Human Performance in Default Reasoning

system of default inference is, nor any other standard with which to compare artificial agents that embody such a system. And so it follows that any judgment of the correctness of a proposed system of default reasoning will be as a test of the system against the way humans perform on that reasoning problem. (This outlook on default reasoning, augmented by many quotations from AI researchers who would appear to believe the same, can be found in Pelletier & Elio 1997).