1989
	Abstract for work in PHILOSOPHY AND AI. 

	AMPLIATIVE INFERENCE, COMPUTATION, AND DIALECTIC




Here it is argued that the difference between expanding logical shorthand and
performing inference is not exhibited in defeasible or inductive reasoning
systems.  Such systems merely introduce more elaborate shorthands.  The
difference can best be understood in terms of the license to use arguments
despite incomplete search for rebuttals.  What is surprising, hence
ampliative, about the inference is the unintended behavior that results from
the dynamics of the reasoner in limited resource situations.

Then I muse about the appropriateness and challenges of a dialectical control
strategy for computing entailments of a defeasible reasoning system in limited
time.  Why is it better than negation as failure?  Why is it better
than lobbying support for propositions to which we are inclined?