Chagatai Khanate

The khanate that stayed on the steppe

Drafted from Biran, Qaidu and the Rise of the Independent Mongol State in Central Asia (1997)formation

Chagatai's inheritance was Transoxiana and the Tarim — the old Karakhanid and Kara Khitai lands, containing Bukhara and Samarkand, two of the wealthiest cities in the Islamic world, and a very large amount of pasture.

The khanate spent a century unable to decide which of those it was. Its rulers were pulled between the settled Muslim towns that generated revenue and the nomadic Mongol military that generated power, and every attempt to move toward the towns — Kebek minting a settled coinage and building a residence near Qarshi in the 1320s, Tarmashirin converting to Islam in the 1330s — produced a revolt from the steppe. Tarmashirin was overthrown for it.

The khanate split in the 1340s into a western half, Transoxiana, that became Muslim and settled, and an eastern half, Moghulistan, that stayed nomadic. This is the case that shows what the Ilkhanate and the Golden Horde avoided by converting decisively: the Chagatai state fractured along exactly the seam that every steppe polity governing farmland has to bridge. The Liao had found the alternative solution two centuries earlier — two administrations, deliberately kept apart — and the Chagatai khans never tried it.

Out of the western half, and out of a minor Turkicised Mongol clan within it, came Timur.

The two countries

The ulus Chagatai received in the 1220s was not one country but two that happened to be adjacent, and the khanate spent its whole existence failing to be both.

South of the Tian Shan and west of the Pamirs lay Transoxiana: Samarkand, Bukhara, Khujand, the Zarafshan valley — irrigated, urban, Persian-speaking, Muslim, and the richest agricultural land between Baghdad and China. North and east lay the steppe and the mountain pasture, the Ili and the Chu and the Yulduz, which grew nothing and supported horses.

The khan's army came from the second. The khan's revenue came from the first. No other Mongol successor state had the two in such equal measure. The Golden Horde had grass and treated its farmers as tributaries at a distance. The Ilkhans had cities and went to live in them. The Chagatai had both, in one inheritance, with a mountain range between.

What the amirs wanted

Not the same thing the khan wanted, and this is the mechanism that wrecked every reign.

A tribal amir with his own following on the pasture had no interest in his khan becoming a settled ruler with a capital, a bureaucracy and a standing revenue, because such a khan would not need him. The amirs' position depended on the khan remaining what he had been: a first among equals who lived in a tent, moved with the seasons, and distributed what he took.

So the khanate developed a rule, unwritten and lethal: a khan who spent too long in Transoxiana, who built, who minted, who governed, was a khan who would be removed. Kebek tried it and is remembered fondly by the townsmen and hardly at all by the Mongols. Tarmashirin tried it and was deposed and killed. When the khanate finally broke in 1347 it broke on exactly that line, and the two halves went in exactly the two directions.

Poorest, and longest

For all that, it lasted half as long again as the Ilkhanate, and its eastern successor outlasted the Golden Horde.

The Ilkhans took the richest country any Mongol conquered, governed it properly, converted, reformed the taxes, patronised the best science in the world — and were gone in seventy-nine years. The Chagatai stayed poor, stayed mobile, stayed divided, and lasted a hundred and twenty-one; and their line was still ruling in the Tarim oases in the seventeenth century.

Being unable to become a state turned out to be survivable. Becoming one did not.

Nothing of science

Drafted from Battuta, The Travels of Ibn Batuta (1829)expansion

Ibn Battuta crossed the Chagatai khanate in the early 1330s, a hundred and ten years after Chinggis had passed through it, and what he records is not a ruined country. It is a country in which some places had recovered and some had not, and he says which.

Bukhara

He came out of Khwarazm after eighteen days of sandy and uninhabited desert, through Kath and Wabkana, and reached Bukhara — the principal city of the country beyond the Oxus.

After it had been ravaged by the Tartars, it almost entirely disappeared: I found no one in it who knew any thing of science.

That is a scholar's judgement and it is the harshest thing in his book. Bukhara had been the intellectual capital of the eastern Islamic world; a traveller who went there specifically to meet its jurists found nobody worth talking to.

Balkh

Across the Oxus, a day and a half over a desert with no house on it, he came to Balkh — and here the destruction was still total.

It now lies in ruins. It has not been rebuilt since its destruction by the cursed Jengiz Khan. The situation of its buildings is not very discernible, although its extent may be traced. It is now in ruins, and without society.

Its mosque had been one of the largest and handsomest in the world, with incomparable pillars — three of which Chinggis had pulled down, because somebody told him the mosque's treasure was hidden inside them.

A hundred and ten years, and the outline of the city was still visible in the ground and nobody was living in it.

Samarkand

And then, twenty days' ride from Balkh, the other half of the answer.

I then proceeded to Samarkand, which is a very large and beautiful city.

No qualification, no ruins, no lament. Samarkand had come back.

Why some and not others

The difference is not how hard each place was hit. It is what each place needed in order to work again.

A city on a river that runs by itself — Samarkand on the Zarafshan, with its canals short and its water arriving without help — can be repopulated. A city that depends on a long artificial system needs the system rebuilt before anyone can live there, and rebuilding it costs more than founding a new town somewhere easier. Balkh's water came a long way. Merv's came further.

And a city whose value was its scholars is a different case again, because the scholars are not replaced by repopulation. They are replaced by two generations of teaching, by libraries, by endowments, and by somebody wanting to come. Bukhara had people in it when Ibn Battuta arrived. What it did not have was anybody who knew anything, because the men who knew things had been killed or had left for Delhi and Cairo and Anatolia, and the books had burned, and the endowments that paid the teachers had been confiscated.

What the khanate did about it

Very little, for a century, and that is the charge against it.

The Ilkhans rebuilt: Ghazan's reforms, the observatory at Maragha, Rashid al-Din's suburb at Tabriz with its hospital and its sixty thousand books. The Chagatai khans mostly did not, because their army did not want them to and their revenue never stayed in one pair of hands long enough.

When the rebuilding of Transoxiana finally came it came under Timur, sixty years after Ibn Battuta's visit — and Timur did it by carrying off the craftsmen of every city he took and putting them to work at Samarkand.

Kebek's coinage

Drafted from Biran, Qaidu and the Rise of the Independent Mongol State in Central Asia (1997)expansion

For seventy years the Chagatai khans governed the richest cities in Central Asia without going near them, and the moment one of them tried to, the trouble started.

The arrangement they inherited

Chagatai's ulus ran from the Ili valley across Transoxiana to the Tarim. It contained Bukhara and Samarkand, and its rulers lived on the steppe north-east of them and came nowhere near.

The towns were farmed out. Mahmud Yalavach and after him his son Mas'ud Beg — Khwarazmian merchants in Mongol service — administered Transoxiana on behalf of the imperial government rather than the local khan, remitting revenue and keeping the cities functioning through a period in which the khans themselves were fighting each other and, repeatedly, the Great Khan.

That produced the strangest fiscal arrangement in the Mongol world: a khanate whose territory contained enormous urban wealth and whose rulers drew very little of it, because the collection apparatus reported somewhere else.

What Kebek did

Kebek came to the throne in about 1318 and made the obvious decision. He built a residence near Qarshi — Qarshi is the Mongol word for palace, and the town is named after it — which meant a khan living in Transoxiana rather than visiting it.

Then he reformed the coinage. A unified silver dinar and its half, struck to a consistent standard across the khanate, replaced the patchwork of local issues; the coins were good enough that the name survives in the Russian denga and eventually in dengi, money.

And he reorganised the provinces into tümens for administration and taxation — which is a nomadic decimal unit being used to govern farmland, and is exactly the kind of hybrid a state in this position produces when it starts trying.

Why it was dangerous

Because every one of those steps moved the centre of gravity from the men on horses to the men with ledgers.

A khan on the steppe is first among the tribal leaders who elected him and depends on distributing what he takes. A khan in a palace at Qarshi with a uniform coinage and a provincial tax administration does not need them in the same way — and they can see it.

Kebek himself was not deposed. His successor Tarmashirin went further, and was overthrown for it in about 1334 by an assembly of the very people the reforms had been making redundant — which is the point at which the khanate stopped being able to hold its two halves together.

What was actually at stake

Not religion, though the sources put it in religious terms, and not culture. Revenue and command.

The eastern nomads were not objecting to Islam as such — they converted eventually themselves. They were objecting to a khan whose income came from somewhere they had no share in, administered by people they had no hold over, which would within a generation have made them irrelevant.

They were right, and the western half is what happened when they lost. By the 1360s the Transoxianan khans were ceremonial and the Turkicised tribal amirs of the region held the power — which is the vacancy an amir of the Barlas walked into.

The feast he cancelled

Drafted from Battuta, The Travels of Ibn Batuta (1829)peak

Ibn Battuta reached the Chagatai khanate in about 1333, at the one moment when it looked as though it was going to work, and he records both the height and the exact constitutional reason it did not last.

The prince

Ala al-Din Tarmashirin he describes as a powerful prince, who has at his command a large army, and is remarkable for the justice of his laws — a Muslim khan who kept the prayers, received a visiting jurist properly, and ruled a territory the traveller places among the four great kingdoms of the world, alongside China, India and Iraq.

He also gives the cities, and the two entries do not match. Bukhara he found wrecked: after the Tartars ravaged it, he writes, it almost entirely disappeared, and I found no one in it who knew any thing of science — a hundred and ten years after the sack, in what had been one of the great centres of Islamic learning. Samarkand, a few days east, he calls a very large and beautiful city.

That is the khanate in two sentences. Enormous wealth, next to a ruin nobody had rebuilt.

The statute

Then he explains, better than any chronicle does, what a Chinggisid khan was actually bound by.

One of the statutes of the yasa, he says, was that on a certain day each year — they call it the feast — the descendants of Chinggis, the governors of the districts, the wives of the nobles and the general officers of the army must all assemble. And if the emperor had altered any of the statutes:

the nobles should stand up and say, Thou hast done so and so upon such and such a day, and hast made an alteration in the statutes of El Yasak, and therefore, thy deposition is a necessary consequence. They are then to take him by the hand, and remove him from the throne, and to place in it another of the descendants of Jengiz Khan.

An annual assembly with the power to depose the sovereign, by naming the specific day on which he broke the law, and to replace him from within the same family. Any noble who had committed a crime was tried on the same occasion.

What Tarmashirin did about it

Now, Tarmashirin had entirely abolished the observance of this day, which gave very great offence.

Some time after Ibn Battuta left, the account continues, the Tartars and their nobles assembled anyway, deposed him, appointed a relative in his place, and pressed the matter so far that Tarmashirin fled and was put to death.

The charge is usually reported as apostasy from Mongol custom, and this version is more precise and more damning. He did not merely stop attending the assembly. He cancelled the one procedure by which he could be held to account, and the people it belonged to held it without him and used it on him.

Why that is the peak and not the recovery

Everything Kebek and Tarmashirin had built — the residence in Transoxiana, the coinage, the provincial administration, the conversion — pointed the khanate at the settled south. All of it made the annual assembly on the steppe less necessary to the khan and more clearly the only leverage the steppe had left.

So the reform programme and the thing that destroyed it were the same programme. A khan who succeeds in becoming a Muslim sovereign of Transoxiana has, by succeeding, given the men who elected him nothing to do but depose him.

Within a decade the khanate had split in two, and neither half was ever again worth a visiting traveller's chapter.

The day they took him by the hand

Drafted from Battuta, The Travels of Ibn Batuta (1829)peak

Ibn Battuta left the Chagatai khanate on good terms with its ruler, and then heard what happened to him. The explanation he was given is the only clear description anybody wrote down of the constitutional machinery of a Chinggisid state.

The book

Everything rested on the yasak — the code attributed to Chinggis, which Ibn Battuta understood as a book, and which he says enacted that whoever opposed any of its statutes should be put out of office.

He may or may not be right that it existed as a written text; that is still argued about. What matters is that everybody in the khanate believed it did, and behaved accordingly.

The day

One of the statutes concerned an annual assembly, which the Mongols called the feast.

On a fixed day each year the descendants of Chinggis, the governors of the districts, the wives of the nobles and the general officers of the army were all to come together. And if the ruler had altered any of the statutes in the year, the nobles were to stand up and say so:

Thou hast done so and so upon such and such a day, and hast made an alteration in the statutes of the yasak — and therefore thy deposition is a necessary consequence.

They were then to take him by the hand, remove him from the throne, and put another descendant of Chinggis in it. On the same occasion, any noble who had committed a crime was to be judged.

An annual review of the sovereign, by an assembly with the power to remove him for breach of the fundamental law, and a standing court for the nobility — described by a Moroccan judge who had just spent some days as the ruler's guest.

What Tarmashirin did

He abolished the day.

Ibn Battuta states it in one flat sentence and lets it do its own work. Now, Tarmashirin had entirely abolished the observance of this day, which gave very great offence.

He had reason. Tarmashirin was a Muslim — by some accounts the first Chagatai khan to make Islam general among his people — who spent his time in Transoxiana, kept the company of doctors of law, and was praised by Ibn Battuta for the justice of his laws and reckoned by him one of the four great kings of the world, whose territories occupy the middle station and to whom the kings of China, India, Iraq and the Turks all send presents.

He was, in other words, becoming a settled Muslim sovereign. The assembly was the institution through which the Mongol amirs of the eastern pasture controlled a khan, and it required him to come east once a year and stand still while they inspected him.

What happened

Some time, therefore, after we had left the country, the Tartars, together with their nobles, assembled and deposed him, appointing for a successor one of his relations; and to such an extent was the matter pressed, that Tarmashirin took to flight and was put to death.

The machinery worked exactly as described, on a ruler who had tried to switch it off.

The quarrel in one sentence

This is the fracture that broke the khanate thirteen years later, and it is worth stating plainly because it looks like a religious quarrel and was not.

The eastern amirs did not depose Tarmashirin for being a Muslim; the khans who followed in the east became Muslims themselves within a generation. They deposed him for governing from the wrong place, out of the wrong institutions, with the wrong people, in a way that made the annual inspection impossible and their consent unnecessary.

Every steppe polity in this history faced the same choice and most of them made it by drifting. The Chagatai are the one case where the two answers were argued out in a formal assembly with a written rule and a named procedure — and where the party of the steppe won, killed the khan, and then split the country in half anyway.

Two halves, two answers

Drafted from Biran, Qaidu and the Rise of the Independent Mongol State in Central Asia (1997)end

The split of the 1340s gave the khanate two futures, and both happened.

The western half, Transoxiana, went settled and Muslim. Its khans became increasingly ceremonial and real power passed to the Turkicised tribal amirs of the region — which is precisely the vacancy that Timur, an amir of the Barlas who could not himself be khan, stepped into. The Timurid empire is the western Chagatai khanate's continuation under management that had stopped pretending.

The eastern half, Moghulistan, stayed nomadic and stayed on the steppe, its rulers keeping Chinggisid descent and converting to Islam later and less completely. It lasted longer as a Chinggisid polity and left far less behind. Its successor states in the Tarim persisted into the seventeenth century.

The two halves are the cleanest controlled comparison there is for the question every steppe conquest state faces. Come inside and you get wealth, literacy, administration, and a ruling family that ceases to command the men who put it there. Stay outside and you keep the army and the legitimacy and remain poor and peripheral.

The Liao tried to have both with two administrations and got two hundred years. The Jin came inside and got a hundred and nineteen. The Yuan came inside half-way and got eighty-nine. Nobody found a third answer.

What actually broke in 1347

Qazan Khan was killed in battle by his own amirs in 1347, and no khan afterwards ruled both halves.

In the west the amir Qazghan took power and did not take the title, ruling through a Chinggisid figurehead — the arrangement that Timur would inherit and perfect. In the east, the Dughlat amirs found a boy who was said to be a son of the old khan, raised him at Aksu, and had a country to run within a few years.

Both halves therefore ended up governed by tribal families holding a Chinggisid in front of them. The difference was what those families wanted.

The western amirs wanted the revenue of Transoxiana, which meant order in the cities, which eventually meant a ruler strong enough to impose it and a khan who was a signature. The eastern amirs wanted the pastures and the right to make khans, and had no cities to speak of until Sultan Said took Kashgar in 1514.

The comparison

The two halves are the cleanest controlled comparison there is for the question every steppe conquest state faces.

Come inside and you get wealth, literacy, administration, and a ruling family that ceases to command the men who put it there. Stay outside and you keep the army and the legitimacy and remain poor and peripheral.

The Liao tried to have both with two administrations and got two hundred years. The Jin came inside and got a hundred and nineteen. The Yuan came inside half-way and got eighty-nine. The Ilkhans came all the way inside and got eighty.

The eastern Chagatai stayed outside, and their line was still ruling in the Tarim when the Dzungars arrived in the seventeenth century — three hundred and fifty years later, having done nothing anybody would call an achievement, and outlasting every one of them.

The amir who could not be khan

Drafted from Dughlat, The Tarikh-i-Rashidi: A History of the Moghuls of Central Asia (1895)afterlife

The western half of the Chagatai inheritance produced the last world conqueror, and he spent his entire career unable to use the only title that would have made his conquests legitimate.

Where he came from

Timur was a Barlas — one of the Turkicised Mongol tribes of Transoxiana, the military class of the western ulus — and he first appears in the record as a junior officer deciding which side to take.

Mirza Haydar, quoting the Zafar-Nama, puts him at the crossing of the Khojand river in March 1360, when Tughluq Timur brought the Moghul army down out of the east to retake Transoxiana. Timur's own chief, Amir Haji Barlas, collected troops from Kesh and Qarshi to oppose them, thought better of it, and withdrew to Khurasan. Timur did not go with him.

That is the whole man in one decision. The country was going to be taken; the question was who would be running it afterwards.

What he inherited and what he could not

He inherited a country that had been without an effective khan for a generation, with a tribal aristocracy accustomed to making and unmaking figureheads, and a Moghul army coming over the mountains every few years to assert a claim that nobody in Transoxiana wanted honoured.

What he could not inherit was the title. Timur was not a descendant of Chinggis Khan, and by the rule that governed the whole Turco-Mongol world he could therefore never be khan.

He did what Edigü would do on the Volga a generation later and what the Manghits at Bukhara would still be doing four centuries later: he kept a khan.

A Chinggisid puppet of the house of Ögödei sat on the throne. Coins were struck in the puppet's name. Decrees went out in the puppet's name. Timur called himself amir — commander — and güregen, son-in-law, because he married a princess of the Chinggisid line and that was the nearest thing to legitimacy available to him.

He conquered from Delhi to Ankara and from Moscow to the Persian Gulf as somebody else's officer.

What he did with Transoxiana

He rebuilt it, by the only method available to a man with an army and no revenue worth the name: he took the craftsmen.

The masons, tile-workers, weavers, armourers and glassmakers of every city he stormed were sent to Samarkand. The wealth of Delhi went to Samarkand. The scholars went to Samarkand. Ibn Battuta had found Bukhara with nobody in it who knew anything about science sixty years earlier; within a generation of Timur his own grandson Ulugh Beg was running an observatory at Samarkand whose star tables were being copied in Europe two centuries later.

It is the exact reverse of the Ilkhanid method. Ghazan tried to make a conquered country productive by reforming the way it was taxed. Timur made one city magnificent by stripping every other city he could reach.

Why he belongs here

Because the Chagatai khanate did not end in 1347. It divided, and both halves went on — and the western half's continuation is not a khanate at all but the Timurid empire.

The elements are all Chagatai elements: the Turkicised Mongol tribal army of Transoxiana, the puppet khan, the Chinggisid legitimism, the amirs who had been deposing khans since Kebek, and the permanent quarrel with the Moghuls in the east. What changed is that one amir was strong enough to stop pretending that the pretence mattered.

His descendants ruled Transoxiana for a century. One of them, driven out of it, took Kabul and then Delhi, and the dynasty he founded in India was still there when the British arrived — and it was called, by everyone including itself, the Mughal: the Persian word for Mongol.

The language, and the name

Drafted from Biran, Qaidu and the Rise of the Independent Mongol State in Central Asia (1997)afterlife

The Chagatai khanate is the least successful of the four Mongol successor states and the only one that gave its name to a literary language, an empire in India, and a modern nation's idea of itself.

Chagatai Turkic

The western half of the khanate produced, over the fourteenth and fifteenth centuries, a literary language: Turkic in the Arabic script, heavily loaded with Persian and Arabic, and called Chagatai after the khanate it grew up in.

It became the literary standard for the whole Turkic-speaking world east of the Ottomans, and it stayed that way for four hundred years. Ali-Shir Nava'i wrote in it at the Timurid court in the 1480s and argued in a famous treatise that it was capable of everything Persian could do — a case that had to be made, in Persian-dominated Central Asia, and that he substantially won.

Babur wrote his memoirs in it. So did the khans of Khiva and the Uzbek courts of Bukhara. Modern Uzbek and modern Uyghur both descend from it directly.

A khanate that could not decide whether to be a steppe polity or a settled one is the ancestor of the literary language of every settled Turkic people in Central Asia, and it acquired that role after it had ceased to exist.

Mughal

The eastern half is where the other name comes from, and the route is roundabout.

Timur came out of the western Chagatai lands, and his descendants ruled Transoxiana until the Shaybanid Uzbeks pushed them out at the beginning of the sixteenth century. One of those displaced Timurid princes, Babur, went south, took Kabul, and then took northern India.

He called his dynasty Timurid and thought of himself as a Chagatai — his memoirs use the word for his own people. Everyone else called the dynasty Mughal, which is simply the Persian for Mongol, and the name stuck to the largest and richest empire in Indian history.

So the term by which a European or an Indian refers to the Delhi and Agra of Akbar and Shah Jahan traces back through a displaced Timurid prince to the second son of Chinggis Khan, whose own khanate had been the poorest of the four.

Moghulistan and after

The eastern half kept the Chinggisid line longest. Its rulers converted to Islam later, moved into the Tarim oases, and the states they founded there — Yarkand above all — went on being ruled by descendants of Chagatai into the seventeenth century, when they passed under the authority of the Dzungars and eventually of the Qing.

The name of the region in Persian and Turkic sources was Moghulistan, the land of the Mongols, and the Turkic-speaking Muslim population of the Tarim was called Moghul by its neighbours for centuries after any of them were.

What the split proved

Come inside and you get the wealth, the literacy and the administration, and your family stops commanding the men who put it on the throne. Stay outside and you keep the army and the legitimacy and remain poor.

The Chagatai khanate is the only one of the four that did both, in two halves, at the same time, and the results are the controlled experiment.

The western half went settled and its khans became ornaments; within a generation the real ruler was an amir who was not a Chinggisid and had to marry one. The eastern half stayed nomadic and kept its Chinggisid khans for another three centuries, and produced nothing anybody outside the Tarim had to take account of.

Both halves were right about the other one.