{"id":"f7216644-f877-4ad6-830d-f1acf9056c8c","arxiv_id":"2411.11836","paper_version":1,"verdict":"CONDITIONAL","confidence":"LOW","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Non-wandering partially hyperbolic diffeomorphisms with quasi-isometric center on closed 3-manifolds are either skew products over a torus Anosov map or discretized Anosov flows, and volume-preserving ones are ergodic.","lead":"This paper studies chaotic maps on 3-dimensional spaces in which the middle direction is only mildly distorted by iteration. It classifies all such maps into two types, proves ergodicity for volume-preserving cases, and links their existence to transitive Anosov flows on the ambient manifold.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Theorem 7.1 hinges on an unproved assertion that lifted center-stable leaves are properly embedded planes; without this, the center-leaf-space conclusion and the resulting Anosov flow construction do not follow.","rationale":"The reader identified exactly the same weakest link: the unproved proper-embeddedness assertion in the proof of Theorem 7.1. My reading of the paper confirms that this assertion is not derived from the surrounding propositions and is used at a critical juncture. The conclusion that each center leaf meets each stable leaf exactly once is not merely a technical convenience; it is the step that produces a Hausdorff one-dimensional center leaf space inside each lifted center-stable leaf. Without that leaf space, the appeal to [FP23a, Theorem 1.1] for uniform quasi-geodesics, and the later construction of a topological Anosov flow, do not follow. I therefore agree with the reader's conditional verdict. I do not recommend REJECT, because the main theorems may still be recoverable: Theorem D's non-solvable case is proven through Section 6's direct accessibility argument and Theorem 5.1 rather than through Theorem 7.1, so the gap, while serious, is not necessarily fatal to every stated theorem. However, as written, the manuscript does not contain a complete proof of Theorem 7.1, and the alternative proof of accessibility in Section 7.2 inherits the same gap. A precise proof or a counterexample is needed, so the verdict should remain CONDITIONAL; since this matches the reader's existing verdict, I mark the final adjustment as UNCHANGED.","tokens_in":28492,"tokens_out":19280,"duration_ms":208967,"concrete_test":"Verify the following formally weaker claim directly from the stated hypotheses: for every leaf L of F~cs and every center leaf γ and stable leaf ℓ contained in L, the intersection γ∩ℓ has at most one point. Attempt to prove this using only completeness of F~cs (Proposition 3.10) and the quasi-isometric center bounds, without invoking proper embeddedness. If the proof requires the lifted leaves to be closed subsets of R^3, check whether any cited result, especially [HHU20, Proposition 6.7] or [FP23a, Theorem 1.1], actually supplies that properness. A concrete counterexample would be a dense cylinder leaf of a minimal F^cs on a solvable 3-manifold whose lift to R^3 is a non-proper plane; constructing or ruling out such an example settles whether Theorem 7.1 needs an extra hypothesis or a missing lemma.","verdict_should_be":"UNCHANGED","load_bearing_attack":"In the proof of Theorem 7.1 (Section 7.1, paragraph beginning 'It implies that, in any leaf of F~cs...'), the manuscript asserts: 'Note that the leaves of F~cs are properly embedded planes in R^3.' This is not proved anywhere. Proposition 3.10 gives completeness of F~cs, and Corollary 3.11 gives that, up to a double cover, each leaf of F^cs is a cylinder or a plane; neither result implies that the lifted leaves are closed, proper subsets of the universal cover. In a minimal foliation of a closed 3-manifold, leaves may be dense, and the corresponding lifted leaves need not be properly embedded. The assertion is load-bearing: it is used to rule out, via Poincaré-Bendixson, the possibility that a center leaf and a stable leaf inside a common F~cs leaf meet in more than one point. That uniqueness of intersection is what lets the proof conclude that the center leaf space inside each F~cs leaf is R, which is then needed to apply [FP23a, Theorem 1.1] to obtain uniform quasi-geodesics and to construct the expansive flow. If the proper-embeddedness claim fails, the conclusion of Theorem 7.1 and the alternative proof of accessibility in Section 7.2 lose their foundation. Since Theorem 7.1 is presented as an independent route toward the main classification results, this is a genuine gap that must be closed before the argument as written can be accepted.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies C^1 partially hyperbolic diffeomorphisms of closed 3-manifolds whose center foliation is quasi-isometric and whose non-wandering set is the whole manifold. The main announced results are: Theorem A, accessibility and stable ergodicity when the fundamental group is not virtually solvable and NW(f)=M; Theorem B, ergodicity equivalent to transitivity for C^2 conservative quasi-isometric-center diffeomorphisms; Theorem C, existence of a transitive Anosov flow when the fundamental group has exponential growth; and Theorem D, a classification into skew products over Anosov automorphisms on T^2 or discretized Anosov flows. The proof strategy combines central unique integrability and completeness of center-stable and center-unstable foliations (Section 3), existence of periodic compact center leaves and their homoclinic structure (Section 4), a transitive classification (Section 5), an accessibility argument using results from the companion paper [FU24] (Section 6), and a non-wandering classification (Section 7).","tokens_in":28741,"tokens_out":6047,"duration_ms":59858,"significance":"If all claims are completed, this would be a substantial contribution: it extends the Hertz-Hertz-Ures ergodicity conjecture to a broad quasi-isometric-center class, gives a topological Anosov flow from partial hyperbolicity under only an exponential-growth hypothesis, and provides a full classification under the non-wandering condition. The paper has real strengths: the careful comparison of quasi-isometric center foliations versus center bundles, the proof that the lifted invariant foliations are complete (Proposition 3.10), the periodic center leaf results of Section 4, and an accessibility proof in Section 6 that is largely independent of the classification. The author also honestly indicates reliance on the companion paper [FU24] and on the in-preparation independent work [EMP]. However, several load-bearing arguments are left unfinished, in particular the proper-embeddedness claim in the proof of Theorem 7.1 and the virtually solvable case of Theorem B. These gaps are significant enough that the manuscript cannot be accepted in its present form.","major_comments":[{"comment":"The proof contains the unproved assertion that the leaves of the lifted foliation F~cs are properly embedded planes in R^3. Proposition 3.10 proves completeness of the lifted foliations, and Corollary 3.11 proves that projected leaves are cylinders or planes; neither statement implies proper embeddedness of the lifted leaves in the universal cover. The subsequent Poincaré-Bendixson argument uses proper embeddedness to rule out the possibility that a center leaf and a stable leaf inside a common F~cs leaf meet in more than one point. Without uniqueness of intersection, the conclusion that the center leaf space inside each F~cs leaf is homeomorphic to R is not justified, and therefore the application of [FP23a, Theorem 1.1] to obtain uniform quasi-geodesics and the construction of the expansive flow do not follow. Since Theorem 7.1 is used in Section 7.2 as an alternative route to accessibility and is presented as a coarser classification under the same hypotheses as Theorem D, this gap must be closed before the arguments as written can be accepted.","section":"Section 7.1, proof of Theorem 7.1"},{"comment":"In the virtually solvable case, the proof relies on two unsupported statements. First, it asserts that \"The set of all su-tori is compact and f-invariant by [Hae]\"; the cited reference [Hae] is a general reference and the compactness of the union of all su-tori is not an obvious consequence. Second, it asserts without proof that \"By transitivity, the map f|γ has irrational rotation number.\" Transitivity of f on M does not by itself imply transitivity of the induced circle map f|γ unless additional argument is supplied. These statements are load-bearing for the equivalence between ergodicity and transitivity in Theorem B, because they are used to conclude that f|γ is ergodic and that an invariant su-saturated set has full or null measure. A complete proof of these steps is needed.","section":"Section 6.3, proof of Theorem B"},{"comment":"The proof of Proposition 4.8 is only a sketch and contains non-obvious claims that are not justified. In particular, the statement that by the Poincaré-Bendixson theorem both c and c' accumulate on γ when F^cs(γ) ∪ F^cu(γ) contains no other compact center leaves requires an argument, and the claim that sγ(x) ∩ uγ(x^1_u) contains finitely many connected center curves needs a transversality and compactness justification that is not supplied. Proposition 4.8 is used in Proposition 5.5 to conclude that every homoclinic center leaf is periodic and hence that all center leaves in F^cs(γ) are periodic, which is then used to apply Theorem 5.3. Since Theorem 5.1 and therefore Theorem D depend on this chain, a complete proof or a precise reference to a full proof in [BZ20] is required.","section":"Section 4.2, Proposition 4.8"}],"minor_comments":[{"comment":"There are numerous typographical errors, including \"relavent\", \"tirvial\", \"pruduct\", \"clsoed\", \"accessbility\", \"devide\", \"contruct\", and \"explicitely\". A careful proofreading pass is needed.","section":"Throughout"},{"comment":"Definitions 2.1 and 2.2 quantify the quasi-isometric inequalities over n ∈ N, while Lemmas 3.1 and 3.2 claim equivalence with inequalities for all n ∈ Z. The proof of (1) implies (2) is not written out for negative iterates; the argument needs to address why quasi-isometry for f implies the corresponding property for f^{-1}, or the statements and definitions should be adjusted consistently.","section":"Section 2.1 and Section 3.1"},{"comment":"The proof says \"(1) to (2) follows by the choice of constants C = A and D = AB.\" This is not transparent; the relationship between the constants and the passage from positive to all integer iterates should be spelled out.","section":"Section 3.1, proof of Lemma 3.1"},{"comment":"In the virtually solvable case, the statement \"The first case cannot occur since it is homotopic to a uniformly hyperbolic map and thus cannot have quasi-isometric center\" is asserted without proof. Since this exclusion is used in the classification, it should either be proved or supported by a precise reference.","section":"Section 7.3, proof of Theorem D"},{"comment":"The reference [Hae] is incomplete: no title, journal, or year is given. The in-preparation work [EMP] is cited for an independent related result; the manuscript should clarify, to the extent possible, the overlap and the provenance of the shared techniques.","section":"References"}],"recommendation":"major_revision","confidential_remarks":"The manuscript's central claims are interesting and the overall architecture is coherent, but the proof has several load-bearing gaps, especially the proper-embeddedness assertion in Theorem 7.1 and the su-torus case in Theorem B. I would recommend asking the author to provide complete arguments for these points and to clarify the dependence on unpublished or in-preparation references before sending the paper back out. The novelty overlap with [EMP] should also be resolved."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Here's the short version: this is the right paper at the right time for the quasi-isometric center class, and if the main claims hold it gives the first full classification for that class as well as the HHU ergodicity conjecture in the no-su-torus case. It deserves a serious referee. But the manuscript as written has a load-bearing gap in the proof of Theorem 7.1, and a few other places where arguments are too compressed.\n\nWhat is genuinely new: the quasi-isometric center condition is strictly weaker than the topologically neutral center of Bonatti-Zhang, and the paper proves classification under NW(f)=M, accessibility in the absence of su-tori, and the Anosov-flow existence criterion. The author explicitly credits the independent parallel work by Espitia-Martinchich-Potrie; that is honest and the right thing to do. The proofs of central unique integrability (Prop 3.3) and completeness of the lifted foliations (Prop 3.10) are real arguments, not black boxes. There is no circularity: reliance on [FU24] for ideal-boundary machinery is a transparency issue, not a logical loop, because the hypotheses there are different.\n\nThe main problem is Theorem 7.1, Section 7.1. The proof says \"the leaves of F~cs are properly embedded planes in R3\" with no supporting argument. Completeness and Corollary 3.11 give that, up to double cover, each leaf of F~cs is a plane or cylinder; they do not give proper embeddedness of the lifted leaves. In minimal foliations, leaves can fail to be closed or proper. The assertion is doing real work: it is what lets the proof use Poincaré-Bendixson to rule out more than one intersection between a center leaf and a stable leaf in the same F~cs leaf, and that uniqueness is what makes the center leaf space R and triggers [FP23a, Theorem 1.1]. As written, that step is not justified. Maybe it can be proved from the quasi-isometric center plus non-wandering, but the referee has to see the argument.\n\nI would not call this fatal to the whole paper, because Theorem A has an independent proof in Section 6 and Theorem D is derived from Theorem A plus Theorem 5.1, not from Theorem 7.1. But Theorem 7.1 is advertised as its own classification result, and the gap must be closed.\n\nSmaller soft spots: Proposition 4.8 is only sketched, and the sketch has choices (\"we can assume y is contained in...\") that need verification. Theorem B asserts, in the su-torus case, that transitivity forces irrational rotation number on the invariant center circle, and cites [Hae] for compactness of the su-torus set; both are plausible but the manuscript should give precise references and more detail.\n\nBottom line: send to a good referee, but with a clear request to resolve the proper-embeddedness claim and to expand the compressed arguments. If those are fixed, this will be a strong contribution to the classification literature.","headline":"A serious classification paper for quasi-isometric center in dimension 3, with a real and repairable gap in Theorem 7.1; worth refereeing, not ready as written.","tokens_in":29278,"tokens_out":3641,"would_cite":true,"duration_ms":34546,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["37D30","37D20","37A25"],"pacs":[],"model":"deepseek-v4-flash","headline":"A quasi-isometric center direction forces every non-wandering partially hyperbolic diffeomorphism of a closed 3-manifold into one of two rigid model classes.","keywords":["partial hyperbolicity","quasi-isometric center","accessibility","ergodicity","Anosov flow","classification","skew product","discretized Anosov flow"],"falsifier":"Find a non-wandering partially hyperbolic diffeomorphism with quasi-isometric center on a closed 3-manifold with non-virtually solvable fundamental group whose lifted center-stable leaf is not properly embedded (for instance, a leaf accumulating on itself), and show that it is neither conjugate to a skew product over an Anosov automorphism of $\\mathbb{T}^2$ nor admits an iterate that is a discretized Anosov flow. A concrete check would be to exhibit a center leaf in the universal cover intersecting a stable leaf in more than one point.","tokens_in":28253,"feed_emoji":"🌀","tokens_out":8813,"duration_ms":67136,"temperature":0.7,"pith_summary":"Partially hyperbolic diffeomorphisms are maps with a $T^s \\oplus T^c \\oplus T^u$ splitting of the tangent bundle, where the stable and unstable directions are uniformly hyperbolic and the center direction is intermediate. This paper studies those in closed 3-manifolds whose center direction is quasi-isometric: the map stretches or shrinks center curves by at most a fixed affine factor. The central result is that, under the non-wandering condition, such a diffeomorphism is, up to a finite lift and iterate, either a skew product over an Anosov automorphism of the 2-torus or a discretized Anosov flow. From this classification the paper derives accessibility and, for volume-preserving maps, stable ergodicity; it also proves that an exponentially growing fundamental group forces the manifold to admit a transitive Anosov flow. The results settle the Hertz–Hertz–Ures ergodicity conjecture and the Pujals classification program within this class.","feed_headline":"Quasi-isometric center classifies 3D partially hyperbolic maps","feed_subtitle":"One center condition splits 3D maps into two rigid families, yielding ergodicity and Anosov flows.","key_machinery":"The load-bearing object is the quasi-isometric center foliation: an invariant $F^c$ for which there exist $A\\ge 1$, $B>0$ with $A^{-1}L(c)-B\\le L(f^n(c))\\le AL(c)+B$ for every center curve $c$ and all $n\\in\\mathbb{Z}$. This condition is shown to force unique integrability of the center bundle (Proposition 3.3), completeness of the lifted center-stable and center-unstable foliations (Proposition 3.10), and existence of at least one periodic compact center leaf (Proposition 4.4). The classification then applies the Bonatti–Wilkinson dichotomy: a periodic compact center leaf with another compact center leaf in its $F^{cs}$ or $F^{cu}$ leaf yields a skew product, while its absence makes all center leaves periodic, producing a discretized Anosov flow. The homoclinic intersection number (Section 4.2) is the counting device that forces periodicity of all center leaves in the non-skew-product branch.","core_discovery":"The core discovery is the rigidity of the quasi-isometric center condition. In a closed 3-manifold, if $f$ is a $C^1$ partially hyperbolic diffeomorphism with a quasi-isometric center foliation and $NW(f)=M$, then up to a finite lift and iterate $f$ is conjugate to a skew product over an Anosov automorphism on $\\mathbb{T}^2$, or some iterate of $f$ is a discretized Anosov flow (Theorem D). This dichotomy yields the other main theorems: Theorem A states that when $\\pi_1(M)$ is not virtually solvable, $f$ is accessible and, if conservative and $C^r$ with $r>1$, stably ergodic; Theorem B states that for $C^2$ conservative maps ergodicity is equivalent to transitivity; Theorem C states that if $\\pi_1(M)$ has exponential growth then $M$ carries a transitive Anosov flow. The message is that a single geometric condition on the center direction collapses the open zoo of partially hyperbolic examples to two model classes, and the usual obstructions to ergodicity and Anosov flows vanish.","pith_inferences":["The author does not state this, but the proper-embeddedness premise in Theorem 7.1 might be derivable from a quasi-isometry estimate on lifted center curves, which would close the only unproved local step.","Although the paper focuses on Definition 2.1, the author notes that the arguments adapt to Definition 2.2; this suggests the classification is really a theorem about quasi-isometric action on the center bundle, not about the foliation per se.","An implicit test of Theorem B is Nassiri's conjecture: if every transitive $C^{1+}$ partially hyperbolic diffeomorphism with one-dimensional center on a closed 3-manifold is quasi-isometric in the center, then it must be ergodic.","The existence result Theorem C leaves open whether the transitive Anosov flow can always be chosen orbit-equivalent to the center foliation of a lift of $f$; the paper's proof only establishes that some transitive Anosov flow exists on the same manifold."],"forward_implications":["The Hertz–Hertz–Ures ergodicity conjecture holds for this class: conservative $C^r$ maps with $r>1$, quasi-isometric center, and no $su$-torus are accessible and stably ergodic.","Every non-wandering partially hyperbolic diffeomorphism with quasi-isometric center on a closed 3-manifold with non-virtually solvable fundamental group is accessible.","If such a diffeomorphism exists on a manifold whose fundamental group has exponential growth, the manifold admits a transitive Anosov flow.","The classification is complete: no partially hyperbolic diffeomorphism with quasi-isometric center and $NW(f)=M$ can fall outside the two model classes up to finite lift and iterate.","For $C^2$ conservative maps with quasi-isometric center, transitivity and ergodicity are equivalent; any non-ergodic example must be non-transitive, as in the direct product of an Anosov map with the identity on a circle."],"supporting_citations":[{"why":"Supplies the Bonatti–Wilkinson dichotomy (Theorem 5.3) that separates skew products from discretized Anosov flows.","marker":"[BW05]"},{"why":"Provides the homoclinic intersection number machinery (Propositions 4.3, 4.5, 4.8) that forces all center leaves periodic when no extra compact leaves exist.","marker":"[BZ20]"},{"why":"Source of completeness and ideal-boundary arguments (Propositions 6.5, 4.3, 4.11, 4.16) used in the accessibility proof.","marker":"[FU24]"},{"why":"Gives accessibility for discretized Anosov flows and shows how accessibility lifts to finite covers, used in Theorem A and Theorem D.","marker":"[FP22]"},{"why":"Proves accessibility for collapsed Anosov flows with non-wandering and non-solvable fundamental group (Theorem 7.2).","marker":"[FP21]"},{"why":"Handles the virtually solvable fundamental group case in the proof of Theorem D.","marker":"[HP15]"},{"why":"Defines collapsed Anosov flows, shows discretized Anosov flows are dynamically coherent, and gives a characterization of topological Anosov flows (Theorem 5.9).","marker":"[BFP23]"},{"why":"Shows Hausdorff leaf space implies leafwise quasi-geodesic, used to build the topological Anosov flow in Theorem 7.1.","marker":"[FP23a]"},{"why":"Establishes that an $su$-torus forces solvable fundamental group and that 3-manifolds admitting partially hyperbolic diffeomorphisms are irreducible.","marker":"[HHU11]"},{"why":"Supplies cylinder-leaf arguments and the lemma that non-compact center leaves intersect stable leaves infinitely often, used in Lemma 4.2.","marker":"[Zha21]"}],"fun_headline_variants":["Center geometry classifies 3D hyperbolic maps into two families","Quasi-isometric center yields ergodicity and Anosov flows","Rigidity of center direction tames 3D partially hyperbolic maps","One condition on center: two rigid model classes in 3D"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The proof of Theorem 7.1 assumes without proof that the lifted center-stable and center-unstable leaves in the universal cover are properly embedded planes, which is needed to conclude that each center leaf meets each stable leaf exactly once and that the center leaf space is $\\mathbb{R}$.","fun_headline_variants_meta":{"raw":{"variants":["Center geometry classifies 3D hyperbolic maps into two families","Quasi-isometric center yields ergodicity and Anosov flows","Rigidity of center direction tames 3D partially hyperbolic maps","One condition on center: two rigid model classes in 3D"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00026,"raw_usage":{"total_tokens":1578,"prompt_tokens":923,"completion_tokens":655,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":539,"completion_tokens_details":{"reasoning_tokens":579}},"tokens_in":539,"tokens_out":655,"duration_ms":5922,"temperature":1.0,"reasoning_tokens":579,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T18:06:09.757804+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Find a non-wandering partially hyperbolic diffeomorphism with quasi-isometric center on a closed 3-manifold with non-virtually solvable fundamental group whose lifted center-stable leaf is not properly embedded (for instance, a leaf accumulating on itself), and show that it is neither conjugate to a skew product over an Anosov automorphism of $\\mathbb{T}^2$ nor admits an iterate that is a discretized Anosov flow. A concrete check would be to exhibit a center leaf in the universal cover intersecting a stable leaf in more than one point.","supporting_citations":[],"review_version":1}