2

A subset of the norm dual of a normed space is weak* compact if and only if it is weak* closed and norm bounded.

This is stated without a proof in:

Infinite Dimensional Analysis A Hitchhiker's Guide

Authors: Aliprantis, Charalambos D., Border, Kim Page 235 6.21.

I know that this is true for Banach spaces and one direction from weak* closed and norm bounded to weak* compact is also true. But I´m not sure if a subset of of the norm dual which is weak* compact is also norm bounded.

If not does exist a counterexample?

David C. Ullrich
  • 89,985
H.K.
  • 61
  • @ David C.Ullrich can u elaborate this? I tried it myself like: weak* compact sets are weak* bounded hence pointwise bounded but for employing UBP we need that X is a Banach space. Am I missing something? – H.K. Feb 27 '20 at 15:13
  • @WoolierThanThou: The uniform boundedness principle isn't valid when $X$ is not complete. – Nate Eldredge Feb 27 '20 at 15:56
  • 1
    @WoolierThanThou: But your sequence $x_n$ lives in $X^*$, not in $X$. – Nate Eldredge Feb 27 '20 at 15:58

1 Answers1

4

Meanwhile I think I found a counterexample

Let $X= c_{00}$ the space of finitly supportet sequences endowed with the sup norm and define for $m\in \mathbb{N}$

$$ \delta_m: c_{00}\to \mathbb{R}; \delta_m((a_n)_n) = ma_m. $$ Then $(\delta_m)_m\subset X'$ converges weakly* to $0$. Hence $\{\delta_m:m\in \mathbb{N}\}\cup \{0\}$ is weak* compact since X' endowed with the weak* topology is a hausdorff TVS. But clearly $\{\delta_m:m\in \mathbb{N}\}$ is not norm bounded since for $m\in \mathbb{N}$ $$ ||\delta_m||_{op}=m. $$

David C. Ullrich
  • 89,985
H.K.
  • 61