
## 6. Unbounded multiplicative functions

1. #### Problem 6.1.

Let $(X, \mu, T)$ be a totally uniquely ergodic $0$-entropy system. We know from recent work of Frantzikinakis-Host that if $g:\N \to \C$ is a bounded multiplicative function, then for all $f \in C(X)$ with $\int_X f d \mu = 0$ and $x \in X$ we have

\begin{equation*} \mathbb{E}^{\log}_{n \in \mathbb{N}} g(n) f(T^nx) = 0. \end{equation*}

Now, if $g: \N \to \C$ is a (possibly unbounded) multiplicative function, show that for all $f \in C(X)$, $x \in X$ and fixed $c \in \C$,

\begin{equation*} \mathbb{E}^{\log}_{n \in \mathbb{N}} \mathbb{1}_{g(n)=c} f(T^nx) = \mathbb{E}^{\log}_{n \in \mathbb{N}} \mathbb{1}_{g(n)=c} \cdot \int_X f d \mu. \end{equation*} This can be done for totally ergodic nilsystems.
• #### Problem 6.2.

Can anything be done for correlations of unbounded functions? What is the behavior in almost all intervals of \begin{equation*} \sum_{x \leq n \leq x + H} g(n), \end{equation*} with $g$ an unbounded multiplicative function such that $g(p) = O(1)$, and $H = \log^A x$ for $A>0$ fixed?

Cite this as: AimPL: Sarnak's conjecture, available at http://aimpl.org/sarnakconjecture.