COMMENT ⊗   VALID 00002 PAGES
C REC  PAGE   DESCRIPTION
C00001 00001
C00002 00002	brown.re1[s88,jmc]	Review of nsf proposal by Frank Brown
C00006 ENDMK
C⊗;
brown.re1[s88,jmc]	Review of nsf proposal by Frank Brown

	I regret that I cannot take the time to give this proposal
a proper independent review.  There are so many new nonmonotonic
systems that I'm far behind, and my sporadic efforts have never
given me a clear picture of Frank Brown's work.  This may well be
my fault rather than his.

	Perhaps I can give some help based on the fact that
I have seen a referee's report of a paper based on what seems
to be essentially the same material as that submitted with the
proposal.  Unfortunately, the referee wasn't able to undertake
reviewing the proposal also.  The referee remarked that 

``The goal of the paper is to show that the author's modal system Z subsumes
parallel circumscription, and also to use Z for the investigation of the
relation between circumscription and other non-monotonic formalisms. The
technical results proved in the paper do not convince me that these goals
are achieved.

The main problem is that it isn't clear how the ``modal definition''
of circumscription is related to McCarthy's original
definition. The author arrives at his modal definition in two steps: he
writes a semi-formal ``semantic definition'' of circumscription, and then
rewrites it in the language of Z. The semantic definition is basically
correct, except that one condition is missing in the definition of leq:
the models j and k should have the same universe. But in what sense can
this definition be ``rewritten'' in Z? In the appendix, Z is defined as an
inference system without any semantics, and I don't understand how the
author uses the language of Z for talking about models of first-order
theories (in the sense of classical model theory as used by McCarthy).''

	It isn't clear to me that these comments are decisive in
evaluating whether the work should be further supported.

	I'm sorry not to be more help, and I hope your other reviews
are more useful.

p.s. The modal approach to nonmonotonic reasoning has value apart from
whether it includes the others.