“…As the following proposition shows, knowledge can be defined in terms of belief if the logic of knowledge is S4.4, but not if the logic of knowledge is S4 and S4.x, where x ranges over {.2, .3, .3.2}. This result can be contrasted with Theorem 4.1 in (Halpern et al, 2009a), from which it follows that the knowledge modality cannot be explicitly defined in the logic (S4.4) K + (KD45) B + {KB1,KB2}. We see once again that the increase in expressivity due to the addition of the interaction axiom KB3 plays an important role in bridging the gap between belief and knowledge.…”