Skew-enriched categories (Q722252)

In the \textit{S. Eilenberg} and \textit{G. M. Kelly} paper [Proc. Conf. Categor. Algebra, La Jolla 1965, 421--562 (1966; Zbl 0192.10604)], categories enriched in a monoidal category \(\mathscr{V}\) did not start out having an ordinary category as part of the structure. An underlying category was eventually constructed with hom sets obtained from the hom objects in \(\mathscr{V}\) by applying the underlying functor \(\mathscr{V}\to \mathrm{Set}\) represented by the tensor unit \(I\in \mathscr{V}\). The definition then included that of ordinary category when \(\mathscr{V}=\mathrm{Set}\). Of course, we already needed to know what a category was to discuss the monoidal base category \(\mathscr{V}\). What the author here has discovered is that this order of development is undesirable when enriching over skew monoidal categories. It comes as a somewhat surprising idea, one not foreseen by the reviewer [\textit{R. Street}, J. Pure Appl. Algebra 217, No. 6, 973--988 (2013; Zbl 1365.18008)] whose definition of enriched category was satisfactory mainly in the case of left normal \(\mathscr{V}\). In [J. Homotopy Relat. Struct. 12, No. 1, 31--81 (2017; Zbl 1417.18001)], \textit{J. Bourke} opened up applications of skew-monoidal category theory to Quillen-style homotopy theory. The present paper advances that direction through the new notion of locally weak comonad in preparation for use in homotopy theory via algebraic weak factorisations. As shown in the reviewer's paper mentioned above, the monoidal versus closed dichotomy sits more naturally in the skew setting than the non-skew. As confirmation of correctness of Campbell's skew-enriched categories concept, he similarly resolves the actegory versus enriched category dichotomy through the introduction of his third new notion: skew enriched proactegory.