Skip to main content
7 events
when toggle format what by license comment
Sep 28, 2022 at 19:54 comment added Jérôme Poineau @TimCampion I don't think $\mathcal{M}(A)$ should be locally connected in general, just as $\operatorname{Spec}(A)$ is not. Yes, $\mathcal{M}(\mathbb{Z})$ is connected and locally connected, which may be seen directly. Actually, all Berkovich spaces over $\mathbb{Z}$ are locally (path-)connected, see Théorème 7.2.17 in arxiv.org/abs/2010.08858.
Sep 28, 2022 at 17:10 vote accept Tim Campion
Sep 28, 2022 at 12:40 comment added Tim Campion Note to self: this answer shows that in the affinoid case, the connected components of $\mathcal M(A)$ are open in $\mathcal M(A)$. This doesn't compete the proof of local connectedness -- one needs to show that each point has a neighborhood basis of connected open sets. But that's exactly what Bekovich's 2.2.8 does.
Sep 28, 2022 at 12:38 comment added Tim Campion Thanks! I was indeed under the impression that $\mathcal M(A)$ should be locally connected for any discrete ring or Banach ring $A$, but I see now that it's important here to restrict to the affinoid case. So for example, the general theorem does not explain the fact (also true, I think!) that $\mathcal M(\mathbb Z)$ is connected and locally connected.
Sep 27, 2022 at 20:47 comment added Jérôme Poineau @Z.M: I am rather familiar with the construction but still not sure what the question is. Is $\mathcal{M}(A)$ locally connected for an arbitrary Banach ring $A$? Too general to be true I guess.
Sep 27, 2022 at 20:22 comment added Z. M I think that the Berkovich spectrum of a ring is simply the usual one, namely, the set of seminorms on a ring with weak topology. For example, $\mathcal M(\mathbb Z)$ looks like a star, with all archimedean and non-archimedean seminorms.
Sep 27, 2022 at 20:10 history answered Jérôme Poineau CC BY-SA 4.0