1
$\begingroup$

Does anybody know if there is a valuation-free proof of the following fact?

Let $K$ be a finite extension of $\mathbb{Q}$. Then, for every $q\in \mathbb{Z}$, the quotient ring $\mathcal{O}_{K}/\langle q \rangle$ is a principal ideal ring.

Thanks for your answers, comments, web links, etc.

$\endgroup$
3
$\begingroup$

Firstly, the Artinian ring $\mathcal O_k/(q)$ can be written as a finite product $\prod R_i$ of local Artinian rings, and one easily checks that $\mathcal O_k/(q)$ is a principal ideal ring if and only if each $R_i$ is.

You somehow need to use the fact that $\mathcal O_K$ is a regular ring of dimension $1$; for higher-dimensional rings the corresponding statement is false; e.g. $R = k[x,y]/(x^2,xy,y^2)$ is not a principal ideal ring.

Lemma. Let $(R,\mathfrak m)$ be a local Artinian ring such that $\dim_{R/\mathfrak m} \mathfrak m/\mathfrak m^2 = 1$. Then $R$ is a principal ideal ring.

Proof. Note that $\mathfrak m$ is principal by Nakayama's lemma, since $\dim \mathfrak m/\mathfrak m^2 = 1$. If $\mathfrak m = (\pi)$, then $\mathfrak m^n = (\pi^n)$, so $\dim \mathfrak m^n/\mathfrak m^{n+1} \leq 1$. If this dimension is $0$, then Nakayama's lemma implies that $\mathfrak m^n = 0$. It suffices to show that the only ideals are the $\mathfrak m^n$.

There exists $n_0 \in \mathbb Z_{>0}$ such that $\mathfrak m^n = 0$ if and only if $n \geq n_0$. Thus, we have a filtration $$0 = \mathfrak m^{n_0} \subsetneq \mathfrak m^{n_0-1} \subsetneq \ldots \subsetneq \mathfrak m \subsetneq R.$$ Since each of the successive subquotients is one-dimensional, we conclude that $\ell(\mathfrak m^n) = n_0-n$ for each $n \in \{0,\ldots,n_0\}$. Moreover, the $R$-module $\mathfrak m^n$ is principal and killed by $\mathfrak m^{n_0-n}$, hence a length computation shows that the surjection $R/\mathfrak m^{n_0-n} \to \mathfrak m^n$ given by $1 \mapsto \pi^n$ is an isomorphism of $R$-modules.

Let $I \subseteq R$ be any ideal, and let $c = \ell(R/I)$ be the colength of $I$. We prove by induction on $c$ that $I = \mathfrak m^c$. The case $c = 0$ is clear. If $c > 0$, then $I \subsetneq R$, so $I \subseteq \mathfrak m$. Because $\mathfrak m \cong R/\mathfrak m^{n_0-1}$, the submodule $I \subseteq \mathfrak m$ corresponds to a submodule $J \subseteq R/\mathfrak m^{n_0-1}$ of colength $c-1$. By the induction hypothesis, we know that $J = \mathfrak m^{c-1}$, hence $I = \mathfrak m^{c-1} \mathfrak m = \mathfrak m^c$, as claimed. $\square$

$\endgroup$
  • $\begingroup$ +1... Thanks a bunch for answering it, @R. van Dobben de Bruyn. $\endgroup$ – Jamai-Con Nov 23 '17 at 22:26
2
$\begingroup$

As Remy observes above, by unique factorization of ideals in $\mathcal{O}_K$ and the chinese remainder theorem we reduce to proving the following.

Proposition: Let $A$ be a Dedekind domain and $\mathfrak{p}$ a prime ideal of $A$. Then for any $n \in \mathbf{N}$, $A/\mathfrak{p}^n$ is a principal ideal ring.

Proof: The only prime ideal of $A$ containing $\mathfrak{p}^n$ is $\mathfrak{p}$ by unique factorization of ideals. Therefore $A/\mathfrak{p}^n$ is local and hence $$ A/\mathfrak{p}^n = (A/\mathfrak{p}^n)_{\mathfrak{p}} = A_{\mathfrak{p}}/\mathfrak{p}^nA_{\mathfrak{p}}. $$ But now $A_{\mathfrak{p}}$ is a DVR and so the right hand side is the quotient of a PID and therefore a principal ideal ring.

$\endgroup$
  • $\begingroup$ +1. Thanks a bunch for your answer, @Ben Lim. $\endgroup$ – Jamai-Con Nov 23 '17 at 22:26

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.