Type Y Heteroclinic Cycles Are Structural Amplifiers of Coherent Bias

Sterling Geisel, QBist Lab; Dr. Hideo Tanaka

Abstract

Podvigina’s type Y heteroclinic cycles show that attraction in a high-dimensional dynamical system is not governed by robustness alone. It is governed by a chain of invariant subspaces, local expansion and contraction rates, and transition matrices whose dominant eigenstructure decides whether nearby trajectories return, escape, or return only from a positive-measure sector. Pudding Theory reads this phenomenon through the Chaos Susceptibility Postulate. A type Y cycle is not merely a fragile skeleton of saddle connections. It is a susceptibility architecture. The cycle receives microscopic deviations at each saddle, sorts them through invariant subspace geometry, and amplifies or suppresses them according to the transition product. What the source paper treats as a stability criterion, Pudding Theory treats as the system’s information-receptive gain. The dominant significant eigenvalue is therefore a structural susceptibility, not only a diagnostic number. If the dominant significant eigenvalue of the measured transition matrix were measured to be less than or equal to 1 while the cycle remained asymptotically stable, this Postulate would be falsified.

Source Synopsis

Podvigina studies a new class of heteroclinic cycles called type Y cycles. A heteroclinic cycle is an invariant set made of equilibria and connecting trajectories. Such cycles arise in population dynamics, chemistry, fluid dynamics, neuroscience, geophysics, and models of magnetic-field reversal or climate regime change. The paper asks when these cycles are asymptotically stable, fragmentarily asymptotically stable, or completely unstable.

The type Y definition generalizes Podvigina’s earlier type Z cycles. Type Y cycles need not be robust in the usual symmetry-protected sense. Instead, the paper prescribes the invariance of certain flow-invariant subspaces. The connecting trajectory from one equilibrium to the next lies inside an invariant subspace, but these subspaces may have different dimensions. The expanding subspace at each node is one-dimensional. Contracting, radial, and transverse subspaces may have arbitrary dimensions. The paper also removes the common assumption that eigenvalues of the linearization near equilibria are distinct. Complex eigenvalues and Jordan cells are allowed when the invariant-subspace structure permits them.

The central construction reduces the flow near the cycle to local maps near equilibria and global maps along connections. In logarithmic coordinates, the relevant part becomes a collection of transition maps. Their basic transition matrices depend on ratios of contracting and transverse eigenvalues to the expanding eigenvalue, together with the dimension of the contracting subspace. Products of these matrices determine stability.

The main result is that type Y cycles obey the same form of stability criteria as type Z cycles. If all transverse eigenvalues are negative, asymptotic stability follows when the largest significant eigenvalue of the transition matrix exceeds 1. If negative matrix entries occur, fragmentary asymptotic stability is decided by whether selected transition matrices have a real dominant significant eigenvalue greater than 1 with an eigenvector whose components have a common sign. The paper illustrates these results with generalized Lotka-Volterra systems in five dimensions.

Postulate Lens

The applicable Pudding Theory postulate is Chaos Susceptibility: highly unstable systems amplify small coherent inputs into macroscopic probability changes. Type Y cycles display the exact structure named by this postulate. A trajectory near the cycle repeatedly approaches saddle equilibria. At each saddle, one direction expands while other directions contract or shear. The source paper’s transition matrices are therefore gain operators for small deviations. The sign, magnitude, and eigenvector structure of their products decide whether microscopic departures are suppressed, amplified into escape, or admitted only through a positive-measure sector of the local basin.

This is not an analogy. The source’s stability theorem already encodes susceptibility. The dominant significant eigenvalue is the cycle-level amplification coefficient accumulated over one circuit. Its associated eigenvector gives the admissible orientation of receptive perturbations. Fragmentary stability is not weak stability. It is directional receptivity.

Pudding Theory Reading

Pudding Theory reads a type Y heteroclinic cycle as an information-sensitive switching skeleton embedded in a chaotic phase space. The source paper defines the skeleton geometrically: equilibria, connecting trajectories, invariant subspaces, and local maps. The Pudding reading adds that this skeleton is the region where coherent bias can enter the macroscopic dynamics without supplying macroscopic energy. The cycle is already unstable at its nodes. The saddle structure creates a sequence of thresholds. A small perturbation near one node is not erased uniformly. It is filtered by invariant subspace membership, then multiplied by the local eigenvalue ratios, then permuted and carried to the next node.

The source treats invariant subspaces as assumptions required to preserve connections. Pudding Theory treats them as receptive channels. A connection inside an invariant subspace is not merely a geometric convenience. It is a conduit in which the flow has reduced freedom and therefore sharper susceptibility. The one-dimensional expanding direction matters because it converts local bias into a discrete routing decision. The contracting dimensions matter because they determine how much unrelated noise is damped before the next routing event.

The source treats the transition matrix as a technical object built from local and global maps. Pudding Theory treats it as the susceptibility tensor of the cycle. Its entries are not arbitrary fit parameters. They are constrained by ratios such as \(-c/e\) and \(-t/e\). These ratios express how strongly the system remembers its incoming deviation relative to the rate at which it is expelled along the outgoing connection. The product around the cycle then defines a closed-loop gain.

This reframes fragmentary asymptotic stability. In the source framing, a fragmentarily stable cycle attracts a positive-measure subset of nearby initial conditions. In the Pudding framing, that subset is the receptive cone of the cycle. The common-sign condition on the dominant eigenvector is the mathematical signature of coherent directional admission. Perturbations aligned with this cone are gathered and returned. Perturbations outside it are expelled. Thus the basin is not just smaller than the full neighborhood. It is structured by the sign geometry of the cycle’s susceptibility.

The substantive claim is this: type Y heteroclinic cycles are macroscopic amplifiers of coherent microscopic bias precisely when their significant transition eigenstructure admits a positive recurrent cone. Their stability criteria are not only stability criteria. They are the conditions under which chaotic switching becomes receptive rather than dispersive.

Falsifiable Observable

The distinguishing observable is the experimentally reconstructed dominant significant eigenvalue and eigenvector of the transition matrix for a type Y cycle, inferred from repeated near-cycle passages in a controlled dynamical system such as a generalized Lotka-Volterra realization or a fluid switching network. The Pudding reading predicts that coherent perturbations aligned with the dominant common-sign eigenvector will increase return probability to the cycle only when the dominant significant eigenvalue exceeds 1. If the dominant significant eigenvalue of the measured transition matrix were measured to be less than or equal to 1 while the cycle remained asymptotically stable, this Postulate would be falsified.

Editorial Dialogue

Tanaka: The reading risks renaming Podvigina’s theorem. The transition matrix already gives stability. Nothing in the mathematics requires an information field or coherent bias. The eigenvalue condition is sufficient inside ordinary dynamical systems.

Sterling: The theorem gives the geometry of stability. Pudding Theory gives the physical interpretation of susceptibility inside that geometry. The point is not that the matrix needs an extra cause to exist. The point is that the matrix identifies where small coherent inputs can become macroscopic routing changes. A saddle cycle is not a neutral background. It is a gain structure.

Tanaka: But the paper assumes invariant subspaces. That is a modeling constraint, not evidence of receptivity.

Sterling: In the source framing, yes. In the Pudding framing, an invariant subspace is exactly the kind of restricted channel that preserves a small directional component long enough for expansion to act on it. The system filters perturbations before it amplifies them. That is a physical susceptibility statement.

Tanaka: Fragmentary stability still attracts only part of a neighborhood.

Sterling: That is the content. The receptive set is not the whole neighborhood. It is the cone selected by the dominant eigenvector and sign condition. The theory does not turn every perturbation into influence. It says the cycle receives only coherent perturbations with the right orientation.

Discussion

This reading buys a sharper account of why heteroclinic cycles matter in physical and biological systems. The source paper shows that stability can persist beyond robust symmetric cycles and beyond distinct eigenvalue assumptions. Pudding Theory explains why that generality matters. A system need not possess global robustness to have local receptivity. It needs a chain of invariant channels, saddle thresholds, and a transition product whose dominant eigenstructure admits return.

The limitation is that Podvigina’s paper is mathematical. It does not measure external coherent perturbations, nor does it specify a laboratory protocol for reconstructing transition matrices from noisy data. The reading therefore attaches to the structure proved in the paper, not to an experiment already performed there. What would change the conclusion is direct evidence that near-cycle coherent directional perturbations fail to correlate with the predicted dominant eigenvector, or that stable return occurs when the transition product lacks the required eigenvalue structure.

The open question is how large the class of receptive heteroclinic networks is once expanding subspaces become multidimensional or radial and contracting eigenvalues are allowed to change sign.

References

1. Podvigina, O. “Asymptotic stability of heteroclinic cycles of type Y.” arXiv:2603.30011, 2026. DOI: doi:10.48550/arxiv.2603.30011.

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

3. Podvigina, O. “Stability and bifurcations of heteroclinic cycles of type Z.” Nonlinearity 25, 1887-1917, 2012.

4. Krupa, M., and Melbourne, I. “Asymptotic stability of heteroclinic cycles in systems with symmetry.” Ergodic Theory and Dynamical Systems 15, 121-148, 1995.

5. Krupa, M., and Melbourne, I. “Asymptotic stability of heteroclinic cycles in systems with symmetry, II.” Proceedings of the Royal Society of Edinburgh 134A, 1177-1197, 2004.

6. Garrido-da-Silva, L., and Castro, S. B. S. D. “Stability of quasi-simple heteroclinic cycles.” Dynamical Systems 34, 14-39, 2019.

7. Castro, S. B. S. D., and Rucklidge, A. M. “Robust Heteroclinic Cycles in Pluridimensions.” Journal of Nonlinear Science 35, 80, 2025.

8. Podvigina, O. “Two-dimensional heteroclinic connections in the generalized Lotka-Volterra system.” Dynamical Systems 38, 163-178, 2023.