🔢 mathematics

On the unit equation $\varepsilon+\delta=n$ in cubic fields

The paper proves that for a non-zero integer $n \neq 2$ , the unit equation $\varepsilon + \delta = n$ has no solutions in units of 100% of cubic fields when ordered by discriminant, confirming a recent conjecture by the authors.

Original authors: Maleeha Khawaja, Samir Siksek

Published 2026-06-02

📖 4 min read🧠 Deep dive

CC BY 4.0

Original authors: Maleeha Khawaja, Samir Siksek

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

Imagine you are a detective trying to solve a very specific, tricky puzzle involving numbers. This paper is about a team of mathematicians (Khawaja and Siksek) who investigated a famous equation: $\varepsilon + \delta = n$ .

Here, $n$ is just a regular whole number (like 3, 5, or 100), and $\varepsilon$ and $\delta$ are special numbers called "units." Think of units as the "VIPs" of a number system—they are numbers that can be multiplied and divided within that system without ever leaving it or creating fractions.

The big question the authors asked is: If we pick a random "world" of numbers (specifically, a cubic field, which is a complex 3-dimensional number system), how often can we find two VIPs that add up to our target number $n$ ?

The Main Discovery: The "Empty Room" Phenomenon

The authors found a surprising statistical pattern. They proved that if you look at 100% of all possible cubic number worlds (when you arrange them by size), you will find zero solutions to this puzzle.

To use an analogy: Imagine there are infinite hotels, each with a different number of rooms (representing different number fields). The authors are asking: "In how many of these hotels can we find two specific guests (the VIPs) who, when they meet in the lobby, have a combined 'weight' exactly equal to $n$ ?"

Their answer is: Almost never. If you walk through a random selection of these hotels, you will almost certainly find that the lobby is empty. The equation simply doesn't work in the vast majority of these number worlds.

The Two Sides of the Story

The paper presents two sides to this story, like a coin:

The Exception (Part I): The authors admit that there are some hotels where the guests do meet. In fact, there are infinitely many cubic fields where a solution exists. However, these are the rare, special cases.
The Rule (Part II): The main headline is that for every single integer $n$ (except for the number 2, which is a special case they handle differently), the equation fails to have a solution in 100% of cubic fields.

How They Compared Different Worlds

To make this even more interesting, they compared these 3-dimensional "cubic" worlds to 2-dimensional "quadratic" worlds (which are simpler, like the world of square roots).

In the 2D world (Quadratic fields): The puzzle is solvable, but only in a very small, finite number of specific worlds. It's like finding a needle in a haystack, but the haystack is small.
In the 3D world (Cubic fields): The puzzle is solvable in an infinite number of worlds, but those solvable worlds are so sparse that if you pick one at random, the odds of it being a solvable one are effectively zero.

The "Fingerprint" Method

How did they prove this? They didn't check every single number world (because there are infinite of them). Instead, they used a clever mathematical tool called a polynomial.

Think of the equation $\varepsilon + \delta = n$ as a machine that generates these number worlds. The authors created a master formula (a polynomial) that acts like a fingerprint scanner.

They analyzed the "shape" of this fingerprint.
They discovered that the fingerprint has a very specific, chaotic structure (mathematically, it relates to a complex curve with no rational points).
Because of this chaotic structure, a famous theorem by Bhargava, Taniguchi, and Thorne tells us that the "machines" (the number fields) generated by this formula are incredibly rare. They are so rare that they make up 0% of the total population when ordered by size.

The Bottom Line

The paper confirms a conjecture the authors made earlier: Hyperbolic curves (a fancy mathematical term for these specific types of number puzzles) are "Diophantine stable."

In plain English: These puzzles are incredibly stubborn. While they can be solved in a few special, constructed cases, if you walk into the universe of cubic number fields and pick one at random, you will almost certainly find that the equation $\varepsilon + \delta = n$ has no solution. The "VIPs" simply refuse to meet in the lobby for 99.99...% of the time.

The authors also provided the computer code they used to verify these calculations, inviting others to check their work, ensuring that this "statistical silence" is a real feature of the mathematical universe.

Technical Summary: On the Unit Equation $\varepsilon + \delta = n$ in Cubic Fields

Problem Statement
The paper investigates the unit equation $\varepsilon + \delta = n$ within the context of cubic number fields, where $\varepsilon$ and $\delta$ are units in the ring of integers $\mathcal{O}_L$ of a cubic field $L$ , and $n$ is a non-zero integer. This equation corresponds to finding $L$ -points on the punctured curve $X = \mathbb{P}^1 \setminus \{0, n, \infty\}$ . Specifically, the authors focus on the set of "new" points, denoted $X(\mathcal{O}_L)^{\text{new}}$ , which are points defined over $L$ but not over any strictly smaller extension of $\mathbb{Q}$ . The central question is the statistical behavior of these solutions when cubic fields are ordered by their absolute discriminant.

Methodology
The authors employ a combination of explicit algebraic parametrization and arithmetic statistics.

Parametrization of Solutions:
The paper first establishes a correspondence between solutions to the unit equation and the roots of specific cubic polynomials.
- For $n \geq 3$ , Lemma 5 demonstrates that if $L = \mathbb{Q}(\varepsilon)$ is a cubic field, the minimal polynomial of $\varepsilon$ must take the form $X^3 + aX^2 - (n^2 + an)X + 1$ for some integer $a$ .
- For $n=2$ and $n=1$ , Lemmas 6 and 7 provide similar explicit parametrizations, identifying specific families of cubic polynomials whose roots generate the relevant fields.
- The authors also analyze the quadratic case (Section 2), proving that for any $n$ , there are only finitely many quadratic fields containing solutions, contrasting this with the cubic case.
Arithmetic Statistics and Galois Theory:
To prove the statistical result, the authors utilize a recent theorem (Theorem 8) derived from their previous work [6] and results by Bhargava, Taniguchi, and Thorne [1]. This theorem provides a criterion for determining when a family of cubic fields defined by a polynomial $F(t, X)$ constitutes a "thin" set (i.e., 0% of all cubic fields ordered by discriminant).
- The authors construct a bivariate polynomial $F(T, X)$ based on the parametrizations found in the lemmas.
- They analyze the discriminant of $F$ with respect to $X$ , denoted $\Delta_X(F)$ . They factor this discriminant into $g(T) \cdot h(T)^2$ , where $g(T)$ is square-free.
- The proof relies on verifying two conditions for $g(T)$ $g (T)$ :
  1. $g(T)$ has no rational roots (verified via the study of an associated genus 1 curve $E: Y^2 + Y = X^3$ ).
  2. The Galois group $G = \text{Gal}(g)$ contains an element that acts freely on the roots of $g$ .

Key Contributions and Results
The paper presents two main results regarding the equation $\varepsilon + \delta = n$ :

Theorem 3(I): There exist infinitely many cubic fields $L$ for which the set of new points $(\mathbb{P}^1 \setminus \{0, n, \infty\})(\mathcal{O}_L)^{\text{new}}$ is non-empty. This is established by the explicit construction of infinite families of cubic fields via the parametrizations in Lemmas 5, 6, and 7.
Theorem 3(II): When cubic fields are ordered by absolute discriminant, 100% of them satisfy $(\mathbb{P}^1 \setminus \{0, n, \infty\})(\mathcal{O}_L)^{\text{new}} = \emptyset$ $(P^{1} ∖ {0, n, \infty}) (O_{L})^{new} = \emptyset$ .
- This result is proven by showing that the family of cubic fields generated by the solutions to the unit equation falls into the "thin" set described by Theorem 8.
- The authors verify that for $n \geq 3$ , the discriminant polynomial $g(T)$ has no rational roots and its Galois group contains a fixed-point-free element. The cases $n=1$ and $n=2$ are noted to be similar but more elementary.

Significance and Claims
The paper positions its findings as further evidence supporting the "statistical Diophantine stability conjecture for hyperbolic curves" (Conjecture 1), which posits that for a hyperbolic curve $X$ , the set of $L$ -new points is empty for 100% of extensions $L$ in an allowable family.

Consistency with Conjecture: The result confirms that for the specific hyperbolic curve $\mathbb{P}^1 \setminus \{0, n, \infty\}$ , the conjecture holds for the family of cubic fields.
Contrast with Quadratic Fields: The authors highlight a distinct difference between quadratic and cubic fields. While the quadratic analogue of Theorem 3(II) holds trivially (as there are only finitely many quadratic solutions), the cubic case requires a non-trivial statistical argument to show that solutions are rare (density 0) despite being infinite in number.
Modesty: The authors frame the work as providing "further evidence" for the broader conjecture rather than a complete proof of the conjecture for all hyperbolic curves. They explicitly reference their previous work [6] for the general framework and the specific case of $n=1$ (Theorem 2), noting that the current paper generalizes this to arbitrary non-zero integers $n$ .

The paper concludes by noting that the supporting computational code (Magma) used to verify the properties of the discriminant polynomials is available in a public repository.

The Main Discovery: The "Empty Room" Phenomenon

The Two Sides of the Story

How They Compared Different Worlds

The "Fingerprint" Method

The Bottom Line

More like this