Great Seljuk Empire

Dandanaqan

Drafted from contributors, Encyclopaedia Iranica · sourceformation

The Seljuks enter this thread from outside it. Every previous polity here grew out of an existing administrative structure — a governorship, a guard corps, a mercenary contract. The Seljuks were Oghuz Turkmen who crossed the Oxus as pastoralists looking for grazing, with a chief's family that had converted to Islam a couple of generations earlier and a following that was not, in any sense the Ghaznavid chancery would recognise, an army.

Mas'ud of Ghazna spent the 1030s trying to manage them as a nomad problem: grant pasture, punish raids, move them on. It did not work, because the thing he was negotiating with had no address.

In 1040 at Dandanaqan, in the desert near Merv, a Ghaznavid army that had been marched around Khurasan without adequate water met the Seljuks and broke. Tughril Beg was proclaimed ruler of Khurasan.

One battle transferred the richest province in the eastern Islamic world from a dynasty at the height of its administrative sophistication to a confederation that had none. It is about as sharp a discontinuity as the region produces, and it is worth noting that it was not the Ghaznavid state that failed — that state went on functioning in India for another century and a half. It was that a settled army lost a battle to people it had regarded as a logistical nuisance.

The followers they could not hold

Drafted from Bosworth, The Cambridge History of Iran, Volume 5 (1968)expansion

The Seljuk family won at Dandanaqan with an army of Oghuz Turkmen and spent the next fifty years trying to work out what to do with them.

The problem is structural and it is the opposite of every other military problem in this thread. The Samanids, Ghaznavids and Buyids had armies they had bought or hired, which could be paid, quartered and — as those pages all record — eventually turned on their employers. The Seljuks had a following. Turkmen pastoralists had come across the Oxus with their households and their flocks because the grazing was better and the leadership was Seljuk, and they expected pasture, plunder and a chief who behaved like one. They were not soldiers in a chancery's sense and could not be discharged.

What a settled state needs from that is exactly what it cannot have. Tughril and his successors set about becoming the thing they had beaten: a Persian bureaucracy, a tax base in irrigated agriculture, an alliance with the urban Sunni establishment, and cities that had to be protected from raiding — which meant protected from their own following. Within a generation the sultans were recruiting slave soldiers of the usual kind and paying them with revenue assignments, because a professional guard obeys and a tribal confederation negotiates.

The Turkmen were therefore surplus, and they went where the grazing and the raiding were: west, into the Armenian highlands and the Byzantine frontier districts, in numbers and largely on their own initiative. Seljuk policy was less to direct them than to point them at somebody else and be relieved when they went.

That is the background to Manzikert, and it is why the battle mattered so much more than its scale. The Turkmen were already on the frontier in force. What Manzikert removed was the thing that had been keeping them out of the country behind it.

The same room, a different man

Drafted from Bosworth, The Cambridge History of Iran, Volume 5 (1968)expansion

In 1055 Tughril entered Baghdad, arrested al-Malik al-Rahim, and ended a hundred and ten years of Buyid control. The caliph al-Qa'im received him and granted him the sultanate.

Read that next to 945 and almost nothing has changed. A military power from outside Iraq takes the city, leaves the caliph in place, and receives from him a formal grant of the authority it already possesses. Same building, same office, same transaction.

Two things had changed, and both mattered. The new masters were Sunni, so the grant could be presented as a restoration rather than a captivity — the caliph delivered from heterodox amirs by an orthodox sultan. And the title was sultan, which under the Seljuks stopped being an honorific attached to a particular strongman and became a standing office, with a chancery, a succession, and a defined relationship to the caliphate: the caliph holds religious authority, the sultan holds temporal power, and each legitimates the other.

That division of labour outlived the Seljuks by centuries and was copied by almost everybody who came after them. It is the most durable thing the dynasty produced, and it is not a piece of territory.

Two thrones, one of them lower

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 watched the arrangement being made physical, and described the furniture.

Tughril rode to the caliph's palace and waited outside on his horse until the gate was opened. He crossed the courtyard; the caliph's nobles came out to meet him; slaves lifted one curtain and then another, and behind the second al-Qa'im was sitting on a throne raised about seven cubits off the ground, in a black headcloth and a black cloak — the Abbasid colours, as Bar Hebraeus is careful to note, the Fatimids in Egypt having taken white — with a gold sceptre in his hand and a eunuch on either side. The sultan bowed to the ground and did homage several times. Then the caliph gave an order and the slaves raised him onto a second throne, lower than the first, about the height of a standing man.

The caliph did not speak to him directly. He spoke to a dragoman, who said to the sultan: the Commander of the Faithful praises your solicitude, applauds your labours, is comforted by your presence and rejoices in you; and behold, he hands over to you all the provinces and countries which God has handed over to him; therefore fear God, do justice in your dominion, and give thanks. The sultan stood, bowed to the ground again, and said that he was the slave and the subject of that command.

Then he was dressed. Seven black robes, one over another; a black headcloth shot with gold; on top of that a crown set with two jacinths, each with fifteen very fine pearls above it, and gold bands at the ends. The collar of gold he would not put round his neck: it was laid on his shoulders instead, and a eunuch carried it in front of him. When he wanted to prostrate himself a third time he found he could not, because the crown was too heavy for him to bend. He was given three purple banners lettered in gold, and a royal seal, and the heralds proclaimed him King of the East and the West. On leaving he asked the caliph to put out his hand, kissed it twice, and passed it across his forehead.

Everything in that room is doing the same work. One throne is higher and one is lower; the higher man does not address the lower man in his own voice; the lower man is given the provinces in a sentence that says God gave them to the man handing them over; and the lower man is dressed in so much of the higher man's authority that he physically cannot bow again.

And the man on the lower throne had the army, and both of them knew it. The arrangement worked for two centuries precisely because neither party tested it. The caliph's grant was the only thing that made a Turkmen chief the lawful ruler of Iraq, and the sultan's soldiers were the only thing that kept a caliph in Baghdad at all. Bar Hebraeus, who was a Christian bishop with no stake in either of them, describes the ceremony as a foreigner describes a rite: attentively, and without saying what it meant.

Manzikert

Drafted from Bosworth, The Cambridge History of Iran, Volume 5 (1968)peak

On 26 August 1071, near Manzikert north of Lake Van, Alp Arslan met a Byzantine army under the emperor Romanos IV Diogenes and destroyed it. Romanos was captured — the first Roman emperor taken alive by a Muslim army — brought before the sultan, treated with public courtesy, given terms, and sent home.

Alp Arslan had not wanted the battle. He was campaigning towards Syria and Fatimid Egypt, which was the direction the Seljuk state's ambitions actually ran; the Byzantine expedition was an interruption he turned to deal with. He died the following year in Transoxiana. Nothing in Seljuk policy after 1071 suggests that anyone in Isfahan regarded Anatolia as the prize.

The consequences are nonetheless about as large as any single day's in the history of the region.

What broke was not the Byzantine army so much as the Byzantine settlement. Romanos returned deposed and was blinded; the empire spent the following decade in civil war, and the rival claimants hired Turkmen bands against each other and opened the passes to do it. The frontier system that had held the Anatolian plateau — the themes, the local levies, the border lords — dissolved in that decade rather than on that afternoon, and the Turkmen who had been pressing on it since the 1050s moved in behind. Within twenty years there was a Turkish sultanate at Nicaea, a hundred miles from Constantinople — the Sultanate of Rum, which lasted two hundred and thirty years.

Three very large things run through the gap that opened here. Byzantium, having lost the manpower and revenue of the plateau, appealed west and got the First Crusade, which arrived in 1095 and took Jerusalem in 1099 from the Fatimids — recorded on their page as the loss of the coast. Anatolia became Turkish-speaking and Muslim, which it had not been, and remained so. And the Seljuk dynasty acquired a branch in the west that outlived the trunk by a century.

None of that was the plan. The sultanate that won at Manzikert was facing the other way.

Nizam al-Mulk's twenty years

Drafted from Bosworth, The Cambridge History of Iran, Volume 5 (1968)peak

Under Alp Arslan and then Malik-Shah, from 1063 to 1092, the empire was administered by one Persian vizier, Nizam al-Mulk, and the peak of the Great Seljuk state is largely his.

Two of his instruments are the reason the dynasty left anything behind at all. The iqta system assigned the revenue of a district to a commander in place of pay, which solved the problem of financing a Turkmen cavalry army out of a settled agrarian economy — and became the standard fiscal-military arrangement across the Islamic world for the next several centuries. The Nizamiyya madrasas, endowed at Baghdad, Nishapur and elsewhere, put state funding behind Sunni jurisprudence and produced a trained, salaried, institutionally connected class of Sunni scholars. Al-Ghazali taught at the Baghdad Nizamiyya.

That is the difference between holding a faith and carrying one. The Seljuks did not convert a population; they built the institutions through which a specific Sunni establishment was propagated and reproduced, in a period when the Fatimids were doing the same for Isma'ilism from Cairo. The two programmes were consciously aimed at each other.

He also wrote the Siyasatnama, a book of statecraft for Malikshah, in the last years of a ministry that ended badly.

The turban and the crown

Drafted from Browne, A Literary History of Persia, Volume II: From Firdawsi to Sa'di (1906)peak

Nizam al-Mulk was dismissed before he was killed, and the dismissal is the part worth having, because it explains the twenty years better than the thirty do.

The pressure came from inside the household. Malikshah's favourite wife, Turkan Khatun, wanted the succession for her small son Mahmud; the vizier was known to favour the elder Barkiyaruq, a boy of twelve or thirteen. Around this ran the ordinary complaint of a long ministry, which was that the vizier's relations held the best posts in the empire. One of his grandsons, governing Merv, behaved badly enough that a victim took the case to the sultan directly, and Malikshah sent Nizam al-Mulk an angry message asking, with heavy irony, whether he was his partner in the throne or his minister.

The old man answered rashly and it became the most quoted sentence of the reign: he who gave you the Crown placed the Turban on my head, and the two are bound together inseparably. Browne, who is more careful with such things than the authors he is reading, notes that the words come from the Rahat al-Sudur and that the fullest account of the quarrel is Ibn al-Athir's, and observes that a remark like that, reported to a sultan with the usual improvements, is not one its author would have made in a calmer hour. Nizam al-Mulk was dismissed in favour of Taj al-Mulk, Turkan Khatun's man, along with a set of other replacements that a court poet promptly turned into a lament: if you were tired of Nizam and Kamal and Sharaf, see what Taj and Majd and Sadid have done to you.

He did not long survive it. On 14 October 1092, travelling with the sultan from Isfahan to Baghdad, he halted near Nahavand — where the last Sasanian army had been destroyed four and a half centuries earlier — broke his fast after sunset, and was walking to his family's tents when a young Daylami came up to him in the posture of a petitioner and stabbed him. The killer belonged to the new order of fida'is that Hasan-i Sabbah had just instituted, and this is generally called their first stroke, though Ibn al-Athir records an earlier murder of a muezzin at Isfahan and thinks the vizier's execution of a carpenter suspected in it was what marked him. It was also widely said that Taj al-Mulk, the rival who had replaced him, had arranged it; if so the arrangement did not pay, because Nizam al-Mulk's servants killed Taj al-Mulk four months later.

Malikshah outlived him by three weeks. He went hunting on 6 November, either took a chill or ate something that disagreed with him, was bled, developed a fever and died on the 19th, not yet forty. Bar Hebraeus, writing from other sources, says his own slave gave him poison. The poet Mu'izzi produced the couplet everyone remembers — one month the aged minister was translated to heaven, the next month the young king followed him, and what impotence the power of God brings upon earthly power — having already written, on the dismissal, that the sultan had set the accursed Taj over his dominions and jeopardised for his sake both Crown and State.

Browne then does the thing he is most useful for. There is a famous set of verses supposed to have been written by the dying vizier and sent to Malikshah: thirty years cleansing the earth of tyranny, a warrant signed by the sultan to carry to the Angel of the Throne, ninety-six years of life ended by violence near Nahavand, and a son commended to the king's care. Browne shows it is fabricated. The last verse is demonstrably by Burhani, Malikshah's own laureate, written to recommend his son Mu'izzi to the royal favour; the Chahar Maqala says Nizam al-Mulk thought little of poets because he had no skill in the art; he was about seventy-five, not ninety-six; and his sons had, as everyone complained, already been given more than enough. A minister who had spent three decades making the state's institutions was handed, within a century of his death, the deathbed poem of somebody else.

The book of government

Drafted from Bosworth, The Cambridge History of Iran, Volume 5 (1968)aside?What aside means

Nizam al-Mulk wrote the Siyasatnama — the book of government — for Malik-Shah towards the end of his thirty years running the empire, and there is almost no other medieval state whose chief minister left this kind of account of how the thing actually worked.

It is a mirror for princes — the same genre Yusuf of Balasaghun was working in at Kashgar a few years earlier, in Turkic — and it belongs to a tradition reaching back through Persian and Arabic advice literature to Sasanian models. What makes it different from most of the genre is that it was written by the man who was actually doing the job, about the machine he had actually built, for the sultan he had actually managed.

So it is specific where the genre is usually pious. It explains how to run a postal and intelligence service and why a ruler who does not have one will be lied to. It explains that assignment-holders must be inspected because a man collecting his own revenue will ruin the district, which is exactly what had happened to the Sawad under the Buyids a century earlier, and which he systematised anyway. It argues for salaried judges, for audiences at which subjects can reach the sultan without an intermediary, and for keeping soldiers of several different origins so that no one group can dictate. It also devotes long stretches to heretics — meaning, in the 1080s, the Nizari Isma'ilis of Alamut, one of whose emissaries is traditionally credited with killing him in 1092.

The strain in the book is the interesting part. Nizam al-Mulk was a Persian administrator serving a Turkish military family, and he is arguing throughout for the bureaucratic, sedentary, tax-collecting order against the instincts of the people he served: for records against improvisation, for salaried officers against tribal claims, for the Sunni scholarly establishment his madrasas produced against everything else. He is not describing the Seljuk state so much as recommending it.

He was assassinated in 1092 and Malik-Shah died a month later. The empire never again had a single undisputed ruler, and the arrangement in this book was running the Islamic world for the next four centuries anyway — which is the peculiar thing about the Great Seljuks. Their state lasted about fifty years in working order. What their vizier wrote down about how to run one is still in print.

A child on the throne

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)contraction

Turkan Khatun got what she had wanted and it lasted about two years.

Bar Hebraeus, who calls her a woman of intelligence and traces her descent to the first king of the Turks, says she administered the kingdom herself after Malikshah's death and had her son Mahmud proclaimed sultan at Baghdad at the age of five. He preserves what the court made of the boy. The caliph's men dressed him in royal apparel and set him on the throne, and he put out neither a hand nor a foot, did not move his eyes, did not bend or salute, and moved no part of himself at all; he sat so exactly like a stone that everyone present marvelled at him. It is offered as evidence of the child's intelligence. It reads now as a description of a frightened five-year-old who had been told not to move.

Other Turks raised his half-brother Barkiyaruq, a son of Malikshah by his cousin. Barkiyaruq had ten thousand men and Mahmud twenty thousand, and when the two armies met, Mahmud's nobles crossed over in the middle of it. Then the dead sultan's brother, Tutush, who held Mosul, Nisibis, Mayyafariqin, Syria and Damascus, came up into Azerbaijan; his nephew went to meet him and the two of them talked in a village near Tabriz called Sa'idabad, after which Tutush went back to Damascus and two of his amirs took service with the nephew. Barkiyaruq went on to Baghdad, where the caliph dressed him in royal apparel, wrote him a patent and proclaimed him sultan — the same ceremony, in the same room, for the third time in forty years, performed now for one claimant among several while the others were still in the field.

In 1094 Turkan Khatun died, and her son had nothing left but Isfahan. Barkiyaruq came to the city with a small escort; Mahmud's officers shut the gates against him, then reconsidered, decided it would be better to let him in and seize him, and opened them. He had been inside one day when his seven-year-old brother caught a violent fever and died of it the same day. So the officers made peace instead and proclaimed the survivor king of Isfahan.

Bar Hebraeus reports the fever without comment, which is his habit and is probably the right treatment. What the sequence establishes is not who poisoned whom but what the succession had become. A sultanate that had been handed from Tughril to Alp Arslan to Malikshah without a war was now decided by which faction's amirs changed sides during the battle, and by whether a child in a besieged city lived through the week.

Barkiyaruq held the title for twelve more years and spent almost all of them fighting relations for it. What his father had ruled was never again ruled by one person.

The period of Sanjar

Drafted from Browne, A Literary History of Persia, Volume II: From Firdawsi to Sa'di (1906)contraction

After the twelve years of fighting that followed Malikshah's death, what the house had instead of an undisputed ruler, for sixty-five years, was Sanjar.

He is the most consequential figure the dynasty produced after Malikshah and the least written about, because his career is a slow one and ends badly.

What he actually did

Sanjar was made king of Khurasan by his brother Barkiyaruq, held it from 1096, and was senior in the house from 1118 until his death in 1157. Browne, reading the Seljuk chroniclers, reports Rawandi's assessment plainly: in length of life and in brilliant achievement Sanjar surpassed all the other Seljuk monarchs.

The record supports it. In forty years he made nineteen conquests. He took Ghazna and installed Bahramshah on its throne, on terms of a thousand dinars a day. He captured Ahmad Khan, the Karakhanid ruler of Samarkand, in 1130. He subdued Sistan and Khwarazm. Khurasan, through the fratricidal wars that convulsed the rest of the family, stayed largely out of it, because he governed it firmly.

And none of that arrested the decline, which is the point.

Why not

Browne lists the reasons and they are structural rather than personal.

The fratricidal wars at the beginning of the period had already dispersed the authority. The revolts of turbulent amirs were constant. And two or three rival powers were always in the field: the kings of the mountains of Ghur in the east, whose rising power would shortly finish the house of Ghazna; and, from Atsiz's accession in 1127, a new dynasty of Khwarazmshahs who became formidable rivals inside a generation — and who were the Seljuks' own appointees.

Then in 1141 Sanjar took an army east against the Qara Khitai and was destroyed at the Qatwan steppe outside Samarkand. Transoxiana passed under the protection of a Buddhist steppe empire, and the news of a Muslim sultan beaten by a non-Muslim power from beyond the Oxus travelled west and became, in Europe, the story of Prester John.

The Ghuzz

The end is worse than the defeat.

In 1153 Sanjar was beaten by the Ghuzz — Turkmen of the same stock his own family had come from a century and a quarter earlier, and whom his officials had been mishandling. Tus and Nishapur were sacked, and many of their inhabitants killed, including, Browne notes, some of the most celebrated of them for learning and piety.

Sanjar himself was taken. For three years he was practically a prisoner in the hands of the Ghuzz: treated, outwardly, with some respect, and unable to go where he wished or to protect his own people while the province he had governed for sixty years was ruined around him.

In the autumn of 1156 a few of his old retainers bribed his custodians and got him out, and brought him safely to Merv, where he began to collect an army. Grief at the desolation of his country, with old age, killed him a few months later. He was buried at Merv, in a building he had put up himself, beside his grandfather Alp Arslan.

What was left

Browne's judgement is that with Sanjar's death the epoch of the Great Seljuks came to an end, and the political record bears him out: what continued for another thirty-seven years was a title, contested among cousins in western Iran, while the atabegs held the provinces and Khwarazm grew.

The devastation the Ghuzz did to Khurasan was frightful, Browne says, and was eclipsed some seventy years later by the Mongols. That is the honest scale of it: the worst thing to happen to eastern Iran in the twelfth century, and a rehearsal.

A family and not a state

Drafted from Bosworth, The Cambridge History of Iran, Volume 5 (1968)end

The Seljuks held their empire as a family possession. Territory was appanaged to brothers, sons and nephews, each with an atabeg — a Turkish commander appointed as guardian to a young prince, running the province in his name.

The arrangement worked while there was a strong sultan to arbitrate. After 1092 there was not, and the mechanism inverted with total predictability: the atabegs kept the provinces and the princes became the formality. The Zangids in Mosul and Aleppo, the Eldiguzids in Azerbaijan and the Salghurids in Fars all begin as guardians of Seljuk children.

Meanwhile the eastern frontier produced the same story in its purest form. The Seljuk sultans appointed a governor to Khwarazm; the office became hereditary; by the 1190s the Khwarazmshahs were stronger than the sultans who had appointed them. In 1194 Tughril III, the last Great Seljuk, was killed in battle against Ala al-Din Tekish of Khwarazm.

It is a rare way for a state to die. The Ghaznavids were ended by a vassal house they had not made; the Samanids by a slave-general who had gone east on his own initiative. The office of Khwarazmshah was a Seljuk appointment from beginning to end, and it took the sultanate's titles along with the sultanate.

Two branches outlived the centre — Rum in Anatolia until 1308, and Kirman — which is why the sultanate went on being a thing that could be claimed long after there was no longer anyone to claim it from.

Three schoolfellows

Drafted from Browne, A Literary History of Persia, Volume II: From Firdawsi to Sa'di (1906)afterlife

The most famous story about the Seljuk age is that Nizam al-Mulk, Omar Khayyam and Hasan-i Sabbah were at school together.

It appears in the preface of almost every edition of FitzGerald's Rubaiyat, which is how most of the English-speaking world encountered Persian history at all. It is a very good story. It is also, on the dates, impossible.

The story

Browne prints the oldest surviving form of it, from Rashid al-Din's universal history, composed in the first quarter of the fourteenth century — two hundred years after all three men were dead:

Now the cause of the enmity and mistrust which existed between the Nizam al-Mulk and Hasan-i Sabbah was that they and Omar Khayyam were at school together in Nishapur, and there, in boyish fashion, conceived for one another a devoted friendship which culminated in their partaking of each other's blood and registering a solemn vow that whichever of them should attain to high rank and lofty degree should protect and help the others.

Nizam al-Mulk becomes prime minister. Omar comes to claim the promise, is offered the governorship of Nishapur, and refuses it — being, the text says, a great man and withal a philosopher and a man of sense — asking instead for a pension, and receiving ten thousand dinars a year from the treasury of Nishapur without deduction. Then Hasan-i Sabbah arrives to claim his share, and the rest of the story is the intrigue that ends with the vizier assassinated by the grand master's men.

Why it cannot be true

Nizam al-Mulk was born in 1017. Hasan-i Sabbah and Omar Khayyam both died between 1123 and 1124, at ages nobody records.

For the three of them to have been schoolfellows, both of the younger men would have had to reach a hundred — and even then, Browne points out, they would still have been considerably younger than the vizier, who appears to have finished his education and entered public life early. The odds against two named individuals both living to a hundred are not good. The objection was pressed by Houtsma and is, on any reading, fatal.

Houtsma's own suggestion is the elegant part. The schoolfellow was probably not the famous Nizam al-Mulk at all, but Anushirvan b. Khalid — a later and much less famous minister, to a Seljuk prince who reigned from 1117 — whose own chronicle says, in the passage where the Assassins first appear, that in his youth he had studied with some of their chief leaders, and particularly with a man of Ray who travelled about the world and whose profession was that of a secretary.

That is Hasan-i Sabbah, described by somebody who had actually known him. Two centuries later the anecdote had been transferred to the celebrated vizier, and a third schoolboy had been added.

What that is an afterlife of

This is the fifty-year state's real inheritance, and it is not the roads or the madrasas.

Everything about the Great Seljuks that a general reader knows arrives through later Persian writers with a shape to impose: Nizam al-Mulk as the wise Persian vizier, Hasan-i Sabbah as the sinister organiser, Omar Khayyam as the sceptical astronomer who wanted a pension rather than power. It is a very satisfying triangle — administration, subversion and detachment, the three things one can do about a state — and it was assembled two hundred years afterwards out of an anecdote about a different man.

The empire itself was a family holding that came apart within a generation of its best sultan. What survived it in working order was the vizier's book about how to run one. What survived it in general circulation was a story about three boys in Nishapur, which is not true, and which is the version that reached England.