1615, and a thirty-year-old bishop of Luçon is given the biggest platform of his life: mouthpiece for the clergy before the three estates of France. It is "an immense responsibility for a man still young and not particularly well known."1
He delivers the Church line on every issue of the day. Then, at the end, he does something that has nothing to do with the Church.
He turns to the throne — where the fifteen-year-old Louis XIII sits beside his mother Marie de' Médicis, regent until he comes of age. Everyone expects the customary words to the young king.
Instead, however, he looked directly at and only at the queen mother. Indeed his speech ended in long and fulsome praise of her, praise so glowing that it actually offended some in the Church. But the smile on the queen's face as she lapped up Richelieu's compliments was unforgettable.1
A year later she made him secretary of state for foreign affairs.
The choice is legible only if you see what he declined.
The king was the future and everybody knew it — Louis would reach his majority, the regency would end, and courting the boy was the obvious long play. Richelieu ignored him completely.
Greene's reading:
he was looking for the weak link in the chain of power, and he saw that it was the queen mother. Not that Marie was obviously weak—she governed both France and her son; but Richelieu saw that she was really an insecure woman who needed constant masculine attention.2
Two separate judgements are packed into that, and only the second is unusual.
That Marie held the power was a fact about the calendar. That her hold on it was psychologically precarious — that a woman governing France needed something she was not getting — is the read, and the evidence for it is that the praise offended his own side. He was willing to pay a cost in the Church to deliver something she had no other supplier for.
The step underneath the step, and it is the most professional thing in the case.
Marie's power ran through Concino Concini — an Italian, her favourite, "or rather her lover, a role that made him perhaps the most powerful man in France."3
Concini was vain and foppish, and Richelieu played him perfectly—attending to him as if he were the king. Within months Richelieu had become one of Concini's favourites.3
As if he were the king. The weak link had a weak link, and Richelieu worked both.
Note the economy: this is the same technique applied recursively, at no extra conceptual cost, and it means Richelieu had two independent routes into the same power. Which turns out to matter within a year.
Then the event nobody had priced:
the young king, who up until then had shown every sign of being an idiot, had Concini murdered and his most important associates imprisoned. In so doing Louis took command of the country with one blow, sweeping the queen mother aside.4
Everything Richelieu had built is now a liability. He was close to Concini, who is dead. He was close to Marie, who is a virtual prisoner in the Louvre. Her advisers are out of favour and some are arrested.
And this is the point in the case where a search method stops being enough, because what he does next is not detection.
If everyone was deserting Marie de Médicis, he would stand by her. He knew Louis could not get rid of her, for the king was still very young, and had in any case always been inordinately attached to her. As Marie's only remaining powerful friend, Richelieu filled the valuable function of liaison between the king and his mother.5
He does not find a new weakness. He identifies an asset that has crashed in price and that he knows will recover, and he buys it when nobody else will hold it. In 1622 she has him made cardinal.
By 1623 Louis is isolated, indecisive, and being told by his mother that Richelieu is his only possible saviour. He "hated the cardinal with a passion, only tolerating him out of love for Marie"6 — and yields.
The moment Richelieu has the premiership:
Now Richelieu no longer needed Marie de Médicis. He stopped visiting and courting her, stopped listening to her opinions, even argued with her and opposed her wishes. Instead he concentrated on the king.7
Eight years of cultivation, ended in a paragraph.
Greene states the principle without comment: *"he discarded the queen mother, moving on to the next weak link in the chain: the king's own character."*2
The rung is disposable by design. And what follows is the chapter's most disturbing sentence, because it is not exploitation of an existing weakness but the deliberate enlargement of one:
All the previous premiers, understanding the king's childishness, tried to keep him out of trouble; the shrewd Richelieu played him differently, deliberately pushing him into one ambitious project after another — such as a crusade against the Huguenots and finally an extended war with Spain. The immensity of these projects only made the king more dependent on his powerful premier.8
Everyone before him protected the child. Richelieu manufactured situations the child could not handle, and on Greene's own account a crusade and a war with Spain are downstream of a premier securing his position.
The chapter ends in triumph: eighteen years, a unified France, "a strong European power for centuries to come."8
The vault holds what happened next. Richelieu and Marillac's Brother covers the Day of the Dupes, 1631 — a plot to remove him that "reached the top of the French government and included the queen mother", and that very nearly succeeded.
So the discarded rung spent the following decade working to destroy him, and on one day in 1631 came within reach of it. Richelieu survived "through a combination of luck and his own manoeuvring" — and luck is the word doing the work.
Greene's version of this method has no cost line. The corpus does: eight years of cultivation, one paragraph of discarding, and a decade of an enemy who knows exactly how you operate because she watched you do it to her. This is the same omission the build has now logged three times — Stavisky cut at 1930, Hartzell cut mid-run, and Richelieu cut at his apex.
Somebody senior has taken an interest in you and it is genuinely useful. They open doors, they say your name in rooms you are not in, and your position currently rests on their goodwill.
At some point you will not need them. That moment is the one this case is about, and the choice made there determines the next decade.
Do not let the relationship become conditional in a way that is visible. The specific failure is not ingratitude, it is timing: cooling off in the same quarter you stop needing something announces the arithmetic to everybody watching, including the next person you will need.
The version that works is unglamorous and cheap: keep going. The same calls, less often. Continue to be useful in the small ways you always were. It costs you an hour a month and it purchases the thing Richelieu paid for in 1631.
And be clear about the second half of this case, because it is not a grey area. Enlarging somebody's dependency on purpose — arranging that they cannot cope without you — is a different act from being useful to them, and it has a specific tell: you find yourself relieved when they take on something too big. Notice the relief. That is the whole diagnosis.
Strongest evidence. A dated sequence spanning 1615–1642 with named actors and verifiable public events — the 1615 Estates-General speech, Concini's murder in 1617, the 1622 cardinalate, the 1624 premiership, the Spanish war. The outcome is not in dispute.
Tension — the account stops at the apex and the corpus holds the reckoning. The Day of the Dupes, 1631, with Marie at the centre.
Tension — the deliberate enlargement of the king's dependency is the chapter's most consequential claim and rests on one interpretive paragraph. Whether Richelieu drove French policy to secure his position, or pursued policies he believed in that happened to have that effect, is not settled by anything Greene presents.
🚩 [POPULAR SOURCE] · 🚩 SINGLE SOURCE · 🚩 SECONDARY WITHOUT PRIMARY — no memoir, no dispatch, no historian cited, for a period with an enormous documentary record. Every reading of an interior state — Marie's insecurity, Louis's childishness, Richelieu's intent — is Greene's.
Open questions. Was the 1615 speech actually the calculated opening Greene describes, or a young cleric's flattery of the person who could advance him, read backwards through what followed? The case's whole architecture depends on the answer and the chapter supplies no evidence either way. And: does the Huguenot crusade survive the dependency reading, given that a French state suppressing an internal armed faction has motives that do not require a premier's ambition?
The two Richelieu pages in this vault are the same career at its beginning and at its near-end, and they belong to different laws.
This one is Law 33 — find the weakness, work the rung. Richelieu and Marillac's Brother is a man who cannot touch an enemy directly because doing so means implicating the queen mother, and who has to find an indirect route. The obstacle in 1631 is the person he used in 1615, and the corpus can state that where neither chapter does.
Within Law 33, this case demonstrates four of the six methods and Greene attributes it to one. The weak link (Marie), the weak link's weak link (Concini), the helpless child (Louis), and filling a void (masculine attention) are all running, sometimes simultaneously. It is less an illustration of a method than a demonstration that the six are not separable in practice — which is an argument against the plan's structure that its own flagship case supplies. See The Strategic Plan for Finding Weakness.
Business — the asset bought when nobody will hold it. The Extrapolation Strategy concerns reading where a position is going rather than where it currently is, and acting on the trajectory while the present price still reflects the visible state.
Richelieu's 1617 move is exactly that, and it is the part of the case that no search method explains.
The insight neither produces alone: the standard account of this chapter is find the weakness and press it. But the decisive act in Richelieu's career was not detection — it was holding a crashed asset through the trough, on a specific and checkable thesis: the king cannot remove his mother, the king is "inordinately attached to her", therefore her price recovers. That is not psychology, it is a structural argument about constraints, and it is the only move in the case that required nerve rather than perception. Which suggests the search plan is the visible half of a method whose actual scarce ingredient is a willingness to be positioned badly for several years — and nothing in Law 33 addresses that at all.
Depth psychology — the dependency deepened rather than met. Mentorship and the Apprentice Knight describes a relationship whose purpose is to be outgrown: the mentor is "what the apprentice will become if the initiation succeeds", and the supply is scaffolding with a removal date.
Richelieu ran the exact inverse on Louis XIII, deliberately, for eighteen years.
What the pairing produces is the diagnostic, and it does not require inspecting anyone's motives. Both the mentor and the premier supply an adult function to someone who lacks it; both produce reliance; both are effective. The difference is the direction of the curve — and it is observable from outside. A relationship in which the junior party's independent capacity grows is one thing; a relationship in which the tasks keep getting larger and the capacity does not is another. Louis XIII in 1642 was less able to govern without Richelieu than he had been in 1624, after eighteen years of the most capable premier in Europe, and that fact — not anybody's intent — is the whole finding.
Sharpest implication. The decisive act in this career was not finding a weakness — it was buying a crashed asset in 1617 when everyone else was selling, on a structural thesis about what the king could not do. The search methods are the visible half; the scarce ingredient was a willingness to be badly positioned for several years. And the method's bill is real and the chapter omits it: the rung he discarded in 1624 was at the centre of the plot that nearly ended him in 1631.
Generative questions.