🔢 mathematics

On de Rham--Witt Cohomology of Classifying Stacks

This paper constructs a sharp example of a proper smooth fourfold over a perfect field of characteristic $p > 0$ exhibiting asymmetric Hodge--Witt numbers in total degree 3, achieved by computing the Hodge--Witt cohomology of the classifying stack $B\alpha_p$ .

Original authors: Shizhang Li, Yuan Yang

Published 2026-04-06

📖 6 min read🧠 Deep dive

CC BY 4.0

Original authors: Shizhang Li, Yuan Yang

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: Finding a "Broken Mirror" in Math

Imagine you are looking at a beautiful, perfectly symmetrical mirror. In the world of mathematics (specifically geometry), there is a famous rule called Hodge Symmetry. It says that if you look at a shape from one angle (say, counting its "holes" in a specific way), the number you get should be exactly the same as if you looked at it from the opposite angle.

For a long time, mathematicians wondered: Does this mirror stay perfect when we change the rules of the game?

In the 1980s, a mathematician named Ekedahl introduced a new set of rules for shapes living in a world with "positive characteristic" (think of this as a universe where numbers wrap around, like hours on a clock, rather than going on forever). He defined new numbers called Hodge–Witt numbers to describe these shapes. He proved that the mirror was perfect for small shapes (like 1D lines or 2D surfaces) and for low-level counts. But he suspected that for larger, more complex shapes (4D objects), the mirror might be cracked.

The authors of this paper found the crack.

They constructed a specific, complex 4-dimensional shape where the symmetry breaks. If you count the "holes" in one direction, you get a different number than if you count them in the opposite direction. This is a big deal because it proves that the "perfect mirror" of symmetry doesn't exist in this specific mathematical universe for high-dimensional objects.

The Tools: The "Magic Calculator" and the "Ghost Shapes"

To find this broken mirror, the authors had to use some very advanced tools. Let's break them down with analogies.

1. The de Rham–Witt Complex: The "Super-Scanner"

Imagine you have a sculpture, and you want to know everything about its structure. In normal math, you use a scanner called the "de Rham complex." But in this weird "clock-number" universe, that scanner is blurry and loses information.

The authors used a de Rham–Witt complex. Think of this as a Super-Scanner that doesn't just take a picture; it takes a picture, then zooms in, then zooms in again, capturing layers of detail that the normal scanner misses. It's a way of looking at a shape that preserves all its hidden "torsion" (twisted, knotted information) that usually gets lost.

2. The Stack $B\alpha_p$ : The "Ghost Shape"

The authors didn't just build a normal sculpture; they built a Classifying Stack (specifically $B\alpha_p$ ).

Analogy: Imagine a normal shape is a solid statue. A "stack" is like a ghostly, infinite library of statues. It's a shape that remembers every possible way you could twist or rotate it, even if those twists aren't visible in a single snapshot.
The specific stack they used, $B\alpha_p$ , is a very tricky, "spooky" shape. It's too weird to be a normal smooth sculpture, so you can't just point to it and say, "Here is the counterexample."

3. The Approximation: "The Shadow Puppet Trick"

Since they couldn't use the ghostly stack directly as a counterexample (because the paper requires a "smooth proper variety," which is a fancy way of saying a nice, solid, finite sculpture), they had to be clever.

They used a technique called Approximation.

Analogy: Imagine you want to prove a shadow puppet is scary, but you can only show a real animal. You find a real animal that casts a shadow exactly like the scary puppet.
The authors took their ghostly stack ( $B\alpha_p$ ) and found a real, solid 4-dimensional sculpture (a "smooth proper fourfold") that mimics the ghost's behavior perfectly up to a certain level of detail. They proved that this real sculpture inherits the "broken symmetry" of the ghost.

The Journey: How They Did It

Here is the step-by-step adventure of the paper:

Building the Engine (Sections 2 & 3):
The authors spent the first half of the paper updating the "engine" of the math. They took old, dusty theories from the 1980s (by Illusie, Raynaud, and Ekedahl) and rebuilt them using modern, high-tech language (called $\infty$ -categories).
- Analogy: They took a vintage car engine and replaced the carburetor with a modern fuel injection system. This allowed them to run the engine much smoother and handle more complex fuel mixtures. They specifically fixed the "Star Product" (a way of combining shapes) and the "Diagonal T-structure" (a way of organizing the data) to work in this modern framework.
Scanning the Ghost (Section 5):
They turned their Super-Scanner on the ghostly stack $B\alpha_p$ .
- They calculated the "Hodge–Witt numbers" for this ghost.
- The Result: They found that for the total degree 3 (counting holes in a specific 3D way), the numbers were asymmetric.
- The Numbers: They found $h^{0,3} = 1$ and $h^{1,2} = -2$ .
- Wait, negative holes? Yes! In this abstract math, these numbers can be negative. It's like saying the shape has a "negative amount of space" in one direction compared to the other. This is the "crack" in the mirror.
The Shadow Puppet (Section 5, Corollary 5.12):
They realized that if they could find a real 4D shape that looked like this ghost, that real shape would also have these weird, negative, asymmetric numbers.
- Using a previous result by other mathematicians (Antieau, Bhatt, and Mathew), they knew such a real shape existed.
- They combined the ghost's data with the real shape's existence to prove: "Yes, there is a real 4D sculpture where the symmetry is broken."

Why Does This Matter?

It settles a 40-year-old question: Ekedahl himself said, "I see no reason why this symmetry should hold generally," but no one had ever found a counterexample. This paper provides the first concrete proof that the symmetry fails in dimension 4.
It's "Sharp": The authors proved that this is the smallest possible dimension where this can happen. You can't find this broken symmetry in 3D or lower. It's the exact tipping point where the math gets weird.
It validates the tools: By successfully using these modern "infinity-category" tools to solve an old problem, the paper shows that this new, complex mathematical language is powerful and necessary for understanding the deep structure of geometry.

Summary in One Sentence

The authors rebuilt an old mathematical engine, used it to scan a "ghostly" shape, found that the shape has a broken symmetry (asymmetric numbers), and then proved that a real, solid 4-dimensional sculpture exists that shares this broken symmetry, finally settling a decades-old mystery about the limits of geometric symmetry.

1. Problem Statement

The paper addresses a fundamental question in the arithmetic geometry of varieties in positive characteristic $p > 0$ : Do Hodge–Witt numbers satisfy Hodge symmetry?

Context: For a smooth proper variety $X$ over a perfect field $k$ of characteristic $p$ , the de Rham–Witt complex $W\Omega^\bullet_{X/k}$ provides a cohomology theory that lifts crystalline cohomology. Associated with this are Hodge–Witt numbers $h^{i,j}_W(X)$ , defined by Ekedahl. These are characteristic $p$ analogues of classical Hodge numbers $h^{i,j}(X)$ .
Known Symmetries: Ekedahl proved that Hodge–Witt numbers satisfy Serre duality ( $h^{i,j}_W = h^{N-i, N-j}_W$ ).
The Open Question: It was unknown whether Hodge symmetry ( $h^{i,j}_W = h^{j,i}_W$ $h_{W}^{i, j} = h_{W}^{j, i}$ ) holds for Hodge–Witt numbers in general.
- Ekedahl proved it holds for $i+j \le 2$ and for all smooth proper varieties of dimension $\le 3$ .
- He conjectured it might fail in higher dimensions but noted no counterexample existed in the literature.
Goal: The authors aim to construct a counterexample to Hodge symmetry for Hodge–Witt numbers, specifically determining the minimal dimension and degree where this failure occurs.

2. Methodology

The authors employ a sophisticated blend of modern homotopy theory, stack geometry, and classical de Rham–Witt theory.

A. Modernizing Ekedahl's Framework

The paper first establishes a rigorous foundation using $\infty$ -categories to reinterpret and extend the work of Illusie, Raynaud, and Ekedahl from the 1980s.

Graded Left $R$ -modules: They work within the $\infty$ -category $DGr(R)$ of left graded modules over the Cartier–Dieudonné–Raynaud ring $R$ .
Completion and Duality: They define a completion functor $\widehat{(-)}$ and a dualizing functor $D_R$ , extending Ekedahl's results to the $\infty$ -categorical setting.
Diagonal $t$ -structure: They extend Ekedahl's "diagonal $t$ -structure" on the category of coherent complexes to the unbounded setting, allowing for precise control over cohomological and grading shifts.
Star Product: They construct a symmetric monoidal structure (the "star product" $\star$ ) on $DGr(R)$ at the $\infty$ -categorical level, which corresponds to the Künneth formula for de Rham–Witt cohomology.

B. Extension to Stacks

The authors generalize the definition of de Rham–Witt cohomology from smooth schemes to smooth geometric Artin stacks.

They define $W\Omega^\bullet_{\mathcal{X}/k}$ for a stack $\mathcal{X}$ via descent (Čech nerve) from smooth covers.
They establish a Künneth formula for the de Rham–Witt cohomology of products of stacks under specific finiteness conditions (Hodge-properness).
They introduce the concept of Hodge $d$ -equivalence and de Rham–Witt $d$ -equivalence, showing that if a map between stacks is a Hodge equivalence, it induces an equivalence on de Rham–Witt cohomology in low degrees.

C. The Counterexample Construction

The core strategy involves computing the cohomology of the classifying stack of the group scheme $\alpha_p$ (the kernel of Frobenius on the additive group $\mathbb{G}_a$ ).

Target Stack: Compute $W\Omega^\bullet_{B\alpha_p}$ .
Approximation: Use the result of Antieau–Bhatt–Mathew (ABM) which states that for any $d$ , there exists a smooth projective variety $X$ and a map $X \to B\alpha_p \times B\mathbb{G}_m$ that is a Hodge $d$ -equivalence.
Transfer: Since $X$ approximates the stack in low degrees, the Hodge–Witt numbers of $X$ in total degree $\le d$ match those of the stack.
Specifics: They choose $d=4$ and compute the cohomology of $B\alpha_p$ using a spectral sequence derived from the Čech nerve of a torsor $E \to E^{(1)}$ (where $E$ is a supersingular elliptic curve).

3. Key Contributions

A. Theoretical Advances

$\infty$ -Categorical Reformulation: The paper successfully lifts the classical theory of de Rham–Witt complexes (specifically the theory of coherent $R$ -modules and dominoes) to the language of stable $\infty$ -categories. This includes defining the diagonal $t$ -structure and the star product in this modern framework.
Stack Cohomology: It provides the first systematic definition and computation of de Rham–Witt cohomology for smooth Artin stacks, establishing necessary finiteness properties and a Künneth formula.

B. Computational Breakthrough

Cohomology of $B\alpha_p$ : The authors explicitly compute the low-degree Hodge–Witt cohomology of $B\alpha_p$ . They utilize the spectral sequence associated with the diagonal $t$ -structure to analyze the terms.
Non-splitting of Extensions: A crucial technical step is proving that a specific short exact sequence in the $E_2$ -page of the spectral sequence (involving elementary dominoes $U_{-1}$ and $U_1$ ) does not split. This non-splitting is essential for generating the asymmetry.

C. The Main Result

The paper constructs a smooth proper 4-fold $X$ over a perfect field $k$ of characteristic $p > 0$ with the following Hodge–Witt numbers in total degree 3:

$h^{0,3}_W(X) = 1$
$h^{1,2}_W(X) = -2$
$h^{2,1}_W(X) = 1$
$h^{3,0}_W(X) = 0$

Significance of the Numbers:

Asymmetry: $h^{1,2}_W \neq h^{2,1}_W$ (specifically $-2 \neq 1$ ). This definitively answers the question of Hodge symmetry in the negative for Hodge–Witt numbers.
Sharpness: The example is sharp.
- Dimension: It is a 4-fold. Ekedahl had previously proven symmetry for all varieties of dimension $\le 3$ .
- Total Degree: The asymmetry appears in total degree 3 ( $i+j=3$ ). Ekedahl proved symmetry for $i+j \le 2$ .

4. Results and Significance

Refutation of Conjecture: The paper disproves the implicit hope that Hodge–Witt numbers might satisfy Hodge symmetry in higher dimensions, confirming Ekedahl's suspicion that no general principle forces this symmetry.
Optimality: The result is optimal. The authors show that one cannot find such a counterexample in dimension 3 or in total degree 2, as these cases are covered by existing theorems.
Methodological Impact: By successfully applying $\infty$ -categorical techniques to de Rham–Witt theory, the paper opens the door for applying modern homotopical algebra tools to classical arithmetic geometry problems. It bridges the gap between the "domino" theory of the 1980s and modern derived algebraic geometry.
Role of Stacks: The work demonstrates that classifying stacks of finite group schemes (like $B\alpha_p$ ) serve as powerful "test objects" for generating counterexamples in the geometry of smooth varieties, a strategy pioneered by Antieau–Bhatt–Mathew and refined here.

In summary, Li and Yang provide a definitive counterexample to Hodge symmetry for Hodge–Witt numbers in the minimal possible dimension (4) and degree (3), utilizing a deep synthesis of classical crystalline cohomology theory and modern $\infty$ -categorical methods applied to algebraic stacks.