Remarks on polynomial count varieties

This short note refutes two natural conjectures regarding polynomial count varieties by demonstrating that a smooth variety with a point count of $q^n$ is not necessarily isomorphic to affine space, and that such varieties do not necessarily have vanishing Hodge numbers for $p \neq q$.

The Gist

Imagine you are a detective trying to solve a mystery about shapes in mathematics. These shapes are called varieties. Usually, when you look at these shapes over different "worlds" (specifically, finite fields, which are like tiny, self-contained universes with a specific number of elements, $q$), counting the points on them is a chaotic mess. The number of points jumps around unpredictably.

However, some special shapes are very well-behaved. No matter which finite universe you visit, the number of points on them follows a perfect, predictable polynomial formula. If you plug the size of the universe ($q$) into the formula, you get the exact count. Mathematicians call these "Polynomial Count Varieties."

For a long time, mathematicians wondered if these well-behaved shapes had to be "simple" shapes, like a flat sheet of paper (affine space) or a perfect sphere. They asked two big questions:

1. The Shape Question: If a shape is smooth and its point-count follows a simple formula (like $q^n$), does that mean the shape is just a flat sheet of paper?
2. The Symmetry Question: If a shape is well-behaved in terms of counting points, does that mean its internal "Hodge structure" (a complex way of describing the shape's holes and symmetries) is perfectly balanced? Specifically, do the "off-diagonal" symmetries vanish?

The authors of this paper, Fernando Rodriguez Villegas and Nicholas Katz, say: "Nope."

Here is the breakdown of their findings using simple analogies:

1. The "Fake Flat" Shapes (Answering Question #1)

Imagine you have a piece of clay. If you squish it into a perfect cube, it's easy to count the atoms on its surface. But what if you have a piece of clay that looks like a cube from the outside (it has the same number of atoms in every tiny universe you check), but inside, it's twisted into a weird knot?

The authors found shapes that are smooth and have a perfect point-counting formula, yet they are not simple flat sheets.

• The Analogy: Think of the Russell Threefold. It's a 3D shape defined by a specific equation. If you count its points in any finite world, the number is exactly $q^3$ (just like a perfect 3D cube).
• The Twist: Even though the math says it looks like a cube, if you look at its actual geometry, it's not a cube. It's a twisted, knotted shape that is "diffeomorphic" to a 6D sphere but not isomorphic to a 3D plane.
• The Lesson: Just because a shape counts like a simple object doesn't mean it is that object. The "fingerprint" (the point count) can be faked by a complex shape.

2. The "Unbalanced" Shapes (Answering Question #2)

Now, imagine a musical instrument. A perfectly balanced instrument produces a pure tone where the left and right speakers are identical. Mathematicians hoped that "Polynomial Count" shapes were like these perfect instruments: their internal symmetries (Hodge numbers) would be perfectly balanced, meaning $p$ would always equal $q$.

The authors proved this is false.

• The Analogy: They built a shape by taking two different pieces and gluing them together (or rather, placing them side-by-side).
  • Piece A: A curve with a hole in it (like a donut with the center removed).
  • Piece B: A 3D space with that same curve removed.
• The Result: When you count the points on this combined shape, it's incredibly simple: it has exactly $q^2$ points. It's a "Polynomial Count" shape.
• The Twist: However, if you look at its internal symmetries (the Hodge numbers), they are messy. There are "off-diagonal" symmetries that shouldn't exist if the shape were perfectly balanced.
• The Lesson: A shape can have a very simple "counting rule" but a very complicated, unbalanced internal structure. The simplicity of the count does not force the internal geometry to be simple.

The Secret Weapon: Newton Polytopes

How did they find these tricky shapes? They used a tool called a Newton Polytope.

• The Analogy: Imagine you have a recipe (a polynomial equation) with ingredients like $x^2y$, $z^3$, etc. If you plot the "exponents" of these ingredients on a graph, they form a geometric shape (a polytope).
• The authors showed that if this geometric shape is a specific type of triangle (a simplex), then the number of solutions to the equation in any finite world follows a predictable pattern, regardless of how twisted the actual shape is.
• They used this to construct "counter-examples" that look simple on the outside (the count) but are wild on the inside.

The Big Takeaway

This paper is a reminder that in mathematics, appearance can be deceiving.

• You cannot judge a shape's complexity just by how many points it has in finite worlds.
• You cannot assume that a "nice" counting formula implies a "nice" geometric shape.

The universe of shapes is much more diverse and surprising than the "simple" examples we usually study. Just because a shape sings a simple song (the polynomial count) doesn't mean it isn't hiding a complex, twisted story inside.

Technical Summary

Based on the provided text, here is a detailed technical summary of the paper "Remarks on polynomial count varieties" by Fernando Rodriguez Villegas and Nicholas M. Katz.

1. Problem Statement and Context

The paper addresses two fundamental questions regarding polynomial count varieties. A variety $X$ over a field is called "polynomial count" if the number of its points over finite fields $\mathbb{F}_q$ is given by a polynomial $C(q)$ (i.e., $\#X(\mathbb{F}_q) = C(q)$).

The authors investigate whether the property of being polynomial count imposes strong geometric or Hodge-theoretic constraints on the variety. Specifically, they address:

1. Geometric Isomorphism: If a smooth variety $X/\mathbb{C}$ is polynomial count with $C(t) = t^n$ (matching the point count of affine space $\mathbb{A}^n$), is $X$ necessarily isomorphic to affine space $\mathbb{A}^n$?
2. Hodge Structure: If $X/\mathbb{C}$ is polynomial count, does its mixed Hodge structure satisfy the "diagonal" condition $h^{p,q} = 0$ unless $p=q$? (This condition holds for varieties with a cellular decomposition, such as toric varieties or Grassmannians).

The authors aim to determine if these intuitive conjectures hold true or if counterexamples exist.

2. Methodology

The authors employ a combination of combinatorial geometry, finite field arithmetic, and algebraic topology (specifically cohomology with compact support and excision sequences).

• Newton Polytopes and Combinatorics:

  • They utilize the Newton polytope $\Delta$ of a defining polynomial $H$.
  • They define combinatorial constants $f_{r,n}$ based on the dimensions of faces of the Newton polytope obtained by setting variables to zero.
  • They introduce a "placeholder polynomial" $F_\Delta(x, y)$ to encode these constants.
  • They establish a criterion for equivalence of lattice polytopes under invertible affine transformations over $\mathbb{Z}$.

• Point Counting Formulas:

  • Using Möbius inversion and properties of the torus $T = (\mathbb{F}_q^\times)^N$, they derive explicit formulas for the number of solutions to equations of the form $H(x_1, \dots, x_N) = 0$.
  • They relate the point count of the full variety $X$ to the point count of its intersection with the torus $X'$ and the faces of the Newton polytope.

• Cohomological Analysis:

  • To address the Hodge number question, they construct specific schemes using disjoint unions and excision sequences in compactly supported cohomology ($H^i_c$).
  • They analyze the mixed Hodge structure of these constructed varieties by relating them to known varieties (like elliptic curves) whose Hodge numbers are well-understood.

3. Key Contributions and Results

A. Counterexample to Geometric Isomorphism (Question 1)

The authors answer No to the question of whether a smooth polynomial count variety with $C(t)=t^n$ must be isomorphic to $\mathbb{A}^n$.

• Theorem 2.1: They provide a general construction for varieties defined by polynomials $H$ whose Newton polytope is equivalent to the standard simplex $\sigma_N$. They prove that such varieties are polynomial count outside a specific integer $D$ (the product of non-zero coefficients).
• Explicit Formulas: They derive exact formulas for $\#X(\mathbb{F}_q)$ and $\#X'(\mathbb{F}_q)$ involving the combinatorial constants $f_{r,n}$ and a specific polynomial $c_n(q)$.
• The Russell Threefold (Example 2.3):
  • They analyze the variety $X: x^2y + x + z^2 + t^3 = 0$ in $\mathbb{A}^4$.
  • They show its Newton polytope is equivalent to $\sigma_4$, implying $\#X(\mathbb{F}_q) = q^3$.
  • Result: While $X$ has the same point count as $\mathbb{A}^3$ (and is diffeomorphic to $\mathbb{R}^6$), it is not isomorphic to $\mathbb{A}^3$. This provides a definitive negative answer to Question 1.
• Generalization (Example 2.4): They generalize this to a family of smooth varieties $X: x^2y + x + \sum x_i^{n_i} = 0$ (with coprime $n_i$) which have point counts $q^{r+1}$ but are not isomorphic to affine space.

B. Counterexample to Diagonal Hodge Structure (Question 2)

The authors answer No to the question of whether polynomial count varieties must have vanishing Hodge numbers $h^{p,q}$ for $p \neq q$.

• Construction (Section 3):
  • They construct a variety $X$ as the disjoint union of an affine elliptic curve $Y = E \setminus \{0\}$ and a specific affine variety $Z$ defined by $z(y^2 - f(x)) = 1$.
  • Using excision sequences, they relate the cohomology of $Y$ and $Z$ to the cohomology of the elliptic curve $E$.
• Hodge Number Calculation:
  • $Y$ contributes non-zero $h^{1,0}$ and $h^{0,1}$ in weight 1.
  • $Z$ contributes non-zero $h^{1,0}$ and $h^{0,1}$ in weight 2.
  • The disjoint union $X$ (or a closed embedding $W$ in $\mathbb{A}^5$) retains these mixed Hodge components.
• Result: The resulting variety is polynomial count (with point count $q^2$ or $q^5 - q^2$ depending on the embedding) but possesses non-zero off-diagonal Hodge numbers ($h^{0,1} \neq 0$ and $h^{1,0} \neq 0$). This disproves the conjecture that polynomial count implies a diagonal Hodge structure.

4. Significance

• Refutation of Conjectures: The paper definitively shows that the arithmetic property of being "polynomial count" is insufficient to determine the geometric isomorphism class of a variety (specifically, it does not imply the variety is affine space) nor does it force the Hodge structure to be diagonal (pure of type $(p,p)$).
• New Class of Examples: The authors provide a robust method for generating counterexamples using Newton polytopes equivalent to simplices and specific cohomological constructions.
• Clarification of "Polynomial Count": The work highlights the distinction between the arithmetic behavior of a variety over finite fields and its complex geometric or topological structure. It demonstrates that varieties can share the same point-counting polynomial as standard spaces (like $\mathbb{A}^n$) while being geometrically distinct (e.g., the Russell threefold).
• Methodological Insight: The paper bridges combinatorial geometry (Newton polytopes) with arithmetic geometry (point counts) and Hodge theory, offering a template for analyzing the relationship between these fields.

In summary, Villegas and Katz demonstrate that the class of polynomial count varieties is much broader and more complex than previously suspected, containing varieties that are arithmetically "simple" (polynomial point counts) but geometrically and topologically "rich" (non-trivial topology and non-diagonal Hodge structures).