King Hiero is having an argument with an enemy, and the enemy — reaching for something that will land — tells him his breath stinks.
Hiero goes home rattled. Not about the enemy. About his wife.
How does it happen that you never told me of this problem?
And she gives an answer that is somehow worse than any excuse: "Sir, I had thought all men breath had smelled so."1
She wasn't hiding it. She had no comparison. The one person closest to him had lived with it so long and so exclusively that she'd filed it under how the world is.
The moral, as Greene reproduces it:
Thus it is plain that faults that are evident to the senses, gross and corporal, or otherwise notorious to the world, we know by our enemies sooner than by our friends and familiars.1
That's a specific claim, and the specificity is what makes it usable. Not enemies are useful. Enemies are your earliest reliable source on a particular class of fact: the ones that are obvious to everyone outside and invisible to everyone inside.
There are two separate failures in the story and it's worth keeping them apart, because they have different fixes.
The one everybody expects: friends withhold. Greene says so a few lines later — friends "cover up their unpleasant qualities so as to not offend each other," and "since honesty rarely strengthens friendship, you may never know how a friend truly feels."2 That's a motivation problem. They know and won't say.
The one the story is actually about: the wife wasn't withholding. She genuinely did not know it was unusual.
That's not tact. It's calibration failure. Proximity destroyed her baseline. She had one long-term sample and no control group, so the thing that would have been glaring to a stranger in four seconds had become, to her, simply what men are like.
The second failure is the dangerous one, because the first has an obvious remedy — ask people who'll be blunt — and the second doesn't. Your blunt friends have the same broken baseline. They'll tell you the truth about what they can see, and this is precisely the category they can't.
The enemy has three properties here, and only one of them is hostility.
They have a fresh baseline — no adaptation, because they haven't been standing next to you for years.
They have no relational cost — nothing to protect by staying quiet, so the calculation that silences a friend doesn't run.
And they are motivated to look. This is the part that makes the enemy better than a neutral stranger. A stranger has the clean baseline but no reason to search. An enemy is actively hunting for the thing that will land, which means they've done the work of finding your most exposed feature and confirming it's real. An insult that misses is useless to them, so they filter for accuracy before delivery.
The delivery is hostile. The selection is rigorous.
That third property deserves its own moment, because it's the counterintuitive one.
The obvious fix for a broken baseline is a fresh pair of eyes — ask someone new. And fresh eyes do solve the calibration problem. What they don't solve is the search problem.
A stranger has no reason to go looking. They'll register whatever is unavoidable and move on, and they'll say nothing about it, because volunteering criticism to someone you've just met is socially expensive and pays nothing.
The enemy has done the search. They have gone looking for your most exposed feature, tested candidates against the question would this land, and discarded the ones that wouldn't. A miss is worthless to them — it makes them look petty and leaves you untouched — so they've pre-filtered for things that are both real and visible.
Which is why hostile testimony beats neutral testimony on this specific job. The enemy is the only party with a fresh baseline who is also paying attention on purpose.
This is the whole practical difficulty, and the story models it in Hiero's reaction.
Hiero doesn't defend himself, and he doesn't dismiss the source. He goes home and checks. Then, finding it true, he's angry at the person who should have told him — not the person who did.
That's an unusual sequence, and it's the technique the page is actually about.
An insult arrives as two things fused: a claim and an intent to wound. Almost everyone processes the intent first, and once you've done that the claim is contaminated — you now have a reason to want it false, and you'll find one.
Splitting them requires a deliberately mechanical move. Take the claim out, strip the adjectives, and ask only: is the factual core of this true? Not is it fair, not was it kind, not did they have standing to say it. Just true or not.
Most of the time it's exaggerated but not empty. The enemy overstates, because overstatement wounds better — but they overstated something real, because a fabricated flaw doesn't land and they know it.
So the yield isn't the accusation. It's the target the accusation was aimed at. They've told you where your unguarded surface is, and they had to do genuine reconnaissance to find it.
The last move in the anecdote is the one worth copying, and it isn't aimed at the enemy at all.
Hiero verifies the claim, finds it true, and then goes and asks his wife why she never said.
That's the second audit, and most people never run it. Having discovered the flaw, the instinct is to fix the flaw — which is right, and insufficient. The flaw is one data point. The silence is the system.
Because whatever produced years of nobody mentioning this is still running, and it is currently concealing everything else in the same category. Fix the breath and you've fixed one thing. Understand the silence and you've found out why you'll be the last to know about the next one, too.
Hiero got a real answer, and it was worse than a lie: she genuinely hadn't known. That told him his problem wasn't a coward in his household. It was that he had no one left with a working comparison.
Someone who doesn't like you said something cutting about your work last week, and you've been carrying it around.
Do this once, on paper, and then stop carrying it.
Write the sentence they said. Now write it again with every loaded word removed — no always, no never, no motive attributed to you. What's left is usually much smaller and much more specific. You bulldoze people in meetings becomes in the meeting on Tuesday, I spoke over two people.
Then check that. Not the feeling — the fact. Did you? You can find out.
If it's false, you're finished, and you can drop it with more confidence than you had before, because you actually looked.
If it's true, the useful question is the Hiero one, and it isn't about your enemy at all. Who else knew, and why didn't they say?
That answer tells you something structural about the people around you — whether you've built a circle that withholds, or one that has simply stopped being able to see you. The first is a courage problem and you can fix it by asking better. The second is a calibration problem, and the only fix is fresh eyes: someone new, someone outside, someone who has no idea what you're normally like.
Strongest support. The anecdote isolates the mechanism cleanly, and the wife's line does something rare — it demonstrates that the failure is perceptual rather than moral, which is a distinction the surrounding chapter never draws.
Tension — this is a third use of enemies, and Greene doesn't count it. Law 2 gives enemies two functions: recruitable labour (Talleyrand and Fouché) and sharpening pressure (Mao). Enemies as an information source is a separate thing entirely — it requires no relationship, no conversion, and no ongoing conflict. It sits in the marginal column and never reaches the Keys to Power.
Tension — the mechanism has no filter. Enemies are motivated to find real flaws and motivated to invent damaging ones. Plutarch's claim only works if hostile testimony skews accurate. Greene supplies no way to sort, and one anecdote where the insult happened to be true is not a general result.
🚩 SINGLE SOURCE · 🚩 SECONDARY WITHOUT PRIMARY — a Plutarch anecdote reproduced without edition or translator, in a marginal column, with no supporting case. [POPULAR SOURCE].
Open questions.
Greene's Never Invoke Gratitude, from The Laws of Human Nature, is one of a family of later-Greene techniques for managing what people around you feel and disclose.
All of them assume the disclosure problem is about willingness. Handle people correctly, remove the threat, and the truth comes out.
Hiero's wife is the counterexample sitting in the margin of the earlier book. There was no unwillingness to manage. She'd have told him instantly if she'd known it was worth telling.
No amount of psychological skill extracts information the informant doesn't possess — and the later Greene's toolkit is built almost entirely for the case where they do.
→ creative-practice · The Validation Lag
That page documents friends staying silent about work for two years, then praising it the moment a newspaper vouched for it. Proximity delays recognition, because close people evaluate against a fixed prior model of who you are.
Run it against Hiero and the two halves of one distortion appear.
The validation lag is proximity suppressing the good news. The bad breath is proximity suppressing the bad news. Same cause — the people nearest you have adapted to you and lost their comparison — expressed in opposite directions.
Which produces something neither page states: your close circle is not biased in your favour or against you. It's low-resolution in both directions, and it systematically reports you as more ordinary than you are — your strengths unremarkable, your flaws unremarkable, everything filed under how they've always been.
That reframes the practical advice. The usual instruction is to find honest friends, on the assumption that honesty is the missing ingredient. It isn't. You need friends with a working baseline, and past a certain number of years nobody close to you has one. Which is an argument for deliberately keeping some relationships shallow enough to stay diagnostic.
→ psychology · Jung's Blind Spot on Aggression and Defense
The psychology page examines a thinker failing to see something specific in his own material — a blind spot in someone whose entire method was the examination of blind spots.
Put beside Hiero, the pairing gives the sharpest form of the problem.
Hiero's blindness is external and correctable: the fault was on the outside of him, visible to any stranger, and one honest sentence fixed it forever.
The Jungian case is worse, because the blind spot is in the instrument. When what you can't see is a feature of how you see, no amount of fresh eyes helps unless those eyes are also willing to contradict your framework — and a framework's adherents are selected for not doing that.
So the two together give a hierarchy of self-knowledge failures with different remedies. Physical and behavioural faults need only a fresh baseline; any stranger will do. Structural faults in your own method need a hostile expert — someone with the competence to see the flaw and the motive to say it, which is a much rarer and more uncomfortable resource than an honest friend.
That is the strongest case for Greene's law that Greene doesn't make. Not that enemies are useful labour, but that for one category of self-knowledge they are the only available instrument.
Sharpest implication. The wife's answer is the whole page: I had thought all men breath had smelled so. The people who know you best have the worst calibration on you, not because they lie but because a long enough sample with no control group stops being data. Which means the standard advice — surround yourself with people who'll tell you the truth — solves the wrong failure, and solves it in a way that guarantees the remaining one goes undetected.
Generative questions.