2

Given a commutative ring $R$ with proper ideal $I$, we can let $N_r=\{r+I^n\}_{n\in\Bbb N}$ and $\mathcal{B}=\bigcup_{r\in R} N_r$, and take the smallest topology containing these sets. I believe this is the $I$-adic topology?

  1. Am I correctly interpreting the definition of the $I$-adic topology?
  2. This definition seems highly dependent on the choice of ideal $I$, in particular, one could take the $0$ ideal? I guess that would give the discrete topology? Are there any classes of ideals where the topology is equivalent?
  3. Is it true that for any proper ideal $I\subseteq R$ that $\hat{R_I}=\displaystyle\projlim_{\longleftarrow} (R/I^n)$ is isomorphic as a topological space to the completion of $R$ with the topology of 1?

What may be confusing me with 1 is that all over the place they say 'take a system of neighbourhoods of 0' (and then use the translations of these), where we haven't yet even picked a topology, so the word neighbourhood means nothing yet (since otherwise you presuppose the notion of openness in your non-topological ring)... But then it further seems like you don't want to take the smallest topology containing $\mathcal{B}$ as I did above, since then you're pretending these sets in $\mathcal{B}$ are open, when maybe you only want them to be neighbourhoods?

  • To define a topology on a group which is compatible with the group structure, it suffuces to specify a filter of neighborhoods of the identity. See, say, Bourbaki, Topologie Générale §III.1.2 Prop 1. – Yai0Phah May 29 '22 at 13:09
  • I will check this out, thanks. @Yai0Phah

    I will update the question after reflecting on this reference.

    – AlgebraQuestion May 29 '22 at 13:37
  • Welcome to MSE. I think you might find useful to wander though the beginning of chapter 10 of Atiyah-Macdonald, as it is a very concise yet informative text about topologies and completions.

    As I understand what you wrote, yes, you interpret definition of $I$-adic topology correctly. Yes, of course, it is highly dependent on the ideal $I$, it even has it in the name: $\mathbf{I}$-adic topology! :)

    I didn't quite get the notation in the third question you asked, but to me it seems that the answer is yes, the construction you mention, topologically, is precisely the completion of $R$

    – Aleksei Kubanov May 31 '22 at 13:42

0 Answers0