The semantics-pragmatics distinction is of great importance within both the philosophy of language and linguistics, but the question of how to really draw this distinction remains disputed. Part of the dispute rests on differing intuitions about the extent to which the context of utterance plays a role in determining the semantic contribution of a given expression. Another part of it stems from the difficulty of accounting for communicative phenomena that appear not to fit neatly on either side of the distinction. In this theoretical context, the present thesis argues for four conclusions: (i) That the currently available empirical evidence on polysemy processing and resolution fails to support a one-representation model of the meanings associated with polysemous expressions. (ii) That the inherent polysemy exhibited by some proper names puts pressure on Kripke’s two main theses about the semantics of proper names. (iii) That the social significance of a slang word is analyzable in terms of the lexical metadata related to it, and the inferences people competent with the word will be disposed to draw on the basis of that metadata. (iv) That slang words are best understood as words that are nonstandard, nontechnical, informal, and part of the in-group language of a social group. These conclusions are of wider importance. Establishing (i) and (ii) mandates rethinking the significance of polysemy vis-à-vis the semantics-pragmatics boundary: Contrary to a widely held view, the phenomenon fails to vindicate any particular way of drawing the semantics-pragmatics boundary and yet, polysemy continues to have bearing on semantic theorizing. Establishing (iii) and (iv) constitutes a major step towards clarifying how the phenomenon of social meaning, which is not easily placed on the semantics-pragmatics boundary, can be integrated into the philosophy of language.