A nurse walks onto the ward, glances at a patient she has been told nothing about, and says: this one is going to crash. Get me a doctor.
The monitors read normal. The numbers are normal. The chart says nothing. Ask her what she saw and she will struggle to answer. Something about the colour. The breathing was wrong. The way the patient was lying. She cannot give you the rule she used, because she used no rule. She walked in and she saw it — the way you notice that a room has gone quiet, or that a friend on the phone is not all right — and twenty minutes later the patient crashes, and she was right, and no one else on the ward would have called it.
Last week we had a name for what she has. We called it the perception of the particular — Aristotle's word, phronēsis — the trained seeing of what this case asks, the thing no rulebook reaches. And we treated it as close to mysterious: a special human capacity, the one thing a machine could never have.
But look at her again, because there is a second way to describe what just happened, and it is not mysterious at all. She has seen ten thousand patients. She has worked that ward for twenty years. She was wrong, early on, and saw what being wrong cost, and slowly the work trained her eye until it does this — sees the crash coming before the numbers do. That is not a gift from beyond nature. That is what getting very good at something looks like. From the inside it can feel like something beyond explanation. From the outside it is the most ordinary thing there is. It is expertise.
And if that is all it is — a very high skill, built the way every skill is built, by doing the thing ten thousand times — then a question we thought we had closed last week is open again. Skills can be taught. Skills can be built. And the thing we were so relieved to think a machine could never have may be exactly the kind of thing it could learn.
This is Philosophy for Us — philosophy for everyone, no degree required.
This is the third of eight episodes on what wisdom is. Last time, Aristotle divided the subject in two: wisdom is not a quantity of knowledge — not sophia, the facts we accumulate — but phronēsis, the perception of the particular, the seeing of what one situation asks. And that handed us a relief we never tested. If wisdom is perceiving the particular, then the machine, for all it knows, is not even in the contest. I asked you to hold that relief and not spend it. Tonight we find out whether it survives.
Because tonight someone argues that the perception of the particular is nothing special, and nothing specially human. It is simply what expertise is, seen from the outside — a high skill, built from experience like any other. And if he is right, the machine question does not stay closed. Tonight it is open again.
Let me build his case, because it is stronger than you might expect, and it challenges the very thing you were most relieved about.
His name is Hubert Dreyfus — an American philosopher who spent his career studying how people get good at things. With his brother Stuart he set out a model of how anyone moves from beginner to expert, and it has one striking feature. At every craft — chess, nursing, flying a plane, judging wine — people pass through the same stages, and the stages run in one direction. They run away from rules.
The beginner needs rules. The student nurse has a checklist: take the pulse, check the chart, look for these five signs. She follows it step by step, because she has nothing else. She is slow, and stiff, and she gets it right by working through the procedure. That is the first stage. As she improves, the rules do not grow more sophisticated — they begin to fall away. She stops checking the list. She starts to notice which things matter in which cases. By the time she is an expert, twenty years in, she is not following rules at all. She walks into the room and responds to the whole situation at once — the colour, the breathing, the wrongness — immediately, without deliberating, often without being able to say why. The rules did not improve. They fell away, and something else took their place.
And here is Dreyfus's claim, the one that matters tonight. That something else — the expert's immediate response to the whole situation — is not a different kind of thing from skill. It is what skill becomes at its highest point. It is the last stage of getting good, not a separate ingredient added at the end. The nurse who sees the crash coming is not exercising a mysterious faculty. She is doing what a master does in any field: responding to a situation she has met, in some form, ten thousand times, with a fluency that has left stated rules so far behind that it looks like perception. And it does not merely look like perception. On Dreyfus's account it is a kind of perception — but a perception that experience built, the way experience builds every expert skill.
Now set that beside Aristotle, and see what it does to him. Aristotle said the wise person does not follow rules; they see the particular. Dreyfus says: yes, exactly — and so does every expert at the top of every craft, and we do not call the chess grandmaster's feel for the board a mysterious human remainder. We call it skill. The taster who names the vineyard from one sip is not perceiving a truth no rule could reach. He has tasted forty thousand glasses, and his palate has been trained. So when Aristotle points at the seasoned judge and says: there, perception, irreducible, special — Dreyfus points at the same person and says: there, expertise, ordinary, the last stage of a training anyone can undergo by doing the work. The same nurse. The same seeing. Two different accounts of what it is.
And I want to be fair to how strong this is, because it is tempting to dismiss and you should not. The account does not ignore what Aristotle saw. It explains it. Aristotle noticed that the wise person cannot state the rule they are using — Dreyfus tells you why: at the expert level there is no rule, only trained response, and trained response was never going to fit into a sentence. Aristotle noticed that the young cannot be practically wise — Dreyfus tells you why: you cannot shortcut ten thousand repetitions; skill takes the time it takes. Aristotle noticed that wisdom comes only through experience — of course it does; that is how every skill comes. Point for point, the skill account explains the very things that made phronēsis look mysterious, and it explains them with no mystery at all. It may be the most unsettling description there is of what that special seeing is underneath. Not a refutation of Aristotle. Something stranger: the same facts, with the mystery removed — if it really has been removed, and not merely hidden. We will come back to that if.
So far this may feel like a quarrel between two philosophers about a nurse. Here is why it is about you — and about the machine you might turn to when you cannot sleep.
Last week the relief ran like this. Wisdom is the perception of the particular. The machine has only the store of facts — the general, the universal, everything there is to know. The particular is exactly where the general runs out. So the machine is not in the contest. The one human thing is safe.
See what the skill account does to that. If the perception of the particular is not a mysterious human remainder but a high skill built from lived experience, then whether a machine could have it is no longer settled. It is open. Skill is the kind of thing you build by training on cases, and training on cases is, as it happens, the one thing these machines do. The nurse got her eye from ten thousand patients. The real, unsettled question is whether a system trained on ten million could acquire something that does the same work. Not imitate it. Have it. If wisdom is expert skill, then the machine's exclusion stops being a law of nature and becomes what it always was: a bet about what a machine can be trained to do. And a bet can lose.
I have to tell you the strangest part, because it is what makes this rival honest rather than glib. The man whose account this is — Dreyfus — spent years using this very picture to argue that machines could not get there. That was his whole campaign. Real intelligence, he said, is embodied and situated, built from a lifetime in a body in a world, and the computers in front of him had none of that, so they would never reach expertise. He may well have been right about those computers. But notice what carries over and what does not. His account of what expertise is — the rules falling away, the trained response to the whole situation — that is the part the rival keeps. His bet about the machines was a claim about a particular technology at a particular time, and the technology changed under it. So the live form of this challenge keeps Dreyfus's account of expertise and bets the other way about the machine. That is allowed. The account does not come with a verdict attached. That is the whole point: it leaves the machine question open, not answered — in either direction.
And I do not want to let that slide into the easy inversion, the one that says: fine, the machine has more data, so the machine is wiser, and that settles it. That is not the claim, and it is only last week's mistake in reverse — reading wisdom straight off the size of the store. The honest claim is smaller and harder: we do not know. If wisdom is a skill, then whether a machine can have it is an open question about machines, not a settled fact about humans. The question you were sure had been closed was never actually settled.
So here is the question, and only your own case can answer it, not me. Take the thing you are expert at — the one field where you stopped following rules years ago and now you simply see it: the work, the child, the kitchen, the code, wherever it is that you walk in and know. Ask it honestly. Is that seeing just a very high skill — built from your own practice, teachable in principle, the kind of thing that could be built into something that is not you? Or is there something in it that no amount of trained pattern-response would ever become — something left over that is not skill at all? Aristotle bets there is a remainder. Dreyfus bets there is not. And the unsettling thing is how hard it is to tell, from the inside, which of them is right about your own case. Look for it. See whether you can find the leftover. You may look hard and still not be able to say either way — and that difficulty is exactly the thing to notice.
Let me push this one step further, because the skill account is not the only way to make wisdom ordinary, and the next way costs you more.
By now you may be resisting the skill account. Something in you wants to say: yes, the nurse's eye is a skill, the taster's palate is a skill — but wisdom, real wisdom, the thing you would want when your life falls apart, is not just a knack. A knack is narrow. Wisdom is about your whole life — what matters, what is worth wanting, how to live. Surely that is not simply expertise at a task.
There is a line of thought, very much alive in philosophy now, that agrees with you it is more than a knack — and then takes it somewhere you will not like. Its answer is that wisdom is neither a mysterious perception nor a mere skill. It is a kind of rationality — a deep one. The leading version is Sharon Ryan's, in her Deep Rationality Theory, and the idea, shared in different forms by Robert Nozick and by Valerie Tiberius, runs like this. To be wise is to have a steady, reliable grasp of what actually matters in a life, and to form your beliefs and your choices and your values well in the light of it — coherently, in response to reasons, across the things that count. Nozick, in his book The Examined Life, put the core of it plainly: wisdom is understanding what is important, and how to respond to it. Not facts about the world. Understanding what is worth it, and acting on that understanding.
That probably sounds agreeable. Here is the cost, and it runs in two directions at once.
If wisdom is deep rationality — a matter of good reasons, coherent values, a record of choosing well — then for the first time in three weeks you can check whether you have it. Not perfectly. But genuinely. You can look at your own life and ask: are my values actually coherent, or do I want contradictory things? Do my choices track what I say matters, or do they track what is easy? When I have been tested, what is my record? On this account wisdom is not sealed away where you can never see it. It can be assessed from the inside, fallibly, by exactly the kind of self-examination you are able to do. After two weeks of being told that wisdom might be something you can never confirm you have, here is an account that says you can confirm it. You can find out. That is the reassurance you have been wanting — a way to know.
And here is the price, in the same breath. Read back what wisdom has just become. Good reasons. Coherent values. Choices that track what matters. A record you can point to. Now ask the question we keep reaching from new directions: is that the kind of thing only a human could have? A system that holds consistent values, that responds to reasons, that can set out why it chose what it chose and show a record of choosing it — that system is not obviously shut out of this. It is a candidate. The very feature that lets you check your own wisdom — that it shows up in reasons and records, out in the open where they can be examined — is exactly what would let a machine present its reasons and its record too. The account that finally lets you confirm your own wisdom is the account that cannot keep the machine out, because it has placed wisdom in precisely the things a machine can have.
So look at the shape of where we stand, and do not look away from it, because this shape will keep returning. The accounts that keep the machine out — the perception of the particular, last week's relief — are the ones that will not confirm your own wisdom, because they place it in something private and untestable. And the account that would confirm your wisdom — wisdom as a rationality you can check — is the one that cannot keep the machine out, because it places wisdom in something public and checkable. You can have the machine kept out, or you can have yourself confirmed. So far tonight, you cannot have both.
So that is the third night. Here is what it asks of you, and it asks more than last week did.
A week ago you left holding a relief: wisdom is the perception of the particular, so the machine is not in the contest, and the one human thing is yours. Tonight that relief is in real trouble — and notice that it is not because anyone proved the machine is wise. It is because two serious accounts of wisdom say the thing you were relying on is not what you thought. The skill account says your special seeing is what expertise looks like from the outside — built from practice, teachable, in principle buildable into something that is not you. The deep-rationality account says wisdom is good reasoning about what matters — reasons and a record, out in the open, the kind of thing a system could present too. Both describe wisdom as a competence — something you build and something you can check. And a competence, unlike a secret, does not depend on whose it is. What you gained is the opposite of comfort. The machine question you thought Aristotle had closed was never closed. No one had checked, and tonight we checked, and it had been open the whole time.
And there is the strange shape underneath it, the one to carry with you: the account that keeps the machine out will not confirm your wisdom, and the account that confirms your wisdom will not keep the machine out. Hold those two facts side by side. We are going to need them.
This is the third of eight, and we have moved somewhere real. We are not collecting theories of wisdom for a catalogue. Each one we meet costs you something different, and tonight's two cost you the easy comfort you walked in with. Next time, someone refuses the whole frame we have been working in. Both of tonight's accounts assumed wisdom is a competence — a skill, a rationality, something you are good at. The next voice says no: wisdom is not a competence at all. It is not a thing you are good at. It is knowing that you do not know — a stance toward your own ignorance. I will not make that case tonight; that is next time's work. But you can already sense the idea behind it, and how strange it is.
Take this with you in the meantime. Go back to the thing you are expert at — the seeing you do without rules. And this week, watch for the leftover. Watch for the moment in your own expert judgment where you catch something that does not feel like skill paying out, does not feel like a reason you could write down or a record you could point to — something else, that you cannot quite name and cannot quite locate. Perhaps you will find it, and Aristotle's remainder is real. Perhaps you will look and look and find that it is skill and nothing more, and the machine has as good a claim to it as you do. Either way, you will have done the thing no one can do for you: looked for the leftover in your own judgment, and felt how hard it is to say whether there is anything there but skill.
Thanks for listening. I'll see you next time.