🔢 mathematics

Towards the Fontaine--Mazur conjecture for GL(2)

This paper proves new cases of the Fontaine--Mazur conjecture for GL(2) by combining a novel modularity result with the geometry of numbers.

Original authors: Jack A. Thorne

Published 2026-08-10

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Jack A. Thorne

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 Great Number Detective Story

Imagine the universe of numbers as a vast, ancient library. Inside this library, there are two distinct but deeply connected sections. One section, called Number Theory, is filled with books about integers, primes, and the hidden patterns that govern how numbers divide and multiply. The other section, Geometry, is filled with blueprints for shapes, spaces, and the curves that exist in higher dimensions. For a long time, mathematicians have suspected that these two sections are actually talking to each other. They believe that every mysterious pattern found in the world of numbers (specifically, certain "representations" which are like complex codes describing how numbers interact) can be traced back to a specific geometric shape, like a curve or a surface.

This idea is known as the Fontaine–Mazur conjecture. Think of it as a detective's rule: "If a number-pattern looks like it was made by geometry, it must have been made by geometry." The challenge is that these patterns are incredibly complex, often hiding in the shadows of "p-adic" numbers (a strange, alternative way of measuring distance between numbers). Proving the rule requires showing that a specific, invisible number-code is actually just a shadow cast by a real, physical geometric object. This paper is about cracking a particularly stubborn case in this investigation, specifically for a type of code involving the number 2 (GL2), where previous methods hit a wall.

The Paper's Big Breakthrough

In this paper, Jack A. Thorne acts as a master detective who combines two very different tools to solve a case that had been stuck for a long time. The case involves proving that certain mysterious number-patterns (called Galois representations) actually come from geometry, just as the Fontaine–Mazur conjecture predicts.

The Problem:
Previously, mathematicians could only solve these cases if the number-patterns were "strong" and "irreducible" (meaning they couldn't be broken down into simpler pieces). It was like trying to identify a person only if they were wearing a bright, unique costume. If the person was wearing a plain, common outfit (a "reducible" pattern), the old methods failed. The paper specifically tackles the difficult scenario where the pattern is reducible, a situation that had resisted solution for general number fields.

The Solution:
Thorne introduces a new strategy that mixes a fresh type of "modularity lifting" theorem with a technique called the "geometry of numbers."

The "Close Enough" Trick (Modularity Lifting): Imagine you have a blurry photo of a suspect (a modular representation) and a slightly different, clearer photo (the representation you want to prove is modular). Old rules said the photos had to be almost identical to prove they were the same person. Thorne's new rule is much more flexible: he proves that if the photos are "close enough" (matching up to a certain level of detail, defined by a specific mathematical constant), they are definitely the same person. Crucially, this "closeness" doesn't depend on the global complexity of the case, but only on local details, making the rule much easier to apply.
The "Geometry of Numbers" Approximation: To use the "close enough" trick, Thorne needs to find a "modular" photo that matches his target. He constructs these approximations using abelian varieties (which are like multi-dimensional donuts with special number-properties). He uses the "geometry of numbers" to ensure these donuts can be built to look exactly like the target pattern in the local neighborhoods where it matters. It's like building a custom-made suit that fits the suspect perfectly, even if the suspect is wearing a disguise.

The Result:
By combining these tools, Thorne proves Theorem B. He shows that for a totally real number field (a specific type of number system), if a continuous, irreducible representation satisfies certain conditions (it's unramified almost everywhere, behaves nicely at prime numbers, and has a specific "odd" determinant), then it is potentially modular.

What does "potentially modular" mean? It means that if you expand your view to a slightly larger number field (a finite extension), this mysterious pattern reveals itself to be the shadow of a genuine geometric object. In simpler terms, the paper proves that these stubborn, reducible patterns are indeed connected to geometry, provided you look at them from the right angle.

Specific Cases:
The paper also deduces Theorem D, which applies this result to the case where the number field is just the rational numbers (the standard fractions and integers). It proves that even when the residual representation is reducible (the "plain outfit" case) and the prime number is 2 (a notoriously difficult case), the pattern is still modular. This fills a significant gap in our understanding, confirming that the connection between numbers and geometry holds even in these tricky, previously unsolved scenarios.

The paper does not claim to solve the entire Fontaine–Mazur conjecture for all possible cases, but it successfully proves new, specific cases that were previously out of reach, particularly those involving reducible representations. The proof is rigorous and relies on constructing specific mathematical objects (moduli spaces of abelian varieties) and using ultrafilters to "patch" together local solutions into a global proof. The confidence is high: the paper presents a complete, logical deduction that these specific representations arise from geometry.

Technical Summary: Towards the Fontaine–Mazur Conjecture for GL2

Problem Statement
The paper addresses the Fontaine–Mazur conjecture for $GL_2$ , specifically Conjecture A, which posits that continuous, irreducible Galois representations $\rho: G_F \to GL_2(\mathbb{Q}_p)$ (where $F$ is a totally real number field) satisfying specific local conditions (unramified almost everywhere, de Rham with distinct Hodge–Tate weights at $p$ , and totally odd determinant) arise from geometry.

A significant gap in the existing literature concerns the case where the residual representation $\bar{\rho}: G_F \to GL_2(\mathbb{F}_p)$ is reducible. While modularity lifting theorems have successfully proven the conjecture for "sufficiently non-degenerate" (e.g., irreducible) residual representations, these methods generally fail when $\bar{\rho}$ is reducible. Previous results in the reducible case, such as those by Skinner and Wiles [SW99], were largely restricted to the case $F = \mathbb{Q}$ and have resisted generalization to arbitrary totally real fields.

Methodology
The author, Jack A. Thorne, introduces a new approach that combines a novel modularity lifting theorem with the "geometry of numbers" to handle the reducible residual case over general totally real fields. The proof strategy relies on two primary technical pillars:

A New Modularity Lifting Theorem (Theorem 4.1):
The paper establishes a modularity lifting result where the required "closeness" between a target representation $\rho$ and a known modular representation $\rho'$ is measured by a congruence modulo $\varpi^C$ . Crucially, the exponent $C$ depends only on local data (ramified places) and the image of the global representation, rather than on subtle global invariants like the rank or cardinality of Selmer groups. This allows the constant $C$ to remain stable under base change to suitable totally real extensions. The proof utilizes a Taylor–Wiles patching argument, constructing a patched deformation ring $R_\infty$ and a patched Hecke algebra $T_\infty$ . A key innovation involves the use of ultrafilters to construct these rings and the application of commutative algebra results (specifically Theorem 2.12 regarding the support of points on irreducible components of spectra) to deduce that if a modular point exists close enough to a target point, the target point must also lie on the same irreducible component, hence is modular.
Construction of Modular Approximations via Geometry of Numbers (Section 3):
To apply the lifting theorem, one must construct a modular representation $\rho'$ that approximates $\rho$ modulo $\varpi^C$ . The paper constructs these approximations using Hilbert–Blumenthal Abelian Varieties (HBAVs).
- The author constructs a moduli space of polarized HBAVs with level structure.
- Using the "geometry of numbers" (specifically results on the distribution of lattice points in convex bodies, following [DH98] and [Wid12]), the paper proves the existence of totally real extensions $F'/F$ and HBAVs over $F'$ whose Tate modules approximate the local restrictions of $\rho$ to a high degree of precision.
- This construction ensures the existence of a modular representation $\rho'$ (arising from the cohomology of these HBAVs) that is congruent to $\rho$ modulo a sufficiently high power of the uniformizer.

Key Contributions and Results

Theorem B (Main Result): The paper proves that for a totally real field $F$ and a prime $p$ , any continuous, irreducible representation $\rho: G_F \to GL_2(\mathbb{Q}_p)$ satisfying the Fontaine–Mazur conditions (unramified almost everywhere, potentially crystalline ordinary of weights $\{0, 1\}$ at $p$ , and totally odd determinant) is potentially modular. This means there exists a finite totally real extension $F'/F$ such that $\rho|_{G_{F'}}$ arises from a cuspidal, regular algebraic automorphic representation of $GL_2(\mathbb{A}_{F'})$ .
- Significance: This result applies even when the residual representation $\bar{\rho}$ is reducible, a case previously inaccessible for general totally real fields.
Theorem D (Application to $\mathbb{Q}$ ): As a corollary, the paper proves the modularity of $\rho: G_\mathbb{Q} \to GL_2(\mathbb{Q}_p)$ under the same conditions, specifically addressing the previously open case where $p=2$ and $\bar{\rho}$ is reducible. This is deduced from Theorem B combined with Serre's conjecture.
Commutative Algebra and Group Determinants: The paper provides refined results on the structure of deformation rings and group determinants (Proposition 2.13, Theorem 2.12), establishing conditions under which points on deformation rings lie on the same irreducible component based on $p$ -adic proximity.

Significance and Claims
The paper claims to provide a new approach to the Fontaine–Mazur conjecture that is robust in the presence of reducible residual representations. By decoupling the required congruence level from global Selmer group sizes and relying instead on local data and the geometry of numbers, the author overcomes a major obstruction in the generalization of modularity lifting theorems to totally real fields.

The work does not claim to prove the full Fontaine–Mazur conjecture (which would require removing the "potentially" in Theorem B or handling non-ordinary cases), but rather establishes a significant new class of cases where potential modularity is guaranteed. The author notes that the potential modularity implies the representation lives in a compatible system, and under additional conditions (existence of a place where the Weil–Deligne representation is not potentially unramified), this implies the representation arises from geometry.

The paper explicitly states that the vast majority of cases for $F=\mathbb{Q}$ were already known via Serre's conjecture and lifting theorems, but this work specifically targets the "open case" of $p=2$ with reducible residual image, extending the reach of modularity results to a broader class of representations and fields.

The Great Number Detective Story

The Paper's Big Breakthrough

More like this