OctoInformBeta

OpenAI preprint, September 2026 · §7. Extraction of common exclusion continua

Lemma 7.2: Unit-distance bounds for weak colorings and support detection from positive averages on shrinking disks

Disk averages constrain density and support.

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

Disk averages describe the local distribution of a color near a point of the complex plane. Let μ\mu be planar Lebesgue measure and let c:ℂ→Fin⁡kc:\mathbb{C}\to\operatorname{Fin}k be a coloring, where Fin⁡k={0,…,k−1}\operatorname{Fin}k=\{0,\ldots,k-1\} is the set of color labels.

The angle circle Θ\Theta is the additive quotient circle of period 2π2\pi, and the unit direction u(t)u(t) associated to t∈Θt\in\Theta is Θ=ℝ/2πℤ,u(t)=eit.\Theta=\mathbb{R}/2\pi\mathbb{Z}, \qquad u(t)=e^{\mathrm{i}t}.(1) Equip the circle in (1) with its normalized angular Haar probability measure σ\sigma.✦ A weak measurable coloring c:ℂ→Fin⁡kc:\mathbb{C}\to\operatorname{Fin}k is a map that is almost everywhere measurable for μ\mu and satisfies the exclusion condition c(z)≠c(z+u(t))for (μ⊗σ)-almost every (z,t)∈ℂ×Θ.c(z)\ne c\bigl(z+u(t)\bigr) \quad\text{for }(\mu\otimes\sigma)\text{-almost every }(z,t) \in\mathbb{C}\times\Theta.(2)✦

Disk Densities and Spatial Support

The color indicator fi:ℂ→ℝf_i:\mathbb{C}\to\mathbb{R} of a label i∈Fin⁡ki\in\operatorname{Fin}k is the function fi(z)=𝟏{w∈ℂ:c(w)=i}(z).f_i(z)=\mathbf{1}_{\{w\in\mathbb{C}:c(w)=i\}}(z).(3)✦ For ε∈ℝ\varepsilon\in\mathbb{R}, the disk measure νε\nu_\varepsilon on ℂ\mathbb{C} is normalized Lebesgue measure on the open disk B(0,ε)B(0,\varepsilon) when ε>0\varepsilon>0, and the zero measure otherwise: νε={μ|B(0,ε)μ(B(0,ε)),ε>0,0,ε≤0.\nu_\varepsilon= \begin{cases} \displaystyle\frac{\mu|_{B(0,\varepsilon)}}{\mu(B(0,\varepsilon))}, &\varepsilon>0,\\[6pt] 0, &\varepsilon\le0. \end{cases}(4)✦ The disk color density piε(q)p_i^\varepsilon(q) at a center q∈ℂq\in\mathbb{C} is the average of the indicator in (3) against the translated disk measure from (4): piε(q)=∫ℂfi(q+v)dνε(v).p_i^\varepsilon(q)=\int_{\mathbb{C}}f_i(q+v)\,d\nu_\varepsilon(v).(5)✦

The spatial essential support Hi⊆ℂH_i\subseteq\mathbb{C} of a label i∈Fin⁡ki\in\operatorname{Fin}k is the support of Lebesgue measure restricted to its color class: Hi=supp⁡(μ|{z∈ℂ:c(z)=i}).H_i=\operatorname{supp}\bigl(\mu|_{\{z\in\mathbb{C}:c(z)=i\}}\bigr).(6)✦

The Claim

Theorem 1. Let c:ℂ→Fin⁡kc:\mathbb{C}\to\operatorname{Fin}k be Borel measurable, with disk densities as in (5) and spatial supports as in (6). If cc is a weak measurable coloring in the sense of (2), ε>0\varepsilon>0 is a radius, q,z∈ℂq,z\in\mathbb{C} satisfy |q−z|=1|q-z|=1, and i∈Fin⁡ki\in\operatorname{Fin}k is a label, then piε(q)+piε(z)≤1.p_i^\varepsilon(q)+p_i^\varepsilon(z)\le1.(7)✦ Let (εn)(\varepsilon_n) be a real sequence converging to zero, and let (an)(a_n) be a sequence in ℂ\mathbb{C} converging to xx. If a label i∈Fin⁡ki\in\operatorname{Fin}k satisfies piεn(an)>0p_i^{\varepsilon_n}(a_n)>0 for every nn, then x∈Hi.x\in H_i.(8)✦

In the paper, the result reads: for ε>0\varepsilon>0, every pair q,zq,z with |q−z|=1|q-z|=1 and every label ii satisfy piε(q)+piε(z)≤1p_i^\varepsilon(q)+p_i^\varepsilon(z)\le1, and if an→xa_n\to x, εn→0\varepsilon_n\to0 and liminf⁡n→∞piεn(an)>0\liminf_{n\to\infty}p_i^{\varepsilon_n}(a_n)>0, then x∈Hix\in H_i.

What Is Checked

The bound (7) is stated over Borel measurable weak colorings of ℂ\mathbb{C} with finitely many labels, while (8) is stated over Borel measurable colorings, sequences of centers converging to a point, and real radii converging to zero with positive disk averages at every index.