Falsifiability

Karl Popper proposed falsifiability as the demarcation criterion separating empirical science from non-science: a scientific theory makes a claim about reality that could be contradicted by observation. A bold theory forbids many possible outcomes, exposing itself to refutation; a theory compatible with every outcome says nothing testable. This grows out of Critical Rationalism‘s logical asymmetry — one genuine counter-example can refute a universal claim, while no quantity of confirming instances can prove it. So evidence works only by refuting rival hypotheses, never by supporting a favored one; there is evidence against, not evidence for. This is why CR rejects induction and justificationism.

CF accepts falsifiability as far as it goes but treats it as a special case of a broader principle. Scientific testing is just one important form of criticism; explanatory and other non-empirical criticism matters even within science. Where Popper limited refutability talk mostly to empirical theories, CF applies binary refutation to all intellectual judgment through its Yes or No approach: every idea is either refuted or non-refuted relative to a goal.

A standing worry is that any refuted theory can be rescued by tacking on ad hoc auxiliary hypotheses so it no longer conflicts with the data. CF’s answer is that such patches wreck the explanation and can be attacked with epistemological criticism, not only topical evidence. CF also drops Popper’s notions of corroboration and degrees of survival under severe tests, denying that criticisms or theories come in strengths.