🔢 mathematics

Quotient branching law for $p$ -adic $(\mathrm{GL}_{n+1}, \mathrm{GL}_n)$ I: generalized Gan-Gross-Prasad relevant pairs

This paper establishes a necessary and sufficient condition for the non-vanishing of the Hom-space between irreducible smooth representations of $\mathrm{GL}_{n+1}$ and $\mathrm{GL}_n$ over a non-Archimedean local field, thereby resolving the quotient branching law problem and providing a generalized Pieri's rule for affine Hecke algebras through the characterization of Bernstein-Zelevinsky derivatives.

Original authors: Kei Yuen Chan

Published 2026-08-07

📖 4 min read🧠 Deep dive

CC0 1.0

Original authors: Kei Yuen Chan

Original paper dedicated to the public domain under CC0 1.0 (http://creativecommons.org/publicdomain/zero/1.0/). ✨ 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 a universe made entirely of invisible, shifting patterns. In this world, mathematicians study "symmetry groups," which are like rulebooks for how these patterns can be rearranged without breaking them. Think of a group as a massive, intricate dance troupe where every dancer knows exactly how to move in relation to everyone else. Now, imagine you have a huge troupe (let's call it the "Big Group") and a slightly smaller one (the "Small Group") that is part of the Big Group's routine. The Big Group is performing a complex, unique dance (an "irreducible representation"). The big question is: if you watch just the Small Group's part of that dance, does it look like a specific, known dance routine that the Small Group already knows how to do?

This is the "branching law" problem. It's like asking: "If I take a giant, complicated recipe and only look at the ingredients used for the dessert, does it match a specific, simple cookie recipe I already have?" Mathematicians care about this because understanding how big, complex systems break down into smaller, simpler parts is key to unlocking the secrets of everything from the structure of the universe to the behavior of subatomic particles. For a long time, mathematicians knew that these smaller parts existed, but they didn't have a perfect, step-by-step rulebook to tell them exactly when a match would happen and when it wouldn't. They had clues, but no complete map.

Enter a new paper by Kei Yuen Chan, which acts like a master key for this specific puzzle involving a famous family of groups called $GL_n$ (General Linear groups). The paper solves a decades-old mystery by providing a precise, "necessary and sufficient" condition. This means the author has found a rule that works 100% of the time: if the rule says "yes," the match is guaranteed; if it says "no," the match is impossible. There is no guessing left.

The paper introduces a clever new way to check for these matches using something called "generalized GGP relevant pairs." To understand this, imagine the Big Group's dance is built from a stack of Lego blocks. The author developed a method to take the Big Group's complex structure apart, layer by layer, using a process called "Bernstein-Zelevinsky derivatives." Think of these derivatives as a special tool that peels off the outermost layers of the Lego tower to reveal what's underneath. The paper proves that if you peel off the right layers in the right order, you can see if the remaining core matches the Small Group's dance.

The author didn't just guess this rule; they proved it rigorously. They showed that a match happens if and only if two specific conditions are met: first, the "peeled" versions of the two dances must look identical in a specific way, and second, the way the layers were peeled must follow a strict "commutative" pattern (meaning the order of peeling doesn't mess up the final result). They also proved that any simple piece you find after peeling can be built by starting with the most basic, fundamental blocks (called "essentially square-integrable representations") and stacking them up.

This result is a big deal because it unifies many different cases that mathematicians had solved separately before. It's like finding a single formula that explains why a square fits in a square hole, why a circle fits in a round hole, and why a triangle fits in a triangular hole, all at once. The paper also connects this to a famous rule in combinatorics called "Pieri's rule," which describes how to add boxes to a grid in a specific way, effectively translating a complex dance problem into a simple game of stacking blocks.

In short, this paper hands mathematicians a complete, foolproof checklist. If you want to know if a complex pattern from a large group contains a specific pattern from a smaller group, you just run the numbers through this new "relevance" test. If the test passes, the connection is real. If it fails, it's not. No more "maybe," no more "sometimes." The mystery of when these mathematical dances align has been solved with absolute certainty.

Technical Summary: Quotient Branching Law for p-adic (GL $_{n+1}$ , GL $_n$ ) I: Generalized Gan-Gross-Prasad Relevant Pairs

1. Problem Statement
Let $G_n = \text{GL}_n(F)$ be the general linear group over a non-Archimedean local field $F$ . The paper addresses the quotient branching law problem: given an irreducible smooth representation $\pi$ of $G_{n+1}$ and an irreducible smooth representation $\pi'$ of $G_n$ , determine when the space of homomorphisms $\text{Hom}_{G_n}(\pi, \pi')$ is non-zero.

While the multiplicity-one theorem (established by Aizenbud-Gourevitch-Rallis-Schiffmann and others) asserts that $\dim \text{Hom}_{G_n}(\pr, \pi') \leq 1$ , the central difficulty lies in determining the precise condition under which this dimension is exactly one. Previous results had solved specific cases (e.g., when $\pi$ is cuspidal, both are generic, or $\pi'$ is trivial), but a unified, necessary, and sufficient condition for general irreducible representations was lacking. This work also addresses the problem of constructing simple quotients of Bernstein-Zelevinsky (BZ) derivatives.

2. Methodology and Framework
The paper employs a combination of representation-theoretic tools and combinatorial structures derived from the Zelevinsky classification:

Bernstein-Zelevinsky (BZ) Theory: The primary tool is the BZ filtration (both left and right), which decomposes representations upon restriction to subgroups. The author utilizes the geometric lemma to analyze the layers of these filtrations.
Derivatives and Integrals: The author extensively uses BZ derivatives ( $D_\Delta$ ) and integrals ( $I_\Delta$ ) associated with segments $\Delta$ . They introduce the concept of "strongly commutative triples" involving these operators.
Generalized GGP Relevance: The author formulates a new combinatorial condition called generalized GGP relevance. This generalizes the original Gan-Gross-Prasad (GGP) relevance condition (defined via Langlands parameters for Arthur-type representations) to all irreducible smooth representations.
Combinatorial Invariants: The paper relies heavily on $\eta$ -invariants (measuring the multiplicity of segments in derivatives) and the theory of minimal multisegments to characterize the structure of representations and their derivatives.
Deformation and Thickening: A key technique involves "thickening" representations (ensuring segments have relative length $\geq 2$ ) to simplify combinatorial arguments, followed by deformation techniques to extend results to the general case.
Rankin-Selberg Integrals: These are used to construct non-zero branching laws (existence proofs) by realizing representations via Whittaker models.

3. Key Contributions and Results

Main Theorem (Theorem 4.1 / Theorem 26.4 & 31.2): The paper establishes a necessary and sufficient condition for $\text{Hom}_{G_n}(\pi, \pi') \neq 0$ .
- Condition: The pair $(\pi, \pi')$ is non-zero if and only if they form a generalized GGP relevant pair.
- Definition of Relevance: $(\pi, \pi')$ $(π, π^{'})$ is relevant if there exist multisegments $m$ $m$ and $n$ $n$ such that:
  1. $D^R_m(\nu^{1/2} \cdot \pi) \cong D^L_n(\pi')$ (where $D^R$ and $D^L$ are right and left derivatives).
  2. $(m, n, \nu^{1/2} \cdot \pi)$ is a strongly RdLi-commutative triple (a condition defined via the compatibility of derivatives and integrals on the representation).
- Refinement: The paper determines the specific layer in the BZ filtration that supports the branching law (the "supporting layer"), providing a refined version of the branching law based on the length of the multisegment $m$ .
Exhaustion Theorem (Theorem 4.2 / Theorem 31.3): The author proves that any simple quotient of a BZ derivative of an irreducible representation can be constructed by a sequence of derivatives of essentially square-integrable representations. This result is significant for understanding the structure of BZ derivatives.
Symmetry and Uniqueness:
- Symmetry (Theorem 18.1): The relevance relation is symmetric: $(\pi, \pi')$ is relevant if and only if $(\pi', \pi)$ is relevant.
- Uniqueness (Theorem 16.2): For a relevant pair, the multisegments $m$ and $n$ satisfying the minimality conditions are unique. This allows for a well-defined "level" of relevance ( $i^*$ -relevance).
Generalized Pieri's Rule: By transferring the Exhaustion Theorem to the setting of affine Hecke algebras of type A, the paper derives a generalization of the classical Pieri's rule for symmetric groups.

4. Significance and Claims
The paper claims to resolve the quotient branching law problem for $p$ -adic general linear groups in full generality. Its significance is articulated as follows:

Unification: It unifies previously known disparate cases (cuspidal, generic, trivial, Steinberg, Arthur-type) under a single combinatorial framework.
Algorithmic Nature: The condition provides a finite, deterministic algorithm to check the branching law, relying on the Langlands correspondence and combinatorial checks of multisegments.
Extension of GGP: It extends the scope of the Gan-Gross-Prasad conjecture beyond Arthur-type representations to all irreducible smooth representations, addressing a gap noted in previous literature.
Structural Insight: The work provides a refined understanding of the BZ filtration layers contributing to branching laws, linking the "supporting layer" directly to the combinatorial data of the relevant pair.
Hecke Algebra Connection: The results offer a bridge between $p$ -adic representation theory and the combinatorics of affine Hecke algebras, generalizing classical symmetric group rules.

The author explicitly states that while the unitary branching law follows from the Kirillov conjecture, this work provides a direct, constructive solution for the $p$ -adic case using BZ theory, and that the techniques may shed light on branching laws for other classical groups (orthogonal, unitary) where the original GGP relevance is not sufficient. The paper does not claim to solve the problem for other groups but suggests the methodology is a potential avenue for future research in those areas.

More like this