# Derivation 18: The Möbius Container ## Claim A bounded, compact, non-orientable surface changes the Kuramoto self-consistency equation from an open dissipative system to a closed resonator. On such a surface, excitations cannot escape — they reflect, interfere, and accumulate until standing waves form at rational phase divisions. The rational divisions are not approximate. They are exact, forced by boundary conditions, and stable enough to persist for arbitrarily many cycles. This is calculable. The boundary conditions are specifiable, the geometry is well-defined, the dynamics are known. What follows is the simulation target: Kuramoto on a compact non-orientable surface, with the minimal parameters to produce one stable rational division from a perturbation of rest. ## Why the container changes everything ### The open-system objection The standard critique of synchronization-as-gravity: beat frequencies dissipate into unbounded space, coupling strength falls off, nothing collects the signal. On R³, this is correct. A finite perturbation spreads, dilutes, and the coherence signal decays as 1/r². ### The compact-surface answer On a compact surface, nothing escapes. Every excitation reflects at the boundary (if there is one) or wraps around (if the surface is closed). The dynamics are: 1. **Reflection**: excitation hits boundary → reverses → encounters itself on the return path 2. **Interference**: reflected and incident waves superpose → standing wave pattern determined by boundary conditions 3. **Accumulation**: energy that would dissipate on an open manifold is trapped → effective coupling strength grows with each traversal 4. **Saturation**: standing wave pattern stabilizes when input matches reflected loss → fixed point of the dynamics This is a resonant cavity. The rational divisions (mode-locked plateaus) are the resonant modes. On an open manifold they are transient; on a compact surface they are attractors. ### Why Möbius specifically A Möbius strip has one boundary and one side. A perturbation traversing the strip encounters itself with reversed orientation after one full traversal. This is forced self-interference — not optional, not tunable, built into the topology. Consequences: - **No inside/outside leakage**: the excitation has nowhere to go except back through the structure - **Orientation reversal**: after one circuit, "phase advance" has become "phase retard." The perturbation interferes destructively with its own image at odd harmonics and constructively at even — this selects specific rational divisions - **Subharmonic generation**: the half-twist means the fundamental round-trip is 2L (twice the strip length), exactly as in Kawano et al. (2025) where the bow contact on a bowed string creates a subharmonic by introducing a reflective boundary with effective frequency ratio 2:1 The Möbius geometry is not decorative. It is the minimal topology that forces self-interference with orientation reversal. ## The simulation target ### Domain A Möbius strip parametrized by: - **Length** L along the non-orientable loop (identified: x ~ x + 2L with orientation reversal at x = L) - **Width** W across the strip (one boundary at w = 0, free or reflecting; identification w → -w at x = x + L gives the twist) For the minimal simulation, W can be collapsed — use the 1D non-orientable circle: the interval [0, L] with identification θ(0) = -θ(L) (antiperiodic boundary condition). This is the simplest compact non-orientable domain. ### Oscillators Place N oscillators uniformly on [0, L]: x_i = (i - 1/2) L/N, i = 1, ..., N Each oscillator has: - Phase θ_i(t) ∈ [0, 2π) - Natural frequency ω_i drawn from distribution g(ω) ### Dynamics Kuramoto with nearest-neighbor coupling on the Möbius domain: dθ_i/dt = ω_i + (K/2)[sin(θ_{i+1} - θ_i) + sin(θ_{i-1} - θ_i)] with the **antiperiodic boundary condition**: θ_{N+1}(t) ≡ θ_1(t) + π (the half-twist) This is the standard Kuramoto ring except the boundary wraps with a π phase shift. The shift is the topological content of the Möbius strip. ### Initial condition: perturbation from rest All oscillators start at θ_i(0) = 0 (uniform phase, "at rest"), with a single localized perturbation: θ_j(0) = ε (one oscillator displaced by ε) The question: does this perturbation grow, propagate, reflect off the antiperiodic boundary, interfere with itself, and produce a stable rational phase division? ### Parameters to determine Three free parameters: 1. **N** — number of oscillators. Derivation 6 says N = 3 is the minimum self-sustaining chain. Start there. 2. **K** — coupling strength relative to frequency spread. The framework gives K_c = 2/(πg(0)) as the synchronization threshold. Scan K/K_c from 0 to 2. 3. **ε** — initial perturbation magnitude. The question is whether any ε > 0 suffices, or whether there is a threshold. The frequency distribution g(ω) should be the framework's self-consistent g* from the rational field equation (Derivation 11), but for the minimal simulation, a Lorentzian of half-width γ centered at ω₀ suffices (this admits the Ott-Antonsen reduction). ## What "one stable rational division" means A rational division p/q is a state where the phase difference between oscillators (or between an oscillator and the mean field) locks to 2πp/q. "Stable" means: 1. The phase difference persists for t → ∞ (or for T >> 1/γ in the dissipative case) 2. Small perturbations around the locked state decay (Lyapunov stable) 3. The winding number of the phase trajectory is exactly p/q (not approximately) On the Möbius domain, the antiperiodic boundary condition θ(L) = θ(0)+π forces the total phase winding across the strip to be (2n+1)π/L for integer n. This quantizes the allowed spatial modes to odd multiples of π/L — the even modes are forbidden by the topology. The first stable rational division is the lowest-energy odd mode: a standing wave with one antinode, winding number 1/2 (half-period across the strip length). ## Analytical predictions ### N = 3 on the Möbius ring Three oscillators with antiperiodic boundary: dθ₁/dt = ω₁ + (K/2)[sin(θ₂ - θ₁) + sin(θ₃ - θ₁ + π)] dθ₂/dt = ω₂ + (K/2)[sin(θ₃ - θ₂) + sin(θ₁ - θ₂)] dθ₃/dt = ω₃ + (K/2)[sin(θ₁ - θ₃ - π) + sin(θ₂ - θ₃)] The +π in the coupling of oscillator 1 to oscillator 3 (and the -π in 3 to 1) is the Möbius twist. Note that sin(α + π) = -sin(α), so the twist converts the attractive coupling to repulsive across the boundary seam. At the fixed point with identical frequencies (ω₁ = ω₂ = ω₃ = ω₀), the uniform state θ₁ = θ₂ = θ₃ is NOT a fixed point (the boundary term gives -sin(π) ≠ 0... actually sin(π) = 0, so the uniform state IS marginally compatible). The first nontrivial solution has: θ₁ = 0, θ₂ = π/3, θ₃ = 2π/3 This is the 1/3 rational division — three oscillators equally spaced in phase, with the boundary twist accommodated by the 2π/3 spacing (since θ₃ + π = 2π/3 + π ≡ 5π/3, and the next oscillator wraps to θ₁ = 0 ≡ 2π, giving a phase jump of 2π - 5π/3 = π/3, consistent). Whether this is stable depends on K. The Jacobian of the N = 3 system at this fixed point determines the critical coupling. ### Critical coupling on the Möbius ring For the standard (periodic) Kuramoto ring with N oscillators and identical frequencies, the fully synchronized state θ_i = const is stable for any K > 0. On the Möbius ring, the antiperiodic boundary shifts the stability threshold. The eigenvalues of the coupling matrix on the Möbius ring are: λ_k = -2K sin²(π(2k+1)/(2N)), k = 0, 1, ..., N-1 (odd modes only, due to antiperiodicity). The most dangerous mode (smallest |λ|) is k = 0: λ₀ = -2K sin²(π/(2N)) Stability requires λ₀ < 0, which holds for all K > 0. But with frequency disorder (ω_i not identical), the effective threshold is: K_c^{Möbius} = Δω / sin(π/(2N)) where Δω is the frequency spread. For N = 3: sin(π/6) = 1/2, so K_c = 2Δω. Compare with the standard ring: K_c = 2Δω/sin(π/N) = 2Δω/sin(π/3) = 2Δω/(√3/2) = 4Δω/√3 ≈ 2.31Δω. The Möbius ring requires **less** coupling to lock — the twist helps, because it introduces a π phase offset that partially compensates the natural frequency differences. ### Minimal parameters for one rational division For N = 3, Lorentzian g(ω) with half-width γ, and initial perturbation ε: 1. **Coupling**: K > K_c = 2γ/sin(π/6) = 4γ (for N = 3 Möbius) 2. **Perturbation**: any ε > 0 suffices — the locked state is a global attractor for K > K_c (Lyapunov, Derivation PRO-LU) 3. **Settling time**: τ ~ 1/(K - K_c) for K near threshold; τ ~ 1/K for K >> K_c 4. **Result**: phase locks to 1/3 rational division (the lowest odd mode on the 3-element Möbius ring) The CPU cycle count to produce one stable division: N_cycles = τ / dt = (1/K) / dt For K = 5γ (moderately above threshold), γ = 1, dt = 0.01: N_cycles ~ 200. Two hundred integration steps. Computationally trivial. ## Coherence length and spiral arm formation ### Local structure before the twist At small N (3, 5), the Möbius twist dominates immediately — every oscillator is within one or two coupling steps of the antiperiodic boundary. The rational divisions appear as global modes from the start. At large N, something different happens. A perturbation at oscillator j propagates outward in both directions at speed v_phase ~ K/γ (coupling strength over frequency spread). The synchronized domain grows as a coherent "arm" — a traveling wave of locked phases surfing along the chain. This arm doesn't know it's on a Möbius strip. It just synchronizes neighbors. The arm learns about the topology when its wavefront has traveled L/2 (half the ring length) and encounters the returning signal — the same perturbation, propagated in the opposite direction, but phase- shifted by π from the twist. At that moment, the growing arm must accommodate the antiperiodic boundary condition. ### The coherence length Define the **coherence length** ξ as the maximum extent of a locally synchronized domain before the twist forces accommodation: ξ = L/2 = N/(2 × lattice spacing) In oscillator units: ξ = N/2 sites. A perturbation starting at site j forms a coherent arm spanning sites j−ξ to j+ξ before the returning wavefront arrives. The **arrival time** of the reflected signal: t_reflect = ξ / v_phase = N / (2 v_phase) where v_phase depends on coupling. For Kuramoto nearest-neighbor with coupling K and frequency spread γ: v_phase ~ K × sin(Δθ_typical) / Δω_typical ~ K / γ So: t_reflect ~ Nγ / (2K) For t < t_reflect, the arm grows freely — the dynamics are indistinguishable from a periodic ring. The arm is a traveling wave with a smooth phase gradient, exactly like a spiral arm in a galaxy: a coherent pattern that propagates because each neighbor locks to the next. ### What happens at t_reflect At t = t_reflect, the two wavefronts meet. The left-propagating front carries phase θ_left. The right-propagating front carries θ_right. The antiperiodic BC requires: θ_left = θ_right + π If the arm has locked to a uniform phase gradient (θ_i = θ_0 + iΔθ), then across the full ring: NΔθ ≡ π (mod 2π) This quantizes the phase gradient: Δθ = (2n + 1)π / N The arm's gradient snaps to the nearest allowed odd mode. The "surf" continues, but now quantized — the wavefronts reinforce on every subsequent traversal because the standing wave satisfies the boundary condition exactly. ### Regimes **Small N (N ≲ 5):** ξ ~ 1–2 sites. No free propagation phase. The twist is felt immediately. Rational divisions emerge as the only possible fixed points. This is the D18 simulation regime. **Intermediate N (N ~ 8–21):** ξ ~ 4–10 sites. Arms form freely, then snap to quantized gradients at t_reflect. The transient looks like galaxy formation: local structure first, then global reorganization when the boundary is felt. **Large N (N ≫ 21):** ξ ≫ 1. Long free-propagation phase. Multiple independent arms can form before any of them reach the boundary. When the reflected signals arrive, the arms must merge and accommodate each other AND the twist. This is the regime where complex mode structure (multiple simultaneous rational divisions, nested tongues) can appear. ### The maximum arm length An arm can persist if its internal coupling is strong enough to maintain coherence across its span. The Kuramoto coherence condition for a chain of n oscillators: K > K_c(n) = 2γ / sin(π / (2n)) For small n: K_c ~ 4γ/π × n (linear growth). For large n: K_c ~ 4γn/π (also linear, sin(π/2n) ~ π/2n). An arm of length n requires coupling K > 4γn/π to stay coherent. Given fixed K, the maximum arm length is: n_max = πK / (4γ) Arms longer than n_max fragment. This sets the **maximum number of coherent spiral arms** on the Möbius ring: N_arms = N / n_max = 4Nγ / (πK) For K/γ = 6 (moderately supercritical): n_max ~ 4.7, so a ring of N = 21 can support ~4 independent arms. Each arm spans ~5 sites and carries a quantized phase gradient. ### Connection to galaxy spiral structure The arms on the Möbius ring are synchronization wavefronts — coherent domains where neighboring oscillators have locked to a common phase gradient. The analogy: | Möbius ring | Spiral galaxy | |---|---| | Synchronized arm | Spiral arm (density wave) | | Phase gradient Δθ | Pattern speed Ω_p | | Arm length n_max | Arm extent (corotation radius) | | Number of arms N/n_max | Number of spiral arms (2, 3, 4) | | Twist at boundary | Global topology of the gravitational field | | Quantized Δθ = (2n+1)π/N | Discrete arm count (not continuous) | The arm count is quantized because the total phase winding must accommodate the antiperiodic boundary condition. The galaxy doesn't "choose" to have 2 arms — the topology forces an integer from the allowed set. ## Simulation specification (coherence length) ``` Domain: 1D Möbius ring, N ∈ {3, 5, 8, 13, 21, 34, 55, 89} Dynamics: Kuramoto nearest-neighbor, antiperiodic BC Parameters: K/γ ∈ {2, 4, 6, 8, 12}, ε = 0.1 Observables: - r(t) trajectory (detect transient vs steady state) - Time to first reflection: t_reflect - Phase gradient before/after reflection - Number of coherent arms at steady state - Arm length distribution Compare: periodic BC at same N — does it form arms at all? Key test: at N = 21, K/γ = 6, predict 4 arms of length ~5 each carrying quantized gradient (2n+1)π/21 ``` ## Connection to existing derivations | This derivation | Builds on | What it adds | |---|---|---| | Antiperiodic boundary | D16 (Möbius topology) | Makes the topology a boundary condition, not a metaphor | | N = 3 minimum | D6 (Planck scale, N = 3) | Tests N = 3 on a new geometry (Möbius vs cylinder) | | Rational divisions | D11 (field equation on tree) | Shows tree structure emerging from dynamics on compact surface | | Global attractor | D-PRO-LU (Lyapunov uniqueness) | Extends uniqueness to non-orientable domain | | Odd mode selection | D16 (half-twist) | Connects topological invariant to spectral selection rule | | Subharmonic | Kawano et al. (2025) | Digital twin of the bowed string subharmonic experiment | ## Simulation specification ``` Domain: 1D ring, N oscillators, antiperiodic BC (θ_{N+1} = θ_1 + π) Dynamics: Kuramoto with nearest-neighbor coupling Frequencies: Lorentzian(ω₀, γ) or self-consistent g* Parameters: N ∈ {3, 5, 8, 13, 21}, K/K_c ∈ [0, 3], ε ∈ [0.01, π] Initial: θ_i = 0 + ε·δ_{i,1} Observables: - Order parameter r(t) = |1/N Σ exp(iθ_i)| - Phase differences Δθ_{i,i+1}(t) - Winding number W(t) = (1/2π) Σ (θ_{i+1} - θ_i) mod 2π - Time to lock: first t where max|dΔθ/dt| < tolerance - Which rational p/q the phase locks to Output: Phase diagram (N, K/K_c) → {locked rational, lock time, mode} Compare: Same parameters on periodic BC (cylinder) — verify Möbius selects odd modes that cylinder does not prefer ``` ## Status **Proposed**: simulation target fully specified. The geometry, dynamics, boundary conditions, parameters, and observables are all explicit. No new mathematics is required — this is a numerical experiment using known equations on a known domain. **What it would establish**: whether the Möbius container produces exact rational divisions from a single perturbation of rest, and whether the odd-mode selection rule holds numerically. If confirmed, the "bounded container" is not a metaphor — it is a resonant cavity for synchronization dynamics, and the rational phase structure is its spectrum. **Open**: - Extension to 2D Möbius strip (finite width W, transverse modes) - Klein bottle (compact, non-orientable, no boundary) as the fully closed variant - Connection to the bowed string: match simulation parameters to Kawano et al. (2025) experimental conditions - Self-consistent g* on the Möbius domain: does the field equation (Derivation 11) have a different fixed point on the non-orientable surface? - The real question: if this is the right container, does it produce the known particle spectrum at the known coupling constants?