English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

A probability measure ν on a product space X × Y is said to be bistochastic with respect to measures λ on X and μ on Y if the marginals π1(ν) and π2(μ) are exactly λ and μ. A solution is presented to a problem of Arveson about sets which are of measure zero for all such ν. If (X,F , λ) and (Y,G, μ) are probability spaces, we shall say that a probability measure σ on the product σ-algebra F⊗G is bistochastic (with respect to λ and μ), and write σ ∈ Bist(λ, μ) if σ has λ and μ as its marginal measures π1(σ) and π2(σ), that is to say, if π1(σ)(A) = σ(A × Y ) = λ(A) and π2(σ)(B) = σ(X × B) = μ(B) for all A ∈ F and B ∈ G. In the course of his study of operator algebras associated with commutative lattices of self-adjoint projections on Hilbert space, Arveson [1] investigated the following problem: Given probability spaces (X,F , λ) and (Y,G, μ), and a subset E of X × Y , when is E a null set for all σ ∈ Bist(λ, μ)? It is clear that a sufficient condition for this to be the case is that there exist F ∈ F , G ∈ G with λ(F ) = μ(G) = 0 and E ⊆ (F × Y ) ∪ (X × G); sets E of this type are said to be “marginally null”. Arveson showed that a converse to this statement is valid under certain circumstances: When X and Y are compact Hausdorff spaces and λ, μ are regular Borel probability measures, then a closed subset of X × Y which is a null set for all σ ∈ Bist(λ, μ) is necessarily marginally null [1, Theorem 1.4.2]. The same conclusion holds if (X,F) and (Y,G) are standard Borel spaces and E is the complement in X × Y of a countable union of Borel rectangles [1, Theorem 1.4.3]. We shall show that in both of these cases the result may be extended to a wider class of sets E. In fact, we shall give a more precise result, expressing in terms of the measures λ and μ the supremum of σ(E) taken over all bistochastic σ. This may be viewed as an extension of a theorem of Sudakov [7, Theorem 9]. We use fairly standard measure-theoretic terminology. A subset of a Cartesian product X × Y is said to be a rectangle if it has the form A× B with A ⊆ X and B ⊆ Y . The product σ-algebra F ⊗ G of σ-algebras F on X and G on Y is the σ-algebra generated by all rectangles A×B (A ∈ F , B ∈ G). If ν is a measure on a σ-algebra containing F ⊗ G the marginals are the measures defined on F and G by π1(ν)(A) = ν(A × Y ), π2(ν)(B) = ν(X × B). In the topological theory we are concerned only with finite positive measures defined on the Borel σ-algebra B(X) Received by the editors August 29, 1994. 1991 Mathematics Subject Classification. Primary 28A35; Secondary 28A12, 47D25. c ©1996 American Mathematical Society

On a measure-theoretic problem of Arveson