arxiv: 2605.05146 · v1 · submitted 2026-05-06 · 🧮 math.CA

Recognition: unknown

Almost Everywhere Convergence of Arithmetic Means of Walsh--Fourier Partial Sums Along Subsequences

Ushangi Goginava

Authors on Pith no claims yet

Pith reviewed 2026-05-08 15:41 UTC · model grok-4.3

classification 🧮 math.CA

keywords Walsh-Fourier seriesalmost everywhere convergencearithmetic meanssubsequencesgrowth conditionsL1 functionspartial sums

0 comments

The pith

Arithmetic means of Walsh-Fourier partial sums converge almost everywhere to f for every f in L1 when the subsequence satisfies the growth condition for all 0<δ<1.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

Gát showed in 2019 that the arithmetic means of Walsh-Fourier partial sums along a subsequence a(n) converge almost everywhere to f for integrable functions, provided a(n) grows at least as fast as (1 + n^{-δ}) a(n) with δ less than 1/2. This paper removes the upper bound on δ and establishes the same almost-everywhere convergence for the full range 0 < δ < 1. A reader cares because the result now applies to a substantially larger family of subsequences while still recovering the original function pointwise almost everywhere for every integrable f.

Core claim

Gát proved that σ_N f → f almost everywhere for f ∈ L¹ under the growth condition a(n+1) ≥ (1 + 1/n^δ) a(n) for 0<δ<1/2. We show that the same conclusion remains valid throughout the full range 0<δ<1.

What carries the argument

The arithmetic means σ_N f = (1/N) ∑_{n=1}^N S_{a(n)} f of the Walsh-Fourier partial sums S_m f, with the growth rate of the increasing sequence a(n) controlling how the partial sums are spaced.

Load-bearing premise

The standard properties of the Walsh orthonormal system suffice to extend the maximal estimates from the δ < 1/2 case to the full range δ < 1.

What would settle it

An explicit f in L¹ together with a sequence a(n) obeying the growth condition for some δ in (1/2, 1) such that σ_N f fails to converge to f almost everywhere would disprove the claim.

read the original abstract

Let $S_m f$ denote the $m$-th partial sum of the Walsh-Fourier series of $f \in L^1$. For an increasing sequence $a=(a(n))_{n \geq 1}$ of positive integers, consider the arithmetic means $$ \sigma_N f:=\frac{1}{N} \sum_{n=1}^N S_{a(n)} f . $$ G\'at proved in 2019 that $\sigma_N f \rightarrow f$ almost everywhere for every $f \in L^1$ under the growth condition $$ a(n+1) \geq\left(1+\frac{1}{n^\delta}\right) a(n), \quad 0<\delta<\frac{1}{2} . $$ We show that the same conclusion remains valid throughout the full range $0<\delta<1$.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

This extends Gát's 2019 result on a.e. convergence of arithmetic means of Walsh-Fourier partial sums to the full range 0<δ<1 under the same subsequence growth condition.

read the letter

The main point is that this paper pushes the allowed growth rate δ on the subsequence a(n) from below 1/2 up to below 1 while keeping the almost-everywhere convergence conclusion for the arithmetic means of the partial sums. It does so by adapting the martingale maximal inequalities and Borel-Cantelli steps that Gát used, relying on the standard L1-boundedness of the Walsh maximal operator and dyadic filtration properties. The extension is genuine because the 2019 paper stopped at 1/2, and the author shows the estimates remain controlled for any fixed δ less than 1 even though the constants now depend on δ. The paper does well by stating the theorem cleanly and sticking to this single improvement without extra claims or new machinery. The soft spots are minor. The constants grow as δ approaches 1, but they stay finite for each fixed value below 1, so the result is still valid. Without the full estimates visible in the abstract it is hard to confirm every step of the adaptation, yet the stress-test finds no inconsistency or divergence in the maximal inequality or the application of Borel-Cantelli. There is no circularity or post-hoc fitting. This is for specialists already working on subsequence convergence for Walsh-Fourier or orthogonal series. A reader who knows the 2019 paper will see the value right away as a direct enlargement of the range. It deserves a serious referee because the claim is precise, the method is standard and reproducible, and the advance is incremental but concrete within its narrow area. I recommend sending it to peer review.

Referee Report

0 major / 2 minor

Summary. The manuscript extends Gát's 2019 theorem by proving that the arithmetic means σ_N f = (1/N) ∑_{n=1}^N S_{a(n)} f of Walsh-Fourier partial sums converge almost everywhere to f for every f ∈ L¹, whenever the index sequence satisfies the growth condition a(n+1) ≥ (1 + n^{-δ}) a(n) for all 0 < δ < 1 (removing the prior restriction δ < 1/2). The argument relies on the dyadic martingale structure of the Walsh system, L¹-boundedness of the maximal partial-sum operator, and an application of the Borel-Cantelli lemma.

Significance. If the derivation holds, the result is a meaningful incremental advance in the theory of a.e. convergence for orthogonal series: it enlarges the admissible range of subsequence growth rates up to nearly linear growth while preserving the parameter-free character of the maximal inequality. The proof re-uses standard Walsh-system and martingale tools without introducing new ad-hoc constants or entities, thereby completing the picture left open by the 2019 work.

minor comments (2)

[Introduction] The introduction would benefit from a one-sentence reminder of the precise statement of Gát's 2019 theorem (including the δ < 1/2 range) before stating the extension, to make the improvement immediately visible to readers.
[Section 2 (Preliminaries)] Notation for the maximal operator M^* f = sup_N |σ_N f| is introduced without an explicit reference to the underlying dyadic filtration; adding a brief sentence linking it to the standard martingale maximal inequality would improve readability.

Simulated Author's Rebuttal

0 responses · 0 unresolved

We thank the referee for the positive evaluation of our manuscript and for recommending acceptance. The report accurately summarizes the extension of Gát's 2019 result to the full range 0 < δ < 1.

Circularity Check

0 steps flagged

No significant circularity

full rationale

The manuscript extends Gát's 2019 result on a.e. convergence of arithmetic means of Walsh-Fourier partial sums under the growth condition a(n+1) ≥ (1 + 1/n^δ) a(n) from δ < 1/2 to the full interval 0 < δ < 1. The argument adapts the existing maximal inequality and Borel-Cantelli application using only the standard dyadic filtration, L1-boundedness of the maximal operator, and the given growth condition; no quantity is defined in terms of itself, no fitted parameter is relabeled as a prediction, and the central claim does not reduce to a self-citation chain. The derivation remains self-contained against external benchmarks.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The proof rests on standard background facts from real analysis and the theory of orthogonal series; no free parameters, ad-hoc axioms, or new postulated entities are introduced.

axioms (1)

domain assumption Standard properties of the Walsh system on [0,1) and the space L1 allowing almost-everywhere convergence statements under subsequence growth conditions
Invoked to support the extension beyond the 2019 range; these are textbook facts in harmonic analysis.

pith-pipeline@v0.9.0 · 5451 in / 1213 out tokens · 44078 ms · 2026-05-08T15:41:34.353367+00:00 · methodology

discussion (0)

Reference graph

Works this paper leans on

10 extracted references

[1]

Durrett, Probability: Theory and Examples (5th ed.)

R. Durrett, Probability: Theory and Examples (5th ed.). Cambridge Series in Statistical and Probabilistic Mathematics, 49. Cambridge University Press, Cambridge, 2019

2019
[2]

N. J. Fine,Cesàro summability of Walsh–Fourier series, Proc. Nat. Acad. Sci. U.S.A.41(1955), 558–591

1955
[3]

Gát,Almost everywhere convergence of Fejér and logarithmic means of subsequences of partial sums of the Walsh–Fourier series of integrable functions, J

Gy. Gát,Almost everywhere convergence of Fejér and logarithmic means of subsequences of partial sums of the Walsh–Fourier series of integrable functions, J. Approx. Theory162(2010), no. 4, 687–708

2010
[4]

Gát, Cesàro means of subsequences of partial sums of trigonometric Fourier series

Gy. Gát, Cesàro means of subsequences of partial sums of trigonometric Fourier series. Constr. Approx. 49 (2019), no. 1, 59–101

2019
[5]

Gát, Almost everywhere divergence of Cesàro means of subsequences of partial sums of trigonometric Fourier series

Gy. Gát, Almost everywhere divergence of Cesàro means of subsequences of partial sums of trigonometric Fourier series. Math. Ann. 389 (2024), no. 4, 4199–4231

2024
[6]

Gát,On the convergence of Fejér means of some subsequences of partial sums of Walsh–Fourier series, Ann

Gy. Gát,On the convergence of Fejér means of some subsequences of partial sums of Walsh–Fourier series, Ann. Univ. Sci. Budapest. Sect. Comput.49(2019), 187–198

2019
[7]

Goginava and F

U. Goginava and F. Mukhamedov, On problems of the divergence of logarithmic means of Fourier series. Proc. Amer. Math. Soc. 153 (2025), no. 6, 2555–2564

2025
[8]

Goginava and G

U. Goginava and G. Oniani, On the divergence of subsequences of partial Walsh-Fourier sums. J. Math. Anal. Appl. 497 (2021), no. 2, Paper No. 124900, 13 pp

2021
[9]

Schipp, W

F. Schipp, W. R. Wade, P. Simon, and J. Pál,Walsh Series. An Introduction to Dyadic Harmonic Analysis, Adam Hilger, Bristol–New York, 1990

1990
[10]

Zalcwasser,Sur la sommabilité des séries de Fourier, Studia Math.6(1936), 82–88

Z. Zalcwasser,Sur la sommabilité des séries de Fourier, Studia Math.6(1936), 82–88. U. Goginava, Department of Mathematical Sciences, United Arab Emirates University, P.O. Box No. 15551, Al Ain, Abu Dhabi, UAE Email address:zazagoginava@gmail.com; ugoginava@uaeu.ac.ae

1936