COMMENT ⊗   VALID 00002 PAGES
C REC  PAGE   DESCRIPTION
C00001 00001
C00002 00002	barwise[w85,jmc]		Barwise on conditionals
C00005 ENDMK
C⊗;
barwise[w85,jmc]		Barwise on conditionals

	Perhaps I should do a note on conditionals.

0. Clearly Barwise's examples are worth thinking about.

1. Barwise is cryptic or even misleading in his references to AI.
Moreover, he doesn't know about the AI strategy of doing the
easy things first.

2. on the ``fleshing out strategy''.  In some circumstances this
works even in the crude form of inserting an extra premiss.
Parametrization by possible worlds or context is a more
sophisticated form of fleshing out.
In other circumstances, maybe all, a non-monotonic rule
applies.
See my notes on p. 20.

3. Barwise also doesn't mention the reification strategy that
permits expressing anything, even situation semantics, in first
order logic.  He also doesn't know about the argument that this
is a correct strategy.

4. Barwise's account of conditions, which I don't fully
follow, may lack sufficient non-monotonicity.  He seems to
attach a pre-ordained set of parameters to a sentence (or is
it a statement?).  Circumscription makes the consequences of a
sentence depend on all the facts that are being taken into account.

5. p30 If Claire is rubbing her eyes, then she is sleepy.
If Claire is rubbing here eyes and there is pollen present,
then she is sleepy.
Barwise wants the first true and informative in the normal
situation and the second false.
Circumscription seems to win here by

clairrubbing ∧ ¬ab aspect1 ⊃ clairsleepy
pollen-present ⊃ ab aspect1

6. Example E. normally numbers are real.