3
$\begingroup$

Let $X$ be a locally compact Hausdorff space. Denote $C_c(X)$ and $C_0(X)$ the space of continuous functions with compact support and vanishing at infinity respectively. By Riesz representation theorem, the dual space $C_0(X)^*$ is isomorphic to the space of finite Radon measures $M(X)$. Since $C_c(X)$ is dense in $C_0(X)$, we actually have $$ C_c(X)^* = C_0(X)^* = M(X) $$ In probability textbooks, the vague topology on $M(X)$, depending on the author, is defined as the weak* topology induced by either $C_c(X)$ or $C_0(X)$. However, these two topologies are in general not the same.

I've heard somewhere that the reason why probabilists don't care about this ambiguity is, because $C_c(X)$ and $C_0(X)$ induce the same vague topology on a closed ball in $M(X)$, and probabilists only care about measures of norm $\leq 1$ anyway.

So is the statement true? And how to prove it?

$\endgroup$
1
  • $\begingroup$ Most probabilists will probably want to use all bounded continuous functions and not restrict themselves to locally compact spaces. $\endgroup$ Feb 18 at 12:15

1 Answer 1

4
$\begingroup$

The dual unit ball of a normed space $E$ is weakly compact à la Alaoglu and hence there is no strictly coarser Hausdorff topology. Applying this to $E=C_0(X)$ you get that the topologies $\sigma(M(X),C_0(X))$ and $\sigma(M(X),C_c(X))$ coincide on all bounded sets of $M(X)$.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.