Recovery recovered

The most controversial condition that the AGM theory of rational belief cha nge places on belief contraction is the recovery condition. The condition i s controversial because of a series of putative counterexamples due (separa tely) to I. Levi and S.O. Hansson. In this paper we show that the conflicts that Levi and Hansson arrange between AGM contraction and our intuitions a bout how to give up beliefs are merely apparent. We argue that these confli cts disappear once we attend more closely to the identification of the beli efs contracted away in each counterexample case. Since, on our view, speake rs' belief contraction intentions are often more complicated than they may first appear, we are led to develop apparatus for thinking about the commun ication and identification of those intentions. Our argument refocuses atte ntion on the difficult question of how to apply the AGM theory to particula r cases.