Filtrations and asymptotic geometry of non-Archimedean norms on section rings

This paper establishes a jointly $d_1$-contracting action of filtrations on the space of graded norms for a polarized variety over a non-Archimedean field, thereby constructing non-Archimedean geodesic rays and infinite-dimensional flats while proving the convergence of relative limit measures along these rays.

The Gist

Imagine you are standing in a vast, infinite landscape made of mathematical shapes. In this world, there are two main types of travelers:

1. The Norms: Think of these as different ways to measure the "size" or "distance" of objects in a space. In our math world, these are like different rulers or maps. Some rulers are standard; others are warped, stretched, or twisted.
2. The Filtrations: These are like "skeletons" or "blueprints" of the space. They represent the underlying structure, stripped of all the specific measurements.

This paper, written by Rémi Reboulet, is about understanding the geography of this landscape. Specifically, it explores how the "skeletons" (filtrations) can act as a compass to guide the "rulers" (norms) across the infinite horizon.

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

1. The Two Worlds: Finite vs. Infinite

The paper starts in a small, manageable room (a finite-dimensional vector space). Here, mathematicians have known for a long time that if you take a ruler and stretch it according to a specific blueprint, you create a straight line called a geodesic. It's like pulling a rubber band taut; it finds the shortest path.

But the author wants to move to a giant, infinite city (a graded algebra or "section ring" associated with a geometric shape called a variety). In this infinite city, the rules of geometry get messy. The question is: Does the same logic apply? Can we still use the blueprints to draw straight lines in this infinite city?

2. The "Gerardin Action": The Magic Compass

The core discovery is a new way to combine a Blueprint (a filtration) with a Ruler (a norm).

• The Analogy: Imagine you have a map of a city (the norm) and a set of traffic rules (the filtration). The author defines a "magic operation" where you apply the traffic rules to the map.
• The Result: When you do this, you don't just get a new map; you get a direction. If you keep applying this operation over and over (scaling it up), you trace out a perfect, straight path called a geodesic ray.
• Why it matters: This proves that the "blueprints" (filtrations) are actually the horizon of the landscape. They are the "points at infinity." If you walk far enough in a straight line, you eventually arrive at a blueprint.

3. The "Radial Isometry": The Speedometer

One of the most beautiful results in the paper is about distance.

• The Analogy: Imagine you are driving a car (a geodesic ray) away from a city center. As you drive further and further out, the landscape starts to look less like the messy city and more like the flat, open horizon.
• The Discovery: The author shows that if you measure the distance between two cars driving on parallel paths, and you divide that distance by how far they have traveled, the number you get stabilizes.
• The Metaphor: It's like looking at two parallel train tracks. From the station, they look far apart. As you zoom out to the horizon, the distance between them seems to shrink relative to the total distance traveled. The paper proves that this "relative distance" at the horizon is exactly the same as the distance between the two blueprints (filtrations) that the trains are following.
• In Math Speak: This is called a Radial Isometry. It means the geometry of the "horizon" perfectly mirrors the geometry of the "directions" you were traveling.

4. The Infinite "Flat Cones"

The paper concludes with a surprising finding about the shape of this infinite landscape.

• The Analogy: In a normal room, if you pick a point, you can draw a straight line in any direction. In this infinite mathematical landscape, the author proves that from any point, you can find not just one, but infinitely many flat, cone-shaped regions where the geometry is perfectly flat (like a sheet of paper).
• The Significance: This suggests the landscape has a hidden, rigid structure (like a crystal or a building made of flat rooms) that we didn't fully understand before. It connects this abstract math to a concept called Bruhat-Tits buildings, which are like geometric puzzles used to understand symmetry in nature.

Summary: What did we learn?

1. The Horizon is Real: The "points at infinity" of this mathematical space are exactly the "filtrations" (the blueprints).
2. Straight Lines Exist: You can use these blueprints to draw perfect straight lines (geodesics) through the infinite space.
3. The Map Matches the Territory: The distance between two blueprints is exactly the same as the "speed" at which two paths diverge as they head toward the horizon.
4. Hidden Flatness: Despite the complexity, the space is filled with flat, infinite cones, giving it a structured, almost architectural feel.

In a nutshell: This paper takes a complex, abstract mathematical space and shows us that it has a clear, navigable structure. It proves that if you know the "blueprints" (filtrations), you can predict exactly how the "rulers" (norms) behave as they stretch out toward infinity. It turns a chaotic, infinite fog into a clear, navigable map.

Technical Summary

1. Problem Statement and Context

The paper addresses the asymptotic metric geometry of the space of non-Archimedean norms on graded algebras, specifically the section ring $R(X, L)$ of a polarized variety $(X, L)$ over a non-Archimedean field $k$.

• Background: The space of norms on a finite-dimensional vector space $V$, denoted $N(V)$, is known as the Goldman–Iwahori space. It is a negatively curved metric space (specifically CAT(0) for the $d_2$ metric) and is closely related to Bruhat–Tits buildings. G´erardin previously described an action of the space of filtrations (identified with trivially-valued norms $N_0(V)$) on $N(V)$, generating geodesic rays.
• The Gap: While the finite-dimensional case is well-understood, the infinite-dimensional setting of graded norms on section rings (which correspond to non-Archimedean plurisubharmonic metrics on $L$) lacks a precise description of its "boundary at infinity."
• Objective: The author aims to generalize G´erardin's construction to the graded setting, establishing a rigorous link between the space of graded norms $N(R)$ and the space of trivially-valued graded norms $N_0(R)$ (which correspond to filtrations of the section ring). The goal is to characterize the asymptotic behavior of geodesic rays and the radial geometry of the space.

2. Methodology

The paper employs a "quantization" strategy, moving from finite-dimensional linear algebra to infinite-dimensional graded algebras.

• Finite-Dimensional Foundation (Section 1):

  • The author first rigorously analyzes the action of $N_0(V)$ on $N(V)$.
  • Key Technical Tools: The paper establishes the continuity of the G´erardin action with respect to the $d_\infty$ (Goldman–Iwahori) and $d_1$ metrics.
  • Approximation: A crucial technique involves approximating non-diagonalizable norms by diagonalizable ones (using the density of $N_{\text{diag}}(V)$) and utilizing ground field extensions to spherical completions to handle diagonalization issues.
  • Determinants and Volumes: The author uses the relationship between the relative volume of norms and the determinant of the induced norm on $\det V$ to prove Lipschitz continuity and convexity properties.

• Graded Extension (Section 2):

  • The results from Section 1 are lifted to the graded setting $N(R)$ via quantization. This involves taking limits of normalized metrics on the graded components $R_m = H^0(X, mL)$ as $m \to \infty$.
  • Spectral Measures: The paper relies on the Chen–Maclean equidistribution theorem, which states that the relative spectral measures of norms on $R_m$ converge weakly to a limit measure $\sigma(\|\cdot\|_\bullet, \|\cdot\|'_\bullet)$.
  • Metric Definitions: The $d_p$ metrics on $N(R)$ are defined via these limit measures: $d_p(\|\cdot\|_\bullet, \|\cdot\|'_\bullet)^p = \int |x|^p d\sigma$.

3. Key Contributions and Results

The paper presents three main theorems (A, B, and C) that characterize the geometry of $N(R)$.

Theorem A: The Geodesic Action

There exists a well-defined action $N_0(R) \times N(R) \to N(R)$, denoted $(\|\cdot\|_0, \|\cdot\|) \mapsto \|\cdot\|_0 \cdot \|\cdot\|$.

• Properties:
  1. The action is jointly 1-Lipschitz with respect to the $d_1$ metric.
  2. The scaling action on $N_0(R)$ (multiplying the filtration by $t$) transforms into geodesic rays in $N(R)$ via this action.
• Significance: This generalizes G´erardin's finite-dimensional result, providing a mechanism to generate geodesic rays in the infinite-dimensional space of metrics.

Theorem B: Radial Isometry and Limit Measures

This theorem describes the asymptotic behavior of geodesic rays. Let $t \mapsto \|\cdot\|_{\bullet, t}$ and $t \mapsto \|\cdot\|'_{\bullet, t}$ be two geodesic rays in $N(R)$ with trivially-valued limits $\|\cdot\|_{0, \bullet}$ and $\|\cdot\|'_{0, \bullet}$ in $N_0(R)$.

• Convergence of Measures: The rescaled relative spectral measures converge weakly:
  $$\lim_{t \to \infty} t^{-1}_* \sigma(\|\cdot\|_{\bullet, t}, \|\cdot\|'_{\bullet, t}) = \sigma(\|\cdot\|_{0, \bullet}, \|\cdot\|'_{0, \bullet})$$
• Radial Isometry: Consequently, the distance between rays grows linearly with $t$, and the asymptotic slope equals the distance between their limits:
  $$\lim_{t \to \infty} t^{-1} d_p(\|\cdot\|_{\bullet, t}, \|\cdot\|'_{\bullet, t}) = d_p(\|\cdot\|_{0, \bullet}, \|\cdot\|'_{0, \bullet})$$
• Significance: This proves that $(N_0(R), d_p)$ is the boundary at infinity of $(N(R), d_p)$. It establishes a precise "radial isometry" analogous to results in Archimedean Kähler geometry (Finski) and non-Archimedean potential theory.

Theorem C: Infinite-Dimensional Flat Cones

Any element in $N(R)$ lies at the apex of infinitely many infinite-dimensional $d_p$-metric flat cones.

• Construction: For any $\|\cdot\|_0 \in N_0(R)$, there is an isometric embedding from the space of convex, decreasing functions on the support of the spectral measure of $\|\cdot\|_0$ (equipped with an $L^p$ norm) into $N(R)$.
• Significance: This reveals a "building-like" structure in the infinite-dimensional setting. While finite-dimensional norms belong to apartments isometric to $\mathbb{R}^n$, graded norms belong to infinite-dimensional flats. This suggests a complex, subtle structure beyond simple Bruhat–Tits buildings.

4. Significance and Implications

• Non-Archimedean Pluripotential Theory: The results provide a metric framework for studying non-Archimedean plurisubharmonic metrics. The space $N(R)$ is identified with these metrics, and the geodesic rays correspond to paths in the space of metrics.
• K-Stability: Filtrations of section rings are central to the study of K-stability (via the work of Boucksom, Jonsson, etc.). This paper rigorously connects the metric geometry of these filtrations to the geometry of the space of metrics, offering new tools for analyzing stability conditions.
• Generalization of Classical Results: The paper successfully extends classical results from finite-dimensional $p$-adic analysis (Goldman–Iwahori, G´erardin) and Hermitian geometry (Darvas, Phong–Sturm) to the infinite-dimensional, non-Archimedean graded setting.
• Open Questions: The author notes that while "flat cones" exist, the full "building" property (that any two points share a common apartment) does not hold in the naive sense, suggesting the geometry is more subtle than standard Euclidean buildings. The paper also highlights open problems regarding $L^p$-CAT(0) conditions and the existence of geodesic lines in general settings.

5. Conclusion

Rémi Reboulet's paper provides a foundational metric study of the space of non-Archimedean norms on section rings. By establishing the action of filtrations, characterizing the boundary at infinity via limit measures, and proving the existence of infinite-dimensional flat structures, the work bridges the gap between algebraic geometry (K-stability), non-Archimedean analysis, and metric geometry. It sets the stage for further investigations into the global geometry of non-Archimedean plurisubharmonic metrics.