🔢 mathematics

Preserving Hodge Vectors of Lattice Polytopes

This paper establishes a formula linking the Hodge vector of a Cayley polytope to the mixed volume of its components and the Hodge vector of a projected polytope, utilizing this result to construct infinitely many high-dimensional lattice polytopes with identical Hodge vectors and to resolve open questions regarding thin polytopes.

Original authors: Vadym Kurylenko, Benjamin Nill

Published 2026-06-24

📖 4 min read🧠 Deep dive

CC BY 4.0

Original authors: Vadym Kurylenko, Benjamin Nill

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 an architect working with a special kind of building block called a lattice polytope. These are shapes made of straight lines and flat faces, but with a very strict rule: every corner (vertex) must sit exactly on a grid point, like a dot on graph paper.

Mathematicians love these shapes because they hold secrets about numbers and geometry. One of the most important secrets they hold is something called a Hodge vector. Think of the Hodge vector as a shape's "fingerprint" or its "DNA." It's a list of numbers that describes the shape's hidden complexity and how it behaves in higher-dimensional spaces.

Usually, if you change the size of a shape or add a new dimension to it, its fingerprint changes completely. However, this paper by Vadym Kurilenko and Benjamin Nill discovers a magical way to build bigger, more complex shapes that keep the exact same fingerprint as a smaller, simpler one.

Here is the breakdown of their discovery using simple analogies:

1. The Problem: Changing the Shape, Keeping the Soul

In the world of these grid-shapes, there is a known trick called a "free join." Imagine taking a shape and sticking a long, thin stick (a line segment) to it. If you do this in a specific way, the new, taller shape keeps the same fingerprint as the original. But this is a bit boring; it's like just adding a handle to a cup. The mathematicians wanted to know: Are there other, more interesting ways to build a bigger shape that still looks exactly the same (mathematically speaking) as the original?

2. The Solution: The "Lawrence Twist"

The authors found a new construction they call a Lawrence Twist (and a more general version called a Generalized Lawrence Twist).

The Analogy:
Imagine you have a flat, 2D drawing of a house (your original shape).

The Old Way (Free Join): You just build a second identical house on top of the first one, connected by a single elevator shaft. It's a tall tower, but it's just two houses stacked.
The New Way (Lawrence Twist): Imagine you take your house and wrap it inside a giant, transparent, multi-layered origami structure. You add "ghost layers" that are perfectly balanced (symmetric) around the center. You twist the structure slightly.

The result is a massive, high-dimensional structure that looks completely different from the outside. It has more corners, more faces, and exists in a much higher dimension. But, if you run the mathematical "DNA test" (calculate the Hodge vector), the result is identical to your original small house.

3. How It Works (The "Cayley" Recipe)

The paper explains that you can create these shapes by mixing different smaller shapes together in a specific recipe called a Cayley polytope.

Think of it like a smoothie. You have a base fruit (your original shape, $P$ ) and you add some "flavor enhancers" (other shapes $P_1, \dots, P_k$ ).
The authors proved that if you choose these flavor enhancers carefully (specifically, if they fit inside a specific lower-dimensional space and have a "mixed volume" of 1), the resulting smoothie (the new giant shape) tastes exactly the same as the base fruit.
The size of the new shape is determined by how many flavor enhancers you add, but the "taste" (the Hodge vector) remains unchanged.

4. Why This Matters

The authors use this discovery to answer some big questions in the field:

Infinite Variety: They showed that you can create infinitely many different high-dimensional shapes that all share the same fingerprint. Before this, people thought there were only a few ways to do this (mostly just stacking shapes). Now we know there are endless variations.
"Thin" Shapes: In this math world, a "thin" shape is one whose fingerprint is essentially empty or zero (it's very simple in a complex way). The paper shows that there are many more of these "thin" shapes than we thought. They aren't just simple stacks; they can be complex, twisted structures that happen to have a zero fingerprint.
Solving Puzzles: They used this method to solve specific questions left open by other mathematicians (like questions from Borger, Kretschmer, Nill, and Selyanin) about whether certain types of shapes could exist. The answer is yes, and here is how to build them.

The Bottom Line

This paper is like discovering a new type of mathematical cloning machine. You can take a small, simple shape and use a "Lawrence Twist" to expand it into a giant, high-dimensional monster. Even though the monster is huge and complex, its mathematical soul (its Hodge vector) is identical to the small original. This proves that the universe of these grid-shapes is much more diverse and flexible than anyone realized.

Technical Summary: Preserving Hodge Vectors of Lattice Polytopes

Problem Statement
In the context of Ehrhart theory and the study of hypergeometric motives, the Hodge vector of a lattice polytope $P$ is defined as the local $h^*$ -vector (or $\ell^*$ -vector) with leading and trailing zeroes removed. This vector corresponds to the coefficient vector of the top-degree part of the Hodge-Deligne polynomial of the primitive cohomology of the associated generic hypersurface in the algebraic torus. A central question in this field, raised by Borger, Kretschmer, and the second author, concerns the existence of lattice polytopes of different dimensions sharing the same Hodge vector.

While classical Ehrhart theory establishes that lattice pyramids preserve the $h^*$ -polynomial (and thus the $h^*$ -vector), the behavior of the Hodge vector under dimension changes is distinct. It is known that the Hodge vector of any lattice pyramid vanishes. Conversely, taking the free join with a lattice interval of length 2 preserves the Hodge vector. The paper addresses the open problem of whether there exist other combinatorial constructions that preserve Hodge vectors, specifically seeking infinitely many high-dimensional lattice polytopes with the same Hodge vector that are not free joins.

Methodology
The authors develop a general construction based on the Cayley polytope and mixed volumes. The core of their methodology involves:

Cayley Polytopes and Projections: Given lattice polytopes $P_1, \dots, P_k$ contained in a $k$ -dimensional rational subspace $U \subseteq \mathbb{R}^d$ and a $d$ -dimensional lattice polytope $P$ , the authors analyze the Cayley polytope $P_1 * \dots * P_k * P$ .
Hodge-Deligne Polynomials of Complete Intersections: The proof relies on a closed formula for the Hodge-Deligne polynomial of generic affine complete intersections in the algebraic torus. This formula is expressed in terms of bivariate $h^*$ -polynomials (mixed $h^*$ -polynomials) of the Newton polytopes involved. This approach generalizes the work of Danilov and Khovanskii [DK87] and utilizes the framework of bivariate Ehrhart theory.
Mixed Volume Calculation: The authors compute the Hodge vector of the resulting Cayley polytope by relating it to the mixed volume $V = MV(P_1, \dots, P_k)$ and the projection of the final polytope $P$ along the subspace $U$ .
Gale Duality and Lawrence Twists: The construction is interpreted through Gale duality. By extending a Gale transform with centrally-symmetric pairs of vectors, the authors define "Lawrence twists." This provides a geometric interpretation of the algebraic construction in the primal space.

Key Contributions and Results

Main Theorem (Theorem 1.2 & 3.1): The paper establishes that for lattice polytopes $P_1, \dots, P_k \subset U$ and $P_{k+1} \subset \mathbb{R}^d$ , the Hodge vector of the Cayley polytope $P_1 * \dots * P_{k+1}$ equals the mixed volume $V = MV(P_1, \dots, P_k)$ multiplied by the Hodge vector of the projection of $P_{k+1}$ along $U$ .
- Specifically, if the mixed volume $V=1$ , the Hodge vector is preserved exactly.
- This result holds under specific conditions regarding the dimensions and "thinness" of the intermediate Cayley sub-polytopes (Theorem 3.1).
Generalized Lawrence Twists: The authors introduce "generalized Lawrence twists" as a construction that generates infinitely many non-isomorphic, high-dimensional lattice polytopes with the same Hodge vector as a base polytope. Crucially, these constructions are distinct from free joins.
- Theorem 3.1 formalizes this, showing that under the condition $V=1$ , the Hodge vector is preserved.
- Corollary 4.4 confirms that standard Lawrence twists (defined via Gale transforms with centrally symmetric vector pairs) preserve Hodge vectors.
Applications to Thin Polytopes:
- Answering Question 5.3: The paper answers a question posed in [BKN24] regarding whether non-free-join thin polytopes must be "trivially thin" (dimension $\ge$ twice the degree). The authors demonstrate that for dimensions $d \ge 5$ , there exist infinitely many non-isomorphic, non-free-join thin polytopes that are not trivially thin. This is achieved by applying generalized Lawrence twists to specific base polytopes (e.g., lattice pyramids over polygons with interior points).
- Bk-Polytopes: The paper provides an alternative explanation for the thinness of $B_k$ -polytopes (defined in [Sel25]), showing they are instances of generalized Lawrence twists where the mixed volume vanishes, thereby forcing the local $h^*$ -polynomial to vanish.
- Nearly Thin Polytopes: The construction yields infinitely many non-free-join "nearly thin" polytopes (Hodge vector $(1)$ ) in every odd dimension $d \ge 3$ .
Explicit Formulas: The paper derives explicit formulas for the bivariate $h^*$ -polynomial of Cayley polytopes and their projections (Corollary 4.5), relating the polynomial of a Cayley sum $P * I$ to the projection of $P$ .

Significance and Claims
The paper claims to resolve the open question of whether there are constructions other than free joins that preserve Hodge vectors. By introducing generalized Lawrence twists, the authors provide a systematic method to generate infinite families of high-dimensional lattice polytopes with identical Hodge vectors.

The significance of this work lies in:

Expanding the Landscape of Thin Polytopes: It demonstrates that the class of thin polytopes is much richer than previously understood, containing non-trivial, non-free-join examples in dimensions $d \ge 5$ .
Connecting Combinatorics and Geometry: It bridges the gap between the combinatorial theory of lattice polytopes (Ehrhart theory, mixed volumes) and the geometric theory of hypergeometric motives (Hodge vectors, Hodge-Deligne polynomials).
Unifying Concepts: It unifies the concepts of Lawrence polytopes, Gale transforms, and Cayley sums under a single framework that preserves Hodge invariants.

The authors note that while their construction is powerful, it does not exhaust all possibilities. They leave open the question of whether other constructions exist for generating infinite families of polytopes with the same Hodge vector, and whether the classification of thin polytopes in dimension $d=4$ is complete. The paper also references a "total twist" construction from the first author's thesis as a sporadic, less flexible alternative that conjecturally preserves Hodge vectors.

1. The Problem: Changing the Shape, Keeping the Soul

2. The Solution: The "Lawrence Twist"

3. How It Works (The "Cayley" Recipe)

4. Why This Matters

The Bottom Line

More like this