10
$\begingroup$

$\newcommand{\orb}{\mathrm{orb}}$Let $T$ ($K$) be the torus (Klein bottle) with one cone point of order $q\geq 2$. The presentation of their orbifold fundamental groups are easy to find. Namely,

$$\pi_1^{\orb}(T)=\{a,b\ |\ (aba^{-1}b^{-1})^q=1\}.$$

$$\pi_1^{\orb}(K)=\{a,b\ |\ (aba^{-1}b)^q=1\}.$$

Want to know the centre of these two groups. Are they trivial?

Improve this question
$\endgroup$

3 Answers 3

Reset to default
9
$\begingroup$

These groups have trivial centers. As one proof, they are both fuchsian and so embed in $\mathrm{PSL}(2, \mathbb{R})$ (well, the orientation preserving subgroups do). However, elements of $\mathrm{PSL}$ that commute must have common fixed points on the "circle at infinity" of the hyperbolic plane.

Improve this answer
$\endgroup$
1
  • 3
    $\begingroup$ Thanks a lot Sam! I was expecting it to be true. For more than one cone points case, I have a direct proof of the triviality of the centre, but this case was troubling. $\endgroup$
    – RKS
    Nov 21, 2022 at 17:52
7
$\begingroup$

The groups have trivial center, as pointed out by Sam Nead. Another, more combinatorial, way to show this is to apply the algorithm from [1], which decides whether any given one-relator group has a non-trivial center or not. If one wishes to be very explicit, one can even compute an explicit generating set for a finite index surface group inside the groups, see [2], particularly §5, where the Reidemeister-Schreier method is used to find a one-relator presentation on $4q-2$ generators for such a finite index subgroup.

${}$

${}$

[1] Baumslag, G.; Taylor, Tekla, The centre of groups with one defining relator, Math. Ann. 175, 315-319 (1968). ZBL0157.34901.

[2] Baumslag, Gilbert; Troeger, Douglas, Virtually free-by-cyclic one-relator groups. I., Fine, Benjamin (ed.) et al., Aspects of infinite groups. A Festschrift in honor of Anthony Gaglione. Papers of the conference, Fairfield, USA, March 2007 in honour of Anthony Gaglione’s 60th birthday. Hackensack, NJ: World Scientific (ISBN 978-981-279-340-9/hbk). Algebra and Discrete Mathematics (Hackensack) 1, 9-25 (2008). ZBL1188.20023.

Improve this answer
$\endgroup$
7
  • 1
    $\begingroup$ Thanks Carl for the detailed reply. I will look at the references. $\endgroup$
    – RKS
    Nov 21, 2022 at 17:53
  • 1
    $\begingroup$ This is a sledgehammer to crack a nut! Plus note that Sam Nead's argument works verbatim for all hyperbolic orbifolds. $\endgroup$
    – HJRW
    Nov 22, 2022 at 8:52
  • 1
    $\begingroup$ @HJRW Agreed on the sledgehammer (for the first part!), but it is such a nice paper that it is worth reading anyway :-) For the second part, it is nice to see more explicit generating sets, I think. $\endgroup$
    – Carl-Fredrik Nyberg Brodda
    Nov 22, 2022 at 9:10
  • 2
    $\begingroup$ @Carl-FredrikNybergBrodda: Indeed! My point is that the judgement of both Gromov and posterity is that Dehn was right. I persist with this point because it's still a serious issue for the field: I regularly see people working on the combinatorial side exaggerating the difficulty of results that are trivial from a geometric point of view. $\endgroup$
    – HJRW
    Nov 22, 2022 at 11:56
  • 2
    $\begingroup$ @HJRW Agreed! (For the most part, with the remainder coming down to my ignorance; I confess I may at times be one of those people you speak of, as much as I try to avoid it...). $\endgroup$
    – Carl-Fredrik Nyberg Brodda
    Nov 22, 2022 at 12:07
2
$\begingroup$

The centre of any non-cyclic one-relator group with torsion is trivial.

Moreover,

the centraliser of any non-identity element of a one-relator group with torsion is cyclic.

This is B. B. Newman’s theorem (1973).

Improve this answer
$\endgroup$
1
  • 1
    $\begingroup$ Thanks Anton for the precise reference. I find this more general fact interesting. $\endgroup$
    – RKS
    Nov 23, 2022 at 15:11

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.