The Setting
Take a unital -representation on a complex Hilbert space of a maximal tensor product of unital -algebras, or a commuting pair of unital -representations of an iterated tensor power and another unital -algebra. The question is whether distinct factors of a tensor power generate commuting von Neumann algebras, and whether, when every bounded linear operator on commuting with all represented elements is a scalar multiple of the identity, the von Neumann algebras generated by individual tensor factors have trivial center.
Maximal Tensor Products and Bicommutants
For unital complex -algebras and , the algebraic tensor product is their tensor product over .✦ The maximal tensor product of two unital complex -algebras and is the unital -algebra formed by the completion of with respect to the maximal -norm (the supremum of all -cross seminorms), into which maps by the canonical tensor completion -homomorphism .✦
Given unital complex -algebras and , the left tensor inclusion is the unital -homomorphism sending each to the image of in (where is the identity of )✦, and the right tensor inclusion is the unital -homomorphism sending each to the image of in (where is the identity of ).✦
For a unital complex -algebra and an integer , the -fold right-nested maximal tensor power is the unital complex -algebra defined recursively by (1): (1)✦ With a unital complex -algebra, , and , the -th tensor-power leg inclusion is the unital -homomorphism defined recursively by (2), where is the identity map on : (2)✦
The triple maximal tensor product of three unital complex -algebras , , and is the right-nested maximal tensor product .✦ Given unital complex -algebras , , and , the first tensor inclusion is the unital -homomorphism ✦, and the second tensor inclusion is the unital -homomorphism formed from and .
For a complex Hilbert space and a unital -subalgebra of the -algebra of bounded linear operators on , the commutant of is the von Neumann algebra on defined by (3): (3)✦ The bicommutant of a unital -subalgebra of the algebra of bounded linear operators on a complex Hilbert space is the von Neumann algebra on .✦ A factor von Neumann algebra on a complex Hilbert space is a von Neumann algebra on satisfying (4), where denotes the identity operator on : (4)✦
The Claim
Theorem 1. Let , , , and be unital complex -algebras, and let be a complex Hilbert space. If is a unital -representation satisfying , then the bicommutants of the left and right tensor legs are factor von Neumann algebras, (5) and (6): (5)✦ (6)✦
If is an integer and and are unital -representations such that for all and and , then for every the -th leg bicommutant is a factor von Neumann algebra (7): (7)✦
For any integer , any unital -representation , and any with , the -th and -th leg bicommutants commute as in (8): (8)✦
For a unital -representation satisfying , the bicommutants of the first and second tensor legs are factor von Neumann algebras, (9) and (10): (9)✦ (10)✦
In the paper, the result reads: for nonzero unital complex -algebras and an irreducible representation of , the weak-operator closures of any partition into nonempty blocks of tensor legs are commuting von Neumann factors. For arbitrary unital complex -algebras, the claims (5), (6), (7), (8), (9), and (10) use bicommutants instead of weak-operator closures, assume the scalar-commutant hypothesis directly in the factor claims rather than irreducibility, and treat binary legs, individual tensor-power legs, and the first two triple legs.
What Is Checked
The claims (5), (6), (7), (8), (9), and (10) are stated over unital complex -algebras , , , and , the maximal tensor product formed as the completion of the algebraic tensor product with its canonical tensor completion -homomorphism, the left and right tensor inclusions and , the right-nested maximal tensor power () with its leg inclusions (), the triple maximal tensor product with its first and second tensor inclusions and , unital -representations on a complex Hilbert space , the commutant and bicommutant von Neumann algebras on , the factor condition (4), and the scalar-commutant hypotheses in (5), (6), (7), (9), and (10) (jointly with a commuting representation in (7)), together with distinct leg indices in (8).