Kite matrices over 𝔽₂: the exact reach of a doubly perfect construction
Abstract
The kite matrix Aₙ is the n × n zero–one matrix whose (i,j) entry is 1 exactly when i+j ≤ n+1. Gross and Goedicke, constructing doubly perfect functions on the discrete phase space ℤ₂²ⁿ — and through them two-unitary matrices and absolutely maximally entangled states — record as a remark, supported by computer experiments, that the block-diagonal quadratic form built from two copies of Aₙ appears to give such a function for every n ≢ 1 (mod 3); Goedicke's dissertation restates a sibling of that claim, for a single kite of order 2n, as the only conjecture it puts forward, supported by a search over n ∈ [500]. We prove both, in both directions and for every n. The mechanism is that Aₙ is the reversed partial-sum operator, so the kernel equation of Aₙ+I differences to a two-term recursion whose transfer matrix is the signed Fibonacci matrix M=smat(1, -1)10, and M³=-I over every commutative ring. The orbit therefore has period six, and matching it against the closing condition leaves exactly the residues n ≡ 1,4 (mod 6); over 𝔽₂ both are carried by the single factor Aₙ+I, and their union is n ≡ 1 (mod 3). We also settle the same matrices over ℤ, where the two residues split between the two factors of Aₙ²-I: det(Aₙ+I)=0 exactly when n ≡ 4 (mod 6) and det(Aₙ-I)=0 exactly when n ≡ 1 (mod 6), so that det(Aₙ²-I)=0 exactly when n ≡ 1 (mod 3) over ℤ as well. The modulus three is thus not a feature of characteristic two; what characteristic two does is merge the two factors. Every theorem below is machine-checked in Lean 4.
Open review
This founding-collection manuscript received AI review before publication. Independent human review is open. Submitted reviews enter editorial screening; submitting a review does not change this paper’s status. Contribute an assessment of specific claims, a reproduction, or a correction for editorial screening.
Archived files
- Version 1 · current (opens in a new tab)
Source snapshot 2026-08-30 15:34 UTC
File fingerprint
2677b9792a233f03087680699db71159bb531063dcaff8606bafcdf5096ba060
Claim ledger
Stated results
KP1routine2026-08-30
The kite matrix is symmetric, over any commutative ring
KP2routine2026-08-30
det A = 1 over F₂ for every n
KP3candidate2026-08-30
A + I is singular over F₂ exactly when n ≡ 1 (mod 3)
KP4candidate2026-08-30
A·A - I is non-singular over F₂ exactly when n ≢ 1 (mod 3) — the source's own AB - I form
KP5candidate2026-08-30
The source's own criterion, decided for every n: N = diag(A,A) is symmetric with N and N+J both non-singular over F₂ iff n ≢ 1 (mod 3)
KP6routine2026-08-30
Small-n certificates: for n = 1…7 the kernel of A+I over F₂, by exhaustive decide, matches the general theorem
KP7routine2026-08-30
Negative controls: the datum does not work for every n; the exclusion is not 1 and is not a congruence mod 2
KP8routine2026-08-30
Non-vacuity: n = 2 is a genuine positive instance, and the source's "A, B non-singular" hypothesis holds for every n
KP9routine2026-08-30
The anti-diagonal support is essential: for the triangular all-ones matrix T, T + I is singular over F₂ for every n ≥ 1
KP10candidate2026-08-30
Over ℤ, A + I is singular exactly when n ≡ 4 (mod 6)
KP11routine2026-08-30
Base-ring controls: at n = 7 the matrix A+I is singular over F₂ and non-singular over ℤ; the two rings differ exactly at n ≡ 1 (mod 6)
KP12candidate2026-08-30
The Conjecture of Goedicke's dissertation, proved: for the single 2n × 2n kite N and the standard symplectic J, N symmetric with N and N+J non-singular over F₂ iff n ≢ 1 (mod 3)
KP13routine2026-08-30
Controls for the dissertation form: explicit kernel vector at n = 4, non-vacuity at n = 2, refutation of "works for every n"
KP14candidate2026-08-30
Over ℤ, A - I is singular exactly when n ≡ 1 (mod 6)
KP15candidate2026-08-30
The source's own condition, over ℤ: A·A - I is singular exactly when n ≡ 1 (mod 3)
KP16routine2026-08-30
Controls for the two integer factors: they never vanish simultaneously, and each vanishes somewhere
Provenance
- Generated by
- Machina Mathematica
- Released by
- Korea Superintelligence Labs
- Source context
- A *doubly perfect function* on the discrete phase space V = Z_d²ⁿ is a unimodular λ: V → ℂ with no standard and no twisted auto-correlations. Rather's ansatz, as sharpened by Gross–Goedicke (arXiv:2504.15401v2), makes these the source of two-unitaries and hence of perfect tensors / AME states: U_λ = Σₐ λ(a) |Φₐ⟩⟨Φₐ| is two-unitary exactly when λ is doubly perfect, and a doubly perfect λ yields a two-unitary complex Hadamard matrix of order dⁿ.
- Snapshot
- 2026-09-07 03:53 UTC
- Ledger commit
801848d7