The following quote is an excerpt from an interview with John Conway:

Coxeter came to Cambridge and he gave a lecture, then he had this problem for which he gave proofs for selected examples, and he asked for a unified proof. I left the lecture room thinking. As I was walking through Cambridge, suddenly the idea hit me, but it hit me while I was in the middle of the road. When the idea hit me I stopped and a large truck ran into me and bruised me considerably, and the man considerably swore at me. So I pretended that Coxeter had calculated the difficulty of this problem so precisely that he knew that I would get the solution just in the middle of the road. In fact. I limped back after the accident to the meeting. Coxeter was still there, and I said, "You nearly killed me." Then I told him the solution. It eventually became a joint paper. Ever since, I've called that theorem "the murder weapon." One consequence of it is that in a group if $a^2 = b^3 = c^5 =(abc)^{-1}$, then $c ^{610} =1$.

Does anyone know where one can find a statement of this theorem or the publication information on the paper Conway and Coxeter wrote?

This quote is from Math. Intelligencer 23 (2001), no. 2,pp.8-9.

  • 48
    $\begingroup$ obligatory xkcd comic $\endgroup$ Sep 29 '17 at 6:35
  • 13
    $\begingroup$ No offense, but mathematicians should be worth less points. $\endgroup$ Sep 29 '17 at 15:50

According to Siobhan Roberts (2006) King of Infinite Space: Donald Coxeter, the Man Who Saved Geometry p.359 (Google Books):

This is a simple statement of Coxeter's “Murder Weapon”: If “$A^p=B^q=C^r=ABC=1$” defines a finite group, then “$A^p=B^q=C^r=ABC=Z$” implies $Z^2=1$. Conway; and Conway, interview, November 26, 2005; Conway with Coxeter and G. C. Shepherd, “The Centre of a Finitely Generated Group,” in Tensor, 1972 (contains the proof for “The Murder Weapon”).

The latter reference apparently means this citation: http://www.ams.org/mathscinet-getitem?mr=333001

MR333001 20F05

Conway, J. H.; Coxeter, H. S. M.; Shephard, G. C. The centre of a finitely generated group. Commemoration volumes for Prof. Dr. Akitsugu Kawaguchi's seventieth birthday, Vol. II. Tensor (N.S.) 25 (1972), 405–418; erratum, ibid. (N.S.) 26 (1972), 477.

Review: https://zbmath.org/0236.20029, written by Coxeter:

The polyhedral group $(l,m,n)$, defined by $a^l=b^m=c^n=abc=1$, is finite if and only if $g>0$, where $g=2(l^{-1}+m^{-1}+n^{-1}-1)^{-1}$, and then its order is $g$. The same condition for finiteness applies also to the binary polyhedral group $\langle l,m,n\rangle$, defined by $a^l=b^m=c^n=abc$, whose order is $2g$. In fact, the central element $z=a^l=b^m=c^n=abc$ is of period $2$. This last statement is not at all obvious. The first step is to observe that the Cayley diagram for $(l,m,n)$ is identical with the coset diagram for $\langle l,m,n\rangle$ relative to the subgroup generated by $z$. This diagram covers the inversive plane, or the $2$-sphere, with a "map" which consists of $g/l$ $l$-gons representing $a^l$, $g/m$ $m$-gons representing $b^m$, $g/n$ $n$-gons representing $c^n$, and $g$ negatively oriented triangles representing $abc$. The procedure is to compare the results of shrinking the peripheral circuit (or any other circuit) to a single point (or "point-circuit") in two distinct ways, corresponding to the two discs into which the circuit decomposes the $2$-sphere. More symmetrically, any point-circuit (such as the point at infinity of the inversive plane) may be continuously deformed into another point-circuit by contracting over all the oriented polygons in the Cayley diagram. The same procedure is applied in more complicated cases. For instance, in the group for which $z$ is central while $a^l=z^p$, $b^m=z^q$, $c^n=z^r$, $abc=z^s$ and $g$ (as defined above) is a positive integer, the period of $z$ is $g|pl^{-1}+qm^{-1}+rn^{-1}-s|$.


A Google search for "Conway and Coxeter murder" came up with the book "King of Infinite Space: Donald Coxeter, the Man Who Saved Geometry".

A search via Google inside this book came up with this on page 359, note 14:

This is a simple consequence of Coxeter's "murder weapon": "If $a^p = b^q = c^r = abc = 1$ defines a finite group, then $a^p=b^q=c^r = abc = z$ implies $z^2 = 1$."

The reference is Conway, Coxeter, and Shepherd, Tensor, 1972.

  • 2
    $\begingroup$ Correct me if I'm wrong, but wouldn't $a^p = b^q = c^r = abc = 1$ and $a^p=b^q=c^r = abc = z$ together imply $z = 1$ trivially? $\endgroup$
    – Arthur
    Sep 29 '17 at 7:07
  • 7
    $\begingroup$ @Arthur The statement means if the group with presentation $\langle a,b,c\,|a^p=b^q=c^r=abc=1\rangle$ is finite, then for any group $G$ and for $w,x,y,z \in G$ if $w^p=x^q=wxy=z$, then $z^2=1$. $\endgroup$
    – Nex
    Sep 29 '17 at 7:15
  • 8
    $\begingroup$ @Nex Aah, so the two equalities are not in the same group. A bit confusing that they're using the same letters, but sure. At least that explains why it's not trivial. $\endgroup$
    – Arthur
    Sep 29 '17 at 7:22
  • $\begingroup$ When you wrote "This is a simple consequence", did you mean to write "This is a simple statement"? $\endgroup$ Sep 29 '17 at 16:24
  • 1
    $\begingroup$ I was quoting the book. $\endgroup$ Sep 29 '17 at 16:33

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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