols went - there is not much point in telling you what was happening in Kirman at the time. Yakut said that though it had been an important town and that it was now half-ruined and almost deserted due to a tax by local Kurds. Mutowfi echoes this, “It was once a large town with walls of 8000 paces but it was now only half-standing”. Here we have a problem with Mutowfi, but quite often you are not quite sure that he is recording contemporary information but actually just repeating the evidence of his sources who of course are writing at a different period.

As for Urmiye, Yakut says that despite its advantages it was not flourishing thanks to the negligence of its ruler. Mostoufi on the other hand writing in the Mongol period calls it a great city with walls of 10,000 paces and large gardens and prosperous environs. Barda on the way to the Caucausus, Yakut noticed that its former splendour had gone and that it was now just a village amidst the ruins. Delakhan, despite the Mongol sack though had survived, returned, and the town quickly took on it former appearance. Hovaar, that’s near Ray in 613, about 6 years before the Mongol invasions he said that it was almost ruined. Salmaas, Yakut found Salmaas partly ruined. Mutowfi noticed that its walls which had fallen into ruin had been restored by the vizier Taj al-Din Ali Shah.

The point of this great cataloguing - and I could go on - that there was quite a lot wrong with the situation before the Mongols arrived and indeed in some cases, although some areas of course remained ruined, in fact they were restored under the Mongols to a better situation than they had been beforehand. Of course one can give equally several examples of places ruined by the Mongols had not been repaired or that formerly flourishing places were now in ruins. I am not trying to say as I did in the beginning that the Mongol invasions didn’t happen. I am just trying to balance this out with some sense of what the situation was like on the eve of their invasion. So as I say this is not the whole story but it shows that in the regions through which the first Mongol invasions passed the notion of a prosperous and populous society needs at best a qualified acceptance.

This also the case with Iraq. This is a picture of Varamin (showing slide) This is the only mosque I believe that was actually started and completed within the Ilkhanid period of Veramin outside Ray and became a flourishing center. The old city of Ray never really recovered.

The breakdown of the caliphate has been thoroughly examined by many writers who note that its collapse was of largely symbolic importance. (slide shown) That fellow down there is the last Abbasid caliph in a very small palace. (slide shown) This is the siege of Baghdad. As we all know Baghdad succumbed rather quickly to Hulegu’s army in 1258. Leaving aside the symbolic significance of the collapse of Baghdad, which needn’t concern us here, there is again the problem of the numbers, of troops, armies and the dead. The figures for those killed are given as between 800,000 and, 2000,000 in various Arab sources. I have already said enough I think about figures such as these, but its worth noticing that none of them are contemporary observers. The only contemporary account of the siege of Baghdad is by Nasiruddin Tusi who has an interesting little sentence which isn’t normally given much emphasis. Which is that, after the city was pillaged for a week the people were given quarter and allowed to return to it. Which suggests that they weren’t all killed. Of course he had his own reasons for playing down the fall of Baghdad, because he had played a fairly large part in getting the Mongols into it in the first place. Nevertheless its interesting that he says what he does. We should also notice that Hulegu orders the viziers and the sahibdivan, that’s like another sort of vizier I suppose, to rebuild Baghdad and reopen the bazaars. Reopen them not rebuild them, as he left the city. In the thesis I mentioned earlier by Woo, there is an interesting coloration between the collapse of Abbasid rule and the frequency of floods in late Abaqa Baghdad which indicates a collapse of the irrigation and canal systems. He shows that repeated flooding culminated in major floods in 1255 and 1256 on the eve of the Mongol conquest which caused heavy damage. The Mongols themselves were compelled to take measures to repair the breaches to prevent further inundations. Subsequently considerable work was done in the reign of Juvaini and others. Woo points out that there were no further floods until 1277 and then again in 1284 and 1286, at a time when Juvaini was fighting and losing the struggle for political survival.

The only serious flood in the whole Ilkhanid period apparently occurred in 1324 from which Woo concludes that the Mongols maintained the canal system in Iraq in good operative condition. I don’t think there is time to go into the obvious objections with this argument here nor to the other elements of the situation in Iraq. Suffice it to say that more work remains to be done on this problem. Before concluding, I would like to mention briefly the historiographical problem that underlies in part this debate as I mentioned earlier. There is a pervasive tendency to exaggerate the ills of the early Mongol period. Early historiography stopped with Juvaini who wrote his book in 1260. His own relations with the conquerors are of themselves of considerable interest. There is nothing till the work of Rashid al-Din who paints a black picture to throw Ghazan’s and his own reforms into greater relief. Forgetting the earlier services of the Juvainis, Nasir al-Din Tusi and others, Rashid al-Din consistently accentuates the horrors of the pre-Islamic phase of Mongol rule which of course was initiated with Ghazan. However both Abaqa and Arghun were well aware of the need to preserve agriculture and to protect traditional society. In fact Rashid refers mainly to the crisis of the 1280’s and 1290’s immediately before Ghazan came to the throne. But even here there were competent administrators such as Sayed ud-Dulah and Sadr ud-Din who took measures against abuses Rashid later claims as his own.

Mustaufi’s geography, some of which I have quoted to you, does not give a picture of a starving and depopulated country, there were many large and flourishing regions among them particularly, Isfahan, was conspicuously prosperous, despite the fact as I mentioned that after the eventual sack following the dispute between the Hanafis and Shafis, Isfahan was reduced to a mound of ashes apparently. Earlier generations didn’t wait for Rashid’s reforms before investing in agriculture. Juvaini in Iraq and Yazd, the Iftiqar family in Qazvin enormously enriched themselves by serving the ruler and obviously enriched themselves through developing the land. Saveh, though devastated by the Mongols, quickly doubled in size when a local malik or ruler found a new town outside the walls which was served by a dam built by Juvaini. This became a residence of many leading people of the bureaucracy. Constructions of khanaqas supported by waqfs were also common for instance at Simnan and elsewhere by the sufi sheikh ‘ala al-Daulet Simnani, a childhood companion of Arghun turned sufi who invested his largesse in real estate as did the viziers as well. So that’s one point that the later sources on the whole play up a very stark contrast between the earlier Mongol rule and then the glorious light and joy that happened in the later period once Ghazan had become a Muslim.

Another element of the historigraphical problem is the question of perception. The historians don’t have a clue about the numbers involved. There was no accurate way of measuring them anyway. The towns and cities destroyed obviously represented large concentrations of people but they were probably a very small proportion of the total population, which in Iran as in most pre-industrial societies was predominantly rural. This was certainly the case in Iran right through till well into the twentieth century. To say the figures who perished in the sieges are swelled by refugees flocking

to the towns from the villages outside is entirely implausible; the towns were death traps. All the evidence is to the contrary, in other words it was much safer in the countryside, and that on the whole the countryside was not molested.

If one were to talk about perception I mean the whole focus of the sources is on the cities. The towns and the cities were the showcases of Islamic civilization and learning, hence the outrage at the trampling of Korans under the Mongol horses hooves in the mosque of Bukhara and the destruction of Merv’s famous libraries. In fact however it is consistently mentioned that the ulema class of the religious scholars and their hangers on largely survived. They were given immunities one way or another. The artisans who you might think were the other sort of worthwhile group of urban society were consistently carried off to carry on their work elsewhere. It is not clear when you have removed these two groups exactly how much there was left. The worst of the situation was being overrun by savages. (Slide of dancing shamans shown)

As the cities fell, so soon they rose again and the most conspicuous signs of destruction were relatively easily repaired. The effects of the Mongol conquests -rather than their subsequent rule - on agriculture is less easy to assess. The only seemingly objective measure is the question of revenue. Mustaufi, the geographer I’ve already mentioned gives a very famous passage about the revenues arriving at the Ilkhanate center at the end of the Ilkhanid period and comparing it with the situation in the Seljuq period and these show an enormous decline of course. But these figures are not particularly instructive and there are great problems of comparability. According to his figures to take just one example the Shabankara’i district of Fars apparently declined 85% since the Seljuq period. However between the reigns of Abaqa and Abu Said, that’s virtually the whole of the Ilkhanid period, more than a dozen Mongol amirs received this district to tax for their own account with their officials. Meanwhile a local dynasty remained in place until the 1340’s and many of its rulers are supposed to have bought prosperity to the region. Is this just relative? or is there a real comparability with earlier times? It seems certain anyway in this case that figures for what is reaching the central divan are hardly relevant to the local state of the countryside.

I think for the contemporary sources as indeed for later generations, part of the question of perception is one that you are seeing an end of a golden age with the Abbasids and this colors all the attitude to what followed it.

We may note finally that revenues from agriculture in the regions around Tabriz and Kashan were later still only a quarter of the product of urban taxes. Though this seems to confirm the decline of agriculture, it also shows the extent of urban regeneration among the Mongols who are anxious to pursue the trade which had bought them West in the first place.


(Please note that this lecture paper is a work-in-progress and is not to be copied, quoted or excerpted from without expressed written permission from Professor Charles P. Melville.)