Let F : 𝒞 → ℰ and G : 𝒟 → ℰ be functors. The comma category F ↓ G is the category with
Objects: (A, B, f) where
f : F(A) → G(B)
Morphisms: (ϕ, ψ) : (A1, B1, f1) → (A2, B2, f2)
making the relevant square commute: f1G(ψ) = F(ϕ)f2
If one of the functors is the identity functor or an inclusion functor, we just write the relevant categories, instead of the functor itself. For example, I𝒞 : 𝒞 → 𝒞 and i : 𝒟 → 𝒞 has the comma category (𝒞 ↓ 𝒟)
Consider the comma category 𝒞 ↓ 𝒞. Then objects are just morphisms in 𝒞 and morphisms are the commuting squares. This is the arrow category.
Let X be about object of 𝒞. Then the comma category 𝒞 ↓ X has objects maps into X and the morphisms are commuting triangles. This is the slice category.
Dually X ↓ 𝒞 is the coslice category
Let F, G : 𝒞 → 𝒟. Then F ↓ G has objects f : F(A) → G(A), which look like components of natural transformations; and the morphisms are commuting squares.
An actual natural transformation is then