The ideal class group action on \(K\)-groups and homology groups (Q741417): Difference between revisions

Let \(R\) be a subring of some commutative ring \(k\) and \(n>0\) an integer. Denote by \(N_n\) the normalizer of \(\mathrm{GL}_n(R)\) in \(\mathrm{GL}_n(k)\). The conjugation action of \(N_n\) on \(\mathrm{GL}_n(R)\) induces actions of the group \(N_n/(k^\times . \mathrm{GL}_n(R))\) on the integral homologies \(H_*(\mathrm{GL}_n(R))\) and \(H_*(GL(R))\) and on the \(K\)-theory \(K_*(R)\) (in degrees \(*>0\)). The article under review first details the construction and studies these actions in general, showing in particular that the induced action on \(K_*(R)\otimes\mathbb{Q}\) is trivial if the groups \(N_n/(k^\times . \mathrm{GL}_n(R))\) are torsion. After, the paper specialises in the case where \(R\) is a Dedekind domain and \(k\) its fraction field. If moreover \(n\) is an odd prime power, the article [\textit{A. J. Berrick}, J. Reine Angew. Math. 569, 55--101 (2004; Zbl 1051.19001)] constructs an isomorphism between \(N_n/(k^\times . \mathrm{GL}_n(R))\) and the subgroup of the ideal class group \(\text{Cl}(R)\) of elements killed by \(n\). In the article under review, the author gives several results that suggest that this action is not so far from being trivial: it is trivial on \(K_i(R)\) when \(k\) is a global field, or when \(i\) is even and the residue field at each maximal ideal is finite in \(R\). Some results for the action on the homology of linear group on \(R\) are also proven, giving triviality of the action in low degree under assumptions. The article ends with the very natural (open) question to know whether one can find an example where the action is not trivial.