← All problems

Fourier restriction conjecture for the sphere (Stein)

4 years old · traced to

Let n≥2n\ge 2. Let σ\sigma be surface measure on the unit sphere Sn−1⊂RnS^{n-1}\subset\mathbb R^n. Define the extension operator, which is the adjoint of restricting g^\widehat g to the sphere, by

Ef(x)=∫Sn−1f(ω) e2πi x⋅ω dσ(ω),x∈Rn.Ef(x)=\int_{S^{n-1}} f(\omega)\,e^{2\pi i\,x\cdot\omega}\,d\sigma(\omega),\qquad x\in\mathbb R^n .

Conjecture. For 1≤p,q≤∞1\le p,q\le\infty, there is a constant C=C(n,p,q)C=C(n,p,q) such that

∥Ef∥Lq(Rn)≤C ∥f∥Lp(Sn−1,σ)for all f∈Lp(Sn−1)\|Ef\|_{L^q(\mathbb R^n)}\le C\,\|f\|_{L^p(S^{n-1},\sigma)}\qquad\text{for all } f\in L^p(S^{n-1})

if and only if

q>2nn−1andq≥n+1n−1 p′,1p+1p′=1.q>\frac{2n}{n-1}\qquad\text{and}\qquad q\ge\frac{n+1}{n-1}\,p' ,\qquad \tfrac1p+\tfrac1{p'}=1 .

Both conditions are necessary. The first is needed because E1=σ^E1=\widehat\sigma decays only like ∣x∣−(n−1)/2|x|^{-(n-1)/2}. The second is the Knapp example, a characteristic function of a small cap. Necessity is classical, so the open content is sufficiency. By interpolation with the trivial estimates, it suffices to prove the endpoint case p=∞p=\infty: ∥Ef∥Lq≲∥f∥L∞\|Ef\|_{L^q}\lesssim\|f\|_{L^\infty} for all q>2n/(n−1)q>2n/(n-1).

By duality, the conjecture is equivalent to the restriction estimate ∥g^∣Sn−1∥Lp′(σ)≤C∥g∥Lq′(Rn)\|\widehat g|_{S^{n-1}}\|_{L^{p'}(\sigma)}\le C\|g\|_{L^{q'}(\mathbb R^n)}. The same statement is conjectured for every compact hypersurface with nonvanishing Gaussian curvature, and for the paraboloid. The case n=2n=2 is a theorem, so the open problem concerns n≥3n\ge3. [1, 2]

Background

Stein observed around 1967 that the Fourier transform of an LrL^r function with r>1r>1 can be meaningfully restricted to a curved hypersurface, even though such a hypersurface has measure zero. He proposed the sharp range above. Curvature is essential: there is no nontrivial restriction to a hyperplane. The problem is central to modern harmonic analysis. It controls Bochner–Riesz means and is closely related to Strichartz estimates for dispersive PDE and to local smoothing for the wave equation. It also implies the Kakeya conjecture: the restriction conjecture in Rn\mathbb R^n implies that Kakeya sets in Rn\mathbb R^n have Hausdorff dimension nn. Wave packets concentrated on 1×R1/21\times R^{1/2} tubes are the bridge between the two problems.

References
  1. E. M. Stein, Harmonic Analysis: Real-Variable Methods, Orthogonality, and Oscillatory Integrals, Princeton Univ. Press (1993), Ch. VIII–IX.
  2. T. Tao, "Some recent progress on the restriction conjecture", in Fourier Analysis and Convexity, Birkhäuser (2004), 217–243. arXiv:math/0303136
  3. C. Fefferman, "Inequalities for strongly singular convolution operators", Acta Math. 124 (1970), 9–36. doi:10.1007/BF02394567
  4. P. A. Tomas, "A restriction theorem for the Fourier transform", Bull. Amer. Math. Soc. 81 (1975), 477–478. doi:10.1090/S0002-9904-1975-13790-6
  5. T. Tao, "A sharp bilinear restriction estimate for paraboloids", Geom. Funct. Anal. 13 (2003), 1359–1384. arXiv:math/0210084
  6. J. Bennett, A. Carbery and T. Tao, "On the multilinear restriction and Kakeya conjectures", Acta Math. 196 (2006), 261–302. arXiv:math/0509262
  7. L. Guth, "A restriction estimate using polynomial partitioning", J. Amer. Math. Soc. 29 (2016), 371–413. arXiv:1407.1916
  8. L. Guth, "Restriction estimates using polynomial partitioning II", Acta Math. 221 (2018), 81–142. arXiv:1603.04250
  9. J. Hickman and K. M. Rogers, "Improved Fourier restriction estimates in higher dimensions", Camb. J. Math. 7 (2019), 219–282.
  10. H. Wang, "A restriction estimate in R3\mathbb R^3 using brooms", Duke Math. J. 171 (2022), 1749–1822. arXiv:1802.04312
  11. H. Wang and S. Wu, "Restriction estimates using decoupling theorems and two-ends Furstenberg inequalities", preprint (2024). arXiv:2411.08871
  12. Wikipedia, Restriction conjecture.

Source: Oberwolfach Reports: 10.4171/OWR/2022/31, page 43

Progress summary

Refreshed
Open · progress reported

The conjecture remains open, despite an unverified three-dimensional claim and a reader-submitted all-dimensional proof that has not been checked.

Stein proposed the sharp restriction range around 1967. The planar case is known, but sufficiency remains open in dimensions n≥3n\ge 3, especially on the Knapp boundary.

Known results

  • Fefferman, Zygmund, and Carleson–Sjölin established the two-dimensional case.
  • Stein–Tomas proved the sharp L2L^2 restriction range, including q≥2(n+1)/(n−1)q\ge 2(n+1)/(n-1).
  • Guth’s polynomial-partitioning methods, followed by Hickman–Rogers, improved higher-dimensional ranges.
  • Wang and Wu reached q>22/7q>22/7 in dimension three; the conjectured threshold there is q>3q>3.

In September 2026, an OpenAI manuscript claimed the bounded-data estimate in dimension three for every p>3p>3; coverage describes it as unreviewed, with no expert verification.

Community submission (unverified)

A submitted proof argues that Pang-hung Chung’s November 8, 2026 preprint establishes the full strong-type range in every dimension n≥3n\ge 3, including the finite Knapp boundary. No independent assessment confirms the argument.

Current status (as of November 2026): The planar case and substantial partial ranges are settled, but the conjecture for n≥3n\ge 3 remains open; the reported three-dimensional advance and the submitted all-dimensional proof are unverified.

Sources

Solutions 1

Pang-hung Chung's paper (arXiv:2611.11060, 8 Nov 2026) claims the sharp strong-type sphere Fourier extension estimate for all n >= 3, including the finite Knapp boundary. The argument combines uniform spherical-graph estimates, transverse product bounds, Whitney decomposition and dyadic pairing, building on existing Euclidean local-smoothing results. Public Lean sources and a successful GitHub CI record accompany the manuscript.See full solutionHide full solution

A proposed solution to JSP-001040: The Fourier restriction conjecture for the sphere

Author: Pang-hung Chung

Paper: The sharp Fourier extension estimate for the sphere

arXiv: https://arxiv.org/abs/2611.11060

First public submission: 8 November 2026, 01:22:36 UTC

Lean proof repository: https://github.com/penghongzhong/sphere-fourier-extension-lean

1. Main theorem

The manuscript presents a proposed proof of the full strong-type Fourier extension conjecture for the unit sphere in all ambient dimensions n≥3n\geq 3, corresponding to JSP-001040.

For the unit sphere Sn−1⊂RnS^{n-1}\subset\mathbb R^n, equipped with Euclidean surface measure σ\sigma, define

Enf(x)=∫Sn−1e2πix⋅ωf(ω) dσ(ω).E_nf(x)=\int_{S^{n-1}}e^{2\pi i x\cdot\omega}f(\omega)\,d\sigma(\omega).

Theorem 1.1 (Sharp sphere Fourier extension estimate).

Let n≥3n\geq3 and 1≤p,q≤∞1\leq p,q\leq\infty. There exists a finite constant Cn,p,qC_{n,p,q} such that

∥Enf∥Lq(Rn)≤Cn,p,q∥f∥Lp(Sn−1,σ)\|E_nf\|_{L^q(\mathbb R^n)} \leq C_{n,p,q}\|f\|_{L^p(S^{n-1},\sigma)}

for every f∈Lp(Sn−1,σ)f\in L^p(S^{n-1},\sigma) if and only if

q>2nn−1,q≥n+1n−1p′,q>\frac{2n}{n-1}, \qquad q\geq\frac{n+1}{n-1}p',

where p′p' denotes the Holder conjugate exponent.

The statement covers the full exponent range and all dimensions n≥3n\geq3 specified in the Justin Sun Prize entry JSP-001040.

2. Proof strategy

The necessity of the two inequalities follows from the radial asymptotic behavior of the Fourier transform of sphere measure and the classical Knapp small-cap example.

For sufficiency, the manuscript develops the following argument:

  1. Establish uniform critical local estimates and diagonal bounds for rescaled spherical graph patches.
  2. Obtain a transverse product estimate through spherical collision geometry and Gaussian energy bounds.
  3. Apply Whitney decomposition to pairs of separated caps.
  4. Sum the resulting estimates at the exact Knapp scaling relation.
  5. Upgrade restricted estimates to strong type using dyadic dual pairing.
  6. Transfer the finite Knapp boundary estimate to the full sphere and combine it with the strict exponent range.

The central boundary result is Theorem 1.2:

2nn−1<q=n+1n−1p′<∞\frac{2n}{n-1} <q=\frac{n+1}{n-1}p'<\infty

implies the corresponding strong-type extension estimate.

The critical local estimate and the strict mixed range are taken from the existing Euclidean local-smoothing work of OpenAI. The manuscript's additional argument addresses the finite Knapp boundary through uniform spherical-graph estimates, transversality and dyadic summation. The paper explicitly cites these external inputs.

3. Lean formalization

The accompanying public repository contains the mathematical statements and formal proof dependencies.

Repository:

https://github.com/penghongzhong/sphere-fourier-extension-lean

Main theorem declarations:

  • JSP001040Manuscript.fullRange
  • JSP001040Manuscript.finiteKnappBoundary

Principal source files:

  • ManuscriptStatement.lean
  • ManuscriptCorrespondence.lean
  • JSP001040FinalAudit.lean

The project uses Lean 4.34.1 and a pinned Mathlib revision.

Reproduction instructions are provided in the repository README.

A successful public GitHub Actions verification record is available at:

https://github.com/penghongzhong/sphere-fourier-extension-lean/actions/runs/42492304844

Fixed Git commit:

644923943fe3057c196dd5bb238106d582378ef0

The formalization includes imported proof sources from the OpenAI mathematical formalization project and other credited dependencies. The repository documents their provenance, theorem correspondence and axiom audits.

4. Request for independent review

I submit this manuscript and its accompanying Lean materials for independent mathematical and formal-verification review under JSP-001040.

I request evaluation of both the mathematical Solution and the Lean Formalization contribution categories.

The submission is a request for review; no prize acceptance or award decision is implied.

  • lean-repository.json12,366 bytesGitHubOpen
  • 2611.11060v1.pdf231,716 bytesOpen