A natural strategy to account for evaluations’ special connection with action is the one suggested by expressivism, which individuates the meaning of a sentence in terms of the attitude it expresses and takes evaluations to express attitudes other than belief. Expressivists have historically focused on solving what I will call their theoretical problem, which is that of accounting for evaluations’ special connection with action without falling prey to Frege-Geach worries. However, I will assume that expressivists nowadays have a solution for this problem that involves understanding the expression of an attitude as the proposal to add that attitude to the common ground. Some characterizations of the common ground, however, make phenomena such as propaganda, in which something that is presented as a pure description is in fact at least partially an evaluation, unintelligible, even if expressivism, with its robust description-evaluation distinction, should look as a specially promising account of them. This problem, which I will call expressivists’ practical problem, results from the unclarity of the notion of common ground involved in the characterization of expression just given. The aim of this talk is to explore how expressivists should construe the common ground in order to solve their practical problem, and in particular to deliver a satisfying account of cases of propaganda. I will suggest that expressivists should rely on something closer to the conversational record, as characterized by Camp, than the common ground as originally introduced by Stalnaker.