The Setting
An array object lifts a measurable automorphism of a space to stationary random arrays indexed by tuples from a finite set ; each entry is a point of a state space and carries a parameter in a space . At order two the parameters form a random matrix. Lemma 3.6 of the paper on Rokhlin’s multiple-mixing problem for one transformation (OpenAI, 2026) describes the geometry of that matrix when the object is saturated: what one row determines about the columns that cross it, and how the row and the column through a diagonal entry depend on each other.
Arrays
Fix a finite index set . An index tuple of order is an element of .✦ An array of order over a space is a function , and the arrays of order form the product .✦ An array of order zero has a single entry, at the empty tuple, through which is identified with as a measurable space.✦ The coordinatewise map induced by a map is .✦
For an axis and a slot , the plane restriction fixing axis at slot sends an array of order to the array of order given by (1)✦ For a subset of , the last-axis planes of along are the family .✦
Array Objects
A standard Borel space is a nonempty measurable space isomorphic to a Polish space with its Borel -algebra.✦ Let be a measurable automorphism of a standard Borel space . Initial array laws over are probability measures on , one for each , each preserved by , that are consistent and proper: for every order , axis , slot and subset , (2)✦
An array object over consists of a standard Borel state space with an automorphism , a standard Borel parameter space with an automorphism , measurable maps and with and , and probability measures on , its laws, such that each is preserved by and lies over the base, , the laws are consistent and proper as in (2), conditionally local, and locally independent: under , read on , the maps and are independent.✦ For Markov kernels from a space to spaces , with in a finite set, the product kernel is the Markov kernel from to that sends to the product measure .✦ Conditional locality of the laws of means that for every and every , under and for almost every parameter array , (3)✦ where is the conditional law under of an array given its parameter array.
An extension of by an array object over the same initial laws is a pair of measurable maps, a state map and a parameter map , that intertwine the automorphisms of the state spaces and of the parameter spaces, satisfy and , and carry laws to laws, for every .✦ The pullback along at order is the linear isometry , , of complex spaces.✦
For each order , an array object carries a data -algebra on for each of data kinds: the parameter field and, for each , the outside field and the enlarged field . The projection onto a data field is the conditional expectation onto , the orthogonal projection of onto the closed subspace of -measurable functions.✦
An array object is saturated if its projections onto data commute with pullback along every extension: for every extension of by an array object , every order , every data field and every , (4)✦
Parameter Laws
The parameter law of order of is the probability measure on , the law of the parameter array of an array drawn from .✦ A parameter array of order one is a tuple , and its plane at a slot is the coordinate , a point of . For , the parameter coordinate kernel is the kernel from to giving the conditional law under of a tuple given its coordinate at : (5)✦
A parameter array of order two is a matrix . The row and the column of a parameter array of order two are parameter arrays of order one, and they meet in the entry , the coordinate of at and of at .
The Claim
Theorem 1. Let be a saturated array object over initial array laws, with an index set of at least two elements, and let be a parameter array of order two drawn from , with rows and columns . For every and every subset of other than itself, (6)✦ and for every , (7)✦
In the paper, the result reads: in the level-two parameter law , for every and every proper subset of , with , and in particular and are conditionally independent given , with the same law .
What Is Checked
The claim is stated over a saturated array object over initial array laws, with an index set of at least two elements, over its parameter law of order two, the plane restrictions that cut a parameter matrix into rows, columns and diagonal entries, the parameter coordinate kernels and the product kernel. Beyond the claim, the library proves that every plane restriction is measurable✦.