An Empirical Test of Patterns for Nonmonotonic Inference

: It is claimed that human inferential apparatus offers interesting ground in order to consider the intuitions of artificial intelligence researchers about the inference patterns a reliable nonmonotonic consequence relationship should satisfy, since (a) human inference appears to be nonmonotonic, and (b) human inference, being adaptive to some extent, generally achieves efficiency. An experiment is conducted to investigate whether human inference tends to be consistent with rationality postulates (System P plus Rational Monotony), especially when it no longer satisfies the Monotony property. The experimental protocol makes use of a possibilistic semantics of plausible rules. Our results appear to be consistent with all the studied properties, except with the CUT property given one kind of material and Left Logical Equivalence which cannot be tested. Moreover, when Monotony was not satisfied by participants' inferences, Cut, Cautious Monotony and And properties were corroborated (R...

• Type: Journal
• Year: 2002
• Where: AMAI
• Authors: Rui Da Silva Neves, Jean-François Bonnefon, Eric Raufaste
• Keywords: AMAI 2002, Artificial Intelligence, Human Inference, Inference, Rational Monotony