Not signed in (Sign In)

Start a new discussion

Not signed in

Want to take part in these discussions? Sign in if you have an account, or apply for one below

• Sign in using OpenID

Site Tag Cloud

Vanilla 1.1.10 is a product of Lussumo. More Information: Documentation, Community Support.

• CommentRowNumber1.
• CommentAuthorUrs
• CommentTimeSep 14th 2010

added stuff to 2-groupoid.

• CommentRowNumber2.
• CommentAuthorTim_Porter
• CommentTimeSep 14th 2010

Fixed two typos there.

• CommentRowNumber3.
• CommentAuthorUrs
• CommentTimeSep 14th 2010

Thanks.

• CommentRowNumber4.
• CommentAuthorTobyBartels
• CommentTimeSep 14th 2010
• (edited Sep 14th 2010)

It seems to me that most of the stuff about strict $2$-groupoids should be at your new article strict 2-groupoid. I have moved it.

• CommentRowNumber5.
• CommentAuthorUrs
• CommentTimeSep 14th 2010

Thanks, good point.

Turns out some people here at ESI are desperate to finally understand what on earth a 2-groupoid might be. It seems the $n$Lab entries in their present state are not helping them much (maybe unsurprisingly). Maybe this motivates somebody to show off his expository skills and expand the entries a bit.

• CommentRowNumber6.
• CommentAuthorTobyBartels
• CommentTimeSep 15th 2010
• (edited Sep 15th 2010)

Well, I could write, at bigroupoid, a detailed definition similar to the one that I wrote at bicategory. I’m not sure that this would be very enlightening, however.

In the meantime, I’ve written a definition paralleling the first definition at bicategory. So this has everything except the coherence laws, written in terms of natural transformations between functors between hom-groupoids.

• CommentRowNumber7.
• CommentAuthorDavidRoberts
• CommentTimeSep 15th 2010

[bigcategory]??

• CommentRowNumber8.
• CommentAuthorTobyBartels
• CommentTimeSep 15th 2010

Fixed.

• CommentRowNumber9.
• CommentAuthorUrs
• CommentTimeSep 15th 2010
• (edited Sep 15th 2010)

Thanks!

i have added now to 2-groupoid a description of 3-coskeletal Kan complexes, mentioning how the existence of the 3-simplices gives the composition operation and the unique 4-simplices filling any 3-sphere of 3-simplices the associativity coherence law.

By Duskin’s result this is equivalent to the definition of a bigroupoid, but much simpler to state.

• CommentRowNumber10.
• CommentAuthorTim_Porter
• CommentTimeSep 15th 2010

I have added a simple example to strict 2-groupoid, in the hope that it helps.

• CommentRowNumber11.
• CommentAuthorTim_Porter
• CommentTimeSep 15th 2010
• (edited Sep 15th 2010)

@Urs Re: No 5 above. Can you tell us what the sticking point for people in understanding 2-groupoid seems to be. Analysing that may help us improve the entries. Perhaps an ’intro’ entry would help with more examples discussed in some detail. (I like the intro entries that have recently been written as they are more discursive and go through things at the ’for dummies’ level… more suitable for some of us :-), i.e. yours truly!)

• CommentRowNumber12.
• CommentAuthorUrs
• CommentTimeSep 15th 2010

Can you tell us what the sticking point for people in understanding 2-groupoid seems to be.

I can give some indications.

The main problem is a general unfamiliarity with the general kind of toolset used to handle these things. There are people who used to entertain attitude that category theory is all just nonsense, and so they struggle now when one comes to a situation where one needs to be able to freely talk about functors and natural transformations, say in the definition of composition in a bigroupoid, or say in the definition of simplicial sets and simplicial homotopy groups.

• CommentRowNumber13.
• CommentAuthorEric
• CommentTimeSep 15th 2010

Can you tell us what the sticking point for people in understanding 2-groupoid seems to be.

As a certified dummy, I can say that thinking of a group as a one-object groupoid is beautiful. It makes everything click. I love it. Plus, once you understand a group as a one-object groupoid, it gives “for free” a pretty decent understanding of groupoids more generally.

However, the same is not necessarily true for 2-groupoids. I don’t think that thinking of a 2-group as a one-object 2-groupoid helps with intuition much because few if any (certainly not me) have any kind of intuition for 2-categories. Once you understand strict 2-categories, you get 2-groupoids for free, but that first step is a barrier.

You can read it over and over, but nothing clicked for me until I started drawing tons of pictures of baby strict 2-categories. Then I started getting some sense for what horizontal composition is. That is kind of the key stumbling block for dummies I think, i.e. horizontal composition. I don’t claim to have 2-groupoids fully under control yet, but horizontal composition is a barrier for dummies. The nLab page is a bit bare on the subject. If I find a moment (it is 10pm and I’m still in the office with no indication I’ll be going home any time soon) I’d try to add some of my hard-earned wisdom.

• CommentRowNumber14.
• CommentAuthorTim_Porter
• CommentTimeSep 15th 2010

@Eric I added something to strict 2-groupoid earlier today and would welcome your reaction to it. It looked at the example of a congruence coming from a group homomorphism. The horizontal composition is from the transitive nature of the congruence and the vertical composition from the group multiplication. Everything is explicit but general (i.e. there are two groups and a homomorphism between them as input, not two specific groups …., as the point might get lost in the example.)

• CommentRowNumber15.
• CommentAuthorTim_Porter
• CommentTimeSep 15th 2010

@Urs I suggest that I (and anyone else who feels like helping) have a go at writing something with such an audience in mind. I have not thought out what to say but they are an important group of people and it would be good to see if we can help them … mostly for the challenge.

• CommentRowNumber16.
• CommentAuthorUrs
• CommentTimeSep 15th 2010

Tim,

yes. It seems the times are such that good exposition in nLab entries of basic concepts of higher categories would find a comparatively large number of thankful readers. Something that wasn’t true in this way a while back, I think.

• CommentRowNumber17.
• CommentAuthorAlexHoffnung
• CommentTimeOct 12th 2010
I think the equivalent definition of a strict 2-groupoid as a Grpd-enriched category is confusing. This is mildy averted on the 2-groupoid page where it says "This is equivalently a certain type of Grpd-enriched category." Anyway, if I am not confused a Grpd-enriched category in the first form of enrichment that comes to mind is more like a (2,1)-category than a (2,0)-category.
• CommentRowNumber18.
• CommentAuthorMike Shulman
• CommentTimeOct 12th 2010

Yes, a Gpd-enriched category is exactly the same as a strict (2,1)-category, so a 2-groupoid = (2,0)-category is a special sort of Gpd-enriched category. Feel free to edit the page to clarify…

• CommentRowNumber19.
• CommentAuthorAlexHoffnung
• CommentTimeOct 13th 2010
So far I only changed the equivalent definition given on the strict 2-groupoid page. I changed "a Grpd-enriched category" to "a Grpd-enriched category in which all 1-morphisms are strictly invertible".

I left the remark on the 2-groupoid page, since it isn't wrong and people can find clarification on the strict 2-groupoid page.

That said, does my change sound correct? Is there a nicer way to say this? Something like "Grpd-enriched groupoid"?
• CommentRowNumber20.
• CommentAuthorTim_Porter
• CommentTimeOct 13th 2010

’Grpd-enriched groupoid’ sounds ok to me. I have a nagging thought that is probably silly. What is the usual definition of a V-enriched groupoid? I presume it must have an inversion morphism on each hom-object $C(a,b)\to C(b,a)$, but does that imply any further equations (other than the obvious one that it is an inversion!). I don’t think so, but I thought I would mention it just in case.

• CommentRowNumber21.
• CommentAuthorDavidRoberts
• CommentTimeOct 13th 2010

It would need to be the inverse! Namely, there would be a diagram expressing: $C(a,b) \stackrel{\Delta}{\to}C(a,b)\otimes C(a,b) \stackrel{id\otimes(-)^{-1}}{\to} C(a,b) \otimes C(b,a) \to C(a,a)$ factors through $I \to C(a,a)$ (and one the other way, to $C(b,b)$), but now that I write this, one needs the diagonal, which clearly doesn’t always exist. It does for the case at hand (as $Gpd$ is cartesian), though.

• CommentRowNumber22.
• CommentAuthorTim_Porter
• CommentTimeOct 13th 2010

@David. Aha! Good point. That may explain my feeling. A nice point would be that in many contexts of homological algebra we have approximations to the diagonal and that must give an interesting structure. Is there a formulation that avoids this, I wonder?

• CommentRowNumber23.
• CommentAuthorMike Shulman
• CommentTimeOct 14th 2010

I deleted one of the definitions of strict 2-groupoid, since it was not equivalent to the others, and added a remark about it to the “Properties” section.

Add your comments
• Please log in or leave your comment as a "guest post". If commenting as a "guest", please include your name in the message as a courtesy. Note: only certain categories allow guest posts.
• To produce a hyperlink to an nLab entry, simply put double square brackets around its name, e.g. [[category]]. To use (La)TeX mathematics in your post, make sure Markdown+Itex is selected below and put your mathematics between dollar signs as usual. Only a subset of the usual TeX math commands are accepted: see here for a list.

• (Help)