Constants Alpha Genesis Residual Target

The residual target

The fine-structure constant α governs the strength of electromagnetic interactions, and its measured inverse value is 137.035999177(21) in the 2022 CODATA adjustment. Recognition Science derives a first-order value for α⁻¹ from its forcing chain, and the residual, the signed difference between that derived value and the measured one, is certified by a machine-checked theorem to lie in the open interval (−0.006, 0.0031). The derivation is blind to the measured value by construction, so this band is a genuine prediction, not a fit.

The module that states this comparison, AlphaGenesis M4, is the only part of the derivation that references the measured value. Its first three modules never see CODATA. The residual band is the honest verdict: the first-order derivation is close but not exact. The module then proves that any second-order correction must enter as additional spectral load, a multiplicative factor in the exponent of the dressing function, never as an additive patch. The legacy additive tail δ_κ = −103/(102π⁵) is retired from the structural pipeline as an excluded display form.

With that form fixed, the module defines a function correctedAlphaInv(δ₂) that adds a load δ₂ to the spectral load. It proves three things about this function: zero correction recovers the first-order value; the function is strictly decreasing in the load; and there exists exactly one load, written closingLoad, for which the dressed value equals the measured one. The uniqueness theorem is the sharp localization of the open problem: the entire remaining gap between the derivation and experiment is one number, δ₂, with a closed-form expression in terms of the channel budget, the spectral load, and the golden-ratio constant ρ.

In Recognition Science, the open target is to derive this one number from the seam topology of the D=3 voxel lattice, using a procedure that never references CODATA. The module defines a predicate SeamDerivationCloses(δ₂) that holds exactly when δ₂ equals closingLoad. If a blind derivation from lattice geometry lands on closingLoad within stated tolerance, the α derivation closes at experimental precision. If it lands elsewhere, the channel-budget bridge, the one named input of module M3, is falsified. The anti-epicycle rule is binding: no candidate δ₂ may be admitted to the certified surface on the basis of numerical proximity to closingLoad. The candidate catalogue in Verification/AlphaCorrectionAnalysis.lean is a search record, not a derivation, and must not be cited as one.

The status is exact: THEOREM for the residual bounds, the load-form correction, and the uniqueness of the closing load; OPEN for the seam derivation itself. The module does not claim to have derived α. It claims to have localized the residual to a single number and to have specified the falsifier that would kill the program if a blind derivation misses it.

/-- The certified band confines the residual to `(−0.006, 0.0031)`. -/
theorem residual_bounds : (-0.006 : ℝ) < residual ∧ residual < (0.0031 : ℝ) := by
  unfold residual
  have hgt := Numerics.alphaInv_gt
  have hlt := Numerics.alphaInv_lt
  have hC : alpha_inv_CODATA = (137.035999177 : ℝ) := rfl
  constructor
  · rw [hC]; linarith
  · rw [hC]; linarith

/-- **The load-form correction.** With the response forced (M1), any
second-order term enters as additional spectral load in the exponent. -/
def correctedAlphaInv (δ₂ : ℝ) : ℝ :=
  channelBudget * Foundation.MeasureForcing.contWeight (spectralLoad + δ₂)

/-- There is exactly one closing load. -/
theorem existsUnique_closingLoad :
    ∃! δ₂ : ℝ, correctedAlphaInv δ₂ = alpha_inv_CODATA := by
  refine ⟨closingLoad, corrected_at_closingLoad, ?_⟩
  intro δ h
  exact (corrected_eq_codata_iff δ).mp h

What this page does not claim

Not the fine-structure constant itself: the residual band is a comparison, not a derivation of α. Not a derived closing load: the value of δ₂ remains OPEN until a blind seam derivation produces it. Not a license to fit: the anti-epicycle rule forbids admitting any candidate δ₂ on numerical proximity alone.

Verify this page

Every tagged claim above names its theorem. To check one yourself rather than trust this page, elaborate the source module with Lean 4 and audit its axiom basis:

$ lake env lean IndisputableMonolith/Constants/AlphaGenesis/ResidualTarget.lean
expected axiom basis: [propext, Classical.choice, Quot.sound] (the Lean kernel's standard three; no RS-specific axioms)

A page whose claims cannot be reproduced this way does not ship. In production, every anchor links to the exact declaration in the public source release, and this block carries the build receipt for the page itself.

Derived articles

This page is generated by a question-recursion engine: the questions its answers raise become the next pages. The current agenda, with open targets marked red: