OctoInformBeta

OpenAI preprint, September 2026 · §1. Introduction

Section 1: polynomial-growth harmonic functions and the comparison count: Intrinsic completeness and volume growth for smooth Cartesian metrics

Harmonic growth and its comparison count.

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

Polynomial growth of a harmonic function is measured against the distance of the metric, while the comparison count depends on dimension and an integer growth order. The library’s main comparison claim supplies a linearly independent family of such functions.

Let nn and kk be nonnegative integers. Cartesian space is ℝn\mathbb R^n✦, and Mn(ℝ)M_n(\mathbb R) is the space of real n×nn\times n matrices.✦ For a metric gg, write gij(x)g_{ij}(x) for its Cartesian coefficients and gij(x)g^{ij}(x) for the entries of their inverse matrix.

Intrinsic Geometry and Differentiation

The Riemannian length Lg(γ)L_g(\gamma) of a C1C^1 curve γ:ℝ→ℝn\gamma:\mathbb R\to\mathbb R^n is the integral in (1): Lg(γ)=∫01∑i,j=1ngij(γ(t))γ̇i(t)γ̇j(t)dt.L_g(\gamma)=\int_0^1 \sqrt{\sum_{i,j=1}^n g_{ij}(\gamma(t))\dot\gamma^i(t)\dot\gamma^j(t)}\,dt.(1)✦ The intrinsic distance dg(x,y)d_g(x,y) between x,y∈ℝnx,y\in\mathbb R^n is the infimum of the lengths of these curves joining xx to yy, as in (2): dg(x,y)=inf⁡{Lg(γ):γ∈C1(ℝ,ℝn),γ(0)=x,γ(1)=y}.d_g(x,y)=\inf\{L_g(\gamma):\gamma\in C^1(\mathbb R,\mathbb R^n),\ \gamma(0)=x,\ \gamma(1)=y\}.(2)✦

The coordinate derivative ∂if\partial_i f of a differentiable real function ff on ℝn\mathbb R^n is its differential applied to the iith coordinate vector eie_i, as in (3): ∂if(x)=Df(x)[ei].\partial_i f(x)=Df(x)[e_i].(3)✦ The Christoffel symbols Γija\Gamma^a_{ij} of gg are the functions in (4): Γija(x)=12∑b=1ngab(x)(∂igjb(x)+∂jgib(x)−∂bgij(x)).\Gamma^a_{ij}(x)=\frac12\sum_{b=1}^n g^{ab}(x) \bigl(\partial_i g_{jb}(x)+\partial_j g_{ib}(x)-\partial_b g_{ij}(x)\bigr).(4)✦ The Laplace-Beltrami operator Δg\Delta_g on smooth real functions is the operator defined by (5): Δgu(x)=∑i,j=1ngij(x)(∂i∂ju(x)−∑a=1nΓija(x)∂au(x)).\Delta_g u(x)=\sum_{i,j=1}^n g^{ij}(x) \left(\partial_i\partial_j u(x)-\sum_{a=1}^n \Gamma^a_{ij}(x)\partial_a u(x)\right).(5)✦

For a radius R>0R>0, the intrinsic ball volume Vg(R)∈[0,∞]V_g(R)\in[0,\infty] is the Riemannian volume of Bg(0,R)={x∈ℝn:dg(0,x)<R}B_g(0,R)=\{x\in\mathbb R^n:d_g(0,x)<R\}.✦ The Euclidean unit ball volume ωn\omega_n normalizes this volume in (6), where dxdx denotes Lebesgue measure: Vg(R)=∫Bg(0,R)det⁡g(x)dx,ωn=∫{x∈ℝn:∑i=1n(xi)2<1}dx.V_g(R)=\int_{B_g(0,R)}\sqrt{\det g(x)}\,dx, \qquad \omega_n=\int_{\{x\in\mathbb R^n:\,\sum_{i=1}^n(x^i)^2<1\}}dx.(6)

The Definition

Definition 1. A smooth Cartesian Riemannian metric gg on ℝn\mathbb R^n is a matrix field g:ℝn→Mn(ℝ)g:\mathbb R^n\to M_n(\mathbb R) satisfying (7): gij∈C∞(ℝn,ℝ)(1≤i,j≤n),g(x)𝖳=g(x),v𝖳g(x)v>0(x∈ℝn,v∈ℝn\{0}).\begin{gathered} g_{ij}\in C^\infty(\mathbb R^n,\mathbb R)\quad(1\le i,j\le n), \qquad g(x)^{\mathsf T}=g(x),\\ v^{\mathsf T}g(x)v>0\quad (x\in\mathbb R^n,\ v\in\mathbb R^n\setminus\{0\}). \end{gathered}(7)✦ An intrinsically complete metric gg is a metric on ℝn\mathbb R^n satisfying the convergence condition (8): Every dg-Cauchy sequence (xj)j≥0 in ℝn has a pointx∈ℝnsuch that limj→∞dg(xj,x)=0.\begin{gathered} \text{Every $d_g$-Cauchy sequence $(x_j)_{j\ge0}$ in $\mathbb R^n$ has a point} \ x\in\mathbb R^n\\ \text{such that }\lim_{j\to\infty}d_g(x_j,x)=0. \end{gathered}(8)✦ The collection Hk(g)H_k(g) of harmonic functions of growth order at most kk is the set in (9): Hk(g)={u∈C∞(ℝn,ℝ):Δgu(x)=0for every x∈ℝn,|u(x)|≤C(1+dg(0,x))kfor every x∈ℝnfor some C≥0}.H_k(g)=\left\{u\in C^\infty(\mathbb R^n,\mathbb R): \begin{array}{l} \Delta_g u(x)=0\ \text{for every }x\in\mathbb R^n,\\ |u(x)|\le C(1+d_g(0,x))^k\ \text{for every }x\in\mathbb R^n\\ \text{for some }C\ge0 \end{array}\right\}.(9)✦ The Euclidean comparison count En,kE_{n,k} is the integer in (10); subtractions in its binomial indices are truncated at zero: En,k=(n+k−1k)+(n+k−2k−1).E_{n,k}=\binom{n+k-1}{k}+\binom{n+k-2}{k-1}.(10)✦ A metric gg has positive subunit asymptotic volume ratio when its normalized intrinsic volume has the limit in (11), taken in [0,∞][0,\infty]: limR→+∞Vg(R)ωnRn=αfor some α∈(0,1).\lim_{R\to+\infty}\frac{V_g(R)}{\omega_n R^n}=\alpha \qquad\text{for some }\alpha\in(0,1).(11)✦

In the paper, the result reads: for a complete connected smooth Riemannian manifold (Mn,g)(M^n,g) without boundary, a base point o∈Mo\in M, and d≥0d\ge0, Hd(M,g)={u∈C∞(M,ℝ):Δgu=0,|u(x)|≤Cu(1+dg(o,x))dfor every x∈M and some finite Cu}H_d(M,g)=\{u\in C^\infty(M,\mathbb R):\Delta_g u=0,\ |u(x)|\le C_u(1+d_g(o,x))^d\ \text{for every }x\in M\text{ and some finite }C_u\}, with hd=dim⁡ℝHd(M,g)h_d=\dim_{\mathbb R}H_d(M,g) and, for integers k≥1k\ge1 and the Euclidean metric gEg_{\mathrm E}, hk(ℝn,gE)=(n+k−1k)+(n+k−2k−1)h_k(\mathbb R^n,g_{\mathrm E})=\binom{n+k-1}{k}+\binom{n+k-2}{k-1}.

What Is Checked

These definitions are stated over smooth positive definite matrix fields on ℝn\mathbb R^n, with nonnegative integer growth order kk and the intrinsic distance dgd_g of (2).