Some sharp $L^2 \to L^p$ decay estimates for $(2+1)$-dimensional degenerate oscillatory integral operators

Shaozhen Xu

REVIEW 1 major objections 4 minor 20 references

Some sharp $L^2 \to L^p$ decay estimates for $(2+1)$-dimensional degenerate oscillatory integral operators

T0 review · 1 major / 4 minor · reviewed 2026-08-12 · deepseek-v4-flash

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

Pith's one-line read Sharp decay rate proven for (2+1)-dimensional degenerate oscillatory operators

desk verdict New theorem, plausible and likely true, but the proof's key L2→L2 endpoint is asserted via an unspecified 'iteration' of a Phong–Stein result and needs real work before the paper is solid. read the letter →

arxiv 2411.15009 v1 pith:QWJHTSU4 submitted 2024-11-22 math.CA

Shaozhen Xu This is my paper · ORCID

classification math.CA MSC 42B20

keywords oscillatoryintegraloperatorsdegeneratephaseL^2toL^pestimatessharpdecaySteincomplexinterpolationPhong-SteintheoremvanderCorputlemmapolynomial

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

This paper proves a sharp $L^2\to L^{2m+2}$ decay estimate for oscillatory integral operators on $\mathbb{R}^2$ whose phase is the polynomial $x^a t^m + y^b t^n$ with $m>n$. The bound decays like $\lambda^{-(1/a+1/\max\{b,n\})/(2(m+1))}$ as the frequency parameter $\lambda$ grows. When $n\le b$, the paper also constructs an example showing the rate cannot be improved. This extends the Stein–Tomas style of argument, which originally gave sharp decay for nondegenerate phases, into a degenerate higher-dimensional setting.

What carries the argument

The machinery is Stein's complex interpolation, applied to an analytic family $T_z$ of operators whose kernel is built from the analytic continuation $\delta_z$ of the Dirac mass. At $\operatorname{Re} z=1$ the family factors as $I_1\circ I_2$, a composition of two $(1+1)$-dimensional oscillatory integral operators, so iterating the Phong–Stein $L^2$ decay theorem [PS94] yields the $L^2\to L^2$ bound (2.2). At $\operatorname{Re} z=-1/m$ the van der Corput lemma gives a uniform $L^1\to L^\infty$ bound. Interpolating between these two endpoints produces the $L^{(2m+2)/(2m+1)}\to L^{2m+2}$ estimate for the $TT^*$ kernel, which dualizes to the stated $L^2\to L^{2m+2}$ decay.

What would settle it

Take $(a,b,m,n)=(1,2,2,1)$ with the cutoff $\psi$ from (2.4) and $f=\chi_{[0,1]}$. For large $\lambda$, numerically evaluate $\|T_\lambda f\|_{L^6(\mathbb{R}^2)}$; the theorem predicts an upper bound of order $\lambda^{-1/4}$, so observing a decay strictly slower than $\lambda^{-1/4}$ would refute it, while matching $\lambda^{-1/4}$ confirms the sharp rate.

Watch

Extended reading notes

Core claim

Theorem 1.1 asserts that for every smooth compactly supported cutoff $\psi$ and every $f\in L^2(\mathbb{R})$, the operator $T_\lambda f(x,y)=\int e^{i\lambda(x^a t^m+y^b t^n)}\psi(x,y,t)f(t)\,dt$ satisfies $\|T_\lambda f\|_{L^{2m+2}(\mathbb{R}^2)}\le C\lambda^{-(1/a+1/\max\{b,n\})/(2(m+1))}\|f\|_{L^2(\mathbb{R})}$, with a constant independent of $f$ and $\lambda$. The proof runs through the $TT^*$ method: it shows the kernel operator $T_\lambda T_\lambda^*$ maps $L^{(2m+2)/(2m+1)}$ into $L^{2m+2}$ with the same decay. The sharpness half constructs an explicit $f$ and cutoff for which the $L^{2m+2}$ norm is at least a constant multiple of $\lambda^{-(1/a+1/b)/(2m+2)}$ when $n\le b$, matching the upper bound because $\max\{b,n\}=b$ there.

Load-bearing premise

The whole argument leans on the unproved assertion that a one-dimensional decay estimate can be applied twice, once to each factor of the product operator, and that the two decay rates add up to the exact exponent in (2.2).

Editorial extensions

If this is right

For parameters satisfying $n\le b$, the decay rate in Theorem 1.1 is optimal; no $L^2\to L^{2m+2}$ estimate can decay faster.
The special case $(a,b,m,n)=(1,2,2,1)$ recovers the previously known sharp $L^2\to L^6$ decay for the operator with phase $x t^2+y^2 t$.
The dual estimate $\|T_\lambda^* g\|_{L^2(\mathbb{R})}\le C\lambda^{-(1/a+1/\max\{b,n\})/(2(m+1))}\|g\|_{L^{(2m+2)/(2m+1)}(\mathbb{R}^2)}$ holds for the adjoint operator.
When $n>b$, the theorem gives a valid upper bound, but the example in the paper no longer matches it, leaving open whether the rate is sharp there.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

If the iteration step (2.2) is made fully explicit, the same factorization idea may extend to phases with more than two monomials, yielding decay rates governed by a sum of one-dimensional Newton-polyhedron exponents.
The sharpness transition at $n=b$ suggests a change in the geometry of the critical points: when the $t^n$ term is the slower-growing one, the decay saturates the bound; when it dominates, the true rate may be better than the proved bound.
A numerical evaluation of the $L^{2m+2}$ norm for large $\lambda$ with the paper's own test function $f=\chi_{[0,1]}$ and cutoff $\psi$ would provide a direct check of the constant and the rate, though it would not settle sharpness for $n>b$.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Desk Editor's Note

read the letter

New theorem, plausible and likely true, but the proof's key L2→L2 endpoint is asserted via an unspecified 'iteration' of a Phong–Stein result and needs real work before the paper is solid.

What is actually new: Theorem 1.1 covers all positive integer exponents (a,b,m,n) with m>n. The best previous results, [Xu23] and [TX24], only handled a=1, m=2, b=2, n=1. The sharpness claim for n≤b is also new. The proof skeleton is Stein's complex interpolation: write TT* as an analytic family, prove an L2→L2 bound and an L1→L∞ bound, then interpolate. The L∞ endpoint is standard and clearly handled (van der Corput plus the Fourier bound for the analytic weight). The lower-bound example for sharpness is explicit and correct for n≤b.

The soft spot. The real problem is (2.2). The paper says 'iterating the (1+1)-dimensional result of Phong and Stein [PS94]' gives ‖I_1∘I_2‖_{L2→L2} ≤ C λ^{-1/a - 1/max{b,n}}. No statement of the 1D theorem is given and no iteration argument appears. This is not a cosmetic omission: I_1∘I_2 is not a tensor product of two independent 1D operators. After the coordinate change u=s+t^m, the amplitude couples u and t (through f_α(u-t^m) or the cutoff), so the product exponent does not obviously follow from applying a 1D result twice. If the correct exponent from the 1D theorem differs from 1/a + 1/max{b,n}, the interpolated exponent in Theorem 1.1 changes. Since (2.2) is the only step in the interpolation carrying a power of λ, the whole theorem rests on it. The paper needs to state the Phong–Stein theorem, prove the iteration, or cite a theorem that covers sums of independent phases with this amplitude coupling.

Smaller issues: the abstract says 'sharp' without qualification, while the theorem proves sharpness only when n≤b. The van der Corput estimate in the L∞ bound is written as |...| ≤ C(1+λ|x^a-z^a|)^{-1/m}, which is trivially 1 when x^a=z^a; the actual decay in that case comes from the y-term. It's a minor point, but the text should phrase the bound in the worst-case sense.

Bottom line: this is a serious paper with a new result and a mostly clear structure. The gap in (2.2) is real and load-bearing. I would send it to a referee with a strong request to justify (2.2), and I would not cite the theorem until the proof is repaired. It is worth a reading-group session to work through the actual Phong–Stein iteration.

Referee Report

1 major / 4 minor

Summary. The paper considers (2+1)-dimensional oscillatory integral operators with polynomial phase x^a t^m + y^b t^n, m>n, and claims sharp L^2 -> L^{2m+2} decay estimates. The proof follows the Stein-Tomas strategy: write T_λ T_λ^* as an integral operator T_E, insert T_E into an analytic family E_α, prove L^2 -> L^2 and L^1 -> L^∞ bounds on two vertical lines, and apply Stein's complex interpolation. The L^2 endpoint is asserted by 'iterating' a (1+1)-dimensional theorem of Phong and Stein, while the L^1 -> L^∞ bound is derived from van der Corput and a Fourier transform estimate. Sharpness is shown by an explicit example when n ≤ b.

Significance. If the main theorem is correct, it provides sharp Stein-Tomas-type decay for a family of degenerate oscillatory integral operators, extending the cases treated in [Xu23] and [TX24] and going beyond what the broad-narrow method has delivered. The interpolation framework and the lower-bound example are clean and credible, and the claimed exponents match the known special cases. However, the central L^2 bound is currently an assertion rather than a proof: the relevant Phong-Stein theorem is not stated, and the iteration is not demonstrated. The paper is short and does not yet meet the standard of a self-contained proof for the main claim.

major comments (1)

[Section 2, Eq. (2.2)] The assertion that 'iterating the (1+1)-dimensional result of Phong and Stein [PS94]' gives ||T_α_E f||_{L^2(R^2)} ≤ C λ^{-1/a - 1/max{b,n}} ||f||_{L^2(R^2)} for Re α=1 is not justified. The specific Phong-Stein theorem is not stated, and no iteration argument is supplied. The operator I_1 is not a tensor product of two independent one-dimensional operators: its amplitude ψ(x,y,t)χ(τ − t^m) couples the (x,τ) and (y,t) factors, so a naive product of one-dimensional bounds cannot be assumed. To make (2.2) load-bearing, the paper must state the one-dimensional theorem precisely and prove the iteration with explicit handling of the coupling cutoff χ(τ − t^m), including any conditions on the amplitude and the dependence on λ. This is not a cosmetic omission: the interpolation, (2.3), and the passage through (2.1) all depend linearly on (2.2), so the entire theorem rests on this step.

minor comments (4)

[Section 2, sharpness paragraph] The optimality proof is written for the case 'b>n', but Theorem 1.1 claims sharpness for 'n≤b'. The argument works verbatim for the boundary case b=n, taking |y| ≲ λ^{-1/b}, so this is a small but real gap in the proof as written.
[Section 2, analytic family] The paper does not verify the admissibility conditions for Stein's complex interpolation theorem, namely that the operator norms on the two vertical lines grow at most exponentially in |Im α|. The factor e^{iα^2} in the definition of δ_α appears designed for this, but the verification is omitted.
[Section 2, Eq. (2.3)] The bound |\hat{δ_α}(t)| ≤ C(1+|t|)^{-Re α} is quoted for Re α = -1/m, where the right-hand side grows like |t|^{1/m}. The cancellation with the van der Corput factor is the key point, but the product bound is stated without details; it would be helpful to show the multiplication explicitly.
[References] The reference [SM93] is listed as 'E. M. Stein and T. S. Murphy', but the cited book 'Harmonic Analysis: Real-Variable Methods, Orthogonality, and Oscillatory Integrals' is by Elias M. Stein alone. Also, for [PS94], a specific theorem number should be given when it is invoked in the proof of (2.2).

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the main estimate is obtained by interpolation from the external Phong–Stein L^2 result and a directly proven L^∞ bound, and sharpness is demonstrated by an explicit counterexample exhibited in the paper.

full rationale

The derivation is self-contained with respect to external benchmarks. The upper bound in Theorem 1.1 follows from Stein complex interpolation between (2.2), an L^2→L^2 estimate obtained by iterating the independent (1+1)-dimensional Phong–Stein theorem [PS94], and (2.3), an L^1→L^∞ estimate proved in the paper via van der Corput and a Fourier bound for the analytic family. The sharpness half is an explicit computation: the paper constructs the amplitude and test function in (2.4), assumes a hypothetical decay exponent σ, obtains the chain λ^{−1/p(1/a+1/b)} ≲ [∫|T_λ f|^p]^{1/p} ≤ C λ^{−σ}, and concludes σ ≤ (1/p)(1/a+1/b). This is a standard lower-bound argument, not a fitted parameter or a renamed input. The citation to [Xu23] for the example is not load-bearing because the example is reproduced in the paper. The only concern is that the iteration of the Phong–Stein result leading to (2.2) is not shown in detail; that is an omitted proof or correctness risk, not circularity, since [PS94] is an external result and (2.2) is not equivalent by construction to the theorem being proved.

Assumptions & free parameters 0 free parameters · 4 assumptions · 0 invented entities

No free parameters are fitted; a, b, m, n are integer exponents in the statement. The proof depends on standard tools (complex interpolation, van der Corput) and on a cited one-dimensional Phong-Stein theorem whose precise formulation and iteration are not fully displayed.

assumptions (4)

standard math Stein's complex interpolation for analytic families of operators
Used in Section 2 to interpolate between the L^2 to L^2 bound (2.2) and the L^1 to L^infinity bound (2.3) at Re α = 0.
domain assumption Phong-Stein sharp L^2 decay for (1+1)-dimensional degenerate oscillatory integral operators [PS94]
Invoked in Section 2 to obtain (2.2) by 'iterating' after the coordinate change w = s + u^m; the precise theorem is not stated.
standard math van der Corput lemma for oscillatory integrals with polynomial phase
Used in Section 2 to bound the one-dimensional t-integral by (1+λ|x^a-x'^a|)^{-1/m} when m>n.
standard math Analytic continuation of the distribution ψ_α and its Fourier transform bound |ψ̂_α(τ)| ≤ C(1+|τ|)^{-Re α}
Taken from Stein's book [SM93, Chapter IX]; used to bound the kernel E_α at Re α = -1/m.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Some sharp $L^2 \to L^p$ decay estimates for $(2+1)$-dimensional degenerate oscillatory integral operators." pith.science (2026). https://pith.science/paper/QWJHTSU4

@misc{pith2026241115009,
  author       = {Pith},
  title        = {Pith review of: Some sharp $L^2 \to L^p$ decay estimates for $(2+1)$-dimensional degenerate oscillatory integral operators},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/QWJHTSU4}},
  note         = {Machine review of arXiv:2411.15009}
}

abstract

We investigate $(2+1)-$dimensional oscillatory integral operators characterized by polynomial phase functions. By employing Stein's complex interpolation, we derive sharp $L^2\to L^p$ decay estimates for these operators.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

20 extracted references · 19 canonical work pages

[1]

Bourgain

J. Bourgain. L^p estimates for oscillatory integrals in several variables. Geom. Funct. Anal. , 1(4):321--374, 1991

work page 1991
[2]

J. L. Connor, P. R. Curtis and W. A. Young. Uniform asymptotics of oscillating integrals: applications in chemical physics. In Wave Asymptotics, Proc. of the Meeting to Mark the Retirement of Fritz Ursell , pages 283--310. Cambridge Univ. Press Cambridge, 1992

work page
[3]

Fefferman

C. Fefferman. Inequalities for strongly singular convolution operators. Acta Math. , 124(1):9--36, 1970

work page 1970
[4]

Greenblatt

M. Greenblatt. A direct resolution of singularities for functions of two variables with applications to analysis. J. Anal. Math. , 92:233--258, 2004

work page 2004
[5]

Greenblatt

M. Greenblatt. Sharp L^2 estimates for one-dimensional oscillatory integral operators with C^ phase. Amer. J. Math. , 127(3):659--695, 2005

work page 2005
[6]

H \"o rmander

L. H \"o rmander. Oscillatory integrals and multipliers on F L^p . Ark. Mat. , 11(1):1--11, 1973

work page 1973
[7]

D. H. Phong and E. M. Stein. Oscillatory integrals with polynomial phases. Invent. Math. , 110(1):39--62, 1992

work page 1992
[8]

D. H. Phong and E. M. Stein. Models of degenerate fourier integral operators and Radon transforms. Ann. of Math. , 140(3):703--722, 1994

work page 1994

Show all 20 references

[9]

D. H. Phong and E. M. Stein. The Newton polyhedron and oscillatory integral operators. Acta Math. , 179(1):105--152, 1997

1997
[10]

D. H. Phong and E. M. Stein. Damped oscillatory integral operators with analytic phases. Adv. Math. , 134(1):146--177, 1998

1998
[11]

D. H. Phong, E. M. Stein, and J. Sturm. Multilinear level set operators, oscillatory integral operators, and Newton polyhedra. Math. Ann. , 319(3):573--596, 2001

2001
[12]

V. S. Rychkov. Sharp L^2 bounds for oscillatory integral operators with C^ phases. Math. Z. , 236(3):461--489, 2001

2001
[13]

A. Seeger. Degenerate fourier integral operators in the plane. Duke Math. J. , 71:685--745, 1993

1993
[14]

E. M. Stein and T. S. Murphy. Harmonic analysis: real-variable methods, orthogonality, and oscillatory integrals , volume 3. Princeton University Press, 1993

1993
[15]

Z. Shi, S. Xu, and D. Yan. Damping estimates for oscillatory integral operators with real-analytic phases and its applications. Forum Math. , 31(4):843--865, 2019

2019
[16]

Tan and S

Y. Tan and S. Xu. Some new decay estimates for (2+ 1)-dimensional degenerate oscillatory integral operators. Arch. Math. , 122(4):437--447, 2024

2024
[17]

A. N. Varchenko. Newton polyhedra and estimation of oscillating integrals. Funct. Anal. Appl. , 10(3):175--196, 1976

1976
[18]

L. Xiao. Endpoint estimates for one-dimensional oscillatory integral operators. Adv. Math. , 316:255--291, 2017

2017
[19]

S. Xu. A sharp decay estimate for degenerate oscillatory integral operators using broad-narrow method. J. Geom. Anal. , 33(4):115, 2023

2023
[20]

C. W. Yang. Sharp L^p estimates for some oscillatory integral operators in R ^1 . Illinois J. Math. , 48(4):1093--1103, 2004

2004

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.

[1] [1]

Bourgain

J. Bourgain. L^p estimates for oscillatory integrals in several variables. Geom. Funct. Anal. , 1(4):321--374, 1991

work page 1991

[2] [2]

J. L. Connor, P. R. Curtis and W. A. Young. Uniform asymptotics of oscillating integrals: applications in chemical physics. In Wave Asymptotics, Proc. of the Meeting to Mark the Retirement of Fritz Ursell , pages 283--310. Cambridge Univ. Press Cambridge, 1992

work page

[3] [3]

Fefferman

C. Fefferman. Inequalities for strongly singular convolution operators. Acta Math. , 124(1):9--36, 1970

work page 1970

[4] [4]

Greenblatt

M. Greenblatt. A direct resolution of singularities for functions of two variables with applications to analysis. J. Anal. Math. , 92:233--258, 2004

work page 2004

[5] [5]

Greenblatt

M. Greenblatt. Sharp L^2 estimates for one-dimensional oscillatory integral operators with C^ phase. Amer. J. Math. , 127(3):659--695, 2005

work page 2005

[6] [6]

H \"o rmander

L. H \"o rmander. Oscillatory integrals and multipliers on F L^p . Ark. Mat. , 11(1):1--11, 1973

work page 1973

[7] [7]

D. H. Phong and E. M. Stein. Oscillatory integrals with polynomial phases. Invent. Math. , 110(1):39--62, 1992

work page 1992

[8] [8]

D. H. Phong and E. M. Stein. Models of degenerate fourier integral operators and Radon transforms. Ann. of Math. , 140(3):703--722, 1994

work page 1994