Formal homotopy quantum field theories. I: Formal maps and crossed \(\mathcal{C}\)-algebras. (Q847565)

A \((d+1)\)-dimensional topological quantum field theory is a linear representation of \((d+1)\)-dimensional cobordism [\textit{M.~Atiyah}, Publ. Math., Inst. Hautes Étud. Sci. 68, 175--186 (1988; Zbl 0692.53053)]. For small values of \(d\), such a theory is completely encoded in very simple algebraic data. For instance, a 1-dimensional TQFT is the same thing as the datum of a finite dimensional vector space endowed with a symmetric non-degenerate inner product, and a 2-dimensional TQFT is the same thing as a Frobenius algebra. It is natural to extend the notion of TQFT, by endowing the manifolds and cobordisms with additional structures, e.g., with a principal \(G\)-bundle. Up to isomorphism, the datum of such a bundle is encoded into a characteristic map, i.e., into the homotopy class of a map from the cobordism manifold to the classifying space \(BG\) of \(G\). More generally, one can define a \((d+1)\)-dimensional homotopy quantum field theory as a linear representation of \((d+1)\)-dimensional cobordism endowed with characteristic maps to a fixed pointed `background' space \(B\) [\textit{V.~Turaev}, Homotopy field theory in dimension 2 and group algebras, \url{arXiv:math/9910010}]. It has been shown in [\textit{G.~Rodrigues}, J. Knot Theory Ramifications 12, No.~3, 287--319 (2003; Zbl 1061.57027)] that a \((d+1)\)-dimensional HQFT with target \(B\) only depends on the \((d+1)\)-type of \(B\), i.e. the algebraic structure involved only knows about the first \((d+1)\)-homotopy groups of \(B\). This in particular means that in order to investigate \((d+1)\)-dimensional HQFTs, one can restrict the targets to be \((d+1)\)-types, i.e., to have vanishing homotopy groups above the \((d+1)^{st}\). In the present paper, the authors investigate 2-dimensional HQFTs. Partial results in this direction were known for \(B\) a connected 1-type, i.e. \(B\cong K(G,1)\) for a discrete group \(G\), and for \(B\) a simply connected 2-type, i.e. \(B\cong K(A,2)\), for \(A\) a discrete abelian group. Namely, it was shown in [\textit{V.~Turaev}, loc. cit.] that 2-dimensional HQFTs with target \(K(G,1)\) are equivalent to crossed \(G\)-algebras, and in [\textit{M.~Brightwell} and \textit{P.~Turner}, J. Knot Theory Ramifications 9, No.~7, 855--864 (2000; Zbl 0999.57030) ] that 2-dimensional HQFTs with target \(K(A,2)\) are equivalent to \(A\)-Frobenius algebras. For a general 2-type \(B\), the algebraic counterpart is that of a crossed module. More precisely, if \(B\) is a 2-type, then there exists a crossed module \(\mathcal{C}\) such that the classifying space \(B\mathcal{C}\) of \(\mathcal{C}\) is weakly equivalent to \(B\) [\textit{R.~Brown} and \textit{P.~J.~Higgins}, Math. Proc. Camb. Philos. Soc. 110, No.~1, 95--120 (1991; Zbl 0732.55007)]. Given a crossed module \(\mathcal{C}\) providing an algebraic model for a 2-type \(B\), the authors introduce a combinatorial model for 2-dimensional cobordisms with target \(B\) and for the corresponding HQFT. These combinatorial versions of homotopy classes of maps to \(B\) are called \textit{formal \(\mathcal{C}\)-maps} in the paper, and the resulting analogues of HQFTs, \textit{formal 2-dimensional HQFTs}, hence the title of the paper. The authors provide a complete classification of formal 2-dimensional HQFTs over a crossed module \(\mathcal{C}\) in terms of crossed \(\mathcal{C}\)-algebras (a common generalization of both crossed \(G\)-algebras and \(A\)-Frobenius algebras). More precisely, the main result in the paper states that there is a canonical bijection between isomorphism classes of formal 2-dimensional HQFTs based on a crossed module \(\mathcal{C}\) and isomorphism classes of crossed \(\mathcal{C}\)-algebras.