Conditionals and monotonic belief revisions: The success postulate (Q2277446)

The author investigates the role played by the postulate of ``success'' in the AGM logic of theory change (i.e. the postulate that K*x is consistent whenever x is) in the generation of Gärdenfors' impossibility theorem (which says that certain of the AGM postulates are inconsistent, up to triviality, with the so-called Ramsey test). He argues that it is indeed the success postulate that should be regarded as responsible for the inconsistency, and that it continues to be so even when weakened in a way in which Gärdenfors at one point suggested. For this reason, he suggests, the postulate of ``success'' should be dropped from the logic of theory change.