🔢 mathematics

Volume polynomials

Based on lectures delivered at the 2025 Summer Research Institute in Algebraic Geometry, these notes survey the realization problems, fundamental inequalities, and applications to algebraic matroids of volume polynomials, a distinguished class of log-concave polynomials.

Original authors: June Huh

Published 2026-06-05

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: June Huh

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 trying to build a 3D object, but you only have a set of blueprints showing the shadows the object casts on different walls. Your goal is to figure out: Can I actually build a real, solid object that casts exactly these specific shadows?

This is the core puzzle June Huh is exploring in this paper. He is studying a special class of mathematical recipes called Volume Polynomials. Think of these recipes as instructions that tell you the size (volume) of a shape when you mix different ingredients together.

Here is a breakdown of the paper's main ideas using simple analogies:

1. The Shadow Problem (Realization)

Imagine you have a mysterious 4-dimensional object (like a hyper-cube). You shine a light on it from different angles, and you measure the area of the shadow it casts on six different flat walls.

The Question: If I give you six numbers representing those shadow areas, can you always build a real, solid object that matches them?
The Answer: Not always! Huh explains that these numbers have to follow a very specific "triangle rule."
- The Analogy: Imagine you have three sticks with lengths equal to the square roots of your shadow numbers multiplied together. If you can't arrange those three sticks to form a triangle (because one is too long), then no such 4D object exists. The numbers are "impossible."
- If the sticks can form a triangle, then a real object exists. If the triangle is "perfect" (not squashed flat), the object can be smooth and round. If the triangle is "squashed" (a flat line), the object must have sharp corners.

2. The "Lorentzian" Recipe Book

Mathematicians have discovered a special category of recipes called Lorentzian Polynomials. Think of these as a "Master List" of all the mathematical patterns that look like they could describe volumes. They have a special property: they are "log-concave," which is a fancy way of saying they are nicely shaped and don't have weird bumps or holes in their graph.

The Big Discovery: Huh and his colleagues found that every real volume polynomial (a recipe for a real physical shape) is on this Master List.
The Catch: Being on the Master List isn't enough to guarantee you have a real shape. Some recipes on the list are "fake"—they look mathematically perfect, but you can't build a physical object to match them.
- Example: In 2D, any recipe on the list works. But in higher dimensions (like 4D), there are recipes that pass the math test but fail the "buildability" test.

3. The Algebraic vs. The Physical

The paper compares two worlds:

The Physical World (Convex Geometry): Trying to build real, solid shapes (like balls, cubes, or pyramids).
The Algebraic World (Projective Geometry): Trying to build shapes out of pure equations and numbers (like curves and surfaces in abstract space).

Huh shows that the Algebraic World is actually more flexible.

The Analogy: Imagine the Physical World is a strict building code. You can only build houses with certain materials. The Algebraic World is like a video game where you can build houses out of anything, even things that don't exist in real life.
Sometimes, a "shadow" pattern is impossible to build in the Physical World (it violates the triangle rule), but it is perfectly fine in the Algebraic World.
However, the Algebraic World has its own strict rules. For example, some patterns work in a world with "characteristic 2" (a specific type of math universe) but fail in a world with "characteristic 3."

4. The "Shadow" of the Recipe (Covolume Polynomials)

The paper also looks at the "reverse" problem. If you have a recipe for a shape, you can create a "dual" recipe (called a Covolume Polynomial) that acts like a shadow of the original recipe.

The Analogy: If the Volume Polynomial is the blueprint for a house, the Covolume Polynomial is the blueprint for the foundation or the negative space around the house.
Huh proves that if you have a valid Volume Polynomial, its dual is also valid. This helps mathematicians check if a recipe is real by looking at its shadow.

5. The "Matroid" Connection

Finally, the paper connects these shapes to Matroids.

The Analogy: Think of a Matroid as a rulebook for a card game. The rules tell you which combinations of cards (or points) are allowed to form a "winning hand" (a basis).
Huh shows that the "allowed combinations" in these volume recipes are exactly the same as the "winning hands" in these card games.
The Big Open Question: If you have a winning hand in the Algebraic card game, is its "dual hand" (the cards you didn't pick) also a winning hand? We know this is true for some card games, but for the most complex ones, we still don't know.

Summary

June Huh is mapping the boundaries between what is mathematically possible and what is physically buildable.

He found a "Triangle Rule" that determines if a set of shadow sizes can form a real 4D object.
He showed that while all real shapes follow a specific "Lorentzian" pattern, not every pattern in that list corresponds to a real shape.
He proved that the rules for building shapes out of pure equations (Algebraic Geometry) are slightly different from the rules for building physical shapes (Convex Geometry), but they are deeply connected.

The paper is essentially a guidebook for mathematicians to know which "volume recipes" are real and which are just mathematical illusions.

Technical Summary: Volume Polynomials

1. Problem Statement and Context
This paper surveys the theory of volume polynomials, a distinguished class of log-concave polynomials arising in convex geometry, algebraic geometry, and combinatorics. The central problem addressed is the realization problem: determining which homogeneous polynomials with nonnegative coefficients can be realized as the volume polynomials of specific geometric or combinatorial objects.

In convex geometry, Minkowski's theorem establishes that the volume of a Minkowski sum of convex bodies varies polynomially. The coefficients of this polynomial are mixed volumes. The paper investigates which such polynomials arise from actual convex bodies in $\mathbb{R}^d$ . In algebraic geometry, volume polynomials are defined via the intersection numbers of semiample divisors on projective varieties. The paper seeks to characterize the set of polynomials realizable in these contexts and understand their relationship to the broader class of Lorentzian polynomials (introduced by Brändén and Huh), which satisfy strong log-concavity properties.

2. Methodology and Framework
The paper employs a synthesis of convex geometry, algebraic geometry, and matroid theory. Key methodological components include:

Lorentzian Polynomials: The paper utilizes the theory of Lorentzian polynomials as a necessary condition for realizability. A polynomial is Lorentzian if its coefficients are nonnegative and its partial derivatives satisfy specific Hessian conditions (at most one positive eigenvalue).
M-convexity and Generalized Permutohedra: The support of a volume polynomial is shown to be an M-convex set, a concept from discrete convex analysis related to the bases of matroids and polymatroids.
Realization Spaces: The paper analyzes the "realization space" of a vector of projection areas or homology classes, characterizing when such spaces are nonempty.
Linear Operators and Duality: The paper introduces covolume polynomials as duals to volume polynomials within the framework of differential operators acting on polynomial rings. It studies linear operators (such as deletion, contraction, and polarization) that preserve the sets of realizable volume and covolume polynomials.
Inequalities: The analysis relies heavily on fundamental inequalities, including the Alexandrov–Fenchel inequality for mixed volumes and the Khovanskii–Teissier inequality for intersection numbers on projective varieties.

3. Key Results and Contributions

3.1. Realization Problems for Projection Volumes
The paper provides a complete characterization for the realizability of projection areas of convex bodies in $\mathbb{R}^4$ .

Theorem 1.1 & 1.2: A vector of six nonnegative numbers $(p_{ij})$ represents the projection areas of a convex body in $\mathbb{R}^4$ if and only if the numbers $\sqrt{p_{12}p_{34}}, \sqrt{p_{13}p_{24}}, \sqrt{p_{14}p_{23}}$ satisfy the triangle inequalities. This condition is equivalent to the validity of the Plücker relation for the Grassmannian $Gr(2,4)$ over the triangular hyperfield, or equivalently, that a specific symmetric matrix has at most one positive eigenvalue.
Higher Dimensions: The paper formulates Conjectures 1.6 and 1.7 for $\mathbb{R}^5$ , proposing that realizability is characterized by the eigenvalue properties of the projection volume matrix (at most one positive eigenvalue for projections to codimension-2 subspaces, and a similar condition for 2D projections). A footnote notes that while the forward implication holds, the converse for $\mathbb{R}^5$ remains an open problem with recent counterexamples to naive formulations.

3.2. Algebraic Realization and Homology Classes
The paper extends these questions to the projective setting, considering homology classes of irreducible surfaces in $(\mathbb{P}^1)^4$ .

Theorem 1.9: A vector of nonnegative rational numbers is realizable as a homology class (up to scaling) if and only if it satisfies the same triangle inequalities as the convex case. However, the paper notes additional obstructions in the algebraic case due to integrality and the specific geometry of the defining equations (e.g., the class $(1,1,1,1,1,3)$ is not realizable despite satisfying the triangle inequalities).
Theorem 1.10: For $(\mathbb{P}^1)^5$ , the realizability of homology classes is linked to the condition that the associated symmetric matrix has at most one positive eigenvalue.

3.3. Characterization of Volume Polynomials

Convex vs. Lorentzian: The paper establishes that the set of volume polynomials of convex bodies is a subset of Lorentzian polynomials. For $d=2$ (quadratic forms) and $n=2$ (bivariate forms), the sets coincide (Shephard's theorem). However, for $d \ge 3$ and $n \ge 3$ , the inclusion is strict.
Counterexamples: The paper presents a specific cubic polynomial (Example 3.2) that is Lorentzian but not a volume polynomial of convex bodies, derived from the reverse Khovanskii–Teissier inequality. This provides a counterexample to Gurvits' conjecture regarding strongly log-concave polynomials.
Algebraic Geometry: The set of realizable volume polynomials over a field $k$ ( $V_n^d(\mathbb{Q}, k)$ ) is shown to depend only on the characteristic of $k$ . The paper proves that any minor of a realizable volume polynomial is also realizable.

3.4. Covolume Polynomials and Duality

Definition: Covolume polynomials are defined as polynomials $g(\partial)$ such that $g(\partial) \circ x^{[\mu]}$ is a realizable volume polynomial.
Characterization (Theorem 4.7): A polynomial $g$ is a realizable covolume polynomial if and only if the operator $g(\partial)$ maps any realizable volume polynomial to another realizable volume polynomial.
Matroid Connection: The paper connects these polynomials to algebraic matroids. Proposition 5.5 states that a polymatroid is algebraic over $k$ if and only if it is the support of a realizable volume polynomial over $k$ . This implies that the class of algebraic polymatroids is closed under taking minors.

4. Significance and Claims
The paper positions volume polynomials as a bridge between the analytic properties of log-concavity (Lorentzian polynomials) and the geometric realizability of mixed volumes and intersection numbers.

Unification: It unifies realization problems in convex geometry (projection areas) and algebraic geometry (homology classes) under a common framework of Lorentzian polynomials and M-convex supports.
Structural Insights: By identifying volume polynomials as a subset of Lorentzian polynomials, the paper clarifies the limitations of log-concavity as a sufficient condition for geometric realization.
New Inequalities: The duality between volume and covolume polynomials allows for the derivation of new inequalities for mixed volumes and intersection numbers via linear operators (e.g., Schubert polynomials acting on volume polynomials).
Open Problems: The paper highlights several significant open questions, including:
- The full characterization of realizability in $\mathbb{R}^5$ (Conjecture 1.6).
- Whether the set of volume polynomials depends on the field $k$ for $d, n \ge 3$ (Conjecture 3.8).
- Whether the dual of an algebraic matroid is always algebraic (Question 5.7).
- The relationship between analytic volume polynomials (on Kähler manifolds) and algebraic volume polynomials.

The author notes that the results are based on lectures given at the 2025 Summer Research Institute, serving as a survey of current knowledge and open directions in the field, rather than a collection of entirely new experimental data. The work relies on and synthesizes results from previous literature (e.g., [HHM+], [BH20], [GHM+]) to provide a coherent theoretical landscape.