Buried in the middle of Law 35's Observance, unmarked and never returned to:
Fouché sensed that in the long run neither side would emerge victorious. Power rarely ends up in the hands of those who start a revolution, or even of those who further it; power sticks to those who bring it to a conclusion. That was the side Fouché wanted to be on.1
Three roles — start, further, conclude — and a claim that the payoff attaches to the third.
The law it sits inside is about timing. This is not about timing. It is a claim about which function in a sequence of change is the one that gets paid, and it is testable in a way most of the book's assertions are not.
The sentence states the outcome and not the mechanism, so the mechanism is worth deriving, because it is not obvious.
Starting requires the qualities that make governing impossible. The people who begin a revolt are selected for willingness to act against a functioning order — which means they are on record, permanently, as the sort of person who does that. The new order needs the opposite reputation.
Furthering is worse, because it has no natural endpoint. The escalator is defined by the escalation; when the escalation stops, so does the role. Robespierre in this very chapter is the case: "He had sent heads on both the right and the left rolling" — and the function that could not stop is the function that was stopped.
Concluding is the only one of the three whose output is a stable state. The person who ends it delivers the thing everyone now wants — quiet — and the delivery is the credential. The starters and furtherers made the quiet valuable; the concluder sells it.
⚠ There is a second reason, less flattering and probably the operative one. By the time a revolution needs concluding, the people capable of contesting the position have mostly been consumed by it. Fouché's rivals were not outmanoeuvred. They were guillotined by each other.
Read it again with an eye on who does the work: those who start… those who further… those who bring it to a conclusion.
The value is created by the first two and captured by the third. That is a straightforward description of expropriation, and Greene states it approvingly without noticing that he has described the mechanism by which risk and reward are separated.
Which makes this the book's clearest statement of a pattern the corpus has now recorded repeatedly: the cat's-paw, the scapegoat, and this. Three chapters describing the same structure at different scales — someone else bears the cost of an act whose benefit accrues elsewhere — and only this one applies it to a whole political period.
⚠ And it supplies the defence, which none of the three chapters do: the position that gets paid is structural, not personal. If you are in the starting or furthering role, no amount of skill moves you to the concluding one, because the roles are defined by the sequence and not by the people in them. The only available move is to not be in the first two.
Fouché is offered as the proof and the fit is imperfect in an instructive way.
He does not conclude the Revolution. He votes for the king's death (furthering), runs the repression at Lyons (furthering), leads the conspiracy against Robespierre (arguably concluding the Terror), then joins the Jacobins, then serves the Directoire, Napoleon, the Bourbons, Napoleon again, and the Bourbons again.
That is not one revolution concluded. It is seven transitions survived, which is a different achievement.
The one moment that does fit the sentence exactly is July 1794. He assembles the coalition against Robespierre — "playing on the widespread fear of Robespierre—everyone was afraid of being the next to go to the guillotine" — and the Terror ends. He concluded that, and it is the single act that makes his later career possible.
So the sentence is right about the mechanism and Fouché is a partial instance. His actual method was to conclude one thing and then survive everything else, and the chapter presents the survival as the same skill as the conclusion. They are not the same skill. ⚠ Concluding requires assembling a majority; surviving requires being absent. Fouché is unusual precisely because he could do both, and the chapter does not register that as remarkable.
The Robespierre operation deserves the mechanism spelled out, because it is the only fully described conclusion in the chapter.
Fouché rallies both the moderates and the Jacobins — factions that agree on nothing. The shared property is not a position. It is "the widespread fear of Robespierre."2
A concluding coalition does not need a common programme. It needs a common threat, and a common threat is far easier to find than a common goal, which is why concluding is more available than it looks.
And it has a schedule attached. Robespierre had been generating that fear continuously by sending heads "on both the right and the left" — the furtherer manufactures the coalition that ends him, at a rate he controls and cannot slow, because slowing is the one thing his role cannot do.
Something in your organisation is being torn up — a restructure, a purge, a change programme — and you are deciding how visible to be.
The three roles are the frame, and the choice is more consequential than the position you take within them.
Do not be the one who starts it, however right you are. Starting requires demonstrating that you will act against the standing order, and that demonstration is permanent and unhelpful afterwards. The person who begins a change is, from that day, the person who begins changes.
Do not be the one who escalates. This is the genuinely dangerous role and it feels like the powerful one, because during the escalation it is. The role has no endpoint that isn't your removal, and the escalation itself accumulates the coalition against you at a rate you cannot reduce.
Be the one who ends it — and note what that actually requires, which is not force. It requires finding the thing everyone is separately tired of. A concluding coalition is assembled from a shared fear, not a shared aim, and the shared fear has usually been supplied by the person you would be replacing.
Then the timing question, which is the only genuinely hard part: too early and there is no coalition, too late and someone else has assembled it. The signal Fouché watched was fear of being next — that is, the point at which the escalation begins consuming people who thought they were safe. Before that there is nothing to build with; after it, briefly, there is a majority.
⚠ And the honest limit: this is a description of how position is captured, not of how anything good is achieved. Everything in it works exactly as well for someone with no aim beyond survival, which is what Fouché had.
Strongest evidence. A clearly stated three-role structure with one fully described instance — the July 1794 coalition, its composition (moderates and Jacobins together), and its binding agent (fear rather than agreement).
Tension — the chapter's exemplar is a partial fit. Fouché concluded one thing and survived six other transitions, and the chapter treats the two as one skill. They require opposite behaviours.
Tension — the sentence describes expropriation and is stated approvingly. The value is created by the first two roles and captured by the third; Greene does not remark on it.
🚩 [POPULAR SOURCE] · 🚩 SINGLE SOURCE — the generalisation is asserted flat, with one supporting case in the same paragraph and no other instance offered anywhere in the chapter. 🚩 CONSENSUS MISREPRESENTATION risk: "power rarely ends up in the hands of those who start a revolution" is the shape of a historical regularity and is offered without a single comparative example.
Open questions. Does the rule hold outside revolutions? Organisational change, product turnarounds and campaigns all have start/further/conclude structure, and the corpus has no case testing whether the payoff distributes the same way when the furthering role does not involve killing people. The consumption-of-rivals mechanism may be doing most of the work, in which case the rule is much narrower than stated.
This is the third structure in the book describing cost borne by one party and benefit collected by another — alongside the scapegoat and the cat's-paw — and it is the only one applied at the scale of a political era rather than a transaction.
But it sits against Law 34 awkwardly. The Strategy of the Crown is about being conspicuously exceptional; the concluder's asset is having been unobjectionable throughout, which requires exactly the invisibility Law 34 forbids. Fouché ran both, and the chapters do not acknowledge that they are asking for opposite things.
⚠ And against the epigraph directly above it: Sertorius argues patience as accumulation — hair by hair, "a steady continuous effort is irresistible." This sentence argues patience as abstention — do nothing until the concluding moment. Two different theories of what patience is for, printed six lines apart. ⚠ Worth flagging: Law 35's epigraph is the first in four clusters that is not crosswise to its law — it argues the law directly, and its disagreement with this sentence is about the mechanism rather than the conclusion.
Business strategy — the entrant who collects what the pioneer built. Star Performance Transfer Risk documents how performance attributed to an individual frequently belongs to the system around them, and does not travel when they do.
Read against the three roles, that produces the corrective this page needs.
The insight neither produces alone: the concluder's advantage looks like a personal quality — better judgement, better timing — and Greene presents Fouché that way. The transfer-risk frame says the opposite: the concluding position is a property of the sequence, and its occupant is largely interchangeable. What made July 1794 available was that Robespierre had frightened everyone, not that Fouché was insightful; anyone standing where he stood, with a coalition already manufactured, could have assembled it. Which yields the practical rule the chapter obscures: you cannot become a concluder by being good at concluding. The role is created by someone else's escalation and allocated by position, so the only actionable move is placement before the sequence starts — and that is a decision made in the boring period, when nothing is happening and the roles look like they do not matter.
Epidemiology — the exhausted host population. Mental Contagion and Mass Delusion treats collective states as running a course with an exhaustion phase, and holds that the phase, not the content, governs what a population will do.
The concluder arrives at exhaustion, and the pairing supplies the timing signal.
What the pairing produces: when to conclude is the whole difficulty — too early and there is no coalition, too late and someone else has it — and the contagion frame gives a specific, observable marker rather than a judgement call. The exhaustion phase begins when the process starts consuming people who believed themselves exempt. That is exactly Fouché's read: "everyone was afraid of being the next to go to the guillotine" — not the enemies of the revolution, everyone. Before that point the fear is distributed by faction and produces no cross-faction coalition; after it, the fear is universal and factional alignment becomes irrelevant, which is why moderates and Jacobins could combine. So the signal is not the intensity of the process but the moment its selection criterion stops being legible to the people inside it — and that moment is visible from outside, well before anyone acts on it.
Sharpest implication. Three roles — start, further, conclude — and the value is created by the first two and captured by the third, which Greene states approvingly without naming it as expropriation. The furthering role is the trap: it has no endpoint that is not your removal, and it manufactures the coalition that ends it, at a rate its occupant cannot slow. The concluding coalition needs a shared fear, not a shared aim, which is why it is more available than it looks — and the position is structural, not personal, so no amount of skill moves you into it. The only real move is placement, made in the boring period before the sequence begins.
Generative questions.