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
There is a canonical way to turn any category in homotopy type theory into a weakly equivalent univalent category. This can be thought of as an analogue of univalence but for isomorphisms instead of equivalence.
Intuitively, the Rezk completion is a “strictification via Yoneda” type result, in the style of strictification for bicategories via Yoneda. One starts with a category that may not have nice strictness properties, embeds it into the category of presheaves, which does have the nice strictness properties, and then restricts to representables, which gives something equivalent to the original category, but retains the nice strictness properties.
Another intuition is that the Rezk completion is a vertical categorification of the construction of a quotient of a type by an equivalence relation by taking equivalence classes. That is, we can think of a type equipped with an equivalence relation as a boolean-enriched groupoid, and a boolean-valued presheaf is equivalently a predicate on that respects the equivalence relation. Then the representable presheaves are those predicates that are furthermore inhabited, i.e., precisely equivalence classes.
We work in a dependent type theory where UIP or axiom K cannot be proven. These results are from UFP13. Note: UFP13 calls a category a “precategory” and a univalent category a “category”, but here we shall refer to the standard terminology of “category” and “univalent category” respectively.
(Theorem 9.9.5 in UFP13) For any category , there is a univalent category and a weak equivalence .
Let be the type of representable objects of , with hom-sets inherited from . Then the inclusion is fully faithful and an embedding on objects. Since is a category by Theorem 9.2.5 (see functor category), is also a category. Let be the Yoneda embedding. This is fully faithful by corollary 9.5.6 (See Yoneda embedding), and essentially surjective by definition of . Thus it is a weak equivalence.
This has the unfortunate side effect of raising the universe level. If is a univalent category in a universe , then in this proof must at least be as large as . Hence the univalent category is in a higher universe than hence must also be in a higher universe and finally is also in a higher universe than .
Now this can all be avoided by constructing a higher inductive type with constructors:
If we ignore the last constructor we could also write the above as .
We now go on to build a univalent category with a weak equivalence by taking the type of objects as and defining hom-sets by double induction. The advantage of this approach is that it preserves universe levels, there are a lot of things to check but it is an easy proof. The kind of proof that is well suited to a proof assistant. For the complete proof see Theorem 9.9.5 of the HoTT book.
The relation between Segal completeness (now often “Rezk completion”) for internal categories in HoTT and the univalence axiom had been pointed out in:
This was developed in
See also:
Univalent Foundations Project, Homotopy Type Theory – Univalent Foundations of Mathematics (2013)
Egbert Rijke, The join construction, 26 Jan 2017, (arXiv:1701.07538)
Last revised on September 13, 2024 at 22:48:14. See the history of this page for a list of all contributions to it.