3
$\begingroup$

H.H. Corson in [C] introduced the following version of Lindelöf property for convex closed subsets of Banach spaces:

A Banach space $X$ has property (C) if every family of convex closed subsets of $X$ with empty intersection contains a countable subfamily with empty intersection.

For instance, all weakly Lindelöf Banach spaces have property (C), but there is an example of a space with property (C) which is not weakly Lindelöf (see [C]).

There are several characterisations of spaces with property (C), eg. in terms of duals (cf. [P]):

A space $X$ has property (C) if and only if for every $A\subseteq B_{X^*}$ and $f\in\overline{A}^{w^*}$ there exists countable $A'\subseteq A$ such that $f\in\overline{\text{conv}}^{w^*}A'$.

The property (C) passes to closed subspaces and quotients. Thus, a natural question concerning products arises:

Assume $X$ has property (C). Does $X\times X$ have it as well?

I haven't been able to find any answers or hints concerning this question. However it seems to be basic... I am especially interested in the case of spaces $C(K)$ of continuous real-valued functions on a compact space $K$ with the supremum norm:

Assume $C(K)$ has property (C). Do $C(K)\times C(K)$ and $C(K\times K)$ have also property (C)?

Thank you very much for any answer or comment.

References

[C] H.H. Corson, The weak topology of a Banach space, Trans. Amer. Math. Soc. 101 (1961), 1–15.

[P] R. Pol, On a question of H.H. Corson and some related problems, Fund. Math. 109 (1980), 143–154.

$\endgroup$
1
  • 1
    $\begingroup$ The analogy with Lindelöfness might suggest it to be false, as Lindelöf is not productive at all (e.g. the Sorgenfrey line). But maybe the structure of convex sets helps... $\endgroup$ Commented Jun 16, 2014 at 6:58

1 Answer 1

2
$\begingroup$

Yes, if $X$ and $Y$ have property (C), then so does $X\oplus Y$ because property (C) is a three-space property; this is a result of Pol (Proposition 1 in the paper you are referring to).

However, I am not sure if it is known whether property (C) passes from $C(K)$ to $C(K\times K)$. (If this is true, it should follow rather easily from the identification $C(K\times K) = C(K)\otimes_{\varepsilon}C(K)$ and the structure of convex sets in the injective tensor products---otherwise perhaps there is a counter-example.)

$\endgroup$
1
  • $\begingroup$ Thanks, Tomek. I have known the result of Pol you refer to, but completely didn't link it with my question. $\endgroup$ Commented Jun 18, 2014 at 23:42

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .