Concepts Meaningful in Approximate Theories

	We contend that many, perhaps most, common sense
concepts are meaningful only in %2approximate theories%1
and lose their meaning when the theory is refined.
Therefore, many philosophical attempts to explicate
the meaning of counterfactual conditionals and statements
of causality among others are misguided and doomed
to result in useless hair splitting.
This idea was first proposed in (McCarthy 1979)
and will be further developed in this paper.

Notes

	Approximate theory in which the counterfactual
"If 4 were a prime, then 2 x 4 would be prime" is false.
This extreme example is intended to put the issue sharply.

	Can we strip the ski instructor dialog example to
its essentials?  Can we criticize David Lewis's definition
with an appropriate example?

	Do all concepts meaningful in approximate theories
depend on counterfactuals?  If so we could imagine a theory
with counterfactuals (or rather conditionals) as abstract
objects.

	Cartesian counterfactuals.  By introducing limiting
adjectives, maybe we can separate the controversial aspects
of the theory from the technical aspects.

	What generalization of cartesian counterfactuals is possible?
What about tree counterfactuals?

	The automaton model of ability.  Is it also a model
of causality?  An intentional action.

	What about more conventional examples like the weight of
a man?  The beliefs of the government and its goals.  Might we
ascribe to a species the goal of survival and suggest that it
"thinks" of modifications that lead towards this goal.

	Hayes suggests that a space may admit a Cartesian structure
only locally, and this seems to be correct.