COMMENT ⊗   VALID 00005 PAGES
C REC  PAGE   DESCRIPTION
C00001 00001
C00002 00002
C00006 00003		Consider a computer program for conducting a reasonable
C00007 00004		This AI research offers the following question to philosophy
C00008 00005
C00009 ENDMK
C⊗;
	Consider a computer program for conducting a reasonable
keyboard mediated conversation in Chinese.  The ultimate goal
would be a program that couldn't be distinguished from an
intelligent Chinese, but AI isn't ready for that yet, so lesser
goals will be considered.
	This AI research offers the following question to philosophy
about ascription of knowledge and belief to machines or other systems.

     Shall we allow any rule for ascribing knowledge and
     belief that makes the behavior subject to useful axioms of
     rationality?

	The scientific alternative is to attempt to define and require
a certain minimal structure before mental qualities are to be ascribed.
There is also a non-scientific alternative that refuses to make
definitions and relies on examples.

Some of
these situations are apparently well structured as in chess,
proving theorems in a given formal axiomatic theory, diagnosing
diseases and recommending treatments within a fixed theory of
disease and recommending whether to honor a proposed charge on a
credit card according to fixed rules.  However, the center of AI
research is the {\it common sense informatic situation}.  Here
there is a database of information about the common sense world
and a description of a particular situation and the goals to be
achieved.