Functional vs. Relational Approaches to Belief Revision

Abstract

In belief revision theory there is an on-going debate between functionalists and relationalists that shows little sign of ending in shared agreement. Roughly, functionalists hold that the result of revising a cognitive state with some new datum is a unique rationally determined belief state. Relationalists disagree. In their view, there may be several rationally admissible results of revising a cognitive state with a given datum. In an attempt to contribute to conceptual clarity three ways of drawing the functionalist-relationalist distinction are identified, giving rise to six non-contradictory overall positions. I proceed to consider arguments in the literature for excluding some of these positions on logical, philosophical or other grounds. Finally, I argue that part of what feeds the functionalist-relationalist controversy is a false dilemma based on an implausible conception of what it means rationally to suspend judgment. Making this precise requires a formal framework that includes a representation of the agent’s research agenda.