Elementary Knot Theory

Marc LackenbyOxford 2017 · DOI · arXiv

Abstract

This is an elementary survey of knot theory. While the field is well-developed, there remain several notoriously intractable problems about knots and links, many of which are surprisingly easy to state. The aim of this paper is to highlight what we still do not understand, as well as to provide a brief survey of what is known regarding algorithmic recognition and fundamental knot properties.

The Equivalence Problem

Knot theory—the study of closed loops in three-dimensional space—has evolved into a sophisticated discipline with deep connections to quantum physics and hyperbolic geometry. Despite this progress, Marc Lackenby argues that we still struggle with the most "elementary" questions. The central challenge is the Equivalence Problem: Given two diagrams of tangled string, is there a foolproof algorithm to decide if they represent the same knot?

While theoretical algorithms exist, they are often computationally massive, involving towers of exponentials. Current research seeks to bridge the gap between these theoretical solutions and practical, efficient algorithms, while cataloging the fundamental properties of knots that remain unproven.

Methodologies in Knot Recognition

Modern knot theory synthesizes several decades of diverse mathematical approaches to tackle the equivalence problem:

  • Normal Surface Theory: This combinatorial method, pioneered by Wolfgang Haken, uses triangulations of 3D space to identify "essential" surfaces that define a knot's structure. By decomposing the space into tetrahedra, researchers can analyze how surfaces intersect these cells.
A normal square in a tetrahedron, a fundamental component of normal surface theory.
  • Geometric Structures: Utilizing Thurston’s Geometrization Conjecture, knot complements are treated as hyperbolic manifolds. This allows for the use of isometry-testing algorithms to distinguish knots.
  • Combinatorial Moves: Researchers analyze Reidemeister moves (local diagram changes) and Pachner moves (triangulation changes) to establish upper bounds on the "distance" between two equivalent knot descriptions.
The knot $5_1$ and its associated hierarchy of surfaces used in algorithmic recognition.
  • Complexity Theory: Knot problems are categorized into classes like NP (where a "yes" answer can be verified quickly) and co-NP (where a "no" can be verified quickly), providing a framework for understanding their computational difficulty.

Fundamental Insights and Open Questions

Several key insights have emerged, yet they often lead to deeper mysteries:

  • The Unknot is Special: Recognizing the "unknot" (a simple circle) is significantly easier than general knot recognition. It is currently known to be in both NP and co-NP, which strongly suggests a polynomial-time algorithm might exist, though one has yet to be discovered.
  • The Additivity Mystery: One of the most intuitive yet unproven conjectures is that the crossing number of a connected sum of two knots is equal to the sum of their individual crossing numbers ($c(K_1 \# K_2) = c(K_1) + c(K_2)$). Currently, we only have a proof that the total is at least $1/152$ of the sum.
  • The Power of Alternating Knots: For knots where the string consistently alternates "over-under," most elementary problems—such as crossing number and primality—are solved. The challenge remains in extending this clarity to non-alternating knots.

Recent Breakthroughs

Recent work has provided significant improvements in complexity bounds:

  • Polynomial Bounds for the Unknot: A major breakthrough is the provision of a polynomial upper bound on the number of Reidemeister moves needed to untangle an unknot. For an $n$-crossing diagram, it can be untangled in at most $(236n)^{11}$ moves. While the exponent is high, this is a landmark shift from previous exponential bounds.
  • Satellite Knot Bounds: For a satellite knot $K$ with companion $L$, the crossing number $c(K)$ is proven to be at least $10^{-13}c(L)$. This establishes a rigorous lower bound where none previously existed.

The Road Ahead

Despite these advancements, several limitations persist. Most algorithms for general knot equivalence remain too slow for practical use on complex knots. Finding a truly efficient (polynomial-time) algorithm for general recognition remains the "holy grail" of the field.

Furthermore, we still lack a confirmed algorithm to compute the unknotting number—the minimum number of times a string must pass through itself to untie a knot. Finally, the Jones Polynomial remains a mystery: it is still unknown if there exists a non-trivial knot that this powerful invariant mistakenly identifies as an unknot.

Summary updated

Abstract

This is an elementary survey of knot theory. It is a well developed, but there remain several notoriously intractable problems about knots and links, many of which are surprisingly easy to state. The aim is to highlight what we still do not understand, as well as to provide a brief survey of what is known.
BibTeX
@inproceedings{lackenby2017elementary,
  author = {Marc Lackenby},
  title = {{Elementary Knot Theory}},
  booktitle = {Oxford},
  year = {2017},
  doi = {10.1093/oso/9780198784913.003.0002},
}

Loading related content...

Figure