🔢 mathematics

The $\rho$ -Fourier transform

This paper proves a major portion of the Braverman-Kazhdan-Ngô conjectures by constructing the $\rho$ -Fourier transform for reductive groups over arbitrary fields and establishing the corresponding $\rho$ -Schwartz space for non-Archimedean fields, using spectral methods.

Original authors: Jayce R. Getz, Armando Gutiérrez Terradillos, Farid Hosseinijafari, Aaron Slipper, Guodong Xi, HaoYun Yao, Alan Zhao

Published 2026-05-21

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Jayce R. Getz, Armando Gutiérrez Terradillos, Farid Hosseinijafari, Aaron Slipper, Guodong Xi, HaoYun Yao, Alan Zhao

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 trying to understand the "music" of a complex system, like a symphony orchestra or a vast, echoing cave. In mathematics, this system is a group (a set of symmetries or transformations), and the "music" is the collection of all possible functions or patterns that can exist within it.

For decades, mathematicians have been trying to build a special "tuning fork" for these groups. This tuning fork is called the $\rho$ -Fourier Transform. Its job is to take a complicated pattern, translate it into a different language (the "frequency" or "spectral" language), and then translate it back perfectly.

Here is a simple breakdown of what this paper achieves, using everyday analogies:

1. The Big Problem: The Missing Tuning Fork

In the world of numbers and shapes, there is a famous tool called the Fourier Transform (think of how your phone converts sound waves into digital data). It works beautifully for simple things like waves on a string.

However, when mathematicians tried to build this tool for complex, high-dimensional symmetries (called reductive groups), they hit a wall. They knew the tool should exist and have specific magical properties (like turning a pattern inside out and then back again perfectly), but they couldn't actually build it. They had a blueprint (a conjecture by Braverman, Kazhdan, and Ngˆo), but no construction crew.

2. The Solution: Building the Bridge

The authors of this paper, led by Jayce Getz and colleagues, finally built the bridge. They didn't just guess; they constructed the tool explicitly.

The Analogy: Imagine you have a giant, foggy room (the group) where you can't see the walls. You want to know what the room looks like from the outside. The authors built a special camera (the $\rho$ -Fourier Transform) that can look at the fog, translate it into a clear blueprint, and then reconstruct the room perfectly.
The Method: Instead of trying to build the camera piece by piece from the ground up, they used a "spectral" approach. They looked at the "echoes" inside the room (the tempered representations). By understanding how these echoes behave, they could define the camera's lens mathematically. They proved that if you look at the echoes correctly, the camera naturally appears.

3. The "Schwartz Space": The Perfect Library

The paper also constructs a special library of functions called the $\rho$ -Schwartz space.

The Analogy: Think of the whole group as a massive, chaotic library containing every possible book (function). Most of these books are messy, torn, or infinite. The authors wanted to find a specific, neat section of the library where the books are:
1. Well-behaved: They don't explode to infinity.
2. Organized: They fit a specific pattern.
3. Self-reflecting: If you use the "tuning fork" (Fourier Transform) on a book in this section, the result is still a book in this same section.

They successfully carved out this special section. They proved that this section is the "Goldilocks zone"—it's not too big (like the whole chaotic library) and not too small (like just a few simple books). It's the perfect size to hold the magic of the Fourier Transform.

4. Two Types of Worlds: Discrete vs. Continuous

The authors had to build their tool for two different types of worlds:

The Non-Archimedean World (Discrete): This is like a world made of distinct, separate blocks (like pixels or integers). Here, they built the library perfectly. Every book fits exactly where it should.
The Archimedean World (Continuous): This is like a smooth, flowing river (like real numbers). Here, the terrain is trickier. They built a very strong "approximation" of the library. It's not quite as perfect as the discrete version, but it's close enough to prove the theory works and to do the heavy lifting.

5. The "Basic Function": The Master Key

A key part of their discovery is something called the basic function ( $b_\rho$ ).

The Analogy: Imagine a master key that fits every lock in the library. The authors found a specific function that, when you apply the Fourier Transform to it, it stays exactly the same. It is the "center of gravity" of their new mathematical universe. They proved this key exists and behaves exactly as the old blueprints predicted.

6. Why This Matters (According to the Paper)

The paper claims that by building this tool, they have:

Proven the Conjecture: They showed that the "tuning fork" exists and works exactly as the famous mathematicians Braverman, Kazhdan, and Ngˆo predicted.
Unified the Theory: Their method works for almost all types of these complex groups, regardless of whether the world is discrete or continuous.
Laid the Foundation: They didn't just find a solution; they built the "firm ground" (the Harish-Chandra space and Plancherel formula) upon which future mathematicians can build more complex theories.

In summary: The paper is like a master architect who finally constructed the missing bridge between two islands of mathematics. They didn't just say the bridge could exist; they built it, showed it's sturdy, and proved it connects the two sides perfectly, allowing traffic (mathematical ideas) to flow freely between them.

Technical Summary of "The $\rho$ -Fourier Transform"

Problem Statement
The paper addresses a conjecture by Braverman, Kazhdan, and Ngˆo regarding the existence of a $\rho$ -Fourier transform and an associated $\rho$ -Schwartz space for a reductive group $G$ over a local field $F$ . Let $\rho: {}^L G \to \mathrm{GL}(V_\rho)$ be a representation of the $L$ -group of $G$ . The conjecture posits the existence of a dense, $G(F) \times G(F)$ -invariant subspace $S_\rho \subset L^2(G(F))$ equipped with a unitary operator $\mathcal{F}_\rho$ (the $\rho$ -Fourier transform) satisfying specific equivariance and functional equation properties. Crucially, for tempered representations $\pi$ , the transform must satisfy $\pi(\mathcal{F}_\rho(f)^\vee) = \gamma(1/2, \pi, \rho, \psi)\pi(f)$ , where $\gamma$ is the local $\gamma$ -factor. Furthermore, the space $S_\rho$ should satisfy analytic conditions regarding local zeta integrals, including convergence, meromorphic continuation, and functional equations involving $L$ -functions, as well as geometric constraints such as being "almost compactly supported" in the non-Archimedean case.

Methodology
The authors employ a spectral approach, utilizing the Harish-Chandra Plancherel theorem as the primary tool for construction. Rather than constructing the transform via integral kernels on the group side directly, they define the Fourier transform spectrally on the space of sections over the tempered dual.

Spectral Construction: The authors define the $\rho$ -Fourier transform $\mathcal{F}_\rho$ on the Harish-Chandra space $C(G(F))$ by declaring that its action on the spectral side (via the Harish-Chandra transform $H_P$ ) is multiplication by the $\gamma$ -factor $\gamma_\rho(\pi)$ . Specifically, $\mathcal{F}_\rho = (\cdot)^\vee \circ H_P^{-1} \circ (\gamma_\rho \cdot) \circ H_P$ .
Local Langlands Correspondence: The construction relies on a weak form of the local Langlands correspondence (LLC) satisfying specific desiderata (LLC1–LLC8). These include compatibility with the LLC for tori and $\mathrm{GL}_n$ , behavior under contragredients, and specific analytic properties of $\gamma$ -factors and $L$ -functions. The authors verify that these conditions hold in known cases (e.g., $G=\mathrm{GL}_n$ ) and assume them generally.
Asymptotic Schwartz Space: To handle the analytic constraints on zeta integrals, the authors first construct an "asymptotic $\rho$ -Schwartz space" $S^{\mathrm{as}}_\rho(G(F))$ . This space is defined as the inverse image under $H_P$ of sections $T$ such that $L_\rho^{-1} T$ is a polynomial (non-Archimedean) or holomorphic (Archimedean) section, subject to specific growth bounds involving the $L$ -functions attached to $\rho$ .
Geometric Control (Non-Archimedean): In the non-Archimedean case, the authors refine $S^{\mathrm{as}}_\rho(G(F))$ to obtain the actual $\rho$ -Schwartz space $S_\rho(G(F))$ . They intersect the asymptotic space with the space of almost compactly supported smooth functions $C^\infty_{\mathrm{ac}}(G(F))$ . Using the Paley-Wiener theorem for $C^\infty_{\mathrm{ac}}(G(F))$ (due to Bernstein and Heiermann) and the extended Bernstein center, they prove that this intersection contains a maximal subspace stable under $\mathcal{F}_\rho$ .
Bounds and Estimates: The paper dedicates significant effort to establishing uniform bounds for $\gamma$ -factors and $L$ -functions over the tempered dual, particularly in the Archimedean case, to ensure the convergence and continuity of the constructed operators.

Key Contributions and Results

Existence of the Transform: The authors prove that the Harish-Chandra space $C(G(F))$ admits a unitary $\rho$ -Fourier transform $\mathcal{F}_\rho$ satisfying the functional equation $\pi(\mathcal{F}_\rho(f)^\vee) = \gamma(1/2, \pi, \rho, \psi)\pi(f)$ for all tempered representations $\pi$ . This confirms Conjecture 1.1 of Braverman-Kazhdan-Ngˆo.
Construction of the Schwartz Space:
- Archimedean Case: They construct the space $S_\rho(G(F)) = S^{\mathrm{as}}_\rho(G(F))$ , proving it satisfies the analytic desiderata (S1) and (S2) of Conjecture 1.5.
- Non-Archimedean Case: They construct $S_\rho(G(F))$ as the maximal $\mathcal{F}_\rho$ -stable subspace of $S^{\mathrm{as}}_\rho(G(F)) \cap C^\infty_{\mathrm{ac}}(G(F))$ . They prove this space satisfies all conditions of Conjecture 1.5, including the existence of a basic function $b_\rho$ (Conjecture S3) and the almost compact support property (Conjecture S4).
Kernel and Basic Function: The paper explicitly constructs the kernel distribution $J_\rho$ and the basic function $b_\rho$ (in the unramified case), showing that $b_\rho$ is a fixed point of the Fourier transform and its zeta integral recovers the $L$ -function $L(s, \pi, \rho)$ .
Uniformity: The methods are shown to proceed uniformly across different types of ramification and for all places of a number field, contrasting with previous works that were often restricted to specific cases or geometrized settings.

Significance and Claims
The paper claims to prove a "large portion" of the Braverman-Kazhdan-Ngˆo conjectures on Schwartz spaces. The primary significance lies in providing a direct, spectral construction of the $\rho$ -Fourier transform, thereby placing the theory on "firm ground." The authors emphasize that their approach does not rely on the geometrization techniques used in previous works (e.g., by Lafforgue or in the equal characteristic case) but instead leverages the Harish-Chandra Plancherel formula and the local Langlands correspondence.

The authors are modest regarding the Archimedean case, noting that while they construct an approximation (the asymptotic space) that satisfies the analytic conditions, the construction of a Schwartz space consisting of almost compactly supported functions in the Archimedean setting remains unclear and requires further work. They also note that their results are conditional on the validity of specific desiderata for the local Langlands correspondence, which are known to hold for $\mathrm{GL}_n$ and other specific groups.

The paper concludes by stating that the constructed spaces satisfy the local conjectures regarding zeta integrals and functional equations, providing a foundation for potential global applications, such as the construction of modulation groups, though the paper itself does not develop these global applications.