Peng Zhou

stream of notes

User Tools

Site Tools


blog:2025-01-09

This is an old revision of the document!


2025-01-09

  • colimit of algebras

colimit

Let's be really naive and simple. What is a colimit? Suppose you have a category $\mathcal{C}$, many objects, morphisms. And then, there is a push-out diagram $A \gets B \to C$. This gang of objects talks with everyone, for example, someone called $X$. $A$ talks with $X$, get a set $Hom(A,X)$, so does $B$, get $Hom(B,X)$. It might be big or small, we don't know, depends on $X$'s relation with all of them. Oh, and don't forget $Hom(C,X)$. So, how do we deal with the three sets? We cannot just union them together, that would be stupid. Well, by pre-composition, we have maps $Hom(A,X) \to Hom(C,X) \gets Hom(B,X)$. So, we can ask for compatible arrows in $Hom(A,X)$ and $Hom(B,X)$, see if their image in $Hom(C,X)$ coincide or not. If they coincide, then we remember that. Suppose $M$ is the would-be colimit, then we have $Hom(M,X)$ be the fibre-product. So, we find $M$ by its relations with everybody, that is $M \in Fun(\mathcal{C}, Set)$. It probably is representable. I am not sure what condition do I need to get an actual $M$, instead of indirect evidence.

blog/2025-01-09.1736497888.txt.gz · Last modified: 2025/01/10 08:31 by pzhou