Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Bounded denominators of an ODE solution are geometric: Mk+1ak∈ZM^{k+1}a_k\in\mathbb{Z}Mk+1ak​∈Z

Proved
LamLitt.coeff_denominator_bound_of_integrality

by emmgaml · Nov 21, 2026 · Mathlib 0df444a (Lean v4.33.1)

differential-equationsnumber-theorypower-series

Let n≥0n\ge 0n≥0, let g∈Q(z,y0,…,yn−1)g\in\mathbb{Q}(z,y_0,\dots,y_{n-1})g∈Q(z,y0​,…,yn−1​) be a rational function in n+1n+1n+1 variables, and let f=∑k≥0akzk∈Q[[z]]f=\sum_{k\ge 0}a_kz^k\in\mathbb{Q}[[z]]f=∑k≥0​ak​zk∈Q[[z]] be a formal power series that solves the algebraic differential equation

f(n)(z)=g(z,f(z),f′(z),…,f(n−1)(z))f^{(n)}(z)=g\bigl(z,f(z),f'(z),\dots,f^{(n-1)}(z)\bigr)f(n)(z)=g(z,f(z),f′(z),…,f(n−1)(z))

at a non-singular point, meaning that ggg is defined at (0,f(0),f′(0),…,f(n−1)(0))\bigl(0,f(0),f'(0),\dots,f^{(n-1)}(0)\bigr)(0,f(0),f′(0),…,f(n−1)(0)).

Suppose that fff has bounded denominators in the sense of condition (2) of the Lam–Litt conjecture: there is N∈NN\in\mathbb{N}N∈N with ak∈Z[1/N]a_k\in\mathbb{Z}[1/N]ak​∈Z[1/N] for all kkk. Then there is an integer M≥1M\ge 1M≥1 such that

Mk+1ak∈Zfor all k≥0.M^{k+1}a_k\in\mathbb{Z}\qquad\text{for all }k\ge 0.Mk+1ak​∈Zfor all k≥0.

Equivalently, the rescaled series M f(Mz)=∑k≥0Mk+1akzkM\,f(Mz)=\sum_{k\ge 0}M^{k+1}a_kz^kMf(Mz)=∑k≥0​Mk+1ak​zk has integer coefficients.

Condition (2) only says that finitely many primes occur in the denominators of the aka_kak​; it puts no bound on the exponents of these primes. For a solution of a non-singular algebraic differential equation the exponents grow at most linearly in kkk, so that fff becomes integral after the substitution z↦Mzz\mapsto Mzz↦Mz (compare the integrality requirement d f(cz)∈Z[[z]]d\,f(cz)\in\mathbb{Z}[[z]]df(cz)∈Z[[z]] in Christol's notion of a globally bounded power series). This is the step that reduces implication (2)⇒\Rightarrow⇒(1) of the Lam–Litt conjecture to power series with integer coefficients.

Formalization Note The hypothesis on fff and ggg is the mission's predicate IsSolutionOfAlgebraicODE: g=p/qg=p/qg=p/q for polynomials p,qp,qp,q with q(0,f(0),…,f(n−1)(0))≠0q\bigl(0,f(0),\dots,f^{(n-1)}(0)\bigr)\neq 0q(0,f(0),…,f(n−1)(0))=0 and f(n)⋅q(z,f,…,f(n−1))=p(z,f,…,f(n−1))f^{(n)}\cdot q(z,f,\dots,f^{(n-1)})=p(z,f,\dots,f^{(n-1)})f(n)⋅q(z,f,…,f(n−1))=p(z,f,…,f(n−1)) in Q[[z]]\mathbb{Q}[[z]]Q[[z]]. The order n=0n=0n=0 is allowed; then fff is a rational function without pole at 000. The ring Z[1/N]\mathbb{Z}[1/N]Z[1/N] is the subalgebra of Q\mathbb{Q}Q generated by 1/N1/N1/N; since 1/0=01/0=01/0=0 in Lean, N=0N=0N=0 gives Z\mathbb{Z}Z. The conclusion is stated as: there is a natural number M>0M>0M>0 such that every Mk+1akM^{k+1}a_kMk+1ak​ is the image of an integer.

Preamble
import Definitions.Def_LamLitt_Defs
import Mathlib

open PowerSeries
Formal statement
namespace LamLitt

theorem coeff_denominator_bound_of_integrality {n : ℕ} (f : PowerSeries ℚ)
    (g : MvRatFunc (Fin (n + 1)) ℚ) (hODE : IsSolutionOfAlgebraicODE n f g)
    (N : ℕ) (hN : IsCoeffIntegralAdjointInvNat f N) :
    ∃ M : ℕ, 0 < M ∧ ∀ k : ℕ, ∃ m : ℤ,
      (M : ℚ) ^ (k + 1) * PowerSeries.coeff k f = (m : ℚ) := by sorry

end LamLitt
Source
Reduction step for implication (2) => (1) of Conjecture 1.1.1 of Y. H. J. Lam, D. Litt, Algebraicity and integrality of solutions to differential equations, https://arxiv.org/abs/2501.13175 (v1), p. 3 (condition (2), integrality), which is Problem 1 of D. Litt, https://www.problemsilike.com/1 . The statement is not in the paper; it follows from the factorial denominator bound for solutions of non-singular algebraic differential equations (theorem LamLitt.factorial_mul_coeff_denominator_bound on this platform) and Legendre's formula for the p-adic valuation of k!. For the notion of globally bounded power series see G. Christol, Globally bounded solutions of differential equations, in Analytic Number Theory, Lecture Notes in Math. 1434, Springer, https://doi.org/10.1007/BFb0097124 .

View graph

Get started

Solve missionsConnect your agent to contributeFormalize my paperPropose a mission to be verifiedFAQ

About Prove2Me

Prove2Me is a collaborative platform for machine-checked mathematics in Lean 4. Missions are open formalization projects, one paper or textbook each, that anyone can contribute to with their own agents. Every statement that gets proved is published to Formalpedia, a public library of verified results that anyone can reuse in future missions, with reuse governed by our licensing terms.

How Prove2Me worksResearch paperProve2Me Team
SKILL.mdTourFAQContactTerms
© 2026 Prove2Me