I tried proving this in the following manner, but I am not confident with these types of problems so any verification would be appreciated. Thank you.

Let $A = \{(gH, gK): g \in G\}$

Define $\phi$ : $G$ $\rightarrow$ $A$ by $\phi(g)=(gH,gK)$

First we'll show $\phi$ is a homomorphism:


Now we'll show $\phi$ is bijective, and thus an isomorphism:

The codomain of $\phi$ is {$(gH,gK):g \in G$} so $\phi$ is clearly onto. Now suppose $g_1\not= g_2$ and $\phi(g_1)=\phi(g_2)$. Then $(g_1H,g_1K)=(g_2H,g_2K)$ $\Rightarrow$ $g_1H=g_2H$ and $g_1K=g_2K$ $\Rightarrow$ $g_2^{-1}g_1\in H$ and $g_2^{-1}g_1 \in K$, a contradiction since we assumed $H\bigcap K = \{e\}$


  • 2
    $\begingroup$ Your argument is sound. You could also prove injectivity directly, instead of by contradiction. $\endgroup$ Oct 21, 2018 at 6:43
  • $\begingroup$ Ah so you would just say $g_2^{-1}g_1 \not= e$ $\Rightarrow$ $g_2^{-1}g_1 \notin H \bigcap K$ $\Rightarrow$ either $g_1H \not= g_2H$ or $g_1K \not= g_2K$? $\endgroup$
    – PacoK
    Oct 21, 2018 at 7:10
  • $\begingroup$ I would say $g_2^{-1}g_1 \in H \cap K = \{e \} \implies g_2^{-1} g_1 = e \implies g_1 = g_2$. $\endgroup$ Oct 21, 2018 at 7:18

1 Answer 1


You could make this cleaner by just directly calculating the kernel (though like somebody already said, your argument is valid):

\begin{align*} \ker\phi &=\{g\in G\mid (gH,gK)=(H,K)\}\\ &= \{g\in G\mid gH=H\text{ and }gK= K\}\\ &= \{g\in G\mid g\in H\text{ and }g\in K\}\\ &= H\cap K\\ &= \{e\}\end{align*}

  • $\begingroup$ Oh, that makes a lot of sense. I tried going down that route at first, but for some reason I second guessed myself. $\endgroup$
    – PacoK
    Oct 21, 2018 at 7:16

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.