1. Too Heavy to Carry

To understand why the world's first paper money appeared in Sichuan, you first have to weigh what money actually weighed in an eleventh-century Sichuanese hand.

After the Northern Song unified the Shu region, it let the local population go on using iron coin, and for a long stretch forbade copper coin from entering the two Chuan circuits at all; later the ban was partly lifted, but iron coin itself was still not allowed to leave the region. Iron is worth little, so buying the same object took more of it, and more iron meant more weight. Merchants who carried copper coin across the border found that a single copper coin bought fourteen iron ones. How many iron coins it took to buy a bolt of cloth, how many porters had to be hired to carry them over a mountain road — these were not abstractions. They were arithmetic you felt in your shoulders.

The reason iron coin was so heavy is simple enough: what a coin is worth depends partly on what it is made of. Copper is worth more than iron, so the same price in iron takes more metal, and more metal means more weight. The histories preserve some startling figures — that buying a single bolt of silk could require hauling in a hundred catties of iron coin. Buying a bolt of silk was an ordinary household transaction, not a great commercial undertaking, and yet this ordinary errand first required solving the separate problem of how to get the money to the shop at all.

Taxes made it worse. At one point the government required that rent payments and monopoly fees be reckoned at a rate of ten iron coins to one copper coin. And copper coin was already scarce in Sichuan — the histories say outright that copper had run dry, and that the people suffered bitterly for it.

The History of Song preserves a few lines that are hard to read past. In the years when the government pushed hardest for tax payment in copper, local officials and clerks seized the opening, layering the conversion rate higher with each pass, so that the people suffered more and more; some dug open ancient graves and smashed Buddhist statues and ritual vessels just to scrape together four or five copper coins, and a great many of them were convicted for it.

These lines deserve a slow reading. An unsuitable currency system, translated down to one actual person, meant that person had to rob someone else's grave, had to smash the Buddha his own household kept, only to assemble a few pieces of copper in the shape the government demanded — and quite possibly be sentenced to punishment for the trouble.

The common scale had not disappeared here. It was still there. It was simply too heavy — heavy enough that a man had to carry it on his back, heavy enough that meeting its specifications could force him into acts he would never otherwise have committed.

Something surfaces here that will keep recurring. Whether a scale is good depends not only on whether it reads true, but on whether it is light enough to carry and easy enough to recognize. In the earlier essays, the difficulty of commensuration always came down to whether a thing could be reckoned clearly at all. In Sichuan the difficulty took a different shape: the reckoning could be done perfectly clearly, and then the money that resulted from it could not be moved. A thing that cannot be moved is no use to anyone, however accurate it is.

A remainder is already buried here. Iron coin could price a bolt of silk down to the last coin; what it could not price was the half-day's labor of hauling that price over a mountain, the wages of the men hired to carry it, the risk of being robbed on the road. None of that entered any ledger, and yet all of it decided whether the transaction could happen at all. This — precisely this piece of unaccounted cost — is what jiaozi (交子) was built to solve.

And so someone thought of a piece of paper.

2. Sixteen Names You Could Trust

The History of Song records the invention in a single sentence: at first, because iron coin was too heavy, the people of Shu privately produced vouchers they called jiaozi — the name means little more than "exchange notes" — to ease trade, run by a consortium of sixteen wealthy households.

Li You's later Facts of the Song Dynasty gives a fuller picture. More than a dozen powerful households in Yizhou jointly guaranteed the notes, printed on uniform paper stock, the face carrying images of houses, trees, and human figures; each shop stamped its own trademark and a secret code number, red ink and black ink crossing in private marks of authenticity. The amount was filled in by hand on the spot, with no ceiling on how many strings of cash — guan — could be written in. Redeeming a note back into cash cost thirty wen per guan in handling fees.

It is worth pausing on what actually appeared on that paper. Uniform stock, fixed images, each house's stamp, a secret code, red and black marks crossing in private patterns — not one of these existed to state what the note was worth. All of them existed to state who had issued it, and whether anyone else could counterfeit it.

Put differently: nearly all the craft that went into that sheet of paper was spent on authentication and signature, not on price. The line for the amount was, if anything, the easiest part — filled in on the spot, any figure would do. What was genuinely hard was making a stranger believe that this particular note was real, and that it truly came from those specific households. Metal coinage never had to work this hard: whether a copper coin is genuine can be checked by weighing it, biting it, looking at its cross-section. Paper cannot be checked that way. A paper note's truth does not live in the paper itself.

Why would a sheet of paper make a total stranger willing to accept it? Not because of the paper. Because of those sixteen names.

This is the crux of the whole matter. Iron coin worked partly because the government recognized it and partly because a coin is, in itself, a lump of iron — whoever holds it, that lump of iron is still there. Paper is not like that. Paper by itself is nothing at all. What it is worth depends entirely on whether the people standing behind it are willing to make good on the account, and able to.

Stripping the metal out of money did not make money need trust any less. Quite the opposite: however much metal was removed, exactly that much weight came to rest on trust.

In the earlier essays, the anonymous scale was always propped up by something personal: clay tablets leaned on seals and witnesses, coinage leaned on an issuer's reputation and a later owner's test-mark, ledgers leaned on a cabinet full of letters and guarantors. With jiaozi this relationship was pushed all the way to its limit, because beyond those few names, the paper itself held nothing whatsoever. It was a pure promise, with no metal underwriting it at all.

Trouble followed close behind. The wealthy merchants who took in iron coin deposits did not leave the coin sitting obediently in a vault; they used it to buy up stock and hoard goods, to acquire shops, houses, gardens, land, and valuables. Forgery appeared. Lawsuits appeared. Runs on the note-issuers appeared, and eventually crowds gathered to fight over the matter. When officials stepped in to mediate, the going rate of redemption was often only seven or eight hundred wen on the guan — and the two or three hundred wen of loss fell on whoever was holding the note.

This lesson would repeat itself, at a far larger scale, many times over the following centuries: the issuer holds someone else's trust in hand, and trust can be spent before it is earned.

The temptation here is not hard to grasp. A pile of iron coin sits in the vault, but the people coming to redeem their notes trickle in a few at a time — a handful today, a handful tomorrow — almost never all on the same day. Given that, why not take part of the reserve and put it into property, since property earns a return, and the return can buy still more property? Every step in this reasoning holds up. The only flaw sits in the word "almost." The one day everyone actually does show up together, every earlier step of the logic fails at once.

On the question of where this paper's roots really lay, researchers split into two camps, and the disagreement runs deep.

One camp stresses networks of trust. He Ping specifically resists reading early jiaozi as credit money in the modern sense; in his account, the crucial thing was not some abstract notion of credit but the relatively closed geography of Sichuan itself, the standing and solvency of the tea-merchant guild, the joint guarantee of the sixteen wealthy households, and the web of local acquaintance. A note could travel from one shop into a stranger's hand not because people had suddenly stopped caring who stood behind it, but because the reputation of that particular circle temporarily supplied a public credibility broad enough to carry it.

The other camp puts more weight on the instrument itself — on debt and claim. On this older reading, jiaozi is best understood as something like a deposit receipt issued by a private vault house: the issuer is the debtor, the holder is the creditor, and once the note has changed hands long enough in the market, it comes to function as money in its own right. In this version, jiaozi's origin lies in the relationship of deposit and redemption, not in some diffuse social trust.

Both accounts are looking at the same piece of paper, and finding different roots. One says it grew out of relationships. The other says it grew out of debt.

The disagreement is not merely scholarly fastidiousness. If the note grew out of relationships, its roots belong to one particular place and one particular circle of people, and it would not work outside Sichuan. If it grew out of debt, its roots belong to a form that can be reproduced anywhere — wherever someone issues a redeemable claim, the same thing can take root. The first account explains why jiaozi happened to appear precisely in Sichuan; the second explains why it could later be taken over by the state and turned into something else. Each theory has hold of one true end of the rope.

There is even dispute, incidentally, about who actually shaped the thing. Many local accounts and later popular histories cast Zhang Yong as the father of jiaozi, pinning his reorganization of the jiaozi shops and his designation of the sixteen guaranteeing households to specific years. Some more recent research pushes back, pointing out that the family records and epitaphs Zhang Yong's own descendants left behind never claim that he invented jiaozi. The safer statement may simply be that Zhang Yong happened to be in office during the period when the folk practice of jiaozi was taking its final shape.

That not even the authorship can be pinned down is itself telling. A thing like this is rarely dreamed up whole by one person on one day. More often, many people are each solving the trouble directly in front of them, and solving it, and solving it again, until it has quietly grown into an institution.

3. The Bronze Seal

The trouble with private jiaozi is stated plainly in the sources: the fortunes of those wealthy households gradually declined, debts went unpaid, and litigation never let up.

In the eleventh month of the first year of Tiansheng — 1023 — the court issued an edict establishing the Yizhou Jiaozi Bureau. The following year, official jiaozi went into circulation, with the first issue fixed at a quota of 1,256,340 strings of cash.

That figure, precise to the last digit, tells you everything about the nature of the change.

The private note had been filled in on the spot, with no ceiling on the amount; this official issue was fixed in advance, down to the unit. The old amount was set by the needs of each individual transaction; the new amount was set by a single administrative document. The same piece of paper went from something that followed the market wherever it needed to go, to something whose total was calculated from above first and handed down afterward. Those two words — fixed quota — are the backbone of the entire official apparatus.

The Facts of the Song Dynasty records the details of the new system: a dedicated office was established, a bronze seal was cast for the Jiaozi Bureau, ledgers were kept, cash was taken in and stored, notes were issued against contract code numbers, and the same thirty wen per guan was still deducted as an ink-and-paper fee.

From this point on, what stood behind the paper was no longer the reputation of sixteen merchants. It was an entire government apparatus.

And that apparatus can be itemized point by point. Denominations were standardized — starting at one to ten guan, later changed to five and ten guan, and later again to one guan and five hundred wen. The cycle was standardized, with printing and exchange happening on fixed schedules. Quantity was standardized, each issue carrying an explicit ceiling. Procedure was standardized, thirty wen on every guan. Issuance and recall were both logged. Even forgery and physical wear were folded into a single procedure of verification by seal-mark and recall by deadline.

Money, at this stage, looked less and less like a piece of metal and more and more like a set of administrative and accounting operations.

This is worth turning over. Coined money — copper, silver — begins with the metal, and institutions arrive afterward to manage it. Jiaozi-type money begins with the institution, and only then does the paper become money at all. Strip away the government's seal, its ledgers, its fixed quota, its exchange schedule, its promise of redemption, and what is left is a printed sheet of paper and nothing more. So it is here, for the first time, that money becomes something wholly manufactured by institutions — its very mode of existing is a set of rules that can be written down, enforced, and also changed.

Reserves appeared too. The History of Song states that, as a rule, producing one issue in the old years required a reserve of 360,000 strings, old funds carried forward to support the new. He Ping, citing Li Gang's writings, offers a different figure: a standing reserve of 1,000,000 guan supporting 3,000,000 guan of jiaozi in circulation.

The two figures do not agree. That mismatch is itself instructive: what later generations think of as "the reserve ratio" was never a single fixed institutional constant, but something that varied by period, by source, and by how the record was kept. What can be said with confidence is this: the Song themselves understood perfectly well that paper money, to function, could not be left entirely unsupported.

On this point the Song said it more clearly than anyone since. Commenting on the founding of the Jiaozi Bureau in the Tiansheng years, Lü Zhong wrote: this was the beginning of paper currency; and yet it had money to put the paper into motion, and paper to regulate the money, mother and child circulating together without either side being favored, so that the people looked on cash as paper and looked on paper as cash.

Mother and child in balanced circulation — four characters that remain, to this day, close to the earliest and most exact description anyone gave of the whole affair.

Paper did not take flight on its own. It needed a mother.

The mother-and-child figure is well chosen, because it does not describe either metal or paper as replacing the other. The mother does not vanish for having a child, and the child is not thereby invalidated for having a mother; the two must exist together, answering to each other, neither able to do without the other. An eleventh-century observer had already grasped that the crux of the matter was never how convenient the paper itself might be, but whether that back-and-forth relationship between paper and coin could hold its footing. Every crisis of the next several centuries would break out precisely along this line.

4. The Loop That Wouldn't Close

Of the whole system, the most ingenious device was what the sources call an issue.

Here is how it worked: every few years a new batch of notes was printed, and at the same time the entire previous batch was called in and invalidated, exchanged one-for-one for the new. Each such batch was called an issue.

The design was genuinely clever. Paper money's greatest hazard is endless printing — the more that gets printed, the less it is worth — but with this system, the field was cleared every few years: the old notes came back and were destroyed, the new ones went out against a fixed quota, and the total in circulation stayed tethered to a controllable rhythm. Issue, circulate, recall, reissue: a loop that could close on itself.

The thinking behind this design runs in the same vein as the efforts, described in the earlier essays, to make a ledger balance — only turned in a different direction. A ledger has to balance in space, its two sides forced into equality. An issue had to balance in time: what went out was required to come back on schedule, and only once it was back could new notes go out again. The former governs whether one set of books adds up; the latter governs whether one round of currency can be called home. But both point at the same underlying wish — that this business ought to have a form that can be brought to a full and final close.

But the more precisely a thing is designed this way, the more precisely the remainder shows its face. When a ledger fails to balance, you can see at a glance how much is missing; when an issue's currency fails to come home, you can see at a glance how much is owed. The finer the graduation, the less anywhere there is for what leaks out to hide. The reason that postponing an issue's exchange reads as so glaring is exactly this: it leaves a loop that was supposed to close standing nakedly open, in plain view of everyone.

But whether the loop actually closes depends on one thing: whether the exchange genuinely happens when it is due.

And whether the exchange genuinely happens when it is due depends on another thing entirely: whether the state, at that moment, has some more urgent use for its money. This is the softest point in the whole clever design. It ties control of the total supply to an action that only the issuer itself can perform, and the moment the issuer is short of funds, the cheapest way out is precisely to decline to perform it. However finely the rule is drawn, if the party enforcing it is the very party most tempted to break it, the loop cannot be relied on.

Someone first tried to export the system. In the early Xining years, the court attempted to transplant the jiaozi law to Hedong and Shaanxi, and ran quickly into opposition from local officials. Wen Yanbo and Zhang Jingxian both argued that the method could work in Shu but not in Shaanxi, and it was soon abandoned.

The reasoning holds up well. In Sichuan, paper and iron coin had been paired together specifically to solve the problems of mountain roads, of heavy loads, of obstacles to regional circulation; move the pairing to a different fiscal structure, a different structure of goods, different regional conditions, and the combination need not hold at all. From the very start, jiaozi's success was never a frictionless template that could simply be copied elsewhere.

This is worth remembering, because later ages like to repeat the glib line that a good method spreads on its own. The record shows something else: the same method worked here and failed there once transplanted, and the objection was not conservatism but the plain fact that local conditions differed. An institution is not a blueprint detached from its soil; it grows in a particular terrain, on a particular currency, within a particular tax system. Transplanting it is possible. It is never automatic.

What actually crushed the loop was military spending.

In the fifth year of Xining, as the twenty-second issue was about to turn over, the following issue had already been overspent, and the government was forced to over-print a twenty-fifth issue to fill the hole left by the twenty-third — so that two issues now circulated side by side. The History of Song does not soften what came next: at that time jiaozi was issued in such quantity, and hard cash was so short, that prices collapsed, and eventually there was no real money behind the notes at all, and the whole method could no longer be carried on.

Looking back from the first year of Daguan, the assessment was blunter still: since the campaigns to take Huang, Kuo, and Xining had begun, jiaozi had been used to subsidize the frontier's expenses, running to twenty times the volume of a single Tiansheng-era issue, and its value had sunk further with every step; by the time old notes came due for exchange, four old guan bought only one new guan.

What does a ratio of four to one actually mean? It means that the paper in your hand, brought in for exchange, is worth exactly one quarter of what it was.

Behind that number is one very specific kind of person. Someone works a hard year and saves a few guan of jiaozi, expecting that next year it will exchange for new notes just as it always has, and when the exchange finally comes, four old guan buys one new one. He did nothing wrong. He did not gamble, did not borrow, did not speculate. He simply accepted the money the government issued, and trusted it. The loss landed on him regardless.

The lesson is easy to see from here: paper money never actually escaped the problem of backing. It is only that once the treasury is desperate to spend, the state grows steadily more willing to push that backing a little further down the road.

Recent scholarship often frames this period as the advance of the fiscal state. Some research finds that by the late Northern Song a substantial share of state revenue already came from non-agricultural, monetized indirect taxation; by around the Southern Song, the volume of redeemable credit instruments in circulation is said to have at times exceeded the scale of annual tax revenue itself, with state finance shifting from a base in taxation to a base in credit. Other work sorts the whole Song paper-money mechanism under four headings — backing, redemption, war, and inflation: in the early period, with both backing and redemption held strict, the system ran well; later, once war finance broke the mechanism, actual issuance at its peak is estimated to have reached 35.4 times what a strict reserve standard would have allowed.

A sharper account still comes from Gao Congming, who locates the truly revolutionary moment not in private jiaozi but in the instant official jiaozi turned into state credit money. In his telling, the key fact is not how private merchants guaranteed each other, but how fiscal spending came increasingly to be settled in jiaozi: once the government began paying for army rations, for the capital behind the Green Sprouts loans, for the ever-normal granary's purchase fund, all in jiaozi, without holding a matching reserve of iron coin, the paper's credit no longer rested mainly on its being exchangeable for iron coin. It rested instead on the state's own promise that it would accept the note, honor it, and let it be used to pay taxes.

It is exactly for this reason that Gao regards official jiaozi as, in a sense, more modern than the private version it replaced.

5. No Capital, No Trust

The Southern Song pushed the experiment further, and the hesitation behind it shows more openly.

In the first year of Shaoxing, with troops garrisoned at Wuzhou and the roads impassable to boats, cash being too heavy to move, the government first issued guanzi — a certificate for moving military funds — to manage the problem. Complaints followed almost immediately: prefectures and counties were using guanzi as a form of grain-purchase capital, effectively as a forced levy, and the monopoly offices in charge of repayment were making good on only a third of it each day, leaving everyone aggrieved.

In the sixth year of Shaoxing, the court briefly established a Jiaozi Bureau at the temporary capital, and it was immediately met with a criticism that stands as the single most important sentence in this entire history. The court, it ran, had set out to manage cash guanzi; the officials in charge had gradually lost sight of the original intent and turned the scheme into jiaozi instead. With the government holding no capital, how could the people be expected to trust it?

Rendered plainly: you, the government, have no reserve in hand — why should ordinary people believe in this paper?

The Jiaozi Bureau was abolished soon after, and the monopoly offices were ordered instead to hold cash reserves against which to print guanzi.

What that sentence exposes runs deeper than it looks. It says that even as paper money grew ever more dependent on the authority of the state, the question of reserves — of whether the thing could actually be redeemed — never left the conversation. The anonymous medium of exchange was expanding, certainly, but the question lurking behind it, of who answers for the paper and with what, had not retreated a single step. It had already surfaced once, on private issuers; now it surfaced again, on the state itself, at a far larger scale.

Put another way: the issuer had grown from sixteen wealthy households into an entire dynasty, larger by some incalculable factor, and yet the underlying question had not changed by one word. Who stands behind the note, with what do they stand behind it, and will they spend what stands behind it before the bill comes due? A larger issuer is, if anything, more dangerous in one respect: when a wealthy household goes bankrupt, there is still a court to sue, still property that can be seized. When a dynasty declines to make good on the account, the note-holder has nowhere left to take his complaint at all.

The mature form of this instrument, huizi, dates only from after the thirtieth year of Shaoxing. The court ordered Qian Duanli to produce huizi and to hold cash reserves against it, circulating both inside and outside the capital; the following year the Huizi Bureau was placed under the Tea Market Administration; and in the thirty-second year of Shaoxing a law against forgery was fixed in place.

Huizi at first circulated only in Liangzhe, later expanding to Huainan, Zhejiang, Hubei, and Jingxi. In many regions tribute payments to the court could be made entirely in huizi; garrison prefectures along the river paid half in cash and half in huizi; and private sales of land, houses, horses, cattle, boats, and carts could likewise be settled entirely in huizi.

This step went deep. Huizi was no longer simply a fiscal instrument; it had entered the trade in major assets — land, livestock, transport. Value was being converted, more thoroughly than ever, onto a uniformly printed sheet of paper.

Under Emperor Xiaozong, institutionalization went further still. New denominations of two hundred, three hundred, and five hundred wen were added; in the fourth year of Qiandao, three years was fixed as the length of one issue, each issue capped at ten million guan, old notes exchanged for new on schedule; a damaged note could still be redeemed as long as the string-and-hundred figures on it remained legible and the seal could be verified; each exchange carried a wear-and-tear fee of twenty wen. In addition, the various tax levies collected by prefectures and counties were reset to seventy percent cash and thirty percent huizi.

Issue period, fixed quota, seal verification, handling fee, the ratio for mixed tax payment — all of it folded into one unified scale of graduated measurement. This was an effort to take something as vague as reputation and turn it into an object the ledgers could actually process.

The seventy-thirty rule is especially telling of the intent behind it. It amounted to the government stipulating, for every tax payment collected, exactly how much of its own paper it would accept in return. The move was shrewd: the real support behind paper money is not how much cash sits in the vault but whether the government itself is willing to take the paper back. As long as it counted at tax time, people had to keep holding it, and it remained money. But the same move was also a lever the government could pull against itself, since the acceptance ratio was set by the government alone — and when it wanted to accept less, changing thirty percent to twenty was simple enough.

And then the whole graduated scale tore down its own scaffolding.

In the third year of Chunxi, the third and fourth issues were each extended by three years; in the first year of Shaoxi, the seventh and eighth issues were extended again, by three years apiece. Officials objected immediately: huizi's issue was supposed to run three years, they said, and now, extended and extended again, it had become nine — on what basis could anyone still be shown that the promise meant something?

The objection lands with precision. The trust behind paper money was never, here, some abstract moral quality. It was this exact and concrete question: did the deadline agreed upon earlier still count for anything?

The issue period had existed to recall old notes and lock the money supply back into an institutional rhythm; once fiscal need forced it to be postponed again and again, that same measuring order tore itself apart with its own hands.

The consequences show up in the numbers. By the first year of Qingyuan, each issue's huizi quota had already risen to thirty million. By the second year of Jiading, the eleventh issue, even after recall, still had more than 13,600,000 guan outstanding; the twelfth issue still had more than 47,000,000 guan outstanding; the thirteenth still had more than 57,000,000. The court was reduced to drawing gold from the sealed treasury, issuing ordination certificates — the documents the government sold to monks, which could themselves be bought and sold — and even drawing on salt-tax revenue, all in an attempt to prop up the price of huizi.

The picture is worth sitting with. What had been set up was a tidy, orderly scale: three years to an issue, ten million to an issue, old exchanged for new on schedule. When the scale could no longer bear its own weight, the state reached for gold, for monks' ordination papers, for the proceeds of the salt monopoly, to prop the paper up from beneath.

Whatever had escaped the unified scale simply resurfaced somewhere else.

And each thing reached for as backing sat a little further from money than the last. Gold, at least, was money's own old trade. Salt revenue meant mortgaging a monopoly's future income. Ordination certificates were stranger still — an identity document the government issued to monks, originally having nothing whatsoever to do with commerce, pressed into service only because it could be sold for cash. When a construct cannot hold itself up, it reaches for whatever it can grab, and the more varied the things it grabs, the more plainly it shows that the original scale no longer works.

The debate around huizi runs more complicated than the one around jiaozi. The traditional account explains its rise as a "coin famine," a shortage of copper cash; other research pushes back, arguing that the total money supply was not necessarily insufficient at all, and that the real trouble looked more like an imbalance in regional distribution, a tax system requiring cash and huizi to be paid together, and the rigid demands of financing war. Other scholars hold that huizi's biggest difference from jiaozi was weaker redeemability — closer, in practice, to a non-convertible paper currency — and that it kept functioning mainly because tax offices accepted it, the government bought it back, and the issue system kept redeeming it on schedule.

So the move from jiaozi to huizi was not a one-way march toward greater modernity for paper money. The more accurate description is that the share of the system given over to anonymity, standardization, and fiscal purpose kept rising, while the old problems — reserves, redemption at the issue date, tax acceptance, official malpractice, market discounting — stayed tangled up in it the whole time. Paper money did not escape personal relationships altogether. It escaped only the personal character of one particular firm, one particular circle of acquaintance — and was immediately tied back to a larger personal character still: whether the state itself would make good on the account.

6. Marco Polo's Alchemy

The Yuan dynasty pushed the experiment to a more radical extreme still.

In the first year of Zhongtong — 1260 — Kublai Khan began issuing jiaochao, backed nominally by silk; the same year saw the Zhongtong Yuanbao banknote, in denominations running from ten wen to two guan across ten grades, each guan equal in value to one liang of the silk-backed jiaochao, and two guan equal to one liang of silver.

In other words, the new Yuan note was anchored to silver from the very start: two guan of paper to one liang of silver. By 1271 the Yuan went further, declaring the Zhongtong note the sole legal currency in the empire, with other metal currencies nominally abolished.

This step differed sharply from the Song experiment. The state was no longer merely managing paper alongside metal; it was trying to make paper occupy the very center of the system.

What Marco Polo witnessed was precisely the most dazzling face of that system. He wrote that Kublai Khan struck currency from the bark of the mulberry tree, stamped with officials' signatures and a red seal, and that with this paper money people anywhere in the empire could buy whatever they wanted. He marveled, too, at how light the paper was — a note with a face value of ten gold coins, he noted, weighed less than a single gold coin — and observed that merchants who sold gold, silver, and pearls to the emperor were then paid back in paper.

To a thirteenth-century Venetian, this naturally looked like a species of alchemy: a sheet of paper, and someone else's gold changes hands.

There is a layer to his astonishment that he himself could probably not have named. In his own homeland, money was gold and silver, and gold and silver were valuable in themselves; here, money was a sheet of paper, valuable because a power almost beyond imagining said it was valuable, and because that power could genuinely make the claim stick for everyone from Dadu to the south of the Yangtze. What he was really watching, then, was less a currency than a demonstration of the reach of government: the fact that a piece of paper could be made to count across thousands of li was itself a display of state power.

More recent scholarship, though, warns against letting Marco Polo's wonder stand in for institutional history itself. Some researchers argue explicitly that Western impressions of Chinese paper money have too often been shaped by Marco Polo's excessively favorable account, obscuring the fact that premodern China's experiments with non-redeemable paper currency also produced plenty of failure, collapse, and retreat. What Marco Polo saw was the most spectacular, most efficient face of empire at its center. The fiscal overreach, the regional discounting, the quiet return of silver, and the eventual collapse at the end of the Yuan — none of that appears anywhere in his exclamations.

And silver, in fact, never really left.

The Yuan banned silver from circulating in private transactions, yet large transactions — especially in real estate — continued to be settled overwhelmingly in silver. The official exchange rate makes the trend plain enough on its own: in 1260 it still stood at two guan to one liang of silver; by 1309 it had moved to twenty-five guan to one liang. On this basis, the average annual depreciation over the period works out to roughly five percent.

This is worth stopping over.

Officially there was only one legal currency, and everything else had been abolished. In practice, when people conducted the transactions that actually mattered, they kept a second scale in mind, and that scale was silver. The state could declare that there was only one scale in existence. It could not make the other one disappear from people's minds.

Notice, too, the occasion on which people reached for the second scale: buying land. For the small daily transaction, paper would do, since it would be spent again soon enough; but for a major holding meant to pass down to one's children, a person had to think further ahead, had to ask whether the thing would still be worth anything in ten or twenty years. The more durably a person wanted to store value, the less willing they were to entrust it to paper. The distinction is telling: paper notes could serve as a medium of exchange, but made a poor container for storing value — and storing value is exactly the thing people care about most.

Some scholarship divides the history of Yuan paper money into three phases: 1260 to 1276, a period of full convertibility into silver; 1277 to 1309, a period of merely nominal convertibility; and 1310 to the end of the dynasty, a period of legally mandated non-convertible paper. In the first phase, excess notes could be recalled with silver; across the latter two, that anchor grew steadily weaker, until it finally became a purely legal fiction backing pure paper.

Paper money did not break free of metal in a single step. It went through a slow retreat — from a silver standard, to a weak silver standard, to fiat paper pure and simple.

7. The Price of Everything, Printed in Advance

The Ming took paper money to another extreme altogether.

In the eighth year of Hongwu — 1375 — the Ming issued the Da Ming Baochao, the Great Ming Circulating Treasure Note. Surviving specimens tell us how large it was: roughly thirty-four centimeters long and just over twenty centimeters wide, printed on paper made from mulberry bark. Across the top, in large characters, ran the words Da Ming Tongxing Baochao; the middle carried the denomination — one guan, say — alongside an image of ten strings of cash; and below ran a long block of legal text whose substance was this: printed with the court's authorization, to circulate alongside copper coin; counterfeiters to be beheaded; informants to be rewarded two hundred and fifty liang of silver, with the criminal's property confiscated.

According to the Collected Statutes of the Ming, one guan of this paper converted to one thousand wen of cash, or to one liang of silver; four guan equaled one liang of gold.

This amounts to printing an entire conversion table directly onto the money.

Printing the conversion table on the note is a gesture heavy with meaning. It announces that this paper is not merely money but the master rule for converting everything into everything else: one guan is a thousand wen is a liang of silver, four guan is a liang of gold, all written right there, no need to consult the market at all. This is, on the long road toward commensuration, one of the rare attempts to print the answer directly onto the instrument. But market prices have never listened to what is printed on paper, and the more rigidly a printed price is fixed, the more conspicuous the gap between it and the real one becomes.

The Ming's early ambition was more direct than either the Song's or the Yuan's. Once the baochao system had taken its final shape, transactions in gold and silver were banned outright, along with barter; copper coin was withdrawn from circulation and at one point banned as well, all in pursuit of a single goal — making the baochao the sole legitimate currency.

This was a more thorough attempt at commensuration than any that came before it. The state tried to fold gold, silver, copper coin, and the whole range of commodity exchange into a single printed note and a single body of law. Paper money here was not merely a medium for the market; it was the state's instrument for compressing every means of payment into one pricing frame.

And it is exactly here that the Ming's problem exploded into view.

One line of scholarship holds that Ming fiscal administration was simply mismanaged — that, unlike the Song, it failed to use paper money effectively for tax collection or to recall it from circulation, and that the baochao depreciated severely as a result; by the early fifteenth century Ming notes had largely stopped functioning, after which the Chinese market shifted to a silver-ingot standard, and Ming fiscal accounting and taxation gradually moved onto a silver basis as well. Other scholarship treats the Ming as a stark case demonstrating that premodern China's experiment with non-convertible paper was, in the end, no more successful than early modern Europe's; by around 1430 the state had already been forced to abandon paper money as the effective core of circulation.

Taken apart, there are at least three parallel explanations for why Ming paper failed.

One stresses a change in fiscal constitution: what had supported paper money under the Song and Yuan was a highly monetized fiscal structure resting on commercial and indirect taxes; the early Ming, by contrast, relied far more on payment in kind, on corvée labor, on forced mobilization. The state issued notes eagerly but recalled and absorbed them through taxation only weakly, so paper flowed out with almost nothing flowing back in.

A second stresses a design flaw built in from the start: the baochao system had been designed from the outset to relieve fiscal pressure, so that money itself was fiscalized, and overissue was practically built into the design.

A third, more commonsensical explanation is not entirely wrong either: forcibly banning silver, banning copper, banning barter cannot cancel out a market's preference for a more reliable store of value — the ban only widens, and widens faster, the gap between the nominal legal currency and what people actually settled their accounts in.

Each of the three explanations has its own logic, and together they come into sharper focus: overissuing money is one problem, failing to recall it is a second, and banning the alternatives too harshly is a third. Issuance, recall, prohibition — all three went wrong, and any one alone might not have been fatal, but together they proved impossible to sustain. Assigning a single culprit is probably not possible. What can be said with confidence is that the system demanded of itself far more than it was ever capable of delivering.

The Ming collapse tells, once more, the same underlying story.

Stripping value away from metal does not mean it thereby escapes reputation. Quite the opposite: once stripped, value presses down more concentratedly on the issuer's reputation, on fiscal capacity, on the credibility of enforcement. The moment a state can no longer sustain a closed loop for its paper across tax payment, redemption, government purchase, and government disbursement, the market moves value back onto silver, onto copper coin, onto whatever else can still be trusted.

The Ming's later turn to a silver standard was not some more advanced, natural victory either. It was a society retreating, after one great paper-money experiment had failed, into another system of pricing and storing value that could better protect itself in practice.

8. What Was Left After the Metal

Taken across the full millennium, the best summary is not to be found in any later theory, but in three sentences that people actually said at the time.

The first belongs to Lü Zhong: having money to put paper into motion, and paper to regulate money, mother and child circulating in balance. Paper and coin, in his account, were never a replacement of one by the other. They were a pairing.

The second is that question from the sixth year of Shaoxing: with the government holding no capital, how could the people be expected to trust it? Without capital, where is trust supposed to come from?

The third runs in the opposite direction, and is no less true: if the government accepts its own paper without difficulty, the people's confidence will naturally follow. If the state is willing to take the paper, the people will be willing to believe in it.

Put the three together and the whole difficulty of the matter is laid out in full. For paper money to stand, one end needs capital and the other needs someone willing to accept it — and of those two ends, one is a thing, and the other is trust.

What is interesting is that there is real tension among the three. The second says that without capital there is no trust; the third says that as long as the government accepts the paper, people will trust it regardless — one plants its root in the vault, the other in the attitude of the magistrate's office. The first sentence sits between them, insisting the two must be held in balance, with neither side allowed to dominate. The Song argued this question for two hundred years, and the three positions that argument produced turn out to be exactly the three positions every later monetary theory would eventually have to reckon with.

Followed all the way through, what actually happened here can be stated quite plainly.

In the earlier essay on double-entry bookkeeping, the scale, for the sake of convenience in record-keeping, actively shed its own body and became a pure, non-circulating number. With jiaozi, it is the circulating currency itself that sheds its metal.

And once the metal is shed, what happens is not that money comes to need people less. It comes to need people more. A copper coin, whoever is holding it, remains a lump of copper — half its value resting in the object itself. A paper note is nothing like that: its value rests one hundred percent on other people — on whether the issuer honors it, whether the government accepts it, whether it will still be redeemable next year.

So the whole business of paper money runs in two directions at once. In circulation it becomes more anonymous — anyone at all can hold a sheet of paper, no questions asked about where it came from. At its foundation, though, it becomes more personal than ever, because beyond that one promise, it has nothing else to stand on.

However much metal is withdrawn, exactly that much weight comes to rest on reputation. One side grows lighter; the other grows heavier.

So the claim that paper money marks the moment personal character exits monetary history has the direction backward. It is closer to the truth to say that personal character was, for the first time, pushed onto the main stage: in the past, the trust placed in an issuer always had metal to share half the weight; from paper money onward, it has to carry the whole weight by itself. And whether it can bear that weight no longer depends on how much ore lies buried in the ground. It depends on whether a regime, in the very years it is most desperate for money, is willing to keep a promise it made years before.

As for the remainder, here it shows a new face.

In the earlier essays, whatever spilled out past the edge of the ledger was either something that could not be measured, or something that should not have been measured, or something measured that ended up touching nothing at all. This time is different. This time what spills out is another scale entirely.

The Yuan declared paper the sole legal currency and abolished everything else, and yet people conducting the transactions that truly mattered — buying land — kept reckoning in silver all the while. The Ming banned gold, banned silver, banned barter, printed its conversion table directly onto the face of the note — and the outcome was that silver, by the fifteenth century, had become once again the true measuring instrument of the entire society.

A construct can declare itself the sole scale. It cannot make the backup scale in people's hearts disappear.

And that backup scale stays silent in ordinary times, surfacing only once the construct itself begins to lose credibility. It does not arrive from outside to topple the construct. It was there the whole time, and it is the construct's own broken faith that calls it back onto the stage.

This is also why such prohibitions always did so little good. They all attempted the same thing: stopping people from using another scale. But people do not turn to another scale because an alternative happens to exist. They turn to it because the scale already in their hands has started reading false. A ban can take away the substitute; it cannot take away the reason people go looking for one — and as long as that reason remains, some substitute will eventually come back, in whatever form it can find.

As for who ends up paying, it is written plainly enough in the sections above. When private jiaozi suffered a run, a guan redeemed for only seven or eight hundred wen, and the loss fell on whoever was holding the note. When jiaozi exchanged four-to-one for new issue, the loss fell on whoever was still holding the old. When huizi's issue was extended again and again, the loss fell on those who had believed in the three-year promise. When Ming notes depreciated, the loss fell on those who had honestly accepted them in the first place.

The issuer can spend someone else's trust in advance. Whoever bears the consequence is always whoever happens to be holding the paper last.

That paper never once carried a column for this.

A jiaozi note carried its face value, its issue number, its code number, the penalty prescribed for forgers, the fee deducted from every guan. Everything was written down — except for one column: if the note cannot be redeemed on schedule, who bears the loss. That column is missing not because anyone forgot it, but because it could not be written. To write it would be to admit that the paper might not be honored, and a paper that admits it might not be honored stops being money the same day. So the one truly essential fact of the whole arrangement had to stay off the page entirely, left to whoever happened to be holding it last.

The ledger has not yet balanced. It is still being kept.