Abstract
For positive integers $m$, we consider the theta functions $f_m(z) := \sum_{mk+1 \text{ square}} q^k$. Utilizing classical identities of Jacobi, it is known that $f_4 \equiv f_6 f_{12} \pmod 2$. This paper provides a complete classification of the triples $(a, b, c)$ for which $f_a \equiv f_b f_c \pmod 2$. We demonstrate that such triples must either belong to two infinite families—$(2q, 4q, 4q)$ or $(4q, 8q, 8q)$ for odd $q$—or exist within a finite set of eight sporadic cases. This classification allows for the formal proof of the Ballantine-Merca Conjecture regarding recurrence relations for the parity of the partition function $p(n)$.
Introduction
The partition function $p(n)$, which represents the number of ways to write a positive integer $n$ as a sum of positive integers, is a cornerstone of additive number theory. Despite its fundamental nature, the parity of $p(n)$—whether it is even or odd—remains one of the most significant unsolved mysteries in the field. While the values of $p(n)$ grow exponentially, their distribution modulo 2 appears almost random, yet structured by deep modular properties.
In 2017, Cristina Ballantine and Mircea Merca proposed a conjecture regarding a specific recurrence relation for the parity of $p(n)$. They suggested that a specific sum of partition values is odd if and only if a related linear expression is a perfect square. This research, conducted by Letong Hong and Shengtong Zhang, settles this conjecture by translating the problem into the classification of identities involving a specific family of theta functions modulo 2.
The Theta Function Framework
The study focuses on the family of theta functions defined by: $$f_m(z) := \sum_{mk+1 = x^2} q^{(x^2-1)/m}$$
These functions are closely related to the generating function for partitions. The core challenge was to identify all instances where the product of two such functions is congruent to a third modulo 2: $$f_a(q) \equiv f_b(q) f_c(q) \pmod 2$$
Prior to this work, only a few sporadic examples were known, such as the Jacobi-derived identity $f_4 \equiv f_6 f_{12} \pmod 2$. The authors set out to determine if these were isolated incidents or part of a larger, discoverable structure.
Methodology
1. The Harmonic Constraint
The authors first established a necessary "harmonic relation" between the parameters. They proved that for the identity $f_a \equiv f_b f_c \pmod 2$ to hold, the parameters must satisfy: $$\frac{1}{a} = rac{1}{b} + \frac{1}{c}$$ This constraint significantly narrowed the search space, transforming an open-ended search into a targeted investigation of specific integer relations.
2. Quadratic Form Analysis
The problem was further reduced to analyzing the number of integer solutions to quadratic equations of the form $bz^2 + cy^2 = N$. By representing the coefficients of the theta function products as counts of lattice points on ellipses, the authors could apply tools from analytic number theory.
3. Analytic Bounding and Weber’s Theorem
To prove that only a finite number of sporadic cases exist beyond the infinite families, the authors employed Weber’s theorem (an extension of Dirichlet’s theorem on primes in arithmetic progressions). By selecting prime numbers with specific Legendre symbols, they demonstrated that for sufficiently large parameters, the required parity conditions would inevitably be violated. This effectively bounded the "sporadic" search to cases where the normalized parameters divide 24.
4. Involution Arguments
For the symmetric case where $b=c$, the authors used an involution argument. By defining a mapping that pairs up solutions to the quadratic form, they showed that the parity of the total solution set depends solely on the "fixed points" of the mapping. This elegant combinatorial approach directly revealed the two infinite families of solutions.
Key Results and Classification
The central result of the paper is the complete classification of all triples $(a, b, c)$ satisfying $f_a \equiv f_b f_c \pmod 2$. The identity holds if and only if the triple is:
- Infinite Family I: $(2q, 4q, 4q)$ for any positive odd integer $q$.
- Infinite Family II: $(4q, 8q, 8q)$ for any positive odd integer $q$.
- Sporadic Cases: One of the following eight triples:
- $(4, 6, 12)$
- $(6, 8, 24)$
- $(8, 12, 24)$
- $(10, 12, 60)$
- $(15, 24, 40)$
- $(16, 24, 48)$
- $(20, 24, 120)$
- $(21, 24, 168)$
By establishing this classification, the authors provided the necessary theoretical machinery to confirm the Ballantine-Merca Conjecture. The conjecture's requirements align precisely with the properties of these theta function identities, proving that the suggested recurrence relations for $p(n)$ parity are indeed correct.
Conclusion
This work provides a definitive answer to the Ballantine-Merca Conjecture and offers a complete map of a specific class of modular identities modulo 2. By bridging additive number theory with the geometry of quadratic forms, Hong and Zhang have clarified a small but significant portion of the mystery surrounding the parity of partitions. Future research may look toward classifying identities involving products of $k$ theta functions ($f_a \equiv f_{b_1} f_{b_2} \dots f_{b_k} \pmod 2$), potentially uncovering even deeper patterns in the behavior of the partition function.
Summary updated