There are three ways a person can come to hold a piece of information, and they are not equal.
People are much more likely to believe something they see with their own eyes than something they are told. They are more likely to believe something they discover than something pushed at them.1
Which produces the technique:
If you plant the false information you desire them to have—with third parties, in neutral territory—when they pick up the clues, they have the impression they are the ones discovering the truth.1
And the scaling rule, which is the sharpest sentence on the page:
The more you can make them dig for their information, the more deeply they will delude themselves.1
Effort is not a cost the target pays to reach the conclusion. It is what makes the conclusion stick.
The case is small, cold, and complete.
During the First World War, Britain and Germany fought a lesser-known campaign for East Africa, where both had colonies. British intelligence there was run by Colonel Richard Meinhertzhagen, whose main rival on the German side was an educated Arab.1
Part of Meinhertzhagen's job was feeding the Germans misinformation, and he could not make it work:
he tried hard to deceive this Arab, but nothing seemed to work—the two men were equals at the game.1
So he stopped trying to deceive the Arab and deceived the Germans about the Arab.
He wrote his opponent a letter. It thanked him for his services as a double agent and for the valuable information he had supplied to the British. He enclosed a large sum of money.1
Then the move that made it work:
he entrusted the letter's delivery to his most incompetent agent.1
The Germans captured the man en route and found the letter. Under torture, he assured them the mission was genuine —
because he believed it was; Meinhertzhagen had kept him out of the loop. The agent was not acting, so he was more than believable.1
The Germans quietly had the Arab shot.1
Every element there is deliberate. The letter was found rather than delivered. It was found on a man selected for incompetence, so that its capture read as an accident. And its bearer could not betray it under torture because he did not know.
Greene names the specific problem the technique solves, and it is a problem about performance rather than about lying:
No matter how good a liar you are, when you deceive, it is hard to be completely natural. Your tendency is to try so hard to seem natural and sincere that it stands out and can be read.1
The trying is the tell. Which is why the answer is not better acting:
That is why it is so effective to spread your deceptions through people whom you keep ignorant of the truth—people who believe the lie themselves.1
The "Man Who Never Was" sidebar in this chapter makes the same point about a British attaché in Spain, told "confidentially" that important papers had been lost and asked to discreetly find out whether the courier's briefcase had been recovered:
The attaché was thus able to act out his part in the fake-out in a very convincing manner by virtue of the fact that for him it wasn't an act.1
One operational note, and it is the difference between a technique and a working method:
When working with double agents of this kind, it is always wise to initially feed them some true information—this will establish the credibility of the intelligence they pass along. After that they will be the perfect conduits for your lies.1
A channel has to earn its reliability before it can be spent, which means planted information is a slow instrument. The lie is cheap; the credibility that carries it is not.
The chapter does not linger on it and the page should.
Meinhertzhagen's operation required two people who did not consent to it. His own agent was selected for incompetence, sent to be captured, and tortured — and the torture was load-bearing, because his sincerity under it was what authenticated the letter. The Arab was shot on the strength of a document about services he had never rendered.
Greene's account records both facts and treats neither as a cost. The technique's central advantage — use people who believe it — is precisely a technique for making other people carry the risk of your deception without knowing they are carrying it.
You need a piece of information to be believed by someone who will scrutinise anything you say directly.
Stop planning what to tell them, and plan where they will find it.
Choose the location first, and choose neutral ground — a third party, a document, an overheard exchange, a channel they already monitor and believe they monitor without your knowledge. The value comes from the discovery feeling unprompted.
Then make it cost them something to get. A fragment that requires assembly, a document that took effort to obtain, a remark that had to be pieced together with something else — because the digging is what converts an item of information into a conclusion of theirs.
Then, wherever possible, route it through someone who believes it. This is the technique's real content, and it is also its ethical line: you are transferring the performance to a person who will not be performing, which works precisely because they cannot warn anyone and cannot be caught acting. Everything that makes this effective is a function of their not knowing.
Then build the channel before you need it. Send true things first. A conduit with no track record carries nothing.
And run the check that Meinhertzhagen's Arab could not: when a valuable piece of information reaches you by a route that feels fortuitous, ask who benefits from your having found it that way. Convenient discoveries are the most reliably manufactured kind, and the sense of having worked for it is not evidence.
Richard Meinertzhagen was a real British intelligence officer in East Africa and Palestine, and this episode is one of several attributed to him. [POPULAR SOURCE]
🚩 The source needs a specific flag. Meinertzhagen's reputation has been substantially revised: he is now widely regarded as a serial fabricator whose most famous exploits — including the celebrated "haversack ruse" — rest largely on his own diaries, which were demonstrably rewritten years afterwards, and whose ornithological collections were shown to contain systematic fraud. The most likely source of this anecdote is the man it makes look brilliant. Greene reports it as fact, and it may well be one.
And the "Man Who Never Was" comparison is doing quiet work. That operation deceived an enemy about an invasion; this one procured the execution of an individual. The chapter presents them as two illustrations of one technique, which is accurate about the mechanism and elides a considerable difference in what was being done.
The technique's success condition is also its blind spot. You cannot easily tell a planted discovery from a genuine one, which is what makes it effective — and it means a practitioner has no reliable way of knowing when the same thing is being done to them. Greene supplies the offensive move and no defensive one; the cui bono check above is this page's addition.
Against The D-Day Deception and the False Mirror this is the mechanism that made the whole operation work, and the D-Day case takes it one step further than this section does. There the Allies planted the truth with a discredited agent, so that Hitler's own act of seeing through a supposed deception was the thing that misled him. Same principle — let them do the work — with the polarity inverted.
Against Lawrence and the Taking of Aqaba the parallel is exact and Greene never notes it: Lawrence planted the Damascus story with an Arab chief secretly in Turkish pay, framed as an indiscretion. Acquired intelligence, from a source the enemy trusted, that the enemy had to work slightly to obtain.
Against Genghis Khan and the Shah of Khwarizm the same principle appears at its strongest: the shah's own son reported, from personal observation, that the Mongols were weak. Nothing was planted at all — the Mongols simply lost a battle and let the conclusion be drawn. That is the limit case of this technique, and it is the most durable version: no fabrication to discover, because there wasn't one.
Psychology → Self-Generated Conclusions Resist Correction. Beliefs a person arrives at themselves are held more confidently and are markedly more resistant to later disconfirmation than beliefs they were told — the effort of generation binds the conclusion to the self in a way that receiving does not.
The insight the pairing produces: this makes effort a design variable rather than an obstacle. The intuition is that you want a deception to be easy to accept; the finding says the opposite — the harder the target works, the more firmly the result is held and the more they will defend it against your later inconsistencies. Which explains why the strongest form of the technique is fragmentary rather than complete, and why a fully-formed story planted somewhere is weaker than three pieces that require assembly. It also names the defence, which is uncomfortable because it targets the thing that feels most like rigour: the conclusions you had to work hardest for are the ones you will be least able to revise, so effort expended is not evidence of soundness.
Business → The Leak as an Instrument. Product timelines, pricing intentions, funding positions and strategic direction all circulate through partners, recruiters, analysts and journalists — channels that carry information a firm would never state directly and that competitors weight heavily because they were not stated directly.
The insight neither domain gives alone: a leak is trusted in inverse proportion to how much the leaker seems to have intended it, which makes the ideal conduit someone who genuinely doesn't know. This is why the most credible commercial signals come from people who are not authorised to send them — a candidate mentioning a team's remit, a partner's account manager explaining a delay, an engineer at a conference. None of them is performing. And the corresponding hygiene is not tighter messaging discipline, which addresses only the deliberate channel, but asking of any competitive intelligence that arrives through a side door: would this have reached me if it were untrue?
The most convincing person to carry your deception is someone who doesn't know it's one — which is why the effective version of this technique always works by putting the risk on somebody who hasn't agreed to take it. And the defensive corollary is worse: the information you had to dig for is the information you will hold most firmly and revise least, so the effort you spent is not evidence that you were right.