Compression: linearity, positivity, and the identity on an invariant subspace #
Compression to K as a ℂ-linear map B(H) → B(K).
Equations
- Orthogonalization.Blocks.compressToₗ K = { toFun := CommutingRepetition.VN.compressTo K, map_add' := ⋯, map_smul' := ⋯ }
Instances For
An operator acting as the identity on K compresses to 1.
Compression preserves positivity (T = √T * √T).
The inverse of a bijective compression #
An element of M supported in z on the right is supported in z on the left too
(z is central).
The preimage in M z = {x ∈ M | x z = x} of an operator on L (chosen by hsurj).
Equations
- Orthogonalization.Blocks.liftOp hsurj y = ⋯.choose
Instances For
Uniqueness: an element of M z compressing to y is liftOp y.
The inverse of a bijective compression, as a ℂ-linear map B(L) → B(H).
Equations
- Orthogonalization.Blocks.liftOpₗ hsurj hinj = { toFun := Orthogonalization.Blocks.liftOp hsurj, map_add' := ⋯, map_smul' := ⋯ }
Instances For
Projections of B(L) lift to projections of M.
The transported functional φ ∘ liftOp on B(L).
Equations
- Orthogonalization.Blocks.liftFunctional hsurj hinj φ = φ ∘ₗ Orthogonalization.Blocks.liftOpₗ hsurj hinj
Instances For
A functional positive on M transports to a functional positive on all of B(L).
The transport theorem #
Transport of Theorem 1.2 through a bijective compression. If z is a central
projection of M, L ≤ range z is an M-invariant subspace, and compression to L maps
{x ∈ M | x z = x} bijectively onto B(L), then Theorem 1.2 holds at z.