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

I want to show $A_n$ can be defined by the following relations $x_1,x_2,\dots,x_{n-2}$: $$ x_1^3;\qquad x_i^2, i>1;\qquad (x_ix_{i+1})^3;\qquad (x_ix_j)^2, j>i+1. $$ I assume $n\geq 3$, and let $x_1=(123)$, $x_i=(12)(i+1,i+2)$ for $2\leq i\leq n-2$. These satisfy the given relations, so $A_n$ is the homomorphic image of $FG^{(n-2)}/K$, where $K$ is the normal subgroup generated by the realtions and $FG^{(n-2)}$ the free group generated by $x_1,\dots,x_{n-2}$.

Now $x_1^3\in K$, and so $x_1^3=1$ in $FG^{(1)}/K$, thus $|FG^{(1)}/K|\leq \frac{3!}{2}=3$. By induction I assume $$ |FG^{(n-3)}/K|\leq \frac{(n-1)!}{2}. $$ If $H$ is the subgroup of $FG^{(n-2)}/K$ generated by $x_1,\dots,x_{n-3}$, I claim $$ FG^{(n-2)}/K=H\cup Hx_{n-2}\cup Hx_{n-2}x_{n-3}\cup\cdots\cup Hx_{n-2}\cdots x_2x_1\cup Hx_{n-2}\cdots x_2x_1^2. $$ This follows since the last relation implies $x_ix_jx_ix_j=1$, or $x_ix_j=x_jx_i$ for $j>i+1$, and the cubed relation implies $$ x_{n-3}x_{n-2}x_{n-3}=x_{n-2}x_{n-3}x_{n-2}. $$ So for any $g\in FG^{(n-2)}/K$, I can write $g=hx_{n-2}x_{j_1}\cdots x_{j_m}$, for $h\in H$. So if $j_1\neq {n-3}$, I can commute to the left of $x_{n-1}$. If $j_1=n-3$, then I look to $j_2$, if $j_2=n-2$, then I can use the second relation above to decrease the number of elements to the right of the first $x_{n-2}$, etc. Either way, it is clear that $g$ has is in one of the cosets of $H$ listed above. So $$ |FG^{(n-2)}/K|\leq (n-1)\frac{(n-1)!}{2}+\frac{(n-1)!}{2}=\frac{n!}{2}. $$

This would show $A_n\simeq FG^{(n-2)}/K$ if the homomorphism is surjective, since I would have an epimorphism from a group of order $\leq n!/2$ onto a group of order $n!/2$. Is the induced homomorphism surjective? I know $A_n$ is generated by $3$-cycles , and even just cycles of form $(1ij)$, but I can't figure a way to get these with the $x_i$ I chose.

Source: Jacobson's Basic Algebra I, pp. 71, #5.

share|cite|improve this question
You can show that $A_n$ is generated by the permutations $x_1,\ldots,x_{n-2}$ by induction. I's true for $n=3$, and to go from $n-1$ to $n$ just show that the generated group is transitive and use the Orbit Stabilizer Theorem. – Derek Holt May 31 '12 at 21:18
Let $G$ be the subgroup of $A_n$ generated by $x_1,\ldots,x_{n-2}$. By inductive assumption, $x_1,\ldots,x_{n-3}$ generate $A_{n-1}$, which has order $(n-1)!/2$, and this subgroup stabilizes the point $n$. It is easy to see that, by using combinations of $x_1,\ldots,x_{n-2}$, we can map the point $n$ to any of the points $1,2,\ldots,n$. In other words $G$ acts transitively on $\{1,2,\ldots,n\}$. So the order of $G$ is $n$ times the order of the stabilizer of the point $n$; i.e. $|G| = n \times (n-1)!/2 = n!/2$, so $G=A_n$. – Derek Holt May 31 '12 at 21:55
I understand now, many thanks @DerekHolt! – Adelaide Dokras May 31 '12 at 21:59
up vote 2 down vote accepted

The cicles $s_i = (i\ i+1)$ generate $S_n$. The function $\mathrm{sign}\colon S_n\to \mathbb{Z}_2$, with kernel $A_n$, sends every generator into $1$. So $\ker(\mathbb{sign})\cong A_n$ is generated by the products $s_is_j$.

Now $s_1s_2 = (1\ 2)(2\ 3) = (1\ 2\ 3) = x_1$. Similarly $s_1s_i = x_{i-1}$ and $s_is_j = x_{i-1}x_{j-1}$. So the set of generator of your presentation comes from the presentation of $S_n$ as a Coxeter group.

The relations of the $S_n$ presentation are $s_i^2$, $(s_is_{i+1})^3$ and $(s_is_j)^2$ in the others cases.

Let's $G$ be a group with your presentation. We can map the generator into the element $x_i = s_1s_{i+1}$ of $S_n$. From the theory of Coxeter groups we know that the function is indeed injective and well defined. So $G\cong f(G)$ where $f$ is the homomorphism defined before. But we have already proved that $f(G) = A_n$.

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.