Geometric Langlands And The Equations Of Nahm And Bogomolny
Edward Witten
TL;DR
The paper presents a gauge-theoretic framework for understanding aspects of geometric Langlands by leveraging twisted $\mathcal N=4$ super Yang–Mills, the Bogomolny and Nahm equations, and dual boundary conditions. It builds a concrete bridge between the cohomology of Hecke-modification moduli $\overline{\mathcal N}(\rho)$ and dual group representations $R^\vee$ through the $\widehat{\mathcal A}$- and $\widehat{\mathcal B}$-models, showing how a principal $SL_2$ action emerges from the grading of cohomology and how universal characteristic classes correspond to Lie-algebra elements under duality. The work details how Wilson and ‘t Hooft operators are related via OPEs, how fusion is compatible with dual actions, and how a three-dimensional theory with a universal kernel unifies boundary-condition dualities across the Langlands correspondence. A key insight is that Nahm’s equations and regular Nahm poles encode the dual boundary data, yielding a unique dual description and a natural map between cohomological and representation-theoretic structures. Overall, the article provides a rigorous gauge-theory route to geometric Langlands, clarifying the role of dualities, boundary conditions, and three-dimensional reductions while connecting universal kernels to boundary theories like $T(G)$.
Abstract
Geometric Langlands duality relates a representation of a simple Lie group $G^\vee$ to the cohomology of a certain moduli space associated with the dual group $G$. In this correspondence, a principal $SL_2$ subgroup of $G^\vee$ makes an unexpected appearance. Why this happens can be explained using gauge theory, as we will see in this article, with the help of the equations of Nahm and Bogomolny. (Based on a lecture at Geometry and Physics: Atiyah 80, Edinburgh, April 2009.)
