Michelangelo quarrelled with Pope Julius II over the building of the pope's marble tomb and walked out of Rome in disgust.1
The pope did not fire him. He "sought him out and in his own haughty way begged the artist to stay."1
Henry Kissinger survived every bloodletting in the Nixon White House. Not because he was the best diplomat available — "there were other fine negotiators" — and not because he and Nixon got on. "They did not."2
Two men, both unremovable, and Greene's point is that they achieved it by opposite methods, only one of which is portable.
Greene's line about Michelangelo is the cleanest statement of the first architecture:
Michelangelo, he knew, could find another patron, but he could never find another Michelangelo.1
Read the asymmetry in that sentence carefully, because it is the whole mechanism.
Michelangelo's supply of patrons: many. Popes, dukes, cities, the Medici. Julius's supply of Michelangelos: one.
That is not a description of talent, it is a description of a market imbalance running in one direction. The artist can walk; the patron cannot replace. And the walking-out is what tests it — Michelangelo actually left, which converted the imbalance from a theory into a demonstration.
Greene calls this intensive power: "depending on one skill, his ability as an artist."3
Kissinger's architecture is the opposite:
Kissinger survived because he entrenched himself in so many areas of the political structure that to do away with him would lead to chaos.2
Not one irreplaceable skill. Many replaceable involvements, arranged so that unpicking them is expensive.
Greene: "He got himself involved in so many aspects and departments of the administration that his involvement became a card in his hand. It also made him many allies. If you can arrange such a position for yourself, getting rid of you becomes dangerous—all sorts of interdependencies will unravel."3
Each individual thing Kissinger did could have been done by someone else. What could not be done by someone else was all of them at once, in that configuration, with those relationships already built.
This is the judgement worth extracting, and Greene states it in one sentence that most readers skim:
Still, the intensive form of power provides more freedom than the extensive, because those who have it depend on no particular master, or particular position of power, for their security.3
No particular master, or particular position.
Work out what that means in practice. Michelangelo's power travels. It is in his hands. If Julius dies, if Rome burns, if the commission is cancelled, the asset walks out of the building attached to the man.
Kissinger's power is located. It exists in a specific administration, in a specific web of departments and allies. Remove him from that building and the entire apparatus evaporates — the interdependencies that protected him were features of a place, not of him.
So the two forms differ on a variable Greene names only glancingly: portability. Intensive power is an asset you own. Extensive power is a position you occupy, and positions have landlords.
Which reverses the intuition about which is safer. Extensive power feels more secure day to day — many anchors rather than one — and is catastrophically fragile to a single event: the administration ends. Intensive power feels precarious — everything rests on one skill — and survives the loss of every particular arrangement.
Then a passage that shifts the law's moral register sharply, and it is easy to read past:
And if, in reality, you are not actually indispensable, you must find a way to make it look as if you are. Having the appearance of specialized knowledge and skill gives you leeway in your ability to deceive those above you into thinking they cannot do without you.4
The appearance of specialized knowledge.
This is a third architecture and it is neither of the two named ones. It is manufactured opacity — arranging matters so that what you do is hard to inspect, and therefore hard to price.
Greene immediately notes the weakness, and the note is more useful than the recommendation:
Real dependence on your master's part, however, leaves him more vulnerable to you than the faked variety, and it is always within your power to make your skill indispensable.4
So the faked version is explicitly the inferior one, and Greene says the real one is always available. Which raises the question of why the fake is offered at all — and the honest answer is that acquiring a genuinely irreplaceable skill takes years, and looking opaque takes an afternoon.
The last architectural variant is thrown away in a sentence and is the most democratic thing in the chapter:
And you do not necessarily have to entwine yourself around the master; another person will do, as long as he or she too is indispensable in the chain.5
You do not have to be irreplaceable to the person with power. You have to be irreplaceable to someone who is.
That is a substantially lower bar, and it is the mechanism that saves the blacklisted screenwriter at Columbia Pictures a few pages later — Lawson was Bogart's writer, and Bogart was the studio's star.
Greene gives it one clause and moves on. It is the only version of this law reachable by someone with no exceptional talent and no institutional position.
You want to be hard to remove.
First, work out which architecture you already have. Ask: if this organisation vanished tomorrow, what would I still own? If the answer is nothing but relationships inside it, your power is extensive and it is not portable.
Second, prefer the intensive form even though it feels riskier. One skill nobody can replace feels like a single point of failure. It is the opposite — it is the only form that survives the failure of everything around it.
Third, test the imbalance rather than assuming it. Michelangelo left Rome. That is what converted I think I am irreplaceable into a fact both parties could see. Most people never run the test and are therefore working from a guess about their own position.
Fourth, if you cannot be irreplaceable, be load-bearing for someone who is. The one-clause version. It requires no exceptional capability, only a genuine and visible link to someone whose removal is unthinkable.
Fifth, be clear with yourself about which of the two you are doing — building a real capability, or arranging to look unauditable. Greene recommends the second when the first is unavailable and admits it is weaker. It is also the one that collapses the moment someone competent actually looks.
And the honest note: the appearance of specialized knowledge strategy is a description of why a great many organisations cannot find out what their own people do, and Greene offers it as career advice without pricing what it costs everyone else.
Strongest support. The distinction is genuinely useful and cleanly drawn, with a case on each side, and Greene's verdict on which is freer follows from a stated reason rather than an assertion. The Michelangelo formulation — he could find another patron, but never another Michelangelo — states a market asymmetry in nine words.
Tension — the fake version undercuts the whole law. If the appearance of indispensability produces the same protection, then the law is not about becoming valuable but about managing perception, and Greene's own caveat (real dependence "leaves him more vulnerable to you") concedes the two are not equivalent while still recommending both.
Tension — Kissinger is used to prove the inferior form and elsewhere to prove the superior one. He is the extensive case here; at Law 7 he is the man who surrendered the China credit upward; at Laws 2, 3 and 4 he runs three further techniques. This is his fifth appearance across the build and no chapter treats the recurrence as evidence about the man.
Tension — neither case tests the failure mode. Michelangelo's portability and Kissinger's locatedness are asserted from successes. Kissinger's power did in fact end with the administrations that housed it, which is the prediction the framework makes and Greene does not check.
🚩 SINGLE SOURCE · 🚩 SECONDARY WITHOUT PRIMARY — the Michelangelo–Julius quarrel is historically attested; Greene's account of the pope's private reasoning ("he knew") is invention, and no source is named for either case. [POPULAR SOURCE]
Open questions.
Set this against Bismarck and the Weak King, the chapter's own flagship, and Greene's two-category scheme turns out not to contain his best case.
Bismarck is not intensive. He has no single irreplaceable skill — plenty of Prussians could draft policy — and Greene never claims he does.
He is not straightforwardly extensive either. His protection did not come from being entangled across many departments; it came from one function, handling the king's enemies, performed for a succession of kings.
So there is a third architecture in the chapter and it is the one the chapter is built on: owning a recurring problem. It is portable in a way extensive power is not — it survived Frederick's death and William's hostility — while requiring no unique talent at all.
That is a better position than either named one, and it explains the anomaly on the Bismarck page: he restored his master's power, violating the judgment's rule about never teaching them enough, and grew safer. Curing a weakness destroys intensive power. It does nothing to a man whose asset is that enemies keep arriving.
Greene has three architectures, names two, and demonstrates the third.
→ creative-practice · Michelangelo and Soderini's Nose
At Law 9, the same artist faces a patron who wrongly declares the David's nose too big — and rather than argue, he climbs the scaffold, taps a chisel on nothing, drops a handful of marble dust, and lets Soderini believe a correction was made.
The same man, in the two chapters, behaves in opposite ways toward a patron, and the pair explains why.
Against Julius II he walks out of Rome in disgust and is begged to return. Against Soderini he performs an elaborate deception to avoid a disagreement about a nose.
Greene never connects them, and the explanation is not temperament — it is that the two patrons held different things. Julius wanted Michelangelo, who could not be sourced elsewhere. Soderini controlled future commissions in Florence, and Greene says so explicitly: to offend him "would put future commissions in jeopardy."
The insight neither page produces alone: intensive power is not general, it is a claim on a specific scarce thing, and it does not transfer to adjacent negotiations. Michelangelo's irreplaceability covered the work itself and gave him nothing at all over the civic patronage system that allocated the work. He could refuse a pope and had to flatter a mayor.
Which is a real limit on the whole law. Be irreplaceable implies a general immunity; the same man demonstrates, one law apart, that it is a narrow and non-transferable one.
→ business · Harry Cohn and the Chain of Dependence
Columbia's president — a Republican with Mussolini's photograph on his wall — refuses to fire a blacklisted communist screenwriter at the height of the HUAC period, because Lawson was Bogart's writer and Bogart was Columbia's star.
This is Greene's one-clause variant made concrete, and putting it beside the two named architectures shows what it solves.
Michelangelo's route requires being irreplaceable — years of acquired capability, available to very few. Kissinger's requires institutional reach — a position senior enough to entangle yourself across departments. Lawson had neither. He was, in Greene's own words, one of many good writers, and politically toxic besides.
The insight the pair produces: indispensability is transitive, and the transitive form has none of the entry costs of the other two. You do not need a unique skill or a senior position; you need one genuine, visible, hard-to-sever link to someone who has one.
And it inverts the usual career instinct. The standard advice is to make yourself directly valuable to the most powerful person available. The chain says: attach to whoever is structurally unremovable, whatever their rank, because their immunity conducts. Bogart was not Cohn's superior. He was simply impossible to replace, and that was enough to protect a man Cohn actively despised.
Sharpest implication. The two architectures differ on portability, and it reverses which one feels safe. Extensive power — many anchors, many allies, entangled everywhere — feels robust daily and is destroyed entirely by one event: the institution ending. Intensive power — everything resting on a single skill — feels precarious and survives the loss of every particular arrangement, because the asset walks out of the building attached to the person. Greene's preference is correct and his reason is the whole content: those with intensive power depend on no particular master.
Generative questions.