Greene ends Law 11 with a warning rather than a technique, and it is aimed at a specific fantasy:
One last warning: Do not imagine that your master's dependence on you will make him love you. In fact, he may resent and fear you.1
That is worth sitting with, because it corrects something almost everyone assumes. The intuitive model of becoming indispensable is that value produces affection — you make yourself useful, they come to appreciate you, and appreciation is what protects you.
Greene says the second half is false and the protection comes from somewhere else entirely.
But, as Machiavelli said, it is better to be feared than loved. Fear you can control; love, never.1
The first sentence is the famous one. The second is Greene's own gloss, and it is doing the real work — it supplies a reason Machiavelli's version does not.
Fear you can control; love, never.
The claim is not that fear is stronger. It is that fear is manufacturable and maintainable, while affection is a state in another person that you can encourage and cannot operate.
Then the conclusion:
Depending on an emotion as subtle and changeable as love or friendship will only make you insecure. Better to have others depend on you out of fear of the consequences of losing you than out of love of your company.1
Here the passage needs reading carefully, because Greene's terms are looser than his argument.
He does not mean intimidation. Nothing in Law 11 involves threatening anybody, and the chapter's exemplar explicitly does not: "Bismarck did not have to bully either Frederick or William into doing his bidding."2
The phrase is "fear of the consequences of losing you."
That is not fear of a person. It is fear of a situation — the state of affairs that arrives when you are gone. Bismarck's kings were not frightened of Bismarck; they were frightened of facing their enemies without him.
Which changes the instruction substantially. It is not be threatening. It is make the counterfactual visible — ensure the other party has a clear picture of what happens in your absence.
That is why the chapter's method is threatened resignation rather than any kind of coercion: "He simply made it clear that unless he got what he wanted he would walk away, leaving the king to twist in the wind."2
The claim's evidence is the chapter's own flagship, and it is unusually good evidence because it contains the control condition.
Frederick William IV was courted "night and day." Whatever their relations, the arrangement worked.
William "disliked Bismarck intensely and had no intention of keeping him around." Greene repeats it: "despite his antipathy toward the man."3
And the outcome was identical. Prime minister. "Whenever the prime minister threatened to resign, the king gave in to him, time after time."3
The affection variable was moved from neutral to actively negative and the arrangement did not change. That is close to a controlled test, and it is the strongest support in the chapter for anything.
The Harry Cohn case runs the same experiment in a different register — a man who called people "Communist bastard" protecting a communist screenwriter — and reaches the same result.
Greene's argument is thin here, so it is worth stating the version his own material supports.
Affection is a stock. It accumulates from past interactions, and like any stock it can be spent, forgotten, or revalued. Law 2's mechanism describes exactly this: gratitude is "a terrible burden that he gladly discards" — a backward-facing claim that the holder is looking for a way to put down.
Dependency is a flow. It is regenerated by circumstance every morning, and it does not require anyone to remember anything or feel anything.
Which is why the condottieri failed with the largest possible affection balance. They had saved cities. Nothing in the ledger of past service transferred into the present, because past service is a stock and the employer's decision was about next year's costs.
Greene's real claim, better stated than he states it: anything pointing backwards decays, and only a live ongoing requirement holds. That covers love, gratitude, loyalty and reputation for past service all at once, and it explains the chapter's cases better than the fear/love binary does.
Worth naming, because the instruction is presented as costless.
A relationship maintained on necessity is one the other party would end if they could. That is the definition of the arrangement — they are there because leaving is expensive, not because they want to be.
So the strategy purchases security by permanently foreclosing the possibility of the other thing. William never came to like Bismarck across nearly thirty years. Nothing in the mechanism produces warmth over time, because the mechanism runs on the absence of alternatives.
And there is a second-order effect the chapter ignores entirely: a party held by necessity is continuously looking for the exit. Every one of those resignation threats taught the king what his vulnerability was. Bismarck's dismissal in 1890, by a third emperor who had not inherited the dependency, is outside Greene's account — and is exactly what this model predicts.
You are trying to make a working relationship durable.
First, stop investing in being liked as a security measure. Being liked is pleasant and worth having for its own sake. As protection it is a stock that decays, and Greene is right that you cannot operate it.
Second, make the counterfactual legible rather than making yourself frightening. The operative fear is what happens without you. Most people are terrible at this — they do valuable work invisibly and assume the value is understood. It usually is not, until they leave.
Third, prefer a live requirement to a record of service. Past contributions are a stock. Ask what someone needs from you next quarter, and be that.
Fourth, accept that dislike is survivable and stop treating it as a crisis. Two kings and a studio head is a reasonable evidence base. If the structure holds, hostility is noise — and trying to convert it into affection adds a variable that can fail.
Fifth, price what you are buying. A relationship held by necessity will end the moment necessity does, and the other party will not be sorry. If the relationship matters to you beyond its function, this instruction is actively working against you.
And the plain note: this is advice to be needed rather than valued, and to prefer it. It is effective. It also describes a working life in which nobody who keeps you around wants to.
Strongest support. The near-controlled test in the chapter's own flagship — the same arrangement surviving a shift from courtship to active antipathy, across a change of monarch — plus the Cohn case reaching the same result independently. Greene rarely has evidence this clean.
Tension — the terms are looser than the argument. Fear in the Machiavelli quotation means fear of a person; fear of the consequences of losing you means fear of a situation. Greene runs them together, and the chapter's own cases only support the second.
Tension — the fear/love binary is the wrong axis. The chapter's evidence supports live requirement versus backward-facing claim, which is a different and better distinction, and which absorbs the condottieri case that the fear/love framing cannot explain.
Tension — the strategy's own failure mode is predicted and omitted. A party held by necessity looks for exits; a successor who does not inherit the need is not held at all. Greene closes on Bismarck at the height and never discusses how the arrangement ends.
🚩 SINGLE SOURCE · 🚩 SECONDARY WITHOUT PRIMARY — the Machiavelli attribution is given without a work, chapter or edition, and "fear you can control; love, never" is Greene's own formulation presented immediately after it, in a way that invites the reader to take both as Machiavelli's. [POPULAR SOURCE]
Open questions.
The instruction sits directly against The Enemy as Untapped Resource at Law 2, and the two together reveal what this book actually thinks emotions are for.
Law 2 has Talleyrand recruiting Fouché, who had tried to have him killed, and Kissinger meeting his own would-be kidnappers. The instruction there is that hostility is workable material — an enemy has motives you can trade with, and the emotion is an input.
Law 11's instruction is that affection is unworkable material — too subtle and changeable to build on.
Put together, the position is consistent and strange: this book treats negative emotions as reliable and manipulable, and positive ones as unreliable and inert.
That is not a claim about human beings; it is a claim about what a strategist can operate. Fear, envy, resentment and need all produce predictable behaviour under pressure. Affection produces behaviour that is generous, variable, and — critically — not responsive to leverage, which from a tactical standpoint reads as noise.
So the corpus's preference for fear over love is a selection effect of its own method, not a finding about which bond is stronger. A book organised entirely around controllable variables will conclude that the controllable emotions matter and the others do not, and Greene never notices that this is a property of his instrument rather than of the world.
→ psychology · Ingratitude and the Burden of the Favor
Law 2's mechanism: receiving a favour implies you were chosen rather than deserving, the injury surfaces slowly as resentment, and a debt is a weight the debtor is looking for a way to put down.
That page explains why the love half of this instruction fails, in a way Greene's own gloss does not.
Greene says affection is unreliable because it is "subtle and changeable" — a claim about volatility. The psychology page says something sharper: gratitude does not merely fade, it actively converts into resentment, because being indebted is unpleasant and the cheapest way to discharge it is to devalue the creditor.
The insight the pair produces: the condottieri were not undone by their employers forgetting. They were undone by the size of what was owed — Siena's assembly killed a man precisely because the debt was too large to pay. Affection and gratitude do not decay to zero and stop; past a certain magnitude they invert.
Which sharpens this instruction considerably. It is not love is a weak bond. It is that an unpayable debt is a motive for removal, so the operator who accumulates enormous claims of past service is building the case against themselves. Necessity has no such property: needing someone next quarter creates no obligation to feel anything, and therefore nothing to escape.
→ history · Bismarck and the Weak King
The case supplies the test, and it also supplies the limit.
Two kings, one courted and one openly hostile, identical outcome across decades. That is as close to a controlled experiment as this book offers, and it establishes that the affection variable genuinely does not drive the result.
But the case also shows what fear-based dependency teaches the dependent party. Every resignation threat that worked was a demonstration to the king of exactly where his vulnerability lay. William gave in "time after time" — and each time learned the shape of the trap more precisely.
The insight neither page yields alone: a relationship held by necessity is under continuous review by the party holding the losing end. It does not decay like affection, but it does something affection never does — it generates, in the other person, a permanent active interest in restructuring their affairs so that it stops being true.
That is why the arrangement ends the way it does: not through a change of heart, but through a successor who arranges not to need him. Fear is more controllable than love and it is also the only bond that motivates the other party to escape it, and Greene's chapter recommends it without mentioning that the client is working against you the entire time.
Sharpest implication. The operative phrase is "fear of the consequences of losing you" — which is not fear of a person at all, but of a situation. That reframes the whole instruction from be threatening to make the counterfactual visible, which is why the chapter's method is a threatened resignation rather than any form of coercion. Most people fail this not by being insufficiently frightening but by doing valuable work invisibly, so the counterfactual is never legible until they have already gone.
Generative questions.