I don't know whether it's correct but I wish to show that $g_1g_2=g_2g_1 \Leftrightarrow g_1=g_2^{-1}$ for any $g_1\neq g_2$ in $\langle x,y : xyx^{-1}y=e\rangle$

I was trying to work be definition, taking $x^{n_1}y^{m_1}...x^{n_k}y^{m_k}$ , $x^{s_1}y^{t_1}...x^{s_j}y^{t_k}$ and getting a contradiction when assuming:

$x^{n_1}y^{m_1}...x^{n_k}y^{m_k} \cdot x^{s_1}y^{t_1}...x^{s_j}y^{t_j} \cdot y^{-m_1}x^{-n_k}...y^{-m_1}x^{-n_1} \cdot y^{-t_j}x^{-s_j}...y^{-t_1}x^{-s_1}$ is the unit element, by using the properties: $xy=y^{-1}x , xyx^{-1} = y^{-1}$ but I didn't succeed to advance from that point.

  • $\begingroup$ Note that $g_1g_2 = g_2g_1$ always holds for $g_1 = g_2$. $\endgroup$ – TastyRomeo Oct 9 '18 at 9:43

The statement $g_1g_2 = g_2g_1 \implies g_1 = g_2^{-1}$ for any $g_1$, $g_2$ is not correct.

There are a few trivial counterexamples:

  • Setting $g_1 = g_2 = x$, then $g_1g_2 = x^2 = g_2g_1$, but $g_1 = x \neq x^{-1} = g_2^{-1}$, since $x$ is an element of infinite order.
  • Set $g_1 = e$ and $g_2 \neq e$.

For a more non-trivial example: set $g_1 = x^2$ and $g_2 = y$. Since $xy = y^{-1}x$ and $xy^{-1} = yx$, we have that $$g_1g_2 = x^2y = x(xy) = x(y^{-1}x) = (xy^{-1})x = (yx)x = yx^2 = g_2g_1,$$ but $g_1 = x^2 \neq y^{-1} = g_2^{-1}$.


If $xyx^{-1}y=e$ then $y^{-1}x=xy$ (and also $yx=xy^{-1}$), so every element of $G=\langle x,y\mid xyx^{-1}y\rangle$ can be represented by $x^my^n$ for some $m,n\in\mathbb{Z}$. Clearly $G'\leq\langle y\rangle$ (the number of $x$ appearing is invariant), and $y^2=[x,y]\neq e$ so $G'\neq\langle e\rangle$.

In fact it isn't hard to show $G'=\langle y^2\rangle$ by calculating the commutator $[x^my^n,x^{m'}y^{n'}]$.


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.