Non-semisimple WRT at the boundary of Crane-Yetter
Benjamin Haïoun
TL;DR
The paper formalizes the long-standing intuition that the Witten–Reshetikhin–Turaev 3-TQFT is a boundary condition for a Crane–Yetter 4-TQFT, extending the framework to non-semisimple settings. It proves Crane–Yetter can be realized as a once-extended TQFT and constructs a boundary condition using skein theory, enabling reconstruction of WRT and its non-semisimple analogues through anomaly-trivializing data. The authors develop a non-semisimple skein-theoretic 4-TQFT $ extsc{S}_{ ext{I}}$ supported by a chromatic, modified-trace input category, and show how the boundary condition yields a projective WRT/DGGPR theory via a resolved anomaly. This work bridges 3D and 4D TQFTs in both semisimple and non-semisimple contexts, offering a robust framework for anomalous and projective TQFTs with potential fully extended generalizations.
Abstract
We prove the slogan, promoted by Walker and Freed-Teleman twenty years ago, that "The Witten-Reshetikhin-Turaev 3-TQFT is a boundary condition for the Crane-Yetter 4-TQFT" and generalize it to the non-semisimple case following ideas of Jordan, Reutter and Walker. To achieve this, we prove that the Crane-Yetter 4-TQFT and its non-semisimple version arXiv:2306.03225 are once-extended TQFTs, using the main result of arXiv:2412.14649. We define a boundary condition, partially defined in the non-semisimple case, for this 4D theory. When the ribbon category used is modular, possibly non-semisimple, we check that the composition of this boundary condition with the values of the 4-TQFT on bounding manifolds reconstructs the Witten-Reshetikhin-Turaev 3-TQFTs and their non-semisimple versions arXiv:1912.02063, in a sense that we make precise.
