REVIEW 2 major objections 4 minor 2 references
Lyapunov exponents of hyperbolic measures and hyperbolic periodic orbits
T0 review · 2 major / 4 minor · reviewed 2026-08-14 · deepseek-v4-flash
Pith's one-line read The paper proves that for any C^{1+α} diffeomorphism preserving an ergodic hyperbolic measure, the full ordered Lyapunov spectrum can be approximated arbitrarily closely by the Lyapunov exponents of a single hyperbolic periodic orbit.
desk verdict Full-spectrum periodic-orbit approximation of Lyapunov exponents is a significant target, and the largest-exponent proof has real content, but the exterior-power step that carries the whole theorem is not justified as written. read the letter →
The pith
A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.
The reading
What carries the argument
The load-bearing object is the Pesin set: the full-measure set of points whose tangent space splits into invariant subbundles on which the derivative has prescribed exponential rates. On this set the proof installs three weighted Lyapunov metrics, norms built from infinite sums that make the derivative's action on each subbundle contract or expand at rates arbitrarily close to the individual Lyapunov exponents; a fourth and fifth metric extend these to neighborhoods. Katok's closing lemma (in the exponential-shadowing form from the shadowing lemma) turns a long recurrent segment in a Pesin block into a hyperbolic periodic orbit, and the Hirsch-Pugh criterion certifies uniform hyperbolicity of that orbit from the block estimates. For the induction, the exterior power $Df^{\wedge i}$ on the bundle of $i$-vectors converts the sum of the $i$ largest (or smallest) exponents of the original system into the largest (or smallest) exponent of the induced action, so the single-exponent theorem can be reapplied to those sums.
What would settle it
Look for an ergodic hyperbolic measure whose Lyapunov spectrum, compared with the closure of the set of Lyapunov spectra of hyperbolic periodic orbits, has a strictly positive gap at some index; Theorem 1.1 says no such measure exists. A concrete starting point is to test the exterior-power step on a low-dimensional example where two different sums of Lyapunov exponents are equal, since Lemma 4.1 requires a gap parameter that such an example would lack.
Extended reading notes
Core claim
The central claim is Theorem 1.1: if $f$ is a $C^{1+\alpha}$ diffeomorphism of a compact $d$-dimensional manifold preserving an ergodic hyperbolic measure $m$ with Lyapunov exponents $\lambda_1 \leq \cdots \leq \lambda_r < 0 < \lambda_{r+1} \leq \cdots \leq \lambda_d$, then for every $\gamma > 0$ there is a hyperbolic periodic point $z$ whose ordered Lyapunov exponents satisfy $|\lambda_i - \lambda_i^z| < \gamma$ for all $i$. The proof first establishes the analogous statement for the largest and smallest exponents, then uses exterior powers of the derivative to approximate sums of the largest or smallest $i$ exponents, and finally inducts over $i$ to recover each individual exponent. The method blends Katok's closing lemma with Pesin set theory, but with a Pesin set built from all individual stable and unstable subbundles and with three distinct Lyapunov metrics chosen so that the norm of the derivative along the shadowing orbit is controlled from both sides by the corresponding exponents. Once a periodic orbit shadows a long recurrent segment of the measure's orbit, Hölder continuity of $Df$ lets the authors transfer these norm estimates from the shadowed segment to the periodic orbit and conclude the exponents match within $\gamma$.
Load-bearing premise
The proof's key step assumes that the exterior-power argument in Lemma 4.1 is rigorous: the induced map on the bundle of $i$-vectors is treated as if it were a diffeomorphism of a compact manifold, and it assumes the required gap between the relevant sums of Lyapunov exponents exists, which need not hold when two such sums coincide.
Editorial extensions
If this is right
- For any hyperbolic ergodic measure, every Lyapunov exponent lies within any prescribed tolerance of the corresponding ordered exponent of some hyperbolic periodic orbit; hence the full spectrum is in the closure of the set of periodic-orbit spectra.
- Once the tolerance is smaller than half the gap between the negative and positive exponents, the approximating periodic orbit has exactly the same number of stable and unstable directions as the measure.
- The earlier result that the smallest absolute Lyapunov exponent can be approximated from above by a periodic orbit is a special case, and the new theorem strengthens it to all exponents individually.
- The proof's estimates are exactly the ones needed for a strong closing lemma that would approximate a recurrent orbit in a Pesin set together with its Oseledec splitting; the authors point to this as the next target.
Reading between the lines
- The exterior-power trick suggests the same approximation should work for sums of any fixed set of exponents corresponding to an invariant subbundle, not just the first or last $i$ exponents; this would say the whole 'exponent profile' of the measure is realizable by periodic orbits.
- Since the proof relies on Hölder continuity of the derivative and on exponential shadowing, a natural test is whether the theorem survives for $C^1$ diffeomorphisms; if it fails there, $C^{1+\alpha}$ is the right regularity class for this closing phenomenon.
- A quantitative version of the argument would relate the period of the approximating periodic orbit to the recurrence time in a Pesin block, potentially giving growth estimates for periodic orbits with prescribed Lyapunov spectra.
Signed reviews
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The paper claims that for a C^{1+α} diffeomorphism of a compact manifold, the Lyapunov exponents of any ergodic hyperbolic measure can be approximated arbitrarily well by the Lyapunov exponents of a hyperbolic periodic orbit (Theorem 1.1). Section 3 proves the largest-exponent case (Theorem 3.1) in detail, using Pesin sets, three adapted Lyapunov metrics, Katok's shadowing/closing lemma, and a cone argument; Theorem 3.2 treats the smallest-exponent case by symmetry with f^{-1}, with the proof omitted. Section 4 attempts to pass from largest/smallest exponents to the full spectrum by applying the Section 3 argument to exterior powers of the derivative cocycle (Lemmas 4.1 and 4.2, both with proofs omitted), then deduces Theorem 1.1 by combining the largest and smallest accumulated sums. The central claim therefore rests on the two exterior-power lemmas, whose proofs are not supplied.
Significance. If Theorem 1.1 were established, it would be a natural and substantial strengthening of Katok's closing lemma, going beyond the known approximation of the extremal absolute Lyapunov exponent and beyond the periodic-measure density results of Hirayama and Liang-Liu-Sun. The proof of Theorem 3.1 is detailed, uses standard machinery (Oseledec theorem, Pesin theory, shadowing), and does not appear to assume the conclusion or fit parameters to data; those are genuine strengths. However, the final step from extremal exponents to the full spectrum is not proven as written, and the stated reduction to Theorem 3.1 is, on its face, incorrect. The significance of the paper is therefore conditional on a rigorous exterior-power argument being supplied.
major comments (2)
- [Section 4, Lemma 4.1] The proof of Lemma 4.1 is omitted; the sentence "By replacing f by f∧i in the proof of Theorem 3.1, and using the Pesin set ∧i, one can prove the Lemma" is not a valid application of Theorem 3.1. The object f∧i is a bundle endomorphism over f, not a diffeomorphism of a compact manifold, and Theorem 3.1 requires a diffeomorphism of a compact manifold. The natural compactification, the induced map on the projectivized bundle P(∧i M), has vertical Lyapunov exponents equal to differences ϑ_a − ϑ_b, which can be larger than the exterior-power sum being approximated. For example, with Lyapunov exponents −5 and 1 and i=2, the target top exterior sum is −4, while the vertical exponent is 6; applying Theorem 3.1 to the induced projective map would approximate the wrong quantity. Thus Lemma 4.1 is not established as written, and Theorem 1.1, which depends on it, is not established.
- [Section 4, Lemma 4.1] Even if one attempted to prove Lemma 4.1 by a direct exterior-power analogue of Theorem 3.1, the proof would face a second obstruction: the Lyapunov spectrum of (m, Df∧i) can have coincident values whenever two distinct i-element subsets of {ϑ_1,...,ϑ_d} have equal sums. The proof of Theorem 3.1 defines q = min_{i≠j} |λ_i − λ_j| and requires q >> ε for the Pesin set construction and the cone estimates (3.1.1)-(3.3.10). For exterior powers the analogous gap is zero in such cases, so the strict separation used in the proof of Theorem 3.1 is unavailable. The manuscript provides no substitute argument for handling multiplicities in the exterior-power spectrum.
minor comments (4)
- [Section 4, proof of Theorem 1.1] In the displayed approximation of sums, the term ∑_{j=d-i+1}^d λ^z_d should read ∑_{j=d-i+1}^d λ^z_j; otherwise the inequality is not the claimed approximation of the periodic orbit's exponents.
- [Section 3, Theorem 3.2] Theorem 3.2 is stated without proof, with only the sentence "We omit the details." The f^{-1} symmetry is plausible and likely acceptable, but since this is a load-bearing statement for the final theorem, the authors should either give the short argument or explicitly state that it is a verbatim repetition of the proof of Theorem 3.1 with f replaced by f^{-1}.
- [Section 4, Lemma 4.2] The displayed definition of the set ˜∧i_k contains the typo ∑_{i}^{k=1} instead of ∑_{k=1}^i in the exponents; as printed, the notation is nonsensical.
- [Throughout] There are several typographical errors, including "Thmeorem 3.1" at the end of Section 3, "Lypunov" in the keywords, and inconsistent capitalization of the exterior-power bundle (Λ^i(M) versus ∧i(M)); these should be corrected.
Circularity Check
No circularity: the approximation theorem derives from Katok's closing lemma/shadowing and standard Oseledec-Pesin theory; no fitted quantity is renamed as a prediction.
full rationale
The paper's derivation is self-contained against standard external results. Theorem 3.1 proves the largest Lyapunov exponent of a hyperbolic ergodic measure is approximated by Lyapunov exponents of hyperbolic periodic orbits via Pesin sets, Lyapunov metrics, Katok's shadowing/closing lemma, and the Hirsch-Pugh hyperbolicity criterion. Theorem 3.2 is the time-reversal analogue. Lemmas 4.1 and 4.2 extend this to exterior powers, and Theorem 1.1 reconstructs the full spectrum from extremal partial sums; no step assumes the target conclusion. The only author self-citation is reference [7] (Liang-Liu-Sun), used to contextualize a known measure-approximation result and explicitly distinguished from the exponent approximation proved here; it is not load-bearing. The omitted proofs of Theorem 3.2 and Lemmas 4.1-4.2 and the claimed replacement of f by its exterior power may be a rigor gap, since the exterior-power cocycle is not literally a diffeomorphism of a compact manifold as required by Theorem 3.1, but that is an issue of proof completeness, not circularity: the target statements are not equivalent by construction to the hypotheses, and no parameter is fitted and then reported as a prediction. Score 0.
Assumptions & free parameters
assumptions (6)
- standard math Oseledec multiplicative ergodic theorem provides Lyapunov exponents and a measurable invariant splitting.
- domain assumption Pesin set theory and Lyapunov metrics, including two-sided estimates and extension to local neighborhoods.
- domain assumption Katok's shadowing lemma and closing lemma for C^{1+α} diffeomorphisms.
- standard math Hirsch-Pugh criterion for uniform hyperbolicity from bounded block matrix entries.
- standard math Poincare recurrence theorem applied to a positive-measure subset of a Pesin block.
- domain assumption C^{1+α} regularity of the diffeomorphism f.
Cite this review
Pith. "Pith review of Lyapunov exponents of hyperbolic measures and hyperbolic periodic orbits." pith.science (2026). https://pith.science/paper/WMP73HIJ
@misc{pith2026190809181,
author = {Pith},
title = {Pith review of: Lyapunov exponents of hyperbolic measures and hyperbolic periodic orbits},
year = {2026},
howpublished = {\url{https://pith.science/paper/WMP73HIJ}},
note = {Machine review of arXiv:1908.09181}
}
read the original abstract
Lyapunov exponents of a hyperbolic ergodic measure are approximated by Lyapunov exponents of hyperbolic atomic measures on periodic orbits.
Reference graph
Works this paper leans on
-
[1]
Lyapunov exponents of hyperbolic measures and hyperbolic periodic orbits
arXiv:1908.09181v5 [math.DS] 1 May 2020 Lyapunov exponents of hyperbolic measures and hyperbolic periodic orbits Zhenqi Wang LMAM, School of Mathematical Sciences, Peking University, Beijing 100871, China E-mail: wangzq@pku.org.cn. Wenxiang Sun ∗ LMAM, School of Mathematical Sciences, Peking University, Beijing 100871, China E-mail: sunwx@math.pku.edu.cn....
work page Pith review arXiv 1908
-
[2]
We set ˜Λ i = ∪k≥1 ˜Λ i k and call it a Pesin set. Clearly m(˜Λ i) = 1 . By replacing f −1 by f −Λ i in the proof of Theorem 3.2 and by using the Pesin set ˜Λ i one can prove Lemma 4.2. We omit the details. Proof of Theorem 1.1 We rewrite the Lyapunov spectrum {λ1, · · · , λt} as ϑ1 ≤ · · · ≤ ϑd. We use the notations in the proofs of Theorems3.1- 3.2 and ...
work page 1980
Reviewed August 14, 2026 · model on record in the stance chip above.
Discussion (0). Continue with ORCID to comment.