🔢 mathematics

On the locally analytic completed cohomology of modular curves

This paper surveys the authors' research concerning the locally analytic vectors within the completed cohomology of modular curves.

Original authors: Lue Pan

Published 2026-03-31

📖 6 min read🧠 Deep dive

CC BY 4.0

Original authors: Lue Pan

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: Mapping the Unseeable

Imagine you are trying to understand a massive, complex city called Modular Land. This city is built on a strange foundation of numbers (specifically, prime numbers like $p$ ). Mathematicians have been trying to build a "map" of this city for decades to solve a famous mystery called the Fontaine-Mazur Conjecture (which is essentially asking: "Which shapes in this city are actually real, and which are just optical illusions?").

For a long time, the map was blurry. Then, a mathematician named Matthew Emerton built a new, high-tech camera called Completed Cohomology. This camera could take a picture of the entire city at once, capturing infinite details. But the picture was still too fuzzy to read the fine print.

Lue Pan's paper is about taking that blurry photo and using a special pair of glasses to make the text perfectly clear. He focuses on a specific type of "text" in the photo called Locally Analytic Vectors.

The Core Concepts (The Analogies)

1. The City and the Infinite Zoom (Modular Curves & Infinite Level)

Think of the Modular Curve as a landscape that changes depending on how close you look.

Low Level: If you look from a satellite, you see broad highways (classical shapes).
Infinite Level: If you zoom in infinitely close (using a tool invented by Peter Scholze), the landscape turns into a fractal—a perfect, self-repeating pattern that looks the same at every scale. This is the "Perfectoid Space."

Pan's work happens at this "infinite zoom" level. It's where the magic of the city's structure is revealed.

2. The Two Languages of the City (Symmetries)

The city speaks two languages simultaneously:

The Local Language ( $GL_2(\mathbb{Q}_p)$ ): This is the language of the immediate neighborhood. It's about how things wiggle and shift right here, right now.
The Global Language ( $G_{\mathbb{Q}}$ ): This is the language of the whole country. It's about how the neighborhood connects to the rest of the world.

The big problem is: How do these two languages talk to each other? Pan's paper shows that they are actually speaking the same dialect, just with different accents.

3. The "Higgs Field" (The Gravity of the City)

To understand the city, Pan introduces a concept called the Higgs Field (borrowed from physics).

Analogy: Imagine the city is covered in a thick, invisible fog. This fog has a "current" or a flow. If you drop a leaf (a mathematical object) into the fog, the current pushes it in a specific direction.
The Discovery: Pan found that this "current" (the Higgs field) acts like a GPS. It tells us exactly where every piece of the puzzle belongs. By following the flow of this fog, he can translate the complex "Local Language" into a simple geometric shape: a Line (specifically, a projective line $\mathbb{P}^1$ ).

4. The "Localization" (Flattening the Map)

This is the paper's biggest breakthrough.

The Old Way: Trying to study the city by walking every street (very hard, very messy).
Pan's Way: He realized that if you look at the "Locally Analytic" parts of the city, they can be flattened out onto a simple line (like unrolling a scroll).
The Metaphor: Imagine you have a crumpled piece of paper with a complex drawing on it. Pan found a way to iron it out perfectly flat onto a table. Once it's flat, the drawing looks like a simple set of instructions (called D-modules).
Why it matters: Instead of solving a 3D puzzle, we can now solve a 1D puzzle. The complex symmetries of the city are now just simple arrows pointing along a line.

What Did He Actually Find? (The Results)

1. The "Sen Operator" is the City's Heartbeat

Pan identified a specific rhythm in the city's data called the Sen Operator.

Analogy: Think of a heartbeat monitor. The "beat" tells you the health of the heart.
The Result: Pan showed that this heartbeat is directly linked to the "weight" of the shapes in the city. If the heartbeat matches a certain pattern, the shape is "real" (classical). If it doesn't, it's a ghost (overconvergent). This helps mathematicians distinguish between real shapes and illusions.

2. The "Fontaine Operator" is the Bridge

He also studied a second tool called the Fontaine Operator.

Analogy: Imagine a bridge connecting two islands. One island is "Real Shapes" (Classical), and the other is "Ghost Shapes" (Overconvergent).
The Result: Pan built a geometric bridge between them. He showed that the "Ghost Shapes" are just the "Real Shapes" that have been stretched or twisted by a specific force. By measuring this twist, he could prove that if a shape looks real enough (satisfies certain conditions), it must be a real shape. This solves a major part of the Fontaine-Mazur conjecture.

3. The "Cousin Complex" (The Family Tree)

He connected his findings to a concept called the Cousin Complex.

Analogy: Imagine a family tree. You have the "Classical" ancestors and the "Overconvergent" descendants. Pan showed that the connection between them isn't random; it follows a strict family rule (a differential equation).
The Result: This allows mathematicians to predict the behavior of the "descendants" just by looking at the "ancestors."

Why Should You Care? (The Impact)

Simplifying the Complex: Pan took a problem that required super-computers and years of calculation and reduced it to a geometric picture on a simple line. He turned a 4D maze into a 2D map.
Solving Old Mysteries: His method proved that certain "ghost" shapes are actually real, confirming long-held guesses by other mathematicians (like Gouvêa's conjecture).
A New Toolkit: He didn't just solve one problem; he built a new machine (the "Geometric Sen Theory") that can be used to solve similar problems in other areas of math, like studying higher-dimensional shapes (Shimura varieties).

The One-Sentence Summary

Lue Pan figured out how to flatten a complex, infinite-dimensional mathematical city onto a simple line, revealing that the hidden rules governing its shapes are actually just simple geometric flows, allowing us to finally distinguish between real mathematical objects and their illusions.

1. Problem Statement

The paper addresses the structure of the completed cohomology of modular curves, denoted $\tilde{H}^i(K_p)$ , introduced by Matthew Emerton. This space serves as a $p$ -adic surrogate for spaces of automorphic forms and carries simultaneous actions of:

The Galois group $G_{\mathbb{Q}} = \text{Gal}(\bar{\mathbb{Q}}/\mathbb{Q})$ .
The group $GL_2(\mathbb{Q}_p)$ .
The Hecke algebra.

The central challenge is to understand the interaction between the $GL_2(\mathbb{Q}_p)$ -action and the Galois action, particularly in the context of the Fontaine-Mazur conjecture and the $p$ -adic local Langlands correspondence. While Emerton established a deep connection via the $p$ -adic local Langlands correspondence, Pan aims to provide a new geometric approach that avoids relying heavily on the classification of $p$ -adic representations, instead utilizing the $p$ -adic geometry of modular curves at infinite level.

2. Methodology

Pan's approach relies on the $p$ -adic geometry of modular curves with infinite level at $p$ , introduced by Peter Scholze. The core methodology involves:

Relative/Geometric Sen Theory: Adapting Sen's method for $p$ -adic Galois representations to the setting of rigid analytic varieties. This involves defining a canonical Higgs field $\theta_V$ on the pro-étale site of a smooth rigid analytic variety.
Localization: Utilizing the Hodge-Tate period morphism $\pi_{HT}: X_{K_p} \to \mathbb{F}\ell \cong \mathbb{P}^1$ , where $X_{K_p}$ is the perfectoid modular curve of infinite level. The author studies the pushforward of the structure sheaf under this map.
Sheaf of Locally Analytic Functions ( $O_{la}$ ): Defining a sheaf $O_{la}$ on the flag variety $\mathbb{P}^1$ consisting of $GL_2(\mathbb{Q}_p)$ -locally analytic sections. The completed cohomology is shown to be isomorphic to the cohomology of this sheaf.
Differential Operators: Identifying the action of the Lie algebra $\mathfrak{gl}_2$ on $O_{la}$ with differential operators (specifically the Sen operator and the Fontaine operator) derived from the geometry of the modular curve.

3. Key Contributions and Results

A. Localization of Locally Analytic Completed Cohomology

The paper establishes that the locally analytic vectors of the completed cohomology, $\tilde{H}^i(K_p)^{la}$ , can be "localized" to the flag variety $\mathbb{P}^1$ .

Theorem 2.11: There are natural isomorphisms:
$\tilde{H}^i(K_p)^{la} \widehat{\otimes}_{\mathbb{Q}_p} \mathbb{C} \cong H^i(\mathbb{P}^1, O_{la})$
where $O_{la}$ is a sheaf on $\mathbb{P}^1$ .
Theorem 2.12 & 2.13: The sheaf $O_{la}$ behaves as a $\mathcal{D}$ -module. The action of the Cartan subalgebra $\mathfrak{h} \subset \mathfrak{gl}_2$ on the cohomology corresponds to the Sen operator. This provides a $p$ -adic Hodge-theoretic interpretation of infinitesimal characters, linking them to Hodge-Tate weights.

B. Hodge-Tate Structure and Overconvergent Modular Forms

The author analyzes the decomposition of the cohomology based on the Hodge-Tate weights (characterized by the Sen operator).

Regular Weights ( $k \neq 1$ ): For weights $k \notin \{0, 1, 2\}$ , the cohomology admits a Hodge-Tate decomposition:
$\tilde{H}^1(\mathbb{C})^{la}_{\mu_k} \cong N_{k,1} \oplus M^\dagger_{2-k}(k-1)$
Here, $M^\dagger_k$ represents overconvergent modular forms. This result characterizes overconvergent eigenforms purely via their Hodge-Tate-Sen weights, confirming a conjecture by Gouvêa.
Irregular Weight ( $k=1$ ): For weight 1, the decomposition fails due to the existence of non-Hodge-Tate representations. Instead, there is an exact sequence relating the cohomology to overconvergent forms and a connecting homomorphism $\delta$ $δ$ .
- Theorem 3.6: A weight 1 overconvergent eigenform with a Hodge-Tate Galois representation must be a classical modular form (an Artin representation). This reproves and generalizes results by Buzzard-Taylor.

C. de Rham Structure and the Fontaine Operator

The paper studies the de Rham structure of the cohomology, governed by the Fontaine operator $N$ (a nilpotent operator).

Geometric Description of $N$ : The Fontaine operator is identified geometrically as a composition of derivations along the modular curve and the flag variety.
- Theorem 4.3: The operator $N$ on cohomology is essentially $H^1(\bar{d}_1 \circ d_1)$ , where $d_1$ and $\bar{d}_1$ are derivations along the modular curve and $\mathbb{P}^1$ respectively.
Classicality Results: This geometric description allows for a proof of classicality for Galois representations that are de Rham, without invoking the $p$ $p$ -adic local Langlands correspondence.
- Theorem 4.7: If a subrepresentation $\rho \subset \tilde{H}^1(K_p)$ is de Rham with weights $0, k$ , it arises from a cuspidal eigenform of weight $k+1$ .

D. Construction of $p$ -adic Local Langlands Representations

The paper provides a geometric construction of the locally analytic representations of $GL_2(\mathbb{Q}_p)$ appearing in the $p$ -adic local Langlands correspondence.

Theorem 4.10 & 4.12: For a classical eigenform $f$ $f$ , the associated locally analytic representation $\Pi_{la}$ $Π_{l a}$ is constructed as the cokernel of a map $i_\pi$ $i_{π}$ induced by the filtration of the de Rham period sheaf.
- In the supercuspidal case, this involves the Drinfeld upper half-plane.
- In the principal series case, this involves the log crystalline cohomology of Igusa curves.
This unifies the description of these representations over different strata of the flag variety $\mathbb{P}^1$ .

4. Significance

New Geometric Perspective: The work shifts the paradigm from representation-theoretic classification (via $p$ -adic local Langlands) to geometric analysis of the Hodge-Tate period morphism. It demonstrates that deep arithmetic properties (like classicality) can be deduced from the geometry of the infinite-level modular curve.
Resolution of Conjectures: It confirms Gouvêa's conjecture regarding the Hodge-Tate weights of overconvergent eigenforms and provides a robust framework for the Fontaine-Mazur conjecture in the context of modular curves.
Generalizability: The methods (Geometric Sen theory, localization to flag varieties) are designed to be generalizable. The author notes that these results have been extended to arbitrary Shimura varieties (by Rodriguez Camargo) and Hilbert modular surfaces (by Jiang), suggesting a unified theory for the $p$ -adic geometry of Shimura varieties.
Bridge between Structures: It explicitly connects the Sen operator (Galois side) and the Fontaine operator (de Rham side) to differential operators on the flag variety, offering a concrete geometric realization of abstract $p$ -adic Hodge theory concepts.

In summary, Lue Pan's work provides a powerful geometric toolkit to analyze the locally analytic completed cohomology of modular curves, revealing that the "locally analytic" structure is naturally encoded in the differential geometry of the Hodge-Tate period map to $\mathbb{P}^1$ .