[Paper Review] Iterated extensions in module categories

Eivind Eriksen|ArXiv.org|Jun 2, 2004

Algebraic structures and combinatorial models12 references3 citations

TL;DR

This paper introduces iterated extensions as a framework to study length categories in module categories over associative algebras over an algebraically closed field. By analyzing extension types and noncommutative deformations, it provides a constructive proof of the characterization of uniserial length categories and applies the method to classify indecomposable graded holonomic D-modules on curves, showing that such categories are uniserial and explicitly parametrized by order and weight data.

ABSTRACT

Let k be an algebraically closed field, let R be an associative k-algebra, and let F = {M_a: a in I} be a family of orthogonal points in R-Mod such that End_R(M_a) = k for all a in I. Then Mod(F), the minimal full sub-category of R-Mod which contains F and is closed under extensions, is a full exact Abelian subcategory of R-Mod and a length category in the sense of Gabriel. In this paper, we use iterated extensions to relate the length category Mod(F) to noncommutative deformations of modules, and use some new methods to study Mod(F) via iterated extensions. In particular, we give a new proof of the characterization of uniserial length categories, which is constructive. As an application, we give an explicit description of some categories of holonomic and regular holonomic D-modules on curves which are uniserial length categories.

Motivation & Objective

• To develop a new method based on iterated extensions to analyze the structure of length categories in module categories.
• To provide a constructive proof of the characterization of uniserial length categories, resolving a long-standing problem in representation theory.
• To classify indecomposable holonomic D-modules on curves by relating them to noncommutative deformations of simple modules.
• To determine conditions under which such module categories are wild, particularly in relation to the presence of a specific quiver substructure.
• To extend the theory to graded modules over differential operator rings, especially for affine monomial curves.

Proposed method

• The paper defines the category $\mathbf{Ext}(\mathcal{F})$ of iterated extensions, where objects are modules equipped with a cofiltration by objects from a family $\mathcal{F}$ of orthogonal simple modules.
• It introduces the extension type as an ordered quiver encoding the sequence of simple modules in a filtration, capturing the structure of iterated extensions.
• The method uses noncommutative deformations of modules to classify the isomorphism classes of extensions, particularly in the context of the first Weyl algebra and differential operator rings.
• It applies the theory to graded holonomic D-modules by showing that the family of simple graded D-modules forms a family of orthogonal points with a $k$-quiver satisfying a specific condition, leading to uniserial structure.
• The paper uses the concept of socle-height and full exact embeddings into $\mathbf{fdMod}(W)$, the category of finite-dimensional modules over the free algebra $k\langle x,y\rangle$, to detect tameness or wildness.
• It proves that if the Gabriel quiver of $\mathbf{Mod}(\mathcal{F})$ contains the quiver $Q_5$, then $\mathbf{Mod}(\mathcal{F})$ is wild, via a full exact embedding from $\mathbf{fdMod}(W)$.

Experimental results

Research questions

• RQ1 Under what conditions is the category $\mathbf{Mod}(\mathcal{F})$ of iterated extensions of a family of orthogonal simple modules a uniserial length category?
• RQ2 How can iterated extensions and noncommutative deformations be used to constructively classify indecomposable modules in such categories?
• RQ3 What is the structure of the category of graded holonomic $D$-modules on curves, and is it tame or wild?
• RQ4 When does the presence of a specific quiver $Q_5$ in the Gabriel quiver imply that $\mathbf{Mod}(\mathcal{F})$ is wild?
• RQ5 How do the results extend to graded modules over differential operator rings of affine monomial curves?

Key findings

• The category $\mathbf{Mod}(\mathcal{F})$ is a full exact Abelian subcategory and a length category when $\mathcal{F}$ is a family of orthogonal simple modules with $\operatorname{End}_R(M_\alpha) \cong k$.
• A constructive proof is given for the characterization of uniserial length categories using iterated extensions and extension types.
• The category of graded holonomic $D$-modules on curves is uniserial, and each indecomposable object is explicitly described as $M(\alpha,n) = D/Dw(\alpha,n)$ for $\alpha = 0,1$ or $D/D(E - \alpha)^n$ for $\alpha \in I \setminus \{0,1\}$.
• For the first Weyl algebra $A_1(k)$ over an algebraically closed field of characteristic zero, the category of graded holonomic $D$-modules is uniserial and all indecomposable modules are parametrized by $\{M(\alpha,n) : \alpha \in I, n \geq 1\}$.
• If the Gabriel quiver of $\mathbf{Mod}(\mathcal{F})$ contains the quiver $Q_5$, then $\mathbf{Mod}(\mathcal{F})$ admits a full exact embedding of $\mathbf{fdMod}(W)$, making it wild in the sense of Klingler and Levy.
• The result extends to $D'$-modules on affine monomial curves, as their holonomic categories are equivalent to those on $A_1(k)$, preserving the uniserial structure and classification.