🔢 mathematics

A short note on Du Bois singularities

This paper investigates the behavior of Du Bois singularities under base change and fiber products in characteristic zero, demonstrating their descent under field extensions and their preservation in products involving varieties with rational singularities.

Original authors: Pat Lank

Published 2026-06-03

📖 4 min read🧠 Deep dive

CC BY 4.0

Original authors: Pat Lank

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 a detective trying to understand the "health" of a geometric shape. In the world of mathematics, these shapes are called varieties. Sometimes, these shapes are perfectly smooth, like a polished marble sphere. Other times, they have kinks, sharp corners, or tears in them. Mathematicians call these imperfections singularities.

For smooth shapes, mathematicians have a well-known toolkit to study them, called the de Rham complex. Think of this toolkit as a high-resolution camera that takes perfect pictures of the shape's geometry. But what happens when the shape is broken? The camera gets blurry.

To fix this, mathematicians invented a special, upgraded camera called the Du Bois complex. It's designed to take clear pictures even of broken, singular shapes. A shape is said to have "Du Bois singularities" if this special camera can still see the shape clearly enough to say, "Yes, this is a valid object, even with its cracks."

This short paper by Pat Lank is like a field guide for two specific scenarios where we want to know if these "cracked" shapes stay healthy when we do two things to them: changing the background and mixing them with other shapes.

1. Changing the Background (Base Change)

Imagine you have a sculpture (your variety) sitting in a gallery in Paris (a field called $k$ ). You decide to ship it to a different gallery in Tokyo (a new field called $L$ ).

The Question: If the sculpture looks "healthy" (has Du Bois singularities) in Paris, does it stay healthy in Tokyo? Conversely, if you see it looks healthy in Tokyo, was it always healthy in Paris?
The Finding: The paper proves that the answer is yes. The "health" of the sculpture doesn't depend on which gallery it's in.
- If it's healthy in the original gallery, it will be healthy in any new gallery you move it to.
- Surprisingly, if you move it to just one new gallery and it looks healthy there, you can be 100% sure it was healthy in the original gallery all along. You don't need to check every possible gallery; one check is enough to confirm the original state.

2. Mixing Shapes (Fiber Products)

Now, imagine you have two sculptures.

Sculpture A is a "Rational Singularity." Think of this as a shape that is very well-behaved. It might have a crack, but it's a "nice" crack that can be easily fixed with a standard repair kit.
Sculpture B has "Du Bois singularities." This shape is a bit more rugged. It has cracks that are harder to fix, but the special Du Bois camera can still handle them.
The Question: If you glue these two sculptures together to make a giant, combined sculpture (a product), will the new giant sculpture still be "healthy" enough for the Du Bois camera?
The Finding: Yes. If you mix a "well-behaved" shape (Rational) with a "rugged but manageable" shape (Du Bois), the result is still a "rugged but manageable" shape. The good behavior of the first shape helps protect the second one, ensuring the whole combination remains healthy.

The Big Unanswered Question

The paper also asks a tricky follow-up question: What happens if you mix two rugged shapes (both having Du Bois singularities)?

The Answer: We don't know for sure yet. It's like asking, "If you glue two very cracked pots together, will the new pot hold water?"
The Exception: The author does know the answer if the shapes are simple, one-dimensional lines (curves). If you glue two cracked lines together, the result is still healthy. But for more complex, multi-dimensional shapes, the mystery remains unsolved.

Summary

In simple terms, this paper tells us that the "Du Bois health check" is very reliable:

It works no matter where you move the shape (changing fields).
It works when you mix a "good" shape with a "Du Bois" shape.
It works for simple lines, but we still need to figure out if it works when mixing two complex "Du Bois" shapes together.

The author uses advanced mathematical tools (like "derived categories," which are like super-complex filing systems for these shapes) to prove these points, ensuring that our understanding of these geometric cracks is solid and consistent.

Technical Summary: A Short Note on Du Bois Singularities

Problem Statement
This note investigates the behavior of Du Bois singularities under fundamental geometric constructions, specifically base change (field extension) and fiber products. While Du Bois singularities are well-defined via the Deligne–Du Bois complex and possess favorable Hodge-theoretic properties, their stability under these operations is not fully understood. The paper addresses two primary questions:

Do Du Bois singularities descend from a field extension? That is, if a variety $Y$ over a field $k$ becomes Du Bois after extending scalars to a field $L$ , is $Y$ itself Du Bois?
Are Du Bois singularities stable under products? Specifically, if one factor has rational singularities and the other has Du Bois singularities, does the product retain Du Bois singularities? Furthermore, does the product of two varieties with Du Bois singularities necessarily have Du Bois singularities?

Methodology
The author works within the framework of derived categories of coherent sheaves ( $D^b_{\text{coh}}(X)$ ) and utilizes the theory of the Du Bois complex $\Omega^\bullet_X$ . The proofs rely on recent developments in the field, specifically:

Murayama's Results: Utilizing [Mur24] regarding the behavior of the Du Bois complex under smooth morphisms and flat base change.
New Characterization: Applying a recent characterization of Du Bois singularities from [LV25], which relates the property to the splitting of the natural morphism $O_X \to R p_* O_E$ in the derived category.
Derived Category Techniques: The arguments employ flat base change, the projection formula, and the analysis of direct summands within triangulated categories. The proofs often involve constructing commutative diagrams of fibered squares and analyzing the cohomology of the resulting complexes.

Key Results

Descent of Du Bois Singularities (Proposition 1.1):
For an embeddable variety $Y$ over a field $k$ of characteristic zero, the following are equivalent:
1. $Y$ has Du Bois singularities.
2. $Y_L = \text{Spec}(L) \times_k Y$ has Du Bois singularities for some field extension $L/k$ .
3. $Y_L$ has Du Bois singularities for every field extension $L/k$ .
  The proof establishes that if the property holds after a single extension, the original variety must possess it. This relies on showing that $O_Y$ is a direct summand of $\Omega^0_Y$ in $D_{\text{qc}}(Y)$ , a condition guaranteed by the behavior of the complex under base change and the new characterization in [LV25].
Stability under Products with Rational Singularities (Proposition 1.2):
Let $Y_1$ and $Y_2$ be embeddable varieties over a field $k$ of characteristic zero. If $Y_1$ has rational singularities and $Y_2$ has Du Bois singularities, then the product $Y_1 \times_k Y_2$ has Du Bois singularities.
The proof constructs a modification $f_1: \tilde{Y}_1 \to Y_1$ where $\tilde{Y}_1$ is smooth. By analyzing the fiber product $\tilde{Y}_1 \times_k Y_2$ , the author shows it has Du Bois singularities (using the smoothness of the projection). A chain of isomorphisms involving the derived pushforward $R(f'_1)_*$ then demonstrates that $Y_1 \times_k Y_2$ inherits the Du Bois property.
The Case of Curves (Remark 2.1):
The paper provides a positive answer to the general product question (Question 1.3) in the specific case where both factors are curves (Krull dimension one). If $X$ and $Y$ are curves with Du Bois singularities, their product $X \times_k Y$ also has Du Bois singularities. This is established by showing that the product of seminormal varieties is seminormal and analyzing the cohomological concentration of the Du Bois complex via a distinguished triangle involving the singular locus and the normalization.

Significance and Claims
The author positions this work as a modest extension of existing stability properties for Du Bois singularities.

The descent result (Proposition 1.1) clarifies that Du Bois singularities are detectable after a single field extension for embeddable varieties, a non-trivial result requiring recent tools.
The product result (Proposition 1.2) is noted as not appearing explicitly in the literature, even in characteristic zero, filling a gap regarding the interaction between rational and Du Bois singularities.
The paper explicitly acknowledges an open problem: it remains unclear whether the hypothesis in Proposition 1.2 can be weakened to allow both factors to have only Du Bois singularities (rather than one having rational singularities). The positive result for curves suggests this might be true in low dimensions, but the general case remains an open question (Question 1.3).

The paper does not propose new applications or future experimental directions but rather consolidates the theoretical understanding of these singularities under standard geometric operations.

1. Changing the Background (Base Change)

2. Mixing Shapes (Fiber Products)

The Big Unanswered Question

Summary

More like this