A topological space is countably compact iff every countably infinite subset has a limit point.

I'm completely stuck on this one. The book is recommending to use the fact that a space is countably compact iff every countable family of closed subsets which has the finite intersection property has a non-empty intersection, but I don't see how to relate this.


You need the $T_1$ axiom for one direction, though. Otherwise let $\{0,1\}$ have the indiscrete topology, and define $X = \mathbb{N} \times \{0,1\}$. There every non-empty set $A$ has a limit point (if $(x,0) \in A$, then $(x,1)$ is a limit point of $A$ and the same with $0$ and $1$ interchanged), but $\{n\} \times \{0,1\}, n \in \mathbb{N}$ is a countable cover without a finite subcover.

Or you could use another (less standard) notion of limit point, where $x$ is a limit point of $A$ iff every neighbourhood $O$ of $x$ has $O \cap A$ infinite. This is usually called an ($\omega$)-accumulation point, or a strong limit point, or some such name.

The proof for $T_1$ spaces can be found here.

For the strong limit point version: suppose $X$ is countably compact (in the cover sense) and let $A$ be an infinite subset of $X$ that has no strong (!) limit point. We can just assume $A$ is countable or we pass to an infinite subset of it. Then by assumption on $A$, for every $x \in X$ we have an open set $O_x$ such that $O_x \cap A$ is finite. To construct a countable cover without a finite subcover define for every finite $F \subseteq A$ the set $O_F = \bigcup \{O_x: O_x \cap A = F \}$. There are at most countably many finite subsets of a countable set, so $\{O_F: F \subseteq A \text{ finite } \}$ is a countable cover of $X$ (every $O_x$ is in exactly one of the $O_F$'s and the $O_x$ were a cover).

Suppose $O_{F_1},\ldots,O_{F_n}$ were a finite subcover of this open cover. There is some $a_0 \in A \setminus \cup_{i=1}^n F_i$, and this $a_0$ cannot be covered by these finitely many sets $O_{F_i}$, by construction. This contradicts the countable compactness, so $A$ must have a strong limit point.

For the reverse direction, suppose every infinite subset has a strong limit point (i.e. $\omega$-accumulation point) and suppose that $\{U_n: n \in \mathbb{N} \}$ is a countable open cover of $X$ without a finite subcover. So pick $x_1 \in U_1$ and note that $U_1$ cannot cover $X$ so we pick $x_2 \notin V_1 := U_1$. But $x_2$ is covered by some $U_{n_2}, n_2 > n_1$, and we define $V_2 = \bigcup_{i=1}^{n_2} U_i$. Then $V_2 \neq X$ (or we'd have a finite subcover), so pick $x_3 \notin V_2$. Again, $x_3 \in U_{n_3}$ for some $n_3 > n_2$, define $V_3 = \bigcup_{i=1}^{n_3} U_i$ etc. By recursion we get a set $x_n$ of all distinct points and increasing open sets $V_n$ such that $x_n \in V_n$ but $x_n \notin V_k$ for $k < n$. As the $n_k$ keep increasing, $\cup_n V_n = \cup_n U_n = X$.

Now $x$ is a strong limit point of $\{x_n: n \in \mathbb{N}\}$, which must exist. Then $x \in V_m$ for some $m$, but this $V_m$ does not contain any of the $x_k$ with $k > m$ so cannot intersect this set in an infinite set, contradiction. So $X$ is countably compact.

So for this stronger notion of limit point the statement holds. Note that in a $T_1$ space, $x$ is a strong limit point of $A$ iff it is a limit point of $A$, so if you know that, you can apply the above proof for the $T_1$ case as well.

  • $\begingroup$ Hi professor Brandsma, I have a problem with a similar proof of this result: could I ask your support? Here I posed a question where I explain my problem. $\endgroup$ Mar 21 '20 at 18:09

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.