ERRATUM: REPRESENTATION-THEORETIC SUPPORT SPACES FOR FINITE GROUP SCHEMES

Abstract. As pointed out to us by Rolf Farnsteiner, the results presented in our paper [1] require a modified definition of “abelian p-point. ” With this modified definition (functionally equivalent to one which we implicitly use), all of the results of our paper become valid. We make explicit this modified definition as well as those arguments which require this new definition and the modification of one proof (of Theorem 4.8) which is required. We recall that we use the notation kG for the group algebra of a finite group scheme G, the dual of the coordinate algebra k[G] of G. As defined below, a “p-point ” α: kZ/pZ → kG is an abelian p-point (as in Definition 3.2 of [1]) with the additional condition that α factors through some subgroup scheme which is not only abelian but also unipotent. Among the numerous equivalent conditions on an abelian finite group scheme C to be “unipotent ” (cf. [2, 8.3]), we implicitly utilize the condition that the group algebra kC be a local algebra (i.e., that C be co-connected). Definition of p-point. Let G be a finite group scheme. A p-point of G is