2
$\begingroup$

A Rajchman measure on the unit circle $\mathbb{T}$ is a Borel probability measure $\mu$ with $\lim_{n\to\infty}\hat{\mu}(n)=0$. Where $\hat{\mu}(n)=\mu(z^n)$ for $n\in\mathbb{Z}$ are Fourier coefficients of $\mu$.

Suppose $(X,\mathcal{B},T,\mu)$ is a measure-theoretic dynamical system consisting of a measure space $(X,\mathcal{B},\mu)$ and an invertible measure preserving map $T:X\to X$. A dynamical system $(X,\mathcal{B},T,\mu)$ is called strong mixing if $$\lim_{n\to\infty} \mu(T^{-n}A\cap B)=\mu(A)\mu(B)$$ for all $A,B\in\mathcal{B}$.

Denote the space $\{f\in L^2(X,\mu)|\int_X f\,d\mu=0\}$ by $L^2_0(X,\mu)$. The dynamical system $(X,\mathcal{B},T,\mu)$ is strong mixing iff $$\lim_{n\to\infty}\int_X f(T^nx)\overline{f(x)}\,d\mu(x)=0$$ for every unit $f\in L^2_0(X,\mu)$.

For every unit $f\in L^2_0(X,\mu)$, one can define a Borel probability measure $\mu_f$ on $\mathbb{T}$ by $$\widehat{\mu_f}(n)=\int_X f(T^nx)\overline{f(x)}\,d\mu(x).$$ So a dynamical system $(X,\mathcal{B},T,\mu)$ is strong mixing iff $\mu_f$ is a Rajchman measure for every unit $f\in L^2_0(X,\mu)$.

Through the above observation, given a strong mixing dynamical system $(X,\mathcal{B},T,\mu)$, for instance, Bernoulli shift $(\{0,1\}^\mathbb{Z},\mathcal{B},S,\mu_p)$ where $\mu_p$ is the product measure of the measure on $\{0,1\}$ giving $\{0\}$ measure $p$ and $\{1\}$ measure $1-p$, each $f\in L^2_0(X,\mu)$ gives rise to a Rajchman measure on $\mathbb{T}$.

Question: For which strong mixing dynamical system $(X,\mathcal{B},T,\mu)$, one can find unit $f\in L^2_0(X,\mu)$ such that $\mu_f$ is a Rajchman measure singular to the Lebesgue measure?

$\endgroup$
1
  • 2
    $\begingroup$ I think you're exactly asking for the strong mixing dynamical systems with singular spectral type. One comment: these must have entropy 0 as any positive entropy measure has Lebesgue spectrum. $\endgroup$ Aug 14, 2015 at 9:13

1 Answer 1

2
$\begingroup$

I think you're exactly asking for the strong mixing dynamical systems with singular spectral type. One comment: these must have entropy 0 as any positive entropy measure has Lebesgue spectrum.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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