1
$\begingroup$

I am reading Strickland - Formal schemes and formal groups. For a functor $X\colon \operatorname{Rings}\to \operatorname{Sets}$, he defines (2.14) the category of $\operatorname{Points}(X)$ in the following way:
Objects: $(R,x)$ with $x \in X(R)$. Morphisms: ring morphisms $f\colon R\to S$ s.t $X(f)(x)=y$.

In remark 2.15 he claims that the category $\mathcal{X}_X$ of affine schemes $Y$ over an affine scheme $X$ and the category of representable functors $Y'\colon \operatorname{Points}(X)\to \operatorname{Sets}$ are equivalent using $Y(S)= \coprod_{z \in X(S)} Y'(S,z)$. I can't really understand how the equivalence follows. Would appreciate any help.

P.S. I understood the equivalence of (a) and (c)—this is usually referred to as the description of a scheme using its functor of points.

$\endgroup$
4
$\begingroup$

You want to see that the category $\mathcal{X}_X$ of affine schemes over $X$ (or affine $X$-schemes) is equivalent to the category of representable functors $Y':\text{Points}(X)\to \text{Set}$.

First, given an affine $X$-scheme $f:Y\to X$, and a point $(x,R)$ with $x\in X(R)$ and $R$ a ring, we associate the set $$f(R)^{-1}(\{x\})=\{ y\in Y(R) \ | \ f(R)(y)=x\},$$ where $f(R):Y(R)\to X(R)$ is the associated map on the $R$-points (usually called just $f$ by abuse of notation). So the functor say $f'$ (bad notation, by sure) we associate is on the objects of $\text{Points}(X)$ defined by $$f'((R,x)):=f(R)^{-1}(\{x\}).$$ (Note that an $X$-scheme is better defined by the morphism $f$ that by the scheme $Y$).

On the contrary, that a functor $Y':\text{Points}(X)\to \text{Set}$ is representable it means (my guess) that there exists a ring $S$ and a point $y\in X(S)$ such that $$Y'(R,x)=\{h:R\to S \ | \ X(h)(x)=y\}.$$ Note that a point $y\in X(S)$ is the same that a map $y:\text{Spec}(S)\to X$. We associate then to $Y'$ such affine $X$-scheme $y$.

$\endgroup$
1
$\begingroup$

Thank you @Xarles for the answer and the corrections and @Lspice for the editting. I think I got it now. For an affine $X$-scheme $f:Y\to X$ with $Y=Spec(R)$ for some ring R, we have the map $f:Spec(R)\to X$. Following the Yoneda lemma, the map $f$ corresponds to an $x_f\in X(R)$. We thus have a point $(R,x_f)$ and can hence associate to it the contravariant functor $Hom_{Points(X)}(\_,(R,x_f)):= Y'$, which is represented by this point. This is one direction $Y\to Y'$.

For the other direction, let $Y': Points(X)\to Set$ be a representable functor with $Y'= Hom_{Points(X)}(\_,(S,y))$ for some point $(S,y)$. We thus have an induced element $y\in X(S)$, which again by Yoneda, corresponds to a map $f_y:Spec(S)\to X$. Alltogether it gives the equivalence. The thing is, I don't fully understand how the identification $Y(S)= \coprod_{z \in X(S)} Y'(S,z)$ fits in my proof. Any ideas?

$\endgroup$

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.