One January morning in 1474, Louis XI of France — the "Spider King," so called because he wove the most intricate plots against his enemies — went into a vehement rant against the Duke of Milan.1
The courtiers listened in amazement as their normally composed and careful king spun out his suspicions. The duke's father had been a friend; the son could not be trusted. He was working against France, breaking the treaty. On and on it went, until Louis suggested he might have to take action.1
Then, to the courtiers' dismay, a man slipped quietly out of the room. It was Christopher da Bollate, the Milanese ambassador — received graciously by the king earlier that morning and then apparently forgotten in the background.1
The outburst threatened a serious diplomatic mess.
Robert Greene's point is that every part of it was deliberate.1
Later that day Louis invited Bollate to his private rooms and, lounging on his bed, began an apparently casual conversation. Drifting into politics, he described himself as a supporter of the Duke of Milan's — he would do anything to help the duke expand his power.1
Then the question, framed as innocent curiosity: "Tell me, Christopher, has it been reported to you what I said this morning in council? Tell me the truth — was it not some courtier who told you?"1
Bollate confessed he had been in the room himself and had heard it firsthand. He protested that the duke was a loyal friend of France. Louis replied that he had his doubts and cause to be angry — then immediately changed the subject to something pleasant, and Bollate eventually left.1
The next day the king sent three councillors to visit the ambassador. Was he comfortable in his lodgings? Happy with his treatment? Was there anything they could do to improve his stay?1
They also wanted to know whether he intended to pass the king's words on to the duke. The king, they explained, considered Bollate a friend and a confidant. He had merely been venting emotion. It meant nothing. Bollate should forget the whole thing.1
None of these men — the councillors, the courtiers, Bollate himself — knew the king had arranged all of it.
Louis was certain the perfidious ambassador, whom he hardly considered a friend let alone a confidant, would report every detail to the duke. He knew the duke was treacherous, and this was precisely how he wanted to deliver a warning.1
It worked. For the next several years the duke was an obedient ally.1
Greene's analysis of the alternatives Louis rejected is the useful part, because each one fails differently.1
Speak politely and diplomatically about his concerns — and the words carry no weight. They seem like whining.
Vent his anger directly at the ambassador — and he looks out of control. A direct thrust is also easily parried: the duke would simply mouth reassurances and the treachery would continue.
Transmit the threat indirectly — and it sticks.
The mechanism is specific and slightly counterintuitive: the fact that the duke was not meant to know made the anger genuinely ominous. It implied Louis was planning something and wanted to keep the duke from suspecting it or reading his true feelings. The threat delivered insidiously made the duke ponder his intentions and instilled an uneasy fear.1
This is the fourth of the chapter's five deterrence methods — playing on natural paranoia — and Greene's statement of its logic is the sharpest line in the chapter:
Make a threat and you may have to live up to it. Making them think you are working treacherously against them is another story entirely.1
A stated threat creates an obligation. An inferred one creates only imagination, and imagination is both free and worse than anything you would actually have done.
Someone has been quietly working against your interests — not openly enough to confront, but consistently enough that you're certain.
The direct move is a conversation, and Greene's read is that it fails twice over. Raise it calmly and it reads as complaint from someone with no leverage. Raise it angrily and you've shown you're rattled, which is information they can use. Either way, you've told them exactly what you know and what you're upset about, and they can now manage you.
The indirect version is to let the concern reach them through a channel you don't appear to control — a shared contact who will mention it, a comment made in a setting where it will travel, a decision you make that implies a reassessment without announcing one.
What makes it work is what you're not doing: you haven't made a demand they can refuse, or a threat you'd have to execute. You've simply let them learn that you might be reconsidering something, and left the rest to them.
The obvious caveat is that this is manipulation with a long history of going wrong. It depends on the message travelling accurately, which you can't control, and on the other party responding with caution rather than by escalating first. Louis was a king with real power behind the implication. The technique carries none of its own.
Louis XI's reputation for elaborate indirect statecraft is well established and the Bollate episode appears in the diplomatic record of the period. [POPULAR SOURCE] for Greene's framing and for the attribution of full premeditation.
Open tension: the account requires that every element was planned — the tirade, the apparent forgetting of the ambassador's presence, the private walk-back, the councillors' visit. That's a strong claim about intent, and the alternative reading is available: a genuinely angry king who then, realizing an ambassador had heard him, spent two days managing the fallout, and whose reputation for cunning caused the sequence to be recorded as design.
Greene needs the premeditated version, and the historical evidence for it is Louis's general reputation rather than anything specific to this incident.
A second tension: the technique's success is measured by the duke's obedience for "several years." That's the only outcome data, and it's consistent with the threat working, with the duke's calculations changing for unrelated reasons, or with the duke having been an ally regardless. A single case with a favourable outcome and no counterfactual is the weakest evidentiary form in the chapter.
Set this beside Strategy 5: The Command-and-Control Strategy, where Greene praises Elizabeth I for the opposite discipline — keeping her opinions so private that nobody outside her inner circle could tell where she stood on any issue, which made her impossible to mirror or manipulate.
Louis's method is to leak a private position deliberately. Elizabeth's is to have no leakable position at all.
What the pairing produces: these are both information-control strategies and they optimize for different threats. Elizabeth was defending against courtiers who would flatter her by reflecting her own views, so opacity was the defense. Louis was projecting toward an external rival who needed to be made uneasy, so a controlled leak was the weapon. The same behavior — being unreadable — is protection in one direction and useless in the other, and Greene never notes that his two admired monarchs are running incompatible policies on the same variable.
Psychology → Ambiguous Threat and Rumination. Research on anxiety consistently finds that ambiguous or unresolved threats produce more sustained worry and behavioral change than clearly specified ones, because a defined threat can be assessed and planned against while an undefined one cannot be closed out.
The insight the pairing produces: Louis's method exploits the asymmetry directly. A stated grievance can be answered — reassurances given, the matter considered closed. An overheard suspicion followed by an official denial cannot be closed at all, because the denial is itself suspicious. The duke was left with a problem that had no resolution available, which is exactly the state that produces sustained caution.
Business → Signalling Through Backchannels. Executives and negotiators routinely move information through intermediaries, off-record conversations and deliberately loose talk rather than official channels.
The insight: the reason is Greene's, and it's about commitment rather than deniability. An official position creates an obligation to act on it and a record to be held to. A backchannel signal changes the other party's behavior while leaving you free to do nothing — which means the technique's value is precisely that it lets you influence without committing, and its risk is that everyone sophisticated knows this and discounts accordingly.
Louis never made a threat, which is why the threat worked. A stated intention is a commitment you may have to honor and a position the other side can respond to. An intention they merely infer costs you nothing, obligates you to nothing, and is filled in by an imagination that will always produce something worse than what you'd actually have done.