I am using the definition of compactly generated space from The Category of CGWH Spaces, which is

In $\mathbf{Top}$, a $k$-closed subset $Y\subset X$ is a set such that $u^{-1}(Y)$ is closed in $C$ for any $u: C\to X$ where $C$ is compact Hausdorff.

A space is compactly generated if all $k$-closed subsets are closed.

locally compact means every point has a local base of compact sets.

This is different from the definition from Wikipedia

So, is any compact space compactly generated?

And, is any locally compact space compactly generated?

  • $\begingroup$ For your first question, if $X$ is compact Hausdorff, consider taking $C=X$ and $u$ the identity. $\endgroup$ May 6, 2014 at 5:12
  • $\begingroup$ Yes, I know this. But I want to know what about non-Hausdorff space $\endgroup$ May 6, 2014 at 5:59
  • $\begingroup$ The term compactly generated is used differently by different authors. Sometimes is denotes a space where a subset is closed if it intersects each compact subset in a closed set, let's call these c-spaces. Other times it denotes a space as you describe it, but these are also called k-spaces. Both compact and locally compact spaces are c-spaces. So you are actually searching for a space that is a c-space but not a k-space. $\endgroup$ May 19, 2014 at 14:48
  • $\begingroup$ It seems that both your question and this require a space with many compact subsets which are not images of compact Hausdorff spaces. $\endgroup$ May 19, 2014 at 14:49
  • $\begingroup$ By the way, is there any explicit example of a compact space, even non-T₀ one, to where a Hausdorff compact can’t be surjectively mapped? $\endgroup$ Dec 17, 2014 at 12:37

1 Answer 1


It seems the following.

We shall use this question by Amathstudent and its answer by Brian M. Scott.

We prove that each weak Hausdorff compactly generated $T_1$ space is $KC$. Since there is a weak Hausdorff compact $T_1$ space, which is not $KC$ (see the space $\Bbb Q^*\times\Bbb Q^*$ in the answer by Brian M. Scott), this space is not compactly generated.

So, let $X$ be a weak Hausdorff compactly generated $T_1$ space and $Y$ be a compact subset of $X$. We claim that $Y$ is $k$-closed subset of $X$. Indeed, let $C$ be a compact Hausdorff space and $u: C\to X$ be a continuous map. Since the space $X$ is weak Hausdorff, the set $u(C)$ is closed in $X$. A set $u(C)\cap Y$ is compact as a closed subspace of a compact space $Y$. By Lemma 1, the space $u(C)$ is Hausdorff. So the set $u(C)\cap Y$ is closed in the space $u(C)$. Since the set $u(C)$ is closed in $X$, the set $u(C)\cap Y$ is closed in the space $X$ too. Since the map $u$ is continuous, the set $u^{-1}(Y)= u^{-1}(Y\cap u(C))$ is closed in $C$. Since the space $X$ is compactly generated, the set $Y$ is closed in $X$. Hence $X$ is a $KC$-space.

  • $\begingroup$ Can you show a k-closed subset in your example that isn’t closed? Is it the diagonal? $\endgroup$ Dec 17, 2014 at 14:45
  • $\begingroup$ @IncnisMrsi The diagonal is not closed as remarked Brian M. Scott, but it should be $k$-closed as I claimed in the old version of my answer. $\endgroup$ Dec 17, 2014 at 14:53
  • 2
    $\begingroup$ @AlexRavsky Brilliant. Just a minor note: weakly Hausdorff spaces are already $T_1$ (any point is closed being the image of some constant function $u\colon K\to X$). What about the second question about locally compact spaces? In $\mathbb{Q}^*\times\mathbb{Q}^*$ no point has a local base of compact sets so it does not disprove the conjecture "locally compact implies compactly generated". $\endgroup$
    – johndoe
    Dec 17, 2014 at 17:22
  • 1
    $\begingroup$ Incidentally, your result shows that not every compact space is homeomorphic to a quotient of some locally compact Hausdorff space (compare this to the fact that every space is homeomorphic to a quotient of some Hausdorff space). $\endgroup$
    – johndoe
    Dec 17, 2014 at 21:18
  • 1
    $\begingroup$ @IncnisMrsi: The diagonal in $\Bbb Q^*\times\Bbb Q^*$ is indeed $k$-closed. This can also be deduced as follows: Since $\Bbb Q^*\times\Bbb Q^*$ is weak Hausdorff, so is its $k$-ification $k(\Bbb Q^*\times\Bbb Q^*)$ (this equippes the space with the final topology with respect to all maps from compact Hausdorff spaces). But a $k$-space is weak Hausdorff iff the diagonal is closed, see Neil Stickland's paper CGWH. That means the diagonal is $k$-closed, but it's not closed as $\Bbb Q^*\times\Bbb Q^*$ is not Hausdorff. $\endgroup$ Dec 17, 2014 at 21:21

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.