2
$\begingroup$

Let $G$ be a finite abelian group and $x,y \in G$ so, such that the order of $x$ and $y$ is coprime. Then the order of the element $xy$ is the product of the orders of $x$ and $y$.

Thoughts. Let $p$ be the order of $x$ and $q$ be the order of $y$. Then all the elements of $\langle xy \rangle$ have the form $x^my^n$ for $m,n \in \mathbb{N}$. Obviously, $(xy)^{pq} = 1$. Thus the order of $xy$ divides $pq$. I hope this is right. Now I would like to use the coprimeness of $p$ and $q$ which means $\mathrm{lcm}(p,q) = pq$. Does from there the statement follow?

$\endgroup$
  • 1
    $\begingroup$ In fact, all elements of $\langle xy \rangle$ have the form $(xy)^n = x^ny^n$ $\endgroup$ – Omnomnomnom Oct 10 '16 at 14:32
3
$\begingroup$

Put $o(x)=m$, $o(y)=n$ and $o(xy)=k$. Since $G$ is abelian, we have $(xy)^{mn}=x^{mn}y^{mn}=(x^m)^{n}(y^n)^m=1$. Hence $k \mid mn$. Since $gcd(m,n)=1$, the subgroups $\langle x \rangle$ and $\langle y \rangle$ have trivial intersection (if $g \in \langle x \rangle \cap \langle y \rangle$, then $o(g)$ divides both $m$ and $n$, so $g=1$). Now assume that $o(xy)=k$, then $x^ky^k=1$, whence $x^k=y^{-k} \in \langle x \rangle \cap \langle y \rangle$. So $x^k=1=y^{-k}$,which means that $m \mid k$ and also since $y^k=1$, $n \mid k$. But $m$ and $n$ are relatively prime, so also $mn=lcm(m,n) \mid k$, and you are done.

$\endgroup$
  • $\begingroup$ Thanks for this really nice answer. $\endgroup$ – TheGeekGreek Oct 10 '16 at 21:30
  • $\begingroup$ You are welcome! $\endgroup$ – Nicky Hekster Oct 10 '16 at 21:48
1
$\begingroup$

Note that since $p$ and $q$ are relatively prime, there exists a solution $n$ to the equation $$ \begin{cases} n \equiv 0 \pmod p\\ n \equiv 1 \pmod q \end{cases} $$ so that $(xy)^n = y$. Similarly, there is an $n$ such that $(xy)^n = x$. Since $\langle xy \rangle$ contains elements of order $p$ and $q$, its order must be at least $\operatorname{lcm}(p,q)$.

In particular: if $ap + bq = 1$ for integers $a$ and $b$, then we may take $n = ap$, then $n = bq$.

$\endgroup$
  • $\begingroup$ Um, pretty sure you need to justify that. How about $(xy)^n=e$ if and only if $x^n=y^{-n}$? Then say something about $\langle x \rangle \cap \langle y \rangle$. $\endgroup$ – Dustan Levenstein Oct 10 '16 at 14:37
  • $\begingroup$ @DustanLevenstein fair point; I'll see if I can fix this efficiently $\endgroup$ – Omnomnomnom Oct 10 '16 at 14:39

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.