COMMENT ⊗   VALID 00002 PAGES
C REC  PAGE   DESCRIPTION
C00001 00001
C00002 00002	.require "let.pub[let,jmc]" source
C00005 ENDMK
C⊗;
.require "let.pub[let,jmc]" source;
∂CSL Dr. Nigel Seel↓Standard Telecommunications Laboratories Ltd
↓London Road↓Harlow↓Essex CM179NA↓ENGLAND∞

Dear Dr. Seel:

	Thanks for your paper "Computational Cybernetics: ...".  I have
two comments.

	First, since I coined the term artificial intelligence in the
middle 1950s I can speak to its motivation.  Avoiding the term
cybernetics, which might have led to Norbert Wiener being consulted
about the appropriate direction of our research, was an important
motivation.  I had read his %2Cybernetics%1 and concluded that he
was on the wrong track for substantially similar reasons to those
mentioned in your paper.  Namely, he didn't face the qualitative
complexity of intelligent behavior.  That rather than questions of
stability, etc. seemed then as now to be more important.  I remain
a skeptic about the relevance of even present day cybernetic research
to AI.	

	Second, I agree with most of the qualitative remarks you make
and in particular with your discussion of the intentional aspect of
systems.  However, I am not optimistic that starting at such a general
level and then specializing will obtain results relevant to AI.  At
least I didn't get anything I considered worth publishing when I took
that approach, admittedly not considering intentionality, in the 1950s.
The one published residue of that effort was "Inversion of functions
defined by Turing machines" in %2Automata Studies%1, Princeton University
Press, edited by Shannon and McCarthy.  That work has not so far
turned out to be continuable.

	For this reason I am skeptical that the approach outlined in
your paper will result in mathematics actually applicable to AI.

.sgn