Kneser-Type Generalization of Pollard's Theorem

• The paper extends Pollard's theorem to arbitrary abelian groups by establishing sharp quadratic lower bounds on popular sumsets.
• It employs additive combinatorial tools such as the Dyson transform and Kneser’s inequalities to derive precise structural characterizations.
• The work refines previous bounds by replacing a -2t² term with a -4t²/3 adjustment, closely approaching the ideal t² threshold in cyclic groups.

The Kneser-type generalization of Pollard’s theorem studies sumset structures in abelian groups and establishes sharp lower bounds on the size of popular sumsets—those elements that can be written multiple times as a sum of elements from two finite sets. Building upon classical results by Kneser for ordinary sumsets and Pollard for popular sumsets in cyclic groups, recent work provides refined quadratic term bounds and structural characterizations for general abelian groups, leveraging techniques from additive combinatorics including the pigeonhole principle, Kneser’s inequalities, and the Dyson transform induction (Grynkiewicz et al., 25 Jan 2026).

1. Fundamental Notions and Key Definitions

Let $(G,+)$ be an abelian group, and $A,B\subseteq G$ be finite subsets. The ordinary sumset is defined as

$A+B := \{ a+b : a\in A,\, b\in B \}.$

For $g\in G$, the number of representations of $g$ as a sum $a+b$ with $a\in A$, $b\in B$ is

$$\mathsf{r}_{A,B}(g) := |\{ (a,b)\in A\times B : a+b = g \}|.$$

For a positive integer $t$, the $t$-popular sumset is

$A+_t B := \{ g\in G : \mathsf{r}_{A,B}(g)\ge t \}.$

In particular, $A+_1B = A+B$. For any subset $X\subseteq G$, its stabilizer (or period) is the subgroup

$\mathsf{H}(X) := \{ h\in G : h + X = X \} \le G.$

If $|\mathsf{H}(X)|>1$, $X$ is periodic; otherwise, aperiodic. For the remainder, let $H := \mathsf{H}(A+_tB)$.

2. Main Theorem and Structural Consequences

The central result, achieved by Grynkiewicz and Wang, extends Pollard’s theorem to arbitrary abelian groups and improves upon previous quadratic lower bounds for the total cardinality of popular sumsets. Given $t \ge 2$ and finite subsets $A,B\subseteq G$ with $|A|, |B| \ge t$, if

$$\sum_{i=1}^t |A+_iB| < t|A| + t|B| - \frac{4}{3} t^2 + \frac{2}{3} t,$$

then there exist subsets $A'\subseteq A$ and $B'\subseteq B$ such that:

• $|A\setminus A'| + |B\setminus B'| \le t-1$,
• $A'+_tB' = A'+B' = A+_tB$,
• $\sum_{i=1}^t |A+_iB| \ge t|A| + t|B| - t|H|$.

Additionally, $|A'|,|B'|\ge t+1$ and $|A'+B'|<|A'| + |B'| - t$. This result refines the best known quadratic term from $-2t^2$ to $-\frac{4}{3}t^2$, approaching the “ideal” $-t^2$ threshold achieved for cyclic groups (Grynkiewicz et al., 25 Jan 2026).

3. Proof Techniques and Underlying Additive-Combinatorial Tools

The proof employs several key ideas from additive combinatorics:

• Pigeonhole Principle: Using the bound $\mathsf{r}_{A,B}(g)\ge |A| + |B| - |G|$.
• Kneser’s Inequality: For sumsets, $|A+B| \ge |A| + |B| - |\mathsf{H}(A+B)|$.
• Structural Proposition (Cleanup Lemma): If $A', B'$ already have $A'+_tB'=A+_tB$ and the difference $|A\setminus A'| + |B\setminus B'| \le t-1$, one can replace them with slightly larger $H$-periodic sets $A'', B''$ with

$|A''+B''| < |A''| + |B''| - (1+\alpha)t$

for $\alpha\ge 0$, and an improved sum of popular sumset sizes.

• Dyson Transform Induction: The central inductive tool iterates on the lexicographically ordered quadruple

$$\left(t,\,\sum_{i=1}^t|A+_iB|,\,-(|A|+|B|),\,\min\{|A|,|B|\}\right).$$

For a well-chosen $z\in A-B$, it replaces

$$A \mapsto A \cup (z+B),\quad B \mapsto A \cap (z+B),$$

reducing the sum $\sum_{i=1}^t|A+_iB|$ while preserving $|A| + |B|$.

Prior results used the quadratic bound $-2t^2$ at three stages; this new argument tightens all three to achieve $-\frac{4}{3}t^2$.

4. Relation to Classical Results and Recent Progress

The progression and interrelations among fundamental results in sumset theory are summarized in the following table:

Theorem	General Group	Popular Sums Bound
Pollard (1974)	Cyclic ($C_p$)	$\geq \min\{tp,\, t|A|+t|B|-t^2\}$
Kneser (1955)	Abelian	$|A+B| \geq |A|+|B|-|H|$
Hamidoune–Serra (2008)	Abelian	$\geq t|A|+t|B|-t^2 - \frac14|H|^2$
Grynkiewicz (2010)	Abelian	$\sum |A+_iB| < t|A|+t|B|-2t^2+3t-2$
Grynkiewicz–Wang (2026)	Abelian	$-\frac{4}{3} t^2 + \frac{2}{3} t$

For $t=2$, the result coincides with the classical $-t^2$ bound, providing a full analog of Pollard’s theorem for any abelian group. In cyclic groups ($G=C_p$), the lower bound reduces exactly to Pollard’s $t|A|+t|B|-t^2$, and is sharp up to structural conditions.

5. Examples and Illustrative Cases

A notable instance is $t=2$. The hypothesis simplifies to

$|A+_1B| + |A+_2B| < 2|A| + 2|B| - 3,$

and the quadratic term matches $-t^2 = -4$ for $t=2$. The conclusion ensures that, after removing at most one element in total from $A$ or $B$, the $2$-popular sumset becomes an ordinary sumset of “almost-full” sets.

For $G=C_p$, the result recovers Pollard’s original bound, since $|H|\in\{1,p\}$, and the structural conclusion becomes vacuous when $H=C_p$ as $A'+B' = G$.

6. Open Directions and Conjectures

The ultimate objective is to establish a unified generalization achieving the “ideal” hypothesis

$\sum_{i=1}^t|A+_iB| < t|A| + t|B| - t^2$

which would enforce strong $H$-periodic structure and yield the lower bound

$\sum_{i=1}^t|A+_iB| \ge t|A| + t|B| - t|H|.$

Grynkiewicz–Wang conjecture a refinement using the Euclidean division $t = s|H| + u$, $0\leq s$, $1\leq u \leq |H|$, to potentially prove

$$\sum_{i=1}^t|A+_iB| \ge t|A| + t|B| - t^2 - u(|H| - u),$$

encompassing both the current bound and the one of Hamidoune–Serra in the regime $|H| < t$. The structural description matching this numerology remains open for $t \geq 3$.

7. References and Foundational Literature

The main developments and proofs are presented in:

• D. J. Grynkiewicz, Structural Additive Theory, Springer (2013), Chapter 12.
• Y.-O. Hamidoune and O. Serra, "A note on Pollard’s theorem," (0804.2593) (2008).
• J. M. Pollard, "A generalisation of the theorem of Cauchy and Davenport", J. London Math. Soc. 8 (1974), 460–462.
• M. Kneser, "Ein Satz über abelsche Gruppen mit Anwendungen auf die Geometrie der Zahlen", Math. Z. 61 (1955), 429–434.
• Recent progress and further elaboration in (Grynkiewicz et al., 25 Jan 2026).