Let $\mu$ be a probability measure on $\mathbb R$ and the Fourier transform of $\mu$ is defined by $$ \widehat{\mu}(\xi) = \int_{\mathbb{R}} e^{-2\pi i \xi x} ~ d \mu(x). $$ Let $\mathcal{Z}(\widehat{\mu})=\{ \xi \in \mathbb R: \widehat{\mu}(\xi)=0 \}$.

Question: how about the set $\mathcal{Z}(\widehat{\mu})$? Can we show that the set $\mathcal{Z}(\widehat{\mu})$ is at most countable?

Try: when $\mu$ has compact support, we can extend $\widehat{\mu}$ to the whole complex plane by $$ \widehat{\mu}(z) = \int_{\mathbb{R}} e^{-2\pi i z x} ~ d \mu(x). $$ Next, using compact support, we can show $\widehat{\mu}(z)$ is an entire function. By the theorem in complex analysis, we can obtain that the set $\mathcal{Z}(\widehat{\mu})$ has no limit point.

I have no idea of the case when the support of $\mu$ is not compact.

  • $\begingroup$ The characteristic function of probability measure can vanish throughout some inerval. Such examples are available in Feller's book. $\endgroup$ Mar 9, 2020 at 10:20
  • $\begingroup$ @KaviRamaMurthy can you give the name of Feller’s book? $\endgroup$
    – wzstrong
    Mar 9, 2020 at 10:22
  • $\begingroup$ I have posted an answer. $\endgroup$ Mar 9, 2020 at 10:26

1 Answer 1


The function $\phi (t)=1-|t|$ for $|t| \leq 1$ and $0$ for $|t| >1$ is the Fourier transform of a probablity measure. This is proved in Feller's book: An Introduction to Probability Theory and its Applications. See the section on 'Special densities. Mixtures' in the chapter on 'Characterisitic Functions'.


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.