Moinard, Yves and Rolland, Raymond Résultats inattendus de la circonscription (premier et second ordre) , RFIA (Reconnaissance des Formes et Intelligence Artificielle) , Paris , 1301--1314 , dec , 1989 Abstract Circumscription, one of the best known forms of nonmonotonic reasoning, minimizes the extension of some predicates. we deal with the first order and the second order versions of circumscription, thus we begin by making precise the differences between these two versions. one problem with circumscription is that it can result in an inconsistency when applied to certain consistent theories. We give some examples proving that, even with simple theories, the second order version is more likely to produce such an unwanted result. Also, we show that the second order version may be equivalent to such an unespected result as a finiteness axiom. the more strikingly unespected results are the cases where the minimization of a predicate ``P'' allows to prove ``for all x, P(x)'', while the original theory did not allow to prove this formula. This may occur with first order and second order versions of circumscription. This web site is maintained by René Quiniou using the Weave system from the Caravel project: http://www-caravel.inria.fr. Last modification: 10-07-2011 09:54:25 Version française