In this paper, I will provide a solution to Fitch's paradox which is based in the failure of epistemic closure. My approach will be semantical, following Beall (2009). Unlike his proposal, my system is logically classical, and relatively conservative with respect to statements about knowledge; all its deviations will be caused by the introduction of modal statements. In order to make this happen, I will introduce abnormal possible worlds (in this, I'm close to Beall), but also open worlds of two different kinds, which will explain the epistemic deviations of the agents.
Fitch's paradox; Knowability; Open worlds; Anti-realism; Epistemic closure