Karl Popper's Epistemology

A theory is scientific not because it can be proved, but because it can be broken — and the same idea, taken seriously, redraws the case for free societies.

Most people think science works by accumulating evidence. Theories that have racked up enough confirmations get promoted to facts. Karl Popper spent fifty years arguing the opposite, with a stubbornness that makes more sense once you realise what was at stake. Theories cannot be confirmed by observations — not by any finite number of them, not in principle. They can only be tentatively held, and only until something breaks them. The job of a scientist is not to prove. It is to try, with as much ingenuity as possible, to refute. This sounds like a technical correction. It quietly rearranges everything.

The problem of induction

Popper started where David Hume had stopped two centuries earlier. Hume had pointed out a puzzle that everyone notices and most people brush past. We generalise from the past to the future. The sun rose every day for as long as anyone can remember, so we expect it tomorrow. We have seen many white swans, so we believe all swans are white. But there is no logical step from “every observed X has been Y” to “every X is Y.” The future is not entailed by the past. To believe it is, you would have to assume that nature is uniform — and that very assumption is what you were trying to justify.

Hume thought induction was a psychological habit, not a logical inference. Popper took him at his word. If induction does not work, then no number of confirmations ever justifies a universal claim. A million white swans do not prove that all swans are white. A single black swan, however — and Australia did indeed have them — refutes it definitively.

This asymmetry is the hinge on which everything turns. Confirmation is hopeless; refutation is conclusive. So the question shifts. If we cannot prove our theories, what can we do with them? Popper's answer was to flip the project of science around. We do not seek confirmation. We seek the most demanding tests we can devise — tests that would expose our theories if they were wrong — and we keep the theories that survive.

Falsifiability

The criterion that emerges, Popper said, is falsifiability. A theory is scientific if and only if it forbids something. It must stick its neck out. It must say: this could happen, but it will not, and if it does, I am wrong.

His paradigm case was Einstein's general relativity. Einstein predicted that starlight passing near the sun would be deflected by a precise, measurable amount during a solar eclipse. Had Eddington's 1919 observations shown no deflection, the theory was finished. The very fact that Einstein's equations forbade certain outcomes — outcomes entirely possible to observe — was what gave the theory its scientific bite. It was a brave theory. It risked itself.

“The criterion of the scientific status of a theory is its falsifiability, or refutability, or testability.”

Compare this to theories that explain everything. As a young man in 1919 Vienna, Popper noticed how readily Adlerian psychology, certain political theories, and pop-Freudianism absorbed every conceivable piece of evidence. Whatever a person did, the theory could narrate it as confirmation. A man saves a child? He is overcoming an inferiority complex. He drowns a child? Same complex, opposite direction. No human action could refute the theory because the theory had been made unfalsifiable. To its admirers this looked like explanatory power. To Popper it looked like the precise opposite — vacuity dressed up in jargon.

This is the core insight. The strength of a theory lies in what it forbids, not in what it accommodates. Astrology, told by a clever interpreter, can fit any planetary configuration to any biography. That flexibility, which feels like a virtue, is the disease. Real theories are inflexible. They cut. They make predictions that could embarrass them. Falsifiability is not a property of where a theory came from, or how plausible it sounds, or whether it uses mathematics. It is a property of its relationship to possible evidence: does the world have a way to tell us no?

Conjecture and refutation

Once the impossibility of confirmation is accepted, the entire shape of inquiry changes. Knowledge does not start from neutral observation. It starts from a problem and a guess.

Popper described the cycle in four steps. We begin with a problem, P₁ — something we do not understand or that an old theory failed to handle. We invent a tentative theory, TT — a bold conjecture, often wildly speculative, that would solve it. We then subject the theory to error elimination, EE — ruthless attempts at refutation, by experiment, by argument, by working out implications and checking them. Whatever survives is kept, but only provisionally. The survival itself defines a new problem, P₂, because the surviving theory has implications that need testing too.

Theories are never proved. The best we can say of one is that it has not yet been killed by serious attempts. Popper called this corroboration to distinguish it from confirmation. A theory that has resisted hard, varied, dangerous tests is more corroborated than one that has resisted easy ones — but it is not closer to certainty. The next test could still demolish it.

What this gives us is an evolutionary epistemology. Conjectures are mutations. Refutations are selection pressure. Knowledge grows by error-elimination, not by accumulation. The reason science makes progress, on this view, is not that it gets things right but that it gets things less wrong, faster, by exposing itself to harder tests than any other tradition has ever willingly accepted.

The open society

Popper noticed that the same structure governs politics. He worked it out during the second world war, in exile in New Zealand, in The Open Society and Its Enemies.

A closed society treats its ruling theory — a religion, a historical destiny, a class struggle, a divine right — as unfalsifiable. It has no procedure for noticing when the theory is wrong. Errors compound. When reality finally forces a confrontation, the system breaks rather than bends.

An open society is the political form of fallibilism. It allows criticism. It permits dissent. Above all, it builds institutions through which leaders can be replaced without violence — which is to say, institutions through which the polity can admit that it was wrong. Popper put it bluntly: the central question of politics is not “Who should rule?” but “How can we arrange things so that bad rulers can be removed without bloodshed?” Democracy's deepest virtue is not that it produces wise leaders. It is that it incorporates an error-correction mechanism.

The connection between Popper's epistemology and his politics is not metaphor. It is the same idea twice. Free criticism is not a luxury. It is the only way any system, of knowledge or of governance, escapes its own mistakes.

Popper's gift was to take a small piece of logic — the asymmetry between proof and refutation — and follow it relentlessly until it had reshaped the philosophy of science, the structure of inquiry, and the case for liberal politics. The result is a stance toward our own beliefs that is at once humble and combative. Hold them tentatively. Test them mercilessly. Replace them when they fail. Keep guessing.