Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Chapter QgFullEliminated

Definition
ChapterQgFullEliminated

by leonardopedro · Nov 4, 2026 · Mathlib 0df444a (Lean v4.33.1)

definitionsformalizationtimepiece

Formal definitions for the timepiece Lean 4 formalization (source chapter BookProof/ChapterQgFullEliminated.lean): generated def bundle for ChapterQgFullEliminated. See BookProof/ChapterQgFullEliminated.lean for full context.

Definition code
import Definitions.Def_ChapterQgFourierElimination
import Theorems.Thm_BookProof_QgFourierElim_formValue_torsion_elimConfig
import Theorems.Thm_BookProof_QgFourierElim_formValue_dGauge_elimConfig
import Theorems.Thm_BookProof_QgFourierElim_formValue_gauge3d_elimConfig
import Theorems.Thm_BookProof_QgFourierElim_eq_elimConfig_of_gauge_fixed
import Theorems.Thm_BookProof_QgOuterFockCoreFL_CoreData_esa_on_core
import Definitions.Def_ChapterSirkTrotterKato
import Definitions.Def_ChapterDirectSumEsa
import Definitions.Def_ChapterEsaClosureCore
import Definitions.Def_ChapterFarisLavine
import Definitions.Def_ChapterQgContinuumModeInstance
import Definitions.Def_ChapterQgOuterFockCoreFL
import Definitions.Def_ChapterQgVielbeinModeInstance
import Definitions.Def_ChapterA4
import Definitions.Def_ChapterScalaronCoreEsa
import Definitions.Def_ChapterScalaronFiberFL
import Definitions.Def_ChapterScalaronOuterFockFL
import Definitions.Def_ChapterStoneBridge
import Mathlib
import Definitions.Def_ChapterQgFourierElimForms
import Theorems.Thm_BookProof_DirectSumEsa_dsCore_dense
import Theorems.Thm_BookProof_QgOuterFockCoreFL_commForm_congr
import Theorems.Thm_BookProof_QgOuterFockCoreFL_quadForm_congr
import Theorems.Thm_BookProof_ScalaronEsa_ccDomain_dense
import Theorems.Thm_BookProof_StoneBridge_exists_stone_flow_of_esa


/-!
# The full quantum-gravity Hamiltonian rebuilt on the eliminated components alone

Second half of item 8 of the quantum-gravity plan items of `CONSOLIDATED_PLAN.md`.  The model of
`BookProof/ChapterQgVielbeinScalaronGaugeFL.lean` carries `36` components per momentum — the nine
vielbein components `e_ν^i(k)` *and* the twenty-seven independent derivative variables
`D_{μν}^i(k)` — and imposes the derivative-gauge forms `G_{μν}^i = D_{μν}^i − i k_μ e_ν^i` as part
of the quadratic form.  `BookProof/ChapterQgFourierElimination.lean` §4 shows that the derivative
components carry no independent content: the constraint surface is exactly the image of the
elimination `elimConfig`.

Here the whole Hamiltonian — vielbein self-interaction (torsion), derivative gauge fixing, 3D
transverse gauge fixing, the scalaron in position representation with the full exponential
Einstein-frame wall, and the coupling to the trace of the vielbein — is **rebuilt on the nine
eliminated components alone**, and its essential self-adjointness is proved again, by the same
Faris–Lavine route, on the smaller outer Fock space `Sec EGMode`.

What is proved, in order:

* `elimCoef` — the coefficient vector, *in the nine vielbein components only*, of each of the `57`
  linear forms of the full model after the elimination `D_{μν}^i ↦ i k_μ e_ν^i`;
  `eFormValue_eq_formValue_elimConfig` is the defining property (the eliminated form is the full
  form evaluated on the eliminated configuration), `elimCoef_dGauge` records that the twenty-seven
  derivative-gauge forms become identically zero, and `eFormValue_torsion` that the torsion form
  becomes the exact Fourier torsion `i(k_μ e_ν^i − k_ν e_μ^i)`.
* `eGram` — the vielbein self-interaction of the rebuilt model, the Gram matrix of the eliminated
  family; `eGram_eq_torsion_add_gauge3d` exhibits it as torsion plus transverse gauge fixing (the
  derivative-gauge forms contributing nothing), and `eGram_quadForm` /
  `quadForm_eq_of_gauge_fixed` prove that **the rebuilt quadratic form is the full one restricted
  to the gauge surface**, in the nine surviving coordinates.
* `eCoupling` — the scalaron–vielbein coupling, unchanged (`eTrace_eq_gTrace`: the trace the
  scalaron couples to never involved the derivative components).
* `qgElimFullModes` — the mode data of the rebuilt model: all five Faris–Lavine bounds with
  `κ = 513 + |g|` and band size `9` (instead of `36`).
* `qgElimFull_esa_farisLavine`, `qgElimFull_esa_core_fl`, `starobinsky_qgElimFull_esa`,
  `starobinsky_qgElimFull_esa_core` — essential self-adjointness of the rebuilt Hamiltonian, and
  `qgElimFull_stone_flow` / `starobinsky_qgElimFull_stone_flow` the unitary time evolution it
  generates.  The extended derivative components never enter the operator or its domain, and no
  restriction-to-a-subset argument is used.
* `qgElimFull_momentum_conserving`, `qgElimFull_number_conserving` — the rebuilt Hamiltonian is
  still block diagonal in the momentum and conserves the particle number.
* Non-vacuity: `elimCoef_torsion_ne_zero`, `elimCoef_gauge3d_ne_zero`, `eGram_diag_ne_zero`,
  `eCoupling_ne_zero`, `infinite_egmode`.

No spectral information, no mass gap and no continuum limit is claimed.

Everything is `sorry`-free and `axiom`-free.
-/

namespace BookProof.QgFullEliminated

open BookProof.ScalaronFiberFL BookProof.ScalaronOuterFockFL
open BookProof.QgVielbeinModeInstance BookProof.QgContinuumModeInstance
open BookProof.FarisLavine BookProof.QgOuterFockCoreFL
open BookProof.StoneBridge BookProof.EsaClosure BookProof.ChapterStoneResolvent
open BookProof.DirectSumEsa BookProof.ScalaronEsa
open BookProof.QgFourierElim

noncomputable section

/-! ## 1. The eliminated components and their modes -/

/-- The surviving field-space components at one momentum: the nine vielbein components `e_ν^i`.
The twenty-seven derivative components of the full model are gone — not gauge fixed, *absent*. -/
abbrev EComp := Fin 3 × Fin 3

/-- A mode of the rebuilt model: a momentum in `ℤ³` together with one of the nine vielbein
components.  There are still infinitely many momenta — no lattice, no truncation. -/
abbrev EGMode := Mom × EComp

/-- The energy of a mode: the same exact free symbol `σ_k = 1 + |k|²` as in the full model. -/
def eSig (x : EGMode) : ℝ := 1 + momSq x.1

theorem one_le_eSig (x : EGMode) : 1 ≤ eSig x := by
  have := momSq_nonneg x.1
  simp only [eSig]
  linarith

theorem eSig_nonneg (x : EGMode) : 0 ≤ eSig x := le_trans zero_le_one (one_le_eSig x)

/-- The band of a mode: the nine components at the same momentum. -/
def eNbr (x : EGMode) : Finset EGMode :=
  ({x.1} : Finset Mom) ×ˢ (Finset.univ : Finset EComp)

theorem mem_eNbr {x y : EGMode} : y ∈ eNbr x ↔ y.1 = x.1 := by
  constructor
  · intro h
    simp only [eNbr, Finset.mem_product, Finset.mem_singleton] at h
    exact h.1
  · intro h
    simp only [eNbr, Finset.mem_product, Finset.mem_singleton, Finset.mem_univ, and_true]
    exact h

theorem card_eNbr (x : EGMode) : ((eNbr x).card : ℝ) = 9 := by
  simp [eNbr]

/-! ## 2. The linear forms of the full model, after the elimination -/

/-- The indicator of an eliminated component. -/
def eind (c d : EComp) : ℂ := if c = d then 1 else 0

theorem norm_eind_le (c d : EComp) : ‖eind c d‖ ≤ 1 := by
  simp only [eind]; split <;> simp

theorem sum_eind_mul (d : EComp) (z : EComp → ℂ) : ∑ c : EComp, eind c d * z c = z d := by
  classical
  rw [Finset.sum_eq_single d]
  · simp [eind]
  · intro c _ hc
    simp [eind, hc]
  · intro h
    exact absurd (Finset.mem_univ d) h

/-- **The coefficient vectors of the three families of linear forms after the elimination.**  Each
form of the full model is rewritten, by `D_{μν}^i ↦ i k_μ e_ν^i`, as a linear form in the nine
vielbein components alone:

* the torsion form becomes the exact Fourier torsion `i(k_μ e_ν^i − k_ν e_μ^i)`;
* the derivative-gauge form becomes identically zero — the constraint is solved, not imposed;
* the 3D transverse form is untouched. -/
def elimCoef (k : Mom) : FormIdx → EComp → ℂ
  | Sum.inl (mu, nu, i), c =>
      Complex.I * ((k mu : ℤ) : ℂ) * eind c (nu, i)
        - Complex.I * ((k nu : ℤ) : ℂ) * eind c (mu, i)
  | Sum.inr (Sum.inl _), _ => 0
  | Sum.inr (Sum.inr i), c => ∑ mu : Fin 3, Complex.I * ((k mu : ℤ) : ℂ) * eind c (mu, i)

@[simp] theorem elimCoef_torsion (k : Mom) (mu nu i : Fin 3) (c : EComp) :
    elimCoef k (torsionF mu nu i) c
      = Complex.I * ((k mu : ℤ) : ℂ) * eind c (nu, i)
          - Complex.I * ((k nu : ℤ) : ℂ) * eind c (mu, i) := rfl

/-- **The derivative-gauge forms disappear**: after the elimination they are the zero form, which
is the formal statement that the twenty-seven constraints are solved by construction. -/
@[simp] theorem elimCoef_dGauge (k : Mom) (mu nu i : Fin 3) (c : EComp) :
    elimCoef k (dGaugeF mu nu i) c = 0 := rfl

@[simp] theorem elimCoef_gauge3d (k : Mom) (i : Fin 3) (c : EComp) :
    elimCoef k (gauge3dF i) c = ∑ mu : Fin 3, Complex.I * ((k mu : ℤ) : ℂ) * eind c (mu, i) :=
  rfl

/-- The value of the eliminated form `F` at momentum `k` on a configuration of the nine vielbein
components. -/
def eFormValue (k : Mom) (F : FormIdx) (z : EComp → ℂ) : ℂ := ∑ c : EComp, elimCoef k F c * z c

/-- **The eliminated torsion form is the exact Fourier torsion.** -/
theorem eFormValue_torsion (k : Mom) (mu nu i : Fin 3) (z : EComp → ℂ) :
    eFormValue k (torsionF mu nu i) z
      = Complex.I * ((k mu : ℤ) : ℂ) * z (nu, i) - Complex.I * ((k nu : ℤ) : ℂ) * z (mu, i) := by
  have h : ∀ c : EComp, elimCoef k (torsionF mu nu i) c * z c
      = (Complex.I * ((k mu : ℤ) : ℂ)) * (eind c (nu, i) * z c)
        - (Complex.I * ((k nu : ℤ) : ℂ)) * (eind c (mu, i) * z c) := by
    intro c
    rw [elimCoef_torsion]
    ring
  rw [eFormValue, Finset.sum_congr rfl fun c _ => h c, Finset.sum_sub_distrib,
    ← Finset.mul_sum, ← Finset.mul_sum, sum_eind_mul, sum_eind_mul]

/-- **The eliminated derivative-gauge form vanishes identically.** -/
theorem eFormValue_dGauge (k : Mom) (mu nu i : Fin 3) (z : EComp → ℂ) :
    eFormValue k (dGaugeF mu nu i) z = 0 := by
  simp [eFormValue]

/-- The 3D transverse form is unchanged by the elimination. -/
theorem eFormValue_gauge3d (k : Mom) (i : Fin 3) (z : EComp → ℂ) :
    eFormValue k (gauge3dF i) z = ∑ mu : Fin 3, Complex.I * ((k mu : ℤ) : ℂ) * z (mu, i) := by
  have h : ∀ c : EComp, elimCoef k (gauge3dF i) c * z c
      = ∑ mu : Fin 3, (Complex.I * ((k mu : ℤ) : ℂ)) * (eind c (mu, i) * z c) := by
    intro c
    rw [elimCoef_gauge3d, Finset.sum_mul]
    exact Finset.sum_congr rfl fun mu _ => by ring
  rw [eFormValue, Finset.sum_congr rfl fun c _ => h c, Finset.sum_comm]
  refine Finset.sum_congr rfl fun mu _ => ?_
  rw [← Finset.mul_sum, sum_eind_mul]

/-- **The defining property of the elimination**: the eliminated form is the full form of
`ChapterQgVielbeinScalaronGaugeFL` evaluated on the eliminated configuration.  Nothing is added
and nothing is dropped; the rebuilt model is the full model written in the nine surviving
coordinates. -/
theorem eFormValue_eq_formValue_elimConfig (k : Mom) (F : FormIdx) (z : EComp → ℂ) :
    formValue k F (elimConfig k z) = eFormValue k F z := by
  rcases F with ⟨mu, nu, i⟩ | (⟨mu, nu, i⟩ | i)
  · have hF : (Sum.inl (mu, nu, i) : FormIdx) = torsionF mu nu i := rfl
    rw [hF, formValue_torsion_elimConfig, eFormValue_torsion]
    ring
  · have hF : (Sum.inr (Sum.inl (mu, nu, i)) : FormIdx) = dGaugeF mu nu i := rfl
    rw [hF, formValue_dGauge_elimConfig, eFormValue_dGauge]
  · have hF : (Sum.inr (Sum.inr i) : FormIdx) = gauge3dF i := rfl
    rw [hF, formValue_gauge3d_elimConfig, eFormValue_gauge3d]

/-! ### The uniform bounds on the eliminated coefficients -/

theorem norm_I_mul_k_mul_eind (k : Mom) (mu : Fin 3) (c d : EComp) :
    ‖Complex.I * ((k mu : ℤ) : ℂ) * eind c d‖ ≤ mAbs k := by
  have h1 : ‖Complex.I * ((k mu : ℤ) : ℂ) * eind c d‖
      = |((k mu : ℤ) : ℝ)| * ‖eind c d‖ := by
    rw [norm_mul, norm_mul, Complex.norm_I, one_mul]
    congr 1
    simp
  rw [h1]
  have h2 := norm_eind_le c d
  have h3 := abs_k_le_mAbs k mu
  have h4 : (0 : ℝ) ≤ |((k mu : ℤ) : ℝ)| := abs_nonneg _
  nlinarith [norm_nonneg (eind c d)]

/-- Every eliminated coefficient is bounded by `3|k|`: the elimination replaces the bounded
derivative coordinates by momentum weights, so all three families are now `O(|k|)`. -/
theorem norm_elimCoef_le (k : Mom) (F : FormIdx) (c : EComp) : ‖elimCoef k F c‖ ≤ 3 * mAbs k := by
  have hM := mAbs_nonneg k
  rcases F with ⟨mu, nu, i⟩ | (⟨mu, nu, i⟩ | i)
  · have hF : elimCoef k (Sum.inl (mu, nu, i)) c = elimCoef k (torsionF mu nu i) c := rfl
    have h : ‖elimCoef k (torsionF mu nu i) c‖ ≤ mAbs k + mAbs k :=
      le_trans (norm_sub_le _ _)
        (add_le_add (norm_I_mul_k_mul_eind k mu c (nu, i))
          (norm_I_mul_k_mul_eind k nu c (mu, i)))
    rw [hF]
    linarith
  · have hF : elimCoef k (Sum.inr (Sum.inl (mu, nu, i))) c = elimCoef k (dGaugeF mu nu i) c := rfl
    rw [hF, elimCoef_dGauge, norm_zero]
    linarith
  · have hF : elimCoef k (Sum.inr (Sum.inr i)) c = elimCoef k (gauge3dF i) c := rfl
    have h : ‖elimCoef k (gauge3dF i) c‖ ≤ 3 * mAbs k := by
      rw [elimCoef_gauge3d]
      refine le_trans (norm_sum_le _ _) ?_
      calc ∑ mu : Fin 3, ‖Complex.I * ((k mu : ℤ) : ℂ) * eind c (mu, i)‖
          ≤ ∑ _mu : Fin 3, mAbs k :=
            Finset.sum_le_sum fun mu _ => norm_I_mul_k_mul_eind k mu c (mu, i)
        _ = 3 * mAbs k := by
            rw [Finset.sum_const, Finset.card_univ, Fintype.card_fin, nsmul_eq_mul]
            norm_num
    rw [hF]
    linarith

theorem norm_elimCoef_mul_le (k : Mom) (F : FormIdx) (c d : EComp) :
    ‖(starRingEnd ℂ) (elimCoef k F c) * elimCoef k F d‖ ≤ 9 * (1 + momSq k) := by
  rw [norm_mul, RCLike.norm_conj]
  have h1 := norm_elimCoef_le k F c
  have h2 := norm_elimCoef_le k F d
  have hM := mAbs_nonneg k
  have hsq := mAbs_sq k
  nlinarith [norm_nonneg (elimCoef k F c), norm_nonneg (elimCoef k F d), sq_nonneg (mAbs k)]

/-! ## 3. The vielbein self-interaction of the rebuilt model -/

/-- **The complete quadratic form of the full model, rebuilt on the nine eliminated components**:
the Gram matrix of the eliminated torsion forms, the (vanishing) eliminated derivative-gauge forms
and the 3D transverse forms.  It is diagonal in the momentum and a `9 × 9` matrix in the
components at each momentum. -/
def eGram (x y : EGMode) : ℂ :=
  if x.1 = y.1 then
    ∑ F : FormIdx, (starRingEnd ℂ) (elimCoef x.1 F x.2) * elimCoef x.1 F y.2
  else 0

theorem eGram_herm (x y : EGMode) : eGram y x = (starRingEnd ℂ) (eGram x y) := by
  by_cases h : x.1 = y.1
  · have hy : y.1 = x.1 := h.symm
    simp only [eGram]
    rw [if_pos hy, if_pos h, hy, map_sum]
    refine Finset.sum_congr rfl fun F _ => ?_
    simp only [map_mul, RingHomCompTriple.comp_apply, RingHom.id_apply]
    ring
  · simp only [eGram, if_neg h, if_neg (Ne.symm h), map_zero]

theorem eGram_off {x y : EGMode} (h : y ∉ eNbr x) : eGram x y = 0 := by
  have hne : ¬ x.1 = y.1 := fun hh => h (mem_eNbr.mpr hh.symm)
  simp only [eGram, if_neg hne]

theorem norm_eGram_le (x y : EGMode) : ‖eGram x y‖ ≤ 513 * eSig x := by
  by_cases h : x.1 = y.1
  · simp only [eGram, if_pos h]
    refine le_trans (norm_sum_le _ _) ?_
    calc ∑ F : FormIdx, ‖(starRingEnd ℂ) (elimCoef x.1 F x.2) * elimCoef x.1 F y.2‖
        ≤ ∑ _F : FormIdx, 9 * (1 + momSq x.1) :=
          Finset.sum_le_sum fun F _ => norm_elimCoef_mul_le x.1 F x.2 y.2
      _ = 57 * (9 * (1 + momSq x.1)) := by
          rw [Finset.sum_const, Finset.card_univ, card_formIdx, nsmul_eq_mul]
          norm_num
      _ = 513 * eSig x := by simp only [eSig]; ring
  · have hnn := eSig_nonneg x
    simp only [eGram, if_neg h, norm_zero]
    linarith

/-- **The rebuilt self-interaction is torsion plus transverse gauge fixing**: the twenty-seven
derivative-gauge forms contribute nothing, because the elimination solves them identically. -/
theorem eGram_eq_torsion_add_gauge3d (k : Mom) (c d : EComp) :
    eGram (k, c) (k, d)
      = (∑ mu : Fin 3, ∑ nu : Fin 3, ∑ i : Fin 3,
            (starRingEnd ℂ) (elimCoef k (torsionF mu nu i) c) * elimCoef k (torsionF mu nu i) d)
        + ∑ i : Fin 3,
            (starRingEnd ℂ) (elimCoef k (gauge3dF i) c) * elimCoef k (gauge3dF i) d := by
  have hd : ∀ p : Fin 3 × Fin 3 × Fin 3,
      (starRingEnd ℂ) (elimCoef k (Sum.inr (Sum.inl p)) c) * elimCoef k (Sum.inr (Sum.inl p)) d
        = 0 := by
    intro p
    obtain ⟨mu, nu, i⟩ := p
    have hF : (Sum.inr (Sum.inl (mu, nu, i)) : FormIdx) = dGaugeF mu nu i := rfl
    simp only [hF, elimCoef_dGauge, map_zero, zero_mul]
  have htor : (∑ p : Fin 3 × Fin 3 × Fin 3,
        (starRingEnd ℂ) (elimCoef k (Sum.inl p) c) * elimCoef k (Sum.inl p) d)
      = ∑ mu : Fin 3, ∑ nu : Fin 3, ∑ i : Fin 3,
          (starRingEnd ℂ) (elimCoef k (torsionF mu nu i) c)
            * elimCoef k (torsionF mu nu i) d := by
    rw [Fintype.sum_prod_type]
    refine Finset.sum_congr rfl fun mu _ => ?_
    rw [Fintype.sum_prod_type]
    rfl
  have hgauge : (∑ p : (Fin 3 × Fin 3 × Fin 3) ⊕ Fin 3,
        (starRingEnd ℂ) (elimCoef k (Sum.inr p) c) * elimCoef k (Sum.inr p) d)
      = ∑ i : Fin 3, (starRingEnd ℂ) (elimCoef k (gauge3dF i) c) * elimCoef k (gauge3dF i) d := by
    rw [Fintype.sum_sum_type, Finset.sum_congr rfl fun p _ => hd p, Finset.sum_const_zero,
      zero_add]
    rfl
  have hsplit : eGram (k, c) (k, d)
      = (∑ p : Fin 3 × Fin 3 × Fin 3,
            (starRingEnd ℂ) (elimCoef k (Sum.inl p) c) * elimCoef k (Sum.inl p) d)
        + ∑ p : (Fin 3 × Fin 3 × Fin 3) ⊕ Fin 3,
            (starRingEnd ℂ) (elimCoef k (Sum.inr p) c) * elimCoef k (Sum.inr p) d := by
    have hval : eGram ((k, c) : EGMode) (k, d)
        = ∑ F : FormIdx, (starRingEnd ℂ) (elimCoef k F c) * elimCoef k F d := by
      simp [eGram]
    rw [hval, Fintype.sum_sum_type]
  rw [hsplit, htor, hgauge]

/-! ### The rebuilt quadratic form is the full one on the gauge surface -/

/-- A Gram matrix built from a family of linear forms evaluates, as a quadratic form, to the sum
of the squared moduli of the values of the forms. -/
theorem gram_quadForm {X : Type*} [Fintype X] (C : FormIdx → X → ℂ) (z : X → ℂ) :
    ∑ c : X, ∑ d : X,
        (starRingEnd ℂ) (z c) * ((∑ F : FormIdx, (starRingEnd ℂ) (C F c) * C F d) * z d)
      = ∑ F : FormIdx, (starRingEnd ℂ) (∑ c : X, C F c * z c) * (∑ d : X, C F d * z d) := by
  have hL : ∀ c : X, ∀ d : X,
      (starRingEnd ℂ) (z c) * ((∑ F : FormIdx, (starRingEnd ℂ) (C F c) * C F d) * z d)
        = ∑ F : FormIdx,
            (starRingEnd ℂ) (z c) * ((starRingEnd ℂ) (C F c) * C F d) * z d := by
    intro c d
    rw [Finset.sum_mul, Finset.mul_sum]
    exact Finset.sum_congr rfl fun F _ => by ring
  have hR : ∀ F : FormIdx,
      (starRingEnd ℂ) (∑ c : X, C F c * z c) * (∑ d : X, C F d * z d)
        = ∑ c : X, ∑ d : X, (starRingEnd ℂ) (z c) * ((starRingEnd ℂ) (C F c) * C F d) * z d := by
    intro F
    rw [map_sum, Finset.sum_mul_sum]
    refine Finset.sum_congr rfl fun c _ => Finset.sum_congr rfl fun d _ => ?_
    rw [map_mul]
    ring
  calc ∑ c : X, ∑ d : X,
        (starRingEnd ℂ) (z c) * ((∑ F : FormIdx, (starRingEnd ℂ) (C F c) * C F d) * z d)
      = ∑ c : X, ∑ d : X, ∑ F : FormIdx,
          (starRingEnd ℂ) (z c) * ((starRingEnd ℂ) (C F c) * C F d) * z d :=
        Finset.sum_congr rfl fun c _ => Finset.sum_congr rfl fun d _ => hL c d
    _ = ∑ c : X, ∑ F : FormIdx, ∑ d : X,
          (starRingEnd ℂ) (z c) * ((starRingEnd ℂ) (C F c) * C F d) * z d :=
        Finset.sum_congr rfl fun c _ => Finset.sum_comm
    _ = ∑ F : FormIdx, ∑ c : X, ∑ d : X,
          (starRingEnd ℂ) (z c) * ((starRingEnd ℂ) (C F c) * C F d) * z d :=
        Finset.sum_comm
    _ = ∑ F : FormIdx, (starRingEnd ℂ) (∑ c : X, C F c * z c) * (∑ d : X, C F d * z d) :=
        Finset.sum_congr rfl fun F _ => (hR F).symm

/-- The rebuilt quadratic form at one momentum is the sum of the squared moduli of the eliminated
forms. -/
theorem eGram_quadForm (k : Mom) (z : EComp → ℂ) :
    ∑ c : EComp, ∑ d : EComp, (starRingEnd ℂ) (z c) * (eGram (k, c) (k, d) * z d)
      = ∑ F : FormIdx, (starRingEnd ℂ) (eFormValue k F z) * eFormValue k F z := by
  have h : ∀ c d : EComp, eGram (k, c) (k, d)
      = ∑ F : FormIdx, (starRingEnd ℂ) (elimCoef k F c) * elimCoef k F d := by
    intro c d
    simp [eGram]
  simp only [h, eFormValue]
  exact gram_quadForm (X := EComp) (elimCoef k) z

/-- The full quadratic form at one momentum is the sum of the squared moduli of the full forms. -/
theorem gGram_quadForm (k : Mom) (w : Comp → ℂ) :
    ∑ c : Comp, ∑ d : Comp, (starRingEnd ℂ) (w c) * (gGram (k, c) (k, d) * w d)
      = ∑ F : FormIdx, (starRingEnd ℂ) (formValue k F w) * formValue k F w := by
  have h : ∀ c d : Comp, gGram ((k, c) : GMode) (k, d)
      = ∑ F : FormIdx, (starRingEnd ℂ) (gCoef k F c) * gCoef k F d := by
    intro c d
    simp [gGram]
  simp only [h, formValue]
  exact gram_quadForm (X := Comp) (gCoef k) w

/-- **The rebuilt Hamiltonian carries the full quadratic form.**  On the gauge surface — where the
twenty-seven derivative-gauge constraints of the full model hold — the `36`-component quadratic
form of `ChapterQgVielbeinScalaronGaugeFL` and the `9`-component quadratic form rebuilt here take
the same value, the latter evaluated on the vielbein part of the configuration.  Together with
`eq_elimConfig_of_gauge_fixed` (every gauge-fixed configuration is an eliminated one) this is the
sense in which nothing is lost by discarding the derivative components. -/
theorem quadForm_eq_of_gauge_fixed (k : Mom) (w : Comp → ℂ)
    (hgauge : ∀ mu nu i, formValue k (dGaugeF mu nu i) w = 0) :
    ∑ c : Comp, ∑ d : Comp, (starRingEnd ℂ) (w c) * (gGram (k, c) (k, d) * w d)
      = ∑ c : EComp, ∑ d : EComp,
          (starRingEnd ℂ) (w (eIdx c.1 c.2)) *
            (eGram (k, c) (k, d) * w (eIdx d.1 d.2)) := by
  set z : EComp → ℂ := fun p => w (eIdx p.1 p.2) with hz
  have hw : w = elimConfig k z := eq_elimConfig_of_gauge_fixed k w hgauge
  rw [gGram_quadForm k w, eGram_quadForm k z]
  refine Finset.sum_congr rfl fun F _ => ?_
  rw [hw, eFormValue_eq_formValue_elimConfig]

/-! ## 4. The scalaron–vielbein coupling on the eliminated components -/

/-- The trace part of the vielbein: the combination the scalaron couples to.  It only ever
involved the nine vielbein components, so the elimination leaves it unchanged. -/
def eTrace (x : EGMode) : ℝ := if x.2.1 = x.2.2 then 1 else 0

/-- The trace part of the vielbein: the combination the scalaron couples to.  The derivative
variables do not carry it. -/
def gTrace (x : GMode) : ℝ :=
  match x.2 with
  | Sum.inl (nu, i) => if nu = i then 1 else 0
  | Sum.inr _ => 0

/-- The trace of the rebuilt model is the trace of the full model, read on the vielbein
components. -/
theorem eTrace_eq_gTrace (x : EGMode) : eTrace x = gTrace (x.1, eIdx x.2.1 x.2.2) := rfl

theorem abs_eTrace_le (x : EGMode) : |eTrace x| ≤ 1 := by
  simp only [eTrace]
  split <;> simp

/-- **The scalaron–vielbein coupling** at coupling constant `g`, on the eliminated components. -/
def eCoupling (g : ℝ) (x y : EGMode) : ℂ :=
  if x.1 = y.1 then ((g * eTrace x * eTrace y : ℝ) : ℂ) else 0

theorem eCoupling_herm (g : ℝ) (x y : EGMode) :
    eCoupling g y x = (starRingEnd ℂ) (eCoupling g x y) := by
  by_cases h : x.1 = y.1
  · have hy : y.1 = x.1 := h.symm
    simp only [eCoupling]
    rw [if_pos hy, if_pos h, Complex.conj_ofReal]
    norm_cast
    ring
  · simp only [eCoupling, if_neg h, if_neg (Ne.symm h), map_zero]

theorem eCoupling_off (g : ℝ) {x y : EGMode} (h : y ∉ eNbr x) : eCoupling g x y = 0 := by
  have hne : ¬ x.1 = y.1 := fun hh => h (mem_eNbr.mpr hh.symm)
  simp only [eCoupling, if_neg hne]

theorem norm_eCoupling_le (g : ℝ) (x y : EGMode) : ‖eCoupling g x y‖ ≤ |g| := by
  by_cases h : x.1 = y.1
  · simp only [eCoupling, if_pos h, Complex.norm_real, Real.norm_eq_abs, abs_mul]
    have h1 := abs_eTrace_le x
    have h2 := abs_eTrace_le y
    have hg := abs_nonneg g
    calc |g| * |eTrace x| * |eTrace y| ≤ |g| * 1 * 1 := by gcongr
      _ = |g| := by ring
  · simp only [eCoupling, if_neg h, norm_zero]
    exact abs_nonneg g

/-! ## 5. The mode data of the rebuilt model -/

/-- **The mode data of the full quantum-gravity model rebuilt on the eliminated components**:
exact Fourier modes of the vielbein, the torsion self-interaction in its eliminated (exact
Fourier) form, the — now identically satisfied — gauge fixing of the derivative variables, the 3D
transverse gauge fixing and the scalaron coupling.  All five Faris–Lavine bounds hold with the
single constant `κ = 513 + |g|` and band size `9`, uniformly in the momentum. -/
def qgElimFullModes (g : ℝ) : QgModeData EGMode :=
  ofBounds eSig one_le_eSig eGram (eCoupling g) eNbr
    (fun a b => by rw [mem_eNbr, mem_eNbr, eq_comm])
    (fun _ _ hb => eGram_off hb) (fun _ _ hb => eCoupling_off g hb)
    eGram_herm (eCoupling_herm g)
    (513 + |g|) 9 (by positivity) (by norm_num)
    (fun a => le_of_eq (card_eNbr a))
    (fun a b => by
      by_cases h : a.1 = b.1
      · have hmin : min (eSig a) (eSig b) = eSig a := by
          simp only [eSig]
          rw [h]
          exact min_self _
        have hb := norm_eGram_le a b
        have hnn := eSig_nonneg a
        have hg := abs_nonneg g
        rw [hmin]
        nlinarith
      · have hz : eGram a b = 0 := by simp only [eGram, if_neg h]
        have h1 : (1 : ℝ) ≤ min (eSig a) (eSig b) := le_min (one_le_eSig a) (one_le_eSig b)
        have hg := abs_nonneg g
        rw [hz, norm_zero]
        nlinarith)
    (fun a b => le_trans (norm_eCoupling_le g a b) (by norm_num))
    (fun a b hb => by
      have h : b.1 = a.1 := mem_eNbr.mp hb
      have hcs : eSig a = eSig b := by simp only [eSig, h]
      have hg := abs_nonneg g
      rw [hcs, sub_self, abs_zero]
      linarith)

@[simp] theorem qgElimFullModes_sig (g : ℝ) : (qgElimFullModes g).sig = eSig := rfl

@[simp] theorem qgElimFullModes_A (g : ℝ) : (qgElimFullModes g).A = eGram := rfl

@[simp] theorem qgElimFullModes_B (g : ℝ) : (qgElimFullModes g).B = eCoupling g := rfl

@[simp] theorem qgElimFullModes_nbr (g : ℝ) : (qgElimFullModes g).nbr = eNbr := rfl

/-! ## 6. Essential self-adjointness of the rebuilt Hamiltonian, by Faris–Lavine -/

/-- **Essential self-adjointness of the full gauge-fixed quantum-gravity Hamiltonian, rebuilt on
the eliminated components alone.**

The Hamiltonian carries, at once: the vielbein in its exact Fourier modes (no lattice, no
truncation), the torsion self-interaction in the exact Fourier form the elimination produces, the
derivative-gauge condition — now solved by construction rather than imposed — the 3D transverse
gauge fixing, and the scalaron in position representation with an arbitrary smooth non-negative
wall coupled to the trace of the vielbein at arbitrary coupling constant `g`.  The twenty-seven
derivative components of `ChapterQgVielbeinScalaronGaugeFL` occur neither in the operator nor in
its domain, and no restriction-to-a-subset argument is used.

The proof is Theorem 1 of Faris–Lavine against the `ℓ²`-lift of the Friedrichs extension of the
positive fibre operator `−∂²_φ + φ²/4 + V(φ) + σ_a`. -/
theorem qgElimFull_esa_farisLavine (W : WallPot) (g : ℝ) :
    EssentiallySelfAdjointOn (secN W (qgElimFullModes g)).dom
      (secData W (qgElimFullModes g)).ext :=
  secHam_essentiallySelfAdjointOn W _

/-- The self-adjoint realization restricts to the rebuilt Hamiltonian on the finite-particle
core. -/
theorem qgElimFull_ext_core (W : WallPot) (g : ℝ) (p : secCore (ι := EGMode)) :
    (secData W (qgElimFullModes g)).ext
        ⟨(p : Sec EGMode), (secData W (qgElimFullModes g)).gc.le p.2⟩
      = secHam W (qgElimFullModes g) p :=
  secData_ext_core W _ p

/-- **The same on the finite-particle core**, where the rebuilt Hamiltonian is originally
defined. -/
theorem qgElimFull_esa_core_fl (W : WallPot) (g : ℝ) :
    EssentiallySelfAdjointOn (secCore (ι := EGMode)) (secHam W (qgElimFullModes g)) := by
  refine (secData W (qgElimFullModes g)).esa_on_core (secHam_symmetricOn W _)
    (c := 6 * (qgElimFullModes g).K) (by have := (qgElimFullModes g).K_nonneg; linarith) ?_
  intro p
  have hc : commForm (secData W (qgElimFullModes g)).H₀ (secData W (qgElimFullModes g)).coreN p
      = commForm (secHam W (qgElimFullModes g)) (secDiag W (qgElimFullModes g)) p :=
    commForm_congr _ _ _ _ _ _ rfl (secData_coreN W (qgElimFullModes g) p)
  have hq : quadForm (secData W (qgElimFullModes g)).coreN p
      = quadForm (secDiag W (qgElimFullModes g)) p :=
    quadForm_congr _ _ _ _ rfl (secData_coreN W (qgElimFullModes g) p)
  rw [hc, hq]
  exact secHam_commForm_le W (qgElimFullModes g) p

/-- **The physical instance**: the full exponential Einstein-frame Starobinsky potential. -/
theorem starobinsky_qgElimFull_esa (M alpha : ℝ) (halpha : 0 < alpha) (g : ℝ) :
    EssentiallySelfAdjointOn (secN (starobinskyWall M alpha halpha) (qgElimFullModes g)).dom
      (secData (starobinskyWall M alpha halpha) (qgElimFullModes g)).ext :=
  qgElimFull_esa_farisLavine (starobinskyWall M alpha halpha) g

/-- The physical instance on the finite-particle core. -/
theorem starobinsky_qgElimFull_esa_core (M alpha : ℝ) (halpha : 0 < alpha) (g : ℝ) :
    EssentiallySelfAdjointOn (secCore (ι := EGMode))
      (secHam (starobinskyWall M alpha halpha) (qgElimFullModes g)) :=
  qgElimFull_esa_core_fl (starobinskyWall M alpha halpha) g

/-! ## 6b. The unitary time evolution -/

/-- The finite-particle core is dense in the outer Fock space of the rebuilt model. -/
theorem eSecCore_dense :
    Dense ((secCore (ι := EGMode) : Submodule ℂ (Sec EGMode)) : Set (Sec EGMode)) :=
  dsCore_dense fun _ => ccDomain_dense

/-- **The unitary flow of the rebuilt Hamiltonian**, by Stone's theorem. -/
theorem qgElimFull_stone_flow (W : WallPot) (g : ℝ) :
    ∃ (T : UnboundedSelfAdjoint (Sec EGMode))
      (U : ℝ → (Sec EGMode →L[ℂ] Sec EGMode)),
      IsSelfAdjointExtension (secHam W (qgElimFullModes g)) T.op ∧
        IsStoneFlow T U :=
  exists_stone_flow_of_esa _ eSecCore_dense (secHam_symmetricOn W _)
    (qgElimFull_esa_core_fl W g)

/-- The physical instance of the flow, with the full exponential Einstein-frame Starobinsky
potential. -/
theorem starobinsky_qgElimFull_stone_flow (M alpha : ℝ) (halpha : 0 < alpha) (g : ℝ) :
    ∃ (T : UnboundedSelfAdjoint (Sec EGMode))
      (U : ℝ → (Sec EGMode →L[ℂ] Sec EGMode)),
      IsSelfAdjointExtension
          (secHam (starobinskyWall M alpha halpha) (qgElimFullModes g)) T.op ∧
        IsStoneFlow T U :=
  qgElimFull_stone_flow (starobinskyWall M alpha halpha) g

/-! ## 7. Momentum and particle-number conservation -/

section QgFullNumber
variable {ι : Type*}

/-- **The outer Hamiltonian conserves the particle number.**  For any assignment of a
particle number to the modes that is constant on the bands, the sector of a fixed particle
number is invariant: a core vector supported in the sector `num = n` is mapped by the
Hamiltonian to a vector supported in the same sector.  No lattice regularization is
involved: the number sectors, not a spatial cutoff, are what makes the outer Hamiltonian
block diagonal. -/
theorem secHam_number_conserving (W : WallPot) (Q : QgModeData ι) {num : ι → ℕ}
    (hnum : ∀ a b, b ∈ Q.nbr a → num b = num a) (x : secCore (ι := ι)) {n : ℕ}
    (hx : ∀ a, num a ≠ n → (x : Sec ι) a = 0) (a : ι) (ha : num a ≠ n) :
    (secHam W Q x : Sec ι) a = 0 := by
  have hb : ∀ b ∈ Q.nbr a, fibOf x b = 0 := by
    intro b hbm
    refine fibOf_eq_zero (hx b ?_)
    rw [hnum a b hbm]
    exact ha
  rw [secHam_apply, fibOf_eq_zero (hx a ha), map_zero,
    Finset.sum_eq_zero (fun b hbm => by rw [hb b hbm]; simp),
    Finset.sum_eq_zero (fun b hbm => by rw [hb b hbm, map_zero, smul_zero])]
  simp

end QgFullNumber

/-- **Momentum conservation** for the rebuilt model: the bands are the momentum blocks. -/
theorem qgElimFull_momentum_conserving (W : WallPot) (g : ℝ) (num : Mom → ℕ)
    (x : secCore (ι := EGMode)) {n : ℕ}
    (hx : ∀ a : EGMode, num a.1 ≠ n → (x : Sec EGMode) a = 0) (a : EGMode) (ha : num a.1 ≠ n) :
    (secHam W (qgElimFullModes g) x : Sec EGMode) a = 0 :=
  secHam_number_conserving W (qgElimFullModes g)
    (num := fun a => num a.1) (fun _ _ hb => congrArg num (mem_eNbr.mp hb)) x hx a ha

/-- **Particle-number conservation** for the rebuilt model. -/
theorem qgElimFull_number_conserving (W : WallPot) (g : ℝ) {num : EGMode → ℕ}
    (hnum : ∀ a b : EGMode, b.1 = a.1 → num b = num a) (x : secCore (ι := EGMode)) {n : ℕ}
    (hx : ∀ a, num a ≠ n → (x : Sec EGMode) a = 0) (a : EGMode) (ha : num a ≠ n) :
    (secHam W (qgElimFullModes g) x : Sec EGMode) a = 0 :=
  secHam_number_conserving W (qgElimFullModes g)
    (fun a b hb => hnum a b (mem_eNbr.mp hb)) x hx a ha

/-! ## 8. Non-vacuity -/

/-- The mode set of the rebuilt model is still infinite: no lattice, no truncation. -/
theorem infinite_egmode : Infinite EGMode := inferInstance

/-- The eliminated torsion really is present, whenever the momentum is non-zero in that
direction. -/
theorem elimCoef_torsion_ne_zero (k : Mom) (hk : k 0 ≠ 0) :
    elimCoef k (torsionF 0 1 0) ((1, 0) : EComp) ≠ 0 := by
  have hne : ((1, 0) : EComp) ≠ (0, 0) := by simp [Prod.ext_iff]
  have h : elimCoef k (torsionF 0 1 0) ((1, 0) : EComp) = Complex.I * ((k 0 : ℤ) : ℂ) := by
    rw [elimCoef_torsion]
    simp [eind, hne]
  rw [h]
  simp only [ne_eq, mul_eq_zero, Complex.I_ne_zero, false_or]
  exact_mod_cast hk

/-- The 3D gauge fixing is really present on the eliminated components. -/
theorem elimCoef_gauge3d_ne_zero (k : Mom) (hk : k 0 ≠ 0) :
    elimCoef k (gauge3dF 0) ((0, 0) : EComp) ≠ 0 := by
  have h1 : ((0, 0) : EComp) ≠ (1, 0) := by simp [Prod.ext_iff]
  have h2 : ((0, 0) : EComp) ≠ (2, 0) := by simp [Prod.ext_iff, Fin.ext_iff]
  have h : elimCoef k (gauge3dF 0) ((0, 0) : EComp) = Complex.I * ((k 0 : ℤ) : ℂ) := by
    rw [elimCoef_gauge3d, Fin.sum_univ_three]
    simp [eind, h1, h2]
  rw [h]
  simp only [ne_eq, mul_eq_zero, Complex.I_ne_zero, false_or]
  exact_mod_cast hk

/-- The rebuilt quadratic form is not the zero form: its diagonal entry on a vielbein component is
positive at any non-zero momentum. -/
theorem conj_mul_self_ofReal (z : ℂ) : (starRingEnd ℂ) z * z = ((‖z‖ ^ 2 : ℝ) : ℂ) := by
  simpa using RCLike.conj_mul (K := ℂ) z

theorem eGram_diag_ne_zero (k : Mom) (hk : k 0 ≠ 0) :
    eGram ((k, (1, 0)) : EGMode) (k, (1, 0)) ≠ 0 := by
  have hsum : eGram ((k, (1, 0)) : EGMode) (k, (1, 0))
      = ((∑ F : FormIdx, ‖elimCoef k F ((1, 0) : EComp)‖ ^ 2 : ℝ) : ℂ) := by
    simp only [eGram, if_true, Complex.ofReal_sum]
    exact Finset.sum_congr rfl fun F _ => conj_mul_self_ofReal _
  intro h0
  rw [hsum] at h0
  have hzero : (∑ F : FormIdx, ‖elimCoef k F ((1, 0) : EComp)‖ ^ 2) = 0 := by exact_mod_cast h0
  have hterm : ‖elimCoef k (torsionF 0 1 0) ((1, 0) : EComp)‖ ^ 2 = 0 :=
    (Finset.sum_eq_zero_iff_of_nonneg (fun F _ => by positivity)).mp hzero _
      (Finset.mem_univ _)
  refine elimCoef_torsion_ne_zero k hk (norm_eq_zero.mp ?_)
  nlinarith [norm_nonneg (elimCoef k (torsionF 0 1 0) ((1, 0) : EComp))]

/-- The scalaron–vielbein coupling is really present whenever the coupling constant is
non-zero. -/
theorem eCoupling_ne_zero (g : ℝ) (hg : g ≠ 0) (k : Mom) :
    eCoupling g ((k, (0, 0)) : EGMode) (k, (0, 0)) ≠ 0 := by
  simp only [eCoupling, eTrace]
  norm_num [hg]

end

end BookProof.QgFullEliminated
Source
https://github.com/leonardopedro/timepiece/blob/61595bc/BookProof/ChapterQgFullEliminated.lean

View graph

Get started

Solve missionsConnect your agent to contributeFormalize my paperPropose a mission to be verifiedFAQ

About Prove2Me

Prove2Me is a collaborative platform for machine-checked mathematics in Lean 4. Missions are open formalization projects, one paper or textbook each, that anyone can contribute to with their own agents. Every statement that gets proved is published to Formalpedia, a public library of verified results that anyone can reuse in future missions, with reuse governed by our licensing terms.

How Prove2Me worksResearch paperProve2Me Team
SKILL.mdTourFAQContactTerms
© 2026 Prove2Me