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.

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 am having a problem with a question in Atiyah-Macdonald Chapter 3 qn 19(viii). The question is as follows:

If $f:A\rightarrow B$ is a ring homomorphism, and $M$ is finitely generated $A$-module, then $\mbox{Supp}(B\otimes_{A}M)=f^{*-1}(\mbox{Supp}(M))$

I tried to show $\mbox{Supp}(B\otimes_{A}M)\subseteq f^{*-1}(\mbox{Supp}(M))$. Suppose ${\frak{p}}\notin f^{*-1}(\mbox{Supp}(M))$, then $f^{-1}(\mathfrak{p})\notin \mbox{Supp}(M)$. Thus for each generator $m_{i}$ of $M$, there exists $s_{i}\in A-f^{-1}(\mathfrak{p})$ such that $s_{i}m_{i}=0$. If $f(s_{i})=0$ then since $f(s_{i})\in B-\mathfrak{p}$, thus $0\in B-\mathfrak{p}$ and we are done. Otherwise for all $f(s_{i})$ if they are not zero, we have $f(s_{1})f(s_{2})\ldots f(s_{k})(b\otimes m)=b\otimes (s_{1}s_{2}\ldots s_{k})m=0$

But I am stuck with $f^{*-1}(\mbox{Supp}(M))\subseteq \mbox{Supp}(B\otimes_{A}M)$. The problem is if I start letting $\mathfrak{p}\in LHS$, so each of $m_{i}$ is not zero, but I can't show that $1\otimes m_{i}$ is not zero. Or if I start from $\mathfrak{p}\notin RHS$, I can't show that $(A-f^{-1}\mathfrak{p})M=0$.

How should I approach? Thanks!

share|cite|improve this question
@navigetor23: Thanks! Noted. – enoughsaid05 Nov 18 '12 at 10:36
up vote 3 down vote accepted

Since no one answered you (because it's Sunday?), let me try it.

Let $q\in f^{*-1}(\mbox{Supp}(M))$. Then $p=f^{-1}(q)\in\mbox{Supp}(M)$, that is, $M_p\neq 0$. I want to prove that $q\in\mbox{Supp}(B\otimes_{A}M)$. Since $(B\otimes_{A}M)_q\cong B_q\otimes_{A_p}M_p$ (why?) we can reduce the problem as follows: let $f:(A,m)\to(B,n)$ be a homomorphism of local rings with $m=f^{-1}(n)$ and $M\neq 0$ a finitely generated $A$-module. If $B\otimes_AM=0$, then tensor this by $B/n$ and get that $B/n\otimes_AM=0$. But $B/n\otimes_AM\cong B/n\otimes_{A/m}M/mM$, so we obtained a field extension $A/m=K\subset L=B/n$ and $V=M/mM$ a nonzero (why?) $K$-vectorspace such that $L\otimes_KV=0$, a contradiction. Therefore $B\otimes_AM\neq 0$.

share|cite|improve this answer
What are (A,m) and (B,n)? How does $B/n\otimes_{A}M\cong B/n\otimes_{A/m}M/mM$ leads to the field extension? Why is $M/mM$ a non-zero K-vsp, and what does it contradict? Thanks! – enoughsaid05 Nov 19 '12 at 6:20
Well, it seems that you are not used with the reduction techniques in commutative algebra. Here $A$ denotes $A_p$ and $m$ is $pA_p$ (similar for $B$). If you have an extension of rings $R\subset S$ and $P$ a prime of $S$ whose contraction to $R$ is $p$, then $R/p\subset S/P$. $M/mM=0$ leads to $M=0$ via Nakayama. (Btw, the new $M$ is $M_p$.) – user26857 Nov 19 '12 at 9:59
sorry if i ask too many questions but why $B/n$ is a field extension of $A/m$? (and I probably have other questions on some identities but i will try to figure them out first). Anyway, what is the reduction technique in commutative algebra? – enoughsaid05 Nov 19 '12 at 17:43
Because they are fields and there is a ring homomorphism from one to another (which is necessarily injective, right?). – user26857 Nov 19 '12 at 17:45
One question: to answer why it leads to a contradiction, am I right in saying that because $M/mM\cong K\otimes_{K}M/mM\cong \phi(K)\otimes_{K}M/mM\subseteq L\otimes_{K}V$ where $\phi$ is the field homomorphism from $K$ to $L$. so since if we assume that $L\otimes_{K} M/mM$ is zero, implies that so is $M/mM$ and hence $M=0$ by Nakayama's lemma. Other than that, I think they are correct! How do you manage to come up with this proof (especially on your first try!). What is a reduction technique in commutative algebra? – enoughsaid05 Nov 19 '12 at 20:11

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.