← All papers
First page of The Zygmund conjecture and Rey's exponential integrability conjecture

The Zygmund conjecture and Rey's exponential integrability conjecture

Henri Martikainen

math.CA Sep 20, 2026 · v2
The main results, including the continuous Zygmund estimate and sparse overlap bounds, are formalized in Lean 4 with Mathlib-only restatements and kernel plus Nanoda checking.
We prove the Zygmund conjecture in all parameters: the maximal operator associated with products of cubes with side lengths $(s_1,\ldots,s_{m-1},φ(s_1,\ldots,s_{m-1}))$, where $φ$ is any positive coordinatewise nondecreasing function, satisfies the weak $L(\log L)^{m-2}$ estimate for every $m\geq3$. The estimate is new for $m>3$. Our approach is to prove Rey's recent conjecture on the exponential integrability of the overlap function $h_{\mathcal G}=\sum_{I\in\mathcal G}\mathbf{1}_I$, with exponent $1/(m-1)$, for sparse incomparable families $\mathcal G$ of $m$-parameter dyadic rectangles, $m\geq2$. The Zygmund conjecture, even in the continuous form above, follows by strengthening a reduction given by Rey.

The Zygmund conjecture asks whether the maximal operator over products of cubes, with the last side length a nondecreasing function of the others, satisfies a weak L(log L)^{m-2} estimate. Before this work the estimate was known only for m=3. A related conjecture of Rey concerns exponential integrability of the overlap function of sparse incomparable families of dyadic rectangles.

Rey's conjecture is proved with exponent 1/(m-1) via an incomparable maximal estimate with constant (p')^{m-1}. The proof uses a product maximal square-function estimate, a projection representing averages through at most m-1 coordinates, and weighted square-function bounds. A strengthened version of Rey's reduction then yields the continuous Zygmund estimate. The results are formalized in Lean 4; the main codebase was produced with an AI model working through specialized subagents.

The Zygmund conjecture is proved for all m≥3, and the bound is new for m>3. The exponent 1/(m-1) for sparse Φ-Zygmund families is shown to be sharp even under incomparability. Seven main results are formalized against independent Mathlib-only statements, confirmed by Comparator, and checked by Lean's kernel and Nanoda.

SettingSparseSparse and incomparable
Bi-parameter1/21 (Rey's theorem)
m-parameter, m≥31/m1/(m-1) (this paper)
Sharp exponential powers for unweighted overlap