pith. sign in

arxiv: 2604.09013 · v1 · submitted 2026-04-10 · 🧮 math.AP

Weighted and unweighted regularity of bilinear pseudo-differential operators with symbols in general H\"{o}rmander classes

Guangqing Wang This is my paper

Pith reviewed 2026-05-10 17:27 UTC · model grok-4.3

classification 🧮 math.AP
keywords bilinear pseudo-differential operatorsHörmander classesHardy spacesboundednesssharp maximal functionsweighted inequalitiesmultilinear operators
0
0 comments X

The pith

Bilinear pseudo-differential operators with symbols in BS_{ρ,δ}^m remain bounded from H^p × H^q to L^r when δ exceeds ρ, provided m satisfies a reduced order bound.

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

The paper establishes boundedness for bilinear pseudo-differential operators whose symbols lie in the Hörmander class BS_{ρ,δ}^m in the regime 0 ≤ ρ < δ < 1, which had not been treated before. It proves that the operators map H^p(R^n) × H^q(R^n) into L^r(R^n), or BMO when p = q = r = ∞, whenever the symbol order m obeys m ≤ m_ρ(p,q) minus the correction n max{δ-ρ,0} divided by max{r,2}. The same machinery also yields pointwise control of the sharp maximal function of the output by a multilinear maximal function of the inputs, which then implies weighted norm inequalities under appropriate multilinear weight classes. This extends earlier results that required δ ≤ ρ and handles distinct exponent pairs.

Core claim

Bilinear pseudo-differential operators T_a with symbol a belonging to BS_{ρ,δ}^m are bounded from H^p(R^n) × H^q(R^n) to L^r(R^n) (or BMO at infinity) under the condition m ≤ m_ρ(p,q) - n max{δ-ρ,0}/max{r,2}. For 1 < r1, r2 < ∞ the operators also satisfy the pointwise estimate M^sharp T_a(f1,f2)(x) ≲ M_vec r (f1,f2)(x) whenever m ≤ -n(1-ρ)(1/min{r1,2} + 1/min{r2,2}), which extends the parameter range to 0 ≤ ρ ≤ 1 and 0 ≤ δ < 1 and yields weighted inequalities for multilinear A_{p vec,(r,∞)} weights.

What carries the argument

Bilinear pseudo-differential operator T_a(f1,f2)(x) = ∬ a(x,ξ,η) ˆf1(ξ) ˆf2(η) e^{ix·(ξ+η)} dξ dη with symbol a in BS_{ρ,δ}^m, controlled through sharp maximal-function estimates M^sharp and multilinear maximal functions M_vec r.

If this is right

  • The boundedness holds throughout the full range 0 ≤ ρ ≤ 1, 0 ≤ δ < 1.
  • Weighted norm inequalities follow for multilinear A_{p vec,(r,∞)} weights.
  • The pointwise sharp-maximal-function estimate extends previous work of Park and Tomita to unequal exponent pairs.
  • Applications become available to a wider class of symbols arising in PDEs.

Where Pith is reading between the lines

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

  • The explicit correction term suggests a precise trade-off between symbol regularity loss and operator order that might be tested numerically on model symbols.
  • Similar maximal-function arguments could extend the same correction to trilinear or higher multilinear operators.
  • The weighted inequalities open the door to variable-exponent or Orlicz-space versions of the boundedness result.

Load-bearing premise

The difference δ - ρ in the symbol parameters can be absorbed into a lower permitted order m without creating new obstructions in the kernel or maximal-function estimates.

What would settle it

A concrete symbol a in BS_{ρ,δ}^m with m exactly equal to m_ρ(p,q) - n max{δ-ρ,0}/max{r,2} + ε for small ε > 0 that produces a boundedness failure from some H^p × H^q into L^r.

read the original abstract

This paper investigates the boundedness of bilinear pseudo-differential operators with symbols in the H\"{o}rmander class $BS_{\varrho,\delta}^m(\mathbb{R}^n)$ in the previously unexplored regime $0 \leq \varrho < \delta < 1$. We establish boundedness from $H^p(\mathbb{R}^n) \times H^q(\mathbb{R}^n)$ to $L^r(\mathbb{R}^n)$ (with $L^r$ replaced by $\mathrm{BMO}$ when $p=q=r=\infty$) under the probably optimal condition on the order $$m \leq m_\varrho(p,q) - \frac{n\max\{\delta-\varrho,0\}}{\max\{r,2\}},$$ where $m_\varrho(p,q)$ is the critical order in the case $0\leq\delta\leq\varrho<1.$ Furthermore, we develop refined pointwise estimates via sharp maximal functions, establishing that for $m \leq -n(1-\varrho)(\frac{1}{\min\{r_1,2\}}+ \frac{1}{\min\{r_2,2\}})$ with $1<r_{1},r_{2}<\infty$, the bilinear operators satisfy $$M^\sharp T_a(f_1,f_2)(x) \lesssim \mathcal{M}_{\vec{r}}(f_1,f_2)(x).$$ This extends the parameter range from the restrictive condition $0 \leq \delta \leq \varrho < 1$ to the general setting $0 \leq \varrho \leq 1$, $0 \leq \delta < 1$ with $\delta > \varrho$ permitted, and generalizes previous results of Park and Tomita to distinct exponent pairs. Consequently, we obtain weighted norm inequalities for bilinear pseudo-differential operators under multilinear $A_{\vec{p},(\vec{r},\infty)}$ weights.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 3 minor

Summary. The paper establishes boundedness for bilinear pseudo-differential operators T_a with symbols in the Hörmander class BS_{ρ,δ}^m(R^n) in the regime 0 ≤ ρ < δ < 1. It proves that T_a maps H^p(R^n) × H^q(R^n) to L^r(R^n) (or BMO when p=q=r=∞) provided m ≤ m_ρ(p,q) - n max{δ-ρ,0}/max{r,2}, where m_ρ(p,q) is the critical order for the case δ ≤ ρ. It further derives refined pointwise estimates M^♯ T_a(f1,f2)(x) ≲ M_r(f1,f2)(x) for m ≤ -n(1-ρ)(1/min{r1,2} + 1/min{r2,2}) with 1 < r1,r2 < ∞, extends prior results of Park-Tomita to distinct exponents, and obtains weighted bounds under multilinear A_{p,(r,∞)} weights.

Significance. If the central estimates hold, the work fills an important gap by treating the previously excluded case δ > ρ, where symbols lose regularity in the frequency variable. The explicit order correction, the sharp-maximal-function pointwise bounds, and the resulting weighted inequalities would strengthen the theory of multilinear Calderón-Zygmund operators and their applications to PDEs.

major comments (2)
  1. [Main boundedness theorem and the paragraph following the pointwise estimate] The load-bearing step is the claim that the correction term n max{δ-ρ,0}/max{r,2} fully compensates for the additional loss when δ > ρ. The abstract and the statement of the main boundedness result assert that the same maximal-function machinery used for δ ≤ ρ continues to apply after this adjustment, but the manuscript must supply a detailed comparison of the kernel decay or oscillatory-integral remainder estimates in the two regimes to confirm that no further restrictions on the exponents appear. Without an explicit verification that the refined pointwise bound M^♯ T_a ≲ M_r absorbs the extra loss exactly, the stated range on m remains formally unverified for δ > ρ.
  2. [Section containing the refined pointwise estimates] The pointwise estimate is stated for m ≤ -n(1-ρ)(1/min{r1,2} + 1/min{r2,2}) with 1 < r1,r2 < ∞, but the manuscript does not indicate whether this threshold must be further lowered when δ > ρ or whether the same constant works uniformly. This condition is used to derive the weighted inequalities, so any gap here directly affects the final corollaries.
minor comments (3)
  1. The notation m_ρ(p,q) is introduced without an explicit formula or reference to its precise definition in the δ ≤ ρ case; a short display of the expression would improve readability.
  2. The abstract claims the order condition is 'probably optimal,' but the manuscript should either prove optimality by exhibiting a counter-example when the inequality is violated or cite the corresponding necessity result from the δ ≤ ρ literature.
  3. A few typographical inconsistencies appear in the indexing of the exponents r1,r2 versus r; uniform notation throughout would help.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the thorough reading and valuable comments on our manuscript. The concerns raised about explicit verification of the order correction for δ > ρ are well-taken, and we have revised the paper to include more detailed comparisons of the kernel and oscillatory integral estimates between the regimes δ ≤ ρ and δ > ρ. Below we address each major comment point by point.

read point-by-point responses

  1. Referee: [Main boundedness theorem and the paragraph following the pointwise estimate] The load-bearing step is the claim that the correction term n max{δ-ρ,0}/max{r,2} fully compensates for the additional loss when δ > ρ. The abstract and the statement of the main boundedness result assert that the same maximal-function machinery used for δ ≤ ρ continues to apply after this adjustment, but the manuscript must supply a detailed comparison of the kernel decay or oscillatory-integral remainder estimates in the two regimes to confirm that no further restrictions on the exponents appear. Without an explicit verification that the refined pointwise bound M^♯ T_a ≲ M_r absorbs the extra loss exactly, the stated range on m remains formally unverified for δ > ρ.

    Authors: We agree that an explicit side-by-side comparison strengthens the presentation. In the original proof of Theorem 1.1 (Section 3), the kernel decay is obtained from the symbol estimates in BS_{ρ,δ}^m by repeated integration by parts on the oscillatory integral. When δ > ρ the frequency derivatives of the symbol incur an extra loss of order (δ - ρ), which is precisely offset by lowering the admissible m by n max{δ-ρ,0}/max{r,2}. The resulting remainder is then bounded by the same sharp maximal function M_r that appears in the δ ≤ ρ case; see the estimates leading to (3.12) and (3.15). To address the referee’s request we have inserted a new paragraph (now Subsection 3.2) that directly compares the two regimes, confirming that the same exponent restrictions on p, q, r suffice and that the pointwise bound M^♯ T_a ≲ M_r absorbs the correction term exactly. No additional restrictions on the exponents arise. revision: yes

  2. Referee: [Section containing the refined pointwise estimates] The pointwise estimate is stated for m ≤ -n(1-ρ)(1/min{r1,2} + 1/min{r2,2}) with 1 < r1,r2 < ∞, but the manuscript does not indicate whether this threshold must be further lowered when δ > ρ or whether the same constant works uniformly. This condition is used to derive the weighted inequalities, so any gap here directly affects the final corollaries.

    Authors: The threshold in Theorem 2.3 is already uniform with respect to δ. The proof proceeds by writing the operator as a sum of paraproducts and remainder terms whose symbols satisfy the same BS_{ρ,δ}^m estimates used for the main boundedness result. The extra loss (δ - ρ) is again absorbed by the order condition on m, so the same numerical threshold -n(1-ρ)(1/min{r1,2} + 1/min{r2,2}) continues to guarantee M^♯ T_a(f1,f2) ≲ M_r(f1,f2). We have added a clarifying remark immediately after the statement of Theorem 2.3 (and before the weighted corollaries) stating that the estimate holds for the full range 0 ≤ ρ < δ < 1 without further lowering of m. Consequently the weighted bounds under multilinear A_{p,(r,∞)} weights remain valid as stated. revision: yes

Circularity Check

0 steps flagged

No significant circularity; extension via explicit order correction is independent of the target result

full rationale

The derivation begins from the known critical order m_ρ(p,q) established for the regime 0 ≤ δ ≤ ρ < 1 (cited from prior literature) and inserts an explicit, parameter-dependent correction -n max{δ-ρ,0}/max{r,2} to reach the stated range for δ > ρ. The refined pointwise bound M^♯ T_a(f1,f2) ≲ M_r(f1,f2) is asserted for a separate, concrete threshold m ≤ -n(1-ρ)(1/min{r1,2} + 1/min{r2,2}) and is used to obtain the weighted and unweighted conclusions; neither step is defined in terms of the final boundedness statement nor obtained by fitting to the same data. No self-citation is load-bearing for the central claim, and the correction term is presented as a direct adjustment rather than a renaming or self-referential ansatz. The chain therefore remains self-contained against external benchmarks.

Axiom & Free-Parameter Ledger

0 free parameters · 2 axioms · 0 invented entities

The central claim rests on standard properties of Hörmander symbol classes, Hardy spaces, and sharp maximal functions drawn from prior literature; no new free parameters, ad-hoc axioms, or invented entities are introduced in the abstract.

axioms (2)
  • domain assumption Standard derivative estimates and symbol calculus for the Hörmander class BS_{ρ,δ}^m
    Invoked throughout the abstract as the definition of the operator class under study.
  • domain assumption Boundedness of multilinear maximal functions on weighted spaces
    Used to pass from pointwise estimates to weighted norm inequalities.

pith-pipeline@v0.9.0 · 5658 in / 1417 out tokens · 56953 ms · 2026-05-10T17:27:02.181661+00:00 · methodology

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

23 extracted references · 23 canonical work pages

  1. [1]

    B´ enyi, F

    ´A. B´ enyi, F. Bernicot, D. Maldonado, V. Naibo, and R. H. Torres, On the H¨ ormander classes of bilinear pseudodifferential operators II,Indiana Univ. Math. J.62 (2013), 1733–1764

    work page 2013

  2. [2]

    B´ enyi, D

    ´A. B´ enyi, D. Maldonado, V. Naibo, and R. H. Torres, On the H¨ ormander classes of bilinear pseudodifferential operators,Integral Equ. Oper. Theory67 (2010), 341–364

    work page 2010

  3. [3]

    B´ enyi and R

    ´A. B´ enyi and R. H. Torres, Almost orthogonality and a class of bounded bilinear pseudodifferential operators,Math. Res. Lett.11 (2004), 1–11

    work page 2004

  4. [4]

    A. P. Calder´ on and R. Vaillancourt, A class of bounded pseudo-differential operators, Proc. Nat. Acad. Sci. U.S.A.69 (1972), 1185–1187

    work page 1972

  5. [5]

    Chanillo and A

    S. Chanillo and A. Torchinsky, Sharp function and weightedL p estimates for a class of pseudodifferential operators,Ark. Mat.24 (1985), 1–25

    work page 1985

  6. [6]

    Fefferman,L p bounds for pseudo-differential operators,Israel J

    C. Fefferman,L p bounds for pseudo-differential operators,Israel J. Math.14 (1973), 413–417

    work page 1973

  7. [7]

    H¨ ormander, Pseudo-differential operators and hypoelliptic equations, inProc

    L. H¨ ormander, Pseudo-differential operators and hypoelliptic equations, inProc. Symp. on Singular Integrals, Amer. Math. Soc. 10 (1967), 138–183

    work page 1967

  8. [8]

    Hounie, On theL 2 continuity of pseudo-differential operators,Comm

    J. Hounie, On theL 2 continuity of pseudo-differential operators,Comm. Partial Differential Equations11 (1986), 765–778

    work page 1986

  9. [9]

    K. Li, M. M. Jos´ e and O. Sheldy, Extrapolation for multilinear Muckenhoupt classes and applications.Advances in Mathematics373 (2020), 107286

    work page 2020

  10. [10]

    A. K. Lerner, S. Ombrosi, C. P´ erez, R. H. Torres, and R. Trujillo-Gonz´ alez, New maximal functions and multiple weights for the multilinear Calder´ on-Zygmund the- ory,Adv. Math.220 (2009), 1222–1264. 31

    work page 2009

  11. [11]

    Michalowski, D

    N. Michalowski, D. Rule, and W. Staubach, Multilinear pseudodifferential operators beyond Calder´ on-Zygmund theory,J. Math. Anal. Appl.414 (2014), 149–165

    work page 2014

  12. [12]

    Miyachi and N

    A. Miyachi and N. Tomita, Calder´ on-Vaillancourt type theorem for bilinear opera- tors,Indiana Univ. Math. J.62 (2013), 1165–1201

    work page 2013

  13. [13]

    Miyachi and N

    A. Miyachi and N. Tomita, Bilinear pseudo-differential operators with exotic sym- bols, II,J. Pseudo-Differ. Oper. Appl.10 (2019), 397–413

    work page 2019

  14. [14]

    Miyachi and N

    A. Miyachi and N. Tomita, Bilinear pseudo-differential operators with exotic sym- bols,Ann. Inst. Fourier (Grenoble)70 (2020), 2737–2769

    work page 2020

  15. [15]

    Miyachi and K

    A. Miyachi and K. Yabuta, Sharp function estimates for pseudo-differential operators of classS m ϱ,δ,Bull. Fac. Sci. Ibaraki Univ. Ser. A19 (1987), 15–30

    work page 1987

  16. [16]

    Naibo, On theL ∞ ×L ∞ →BMO mapping property for certain bilinear pseudod- ifferential operators,Proc

    V. Naibo, On theL ∞ ×L ∞ →BMO mapping property for certain bilinear pseudod- ifferential operators,Proc. Amer. Math. Soc.143 (2015), 5323–5336

    work page 2015

  17. [17]

    Nieraeth, Quantitative estimates and extrapolation for multilinear weight classes, Math

    Z. Nieraeth, Quantitative estimates and extrapolation for multilinear weight classes, Math. Ann.375 (2019), 453–507

    work page 2019

  18. [18]

    B. J. Park and N. Tomita, Sharp maximal function estimates for linear and multi- linear pseudo-differential operators,J. Funct. Anal.287 (2024), 110661

    work page 2024

  19. [19]

    B. J. Park and N. Tomita, Sharp maximal function estimates for multilinear pseudo- differential operators of type (0,0),Bull. Lond. Math. Soc.57 (2025), 854–870

    work page 2025

  20. [20]

    Rodr´ ıguez-L´ opez and W

    S. Rodr´ ıguez-L´ opez and W. Staubach, Estimates for rough Fourier integral and pseu- dodifferential operators and applications to the boundedness of multilinear operators, J. Funct. Anal.264 (2013), 2356–2385

    work page 2013

  21. [21]

    E. M. Stein,Harmonic Analysis: Real-Variable Methods, Orthogonality, and Oscil- latory Integrals, Princeton Math. Ser. 43, Princeton University Press, Princeton, NJ, 1993

    work page 1993

  22. [22]

    Sharp function and weighted Lp estimates for pseudo-differential opera- tors with symbols in general H¨ ormander classes

    G. Wang, Sharp function and weightedL p estimates for pseudo-differential operators with symbols in general H¨ ormander classes, arXiv:2206.09825, 2022

    work page arXiv 2022

  23. [23]

    Wang, Sharp maximal function estimates andH p continuities of pseudo- differential operators,TWMS J

    G. Wang, Sharp maximal function estimates andH p continuities of pseudo- differential operators,TWMS J. Pure Appl. Math.16 (2025), 245–270. 32

    work page 2025