COMMENT ⊗   VALID 00003 PAGES
C REC  PAGE   DESCRIPTION
C00001 00001
C00002 00002	.cb THE CASE FOR ONTOLOGICAL ENRICHMENT
C00005 00003	Notes:
C00006 ENDMK
C⊗;
.cb THE CASE FOR ONTOLOGICAL ENRICHMENT


	Common sense language seems to refer to many kinds of entities
that philosophers, philosophers of science and scientists find superfluous
and advise against including in scientific or philosophical theories.
Some of them are ⊗thoughts, ⊗voices, ⊗redness, ⊗sakes and ⊗behalfs, %2things
wrong with the boat%1 and %2dormitive virtues%1.  We will argue that,
in the main, common sense is right and philosophy is mistaken.  What
some call "ontological bloat", we will call "ontological enrichment".
Our goal is artificial intelligence, and we will argue that an
intelligent computer program must admit and quantify over many
kinds of entity and must be able to introduce new kinds of entities
quite freely in reaction to experience and conversation.

	For the benefit of the AIer not familiar with philosophical
terminology, we give the meanings of a few terms.  The philosopher
should also take note that they are used in this paper with no
more baggage than is contained in this brief explanation.
In classical philosophical terminology, ⊗ontology is the study
of what exists.  Quine proposes to make this precise by saying
that the ontology of a theory is given by specifying the ranges
of its bound variables, e.g. ⊗redness is in the ontology of
a theory if ⊗"redness" may be substituted for ⊗x in some
sentence beginning %2∀x%1.  We expect to go along with Quine
in this usage.  Another useful term is ⊗reify, i.e. %1to make
a thing out of%1.  ⊗"Re" is the Latin word for ⊗"thing".  We
reify the predicate ⊗red to get the constant ⊗redness.  This
is the predicate calculus analog of turning the adjective "red"
into the noun "redness".

Notes:

	We need a language permitting the frivolous introduction
of ⊗dints, ⊗sakes and ⊗voices without (1) any investigation of their
exact ontological status and without (2) risk of paradox.
It may turn out that as with common sense, paradoxical results
can be obtained by use of ordinary circumscriptions, but are
quickly rejected.  Certainly non-monotonic reasoning is involved.

%3Dennett, Daniel C. (1968)%1: %2Content and Consciousness%1,
Routledge and Kegan Paul.

	I'm now not so sure that the criterion for admission to
the ontology is substitutability for bound variables.