Skip to main content
19 events
when toggle format what by license comment
Jul 26, 2021 at 8:35 vote accept AGenevois
Jul 26, 2021 at 8:35 answer added AGenevois timeline score: 1
Jul 14, 2021 at 21:20 comment added HJRW @YCor, Anthony’s answer also answers this subquestion, since all its noncyclic quotients are dihedral. I suspect the fundamental group of the surface of Euler characteristic -1 also fails to be residually torsion-free nilpotent, but don’t know a proof.
Jul 14, 2021 at 19:39 answer added AGenevois timeline score: 7
Jul 13, 2021 at 6:59 comment added YCor @MarkSapir isn't the (sub)question rather whether torsion-free f.g. subgroups of RACG residually (torsion-free nilpotent)?
Jul 13, 2021 at 6:14 answer added HJRW timeline score: 11
Jul 12, 2021 at 18:58 comment added AGenevois @MattZaremsky: Thank you. In fact, RAAGs satisfy a strong Tits alternative: every finitely generated subgroup either is abelian or it surjects onto $\mathbb{F}_2$. For RACGs, you have something similar, every finitely generated subgroup either is virtually abelian or it surjects onto a non-elementary free product of $\mathbb{Z}$ and $\mathbb{Z}/2\mathbb{Z}$. So you get a similar obstruction: for finitely generated subgroups, the abelianisation cannot be cyclic.
Jul 12, 2021 at 18:11 comment added Matt Zaremsky @MarkSapir: That sounds right, so I think my one trick for seeing something doesn't embed in a RAAG also makes it not embed in a RACG, and hence is not much help for this question.
Jul 12, 2021 at 16:48 comment added markvs @MattZaremsky: Aren't RACG residually nilpotent? Also see link.springer.com/content/pdf/10.1007/s00013-019-01332-7.pdf, Theorem 1 (ii).
Jul 12, 2021 at 16:12 comment added Matt Zaremsky @AGenevois: The proof I know that $B_4$ (even $B_3$) doesn't embed into any RAAG (as told to me by Thomas Koberda) is that thanks to residual torsion-free nilpotence, any non-abelian subgroup of a RAAG must surject onto $\mathbb{Z}^2$, and of course braid groups can't do this.
Jul 12, 2021 at 16:03 comment added YCor Sorry for adding mess, indeed my comment was irrelevant since $B_4$ is not a RAAG.
Jul 12, 2021 at 15:49 comment added AGenevois @MattZaremsky: I see. Is it easy to prove that $B_4$ does not embed in a RAAG?
Jul 12, 2021 at 15:33 comment added Matt Zaremsky I think @YCor's point was that $B_4$ can't embed in a RAAG, so one could wonder whether it embeds in a RACG and hence serves as a counterexample, but "probably" $B_4$ also can't embed in a RACG. (Actually all of this could be said already for $B_3$.) (For the record I can't think of any other candidate counterexamples.)
Jul 12, 2021 at 15:05 comment added AGenevois @YCor: I don't see the connection between your comment and my question. Am I missing something?
Jul 12, 2021 at 15:03 comment added AGenevois @MarkSapir: Sure, but that only proves that torsion-free subgroups of RACGs are virtually subgroups of RAAGs. My question asks precisely whether or not the virtually can be removed.
Jul 12, 2021 at 14:13 comment added YCor I guess the braid group $B_4$ is not isomorphic to any subgroup of any Coxeter group. (If it were, we would deduce its linearity, which was hard to establish. Anyway that's not a proof).
Jul 12, 2021 at 14:12 comment added YCor f.g. RAACG are virtually torsion-free. But not orbitrary RAACG, since these include the group of finitely supported permutations of the integers.
Jul 12, 2021 at 13:37 comment added markvs RAACGs are not virtually torsion-free? Did you see this: core.ac.uk/download/pdf/82768543.pdf
Jul 12, 2021 at 13:10 history asked AGenevois CC BY-SA 4.0