OctoInformBeta

OpenAI preprint, September 2026 · §4. A planar surface cannot have one nontrivial boundary

§4 (setting): word disks, slots, bands, and boundary cycles: Finite permutations describe the disk arrangement, with planarity given by the Euler equality

Signed disks and their boundary cycles.

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

A disk boundary has positions at which bands attach, separated by coefficient labels in a group AA. The disk successor and the band pairing determine the boundary circuits of the resulting assembly.

Let II be a finite set of word positions, KK a finite set, and ρ\rho a permutation of II. Write a:I→Aa:I\to A for the coefficients and s:I→ℤs:I\to\mathbb Z for the exponents at the positions, and put n=|I|n=|I| and k=|K|k=|K|. The sets K+K_+ and K−K_- are two tagged copies of KK; their elements are written (+,u)(+,u) and (−,u)(-,u).

Let DD be a finite set of positions, let σ,θ\sigma,\theta be permutations of DD, and let η:D→ℤ\eta:D\to\mathbb Z assign traversal exponents. For a permutation π\pi of DD, the set 𝒞(π)\mathcal C(\pi) of cycles is the quotient in (1), where two positions belong to the same cycle when one is an integer iterate of the other: d∼πe⇔e=πj(d) for some j∈ℤ,𝒞(π)=D/∼π.d\sim_\pi e\ \Longleftrightarrow\ e=\pi^j(d)\text{ for some }j\in\mathbb Z, \qquad \mathcal C(\pi)=D/{\sim_\pi}.(1)✦ The class count κ(r)\kappa(r) of an equivalence relation rr on DD is the number of equivalence classes given in (2): κ(r)=|D/r|.\kappa(r)=|D/r|.(2)✦ The cycle count N(π)N(\pi) of a permutation π\pi of DD is the number of its cycles, given in (3): N(π)=|𝒞(π)|.N(\pi)=|\mathcal C(\pi)|.(3)✦

For a finite set VV of disks and a function ε:V→ℤ\varepsilon:V\to\mathbb Z, the oriented vertex successor ν\nu on V×IV\times I is the map from V×IV\times I to itself given by (4): ν(v,i)={(v,ρ(i)),ε(v)=1,(v,ρ−1(i)),ε(v)≠1.\nu(v,i)= \begin{cases} (v,\rho(i)),&\varepsilon(v)=1,\\ (v,\rho^{-1}(i)),&\varepsilon(v)\ne1. \end{cases}(4)✦

The Claim

Definition 1. The word disk slot set SS associated to II and KK is the set in (5): S=(K+⊔K−)×I.S=(K_+\sqcup K_-)\times I.(5)✦ The oriented disk successor σS\sigma_S of the word disk slot set SS is the permutation of SS given in (6): σS(+,u,i)=(+,u,ρ(i)),σS(−,u,i)=(−,u,ρ−1(i)).\sigma_S(+,u,i)=(+,u,\rho(i)), \qquad \sigma_S(-,u,i)=(-,u,\rho^{-1}(i)).(6)✦ The inverse coefficient label function λ:S→A\lambda:S\to A is defined by (7): λ(+,u,i)=ai−1,λ(−,u,i)=aρ−1(i).\lambda(+,u,i)=a_i^{-1}, \qquad \lambda(-,u,i)=a_{\rho^{-1}(i)}.(7)✦ The signed traversal exponent τ:S→ℤ\tau:S\to\mathbb Z is the function in (8): τ(+,u,i)=si,τ(−,u,i)=−si.\tau(+,u,i)=s_i, \qquad \tau(-,u,i)=-s_i.(8)✦

Definition 2. The component relation rσ,θr_{\sigma,\theta} of permutations σ\sigma and θ\theta of DD is the smallest equivalence relation on DD containing the cycle relations of σ\sigma and θ\theta, as in (9): rσ,θ=∼σ∨∼θ.r_{\sigma,\theta}={\sim_\sigma}\vee{\sim_\theta}.(9)✦ A pair (σ,θ)(\sigma,\theta) of permutations of DD is planar when the Euler equality (10) holds: N(σ)+N(θ)+N(σθ)=|D|+2κ(rσ,θ).N(\sigma)+N(\theta)+N(\sigma\theta) =|D|+2\kappa(r_{\sigma,\theta}).(10)✦ Here permutation products act from right to left, so the boundary permutation is σθ:d↦σ(θ(d))\sigma\theta:d\mapsto\sigma(\theta(d)).

Definition 3. A disk atlas (ι,ε)(\iota,\varepsilon) for (ρ,a,s)(\rho,a,s) in (D,σ,η)(D,\sigma,\eta) consists of an injective map ι:V×I→D\iota:V\times I\to D and a polarity function ε:V→ℤ\varepsilon:V\to\mathbb Z on a finite set VV of disks, satisfying (11), with ν\nu as in (4): ε(v)∈{1,−1}(v∈V),σ(ι(v,i))=ι(ν(v,i)),η(ι(v,i))=ε(v)si(v∈V,i∈I).\begin{gathered} \varepsilon(v)\in\{1,-1\}\qquad(v\in V),\\ \sigma(\iota(v,i))=\iota(\nu(v,i)),\qquad \eta(\iota(v,i))=\varepsilon(v)s_i \qquad(v\in V,\ i\in I). \end{gathered}(11)✦

Theorem 1. Let VV be a finite set and let ε:V→{1,−1}\varepsilon:V\to\{1,-1\} have sum zero. Put P={v∈V:ε(v)=1}P=\{v\in V:\varepsilon(v)=1\}, and write P+,P−P_+,P_- for two tagged copies of PP. There is a bijection pairing the positive and negative disks as specified in (12): β:P+⊔P−→∼V,ε(β(+,v))=1,ε(β(−,v))=−1(v∈P).\begin{gathered} \beta:P_+\sqcup P_-\xrightarrow{\ \sim\ }V,\\ \varepsilon(\beta(+,v))=1,\qquad \varepsilon(\beta(-,v))=-1 \qquad(v\in P). \end{gathered}(12)✦

In the paper, the result reads: for any group AA, an infinite cyclic group with generator tt, and a cyclic word w=ts1a1⋯tsnanw=t^{s_1}a_1\cdots t^{s_n}a_n in their free product, with ai∈Aa_i\in A, si∈{1,−1}s_i\in\{1,-1\}, ∑isi=±1\sum_i s_i=\pm1 and identity coefficients allowed, positive disks read ww and negative disks read w−1w^{-1}, every slot is used once by an oriented band joining opposite traversal signs, and, with σ\sigma the disk successor, θ\theta the band pairing, and cdc_d the geometric corner labels, boundary circles are cycles z=(d1,…,dℓ)z=(d_1,\ldots,d_\ell) of σθ\sigma\theta with boundary word rz=cθd1⋯cθdℓr_z=c_{\theta d_1}\cdots c_{\theta d_\ell}, while a connected planar assembly has k≥1k\ge1 disks of each kind and q=nk−2k+2q=nk-2k+2 boundary circles.

What Is Checked

The definitions are stated over arbitrary coefficient groups and permutations of finite sets, and the balance assertion is stated for a finite set of disks whose polarities lie in {1,−1}\{1,-1\} and have sum zero.