Bilinear ergodic Hilbert variation for commuting transformations as a consequence of annular triangular variation
Read the paper
Open the original PDF · Download PDF · Zenodo DOI: 10.5281/zenodo.23204591
Original Zenodo files: README.md | verification-supplement.zip | source.zip | paper.pdf
The PDF is an unchanged copy of the published Zenodo file. The printed paper is dated October 6, 2026 and carries no printed version label; 1.0.0 is the Zenodo version.
Abstract
We give an explicit restriction and transference consequence of OpenAI's October 5, 2026 full annular variation theorem for the triangular Hilbert transform. For arbitrary commuting invertible probability-preserving transformations S,T and complex f,g ∈ L3, the symmetric bilinear Hilbert sums have bounded pointwise r-variation in L3/2 for every r > 2, and converge almost everywhere and in L3/2. The restriction uses boxes of fixed positive area and a uniformly summable kernel error. The analytic breakthrough is the cited continuous theorem; the present note supplies the discrete restriction and full measure-theoretic transfer.
What the result says
Two commuting transformations describe two compatible ways to move through a probability space. A bilinear ergodic Hilbert sum combines the values of two functions along those motions, using the signed weight 1/n and truncating symmetrically in positive and negative time. This note shows that the sums settle to a limit and quantitatively controls their fluctuations.
The argument is a consequence of OpenAI's full annular variation theorem for the triangular Hilbert transform. The note adds an explicit passage from that continuous estimate to lattice sums and then to arbitrary commuting probability-preserving transformations.
Pointwise variation and convergence
Let S and T be commuting invertible measurable transformations of a probability space, with measurable inverses, preserving the probability measure. For complex f,g ∈ L3, set H0 = 0 and
HN(f,g)(x) = ∑0<|n|≤N f(Snx)g(Tnx)/n.
For every real r > 2 there is a finite constant Dr, independent of the space and transformations, such that
‖Vr(HN : N ≥ 0)‖3/2 ≤ Dr ‖f‖3 ‖g‖3.
Here Vr is the supremum, over all finite increasing integer lists 0 ≤ N0 < ⋯ < NJ, of (∑j=1J|HNj − HNj−1|r)1/r. The supremum is taken pointwise before the L3/2 norm; each point may choose its own list.
The maximal function supN≥0|HN| has the same bound. The sums converge almost everywhere and in L3/2 to a bounded complex bilinear operator H(f,g). They also satisfy the stronger maximal-tail conclusion
‖supN≥m|HN(f,g) − H(f,g)|‖3/2 → 0 as m → ∞.
No ergodicity or freeness assumption is required.
From continuous annuli to ergodic sums
The continuous analytic breakthrough is OpenAI's full annular variation theorem, Theorem 1.1 in the manuscript dated October 5, 2026, at source commit adc7f1241b42e322a6451854ab7e4b4c146bf78a. It controls the full positive-scale, output-dependent partition supremum for the triangular Hilbert transform at L3 inputs. The present note uses that theorem as its analytic input.
The note's contribution is an explicit restriction and transference proof. Fixed cells of positive area turn lattice values into functions on the plane. Half-integer radial endpoints select precisely the positive and negative lattice terms. A signed kernel comparison has a uniformly summable O(k−2) error, so it controls the entire variation of the error. Integrating over positive-area output cells avoids evaluating an almost-everywhere estimate on a null set.
A finite-menu, finite-window form of Calderón transference then applies the lattice estimate to orbit arrays. Keeping the endpoint horizon fixed while the orbit window expands preserves the pointwise supremum and makes the limiting argument explicit. The maximal estimate and finite variation supply almost-everywhere convergence, norm convergence, and the maximal-tail conclusion.
Scope and limitations
The result concerns complex L3 inputs, symmetric odd Hilbert truncations, commuting transformations, and r > 2. It gives no r = 2 estimate, noncommuting theorem, extension to other input exponents, or one-sided commuting Cesàro convergence statement. The supporting kernel checks address the discrete cell bookkeeping; they do not establish the continuous analytic theorem.
Credited prior work
The continuous theorem is credited to OpenAI. The note makes no independent triangular-Hilbert breakthrough or first-priority claim. Its explicit kernel argument uses established restriction and transference ideas:
- A. P. Calderón, Ergodic theory and translation-invariant operators (1968), for classical transference.
- C. Demeter and C. Thiele, On the two dimensional bilinear Hilbert transform, Section 6 and footnote 36, for the continuous-to-lattice and orbit-array reductions.
- P. Durcik, V. Kovač, K. A. Škreb and C. Thiele, Norm-variation of ergodic averages with respect to two commuting transformations, Section 5, for positive-area phase integration.
- C. Demeter, Pointwise convergence of the ergodic bilinear Hilbert transform, Theorem 1.2, for Hilbert convergence along opposite powers of one transformation.
- L. Becker and P. Durcik, The shifted bilinear Hilbert transform, Corollary 1.4, for r > 2 variation of bilinear averages along opposite powers of one transformation. That setting differs from arbitrary commuting transformations.
Research disclosure
AI tools were used extensively in research, drafting and verification. This preprint has not undergone conventional human peer review or refereeing. Automated audits are not human peer review or formal certification; neither this note nor the full annular variation theorem is claimed formally verified. The cited Lean artifacts concern an older maximal theorem.
Substantive collaboration
I welcome independent verification of the argument, a substantive mathematical extension, or work toward a journal submission. Any coauthorship would reflect actual contributions to the research.
Citation
Alec Kriebel. Bilinear ergodic Hilbert variation for commuting transformations as a consequence of annular triangular variation. Version 1.0.0, Zenodo, October 6, 2026. Preprint. doi:10.5281/zenodo.23204591.
@misc{kriebel2026bilinearergodichilbert,
author = {Kriebel, Alec},
title = {Bilinear ergodic Hilbert variation for commuting transformations as a consequence of annular triangular variation},
year = {2026},
month = oct,
version = {1.0.0},
publisher = {Zenodo},
doi = {10.5281/zenodo.23204591},
url = {https://doi.org/10.5281/zenodo.23204591},
note = {Preprint; published October 6, 2026}
}