2-category theory MOC

2-category

A 2-category1 is motivated from several perspectives:

  1. A 2-category is a category in the next dimension: While categories have a set2 of morphisms between objects, a 2-category has a category of morphisms between objects. Because the objects of each “hom-category” need not form a set, the “identities” involved are higher isomorphisms having coherence relations of their own, similar to the associator and unitors of a monoidal category.
  2. A 2-category is the oidification of a monoidal category, so that each object has a monoidal category of endomorphisms.

Thus, in terms of collections, a 2-category ℭ is a mathematical object consisting of: #m/def/cat/2

  1. a collection of objects, ℭ0;
  2. for any 𝐴,𝐵 ∈ℭ0, a category ℭ(𝐴,𝐵);
Notation

We call the elements of the collection ℭ1(𝐴,𝐵) :=ℭ(𝐴,𝐵)0 1-morphisms, and notate 𝑓 ∈ℭ1(𝐴,𝐵) with

𝑓:𝐴→𝐵.

For any 𝑓,𝑔 ∈ℭ1(𝐴,𝐵) we call the elements of the set ℭ2(𝐴,𝐵)(𝑓,𝑔) :=ℭ(𝐴,𝐵)(𝑓,𝑔) 2-morphisms, and notate 𝛽 ∈ℭ2(𝐴,𝐵)(𝑓,𝑔) with

𝛽:𝑓⇒𝑔:𝐴→𝐵.
  1. for any 𝐴,𝐵,𝐶 ∈ℭ0, a 1-composition bifunctor; (∘):ℭ(𝐵,𝐶)×ℭ(𝐴,𝐵)→ℭ(𝐴,𝐶);
  2. for any 𝐴 ∈ℭ0, a 1-identity 1𝐴 ∈ℭ1(𝐴,𝐴), sometimes by abuse denoted 𝐴;
  3. for any 𝐴,𝐵,𝐶,𝐷 ∈ℭ0, a natural isomorphism called the associator with components 𝛼ℎ,𝑔,𝑓 :(ℎ ∘𝑔) ∘𝑓 ⇒ℎ ∘(𝑔 ∘𝑓) :𝐴 →𝐷 for indices ℎ ∈ℭ1(𝐶,𝐷), 𝑔 ∈ℭ1(𝐵,𝐶), and 𝑓 ∈ℭ1(𝐴,𝐵);
  4. for any 𝐴,𝐵 ∈ℭ0, a natural isomorphism called the left-unitor with components 𝜆𝑓 :𝟙𝐵 ∘𝑓 ⇒𝑓 :𝐴 →𝐵 for index 𝑓 ∈ℭ(𝐴,𝐵); and
  5. for any 𝐴,𝐵 ∈ℭ0, a natural isomorphism called the right-unitor with components 𝜌𝑓 :𝑓 ∘𝟙𝐴 ⇒𝑓 :𝐴 →𝐵 in index 𝑓 ∈ℭ(𝐴,𝐵);

satisfying the so-called triangle identity

c|https://q.uiver.app/#q=WzAsMyxbMCwwLCIoZyBcXGNpcmMgQikgXFxjaXJjIGYiXSxbMSwxLCJnIFxcY2lyYyBmIl0sWzIsMCwiZyBcXGNpcmMgKEIgXFxjaXJjIGYpIl0sWzAsMiwiXFxhbHBoYV97ZyxCLGZ9IiwwLHsibGV2ZWwiOjJ9XSxbMCwxLCJcXHJob19nIFxcY2lyYyBmIiwyLHsibGV2ZWwiOjJ9XSxbMiwxLCJnIFxcY2lyYyBcXGxhbWJkYV9mIiwwLHsibGV2ZWwiOjJ9XV0=

and pentagon identity

c|https://q.uiver.app/#q=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

for any 𝐴𝑓→𝐵𝑔→𝐶ℎ→𝐷𝑖→𝐸 in ℭ. Together these diagrams ensure that the operation of ( ∘) is unital associative up to canonical natural isomorphism, by the Coherence theorem for 2-categories.

Examples

See also


#state/tidy | #lang/en | #SemBr

Footnotes

  1. What we call a 2-category is traditionally referred to as a bicategory; 2-category traditionally refers to a strict 2-category. ↩

  2. Which is the same as a (0,0)-category. ↩