Abbasid Caliphate

What the revolution actually changed

Drafted from Kennedy, The Prophet and the Age of the Caliphates: The Islamic Near East from the Sixth to the Eleventh Century (2004)formation

The Abbasids came to power in 750 on a Khurasani army raised in the name of an unnamed member of the Prophet's family, and the first thing they did with it was disappoint everyone who had been imagining a different name. The second was to have Abu Muslim, who had organised the whole thing, killed in 755.

What changed was the centre of gravity. The Umayyads had ruled Syria-first, through Syrian tribal levies, facing Byzantium. The Abbasids ruled from Iraq, through a Khurasani army and a salaried administration, facing east. Baghdad — founded in 762 on an empty site chosen for its rivers and its distance from Kufa's politics — was a purpose-built imperial capital in a way Damascus never was, and within fifty years it was probably the largest city west of China.

The other change was who could hold office. The Umayyad settlement had an Arab military class at the top of it and everyone else attached sideways as clients. The Abbasid administration was staffed substantially by Persians, most famously the Barmakid family of viziers, and it ran on a Sasanian bureaucratic inheritance that the Umayyads had used but never absorbed. The conquest aristocracy stopped being the state.

It is also the arrangement the next four centuries in the east run on. The Tahirids, the Saffarids, the Samanids, the Ghaznavids and the Seljuks all hold their provinces by an Abbasid grant, or claim to, and all name the caliph in the Friday sermon. None of them is a successor to the Abbasids and none ever said it was. They are delegates who outlast the delegation, which is a different and stranger relationship, and it is the one that keeps the office alive for five hundred years after it stops governing anything.

The round city

Drafted from Strange, Baghdad during the Abbasid Caliphate, from contemporary Arabic and Persian sources (1900)expansion

The dynasty needed a capital. Kufa had made and unmade caliphs for a century and could not be trusted; Damascus was the losing side's city; the two provisional seats the first Abbasids tried had already produced an insurrection. In 762 al-Mansur chose a site on the west bank of the Tigris, at the point where the canal system of the Euphrates comes closest to the river, and built something nobody in Islam had built before.

Cotton soaked in naphtha

Craftsmen and labourers were brought from Syria, Mosul, the Jibal and Babylonia, with architects and land surveyors, and four overseers were put over them. One of the four was Abu Hanifa, the jurist whose school is the oldest of the four Sunni traditions of law — and what he is credited with on the site is a piece of administration: instead of counting the bricks one by one, he measured the stacks with a graduated rod and computed the number.

The plan was laid out on the ground in lines of cinders, and to make the outline legible they set balls of cotton soaked in naphtha along it and set them alight. The foundations were dug on the burning line.

The shape

What went up was a circle. Two great concentric ramparts with a water-filled ditch outside them, and a third wall inside enclosing the central area — four gates in each ring, equidistant, so that four highroads ran from the centre like the spokes of a wheel and passed through five gateways each before crossing the ditch.

The gates were named for where their roads went: Kufa to the south-west and Basra to the south-east, both opening on the Sarat canal; Khurasan to the north-east on the Tigris, at the main bridge of boats; Syria to the north-west, on the Anbar road. In the centre stood the caliph's palace, called the Golden Gate, and beside it the great mosque, and nothing else at first.

The Arabic writers say what the point was, and al-Mansur said it himself: the sovereign should live at the centre of everything and equidistant from all of it. A circular city with four gates was a novelty in Islam, and probably came from Persia. Every city the conquerors had founded until then had grown out of a camp. This one was a diagram.

What the measurements do not agree about

Nothing of it survives, and the figures in the sources vary in the way figures do.

Ya'qubi gives five thousand ells from gate to gate, about 2,500 yards, which makes the outer circle around the ditch nearly 3,200 yards across; the central area alone was some 2,000 yards, over a mile. Yaqut gives an Arab mile between gates, which is close enough. Khatib al-Baghdadi reports the caliph al-Mu'tadid, a century and a half later, pointing out the old limits as covering two Arab miles in every direction — and elsewhere Khatib passes on a tradition that would make the city oval and another that halves the diameter, neither of which anybody else supports.

The bricks are more certain than the plan. They were sun-dried and enormous: cubes an ell — eighteen inches — on every side, weighing two hundred ratls, and half-bricks eighteen inches square and nine thick at a hundred.

Within fifty years the round city had been besieged by one caliph's army against another's and was largely a ruin, and Baghdad had spread across the river and around it into the sprawling two-banked city that is meant when the name is used. The diagram lasted no time at all. The place it was drawn on has been a city ever since.

Harun al-Rashid, and what came after him

Drafted from El-Hibri, The Abbasid Caliphate: A History (2021)peak

Harun al-Rashid's reign (786–809) is the one everybody has heard of, and the reputation is a problem for reading it: the Thousand and One Nights made him a figure of legend three centuries later, and the legend has been read backwards into the reign ever since.

What the reign was, in institutional terms, is the point at which the Abbasid state was wealthy and coherent enough to be administered by professionals — the Barmakid viziers, until Harun destroyed them in 803 for reasons the sources report differently and none of them convincingly. It is also the point at which the caliphate was already delegating: Ifriqiya was handed to the Aghlabids as a hereditary governorship in 800, in exchange for tribute and the khutba. The empire was being converted into a system of clients while it was still at its height.

Harun's succession arrangement is what broke it. He divided the inheritance between two sons — al-Amin in Iraq, al-Ma'mun in Khurasan — with terms deposited at the Ka'ba, and the arrangement lasted until he died. The fourth fitna that followed was not a frontier revolt or a sectarian rising. It was the empire's two halves fighting each other with their own armies, and it ended in 813 with Baghdad besieged and sacked by Khurasani troops, and al-Amin killed.

Al-Ma'mun won, and won as a Khurasani candidate. He stayed in Merv for six years before returning to a capital that had learned its garrison could be beaten by the provinces. Every subsequent Abbasid attempt to solve the army problem starts from that fact.

The first siege

Drafted from Strange, Baghdad during the Abbasid Caliphate, from contemporary Arabic and Persian sources (1900)peak

Harun died in 809 having divided the empire between two sons. Within three years one of them was besieging the other inside the city their great-grandfather had built, and the Round City did not survive it.

Al-Amin held Iraq and the capital. Al-Ma'mun held the east and preferred, Le Strange notes drily, to remain safely in far-off Khurasan; he sent two generals instead. Harthama came at Baghdad from the east. Tahir — who would found the Tahirid dynasty on the strength of what he did here — crossed the Tigris at Ctesiphon, came up the great Kufa road, and invested the western side.

What a siege did to a city with no walls

East Baghdad had no town wall at all. Its people built barricades across the roads where they left the city and used the line of their own houses and garden walls as the outer defence.

On the west Tahir set up outside the Anbar Gate and began on the suburbs. The Harbiya quarter was largely brought down by his catapults, and the ruin ran from the Tigris bank at Baghtiyin round past the Syrian Gate to the Kufa Gate and the line of the Sarat canal. Fire finished what the catapults began. The great mills where the two Sarat canals met were partly destroyed. The suburbs along the Karkhaya canal were laid in ruins.

Then it simply went on. The siege dragged from month to month and the inhabitants, in Le Strange's phrase, suffered horribly. Zubayda — Harun's widow, the most famous woman of the age — was driven out of her own palace near the Qatrabbul Gate and went to join her son inside the Round City.

The methodical part is the worst of it. Tahir moored a fresh bridge of boats upstream so he could move reinforcements between the two armies. He destroyed the Old and the New bridges over the Sarat, cutting the highroads out of the Kufa and Basra gates. When Harthama stormed the Khurasan Gate and took the eastern half of the city, his first act was to cut the main bridge of boats, so that the Round City was isolated.

The end of it

The merchants of West Baghdad began to parley when they saw the defence failing, and al-Amin's troops deserted in increasing numbers.

He retired with his mother into the Palace of the Golden Gate at the centre of the Round City, keeping a way out to the river through the Khuld palace and its gardens. Then Harthama planted catapults in the Mukharrim quarter across the water, the stones made the Khuld untenable, and its garrison abandoned it.

Al-Amin decided to surrender, and — fearing Harthama less than Tahir — set out secretly at night to be rowed across to the eastern camp. By accident or by treachery the boat went over. He swam back to the western bank, where Tahir's troops took him, and Tahir had him killed at once in a garden near the Anbar Gate. The head went to Ma'mun in Khurasan as proof that the war was over.

What it cost

Ma'mun arrived some months later and rebuilt much of the capital. The Round City never recovered.

He did not live in it. When he was in Baghdad at all he stayed on the east bank, in the old Barmakid palace below the Mukharrim quarter. His brother al-Mu'tasim would leave the city altogether for Samarra within twenty years.

So the purpose-built capital lasted fifty years as the thing it was designed to be. It was not destroyed by an enemy of the dynasty or by a frontier the caliphate could not hold. It was destroyed by the Abbasid army, on the orders of an Abbasid caliph, to remove an Abbasid caliph, while the man who ordered it stayed two thousand miles away and waited for the head to be sent to him.

Two centuries of translation, and who paid for it

Drafted from Gutas, Greek Thought, Arabic Culture: The Graeco-Arabic Translation Movement in Baghdad and Early 'Abbasid Society (1998)peak

Between the middle of the eighth century and the end of the tenth, very nearly the whole surviving corpus of Greek secular science was translated into Arabic: medicine, astronomy, mathematics, physics, botany, alchemy, philosophy, most of Aristotle, most of Galen, Euclid, Ptolemy. Not the poetry, not the drama, not the history — the useful books. It is the largest transfer of knowledge between two languages before the printing press.

The version usually told makes it a story about individuals: a curious caliph, a famous translator, a house of wisdom. Gutas's argument is that the individuals are real and the explanation is not. A two-century programme, employing hundreds of people, funded well enough that Hunayn ibn Ishaq could be paid by weight for what he produced, is a social phenomenon with a demand behind it, and the demand was the caliphate's.

Several demands, in fact, and they are worth separating. The court needed astrology, which needs astronomy, which needs Ptolemy and the mathematics under him. The administration inherited a Sasanian bureaucratic culture in which scientific patronage was part of what a king did. Physicians were wanted and Galen was where medicine was. And there was an ideological argument, which is the part that makes this Abbasid rather than merely eighth-century: the claim, promoted in early Abbasid circles, that the sciences had been Persian, that Alexander had scattered them to the Greeks, and that recovering them in Arabic was therefore a restoration rather than a borrowing. A dynasty that had come to power on Khurasani arms and ruled through a Persianate administration had good reason to find that argument congenial.

Two things follow. The first is that the movement ended, around the close of the tenth century, and not because anybody suppressed it: the useful books had been done and the Arabic scientific tradition had outgrown them. The second is that almost everything later courts in the east had to think with came out of this. When the Ilkhanate's vizier writes a world history, when the Nizamiyya madrasas endow a scholarly class, when al-Biruni measures India for a Ghaznavid patron, the intellectual equipment was assembled in Baghdad on Abbasid money in these two centuries.

The last caliph who tried to say what Islam was

Drafted from Hinds, University of Cambridge Oriental Publications 37 (1986)contraction

In 833, a few months before he died, al-Ma'mun sent instructions to his governor in Baghdad. Judges, jurists and traditionists were to be examined on a point of doctrine — whether the Qur'an was created or eternal — and those who would not affirm that it was created were to be removed from office, refused as witnesses, and if necessary imprisoned. The examination is called the mihna, and it ran under three caliphs for about fifteen years.

The doctrine is not the point, and the caliphs' opponents understood that better than most later accounts do. What the mihna asserted was jurisdiction: that the caliph could determine a question of belief and require the religious professionals to accept his determination. The Umayyads had made the earlier version of the same claim, styling themselves deputy of God; the mihna is that claim made operational, by an Abbasid, against named people, with penalties.

It failed, and it failed in the most damaging possible way. Ahmad ibn Hanbal refused, was flogged and imprisoned, and did not yield, and his refusal made him the most respected religious figure of the age. Around 848 al-Mutawakkil abandoned the examination and reversed the policy. No caliph tried again.

What was settled in those fifteen years was not the createdness of the Qur'an — that question simply stopped being asked at swordpoint. What was settled was who had the right to answer it. After the mihna the definition of Islam belonged to the scholars, as a class the caliph paid, appointed and could not overrule, and the caliphate held political authority and a religious title with the content taken out of it.

So the caliphate began contracting before its territory did. The office lost something in 848 that it never recovered, and that mattered more in the long run than the provinces it was about to lose to its own guard. Every later ruler who sent to Baghdad for a caliphal investiture was buying the political endorsement of an institution that had already conceded the religious argument.

The soldiers at Samarra

Drafted from El-Hibri, The Abbasid Caliphate: A History (2021)contraction

Al-Mu'tasim's solution to an army with local loyalties was an army with none: Turkish slave soldiers, bought young from the Central Asian steppe, converted, trained, and attached to the caliph personally rather than to a tribe or a province. It is one of the most consequential administrative decisions in Islamic history, and it is the origin of the institution that produces the Ghaznavids, the Mamluks and eventually the janissaries.

In 836 he moved the whole apparatus sixty miles up the Tigris to a new capital at Samarra, because Baghdad's population and his new troops could not be kept in the same city without fighting. Samarra was a caliphal city in a way Baghdad had stopped being: enormous, planned, and containing almost nobody the caliph had not brought there.

Within thirty years the guard was choosing the caliph. Between 861 and 870 four caliphs were raised and destroyed by their own Turkish commanders — a period the sources call the anarchy at Samarra — and the office never recovered its independence from the men who held its weapons. The army the caliphs bought to free themselves from provincial politics became the only politics there was.

The provinces drew the obvious conclusion. The Tahirids were already governing Khurasan hereditarily; the Saffarids took Sistan without asking; the Samanids consolidated Transoxiana; the Buyids took Baghdad itself in 945 and kept the caliph. What was left in each case was the grant and the khutba — the caliph's name on the coinage and in the Friday prayer, and nothing behind it.

Freedom, wealth, and slaves

Drafted from Noldeke, Sketches from Eastern History (1892)contraction

East of Basra the ground was flats and ditches, and on them rich Basran contractors worked gangs of black slaves, most of them brought from the East African coast — the land of the Zanj. The job was to strip off the nitrous surface soil, both to reach the fertile ground beneath and to recover the saltpetre in the crust.

Nöldeke's observation about the arrangement is the beginning of everything that followed. Industrial labour on that scale, in the open country, was rare in the Islamic world; the work was very hard and the supervision had to be strict; and the attachment that elsewhere bound a household slave to the family he had grown up in was entirely absent. What forms instead, among masses of people working together, is a common sense of grievance and eventually a sense of their own strength. He compares it to the servile wars of the late Roman republic, and the comparison is exact.

What he actually promised

In September 869 a man calling himself Ali ibn Muhammad came into the saltpetre district under cover of acting as a business agent, and began working on the gangs. He claimed descent from the Prophet's son-in-law; whether the claim was true was disputed at the time and cannot be settled now.

What he offered them is the part that gets left out.

He told the slaves how badly they were being treated, and promised that if they joined him they would have freedom, wealth — and slaves. Nöldeke states the consequence plainly: he did not preach universal equality, but reserved supremacy for the particular class he was addressing.

The theology was chosen to fit. Being an Alid, he might have been expected to argue a hereditary right and found a Shi'i movement. Instead he adopted the doctrine of the Kharijites, the Alids' bitterest opponents, who held that the office should go to the best man though he were an Abyssinian slave. To men in a ditch, a claim about somebody's grandfather was worth nothing and that sentence was worth everything. He put a Kharijite verse on his banner and signed himself on it simply Ali, son of Muhammad, with no allusion to the descent he claimed elsewhere.

Basra

The rising took southern Iraq for fifteen years. The rebels built a capital below Basra on a salt flat where nothing grew, called al-Mukhtara, the elect city — palm-branch and mud huts, sun-dried brick palaces, prisons for the captives, mosques — on ground so changed by the shifting waters that its site will probably never be found.

In 871 they took Basra itself, the second commercial city of the empire. The chief had vowed vengeance on it. His general Ali ibn Aban let a delegation through to plead for quarter and was replaced for it by a less soft-hearted man. Inhabitants were reportedly drawn together at appointed places by offers of safety so that they could be killed more conveniently. The lowest figure the chroniclers give for the dead is three hundred thousand — a number of the kind that period produces and not one anybody counted, but nothing in the tradition suggests a small massacre.

The women and children were taken as slaves. Women of the houses of Ali and of the reigning Abbasid family were sold to the highest bidder. And, Nöldeke reports, many of the Zanj received ten slaves or more as their share.

The promise had been freedom, wealth and slaves, and at Basra it was kept in those three terms exactly.

Fifteen years

The caliphate could not finish it quickly. Army after army was beaten in night attacks in country nobody from Baghdad could read: tidal channels, broken embankments, marsh that had been arable within living memory, and canals around Basra which one source claims numbered more than a hundred and twenty thousand.

What ended it was al-Muwaffaq, the caliph's brother, who ran the war as a siege: a new city built opposite the rebel capital, the waterways closed, and the supplies that traders and bedouin had been carrying in within sight of the army finally cut. Al-Mukhtara was stormed in 883. How the chief died is unknown — Nöldeke inclines to suicide, partly because no soldier ever claimed the reward for killing him. The head reached Baghdad on a pole in November 883.

Thousands of the Zanj surrendered and were pardoned and taken into the caliph's service, because it would have been senseless to drive them to desperation when they could be used as soldiers. Others died of thirst in the desert or were enslaved again by bedouin. A contemporary source describes the ones who entered imperial service as barbarians who spoke no Arabic and ate carrion, which is what the winners thought of the men they had just recruited.

What it did to the caliphate

Fifteen years, the second city of Iraq sacked, the Gulf trade routes closed, the sugar and saltpetre economy of the south wrecked, and the campaign in the end conducted and won not by the caliph but by his brother.

The pattern is the one that runs through the whole of this stretch. Authority worked when a capable man in the family took it and spent everything he had on it; the office itself supplied legitimacy and very little else. Al-Muwaffaq saved Iraq for his brother and never became caliph. Within a generation the men running the empire on the caliph's behalf were doing it without the family connection, and within two the arrangement had a name.

The vizier and his property

Drafted from al-Rudhrawari, The Eclipse of the 'Abbasid Caliphate: Original Chronicles of the Fourth Islamic Century (1921)contraction

By the reign of al-Muqtadir, which began in 908, the caliphate had a problem that no amount of authority could solve: it could not pay the troops, and the only reliable money left in the empire was money already in the hands of its own officials.

What it did about that is the clearest picture we have of a state consuming itself, and Miskawayh, working from the records of the men who ran it, sets it out in the language of accountancy.

The riot

Abu Ali al-Khaqani, made vizier without experience and bewildered by the job, could not pay the stipendiaries and the chief officers in full. So they rioted. They occupied the Oratory, took most of the officers with them, and abused him to his face.

Al-Muqtadir ordered him to pay. He excused himself on the grounds of deficient supplies and reduced revenue — and then said the interesting thing: the money extracted from Ibn al-Furat, the previous vizier, and from his dependents had gone into the caliph's private treasury, and the treasurer would not release any of it.

The caliph then ordered half a million dinars paid out of that private treasury to satisfy the rioters. And appointed a postmaster of Baghdad who was also to be overseer of the vizier, the army, the officials, the judges and the police — which is what a state does when it has stopped trusting its own executive.

Office as a source of extraction

The mechanism this produced is the thing to understand.

A vizier's tenure ended in his arrest and the confiscation of his estate. That confiscated estate was then a significant part of the next year's revenue. So the office was worth holding for what could be accumulated in it, and worth ending for what could be taken back — and everyone involved knew both halves.

Ibn Thawaba, seeing how weak Khaqani was, wrote to the caliph through Umm Musa, one of the palace women, offering to extract vast sums the vizier had overlooked: seven hundred thousand dinars from Muhammad ibn Ali al-Madharai and his brother alone. He was given full powers, and used them to extort with violence, dismiss the vizier's appointees and install his own. After which, Miskawayh says, various knaves wrote in through the same channel asking for posts and promising money, and the vizier was ordered to appoint them, so that the administration was weakened and the knaves extorted money by every form of violence.

The bureaux were also cooking the books downward. Hamid ibn Abbas had farmed the revenue of Wasit for four years; the clerks drew up his account at 240,000 dinars a year plus 24,000 kurr of barley for the cavalry, when what he had actually contracted for was the previous year's register with a slight addition.

What Mu'nis understood

The sharpest moment is a piece of advice.

Seeing the disorder, al-Muqtadir proposed bringing Ibn al-Furat back as vizier. Mu'nis the eunuch — the commander who was the real power for most of the reign — advised against it, and his reason was not administrative. It would cause a scandal, he said, if the provincial governors learned that the Sultan had dismissed a vizier and then been compelled to restore him after a few months; and the Sultan's action would be attributed simply to the desire to seize the vizier's goods.

Not that it would be unjust. That it would be understood. The centre had reached the point where its provinces read every appointment as a plan to confiscate somebody, and the man closest to the caliph knew they did.

A government financed by robbing its own servants cannot recruit good servants, and cannot be believed by the provinces that pay it. Al-Muqtadir was murdered by his own troops in 932. Within four years the caliphs had handed military and fiscal power to a commander-in-chief, and within thirteen a Daylami soldier had walked into Baghdad and taken the city from all of them.

945, and the caliph they kept

Drafted from El-Hibri, The Abbasid Caliphate: A History (2021)contraction

In December 945 Ahmad ibn Buya, a Daylami soldier from the Caspian highlands, entered Baghdad and took the city. He deposed the caliph al-Mustakfi the following year and had him blinded. He did not abolish the caliphate. He took the title amir al-umara, installed the next Abbasid in the palace on a pension, and ruled Iraq and western Iran for the rest of his life with a powerless Sunni caliph in his keeping.

The Buyids were Shi'i. They had every doctrinal reason to end the Abbasid line and no military reason not to. That they kept it for a hundred and ten years is the fact that explains what the office had by then become. The caliph's grant was the currency in which authority was quoted across the whole east — the Samanids in Bukhara, the Ghaznavids after them, and every governor who wanted his seizure to read as an appointment. A Buyid amir who destroyed the institution would have destroyed the thing that made his own rule legible to everyone outside Iraq, and put nothing in its place.

So the office survived precisely because it had been emptied. It was worth more as a stamp than as a rival, and a stamp is cheap to keep.

Whether the Buyids ever seriously weighed replacing the Abbasids with an Alid caliph of their own is contested; the sources that report the deliberation are later and hostile, and there is no settling it. What is not contested is the outcome. For over a century the most powerful men in Iraq governed in the name of a prisoner whose signature they needed and whose authority they had taken.

The forty-five years when it nearly worked

Drafted from El-Hibri, The Abbasid Caliphate: A History (2021)contraction

In 1055 Tughril Beg entered Baghdad, expelled the Buyids and restored the caliph — to exactly the position he had held before. The Great Seljuk sultans were Sunni and presented themselves as the caliphate's protectors, which changed the theology of the arrangement and nothing about its arithmetic. The sultan held the army and the provinces; the caliph held the investiture that made the sultan's holding lawful, and the arrangement ran, with interruptions, for over a century.

A second name in the pulpits

That arrangement was tested almost as soon as it was made. In 1058 a Turkish commander named al-Basasiri, acting for the rival caliphate at Cairo, took control of Baghdad while Tughril Beg was away putting down a revolt by his own brother. For the better part of a year the Friday sermon in every mosque of the Abbasid capital named not al-Qa'im but the Fatimid imam al-Mustansir as commander of the faithful, and al-Qa'im himself was held as a hostage downriver at Anah, a bargaining piece in an argument over whose name Baghdad's own pulpits belonged to. Tughril, once free to deal with it, did not negotiate with Basasiri at all — he went directly to the officer holding the caliph and had him released. The Fatimids had reached Baghdad through one disaffected general and could not keep it: holding Iraq would have needed an army they had never built that far east. But the arrangement Tughril had made only three years earlier survived its first real test by the width of one man's loyalty, not by its own strength.

What is usually missed is what happened when the protector weakened. Seljuk power fragmented after 1092 and again after 1157, and each time the caliphs took back not just prestige but ground — tax, garrisons, an actual administration in Iraq. By the time al-Nasir li-Din Allah came to the throne in 1180 there was something real to govern, and he governed it for forty-five years, the longest reign in the dynasty.

Al-Nasir is the most interesting Abbasid after al-Ma'mun and the least written about here until now. He rebuilt a territorial state in Iraq, and he tried to rebuild the caliph's authority over rulers he could not coerce by reorganising the futuwwa — the urban chivalric associations — into an order with himself at its head, and inviting princes to join it. It was an attempt to make the caliphate a network of personal allegiance rather than a fiction of delegation.

It very nearly worked, and then it met Khwarazm. Muhammad II repudiated al-Nasir in 1217 and marched on Baghdad; the campaign broke in the mountains before it arrived. Within four years the Khwarazmshah's empire had been destroyed by the Mongols, and the caliphate had thirty-seven years left.

The point for the next chapter is that what ended in 1258 was not a shadow. It was a functioning state, smaller than it had been and much larger than it had been in 945.

The grant outlived the debtors

Drafted from contributors, Encyclopaedia Iranica · sourceend

An office with no army spent its last three centuries investing the men who had one — Buyid amirs, then Seljuk sultans, each ruling Iraq and the Iranian east by force and asking Baghdad to say so in form, because conquest alone could not make a seizure read as succession. That arrangement, not a standing army or a governed territory, is what the Abbasid caliphate actually was for most of its existence, and the arithmetic underneath it is worth doing once, end to end, before the narrative of how it stopped.

Tahirid, Saffarid, Samanid, Ghaznavid, Seljuk, Khwarazmshah: every dynasty of consequence in the Iranian east for four hundred years held its authority on paper from an institution that did not fall until 1258 — and every one of them had already fallen first. The Khwarazmshahs, the last of them, were destroyed by the same Mongol advance that would take Baghdad, twenty-seven years earlier. An office that had not commanded an army of its own since the ninth century outlived every army that ever claimed to rule in its name.

The chief town of a province

Drafted from Strange, Baghdad during the Abbasid Caliphate, from contemporary Arabic and Persian sources (1900)afterlife

Baghdad had been besieged and taken once before, in 813, by its own dynasty's army against itself; the Round City fell and the capital rebuilt around it. This was the second time, and the army outside the walls was not the caliph's own.

The siege began in the middle of Muharram 656 — January 1258. Hulagu commanded the centre in person and camped east of the city; his right lay before the Gate of the Sultan's Market on the north, his left before the Kalwadha Gate at the southern end of East Baghdad, and detachments that had crossed the Tigris upstream came round by Anbar and the Isa canal to take the western side. The main assault was aimed at the Persian Bastion and the Halbah Gate.

Inside, the city helped. Karkh and the quarter around the shrine of the Imam Musa were Shi'i, and to demonstrate what they thought of a Sunni caliph they corresponded with the besiegers.

After about fifty days East Baghdad was stormed. Al-Musta'sim was brought out of the Kalwadha Gate to stand as a suppliant in front of Hulagu, and was killed shortly afterwards in the camp near that gate, with his sons; the authorities differ on the details and agree on the fact. By one report he was rolled in a carpet and trampled, so that no royal blood touched the ground. Hulagu moved into the Ma'muniyya quarter, which the Persian historian of the conquest renders, deliberately, as the Monkey-house.

The sack lasted forty days. A large part of the population was killed, and a fire took the caliph's mosque, the shrine of Musa al-Kazim, the tombs of the caliphs at Rusafa, and most of the streets and houses. Then the army moved on to the next conquest, Hulagu leaving orders that the mosque and the shrine were to be rebuilt.

What the city became

The rest is a list, and the list is the point. From being the real or nominal capital of Islam, Baghdad became the chief town of the province of Arab Iraq.

Hulagu's descendants held it for not quite a century. Then the Jalayirids, whose chief made it his residence in 1340. Then Timur, who occupied it in 1393, stayed two months, and on leaving instructed his lieutenant to rebuild a city that had mostly fallen into ruin. Then the Turkomans of the Black Sheep from 1411, and the White Sheep who dispossessed them in 1469, and Shah Isma'il's Persians in 1508, and the Ottomans in 1534, and the Persians again in 1623 through the treachery of a janissary officer, and Murad IV in 1638, after which it was the seat of a Turkish pasha.

Nine changes of master in four centuries, and not one of them made anything of it except a provincial capital.

The last man to see the mosques

Ibn Battuta was in Baghdad in 1327, sixty-nine years after the siege, and he appears to have been the last writer who saw the three great mosques of the old capital still standing: al-Mansur's on the west bank, the mosque of Rusafa on the east, and the Seljuk sultan's mosque a mile from it.

All three have since disappeared entirely, along with every trace of the great hospital, which in Ibn Battuta's day was already a ruin on the river bank where the old bridge of boats had crossed.

A city that had been laid out as a circle with the ruler at its centre, and had grown into the largest thing of its kind west of China, is now recoverable mostly as a list of names in books, matched up against ground where somebody would have to dig.