Can automorphism groups ever be infinitely generated in more than a finite number of dimensions? (Q1825279)

The authors survey examples of infinite generation of automorphism groups of (finitely generated) free objects. These exhibit interesting behavior typified by the free metabelian groups \(M(n)=F[[F,F],[F,F]]\), where F is the free group of rank n. For \(n=2\) or \(n\geq 4\), Aut(M(n)) is finitely generated, but for \(n=3\) it is not. Another example is that \(GL_ n({\mathbb{Z}}[x,x^{-1},y,y^{-1}])\) is finitely generated except when \(n=2.\) Related to questions of finite generation is the concept of tame elements. An automorphism of X(n) (the free object of rank n in a variety X of groups) is called tame if it lifts to an automorphism of the free group of rank n, and an element of \(GL_ n(R)\) is tame if it is in the subgroup generated by elementary and diagonal matrices. In examples such as \(GL_ 2({\mathbb{Z}}[x,x^{-1},y,y^{-1}])\), there are matrices that are not tame, but under the natural stabilization inclusion into \(GL_ 3({\mathbb{Z}}[x,x^{-1},y,y^{-1}])\), they become tame. Motivated by many examples and results, the authors raise two general questions. First, if the free objects X(n) do not satisfy a power law, e.g. if the X(n) are torsion-free groups, does there exist a critical dimension N depending on X such that Aut X(n) is finitely generated for all \(n\geq N_ X?\) Second, is there a critical dimension above which the automorphism groups become tame? A similar and somewhat more detailed survey by these same authors appears as: The tame range of automorphism groups and \(GL_ n\) [Group Theory, Proc. Conf., Singapore 1987, 241-251 (1989; Zbl 0661.20022)].