COMMENT ⊗   VALID 00002 PAGES
C REC  PAGE   DESCRIPTION
C00001 00001
C00002 00002	\input buslet[1,ra]
C00004 ENDMK
C⊗;
\input buslet[1,ra]
\jmclet
\vskip 30pt
\address 
Prof. Ronald M. Lee
College and Graduate School of Business
The University of Texas as Austin
Austin, Texas 78712-1175

\body
Dear Prof. Lee:

	Thanks for your January 15 letter.

	My comment about Prolog was not a suggestion to use Lisp which is
even more unsuitable.  (CBCL doesn't use Lisp; it merely uses lists as
a syntactic medium).  It
was rather an objection to using a programming language to express
declarative information.  Prolog is an executable subset of first order
logic.  In order to achieve executability it sacrifices some
expressiveness.  A communication language doesn't need direct
executability, so why restrict yourself to Horn clauses with an unusual
(if any) interpretation of negation and give up full quantifiers?

	Secondly, unboundedness of discourse doesn't preclude formalization,
but it does require non-monotonic reasoning.  It may also require
formalization of the notion of context.  Unfortunately, I don't yet have
a paper on this later topic.

\closing
Sincerely,       

John McCarthy    
\annotations
cc: S. Kimbrough
\ \ \ \ T. Winograd
\bigskip   
JMC/ra 
\endletter
\end

Sincerely,