Construction 15.1. Suppose a category equipped with a subcategory that contains all the objects of (i.e., a relative category in the terminology of [6]). We call the morphisms of weak equivalences. In this situation, one may form the hammock localization of DwyerβKan [16]; this is a simplicial category. One may apply to each mapping space a fibrant replacement that preserves products (e.g., ) to obtain a category enriched in Kan complexes, which we shall denote . We may now apply the simplicial nerve construction [28, 1.1.5.5] to obtain a -category , which we shall denote simply by . We shall call the -category underlying the relative category .
15. Epilogue: Model categories of -categories[0MM3]
We conclude with a brief discussion of model categories of -categories, in which we describe some interactions between our results here and those of Bergner, Lurie, Rezk, and Simpson. We first note that a spate of further corollaries to our main results can be obtained by employing the following.
Remark 15.2. When and are model categories, and is a Quillen equivalence between them, there is [17] an induced equivalence of hammock localizations , and thus of underlying -categories .
Example 15.3. The -category underlying the relative category of -relative categories [5] is a theory of -categories.
Definition 15.4. Let us call a model category a model category of -categories if its underlying -category is a theory of -categories.
Example 15.5. By [25], the Joyal model category of simplicial sets is a model category of -categories. More generally, all of the following model categories are model categories of -categories:
We may now use the construction above to find more examples of theories of -categories.
Example 15.7 ([29, Proposition 1.5.4]). Let be a left proper simplicial combinatorial model category which is an absolute distributor ([29, Definition 1.5.1]). Then the category , of simplicial objects in , admits the -enriched complete Segal model structure , which is again left proper, simplicial, combinatorial, and an absolute distributor. If is a model category of of -categories, then is a model category of -categories.
The condition of being an absolute distributor is needed in order to formulate the correct notion of complete -enriched Segal object. We refer the reader to [29] for details, but note that being an absolute distributor is a property of the underlying -category of the given model category. In particular it is preserved under any Quillen equivalence.
Example 15.8. Suppose that is a model category satisfying the following list of conditions.
- ββ(M.1)
The class of weak equivalences of are closed under filtered colimits.
- ββ(M.2)
Every monomorphism of is a cofibration.
- ββ(M.3)
For any object of , the functor preserves colimits.
- ββ(M.4)
For any cofibrations and , the pushout product
is a cofibration that is trivial if either or is.
- ββ(M.5)
The -category is a homotopy theory of -categories.
Work of Bergner [12] and Lurie [29], combined with 14.6 above, shows that each of the following is an example of a model category of -categories:
- β’
the projective or (equivalently) the injective model category [29, 2.2.16, 2.3.1, 2.3.9] of -enriched preSegal categories, and
- β’
the model category [28, A.3.2] of categories enriched in .
Moreover, following Simpson [37] the injective (aka Reedy) model category of Segal -categories [22, 32, 37] satisfies conditions (M.1-4); indeed, the most difficult of these to verify is (M.4), which Simpson does in [37, Th.Β 19.3.2 (using Corollary 17.2.6)].
Thus, for example, the injective and projective model categories of -enriched Segal categories as well as the model category of categories enriched in are seen to be model categories of -categories. Indeed very recent work of Bergner and Rezk [13] discusses these model categories in detail and links them by an explicit chain of Quillen equivalences.
Additionally, we see that the injective model category of Segal -categories is also a model category of -categories, as is the model category of categories enriched in Segal -categories.
A partial converse to 15.2 holds, which allows one to deduce Quillen equivalences between these various model categories.
Lemma 15.9 ([28, A.3.7.7]). Two combinatorial model categories and are connected by a chain of Quillen equivalences if and only if and are equivalent -categories.
From this it follows that if and are combinatorial model categories with the property that both and are theories of -categories, then and are connected by a chain of Quillen equivalences. This applies to all of the model categories of -categories mentioned above.
A zig-zag of Quillen equivalences can be a troublesome gadget to work with. It is usually far more informative to have a single direct and explicit Quillen equivalence between competing model categories of -categories. While our techniques do not generally provide such a direct Quillen equivalence, we do offer the following recognition principle.
Proposition 15.10. Let and be two model categories of -categories and let be a Quillen adjunction between them. Then is a Quillen equivalence if and only if the left derived functor preserves the cells up to weak equivalence.
Proof. A Quillen equivalence induces an equivalence of -categories. By Lemma 10.2 and Lemma 4.8 any such equivalence necessarily preserves the cells up to equivalence. Conversely, as the left-derived functor preserves (homotopy) colimits and and are generated under (homotopy) colimits by the cells (Axiom C.2), it follows that induces an equivalence of -categories. In particular it induces an equivalence of homotopy categories, and hence is a Quillen equivalence. β
In particular the above applies when the cells are fibrant-cofibrant objects of and which are preserved by itself.
Example 15.11. The standard Quillen adjunction (cf. [29, Lemma 2.3.13]) from Segal -categories to -fold complete Segal spaces is a Quillen equivalence.
Example 15.12. The functor induces a Quillen equivalence between the model category of complete Segal -spaces [34] and the model category of -fold complete Segal spaces [29, 1.5.4]. (See also BergnerβRezk [14]).
A category with a specified subcategory of weak equivalences is a relative category, and hence gives rise to a homotopy theory. Thus any theory of -categories arising this way may, in principle, be compared using our axioms. We therefore end with the following.
Conjecture 15.13. The -category underlying Verityβs -trivial weak complicial sets [39, 40] is a homotopy theory of -categories. The relative category consisting of Bataninβs -categories [7] such that every -cell is an equivalence for , together with the class of morphisms which are essentially -surjective for all is a theory of -categories.
Original source: arXiv:1112.0040v6
Original source Β· 1112.0040v6