As I was trying to understand the category $Ind(C)$ of diagrams of the form $I \to C$, where $I$ is a small filtered $(0,1)$-category, I wondered whether it is possible to define morphisms directly, without the limit-colimit construction (see https://ncatlab.org/nlab/show/ind-object), and then try to show that its equivalent to the common definition as $\text{lim$_i$ colim$_j$}\,C(F(i),G(j))$. A natural choice would be to define a morphism $(F:I\to C) \to (G:J \to C)$ as a pair $(f,g)$, where $f:I \to J$ is a functor in the $(1,2)$-category of all poset-categories, and $g:F \Rightarrow G \circ f$ is a natural transformation. Then, the functor $C \hookrightarrow Ind(C)$ is clearly fully faithful. While trying to show that each object in $Ind(C)$ is a colimit of itself, I recognized that I should rather look at 2-colimits, but then I should define what a 2-morphism is. I tried several things but I'm not an expert for higher categories, so unfortunately I lack the necessary intuition... do you know a canonical choice for 2-morphisms in $Ind(C)$?

  • 2
    $\begingroup$ This paper might be relevant to your question: arxiv.org/abs/1406.6229 $\endgroup$ – Yonatan Harpaz Apr 27 '16 at 20:44
  • $\begingroup$ Thx! This is exactly the kind of result I desired! $\endgroup$ – Bipolar Minds Apr 27 '16 at 21:38

On a more general note, you're (gradually) building a subcategory of the 2-fiber $\textbf{Cat}/C$. An excellent reference for those is the following paper, which helped me a lot when I was looking into this stuff:

A Bullejos and M Cegarra, On the geometry of 2-categories and their classifying spaces, K-Theory 29 (2003) 211-229.

The journal where this paper appeared is now famously defunct, but you can find a pdf of the article here. The explicit answer to your question regarding 2-morphisms is as follows. Let $F$ and $G$ be as in your question and consider a pair of 1-morphisms $(f,g)$ and $(f',g')$, also as in your question. A 2-morphism $(f,g) \Rightarrow (f',g')$ is a natural transformation $\beta:f \Rightarrow f'$ (i.e., a 1-morphism in the category of functors from $I$ to $J$) so that the following holds: $$(G\circ\beta) * g = g'$$

Here $*$ stands for vertical composition while $\circ$ is horizontal composition. Pictorially, we have:

Fiber 2-morphism

| cite | improve this answer | |
  • $\begingroup$ Wow, thx a lot, I will check that out! $\endgroup$ – Bipolar Minds Apr 28 '16 at 17:41

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.