gpt-6.2-sol high: bw-density
Reviewer record, not the publication verdict. The response includes unaccepted proposals and source limits. Missing antecedent proofs are not automatically defects in the paper. The initial request for Lemmas 8.4–8.5 was our inventory error; the journal ends at Lemma 8.3. The curated report records final dispositions.
Completed visible assessment, preserved as a readable check record. It is a model assessment, not a formal certificate. This document contains no private reasoning stream.
Proposition 2.4 is false as quantified, with a counterexample in a flat foliation box; its intended applications supply the missing positivity and nesting. The julienne arguments have repairable local gaps, especially the propagation of terminal intersections and the lower radius bound in Lemma 8.2. The new local justifications are separate from the supplied proof and conditional on its uniform plaque inputs. The implication from the repaired density chain to genuine bisaturation and then to ergodicity checks conditionally. The Kolmogorov conclusion needs a separate Pinsker input. No counterexample to Theorem 5.1 or to either principal theorem was found. The historical external-input cutoff is 2005-10-11 (received date, journal printed p. 489 and packet policy). Later material was not used to close obligations. This is a source-only review; later knowledge in model weights cannot be excluded, and no unsupported external theorem was treated as verified.
D1
Checked statement: Journal Sections 1–2 on the supplied pages: local contraction and cocycle iteration, Lebesgue density sequences, internesting, the two bounded-Jacobian properties, and the Cavalieri estimates and their quantified conclusions.
Returned assessment: claim_false
The false assertion is Proposition 2.4 as stated, not the fiber-volume estimate it proves. For Lemma 1.1, keep the orientation of local-leaf membership in the hypothesis and carry that membership as part of the induction. The one-step bound (9) and forward invariance of genuine local stable plaques give the forward product. For the inverse, the factor at step j is ν̂(p_{-j-1}), so the product is ν̂_{-n}(p)^{-1}. The reason ambient distance can be used in (9) is not simply that intrinsic distance dominates ambient distance: on sufficiently small uniformly C¹ plaques their ratio is uniformly as close to 1 as needed, and there is a uniform strict margin in the derivative bound. Apply the derivative estimate to paths in the local plaque, including a small neighborhood of the endpoints. This uses the genuine strong plaques, not fake ones. The assertion on p. 459 that every Lebesgue density sequence is regular can be checked directly. If the successive volume ratios have infimum zero, select increasing indices n_i with square-root ratios ε_i summable and tending to zero. Choose centered balls of volumes √(m(Y_{n_i})m(Y_{n_i+1})), and let H be the union of these balls with Y_{n_i+1} removed from the respective ball. In those balls the fraction of H tends to 1. In Y_n, terms with n_i<n vanish by nesting, and the fraction contributed by any remaining term is at most ε_i, so m(H:Y_n) tends to zero. The complement contradicts the density-test property. For Lemma 2.1, if a is the regularity constant for Y and k is the shift, then m(Z_{n+1})≥m(Y_{n+k+1})≥a^{2k+1}m(Y_{n-k})≥a^{2k+1}m(Z_n) for n≥k; positivity treats the finitely many earlier ratios. The two complement-density comparisons follow from the same inclusions and bounded ratios of volumes at fixed shifts. The journal's equivalence of (11) and (12) uses the quantified transversal and bounded-Jacobian versions. For (12) to (11), apply (12) on both transversals to the union of plaques over a measurable base. The two integrals compare the two base measures weighted by the same strictly positive plaque volume; comparison for all base subsets gives comparison of the unweighted measures by integrating the reciprocal weight. For the converse, in a small chart use smooth parallel slices transverse to the plaques. On each plaque the slice coordinate has a leaf-volume Jacobian bounded above and below by the cone and metric bounds. Integrate first in that coordinate, use (11) to compare the measure on each slice to the measure on the given transversal, and use ambient Fubini in the slice chart. This checks the local mechanism; it does not establish that the dynamical strong foliations satisfy either property. In Proposition 2.3, integrate the upper and lower fiber bounds. In Proposition 2.7, integrate over the base of the complement: each fiber there is either the whole Y_n fiber or empty, giving the factor C²c in both directions. Proposition 2.5 then replaces X by its saturated representative in ambient volume and uses the common base; it does not assert that an arbitrary null change of X has null intersection with every transversal.
Premises used
- Compactness, strict derivative inequalities, and uniformly C¹ local plaques for the chord estimates; the existence theorem for the genuine plaques is not re-proved here.
- The regularity definition on journal p. 459 requires both nesting and strictly positive measure. The density definition on that page requires a positive denominator.
- The Cavalieri comparisons require one uniform disintegration constant for the box and transversal in use and one fiber-comparability constant independent of rank.
Missing inputs
- Journal pp. 456–457 were not supplied as images. The conventions and initial setup preceding p. 458 were consulted in the October 2006 extraction for context only.
- The genuine strong-foliation bounded-Jacobian theorem, including the pointwise-rate and C¹⁺δ setting, is not supplied; the historical input is recorded under D3.
Scope changes
- The reference to 'Propositions 2.1–2.5' was interpreted using the printed numbering: Lemma 2.1, Corollary 2.2, and Propositions 2.3–2.5. Proposition 2.7 was checked because it proves Proposition 2.5 and is used again in Section 8.
- The equivalence verified is the journal's uniform bounded-Jacobian formulation; the author extraction's sentence about a converse is not substituted for the journal statement.
Located concern
BW journal (2010), printed p. 458, proof of Lemma 1.1
The final factor is printed as ν(p_{k+1}), not the cocycle ν_{k+1}(p). The proof also reverses the orientation of the initial local-plaque membership without needing to do so.
The preceding line multiplies ν(p_k) by ν_k(p). Its product is ν_{k+1}(p). The hypothesis is q∈W^s_loc(q′), which can be carried forward in that same orientation.
Correct the subscript and keep q_j in the local plaque through q′_j throughout the induction.
No change to the stated contraction estimates.
Located concern
BW journal (2010), printed pp. 459 and 462, definition of regularity and Proposition 2.4
The stated hypotheses imply a lower bound on the successive volume ratio when volumes are positive, but do not imply either positivity or nesting.
In a small Euclidean box with the vertical-line foliation and horizontal transversal, take Y_n to be the vertical segment over 0 of length 2^{-n} times a fixed small positive length. The foliation satisfies (12), the base is {0}, the fibers are 1-uniform, (i) reads 0≥δ·0, and (ii) holds with δ=1/2. The sets are nested but have zero ambient volume. Independently, alternating two disjoint positive-width rectangles of equal width and fiber length satisfies (i)–(ii) with δ=1 and fails nesting.
Require a nested sequence of measurable sets and positive base measure at rank zero; (i) propagates positivity, and Proposition 2.3 converts it to positive ambient measure. Alternatively state only the volume inequality that the proof establishes.
This is a false auxiliary statement as printed. The applications use nested families with positive-volume bases; they do not need regularity for arbitrary families satisfying only (i)–(ii).
Located concern
BW journal (2010), printed pp. 460–464, equations (11)–(12) and Propositions 2.3–2.7
The estimates are defined on sufficiently small boxes, but the Cavalieri formulations say 'a foliation box.' Also, c-uniformity alone does not ensure that the densities appearing in Propositions 2.5 and 2.7 are defined.
The defining constants in (11)–(12) come with a diameter threshold. A singleton positive-fiber base, as in the vertical-segment example, satisfies c-uniformity but has zero transversal and ambient measure.
State the propositions for a box and transversal on which (12) holds with one constant, and require positive density denominators. The local applications can be made in sufficiently small boxes, discarding finitely many ranks for limit statements.
The fiber integrations and the limit transfers are valid with these qualifications; there is no counterexample here to the positive-volume limit comparisons.
D2
Checked statement: Propositions 5.3–5.5: rank shifts under genuine stable holonomy, distortion and regularity of genuine and fake unstable fibers and fake cu-juliennes, and the density comparison through the six families in Section 8.
Returned assessment: repairable_gap
On p. 474 the stable-side choice is consistent: ν<γ min{1,γ̂} follows from domination and ν<γγ̂. Choose σ strictly between ν/γ and min{1,γ̂}, then τ strictly between ν and σγ. In particular τ<1 and τν̂<σ, the latter using ν̂<γ̂ and γγ̂<1, not the other bunching inequality. Compactness makes all strict cocycle-ratio bounds uniform. Fake unstable and genuine unstable fibers each nest because one backward step contracts their terminal balls and τ_{n+1}<τ_n; fake cu-juliennes also have nested center bases. Uniform plaque graphs give two-sided leaf-ball volumes comparable to radius to the leaf dimension. Together with backward contraction and the shrinking center radii this gives shrinking, positive-volume cu-juliennes on their fake cu plaques. In Section 6 the hats matter. Lemma 6.1 concerns fake stable holonomy of center plaques in a common fake cs plaque; Lemma 6.2 uses the fake stable relation between q and q′ but genuine stable holonomy between y and y′. The journal already gives q′∈S_{n-k₁} in Lemma 6.2; no same-rank assumption is needed. The terminal product and pullback argument given below fills in the forward-domain and rank estimates. Once the center displacement is O(σ_n) and the terminal unstable displacement is O(τ_n), a single rank shift absorbs both constants. Applying the inclusion with x and x′ exchanged gives the other inclusion because the same genuine local holonomies are inverses. In Section 7, the ingredient behind Lemma 7.1 is the uniform estimate on the sum of powers of orbit separations in the thin tubes. For a uniformly Hölder one-step Jacobian on each orbit chart, bounded away from zero, the difference of the two log products is bounded by a constant times this sum. The same proof works for the genuine unstable Jacobian with the genuine unstable distribution, not by measuring genuine fibers with the fake distribution. After the leafwise change of variables, the two measure ratios compare by the bound on the ratio of the maximum and minimum Jacobians. Terminal balls of the same radius thus give uniform fibers. The one-step inverse has a uniform lower leaf-volume multiplier, and inf τ>0 supplies the consecutive volume lower bound. Since q can be taken in the smaller center base at rank n+1 for both comparisons, Proposition 2.4 applies to the fake cu volume once positivity and nesting are made explicit. This step uses the bounded-Jacobian property of fake u inside fake cu, not genuine unstable absolute continuity. In Section 8, round balls test all radii because successive radii have a positive lower ratio. The center–stable base and C_n are comparable to plaque and ambient balls, respectively, by the two successive local products of uniformly transverse small graphs. C_n and D_n have the same base on fake cs with genuine unstable fibers; replacing X by its essential unstable saturate for the ambient comparison is legitimate. D_n to E_n is a geometric change from genuine to fake unstable fibers, not an assertion of saturation for the fake foliation. E_n to F_n uses Proposition 5.3 with genuine stable holonomy; a common stable base is available on the genuine stable plaque through p. Shift both the julienne rank and the stable base radius. The F_n to G_n metric comparison needs the extra argument below. G_n has genuine stable fibers of volume comparable to σ_n^{dim E^s}; Proposition 5.4 gives regularity of its transverse base. This yields the regular sequence required for the middle internesting transfers. The last Cavalieri comparison uses full genuine stable saturation and the base on the fake cu plaque. No density comparison for arbitrary sets on fake cu transversals follows.
Premises used
- Conditional geometric inputs: uniform local cone graphs, local invariance and rates from Proposition 3.1, genuine local strong plaque estimates, and the thin-neighborhood bounds used in Section 4.
- C² branch: fake stable and unstable C¹ subfoliation inside the corresponding fake weak plaques under ordinary center bunching. C¹⁺δ branch: the analogous conclusion under strong bunching. The fake tangent Hölder estimates are needed in both.
- For the local graph arguments, in coordinates adapted to complementary planes the two graph slopes can be made less than a common η<1 in a box with a margin. Substitution gives the intersection by contraction; solving the two displacement inequalities bounds the lengths to the intersection by a uniform multiple of the initial separation. Within a weak plaque the same argument uses its center and strong directions.
- Genuine Wᵘ and Wˢ bounded-Jacobian disintegration on the uniformly transverse C¹ plaques, and the corrected local versions of the D1 measure lemmas.
Missing inputs
- Journal images for printed pp. 465–470, containing the rest of Proposition 3.1 and the thin-neighborhood proof, were not supplied. The relevant author-version text was consulted, but the construction and full uniform regularity estimates were not certified against the journal.
- No supporting holonomy note is supplied. The two regularity imports and the historical status of the C¹⁺δ import are recorded in D3.
- Section 8 on the supplied journal pages ends with Lemma 8.3. No Lemmas 8.4 or 8.5 appear in the assigned pages.
Scope changes
- Where a comparison is needed only for density limits, tail comparisons are used explicitly; this does not certify every initial rank of a statement made without working-domain qualifications.
- The comparison with the older Pugh–Shub juliennes on printed pp. 475–476 and the Vitali claim in the remark were not used or audited. The journal's own rank-comparison proof was checked instead.
Located concern
BW journal (2010), printed pp. 478–479, Lemmas 6.1–6.2 and proof of Proposition 5.3
The proof suppresses checks that the intersections can be iterated and that estimates at rank n give membership at a lower rank. The terminal bound should not require y_n and y′_n to belong to one fake cs plaque.
The y-to-y′ leg is genuine stable, not fake stable. The argument on p. 478 invokes the forward orbit of z′ without constructing it in the controlled boxes. Also σ_{n-k₁}/σ_n is σ_{-k₁}(p_n), not in general σ_{-k₁}(p). The p′ in the final line on p. 478 is not the center x′ used for the base on p. 479.
New local justification, conditional on the stated plaque estimates: use sup σ<1 for the uniform center shift. Track q,y in the controlled boxes. Track q′ along the fake stable plaque of q, and y′ along the genuine local stable plaque of y: one-step contraction keeps each in an enlarged box with a fixed margin and gives O(ν_j) separation. Thus fake cu(q′_j) contains y′_j by local invariance. At time n intersect fake u(y′_n) with fake cs(q′_n); the local product bounds its displacement by O(d(y′_n,q′n))=O(τ_n), and coherence puts this intersection on fake c(q′n). Pull it back along fake u; contraction keeps it close to y′j and in the boxes, so backward center invariance identifies its initial point z′. Backward center growth gives d(q′,z′)≤Cτ_n/γ_n≤Cσ_n. After k backward steps the unstable displacement divided by τ{n-k} is at most C[sup(τν̂)]^k. The center displacement and the offset σ{n-k₁} from x′ are absorbed by σ{n-k} using sup σ<1. Use x′ in the base inclusion.
This supplies the missing local justification for the internesting in Proposition 5.3, given the regularity and plaque inputs. It does not replace genuine stable holonomy by fake stable holonomy.
Located concern
BW journal (2010), printed pp. 476 and 480–482, Proposition 5.4 and Section 7
There are domain qualifications in the measure argument: Lemma 7.2 permits a null denominator, and Proposition 5.4 uses Ĵᵘ_{n+1}(q) for q∈S_n although the definition initially gives rank n+1 for q∈S_{n+1}. The base-volume display on p. 481 omits the measures on its right-hand side.
X′ may be empty in Lemma 7.2. S_{n+1} is generally a strict subset of S_n. The last displayed quotient before Section 8 has sets rather than volumes on the right.
Require positive denominator in the ratio, or cross-multiply. For the stronger quantifier of Proposition 5.4, explicitly extend the pullback definition one step for q∈S_n: q_n is in a fixed small box, so choose the working scale with room for f of that box and its terminal plaques. One backward contraction places the new terminal plaque in the old one at time n; consequently Ĵᵘ_{n+1}(q)⊂Ĵᵘ_n(q), as needed for Lemma 7.2. For the regularity application alone q∈S_{n+1} suffices. Put induced center volumes on the right of the display.
The required positive-volume comparisons and regularity follow under the local geometric inputs; the published ratio is not literally defined for every permitted X′.
Located concern
BW journal (2010), printed pp. 476, 481–482 and 484–485, scope of Proposition 5.5 and Lemma 8.1
The displayed orbit identity for q̂ needs a localization argument. The rank families also have a working-domain restriction: the proof uses x∈Wˢ(p,1), though the statement uses Wˢ(p), and the radius-σ_n stable base about x need not be contained in Wˢ(p,1) at the first ranks.
Transversality at time n alone bounds the distance to a terminal intersection; it does not say that the intersection is the n-th iterate of the initial one. The unions in the opening display of Lemma 8.1 also have fiber argument x instead of their bound variable q.
New local justification: because q and x stay in the controlled boxes, their fake cs relation propagates. Intersect the alternate unstable plaque through y_n with fake cs(x_n) at time n. The graph bound gives distance O(τ_n) from y_n, since q_n is a point of that cs plaque at distance O(τ_n). Pull back along the alternate unstable plaque, genuinely or locally invariant as appropriate; contraction keeps the pullback close to y_j at every step, and local cs invariance identifies the initial intersection. Both initial unstable distances are O(ν̂_nτ_n)=o(σ_n), uniformly. Because the cs bases contain and are contained in cs balls with radii comparable to σ_n, a fixed shift contains the new base point. The same backward contraction absorbs the terminal constant in the lower-rank unstable fiber. Use q in the union displays. For a fixed interior x, use the tail on which its stable base stays inside Wˢ(p,1); for uniform working domains, build the thin tubes over a slightly larger stable plaque and restrict x to the interior plaque.
This repairs the genuine-to-fake switch without a measure-theoretic assumption on the fake unstable foliation. The local interior statement is sufficient for Theorem 5.1; this audit does not certify a larger-domain reading of Proposition 5.5.
Located concern
BW journal (2010), printed p. 485, Lemma 8.2
The boundary estimate in the F_n–G_n comparison does not follow just from the two O(σ_n) diameters and d(x,x′)=σ_n. The claimed same-rank julienne membership of q′ is not supplied by Proposition 5.3.
The triangle inequality from those bounds gives an upper bound but need not give a positive lower bound. The fiber definition gives q′ in fake cu(x′), and Proposition 5.3 gives membership in a shifted julienne there, not in general at rank n.
New local argument: put r=σ_n and take q in the rank-n julienne at x. The graph intersection estimate for the genuine stable plaque through q and fake cu(x′), with d(x,x′)≤r, gives d(q,q′)+d(x′,q′)=O(r), since d(q,x)=O(r). In coordinates at x project along the fake cu tangent at x to the genuine stable plane there. By uniform tangent continuity, the projections of q−x and q′−x′ are o(r) on these O(r) plaques. For x′ on the boundary of the stable base disk, the projection of x′−x has norm at least c r, by the genuine stable graph bound. Hence the projection of q′−q has norm at least c r−o(r). The local product identifies fiber boundaries with base boundaries; the fiber contains q. Its inner radius is therefore bounded below by a multiple of r and its outer radius bounded above by a multiple of r. Fixed rank shifts absorb both constants on the tail, with base inclusion supplied by nesting. No same-rank membership of q′ is needed.
This repairs the lower bound required for the density transfer, given the uniform plaque inputs. The defect is in the proof of the comparison, not a counterexample to the comparison.
D3
Checked statement: Deduction of genuine stable and unstable saturation of ambient density points, essential bisaturation, ergodicity, and the Kolmogorov property, with the two differentiability branches and historical inputs separated.
Returned assessment: unresolved
The implication to Theorem 5.1 checks conditionally. Let Aˢ be a measurable genuine stable saturate agreeing almost everywhere with A. It remains essentially genuine unstable saturated. Its ambient density points are unchanged by the null modification. Proposition 5.5 applies to Aˢ itself on the fake cu slices, without transferring a null modification to those slices. Genuine stable holonomy between the two fake cu plaques takes the local trace of Aˢ, and its complement, to the corresponding trace on the target. Two-sided bounded Jacobians transfer small complement density. Proposition 5.3 and the regularity in Proposition 5.4 transfer it to the target juliennes by Lemma 2.1; Proposition 5.5 recovers ambient density there. Local stable propagation yields whole-leaf saturation by subdividing stable paths. For f^{-1}, exchange stable and unstable, use ν̂, γ̂, γ, ν at f^{-1}(p) in place of ν, γ, γ̂, ν̂, and exchange the lower-bound functions in the same way. The two-sided bunching conditions and regularity persist. No invariant probability or essential accessibility is needed for Theorem 5.1. The density representative is Borel: in small balls use the continuous dependence of their measures, including the measure of intersection with a fixed Borel set, on center and radius (boundaries have zero volume), and use countably many radii. The Lebesgue density theorem then gives Corollary 5.2. For an invariant probability equivalent to Riemannian volume, forward limsup averages of a continuous test function are exactly stable saturated as functions, by uniform continuity and forward contraction; the backward version is exactly unstable saturated. Birkhoff in the two directions gives the same conditional expectation on the common invariant σ-algebra. Thus corresponding sublevel sets agree modulo null sets, and their intersection is bi-essentially saturated. Corollary 5.2 and essential accessibility force zero or full measure for each rational sublevel; the averages are almost surely constant. Conditional expectations of continuous functions being constant imply triviality of the invariant σ-algebra by L¹ approximation of invariant indicators. The K conclusion is a separate implication: it requires that every Pinsker set have both essential saturates. Ergodicity alone does not supply that input. With that input, Corollary 5.2 and essential accessibility make the Pinsker algebra trivial.
Premises used
- C² branch: ordinary center bunching suffices for the geometric rank inequalities and, conditionally, the fake C¹ strong subfoliations. For the stable argument use ν<γγ̂; for the unstable argument use its inverse analogue.
- C¹⁺δ branch: the author context specifies lower bounds μ, μ̂ on the strong rates in the indicated directions, domination tests ν/γ<μ^θ and ν̂/γ̂<μ̂^θ, and ν^θ,ν̂^θ<γγ̂ for θ∈(0,δ). These stronger conditions are used for the fake regularity, not for a different density proof. They imply ordinary bunching: θ≥1 would force both γ>1 and γ̂>1 by μ<ν and μ̂<ν̂, contrary to γγ̂<1; hence ν≤ν^θ and ν̂≤ν̂^θ.
- Birkhoff pointwise ergodic theorem: for an invertible probability-preserving transformation and an L¹ test function, the forward averages converge almost everywhere to the conditional expectation on the invariant σ-algebra, and the same result for the inverse gives the same limit. available_from: 1931; date_evidence: bibliographic identification of Birkhoff's 'Proof of the ergodic theorem' and the invocation on journal p. 472; date_basis: original publication year, not a supplied copy; version: classical pointwise ergodic theorem, external text not inspected.
- Lebesgue density theorem: for Riemannian volume and a measurable set the small-ball density-one representative agrees up to a null set. available_from: classical pre-cutoff result, exact original version/date not verified from packet; date_evidence: formulation invoked on journal pp. 459 and 472; date_basis: classical attribution, with no dated source text supplied; version: Euclidean density theorem localized in smooth charts.
- Kolmogorov criterion used on journal p. 472: for the invertible measure-preserving system in question, the maximal zero-entropy factor being trivial is the Pinsker formulation of the K property. No extra K criterion from a later source is used.
Missing inputs
- Genuine strong foliations: needed theorem is existence of invariant strong plaques and, for C¹⁺δ partially hyperbolic maps on compact M with pointwise rates, local absolute continuity with two-sided uniformly bounded volume factors, including uniformly transverse C¹ transversals. available_from: BP74 (1974) and PS72 (1972) for versions described on journal p. 461 with constant rates; for the needed pointwise Pesin-theory application, not established from supplied theorem text. date_evidence: journal pp. 461 and 488 and the 2004 [Pes04] bibliographic entry; date_basis: publication years of cited versions, not verification that a given version contains the required formulation; version: exact pre-cutoff Pesin theorem and application not supplied. The date of the cited AV08 version is 2008, after cutoff.
- Fake plaque construction and Hölder estimates: needed are coherent, uniformly controlled plaques for the fiberwise orbit extension of a partially hyperbolic map, with local invariance and uniformly Hölder tangent spaces at C¹⁺δ. available_from: HPS77 (1977), KH95 (1995), PSW97 (1997) and PSW00 correction (2000) are pre-cutoff cited versions; date_evidence: journal bibliography pp. 488–489 and author proof context; date_basis: publication years; version: cited book and corrected journal versions, not supplied. The exact graph-transform and Hölder theorems and the passage to the disjoint, noncompact family with uniform estimates remain conditional.
- C² center-bunched branch: needed is uniform C¹ strong holonomy/subfoliation inside fake cs and fake cu leaves of the extended system, including C¹ holonomy on center transversals for Lemma 6.1. available_from: 1997/2000 for the cited PSW paper and correction, not independently established for this precise application; date_evidence: journal p. 489 and author proof context; date_basis: publication years; version: corrected PSW97 result. Supply the exact theorem and verify the uniform extension and the regularity hypotheses on the weak leaves; the author proof flags that the weak-leaf manifold is not automatically of class C².
- C¹⁺δ strongly bunched branch: needed is the supporting note's strong holonomy regularity with the stated pointwise rate and Hölder hypotheses, applied to the same extended orbit family. available_from: 2006 for the cited [BW06b] version; a matching version available by 2005-10-11 is unestablished. date_evidence: journal p. 488 lists a 2006 preprint; no note or dated earlier version is supplied. date_basis: bibliography year only; version: cited 2006 preprint, theorem numbering and text not inspected. Supply a dated pre-cutoff version containing the result, or a self-contained argument for this branch.
- Pinsker step: needed is that every Pinsker set of the invariant-probability system has separate measurable genuine stable and unstable saturates modulo null sets. available_from: 1974 for the named source; date_evidence: [BP74] item on journal p. 488 and explicit Proposition 5.1 attribution on p. 472; date_basis: original and translation publication year; version: Brin–Pesin Proposition 5.1, source text not supplied. Verify exact regularity, pointwise versus constant rates, and permitted invariant density, or give the required extension to each branch. The journal paraphrase does not establish those hypotheses.
Scope changes
- The journal introduction containing the full principal theorem statements was not supplied as images. The printed deductions on pp. 471–477 are primary; the October 2006 introduction is used for context on the strong-bunching branch only.
- The supporting note's Lemma 1.1 was not audited. The only Lemma 1.1 checked here is the journal's local contraction lemma.
- Retrospective note: the journal's AV08 reference is recorded as post-cutoff, not used to close the historical genuine-foliation input. No later proof of the principal result was used.
Located concern
BW journal (2010), printed pp. 472 and 488, proof of Theorem 0.1 and [BP74]
The quoted Pinsker input is indispensable for K and its scope has not been verified from the primary statement.
The proof permits any invariant probability equivalent to volume and the density theorem includes both smoothness branches. The packet contains the proposition number and paraphrase, not the proposition's exact hypotheses or proof.
Supply the 1974 text and check its application to the invariant measure and regularity in each branch, with a pre-cutoff or self-contained extension if needed.
K remains conditional independently of the Hopf deduction of ergodicity. This is not evidence that the Pinsker assertion is false.
Located concern
BW journal (2010), printed pp. 473 and 488–489, fake regularity interface and [BW06b]
Historical availability of the lower-regularity input is not established by the cited version.
The journal uses uniform C¹ fake subfoliation at p. 473 for both branches and cites a 2006 holonomy-note version in its bibliography. The cutoff is the received date, 2005-10-11, printed on p. 489.
Provide a dated pre-cutoff note containing the needed statement and proof, and verify its hypotheses for fake foliations. The publication's citation date alone neither admits the cited version nor disproves the existence of an earlier one.
The original proof obligation specific to C¹⁺δ remains historically unresolved. The C² branch uses a different regularity input.
Outside the audit
- Full published proofs of Proposition 3.1 and Section 4; relevant journal images were not supplied, and author text was used only as context.
- Exact external theorem texts for graph transforms, bundle and holonomy regularity, pointwise-rate genuine absolute continuity, and Pinsker saturation; no external sources were inspected.
- The separate holonomy note, including its own Lemma 1.1. No note pages are in this packet.
- Lemmas 8.4 and 8.5: no such statements appear in the supplied Section 8.
- Comparisons to Pugh–Shub postjuliennes and Vitali bases, the historical claims in the introduction, and the open questions and examples.
Additional supporting archive items are identified in the original report. View the manuscript hashes and repository snapshot.