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.
added remark (here) that the arrow category is the Grothendieck construction on the slice categories (copied over just with mild adjustments of notation from the same example that I just added to Grothendieck construction)
Is there a specific use or characterization of the arrow category of a category of representations Rep(A) of some algebra A? I think I’ve seen at some point someone discuss arrow categories of representations of quivers but perhaps the case of algebra reps is more familiar?
There is a somewhat tautological “meaning” of the arrow category of a category of representations: It’s the category of all “intertwiners”/”natural transformations” without restriction on the (co)domains being intertwined.
Maybe it helps to unwind a little:
Let 𝒞 be a small category, which may be a delooping groupoid if we are talking about group representations, or the free category of a graph if we are talking about quiver representations, etc.
Then for some coefficient category 𝒟 (like vector spaces), the representations of 𝒞 in 𝒟 are the functors 𝒞→𝒟, and the representation category is the functor category Rep(𝒞;𝒟)≡Func(𝒞,𝒟).
Now an arrow category, of course, is a functor category out of an interval category I, and since Cat is cartesian closed we have
Arr(Rep(𝒞;𝒟))≡Arr(Func(𝒞,𝒟))≡Func(I,(Func(𝒞,𝒟)))≃Func(I×𝒞,𝒟)≡Rep(I×𝒞;𝒟).On the right I×𝒞 is the “directed cylinder diagram” over 𝒞 (two copies of 𝒞, all directed pairs of copies of objects connected by a morphism, such that all resulting squares commute). Images of I×𝒞 in 𝒟 are pairs of functors 𝒞→𝒟 with components of a natural transformation between them.
In particular, if 𝒞 is a free category on a graph, then so is I×𝒞.
P. S. The same argument applies to representations of algebras, if we regard these are Vect-enriched categories and speak about Vect-enriched functor categories, etc.
Hmm, so in the case of algebras is it something more familiar like Mod(End(A)) (though with which algebra structure?)? For example, if 𝒞=Mod(A) for A an algebra, then the monoidal category of right exact endofunctors of Mod(A) is equivalent to Bimod(A). Do you see a somewhat more familiar characterization of these intertwiners?
Not sure if I understand this question now.
Let me highlight that in the language I used above, we have
Mod(A)≃Func(BA,Vect),with BA the one-object Vect-enriched category with endomorphism of that one object being A, and with the Vect-enriched functor category on the right.
RIght, and then the arrow category is the representations of the algebroid (BA→BA), but does that algebroid have a name? For example, if I take a groupoid 𝒢 and do [𝒢,Vect], these are the representations of a weak Hopf algebra, the groupoid algebra k[𝒢].
I would call this the “directed cylinder” over BA. Don’t know if in other areas this goes by another name.
1 to 9 of 9