$\newcommand\End{\operatorname{End}}$Let $G$ be an Abelian group.

Are there sufficient conditions for the existence of an isomorphism $G\cong\End(G)$, where $\End(G)$ is considered a group under addition? This is true for $\mathbb Z$ and $\mathbb Q$, but not (as far as I know) for $\mathbb R$, where you can construct monstrous additive functions from the Axiom of Choice.

Side questions for people to consider: If $R$ is a ring, when will it be isomorphic (as a ring) to the ring $\End_{\text{group}}(R)$ of group endomorphisms? When will it be isomorphic to the ring $\End_{\text{ring}}(R)$ of ring endomorphisms?

  • $\begingroup$ Regarding the second part of the side question: the ring endomorphisms of a ring don't form a ring. $\endgroup$ – Jeremy Rickard Apr 11 '15 at 10:09
  • 1
    $\begingroup$ With composition as multiplication, I'm quite sure they do. $\endgroup$ – Gaussler Apr 11 '15 at 10:26
  • 1
    $\begingroup$ @Gaussier For $R = ℤ$, $f = \mathrm{id}_ℤ + \mathrm{id}_ℤ$ is not multiplicative (nor does it preserve $1$), because $f(1·1) = 2 ≠ 4 = f(1)·f(1)$ (that’s right – $\mathrm{Ring}$ is not an additive category). $\endgroup$ – k.stm Apr 11 '15 at 11:11
  • $\begingroup$ Ah, you're right. Damn rings with one! $\endgroup$ – Gaussler Apr 11 '15 at 11:13
  • $\begingroup$ @Gaussler Yeah, but it doesn’t work for rings without one (like $2ℤ$) either. $\endgroup$ – k.stm Apr 11 '15 at 11:18

A few partial answers:

1) if $G$ is a finitely generated abelian group, then $G\simeq\mathrm{End}(G)$ iff $G$ is cyclic.

indeed, if $G$ is infinite and has an element of prime order $p$, then $\mathrm{End}(G)$ has more elements of order $p$ than $G$. A similar counting argument also works when $G$ is finite and non-cyclic. And if $G$ is free abelian of rank $r\ge 2$, then $\mathrm{End}(G)$ is free abelian of rank $r^2$.

2) if $G$ is a subring of $\mathbf{Q}$ (e.g., $\mathbf{Z}[1/n]$ or $\mathbf{Q}$ itself), then $G\simeq\mathrm{End}(G)$. On the other hand, there are subgroups of $\mathbf{Q}$ such that $G$ is not isomorphic to $\mathrm{End}(G)$: for instance the subgroup of $\mathbf{Q}$ generated by the $1/p$ for prime $p$ has $\mathrm{End}(G)\simeq\mathbf{Z}$.

3) there are some more examples: for instance if $A$ is a subring of $\mathbf{Q}$ containing $1/n$ then $G=A\times(\mathbf{Z}/n\mathbf{Z})$ satisfies $G\simeq\mathrm{End}(G)$.

4) if $G\simeq H^{(X)}$ for some nonzero countable group $H$ and infinite set $X$ (this is the set of finitely supported functions $X\to H$), then $G$ is not isomorphic to $\mathrm{End}(G)$, just because $\mathrm{End}(G)$ has cardinal greater than $G$: indeed the cardinal of $G$ is that of $X$, while the cardinal of $\mathrm{End}(G)$ is that of $2^X$ (because $\mathrm{End}(G)$ contains $\mathrm{End}(H)^X$).


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.