5

In Braid Groups of Kassel, Turaev, it mentions that $\mathcal{B}_n$ is a residually finite group. The definition that they give as a residually finite group is a group $G$ such that for each $g\in G-\{e_G\}$ ($e_G$ the identity of $G$), there exists an homomorphism $f$ to a finite group $H$ such that $f(g)\neq e_H$. My question is:

How can I obtain the group $H$ for a given element $g\in \mathcal{B}_n$ and the homomorphism that fulfills this?

I hope you can help me. Nice Holidays.

iam_agf
  • 5,440
  • 1
  • 19
  • 45
  • You need $H$ to be a *finite* group. – Derek Holt Dec 15 '15 at 19:08
  • Someone explained this to me once. Something along the lines of: "$B_n$ is a subgroup of $\text{Aut}(F_n)$, which is much easier to show is residually finite." I don't remember how to do either half of that argument. –  Dec 15 '15 at 19:18
  • 1
    Well if $g$ is not a pure braid, the obvious quotient map to the symmetric group will do. For pure braids, I assume you'll want to comb the braid. So use the fact that $P_n=F_n\rtimes P_{n-1}$ where $F_n$ is the free group on $n$ generators (which is residually finite) and then use induction – Dan Rust Dec 15 '15 at 19:18
  • 1
    @DanRust I just posted the same argument. :) – Cheerful Parsnip Dec 15 '15 at 19:19

2 Answers2

2

If the braid is not pure, then you can detect it by a homomrphism to the symmetric group. So this reduces to showing that the pure braid group $P_n$ is residually finite. $P_n$ is an iterated semi-direct product of free groups. (This is called combing the braid.) The result now follows because free groups are residually finite. https://mathoverflow.net/questions/20471/why-are-free-groups-residually-finite

Cheerful Parsnip
  • 26,630
  • 2
  • 61
  • 122
  • By the Symmetric Group which one you mean? $S_n$ for the group $\mathcal{B}_n$? I'm looking for an explicit example of this homomorphisms between a $\mathcal{B}_n$ to that $H$ group mentioned in the question. Can be this possible? – iam_agf Dec 15 '15 at 19:55
  • @MonsieurGalois The group H depends on the element. So, yes, for many elements you can use the homomorphism to $S_n$, but it won't work on the pure braid subgroup, when you have to start using different homomorphisms to other finite groups. – Cheerful Parsnip Dec 16 '15 at 00:37
  • For the group $P_n$ which group(s) can be used? – iam_agf Dec 16 '15 at 00:51
  • @MonsieurGalois: $P_n$ is an iterated semidirect product of free groups. In the link I gave, the groups $H$ are constructed for free groups. – Cheerful Parsnip Dec 16 '15 at 00:53
2

Another possibility is to embed $\mathcal{B}_n$ into the automorphism group $\mathrm{Aut}(\mathbb{F}_n)$ of the free group $\mathbb{F}_n$. Now, Baumslag gave a very short prove of the fact that, for any finitely generated residually finite group $G$, $\mathrm{Aut}(G)$ is also residually finite. The conclusion follows from the residual finiteness of finitely generated free groups (see for example the beautiful proof of Stallings in Topology of finite graphs), since a subgroup of a residually finite group is clearly residually finite itself.

For more details, see Basic results on braid groups and the references therein.

Seirios
  • 32,479
  • 5
  • 76
  • 139