model structure on relative categories



Model category theory

model category



Universal constructions


Producing new model structures

Presentation of (,1)(\infty,1)-categories

Model structures

for \infty-groupoids

for ∞-groupoids

for rational \infty-groupoids

for nn-groupoids

for \infty-groups

for \infty-algebras



for stable/spectrum objects

for (,1)(\infty,1)-categories

for stable (,1)(\infty,1)-categories

for (,1)(\infty,1)-operads

for (n,r)(n,r)-categories

for (,1)(\infty,1)-sheaves / \infty-stacks



The model category structure on the category of categories with weak equivalences is a model for the (∞,1)-category of (∞,1)-categories.

Every category with weak equivalences CC presents under Dwyer-Kan simplicial localization a simplicially enriched category or alternatively under Charles Rezk‘s simplicial nerve a Segal space, both of which are incarnations of a corresponding (∞,1)-category C\mathbf{C} with the same objects of CC, at least the 1-morphisms of CC and such that every weak equivalence in CC becomes a true equivalence (homotopy equivalence) in C\mathbf{C}.

By the work of Barwick and Kan, there is a model structure on relative categories presenting the (∞,1)-category of simplicial spaces.


For the purposes of the present entry, a category with weak equivalences means the bare minimum of what may reasonably go by that name:

Definition A relative category (C,W)(C,W) is a category CC equipped with a choice of wide subcategory WW.

A morphism in WW is called a weak equivalence in CC. Notice that we do not require here that these weak equivalence satisfy 2-out-of-3, nor even that they contain all isomorphisms.

A morphism (C 1,W 1)(C 2,W 2)(C_1,W_1) \to (C_2,W_2) of relative catgeories is a functor C 1C 2C_1 \to C_2 that preserves weak equivalences.

Write RelCatRelCat for the category of relative categories and such morphisms between them.

Model category structure


The subdivided nerve of Barwick-Kan induces a right transferred model structure on RelCatRelCat such that there is a Quillen equivalence

K ξ:sSet Reedy Δ op(RelCat,BK):N ξ K_\xi : sSet^{\Delta^{\op}}_{Reedy} \leftrightarrows (RelCat, BK) : N_\xi

Both functors preserve and reflect weak equivalences, and the adjunction unit and counit are natural weak equivalences. The simplicial nerve NN is naturally weakly equivalent to N ξN_\xi.

Furthermore, this all remains true if the Reedy model structure on sSet Δ opsSet^{\Delta^{\op}} is replaced with any of its Bousfield localizations.

The model structure on RelCat is left proper. It is also right proper when the model structure on sSet Δ op sSet^{\Delta^{\op}} is right proper.


The existence and properness of the model structure and the Quillen equivalence is Theorem 6.1 of Barwick-Kan, as is the fact N ξN_\xi preserves and reflects weak equivalences.

Proposition 10.3 shows the adjunction unit is a natural weak equivalence. That the counit is a natural weak equivalence follows by considering the triangle identity, 3-for-2, and the fact N ξN_\xi reflects weak equivalences

N ξCηN ξCN ξK ξN ξCN ξεCN ξC N_\xi C \xrightarrow{\eta N_\xi C} N_\xi K_\xi N_\xi C \xrightarrow{N_\xi \varepsilon C} N_\xi C

Finally, by considering the diagram

X η N ξK ξX f Y η N ξK ξY \array{ X & \stackrel{\eta}{\to} & N_\xi K_\xi X \\ \downarrow f && \downarrow \\ Y & \stackrel{\eta}{\to} & N_\xi K_\xi Y }

we see ff is a weak equivalence iff N ξK ξfN_\xi K_\xi f is, and iff K ξfK_\xi f is.

In particular, Rezk’s complete Segal space structure on bisimplicial sets is a Bousfield localization of the Reedy model structure, so we can define:


Let (RelCat,Rezk)(RelCat, Rezk) denote the model structure corresponding to the complete Segal spaces.


The localization L(RelCat,Rezk)(,1)CatL(RelCat, Rezk) \simeq (\infty,1)Cat is the ∞-localization functor (C,W)L(C,W)(C,W) \mapsto L(C, W) inverting the weak equivalences


We will see below that we can compute this through the map (C,W)(NC,NW)(C,W) \to (NC, NW) to marked simplicial sets. Suppose we’re given a fibrant replcaement (NC,NW)Y (NC, NW) \to Y^\natural. Since NCNC is a quasi-category, composition induces an equivalence for every quasi-category ZZ

Fun(Y,Z)Map (Y ,Z )Map ((NC,NW),Z )Fun NW(NC,Z) Fun(Y, Z) \simeq Map^\flat(Y^\natural, Z^\natural) \to Map^\flat((NC, NW), Z^\natural) \simeq Fun_{NW}(NC, Z)

thus YY satisfies the universal property of L(NC,NW)L(NC, NW).

It is shown in Meier that categories of fibrant objects are fibrant in this model structure.

Compatibility with other models for (∞,1)-categories

One often uses the hammock localization L H:RelCatsSetCatL^H : RelCat \to sSetCat, where sSetCatsSetCat is given the Bergner model structure whose weak equivalences are the Dwyer-Kan equivalences: i.e. the local weak homotopy equivalences.


A functor ff in RelCatRelCat is a Rezk equivalence iff L H(f)L^H(f) is a Dwyer-Kan equivalence.

This is main theorem 1.4 of Barwick-Kan.

The idea underlying marked simplicial is directly analogous to the idea underlying relative categories. In fact, the functor RelCatsSet +:(C,W)(NC,NW)RelCat \to sSet^+ : (C,W) \to (NC, NW) preserves and reflects equivalences, since


Let RR be a fibrant replacement in sSetCatsSetCat. Then the natural transformation (NC,NW)NRL H(C,W) (NC, NW) \mapsto NRL^H(C,W)^\natural of marked simplicial sets is a natural weak equivalence.

This is theorem 1.2.1 of Hinich

Nerve functors

The compatibility of the various nerve and simplicial localization functors is in section 1.11 of


Last revised on February 28, 2021 at 09:26:00. See the history of this page for a list of all contributions to it.