Belief Revision and Language Splitting
- DIMACS Center - Room 431
- Busch Campus
- Piscataway, New Jersey
- March 15, 2:30 p.m.
The problem of belief revision goes back to a now classic paper due to
Alchourron, Gaerdenfors and Makinson (JSL-85). The central problem is
how to revise a theory T when a piece of information A inconsistent
with T is received. Clearly T must be contracted to a smaller theory
T' which IS consistent with A and then A added to T'. However, various
proposals for defining this process have run into trouble, either by
being too flexible and allowing implausible update operators or by
allowing ONLY the trivial update.
We propose axioms for update operators which are consistent with the
AGM axioms and which block the trivial update. The axioms are based
on the notion of splitting languages. In particular we show that an
agent's information can be split into various subject matters in a
precise and well defined way and show how this splitting can be used
to obtain a belief revision process which satisfies the AGM axioms,
and is provably non-trivial.
The talk will only assume familiarity with elementary Logic.