{"id":"1db09f5f-c995-4d98-be76-8a942e7ebb49","arxiv_id":"1908.09181","paper_version":5,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"For C^{1+α} diffeomorphisms, every hyperbolic ergodic measure's full Lyapunov spectrum can be approximated by the Lyapunov spectrum of a hyperbolic periodic orbit.","lead":"This paper proves that the Lyapunov exponents of a hyperbolic ergodic measure can be approximated arbitrarily closely by the Lyapunov exponents of a single hyperbolic periodic orbit. It extends a classical closing-lemma result from one extremal exponent to the entire spectrum.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Lemma 4.1's exterior-power reduction is not justified: the induced projective-bundle dynamics has vertical Lyapunov exponents that can dominate the exterior-power sum, and the proof is omitted.","rationale":"The reader's weakest assumption is exactly the exterior-power reduction, and I agree: this is the most load-bearing gap. Unlike the main proof of Theorem 3.1, which is largely written out, Lemmas 4.1 and 4.2 are asserted with one-sentence reductions and no details. The issue is not merely cosmetic: f∧i is not a map satisfying Theorem 3.1, and the standard way to make it one—passing to the projectivized bundle—changes the Lyapunov spectrum by introducing vertical exponents. These can dominate the intended exterior sum, so even a careful reader cannot fill the gap without proving a genuinely new cocycle version of Theorem 3.1. The multiplicity issue (coinciding exterior sums) is an additional obstruction to the proof's cone argument. The result may well be true and fixable—the main theorem is plausible and the core argument in Theorem 3.1 is credible—so I would not reject the paper. But the full claim is unsupported by the present manuscript, so CONDITIONAL is appropriate. No change to the reader's verdict.","tokens_in":15222,"tokens_out":8422,"duration_ms":95036,"concrete_test":"Provide the missing proof of Lemma 4.1: specify the compact phase space, the ergodic invariant measure, and verify that its largest Lyapunov exponent equals the top exterior sum. A minimal analytical check: take the linear Anosov automorphism of T^3 with exponents −2, −1, 1 and i=2, and compute the Lyapunov exponents of the induced map on P(∧2 T^3) with respect to the measure supported on the top exterior Oseledets section. If any vertical exponent (e.g., 3) exceeds the target sum 0, the naive reduction fails.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Theorem 1.1 rests on Lemma 4.1, whose proof is 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. We omit the details.\" This replacement is not a direct application: Df∧i is a cocycle over f, not a diffeomorphism of a compact manifold, so Theorem 3.1 does not apply. The natural compactification is the projectivized bundle P(∧iM) with induced map F. To invoke Theorem 3.1 one needs an F-invariant ergodic hyperbolic measure whose largest Lyapunov exponent is the desired sum of the top i Lyapunov exponents. The only evident invariant measure is supported on a measurable section over the top exterior Oseledets subspace, but F acts on vertical directions by the adjoint cocycle, whose Lyapunov exponents are differences ϑ_a − ϑ_b. These can exceed the sum: for exponents −5 and 1 and i=2, the target sum is −4 while the vertical exponent is 6. Applying Theorem 3.1 to F would approximate the wrong quantity. Moreover, exterior sums can coincide (e.g., ϑ_1+ϑ_4=ϑ_2+ϑ_3), so the largest exterior exponent has multiplicity greater than one and the gap parameter q = min_{i≠j}|ϑ_i−ϑ_j| used in the proof of Theorem 3.1 is zero; the cone construction (3.3.1)-(3.3.10) needs strict separation. Thus Lemma 4.1, and with it Theorem 1.1, is not established as written.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","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.","tokens_in":15554,"tokens_out":5959,"duration_ms":64631,"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":[{"comment":"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":"Section 4, Lemma 4.1"},{"comment":"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.","section":"Section 4, Lemma 4.1"}],"minor_comments":[{"comment":"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":"Section 4, proof of Theorem 1.1"},{"comment":"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":"Section 3, Theorem 3.2"},{"comment":"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.","section":"Section 4, Lemma 4.2"},{"comment":"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.","section":"Throughout"}],"recommendation":"major_revision","confidential_remarks":"The paper contains a detailed proof of the extremal Lyapunov exponent approximation, but the proof of the main theorem is incomplete in a load-bearing way. The exterior-power reduction in Section 4 is not a routine application of Theorem 3.1, and the standard route via the projectivized bundle appears to approximate the wrong quantity. I would encourage the authors to provide a complete proof of Lemmas 4.1 and 4.2, or to prove Theorem 1.1 by a different method; until then, the main result should not be regarded as established. The manuscript also appears to be an early arXiv version, and the authors should ensure the submitted version is fully polished."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Worth your time. The paper proves a natural strengthening of Katok's closing lemma: every individual Lyapunov exponent of a C^{1+α} hyperbolic ergodic measure is approximated by the exponents of a hyperbolic periodic orbit. If the theorem stands, it goes beyond Katok-Mendoza and Barreira-Pesin, which only handle the smallest absolute exponent from above. Section 3 is genuine work, not a routine transcription. The proof of Theorem 3.1 uses three adapted Lyapunov metrics, a cone argument, and Katok shadowing on finite Pesin blocks, and the estimates are written out in detail. They look plausible and are the strongest part of the paper.\n\nThe soft spot is the reduction. Theorem 1.1 is assembled from Theorem 3.1, Theorem 3.2, and Lemmas 4.1 and 4.2, and the lemmas have essentially no proofs. Lemma 4.1 says to apply Theorem 3.1 to the i-th exterior power. That is not a direct application: Df∧i is a cocycle over f, not a diffeomorphism of a compact manifold, and Theorem 3.1 is stated for the latter. The natural compactification is the projectivized bundle, but the vertical directions there have Lyapunov exponents equal to differences of the original exponents, and those can dominate the exterior sum being approximated. For exponents -5 and 1 in dimension 2, the target two-wedge sum is -4 while the vertical exponent is 6, so applying Theorem 3.1 to the projectivized map would approximate the wrong quantity. Moreover, exterior sums can coincide, so the strict gap q used in the proof of Theorem 3.1 can be zero, breaking the cone separation. The one-sentence 'we omit the details' does not cover this. I see no circularity and no data-fitting, and the use of Oseledec's theorem and Katok's closing lemma is standard. But the full theorem is conditional on a nontrivial technical step that is not proven.\n\nThis paper is for people working on closing lemmas and periodic-orbit approximations in nonuniform hyperbolicity. It deserves a serious referee. The main idea is credible, and the largest-exponent argument may be reusable even if the full claim needs repair. The referee should press hard on Lemmas 4.1 and 4.2 and ask for a rigorous proof of the exterior-power reduction or a different route to the full spectrum. I would not cite the full theorem yet.","headline":"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.","tokens_in":16071,"tokens_out":3834,"would_cite":false,"duration_ms":40206,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["37D25","37C40"],"pacs":[],"model":"deepseek-v4-flash","headline":"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.","keywords":["Lyapunov exponents","hyperbolic ergodic measure","hyperbolic periodic orbits","Pesin set","closing lemma","Oseledec splitting","nonuniform hyperbolicity","exterior power"],"falsifier":"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.","tokens_in":15018,"feed_emoji":"🔁","tokens_out":7321,"duration_ms":70464,"temperature":0.7,"pith_summary":"This paper proves that the full Lyapunov spectrum of an ergodic hyperbolic measure can be replicated, to any desired accuracy, by the Lyapunov exponents of a single hyperbolic periodic orbit. Lyapunov exponents are the asymptotic growth rates of tangent vectors under iteration; for a periodic orbit they are simply the logarithms of the moduli of the eigenvalues of one return derivative. The authors show that the ordered list of the measure's exponents can be matched pointwise by such periodic-orbit exponents. This matters because it says that the statistical, non-periodic behavior of a nonuniformly hyperbolic system leaves a periodic fingerprint, and it strengthens the classical closing lemma from a statement about individual recurrent points to one about the whole spectrum.","feed_headline":"Periodic orbits reproduce every Lyapunov exponent","feed_subtitle":"For any hyperbolic ergodic measure, the whole spectrum can be matched within any tolerance by one periodic orbit","key_machinery":"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.","core_discovery":"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 $\\gamma$.","pith_inferences":["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."],"forward_implications":["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."],"supporting_citations":[{"why":"Supplies the closing lemma that produces a hyperbolic periodic orbit shadowing a recurrent segment of a Pesin set; the proof's Step 1 re-derives this with estimates adapted to the individual subbundles.","marker":"[5]"},{"why":"Gives the Oseledec theorem, which supplies the Lyapunov exponents and the invariant splitting that are the object being approximated.","marker":"[8]"},{"why":"Supplies the criterion for uniform hyperbolicity used to verify that the periodic orbit found by shadowing is hyperbolic under the Lyapunov metric.","marker":"[4]"},{"why":"Supplies the shadowing lemma and its convenient exponential closing-lemma form that the paper uses to get the sharp tracking estimates.","marker":"[12]"},{"why":"Provides the foundational construction of Pesin sets and Lyapunov metrics, which the proof adapts into its three individual metrics.","marker":"[9]"},{"why":"Records the known approximation of the smallest absolute Lyapunov exponent from above, the result that Theorem 1.1 generalizes.","marker":"[6]"}],"fun_headline_variants":["One periodic orbit approximates all Lyapunov exponents of any hyperbolic measure","Hyperbolic measure's full Lyapunov spectrum approximated by one periodic orbit","Approximating every Lyapunov exponent by a hyperbolic periodic orbit","Single periodic orbit approximates any hyperbolic measure's Lyapunov spectrum"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"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.","fun_headline_variants_meta":{"raw":{"variants":["One periodic orbit approximates all Lyapunov exponents of any hyperbolic measure","Hyperbolic measure's full Lyapunov spectrum approximated by one periodic orbit","Approximating every Lyapunov exponent by a hyperbolic periodic orbit","Single periodic orbit approximates any hyperbolic measure's Lyapunov spectrum"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001563,"raw_usage":{"total_tokens":6177,"prompt_tokens":814,"completion_tokens":5363,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":430,"completion_tokens_details":{"reasoning_tokens":5286}},"tokens_in":430,"tokens_out":5363,"duration_ms":34473,"temperature":1.0,"reasoning_tokens":5286,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T11:20:12.850052+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"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.","supporting_citations":[],"review_version":1}