betaFeedback

WorksMathematics

The Erdős similarity problem for polynomials and analytic maps

Abstract

In this paper, we study the Erdős similarity problem by replacing affine copies with images of polynomial maps or analytic maps. This is inspired by a connection between a partial converse of Szemerédi's theorem and the Erdős similarity problem and its variants, as well as a version of polynomial Szemerédi's theorem by Bergelson and Leibman. In particular, we strengthen several known results in the literature. We show that a large class of sequence are not polynomially universal in this sense. First, for sequences satisfying suitable gap hypotheses, an explicit Cantor-type witness is constructed. Second, for sets satisfying certain growth conditions, a random Cantor-type witness is constructed, extending work of Kolountzakis. This also extends certain results in recent work of Cuéllar-Iosevich-Kulkarni-Aravena-Yavicoli to the polynomial setting. Finally, we show for arbitrary sequences AA with zero as an accumulation point that the (d+3)(d+3)-fold sumset (d+3)A(d+3)A is not polynomially universal when restricted to polynomials of degree at most dd; this complements classical work of Bourgain.

Introduction

In 1920, Steinhaus [24] showed that if AA is a finite subset of R\mathbb{R}, then every measurable set E⊂RE \subset\mathbb{R} of positive Lebesgue measure contains a nontrivial affine copy of AA. In 1974, Erdős [10] conjectured that this finite-pattern phenomenon breaks down completely for infinite sets. This is known as the Erdős similarity conjecture; see also Erdős Problem #120 [2].

Conjecture 1.1 (Erdős similarity conjecture). If AA is an infinite subset of R\mathbb{R}, then there is a set E⊂[0,1]E \subset[0,1] with ∣E∣>0|E| > 0 such that EE does not contain any nontrivial affine copy of AA.

To study Conjecture 1.1, it suffices to consider sequences of positive real numbers decreasing to 0 [25]. Although the conjecture remains widely open, several classes of sequences are known to satisfy the conjecture. Eigen [9] and Falconer [11] independently proved the conjecture for sublacunary sequences, that is, sequences A=(an)n=1∞A = (a_n)_{n=1}^{\infty} satisfying an+1/an→1a_{n+1}/a_n \to1. Humke and Laczkovich [16] showed that the condition lim⁡n→∞an+1/an=1\lim_{n\to\infty} a_{n+1}/a_n = 1 can be replaced with lim sup⁡n→∞an+1/an=1\limsup_{n\to\infty} a_{n+1}/a_n = 1, and am−am+1≤an−an+1a_m-a_{m+1} \le a_n-a_{n+1} for all m>nm>n. Kolountzakis [20] proved a more flexible criterion in terms of large finite subsets with large normalized gaps. Bourgain [3] proved the conjecture for sets with special additive structure, for instance sums of three infinite sets. We refer to [25, 19] for surveys of the Erdős similarity problem. Very recently, OpenAI [23] claimed a proof of the conjecture for geometric progressions. See [17, 18] for expositions and extensions. Nevertheless, The Erdős similarity conjecture remains open in general.

We study a variant of the Erdős similarity problem in which affine copies are replaced by polynomial copies, and more generally by analytic copies. We now explain the motivation for this variant.

The Erdős similarity problem is naturally connected with finite-pattern recurrence phenomena. As Bourgain [3] observed, Erdős’s question was likely motivated by the finite-pattern phenomena underlying Szemerédi’s theorem [26] and its extension by Furstenberg and Katznelson [14]. Szemerédi’s theorem guarantees arbitrarily long arithmetic progressions in every set of integers of positive upper density, while Steinhaus’ theorem implies that every measurable subset of R\mathbb{R} of positive measure contains a nontrivial affine copy of every finite configuration. The Erdős similarity problem asks whether this universality breaks down for infinite configurations. The Erdős similarity problem in the large is particularly reminiscent of Szemerédi’s theorem, as it asks whether a uniform largeness condition forces the presence of infinite patterns [4, 5, 21, 15]. It was recently resolved by Yip [28], who proved that for every unbounded set A⊂RA \subset\mathbb{R} and every 0<ε<10 < \varepsilon< 1, there exists a closed set S⊂RS \subset\mathbb{R} having measure at least 1−ε1-\varepsilon in every unit interval and containing no nontrivial affine copy of AA.

The polynomial version of the problem is motivated by the corresponding polynomial recurrence theory. A real-line polynomial Szemerédi theorem of Bergelson and Leibman [1 Theorem B”] says that if P∈R[x]P \in\mathbb{R}[x] satisfies P(0)=0P(0)=0, if a1,…,ak∈Ra_1,\ldots,a_k \in\mathbb{R}, and if E⊂RE \subset\mathbb{R} has positive upper Banach density, then there are a positive integer nn and a point x∈Rx \in\mathbb{R} such that

x+P(naj)∈E(1≤j≤k).x + P(na_j) \in E \qquad(1 \leq j \leq k).

Equivalently, with

Q(t)=x+P(nt),Q(t) = x + P(nt),

the set EE contains the finite polynomial image

{Q(a1),…,Q(ak)}.\{Q(a_1),\ldots,Q(a_k)\}.

Thus finite polynomial copies of arbitrary finite configurations are unavoidable in positive-density sets. We ask whether infinite polynomial copies can nevertheless be avoided by a set of positive measure.

This perspective is part of a broader circle of continuous analogues of classical results from additive combinatorics. For example, polynomial Roth-type theorems on sets of fractional dimension have been studied in [7, 13], and continuous Furstenberg–Sárközy-type problems have been studied in [8, 22]. A more familiar polynomial Szemerédi theorem of Bergelson and Leibman [1 Theorem B’] concerns configurations u+pi(n)u+p_i(n) generated by several polynomials p1,…,pkp_1,\ldots,p_k. By the Lebesgue density theorem, this also has a continuous Steinhaus-type analogue; see for instance [7 p. 6973].

It is therefore natural to ask for an Erdős similarity analogue in which affine maps are replaced by polynomial maps, and more generally by analytic maps. This direction is also motivated by the bi-Lipschitz variant studied recently by Feng, Lai, and Xiong [12]: since every nontrivial affine map is bi-Lipschitz, their problem already contains the classical affine setting and asks for avoidance under a larger class of transformations.

More precisely, we study the following polynomial and analytic analogue of the Erdős similarity problem.

Question 1.2 (Polynomial and analytic Erdős similarity problem). Let (an)n=1∞(a_n)_{n=1}^{\infty} be a decreasing sequence of positive numbers with an→0a_n \to0. Does there exist a measurable set E⊂[0,1]E \subset[0,1] with ∣E∣>0|E| > 0 such that

{P(an):n∈N}⊄E\{P(a_n): n \in\mathbb{N}\} \not\subset E

for every nonconstant polynomial P∈R[x]P \in\mathbb{R}[x]? Can one choose EE to avoid the images of (an)(a_n) under every nonconstant real-analytic function defined near 00?

We prove positive results for Question 1.2 by three complementary methods. The first is an explicit Cantor-type construction. It is deterministic, works particularly well for sequences satisfying a global gap-comparability condition, and gives a strong tail-avoidance conclusion with a set of measure arbitrarily close to one.

We first state the result obtained by the explicit construction.

Theorem 1.3. Let (an)n=1∞(a_n)_{n=1}^{\infty} be a decreasing sequence of positive numbers such that an→0a_n \to0,

lim sup⁡n→∞an+1an=1,\limsup_{n\to\infty} \frac{a_{n+1}}{a_n}=1,

and there is a constant C>0C > 0 such that

am−am+1≤C(an−an+1)(m>n).a_m-a_{m+1} \le C(a_n-a_{n+1}) \qquad(m>n).

Then for each ϵ>0\epsilon> 0, there is a set E⊂[0,1]E \subset[0,1] with ∣E∣≥1−ϵ|E| \ge1-\epsilon such that EE does not contain

{f(an):n≥N}\{f(a_n):n\ge N\}

for any nonconstant function f:R→Rf:\mathbb{R}\to\mathbb{R} analytic at 00 and any positive integer NN.

The second is a random Cantor-type construction, in the spirit of Kolountzakis [20]. It is less explicit but more flexible: it applies under a local finite-block hypothesis, treats polynomial maps by counting finitely many interval patterns, and treats analytic maps by adding a Taylor-truncation estimate. The following theorem is a strengthening of a result of Kolountzakis [20], Theorem 3.

Theorem 1.4. Let

A=(an)n=1∞,a1>a2>⋯>0,an→0,A=(a_n)_{n=1}^{\infty}, \qquad a_1>a_2>\cdots>0,\qquad a_n\to0,

and identify AA with its underlying set. Suppose that there are finite subsets

Bν={bν,1>⋯>bν,Nν>0}⊂AB_\nu=\{b_{\nu,1}>\cdots>b_{\nu,N_\nu}>0\}\subset A

such that

Nν→∞,bν,1→0,N_\nu\to\infty,\qquad b_{\nu,1}\to0,

and, writing

Δν=min⁡1≤i<Nνbν,i−bν,i+1bν,1,\Delta_\nu=\min_{1\le i<N_\nu}\frac{b_{\nu,i}-b_{\nu,i+1}}{b_{\nu,1}},

one has

−log⁡Δν=o(Nν).-\log\Delta_\nu=o(N_\nu).

Then there exists a measurable set E⊂[0,1]E\subset[0,1] of positive Lebesgue measure such that

P(A)⊄EP(A)\not\subset E

for every nonconstant real polynomial PP.

Moreover, if the sets BνB_\nu may be chosen so that

(1+∣log⁡Δν∣log⁡(1/bν,1))(log⁡Nν+∣log⁡Δν∣)=o(Nν),(1)\left(1+\frac{|\log\Delta_\nu|}{\log(1/b_{\nu,1})}\right)(\log N_\nu+|\log\Delta_\nu|)=o(N_\nu), \tag*{(1)}

then EE may be chosen so that

f(A)⊄Ef(A)\not\subset E

for every nonconstant real-analytic function ff defined in a neighbourhood of 00.

Remark 1.5. This formulation is equivalent, for the polynomial conclusion, to Kolountzakis’s large-gap hypothesis that AA contains arbitrarily large finite subsets BB with −log⁡Δ(B)=o(∣B∣)-\log\Delta(B)=o(|B|). Indeed, since AA is a sequence tending to 00, one may pass to terminal portions of such good blocks to ensure that their largest elements tend to 00, without worsening the normalized gap condition.

The following consequence gives convenient hypotheses which imply the analytic part of the random Cantor theorem. In particular, it contains the sublacunary and gap-comparable classes mentioned above.

Corollary 1.6. Let

A=(an)n=1∞,a1>a2>⋯>0,an→0,A=(a_n)_{n=1}^{\infty}, \qquad a_1>a_2>\cdots>0, \qquad a_n\to0,

and identify the sequence with its underlying set. Assume that at least one of the following conditions holds.

  1. For each nn, let ℓ(n)\ell(n) be the smallest integer such that

aℓ(n)≤an4,a_{\ell(n)}\le\frac{a_n}{4},

and put T(n)={n,n+1,…,ℓ(n)}T(n)=\{n,n+1,\ldots,\ell(n)\}. Then

lim inf⁡n→∞max⁡m∈T(n)(am−am+1)an=0.\liminf_{n\to\infty}\frac{\max_{m\in T(n)}(a_m-a_{m+1})}{a_n}=0.

2.

an+1an→1.\frac{a_{n+1}}{a_n}\to1.

3.

lim sup⁡n→∞an+1an=1,\limsup_{n\to\infty}\frac{a_{n+1}}{a_n}=1,

and there exists C>0C>0 such that

am−am+1≤C(an−an+1)(m>n).a_m-a_{m+1}\le C(a_n-a_{n+1}) \qquad(m>n).

Then there exists a measurable set E⊂[0,1]E\subset[0,1] of positive Lebesgue measure such that

f(A)⊄Ef(A)\not\subset E

for every nonconstant real-analytic function ff defined in a neighbourhood of 00.

We now discuss another corollary to this theorem; this crucially uses an estimate in [6], which shows that certain lacunary sequences cannot have too large additive energy.

Let us begin by setting up the framework. We call a sequence (bi)i≥1⊂(0,∞)(b_i)_{i\ge1}\subset(0,\infty) lacunary with ratio q∈(0,1)q\in(0,1) if bi+1≤qbib_{i+1}\le qb_i for all ii, and its associated counting function is

I(W):=#{i:bi≥exp⁡(−W)},W>0.I(W):=\#\{i:b_i\ge\exp(-W)\}, \qquad W>0.

Corollary 1.7. Suppose S1S_1 and S2S_2 are two lacunary sequences with counting functions I(W)I(W) and J(W)J(W), respectively. Then if

lim sup⁡W→∞I(W)J(W)W=∞\limsup_{W\to\infty}\frac{I(W)J(W)}{W}=\infty

then neither S1+S2S_1+S_2 nor S1−S2S_1-S_2 are polynomial universal.

References

  1. [1]V. Bergelson and A. Leibman. Polynomial extensions of van der Waerden’s and Szemerédi’s theorems. J. Amer. Math. Soc., 9(3):725–753, 1996.DOI
  2. [2]T. F. Bloom. Erdős problem #120. https://www.erdosproblems.com/120.
  3. [3]J. Bourgain. Construction of sets of positive measure not containing an affine image of a given infinite structures. Israel J. Math., 60(3):333–344, 1987.DOI
  4. [4]L. Bradford, H. Kohut, and Y. Mooroogen. Large subsets of Euclidean space avoiding infinite arithmetic progressions. Proc. Amer. Math. Soc., 151(8):3535–3545, 2023.arxiv.org/abs/2205.04786
  5. [5]A. Burgin, S. Goldberg, T. Keleti, C. MacMahon, and X. Wang. Large sets avoiding infinite arithmetic / geometric progressions. Real Anal. Exchange, 48(2):351–364, 2023.DOI
  6. [6]N. M. Cuéllar, A. Iosevich, N. Kulkarni, I. R. Aravena, and A. Yavicoli. Non-universality of sumsets of lacunary sequences and arbitrary sets, 2026.
  7. [7]P. Durcik, S. Guo, and J. Roos. A polynomial Roth theorem on the real line. Trans. Amer. Math. Soc., 371(10):6973–6993, 2019.DOI
  8. [8]P. Durcik, V. Kovač, and M. Stipčić. A strong-type Furstenberg-Sárközy theorem for sets of positive measure. J. Geom. Anal., 33(8):Paper No. 255, 16, 2023.DOI
  9. [9]S. J. Eigen. Putting convergent sequences into measurable sets. Studia Sci. Math. Hungar., 20(1-4):411–412, 1985.
  10. [10]P. Erdős. Remarks on some problems in number theory. Math. Balkanica, 4:197–202, 1974.
  11. [11]K. J. Falconer. On a problem of Erdős on sequences and measurable sets. Proc. Amer. Math. Soc., 90(1):77–78, 1984.
  12. [12]D.-J. Feng, C.-K. Lai, and Y. Xiong. Erdős Similarity Problem via Bi-Lipschitz Embedding. Int. Math. Res. Not. IMRN, (17):12327–12342, 2024.arxiv.org/abs/2312.01319
  13. [13]R. Fraser, S. Guo, and M. Pramanik. Polynomial Roth theorems on sets of fractional dimensions. Int. Math. Res. Not. IMRN, (10):7809–7838, 2022.arxiv.org/abs/1904.11123
  14. [14]H. Furstenberg and Y. Katznelson. An ergodic Szemerédi theorem for commuting transformations. J. Analyse Math., 34:275–291, 1978.DOI
  15. [15]X. Gao, Y. Mooroogen, and C. H. Yip. On an Erdős similarity problem in the large. Bull. Lond. Math. Soc., 57(6):1801–1818, 2025.DOI
  16. [16]P. D. Humke and M. Lazckovich. A visit to the Erdős problem. Proc. Amer. Math. Soc., 126(3):819–822, 1998.
  17. [17]A. Josevich and A. Yavicoli. The geometric case of the Erdős similarity conjecture. a human exposition of the openai proof, 2026. arXiv:2611.02454.
  18. [18]K. Jiang and L. Xi. The Erdős similarity problem for null sequences with adjacent ratios bounded below, 2026. arXiv:2610.11449.arxiv.org/abs/2610.11449
  19. [19]Y. Jung, C.-K. Lai, and Y. Mooroogen. Fifty years of the Erdős similarity conjecture. Res. Math. Sci., 12(1):Paper No. 9, 22, 2025.arxiv.org/abs/2412.11062
  20. [20]M. N. Kolountzakis. Infinite patterns that can be avoided by measure. Bull. London Math. Soc., 29(4):415–424, 1997.DOI
  21. [21]M. N. Kolountzakis and E. Papageorgiou. Large sets containing no copies of a given infinite sequence. Anal. PDE, 18(1):93–108, 2025.DOI
  22. [22]B. Kuca, T. Orponen, and T. Sahlsten. On a continuous Sárközy-type problem. Int. Math. Res. Not. IMRN, (13):11291–11315, 2023.DOI
  23. [23]OpenAI. The geometric case of the Erdős similarity conjecture, 2026. available on GitHub.
  24. [24]H. Steinhaus. Sur les distances des points dans les ensembles de mesure positive. Fundamenta Mathematicae, 1(1):930–104, 1920.DOI
  25. [25]R. E. Svetic. The Erdős similarity problem: a survey. Real Anal. Exchange, 26(2):525–539, 2000/01.
  26. [26]E. Szemerédi. On sets of integers containing no k elements in arithmetic progression. Acta Arith., 27:199–245, 1975.
  27. [27]T. Tao. Exploring the toolkit of Jean Bourgain. Bulletin of the American Mathematical Society, 58(2):155–171, 2021.arxiv.org/abs/2009.06736
  28. [28]C. H. Yip. A resolution of the Erdős similarity problem in the large, 2026. arXiv:2611.14056.arxiv.org/abs/2611.14056

Paper details

vlnc:W_wsfcv8wd
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 2

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Mathematics

Not recorded for this paper

View statement
Looking for related papers…
Contents