Let $g$ be a positive integer, and let $G$ be a commutative group with the following constraint on its torsion subgroup: there is an injection $G[\operatorname{tors}] \hookrightarrow (\mathbb{Q}/\mathbb{Z})^{2g}$. Must there be subgroups $G_1,\ldots,G_g$ of $G$ such that

(i) $G = G_1 \times \ldots \times G_g$ (internal direct product), and
(ii) For all $1 \leq i \leq g$, there is an injection $G_i[\operatorname{tors}] \hookrightarrow (\mathbb{Q}/\mathbb{Z})^2$?

Motivation: If this is true, then it reduces the "Inverse Mordell-Weil Problem for Abelian Varieties" to the "Inverse Mordell-Weil Problem for Elliptic Curves".

Thus although the given question certain has an affirmative answer in many special cases -- e.g. it is a triviality if $G$ is finitely generated -- I am not really interested in that. But it would be "lucky for me" if the answer turns out to be affirmative in the general case, so it's worth asking.

Added: This previous question contains some information on when the torsion subgroup of a commutative group is a direct summand.

  • $\begingroup$ I've been told the torsion subgroup of an abelian group need not be a direct summand, but I don't think I actually know any examples. It seems like a reasonable strategy here to construct a counterexample would be to find a group with torsion subgroup, say, $(\mathbb{Q}/\mathbb{Z})^3$ which isn't a direct summand. $\endgroup$ Jun 29, 2013 at 4:01
  • 2
    $\begingroup$ @QiaochuYuan That's impossible. $(\mathbb{Q}/\mathbb{Z})^3$ is divisible and hence injective as $\mathbb{Z}$-module. $\endgroup$ Jun 29, 2013 at 4:02
  • $\begingroup$ I'm a bit confused about the motivation; in the Mordell-Weil situation the group is finitely generated, right? Or is the idea that you want to show any such group can arise as the $K$-points of an ab. var. for some possibly infinite extension $K$? In that case, you may assume $G[tors]$ is finite, basically since $\mathbb{Q}_p/\mathbb{Z}_p$ is divisible. $\endgroup$ Jun 29, 2013 at 4:03
  • $\begingroup$ @Qiaochu: indeed the torsion subgroup need not be a direct summand. I asked an MO question about this a while back. The easiest example is probably $\prod_{\ell} \mathbb{Z}/\ell \mathbb{Z}$. The precise result is a theorem of Baer: for a torsion commutative group $T$, every (commutative!) extension of a torsionfree group by $T$ is split iff $T$ is the direct sum of a divisible $T_1$ and a group $T_2$ of bounded exponent (i.e., $T_2 = T_2[n]$ for some $n$). $\endgroup$ Jun 29, 2013 at 4:27
  • $\begingroup$ @Kevin: Here, by "Mordell-Weil group" I mean any group of the form $A(K)$ for an abelian variety over any field $K$. (And it is not enough to assume that the torsion subgroup is finite. In fact that case is significantly easier than the general case, because the torsion subgroup is necessarily a direct factor.) $\endgroup$ Jun 29, 2013 at 4:29

1 Answer 1


The answer is negative.

Define $S=\bigoplus_p C_p$, where $C_p$ is cyclic of order $p$ and $p$ ranges over prime numbers. Fix an integer $n\ge 1$. Define $P=\prod_p C_p$, and consider a maximal $\mathbf{Z}$-free family in $P^n/S^n$, generating a free abelian group (of continuum rank) $L/S^n$. Fix a prime $q$ and let $U/S^n$ be (*) a subgroup of $P^n/S^n$ isomorphic to $\mathbf{Z}_q$ (the group of $q$-adic numbers), with $L\subset U$.

Let's check that for every direct decomposition $U=V\oplus W$, either $V$ or $W$ is finite. Since $T(U)\simeq S^n$, this implies that $U$ answers negatively your question (as soon as $n\ge 3$).

First observe that $\mathbf{Z}_q$ is indecomposable (i.e. has no nontrivial decomposition as direct product, easy exercise (**)). Since $U/T(U)\simeq\mathbf{Z}_q$, it follows that either $V$ or $W$, say $W$, is torsion. So $V/T(V)\simeq\mathbf{Z}_q$.

Write $T(V)=\bigoplus A_p$ and $T(W)=\bigoplus B_p$, so that $A_p\oplus B_p=C_p^n$ for all $p$.

We have $P^n=(\prod A_p)\times (\prod B_p)$. So $P^n/T(V)=(\prod A_p/\bigoplus A_p)\times \prod B_p$. Since there is no nonzero homomorphism $\mathbf{Z}_q\to C_p$ for $p\neq q$ and since the subgroup $V/T(V)$ of $P^n/T(V)$ is isomorphic to $\mathbf{Z}_q$, the projection of $V/T(V)$ into $\prod B_p$ is contained in $B_q$. Thus in the above decomposition, $V\subset (\prod A_p)\times B_q$ and $W\subset (\bigoplus A_p)\times (\bigoplus B_p)$. So $U\subset (\prod A_p)\times (\bigoplus B_p)$. Since $P^n/U$ is torsion (because it is a quotient of $P^n/M$ which is torsion), it follows that $\prod B_p/\bigoplus B_p$ is torsion. This means that $B_p=0$ for large $p$. It follows that $W$ is finite.

Here are the two easy verifications:

(*): pick in $\mathbf{Z}_q$ a maximal $\mathbf{Z}$-free family, generating an abelian free subgroup $Z$; since $P^n/S^n$ is a rational vector space, any isomorphism $Z\to L/S^n$ (it exists since both groups are free abelian of continuum rank) can be extended to an injective homomorphism $\mathbf{Z}_q\to P^n/S^n$.

(**): $\mathbf{Z}_q$ is $p$-divisible for all $p\neq q$ and $\mathbf{Z}_q/q\mathbf{Z}_q$ has order $q$, so for every direct decomposition of $\mathbf{Z}_q$, at least one factor is divisible. But $\mathbf{Z}_q$ does not contain any copy of $\mathbf{Q}$.

  • $\begingroup$ Thanks very much. (This is very picky, but: I think you're not giving a counterexample but rather a negative answer to a question.) $\endgroup$ Jun 29, 2013 at 17:52
  • $\begingroup$ you're right; I'll fix this if I need to edit something else. $\endgroup$
    – YCor
    Jun 29, 2013 at 18:55
  • 1
    $\begingroup$ Now fixed (after 5 years of intense thinking...) $\endgroup$
    – YCor
    May 22, 2018 at 10:58
  • $\begingroup$ With some more efforts maybe one could get a countable example. We can replace the $p$-adics by rational $p$-adics, but then argument does not carry over when we use that $P^n/M$ is torsion... but it maybe possibly still holds that every decomposition has a finite summand. $\endgroup$
    – YCor
    May 22, 2018 at 11:01

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.