Relative Szemerédi Theorem

• Relative Szemerédi theorem is a fundamental result in additive combinatorics ensuring that a positive-density subset of a sparse, pseudorandom ambient set contains arithmetic progressions.
• It employs a transference principle and pseudorandom linear forms condition to extend dense combinatorial phenomena to structured, sparse environments.
• Recent advances generalize the theorem to settings like prime numbers, hypergraphs, and fractals, demonstrating its versatility in arithmetic and ergodic theory.

A relative Szemerédi theorem is a fundamental result in additive combinatorics stating that if a subset of a sparse but sufficiently pseudorandom set (the “ambient” set) has positive density—measured relative to that ambient set—then it must contain nontrivial arithmetic progressions of prescribed length. Introduced in the context of the Green–Tao theorem to detect arbitrarily long arithmetic progressions within the primes, the relative Szemerédi paradigm has become a central organizing principle for sparse and structured environments. The qualitative strength and breadth of these theorems derive from a transference principle: dense combinatorial phenomena in uniform settings transfer to structured yet sparse settings, provided robust forms of pseudorandomness hold. Modern developments have established these results in a range of settings—including finite abelian groups, hypergraphs, locally compact groups, approximate lattices, cut-and-project sets, and probabilistic or random environments—using an overview of hypergraph regularity, ergodic theory, and arithmetic transference arguments.

1. Formulations and Key Theorems

The classical Szemerédi theorem asserts that any subset of the integers (or of a finite abelian group) of positive density contains arbitrarily long arithmetic progressions. The relative Szemerédi theorem replaces the full ambient group by a sparse subset $S$, under a pseudorandomness assumption. The classical instantiation is as follows (Conlon et al., 2013):

• Let $S \subseteq \mathbb{Z}_N$ be a sparse, pseudorandom set, and $A \subseteq S$ of positive relative density: $\frac{|A|}{|S|} \geq \delta > 0$.
• If $S$ satisfies the $k$-linear forms condition (a pattern-counting pseudorandomness), then $A$ contains a $k$-term arithmetic progression.

A precise weighted form is (see (Conlon et al., 2013, Zhao, 2013)): for a majorizing weight $v : \mathbb{Z}_N \to \mathbb{R}_{\ge 0}$ satisfying linear forms conditions, and $0 \leq f \leq v$ with $S \subseteq \mathbb{Z}_N$0,

$S \subseteq \mathbb{Z}_N$1

In more recent advances, these relative results have been extended to multidimensional patterns in the primes (Tao et al., 2013), to recurrence and patterns in approximate lattices (Björklund et al., 2024), to sparse random environments (Briët et al., 2023, Frantzikinakis et al., 2013), and to group settings far beyond $S \subseteq \mathbb{Z}_N$2.

2. Pseudorandomness: The Linear Forms Condition

Fundamental to the relative Szemerédi theorem is a precise pseudorandomness requirement on the ambient set or its majorizing measure. The principal condition is the $S \subseteq \mathbb{Z}_N$3-linear forms condition:

• For $S \subseteq \mathbb{Z}_N$4, the $S \subseteq \mathbb{Z}_N$5-linear forms condition requires that, for any appropriate system of affine-linear forms $S \subseteq \mathbb{Z}_N$6 associated to arithmetic progression patterns,

$S \subseteq \mathbb{Z}_N$7

where the expectation ranges over all relevant parameters. For $S \subseteq \mathbb{Z}_N$8, this means including all products of the form $S \subseteq \mathbb{Z}_N$9, $A \subseteq S$0, etc., including all subproducts when factors are erased (Conlon et al., 2013, Zhao, 2013).

Importantly, the modern relative Szemerédi theorem requires only the linear forms condition, dispensing with the earlier Green–Tao correlation condition (Conlon et al., 2013). This allows the relative theorem to apply to ambient sets of density as low as $A \subseteq S$1 for some constant $A \subseteq S$2, a substantial strengthening over earlier work.

3. Structural Tools: Transference, Hypergraph Regularity, and Counting Lemmas

Relative Szemerédi theorems combine transference techniques (reducing sparse to dense settings via pseudorandom majorants or weights) with hypergraph regularity and counting lemmas adapted to sparse environments (Conlon et al., 2013, Zhao, 2013). Key elements include:

• Dense Model Theorem: Given $A \subseteq S$3 and $A \subseteq S$4 pseudorandom, there exists a bounded “dense model” $A \subseteq S$5 indistinguishable from $A \subseteq S$6 in appropriate cut/discrepancy norms (Zhao, 2013).
• Counting Lemma: If $A \subseteq S$7 and $A \subseteq S$8 are close in cut norm, progression counts are asymptotically preserved; i.e.,

$A \subseteq S$9

• Sparse Hypergraph Removal Lemma: The extension of the removal lemma to sparse, pseudorandom hypergraphs using regularity (approximate decomposition into bounded-complexity parts), a crucial ingredient for higher-complexity configurations (Conlon et al., 2013).

The combination of these tools yields both qualitative existence results and, through recent advances, effective quantitative lower bounds for progression counts, often matching the best-known dense-case constants up to negligible error (Zhao, 2013, Rimanic et al., 2017).

4. Quantitative and Random Versions

An active direction is to quantify progression counts or to study arithmetic structure in randomly generated ambient sets or along random differences. Highlights include:

• Quantitative Relative Theorems in Primes: For the primes up to $\frac{|A|}{|S|} \geq \delta > 0$0, if a subset has relative density at least $\frac{|A|}{|S|} \geq \delta > 0$1, it must contain long progressions, following from a quantified relative Szemerédi theorem together with optimized sieve majorants (Rimanic et al., 2017).
• Thresholds for Random Differences: For random sparse subsets $\frac{|A|}{|S|} \geq \delta > 0$2, the critical threshold for intersectivity (i.e., forcing progressions in every dense set using only differences from $\frac{|A|}{|S|} \geq \delta > 0$3) is $\frac{|A|}{|S|} \geq \delta > 0$4 for $\frac{|A|}{|S|} \geq \delta > 0$5-term progressions, improving and unifying previous bounds via combinatorial and matrix-analytic techniques (Briët et al., 2023, Frantzikinakis et al., 2013).
• Finitary and Pointwise Variants: Probabilistic statements about random sets guarantee, almost surely and with high probability, progression-rich structure in all positive-density subsets, extending to ergodic averages and pointwise convergence (Frantzikinakis et al., 2013).

5. Extensions to Non-Standard and Geometric Settings

Relative Szemerédi theory has evolved to encompass ergodic and geometric settings, including:

• Approximate Lattices and Cut-and-Project Sets: In a general locally compact second countable abelian group $\frac{|A|}{|S|} \geq \delta > 0$6, for approximate lattices $\frac{|A|}{|S|} \geq \delta > 0$7, subsets of positive upper Banach density exhibit syndetic sets of recurrence parameters for generalized patterns. The key is a transverse Furstenberg correspondence principle and multiple recurrence theorem for cross-sections, strengthening the paradigm for sparse, structured subsets outside pure group settings (Björklund et al., 2024).
• Cantor and Fractal Step Sets: Results along sparse digit-restricted Cantor sets (integer sequences constructed via digit restrictions) extend Szemerédi-type recurrence and progression existence to fractal arithmetic environments, often via extensions of IP-set ergodic theorems (Burgin et al., 17 Feb 2026).
• Fractal-Structural Analogues: Work on Erdős semigroups and subsets of $\frac{|A|}{|S|} \geq \delta > 0$8 with invariant and self-similar structure provides a geometric model for relative Szemerédi phenomena in terms of Hausdorff dimension and sumset structure (Yu, 2018).

6. Applications and Impact in Arithmetic and Ergodic Theory

Relative Szemerédi theorems are instrumental in a diverse set of applications:

• Underpinning the Green–Tao theorem that the primes contain arbitrarily long arithmetic progressions, and its extensions to multidimensional constellations (Tao et al., 2013), by providing the transference principle needed to deduce structure in the arithmetic setting.
• Enabling arithmetic removal lemmas, sparse variant extremal results, and connections to locally decodable codes and complexity theory via random differences threshold results (Briët et al., 2023).
• Generalizing multiple recurrence and combinatorial structure to nonstandard groups, cut-and-project sets, and fractal combinatorics.
• Providing the foundational framework for further generalizations involving polynomial configurations, higher-order recurrence, and advanced analytic number theory.

7. Principal Technical Innovations and Recent Directions

The evolution of relative Szemerédi theory has been driven by several technical and conceptual advances:

• The removal of correlation and strong pseudorandomness conditions in favor of linear forms conditions, expanding the applicability to sparser ambient sets (Conlon et al., 2013).
• The introduction and refinement of the arithmetic transference mechanism, leveraging dense model theorems and discrepancy/cut norms for fine control in the sparse regime (Zhao, 2013, Rimanic et al., 2017).
• Higher-dimensional and non-commutative variants via weighted and hypergraph-based analogues of the correspondence principle (Tao et al., 2013, Björklund et al., 2024).
• Fractal and geometric approaches connecting recurrence and combinatorial largeness to dimension theory and sumset structure, illuminating “relative largeness” in new contexts (Yu, 2018, Burgin et al., 17 Feb 2026).

A plausible implication is that the flexible machinery of the relative Szemerédi paradigm, encompassing hypergraph regularity, ergodic theory, arithmetic transfer, and combinatorial geometry, will continue to facilitate generalizations to new classes of sparse, pseudorandom, or structured environments, including multidimensional and polynomial patterns, and inform quantitative applications in number theory and theoretical computer science.