Skip to content

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

OAI-26 family 078: the three-dimensional Bochner–Riesz conjecture

Editorial status: SOURCE_CRITIC_COMPLETE; protected acceptance is determined by the admissions ledger. Mathematical status: NO_CERTIFICATION. This article describes a theorem claimed in a pinned paper without independently verifying its proof.

The question in plain language

Fourier analysis decomposes a function into frequencies. The Bochner–Riesz operator reconstructs it using frequencies inside a ball while gradually tapering the contribution of frequencies near the ball's boundary. The fundamental question is when this reconstruction acts as a bounded operator on the space of functions whose absolute values have an integrable power of order \(p\).

Family 078 reports a solution to this boundedness problem in three-dimensional Euclidean space, with a precise threshold for the degree of tapering. The strict inequality at the threshold is essential.

A small example

For a function on three-dimensional space, a Bochner–Riesz multiplier of order \(\delta\) weights frequency \(\xi\) by

\[ (1-|\xi|^2)_+^\delta, \]

where the positive part sets the weight to zero outside the unit frequency ball. Larger \(\delta\) makes the transition at the boundary smoother; the challenge is controlling the resulting operator on \(L^p\).

The publisher's claimed threshold is

\[ \delta>\max\left\{3\left|\frac1p-\frac12\right|-\frac12,\ 0\right\}. \]

For example, at \(p=2\), any positive \(\delta\) satisfies this sufficient condition. At \(p=1\) or \(p=\infty\), the stated sufficient condition requires \(\delta>1\). These examples illustrate the formula but are not independent verification of its sharpness.

What is known

The protected catalogue lists one Family 078 manuscript, Bochner–Riesz multipliers in three dimensions (September 24, 2026).

The primary PDF, Theorem 1.1, claims that the multiplier \(T_\delta\) is bounded on \(L^3(\mathbb{R}^3)\) for every positive fixed order \(\delta\). The paper then states that complex-order interpolation and duality (see Corollary 12.4) imply that, for every \(1\le p\le\infty\), the multiplier extends boundedly on \(L^p(\mathbb{R}^3)\) whenever

\[ \delta>\max\left\{3\left|\frac1p-\frac12\right|-\frac12,\ 0\right\}. \]

The Fourier multiplier is \(\left(1-|\xi|^2\right)_+^\delta\), using the Fourier transform with phase \(e^{-2\pi i x\cdot\xi}\). The order is fixed before the constant is chosen; the operator bound may depend on \(p\) and \(\delta\). Scaling gives uniform bounds for the associated multipliers at radius \(R\) for every \(R>0\). The paper asserts boundedness only in the strict range, not equality at the threshold and not higher-dimensional analogues.

The broader notion of a Bochner–Riesz conjecture includes endpoint and dimension questions; the indexed claim must not be silently enlarged to settle them.

GCL verification state

Source pin: openai/math@1581d4cbdd69a0faf34b2089614fe2a814316a6a. The 114-page primary PDF was retrieved and its bytes reproduced pinned Git blob 2dd11d2a25a1d2fc0dd86bfca549162f970d8d5b (1,044,335 bytes, SHA-256 ea18d159eefcdc7fff031ab14035ecba7205de156a7fcd747a6a9c6bd0d6762a). The abstract, Theorem 1.1, displayed inequality (1.3) and opening proof outline were inspected.

This source-critical pass checks the claimed bounds and their quantifiers, but it does not independently verify the harmonic-analysis proof, packet estimates, interpolation argument, endpoint treatment, or uniformity of the bounds. No Family 078-specific Lean theorem, Comparator acceptance, proof-semantic equivalence audit, or MATHCERT certificate is established.

What remains

The bounded source critic has checked the main theorem, the Fourier multiplier convention, the strict-order criterion, and the claim of radius uniformity by scaling. Full proof criticism must still test the entropy and planar projection arguments, positive transport and packet uncertainty estimates, oscillatory-integral reduction, complex interpolation, and the claimed endpoint bounds.

The updated exact article hash, mathematical Markdown, compendium tests, generated HTML, and protected merge/readback are documentary acceptance controls. These do not establish theorem correctness or source-to-Lean equivalence.

Sources and claim boundaries

Boundary: NO_CERTIFICATION. This is a source-qualified report of a publisher-claimed theorem, not proof certification.


Indexed primary manuscript paths: