🔢 mathematics

Cubic Polynomials and Sums of Two Squares

This paper establishes a quantitative lower bound for the frequency of irreducible monic cubic polynomials with negative discriminant taking values that are sums of two squares, thereby resolving a question posed by Grechuk regarding the infinitude of such values through the application of two-dimensional unit arguments and the arithmetic of degree six number fields.

Original authors: Siddharth Iyer

Published 2026-05-19

📖 6 min read🧠 Deep dive

CC BY 4.0

Original authors: Siddharth Iyer

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

The Big Picture: The "Magic Recipe" Hunt

Imagine you have a magical recipe book. Each recipe is a specific mathematical formula called a cubic polynomial. If you plug in a whole number (like 1, 2, 3, or -5) into this recipe, it spits out a result.

The question this paper asks is: How often does this recipe produce a number that can be built by adding two perfect squares together?

In math terms, a number is a "sum of two squares" if it looks like $A^2 + B^2$ (for example, $5 = 1^2 + 2^2$ , or $13 = 2^2 + 3^2$ ). Some numbers, like 3 or 7, cannot be made this way.

The author, Siddharth Iyer, is trying to prove that for certain types of these magical recipes, the "sum of two squares" numbers appear frequently enough to be considered infinite. He doesn't just say "there are infinitely many"; he gives a specific estimate of how many you can find if you look at all numbers up to a certain size.

The Main Characters

The Polynomial ( $P(x)$ ): Think of this as a machine. You put a number $n$ in, and it outputs $n^3 + \text{some other stuff}$ . The paper focuses on machines where the output is an "irreducible" cubic (meaning the machine's formula can't be broken down into simpler, smaller machines).
The Target: We want the machine's output to be a "Sum of Two Squares."
The Problem: It's hard to predict when a machine will spit out a "Sum of Two Squares." Sometimes they appear, sometimes they don't.

The Author's Strategy: Building a "Bridge"

The author realizes that checking every single number one by one is impossible. Instead, he builds a bridge to a different, easier-to-understand world.

The Old World (Integers): This is where we count $1, 2, 3...$ and check if $P(n)$ is a sum of squares.
The New World (Complex Numbers & Units): The author creates a special mathematical landscape involving "complex numbers" (numbers with a real part and an imaginary part, like $3 + 4i$ ). In this landscape, there are special numbers called units.

The Analogy:
Imagine you are trying to find rare gold coins (sums of two squares) in a giant, dark cave (the integers). It's hard to see them.
Instead, the author builds a tunnel (the bridge) that leads to a bright, well-lit room (the degree-six number field). In this bright room, the gold coins are actually just shiny mirrors reflecting light.

By studying the "shiny mirrors" (the units) in the bright room, he can prove that there must be a lot of gold coins in the dark cave.

The "Two-Dimensional Unit" Argument

The core of the proof relies on a concept called Units. In the bright room, these units act like gears in a clock.

Some gears spin in a way that changes the size of things (getting bigger or smaller).
Other gears spin in a way that keeps the size exactly the same (like a circle).

The author proves that in this specific mathematical room, there are two independent gears (a two-dimensional unit group) that can be combined in infinite ways. By turning these gears just right, he can generate a massive family of numbers that are guaranteed to be "sums of two squares."

He shows that these generated numbers are "sparse" enough that they don't overlap too much, but "dense" enough that they cover a significant portion of the number line.

The Results: What Did He Find?

The paper proves a specific lower bound. If you take a polynomial that fits certain rules (like having specific even/odd coefficients), and you look at all numbers up to a huge number $X$ :

The Result: The number of times the polynomial produces a "sum of two squares" is roughly $X^{1/3}$ .
What that means: If you look at the first 1,000,000 numbers ( $10^6$ ), the formula predicts you will find roughly $(10^6)^{1/3} = 100$ solutions. If you look at the first 1 billion ( $10^9$ ), you'll find roughly 1,000 solutions.

This answers a question posed by another mathematician (Grechuk) who asked: "Are there infinitely many numbers where $n^3 - 2$ is a sum of two squares?"
The Answer: Yes, and here is exactly how many you can expect to find.

The "Magic Substitution" (The Polynomial Trick)

The paper also shows something cool: You can actually write down a new, complicated polynomial formula (let's call it $R(t)$ ) such that if you plug any integer $t$ into it, the result $R(t)^3 - 2$ is guaranteed to be a sum of two squares.

Analogy: Imagine someone asks, "Can you make a machine that always prints a number made of two squares?"
The Paper's Answer: "Yes, here is a very complex machine (a degree-9 polynomial) that does exactly that."
The paper provides the specific, messy coefficients for this machine, showing it's not just a theoretical possibility, but a concrete construction.

Limitations and Boundaries

The author is careful to say what his bridge doesn't reach:

Three Real Roots: The bridge only works if the polynomial has one real root and two "imaginary" roots. If the polynomial has three real roots, the bridge collapses.
Other Shapes: The paper focuses on "sums of two squares" ( $x^2 + y^2$ ). It doesn't fully solve the problem for other shapes like $x^2 + 3y^2$ , though the author suggests the method might be adaptable.
Optimality: The author admits that for some specific polynomials, you might find even more solutions than his formula predicts (up to $X^{1/2}$ ), but for a general case, $X^{1/3}$ is the safe, proven lower limit.

Summary

In simple terms, Siddharth Iyer built a mathematical bridge from a difficult problem (finding sums of squares in cubic polynomials) to an easier problem (counting special gears in a complex number system). By proving that these gears can be turned in infinite, non-overlapping ways, he proved that the original problem has infinitely many solutions and gave a precise estimate of how frequent they are. He also constructed a specific, complex formula that acts as a "guaranteed generator" for these solutions.

Technical Summary: Cubic Polynomials and Sums of Two Squares

Problem Statement
This paper addresses the frequency with which an irreducible monic cubic polynomial $p(x) \in \mathbb{Z}[x]$ with negative discriminant takes values that are sums of two squares. Specifically, let $B_p(X)$ denote the set of integers $n \le X$ such that $p(n)$ is a sum of two squares. The central question, motivated by a query from Grechuk (2021) regarding the infinitude of such values for $p(x) = x^3 - 2$ , is to establish a quantitative lower bound for $|B_p(X)|$ . The paper seeks to determine if there are infinitely many such integers and, if so, to provide an asymptotic estimate for their density.

Methodology
The proof strategy relies on constructing a "sufficiently dense" family of degree-six integer-valued polynomials $R(t)$ such that $p(R(t))$ is a sum of two squares for every integer $t$ . The methodology proceeds through the following steps:

Algebraic Reformulation: The problem is translated into the ring of integers $\mathbb{Z}[\theta]$ , where $\theta$ is a root of $p(x)$ . The condition $p(n) \in \mathbb{Z}^2$ (sum of two squares) is linked to the existence of elements in $\mathbb{Z}[\theta]$ whose norm-like structure relates to sums of squares. Specifically, the paper investigates when $q(n - \theta)$ is a sum of two squares in $\mathbb{Z}[\theta]$ for some $q \in \mathbb{Z}^2 \setminus \{0\}$ .
Unit Arguments in Degree-Six Fields: A critical component of the analysis involves the arithmetic of the degree-six number field $K = \mathbb{Q}(\theta, i)$ . Since $p(x)$ has one real root and two complex conjugate roots, $K$ is a CM-field. By Dirichlet's Unit Theorem, the unit group of the ring of integers $\mathcal{O}_K$ has rank 2. The proof utilizes a two-dimensional unit argument, decomposing units into a finite set of torsion elements and powers of two fundamental units: one with complex modulus $\neq 1$ and one with modulus $1$.
Counting Lattice Points: The authors define sets $R_{\theta, q}(Y_1, Y_2)$ representing pairs of elements in $\mathbb{Z}[\theta]$ with bounded coefficients whose squares sum to $q(n-\theta)$ . The core of the proof involves establishing lower bounds for the cardinality of these sets under specific parity conditions on the coefficients of $p(x)$ .
Parity Conditions: The analysis is split into two cases based on the coefficients $a_2, a_1, a_0$ $a_{2}, a_{1}, a_{0}$ of $p(x) = x^3 + a_2x^2 + a_1x + a_0$ $p (x) = x^{3} + a_{2} x^{2} + a_{1} x + a_{0}$ :
- Case 1: $a_2^2 - a_1$ is even, and $a_1a_2 - a_0$ is odd.
- Case 2: $a_0 \equiv 2 \pmod 4$ , and $a_1, a_2 \equiv 0 \pmod 4$ .

Key Contributions and Results
The paper establishes the following main theorem:

Theorem 1.1: If $p(x) = x^3 + a_2x^2 + a_1x + a_0$ is an irreducible cubic polynomial satisfying either Case 1 or Case 2 above, and has one real and two complex roots, then:
$|B_p(X)| \gg X^{1/3 - o(1)}$
This result provides a quantitative answer to Grechuk's question, confirming that for $p(x) = x^3 - 2$ (which satisfies Case 2), the number of integers $n \le X$ such that $n^3 - 2$ is a sum of two squares is at least $X^{1/3 - o(1)}$ .

Additionally, the paper provides explicit polynomial substitutions $R(t), S_1(t), S_2(t)$ demonstrating that $R(t)^3 - 2 = S_1(t)^2 + S_2(t)^2$ holds identically. These polynomials were derived using the system $T$ described in the proof of Lemma 1.6.

Significance and Limitations
The paper claims significance in providing a quantitative lower bound for a class of Diophantine equations involving cubic polynomials and sums of two squares, a problem previously lacking such estimates. The result builds upon the work of Friedlander and Iwaniec regarding quadratic polynomials but extends the difficulty to the cubic case, where the geometry of the associated Châtelet surface is more complex.

The authors note several limitations and modest claims:

Optimality: The bound $X^{1/3 - o(1)}$ is not always optimal. If a polynomial substitution exists such that $p(q_0(x)) = q_1(x)^2 + q_2(x)^2$ for a quadratic $q_0$ , the count would be $X^{1/2}$ . The authors found such substitutions for specific cases (e.g., $x^3 + 17$ ) but could not find them for arbitrary monic cubics, leaving the general case open.
Scope: The arguments are specific to irreducible monic cubic polynomials with negative discriminant (one real, two complex roots). The paper explicitly states that the arguments cannot be straightforwardly generalized to polynomials with three distinct real roots.
Generalization: While the lower bound arguments are expected to generalize to other quadratic forms $x^2 + ny^2$ with negative discriminant, the theory breaks down when $n$ is negative due to insufficient information about the density of units in the relevant rings.
Coverage: The current results cover approximately 26.5625% of degree-three monic irreducible polynomials with negative discriminant (those satisfying the specific parity conditions). The authors suggest that covering all cases would require more complex parity arguments involving sets $R_{\theta, 2^t}$ , which are not pursued in this work.

The paper concludes by posing Question 1.2: whether for any irreducible monic cubic polynomial, there exists an integer-valued quadratic polynomial $q_0$ and rational polynomials $q_1, q_2$ such that $p(q_0(x)) = q_1(x)^2 + q_2(x)^2$ .