Popper's Demarcation Criterion

In this paper I shall venture into an area with which I am not very familiar and in which I feel far from confident; namely into phenomenology. My main motive is not to get away from standard, boring, methodological questions like those of induction and demarcation; but the conviction that a phenomenological account of the empirical basis forms a necessary complement to Popper's falsificationism. According to the latter, a scientific theory is a synthetic and universal, hence unverifiable proposition. In fact, in order to be technologically useful, a scientific hypothesis must refer to future states-of-affairs; it ought therefore to remain unverified. But in order to be empirical, a theory must bear some kind of relation to factual statements. According to Popper, such a relation can only be one of potential conflict. Thus a theory T will be termed scientific if and only if T is logically incompatible with a so-called basic statement b, where b is both empirically verifiable and empirically falsifiable. (We shall see that neither the verifiability nor the falsifiability of b was meant, by Popper, in any literal sense.) In other words: T is scientific if it entails ⌝b; where b, hence also ⌝b, is an empirically decidable proposition.

This demarcation criterion which provides the best explication of the scientific character of theories, not to say an essential definition of empirical science, has been subjected to two very different types of criticism; the first is levelled at its theoretical and the second at its empirical aspect. Let us start with the first kind of criticism which is often referred to as the Duhem–Quine problem.