Skip to content
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
Original file line number Diff line number Diff line change
Expand Up @@ -89,4 +89,89 @@ theorem reidemeister3Connected_arcPartition_witness :
⟨1, 2, 10, 10⟩, ⟨3, 4, 5, 6⟩], numEdges := 10 } := by
decide

/-! ## 4. Controle negatif : le mineur designe s'effondre sur le temoin Y

La section 3 etablit la **premiere** de l'argument de reindexation (la
partition d'arcs est preservee). Le controle de **conclusion** attendu — le
polynome signe du temoin est le meme des deux cotes — est **refute sur ce
temoin**, et c'est une information de cadrage, pas un echec de preuve :

* cote X, le mineur designe vaut `t³ - t²` (chiralite toute positive) ;
* cote Y, le mineur designe est **identiquement nul** — et la sonde Python
fidele a la construction (validee sur les valeurs kernel de `trefoil` et
`figureEight` signe) mesure la meme nullite pour chacune des cinq colonnes
supprimables et chacune des 32 chiralites.

La cause se lit sur les lignes de la matrice : la chirurgie reecrit
`(9, 8, 5, 6)` en `(7, 8, 1, 6)`, dont toutes les etiquettes non singleton
(`7`, `8`, `6`) tombent dans la grande classe du temoin — sa ligne, comme
celle du kink `(1, 2, 10, 10)` (kink R1 exact : `e3 = e4`), ne porte plus que
sur la colonne de l'arc singleton `{1}` et sur la colonne supprimee de la
grande classe. Deux lignes proportionnelles : le rang tombe a 3 et tout
mineur `4 × 4` s'annule.

Le classique « tout mineur `(n-1) × (n-1)` de la matrice d'Alexander vaut
`± t^k · Δ` » suppose la matrice de rang `n - 1` ; le kink ajoute une
relation, et la normalisation designee (premiere ligne et derniere colonne
supprimees, **fixes**) ne survit pas a la chirurgie. Consequence pour la
section 1 : l'argument de reindexation du cas `i ≥ 1` devra soit s'entenir
de diagrammes sans kink, soit rendre le couple (ligne, colonne) supprime
adaptatif.
-/

/-- Determinant 4×4 generique, meme esprit que `det_two_aux` / `det_three_aux`
de `Conway.lean` : expansion de Laplace le long de la premiere colonne, les
mineurs 3×3 etant traites par `det_three_aux`. -/
theorem det_four_aux (A : Matrix (Fin 4) (Fin 4) (Polynomial ℤ)) :
A.det = A 0 0 * (A.submatrix (Fin.succAbove 0) (Fin.succAbove 0)).det
- A 1 0 * (A.submatrix (Fin.succAbove 1) (Fin.succAbove 0)).det
+ A 2 0 * (A.submatrix (Fin.succAbove 2) (Fin.succAbove 0)).det
- A 3 0 * (A.submatrix (Fin.succAbove 3) (Fin.succAbove 0)).det := by
rw [Matrix.det_succ_column_zero]
simp (config := { decide := true }) [Fin.sum_univ_succ]
simp (config := { decide := true }) [det_three_aux, Matrix.submatrix_apply,
Fin.succAbove]
ring

/-- Controle de la tranche 3 : cote X, le mineur designe du polynome signe
(chiralite toute positive) vaut `t³ - t²` — il n'est pas degenere. C'est la
contrepartie positive du controle negatif suivant : c'est bien la chirurgie
qui effondre le mineur, pas une degenerescence anterieure. -/
theorem reidemeister3Connected_alexanderSigned_witness_X :
alexanderPolynomialSigned
{ crossings := [⟨1, 2, 7, 8⟩, ⟨3, 7, 9, 4⟩, ⟨9, 8, 5, 6⟩,
⟨1, 2, 10, 10⟩, ⟨3, 4, 5, 6⟩], numEdges := 10 }
[true, true, true, true, true]
= Polynomial.X ^ 3 - Polynomial.X ^ 2 := by
simp only [alexanderPolynomialSigned]
simp (config := { decide := true })
rw [det_four_aux]
simp only [det_three_aux]
simp only [Matrix.submatrix_apply, Fin.succAbove, Matrix.of_apply]
simp (config := { decide := true }) [alexanderEntrySigned, alexanderEntry,
alexanderEntryNeg]
ring

/-- Controle negatif de la tranche 3 : cote Y, **le meme mineur designe est
identiquement nul** — la chirurgie R3 du temoin aligne la ligne du
croisement reecrit `(7, 8, 1, 6)` sur celle du kink `(1, 2, 10, 10)`
(toutes deux ne portent plus que sur l'arc singleton `{1}` et la grande
classe supprimee), le rang tombe a 3 et le determinant s'annule (sonde :
nullite pour les cinq colonnes supprimables et les 32 chiralites). La
normalisation designee ne survit pas a la chirurgie sur un temoin porteur
de kink — cf section 4 du module. -/
theorem reidemeister3Connected_alexanderSigned_witness_Y_zero :
alexanderPolynomialSigned
{ crossings := [⟨3, 4, 9, 7⟩, ⟨9, 2, 5, 8⟩, ⟨7, 8, 1, 6⟩,
⟨1, 2, 10, 10⟩, ⟨3, 4, 5, 6⟩], numEdges := 10 }
[true, true, true, true, true]
= 0 := by
simp only [alexanderPolynomialSigned]
simp (config := { decide := true })
rw [det_four_aux]
simp only [det_three_aux]
simp only [Matrix.submatrix_apply, Fin.succAbove, Matrix.of_apply]
simp (config := { decide := true }) [alexanderEntrySigned, alexanderEntry,
alexanderEntryNeg]

end Knots
Original file line number Diff line number Diff line change
Expand Up @@ -92,4 +92,88 @@ theorem reidemeister3Connected_arcPartition_witness :
⟨1, 2, 10, 10⟩, ⟨3, 4, 5, 6⟩], numEdges := 10 } := by
decide

/-! ## 4. Negative control: the designated minor collapses on the Y witness

Section 3 establishes the **first premise** of the reindexation argument (the
arc partition is preserved). The expected **conclusion** control — the signed
polynomial of the witness is the same on both sides — is **refuted on this
witness**, and that is framing information, not a proof failure:

* on the X side, the designated minor is `t³ - t²` (all-positive chirality);
* on the Y side, the designated minor is **identically zero** — and the Python
probe faithful to the construction (validated against the kernel values of
the signed `trefoil` and `figureEight`) measures the same nullity for each
of the five deletable columns and each of the 32 chirality assignments.

The cause reads off the matrix rows: the surgery rewrites `(9, 8, 5, 6)` into
`(7, 8, 1, 6)`, whose non-singleton labels (`7`, `8`, `6`) all fall into the
witness's big class — its row, like the kink row `(1, 2, 10, 10)` (an exact
R1 kink: `e3 = e4`), now only touches the singleton arc column `{1}` and the
deleted big-class column. Two proportional rows: the rank drops to 3 and
every `4 × 4` minor vanishes.

The classical "every `(n-1) × (n-1)` minor of the Alexander matrix equals
`± t^k · Δ`" assumes the matrix has rank `n - 1`; the kink adds a relation,
and the designated normalization (first row and last column deleted,
**fixed**) does not survive the surgery. Consequence for section 1: the
reindexation argument of the `i ≥ 1` case must either restrict to kink-free
diagrams or make the deleted (row, column) pair adaptive.
-/

/-- Generic 4×4 determinant, same spirit as `det_two_aux` / `det_three_aux`
from `Conway.lean`: Laplace expansion along the first column, the 3×3
minors being handled by `det_three_aux`. -/
theorem det_four_aux (A : Matrix (Fin 4) (Fin 4) (Polynomial ℤ)) :
A.det = A 0 0 * (A.submatrix (Fin.succAbove 0) (Fin.succAbove 0)).det
- A 1 0 * (A.submatrix (Fin.succAbove 1) (Fin.succAbove 0)).det
+ A 2 0 * (A.submatrix (Fin.succAbove 2) (Fin.succAbove 0)).det
- A 3 0 * (A.submatrix (Fin.succAbove 3) (Fin.succAbove 0)).det := by
rw [Matrix.det_succ_column_zero]
simp (config := { decide := true }) [Fin.sum_univ_succ]
simp (config := { decide := true }) [det_three_aux, Matrix.submatrix_apply,
Fin.succAbove]
ring

/-- Tranche 3 control: on the X side, the designated minor of the signed
polynomial (all-positive chirality) equals `t³ - t²` — it is not
degenerate. This is the positive counterpart of the negative control
below: it is the surgery that collapses the minor, not a prior
degeneracy. -/
theorem reidemeister3Connected_alexanderSigned_witness_X :
alexanderPolynomialSigned
{ crossings := [⟨1, 2, 7, 8⟩, ⟨3, 7, 9, 4⟩, ⟨9, 8, 5, 6⟩,
⟨1, 2, 10, 10⟩, ⟨3, 4, 5, 6⟩], numEdges := 10 }
[true, true, true, true, true]
= Polynomial.X ^ 3 - Polynomial.X ^ 2 := by
simp only [alexanderPolynomialSigned]
simp (config := { decide := true })
rw [det_four_aux]
simp only [det_three_aux]
simp only [Matrix.submatrix_apply, Fin.succAbove, Matrix.of_apply]
simp (config := { decide := true }) [alexanderEntrySigned, alexanderEntry,
alexanderEntryNeg]
ring

/-- Tranche 3 negative control: on the Y side, **the same designated minor is
identically zero** — the witness's R3 surgery aligns the rewritten
crossing's row `(7, 8, 1, 6)` with the kink row `(1, 2, 10, 10)` (both now
only touch the singleton arc `{1}` and the deleted big class), the rank
drops to 3 and the determinant vanishes (probe: null for all five
deletable columns and all 32 chirality assignments). The designated
normalization does not survive the surgery on a kink-carrying witness —
cf section 4 of the module. -/
theorem reidemeister3Connected_alexanderSigned_witness_Y_zero :
alexanderPolynomialSigned
{ crossings := [⟨3, 4, 9, 7⟩, ⟨9, 2, 5, 8⟩, ⟨7, 8, 1, 6⟩,
⟨1, 2, 10, 10⟩, ⟨3, 4, 5, 6⟩], numEdges := 10 }
[true, true, true, true, true]
= 0 := by
simp only [alexanderPolynomialSigned]
simp (config := { decide := true })
rw [det_four_aux]
simp only [det_three_aux]
simp only [Matrix.submatrix_apply, Fin.succAbove, Matrix.of_apply]
simp (config := { decide := true }) [alexanderEntrySigned, alexanderEntry,
alexanderEntryNeg]

end Knots_en
Loading