It is well known that a group object in a category $C$ (with terminal object $1$ and such that any two objects of $C$ have a product) is defined as an object $G$ in $C$ with the following morphisms:

$m:G \times G \rightarrow G$, $e:1 \rightarrow G$, $\mathit{inv}: G \rightarrow G$

satisfying some conditions modelled on the group axioms such that $m$ behaves as the multiplication map, e behaves as the identity and $\mathit{inv}$ behaves as the inverse map.

The group action (right) of a group object $G$ on an object $X$ in $C$ can be defined as a morphism $\rho:X \times G \rightarrow X$ in $C$ such that the following two diagrams are commutative:

enter image description here enter image description here

where $\mathit{id}_X$, $\mathit{id}_G$ are identity morphisms at $X$ and $G$ respectively and $\mathit{pr}_1$ is the first projection on $X$ from the product $X \times 1$.

My Question are the following:

(1) What is the analogue of the above notion in a $2$-category? I couldn't find any literature in this direction.

So I tried to guess it's definition roughly in the following way:

Let $C$ be a 2-category (with terminal object $1$ and such that any two objects of $C$ have a product in the context of 2- category as mentioned in https://ncatlab.org/nlab/show/2-limit#2limits_over_diagrams_of_special_shape ). I define a group object in the 2-category $C$ as an object $G$ with the following 1-morphisms:

$m \in C(G \times G ,G)$, $e \in C(1,G)$, $inv \in C(G,G)$ satisfying some conditions similar as above but in this case every equality in the conditions will be replaced by an invertible 2-morphism satisfying certain appropriate coherent conditions.

Correspondingly the action of $G$ on an object $X$ in $C$ will be defined exactly in the same way but the above two diagrams will be commutative only upto invertible 2-morphisms.

Is my guess correct?

Even if it is correct but writing all the details (taking care of all the invertible 2 morphisms) seems very complicated to me and seems an inappropriate definition to work with.(Both in the context of strict 2-category and bicategory)

So what should be an appropriate definition for a group object in a 2-category and its corresponding action on an object? (Both in the context of strict 2-category and bicategory)


It is well known that the Strict 2-Group is a group object in Cat(when Cat is considered as a 1 -category or a usual category ).

But then

(2) What is the group object in Cat (when Cat is considered as strict 2-category)?

I would be also very grateful if someone can refer some literatures in this direction.

Thank you.

  • 1
    $\begingroup$ 4 upvotes, 4 bookmarks is a good number :D Hope some one answers your question.. $\endgroup$ May 9 '20 at 5:01
  • $\begingroup$ @PraphullaKoushik I also hope that :D $\endgroup$ May 9 '20 at 5:15

I'll have a go at answering your question (although higher category-theory is absolutely not my area of expertise).

The conditions for $(G,m,e,inv)$ to be a group object is stipulated by the following relations

  1. $m\circ (e\times inv)\circ\Delta=m\circ (inv\times e)\circ\Delta=id_G$
  2. $m\circ (m\times id_G)=m \circ (id_G\times m)$
  3. $m\circ(e\times id_G)=m\circ(id_G\times e)=id_G$

where $\Delta$ is the diagonal map and $(e,id_G)$ is the obvious map $G\to G\times G$.

Writing these out in full is a bit pedantic but it's important. The reason for this is that we can see that adding 2-morphisms to obtain a strict 2-category doesn't change anything about the structure of $G$. In particular we don't gain any structure from the 2-category case. Thus the strict 2-groups are actually group objects in $\mathbf{Cat}$ viewed as a strict 2-category.

There are other ideas for "group-like" things in higher categories. You can ask for things like weak 2-groups where we consider weak 2-categories and thus only require the composition of morphisms to be associative and unital up to some 2-isomorphism. In particular a weak 2-group is a monoidal category with all morphisms invertible such that for any object $x$ we have $x\otimes x^{-1}$ and $x^{-1}\otimes x$ are only isomoprhic to the tensor unit 1. (This corresponds to (1).) We can also define the notion of a coherent 2-group where we make specific choices for $x^{-1}$ and specific isomorphisms $x\otimes x^{-1}\to 1$ and $1\to x\otimes x^{-1}$ such that these form an adjoint pair. All of this is developed in this paper of Baez's & Lauda's (see definition 20).

In particular, a group object (or a strict 2-group) in a 2-category with finite products is a coherent 2-group $G$ except the natural isomorphisms defining the "coherent structure" of $G$ are all simply the identity (see Definition 29 of the above paper).

Another reference which spells out different group-like structures in 2-categories as well as how these things act on categories $X$ is given in this paper by Morten's & Picken's.

edit: edited for clarity (see comments below)

  • 1
    $\begingroup$ @AdittyaChaudhuri Note that in Definition 29 at the end they state that a strict 2-group in a 2-category is "...a group in the underlying category of this 2-category" which is what I was trying to communicate above. I'm glad you are satisfied! :) $\endgroup$
    – asd
    May 10 '20 at 16:40
  • 1
    $\begingroup$ Thanks for explaining. $\endgroup$ May 10 '20 at 16:42
  • 1
    $\begingroup$ For the sake of completeness of your answer can you please add the notion of definition 29 in arxiv.org/pdf/math/0307200.pdf along with the reference? Hope this will help the future readers. Thank you. $\endgroup$ May 10 '20 at 16:54
  • 1
    $\begingroup$ @AdittyaChaudhuri Let me know if these edits work and whether you'd like me to make any more changes. And thanks for the feedback! $\endgroup$
    – asd
    May 10 '20 at 17:17
  • 1
    $\begingroup$ It looks fine. :) $\endgroup$ May 10 '20 at 17:30

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.