2
$\begingroup$

Is it possible to define (possibly derived) categories of constructible sheaves over sites more general than those of open subsets of topological spaces while still retaining essential features, like how if $\Lambda$ is a Noetherian ring then constructible sheaves of $\Lambda$-modules will span a thick subcategory of the abelian category of sheaves of $\Lambda$-modules ? Will such a generalisation of constructible sheaves involve defining stratifications of sites/sheaf topoi, and if so, how would we go about defining such notions of stratifications ?

In particular, I am interested in "direct" definitions of constructible étale, lisse-étale, and fppf sheaves over a given scheme/algebraic space/Artin stack $X$. Certain restrictions that I am willing to work under are that $X$ is (locally) of finite type over some base scheme $S$ that is affine, regular, Noetherian, of dimension $\leq 1$, and of characteristic $p \geq 0$, and that the ring of coefficient $\Lambda$ is a Gorenstein local ring of dimension $0$ and characteristic $\ell \not = p$.

$\endgroup$
3
  • $\begingroup$ Am I mistaken? Constructible sheaves are never defined over sites of "open subsets of topological spaces" for schemes of characteristic $p$, but rather on its étale site? Anyway, if you are looking for a modern reference, maybe you could consult The pro-étale topology for schemes and Constructible sheaves on schemes and a categorical Künneth formula. $\endgroup$
    – Z. M
    Commented Feb 15, 2022 at 18:56
  • $\begingroup$ @Z.M Thank you for the references. In response to your question: I might very well be wrong, but aren't constructible sheaves on a given scheme defined with respect to a stratification of that scheme into Zariski-locally closed subschemes (which would make constructibility not a "purely étale" notion) ? $\endgroup$ Commented Mar 5, 2022 at 23:50
  • $\begingroup$ It does not seem possible to say something "purely étale": an étale morphism is an open immersion if and only if it is a monomorphism. $\endgroup$
    – Z. M
    Commented Mar 6, 2022 at 1:03

0

You must log in to answer this question.