OctoInform Beta

Rokhlin’s multiple-mixing problem

Let (Ω,ℱ,μ)(\Omega,\mathcal F,\mu) be a probability space, and let TT be an invertible measurable transformation of Ω\Omega with measurable inverse that preserves μ\mu. The transformation TT is mixing if μ(A∩T−nB)→μ(A)μ(B)\mu(A\cap T^{-n}B)\to\mu(A)\mu(B) as |n|→∞|n|\to\infty, for all A,B∈ℱA,B\in\mathcal F. For an integer k≥2k\ge2, the transformation TT is mixing of order kk if μ(A1∩T−n1A2∩⋯∩T−(n1+⋯+nk−1)Ak)→μ(A1)⋯μ(Ak)\mu\bigl(A_1\cap T^{-n_1}A_2\cap\cdots\cap T^{-(n_1+\cdots+n_{k-1})}A_k\bigr)\to\mu(A_1)\cdots\mu(A_k) as min⁡(n1,…,nk−1)→∞\min(n_1,\ldots,n_{k-1})\to\infty through positive integers, for all A1,…,Ak∈ℱA_1,\ldots,A_k\in\mathcal F. It is mixing of all orders if it is mixing of order kk for every k≥2k\ge2. The question is whether mixing implies mixing of all orders.

Rokhlin introduced mixing of higher orders. He proved mixing of higher orders for ergodic surjective continuous endomorphisms of compact commutative groups with Haar probability. In the account of de la Rue, the question whether mixing implies mixing of order 33 is Rokhlin’s question of 1949. Ryzhikov surveys Rokhlin’s question in its formulation for all orders. Kalikow proved that mixing implies mixing of order 33 for rank-one transformations. Host obtained mixing of all orders for mixing transformations with singular spectrum. Ryzhikov obtained it for mixing automorphisms of finite rank. More recently, Kanigowski and Ravotti proved that mixing locally uniformly shearing flows are mixing of all orders.

As its main theorem, the preprint proves that every invertible mixing probability-preserving transformation is mixing of all orders. A corollary in the preprint, which the Lean library does not contain, extends the conclusion to mixing endomorphisms of arbitrary probability spaces and to mixing flows whose Koopman operators are strongly continuous on L1L^1.

The Lean library MultipleMixing proves the main theorem on an arbitrary measurable space, with no hypothesis beyond the preservation of μ\mu and the mixing of TT. Lean has checked every result of the Lean library, so the Lean library is the reference. Each result on these pages is stated as the Lean library states it and is linked to the Lean declaration behind it. Where the preprint states a result differently, the preprint’s statement stands beside the Lean library’s. The preprint is a guide only: it gives the results their order and their names.

The Lean library: MultipleMixing

MultipleMixing is a Lean library: 32 items, in seven groups. Each has a page written by OctoInform, listed below.

§1. Introduction 2 items
  1. §1 (setting): mixing and mixing of order k Mixing and mixing of order kk of an invertible transformation. §1
  2. Theorem 1.1 [Main theorem] Invertible mixing transformations are mixing of all higher orders. §1
§2. A finite-alphabet witness of zero entropy 1 item
  1. Proposition 2.1 [Zero-entropy witness] No finite-alphabet witness has a function of mean zero and zero entropy. §2
§3. Ordered arrays and conditional locality 9 items
  1. §3.1 (setting): the initial array laws ρ_n Initial array laws and their construction by weak ultralimits. §3.1
  2. Definition 3.1 [An array object] Array objects over the initial laws and their extensions. §3
  3. §3.2 (setting): the data fields 𝒫_n, 𝒪_{n,j}, 𝒢_{n,j} Data fields of an array object and their projections. §3.2
  4. Proposition 3.2 [Saturation] Every array object has a saturated extension. §3
  5. Lemma 3.3 [Local conditioning] Conditioning on local observations preserves conditional independence. §3
  6. Lemma 3.4 [Rolling an axis into the local state] Rolling an axis of a saturated object gives an extension. §3
  7. §3.4 (setting): the square The kernel and the parameter law of an array object. §3.4
  8. Proposition 3.5 [Square locality] The square of a saturated array object is conditionally local. §3
  9. Lemma 3.6 [Parameter geometry] Given a row, proper sets of columns are conditionally independent. §3
§4. The operator square 8 items
  1. Definition 4.1 [Active space, exact tuples and lines] Active spaces and exact lines of a tuple law. §4
  2. Lemma 4.2 [Exact lines] Independence of all but one coordinate makes exact lines rigid. §4
  3. Lemma 4.3 [The rectangle identities] Row and column operators almost surely form a rectangle. §4
  4. Lemma 4.4 [Active-space reduction] Rectangle operators intertwine the projections onto active spaces. §4
  5. Lemma 4.5 [Fixed squares] Fixed injective squares of random operator rectangles. §4
  6. Lemma 4.6 [Normalization of the polar rectangle] Polar rectangles of a saturated object normalize into one abelian group. §4
  7. Lemma 4.7 [Projection of tensor contractions] Active projections pass through contractions in saturated array objects. §4
  8. Proposition 4.8 [Common exact lines] Exact lines common to row and column laws. §4
§5. Measurable labels and finite blocks 5 items
  1. Lemma 5.1 [Active-space exactness] Active spaces as closed spans of exact lines. §5
  2. Proposition 5.2 [Measurable local labels] Exact lines of a saturated array object carry measurable labels. §5
  3. Lemma 5.3 [Counting lift] Parameter laws lifted to counting measures on labels. §5
  4. Lemma 5.4 [Finite label blocks] Lines fall into finite blocks under a measurable labelling. §5
  5. Proposition 5.5 [A block carrying the witness] Correlation on a saturated object forces a positive block. §5
§6. Cancellation in one time dimension 6 items
  1. §6 (setting): projective characters An abelian group represented by unimodular functions up to phases. §6
  2. Proposition 6.1 [Orbit cancellation] Centered zero-entropy observables cancel against projective characters on average. §6
  3. Definition 6.2 [Dissociated sequence] Sequences whose signed sums never vanish. §6
  4. Lemma 6.3 [Entropy cancellation] Zero-entropy processes decorrelate on average from dissociated projective characters. §6
  5. Lemma 6.4 [Orbit alternatives] Orbits of automorphisms: dissociated progressions or a unipotent power. §6
  6. Lemma 6.5 [Polynomial cancellation] Mixing cancels centered correlations twisted along a polynomial. §6
§7. Proof of the main theorem 1 item
  1. §7 (proof of Theorem 1.1): assembly Zero-entropy correlations have no positive lower bound. §7

No item matches.