Popper's Falsificationism

Core Idea

Karl Popper proposed that science be demarcated by falsifiability: a theory is scientific if there exist possible observations that would prove it false. Instead of seeking confirmation, Popper advocated bold conjecture and criticism: propose ambitious theories, vigorously attempt to refute them, and only accept provisionally until falsified. This avoids logical problems of induction while preserving empiricism. Popper argued that scientific progress comes through elimination of false theories, not accumulation of confirmed instances.

How It's Best Learned

Compare falsifiability with verifiability on examples like 'God exists', 'the sun will rise tomorrow', quantum mechanics, and psychoanalysis. Examine why some unfalsifiable theories are considered unscientific.

Explainer

You already know Hume's problem of induction: no finite number of observations, however large, can logically justify a universal generalization. We've seen a million white swans, but the next swan could be black — logic gives us no guarantee. This seems to undermine science at its foundations, since science depends on universal laws derived from finite observation. Popper's falsificationism is, in part, a bold response to this problem: instead of solving induction, he proposed abandoning it as the basis for scientific method.

Popper's key insight is logical asymmetry: while a thousand confirmations can never prove a universal theory, a single counterexample can refute it. If Newton's theory predicts that planets move in ellipses, and one planet is observed to deviate, the theory is falsified — definitively, by deductive logic (modus tollens: if theory T, then observation O; not-O; therefore not-T). Popper concluded that the proper aim of science is not to confirm theories but to conjecture boldly and attempt falsification. Science advances by eliminating false theories through critical testing, not by accumulating confirmations.

This gives Popper his demarcation criterion: a theory is scientific if and only if it is falsifiable — if there exist possible observations that would contradict it. Unfalsifiable claims aren't necessarily wrong or meaningless, but they aren't scientific. By this criterion, Freudian psychoanalysis and Adlerian psychology fail: they can accommodate any observation. A patient who resists treatment shows repression; a patient who responds well shows the therapy worked. No possible observation rules out the theory. In contrast, Einstein's general relativity made the bold prediction that starlight would bend around the sun by a specific angle — a claim that could easily have been refuted, and which therefore counts as genuinely scientific.

The concept of corroboration replaces confirmation in Popper's framework. A theory isn't verified when it survives testing — it is *corroborated*, meaning it has survived attempts to falsify it so far. Corroboration is not inductive support; it says nothing about the probability that the theory is true. We accept theories provisionally, as our best current conjectures, knowing they may be overturned. Science is not a body of proven truths but a self-correcting process of conjecture and criticism.

The central objection is the Duhem-Quine problem: theories are never tested in isolation. Any prediction requires auxiliary assumptions (about instruments, initial conditions, background theories). When an observation conflicts with a prediction, you can always save the theory by revising an auxiliary assumption instead of rejecting the core theory. Popper acknowledged this but argued that such "immunizing stratagems" are methodologically forbidden — a scientist must commit in advance to what would count as falsification. Critics like Kuhn and Lakatos argued that actual scientific practice is far more complex, and that the history of science shows productive research programs surviving apparent falsifications through exactly the moves Popper prohibits.