Background and Context
This is a saying with no moment — no room, no speaker, no occasion it can be pinned to. What it has instead is a use, and the use is legible in its grammar. Every one of the four lines ends in an order: teach, awaken, follow, shun. Nobody composes a sentence that way for contemplation. That is a sentence built for someone standing in front of a group of people and needing to decide, quickly and without a test, what each of them is and what to do about it. The proverb belongs to the vast body of instructional wisdom that circulated by mouth in cultures where teaching was a face-to-face craft with no curriculum behind it, and where a teacher's scarcest resource was not knowledge but attention — whom to spend it on. It survived because that problem never went away.
Interpretation
Strip the poetry and what remains is a truth table. Two binary variables — does he know, and does he know whether he knows — generating exactly four cells, each assigned a name and a verb. The compression is remarkable: a complete taxonomy of epistemic states in thirty-odd words, with the response already baked in. Nothing is left for the reader to work out except which person is standing in front of him.
The choice of variables is where the real argument sits. The first is ordinary — knowledge, the thing everyone thinks the proverb is about. The second is metacognition: not what you know but what you know about the state of your own knowing. And the proverb's entire claim is that the second variable dominates the first.
Watch how it does this. Take the two men who know nothing — identical on the first variable, indistinguishable in content. One is called a student and handed a teacher. The other is called a fool and sent away. Nothing about their ignorance differs. What differs is whether either can locate the edge of it. The proverb spends its harshest word and its most generous one on two people whose actual knowledge is the same, which is as clear a statement of priority as a sentence can make: awareness is load-bearing; content is not.
The second line is the strangest and the least discussed. He who knows and knows not that he knows is asleep. This is a person in possession of something he cannot find — competence with no receipt, understanding that has never been converted into a claim he could make out loud. The verb assigned to him is awaken, and the choice is exact. You cannot teach him; he already has it. You can only put him in a relation to what he is already carrying. This is unclaimed competence, and it is the only one of the four conditions where the remedy costs nothing but attention.
The fourth line is the one that makes the proverb famous and the one that makes it uncomfortable. Knows not, and knows not that he knows not — the state now widely called unknown unknowns, and the reason it terminates in shun rather than teach is mechanical rather than moral. Instruction requires the student to first register a gap; the whole transaction runs on the learner's own sense of absence. A person who perceives no absence has nothing for the teaching to attach to. The proverb is not calling him stupid. It is making a claim about the physics of instruction: you cannot fill a container that reports itself full. The brutality is a budget decision dressed as a verdict.
And then there is the trap, which is the deepest thing in the four lines and is never stated.
The proverb places the reader outside the matrix, as the sorter — the one who inspects four men and applies four verbs. But the reader is a person, and every person occupies one of these cells. Which one? The taxonomy offers no procedure for turning itself around. Worse: the single cell you would most need to detect in yourself is defined precisely by the impossibility of detecting it. The fourth man cannot use this proverb, because using it requires the faculty he is defined by lacking — and he will read it, recognize the fourth man immediately, and assign him to someone else. Every reader of these four lines finishes them feeling like the third man. That is not a flaw in the proverb. It is the proverb's most honest feature, visible only to a reader who notices what just happened to him while he was reading.
Finally, the order carries an argument of its own. Read as filed, the sequence descends from the teachable to the unreachable and closes on a door shutting — a warning. Read in the arrangement that appears in the oldest printings, which open on the fool and climb to the wise, it becomes a ladder, an ascent, an account of how a person moves upward through the states. Same four cells, two different sentences. What the proverb means is partly a function of which end of it you are left holding.
Current Relevance
The fourth cell acquired a research literature. Justin Kruger and David Dunning's "Unskilled and Unaware of It," published in the Journal of Personality and Social Psychology in 1999, reported that low performers across several domains systematically overestimated their competence, and offered an explanation the proverb had already assumed: the skills required to perform well and the skills required to recognize good performance are the same skills, so the person who lacks them is deprived of the instrument that would reveal the lack. Whatever one makes of the finding's contested statistical basis, the phenomenon it named — confident incompetence as a structural condition rather than a character flaw — is the fourth line restated in the vocabulary of experimental psychology, arriving roughly a century after the proverb's first English printing.
The vocabulary went further into public life than the research did. Donald Rumsfeld's 2002 formulation of "known unknowns" and "unknown unknowns," delivered at a Pentagon briefing and widely mocked before it was widely adopted, is a two-by-two grid on precisely the proverb's axes, and it entered risk management, intelligence analysis, and project planning as a standard framework. The proverb's structure had been rediscovered by people who almost certainly did not know they were rediscovering it.
The sharpest contemporary parallel, though, is a machine. A large language model generates fluent, well-formed, confident text whether or not it has the underlying knowledge, and its failure mode is not silence but articulate invention. In the proverb's terms, an unconstrained generative system is the fourth man rendered in software — it does not know, and possesses no representation of the fact that it does not know, and so cannot signal the gap that would let a user route around it. The entire technical effort now going into calibration, uncertainty estimation, and citation-grounding is an attempt to build the second variable into a system that started out with only the first. That is a striking place for a folk proverb about teaching to end up.
There is a social shift as well. When this saying was in oral circulation, knowing that you know not was a private state requiring a teacher to resolve. It is now a search query. The student cell has become nearly free — a person who can locate the boundary of his knowledge can cross it in seconds. Which raises the value of the second variable enormously, because it is now the only expensive part: retrieving an answer is trivial, and knowing that you needed one is the whole game.
Impact and Legacy
The proverb's most interesting legacy is that its structure keeps being independently reinvented by people building models of competence, and that the reinventions sometimes reverse its judgments.
The clearest case is the four stages of competence — unconscious incompetence, conscious incompetence, conscious competence, unconscious competence — a training model generally credited to Noel Burch at Gordon Training International in the 1970s and now standard in instructional design, coaching, and clinical education. Its four states map cell-for-cell onto the proverb's four men. But its top stage is unconscious competence: mastery so thoroughly internalized that it has become automatic and can no longer be narrated. That is the proverb's second man, the sleeper — and where the proverb says awaken him, the modern model treats his condition as the summit. The same epistemic state is read as a deficiency in one framework and as expertise in the other. The disagreement is real and instructive: unarticulated knowledge is a limitation if the person needs to teach, transfer, or defend it, and an achievement if he only needs to use it.
The Johari window, devised by Joseph Luft and Harrington Ingham in 1955 as a tool for self-awareness in group settings, is another four-cell grid built on a nearly identical pair of axes — what is known to self against what is known to others — and became a fixture of management training on its own track.
The proverb's other legacy is as a magnet for attribution. Having no author, it has been assigned to a series of them: Persian, Sanskrit, Confucius, and — four decades after his death — Bruce Lee. This is a common fate for ownerless wisdom, which tends to accrete famous names as it circulates, and it is why the saying appears in a hundred aggregators and almost no scholarly citations. What is worth noting is that the accretion is itself an index of reach: a proverb only collects false parents once it has traveled far enough that people want to know whose it is.
Contrasting Views or Controversies
The strongest objection is to the fourth verb. Shun him licenses writing a person off entirely on the basis of an assessment the assessor makes unilaterally, with no appeal and no test specified. Who decides that a man does not know that he does not know? Someone with power over the answer. Historically, the confident dismissal of the unteachable has run reliably in the direction of existing hierarchy — the poor, the foreign, the differently educated, the merely disagreeable. And the empirical case for it is weak: research on intellectual humility suggests that people who are confidently wrong are frequently movable, just not by direct correction, which triggers defense rather than revision. The proverb's own mechanism-based logic points to a different conclusion than its verb does. If the fourth man cannot register a gap, the response is not to abandon him but to create one — a demonstration, a failure, an experience — which is a slower and harder form of teaching, not the absence of teaching.
The second objection is that the binary is false at the root. Knowledge is not a possession one has or lacks. It is graded, domain-specific, and contextual — and any actual person is all four of these men simultaneously, sorted differently by subject. A physician may be the wise man in her clinic, the student in a courtroom, the sleeper about the thing she does so habitually she has stopped noticing it is a skill, and the fool about something she has never been given occasion to doubt. The proverb collapses a person into a cell. That is what makes it memorable and what makes it wrong: it assigns identities where it should be assigning descriptions of a moment.
Third, the empirical support has been substantially contested. The Dunning–Kruger effect in its popular form — "the incompetent are the most confident" — has been challenged on the grounds that much of the characteristic pattern reproduces from regression to the mean and a general better-than-average bias, and can be generated from randomly simulated data with no psychological mechanism at all. The underlying observation about metacognitive limits survives in weaker form, but the proverb should not be treated as having been validated by science; it has been echoed by science, which is a different thing.
The fourth objection is the least noticed and possibly the most serious: follow him is worse advice than shun him. The proverb's only instruction regarding the wise is deference — and its stated marker for wisdom is that the man knows, and knows that he knows. But that description is externally indistinguishable from the fourth man's. Both present as certainty. The proverb supplies no test whatsoever for telling the third case from the fourth, and then instructs the reader to follow one of them. A rule that says follow the person who is confident he knows is the exact rule by which demagogues, quacks, and gurus are acquired. To apply the taxonomy safely, a reader would need enough knowledge to grade the knowledge in front of him — which is to say the proverb quietly presupposes the wisdom it purports to help him locate.
Practical Application
- For a teacher, manager, or clinician: Diagnose the second variable before the first. The useful opening question is not "what do you know about this?" but "what part of this are you unsure of?" — an answer that reveals whether the person can find their own boundary, which determines the entire shape of the instruction that follows. A confident, fluent, wholly wrong answer and a hesitant, partial, correct one are opposite problems requiring opposite responses.
- For anyone leading a team: The sleeper is the highest-return cell and the one nobody audits. Ask people what part of their job they find easy — the answer is frequently a genuine expertise they have discounted precisely because it costs them nothing. Unclaimed competence does not appear in performance reviews, does not get taught to anyone else, and walks out the door when its owner does, still unlabeled.
- For self-assessment, which the proverb cannot help you with directly: Use proxies, since introspection is the one instrument the fourth cell disables. Write predictions down with confidence levels and check them later; the calibration record is external and cannot be talked out of. Actively seek the strongest disagreement rather than the friendliest. And treat the complete absence of felt uncertainty on a contested subject as a warning sign rather than a credential — that flat, frictionless feeling of obviousness is the fourth man's only symptom.
- For handling the fourth verb honestly: Read shun as an allocation rule, not a moral judgment. Instructional attention is finite, and spending it on someone who registers no gap is spending it on nothing. But allocation is reversible in a way that dismissal is not — the correct posture is to redirect effort toward creating the gap (a demonstration, a consequence, a question he cannot answer) or toward someone who can use the effort now, while leaving the door open.
- For working with AI systems: Demand the second variable explicitly, because these tools do not supply it by default. Ask what a model is uncertain about; require sources you can check; treat fluency as evidence of nothing. The proverb's most practical modern application is that the confident answer and the correct answer are produced by the same machinery, and only an external check can separate them.
Background on the Author
There is no author, and for a proverb this is not a gap in the record but a fact about how the thing was made. Sayings of this kind are not composed and then transmitted; they are transmitted into existence — smoothed across generations of retelling until the awkward words fall out and what remains is the version that fits the mouth. Any single instance is a snapshot of a form still in motion, which is why every recorded printing of these four lines differs somewhat from every other.
What authorizes an unowned saying is not credentials but selection. A proverb survives because people keep finding it worth repeating, and the filter is ruthless in a narrow way: it selects for memorability, brevity, and usefulness in the mouth of whoever is doing the repeating. That is real evidence of something — this particular arrangement has outcompeted every rival formulation of the same idea for well over a century in English alone — but evidence of use, not of truth. Sayings persist because they are easy to carry.
The text does, though, carry the fingerprint of who was carrying it. Four imperatives, one per line, addressed to a second person who is looking at someone else: this is not the syntax of a philosopher working out a theory of knowledge. It is the syntax of a working teacher with a room in front of him and limited hours. A philosopher would have interrogated the fourth cell's paradox — how a taxonomy of self-knowledge could be applied by a self. The proverb declines to, because the person it was built for did not need that question answered. He needed to know who to spend the afternoon on. The proverb's authority is the accumulated judgment of everyone who found it useful for that, and its blind spot is the question none of them had time to ask.