On uniform exponential growth for linear groups. (Q1770276): Difference between revisions

Given a group \(\Gamma\), a subset \(S\), and a natural number \(n\), let \(B_S(n)\) denote the set of products of at most \(n\) elements in \(S\cup S^{-1}\). For any finite \(S\) there is the limit \(\omega_S(\Gamma)=\lim|B_S(n)|^{1/n}\) as \(n\to\infty\). If \(\omega_S(\Gamma)>1\) for a finite generating set, then this is true for every finite generating set and \(\Gamma\) is said to be of exponential growth. If the numbers \(\omega_S(\Gamma)\) are separated from 1, then \(\Gamma\) is said to be of uniform exponential growth. The main result of the paper has the following corollary: For any \(m\) and any finitely generated subgroup \(\Gamma\) of \(\text{GL}_m(\mathbb{C})\) the following 3 conditions are equivalent: (1) \(\Gamma\) is not virtually nilpotent, (2) \(\Gamma\) is of uniform exponential growth, (3) \(\Gamma\) is of exponential growth. -- It had been known before that (2) and (3) are equivalent for some \(\Gamma\) but not equivalent for other finitely generated \(\Gamma\).