Knot Invariants from Four-Dimensional Gauge Theory

TL;DR

The paper directly analyzes four-dimensional gauge theory equations to recover the Jones polynomial by counting classical solutions, avoiding reliance on dualities and highlighting a deep link to two-dimensional Virasoro conformal blocks through opers and degenerate fields. By deforming t and introducing controlled symmetry breaking, the authors reduce the problem to three-dimensional Higgs-bundle data and connect to the Gaudin model via Bethe equations, establishing a bridge between 4D gauge theory, Chern-Simons theory, and conformal field theory. They develop a robust framework using Morse theory and thimbles to compute braid-group representations and embed these into free-field conformal blocks, ultimately reproducing Jones-type invariants through a braid-group action on integration cycles. The work also integrates Khovanov homology through higher-dimensional perspectives and brane constructions, revealing a rich interplay among opers, Miura structures, Gaudin integrable systems, and S-duality, with implications for geometric Langlands and the AGT correspondence. Overall, the paper provides a comprehensive, multi-faceted route from 4D BPS equations to knot invariants and their categorifications, via a cohesive network of gauge theory, integrable systems, and conformal field theory.

Abstract

It has been argued based on electric-magnetic duality and other ingredients that the Jones polynomial of a knot in three dimensions can be computed by counting the solutions of certain gauge theory equations in four dimensions. Here, we attempt to verify this directly by analyzing the equations and counting their solutions, without reference to any quantum dualities. After suitably perturbing the equations to make their behavior more generic, we are able to get a fairly clear understanding of how the Jones polynomial emerges. The main ingredient in the argument is a link between the four-dimensional gauge theory equations in question and conformal blocks for degenerate representations of the Virasoro algebra in two dimensions. Along the way we get a better understanding of how our subject is related to a variety of new and old topics in mathematical physics, ranging from the Bethe ansatz for the Gaudin spin chain to the $M$-theory description of BPS monopoles and the relation between Chern-Simons gauge theory and Virasoro conformal blocks.

Figures (29)

• Figure 1: A knot has been placed at the boundary of the four-manifold $M_4=W\times{\mathbb R}_+$.
• Figure 2: Stretching a knot in one direction -- here taken to be the $x^1$ direction -- to reduce to a situation that almost everywhere is nearly independent of one coordinate. After much stretching, the knot is everywhere nearly independent of $x^1$, except near the finite set of critical values of $x^1$ at which a pair of strands appears or disappears. (In the figure, these occur only at the top and bottom.)
• Figure 3: A braid in $I\times C$; by gluing together the top and bottom, one can make a closed braid in $S^1\times C$. After much stretching, a braid can be described by adiabatic evolution in $x^1$, with no exceptional values where this description breaks down.
• Figure 4: As a knot is stretched along the boundary, a solution of the supersymmetric equations might become delocalized in the $y$ direction, normal to the boundary. This is schematically indicated here; the shaded region indicates the spatial extent of a solution -- that is, of the region over which the chosen solution deviates significantly from the one that describes the vacuum in the absence of knots -- and its thickness is proportional to the amount that the knot has been stretched.
• Figure 5: In a time-independent situation, we look for solutions on a three-manifold $M_3=C\times {\mathbb R}_+$, where $C$ (taken here to be a two-sphere) is a Riemann surface. Knots are placed at points on the boundary of $M_3$, labeled here as $z_1,\dots,z_4$.