Non-semisimple Crane-Yetter theory varying over the character stack

TL;DR

Problem: extend four-dimensional Crane–Yetter theory to non-semisimple data and one that varies over the character stack, capturing a framing-invariant structure as a relative theory. Approach: build a relative 4D TQFT from non-semisimple data, using Morita theory, monadic reconstructions, and a sequence of relative dualizability, nondegeneracy, factorizability, and cofactorizability results. Contributions: establish a relative invertibility property, yield a canonical line bundle on 3-manifolds and a stacky unicity for skein algebras, and realize the character stack as a gauged symmetry defect. Significance: links higher category theory, non-semisimple quantum invariants, and gauge-theory defects, with potential impact on skein theory and topological phases.

Abstract

We construct a relative version of the Crane-Yetter topological quantum field theory in four dimensions, from non-semisimple data. Our theory is defined relative to the classical $G$-gauge theory in five dimensions -- this latter theory assigns to each manifold $M$ the appropriate linearization of the moduli stack of $G$-local systems, called the character stack. Our main result is to establish a relative invertibility property for our construction. This invertibility generalizes the key invertibility property of the original Crane-Yetter theory which allowed it to capture the framing anomaly of the celebrated Witten-Reshetikhin-Turaev theory. In particular our invertibilty statement at the level of surfaces implies a categorical, stacky version of the unicity theorem for skein algebras; at the level of 3-manifolds it equips the character stack with a canonical line bundle. Regarded as a topological symmetry defect of classical gauge theory, our work establishes invertibility of this defect by a gauging procedure.