In the context of a bigger proof, I would like the following to hold true:

Claim. Let $(\mu_j)_{j\in\mathbb{N}}$ be a sequence of positive Borel measures on $\mathbb{C}$ that converges weakly toward $\mu$, a probability measure on $\mathbb{C}$, that is for all $\varphi\in C_0^\infty(\mathbb{C})$, one has: $$\lim_{j\to+\infty}\int_\mathbb{C}\varphi\,\mathrm{d}\mu_j=\int_\mathbb{C}\varphi\,\mathrm{d}\mu.$$ Then, the following limit holds: $$\lim_{j\to+\infty}\mu_j(\mathbb{C})=\mu(\mathbb{C}).$$

I know that for each positive Borel measure $\nu$ on $\mathbb{C}$, one has: $$\nu(\mathbb{C})=\sup\left\{\int_\mathbb{C}\varphi\,\mathrm{d}\mu_j;\varphi\in C_0^\infty(\mathbb{C}),0\leqslant\varphi\leqslant 1\right\}.$$ Therefore, I kind of want to invert $\lim$ and $\sup$, but this does not really make sense (to me).

Any enlightenment will be greatly appreciated!

  • 2
    $\begingroup$ It's not true without more assumptions. Let $\mu_j$ be a point mass at $z_j$, where $|z_j| \to \infty$, and let $\mu$ be the zero measure. $\endgroup$ – Nate Eldredge Oct 16 '17 at 17:04

I don't think this is the case. Take any sequence of measures $\mu_n$ that weakly approximates the unit point mass $\mu = \delta_0$ at the origin. Now I could perturb my approximating sequence by putting some additional mass to $\mu_n$, supported in the annulus $A_n$ consisting of complex numbers of modulus in between $n$ and $n+1$. Note that the perturbed sequence still converges weakly to $\delta_0$, but the total mass of each $\mu_n$ has changed.

Some kind of restriction prohibiting the mass from "escaping to infinity" will be necessary.


It will hold for bounded sets $C$, but not necessarily for the whole space. You can approximate a bounded set from above by an open subset and from below by a compact subset. Then fit between the compact and the open a smooth function.


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.