66.8 Constructible sets
Lemma 66.8.1. Let $S$ be a scheme. Let $X$ be an algebraic space over $S$. Let $E \subset |X|$ be a subset. The following are equivalent
for every étale morphism $U \to X$ where $U$ is a scheme the inverse image of $E$ in $U$ is a locally constructible subset of $U$,
for every étale morphism $U \to X$ where $U$ is an affine scheme the inverse image of $E$ in $U$ is a constructible subset of $U$,
for some surjective étale morphism $U \to X$ where $U$ is a scheme the inverse image of $E$ in $U$ is a locally constructible subset of $U$.
Proof.
By Properties, Lemma 28.2.1 we see that (1) and (2) are equivalent. It is immediate that (1) implies (3). Thus we assume we have a surjective étale morphism $\varphi : U \to X$ where $U$ is a scheme such that $\varphi ^{-1}(E)$ is locally constructible. Let $\varphi ' : U' \to X$ be another étale morphism where $U'$ is a scheme. Then we have
\[ E'' = \text{pr}_1^{-1}(\varphi ^{-1}(E)) = \text{pr}_2^{-1}((\varphi ')^{-1}(E)) \]
where $\text{pr}_1 : U \times _ X U' \to U$ and $\text{pr}_2 : U \times _ X U' \to U'$ are the projections. By Morphisms, Lemma 29.22.1 we see that $E''$ is locally constructible in $U \times _ X U'$. Let $W' \subset U'$ be an affine open. Since $\text{pr}_2$ is étale and hence open, we can choose a quasi-compact open $W'' \subset U \times _ X U'$ with $\text{pr}_2(W'') = W'$. Then $\text{pr}_2|_{W''} : W'' \to W'$ is quasi-compact. We have $W' \cap (\varphi ')^{-1}(E) = \text{pr}_2(E'' \cap W'')$ as $\varphi $ is surjective, see Lemma 66.4.3. Thus $W' \cap (\varphi ')^{-1}(E) = \text{pr}_2(E'' \cap W'')$ is locally constructible by Morphisms, Theorem 29.22.3 as desired.
$\square$
Definition 66.8.2. Let $S$ be a scheme. Let $X$ be an algebraic space over $S$. Let $E \subset |X|$ be a subset. We say $E$ is étale locally constructible if the equivalent conditions of Lemma 66.8.1 are satisfied.
Of course, if $X$ is representable, i.e., $X$ is a scheme, then this just means $E$ is a locally constructible subset of the underlying topological space.
Comments (2)
Comment #5096 by Klaus Mattis on
Comment #5305 by Johan on