Pith. sign in

REVIEW 5 major objections 5 minor 42 references

Simple proof of Bourgain bilinear ergodic theorem and its extension to polynomials and polynomials in primes

T0 review · 5 major / 5 minor · reviewed 2026-08-14 · deepseek-v4-flash

Pith's one-line read This paper proves that bilinear ergodic averages along two arbitrary integer-valued polynomials, and also along prime-indexed polynomial steps, converge almost everywhere.

desk verdict The claimed polynomial bilinear theorem is not established: the rotated oscillation estimates are asserted without proof and the interpolation step is flawed, though the linear Bourgain proof is mostly sound. read the letter →

arxiv 1908.02281 v1 pith:ZWQ4SA4L submitted 2019-08-05 math.DS

classification math.DS MSC 37A3028D0505D1011B3011N3737A45
keywords bilinearergodicaveragespolynomialprimenumbersalmosteverywhereconvergenceoscillationmethodmaximalinequalitiestransferenceprinciplespectralisomorphism
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

This paper gives a proof that bilinear ergodic averages taken along two arbitrary non-constant integer-valued polynomials $P,Q$ converge almost everywhere, both for ordinary time steps and for prime steps. In full, for any invertible measure-preserving transformation $T$ and any $f\in L^r$, $g\in L^{r'}$ with $\frac{1}{r}+\frac{1}{r'}=1$, the averages $\frac{1}{N}\sum_{n=1}^N f(T^{P(n)}x)g(T^{Q(n)}x)$ and their prime-indexed versions converge for almost every $x$. These are nonconventional ergodic averages, so the result is a pointwise, polynomial version of double recurrence. The proof is a modern simple proof based on the simple Lebesgue spectrum of the integer shift, a Fourier identity that rewrites the bilinear averages as integrals of linear averages of rotated functions, and maximal and oscillation estimates for polynomial and prime-polynomial averages; a maximal inequality for the supremum of these averages is isolated as a result of independent interest.

What carries the argument

The load-bearing object is the identity (7.5), combined with the oscillation method. The identity is $$\frac{1}{N}\sum_{n=1}^N f(x+P(n))g(x-P(n)) = \int_{-\pi}^{\pi} \hat f(\$\theta$)\left(\frac{1}{N}\sum_{n=1}^N g_\$\theta$(x-P(n))\right)$e^{{2ix\theta}}$\,d\$\theta$,$$ with $g_\theta(x)=g(x)e^{ix\theta}$. It turns a product of two orbit segments into a continuum of ordinary linear averages of a single rotated function, so estimates from the linear theory can be integrated against $|\hat f|$. The oscillation method is the scheme of Lemma 3 by which a bound on the sum over dyadic blocks of the $\ell^2$ size of the maximal oscillation between block endpoints forces almost-everywhere convergence; the polynomial and prime-polynomial versions of these oscillation estimates are Lemmas 6 and 7, whose proofs rest on the discrete maximal inequality and the circle method for exponential sums. A spectral fact about the integer shift — that it has simple Lebesgue spectrum — is what, in this argument, upgrades these estimates to the rotated, uniformly-in-$\theta$ form (8.3)-(8.4) needed for arbitrary $P,Q$.

What would settle it

Choose $R(n)=n^2$, $Q(n)=n$, and for growing $K$ compute, on finite subsets of $\mathbb Z$, the quantity $\sum_{k=1}^K \|\sup_{N_k\le N\le N_{k+1}, N\in S_\rho} |\frac{1}{N}\sum_{n\le N} e^{in^2\theta}g(x+n)-\frac{1}{N_{k+1}}\sum_{n\le N_{k+1}} e^{in^2\theta}g(x+n)|\|_{\ell^2(\mathbb Z)}$. If there are sequences $g\in\ell^2(\mathbb Z)$ and phases $\theta$ for which this grows like $\sqrt{K}\|g\|_2$ rather than $o(\sqrt{K})\|g\|_2$, the uniform rotated inequality (8.3) fails, and with it the paper's route from the proved case $Q=-P$ to arbitrary polynomial pairs.

Watch

Extended reading notes

Core claim

The central claim is Theorem 2.1: for any non-constant integer-valued polynomials $P(n),Q(n)$ and any invertible measure-preserving transformation $T$ of a probability space, the averages $\frac{1}{N}\sum_{n=1}^N f(T^{P(n)}x)g(T^{Q(n)}x)$ and $\frac{1}{\pi_N}\sum_{p\le N} f(T^{P(p)}x)g(T^{Q(p)}x)$ converge almost everywhere whenever $\frac{1}{r}+\frac{1}{r'}=1$ and $f\in L^r$, $g\in L^{r'}$. The discovery is that this family of results can be reached from a single oscillation scheme: prove the convergence on the integer lattice $\ell^2(\mathbb Z)$ via Fourier analysis and maximal inequalities, then transfer to an arbitrary dynamical system by a bilinear transference principle. The decisive mechanism is the identity (7.5), which rewrites the bilinear average $f(x+P(n))g(x-P(n))$ as an integral of linear averages of rotated functions $g_\theta(x)=g(x)e^{ix\theta}$; this reduces the bilinear problem to a family of linear polynomial averages. To pass from the symmetric case $Q=-P$, where (7.5) applies directly, to arbitrary pairs $P,Q$, the paper states a rotated version of its polynomial and prime-polynomial oscillation estimates, uniformly in the rotation parameter $\theta$ (equations (8.3)-(8.4)).

Load-bearing premise

The proof's bridge to the general case of two distinct polynomials $P$ and $Q$ is an assertion after (8.2): the oscillation inequalities remain true when the summands are multiplied by a phase $e^{iR(n)\theta}$, uniformly over all phases $\theta$. The paper gives no proof or reference for this rotated version, so everything beyond the symmetric case $Q=-P$ stands on that unproved estimate.

Editorial extensions

If this is right

  • For any non-constant integer-valued polynomials $P,Q$, the averages $\frac{1}{N}\sum_{n=1}^N f(T^{P(n)}x)g(T^{Q(n)}x)$ converge for almost every $x$, for all $f\in L^r$ and $g\in L^{r'}$ with $\frac{1}{r}+\frac{1}{r'}=1$.
  • The same almost-everywhere convergence holds when the time parameter runs over primes: $\frac{1}{\pi_N}\sum_{p\le N} f(T^{P(p)}x)g(T^{Q(p)}x)$ converges for almost every $x$.
  • Theorem 2.2 yields a strong maximal inequality for the supremum in $N$ of the polynomial bilinear averages, with $L^1$ norm bounded by $C_r\|f\|_r\|g\|_{r'}$; this is the mechanism that passes from dense classes of functions to all admissible $L^r\times L^{r'}$ pairs.
  • The same oscillation method gives a unified treatment of Birkhoff's theorem, the homogeneous bilinear ergodic theorem, and their polynomial and prime-polynomial extensions, with a bilinear transference principle carrying the integer-lattice estimates to arbitrary measure-preserving systems.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • If the rotated oscillation inequalities (8.3)-(8.4) are supplied with proof, the same Fourier-rotation scheme would plausibly extend to pairs of sequences whose exponential sums obey uniform Weyl-type estimates, including multi-dimensional $\mathbb Z^d$ actions, not just polynomial powers of one transformation.
  • A finite-range numerical check of the uniform-in-$\theta$ oscillation bound for, say, $R(n)=n^2$ and $Q(n)=n$ would provide evidence about whether the asserted rotated estimates are plausible; such a check cannot prove almost-everywhere convergence, but a clear violation would show the gap at (8.3)-(8.4) is genuine.
  • The paper's final conjecture about averages with operators in the weak closure of $T$ would become a natural target if the rotated estimates are established, since the same integral-of-linear-averages mechanism could then handle a much wider class of exponent sequences.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

5 major / 5 minor

Summary. The paper claims to give a simple proof of Birkhoff's ergodic theorem, a proof of Bourgain's homogeneous bilinear ergodic theorem, and an extension of these results to averages along arbitrary non-constant integer-valued polynomials and along primes evaluated at polynomials. The main theorem (Theorem 2.1) asserts almost everywhere convergence of (1/N)Σ f(T^{P(n)}x)g(T^{Q(n)}x) and the prime analogue for f∈L^r, g∈L^{r'} with 1/r+1/r'=1 for any invertible measure-preserving T. The proof strategy follows Bourgain's oscillation method: a spectral identity reduces the bilinear average to an integral of a linear average of a rotated function g_θ; oscillation lemmas (Lemmas 6 and 7, borrowed from Bourgain and Nair) are then invoked; and Theorem 2.2 claims a bilinear maximal inequality that is used to pass from L^∞ to L^r×L^{r'}. Section 7 derives the convergence only in the special case Q = -P (Equation (7.5)), while Section 8 asserts, without proof, a 'rotated version' of Lemmas 6 and 7 (equations (8.3)-(8.4)) that would handle general polynomial pairs, and also claims a bilinear interpolation step that is not carried out.

Significance. If the main theorem were fully proved, it would be a significant contribution: it would generalize Bourgain's bilinear ergodic theorem to polynomial iterates and to primes along polynomials, with a claimed sharp L^r×L^{r'} range. The paper also presents an expository proof of Birkhoff's theorem via Bourgain's oscillation method and cites several machine-checkable or reproducible lemmas from the literature. However, the central new claims are not established: the key estimates for general polynomial pairs are asserted rather than proved, and the interpolation argument does not mathematically close. The paper's strengths are its expository portions and the recognition of the role of the shift's simple Lebesgue spectrum, but the novelty of the extension is unsupported.

major comments (5)
  1. [Section 7, Equation (7.5)] The proof of convergence in Section 7 is carried out only for the special case Q = -P. The identity (7.5) expresses the average along P(n) and -P(n), and the subsequent estimates (7.6)-(7.8) concern g(x-P(n)). Theorem 2.1 is stated for arbitrary non-constant integer-valued polynomials P and Q; the passage from Q = -P to general P,Q is not present in Section 7 and is deferred to the assertions in Section 8. This is a load-bearing gap for the main theorem.
  2. [Section 8, Equations (8.3)-(8.4)] The 'rotated version' of Lemmas 6 and 7 is asserted with no proof and no reference. The statements (8.3)-(8.4) claim a uniform-in-θ oscillation estimate for averages weighted by e^{iR(n)θ} along a polynomial Q, and the same for prime averages. This is precisely the estimate needed to control the integral in (7.6) when the two polynomials P and Q are distinct: the phase e^{iR(n)θ} arises from the second polynomial and changes the Weyl sums fundamentally. Bourgain's lemmas, as cited in [6,7,33], do not contain such a uniform oscillatory estimate. The phrase 'it can be seeing that' is not a proof. Since (8.3)-(8.4) is the only bridge from the proved case Q = -P to arbitrary P,Q, the main theorem for general polynomial pairs is unsupported.
  3. [Section 8, Equation (8.2) and 'bilinear interpolation'] The proof of Theorem 2.2 does not yield the claimed maximal inequality. Equation (8.2) establishes an ℓ1×ℓr → ℓr bound, and the text notes the trivial ℓ∞×ℓ∞ → ℓ∞ bound. The claimed conclusion for f∈L^r, g∈L^{r'} with 1/r+1/r'=1 would require an endpoint such as ℓ1×ℓ∞ → ℓ1 or an interpolation that produces ℓr×ℓ^{r'} → ℓ1. The stated endpoints do not give this: bilinear interpolation between ℓ1×ℓr → ℓr and ℓ∞×ℓ∞ → ℓ∞ yields bounds with norms on the right in a range that does not include the conjugate-exponent pair claimed in Theorem 2.1. The remark at the end of Section 8 concedes that the claimed range is 1 ≤ 1/r+1/r' < 3/2, which is not the same as the conjugate-exponent range in Theorem 2.1. Thus the maximal inequality, and hence the L^r×L^{r'} extension, is not established.
  4. [Section 8, interpolated application to primes] The proof of the prime version of Theorem 2.2 and Theorem 2.1 is not written out. The text says 'Applying the same machinery' and 'The case r = +∞ can be handled in the same manner', but no prime analogue of (8.1)-(8.2) is proved. Since the prime maximal inequality (7.4) is cited from Nair and the rotated prime estimate (8.4) is asserted without proof, the prime part of the main theorem inherits the same gaps and adds no independent derivation.
  5. [Section 6, Equation (6.7)] In the proof of Bourgain's bilinear ergodic theorem, the passage from the ℓ2(Z) estimate (6.5)-(6.6) to the L^2(X) estimate (6.7) is only sketched as 'using carefully similar arguments to that in the proof of Propositions 1 and 2'. Since the transference principle in Propositions 1 and 2 is stated for the homogeneous averages with the form f(T^n x)g(T^{-n}x), and the oscillation estimate requires control along a sequence of intervals, the transference of the oscillation norm is not immediate. This leaves the proof of the classical Bourgain theorem itself incomplete, although this portion is not the main novelty.
minor comments (5)
  1. [Throughout] There are numerous typographical errors and notation inconsistencies (e.g., 'Garcia' for 'Garsia', 'Etamedi' for 'Etemadi', 'curial' for 'crucial', 'Cauchy-Cauchy' repetitions, inconsistent use of [−π,π) and [−1/N,1/N] in Lemma 4 and its proof). These should be corrected.
  2. [Section 2, Theorem 2.2 statement] The displayed norm in Theorem 2.2 contains a misprint: it writes ||f||_r ||f||_{r'} instead of ||f||_r ||g||_{r'}. This should be fixed.
  3. [Section 5, Lemma 4 and Lemma 5] Lemma 4 is stated for g positive integrable on the circle, while the proof of Lemma 5 appears to prove the stronger estimate for arbitrary θ with a uniform constant; the relationship between the two lemmas and the role of the spectral measure σ_f (which is not necessarily absolutely continuous with respect to Lebesgue measure) should be clarified. The estimate is used at a point where absolute continuity of σ_f is asserted without proof.
  4. [Section 5, Theorem 5.1 vs Theorem 5.2] In the proof of Theorem 5.2, the spectral transfer uses σ_f and claims its absolute continuity with respect to Lebesgue measure; this is not generally true for arbitrary f∈ℓ2(Z). The argument appears to rely instead on the fact that the shift has Lebesgue spectrum, but the exposition is not precise about how σ_f is handled.
  5. [References] The reference [2] is cited as the author's own work, but the connection to the present proof is not explained. Several references (e.g., [13], [37]) are cited for standard lemmas; a precise pointer to the claimed rotated estimates (8.3)-(8.4) would be essential, but none is given.

Circularity Check

0 steps flagged · score 2.0 of 10

No circular step exhibited; the lone self-citation is a peripheral pointer, while the load-bearing inputs are external Bourgain and Nair lemmas.

full rationale

The derivation chain does not reduce to its own inputs. Section 7's polynomial case rests on Lemma 6 and Lemma 7, whose proofs are explicitly referred to the independent works of Bourgain and Nair: "The proof of the Lemma 6 and Lemma 7 is based essentially on the Hardy-Littlewood circle method. For their proof, we refer to [6], [7], [33, Lemmas 4 and 5]." The Fourier identity (7.5) is likewise attributed to Bourgain. The only self-citation is the introduction's pointer "For a finitary simple proof of it, we refer to [2]", where [2] is the author's own paper; that citation is not used in the proof of Theorem 2.1 or Theorem 2.2 and hence is not load-bearing. The rotated oscillation estimates (8.3)-(8.4) are asserted without proof, and the bilinear interpolation in Section 8 is only sketched, so the general P,Q case has a serious correctness gap; however, this is an unproved assertion rather than a circular derivation, since the estimates are not derived from the theorem being proved and no fitted parameter is renamed as a prediction.

Assumptions & free parameters 0 free parameters · 4 assumptions · 0 invented entities

The paper's main theorem rests on two unproved assumptions: the rotated oscillatory estimates (8.3)-(8.4) and the bilinear interpolation in Section 8. The other axioms are standard references to known results in ergodic theory and Fourier analysis.

assumptions (4)
  • domain assumption Shift map on Z has simple Lebesgue spectrum
    Used throughout as the spectral foundation for ell^2(Z); cited but not derived.
  • domain assumption Hardy-Littlewood maximal inequality and circle method estimates (Lemmas 6 and 7) cited to [6],[7],[33],[37]
    Bilinear polynomial proof leans on these unproved-in-paper estimates.
  • ad hoc to paper Rotated oscillation inequalities (8.3)-(8.4) hold uniformly in theta
    Asserted without proof in Section 8; this is exactly what converts the Q = -P case to arbitrary P,Q.
  • ad hoc to paper Bilinear interpolation yields maximal inequality for 1 <= 1/r + 1/r' < 3/2
    Claimed in the Remarks of Section 8 without details; needed to pass from L^infinity to L^r times L^{r'}.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Simple proof of Bourgain bilinear ergodic theorem and its extension to polynomials and polynomials in primes." pith.science (2026). https://pith.science/paper/ZWQ4SA4L

@misc{pith2026190802281,
  author       = {Pith},
  title        = {Pith review of: Simple proof of Bourgain bilinear ergodic theorem and its extension to polynomials and polynomials in primes},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/ZWQ4SA4L}},
  note         = {Machine review of arXiv:1908.02281}
}
abstract

We first present a modern simple proof of the classical ergodic Birkhoff's theorem and Bourgain's homogeneous bilinear ergodic theorem. This proof used the simple fact that the shift map on integers has a simple Lebesgue spectrum. As a consequence, we establish that the homogeneous bilinear ergodic averages along polynomials and polynomials in primes converge almost everywhere, that is, for any invertible measure preserving transformation $T$, acting on a probability space $(X, \mathcal{B}, \mu)$, for any $f \in L^r(X,\mu)$ , $g \in L^{r'}(X,\mu)$ such that $\frac{1}{r}+\frac{1}{r'}= 1$, for any non-constant polynomials $P(n),Q(n), n \in \mathbb{Z}$, taking integer values, and for almost all $x \in X$, we have, $$\frac{1}{N}\sum_{n=1}^{N}f(T^{P(n)}x) g(T^{Q(n)}x),$$ and $$\frac{1}{\pi_N}\sum_{\overset{p \leq N}{p\textrm{~~prime}}}f(T^{P(p)}x) g(T^{Q(p)}x),$$ converge. Here $\pi_N$ is the number of prime in $[1,N]$.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

42 extracted references · 42 canonical work pages

  1. [1]

    Aaronson, An introduction to infinite ergodic theory

    J. Aaronson, An introduction to infinite ergodic theory. Mathem atical Sur- veys and Monographs, 50. American Mathematical Society, Provid ence, RI, 1997

  2. [2]

    E. H. el Abdalaoui, On the homogeneous ergodic bilinear averages with M¨ obius and Liouville weights, arXiv:1706.07280v3 [math.CA]

  3. [3]

    Beckner, Inequalities in Fourier analysis, Annals of Mathematic s, Second Series, 102 (1): 159-182

    W. Beckner, Inequalities in Fourier analysis, Annals of Mathematic s, Second Series, 102 (1): 159-182

  4. [4]

    Bellow, Transference principles in ergodic theory, Harmonic An alysis and Partial Differential Equations, edited by Michael Christ, Carlos E

    A. Bellow, Transference principles in ergodic theory, Harmonic An alysis and Partial Differential Equations, edited by Michael Christ, Carlos E. K enig, and Cora Sadosky, Univ. Chicago Press, Chicago, Illinois, (1999), 2 7-39

  5. [5]

    Bourgain, On the maximal ergodic theorem for certain subset s of the integers, Israel Journal of Mathematics, Vol

    J. Bourgain, On the maximal ergodic theorem for certain subset s of the integers, Israel Journal of Mathematics, Vol. 61, (1) , 1988, 39 -72

  6. [6]

    Bourgain, On the pointwise ergodic theorem on Lp for arithmetic sets, Israel Journal of Mathematics, Vol

    J. Bourgain, On the pointwise ergodic theorem on Lp for arithmetic sets, Israel Journal of Mathematics, Vol. 61, (1), 1988, 73-84

  7. [7]

    Bourgain, An approach to pointwise ergodic theorems, Spring er L.N.M., Vol

    J. Bourgain, An approach to pointwise ergodic theorems, Spring er L.N.M., Vol. 1317, 204-223

  8. [8]

    Bourgain, Pointwise ergodic theorems on arithmetic sets, with an appen- dix on return time sequences (jointly with H

    J. Bourgain, Pointwise ergodic theorems on arithmetic sets, with an appen- dix on return time sequences (jointly with H. Fnrstenberg, Y. Katz nelson, D. Ornstein), Publications Math´ ematiques de l’I.H.E.S., (69), 1989, 5 -45

Show all 42 references
  1. [9]

    Bourgain, Double recurrence and almost sure convergence, J

    J. Bourgain, Double recurrence and almost sure convergence, J. Reine Angew. Math., 404 (1990), pp. 140-161

  2. [10]

    A. P. Calderon, Ergodic theory and translation-invariant oper ators, Proc. Nat. Acad. Sci. USA 59 (1968), 349-353

  3. [11]

    Carleson, On convergence and growth of partial sums of Fo urier series, Acta Math

    L. Carleson, On convergence and growth of partial sums of Fo urier series, Acta Math. 116 (1966), 135157

  4. [12]

    Cuny and M

    C. Cuny and M. Weber, Ergodic theorems with arithmetical weights , to ap- pear in Isr. J. Math., available at http://arxiv.org/pdf/1412.7640. WITH OSCILLATION METHOD. 29

  5. [13]

    Demeter, Pointwise convergence of the ergodic bilinear Hilber t transform, Illinois J.Math

    C. Demeter, Pointwise convergence of the ergodic bilinear Hilber t transform, Illinois J.Math. 51.4 (2007), pp. 1123-1158

  6. [14]

    Demeter, T

    C. Demeter, T. Tao, C. Thiele, Maximal multilinear operators. Tr ans. Amer. Math. Soc. 360 (2008), no. 9, 4989-5042

  7. [15]

    Etemadi, An elementary proof of the strong law of large numb ers

    N. Etemadi, An elementary proof of the strong law of large numb ers. Z. Wahrsch. Verw. Gebiete 55 (1981), no. 1, 119-122

  8. [16]

    Furstenberg, Recurrence in ergodic theory and combinato rial number theory, M

    H. Furstenberg, Recurrence in ergodic theory and combinato rial number theory, M. B. Porter Lectures. Princeton University Press, Prin ceton,N.J., 1981

  9. [17]

    V. F. Gaposhkin, A theorem on the convergence almost everyw here of a sequence of measurable functions, and its applications to sequenc es of sto- chastic integrals. Math. USSR Sbornik 33 (1977), -1-17

  10. [18]

    V. F. Gaposhkin, Individual ergodic theorem for normal opera tors on L2 , Funct. Anal. Appl., 15 (1981), 14-18

  11. [19]

    Garcia, Topics in almost everywhere convergence , Markham Publishing Company, Chicago, 1970

    A. Garcia, Topics in almost everywhere convergence , Markham Publishing Company, Chicago, 1970

  12. [20]

    Grafakos, Modern Fourier analysis, Third edition

    L. Grafakos, Modern Fourier analysis, Third edition. Graduate Texts in Mathematics, 250. Springer, New York, 2014

  13. [21]

    Grafakos, T

    L. Grafakos, T. Tao, and E. Terwilleger, Lp bounds for a maximal dyadic sum operator, Math. Z. 246 (2004), no. 1-2, 321-337

  14. [22]

    P. Hall, C. C. Heyde, Martingale limit theory and its application. Pro ba- bility and Mathematical Statistics. Academic Press, Inc. [Harcourt Brace Jovanovich, Publishers], New York-London, 1980

  15. [23]

    G. H. Hardy and J. E. Littlewood, A maximal theorem with functio n- theoretic applications, Acta Mathematica, 54 (1930), 81-116

  16. [24]

    G. H. Hardy, J. E. Littlewood and G. P´ olya, Inequalities, secon d ed., Cam- bridge, at the University Press, 1952

  17. [25]

    R. A. Hunt, On the convergence of Fourier series, Orthogona l Expansions and their Continuous Analogues (Proc. Conf., Edwardsville, Ill., 1967 ), Southern Illinois Univ. Press, Carbondale, Ill., 1968, pp. 235255

  18. [26]

    Jones, Z

    R. Jones, Z. Kaufman, J. Rosenblatt, and M. Wierdl, Oscillation in ergodic theory, Ergodic Theory Dynam. Systems, 18 (1998), 889935

  19. [27]

    Katznelson and B

    Y. Katznelson and B. Weiss, A simple proof of some ergodic theor ems, Israel J. Math., 42 (1982), no. 4, 291296

  20. [28]

    Keane, K

    M. Keane, K. Petersen, Easy and nearly simultaneous proofs o f the er- godic theorem and maximal ergodic theorem. (English summary) Dyn amics & stochastics, 248251, IMS Lecture Notes Monogr. Ser., 48, Ins t. Math. Statist., Beachwood, OH, 2006

  21. [29]

    Lacey and C

    M. Lacey and C. Thiele, A proof of boundedness of the Carleson operator, Math. Res. Lett. 7 (2000), no. 4, 361370

  22. [30]

    Mirek and B

    M. Mirek and B. Trojan, Cotlar’s ergodic theorem along the prime numbers, J. Fourier Ana. and App. 21 (2015), 822848

  23. [31]

    M. G. Nadkarni, Basic ergodic theory, Reprint of the 1997 original. Texts and Readings in Mathematics, 15. Hindustan Book Agency, New Delhi, 201 3

  24. [32]

    M. G. Nadkarni, Spectral theory of dynamical systems, Reprint of the 1998 original. Texts and Readings in Mathematics, 15. Hindustan Book Age ncy, New Delhi, 2011. 30 E. H. EL ABDALAOUI

  25. [33]

    Nair, On polynomials in primes and J

    R. Nair, On polynomials in primes and J. Bourgains circle method app roach to ergodic theorems II, Stud. Math. 105 (1993), no. 3, 207–233

  26. [34]

    Parry, Topics in ergodic theory, Reprint of the 1981 original

    W. Parry, Topics in ergodic theory, Reprint of the 1981 original. Cambridge Tracts in Mathematics, 75. Cambridge University Press, Cambridge , 2004

  27. [35]

    Riesz, Sur un Th´ eor` eme de maximum de Mm

    F. Riesz, Sur un Th´ eor` eme de maximum de Mm. Hardy et Littlewo od, J. London Math. Soc., 7 (1), 1932 10-13

  28. [36]

    Simon, Harmonic analysis

    B. Simon, Harmonic analysis. A Comprehensive Course in Analysis, Part 3. American Mathematical Society, Providence, RI, 2015

  29. [37]

    Thouvenot, La convergence presque sˆ ure des moyennes ergodiques suiv- ant certaines sous-suites d’entiers (d’apr` es Jean Bourgain)

    J-P. Thouvenot, La convergence presque sˆ ure des moyennes ergodiques suiv- ant certaines sous-suites d’entiers (d’apr` es Jean Bourgain). (French) [Almost sure convergence of ergodic means along some subsequences of in tegers (after Jean Bourgain)] S´ eminaire Bourbaki, Vol. ...

  30. [38]

    Weber, Dynamical Systems and Processes, European Mathematical So- ciety Publishing House, IRMA Lectures in Mathematics and Theoretic al Physics 14, 2009

    M. Weber, Dynamical Systems and Processes, European Mathematical So- ciety Publishing House, IRMA Lectures in Mathematics and Theoretic al Physics 14, 2009

  31. [39]

    Wiener, The Fourier integral and certain of its applications, Dover publi- cations, (1958)

    N. Wiener, The Fourier integral and certain of its applications, Dover publi- cations, (1958)

  32. [40]

    Wierdl, Pointwise ergodic theorem along the prime numbers

    M. Wierdl, Pointwise ergodic theorem along the prime numbers. Israel J. Math. 64 (1988), no. 3, 315-336

  33. [41]

    Williams, Probability with martingales, Cambridge Mathematical Text- books

    D. Williams, Probability with martingales, Cambridge Mathematical Text- books. Cambridge University Press, Cambridge, 1991

  34. [42]

    Zygmund, Trigonometric series vol

    A. Zygmund, Trigonometric series vol. I & II, second ed., Cambridge Univ. Press, Cambridge, 1959. University of Rouen Normandy, LMRS UMR 6085 CNRS-UNIV, F768 01 Saint-´Etienne-du-Rouvray, France E-mail address : elhoucein.elabdalaoui@univ-rouen.fr

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.