homotopy theory, (∞,1)-category theory, homotopy type theory
flavors: stable, equivariant, rational, p-adic, proper, geometric, cohesive, directed…
models: topological, simplicial, localic, …
see also algebraic topology
Introductions
Definitions
Paths and cylinders
Homotopy groups
Basic facts
Theorems
The notion of a relative category is an extremely weak version of a category with weak equivalences, providing the bare minimum needed to present an (∞,1)-category. The idea was first explored in a series of papers Dwyer & Kan 1980 on simplicial localization (see there for more). A model category structure on relative categories was constructed by Barwick & Kan 2012, presenting the (∞,1)-category of (∞,1)-categories.
More generally, the notion of relative category captures the idea of having a distinguished wide subcategory of morphisms that should be preserved by functors. Relative categories are occasionally used for this purpose, unrelated to any higher categorical purpose.
A relative category $C$ is a pair $(und C, weq C)$, where $und C$ is a category and $weq C$ is a wide subcategory. A morphism in $weq C$ is said to be a weak equivalence in $C$. A relative functor $f : C \to D$ is a functor $f \colon und C \to und D$ that preserves weak equivalences in the obvious sense.
The homotopy category of a relative category $C$ is the ordinary category $Ho C$ obtained from $und C$ by freely inverting the weak equivalences in $C$.
A semi-saturated relative category is a relative category $C$ such that every isomorphism in $und C$ is a weak equivalence in $C$.
A category with weak equivalences is a semi-saturated relative category $C$ such that $weq C$ has the two-out-of-three property.
A homotopical category is a relative category $C$ such that $weq C$ has the two-out-of-six property; note that this automatically makes $C$ a category with weak equivalences.
A saturated homotopical category is a relative category $C$ such that a morphism is a weak equivalence if and only if it is invertible in $Ho C$; note that any such relative category must be a homotopical category in particular.
The semi-saturation condition is precisely the condition on the inclusion $weq C \to und C$ ensuring it is a monomorphism in the (2,1)-category of 1-categories.
Any ordinary category $C$ gives rise to three relative categories in a functorial way:
A minimal relative category $min C$, in which the only weak equivalences are identities.
A minimal homotopical category $min^+ C$, in which the only weak equivalences are isomorphisms.
A maximal homotopical category $max C$, in which every morphism is a weak equivalence.
Let $\mathbf{RelCat}$ be the large category of small relative categories and relative functors. It is a locally finitely presentable cartesian closed category, and we refer to the exponential object $[C, D]_h$ as the relative functor category. $\mathbf{RelCat}$ is, in particular, a (strict) 2-category.
Let $\mathbf{SsRelCat}$ be the full subcategory of semi-saturated relative categories. The inclusion $\mathbf{SsRelCat} \hookrightarrow \mathbf{RelCat}$ has a left adjoint that preserves finite products, so $\mathbf{SsRelCat}$ is a reflective exponential ideal of $\mathbf{RelCat}$.
There are then the following strings of adjunctions:
Moreover, because $min^+$ embeds $\mathbf{Cat}$ as a reflective exponential ideal in $\mathbf{SsRelCat}$ and in $\mathbf{RelCat}$, the functor $Ho : \mathbf{SsRelCat} \to \mathbf{Cat}$ preserves finite products.
The application of relative categories as expressing the idea of having a subcategory of distinguished morphisms can be neatly packaged into viewing relative categories as enriched categories over the category of sets with distinguished subsets. An equivalent formulation is that of an M-category (see also at F-category for a 2-category-theoretic version of this idea).
Let $PairSet$ be the cartesian closed category whose objects are pairs of small sets $(X, A)$ such that $A \subseteq X$, and whose morphisms $(X, A) \to (Y, B)$ are functions $f : X \to Y$ with $f(A) \subseteq B$.
$RelCat$ is isomorphic to the category of small $PairSet$-enriched categories.
With minimal rephrasing, a small $PairSet$-enriched category consists of the data
that is subject to identity and associativity relations. It’s immediate that this is exactly the same data as a relative category $(C,W)$. And under this identification, enriched functors and relative functors are the same thing.
A relative category can be equipped with a weak model structure, left or right semimodel structure, or a model structure. (These are listed from less restrictive to more restrictive structures.)
These structures do not change the underlying (∞,1)-category. However, the do provide constructions to perform computations in the underlying (∞,1)-category. Different structures yield different constructions, but all resulting answers are weakly equivalent.
The theory of relative categories presents a theory of (∞,1)-categories in the following sense:
There exists an adjunction
such that every left Bousfield localization of the Reedy model structure on the category $\mathbf{ssSet}$ of bisimplicial sets induces a cofibrantly-generated left proper model structure on $\mathbf{RelCat}$ making the adjunction a Quillen equivalence. In particular, there exists a model structure on $\mathbf{RelCat}$ that is Quillen equivalent to the model structure for complete Segal spaces on $\mathbf{ssSet}$.
This is discussed in further detail at model structure on categories with weak equivalences.
The terminology “relative categories” is due to
while the basic idea goes back, at least, to:
in the context of simplicial localization (cf. the references at “category with weak equivalences”).
Last revised on May 1, 2023 at 16:55:19. See the history of this page for a list of all contributions to it.