Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

I have a problem in understanding the proof of the following theorem:

Let $I\subseteq\mathbb{C}[[x_1,...,x_n]]$ be an ideal. Then there exists a $k\in\mathbb{N}$ and a linear coordinate change $\phi:\mathbb{C}[[x_1,...,x_n]]\to\mathbb{C}[[x_1,...,x_n]]$ such that $\mathbb{C}[[x_1,...,x_k]]\subseteq\mathbb{C}[[x_1,...,x_n]]/\phi(I)$ and $\mathbb{C}[[x_1,...,x_n]]/\phi(I)$ is finite as $\mathbb{C}[[x_1,...,x_k]]$-module.

As for the proof, assume $I\neq0$. Let $0\neq f\in I$, then one finds a linear coordinate change $\phi_1:\mathbb{C}[[x_1,...,x_n]]\to\mathbb{C}[[x_1,...,x_n]]$ such that $\phi_1(f)$ is $x_n$-regular. By the Weierstraß Preparation Theorem, there is a unit $u$ and a Weierstraß polynomial $p$ w.r.t. $x_n$ such that $u\phi_1(f)=p$. In particular, $\mathbb{C}[[x_1,...,x_{n-1}]]\hookrightarrow\mathbb{C}[[x_1,...,x_n]]/p$ is finite. Hence $\mathbb{C}[[x_1,...,x_{n-1}]]\to\mathbb{C}[[x_1,...,x_n]]/\phi_1(I)$ is finite.

The bold part is where I am still stuck. First, finite means the right hand side is a finitely generated module over the left hand side, is this correct (or is it over the image of the lhs)? Assuming this, the first part is due to $p$ being a Weierstraß polynomial w.r.t. $x_n$, i.e. $p$ has terms containing $x_n$ only up to a certain order.

Edit: If $f=\sum_{\mu\geq m}f_\mu$ is the homogeneous decomposition of $f$ with $f_m\neq 0$, take any $(a_1,...,a_{n-1})\in\mathbb{C}^{n-1}$ with $f_m(a_1,...,a_{n-1},1)\neq 0$, and define $\phi_1(x_i):=x_i+a_i x_n$ for $i<n$, $\phi_1(x_n):=x_n$. Then $\phi_1$ is what we wanted. But still, I don't get why this $\phi_1$ seems to work for all $f\neq 0$ in $I$ simultaneously. Or is this not what we need for the proof to work? How can I 'patch' these to get the result about $\mathbb{C}[[x_1,...,x_n]]/\phi_1(I)$?

And does it have anything to say that once the arrow $\hookrightarrow$ is used (which indicates injectivity to me), and in the next sentence it's just a 'normal' map?

Thank you in advance!

share|cite|improve this question
The second map is finite because it is the composition of the first one with the canonical surjection $C[[x_1,...,x_n]]/(p) \to C[[x_1, ..., x_n]]/(\phi_1(I)$. As both maps are finite, the composition is finite. – user18119 Aug 26 '12 at 21:25

About the first question: A ring $B$ is finite over a ring $A$ iff $B$ is a finite $A$-module. Now, let $A = \mathbb C[[x_1,\ldots,x_{n-1}]]$. Then $$C[[x_1,\ldots,x_n]]/\langle p \rangle \cong \bigoplus_{i=0}^{\deg p - 1} A x_n^i$$ as $A$-module, so it is obviously finite over $A$.

For the second part, I'm not 100% sure, but I think you simply apply the $\phi_1$ that was used for the construction of $p$ above and apply it to all elements of the ideal. Since it is linear, the result is again an ideal.

share|cite|improve this answer
Hello @Johannes and thanks so far! I think I have to check up on how such a $\phi_1$ can be constructed to see if for any element of the ideal, the same $\phi_1$ makes all of them regular w.r.t. $x_n$. Why / how do you use linearity there to say that $\phi_1(I)$ is again an ideal? Wouldn't surjectivity alone be enough? – InvisiblePanda Apr 1 '12 at 8:43
I'll have to look up the details for that. Indeed, $\phi_1$ surjective implies $\phi_1(I)$ ideal, but for the construction, it was somehow useful to assume $\phi_1$ linear. Actually, in the commutative algebra class I took, we proved that $\phi_1$ was given by a block matrix of the form $$\left(\begin{array}{c|ccc} 1 & & * & \\\hline & 1 & & * \\ 0 & & \ddots & \\\ & 0 & & 1 \end{array} \right)$$ – Johannes Kloos Apr 1 '12 at 8:53
Just a short idea: you can probably build $\phi_1$ inductively in the given form, by extending the matrix by one row and one column for each $x_k$. I guess there should be some argument for this in the Greuel/Pfister book somewhere... – Johannes Kloos Apr 1 '12 at 9:01
Hello @Johannes, sorry I didn't reply by now. I don't know if $\phi_1$ should be built inductively, since it is only meant to work for $x_n$. Indeed, the actual $\phi$ in the claim is built inductively. I will soon edit the above post to include the construction of $\phi_1$. Hopefully - as it often is - the actual idea how it works comes to me while writing ;) if not, we'll see. Thanks so far! The Greuel / Pfister book was a nice reference, shame on me I didn't think of it, since the class I took was held by Prof. Pfister... – InvisiblePanda Apr 16 '12 at 5:31
@Randal'Thor: If you want, I can send you a copy of my notes from Thomas Markwig's lecture - since he's in the same group as Pfister, I suppose the proof (or a variation of it) should be in there. – Johannes Kloos Apr 16 '12 at 9:53

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.