COMMENT ⊗   VALID 00005 PAGES
C REC  PAGE   DESCRIPTION
C00001 00001
C00002 00002	#. Airline Reservation Systems
C00003 00003	Metaphysics and epistemology - 
C00007 00004	#. The "argument from design".
C00009 00005	~	This  kind  of  teleological  analysis  is  often  useful  in
C00014 ENDMK
C⊗;
#. Airline Reservation Systems



Metaphysics and epistemology - 

	This paper following (McCarthy and Hayes 1969) has used these
terms in ways that correspond reasonably well to the dictionary
definitions,  but which need to be explained.

	An intelligent computer program must be equipped with certain
framework within which facts of science and of common sense can be
put.  The necessary notions include time (at least in the form
of a precedence relation between events), events, states, objects,
the persistence of objects in time, causality, ability, the
creation and destruction of objects, the locaation and motion
of objects in space, mass terms, and actors with their ability
and inability to achieve goals.  I have called this domain
metaphysics, because its general formulation should be neutral
between physical theories.  This framework must allow the expression
of facts of particular sciences and also facts of what may be
called %2common sense physics%1 - the knowledge that allows us,
for example, to predict the effects of spilling a cup of coffee
on the table without using the science of hydrodynamics.  Common
sense science gives less information than science proper, but it
demands less input data and less understanding as well.
Formalizing common sense science encounters difficulties not
met in science proper, because of the need to use extremely
incomplete information and imprecise theories.

Much recent analytical philosophy can be regarded as common sense
science, but the problem of partial information hasn't really
been directly faced.

	The word ⊗epistemology has always concerned the study of
what can be known, but this is ordinarily taken as referring to
very general knowledge.  For AI, we need to study what can
be known under particular circumstances of observation and previous
knowledge and how this knowledge can be represented in the
memory of a computer, and how new knowledge can be obtained.
The representation of facts what other's know is particularly
important, and again analytical philosophy has studied many
of the questions important for AI.
Two technical issues that have been studied in the Stanford Artificial
Intelligence Laboratory are "knowing what" as compared to
"knowing that" and the formalization of a theory of non-knowledge.
Giving sufficient conditions for proving that someone does not
know a fact.
#. The "argument from design".

	From the AI point of view, the most powerful argument for
materialism may be called %2"the argument from design"%1.  Namely,
if one sets out to make an intelligent machine, there seems to be
no way to do it without producing something with the following
characteristics:

	1. There is no way to assure that the world
can be defined in terms of its sense impressions, because there
is no way to insure that it can discover everything we humans
can discover.  Thus even if I am some kind of empiricist, my
intelligent program can't correctly (from my notion of correct)
be an empiricist.

	2. Suppose we provide it with a certain way of representing
general hypotheses about the world.  There is no way we can insure
that the basic concepts we provide it with are more than approximations.

~	This  kind  of  teleological  analysis  is  often  useful  in
understanding natural organisms as well as machines.  Here  evolution
takes  the place  of  design and  we  often understand  the  function
performed by  an organ before we  understand its detailed physiology.
Teleological analysis  is  applicable to  psychological  and  social
phenomena in  so far as  these are designed  or have been  subject to
selection.

	However,  teleological analysis fails when applied to aspects
of nature which  have neither been  designed nor produced by  natural
selection from a population.   Much medieval science was based on the
Judeo-Christian-Moslem  hypothesis that  the details of  the
world were designed by God  for the benefit of man.   The strong form
of  this hypothesis was abandoned  at the time of  Galileo and Newton
but occasionally recurs.   Barry Commoner's (1972) axiom of  ecology
"Nature knows best"  seems to be mistakenly based on  the notion that
nature  as a  whole  is the  result of  an evolutionary  process that
selected the "best nature".~
	
#. %3Actions%1.  We want to distinguish the actions of a being
from events that occur in its body and that affect the outside
world.  For example, we wish to distinguish a random twitch
from a purposeful movement.  This is not difficult %2relative
to a theory of belief that includes intentions%1.  One's purposeful
actions are those that would have been different had one's intentions
been different.  This requires that the theory of belief have sufficient
Cartesian product structure so that the counterfactual conditional
`"if its intentions had been different" is defined in the theory.
As explained in the section on definitions relative to an approximate
theory, it is not necessary that the counterfactual be given a
meaning in terms of the real world.

	#. If, as we conjecture, most common sense and even scientific
terms are meaningful only in approximate theories, then a philosophical
method as old as Socrates needs to be re-examined.  This method
involves attacking a common sense notion by introducing examples that
have not been encountered in the previous usage of the notion and
showing that in these cases the notion gives unacceptable results.
But it may be that any theory of ⊗just or ⊗unjust actions must
be based on an approximate model of the world, and anomalies can
always be found.  Thus philosophical analysis cannot invalidate a
notion by finding limits on its applicability; that requires
a better and more general notion.
Even when found, the improved notion will still be limited, and
there may never be a notion of just action
founded on quantum physics and chemistry.  Notice that this conclusion
does not depend on considerations of ⊗emergent ⊗phenomena.