Discrete groups of slow subgroup growth (Q677419): Difference between revisions

Given a finitely generated, residually finite group \(G\) and a positive function \(f\) on the integers, one says that \(G\) has (subgroup) growth type \(f\), if \[ 0<\limsup_n\frac{\log\#\{\text{subgroups of \(G\) of index }n\}}{\log f(n)}<\infty. \] Groups with growth type \(\leq n\) have polynomial (subgroup) growth. It is known that for linear groups, the slowest possible non-polynomial growth is \(n^{\log n/\log\log n}\) [see \textit{A. Lubotzky}, Invent. Math. 119, No. 2, 267-295 (1995; Zbl 0848.20036)], and this was so far also the slowest nonpolynomial growth known to occur. In their paper, the authors construct an example which has growth type \(n^{\log n/(\log\log n)^2}\), and they conjecture that no slower non-polynomial growth is possible. They also give another example, which has growth type \(n^{\log n}\); this is also the first group known to exhibit this growth type. The examples given by the authors have as their profinite completion essentially an infinite product of pairwise non-isomorphic alternating groups (in the first case) or finite simple groups of Lie type (in the second case). The proof then relies on asymptotic results on the number of subgroups of index \(n\) in alternating groups (resp. finite simple groups of Lie type).