REVIEW 1 major objections 7 minor 23 references
The Hilbert transform along the parabola, the polynomial Carleson theorem and oscillatory singular integrals
T0 review · 1 major / 7 minor · reviewed 2026-08-14 · deepseek-v4-flash
Pith's one-line read The paper proves uniform L^p bounds for maximal modulations of the Hilbert transform along the parabola when restricted to lines.
desk verdict A genuinely new and plausible line-restricted bound for the parabolic Carleson problem, with one external step that needs a referee's scrutiny. 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 load-bearing object is the linearized phase operator $C_R$: via Kolmogorov--Seliverstov linearization, the suprema over $N$ and $b$ become a single operator with measurable functions $N(x)$ and $b(x)$, and the phase is $b(x)[t+1]^{1/2}$. The proof splits $|t|$ into dyadic scales and applies two mechanisms: for $|t|\le b(x)^{-1/6}$ it replaces the phase by its degree-five Taylor polynomial $P_{b(x)}(t)=b(x)\left(1+\tfrac{t}{2}-\tfrac{t^2}{4}+\tfrac{3t^3}{8}-\tfrac{15t^4}{16}+\tfrac{105t^5}{32}\right)$, reducing the operator to a maximally truncated polynomial Carleson operator of degree at most five; for larger $|t|$ it uses the $TT^*$ method, bounding $S_j(S_j)^*$ through oscillatory integrals whose phase is controlled by the vector $Q=(\varphi'',-\tfrac{2}{3}\varphi''')$, represented as a matrix times a vector so that determinant and norm bounds yield stationary-phase decay like $2^{-j/200}$.
What would settle it
The decisive check is the stationary-phase estimate in Lemma 2 at the boundary $|\xi'| = 2^{-j/100}$, $h=1$, with $b(x)=b(y)=10$: the determinant lower bound used to prove $|Q(s')| \gtrsim 2^{j/3}$ must have an implied constant independent of $h$; a direct computation showing the constant grows with $j$ would make the claimed $2^{-j/9}$ decay false and would collapse the proof of Theorem 2.
Extended reading notes
Core claim
The paper's central discovery is that maximal modulations of each one-dimensional restriction of the parabolic Hilbert transform are uniformly $L^p$-bounded. In symbols, for the multiplier $m_2$ of the parabolic Hilbert transform and its line restrictions $m_{a,b}(\eta)=m_2(a\eta+b,\eta)$ (with $m_{+\infty,b}(\eta)=m_2(\eta,b)$), the operators $C_{a,b}f(x)=\sup_{N\in\mathbb{R}}|\mathcal{F}^{-1}(m_{a,b}\widehat{M_N f})(x)|$ satisfy $\sup_a\|\sup_b C_{a,b}\|_{p\to p}<\infty$ for $1<p<\infty$. The route is a reduction (Proposition 1) to the truncated operators $C_R f(x)=\sup_{N,b}\left|\int_{-R}^R f(x-t)e^{iN t}e^{ib[t+1]^{1/2}}\frac{dt}{t}\right|$, followed by a proof (Theorem 2) that these are bounded independently of $R$: on a shrinking interval around the origin the phase is compared to a degree-5 Taylor polynomial, so the polynomial Carleson theorem applies, while on the remaining intervals oscillation is strong enough that $TT^*$ and stationary phase give summable dyadic decay.
Load-bearing premise
The load-bearing premise is that the maximally truncated polynomial Carleson theorem covers degree-5 operators whose coefficients are measurable functions of $x$ (the linearized phase $b(x)$), after the phase $b(x)\sqrt{t+1}$ is replaced by its Taylor polynomial on $[-b(x)^{-1/6}, b(x)^{-1/6}]$; if that theorem does not apply to measurable-coefficient operators, or if the error estimate (9) fails uniformly, Theorem 2 does not follow.
Editorial extensions
If this is right
- Theorem 1 gives uniform $L^p$ bounds for maximal modulations along every line in the multiplier plane, so the constant in the $L^p$ estimate does not depend on the slope or intercept of the line.
- If the same argument can be extended from lines to arbitrarily thin strips with constants independent of width, the limit argument in Section 4 would yield $L^2$ bounds for the full parabolic Carleson operator $C_2$.
- The dyadic split at $b(x)^{-1/6}$ and the degree-five Taylor phase make concrete the paper's suggested principle: maximal operators with phases $N\cdot t + b\,\eta(t+1)$ should be $L^p$ bounded whenever $\eta$ is smooth away from the origin with controlled derivatives.
- For monomial curves $(t,t^m)$, the reductions are expected to carry through with phase $[t+1]^{1/m}$, giving line-restricted uniform bounds for higher-order parabolic Hilbert transforms.
- Proposition 5 shows a fixed-degree polynomial approximation of $\sqrt{t+1}$ over a fixed interval forces the degree to grow with $b$, so the shrinking interval in the proof is not an artifact: some cutoff is necessary in this approach.
Reading between the lines
- A natural next test is whether the two-regime split persists in a full time-frequency proof of the open parabolic Carleson question; the line restriction may be exactly the regime where a one-dimensional polynomial Carleson theorem suffices, and a two-dimensional analogue would be needed only for genuinely two-dimensional strips.
- The asymptotic threshold $d\ge 4$ suggests a trade-off between polynomial degree and interval size: choosing a higher-degree Taylor polynomial shrinks the Carleson interval and moves more of the analysis into the oscillatory regime, which might be optimizable for better decay exponents.
- One could test the conjectured general principle numerically for smooth compactly supported $f$ and slowly varying $b(x)$: if the constants in the model operator $C_R$ drift with $R$ in simulations, the uniform $R$ statement in Theorem 2 would be suspect even though the proof's dyadic bounds appear summable.
Signed reviews
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. This paper proves uniform Lp estimates (1<p<∞) for the maximal modulation operators C_{a,b}f(x)=sup_N |T_{a,b}(e^{iN⋅}f)(x)|, uniformly in the line parameter a (including a=+∞) and in the intercept b, where T_{a,b} are the one-dimensional operators whose multipliers are the restriction of the parabolic Hilbert transform multiplier m_2(ξ,η) to the line (aη+b,η). The proof proceeds in two steps: Proposition 1 reduces the original two-parameter problem to the boundedness of a model operator C^R with phase b(x)[t+1]^{1/2} and a linear modulation e^{iN(x)t}; Theorem 2 then proves the required uniform bound for C^R. The proof of Theorem 2 splits into a small-b regime (b(x)≤10), handled by the classical maximally truncated Carleson theorem plus TT* estimates with exponential decay in dyadic scales, and a large-b regime (b(x)>10), where the phase is replaced by its degree-five Taylor polynomial on the interval |t|≤b(x)^{-1/6}, invoking the polynomial Carleson theorem, while the remaining layers are treated by TT* estimates. The paper also proves (Proposition 5) that no fixed finite-degree polynomial approximation can replace the Taylor step uniformly in b, so the degree-five choice is in a sense necessary.
Significance. If the proof is correct, this is a substantive advance on the parabolic Carleson problem of Question 1: it gives uniform Lp bounds for maximal modulations of the full family of line restrictions of the parabolic Hilbert transform, going beyond the partial results of Roos [18] and the restricted-phase results of Guo–Pierce–Roos–Yung [9]. The method is a clean and instructive combination of the polynomial Carleson theorem with TT* oscillatory estimates, and the explicit exponential decay in Propositions 3 and 4 is presented in sufficient detail to be checkable. Proposition 5 is an elegant and rigorous obstruction result. The paper is honest about its reliance on external results, though that reliance should be spelled out more precisely.
major comments (1)
- [Section 3, Part 2, Eq. (9)] The claim that the local integral in Eq. (9) is bounded by a maximally truncated polynomial Carleson operator of degree ≤5 in the sense of [14,23] is load-bearing for Theorem 2 and is not fully verified. The manuscript does not quote the exact theorem it relies on. Please state the theorem (or a precise corollary) and explicitly check that it applies to: (i) the x-dependent truncation radius b(x)^{-1/6} (if the cited theorem is stated only for constant truncation R, the standard reduction to x-dependent cutoffs must be supplied); (ii) polynomial coefficients that are measurable functions of x after the Kolmogorov–Seliverstov linearization; (iii) the full phase P_{b(x)}(t)+N(x)t of degree 5, with maximal truncation. Without this verification, Eq. (9) leaves a gap in the proof of Theorem 2.
minor comments (7)
- [Title and Abstract] There are several typographical errors ('Hilber t', 'P ar abola', 'POL YNOMIAL', 'OSCILLA TOR Y', 'Ackowledgements'); please correct them.
- [Section 3, Part 2] The displayed definition of φ_{b(x)}(t) has garbled summation limits; please rewrite it with clear bounds, for instance j ranging from (2−1/6)⌊log_2 b(x)⌋ to 2⌊log_2 b(x)⌋−3 (or whatever is intended).
- [Near Eq. (9)] The Taylor remainder bound is stated for t∈[−1/2,1/2], but the integration interval is [−b(x)^{-1/6}, b(x)^{-1/6}], which exceeds [−1/2,1/2] when b(x)<64. Since b(x)^{-1/6}<1, the estimate still holds with an absolute constant, but the interval should be adjusted and the constant justified.
- [Lemma 2] The reference 'Proposition 2 in Chapter VIII of [20]' should be given a precise number or title to avoid collision with Proposition 2 of this paper.
- [Proposition 2] The phrase 'both bounded in Lp' should explicitly state that the two choices are [u]^{1/2}=|u|^{1/2} and sign(u)|u|^{1/2}.
- [References] References [14] and [23] are listed as preprints; if published versions exist, they should be updated.
- [Section 2] Several changes of variables in the reduction are stated without showing Jacobians and interval endpoints; expanding them would improve readability, though I did not find an error in the stated equivalences.
Circularity Check
No significant circularity: Theorem 1 is derived from Theorem 2, which is proved using independent external Carleson and polynomial Carleson theorems.
full rationale
The paper's derivation chain is linear rather than circular. Theorem 1 is reduced in Section 2 to uniform Lp bounds for the operators C_R defined in (4), and Proposition 1 explicitly assumes those bounds as a hypothesis. Theorem 2 then proves the bounds using a Kolmogorov-Seliverstov linearization, a dyadic decomposition, TT* estimates for oscillatory integrals, and the polynomial Carleson theorems of Lie [14] and Zorin-Kranich [23]. The use of those theorems is not a self-citation: Lie and Zorin-Kranich are external authors, and the cited results are independent, parameter-free theorems. In Part 2, after replacing the phase b(x)sqrt(t+1) by the degree-five Taylor polynomial P_{b(x)}(t), the resulting operator is genuinely of polynomial Carleson type: at each x the phase is a degree-at-most-five polynomial, so the pointwise value is dominated by the supremum over all such polynomials; the constant term of P_{b(x)} only contributes a harmless phase factor. No parameter is fitted from the target conclusion, no quantity called a prediction is actually an input in disguise, and no load-bearing step reduces to an assumption of Theorem 1 itself. The limiting argument in Proposition 2 uses Theorem 2 for the untruncated limit properly, and Proposition 5 is an independent impossibility remark rather than a disguised assumption. Possible concerns about the exact scope of the maximally truncated polynomial Carleson theorem or about the uniformity constants are correctness or verification risks, not circularity. The paper contains no self-citations and no definitional equivalence between its conclusions and its hypotheses, so the appropriate circularity score is 0.
Assumptions & free parameters
assumptions (6)
- domain assumption Maximally truncated polynomial Carleson theorem is Lp bounded for all p in (1, infinity).
- standard math Classical maximally truncated Carleson operator is Lp bounded.
- standard math Hardy-Littlewood maximal function is bounded on Lp for 1 < p <= infinity.
- standard math Stationary phase estimates for oscillatory integrals with lower bounds on second or third derivatives.
- standard math Marcinkiewicz interpolation theorem.
- standard math Lp boundedness of the Hilbert transform and Fourier inversion on the Schwartz class.
Cite this review
Pith. "Pith review of The Hilbert transform along the parabola, the polynomial Carleson theorem and oscillatory singular integrals." pith.science (2026). https://pith.science/paper/CDFTVVK4
@misc{pith2026190801833,
author = {Pith},
title = {Pith review of: The Hilbert transform along the parabola, the polynomial Carleson theorem and oscillatory singular integrals},
year = {2026},
howpublished = {\url{https://pith.science/paper/CDFTVVK4}},
note = {Machine review of arXiv:1908.01833}
}
abstract
We make progress on an interesting problem on the boundedness of maximal modulations of the Hilbert transform along the parabola. Namely, if we consider the multiplier arising from it and restrict it to lines, we prove uniform $L^p$ bounds for maximal modulations of the associated operators. Our methods consist of identifying where to use effectively the polynomial Carleson theorem, and where we can take advantage of the presence of oscillation to obtain decay through the $TT^*$ method.
Reference graph
Works this paper leans on
-
[18]
Roos, Bounds for anisotropic Carleson operators
J. Roos, Bounds for anisotropic Carleson operators. J. Fourier Anal. Appl., Online version (10-Dec- 2018), 1–32
work page 2018
-
[9]
S. Guo, L. Pierce, J. Roos and P.L. Yung, Polynomial Carleson operators along monomial curves in the plane. J. Geom. Anal., 27 (2017), n. 4, 2977–3012
work page 2017
-
[1]
Carleson, On convergence and growth of partial sums of Fourier series
L. Carleson, On convergence and growth of partial sums of Fourier series. Acta Math., 116 (1966), 135–157
work page 1966
-
[2]
Christ, Hilbert Transforms Along Curves: I
M. Christ, Hilbert Transforms Along Curves: I. Nilpotent Groups. Ann. Math., 122 (1985), n. 3, 575–596
work page 1985
- [3]
- [4]
-
[5]
Fefferman, ointwise convergence of Fourier series
C. Fefferman, ointwise convergence of Fourier series. Ann. Math., 98 (1973), n. 3, 551–571
work page 1973
-
[6]
Grafakos, Modern Fourier Analysis, 3rd edition
L. Grafakos, Modern Fourier Analysis, 3rd edition. Springer-Verlag New York (2014)
work page 2014
Show all 23 references
-
[7]
Guo, A remark on oscillatory integrals associated with fewnomia ls
S. Guo, A remark on oscillatory integrals associated with fewnomia ls. New York J. Math., 23 (2017), 1733–1738. HILBERT TRANSFORM ALONG THE PARABOLA 19
2017
-
[8]
Guo, Oscillatory integrals related to Carleson ’s theorem: frac tional monomials
S. Guo, Oscillatory integrals related to Carleson ’s theorem: frac tional monomials. Comm. Pure Appl. Anal., 15 (2016), no. 3, 929–946
2016
-
[10]
S. Guo, J. Hickman, V. Lie and J. Roos, Maximal operators and Hilbert transforms along variable non-flat homogeneous curves. Proc. London Math. Soc., 115 (2017), n. 1, 177–219
2017
-
[11]
Hunt, On the convergence of Fourier series
R. Hunt, On the convergence of Fourier series. Orthogonal Expansions and their Contin-uous Ana- logues (Proceedings of Conference, Edwardsville,Illinoi s, 1967) (1968), 235–255
1968
-
[12]
Lacey, C
M. Lacey, C. Thiele, A proof of boundedness of the Carleson operator. Math. Res. Lett., 7 (2000), 361–370
2000
-
[13]
Lie, The (weak-L2) boundedness of the quadratic Carleson operator
V. Lie, The (weak-L2) boundedness of the quadratic Carleson operator. Geom. Funct. Anal., 19 (2009), n. 2, 457–497
2009
-
[14]
Lie, The polynomial Carleson operator
V. Lie, The polynomial Carleson operator. preprint available at arXiv:1105.4504v3
-
[15]
Nagel, N
A. Nagel, N. Rivi´ ere, S. Wainger, On Hilbert transforms along curves. Bull. Amer. Math. Soc., 80 (1974), n. 1, 106–108
1974
-
[16]
Nagel, N
A. Nagel, N. Rivi´ ere, S. Wainger, On Hilbert Transforms Along Curves II. Amer. J. Math., 98 (1976), n. 2, 395–403
1976
-
[17]
Pierce, P.L
L. Pierce, P.L. Yung, A polynomial Carleson operator along the paraboloid. Rev. Math. Iberoam., 35 (2019), n. 2, 339–422
2019
-
[19]
Seeger, T
A. Seeger, T. Tao, J. Wright, Singular maximal functions and Radon transforms near L1. Amer. J. Math., 126 (2004), n. 3, 607–647
2004
-
[20]
Stein, Harmonic Analysis: Real variable methods, Orthogonality, and Oscilatory integrals, Prince- ton University Press, Princeton, NJ, 1993
E. Stein, Harmonic Analysis: Real variable methods, Orthogonality, and Oscilatory integrals, Prince- ton University Press, Princeton, NJ, 1993
1993
-
[21]
Stein, Oscillatory integrals related to Radon-like transforms
E. Stein, Oscillatory integrals related to Radon-like transforms. In Proceedings of the Conference in Honor of Jean-Pierre Kahane, Orsay, Special Issue (1993), pp. 535–551
1993
-
[22]
Stein, S
E. Stein, S. Wainger, Oscillatory integrals related to Carleson ’s theorem. Math. Res. Lett., 8 (2001), 789–800
2001
-
[23]
Zorin-Kranich, Maximal polynomial modulations of singular integrals
P. Zorin-Kranich, Maximal polynomial modulations of singular integrals. preprint available at arXiv:1711.03524v5 Mathematisches Institut der Universit ¨at Bonn, Endenicher Allee 60, 53115 Bonn, Ger- many E-mail address : joaopgramos95@gmail.com
Reviewed August 14, 2026 · model on record in the stance chip above.
Discussion (0). Continue with ORCID to comment.