Take the 2-minute tour ×
Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

I am trying to find the proof of :

The fibers of a finite morphism $\phi: X \rightarrow Y$ ($X,Y$ affine) are all finite.

Here, a morphism is called finite if $K[X]$ is integral over the image of $K[Y]$ under the comorphism $\phi^*$ of $\phi$.

If I can suppose this morphism to be dominant of irreducible varieties, then for any closed subset $V \subseteq Y$, the dimension of $\phi^{-1}(V)$ is no less than the dimension of $V$. Do I have to prove in this case that for any closed irreducible subset $V$ in $Y$, the reverse image in $X$ has equal dimension with $V$? How can I use the fact of $K[x]/\phi^*K[Y]$ being integral?

Thanks very much.

share|improve this question
1  
This is a question in commutative algebra, really: all you are asking is, if $B$ is an integral extension of $A$, why are there only finitely many primes of $B$ lying above a given prime of $A$? The answer has very much to do with integrality and is local on $A$, so it suffices to consider the case where $A$ is a local ring... –  Zhen Lin Feb 23 '12 at 7:54
    
@Zhen Lin: Thanks for reminding me of the essence of the question. Also, thank you for editing the tag~ –  ShinyaSakai Feb 27 '12 at 16:04

1 Answer 1

up vote 3 down vote accepted

You are asking why a finite morphism is quasi-finite. In the commutative-algebra language : For an algebra $A$, a finite $A$-algebra $B$, and a prime ideal $p$ of $A$, there finitely many prime ideals $q$ of $B$ such that $A\cap q = p$.

As Zhen Lin said, this is a local property. Indeed, is $S$ is a multiplicative subset of an algebra $A$, and if $B$ is a finite algebra over $A$, then $S^{-1} B$ is finite over $S^{-1} A$. More over, if $S\cap p = \varnothing$, then the prime ideals of $B$ above $p$ correspond bijectively to the prime ideals of $S^{-1}B$ above $pS^{-1}A$.

So we can assume that $A$ is local, with maximal ideal $p$. The prime ideals of $B$ above $p$ are prime ideals of $B/pB$. But this algebra $B/pB$ is a finite-dimensional $A/p$-vector space, and thus it is an noetherian and artinian algebra, and thus it has only a finite number of prime ideal [Bourbaki AC chap. II, §2, n˚ 5, prop. 9].

Edit — Thanks to Zach N for having pointed out the ridiculous mistakes I made. I don't know if there exists a direct proof for the last point.

share|improve this answer
    
Thanks a lot for the answer~ Would you please tell me why is $B/pB$ finite-dimensional over $A/p$? Also, as you said, "a finite $A$-algebra, $B$", is $K[X]$ necessarily finite over $\phi^*K[Y]$? I was just assuming $K[X]/\phi^*K[Y]$ to be integral. Thanks again~ –  ShinyaSakai Feb 27 '12 at 16:02
    
@ShinyaSakai — If you consider finitely generated $k$-algebra, then yes, integral is equivalent to finite. So the $A$-module $B$ is finite. Of course, a generating family of $B$ over $A$ is, after reduction, a generating family of $B/pB$ over $A$, hence over $A/p$. The point is that $A/p$ is a field, so you have a vector-space. Am I clear ? –  Lierre Feb 28 '12 at 9:28
    
@Lierre: I don't follow the last part of your argument. I think you want to argue that B/pB is artinian since B/pB is finite over A/p, so B/pB has only finitely many maximal (= prime, for artinian rings) ideals. Surely you don't mean to say that a finite-dimensional vector space can have only finitely many subspaces? Am I missing something? –  Zach N Jun 5 '12 at 18:03
    
I should point out that you can prove directly (perhaps this is what you were trying to get at) that a ring finite over a field has only finitely many prime ideals, but I think the argument is still nontrivial. –  Zach N Jun 5 '12 at 18:14
    
@ZachN — Oh... What the heck did I say ?? Thanks for pointing out that. –  Lierre Jun 5 '12 at 19:00

Your Answer

 
discard

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.