The Diagrammatic Spherical Category

Here is an explanation of the paper "The Diagrammatic Spherical Category" by Tasman Fell, translated into everyday language with creative analogies.

The Big Picture: Solving a Cosmic Puzzle

Imagine you are trying to understand the behavior of a massive, complex machine (a mathematical object called a reductive algebraic group). This machine has thousands of gears, and you want to know exactly how each gear spins. In math, these "spins" are called characters.

For decades, mathematicians had a map to find these spins, called Lusztig's Conjecture. It was like a GPS that worked perfectly in sunny weather (large prime numbers). But recently, someone found that the GPS breaks down in the rain (small prime numbers). The old map was wrong.

The new map requires a different kind of navigation tool. Instead of standard coordinates, we need a special, flexible tool called the Spherical Module. The problem? This tool is incredibly hard to build using traditional algebraic bricks. It's like trying to build a skyscraper out of wet clay; it's messy and hard to calculate.

Tasman Fell's paper says: "Let's stop using wet clay. Let's build this skyscraper out of LEGOs."

The LEGO Analogy: Diagrammatic Categories

Fell introduces a new way to build these mathematical objects using diagrams. Instead of writing long, confusing equations, you draw pictures:

Strings represent the building blocks.
Dots and junctions (where strings meet) represent operations.
Walls represent special boundaries in the system.

This is called a Diagrammatic Category. It's like a visual programming language. You can slide these strings around (isotopy), and as long as you follow the local rules (like "two strings crossing must equal three strings crossing"), the math works out. This is much easier to compute with than the "wet clay" of traditional algebra.

The Three Main Achievements

The paper accomplishes three major things, which we can think of as three steps in a construction project:

1. The "Double-Leaf" Blueprint (The Basis)

The Problem: You have a box of LEGO pieces (morphisms), but you don't know which specific combination of pieces creates a unique structure. You might have 100 different ways to build a tower, but are they all unique, or are some just copies of others?

The Solution: Fell creates a specific set of "Double-Leaf" diagrams.

Imagine a Light-Leaf as a single path from the bottom of your diagram to the top, following a specific set of rules (like a hiking trail).
A Double-Leaf is simply taking one path going up and one path going down and gluing them together.
The Breakthrough: Fell proves that if you take all possible combinations of these Double-Leaves, you get a perfect set of unique building blocks. You don't need any other pieces to build the structure, and none of these pieces are redundant. It's like finding the "atomic" set of LEGO instructions that can build anything in the set without waste.

2. The Translation Guide (Categorification)

The Problem: We built our LEGO tower, but does it actually match the "wet clay" skyscraper we were trying to model? Does our visual map correspond to the real mathematical object?

The Solution: The paper proves that the LEGO tower (the Diagrammatic Category) and the Skyscraper (the Algebraic Spherical Category) are identical twins.

They have the same number of rooms.
They have the same connections.
If you take the LEGO tower apart and count the pieces, it matches the algebraic formula perfectly.
This means we can now use the easy LEGO diagrams to solve the hard algebraic problems.

3. The Universal Translator (Equivalence)

The Problem: There are different ways to build these structures. Some mathematicians use "Singular Soergel Bimodules" (a very technical algebraic method). Others use diagrams. Are they compatible?

The Solution: The paper shows that the Diagrammatic Category is not just similar to the Algebraic one; it is equivalent.

Think of it like having a dictionary between English and French. You can write a sentence in English (Diagrams), translate it to French (Algebra), and get the exact same meaning.
This allows mathematicians to switch between the visual, easy-to-compute diagrams and the rigorous, established algebra whenever it's convenient.

Why "Spherical" and "Walls"?

You might wonder about the name "Spherical." In this context, it doesn't mean a ball. It refers to a specific type of symmetry where you are looking at the system from a specific angle (defined by a subset of generators $J$ ).

The "Wall" in the diagrams is a crucial innovation.

Imagine you are building a tower, but on the left side, there is a solid wall.
Some of your strings (representing specific mathematical operations) can "plug into" this wall.
This wall acts like a filter or a boundary condition. It forces the math to behave in a specific way that matches the "Spherical Module." Without this wall, you'd just be building a standard tower (the Hecke category). With the wall, you build the special "Spherical" tower.

The "Light-Leaf" Algorithm

How did he find the perfect Double-Leaf blocks? He used an algorithm called Light-Leaves.

Imagine you are walking through a forest (the mathematical space).
You have a map that tells you which path to take based on the terrain (the Bruhat order).
Sometimes you go up a hill ( $U$ ), sometimes down ( $D$ ), and sometimes you hit a dead end and have to turn around ( $X$ ).
The "Light-Leaf" is the specific path you draw based on these rules.
By gluing a path going up with a path going down, you create a "Double-Leaf."
The genius of the paper is proving that every possible way to move between two points in this mathematical forest can be broken down into a sum of these specific Double-Leaf paths.

The Takeaway

Before this paper, calculating the properties of these complex mathematical objects in "rainy weather" (characteristic $p$ ) was a nightmare. You had to use heavy, slow algebraic methods that often failed or were impossible to compute.

Tasman Fell's paper provides a new toolkit:

Visuals: It replaces heavy algebra with string diagrams (LEGOs).
Efficiency: It gives a precise list of "atomic" moves (Double-Leaves) that you can use to build anything.
Reliability: It proves that these visual moves are mathematically identical to the heavy algebraic ones.

In short, the paper turns a dark, foggy maze into a well-lit room with a clear set of instructions, allowing mathematicians to finally compute the characters of these groups correctly, even when the old maps failed.

Here is a detailed technical summary of the paper "The Diagrammatic Spherical Category" by Tasman Fell.

1. Problem Statement and Context

The Core Problem:
In the representation theory of reductive algebraic groups over fields of characteristic $p$ , determining the characters of simple modules is a central challenge. Lusztig's conjecture (1980) proposed that these characters could be computed using affine Kazhdan-Lusztig polynomials. However, Williamson (2017) provided counter-examples, showing the conjecture fails for certain $p$ . The correct formula requires $p$ -Kazhdan-Lusztig polynomials, which are significantly harder to compute as they lack a known recursive formula.

The Gap:
Recent work by Riche and Williamson established that simple characters can be computed using the $p$ -canonical basis of the spherical module $M(J)$ over the Hecke algebra. To compute this basis, one must perform object decompositions in a spherical category (a categorification of $M(J)$ ).
While the diagrammatic Hecke category (introduced by Elias and Williamson) successfully categorifies the Hecke algebra and is well-behaved in characteristic $p$ , a corresponding diagrammatic spherical category was missing. Existing algebraic models (Singular Soergel Bimodules) are difficult to compute with directly in characteristic $p$ .

Objective:
The paper aims to construct a diagrammatic spherical category, denoted $\mathcal{M}_{BS}(J)$ , which serves as a diagrammatic categorification of the spherical module $M(J)$ , analogous to how the diagrammatic Hecke category categorifies the Hecke algebra.

2. Methodology

The author employs a combination of diagrammatic algebra, localization techniques, and comparison with algebraic models.

A. Construction of the Diagrammatic Category ( $\mathcal{M}_{BS}(J)$ )

Objects: Indexed by expressions $x$ in the Coxeter system $(W, S)$ .
Morphisms: Colored string diagrams with a "wall" on the left. Strands corresponding to generators in a finitary subset $J \subset S$ can "plug into" this wall.
Relations: The category inherits relations from the standard diagrammatic Hecke category (Elias-Williamson) and adds specific "wall relations" governing how strands interact with the wall (e.g., removing a strand plugged into the wall via a Demazure operator).
Functor: A functor $\tilde{A}: \mathcal{M}_{BS}(J) \to J\text{BSBim}$ (Singular Bott-Samelson bimodules) is defined, mapping diagrams to algebraic bimodules.

B. The Double-Leaves Basis

The core technical innovation is the construction of a basis for morphism spaces, called the double-leaves basis ( $SDL_{x,y}$ ).

Spherical Light-Leaves: The author adapts the "light-leaves" algorithm (Elias-Williamson) to the spherical setting. Given an expression $x$ and a subexpression $e$ , a "coset stroll" is computed. Based on whether the stroll stays within the minimal coset representatives ( $JW$ ) or leaves it, specific diagrammatic moves (Up, Down, or "X" moves involving the wall) are applied to construct a map $SLL_{x,e}$ .
Double-Leaves: A morphism between objects $x$ and $y$ is constructed by composing a light-leaf $SLL_{x,e}$ with the inverted light-leaf $SLL_{y,f}$ , provided they meet at the same minimal coset representative $z$ (i.e., $W_J x e = W_J y f = W_J z$ ).
$SDL_{f,e} := SLL_{y,f} \circ SLL_{x,e}$

C. Localization and Linear Independence

To prove the double-leaves form a basis, the author uses localization (tensoring with the field of fractions $Q$ of the polynomial ring $R$ ).

The category is localized to $\mathcal{M}_{BS, Q}$ .
A commutative diagram is established linking the diagrammatic category, its Karoubi envelope, and the category of Standard Bimodules ( $J\text{StdBim}_Q$ ).
Key Insight: In the localized category, objects decompose into direct sums of "reflection indices." The author proves that the localized double-leaves act as upper-triangular matrices with respect to a specific partial order (path-dominance order on subexpressions) on the set of pairs of subexpressions.
Since the diagonal entries are non-zero, the set is linearly independent.

D. Spanning

The spanning property is proven by adapting the non-spherical proof from Elias-Williamson. The author shows that any diagram can be decomposed into a sum of "negative-positive" diagrams. Through a careful inductive argument involving "sweeps" and specific choices of non-spherical light-leaves that mimic spherical ones, any morphism is shown to lie in the span of the double-leaves.

3. Key Contributions and Results

Theorem 1.1: Basis for Morphism Spaces

The collection of spherical double-leaves $SDL_{x,y}$ forms a basis for $\text{Hom}_{\mathcal{M}_{BS}(J)}(x, y)$ as a right $R$ -module (where $R = \text{Sym}(\mathfrak{h}^*)$ ). This provides a concrete, computable basis for the diagrammatic category.

Theorem 1.2: Categorification of the Spherical Module

There is an isomorphism of $H$ -modules between the split Grothendieck group of the Karoubi envelope $\mathcal{M}(J) = \text{Kar}(\mathcal{M}_{BS}(J))$ and the spherical module $M(J)$ :
$\text{ch}: [\mathcal{M}(J)] \xrightarrow{\sim} M(J)$
This confirms that the constructed category correctly categorifies the spherical module.

Theorem 1.3: Equivalence to Algebraic Spherical Category

Assuming the realization $\mathfrak{h}$ is reflection-faithful, the diagrammatic spherical category $\mathcal{M}(J)$ is equivalent to the algebraic spherical category (the Karoubi envelope of Singular Bott-Samelson bimodules, $\text{Kar}(J\text{BSBim})$ ).

Furthermore, this equivalence respects the module category structure over the Hecke category (diagrammatic) and the Soergel bimodule category (algebraic).
This result bridges the gap between the computationally friendly diagrammatic approach and the rigorous algebraic definition.

Additional Result: Equivalence to Abe's Category

The paper also establishes an equivalence between $\mathcal{M}(J)$ and Abe's category of one-sided Soergel bimodules ( $J\text{Sbimod}$ ), which is well-defined even without the reflection-faithful assumption. This suggests the diagrammatic category is robust across different realizations.

4. Significance

Computational Tool for $p$ -Kazhdan-Lusztig Theory:
The construction provides a practical, diagrammatic framework for computing the $p$ -canonical basis of the spherical module. Since $p$ -Kazhdan-Lusztig polynomials are notoriously difficult to compute recursively, having a diagrammatic category where object decompositions can be performed (potentially via computer algorithms) is a major step forward in understanding modular representation theory.
Extension of Elias-Williamson Theory:
This work extends the powerful "Soergel Calculus" (diagrammatic Hecke categories) to the singular setting (spherical modules). It generalizes Elias's Type A construction to all Coxeter types, unifying the theory.
Resolution of Structural Questions:
By proving the equivalence between the diagrammatic and algebraic spherical categories (under reflection-faithfulness), the paper validates the diagrammatic approach as a rigorous alternative to algebraic bimodules. It confirms that the "wall" relations correctly capture the algebraic structure of singular Soergel bimodules.
Foundation for Future Work:
The "double-leaves" basis and the specific handling of the wall relations provide the necessary machinery to tackle the decomposition of objects in the spherical category, which is the prerequisite for calculating simple characters in characteristic $p$ via the Riche-Williamson formula.

In summary, Tasman Fell successfully constructs a diagrammatic model for the spherical module, proves it possesses a computable basis (double-leaves), and demonstrates its equivalence to established algebraic models, thereby providing a vital new tool for the study of modular representation theory.