To solve the problem of counterpossibles (i.e., counterfactuals with necessarily false antecedents), many philosophers have been arguing that one needs to invoke impossible worlds. This extension of the semantics of modality should save the analysis of counterfactuals from being insensitive to the problem of counterpossibles. The aim of this paper is to show why the theories of impossible worlds do not fully solve the problem of counterpossibles, but merely shift it. Moreover, by distinguishing two types of languages, we will show that some expectations about a proper theory of counterfactuals might be too great.