Picasso is working with the art dealer Paul Rosenberg, and giving him a fair amount of latitude in handling the paintings. Then one day, "for no apparent reason," he tells the man he will no longer give him any work to sell.1
No explanation. No grievance. Nothing.
Picasso's own account of what happened next is the best part, and it is worth having in his voice:
"Rosenberg would spend the next forty-eight hours trying to figure out why. Was I reserving things for some other dealer? I'd go on working and sleeping and Rosenberg would spend his time figuring. In two days he'd come back, nerves jangled, anxious, saying, 'After all, dear friend, you wouldn't turn me down if I offered you this much [naming a substantially higher figure] for those paintings rather than the price I've been accustomed to paying you, would you?'"2
I'd go on working and sleeping. That clause is the whole strategy and it is doing more work than the withdrawal is.
Two men spend the same forty-eight hours. One paints and sleeps. The other does not sleep.
Notice that Picasso is not performing anything during those two days. There is no campaign, no signalling, no follow-up. He made one statement and then went back to his life, and the entire mechanism ran inside somebody else's head at no ongoing cost to him.
That is the unusual property here, and it separates this from most of the manoeuvres in the corpus. Talleyrand has to be in the room. Duveen has to keep a payroll running. Deioces has to build a palace. Picasso's move has no maintenance cost at all — the target does the labour, and the labour is what produces the price.
Greene's framing of the general principle: "People are always trying to read the motives behind your actions and to use your predictability against you. Throw in a completely inexplicable move and you put them on the defensive."3
But defensive undersells what actually happened. Rosenberg was not defending. He was working — generating hypotheses, discarding them, and pricing each one. And the hypothesis he settled on was expensive.
The critical detail is that Picasso gave no reason, and this is not modesty about his motives. A stated reason would have ended the process in about ten minutes.
Suppose he had said your commission is too high. Rosenberg then has a defined problem with a defined solution, and can negotiate it — probably downward, since a named grievance is a bounded one.
Suppose instead he had said I have another dealer. Rosenberg can verify it, or call the bluff, or decide the other dealer is welcome to the work.
Say nothing, and Rosenberg has to supply the reason himself — and he will supply the one he most fears. He is not reasoning toward the truth; he is reasoning toward the worst case that fits the available facts, because the worst case is the one it would be expensive to be wrong about.
Which is why the offer that came back was substantially higher rather than marginally so. Rosenberg was not pricing the paintings. He was pricing the disappearance of Picasso, which is a different and much larger number.
Greene gives Picasso a general statement of method a few lines earlier, and it contains a condition most readings skip:
"The best calculation is the absence of calculation. Once you have attained a certain level of recognition, others generally figure that when you do something, it's for an intelligent reason. So it's really foolish to plot out your movements too carefully in advance. You're better off acting capriciously."4
Read the middle clause. Once you have attained a certain level of recognition.
That is the entire mechanism, stated by the practitioner, and it is a claim about the audience's prior rather than about the action. An inexplicable move by an established figure is read as a reason I cannot see. The identical move by an unestablished one is read as no reason — which is to say, as unreliability.
So capriciousness is not a technique that generates standing. It is a dividend paid on standing you already have, and it is paid in the form of other people volunteering to do your thinking for you.
This is the same precondition Law 16's Reversal states — leave too early and you are simply forgotten — arriving from a different chapter and a different practitioner. Greene has both and connects neither.
There is a second thing in the transaction worth separating, because it explains why the price moved so much.
Rosenberg's business was not these paintings. It was the relationship, which is an ongoing claim on future work by an artist whose output was appreciating. The paintings in front of him were an instalment.
So when supply stopped without explanation, the threatened loss was not a shipment. It was the annuity — and the annuity is worth vastly more than any single consignment, which is why he came back offering substantially above the accustomed price for those paintings. He overpaid for the instalment in order to protect the stream.
That is a structural feature of any relationship with recurring supply, and it means the leverage available to the supplier is not proportional to the current transaction. It is proportional to everything the other party expects to receive afterwards — a much larger number, and one the supplier can invoke without ever mentioning it.
Someone you supply has become comfortable.
Not hostile, not exploitative — comfortable. The terms were set a while ago, they have not been revisited, and there is no natural moment at which they would be.
The move here is smaller and stranger than a negotiation. You stop, and you do not explain.
Not dramatically. No announcement of grievance, no list of what has been unfair, no invitation to discuss. Those are all ways of handing over a defined problem, and a defined problem gets solved on the smallest terms that will close it.
Then — and this is the part that is genuinely hard — you go back to work. Picasso painted and slept. The mechanism does not need tending, and tending it is how you convert an unreadable act into a readable one. Every follow-up message you send is a clue that narrows what they have to imagine.
Two honest conditions before you try it.
Have you attained a certain level of recognition — in their eyes, in this relationship? If they have reason to assume your actions have reasons, the silence reads as strategy. If they do not, the same silence reads as flakiness, and the outcome is not a higher offer but a replacement supplier.
And can you actually absorb the pause? Picasso could stop selling to Rosenberg indefinitely. If the stoppage costs you more than it costs them, you are not withdrawing supply — you are starting a waiting contest you have already lost, and the person who blinks does so from a weaker position than the one they left.
The strongest evidence is that Picasso narrates it himself, in the first person, including his own reasoning and his own idleness during the two days. That is unusually direct for this book — most of Law 17's material is Greene reconstructing someone's interior.
Tension: Picasso's stated theory is not what Picasso did. "It's really foolish to plot out your movements too carefully in advance. You're better off acting capriciously." But the Rosenberg move was not capricious — he describes the target's reaction in precise, predicted detail, including the timeline. This is a calculated act described as an absence of calculation, and the description is itself part of the technique, since a man who is genuinely capricious cannot be negotiated with at all.
Tension: the chapter files this under terror. Law 17 is Keep Others in Suspended Terror, and the material around this case is about unnerving opponents and keeping them off-balance. Rosenberg was not terrorised; he was puzzled, and the puzzlement was commercial. Greene runs a pricing manoeuvre and an intimidation doctrine under one heading.
🚩 SINGLE SOURCE · 🚩 SECONDARY WITHOUT PRIMARY — the Picasso quotations are given without source, date, or interlocutor. The quoted passage has the texture of a recorded conversation, but no interview, memoir or biography is named, and the "substantially higher figure" is Greene's bracketed paraphrase rather than a reported number. [POPULAR SOURCE] · [PLAUSIBLE — needs corroboration]
Open questions. How many times could this be run on the same person before Rosenberg learned the pattern and simply waited — and does the technique have a use limit that Greene's single-instance telling conceals? And: Picasso is describing a manoeuvre he performed on a business partner, to a listener, apparently with some relish. What does it do to the relationship's other transactions once the story is in circulation?
Against Picasso and the Refusal of Predictability — the same artist at Law 6, and a genuinely different operation — the pair separates two things Greene treats as one temperament.
At Law 6, Picasso disrupts his own work: if a style becomes too attached to his name he produces something ugly and disturbing rather than let viewers grow familiar. The cost is borne by him, in the work itself, and the target is the public's model of the artist.
Here he disrupts a commercial relationship, at no cost to himself, and the target is one man's model of his own security.
What the pair produces: the same man is running unpredictability as an artistic discipline and as a pricing lever, and only one of them costs him anything. Which suggests the two should not be read as a single character trait at all. The Law 6 version requires genuinely sacrificing an established position; the Law 17 version requires only silence. Greene presents both as evidence of the same restless intelligence, and the second is considerably cheaper than the first.
Within Law 17, this case sits oddly beside Fischer at Reykjavik. Fischer's unpredictability cost him a forfeited game and very nearly the match; Picasso's cost him two days of ordinary work. The chapter's flagship case is expensive and its cleanest case is free, and the variable is whether the operator has anything at stake in the interval.
Business — Customer-Says-It Truth Asymmetry
That page's principle: when the operator says it, the prospect discounts it; when the prospect says it, it is true. So you engineer the conversation until the conclusion comes out of their mouth — ask what a fifteen percent cost cut would mean, and let them say that would save us money.
Rosenberg named the higher figure himself. Picasso never asked for a raise.
The insight the pair produces is about who bears the burden of the number. In the sales version, the operator still runs the conversation — asks the questions, shapes the path, and is present throughout. Picasso was absent, and absence removes the last trace of authorship. There is no question to trace back, no framing to resent later, and no moment at which Rosenberg can identify himself as having been led.
Which means the offer is not merely accepted as true — it is defended as his own. A price you named is a price you will justify to yourself, and that is a materially stronger position than a price you were talked into. The sales page's mechanism is strongest precisely where the operator does least, and the limiting case is doing nothing at all, which is what Picasso did.
Psychology — The Question You Actually Answered
Faced with a hard question you cannot answer — is this car sound — the mind silently swaps in an easier one it is equipped for, answers that, and reports the result with full confidence. You never notice the swap.
Rosenberg's hard question was why did Picasso stop? — genuinely unanswerable, since the only person with the information had gone quiet.
So he answered a different one: what would be bad enough to explain this? And that substitution has a direction built into it. The substituted question is not what is most likely; it is what would I most need to prevent — because the whole point of the substitution is to produce an actionable answer, and actionable answers are about risks, not about probabilities.
What the pair produces: silence does not merely fail to inform, it systematically biases what gets invented toward the expensive end. That explains the size of the offer, which a simple uncertainty account cannot. Rosenberg did not split the difference between benign and catastrophic explanations; he priced the catastrophic one. And it generalises past dealers: any withheld explanation is filled in by someone whose substituted question is what should I be protecting myself from, which is why unexplained silences are read as worse news than the truth almost always is.
Sharpest implication. The withdrawal cost Picasso nothing and Rosenberg two sleepless days — and the price moved by the amount of Rosenberg's labour, not the amount of Picasso's. The technique's whole yield comes from work performed by the target, which means it is available only to someone the target already believes has reasons.
Generative questions.