A troubadour is riding between castles in medieval southern France, reciting poetry and playing the perfect knight. At Javiac he falls in love with the lady of the house, Guillelma, and over months of songs and chess she falls in love with him.1
His friend Pierre, travelling with him, is in love with a different lady at the same castle. Pierre and his lady have a violent quarrel; she dismisses him; Guillaume patches it up. And afterwards Pierre says something that ruins Guillaume's life.
He says his love has increased tenfold. That there is no stronger love than the love that follows reconciliation. "The stronger and longer the disagreement, he told Guillaume, the sweeter the feeling that comes with peace and rapprochement."2
Guillaume, whose entire vocation is experiencing all the joys and sorrows of love, decides he wants that one too.
So he manufactures a quarrel. He feigns great anger with Guillelma, stops writing to her, and abruptly leaves the castle — staying away even through the festivals and the hunts.3
Guillelma sends messengers to find out what has happened. He turns them away. He is expecting anger — anger is the point, because anger is what he will have to beg his way out of.
Instead: "his absence had the opposite effect: It made Guillelma love him all the more."4
Now she pursues him. Messengers, love notes — and Greene notes how unusual this was: "a lady never pursued her troubadour."5
And Guillaume does not like it. Her forwardness makes him feel she has lost some of her dignity. "Not only was he no longer sure of his plan, he was no longer sure of his lady."6
Read that carefully, because it is the finding. The technique succeeded completely and the operator experienced the success as a loss. He did not get more of what he wanted. He got a different thing, in larger quantity, and it devalued the person supplying it.
After several months she gives up and stops sending messengers. And now Guillaume starts to wonder — perhaps she is angry after all? Perhaps the plan worked?
"So much the better if she was."7
He puts on his best robe, decks his horse in its fanciest caparison, chooses a magnificent helmet, and rides to Javiac to be forgiven.
She hears he has returned, rushes out, kneels before him, drops her veil to kiss him, and "begged forgiveness for whatever slight had caused his anger."8
His plan has failed completely. She was never angry. There is no reconciliation to be had because there was never a rupture — only his own.
So, still desperate for the sensation, he tries once more: he drives her away with harsh words and threatening gestures. She leaves, vowing never to see him again.9
By the next morning he regrets it. He rides back and she will not receive him; her servants chase him across the drawbridge and over the hill.
Strip the courtly setting and the structure is precise, and unusual for this book: an operator running an intervention on a system he is inside, for an outcome that is his own emotional state.
That is what makes the case go wrong in a way none of Law 16's other cases do. Deioces withdrew to obtain a kingdom — an external object, unaffected by how he felt about it. Guillaume withdrew to obtain the feeling of reconciliation, which is a state that exists only inside the relationship he was manipulating.
So each move degraded the thing he was trying to harvest.
The first withdrawal produced devotion rather than anger — and devotion, freely offered, is exactly what makes reconciliation impossible, because there is nothing to reconcile.
Her pursuit lowered his estimate of her, which lowered the value of whatever feeling she could give him.
The second, deliberate cruelty finally produced a real rupture — and a real rupture is not a staged one. It could not be lifted by riding up in a good helmet.
Reconciliation is a by-product, not a substance. Reconciliation feels the way it does because the danger of loss was real. Manufacture the danger and you get either no danger (she isn't angry) or real danger (she won't see you). There is no third setting, and the whole plan depended on one.
Then the part Greene reports flatly and does not interpret.
Guillaume collapses in his chamber and cries. Over the next year, unable to see her, he "experienced the absence, the terrible absence, that can only inflame love." He writes one of his most beautiful poems, "My song ascends for mercy praying." He sends many letters explaining what he did and begging forgiveness.10
Eventually Guillelma, remembering his songs and his figure and his skill in dancing and falconry, finds herself wanting him back. As penance she orders him to remove the fingernail from the little finger of his right hand and send it to her with a poem describing his miseries.
He does. And "Guillaume de Balaun was able to taste the ultimate sensation—a reconciliation even surpassing that of his friend Pierre."11
He got what he wanted, and every step that produced it was a step he did not plan. The staged absence failed. The real absence — undesired, involuntary, a year long, and costing him a fingernail — worked. The chapter files this under a law about deliberate withdrawal.
Greene's interpretation extracts the general rule and does not touch the irony: "At the start of an affair, you need to heighten your presence in the eyes of the other. If you absent yourself too early, you may be forgotten. But once your lover's emotions are engaged, and the feeling of love has crystallized, absence inflames and excites."12
That is a real sequencing rule and the case does support it.
He also names a second mechanism: "Giving no reason for your absence excites even more: The other person assumes he or she is at fault."13 Guillelma did exactly this — she knelt and apologised for a slight that never occurred.
And the reverse: "the more Guillelma pursued Guillaume, the less he loved her—she had become too present, too accessible, leaving no room for his imagination and fancy."14
But the interpretation stops before the part that matters, and states the law's own condition as a general truth about lovers rather than as what it is here: the operator was subject to the mechanism he was operating. He ran it on her, it ran on him, and it worked on him harder.
You want something from a relationship that the relationship is not currently producing.
Not a favour — a feeling. To be missed. To be chosen again. To have someone realise what you are worth. And you have noticed that if you pulled back, you would probably get it.
You might. Guillaume did, eventually, and this is the case worth holding onto before you start.
Run the check he did not. Ask what specifically you are trying to harvest, and then ask whether it is a thing the other person supplies or a by-product of a situation being real. Attention, help, time — those are supplied, and withdrawal genuinely changes their price. Relief, remorse, the sweetness of being forgiven — those are by-products of actual risk, and manufacturing the risk destroys the ingredient. There is no dial between she isn't angry and she won't see you.
Then the second check, which is the one the case is really about. You are inside the system you are adjusting. Guillaume's plan assumed his own valuation would hold steady while hers moved. It did not — her pursuit lowered his estimate of her, and by the time the manoeuvre resolved he no longer wanted the thing he had been engineering. Before you withdraw, ask what you will think of them if it works, because a person who chases you is a different person from the one you withdrew from.
And notice what actually produced the outcome here: a year of involuntary absence he would never have chosen, a genuine apology, and a real cost paid. If that is what it takes, the staged version was never going to be a shortcut to it.
The strongest thing about this case is that it contains its own control. The staged absence and the real absence are run on the same two people, in sequence, with opposite outcomes — which is better evidence than the chapter's other cases, where the withdrawal has no comparison.
Tension: the case refutes what it is used to prove. Filed under a law about deliberate withdrawal, and the deliberate withdrawal fails twice. What works is an involuntary year-long separation the operator would never have chosen and could not have engineered.
Tension: the operator is inside the mechanism. Greene's interpretation treats absence as a tool applied to a target. Guillaume's own valuation moved during the operation, in the direction that defeated it, and the interpretation does not register this even while reporting it.
Tension: the mechanism's success and the operator's goal came apart. It made Guillelma love him all the more is the law working perfectly, and it is the sentence in which the plan dies.
🚩 SINGLE SOURCE · 🚩 SECONDARY WITHOUT PRIMARY — no source is named. The material derives from the Occitan vidas and razos, medieval prose lives of the troubadours, which are themselves later narrative elaborations around the poems rather than biography. Greene narrates internal states — what Guillaume expected, felt and concluded — as fact. [POPULAR SOURCE] · [PLAUSIBLE — needs corroboration]
Open questions. Is the razo tradition even trying to report events, or is it constructing a backstory that explains a surviving poem — in which case the whole case is an origin story reverse-engineered from "My song ascends for mercy praying"? And the one the chapter raises without noticing: if the operator's own valuation is in play, is there any withdrawal that is purely tactical, or does every one of them change what you want by the time it resolves?
Against Deioces, the chapter's other long case, this is a near-perfect inverse and the pairing produces the law's real content.
Both men engineered an absence deliberately. Deioces got a kingdom; Guillaume got a devotion he did not want, then a rupture he could not undo. The difference is not skill or intent — it is that the outcome depends entirely on what the other party's imagination supplies while you are gone, and the operator does not control that. Deioces' population supplied divinity. Guillelma supplied guilt and love. Guillaume wanted anger, and there is no version of withdrawing that specifies which one you get.
Greene presents absence as a lever with a predictable output. His two flagship cases jointly demonstrate that it is a lever with an unspecified output, and that the operator's own preferences do not enter into it.
Against Absence and Withdrawal as Active Strategy — the vault's existing page, corrected and enriched in this build — Guillaume is the case that page most needs and does not have. That page distinguishes tactical from panicked retreat and locates the difference in the operator's intent and signalling. Guillaume's retreat was fully tactical, correctly signalled, and produced neither result on the menu. Intent does not determine which of absence's effects arrives.
Psychology — Loss Aversion and Asymmetric Valuation
Losses loom larger than equivalent gains, consistently and by a measurable factor. The same quantity of value hurts more leaving than it pleases arriving.
That is the engine under the whole law, and it explains why Guillaume's plan was structurally incoherent rather than merely unlucky.
Absence works — when it works — because it converts having you into losing you, moving the other person from the flat gain side of the value function to the steep loss side. Guillelma's love intensified because his departure registered as a loss, exactly as predicted.
But the feeling Guillaume wanted to harvest lives on the same curve, and requires the loss to be real. Reconciliation is sweet in proportion to how much was genuinely at risk. So he needed her to experience a real loss while he experienced none — and the case shows that is not available, because he was in the relationship too. The year of actual absence delivered the sensation precisely because it cost him a year, a poem and a fingernail. The intensity he was chasing is not an output of the technique; it is a receipt for what was actually spent. Which generalises past troubadours: any attempt to manufacture the feeling of a high-stakes situation without the stakes is trying to buy the receipt without the purchase, and the two are the same object.
Creative practice — Picasso and Rosenberg
Picasso withdrew his paintings from Rosenberg for no stated reason and let the dealer's own imagination raise the price — forty-eight hours of a man working out why, and then a substantially higher offer. "The best calculation is the absence of calculation."
Structurally identical to Guillaume: withdraw, give no reason, let the imagination fill the gap. And the outcomes diverge completely, which is the useful part.
Picasso wanted a number. Guillaume wanted a feeling. A number is external — Rosenberg's anxiety converts into a price, and Picasso's own state during those forty-eight hours is irrelevant to whether the manoeuvre worked. He could be indifferent, and was.
The insight the pair produces: withdrawal is reliable exactly to the extent that the payoff is external to the operator. Picasso is outside his own mechanism; Guillaume is inside his. That is a checkable condition and it sorts every case in Law 16 cleanly — Deioces wanted a throne and got one, Duveen wanted prices and got them, Charles V wanted a reputation and got one. Every success in the chapter is someone harvesting an external object; the one failure is the man harvesting his own emotional state. Greene has all the cases and never states the rule, and it is the rule that tells you in advance whether the technique is available to you.
Sharpest implication. Absence does not produce a specified feeling in another person — it produces a gap, and what fills the gap is theirs. Guillaume wanted anger and received love; Deioces expected need and received worship. The operator chooses only whether to leave. Everything that makes the technique worth doing is supplied by someone whose imagination you do not control.
Generative questions.