The AJ conjecture and connected sums of torus knots

This paper verifies the AJ conjecture for connected sums of torus knots $T(p,q)\#T(a,b)$ where $p$ and $a$ share the same sign, while identifying new cases with repeated factors in the recurrence polynomial that necessitate a slight modification to the conjecture.

The Gist

Imagine you are a detective trying to solve a mystery about knots. But these aren't the shoelace knots you tie in the morning; they are mathematical knots floating in a three-dimensional universe called $S^3$.

This paper, written by Xingru Zhang, is about cracking a specific code that links two different ways of describing these knots. The author proves that for a specific type of knot combination, the code works perfectly—except for one tiny, surprising glitch that forces us to rewrite the rulebook slightly.

Here is the breakdown using everyday analogies:

1. The Two Languages of Knots

To understand the problem, imagine every knot has two "languages" it speaks:

• Language A (The A-Polynomial): Think of this as the knot's DNA. It's a static blueprint that tells you the knot's shape and how it twists. It's like a fingerprint; no two different knots have the exact same A-polynomial.
• Language B (The Colored Jones Polynomial): This is the knot's musical score. It's a complex song that changes depending on how you "listen" to it (mathematically, based on a number $n$). If you play the song, you can hear the rhythm of the knot's structure.

2. The AJ Conjecture: The Great Translator

For decades, mathematicians have suspected a deep connection between these two languages. The AJ Conjecture is the hypothesis that says: "If you take the musical score (Language B), slow it down to a specific tempo (set a variable $t = -1$), and strip away the noise, you will get the DNA blueprint (Language A)."

It's like saying: "If you take a symphony, slow it down to a single note, and remove the reverb, you'll hear the exact melody written in the composer's original notebook."

3. The Experiment: Tying Knots Together

The author focuses on Torus Knots. Imagine a rubber band wrapped around a donut (a torus). If you wrap it $p$ times around the hole and $q$ times around the tube, you get a knot $T(p, q)$.

The paper investigates what happens when you tie two of these knots together in a row (a "connected sum"). It's like taking two different rubber band sculptures and fusing them into one long, complex sculpture.

The author asks: Does the AJ Conjecture still hold when we fuse two knots together?

4. The Discovery: The "Echo" Glitch

The author proves that for most combinations of these knots, the conjecture holds true. The "musical score" does indeed translate perfectly into the "DNA blueprint."

However, the author found a very specific, rare scenario where things get weird.
Imagine you have two knots that look different but happen to have the same "total twist" (mathematically, $pq = ab$). When you tie these specific twins together, the translation process produces a glitch.

• The Glitch: When you translate the music to the DNA, you don't just get the blueprint; you get the blueprint with a repeated echo.
• The Metaphor: Imagine you are trying to identify a person by their voice. Usually, the voice is unique. But in this specific case, the voice sounds like two identical twins speaking in perfect unison. The translation software sees "Twin A" and "Twin A" again, creating a duplicate factor in the DNA code.

5. The Solution: Updating the Rulebook

Because of this "echo," the original AJ Conjecture needed a tiny tweak.

• Old Rule: "The translation equals the DNA."
• New Rule: "The translation equals the DNA, after you remove any duplicate echoes."

The author shows that once you delete the repeated parts of the code, the AJ Conjecture works perfectly again.

Why Does This Matter?

In the world of mathematics, finding a case where a famous rule breaks (or needs a tiny patch) is a huge deal.

1. It's the First Time: The author notes these are likely the first examples of knots that produce this specific "duplicate echo" phenomenon.
2. It Refines Our Tools: It tells mathematicians that their tools for studying knots are robust, but they need to be careful about "repeated factors" when dealing with complex knot combinations.
3. It Solves a Puzzle: It confirms that the deep connection between the "musical score" and the "DNA" of knots is real, even for these complicated fused knots, provided we know how to clean up the static.

Summary

Xingru Zhang took two complex mathematical knots, tied them together, and checked if their musical patterns matched their DNA blueprints. They mostly matched perfectly, but in one special case, the music created a "double echo." The paper proves that if you ignore the echo, the match is perfect, and suggests a small update to the mathematical rulebook to account for this interesting glitch.

Technical Summary

Here is a detailed technical summary of the paper "The AJ Conjecture and Connected Sums of Torus Knots" by Xingru Zhang.

1. Problem Statement

The paper addresses the AJ Conjecture, formulated by Stavros Garoufalidis, which posits a deep relationship between two fundamental knot invariants:

1. The A-polynomial ($A_K(M, L)$): A two-variable polynomial in $\mathbb{Z}[M, L]$ derived from the $SL(2, \mathbb{C})$ character variety of the knot complement.
2. The Recurrence Polynomial ($\alpha_K(t, M, L)$): A non-commutative polynomial in the quantum torus $\mathcal{T}$ that annihilates the sequence of colored Jones polynomials $J_K(n)$.

The Conjecture: For any knot $K$, the recurrence polynomial evaluated at $t = -1$, denoted $\alpha_K(-1, M, L)$, is equal to the A-polynomial $A_K(M, L)$ up to a non-zero factor depending only on $M$.

The Specific Challenge: While the conjecture has been verified for simple knots (2-bridge, torus, pretzel, cabled), its validity for connected sums of knots remains largely open. This paper investigates the conjecture for connected sums of torus knots, $K = T(p, q) \# T(a, b)$. A critical issue arises when the product of the parameters of the two knots is equal ($pq = ab$) but the knots are distinct ($p \neq a$). In these specific cases, the standard formulation of the conjecture appears to fail because $\alpha_K(-1, M, L)$ contains repeated factors involving $L$, whereas the A-polynomial is defined to have no repeated factors.

2. Methodology

The author employs a constructive algebraic approach to verify the conjecture. The methodology consists of three main steps for each case of the connected sum:

1. Construction of a Candidate Annihilator:

   • Using the known recurrence relations for individual torus knots $T(p, q)$ and $T(a, b)$ (derived from previous literature like [14] and [11]), the author derives a recurrence relation for the product $J_{T(p,q)}(n)J_{T(a,b)}(n)$.
   • By applying operators from the quantum torus $\mathcal{T}$ (specifically shift operators $L$ and multiplication operators $M$) and utilizing the connected sum formula for colored Jones polynomials ($[n]J_{K_1 \# K_2}(n) = J_{K_1}(n)J_{K_2}(n)$), a candidate annihilator $\alpha_K(t, M, L)$ is constructed.
   • This involves complex algebraic manipulations to eliminate intermediate terms and isolate a polynomial in $L$ that annihilates the sequence.

2. Verification of the Limit at $t = -1$:

   • The candidate polynomial $\alpha_K(t, M, L)$ is evaluated at $t = -1$.
   • The author compares the resulting polynomial with the known A-polynomial of the connected sum $T(p, q) \# T(a, b)$ (calculated using the formula by Hoste and Shanahan).
   • Crucial Modification: In cases where $pq = ab$ but $p \neq a$, the author demonstrates that $\alpha_K(-1, M, L)$ contains a repeated factor (specifically involving $L$). The paper proposes a modified AJ conjecture: $\alpha_K(-1, M, L)$, after removing all repeated factors, equals $A_K(M, L)$ up to a factor of $M$.

3. Proof of Minimality (L-degree):

   • To confirm that the constructed $\alpha_K$ is indeed the recurrence polynomial (and not just an annihilator), the author must prove that its degree in $L$ is the minimum possible among all non-zero annihilators.
   • This is achieved by assuming a lower-degree annihilator exists, setting up a system of homogeneous linear equations for the coefficients, and proving that the only solution is the trivial one (all coefficients are zero).
   • This step involves analyzing the degrees of the colored Jones polynomials (using formulas for lowest and highest degrees in $t$) and evaluating the resulting linear systems at $t = -1$ (or using limits where direct evaluation yields zero).

3. Key Contributions and Results

Main Theorem (Theorem 1.1):
The AJ conjecture holds for the connected sum of two torus knots $T(p, q) \# T(a, b)$ provided that $p$ and $a$ have the same sign.

Discovery of Repeated Factors:
The paper identifies a novel phenomenon:

• For $K = T(p, q) \# T(a, b)$, if $pq = ab$ but $p \neq a$ (e.g., $T(3, 2) \# T(6, 1)$ is not possible as $q \ge 2$, but $T(4, 3) \# T(6, 2)$ where $12=12$), the recurrence polynomial$\alpha_K(-1, M, L)$contains a **repeated factor** involving$L$.
• Specifically, the factor $(L^2 - M^{-2ab})$ appears twice in the evaluation.
• Significance: These are the first known examples of knots where the recurrence polynomial at $t=-1$ has repeated factors. This necessitates the modification of the AJ conjecture to include the "removal of repeated factors" clause.

Case Analysis:
The proof is divided into seven cases based on the parameters $p, q, a, b$:

1. General Case ($|p|>q>2, |a|>b>2, pq \neq ab$): Standard verification; no repeated factors.
2. Identical Knots ($T(p, q) \# T(p, q)$): Verification holds; degree is 5.
3. Equal Product, Distinct Knots ($pq = ab, p \neq a$): The critical case. The author proves the existence of the repeated factor and verifies the modified conjecture.
4. Mixed Cases ($b=2$ or $q=2$): The author handles cases where one or both knots are of the form $T(p, 2)$ (trefoil-like or similar), covering subcases where $pq = 2a$ or $pq \neq 2a$.
5. Sign Constraints: The results are proven specifically when $p$ and $a$ have the same sign.

4. Significance

• Extension of the AJ Conjecture: The paper significantly broadens the scope of the AJ conjecture from simple knots to connected sums, a more complex class of knots.
• Refinement of the Conjecture: The discovery of repeated factors in the recurrence polynomial is a major theoretical contribution. It suggests that the relationship between the colored Jones polynomial and the A-polynomial is slightly more subtle than originally stated, requiring the removal of repeated factors to match the geometric invariant (A-polynomial).
• Technical Advancement: The paper provides a rigorous framework for handling the algebraic complexity of connected sums, utilizing the quantum torus and recurrence relations in a systematic way. The method of proving minimality via degree analysis and linear system determinants serves as a template for future investigations into other composite knots.
• Counter-Examples to Naive Formulations: By showing that $\alpha_K(-1, M, L)$ is not always square-free, the paper prevents future researchers from incorrectly assuming the conjecture holds in its original, unmodified form for all knots.

In conclusion, Xingru Zhang's paper successfully verifies the AJ conjecture for a large class of composite knots while simultaneously identifying and correcting a subtle flaw in the original formulation of the conjecture regarding repeated factors in the recurrence polynomial.