Sitting in the margin of Law 30, next to the Houdini material, is a passage by Baltasar Gracián that makes a stranger claim than the chapter it decorates.
*Keep the extent of your abilities unknown. The wise man does not allow his knowledge and abilities to be sounded to the bottom, if he desires to be honored by all. He allows you to know them but not to comprehend them. No one must know the extent of his abilities, lest he be disappointed. No one ever has an opportunity of fathoming him entirely. For guesses and doubts about the extent of his talents arouse more veneration than accurate knowledge of them, be they ever so great.1
Read that last clause again, because it is doing something the rest of the chapter never attempts.
Be they ever so great. Gracián is saying that being accurately known costs you veneration even if what is known is magnificent. The penalty is not for being ordinary. It attaches to being measured.
The chapter's main argument is Castiglione's: conceal the labour, because visible effort lets people conclude anyone could have done it with the same hours.
Gracián's argument is about the ceiling, and it survives the total absence of labour. Suppose everyone knows exactly how you did it and exactly how much work it took, and it was magnificent. Castiglione's discount has already been fully applied. Gracián says there is a second, separate loss still to come: once they know how far you go, the wondering stops, and the wondering was worth something on its own.
You can hold either one without the other. A famous surgeon whose hours nobody sees but whose limits everybody knows has done Castiglione and not Gracián. A visibly hardworking person of unknown depth has done Gracián and not Castiglione.
Greene fuses them into one instruction, and in fusing them loses the more interesting claim — because the second is much harder to obey. You can hide your drafts. You cannot easily hide your ceiling, since every completed task is a data point about where it is.
Why would an accurate estimate be worth less than an uncertain one?
Because veneration — Gracián's word, and it is chosen — is not a response to a quantity. It is a response to an open question.
A known ability, however large, is finite. It can be planned around, budgeted, compared, exhausted. People allocate you tasks up to your limit and no further, and the limit is now a fact about the organisation.
An unfathomed ability generates a different behaviour: nobody knows where the edge is, so nobody knows what would be safe to attempt against you. That produces caution, and caution is indistinguishable from respect at a distance.
Gracián's honesty is in "lest he be disappointed." The mechanism assumes the audience's estimate of an unfathomed person runs higher than the truth — that uncertainty resolves upward. That is an empirical claim and it is not always true; uncertainty about an unknown quantity can just as easily resolve to nothing much there. What makes it resolve upward is having demonstrated enough to establish the floor. He allows you to know them but not to comprehend them — the first half is load-bearing and easy to skip.
Thirty-seven lines below the marginal, in the closing paragraph of the Keys, in Greene's own voice:
*Finally, because you achieve your accomplishments with grace and ease, people believe that you could always do more if you tried harder. This elicits not only admiration but a touch of fear. Your powers are untapped—no one can fathom their limits.2
Fathom is Gracián's verb. No one can fathom their limits is Gracián's claim, restated. There is no attribution.
This is the second confirmed lift from a marginal in two clusters. Law 28's Image — "The Lion and the Hare" — comes from a Cardinal de Retz marginal forty lines above it, also unattributed. Two chapters running, the same operation: a marginal supplies the sharpest formulation in the chapter, and it reappears a few dozen lines later as the author's own conclusion.
Worth being precise about what this is and is not. It is not plagiarism in the sense of hidden theft — the Gracián passage is printed on the same spread, attributed, in a larger typeface than the footnotes of most books. A reader who reads the margins will see it.
What it is, is the marginal doing the intellectual work and the body taking the credit for the conclusion — which is a fairly exact instance of what this chapter recommends. Greene has concealed the labour of his own argument by putting its source in the decorative column, and the effect on the reader is that his closing line arrives as insight rather than as citation.
The strategy has a failure mode Gracián acknowledges in four words and does not develop: "lest he be disappointed."
An unfathomed reputation is a liability you cannot discharge, because you can never find out what you are actually capable of in public. Every real test risks establishing the limit, and establishing the limit is the thing being avoided. So the person practising this seriously declines the tasks that would measure them — and declining is cheap, and looks like discernment, and compounds.
Over a career that produces a specific character: someone whose reputation exceeds their demonstrated output, who is thought to be holding something back, and who increasingly is not, because the muscle has not been used. The veneration survives, and the capability underneath it quietly stops being tested.
Houdini is the counter-instance and it is why he belongs in this chapter rather than merely in it. He was unfathomed — "to this day no one knows" — and he took a public, adversarial, potentially disqualifying test every night. The reputation was unfathomed and the capability was continuously verified. Those two are hard to hold together and almost nobody does.
You are asked what you can do.
The instinct is to be helpful and specific: here is my range, here is what I've handled, here is roughly where I top out. That is honest and it is also the moment your ceiling enters the record, and from then on you will be allocated work up to it and no further.
The Gracián version has two halves and people only ever do one.
Establish the floor. Do something visible, completed, unambiguously good. This is not optional and it is the half that gets skipped by people who mistake this law for mystique. Without a demonstrated floor, unfathomed reads as unproven, and uncertainty resolves downward.
Then decline to specify the rest. Not evasively — just stop where the answer stops being about work you have done. That's the sort of thing I've handled. Full stop. Do not extrapolate on your own behalf, and do not disclaim either, because a disclaimer is also a measurement.
Then the part that keeps this honest: privately, take the tests anyway. Somewhere the outcome is not on the record — a side project, a hard problem nobody assigned, a thing you might fail at. Otherwise you are running Gracián's strategy on yourself, and in ten years the unfathomed depth will be a room nobody has been into, including you.
Strongest evidence. A complete, named, dated passage from a source with standing — Gracián (1601–1658), whose Oráculo Manual is a genuine antecedent of this entire genre. Rare in this book for being an argument rather than an aphorism: it states a claim, gives a mechanism, and names the failure condition.
Tension — the passage is stronger than the law it decorates and the law absorbs it without credit. Documented above. Second lift in two clusters.
Tension — Gracián and Castiglione are fused. Concealing labour and concealing the ceiling are separable operations with different difficulties. The chapter treats them as one.
Tension — the mechanism assumes uncertainty resolves upward. It does so only above an established floor, which Gracián states ("allows you to know them") and Greene's restatement drops.
🚩 [POPULAR SOURCE] · 🚩 SECONDARY WITHOUT PRIMARY — no work is named, no edition, no translator. Gracián is dated but the passage is untraceable as printed. ⚠ Gracián's second appearance in five laws — he supplies Law 28's Authority ("it would be safer to do nothing"), which also argues against the law it closes.
Open questions. Is there any way to maintain an unfathomed reputation while continuously testing the capability, other than Houdini's — a profession that supplies adversarial public tests as a matter of routine? And does the strategy work at all for someone whose output is legible and countable?
Gracián appears twice in this cluster region and both times he undercuts his host chapter.
At Law 28 he is the Authority for Enter Action with Boldness, and his passage closes: "it would be safer to do nothing." Here at Law 30 he supplies a claim sharper than the chapter's own, which Greene then restates as his own conclusion.
Two chapters, two Graciáns, both of them quietly out of step with the law they are printed inside. He is the most consistently disruptive presence in this stretch of the book, and he is always in the margin.
Against the chapter's other cases, this page explains something they leave unexplained. Talleyrand and Manufactured Clairvoyance produced a reputation for second sight; Houdini and the Birmingham Manacles produced one for occult power. Neither man's audience concluded well-informed or extremely well-practised. Gracián's mechanism is why: an unfathomed capability does not sit as an open question in the audience's mind. It gets filled, and the filling is generous, and it is not chooseable by the person it happens to.
Behavioural economics — the cost of being priced. Costly Signaling holds that a signal's credibility comes from what it cost the sender, and that reliable signals are ones the unqualified cannot afford to fake.
Gracián's move is the inverse and it exposes a gap in the framework. He is not signalling a quantity at all — he is withholding the quantity while signalling that there is one. In signalling terms this should not work: an unverified claim of depth is exactly what the unqualified would also emit, so it ought to carry no information.
The insight neither produces alone: it works because of the floor. He allows you to know them but not to comprehend them — a verified sample plus a refusal to bound the rest. The sample is the costly signal; the refusal converts a point estimate into a distribution, and audiences appear to take the sample as a draw from an unknown range rather than as the range itself. That is a specific and slightly irrational inference, and it means the strategy is not signal-free after all — it is one honest costly signal, followed by strategic silence about the sample size, which is a distinguishable technique and a much more usable one than "be mysterious."
Eastern practice — the teacher whose depth is deliberately not displayed. How to Evaluate a Spiritual Teacher confronts the problem from the other side: a student trying to assess someone whose claimed attainment is by nature unverifiable, in a field where displaying depth is itself considered evidence against it.
That is Gracián's technique operating as an entire epistemic environment — and it shows what the strategy does when it becomes the norm rather than an individual's edge. When unfathomability is the expected posture, it stops discriminating: the genuinely deep and the empty produce identical surfaces, and the audience loses any way to tell.
What the pairing produces is the boundary condition on this law. Gracián's advice is positional — it pays because most people around you are legible, and your opacity stands out against that background. In a population where everyone is unfathomed, veneration accrues to nobody and the assessment problem lands entirely on the audience. Which predicts something checkable: the technique should be most powerful in transparent, measured, output-legible environments, and nearly worthless in fields that already run on mystique. The consultant benefits; the guru does not. And Greene, writing a manual that recommends it to everyone, is describing a strategy that stops working in proportion to how many readers take it.
Sharpest implication. There are two different concealments in this chapter and only one is discussed. Hiding your labour is Castiglione and it is easy. Hiding your ceiling is Gracián and it is hard, because every completed task is a data point — and it is the one that pays "be they ever so great," since the penalty attaches to being measured rather than to being small. But it only works above a demonstrated floor, and the person who skips the floor has not acquired mystique, only unprovenness.
Generative questions.