Polynomial monads and delooping of mapping spaces (Q2180250)

Finitary polynomial monads, also known as \(\Sigma\)-free colored operads or multicategories, are generalizations of (small) categories in which morphisms can have several outputs rather than one. Small categories have a rich homotopy theory based on many different ingredients developed throughout the years. In this article, the authors generalize several of these ingredients to polynomial monads: \begin{itemize} \item The Grothendieck construction: it turns a polynomial monad \(T\) and a categorical \(T\)-algebra \(A\) (i.e. a \(T\)-algebra in \(\mathsf{Cat}\)) into a polynomial monad \(\int A\) over \(T\) such that sections of \(\int A \to T\) are internal \(T\)-algebras in \(A\). This construction is the right adjoint to the classifier construction \(T^{(-)}\), where \(T^S\) classifies internal \(S\)-algebras inside categorical \(T\)-algebras. \item Quillen's Theorem A: it implies that a morphism of polynomial monads \(f : T \to S\) induces an equivalence between nerves \(N(R^S) \to N(R^T)\) (for any polynomial monad \(R\)) if and only if \(N(T^S)\) is contractible -- which the authors define as \(f : S \to T\) being homotopy cofinal. \item A criterion for the homotopy cofinality in terms of preservation of homotopy limits. \item Thomason's theorem: given a presheaf \(F\) of polynomial monads over a polynomial monad \(D\), the nerve of the classifier \(D^{\oint F}\) can be expressed as the homotopy colimit of the nerves of \(D^{F(a)}\), where \(\oint F\) is the twisted Boardman--Vogt tensor product which represents families of \(F(a)\)-algebras. \end{itemize} As an application, the authors study the same \(\overline{\mathrm{Emb}}(\mathbb{R}, \mathbb{R}^m)\) of long knots modulo immersions. Using their categorical methods, they reprove a theorem of Dwyer-Hess and Turchin which states that this space is the double loop space of the derived mapping space of morphisms \(\mathcal{D}_1 \to \mathcal{D}_m\) between little disks operads. Starting from a result of Sinha that \(\overline{\mathrm{Emb}}(\mathbb{R}, \mathbb{R}^m)\) can be expressed as the totalization of the Kontsevich operad \(\mathcal{K}\) and thus as a derived mapping space of weak \(\mathsf{Ass}\)-bimodules, the proof works in two steps: \begin{itemize} \item The mapping space of weak bimodule morphisms \(\mathsf{Ass} \to \mathcal{O}\) is homotopy equivalent to the loop space of the mapping space of bimodule morphisms \(\mathsf{Ass} \to \mathcal{O}\). In the authors' formalism, this follows from the fact that the morphism \(\mathsf{WBimod}_+ \to \mathsf{Bimod}_{**}\) from the \(+\) construction of the polynomial monad of weak \(\mathsf{Ass}\)-bimodules to polynomial monad of double pointed \(\mathsf{Ass}\)-bimodules is homotopy cofinal. \item The mapping space of bimodule morphisms \(\mathsf{Ass} \to \mathcal{O}\) is homotopy equivalent to the loop space of the mapping space of operad morphisms \(\mathsf{Ass} \to \mathcal{O}\). In the authors' formalism, this follows from the fact that the morphism \(\mathsf{Bimod}_+ \to \mathsf{NOp}_{**}\) from the \(+\) construction of the polynomial monad of \(\mathsf{Ass}\)-bimodules to polynomial monad of double pointed operads is homotopy cofinal. \end{itemize}