Topological generation of exceptional algebraic groups

TL;DR

This paper investigates conditions under which tuples of elements from conjugacy classes generate dense subgroups in simple algebraic groups, especially exceptional types, and applies these results to problems in representation theory and finite group generation.

Contribution

It establishes density results for generating subgroups in exceptional algebraic groups and proves a conjecture on random generation of finite exceptional groups.

Findings

Density of generating tuples in exceptional groups for t ≥ 5

Existence of dense generating tuples in G₂ with t ≥ 4

Validation of a conjecture on random (r,s)-generation of finite exceptional groups

Abstract

Let $G$ be a simple algebraic group over an algebraically closed field $k$ and let $C_1, \ldots, C_t$ be non-central conjugacy classes in $G$. In this paper, we consider the problem of determining whether there exist $g_i \in C_i$ such that $\langle g_1, \ldots, g_t \rangle$ is Zariski dense in $G$. First we establish a general result, which shows that if $\Omega$ is an irreducible subvariety of $G^t$, then the set of tuples in $\Omega$ generating a dense subgroup of $G$ is either empty or dense in $\Omega$. In the special case $\Omega = C_1 \times \cdots \times C_t$, by considering the dimensions of fixed point spaces, we prove that this set is dense when $G$ is an exceptional algebraic group and $t \geqslant 5$, assuming $k$ is not algebraic over a finite field. In fact, for $G=G_2$ we only need $t \geqslant 4$ and both of these bounds are best possible. As an application, we show that many faithful representations of exceptional algebraic groups are generically free. We also establish new results on the topological generation of exceptional groups in the special case $t=2$, which have applications to random generation of finite exceptional groups of Lie type. In particular, we prove a conjecture of Liebeck and Shalev on the random $(r,s)$-generation of exceptional groups.