The spectrum survives complexification #
For a real Hilbert space E, complexify : (E →L[ℝ] E) → (Eℂ →L[ℂ] Eℂ) is an injective
unital ⋆-algebra map (complexifyStarAlgHom). This file proves that it also reflects
invertibility, and deduces that it preserves the real spectrum on the nose:
isUnit_complexify_iff:IsUnit (complexify T) ↔ IsUnit T;spectrum_complexify:spectrum ℝ (complexify T) = spectrum ℝ T;mem_spectrum_complexify_iff: the real points of the complex spectrum.
The two halves of the first statement are asymmetric:
- forward is formal — a unital ring map carries units to units;
- backward is the content — an inverse of
complexify Tmust be shown to come from a real operator, andrealifyis that operator. The proof does not need the usual "the inverse of a conjugation-fixed operator is conjugation-fixed" argument: evaluating both unit equations atofReal xand taking real parts gives the two real identities directly.
Which real algebra structure #
spectrum ℝ on Eℂ →L[ℂ] Eℂ is taken with respect to the scoped complexOperatorRealAlgebra
(Algebra.complexToReal), the one the real continuous functional calculus of
Complexification/FunctionalCalculus.lean is registered against. A globally available second
real structure, ContinuousLinearMap.algebra, exists on the same type and is only
propositionally equal to it, so a consumer of spectrum_complexify must have this namespace's
scope open. See algebraMap_complexOperator.
Provenance #
The mathematics is moved, not restated, from TauCeti.DavisKahan.Experimental.Foundation.
RealComplexification in DavisKahan/SpectralTheory/Complexification/Spectrum.lean, which
depended on nothing in the paper library and is where this argument was first proved. That
module still carries its own copies in its own namespace; re-grounding it on these is
separate, mechanical work. A third copy, in DavisKahan/Experimental/MathAhead/
HiddenFoundations/RealSylvesterDescent.lean, was deleted with that staging file on
2026-08-27.
Complexification is a unital ring map, so it carries units to units.
Complexification reflects invertibility.
The inverse of complexify T restricts to an inverse of T: realify of it works on both
sides, because evaluating each unit equation at ofReal x and taking real parts gives exactly
the two real identities.
Invertibility is preserved and reflected by complexification.
The real points of the complex spectrum survive complexification.
The real spectrum survives complexification, as an equality of subsets of ℝ.
This is the statement that lets the real continuous functional calculus of the complexified
algebra be read as a calculus for the real operator: the two symbol algebras
C(spectrum ℝ T, ℝ) and C(spectrum ℝ (complexify T), ℝ) are the same object.