REVIEW 3 major objections 4 minor 7 references
Webs Generated by Products of convex and homogeneous Foliations on $\mathbb{P}^2$
T0 review · 3 major / 4 minor · reviewed 2026-08-12 · deepseek-v4-flash
Pith's one-line read The paper proves that products of convex reduced foliations (or convex homogeneous foliations) with invariant lines have flat Legendre transforms on the dual projective plane, yielding new flat webs.
desk verdict A useful generalization of Bedrouni's flatness results, but Theorem A as stated needs a degree bound or a low-degree argument. read the letter →
The pith
A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.
The reading
What carries the argument
The Legendre transform of a web sends a $d$-web of degree $k$ on $\mathbb{P}^2$ to a $k$-web of degree $d$ on the dual plane by mapping each line to its tangency points; a web is flat when its curvature $K(W)$ vanishes, which on $\mathbb{P}^2$ is equivalent to $K$ being holomorphic along every irreducible component of the discriminant. The carrying objects are the local decompositions $\operatorname{Leg} F = W_\nu \boxtimes W_{d-\nu}$ around a radial singularity, where $W_\nu$ is a totally invariant radial foliation of minimal multiplicity, together with the discriminant formula $\Delta(\operatorname{Leg}(F_1 \boxtimes F_2)) = \Delta(\operatorname{Leg} F_1) \cup \Delta(\operatorname{Leg} F_2) \cup G_{F_i}(C_{\mathrm{inv}}) \cup G_{F_i}(C_{\mathrm{tr}})$. These tools reduce flatness of a product to holomorphy of curvature along the dual lines of radial singularities, checked through pole-order estimates and the barycenter criterion.
What would settle it
Compute the curvature of $\operatorname{Leg}(F_1 \boxtimes F_2)$ for a pair of degree-2 reduced convex foliations satisfying the invariant-line tangency condition; if a curvature pole appears along a dual radial line, Theorem A as stated is false, and if it does not, the missing degree bound is harmless.
Extended reading notes
Core claim
On the paper's own terms, the central claim is that flatness of the dual web is forced by the tangency geometry: when two convex foliations meet only along common invariant lines, the curvature of the Legendre transform of their product has no poles. The proof shows the discriminant of $\operatorname{Leg} W$ is exactly the union of the dual lines of radial singularities (plus the dual of the line at infinity in the homogeneous case), and at each such component the curvature extends holomorphically because the Legendre transform decomposes locally as a radial foliation times a transverse web. Consequently the product web has zero curvature on the whole projective plane. This gives new examples of flat webs, and Example 3.6 shows the invariant-line condition is necessary: a transversal tangency component makes the dual curvature develop a pole.
Load-bearing premise
The proof of Theorem A relies on a local decomposition of the Legendre transform near a radial singularity that the cited source states only for foliations of degree at least 3, while the theorem itself imposes no degree bound.
Editorial extensions
If this is right
- Any finite collection of reduced convex foliations with pairwise invariant-line tangencies, together with any common invariant lines, yields a flat web on the dual projective plane.
- For Fermat foliations, the pairs $F_l, F_d$ with $d = 2l-1$ give explicit flat webs, and Proposition 2.4 shows these are the only Fermat pairs whose tangency locus is reduced and made of invariant lines.
- For convex homogeneous foliations, the forms $\omega_d = y^d\,dx - x^d\,dy$ with degrees $d, d+1, d+2$ combine with the invariant lines $L_\infty, x=0, y=0, y=x$ to produce flat webs.
- Example 3.6 demonstrates that a transversal tangency component produces a non-flat dual web, so the invariant-line hypothesis is necessary in both theorems.
- Because flatness is a necessary condition for a web to have maximal rank, these constructions supply candidates for extremal webs on $\mathbb{P}^2$.
Reading between the lines
- The theorem statements do not include a degree bound, but the proof of Proposition 2.2 explicitly assumes $d_1, d_2 \ge 3$; a separate argument would be needed to cover degree-2 reduced convex foliations, since the cited radial decomposition may degenerate there.
- The same discriminant mechanism suggests a general criterion: for a product of foliations, a transversal component in the tangency locus should force a curvature pole at its Gauss image, making flatness equivalent to the tangency locus being entirely invariant.
- The homothety invariance used in Lemma 3.1 reduces flatness for homogeneous products to checking holomorphy away from one point; this could be turned into an algebraic test for flatness of arbitrary homogeneous foliation products.
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The paper studies flatness of the Legendre transform of webs on P^2 formed as products of invariant lines and convex foliations. Theorem A claims that if F_1,...,F_n are reduced convex foliations whose pairwise tangency loci are unions of invariant lines, and l_1,...,l_k are common invariant lines, then Leg(l_1 ⊠ ... ⊠ l_k ⊠ F_1 ⊠ ... ⊠ F_n) is flat. Theorem B makes the analogous claim for convex homogeneous foliations with tangency loci L_∞ ∪ C^inv_{ij}. The proof strategy is to reduce flatness of a product to flatness of its pairs and triples via curvature decompositions (Propositions 1.5–1.7), to identify the discriminant of the Legendre transform (Propositions 2.1, 3.2), and to analyze local decompositions near radial singularities using results of Marín–Pereira and Bedrouni. The paper also gives Fermat and homogeneous examples satisfying the hypotheses, and an example showing the invariant-tangency hypothesis is necessary.
Significance. If the stated theorems are correct, the paper provides new families of flat webs on P^2 and generalizes results of Bedrouni; the reduction to pair and triple flatness is a useful organizing principle, and the concrete examples are valuable. The paper is honest about many of its ingredients and cites the relevant literature. However, the main theorem as stated is not fully proven: the proof of the two-foliation case explicitly assumes degrees at least 3, while Theorem A imposes no degree bound, and the Fermat example includes degree-2 factors. In addition, Lemma 1.4, which is used in the proofs of Propositions 2.5 and 2.6, is asserted without proof. These are load-bearing gaps for the central claim, although they appear fixable within the scope of the manuscript.
major comments (3)
- [Theorem A / Proposition 2.2] Theorem A is stated for reduced convex foliations of arbitrary degree, but its proof depends on Proposition 2.2, which explicitly assumes d_1, d_2 ≥ 3. The proof of Proposition 2.2 imports [6, Proposition 3.3] to write Leg F_i = W_{ν_i} ⊠ W_{d_i−ν_i} near a radial singularity, and no separate argument is given for the degree-2 case. This is not a vacuous gap: Example 2.3 applies Proposition 2.2 to Fermat foliations F_l and F_d with d = 2l − 1, which for l = 2 includes the reduced convex Fermat foliation F_2. Thus Theorem A as written includes cases that the proof does not cover.
- [Lemma 1.4] Lemma 1.4 is stated as 'an adaptation' of [2, Proposition 2.9] and no proof is supplied. The lemma is then used in Propositions 2.5 and 2.6 to control curvature in the presence of an additional tangent foliation, and these propositions feed directly into Theorem A. Since the lemma is not identical to the cited result, the flatness claim depends on an unproved local statement; a proof or a precise reduction to [2, Proposition 2.9] is needed.
- [Theorem B / Propositions 3.2–3.3] The homogeneous case is delegated to 'a similar analysis to that done in Proposition 2.2' (Propositions 3.2 and 3.3). If the degree-2 issue in Proposition 2.2 is not resolved, Theorem B inherits the same gap for homogeneous factors of degree 2. Moreover, the local analysis near s ∈ Σ^rad_{H_1} ∪ Σ^rad_{H_2} is not written out explicitly, so the reader cannot verify that the cited results apply without the same degree restrictions.
minor comments (4)
- [Throughout] There are numerous typos and small grammatical errors, e.g., 'convex reduced' vs 'reduced convex', 'in a union' for 'is a union', and 'F ALLA LUZA' in the author header; these should be corrected in a revision.
- [Proposition 1.3 proof] The notation for the local equations of the webs, especially the displayed formula for T W_d|_U, is hard to parse; writing the equations with explicit coefficient functions and clarifying the meaning of '1/ν g_j(z,w)' would improve readability.
- [Proposition 2.5 proof] In the case s ∈ Sing F_2 ∩ Sing F_3, after writing Leg F_2 = F_2 ⊠ W_{d_2−1}, the symbol F_2 is overloaded: it denotes both the original foliation and a local foliation tangent to s. Using a different letter for the local foliation would avoid confusion.
- [Example 2.3] The phrase 'the unity group composed of (l−1) roots of unity' should be 'the group of (l−1)th roots of unity'; also, the inclusion I_{F_l} ⊆ I_{F_d} is asserted and used to conclude that the tangency locus is contained in the inflection divisor, which is correct but deserves a brief justification.
Circularity Check
No circularity: the proof chain is self-contained against external flatness results; the flagged issues are coverage gaps, not circular reasoning.
full rationale
Walking the derivation chain of Theorems A and B, I find no step in which a claimed prediction is equivalent by construction to its input. Flatness of Leg W is reduced, via a curvature decomposition formula from [1, Lemma 2.1] and Propositions 1.5-1.7, to flatness of subproducts such as Leg F_i ⊠ Leg F_j and Leg l_i ⊠ Leg F_j, and those subproducts are established using Proposition 2.1/3.2 together with external local results [6, Prop. 2.6, Prop. 3.3, Thm. 4.2], [2, Prop. 3.9, Thm. 3.4, Rem. 3.10], and [3, Lemma 6.8, Lemma 3.2]. None of these inputs is the target flatness statement, and the paper does not fit parameters or rename a known result as a prediction. The only citation with author overlap is [4] (Beltran-Falla Luza-Marin), used as a barycenter criterion for degree-one 3-webs in a subcase of Proposition 2.6(iii) and in Example 3.6; it is a published, parameter-free external theorem and is not load-bearing for the main construction. Two non-circular gaps are worth flagging explicitly: Theorem A is stated for reduced convex foliations of arbitrary degree, whereas the proof of Proposition 2.2 assumes "respective degrees d1,d2 ≥ 3" and invokes [6, Proposition 3.3]; the degree-2 case such as the Fermat foliation F_2 in Example 2.3 is therefore not covered by the argument as written. Also, Lemma 1.4 is stated without proof as "an adaptation of [2, Proposition 2.9]". These are completeness/correctness concerns, not circularity. The derivation is self-contained against external benchmarks, so the circularity score is 0.
Assumptions & free parameters
assumptions (4)
- standard math Validity of the external curvature results [6, Lemma 2.2, Prop 2.6, Prop 3.3, Theorem 4.2] used to regulate the poles of the curvature along radial singularities.
- domain assumption Correctness of Bedrouni's preprint results [1, Lemma 2.1, Theorem 2] and [2, Prop 3.9, Remark 3.10, Theorem 3.4] under the hypotheses used in Theorems A and B.
- ad hoc to paper Lemma 1.4, an adaptation of [2, Proposition 2.9] to include a tangent foliation, is true.
- ad hoc to paper The degree condition d_i ≥ 3 is either satisfied or unnecessary for the foliations in Theorem A.
Cite this review
Pith. "Pith review of Webs Generated by Products of convex and homogeneous Foliations on $\mathbb{P}^2$." pith.science (2026). https://pith.science/paper/DU5LZW5H
@misc{pith2026241109602,
author = {Pith},
title = {Pith review of: Webs Generated by Products of convex and homogeneous Foliations on $\mathbbP^2$},
year = {2026},
howpublished = {\url{https://pith.science/paper/DU5LZW5H}},
note = {Machine review of arXiv:2411.09602}
}
read the original abstract
This paper investigates flat webs on the projective plane. We present two methods for constructing such webs: the first involves taking the product of finitely many convex reduced foliations and invariant lines, while the second consists of taking the product of finitely many convex homogeneous foliations and invariant lines. In both cases, we demonstrate that the dual web is flat.
Reference graph
Works this paper leans on
-
[1]
Platitude des tissus duaux de certains pr \'e -feuilletages convexes du plan projectif complexe
Bedrouni BEDROUNI 2024 BEDROUNI, S. Platitude des tissus duaux de certains pr \'e -feuilletages convexes du plan projectif complexe. arXiv preprint arXiv:2405.05464, 2024
arXiv 2024
-
[2]
Pre-foliations of co-degree one on P _ c ^2 with a flat legendre transform
Bedrouni BEDROUNI 2023 BEDROUNI, S. Pre-foliations of co-degree one on P _ c ^2 with a flat legendre transform. arXiv e-prints, p. arXiv--2309, 2023
work page 2023
-
[3]
Tissus plats et feuilletages homog \`e nes sur le plan projectif complexe
Bedrouni e Mar \' n BEDROUNI; MAR \'I N 2018 BEDROUNI, S.; MAR \'I N, D. Tissus plats et feuilletages homog \`e nes sur le plan projectif complexe. Bull. Soc. Math. France, v. 146, n. 3, p. 479--516, 2018
work page 2018
-
[4]
Beltr \'a n, Luza e Mar \' n BELTR \'A N; LUZA; MAR \'I N 2014 BELTR \'A N, A.; LUZA, M. F.; MAR \'I N, D. Flat 3-webs of degree one on the projective plane. In: Annales de la Facult \'e des sciences de Toulouse: Math \'e matiques . [S.l.: s.n.], 2014. v. 23, n. 4, p. 779--796
work page 2014
-
[5]
Ince INCE 1944 INCE, E. L. Ordinary differential equations. [S.l.]: Dover Publications, 1944
work page 1944
-
[6]
Marín e Pereira MARíN; PEREIRA 2013 MARíN, D.; PEREIRA, J. V. Rigid flat webs on the projective plane. Asian Journal of Mathematics, International Press of Boston, v. 17, n. 1, p. 163--192, 2013
work page 2013
-
[7]
Pereira PEREIRA 2001 PEREIRA, J. V. Vector fields, invariant varieties and linear systems. In: Annales de l'institut Fourier. [S.l.: s.n.], 2001. v. 51, n. 5, p. 1385--1405
work page 2001
Reviewed August 12, 2026 · model on record in the stance chip above.
Discussion (0). Continue with ORCID to comment.