1
$\begingroup$

I am struggling with understanding this proof. I am confused as to why they took $a_{n+1}$ and how they got $a_{n+1}= b_1a_1+...+ b_na_n$, $a_1 = ... = a_0 = 0,$ and $a_{n+1} = 1$ gives $1 = 0$ in $\mathbb{R}$. Please help with explaining the steps of this proof in more detail. I am also not sure how to tie in that the ideals are infinitely generated and a proper susbset of each other.

Let $S = \mathbb{R}[a_0,a_1,a_2...]$ be a polynomial ring over R in infinitely many variables. Prove $S$ is non-Noetherian.

Proof: We want to show that S in non-Noetherian.Take the infinite chain of ideals: $<a_0>\subset<a_0, a_1>\subset...\subset<a_0,...,a_n>\subset ...$. Suppose that for any $n, a_{n+1}$ were an element of $<a_1,...,a_n>$. In other words, there exist polynomials $b_1,...,b_n$ such that $a_{n+1} = b_1a_1+...+ b_na_n$. Setting $a_1 = ... = a_0 = 0,$ but $a_{n+1} = 1$ gives $1 = 0$ in $\mathbb{R}$, a contradiction. So, S is non-noetherian.

$\endgroup$
  • $\begingroup$ There appear to be several typos in what you've written. Is this what is written in the text? $\endgroup$ – RideTheWavelet May 2 '18 at 13:34
  • $\begingroup$ I don't see how letting $a_{n+1}=1$, and $a_1=\dots =a_n=0$ is valid. Wouldn't this be like saying saying "Assume, $x=y$, now let $x=1$ and $y=0$"? $\endgroup$ – cansomeonehelpmeout May 2 '18 at 13:55
  • 2
    $\begingroup$ It is valid by the universal property of the polynomial ring, i.e. you can define a homomorphism out of it by specifying it on the variables. $\endgroup$ – Leon Hendrian May 2 '18 at 13:57
  • $\begingroup$ I'm confused as to how $a_{n-1}=1$ and $a_0 = ... = a_n = 0.$ Can you explain how they were able to make these distinctions? $\endgroup$ – Hannah Diaz May 2 '18 at 14:11
  • $\begingroup$ @LeonHendrian Thank you, you're right. I wasn't thinking about the evaluation homomorphism. $\endgroup$ – cansomeonehelpmeout May 2 '18 at 15:44
2
$\begingroup$

If $S=\mathbb{R}[a_0,a_1,\dots]$ were Noetherian then any chain of ideals $$I_0\subset I_1\subset\dots$$ would stabilize, that is, for some $n$ we would have $$I_n=I_{n+1}=I_{n+2}=\dots$$ To prove that $S$ is not Noetherian, we first assume that it is, and then see if we get a contradiction.

First we need a chain of ideals, so let's look at the chain: $$(a_0)\subset(a_0,a_1)\subset (a_0,a_1,a_2)\subset\dots$$ If we assume that $S$ is Noetherian, then this chain eventually stabilizes, so we would have, for some $n$, that $$(a_0,a_1,\dots,a_n)=(a_0,a_1,\dots,a_n,a_{n+1})$$

This would mean, in particular, that $a_{n+1}\in (a_0,a_1,\dots,a_n)$ so it should be possible to write $$a_{n+1}=\sum_{i=0}^{n}b_ia_i$$ Now, every polynomial ring has an evaluation homomorphism (this is what first confused me, but was clearified by @Leon Hendrian) that maps each $a_i$ to some element in the ring. (For instance, the evaluation homomorphism for $R[x]$, $R$ is any ring, is $\phi_a:R[x]\rightarrow R$, such that $f(x)\mapsto f(a)$, for some $a\in R$).

Using this homomorphism we can evaluate $a_{n+1}=\sum_{i=0}^nb_ia_i$ at $a_0=a_1=a_2=\dots =a_n=0$ and $a_{n+1}=1$. This gives $1=0$ which is a contradiction. Therefore $(a_0,a_1,\dots,a_n)\neq (a_0,a_1,\dots,a_n,a_{n+1})$ and so $S$ is not Noetherian.

$\endgroup$
0
$\begingroup$

When you write $a_n+1$, I assume that you mean $a_{n+1}$. Actually, this might already solve your confusion.

The logic of the proof you wrote is as follows: Assume $<a_0, \ldots, a_n> = <a_0, \ldots, a_{n+1}>$. This means that $a_{n+1} \in <a_0, \ldots, a_n>$, which means that $a_{n+1}=\sum_{i=0}^n b_ia_i$. But then, evaluate both sides on $a_0=\ldots=a_n=0, a_{n+1}=1$ to get the contradiction.

$\endgroup$
  • $\begingroup$ @Hannah Diaz - You're saying that two polynomials are equal. So, they must take the same values for any assignment of values to the variables. In particular, equality must hold if $a_1= \dots a_n=0$ and $a_{n+1}=1.$ $\endgroup$ – Chris Leary May 2 '18 at 14:14
  • $\begingroup$ But how does this suggest it’s not Noetherian? $\endgroup$ – Hannah Diaz May 2 '18 at 15:05
  • $\begingroup$ It proves that $<a_0, \ldots, a_n> \neq <a_0,\ldots,a_{n+1}>$ for all $n$, therefore the chain of ideals is indeed an infinite ascending chain of ideals, which directly contradicts Noetherianness. $\endgroup$ – Leon Hendrian May 2 '18 at 16:32

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.