§5.4, Corollary 1 — Norm-preserving Hahn–Banach extension
ProvedVectorSpaceOpt.hahn_banach_norm_preserving_extensionLet X be a real normed space, let M be a real linear subspace of X, and let f : M → ℝ be continuous linear. Then there exists a continuous linear functional F : X → ℝ whose restriction to M is f and whose operator norm is unchanged:
Thus extension loses neither values nor norm information. This is the norm-controlled Hahn–Banach interface used to turn functionals constructed on one-dimensional spaces or quotient directions into global dual certificates, and it is reusable independently of the later distance theorem.
import Mathlib
namespace VectorSpaceOpt
theorem hahn_banach_norm_preserving_extension
{X : Type*} [NormedAddCommGroup X] [NormedSpace ℝ X]
(M : Submodule ℝ X) (f : M →L[ℝ] ℝ) :
∃ F : X →L[ℝ] ℝ, (∀ m : M, F m = f m) ∧ ‖F‖ = ‖f‖ := by sorry
end VectorSpaceOptRead-back
What the Lean code literally says, in plain math · gpt-5
Let be any real normed vector space and let be any real linear subspace of . For every continuous real-linear functional , where has its induced norm, there exists a continuous real-linear functional such that for every (viewing also as a point of ) and in operator norm. No closedness of , completeness of , or nontriviality assumption is present, and the statement includes zero functionals and the zero and whole subspaces.
Confirmed by the mission captain (proposal self-audit).