In "A First Look at Rigorous Probability Theory" by J. S. Rosenthal there is the following exercise:

Prove that weak limits, if they exist are unique. That is, if $\mu, \nu, \mu_1, \mu_2, \ldots$ are probability measures [on $(\Bbb{R},\mathcal{B}(\Bbb{R})$] and $\mu_n \Rightarrow \mu$, and also $\mu_n \Rightarrow \nu$, then $\mu = \nu$

Now the official solution:enter image description here

Ok, I had instead the following idea:

If $\int f \, d\mu_n \overset{n\rightarrow\infty}{\longrightarrow} \int f \, d\mu$ and $\int f \, d\mu_n \overset{n\rightarrow\infty}{\longrightarrow} \int f \, d\nu$ for any real continuous bounded function $f$, for any such function it must also hold (the limits are just limits of real numbers!) that $$\int f \, d\mu = \int f \, d\nu\, .$$

Now using those "trapezoid-like functions" $f_n$ enter image description here

we have for an half-open interval $(a, b] \subset \Bbb{R}$ pointwise limit $$\lim_{n\rightarrow \infty} f_n = \mathbf{1}_{(a, b]} \, . $$ Now using dominated convergence $$ \mu((a, b]) = \int \mathbf{1}_{(a, b]} \, d\mu = \int \lim_{n\rightarrow \infty} f_n \, d\mu = \lim_{n\rightarrow \infty} \int f_n \, d\mu = \lim_{n\rightarrow \infty} \int f_n \, d\nu = \int \lim_{n\rightarrow \infty} f_n \, d\nu = \int \mathbf{1}_{(a, b]} \, d\nu = \nu((a, b])\,. $$

The crucial step here was $\lim_{n\rightarrow \infty} \int f_n \, d\mu = \lim_{n\rightarrow \infty} \int f_n \, d\nu$. Contrary to the indicator function, the $f_n$ are continuous, so this follows, as we have seen, from $\mu$ and $\nu$ being weak limits of the same sequence of measures.

Since the measures $\mu, \nu$ are probability measures and so $\sigma$-finite and have the same value on the semi-ring of half-open intervals, by Carathéodory's extension theorem we conclude that $\mu =\nu$.

Would that solution be right, too?


Your solution is right; the only little thing is that you can justify what you call "crucial step" by the displayed equation before the picture of the function $f_n$.


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.