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.
1 to 8 of 8
Let F,G∈[C,D] be two functors commuting with small colimits, and let α:F→G be a natural transformation between them.
Then according to Cisinski’s book Astérisque 308 Corollaire 1.1.8, the natural transformation α induces a functor H:D→Arr(D) that commutes with small colimits as well. That’s it. That’s the only information he gives about H. I assume that it’s trivial because he doesn’t go into more detail, but I have no idea what it should be.
If I had to guess, and this is very speculative, I would say that for X in D, we can look at the diagram induced by α, that is, the diagram defined by {αZ:X→G(Z)|Z∈F−1(X)} and take
H(X):=X→colimZ∈F−1(X)G(Z),the induced map to the colimit.
The natural transformation induces a functor C′→Arr(C) (switching to his notation, C replaced by C’ and D replaced by C) - this is standard and trivial. I wonder if it is a typo?… It is a typo, because by the following paragraph, he refers to F′=H−1(˜F), using the notation of the corollary this is a class of arrows in C′, hence H is a functor from C′ to Arr(C). The only potentially tricky bit is showing that H commutes with inductive limits. I leave it as an exercise.
Ah, so it was a typo!
Hmm… Then in Proposition 1.1.7, how is f↦˜f an endofunctor of Arr(C)? It’s sending arrows in Arr(C) to objects in Arr(C). Rather, it seems like a functor Arr(Arr(C))→Arr(C), which is not an endofunctor. Is this also a mistake, or am I missing the point…
(Comment deleted.)
Alright, I’ve concluded that calling it an endofunctor is incorrect. If it is a functor Arr(Arr(C))->Arr(C), this makes sense with the proof, since it turns a retract of squares into a retract of arrows, and it preserves identities blah blah.
Hmm… indeed!
Alright, I e-mailed Cisinski about the errors. I’m probably the first person who didn’t skip through this section.
1 to 8 of 8