The Setting
Rokhlin asked whether every mixing transformation is mixing of all orders. Let be a standard Borel space, a nonempty measurable space isomorphic to a Polish space with its Borel sets.✦ A measurable automorphism of preserving a probability measure is mixing when, for all measurable sets and , (1) For a finite index set of slots, functions on bounded by , one for each , and a tuple of times, the correlation along is . The main theorem concerns these correlations for when, for some index , the factor is a centered function of a finite-valued observable of word-entropy rate zero, notions defined next.
A finite-valued observable is a measurable map from to a finite set . For a sequence of maps from to , the word map of length sends to the word in ✦, and the word entropy of at length is the Shannon entropy , with , of the law of that word.✦ The observable has word-entropy rate zero when (2) and for a real function on the function is centered when . The statements below assemble the main theorem; the sections before them set out the arrays, array objects and Hilbert codes they are stated over.
Arrays
Fix the finite index set . An array of order over a space is a function from the index tuples to , so that the arrays of order form the product .✦ The only index tuple of length zero is the empty one, and the arrays of order zero form a space measurably isomorphic to , an array corresponding to its one entry; through this identification an array of order one is the -tuple of its entries.✦ For an axis and an index , the plane restriction fixes the axis at : (3)✦ The tuple of planes of along a subset of is the family of its planes that fix the last axis.✦ A map acts on arrays coordinatewise, the array being .✦
Array Objects and Saturation
Let be a measurable automorphism of a standard Borel space . Initial array laws of over are probability measures on , one for each order , each invariant under , that are consistent and proper: for every order , every axis , every index and every subset of other than itself, (4)✦
An array object over initial array laws of consists of a standard Borel space of states and a standard Borel space of parameters with measurable automorphisms and , measurable maps and with and , and probability measures on , its laws, with the properties below.✦ The parameter law of at order is the image , the law of the parameter array , and the own kernel of is the Markov kernel from to giving the conditional law under of an array given its parameter array. Each is invariant under and lies over , in that ; the laws are consistent and proper as in (4); they are conditionally local, in that for every order , every subset of other than and almost every parameter array of order , (5) and and are independent under the local law of , the measure read on .
An extension of an array object by an array object over the same initial laws, with states , , parameters , , base maps , , parameter maps , and laws , , is a pair of measurable maps, on states and on parameters, that intertwine the automorphisms, satisfy and , and carry laws to laws, for every .✦ The pullback along at order is the linear isometry of complex spaces (6)✦
A data kind over is a label naming one of data fields: the parameter field, and for each an outside field and an enlarged field.✦ The parameter field , the outside field and the enlarged field are the -algebras on that an array object comes with, for every and . The data -algebra of of kind at order is the -algebra on , among the parameter, outside and enlarged fields of , that names.✦ The projection onto data of of kind at order , written , is the orthogonal projection of onto the functions measurable for the data -algebra of of kind at order .✦ An array object is saturated when its projections onto data commute with pullback along every extension: for every array object over the same initial laws, every extension of by , every order and every data kind , (7)✦
Hilbert Codes
Let be a finite measure on a standard Borel space . The test set algebra of is the algebra of sets generated by the standard generating sequence of its -algebra, a fixed sequence of measurable sets generating it.✦ A Hilbert test key on is a finite list of pairs of two rational numbers and a set of the test set algebra✦; its Hilbert test function is the simple function ✦, and its Hilbert test vector is the class of that function in .✦ A fixed enumeration runs through all test keys✦, and the Hilbert test sequence of lists their test vectors in that order.✦
A real sequence is a distance code for a matrix of distances when, for all indices and , (8)✦
Let be a kernel of finite measures from a space to . A Hilbert code for is a pair of a point of , its base point, and a distance code for the distances between test vectors.✦ The decoded vector of a Hilbert code for is a vector of chosen with for every , and the encoding of a vector of is the code .✦ The approximation index of at precision is the least with ✦; the approximant of of level is the test function of the key at that index✦, and the representative of is the pointwise limit of its approximants as .✦ The circle representative sends to the point of the unit circle attached to , which is itself when ✦; the circle carries the Borel -algebra it inherits from .✦
A vector of a complex inner product space is canonical in its ray with respect to a sequence when the first not orthogonal to has .✦ A Hilbert code for is canonical when its decoded vector is canonical in its ray with respect to the test sequence of ✦, and unimodular when almost everywhere for .✦
Projective Circle Groups
The measurable circle subgroup is the group of measurable maps from to the unit circle under pointwise multiplication.✦ The constant phase subgroup of a measure on is the subgroup of the maps in that are -almost everywhere equal to a constant✦, and the projective circle group of is the quotient (9) The projective circle class of a measurable map is its image in ✦, and the projective class at of a Hilbert code for is the class in .✦ The projective representative of an element of is a map chosen in the class , the constant when ✦, and the projective character of is read as a complex function on .✦
For an abelian group , a family of complex functions on is a family of projective characters of when each is -measurable with almost everywhere, almost everywhere, for every , and for all and in there is a constant of modulus one with (10)✦
Let be a set of pairs of a point of and a Hilbert code for . The local projective group of at a point of is the subgroup of generated by the projective classes at of the codes with .✦ The local group code of an element of is a code chosen with and projective class when there is one, and the encoding of the zero vector otherwise.✦ For a map from the pairs of a point of and a Hilbert code for to a set , the local group block of an element of is the value of at the pair of and the local group code of .✦
Exact Codes and Fibre Sums
The absolute coefficient of a complex function on at a pair of a point of and a Hilbert code for is the modulus of the correlation of with the circle representative : (11)✦
Let be a kernel from a space to and let assign to each point of an -tuple of points of . The exact code mean of a pair of a point of and an -tuple of Hilbert codes for is .✦ The exact code set for , and is the set of the pairs in which every is canonical, unimodular and based at , and whose exact code mean has modulus .✦ For a space and maps and from to , the exact square set for , , , and is the set of pairs of a point of and an matrix of Hilbert codes such that, for every , the pair lies in the exact code set and, for every , the pair lies in the exact code set.✦
Let be a space. A Borel fibre enumeration of a set is a sequence of measurable sets and measurable maps such that for , every point of is for some with , and with forces .✦ The fibre sum of a real function on along is the function on whose value at a point of is the sum of over the points of above : (12) For a measurable map from to a set and a value in , the level restriction is the Borel fibre enumeration of with domains and the same maps .✦
Raw Labels
For an array object , identify parameter arrays of order zero with points of , so that is a kernel from to . The raw label space of is the product of with the Hilbert codes for .✦ The raw tuple of a state array of order one is the -tuple of its entries✦, and the tuple parameter of a parameter array of order one is likewise .✦ The raw tuple kernel of , from to , is the image of under the map sending a state array of order one to its raw tuple.✦ For a kernel from to and an index , the raw line graph at is the set of pairs of a parameter array of order one and a code for with for some tuple such that lies in the exact code set for , and the tuple parameter.✦ The raw exact square of for a kernel from to is the exact square set for , and the tuple parameter, where is the set of parameter arrays of order two and the rows and columns of are read on its planes and .✦
The Claim
Theorem 1. (a) Orbit cancellation. Let be a Markov kernel from a space to a measurable space with a countably generated -algebra; let be a Borel fibre enumeration of a set of pairs of a point of and a Hilbert code for , a measurable map from these pairs to a measurable space whose points are measurable, and a bijection of . Let and be points of and a value in , and write for the local group block of . Suppose that the block is finite; that some group automorphism of has and some group isomorphism has for every ; that some bijection from onto sends every code to its projective class at ; and that the projective characters of the elements of , indexed by written additively, form a family of projective characters of on . Let be a probability measure on a measurable space , a measure-preserving map from to , and maps of preserving with and as for all measurable sets and . Let be a measurable map from to a finite set with under , a complex function on with , a map of and a complex function on with for every and every . Then (13)✦
(b) Bounded Cesàro averages. Let be a probability measure on a space , and real numbers, and a sequence of -measurable real functions on with almost everywhere and for every , whose Cesàro averages tend to for -almost every . Then (14)✦
(c) Saturated objects. Suppose . Let be a measurable automorphism of a standard Borel space that preserves a probability measure and is mixing, a finite-valued observable on of word-entropy rate zero, and a complex function on its values with . Let be a saturated array object over initial array laws of , with state space , base map and laws , such that is the image under of the local law of ; fix , and let be measurable complex functions on with for every and . Then the correlation of order one vanishes, read on the planes of the array or on its entries. The library proves it twice. Along the raw route it assumes in addition a Markov kernel from to equal to the raw tuple kernel of at almost every parameter array of order one for the parameter law of , under which, for every and every , the entry has law and the entries other than are independent; a set of raw labels such that, for almost every parameter array of order one for the same law and every , the codes with in the raw line graph at for are exactly those with ; Borel fibre enumerations of , of the exact code set for , and the tuple parameter, and of the raw exact square of for ; and two indices of . It then gives (15)✦ and without these additional data it gives (16)✦
(d) Initial laws. With and , , , , and as in (c), suppose that is the law read on , and let be measurable complex functions on with for every and . Then (17)✦
(e) Compact systems. Let be a compact metrizable space, with its Borel sets a standard Borel space, and a homeomorphism of that preserves a Borel probability measure and is mixing, suppose , let be as in (c) and a real function on its values with centered, fix , let be continuous real functions on with for every and , and let be a sequence of tuples of natural numbers whose proper sublayouts mix: for every subset of and all measurable sets , , as . Then for every there is an index with (18)✦
In the paper, the result reads: in each typical fibre, orbit cancellation applied to every label of the positive block gives , and bounded convergence makes the integrated averages tend to , contradicting .
What Is Checked
The claims are stated over arrays and their plane restrictions, initial array laws of a mixing automorphism, saturated array objects with their laws, data fields, projections onto data and extensions, the word entropy of a finite-valued observable, Hilbert codes with their circle representatives and absolute coefficients, and Borel fibre enumerations with their fibre sums and level restrictions. About these constructions the library proves that at every order the coordinatewise state map of an extension is measure preserving from to ✦, and that the circle representative of every Hilbert code is a measurable map into the unit circle.✦