Acquaintance and awareness in belief representation

A (partial) solution to the well-known problem of "logical omniscience" in modal-like logics for knowledge and belief is suggested; it is based on a notion of "awareness of a proposition" inspired by the idea of "explicit belief" (i.e. "c1assical" belief modulo awareness). Our formal treatment relies on the idea that an agent is aware of a proposition ? off he is "capable of attributing a meaning to ?": this idea has lead us to develop a logic for belief where the fact that an agent is aware of a proposition may be characterized in terms of its "terminological content" alone. Following the introduction of the awareness requirement, our notion of explicit belief does not suffer from logical omniscience: in particular, it is not closed with respect to logical equivalence and "valid implication". We argue that this is a useful feature when such logics are applied to e.g. user modelling and, in general, to the specification of intelligent rational agents.