Adam Grobler

In the paper I explore some hints one can find in an updated version of Ajdukiewicz’s radical conventionalism that may help to resolve some controversies within hinge epistemology, i.e. a family of positions that invoke Wittgenstein’s idea of groundless grounds of knowledge. In particular I put into doubt whether there is a real difference between pragmatic and transcendental justification of hinges, I reject epistemological disjunctivism, and I argue for anti-realistic reading of truth in a context determined by particular hinges while retaining the classical notion of truth as a regulative idea of knowledge that governs the possible revisions of hinges.

In the paper I present my path towards the sandwich theory of knowledge, which I have been travelling for about three decades. It started from a modest remark by R. Wójcicki who drew my attention to the concept of presupposition. This concept played an important role in my analyses of the growth of science, which I have been occupied with for much of my career. At that time I was also impressed by A. Wiśniewski’s erotetic logic. Later, I took more interest in epistemology, first for teaching, subsequently also for study. I point also to other sources of inspiration: Ajdukiewicz, Hintikka, Popper, Putnam, and Wittgenstein.

To prevent possible misunderstanding, let me start from one crucial distinction between the logic of science on the one hand and the use of logic in science on the other. The very phrase “inference to the best theory” and the style of its further explication may suggest that it is the former that is at issue. In contrast, the title of the present workshop, “Induction and deduction in the sciences”, suggests the latter to be its focus. And this I find to point to the more adequate way of thinking about the relation between logic and science. I do not believe in “the logic of science” as something that can possibly be read off science in a manner analogous to that of reading logic off mathematics: the classical logic off the classical mathematics or the intuitionistic logic off the intuitionistic mathematics

To prevent possible misunderstanding, let me start from one crucial distinction between the logic of science on the one hand and the use of logic in science on the other. The very phrase “inference to the best theory” and the style of its further explication may suggest that it is the former that is at issue. In contrast, the title of the present workshop, “Induction and deduction in the sciences”, suggests the latter to be its focus. And this I find to point to the more adequate way of thinking about the relation between logic and science. I do not believe in “the logic of science” as something that can possibly be read off science in a manner analogous to that of reading logic off mathematics: the classical logic off the classical mathematics or the intuitionistic logic off the intuitionistic mathematics.

It is claimed that Kuipers' approach to explanation opens the possibility for a further refinement of his own refined HD method for the evaluation of theories. One severe problem for the HD method, refined or not, is theory-ladeness. Given that experimental results are theory-laden, the comparative evaluation of alternative hypotheses is always relative to background knowledge. This difficulty can be avoided by supplementing HD considerations with the principle of inference to the best explanation. The authors sketch a program for doing this. The general idea plays on some similarities between Kuipers' account of explanation and Lipton's. The former, however, is considered more flexible than the latter, which makes it even more attractive for the purpose under consideration.

In the early sixties there broke out a fierce controversy concerning rationality in science which was labelled as the Popper-Kuhn controversy. It can be conceived in terms of the rationalism-relativism opposition. This may seem dubious, for the proper contrast to rationalism is irrationalism, and the one to relativism is absolutism. What is at issue, however, is whether scientific change comes about in consequence of argument or in consequence of-to use Kuhn's favourite dictum-conversion. The notion of argument does not involve here anything absolute, while the one of conversion is intended to cover not merely the idea of irrationality of change, but the idea of self-authentication and constitution as one among many possible patterns of rationality. Thus the main message is that of break of communication between otherwise rational agents and not of their apparent irrationality. These are my reasons for thinking this unusual opposition better than the usual ones pointing at the substance of the controversy. I think also this opposition has in fact been implicit in the resulting debate. This explains, for instance, why in the first decade of the debate Feyerabend was commonly perceived as Kuhn's comrade and, consequently, largely misunderstood. Conflation of two different oppositions into one, on the other hand, results in some obscurity. This is perhaps why different commentators have rather freely qualified some participants of the debate to one or other faction. Among those who have been subjected to such a treatment there are found, in the first place, the authors of such mediational conceptions of the seventies as Lakatos (methodology of scientific research programmes), Sneed and Stegmiiller (non-statement view), Laudan (research traditions), and, in my country, Amsterdamski (ideals of science). It may well have been expected that Larry Laudan, who has recently put forward a new mediational conception of his own (see Laudan [1984]), again will be treated in the same way. This is due to the clear Popperian and Kuhnian inspirations of his work and his fidelity to his masters which is surprisingly great consider