4
$\begingroup$

I want to show that direct limit commutes with direct sums in the category of abelian groups. (I need this when studying algebraic topology) To do this, I first need a concise setting to start a proof, but I got stuck with this.

Let $\{G_\alpha , f_\alpha^\beta \}$ and $\{H_i,g_i^j\}$ be directed systems of abelian groups, with directed index sets $A$ and $I$, respectively. Then we have to show that the direct sum of the direct limit groups of these two directed systems are isomorphic to the direct limit group of the system $\{G_\alpha \oplus H_i, f_\alpha^\beta \oplus g_i^j \}$.

My question is

  1. Actually I'm not sure that my assertion is true. Is it true?

  2. Is my setting valid?

  3. If 1 is true, then it seems to be this isomorphism is natural. Is it right?

Thanks in advance

$\endgroup$

1 Answer 1

4
$\begingroup$

Denote the direct limit of $\left\{G_\alpha , f_\alpha^\beta \right\}$ as $G$, and the direct limit of $\left\{H_i,g_i^j\right\}$ as $H$. Then we claim that $G\oplus H$ is the direct limit of $\left\{G_\alpha \oplus H_i, f_\alpha^\beta \oplus g_i^j \right\}$. And your questions are answered by this claim.

The morphisms $G_\alpha\oplus H_i\rightarrow G\oplus H$ are induced by the morphisms $G_\alpha\rightarrow G$ and $H_i\rightarrow H$.

Suppose there is a compatible system of morphisms $G_\alpha\oplus H_i\rightarrow M$. Since $\operatorname{Hom}(G_\alpha\oplus H_i,M)\cong\operatorname{Hom}(G_\alpha, M)\oplus\operatorname{Hom}(H_i, M)$, this induces compatible systems of morphisms $G_\alpha\rightarrow M$ and $H_i\rightarrow M$. Then by the universal property of the direct limit, these morphisms factor through $G_\alpha\rightarrow G$ and $H_i\rightarrow H$. This immediately implies that the morphisms $G_\alpha\oplus H_i\rightarrow M$ factor through $G_\alpha\oplus H_i\rightarrow G\oplus H$. Therefore $G\oplus H$ is the direct limit of $\left\{G_\alpha \oplus H_i, f_\alpha^\beta \oplus g_i^j \right\}$ as claimed.


Hope this helps.

$\endgroup$
1
  • 1
    $\begingroup$ Does this result also hold true if we have a direct sum of an infinite number of abelian groups? $\endgroup$ Dec 20, 2020 at 15:53

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .