To propose a revisionist ontology of art one has to hold that our everyday intuitions about the identity and persistence conditions of various kinds of artworks can be massively mistaken. In my presentation I defend this view: our everyday intuitions about the nature of art can be (and sometimes are) mistaken. First I reconstruct an influential argument of Amie L. Thomasson (2004; 2005; 2006; 2007a; 2007b) against the fallibility of our intuitive judgments about the identity and persistence conditions of various (...) kinds of artworks. Second, I present three objections to this account: two of them concern the semantic and pragmatic rules regulating the use of art-kind terms, while the third one is based on the assumption that the history of art partly comprises a series of successful attempts of transgression of artistic conventions and expectations, therefore our artistic intuitions are dynamic. Taking this point I finally argue that in philosophy of art we need a “reverse” methodology: first we have to provide a general definition, containing all the sufficient and necessary conditions of artworks (“being an artwork”) in any period within the history of art. Only after completing this task are we ready to answer the metaphysical question about the ontological status of works of art. (shrink)