The Hilbert-space complexification of a real inner product space #
For a real inner product space X, the complexification X_ℂ = X ⊕ i X
(TauCeti.Complexification X) carries the complex inner product
⟪x + i y, u + i v⟫ = (⟪x, u⟫ + ⟪y, v⟫) + i (⟪x, v⟫ - ⟪y, u⟫),
whose norm is the L² norm ‖x + i y‖ = √(‖x‖² + ‖y‖²). This is not the Taylor norm that
TauCeti.Analysis.Normed.Module.Complexification puts on Complexification X: the two norms are
equivalent but differ in general (for orthogonal unit vectors x and y, the Taylor norm of
x + i y is 1 and its L² norm is √2). The Hilbert structure therefore lives on the type
synonym WithLp 2 (Complexification X), just as Mathlib's L² product of two inner product
spaces lives on WithLp 2 (E × F). As a real normed space, X_ℂ is isometric to the L²
product WithLp 2 (X × X), so it is complete whenever X is.
A bounded real operator T : X →L[ℝ] Y complexifies to the bounded complex operator
T_ℂ (x + i y) = T x + i T y, the WithLp lift of ContinuousLinearMap.complexify. For the
L² norm, complexification is isometric and commutes with composition and with adjoints. For a
real Hilbert space X, it is therefore an isometric ⋆-algebra embedding of the real
C⋆-algebra X →L[ℝ] X into the complex C⋆-algebra X_ℂ →L[ℂ] X_ℂ. This is the standard
device for transferring complex spectral theory, such as the continuous functional calculus of a
self-adjoint operator, to operators on real Hilbert spaces.
Main declarations #
WithLp.instComplexificationInnerProductSpace: the complex inner product spaceWithLp 2 (Complexification X), withWithLp.complexification_re_innerandWithLp.complexification_im_innercomputing the inner product from real and imaginary parts.WithLp.complexification_norm_sq_eq:‖x + i y‖² = ‖x‖² + ‖y‖².WithLp.complexificationEquivProd: the real-linear isometric equivalenceWithLp 2 (Complexification X) ≃ₗᵢ[ℝ] WithLp 2 (X × X), andWithLp.instComplexificationCompleteSpaceas a consequence.ContinuousLinearMap.complexifyL2: the complexification of a bounded real operator, withContinuousLinearMap.eq_complexifyL2_iff(it is the unique complex-linear extension),ContinuousLinearMap.norm_complexifyL2,ContinuousLinearMap.complexifyL2_compandContinuousLinearMap.adjoint_complexifyL2.ContinuousLinearMap.complexifyL2ₗᵢ: complexification as a real-linear isometry of operator spaces.ContinuousLinearMap.complexifyL2StarAlgHom: complexification of the bounded operators on a real Hilbert space, as a⋆-algebra homomorphism overℝ.ContinuousLinearMap.isClosedEmbedding_complexifyL2StarAlgHom: complexification of bounded operators is a closed embedding.ContinuousLinearMap.isSelfAdjoint_complexifyL2_iff: complexification preserves and reflects self-adjointness.
References #
- G. A. Muñoz, Y. Sarantopoulos, A. Tonge, Complexifications of real Banach spaces, polynomials and multilinear maps, Studia Math. 134 (1999), 1–33, for the norms a complexification can carry.
Acknowledgements #
The Hilbert-space complexification, the complexification of bounded operators and its
compatibility with adjoints also appear in ForTauCeti/Analysis/InnerProductSpace/Complexification/
of the AIQ DKPS formalization
(Kitware, Inc.; Apache-2.0), on the carrier WithLp 2 (E × E). This file builds the same
structure on top of Tau Ceti's algebraic Complexification.
The L² norm ‖x + i y‖ = √(‖x‖² + ‖y‖²) on the complexification, induced from the L²
product WithLp 2 (X × X) along the coordinates z ↦ (re z, im z).
The real embedding x ↦ x + i 0 is isometric for the L² norm.
As a real normed space, the L² complexification of X is the L² product X × X, via
x + i y ↦ (x, y). Real scalars act on the complexification by restricting the complex ones.
Instances For
The L² complexification of a real Banach space is complete.
Complex scaling multiplies the L² norm of the complexification by the modulus.
The complex inner product ⟪x + i y, u + i v⟫ = (⟪x, u⟫ + ⟪y, v⟫) + i (⟪x, v⟫ - ⟪y, u⟫) on
the complexification of a real inner product space, whose norm is the L² norm.
The imaginary part of the inner product on the complexification.
The complex subspace whose real and imaginary parts lie in a given real subspace, on the Hilbert-norm complexification.
Instances For
Complexifying a real subspace preserves and reflects density.
The complexification T_ℂ (x + i y) = T x + i T y of a bounded real operator, as a bounded
complex-linear operator between the L² complexifications.
Instances For
The complexification acts on the underlying vector of WithLp 2 (Complexification X) by
ContinuousLinearMap.complexify.
The complexification agrees with T on real vectors.
The complexification is the unique complex-linear extension of T.
Complexification sends the zero operator to zero.
Complexification is additive.
Complexification preserves identity maps.
Complexification preserves composition of bounded operators.
Complexification preserves the operator norm.
The adjoint of a complexification is the complexification of the adjoint.
Complexification is real-homogeneous.
Complexification of bounded operators, as a real-linear isometry.
Instances For
Complexification of the bounded operators on a real Hilbert space, as a homomorphism of
real ⋆-algebras.
Instances For
Complexification of the bounded operators on a real Hilbert space is a closed embedding.
A bounded operator on a real Hilbert space is self-adjoint exactly when its complexification is.