This is Exercise 8.28 of Gallian's "Contemporary Abstract Algebra".

Answers that use only methods from the textbook prior to the exercise are preferred.

Here $G_1\oplus G_2$ is the external direct product of $G_1$ by $G_2$.

Here $G_1\le G_2$ means $G_1$ is a subgroup or equal to the group $G_2$.

The Question:

Find a subgroup of $\Bbb Z_4\oplus\Bbb Z_2$ not of the form $H\oplus K$ for some $H\le \Bbb Z_4, K\le \Bbb Z_2$.


I must confess: I cheated here a little bit by looking up the subgroups of $\Bbb Z_4\times \Bbb Z_2$. But notice the difference in notation! I think in terms of just plain old direct products (because aren't internal and external direct products equivalent? Yes! But this is not established in the textbook yet; indeed, the former is not even mentioned at this point).

It appears to me to be a trick question. Here are the subgroups of $\Bbb Z_4\times \Bbb Z_2$. Where is the subgroup of the desired form?

My guess is that there's some technical aspect of external direct products that is being emphasised here.

Please help :)

  • $\begingroup$ What does "internal direct product" mean? Are you assuming $\mathbb Z_4,\mathbb Z_2$ are subgroups of some larger group? $\endgroup$ – Wojowu Mar 3 at 15:29
  • $\begingroup$ Sorry, @Wojowu: I got them the wrong way around! Hang on a second while I edit the question . . . $\endgroup$ – Shaun Mar 3 at 15:31
  • $\begingroup$ Also, the symmetric groups by Cayley's theorem, I would imagine, @Wojowu. $\endgroup$ – Shaun Mar 3 at 15:33
  • 1
    $\begingroup$ Internal direct products are mentioned in Section 9 of the textbook, @Wojowu; I'm on Section 8. $\endgroup$ – Shaun Mar 3 at 15:37
  • 1
    $\begingroup$ @Shaun In a small case like this, "shut up and calculate" might be my preferred approach. But there are some theoretical considerations which may help it go faster. For instance, if a subgroup is of the form $H\oplus K$ and contains $(a,b)$, then it must also contain $(a,0)$ and $(0,b)$. So taking an element of the form $(a,b)$ with $a,b\neq0$ and looking at the cyclic group generated by that element would be my first attempt. $\endgroup$ – Arthur Mar 3 at 16:02

The cyclic group generated by $(1,1)$ fits. It maps surjectively under both projections, but it is not the whole group, so you cannot write it as the direct product of two subgroups.


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.