I. The Poker

October 25, 1946. King's College, London. The Cambridge Moral Sciences Club.

Popper has been invited to give a talk. Wittgenstein is chairing. Russell is in the audience. This is the most famous room in twentieth-century philosophy.

What exactly happened that evening depends on whom you ask. But one detail appears in every account: Wittgenstein picks up a fireplace poker, waves it, and argues with Popper. The argument: whether genuine philosophical problems exist, or whether philosophy is merely a confusion of language. Wittgenstein says no real problems. Popper says yes. Popper gives examples. Wittgenstein drops the poker and walks out.

Russell, sitting there, reportedly smiles.

The poker story has been retold endlessly, each version different. But the scene matters more than the details: the two most combative philosophers of the twentieth century, in one room, arguing, one of them holding a fireplace poker. Popper spent his entire life arguing. With Wittgenstein. With the Vienna Circle. With Kuhn. With Adorno. His student Feyerabend later said Popper's seminars were not discussions — they were trials. Popper was the judge. Everyone else was the defendant.

Why so combative?

Because he believed one thing: your theory must be capable of being overthrown. If your theory cannot possibly be overthrown, it says nothing. You think you're saying something profound. You are saying something that can never be wrong — which means it can never be right either.

That sounds like common sense. But it reshaped the standard of knowledge for an entire century.

II. Vienna, and the Circle He Didn't Join

Popper was born in Vienna in 1902. Jewish family, later converted to Lutheranism. He grew up in the most intellectually dense city in Europe. Freud was there. Wittgenstein came from there. Mach's influence was still in the air.

In the 1920s, the Vienna Circle formed. Schlick, Carnap, Neurath. Their program was called logical positivism, and its core claim was simple: a proposition is meaningful if and only if it can be verified by experience. Propositions that cannot be verified — about God, about morality, about metaphysics — are not wrong. They are meaningless. They don't even qualify as wrong.

This was Comte's positivism upgraded. Comte said the unobservable doesn't count as knowledge. The Vienna Circle said the unverifiable doesn't even count as meaningful. Posterior colonizing prior, colonized all the way down to the root of language.

Popper was close to the Vienna Circle but never a member. He watched from the edge. He thought they had gotten one crucial thing wrong.

The problem with verificationism? Popper gave an example: all swans are white. You've seen a thousand white swans, ten thousand white swans. Can you say "all swans are white" has been verified? No. Because the next one might be black. You can never verify a universal claim through finite observation.

But you can overthrow it with one black swan.

What verification cannot do, falsification can. You cannot prove a theory right, but you can prove it wrong. A good scientific theory is not a verified theory — it is a theory that has not yet been overthrown, and stands ready to be overthrown at any moment.

1. The Logic of Scientific Discovery (Logik der Forschung). This book turned verificationism on its head.

III. Conjecture First

Comte said: start from observation. You observe the world, extract regularities, and regularities become knowledge. Posterior to knowledge, a straight line.

Popper said: wrong. Science never starts from observation. You cannot observe without a prior question. Observe what? Record what? Ignore what? These choices presuppose a framework. Observation without theory is blind — you see everything and perceive nothing. Like the Aleph.

Science starts from conjecture. You first propose a bold hypothesis — prior, unverified, possibly wrong. Then you try to destroy it. You design experiments specifically to find evidence that would kill the hypothesis. If you find it, the hypothesis dies. If you don't, it survives for now — but it can die at any time.

This is Popper's "conjectures and refutations." Science is not a process of accumulating observations into knowledge. Science is a process where conjecture (the prior) submits to trial (the posterior). The posterior is not the starting point. The posterior is the judge. The prior is back in the driver's seat.

Comte's posterior-colonizing-prior was completely reversed.

This is not a small adjustment. Think about what it means: if science starts from conjecture, then the scientist's most important ability is not observation but the capacity to produce good conjectures. What makes a good conjecture? Popper says: the bolder the better, the more easily overthrown the better. A theory that can be destroyed in a thousand ways is worth more than a theory that can never be destroyed at all. Because the former says a great deal (and so can be tested), while the latter says nothing (and so can never be wrong).

Einstein's general relativity predicted that light bends near the sun. The 1919 solar eclipse observation confirmed the prediction. Popper was deeply moved — not because Einstein was right, but because Einstein could have been wrong. If the light hadn't bent, general relativity was dead. Einstein staked everything. That is science.

What about Freud? Popper said: bring any patient, Freud can explain. Bring the opposite patient, Freud can explain that too. Bring any patient whatsoever, Freud has an explanation. So what has Freud actually said? Nothing. Because a theory that explains everything excludes nothing. And what excludes nothing has no content.

What about Marx? Early Marxism made predictions — capitalism necessarily leads to proletarian revolution. That is testable. But when the predictions failed, Marxists didn't concede the theory was wrong. They added patches. "Conditions weren't ripe." "Imperialism delayed the contradictions." "Ideology blinded the working class." Each patch made the theory more "correct" but also more hollow. Eventually it explained everything. And explaining everything meant saying nothing.

IV. The Line He Drew

Popper drew a line. On this side: science. On the other side: not science.

The line is called the demarcation criterion. Falsifiable propositions are science. Unfalsifiable propositions are not science.

Note carefully what he did not say. He did not say "the unfalsifiable is wrong." He said "the unfalsifiable does not belong to science." Metaphysics is not wrong — it may be meaningful, may be valuable, but it is not science. Ethics is not wrong — but it is not science. Aesthetics is not wrong — but it is not science.

Popper himself was more open than the line suggests. His late "World 3" theory — that objective knowledge exists independently of any knower — is itself difficult to falsify. He acknowledged that metaphysical research programs can inspire science. He even conceded that evolutionary theory doesn't fully meet the falsifiability criterion in the strict sense.

But culture doesn't hear nuance. Culture hears slogans.

Popper drew a line. Culture turned that line into a wall. "Unfalsifiable" went from being a description to being a verdict. To call something "unfalsifiable" in academic circles became equivalent to saying "not worth taking seriously." Freud was out. Jung was out. Most of philosophy was out. Most of the humanities were out. Falsifiability shifted from a demarcation criterion to a value criterion — things on the wrong side of the line weren't just "not science," they were practically "not knowledge."

This is Comte's ghost returning. Popper reversed Comte's direction (the prior back in the driver's seat), but Popper's boundary resurrected Comte's posture (what falls outside doesn't count) in the culture. The prior can conjecture now, but only conjectures that produce falsifiable propositions are taken seriously. You are allowed to be bold, but your boldness must stay within the permitted territory.

The cracks were grouted again. The material changed — from Comte's concrete to Popper's epoxy resin. More sophisticated, more refined. Still grout.

V. His Own Cracks

Popper's system has a problem he could never resolve.

Where do scientific discoveries come from?

Popper says: conjectures. Bold conjectures. Fine. But where do conjectures come from? Popper essentially declines to answer. He says that belongs to "the psychology of discovery," not "the logic of discovery." He cares about how conjectures are tested after they're proposed, not how they emerge in the first place.

But that is precisely the most important question.

Where did Einstein's general relativity come from? Not from observation — no one observed spacetime curvature and then derived equations from it. Not from pure logic — you can't deduce it. It came from somewhere. From somewhere inside Einstein's mind. From an intuition about the structure of spacetime. From a thought experiment: what would I see if I rode on a beam of light?

Where did McClintock's discovery of transposable elements come from? She said it herself: a feeling for the organism. Not observation first. Feeling first, observation after.

Where did Yukawa's meson hypothesis come from? He said Zhuangzi influenced his physical intuition. A classical Chinese philosopher influencing a Japanese physicist's prediction of a particle. Where does that fit in Popper's system?

Popper says the source of conjectures doesn't matter; what matters is whether they can be tested. But if the source doesn't matter, you can never explain why some people produce brilliant conjectures and most don't. You can't explain scientific creativity. You can't explain how scientists actually work.

Kuhn drove a wedge into this crack. Kuhn said: scientists don't spend every day trying to overthrow theories. Scientists spend most of their time in "normal science" — solving puzzles within a paradigm. Paradigms aren't overthrown by a single counterexample (Popper's account); they are replaced by a new paradigm when anomalies accumulate past a tipping point. And the acceptance of a new paradigm isn't a purely rational process — it involves generational turnover, aesthetic preference, power dynamics.

Feyerabend went further. He said: look at the history of science and you'll find that no methodological rule has ever been consistently followed. Galileo violated every methodological standard of his time. Science advances not by following rules but by breaking them. "Anything goes" — this phrase made Popper furious.

Popper argued with Kuhn. Argued with Feyerabend. He argued his whole life. But he couldn't argue away a fact: what scientists actually do is not what Popper says they should do. Conjectures don't fall from the sky. They come from intuition, from aesthetic sense, from things that cannot be articulated. And those inarticulate things are precisely what Popper's line places on the other side.

VI. Popper and Comte

The previous essay described Comte's posterior-colonizing-prior. Popper reversed this. The prior is back in the driver's seat. That reversal is real.

But what happens after the reversal?

Comte said: only the observable counts as knowledge. Beyond the boundary lies immaturity. Popper said: only the falsifiable counts as science. Beyond the boundary lies non-science.

Two lines drawn at different positions. But drawn the same way: take a pen, draw a line on the map of cognition, and declare what lies beyond doesn't count. Comte drew with observation. Popper drew with falsification. Both stood on their side and said: over here, it's clean.

Where Popper surpasses Comte: he admitted there are things on the other side. Comte said nothing was there (the cracks are an illusion). Popper said metaphysics, ethics, aesthetics are there — they may be meaningful, they just don't belong to science. This is a significant difference. Popper didn't grout the cracks. He left them open. He only said: science doesn't work in the cracks.

But culture turned his "science doesn't work there" into "nobody should work there." The ghost of posterior-colonizing-prior returned — not Comte's version (posterior directly denying the prior) but Popper's version (the prior may exist, but only the prior that submits to posterior judgment has value). The prior came back, but domesticated. You may conjecture, but your conjectures must be testable. Untestable conjectures aren't illegal; they're just ignored.

From Comte to Popper, the treatment of the cracks was upgraded — from "don't exist" to "exist but aren't science." In practice, though, "isn't science" has roughly the same effect as "doesn't exist."

VII. Where He Stands in This Round

Comte is the target. Popper is the upgraded target.

Comte grouted the cracks — declared them nonexistent. Popper didn't grout them — acknowledged they're there, but drew a line between the cracks and science. The line itself is far thinner than Comte's concrete. But culture thickened it.

Starting from the third essay, the figures in this round will no longer be "opposing Comte" or "opposing Popper." They will be speaking from inside the cracks. Dirac isn't opposing anyone. He is simply working equations, and the equations tell him there is antimatter on the other side. Qu Yuan isn't refuting any epistemology. He is simply writing poetry, and the poetry itself is a form of cognition. Fichte isn't debating Popper. He is simply asking a question Popper never asked: where is the knower herself?

Popper's system can accommodate Einstein. It cannot accommodate Qu Yuan. It cannot accommodate McClintock's feeling for the organism. It cannot accommodate Yukawa's physical intuition drawn from Zhuangzi. It cannot accommodate Polanyi's tacit knowing — the things you know but cannot say.

And that is precisely where this round ends. Polanyi will say: what you cannot articulate is not the residue of knowledge. It is the foundation. Everything you can say — including Popper's entire system — is built on things you cannot.

From Comte's "what you can't say doesn't count," to Popper's "what you can't say isn't science," to Polanyi's "what you can't say is the foundation." Three steps. Comte stands on the first. Popper stands on the second. He took a giant step — reversed the direction, acknowledged the cracks exist. But he stopped at the second step.

VIII. The Bridgehead

Popper arrives fast. He doesn't walk — he charges. He carries nothing in his hands, but his whole person is a fireplace poker: rigid, hot, ready to prod.

He spots Comte immediately. Comte is there, spreading his calendar on the railing. Popper walks over. Glances at it.

"Is Freud in your calendar?"

Comte doesn't respond. Comte doesn't argue with people. Comte argues with the world.

Popper doesn't care. He turns to face everyone on the bridge. "Conjecture first. Observation second. Theory doesn't grow out of experience — you propose it yourself, then you let the world hit you in the face. If the world can't hit you in the face, your theory is empty."

Someone laughs. Possibly Hume. Hume always thought causation was habit rather than necessity — Popper's conjectures-and-refutations has a family resemblance to Hume's skepticism.

Kant stands at a distance. Popper's relationship with Kant is complicated. He inherited Kant's caution about the limits of reason, but he wouldn't accept Kant's a priori categories — he thought those should be overthrowable too.

Popper finds a position and stands. Next to Comte, but facing the opposite direction. Comte looks at the bridge surface. Popper looks toward both ends. He knows there are things at both ends. He just doesn't think science should go there.

Night deepens. The bridge grows quiet. Popper is still standing. He is waiting. For what? For someone to refute him. If nobody refutes him, he will grow uneasy. A theory that goes unrefuted is a theory without content. A person who goes unrefuted is the same.

From somewhere far away, a sound reaches the bridge. Not speech. Something more like a poem. Or an equation. Or a feeling for an organism.

Popper hears it. He frowns.

The sound comes from the cracks. From the other side of the line he drew. He knows something is there. He just doesn't know what to call it.^[1][2]

Han Qin (秦汉) · @hqinjarsy · Great Lives