COMMENT ⊗   VALID 00002 PAGES
C REC  PAGE   DESCRIPTION
C00001 00001
C00002 00002	Notes for talk May 14
C00005 ENDMK
C⊗;
Notes for talk May 14

1. Why non-monotonic reasoning and why it can be discussed
epistemologically.  Why we say "non-monotonic reasoning" rather
than "non-monotonic logic".

I will let you feel the trunk and one ear of the elephant.

defeasible presumptions

[unsafe because there may be a train]

circumscribing what objects exist

Avoiding making distinctions till they are necessary
Necessary if philosophy is to be realistic and AI is to be possible.

2. Examples
	[noon ∧ M sunny ⊃ sunny] ∧ noon ∧ [eclipse ⊃ not sunny]
	the meeting will be on Wednesday unless there is reason to do different
	the story of the peas
	isflier tweety
	M istrain ⊃ unsafe
	non ex boats and helicopters
	Mary's home town - the cases where it isn't well defined
must be treated in an ad hoc way
	attempting to bribe a public official
		a. didn't know he was an official
		b. mistakenly thought he was an official
		c. advertised for a bribee
It would be desirable to have a purely de re example where a distinction
is not imagined until the necessity for it arises.  Zenon suggests
"mammal" but here we can suppose that the ambiguities are known in
advance.  Best one in which the audience doesn't see the ambiguity
till it is pointed out.

3. Formalisms
	McDermott, Reiter, McCarthy
	Davis and Stalnaker comments
	Weyhrauch and Winograd

4. Heuristic aspects of non-monotonic reasoning
	refer to Reiter, microplanner, Doyle, and Moore

5. The problem of formalizing "natural kinds".