13
$\begingroup$

Given a topological ring $R$, under what conditions and in what way, can one induce a topology on the $R$-points of a scheme $X$? For example, if $X$ is $P^n$ or $A^n$, one has natural topology on the $R$-points.

If $G$ is a group scheme/A and $R$ is $A$-algebra (still a topological ring), will the induced topology on $G\left(R\right)$ (as above) automatically make $G$ into a topological group.

For number theorists, if $G$ is an algebraic group/Q, we can consider the adelic points $G\left(A_{K}\right)$ for any number field $K$. Is the induced topology on $G\left(A_{K}\right)$ that of a restricted direct product? Under what conditions will $G\left(A_{K}\right)$ be locally compact or satisfy other nice properties?

Improve this question
$\endgroup$

3 Answers 3

Reset to default
17
$\begingroup$

Brian Conrad has some notes on this on his website ("Some notes on topologizing the adelic points of schemes, unifying the viewpoints of Grothendieck and Weil"). The short version is that if X is affine, you can topologize X(R) in a natural, functorial way (specifically, the weakest topology such that the functions X(R)-->R induced by elements of the structure sheaf are continuous). If X isn't affine, you have to be more careful, because the units of R might not be open in X and might not be a topological group wrt the subspace topology. But those are the only problems, and if your ring doesn't have those problems, you can glue the naturally topologized affines and everything is functorial. Happily for number theorists, the adeles are fairly nice, and for a finite-type separated K-scheme X, X(A_K) can be naturally topologized, and it is locally compact and Hausdorff.

Improve this answer
$\endgroup$
2
  • $\begingroup$ Does this all agree with, or can this be made to agree with, the "classical" topology on C-points? $\endgroup$
    – Kevin H. Lin
    Oct 11, 2009 at 15:51
  • $\begingroup$ Yes, it is all consistent with the case when R is a topological field. This stuff is all addressed in the notes which Rebecca mentions. But it is probably more instructive to figure it out for yourself. $\endgroup$
    – BCnrd
    Feb 23, 2010 at 4:14
3
$\begingroup$

For adelic points of X (or G), one can first topologize X(Q_p) so that it becomes a p-adic analytic variety, and for almost all p one can define an open subset X(Z_p). Then take X(A) to be the restricted product.

Improve this answer
$\endgroup$
1
$\begingroup$

See also Andrei Jorza's thesis http://www.its.caltech.edu/~ajorza/notes/bsd.pdf, pp.16 ff.

Improve this answer
$\endgroup$
1
  • $\begingroup$ This link seems to be broken. Would you mind updating it ? Thank you! $\endgroup$
    – Dat Minh Ha
    Dec 28, 2022 at 4:50

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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