Danishmendid Dynasty

The other Turkish state

Drafted from (eds), The Seljuks of Anatolia: Court and Society in the Medieval Middle East (2013)formation

Anatolia after Manzikert had two Turkish states, and for most of the first century the one nobody remembers was the stronger.

Manzikert did not hand Anatolia to anybody. It removed the Byzantine frontier system, and what moved into the country behind it was not an army under a command but a number of Turkmen groups under their own leaders, each taking what it could hold. Two of them became states. The house of Seljuk took the west and Konya. A commander remembered as Danishmend Gazi — whose origins the sources cannot agree on, making him variously a Turkmen emir, an Armenian, a kinsman of the Great Seljuk sultan and a veteran of Manzikert — took the centre and north-east, with Sivas, Niksar, Tokat and eventually Malatya and Kayseri.

For most of the twelfth century the Danishmendids were the stronger. They captured Bohemond of Antioch in 1100 and held him three years — the sources differ over whether the captor was the founder or the son who succeeded him, which is of a piece with everything else about this dynasty's early generations; a Danishmendid and Seljuk alliance destroyed the crusading army of 1101; and at their height they held more of Anatolia than Konya did, with the caliph reportedly sending Emir Gazi the insignia of a sultan.

What that meant on the ground is best shown by the money. Danishmendid coins carry inscriptions in Greek, and some of them carry a figure on horseback spearing a dragon, which is Saint George. A Turkish Muslim dynasty in central Anatolia was minting in the language of its subjects with the imagery of their religion within sixty years of arriving.

That is what the Turkish settlement of Anatolia actually looked like at the start. Not a replacement of one population by another, but a small armed elite governing Greeks and Armenians and using whatever vocabulary of authority was already lying about.

Tanismanes

Drafted from Komnene, The Alexiad (1928)expansion

Anna Komnene calls him the Sultan Tanisman, which is a Greek princess's ear rendering a Persian word — dānishmand, the learned one — that a Turkmen warlord had taken as a name. She is describing the summer of 1097 and she is describing a defeat.

The First Crusade had beaten Kilij Arslan outside Dorylaeum and was marching south-east across the plateau. Near Herakleia, Anna says, Tanisman and Hasan met them with an army she puts at eighty thousand — her figures are never usable and her point is that this was the main Turkish effort — and there was a fierce contest in which neither side would give way, until Bohemond came off the right wing and rode straight at the sultan, like a lion rejoicing in his strength, as the poet says.

The Turks broke. Anna adds, with the satisfaction of somebody who disliked Bohemond intensely and disliked the Turks more, that afterwards the barbarian power collapsed and the survivors scattered, leaving their wives and children behind, not daring to face the Latins again.

Three years later

She was wrong about the not daring, and she does not tell the story that proves it.

In 1100 Bohemond, by then prince of Antioch, was taken prisoner by the Danishmendids campaigning in the north, and held for three years; the crusade of 1101, raised partly to retrieve him, was destroyed in Anatolia by a Danishmendid and Seljuk alliance — the two Turkish states cooperating, which they did rarely and always against somebody from outside. None of that is in the Alexiad. Anna's Bohemond is a figure she needs upright, as her father's great adversary, and a Bohemond in a Turkish prison for three years is not useful to the book.

The gap is worth noticing on its own account. Within five years, the man she has riding down the Turkish army like a lion was a captive in Niksar and the western army sent to fetch him had been annihilated on the road — and the historian closest to the events, writing in Greek about the Turkish frontier, leaves it out.

The rebel who sent for him

The other appearance the Danishmendids make in the Alexiad says more about their standing than either battle.

Gregory Taronites was the Byzantine duke of Trebizond and he rebelled. He imprisoned the outgoing governor and a number of the leading men of the city; he wrote the emperor a letter of many pages abusing the Senate, the army and Alexios's own relations; and when the emperor's nephew came against him with a force, Gregory withdrew to the fortress of Colonea in order to call Tanismanes to his aid.

That sentence is the measure of the thing. Twenty-five years after the frontier dissolved, a Byzantine imperial governor in revolt against his own emperor looked round for a patron strong enough to protect him, and the man he sent to was the Turkmen emir at Sivas.

He did not get there in time. Alexios's men caught him on the road, knocked him off his horse with spears, and brought him to Constantinople, where his hair and beard were shaved to the skin and he was paraded through the Agora and shut in the tower of Anemas. The emperor's clemency, Anna notes, was such that he was eventually pardoned and loaded with honours.

What the empire had lost

Anna's whole book is an argument that her father held the situation together, and on Anatolia it is an argument she loses in the details.

The Byzantine recovery of the west coast was real and it stopped at the plateau. What lay beyond was a country in which two Turkish dynasties fought each other for the interior, took western princes prisoner and ransomed them, and were solicited as allies by the empire's own provincial governors. The Danishmendids at this point held Sivas, Niksar, Tokat, Malatya and Kayseri — which is most of the ground between the Seljuks at Konya and the Armenian and Georgian marches — and they had got there in a generation.

The city he fed

Drafted from Abu'l-Faraj), The Chronography of Gregory Abu'l Faraj, the Son of Aaron, the Hebrew Physician, Commonly Known as Bar Hebraeus (1932)expansion

Bar Hebraeus's own family came from Melitene, and his account of how the Danishmendids took it is the most useful thing written about them by anybody, because it was written by a Christian who thought the Turkish conqueror had been the best thing to happen to the city in a generation.

The situation first. Melitene was held by Gabriel, a Greek, and every summer the Danishmendid amir came over from Sivas, wasted the country, ate the crops and went home for the winter. Gabriel, unable to stop this, sent three times to the Franks promising them the city. So Bohemond of Antioch set out for Melitene, trusting him.

What happened next is a conspiracy nobody in the standard account of the Crusades is on the right side of. The Armenian lords of the hill country — including one Bar Hebraeus names as Khoj Basil, glossing the title helpfully as thief — were afraid that if the Franks took Melitene they would be pushed out of their own places, so they sent secretly to the Danishmendid asking him to lay an ambush. Gabriel, who had invited Bohemond in the first place, met him at a village above the city and spun the days out to give the Turks time to arrive. The ambush worked. Bohemond was put in fetters and sent to Sivas.

Melitene itself held out until two of its own soldiers, enraged past bearing by Gabriel robbing the people inside the walls, opened it to the Turks on Wednesday 18 September 1102 — a date Bar Hebraeus gives precisely, noting that some Arabic manuscripts say the year before.

And then the passage that is the reason to read him. When the Turks went into the city they plundered all its wealth, because the amir had already assigned the wealth to his troops — with the exception of the inhabitants. He did not leave one soul to perish; he took the people upon himself, sent them back to their own houses, and brought wheat and oxen and other necessaries from his own country and gave them to them. Very many blessings came to Melitene in his days. He appointed a governor over it named Basil, a man who was just and God-fearing.

Gabriel, meanwhile, was handed over to justice by Turks and Christians together, made to remember the murders he had ordered, and taken in front of a fortress where his wife was, and killed.

Three things in that sequence are worth holding.

The first is the Armenians' calculation. Faced with a Frankish crusader and a Turkish amir, the local Christian lords chose the Turk, because the Turk was raiding their crops and the Frank would have taken their castles. That is the whole political logic of eleventh-century Anatolia in one decision, and it is why the Danishmendids lasted eighty years.

The second is the distinction between the wealth and the people. A conquering army was paid out of the city's goods and forbidden the city's population, and the population was then subsidised out of the conqueror's own province until it could feed itself. That is a state behaving like a state, in its second generation, in country it had entered as nomads.

The third is who says so. Bar Hebraeus had no reason to flatter the family that took his home town, and the ruler he describes as wicked in this story is the Christian one.

The insignia from Baghdad

Drafted from (eds), The Seljuks of Anatolia: Court and Society in the Medieval Middle East (2013)peak

Emir Gazi ruled from about 1104 to 1134 and spent thirty years holding off three powers at once, which is not something a Turkmen principality was supposed to be able to do.

To the west was the Sultanate of Rum, contesting the same plateau. To the north and west was Byzantium under John II Komnenos, who campaigned in Anatolia repeatedly in the 1130s with the serious intention of recovering it and who was a considerably better soldier than his reputation suggests. To the south were the crusader principalities of Antioch and Edessa, whose whole northern frontier ran against Danishmendid country.

Emir Gazi fought all three and lost none of them decisively. He took Malatya in 1103, took Ankara and Kastamonu from the Seljuks, beat a crusading army, held the Byzantine campaigns to sieges of individual towns, and at his death the Danishmendids controlled more of Anatolia than Konya did.

The recognition followed. The sources report the caliph in Baghdad sending him the insignia of a sultan — a diploma, a standard, a robe.

That transaction had been running for two centuries by then, from Baghdad outward to whoever actually held power in the provinces: the caliph confirmed what he could not prevent, and the recipient got a piece of paper saying that the head of Sunni Islam regarded him as lawful. It had reached Bukhara, Ghazna and Isfahan. Sivas is about as far west as it ever went. A Turkmen emir in central Anatolia was worth investing, and being invested was worth having.

He is also the last Danishmendid to hold all of it at once. His son Melik Mehmed held it too, and on Mehmed's death in 1142 the realm divided and never reassembled. The dynasty's high point is a single long reign, and everything after it is the same territory administered by people who were not on speaking terms.

Saint George on a Muslim coin

Drafted from (eds), The Seljuks of Anatolia: Court and Society in the Medieval Middle East (2013)aside?What aside means

Danishmendid coins are strange objects, and strange in a way that carries more information than most of the chronicles about them.

They carry legends in Greek. Not Greek alongside a dominant Arabic, but Greek as the principal language on many issues — the ruler's name and titles rendered in Greek letters, sometimes with the Greek word for great emir, sometimes with Byzantine imperial formulae. Some carry Arabic on the reverse and some do not. And a number carry a figure on horseback spearing a serpent or dragon, which is the standard Byzantine image of Saint George.

The obvious reading is that these are coins for Christians to use, and that is part of it: the Danishmendids ruled a population that was overwhelmingly Greek and Armenian and had used Byzantine coin for six centuries, and money that looked unfamiliar would not have circulated.

The less obvious reading is the one the numismatists press, and it is about authority rather than commerce. A ruler putting Byzantine formulae and Byzantine saints on his coinage in the twelfth century is not only accommodating his subjects, he is claiming the thing those images meant — lawful rule over this country, in the vocabulary the country already had. The Seljuks of Rum did the same sort of thing from the other direction, taking Persian throne-names out of the Shahnama while marrying Byzantine princesses and minting with Greek legends of their own.

Set that beside the two great set-piece images of this period. Manzikert and the crusades are wars of religion in the sources on both sides. The coins are what the same decades looked like at the level of a market in Sivas, where a Muslim Turkish emir paid his troops in money bearing a Christian saint, and nobody appears to have found it worth remarking on.

It is the reason objects are worth more than chronicles on a frontier. The chronicles record what people said about each other, in front of audiences who expected them to say it. The coins record what they handed over for bread.

Two branches

Drafted from (eds), The Seljuks of Anatolia: Court and Society in the Medieval Middle East (2013)contraction

Melik Mehmed died in 1142 and the realm divided between two branches, one at Sivas and one at Malatya. That is the whole of the contraction; everything else follows from it.

They did not reunite, and the reason they did not is that the division suited each of them better than a reunion would have. A Danishmendid at Malatya facing Nur al-Din of Aleppo had a different set of problems from a Danishmendid at Sivas facing Byzantium and Konya, and each could get better terms from his own neighbours by dealing separately than either could by subordinating himself to a cousin.

So each dealt separately. With the Seljuks. With Manuel Komnenos, who campaigned in Anatolia through the 1140s and 1150s with the serious intention of recovering it. With Nur al-Din, who was assembling northern Syria in exactly these decades and had an interest in a weak neighbour on his northern flank.

Every one of those relationships was rational for the branch that entered it and fatal for the dynasty as a whole. Three larger powers were each able to buy one Danishmendid against another, and none of the three had to fight both at once.

The arithmetic

It is worth stating what was lost, because the state that split in 1142 was not a small one.

Emir Gazi at his death held Sivas, Niksar, Tokat, Amasya, Kayseri, Ankara, Kastamonu and Malatya — more of Anatolia than Konya did, with the caliph's diploma to say it was lawful. Thirty-six years later there was nothing.

Kilij Arslan II took the Sivas branch by 1175 and Malatya in 1178. He did it in pieces, over years, without a decisive battle, because there was no longer a Danishmendid army to fight — only a succession of towns each held by somebody with his own arrangement to protect, most of whom eventually judged that Seljuk service was a better arrangement than the one they had.

They were not wrong about that either. The surviving Danishmendids went into the Seljuk administration as commanders and provincial governors and their descendants are visible there for generations. The family did well out of the end of its own state, which is the ordinary result of a dynasty being absorbed rather than destroyed and is the reason so few of them fight to the last.

1178

Drafted from (eds), The Seljuks of Anatolia: Court and Society in the Medieval Middle East (2013)end

Malatya fell to Kilij Arslan II in 1178 and there was one Turkish state in Anatolia instead of two.

There is no battle to date it to and no last stand. The end of the Danishmendids is a series of towns changing hands over about fifteen years, with the men holding them transferring their service, and it is a very ordinary way for a medieval dynasty to disappear. What makes it worth setting down is what was lost with it, which is not territory.

The other outcome

For a century Anatolia had two Turkish states and it was not obvious which of them would end up as the country.

The Danishmendids held the richer half. They had the caliph's investiture and Konya did not. They beat crusaders, took a prince of Antioch prisoner, fought Byzantium to a standstill under its most capable twelfth-century emperor, and were solicited as allies by the empire's own rebel governors. Had the succession of 1142 gone differently there is no obvious reason the peninsula would not have been organised from Sivas — in which case the road to the Aegean would have run the other way, and the frontier principality on the Byzantine march that eventually produced the Ottomans might never have had a reason to exist.

That is speculation and it should be marked as such. What is not speculation is that the outcome turned on an inheritance rather than on any of the things the two states were fighting about.

What Rum inherited

A working country. The Danishmendid cities — Sivas, Tokat, Amasya, Kayseri, Malatya — became the eastern half of the Seljuk sultanate and stayed its economic centre, sitting on the routes east to Tabriz and south to Aleppo. The Danishmendid administrative and military class became Seljuk officers. The caravan roads that Rum is remembered for run through territory it took from this dynasty.

It is the second time this had happened in a century in the same country and by the same method. The Turkish states in Anatolia were built on top of a Byzantine one whose towns, taxpayers and roads they kept using, and then one of them was built on top of the other.

And a book

One thing outlasted the state, and it was still being read two hundred years later by people doing the same job on a different frontier.

The Danishmendname is a Turkish epic romance, written down in the fourteenth century out of older oral material, which turns Danishmend Gazi into the archetypal frontier warrior: a holy fighter on a Christian border, with named companions, sieges, single combats and conversions. It is not history and was never offered as history. It is a model of conduct.

Its audience is the point. It was read most enthusiastically in the fourteenth-century Turkmen beyliks and at the early Ottoman court — men on the Aegean and Byzantine marches who were doing in their own generation what the book described, and who wanted a template with a pedigree.

A dynasty that lost its state in 1178 supplied the vocabulary in which the men who eventually took the rest of Anatolia understood themselves.

The road in 1332

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

A hundred and fifty years after Malatya fell, a Moroccan lawyer walked through what had been Danishmendid country and wrote down what was there.

He came up from the coast into the interior and went, in order, to Aksaray — which he calls one of the finest districts of Rum — then Niğde, then Kayseri, then Sivas, then Amasya, then Sonusa, then Gümüş, then Erzincan, then Erzurum. Every one of those places is Danishmendid country. Every one of them, he says, is subject to the king of Iraq: the Ilkhan at Tabriz.

Sivas he found large, and now the rendezvous of the greater part of the king's army. The city Emir Gazi had ruled from, that a Byzantine duke in revolt had sent to for protection, was the assembly ground of a Mongol force based eight hundred miles away.

The brotherhood at Erzurum

What Ibn Battuta actually found holding those towns together was not a state at all.

In every Turkmen town, he says, there was a brotherhood of young men — the akhis — and no people on earth were more courteous to strangers, more ready to supply them with food, or more opposed to oppressors. A man of some trade, around whom the young men of that trade and any destitute strangers collected, would be made their president; he built a cell and put a horse and a saddle in it; in the evening the members brought in everything they had earned that day and it was sold to buy food and fruit for the house. A traveller arriving was taken in and kept until he left the country.

At Erzurum, Ibn Battuta met one of these brothers who he was told was over a hundred and thirty years old, still in possession of all his faculties and able to walk wherever he wished. He took his blessing and went on.

The number is not usable and the institution is. When Anatolia's states failed — first the Danishmendids, then the Seljuks, then the Ilkhanid administration itself — what kept the towns functioning was a network of craft brotherhoods that fed strangers and stood against the powerful, run out of a room with a horse in it.

That outlived both sultanates. It is also the same answer, in a different key, as the frontier epic: both were ways of organising men where there was no government worth the name, and Anatolia spent a great deal of its history in exactly that condition.