An emperor has imposed reforms on a country that did not ask for them. The country is unsettled and wants order, and the order it wants is the one Confucius represents.
Confucius has been dead for five hundred years and cannot be consulted.
So the emperor consults him anyway. Greene's account is four sentences and it is the most naked passage in Law 45:
"He first confiscated them, then had his scribes insert passages into them that seemed to support the changes he had been imposing on the country. When he released the texts, it seemed that Confucius sanctioned Wang's reforms, and the people felt comforted and accepted them more easily."1
And the gloss Greene draws from it: *"The past is a text in which you can safely insert your own lines."*2
The three-step sequence is the technique, and each step is doing something the others cannot.
Confiscate. You must control the physical object before you can alter it, and confiscation is the only move that produces exclusive access while looking like custody rather than seizure. Wang took the texts as an emperor protecting a treasure.
Interpolate. The insertions "seemed to support the changes" — not to command them, not to prophesy them. Support is the right calibration: a text that turned out to contain explicit endorsement of a specific contemporary policy would be obviously wrong. General sanction is both more credible and more useful.
Release. This is the step that does the persuading, and it is the one most easily missed. The reappearance of the texts is not experienced as an emperor publishing his own justification. It is experienced as a restoration — a lost inheritance returned to the people by a ruler who preserved it.
So the forgery arrives wrapped in an act of generosity, and the generosity is what makes the content credible. Nobody examines a gift for insertions.
The vulnerability Wang exploited is stated in one clause and it is the precondition for everything else.
Two centuries earlier Emperor Ch'in had the writings of Confucius burned. Then: "A few years later, word had spread that certain texts had miraculously survived, hidden under the scholar's house."1
A text with a miraculous recovery narrative has one distinctive property: its provenance cannot be checked. There is no chain of custody, no comparison copy, no institution that held it. Its authority rests entirely on the story of its survival.
And that single property does both jobs at once. It is what made the texts believable — a document that survived Ch'in's fire against the odds is precious, and its very improbability is the source of its aura. It is also what made them alterable, because a document with no verifiable history cannot be shown to have been changed. There is nothing to compare the new version to.
Which yields a general rule worth carrying: the documents most venerated for having narrowly survived are the documents most open to being edited, and for exactly the same reason. Veneration and vulnerability come from the same absence.
Greene supplies the detail that makes this the strangest case in the chapter, and then walks past it:
"These texts may not have been genuine, but they gave Wang his opportunity."1
So the sequence is: Confucius's writings are burned. Texts appear, claiming miraculous survival, which were probably not his. Wang confiscates those probably-false texts and inserts further false passages into them. The people are comforted by the result and accept his reforms.
Two strata of forgery, the second built on the first, and the reassurance the population felt was produced entirely by material Confucius never wrote.
The consequences are worth stating carefully, because the obvious reading is too cynical and the interesting one is not.
The obvious reading is that the whole thing was fake and the people were fools. But the values in the recovered texts were presumably recognisably Confucian — that is why the forgery worked and why nobody rejected them on sight. Whoever produced the first layer was reconstructing a tradition they knew, and the tradition was real even if the document was not.
Which means Wang's insertions were not smuggled into an authentic text. They were smuggled into a reconstruction, and a reconstruction has no fixed edge. There is no original to violate, so there is no boundary at which addition becomes falsification.
That is the uncomfortable finding: a text assembled from memory and reverence is not merely easier to forge than an authentic one — it is unclear what forging it would even mean. Wang did in bad faith what the first forgers may have done in good faith, and the two operations are indistinguishable in the object they produced.
One phrase changes what the technique is for: the insertions supported "the changes he had been imposing on the country."1
Past continuous. The reforms were already in place. Wang was not seeking authorisation to act; he was seeking acceptance of what he had already done.
That makes this a pacification instrument rather than a permission instrument, and the two have different requirements. Permission must arrive before the act and must be specific enough to cover it. Pacification arrives afterwards and needs only to remove the sense that something illegitimate has happened.
Greene's description of the result is precise about this: the people "felt comforted and accepted them more easily."1 Not convinced. Not persuaded that the reforms were correct. Comforted — the reforms stopped feeling like a rupture.
Which explains why a vague general sanction was sufficient. A population that has already absorbed a change is not evaluating a proposal; it is looking for evidence that the change belongs to the same world as everything else. Any credible continuity will do, and it does not have to be argued, only exhibited.
Law 45 contains, in three separate passages, a complete two-thousand-year sequence about the same texts, and never assembles it.
Emperor Ch'in burns the works of Confucius.3 Texts "miraculously" reappear. Wang Mang confiscates and interpolates them to sanction his reforms.1 Nineteen centuries later, Mao rehabilitates Ch'in — specifically crediting the burning among his achievements — and then runs a national campaign against Confucius to destroy a rival.3 4
Four operations on one philosophical inheritance: destruction, reconstruction, forgery, and finally the rehabilitation of the destroyer.
The chapter uses these as three unconnected illustrations. Assembled, they say something the chapter does not: the Confucian corpus functions less as a fixed body of teaching than as a durable political instrument that successive rulers modify, and its authority survives every modification. Wang inserted lines and it remained authoritative. Ch'in burned it and it came back. Mao attacked it and it is still there.
That durability is the real subject, and it is invisible from inside any one of the three cases.
Greene's gloss is worth examining word by word, because one adverb carries the argument:
"The fact that the past is dead and buried gives you the freedom to reinterpret it. To support your cause, tinker with the facts. The past is a text in which you can safely insert your own lines."2
Safely is the claim. Not effectively, not credibly — safely. The argument is about risk, and the risk is low because the injured party is dead.
That is true as far as it goes, and it is the same reasoning that makes historical arguments cheaper than present ones throughout this chapter. Confucius cannot issue a correction.
But the chapter's own material shows where the safety ends, and it is not with the dead. It is with the living who care about the dead. Nothing happened to Wang from Confucius; the danger to a ruler who tampers with a revered text comes from scholars, from rivals with a use for the accusation, and from the fact that the tampering, once suspected, retroactively contaminates everything else he claimed.
Greene's own preferred technique elsewhere in the chapter avoids this entirely. Mao's Water Margin borrowing asserted nothing false and could not be exposed. Wang's version is the same operation with an exposure attached — and the chapter recommends it in the more dangerous form, on the grounds that the person you are lying about cannot object.
You have already made the change, and the room has absorbed it without accepting it. Nobody is fighting you. Something is unsettled.
The instinct is to justify — an explanation, a memo, the reasoning laid out. Watch what that does. A justification issued after the fact concedes that a justification was required, and it puts your reasoning in the room as a thing to be argued with. You have reopened a settled matter and taken the weaker side of it.
So look for continuity instead of argument. Find where this change already agrees with something the group considers its own — a principle in the founding documents, a decision made three years ago, something in the way this place has always described itself.
If it is genuinely there, exhibit it. Not as a defence. As a passing observation: this is the same thing we did in 2021. That is the whole intervention, and its power is that it is not addressed to anyone's objection.
Now the line, and Wang Mang is on the wrong side of it. If it is not genuinely there, do not insert it.
The temptation will be small and specific — restating an old decision slightly more generally than it was made, describing a past principle in words nobody used at the time, letting a document be summarised in a way it does not quite support. Each is minor and each is the same operation Wang ran: adding a line to a text and releasing it as though it had always been there.
The reason to refuse is not only integrity. It is that the exposure is asymmetric and delayed. The insertion buys a small amount of comfort now; discovery costs you every other continuity claim you have ever made, at a moment you will not choose. Wang's technique worked precisely because a population absorbing a change wants very little — any credible continuity will do. If very little is enough, it is not worth manufacturing.
Strongest evidence is the concession that the texts "may not have been genuine."1 Greene had no argumentative reason to include it and it complicates his own example, which is why it can be relied on and why it is the most interesting fact in the passage.
Tension one — the chapter's safely is contradicted by its own better technique. Sourcing (Mao and The Water Margin) asserts nothing false and has no exposure. Interpolation asserts something false about a revered text and has a permanent one. The chapter recommends the second without weighing them against each other.
Tension two — a pacification instrument is offered as though it were a permission instrument. The reforms were already imposed. Nothing here shows the technique enabling a change; it shows it settling one.
🚩 SECONDARY WITHOUT PRIMARY / SINGLE SOURCE. Wang Mang's reign is dated (A.D. 8–23) and nothing else is sourced: no historian, no named text, no scribe, no specific interpolated passage, and no evidence for the claim that the interpolation happened at all. This is a serious allegation of textual forgery against a named historical figure, asserted flatly. [LOW CONFIDENCE]
🚩 The claim that the recovered texts "may not have been genuine"1 is itself unsourced, and it is doing heavy work on this page. Treat the two-layer reading as following from Greene's account rather than as established.
Open question — did the forgery pacify anyone, or did the reforms simply bed in? The chapter asserts that people "felt comforted." A population accommodating to an imposed change over time is the null hypothesis and is not considered.
Open question — what is the modern equivalent of a text with no chain of custody? Wang needed a document whose provenance could not be checked. Whether that vulnerability has narrowed or widened is not something this case can settle.
Within Law 45 this is the technique in its most literal form, and the chapter's three other historical operations sort around it by how much has to be untrue.
Mao and The Water Margin asserts nothing false — the novel really does exalt the band above the family. The Ch'in rehabilitation advances a contested claim about a real reputation. Wang Mang manufactures the evidence. And Cromwell runs the whole thing in reverse, circulating stories to discredit the past rather than to borrow from it.
Four positions on one axis, in one chapter, unsorted — and the chapter's own instruction, tinker with the facts, sits at the dishonest end while its best example sits at the honest one.
The direct link is to The Anti-Confucius Campaign, because it closes the loop on the same corpus. Ch'in burns it, Wang forges into it, Mao rehabilitates the burner and then attacks the text again. Greene tells all three and connects none.
Buddhist Canon: "Authentic" Means Correctly Attributed, Not True — the distinction between a librarian's authentic (the name on the spine is the person who wrote it) and a practitioner's (the teaching does what a teaching is supposed to do), and a canon tradition that has had to decide which one it means.
Wang Mang's operation lands squarely on that distinction and shows why it is not merely a scholarly nicety.
In the librarian's sense, the interpolated passages are straightforwardly fraudulent: Confucius did not write them. In the practitioner's sense the question is different and much harder — do they teach what Confucius taught? If Wang's scribes inserted material consistent with the tradition, then by a pragmatic standard the text is no less authentic than before, and the population's comfort was not misplaced.
And the case will not resolve cleanly, because the texts were already a reconstruction. There is no attributional baseline to appeal to. The librarian's test cannot be run at all.
What the pairing produces: traditions that adopt the pragmatic standard are structurally unable to detect this attack, and traditions that adopt the attributional one are structurally unable to survive their own transmission history. The first has no test for insertion; the second discovers that most of what it reveres fails the test. Wang Mang exploits the first horn. Greene's "tinker with the facts" is advice that only works inside a tradition that has quietly chosen it.
Forgery Has No Meaning (Dukes's Counter-Philological Doctrine) — the position that the philologist's verdict of forgery, reached by spotting anachronisms and inconsistencies, misunderstands what transmitted texts are and what they are for.
Read against Wang Mang, this doctrine looks alarming — and the alarm is the useful part.
The counter-philological argument has real force in its own setting. Living traditions accrete; texts are updated by practitioners who consider themselves transmitters rather than authors; and a scholar circling an anachronism and writing forgery is applying a standard the tradition never adopted. Much of what looks like fabrication is ordinary transmission.
Wang Mang is the case that shows the doctrine cannot be load-bearing on its own. His scribes were doing, formally, exactly what a transmitting tradition does: adding material to a text in a way consistent with its idiom. The formal operation is identical. What differs is intent and beneficiary — the insertions served a living emperor's political programme.
And intent is precisely what philology cannot see and what the counter-doctrine declines to weigh. So the two positions have complementary blind spots: the philologist can detect the insertion and cannot tell whether it was transmission or fraud; the counter-philologist can explain why insertion is normal and has no way to flag the case where it was an attack.
Which suggests the missing test is neither textual nor doctrinal but circumstantial — who benefits, and were they holding the manuscript? Wang confiscated the texts before releasing them. That fact is not in the document and no amount of reading will recover it.
Sharpest implication. The property that made the recovered texts venerable — a miraculous survival, no chain of custody, nothing to compare them to — is the identical property that made them alterable. Veneration and vulnerability came from the same absence, and they cannot be separated, because a document's aura of narrow escape is the gap in its provenance. Which means the texts a community holds most sacred on account of how nearly they were lost are structurally the ones most open to having lines added, and the community has no instrument for noticing.
Generative questions.
If a reconstructed text has no fixed edge, at what point does addition become falsification — and is that a question about the document or only about the adder's intent?
Wang's insertions pacified a population that had already absorbed the change. How much of what looks like persuasion is retroactive comfort, and would we recognise the difference from inside?
The missing test is circumstantial — who held the manuscript. What are the modern equivalents of confiscation, where someone acquires exclusive custody of a record before it is released?