Let $X$ be a metric space in which every infinite subset has a limit point. Prove that $X$ is compact.

The following is my proof I'd like to know if it is correct.


I will use the fact that if $X$ is a metric space in which every infinite subset has a limit point, then $X$ has a countable base, so every open cover of $X$ has a countable subcover $\{G_n\}$. To lead contradiction, suppose that there is no finite subcollection of $\{G_n\}$ which covers $X$. Then the complement $F_n$ of $G_1 \cup \dots \cup G_n$ is nonempty for each $n$, while clearly $\bigcap F_n$ is empty. So we can choose $x_1$ that doesn't belong to $G_1$ and $x_2$ distinct from $x_1$ that doesn't belong to $G_1 \cup G_2$ and recursively, get a sequence of distinct points $\{x_n\}$ such that each $x_n$ is from $F_n$. Let's call this set $E$ then $E$ is infinite (since if it is finite then $E=\{x_1,\dots, x_k\}$. Then each $x_i$ is covered by some $G_m$, we can take a union of at most $k$ such covers and since it is only a finite union, by assumption, there exists another point outside the union that belongs to $E$.)

Now by hypothesis, $E$ has a limit point $z$. Then, $z$ belongs to some open cover $G_m$ and so there is an open ball $B(z;d)\subset G_m$. But then, by our construction, all $x_n$ for which $n \ge m$, do not belong to $B(z;d)$. So there is an open ball containing $z$ which doesn't contain infinitely many points of $E$, which contradicts the assumption that it is a limit point. Thus, by way of contradiction, there must be a finite subcover. QED.

This is my proof but I have a couple points that I'm not sure of. First, how do I guarantee that I can select all distinct points of $x_n$ by the above process? Also, is my proof that $E$ must be an infinite set correct? Is there anything incorrect about the rest of the proof? Thanks.


If no finite subcollection of $\{G_{n}\}$ covers X, then $F_{1} = G_{1}^{c}\neq\emptyset$, so choose $x_{1}\in F_{1}$. Having chosen $x_{1}, . . .,x_{j} \in X$, choose $x_{j+1} \in X$, if possible, so that $x_{j+1} \in F_{j+1} \setminus \{x_{1}, ...,x_{j}\}$. If I prove that this process must not stop after a finite number of steps, then you will get distinct $x_{n}'s$ and the set E will be infinite.

suppose this process stops after a finite number of steps, say k, then $F_{k+1}\setminus \{x_{1},...,x_{k}\}=\emptyset$. Then $G_{1}\cup ....\cup G_{k+1}\cup\{x_{1},...,x_{k}\}=X$. Since $\{G_{n}\}$ covers X, there exists $G_{m_{i}}$ such that $x_{i}\in G_{m_{i}}$ for $i = 1,...,k$. Now $G_{1}\cup ...\cup G_{k+1}\cup G_{m_{1}}\cup ...\cup G_{m{k}} = X$, a contradiction.


Here's part of my proof:

Suppose that E contains a point of each $F_{n}$. If E is infinite, then it has a limit point $p$. Notice that if we had $p\in G_{n}$ for some $n$, then there would be a neighborhood with no points of $F_{n+1}$. Then $p\notin G_{n}\forall n$, that is, $p\in \bigcap F_{n}$.

Hint: Consider the set $\bigcup_{n}F_{n}$. What can be said about the relationship between $F_{n}$ and $F_{n+1}$? After answering that, consider the set $\bigcap_{n}F_{n}$ to get a contradiction.


Any sequence will then have a limit point, so a subsequence of that sequence will converge to the limit point.


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.