From: John Stick (johnstick@worldnet.att.net)
Date: Tue Nov 29 2005 - 12:03:20 MST
Deontic logic, like modal logic, is only a minor variant of standard logic.
I would doubt that it would add anything worthwhile to your AI system: the
addition is so minor and besides, it is moving in the wrong direction. In
effect they extend the predicate calculus by each taking a single predicate
(necessity for modal logic, morally required for deontic logic),
transforming the predicate into an operator and hardwiring an inference rule
for the predicate/operator into the system. If you are trying to build a
general AI, how is that helpful?
My background is in philosophy. Modal logic and deontic logic appealed to
some philosophers because the operator added was thought to be the essential
concept in an important philosophical field, and so hardwiring an inference
rule for it, if inelegant, seemed useful. But I never found that it
actually helped. For the field I worked in, moral and legal philosophy,
deontic logic was, in my controversial opinion, a blind alley. But even if
I am wrong about that, it is just so narrow an extension of the standard
logic you will build into your system anyway.
Have you looked at Ben's predicate term logic? Are you adding a Bayesian
analysis module? A fuzzy logic component of some sort? pattern matching?
These all reach far beyond standard logic. Compared to them, the
usefulness added by modal/deontic logic is infinitesimally small.
This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:53 MDT