Behavioral
Behavioral

The Strategic Plan for Finding Weakness

Behavioral Mechanics

The Strategic Plan for Finding Weakness

Law 33 heads its own list with a title the rest of the book does not use:
developing·concept·1 source··Aug 9, 2026

The Strategic Plan for Finding Weakness

Six Methods, Presented as a Plan

Law 33 heads its own list with a title the rest of the book does not use:

FINDING THE THUMBSCREW: A Strategic Plan of Action1

Then: "In planning your assault, keep these principles in mind"2 — and six bolded methods.

Method What it searches Named source
Pay Attention to Gestures and Unconscious Signals involuntary emission Freud; Talleyrand
Find the Helpless Child pre-defensive residue
Look for Contrasts the concealed opposite
Find the Weak Link the wrong person
Fill the Void insecurity, unhappiness
Feed on Uncontrollable Emotions disproportionate affect

It is the most systematic thing in this stretch of the book, and it is the only place Greene calls a list a plan.

They Are Not Six of the Same Thing

The word plan implies a sequence and there is none, but the deeper problem is that the six are not comparable operations.

Four are search procedures — ways of looking. Gestures, the helpless child, contrasts, and the weak link all tell you where to point your attention and what to treat as signal.

Two are exploitation procedures — ways of acting. Fill the void and feed on uncontrollable emotions do not help you find anything; they tell you what to do once you have.

So a genuine plan would run gestures → contrasts → helpless child → weak link, producing a finding, and then branch into fill or feed depending on what the finding was. Greene presents all six flat, in one register, as though they were alternatives.

Fill and Feed Run on Different Clocks

That matters because the two exploitation methods have completely different durations. Filling a void is an ongoing supply relationship — Duveen and Arabella Huntington, running for years. Feeding an uncontrollable emotion is a discharge — Caesar's javelins at Pompey's cavalry, over in minutes. Choosing between them is the actual strategic decision in this law, and the list does not present it as a decision at all.

The Search Methods Share One Assumption

Read the four search procedures together and they turn out to be one idea applied four ways.

  • Gestures: what leaks out below conscious control.
  • The helpless child: what was laid down before the defences existed.
  • Contrasts: what the visible trait was built to cover.
  • The weak link: who in the structure was never fortified.

Every one of them looks for something that was never defended, either because it could not be, was formed too early to be, is hiding behind something else, or belongs to somebody nobody thought to protect.

Greene states the principle once, in the gestures section, and does not generalise it: "Since we all try to hide our weaknesses, there is little to be learned from our conscious behavior. What oozes out in the little things outside our conscious control is what you want to know."3

That is the whole search doctrine in one sentence, and it makes the four methods a single method with four collection points.

Analytical Case Study: Schopenhauer Supplies the Seventh and Greene Uses It Without Listing It

In the margin, thirty lines after the list, a complete passage:

*A man shows his character just in the way in which he deals with trifles—for then he is off his guard.… if these defects show themselves in small things, or merely in his general demeanour, you will find that they also underlie his action in matters of importance, although he may disguise the fact.*4

Arthur Schopenhauer, named and dated. That is a seventh method, and it is not any of the six: it does not look at gestures, childhood, opposites or org charts. It looks at low-stakes behaviour, on the argument that guard is only mounted where stakes are visible.

And Greene runs it in the body without naming it. His Lustig interpretation reads Loller entirely off trifles: *"Loller, for instance, overtipped waiters, seemed nervous in conversation with the concierge, talked loudly about his business."*5

Overtipping a waiter is exactly Schopenhauer's case — an unimportant transaction where nobody is performing, revealing a need for the small recognition that a tip buys.

The chapter's most-used diagnostic is in the margin and not in the plan. Third instance in this cluster of the marginal column carrying the operative content. See Character Shows in Trifles.

The Plan Has No Stopping Condition

The structural defect, and it is the same one Law 31's seven methods had.

A search procedure needs a termination rule: how do you know you have found it, and how do you know when there is nothing to find?

Greene supplies neither. The premise — "they all have a weakness"6 — guarantees a target exists, which means a failed search is always attributable to insufficient looking. There is no state of the world in which the method returns nothing here.

That makes the plan unfalsifiable in operation, and it has a practical consequence rather than merely a logical one: an operator running it will keep going until they find something, and something is always available. A person who is slightly vain, mildly insecure, occasionally impatient — everyone reads as a hit.

The method's yield is guaranteed and therefore uninformative, and the chapter's cases are all instances where the reading happened to be right and large. There is no case of a misdiagnosis, which is exactly what you would expect from a corpus selected on outcome.

Gracián's Competing List

The Authority offers a different structure and the chapter does not notice:

All men are idolaters, some of fame, others of self-interest, most of pleasure. Skill consists in knowing these idols in order to bring them into play.7

Three categories, ranked by frequency, closed.

That is a target list where Greene's is a search list, and a target list solves the stopping problem: you are looking for which of three, and when you know, you stop. It also makes a falsifiable claim — most of pleasure — that could be wrong.

Greene's title comes from this Authority and his method does not. See The Thumbscrew.

Implementation Workflow

You need to understand somebody and the ordinary channels are giving you nothing.

Do not run six methods. Run the one principle they share: stop attending to anything they have prepared.

Prepared material is the presentation, the stated position, the answer to the question you asked. All of it is guarded and all of it is uninformative, and most people spend their attention there because it is what is offered.

Attend instead to four unguarded surfaces, which cost nothing to observe:

What they do in low-stakes moments. How they treat someone who cannot help them. What they do when a small thing goes wrong. Schopenhauer's trifles, and the highest-yield of the four.

What they bring up unprompted. Not the topic you raised — the one they return to. Repetition is not guarded because nobody notices themselves repeating.

Where the reaction exceeds the event. Disproportion is the single clearest signal in this whole apparatus, and it does not require interpretation, only arithmetic: was that response the size of that stimulus?

What is absent. The thing nobody in their position has and they never mention.

Then the discipline the chapter has none of: write down what you think you found and what would show you were wrong. If you cannot state a disconfirming observation, you have not made a reading — you have made an assumption that will now organise everything you notice.

And run the test from the parent page: could you say it to them? If the reading is correct and speaking it would help them, you are doing something defensible. If its whole value depends on their not knowing, be clear with yourself about which of the six you are actually running.

Evidence, Tensions, Open Questions

Strongest evidence. Six methods, five with a named case in the same chapter, spanning conversation, court structure, commerce and battlefield. Freud's line supplies a mechanism for the first and Greene's Lustig interpretation demonstrates a seventh.

Tension — the six are not comparable operations. Four search, two exploit; a real plan would sequence them and branch. Documented above.

Tension — no stopping condition, and the premise guarantees a hit. The method cannot return nothing, which makes its yield uninformative and its failures invisible.

Tension — the chapter's most-used diagnostic is in the margin. Schopenhauer's trifles, unlisted, and running under the Lustig case.

Tension — the Authority supplies a competing, closed, falsifiable taxonomy. Three idols against six procedures.

🚩 [POPULAR SOURCE] · 🚩 SINGLE SOURCE · 🚩 SECONDARY WITHOUT PRIMARY — the Freud quotation is given with no work, no date and no source; it is widely circulated in this form and its provenance is not established here. 🚩 No case of a failed or mistaken diagnosis appears anywhere in the chapter.

Open questions. Which of the six actually carries the yield? The cases suggest fill the void and contrasts do nearly all the work, and the other four are elaboration. And: is there any documented instance in the corpus of an operator reading someone wrong and acting on it — because a method with no recorded failures is a method whose error rate nobody has measured.

Author Tensions & Convergences

Greene's own later book runs this same list from the other side, and the pairing is the sharpest available.

The Laws of Human Nature takes the identical inventory — the childhood residue, the compensating opposite, the uncontrollable emotion, the unfilled void — and presents it as self-knowledge, with the reader as the person who has them. This chapter presents it as reconnaissance, with the reader as the person opposite.

Neither book's psychology needs modification to serve the other, which is a disclosure rather than a contradiction: this map is use-neutral, and the chapter that reads it as a plan of assault has added nothing to the map to make it one.

Within Law 33, note the structural echo of Law 31. The Six Methods of Controlling Options is also a flat list of named techniques, also unranked, also with the organising criterion supplied elsewhere in the chapter and never applied. Two laws, two lists, the same failure of assembly — and in both, Greene's own text contains the missing structure in fragments.

Cross-Domain Handshakes

Profiling practice — the same collection points, with an error rate attached. Reading People: Nonverbal Cues covers the systematic reading of involuntary signals, and — crucially — the discipline that goes with it: baselines, clusters, and the standing warning that a single cue means nothing.

Greene's gestures method has none of that apparatus. It says look at what leaks and gives no way to distinguish a signal from a person who is simply cold, tired or foreign.

The insight the pairing produces: trained observation practice treats the base rate as the whole problem — you cannot read a deviation without knowing what that person's normal is, so the first and largest cost of any such method is establishing a baseline you did not have. Greene's cases all quietly satisfy this and never mention it: Richelieu studied the court "as if it were the machinery of a watch", Duveen spent years with Arabella Huntington before selling her anything, Lustig watched Loller for days. The chapter presents the reading and omits the surveillance that made it possible, which is why the methods look fast and are not.

Clinical assessment — the reading you are required to state. Erickson and the Utilization Technique describes a practitioner reading resistance with great precision and then using it toward what the patient came for — the reading is returned to the person, in the form of an intervention they can accept or refuse.

That is the same diagnostic skill under an accountability structure, and the comparison isolates what the accountability actually buys.

What the pairing produces: a clinician who reads someone wrong finds out, because the intervention fails visibly and the patient comes back or does not. An operator running Greene's plan never finds out, because a wrong reading produces a failed manoeuvre that is attributable to a hundred other things, and the diagnosis is never tested independently of the action taken on it. So the two practices differ less in ethics than in error correction: one is a discipline with feedback and the other is a discipline without it, and a discipline without feedback does not improve with experience — it only becomes more confident. That is the strongest practical argument against this chapter and it is not a moral one.

The Live Edge

Sharpest implication. The four search methods are one method with four collection points, and Greene states the principle once without generalising it: stop attending to anything the person has prepared. But the plan has no stopping condition and a premise that guarantees a hit — everyone reads as slightly vain, mildly insecure, occasionally impatient — so the method's yield is certain and therefore uninformative, and its failures are structurally invisible.

Generative questions.

  • The chapter's most-used diagnostic is Schopenhauer's trifles, running under the Lustig case and absent from the plan. Which of the listed six actually carries any yield, and would the law survive being cut to two?
  • A discipline without feedback does not improve, it only becomes more confident. Is there anything in the corpus that would let an operator discover they had read someone wrong?
  • Greene's title comes from an Authority whose taxonomy is closed, ranked and falsifiable, and his own method is open, unranked and not. Why take the phrase and leave the structure?

Connected Concepts

Footnotes

domainBehavioral Mechanics
developing
sources1
complexity
createdAug 9, 2026
inbound links14
next in Robert Greene
Gestures and Unconscious Signals
> As Sigmund Freud remarked, "No mortal can keep a secret. If his lips are silent, he chatters with his fingertips; betrayal oozes out of him at every pore."