Finite Length Subrepresentation

• Finite Length Subrepresentation is a concept where a submodule possesses a finite composition series with irreducible successive quotients.
• It plays a crucial role in automorphic, p-adic, and smooth representation theories, underpinning key results in the Langlands program and Harish–Chandra module analysis.
• Explicit construction methods such as patching, diagram techniques, and algebraic decomposition provide precise filtration and multiplicity control for these representations.

A finite length subrepresentation is a submodule of a representation possessing a composition series of finite length, i.e., it admits a finite filtration whose successive quotients are irreducible. This property arises prominently in the study of automorphic, $p$-adic, and smooth representations of reductive groups over local and global fields, with structural implications in the Langlands program, the analysis of Harish–Chandra modules, and the geometry of spherical spaces. In these contexts, finite length controls the complexity and decomposability of representations, ensuring manageable behavior in cohomological and module-theoretic problems.

1. Definitions and General Background

A smooth or locally analytic representation $\pi$ of a (pro-)finite or reductive $p$-adic group $G$ over a field $k$ is said to be of finite length if there exists a filtration

$$0 = \pi_0 \subset \pi_1 \subset \dots \subset \pi_n = \pi$$

such that each successive quotient $\pi_i/\pi_{i-1}$ is irreducible. The length of $\pi$, denoted $\operatorname{length}_G(\pi)$, is the maximal $n$ for which such a composition series exists. A subrepresentation (submodule) $\pi' \subset \pi$ is a finite length subrepresentation if it itself admits such a finite filtration.

Harish–Chandra modules are $(\mathfrak{g},K)$-modules (for a real semisimple Lie group $G$) underlying admissible representations and equipped with moderate growth conditions. In spherical and automorphic settings, finite length subrepresentations arise within spaces of $C^\infty(G/H)$, completed cohomology of Shimura varieties, and patched modules linked to congruences of Hecke eigenforms (Krötz et al., 2013, Breuil et al., 13 Dec 2025, Bertoletti, 26 May 2025, Breuil et al., 7 Jan 2025).

2. Spherical Spaces and Harish–Chandra Modules

Let $G$ be a connected real semisimple Lie group, $H$ a closed subgroup such that the minimal parabolic subgroup $P$ of $G$ satisfies $P \cdot H$ open in $G$, so $Z=G/H$ is a real spherical space (Krötz et al., 2013). Fix a Harish–Chandra module $V$. The space $$V^{-\infty,H} = \mathrm{Hom}_H(V^\infty, \mathbb{C})$$ of $H$-fixed distribution vectors plays a central role; dimension bounds for this space yield direct control over the structure and complexity of subrepresentations.

A principal result is the finite multiplicity theorem: there exists an explicit subalgebra $\mathfrak{q}_1 \subset \mathfrak{q}$ (built via intersections and sums involving $\mathfrak{h}$, $\mathfrak{n}$, $\mathfrak{m}$, etc.) such that $V/\mathfrak{q}_1 V$ is finite-dimensional and

$$\dim V^{-\infty,H} \leq \dim (V/\mathfrak{q}_1 V)^{L \cap K \cap H}.$$

Consequently, any $H$-spherical irreducible $V$ admits an embedding into an induced module $\overline{\operatorname{Ind}}_Q^G \tau$, where $Q=L U$ and $\tau$ is a finite-dimensional irreducible $L$-representation. Standard arguments using Knapp–Stein operators and Langlands classification yield that Harish–Chandra submodules of $C^\infty(G/H)$ have finite Jordan–Hölder length and decompose with finite multiplicity (Krötz et al., 2013).

3. Finite Length in Smooth and Locally Analytic $p$-adic Representation Theory

Completed Cohomology and Crystalline Representations

Consider $$r: \mathrm{Gal}(\overline{\mathbb{Q}}_p/\mathbb{Q}_p) \to \mathrm{GL}_n(E)$$ a crystalline $p$-adic Galois representation associated to a definite unitary group $G$ which at $p$ splits as $\mathrm{GL}_n(K)$. In the completed cohomology space $\widehat S_{\xi, \tau}(U^p, E)$, the locally analytic Hecke eigenspace $\Pi_\mathrm{an}$ attached to $r$ admits an explicit finite length subrepresentation $\Pi_\mathrm{fin}$, canonically constructed as a universal extension of the locally algebraic constituent $\pi_\mathrm{alg}(D) \otimes \epsilon^{n-1}$ by a direct sum of locally analytic “companion” representations $C(I, s_{|I|,\sigma}) \otimes \epsilon^{n-1}$, indexed by eigenvalue data and Hodge filtration steps. The tautological nature of $\Pi_\mathrm{fin}$, and its explicit filtration by irreducible and non-split analytic blocks, ensure that $\Pi_\mathrm{fin}$ determines and is determined by the filtered Frobenius module $D = D_{\mathrm{cris}}(r)$ (Breuil et al., 13 Dec 2025).

Mod $p$ Representations of $\mathrm{GL}_2(K)$

For $G = \mathrm{GL}_2(K)$ ($K$ a finite unramified extension of $\mathbb{Q}_p$), under mild genericity assumptions, the smooth representations $\pi$ arising in towers of Hecke eigenspaces of the mod $p$ étale cohomology of Shimura curves possess finite length. The length is bounded above by $r(f+1)$, where $r$ is a socle multiplicity, and $f = [K:\mathbb{Q}_p]$ (Bertoletti, 26 May 2025). Even in the absence of multiplicity one at the tame level, Jordan–Hölder constituents and block structures are fully controlled. Exactness properties of socle and congruence subgroups (Iwahori, $K_1$-invariants), and Cohen–Macaulay duality of module duals, yield a fine classification of explicit subrepresentations and subquotients.

In the minimal global situation with multiplicity one ($r=1$), each smooth admissible $\mathrm{GL}_2(K)$-representation attached to an absolutely irreducible Galois representation decomposes with explicit bounds. In the irreducible case, $\pi$ is irreducible supersingular (length one); in split and non-split cases, $\pi$ is a direct sum (or uniserial extension) of principal series and supersingular blocks, with overall length at most $f+1$ (Breuil et al., 7 Jan 2025).

4. Explicit Construction and Internal Structure

The finite length property of subrepresentations is often established via explicit module-theoretic and homological techniques:

• Patching and Eigenvariety Methods: Constructing $\Pi_\mathrm{fin}$ by patching completed cohomology and trianguline local models, injecting the resulting finite length subspace into the full locally analytic representation using smoothness properties at classical and companion points (Breuil et al., 13 Dec 2025).
• Diagram and Filtration Techniques: For $\mathrm{GL}_2(K)$, subrepresentations are classified via tuples of linear subspaces attached to blocks in a diagrammatic model, and every subquotient is shown to be generated by its $K_1$-invariants. Socles and Iwahori-invariants admit exactness properties, and the entire filtration structure is encoded by combinatorial data (Bertoletti, 26 May 2025, Breuil et al., 7 Jan 2025).
• Algebraic Decomposition: In real spherical settings, algebraic decomposition formulas for universal enveloping operators allow reduction to finite systems governing the space $V/\mathfrak{q}_1 V$, leading to concrete dimension bounds and subrepresentation theorems (Krötz et al., 2013).

5. Consequences and Applications

The finite length subrepresentation property effects several structural consequences:

• Submodules of $C^\infty(G/H)$ with Harish–Chandra structure are necessarily Artinian and Noetherian, decomposing with finite multiplicities into irreducible constituents. The same holds for categories generated by H-spherical modules (Krötz et al., 2013).
• In the $p$-adic analytic and mod $p$ settings, all subrepresentations admit finite Jordan–Hölder filtrations, which forms a foundation for the mod $p$ local Langlands correspondence for $\mathrm{GL}_n(K)$. When tame multiplicity one fails, the theory still persists by managing arbitrary multiplicities and classifying submodules combinatorially (Bertoletti, 26 May 2025).
• Explicit construction of finite-length locally analytic subrepresentations attached to Galois representations provides a canonical bridge between Hodge filtration and automorphic representation theory, enabling precise extraction of arithmetic data from representation-theoretic invariants (Breuil et al., 13 Dec 2025).

6. Case Studies and Examples

A prototypical example arises for $G = \mathrm{SL}(2,\mathbb{R})$, $H = AN$, where the real spherical setting is trivial, and each irreducible Harish–Chandra module $V$ is either length one (irreducible principal and discrete series) or two (reducible principal series), with dimension of $H$-fixed distribution vectors always equal to one for spherical representations (Krötz et al., 2013).

In the $p$-adic and automorphic setting for $G = \mathrm{GL}_2(K)$, each submodule in the étale cohomology Hecke eigenspaces or completed cohomology local factors is “built” via explicit choices of tuples of linear subspaces, with length and block structure governed by the degree $f$ and multiplicity $r$. Jordan–Hölder constituents are supersingular or principal series, and socle functors enjoy exactness on all subobjects (Bertoletti, 26 May 2025, Breuil et al., 7 Jan 2025).

7. Key Lemmas, Formulas, and Analytic Tools

Several analytic and algebraic ingredients are essential in proving the finite length property:

• ODE Asymptotics and Dimension Bounds: Solutions of regular singular ODEs on polydiscs yield expansions governing decay properties of matrix coefficients, which force finite bounds under growth estimates (Krötz et al., 2013).
• Cohen–Macaulay Duality: Cohen–Macaulay property of module duals ensures purity and self-duality, controlling socle and cycle structure in graded modules (Bertoletti, 26 May 2025, Breuil et al., 7 Jan 2025).
• Exactness Properties: For $n \leq 3$ (or $f+1$), socle and congruence subgroup-invariants functors are exact and split, enabling fine control over subrepresentation generation and filtration (Bertoletti, 26 May 2025, Breuil et al., 7 Jan 2025).

A plausible implication is that in all contexts in which explicit functorial or diagrammatic control exists together with appropriate genericity conditions, the finite length subrepresentation phenomenon will persist, ensuring manageability and decomposability of automorphic and smooth representations for deeper arithmetic and geometric applications.