The Setting
Take a unital complex -algebra and a finite-dimensional real subspace of selfadjoint elements of such that every unital -homomorphism from into a nonzero unital complex -algebra sends some two elements of to noncommuting elements. How does this noncommutativity yield arbitrarily many elements in some finite tensor power whose images remain -linearly independent under commuting Hilbert-space representations?
Definitions
For a complex Hilbert space , write for the unital complex -algebra of bounded linear operators on , with identity operator . A factor von Neumann algebra on is a von Neumann algebra on such that every operator belonging to both and its commutant equals for some .
Writing for the maximal tensor product of unital complex -algebras, for any unital complex -algebra the maximal tensor power sequence is the sequence of unital complex -algebras defined recursively by (1)✦ so that is the right-nested -fold maximal tensor power of .
Given an integer , real vector spaces , a unital associative real algebra , and a family of -linear maps (), the tensor multiplication map is the -linear map satisfying (2)✦ for all (), mapping the empty tensor to the unit of when . For a real vector space and an integer , the -fold real tensor power is denoted .
The Claim
Theorem 1. Let be a unital complex -algebra, and let be a finite-dimensional real subspace of selfadjoint elements such that for every nonzero unital complex -algebra and every unital -homomorphism , there exist with . Fix an integer .
There exist an integer and selfadjoint elements such that, for every nonzero complex Hilbert space , every unital complex -algebra , and all unital -homomorphisms and satisfying for all and and such that every commuting with and is a complex scalar multiple of the identity operator on , (3)✦
There exist an integer and elements such that, for every nonzero complex Hilbert space , all factor von Neumann algebras on with for all distinct , and all unital -homomorphisms with for each , writing for the family of -linear restrictions to , (4)✦
In the paper, the result reads: for and as in Lemma 4.3 and each , there exist and selfadjoint elements occupying the first legs whose images are -linearly independent in every irreducible representation of for every nonzero unital -algebra . Instead, the theorem above takes explicitly, uses the -fold power () in (3) alongside the factor formulation (4), and replaces irreducible representations of for nonzero by commuting unital -homomorphisms of and a unital complex -algebra on a nonzero complex Hilbert space with scalar joint commutant, without requiring to be nonzero.
What Is Checked
The claims (3) and (4) are stated over a unital complex -algebra , a finite-dimensional real subspace of selfadjoint elements of with at least two elements having noncommuting images under every unital -homomorphism from into a nonzero unital complex -algebra, an integer , and, for a nonzero complex Hilbert space , both the maximal tensor powers (1) under commuting unital -homomorphisms with a unital complex -algebra into having scalar joint commutant and the tensor multiplication map (2) on real tensor powers of for unital -homomorphisms from with ranges in pairwise commuting factor von Neumann algebras on .