Compactifications of spaces of symmetric matrices and pointed Kontsevich spaces of isotropic Grassmannians

This paper constructs and analyzes the birational geometry of a Kausz-type compactification $\mathcal{TL}_n$ of the space of symmetric matrices by realizing it as an evaluation fiber in the Kontsevich space of pointed conics on the Lagrangian Grassmannian, while also presenting analogous results for orthogonal Grassmannians.

The Gist

Imagine you are an architect trying to design a perfect, infinite city. In mathematics, this "city" is a space of symmetric matrices—a vast collection of grids of numbers that have a special symmetry (like a mirror image across the diagonal).

The problem is that this city is incomplete. If you walk far enough in certain directions, you fall off the edge into "infinity," where the rules break down. Mathematicians call this "compactification": the art of building a fence, a wall, or a new neighborhood to catch those falling edges and make the space complete, smooth, and usable again.

This paper, written by Fang, Massarenti, and Wu, is about building two very specific, beautiful types of these "complete cities" and understanding their blueprints.

Here is the story of their work, broken down into simple concepts:

1. The Two Cities They Built

The authors constructed two related "cities" (mathematical varieties):

• City $T L_n$ (The Symmetric City): This is a new, fancy version of the space of symmetric matrices. They built it by taking a standard, well-known geometric shape (the Lagrangian Grassmannian) and performing a series of precise "blow-ups."
  • The Analogy: Imagine a smooth, flat sheet of paper. If you pinch a point and pull it up into a cone, you've created a new shape. The authors did this repeatedly at specific "pinch points" (called osculating loci) to resolve singularities and create a smooth, complex structure. They call this a Kausz-type compactification, named after a mathematician who first built a similar structure for general matrices.
• City $T O_n$ (The Skew City): They also built a twin city for "skew-symmetric" matrices (where the numbers flip signs across the diagonal). This is related to orthogonal geometry.

2. The Secret Connection: The "Pointed" Map

The most exciting part of the paper is how they connected their new city ($T L_n$) to a famous concept in modern geometry called Kontsevich spaces.

• The Concept: Imagine you are drawing curves (like conics or lines) on a canvas. A "Kontsevich space" is a giant catalog of all possible ways to draw these curves, including broken or degenerate ones.
• The Twist: Usually, we just look at the curves. But what if we put a "pin" (a marked point) on the curve? This is a "pointed" curve.
• The Discovery: The authors realized that their new city, $T L_n$, is actually a slice of a much larger catalog of pointed curves.
  • The Analogy: Imagine a massive library of all possible movies (the Kontsevich space). If you ask the librarian, "Show me all movies where the first scene is at location A and the second scene is at location B," you get a specific shelf of movies. The authors proved that their city $T L_n$ is exactly that specific shelf.
  • Why it matters: Because $T L_n$ is a "slice" of the bigger catalog, they can use the simple, clean rules of $T L_n$ to solve complicated problems about the whole library of pointed curves.

3. The "Magic" Properties of the City

Once they built $T L_n$, they analyzed its "geography" (its birational geometry). They found it has some superpowers:

• It's a "Mori Dream Space": This is a fancy way of saying the city is incredibly well-organized. You can predict exactly how to transform it, where the roads go, and how to build new neighborhoods without running into dead ends. It's like a city with a perfect, infinite map.
• It's Rigid (Stable): The authors proved that this city cannot be slightly wiggled or deformed. If you try to push a wall, it doesn't bend; it's locked in place. This is rare and valuable in geometry.
• It's "Fano" (or close to it): In geometry, being "Fano" is like having a very positive, energetic curvature. It means the space is "nice" and behaves well. They found that for small dimensions, the city is perfectly Fano, and for larger ones, it's "weakly" Fano (still very nice, just slightly less energetic).

4. Solving the "Pointed Conic" Puzzle

The ultimate goal was to understand the moduli space of pointed conics (curves shaped like circles or ellipses with a marked point) inside the Lagrangian Grassmannian.

• The Problem: Before this paper, mathematicians didn't have a complete map of the "divisor cones" (the boundaries and limits) for these pointed spaces, especially for higher dimensions. It was like trying to navigate a foggy forest without a compass.
• The Solution: Because $T L_n$ is a slice of this space, the authors used the clear, sunny map of $T L_n$ to draw the map for the whole pointed conic space.
• The Result: They successfully calculated the exact boundaries, the "Fano index" (a measure of the space's energy), and the group of symmetries (how you can rotate or flip the space without changing it) for these pointed conic spaces for all dimensions.

Summary in a Nutshell

The authors built a perfect, smooth model city ($T L_n$) to represent a complex mathematical space. They discovered that this model city is actually a key that unlocks the secrets of a much larger, messier library of "pointed curves." By studying the clean model, they were able to draw a complete, precise map of the messy library, solving long-standing questions about the geometry of these spaces.

The Takeaway: Sometimes, to understand a giant, complicated problem, you need to build a smaller, perfect model of it first. This paper shows us how to build that model and use it to see the whole picture clearly.

Technical Summary

Here is a detailed technical summary of the paper "Compactifications of Spaces of Symmetric Matrices and Pointed Kontsevich Spaces of Isotropic Grassmannians" by Hanlong Fang, Alex Massarenti, and Xian Wu.

1. Problem Statement and Motivation

The paper addresses two interconnected problems in modern algebraic geometry:

1. Compactification of Symmetric Matrices: Constructing and analyzing a specific compactification of the space of symmetric matrices (or the open orbit in the Lagrangian Grassmannian) that generalizes the Kausz compactification of the general linear group.
2. Birational Geometry of Pointed Kontsevich Spaces: Providing an explicit, uniform description of the birational geometry (divisor cones, automorphism groups, Fano properties) of pointed moduli spaces of stable maps, specifically $M_{0,1}(LG(n, 2n), 2)$, the space of pointed conics in the Lagrangian Grassmannian.

While the birational geometry of unpointed Kontsevich spaces (e.g., $M_{0,0}$) is well-understood via connections to GIT quotients and wonderful compactifications, the pointed case ($M_{0,k}$ with $k \ge 1$) is significantly more complex. Pointed spaces often fail to be spherical varieties under natural group actions, making standard combinatorial tools insufficient. The authors aim to bridge this gap by realizing pointed Kontsevich spaces as evaluation fibers of specific compactifications.

2. Methodology

The authors employ a multi-faceted approach combining:

• Iterated Blow-ups: They construct the compactification $T L_n$ by performing a sequence of blow-ups on the Lagrangian Grassmannian $LG(n, 2n)$ along specific "osculating loci" associated with two transverse Lagrangian points $p_+$ and $p_-$.
• Spherical Variety Theory: They analyze $T L_n$ as a toroidal spherical variety for the action of the symplectic group $Sp_{2n}$ (specifically the stabilizer of the pair $(p_+, p_-)$). This allows the use of combinatorial tools (colored fans, Cox rings) to describe divisor cones.
• Modular Interpretation: They identify $T L_n$ as a general fiber of an evaluation map within a Kontsevich moduli space $M_{0,3}(LG(n, 2n), n)$. This links the geometry of the compactification directly to the geometry of stable maps.
• Hilbert Quotients: They relate the construction to the Hilbert quotient of $LG(n, 2n)$ under a $\mathbb{C}^*$-action, showing that the compactification serves as the universal family over the space of complete quadrics.
• Computational Algebra: They utilize Maple scripts to compute generators for effective, nef, and movable cones for specific cases ($n=2, 3$) to verify general patterns.

3. Key Contributions and Results

A. The Kausz-Type Compactification $T L_n$

The authors construct a smooth projective variety $T L_n$ which compactifies the space of symmetric matrices (identified with the open orbit in $LG(n, 2n)$).

• Construction: $T L_n$ is obtained by iteratively blowing up $LG(n, 2n)$ along the strict transforms of osculating loci $Z^\pm_k$ at two fixed points $p_\pm$.
• Structure: $T L_n$ is a toroidal spherical variety for the group $G = \text{Stab}_{Sp_{2n}}(p_+, p_-)$.
• Geometry:
  • It admits a flat morphism $P_n: T L_n \to CQ_n$ to the space of complete quadrics $CQ_n$.
  • $P_n$ is the universal family over the Hilbert quotient $LG(n, 2n) // \mathbb{C}^*$.
  • The Picard group is generated by the pullback of the Plücker class $H$ and the exceptional divisors $D^\pm_i$.
• Rigidity: The authors prove that $H^k(T L_n, T_{T L_n}) = 0$ for all $k > 0$. Consequently, $T L_n$ is locally rigid (admits no non-trivial local deformations).
• Fano Property: $T L_n$ is weak Fano for all $n \ge 2$, and Fano if and only if $n=2$.
• Automorphisms: The group of pseudoautomorphisms coincides with the biregular automorphism group:
  $$\text{PsAut}(T L_n) \cong \text{Aut}(T L_n) \cong (GL_n / \{\pm I\}) \rtimes S_2$$

B. Birational Geometry of Pointed Conics $M_{0,1}(LG(n, 2n), 2)$

Using the identification of $T L_n$ as a fiber of the evaluation map, the authors derive the birational geometry of the moduli space of pointed conics.

• Divisor Cones: They provide explicit descriptions of the Effective, Nef, and Movable cones.
  • Effective Cone: Generated by $H_1$ (pullback of Plücker), $\Delta_{1|1}$ (boundary divisor), and $D_{unb}$ (unbalanced divisor).
  • Nef Cone: Generated by $H_1$, $H_{\sigma_2}$ (Schubert class), and $T$ (tangency divisor).
• Fano Classification:
  • $M_{0,1}(LG(n, 2n), 2)$ is a Mori Dream Space for all $n \ge 2$.
  • It is a Fano variety if and only if $2 \le n \le 5$ (with Fano index 1).
  • It is weak Fano if and only if $2 \le n \le 6$.
• Automorphisms: The automorphism group is induced by the target:
  $$\text{Aut}(M_{0,1}(LG(n, 2n), 2)) \cong PSp_{2n}$$

C. Extension to Orthogonal Grassmannians

The methodology is extended to the orthogonal setting ($OG(n, 2n)$). The authors construct analogous compactifications $T O_n$ related to spaces of complete skew-forms.

• Result: These varieties are also smooth, projective, spherical, and weak Fano. They are locally rigid and serve as universal families for the Hilbert quotients of orthogonal Grassmannians.

4. Significance

• Uniform Description of Pointed Spaces: This work provides one of the few uniform, computable descriptions of the birational geometry of a family of pointed Kontsevich spaces targeting higher-rank homogeneous varieties. It overcomes the difficulty that pointed spaces are often not spherical themselves by embedding them into a spherical framework ($T L_n$).
• New Compactification: The construction of $T L_n$ offers a new, explicit compactification of the space of symmetric matrices that is distinct from, yet related to, the De Concini–Procesi wonderful compactification. It fills a gap in the theory of compactifications for homogeneous spaces with non-trivial centers.
• Rigidity and Moduli: The proof of local rigidity for $T L_n$ and the complete classification of Fano indices for $M_{0,1}$ contribute significantly to the minimal model program (MMP) and the classification of higher-dimensional varieties.
• Computational Framework: The paper provides a concrete algorithm (implemented in Maple) to compute the cones of divisors and curves for these varieties, facilitating further research in Gromov-Witten theory and birational geometry.

In summary, the paper successfully links the geometry of symmetric matrices, spherical varieties, and stable maps, offering a comprehensive structural analysis of both the compactification $T L_n$ and the pointed moduli space $M_{0,1}(LG(n, 2n), 2)$.