🔢 mathematics

The Pandharipande-Thomas rationality conjecture for superpositive curve classes on projective complex 3-manifolds

This paper proves the Pandharipande-Thomas rationality conjecture for generating functions of PT invariants with descendent insertions on projective complex 3-manifolds for superpositive curve classes by applying a newly developed theory of enumerative invariants in abelian categories and wall-crossing formulae.

Original authors: Reginald Anderson, Dominic Joyce

Published 2026-04-08

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Reginald Anderson, Dominic Joyce

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 an architect trying to count the number of ways you can build specific types of structures using a limited set of Lego bricks. In the world of mathematics, specifically in a field called Algebraic Geometry, the "bricks" are shapes called curves, and the "structures" are complex 3-dimensional spaces (like a twisted, multi-dimensional donut).

This paper, written by Reginald Anderson and Dominic Joyce, is about solving a massive counting puzzle involving these curves. Here is the breakdown of what they did, using simple analogies.

1. The Big Problem: The "Infinite" Counting Machine

Mathematicians have a machine called a Pandharipande–Thomas (PT) invariant. Think of this machine as a super-complex calculator that takes a specific type of curve (let's call it a "shape") and a number (let's call it "size") and spits out a number representing how many ways that shape can exist in the 3D space.

The problem is that if you ask the machine for every possible size, it generates an infinite list of numbers. Mathematicians want to know: Is there a simple, neat formula (a rational function) that describes this entire infinite list?

Usually, these lists are chaotic. But for certain "nice" shapes, mathematicians suspect there is a hidden order—a simple formula that generates the whole list, like a recipe that can bake an infinite number of cookies.

2. The Special Ingredients: "Superpositive" Curves

Not all curves are created equal. Some are "messy" and hard to count. The authors focus on a special group of curves they call "superpositive."

The Analogy: Imagine you are sorting marbles. Some marbles are heavy and sink (positive), some are light and float (negative), and some are just weird.
The Rule: A "superpositive" curve is like a stack of marbles where every single marble in the stack is heavy. You can't break the stack apart and find a light marble hiding inside.
Why it matters: If a curve is "superpositive," it behaves very predictably. It's like a well-organized library where every book is in its correct spot, making it much easier to count them.

3. The Main Discovery: The "Rationality" Conjecture

The authors prove a famous guess (conjecture) made by two other mathematicians, Pandharipande and Thomas.

The Guess: If you take a "superpositive" curve and run it through the counting machine, the resulting infinite list of numbers isn't random chaos. It is actually the Laurent expansion of a rational function.
Translation: This is a fancy way of saying: "There is a simple fraction formula (like $\frac{A}{B}$ ) that generates this entire infinite list."
The "Poles": The authors also prove that the "glitches" in this formula (where the numbers go crazy, called "poles") only happen at very specific, predictable spots (like $q=0$ or specific roots of unity). It's like saying a broken clock only stops at 12:00 or 6:00, never at random times.

4. How They Did It: The "Wall-Crossing" Bridge

How did they prove this? They didn't just stare at the curves. They built a bridge between two different worlds of mathematics.

World A (The PT World): This is where the "stable pairs" (our Lego structures) live. It's complicated and hard to count directly.
World B (The DT World): This is a different way of counting the same things, using "Donaldson–Thomas invariants." Think of this as a different language that describes the same Lego structures but is easier to analyze in certain ways.

The Wall-Crossing Formula:
Imagine you are walking through a landscape. Sometimes, the terrain changes (a "wall"), and the rules of how you walk change slightly. The authors used a powerful mathematical tool called a Wall-Crossing Formula.

This formula acts like a translator. It tells you exactly how to convert a count from World A (PT) into World B (DT) and back again.
They showed that in the "superpositive" world, the counts in World B have a very nice, repeating pattern (like a polynomial that repeats every few steps).
Because the pattern in World B is so nice, and the translator (the Wall-Crossing Formula) is reliable, the pattern in World A must also be nice.

5. The "Vertex Algebra" Secret Sauce

To make the translator work, they had to use some heavy machinery called Vertex Algebras and Lie Algebras.

The Metaphor: Imagine the space of all possible shapes is a giant, multi-dimensional dance floor.
Vertex Algebras are like the choreography rules that tell the dancers how to interact when they bump into each other.
Lie Algebras are the rules for how the dance floor itself changes shape.
The authors used these complex rules to show that the "dance" of the curves follows a strict, predictable rhythm, which guarantees that the final counting formula is rational (simple).

6. What They Didn't Solve

The paper is a huge success, but it's not a "perfect" solution.

They proved the formula exists and where the glitches are.
However, they couldn't prove a specific symmetry rule (that the formula looks the same if you flip it upside down).
Why? Their "translator" (the Wall-Crossing Formula) has a slight technical limitation. It's like having a dictionary that translates English to French perfectly, but it doesn't have a section for translating idioms. Another team of mathematicians (Karpov and Moreira) is working on a slightly different dictionary that might solve that missing piece.

Summary

In short, Anderson and Joyce proved that for a specific, well-behaved class of 3D shapes, the infinite list of ways to build them follows a simple, predictable mathematical recipe. They did this by building a bridge between two different counting methods and using complex "dance rules" (algebra) to show that the chaos of the infinite list is actually just a hidden, elegant pattern.

This is a major step forward in understanding the deep, hidden order of the universe's geometric shapes.

1. Problem Statement

The paper addresses the Pandharipande–Thomas (PT) rationality conjecture regarding the generating functions of enumerative invariants on smooth projective complex 3-folds ( $X$ ).

Context: PT invariants count stable pairs $(F, s)$ on $X$ , where $F$ is a pure 1-dimensional coherent sheaf and $s: \mathcal{O}_X \to F$ is a section with 0-dimensional cokernel. These invariants are defined for curve classes $\beta$ and Euler characteristics $n$ .
The Conjecture: Pandharipande and Thomas conjectured that the generating functions of these invariants (with descendent insertions) are rational functions in the variable $q$ . Furthermore, their poles should occur only at $q=0$ and roots of unity, and they should satisfy specific symmetry properties (inversion formulas).
Previous Status: The conjecture was known for:
- Calabi-Yau 3-folds (Bridgeland, Toda).
- Toric 3-folds (Pandharipande–Pixton).
- Fano or semi-Fano cases with restricted insertions (Pardon).
The Gap: A general proof for superpositive curve classes on arbitrary projective 3-folds, particularly handling the complexity of the underlying algebraic structures and equivariant settings, remained open.

Key Definitions:

Positive Class: $\beta$ is positive if $c_1(X) \cdot \beta > 0$ .
Superpositive Class: $\beta$ is superpositive if every effective summand (factor) of $\beta$ is positive. (Note: If $X$ is Fano, all effective classes are superpositive).

2. Methodology

The authors employ a sophisticated framework developed by the second author (Joyce) involving vertex algebras, Lie algebras on moduli stacks, and wall-crossing formulas.

A. Algebraic Framework: Vertex and Lie Algebras

The core machinery relies on the homology of moduli stacks of objects in the derived category $D^b_{coh}(X)$ .

Moduli Stacks: Let $\mathcal{M}$ be the moduli stack of objects in $D^b_{coh}(X)$ and $\mathcal{M}^{pl}$ be the "projective linear" moduli stack (quotient by the $\mathbb{G}_m$ action).
Vertex Algebra Structure: Joyce previously established that the homology $H_*(\mathcal{M})$ carries a graded vertex algebra structure.
Lie Algebra Structure: The homology $H_*(\mathcal{M}^{pl})$ carries a graded Lie algebra structure.
Connection: There is a morphism from the vertex algebra to the Lie algebra (via a quotient by the image of a translation operator $D$ ). This allows computations in the complex vertex algebra to be lifted to the Lie algebra, where wall-crossing formulas are naturally expressed.

B. Wall-Crossing Formulas

The proof utilizes wall-crossing formulas that relate invariants under changes of stability conditions.

Stability Conditions: The authors compare stability conditions defined by a Kähler class $\omega$ (Gieseker stability) against "limit" stability conditions.
The Identity: They establish an identity (Theorem 2.16) relating the PT invariants (counting stable pairs) to 1-dimensional Donaldson–Thomas (DT) invariants (counting semistable sheaves).
- Specifically, the PT class $[P_n(X, \beta)]^{virt}$ is expressed as a Lie bracket sum involving DT invariants $[M^{ss}_{(\beta, n)}(\mu_\omega)]_{inv}$ and lower-order PT classes.
- This is a recursive relation: PT invariants for a class $\beta$ are determined by DT invariants and PT invariants for "smaller" curve classes (factors of $\beta$ ).

C. Inductive Strategy on Curve Classes

The proof proceeds by induction on the number of effective factors of the curve class $\beta$ .

Base Case: Irreducible classes.
Inductive Step: For a reducible $\beta$ , the wall-crossing formula expresses the PT invariant as a sum of products of invariants for factors $\beta_i$ .
Periodicity: A crucial observation is that for superpositive classes, the DT invariants exhibit quasi-polynomial behavior in the Euler characteristic $n$ . This is derived from the periodicity of the moduli spaces under tensoring with an ample line bundle $L$ .

D. Equivariant Cohomology

The paper extends these results to $G$ -equivariant homology (where a linear algebraic group $G$ acts on $X$ ).

They work with truncated equivariant homology $H_{*, \leq N}^G$ to handle convergence issues in infinite sums arising from the vertex algebra operations.
They prove that the rationality holds for these truncated series, though they note technical limitations prevent proving the full non-truncated equivariant rationality in the general case.

3. Key Contributions

Proof of Rationality for Superpositive Classes: The authors prove that for any superpositive effective curve class $\beta$ , the generating function of PT invariants (with descendent insertions) is a rational function.
Pole Analysis: They confirm that the poles of these rational functions occur only at $q=0$ and roots of unity.
Generalization to Descendents: Unlike some previous results restricted to "primary" insertions (no descendents), this result holds for general descendent insertions $\tau_k(\eta)$ .
Structural Insight: The paper provides a deep structural link between PT invariants and 1-dimensional DT invariants via the Lie algebra of the moduli stack, generalizing the Bridgeland–Toda approach for Calabi-Yau manifolds to the Fano/superpositive setting.
Equivariant Results: The results are extended to the equivariant setting (truncated), providing a framework for computing invariants on toric varieties or other spaces with group actions.

4. Main Results

Theorem 1.4: Conjecture 1.3 (parts a, b, d, e) holds when $\beta$ is a superpositive effective curve class.

(a) The generating function is a rational function.
(b) Poles are at $q=0$ and roots of unity.
(d) The equivariant generating function is rational.
(e) Equivariant poles are at $q=0$ and roots of unity.

Theorem 1.8 (The Core Technical Result):

(a) The formal power series $\sum [P_n^{alg}(X, \beta)]^{virt} q^n$ is the Laurent expansion of a rational function with poles only at $0$ and roots of unity.
(b) The truncated equivariant series is also the expansion of a rational function with the same pole constraints.

Proposition 3.6: The PT virtual classes $[P_n(X, \beta)]^{virt}$ , viewed as elements in the homology of the moduli stack, are piecewise quasi-polynomial functions of $n$ . This is the engine driving the rationality proof.

5. Limitations and Open Questions

Symmetry Conjectures (Parts c, f): The authors do not prove the symmetry properties (inversion formulas) of the conjecture (e.g., $PT(q^{-1}) = \dots PT(q)$ $P T (q^{- 1}) = \dots P T (q)$ ).
- Reason: The wall-crossing technique used (based on the theory in [28]) relies on specific stability conditions that do not preserve the necessary symmetries (like Verdier duality) required to prove the inversion formula. The authors note that a different approach (potentially using Toda's limit stability conditions) might be needed, but current technical limitations in the framing functor setup prevent this.
Full Equivariant Rationality: The equivariant result (Theorem 1.8b) is proven only for truncated homology ( $N < \infty$ ). The authors suspect the full non-truncated equivariant generating function might not be rational due to the degree of poles growing with the truncation level $N$ .

6. Significance

This paper represents a major advancement in the enumerative geometry of 3-folds.

Unification: It unifies the study of PT invariants across Calabi-Yau, Fano, and general projective 3-folds under the umbrella of "superpositive" classes.
Methodological Shift: It demonstrates the power of Joyce's vertex algebra/Lie algebra machinery in solving rationality problems that were previously intractable using standard geometric methods.
Foundation for Computation: By establishing the piecewise quasi-polynomial nature of the invariants, the paper provides a concrete algorithmic path for computing these invariants for specific classes, which is further explored in a sequel by the first author.
Independence: The work runs parallel to and independently confirms similar results by Karpov and Moreira (who use a K-theoretic approach), reinforcing the validity of the rationality conjecture in this regime.

In summary, Anderson and Joyce have successfully proven the rationality and pole structure of Pandharipande–Thomas invariants for a broad and significant class of curve classes on projective 3-folds, utilizing a deep interplay between moduli space geometry, vertex algebras, and wall-crossing phenomena.