This paper considers the connection between concepts, conceptual schemes and grammar in Wittgenstein’s last writings. It lists eight claims about concepts that one can garner from these writings. It then focuses on one of them, namely that there is an important difference between conceptual and factual problems and investigations. That claim draws in its wake other claims, all of them revolving around the idea of a conceptual scheme, what Wittgenstein calls a ‘grammar’. I explain why Wittgenstein’s account does not fall prey to Davidson’s animadversions against the idea of a conceptual scheme as a force operating on a pre-conceptual content. In the sequel I deny that the distinction between grammatical and empirical propositions disappears in the last writings: it is neither deliberately abandoned, nor willy-nilly undermined by the admission of hinge propositions in On Certainty or by the role accorded to agreement in judgement.