{"id":"a7250e3f-dcea-4d4d-9ea9-58a0fc4d57a4","arxiv_id":"2411.10339","paper_version":1,"verdict":"ACCEPT","confidence":"MODERATE","novelty_score":8.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Cantat and Dujardin prove that loxodromic automorphisms of C^2 have no smooth or rectifiable Julia slices, preserve no invariant real-analytic foliation, and are rigid under real-analytic conjugacy when a saddle point has non-real eigenvalues.","lead":"This mathematics paper proves several rigidity theorems for polynomial automorphisms of C^2, a class of chaotic dynamical systems in two complex dimensions. It shows that their Julia sets cannot be smooth curves or foliated surfaces, and that real-analytic equivalence forces polynomial equivalence under a mild eigenvalue condition.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Proposition 2.5 Step 3 appears to skip the case where all fixed points of f^3 are saddles; the Bedford-Kim contradiction is only derived when a non-saddle fixed point exists.","rationale":"The reader correctly identified the dependence on Bedford-Kim [4, Prop. 5.1 and 6.1] as the most delicate external input. My stress-test agrees that this is the pivotal point, but it locates the risk more precisely: not in the possibility that [4] is erroneous, but in the way Proposition 2.5 invokes it. In Step 3, the only case that is actually shown to satisfy the 'd−1 saddles with unstable multiplier +d' hypothesis is the case where a non-saddle fixed point exists. The all-saddles case is not handled, and Lemma 2.4 allows both +d and −d unstable multipliers. Without an additional argument ruling out mixed signs, the appeal to [4, Prop. 5.1] does not cover the full set of configurations that can arise. This is not a manufactured objection: the text explicitly conditions the key multiplier conclusion on the existence of a non-saddle point and then draws the unconditional-looking conclusion. Corollary 2.6 depends on Proposition 2.5, and Theorem 2.7 depends on Corollary 2.6, so the gap affects the central claim. I do not recommend rejection, because the missing case may be closed by a short counting or index argument, or by replacing f^3 with a suitable iterate while controlling non-saddle periodic points; hence CONDITIONAL is the appropriate verdict rather than REJECT or UNCHANGED.","tokens_in":43157,"tokens_out":16495,"duration_ms":160269,"concrete_test":"Re-derive Proposition 2.5 Step 3 in the case where all d fixed points of f^3 are saddles. Write their unstable multipliers as ε_i d, ε_i ∈ {±1}. Check, using the fixed-point index identity and the coefficient equations for a composition of Hénon maps, whether a mixed-sign configuration can occur with |Jac(f)| < 1. If an index/Newton-sum argument forces at least d−1 signs equal (or forces all signs to become + after passing to a suitable even iterate without creating new non-saddle fixed points), the gap closes; if a degree-3 or degree-4 dissipative example with mixed signs exists, then Proposition 2.5 and Theorem 2.7 require a new argument.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The proof of Theorem C (Theorem 2.7) funnels through Proposition 2.5, whose Step 3 contains a logical branch that is not justified. After replacing f by f^3, the proof shows at most one non-saddle fixed point. It then argues: if such a non-saddle exists, every saddle fixed point has unstable multiplier +d, hence at least d−1 fixed saddles share multiplier +d, contradicting [4, Prop. 5.1]. But if no non-saddle fixed point exists, this inference is unavailable: all d fixed points are saddles, yet Lemma 2.4 only gives unstable multiplier ±d for each, and there is no reason given that d−1 of them must equal +d. The text writes 'In particular f has d distinct fixed points and at least d−1 of them are saddles with unstable multiplier equal to d' as if it followed unconditionally, but the preceding conditional was 'if such a non saddle point exists'. Since Proposition 2.5 is used to prove Corollary 2.6, and Corollary 2.6 is the contradiction completing Theorem 2.7, this unhandled mixed-sign case is load-bearing. The gap is not that Bedford-Kim is wrong; it is that the written argument does not show the Bedford-Kim hypotheses are met in the all-saddles branch. If mixed signs ±d are impossible for a dissipative unstably linear Hénon composition, an unstated sign/index argument is needed; if they are possible, the contradiction does not fire.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper establishes several rigidity results for loxodromic polynomial automorphisms of C^2. Theorem C states that no unstably linear loxodromic automorphism exists; this is the key input for Theorem B, which rules out global invariant real-analytic foliations with complex leaves (and hence holomorphic foliations), and for Theorem A, which says that a real-analytic conjugacy between two loxodromic automorphisms, under a non-real multiplier assumption at a saddle point, implies polynomial conjugacy up to complex conjugation. The paper also proves Theorem D, which states that under certain hypotheses (hyperbolicity with distinct Lyapunov exponents, or a saddle with Lyapunov exponent exceeding that of the maximal entropy measure), the unstable and stable multipliers cannot all lie in a fixed number field. The proofs combine Bedford-Smillie theory, quasi-expansion, Bedford-Kim multiplier rigidity, Pesin theory, equidistribution of periodic points, and arithmetic specialization.","tokens_in":43450,"tokens_out":24368,"duration_ms":203093,"significance":"These are substantial results in higher-dimensional holomorphic dynamics. Theorem C resolves a question analogous to Fatou's one-dimensional classification in the setting of Hénon maps, and it powers the foliation and conjugacy rigidity theorems. The proofs are detailed and draw on a broad range of techniques. The paper also contributes a useful reinforcement of the equidistribution theorem (Appendix A) and several independent results (e.g., Theorem 6.1 on homoclinic shadowing). The dependence on the Bedford-Kim multiplier theorems is explicit and appropriately cited. Given the depth and breadth, the paper is likely to have a lasting impact on the field.","major_comments":[{"comment":"In Step 3 of Proposition 2.5, the proof asserts that if a non-saddle fixed point exists, then 'f has d distinct fixed points and at least d−1 of them are saddles with unstable multiplier equal to d'. This assertion requires that the d fixed points (counted with multiplicity for a Hénon composition) are distinct. A non-saddle fixed point with eigenvalue 1 can be multiple; for example, a dissipative Hénon map f(z,w) = (aw + z^2 + c, z) has a double fixed point when (1−a)^2 = 4c, and that point has eigenvalue 1. The text does not justify why such a multiple non-saddle fixed point cannot occur here, and without distinctness the Bedford-Kim theorem [4, Prop. 5.1] (which concerns d−1 saddles among d fixed points) may not apply. Please provide a proof of distinctness or state the exact form of [4, Prop. 5.1] and show its hypotheses are met.","section":"§2.3, Proposition 2.5, Step 3"},{"comment":"Proposition 2.5 concludes that every periodic point is a saddle. The proof, however, only performs the fixed-point argument for f^3 and derives a contradiction from the existence of a non-saddle fixed point. A non-saddle periodic point of period n>1 would be a non-saddle fixed point of f^n, but the proof does not explicitly apply the argument to all iterates f^{3n}. Since Corollary 2.6 uses the full conclusion to contradict the existence of an attracting point (which could have period >1), this iteration step should be stated and justified.","section":"§2.3, Proposition 2.5, conclusion"},{"comment":"In the proof of Proposition 5.2, the sentence 'Since f is not unstably real, Proposition 2.2 shows that the local F saturation of J^+_τ is R-Zariski dense' is too quick. Proposition 2.2 gives that J^+_Γ is not contained in any C^1 curve for any transversal Γ; it does not directly imply that the F-saturation is not contained in any proper real-analytic subvariety of C^2. The argument should be expanded to explain why a proper real-analytic subvariety containing the saturation would force J^+_τ to lie in a C^1 curve (for instance via Levi-flatness), or a different justification for the R-Zariski density should be given.","section":"§5.1, Proposition 5.2"}],"minor_comments":[{"comment":"The reference to 'Lemma 2.4.(3)' for the multiplier assertion should be 'Lemma 2.4.(2)', since item (3) concerns unstable connectedness, not the value of λ^u(p).","section":"Page 9, Step 3 of Proposition 2.5"},{"comment":"The reference to 'Lemma 2.4.(4)' should be 'Lemma 2.4.(3)', as Lemma 2.4 has only three numbered items.","section":"Page 10, Corollary 2.6"},{"comment":"There is a duplicated word in 'then the the limit pψ' which should be 'then the limit pψ'.","section":"Page 12, Lemma 2.13 proof"},{"comment":"The phrase 'a subset of the place R^2' should read 'a subset of the plane R^2'.","section":"Page 18, §3.4.1"},{"comment":"The passage about R-Zariski density is terse and should be expanded for readability, in addition to the technical point raised in the major comments.","section":"Page 24, Proposition 5.2 proof"},{"comment":"The word 'migth' should be 'might'.","section":"Page 36, Theorem 8.3 proof"},{"comment":"The word 'convergene' should be 'convergence'.","section":"Page 41, Appendix A"},{"comment":"The name 'Xenelkis de Hénon' in the reference to Question 31 in [62] appears to be a typo or misattribution and should be corrected.","section":"Page 2, Introduction"},{"comment":"The word 'sadlle' should be 'saddle'.","section":"§2.6, Question 2.16"}],"recommendation":"major_revision","confidential_remarks":"The paper is dense and relies heavily on [4], [12], and the authors' prior works. The main external dependency is the Bedford-Kim multiplier theorem; if that theorem has unstated hypotheses, Theorems A-C would be affected. I did not find internal inconsistencies beyond the gaps in Proposition 2.5 noted above and the terse passage in Proposition 5.2. The paper is suitable for a top journal if the Proposition 2.5 issues are resolved."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Read this one. The theorems are genuinely new and the overall architecture is convincing, but there is a hole in the proof of Proposition 2.5 (Step 3) that needs patching before Theorem C is airtight.\n\nWhat is good: Theorem A extends Friedland-Milnor to real-analytic conjugacies; Theorem B goes beyond Brunella's algebraic case; Theorem C and the rectifiability results are the kind of clean statements the field wants. The paper does a lot of work carefully: the normalization, the quasi-expansion lemmas, and the use of Bedford-Kim, [12], [2], and the authors' own prior results. The reliance on external results is explicit and appropriate. The writing is unusually honest about sketches and open questions.\n\nThe soft spot: the stress-test note about Proposition 2.5 Step 3 lands. After replacing f by f^3, the proof shows at most one non-saddle fixed point. It then argues: if such a non-saddle exists, every saddle fixed point has unstable multiplier +d, so d-1 saddles share multiplier +d, contradicting Bedford-Kim. But if all fixed points are saddles, the conditional conclusion does not apply. Lemma 2.4 only gives lambda^u(p) = +/-d for each. Step 1 propagates either \"one side in K^+\" or \"both sides in escape\", but not the sign itself. In the all-saddles branch, you could have a mix of +d and -d saddles; with d large (after passing to f^3, the degree is d^3 >= 8), the pigeonhole does not force d-1 of one sign. So the Bedford-Kim contradiction does not fire as written. The text's \"In particular f has d distinct fixed points and at least d-1 of them are saddles with unstable multiplier equal to d\" is not justified in that branch.\n\nIs it fatal? Not necessarily. A short argument showing the sign is constant in the all-saddles case (e.g., using the propagation of escape half-planes plus the dissipativity c != 0 from Lemma 2.9) might fix it. But as it stands, the proof of Corollary 2.6 and hence Theorem 2.7 has a gap. The authors need to address this.\n\nMinor notes: Proposition 8.6 is a sketch, and Lemma 8.5 has a delicate substantiality point, but those are secondary. The appendix's diagonal extraction is terse but plausible.\n\nBottom line: this paper deserves a serious referee. I would send it, but ask the referee to focus on Proposition 2.5 Step 3. If the authors close that gap, the paper is a strong accept; if not, the main theorem C is unproven. I would cite it now for the statements, with a mental asterisk until the gap is resolved.","headline":"Strong, important paper with a real gap in Proposition 2.5's all-saddles case; likely fixable, but the written proof of Theorem C is incomplete.","tokens_in":43987,"tokens_out":9606,"would_cite":true,"duration_ms":84943,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["32H50","37F10","37F80","37D20"],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper proves that no loxodromic polynomial automorphism of C2 is unstably linear, so Julia slices are never smooth curves, and derives foliation and conjugacy rigidity.","keywords":["polynomial automorphisms of C^2","Julia sets","rigidity","loxodromic automorphisms","real-analytic conjugacy","multipliers","foliations","Hausdorff dimension"],"falsifier":"Construct, or find computationally, a loxodromic polynomial automorphism of $\\mathbb C^2$ with a saddle fixed point whose unstable multiplier is $\\pm d$ and whose Julia slice along its unstable manifold is a straight line; or exhibit a dissipative composition of three Hénon maps of degree $d$ with $d-1$ fixed points sharing unstable multiplier $d$. Either example would refute Theorem C or the multiplier input [4] on which it depends.","tokens_in":42950,"feed_emoji":"🌀","tokens_out":8613,"duration_ms":76048,"temperature":0.7,"pith_summary":"This paper establishes several rigidity theorems for polynomial automorphisms of $\\mathbb C^2$ with positive entropy (called loxodromic). Its central result is that no such map can be 'unstably linear': after straightening the unstable manifold of a saddle point, the forward Julia set can never be a straight line; in particular, a complex slice of the forward or backward Julia set is never a $C^1$ curve, and never a rectifiable curve. The authors use this to show that no loxodromic automorphism preserves a global real-analytic foliation with complex leaves, hence no global holomorphic foliation, and that any real-analytic conjugacy between two loxodromic automorphisms, under a mild non-real-multiplier hypothesis, is a polynomial conjugacy, possibly composed with complex conjugation. A final arithmetic theorem says that under natural hypotheses the saddle multipliers cannot all lie in a fixed number field.","feed_headline":"Julia slices of C2 automorphisms are never smooth curves","feed_subtitle":"Rigidity results rule out invariant foliations and force real-analytic conjugacies to be polynomial.","key_machinery":"The central objects are the notions 'unstably real' and 'unstably linear', defined through unstable manifold parametrizations $\\psi_p^u:\\mathbb C\\to W^u(p)$: the map is unstably real if $(\\psi_p^u)^{-1}(J^+ \\cap W^u(p))$ is contained in a line, and unstably linear if it is a line. The proof machinery is the quasi-expansion theory of [12], which organizes the family of normalized unstable parametrizations into a normal family and shows that unstably linear maps are quasi-expanding; an argument adapted from [13] plus hyperbolicity criteria [2] upgrades this to uniform hyperbolicity, and the final contradiction comes from the fixed-point multiplier theorem [4, Props. 5.1 and 6.1]. The identity driving the multiplier estimates is that $G^+\\circ\\psi_p^u$ is proportional to $|\\operatorname{Im}\\zeta|$ on each half-plane, so invariance under multiplication by $\\lambda^u$ forces $\\lambda^u=\\pm d^n$. For the rectifiable-slice and foliation theorems, the key mechanism is the comparison between harmonic measure and the Laplacian of the Green function $G^+$ on a transversal: rectifiability would produce tangents at almost every point, forcing unstable linearity, while saturation by a hypothetical foliation would extend the local picture to a global holomorphic foliation, contradicting [15].","core_discovery":"At the heart of the paper is Theorem C: unstably linear loxodromic automorphisms of $\\mathbb C^2$ do not exist, where 'unstably linear' means that for every saddle periodic point $p$, the pullback $(\\psi_p^u)^{-1}(J^+ \\cap W^u(p))$ is a line through the origin in the unstable parametrization. The proof combines quasi-expansion theory (Lemma 2.11 shows that unstably linear maps are quasi-expanding with factor $d$) with a rigidity result on fixed-point multipliers to reach a contradiction: an unstably linear dissipative map would force at least $d-1$ of its $d$ fixed points to have unstable multiplier $d$, which the cited multiplier theorem rules out. Along the way the authors show that all periodic points must be saddles and that both sides of $W^u(p) \\setminus J^+$ avoid the filled Julia set, and they upgrade quasi-expansion to uniform hyperbolicity before invoking the multiplier contradiction. A separate rectifiability argument (Theorem 3.1) rules out rectifiable Julia slices without passing through linearity, and Theorem B rules out invariant real-analytic foliations; Theorem A then converts real-analytic conjugacies into holomorphic or anti-holomorphic ones, where the earlier rigidity result [19] applies.","pith_inferences":["The non-existence of unstably linear maps suggests a two-dimensional analogue of Fatou's one-dimensional trichotomy: since smooth Julia slices cannot occur, the only possible irregularity of unstable slices is total disconnectedness versus Hausdorff dimension strictly greater than one; this dichotomy is consistent with Corollary 3.9.","If the multiplier-field theorems are sharpened, they point to a full multiplier-spectrum rigidity: the collection of all saddle multipliers should determine the automorphism up to polynomial conjugacy and complex conjugation, with no Chebyshev or monomial exceptional classes in $\\mathbb C^2$.","A concrete testable extension is numerical: for dissipative Hénon maps sufficiently close to one-dimensional polynomials, the saddle multipliers should be provably outside any fixed number field; high-precision computation could search for algebraic relations among multipliers in perturbative families.","The mechanism behind Theorem A—pulling back the complex structure and deriving an invariant foliation—might extend to smooth (not necessarily real-analytic) conjugacies if the real-analytic foliation arguments can be replaced by elliptic-regularity or quasiconformal techniques; the paper does not claim this extension."],"forward_implications":["A complex slice $J^+_\\Gamma$ of the forward Julia set along any transversal disk $\\Gamma$ is never a $C^1$ curve, and a nontrivial component never admits a tangent at a transverse intersection with a stable manifold (Corollary 2.8).","The same slices are never rectifiable curves; consequently a Jordan-arc slice, if it exists, must have infinite one-dimensional Hausdorff measure (Corollary 3.2 and Theorem 3.1).","No loxodromic automorphism of $\\mathbb C^2$ preserves a global holomorphic foliation, nor a global real-analytic foliation with complex leaves (Theorem 4.3 and Theorem B).","A real-analytic conjugacy between loxodromic automorphisms, when one saddle point has non-real stable and unstable multipliers, is necessarily a polynomial automorphism up to complex conjugation (Theorem A).","Under the stated hyperbolicity or Lyapunov-exponent hypotheses, the full set of unstable (and stable) multipliers cannot be contained in any fixed number field (Theorems 8.1 and 8.3)."],"supporting_citations":[{"why":"Supplies the multiplier rigidity statements (Propositions 5.1 and 6.1) that produce the final contradiction in Theorem C.","marker":"[4]"},{"why":"Supplies quasi-expansion theory and normal families of unstable parametrizations used to upgrade unstably linear maps to hyperbolic ones.","marker":"[12]"},{"why":"Supplies the theorem on algebraic foliations used to rule out global holomorphic invariant foliations.","marker":"[15]"},{"why":"Provides the laminar current structure, stable and unstable manifolds, and homoclinic intersection facts underlying the Julia set geometry.","marker":"[7]"},{"why":"Is the earlier holomorphic-conjugacy rigidity result that Theorem A ultimately reduces to.","marker":"[19]"},{"why":"Provides equidistribution of saddle periodic points and Lyapunov exponent averages used in Theorems D and A.1.","marker":"[6]"},{"why":"Normalizes loxodromic automorphisms as compositions of Hénon maps and formulates the conjugacy rigidity problem.","marker":"[38]"},{"why":"Supplies the one-dimensional multiplier-field rigidity argument adapted in Section 8.","marker":"[45]"},{"why":"Provides transversality and Zariski-density techniques used in Sections 2 and 6.","marker":"[29]"},{"why":"Supplies weak stability and persistence of saddle points used in the specialization arguments of Theorem 8.3.","marker":"[30]"}],"fun_headline_variants":["No smooth slices for Julia sets of C2 maps","C2 automorphisms: no smooth Julia slices, no foliations","Rigid C2 automorphisms: no smooth slices or invariant foliations","Under mild conditions, real-analytic conjugacy is polynomial","C2 automorphisms' Julia slices: no smoothness, no rectifiability"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The whole chain of theorems rests on an external multiplier rigidity statement [4, Propositions 5.1 and 6.1] asserting that a dissipative composition of at least three Hénon maps cannot have $d-1$ of its $d$ fixed points with the same unstable multiplier $d$; if that statement is false or has extra unstated hypotheses, Theorem C and consequently Theorems B and A would fail.","fun_headline_variants_meta":{"raw":{"variants":["No smooth slices for Julia sets of C2 maps","C2 automorphisms: no smooth Julia slices, no foliations","Rigid C2 automorphisms: no smooth slices or invariant foliations","Under mild conditions, real-analytic conjugacy is polynomial","C2 automorphisms' Julia slices: no smoothness, no rectifiability"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001191,"raw_usage":{"total_tokens":4901,"prompt_tokens":919,"completion_tokens":3982,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":535,"completion_tokens_details":{"reasoning_tokens":3888}},"tokens_in":535,"tokens_out":3982,"duration_ms":29456,"temperature":1.0,"reasoning_tokens":3888,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T19:43:47.321655+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Construct, or find computationally, a loxodromic polynomial automorphism of $\\mathbb C^2$ with a saddle fixed point whose unstable multiplier is $\\pm d$ and whose Julia slice along its unstable manifold is a straight line; or exhibit a dissipative composition of three Hénon maps of degree $d$ with $d-1$ fixed points sharing unstable multiplier $d$. Either example would refute Theorem C or the multiplier input [4] on which it depends.","supporting_citations":[{"cited_title":"No smooth Julia sets for polynomial diffeomorphisms of C2 with positive entropy","cited_arxiv_id":null,"evidence_quote":"Supplies the multiplier rigidity statements (Propositions 5.1 and 6.1) that produce the final contradiction in Theorem C."},{"cited_title":"Polynomial diffeomorphisms of C2","cited_arxiv_id":null,"evidence_quote":"Supplies quasi-expansion theory and normal families of unstable parametrizations used to upgrade unstably linear maps to hyperbolic ones."},{"cited_title":"Minimal models of foliated algebraic surfaces.Bull","cited_arxiv_id":null,"evidence_quote":"Supplies the theorem on algebraic foliations used to rule out global holomorphic invariant foliations."},{"cited_title":"Polynomial diffeomorphisms of C2","cited_arxiv_id":null,"evidence_quote":"Provides the laminar current structure, stable and unstable manifolds, and homoclinic intersection facts underlying the Julia set geometry."},{"cited_title":"Holomorphically conjugate polynomial automorphisms of C2 are polyno- mially conjugate","cited_arxiv_id":null,"evidence_quote":"Is the earlier holomorphic-conjugacy rigidity result that Theorem A ultimately reduces to."},{"cited_title":"Distribution of periodic points of polynomial diffeomor- phisms of C2","cited_arxiv_id":null,"evidence_quote":"Provides equidistribution of saddle periodic points and Lyapunov exponent averages used in Theorems D and A.1."},{"cited_title":"Dynamical properties of plane polynomial automorphisms","cited_arxiv_id":null,"evidence_quote":"Normalizes loxodromic automorphisms as compositions of Hénon maps and formulates the conjugacy rigidity problem."},{"cited_title":"Rational maps with rational multipliers","cited_arxiv_id":null,"evidence_quote":"Supplies the one-dimensional multiplier-field rigidity argument adapted in Section 8."},{"cited_title":"The dynamical Manin-Mumford problem for plane polynomial automor- phisms","cited_arxiv_id":null,"evidence_quote":"Provides transversality and Zariski-density techniques used in Sections 2 and 6."},{"cited_title":"Stability and bifurcations for dissipative polynomial automorphisms of C2","cited_arxiv_id":null,"evidence_quote":"Supplies weak stability and persistence of saddle points used in the specialization arguments of Theorem 8.3."}],"review_version":1}