The project aims at systematically examining the relationship between three debates in three different disciplines: the problem of logical constants in philosophy of logic, the debate on the functional vs. lexical distinction in linguistics, and the debate on the semantics vs. pragmatics distinction in philosophy of language. Our hypothesis is that these seldom-connected debates actually address the same set of underlying issues. In particular, we will argue that to give an account of semantic content, and to provide a principled answer to the question of where to draw the line between semantics and pragmatics, is correlated, on the one hand, with identifying criteria that distinguish the logically valid inferences from other inference patterns, and, on the other, with an account of the lexical and functional properties of the expressions of a given language. We will further examine whether the distinction between logical vs. non-logical expressions maps onto the distinction between functional vs. lexical (or 'substantive') categories. If our hypotheses are correct, then not only will the language of logic and natural language turn out to be intimately connected, but we will also gain new insight into the distinction between function words and content words. While our main goal is to address conceptual issues, we will flesh out and test our hypotheses through three case studies, involving respectively indexical pronouns, prepositions, and degree modifiers. The latter two will allow us to enrich our dominantly theoretical approach with standard methods from corpus linguistics, as well as statistical methods, such as distribution modelling.