Smallest noncyclic quotients of braid and mapping class groups (Q6142257)

The main results of this paper identify the smallest noncyclic quotients of braid groups and mapping class groups as the obvious candidates. For braid groups, this was conjectured by Margalit (and improves various recent results), and for mapping class groups this reproves (and improves) a result of \textit{D. Kielak} and \textit{E. Pierro} [Groups Geom. Dyn. 14, No. 2, 489--512 (2020; Zbl 1512.57022)] with much simpler methods. To be a bit more precise, the \(n\)-strand braid group has an obvious map to the symmetric group \(S_n\) of order \(n!\) by sending a braid to the induced permutation of endpoints. The first main result (Theorem 1) shows that any non-cyclic quotient of the \(n\)-strand braid group (for \(n=3\) or \(n\geq 5\)) has order at least \(n!\), and the lower bound is only achieved by post-composing the obvious map to the symmetric group \(S_n\) with an automorphism. The mapping class group case (Theorem 4) is similar, with action on mod-\(2\)-homology of the surface taking the place of the symmetric group. For surfaces with punctures and boundary components (Theorem 5) the smaller of the two phenomena happens. The proofs use a clever inductive application of the orbit stabiliser theorem: first conjugate elements \(\rho_i\) of the braid (or mapping class) group are found that are necessarily distinct in non-cyclic quotients (Lemmas 8 and 11). Then, using the orbit stabiliser theorem on the conjugation action, this yields lower bounds on the size of the centraliser of the images of the \(\rho_i\). The centralisers of the \(\rho_i\) in the braid groups are braid groups of lower complexity, then allowing for an inductive argument.