This is related to an older question.

Let $(X,\tau)$ be a topological space. Trivially, the indiscrete topology $\{\emptyset, X\}$ is a connected subtopology of $\tau$.

Is there a connected topology $\tau_c\subseteq \tau$ on $X$ such that whenever $\tau_c'$ is a connected topology on $X$ such that $\tau_c\subseteq\tau_c'\subseteq \tau$ then $\tau_c=\tau_c'$?

  • 1
    $\begingroup$ The answer I gave below is incorrect; as pointed out by a comment to it, the union of a chain of topologies is not necessarily a topology. I think you should un-accept it. $\endgroup$ – user44191 Dec 3 '18 at 19:15
  • $\begingroup$ Thanks for putting the effort into it... I also missed the fact that the union of a chain of topologies need not be a topology! Can you delete your answer and post as a comment, why a straightforward application of Zorn's lemma doesn't work? $\endgroup$ – Dominic van der Zypen Dec 3 '18 at 20:27
  • 1
    $\begingroup$ The straightforward application of Zorn's lemma doesn't work because one of the conditions on topologies uses infinitely many sets (in fact, arbitrarily many). An example: let $X = \mathbb{N}, \tau_n$ be the topology that doesn't distinguish between $m \geq n$, and otherwise is discrete. Then $\cup \tau_n$ is the set of "eventually constant" sets (and even is countable). But it does include each of the point-sets, which contradicts the "arbitrary unions" condition on topologies. Therefore, Zorn's lemma can't be directly applied. $\endgroup$ – user44191 Dec 3 '18 at 20:46
  • 1
    $\begingroup$ @ForeverMozart -- works fine if $\tau$ is connected, but if it isn't? $\endgroup$ – Dominic van der Zypen Dec 4 '18 at 7:56
  • 4
    $\begingroup$ Split the real line into countably many disjoint dense sets $A_n$, e.g., $A_n=\mathbb{Q}+n\pi$ for $n\ge1$ and $A_0=\mathbb{R}\setminus\bigcup_{n\ge1}A_n$. Let $\tau_n$ be the topology generated by the usual topology together with $\{A_i:i\le n\}$. Each topology is connected, yet the union generates a topology with a pairwise disjoint cover by open sets. $\endgroup$ – KP Hart Dec 5 '18 at 8:32

Your Answer

By clicking "Post Your Answer", you acknowledge that you have read our updated terms of service, privacy policy and cookie policy, and that your continued use of the website is subject to these policies.

Browse other questions tagged or ask your own question.