I can't see where I've gone wrong in the following proof. The statement above is wrong, as any topological space can be endowed with a constant sheaf.

Proof. That any set $X$ endowed with the trivial topology admits a constant sheaf is obvious: for any set $E,$ let $\mathscr{F}(X) = E.$

Conversely, suppose that $X$ admits a constant sheaf. Then since $X$ is irreducible, for every nonempty open set $U$ of $X,$ the restriction $\mathscr{F}(X)\to \mathscr{F}(U)$ is a bijection$^1$. We recall that a sheaf is a functor, $\mathscr{F}\colon \mathbf{Ouv}_X^{\text{op}} \to \mathbf{Set},$ where $\mathbf{Ouv}_X$ is the category whose objects are open $U\subseteq X$ and whose morphisms are the inclusions. Functors preserve isomorphisms; therefore $\mathscr{F}(X) \cong_{\mathbf{Set}} \mathscr{F}(U)$ implies that $X \cong_{\mathbf{Ouv}_X} U,$ i.e., $X = U.$

  1. Exercise $2.13$ in Algebraic Geometry I: Schemes, With Examples and Exercises by Ulrich Görtz and Torsten Wedhorn.
  • 5
    $\begingroup$ "Functors preserve isomorphisms". Yes. But you actually use it the other way around. $\endgroup$
    – user301452
    Jan 27, 2018 at 20:06
  • 3
    $\begingroup$ For example, if $C$ is any category, the functor $F \colon C \to \mathrm{Set}$ sending any object of $C$ to a fixed singleton set is a perfectly good functor. Further, $F(A) \cong F(B)$ for any objects $A, B$ in $C$, but $A$ and $B$ need not be isomorphic. $\endgroup$ Jan 27, 2018 at 20:08

1 Answer 1


It is not true that $\mathscr{F}(X)\cong\mathscr{F}(U)$ implies $X\cong U$! "Functors preserve isomorphisms" goes the other direction: if $F$ is a functor and $f:A\to B$ is an isomorphism in the domain category, then $Ff:FA\to FB$ is also an isomorphism, so if $A\cong B$, then $FA\cong FB$. The converse is not necessarily true.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.