Let X and Y be topological spaces. Prove that if $X \times Y$ is compact with respect to the product topology then X and Y are compact.

I've been stuck on this for a hour or so and I think what I'm missing is a subtle connection.


Suppose that $X \times Y$ is a compact product space. This implies that there is a collection $C_{X \times Y}=\left \{ U_{\alpha}:\alpha \in J \right \}$ of open subsets $U_{\alpha}$ that is an open cover of $X \times Y$.

In fact, every open cover of $X \times Y$ contains a finite subcollection that covers $X \times Y$.


The product topology on the product space $X \times Y$ is generated by the basis $\bar{B}=\left \{ T \times U:T \in \tau, U \in \nu \right \}=\tau \times \nu$

This implies that any element $\left ( x,y \right )$ in the basis element B of $\bar{B} \subseteq U_{\alpha}$ for any open set $U_{\alpha}$.

To show that X is compact, we need to show that there is an open cover of X that contains X and that every open cover of X contains a finite subcollection that covers X. The same applies to Y.

Any useful help to take me further is much appreciated.

  • 1
    $\begingroup$ Aren't the projections $\pi_b : \prod_{a\in A}X_a \to X_b$ continuous surjections? $\endgroup$ – MPW Oct 13 '16 at 13:19
  • 1
    $\begingroup$ You will need to assume $X,Y$ are nonempty. The projection $\varnothing \times [0,1] \to [0,1]$ may not be surjective... $\endgroup$ – GEdgar Oct 13 '16 at 13:26
  • $\begingroup$ @GEdgar: But in that case, the assertion is false. $\varnothing \times X$ is compact (since it s empty) for any $X$, whether $X$ is compact or not. So that would have to be implicit in OP's assertion if it is to be true. $\endgroup$ – MPW Oct 13 '16 at 13:33
  • $\begingroup$ Thus, my comment is to the OP, not MPW. $\endgroup$ – GEdgar Oct 13 '16 at 13:37

To proceed using open covers: Let $\{U_{\alpha} : \alpha \in J\}$ be an open cover of $X$ (say). The family $\{U_{\alpha} \times Y : \alpha \in J\}$ is an open cover of $X \times Y$. Since $X \times Y$ is compact, there exists a finite subcover, say $\{U_{i} \times Y : i = 1, \dots, n\}$. The collection $\{U_{i} : i = 1, \dots, n\}$ is a cover of $X$. Since $\{U_{\alpha} : \alpha \in J\}$ was an arbitrary cover, $X$ is compact. Then argue similarly for $Y$.

| cite | improve this answer | |
  • $\begingroup$ Promising and I'm leaning towards your contribution. Is it valid to assume the statement to be true for what we want to proof? I've always understand that to show that a space X is compact, we have to show first(conventionally) that the space X has an open cover. But it seems that we could actually validly assume space X has an open cover, then, via a finite subcollection of the open cover determine if space X is compact. $\endgroup$ – Mathematicing Oct 13 '16 at 13:49
  • 1
    $\begingroup$ Every topological space $X$ admits an open cover: Take $X$ itself (a cover containing one set), or use the axioms of a topology to pick, for each $x$ in $X$, an open set $U_{x}$ containing $x$, and consider the resulting family $\{U_{x} : x \in X\}$. $\endgroup$ – Andrew D. Hwang Oct 13 '16 at 14:00

Hint: The projection map $\pi_X\colon X\times Y\to X$ is continuous and surjective; the image of a compact space under a continuous map is…

| cite | improve this answer | |
  • $\begingroup$ Doesn't the projection map maps a set X to a set of equivalence class X*? $\endgroup$ – Mathematicing Oct 13 '16 at 13:26
  • $\begingroup$ Different use of the word "projection". $\endgroup$ – GEdgar Oct 13 '16 at 13:28
  • $\begingroup$ Which one is it then? $\endgroup$ – Mathematicing Oct 13 '16 at 13:28
  • $\begingroup$ Well, the image of a compact space under a continuous map is compact. is this just any arbitrary 'projection' map? $\endgroup$ – Mathematicing Oct 13 '16 at 13:34
  • $\begingroup$ @Mathematicing $\pi_X(x,y)=x$ $\endgroup$ – egreg Oct 13 '16 at 14:49

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.