Pseudohyperbolicity in Dynamical Systems

• Pseudohyperbolicity is a dynamical property that generalizes uniform hyperbolicity by permitting local deviations while ensuring a robust invariant splitting of the tangent space.
• It is characterized by a continuous decomposition into expanding and contracting subspaces verified via Lyapunov exponents, covariant Lyapunov vectors, and angle conditions.
• Applications include robust chaos in attractors, analysis of pseudohyperbolic metrics in complex analysis, and numerical tests ensuring structural stability in networked systems.

Pseudohyperbolicity refers to a robust dynamical property generalizing the concept of uniform hyperbolicity in the context of smooth dynamical systems, particularly in the theory of attractors, ergodic theory, and geometric function theory. Unlike strict hyperbolicity, pseudohyperbolicity allows for certain local violations of expansion and contraction, provided these are balanced so that the global splitting into expanding and contracting subspaces remains robust and invariant under small perturbations.

1. Core Definition and Conceptual Foundation

The modern definition of pseudohyperbolicity arises from the study of invariant sets (typically attractors) for smooth flows or diffeomorphisms on a finite-dimensional manifold. Formally, an invariant set $A$ is pseudohyperbolic if, along every trajectory $x(t)$, the tangent space $T_x M$ admits a continuous, invariant direct sum decomposition

$T_x M = E_1(x) \oplus E_2(x),$

with dimensions $p$, $N-p$ respectively, such that:

• Domination: The minimal angle $\alpha(x)=\angle(E_1(x),E_2(x))$ obeys $\inf_{x \in A} \alpha(x) > 0$.
• Volume Expansion: The sum $S_p = \lambda_1 + \ldots + \lambda_p > 0$, where the $\lambda_i$ are the Lyapunov exponents in non-increasing order.
• Uniformity: Both the splitting and the volume expansion are continuous and persist under small $x(t)$0 perturbations (Karatetskaia et al., 2024).
• Contraction in $x(t)$1: All directions in $x(t)$2 exhibit negative Lyapunov exponents.

A key distinction from uniform hyperbolicity is that while every direction in $x(t)$3 need not be expanding individually, the $x(t)$4-volume in $x(t)$5 grows on average. Conversely, $x(t)$6 may momentarily expand but is contracting on average.

2. Pseudohyperbolic Metrics in Complex Analysis

In classical one-variable complex analysis, the term "pseudohyperbolic" also denotes a metric structure in the unit disk $x(t)$7. The pseudohyperbolic metric

$x(t)$8

is Möbius-invariant and intimately linked to the Schwarz–Pick lemma. The holomorphic isometries of $x(t)$9 under $T_x M$0 are precisely the conformal automorphisms (disk Möbius maps). The pseudohyperbolic balls (or disks) $T_x M$1 map to Euclidean circles lying strictly inside $T_x M$2, with explicit formulas for their centers and radii in terms of $T_x M$3 and $T_x M$4. These pseudohyperbolic disks provide a natural structure for studying approach regions, boundary regularity, and value distribution theory, and their union forms a domain sharply characterized by conformal geometry (Mortini et al., 2015).

3. Diagnostic Criteria and Numerical Tests

For dynamical systems, the verification of pseudohyperbolicity involves both Lyapunov spectrum analysis and geometric criteria on the tangent bundle splitting:

• Spectral Criteria: Identify $T_x M$5 so that $T_x M$6 (block domination), $T_x M$7 (volume expansion in $T_x M$8), and all exponents in $T_x M$9 are negative.
• Non-Tangency: Compute the principal angles (via Covariant Lyapunov Vectors, CLVs) between the subspaces $T_x M = E_1(x) \oplus E_2(x),$0 and $T_x M = E_1(x) \oplus E_2(x),$1, ensuring $T_x M = E_1(x) \oplus E_2(x),$2.
• Test Algorithm: Numerically integrate the variational equations, compute Lyapunov exponents, reconstruct CLVs or Lyapunov subspaces, and check prescribed thresholds $T_x M = E_1(x) \oplus E_2(x),$3 for spectral gaps, $T_x M = E_1(x) \oplus E_2(x),$4 for volume expansion, and $T_x M = E_1(x) \oplus E_2(x),$5 for minimal angle. The attractor is pseudohyperbolic if all thresholds are achieved throughout the attractor (Karatetskaia et al., 2024).

Criterion	Analytical/Numerical Condition	Threshold (Typical)
Lyapunov gap	$T_x M = E_1(x) \oplus E_2(x),$6	$T_x M = E_1(x) \oplus E_2(x),$7
Volume expansion	$T_x M = E_1(x) \oplus E_2(x),$8	$T_x M = E_1(x) \oplus E_2(x),$9
Minimal angle	$p$0	$p$1

The method has been systematically applied for flows (e.g., generalized Lorenz systems) and discrete-time maps (e.g., Hénon-type) (Kuptsov et al., 2018, Karatetskaia et al., 2024).

4. Fine-Scale Structure: Instant Lyapunov Exponents and Local Violations

To probe the fine-scale behavior of pseudohyperbolic attractors, instant Lyapunov exponents (ILE) are introduced:

• IBLE (Instant Backward Lyapunov Exponent): At time $p$2,

$p$3

where $p$4 is the $p$5-th backward Lyapunov vector, and $p$6 the Jacobian. The $p$7-volume growth in $p$8 is $p$9, which may become negative transiently despite positive long-time averages.

• ICLE (Instant Covariant Lyapunov Exponent):

$N-p$0

where $N-p$1 is the $N-p$2-th CLV. This measures the instantaneous expansion/contraction of a CLV, which can also show reversals.

These instant exponents directly reveal local "violations" of averaged expansion/contraction in their respective subspaces. For example, the expanding subspace $N-p$3 can support local volume contraction, and the stable $N-p$4 can momentarily exhibit expansion. The domination gap $N-p$5 may also transiently become negative, indicating that instantaneous contraction in $N-p$6 exceeds that in $N-p$7 (Kuptsov et al., 2018).

5. Pseudohyperbolicity in Robust Chaos and Structural Stability

Pseudohyperbolic attractors are structurally stable under small $N-p$8 perturbations. The regularity of the splitting and the uniform lower bound on angle ensures that Lyapunov exponents and subspace structure survive under perturbations, supporting the persistence of chaos. This robustness:

• Withstands time-dependent forcings: Periodic or quasiperiodic perturbations do not destroy the pseudohyperbolic splitting or positive exponents.
• Is inherited by networks: Clusters of weakly coupled systems with pseudohyperbolic attractors retain pseudohyperbolicity for the full network, provided the inter-cluster coupling is sufficiently weak (Karatetskaia et al., 2024).

A notable application is in networks of coupled phase oscillators, where explicit analytic conditions on the coupling function yield pseudohyperbolicity. For instance, in 4-oscillator Kuramoto-type models, chaos can be rendered robust via parameter regimes ensuring a suitable Lyapunov spectrum and non-tangency, verified using the outlined numerical tests (Karatetskaia et al., 2024).

6. Geometric Theory and Pseudohyperbolic Metric Structures

In function theory, the pseudohyperbolic metric's geometric structure underpins analysis of function-theoretic boundary behavior. The envelope of all pseudohyperbolic disks of fixed radius and real centers $N-p$9 in $\alpha(x)=\angle(E_1(x),E_2(x))$0 is explicitly characterized. Under conformal mapping, this region corresponds to an infinite wedge in the right half-plane, whose preimage is bounded in $\alpha(x)=\angle(E_1(x),E_2(x))$1 by two circular arcs symmetric about the real axis:

$\alpha(x)=\angle(E_1(x),E_2(x))$2

meeting the real axis at $\alpha(x)=\angle(E_1(x),E_2(x))$3 under fixed angle $\alpha(x)=\angle(E_1(x),E_2(x))$4 with $\alpha(x)=\angle(E_1(x),E_2(x))$5, $\alpha(x)=\angle(E_1(x),E_2(x))$6. This construction, originated in (Mortini et al., 2015), yields precise control over "Stolz regions" and is central to non-tangential limit results and Julia–Carathéodory regularity theory.

7. Illustrative Examples and Applications

Specific dynamical models exhibiting pseudohyperbolic attractors include:

• Lorenz flows and generalized Lorenz systems, where $\alpha(x)=\angle(E_1(x),E_2(x))$7 but $\alpha(x)=\angle(E_1(x),E_2(x))$8 frequently assumes negative values, reflecting transient local contraction within the "expanding" subspace.
• Higher-dimensional maps (e.g., Hénon-like), where local domination reversals occur but the global pseudohyperbolic structure remains.
• Networks of oscillators: For instance, in the 4-phase oscillator Kuramoto model, explicit parameter regimes produce 4-winged Lorenz-like pseudohyperbolic attractors, which have been verified both analytically (via normal form analysis) and numerically, with detailed Lyapunov diagrams mapping the domains of robust chaos (Karatetskaia et al., 2024).

A plausible implication is that pseudohyperbolicity provides a unifying framework for the geometric and analytic study of robust chaos, bridging ergodic theory, smooth dynamical systems, and geometric complex analysis.

---

References:

• (Mortini et al., 2015) "On a family of pseudohyperbolic disks"
• (Kuptsov et al., 2018) "Lyapunov analysis of strange pseudohyperbolic attractors: angles between tangent subspaces, local volume expansion and contraction"
• (Karatetskaia et al., 2024) "Robust chaos in a totally symmetric network of four phase oscillators"