MathOverflow is a question and answer site for professional mathematicians. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

I have a signed measure $\mu$ on a convex subset $C\subset \mathbb{R}^n$, and I want to prove that $\mu$ is a probability measure, most importantly that it is positive everywhere.

I do know that $\int f(x)d\mu(x)\geq 0$ for any positive CONVEX function $f$. So if I could get this inequality for indicator functions I'd be done.

Do you know if this suffices to get that the measure is positive, or maybe have a counterexample?

share|cite|improve this question
If $C=(a,b)$ is an interval, your assumption is that $\langle \mu,f\rangle\ge0$ for every $f$ such that $f''\ge0$ in distributional sense. Therefore an equivalent statement is that $\mu=N''$ where $N\ge0$ and $N(a)=N(b)=0$. If $n\ge2$, a characterization must be more involved. – Denis Serre Oct 13 '11 at 6:40
up vote 10 down vote accepted

A counterexample is a signed measure on the interval $I:=[-1,1]$ concentrated in the points $\{-1\}$, $\{0\}$, $\{1\}$ with weights respectively $1/2$, $-1$, $1/2$. (Thus $\int_If d\mu= f(1)/2+f(-1)/2 - f(0)\ge0$ is just the convexity inequality).

share|cite|improve this answer
Thanks, that's a very simple example. I actually had the problem stated otherwise before, and I thought that the formulation I put here was simpler, but equivalent, but I see that was not really the case. Maybe I'll put out another question with the alternative formulation. – Henrique de Oliveira Oct 12 '11 at 18:35

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

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