Gang-Kim-Yoon integrality conjectures on adjoint Reidemeister torsions for torus knots

This paper proves the Gang-Kim-Yoon integrality conjecture for all torus knots and non-negative integers $g$ by introducing Verlinde numbers derived from the modular S-matrix, establishing their recursion formulas, and demonstrating how adjoint Reidemeister torsions can be recovered from the Hessian of a birational model of the character variety.

The Gist

Imagine you have a complex, knotted piece of string floating in space. In the world of mathematics, this is called a torus knot. Now, imagine trying to understand the "shape" of the empty space surrounding this knot. Mathematicians use special tools called Reidemeister torsions to measure the "twist" and "tension" of this invisible space.

Think of these torsions like the unique "fingerprint" or "vibe" of the knot's surrounding space. If you look at the knot from different angles (represented by different mathematical representations), you get different values for this twist.

The Big Mystery

A few years ago, a group of mathematicians (Gang, Kim, and Yoon) made a bold guess, or conjecture. They wondered: If you take all these different "twist" values, raise them to a specific power, and add them all up, do you get a whole number?

In the real world, adding up measurements often gives you messy decimals (like 3.14159...). But in this mathematical universe, they suspected the answer would always be a clean, whole number (like 1, 2, or 100), no matter how complex the knot or how high the power you chose.

The Solution: A New Kind of "Recipe"

In this paper, authors Yuji Terashima and Yoshikazu Yamaguchi prove that this guess is true for all torus knots. They didn't just check a few examples; they found a universal rule that works for every single one.

Here is how they did it, using some creative mathematical "tools":

1. The "Magic Matrix" (The S-Matrix)
To solve the puzzle, the authors introduced a special grid of numbers called a modular S-matrix. Think of this matrix as a giant, magical recipe book. In physics, similar books are used to predict how particles interact. Here, the authors adapted this "recipe book" specifically for knots. It helps translate the messy, twisting geometry of the knot into a structured list of numbers.

2. The "Verlinde Numbers" (The Counting Game)
Using this recipe book, they defined new numbers called Verlinde numbers. You can think of these as a special way of counting the "energy" or "weight" of the knot's space.

• The Analogy: Imagine you have a bag of marbles, each with a different color and weight. The Verlinde number is a specific way of weighing the whole bag. The authors showed that if you follow their specific counting rules, the total weight always comes out to a whole number.

3. The "Blowing Up" Trick (Geometry)
To make sense of the knot's shape, the authors used a technique called "blowing up."

• The Analogy: Imagine a crumpled piece of paper with a sharp point (a singularity). If you gently blow air into that point, it smooths out into a nice, round surface. The authors did this mathematically with the knot's shape. They turned a jagged, singular curve (called a Chebyshev curve) into a smooth, clean surface.
• On this smooth surface, they found that the "twist" of the knot (the Reidemeister torsion) is directly related to the curvature of the surface at specific points. It's like measuring how bumpy a hill is to determine how fast a ball would roll down it.

4. The "Recursive Ladder" (The Proof)
The final piece of the puzzle was a recursion formula.

• The Analogy: Imagine a ladder. To know the height of the 10th rung, you don't need to measure from the ground every time; you just need to know the height of the 9th rung and add the height of one step.
• The authors showed that the "Verlinde numbers" for a complex knot (a high rung) can be built up step-by-step from simpler numbers (lower rungs).
• They proved that the very first step (the bottom rung) is always a whole number (specifically, 1). Because every step up the ladder preserves this "whole number" quality, the final answer at the top must also be a whole number.

The Conclusion

The paper confirms that for any torus knot, if you take the "twist" measurements, power them up, and add them together, the result is always an integer.

They achieved this by:

1. Smoothing out the knot's geometry to see its true shape.
2. Using a "recipe book" (S-matrix) to translate geometry into numbers.
3. Showing that these numbers follow a strict "ladder" rule that guarantees the final sum is always a whole number.

This discovery connects the abstract world of knot geometry with the structured world of number theory, showing that even in the most twisted spaces, there is an underlying order that results in clean, whole numbers.

Technical Summary

Technical Summary: Gang-Kim-Yoon Integrality Conjectures on Adjoint Reidemeister Torsions for Torus Knots

Problem Statement
The paper addresses a conjecture proposed by Gang, Kim, and Yoon (GKY) regarding the integrality of sums of adjoint Reidemeister torsions. Specifically, for a 3-manifold and a non-negative integer $g$, the conjecture posits that the sum of the $(g-1)$-st powers of the adjoint Reidemeister torsions, taken over irreducible representations of the fundamental group on a specific level set of the trace function, yields an integer. This conjecture is grounded in the 3d–3d correspondence, which relates the geometry of a 3-manifold to 3-dimensional gauge theory, suggesting that this sum corresponds to the index of a gauge theory, which is inherently integral. While the conjecture is motivated by physical dualities, a rigorous mathematical proof for general cases was lacking. This paper focuses on proving the conjecture for the complements of torus knots for all $g \geq 0$.

Methodology
The authors employ a combination of algebraic geometry, representation theory, and combinatorial techniques derived from conformal field theory:

1. Birational Models via Chebyshev Curves: The authors construct a rational model for the character variety $X(G_{p,q})$ of a $(p,q)$-torus knot group. They utilize Chebyshev curves defined by $F(X, Y) = C_p(X) - C_q(Y) = 0$, where $C_k$ are Chebyshev polynomials of the first kind. By blowing up the singular points of these curves in $\mathbb{C}^2$, they obtain a resolution $Y_{p,q}$ consisting of a non-singular curve $D_{p,q}$ and exceptional lines. They establish that the reducible part of the character variety is birational to $D_{p,q}$, while the irreducible part is isomorphic to $Y_{p,q} \setminus D_{p,q}$.

2. Recovering Torsion from Hessians: A key technical step involves recovering the adjoint Reidemeister torsion $T_\mu(\rho)$ from the geometry of the Chebyshev curve. The authors demonstrate that $T_\mu(\rho)$ on a component corresponding to a critical point $(2\cos(\pi a/p), 2\cos(\pi b/q))$ is proportional to the Hessian (or Jacobian determinant) of the defining polynomial $F(X,Y)$ at that point. Specifically, $T_\mu(\rho) = -\frac{1}{4pq} \text{Hess}(F)$ at the critical point.

3. Modular S-Matrices and Verlinde Numbers: The authors introduce a modular S-matrix for torus knots, defined as $S_{(i,j),(a,b)} = \sqrt{\frac{8}{pq}} \sin(\frac{ia\pi}{p}) \sin(\frac{jb\pi}{q})$. This matrix is a generalization of the modular S-matrix found in $SU(2)$ Wess-Zumino-Witten (WZW) models. Using this matrix, they define "Verlinde numbers" $d(g, n)$ for torus knot exteriors, which generalize the classical Verlinde numbers. The sum of the torsion powers in the conjecture is shown to be proportional to $d(g, 0)$.

4. Recursion and Fusion Rules: To prove integrality, the authors derive recursion formulas (Fusion rules) for the Verlinde numbers $d(g, n)$. These rules allow the computation of $d(g, n)$ for higher genera $g$ based on lower genus values and specific initial conditions.

Key Contributions and Results

• Proof of Integrality: The primary result is the proof that for any $(p,q)$-torus knot exterior and any $g \geq 0$, the sum $\sum_{[\rho] \in \text{tr}_\mu^{-1}(c)} (2 T_\mu(\rho))^{g-1}$ is an integer.
• Initial Values and Recursion: The authors establish that the sequence of these sums starts with the value 1 at $g=0$ (i.e., $\sum (2 T_\mu(\rho))^{-1} = 1$). They prove that the Verlinde numbers $d(g, 0)$ lie in the set $(1/2)^{g-2}\mathbb{Z}$. Combined with the factor $2^{g-2}$ in the definition of the sum, this ensures the final result is an integer.
• Geometric Interpretation: The paper provides a concrete geometric realization of the character variety for torus knots via the resolution of Chebyshev curves and explicitly links the adjoint Reidemeister torsion to the Hessian of the defining polynomial of these curves.
• Generalization of WZW Models: The work extends the framework of Verlinde numbers and modular S-matrices from standard WZW models to the context of torus knot complements, offering a new algebraic structure for these topological objects.

Significance
The paper claims significance primarily in the mathematical verification of the GKY integrality conjecture for the specific and important class of torus knots. By proving the conjecture, the authors validate the connection between the index of 3-dimensional gauge theories and the sum of adjoint Reidemeister torsions in this context.

The work also contributes to the broader program of associating modular tensor categories to 3-manifolds and establishing correspondences between knots and gauge theories (knot-quiver and knot-gauge correspondences). The introduction of the modular S-matrix for torus knots and the birational model of their character varieties provides tools that may be applicable to further research linking knot theory with gauge theory physics. The paper does not propose new experimental applications but rather solidifies the theoretical underpinnings of existing physical conjectures within a rigorous mathematical framework.