On the (non-)contractibility of the order complex of the coset poset of a classical group. (Q661895): Difference between revisions

Let \(G\) be a finite group. There is a zeta function \(P_G(s)\) associated with \(G\), with the property that for all \(t\in\mathbb N\), \(P_G(t)\) is the probability that \(t\) random elements of \(G\) generate \(G\). It is defined as \(P_G(s)=\sum_{H\leq G}\mu_G(H)/|G:H|^s\) where \(\mu_G\) is the Möbius function on the subgroup lattice of \(G\). In a 2000 paper [J. Algebra 225, No. 2, 989-1012 (2000; Zbl 0973.20016)], \textit{K. S. Brown} attributes to S. Bouc the idea that \(P_G(-1)\) is in fact the reduced Euler characteristic of the order complex of the coset poset of \(G\). The coset poset consists of all cosets of all proper subgroups of \(G\), ordered by inclusion, and the order complex of any poset \(P\) is the simplicial complex whose \(k\)-dimensional faces are the ordered subsets of \(P\) that have size \(k+1\). Brown conjectured that the order complex of the coset poset is never contractible, and in fact that its reduced Euler characteristic is never zero. This fascinating conjecture remains open. In the present paper, the author obtains a significant result in this contest. He proves that if \(G\) is an almost simple finite classical group that does not contain a graph automorphism of its socle, then \(P_G(-1)\neq 0\).