COMMENT ⊗   VALID 00002 PAGES
C REC  PAGE   DESCRIPTION
C00001 00001
C00002 00002	Comments on your draft thesis proposal:
C00005 ENDMK
C⊗;
Comments on your draft thesis proposal:
.item←0;

	#. The proposal raises interesting issues as it stands.  I
hope you will show it to Weyhrauch, Glassmire, Goad, Bulnes,
Bob Smith, Freeman Rawson, and Douglas Hofstadter, Suppes, and
possibly Hintikka.

	#. It seems that the proposal is for a theoretical thesis.
I think this is reasonable, but I would like you to do everything
possible so that the ideas will enter the AI world and not be
solely of interest to philosophers.  This may involve implementing
some of the problem solving methods, but it may be better to work out
some detailed examples that demonstrate your methods
and put them in a form in which they can be a challenge to
system builders.

	#. I still think the syntactic approach to knowledge is
better precisely because the laws are sentences in the logic
and not built in and hence are changeable.  However, I recognize
the difficulties you have found, and I will try to solve them.
I think that my notion of extensional forms may solve the problems
about formalizing substitution into quantified expressions
that worry you.  This is because when λ is a function, the bound
variables actually disappear.

	#. I would like to see some trial examples of the reasoning.

	#. It is not obvious to me that loops and recursion don't
introduce new uses of knowledge.  We have both "it will eventually
know" and "it knows that eventually" and "if I do X I will never
know" as well as knowing that no matter how many times you rearrange
it, it's still only fifty cents.  I will go along with leaving these
things out, but you might think some more about what you may be
losing by doing so.

	#. Work out my telephone example using only ⊗knowing ⊗that 
in order to verify that it doesn't become to clumsy.