Dynamic doxastic logic: why, how, and where to? (Q885520)

Dynamic doxastic logic (acronym DDL) seeks to reconstruct and extend AGM-style belief revision theory by bringing the belief revision operation into the object language (treating it as a modal operator like that of dynamic logic), at the same time enriching the object language with operators for belief and knowledge (in the manner of epistemic/doxastic logic). The authors review the present state of play in this programme, with special attention to the light that it may throw on the differences between AGM-type revision and KGM-type update, as well as on the import of Gärdenfors' well-known impossibility theorem concerning the Ramsey test for conditionals in terms of revision. The paper ends with a list of questions and projects whose resolution may take the subject further. One of the most interesting sections of the paper compares Gärdenfors' impossibility theorem with the even more celebrated impossibility result of Arrow in the theory of collective preference. It is suggested that the two impossibilities manifest a common underlying pattern, and in support of this claim the authors construct a translation map from the structures used in collective preference theory to what might be called ``bi-systems-of-spheres'' for belief revision in a language admitting a non-classical conditional connective in addition to the usual Boolean ones. It is claimed that three assumptions used by Arrow for his impossibility theorem (the Pareto condition, independence of irrelevant alternatives, non-dictatorship) translate into conditions on such bi-systems-of-spheres, leading to a result for conditionals and revision with some similarity to Gärdenfors' impossibility theorem. Unfortunately, while the translation of the Pareto condition comes out nicely as one half of the Ramsey test, the translations of the other two conditions (especially the independence of irrelevant alternatives) are rather more complex and difficult to interpret. The presentation is only an outline, with proofs omitted. The discussion, carried out in section 3.2, appears to be quite independent of considerations concerning dynamic and epistemic/doxastic operators in the remainder of the paper, and indeed of the programme of DDL. It is to be hoped that the authors will take the matter further and in detail in a future publication.