Want to take part in these discussions? Sign in if you have an account, or apply for one below
Vanilla 1.1.10 is a product of Lussumo. More Information: Documentation, Community Support.
There is a natural bijection between functors I→C and morphisms in C, namely the map that sends a functor I→C to the image in C of the morphism 0→1 in I.
Funct(I,C) is not itself a map from I to C; it is a category (the arrow category itself) whose objects are maps (functors) from I to C. So an object of Funct(I,C) is a functor F from I to C.
And what does this functor F:I→C consist of? I has two objects, 0 and 1; so we need two objects of C, F(0) and F(1). Similarly, I has three morphisms, id0:0→0, i:0→1, and id1:1→1; so we need three morphisms of C, F(id0):F(0)→F(0), F(i):F(0)→F(1), and F(id1):F(1)→F(1). Since the functor F must preserve identities, we need F(id0)=idF(0) and F(id1)=idF(1), so the data remaining is precisely that in the morphism f(i):F(0)→F(1). It now remains to check that the functor F preserves composition, but it does.
So a functor from I to C consists of precisely the same information as a morphism in C.
We still have to check that the morphisms in Funct(I,C) (which are natural transformations) correspond in this way to commutative squares in C, as claimed; I leave that to you.
One problem with this article is that it still shows our early tendency to give everything the slickest definition possible. I've rewritten it in a way that I think is more comprehensible.
yes, I see. While the previous definition is good for the brain if one has faith enough to pursue it, it does leave one with the feeling that in CT the easy is explained by the more complex. With your new definition the intuitive is explained first, then one gets an identity statement with a purer, though more complex way of stating things.
1 to 6 of 6