COMMENT ⊗   VALID 00002 PAGES
C REC  PAGE   DESCRIPTION
C00001 00001
C00002 00002	PATTERNS AND CONCEPTS
C00004 ENDMK
C⊗;
PATTERNS AND CONCEPTS

by John McCarthy, Stanford University


	Papers reporting research in pattern recognition or concept
formation often use very limited notions of pattern or concept without
even hinting that there are patterns and concepts that don't fit
under their notions.  Some even give formal definitions of what
a concept is that are extremely limited in what they cover.

	This paper gives some examples of such limited notion, gives
some examples of patterns and concepts that don't fit them, and
gives a more general notion of pattern that includes all examples
that I know about.  However, since it has an arbitrary first order
language as a parameter, what patterns and concepts can be defined
depends on the language.  To some extent this passes the buck to
the inventor of the language, but we shall argue that this is the
way it should be.

	The works we shall criticize include (Bruner 1956), (Hunt 197x)
(Winston 197x) and (Mitchell 1981).

	None of the above notions can handle the following pattern
which we present in a form suitable for an intelligence or scholastic
aptitude test.