7
$\begingroup$

I wonder if there is a compact and locally Hausdorff space $X$ which is not locally compact, in the sense that every point has a neighborhood base consisting of compact sets.

A space is called locally Hausdorff if every point has a neighborhood which is a Hausdorff space under the subspace topology. Such a space need not be Hausdorff. However, if each point has a closed Hausdorff neighborhood, then the space is Hausdorff. An example is the real line with two origins. If one restricts this space between $-1$ and $1$, it becomes compact. Unfortunately, it is also locally compact.

If $X$ is Hausdorff and compact, then it is also locally compact, so the Hausdorff neighborhoods must all be proper subsets of $X$.

Does anyone have an idea?

$\endgroup$
5
$\begingroup$

Here’s a counterexample to the conjecture.

Let $Y=\Bbb N\times\Bbb N$, let $p$ and $q$ be distinct points not in $Y$, and let $X=Y\cup\{p,q\}$. Points in $Y$ are isolated. For each $k\in\Bbb N$ the set

$$B_p(k)=\{p\}\cup\left\{\langle m,n\rangle\in Y:n\ge k\right\}$$

is a basic open nbhd of $p$. For each $k\in\Bbb N$ the set

$$B_q(k)=\{q\}\cup\left\{\langle m,n\rangle\in Y:m\ge k\right\}$$

is a basic open nbhd of $q$. For convenience, for $n\in\Bbb N$ let $S_n=\{n\}\times\Bbb N$.

Each open nbhd of $p$ contains all but finitely many points of each $S_n$, and each open nbhd of $q$ contains all but finitely many of the sets $S_n$. Thus, every open set containing both $p$ and $q$ contains all but finitely many points of $Y$, and $X$ is compact. $Y$ is Hausdorff, as are the sets $B_p(k)$ and $B_q(k)$, so $X$ is locally Hausdorff. However, none of the sets $B_p(k)$ contains a compact nbhd of $p$, so $X$ is not locally compact. (Similarly, $X$ fails to be locally compact at $q$.)

$\endgroup$
  • $\begingroup$ That's a great counterexample. It's like taking $\mathbb N$ with two points at infinity and then blowing it up, so it ceases being locally compact. Thanks a lot! $\endgroup$ – Stefan Hamcke Mar 22 '13 at 15:13
  • $\begingroup$ @Stefan: You’re welcome! It’s sort of turning the points at infinity crosswise to each other. $\endgroup$ – Brian M. Scott Mar 22 '13 at 17:08

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.