Fibrations and Yoneda's lemma in an \(\infty\)-cosmos (Q338225): Difference between revisions

The intention for \((\infty,1)\)-categories is that they should be weak infinite-dimensional categories with all \(n\)-morphisms invertible for \(n>1\). There are a number of concrete models of \((\infty,1)\)-categories, one such being weak Kan complexes (also called quasi-categories). A lot of category theory is by now adapted to \((\infty,1)\)-categories by André Joyal and by Jacob Lurie, working with specific models. Töen and, separately Barwick and Schommer-Pries, have shown that all models of \((\infty,1)\)-categories have the same homotopy theory in a specific sense. The present paper makes significant progress in showing that the category theory of \((\infty,1)\)-categories is model invariant. An environment for developing modifications of ordinary category theory is provided by a 2-category with the structure of a ``cosmos'' in the reviewer's terminology. Indeed, Joyal has shown that some aspects of \((\infty,1)\)-categories can be understood at a first level using an ordinary 2-category of \((\infty,1)\)-categories, called the homotopy 2-category. To obtain the higher level information, a richer structure than ordinary 2-category is needed, from which the homotopy 2-category can be derived. The authors' answer is an \(\infty\)-cosmos: this is a category enriched in simplicial sets with two classes of distinguished morphisms called weak equivalences and isofibrations. The axioms are much like those for a category of fibrant objects. Each of complete Segal spaces, Segal categories, and marked simplicial sets has its own \(\infty\)-cosmos. The basic theory of adjunctions, limits and colimits, and monadicity, are shown to apply in the homotopy 2-category of any \(\infty\)-cosmos. The axioms imply that the 2-categorical notion of equivalence, interpreted in the homotopy 2-category, precisely coincides with the model categorical notion of weak equivalence: so the theory is weak equivalence invariant. An \(\infty\)-cosmos is not meant to axiomatize the collection of \((\infty,1)\)-categories. Slices of an \(\infty\)-cosmos again form an \(\infty\)-cosmos. Some models of \((\infty,n)\)-categories also form an \(\infty\)-cosmos; so the present paper also begins the basic category theory of \((\infty,n)\)-categories. Indeed, moving forgetfully from \((\infty,n)\)-categories to \((\infty,1)\)-categories gives a nice example of the authors' notion of ``functor'' between \(\infty\)-cosmoi. These functors always induce a bicategorical equivalence on the homotopy 2-categories thereby confirming the model invariance of the \(\infty\)-cosmos framework.