De dicto reading of predicate symbols in artificial intelligence

After an introductory discussion on the role of logics for mental attitudes in computational models of dialogue, we present the M-KRYPTON representation language. M-KRYPTON has been built in order to provide a tool for reasoning about attitude reports in the presence of multiple cognitive agents, and where an attitude report is interpreted de dicto, i.e. relative to (the reporter's view of) the cognitive space of the agent the attitude is attributed to. M-KRYPTON may be seen as an attempt to combine insights from AI knowledge representation and doxastic modal-like logics. On one side, hybrid knowledge representation systems embody a powerful representational paradigm, accounting for both belief (or, popularly, "knowledge") about the "necessary" nature of concepts relevant to the domain of discourse and belief in "contingent" facts about them. On the other side, the possible worlds semantics typical of modal logic is easily tailored to model various notions of a doxastic nature that are relevant to the modelling of cognitive agents. By recasting the KRYPTON hybrid KR system in terms of possible worlds semantics we have obtained a semantic account that, besides being "functionally" equivalent to the original one, is easily extensible to deal with operators for mental attitudes. In particular, we have concentrated on adding to KRYPTON the possibility of representing beliefs about the beliefs of multiple agents, where such agents may believe in propositions either of a "necessary" or of a "contingent" nature.