Avila–Viana Invariance Principle

• Avila–Viana Invariance Principle is a foundational rigidity concept in smooth ergodic theory that details invariant conditional measures along center foliations.
• It establishes u-invariance of conditional measures, linking vanishing center Lyapunov exponents with the equality of unstable and metric entropy.
• The principle extends C2 results to C1 systems using techniques like fake foliation charts and Hölder estimates, broadening its applicability in dynamical systems.

The Avila–Viana Invariance Principle is a foundational concept in smooth ergodic theory and dynamical systems, providing a precise characterization of how invariant measures disintegrate along center directions of partially hyperbolic diffeomorphisms—particularly when the center Lyapunov exponent vanishes or is nonpositive. At its core, the principle imposes a rigidity condition: conditional measures along invariant foliations (typically the center foliation) are forced to be invariant under holonomy maps relating nearby leaves. This rigidity has deep implications for entropy, measure classification, and statistical properties of both hyperbolic and partially hyperbolic systems.

1. Mathematical Statement of the Avila–Viana Invariance Principle

The invariance principle applies to diffeomorphisms $f: M \to M$ with an invariant splitting of the tangent bundle $TM = E^u \oplus E^c \oplus E^s$, where $E^c$ is the center bundle. If $m$ is an $f$-invariant ergodic measure with vanishing center Lyapunov exponent, the principle asserts the invariance of the conditional measures $m^c_x$ along center leaves under unstable holonomies $H^u_{x \to y}$: $(H^u_{x \to y})_* m^c_x = m^c_y$ for $m$-almost every $x$, $y$ lying in the same center-unstable plaque. This property is referred to as $u$-invariance of the family $\{m^c_x\}$.

In the case of $C^1$ diffeomorphisms with dominated splitting, and particularly for systems where $\dim E^c \leq 1$, the invariance principle continues to hold due to one-dimensional geometric control, even when holonomies are only Hölder continuous rather than Lipschitz, as detailed in (Gan et al., 21 Sep 2025).

2. Relationship to Entropy and Measure Rigidity

The invariance principle interacts closely with entropy formulas, notably those of Ledrappier–Young. When the center Lyapunov exponent is zero and the conditional measures are invariant under holonomy, the unstable entropy coincides with the total metric entropy: $h_m(f) = h^u_m(f)$ This equivalence is established for $C^1$ diffeomorphisms with one-dimensional center and dominated splitting, extending classical $C^2$ results. The invariance principle becomes a key ingredient in classifying measures of maximal entropy, partitioning them into "rotation type" (where the measure is rigidly invariant along center leaves and the dynamics is isometric) and "hyperbolic type" (nonzero center exponent, exhibiting hyperbolic behavior). The dichotomy is fundamental to understanding ergodic and statistical properties.

3. Holonomy Regularity and One-Dimensional Centers

In higher regularity ($C^{1+\alpha}$ or $C^2$), holonomies between center leaves can be taken to be Lipschitz, enabling direct application of maximal function theory and density arguments. For $C^1$ systems, holonomies are typically only Hölder, but in dimension one, covering and measure-theoretic techniques (such as uncentered maximal functions and adapted foliation charts) suffice to establish the necessary invariance properties. This dimensional restriction is critical: higher-dimensional centers would generally not inherit this rigidity due to weaker control on geometric regularity, unless additional structural properties are imposed.

4. Applications and Consequences in Dynamics

The principle yields several consequences:

• Rigidity of Measures: When the invariance holds, the structure of conditional measures on center leaves is highly constrained, often supporting only finitely many ergodic measures of maximal entropy.
• Classification of Dynamical Types: Systems with vanishing central exponent and $u$-invariant measures can often be classified as "rotation type" (isometric extension of Anosov), while those lacking this property must have center exponents bounded away from zero ("hyperbolic type"). This dichotomy has practical implications for orbit structure and mixing.
• Extension to Lower Regularity: The principle allows extension of results from $C^2$ to $C^1$ diffeomorphisms for partially hyperbolic systems, as established in (Gan et al., 21 Sep 2025), with corresponding persistence of measure-theoretic and entropy properties.

5. Technical Innovations in the $C^1$ Setting

Key technical elements allowing the invariance principle in $C^1$ regularity include:

• Use of "fake foliation charts" at scales determined by hyperbolicity and domination, circumventing the need for classical Lyapunov charts.
• Application of Lebesgue density arguments via the weak $(1,1)$ property of maximal operators in one dimension.
• Control of holonomy regularity via Hölder estimates:

$|^u_x(p) - ^u_x(q)| \leq C_2 |p - q|^\alpha$

where $\alpha \to 1$ as the system approaches higher regularity.

These tools enable proof of invariance and entropy equality even in the absence of the smoothness traditionally required.

6. Broader Connections and Comparative Principles

The Avila–Viana invariance principle fits into a spectrum of invariance results across ergodic theory and probability, standing alongside probabilistic invariance principles for random fields (Wang et al., 2011), Wasserstein convergence refinements (Liu et al., 2022), vector-valued invariance principles (Su, 2019), and operator-theoretic invariance laws (Makarov et al., 2021). While differing in technical context—projective criteria, spectral theory, and martingale approximation vs. geometric and measure-theoretic holonomy invariance—the underlying theme is the emergence of rigidity in the statistical or geometric structure under mild regularity and dynamical hypotheses.

7. Implications and Open Directions

The Avila–Viana invariance principle has direct implications for:

• Uniqueness and classification of measures of maximal entropy
• Exactness and mixing properties in the presence of dominated splitting
• Extension of measurable classification theorems to otherwise unreachable regularity classes ($C^1$)
• Potential applications to random or non-autonomous dynamical systems via connections with approximate invariance and probabilistic limit laws

A plausible implication is that future work may generalize the principle to higher-dimensional centers under additional geometric or dynamical constraints or may extend the scope to noncompact or random settings by leveraging probabilistic approximation theorems.

---

In summary, the Avila–Viana invariance principle constitutes a central rigidity result in modern smooth dynamics, linking entropy theory, conditional measure invariance, and the classification of partially hyperbolic systems, with a robust geometric and measure-theoretic foundation that admits meaningful extension to lower regularity and new dynamical contexts (Gan et al., 21 Sep 2025, Tahzibi et al., 2016, Marin, 2015).