I was reading the proof here on the claim that a closed subset of a compact set is compact, which reads:

Say F ⊂ K ⊂ X where F is closed and K is compact. Let $\{V_α\}$ be an open cover of F. Then $F^c$ is a trivial open cover of $F^c$. Consequently { $F^c$} ∪ $\{V_α\}$ is an open cover of K. By compactness of K it has a finite sub-cover – which gives us a finite sub-cover of F.

The proof has to add $\{F^c\}$ to an open cover of of F so that it covers K. What about the open cover $\{V_α\}$ without $\{F^c\}$? How can one be sure that it has a finite subcover for F since it is not necessarily true that $\{V_α\}$ covers K?


I want to put the question differently. How do I know that there is not an open cover that covers F but not K?

  • $\begingroup$ But when you add the set $F^{c}$, that's just adding one more set. $\endgroup$ – quasi Nov 6 '17 at 7:48
  • $\begingroup$ I misread your statement. But you are sure, because as quasi has remarked,you have just added one set, so you can subtract it as well (if the process of going to a finite subcover hasn't done that for you already). $\endgroup$ – B. Pasternak Nov 6 '17 at 7:49
  • $\begingroup$ @quasi The open cover without $\{F^c\}$ should have a finite subcover as well, right? $\endgroup$ – user1691278 Nov 6 '17 at 7:49
  • $\begingroup$ @B.Pasternak I thought you need to show that every open cover has a finite subcover. The open cover $\{V_α\}$ without $\{F^c\}$ should have a finite subcover too. How can I show that? $\endgroup$ – user1691278 Nov 6 '17 at 7:50
  • $\begingroup$ @quasi I updated the question. $\endgroup$ – user1691278 Nov 6 '17 at 7:54

How do I know that there is not an open cover that covers $F$ but not $K$?

You do not know. For example if $F = K$, then every open cover of $F$ also covers $K$. But this is none of your business.

I guess I understand your confusion now. Let me spell out the proof.

Your assumption is that $\{ V_\alpha\}_\alpha$ covers $F$. So $\{ V_\alpha\}_\alpha \cup F^c$ covers $K$. From there you can use the assumption that $K$ is compact and pick a finite subcover. The finite subcover looks like

$$ \{ V_{\alpha_1}, V_{\alpha_2} , \cdots ,V_{\alpha_k}, F^c\}.$$

This covers $K$ and in particular covers $F$. But now we can take $F^c$ away: we know by definition of $F^c$ that $F\cap F^c$ is empty anyway. Thus

$$ \{ V_{\alpha_1}, V_{\alpha_2} , \cdots ,V_{\alpha_k}\}$$

covers $F$ and is a finite subcover of $\{V_\alpha\}_\alpha$.

  • $\begingroup$ Well said. Thank you! $\endgroup$ – user1691278 Nov 6 '17 at 8:07

You don't have to know that $V_{alpha}$ doesn't cover $K$, if it happens to cover $K$ then still adding $F^C$ wouldn't alter anything - the resulting cover would still cover $K$.

Now of course in the subcover of $K$ including $F^C$ we must know that the subcover excluding $F^C$ would cover $F^C$, but this is rather obvious. If $x\in F$ then it would be in at least one of the set in the subcover, but it can't be in $F^C$ by definition so it must be in at least another of the sets in the subcover.

  • $\begingroup$ Well, $V_\alpha$ has to produce a finite subcover for F without having $F^C$ added to it, right? Assuming that $V_\alpha$ is an open cover for F but not K. $\endgroup$ – user1691278 Nov 6 '17 at 8:01
  • $\begingroup$ @user1691278 Yes, the finite subcover of $K$ will produce a subcover of $F$ by removing $F^C$ - I've added a comment about that in my answer. $\endgroup$ – skyking Nov 6 '17 at 8:20

The idea of the proof is that, according to the definition of compactness, every open cover for $K$ has a finitie subcover. Now we found an open cover for $K$, containing only one set more than the open cover for $F$, we conclude that this set has a finite subcover. Since $F^c \not\subset F$, we know that there exists a subset of the finite subcover for $K$ also covering $F$, that does not contain $F^c$, and is finite since it contains as most as many sets as the finite cover for $K$.

To answer your update, there exist open covers for $F$ not covering $K$, but by construction, adding $F^c$ makes suren the open cover also covers $K$. That makes the possibility you're asking about irrelevant.


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.