A party at a New York tycoon's house. The dealer Joseph Duveen has recently sold the man a Dürer for a large sum, and the painting is not yet hung.
The tycoon's daughter, wanting to impress a confident young French art critic among the guests, shows it to him.
He studies it for a while. Then: "You know, I don't think this Dürer is right."1
He follows her as she hurries to her father. The magnate, badly unsettled, turns to Duveen.
Duveen laughs.
"How very amusing," he said. "Do you realize, young man, that at least twenty other art experts here and in Europe have been taken in too, and have said that painting isn't genuine? And now you've made the same mistake."1
The critic apologised.
Look at what that sentence actually does, because it is a small machine.
It does not argue. No provenance, no brushwork, no documentation, no counter-evidence of any kind is offered.
It supplies a number. At least twenty. Specific enough to sound like a record, vague enough to be unverifiable.
It relocates the critic. He is no longer a man making a judgement; he is the twenty-first entry on a list of people who made a known error.
And it inverts the direction of expertise. The critic came in holding the sophisticated position — spotting a fake is the connoisseur's move. Duveen makes doubt the naive position, the thing beginners do.
Nothing in the text suggests the twenty experts existed.
Greene gives no source for them, and Duveen produces the figure instantly, in a social setting, under pressure, to a stranger. The number is doing rhetorical work, not evidentiary work.
Which makes this a specific and reusable move: invent a prior consensus that has already dismissed the objection being raised. The objection is not answered; it is dated. It becomes a mistake other people already finished making.
And there is no way to check it in the room. Asking which twenty? would itself be a further display of the naivety just diagnosed.
Here is the part that makes the anecdote worth a page rather than a paragraph. Greene follows it immediately with this:
"Duveen knew that the art market was flooded with fakes, and that many paintings had been falsely ascribed to old masters. He tried his best to distinguish the real from the fake, but in his zeal to sell he often overplayed a work's authenticity. What mattered to him was that the buyer believed he had bought a Dürer."2
Read it carefully.
The market is "flooded with fakes." Duveen "often overplayed" authenticity. And what mattered was not that the painting was a Dürer but that the buyer believed it was.
So Greene has told us: a knowledgeable young expert made a judgement that was probably correct, and was talked out of it by a man who knew the market was full of forgeries and whose incentive ran entirely one way.
And it is filed as a technique to admire.
Trace the incidence, because three parties are affected and only one is discussed.
The critic loses immediately and publicly. He apologises for being right — and, worse, learns a lesson about the cost of voicing an accurate judgement in a room where a bigger reputation is exposed.
The tycoon is the one with money at stake, and he is the person the reassurance is aimed at. He keeps the painting and the good feeling, and never finds out.
Duveen keeps the sale, the relationship, and his standing as the man whose attributions are not questioned.
Greene's closing line is the operating instruction: "it is important to be able to play the professor when necessary and never impose such an attitude for its own sake."2
When necessary — necessary for what? In this case, necessary to prevent a customer discovering he had been sold a fake.
The chapter names the skill and never names the transfer. The person who bore the cost of Duveen's performed authority was the man who had already paid for the painting.
Law 21 says seem dumber than your mark. This is the Reversal, and Greene is explicit that it inverts:
"There is, however, one situation where it pays to do the opposite—when you can cover up a deception with a show of intelligence. In matters of smarts as in most things, appearances are what count. If you seem to have authority and knowledge, people will believe what you say."3
So the same goal — successful deception — is served by opposite means, and the stated criterion is when you can cover up a deception with a show of intelligence, which reduces to: use whichever one works.
That is not a criterion. It is a description of the outcome, offered as the rule for choosing.
Someone junior has just questioned something you sold, built, or signed off — and they might be right.
Notice the sequence in your body, because it arrives in this order: first the flush, then the impulse to establish that you have been doing this for twenty years.
That impulse is Duveen's move, and it is available to you whether or not you deserve it. Performed authority works regardless of whether the authority is real. That is precisely why it is unsafe.
So run the check before you answer. What would it cost me to be wrong here, and who would pay it? If the answer is that someone else eats the loss while you keep the standing, you are in the exact position Duveen was in, and the reassuring sentence is already forming.
Then answer the question rather than the questioner. What are you seeing? — and mean it, because a specific objection is either answerable or it isn't, and finding out which is the entire job.
And if you catch yourself reaching for a number — twenty experts, everyone in the field, we've looked at this a hundred times — stop. An unverifiable count is the signature of this move, and you will recognise it in others the moment you have caught it in yourself.
The strongest thing here is Greene's own concession, which converts an anecdote about confidence into evidence about the art market: "flooded with fakes", "often overplayed a work's authenticity", and the buyer's belief as the operative goal.
Tension: the case is presented as a skill and reads as a documented harm. A correct expert judgement is suppressed, and the person carrying the financial exposure is deliberately kept from it.
Tension: the Reversal supplies no criterion. Seem dumber and seem authoritative are opposite instructions for the same objective, separated only by "when you can."
Tension: this contradicts the chapter's core claim about vanity. Law 21 says never impugn a person's intelligence — "an unforgivable sin." Duveen impugns the critic's intelligence directly, in company, and it works perfectly. The chapter's iron rule is broken by its own Reversal without comment.
🚩 SINGLE SOURCE · 🚩 SECONDARY WITHOUT PRIMARY — no date, no name for the critic or the tycoon, no source for the exchange, and the twenty experts are uncorroborated within the text itself. [POPULAR SOURCE]
Open questions. Was the painting genuine? Greene's framing implies strongly that it may not have been and never says. And: the young critic apologised for an accurate reading. How much expert disagreement disappears this way — not refuted, but socially priced out of the room?
Against the law this reverses, the pair exposes what the chapter is really organised around. Both halves are about managing the perceived intelligence gap — Arnold and Slack widen it downward, Duveen widens it upward. The constant is not humility or authority but control of the comparison, and Greene never states that this is the actual variable.
Set beside Duveen and Henry Ford the corpus is building a consistent operator across five appearances. There, five dealers who hated each other pooled money to work a single buyer. Here, one dealer defends a sale by manufacturing a consensus. Both are about controlling what the market appears to believe — and Greene distributes them across separate laws as separate lessons.
And against the Claridge's elevator the pattern holds: Duveen's method is never connoisseurship. It is staging. The vault now has five Duveen episodes and not one of them turns on his judgement about a painting being correct.
Psychology — The Trial Where the Experts Couldn't Agree What a Person Is
That page sets four eminent authorities on coercion against two forensic psychiatrists at the Patricia Hearst trial — same defendant, same facts, opposite conclusions — decided by which model the jury found more believable rather than by which was truer.
Duveen's room is the same structure with the sample size cut to two.
What the pair produces: when qualified experts disagree, the resolution mechanism is not epistemic, and the deciding party is always a layperson. The Hearst jury could not evaluate competing theories of the self; the tycoon could not evaluate competing attributions of a Dürer. In both cases the verdict goes to the more confident performance, because confidence is the only variable a non-expert can actually assess.
Which means performed authority is not a distortion of expert dispute — it is the mechanism by which such disputes are routinely settled. And it locates the real defect in Duveen's room: not that he performed, but that only one side had any incentive to. The critic had nothing to gain from being right and everything to lose from being rude, so the dispute was resolved before it was heard.
Behavioral mechanics — The Four Fake Forms of Authority
That page inventories authority that is displayed rather than possessed — the forms that borrow the signals of standing without the substance underneath.
Duveen's "twenty other art experts" is a clean specimen: authority claimed by asserting an invisible consensus, in a setting where checking is socially impossible.
What the pair produces: the fake and the real forms are indistinguishable at the point of use, and the tell is structural rather than behavioural. You cannot detect the fake by watching the person — Duveen's laugh, timing and specificity are exactly what a genuinely confident expert would produce.
What you can check is whether the claim is constructed so that verification is possible. Real authority names a checkable thing: this provenance, that document, this comparison. Duveen names a number of unidentified people in unspecified places at unspecified times — a claim engineered to be unfalsifiable in the room where it is made. The signature is not overconfidence; it is unfalsifiability delivered confidently, and that is a test you can apply without knowing anything about Dürer.
Sharpest implication. Duveen does not rebut the critic's judgement — he redates it, converting a live objection into a mistake other people have already made and moved past.
An objection that has been placed in the past does not need answering, and the person raising it has to argue their way back into the present before they can even state it again.
Generative questions.