Dynamic doxastic logic (DDL) is the modal logic of belief change. In
basic DDL a modal operator [*?] carries the informal meaning “after the
agent has revised his beliefs by ?” or “after the agent has accepted the
information that ?”; it is assumed that the arguments of the star
operator * are pure Boolean formulæ. That assumption is discarded in
full DDL where any pure doxastic formula may be an argument. As noted by
other authors, a straight-forward extension of the theory from basic DDL
to full DDL invites problems of the kind first discussed by G. E. Moore.
In this paper it is argued that a way to escape those problems is to
redefine revision in a way that seems appropriate for this semantically
richer context. The paper deals only with the one-agent case, but the
approach can be extended to the case of multiple agents.