Let $\mathbb{Q}$ be the rationals, and let $\tau$ be the Euclidean topology on $\mathbb{Q}$. Is there a topology $\tau' \subseteq \tau$ such that $(\mathbb{Q},\tau')$ is connected and $T_2$?


Using Sierpinski's topological characterization of $\mathbb Q$ (as a unique countable regular second countable space without isolated points), it can be shown that $\mathbb Q$ is homeomorphic to the set $\mathbb N$ endowed with the Furstenberg topology $\tau$ generated by the base consisting of all possible arithmetric sequences $a+\mathbb N_0b:=\{a+bn:n\ge 0\}$ with $a,b\in\mathbb N$.

The Furstenberg topology contains the Golomb topology $\tau'$ on $\mathbb N$, which is generated by the base consisting of arithmetric sequences $a+\mathbb N_0b$ where $a,b$ are comprime.

It is well-known (and can be easily shown using the Chinese Remainder Theorem) that the Golomb topology is Hausdorff and connected.

  • $\begingroup$ Why do you call this "Furstenberg topology"? it seems it's plainly the profinite topology. $\endgroup$ – YCor May 19 '18 at 15:04
  • $\begingroup$ Why do you answer the question instead of closing it as a duplicate of the other question where you yourself have given the accepted answer? $\endgroup$ – Alex M. May 19 '18 at 15:18
  • $\begingroup$ @AlexM. Yes, you are right. Simply I do not remember all answers I give though I realize that this is very close to what was have been discussed earlier (especially the Golomb topology). I thought that the system shows all the related questions and answers at the moment of posing the question. So, the author always has a possibility to look and analyse what has already been asked and answered prior to posting his own question. $\endgroup$ – Taras Banakh May 19 '18 at 15:22
  • 1
    $\begingroup$ @YCor The term "Furstenberg topology" is standard and well-accepted in this area, see the seminal paper of Golomb (dml.cz/bitstream/handle/10338.dmlcz/700933/…) and the recent paper (alpha.math.uga.edu/~pete/CLLP_November_30_2017.pdf) $\endgroup$ – Taras Banakh May 19 '18 at 15:24
  • $\begingroup$ @YCor, this is the pro-group topology on the non-negative integers. The full profinite topology is discrete since it allows homomorphisms to any finite monoid not just fine groups. $\endgroup$ – Benjamin Steinberg May 19 '18 at 15:35

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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