Linear stable range for homology of congruence subgroups via FI-modules (Q2323546)

This article is a contribution to the application of functor categories -- more precisely, \(FI\)-modules (that is, functors from the category \(FI\) of finite sets with injections to modules over a fixed commutative ring \(\Bbbk\)) -- to \textit{representation stability} (that is, regularity phenomena in sequences of representations of nice groups) for the homology of congruence subgroups of general linear groups. Its main theoretical result (\textit{Theorem 5}) relates, for a chain complex of \(FI\)-modules, the fundamental homological invariants \(t_i\) of each term of the complex and the ones of its homology. Here, for an \(FI\)-module \(F\), \(t_i(F)\) is an integer with controls the size in homological degree \(i\) of a projective resolution of \(F\). Indeed the only used invariants are \(t_0\) and \(t_1\) (which allow to control all higher \(t_i\) thanks to [\textit{T. Church} and \textit{J. Ellenberg}, Geom. Topol. 21, No. 4, 2373--2418 (2017; Zbl 1371.18012)]), but the proof does need to consider all \(t_i\). Using a result of [\textit{T. Church} et al., Adv. Math. 333, 1-40 (2018; Zbl 1392.15030)], it allows to get the inequalities \[t_0\big(H_k(\mathrm{GL}(R,I);\Bbbk)\big)\le 4k+2d+1\quad\text{and}\quad t_1H_k(\mathrm{GL}(R,I);\Bbbk)\le 4k+2d+6\] where \(R\) is a ring satisfying Bass's stable range condition \(SR_{d+2}\), \(I\) is a two-sided ideal of \(R\), \(\mathrm{GL}_n(R,I)\) is, for each \(n\), the congruence subgroup \(\mathrm{Ker}\,(\mathrm{GL}_n(R)\to \mathrm{GL}_n(R/I))\), giving rise to a \(FI\)-group \(\mathrm{GL}(R,I)\) and so, after taking group homology, to a \(FI\)-module \(H_k(\mathrm{GL}(R,I);\Bbbk)\) for each integer \(k\). These linear bounds improve the quadratic ones of [\textit{T. Church} et al., Adv. Math. 333, 1--40 (2018; Zbl 1392.15030)]. It gives qualitative informations on the symmetric sequences \(\big(H_k(\mathrm{GL}_n(R,I);\Bbbk)\big)_n\) (the symmetric group \(\mathfrak{S}_n\) acting on the homology \(H_*(\mathrm{GL}_n(R,I);\Bbbk)\) through conjugation by permutation matrices). The idea to study the homology of congruence groups as symmetric sequence to catch a part of its structure (the computation being almost always out of reach) is related to the problem of excision in higher algebraic \(K\)-theory and goes back to \textit{R. Charney} [Commun. Algebra 12, 2081--2123 (1984; Zbl 0542.20023)]. It has been studied with the point of view of representation stability since the work of \textit{A. Putman} [Invent. Math. 202, No. 3, 987--1027 (2015; Zbl 1334.20045)]. The bounds given in the article under review are probably close to optimality.