Inverse theorems and non-arithmetic critical pairs for restricted sumsets in Z p
Listen
Radio episode about this paper
Transcript
Introduction to the show: ident: Mathematics. The day in mathematics: new theorems, new proofs, and what they open.
Ada: We're looking at the paper: "Inverse theorems and non-arithmetic critical pairs for restricted sumsets in Z p".
Paper summary: Ada: So we're looking at this paper, "Inverse theorems and non-arithmetic critical pairs for restricted sumsets in Z p," by Lia and Lianga. It basically tackles a question that came up from the work of Alon, Nathanson, and Ruzsa back in one thousand nine hundred ninety-six about finding all sets A and B where the sumset size is exactly what you'd expect when they are not arithmetic progressions <ref:2609.10148#pg1>.
Emmy: That's right. The whole idea stems from this earlier work where they proved that if one of the sets, say A, is an arithmetic progression, then the other set B has a very specific structure—it has to be either the first few terms or the last few terms of A. But this paper looks at what happens when you hit a boundary condition.
Kolya: What's that boundary condition specifically? Is it when the sumset size hits p minus two, and we are in that situation where A plus B equals p ? I mean, that feels like a critical point where things might break down or behave differently than usual.
Ada: Exactly. The main purpose of this paper is to show that the conjecture proposed by Liu and Qian fails specifically in this boundary case where the total size of A and B equals the prime number p, and you get exactly p minus two for your restricted sumset size.
Emmy: And they do that by constructing infinite families of sets A and B that are not arithmetic progressions but still hit that exact target of p minus two for every prime number p greater than or equal to eleven <ref:2609.10148#pg1>.
Kolya: Infinite families, huh? That sounds powerful. So what does this mean for the general behavior of these sumsets in Z p? Does it suggest that when you're right on that edge, like A plus B equals p, the sets can be quite chaotic and not have that neat arithmetic progression structure we usually expect?
Ada: It suggests that the boundary case is fundamentally different from what happens in the non-boundary case. The paper constructs these counterexamples using two different infinite families, Theorem two point one and Theorem two point three, which are distinct depending on whether you're looking at the prime number eleven or other primes greater than it <ref:2609.10148#pg3>.
Emmy: It's interesting how they use these constructions to show that when you step outside the condition of A + B being less than p minus one, the structure collapses in a way that allows for these non-arithmetic pairs to exist.
Paper summary: Kolya: So, if I'm just listening, what does this mean for someone who is trying to apply these sumset theorems in a real application? Like, if I'm working with data sets and I see my sumset size hitting this specific value, do I need to be worried that it might not be an arithmetic progression?
Ada: You should be. Because the paper shows that you can construct scenarios where you have A plus B equal to p, yet your restricted sumset size is still p minus two, and those sets aren't structured as simple arithmetic progressions.
Emmy: They then move on to formulate a refined inverse theorem under a slightly stronger condition, specifically when the total size of A and B is less than or equal to p minus one. This theorem, called Theorem A, states that if you have that tighter constraint on the total size of A plus B, and you're looking at equality for the sumset size being p minus two, then both sets must be arithmetic progressions with the same common difference.
Kolya: That sounds like a strong statement. So they show that when we add just a little bit of slack to the total size constraint—making it less than or equal to p minus one—then the structure is forced back into being arithmetic progressions, provided B is at least three and A is at least B plus three.
Ada: That's the core difference between the boundary case and the non-boundary case. In that smaller, constrained area, everything seems to behave as expected, meaning you get those nice arithmetic progression results for equality in Theorem A.
Emmy: And they prove that this necessity—that if you have that sumset size p minus two under those conditions, then A and B must be arithmetic progressions with the same common difference—works because of Lemma three point two <ref:2609.10148#pg2>. That lemma shows that if one set isn't an arithmetic progression, you get a contradiction when looking at the size of the sumset.
Kolya: So for someone who cares about the tail behavior, what does this refinement tell us? It tells us that if we can bound our sets tightly enough, we can guarantee that whatever structure we find is actually an arithmetic progression. But it doesn't guarantee that when we are on the absolute edge of the prime size, things stay simple.
Paper summary: Ada: That's right. The refined inverse theorem under A + B less than or equal to p minus one is what gives us a clean characterization, but Theorem B gives us those infinite families that show why we can't just assume that structure holds when the boundary condition is exactly met.
Emmy: The construction of these families, Theorem two point one and Theorem two point three, which are distinct for the prime eleven case, really drives home how sensitive this problem is to whether you are exactly at p or slightly below it <ref:2609.10148#pg3>. It shows the complexity increases right at that boundary line between being constrained and being on the edge of a prime modulus.
Kolya: So, to summarize what we've heard about "Inverse theorems and non-arithmetic critical pairs for restricted sumsets in Z p," we have an inverse theorem that works nicely when the total size is less than or equal to p minus one, forcing structure, and counterexamples that prove the conjecture fails right at the boundary when the sizes add up exactly to p.
Ada: That's it. The title itself points out two big things: inverse theorems and non-arithmetic critical pairs. And those families of counterexamples are what make this paper so interesting because they show that for every prime number eleven and beyond, there are sets that break the expected pattern on the edge of the constraint.
Emmy: It’s a deep dive into how algebraic geometry—the structure behind these sets—interacts with additive number theory when you're dealing with restricted sums in a cyclic group. It shows that even when you have strong lower bounds like Cauchy-Davenport, there are specific configurations near the boundary where the expected arithmetic progression structure doesn't automatically appear.
Kolya: For me, it means that if we are looking at random sets of this size, we can't just assume they will behave nicely unless we can prove they fall into that less constrained regime where Theorem A applies. The caveat is always there: the boundary case is special.
Ada: Right. So the main implication for the field is this: you need to be very careful about where you set your constraints on the size of your sets when dealing with sumsets in Z p, because that exact boundary condition can admit non-arithmetic pairs, which contradicts what everyone thought before.
Conclusion: Ada: So we're wrapping up this discussion on "Inverse theorems and non-arithmetic critical pairs for restricted sumsets in Z p."
Emmy: It's a paper that digs into those tricky boundary conditions we talked about earlier, specifically when the total size of two sets A and B equals the prime number p.
Kolya: So at the heart of it is showing that even at this exact boundary, you can still find pairs of sets that aren't simple arithmetic progressions.
Ada: Exactly. The authors construct infinite families of these non-arithmetic critical pairs for every prime number p bigger than eleven.
Emmy: They build two distinct families, one for the case where three times some integer is one modulo p, and another where two times some integer is negative one modulo p.
Kolya: So what does this tell us about the general behavior of these sumsets? Does it mean that when you're right on that edge, like when A plus B equals p, the sets can be much more complicated than we usually expect?
Ada: It means the boundary case is fundamentally different from the non-boundary case. The conjecture about how these pairs should look fails precisely at this line.
Emmy: And they also give us a refined inverse theorem for when you keep things slightly away from that edge, specifically when the total size of A plus B is less than or equal to p minus one.
Kolya: That means in that smaller space, if you see an equality in the sumset size, the sets must be arithmetic progressions with the same common difference. That sounds like a strong guarantee for structure.
Ada: It is strong. But that's only under those tighter constraints we just talked about. The paper shows that when you loosen those constraints and hit that exact boundary of p, you lose that guarantee of arithmetic structure entirely.
Emmy: So the main point is this paper proves the conjecture fails at the boundary, while providing a clean characterization for everything just below it.
Kolya: It really highlights how important it is to precisely define your constraints in these kinds of additive problems. If you miss that boundary condition by even one unit, you get a completely different kind of behavior.
Jiantao Lia, Xinyi Liang
School of Mathematics and Statistics, Liaoning University
math.NT, math.RA
Submitted: 2026-09-09
Updated: 2026-11-17
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Importance score: 89/100
The gist: Infinite families of non-arithmetic critical pairs are constructed in the boundary case where A + B = p and A+bB = p − 2 for every prime p ≥ 11…
Key concepts
- Critical Pair
- A pair of subsets (A, B) in Zp that achieves a specific size for their restricted sumset, |A+bB| = |A| + |B| - 2. These pairs are central to understanding when the Cauchy-Davenport theorem reaches its lower bound.
- Arithmetic Progression
- A set of numbers where the difference between consecutive elements is constant. The paper shows that for non-boundary cases (|A|+|B| <= p-1), critical pairs must be arithmetic progressions with the same common difference to satisfy the equality condition.
- Boundary Case
- The situation in Zp where the sum of set sizes equals the prime modulus, |A| + |B| = p. The paper demonstrates that this specific boundary case allows for non-arithmetic critical pairs, contradicting existing conjectures.
Terminology
Summary
Infinite families of non-arithmetic critical pairs are constructed in the boundary case where A + B = p and A+bB = p − 2 for every prime p ≥ 11, showing that this boundary case is fundamentally different from the non-boundary case The main purpose of this paper is to show that this conjecture fails in the boundary case A + B = p
The Failure of Conjecture 1.3
The authors construct infinite families of non-arithmetic critical pairs (A, B) with A + B = p and A+bB = p − 2 for every prime p ≥ 11 These families reveal that the boundary case is fundamentally different from the non-boundary case The construction involves two infinite families, Theorem 2.1 and Theorem 2. p. 0 p. 03, which are distinct for p = 11
The Refined Inverse Theorem
A refined inverse theorem is formulated under the stronger hypothesis A + B ≤ p − 1 Theorem A states that if B ≥ 3, A ≥ B + 3, and A + B ≤ p − 1, then the equality holds if and only if A and B are arithmetic progressions with the same common difference This theorem excludes the boundary case
Proof of Necessity via Lemma 3.2
The only if
direction of Theorem 3.1 relies on Lemma 3. p. 52, which states that if A is not an arithmetic progression and A+bB0 ≤ k + l0 − 2, then a contradiction arises The proof shows that if A is not an arithmetic progression, the set A+C will contain at least one new element for each element in C, leading to A+bB ≥ A + B − 1, which contradicts the assumption
Proof of Sufficiency via Case Analysis
The "if" direction of Theorem 3.1 is proven by direct computation, showing that if A and B are arithmetic progressions with the same common difference and B consists precisely of the first B terms or the last B terms of A, then A+bB = A + B − 2 This is demonstrated by considering two cases for B: Case 1 where B = [0, l − 1] and Case 2 involving a reflection map
The Role of Previous Results
The proof of the only if
direction uses Theorem 1.2 to establish that if exactly one set is an arithmetic progression, it forces the other to also be one This confirms that both A and B must be non-arithmetic in the critical case The proof of the "if" direction utilizes Theorem 1.2 together with a new combinatorial counting argument
The Cauchy–Davenport Context
The study builds upon the classical Cauchy–Davenport theorem, which states that A + B ≥ min(p, A + B − 1) The restricted sumset is defined by A+bB = A+bB where equal summands are forbidden This study also relates to the Erdős–Heilbronn conjecture, which states that for any nonempty A ⊂ Zp, A+bA ≥ min(p, 2A − 3)
Summary of Key Findings
The paper establishes that the boundary case A + B = p admits non-arithmetic critical pairs, contradicting Conjecture 1.3 It then proves a refined inverse theorem for the non-boundary case A + B ≤ p − 1 This refined theorem characterizes critical pairs when both sets are arithmetic progressions The construction of infinite families provides a strong counterexample to the conjecture in the boundary case
References
[1] A. Cauchy, Recherches sur les nombres, J. p. 7 Éc. Polytech. 9 (1813), 99–116
[2] H. Davenport, On the addition of residue classes, J. p. 7 Lond. Math. Soc. 10 (1935), 30–32
[3] A. G. Vosper, The critical pairs of subsets of a group of prime order, J. p. 7 Lond. Math. Soc. 31 (1956), 200–205
[4] A. G. Vosper, Addendum to “The critical pairs of subsets of a group of prime order”, J. p. 7 Lond. Math. Soc. 31 (1956), 280–282
[5] P. Erdős, Some problems in number theory, in: Computers in Number Theory, Academic Press, New York, 1971, pp. p. 7 405–414
[6] P. Erdős and R. p. 0 L. Graham, Old and New Problems and Results in Combinatorial Number Theory, L’Enseignement Mathématique, Geneva, 1980
[7] J. A. Dias da Silva and Y. p. 0 O. Hamidoune, Cyclic spaces for Grassmann derivatives and additive theory, Bull. p. 8 Lond. Math. Soc. 26 (1994), 140–146
[8] N. Alon, M. B. Nathanson, and I. p. 0 p. 0 Z. Ruzsa, Adding distinct congruence classes modulo a prime, Amer. p. 8 Math. Monthly 102 (1995), 250–255
[9] N. Alon, M. B. Nathanson, and I. Z. Ruzsa, The polynomial method and restricted sums of congruence classes, J. p. 8 Number Theory 56 (1996), 404–417
[10] G. Károlyi, An inverse theorem for the restricted set addition in abelian groups, J. p. 8 Algebra 290 (2005), 557–593
[11] P. Liu and G. Qian, On the critical pairs of restricted sumsets in Z/pZ, Finite Fields Appl. p. 8 116 (2026), 102867
--- Page 3 ---
Let p be a prime and let Zp = Z/pZ be the cyclic group of order p For nonempty subsets A, B ⊂ Zp, the sumset is A + B = [a + b: a ∈ A, b ∈ B] The classical Cauchy–Davenport theorem states that A + B ≥ min(p, A + B − 1) The corresponding inverse problem, solved by Vosper [3, 4], asserts that equality holds if and only if A and B are arithmetic progressions with the same difference, provided A, B ≥ 2 and A + B ≤ p − 2 The restricted sumset is defined by A+bB = [a + b: a ∈ A, b ∈ B, a ≠ b] The study of restricted sumsets has a rich history, beginning with the Erdős–Heilbronn conjecture [5, 6], proved by Dias da Silva and Hamidoune [7], which states that for any nonempty A ⊂ Zp, A+bA ≥ min(p, 2A − 3)
--- Page 4 ---
Theorem 1.1 (Alon–Nathanson–Ruzsa). p. 0 Let A, B ⊂ Zp be nonempty with A ≠ B. p. 0 Then A+bB ≥ min(p, A + B − 2) The inverse problem for Theorem 1. p. 01 asks for a description of all critical pairs (A, B) attaining equality In the symmetric case A = B, Károlyi [10] proved that if A ≥ 5 and A+bA < p, then equality A+bA = 2A − 3 holds if and only if A is an arithmetic progression The asymmetric case has remained largely open A significant advance was made by Liu and Qian [11], who solved the inverse problem under the assumption that at least one of the sets is an arithmetic progression Their main result is the following
--- Page 5 ---
Theorem 1.2 (Liu–Qian). p. 0 Let A, B ⊂ Zp with A = k, B = l, where l ≥ 3 and k ≥ l + 1 Suppose that A+bB = k + l − 2 ≤ p − 2 (i) If A is an arithmetic progression and k ≥ l + 3, then B is an arithmetic progression consisting precisely of the first l terms or the last l terms of A (ii) If B is an arithmetic progression, then A is an arithmetic progression for which the terms of B constitute exactly its first l terms or its last l terms In the same paper, Liu and Qian proposed the following conjecture for the general case
--- Page 6 ---
Conjecture 1.3 (Conjecture 1. p. 07 in [11]). Let A, B ⊂ Zp. p. 0 Suppose that B = l ≥ 3, A = k ≥ B + 3, A + B ≤ p Then A+bB = k + l − 2 if and only if A and B are arithmetic progressions with the same common difference, and B consists precisely of the first l terms or the last l terms of A The main purpose of this paper is to show that Conjecture 1. p. 13 is false in the boundary case A + B = p Unlike the non-boundary case, where the conjecture is expected to hold
--- Page 7 ---
Motivated by these counterexamples, we then formulate a refined inverse theorem under the stronger hypothesis A+B ≤ p − 1 Under this condition, we have the following theorem
--- Page 8 ---
Theorem A (Refined inverse theorem). Let A, B ⊂ Zp be nonempty subsets such that B ≥ 3, A ≥ B + 3, and A + B ≤ p − 1 Then A+bB = A + B − 2 if and only if A and B are arithmetic progressions with the same common difference, and B consists precisely of the first B terms or the last B terms of A
--- Page 9 ---
Theorem B (Infinite families of counterexamples). For every prime p ≥ 11, there exist (at least two) infinite families of non-arithmetic subsets A, B ⊂ Zp with A + B = p and A+bB = p − 2
--- Page 10 ---
Theorem 2.1 (First family). p. 2 Let p ≥ 11 be a prime. p. 0 p. 0 Let c be an integer such that 3c = 1 (mod p) (i. p. 2e., c = 1/3 in Zp) Define B = [0, 1, c], A = Zp - [-1, -c, 2c] Then A = p − 3, B = 3, A + B = p and A+bB = p − 2 Moreover, neither A nor B is an arithmetic progression
--- Page 7 ---
Example 2.2. For p = 11, the inverse of 3 is 4, so c = 4 Then B = [0, 1, c] and A = Zp - [-1, -c, -2]
--- Page 9 ---
Theorem 2.3 (Second family). p. 3 Let p ≥ 11 be a prime. Let c be an integer such that 2c = -1 (mod p) (i. p. 2e., c = (p − 1)/2) Define B = [0, 1, c], A = Zp - [-1, -c, -2] Then A = p − 3, B = 3, A + B = p and A+bB = p − 2 Moreover, neither A nor B is an arithmetic progression
--- Page 7 ---
Example 2.4. For p = 11, c = 5 Then B = [0, 1, c] and A = Zp - [-1, -c, -2]
--- Page 7 ---
Remark 2.5. The two families in Theorem 2. p. 21 and Theorem 2.3 are distinct for instance, for p = 11, the first family gives B = [0, 1, c] and A = [0, k − l - 1] while the second family gives B = [0, 1, c] and A = Zp - [-1, -c, -2]
--- Page 7 ---
Let p be a prime and let Zp = Z/pZ be the cyclic group of order p For nonempty subsets A, B ⊂ Zp, the sumset is A + B = [a + b: a ∈ A, b ∈ B] The classical Cauchy–Davenport theorem states that A + B ≥ min(p, A + B − 1)
Green — the paper says it, and the page is printed. Red — no passage carries it, so treat that sentence as unverified. The same check runs on your own text: check a summary against its source, $5.
Improvements for AI systems
-
Improved AI system can solve restricted sumset inverse problems for prime order groups Zp when A + B <= p-1 by directly applying Theorem 3.1 to characterize critical pairs (A, B) attaining equality, ensuring the resulting characterization is valid and excludes the boundary case where counterexamples exist.
-
Improved AI system can identify non-arithmetic critical pairs in the boundary case |A| + |B| = p by systematically searching for structures similar to those described in Theorem 2.1 and Theorem 2.3, specifically those involving sets B of size three whose elements are related to inverses or specific linear congruences modulo p.
-
Improved AI system can rigorously verify the
only if
direction of the refined inverse theorem by employing Lemma 3.2 to prove that if |A+bB| = |A| + |B| - 2, then A and B must be arithmetic progressions with the same common difference, and B must consist of precisely its first or last terms.
Related papers
- K ab has the Bogomolov property for canonical heights
- Quasi-Jacobi forms, Appell-Lerch functions, and false theta functions as q-brackets of functions on partitions
- Wasserstein rates of all orders for the Sato-Tate law of Kloosterman sums
- Completing the classification of torsion over sextic fields for elliptic curves defined over Q
- Riemann functions for almost primes and their counting errors
- Topological flatness of orthogonal spin local models