A popular view in metasemantics is the view that a speaker’s dispositions regarding the use of a symbol determine the meaning of that symbol for the speaker. Kripke (Wittgenstein on rules and private language, Harvard University Press, Cambridge, 1982) arguments against simple versions of semantic dispositionalism have inspired ever new versions. A recent account in the literature, due to Warren (Noûs 54(2):257–289, 2020) offers a sophisticated version of semantic dispositionalism whereby certain conditions are imposed on speaker’s dispositions to count as (...) meaning determining—conditions we can refer to as ‘standard’. In this paper, I argue that there are a number of cases that suggest that even under such conditions, a speaker’s meaning and their dispositions can come apart, and that this suggests that dispositional accounts of semantic content presuppose that there are semantic norms, independent of a speaker’s dispositions, and thus do not explain semantic content. I conclude that, since these problems generalise to non-standard dispositionalist accounts, the dispositionalist’s prospects are dim. (shrink)