Han Empire

The inheritance

Drafted from (eds), The Cambridge History of China, Volume 1: The Ch'in and Han Empires, 221 BC-AD 220 (1986)formation

The Han did not build the Chinese empire. The Qin did, in about fifteen years, by force, and then fell apart within four years of its founder's death. What the Han did was work out how to keep one — which turned out to be the harder problem and took them a century.

What was on the table in 206 BC

When the Qin collapsed, the obvious thing to do was go back. The pre-Qin order had been a set of hereditary kingdoms under a nominal Zhou king, and most of the men fighting over the wreckage wanted that restored with themselves in it.

Liu Bang — a village headman of no family, which is worth stating plainly because nothing like it happens again in Chinese history for fifteen hundred years — won the war and then did neither thing. He kept the Qin apparatus of commanderies and counties under centrally appointed officials for about half the empire, and handed the other half out as kingdoms to his relatives and his generals.

That was a compromise made under duress and it contained its own crisis. The kingdoms grew, the generals were purged one by one, the kingdoms passed to Liu relatives instead, and in 154 BC seven of them revolted at once. The revolt was crushed inside three months, and after it the kings kept their titles and their revenues and lost their governments. The commandery system quietly ate the rest of the country.

The argument about the Qin

The other thing the early Han had to do was explain why it was not the Qin, and the explanation became the political theory of Chinese empire.

The verdict the Han settled on was that the Qin had failed through harshness: too much law, too much labour service, too much punishment, no benevolence, and therefore the loss of Heaven's mandate. Jia Yi's essay on the faults of Qin, in the 170s BC, is the canonical statement, and its argument is that an empire taken by force cannot be held by the same means.

That verdict is a piece of dynastic self-justification and it is also the foundation of a doctrine with teeth. If Heaven withdraws its mandate from a ruler who governs badly, then rebellion against a bad ruler is not treason but evidence, and the dynasty that wins was owed the throne. Every Chinese dynasty afterwards ruled on that proposition and every one of them was eventually overthrown by it.

The apparatus

Underneath the theory, what the early Han kept was Qin machinery: the commandery and county administration, the standardised script, the registers, a uniform law, weights and measures, and the roads.

What it added was people to run them who were not soldiers. From about 130 BC the state established an Imperial Academy, began recruiting officials partly on recommendation and examination in a canon of texts, and made a body of learning the qualification for office.

That is the invention. The Qin had an administrative machine and staffed it with clerks and officers. The Han built a class whose claim to govern was that they had read the right books, and who therefore had views about how the empire ought to be run and a vocabulary for arguing about it with the throne.

The argument started almost immediately, and eighty years later it produced the most fully documented policy debate in the ancient world.

The singing arrow

Drafted from Parker, A Thousand Years of the Tartars (1895)expansion

The Han spent its first seventy years paying a neighbour it could not beat, and the man who built that neighbour into an empire did it by teaching his guard to shoot whatever he shot at.

Tuman's mistake

North of the Ordos the Xiongnu were ruled by a shanyu — Parker renders it Jenuye — called Tuman, who in his old age fell under the influence of a much loved young queen and agreed to make her son his heir in place of the legitimate one.

That legitimate son was Modu, and he was an able captain. To be rid of him, Tuman sent him as a hostage to the Yuezhi and then attacked them, expecting them to kill the boy in their indignation. Modu took one of their fastest horses and rode home.

His father was so pleased with the feat that he gave him command of ten thousand troopers.

The arrow

Modu was not pleased. He invented a new kind of arrow — a whistling arrow, of the sort the early Japanese called nari-kabura — and made a rule: whatever he shot at with it, every man of his was to shoot at instantly.

He tested them on his own best horse. Several troopers who did not shoot were executed. He tested them on his favourite wife. Several more were executed.

Then he waited until his father was out hunting, and fired a whistling arrow at him. Tuman fell dead, pierced through and through, and Modu was proclaimed in his place; a general massacre of his father's family and household followed.

Rope

What he did next is the part worth reading, because it is not the behaviour of a berserker.

The Donghu — Tungus, on the other side of a three-hundred-mile strip of neutral desert — heard about the parricide and sent to demand his finest horses as the price of leaving him alone. Modu publicly rebuked the members of his council who wanted war and sent the horses.

The Donghu, encouraged, demanded one of his favourite wives. The council was appalled. He sent her.

Then, having concluded that he had no stomach for a fight, they massed troops and demanded the neutral strip itself. Some of his councillors were inclined to give up what they called a useless piece of land.

He had them beheaded, declared war, and announced that any able-bodied man who did not report to headquarters at once would be executed. The Donghu, who had taken no precautions because they had decided he was a coward, were broken in a single campaign; their herds were driven off and most of the people enslaved. A remnant escaped north-east of what is now Peking, where — Parker notes, because he is writing a thousand years of this — they became a formidable power of their own in due course.

What Modu had made

An empire of the steppe with a chain of command that reached from the shanyu to the individual trooper, held by terror and drilled to act on a signal.

Parker's observation about how such things are assembled is worth having, because it explains why the same ground produces one of these every few centuries. After a great battle the captured young men became warriors, usually under their own chiefs but subject to the conqueror; the old men were turned out to tend the flocks; and the flocks changed masters for a few years until the next revolution. Master and slave lived much the same way. Languages mixed, tribes merged, and no horde can be given a fixed location or an exclusive nationality.

That is the thing the Han now had to deal with on a frontier two thousand miles long, and it had just finished its own civil war.

Teeth and hair

Drafted from Parker, A Thousand Years of the Tartars (1895)expansion

In 200 BC the founder of the Han dynasty took an army of three hundred and twenty thousand men north to deal with the Xiongnu and spent seven days surrounded, cut off from his own army, in a fort near modern Datong.

Pingcheng

It was a bitter winter and about a quarter of the Chinese host had lost fingers to frostbite before the fighting started.

Modu used the standard Xiongnu method — feigned defeat, best troops out of sight, and a long enticing retreat northwards — and Gaozu, who had outrun the bulk of his own infantry, walked into it. Three hundred thousand Xiongnu horse closed round him.

Parker preserves the detail that makes it real: the white, grey, black and chestnut horses were massed in four separate bodies, one at each point of the compass. It must have been, he says, a very picturesque affair.

Then Modu hesitated, suspecting a trap, and lost his chance — as Attila would at the Catalaunian Plains six centuries later. How the emperor got out is not recorded, but it is clearly hinted that discreditable negotiations were conducted with the shanyu's consort, and that one corner of the encircling line was left unguarded. Gaozu rode out through it.

The price

What followed was the settlement that governed the next seventy years, and it is worth stating in its own terms because later Chinese history is careful not to.

A princess of the imperial blood was sent to the shanyu in marriage. With her went an annual subsidy: silk piece-goods, waste silk for wadding garments, rice wine and choice foodstuffs.

The man who proposed this policy was made the envoy who carried it out, and the official rationale was long-term — the children of a Chinese consort would one day be shanyu, and the empire would have a claim on them. Parker points out what actually happened to that reasoning five hundred years later, when Xiongnu princes descended from Han consorts claimed the throne of China as the only surviving legitimists, and got it.

The subsidy moderated the raiding. It did not stop it.

The letter

Gaozu died. His widow, the empress dowager Lü, put her own son on the throne, had her rival killed with great cruelty, and then ruled for nearly a decade in her own name as a recognised monarch.

Modu — prompted, Parker suspects, by one of the many Chinese renegades in his service — wrote to her offering his hand and heart.

The letter caused a tremendous flutter in the imperial council, and the question was whether to send back a civil answer or the envoy's head. There were braggart generals ready to fan the dowager's fury. What cooled her was one cautious old statesman reminding her that boys in the streets were still singing a popular song about her late husband's narrow escape from the beleaguered town.

So a very diplomatic reply went north. It thanked the shanyu for the honour, and pointed out that the condition of the dowager's teeth and hair was unequal to the task of securing his affections. Two royal carriages and teams of horses went with it.

Why this is the beginning of the record and not an anecdote

Because it fixes what the Han was for the next seven decades: a state that had unified China, inherited the Qin's machinery, and could not defend its own northern frontier except by payment.

Every institution the dynasty is remembered for was built in the reaction against that. The state monopolies were created to finance the war that ended the subsidy. The road west was opened to find allies against the Xiongnu. The conscript armies, the frontier colonies, the horse-breeding programmes and the enormous deficits all begin with the seven days at Pingcheng.

A great deal of Chinese imperial history is the working out of one bad afternoon in the snow.

The road west

Drafted from (eds), The Cambridge History of China, Volume 1: The Ch'in and Han Empires, 221 BC-AD 220 (1986)expansion

The Han went west for a military reason and got a trade route by accident, and the accident lasted longer than the reason.

Zhang Qian

In about 139 BC Emperor Wu sent an envoy to find the Yuezhi, a people the Xiongnu had driven out of the Gansu corridor a generation earlier, and propose an alliance against them.

Zhang Qian was captured by the Xiongnu almost at once and held for about ten years. He escaped, went on west anyway, found the Yuezhi in Bactria — settled, prosperous and entirely uninterested in a revenge war three thousand kilometres behind them — and came home after thirteen years with two of the hundred men he had left with.

He had failed. What he brought back instead was a description: Ferghana and its horses, Bactria, Parthia, Sogdiana, a settled and wealthy world west of the desert that nobody at Chang'an had known was there, and — the detail that mattered most to the court — Chinese goods on sale in Bactrian markets, which meant a route already existed.

What was actually being bought

The Han did not go west for silk. It went west for horses, and for the flank.

The horses are literal. Ferghana bred an animal larger than anything in China, and Wu wanted them badly enough to send two expeditions across the Taklamakan — the second with sixty thousand men — to get them, in a war fought over bloodstock. The strategic argument was better: the Gansu corridor separated the Xiongnu from the Qiang peoples to their south, and holding the Tarim oases cut the Xiongnu off from the settled communities whose grain and manufactures a chanyu had to distribute in order to be obeyed.

Wu took the corridor in the 120s BC and planted four commanderies across it. Everything else follows from that: garrison colonies, a chain of posts and beacon towers running out into the desert, and a permanent contest with the Xiongnu for the oases.

The cost

It nearly bankrupted the state, and the record of it is unusually good because the Han argued about the bill in public.

The campaigns ran for decades and consumed the treasury. The horse losses were catastrophic. To pay for it the government took over the production of salt and iron, sold offices and titles, debased the coinage, and imposed a property tax on merchants that was enforced by inviting informers to denounce them for a share of the proceeds.

By the end of Wu's reign the empire was larger, the frontier was quieter and the fiscal system had been rebuilt around monopolies. In 89 BC the emperor issued an edict of repentance, abandoning further expansion.

What was left when the reason went away

The Xiongnu split in the first century BC and stopped being the threat the road had been opened against. The road stayed open.

Silk went west along it in quantity, through Parthian and Sogdian intermediaries, to Rome, where the elder Pliny complained about the drain of gold. The two empires knew of each other at several removes and never made contact: an envoy sent west in 97 got as far as the Persian Gulf and was talked out of going further by Parthian intermediaries with every interest in keeping the trade indirect.

Buddhism came the other way, along the same route, into an empire that had opened it to buy cavalry.

Something in the sea

Drafted from Hirth, China and the Roman Orient: Researches into their Ancient and Mediaeval Relations as Represented in Old Chinese Records (1885)expansion

In AD 97 the general Ban Chao, who had spent thirty years making the Tarim basin Chinese, sent an officer named Gan Ying west to find the country the Chinese called Da Qin. He reached the Persian Gulf, asked about the crossing, and turned round.

What he was told

The account is in the Hou Han Shu, and Hirth translates it without softening anything.

Gan Ying arrived at Tiaozhi, on the coast of the great sea, and was about to take his passage. The sailors of the western frontier of Parthia told him:

The sea is vast and great; with favourable winds it is possible to cross within three months; but if you meet slow winds, it may also take you two years. It is for this reason that those who go to sea take on board a supply of three years' provisions. There is something in the sea which is apt to make man home-sick, and several have thus lost their lives.

When Kan-ying heard this, he stopped.

Who was telling him

Parthian sailors, on the Parthian shore, about a voyage Parthian middlemen made their living from.

The route from Tiaozhi to the Syrian port of Aelana at the head of the Gulf of Aqaba was regular traffic; Gan Ying was talking to the very men who made it. A crossing described as three months with a fair wind and two years with a foul one, requiring three years' stores, is not a description of that voyage. It is a description designed to stop somebody making it.

The Parthians had every reason. Chinese silk reached Rome through their hands and was taxed on the way, and a direct Han embassy to the Roman east would have made that unnecessary. The one sentence about something in the sea that makes men homesick, and has killed several, is the flourish of a man who knows his audience is a long way from home.

Hirth's judgement is that Gan Ying, having talked to the sailors who actually made that journey, had the best opportunity anyone had for collecting information about the object of his mission. He collected it, and it was wrong on purpose.

What the Han knew anyway

Enough to want to go.

The Chinese histories record an embassy in AD 87 offering lions and a hornless unicorn, and in 101 the Parthian king sending lions and large birds from Tiaozhi which were thereafter called Parthian birds — ostriches. They record the road west in stages with distances, from Parthia to Aman to Sibin to Yulo on the extreme western frontier, and then south by sea to Da Qin.

They also record, later, that Da Qin was governed from a great city, that its people were tall and honest and resembled the Chinese — which is why it is called Da Qin, the great Qin — that they made money of gold and silver, and that they had always wanted to send embassies to China and been prevented by the Parthians, who wanted to keep the silk trade in their own hands.

That last claim is in the Chinese record, made by the Chinese, about the people who had just talked their envoy out of crossing.

What did not happen

The two ends of Eurasia were never joined.

An embassy arrived at the Han court in AD 166 claiming to come from the king of Da Qin, Andun — Marcus Aurelius Antoninus — and brought ivory, rhinoceros horn and tortoiseshell, which the Chinese officials noted were ordinary tropical goods and not the treasures a great empire would send. The likeliest explanation is a party of merchants improving their standing at Canton.

So the record of contact between Rome and Han China amounts to a mission turned back by a shipping line's sales talk, and a probable fraud sixty-nine years later. Two empires of comparable size, each of which knew the other existed, each of which had a name for the other, neither of which ever heard from the other directly.

When Kan-ying heard this, he stopped. Very few sentences in any record cost so much.

What four centuries built

Drafted from (eds), The Cambridge History of China, Volume 1: The Ch'in and Han Empires, 221 BC-AD 220 (1986)peak

The Han inherited the unified state the Qin had built by force and made it durable, which is the harder achievement. Over four centuries they established the machinery that Chinese empire used until the twentieth: a salaried bureaucracy recruited partly on merit, a state monopoly on iron and salt, official histories, and Confucian doctrine as the training and the justification of officials.

The machine

The list is worth being specific about, because every item on it outlived the dynasty by a millennium and a half.

A salaried civil service, paid in grain and cash rather than in land, appointed and transferable by the centre and therefore not hereditary. A territory divided into commanderies and counties with a governor in each, appointed from the capital. A census of households and individuals, taken regularly, on which the tax and the conscription were assessed — the earliest such counts anywhere that survive as figures. A uniform written law. A standing granary system for price stabilisation. A state monopoly on salt and iron. An imperial academy teaching a fixed canon. And an official history of the preceding dynasty, compiled by the succeeding one, which became the obligation of every dynasty afterwards and is the reason China has a continuous documentary record that no other civilisation can match.

What the Han added to the Qin

The Qin had built most of the apparatus and had lasted fifteen years.

What the Han added was a theory of why the apparatus should be obeyed, and it was explicitly a theory about the Qin's failure: that an empire taken by force cannot be held by the same means. The doctrine that emerged — a Confucian ethics of government grafted onto Legalist machinery — made the official's education, rather than his birth or his troops, the ground of his authority, and made the ruler answerable to something other than his own will.

It was frequently a fiction in practice. It was still the language in which every subsequent dynasty had to justify itself, including every foreign one.

The dividing line

The dynasty is conventionally split in two, and the break is real.

The Western Han ruled from Chang'an until AD 9, when the regent Wang Mang took the throne outright and declared a new dynasty. He held it for fourteen years, attempted a radical programme — land nationalisation, abolition of slavery, currency reform — and was destroyed by rebellion, famine and the Yellow River changing its course. A Han claimant restored the dynasty in AD 25 and ruled from Luoyang.

The Eastern Han was never the same state. The great landholding families that had raised the restoration were the price of it; they were never brought back under control; and by the second century the provinces were run by magnates with private armies and a court paralysed by the struggle between eunuchs and officials.

The dynasty fell apart through that century into warlordism and ended in 220, pulled to pieces by its own generals and provincial families. The three kingdoms that followed each claimed to be its legitimate continuation, and one of them took the throne by an abdication ceremony rather than a conquest — which is itself a Han legacy, and the form every later usurpation would use.

Salt and iron

Drafted from (eds), The Cambridge History of China, Volume 1: The Ch'in and Han Empires, 221 BC-AD 220 (1986)peak

From 117 BC the Han state took the production and sale of salt and of iron, and later of liquor, out of private hands. It is the most thoroughly argued fiscal institution anywhere in the ancient world, and the reason we can argue about it is that both sides of the argument survive.

What was taken over

Emperor Wu had been at war with the Xiongnu for twenty years by then, and the ordinary taxes did not cover it. The land tax stood at a thirtieth of the harvest and was deliberately light; the poll tax fell on people who could not pay much more.

Salt and iron were different. Everybody needs salt, nobody can make it at home, and it is produced at a small number of identifiable places — coastal pans and inland wells. Iron is the same shape of commodity. A state that controls a few hundred production sites can tax an entire population without knowing where any of them live.

That is the whole idea, and it worked well enough that versions of it were still being used in China two thousand years later.

The transcript

In 81 BC, six years after Wu died, the court held a debate on whether to keep the monopolies.

On one side were the officials who ran them, arguing fiscal necessity and frontier defence. On the other was a delegation of Confucian scholars, arguing that the state had gone into business, was competing with its own people, and had made profit a respectable motive at court. A version of the proceedings was written up a generation later as the Discourses on Salt and Iron, and it is a substantial book.

The scholars lost on salt and iron and won on liquor.

It is worth appreciating how unusual that record is. For most of the states on these pages, we know what was taxed and can only infer why. Here the fiscal policy of an empire was put on trial by its own officials, in front of the throne, and somebody wrote down what both sides said.

Recruiting a bureaucracy

Drafted from (eds), The Cambridge History of China, Volume 1: The Ch'in and Han Empires, 221 BC-AD 220 (1986)aside?What aside means

The Han began the practice that separates Chinese imperial administration from every other kind there has been: recruiting officials by examination on a body of texts, rather than by birth, purchase, military service or personal following.

Under the Han it was partial — recommendation by local officials mattered more than examination, and great families dominated office in practice — and the system only took its mature form centuries later. But the principle was established, along with the Imperial Academy to teach the canon and the idea that the texts an official had studied were the qualification for governing.

Set that against how other empires solved the same problem. The Abbasids ran on a Persian secretarial class and then on Turkish slave soldiers. The Umayyads ran on Syrian tribal levies. The Mamluks were governed by their own purchased troops; the Ottomans by a slave-recruited administration; the Mongols by whoever could count. Every one of those solved the problem of finding reliable servants by controlling the servant's origin — buy him young, take him from a tribe, import him from elsewhere.

China solved it by controlling the servant's education. It is the only durable alternative anybody found, and it is why the Chinese state survived conquest by the Khitan, the Jurchen, the Mongols and eventually the Manchus with its administrative machinery substantially intact. Every one of those conquerors had to decide what to do about a bureaucracy that regenerated itself from books, and every one of them decided to keep it.

The conscript empire

Drafted from (eds), The Cambridge History of China, Volume 1: The Ch'in and Han Empires, 221 BC-AD 220 (1986)contraction

The Han put a peacetime military obligation on its whole adult male population that almost no pre-modern state has matched.

Every registered adult male owed two years: one of training and service in his home commandery, one on the frontier or in the capital guard, with a liability to be recalled until he was in his fifties. The registers that made it possible are the same registers the poll tax was assessed on, and a census of 2 CE survives that counts fifty-seven million people in something over twelve million households.

That is the earliest population figure of that scale from anywhere on earth, and it should be read for what it is: a tax document. It counts households the state could reach, not people who existed, and the difference between those two numbers is the entire fiscal history of the dynasty.

What it was spent on

The Xiongnu, mostly, and the two states are a matched pair: a settled empire that could raise enormous numbers of infantry cheaply and could not force a battle, against a mobile confederation that could always decline one.

Wu's answer was to buy cavalry — horses from Ferghana, and eventually Xiongnu cavalry who had surrendered — and to push the frontier west along the Gansu corridor, which is what put Han garrisons in the oases and opened the road west.

The failure mode

Conscription of that kind depends on a registered smallholding peasantry, and the Han did not keep one. Land accumulated: by the first century CE much of the population was tenants of large estates, off the registers and unavailable to the state, and the militia system decayed accordingly.

What replaced it, by the end, was provincial commanders raising their own troops to suppress the Yellow Turban rising. They did suppress it, and then would not disband, and the empire was divided between them.

The Tang ran the same sequence seven hundred years later with different names: a militia resting on an allotted-land system, the land system eroding, professional regional armies replacing it, and one of their commanders taking the capital. Two of the longest dynasties in Chinese history, ended the same way by the same mechanism — and the second one had the first one's history written down and on the shelf.

The Yellow Turbans, and the abdication

Drafted from (eds), The Cambridge History of China, Volume 1: The Ch'in and Han Empires, 221 BC-AD 220 (1986)end

The Han ended twice: once in fact, in 189, and once on paper, in 220, and the thirty-one years between the two are the useful part.

What was already broken

By the middle of the second century the court was a three-way fight between the emperor's consort families, the eunuchs of the palace, and the scholar-officials who had been trained to think of themselves as the empire's conscience.

The officials lost. In the proscriptions of the 160s and 170s hundreds were banned from office, imprisoned or killed for factional association, and a class that had been the state's administrative backbone was taught that the throne was its enemy. That lesson did not unlearn itself.

Underneath the court, the fiscal base had been eroding for a century in the way these systems erode: land accumulating in great estates, the households on them disappearing from the tax registers, and the registered population — the people who paid and served — shrinking while the actual population did not.

184

The Yellow Turbans were a religious movement, which is the part usually skipped over.

Zhang Jue's Way of Great Peace was a healing cult with a doctrine, an organisation of thirty-six divisions across eight provinces, and a date: the jiazi year, 184, at which the blue heaven would die and the yellow heaven be established. Hundreds of thousands rose in the same season across half the empire.

It was suppressed within about a year, and suppressing it destroyed the dynasty.

To fight a rising on that scale the court gave provincial governors military authority and let commanders raise their own troops locally. The armies that resulted belonged to the men who had raised them. Within five years those men were the real government, and in 189 one of them marched into the capital, massacred the eunuchs, deposed one emperor, installed another and burned Luoyang.

The emperor who lasted thirty years

Emperor Xian was put on the throne as a child by a warlord and spent his entire reign in the custody of one commander or another. From 196 he was held by Cao Cao, who used him exactly as such a person is used: every campaign Cao Cao fought was fought under an imperial commission, against rebels, on behalf of a dynasty he was in the process of replacing.

That is worth being clear about, because it explains why the formal ending is so late. The Han name was worth having. A general who held the emperor could issue lawful orders to men who would not have obeyed him otherwise, and the whole point of the mandate doctrine — that legitimate authority is a real thing and not merely a possession — made the puppet valuable.

In 220 Cao Cao died and his son stopped bothering. Emperor Xian abdicated in his favour, in a carefully staged ceremony with three ritual refusals, and was given a fief and allowed to live. He outlived the man who took his throne by several years.

What ended

Not the country, not the institutions and not the idea. The commandery system, the registers, the examinations, the canon and the mandate doctrine all went straight into the states that fought over the wreckage, and all of them claimed to be continuing something.

What ended was four hundred years of one family, and the population registers. The census of 2 CE counted fifty-seven million people. The next one anybody was in a position to take, in 280, counted about sixteen million.

Almost nobody now reads that as a death toll of forty million. The state that did the counting had come apart, and the great estates that had been absorbing households for two centuries had no interest in reporting them. Uncounted is not dead. But the fact that the number could fall that far, and that no government could establish what had actually happened to its own people, is a better measure of what the end of the Han was than any battle in it.

Two claims

Drafted from (eds), The Cambridge History of China, Volume 1: The Ch'in and Han Empires, 221 BC-AD 220 (1986)afterlife

Two later states claimed to be the Han, and the second of them is one of the strangest claims of legitimacy anybody has ever made.

Shu Han

When the empire broke into three in 220, the ruler of the western kingdom, Liu Bei, was a member of the Han imperial clan by a distant line, and named his state Han. He was not claiming to succeed the dynasty; he was claiming to be it, continuing, with the other two kingdoms as usurpers.

Whether the claim was any good is a separate question from whether it was made. Chinese historiography spent the next fifteen hundred years arguing about which of the three kingdoms held the legitimate line — Chen Shou put the Wei first, Zhu Xi put Shu Han first, and the argument was still live in the Qing — which is about as clear a sign as there is that people took it seriously.

Liu Yuan

In 304 a Xiongnu leader declared a dynasty in Shanxi and called it Han.

His argument was this. The Han had bought peace on the northern frontier for generations with the heqin treaties — a Chinese princess sent north to marry a Xiongnu ruler, with silk and grain. His line descended from one of those marriages. He therefore stood in the maternal line of the Han imperial house, and when the Chinese dynasty then reigning collapsed, he was as entitled to the mandate as anyone.

He sacrificed to the Han emperors, took a Han reign style, and his successors held both Chinese capitals within a decade.

The Han and the Xiongnu are usually read as a frontier: the empire and the confederation, the wall and the steppe. What this claim says is that the frontier was also a family relationship, created deliberately by the Han as an instrument of policy, and that four hundred years later somebody presented the bill.

It is about the only claim to an empire ever built out of a treaty rather than out of a conquest or a line of rulers — and the treaty was one the claiming side had been paid to sign.