🔢 mathematics

Integral Shell Polytopes of Composition Algebras

This paper investigates integral shell polytopes in real composition algebras, revealing that while classical Hurwitz systems recover familiar root polytopes, the Okubo algebra yields a distinct two-adic hierarchy of cubic-lattice orbits that, through maximal-isotropic gluing, reconstructs the full $E_8$ Gosset polytope.

Original authors: Daniele Corradetti

Published 2026-05-12

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Daniele Corradetti

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 have a magical, multi-dimensional building made of math. In this building, every room represents a specific "size" or "weight" of a number. Usually, when mathematicians look at these buildings, they only care about the very first floor—the smallest, most basic rooms. They assume that if you know the first floor, you know the whole building.

This paper, written by Daniele Corradetti, is about checking the entire building, floor by floor, to see what happens when you change the rules of how the building is constructed.

Here is the story of the paper, broken down into simple concepts:

1. The Two Types of Buildings

The author compares two different types of mathematical structures:

The Classic Building (Octonions): Think of this as a famous, perfectly symmetrical castle. It has a "front door" (a unit element) and follows strict, alternating rules. When you look at the first floor of this castle, you find a beautiful, complex shape called the Gosset Polytope (which has 240 corners). This shape is famous and well-known.
The New Building (Okubo Algebra): This is a strange, new type of castle. It doesn't have a front door (no unit element) and its rules are a bit more flexible. The author asks: "If we build a castle with these new rules, does the first floor look like the famous Gosset castle?"

2. The Surprise: The First Floor is Different

The answer is a big no.

In the Classic Building, the first floor is full of 240 points, forming that famous complex shape.
In the New (Okubo) Building, the first floor is empty. There are no rooms there at all!
The first time you find any rooms in the New Building is on the 4th floor. Even then, it's not the complex Gosset shape. It's a much simpler shape: an 8-dimensional cross (like a star with 8 arms pointing in opposite directions). It only has 16 corners.

3. The "Conductor" Filter

Why is the New Building so different? The paper explains that the rules of the Okubo algebra act like a strict filter or a sieve.

Imagine the Classic Building is a sieve with large holes. It lets the smallest numbers (norm 1) pass through immediately.
The Okubo Building has a sieve with tiny holes. It blocks all the small numbers. It only lets numbers through if they are divisible by 4.
Because of this filter, the "first floor" (norm 1) is empty. The first things that get through are the "norm 4" items, which form that simple cross shape.

4. The Hidden Ladder (The Intermediate Step)

The author discovers a secret ladder that connects the New Building to the Classic one.

If you take the New Building and shrink it down by half (dividing everything by 2), it transforms into something very familiar: a scaled-up grid (like a giant 3D graph paper extended into 8 dimensions).
On this grid, the shapes become very easy to understand. The "cross" on the first floor is just a simple set of points. The next floor up (which had 112 points) turns out to be a specific pattern of points on this grid.
The paper shows that every single layer of this New Building can be described as a collection of points moving in specific patterns on this grid. It's like realizing that a complex dance routine is just a series of simple steps on a dance floor.

5. Rebuilding the Masterpiece

So, does the New Building ever look like the Classic Gosset castle?

Not directly. The New Building starts with a simple cross, then a slightly more complex shape, then bigger and bigger groups of points.
But... if you take this "shrunk" grid version of the New Building and perform a specific mathematical "gluing" operation (connecting specific missing pieces together), you can rebuild the full Gosset castle with all 240 corners.
The paper concludes that the New Building doesn't replace the Classic one; it reveals a hidden, step-by-step path to it. The Classic castle is the final destination, but the New Building shows you the intermediate stops (the cross, the grid, the gluing) that you usually skip over.

Summary

The paper is a detective story about mathematical shapes. It proves that while the famous "Classic" math structures give you a complex shape immediately, a "New" type of structure forces you to go through a simpler, grid-like phase first.

Old View: "The first layer is the Gosset shape."
New View: "The first layer is empty, the second is a cross, the third is a grid pattern, and only after gluing them all together do you get the Gosset shape."

The author uses this to show that even in advanced math, looking at the "middle steps" (the intermediate layers) reveals a richer, more detailed picture than just looking at the start or the finish.

Technical Summary: Integral Shell Polytopes of Composition Algebras

Problem Statement
The paper investigates the finite metric configurations arising from integral elements of fixed norm within real composition algebras. While it is well-established that unital composition algebras (Hurwitz algebras: $\mathbb{R}, \mathbb{C}, \mathbb{H}, \mathbb{O}$ ) give rise to familiar root systems and regular polytopes (e.g., $E_8$ from the Coxeter-Dickson octonions), the behavior of non-unital composition algebras, specifically the Okubo algebra, remains less understood in this context. The central problem is to determine whether the integral closure of the Okubo algebra yields the $E_8$ Gosset polytope directly, or if it produces a distinct hierarchy of shell polytopes. The author distinguishes between the algebraic structure, the integral order, the underlying metric shadow (lattice), and the resulting shell polytopes, noting that these distinctions collapse in classical Hurwitz cases but separate significantly in the Okubo case.

Methodology
The study employs a "shell-polytope" viewpoint, defined by fixing a norm value $N$ and analyzing the convex hull of the corresponding integral elements.

Definitions: An integral system $\Lambda$ is defined as a full-rank discrete $\mathbb{Z}$ -lattice within a composition algebra where the norm is integral. The norm shell $S_N(\Lambda)$ is the set of vectors with norm $N$ , and the shell polytope $P_N(\Lambda)$ is their convex hull.
Metric Shadow: The analysis focuses on the "metric shadow" of the algebraic order—the lattice generated by the integral basis under the trace/norm form, rather than the full multiplicative structure.
Comparative Analysis: The paper compares classical Hurwitz systems with the Okubo algebra. For the Okubo case, the author utilizes a specific arithmetic model where the product forces arithmetic over $\mathbb{Q}(\sqrt{3})$ .
Computational Verification: Exact enumeration of integer quadratic forms is used to count shell vectors. Root system identification is performed strictly via Cartan matrices and reflection orbit closure, not merely by vertex counts.
Lattice Isometry and Gluing: The intermediate lattice $M = (1/2)LOk$ is analyzed to determine its isometry class and its relationship to the unimodular $E_8$ lattice via 2-adic gluing.

Key Contributions and Results

Reinterpretation of Classical Tables: The paper proposes a revised interpretation of classical integral systems based on the "first non-empty norm shell." Under this convention, systems like the Gaussian integers and Cayley-Graves octonions are shown to have first shells of type $A_1^2$ and $A_1^8$ respectively, rather than the traditionally cited non-simply-laced types ( $C_2, C_8$ ), which require higher shells to manifest.
The Okubo Conductor Lattice ($LOk$): The integral order of the Okubo algebra does not yield $E_8$ $E_{8}$ directly. Instead, it produces a conductor sublattice $LOk \subset E_8$ $L O k \subset E_{8}$ with index $[E_8 : LOk] = 2^{12}$ $[E_{8} : L O k] = 2^{12}$ and determinant $2^{24}$ $2^{24}$ .
- Norm Divisibility: A critical result is that $LOk$ contains no elements of norm 1. The first non-empty shell occurs at norm $N=4$ (squared length 8).
- Shell Hierarchy:
  - $S_4(LOk)$ : Contains 16 vectors forming an 8-dimensional cross-polytope ( $A_1^8$ ).
  - $S_8(LOk)$ : Contains 112 vectors forming the $D_8$ root polytope.
  - Higher shells ( $S_{12}, S_{16}$ ) contain 448 and 1136 vectors respectively.
Identification of the Intermediate Lattice ( $M$ ): The lattice $M = (1/2)LOk$ is identified as isometric to the rescaled cubic lattice $\sqrt{2}\mathbb{Z}^8$ $2 Z^{8}$ .
- This identification allows for a complete combinatorial description of the shells as unions of orbits under the hyperoctahedral group $W(B_8)$ .
- For example, the shell $S_{12}(LOk)/2$ corresponds to the single $W(B_8)$ -orbit of type $(\pm 1, \pm 1, \pm 1, 0, 0, 0, 0, 0)$ .
Recovery of $E_8$ : The full $E_8$ Gosset polytope (240 vertices) is not the first shell of $LOk$. It is recovered only after passing to the intermediate lattice $M$ and performing a maximal-isotropic gluing along the quotient group $(\mathbb{Z}/2\mathbb{Z})^4$ .

Significance and Claims
The paper claims that the Okubo algebra does not provide a direct alternative realization of the $E_8$ root system. Instead, it reveals a "two-adic hierarchy" where the geometry is filtered through a conductor lattice.

Separation of Phenomena: The work demonstrates that in non-unital composition algebras, the algebraic integral closure, the metric lattice, and the root system are distinct stages. The "separation" of these stages is the primary phenomenon measured.
Cubic Lattice Bridge: A significant finding is that the Okubo hierarchy, when normalized, maps to the classical cubic lattice $\sqrt{2}\mathbb{Z}^8$ . This provides an elementary combinatorial explanation for the shell counts and symmetries observed in the Okubo system, which were previously viewed as mysterious or distinct from standard root systems.
Strict Classification: The paper emphasizes a rigorous distinction between "lattice-shell polytopes" (finite convex hulls of lattice points) and "root polytopes" (convex hulls of verified root systems). While $S_4$ and $S_8$ of the Okubo system correspond to $A_1^8$ and $D_8$ root systems, higher shells are classified as unions of $W(B_8)$ orbits but are not assigned root system types as they lack the necessary reflection structure.

The author concludes that the study of integral shell polytopes offers a refined language to compare classical and non-unital composition algebras, revealing that the $E_8$ geometry in the Okubo case is an emergent property of a gluing process rather than a direct consequence of the algebra's minimal norm elements.