Higher operad structure for Fukaya categories

Here is an explanation of Hang Yuan's paper, "Higher Operad Structure for Fukaya Categories," translated into simple, everyday language using creative analogies.

The Big Picture: Building a Universal Lego Kit

Imagine you are an architect trying to describe how to build complex structures. In mathematics, specifically in the fields of symplectic geometry (which studies shapes in space) and algebra, there are many different types of "instruction manuals" for building things.

For a long time, mathematicians had a very popular instruction manual called an Operad. Think of an operad as a specific type of Lego kit where you have a pile of bricks, and you can snap them together in a specific way to make a tower. This kit was perfect for describing one specific kind of mathematical structure called an $A_\infty$ -algebra.

However, in the world of symplectic geometry (specifically in something called Fukaya categories), the "bricks" are much more complicated. They aren't just single towers; they are entire cities with roads, intersections, and different types of buildings. The old Lego kit (the standard operad) was too simple. It couldn't capture the complexity of these cities.

Hang Yuan's paper introduces a new, super-charged instruction manual called an "fc-multicategory."

Think of an fc-multicategory as a 2D Lego Kit.

Standard Operad (1D): You have a stack of bricks. You put one on top of another. It's a straight line.
fc-multicategory (2D): You have a flat surface. You can have "roads" (horizontal lines) and "buildings" (vertical lines), and you can place "tiles" (operations) that connect them. You can paste these tiles together in complex, grid-like patterns.

This new kit is flexible enough to describe not just simple towers, but entire cities, intersections, and even curved roads.

The Problem: The Geometry is Messy

In symplectic geometry, mathematicians study pseudo-holomorphic polygons.

The Analogy: Imagine you are drawing shapes on a rubber sheet (a surface) that is stretched over a bumpy landscape. You have to draw a shape (a polygon) such that its edges stick to specific lines (Lagrangian submanifolds) drawn on the landscape.
The Issue: Sometimes the landscape has self-intersections (the lines cross over themselves). When your shape hits these crossing points, it creates "corners" or "kinks."
The Old Way: Mathematicians used to try to flatten these kinks out and pretend the lines were simple. They would count the shapes and write down a list of rules (formulas) for how they fit together. But this often threw away important information, like where exactly the shape touched the crossing lines or which path it took around a bump.

Yuan's Insight: Instead of flattening the problem, let's embrace the messiness. Let's treat the crossing lines as a map with different "zones" (connected components). Let's treat the shapes as tiles that fit into a grid defined by these zones.

The Solution: The "fc-multicategory" Map

Yuan shows that the collection of all these weird, kinked shapes naturally forms a structure called an fc-multicategory.

Here is how the analogy works:

The 0-Cells (The Cities): These are the different connected pieces of the landscape (the Lagrangian submanifolds).
The 1-Cells (The Roads): These are the points where the landscape pieces cross each other.
The 2-Cells (The Tiles): These are the pseudo-holomorphic polygons themselves. They are the "glue" that connects the roads.

The magic is that you can glue these tiles together. If you have a shape that goes from City A to City B, and another from City B to City C, you can paste them together to make a shape from A to C. This "pasting" operation is exactly what the fc-multicategory structure describes.

The "Shrinking" Trick: From Geometry to Algebra

The paper's second major goal is to turn this geometric map into an algebraic one.

The Analogy: Imagine you have a giant, detailed map of a city with every street, park, and building drawn in high definition. This is the geometric moduli space. It's beautiful, but too big to carry in your pocket.
The Trick: Yuan proposes "shrinking" every single street and building down to a single point.
- If you shrink a whole neighborhood to a dot, you lose the details of the streets, but you keep the connections.
- If you do this to the fc-multicategory, you get a Differential Graded (dg) fc-multicategory.

This "shrunk" version is the algebraic instruction manual. It tells you exactly how to combine your mathematical "bricks" (operations) without needing to draw the whole city map every time.

Why This Matters: One Kit for All Structures

Before this paper, if you wanted to study:

An $A_\infty$ -algebra (a simple tower),
An $A_\infty$ -category (a city with many buildings),
An $A_\infty$ -bimodule (a bridge between two cities),
Or an $A_\infty$ -module (a side-structure attached to a city),

You had to write down a different set of rules for each one. It was like having a different instruction manual for a tower, a different one for a house, and another for a bridge.

Yuan's paper says: "No, you only need one master kit."

By using the fc-multicategory framework:

An algebra is just a kit with one city.
A category is a kit with many cities.
A bimodule is a kit with two cities and a bridge.

They are all just different ways of using the same underlying "pasting" rules.

The "Curved" Twist

In symplectic geometry, sometimes the shapes don't close perfectly; they have a little "curvature" or energy leak. In algebra, this is called a curved structure.

The Old View: Curved structures were messy and hard to fit into the standard Lego kits.
Yuan's View: The fc-multicategory is so flexible that it can naturally handle these "curved" tiles. It can label them with extra information (like "energy level" or "homotopy class") so you don't lose track of where the leak happened.

Summary

Hang Yuan has built a universal translator.

Geometry: He took the messy, complex shapes of symplectic geometry (polygons on crossing lines) and organized them into a neat, 2D grid structure called an fc-multicategory.
Algebra: He showed how to "shrink" this grid into a powerful algebraic tool (a dg fc-multicategory).
Unification: He proved that almost every complex structure mathematicians use in this field (algebras, categories, modules, bimodules) is just a specific way of using this one universal tool.

In short: He took a pile of different, confusing instruction manuals and realized they were all just different chapters of the same, massive, super-flexible encyclopedia. This makes it much easier for mathematicians to understand how these different structures relate to one another and to the geometry that inspired them.

Here is a detailed technical summary of the paper "Higher operad structure for Fukaya categories" by Hang Yuan.

1. Problem Statement

The paper addresses two fundamental gaps in the algebraic formulation of symplectic geometry, specifically within the context of Fukaya categories and $A_\infty$ -structures:

Geometric Limitation of Standard Operads: While the standard $A_\infty$ -operad (derived from Stasheff associahedra) successfully encodes $A_\infty$ -algebras, it is insufficient for capturing the full complexity of moduli spaces of pseudo-holomorphic curves with Lagrangian boundary conditions. Specifically, standard operads assume a single object (or a fixed set of colors) and do not naturally accommodate:
- Multiple Lagrangian submanifolds (possibly immersed).
- The topological data of boundary paths connecting different Lagrangian components.
- The distinction between "Floer theory for a pair" and the "Fukaya category" of a set of objects.
- Curved $A_\infty$ -structures (where $m_0 \neq 0$ ) arising from non-exact Lagrangians.
Lack of Uniform Operadic Formulation for Variants: Variants such as $A_\infty$ -categories, $A_\infty$ -bimodules, and $A_\infty$ -multi-modules are typically defined via ad-hoc collections of multilinear operations and component-wise identities. This obscures the conceptual unity suggested by the operadic view of $A_\infty$ -algebras and discards useful geometric data (like the homotopy class of boundary loops).

Core Question: Can one extract a "higher" operadic structure from the moduli spaces of pseudo-holomorphic polygons that unifies these diverse $A_\infty$ -structures and retains their geometric origin?

2. Methodology

The author employs a synthesis of higher category theory and symplectic geometry:

fc-multicategories (Virtual Double Categories): The paper utilizes the framework of fc-multicategories (introduced by Leinster and Cruttwell/Shulman). Unlike standard operads (indexed by trees) or multicategories (indexed by lists), fc-multicategories are indexed by 2-dimensional pasting diagrams.
- They consist of 0-cells (objects), horizontal 1-cells (edges), vertical 1-cells (morphisms between objects), and 2-cells (operations).
- The "fc" stands for the free category monad, allowing the structure to encode the composition of paths (horizontal edges) and operations (2-cells) simultaneously.
Geometric Construction:
- The author constructs an fc-multicategory, denoted $\mathcal{M}_\iota$ , where the 0-cells are connected components of a Lagrangian immersion $\iota: L \to X$ , the horizontal 1-cells are intersection points (elements of $L \times_X L$ ), and the 2-cells are moduli spaces of pseudo-holomorphic polygons.
- Crucially, the boundary of a pseudo-holomorphic polygon is lifted to the domain $L$ via a map $\gamma$ . This allows the structure to track the specific path-connected components of $L$ traversed by the boundary, refining the standard relative homology class.
Differential Graded (dg) Variants:
- The paper develops a theory of dg fc-multicategories, where the 2-cells form cochain complexes equipped with a differential satisfying a Leibniz rule with respect to composition.
- This allows the translation of geometric moduli spaces (which satisfy Gromov compactness and boundary relations) into algebraic structures where the $A_\infty$ relations arise as the condition $\delta^2 = 0$ .
Algebras over dg fc-multicategories:
- The author defines an algebra over a dg fc-multicategory as a morphism from the fc-multicategory to an endomorphism fc-multicategory of a cochain complex (or a collection of complexes).
- This generalizes the standard definition of an $A_\infty$ -algebra as a morphism $A_\infty \to \text{End}(X)$ .

3. Key Contributions

A. The fc-multicategory Structure on Moduli Spaces (Theorem 1.2)

The paper establishes that the collection of moduli spaces of pseudo-holomorphic polygons bounded by a Lagrangian immersion $\iota: L \to X$ naturally forms a topological fc-multicategory $\mathcal{M}_\iota$ .

0-cells: Connected components of $L$ ( $V_\iota = \pi_0(L)$ ).
Horizontal 1-cells: Points in the fiber product $L \times_X L$ (intersection points).
2-cells: Isomorphism classes of pseudo-holomorphic polygons with boundary paths $\gamma$ in $L$ .
Significance: This structure captures the "gluing" of polygons (composition of 2-cells) which corresponds to the gluing of pseudo-holomorphic curves in symplectic geometry. It naturally handles multiple Lagrangians and immersed cases, which standard multicategories cannot do without artificial labeling.

B. Refined Labeling via Mapping Cones (Section 4.3)

The author introduces a refined labeling system $S_\iota$ based on the mapping cone of the immersion $\iota$ .

Instead of just using the relative homology class $[u] \in H_2(X, \iota(L))$ , the labeling records the boundary paths $\gamma_j$ as 1-chains in $C_1(L)$ .
This provides a finer invariant than standard relative homology, essential for defining curved $A_\infty$ -structures and handling the "divisor axiom" in applications.

C. Uniform Algebraic Formulation (Theorem 1.5)

The paper proves that a wide range of $A_\infty$ -type structures are simply algebras over specific dg fc-multicategories. This unifies the following under a single framework:

$A_\infty$ -algebras: Recovered as algebras over a dg fc-multicategory with a single 0-cell (reducing to the standard $A_\infty$ -operad).
$A_\infty$ -categories: Recovered as algebras over a dg fc-multicategory where the underlying graph is $V \times V$ .
$A_\infty$ -modules (Left/Right): Recovered by extending the graph to include a "source" or "sink" vertex.
$A_\infty$ -bimodules: Recovered as algebras over a dg fc-multicategory with two distinct loops and a connecting edge.
$A_\infty$ -multi-modules: Generalized to arbitrary partitions of the object set.

D. Factor-Closedness and Gromov Compactness (Section 4.4)

The paper demonstrates that the factor-closedness condition for fc-submulticategories (Definition 3.6) is the operadic reflection of Gromov compactness.

If a subcollection of moduli spaces is closed under Gromov limits (a standard requirement for defining invariants), it corresponds to a factor-closed sub-fc-multicategory.
This provides a rigorous categorical justification for why certain sub-collections of Lagrangians (e.g., finite subsets) yield well-defined algebraic structures.

4. Results

Theorem 1.2: The moduli spaces of pseudo-holomorphic polygons bounded by a Lagrangian immersion form a topological fc-multicategory.
Theorem 1.5: There exist dg fc-multicategories whose algebras recover $A_\infty$ -algebras, $A_\infty$ -categories, $A_\infty$ -modules, and $A_\infty$ -bimodules.
Proposition 4.5: The moduli space fc-multicategory $\mathcal{M}_\iota$ admits a natural labeling by the fc-multicategory $S_\iota$ derived from the mapping cone, refining the homological data.
Proposition 4.6 & 4.7: Specific sub-collections of moduli spaces (e.g., those involving a subset of Lagrangians or specific partitions) form factor-closed fc-submulticategories, ensuring the consistency of algebraic operations under Gromov limits.
Section 6: Explicit constructions showing how the standard $A_\infty$ relations (e.g., $\sum (-1)^{\dots} m_{r+1+t} \circ m_s = 0$ ) arise directly from the condition $\delta^2 = 0$ in the dg fc-multicategory setting.

5. Significance

Conceptual Unification: The paper resolves the fragmentation in the literature where $A_\infty$ -algebras, categories, and modules are treated as distinct, ad-hoc definitions. It shows they are all instances of the same categorical concept: algebras over dg fc-multicategories.
Geometric Fidelity: By using fc-multicategories, the framework retains the geometric data of boundary paths and intersection components that are often lost in purely algebraic formulations. This is crucial for applications involving curved $A_\infty$ -algebras and the divisor axiom in mirror symmetry.
Flexibility for Immersed Lagrangians: The framework naturally handles Lagrangian immersions (self-intersections), which are increasingly important in modern symplectic topology (e.g., in the study of exact Lagrangian fillings and wall-crossing phenomena).
Foundation for Future Work: By establishing the link between Gromov compactness and factor-closedness, the paper provides a robust categorical language for constructing Fukaya categories of infinite collections of Lagrangians via homotopy colimits.

In summary, Hang Yuan's work elevates the operadic perspective from a tool for $A_\infty$ -algebras to a comprehensive higher operadic framework capable of encoding the full geometric and algebraic complexity of Fukaya categories.