SPEAKER: Jim Delgrande

TITLE: Modelling Belief Change Operators Using Ordinal Conditional Functions

DATE: Thursday, October 11, 2001, 12:30 p.m. - 1:30 p.m; ASB 9705

ABSTRACT:

Ordinal conditional functions have emerged as a central means for modelling belief change operators. (As well, ordinal conditional functions generalize major approaches to modelling default, counterfactual, and deontic reasoning, among others.) The main idea is that, given a knowledge base K, interpretations in the language are assigned a number representing their "closeness" (or perhaps "degree of belief") with respect to K. To revise K by a sentence A, for example, one simply selects those interpretations that are closest to K and in which A is true. The result is an appealing and intuitive modelling for belief revision, and a general approach upon which numerous specific approaches have been based. However, it is now becoming clear that ordinal conditional functions have their limitations, and in particular it is not clear how repeated (or iterated) belief change operators can be modelled.

In this talk I'll briefly introduce the area of belief revision, in particular the dominant AGM approach (named after the founders, Alchourron, Gardenfors and Makinson). After this, ordinal conditional functions will be discussed, and their advantages and limitations illustrated. To conclude (and assuming sufficient time) some open problems will be given. This talk will be informal for the most part, focusing on intuitions rather than formal details.