DDL unlimited: Dynamic doxastic logic for introspective agents (Q1583791): Difference between revisions

This is a paper in the logic of belief and belief change. The authors consider a liberalization of the syntax of \textit{K. Segerberg's} dynamic doxastic logic [in: H. J. Ohlbach et al. (eds.), Logic, language and reasoning, Trends Log. Stud. Log. Liber. 5, 135-147 (1999; Zbl 0957.03023)] so that the belief operator may be nested and applied also to statements of belief change correctness. The development of a Segerberg-style semantics for this language is explored, i.e., a semantics in which a doxastic state is postulated to be a so-called hypertheory, a family \({\mathbf H}\) of closed sets such that the intersection of \({\mathbf H}\) and the underlying universe belong to \({\mathbf H}\). This development is hampered by a number of unacceptable, `paradoxical' formulas emerging as valid. A remedy free of the observed shortcoming is presented. A distinction is drawn between states in which a belief is held (points of evaluation) and states about which certain things are believed (points of reference). Formulas are then evaluated at pairs \((x,y)\), where \(x\) is the point of evaluation and \(y\) the point of reference. In this two-dimensional semantics a distinction can be made between weak and strong validity. Certain formulas problematic within the one-dimensional semantics receive a new, unproblematic reading. Moreover, other problems can be avoided by assuming that various introspection principles are only weakly valid. The authors do not consider metalogical properties of their two-dimensional dynamic logic.