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've been reading this article (, page 7, paragraph 2) about a generalized Goursat lemma and in the article the author determines the cyclic subgroups of $A \times B$. In the proof he makes the following statement (sort of):

Suppose we are given finite cyclic groups $\overline{G}_1, \overline{G}_2$ of orders $m=m_1d$ and $n=n_1d$, where $d=gcd(m, n)$. Let $G_1, G_2$ be subgroups of $\overline{G}_1, \overline{G}_2$ of coprime orders $m_1, n_1$. Suppose there exists an isomorphism $\theta: \overline{G}_1/G_1\rightarrow \overline{G}_2/G_2$. If $\alpha$ is a generator of $\overline{G}_1$ and $\beta G_2=\theta(\alpha G_1)$, then both $\alpha G_1$ and $\beta G_2$ are of order $d$, whence $\beta$ is of order $d|G_2|=dn_1=n$.

Why must $\beta$ be of order $n$?? All I can prove is that the order of $\beta$ is of the form $d\cdot r$, where $r|n_1$.

share|cite|improve this question

Admittedly this step is weak: without further explanation the claim may be false. However, we can fix things. Note that the condition $\beta G_2=\theta(\alpha G_1)$ does not determine the element $\beta$ uniquely. Only its coset modulo $G_2$ is known. It turns out that we can choose $\beta$ within this coset in such a way that $\beta$ does have order $n$.

Let $\overline{G}_2$ be generated by an element $y$ of order $n_1d$, so $G_2=\langle y^d\rangle$. The condition $\beta G_2=\theta(\alpha G_1)$ tells us that $\beta G_2$ generates the quotient group $\overline{G}_2/G_2$, so $\beta\in y^i G_2$ for some integer $i, 0<i<d, (i,d)=1$. The choices that were made earlier ($\alpha$, $\theta$ and $y$) will determine $i$, but only modulo $d$. Let $n^*$ be the product of those prime factors of $n$ that are not factors of $d$. By the Chinese remainder theorem we can find an integer $\ell$ such that $\ell\equiv i\pmod d$ and $\ell\equiv 1\pmod {n^*}$. Then $(\ell,n)=1$, because otherwise $\ell$ would have common prime factors with either $d$ or $n^*$.

The former congruence says that $y^\ell G_2=y^i G_2=\theta(\alpha G_1)$, and the conclusion $(\ell,n)=1$ says that $y^{\ell}$ has order $n$. Therefore we can choose $\beta=y^\ell$.

share|cite|improve this answer
IOW: The natural projection $p:\mathbf{Z}_n\rightarrow\mathbf{Z}_d$ of residue class rings remains surjective, when restricted to the group of units $p: \mathbf{Z}_n^*\rightarrow\mathbf{Z}_d^*,$ and by lifting correctly we can be sure that $\beta$ is a generator. – Jyrki Lahtonen Sep 26 '11 at 19:33
I only glanced at the paper before I was told you had posted an answer, but it seems like on that particular page the authors are assuming that the group $G\lt A\times B$ is cyclic; $G_1$, $\overline{G_1}$, $G_2$, and $\overline{G_2}$ are the groups you get from Goursat's Lemma ($\overline{G_i}$ is the projection of $G$ onto the $i$th component, $G_i$ is the intersection of $G$ with the $i$th component of the product). So perhaps the assumption that the subgroup $G$ is cyclic comes into play (if I'm correct that this is an overarching assumption). – Arturo Magidin Sep 26 '11 at 19:54
I'll try to take a look as soon as I have the time. Probably not just right now, though... – Arturo Magidin Sep 27 '11 at 2:11
After looking at it more carefully, the paragraph in question is trying to establish that $G$ is cyclic (it is part of an "if and only if"). And I think you were spot on in terms of what the authors intended (namely, that by switching $\beta$ if necessary they could select it of the appropriate order). – Arturo Magidin Sep 27 '11 at 3:49

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.