Pith. sign in

REVIEW 2 major objections 5 minor 32 references

This paper proves that averaging directional triangular Hilbert transforms against any odd L^q function on the circle yields a bounded twisted paraproduct on the full expected exponent range, with a sharp additional condition 1/p+1/q<2 gove

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Rough L^q averages of directional triangular Hilbert transforms are bounded in the full p≥1 range matching smooth kernels, with a sharp balance condition 1/p+1/q<2 for p<1.

T0 review reviewed 2026-08-01 challenge →

load-bearing objection A genuinely new sharp result for rough averages of triangular Hilbert transforms, but the exponential decay that makes it work is imported from the authors' own unreviewed preprint, so the paper is not yet self-contained. the 2 major comments →

arxiv 2607.20206 v1 pith:XKVCW6KQ submitted 2026-07-22 math.CA

Rough averages of triangular Hilbert transforms

Fred Yu-Hsiang Lin , Lenka Slav\'ikov\'a This is my paper
classification math.CA MSC 42B1542B2042B25
keywords triangular Hilbert transformtwisted paraproductrough kernelsL^q angular kernelssmoothing inequalityquasi-Banach estimatesbilinear singular integralssharp balance condition
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper studies the twisted paraproduct T_Ω whose kernel is a rough homogeneous function: the angular part Ω lies in L^q(S^1) for some q>1 and has zero mean, but is not assumed smooth. It proves that for every 1

Core claim

The central claim is a sharp balance condition between the integrability exponent q of the rough angular kernel and the target Lebesgue exponent p. For p≥1 the rough twisted paraproduct is bounded on the same exponent range as the smooth-kernel twisted paraproduct. For 0<p<1, boundedness for every mean-zero Ω∈L^q(S^1) holds if and only if 1/p+1/q<2, with explicit test functions showing unboundedness when 1/p+1/q≥2. The proof decomposes the rough kernel into low-, mid-, and high-frequency pieces: the high-frequency piece decays exponentially in the frequency parameter via a smoothing inequality for bilinear multipliers, and the mid-frequency piece is controlled by logarithmic estimates for sh

What carries the argument

A dyadic frequency decomposition of the rough kernel, split into low-, mid-, and high-frequency regimes. Low frequencies follow from the known boundedness of the twisted paraproduct with smooth kernels. The high-frequency (diagonal) piece is controlled by a companion bilinear Fourier multiplier theorem of smoothing-inequality type, which yields exponential decay in the frequency parameter j for the L^2×L^2→L^1 norm. The mid-frequency (off-diagonal) piece is bounded by reduction to shifted twisted paraproducts, whose logarithmic growth in the shift parameters is tamed by shifted square-function estimates. A fiberwise Calderón–Zygmund decomposition then extends the strong L^p bounds down to th

Load-bearing premise

The exponential decay of the high-frequency part is imported as a black box from a companion Fourier multiplier theorem; if that theorem fails or its hypotheses are not fully met by this symbol class, the decay and both main theorems collapse.

What would settle it

Compute the multiplicative-derivative norms appearing in the companion multiplier theorem for the specific symbol xK_0(ξ,η)ψ(2^{-j}ξ)ψ(2^{-k}η) with mean-zero Ω∈L^q(S^1), q>1; a single symbol with infinite norm would invalidate Proposition 2.2. Alternatively, for a non-smooth Ω such as Ω(θ)=sgn(θ1) on a small arc, evaluate the single-scale high-frequency norm ratio 2^{δj}||T^1_j|| against j and check whether it stays bounded as j grows.

Watch this falsifier. Get emailed when new claim-graph text bears on it.

If this is right

  • The averaged operators A_Ω are bounded for every odd Ω∈L^q(S^1), q>1, for all 1≤p<2, even though the individual directional Hilbert transforms H_θ are not known to be bounded.
  • Since a difference of point masses recovers a single directional Hilbert transform, the results show that sufficiently rich averaging over directions regularizes these operators into bounded ones.
  • For p<1, the condition 1/p+1/q<2 is a complete iff characterization, so any future extension of the quasi-Banach range must either weaken the kernel class or improve the balance.
  • The high-frequency part is bounded for p>1 regardless of the smooth-kernel twisted paraproduct, and actually holds for a wider range than is currently known for smooth kernels.
  • The counterexamples work near any line through the origin (θ1=βθ2), not just the diagonal, so the failure at the boundary is not a special artifact.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • If the companion smoothing inequality were replaced by the wavelet-decomposition route the paper sketches, one might obtain explicit constants and possibly a cleaner proof for p>1; this is not claimed by the paper.
  • The balance condition 1/p+1/q<2 is also the known threshold for rough Coifman–Meyer operators, suggesting a general principle for rough homogeneous bilinear singular integrals; the paper does not make this comparison.
  • The logarithmic shift estimates are likely not optimal in the second variable; sharpening them could transfer to improved bounds for double and cubic ergodic averages, a connection the paper mentions but does not pursue.
  • One could test whether the p≥1 range extends beyond p<2 if the smooth-kernel twisted paraproduct is ever shown bounded for larger p; the paper's reduction indicates the high-frequency part would not obstruct such an extension.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit.

Referee Report

2 major / 5 minor

Summary. The paper studies bilinear operators obtained by averaging directional variants of the triangular Hilbert transform over the circle with respect to a rough odd weight Ω. Equivalently, these are twisted paraproducts whose symbol is the Fourier transform of the rough homogeneous kernel Ω(x/|x|)/|x|^2. The main results are Theorem 1.1, giving L^{p_1}×L^{p_2}→L^p bounds for all 1<p_1,p_2<∞, 1≤p<2 whenever Ω∈L^q(S^1), q>1, and Theorem 1.2, establishing equivalence between the balance condition 1/p+1/q<2 and boundedness of T_Ω for all mean-zero Ω∈L^q(S^1) in the quasi-Banach range 0<p<1. The proof decomposes T_j into low-frequency, diagonal high-frequency, and off-diagonal mid-frequency pieces. The low-frequency piece uses the Calderón–Zygmund twisted paraproduct, the diagonal piece is controlled by a smoothing inequality imported from the authors' preprint [23], the mid-frequency piece is reduced to shifted twisted paraproduct estimates, and the quasi-Banach range is handled by a fiberwise Calderón–Zygmund argument. Sharpness is shown by explicit test functions for the condition 1/p+1/q≥2.

Significance. If the results are correct, they are a notable advance on a difficult class of bilinear singular integrals. The sharp balance condition 1/p+1/q<2 for p<1 is new for the twisted paraproduct with rough kernels and matches the expected analogy with rough Coifman–Meyer operators. The shifted twisted paraproduct estimates in Section 4 are also of independent interest and give a clean structural reduction to the smooth Calderón–Zygmund case. The counterexamples in Section 6 are explicit, self-contained, and internally consistent; no fitted constants or ad hoc exclusions appear. The main reservation is the black-box use of the Hsu–Lin smoothing inequality [23, Theorem 1.6], which is a preprint by one of the present authors and is not proved in the paper. Since Proposition 2.2 is the only source of exponential j-decay for the diagonal piece, this dependency is load-bearing. The alternative wavelet route is only sketched in Remark 3.4 and is not a complete substitute. With a full proof of Theorem 3.2, or a full wavelet proof of Lemma 3.1, the central claims would be substantially supported.

major comments (2)
  1. [Section 3, Proposition 2.2 and Lemma 3.1] The exponential decay estimate (2.7) for the diagonal high-frequency piece T^1_j is obtained entirely by applying Theorem 3.2, i.e. [23, Theorem 1.6], as a black box. Corollary 3.3 derives from it the decay (3.8) for the compactly supported symbol m; without Theorem 3.2 the estimate (3.9) has no basis. Since [23] is an unreviewed preprint by one of the present authors, and since the paper neither proves the u- and U-norm bounds (3.4)–(3.5) directly nor supplies the full proof of Theorem 3.2, the central summability in j rests on an externally imported result. Remark 3.4 explicitly says only that 'one may prove' (3.11), so it is not a substitute. This must be resolved: either include a complete proof of Theorem 3.2, or give a complete wavelet-based proof of Lemma 3.1, or otherwise verify the hypotheses for the specific symbol class used.
  2. [Section 5, proof of Proposition 2.6] The interpolation step in Proposition 2.6 relies on applying Lemma 5.1 with p_0=p/(1-p+ε). Under the standing assumption 1/p+1/q<2 one necessarily has p>1/2, so for sufficiently small ε one indeed has p_0>1 and the lemma applies; however, the manuscript does not spell out that 1/p+1/q<2 forces p>1/2. This is a minor omission, not a fatal gap, but the reader should be told explicitly that the argument only needs to cover p>1/2 and why the choice of ε is possible under the assumption.
minor comments (5)
  1. [Abstract / title] The title contains a typo: 'Hilber T' should be 'Hilbert transform'.
  2. [Section 3, Remark 3.4] The remark announces an alternative wavelet proof but does not provide it. Either move it to a 'further directions' paragraph or supply the proof; in its current form it should not be cited as a completed alternative.
  3. [Section 4.2, Lemma 4.3] In the proof, the assertion that c(1+|v|)^{-L}K_v is a Calderón–Zygmund kernel of order L is compressed. The verification is routine but a sentence indicating how the factor (1+|v|)^L arises from the oscillation term e^{-2πi2^k vη} would improve readability.
  4. [Section 5, proof of Proposition 2.6] The statement 'by lowering ε further, we ensure that the point lies inside the triangle' is plausible but not demonstrated. A short algebraic check with the three vertices would make the interpolation step transparent, especially since the reader must reconcile the quasi-Banach target exponent with the bilinear Marcinkiewicz theorem.
  5. [Section 6] In the limiting case 1/p+1/q=2, the authors choose α>1 so that 1/p+1-2α>0. Since this is only possible when p<1, it would help to state explicitly why such an α exists and that the constructed Ω depends on q through the exponents in (6.1).

Circularity Check

0 steps flagged

No significant circularity: the proof imports a general smoothing theorem from a self-authored preprint, but the theorem is not equivalent to the claimed bounds and no fitted or self-defined quantity drives the conclusions.

full rationale

The derivation chain is a genuine decomposition argument. The only place one might suspect circularity is Proposition 2.2, whose proof is described as 'based on a bilinear Fourier multiplier theorem from [23]', namely Theorem 3.2 ('Hsu, Lin [23, Theorem 1.6]'). This is a load-bearing citation of a preprint by the present first author, and the paper does not reproduce the proof of Theorem 3.2; Remark 3.4 only sketches an alternative wavelet route ('one may prove'). That makes the argument conditional on an external verification, but it is not circular. Theorem 3.2 is a parameter-free assertion about general bilinear Fourier integral operators, with hypotheses (3.4)-(3.5) that do not mention the twisted paraproduct, the target exponents, or the kernel Omega; Corollary 3.3 and Lemma 3.1 show how the abstract estimate yields the desired 2^{-delta j} decay. Thus the cited result is independent content, albeit from a self-authored preprint. The other main steps are internally proved: the low-frequency Proposition 2.1 uses the Calderon-Zygmund twisted paraproduct bounds [24] and the derivative estimate (2.3); the mid-frequency Proposition 2.4 is proved via shifted kernels and Lemma 4.2; Proposition 2.6 is a fiberwise Calderon-Zygmund decomposition; and the counterexample Proposition 2.7 is computed explicitly from test functions (6.2)-(6.7). No parameter is fitted to data, no operator is defined in terms of the target bound, and no uniqueness theorem or ansatz is smuggled in through self-citation. The observed self-citation ([23], and the incidental reference [28]) is therefore a dependency/verification risk rather than circularity.

Axiom & Free-Parameter Ledger

0 free parameters · 6 axioms · 0 invented entities

The paper contributes no free parameters or invented entities; all constants are structural. The main external dependencies are the established Kovač/Muscalu/interpolation machinery and the Hsu–Lin preprint [23], which carries the high-frequency part and is the least vetted item in the ledger.

axioms (6)
  • standard math Kovač's theorem: the twisted paraproduct with any Calderón–Zygmund kernel of order L is bounded for 1<p1,p2<∞, 0<p<2 with 1/p=1/p1+1/p2 [24].
    Used in Proposition 2.1 for the low-frequency part and in Lemma 4.3 for the shifted twisted paraproducts (Section 4.2).
  • domain assumption Hsu–Lin smoothing inequality [23, Theorem 1.6]: for fiberwise bilinear Fourier integral operators, ||T_m||_{L^2×L^2→L^1} is controlled by mixed multiplicative-derivative norms of the symbol.
    Imported without proof from a preprint by the present author Lin; it drives Proposition 2.2 through Lemma 3.1 (Section 3).
  • standard math Muscalu's shifted square function estimate (4.1): the ℓ^2 square function over shifted Littlewood–Paley projections is bounded by a logarithmic factor log(2+|v|) times the L^r norm.
    Used in Lemmas 4.1 and 4.3 to obtain logarithmic-in-shift estimates for shifted kernels.
  • standard math Multilinear Marcinkiewicz interpolation theorem [20].
    Used in the proofs of Proposition 2.3 and Proposition 2.6 to interpolate between endpoint estimates.
  • standard math Duoandikoetxea decay estimate (2.1) for Fourier transforms of rough homogeneous kernels: |\hat K_0(ξ,η)| ≤ C||Ω||_{L^q}|(ξ,η)|^{-δ} for δ<1/q'.
    Used throughout to control the symbol of the kernel in the frequency decomposition.
  • standard math Bernicot's fiberwise Calderón–Zygmund decomposition [2], with the adaptations of He–Park [21] and Bhojak–Shrivastava [3].
    Used in Section 5 to prove the weak-type estimates needed for the p<1 extension.

reviewed 2026-08-01 · how reviews work

0 comments
Cite this review

Pith. "Pith review of Rough averages of triangular Hilbert transforms." pith.science (2026). https://pith.science/paper/XKVCW6KQ

@misc{pith2026260720206,
  author       = {Pith},
  title        = {Pith review of: Rough averages of triangular Hilbert transforms},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/XKVCW6KQ}},
  note         = {Machine review of arXiv:2607.20206}
}
Share X Bluesky LinkedIn Reddit HN
read the original abstract

We study a bilinear singular integral operator obtained by taking rough averages of certain directional variants of the triangular Hilbert transform. This operator can be interpreted as the twisted paraproduct with a rough homogeneous kernel. Under a balance condition on the $L^q$ integrability of the kernel on the unit sphere, we establish a range of $L^{p_1} \times L^{p_2} \rightarrow L^p$ bounds for this operator. Our results are optimal within the known boundedness range for the twisted paraproduct with a smooth kernel.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

32 extracted references · 4 linked inside Pith

  1. [1]

    Becker, P

    L. Becker, P. Durcik,The shifted bilinear Hilbert transform, preprint, arXiv:2603.20173

    arXiv

  2. [2]

    Bernicot,Fiber-wise Calder´ on-Zygmund decomposition and application to a bi- dimensional paraproduct, Illinois J

    F. Bernicot,Fiber-wise Calder´ on-Zygmund decomposition and application to a bi- dimensional paraproduct, Illinois J. Math.56(2012), no. 2, 415–422

    2012

  3. [3]

    Bhojak, S

    A. Bhojak, S. Shrivastava,An alternate approach to bilinear rough singular integrals, preprint, arXiv:2508.19181

    Pith/arXiv arXiv

  4. [4]

    Carleson,On convergence and growth of partial sums of Fourier series, Acta Math.116 (1966), 135–157

    L. Carleson,On convergence and growth of partial sums of Fourier series, Acta Math.116 (1966), 135–157

    1966

  5. [5]

    Christ,A class of singular bilinear maximal functions, J

    M. Christ,A class of singular bilinear maximal functions, J. Funct. Anal.287(8), (2024)

    2024

  6. [6]

    Christ, P

    M. Christ, P. Durcik, J. Roos,Trilinear smoothing inequalities and a variant of the triangular Hilbert transform, Adv. Math.390, (2021)

    2021

  7. [7]

    Dosidis, L

    G. Dosidis, L. Slav ´ ıkov´ a,Multilinear singular integrals with homogeneous kernels nearL 1, Math. Ann.389(2024), no. 3, 2259–2271

    2024

  8. [8]

    Duoandikoetxea,Fourier analysis, Graduate Studies in Mathematics,29, American Math- ematical Society, Providence, RI, 2001

    J. Duoandikoetxea,Fourier analysis, Graduate Studies in Mathematics,29, American Math- ematical Society, Providence, RI, 2001

    2001

  9. [9]

    Durcik, V

    P. Durcik, V. Kovaˇ c, K. A.ˇSkreb, C. Thiele,Norm variation of ergodic averages with respect to two commuting transformations, Ergodic Theory Dynam. Systems39(2019), no. 3, 658– 688

    2019

  10. [10]

    Durcik, V

    P. Durcik, V. Kovaˇ c, C. Thiele,Power-type cancellation for the simplex Hilbert transform, J. Anal. Math.139(2019), no. 1, 67–82

    2019

  11. [11]

    Durcik, J

    P. Durcik, J. Roos,Averages of simplex Hilbert transforms, Proc. Amer. Math. Soc.149 (2021), no. 2, 633–647

    2021

  12. [12]

    Durcik, K

    P. Durcik, K. A. ˇSkreb,Norm-variation of cubic ergodic averages, J. Math. Anal. Appl.562 (2026), no. 2, Paper No. 130689, 22 pp

    2026

  13. [13]

    Fefferman,Pointwise convergence of Fourier series, Ann

    C. Fefferman,Pointwise convergence of Fourier series, Ann. of Math. (2)98(1973), no. 3, 551–571

    1973

  14. [14]

    Gaitan, V

    A. Gaitan, V. Lie,Non-zero to zero curvature transition: Operators along hybrid curves with no quadratic (quasi-)resonancesAdv. Math.478, (2025). ROUGH A VERAGES OF TRIANGULAR HILBERT TRANSFORMS 25

    2025

  15. [15]

    Grafakos,Modern Fourier Analysis.,3rd Ed., GTM250, Springer, 2014

    L. Grafakos,Modern Fourier Analysis.,3rd Ed., GTM250, Springer, 2014

    2014

  16. [16]

    Grafakos, D

    L. Grafakos, D. He, P. Honz ´ ık,Rough bilinear singular integrals, Adv. Math.326(2018), 54–78

    2018

  17. [17]

    Grafakos, D

    L. Grafakos, D. He, P. Honz ´ ık, B. Park,Multilinear rough singular integral operators, J. Lond. Math. Soc. (2)109(2024), no. 2, Paper No. e12867, 35 pp

    2024

  18. [18]

    Grafakos, D

    L. Grafakos, D. He, L. Slav ´ ıkov´ a,L2ˆL 2 ÑL 1 boundedness criteria, Math. Ann.376 (2020), no. 1–2, 431–455

    2020

  19. [19]

    Grafakos and X

    L. Grafakos and X. Li,Uniform bounds for the bilinear Hilbert transforms, I, Ann. of Math. (2)159(2004), no. 3, 889–933

    2004

  20. [20]

    Grafakos, L

    L. Grafakos, L. Liu, S. Lu, and F. Zhao,The multilinear Marcinkiewicz interpolation theorem revisited: The behavior of the constant, J. Funct. Anal.262(2012) 2289–2313

    2012

  21. [21]

    D. He, B. J. Park,Improved estimates for bilinear rough singular integrals, Math. Ann.386 (2023), no. 3–4, 1951–1978

    2023

  22. [22]

    Honz ´ ık, S

    P. Honz ´ ık, S. Lappas, L. Slav ´ ıkov´ a,Bilinear singular integral operators with kernels in weighted spaces, preprint, arXiv:2412.07014

    Pith/arXiv arXiv

  23. [23]

    M. Hsu, F. Lin,Smoothing inequalities for corner-type bilinear averages: geometric charac- terization and applications, preprint, arXiv:2410.15791

    Pith/arXiv arXiv

  24. [24]

    Kovaˇ c,Boundedness of the twisted paraproduct, Rev

    V. Kovaˇ c,Boundedness of the twisted paraproduct, Rev. Mat. Iberoam.28(2012), no.4, 1143–1164

    2012

  25. [25]

    Kovaˇ c, C

    V. Kovaˇ c, C. Thiele, P. Zorin-Kranich,Dyadic triangular Hilbert transform of two general functions and one not too general function, Forum Math. Sigma3(2015), Paper No. e25, 27 pp

    2015

  26. [26]

    Lacey, C

    M. Lacey, C. Thiele,A proof of boundedness of the Carleson operator, Math. Res. Lett.7 (2000), no. 4, 361–370

    2000

  27. [27]

    Li,Uniform bounds for the bilinear Hilbert transforms, II, Rev

    X. Li,Uniform bounds for the bilinear Hilbert transforms, II, Rev. Mat. Iberoam.22(2006), no. 3, 1069–1126

    2006

  28. [28]

    Lin,On the family of singular Brascamp-Lieb inequalities with dimension datump1,2,2,1q, Mathematika71(2025), no

    F. Lin,On the family of singular Brascamp-Lieb inequalities with dimension datump1,2,2,1q, Mathematika71(2025), no. 1, Paper No. e70003, 13 pp

    2025

  29. [29]

    Muscalu,Calder´ on Commutators and the Cauchy integral on Lipschitz curves revisited: I

    C. Muscalu,Calder´ on Commutators and the Cauchy integral on Lipschitz curves revisited: I. First commutator and generalizations, Rev. Mat. Iberoam.30(2014) 1413–1437

    2014

  30. [30]

    Thiele,A uniform estimate, Ann

    C. Thiele,A uniform estimate, Ann. of Math. (2)156(2002), no. 2, 519–563

    2002

  31. [31]

    Uraltsev and M

    G. Uraltsev and M. Warchalski,The full range of uniform bounds for the bilinear Hilbert transform, arXiv:2205.09851, 2022

    Pith/arXiv arXiv 2022

  32. [32]

    Zorin-Kranich,Cancellation for the simplex Hilbert transform, Math

    P. Zorin-Kranich,Cancellation for the simplex Hilbert transform, Math. Res. Lett.24(2017), no. 2, 581–592. Department of Mathematics, Purdue University, Mathematical Sciences Bldg, 150 N University St, West Lafayette, IN 47907 USA Email address:lin2311@purdue.edu Department of Mathematical Analysis, F aculty of Mathematics and Physics, Charles University,...

    2017

This paper was first reviewed by deepseek-v4-flash on August 1, 2026.