Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

This comes from Artin Second Edition, page 219. Artin defined $G = \langle x,y\mid x^3, y^3, yxyxy\rangle$, and uses the Todd-Coxeter Algorithm to show that the subgroup $H = \langle y\rangle$ has index 1, and therefore $G = H$ is the cyclic group of order 3.

That being the case, $x$ cannot be either $y$ or $y^2$, for then the third relation would not be satisfied. So the relation $x=1$ must follow from the given relations. Is there another way of seeing this besides from the Todd-Coxeter algorithm?

share|cite|improve this question
I think you mean $G/H$ rather than $G=H$? – Matthew Kahle Jan 11 '12 at 22:34
I mean that $H$ is defined as a subgroup of $G$, and in fact, $G = H$. Was that unclear? – Dustan Levenstein Jan 11 '12 at 22:38
@DustanLevenstein: Have you accomplished this problem with Todd-Coxeter by yourself already? I did it using this method and see that there are 2 coincidences and that's why $x$ gets vanished. Would u like see my attempts? However, I see you got the answer greatly long ago. – Babak S. Jul 28 '14 at 23:13
@B.S. I don't have my copy of Artin handy, but I believe when I asked this problem, it was because he reduced it explicitly using Todd-Coxeter, and not by directly manipulating the relations. – Dustan Levenstein Jul 28 '14 at 23:48
@DustanLevenstein: So in this problem you were looking for a another way than TC approach? Anyway, I am glad if I could post a TC approach here. :-) – Babak S. Jul 29 '14 at 0:01
up vote 21 down vote accepted

Let's see. We have $yxyxy=1$, so (multiplying by $y$ on the left) $y=y^2xyxy$, so (cancelling $y$ on the right) $y^2xy=x^{-1}$.

Also, $yxyxy=1$, so $yxyxy^2=y$, or $xyxy^2=1$. So $yxy^2=x^{-1}$.

It follows that $y^2xy=yxy^2$, or $yx=xy$. (So the group is Abelian.)

But then $1=yxyxy=x^2y^3=x^2$. Since $x^3=1$ as well, we finally conclude $x=1$.

share|cite|improve this answer
yeah, I definitely knew it would be sufficient to show that $x$ and $y$ commute. Nicely done! – Dustan Levenstein Jan 11 '12 at 23:24

Here's one way -- quite a bit ad hoc.

The basic idea is write all the elements in terms of $z := yx$. From the third relation, we can see that $z^2 y = 1$, or $z^2 = y^{-1} = y^2$. Therefore, $y = y^4 = z^4$. Now, we can also write $x$ in terms of $z$: $$x = y^{-1} z = y^2 \cdot z = z^8 \cdot z = z^9 .$$ Now $x$ and $y$ commute, both being powers of $z$. (It is a simple exercise to show that $x = 1$ from this. The last line in Andres' answer explains this.)

Here's an alternative approach, which is what I originally followed. Armed with these two identities, we can rewrite all three given relations entirely in terms of $z$:

  • $(z^{9})^3 = 1$;
  • $(z^{4})^3 = 1$; and
  • $z^2 \cdot z^4 = 1$.

From these observations, since $\gcd(27, 12, 6)=3$, we get that $z^3 = 1$. Finally, plugging this back, we obtain $x = z^9 = 1$ and $y = zx^{-1} = z$.

share|cite|improve this answer

I computed the powers of $z = xy$:

$xy$, $y^2$, $x$, $y$, $xy^2$, $x^2$, $xy$, $y^2$, $1$.

This in particular gives $z^7 = z$ so $z^6 = x^2 = 1$, and $x^3 = x^2 = 1$ gives $x = 1$.

All you need from this though is

$z^2 = y^2$
$z^3 = x$

So $1 = y^6 = (z^2)^3 = (z^3)^2 = x^2$, then the last step of $x^3 = x^2 = 1 $ implies $x = 1$.

I posted this solution since the idea is just simplying that "playing around" with powers of $xy$ is algebraically easy and lets you do stuff. So I think computing up to its "period" of 9 is more insightful than just listing the final result.

share|cite|improve this answer

In other words, assuming that $x^3=y^3=yxyxy=e$, the goal is to prove that $x=e$.

Note that $xyx=y^2(yxyxy)y^2=y^4=y$ hence $xy=(xyx)x^2=yx^2$ $(*)$.

Imagine one wants to carry every $y$ in $x=xy^3$ to the leftmost end of the product. Using $(*)$ twice, one first gets $$ x=xy^3=(xy)y^2=(yx^2)y^2=yx(xy)y=yx(yx^2)y=y(xy)x(xy), $$ and, again using $(*)$ twice, $$ x=y(yx^2)x(yx^2)=y^2x^3yx^2=y^3x^2=x^2. $$ Thus, $x=x^2$ and $x=e$.

share|cite|improve this answer

Applying the way Artin used, you see from the following TC algorithm that $[G:H]=1$.

enter image description here

share|cite|improve this answer

$e=y^{-1}(yxyxy)y=xyxy^{-1}$ so $y=xyx$ and $xy=yx^{-1}$; therefore $e=y^3=(xyx)^3=xyx^{-1}yx^{-1}yx=xxyxyyx=x^{-1}yxy^{-1}x$. Cancelling on the left and right yields $x=e$ and the result follows.

share|cite|improve this answer

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

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