🔢 mathematics

A geometric proof of Lagrange's theorem for continued fractions

This paper extends Lagrange's theorem on continued fractions to various higher-dimensional and non-commutative settings by establishing a geometric equivalence between finite or eventually-periodic expansions and the fixed points of parabolic or loxodromic elements within appropriate modular groups.

Original authors: Anton Lukyanenko, Joseph Vandehey

Published 2026-03-16

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Anton Lukyanenko, Joseph Vandehey

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 trying to describe a number using a special kind of recipe. Instead of writing $3.14159...$, you write a list of instructions: "Start with 3, then take the reciprocal of the remainder, add 7, take the reciprocal of that, add 15..." This is called a Continued Fraction.

For a long time, mathematicians knew two big secrets about these recipes for regular numbers (like 1, 2, 3, or $\pi$ ):

Rational numbers (fractions like $3/4$ ) always have recipes that stop after a few steps.
Quadratic surds (roots of simple equations like $\sqrt{2}$ or the Golden Ratio) always have recipes that repeat forever, like a broken record.

This paper, written by Anton Lukyanenko and Joseph Vandey, asks a bold question: Do these rules still work if we leave the flat 2D world of regular numbers and move into higher dimensions, like 3D space, 4D space, or even stranger mathematical universes?

The answer is yes, but to prove it, the authors had to stop thinking about algebra and start thinking about geometry.

The Big Idea: A Geometric Dance

Instead of crunching numbers, the authors imagine these numbers as dancers on a hyperbolic dance floor (a curved, infinite space).

The Modular Group (The Dance Troupe): Imagine a group of dancers (matrices) who can move the floor around. They can slide things (translations) or flip them inside out (inversions).
The Digits (The Moves): When you calculate a continued fraction, you are essentially watching how a point moves under the influence of these dancers.
The Goal: We want to know: "If a dancer's path stops, what kind of dancer are they? If their path loops forever, what kind of dancer are they?"

The Two Types of Dancers

The paper proves that the type of dancer (the mathematical transformation) perfectly matches the type of number's recipe:

The "Parabolic" Dancer (The Finite Recipe):
- What they do: This dancer moves in a way that eventually pushes everything to a single point (infinity).
- The Result: If a number is "fixed" by this dancer (meaning the dancer's move leaves the number unchanged), that number has a finite continued fraction.
- Analogy: Think of a ball rolling down a hill that eventually hits a wall and stops. The ball represents a rational number. The wall is the "infinity" point. The paper shows that in these higher-dimensional worlds, if your number hits the wall, it must be a "rational" type of number.
The "Loxodromic" Dancer (The Repeating Recipe):
- What they do: This dancer spins and stretches the space. They have two special points they love: one they pull toward (attracting) and one they push away (repelling).
- The Result: If a number is "fixed" by this dancer, its continued fraction repeats forever.
- Analogy: Think of a spiral staircase. If you keep walking up, you never stop, but you keep passing the same railing patterns over and over. The paper proves that in 3D, 4D, and even "Quaternion" worlds, if a number is stuck in this spiral, it is the higher-dimensional equivalent of a square root (a quadratic surd).

Why This is a Big Deal

Usually, to prove things about repeating patterns in math, you need to use heavy algebra (like solving complex equations). But the authors realized that in these weird, high-dimensional spaces (like the Heisenberg group or Octonions), the usual algebraic rules get messy or don't exist at all.

So, they used geometry instead.

They imagined "Ford Circles" (or in higher dimensions, "Ford Bubbles") floating in this space.
They showed that if a number's recipe goes on forever without repeating, the bubbles would have to get infinitely small or infinitely large in a way that breaks the rules of the dance floor.
Because the dance floor is "discrete" (it has a grid-like structure), the bubbles can't behave wildly forever. They must eventually settle into a pattern.

The "Magic" of Higher Dimensions

The paper extends these rules to:

3D Space: Numbers that look like $(x, y, z)$ .
4D Space (Quaternions): Numbers with four parts, used in computer graphics and physics.
8D Space (Octonions): Even stranger numbers.

They even found a funny little pattern: In 2D, the number $i$ (the square root of -1) has a special relationship where $i + 1/i = 0$ . The authors found similar "magic" identities for 3D and 4D numbers, showing that these strange higher-dimensional worlds still have their own hidden symmetries.

The Takeaway

This paper is like finding a universal law of motion. It says: "No matter how many dimensions you add to your universe, the relationship between 'stopping' and 'looping' in these number recipes remains the same."

Finite Recipe = You are a "Rational" point fixed by a "Parabolic" move.
Repeating Recipe = You are a "Quadratic" point fixed by a "Loxodromic" spin.

The authors didn't just check the math; they drew a map of the geometry to prove it, showing that even in the most abstract, high-dimensional corners of mathematics, the old rules of Lagrange (a famous mathematician from the 1700s) still hold true, just viewed through a new, geometric lens.

1. Problem Statement

The paper addresses the generalization of Lagrange's Theorem and the characterization of rational numbers within the context of Continued Fractions (CFs) beyond the standard real case.

Classical Context: In standard real CFs, a number has a finite expansion if and only if it is rational, and an eventually periodic expansion if and only if it is a quadratic surd (a root of a non-degenerate quadratic equation with integer coefficients).
The Challenge: Extending these results to higher-dimensional and non-commutative settings (e.g., complex numbers, 3D Euclidean space, quaternions, octonions, and the Heisenberg group) is difficult.
- It is non-trivial to define what constitutes a "quadratic surd" in these generalized spaces.
- Existing algebraic proofs often rely on specific properties of the underlying ring (e.g., commutativity, existence of a multiplicative norm, or ordered structures), which do not exist in all target spaces (like the Heisenberg group or octonions).
Goal: To provide a unified, geometric proof that characterizes finite and periodic CF expansions in terms of the dynamics of the associated modular group acting on a hyperbolic space, applicable to a broad class of "Iwasawa Continued Fractions."

2. Methodology

The authors employ a geometric approach based on the theory of discrete groups acting on hyperbolic spaces, specifically utilizing the Iwasawa CF framework.

Iwasawa CF Framework: The authors define a generalized CF algorithm using:
- An ambient space $X$ (modeled on $k^n \times \text{Im}(k)$ for a division algebra $k$ ).
- A lattice of isometries $Z$ (providing the "digits").
- A fundamental domain $K \subset X$ .
- An inversion map $\iota$ (generalized Koranyi inversion).
- The Generalized Modular Group $M = \langle Z, \iota \rangle$ .
Hyperbolic Geometry: The space $X$ is viewed as the boundary of a negatively curved rank-one symmetric space (hyperbolic space) $\mathcal{H}$ . The group $M$ acts as isometries on $\mathcal{H}$ .
Classification of Isometries: The proof relies on classifying elements of $M$ $M$ into three types based on their fixed points on the boundary $\partial \mathcal{H}$ $\partial H$ :
1. Parabolic: Exactly one fixed point (neither attracting nor repelling).
2. Loxodromic: Exactly two fixed points (one attracting, one repelling).
3. Elliptic: Fixed points inside $\mathcal{H}$ .
Key Geometric Tools:
- Generalized Ford Circles (Horoballs): The authors utilize the fact that for a lattice $M$ , there exist disjoint horoballs (generalized Ford circles) centered at the parabolic fixed points.
- Expansion Properties: The "properness" condition of the CF ensures that the inversion map $\iota$ is uniformly expanding on the fundamental domain $K$ .
- Compactness Arguments: They use the discreteness of $M$ and the compactness of certain sets in hyperbolic space to prove finiteness of orbits.

3. Key Contributions

A. Unification of Finite and Periodic Expansions

The paper establishes a direct equivalence between the arithmetic nature of the expansion and the dynamical nature of the modular group elements:

Finite Expansions: A point $p$ has a finite CF expansion if and only if it is a fixed point of a parabolic element of $M$ .
Eventually Periodic Expansions: A point $p$ has an eventually periodic CF expansion if and only if it is a fixed point of a loxodromic element of $M$ .

This result holds for proper and discrete Iwasawa CFs, covering:

Real nearest-integer CFs.
Complex CFs (Hurwitz).
3D CFs (Euclidean).
Quaternionic CFs.
Octonionic CFs.
Heisenberg CFs.

B. Geometric Proof of Lagrange's Theorem

The authors provide a new proof of Lagrange's Theorem (periodic $\iff$ quadratic surd) that avoids algebraic assumptions about the underlying ring structure.

Mechanism: They show that if a point is a fixed point of a loxodromic element, the orbit under the Gauss map must eventually repeat due to the discreteness of the modular group. Specifically, they construct a sequence of matrices mapping a point into a compact region of hyperbolic space; since $M$ is discrete, only finitely many such matrices exist, forcing a repetition in the orbit.
Advantage: This method works for non-commutative algebras (quaternions, octonions) and non-abelian groups (Heisenberg) where traditional algebraic proofs (relying on discriminants or bounded coefficients) fail.

C. Characterization of "Quadratic Surds"

The paper redefines "quadratic surds" in higher dimensions as fixed points of loxodromic elements.

Real/Complex Cases: They reconcile this with classical theory, showing that fixed points of loxodromic matrices in $SL(2, \mathbb{Z})$ or $SL(2, \mathbb{Z}[i])$ are exactly the roots of non-degenerate quadratic equations.
Higher Dimensions: They utilize Clifford Algebra formalism (Ahlfors/Vahlen) to represent these transformations.
- For $\mathbb{R}^3$ (Quaternions), they identify the modular group with $SL(2, \mathbb{Z}_4)$ (Clifford matrices over Hurwitz integers) and use Parker and Short's classification to identify loxodromic elements.
- They demonstrate that fixed points of these loxodromic matrices correspond to roots of "non-degenerate quadratics" defined via Clifford algebra.

4. Key Results

Theorem 1.1: The main theorem stating the equivalence between finite/periodic expansions and parabolic/loxodromic fixed points for proper, discrete Iwasawa CFs.
Lemma 3.1 & 3.6: Rigorous proofs establishing the bidirectional implications (e.g., finite expansion $\implies$ parabolic fixed point; parabolic fixed point $\implies$ finite expansion via the expansion of horoballs).
Theorem 4.1 & 4.4: Explicit proofs for Real and Complex cases showing the equivalence between loxodromic fixed points and quadratic surds.
Proposition 4.7: Application of Parker and Short's classification to categorize quaternionic matrices, confirming that the modular group elements corresponding to periodic CFs are indeed loxodromic.
Proposition 5.1 & 5.2: A discussion on generalized identities (e.g., $i + 1/i = 0$ ) in quaternions, showing that while certain periodic identities hold in dimensions 2, 3, and 4, they fail for $d > 4$ .

5. Significance

Generalization: This is the first work to successfully extend Lagrange's Theorem to octonionic and Heisenberg continued fractions, spaces where the lack of associativity or commutativity previously obstructed such results.
Methodological Shift: By shifting from algebraic manipulation (discriminants, coefficients) to geometric dynamics (isometries, horoballs, discreteness), the authors bypass the need for the underlying space to be a commutative ring or an ordered field.
New Definition of Quadratic Surds: The paper provides a robust, geometric definition of "quadratic surds" in higher dimensions: they are simply the fixed points of loxodromic isometries in the associated modular group.
Unification: It unifies the study of CFs across real, complex, and non-commutative settings under a single geometric framework (Iwasawa CFs), highlighting the deep connection between number theory and hyperbolic geometry.

In summary, the paper successfully generalizes one of the most fundamental theorems in the theory of continued fractions to high-dimensional and non-commutative settings by leveraging the geometric properties of discrete groups acting on hyperbolic spaces.