Abstract

In this paper I try to explain why Lakatos’s conventionalist view must be replaced by a phenomenological conception of the empirical basis; for only in this way can one make sense of the theses that the hard core of an RP can be shielded against refutations; that this metaphysical hard core can be turned into a set of guidelines or, alternatively, into a set of heuristic metaprinciples governing the development of an RP; and that a distinction can legitimately be made between novel predictions and facts to which a theory might have been adjusted post hoc. Two basic metaprinciples are finally examined: the Correspondence Principle and various symmetry requirements; both of these heuristic devices illustrate the fundamental role which, according to Lakatos, mathematics plays in the progress of empirical science.