This theorem is from Matsumura (p.34)

Let $k$ be a field and $A$ an integral domain which is finitely generated over $k$. Then $\dim A = \operatorname{trdeg}_k A$ (where $\operatorname{trdeg}_k A$ is the transcendence degree of the field of fractions of $A$ over $k$).

I've been trying to read this proof but there are so many things that are confusing me, so I would appreciate any kind of help.


Let $A = k[X_1, ... , X_n]/P$, and set $r = \operatorname{trdeg}_k A$

Question 1: From what I understand $A = k[a_1, ... , a_n]$ for some $a_1, ... , a_n$ since it is finitely generated over $k$, and we have $f: k[X_1, ... ,X_n] \rightarrow k[a_1, ... , a_n]$ sending $X_i$ to $a_i$. So the first isomorphism theorem gives us $k[X_1, ... , X_n]/\ker f \simeq k[\alpha_1, ... ,\alpha_n]$. Since $A$ is an integral domain, $\ker f$ must be prime and we just call it $P$, right?

Question 2: Since $A$ is a module over $k$, the dimension of $A$ is defined to be the Krull dimension of $k/\operatorname{ann}(A)$, right?

To prove that $r \geq \dim A$ it is enough to show that if $P$ and $Q$ are prime ideals of $k[X] = k[X_1, ... ,X_n]$ with $Q \supset P$ and $Q \not= P$, then $$\operatorname{trdeg}_k k[X]/Q < \operatorname{trdeg}_k k[X]/P.$$

Question 3: Why would it be enough to show that? How does that say anything about the Krull dimension of $k/\operatorname{ann}(A)$?

  1. True.

  2. False: $\operatorname{ann}_k(A)=0$. The considered Krull dimension is that of $A$ as a commutative ring.

  3. Let's say $n=\dim A$. Then there is a chain of prime ideals $(0)=p_0\subset\cdots\subset p_n$ in $A$. Now note that $p_i=P_i/\ker f$, where $P_i$ is a prime ideal in $k[X]$. Moreover, $A/p_i=k[X]/P_i$. In particular, $P_0=\ker f$ and $A=k[X]/P_0$. From the inequality mentioned by the proof we get $$\operatorname{trdeg}_k k[X]/P_n+n\le \operatorname{trdeg}_k k[X]/P_0=\operatorname{trdeg}_kA.$$ Now use that $\operatorname{trdeg}_k k[X]/P_n=0$ (why?).

| cite | improve this answer | |
  • $\begingroup$ @user26857 $tr \deg_k k[X]/P_n = 0$ because $P_n$ is maximal, meaning that $P_n = (X)$...so we have $tr \deg_k k = 0$. But I have one more question (if you don't mind). The inequality tells us that $tr \deg k[X]/P_n \leq tr \deg k[X]/P_0$, right? It does not tell us that $tr \deg k[X]/P_n + n \leq tr \deg_k k[X]/P_0$. $\endgroup$ – Artus Oct 18 '14 at 18:41
  • $\begingroup$ @Artos In your question it is written $<$, not $\le$, right? Or if $a,b$ are integers and $a<b$ then $a+1\le b$. (Btw, $P_n$ is maximal, but isn't necessarily $(X)$.) $\endgroup$ – user26857 Oct 18 '14 at 18:44
  • $\begingroup$ @user26857 Oh...I'm sorry, I forgot about the chain length...thanks $\endgroup$ – Artus Oct 18 '14 at 18:48

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.