🔢 mathematics

Middle convolution for Lie algebra representations

This paper introduces a Lie algebra analogue of the middle convolution functor that generalizes the Long-Moody functor, recovers the Dettweiler-Reiter and Haraoka convolutions as special cases, and establishes a Riemann-Hilbert correspondence between Lie algebra representations and local systems on hyperplane arrangement complements.

Original authors: Kazuki Hiroe

Published 2026-05-12

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Kazuki Hiroe

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 a master architect working with complex blueprints. These blueprints aren't for buildings, but for mathematical structures called "Lie algebras." These algebras are like rulebooks that describe how different mathematical objects interact, twist, and turn.

This paper, written by Kazuki Hiroe, introduces a new tool called the "Middle Convolution." Think of this tool as a magical machine that takes one specific blueprint (a representation of a Lie algebra) and transforms it into a new, often more interesting blueprint.

Here is a breakdown of what the paper does, using everyday analogies:

1. The Big Idea: A Universal Transformer

The author is building a "universal translator" for these mathematical rulebooks.

The Problem: Mathematicians already had similar machines for specific types of problems (like solving equations on a sphere or dealing with braids). But they didn't have a single machine that worked for the broad, abstract category of Lie algebras.
The Solution: Hiroe builds a "Middle Convolution" machine that works on a wide variety of Lie algebras, including Free Lie algebras (the most basic building blocks), Drinfeld-Kohno algebras (which describe how braids untangle), and Holonomy algebras (which describe the geometry of spaces with holes, like a donut or a room with pillars).

2. The "Long-Moody" Connection: The Parent-Child Analogy

The paper connects this new machine to an older, famous tool called the Long-Moody functor.

The Analogy: Imagine you have a family of braid groups (think of them as families of knotted ropes). The Long-Moody functor is like a way to take a representation (a description) of a larger family (say, a family with $n+1$ members) and shrink it down to create a representation for a smaller family (with $n$ members).
The Innovation: Hiroe shows that his new "Middle Convolution" is actually a deformed version of this shrinking process. It's like taking the Long-Moody machine, adding a "knob" (a parameter called $\lambda$ ), and turning it. When you turn the knob to zero, you get the original Long-Moody machine back. This proves the new tool is a natural, logical extension of the old one.

3. The "Dettweiler-Reiter" Connection: The Special Case

There is another famous machine called the Dettweiler-Reiter additive middle convolution, used for solving specific types of differential equations (Fuchsian systems).

The Analogy: Think of the Dettweiler-Reiter machine as a specialized screwdriver that only fits one specific type of screw.
The Claim: Hiroe demonstrates that his new Middle Convolution machine is the universal power drill. If you set the settings on his drill just right (by removing a specific part of the algebra), it behaves exactly like the old specialized screwdriver. This means his new tool doesn't just replace the old one; it encompasses it.

4. The Geometric Twist: The "Y-Closure" and the Room with Holes

The paper gets very geometric in the second half. It looks at hyperplane arrangements.

The Analogy: Imagine a room filled with invisible, flat walls (hyperplanes) cutting through space. The "complement" is the empty space left over where you can walk around.
The "Y-Closure": Sometimes, if you look at these walls from a specific angle (a line called $Y$ ), the room looks like a bundle of fibers. The author introduces a concept called the $Y$ -closure, which is like adding a few extra "ghost walls" to the room to make the geometry perfectly tidy and symmetrical.
The Result: The paper proves that the Middle Convolution machine works perfectly on the "Holonomy Lie algebra" (the rulebook for this room) and produces the exact same result as a machine built specifically for "logarithmic connections" (a way of measuring how things change as you move around the walls). It's like proving that two different maps of the same city lead to the same destination.

5. The Riemann-Hilbert Correspondence: The Rosetta Stone

Finally, the paper connects two different languages of mathematics:

Algebraic Language: Describing the room using Lie algebras (rules and equations).
Geometric Language: Describing the room using "local systems" (patterns of how things twist and turn as you walk around).

The Claim: The author builds a "Rosetta Stone" (a correspondence) between these two languages. He shows that if you take a rulebook, run it through his Middle Convolution machine, and then translate it into the geometric language, you get the exact same result as if you had translated the rulebook into geometry first and then run it through the geometric version of the Middle Convolution machine.

Summary

In simple terms, this paper says:

"We have built a new, powerful mathematical machine called the Middle Convolution. It is a generalization of several older, famous machines. It works on a wide variety of abstract rulebooks (Lie algebras). We proved that it behaves exactly like the older machines when you use them in their specific contexts, and we showed that it perfectly translates between the algebraic rules of these systems and the geometric shapes they describe."

The paper is a theoretical bridge, connecting different islands of mathematics (algebra, geometry, and topology) with a single, robust tool.

Technical Summary: Middle Convolution for Lie Algebra Representations

Problem Statement
The middle convolution functor, originally introduced by Katz for local systems on punctured Riemann spheres, has become a fundamental tool in the study of ordinary differential equations, rigid local systems, and the additive Deligne-Simpson problem. While algebraic versions of this functor have been established for Fuchsian systems (Dettweiler-Reiter) and logarithmic connections on hyperplane arrangement complements (Haraoka), a unified Lie algebraic framework encompassing these constructions was lacking. Specifically, there was a need to define a middle convolution functor directly on the category of modules over Lie algebras that serve as infinitesimal analogues of braid groups and fundamental groups of arrangement complements, such as free Lie algebras, Drinfeld-Kohno Lie algebras, and holonomy Lie algebras.

Methodology
The author constructs a "middle convolution functor" ( $mc_\lambda$ ) defined on the category of finite-dimensional modules over the universal enveloping algebra of a semidirect sum Lie algebra $L_n \rtimes_{\theta_g} \mathfrak{g}$ . Here, $L_n$ is a free Lie algebra, and $\mathfrak{g}$ is a Lie algebra acting on $L_n$ via a homomorphism $\theta_g$ derived from the infinitesimal braid action.

The construction proceeds in several steps:

Deformed Long-Moody Functor: The author first introduces a one-parameter deformation ( $LM^{df}_\lambda$ ) of the Long-Moody functor. While the classical Long-Moody functor constructs representations of braid groups from larger braid groups using augmentation ideals of group rings, its infinitesimal Lie algebra analogue is defined via the first homology group $H_1(L_{n+1}, V \otimes K_\lambda)$ , where $K_\lambda$ is a one-dimensional module defined by a scalar $\lambda$ .
Definition of Middle Convolution: The middle convolution functor $mc_\lambda$ is defined as the cokernel of a specific map involving the direct sum of homology groups $H_1(L(x_i), V \otimes K_\lambda)$ mapping into $H_1(L_{n+1}, V \otimes K_\lambda)$ . This construction effectively quotients out "trivial" submodules and invariant subspaces, analogous to the geometric middle convolution.
Generalization and Compatibility: The paper establishes that this Lie algebraic functor generalizes the Long-Moody functor (in the limit $\lambda \to 0$ ) and recovers the Dettweiler-Reiter additive middle convolution when $\mathfrak{g}$ is trivial.
Geometric Realization: For the holonomy Lie algebra $h(M(\mathcal{A}))$ of the complement of a hyperplane arrangement $\mathcal{A}$ , the author utilizes the equivalence between logarithmic Pfaffian systems and modules over the holonomy Lie algebra (Kohno's theorem). By considering the $Y$ -closure of an arrangement (where $Y$ is a line through the origin), the author demonstrates that the Lie algebraic middle convolution corresponds to Haraoka's middle convolution for logarithmic connections.
Riemann-Hilbert Correspondence: Finally, the paper combines the de Rham functor with the equivalence of categories to establish a correspondence between the Lie algebraic middle convolution and the topological middle convolution for local systems ( $MC_\chi$ ).

Key Contributions and Results

Definition of Lie Algebra Middle Convolution: The paper defines $mc_\lambda$ as an endofunctor on $U(L_n \rtimes_{\theta_g} \mathfrak{g})\text{-mod}$ . It is shown to be a natural generalization of the infinitesimal Long-Moody functor.
Recovery of Known Constructions:
- Long-Moody: The functor $mc_\lambda$ factors through the deformed Long-Moody functor $LM^{df}_\lambda$ . In the case where $\mathfrak{g} = P_n$ (Drinfeld-Kohno algebra) and $\lambda = 0$ , $mc_0$ recovers the original Long-Moody functor.
- Dettweiler-Reiter: When $\mathfrak{g} = \{0\}$ , the functor $mc_\lambda$ recovers the Dettweiler-Reiter additive middle convolution for Fuchsian systems.
- Haraoka: For the holonomy Lie algebra of a hyperplane arrangement complement, $mc_\lambda$ is compatible with Haraoka's middle convolution for logarithmic connections on the $Y$ -closure of the arrangement.
Composition Law: The paper proves that the middle convolution functor satisfies the composition law $mc_\lambda \circ mc_\mu \cong mc_{\lambda+\mu}$ and $mc_{-\lambda} \circ mc_\lambda \cong \text{id}$ for modules satisfying specific genericity conditions (irreducibility and non-resonance conditions on eigenvalues). This mirrors the fundamental properties of Katz's original functor.
Preservation of Irreducibility: It is shown that if a module $V$ is irreducible as a $U(L_n)$ -module, then $mc_\lambda(V)$ remains irreducible.
Riemann-Hilbert Correspondence: Theorem 5.2 establishes an isomorphism between the result of applying the Lie algebraic middle convolution (followed by the de Rham functor) and the result of applying the topological middle convolution for local systems ( $MC_\chi$ ), provided the character $\chi$ is related to $\lambda$ by $\chi(1) = \exp(2\pi i \lambda)$ and certain eigenvalue conditions are met.

Significance
The paper provides a unified algebraic framework for the middle convolution operation across different contexts: braid group representations, Fuchsian systems, and logarithmic connections on hyperplane arrangement complements. By formulating the operation directly on Lie algebra modules, it bridges the gap between the categorical study of Long-Moody functors and the analytic theory of rigid local systems. The establishment of the Riemann-Hilbert correspondence in this setting confirms that the Lie algebraic construction is the correct infinitesimal counterpart to the topological middle convolution, thereby extending the applicability of middle convolution techniques to the representation theory of holonomy Lie algebras and the study of differential equations with logarithmic singularities.