Uyghur Khaganate

The ones who were owed

Drafted from Golden, An Introduction to the History of the Turkic Peoples (1992)formation

The Uyghurs did not arrive from outside. They had been inside the Türk khaganate the whole time, as one of the Tokuz Oghuz — the Nine Oghuz — and the Orkhon inscriptions name them among the peoples that kept rising against the qaghan who was addressing them.

Bilge Qaghan's own words, cut in the 730s: The Tokuz-Oguz people was my own people. As Heaven and Earth were in a turmoil they rose against us. Within one year we fought five times.

That is the last generation of Türk rule describing, on a monument, the people who would end it fourteen years later.

744

The second Türk khaganate came apart in the 740s in the ordinary way — a disputed succession, then a coalition. The Uyghurs, the Basmil and the Karluks combined against the reigning qaghan; the Basmil took the title first; the Uyghurs then took it from the Basmil and drove the Karluks west.

By 744 the Uyghur leader held the Ötükan — the mountain forest Bilge had told his people never to leave — and the Türk khaganate was finished.

What they inherited, and what they changed

They took the ground, the title, the runiform script and the political vocabulary. Their early inscriptions are in the same alphabet, on the same kind of stone, in the same idiom.

They also took the strategic position, and read it differently. Bilge's advice had been to stay in the forest, trade outward, and treat Chinese generosity as a weapon. The Uyghurs did the opposite on purpose: they went into the Tang relationship as deeply as they could get, and made it pay.

That decision is the whole of their history and it has two halves. It made them rich, literate and — eventually — a settled people with towns. It also made them dependent on a neighbour, which is exactly what the stones on the same river had warned against, and the warning was correct.

Karabalgasun

The clearest evidence of the difference is that they built a city.

The Türks had no capital. The Uyghurs put up Ordu-Baliq at Karabalgasun on the Orkhon: walls, a citadel, watchtowers, workshops, a quarter for foreign merchants, and a palace complex. The ruins cover several square kilometres and they are the first permanent capital anybody built on the Mongolian steppe.

A steppe empire that builds a walled capital has decided something. It has decided that what it holds is worth defending in one place, that the wealth coming in is worth storing rather than distributing, and that the people who manage it need somewhere to live. Every one of those is a break with the way the Türks had governed the same ground from the same valley.

The price of horses

Drafted from Golden, An Introduction to the History of the Turkic Peoples (1992)expansion

In 755 An Lushan revolted and the Tang empire came apart. Chang'an fell, Luoyang fell, the emperor fled to Sichuan, and the dynasty survived because it bought help.

The Uyghurs sold it.

What the cavalry cost

Uyghur horsemen fought for the Tang against the rebels and were decisive in the recovery of both capitals. What they were paid, and how they behaved while being paid, is recorded in the Chinese sources with an unusual amount of feeling — a dynasty writing down the terms of its own rescue is not a flattering genre.

They took an agreed right to plunder a recaptured city. They took silk in quantity. And, permanently, they took a trade: horses for silk, at a fixed ratio, in quantities the Uyghurs set.

The horse-and-silk exchange is the heart of the ninety-six years. Tens of thousands of horses a year went east; hundreds of thousands of bolts of silk came west. The Chinese sources say the horses were often poor and the ratio never favourable, and that the court could not refuse — partly because it needed cavalry, partly because the alternative was a khaganate with a grievance on the northern frontier.

What silk is for

A steppe empire does not want silk to wear. It wants silk because silk is currency.

A bolt of Chinese silk is portable, durable, divisible, universally valued from the Yellow River to the Mediterranean, and — crucially for a qaghan — it is something he can hand to a subordinate chief in front of witnesses. Steppe authority is held together by distribution: the ruler who can pay is the ruler who is obeyed, and the ruler who runs out of things to hand over stops being one.

So the Tang subsidy was not tribute in the sense of humiliation money. It was the raw material of Uyghur politics, and the khaganate needed it in bulk, annually, in order to remain a khaganate at all.

That is why the terms could not be softened. A ruler who accepted a smaller consignment would be distributing less at home.

The merchants who came with it

Converting silk into everything else required people who knew the roads, and those people were Sogdians.

Sogdian merchants from Samarkand and Bukhara had run the Silk Road caravan network for centuries, with colonies from the Crimea to Chang'an. Under the Uyghurs they moved into the khaganate itself: as traders, as moneylenders to the Chinese court on Uyghur account, and as the literate class the khaganate had not previously possessed.

They brought a script — the Sogdian alphabet, which the Uyghurs eventually adopted in place of the runiform letters. They brought commercial law and accounting. And they brought a religion, which is the subject of the next chapter and the strangest decision any steppe ruler ever made.

The shape of the bargain

By about 780 the Uyghur khaganate was the richest steppe polity that had ever existed, with a walled capital, a foreign merchant class, a literate chancery, and an annual income denominated in Chinese silk.

Every one of those advantages was a hostage. The capital could be found and burned. The merchants would go wherever the money was. And the income depended on the continued existence of a Chinese dynasty that resented paying it, on a route that ran through country the khaganate did not control.

An empire that chose Manichaeism

Drafted from Golden, An Introduction to the History of the Turkic Peoples (1992)peak

The remarkable thing about this khaganate is the religion. Around 762 Bogu Khagan adopted Manichaeism — a dualist faith of Mesopotamian origin, persecuted in the Roman and Sasanian empires and by then a minority tradition surviving largely among Sogdian merchants — and made it the state cult of a steppe empire. It is the only state in history to have done so. The Sogdian commercial network that ran the Silk Road was the medium, and the khaganate's literate administration came with it.

Why a steppe ruler would do it is worth thinking about rather than filing as an eccentricity. Manichaeism arrived with the Sogdian merchants who ran the khaganate's commerce and staffed its chancery, and adopting it bound that class to the throne. It was a literate religion with a scripture, which gave the court a clerisy. And it belonged to nobody else's empire: it was not Chinese Buddhism, not Tang state ritual, not the faith of any neighbour who could claim authority over a convert.

That is the same calculation the Khazar khagans made on the Volga at almost exactly the same time, and reached by two steppe states that had no contact with each other. A power between larger religious empires does not want the religion of either.

The cost was that it made the khaganate stranger to its subjects than to its merchants. Manichaeism forbade meat and required a class of Elect who did no work — an awkward proposition on the steppe — and it never took hold below the court and the trading towns. When the khaganate fell, the religion went with it everywhere except among the Uyghurs who had already stopped being nomads.

840

Drafted from Golden, An Introduction to the History of the Turkic Peoples (1992)end

The khaganate ended in about three years, and the sequence is a textbook case of what happens when every one of a state's advantages is also a dependency.

The three winters

The 830s went wrong on the ground first. The sources record severe winters, disease among the herds, and famine — the zud that periodically empties the Mongolian steppe of animals and, with them, of the wealth that holds a confederation together.

Then the politics. A succession dispute at the top produced the ordinary thing: a faction, a murdered qaghan, and a defeated commander who went and found someone else to fight with.

The Kirghiz

The Yenisei Kirghiz had been subjects of the Türks and then of the Uyghurs, in the forested country north-west of the Ötükan, and they had been fighting the khaganate on and off for two decades. In 840 an Uyghur general defected to them and brought them down the valley.

They took Karabalgasun, killed the qaghan, and burned the city.

Then they did not stay. The Kirghiz had no interest in ruling the Orkhon and no intention of running an empire from a walled town on it; they went back north with what they had taken, and the Mongolian steppe was left without a central power for the first time in three hundred years and stayed that way until the Mongols.

Why a capital was the wrong asset

The Türks had nothing to burn. Their qaghans lived where the grass was, and a Chinese army that marched into the Ötükan found a valley.

The Uyghurs had put everything in one place: the treasury, the chancery, the merchant quarter, the temples, the workshops, the stored silk. It made them formidable while it stood and it made the war a single objective. One city, one season, one result.

That is the sharpest lesson of the ninety-six years, and it was not lost on anybody afterwards. When the Mongols built a capital at Karakorum four hundred years later, they put it about twenty-five miles from these ruins, in the same valley, for the same reasons — and Karakorum was burned too.

The scattering

The population went in three directions.

Some went south to the Chinese frontier and were massacred or absorbed. Some went west into the Gansu corridor and founded a kingdom at Ganzhou. The largest group went south-west into the Tarim basin, to the oasis towns around Turfan and Kucha, and founded the kingdom of Qocho.

That third group stopped being nomads. Within a couple of generations they were farming irrigated land, living in walled towns, writing in the Sogdian-derived alphabet, and — increasingly — Buddhist rather than Manichaean.

They are the reason the end of this khaganate is not the end of the story. What was destroyed in 840 was a steppe empire. What survived it was a literate settled people who spent the next four centuries being indispensable to everybody who came after.

The alphabet the Mongols borrowed

Drafted from Golden, An Introduction to the History of the Turkic Peoples (1992)afterlife

The khaganate on the Mongolian steppe lasted ninety-six years. What followed it mattered more.

When the Kyrgyz broke it in 840, Uyghur groups migrated south-west into the Tarim basin and the Gansu corridor and founded settled kingdoms, principally Qocho. There they stopped being nomads: irrigated agriculture, walled towns, a manuscript culture in several religions — Manichaean, then increasingly Buddhist, with Nestorian Christian communities alongside.

They also became the literate class that every subsequent steppe empire hired. When Chinggis Khan needed an administration and a script, he took the Uyghur alphabet — itself descended from Sogdian, and ultimately from Aramaic — and had Mongolian written in it. Uyghur scribes staffed the Mongol chancery. Mongolian is still written in a direct descendant of it in Inner Mongolia today, and Manchu was written in a descendant of that.

Strictly, none of it is the khaganate's doing. The adaptation to Mongolian happened two centuries after 840, under a different empire, by people whose ancestors had left the Orkhon and become farmers. What the khaganate did was adopt a merchant alphabet from Samarkand for its own use, which is the least dramatic decision in its history and the one still visible on a street sign in Hohhot.