Let $X$ be a compact metric space and $\mathcal{M}(X)$ be the set of probability measures on $X$ equipped with the topology of weak convergence. So $\mathcal{M}(X)$ itself can be viewed as a compact metrizable space. Let $A\subseteq X$ be any measurable set in $X$, is the map $$\mathcal{M}(X)\rightarrow [0,1]~,~~P\mapsto P(A)$$ measurable?


The map $P\mapsto \int f \,dP$ from $(M(X),{\cal B}(M(X)))$ to $(\mathbb{R},{\cal B}(\mathbb{R}))$ is measurable for all bounded, Borel $f$ on $X$. In particular, for $f={\bf 1}_A,$ the map $P\mapsto P(A)$ is measurable.

Proof: If $f$ is a continuous function, then $P\mapsto \int f \,dP$ is continuous by the definition of weak convergence of measures, and hence is also measurable.

Now apply the functional Monotone Class Theorem. We use Theorem 2 at the link, with $${\cal H}=\left\{f\in {\cal B}_b(X): P\mapsto \int f\,dP\mbox{ is measurable}\right\},$$ and conclude that ${\cal H}={\cal B}_b(X).$

  • $\begingroup$ Great! I didn't know about the functional Monotone Class Theorem. Thank you. $\endgroup$ – user240643 May 19 '16 at 17:32

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.