OctoInformBeta

OpenAI preprint, September 2026 · §9. Rational values and curve alternatives

Proposition 9.2: For prime ideals over ℂ\mathbb{C} of affine dimension at most 11 and bounded degree under norm-form or divisor relations

Point sparsity or linear weighted Hilbert growth in polynomial quotients.

Hover over a blue or green term in the text, or a node in the graph, to find it in the other. Select it to read its declaration both in Lean and in words.

The Setting

Given an integer N≥0N \ge 0, take a finite set of integer tuples (z0,…,zN−1)∈ℤN(z_0, \dots, z_{N-1}) \in \mathbb{Z}^N lying in a weighted box of scale X≥2X \ge 2 on the zero locus of a prime ideal I⊆ℂ[X0,…,XN−1]I \subseteq \mathbb{C}[X_0, \dots, X_{N-1}] of dimension at most 11 and bounded degree. For any exponent ε>0\varepsilon > 0, the question is whether either the point count is bounded by a constant multiple of XεX^\varepsilon, or the dimension of the space of weighted functions on ℂ[X0,…,XN−1]/I\mathbb{C}[X_0, \dots, X_{N-1}]/I grows at least linearly with the weight threshold.

Two algebraic configurations give rise to this alternative. In the first, with exponent k≥2k \ge 2, a family of d≥2d \ge 2 linear forms built from complex embeddings of a ℚ\mathbb{Q}-linearly independent tuple satisfies kk-th power product relations modulo II against a transcendental quotient element and distinct complex shifts; in the second, for an integer k≥2k \ge 2 and a primitive polynomial f∈ℤ[X]f \in \mathbb{Z}[X] squarefree over ℂ\mathbb{C}, a prime ideal I⊆ℂ[X0,X1,X2]I \subseteq \mathbb{C}[X_0, X_1, X_2] containing f(X0)−X1X2kf(X_0) - X_1 X_2^k carries integer triples (z0,z1,z2)(z_0, z_1, z_2) with X<z0≤2XX < z_0 \le 2X and z2=p>Xz_2 = p > X a prime dividing f(z0)f(z_0).

Definitions

Given a field KK, an integer N≥0N \ge 0, and an ideal I⊆K[X0,…,XN−1]I \subseteq K[X_0, \dots, X_{N-1}], a point z=(z0,…,zN−1)∈ℤNz = (z_0, \dots, z_{N-1}) \in \mathbb{Z}^N on II is an integer tuple such that evaluating every polynomial p∈Ip \in I at zz yields 00 in KK.✦

For a field KK, a set of variables σ\sigma, and an integer T≥0T \ge 0, the space of polynomials of bounded degree 𝒟T⊆K[σ]\mathcal{D}_T \subseteq K[\sigma] is the KK-subspace of polynomials supported on monomials of total degree at most TT.✦ Given an ideal I⊆K[X0,…,XN−1]I \subseteq K[X_0, \dots, X_{N-1}] over a field KK, an affine Hilbert polynomial PI∈ℚ[t]P_I \in \mathbb{Q}[t] of II is a rational univariate polynomial equal at t=Tt = T to the KK-dimension of the image of 𝒟T⊆K[X0,…,XN−1]\mathcal{D}_T \subseteq K[X_0, \dots, X_{N-1}] in K[X0,…,XN−1]/IK[X_0, \dots, X_{N-1}]/I for all sufficiently large integers TT.✦

The affine dimension dim⁡aff(I)\dim_{\mathrm{aff}}(I) of an ideal I⊆K[X0,…,XN−1]I \subseteq K[X_0, \dots, X_{N-1}] with affine Hilbert polynomial PIP_I over a field KK is the degree deg⁡PI\deg P_I (taken to be 00 when PI=0P_I = 0).✦ For an ideal I⊆K[X0,…,XN−1]I \subseteq K[X_0, \dots, X_{N-1}] (with N≥0N \ge 0 and KK a field) having affine Hilbert polynomial PIP_I, the affine degree deg⁡aff(I)\deg_{\mathrm{aff}}(I) of II is the rational number (dim⁡aff(I))!(\dim_{\mathrm{aff}}(I))! multiplied by the coefficient of tdim⁡aff(I)t^{\dim_{\mathrm{aff}}(I)} in PIP_I.✦

The coordinate linear form Lx,c∈K[X0,…,XN−1]L_{x,c} \in K[X_0, \dots, X_{N-1}] associated with an integer d≥1d \ge 1, indices x:{0,…,d−1}→{0,…,N−1}x : \{0, \dots, d-1\} \to \{0, \dots, N-1\}, and coefficients c=(c0,…,cd−1)∈Kdc = (c_0, \dots, c_{d-1}) \in K^d over a field KK is the polynomial given by (1): Lx,c=∑0≤j<dcjXx(j).L_{x,c} = \sum_{0 \le j < d} c_j X_{x(j)}.(1)✦

For a finite variable set σ\sigma, a real weight function ω:σ→ℝ\omega : \sigma \to \mathbb{R}, and a real threshold T∈ℝT \in \mathbb{R}, the set of weighted monomials ℳω(T)⊆ℕσ\mathcal{M}_\omega(T) \subseteq \mathbb{N}^\sigma is the finite set of nonnegative integer exponent vectors on σ\sigma associated with ω\omega and TT.✦ Given a finite variable set σ\sigma, a field KK, an ideal I⊆K[σ]I \subseteq K[\sigma], a weight function ω:σ→ℝ\omega : \sigma \to \mathbb{R}, and a real threshold T∈ℝT \in \mathbb{R}, the real-weighted space of weighted functions ℱω(I,T)\mathcal{F}_\omega(I, T) of II is the KK-subspace of K[σ]/IK[\sigma]/I spanned by the residue classes modulo II of the monomials ∏i∈σXiai\prod_{i \in \sigma} X_i^{a_i} as a=(ai)i∈σa = (a_i)_{i \in \sigma} ranges over ℳω(T)\mathcal{M}_\omega(T).✦ For any ideal I⊆K[σ]I \subseteq K[\sigma] over a field KK in finitely many variables σ\sigma, weight function ω:σ→ℝ\omega : \sigma \to \mathbb{R}, and threshold T∈ℝT \in \mathbb{R}, the weighted Hilbert function Hω(I,T)H_\omega(I, T) of II is the KK-dimension dim⁡Kℱω(I,T)\dim_K \mathcal{F}_\omega(I, T).✦ The monomial weight of a nonnegative integer exponent vector a∈ℕσa \in \mathbb{N}^\sigma with respect to a real weight function ω:σ→ℝ\omega : \sigma \to \mathbb{R} on a finite variable set σ\sigma is the sum over all i∈σi \in \sigma of the value of ω\omega at ii multiplied by the ii-th exponent of aa.✦

The Claim

Theorem 1. First, let N≥0N \ge 0 and d,k≥2d, k \ge 2 be integers, let ω:{0,…,N−1}→(0,∞)\omega : \{0, \dots, N-1\} \to (0, \infty) be a weight function, let x:{0,…,d−1}→{0,…,N−1}x : \{0, \dots, d-1\} \to \{0, \dots, N-1\} be a tuple of coordinate indices, and let U,V,W,C0>0U, V, W, C_0 > 0 be real numbers satisfying ωx(j)=U\omega_{x(j)} = U for all 0≤j<d0 \le j < d and ωj≤W\omega_j \le W for all 0≤j<N0 \le j < N, together with an integer e≥0e \ge 0 and a real number ε>0\varepsilon > 0. Suppose that KK is a field of characteristic 00, v=(v0,…,vd−1)∈Kdv = (v_0, \dots, v_{d-1}) \in K^d is linearly independent over ℚ\mathbb{Q}, σ0,…,σd−1:K→ℂ\sigma_0, \dots, \sigma_{d-1} : K \to \mathbb{C} are ring homomorphisms, θ0,…,θd−1∈ℂ\theta_0, \dots, \theta_{d-1} \in \mathbb{C} are pairwise distinct, X≥2X \ge 2 is real, and 𝒮⊆ℤN\mathcal{S} \subseteq \mathbb{Z}^N is a finite set satisfying |zj|≤C0Xωj|z_j| \le C_0 X^{\omega_j} for all z∈𝒮z \in \mathcal{S} and 0≤j<N0 \le j < N. Suppose further that I⊆ℂ[X0,…,XN−1]I \subseteq \mathbb{C}[X_0, \dots, X_{N-1}] is a prime ideal with dim⁡aff(I)≤1\dim_{\mathrm{aff}}(I) \le 1 and deg⁡aff(I)≤e\deg_{\mathrm{aff}}(I) \le e on which every z∈𝒮z \in \mathcal{S} is a point. Let b0,…,bd−1,q∈ℂ[X0,…,XN−1]b_0, \dots, b_{d-1}, q \in \mathbb{C}[X_0, \dots, X_{N-1}] be polynomials, and write p‾\bar{p} for the residue class modulo II of any p∈ℂ[X0,…,XN−1]p \in \mathbb{C}[X_0, \dots, X_{N-1}] and σi(v)=(σi(v0),…,σi(vd−1))∈ℂd\sigma_i(v) = (\sigma_i(v_0), \dots, \sigma_i(v_{d-1})) \in \mathbb{C}^d. Suppose that b‾i∈ℱω(I,V)\bar{b}_i \in \mathcal{F}_\omega(I, V) and Lx,σi(v)¯kb‾i=q‾−θi\overline{L_{x,\sigma_i(v)}}^{\,k} \bar{b}_i = \bar{q} - \theta_i in ℂ[X0,…,XN−1]/I\mathbb{C}[X_0, \dots, X_{N-1}]/I for every 0≤i<d0 \le i < d, while q‾\bar{q} and each X‾x(s)\bar{X}_{x(s)} for s∈{0,…,d−1}s \in \{0, \dots, d-1\} are transcendental over ℂ\mathbb{C}. Then there exist real constants C1>0C_1 > 0 and H≥0H \ge 0, depending only on N,d,k,ω,x,U,V,W,C0,eN, d, k, \omega, x, U, V, W, C_0, e, and ε\varepsilon, such that (2) holds: |𝒮|≤C1XεorHω(I,T)≥min⁡(2U,max(1U,(d−1)(k−1)V))T−Hfor all T≥0.|\mathcal{S}| \le C_1 X^\varepsilon \qquad\text{or}\qquad H_\omega(I, T) \ge \min\!\left(\frac{2}{U},\, \max\!\left(\frac{1}{U},\, \frac{(d-1)(k-1)}{V}\right)\right) T - H \quad \text{for all } T \ge 0.(2)✦

Second, let f∈ℤ[X]f \in \mathbb{Z}[X] be a primitive polynomial that is squarefree over ℂ\mathbb{C}, let k≥2k \ge 2 and e≥0e \ge 0 be integers, let ω:{0,1,2}→(0,∞)\omega : \{0, 1, 2\} \to (0, \infty) be a weight function with ω0=1\omega_0 = 1, and let C0>0C_0 > 0 and ε>0\varepsilon > 0 be real numbers. Suppose that X≥2X \ge 2 is real, 𝒮⊆ℤ3\mathcal{S} \subseteq \mathbb{Z}^3 is a finite set on which the projection z↦z0z \mapsto z_0 is injective and every z=(z0,z1,z2)∈𝒮z = (z_0, z_1, z_2) \in \mathcal{S} satisfies |zj|≤C0Xωj|z_j| \le C_0 X^{\omega_j} for j∈{0,1,2}j \in \{0, 1, 2\}, X<z0≤2XX < z_0 \le 2X, and z2=pz_2 = p for some prime p>Xp > X dividing f(z0)f(z_0). Suppose further that I⊆ℂ[X0,X1,X2]I \subseteq \mathbb{C}[X_0, X_1, X_2] is a prime ideal with dim⁡aff(I)≤1\dim_{\mathrm{aff}}(I) \le 1 and deg⁡aff(I)≤e\deg_{\mathrm{aff}}(I) \le e on which every z∈𝒮z \in \mathcal{S} is a point, and that f(X‾0)=X‾1X‾2kf(\bar{X}_0) = \bar{X}_1 \bar{X}_2^k in ℂ[X0,X1,X2]/I\mathbb{C}[X_0, X_1, X_2]/I. Then there exists a constant C2≥1C_2 \ge 1, depending only on f,k,e,ω,C0f, k, e, \omega, C_0, and ε\varepsilon, such that (3) holds: |𝒮|≤C2XεorHω(I,T)≥2T−2(ω1+ω2+1)for all T≥1.|\mathcal{S}| \le C_2 X^\varepsilon \qquad\text{or}\qquad H_\omega(I, T) \ge 2T - 2(\omega_1 + \omega_2 + 1) \quad \text{for all } T \ge 1.(3)✦

In the paper, the result reads: putting g=(d−1)(k−1)+𝟏d=4g = (d-1)(k-1) + \mathbf{1}_{d=4} and δ1(U,V)=max⁡{U/2,min⁡(U,V/g)}\delta_1(U, V) = \max\{U/2, \min(U, V/g)\}, for fixed EE and every geometrically irreducible curve CC of degree at most EE defined over ℚ\mathbb{Q} in the coordinates aikbi=n−θia_i^k b_i = n - \theta_i of Proposition 6.2, and for every ε>0\varepsilon > 0, either |S∩C|≪f,E,εXε|S \cap C| \ll_{f,E,\varepsilon} X^\varepsilon or HC(T)≥T/δ1(U,V)−Of,E(1)H_C(T) \ge T / \delta_1(U, V) - O_{f,E}(1) uniformly for U,VU, V in a compact subset of (0,∞)(0, \infty) and under unimodular coordinate changes, where ee, Hω(I,T)H_\omega(I, T), qq, Lx,σi(v)L_{x,\sigma_i(v)}, and 𝒮\mathcal{S} stand for EE, HC(T)H_C(T), nn, aia_i, and S∩CS \cap C, respectively.

What Is Checked

The proposition is stated over prime ideals of complex polynomial rings of affine dimension at most 11 and bounded affine degree, finite sets of points on them, and the weighted Hilbert function of real-weighted spaces of functions, alongside coordinate linear forms in the first part and primitive integer polynomials squarefree over ℂ\mathbb{C} in the second. For every integer N≥0N \ge 0 and every field KK, every ideal I⊆K[X0,…,XN−1]I \subseteq K[X_0, \dots, X_{N-1}] admits an affine Hilbert polynomial of degree at most NN.✦