🔢 mathematics

Deletion-contraction properties of graphically stable spaces

This paper investigates the intersection numbers and Grothendieck classes of graphically stable spaces by establishing deletion-contraction relations that connect their geometry to chromatic polynomials and combinatorial invariants, thereby deriving string and dilaton equations and extending these results to higher genera under specific conditions.

Original authors: Andy Xiaoan Yang

Published 2026-08-13

📖 4 min read🧠 Deep dive

CC BY 4.0

Original authors: Andy Xiaoan Yang

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 designing a city where the buildings are not made of brick and mortar, but of flexible, stretchy rubber sheets. In this city, the "streets" are paths that can twist and turn, and the "landmarks" are special points marked with flags. Sometimes, these landmarks want to get closer together, maybe even bump into each other. In the world of mathematics known as algebraic geometry, there is a famous rulebook called the Deligne–Mumford compactification. It tells us exactly what happens when these landmarks crash: they form a "node," a tiny, sharp pinch in the rubber sheet, creating a new kind of shape. This rulebook is essential for understanding the geometry of curves, which are the fundamental shapes in this mathematical universe.

However, what if we wanted to write a different rulebook? What if we decided that some landmarks are allowed to crash into each other freely, while others are strictly forbidden from touching? This is where "graphically stable spaces" come in. Imagine a graph as a simple drawing of dots (landmarks) connected by lines (rules). If two dots are connected by a line, they are "best friends" and must stay apart. If there is no line between them, they are "strangers" who are allowed to collide and merge. By changing the pattern of lines in the graph, we create a whole new family of cities, each with its own unique rules for how the rubber sheets can fold and pinch. Mathematicians care about these spaces because they act as a bridge between pure geometry and the combinatorial world of graphs, helping us count and understand complex shapes in ways that were previously impossible.

Now, enter the story of this paper. The author, Andy Xiaoan Yang, discovers a magical "deletion-contraction" key that unlocks the secrets of these graph-based cities. In the world of graph theory, there is a famous trick: to understand a complex graph, you can either delete a line (pretend it doesn't exist) or contract it (squish the two dots it connects into one). Yang proves that this same trick works perfectly for the geometry of these rubber-sheet cities. Whether you are counting the "twists" in the rubber (called $\psi$ -class integrals) or trying to figure out the total "size" of the city (its Grothendieck class), you can break the problem down into smaller, simpler pieces using this deletion-contraction rule.

The paper shows that the geometry of these spaces is governed by a hidden structure that looks exactly like the rules for coloring a map. Specifically, the author finds that the number of ways to twist the rubber sheets or the "size" of the space can be calculated using something called the "chromatic polynomial." Think of the chromatic polynomial as a magic formula that tells you how many ways you can paint a graph with a certain number of colors so that no two connected dots share the same color. Yang proves that if you plug the right numbers into this coloring formula, you get the exact answers for the geometric questions. For example, in the simplest case (genus 0), the entire "size" of the city is determined by a specific polynomial derived from the graph's coloring rules.

The paper also tackles some tricky edge cases. It rules out the idea that you can simply ignore the rules when two points have no "twist" assigned to them; sometimes, you have to add a tiny "leaf" (a new dot) to the graph to make the math work correctly. Furthermore, the author connects these geometric findings to other areas of math and physics, such as the "scattering potentials" used in particle physics to predict how particles bounce off each other. The paper suggests that the number of critical points in these physics formulas is actually the same as the number of ways to color the graph, a connection that was previously just a guess but is now backed by solid proof in many cases.

In short, this paper doesn't just solve a few isolated math problems; it provides a unified framework. It shows that the chaotic-looking world of these flexible rubber-sheet cities is actually organized by the simple, rigid logic of graph theory. By using the deletion-contraction method, the author turns incredibly complex geometric calculations into manageable algebraic puzzles, revealing that the deep structure of these spaces is written in the language of graph coloring. While the paper proves these results rigorously for many cases, it also points out that for some very complex, high-genus scenarios, the full picture is still being explored, leaving the door open for future explorers to use this new key to unlock even more mysteries.

Technical Summary: Deletion–Contraction Properties of Graphically Stable Spaces

Problem Statement
The paper investigates the geometry and topology of graphically stable spaces, denoted $M_{g,G}$ . These spaces provide a modular compactification of the moduli space of curves $M_{g,n}$ , parametrizing $n$ -marked nodal curves of genus $g$ where the permitted collisions of markings are dictated by a finite simple graph $G$ . Specifically, a set of markings is allowed to coincide if and only if the corresponding vertices in $G$ form an independent set. While the case where $G$ is the complete graph $K_n$ recovers the classical Deligne–Mumford space $M_{g,n}$ , and $G$ being a complete multipartite graph relates to Hassett's weighted stable curves, the general case $M_{g,G}$ offers a broader class of alternative compactifications.

The central problem addressed is the lack of a unified structural framework for computing intersection numbers of $\psi$ -classes (cotangent classes) and Grothendieck classes (classes in the Grothendieck ring of varieties) on these spaces. Existing methods often rely on wall-crossing formulas or specific combinatorial descriptions for special cases (e.g., heavy/light spaces), but a general principle governing the dependence of these invariants on the underlying graph structure $G$ has been elusive.

Methodology
The author's primary methodological insight is the application of deletion–contraction relations from graph theory to the geometry of $M_{g,G}$ . The paper establishes that the geometric invariants of $M_{g,G}$ satisfy recursive relations analogous to the deletion–contraction relation for the chromatic polynomial $\chi_G(x)$ .

Wall-Crossing and Induction: The proofs rely heavily on the wall-crossing formula relating integrals on $M_{g,G}$ to those on $M_{g,n}$ (via the morphism $\rho_G: M_{g,n} \to M_{g,G}$ ). By analyzing the behavior of $\psi$ -classes under this morphism and utilizing the combinatorics of $G$ -stable partitions, the author derives recursive formulas.
Graph-Theoretic Reduction: The core technique involves reducing integrals over $M_{g,G}$ to sums of integrals over $M_{g, G-e}$ (deletion) and $M_{g, G/e}$ (contraction). Special care is taken to handle cases where the resulting graphs fail the stability condition (e.g., becoming bipartite in genus 0), requiring modifications such as attaching leaves (denoted $G \odot e$ ) or adjusting $\psi$ -class exponents.
Motivic Integration: For Grothendieck classes, the paper utilizes the cut-and-paste property of the Grothendieck ring $K_0(\text{Var}_k)$ . By stratifying the boundary of the partial compactification and relating the classes of configuration spaces $\text{Conf}_G(\mathbb{P}^1)$ to the chromatic polynomial, the author derives explicit polynomial expressions for the classes $[M_{0,G}]$ .

Key Contributions and Results

Deletion–Contraction for $\psi$ -Class Integrals (Theorem A): The paper proves that for a $g$ -stable graph $G$ and an edge $e=ab$, the integral of $\psi$ -classes satisfies a deletion–contraction relation.
- If $(k_a, k_b) \neq (0,0)$ , the integral over $M_{g,G}$ equals the sum of the integral over $M_{g, G-e}$ and a modified integral over $M_{g, G/e}$ where the exponents at the contracted vertex are summed and reduced by 1.
- If $(k_a, k_b) = (0,0)$ , the relation involves an auxiliary graph $G \odot e$ (obtained by attaching a leaf to the contracted vertex) to handle the vanishing of certain terms.
- This framework unifies and generalizes classical string and dilaton equations for $M_{g,n}$ .
Connection to Chromatic Polynomials: A major result is the explicit expression of various $\psi$ -class integrals in terms of the chromatic polynomial $\chi_G(x)$ and its derivatives.
- Genus 1: The integral $\int_{M_{1,G}} \psi_1 \cdots \psi_n$ is proportional to the derivative of $\chi_G(x)$ at $x=0$ .
- Genus 0: Integrals over spaces like $M_{0, G \sqcup K_3}$ and $M_{0, G * K_3}$ are shown to equal evaluations of $\chi_G(x)$ at specific points (e.g., $x=2, x=-1$ ).
- This establishes that the intersection theory of $M_{g,G}$ completely determines the chromatic polynomial of $G$ .
Deletion–Contraction for Grothendieck Classes (Theorem D): The paper establishes the relation $[M_{g,G}] = [M_{g, G-e}] - [M_{g, G/e}]$ in the Grothendieck ring $K_0(\text{Var}_k)$ , provided the graphs remain $g$ -stable.
- Explicit Formula for Genus 0 (Theorem E): For a 0-stable graph $G$ , the Grothendieck class is given by $[M_{0,G}] = \frac{\chi_G(L+1)}{L^3 - L}$ , where $L = [\mathbb{A}^1]$ . This expresses the class entirely in terms of the chromatic polynomial.
Topological Invariants: By specializing the Grothendieck class formula to different fields and weight filtrations, the paper derives formulas for:
- The complex Euler characteristic $\chi(M_{0,G}(\mathbb{C}))$ .
- The top-weight Euler characteristic $\chi_{\text{top-wt}}(M_{0,G}(\mathbb{C}))$ , identified with Crapo's $\beta$ -invariant.
- The compactly supported Euler characteristic over $\mathbb{R}$ , linked to acyclic orientations with a unique source.
- Point counts over finite fields $\mathbb{F}_q$ .
Higher Genus Results: The paper extends these results to genus $g \geq 1$ . For genus 1, explicit formulas for Euler characteristics are derived in terms of $\chi_G(x)$ and its derivatives. For $g \geq 2$ , under the condition that the chromatic number $\chi(G)$ is sufficiently large (specifically $\chi(G) > 2g+2$ ), the Euler characteristics are expressed using Bernoulli numbers and evaluations of $\chi_G(x)$ at negative integers.
New Combinatorial Formulas: The work yields a new formula for Crapo's $\beta$ -invariant of graphs, relating it to the alternating sum of the number of edges in $G$ -stable trees (Corollary H).

Significance and Claims
The paper claims that the deletion–contraction principle serves as a "natural organizing principle" for the study of graphically stable spaces across all genera. Its significance lies in:

Unification: It provides a single framework that recovers known results for $M_{g,n}$ and Hassett spaces while extending them to a large class of alternative compactifications.
Bridge between Geometry and Combinatorics: It demonstrates a deep, explicit correspondence between the intersection theory of moduli spaces of curves and classical graph invariants (specifically the chromatic polynomial). This allows geometric quantities to be computed via purely combinatorial means and vice versa.
Resolution of Conjectures: The results confirm a conjecture by Blankers, Gillespie, and Levinson regarding the count of critical points of CHY scattering potentials (Proposition 3.32) by linking them to the Euler characteristic of $M_{0,G}$ .
Generalization of Wall-Crossing: The paper generalizes wall-crossing techniques from Hassett spaces to the broader context of graphically stable spaces, offering alternative proofs for existing theorems and new tools for computing topological invariants.

The author notes that while the deletion–contraction relations hold generally, the stability conditions on the graphs are crucial; the paper carefully delineates the cases where the standard relations apply and where modifications (like the $G \odot e$ construction) are necessary. The work also posits a conjecture (Conjecture I) equating integrals over $M_{g,G}$ with specific pullback integrals over $M_{g,n}$ , suggesting a deeper structural link between the tautological rings of these spaces, which has been verified for several infinite classes of graphs.

More like this