Orbifold construction for topological field theories (Q1621595)

Turaev defined a homotopy quantum field theory (HQFT) with target space, $T$, to be a symmetric monoidal functor, $Z$, similar to those used for TQFTs, except that whilst the domain category for a TQFT is a category, $\mathrm{Cob}(n)$, having as objects $(n-1)$-dimensional manifolds, $X$, $Y$, etc., and as morphisms $n$-dimensional (co)bordisms between them, the objects of the domain category, $T$-$\mathrm{Cob}(n)$, for a HQFT are pairs $(X,g)$, where $X$ is still an $(n-1)$-dimensional manifold, but one also has a continuous map, $g:X\to T$, where $T$ is the fixed `target space' of the theory. Cobordisms between pairs also come equipped with functions to $T$, compatibly with those on the domain/input and codomain/output objects. In addition a homotopy invariance condition must be satisfied. The category, $\mathrm{HSym}(T$-$\mathrm{Cob}(n), \mathrm{Vect}_K)$, of homotopy quantum field theories with target $T$, then forms a groupoid. Following several previous authors, the authors here define a $G$-equivariant TQFT to be a HQFT with target a classifying space, $BG$, where $G$ is a finite group. An object of $G$-$\mathrm{Cob}(n)$ (short for $BG$-$\mathrm{Cob}(n)$) is then, equivalently, a manifold, $X$, with a specified principal $G$-bundle defined on it, induced from the universal principal $G$ bundle on $BG$, by pullback along $g$. The aim of the paper under review is then to explore, in depth, the effects of changing the finite group along a morphism $\lambda:G\to H$. This yields a pushforward operation, $\lambda_*$ from $\mathrm{HSym}(G$-$\mathrm{Cob}(n), \mathrm{Vect}_K)$ to $\mathrm{HSym}(H$-$\mathrm{Cob}(n), \mathrm{Vect}_K)$ and explicit formulae are derived for $\lambda_*$. The important case when $H$ is the trivial group, $1$, is then the orbifold construction mentioned in the title of the paper. This construction unifies and generalises several known algebraic notions of orbifoldisation.