Integrals of stable envelopes for cotangent bundles to Grassmannians

This paper computes the $\mathbb{C}^*_\hbar$-equivariant integral of cohomological stable envelopes for cotangent bundles to Grassmannians by deriving a new combinatorial formula for the resulting integer coefficients, which generalize binomial coefficients and are conjectured to relate to curve counting phenomena in 3d mirror symmetry.

The Gist

Imagine you are standing in a vast, multi-dimensional garden called The Grassmannian. This garden isn't made of flowers, but of geometric shapes and symmetries. In this garden, there are special "landmarks" called fixed points. These are spots that don't move even when you spin the whole garden around.

Now, imagine a powerful force called a Stable Envelope. Think of this as a magical spotlight or a "shadow-casting" tool. When you shine this spotlight on one specific landmark, it casts a complex, multi-layered shadow across the entire garden. This shadow isn't just a flat shape; it's a mathematical object that holds deep secrets about the garden's structure.

The authors of this paper, Matthew, Kartik, and Reese, are mathematicians who asked a very specific question: "If we measure the total 'amount' of this shadow, what number do we get?"

Here is the breakdown of their journey, explained simply:

1. The Problem: Measuring the Unmeasurable

Usually, calculating the total "size" of these shadows is incredibly hard because the garden is infinite and the math involves complex, moving variables (like wind speed or temperature).

• The Trick: The authors decided to "freeze" the wind. They turned off all the moving parts of the garden except for one specific dial (called $\hbar$).
• The Result: When they did this, the messy, infinite shadows collapsed into simple, clean numbers. These numbers are always integers (whole numbers like 1, 2, 3) multiplied by a power of that dial.

2. The Discovery: A New Kind of Pascal's Triangle

When the garden is small (specifically, when $k=1$), the numbers they found are the famous Binomial Coefficients. You might know these as the numbers in Pascal's Triangle:

• Row 1: 1
• Row 2: 1, 1
• Row 3: 1, 2, 1
• Row 4: 1, 3, 3, 1

These numbers are famous because each one is the sum of the two numbers directly above it. It's a simple, beautiful pattern.

The Big Surprise:
When the authors looked at larger, more complex gardens (where $k > 1$), they found that the shadows still produced whole numbers! But these numbers didn't follow the simple "sum of two above" rule.

• They discovered a new, complex pattern.
• Instead of adding two neighbors, these new numbers seem to follow a rule where you add four neighbors (or even more, depending on the size of the garden).
• They call this a "2k-neighbor addition" rule. It's like a 3D version of Pascal's Triangle, or a pyramid of numbers where each block is built from a cluster of blocks above it.

3. The Analogy: The "Shadow Garden"

Imagine you have a stack of transparent sheets.

• Sheet 1 (Simple): You draw a simple triangle. The numbers are 1, 2, 1.
• Sheet 2 (Complex): You draw a pyramid. The numbers are bigger and more scattered.
• The Magic: The authors found a formula to predict exactly what number goes in any spot on any sheet, just by knowing the coordinates of that spot.

They also found that if you arrange these numbers correctly, they form a shape that can be divided by a specific algebraic "key" (like a common factor), revealing an even simpler, "reduced" pyramid underneath.

4. Why Does This Matter? (The Mirror World)

The paper mentions 3D Mirror Symmetry. Think of this as a magical mirror.

• On one side of the mirror, you have our garden (the Grassmannian) with its complex shadows.
• On the other side of the mirror, there is a completely different world (the "Mirror Dual").
• The authors believe that the whole numbers they calculated are actually counting something in that mirror world.
• Specifically, they might be counting the number of ways to draw specific curves or paths in the mirror universe. It's like solving a puzzle on one side of the mirror by looking at the reflection on the other side.

5. The Limits and the Future

The authors tried to extend this magic to other types of gardens (called "Bow Varieties").

• The Good News: For most standard gardens (Quiver Varieties), the magic works perfectly. The shadows always collapse into nice whole numbers.
• The Bad News: For some weird, twisted gardens (Bow Varieties), the magic breaks. The shadows don't collapse into whole numbers; they explode into messy fractions.
• The Conjecture: They suspect that the gardens where the magic works are exactly the ones that are "equivalent" to standard gardens. If the shadow calculation fails, it's a sign that the garden is too twisted to have a simple mirror twin.

Summary

In short, this paper is about finding order in chaos.

1. They took a complex geometric object (a shadow in a high-dimensional garden).
2. They simplified it to find that it always results in a whole number.
3. They discovered these numbers follow a new, complex version of Pascal's Triangle.
4. They believe these numbers are counting hidden paths in a parallel "mirror" universe.

It's a bit like finding that if you count the leaves on a very strange tree in a specific way, you always get a number that tells you how many birds are nesting in a tree on the other side of the planet. The math is the bridge between the two worlds.

Technical Summary

1. Problem Statement and Context

The paper investigates the cohomological stable envelopes introduced by Maulik and Okounkov for the cotangent bundle to the Grassmannian, denoted as $X_{k,n} = T^*\text{Gr}(k, n)$. These varieties are fundamental examples of Nakajima quiver varieties and play a central role in 3D mirror symmetry, where stable envelopes relate enumerative geometry on a space $X$ to its mirror dual $X^\vee$.

The specific problem addressed is the computation of the non-equivariant limit of the integral of a stable envelope class.

• The Integral: The authors define the integral $\int_{X_{k,n}} \text{Stab}(p)$ via equivariant localization over a subtorus $C^*_\hbar \subset T$, where $T = (C^*)^n \times C^*_\hbar$.
• The Limit: They study the limit of this integral as the equivariant parameters $a = (a_1, \dots, a_n)$ associated with the torus $(C^*)^n$ approach zero ($a \to 0$).
• The Goal: To derive a closed-form combinatorial formula for the resulting integers (which are multiples of powers of $\hbar$) and to understand their combinatorial structure, particularly as generalizations of binomial coefficients.

2. Methodology

The authors employ a multi-step approach combining algebraic geometry, equivariant localization, and asymptotic analysis of special functions.

A. Equivariant Localization

Since the projection $X_{k,n} \to \text{pt}$ is not proper, the integral is defined via localization.

1. Reduction to Fixed Points: The integral over $X_{k,n}$ is computed by localizing over the full torus $T$. By the Atiyah-Bott localization formula, the integral is expressed as a sum over the fixed points $p_J \in X_{k,n}^T$:
   $$\int_{X_{k,n}} \text{Stab}(p_I) = \sum_{p_J \in X_{k,n}^T} \frac{\text{Stab}(p_I)|_{p_J}}{e(T_{p_J}X_{k,n})}$$
   where $e(T_{p_J}X_{k,n})$ is the equivariant Euler class of the tangent space.
2. Weight Functions: Instead of computing stable envelopes directly, the authors utilize weight functions $W(p_I)$ introduced by Rimanyi, Tarasov, and Varchenko. These rational functions represent the stable envelope classes in equivariant cohomology.
3. Limit Computation: The authors compute the limit of the localization sum as $a \to 0$. They prove that despite the denominators containing terms like $(a_i - a_j)$ which vanish in the limit, the sum of rational functions simplifies to a well-defined integer (times a power of $\hbar$).

B. Asymptotic Analysis

To evaluate the limit rigorously, the authors introduce a scaling substitution $a_s = s z \hbar$ and take the limit $z \to 0$ (equivalently $u = 1/z \to \infty$).

• They express the summands using Gamma functions arising from the product terms in the weight functions and Euler classes.
• Using the asymptotic expansion of the ratio of Gamma functions $\frac{\Gamma(u+a)}{\Gamma(u+b)} \sim u^{a-b}$, they isolate the coefficient of the leading term ($u^0$) to extract the integer value.

C. Combinatorial Proofs

The paper includes a direct combinatorial proof (Appendix A) demonstrating that the non-equivariant limit exists. This involves showing that the sum of terms vanishes whenever $a_x = a_y$, implying that the Vandermonde determinant divides the numerator, thus canceling the singularities in the denominator.

3. Key Contributions and Results

A. Main Theorem: Closed-Form Formula

The primary result is Theorem 3.1, which provides an explicit combinatorial formula for the integral:
$$\int_{X_{k,n}} \text{Stab}(p_I) = (-1)^{k(n-k)-|I|} \sum_{J \ge I} A_J B_J$$
where the sum is over fixed points $p_J$ in the attracting set of $p_I$ (ordered by a specific partial order), and $A_J, B_J$ are explicitly defined combinatorial terms involving factorials, multinomial coefficients, and generalized Bernoulli polynomials.

• Integrality: The authors prove that this rational expression always evaluates to an integer times $\hbar^{k(n-k)}$.
• Recovery of Binomials: For the case $k=1$ (cotangent bundle to projective space), the formula recovers the standard binomial coefficients $\binom{n}{i-1}$.

B. Combinatorial Interpretations

The paper explores the combinatorial structure of these integers for $k > 1$:

1. Generalized Pascal's Simplex: The authors conjecture that the integers for $X_{k,n}$ form a $(k+1)$-simplex structure. They propose a "2k-neighbor addition" recurrence relation, where an entry in layer $n$ is the sum of $2k$ "nearby" entries from layer $n-1$, potentially with correction terms for boundary cases.
2. Case $k=2$: They prove the recurrence for $k=2$, showing that the integers form a 3-simplex (pyramid) structure. They introduce a "reduced" simplex where the correction terms vanish, obeying a clean 4-neighbor addition rule.
3. Log-Concavity: They conjecture that specific subsequences of these integers are log-concave, generalizing a known property of binomial coefficients.

C. Varchenko's Path Conjecture

The paper outlines a new conjectural method (attributed to A. Varchenko) for computing these integrals using paths to Young diagrams.

• For a partition $\lambda$ fitting in an $(n-k) \times k$ box, they define a sum $V(\lambda)$ over paths removing boxes from $\lambda$ and its complement.
• They conjecture that $V(\lambda)$ equals the localization sum for the corresponding fixed point, providing a purely combinatorial alternative to the localization formula.

D. Generalizations to Type A Varieties

The authors investigate whether these results extend to other Type A varieties:

• Quiver Varieties: They conjecture that the non-equivariant limits exist for all Type A Nakajima quiver varieties.
• Bow Varieties: They find that for general bow varieties, the non-equivariant limit does not necessarily exist.
• Hanany-Witten Equivalence: They propose a striking conjecture (Conjecture 5.2): A bow variety admits non-equivariant limits for all stable envelope integrals if and only if it is Hanany-Witten equivalent to a quiver variety. This suggests a deep link between the analytic existence of these limits and the geometric classification of the varieties.

4. Significance

• Mathematical Physics: The results provide concrete data for 3D mirror symmetry. The non-equivariant limits of stable envelopes are expected to correspond to curve counting invariants (vertex functions) on the mirror dual.
• Combinatorics: The discovery of a new family of integers generalizing binomial coefficients, organized into higher-dimensional simplices with complex recurrence relations, opens new avenues in algebraic combinatorics.
• Geometric Classification: The conjecture linking the existence of these limits to Hanany-Witten equivalence offers a potential new criterion for distinguishing between different classes of symplectic resolutions (quiver vs. bow varieties).
• Methodological Advancement: The paper demonstrates a robust technique for computing non-equivariant limits of stable envelopes by combining localization with asymptotic analysis of Gamma functions, a method that could be applicable to other equivariant integrals in geometric representation theory.

In summary, this paper bridges the gap between the abstract theory of stable envelopes and concrete combinatorial formulas, revealing rich arithmetic structures and proposing deep connections between the geometry of quiver/bow varieties and the analytic properties of their associated integrals.