Popper rejected the most characteristic doctrine of empiricism, in arguing (following Kant) that all descriptions of experiences involve selection and interpretation in terms of some prior conceptual framework, or theory. The model of scientific advance as an inductive process of generalization from particular experiences must therefore be rejected. Popper's alternative model is eloquently captured in the title of one of his booksConjectures and Refutations (1963). Scientific theories are invented, by a process which cannot be captured by any logical scheme. Once invented, their scientific status is established by their fruitfulness in allowing the deduction of hypotheses which are empirically contentful. By this, Popper means that they should be highly improbable (in the sense that they rule out as impossible many happenings which might otherwise seem possible), and at the same time be clear and unambiguous in specifying what they rule out. In Popper's version, the empirical testing of a theory is not a matter of finding evidence to support or confirm it, but rather a matter of systematically attempting to show it to be falsea logic of refutation or falsification. In this way Popper avoids the problem of induction which had bedevilled the attempt to justify science in terms of the idea of empirical verification. Popper's position is based on recognition of a very simple asymmetry between the logic of verification and that of falsification in relation to the law-like generalizations of science: universal claims always go beyond what is strictly justified by the (finite) body of evidence for them, but may be decisively refuted by a single counter-instance.
But the situation is more complex than this. Most especially, although the logic of falsification may be simple, its methodology is not. An observation which appears to challenge an established theory may itself be challenged as fraudulent, methodologically suspect, and so on, and will always leave advocates of the theory a range of choices to modify their theory, short of wholesale abandonment. Popper is fully aware of this, and is inclined to present falsificationism as a normative injunction, rather than as a description of the actual practice of scientists. Nevertheless, choice between rival theories is never an arbitrary matter. Although
all scientific knowledge must be considered provisional (there is no conclusive proof or disproof),
scientists properly prefer, among rival theories which are so far unfalsified and account for the known facts, that theory which has most empirical content.
[
falsification - reformatted for emphasis]
