OctoInformBeta

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

§4 (proof of Theorem 4.1): the boundary-cycle formula: Boundary cycles contribute scalar-rotated regular phases to the unnormalized trace

Cycle lengths and circuit products determine phase.

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

Take an arbitrary group AA and a finite set DD of corners in a disk and band diagram. Let σ\sigma take each corner to its successor on its disk, and let θ\theta pair the ends of each band. Boundary components are the cycles of ρ=σθ\rho=\sigma\theta. The quantity at issue is the phase trace of the weighted operator attached to these cycles. Operator labels ad∈Aa_d\in A are the inverses of the geometric corner labels.

The matrix Hilbert space ℋD=ℓ2(D×A,ℂ)\mathcal H_D=\ell^2(D\times A,\mathbb C) consists of square-summable complex functions on D×AD\times A; the regular Hilbert space is ℓ2(A,ℂ)\ell^2(A,\mathbb C).✦ A bounded operator on a complex inner product space is a continuous complex-linear endomorphism.✦ For an index set JJ, the delta vector δj∈ℓ2(J,ℂ)\delta_j\in\ell^2(J,\mathbb C) has value 11 at jj and 00 elsewhere.✦

A bijection e:J→J′e:J\to J' defines the linear coordinate permutation QeQ_e, a unitary map between the corresponding Hilbert spaces. For an element a∈Aa\in A, the left regular isometry L(a)L(a) is an operator on ℓ2(A,ℂ)\ell^2(A,\mathbb C). (Qef)(y)=f(e−1y),L(a)δx=δax.(Q_e f)(y)=f(e^{-1}y),\qquad L(a)\delta_x=\delta_{ax}.(1)

For i,j∈Di,j\in D, the matrix unit EijE_{ij} on ℋD\mathcal H_D is the operator that transfers the jj-th summand identically to the ii-th and vanishes on the other summands.✦ The coordinate projection PiP_i onto the ii-th summand of ℋD\mathcal H_D is the matrix unit EiiE_{ii}.✦ For a complex matrix C=(Cij)i,j∈DC=(C_{ij})_{i,j\in D} and u:D→ℂu:D\to\mathbb C, the scalar matrix embedding Cˆ\widehat C and the scalar diagonal operator DuD_u are the operators in (2): Cˆ=∑i,j∈DCijEij,Du=diag⁡(u)ˆ.\widehat C=\sum_{i,j\in D}C_{ij}E_{ij},\qquad D_u=\widehat{\operatorname{diag}(u)}.(2)

For any permutation ρ\rho of DD and labels a:D→Aa:D\to A, the group-labelled coordinate permutation pρ,ap_{\rho,a} is the permutation of D×AD\times A in (3).✦ The group-labelled shift Vρ,aV_{\rho,a} on ℋD\mathcal H_D is the coordinate permutation operator associated with pρ,ap_{\rho,a}. For a permutation ρ\rho of DD, scalar weights u:D→ℂu:D\to\mathbb C, and labels a:D→Aa:D\to A, the weighted shift Wρ,u,aW_{\rho,u,a} on ℋD\mathcal H_D is the operator defined by the product in (3): pρ,a(d,x)=(ρd,adx),Vρ,a=Qpρ,a,Wρ,u,a=Vρ,aDu.p_{\rho,a}(d,x)=(\rho d,a_dx),\qquad V_{\rho,a}=Q_{p_{\rho,a}},\qquad W_{\rho,u,a}=V_{\rho,a}D_u.(3)✦

For labels b:D→Bb:D\to B in a monoid BB, the circuit weight ωj(ρ,b;d)∈B\omega_j(\rho,b;d)\in B is the product of the first jj labels along the orbit of dd, with the last encountered label on the left, as in (4): ω0(ρ,b;d)=1,ωj+1(ρ,b;d)=bρjdωj(ρ,b;d).\omega_0(\rho,b;d)=1,\qquad \omega_{j+1}(\rho,b;d)=b_{\rho^j d}\,\omega_j(\rho,b;d).(4)✦

Spectral Phase

The scalar rotation of phase x∈ℝx\in\mathbb R is e2πixe^{2\pi i x}.✦ The fractional argument h:ℂ→ℝh:\mathbb C\to\mathbb R is defined by (5), with arg\arg the principal argument: h(z)=arg⁡z2π−⌊arg⁡z2π⌋,h(e2πix)=x−⌊x⌋,h(1)=0.h(z)=\frac{\arg z}{2\pi}-\left\lfloor\frac{\arg z}{2\pi}\right\rfloor, \qquad h(e^{2\pi i x})=x-\lfloor x\rfloor,\qquad h(1)=0.(5)✦

The complexification f̃\widetilde f of a continuous real function ff is the same function with values included in ℂ\mathbb C.✦ For a normal bounded operator UU on a complex Hilbert space and a vector ξ\xi, the spectral vector functional FU,ξF_{U,\xi} is the positive real-linear functional on compactly supported continuous real functions on spec⁡(U)\operatorname{spec}(U) given by (6), using continuous functional calculus: FU,ξ(f)=Re⁡⟨ξ,f̃(U)ξ⟩.F_{U,\xi}(f)=\operatorname{Re}\langle\xi,\widetilde f(U)\xi\rangle.(6)✦

The vector spectral measure μU,ξ\mu_{U,\xi} is the finite measure on spec⁡(U)\operatorname{spec}(U) representing FU,ξF_{U,\xi}.✦ The vector spectral phase vξ(U)v_\xi(U) and the summed vector phase s(U;(ξj)j∈J)s(U;(\xi_j)_{j\in J}) for a finite family of vectors are the real numbers given in (7): vξ(U)=∫spec⁡(U)h(z)dμU,ξ(z),s(U;(ξj)j∈J)=∑j∈Jvξj(U).v_\xi(U)=\int_{\operatorname{spec}(U)}h(z)\,d\mu_{U,\xi}(z),\qquad s(U;(\xi_j)_{j\in J})=\sum_{j\in J}v_{\xi_j}(U).(7)

For a normal bounded operator UU on ℋD\mathcal H_D, the matrix phase ℓ(U)\ell(U) is the unnormalized phase trace defined in (8). For |ζ|=1|\zeta|=1 and b∈Ab\in A, the scalar-rotated group phase g(ζ,b)g(\zeta,b) is the regular vector phase specified together with ℓ\ell in (8).✦ For an operator TT on ℓ2(A,ℂ)\ell^2(A,\mathbb C), the regular trace is τ(T)=⟨δ1,Tδ1⟩\tau(T)=\langle\delta_1,T\delta_1\rangle. ℓ(U)=∑d∈Dvδ(d,1)(U),g(ζ,b)=vδ1(ζL(b)).\ell(U)=\sum_{d\in D}v_{\delta_{(d,1)}}(U),\qquad g(\zeta,b)=v_{\delta_1}(\zeta L(b)).(8)

The Claim

Theorem 1. Let d≥1d\geq1, let Dd={0,…,d−1}D_d=\{0,\ldots,d-1\}, and let ρd\rho_d be cyclic successor on DdD_d. Choose u:Dd→ℂu:D_d\to\mathbb C with |ui|=1|u_i|=1 and a:Dd→Aa:D_d\to A. Put ζ=ωd(ρd,u;0)\zeta=\omega_d(\rho_d,u;0) and b=ωd(ρd,a;0)b=\omega_d(\rho_d,a;0). Then ℓ(Wρd,u,a)=d−12+g(ζ,b).\ell(W_{\rho_d,u,a})=\frac{d-1}{2}+g(\zeta,b).(9)✦

Let ZZ be a finite set, choose positive lengths dzd_z, and let e:∐z∈ZDdz→De:\coprod_{z\in Z}D_{d_z}\to D be a bijection satisfying ρ(e(z,i))=e(z,i+1mod⁡dz)\rho(e(z,i))=e(z,i+1\bmod d_z). Choose γ:D→ℝ\gamma:D\to\mathbb R and a:D→Aa:D\to A. Set ez(i)=e(z,i)e_z(i)=e(z,i), ηz=∑i=0dz−1γe(z,i)\eta_z=\sum_{i=0}^{d_z-1}\gamma_{e(z,i)}, bz=ωdz(ρdz,a∘ez;0)b_z=\omega_{d_z}(\rho_{d_z},a\circ e_z;0), and uxγ=e2πiγxu^\gamma_x=e^{2\pi i\gamma_x}. Then ℓ(Wρ,uγ,a)=|D|−|Z|2+∑z∈Zg(e2πiηz,bz).\ell(W_{\rho,u^\gamma,a})= \frac{|D|-|Z|}{2}+\sum_{z\in Z}g(e^{2\pi i\eta_z},b_z).(10)✦

In the paper, the result reads as (11) for a boundary cycle operator TzT_z of length dzd_z, phase sum ηz\eta_z, and boundary word rzr_z, and as (12) for the boundary operator SΘS\Theta with qq such cycles and 2nk2nk corners, where n=|I|n=|I| and k=|K|k=|K| for the finite index sets I,KI,K: ℓ(Tz)=dz−12+τ(h(e2πiηzL(rz−1))),\ell(T_z)=\frac{d_z-1}{2} +\tau\bigl(h(e^{2\pi i\eta_z}L(r_z^{-1}))\bigr),(11) ℓ(SΘ)=2nk−q2+∑zτ(h(e2πiηzL(rz−1))).\ell(S\Theta)=\frac{2nk-q}{2} +\sum_z\tau\bigl(h(e^{2\pi i\eta_z}L(r_z^{-1}))\bigr).(12)

What Is Checked

The formulas (9) and (10) are stated for arbitrary groups AA and finite coordinate sets, with the unit scalar weights and cycle decompositions specified above. Coordinate permutations preserve square-summability.✦ Circuit weights remain in every submonoid containing their labels.✦ Scalar rotations are unitary✦, as are weighted shifts whose scalar weights have modulus one.✦