9
$\begingroup$

Let $G$ be a group generated by $a_0, a_1, a_2$ with relations:

$a_0 a_1 a_0^{-1}=a_1^4$

$a_1 a_2 a_1^{-1}=a_2^4$

$a_2 a_0 a_2^{-1}=a_0^4$

I am wondering if $BS(1,4)=\langle a,b:bab^{-1}=a^4\rangle$ is embedded into G via $a\mapsto a_1$, $b\mapsto a_0$

Remark: the group is constructed in analogy to Higman group

$\endgroup$
7
  • $\begingroup$ The preview and the actual parser treat html tags differently; I've substituted with \langle and \rangle. $\endgroup$ Nov 7, 2010 at 0:48
  • $\begingroup$ Please check, I've edited the question $\endgroup$ Nov 7, 2010 at 0:49
  • $\begingroup$ In the standard Higman group (the one without finite quotients), there are 4 generators. You really want to have only 3? $\endgroup$
    – user6976
    Nov 7, 2010 at 0:58
  • $\begingroup$ @Mark: yes, I really want to have 3 relations. For 4 the answer is "yes". $\endgroup$ Nov 7, 2010 at 1:03
  • $\begingroup$ @Jon: the group is a possible candidate for example of non-sofic group $\endgroup$ Nov 7, 2010 at 1:07

2 Answers 2

8
$\begingroup$

This is a fundamental group of a graph of groups with B-S groups as vertex groups and cyclic groups as edge groups. So yes, each Baumslag-Solitar group naturally embeds in your group.

Edit. It is not a graph of groups but a complex of groups: triangle with B-S groups at vertices, cyclic groups as edges and 1 in the triangle. So you need to check the Haefliger condition to prove that it is developable. One can also (better) use Stallings, John R. Non-positively curved triangles of groups. Group theory from a geometrical viewpoint (Trieste, 1990), 491–503, World Sci. Publ., River Edge, NJ, 1991.

Edit 2. No, this method does not work. To compute the Gersten-Stalling angle, one needs to count only the alternating length, not the total length of the word. The alternating length of each relator is not 7 but 4, so each angle is $\pi/2$ and the total angle is $3\pi/2$, too large.

Edit 3. In fact the group is finite, so none of the B-S groups embeds. See this paper (the link was sent to me by Dani Wise).

$\endgroup$
4
  • $\begingroup$ Thank you for your suggestions. I have to note that the following: $a_0 a_1 a_0^{-1}=a_1^2$; $a_1 a_2 a_1^{-1}=a_2^2$; $a_2 a_0 a_2^{-1}=a_0^2$ is trivial group, so it is essential that we have power 4. It might be that 3 will also work $\endgroup$ Nov 7, 2010 at 1:52
  • $\begingroup$ You need to compute the Gersten-Stallings angles at each corner of the triangle. For example two sides have $\langle a_0\rangle$ and $\langle a_1 \rangle$ on the edges. Then compute the shortest word in $a_0, a_1$ that is equal to 1. In your case, I am sure, it is 7 (the length of the defining relation involving these two letters). Hence the angle is $2\pi/7$. The sum of all three angles is then $6\pi/7\lt \pi$. Hence by the Gersten-Stallings-Haefliger... the complex of groups is developable, and the B-S group embeds. $\endgroup$
    – user6976
    Nov 7, 2010 at 1:56
  • $\begingroup$ Note that in your other example with $2$ instead of $4$, the angles are $2\pi/5$, the sum is $6\pi/5\gt \pi$ and the theorem does not apply. $\endgroup$
    – user6976
    Nov 7, 2010 at 1:58
  • $\begingroup$ @Mark: thank you so much for the paper! $\endgroup$ Nov 7, 2010 at 4:31
3
$\begingroup$

A few minutes of running GAP shows that the group has order $6751269$. Hence, as the final edit of the accepted answer points out, the group is finite, and no Baumslag-Solitar group embeds in it.

$\endgroup$
2
  • $\begingroup$ $3^97^3$... I wonder if this is an easily characterizable group? $\endgroup$ May 19, 2021 at 17:38
  • $\begingroup$ @StevenStadnicki All I know for sure (right now) is that it is solvable :-) $\endgroup$ May 19, 2021 at 18:57

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.