Let $(X, \tau) $ a compact Hausdorff topological space such that for all $x \in X$ exists $U_x \subset \tau$ countable such that $\bigcap U_x =\{ x \} $. How can I prove that $(X, \tau) $ is a first-countable space?
-
Hint use $\cup X= \cap X^c$ and properties of open covers of compact sets – Elad Jun 16 '17 at 21:08
-
@Elad I didn't get what do you mean by that. Can you explain it to me? – user435803 Jun 16 '17 at 21:14
-
1Use the formula that connects unions and intersections. Once you do that use the finite open covering in a compact space. Think how to apply those hints to your case – Elad Jun 16 '17 at 21:24
-
@Elad I don't see how it works. $\forall V\in U_x, V^c$ are closed sets. – MaudPieTheRocktorate Jun 16 '17 at 21:27
-
2Possible duplicate of If every point of a compact Hausdorff space is the intersection of a nested sequence of open sets then is the space first-countable? – MaudPieTheRocktorate Jun 16 '17 at 21:41
-
@MaudPieTheRocktorate thank you – user435803 Jun 16 '17 at 21:48
1 Answers
Define $\mathcal{B}$ to be all finite intersections of members of $U_x$ so $$\mathcal{B} = \{\cap \mathcal{F}: \mathcal{F} \subseteq \mathcal{U_x} \text{ finite }\}$$ and note that this is still a countable family of open neighbourhoods of $x$, and moreover it's closed under finite intersections.
Also note that Hausdorfness implies that
$$\cap \{\overline{O} \in U_x\} = \cap_n \{\overline{B}: B \in \mathcal{B}\} = \{x\}$$
(any $q \neq x$ has a disjoint neighbourhood with some neighbourhood from $U_x \subseteq \mathcal{B}$ and so does not lie in either intersection.) Now let's see we have a local base at $x$:
Suppose that $x \in O$, where $O$ is open. We want to show that there is some member $B \in \mathcal{B}$ such that $x \in B \subseteq O$.
Suppose that this were not the case. Then $\{X\setminus \overline{B}: B \in \mathcal{B}\} \cup\{O\}$ is an open cover of $X$ (if $p$ is in $X$ and $p \notin X\setminus \overline{B}$ for all $B \in \mathcal{B}$, then $p =x$ by the intersection property and so $p \in O$). So finitely many of them cover $X$ by compactness, say the sets $O$ and $X\setminus \overline{B_{n_1}}, \ldots, X\setminus \overline{B_{n_k}}$, for $B_{n_i} \in \mathcal{B}, i=1\ldots k$ But that these sets are a cover of $X$ implies that $\mathcal{B} \ni \cap_{i=1}^k B_{n_i} \subseteq O$ (if $p \in X$ is in all of the $B_{n_i}$, it's in none of the $X\setminus \overline{B_{n_i}}$, so it must be in $O$). This contradiction shows the claim.
- 242,131