Consider the following theorem :

Let $G$ be a group. If $H$ is normal in $G$, and $G$ is solvable then $\tfrac{G}{H}$ is solvable.

According to my lecture notes the proof proceeds as follows:

If $H$ is normal in G and $G^n=1$ for some $n\in \Bbb N$.

$\Rightarrow (\tfrac{G}{H})^n=\tfrac{G^nH}{H}=\tfrac{H}{H}=1$, which shows that $\tfrac{G}{H}$ is solvable.

My questions:

i) Are we allowed to say that $(\tfrac{G}{H})^n=\tfrac{G^nH}{H}$ Because H is a normal subgroup and $G^n$ ( as it is contained in H) must be a subgroup of the normaliser $\Rightarrow G=G^nH$. It reminds me of the fratini argument but then that is only true for P-groups and even still $G^n$ is only a subgroup of the normaliser not the whole thing. Is there a better/correct reason for why this is true?

ii) I do not see how $\tfrac{H}{H}=1$ sure this would be true if they were numbers but it seems non-sensical if they are groups. Could anyone please explain this ?

  • 1
    $\begingroup$ $1$ is short for the trivial group $\{1\}$ or $\{e\}$. $\endgroup$
    – Randall
    Jan 11, 2019 at 17:33
  • $\begingroup$ @Randall still though how is $H/H=1$, I would of thought taking right cosets of H with elements of H would just return the group H itself ? $\endgroup$
    – excalibirr
    Jan 11, 2019 at 17:34
  • $\begingroup$ @can'tcauchy No How many cosets of $H$ in $H$ are there? $\endgroup$ Jan 11, 2019 at 17:35
  • $\begingroup$ @can'tcauchy well, if you want, $H/H =\{H\}$, the trivial group. $\endgroup$
    – Randall
    Jan 11, 2019 at 17:40

1 Answer 1


i) We're allowed to say it if we can prove it (as is the case with everything in mathematics). A proof could go as follows : by induction: it's clear for $n=1$; now if it's true for $n$, let $x\in (G/H)^{n+1} = D((G/H)^n)$. Then $x = [a,b]$ for some $a,b\in (G/H)^n$. By the induction hypothesis, $a$ and $b$ can be written as the class mod $H$ of $t,s \in G^n$. So $x$ is the class mod $H$ of $[t,s]\in G^{n+1}$, so $x\in G^{n+1}H/H$.

Conversely, let $x \in G^{n+1}H/H$. Then $x$ is the class mod $H$ of some $y\in G^{n+1}$. But then $y=[a,b]$ for some $a,b\in G^n$. By the induction hypothesis, the classes of $a,b$ mod $H$ are in $(G/H)^n$, so the class of $[a,b]$ is in $(G/H)^{n+1}$, so $x=$ the class of $y= $ the class of $[a,b] \in (G/H)^{n+1}$.

The result follows.

ii) There is only one class of $H$ mod $H$, as is the case for any group $G$. Indeed, if $x,y\in H$, then $xH= H = yH$.Therefore $H/H$ is a group with one element, there is only one up to isomorphism and we denote it by $1$.


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.