The Community at Medina

An arbitrator, invited

Drafted from Watt, Muhammad at Medina (1956)formation

In September 622 a preacher who had failed in his own city moved two hundred miles north to an oasis that had sent for him.

That is the least dramatic way to state it and it is the most useful, because what Yathrib wanted was specific. The oasis was two confederations of Arab clans, the Aws and the Khazraj, with several Jewish clans among them, and it had been feuding for years without any of its chiefs being able to stop it. What it lacked was an arbitrator nobody's cousin — someone from outside with standing enough to be obeyed and no blood in the local quarrels. Muhammad had been preaching in Mecca for a decade against increasing hostility, and had the standing.

How little he was given

The most striking thing about the arrangement is how weak the position was.

Watt's reading of the founding agreement is that Muhammad's formal powers in 622 were so slight that they cannot have been much less at the very start. The document gives him exactly one operative function: disputes are referred to him. He is named in the preamble as Muhammad the prophet. His own following, the emigrants from Mecca, appears in the list as one group alongside the eight Medinan clans, which puts him level with the clan chiefs — first among them in order of mention, and perhaps in honour, but level. He is not the ruler of Medina. He is one important man among a number of important men, and for the first year several of them were probably more influential than he was.

Nor was the right of arbitration much of a power in itself. A man to whom disputes are referred has authority only if his settlements are accepted, and they are accepted only if he finds answers people can live with. The office was an opportunity to acquire influence, not a grant of it.

What that looked like in practice

Two episodes show the constraint working.

When Abdullah ibn Ubayy, the most powerful man in Medina, spread the slander against Aisha, Muhammad could not simply act against him. He had to convene the Ansar and ask permission of the men who might feel bound to avenge any injury done to Ibn Ubayy. He got his way, but he had to ask.

Later, when the fate of the Jewish clan of Qurayza was decided after the siege, Muhammad did not pronounce the judgement himself. If he had ordered blood shed, honour might have obliged Qurayza's confederates — Muslims by then — to avenge it. The decision was handed to the chief of the clan they had been allied to.

Both are cases of a man routing around a limit rather than overriding it. The clan chiefs of Medina kept their power for years, and it bounded what he could do; his authority grew inside those limits rather than by abolishing them.

Ten years later the same man's writ ran over most of Arabia. That distance, travelled without a bureaucracy or a standing army, is the subject of everything that follows.

A writing between the emigrants and the Ansar

Drafted from Watt, Muhammad at Medina (1956)formation

The document usually called the Constitution of Medina survives because Ibn Ishaq copied it. He says nothing about where he got it or when it came into force, and its position near the start of his Medinan narrative is probably just where logic put it rather than where evidence did. Watt treats it as an ancient document nonetheless, and the internal reason is strong: it gives Muhammad so little that it is hard to see a later generation composing it.

It opens by naming what it is making:

This is a writing of Muhammad the prophet between the believers and Muslims of Quraysh and Yathrib and those who follow them and are attached to them and who crusade along with them. They are a single community distinct from other people.

The last verb is jahadu, which Watt renders as crusade and which later vocabulary would handle differently; here it simply marks out those who campaign with the community from those who merely live alongside it.

Ummah, the word translated community, is doing the founding work. The clause asserts that a body of people drawn from two different towns and a dozen unrelated clans is now one thing.

What it actually regulates

Most of the clauses are about killing, and that is not a curiosity — it is the whole design.

Nine near-identical articles set out that each clan continues to pay blood-money and ransom its captives according to their former condition, meaning by the arrangements already in force. Nothing is centralised. The emigrants from Mecca are simply added to the list as a tenth unit with the same obligations, which is how a body of refugees with no land became a clan for legal purposes.

What is new sits on top of that. Believers do not abandon a debtor among them. No believer kills a believer over an unbeliever. God's protection is one, and protection granted by the least of them binds all of them. Where a man is killed wrongfully the whole community is against the killer, nothing is permissible to them except to oppose him — which takes the most dangerous decision in a feuding society, whether to avenge or accept payment, and puts it beyond a single clan's discretion.

The old order is left standing and a second order is built across it. That is why the document works: nobody had to give anything up.

The Jews are inside it

The clauses that surprise modern readers are the ones about the Jewish clans, and they are unambiguous:

The Jews of Banu Awf are a community along with the believers. To the Jews their religion and to the Muslims their religion.

The formula is then repeated for Najjar, Harith, Saida, Jusham, Aws, Thalaba and their clients. They bear their own expenses and the Muslims bear theirs; between them there is help against anyone who attacks the people of the document, and sincere friendship, and honourable dealing, not treachery. The valley of Yathrib is declared sacred to all of them. Whoever attacks the oasis is to be resisted by all of them.

This is a defensive league of religiously distinct parties, not a confessional state. What happened to that arrangement over the following five years is a separate and much darker story, and it is told later.

The two clauses that mattered

Buried in the list are the provisions that turned an arbitrator into a government.

Whenever among the people of this document there occurs any incident or quarrel from which disaster is to be feared, it is to be referred to God and to Muhammad, the Messenger of God.

And: no member of the community may go out to war without Muhammad's permission.

One clause makes him the final court. The other gives him a monopoly on the decision to fight, which in a society where any clan could start a war by raiding was the single most consequential thing to control. Neither reads like much. Between them they are the whole apparatus of the state he had, and for several years they were enough.

The quarrel over the spoils

Drafted from Watt, Muhammad at Medina (1956)expansion

The emigrants had a problem nobody records anyone solving: they had to eat.

Seventy or so Meccan Muslims had moved to an oasis whose land was already owned and already worked. They did not take up farming. They could not live indefinitely on Medinan hospitality, though for the first two years they largely did — one of the complaints levelled at Muhammad by men trying to prove they were not Muslims was that he and his followers were a burden on the town. Some of them traded, and a few were very good at it: Abd al-Rahman ibn Awf turned down the offer of half a Medinan's property, asked to be shown the market instead, and came back the same day with butter and cheese he had made a profit on. Umar missed an instruction from Muhammad because he was in the middle of a transaction. Ali gathered rushes to sell.

But the Medinan market already had a trading clan working it, and the obvious alternative was the one that led straight to war: the Meccan caravans passing within reach on the Syrian road.

Badr

In 624 a raiding party went out after a caravan and met the force Mecca had sent to protect it. The Muslims won.

What matters here is the argument afterwards. There was a dispute over dividing the spoil, and it was settled by a ruling that Muhammad could dispose of the whole as he chose. Out of that came the arrangement that held for the rest of his life and long after: a fifth — the khums — went to Muhammad for public purposes, and the rest was divided equally among the men who had fought, a horseman drawing two extra shares for his horse.

That fifth is the beginning of the state's finances. Until then the community had money in the sense that individuals in it had money. Now there was a fund held by the head of the community, spent on its behalf, with a collecting agent appointed to handle it — a man named Mahmiya ibn Jaz, of the Zubayd. A treasury with a treasurer is a different kind of institution from a preaching.

An old shape, filled differently

The proportion was not invented. An Arabian tribal chief customarily took a quarter of what his tribe seized, and Muhammad took a fifth; he also took the safi, the chief's traditional right to pick one item out before the division began, and at Badr he picked a sword.

So the office he was building was recognisable. He drew what a chief drew, a little less of it, in the manner chiefs drew it. What was different was where it went. A chief's quarter was his; the fifth was for public purposes. The same custom, pointed at a different object, is most of how the institutions of this community were made — and when he later dealt with nomad tribes far from Medina, what he required of them was this same fifth, his share and his pick, in terms every one of them already understood.

Have you forgotten what he told you

Drafted from Watt, Muhammad at Medina (1956)expansion

Badr had been won by a raiding party that stumbled into a fight it did not plan for. Uhud, the following year, was a battle Medina chose, and the choosing is where the authority the Constitution had built first came under real pressure — before a single blow was struck.

The argument over how to fight

Muhammad's own preference, when word came that Mecca was marching on Medina with three thousand men, was to stay inside the town and let its houses and lanes do the fighting. Men who had missed Badr and wanted their own chance at the enemy argued him out of it, and he gave way, put on his armour, and led about a thousand men out to open ground below Mount Uhud.

He had not finished losing men before the battle even began. Abdullah ibn Ubayy, the most powerful chief in Medina and no friend of the arrangement he had been outmanoeuvred into accepting, pulled roughly a third of the force back to the town, unconvinced by the plan he had been overruled on. A community that had agreed nobody could go to war without Muhammad's permission had just demonstrated that permission and participation were not the same thing: a chief could still take his own men home.

The order, and what happened to it

Muhammad posted fifty archers on a rise covering the army's exposed flank and gave a direct order, plainly enough that no one afterwards claimed to have misunderstood it: hold the position, whatever happened in front of you, and do not come down.

The battle opened well. The Meccan line broke and its camp lay open, and the archers watched their own side start taking the plunder below. Most of them left the hill for it. The booty! O people, the booty! Your companions have won — what are you waiting for now? Their own commander, Abdullah ibn Jubayr, tried to hold the rest to the order he had been given and was asked, in effect, why it still mattered. Have you forgotten what the Messenger of God told you? He had not forgotten. It did not matter. Enough of them went down the hill that the flank Muhammad had told them to cover was open, and Khalid ibn al-Walid, commanding the Meccan cavalry, rode round it.

The reversal

What had been a Muslim victory became a rout inside the same afternoon. Khalid's horsemen came in from the rear the archers had been posted to deny them, the Meccan line that had broken reformed, and the Muslim force that had been advancing found itself fighting in a circle with no clear front.

Muhammad went down in the middle of it. A blow drove two rings from his helmet into his cheek; a thrown stone split his lip; he lost a tooth and fell into one of the pits an apostate named Abu Amir had dug against exactly this kind of scramble. Ali pulled him out and Talha took his weight while Abu Ubaydah worked the iron rings loose from his face. And into that confusion went the report that mattered more than any wound: that he was dead.

What the rumour actually threatened

It is worth being exact about why that report was dangerous, because it was not simply bad news.

The whole apparatus the Constitution had built — the referral of disputes, the monopoly on the decision to fight, the ummah itself — ran through one man. There was no office to inherit it, no procedure for a successor, nothing but the arrangement he personally held together. A false report of his death was not a rumour about a battle. For as long as it was believed, it was a test of whether the thing built in the Constitution could survive for even an afternoon without the one person the whole document was written around. Umar ibn al-Khattab, hearing it and refusing to accept it, is recorded rallying men on exactly that logic before anyone had confirmed anything either way.

Muhammad was alive, badly hurt, and got back up the mountain with the survivors before the Meccan cavalry — unable or unwilling to press the advantage into the high ground — turned for home. Around seventy Muslims were dead, among them his uncle Hamza, killed by a thrown javelin and mutilated afterward. Meccan losses were a third of that.

What survived it

Nothing about the authority collapsed, which is the fact worth sitting with longer than the battle itself.

The disobedience on the hill was not punished by exile or execution; it was named, in the revelation that followed, as exactly what it was — men who wanted the plunder more than the order — and left there. Abdullah ibn Jubayr kept his place among the community afterward. The archers who ran for the camp were not expelled the way Qaynuqa already had been, the year before, for a market brawl. What Uhud tested was not obedience in the abstract but whether an authority built on consent rather than force could take an open, undisguised act of disobedience, in the middle of a battlefield reversal that nearly cost the arbitrator his life, and still be standing the following week to arbitrate the next dispute.

It was. The clause about permission to fight had never been tested by victory. It was tested by a rout, by a false report of the founder's death, and by fifty men choosing loot over an order given to their faces — and the community that had agreed to obey him went on agreeing to obey him. That is a stronger proof of what the Constitution had actually built than Badr ever supplied.

The opportunity that was lost

Drafted from Watt, Muhammad at Medina (1956)aside?What aside means

The agreement that founded the community named the Jewish clans of the oasis as members of it, on their own religion and their own property. Within five years two of those clans had been expelled and the third destroyed. What happened in between is the hardest part of the Medinan decade to write about, and it cannot be left out, because the founding document does not make sense without it.

What happened

The first was Qaynuqa, in 624, shortly after Badr. The occasion was squalid: in the Qaynuqa market someone pinned an Arab woman's skirt so that she was exposed when she stood, a Muslim who saw the joke killed the man responsible, and the man's people killed him back. The clan withdrew into its strongholds, Muhammad treated the killing as grounds for war, and after a fifteen-day siege they surrendered and were expelled. They were allowed to take their families and three days to collect debts owed them, and had to leave their arms behind.

The politics beneath it are visible in one detail. Qaynuqa were confederates of Abdullah ibn Ubayy, the most powerful man in Medina, and expected him to protect them. He spoke for them, and it achieved nothing; when he tried to force his way in to Muhammad, a guard shoved him into a wall hard enough to bloody his face, and he could neither avenge it nor claim compensation. Watt's point is that Qaynuqa accepted exile only when they understood what that told them. The expulsion measured how much the balance in the oasis had shifted since the document was written.

Nadir followed, and their date-lands went not to the Muslims at large but to the emigrants, along with two of the poorest of the Ansar — the first landed property the community disposed of. Then in 627, after the siege of Medina, came Qurayza, and it did not end in exile.

Watt's reading

Watt is explicit that the fundamental cause was not the money.

He does not deny the money existed. The wealth of the Jewish clans, he writes, was of great benefit to Muhammad and considerably eased his financial position, and the prospect of it may well have influenced the timing of the attacks. But making it the whole explanation he calls unduly materialistic. The quarrel, on his reading, was theological on both sides and inescapable once joined: the Jews of Medina held that God had chosen them and revealed Himself through them, and Muhammad's claim to receive messages could not be accommodated to that; while Muhammad had come to see his prophethood as the only possible basis for uniting the Arabs, which made public criticism of it a threat to the entire political experiment rather than a difference of opinion.

He is unusually blunt about the rejection being reasonable. It was altogether in keeping with the traditional outlook of Jewry that the Jews of Medina should reject Muhammad, he writes, and even leaders more far-sighted than the ones they actually had would have done the same.

And then he writes the sentence the chapter exists for:

In the early months at Medina the seeds were sown of a great tragedy; a great opportunity was lost.

His counterfactual is that at several points terms were available that would have given the Jewish clans religious autonomy inside the community, and that had they been taken, the Jews might have become partners in the Arab empire and Islam a sect of Jewry.

That is one scholar writing in 1956, and it is speculation, which he says himself. It is included here because it is the judgement of the standard study of this polity, and because the alternative — telling the story of the founding agreement without saying what became of many of the parties to it — would be worse than partisan. It would be incomplete.

Muhammad son of Abdullah

Drafted from al-Waqidi, The Life of Muhammad: al-Waqidi's Kitab al-Maghazi (2011)expansion

In 628 Muhammad went towards Mecca with his followers in pilgrim dress and stopped at a place called al-Hudaybiyya, outside the sacred area. Quraysh sent Suhayl ibn Amr to negotiate. What they produced is the clearest evidence available that the thing at Medina had become a state, and it survives with the argument still in it.

The wording

Ali was called to write, and Muhammad dictated the opening: In the name of God, most gracious most merciful.

Suhayl stopped him. I do not know the Merciful, he said. Write as we write: in the name of God. The Muslims present objected loudly. Suhayl said that in that case there would be no agreement about anything. Muhammad told Ali to write it Suhayl's way.

Then he dictated this is what the Messenger of God has agreed to, and Suhayl stopped him again — and his objection was courteous and completely uncompromising:

If I knew that you were the Messenger of God, I would not differ from you and I would follow you. I prefer your name. Do you then detest your name and the name of your father, Muhammad b. Abdullah?

This time the protest was worse. Usayd ibn Hudayr and Sad ibn Ubada took hold of the scribe's hand and told him not to write anything but Muhammad the Messenger of God, or the sword would settle it — why would we permit this insult to our religion? Muhammad signalled them to be quiet. A Meccan envoy watching this turned to his colleague and said he had never seen a people so concerned about their religion.

And then Muhammad gave way a second time. I am Muhammad son of Abdullah, he said. Write: in the name of God, this is what was agreed between Muhammad son of Abdullah and Suhayl.

He signed as a man with a father, not as a prophet.

The terms

Ten years of no fighting. Both sides secure, with no bribery and no treachery. Any Arab tribe that wished could enter Muhammad's agreement or Quraysh's, and Khuzaa immediately joined his while Bakr joined theirs.

Then the clause that nearly broke it: anyone who came to Muhammad from Mecca without their guardian's permission was to be sent back, while anyone who went from Muhammad to Quraysh was not. The asymmetry is total and Muhammad accepted it. He also accepted that he would turn round and go home that year without performing the pilgrimage, and could return the following year for three days only, carrying no weapons but a sheathed sword.

Nine men witnessed it and their names were written on the face of the document. Suhayl said he would keep it. Muhammad said he would. They argued, a second copy was made, and each side went away holding one.

Why it mattered

Read as religion the agreement is a humiliation, and some of his own men read it that way at the time.

Read as politics it is the moment Quraysh conceded the argument. For six years they had treated Muhammad as a renegade to be crushed. At Hudaybiyya they sent a plenipotentiary, negotiated clause by clause, took witnesses, executed the document in duplicate and bound themselves for ten years. States do that; states do it with other states. Quraysh had recognised a rival power in Arabia even while refusing to write down what it called itself.

Within two years Mecca was his.

The peace, and who was required to believe

Drafted from Watt, Muhammad at Medina (1956)peak

Mecca submitted in 630 without a battle, and Muhammad required nothing of its people. No profession of faith was demanded of the city that had driven him out. Many Meccans fought alongside him at Hunayn weeks later while still not Muslims.

That is worth sitting with, because it cuts against the shape the story is usually given. What was being extended over Arabia in these years was not, in the first instance, a religion. It was a peace.

The Pax Islamica

Watt's name for it is the Pax Islamica, and his argument is that it grew out of the problem the oasis had hired Muhammad to solve. Medina's clans had been feuding; the founding agreement stopped them. Feuding was not a Medinan peculiarity but the ordinary condition of Arabia, and once it became visible that somebody had a working method for suspending it, other tribes wanted in.

So the system spread partly by being useful. In the early Medinan years Muhammad made alliances with neighbouring tribes on what Watt calls a purely secular basis — no religious condition at all. Only gradually, as the zone of peace widened and his position strengthened, did he begin requiring belief in God and recognition of himself as prophet as a condition of alliance. The tightening came after Hudaybiyya, when nomads he had counted on failed to turn out; from then, accepting Islam also meant accepting that the prophet could give orders.

And even then it was not uniform. He went on making purely secular alliances with groups from whom no religious demand was made at all, and this was the normal arrangement with distant and powerful tribes, right up to his death. There were two kinds of membership, and the sources describe both.

The Arabs, as an idea

The unification is easier to see now than it was then.

The word Arab is barely present in pre-Islamic poetry. The adjective Arabic is first attested in the Qur'an, where it is used to contrast clear speech with the chewed or indistinct speech of the foreigner — the ajami — and otherwise as a description of the revelation itself. There is no noun for the Arabs. What the word al-arab means in the later passages is the nomadic tribes as opposed to townsmen.

So the category was still forming. An Arabic Qur'an addressed to the clear-speakers implies a body of people set against Abyssinians, Byzantines and Persians, and Watt thinks a good deal of the appeal of the new religion over Christianity or Judaism was exactly that: it was not foreign.

Muhammad had begun by understanding himself as sent to his own tribe. The widening happened in steps that the sources do not mark clearly, and by the end he was the head of a system covering most of the peninsula. Whether he set out to unify the Arabs is doubtful. He set out to stop people killing each other, and unification is what that turned out to require.

The problem peace created

Drafted from Watt, Muhammad at Medina (1956)peak

By 631 Muhammad had a difficulty that success had manufactured, and Watt argues he could see it coming for some time before it arrived.

Arabia's tribes had raided each other. The razzia was not only violence: it was a redistribution mechanism and a demographic one. A tribe that fell short in a bad season made the difference up from its neighbour. Raiding also killed people, which in a country of thin margins was part of how the population stayed within what the land could carry. And it absorbed the energies of a great many young men who had nothing else to do.

The Pax Islamica switched that off. Tribes inside the system stopped raiding one another. The consequences followed immediately: more people, because fewer were dying, and no way for a tribe in temporary need to make good its deficit. There was also, as Watt puts it, the psychological problem — an outlet was required for energies that had previously gone into the razzia.

If the peace was going to last, the standard of living had to be maintained, and that meant finding a new source of wealth.

The options

Trade was the obvious candidate and Watt thinks it was considered and rejected. The Medina–Syria traffic was small; the sources say so little about it that it cannot have been important. Mecca's commerce, once the city came under his rule, was not remotely large enough to support the numbers now looking to him. And Watt adds a reason that is not economic at all: trade carried the risk of reviving exactly the religious attitude that had been the fault of pagan Mecca.

Booty was the other candidate, and it had been doing this work already. In the early Medinan years plunder mattered enormously, particularly to the emigrants, and it is likely that a good many men attached themselves to Muhammad because of it.

But booty had a structural flaw that got worse as he succeeded. Raiding requires prosperous non-Muslims within reach, and every year there were fewer. As the peninsula came inside the peace, the supply of legitimate targets contracted toward zero. Watt states the conclusion plainly: if the whole of Arabia became Muslim, raiding would remain possible only on the northern frontier.

What that means

This is the least mystical explanation available for what happened next, and it comes from inside Muhammad's own lifetime rather than being read back from the conquests.

A state had been built that abolished the internal economy of raiding and put nothing in its place except an army with no one left to fight. The expedition to Tabuk in 631, the northernmost thing he ever attempted, points where the pressure pointed. He died the following year with the problem unresolved and the force intact.

The men who inherited it solved it in the direction the geography left open.

What existed in 632

Drafted from Watt, Muhammad at Medina (1956)end

Muhammad died at Medina in June 632, having named nobody to follow him and having built nothing that would produce a successor on its own.

It is worth asking precisely what it was that he left, because the answer is stranger than either "a religion" or "a state" suggests.

A tribe that was not one

Arabian political thinking ran on the tribe. A tribe was a blood group; it was the whole of a man's social security, and in exchange it took his highest loyalty. Tribes could confederate, usually for a limited purpose such as fighting another confederation, and outsiders could be attached as confederates, protected neighbours or clients — but none of that displaced kinship as the organising idea. The main tribes were sovereign. Arabs knew the Byzantine, Abyssinian and Persian empires existed, had a notion of kingship and disliked it, and understood foreign rulers by analogy to tribal chiefs anyway: in the story of Muhammad's letter to Heraclius, the emperor behaves like the chief of an Arab tribe.

The ummah took that shape and changed one thing. It was based on religion rather than on kinship. Watt notes this is nowhere stated as a theory and everywhere assumed in practice.

Everything else follows from it. If Muhammad is a messenger there is a message, and a message means God is issuing directions about practical affairs; so on matters of principle Muhammad does not decide, he announces. God is the head of the community. The founding document says the security members enjoy is the dhimmah of God — His guarantee — and in the treaties Muhammad's name is coupled with God's: groups are given the protection of God and of Muhammad together.

Not a collection of individuals

The other structural fact is that nobody belonged to this community on their own.

You were a member because your group was. A man was in because he belonged to a clan named in the founding document, or to a tribe that had since made a treaty, or because he had left his tribe and come to Medina and been classed among the emigrants. Watt draws the consequence: this is why later generations were so eager to establish that their tribe had sent a delegation to Muhammad, and why non-Arabs who became Muslims had to be enrolled as clients of an Arab tribe. There was no procedure for simply joining.

By the end there was not even a settled name for the thing. Ummah drops out of use once Mecca is in and the tribes are confederated; the treaties occasionally say jamaa or hizb Allah. Watt's explanation is that no official term was needed, because diplomacy was conducted in the name of God and Muhammad and that was sufficient designation.

The gap

Put those together and the problem of June 632 is visible.

The community's head was God; its intermediary was a man who announced what God commanded; and its membership was a bundle of treaties made with that man personally. He then died, and the one function nobody could inherit was the one the whole arrangement rested on.

The treaties were the immediate crisis, because a great many tribes took the reasonable view that an agreement made with Muhammad had ended with Muhammad. Whether they were right was decided by force over the following two years.

Khaybar, and the shape of an empire

Drafted from Watt, Muhammad at Medina (1956)afterlife

Within twenty years of Muhammad's death Arab armies held Syria, Egypt and the Iranian plateau. The institutions they governed those places with were not designed for them. They were worked out in an oasis for a population of a few thousand, and then applied to two former empires without much alteration, because they were what existed.

The Khaybar arrangement

The clearest case is a settlement made in 628 at an oasis north of Medina.

When Khaybar fell, its Jewish cultivators were not expelled and the land was not distributed to the conquerors. The previous owners stayed on it and worked it, handing over a share of what it produced. Watt notes that this was a new arrangement, and that it became the system followed in the conquests after Muhammad's death.

It is difficult to overstate what that decision did. The alternative — settling the victorious army on the land it had taken — was the normal outcome of conquest, and it would have turned the conquerors into a scattered landholding class inside a far larger population, to be absorbed by it in a few generations. What happened instead was that the conquerors stayed together in garrisons and drew the surplus of somebody else's agriculture. The Sawad of Iraq and the Nile valley were run on the principle first applied to a date plantation in the Hijaz.

The fifth, and the stipend

The other inheritance is fiscal.

The khums had been instituted after Badr, when a quarrel over dividing spoils was settled by giving Muhammad a fifth for public purposes. He required the same fifth of every tribe that submitted, along with his share and his pick, and it was collected by an appointed agent. When the conquests began, the fifth simply scaled: the same fraction of vastly larger sums, taken by the same right.

And on the other side of the ledger, membership was already tied to payment. Watt observes that once Umar fixed stipends in 636, every full member of the community drew one — and that some such arrangement was presumably in force before that, though the details are lost. Combined with the rule that nobody belonged as an individual, this produced the register that governed the caliphate for a century: to be in the community was to be in a group that had treated with Muhammad, and to be in that group was to be on the roll.

What it means for the conquests

The caliphate did not improvise its way through the seventh century. It applied a small number of arrangements, all of them worked out under Muhammad and most of them worked out for reasons that had nothing to do with governing conquered provinces, to a situation none of them had anticipated.

Some of them fitted remarkably well. The garrison-and-stipend model held an enormous conquest together with very few men. Others did not survive contact. The assumption that plunder would keep arriving was written into the stipend roll and became untenable the moment the frontiers stopped moving, which is the fiscal crisis the caliphate's first civil war was conducted over.

The oasis outlived itself by a very long way.