Definite expressions have been studied extensively in the semantics and pragmatics literature, with different division lines drawn between what goes in the lexical meaning of the expressions and what is derived from other conversational mechanisms. In this talk, I explore what empirical grounds we can cover by minimizing the semantics and incorporating language general mechanisms. Specifically, I argue that the semantic space of definiteness should be divided along two dimensions, the content of the linguistic description and the presence of a linker to the actual world. This division is assumed to be universal, though languages morphosyntactically realize the space in different ways. I explore how this view can derive observed anti-uniqueness effects of demonstratives and different presuppositions of definite descriptions.