2
$\begingroup$

I'm working with Eberlein- Smulian Theorem fromm the book "Topics in Banach Space Theory". During the proof I have seen that there is used a lot the concept of weak topology and weak star topology. For a given Banach space X according to the definitions the weak topology is defined on X and the weak star topology is defined on the dual space X* . While trying to understand the proof I see that for a set A in X there is considered the weak* closure W of A in X** . *I don't understand how we can speak about the weak and weak-star closure of a set A in X, since the weak-star closure, to me, makes sense only for sets in X **. I am really struggling with this issue! Is anybody who can tell me which is the clue here? The set A to which I am referring is used to prove that (iii) implies (i) in the last paragraph of the theorem. Any advice would be really appreciated! enter image description here

$\endgroup$
3
  • $\begingroup$ The author is thinking of $X$ embedded in $X^{\ast\ast}$ via the natural map $x \mapsto x^{\ast\ast}$, where $x^{\ast\ast}(f) = f(x)$ $\endgroup$ – Prahlad Vaidyanathan Oct 23 '13 at 17:46
  • $\begingroup$ I have seen this embedding to be mentioned in several places. The nature of X and X** is different, so how is this embedding constructed ? Thank You! $\endgroup$ – Kristina Dedndreaj Oct 23 '13 at 20:41
  • $\begingroup$ The embedding is precisely as @PrahladVaidyanathan gave it above. In more detail: If $x\in X$ then you embed that as $x^{**}\in X^{**}$ via the given formula. So $x^{**}$ needs to be a linear functional on $X^*$; you pick some $f\in X^*$ to apply $x^{**}$ to, and discover that now you have $x\in X$ and $f\in X^*$, and you combine them in the only way that makes sense: $f(x)$. Thus arriving at $x^{**}(f)=f(x)$. $\endgroup$ – Harald Hanche-Olsen Oct 23 '13 at 20:54
1
$\begingroup$

The short answer relies on two facts:

  1. $X^{**}$ is a dual space, so it has a weak* topology.
  2. $X$ can be considered a subspace of $X^{**}$.

As a side remark, the weak topology on $X$ coincides with the subscpace topology inherited from $X^{**}$ with the weak* topology, since both are derived from the functionals in $X^*$.

$\endgroup$
1
  • $\begingroup$ Yes, now I think I got the point! So, what is meant by weak* topology during the proof is the weak* topology on X**, since X can be embedded in X** and as you claim, X can be viewed as a linear subspace of X**. So, for a given a Normed Space X when speaking about weak* toplogy we may refer to it as the weak* toplogy defined on X** when we want to consider certain properties of sets in X such as compactness and closure. Now it makes sense to me talking about weak* closure and compactness of a set A in X. It seems like I'm repeating myself but I need to know if I got it right or not! Thank You! $\endgroup$ – Kristina Dedndreaj Oct 23 '13 at 21:14

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.