Keyboard shortcuts

Press ← or → to navigate between chapters

Press ? to show this help

Press Esc to hide this help


slug: oeis-a396806-iterate-exponential-mod-six bibkey: hanna2026a396806 doi: null url: https://oeis.org/A396806 triage: theorem motivation_gids:

  • D5/S1/Recurrence/Residue/IterateExponentialParity

A396806 coefficients modulo six

Problem

Paul D. Hanna’s OEIS A396806 entry, dated June 9, 2026, states:

E.g.f. satisfies A(x) = x*exp( A^6(x) ). Conjecture: a(n) == [1,2,0] repeating (mod 3) for n >= 1. Conjecture: a(n) == [1,2,3,4,5,0] repeating (mod 6) for n >= 1.

The formula field defines A^6 as the sixth compositional iterate and gives A(x)=Sum_{n>=1} a(n)x^n/n!. For the unique zero-constant rational formal solution A=Xexp(A^{[6]}), the assertion is forall n : Nat, 1 <= n -> Nat.ModEq 6 (IterateExponentialParity.aK 6 n) n. It entails the quoted modulo-three clause because three divides six.

Motivation

This first-tier external target was preregistered in issue #7920. The existing IterateExponentialParity module settles the parity comment and identifies the coefficient sequence. The separate A396803 result has iterate count three.

Gap

OEIS revision 11, retrieved September 15, 2026, still labels both residue comments Conjecture. The refreshed source and all-state repository searches found no prior modulo-three or modulo-six settlement for A396806. The earlier direct-cross-reference, arXiv, MathOverflow, Math StackExchange and external GitHub sequence-code searches are reused with the bounds stated in #7920; they were not rerun in this refresh. Inaccessible general search engines are not negative evidence. No exhaustive literature coverage or priority is claimed.

Route

In integral EGF coordinates over ZMod 3, set l(n)=n and T(f)(n)=sum_j binomial(n,j)*f(j)*j^(n-j). Lucas decomposition and the binomial theorem give, for all j,q, T^[j](l)(3q)=0 and T^[j](l)(3q+1)=(j+1)^q. Integer composition associativity and coefficient reduction identify the sixth compositional iterate of l with T^[5](l). Its residue-zero terms vanish; its residue-one terms equal one at q=0 and zero otherwise. Residue-two terms remain unrestricted.

For every inner sequence with exactly these support properties, the outer exponential recurrence gives value one at indices 3q and 3q+1: the two contributions S and 2S cancel across the three-index block. The leading index factor then makes stepK 6 preserve l modulo three. Induction through approximationK yields the modulo-three congruence for aK 6. The existing parity_mod_two theorem and the coprime moduli two and three give modulo six.

For the comparison series F=X*exp(X), the stronger identity F^{[6]}=X modulo three is not used: its degree-eight integral EGF coefficient has nonzero residue two. That auxiliary identity’s failure does not refute the source conjecture.

Falsifier

A positive n with aK 6 n % 6 different from n % 6 would contradict the assertion. A_equation, fixed_unique and eCoeff_encode identify the source sequence independently of finite term agreement.

Evidence

The sole result is D5/S1/Recurrence/Residue/IterateExponentialModSix.result. It proves the exact positive-index congruence for the existing natural sequence. The universal proof uses only Classical.choice, Quot.sound and propext; all auxiliary claims are local and their hypotheses are discharged. The Scribe statement expresses the equivalent remainder equality. No finite table or numerical premise supplies the all-index conclusion.

Triage

theorem. The single result proves the source modulo-six assertion and therefore also its modulo-three assertion. Together with the existing parity theorem, these settle the three conjecture comments in the inspected A396806 snapshot. A396805’s modulo-three question remains separate; its modulo-five assertion has previously recorded non-kernel counterexamples.

ASSUMED-UNVERIFIED

Literature coverage is limited to the stated searches; the Lean kernel does not authenticate the external page. The source-to-formal correspondence is independently reviewed through the equation, composition convention and factorial-normalized coefficients. No analytic convergence, global priority or universal congruence for every iterate count is asserted.