Graphical combinatorics and a distributive law for modular operads (Q2237392)

The combinatorics of category-like structures is governed by different gadgets; ordinary categories by finite ordinals (thought of as linear trees), ordinary operads by finite rooted trees, and modular operads by graphs that admit cycles. In each case, collating the combinatorial objects to form a category, and taking prehseaves on it, leads to a topos in which an \(\infty\)-version of the original structures can be interpreted. E.g., Joyal's \(\infty \)-categories are simplicial sets satisfying a filling condition. In this context, a nerve theorem refers to identifying the original structures inside the presheaf category by combinatorial means; e.g., the nerve of an ordinary category is essentially the \(\infty\)-categories in which the filling condition is satisfied uniquely. In the case of ordinary categories and operads a verification by hand of the nerve theorem is relatively straightfoward, owing to the combinatorics of linear trees and finite rooted trees being not too complex. Moreover, Weber's general machinery furnishes a nerve theorem for the algebras of a suitable monad with arities. Categories and opeards are algebras for such a monad, and so their nerve theorem follows from general principles. The situation for modular operads is starkly different. The combinatorial complexity resulting from the cycles in the graphs is well known to induce many difficulties. The two results of the article are the existence of a monad on a suitable category whose algebras are modular opeards, and obtaining a nerve theorem. Variants of such results for modular operads have been obtained in [\textit{P. Hackney} et al., Adv. Math. 365, Article ID 107044, 61 p. (2020; Zbl 1437.18015)] and [\textit{P. Hackney} et al., Adv. Math. 370, Article ID 107206, 38 p. (2020; Zbl 1482.18014)]. However, the techniques developed here are of no lesser importance than these results. The author identifies the main combinatorial obstacle as arising from contracting units and obtaining exceptionally problematic loops. The next step is thus to split up the monad construction into two monads: one whose algebras are non-unital modular operads, and another whose algebras add unites. The two are then combined via the general theory of distributive laws for monads. A result of the monadic approach taken is an elucidation of the combainatorics underlying modular operads, as well as rendering the latter amenable to general machinery, such as Weber's nerve theorem. There results a much more well rounded understanding of the theory whereby desired results, such as the nerve theorem, follow expectedly from general principles in a non ad hoc fashion. The article is very well written and its length is accounted for by the inherent combinatorial complexities involved, as well as to the much care taken in order to make the article readable and accessible. In particular, it contains much of the background information on monads, distributivity laws, Weber's nerve theorem, and the category of graphical interfaces forming the carrier category for the monad.