The solution to a conjecture of Tits on the subgroup generated by the squares of the generators of an Artin group (Q5944967)

Given a finite set \(S\) and, for each pair \((s,t)\in S\times S\) a number \(m_{s,t}=m_{t,s}\in\{1,2,3,\dots,\infty\}\) (with \(m_{s,t}=1\) iff \(s=t\)), the group \(A\) with presentation \[ \langle S\mid stst\cdots=tsts\cdots\text{ for each }(s,t)\in S\times S\rangle \] (where the expressions \(stst\cdots\) and \(tsts\cdots\) each have length \(m_{s,t}\)) is called an `Artin group'. The set \(S\) is called the `standard generating set' of \(A\). The classical braid groups are examples of Artin groups. Let \(A\) be an Artin group with standard generating set \(S\). For each \(s\in S\), let \(T_s=s^2\), and let \(H\) be the subgroup of \(A\) generated by \(\{T_s\mid s\in S\}\). It was conjectured by Tits (as reported by \textit{K. I. Appel} and \textit{P. E. Schupp} [Invent. Math. 72, 201-220 (1983; Zbl 0536.20019)]) that the set \(\{T_sT_t=T_tT_s\mid(s,t)\in S\times S,\;m_{s,t}=2\}\) forms a complete set of defining relations for \(H\). This so-called `Tits conjecture' was subsequently proved in a number of special cases by various authors. In the present work, the authors prove the Tits conjecture for all Artin groups. In fact, they prove the following stronger result: for each \(s\in S\), let \(n_s\geq 2\) be an integer, and let \(R_s=s^{n_s}\). Then the set \(\{R_sR_t=R_tR_s\mid(s,t)\in S\times S,\;m_{s,t}=2\}\) forms a complete set of defining relations for the subgroup \(G\) of \(A\) generated by the \(R_s\). The result is proved first for Artin groups of `small type' (i.e., where each \(m_{s,t}\leq 3\)), by means of a representation of \(G\) in the mapping class group of a surface. The general case is then handled by mapping an arbitrary Artin group into one of small type, and studying the image of \(G\) under this map.