Abstract

In this paper, one aspect of Carnap's philosophy is examined, namely the relations of Carnap's various views and the rationality of theory acceptance in science. Irzik, Friedman and others have shown already that the so called “standard account” – presenting Carnap as an “arch rationalist”- is over-simplified. Friedman’s earlier view was criticized by Irzik to be too relativistic. I agree with these critiques; however I attempt to show that even Irzik’s and Friedman’s later view – which converge to each other - are not adequate. Their argument based on linguistical framework is examined, but several defects are shown. I point out that linguistical frameworks can be wider and narrower, which makes the conclusion invalid. Carnap’s view on theory acceptance is investigated and we find that Carnap accepted algorithmic evaluation of degree of confirmation, but rejected a binary theory choice. I argue that the reason for this is avoidance of information loss and not framework relativity. Irzik’s and Friedman’s term of “instrumental rationality” is analyzed, and some conceptual problems are indicated. I reason that Carnap’s conventionalism has to be regarded with keeping in mind his distinction of synthetical and analytical questions. Friedman’s hierarchy of frameworks, - which he constructed as an extension of Carnap’s and Kuhn’s theory - is discussed, and I propose another hierarchy of frameworks and give some arguments why it is more fruitful than Friedman’s.