🔢 mathematics

Representability of codimension three cycles

This paper introduces and develops the concept of representability for codimension three cycles on a fourfold by characterizing them through zero cycles modulo rational equivalence on surfaces.

Original authors: Kalyan Banerjee

Published 2026-04-23

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Kalyan Banerjee

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

The Big Picture: Sorting the Messy Room

Imagine you have a massive, complex 4-dimensional room (a mathematical object called a fourfold). Inside this room, there are all sorts of "shapes" or "cycles" floating around. Some are tiny dots (0-dimensional), some are lines, some are sheets, and some are 3D volumes.

Mathematicians want to understand these shapes. Specifically, they want to know: "Can we describe all the complicated shapes in this room using only simple, familiar building blocks?"

In this paper, the author focuses on 3D shapes (codimension 3) inside a 4D room. The goal is to see if these complex 3D shapes can be "represented" (or built) using simpler pieces coming from curves (1D) and surfaces (2D).

The Core Concept: "Representability"

Think of Representability like a translation service.

The Problem: You have a huge library of complex, abstract books (the 3D shapes in your 4D room).
The Goal: Can you translate every single one of these complex books into a language made only of short stories written on curves (like a string of beads) or surfaces (like a sheet of paper)?

If the answer is YES, we say the group of shapes is "representable." It means the complex stuff isn't actually that complex; it's just a big pile of simple stuff glued together.
If the answer is NO, it means there is something "wild" or "mysterious" in the room that cannot be broken down into those simple pieces.

The History: The "Mumford" and "Bloch" Rules

The paper stands on the shoulders of giants:

Mumford's Discovery: He found that if a 2D surface (like a fancy piece of fabric) has a certain "twist" (geometric genus > 0), you cannot describe all its shapes using just curves. It's too wild.
Bloch's Conjecture: He guessed the opposite: If a surface is "simple" (genus = 0), then you can describe all its shapes using curves.

This paper asks: What happens in 4D? specifically for 3D shapes.

The New Idea: "Weak Representability up to Dimension Two"

Banerjee introduces a slightly looser definition called "Weak Representability up to Dimension Two."

Imagine you are trying to build a giant sculpture (the 3D shape).

Old Rule: You must build it entirely from a single type of Lego brick (a curve).
New Rule (Banerjee): You are allowed to use Curves (1D) AND Surfaces (2D) as your building blocks.

The paper asks: Can we build every 3D shape in our 4D room using a mix of curves and surfaces?

The Main Discovery: The "Fano" Shortcut

The author proves a powerful theorem:
If your 4D room has a special property (it's a "Fano" variety, which is a very nice, well-behaved type of space), and if the 3D shapes inside it are generated by "lines" (linear subspaces), then YES, they are representable.

The Analogy:
Imagine your 4D room is a giant warehouse filled with 3D boxes.

Banerjee finds a special "blueprint" hidden in the room called the Fano Variety. Think of this as a map of all the straight lines that can fit inside your warehouse.
He shows that if you take a slice of this map (a "hyperplane section"), you get a Surface (like a sheet of paper).
The Magic: He proves that you can take this single Surface, draw lines on it, and "project" them into your 4D room to recreate every single 3D shape you need.

In short: You don't need a million different tools. You just need one specific Surface (derived from the lines in the room) and one specific Curve, and you can build the whole world of 3D shapes.

Why Does This Matter? (The Rationality Test)

Why do mathematicians care if we can build shapes out of curves and surfaces?

It helps solve the "Rationality Problem."

Rationality is like asking: "Is this 4D room just a distorted version of a perfect, empty 4D cube ( $P^4$ )?"
If a room is "rational," it means it's essentially simple, just bent out of shape.
If a room is "irrational," it has a hidden, complex structure that can't be smoothed out.

Banerjee shows that if a 4D room is "rational" (like a distorted cube), its 3D shapes must be representable by curves and surfaces.

The Catch: If you find a 4D room where the 3D shapes cannot be built from curves and surfaces, you have proven that the room is not rational. It has a hidden complexity that makes it fundamentally different from a simple cube.

The "Finite Dimensionality" Metaphor

The paper also talks about "Finite Dimensionality."
Imagine you are trying to fill a bucket (the group of shapes) with water using cups of different sizes.

If you need an infinite number of different cup sizes to fill the bucket, the bucket is "infinite dimensional" (chaotic).
If you can fill the bucket using only a specific set of cup sizes (or a specific pattern of cups), it is "finite dimensional."

Banerjee proves that for these specific 4D rooms, the "bucket" of 3D shapes is finite dimensional. It behaves in a predictable, manageable way, which confirms it can be built from the simpler curves and surfaces.

Summary in One Sentence

Kalyan Banerjee proves that for a specific type of 4-dimensional space, all the complex 3D shapes inside it can be perfectly reconstructed using a combination of simple curves and surfaces, which helps mathematicians determine whether that space is "simple" (rational) or "complex" (irrational).

1. Problem Statement and Context

The paper addresses a fundamental question in the theory of Chow groups: the representability of algebraic cycles.

Background: For a smooth projective variety $X$ over $\mathbb{C}$ , let $CH^p(X)$ be the group of codimension $p$ cycles modulo rational equivalence, and $CH^p(X)_{alg}$ (denoted as $A^p(X)$ in the paper) be the subgroup of algebraically trivial cycles.
Classical Representability: A group $A^p(X)$ $A^{p} (X)$ is "representable" if there exists a smooth projective curve $C$ $C$ and a correspondence $\Gamma$ $Γ$ such that the induced map $J(C) \to A^p(X)$ $J (C) \to A^{p} (X)$ is surjective.
- Mumford's Theorem: For surfaces with geometric genus $>0$ , $A^2(S)$ is not representable.
- Bloch's Conjecture: For surfaces with geometric genus $=0$ , $A^2(S)$ is representable.
The Gap: While representability for zero-cycles (codimension $n$ ) and codimension 1 cycles is well-studied, the behavior of codimension 3 cycles on a fourfold ( $X$ is a 4-fold, $p=3$ ) is less understood.
The Specific Problem: The author seeks to define and analyze a notion of "weak representability up to dimension two" for $A^3(X)$ . This concept asks whether $A^3(X)$ can be generated by correspondences from curves ( $A^1(C)$ ) and surfaces ( $A^2(S)$ ), rather than just curves. This is crucial for distinguishing rational varieties from unirational ones, as standard representability fails to detect non-rationality in unirational fourfolds.

2. Methodology

The paper employs a combination of algebraic geometry techniques, specifically focusing on Chow varieties, Fano varieties of linear subspaces, and Roitman-Voisin arguments.

A. Definitions and Framework

Weak Representability up to Dimension Two: $A^p(X)$ is weakly representable up to dimension two if there exist finitely many curves $C_i$ and surfaces $S_j$ with correspondences $\Gamma_i, \Gamma'_j$ such that the sum of induced maps:
$\bigoplus_i \Gamma_i + \bigoplus_j \Gamma'_j : \bigoplus_i A^1(C_i) \oplus \bigoplus_j A^2(S_j) \to A^p(X)$
is surjective.
Finite Dimensionality via Chow Varieties: The author introduces a second notion of representability based on the surjectivity of maps from products of Chow varieties.
- Let $C^p_d(X)$ be the Chow variety parameterizing codimension $p$ cycles of degree $d$ .
- Consider the map $\theta^p_d: C^p_d(X) \times C^p_d(X) \to A^p(X)$ defined by $(Z_1, Z_2) \mapsto [Z_1 - Z_2]$ .
- $A^p(X)$ is "finite dimensional up to dimension two" if there exists an integer $g > 1$ such that the union of images $\bigcup_d \text{im}(\theta^p_{gd})$ covers $A^p(X)$ .

B. Key Assumptions

The main theorems rely on two primary assumptions:

Generation by Linear Subspaces: The group $A^p(X)$ is generated by linear subspaces. Specifically, the natural map from the Chow group of the Fano variety of linear subspaces $F(X)$ to $A^p(X)$ is surjective.
Equivalence of Equivalences: Algebraic and homological equivalence coincide for codimension $p$ cycles on $X$ .

C. Analytical Tools

Baire Category Argument: Used to handle countable unions of Zariski closed subsets in the product of Chow varieties. This ensures that if a map is surjective over a countable union, there exists a specific component with maximal dimension properties.
Dimensional Analysis of Fibers: The author analyzes the dimension of fibers of the map $\theta^p_d$ . By comparing dimensions of symmetric powers of the Fano variety and the images, the author establishes bounds on the "size" of the Chow group.
Intersection Theory on Fano Varieties: The proof utilizes the geometry of the Fano variety $F(X)$ , specifically its hyperplane sections. The author proves that if a group is finite dimensional up to dimension two, the generating cycles can be restricted to a surface (a complete intersection of hyperplane sections) within $F(X)$ .

3. Key Contributions and Results

Theorem 1.2 (Main Structural Result)

Under the assumptions that $A^p(X)$ is generated by linear subspaces and is "finite dimensional up to dimension two" (in the sense of the countable union of Chow variety products), the author proves:

There exists a smooth projective surface $S$ (specifically a hyperplane section of the Fano variety $F(X)$ ) and a correspondence $\Gamma$ on $S \times X$ .
The induced map $\Gamma^*: A^2(S) \to A^p(X)$ is surjective.
Significance: This reduces the representability of high-codimension cycles on a fourfold to the study of zero-cycles on a specific surface derived from the geometry of $X$ .

Theorem 2.1 (Birational Invariance)

The property of $A^3(X)$ being "representable up to dimension two" is a birational invariant for smooth projective fourfolds.

Proof Logic: Using the blow-up formula, $A^3(\tilde{X}) \cong A^3(X) \oplus A^2(S) \oplus A^1(C)$ . Since $A^2(S)$ and $A^1(C)$ are inherently representable up to dimension two, the property holds for $\tilde{X}$ if and only if it holds for $X$ .
Implication: This allows the concept to be used as a potential obstruction to rationality. If a unirational fourfold $X$ has $A^3(X)$ that is not representable up to dimension two, $X$ cannot be rational.

Theorem 2.6 & 2.7 (Equivalence of Notions)

Theorem 2.6: If linear subspaces generate $A^p(X)$ , then $A^p(X)$ is "finite dimensional up to dimension two" if and only if it has "finite dimension" (a specific numerical bound on the image of the Chow variety maps).
Theorem 2.7: For a smooth projective fourfold where algebraic and homological equivalence coincide, $A^3(X)$ is representable up to dimension 2 if and only if it is finite dimensional up to dimension 2.

Application to Fano Fourfolds

The paper applies these results to smooth projective Fano fourfolds where $A^3(X)$ is generated by differences of rational curves. Citing results from [TZ], the author concludes that for such varieties, $A^3(X)$ is representable up to dimension two.

4. Significance and Impact

Extension of Mumford-Roitman Theory: The paper successfully extends the classical theory of representability (originally focused on zero-cycles and curves) to codimension 3 cycles on fourfolds.
New Obstruction to Rationality: By establishing that "representability up to dimension two" is a birational invariant, the paper provides a new tool to test the rationality of unirational fourfolds. Since unirational varieties are known to have $A^3(X)$ representable up to dimension two, finding a unirational fourfold where this property fails would prove non-rationality (though the paper suggests this property often holds for standard examples).
Geometric Reduction: The result that $A^3(X)$ can be dominated by $A^2(S)$ (where $S$ is a surface inside the Fano variety) is a powerful geometric reduction. It shifts the problem from the 4-fold $X$ to the geometry of a surface $S$ , which is more tractable.
Methodological Rigor: The paper rigorously connects the abstract notion of "weak representability" with the concrete geometry of Chow varieties and symmetric powers, using sophisticated dimension counting and intersection theory arguments (Lemma 2.8) to prove the existence of the dominating surface.

In summary, Banerjee's work provides a robust framework for understanding the structure of codimension 3 cycles on fourfolds, linking them to the geometry of Fano varieties and offering a refined birational invariant for the study of rationality.