3

I have some question on functional analysis. Recently, I'm reading an article of Coifman and Weiss, "Extensions of hardy spaces and their use in analysis".

They proved some important theorem to me by using the following functional analysis fact without proof.

Exercise (Dunford&Schwartz, p.439, #41). Let $X$ be a locally convex linear topological space and let $G$ be a linear subspace of $X^*$. Then $G$ is $X$-dense in $X^*$ if and only if $G$ is a total set of functionals on $X$.

In the case of Banach space, I can prove the one direction ($G$ is a total implies $G$ is dense in $X^*$ in the sense of weak* topology).

First, I proved that if $N$ is a subspace of $X^*$, then $(^\bot N)^\bot = \overline{N}^*$, where $-^*$ denotes the weak* closure. Here I follow Rudin's notation. So the result of one direction is proved one takes $^\bot N=\{0\}$.

But I fail to prove its reverse direction and I cannot extend to the locally convex linear topological space case.

Thank you in advance.

Will Kwon
  • 1,175

1 Answers1

3

Assume $G$ is not total. Then there is some $x\in X$ ($x\neq 0$) such that $x^*(x)=0$ for all $x^*\in G$. Thus the set $\{x^*\in X^*:x^*(x)\neq 0 \}$ is a weak$^*$ open set which is disjoint from $G$, so $G$ is not weak$^*$ dense.

Aweygan
  • 23,232