A. G. Benedek:

It has been apparent for some time now that traditional models of philosophy of science provide poor explanations for the emergence of social norms of information exchange in the knowledge based economy. The paper studies the underlying idea of learning in models of bounded rationality that may shed new light on traditional normative epistemological concepts of obtaining knowledge from the vantage point of game theory.

Modern philosophy of science and contemporary game theory share the view that rational agents cannot avoid accepting or rejecting certain norms, or rules in the process of obtaining information. Important results in both fields model paradigmatic types of interaction for the improvement of knowledge as social institutions. According to the shared, common view many, even if not all, of the real life encounters of learners can be analyzed in terms of models of rational choice in which agents use theories which comply with normative assumptions in order to look forward in competitive situations of social interactions. The assumptions come not only from the game theoretic framework but also from certain recalcitrant epistemological principles of traditional philosophy of science.

First I reconsider some of the classical choice theoretic assumptions from the point of view of learning theory and information transmission and argue that they are hardly tenable for the characteristic social interactions of the information society. Using some examples of "repeated knowledge games" and social situations where we "bargain for knowledge" I suggest alternative epistemological principles more suitable for the description of cooperative and non-cooperative games of obtaining information. In doing so I abandon the game theoretic assumption that preferences of learning agents are prefixed and independent of the obtained information. Finally the paper explores some consequences of imperfect information transmission for social conceptions of knowledge within the class of models of bounded rationality.