An infinity operad of normalized cacti (Q2145910)

A \textit{cactus} is a configuration of labelled circles of various lengths, attached to each other in a treelike manner, and with a specified global base-point. The space of cacti with a given number, \(n\), of lobes will be denoted \(\mathrm{Cact}(n)\). A cactus in \textit{normalised} if the length of each of the circles is 1, and \(Cact^1(n)\) denotes the subspace of \(\mathrm{Cact}(n)\) determined by the normalised cacti. Cacti were introduced by \textit{A. A. Voronov} [Proc. Symp. Pure Math. 73, 81--103 (2005; Zbl 1083.18005)]. \textit{R. M. Kaufmann} [Algebr. Geom. Topol. 5, 237--300 (2005; Zbl 1083.55007)] proved that \(\mathrm{Cact}^1(n)\) has the structure of a CW-complex with the \(k\)-cells indexed by black-white trees having \(k\) white edges. He showed that \(\mathrm{Cact}^1(n)\simeq \mathrm{Cact}(n)\simeq \mathrm{Disk}_2(n)\), the \(n\) component of the little 2-disc operad. This gives an operad structure on \(\mathrm{Cact}\) making it equivalent to the little 2-disc operad. This structure does not, however, restrict to an operad structure on \(\mathrm{Cact}^1\) as the induced structure fails to be associative, yet given the explicit CW-complex structure on it, a better understanding of this lack of associativity would make it a very useful tool to have in the study of the little 2-disc operad and thus to various related constructions in geometry and topology. This paper looks at the question as to whether or not \(\mathrm{Cact}^1\) is an operad up to coherent homotopy, \emph{i.e.}, is it an \(\infty\)-operad? The authors show that it is. The higher homotopies are encoded by a topological operad of bracketed trees which is related to both an enrichment of the dendroidal category and to the Boardman-Vogt \(W\)-construction on the operad of operads