Each function $f\in\mathbb{C}(t)$ can be rewritten in the form $f = a_{k}t^{k}+\ldots+a_0+a_1t+\ldots$, $k\in\mathbb{Z}$ and it is possible to define the topology with the open prebase at zero

$V_{n,v,\varepsilon} = \{f=a_kt^k+\ldots+a_0+a_1t+\ldots\in \mathbb{C}(t)| \sum_{k\leq m<n}|m|^v|a_m|<\varepsilon\}$

Question. Is this Lindelöf topology?

  • 14
    $\begingroup$ I don't think there's a natural locally compact topology on $\mathbb{C}(t)$: the fields admitting such a thing are called local fields and are completely classified (this might be possible to fix by asking for a weaker compatibility of the topology with the field structure, but I'm personally .skeptical) $\endgroup$ Oct 29, 2020 at 16:28
  • $\begingroup$ Thank you. What about $\sigma$ compact topology? $\endgroup$
    – solver6
    Oct 29, 2020 at 16:40
  • 3
    $\begingroup$ Each one of your prebasic open sets is $t$-adically open (the set contains the open disk of radius $p^{-n}$ around each of its points). But for every $n$, the $t$-adic disk of radius $p^{-n}$ centered at $0$ contains $V_{n+1,n+2,\{1\}}$. So it seems to me that this topology is the same as the $t$-adic topology. $\endgroup$ Oct 29, 2020 at 18:14
  • 1
    $\begingroup$ This seems vaguely like dealing with the box topology instead of the product topology, and probably runs into similar issues. $\endgroup$
    – LSpice
    Oct 29, 2020 at 19:16
  • 1
    $\begingroup$ What if we consider the inverse limit of topological rings $\mathbf{C}[t]/t^n$ each with the Euclidean topology? This gives a mixture of the $t$-adic and Euclidean topologies on $\mathbf{C}[[t]]$. I didn't check, but it may be that this topology does not extend to a ring topology $\mathbf{C}((t))$. $\endgroup$ Nov 1, 2020 at 8:53

1 Answer 1


Let $\|f\|_{n, m} = \sum_{i<n}(|i|^m+10)|a_i|$ denote a family of seminorms on $\mathbb{C}(t)$. Note that for any $\{(n_v, m_v)\}_{v\text{ in finite subset of }\mathbb{N}}$ there is a seminorm $\|.\|_{n,m}$ such that $\|f\|_{n, m}\geq\sup_v\|f\|_{n_v, m_v}$. Also these seminorms define the same topology as stated in the problem. Resulting topological space is separabel so for some countable set of $F_i\in\mathbb{C}(t)$ this set is dense. Let $\{U_{\lambda}\}$ be some open cover of this topological space. For each $F_i$ and $(n, m)$ define $\varepsilon_{i, n, m}$ as supremum of $\varepsilon >0$ such that the $\varepsilon$ ball around $F_i$ in $\|.\|_{n, m}$ lies inside of $U_{\lambda}$ for some $\lambda$ and let $\{\|f\|_{n, m}<\varepsilon_{i, n, m}/2\}\subset U_{\lambda_{i, n, m}}$ in result $\cup_{i, n, m}U_{\lambda_{i, n, m}}$ is countable open subcover.


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.