On the neighborhood of knots

TL;DR

This paper introduces a novel framework for analyzing knots through their neighborhoods in the space of curves, using knotoid spectra and extended invariants to distinguish knots and explore their properties.

Contribution

It develops the concept of knotoid spectra for knots, demonstrating their ability to distinguish knots with large Gordian distance and to extend invariants for better discrimination.

Findings

Pure knotoids can distinguish knots with Gordian distance greater than one.

Neighborhood invariants can distinguish the unknot from non-trivial knots satisfying the cosmetic crossing conjecture.

Invariants extended to neighborhoods may outperform traditional invariants in knot discrimination.

Abstract

This manuscript introduces a new framework for the study of knots by exploring the neighborhood of knot embeddings in the space of simple open and closed curves in 3-space. The latter gives rise to a knotoid spectrum, which determines the knot type via its knot-type knotoids. We prove that the pure knotoids in the knotoid spectra of a knot, which are individually agnostic of the knot type, can distinguish knots of Gordian distance greater than one. We also prove that the neighborhood of some embeddings of the unknot can be distinguished from any embedding of any non-trivial knot that satisfies the cosmetic crossing conjecture. Topological invariants of knots can be extended to their open curve neighborhood to define continuous functions in the neighborhood of knots. We discuss their properties and prove that invariants in the neighborhood of knots may be able to distinguish more knots than their application to the knots themselves. For example, we prove that an invariant of knots that fails to distinguish mutant knots (and mutant knotoids), can distinguish them by their neighborhoods, unless it also fails to distinguish non-mutant pure knotoids in their spectra. Studying the neighborhood of knots opens the possibility of answering questions, such as if an invariant can detect the unknot, via examining possibly easier questions, such as whether it can distinguish height one knotoids from the trivial knotoid.