6
$\begingroup$

Let $G$ be a finitely generated abelian group and $H$ be a subgroup. Let subscript $t$ denote the torsion subgroup. If $G/G_t$ is free of rank $n$ and $H/H_t$ is free of rank $m$, it is easy to embed $H/H_t\hookrightarrow G/G_t$ and deduce that $m\le n$. Now the question is that I want to show that $(G/H)/(G/H)_t$ is free of rank $n-m$.

This is harder than it looks and I have not succeeded in finding a proof after many hours.

[EDIT] I'm looking for a group theory proof.

$\endgroup$
2
  • $\begingroup$ Try using the elementary divisor structure theory. $\endgroup$ – i. m. soloveichik May 1 '14 at 14:08
  • $\begingroup$ It is definitely possible to derive this from a known theorem about the basis of a subgroup of a finitely generated free abelian group. $\endgroup$ – Dan Shved May 13 '14 at 20:30
6
$\begingroup$

The rank of $G/G_t$ is the dimension of $G\otimes\mathbb{Q}$ as a vector space.

From the exact sequence $0\to H\to G\to G/H\to 0$, you get the exact sequence $$ 0\to H\otimes\mathbb{Q}\to G\otimes\mathbb{Q}\to (G/H)\otimes\mathbb{Q}\to 0 $$

$\endgroup$
2
  • $\begingroup$ why the rank of $G/G_t$ is the dimension of $G\otimes\mathbb{Q}$ as a vector space ? $\endgroup$ – WLOG May 1 '14 at 16:25
  • 2
    $\begingroup$ @WLOG $G\otimes\mathbb{Q}$ is isomorphic to $(G/G_t)\otimes\mathbb{Q}$ (easy); since $G/G_t\cong\mathbb{Z}^n$ where $n$ is the rank, we're done. $\endgroup$ – egreg May 1 '14 at 16:27
0
$\begingroup$

The rank is the dimension of G⊗Q because it is isomorphic to (G/Gt)⊗Q as G/Gt=Z^n as n is the rank

$\endgroup$
6
  • $\begingroup$ But why is it this?!? $\endgroup$ – user1729 May 20 '14 at 13:13
  • $\begingroup$ Because it is isomorphic to (G/Gt)⊗Q as G/Gt=Z^n as n is the rank. $\endgroup$ – user150369 May 20 '14 at 13:21
  • $\begingroup$ Then say so in your answer! $\endgroup$ – user1729 May 20 '14 at 13:24
  • $\begingroup$ My apologies, I will edit it. Thank you. $\endgroup$ – user150369 May 20 '14 at 13:27
  • $\begingroup$ Also, this is the same as egreg's answer, unless I am missing some subtlety? $\endgroup$ – user1729 May 20 '14 at 13:38

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.