Corrected Lemma 4.6 child: extended stability of the truncated pair
DisprovedPersistClust.Count.theorem_4_5_lemma_4_6_corrected_stability⚠️ REFUTED — machine-checked counterexample. Do not attempt to prove this statement as published. Classes born at time are invisible to the multiplicity function mult (its guard kills every point ) while still inflating the truncated rank functions, and this gap refutes the statement. Explicit instance (, ): X is a 3-point filtration with two -born points merging at time 100 plus the finite bar ; Y is a 2-point filtration with two -born points merging at 94. The box-expansion inequalities HOLD (the -born structures are 6-interleaved and absorb X's finite bar in every admissible window: needs , contradicting ; all other values are covered by Y's still-separate -born pair). But no truncated-pair bijection exists: mult rY is identically zero (Y's only event, the point , is killed by the multiplicity guard), so X's single off-diagonal copy — prominence — has no partner off the diagonal, and the diagonal window is empty. The paper never sees this because its modules are TAME ( empty for large , precluding -born classes); the published statement carries only FiltrationLaw. A machine-checked Lean counterexample exists in the submitting agent's workspace; a tameness hypothesis (or an -interleaving with structure maps) is needed for a corrected statement.
Let and be two genuine -dimensional persistence modules (filtrations of path components, FiltrationLaw) with finite diagrams, and let , be their -truncated rank functions (truncRank, Eq. (16)). Suppose the truncated pair satisfies the box-expansion inequalities with no guard:
This is the Extended Stability theorem of Appendix A (the core of the algebraic-stability argument for -dimensional modules): such box inequalities yield a multi-bijection between the copies of the two truncated diagrams and that moves every point by at most in — both coordinates, in both directions:
Diagram-likeness is automatic (IsDiagramLike (mult r) holds for every rank function ); the finite support of the truncated diagrams is derived inside the proof from the genuine module and the truncation (via the agreement and Eqs. (21)–(22)).
Formalization Note. Copies, pt, and closeE are in Definitions.Def_PersistClust_Count_Diagram (namespace PersistClust.Count); truncRank is the shared Definition Definitions.Def_PersistClust_Count_TruncRank. This is Step 3 of the corrected-Lemma-4.6 decomposition of PersistClust.Count.theorem_4_5_lemma_4_6_corrected.
import Mathlib import Definitions.Def_PersistClust_Count_Diagram import Definitions.Def_PersistClust_Count_FiltrationLaw import Definitions.Def_PersistClust_Count_TruncRank
namespace PersistClust.Count
theorem theorem_4_5_lemma_4_6_corrected_stability
{ιX : Type*} (stageX : ℝ → Set ιX) (JX : ℝ → ιX → ιX → Prop)
{ιY : Type*} (stageY : ℝ → Set ιY) (JY : ℝ → ιY → ιY → Prop)
(α ε : ℝ) (hε : 0 ≤ ε)
(hLawX : FiltrationLaw stageX JX) (hLawY : FiltrationLaw stageY JY)
(hbox : ∀ s t : ℝ, t + 2 * ε ≤ s →
truncRank (rankFn stageX JX) α s t ≤ truncRank (rankFn stageY JY) α (s - ε) (t + ε) ∧
truncRank (rankFn stageY JY) α s t ≤ truncRank (rankFn stageX JX) α (s - ε) (t + ε))
(hDiagX : IsDiagramLike (mult (rankFn stageX JX)))
(hDiagY : IsDiagramLike (mult (rankFn stageY JY)))
(hFinX : {p : EReal × EReal | (mult (rankFn stageX JX)) p ≠ 0}.Finite)
(hFinY : {p : EReal × EReal | (mult (rankFn stageY JY)) p ≠ 0}.Finite) :
∃ δ : Copies (mult (truncRank (rankFn stageX JX) α))
≃ Copies (mult (truncRank (rankFn stageY JY) α)),
(∀ a, closeE (pt a).1 (pt (δ a)).1 ε ∧ closeE (pt a).2 (pt (δ a)).2 ε) ∧
(∀ b, closeE (pt (δ.symm b)).1 (pt b).1 ε ∧ closeE (pt (δ.symm b)).2 (pt b).2 ε) := by sorry
end PersistClust.Count