The semantics of definite descriptions has been a central topic in philosophy of language ever since Russell’s landmark paper ‘On Denoting’(1905). Russell argued that definite descriptions should not be seen as referential expressions, but instead as quantificational expressions. In other words, a sentential utterance containing a definite description should be understood as expressing a general/object-independent proposition. A problem arises with the view once we consider the fact that definite descriptions are used frequently and consistently to refer to particular individuals. Through this observation, Donnellan (1966) argued that definite descriptions would be better understood as having two distinct uses, one referential and one attributive/quantificational. We can call this the ambiguity problem in definite descriptions. The following thesis will argue that the ambiguity problem disappears once we take seriously the grammar that underpins sentential utterances containing definite descriptions, and that the semantics of the definite article is determined in part by the grammatical topology of determiner phrases and in part by what grammatical environments determiner phrases can be felicitously placed. The thesis, therefore, is that the semantics of definite descriptions is grounded in grammatical facts.