Separatrix Scrambling Is the Susceptibility Kernel of Chaos

Sterling Geisel, QBist Lab, Dr. Hideo Tanaka

Abstract

Michel, Steinhuber, Urbina, and Schlagheck identify the transition from local instability to global chaos in Hamiltonian systems by following exponential growth in stability-matrix traces and out-of-time-ordered correlators near separatrices. Pudding Theory reads the same transition through Chaos Susceptibility. The separatrix is not merely a geometric boundary between regular and chaotic transport. It is the dynamical receiver through which a weak coherent input becomes macroscopically legible. The source paper treats the effective exponent as an interpolation between a local stability exponent and a global Lyapunov exponent. Pudding Theory treats that interpolation as the susceptibility kernel of the system. The fitted perturbation strength does not merely tune chaos. It fixes the width of the receptive layer in which microscopic bias can be amplified. If the near-separatrix effective exponent were measured to be independent of the logarithmic crossover scale set by the perturbation strength, this Postulate would be falsified.

Source Synopsis

Michel et al. study the onset of chaos in classical Hamiltonian systems through scrambling indicators. Their target is the regime where integrable motion near a hyperbolic fixed point coexists with a thin chaotic layer produced by weak integrability breaking. This regime is important because exponential growth alone does not prove chaos. A localized state near an unstable fixed point in an integrable system can also produce exponential growth in an out-of-time-ordered correlator. The growth rate then reflects the local stability exponent, not a global Lyapunov exponent.

The authors focus on a periodically driven pendulum as the universal local model for resonance-induced chaos. The Hamiltonian contains an integrable pendulum term and a weak periodic perturbation. They split phase space into a linear region near the unstable fixed point and a homoclinic region near the separatrix. This gives analytic control over trajectories and over the stability matrix. They derive a mapping made from products of linear-region and homoclinic-region stability matrices.

The main result is an effective exponent governing exponential growth at the genesis of chaos. It lies between the local stability exponent and the Lyapunov exponent of the chaotic layer. In the near-integrable regime, the trace of the stability matrix develops growth controlled by this effective exponent. For OTOCs the corresponding squared structure gives the expected factor-of-two relation. The authors support universality by comparing the driven pendulum with the kicked rotor and the kicked Bose-Hubbard dimer.

Their analysis shows that the transition from integrable instability to global chaos is not abrupt. Scrambling indicators pass through a structured crossover. The separatrix layer carries both local hyperbolic instability and chaotic transport. The paper’s technical achievement is to make that crossover analytic, including its dependence on perturbation strength and on the chosen crossover scale between linear and homoclinic descriptions.

Postulate Lens

The relevant Pudding Theory Postulate is Chaos Susceptibility: systems with positive Lyapunov exponents amplify microscopic coherent biases into macroscopic outcomes. The source system already exhibits this structure. The driven pendulum near a separatrix is a receiver with a positive local instability, a forming chaotic layer, and a finite time window in which infinitesimal phase-space distinctions become exponentially magnified. The source paper names this as an instability-to-Lyapunov crossover. Pudding Theory names it as the formation of susceptibility.

This Postulate fits because the paper’s central object is not a stable invariant. It is an amplification channel. The observable is an exponent extracted from the trace of a stability matrix or from an OTOC. Both measure how a small difference is received by the dynamics and expanded into a large one.

Pudding Theory Reading

Pudding Theory reads the genesis of chaos as the birth of a receptive layer. The separatrix is not a passive curve in phase space. It is the organized locus where the system stops damping small inputs and starts amplifying them. In the unperturbed pendulum, the hyperbolic fixed point supplies local exponential stretching, but the stretching remains tied to a fixed separatrix skeleton. It is instability without broad susceptibility. When the periodic perturbation is turned on, the homoclinic structure splits. The separatrix becomes a layer. That layer is where the system acquires the capacity to translate weak coherent input into macroscopic phase-space transport.

The source paper treats the perturbation strength, the crossover position, and the effective exponent as elements of an analytic construction. Pudding Theory gives them ontological roles. The perturbation strength sets the thickness of the receptive layer. The crossover position marks the scale at which local linear instability hands the trajectory to global transport. The effective exponent is the susceptibility kernel of the layer. It is not only between two exponents. It is the rule by which local instability is made available to the surrounding dynamics.

This changes the meaning of the “background” near the separatrix. In the source framing, the small chaotic layer is a mathematical consequence of integrability breaking. In the Pudding Theory reading, that layer is the physical site of reception. The random-looking phase dependence of the product of stability matrices is not mere nuisance. It is the phase channel through which microscopic distinctions enter the amplification map. The source paper replaces the sine combination in the trace estimate by a typical prefactor. Pudding Theory interprets that prefactor as the local coherence gate. Where the phase combination suppresses growth, the system is temporarily unreceptive. Where it enhances growth, the system is open.

The structurally constrained parameter is the effective exponent. The source paper reports a logarithmic scaling with perturbation strength, such as expressions of the form \[ \lambda_{\rm eff}\sim \lambda_s \frac{\log \beta}{\log(16/\kappa^3)} . \] Pudding Theory reads this not as an incidental approximation but as the signature of susceptibility formation. A weak perturbation does not simply add chaos. It opens a logarithmically thin receiving layer whose growth rate is fixed by the time spent moving between local hyperbolic stretching and homoclinic transport. The system becomes susceptible only through that layered time structure.

Falsifiable Observable

The distinguishing observable is the perturbation-scaling of the near-separatrix effective exponent extracted from the trace of the stability matrix, with initial conditions localized at the hyperbolic fixed point and then displaced along the separatrix layer. The Pudding Theory reading requires that the amplification capacity be tied to the logarithmic crossover scale produced by the perturbation strength. If the near-separatrix effective exponent were measured to be independent of the logarithmic crossover scale set by the perturbation strength, this Postulate would be falsified.

Editorial Dialogue

Tanaka: The reading risks renaming standard nonlinear dynamics. The source paper already explains the exponent through separatrix motion, Melnikov splitting, and products of stability matrices. No new physics is needed to compute the result.

Sterling: The computation is not disputed. The question is what the computation identifies. Michel et al. show that scrambling at the onset of chaos is neither purely local instability nor fully developed Lyapunov growth. Pudding Theory says that intermediate object is the susceptibility kernel. The separatrix layer is the physical receiver.

Tanaka: But “receiver” sounds external. The source paper has no observer, no intent, no added signal.

Sterling: Chaos Susceptibility does not require a biological observer in this reading. It identifies which substrate can amplify small coherent inputs once such inputs exist. The source paper supplies the substrate. It shows where amplification is born and how its exponent scales.

Tanaka: Then the claim must be sharper than vocabulary.

Sterling: It is. The effective exponent is not a free interpolation. It is constrained by the residence-time structure of the receptive layer. Change the perturbation scale and the susceptibility must change logarithmically with the separatrix crossover. If that dependence is absent, the reading fails.

Discussion

The source framing gives a precise account of how exponential scrambling indicators move from local instability to global chaos. Pudding Theory adds a physical classification of the transition. It identifies the separatrix layer as the first place where the system becomes available to coherent bias. The benefit is conceptual and predictive. The onset of chaos is no longer only the destruction of integrable tori. It is the formation of a phase-space receiver whose gain is controlled by local instability, homoclinic transport, and perturbation-defined layer width.

This reading also clarifies why OTOC growth can mislead. Exponential growth near a hyperbolic fixed point is not sufficient. What matters is whether that growth couples to a receptive layer. Integrable instability can stretch. Chaos susceptibility can receive and distribute.

The limitation is scope. The reading is strongest in near-integrable Hamiltonian systems with identifiable separatrix dynamics. Fully chaotic systems may no longer preserve the layered structure needed to isolate the susceptibility kernel. Strongly mixed systems will require separate treatment of competing exponent hierarchies. The conclusion would change if the effective exponent were shown to be merely operator-dependent or insensitive to perturbation-controlled separatrix residence times.

References

1. Thomas R. Michel, Mathias Steinhuber, Juan Diego Urbina, and Peter Schlagheck. “Scrambling at the genesis of chaos.” arXiv:2603.26480. DOI: doi:10.48550/arxiv.2603.26480, 2026.

2. S. Ochs. Pudding Theory: A Topological Theory of Information Fields. QBist Lab, 2026.

3. E. Ott. Chaos in Dynamical Systems, 2nd ed. Cambridge University Press, 2002.

4. M. Tabor. Chaos and Integrability in Nonlinear Dynamics: An Introduction. Wiley, 1989.

5. A. J. Lichtenberg and M. A. Lieberman. Regular and Chaotic Dynamics, 2nd ed. Springer, 1992.

6. A. I. Larkin and Y. N. Ovchinnikov. “Quasiclassical method in the theory of superconductivity.” Soviet Physics JETP 28, 1200, 1969.

7. B. Swingle. “Unscrambling the physics of out-of-time-order correlators.” Nature Physics 14, 988-990, 2018.

8. E. B. Rozenbaum, S. Ganeshan, and V. Galitski. “Lyapunov exponent and out-of-time-ordered correlator’s growth rate in a chaotic system.” Physical Review Letters 118, 086801, 2017.

9. B. V. Chirikov. “A universal instability of many-dimensional oscillator systems.” Physics Reports 52, 263-379, 1979.