{"id":"523b80b1-1a4d-497b-ba72-9157b66bc4b6","arxiv_id":"2411.13246","paper_version":4,"verdict":"ACCEPT","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"The dual web of every reduced convex pre-foliation on the complex projective plane is flat.","lead":"The paper proves that every reduced convex pre-foliation on the complex projective plane has a flat dual web. This extends earlier flatness results from invariant lines to arbitrary reduced invariant curves.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Proposition 7.1(3) is the load-bearing reduction in Theorem 1: it forces every bad singularity on a non-linear invariant component to be radial, but its proof depends on an unstated Camacho-Sad product identity and on [12, Cor. 3.8].","rationale":"The reader's weakest-assumption analysis correctly identifies Proposition 7.1(3) as the structural hinge: all later reductions in Theorem 1 assume that non-linear invariant components contribute nothing to Ξ_C^F. I read the proof as a coherent chain of local decomposition lemmas, curvature formulas, and a final degree argument; the local Lemmas 4.1, 4.3, 5.1, and 5.3 are internally plausible, and the canonical-divisor argument in the n=2 case is sound provided the pole-order estimate from [11, Prop. 2.2] holds. No internal inconsistency was found in the main argument. The genuine soft spot is the passage from Camacho-Sad index bounds to simplicity in Proposition 7.1(2), because the product identity is not derived in the text and the cited results come from separate classifications of convex foliations of degrees 4 and 5. Since this is a verification issue rather than a demonstrated error, the reader's ACCEPT verdict with moderate confidence remains appropriate. The concrete test on F_4^0 would either confirm the proposition on a nontrivial reduced convex foliation or expose a gap in the cited chain.","tokens_in":18665,"tokens_out":28833,"duration_ms":328755,"concrete_test":"For the Fermat foliation F_4^0 of §8.1, compute Sing(F_4^0) directly from the 1-form, determine the radial set from ν=1 and τ≥2, and verify Proposition 7.1(3): every singular point lying on a non-linear invariant curve C_λ must be radial. Simultaneously, at an intersection of two invariant lines of F_4^0, compute the two Camacho-Sad indices and check whether the product identity CS(ℓ1)CS(ℓ2)=1 holds without first assuming simplicity. If the product identity fails for any non-reduced local model, or if a non-radial singularity is found on some C_λ, Proposition 7.1(3) needs an additional argument before Theorem 1 relies on it.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The main proof reduces flatness of LegF to control of the exceptional set Ξ_C^F = SingC \\ Σ_rad^F. For a non-linear invariant component C0, the only tool excluding points of SingF∩C0 from Ξ_C^F is Proposition 7.1(3), which asserts that every such singularity is radial. Its proof has two externally cited ingredients: (a) every non-radial singularity has its two Camacho-Sad indices outside R_+ (Proposition 7.1(2), via [6, Prop. 3.2] and [8, Cor. 3.3]), and (b) a simple singularity has no invariant curve other than its two invariant lines ([12, Cor. 3.8]). Step (a) uses the identity CS(F, ℓ1) CS(F, ℓ2) = 1. The text does not prove this identity for possibly non-reduced singularities, although it is exactly simplicity that step (a) is meant to establish. If the identity fails outside the simple case, or if [8, Cor. 3.3] does not apply to every reduced convex foliation rather than only degree-5 models, then Proposition 7.1(2) does not prove simplicity and Proposition 7.1(3) loses its basis. The rest of the proof, including the n=2 pole-order argument and the n≥3 curvature decomposition, depends on Ξ_C^F having no contribution from non-linear components.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proves Theorem 1: if F = C ⊠ F is a reduced convex pre-foliation of degree d ≥ 3 on P^2_C, with C invariant by F, then the dual Legendre web LegF is flat. The proof computes the discriminant of LegF, gives local normal forms of LegF near the components of that discriminant (Lemmas 4.1, 4.3, 5.1, 5.3), establishes a curvature criterion (Theorem 6.1 and Corollary 6.4), and then reduces flatness to controlling the exceptional set Ξ_C^F = SingC \\ Σ_rad^F. The final reduction uses Proposition 7.1 to force singularities on non-linear invariant components to be radial, and then treats the cases where C has zero, one, two, or at least three invariant lines. The paper also provides examples for Fermat, Hesse, and Hilbert modular foliations and states two conjectures.","tokens_in":18935,"tokens_out":21020,"duration_ms":241785,"significance":"If the proof is correct, the theorem is a genuine extension of the previous results in [10] and [11] from invariant curves that are unions of lines to arbitrary reduced invariant curves in a reduced convex pre-foliation. The curvature criterion in Corollary 6.4 and the explicit discriminant decomposition of Lemma 3.1 are useful tools. The paper is written in the standard style of the subject and contains substantial, concrete examples. The main caveat is that the final reduction depends on a Camacho–Sad product identity that is asserted without proof or reference, and the manuscript leans on the unpublished preprint [11] for several central curvature decompositions.","major_comments":[{"comment":"The identity CS(F, ℓ_m^(1), m) · CS(F, ℓ_m^(2), m) = 1 is stated without proof or reference. This is not a general identity for arbitrary pairs of invariant curves through a singular point, and at this stage of the argument simplicity of m has not yet been established, so the identity cannot be assumed as a consequence of the conclusion it is meant to prove. Since part (3) of Proposition 7.1, and hence the reduction preceding Theorem 1, relies on the simplicity statement, this is load-bearing. Please supply a proof of the product formula under the hypotheses of part (2), or cite an exact statement in [6] or [8] that implies it.","section":"§7, Proposition 7.1(2)"},{"comment":"The proof uses [11, Lemma 2.1] and [11, Theorem 1] as black boxes, but [11] is an unpublished arXiv preprint by the author. In particular, the curvature decomposition used in the n ≥ 3 case of Theorem 1 is exactly [11, Lemma 2.1], and the case where C is a union of invariant lines is delegated to [11, Theorem 1]. For a journal publication, the necessary statements from [11] should either be included in the paper or the reference should be updated to a published version.","section":"§7, proof of Theorem 1; §6, Proposition 6.2"},{"comment":"The step 'En reprenant un argument de la démonstration de [11, Proposition 2.2]' is vague at the precise point where the proof needs the pole order of K(LegF) along ˇs to be at most one. That bound is essential for the canonical-divisor argument that yields K(LegF) ≡ 0. Please state the quoted proposition or spell out the argument, since the conclusion of the whole n = 2 case depends on it.","section":"§7, proof of Theorem 1, n = 2 case"}],"minor_comments":[{"comment":"The proof cites [16, Lemme 2.5] for the inequality mult(Δ(W), C0) ≥ n − 1 for a totally invariant component; for readability, please state the needed form of that lemma, since it is used twice in the proof.","section":"§4, Lemma 4.3"},{"comment":"The sentence about the validity of Conjecture 1 for d ∈ {2,3} is terse: it would help to indicate explicitly which cited result covers which of the three alternatives in the dichotomy.","section":"§9, Conjecture 1"},{"comment":"In the display after the definition of F_0^d, the notation for the forms defining H_0^d, H_1^d and F_1^d is easy to confuse because ω^d_1 and Ω^d_1 differ only by a font; please disambiguate the names of the forms and the foliations.","section":"§8, Section 8.1"}],"recommendation":"major_revision","confidential_remarks":"The central claim is plausible and the paper is carefully organized, but the referee report above identifies a load-bearing step in Proposition 7.1(2) that the text does not justify. I would also ask the editor to verify the status of [11], since several curvature decompositions and the all-lines case are taken from that unpublished preprint. The paper is otherwise a strong contribution to the flatness program for dual webs of pre-foliations."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Samir has a real result here. The theorem—reduced convex pre-foliations of degree at least 3 have flat dual Legendre webs—extends the Marín–Pereira theorem from foliations to arbitrary reduced invariant curves, and it covers the non-line components that his earlier work left open. The paper earns the result: the discriminant formula in Lemma 3.1, the local splittings in Lemmas 4.1/4.3/5.1/5.3, and the curvature characterization in Theorem 6.1 all work together, and Theorem 6.1's explicit exceptional set is independently useful. The Section 8 examples are concrete and checkable.\n\nThe proof is coherent. The main caveats are proportionate, not fatal: the paper relies heavily on the author's own unpublished preprint [11] for the curvature decomposition lemma and the line-only case; that is honest but makes verification slower. More substantively, Proposition 7.1(3) carries a lot of weight—it forces every singularity on a non-linear invariant component to be radial—and its proof is compressed. The Camacho–Sad product identity used in point 2 is stated without proof, and the text does not spell out why [8, Cor. 3.3] applies to every reduced convex foliation rather than only the degree-5 models. The stress-test note is right to flag this passage. I did not find a counterexample, and the identities look like standard tools in this area, but a referee should ask the author to expand that step.\n\nThe self-citation pattern is heavy but appropriate; the cited results are the actual tools. This is a paper for specialists in web geometry and holomorphic foliations, and it deserves referee time. I would send it to peer review.","headline":"A credible proof of the natural generalization of Marín-Pereira to reduced convex pre-foliations; referee it, but ask for the compact proof of Proposition 7.1(3) to be expanded.","tokens_in":19474,"tokens_out":3485,"would_cite":true,"duration_ms":36378,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["14C21","32S65","53A60"],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper proves that every reduced convex pre-foliation of degree at least 3 on the complex projective plane has a flat Legendre dual web: its Blaschke curvature vanishes identically.","keywords":["pre-foliation","convex pre-foliation","reduced convex foliation","dual web","Legendre transform","Blaschke curvature","flat web","projective plane"],"falsifier":"Compute the residue of the Blaschke curvature $K(\\mathrm{Leg}\\,\\mathscr{F})$ along the dual line $\\check{s}$ of the intersection point of two invariant lines in one of the explicit reduced convex pre-foliations of Section 8, with the curve completed by one non-line invariant component. The theorem forces the residue to vanish, since the curvature has at most a simple pole there and the canonical-divisor argument kills it; any nonzero residue would be a direct counterexample to Theorem 1. Equivalently, exhibiting a reduced convex foliation with a non-radial singularity on an invariant curve of degree at least 2 would destroy Proposition 7.1(3), the step on which the proof's reduction relies.","tokens_in":18443,"feed_emoji":"🕸️","tokens_out":18868,"duration_ms":164815,"temperature":0.7,"pith_summary":"This paper proves that every reduced convex pre-foliation of degree $d\\ge 3$ on the complex projective plane has a flat dual web. A pre-foliation is the data of a reduced algebraic curve together with a holomorphic foliation that leaves the curve invariant; the dual web is the $d$-web on the dual plane obtained by Legendre transform, whose leaves are the tangents to the leaves of the foliation and of the invariant curve. Flatness means the web's Blaschke curvature, the obstruction to being locally equivalent to the standard family of parallel lines, vanishes identically. The result extends earlier work where the invariant curve was a single line or a union of lines to arbitrary reduced invariant curves. It says that this entire family of dual webs carries no curvature, so locally each one looks like the trivial web.","feed_headline":"The dual web of a reduced convex pre-foliation is flat","feed_subtitle":"The Legendre dual of any reduced convex pre-foliation has vanishing curvature.","key_machinery":"The central objects are the Legendre transform web $\\mathrm{Leg}\\,\\mathscr{F}$ and its Blaschke curvature $K$. The web is the $d$-web on the dual projective plane defined by the implicit differential equation obtained by substituting the equation of a variable line into the 1-form defining the foliation; a web is flat when $K$ vanishes. The proof is carried by four interlocking tools: the discriminant formula $\\Delta(\\mathrm{Leg}\\,\\mathscr{F})=\\check{\\mathcal{C}}\\cup\\check{\\Sigma}^{\\mathcal{C}}_{\\mathcal{F}}\\cup\\check{\\Sigma}^{\\mathrm{rad}}_{\\mathcal{F}}$ for reduced convex pre-foliations (Corollary 3.2); local decompositions of $\\mathrm{Leg}\\,\\mathscr{F}$ near components of its discriminant, such as $W_n\\boxtimes W_\\tau\\boxtimes W_{d-n-\\tau}$ near the dual of a singular point and $\\mathcal{F}_0\\boxtimes W_2\\boxtimes W_{d-3}$ near the dual of a non-line component of $\\mathcal{C}$; a curvature decomposition formula expressing $K$ of a product of webs as a combination of curvatures of smaller webs; and the structural Proposition 7.1, which says that in a reduced convex foliation every non-radial singularity is simple and has no invariant curve other than its two invariant lines. Together these reduce flatness of the full web to holomorphy of the curvature along a finite set of dual lines, and then to zero.","core_discovery":"The central claim is Theorem 1: if $\\mathscr{F}=\\mathcal{C}\\boxtimes\\mathcal{F}$ is a reduced convex pre-foliation of degree $d\\ge 3$ on $\\mathbb{P}^{2}_{\\mathbb{C}}$, then the $d$-web $\\mathrm{Leg}\\,\\mathscr{F}$ is flat. The proof establishes a finer statement (Theorem 6.1 and Corollary 6.4): for a convex pre-foliation the curvature $K(\\mathrm{Leg}\\,\\mathscr{F})$ is holomorphic on the dual plane away from the duals of the non-radial singular points of $\\mathcal{C}$ and the duals of the non-simple singularities of $\\mathcal{F}$; in the reduced convex case only the lines dual to $\\Xi^{\\mathcal{C}}_{\\mathcal{F}}=\\mathrm{Sing}\\,\\mathcal{C}\\setminus\\Sigma^{\\mathrm{rad}}_{\\mathcal{F}}$ can obstruct flatness, and the web is flat exactly when the curvature is holomorphic along them. The remainder of the proof shows those lines cannot carry poles: when the curve has at most one invariant line the exceptional set is empty; with exactly two invariant lines the possible pole at the dual of their intersection has at most order one and is killed by a degree argument on the canonical divisor; with three or more invariant lines a decomposition formula reduces the curvature to previously settled cases. The theorem thereby generalizes the flatness result for the foliation-only and line-only settings to an arbitrary reduced invariant curve.","pith_inferences":["Inference: the curvature-holomorphy criterion of Theorem 6.1 gives a concrete route toward the paper's Conjecture 2: for a convex but not reduced pre-foliation, one need only compute residues of $K$ along the duals of the non-radial singular points of $\\mathcal{C}$ and the non-simple singularities of $\\mathcal{F}$.","Inference: since Section 8 obtains new flat webs by degenerating already flat webs to the boundary of the automorphism-group orbit, flatness of Legendre dual webs is plausibly a closed condition in families of convex pre-foliations; proving that directly would turn the examples into a general principle.","Inference: the same architecture — a discriminant formula, subweb decompositions, a curvature splitting formula, and a canonical-divisor pole count — may adapt to Legendre-type transforms of foliations with invariant divisors in higher dimensions, where a flatness notion for webs on dual spaces is less studied."],"forward_implications":["For every reduced convex pre-foliation of degree $d\\ge 3$, the dual $d$-web is flat; in particular the flatness result now covers arbitrary reduced invariant curves, not just unions of invariant lines.","The discriminant of the dual web has the explicit form $\\Delta(\\mathrm{Leg}\\,\\mathscr{F})=\\check{\\mathcal{C}}\\cup\\check{\\Sigma}^{\\mathcal{C}}_{\\mathcal{F}}\\cup\\check{\\Sigma}^{\\mathrm{rad}}_{\\mathcal{F}}$, so the only possible polar locus is the dual curve plus duals of radial singularities and singular points of the invariant curve.","Flatness of $\\mathrm{Leg}\\,\\mathscr{F}$ is equivalent to holomorphy of its Blaschke curvature along the duals of the singular points of $\\mathcal{C}$ that are not radial singularities of $\\mathcal{F}$; the theorem shows that, for reduced convex pre-foliations, those dual lines carry no poles.","For a reduced convex foliation, every non-radial singularity is simple, has exactly two invariant lines, and lies on no invariant curve of degree at least 2; any non-line invariant curve meets the foliation only in radial singularities.","The explicit families treated in Section 8 give concrete flat dual webs: every invariant algebraic curve for the degree-$d$ homogeneous foliation, for the two Hessian foliations, and for the degree-5 foliation whose only invariants are lines."],"supporting_citations":[{"why":"Defines the Legendre transform and Blaschke curvature, and proves flatness of the dual web for reduced convex foliations, the foliation-only case this paper extends.","marker":"[16]"},{"why":"Proves the union-of-invariant-lines case and supplies the curvature decomposition formula (Lemma 2.1) used to handle three or more lines.","marker":"[11]"},{"why":"Proves the single-invariant-line case and supplies the subweb-curvature holomorphy criteria (Proposition 3.9 and Remark 3.10) used in Theorem 6.1.","marker":"[10]"},{"why":"Shows that each non-radial singularity of a reduced convex foliation has exactly two invariant lines and gives homogeneous convex models used to control indices.","marker":"[6]"},{"why":"Provides the bound on Camacho–Sad indices for those homogeneous models that rules out real non-negative indices at non-radial singularities.","marker":"[8]"},{"why":"Gives the result that a simple singularity has no invariant curve other than its two invariant lines, which yields Proposition 7.1(3).","marker":"[12]"},{"why":"Supplies the discriminant formula for the dual web of a foliation used in Lemma 3.1.","marker":"[3]"},{"why":"Shows reduced convex foliations have only non-degenerate singularities, removing the $\\Sigma^{\\nu\\ge 2}$ term from the discriminant in the reduced case.","marker":"[4]"},{"why":"Provides the web-geometry facts on dual curves and discriminant multiplicities used in Lemmas 4.1 and 4.3.","marker":"[18]"},{"why":"Shows a radial singularity has locally distinct branches, used in Lemma 6.3 to conclude it is an ordinary multiple point of the invariant curve.","marker":"[19]"}],"fun_headline_variants":["Dual web flatness proven for reduced convex pre-foliations","Reduced convex pre-foliations yield flat Legendre dual webs","Flatness of dual webs: new result for reduced convex pre-foliations","Curvature zero: dual webs of reduced convex pre-foliations","Theorem: Legendre duals of reduced convex pre-foliations are flat"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The proof assumes that in a reduced convex foliation a non-radial singular point is simple, with only the two invariant lines through it, so that an invariant curve of degree at least two cannot pass through such a point; if that failed, the exceptional set of dual lines would be nonempty and the reduction would break.","fun_headline_variants_meta":{"raw":{"variants":["Dual web flatness proven for reduced convex pre-foliations","Reduced convex pre-foliations yield flat Legendre dual webs","Flatness of dual webs: new result for reduced convex pre-foliations","Curvature zero: dual webs of reduced convex pre-foliations","Theorem: Legendre duals of reduced convex pre-foliations are flat"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00025,"raw_usage":{"total_tokens":1586,"prompt_tokens":1008,"completion_tokens":578,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":624,"completion_tokens_details":{"reasoning_tokens":482}},"tokens_in":624,"tokens_out":578,"duration_ms":5567,"temperature":1.0,"reasoning_tokens":482,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T16:40:04.248160+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute the residue of the Blaschke curvature $K(\\mathrm{Leg}\\,\\mathscr{F})$ along the dual line $\\check{s}$ of the intersection point of two invariant lines in one of the explicit reduced convex pre-foliations of Section 8, with the curve completed by one non-line invariant component. The theorem forces the residue to vanish, since the curvature has at most a simple pole there and the canonical-divisor argument kills it; any nonzero residue would be a direct counterexample to Theorem 1. Equivalently, exhibiting a reduced convex foliation with a non-radial singularity on an invariant curve of degree at least 2 would destroy Proposition 7.1(3), the step on which the proof's reduction relies.","supporting_citations":[{"cited_title":"Bedrouni and D","cited_arxiv_id":null,"evidence_quote":"Defines the Legendre transform and Blaschke curvature, and proves flatness of the dual web for reduced convex foliations, the foliation-only case this paper extends."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Proves the union-of-invariant-lines case and supplies the curvature decomposition formula (Lemma 2.1) used to handle three or more lines."},{"cited_title":"Alors toutes les singularités de F sur C sont radiales, i.e","cited_arxiv_id":null,"evidence_quote":"Proves the single-invariant-line case and supplies the subweb-curvature holomorphy criteria (Proposition 3.9 and Remark 3.10) used in Theorem 6.1."},{"cited_title":"Nous le particularisons ensuite au cas où F est convexe réduit","cited_arxiv_id":null,"evidence_quote":"Shows that each non-radial singularity of a reduced convex foliation has exactly two invariant lines and gives homogeneous convex models used to control indices."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the bound on Camacho–Sad indices for those homogeneous models that rules out real non-negative indices at non-radial singularities."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Gives the result that a simple singularity has no invariant curve other than its two invariant lines, which yields Proposition 7.1(3)."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the discriminant formula for the dual web of a foliation used in Lemma 3.1."},{"cited_title":"Le lemme suivant traite le cas de la droite ˇs duale d’un point s ∈ C ( ˇs ⊂ ∆ (LegC ) ⇐ ⇒s ∈ SingC )","cited_arxiv_id":null,"evidence_quote":"Shows reduced convex foliations have only non-degenerate singularities, removing the $\\Sigma^{\\nu\\ge 2}$ term from the discriminant in the reduced case."},{"cited_title":"Bedrouni and D","cited_arxiv_id":null,"evidence_quote":"Provides the web-geometry facts on dual curves and discriminant multiplicities used in Lemmas 4.1 and 4.3."},{"cited_title":"Bedrouni and D","cited_arxiv_id":null,"evidence_quote":"Shows a radial singularity has locally distinct branches, used in Lemma 6.3 to conclude it is an ordinary multiple point of the invariant curve."}],"review_version":1}