# How to construct the espace étalé (space of sections) for an arbitrary category?

I want to consider the sheaf valued in an arbitrary category (not only of sets, groups, modules and so on) on a topological space, using the language of étalé space.

In all references I am reading (Hartshorne's book, Harris' book, Vakil's note), only the sheaf of sets is considered in this way, although it is not hard to show that the same discussion works for sheaf of groups or modules. But how about the sheaf taking values in an arbitrary category?

For example, let $$\mathcal{F}$$ be a presheaf of sets on a topological space $$X$$. We can construct a topology on the disjoint union $$F$$ of $$\mathcal{F}_p$$ for all point $$p \in X$$ as well as a continuous map $$\pi: F \rightarrow X$$. Then the sheaf $$\mathcal{F}'$$ of sections of $$\pi$$ is the sheafification of $$\mathcal{F}$$.

But I cannot do the same discussion as above in the case where $$\mathcal{F}$$ is a presheaf valued in an arbitrary $$\mathbf{C}$$, because the construction of $$\mathcal{F}'$$ is based on topology but not on category.

That is, since $$\mathcal{F}$$ is a presheaf valued in the category $$\mathbf{C}$$, every $$\mathcal{F}(U)$$ and every $$\mathcal{F}_p$$ is an object of $$\mathbf{C}$$. Then if we assume that $$\mathbf{C}$$ has coproduct, the topological space $$F =\coprod_{p \in X} \mathcal{F}_p$$ can be also regarded an object of $$\mathbf{C}$$.

But $$\mathcal{F}' (U)$$ is defined to be the collection of sections of the map $$\pi: F \rightarrow X$$ on $$U$$, and none of categorical information is involved in this step. Then how can I claim that $$\mathcal{F}'$$ is a sheaf not only of sets but also valued in $$\mathbf{C}$$?

Of course, for the case of category of groups, rings, or $$R$$-modules, we can define addtions, productions, or $$R$$-actions of the sections, then I can show that $$\mathcal{F}'$$ is what I want. But what is the correct construction for an arbitrary $$\mathbf{C}$$?

• Does this help? mathoverflow.net/q/4474/73967 – Samantha Y Dec 7 '18 at 6:05
• Samantha, thank you! I am still learning some details in those answers, but I think that it is want I need. – Zhang Kongzheng Dec 7 '18 at 11:58
• You are welcome. For what it's worth, I also found math.stackexchange.com/questions/222896/… Perhaps this will be more accessible. – Samantha Y Dec 7 '18 at 20:10
• There's no such notion for a general codomain category. In particular, notice that $\coprod \mathcal F_p$ is not a topological space in any natural way! – Kevin Carlson Dec 8 '18 at 2:55
• Since $F$ is an object of $C$ and $X$ is a topological space, how are you defining the map $F\to X$? Do you assume that $C$ is some kind of concrete category? Arbitrary $C$ is too general, since I can take, for instance $C$ could be (the coproduct completion of) a combinatorially-defined countable category whose objects have no geometric interpretation. – David Roberts Dec 12 '18 at 20:38