3

Let $(X, \tau)$ be a regular, scattered and Lindelöf space with countable pseudocharacter. Is it true that $X$ needs to be countable?

What I have tried (using the Theorem 2.2 from M. E. Gewand, “The Lindelöf degree of scattered spaces and their products,” Journal of the Australian Mathematical Society (Series A) 37 (1984), 98–105): Let $(X, \tau)$ be a regular scattered Lindelöf space with countable pseudocharacter.
(Before going on, $X_{\delta}$ denotes the topological space of $X$ with the topology generated by its $G_{\delta}$ sets under the topology $\tau$).

As a consequence of the Theorem 2.2 ($X$ is $T_3$, scattered and Lindelöf), $X_{\delta}$ is Lindelöf. Now, as each point of $X$ is a $G_{\delta}$ set, then $X_{\delta}$ is discrete and $\mathcal{A} = \{ \{x\} : x \in X \}$ is an open cover of $X_{\delta}$. Finally, as $X_{\delta}$ is Lindelöf, $X$ needs to be countable.

amWhy
  • 210,739
Almanzoris
  • 1,477
  • 1
    Was this your assumption, or did you find this as a statement / question somewhere? – Ulli Aug 05 '24 at 20:05
  • I was reading the question from MO: https://mathoverflow.net/q/416331 and the author said: "It is known that any first countable, T3, Lindelöf and scattered space is countable".

    I wasn't aware of this result, so I tried to look for its proof and I found that it appears stated as the Corollary 2.5 (with a typo) in the citation I make in this question, but its proof would say that it's a consequence of the Theorem 2.2 from that same citation. In this question, I show my attempt to prove the corollary using the theorem, but I wanted to ask for clarifications or reviews and more direct proofs.

    – Almanzoris Aug 05 '24 at 20:58
  • 1
    Thank you for your comment. About your question: yes, of course, your derivation of the corollary is correct. Moreover, PatrickR gave a direct proof. – Ulli Aug 06 '24 at 06:10
  • You're welcome. Thank you too for the review. – Almanzoris Aug 06 '24 at 11:23

1 Answers1

2

Here is a direct proof.

Lemma: Every Lindelöf locally countable space is countable.

This is easy to check (for example see Must scattered spaces with points $G_\delta$ be locally countable?).

Theorem: Every regular Lindelöf scattered space with each point a $G_\delta$-set is countable.

Proof: Suppose $X$ is regular Lindelöf scattered and each point is a $G_\delta$-set. Scattered spaces are $T_0$ (see here), and $T_0$ regular spaces are $T_3$. So $X$ is Hausdorff and $T_1$.

Let $B$ be the set of points $x\in X$ at which $X$ is locally countable (i.e., such that $x$ has a countable nbhd). The set $B$ is open in $X$. So its complement $A=X\setminus B$ is closed, hence Lindelöf. Suppose by contradiction that $A$ is not empty. By the scattered property, there is a point $x\in A$ that is isolated in $A$. So there is an open set $U\subseteq X$ with $U\cap A=\{x\}$.

Since the singleton $\{x\}$ is a $G_\delta$, we have $\{x\}=\bigcap_{n=1}^\infty G_n$ for some open sets $G_n\subseteq U$. By regularity of $X$ we can also assume $\overline G_1\subseteq U$.

Each of the $\overline G_1\setminus G_n$ is closed in $X$, hence Lindelöf. And it is also locally countable as a subset of the locally countable open set $U\setminus\{x\}\subseteq B$. Thus $\overline G_1\setminus G_n$ is countable by the Lemma. And since the set $G_1\setminus\{x\}$ is covered by these countably many countable sets, it is also countable. But then $G_1$ is a countable nbhd of $x$, which contradicts the fact that $x\in A$. Therefore $A=\emptyset$ and $X$ is locally countable. Finally, one more application of the Lemma shows that $X$ is countable.

PatrickR
  • 7,165
  • 1
    Very nice proof (+1). I was just wondering whether regular can be weakened to Hausdorff. According to this paper it cannot! – Ulli Aug 05 '24 at 12:42
  • 1
    This space https://topology.pi-base.org/spaces/S000171 (which isn't regular) is functionally Hausdorff (so also Hausdorff), scattered, Lindelöf and with countable pseudocharacter, but it has uncountable cardinality. – Almanzoris Aug 05 '24 at 21:11
  • 1
    @Ulli Thanks for the very interesting reference. – PatrickR Aug 06 '24 at 00:55
  • 1
    @Almanzoris: That's very strange: before finding the above cited paper, I also looked up pi-base with this query. But it didn't provide Brian's example, although it matches all criteria. – Ulli Aug 06 '24 at 06:20
  • Yeah, it didn't provide it, because Brian's example didn't have the regular property included (as not being regular) in its properties list yet. This theorem has been added to pi-Base after the question has been answered and, as a result, Brian's example has got that deduced from it and from the other properties that it had already in its own list. – Almanzoris Aug 06 '24 at 11:37
  • 1
    @Almanzoris: Brian's example is listed as non-regular (at least just now). However, the query still does not provide it. – Ulli Aug 06 '24 at 14:53
  • 1
    @Ulli Have you synced the main branch in the advanced section of pi-Base?? – Almanzoris Aug 06 '24 at 15:17
  • 1
    @Almanzoris: ah, now I see it. I didn't know that one has to reset pi-base. Anyway, that I didn't see it in the beginning had the big advantage that I kept searching for examples and so found this quite interesting paper cited above. THANK YOU VERY MUCH FOR YOUR SUPPORT! – Ulli Aug 06 '24 at 17:31
  • @Ulli jajaja, it also happened to me the first time. You're welcome. And thank you for having provided such rich paper. Have a great day, sir. – Almanzoris Aug 06 '24 at 17:39
  • 1
    @Almanzoris fyi, I have asked https://math.stackexchange.com/questions/4955301 for a related question. It should be answerable with similar ideas. – PatrickR Aug 06 '24 at 18:21
  • Thank you for the information. Also, despite I hadn't said it yet, thank you for your great proof of your answer to this question, Patrick. – Almanzoris Aug 07 '24 at 13:14