Keyboard shortcuts

Press ← or → to navigate between chapters

Press ? to show this help

Press Esc to hide this help

Square-Denominator Reversion and Dyadic Support

Abstract

Hanna’s A380678 series has odd coefficients exactly at the indices of A027383.

Paul D. Hanna’s entry hanna2025a380678 specifies A(x-A(x)^2/(1-A(x)^2))=x. Its conjecture says that, for n at least one, a(n) is odd exactly when n=32^m-2 or n=42^m-2 for a natural m. The equivalent formulas below use n+2, since the two products are at least three and four respectively.

A denotes generatingSeries and T denotes its private approximations, all over the integers. X is the indeterminate. The operator coeff(n,f) extracts a coefficient, mk constructs a series from its coefficient function, and subst(f,g) composes f with g. The operation invOfUnit(f,1) is the formal power-series inverse with prescribed constant unit one; every denominator used here has constant coefficient one. All indices and exponents are natural numbers.

P denotes the imported lacunarySeries from D5.S1.Recurrence.Invariants.QuadraticReversionDyadicSupportParity. It has coefficients in ZMod(2), with coefficient one exactly when n+2=32^m or n+2=42^m for some natural m, and zero otherwise. The operator map applies its ring homomorphism coefficientwise; intCastRingHom(ZMod(2)) is the canonical map from the integers.

Definition 1.1 (The integral generating series).

Formalization. D5/S1/Recurrence/Parity/QuadraticSquareReversionDyadicSupportParity.generatingSeries (✓ std3).

Source. Repository-derived.

Commentary.

For zero-constant f and g, clearing the unit denominators factors the difference of f^2/(1-f^2) and g^2/(1-g^2) as (f-g)(f+g) times the two inverses. The factor f+g has zero constant coefficient, so the inner arguments gain one degree of agreement. Substitution by an argument with linear coefficient one preserves the first coefficient of a difference. Thus the displayed transformation gains one degree of agreement, and its diagonal coefficient limit is well defined.

Definition 1.2 (The coefficient sequence).

Formalization. D5/S1/Recurrence/Parity/QuadraticSquareReversionDyadicSupportParity.a (✓ std3).

Source. Repository-derived.

Commentary.

The integer a(n) is the coefficient of degree n in A.

Theorem 1.3 (The OEIS equation and normalization).

Proof. Machine-checked in Lean as D5/S1/Recurrence/Parity/QuadraticSquareReversionDyadicSupportParity.generating_equation (✓ std3). ∎

Source. Repository-derived.

Commentary.

The coefficient limit is fixed by f+X-subst(f,X-f^2/(1-f^2)). Rearranging gives exactly the functional equation in hanna2025a380678. The approximations have zero constant coefficient, and their stable linear coefficient is one.

Theorem 1.4 (Uniqueness of the normalized solution).

Proof. Machine-checked in Lean as D5/S1/Recurrence/Parity/QuadraticSquareReversionDyadicSupportParity.generating_unique (✓ std3). ∎

Source. Repository-derived.

Commentary.

Every zero-constant solution is fixed by the same transformation. Induction on the degree of agreement proves equality with A. In particular, this applies under both displayed normalization hypotheses.

Theorem 1.5 (Reduction to the imported support series).

Proof. Machine-checked in Lean as D5/S1/Recurrence/Parity/QuadraticSquareReversionDyadicSupportParity.mod_two_identity (✓ std3). ∎

Source. Repository-derived.

Commentary.

Mapping the integral equation commutes with substitution and the inverse of its unit denominator. In characteristic two, 1-f^2=(1-f)^2. Cancellation of this unit shows that its inverse is invOfUnit(1-f,1)^2. Therefore the imported lacunary_reversion identity for P is also the square-denominator equation. The imported lacunary_quadratic identity gives constant coefficient zero. Applying the same degree-contraction uniqueness argument over ZMod(2) identifies the reduction of A with P.

Theorem 1.6 (Hanna’s parity conjecture).

Proof. Machine-checked in Lean as D5/S1/Recurrence/Parity/QuadraticSquareReversionDyadicSupportParity.hanna_conjecture (✓ std3). ∎

Resolves. Problems/oeis-a380678-quadratic-square-reversion-dyadic-support-parity (proved) by D5/S1/Recurrence/Parity/QuadraticSquareReversionDyadicSupportParity.hanna_conjecture.

Citation. Paul D. Hanna (2025). OEIS A380678, g.f. satisfying A(x - A(x)^2/(1 - A(x)^2)) = x. URL: https://oeis.org/A380678.

Commentary.

Extracting coefficient n from the reduction identity gives the indicator of the two dyadic families. Mathlib’s intCast_eq_one_iff_odd identifies an integer’s image being one in ZMod(2) with its being odd.

References

  • Truth anchor: D5/S1/Recurrence/Parity/QuadraticSquareReversionDyadicSupportParity.a
  • Truth anchor: D5/S1/Recurrence/Parity/QuadraticSquareReversionDyadicSupportParity.generatingSeries
  • Truth anchor: D5/S1/Recurrence/Parity/QuadraticSquareReversionDyadicSupportParity.generating_equation
  • Truth anchor: D5/S1/Recurrence/Parity/QuadraticSquareReversionDyadicSupportParity.generating_unique
  • Truth anchor: D5/S1/Recurrence/Parity/QuadraticSquareReversionDyadicSupportParity.hanna_conjecture
  • Truth anchor: D5/S1/Recurrence/Parity/QuadraticSquareReversionDyadicSupportParity.mod_two_identity
  • Dependency: D5/S1/Recurrence/Invariants/QuadraticReversionDyadicSupportParity