4
$\begingroup$

I have two questions, one general and the other particular to the case I am interested in.

The 'homotopically correct' notion of equivalence of categories is an adjoint equivalence (from one point of view, at least). Given an (edit: equivalence invariant) algebraic structure on a groupoid $G$, say given by some 2-monad $M$ on the 2-category of groupoids, and an adjoint equivalence $f\colon X \to G$, consider the 2-groupoid of lifts of the algebraic structure along $f$, namely lifts of $f$ through the forgetful functor from the 2-category of $M$-algebras. Are there results that say this 2-groupoid is contractible? It seems very likely and also likely to be proven somewhere.

Secondly, I want to apply this to coherent 2-groups, which I guess are algebras for a 2-monad on groupoids, but would like to make sure.

$\endgroup$
2
  • $\begingroup$ Is this reference, arXiv:1403.3254, on fibrations of ordered groupoids, relevant to your question? $\endgroup$ Commented Mar 30, 2014 at 14:02
  • $\begingroup$ @RonnieBrown I don't think so, I'm afraid. $\endgroup$
    – David Roberts
    Commented Mar 30, 2014 at 20:59

2 Answers 2

2
$\begingroup$

Here is another answer which is perhaps more "in the spirit" of the question.

Suppose that $U: \mathcal{A} \to \mathcal{C}$ is a functor of 2-categories. Let $x$ and $y$ be objects of $\mathcal{C}$, and $f: x \to y$ a fixed morphism. Suppose that we have a lift $Y \in \mathcal{A}$ of $y \in \mathcal{C}$, that is an object such that $U(Y) = y$.

We can define the 2-category $Lift(f; Y)$ of lifts of $f$ relative to $Y$. It is the 2-category whose objects are arrows $F: X \to Y$ such that $U(F) = f$, whose morphisms are the obvious triangles (with 2-isomorphisms witnessing the commutativity of the triangle). The top edge of the triangle maps via $U$ to the identity morphism of $x$. The 2-morphisms are maps of the top edges of triangles such that the obvious diagram commutes and such that the 2-morphisms map via $U$ to the identity 2-morphism of the identity 1-morphism of $x$.

I think this is the 2-category of lifts that you meant, or perhaps you meant to restrict just to the invertible part of it. Either way...

Now suppose that:

  1. The functor $U$ is conservative in the following sense: a 1-morphism of $\mathcal{A}$ is an equivalence in $\mathcal{A}$ if and only if its image under $U$ is an equivalence in $\mathcal{C}$; a 2-morphism of $\mathcal{A}$ is an isomorphism if and only if its image under $U$ is an isomorphism; and two 2-morphisms of $\mathcal{A}$ are equal if and only if they are equal after applying $U$. (This is really an analog of being conservative and faithful).

  2. You can lift your structure along equivalences and invertible transformations. Meaning that if $f:x \to y$ is in $\mathcal{C}$ and $Y \in \mathcal{A}$ lifts $y$, then there exists some lift of $f$ to $\mathcal{A}$. Moreover if $\alpha: f \cong g$ is a 2-isomorphism in $\mathcal{C}$ and you can lift $f$ and $g$ to parallel arrows in $\mathcal{A}$, then there exists some compatible lift of $\alpha$ to $\mathcal{A}$.

  3. The morphism $f$ is an equivalence.

Then the 2-category $Lift(f; Y)$ is contractible.

I already gave an example where $\mathcal{A}$ is the category of algebras for a 2-monad where there can fail to be lifts.

However in your situation for coherent 2-groups you know that you can transfer the structure along equivalences and 2-isomorphism. Moreover we know that the forgetful functor from coherent 2-groups to groupoids is conservative in the sense I describe above. Hence the 2-category $Lift(f; Y)$ of lifts is contractible in this case.

$\endgroup$
3
  • $\begingroup$ Ah, excellent. Thank you, and apologies for the slightly grumpy tone of my other comments! $\endgroup$
    – David Roberts
    Commented Apr 2, 2014 at 1:41
  • $\begingroup$ Amongst the properties mentioned by Chris above, the only one that may fail for the strict algebras of a 2-monad is the ability to transport structure along equivalences. He mentioned flexible 2-monads: for a flexible 2-monad each pseudoalgebra is isomorphic to a strict algebra, with the same underlying object. Since one can transport pseudoalgebra structure along equivalences in general, the same holds for strict algebras for flexible 2-monads. This is discussed in "Monoidal functors generated by adjunctions..." by Kelly and Lack, for instance. $\endgroup$
    – john
    Commented Apr 3, 2014 at 7:51
  • 1
    $\begingroup$ That coherent 2-groups are the algebras for a flexible 2-monad on groupoids is not hard to see using presentations of 2-monads - it can be built as a flexible (indeed pie) colimit of free 2-monads on signatures. $\endgroup$
    – john
    Commented Apr 3, 2014 at 7:54
1
$\begingroup$

This is true in your example of coherent 2-groups. There are many sources that show that you can transfer monoidal structures along adjoint equivalences. You just need to adapt those to also transfer the coherent inverses, which should not be too difficult conceptually but might be tedious.

However the general statement is false, I believe.

Consider the 2-monad whose algebras are strict group objects. In other words you have a strict monoidal structure and strict inverses. Note that every coherent 2-group is equivalent to one of these. If it were true that you could transfer this structure along an adjoint equivalence, then this would imply that any coherent 2-group is equivalent to a 2-group which is both strict and skeletal. You first replace by a strict 2-group and then you apply the transference result (with respect to strict group objects) with X skeletal.

I suspect that it might be true if your 2-monad satisfies some kind of cofibrancy (aka flexible) condition. Have you looked in Blackwell-Kelly-Power?

$\endgroup$
4
  • 1
    $\begingroup$ If I recall correctly, folklore says you can transport pseudo-algebra structures along (adjoint) equivalences. $\endgroup$
    – Zhen Lin
    Commented Mar 30, 2014 at 16:47
  • $\begingroup$ That wasn't my question, as I know I can transfer 2-group structure across an adjoint equivalence. Rather, I want to consider the whole 2-category of transferred structures, and whether this is contractible. And as @Zhen Lin said, I'm really interested in pseudoalgebras, since these are the appropriate notion of algebra in a 2-category. $\endgroup$
    – David Roberts
    Commented Mar 30, 2014 at 20:57
  • $\begingroup$ I believe I gave an example of structure such that the 2-category of transferred structures along certain f's is empty, hence not contractible. $\endgroup$ Commented Mar 30, 2014 at 22:57
  • $\begingroup$ OK, but it's not in the spirit of the question :-) Can I retcon to 'algebraic structure which is invariant under equivalence of categories'? $\endgroup$
    – David Roberts
    Commented Mar 31, 2014 at 1:56

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .