Modality in Slavic and semantic maps Ferdinand de Haan 1. Introduction The most important issue facing the study of modality in the 21st century is to agree on a comprehensive framework that can deal with synchronic and diachronic issues. 1 Many frameworks have been proposed, ranging from the very formal to the very ad hoc, but all suffer from insurmountable problems. Most simply do not deal with the wide range of data and the equally wide range of meanings that modal elements in natural languages have to offer. In many cases, attention is focused on just the so-called core-meanings of the modals (such as necessity or permission) and more peripheral functions (such as finer shades of interpretation of the core functions) are disregarded or subsumed under the core functions. This can lead in some cases to confusion, disagreement or misunderstanding about the nature of certain grammatical categories. A case in point is the distinction between epistemic modality and evidentiality, a topic that is currently being hotly debated in the literature 2 and which will be discussed in section 4.3 of the present paper. This paper is meant to contribute to the theory of semantic maps, visual representations of the semantic range of linguistic material. It is argued that semantic maps are very useful for describing the semantic range of modal elements that occur in natural languages. The paper builds on the pioneering work done in this area by Anderson (1986) and, more recently, by Van der Auwera and Plungian (1998) and Hansen (2001). The focus is on modal expressions in Slavic, but other languages will be taken into account as well. The level of detail in studies such as Hansen (2001) and in various papers presented at the Regensburg Workshop allow us to start working on the modal map of the Slavic languages and how it compares to that of other language 1 I am very grateful to the organizers and participants of the Regensburg Workshop on Modality in Slavic for their invitation and their thoughtful discussions, especially to Björn Hansen, Johan van der Auwera, and Ruprecht von Waldenfels. None of them is responsible for any errors remaining. 2 See among others Palmer (1986 [2001]), Willett (1988), De Haan (1999, 2005a), Aikhenvald and Dixon (2003).