Semantic compatibilism is a meta-philosophical strategy of reconciliation between our best philosophical theories and recalcitrant common sense beliefs. It states that a recalcitrant assertion of a common belief is reconciled with our theory in case it is shown that the proposition really expressed by the ordinary assertion is compatible with the theory. Semantic compatibilists traditionally have appealed to a series of semantic-pragmatic phenomena to justify this kind of reconciliation: from so-called “reconciling paraphrases” to contextual variation, from joint-carving quantifiers to loose talk. The main aim of semantic compatibilism, according to P. Van Inwagen, is that of avoiding global scepticism: if we cannot trust our common sense, this means that our capacity of forming knowledge is globally damaged, so every revisionist metaphysician should also be a compatibilist. In our paper we criticise semantic compatibilism at this foundational level: on the one hand, we defend the view that being revisionist with respect to a certain area of discourse (for example, being eliminativists with respect to composite artefacts) does not lead necessarily to global scepticism, on the other we claim that it is just semantic compatibilism to be at risk of being a source of a sceptical attitude.