1
$\begingroup$

Problem Statement

Suppose $G_1, G_2$ are groups with $N_1 \lhd G_1$ and $N_2 \lhd G_2$ normal subgroups. Prove that $N_1 \times N_2 \lhd G_1 \times G_2$ and $(G_1 \times G_2)/(N_1 \times N_2) \cong G_1/N_1 \times G_2/N_2$.


Preamble

For this, I have a proof outlined and would like feedback on whether it's correct and where it could be simplified - all that good stuff. I will not include the proof that $N_1 \times N_2 \lhd G_1 \times G_2$ since that's a straightforward application of the definition of a normal subgroup as well as the computation of the direct product.

The proof of the second statement is, I believe, a case of the First Isomorphism Theorem. To that end, I endeavored to find a surjective homomorphism from $G_1 \times G_2$ to $G_1/N_1 \times G_2/N_2$ whose kernel was the normal subgroup $N_1 \times N_2$. I believe my proof is correct though my grasp of the material, specifically operating with quotient groups, is less intuitive than I would like. The possibility that I've made some oversight or mistake in my logic is a very real one. Due to this lack of intuition the proof will include a bit of my thought process as well, so it will be longer than is necessary but I want to include the thinking so that if something is wrong it can be clear where that misinterpretation is coming from. In particular my proof that the kernel of the homomorphism is the normal subgroup will include this extended thought process.


The Proof

Let $\eta: G_1 \times G_2 \to G_1/N_1 \times G_2/N_2$ be defined by $$\eta((g_1,g_2)) = (g_1N_1, g_2N_2)$$

  1. Now this mapping is a homomorphism since for any $(g_1, g_2),(h_1, h_2) \in G_1 \times G_2$: $$\eta((g_1, g_2)(h_1, h_2)) = ((g_1h_1)N_1, (g_2h_2)N_2) = ((g_1N_1)(h_1N_1), (g_2N_2)(h_2N_2)) = \eta((g_1,g_2))\eta((h_1,h_2))$$ Where the second equality holds since the product of two left cosets is itself a left coset.

  2. $\eta$ is surjective since for any $(g_1N_1, g_2N_2) \in G_1/N_1 \times G_2/N_2$ there exists a $(g_1,g_2) \in G_1 \times G_2$ such that $\eta((g_1,g_2)) = (G_1N_1,g_2N_2)$

  3. The kernel of $\eta$ is a subgroup of $G_1 \times G_2$ such that all elements in the kernel get sent to the identity in $G_1/N_1 \times G_2/N_2$.

    • The identity in $G_1/N_1 \times G_2/N_2$ is $N_1 \times N_2 = (n_1 , n_2)\ \text{for any}\ n_1,n_2 \in N_1 \times N_2$ since $$(g_1N_1,g_2N_2)(n_1,n_2) = (g_1N_1n_1, g_2N_2n_2) = (g_1N_1, g_2N_2)$$ Since $N_1,N_2$ are subgroups and hence closed under products and since all right identities are left identities we can see $N_1 \times N_2$ is the identity.
    • Now we can take any $(n_1,n_2) \in N_1 \times N_2$ and see that $$\eta((n_1,n_2)) = (n_1N_1,n_2N_2) = (N_1,N_2)$$ Since $(n_1,n_2)$ is an arbitrary element of the group $N_1 \times N_2$ we can see that $ker(\eta) = N_1 \times N_2$.

Thus, we can appeal to the First Isomorphism Theorem to see that we are guaranteed an isomorphism between $(G_1 \times G_2)/(N_1 \times N_2)$ and $G_1/N_1 \times G_2/N_2$ and hence $(G_1 \times G_2)/(N_1 \times N_2) \cong G_1/N_1 \times G_2/N_2. \square$

$\endgroup$
2
  • 1
    $\begingroup$ It's all good. One small thing is in the last step, strictly speaking you showed $N_1 \times N_2 \subseteq ker(\eta)$ but not the other way around. $\endgroup$ Commented Jun 26, 2022 at 0:31
  • $\begingroup$ @WhatsUp Ah yes, good catch it's not the first time I've only showed one-way containment when trying to prove equality of sets. I'll endeavor for it to be the last however. Thank you for that. $\endgroup$ Commented Jun 26, 2022 at 1:02

1 Answer 1

1
$\begingroup$

The notation in part 3 can be made more precise:

  1. Elements of $G_1/N_1 \times G_2/N_2$ are of the form $(g_1N_1, g_2N_2)$ for $g_1\in G_1, g_2\in G_2$, so while $(N_1, N_2)\in G_1/N_1 \times G_2/N_2$, it wouldn't make sense to say $N_1\times N_2\in G_1/N_1 \times G_2/N_2$.

  2. If we wanted to expand, we could write $(N_1, N_2)$ as $(\{n_1: n_1\in N_1\},\{n_2: n_2\in N_2\})$ or $(n_1 N_1, n_2 N_2)$ for some $n_1\in N_1, n_2\in N_2$, but we can't just write $(n_1, n_2)$.

  3. Similarly, to verify the identity element, I would just write $$(g_1 N_1, g_2 N_2)(N_1, N_2)=(g_1 N_1, g_2 N_2)(e_1N_1, e_2N_2)=((g_1e_1) N_1, (g_2 e_2)N_2)=(g_1 N_1, g_2 N_2),$$ where $e_1$ and $e_2$ are the identities of $G_1$ and $G_2$.

Finally, the last bullet is missing some details that WhatsUp pointed out.

Other than that, your solution looks good!

$\endgroup$
1
  • $\begingroup$ Great! Thanks for pointing that out, I'm still relatively new to proof-based math and so tightening up the rigor of my statements will be an ongoing process, accelerated by kind folks such as you! So, thanks again. $\endgroup$ Commented Jun 26, 2022 at 1:07

You must log in to answer this question.

Start asking to get answers

Find the answer to your question by asking.

Ask question

Explore related questions

See similar questions with these tags.