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.

November 29, 2007

On BV Quantization, Part VIII

Posted by Urs Schreiber

A good way to understand the relevance and meaning of the trinity of

ghostsfieldsanti-fields

in the BV-formulation of quantum field theory (part I, II, III, IV, V, VI, VII) is to interpret the story of the charged nn-particle in the world of Lie \infty-algebroids (as opposed to Lie \infty-groupoids), which have a restriction on the sign of the degrees of their morphisms, but then form the configuration space of fields by taking the internal hom in the category of arbitrarily graded things.

This is, to my mind, how to interpret one of the crucial points of AKSZ: The Geometry of the Master Equation and Topological Quantum Field Theory.

Here I start to talk about that, but it will probably take me more than one installment. Progress will be documented in On Lie \infty-modules and the BV complex.

Recall, maybe from QFT of the Charged nn-Particle: Definition, that quantum (field) theory is the study of the following setup:

- a Lie \infty-category par\mathrm{par}, called the parameter space or the nn-particle

- a Lie \infty-category tar\mathrm{tar}, called the target space

- an \infty-functor tarphas\mathrm{tar} \to \mathrm{phas} called the background field

(and possibly a symplectic structure on some of the \infty-groupoids, but this will be the topic of later installments)

which allows to form the configuration space

conf:=hom(par,tar). \mathrm{conf} := \mathrm{hom}(\mathrm{par}, \mathrm{tar}) \,.

The background field transgresses to this configuration space, where it then forms the (non-kinetic part of) the action of the nn-particle.

In practice, instead of Lie \infty-groupoids one will often turn to the corresponding Lie \infty-algebroids (see for instance On Lie NN-tegration and Rational Homotopy Theory).

Depending on various conventions, these are structures on (co)chain complexes of non-negative degree, and that makes good sense:

the stuff in degree kk are the tangents to the space of kk-morphisms of the original \infty-groupoid.

But when we forget categories and just look at chain complexes, then there is nothing more natural than allowing arbitrary gradings. This hence corresponds to considering categories with negative morphisms.

Luckily the Wizards have thought of that before. And have come up with a name for them: \mathbb{Z}-categories.

As far as I understand, there are currently precisely two persons on this planet who have seriously thought about \mathbb{Z}-categories: John Baez and Jim Dolan. Though the point is that in disguise at least \mathbb{Z}-groupoids are well known to topologists: \mathbb{Z}-groupoids correspond to spectra.

But let’s not worry about \mathbb{Z}-categories too much at this point. While they nicely highlight the depth of the step from non-negative to arbitrary grading, performing that step itself in the linearized world of complexes and dg-algebras is very simple. (Maybe too simple, actually. Nobody seems to discuss the deep importance of this simple step.)

Then we obtain a picture of this sort:

I have tried to summarize some of the properties of the category of (co)chain complexes that we need in section 2.

So before actually going in to the AKSZ description of the charged nn-particle, let’s have another look at our toy example setup from BV for Dummies and On Noether’s Second (BV, Part VI).

Recall that there we were dealing with a setup where

- configuration space conf=X\mathrm{conf} = X

is an ordinary manifold, and where the action

- SC(X)S \in C(X)

is an ordinary function on that manifold.

This induced for us the BV-complex which was a dg-algebra extension of this simple situation

0 d| M 2 * d| M 1 * d| M 0 * d| V 1 * d| 0 0 ker(dS()) Γ(TX) dS() C (X) ρ C (X)g * 0 2 1 0 1 antighosts antifields fields ghosts (three vector space) (Liealgebra) ( 3module ) \array{ 0 &\stackrel{d|}{\to}& M_{-2}^* &\stackrel{d|}{\to}& M_{-1}^* &\stackrel{d|}{\to}& M_{0}^* &\stackrel{d|}{\to}& V_{1}^* &\stackrel{d|}{\to}& 0 \\ \\ 0 &\to& \mathrm{ker}(d S(\cdot)) &\hookrightarrow& \Gamma(T X ) &\stackrel{d S(\cdot)}{\to}& C^\infty(X) &\stackrel{\rho}{\to}& C^\infty(X) \otimes g^* &\to& 0 \\ \\ && -2 && -1 && 0 && 1 \\ && antighosts && antifields && fields && ghosts \\ \\ && (three- &-& -vector- &-& space) && (Lie algebra) \\ \\ && (--- &-& -3-module- &-& ---) && }

(Here the d|d| denotes the result of restricting domain and codomain of dd to be V *M *V^* \otimes M^* (i.e. just the first tensor power). )

(which I have taken from Modules for Lie \infty-Algebras).

In order to fully appreciate the AKSZ-BV construction for the charged quantum nn-particle, we need to familiarize ourselves with the dual picture of this BV complex (section 4.3.5):

By simply working out what the dual differential is, one finds

With that simple example in mind, we should now see what happens when we form the configuration space of worldvolume fields of the charged nn-particle using the internal hom in arbitrarily graded chain complexes.

But this I’ll discuss next time.

Posted at November 29, 2007 8:46 PM UTC

TrackBack URL for this Entry:   http://golem.ph.utexas.edu/cgi-bin/MT-3.0/dxy-tb.fcgi/1518

50 Comments & 4 Trackbacks

Re: On BV Quantization, Part VIII

Cool stuff! You are clearly becoming quite a WizardTM yourself.

One pathetically minor question: when you write

Lie:GpdLieAlg *Lie : \infty Gpd \to \infty LieAlg^*

what’s the meaning of that little asterisk? It doesn’t mean ‘dual’, does it? Is it hinting at a footnote I can’t find??

Posted by: John Baez on November 30, 2007 2:05 AM | Permalink | Reply to this

Re: On BV Quantization, Part VIII

John said:

One pathetically minor question: when you write

Lie:GrpdLieAlg * Lie : \infty Grpd \to \infty LieAlg^*

whats the meaning of that little asterisk?

Sorry, that’s a typo. .

It’s a remnant of a thought process in which I first wanted to talk about a functor from \infty-groupoids to Chevalley-Eilenberg-algebras of Lie \infty-algebras.

But, while technically important, this is not something that should be dealt with in a semi-sentence in the caption of the above figure.

So I have erased that little star now. Thanks for catching that.

Posted by: Urs Schreiber on November 30, 2007 11:09 AM | Permalink | Reply to this

Re: On BV Quantization, Part VIII

I thought I’d already asked this, but in this context
why would you want to replace Hom by a \otimes?

and what does the word `induces’ as in AKSZ induces
mean?

abd remind me what is in AKSZ beyond BV
except for extension of the `manifold’ language?

Posted by: jim stasheff on November 30, 2007 2:06 AM | Permalink | Reply to this

Re: On BV Quantization, Part VIII

I thought I’d already asked this, but in this context why would you want to replace “HomHom” by a “\otimes”?

I gather you are asking why I care about the fact that – if everything is suitably finite – chain complexes are compact closed such that the internal hom satisfies the condition

hom(V,W)WV * \mathrm{hom}(V,W) \simeq W \otimes V^*

familiar from the category of vector spaces (as we have discussed at some length here and here, which I summarize in detail in section 2.3 of my notes, and for detailed discussion of which I am very grateful to Todd Trimble, who helped me look at it more systematically, and, finally ;-), also to John Baez)?

If so, the answer is: because it is a very handy fact, for instance for the following reasons:

- It is exactly the formal version of the way many people write a superfield ϕ:VW\phi : V \to W - not as an element of a hom-space in general, but explicitly as an element of WV *W \otimes V^*, namely an expression they write like ϕ=ϕ n(x)+ϕ n1(x 1)+ϕ n+1(x 1)+ \phi = \phi^n(x) + \phi^{n-1}(x_1) + \phi^{n+1}(x_{-1}) + \cdots with ϕ iW\phi^i \in W and x iV *x_i \in V^*. Like on the bottom of p. 11 of Dmitry’s review of AKSZ-BV.

- Moreover, using this fact we can transate Chevalley-Eileneberg algebras of Lie \infty-algebras, which come to us as complexes of homs CE(g)=(hom(gg,B)d CEhom(ggg,B)) CE(g) = ( \cdots \to hom(g \wedge g, B) \stackrel{d_{CE}}{\to} hom(g \wedge g \wedge g, B) \to \cdots ) into qDGCAs on (g *B) \wedge^\infty ( g^* \otimes B) which is not only convenient for practical computations but also – as I am beginning to try to argue in Something Like Lie-Rinehart \infty-pairs and the BV-complex – apparently a way to understand the BV-complex as a Lie Rinehart \infty-pair and, in the process, obtain one explanation/interpretation of the meaning of arbitrarily (positively and negatively) graded dg-algebras (“Lie \mathbb{Z}-groupoids”, if you like): the positive part, g *g^*, is a Lie \infty-algebra which acts on the chain complex BB in the non-positive degree.

That’s all. I do understand that hom(V,W)WV *\mathrm{hom}(V,W) \simeq W \otimes V^* relies on the complex being bounded and of finite rank in each degree (as Todd kept emphasizing, for instance here). But it’s still useful.

Posted by: Urs Schreiber on November 30, 2007 11:43 AM | Permalink | Reply to this

Re: On BV Quantization, Part VIII

Oh, OK for that application
but it really should be
(∧ ∞g)*⊗B)
instead of
∧ ∞(g *⊗B)
and that would avoid some of the technical conditions you and Todd mention

Posted by: jim stasheff on November 30, 2007 1:50 PM | Permalink | Reply to this

Re: On BV Quantization, Part VIII

but it really should be ( g *)B (\wedge^\infty g^*) \otimes B instead of (g *B) \wedge^\infty (g^* \otimes B )

Thanks for mentioning this! I tried to discuss this issue before, but failed to make myself clear:

I understand that (when everything is suitably finite), from hom(gg,B)d CEhom(ggg,B) \cdots \to hom(g \wedge g,B) \stackrel{d_{CE}}{\to} hom(g \wedge g\wedge g,B) \to \cdots we get, at face value, a differential on (Ig *)B. \wedge^\infty (I \oplus g^*) \otimes B \,.

But this extends, trivially, to all of (g *B) \wedge^\infty ( g^* \oplus B) doesn’t it?

And after being puzzled about what to make of this for a while, it seemed to occur to me that what is actually going on in BV formalism is that we form the full (g *B) \wedge^\infty ( g^* \oplus B) and then put a differential on that.

So this struck me as relevant. In particular, since the expression (g *B) \wedge^\infty ( g^* \oplus B) is in a way quite more “elegant” or “natural” than (Ig *)B. \wedge^\infty (I \oplus g^*) \otimes B \,. The reason is that, as it happens,

- g *g^* is always of positive degree

- while BB is always of non-negative degree

- such that they exactly fit together to form a complex

g *B g^* \oplus B

of arbitrary degree, with no overlap.

This seems to be good, because also when we are forming, for instance, the internal hom hom(V,M) \mathrm{hom}(V,M) we obtain one single thing in arbitrary degree. If we like, we can separate this as g *Bg^* \oplus B, and this way identify ghosts, fields and antifields. But the way it arises is as one unified thing, and it would seem unnatural (and, I think, for the general case even impossible, but this I am not sure about right now) to have the dg-structure on hom(V,W)\mathrm{hom}(V,W) equivalently encoded in one on (Ig *)B\wedge^\infty(I \oplus g^*)\otimes B.

So, you see, this issue you address is one I kept wondering about when starting to think about this a while ago. Let me put it like this:

how do we say “Lie \infty-algebra” in an arrow-theoretic way internal to cochain complexes? We say:

there is a positively graded cochain complex g * g^* together with an extension of the differential on (Ig *)\wedge^\infty(I \oplus g^*) to a complex ( (Ig *),d CE) (\wedge^\infty(I \oplus g^*), d_{CE}) such that the canonical projection ( (Ig *),d CE) g * \array{ (\wedge^\infty(I \oplus g^*), d_{CE}) \\ \downarrow \\ g^* } still exists (i.e is still a cochain map).

This is equivalent to saying that we define a list of maps d 0:g *g *[1] d_0 : g^* \to g^*[-1] d 1:g *(g *g *)[1] d_1 : g^* \to (g^*\wedge g^*)[-1] d 2:g *(g *g *g *)[1] d_2 : g^* \to (g^*\wedge g^* \wedge g^*)[-1] \vdots of “co-arity” 0n<0 \leq n \lt \infty such that d 0:g *g *[1] d_0 : g^* \to g^*[-1] coincides with the differential on g *g^* and which together form a map d:g * (Ig *)[1] d : g^* \to \wedge^\infty(I \oplus g^*)[-1] which we extend to a differential on all of d: (Ig ) (Ig ). d : \wedge^\infty(I \oplus g^) \to \wedge^\infty(I \oplus g^) \,.

Sorry for boring you with this, this is just a triviality formulated in a verbose way. But I am just trying to make myself clear, which I think I failed to do before.

Namely if we now think about what

a) people do in the BV literature

b) people do in the lietarure on L L_\infty-modules

we can think of both a) and b) as summarized neatly by just repeating the above construction, with the non-negatively graded cochain complex Ig *I \oplus g^* everywhere replaced by an arbitrary cochain complex g *B. g^* \oplus B \,.

Maybe I am seeing ghosts (pun intended). But right now I am pretty fond of this (admittedly simple) idea (but also simple ideas need to be thought).

There seems to be a triangle emerging

groupoids LieRinehartmodules BVcomplexes \array{ && \mathbb{Z}-groupoids \\ & \nearrow && \searrow \\ &Lie-Rinehart \infty-modules & \to & BV-complexes }

Somehow.

Posted by: Urs Schreiber on November 30, 2007 3:32 PM | Permalink | Reply to this

Re: On BV Quantization, Part VIII

It only occurs to me now that Jim’s remark above to which I gave a long-winded reply is apparently caused by another typo of mine.

Argh.

It seems here I wrote (g *B) \wedge^\infty(g^* \otimes B) instead of (g *B). \wedge^\infty(g^* \oplus B) \,.

Sorry for that.

Posted by: Urs Schreiber on November 30, 2007 7:22 PM | Permalink | Reply to this

Re: On BV Quantization, Part VIII

My comment also meant something more substantive: i said


(Lambda g)* otimes B

instead of

(Lambda g*) otimes B

and that would avoid some of the technical conditions you and Todd mention.
N.B. g)* not g*)
ie. alternating multilinear functions
ass in the original Chevalley-Eilenberg

Posted by: jim stasheff on December 1, 2007 1:36 AM | Permalink | Reply to this

Re: On BV Quantization, Part VIII

Urs wrote:

they write like
ϕ=ϕ n(x)+ϕ n−1(x 1)+ϕ n+1(x −1)+⋯
with ϕ i∈W and x i∈V *.

boy, did that confuse me
I thought that phi(x) meant the usul functional notation
but what Dmitry does is use (du)^i
to avoid confusion with du^i

Posted by: jim stasheff on November 30, 2007 2:03 PM | Permalink | Reply to this

Re: On BV Quantization, Part VIII

boy, did that confuse me I thought that phi(x) meant the usul functional notation but what Dmitry does is use (du)^i to avoid confusion with du^i

Actually, I need to take back what I said about how that is an example of using V *WV^* \otimes W instead of hom(V,W)\mathrm{hom}(V,W).

Actually, what Dmitry has there is rather the hom(V,W)\mathrm{hom}(V,W) version, obscured through the fact that he writes

ξ a=A adu+ \xi^a = A^a d u + \cdots

instead of

ϕ *:ξ aA adu+. \phi^* : \xi^a \mapsto A^a d u + \cdots \,.

I meant to point to some other literature, where people actually do use the V *WV^* \otimes W-notation. Let me try to seach for a better example…

Posted by: Urs Schreiber on November 30, 2007 7:34 PM | Permalink | Reply to this

Re: On BV Quantization, Part VIII

and what does the word ‘induces’ as in “AKSZ induces” mean?

Okay, thanks for asking. This is the part I still needed to talk about in more detail. I became too tired and ended the entry before actually saying anything AKSZ-wise.

So this is what I mean by “induces”:

Let for instance

par=TX=Lie(Π(X))\par = T X = Lie(\Pi(X)) be the tangent algebroid of a manifold, and let tar=Poiss(Y,ω) \mathrm{tar} = Poiss(Y,\omega) be the Poisson algebroid on some other manifold YY with symplectic form ω\omega, then the external Hom Hom(par,tar) \mathrm{Hom}(par,tar) is precisely the space of Lie algebroid morphisms from TXT X to Poiss(X,ω)\mathrm{Poiss}(X,\omega).

These algebroid morphisms turn out to be that subspace of the space of all maps from par to tar that can be regarded as the on-shell fields with respect to the action of the Poisson Sigma-model.

So, then one might want to find the corresponding BV formalism. What are the ghosts, what are the anti-ghosts, what is the BV differential, etc.

Now, AKSZ answer this question as follows: they say: the BV complex is simply obtained by not taking the external Hom, but by taking the hom hom(par,tar) \mathrm{hom}(par,tar) internal to arbitrarily graded complexes.

This contains the external Hom we had above, hence the solutions to the classical equations of motion, as the closed elements in degree 0 Hom(par,tar)=Z 0(hom(par,tar)). \mathrm{Hom}(par,tar) = Z^0(hom(par,tar)) \,. On top of that, it contains stuff in positive degree. This are the ghosts. And it contains stuff in negative degree. This are the anti-ghosts. It all drops out automatically from just forming the inner hom.

Compare the first paragraph in section 4 of Dmitry’s review.

And of course that does not depend on the Poisson sigma model. Choose another target, and get the BV complex for the corresponding field theory.

That’s what I mean by “AKSZ induces” the BV complex: we just form the configuration space of fields internally in arbitrarily grading conf=hom(par,tar) \mathrm{conf} = hom(par,tar) and by just turning the crank out drop all ghost, fields and the BV differential on them.

P.S. I should add that I am glossing over the following point:

I keep talking about the homs in chain complexes. But what we really need here are homs in monoids in chain complexes, namely dg-algebras. This doesn’t change much about the general point here, but I need to eventually discuss this in more detail.

And at some point I will actually want to ask the wizards: what is known, from an abstract category-theory point of view, about the category SymMonoids(Ch (A)) \mathrm{SymMonoids}(Ch^\bullet(A)) of symmetric monoids internal to chain complexes ??

And in particular, that sub-category which has only qDGCAs as objects, namely dg-algebras which are free as graded commutative algebras.

Posted by: Urs Schreiber on November 30, 2007 12:09 PM | Permalink | Reply to this

Re: On BV Quantization, Part VIII

Urs writes:

Thats what I mean by
AKSZ induces the BV complex: we just form the configuration space of fields internally in arbitrarily grading
conf=hom(par,tar)
and by just turning the crank out drop all ghost, fields and the BV differential on them.


P.S.
I should add that I am glossing over the following point:


I keep talking about the homs in chain complexes. But what we really need here are homs in monoids in chain complexes, namely dg-algebras. This doesn’t change much about the general point here, but I need to eventually discuss this in more detail.

I suspect you will need A_\infty maps, not just homs in dgas.

Like wise I’m amazed if the BV differential drops out, but I better read Dmitry first.

Posted by: jim stasheff on November 30, 2007 2:00 PM | Permalink | Reply to this

Re: On BV Quantization, Part VIII

Urs wrote:

And at some point I will actually want to ask the wizards: what is known, from an abstract category-theory point of view, about the category SymMonoids(Ch (A)) \mathrm{SymMonoids}(Ch^\bullet(A)) of symmetric monoids internal to chain complexes??

Okay, I can take a hint — you know I feel duty-bound to reply to any post that contains the word ‘wizard’, and you want me to say something helpful before you struggle through an enormous nasty calculation.

I’ve never heard of a “symmetric monoid”. You probably mean either “commutative monoid” or its categorified version, “symmetric monoidal category”.

I’m guessing you mean the former, though I have no idea what ‘Λ \Lambda^\infty’ means, and that would be a helpful clue.

We can talk about commutative monoids internal to any symmetric monoidal category CC.

(For example: when CC is the category of chain complexes with its usual symmetric monoidal structure, these are supercommutative differential graded algebras.)

We always get a category CommMon(C)CommMon(C) of commutative monoids in CC, with monoid homomorphisms as the morphisms. This category is always symmetric monoidal in its own right.

We can talk about the free commutative monoid SXS X on an object XX whenever CC also has coequalizers and (countable) coproducts:

SX=1+X+X 2/2!+S X = 1 + X + X^{\otimes 2}/2! + \cdots

You just take the symmetrized tensor powers of XX and add them up.

(For example, when CC is the category of chain complexes, SXS X is the free supercommutative differential graded algebra on the chain complex XX.)

So, under these hypotheses, we get a ‘free commutative monoid’ functor

S:CCommMon(C)S: C \to CommMon(C)

which is left adjoint to the forgetful functor

U:CommMon(C)CU: CommMon(C) \to C

There’s a lot more one can say about all this, especially if we keep increasing the number of hypotheses on our category CC. But, I’ll stop here and check to see if this is at all relevant to your interests.

By the way, the ‘free commutative monoid’ functor

SX=1+X+X 2/2!+S X = 1 + X + X^{\otimes 2}/2! + \cdots

really deserves to be called e Xe^X, for obvious reasons. It’s a categorification of the exponential function. And, in the case when C=HilbC = Hilb, a very similar functor (but not quite the same, for annoying technical reasons) is called taking the ‘Fock space’ of XX. So, taking the Fock space of a Hilbert space is like exponentiation!

I was taught this by Irving Segal, who had no interest whatsoever in category theory.

I later went crazy iterating this Fock space construction in my webpage on nth Quantization. If we start with the 0-dimensional Hilbert space, and take its Fock space, and the Fock space of that, and so on, we get some very interesting and famous things.

This webpage ends with a puzzle that nobody has ever solved to my satisfaction: what’s a nice description of the Fock space of the Fock space of the Fock space of the Fock space of the 0-dimensional Hilbert space?

But, I digress… you haven’t seen anyone iterating BV quantization, have you?

Speaking of iterating, here’s an easier puzzle: what’s CommMon(CommMon(C))CommMon(CommMon(C)) like?

Posted by: John Baez on December 1, 2007 6:36 PM | Permalink | Reply to this

Re: On BV Quantization, Part VIII

We can talk about the free commutative monoid SX on an object X whenever C also has coequalizers and (countable) coproducts:

SX=1+X+X⊗2/2!+⋯

You just take the symmetrized tensor powers of X and add them up.

when can you get the corresponding symmetric comonoid replacing e.g. X⊗2/2! by the symmetric invariants?

Posted by: jim stasheff on December 1, 2007 8:24 PM | Permalink | Reply to this

Re: On BV Quantization, Part VIII

Jim wrote:

when can you get the corresponding symmetric comonoid replacing e.g. X⊗2/2! by the symmetric invariants?

Just remove the prefix “co” wherever it appeared and insert wherever it didn’t!

I formed the free commutative monoid on XX by taking the coequalizers of the symmetric group actions on X nX^{\otimes n} (the symmetric coinvariants) and then taking the coproduct of these for all nn.

So, we form the free commutative comonoid on XX by taking the equalizers of the symmetric group actions on X nX^{\otimes n} (the symmetric invariants) and then taking the product of these for all nn.

(If some wise guy says I should’ve removed the ‘co’ in ‘commutative’, I’ll slap ‘em.)

The point is that a commmutative comonoid in a symmetric monoidal category CC is the same as a commutative monoid in the opposite category, C opC^{op}.

should it be denoted 1/(1X)1/(1-X)?

No, 1/(1X)1/(1-X) is a good name for the free monoid on an object XX, since this is just

1+X+X 2+1 + X + X^2 + \dots

No symmetric group actions are involved here! We can do this in any monoidal category with (countable) coproducts.

Posted by: John Baez on December 1, 2007 9:03 PM | Permalink | Reply to this

Re: On BV Quantization, Part VIII

No, 1/(1X)1/(1&#8722;X) is a good name for the free monoid on an object X, since this is just 1+X+X 2+1 + X + X^2 + \ldots No symmetric group actions are involved here! We can do this in any monoidal category with (countable) coproducts.

There’s an important caveat here: the monoidal product XYX \otimes Y must preserve countable coproducts in each of its separate arguments XX and YY in order for this nice construction to work. In practice this is often automatic, especially in cases where the monoidal category is biclosed (which means XX \otimes - and Y- \otimes Y are left adjoints to appropriate Hom-functors – in that case, the left adjoints preserve any colimits which happen to exist, including of course countable coproducts).

But for the same reason, you typically have to be careful in how you construct cofree comonoids in a monoidal category. Even if the monoidal category CC has countable products (so that C opC^{op} has countable coproducts), it is typically not the case that the tensor product will preserve them in separate arguments, which is what is needed in order for the dual of John’s construction to apply.

In rough outline, according to the paper by Getzler and Goerss that I referred to here, the cofree comonoid Cof(V)Cof(V) on a chain complex VV of finite type is given by taking the continuous dual of the profinite completion of the tensor algebra (see page 5). For a general chain complex VV, they give a proof that Cof(V)Cof(V) is given by the filtered colimit formula

Cof(V)=colim αCof(V α)Cof(V) = colim_{\alpha} Cof(V_{\alpha})

where the V αV_{\alpha} range over subcomplexes of finite type. (I’ll assume for now that this is a correct description, although I haven’t checked all the details.)

Posted by: Todd Trimble on December 1, 2007 10:15 PM | Permalink | Reply to this

Re: On BV Quantization, Part VIII

Whoops. And contrary to my original post, I guess you also need the tensor product to distribute over some colimits to get

SX=1+X+X 2/2+ S X = 1 + X + X^{\otimes 2}/2 + \cdots

to be a (commutative) monoid.

Posted by: John Baez on December 2, 2007 12:03 AM | Permalink | Reply to this

Re: On BV Quantization, Part VIII

John wrote:

By the way, the ‘free commutative monoid’ functor SX=1+X+X 2/2!+.. SX=1+X+X^{\otimes 2}/2!+ .. really deserves to be called e Xe^X, for obvious reasons. It’s a categorification of the exponential function. And, in the case when C=Hilb, a very similar functor (but not quite the same, for annoying technical reasons) is called taking the ‘Fock space’ of X. So, taking the Fock space of a Hilbert space is like exponentiation!

I know it’s pretty shameless chiming in to talk about my own work, but it’s a pretty fact that this Fock space endofunctor will in fact translate product to tensor product, for sufficiently well-behaved cases: there are canonical isomorphisms i A,B:S(AB)S(A)S(B)i_{A,B}:S(A \oplus B) \simeq S(A) \otimes S(B). This just follows from the fact that SS is an adjoint functor, constructing the free cocommutative comonoid, so must preserve products — and the tensor product is the categorical product in the category of commutative comonoids.

Anyway, the nice thing is that these Fock spaces will sometimes come with raising and lowering operators, satisfying the commutation relations — and these are canonically defined precisely when the isomorphisms i A,Bi_{A,B} are unitary!

Posted by: Jamie Vicary on December 1, 2007 9:35 PM | Permalink | Reply to this

Re: On BV Quantization, Part VIII

Okay, I can take a hint

Thanks! Hope I wasn’t too obnoxious.

I’ll stop here and check to see if this is at all relevant to your interests.

Yes, it is!

So let me indicate where I want to go:

above I admitted that I was being somewhat vague when I said that AKSZ form the internal hom in the world of dg-algebras.

It is sold that way, and it is “certainly morally true” (whatever that means). But now I want to get it completely straight. Or as straight as possible at least.

So: I want to understand the (if any) closed structure on

- dg-algebras

- quasi-free dgc-algebras (free graded-commutative algebras with differentials)

- dg-coalgebras

- quasi- free dg-coalgebras

and I want to undertand how the various more or less obvious constructions that one sees appear in the literature, or that one can dream up onself, fit into the general abstract nonsense.

I was busy most of the day with teaching duties. Now I will go to Todd’s detailed description of the closed structure on coDGAs and try to absorb that (thanks Todd! where would I be without the Café??).

A first glance revealed that it might take me a while to unwrap Todd’s description to the degree that I can see whether and how the expectation that I mention at then end of my previosu comment is realized here. All help is appreciated.

But, I digress… you haven’t seen anyone iterating BV quantization, have you?

I have seen BV quantization applied to the 2-particle… and then to the second quantization of the 2-particle!

(Ordinarily called “the string” and “string field theory”.)

In fact, Zwiebach’s BV treatment of bosonic string field theory seems to be one of the two outstanding successes of the BV formalism (next to the Poisson-sigma model quantization and everything related to that.)

But of course that’s not the explicit “iteratation” that you (and I! :-) want to see.

My personal idea is this:

“quantization” of sufficiently nice setups (namely sigma-models of an nn-dimensional thing propagating on something and coupled to an nn-background bundle) should really be a functor which eats transport functors and spits out quantum propagation functors. But the latter should be “transport functors” again. So we may try to feed them back into the machinery.

BV-formalism is a way to describe how this machine works. So once we really understand it, we’ll iteratre it.

In the comment second quantization to the thread The nn-Café Quantum Conjecture I made some comments on this idea of iterating this endofunctor transportfunctorsquantizationtransportfunctors transport functors \stackrel{quantization}{\to} transport functors and how this does seem to have a good chance to indeed capture the idea of second quantization.

Posted by: urs on December 3, 2007 7:49 PM | Permalink | Reply to this

Re: On BV Quantization, Part VIII

Urs wrote:

P.S. I should add that I am glossing over the following point: I keep talking about the homs in chain complexes. But what we really need here are homs in monoids in chain complexes, namely dg-algebras. This doesn’t change much about the general point here, but I need to eventually discuss this in more detail. And at some point I will actually want to ask the wizards: what is known, from an abstract category-theory point of view, about the category SymMonoids(Ch *(A))SymMonoids(Ch^{*}(A)) of symmetric monoids internal to chain complexes ??

As is often the case, I’m not following very well what is really going on in this discussion, so please treat this as a surface-level reaction. It sounds as if you hope there’s an internal HomHom, internal to the category of commutative dga’s (commutative monoids in the symmetric monoidal category of chain complexes), i.e., that there’s a closed category structure on the category of commutative dga’s.

That would be too much to hope for, I think, but it reminds me of some facts which I find very interesting:

  • The category CoDGACoDGA of cocommutative comonoids in the category of chain complexes is complete and cocomplete, and the forgetful functor to the category of chain complexes, U:CoDGACh,U: CoDGA \to Ch, admits a right adjoint (cofree cocommutative coalgebra construction). See this ps file by Ezra Getzler and Paul Goerss for details.
  • The category CoDGACoDGA is cartesian closed (the cartesian product is the tensor product in ChCh). I first learned this from Jim Dolan; I can give a proof of this and the following statement on request.
  • The category of commutative dga’s is enriched in the cartesian closed category CoDGACoDGA.

The last part means that if AA and BB are two commutative dga’s, there is a cocommutative comonoid Hom(A,B)Hom(A, B) in ChCh [which Martin Hyland once told me was called a “measure coalgebra”, if I remember correctly], defined by the following universal property: if CC is a cocommutative comonoid in ChCh, then the chain complex hom(C,B)hom(C, B) carries a commutative dga-structure, and there is a natural isomorphism

DGA(A,hom(C,B))CoDGA(C,Hom(A,B)).DGA(A, hom(C, B)) \cong CoDGA(C, Hom(A, B)).

I’m also vaguely wondering whether this can be at all related to what Jim Stasheff was just saying about the need to pass to A A_{\infty}… I’d have to think further before I can amplify.

Posted by: Todd Trimble on December 1, 2007 6:48 PM | Permalink | Reply to this

Re: On BV Quantization, Part VIII

I think you answered my last question before I answered it in re: symm comonoid

should it be denoted 1/(1-x)?

but the A_\infty question about maps is
the issue some would hide in the language of derived cats

Posted by: jim stasheff on December 1, 2007 8:27 PM | Permalink | Reply to this

what’s in a name?

Speaking of derived cats, suppose Newton had one and the word fluxion had caught on,
would we now have fluxed cats? financial fluxions might have seemed less esoteric to the casual investor

thoughts provoked by reading Neal Stephenson’s
Con-fusion

Posted by: jim stasheff on December 3, 2007 2:19 PM | Permalink | Reply to this

Re: On BV Quantization, Part VIII

It sounds as if you hope there’s an internal Hom\mathrm{Hom}, internal to the category of commutative dga’s (commutative monoids in the symmetric monoidal category of chain complexes), i.e., that there’s a closed category structure on the category of commutative dga’s.

Yes, that’s the kind of thing I am looking for.

Let’s think rational homotopy theory: DGAs are much like spaces. The morphisms between one space and another should be a space itself.

That would be too much to hope for, I think, but it reminds me of some facts which I find very interesting:

Great, thanks.

The category CoDGA of cocommutative comonoids in the category of chain complexes

I’d be just as happy, or even happier, to discuss CoDGAs instead of DGAs. When everything is suitably finite both should be equivalent (I guess in the ps that you linked to this statement is prop 1.7) and people tend to want to stick with the DGAs as long as possible (or even longer).

But if its just the issue of CoDGAs versus DGAs that prevents me from getting the closed structure that I am looking for, then I won’t hesitate to stick with CoDGAs.

The category CoDGA is cartesian closed

Cool. That’s what I want!

I first learned this from Jim Dolan; I can give a proof of this and the following statement on request.

I’d be very grateful if you indeed did provide the proof. Thanks!

(So this means that this statement is not to be found in the literature??)

I’ll mostly need this statement for CoDGAs which are free as CoGCAs, i.e. for which the underlying graded coalgebra (forgetting the codifferential) is free as a graded commutative coalgebra. But the more general the statement you can provide, the better.

For this last case, I know the following:

let V\vee^\bullet V be the free graded commutative coalgebra generated from a graded vector space VV. Write ( V,D)(\vee^\bullet V, D) for the CoDGA obatined by equiupping this coalgebra with a codifferential (degree ±1\pm 1, depending on your conventions) such that D 2=0D^2 = 0.

Then a morphism ( V,D V)( W,D W) (\vee^\bullet V, D_V) \to (\vee^\bullet W, D_W) is supposed to be a linear map which is both, a morphism of graded commutative coalgebras and a chain map.

The fact that it is a morphism of free CoGCAs implies that it is entirely fixed by its value on “cogenerators”, hence by maps f: VW f : \vee^\bullet V \to W (when beginning to think about this it helps to consider the same statement after dualizing everything, where it is more familiar).

Now, for the external hom these maps would be required to respect the differentials [D,f]=d Wfd VF=0[D,f] = d_W f - d_V F = 0. But for the internal hom we don’t put any restriction on ff and instead have the commutator with the codifferentials act as a new codifferential.

This way we should get a new CoDGA (hom Vect( ,W),[D,]) (\mathrm{hom}_{Vect}(\vee^\bullet,W), [D,-])

and I am thinking that this, or something closely related, should be the internal hom.

But there must be some things missing here. Not sure if the above will in general be free, for instance.

Anyway, this is roughly my present understanding. Now I would like to understand this more deeply. Thanks for whatever help you can offer.

Posted by: urs on December 3, 2007 12:59 PM | Permalink | Reply to this

Re: On BV Quantization, Part VIII

Urs wrote:
Let’s think rational homotopy theory: DGAs are much like spaces. The morphisms between one space and another should be a space itself.

There are more things in the world of rational dgca’s than are dreamt of in
spaces.

Schlessinger and I address this in
“The Lie Algebra Structure of Tangent Cohomology and Deformation Theory,”

Journal of Pure and Applied Algebra 38 (1985) 313-322. (with Michael Schlessinger)
and
in the expanded but never finished `cahier secret’ version
available on demand

Hom(A,B) is fine as a dgca but does not represnet a space in general becasue of those pesky maps of negative degree
so we get the appropriate space by truncating, but not jsut getting rid of the negative stuff - degree 0 gets handled in an ad hoc manner to get a space

Posted by: jim stasheff on December 3, 2007 3:16 PM | Permalink | Reply to this

Re: On BV Quantization, Part VIII

Jim Stasheff wrote:

Hom(A,B) is fine as a dgca

When you say this, precisely what product structure Hom(A,B)Hom(A,B)Hom(A,B)Hom(A,B)\otimes Hom(A,B) \to Hom(A,B) are you referring to?

but does not represnet a space in general becasue of those pesky maps of negative degree

Okay, thanks. So maybe it represents, in general, instead a spectrum??

so we get the appropriate space by truncating, but not jsut getting rid of the negative stuff - degree 0 gets handled in an ad hoc manner to get a space

I have the papers you mentioned here, may need to look at them more closely, though.

Posted by: urs on December 3, 2007 8:33 PM | Permalink | Reply to this

Re: On BV Quantization, Part VIII

I’d be very grateful if you indeed did provide the proof [of the fact that CoDGACoDGA is cartesian closed]. Thanks!

Okay, here goes. This proof is supposed to be very “soft” and general. But, you may need to grab pencil and paper to do some side calculations in order to follow this. (We might also humbly beseech the Grand Wizard to sprinkle a little magic sugar over this to make it go down easier, but he may be busy on other errands of mercy.)

By the way, in response to

(So this means that this statement is not to be found in the literature??)

I don’t know. But maybe Jim Dolan knows. I have a feeling this result may have been known to Tom Fox (a categorist at McGill) back in the 80’s, when he was publishing stuff on cocommutative comonoid theory.

First, for any symmetric monoidal category CC, the category CocCoc of cocommutative comonoids has cartesian products given by the monoidal product in CC. That is, if XX and YY are cocommutative comonoids, then their categorical product is XYX \otimes Y with the evident structure of cocommutative comonoid. The projections XYXX \otimes Y \to X, XYYX \otimes Y \to Y are formed with the help of the counit XIX \to I, YIY \to I, and the comultiplication XXXX \to X \otimes X plays the role of diagonal map. This is all well known.

Thus, to prove cartesian closure, we must construct, given two cocommutative comonoids YY and ZZ, a cocommutative comonoid Hom(Y,Z)Hom(Y, Z) together with an isomorphism

Coc(XY,Z)Coc(X,Hom(Y,Z))Coc(X \otimes Y, Z) \cong Coc(X, Hom(Y, Z))

natural in XX. (Then, by the usual yoga of representable functors, there is a canonical way of making Hom(Y,Z)Hom(Y, Z) functorial in its arguments YY and ZZ in such a way that the naturality extends to YY and ZZ as well.)

Now suppose that CC is symmetric monoidal closed. The extra key assumptions we need to prove CocCoc cartesian closed is that it has equalizers and that the underlying functor

U:CocCU: Coc \to C

has a right adjoint CofCof (the cofree cocommutative comonoid functor). Happily, these assumptions are satisfied in the case of present interest (CC = chain complexes), and for many other symmetric monoidal closed abelian categories, I believe.

Start with a map f:XYZf: X \otimes Y \to Z in CocCoc. This means we have two commutative diagrams in CC:

XY δ XY XYXY f ff Z δ Z ZZ \array{ X \otimes Y & \stackrel{\delta_{X \otimes Y}}{\to} & X \otimes Y \otimes X \otimes Y \\ f \; \downarrow & & \downarrow \; f \otimes f \\ Z & \stackrel{\delta_Z}{\to} & Z \otimes Z }

and

XY ε Xε Y I f ε Z Z \array{ X \otimes Y & \stackrel{\varepsilon_X \otimes \varepsilon_Y}{\to} & I \\ f \; \downarrow & \nearrow \; \varepsilon_Z & \\ Z & & }

where the comultiplication δ XY\delta_{X \otimes Y} in the first diagram is the composite

XYδ Xδ YXXYY1σ1XYXY.X \otimes Y \stackrel{\delta_X \otimes \delta_Y}{\to} X \otimes X \otimes Y \otimes Y \stackrel{1 \otimes \sigma \otimes 1}{\to} X \otimes Y \otimes X \otimes Y.

Let Z YZ^Y denote the internal hom in CC, and let f^:XZ Y\hat{f}: X \to Z^Y denote the transform of f:XYZf: X \otimes Y \to Z under the symmetric monoidal closed structure on CC. A little calculation shows that the square above transforms into a square in CC:

X δ X XX f^ ϕ Z Y δ Z Y (ZZ) Y \array{ X & \stackrel{\delta_X}{\to} & X \otimes X \\ \hat{f} \; \downarrow & & \downarrow \; \phi \\ Z^Y & \stackrel{\delta_{Z}^Y}{\to} & (Z \otimes Z)^Y }

where ϕ\phi is the composite

XXf^f^Z YZ Y 1(ZZ) YY(ZZ) δ Y(ZZ) YX \otimes X \stackrel{\hat{f} \otimes \hat{f}}{\to} Z^Y \otimes Z^Y \stackrel{\otimes_1}{\to} (Z \otimes Z)^{Y \otimes Y} \stackrel{(Z \otimes Z)^{\delta_Y}}{\to} (Z \otimes Z)^Y

in which 1\otimes_1 is a canonical map which expresses the enriched functoriality of the monoidal product.

Let f˜:XCof(Z Y)\tilde{f}: X \to Cof(Z^Y) be the unique map in CocCoc which lifts the arrow f^:XZ Y\hat{f}: X \to Z^Y through the counit π:Cof(Z Y)Z Y\pi: Cof(Z^Y) \to Z^Y of the adjunction UCofU \dashv Cof. The trick is to see that commutativity of the square above translates into saying that this f˜\tilde{f} equalizes a certain pair of arrows (Φ,Ψ)(\Phi, \Psi) in CocCoc of the form

Cof(Z Y)Cof((ZZ) Y).Cof(Z^Y) \stackrel{\to}{\to} Cof((Z \otimes Z)^Y).

Similarly (and I’ll leave this part to you), commutativity of the triangle above translates into saying that f˜\tilde{f} equalizes a certain pair of arrows in CocCoc of the form

Cof(Z Y)Cof(I Y).Cof(Z^Y) \stackrel{\to}{\to} Cof(I^Y).

Thus, the exponential Hom(Y,Z)Hom(Y, Z) in CocCoc we are after is ultimately expressed as an equalizer in CocCoc of a certain pair of comonoid maps of the form

Cof(Z Y)Cof((ZZ) Y)Cof(I Y)Cof(Z^Y) \stackrel{\to}{\to} Cof((Z \otimes Z)^Y) \otimes Cof(I^Y)

(where we are using the fact that the right side is a cartesian product in CocCoc).

The construction of Φ\Phi and Ψ\Psi is a bunch of abstract nonsense. The map

Φ:Cof(Z Y)Cof((ZZ) Y)\Phi: Cof(Z^Y) \to Cof((Z \otimes Z)^Y)

is the unique comonoid map which lifts the long horizontal composite Cof(Z Y)(ZZ) YCof(Z^Y) \to (Z \otimes Z)^Y you see at the bottom of

X δ X X 2 f˜ f˜f˜ Cof(Z Y) δ Cof(Z Y) 2 ππ (Z Y) 2 1 (ZZ) YY (ZZ) δ Y (ZZ) Y \array{ X & \stackrel{\delta_X}{\to} & X^{\otimes 2} & & & & & & \\ \tilde{f} \; \downarrow & & \downarrow \; \tilde{f} \otimes \tilde{f} & & & & & & \\ Cof(Z^Y) & \stackrel{\delta}{\to} & Cof(Z^Y)^{\otimes 2} & \stackrel{\pi \otimes \pi}{\to} & (Z^Y)^{\otimes 2} & \stackrel{\otimes_1}{\to} & (Z \otimes Z)^{Y \otimes Y} & \stackrel{(Z \otimes Z)^{\delta_Y}}{\to} & (Z \otimes Z)^Y }

If you look at it carefully, you should be able to see that the composite arrow from X 2X^{\otimes 2} to (ZZ) Y(Z \otimes Z)^Y in this last diagram is the map ϕ\phi we had above in the square. It follows that

Φf˜:XCof((ZZ) Y)\Phi \circ \tilde{f}: X \to Cof((Z \otimes Z)^Y)

is the unique comonoid map which lifts one of the legs of that square, namely

Xδ XXXϕ(ZZ) Y,X \stackrel{\delta_X}{\to} X \otimes X \stackrel{\phi}{\to} (Z \otimes Z)^Y,

through the counit π:Cof((ZZ) Y)(ZZ) Y\pi: Cof((Z \otimes Z)^Y) \to (Z \otimes Z)^Y of the adjunction UCofU \dashv Cof.

Similarly, using the π\pi-naturality square

Cof(Z Y) π Z Y Cof(δ Z Y) δ Z Y Cof((ZZ) Y) π (ZZ) Y, \array{ Cof(Z^Y) & \stackrel{\pi}{\to} & Z^Y \\ Cof(\delta_{Z}^Y) \; \downarrow & & \downarrow \; \delta_{Z}^Y \\ Cof((Z \otimes Z)^Y) & \stackrel{\pi}{\to} & (Z \otimes Z)^Y, }

it is easy to check, if we put Ψ=Cof(δ Z Y)\Psi = Cof(\delta_{Z}^Y), that Ψf˜:XCof((ZZ) Y)\Psi \circ \tilde{f}: X \to Cof((Z \otimes Z)^Y) is the unique comonoid map which lifts the other leg of that square,

Xf^Z Yδ Z Y(ZZ) Y.X \stackrel{\hat{f}}{\to} Z^Y \stackrel{\delta_{Z}^Y}{\to} (Z \otimes Z)^Y.

If you’ve followed me thus far, then hopefully it will be clear that f˜\tilde{f} equalizes Φ\Phi and Ψ\Psi iff that square above (translating the fact that f:XYZf: X \otimes Y \to Z preserves comultiplication) commutes. Thus, modulo the additional exercise which I left to you of carrying out a similar construction which translates preservation of the counit, we are finally done (!).

Posted by: Todd Trimble on December 3, 2007 6:51 PM | Permalink | Reply to this

Re: On BV Quantization, Part VIII

Todd,

thanks for taking the time to spell this out in so much detail.

I am afraid I nevertheless have a rather dumb question, which probably will show that I am missing the main point you were making. Since I can’t help it, here I go:

You start with a morphism f:XYZ f : X \otimes Y \to Z in CocCoc.

We are looking for the transform of that to XHom Coc(Y,Z) X \to \mathrm{Hom}_{Coc}(Y,Z) and in fact for the definition of Hom Coc(Y,Z)\mathrm{Hom}_{Coc}(Y,Z).

Right?

You start by considering the transform f^:XZ Y \hat f : X \to Z^Y of ff with respect to the closed structure of the ambient category CC.

Then you consider its unqiue lift f˜:XCof(Z Y) \tilde f : X \to Cof(Z^Y) such that Cof(Z Y) f˜ π X f^ Z Y, \array{ Cof(Z^Y) \\ \uparrow^{\tilde f} & \searrow^{\pi} \\ X &\stackrel{\hat f}{\to}& Z^Y } \,, if I understand correctly.

At this point something seems to be clear to you which is not clear to me.

Because, if I understand correctly, in the remainder you demonstrate that f˜\tilde f indeed exists (by showing that it is an equalizer and using the assumption that all equalizers exist). But it is not clear to me how the mere existence of f˜\tilde f solves our problem. Why does it?

You are not saying that Hom Coc(Y,Z)Cof(Z Y) Hom_{Coc}(Y,Z) \simeq Cof(Z^Y) are you?

Posted by: urs on December 3, 2007 9:14 PM | Permalink | Reply to this

Re: On BV Quantization, Part VIII

No, I’m not claiming Cof(Z Y)Cof(Z^Y) is the desired internal hom (nor that it is equivalent to the desired hom). The internal hom is a subcomonoid of Cof(Z Y)Cof(Z^Y), namely an equalizer of the form

Hom Coc(Y,Z)Cof(Z Y)Cof((ZZ) Y)Cof(I Y)Hom_{Coc}(Y, Z) \hookrightarrow Cof(Z^Y) \stackrel{\to}{\to} Cof((Z \otimes Z)^Y) \otimes Cof(I^Y)

where each of the parallel arrows Cof(Z Y)Cof((ZZ) Y)Cof(I Y)Cof(Z^Y) \to Cof((Z \otimes Z)^Y) \otimes Cof(I^Y) is determined by a pair of maps, one to Cof((ZZ) Y)Cof((Z \otimes Z)^Y) and the other to Cof(I Y)Cof(I^Y) (since \otimes is the cartesian product in CocCoc). The bulk of my comment was about constructing the parallel pair

(Φ,Ψ):Cof(Z Y)Cof((ZZ) Y)(\Phi, \Psi): Cof(Z^Y) \stackrel{\to}{\to} Cof((Z \otimes Z)^Y)

and I left it as an exercise to construct the other parallel pair

Cof(Z Y)Cof(I Y).Cof(Z^Y) \stackrel{\to}{\to} Cof(I^Y).

Sorry if that was unclear. Let me know if there’s anything else bugging you.

I’m not sure about homs of co-dga’s which are free as graded-commutative algebras, but I can try thinking about it.

Posted by: Todd Trimble on December 3, 2007 10:22 PM | Permalink | Reply to this

Re: On BV Quantization, Part VIII

No, I’m not claiming Cof(Y,Z)Cof(Y,Z) is the desired internal hom (nor that it is equivalent to the desired hom).

Oh, good.

The internal hom is a subcomonoid of Cof(Y,Z)Cof(Y,Z)

All right, now I see where I got astray when following your exposition

Sorry if that was unclear.

My fault. Thanks for your patience.

Let me know if there’s anything else bugging you.

What’s bugging me personally now is that I cannot yet see what the internal hom thus constructed looks like concretely.

Here is a very simple example which I should eventually try to work out:

let the ambient category CC be chain maps internal to plain old finite dimensional real or complex vector spaces.

Then for (g,[,])(g, [\cdot,\cdot]) any Lie algebra, we obtain an object L g L_g in Coc(C)Coc(C) as follows:

as a graded cocommutative coalgebra L gL_g is the free graded cocommutative coalgebra generated from g[1]g[1], i.e. from the vector space underlying the Lie algebra gg, regarded as being in degree 1.

The degree -1 codifferential on the coalgebra acts on second tensor powers as the Lie bracket D:xy[x,y] D : x \vee y \mapsto [x,y] and is extended as a coderivation to the rest.

(This coalgebra L gL_g is an ordinary Lie algebra, regarded as an L L_\infty-algebra.)

Now let gg and hh be any two Lie algebras. I would like to have an explicit description of the internal hom in Coc between L gL_g and L hL_h, i.e. of the codifferential coalgebra Hom Coc(L g,L h). \mathrm{Hom}_{Coc}(L_g,L_h) \,.

If and when I understand this, I next want to understand the corresponding question for the next more nontrivial example: if g (2)g_{(2)} is an infinitesimal crossed module (crossed module of Lie algebras), there is accordingly a cocommutative coalgebra L g (2)L_{g_{(2)}} associated with it. For g (2)g_{(2)} and h (2)h_{(2)} arbitrary such crossed modules, I would like to see explicitly what the codifferential coalgebra Hom Coc(L g (2),L h (2)) Hom_{Coc}(L_{g_{(2)}},L_{h_{(2)}}) is.

After I understand this, I suppose it should be easy to see the more general cases:

for g g_{\infty} any L L_\infty-algebra, there is a corresponding codifferential coalgebra L g L_{g_\infty}. For g g_\infty and h h_\infty any two L L_\infty-algebras, I would like to know what the internal hom Hom Coc(L g ,L h ) Hom_{Coc}(L_{g_\infty}, L_{h_\infty}) is explicitly.

For the L L_\infty-algebras the corresponding coalgebras will be freely generated (as coalgebras, not as codifferential algebras) from \mathbb{N}-graded vector spaces in positive degree.

The next interesting step is to see what happens when generators in degree 0 appear.

So then, next, I would want to understand the following: for (g,B)(g,B) and (g,B,)(g',B',) two Lie-Rinehart pairs, there should be codifferential coalgebras L (g,B)L_{(g,B)} and L (g,B)L_{(g',B')} corresponding to them. I want to know explicitly what the internal Hom Coc(L (g,B),L (g,B)) Hom_{Coc}(L_{(g,B)}, L_{(g',B')}) is.

When I understand all this, it should be easy to understand what happens in the above example when the gg in the Lie Rinehart pair (g,B)(g,B) is replaced by an arbitrary L L_\infty algebra.

I’ll think about all this. But progress will be slow. I have a guest this week (Igor Baković is visiting) and a couple of other duties and obligations. But this is most interesting. I am so glad you told me about this internal hom in Coc.

Posted by: Urs Schreiber on December 4, 2007 11:58 AM | Permalink | Reply to this

Re: On BV Quantization, Part VIII

I have a spare minute.

For the entertainment of everybody who wants to watch it, I start, very modestly, to work out the answer to the exercise I stated above, which says:

work out what Todd’s general abstract description of the internal Hom in cocommutative comonads internal to chain complexes (aka codifferential co-commutative coalgebras) amounts to explicitly in special cases.

So, my ambient category is that of chain complexes of (finite dimensional, say, real, say) vector spaces.

Let YY and ZZ be any two cocommutative coalgebras internal to that category, and let f,g:YZ f,g : Y \to Z be two coalgebra homomorphisms (in the following, to be precise, to be regarded as their “names” f,g:IZ Yf,g : I \to Z^Y).

The first thing I want to figure out is what the two morphisms that Todd named Φ:Cof(Z Y)Cof((ZZ) Y)\Phi : Cof(Z^Y) \to Cof((Z \otimes Z)^Y) and Ψ:Cof(Z Y)Cof((ZZ) Y) \Psi : Cof(Z^Y) \to Cof((Z \otimes Z)^Y) do to these.

I think we have Φ:fg((fg)δ Y:Yδ YYYfgZZ) \Phi : f \vee g \mapsto ( (f\otimes g) \circ \delta_Y : Y \stackrel{\delta_Y}{\to} Y \otimes Y \stackrel{f \otimes g}{\to} Z \otimes Z ) and Ψ:f(δ Zf:YfZδ ZZZ). \Psi : f \mapsto (\delta_Z \circ f : Y \stackrel{f}{\to} Z \stackrel{\delta_Z}{\to} Z \otimes Z ) \,.

Here I am writing fg=fg+(1) |g||f|gff \vee g = f \otimes g + (-1)^{|g|\cdot |f|} g \otimes f for the symmetrized tensor product appearing in Cof(Z Y)Cof(Z^Y). I am also beginning to make use of the fact that a morphism of cofree coalgebras Cof(A)Cof(B)Cof(A) \to Cof(B) is specified by its value Cof(A)B Cof(A) \to B “on cogenerators”.

(This statement, and its meaning, are best remembered by thinking of dualizing everything in sight, such that it says that a morphism Fr(B *)Fr(A *)Fr(B^*) \to Fr(A^*) of free algebras is specified by its value B *Fr(A *) B^* \to Fr(A^*) on generators in B *B^*of Fr(B *)Fr(B^*).)

Okay, now I am hoping this explicit description of the nature of Ψ\Psi and Φ\Phi in this case will help me figure out their equalizer…

Posted by: Urs Schreiber on December 4, 2007 7:30 PM | Permalink | Reply to this

Re: On BV Quantization, Part VIII

While walking back home yesterday, I thought more about this.

Apart from a moderately thorny issue of prefactors, the situation seems to be this:

Given a homogeneous element fCof(Z Y)f \in Cof(Z^Y), hitting it with Ψ\Psi produces δ Zf. \delta_Z \circ f \,. On the other hand, hitting fff \vee f with Φ\Phi produces (ff)δ Y. (f \otimes f)\circ \delta_Y \,. Hence precisely when ff has the property that it is a morphism of comonoids, i.e. that (ff)δ Y=δ Zf (f \otimes f)\circ \delta_Y = \delta_Z \circ f do both terms equal each other.

Using this logic, it seems that we can build an element in Cof(Z Y)Cof(Z^Y) which equalizes Ψ\Psi and Φ\Phi by writing e(f):=f+c 1ff+c 2ffff+ e(f) := f + c_1 \, f \vee f + c_2 \; f \vee f \vee f \vee f + \cdots with ff a comonoid morphism and with the constant prefactors c nc_n determined recursively by requiring Ψ(e(f))=Φ(e(f)). \Psi(e(f)) = \Phi(e(f)) \,.

I need to think about this more carefully, but this looks good to me.

Then I need to figure out how to prove that the elements of the form e(f)e(f) exhaust the collection of equalizing elements completely, as it seems it should be.

I have a general question about equalizers in Coc: Todd emphasized that for suitably well behaved ambient categories CC, Coc has all equalizers. Can we say anything about what they look like explcitily?

Suppose I have found the equalizer as (in my case here) a vector space. How do I systematically find the coproduct and the codifferential on it? Probably this is simply induced from the coproduct and codifferential on, in our case, Cof(Z Y)Cof(Z^Y).

That would probably amount then to something like δ Hom Coc(Y,Z):e(f)e(f)e(f) \delta_{Hom_{Coc}(Y,Z)} : e(f) \mapsto e(f) \otimes e(f) if the above is right and unless my educated guessing is not educated enough.

If this analysis is roughly right so far I’d be rather satisfied, because this is the kind of result I am expecting to see.

All comments are very welcome. Unfortunately I need to take care of something else now…

Posted by: Urs Schreiber on December 5, 2007 10:37 AM | Permalink | Reply to this

Re: On BV Quantization, Part VIII

In case it helps, let me try to summarize what the point of my long proof was supposed to be.

We define the internal hom in CocCoc as an certain equalizer in CocCoc:

Hom Coc(Y,Z)Cofree(Z Y)Cofree((ZZ) Y)Cofree(I Y).Hom_{Coc}(Y, Z) \hookrightarrow Cofree(Z^Y) \stackrel{\to}{\to} Cofree((Z \otimes Z)^Y) \otimes Cofree(I^Y).

To see this works, consider a comonoid map XHom Coc(Y,Z)X \to Hom_{Coc}(Y, Z). By the equalizer property, this amounts to a comonoid map f˜:XCofree(Z Y)\tilde{f}: X \to Cofree(Z^Y) which equalizes that parallel pair coming out of Cofree(Z Y)Cofree(Z^Y). By a series of manipulations which was the main part of the proof, f˜\tilde{f} equalizes that pair iff the corresponding map in CC,

f:XYZ,f: X \otimes Y \to Z,

preserves the comultiplication and counit maps (i.e., is a morphism of comonoids).

Posted by: Todd Trimble on December 3, 2007 10:41 PM | Permalink | Reply to this

Re: On BV Quantization, Part VIII

I should have phrased that differently. On rereading what I just wrote I get the impression that I seemed to have followed your reasoning all right. But if the result is really that Hom Coc(Y,Z)Cof(Z Y)Hom_{Coc}(Y,Z) \simeq Cof(Z^Y) then this is rather unexpected to me.

The problem might actually be this: what I really want to understand are the internal Homs for “CoqDGCA”s, namely CoDGAs which are free as graded-commutative algebras (not, though, necessarily, as differential algebras).

Presumeably for those the discussion is rather different?

Posted by: urs on December 3, 2007 9:28 PM | Permalink | Reply to this

Re: On BV Quantization, Part VIII

One stupid question, just to make sure:

the counit π X:Cof(X)X\pi_X : Cof(X) \to X of the adjunction UCofU \dashv Cof

simply projects Cof(X)=IXCof(X) = I \oplus X \oplus \cdots onto the second summand.

Right?

Posted by: Urs Schreiber on December 4, 2007 12:11 PM | Permalink | Reply to this

Re: On BV Quantization, Part VIII

Something like this, yes; that’s the right idea. It would be a good idea to take a hard look at the explicit formula for Cof(X)Cof(X), the cofree cocommutative coalgebra construction, so as to fill in that ellipsis “…”; as far as I can tell, the construction is a little bit tricky for reasons given here.

Posted by: Todd Trimble on December 4, 2007 1:11 PM | Permalink | Reply to this

more about Hom

Todd’s comment
The last part means that if A and B are two commutative dgas, there is a cocommutative comonoid Hom(A,B) in Ch

reminded me to emphasize the importance of
the algebra structure (convolution product) on Hom(C,B) where
B is a dga and
C is a dg coalg

Is the categorical version of this written?

Posted by: jim stasheff on December 3, 2007 2:10 PM | Permalink | Reply to this

Re: more about Hom

Absolutely right. The monoid multiplication

Hom(C,B)Hom(C,B)Hom(C,B)Hom(C, B) \otimes Hom(C, B) \to Hom(C, B)

takes ‘ff\otimesgg’ (where ‘ff’: IHom(C,B)I \to Hom(C, B) is the name of f:CBf: C \to B) to (the name of)

CδCCfgBBμBC \stackrel{\delta}{\to} C \otimes C \stackrel{f \otimes g}{\to} B \otimes B \stackrel{\mu}{\to} B

and this ‘elementwise’ construction internalizes so as to work generally in any symmetric monoidal closed category, in which CC is a cocommutative comonoid and BB is a commutative monoid. This remark is implicit in my rough description of the enrichment of DGA in CoDGA discussed here.

Posted by: Todd Trimble on December 3, 2007 7:14 PM | Permalink | Reply to this

Re: more about Hom

It also works without comm or cocomm
at least for algebras
so I doubt it’s worse for monoidal cats
?even without symm?

the only place I know of where cocomm is relevant is for Hom(C,L) where L is Lie

Posted by: jim stasheff on December 3, 2007 7:50 PM | Permalink | Reply to this

Re: more about Hom

Right again – I threw in ‘(co)commutative’ because we happened to be talking about that, but it works more generally as you say.

But to internalize the construction, you do need a symmetry or braiding for the monoidal structure, as far as I can tell. The desired map

B CB CB CB^C \otimes B^C \to B^C

is, by \otimes-Hom adjunction, the transform of a composite map

B CB CC11δB CB CCC1σ1B CCB CCevalevalBBμBB^C \otimes B^C \otimes C \stackrel{1 \otimes 1 \otimes \delta}{\to} B^C \otimes B^C \otimes C \otimes C \stackrel{1 \otimes \sigma \otimes 1}{\to} B^C \otimes C \otimes B^C \otimes C \stackrel{eval \otimes eval}{\to} B \otimes B \stackrel{\mu}{\to} B

where you need to commute CC past B CB^C to get the definitions to work out. (Similarly, to define even nonpermutative operads internally in a monoidal category, one needs some braiding isomorphisms.)

Actually, this ‘principle’

  • Hom(cothing,thing)Hom(co-thing, thing) is a thing

is a pretty general phenomenon which may be worth talking about some time.

Posted by: Todd Trimble on December 3, 2007 8:17 PM | Permalink | Reply to this

Re: On BV Quantization, Part VIII

and remind me what is in AKSZ beyond BV except for extension of the ‘manifold’ language?

To me, the main point is this:

AKSZ show that one powerful way to solve the BV construction, i.e. to find the right BV complex together with an action that satisfies the master equation, is to look at quantum theories that explicitly arise as sigma-models of maps from par par to tar tar and then to form the configuration space conf=hom(par,tar) conf = hom(par,tar) of fields as the hom internal to arbitrary graded things.

This strikes me as an important insight. All fundamental quantum mechanical systems are of sigma-model type. (And this does include all gauge theories, of course, as you emphasized in 1987 with Bonora, Cotta-Ramusino and Rinaldi: a gauge theory with gauge \infty-group GG is simply a σ\sigma-model with target the one-object \infty-groupoid tar=BG. tar = \mathbf{B}G \,. In particular (super)gravity belongs to this class, where G=ISO^(n,m)G = \hat ISO(n,m) is some extension of the Poincaré-group (for instance to a Lie 3-group by using the sugra 4-cocycle).

Posted by: Urs Schreiber on November 30, 2007 1:35 PM | Permalink | Reply to this

Re: On BV Quantization, Part VIII

Urs wrote:

All fundamental quantum mechanical systems are of sigma-model type. (And this does include all gauge theories, of course, as you emphasized in 1987 with Bonora, Cotta-Ramusino and Rinaldi: a gauge theory with gauge \infty-group GG is simply a σ\sigma-model with target the one-object \infty-groupoid

tar=BG tar= \mathbf{B} G

I have a question about this that keeps bugging me — maybe you know the answer by now.

I know that, at least when GG is a compact Lie group, the universal GG-bundle over BGB G is equipped with a connection, so a map f:XBGf: X \to B G gives a principal GG-bundle over XX with connection. And, I vaguely recall that someone showed we can get any connection this way, as long as the dimension of XX is finite.

But, if I have a principal GG-bundle over XX with connection, it doesn’t seem to give a unique map f:XBGf: X \to B G.

So, to what extent is the groupoid of principal GG-bundles with connection over XX ‘the same as’ the space hom(X,BG)hom(X, B G)?

If I’d known a really satisfactory answer to this, I might never have developed that ‘principal GG-bundles with connection over XX are smooth anafunctors PXGP X \to G’ baloney.

Posted by: John Baez on December 1, 2007 9:14 PM | Permalink | Reply to this

Re: On BV Quantization, Part VIII

First of all, the bijection is between
equivalence classes of G-bundles over X
and
homotopy classes of maps [X, BG].

The appropriate notion of equivalence of connections on equivalent bundles is??
and the universal bundle with connection
is proved to exist
but I’m not sure what it looks like,
(netiher of the proofs - Quillen or Narasimhan and…
do I find pleasing)
nor how much further it is well behaved

Posted by: jim stasheff on December 2, 2007 12:46 AM | Permalink | Reply to this

Re: On BV Quantization, Part VIII

If I’d known a really satisfactory answer to this, I might never have developed that ‘principal GG-bundles with connection over XX are smooth anafunctors PXGP X \to G’ baloney.

I got the vague impression from what some people mumbled that using something like smooth homotopy classes of maps XBGX \to B G, we can classify GG-bundles with connection. But don’t take my word for it. I don’t know.

And, worse, I don’t care. :-) Because my impression is that even if it is possible, it is not the way to go.

Instead, when I made my statement that gauge theories are σ\sigma-models with target BG\mathbf{B}G I did mean it in the groupoid picture!

The problem is that you all convinced me not to write ΣG\Sigma G for the 1-object groupoid obtained from GG, but BG\mathbf{B}G. And, sure enough, already the second time I do so it leads to misunderstandings: you think I am talking about the huge space BGB G, where I am really talking about the nice and small groupoid {g|gG}\{\bullet \stackrel{g}{\to} \bullet | g \in G\}.

I think:

we should conceive QFTs of sigma-model kind as being about morphisms of suitably smooth \infty-groupoids.

Under certain circumstances it may be worthwhile to take geometric realizations everywhere and send {g|gG}\{\bullet \stackrel{g}{\to} \bullet | g \in G\} to the space BGB G. But under many other circumstances it may not. This happens usually when a connection enters the game.

Because, as the name suggests: where bundles give rise to cohomology classes, bundles with connection give rise to differential cohomology classes, and all things differential will be hard to see once we divide out continuous homotopies.

And of course, for the BV exercise which we are talking about here, it is rather backwards to start by thinking about maps to the space BGB G. Rather, we want to think of morphisms of Lie groupoids to the Lie groupoid {g|gG}\{\bullet \stackrel{g}{\to} \bullet | g \in G\} and then hit that with the Lie-functor which sends it to the world of L L_\infty-algebras etc.

Posted by: urs on December 3, 2007 5:33 PM | Permalink | Reply to this

Re: On BV Quantization, Part VIII

Although I’m happy to live in the L_\infty (drived??) world, for a large class of smooth spaces, there is a smooth approx thm that says docnitnuously homtopic = smoothly so.

Posted by: jim stasheff on December 24, 2007 1:02 PM | Permalink | Reply to this

Re: On BV Quantization, Part VIII

John wrote:

I know that, at least when G is a compact Lie group, the universal G-bundle over BG is equipped with a connection, so a map f:X→BG gives a principal G-bundle over X with connection. And, I vaguely recall that someone showed we can get any connection this way, as long as the dimension of X is finite.

This sounds related to Mark Mostow’s work on smooth spaces in

MR0587553 (84e:57030)
Mostow, Mark A.
The differentiable space structures of Milnor classifying spaces, simplicial complexes, and geometric realizations.
J. Differential Geom. 14 (1979), no. 2, 255–293.

I won’t get a chance to look at that paper for a couple of weeks, but I thought I’d mention it in case it rings a bell for you.

Posted by: Dan Christensen on December 22, 2007 3:17 AM | Permalink | Reply to this

Re: On BV Quantization, Part VIII

Thanks, Dan
It’s another of those papers I probably enjoyed when it appeared but had forgotten.
I will not have access to it til late January so if anyone can enlighten me further…

Posted by: jim stasheff on December 22, 2007 1:54 PM | Permalink | Reply to this

Re: On BV Quantization, Part VIII

Dan wrote:

I won’t get a chance to look at that paper for a couple of weeks, but I thought I’d mention it in case it rings a bell for you.

Thanks! Getting a little time to read blogs during the holidays?

As you know, this paper defines a concept of ‘smooth space’ in terms of a sheaf of real-valued functions on that space, which are declared to be smooth. It shows that if GG is a Lie group (or maybe any smooth group), the usual simplicial model for BGB G becomes a smooth space, with the universal bundle EGBGE G \to B G having a connection on it. This lets you describe real characteristic classes as closed differential forms on BGB G, built from the curvature of this connection.

But, I don’t recall him showing that any connection on any (sufficiently nice) GG-bundle is a pullback of this standard one. I may just have missed it…

… but for now, if I ever get time to dig into this issue, I’m gonna try to find those papers by Quillen and Narasimhan that Jim Stasheff alluded to.

I should, however, reread that paper by Mostow, just because it’s so nice.

Posted by: John Baez on December 23, 2007 11:47 PM | Permalink | Reply to this

Re: On BV Quantization, Part VIII

Urs wrote:

I got the vague impression from what some people mumbled that using something like smooth homotopy classes of maps XBGX \to B G, we can classify GG-bundles with connection. But don’t take my word for it. I don’t know.

It’s been a long time since we had this conversation, but just for the record, let me add:

This ain’t true, at least not in any sense I enjoy. But maybe I should be precise, since Jim wondered what I mean between bundles with connection. Here’s what I mean. Given two principal GG-bundles PXP \to X and P XP^' \to X with connections AA, resp. A A^', I say they’re isomorphic if there’s a GG-bundle isomorphism f:PP f: P \to P^', lifting the identity map on XX, such that:

f *A =A f^* A^' = A

This really says our GG-bundles with connection are the same in every practical sense.

This equivalence relation on principal GG-bundles with connection is bound to be much less coarse than ‘coming from smoothly homotopic maps to BGB G’. In fact, if the smooth structure on BGB G is nice, it’s quite likely that smooth maps from a manifold into BGB G are smoothly homotopic iff they’re homotopic. If so, two principal GG-bundles with connection come from smoothly homotopic maps to BGB G iff their underlying principal GG-bundles are isomorphic! The connection becomes completely irrelevant!

And, worse, I don’t care. :-)

Well, then it’s okay that you don’t know. :-)

I guess the problem is that for many years, I wanted to think of a principal GG bundle with connection over XX as a map from XX to something.

By now, I know that a principal GG bundle with connection over XX is the same as a map from PXP X to something. Here PXP X is the smooth path groupoid of XX.

Theorem: isomorphism classes of principal GG-bundles with connection over XX are in 1-1 correspondence with isomorphism classes of smooth anafunctors from PXP X to GG.

This is great and probably just what we need. A map from PXP X to GG really is a kind of map from XX to something.

But, I still keep wondering if can delete the phrase ‘path groupoid of’, basically by using an adjoint functor:

hom(PX,G)hom(X,BG)??? hom(P X , G) \cong hom(X , B G) ???

Here PP is the ‘path groupoid’ functor going from smooth spaces to smooth groupoids, and BB is my hoped-for, but apparently nonexistent, adjoint functor going back from smooth groupoids to smooth spaces. On the left ‘hom’ means ‘smooth anafunctors’, while on the right it means ‘smooth maps’. Or something like that.

Note: no taking homotopy classes anywhere!

I’ve sort of given up on this idea, but it still haunts me…

Posted by: John Baez on December 24, 2007 12:35 AM | Permalink | Reply to this

Re: On BV Quantization, Part VIII

You might be on to something - hope some bright young thing will succeed. I’m willing to believe a connection on a specific bundle may change things fro the homotopy point of view. After all, it gives a unique path lifted morphism PX to G, not just the infty version without using a connection. But notice - BG is well defined up to homotopy.
Which example would make things work sans homotopy??

Posted by: jim stasheff on December 24, 2007 1:08 PM | Permalink | Reply to this

Re: On BV Quantization, Part VIII

On the issue of universal connections and classifying bundles with connections by maps of spaces, I have a reply, which however I decided to post to another thread. Please see this.

Posted by: Urs Schreiber on December 27, 2007 1:42 PM | Permalink | Reply to this
Read the post The Principle of General Tovariance
Weblog: The n-Category Café
Excerpt: Landsmann proposes that physical laws should be formulated such that they may be internalized into any topos.
Tracked: December 5, 2007 5:11 PM
Read the post Lie oo-Connections and their Application to String- and Chern-Simons n-Transport
Weblog: The n-Category Café
Excerpt: A discussion of connections for general L-infinity algebras and their application to String- and Chern-Simons n-transport.
Tracked: December 25, 2007 7:37 PM
Read the post Comparative Smootheology
Weblog: The n-Category Café
Excerpt: A survey and comparison of the various notions of generalized smooth spaces, by Andrew Stacey.
Tracked: January 4, 2008 10:33 AM
Read the post Frobenius algebras and the BV formalism
Weblog: The n-Category Café
Excerpt: Bruce Bartlett is looking at the latest article by Cattaneo and Mnev on BV-quantization of Chern-Simons theory.
Tracked: November 14, 2008 1:27 PM

Post a New Comment