Skip to content

Documentary state: EDITORIALLY_ADMITTED. The manuscript has a protected editorial admission, not a theorem certification.

OAI-26 family 077: Fourier restriction for positively curved surfaces

Editorial status: source-pinned candidate. Mathematical status: NO_CERTIFICATION. This article explains claims in two pinned manuscripts; it does not independently validate their proofs.

The question in plain language

Fourier extension transforms information assigned to a curved surface in frequency space into a wave across ordinary three-dimensional space. The question is whether controlling the size of the input on the surface also controls the size of the resulting wave throughout space. The curvature is essential: when wave packets travel in different directions, their overlap can be limited, but controlling complicated concentrations of many packets remains difficult.

A small example

For constant input on the unit sphere, the extension integral can be evaluated explicitly as a radial oscillating function decaying approximately like the reciprocal of distance. In three dimensions, its power is integrable at infinity precisely for exponents greater than three. This explains why the claimed exponent range is strictly above three and why the endpoint cannot simply be appended to the theorem. This example does not prove a bound for arbitrary bounded input.

What is known

The pinned Elliptic capacity propagation and Fourier restriction to the sphere manuscript states that, for every real exponent strictly greater than three, the sphere extension operator maps all bounded measurable complex-valued inputs on the two-sphere into the corresponding power-integrable functions on three-dimensional space, with a constant depending only on the exponent. Its introduction explicitly excludes the endpoint at three. It describes an elliptic-capacity method for joint position-direction concentrations, weighted hairbrush estimates, finite packet arrays, conditional projection arguments, and induction with positive power losses. The manuscript also states a sphere-specific factorization consequence for finite-power input norms and a related Kakeya maximal estimate, which require their own source audit.

The companion Diagonal Fourier extension for positively curved surfaces in three dimensions claims the diagonal bound for every finite exponent greater than three on any compact smooth positively curved surface, possibly with smooth boundary. Unlike the first theorem, this is not restricted to the sphere and it measures the input and output in the same power norm. Its introduction relies on a full array packet theorem and a Kakeya maximal theorem as distinct inputs, then describes elliptic phase continuation, capacity decomposition, sparse-ball epsilon removal, and finite chart transfer. Its additional local smoothing and oscillatory-integral consequences have further specified hypotheses.

The two claims must not be conflated: bounded sphere data and power-integrable input on arbitrary positively curved surfaces are different statements. Both belong to the publisher's September 24, 2026 source package.

GCL verification state

The source lock is openai/math@1581d4cbdd69a0faf34b2089614fe2a814316a6a. Both exact-pinned TeX introductions and primary theorem statements were read. No independent proof replay or theorem certification was performed. The excerpts give a source-grounded description, not a judgment that their analytic estimates are correct.

What remains

Independently inspect the elliptical capacity exponent and weighted hairbrush argument, joint and conditional projection laws, packet-frame normalization, masks and finite-array propagation, and the step from packet concentration to the sphere bound. For the companion verify hypotheses of the packet and Kakeya inputs, the chartwise elliptic continuation and boundary treatment, the uniform sparse-ball estimate and epsilon removal, and the passage from local graph estimates to the claimed diagonal surface estimate. Separately audit the exponent dependence and the asserted local-smoothing corollary. Editorial admission needs the bounded critic, article digest, regression replay, strict MkDocs check, protected merge and current-main readback.

Sources and claim boundaries

NO_CERTIFICATION.


Indexed primary manuscript paths: