COMMENT ⊗   VALID 00002 PAGES
C REC  PAGE   DESCRIPTION
C00001 00001
C00002 00002	The Ontology of Artificial Intelligence
C00006 ENDMK
C⊗;
The Ontology of Artificial Intelligence
 
	Ontology is the branch of philosophy concerned with what entities

exist.  Quine (19xx) proposed that when the use of a logical
language is in question, one is "ontologically committed"
to those classes of objects which can be values of bound variables,

and Quine's proposal seems to suit our present purposes.
Philosophy, especially in this century, has tended to be ontologically

stingy.  Occam's razor suggests that entities not be multiplied
beyond necessity, and philosophers have found many of the kinds
of objects suggested by ordinary language to be unnecessary.
 
	This paper gives reasons why artificial intelligence requires
a more generous ontology and gives examples of the heuristic
use of entities like reasons, things wrong with the boat, and
other rather abstract entities.
 
	For example, we may say that a boat may be used to cross a river

unless something prevents it, using a logical expression amounting to

qEx.(x prevents using the boat).  The variable  x  must then
be instantiated to entities that can prevent using a boat.  From
a strictly extensional logical point of view, the sentence seems
to convey no information, being equivalent to  "The boat can
be used to cross the river unless it can't".  Indeed people
frequently criticize each other's statements on the ground that
they are content-free in this manner.
 
	When we admit circumscription (and perhaps other non-monotonic
modes of reasoning) as a rule of conjecture, the two statements
are not equivalent, and the class of "things that can prevent
the boat from being used to cross the river" becomes important.
Thus consider th@ question "What else is preventing the boat
from being used to cross the river".  It is an ordinary and
useful common sense question, and its posing suggests that
things preventing the boat from crossing the river can be
counted - at least some of the time.