Let $G$ be a finite non-commutative group of order $N$, and let $x, y \in G$. Let $a$ and $b$ be the orders of $x$ and $y$, respectively. Can we say anything non-trivial about the order of $xy$ in terms of $a$ and $b$? If it helps, you may assume that $a$ and $b$ are coprime.

  • $\begingroup$ No, it is more-or-less answered here: mathoverflow.net/questions/130697/… $\endgroup$
    – Bugs Bunny
    Mar 31, 2020 at 6:57
  • $\begingroup$ Only in the comments, to be fair... $\endgroup$
    – Bugs Bunny
    Mar 31, 2020 at 6:59
  • 8
    $\begingroup$ To be more explicit: for all three positive integers $a,b,c$, there exist a finite group and elements $x,y$ such that $x,y,xy$ have order $a,b,c$, with the only trivial restriction: if one of $a,b,c$ is $1$, then the other two are equal. $\endgroup$
    – YCor
    Mar 31, 2020 at 7:56
  • 2
    $\begingroup$ Let me also mention that the case when $a,b,c$ are all even is trivial: one can then find three such elements in the direct product of three dihedral groups $D_a\times D_b\times D_c$, each with generators of order two $u,v$, namely $x=(uv,u,v)$, $y=(v,uv,u)$, $xy=(u,v,vu)$. $\endgroup$
    – YCor
    Mar 31, 2020 at 10:35
  • $\begingroup$ Given $a,b,c\ge 2$, estimating/computing the smallest size $n$ such that there is such a pair in the symmetric group $S_n$, looks like an interesting problem. For $a,b,c$ even, the above gives an upper bound of $a+b+c$. $\endgroup$
    – YCor
    Mar 31, 2020 at 10:42

2 Answers 2


The following theorem (which does not take the order $N$ of the group $G$ into account) shows that all possible combinations of $a$, $b$ and the order of $xy$ are possible. See Theorem 1.64 from Milne's course notes on group theory.

For any integers $a,b,c > 1$, there exists a finite group $G$ with elements $x$ and $y$ such that $x$ has order $a$, $y$ has order $b$, and $xy$ has order $c$.

  • $\begingroup$ :what can we say about such a group,....it was asked in exam that "consider such group with above property,then G, (1)cannot be cyclic,(2) G has to non abelion......which one is true,1) or 2)? $\endgroup$ Dec 25, 2021 at 10:23

I strongly suspect that the answer is "very little", unless you are looking for bounds on the order of $xy$ in terms of $N$ or something like that.

Example. Let's focus on the example $a = 2$ and $b = 3$. It is well-known that \begin{align*} \mathbf Z/2 * \mathbf Z/3 &\stackrel\sim\to \operatorname{PSL}_2(\mathbf Z)\\ x &\mapsto \begin{pmatrix}0 & -1 \\ 1 & 0\end{pmatrix},\\ y &\mapsto \begin{pmatrix}0 & 1 \\ -1 & -1\end{pmatrix}, \end{align*} where $x$ and $y$ are the generators with $x^2 = 1$ and $y^3 = 1$. The product $xy$ maps to $$\begin{pmatrix}0 & -1 \\ 1 & 0\end{pmatrix}\begin{pmatrix}0 & 1 \\ -1 & -1\end{pmatrix} = \begin{pmatrix}1 & 1 \\ 0 & 1\end{pmatrix},$$ which has infinite order. For any $n \in \mathbf N$, the surjection $$\operatorname{PSL}_2(\mathbf Z) \twoheadrightarrow \operatorname{PSL}_2(\mathbf Z/n)$$ gives a finite group of order roughly $n^3$ in which $xy$ has exact order $n$.

Example. Another example, still with $a = 2$ and $b = 3$, is the group $$G = \mathbf Z/n \wr S_3 = (\mathbf Z/n)^3 \rtimes S_3,$$ where multiplication is defined by $$(a_1,a_2,a_3,\sigma)(b_1,b_2,b_3,\tau) = (a_1+b_{\sigma(1)},a_2+b_{\sigma(2)},a_3+b_{\sigma(3)},\sigma\tau).$$ Then one easily checks that $x = (1,-1,0,(12))$ has order $2$ and $y = (0,0,0,(123))$ has order $3$, and $xy = (1,-1,0,(23))$ has order $2n$. This time we get any even number as the order of $xy$, inside a group of order $6n^3$.


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.