The Chevalley-Shephard-Todd theorem for finite linearly reductive group schemes (Q442413): Difference between revisions

Given a field \(k\) and a faithful action of a finite group \(G\) on a \(k\)-vector space \(V\), \(g \in G\) is a pseudoreflection if \(V^g\) is a hyperplane. The classical Chevalley-Shephard-Todd theorem asserts that if \(|G| \not | \; \, \mathrm{char}(k)\), then \(k[V]^G\) is polynomial if and only if \(G\) is generated by pseudoreflections. In this paper, the author generalizes this theorem to the case of finite linearly reductive groups schemes. Let \(G\) be a finite linearly reductive group scheme acting faithfully on a \(k\)-vector space \(V\). A subgroup scheme \(N\) of \(G\) is called a pseudoreflection if \(V^N\) has codimension \(1\) in \(V\), and \(G\) is said to be generated by pseudoreflections if \(G\) is the intersection of the subgroup schemes which contain all of the pseudoreflections of \(G\). The Chevalley-Shephard-Todd theorem is generalized (in the case of algebraically closed field) to Theorem 1.3. If \(k\) is algebraically closed, then \(k[V]^G\) is polynomial if and only if \(G\) is generated by pseudoreflections. The author prove Theorem 1.3 in a more general context. The ``if'' part of Theorem 1.3 holds for arbitrary fields which are not algebraically closed. The ``only if'' part does not hold for arbitrary fields and linearly reductive group scheme, it holds for arbitrary fields and a smaller class of group schemes, which are called \textit{stable}; see Definition 1.4 for a precise definition. However, over algebraically closed fields, the class of stable group schemes coincides with that of finite linearly reductive group schemes. See Theorem 1.6 for the generalization of Theorem 1.3. Theorem 1.6 can be generalized to action of a finite linearly reductive group scheme on a smooth scheme. Let \(U\) be a smooth affine scheme, \(x \in U(K)\) a fixed field-valued point. A subgroup scheme \(N\) of \(G\) is a pseudoreflection at \(x\) if \(N_K\) is a pseudoreflection with respect to the induced action of \(G_K\) on the cotangent space at \(x\). Theorem 1.6 thus gives a criterion on the smoothness of an image of \(x\) in \(U/G\). The second main result of the paper is Theorem 1.9: Let \(U\) be a smooth affine \(k\)-scheme with a faithful action by a stable group scheme \(G\) over \(\mathrm{Spec}k\). Suppose \(K/k\) is a finite separable field extension and \(G\) fixes a point \(x \in U(K)\). Let \(M = U/G\), let \(M^0\) be the smooth locus of \(M\), and let \(U^0 = U \times_M M^0\). If \(G\) has no pseudoreflections at \(x\), then after possibly shrinking \(M\) to a smaller Zariski neighborhood of the image of \(x\), we have that \(U^0\) is a \(G\)-torsor over \(M^0\). As an application of Theorem 1.9, the author generalizes a well-known result which says that schemes with quotient singularities prime to the characteristic are coarse spaces of smooth Deligne-Mumford stacks, to schemes which are étale-locally the quotient of a smooth scheme by a finite linearly reductive group scheme. The author shows that every such scheme is the coarse space of a smooth tame Artin stack (as defined by Abramovich, Olsson, and Vistoli).