The project investigates the relation between formal representations of contexts, and the real-life contexts they purport to represent. It is widely acknowledged that such representations are needed in formal semantics for natural languages in order to handle context sensitive expressions, but the question of what it takes for a formal context representation to count as representing a particular real-life context has to a large extent remained unanswered. The question is important, since it concerns the relation between theories of natural language and the phenomena they are supposed to explain. Recent work in philosophy, linguistics and cognitive psychology, and collaborations across these disciplines, have opened up new perspectives on this important issue, and the present investigation aims to take this development one step further. The key idea is that formal context representations should be taken to represent language users' Conceptions of Situations (COS), i.e. complex mental states that may consist of anything from simple first-order beliefs to complicated higher-order intentions. The notion of COS will be developed within a novel theoretical framework, which departs from philosophical tradition in content as well as methodology. The project incorporates ideas and methods from linguistics and psychology, as well as logic and philosophy, and is part of a more comprehensive inquiry into the role of these disciplines in the study of linguistic communication.