"The scientific method — apart from a narrowly agnostic and pragmatist point of view — is therefore by itself incomplete and insufficient: it demands in order to make contact with reality the complement of some metaphysic or other."