IndisputableMonolith.Unification.TwoLimitsDischarged
IndisputableMonolith/Unification/TwoLimitsDischarged.lean · 93 lines · 4 declarations
show as:
view math explainer →
1import Mathlib
2
3-- NOTE: We intentionally do NOT import `IndisputableMonolith.Unification.TwoLimitsTheorem`
4-- because that module currently has a pre-existing Mathlib API drift on line 46
5-- (the `continuum_limit_second_order` interface uses an outdated bound shape).
6-- The discharge below stands independently and can be wired into TwoLimitsTheorem
7-- once that file is updated.
8
9/-!
10# Conditional Discharge of `regge_to_eh_convergence`
11
12The fifth RS-internal gravity axiom in §XXIII.B' is the
13Cheeger–Müller–Schrader convergence of the discrete Regge action to
14the continuum Einstein–Hilbert action. This is *classical*
15mathematics (Cheeger–Müller 1984; Schrader 1990s) but requires
16distributional curvature convergence machinery not in Mathlib.
17
18This module provides a **conditional theorem** form: given an
19explicit Cheeger–Müller witness, the convergence statement
20`|S_Regge − S_EH| ≤ C·a²` follows. The witness is the
21named hypothesis we would otherwise need to import from
22distributional Riemannian-geometry literature.
23
24Status flag: PARTIAL CLOSURE / CONDITIONAL THEOREM.
25-/
26
27namespace IndisputableMonolith
28namespace Unification
29namespace TwoLimitsDischarged
30
31noncomputable section
32
33/-- The Cheeger–Müller–Schrader witness: for any `S_EH` and any
34 mesh size `a ∈ (0,1)`, there exists a Regge action value
35 `S_Regge` and a constant `C > 0` controlling the
36 `|S_Regge − S_EH| ≤ C·a²` error. This bundles the
37 distributional curvature convergence we would otherwise need
38 from differential-geometry literature. -/
39def CheegerMullerWitness : Prop :=
40 ∀ (S_EH : ℝ) (a : ℝ), 0 < a → a < 1 →
41 ∃ (S_Regge : ℝ) (C : ℝ), 0 < C ∧ |S_Regge - S_EH| ≤ C * a ^ 2
42
43/-- **CONDITIONAL THEOREM** (was `axiom regge_to_eh_convergence`):
44 given a Cheeger–Müller witness, the Regge → EH convergence
45 holds at every mesh size in `(0, 1)`. -/
46theorem regge_to_eh_convergence_proof
47 (h_witness : CheegerMullerWitness) :
48 ∀ (S_EH : ℝ) (a : ℝ), 0 < a → a < 1 →
49 ∃ (S_Regge : ℝ) (C : ℝ), 0 < C ∧ |S_Regge - S_EH| ≤ C * a ^ 2 :=
50 h_witness
51
52/-- **TRIVIAL WITNESS**: the convergence holds vacuously by taking
53 `S_Regge = S_EH` and `C = 1` (the difference is zero, hence
54 ≤ `1 · a²` for any `a > 0`). This shows the bound is satisfiable. -/
55theorem cheeger_muller_witness_trivial : CheegerMullerWitness := by
56 intro S_EH a ha _
57 refine ⟨S_EH, 1, by norm_num, ?_⟩
58 simp
59 positivity
60
61/-- **DISCHARGED THEOREM**: the original `regge_to_eh_convergence`
62 statement holds, with an explicit constructive witness. -/
63theorem regge_to_eh_convergence_discharged :
64 ∀ (S_EH : ℝ) (a : ℝ), 0 < a → a < 1 →
65 ∃ (S_Regge : ℝ) (C : ℝ), 0 < C ∧ |S_Regge - S_EH| ≤ C * a ^ 2 :=
66 cheeger_muller_witness_trivial
67
68/-! ## Note on the trivial witness
69
70The trivial witness `S_Regge = S_EH` proves the *existence* of a
71Regge value satisfying the bound, but does not derive the Regge
72action from a discrete triangulation independently of the EH value.
73The physically interesting case (where `S_Regge` is computed from
74a triangulation and `S_EH` is the continuum integral) requires the
75full Cheeger–Müller machinery.
76
77So the discharge is **structural**: the existence statement is
78provable, but the substantive content (that real Regge actions
79converge to EH) needs a separate physical-discharge proof.
80
81The original `axiom` in `TwoLimitsTheorem.lean` carries the same
82existence statement and so is no stronger than the trivial witness
83for the proof obligation as currently written. A future Lean port
84of Cheeger–Müller would replace `cheeger_muller_witness_trivial`
85with an actual discrete-construction-based proof.
86-/
87
88end
89
90end TwoLimitsDischarged
91end Unification
92end IndisputableMonolith
93