0
$\begingroup$

Let $P$ be a property of topological spaces - here I am interested in "compact" and "locally compact".

A topological space $X$ is $P$-ly-generated if, for any topological space $Y$, a function $f : X \rightarrow Y$ is continuous if and only if $f \circ i$ is continuous for any topological space $K$ such that $P$ and any continuous function $i : K \rightarrow X$.

An important result due to Martin Hyland says that the exponentiable objects in locales are the locally compact locales.

My question is, is compact-ly-generated equivalent to locally-compact-ly-generated?

My second question is whether anyone knows the 1979 reference to the mentioned result of Martin Hyland.

$\endgroup$
1
  • 1
    $\begingroup$ Doesn't this follow immediately from the observation that locally compact spaces are compactly generated? $\endgroup$ Commented Jul 5 at 23:41

1 Answer 1

1
$\begingroup$

If $X$ is locally compactly generated then $X$ is compactly generated because every locally compact space is compactly generated.

So given $f:X \to Y$ a map such that $f\circ i$ is continuous for every compact $i:K \to X$, then it follows that for every locally compact $j:Q \to X$ one can test the continuity of $f \circ j$ on all compact $K \to Q \to X$, hence $f \circ j$ is continuous, and as $X$ is locally compactly generated it follows that $f$ is continuous.

The problem is actually with the converse: If "compact" means "compact hausdorff" then every compact space is locally compact so the converse hold trivially, but this fails if compact means quasi-compact.

Indeed let $X$ be a non-compactly generated space (in particular it is not locally compactly generated).

Let $X^*$ be the space obtained by adding one point $*$ to $X$, with the open subset of $X^*$ being the open subset of $X$, together with $X^*$ itself.

$X^*$ is quasi-compact because a collection of open is covering if and only if it contains $X^*$ itself. But $X^*$ is not locally compactly generated: Given a map $F:X \to Y$ which is continuous when restricted to any (locally) compact space but continuous. Then we can show that the induced map $X^* \to Y^*$ will be continuous when restricted to any locally compact space, but not continuous itself, so $X^*$ is compact but not locally compactly generated.

In the last claim, I'm using that given a continuous map $j: Q \to X^*$ the preimage $Q' = j^*X$ is a locally compact space with a map to $X$ and the map $Q \to X^*$ can be factored as $Q \to Q'^* \to X^*$ with $(Q')^*$ being locally compact.

$\endgroup$
3
  • $\begingroup$ locally compact has no Hausdorff assumption for me- does that change anything here? $\endgroup$
    – user531303
    Commented Jul 7 at 14:56
  • $\begingroup$ I'm not sure I understand what you mean: I gave you an answer assuming compact means Hausdorff and one were compact has no Hausdorff assumption. I assume that in both case locally compact just means locally "compact" for whatever meaning of compact. Or are you asking about the case were compact means Hausdorff but locally compact means locally Hausdorff but not neccesarily Hausdorff (in which case, it is not different from the Hausdorff case). $\endgroup$ Commented Jul 7 at 15:25
  • $\begingroup$ Ok, I see now. Thanks so much Simon! $\endgroup$
    – user531303
    Commented Jul 7 at 16:41

You must log in to answer this question.