🔢 mathematics

Schwartz spaces on L-monoids: non-Archimedean

This paper completes the Braverman-Kazhdan-Ngô program over non-Archimedean local fields by establishing the existence of Schwartz spaces on L-monoids, relying on the local Langlands conjecture for tempered representations and specific assumptions on $\gamma$ -factors, which renders the program unconditional for general linear groups.

Original authors: Chun-Hsien Hsu, HaoYun Yao

Published 2026-07-31

📖 9 min read🧠 Deep dive

CC BY 4.0

Original authors: Chun-Hsien Hsu, HaoYun Yao

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

=== DRAFT ===
Imagine a vast, invisible landscape where numbers dance and shapes twist in ways we can't quite see. This is the world of number theory, specifically the branch that studies how numbers behave in "local" neighborhoods—tiny, self-contained universes called local fields. In this world, mathematicians are obsessed with a special kind of symmetry called "Langlands functoriality." Think of it as a universal translator that allows different groups of numbers to speak to one another, revealing hidden patterns that connect seemingly unrelated mathematical objects. To make this translation work, mathematicians need a special tool: a "Schwartz space." If you imagine the numbers as a chaotic crowd, a Schwartz space is a perfectly organized, quiet room where the noise is controlled, and the most important signals (called "L-functions") can be heard clearly. These signals are like the DNA of numbers, encoding deep secrets about prime numbers and the structure of the universe. For decades, mathematicians have been trying to build the perfect room for these signals, but in certain complex situations, the walls kept collapsing, and the noise got too loud.

This paper, written by Chun-Hsien Hsu and Haoyun Yao, steps into that chaotic construction site to finish a massive project known as the Braverman-Kazhdan-Ngô program. The authors are tackling a specific problem: how to define this perfect "room" (the Schwartz space) when working with non-Archimedean local fields—a type of number system that behaves very differently from the real numbers we use every day, more like a digital grid than a smooth line. They are building on a foundation laid by previous giants in the field, who had already figured out how to handle the "easy" cases but left the "hard" cases (where the geometry gets isotropic, or slippery) unfinished. The authors' main finding is that they have successfully constructed this missing room. They proved that three different ways of trying to build this space actually result in the exact same thing, confirming that they have found the correct, stable structure. They also showed that inside this room, there is a special "basic function" that acts like a master key, and they demonstrated how to use it to unlock the greatest common divisors of complex mathematical integrals. Crucially, they have proven this construction works perfectly and unconditionally for general linear groups (a specific, well-behaved family of number groups). For other, more twisted groups, the program is complete only if we accept a few widely believed assumptions about how certain mathematical factors behave. In short, they have completed the blueprint for a crucial part of the mathematical universe, turning a shaky scaffold into a solid, working machine for the most important cases, while paving the way for the rest.

The Story of the Perfect Room

Imagine you are trying to organize a massive, chaotic library. The books are not just paper and ink; they are living, breathing mathematical functions that represent the deep symmetries of numbers. Your goal is to find a specific section of the library—a "Schwartz space"—where these functions behave nicely. In this space, if you take a function and perform a "Fourier transform" (a magical operation that swaps the function's perspective, like looking at a sculpture from the front and then from the back), the result should still be a well-behaved function in the same library.

For a long time, mathematicians knew how to build this library for simple, straight-line groups. But when the groups got more complex—specifically, when they had "isotropic" parts, which is a fancy way of saying they had directions that could stretch infinitely without hitting a wall—the old blueprints failed. The functions would start behaving badly, leaking out of the library or becoming too wild to control.

Hsu and Yao stepped in to fix this. They didn't just patch the holes; they realized that three different groups of mathematicians had been trying to build the same room using three different sets of rules.

The Asymptotic Group: They tried to build the room by looking at how functions behave as they get very large (asymptotic behavior).
The Compatible Spectrum Group: They tried to build it by ensuring the "frequencies" of the functions matched up perfectly across different layers of the library.
The Analytic Group: They tried to build it by strictly controlling how fast the functions could grow or shrink.

The authors proved a stunning result: All three groups were actually building the exact same room. They showed that if a function fits the rules of one group, it automatically fits the rules of the other two. This means they finally found the correct definition of the Schwartz space for these tricky, non-Archimedean fields.

The Magic of the "Basic Function"

Inside this newly confirmed room, the authors discovered a very special object called the basic function (denoted as $b_\rho$ ). You can think of this as the "master key" or the "perfect seed" of the library.

It is supported on a specific geometric shape called an L-monoid. Imagine the library isn't just a flat room but a multi-dimensional shape with corners and edges. The basic function lives entirely within this shape, which is constructed from the specific symmetries ( $\rho$ ) the mathematicians are studying.
When you apply the Fourier transform to this basic function, it stays exactly the same (if the conditions are right). It is a perfect reflection of itself.
Most importantly, this function allows mathematicians to calculate the greatest common divisor (GCD) of a whole bunch of complicated integrals. In the world of number theory, finding the GCD of these integrals is like finding the common thread that ties all the numbers together. The authors showed that by using this basic function, you can extract the "L-factors"—the fundamental DNA of the numbers—directly from the chaos.

The "Slippery" Problem and the Solution

One of the biggest headaches in this field was a "pathological" problem. Sometimes, a function would look like it belonged in the library, but when you tried to translate it (using the Fourier transform), it would break the rules. It was like a book that looked fine on the shelf but turned into a pile of confetti when you opened it.

The authors identified exactly why this happened. It turned out that the old definitions didn't account for functions that were "compatible" across different layers of the library. They introduced a new condition called compatible spectrum. This condition ensures that if a function behaves well in one part of the library, it behaves well in all the connected parts. By adding this rule, they filtered out the "confetti" functions and kept only the sturdy, well-behaved ones.

They also proved that this new, corrected library is stable. If you take a function from the library, translate it, and put it back, it stays in the library. This stability is crucial because it means the mathematical machinery works reliably.

What About the Rest of the Universe?

The paper is very careful about what it claims to have solved.

What is Proven: The construction of the Schwartz space is complete and proven unconditionally for general linear groups. For these groups, the "basic function" exists, the Fourier transform works perfectly, and the GCD of the integrals is exactly the L-factor.
What is Assumed: For other, more complex groups (like classical groups), the proof relies on a few "natural assumptions" about how certain mathematical factors (called $\gamma$ -factors) behave. The authors believe these assumptions are true (and they are true for general linear groups), but they haven't proven them from scratch for every possible group. They leave this as a task for future work.
What is Suggested: The authors have a strong hunch (a conjecture) that the "basic function" and the Schwartz space are "local." This means the function's behavior in one small area depends only on the geometry of that specific area, not on the whole universe. They haven't proven this yet, but they have verified it for some specific examples, like tori (donut-shaped groups) and symmetric powers of $GL_2$ .

The Geometry of the Numbers

To make all this work, the authors had to build a new kind of geometric map. They used affine toric varieties, which are shapes built from cones and rays. Imagine a star-shaped cookie cutter. The "L-monoid" is the shape you get when you press this cutter into the dough of your number system. The authors showed that the "basic function" lives entirely inside this cookie shape. This connection between abstract algebra (the groups) and geometry (the shapes) is the heart of the paper. It shows that the rules for organizing these number functions are dictated by the shape of the underlying geometric space.

Why Should You Care?

You might wonder, "Why do we need a perfectly organized room for number functions?" The answer lies in the Langlands Program, one of the biggest unsolved mysteries in mathematics. This program tries to connect two completely different worlds: the world of numbers (arithmetic) and the world of symmetry (representation theory). The "Schwartz space" is the bridge between them. Without a stable, well-defined bridge, the connection is shaky, and we can't prove the deep theorems that link prime numbers to symmetries.

By completing this part of the bridge, Hsu and Yao have given mathematicians a solid footing to walk across. They have shown that for a huge class of groups, the bridge is not just a theory; it is a concrete, working structure. This paves the way for proving even more profound results about the nature of numbers, potentially unlocking secrets about prime numbers that have remained hidden for centuries.

In the end, this paper is a triumph of organization. It took a messy, confusing situation where functions were leaking and breaking, and it built a sturdy, geometrically sound room where everything fits perfectly. It's a reminder that even in the most abstract corners of mathematics, the right definition can turn chaos into clarity.

Technical Summary: Schwartz Spaces on L-Monoids: Non-Archimedean

Problem Statement
This paper addresses the completion of the Braverman-Kazhdan-Ngô (BKN) program over non-Archimedean local fields. The program aims to associate a Fourier theory to a representation $\rho$ of the $L$ -group $L G$ of a connected reductive group $G$ , generalizing Tate's thesis and the Godement-Jacquet theory for $GL_n$ . Specifically, the goal is to construct a "Schwartz space" $\mathcal{S}_\rho(G(F))$ and a unitary Fourier transform $\mathcal{F}_{\rho, \psi}$ such that local $L$ -factors arise as the greatest common divisors (gcd) of local zeta integrals.

Previous work by GGH+26 constructed an "asymptotic $\rho$ -Schwartz space" $\mathcal{S}^{as}_\rho(G(F))$ and a Fourier transform on the Harish-Chandra Schwartz space $\mathcal{C}(G(F))$ . However, the authors identify a critical obstruction: when $G/Z_G$ is isotropic, $\mathcal{S}^{as}_\rho(G(F))$ contains functions with pathological analytic behaviors. Specifically, for non-supercuspidal discrete series, the zeta integrals of functions in $\mathcal{S}^{as}_\rho(G(F))$ may fail to satisfy compatibility conditions required for the Fourier transform to preserve the space, and the space may not coincide with the expected geometric support on the associated $L$ -monoid.

Methodology
The authors employ a combination of spectral analysis, harmonic analysis on reductive groups, and algebraic geometry of toric varieties.

Spectral Construction and Compatibility: The paper introduces two alternative characterizations of the correct Schwartz space to resolve the obstructions found in $\mathcal{S}^{as}_\rho(G(F))$ :
- Compatible Spectrum ( $C_{cs}$ ): A subspace of functions whose zeta integrals satisfy specific compatibility conditions relating induced representations from different Levi subgroups (Definition 3.9). This ensures that matrix coefficients of subrepresentations and their lifts yield consistent zeta integrals.
- Analytic Definition ( $S_\rho$ ): A space defined by the absolute convergence of zeta integrals in positive cones and the meromorphic extension of these integrals divided by $L$ -factors (Definition 4.3).
Equivalence of Definitions: The core technical effort involves proving that three distinct definitions of the Schwartz space coincide (up to a central shift $|\nu|^{-1/2}$ ):
- The intersection of the asymptotic space and the compatible spectrum space: $\mathcal{S}^{as}_\rho(G(F)) \cap C_{cs}(G(F))$ .
- The intersection of the asymptotic space and the almost compactly supported space: $\mathcal{S}^{as}_\rho(G(F)) \cap C^\infty_{ac}(G(F))$ .
- The analytically defined space $S_\rho(G(F))$ .
  This equivalence is established using a modification of the matrix Paley-Wiener theorem (Heiermann) and Harish-Chandra's Plancherel theorem, specifically analyzing the support of functions on the Cartan decomposition.
Assumptions on Local Langlands Correspondence (LLC): The proofs rely on the tempered Local Langlands Correspondence (LLC1–LLC6) for all Levi subgroups. Crucially, the stability of the Schwartz space under the Fourier transform $\mathcal{F}_{\rho, \psi}$ is shown to be equivalent to a new assumption, (LLC7), concerning the holomorphy and equality of $\gamma$ -factors for non-supercuspidal discrete series and their supercuspidal supports.
Geometric Realization: The paper studies the underlying geometry of the Schwartz space by associating $\rho$ with a normal reductive monoid $X_\rho$ (an $L$ -monoid). The authors analyze the support of functions in $S_\rho(G(F))$ and show they are compactly supported within the $F$ -points of this monoid, $X_\rho(F)$ .

Key Contributions and Results

Theorem 1.2 (Equivalence of Spaces): The authors prove that the analytically defined space $S_\rho(G(F))$ is equal to $|\nu|^{-1/2}(\mathcal{S}^{as}_\rho(G(F)) \cap C_{cs}(G(F)))$ . Furthermore, they establish that functions in this space have compact support in the $L$ -monoid $X_\rho(F)$ .
Theorem 1.4 (GCD and Functional Equations): Assuming (LLC7), the paper proves that local $L$ -factors $L(s, \pi, \rho)$ are realized as the greatest common divisors of the zeta integrals $Z(s, f, c)$ for $f \in S_\rho(G(F))$ and matrix coefficients $c \in C(\pi)$ . Specifically, the span of these zeta integrals forms the fractional ideal generated by $L(s, \pi, \rho)$ .
Existence of Basic Functions: The authors construct a "basic function" $b_\rho \in S_\rho(G(F))$ such that its zeta integral recovers the $L$ -factor $L(s, \pi, \rho)$ for unramified representations.
Fourier Transform Stability: The paper demonstrates that the stability of the Schwartz space under the Fourier transform $\mathcal{F}_{\rho, \psi}$ is equivalent to assumption (LLC7). It is noted that (LLC7) holds unconditionally for general linear groups ( $GL_n$ ) due to the Bernstein-Zelevinsky classification.
Naïve Schwartz Space Limitation: The authors show that the naïve Schwartz space $C^\infty_c(G(F)) + \mathcal{F}_\rho(C^\infty_c(G(F)))$ is generally a proper subspace of the correct Schwartz space (unless $G$ has split rank 1), as it fails to contain the basic function $b_\rho$ .
Geometric Support: For tori and specific examples ( $GL_2$ symmetric powers), the paper verifies that the Schwartz space consists of functions supported on the integral points of the associated affine toric variety or reductive monoid.

Significance and Claims
The paper claims to complete the BKN program over non-Archimedean local fields under the assumption of the tempered Local Langlands Correspondence and the specific condition (LLC7) on $\gamma$ -factors.

Unconditional for $GL_n$ : Since (LLC7) is known for general linear groups, the program is stated to be complete unconditionally for $GL_n$ .
Resolution of Pathologies: The work resolves the "pathological examples" identified in previous literature where the asymptotic Schwartz space failed to behave correctly for isotropic groups, providing a rigorous definition of the Schwartz space that aligns with the expected geometric support on $L$ -monoids.
Foundation for Global Theory: By establishing the local functional equations and the gcd property of $L$ -factors, the paper provides the necessary local ingredients for the global Langlands functoriality conjectures within the BKN framework.

The authors modestly note that the "locality" of the Schwartz space (i.e., whether it is a module over smooth functions on the monoid) remains a conjecture (Conjecture 1.5 and 1.6), which they verify for split tori and specific examples but leave for future work for general groups. The paper does not claim to prove (LLC7) for classical groups beyond $GL_n$ , noting this as a direction for future research.