4
$\begingroup$

Are there any interesting cases (interesting here is interpreted rather loosely here) where you can show $X$ has property $P$ whenever all $X(K)$ have property $P$ where $K$ runs through all fields?

To clarify, let me state different versions of the questions.

Assume $X$ is defined over a base field $k$ and $K$ runs through all field extensions of $k$.

Version 1 Consider the sets $X(K)$ as having just a topology and assume $X$ is reduced.

I'm concerned with topological properties: being irreducible, being connected, counting irreducible or connected components, being an open subscheme of another scheme, being non empty,...

Version 1' If $X$ is not necessarily reduced what properties are insensitive to being reduced or not?

Version 2 Consider $X(K)$ as a topological space with the additional data of a sheaf of rings where for every $U(K) \subset X(K)$ we consider the ring generated by set maps $r_K: U(K) \to K$ for $r \in \mathcal{O}_X(U)$ where $r_K$ is defined by setting $r_K(\phi) = \phi^\#(r)$ for every $\phi \in U(K)$.

I'm more interested in Version 1 but perhaps more can be said about Version 2.

Motivation: Sometimes the field valued points of a scheme are much easier to understand than general $R$ points. For example a split reductive group $G$ you have the Bruhat decomposition of $G(K)$ which tell you the double cosets for a Borel $B \subset G$ are indexed by permutation matrices. But in general the cosets in $G(R)$ for a $K$-alebra $R$ are more complicated.

$\endgroup$
8
  • $\begingroup$ If X is a scheme, then the field valued points cannot distinguish between X and red X, the underlying reduced subscheme. $\endgroup$ Commented Jun 15, 2015 at 23:30
  • 3
    $\begingroup$ X(K) is just a set so not clear what properties it has. Do you want to work with properties of the functor $X(K)$? $\endgroup$
    – Will Sawin
    Commented Jun 16, 2015 at 1:44
  • $\begingroup$ @DanielBarter, so maybe the question makes more sense if one stipulates that the scheme is reduced? In that case, for example, is there any sense in which we know 'everything' about the scheme once we know its functor of field points? $\endgroup$
    – LSpice
    Commented Jun 16, 2015 at 2:06
  • $\begingroup$ @WillSawin, the post mentions the example of reductive groups. Maybe one more generally wants to understand $X$ in terms of additional structure on $X(K)$—for example, group structure, or, if $K$ is a special field (say, $p$-adic or Archimedean), topology? $\endgroup$
    – LSpice
    Commented Jun 16, 2015 at 2:07
  • $\begingroup$ @solbap. One example you probably already know. If $X_{/\mathbf{C}}$ is proper, then the analytic space $X(\mathbf{C})$ "knows" $X$ by GAGA theorems. $\endgroup$ Commented Jun 16, 2015 at 2:18

0

Browse other questions tagged .