I'm somewhat confused over the definition of Artinian and Noetherian rings.

A Noetherian ring is a ring in which there are no infinite chains of nested ideals. That is, if $I_i$ are some ideals in a ring $R$, which satisfy the condition $I_1\subset I_2 \subset ... \subset I_n \subset...$, then $\exists$ integer $N$ such that $I_N = I_m$ for all $m\ge N$.

An Artinian ring is a ring in which there are no infinite chains of nested ideals, but... if $I_i$ are some ideals in a ring $R$, which satisfy the condition $I_1\supset I_2 \supset ... \supset I_n \supset...$, then $\exists$ integer $N$ such that $I_N = I_m$ for all $m\ge N$.

But what is the difference between the Noetherian and the Artinian conditions? Can't we rewrite the ascending chain in the Noetherian case as a descending chain, and vice versa?

  • 3
    $\begingroup$ It suffices to give an example of one that is not the other. $\mathbb{Z}$ is Noetherian but not Artinian. There is no infinite ascending chain, but $\ldots \subset(8)\subset(4)\subset(2)$ is a descending chain. Ponder these two statements for a while. $\endgroup$ – Bob Jones Feb 26 '17 at 23:30
  • $\begingroup$ @BobJones Do you mean there is no infinite descending chain? I thought an ascending chain is a chain which begins from a smallest set and ascends to a bigger one. $\endgroup$ – sequence Feb 27 '17 at 1:02
  • 2
    $\begingroup$ You are correct in your definition. But do you believe $(2)\supset (4)\supset (8)\supset\ldots$ is ascending or descending? $\endgroup$ – Bob Jones Feb 27 '17 at 1:24
  • 2
    $\begingroup$ Exactly. So $\mathbb{Z}$ HAS infinitely descending ideal chains, namely the one I just gave. So it is not Artinian. It is a PID, though, so it is Noetherian. $\endgroup$ – Bob Jones Feb 27 '17 at 1:31
  • 5
    $\begingroup$ It turns out that artinian rings are noetherian, but it is far from obvious. $\endgroup$ – rschwieb Feb 27 '17 at 2:46

No, you can't just rewrite an infinite ascending chain as an infinite descending chain or vice versa. For instance, suppose you have an ascending chain $$I_1\subset I_2 \subset \dots \subset I_n \subset\dots$$ and want to get a descending chain $$J_1\supset J_2 \supset \dots \supset J_n \supset\dots$$ What are you going to define $J_1$ to be? You can't make it $I_1$, since that's the smallest ideal in the chain, and $J_1$ needs to be the largest instead. If you made $J_1$ be $I_2$, you could then make $J_2$ be $I_1$, but then you wouldn't be able to define $J_3$. If you made $J_1$ be $I_3$ you could define $J_2=I_2$ and $J_3=I_1$, but then you would have no way to define $J_4$. And so on.

For a similar phenomenon that should be familiar, there is an infinite ascending chain of natural numbers, namely $0\leq 1\leq 2\leq 3\leq\dots$. But there is no infinite descending chain of natural numbers, since if you start such a chain at $n$, you can only ascend $n$ times before you run out of natural numbers!

  • $\begingroup$ My understanding was that we could rewrite $I_1\subset I_2 \subset \dots \subset I_n \subset\dots$ as $...\supset I_n\supset \dots\supset I_2 \supset I_1$, hence my confusion. Also, if $R$ is Noetherian then $I_1\subset I_2 \subset I_n$ can be rewritten as $I_n\supset \dots\supset I_2 \supset I_1$, so it is also Artinian. $\endgroup$ – sequence Feb 27 '17 at 1:08
  • $\begingroup$ You can write that, but that's not an "infinite descending chain" for the purposes of the definition of an Artinian ring. $\endgroup$ – Eric Wofsey Feb 27 '17 at 1:11
  • $\begingroup$ So the idea is that every infinite ascending or descending chain must have a beginning element? Then this greatly clarifies the definition. Unfortunately, this important note was not given in my book or elsewhere I searched. $\endgroup$ – sequence Feb 27 '17 at 1:13
  • $\begingroup$ Well, the definition you quoted does not actually use the term "infinite descending chain". It just writes $I_1\supset I_2 \supset ... \supset I_n \supset...$. Notice that this is simply not the same as $...\supset I_n\supset \dots\supset I_2 \supset I_1$. $\endgroup$ – Eric Wofsey Feb 27 '17 at 1:14
  • 2
    $\begingroup$ No? As I've said, you can't interchange infinite ascending chains and infinite descending chains to deduce Artinian from Noetherian or vice versa. $\endgroup$ – Eric Wofsey Feb 27 '17 at 1:29

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.