Symmetric homotopy theory for operads (Q1983572)

This article is a foundational treatment of the homotopy theory of symmetric operads, which are algebraic objects that encode algebraic structures (such as associative algebras, commutative algebras, etc.). Let us now give some context to explain the crucial difference between this article and the classical treatment of homotopy theory for symmetric operads. In a symmetric operad, the symmetric groups \(\Sigma_n\) act on the elements of the operad. This action is reflected in the algebraic structures: for example, in a commutative associative algebra (i.e., an algebra over the operad \(\mathsf{Com}\)), the commutativity of the binary product is encoded as the fact that the generating binary operation \(\mu \in \mathsf{Com}(2)\) is invariant under the action of \(\Sigma_2\). In the classical versions of the constructions of homotopy theory of symmetric operads (the bar-cobar adjunction, the homotopy transfer theorem, or Koszul duality), these symmetric group actions are ``baked'' into the theory, in that they are never relaxed up to homotopy, unlike e.g. the operadic composition. To illustrate this, consider that Koszul duality of operads produces a resolution of the operad \(\mathsf{Com}\) which encodes \(\mathsf{C}_\infty\)-algebras, i.e., spaces equipped with a binary product which is associative up to homotopy but strictly commutative. This stands in contrast to the notion of \(\mathsf{E}_\infty\)-algebras, which are equipped with a binary product that is associative \textit{and} commutative up to homotopy. This notion of \(\mathsf{E}_\infty\)-algebras appears naturally in the study of iterated loop spaces [\textit{J. M. Boardman} and \textit{R. M. Vogt}, Bull. Am. Math. Soc. 74, 1117--1122 (1968; Zbl 0165.26204); \textit{J. P. May}, The geometry of iterated loop spaces. Berlin-Heidelberg-New York: Springer-Verlag (1972; Zbl 0244.55009)], and it is only equivalent to the notion of \(\mathsf{C}_\infty\)-algebras in characteristic zero. In the article under review, the authors develop a homotopy theory for symmetric operads which allows the relaxation of symmetric group actions up to homotopy. Their key technical theorem is that the operad \(\mathcal{O}\), whose algebras in \(\mathbb{N}\)-modules (rather than \(\mathbb{S}\)-modules as usual) are nonunital operads, is Koszul in the sense of \textit{J. Hirsh} and \textit{J. Millès} [Math. Ann. 354, No. 4, 1465--1520 (2012; Zbl 1276.18009)]. This allows the authors to prove that their version of the bar-cobar adjunction produces resolutions of operads as \(\mathbb{N}\)-modules, i.e., it also resolves the action of the symmetric groups actions. Finally, they upgrade the homotopy transfer theorem. This theorem usually allows one to transfer an operad structure on some operad \(\mathcal{P}\) onto any equivalent \(\mathbb{S}\)-module. In this article, the authors prove that one can, in fact, transfer such a structure in the category of \(\mathbb{N}\)-modules.