Khazar Khaganate

The third option

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

The Khazars came out of the wreck of the western Turkic khaganate. When that state came apart in the 650s under Tang pressure, the ruling house on the lower Volga did not disappear with it: it kept the Ashina claim, kept the title qaghan, and turned a wing command into a state of its own between the Caspian and the Black Sea.

That is a common enough way for a steppe polity to begin. What the Khazars did with the position afterwards is not.

The position

The lower Volga is not a place anybody would choose for a capital on agricultural grounds. It is a place you choose if what you want to tax is movement.

Four systems crossed there. The Volga carried furs, wax, honey and slaves down from the northern forests. The Caspian carried the trade of Iran and Khurasan. The steppe road ran east into Central Asia and west into the Pontic grasslands and beyond them to the Danube. And the Caucasus passes, above all the one at Derbend, were the gate between all of that and the Muslim world.

A state sitting on the junction of those four does not need to produce anything. It needs to be trusted by everyone using them, and it needs to be able to shut the Caucasus.

Between two universalisms

By the eighth century the Khazars had a Christian empire on one side and a Muslim caliphate on the other, and both of those states were organised around the proposition that everyone ought eventually to belong to them.

A pagan steppe khaganate between them was a natural object of missionary attention from both, and conversion to either would have meant something specific and unattractive: a Christian Khazar khagan becomes a junior partner of Constantinople, a Muslim one becomes a client of the caliph. Both religions came with a hierarchy attached, and in both hierarchies the convert enters at the bottom.

Somewhere in the eighth or ninth century the ruling elite converted to Judaism instead.

The evidence is real and comes from three directions that had no reason to agree: Arab geographers, Byzantine accounts, and a Hebrew correspondence between a Khazar king and a Jewish official in Muslim Spain. How deep the conversion went below the elite is not recoverable and was probably not very deep.

The obvious reading is the political one. A monotheism with no imperial sponsor made the khaganate legible to both neighbours — a people of the book, with a law and a scripture, arguable with in terms both understood — without making it subordinate to either. It is the third option, and the Khazars are the only state that ever took it.

What it bought them

Three hundred years, in the most contested corridor in western Eurasia, as a power both empires had to negotiate with rather than absorb.

It also cost them a place in anybody's story afterwards. A state that stops the Arab advance into eastern Europe is a hero of Christendom; a state that holds the Caucasus against Byzantium is a bulwark of Islam. A state that did both while being neither belongs to no later tradition, and gets remembered by nobody's chroniclers except as a neighbour.

Seven judges

Drafted from al-Masudi, Les Prairies d'or, volume II (1863)expansion

Al-Masudi, writing at Baghdad in 943, gives the fullest description anybody wrote of the Khazar capital while it was standing, and what he describes is not a nomad camp. It is a constitutional arrangement.

The city

Atil, he says, is cut in three by a great river coming down from the high country of the Turks, one branch of which runs off towards the Bulghars and into the sea of Maeotis. The town is built on both banks, and in the middle of the river is an island holding the seat of government. The king's castle stands at one end of the island, joined to one bank by a bridge of boats.

The population is Muslims, Christians, Jews and pagans. The king, his court and everyone of Khazar race practise Judaism, which had become the dominant religion of the state, he says, since the caliphate of Harun al-Rashid.

He adds a specific and datable reason why there were more Jews there than there used to be: in the year 332 of the Hijra — the year he was writing — the emperor Romanos at Constantinople was persecuting the Jews of his empire to convert them to Christianity, and a great many had left Rum for the Khazars.

The seven

Then the sentence that makes the record worth having:

It is an invariable rule in the capital of the Khazars that there should be seven judges in all: two for the Muslims, two for the Khazars, who decide according to the Torah, two for the Christians, who decide according to the Gospel, and one for the Slavs, the Rus and the other pagans.

Not toleration as a policy announced from above. A fixed bench, with a fixed number of seats, allocated by community, each applying its own law — and one seat for people with no book at all.

Nothing like that existed in either of the empires on the Khazars' borders. The caliphate had a developed law for non-Muslim communities and it was a law of subordinate status. Byzantium had no equivalent at all. Here four legal traditions sat at the same level in the same city because the city's business required all four to keep coming.

The Muslim guard, and its terms of service

The most extraordinary thing al-Masudi reports is the army.

The Khazar king's guard was Muslim: the Larisiyya, originally from the neighbourhood of Khwarazm, who had come to the Khazar realm not long after the appearance of Islam, fleeing war and plague. Seven thousand of them served as the king's horse-archers, in cuirass, helmet and mail, some with lances, equipped as Muslims usually are. They had judges of their own religion.

They had also negotiated their terms, and al-Masudi lists them: free exercise of their religion, permission to build mosques, the call to prayer to be made publicly, and the vizier always to be chosen from among them. The vizier in his own day, he notes, was a Muslim named Ahmad ibn Kuwayh.

And one clause more:

It was agreed that whenever the king of the Khazars should be at war with the Muslims, those who serve in his army should stand aside and not fight their co-religionists, but should march against all other infidels.

A Jewish king, whose standing army is Muslim, whose chief minister is constitutionally required to be Muslim, and whose army has a written exemption from fighting Muslims.

What kind of state that is

Not an empire in any sense the word usually carries. A state of this shape cannot conquer in a direction its own guard is excused from fighting in, and it does not need to.

What it can do is hold a junction and charge for it, and guarantee to four different legal communities that their contracts will be enforced by a judge they recognise. That is the whole product, and it was worth enough that people came from Khwarazm and from Constantinople to be inside it.

The wall in the Caucasus

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

The Arab-Khazar wars of the seventh and eighth centuries are the least discussed of the caliphate's frontiers and arguably decided the most. Umayyad armies pushed north through the Caucasus repeatedly; the Khazars pushed south; the fighting was severe, and in 730 a Khazar force destroyed an Arab army and killed its commander in Azerbaijan. Neither side broke through permanently, and the frontier settled in the Caucasus rather than on the Volga or the Don.

The consequence is that the caliphate did not expand into eastern Europe. The steppe north of the Black Sea, and the Slavic populations beyond it, developed outside the Islamic world for reasons that have a good deal to do with Khazar military capacity.

The khaganate's other distinction is that its ruling elite converted to Judaism, at some point in the eighth or ninth century. The evidence is real — Arab, Byzantine and Hebrew sources, and a correspondence between a Khazar king and a Jewish official in Cordoba — though the depth and extent of the conversion below the elite is not recoverable. The obvious reading is political: a power sitting between a Christian empire and a Muslim one, whose commerce depended on being trusted by both, took the third option.

Kievan Rus took Sarkel on the Don and then Atil itself between about 965 and 969, and the khaganate stops appearing in the sources as a power.

A letter to Cordoba

Drafted from Golden, An Introduction to the History of the Turkic Peoples (1992)aside?What aside means

In the 950s Hasdai ibn Shaprut, a Jewish physician and diplomat in the service of Abd al-Rahman III at Cordoba, wrote to the king of the Khazars asking whether it was true that there existed a sovereign Jewish kingdom. A reply survives, attributed to King Joseph, describing the conversion, the royal genealogy and the extent of the realm.

The authenticity of the exchange has been argued over for a century and the current position is broadly that it is genuine, with later editing. What is striking regardless is the geography of it: a Jewish official of an Umayyad caliph in Spain corresponding with a Turkic khagan on the Volga, in Hebrew, about a conversion that had happened two centuries earlier.

The Khazar conversion has attracted more speculation than evidence, most of it in the twentieth century and much of it politically motivated in one direction or another. What the sources support is an elite conversion, a court that adjudicated between Muslim, Christian, Jewish and pagan communities, and a capital where all four had standing.

The polity that stopped the Arab advance into eastern Europe therefore did it while being neither Christian nor Muslim, which is why neither surviving tradition made much of it. States get remembered by their heirs, and this one had none who wanted the inheritance.

Five hundred ships

Drafted from al-Masudi, Les Prairies d'or, volume II (1863)end

Al-Masudi tells one story about the Khazars at length, and it is the story of how the arrangement broke.

The bargain at the mouth of the river

Some time after the year 300 of the Hijra, about five hundred Rus ships, each carrying a hundred men, came down to the channel that connects the Black Sea with the Caspian.

At that channel the Khazars kept a strongly entrenched post whose whole job was to stop anyone coming through from either direction. The Rus sent to the king and offered him a share of whatever they took. He agreed and let them past.

They spent several months on the Caspian, and al-Masudi is plain about what that meant: they raided the coasts, killed and drowned thousands of Muslims, and nobody dared put to sea against them, so the very populous shores simply fortified themselves and waited. Then, glutted and tired of it, the Rus came back to the river mouth and sent the king his agreed portion of the spoil.

He took it.

Al-Masudi adds, in passing, the detail that explains why the Khazars had to bargain rather than fight: this prince possesses no ships, and his subjects are not accustomed to navigation. A state on the Volga and the Caspian with no fleet.

What the guard did

Then the Larisiyya and the other Muslims of the country heard what had happened, and went to the king.

Let us take vengeance on this people that has ravaged the country of our brothers, shed their blood and dragged their women and children into slavery.

He could not restrain them. So he did the only thing left: he sent word to the Rus that they were about to be attacked.

The Muslims came down the river and met them. The Rus were put to the sword or drowned. And, al-Masudi says, since the news of that disaster they have never repeated the attempt.

Read it twice

The first reading is that the Khazar system worked. A confessional guard with a written exemption acted on it, the king honoured the exemption, warned the party he had contracted with, and stood back.

The second reading is that the state had stopped being able to decide anything. Its king took a cut of a raid on his neighbour's subjects, could not prevent his own army from avenging them, and ended by tipping off the raiders against his own troops. Every party in that paragraph is doing something reasonable and the sovereign is not governing.

The Rus came back

They came back overland. Around 965 to 969 the Rus under Sviatoslav of Kiev took Sarkel on the Don and then Atil itself, and the khaganate stops appearing in the sources as a power.

There is no long decline recorded and no great battle described. A state whose product was a guaranteed crossing, in a corridor where the crossing had become contested by people with better boats and no interest in guaranteeing anything, simply stopped being purchased from.

Ibn Hawqal, passing through afterwards, describes the place as wrecked. What happened to the population is not recorded anywhere reliable, which has left a vacuum that two centuries of speculation have been poured into.

The vacuum

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

Most states that end leave a successor arguing about them. This one left a hole, and the hole has been filled, repeatedly, by people who needed it to contain something.

What actually followed

Politically, the answer is straightforward and unglamorous. The Pontic steppe passed to the Pechenegs and then the Cumans; the Volga route passed to the Rus, who had just taken it by force and had no reason to keep paying anybody for it; the Volga Bulghars, who had been Khazar tributaries and had converted to Islam in 922, went on as an independent Muslim state on the middle Volga until the Mongols.

Khazar communities are mentioned here and there for a century or so afterwards — in Crimea, on the Taman peninsula, in a Byzantine document or two — and then they stop being mentioned. There is no Khazar successor state, no Khazar diaspora that anybody traced at the time, and no later dynasty anywhere that claimed descent from them, which is unusual: steppe pedigrees were valuable and were routinely claimed.

Why the silence is loud

Consider what the khaganate did. It held the Caucasus against the Umayyads for about eighty years, in a war severe enough that in 730 a Khazar force destroyed an Arab army and killed its commander in Azerbaijan. The caliphate never got past it, and the steppe north of the Black Sea — and the Slavic populations beyond — developed outside the Islamic world.

Any state that had done that as a Christian power would be a fixture of European historical memory. As a Muslim power it would be a fixture of the other one. Being neither, it belongs to no continuous tradition that had an interest in remembering it, and the Rus who destroyed it wrote the account that survives best.

What the vacuum attracted

An unremembered Jewish kingdom of uncertain fate is an almost perfect vessel, and the twentieth century treated it as one.

It has been used to argue that Ashkenazi Jews descend from converted Turks rather than from the ancient Near East, a claim advanced with various motives — some of them by people trying to detach European Jews from any connection to Palestine, some by people trying to detach them from Europe. It has been used in the other direction as evidence of Jewish statehood in exile. Genetic and philological work has not supported the descent claim, and the argument has continued regardless, because it was never really about the Khazars.

What the sources actually support is narrower and more interesting than any of it: an elite conversion, a court that adjudicated between four communities, a capital where all four had standing and their own judges, and a Muslim guard with a written exemption from fighting Muslims.

The one thing that lasted

The word qaghan came west with them. Rus princes used it of themselves for a while afterwards — it appears for Vladimir and Yaroslav in eleventh-century Kievan texts — which is the ordinary way a conquering power borrows the vocabulary of the state it replaced.

A title, in somebody else's chancery, for about a century. That is the inheritance, and it is a fair measure of how completely a state can vanish while its neighbours go on standing exactly where it put them.