🔢 mathematics

Rational singularities and $q$ -birational morphism

This paper generalizes the concept of rational singularities to any reflexive sheaf of rank 1, establishes its connection to existing definitions, and introduces the dual notion of $(B_{q+1})$ to prove new theorems regarding $q$ -birational morphisms.

Original authors: Donghyeon Kim

Published 2026-06-03

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Donghyeon Kim

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 understand the structural integrity of a building. Sometimes, a building looks perfect from the outside, but if you look closely at the foundation or the corners, you might find cracks, weak spots, or "singularities" where the structure breaks down. In algebraic geometry, mathematicians study shapes (called varieties) that can have these mathematical "cracks."

This paper by Donghyeon Kim is like a new manual for inspecting these cracks, but with a twist: instead of just looking at the building itself, the author introduces a way to inspect specific "layers" or "sheets" wrapped around the building (called reflexive sheaves of rank 1).

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

1. The Problem: What is a "Rational Singularity"?

In the old days, mathematicians had a strict rule for deciding if a building was "healthy" (having rational singularities). They would send a team of inspectors (a "resolution") to smooth out all the cracks. If, after smoothing, the team could perfectly reconstruct the original building's blueprint without losing any information, the building was considered healthy.

However, this old rule was a bit rigid. It only worked well for the building itself or very specific types of cracks.

2. The New Tool: The "Double-Back" Mirror

Kim introduces a new, more flexible way to check for health. He uses a concept called the Double Dual.

The Analogy: Imagine you have a piece of fabric (the sheaf). If you look at it in a mirror, you get a reflection. If you look at that reflection in a second mirror, you get a "double reflection."
The Insight: For most healthy fabrics, the original fabric and the double reflection are identical. Kim uses this "double reflection" to define what it means for a specific layer of the building to be healthy, even if the building itself is a bit weird. This allows him to check the health of specific "sheets" (divisors) wrapped around the shape, not just the shape itself.

3. The "Weakly Rational" vs. "Rational" Distinction

The author creates two levels of health:

Weakly Rational: The building passes a basic test. The inspectors can smooth it out, and the "double reflection" matches up perfectly.
Rational: The building passes the basic test AND it is structurally sound in a deeper way (mathematically, it is "Cohen-Macaulay," which is like saying the building is solid all the way through, not just on the surface).

4. Measuring the "Distance" from Perfection

Sometimes a building isn't perfect, but it's close. How do we measure how close it is?

The Analogy: Imagine a "Non-Rational Locus" is a map showing exactly where the cracks are.
The New Metric: Kim introduces a new ruler called $(RS_q)$ and $(KV_q)$ .
- Think of $(RS_q)$ as a safety zone. If you stay within a certain distance ( $q$ ) from the center of the building, the cracks don't matter; the structure is safe.
- If the cracks are far away (high codimension), the building is considered "mostly rational."

5. The "Dual" Concept: $(B_{q+1})$

Mathematics often works in pairs (like a key and a lock). The paper introduces a new concept called $(B_{q+1})$ , which is the "dual" (the opposite side of the coin) to a well-known concept called $(S_{q+1})$ .

The Analogy: If $(S_{q+1})$ is checking if the building is solid from the inside out, $(B_{q+1})$ is checking if the building is solid from the outside in.
The author proves that if a building has a certain level of safety zone $(RS_q)$ , then checking the "outside-in" solidity $(B_{q+1})$ is the same as checking the "inside-out" solidity $(S_{q+1})$ . This unifies two different ways of measuring structural integrity.

6. The "q-Birational" Morphism: The Elevator Test

The paper also looks at how two buildings relate to each other when one is a "smoothed out" version of the other.

The Analogy: Imagine an elevator that only stops at certain floors. A $q$ -birational morphism is like a special elevator where the "stops" (the places where the building changes) are very high up (far from the ground floor).
The Result: If the elevator stops are high enough (codimension $\ge q+1$ ), then the "vibrations" (mathematical cohomology) don't travel down to the lower floors. This means the lower floors of the smooth building and the rough building are identical. This proves that if you smooth out a building in a specific way, the "deep" structure remains unchanged.

Summary of Key Findings

Generalization: You can now check the "health" of any single-layer sheet wrapped around a shape, not just the shape itself.
Robustness: If you have a healthy building, and you take a "finite cover" (like looking at a building through a kaleidoscope that repeats the image), the original building is still healthy.
Positive Characteristic: The paper shows that even in "weird" mathematical worlds (positive characteristic fields), certain types of buildings (strongly F-regular) are guaranteed to be healthy.
The Big Theorem: If a building is "mostly smooth" (has a large safety zone) and its layers are "mostly solid," then the whole building is structurally sound (Cohen-Macaulay).

In short: Donghyeon Kim has built a more flexible, powerful toolkit for mathematicians to inspect the structural integrity of complex geometric shapes, allowing them to measure exactly how "broken" a shape is and proving that under certain conditions, smoothing out the shape preserves its deep structural secrets.

Technical Summary of "Rational Singularities and q-Birational Morphisms" by Donghyeon Kim

Problem Statement
The paper addresses the generalization of the classical notion of rational singularities, traditionally defined for normal varieties (specifically for the structure sheaf $\mathcal{O}_X$ ), to the setting of any reflexive sheaf of rank 1 on a normal variety. While standard definitions (e.g., in [Kov11] and [Kol13]) rely on strict transforms of divisors and log resolutions, the author seeks a definition that is more intrinsic to the sheaf itself, utilizing double duals rather than strict transforms. Furthermore, the paper aims to quantify the "distance" of a singularity from being rational by introducing a hierarchy of conditions ( $RS_q$ , $KV_q$ ) and a dual notion to Serre's condition ( $S_{q+1}$ ), termed ( $B_{q+1}$ ). Finally, the work investigates the behavior of these singularities under finite morphisms and $q$ -birational morphisms.

Methodology
The methodology relies heavily on derived category machinery, Grothendieck duality, and local duality.

Generalized Definition: The author defines rational singularities for a reflexive sheaf $\mathcal{F}$ of rank 1 via a natural map $\theta_{\mathcal{F}, f}$ in the derived category, relating the derived pushforward of the dual of the pullback of $\mathcal{F}$ to the derived Hom of $\mathcal{F}$ with the dualizing complex.
Double Duals: Instead of using strict transforms of divisors (as in [Kov11]), the paper utilizes the double dual operation ( $(-)^{DD}$ ) to define the relevant sheaves on a resolution. This avoids the necessity of defining "normal pairs."
Duality and Vanishing: The proofs frequently employ local duality (Theorem 2.4) and Grothendieck duality (Theorem 2.5) to translate vanishing conditions on higher direct images ( $R^i f_*$ ) into conditions on local cohomology ( $H^i_x$ ).
Spectral Sequences: The Leray spectral sequence and edge maps are used extensively to relate the cohomology of the base space to the resolution, particularly in establishing vanishing theorems for $q$ -birational morphisms.
Finite Morphisms: The stability of rational singularities under finite étale and flat morphisms is analyzed using the splitting properties of pullbacks and the behavior of reflexive sheaves under these maps.

Key Contributions and Results

Generalized Definition of Rational Singularities:
The paper introduces a definition of rational singularities for any reflexive sheaf $\mathcal{F}$ of rank 1 (Definition 1.2). A resolution $f: X' \to X$ is a "weakly rational resolution" if a specific natural map $\theta_{\mathcal{F}, f}$ is a quasi-isomorphism, which is equivalent to the vanishing of higher direct images of the double dual of the pullback ( $R^i f_* (f^* \mathcal{F})^{DD} = 0$ for $i \ge 1$ ). $\mathcal{F}$ has "rational singularities" if it is weakly rational and Cohen-Macaulay (CM).
Comparison with Existing Notions:
The author clarifies that if a pair $(X, D)$ has rational singularities in the sense of [Kov11], then $-D$ has rational singularities in the new sense. The new definition is presented as minimal regarding the vanishing requirements and CM-ness.
Existence in Specific Geometric Settings:
- Characteristic 0: Theorem 1.5 proves that any $\mathbb{Q}$ -Cartier Weil divisor $D$ on a dlt pair $(X, \Delta)$ has rational singularities. This generalizes Corollary 5.25 in [KM98] without requiring cyclic covering theory.
- Positive Characteristic: Theorem 1.6 establishes that on a strongly $F$ -regular variety, any Weil divisor $D$ (where $rD$ is Cartier and $\gcd(r, p)=1$ ) has rational singularities.
Stability under Finite Morphisms:
Theorem 1.7 states that if $p: Y \to X$ is a finite étale morphism in characteristic 0, and $p^* \mathcal{F}$ has (weakly) rational singularities, then $\mathcal{F}$ also has (weakly) rational singularities. This provides a partial answer to a question posed by Kollár in [Kol13]. Theorem 1.8 extends this to finite flat morphisms under the condition that $\mathcal{F}$ is $(KV_{\dim X})$ .
The Notion of $(B_{q+1})$ and CM-ness:
The paper introduces $(B_{q+1})$ as a dual condition to Serre's $(S_{q+1})$ , defined by the vanishing of local cohomology in specific high degrees ($dim X - q < i < dim X$). Theorem 1.14 establishes the equivalence between a sheaf being $(S_{q+1})$ , the vanishing of certain higher direct images, and its dual being $(B_{q+1})$ .
- Corollary 1.15: Under specific conditions ( $X$ is $(R_q)$ , all Weil divisors are $(KV_q)$ and $(S_{q+1})$ , and $q \ge \lceil (\dim X + 1)/2 \rceil$ ), the paper proves that any Weil divisor on $X$ is CM, implying $X$ itself is CM.
$q$ -Birational Morphisms:
The paper utilizes the notion of $q$ -birational morphisms (where the exceptional locus has codimension 1 and the center has codimension $\ge q+1$ ). Theorem 1.19 proves a vanishing theorem: if $f: X' \to X$ is $q$ -birational and $D$ is an anti- $f$ -nef Cartier divisor such that $f_* D$ is $\mathbb{Q}$ -Cartier and $(S_{q+1})$ , then $R^i f_* \mathcal{O}_{X'}(D) = 0$ for $1 \le i < q$ .

Significance and Claims
The paper claims that its primary significance lies in providing a unified and intrinsic framework for rational singularities applicable to reflexive sheaves of rank 1, rather than just the structure sheaf or reduced divisors. By utilizing double duals, the author removes the dependency on the concept of "normal pairs" found in previous literature.

The work demonstrates that the property of having rational singularities is robust under finite étale morphisms in characteristic 0, addressing an open question in the field. Furthermore, the introduction of the $(B_{q+1})$ condition offers a new tool for analyzing the Cohen-Macaulay property of divisors on normal varieties, particularly in the context of $q$ -birational morphisms. The results generalize known theorems regarding klt and strongly $F$ -regular varieties, extending their applicability to arbitrary reflexive sheaves of rank 1. The author notes that while the proofs for characteristic 0 are complete, the extension of the finite morphism stability results to positive characteristic remains an open difficulty due to the potential lack of splitting in the relevant maps.