2
$\begingroup$

Engelking in his "General Topology" states that $T_0$ separation axiom is not preserved under open closed continuous maps.

But I can't find any example of open closed continuous image of $T_0$-space that is not $T_0$. Or, equivalently, an example of quotient space of $T_0$-space by open closed equivalence relation that is not $T_0$.

$\endgroup$
  • 2
    $\begingroup$ Perhaps you could share your thoughts about approaching the problem, even if in an unsuccessful way. Merely stating "I need X" might strike Readers as imperious. $\endgroup$ – hardmath Aug 27 '15 at 18:20
  • $\begingroup$ Also, it would help if you included your definition of $T_0$ (or at least provided a link). $\endgroup$ – Omnomnomnom Aug 27 '15 at 18:31
  • 1
    $\begingroup$ @Omnomnomnom: There is only one definition of $T_0$: if $x$ and $y$ are distinct points, then at least one of them has a nbhd that does not contain the other. $\endgroup$ – Brian M. Scott Aug 27 '15 at 18:35
  • $\begingroup$ @Brian ah... I find the $T_n$ notation confusing. Thanks. $\endgroup$ – Omnomnomnom Aug 27 '15 at 18:38
5
$\begingroup$

Let $f:X \to Y$ be the desired open and closed continuous surjection. Since $Y$ is not $T_0$, it contains two topologically indistinguishable points. If we restrict $f$ to such two-point subset of codomain, we still get a desired continuous surjection, so we may assume that $Y$ is two-point indiscrete space. Hence, it is enough to find a $T_0$ space and its decomposition into two sets $A$, $B$ such that neither $A$ nor $B$ contains a nonempty closed or open subset of $X$. In particular, $X$ cannot contain isolated point nor closed point. In particular, it cannot be $T_1$ and it cannot be finite. So let's take some easy example of infinite $T_0$ space that is not $T_1$ and let's try to find a good decomposition…

$\endgroup$
  • $\begingroup$ Do you have reason to believe this is possible? $\endgroup$ – Cameron Buie Aug 27 '15 at 19:23
  • $\begingroup$ @CameronBuie: I tried to follow the logic of what is necessary, so if it is not possible, then it is not possible to satisfy the OP's requirement. And I think I have an example. What is “the easiest” example of infinite $T_0$ space that is not $T_1$? (I would have written a complete solution, I just thought that my answer could start as a hint.) $\endgroup$ – user87690 Aug 27 '15 at 19:28
  • $\begingroup$ I'm unaware of any $T_0,$ non-$T_1$ spaces with neither closed points nor open points, so I'm at a loss. $\endgroup$ – Cameron Buie Aug 27 '15 at 19:35
  • $\begingroup$ @CameronBuie: An example of a $T_0$ space that is not $T_1$ that we had in our lecture was $ω$ with final segments as open sets… $\endgroup$ – user87690 Aug 27 '15 at 19:40
  • $\begingroup$ Under that topology, $0$ is a closed point. $\endgroup$ – Cameron Buie Aug 27 '15 at 19:45
2
$\begingroup$

@user87690, many thanks for the helpful hint) Maybe, one of the simplest examples is the following.

On the set $\mathbb Z$ of all integers with left order topology which is obviously $T_0$ consider equivalence relation with partition consisting of the set of odd and the set of even integers. This relation is open and closed, since saturation of any nonempty set is $\mathbb Z$. So there are only two saturated open sets $\varnothing$ and $\mathbb Z$. Hence, the corresponding quotient space is two point indiscrete and therefore it is not $T_0$.

$\endgroup$
  • $\begingroup$ Yes, that's exactly the example I've had in my mind. :-) $\endgroup$ – user87690 Aug 28 '15 at 10:23

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.