Skip to the Main Content

Note:These pages make extensive use of the latest XHTML and CSS Standards. They ought to look great in any standards-compliant modern browser. Unfortunately, they will probably look horrible in older browsers, like Netscape 4.x and IE 4.x. Moreover, many posts use MathML, which is, currently only supported in Mozilla. My best suggestion (and you will thank me when surfing an ever-increasing number of sites on the web which have been crafted to use the new standards) is to upgrade to the latest version of your browser. If that's not possible, consider moving to the Standards-compliant and open-source Mozilla browser.

October 16, 2007

Loday and Pirashvili on Lie 2-Algebras (secretly)

Posted by Urs Schreiber

Zoran Škoda made me aware of

J. L. Loday, T. Pirashvili
The tensor category of linear maps and Leibniz algebras
Georgian Math. J. 5 3, 1998, 263-276 .

Even though the authors do not use that term, this is about (strict) Lie 2-algebras, namely Lie algebras internal to “Baez-Crans 2-vector spaces”, as well as more general 2-algebras: associative, Hopf, etc, all internal 2Vect2\mathrm{Vect}.

Interestingly, they conceive 2Vect2\mathrm{Vect} entirely in terms of 2-term chain complexes, but consider on 2Term2\mathrm{Term} the non-standard monoidal structure which makes it equivalent even as a symmetric monoidal category to 2Vect2\mathrm{Vect}.

This non-standard monoidal structure is easy to figure out, but I think it is worthwhile making it explicit. Loday and Pirashvili make great use of it, in particular in that they prove that with that structure 2Term2\mathrm{Term} becomes cartesian closed and explicitly compute the internal hom.

The issue of finding this non-standard monoidal structure on 2Term2\mathrm{Term} is what Dmitry Roytenberg is referring to in the first paragraph on p. 4 of his article on weak Lie 2-algebras (pdf, html).

So it’s maybe worthwhile making explicit a couple of easy but useful facts here. That’s what I shall try to do in the following.

The 2-category 2Vect 2\mathrm{Vect} (in the present context) is that of categories internal to vector spaces.

The 2-category 2Term 2\mathrm{Term} is that of chain complexes of length 2.

Notice that such chain complexes are really nothing but morphisms V 1fV 0 V_1 \stackrel{f}{\to} V_0 of vector spaces. So an object in 2Term2\mathrm{Term} is nothing but a linear map.

Morphisms of 2-term chain complexes are nothing but commuting squares between such linear maps

V 1 f V 0 V 1 f V 0. \array{ V_1 &\stackrel{f}{\to}& V_0 \\ \downarrow &&\downarrow \\ V'_1 &\stackrel{f'}{\to}& V'_0 } \,.

Therefore Loday and Pirashvili call this category not 2Term2\mathrm{Term} but

LM L M

(“Linear Maps”).

They would however probably have thought of a different name than that had they thought of the next level categorification of their setup. That they didn’t is apparently largely due to the fact, as far as I can see at least, that one of main facts they are interested in in this paper is that a couple of ordinary non-Lie algebras, like bialgebras, Leibniz algebras, become Lie objects internal to 2Term2\mathrm{Term}.

We will of course rephrase these statements in the form

Every Leibniz algebra gg gives rise to a strict Lie 2-algebra gg Lie. g \to g_{\mathrm{Lie}} \,.

Here g Lieg_{\mathrm{Lie}} is the Lie algebra obtained from gg by antisymmetrization.

So, recall first the standard fact described in HDA VI:

There is a (slightly non-canonical) equivalence 2Vect2Term. 2\mathrm{Vect}\simeq 2\mathrm{Term} \,.

Given a 2-vector space VV, we define the corresponding 2-term chain complex to be ker(s)tObj(V). \mathrm{ker}(s) \stackrel{t}{\to} \mathrm{Obj}(V) \,.

The 2-category 2Vect2\mathrm{Vect} has an obvious monoidal structure obtained by tensoring internal to Vect\mathrm{Vect}.

Let V 1V 2V_1 \otimes V_2 be the tensor product of two 2-vector spaces. Then, since ker s V 1V 2=0 f t(f)Id a 2Id a 10 f t(f) \mathrm{ker}_{s_{V_1 \otimes V_2}} = \left\langle \array{ 0 \\ \downarrow^f \\ t(f) } \otimes \mathrm{Id}_{a_2} \;\oplus\; \mathrm{Id}_{a_1} \otimes \array{ 0 \\ \downarrow^{f'} \\ t(f') } \right\rangle we find that V 1V 2V_1 \otimes V_2 corresponds to the 2-term chain complex given by ker(s 1)Obj(V 2)Obj(V 2)ker(s 2)t 1IdIdt 2Obj(V 1)Obj(V 2). \mathrm{ker}(s_1)\otimes \mathrm{Obj}(V_2) \oplus \mathrm{Obj}(V_2) \otimes \mathrm{ker}(s_2) \stackrel{t_1\otimes \mathrm{Id}\oplus \mathrm{Id}\otimes t_2}{\to} \mathrm{Obj}(V_1) \otimes \mathrm{Obj}(V_2) \,.

Equipped with this non-obvious tensor product

(V W)(V W):=((VW)(WV) WW) \left( \array{ V \\ \downarrow \\ W } \right) \otimes \left( \array{ V' \\ \downarrow \\ W' } \right) := \left( \array{ (V \otimes W' ) \oplus (W \otimes V') \\ \downarrow \\ W \otimes W' } \right)

2Term2\mathrm{Term} becomes a symmetric monoidal closed category. The equivalence 2Vect2Term2\mathrm{Vect} \simeq 2\mathrm{Term} extends to an equivalence of symmetric monoidal categories.

The internal Hom-object is define din the very last lines of Loday&Pirashvili’s paper.

By the equivalence, this then also determines the internal hom in Baez-Crans 2-vector spaces 2Vect. 2\mathrm{Vect} \,.

Loday and Pirashvili have a couple of further interesting observations concerning various kinds of 2-algebras as algebras internal to Baez-Crans 2-vector spaces.

Posted at October 16, 2007 2:02 PM UTC

TrackBack URL for this Entry:   http://golem.ph.utexas.edu/cgi-bin/MT-3.0/dxy-tb.fcgi/1463

2 Comments & 1 Trackback

Re: Loday and Pirashvili on Lie 2-Algebras (secretly)

I find this very intriguing. I have a couple of naive questions. First I could not see how to write down a dual. Did I miss something? Also you can iterate this and get a category whose objects are commutative cubes and with a symmetric tensor product that needs some cunning notation to write down. Do these have any significance?

Posted by: Bruce Westbury on October 17, 2007 9:46 PM | Permalink | Reply to this

Re: Loday and Pirashvili on Lie 2-Algebras (secretly)

First I could not see how to write down a dual.

Are you asking about dual 2-vector spaces? You’ll want to fix a “valuation object” first: a dual 2-vector space should be a hom-space of morphisms from a 2-vector space into some canonical 1-dimensional 2-vector space.

In the context of Baez-Crans 2-vector spaces there are maybe 1.5 different obvious choices one could make here. Fix one, and I’d think that, by the closedness result, you’d get a nice notion of dual 2-vector spaces.

Also you can iterate this and get a category whose objects are commutative cubes

Squares might have their use, too (as do crossed squares in the theory of nn-groups) but they are expected to give rise to 3-term chain complexes (as do crossed squares in the theory of nn-groups).

The main generalization here is:

nVectnTerm. n\mathrm{Vect} \simeq n\mathrm{Term} \,.

So it’s not really cubes, but nn-term chain complexes that generalize what we are talking about here.

A Lie algebra structure on an nn-term chain complex yields an nn-term L L_\infty-algebra.

See John and Alissa’s HDA VI for more.

Posted by: Urs Schreiber on October 17, 2007 11:22 PM | Permalink | Reply to this
Read the post BV for Dummies (Part V)
Weblog: The n-Category Café
Excerpt: Some elements of BV formalism, or rather of the Koszul-Tate-Chevalley-Eilenberg resolution, in a simple setup with ideosyncratic remarks on higher vector spaces.
Tracked: October 30, 2007 10:26 PM

Post a New Comment