The Ratner property, a quantitative form of divergence of nearby trajectories, is a central feature in the study of parabolic homogeneous flows. Discovered by Marina Ratner and used in her 1980th seminal works on horocycle flows, it pushed forward the disjointness theory of such systems. In this paper, exploiting a recent variation of the Ratner property, we prove new disjointness phenomena for smooth parabolic flows beyond the homogeneous world. In particular, we establish a general disjointness criterion based on the switchable Ratner property. We then apply this new criterion to study disjointness properties of smooth time changes of horocycle flows and smooth Arnol'd flows on T 2 , focusing in particular on disjointness of distinct flow rescalings. As a consequence, we answer a question by Marina Ratner on the Möbius orthogonality of time-changes of horocycle flows. In fact, we prove Möbius orthogonality for all smooth time-changes of horocycle flows and uniquely ergodic realizations of Arnol'd flows considered.
Dedicated to the memory of Marina RatnerContents 49 7.3. Slow shearing properties: proof of Lemma 7.2 51 7.4. Splitting of orbits: proof of Lemma 7.3 53 8. Disjointness of time changes of horocycle flows and Arnol'd flows (proof of Theorem 1.3) 60 9. Time changes of horocycle flows and Sarnak's conjecture (answer to M. Ratner's question) 64 Appendix A. Consequences of shearing for time changes of horocycle flows 66 Appendix B. Ergodic averages for horocycle flows 71 Appendix C. Strong MOMO and USIC properties 73 Acknowledegements 75 References 75 8 ADAM KANIGOWSKI, MARIUSZ LEMAŃCZYK, AND CORINNA ULCIGRAI flows (Theorem 1.3).Finally, in Section 9 we discuss consequences of our main results to the problem of Möbius disjointness. Three appendices follow. In the first two, we give the proofs of two technical results which are used in the proof of Theorem 1.1: in the first one, Appendix A, we prove some quantitative shearing properties for time-changes of horocycle flows; in Appendix B we state a result on ergodic averages of a special family of smooth functions which can be deduced from [6] (we thank G. Forni for the proof). Finally, in Appendix C, we present a missing link in the literature, namely, the equivalence between two different kinds of convergence on short intervals.2. Preliminaries: definitions, notation and some basic facts 2.1. Flows, joinings and disjointness. Assume that (Z, D, κ) is a probability standard Borel space. If Z is additionally a metric space (with a metric d), then we will also write (Z, D, κ, d) to emphasize the role of d. By Aut(Z, D, κ), we denote the set of all automorphisms of (Z, D, κ), i.e. R ∈ Aut(Z, D, κ) if R : Z → Z is a bi-measurable, measure-preserving κ-a.e. bijection. Each R ∈ Aut(Z, D, κ) determines a unitary operator, also denoted by R, on L 2 (Z, D, κ) given by Rf := f • R for f ∈ L 2 (Z, D, κ). Endowed with the weak operator topology of unitary operators, Aut(Z, D, κ) becomes a Polish group. We will be mainly deal with flows, i.e. with measurable, measure-pr...