Abstract
We study a particular category ${\cal{C}}$ of $\mathfrak{gl}_{\infty}$-modules and a subcategory ${\cal{C}}_{int}$ of integrable $\mathfrak{gl}_{\infty}$-modules. As the main results, we classify the irreducible modules in these two categories and we show that every module in category ${\cal{C}}_{int}$ is semi-simple. Furthermore, we determine the decomposition of the tensor products of irreducible modules in category ${\cal{C}}_{int}$.
Introduction
In the study of Lie algebras, the transition from finite-dimensional structures to infinite-dimensional ones often introduces significant complexity. While the representation theory of $\mathfrak{gl}_n$ (the Lie algebra of $n \times n$ matrices) is well-understood and characterized by elegant results like Weyl's Theorem on complete reducibility, the infinite-dimensional counterpart $\mathfrak{gl}_{\infty}$—representing doubly infinite matrices with finitely many non-zero entries—is considerably more "wild."
Standard frameworks, such as Category $\mathcal{O}$, are frequently employed to manage highest-weight modules. However, these frameworks often fail to capture "natural" modules, such as the infinite-dimensional vector space $\mathbb{C}^\infty$ itself. This paper addresses this gap by defining a specific category $\mathcal{C}$ of $\mathfrak{gl}_{\infty}$-modules that preserves the desirable properties of finite-dimensional representations while accommodating the infinite nature of the algebra.
Defining Category $\mathcal{C}$
The core innovation of this work lies in the definition of category $\mathcal{C}$. A module $W$ belongs to $\mathcal{C}$ if it satisfies a "column-finiteness" condition: for any vector $w \in W$, the action of the matrix units $E_{m,n}w$ must be zero for all but a finite number of column indices $n$.
This constraint is the "secret sauce" of the research. It ensures that although the algebra is infinite, the action on any individual vector is effectively finite-dimensional. This allows the authors to bridge the gap between the infinite-dimensional algebra and more manageable finite-dimensional subalgebras.
Methodology and Classification
To analyze this category, the authors employ several sophisticated algebraic techniques:
- Triangular Decomposition: For any finite subset $S$ of integers, the authors associate a triangular decomposition. This allows them to partition the infinite algebra into a finite-dimensional core, denoted $\mathfrak{gl}_S$, and two abelian "tails."
- Generalized Verma Modules: The authors construct modules by "inducing" them from irreducible representations of the finite-dimensional subalgebra $\mathfrak{gl}_S$. By taking the unique irreducible quotient of these induced modules, they derive the fundamental building blocks of the category.
- Weight Space Analysis: By utilizing $S$-singular vectors (vectors annihilated by specific portions of the algebra), the authors uniquely identify and classify the irreducible modules within $\mathcal{C}$.
The primary classification result states that every irreducible module in $\mathcal{C}$ is isomorphic to a quotient $L(S, U)$, where $S$ is a finite set of indices and $U$ is an irreducible module for the finite-dimensional algebra $\mathfrak{gl}_S$.
Integrability and Semi-simplicity
A major focus of the paper is the subcategory $\mathcal{C}_{int}$ of integrable modules. The authors provide a precise criterion for integrability: the "highest weight" $\lambda$ must consist of non-negative integers in a non-increasing sequence, corresponding to dominant integral weights in the finite-dimensional theory.
By decomposing the action of $\mathfrak{gl}_{\infty}$ into numerous copies of $\mathfrak{sl}_2$—whose representation theory is perfectly understood—the authors prove a significant result: every integrable module in category $\mathcal{C}$ is semi-simple. In many infinite-dimensional contexts, modules can be "tangled" (indecomposable but not irreducible). In $\mathcal{C}_{int}$, however, they always resolve into a clean direct sum of irreducible components.
Tensor Product Decomposition
The paper further explores how these modules interact through tensor products. In infinite-dimensional settings, tensor products often result in infinite sums or intractable structures. However, the authors prove that the tensor product of any two irreducible modules in $\mathcal{C}_{int}$ is a finite sum of irreducible modules.
This result demonstrates a "stabilization" effect: if one considers a sufficiently large finite subset of indices, the interaction between two infinite-dimensional modules becomes identical to the interaction between their finite-dimensional $\mathfrak{gl}_n$ counterparts. This mirrors the classical Littlewood-Richardson rules found in finite-dimensional theory.
Conclusion and Future Directions
While category $\mathcal{C}$ provides a robust and "clean" framework for $\mathfrak{gl}_{\infty}$, it intentionally excludes non-trivial highest-weight and lowest-weight modules. Future research may involve synthesizing these results with other categorical approaches, such as those by Penkov and Serganova, which focus on integrable modules with finite-dimensional weight spaces.
Additionally, the authors note that this work was motivated by the study of quantum vertex algebras. Having established the representation-theoretic foundation for $\mathfrak{gl}_{\infty}$ within this category, the path is now open for explicit applications in the construction and analysis of vertex operator algebras.
Summary updated