Normal subgroup generated by a plane polynomial automorphism (Q601757)

Let \(\mathrm{Aut}[\mathbb{C}^2]\) be the group of polynomial automorphisms of \(\mathbb{C}^2\). This group has a normal subgroup \(G\), which consists of automorphisms having Jacobian determinant~\(1\). So, the interesting question about simplicity and \(\mathbb{C}^2\) it to ask whether \(G\) is a simple group. \textit{V. I. Danilov} [Mat. Zametki 15, 289--293 (1974; Zbl 0309.12104)] proved that \(G\) is not simple. More precisely, he showed that the normal subgroup generated by \(f^{13}\), where \(f(x,y)=(y,-x+3y^5-5y^4)\), is not the whole group \(G\). Using similar methods, the authors refine the method of Danilov, and describe which are the elements of \(G\) which have a normal closure distinct from \(G\), proving that it depends of the length of these in the amalgamated product structure of \(G\). This refinement is new, and the technique is more detailed than in the short note of Danilov, which uses mainly results of \textit{P. E. Schupp} [Math. Ann. 193, 255--264 (1971; Zbl 0209.05204)]. This allows the reader to understand more easily why \(G\) is not simple. Precisely, the authors show that the normal subgroup generated by an element of length at most \(8\) is the whole group \(G\), and that this is false for any \textit{general} element of length at least \(14\). The cases \(10\), \(12\) are still open (note that the minimal length of an element is always even).