Eqs. (16)–(20) — symmetric tridiagonal matrix entries
ProvedQuantumSixJ.Network.matrix_entriesLet be a -linear preadditive monoidal category, , with additive and -linear tensor product. The chosen objects for natural labels satisfy and have exact self-pairings. Assume SuQ2Axioms: unique scalar endomorphisms of each , zero morphisms between different labels, the admissible fusion decompositions, and the loop values . Choose trivalent vertices , and assume VertexNormalized W: theta normalization and cyclicity. Scalars act through , fixing . Composition means first , then .
Also assume the separate initial data TetSeed W: for every natural , . These are seed values, not an all-channel network-to-sum identity.
Let . Prove all of the following, for every natural choice of external and channel labels:
- Identify each of the two literal four-vertex closures in the formal statement with the graphical networks and , respectively. The first closure inserts a split on the strand of and fuses to ; the second closes the triangle obtained by splitting to , fusing to , and then to .
- Establish the exact factorization
- Independently show when is inadmissible, so the central zero-theta case is included rather than discarded by cancellation.
- Establish the complete numerical matrix evaluation
Thus the categorical matrix is symmetric and tridiagonal, including both off-diagonals, all inadmissible channels, and . The closure identifications and numerical evaluation are conclusions of the theorem, not additional seed assumptions.
Independent mathematical read-back
Horizontal matrix entries
Setting and all assumptions
Let , , with indeterminate and coefficient embedding fixing . Let be a category in arbitrary object and morphism universes, assumed preadditive, -linear, monoidal, monoidally preadditive, and monoidally -linear. No braiding, symmetric structure, dagger, positivity, or concrete representation category is assumed.
There are labelled objects , , a chosen isomorphism , and chosen exact self-pairings giving cups and caps satisfying both snake identities. Assume:
- Every endomorphism of is a unique -scalar multiple of its identity.
- Every morphism between differently labelled objects is zero.
- For each , there exist maps , , for all natural , satisfying on admissible , and over that finite range. These maps need not be the chosen trivalent maps.
- .
Here means first , then . The notation inserts the canonical associators and unitors for composition and never exchanges strands. Tensor factors have exactly their displayed order; and mean left and right whiskering.
Choose covectors for every natural triple. Their cup-bent forms are defined exactly by
Thus , , and . In particular the covector used in has reversed labels ; this is not complex conjugation.
There are exactly two vertex-normalization assumptions:
where
Define the ordered four-leg states and covectors
The states have outputs , in that order. For signed , define
The boundary permutation is , and the last channel is the inner rung. In addition to category axioms and vertex normalization, assume the separate seed
This evaluates only the already-defined tetrahedral networks with last channel 2. It neither assumes the closure identifications below nor assumes a full all-channel scalar evaluation.
Scalar quantities, including the seed
An integer triple is admissible when its entries are nonnegative, satisfy every triangle inequality, and have even sum. Write for admissibility of both and . Define
For admissible , with , set
otherwise set it to zero. Set
and, on admissible ,
with value zero otherwise. The last bracket is , not a half with the 2 inside its numerator.
For the seed define
When are all admissible,
it is zero otherwise. Halves are integer divisions, exact on the admissible branch.
The operator and five asserted conclusions
The operator acts on the first two boundary legs, without exchanging them:
The rung is split from and fused into . Write . Thus is the input-state channel and the final covector channel. All labels in the following five conclusions range over the natural numbers, with no extra admissibility or nonvanishing hypotheses except the explicit premise in item 4.
- The first literal four-vertex closure is identified with a tetrahedron:
There is no factor in this closure identity. Its right side expands to : the first channel is , not or . 2. Define a morphism by
Then the second literal closure satisfies
Its right side is , with first channel . Again there is no scalar prefactor. 3. The complete matrix entry factors as
The two tetrahedra are endomorphisms of , composed in exactly the displayed order. The denominator is , not an external-channel norm or a square root. 4. If is not admissible, then . 5. The same entry has the explicit numerical value
where the independent scalar definition is
In particular , even when horizontal channel 0 is admissible. The neighboring entries use the norm at the larger channel , not the smaller one. They are symmetric under , and zero off the diagonal and step-2 neighbors. There is no normalization by in the matrix-entry conclusion itself.
The two literal closure identifications in items 1 and 2 are conclusions to establish. They are not supplied as conditional inputs, and they do not follow merely by substituting a name into the seed. The seed assumes scalar evaluations of the canonical right-hand tetrahedra; identification of the literal left-hand composites with those networks remains part of what is asserted. The five statements are conjoined, not a chain of implications adding earlier statements as hypotheses. No rung-eigenvalue theorem, recurrence, basis-completeness statement, or all-channel tetrahedral evaluation is assumed.
Total division and the central-channel-zero clause
Every scalar quotient is a total field quotient: , so . In particular item 3 is asserted even when ; it is not restricted to the admissible central channel and carries no hidden denominator-nonzero premise. Since is defined to be zero on inadmissible triples, its right-hand side then vanishes under total division. Item 4 expressly asserts that the categorical left-hand side vanishes too. Although item 3 together with the zero extension implies that same equality, item 4 remains an explicit conjunct; cancellation by the central theta is not a valid proof of this case. For natural , central admissibility allows precisely step-2 neighboring channels and equal positive channels; is inadmissible. Its zero case is therefore relevant to the exceptional diagonal entry.
The signed factorial zeros likewise cause invalid summands to vanish through total division, not through an implicit nonzero-denominator condition. These rational-function conventions make no claim about specialization at complex poles or roots of unity. The five identities are the asserted statement; no proof of them is supplied here.
import Definitions.Def_QuantumSixJNetwork open CategoryTheory MonoidalCategory open scoped BigOperators QuantumSixJ universe v u
namespace QuantumSixJ.Network
theorem matrix_entries {C : Type u} [Category.{v} C] [Preadditive C] [Linear K C]
[MonoidalCategory C] [MonoidalPreadditive C] [MonoidalLinear K C]
(S : SuQ2Data C) (W : Vertex S) (hS : SuQ2Axioms S) (hW : VertexNormalized W) (hSeed : TetSeed W) :
(∀ a d j k : ℕ,
(W.vertexUp a k d ⊗≫ (W.split a a 2 ▷ (S.V k ⊗ S.V d)) ⊗≫
(S.V a ◁ (W.fuse 2 k j ▷ S.V d)) ⊗≫ W.vertex a j d) =
W.tet a a k j d 2) ∧
(∀ b c j k : ℕ,
let q : S.V 2 ⊗ S.V k ⟶ S.V j :=
(S.V 2 ◁ W.split k b c) ⊗≫ (W.fuse 2 b b ▷ S.V c) ≫ W.fuse b c j
(S.cup j ≫ ((W.split j 2 k ≫ q) ▷ S.V j) ≫ S.cap j) =
W.tet b b k j c 2) ∧
(∀ a b c d j k : ℕ,
W.H a b c d k ≫ W.Kop a b c d ≫ W.Hcov a b c d j =
(ι (⟬(a : ℤ)⟭ * ⟬(b : ℤ)⟭) / ι (θ (2 : ℤ) k j)) •
(W.tet a a k j d 2 ≫ W.tet b b k j c 2)) ∧
(∀ a b c d j k : ℕ, ¬ Admissible (2 : ℤ) k j →
W.H a b c d k ≫ W.Kop a b c d ≫ W.Hcov a b c d j = 0) ∧
(∀ a b c d j k : ℕ,
W.H a b c d k ≫ W.Kop a b c d ≫ W.Hcov a b c d j =
ι (L a b c d j k) • 𝟙 (𝟙_ C)) := by sorry
end QuantumSixJ.NetworkConfirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.