Morphisms of locally ringed spaces and functors of points

137 Views Asked by At

I'm learning algebraic geometry and I'm trying to reconcile the locally ringed space and functor of points perspectives. Often when one defines a morphism $X \to Y$ of schemes thought of as locally ringed spaces they begin by saying where $x \in X$ goes to $y \in Y$, and then show that the resulting map is actually continuous, and then a morphism of locally ringed spaces.

(From now on, let's just suppose that $X = \operatorname{Spec} A$ is an affine scheme.)

On the other hand if $X$ and $Y$ are schemes thought of as functors of points, a morphism $X \to Y$ is just a natural transformation of functors from rings to sets. I figure that a recipe to take the point $x$ of a locally ringed space $X$ to a point $y \in Y$ of another translates into the functor of points perspective as a rule which produces from any field point $\operatorname{Spec} K \to X$ a field point $\operatorname{Spec} K \to Y$.

Of course, this is on its face not a rule to perform this translation for an arbitrary $R$-points $\operatorname{Spec} R \to X$. Nonetheless, by the equivalence between the two perspectives we should have:

  1. that any natural transformation $X \to Y$ of functors of points is completely determined by what happens to the field points plus some additional gluing data (encoded in the LRS language as a morphism of structure sheaves), and
  2. the conditions "is continuous", and "is a morphism of locally ringed spaces", should translate into conditions on the map of field points and additional gluing data required for the rule to assemble into an entire natural transformation $X \to Y$.

Is someone able to explain how this correspondence goes for (1) and (2) (or direct me to a reference)? Is there a convenient way to formulate how the additional gluing data (besides what happens to field points) is expressed?