On the topology of conormal complexes and posets of matroids

This paper introduces the biflats complex of a matroid and demonstrates through elementary collapses that it is simple homotopy equivalent to both the conormal complex and the simplicial join of the Bergman complexes of the matroid and its dual, while also analyzing their distinct combinatorial properties.

The Gist

Imagine you have a complex puzzle made of interlocking shapes. In the world of mathematics, specifically in a field called combinatorics, these puzzles are often built from objects called matroids. You can think of a matroid as a generalized way of describing "independence," like how some wires in a circuit can be removed without breaking the connection, or how some vectors in space don't add any new direction.

This paper, written by Anastasia Nathanson and Ethan Partida, is about exploring the hidden "shape" (topology) of these puzzles. They introduce a new way to look at these shapes and prove that three different-looking versions of the puzzle are actually the same shape underneath, just stretched or squashed in different ways.

Here is the breakdown using simple analogies:

1. The Three Characters in the Story

The paper focuses on three specific structures related to a matroid. Imagine them as three different maps of the same territory:

• The Bergman Complex (The Classic Map): This is the old, well-known map. It's built from the "flats" of the matroid (think of these as the stable, solid ground). It's a nice, tidy map, but it only tells half the story.
• The Conormal Complex (The New, Messy Map): This is a newer, more complicated map introduced by recent mathematicians. It tries to capture a "Lagrangian" view (a fancy word for a specific kind of balance between two sides). The authors admit this map is "unruly." It's messy, has holes, and doesn't follow the neat rules of the Classic Map.
• The Biflats Complex (The Master Blueprint): This is the big, all-encompassing map the authors created. It contains everything. It's like a giant warehouse that holds both the Classic Map and the Messy Map inside it, along with a bunch of extra junk.

2. The Big Discovery: "Simple Homotopy Equivalence"

The main goal of the paper is to prove that the Messy Map (Conormal) and the Classic Map (Bergman) are actually the same shape.

In math, saying two things are "homotopy equivalent" is like saying you can stretch a rubber band into a circle or a square without tearing it. They are topologically the same.

But the authors go a step further. They prove they are "simple homotopy equivalent."

• The Analogy: Imagine you have a block of clay (the Master Blueprint/Biflats Complex).
  • You can carve away pieces of clay to turn it into the Messy Map.
  • You can also carve away different pieces of clay to turn it into the Classic Map.
  • Because you can get to both shapes by just removing pieces (collapsing) without ever having to glue anything back together or tear the clay, they are "simply" equivalent. It's a very strong, combinatorial proof that they are the same.

3. The "Biflat" Concept: The Two-Sided Coin

To build this Master Blueprint, the authors invented a new object called a biflat.

• The Analogy: Imagine a standard "flat" (a piece of ground) is a coin showing "Heads."
• A biflat is a coin that shows "Heads" on one side and "Tails" on the other, but with a catch: The "Heads" side must come from the original matroid, and the "Tails" side must come from its "dual" (a mirror-image version of the matroid).
• Crucially, the two sides must cover the whole coin (the ground set).
• The Biflats Complex is just a collection of all these two-sided coins arranged in a specific order.

4. The "Collapse" Magic

The paper's main technical achievement is showing how to turn the Master Blueprint into the other two maps. They do this using elementary collapses.

• The Analogy: Think of the Master Blueprint as a house made of cardboard boxes stacked in a specific way.
  • Some boxes are "free faces"—they are sticking out and not holding up anything else.
  • The authors show you can pull these boxes out one by one (a "collapse").
  • If you pull them out in the right order, the house shrinks down perfectly into the Messy Map.
  • If you pull them out in a different right order, the house shrinks down perfectly into the Classic Map.

This proves that even though the Messy Map looks weird and the Classic Map looks tidy, they are both just "collapsed" versions of the same giant structure.

5. Why Does This Matter?

• Solving Old Mysteries: Mathematicians had long suspected these shapes were related, but they didn't have a direct, step-by-step recipe to prove it. This paper provides that recipe.
• New Tools: By creating the "Biflats Complex," the authors gave mathematicians a new, powerful tool to study these shapes. Even though the Conormal Complex is messy, knowing it comes from a structured "Biflats" world helps us understand its properties.
• Uniform vs. Non-Uniform: The paper notes that if your matroid is "uniform" (very simple and symmetrical, like a perfect sphere), all three maps are actually identical. But if the matroid is complex and irregular, the maps look different but are still topologically the same.

Summary

The authors built a giant, messy warehouse (the Biflats Complex) containing all the information about a matroid's shape. They then showed that you can walk through this warehouse and exit through two different doors:

1. One door leads to the Conormal Complex (the new, complex map).
2. The other door leads to the Join of Bergman Complexes (the classic, tidy map).

Because you can get to both doors by just walking straight through and removing obstacles (collapsing), the two maps are fundamentally the same shape. This is a beautiful example of how complex, messy mathematical objects can be simplified to reveal a hidden, elegant symmetry.

Technical Summary

1. Problem Statement and Motivation

The paper addresses the combinatorial and topological structure of matroids, specifically focusing on the relationship between the conormal complex ($\Delta_{M, M^\perp}$) and the join of Bergman complexes ($\Delta_M * \Delta_{M^\perp}$).

• Background: The Bergman fan of a matroid $M$ is a well-studied object whose support is the tropical linear space $\text{trop}(M)$. Recently, Ardila, Denham, and Huh (ADH) introduced the conormal fan, a Lagrangian analog with support $\text{trop}(M) \times \text{trop}(M^\perp)$.
• The Gap: While ADH established that the conormal complex and the join of Bergman complexes are simple homotopy equivalent (via a sequence of edge subdivisions and inverse subdivisions), their proof relied on geometric operations.
• The Challenge: The conormal complex is combinatorially "unruly." Unlike the Bergman complex (which is the order complex of the lattice of flats), the conormal complex is generally not a flag complex and cannot be realized as the order complex of a poset. Furthermore, the join of Bergman complexes is not naturally a subcomplex of the conormal complex.
• Objective: The authors aim to provide a purely combinatorial proof of the simple homotopy equivalence between these two complexes by constructing a larger, unifying simplicial complex and demonstrating explicit sequences of elementary collapses from this larger complex onto both target complexes.

2. Methodology and Key Definitions

The authors introduce a new combinatorial object, the biflat, to serve as the bridge between the two complexes.

A. The Biflat and Biflats Complex

• Biflat ($F|G$): A pair consisting of a flat $F$ of matroid $M$ and a flat $G$ of the dual matroid $M^\perp$ such that $F \cup G = E$ (the ground set), with $F, G \neq \emptyset$ and at least one being a proper flat.
• Poset of Biflats ($BF_{M, M^\perp}$): Ordered by $F|G \leq F'|G'$ if $F \subseteq F'$ and $G \supseteq G'$.
• Biflats Complex ($\Delta(BF_{M, M^\perp})$): The order complex of the poset of biflats. This complex serves as the "parent" structure containing both the conormal complex and the join of Bergman complexes as subcomplexes.
  • Properties: This complex is generally not pure (it has facets of different dimensions) and not shellable.

B. Distinguished Subcomplexes

1. Unmixed Biflats Complex ($\Delta(BF^u_{M, M^\perp})$):
   • Defined by restricting the poset to unmixed biflats (where either $F=E$ or $G=E$).
   • Result: This complex is isomorphic to the join of the Bergman complexes ($\Delta_M * \Delta_{M^\perp}$).
2. Conormal Complex ($\Delta_{M, M^\perp}$):
   • Defined as the subcomplex of the biflats complex consisting of biflags.
   • Biflag: A bichain satisfying a "gap condition" (specifically, the union of the flat and dual flat components in the chain never equals the ground set $E$ until the very end).
   • This definition ensures the conormal complex captures the specific combinatorial structure of the conormal fan.

C. Proof Strategy: Elementary Collapses

The core methodology relies on simple homotopy theory. Two complexes are simple homotopy equivalent if they can both be collapsed from a common larger complex via a sequence of elementary collapses.

• Elementary Collapse: Removing a pair of faces $(\sigma, \tau)$ where $\sigma$ is a maximal face, $\tau$ is a free face of $\sigma$ (contained in no other maximal face), and $\tau$ is a codimension-1 face of $\sigma$.
• The authors construct two distinct sequences of collapses starting from the Biflats Complex:
  1. Collapsing onto the Unmixed Biflats Complex.
  2. Collapsing onto the Conormal Complex.

3. Key Contributions and Results

Theorem 1: Collapse to the Join of Bergman Complexes

• Statement: There exists an explicit sequence of elementary collapses from the biflats complex $\Delta(BF_{M, M^\perp})$ onto the unmixed biflats complex $\Delta(BF^u_{M, M^\perp})$.
• Mechanism: The authors order the mixed biflats (where $F \neq E$ and $G \neq E$) using a linear extension of the poset. They iteratively remove these mixed biflats.
• Key Lemma: For any mixed biflat $F|G$ in the current poset, the subposet of elements strictly smaller than it has a unique maximal element (specifically $F|E$). This allows the application of a standard lemma (Lemma 2.7) to collapse the complex by removing the element.
• Outcome: Since $\Delta(BF^u_{M, M^\perp}) \cong \Delta_M * \Delta_{M^\perp}$, this proves the biflats complex collapses to the join of Bergman complexes.

Theorem 2: Collapse to the Conormal Complex

• Statement: There exists an explicit sequence of elementary collapses from the biflats complex $\Delta(BF_{M, M^\perp})$ onto the conormal complex $\Delta_{M, M^\perp}$.
• Mechanism: The authors identify the faces of the biflats complex that are not in the conormal complex (bichains that are not biflags). They focus on non-repeating bichains (chains where no flat or dual flat appears twice).
• Ordering: They define a specific partial order $\preceq$ on these non-repeating bichains.
• Key Lemma (Lemma 4.10): For any non-repeating bichain $F|G$ in the current complex, the link of $F|G$ contains a cone vertex (a specific biflat $F_1|G_2$).
• Outcome: The existence of a cone vertex in the link implies the complex can be collapsed onto the deletion of that face. By iterating this process according to the order $\preceq$, they remove all non-biflag chains, leaving exactly the conormal complex.

Corollary 3: Simple Homotopy Equivalence

• Combining Theorems 1 and 2, the authors conclude that $\Delta_{M, M^\perp}$ and $\Delta_M * \Delta_{M^\perp}$ are simple homotopy equivalent. This recovers the result of ADH but via a purely combinatorial, constructive proof involving elementary collapses rather than geometric subdivisions.

4. Combinatorial Properties and Differences

The paper highlights that while these complexes are homotopy equivalent, they possess distinct combinatorial properties, summarized in Table 1 of the paper:

Property	Conormal Complex ($\Delta_{M, M^\perp}$)	Biflats Complex ($\Delta(BF)$)	Unmixed Biflats ($\Delta(BF^u)$)
Dimension	$|E|-3$	$|E|-2$	$|E|-3$
Pure	Yes	No (generally)	Yes
Flag	No (generally)	Yes	Yes
Shellable	Unknown (likely yes)	No (generally)	Yes
Homeomorphic to $\Delta_M * \Delta_{M^\perp}$	Yes	No	Yes

• Significance of Differences: The biflats complex is the "largest" object, containing the necessary structure to perform the collapses. The conormal complex is "smaller" and more restrictive (not flag), while the unmixed complex is "cleaner" (pure, flag, shellable) but requires the collapse of the biflats complex to reach it.

5. Significance and Future Directions

• Combinatorial Rigor: The paper provides a rigorous, purely matroid-theoretic proof of a topological equivalence that was previously understood through tropical geometry. This makes the result accessible to combinatorialists without requiring deep knowledge of tropical fans.
• Structural Insight: By introducing the poset of biflats, the authors unify the study of the conormal fan and the product of Bergman fans, revealing the underlying combinatorial mechanism (elementary collapses) that connects them.
• Counterexamples: The paper demonstrates that the conormal complex is not always a flag complex and that the biflats complex is not shellable, correcting potential assumptions about the "niceness" of these Lagrangian objects.
• Open Questions:
  1. Global Deformation: Can these results be extended to "global" objects like the bipermutohedral complex?
  2. Equivariance: Can the collapses be made Aut(M)-equivariant? (i.e., preserving the symmetry of the matroid during the collapse).

In summary, Nathanson and Partida successfully bridge the gap between the geometric intuition of Lagrangian matroid theory and the discrete world of posets and simplicial complexes, offering a new, constructive perspective on the topology of matroid invariants.