The conceptual world of AI is inhabited by a number of epistemic puzzles whose role is to provide a test harness environment for various methods and algorithms. In our paper we focus on those puzzles in which agents either collaborate or compete with one another in order to adopt their epistemological situations to their environment. Our goal is to devise a formal model for epistemic interactions and a family of reasoning mechanisms that would solve those puzzles. Once specified in ract manner, they are implemented in the Prolog environment. Key words: knowledge representation, ontology, epistemic change