Let $R$ be a ring (assumed associative and unital) whose additive group is a finitely generated abelian group. As a reduction step in a paper I'm working on, we need to know that $R$ is a quotient of another ring $S$ whose additive group is a finite rank free abelian group. We believe we have a proof, but it is rather involved. Is this something that has appeared already or at least has a reasonably short proof? (The current proof reduces to showing it for finite rings and then proves it by using the basic ring associated to $R$.)

When $R$ is commutative, this has a quick proof: $R$ is generated by $r_1,\dots,r_n$ which are integral over ${\bf Z}$, and hence $R$ is a quotient of ${\bf Z}[x_1,\dots,x_n]/(f_1(x_1),\dots,f_n(x_n))$ where each $f_i$ is a monic univariate polynomial.

So I'm wondering if there is something analogous in the general case.

  • 2
    $\begingroup$ It may be good to say at the start that $R$ is "an associative ring" (with identity?) rather than just "a ring". $\endgroup$
    – nfdc23
    Jun 1, 2017 at 15:00

1 Answer 1


For a finite ring $R$ you can write it as a quotient of the monoid ring $\mathbb ZR$ (with respect to the multiplicative structure) which has a finitely generated free additive group.

  • $\begingroup$ For the second part, I write $1=t+f$ where $t$ is torsion and $f$ belongs to the torsionfree part. Are you saying that $t$ is an idempotent? Could you please explain why? $\endgroup$
    – Steven Sam
    Jun 1, 2017 at 17:34
  • $\begingroup$ If you square then since the torsion elements are an ideal you get $f^2=f$. Therefore $t=1-f$ is idempotent. I did forget to check it is central $\endgroup$ Jun 1, 2017 at 17:52
  • 1
    $\begingroup$ Either way, doing the finite case was where our proof was more inefficient, so you've already helped trim the proof substantially with the monoid ring observation. $\endgroup$
    – Steven Sam
    Jun 1, 2017 at 18:14
  • 1
    $\begingroup$ $t$ can be 0. That is true. But $f$ and $t$ are idempotent. If $f$ Is 0 this is clear. Else $1=1^2=f^2+tf+ft+t^2$ and $tf+ft+t^2$ is torsion. So $f^2$ must be the torsion-free part of $1$. Thus $f^2=f$. $\endgroup$ Jun 1, 2017 at 18:21
  • 1
    $\begingroup$ I see that the above argument is slightly off because $f^2$ could have a further torsion part. $\endgroup$ Jun 1, 2017 at 18:32

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.