Transverse knots determined by their cyclic branched covers

This paper constructs new examples of non-isotopic transverse knots with contactomorphic cyclic branched covers while simultaneously proving that the transverse isotopy classes of many other transverse knots are uniquely determined by the contactomorphism type of their cyclic branched covers.

The Gist

Imagine you have a piece of string tied into a knot. In the world of mathematics, specifically in a field called contact topology, we don't just look at the shape of the knot; we also look at how it "twists" through an invisible, swirling wind field that fills the entire universe (mathematicians call this a contact structure).

When a knot moves through this wind in a specific, positive direction, we call it a Transverse Knot.

This paper asks a fascinating question: Can you identify a knot just by looking at the "shadows" it casts?

The Shadow Analogy: Cyclic Branched Covers

To understand the "shadows" the authors are talking about, imagine you have a magical camera. Instead of taking a flat photo, this camera takes a 3D hologram of the knot, but it wraps the universe around the knot $n$ times.

• The Knot: Your original string.
• The Shadow (Cyclic Branched Cover): A new, complex 3D world created by wrapping space around your knot $n$ times.
• The Contact Structure: The "wind" inside this new shadow world.

The authors are asking: If two different knots cast the exact same "windy shadow" (are contactomorphic), are the original knots actually the same?

The Main Discovery: Some Knots are Unique, Others are Imposters

The paper finds that the answer depends entirely on the type of knot you are looking at.

1. The "Honest" Knots (The Determined Ones)

Some knots are so unique that their shadows are like fingerprints. If you see a specific shadow, you know exactly which knot made it.

• The Torus Knots: Think of these as knots wrapped neatly around a donut (like a pretzel). The authors prove that if you take a torus knot and wrap space around it 3 or more times, the resulting shadow is unique to that specific knot. No other knot can fake it.
• The Figure-Eight Knot: This is another famous knot. The authors show that every time you wrap space around it (for any number of times), the shadow is unique. It's impossible to trick the shadow.
• The General Rule: For most "prime" knots (knots that can't be broken down into smaller knots), there are only a very few specific numbers of wraps where a "fake" knot could mimic the shadow. If you check the shadows for just three different numbers of wraps, you can be 100% sure of the knot's identity.

2. The "Imposters" (The Twins)

On the flip side, the authors also found knots that are smooth imposters.

• Imagine two knots, Knot A and Knot B. They are made of different strings and are twisted differently in the wind (so they are mathematically distinct).
• However, when you put them through the "magic camera" to create their shadows, the shadows look identical.
• The authors constructed a specific pair of knots (using a method involving two linked loops) that are smooth imposters. They look different in the real world, but their $n$-fold shadows are indistinguishable.

The "Twins" Concept

The authors use the term "Twins" to describe these pairs.

• Transverse Twins: Two knots that are different but cast the same shadow.
• The Big Surprise: While we knew some knots could be twins, this paper shows that for many famous knots (like the torus knots), twins do not exist. They are "orphaned" in the sense that no other knot can mimic their shadow.

Why Does This Matter?

Think of it like a security system.

• Old Security: We knew that some knots could break the security system by wearing a "shadow mask" (being twins).
• New Security: This paper proves that for many high-security knots (like the torus knots), the shadow mask doesn't work. The shadow is a perfect, unforgeable ID card.

Summary in a Nutshell

1. The Setup: We study knots in a swirling wind field.
2. The Test: We wrap the universe around the knot $n$ times to create a "shadow world."
3. The Result:
   • For Torus Knots and Figure-Eight Knots, the shadow is a perfect ID. If the shadows match, the knots are identical.
   • For Composite Knots (knots made of smaller knots), the rule is slightly more complex, but generally, the shadow is still a very strong identifier.
   • However, the authors also built a counter-example: a pair of knots that are different but have identical shadows, proving that the "shadow test" isn't perfect for every knot in the universe.

In short, the paper maps out the "fingerprint zone" of knots, showing us which knots are so unique that their shadows give them away, and which knots are sneaky enough to wear a disguise.

Technical Summary

1. Problem Statement and Context

The paper investigates the extent to which the contactomorphism type of the $n$-fold cyclic branched cover of a transverse knot determines the transverse isotopy class of the knot itself.

• Setting: The study takes place in the standard tight contact 3-sphere $(S^3, \xi_{\text{std}})$.
• Object: For a transverse knot $T$, $\Sigma_n(T)$ denotes the $n$-fold cyclic branched cover of $S^3$ along $T$. This manifold carries a preferred contact structure $\xi_n(T)$.
• Core Question: If two transverse knots $T$ and $T'$ have contactomorphic cyclic branched covers $(\Sigma_n(T), \xi_n(T)) \cong (\Sigma_n(T'), \xi_n(T'))$, are $T$ and $T'$ necessarily transversely isotopic?
• Background: Harvey–Kawamuro–Plamenevskaya [HKP09] previously showed that the answer is generally no: there exist non-isotopic transverse knots (even with the same smooth type and self-linking number) that share contactomorphic cyclic branched covers for all $n > 1$.
• Goal: This paper aims to identify specific classes of knots where the answer is yes (determination holds) and to construct new examples where the answer is no (non-determination), specifically focusing on knots that are smoothly non-isotopic.

2. Methodology and Tools

The authors employ a combination of contact topology, knot theory, and 3-manifold topology.

• Homotopical Invariants:
  • $d_3$-invariant: Used to distinguish contact structures on rational homology spheres.
  • $\Gamma$-invariant (Gompf): A refinement of the Euler class for 2-plane fields, dependent on spin structures.
  • Tristram–Levine Signatures: Used to compute the $d_3$-invariant of the branched cover via Lemma 3.1.
• Surgery Diagrams: The authors utilize contact $(\pm 1)$-surgery diagrams to represent the cyclic branched covers. They rely on the algorithm from [HKP09] which converts braid words representing transverse knots into Legendrian surgery diagrams.
• Connected Sum Decomposition:
  • Lemma 4.1: Establishes that the cyclic branched cover of a connected sum $T_1 \# T_2$ is the connected sum of the individual cyclic branched covers: $\Sigma_n(T_1 \# T_2) \cong \Sigma_n(T_1) \# \Sigma_n(T_2)$.
  • This allows the reduction of problems involving composite knots to their prime factors.
• Nakanishi–Sakuma Construction: Adapted from the smooth setting [Nak81, Sak81], this construction uses a two-component link $L = L_1 \cup L_2$ with linking number $\pm 1$ to generate two distinct knots $K_1, K_2$ that share the same cyclic branched cover.
• Alexander Polynomials: Used to distinguish the smooth isotopy types of the constructed knots in the counter-example section.

3. Key Contributions and Results

A. Determination Results (Positive Cases)

The paper proves that for many "simple" or "prime" knots, the transverse isotopy class is determined by the contactomorphism type of their cyclic branched covers.

• Theorem 1.1 (Prime, Transversely Simple Knots):

  • If $K$ is a prime, transversely simple knot, then any transverse realization $T$ admits at most two odd primes $p$ for which a transverse $p$-twin (a non-isotopic knot with the same cover) exists.
  • For any specific odd prime $p$, there are at most finitely many transverse $p$-twins.
  • Corollary: $T$ is uniquely determined by the contactomorphism types of its cyclic branched covers for three distinct odd primes.

• Theorem 1.2 (Torus Knots):

  • Transverse torus knots admit no transverse $n$-twins for any $n \ge 3$. They are fully determined by their $n$-fold cyclic branched covers for $n \ge 3$.

• Theorem 1.3 (Figure-Eight Knot):

  • The transverse figure-eight knot admits no transverse twin for any $n \ge 2$.

• Theorem 1.5 (Connected Sums of Positive Torus Knots):

  • A transverse realization of a connected sum of positive torus knots (with maximal self-linking number) admits no transverse $n$-twins for any $n \ge 3$.

B. Non-Determination Results (Counter-Examples)

The paper constructs new examples of non-isotopic transverse knots sharing cyclic branched covers, extending previous work.

• Theorem 1.6 (Smoothly Non-Isotopic Twins):

  • For every $n \ge 2$, there exist hyperbolic transverse knots $T_1$ and $T_2$ that are smoothly non-isotopic but share contactomorphic $n$-fold cyclic branched covers.
  • Construction: Based on a Legendrian link $L$ (derived from a $T(2,4)$ torus link with a band). The authors compute the Alexander polynomials of the resulting knots $T_1$ and $T_2$ and show they differ (specifically, the breadth of the polynomial differs by a factor of $n$), proving they are not smoothly isotopic.

• Theorem 1.7 (Overtwisted Covers):

  • If two transverse knots are smoothly isotopic and have the same self-linking number, their $n$-fold cyclic branched covers have the same homotopical invariants ($d_3$ and $\Gamma$).
  • Consequently, if these covers are overtwisted, they are contactomorphic. This provides a mechanism for generating contactomorphic covers from smoothly isotopic knots.

4. Significance and Implications

1. Refining the Classification of Transverse Knots: The paper clarifies the boundary between when topological invariants (branched covers) are sufficient to classify transverse knots and when they fail. It shows that while "twins" exist, they are rare for prime knots and specific classes like torus knots.
2. Smooth vs. Transverse Distinction: The construction in Theorem 1.6 is significant because it produces knots that are not just transversely non-isotopic (like the examples in [HKP09]) but smoothly non-isotopic. This demonstrates that the contact structure on the branched cover can fail to detect even the underlying smooth knot type in certain hyperbolic cases.
3. Connection to Smooth Knot Theory: The results heavily rely on the resolution of the Smith Conjecture and the classification of smooth twins (e.g., prime knots having at most two smooth $p$-twins). The paper successfully translates these smooth constraints into the transverse contact setting.
4. Computational Invariants: The paper provides explicit computational methods using the $d_3$-invariant and Alexander polynomials to distinguish or identify transverse twins, offering practical tools for future research in contact topology.

5. Open Questions

The authors conclude by listing several open questions, including:

• Whether there exists a transversely non-simple knot where every realization is determined by a cyclic branched cover.
• Whether non-isotopic transverse knots with the same smooth type and self-linking number can have non-contactomorphic cyclic branched covers for some $n$.
• The existence of "transversely universal" knots determined by their branched covers.

In summary, this paper establishes that while cyclic branched covers do not universally determine transverse knots, they are a powerful invariant for prime and simple knots, while simultaneously providing a robust construction for generating counter-examples involving hyperbolic, smoothly distinct knots.