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

Let $\phi,\psi: Y \rightarrow X$ be two morphisms between prevarieties. Prove that $\{ y \in Y| \phi(y)=\psi(y) \}$ is locally closed in $Y$.

In order to prove this, I denote the set $\{ y \in Y| \phi(y)= \psi(y) \}$ as $Z$, and define a map $f: Y \rightarrow X \times X, y \mapsto (\phi(y),\psi(y))$, then $Z$ is the inverse image of $\Delta(X) =\{ (x,x)| x \in X \} \subseteq X \times X$.

Is the map $f$ a morphism between $Y$ and $X \times X$? Is $\Delta(X)$ constructible in $X \times X$? If both of the questions have affirmative answers, the problem is settled.

Thank you very much.

share|cite|improve this question
Typo? If $\phi, \psi: X \to Y$ the map $f$ as written above isn't defined ... it should be $\phi, \psi\colon Y \to X$. – martini Sep 20 '11 at 15:13
@martini: This was a type error. Thank you very much. – ShinyaSakai Sep 20 '11 at 15:20
up vote 3 down vote accepted

Yes, the set $Ker(\phi,\psi)=\lbrace y\in Y|\phi (y)= \psi (y) \rbrace $ is a locally closed subset of $Y$.
Moreover it is closed if $X$ is a variety, that is if $X$ a separated prevariety.
And, yes: your map $f: Y \rightarrow X \times X, y \mapsto (\phi(y),\psi(y))$ is a morphism and $\Delta(X) \subset X \times X$ is locally closed and so certainly constructible.
This is proved in Mumford's The Red Book of Varieties and Schemes (II, §6, Proposition 4, page 118).
If $X$ is not separated, however, the subset of coincidence $Ker(\phi,\psi)$ need not be closed:

An example of non-closedness of the coincidence set Consider the notorious non-separated prevariety $X$ equal to the affine line $\mathbb A^1$ with two origins $O_1,O_2$, so that you have the open covering $X=U_1\cup U_2$ with $U_1=X\setminus \lbrace O_2 \rbrace, U_2=X\setminus \lbrace O_1 \rbrace$. There are obvious morphisms $\phi_1,\phi_2 :Y=\mathbb A^1 \to X$ with respective image $U_1,U_2$. Their set of coincidence is $Ker(\phi_1,\phi_2)=\mathbb A^1\setminus \lbrace O \rbrace$, which is not closed in $\mathbb A^1$.

Edit: Reminder on "locally closed" As an answer to the question in ShinyaSakai's comment, and since the notion "locally closed" is not so well-known, let us recall that given a topological space $T$ and a subspace $S\subset T$, the following are equivalent:

a) $S$ is the intersection $S=C\cap O $ of a closed subset $C$ and an open subset $O$ of $T$.
b) $S$ is relatively open in its closure $\bar S$.
c) Every point $s\in S$ has an open neighbourhood $U\subset T$ such that $S\cap U$ is closed in $U$.

Such a subset $S$ is said to be locally closed in $T$, which is quite reasonable in view of condition c).

Caveat In c), beware that the condition is to hold only for points $s\in S$: if it holds for all $t\in T$, the subset $S$ will be closed in $T$ ("closedness can be checked on an open covering").

share|cite|improve this answer
Thank you very much for your answer. I downloaded the book and read the proof. But it is still not clear to me. It is proved that $Z \cap U_2 = V(A)$. If I am not mistaken, $U_2$ is open, and $V(A)$ is closed. Is every set whose intersection with an open set closed a locally closed set? If not, why is the proof completed with $Z \cap U_2 = V(A)$? Thanks a lot~ – ShinyaSakai Sep 21 '11 at 14:01
Dear @ShinyaSakai, Mumford proves that every point $x\in Z$ in the coincidence set $Z$ has an open neighbourhood $U_2$ such that $U_2\cap Z$ is closed in $U_2$.This is enough to prove that $Z$ is locally closed (so that the terminology is actually excellent!) See the edit in my answer. – Georges Elencwajg Sep 22 '11 at 6:38
Dear @Georges Elencwajg, thank you really very very much for the detailed explanation. Now I understand the proof. – ShinyaSakai Sep 23 '11 at 17:01

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.