Spherical 2-categories and 4-manifold invariants (Q1290945)

Following on from the many constructions of invariants of 3-manifolds based on quantum group (Hopf algebra) structures [\textit{N. Yu. Reshetikhin} and \textit{V. G. Turaev}, Commun. Math. Phys. 127, No. 1, 1-26 (1990; Zbl 0768.57003); \textit{V. G. Turaev} and \textit{H. Wenzl}, Int. J. Math. 4, No. 2, 323-358 (1993; Zbl 0784.57007); \textit{D. S. Freed}, Commun. Math. Phys. 159, No. 2, 343-398 (1994; Zbl 0790.58007), and others], categories of their representations [\textit{V. G. Turaev} and \textit{O. Y. Viro}, Topology 31, No. 4, 865-902 (1992; Zbl 0779.57009), and others] or suitable category structures [\textit{S. Gelfand} and \textit{D. Kazhdan}, Geom. Funct. Anal. 6, No. 2, 268-300 (1996; Zbl 0870.57017); \textit{J. W. Barrett} and \textit{B. W. Westbury}, Trans. Am. Math. Soc. 348, No. 10, 3997-4022 (1996; Zbl 0865.57013), and others], there has been much speculation as to what corresponding data would be needed in order to construct an analogous 4-manifold invariant. \textit{L. Crane, L. H. Kauffman} and \textit{D. Yetter} [World Scientific Ser. Knots Everything 3, 131-138 (1993; Zbl 0841.57029)] tried to generalize the Turaev-Viro approach by replacing quantum \(6j\)-symbols by quantum \(15j\)-symbols, only to discover a state-sum description of the signature [\textit{J. Roberts}, Topology 34, No. 4, 771-787 (1995; Zbl 0866.57014)]. \textit{L. Crane} and \textit{I. B. Frenkel} [J. Math. Phys. 35, No. 10, 5136-5154 (1994; Zbl 0892.57014)] proposed the formalism of categorification and used it to suggest that a ``Hopf category'' would be the appropriate structure. The only example of a Hopf category which has been worked out is \(C(D(G))\), a categorification of the quantum double of a finite group; calculations of the associated 4-manifold invariants are in progress. Although they suggested that Lusztig's canonical bases for representations of quantum groups could be used to generate interesting examples, no interesting concrete calculations have been completed. Rather than categorifying Hopf algebras, the paper under review attempts to categorify the abstract structure of their category of representations. Namely, it proposes a concrete suggestion as to what should play the role for 4-manifolds of the structure of a braided balanced monoidal tensor category, from which 3-manifold invariants can be constructed, namely a ``non-degenerate finitely semi-simple semi-strict spherical 2-category of non-zero dimension''. A definition of this structure is given, starting from a strict 2-category with semi-strict monoidal structure with duals, trace-functors and a linear structure (so that the Hom spaces are 2-vector spaces). The derived invariant of a triangulated 4-manifold \((M,T)\) is constructed as a state sum, \(I_C(M,T)=K^{-v}\sum_\ell Z_C(M,T,\ell)\prod_e{\dim_q(\ell(e))^{-1}} \prod_f\dim_q(\ell(f))\), in which a state \(\ell\) is a labelling of edges by simple objects, and of triangles by simple 1-morphisms while the 2-category structure \(C\) provides linear maps associated with 4-dimensional simplices whose contraction is \(Z_C\). Unfortunately, the only example worked out in the paper of the appropriate \(2\)-category structure, \(C\), is derived from a finite group, which generates a theory in four dimensions analogous to Dijkgraaf-Witten theory in 3-dimensions, where \(I_C\) counts homomorphisms from the fundamental group into \(G\). However the author is hopeful that other examples can be derived from the 2-category of representations of the Hopf category \(C(D(G))\) of Crane-Frenkel, as well as from other (yet to be defined) Hopf categories.