Refutation

Popper grew up in post-1918 Vienna, where Einstein’s relativity, Freud’s and Adler’s psychoanalysis and Marxism all claimed scientific standing. He was struck that the 1919 eclipse observations could have refuted Einstein, while the others seemed able to explain any outcome. He argued against the Vienna Circle’s verificationism, left for New Zealand in 1937 ahead of the Nazi annexation of Austria, wrote The Open Society there during the war, and taught at the London School of Economics from 1946.

• 1934 · Logik der Forschung

  Falsifiability

  A theory is scientific to the extent that it rules out possible observations and so could be shown false. Confirmations are cheap if you look for them; bold predictions that risk failure are what count.

• 1934 · Logik der Forschung

  No rescue by patching

  A theory that meets a refuting result should not be saved by assumptions invented only to protect it, because each such rescue reduces what the theory says.

• 1945 · The Open Society and Its Enemies

  Against prophecy, for piecemeal reform

  Claims to know the course of history are unscientific. Policy should proceed by small, correctable steps, and institutions should let bad rulers be removed without bloodshed.

• 1963 · Conjectures and Refutations

  Conjectures and refutations

  Knowledge grows by bold guesses exposed to the strongest attempts at refutation, not by piling up confirmations.

1. What conceivable observation would count against this claim?
2. What result would show this is wrong, and has anyone tried to get it?
3. Does the claim forbid anything, or could it absorb every possible outcome?
4. When it failed before, was it revised honestly or rescued by an assumption added after the fact?

A failed prediction never refutes a theory on its own: the fault may lie in an auxiliary assumption or the measurement, and Kuhn and Lakatos argued that scientists reasonably keep theories through anomalies.