Post

Science basically works by dis-proof. Theories either match empirical evidence or they don't. A theory may stand up to the accumulated evidence for years. But when you get data that does not fit the current theory, you need a new theory. All knowledge is contingent on the next experiment.

When a theory is consistent with empirical evidence it is regarded as a good theory. Simple theories are preferred to complex concepts. Experiments are constructed to test theories to destruction. If evidence inconsistent with a theory accumulates then a point comes when you need a new theory.