Attunement to Constraints in Nonmonotonic Reasoning

This paper began as an attempt to use domain theory to throw some light on some of the semantic problems in the area of non-monotonic logic within artificial intelligence. After having thought about some of the issues involved, though, it seemed to us that the examples and methods we use might be of interest in a broader setting. We are offering the results of our investigations in that spirit, hoping that this will be a start on the problem of putting the work in standard AI approaches to non-monotonicity together with current work on information flow (Seligman and J.Barwise 1993.) We are interested in the subject of default or "normal" inferencing. On the surface of things, this can be exemplified within propositional logic using a non-monotonic consequence relation ff j¸ fi between sentences f