0
$\begingroup$

Using exact sequences, it's fairly easy to prove the converse, but I can't figure out how to prove this statement.

Suppose we have a descending chain $N_1\supset N_2\supset\cdots$ of $R$-submodules of $M_1\times M_2$. Then if $N_i'=\pi_1(N_i)$, $N_i''=\pi_2(N_i)$ are the projections for all $i$, then I see that $\exists\,n_0$ s.t. $n\geq n_0 \implies N_n'=N_{n_0}', N_n''=N_{n_0}''$, so, looking at $N_{n_0}\supset N_{n_0 +1}\supset\cdots$, we have a descending chain of submodules with the same projection. Unfortunately I can't figure out how to get a contradiction from here.

I might be applying a topological intuition where it doesn't really apply.

$\endgroup$
3
  • $\begingroup$ Prove that if you have an exact sequence $0 \to E' \to E \xrightarrow{\pi} E'' \to 0$ and submodules $F \subset G \subset E$ such that $F \cap E' = G \cap E'$ and $\pi(F) = \pi(G)$ then $F = G$. $\endgroup$ – Hoot Jun 27 '16 at 16:48
  • $\begingroup$ Given a short exact sequence $0\to M' \to M \to M'' \to 0$ of $R$-modules, if $M'$ and $M''$ are Artin, then so is $M$. If you can prove this result, then you can apply this result to the short exact sequence $0 \to M_1 \to M_1 \oplus M_2 \to M_2 \to 0$ (Note that $M_1 \oplus M_2 \cong M_1 \times M_2$ $\endgroup$ – user348338 Jun 28 '16 at 2:49
  • $\begingroup$ I mean I already got it. I'll post an answer then. $\endgroup$ – Monstrous Moonshine Jun 28 '16 at 3:40
0
$\begingroup$

Consider the short exact sequence $0\xrightarrow{}M_1\xrightarrow{}M_1\oplus M_2\xrightarrow{\pi}(M_1\oplus M_2)/M_1\xrightarrow{}0$, then $M_1$ and $(M_1\oplus M_2)/M_1$ are both Artinian.

Let $N_1\supset N_2\supset\cdots$ be the descending chain, let $N_i' = N_i\cap M_1$, let $N_i'' = (N_i+M_2)/M_2$, then it is easy to see that $\{N_i'\}_{i=1}^\infty$ and $\{N_i''\}_{i=1}^\infty$ are descending sequences of $R$-submodules of $M_1$ and $(M_1\oplus M_2)/M_1$, respectively. Fix $n_0$ s.t. $\{N_i'\}_{i=n_0}^\infty$ and $\{N_i''\}_{i=n_0}^\infty$ are constant. Then we use the following lemma:

Lemma (due to Hoot): If $0 \to E' \to E \xrightarrow{\pi} E'' \to 0$ is an exact sequence of $R$-modules, and $F\subset G\subset E$ are submodules s.t. $F \cap E' = G \cap E'$ and $\pi(F) = \pi(G)$, then $F=G$.

Proof: By exactness, we set $E'\subset E$ and $E''=E/E'$. Suppose $x\in G$, then $\overline{x}\in\pi(G)=\pi(F)=(F+E')/E'\implies x\in F+E'$. Let $x=x_1+x_2,\,x_1\in F\subset G,\,x_2\in E'$. Then $x_2=x-x_1\in E'\cap G=E'\cap F\implies x\in F$.

Now, $0\xrightarrow{}M_1\xrightarrow{}M_1\oplus M_2\xrightarrow{\pi}(M_1\oplus M_2)/M_1\xrightarrow{}0$ is exact, and for $i\geq n_0$, $N_i\subset N_{n_0}\subset M_1\oplus M_2$, $N_i\cap M_1 = N_i' = N_{n_0}' = N_{n_0}\cap M_1$, $\,\pi(N_i) = N_i'' = N_{n_0}'' = \pi(N_{n_0})$, so $N_i=N_{N_0}\,\forall\,i\geq n_0$. Then $\{N_i\}_{i=1}^\infty$ is eventually constant, and $M_1\oplus M_2$ is therefore Artinian.

$\endgroup$

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.