🔢 mathematics

Divisors on coherent schemes and homogeneous spaces

This paper develops a valuative divisor theory on locally coherent schemes to establish a non-Noetherian Ramanujam–Samuel theorem and construct ample invertible sheaves on group homogeneous spaces over Prüfer bases, thereby generalizing Raynaud's quasi-projectivity results without requiring smoothness hypotheses.

Original authors: Ning Guo

Published 2026-03-24

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Ning Guo

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 build a perfect, stable structure (a mathematical object called a "scheme") on a foundation that is a bit wobbly. In the old days of mathematics, architects only built on solid, predictable ground (called "Noetherian" rings). But modern mathematics, especially in fields like cryptography and number theory, demands building on much stranger, more fluid foundations (called "non-Noetherian" or "Prüfer" rings).

The problem? The old blueprints (standard mathematical rules) break down on this wobbly ground. You can't just copy-paste a design from a solid foundation to a fluid one; the walls might collapse, or the doors might disappear.

This paper, written by Ning Guo, is a new construction manual for building on this fluid ground. Here is the breakdown of what the author does, using simple analogies:

1. The Broken Compass: Fixing "Divisors"

The Problem: In math, a "divisor" is like a fence or a boundary line you draw on your land to mark where a field ends. On solid ground, there's a perfect rulebook (the Weil–Cartier correspondence) that says: "If you draw a fence, you automatically get a specific type of wall (an invertible sheaf)."
The Breakdown: On the wobbly, non-Noetherian ground, this rulebook is useless. If you try to draw a fence, the wall might not form, or it might turn into a ghost.
The Solution: Guo invents a new kind of compass called "Valuative Divisors." Instead of counting fence posts with whole numbers (1, 2, 3), he uses a continuous scale (like a ruler that can measure infinitely small fractions). This new compass works perfectly on the wobbly ground, allowing him to draw boundaries that actually result in solid walls. He proves that on this new ground, every valid boundary corresponds to a valid wall, just like in the old days.

2. The Shape-Shifting Terrain: "Morphisms of (N)-Type"

The Problem: When you build a structure that spans different types of ground (some solid, some fluid), the shape of your building can change unpredictably as you move from one section to another. A smooth hallway might suddenly turn into a jagged cave.
The Solution: Guo defines a special category of building projects called "(N)-type". Think of this as a "Safety Certification."

If a project is (N)-type, it guarantees that while the ground might change, the essential shape of the building remains stable and predictable.
He shows that many famous mathematical structures (like curves and surfaces used in cryptography) naturally fit into this "Safety Certified" category. This allows him to ignore the scary, jagged parts and focus on the stable core.

3. The Magic Mirror: Group Actions and "Picard-Admissibility"

The Problem: Imagine your building has a magical property: if you rotate the whole structure, the walls should still line up perfectly. In math, this is a "group action." Usually, checking if your walls line up after a rotation is incredibly hard, especially on wobbly ground.
The Solution: Guo introduces a concept called "Picard-admissibility."

Think of this as a "Magic Mirror." Instead of checking the whole building, you only need to look at the reflection in the mirror (the "generic fibre" or the top layer of the building).
If the reflection looks perfect, the whole building is perfect.
This is a huge shortcut. It means he doesn't have to check every single brick on the fluid ground; he just checks the top layer, and the math guarantees the rest follows suit.

4. Building the Ultimate Structure: Ample Sheaves

The Goal: The ultimate goal of this paper is to prove that these structures are "Quasi-Projective."

Analogy: In simple terms, this means the structure is "nice enough" to be built inside a standard, well-behaved city (projective space). If a structure is not quasi-projective, it's like a house built in a swamp that can't be connected to the city grid.
The Breakthrough: Raynaud (a famous mathematician from the 1970s) proved this for solid ground. Guo proves it for the wobbly ground.
How? He uses the "fences" (valuative divisors) he invented in Step 1. He looks at the boundaries where the group action (the rotation) stops. He builds a "fence" right along these boundaries. This fence acts as a "magnet" that pulls the whole structure into the city grid, proving it is quasi-projective.

5. The "Rigid Extension": Saving the Blueprints

The Problem: Imagine you have a perfect blueprint for a house on a solid island. You want to build the same house on the fluid mainland. Usually, the blueprint fails because the ground shifts.
The Solution: Guo proves a "Rigid Extension" theorem.

If you have a "perfect" (ample) blueprint for the top layer of your house (the generic point), you can extend it to the whole fluid foundation without changing the design.
It's like having a blueprint that is "self-healing." No matter how the ground shifts, the blueprint automatically adjusts to fit the new terrain while keeping the house standing tall and beautiful.

Summary

Ning Guo's paper is a masterclass in mathematical adaptation.

He fixed the tools: Created new ways to measure boundaries on unstable ground.
He found the safe zones: Identified which structures naturally stay stable even when the ground changes.
He found a shortcut: Proved that checking the top layer is enough to guarantee the whole structure works.
He built the bridge: Showed that structures on this weird, fluid ground can still be connected to the "standard city" of mathematics.

This work is crucial because it allows modern mathematicians to use powerful geometric tools in areas like perfectoid spaces and arithmetic geometry, which are essential for understanding deep secrets of numbers and cryptography, but which previously were too "wobbly" for standard math to handle.

Based on the paper "Divisors on Coherent Schemes and Homogeneous Spaces" by Ning Guo, here is a detailed technical summary covering the problem, methodology, key contributions, results, and significance.

1. Problem Statement

The paper addresses fundamental obstacles in extending classical algebraic geometry results regarding homogeneous spaces and invertible sheaves (line bundles) to non-Noetherian bases.

The Context: Classical results, such as Raynaud's quasi-projectivity theorem for homogeneous spaces over normal bases, rely heavily on Noetherian assumptions. These assumptions facilitate the use of limit arguments and the classical Weil–Cartier correspondence (linking Weil divisors to Cartier divisors).
The Obstacles:
1. Failure of Limit Arguments: In non-Noetherian settings (e.g., perfectoid rings, Prüfer rings, valuation rings of higher rank), a limit of positive cycles of codimension 1 does not necessarily correspond to a Cartier divisor on the limit scheme.
2. Breakdown of Weil–Cartier Correspondence: The classical bijection between Weil divisors and Cartier divisors fails in non-Noetherian contexts because maximal ideals in local rings are not necessarily principal, and the structure sheaf may not be coherent.
3. Fibre Degeneration: In arithmetic geometry, special fibres often degenerate (e.g., becoming non-reduced or non-normal), rendering standard smoothness hypotheses insufficient for controlling the descent of polarizations.

The central question is: How can one construct ample invertible sheaves and extend generic polarizations on group homogeneous spaces over non-Noetherian, coherent bases without relying on standard limit arguments or smoothness hypotheses?

2. Methodology

The author develops a new framework based on locally coherent schemes and valuative divisor theory.

Locally Coherent Schemes: The paper restricts attention to schemes where the structure sheaf is coherent as a module over itself. This class includes Noetherian schemes but also covers important non-Noetherian cases like Prüfer schemes and rings of algebraic integers.
Valuative Divisors: Instead of using discrete integer multiplicities (as in classical Weil divisors), the author introduces valuative Weil divisors. These assign coefficients from the totally ordered value groups ( $\Gamma_x$ ) of the local valuation rings at codimension-1 points.
Morphisms of (N)-type: To handle fibre degenerations without assuming global smoothness, the paper introduces a class of morphisms called (N)-type. A flat, locally finitely presented morphism $X \to S$ $X \to S$ is of (N)-type if:
- Generic fibres are geometrically normal.
- Fibres over points $s$ where the weak dimension of $\mathcal{O}_{S,s}$ is 1 are geometrically reduced.
- This class includes universal curves over moduli stacks, Schubert schemes, and semistable degenerations.
Picard-Admissibility: The author utilizes the concept of Picard-admissibility for invertible sheaves under group actions. This property ensures that the group action induces a homomorphism into the relative Picard functor, effectively generalizing the "Theorem of the Square."

3. Key Contributions and Results

A. Valuative Divisor Theory (Section 2)

Theorem A (Correspondence): For a locally coherent, topologically locally Noetherian, normal scheme $X$ , there is an exact, order-preserving bijection between the monoid of effective valuative divisors and the monoid of rank-one reflexive ideal sheaves.
Generalized Weil–Cartier: If $X$ is regular, the classical isomorphism $\text{Cl}(X) \simeq \text{Div}(X)$ is recovered.
Non-Noetherian Ramanujam–Samuel Theorem: The paper establishes that for smooth schemes over normal bases, relative effective Cartier divisors correspond precisely to flat one-codimensional cycles. Crucially, over Prüferian bases, effective Cartier divisors on generic fibres uniquely extend to the whole space via schematic closures.

B. Descent and Extension of Sheaves (Sections 3 & 7)

Rigid Extension of Polarizations (Theorem C): The paper proves that if an invertible sheaf $L_\eta$ $L_{η}$ on the generic fibre of a homogeneous space over an integral Prüfer scheme is generated by global sections (resp. ample), it extends to an $S$ $S$ -semiample (resp. $S$ $S$ -ample) sheaf on the whole space.
- This bypasses the failure of standard spreading-out arguments.
- It shows that any invertible extension of an ample generic sheaf is automatically ample.
Descent of Trivial Sheaves: Using the properties of (N)-type schemes, the author proves that generically trivial invertible sheaves descend to the base under specific conditions, establishing injectivity of $H^1_{\text{ét}}(X, \mathbb{G}_m) \to H^1_{\text{ét}}(X_\eta, \mathbb{G}_m)$ .

C. Ampleness and Homogeneous Spaces (Sections 5 & 6)

Theorems of the Cube and Square: The author proves these theorems for (N)-type schemes without smoothness hypotheses. This is achieved by reducing Picard-admissibility to the generic fibre using descent properties.
Effective Ampleness via Orbit Boundaries (Theorem B):
- Let $X$ be an $S$ -smooth scheme with a $G$ -action and a quasi-affine open $U$ .
- The one-codimensional irreducible components of the orbit boundary $G \cdot U \setminus U$ span an effective divisor $D$ .
- The line bundle $\mathcal{L} = \mathcal{O}_X(D)$ is $S$ -ample.
- This provides an explicit construction of ample sheaves directly from orbit geometry, generalizing Raynaud's results to non-Noetherian bases.

D. Quasi-Projectivity (Section 6)

Corollary 6.14: For a normal scheme $S$ and an $S$ -smooth group scheme $G$ with connected fibres, any $G$ -homogeneous space $X$ (locally of finite type) is locally $S$ -quasi-projective.
This recovers Raynaud's classical theorem but removes the Noetherian hypothesis, proving it for bases like $\text{Spec}(\mathbb{Z})$ or perfectoid rings.

4. Significance

Foundational Shift: The paper successfully transitions the theory of divisors and homogeneous spaces from the Noetherian world to the coherent/non-Noetherian world. This is essential for modern arithmetic geometry, particularly in the study of perfectoid spaces and $p$ -adic Hodge theory where base rings are often non-Noetherian valuation rings.
Resolution of Obstructions: By replacing limit arguments with valuative divisors and schematic closures, the author resolves the "divisorial limit obstruction" and the "fibre degeneration obstruction" that previously blocked generalizations of Raynaud's work.
Explicit Constructions: The construction of ample sheaves from orbit boundaries (Theorem B) provides a concrete, geometric method to generate polarizations, which is vital for moduli problems and the study of integral models.
Robustness: The introduction of (N)-type morphisms creates a robust category of schemes where theorems of the cube and square hold even when fibres degenerate, broadening the scope of applicability for group scheme actions in arithmetic contexts.

In summary, Ning Guo's work provides a rigorous, non-Noetherian foundation for the theory of divisors on homogeneous spaces, enabling the extension of classical ampleness and quasi-projectivity results to the most general arithmetic bases currently used in advanced number theory and algebraic geometry.