I am reading through Munkres pg 99 and it says without proof, "the real line in the finite complement topology is nota hausdorff space" i am having trouble getting started with proving this claim .. thank you for a push in the right direction.

I think this would be proved by saying along the lines that (X- $U_{2}\bigcup U_{1})$ = (X - $U_{1}$)$\bigcap (X - U_{2}) \neq \emptyset$ then not hausdorff. ?

  • 3
    $\begingroup$ Try to think of 2 open non-empty sets in this topology that don`t intersect. $\endgroup$
    – Pedro
    Oct 1 '14 at 14:32
  • 2
    $\begingroup$ What is $U_1$ and $U_2$ $\endgroup$
    – Pedro
    Oct 1 '14 at 14:38

Suppose that $X$ is equipped with the cofinite topology and that $x,y$ are distinct elements.

If it is Hausdorff then open sets $U,V$ must exist with $x\in U$, $y\in V$ and $U\cap V=\emptyset$.

Then $U\subseteq V^c$ tells us that $U$ is finite and consequently $X=U\cup U^c$ is finite.

The real line however is not finite, so...

  • $\begingroup$ does the finite complement topology have the T1 axiom? $\endgroup$ Oct 1 '14 at 15:02
  • 1
    $\begingroup$ If $x\neq y$ then $\{y\}^c$ is an open set that contains $x$ and does not contain $y$. So the topology is indeed $T_1$. $\endgroup$
    – drhab
    Oct 1 '14 at 16:28
  • $\begingroup$ @ arcolombo Also note T_2 implies T_1 implies T_0. $\endgroup$ Jan 5 '17 at 18:48

This is not true for $\mathbb{R}$ or any other infinite space $X$ with the finite complement (cofinite) topology.

Normally, if you wanted to prove $X$ is not Hausdorff, you would just need to prove that there is one specific pair of $p_1$ and $p_2$ where the two points cannot be separated by disjoint open sets around each one. However, in this case if $X$ is infinite, you can't do this for any $p_1$ and $p_2$ pair.

If you had $p_1$ and $p_2$ points in $X$ and $U$ any neighborhood of $p_1$, there is no neighborhood of $p_2$ in $X\ \backslash\ U$ because $X\ \backslash\ U$ is finite (and so could not contain any cofinite neighborhood for us to put around $p_2$).

The $\backslash$ symbol for me means complement, i.e. $X\ \backslash\ U=X\cap U^c$ which (since $X$ is the whole space) is $U^c$.

In fact, ignoring $p_1$ and $p_2$, it is impossible to have two two disjoint nonempty open sets $U_1$ and $U_2$ (by virtue of being open and disjoint, each one seems to insist that the other is finite).

This cofinite topology is nice, as an example, because it is a Fréchet space (aka T1, where Hausdorff is T2). To be T1, you must satisfy the weaker condition: For every $p_1$ and $p_2$ in $X$, there is a neighborhood of $p_1$ not containing $p_2$ and a neighborhood of $p_2$ not containing $p_1$.

Notice that the difference between this and Hausdorff is that $U_1$ and $U_2$ don't have to be disjoint -- so long as they don't contain both points.

Given points $p_1$ and $p_2$, you can take $U_1$ to be $X\ \backslash\ \{p_2\}$ and likewise for $U_2$. Clearly $U_1$ is cofinite and doesn't contain $p_2$ (its complement is just $\{p_2\}$ after all!). Likewise for $U_2$.

But of course, $U_1$ and $U_2$ have significant overlap!


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.