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.

The following question occurred to me while doing exercises in Hartshorne. If $A \to B$ is a homomorphism of (commutative, unital) rings and $f : \text{Spec } B \to \text{Spec } A$ is the corresponding morphism on spectra, does $f$ bijective imply that $f$ is a homeomorphism? If not, can anyone provide a counterexample? The reason this seems reasonable to me is because I think that the inverse set map should preserve inclusions of prime ideals, which is the meaning of continuity in the Zariski topology, but I can't make this rigorous.

share|improve this question
    
@Arturo: bijectivity on prime ideals is a weaker condition than bijectivity on the elements of the rings themselves. –  Qiaochu Yuan Jun 17 '11 at 19:07
    
@Qiaochu: Oh, I see; the only $f$ here is the induced map on Specs. Sorry! I thought "$f$ bijective" meant the original map $A\to B$ was bijective. I'll delete my first comment. –  Arturo Magidin Jun 17 '11 at 19:09
    
This is an interesting question! Justin's intuition for why it should be true makes sense to me. Just wanted to mention a random thought: this seems to be, in some sense, asking for a version of invariance of domain to hold for schemes. –  Zev Chonoles Jun 17 '11 at 19:28
    
@Zev: I hadn't thought of that. An analogous statement which I did have in mind is that a bijective continuous map of compact Hausdorff spaces is a homeomorphism. –  Justin Campbell Jun 17 '11 at 19:35
    
One observation is that you can reduce to the case where $A\to B$ is an embedding: the result clearly holds if $A\to B$ is onto, because then $P\subseteq Q$ implies $\varphi(P)\subseteq \varphi(Q)$. The map $\mathrm{Spec}(B)\to\mathrm{Spec}(A)$ factors through $\mathrm{Spec}(\varphi(A))$, and the conditions ensure that the contraction map to $\varphi(A)$ is also a bijection, while the map $\mathrm{Spec}(\varphi(A))\to\mathrm{Spec}(A)$ is also a bijection and thus a homeomorphism. So you want to know if the bijection of primes means if $A\subseteq B$, $P\cap A\subseteq Q\cap A$ then $P\leq Q$. –  Arturo Magidin Jun 17 '11 at 23:33

1 Answer 1

up vote 7 down vote accepted

No. Let $A$ be a DVR. Let $k$ be the residue field, $K$ the quotient field. There is a map $\mathrm{Spec} k \sqcup \mathrm{Spec} K \to \mathrm{Spec} A$ which is bijective, but not a homeomorphism (one side is discrete and the other is not). Note that $\mathrm{Spec} k \sqcup \mathrm{Spec}K = \mathrm{Spec} k \times K$, so this is an affine scheme.

As Matt E observes below in the comments, one can construct more geometric examples of this phenomenon (e.g. the coproduct of a punctured line plus a point mapping to a line): the point is that things can go very wrong with the topology.

share|improve this answer
    
Incidentally, this also gives an example of noetherian topological spaces $X, Y$ and a surjective map $f: X \to Y$ such that the dimension of $Y$ is greater than that of $X$ (this can't happen for varieties over an algebraically closed field, for instance). –  Akhil Mathew Jun 18 '11 at 0:14
1  
Dear Akhil, One could also give a more geometric example of the same kind, i.e. one that involves varieties rather than "exotic" schemes like the Spec of a DVR. E.g. $\mathbb A^1$ minus a closed point union $\Spec k$ mapping to $\mathbb A^1$ in the obvious way. Best wishes, –  Matt E Jun 18 '11 at 0:24
    
@Matt: Dear Matt, thanks for pointing that out. –  Akhil Mathew Jun 18 '11 at 0:33
2  
Dear Akhil, Of course my example doesn't exhibit the dimension-theoretic feature that yours does! (This goes hand-in-hand with its "less exotic" nature.) Best wishes, –  Matt E Jun 18 '11 at 1:00
    
Thanks for the simple examples, Akhil and Matt! –  Justin Campbell Jun 18 '11 at 5:15

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.