What One May Come to Know
Johan van Benthem

Abstract:
Fitch's 'paradox of the Knower' shows that verificationism in the
appealing version "everything that is true may be known" is
inconsistent. We give an even stronger dynamic version: "everything
that is true may come to be known", and point out its connections with
the 'learning problem' for statements in update logics. Moreover, we
show how different natural versions of the verificationist thesis hold
for different classes of epistemic statement, and that these classes
are all decidable in basic epistemic S5.