In a number works Jerry Fodor has defended a reductive, causal and referential theory of cognitive content. I argue against this, defending a quasi-Fregean notion of cognitive content, and arguing also that the cognitive content of non-singular concepts is narrow, rather than wide.