Qara Khitai

A dynasty that walked west

formationDrafted from Biran, Cambridge Studies in Islamic Civilization (2005)

The Liao were the Khitan dynasty that ruled the northern Chinese frontier for two centuries under two administrations at once — one for the steppe, one for the settled prefectures. In 1125 the Jurchen destroyed it and the Jin took its place.

Yelu Dashi was a member of the Liao imperial clan who had passed the Chinese examinations. When the state collapsed he did not surrender. He rode north-west with a couple of hundred followers, gathered what Khitan garrisons and allied tribes remained on the steppe frontier, and kept going — across Mongolia, past the Uyghurs of the Tarim, into the Semirechye. In 1134 he took Balasaghun from the eastern Karakhanids, whose ruler had invited him in to deal with unruly Qarluq and Qangli tribes and discovered he had invited in a state.

Biran, Cambridge Studies in Islamic Civilization (2005) · Bernard Gagnon · CC0

What he founded there is one of the odder objects in medieval Eurasia. It was a Chinese dynasty. The rulers took the title gurkhan, universal khan, but also Chinese reign names; the chancery worked in Chinese as well as Persian; the coinage was struck in Chinese form; the dynastic name they used of themselves was still Liao. Two thousand miles from the nearest Chinese-speaking administration, ruling a Muslim, Turkic and Persian-speaking population that had no reason to be impressed by any of it, they kept the whole apparatus.

It is a state that turns up constantly in other people's histories as the power in the background. The Karakhanids, the first Turkish Muslim dynasty, spent their last seventy years collecting taxes for it. The Khwarazmshahs made their empire by throwing its suzerainty off. It is the thing the Ghurids and the Chagatai inheritance both keep gesturing at, and it is almost never described in its own right.

Two hundred horsemen

expansionDrafted from Bretschneider, Mediaeval Researches from Eastern Asiatic Sources, volume I (1888)

The Liao dynastic history contains a short account of the man who refounded it two thousand miles away, and Bretschneider translated it in full. It is the Qara Khitai founding as told by the Chinese historians of the state Yelü Dashi walked out of, and it is better than any summary of it.

The man

He was a descendant in the eighth generation of Tai-tsu. He was well versed in the literature of the Liao, as well as in that of the Chinese; and besides this he was a clever horseman and skilful archer. In 1115 he received the degree of shi, and was admitted a member of the Hanlin academy.

A Khitan prince of the imperial line with a Chinese doctorate and a seat in the national academy, who could also shoot from the saddle. That combination is the whole of what he did next.

The answer to the emperor

When the Jurchen broke the Liao, the emperor Tianzuo abandoned his throne and took to roving about. Dashi assembled the dignitaries and put a prince of the imperial family on the throne instead. The emperor, when he heard, had the regent put to death and reprimanded Dashi for raising a rival while the legitimate emperor was still alive.

Dashi answered him:

Even at the time when the empire was still in full vigour, your majesty had not been able once to repulse the enemy, but had fled far away, abandoning the people in their calamities; and had I then placed Shun on the throne, the reigning family would still have been descendants of our ancestor Tai-tsu. Has it not been the more dignified course now thus to act, instead of imploring the enemy for mercy?

To this the emperor had nothing to reply. He entertained Ta-shi with a feast and pardoned him.

Dashi did not feel safe. He killed two of the court dignitaries, declared himself sovereign, and left by night at the head of two hundred well-armed horsemen, heading north.

Bretschneider, Mediaeval Researches from Eastern Asiatic Sources, volume I (1888) · Hu Gui · Public domain

What two hundred horsemen accumulate

Three days out, across the Etsina river, an Ongut chief gave him four hundred horses, twenty camels and about a thousand sheep. He went on west to a city called Ketun and rested, and there he assembled the chiefs of seven cities and eighteen tribes, whose names the Liao history lists.

They raised more than ten thousand horsemen. He appointed officers and armed them.

In the second month of 1121 he sacrificed a black bull and a white horse to heaven, to earth and to his ancestors, put his troops in order and set out west.

The letter to the Uyghurs

Before he moved he had written to the Uyghur king, and the letter is a small masterpiece of the diplomatic use of history.

He reminded the king that his own ancestor, the first Liao emperor, had passed through Uyghur country and sent word to the king's ancestor: do you still cherish the remembrance of your ancient country? I shall occupy it for you if you wish to return. The Uyghur had thanked him and declined, saying his people had been away more than ten generations, were quite happy where they were, and had no desire to go back.

This proves that we have been for a long time on good terms with your people. Therefore: I intend to advance west, I ask permission to pass through your country, do not entertain suspicion.

The king came out to meet him, spent three days entertaining him, gave him six hundred horses, a hundred camels and three thousand sheep, handed over sons and grandsons as hostages, declared himself a vassal, and escorted him to the border of his realm.

That exchange cost nothing and bought a safe corridor, remounts and a client — using an offer his great-great-grandfather had made three hundred years earlier and which had been refused.

Ten thousand li

From there, the history says, he vanquished everyone he met and kept the peace with those who surrendered voluntarily, and the army advanced ten thousand li, overthrowing kingdoms and taking camels, horses, cattle, sheep and goods beyond counting.

At Samarkand he met the combined army of all the western kingdoms, a hundred thousand men. What he said to his troops before it is the last thing in the Chinese account worth quoting, and it is a professional's assessment:

We see before us a numerous host, but it wants an able commander. When they are attacked in front, the rear will fail to come to the rescue, and there can be no doubt our army will gain the victory.

He split into three: twenty-five hundred against the right flank, twenty-five hundred against the left, and the main body at the centre, all going in together. The enemy army was destroyed, and for ten li the ground was covered with bodies.

He rested at Samarkand for ninety days.

The day Transoxiana stopped being Muslim-ruled

peakDrafted from Biran, Cambridge Studies in Islamic Civilization (2005)

In September 1141, on the Qatwan steppe north of Samarkand, the Qara Khitai destroyed the army of Sanjar, the last great Seljuk sultan. It was the heaviest defeat any Muslim power suffered before the Mongols, and it is why this dynasty turns up so often as somebody else's overlord.

The consequences were immediate and structural. Transoxiana — Bukhara, Samarkand, the country the Samanids and Karakhanids had made — became tributary to a non-Muslim power for the first time since the Arab conquests. The Karakhanids continued to reign and began to remit. The Khwarazmshahs became tributaries too and stayed that way for two generations. Seljuk authority east of Khurasan never recovered.

The other consequence happened in Europe and is one of the strangest in medieval history. Word travelled west that a great king from the far east, not a Muslim, had destroyed a Muslim empire in a huge battle beyond Persia. Within a few years chroniclers in the Latin kingdoms were reporting a Christian priest-king in the east — Prester John — who would come to the aid of the Crusader states, and the letters and expectations attached to that figure ran for four hundred years. The scholarship generally traces the story to news of Qatwan.

What the victory did not produce is an occupation. The Qara Khitai took tribute, kept a resident to collect it, left the local dynasties on their thrones, and did not garrison the cities or interfere with their religion. That is the subject of the next chapter, and it is the reason this empire is remembered by its subjects less bitterly than almost anyone else who beat them.

The empress who ruled for her son

contractionDrafted from Bretschneider, Mediaeval Researches from Eastern Asiatic Sources, volume I (1888)

Yelü Dashi died in 1135, twenty years into the empire he had built by walking his people three thousand miles west rather than surrender to the Jin. His own son Yilie was not yet of age, and the Liao shi records what happened next without treating it as anything unusual: the widowed empress took the regency, ruled the empire in her own name for seven years, and handed it to her son only when he was old enough to want it.

Tabuyan

The empress, named Tabuyan and given the Chinese title Kan tien huang hou, governed under her own reign-name, Xian Qing, from 1135 until 1142. The Liao shi says nothing about her policies in the detail it gives Yelü Dashi's own campaigns, which is itself a kind of evidence: a seven-year regency that produced no crisis worth narrating is a regency that worked. The empire she handed over was intact, still collecting tribute from the Khwarazmian shahs and the Kara-Khanid princes Qatwan had made vassals, with no successor's legitimacy seriously questioned when Yilie finally took the throne himself in 1142.

Yilie's own reign

Yilie ruled for thirteen years under the reign-name Shao Xing, long enough to order a census of his own empire - eighty-four thousand five hundred families counted across the whole of it, the only such figure the Chinese sources preserve for the Qara Khitai at any date. He died in 1155 and was canonised, in the Chinese fashion his dynasty had kept even three thousand miles from where it was first used, as Renzong.

His own son was still a minor. For the second time in twenty years, the throne of an empire built by conquest passed not to a man but to a woman ruling in her own name - this time not a widow, but Yilie's own younger sister.

Killed by arrows

contractionDrafted from Bretschneider, Mediaeval Researches from Eastern Asiatic Sources, volume I (1888)

Pusuwan held the Qara Khitai throne for fourteen years, longer than either of the two emperors she ruled between, and the Liao shi records her downfall in as much detail as it gives any single event in the dynasty's whole reign in the west - because the men who finally killed her were the same family that had made her empire's conquests possible in the first place.

The regent

When Yilie died in 1155 leaving a son too young to rule, the regency passed to his own younger sister, Pusuwan, who took the title Chengtian Taihou and reigned under her own era-name, Chongfu, for fourteen years - the longest single tenure of anyone who ever held the Qara Khitai throne apart from Yelü Dashi himself. She had married Xiao Duolu, son of Xiao Wa-li-la, the general who had led Yelü Dashi's own cavalry divisions at the founding of the empire in Central Asia and whose family's loyalty had underwritten every gurkhan's authority since.

The affair the general would not tolerate

Pusuwan took her husband's own younger brother, Pugu Dishali, as a lover, and when the affair threatened exposure she had her husband Xiao Duolu murdered to clear the way - after first, with what the sources treat as a particularly pointed insult, granting him the honorary title Prince of Dongping. It was a miscalculation. Xiao Wa-li-la, the murdered man's own father and the dynasty's most senior surviving general, surrounded the palace with troops rather than accept the killing of his son. Pusuwan and her lover were both shot to death with arrows.

What survived her

The dynasty did not merely survive the killing of a reigning gurkhan by its own senior general - it settled immediately, without a further succession crisis, on Yilie's second son, Zhilugu, who took the throne under the era-name Tianxi and, by the Liao shi's own count, held it for the longest reign of any Qara Khitai ruler after the founder - the Chinese and Persian chronicles differ by several years on exactly when that reign ended, which is a dating quarrel rather than a competing account of the reign itself. What Xiao Wa-li-la's intervention actually protected was not one woman's authority but the principle that had let two of them hold it in the first place: that the throne belonged to Yelü Dashi's own bloodline, male or female, regent or empress, and that the loyalty the conquest generation owed it did not depend on the sex of whoever currently held it. Zhilugu's own reign would end differently - not at the hands of a general defending the dynasty, but at the hands of a refugee it had sheltered.

How to rule what you cannot outnumber

aside?What aside meansDrafted from Biran, Cambridge Studies in Islamic Civilization (2005)

The census of the Qara Khitai heartland records something like eighty-four thousand households. That is a small people holding suzerainty over Transoxiana, the Semirechye and the Tarim, and everything characteristic about how they governed follows from it.

They ruled indirectly and almost entirely through people who were already there. The Karakhanids kept their thrones at Samarkand and Kashgar, the Khwarazmshahs kept Khwarazm, the Buddhist Uyghur kingdom of Qocho kept the Tarim, and each remitted tribute and received a Qara Khitai resident to make sure of it. There was no attempt to settle Khitans on the land, no redistribution of estates, and — this is the part the Muslim sources consistently notice — no interference in religion.

The rulers were Buddhist. Their subjects were overwhelmingly Muslim, with Nestorian Christian, Buddhist and Jewish communities among them. Mosques functioned, qadis sat, the law was applied, and the Friday sermon was preached for local rulers rather than for the gurkhan. Muslim writers who had every reason to describe infidel rule as a calamity mostly describe it as unobjectionable, and several are notably warmer than that.

Set that against the century it happened in. The Almohads were building a caliphate on an infallible mahdi. The Fatimids had run a state whose foreign policy was a missionary organisation. The Abbasids had held the mihna, an examination of judges on a point of doctrine. The Qara Khitai had no religious programme at all — not tolerance as a principle argued for, but the practical indifference of a ruling group too small to impose anything and too sensible to try.

One further oddity belongs in the same chapter. They kept Chinese imperial forms nobody around them could read — reign titles, coin legends, a chancery working in a language with no speakers for two thousand miles. That is a claim addressed backwards, to a dynasty that no longer existed, rather than outwards to anyone who was actually present to be impressed.

Biran, Cambridge Studies in Islamic Civilization (2005) · Unknown moneyer · Public domain

The refugee who wrecked it

endDrafted from Biran, Cambridge Studies in Islamic Civilization (2005)

Kuchlug was a prince of the Naiman, a confederation on the western Mongolian steppe that Chinggis Khan broke in 1204. He fled west, was received at the Qara Khitai court, married the gurkhan's daughter, and in 1211 took the throne from his father-in-law.

Then he did the one thing his predecessors had never done. Kuchlug — raised Nestorian Christian, converted to Buddhism, and on any account a man in a hurry — began persecuting his Muslim subjects: closing mosques and madrasas in the Tarim oases, requiring conversion to Buddhism or Christianity, and reportedly having the imam of Khotan crucified on the door of his own school. Eighty years of a religious policy that had made an infidel suzerainty tolerable were reversed in about seven.

The Mongols arrived in 1218 under Jebe with instructions that are the point of the whole episode: proclaim the restoration of religious freedom. The oasis populations of the Tarim, faced with a choice between a Buddhist persecutor and a pagan army, treated the Mongol advance as a deliverance and largely did not resist it. Kuchlug was hunted into the Pamirs and killed. The empire ended without a battle worth the name.

Two things follow. The Mongols acquired Transoxiana's approaches cheaply and with a reputation for religious tolerance they traded on across the whole conquest — a reputation earned, in this particular case, by being less bad than a usurper. And Khwarazm, which had spent the previous decade throwing off Qara Khitai suzerainty, discovered in the same year that the buffer between it and the steppe was gone.

Calling it a conquest is accurate and thin. An empire that had governed by not asking anything of its subjects was destroyed by a ruler who asked, and then finished off by an army its subjects preferred to him.

The name that stuck

afterlifeDrafted from Bretschneider, Mediaeval Researches from Eastern Asiatic Sources, volume I (1888)

The empire lasted about eighty-five years and left three things: a reputation the Mongols spent, a name for China in half the languages of Eurasia, and a figure Europe waited four hundred years for.

Cathay

The Khitan ruled the northern Chinese frontier for two centuries, and everyone who reached China across the steppe reached it through them, so Khitan became the word for the country beyond.

It went west with the Qara Khitai and stayed. Russian still calls China Kitai. Mongolian, Turkic and Persian all carry versions of it. It reached Latin Christendom through the Mongol-era travellers as Cathay, and it was still being used as though it named a country distinct from China three hundred years later — which is why Portuguese and Jesuit expeditions went looking for Cathay after Europeans were already living in Canton.

A dynasty that had ceased to exist in China in 1125 gave the country its name in the language of everyone who approached it from the west, and the name outlived every state that ever used it.

Prester John

The other export was a rumour.

Word travelled west after Qatwan that a great king from the far east, not a Muslim, had destroyed a Muslim empire in an enormous battle beyond Persia. Within a few years chroniclers in the Latin kingdoms were reporting a Christian priest-king in the east who would come to the aid of the Crusader states, and the letters, maps and expectations attached to that figure ran for four hundred years — attaching themselves in turn to the Mongols, and eventually to Ethiopia.

Bretschneider, Mediaeval Researches from Eastern Asiatic Sources, volume I (1888) · Abraham Ortelius · Public domain

The scholarship generally traces the story to news of 1141. Europe spent four centuries waiting for rescue from a Buddhist Khitan who had never heard of it.

The reputation the Mongols inherited

The most consequential legacy is the one that is hardest to see, because it looks like a Mongol achievement.

For eighty years the Qara Khitai had governed Transoxiana and the Tarim by leaving everything alone: local dynasties on their thrones, a resident to collect the tribute, mosques open, qadis sitting, no settlement of Khitans on the land and no interference in religion. Muslim writers who had every reason to call infidel rule a calamity mostly describe it as unobjectionable.

Then Kuchlug reversed it in seven years, and when Jebe's army arrived in 1218 proclaiming the restoration of religious freedom, the oasis populations treated a pagan Mongol advance as a deliverance.

The Mongols got Transoxiana's approaches without a battle, and they got something more useful: a demonstration, in front of the whole Muslim east, that they were the ones who let people pray. They traded on that reputation across the entire conquest. It was earned, in this case, by being less bad than a usurper who had wrecked an arrangement somebody else had built.

The women

One thing the Muslim chroniclers record with notably little comment, and it is worth setting down.

The throne passed twice to women ruling in their own right: Yelü Dashi's widow as regent for their son, and later his daughter Pusuwan as gurkhan. Both issued orders, both were obeyed, and the Persian and Arabic historians who describe them do so in the ordinary register they use for any sovereign.

For a Buddhist steppe dynasty exercising suzerainty over the Muslim heartland of Central Asia, that is the most quietly remarkable fact about it, and nobody at the time appears to have thought it worth an argument.