Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Freidel does it: BeeF, it's what's for dinner!

  1. Jul 7, 2006 #1

    marcus

    User Avatar
    Science Advisor
    Gold Member
    2015 Award
    Dearly Missed

    Freidel does it: "BeeF, it's what's for dinner!"

    This looks like the Quantum Gravity+Matter breakthrough I've been expecting for a few years. The paper's arxiv number is easy to remember----just think of July 14 as Bastille Day and write this year's Quatorze Juillet holiday as 0607014

    http://arxiv.org/gr-qc/0607014

    It is a paper by Laurent Freidel and two other gentlemen---a Pole and Russian, I believe---named Jerzy Kowalski-Glikman and Artem Starodubtsev.

    Jerzy K-G is at Wroclaw University, Artem S is in Renate Loll's group at Utrecht, and Laurent Freidel, as I guess everybody knows, is jointly at Perimeter and the Lyon Ecole Normale.

    The apparent breakthrough is not a sure thing. All I have to go on right now (besides some second-hand information) is a sense of confidence I get from reading the paper. It is an order of magnitude more on top of things than what I've met in the past.

    Let's explicate the paper: this thread can be devoted to doing that---and also to examining the followups. there are two or more other Freidel et al papers expected to follow this one.
     
    Last edited: Jul 8, 2006
  2. jcsd
  3. Jul 7, 2006 #2

    marcus

    User Avatar
    Science Advisor
    Gold Member
    2015 Award
    Dearly Missed

    the paper is based on some unfamiliar things-----BeeF, the deSitter group SO(4,1) and it's algebra so(4,1)
    We encountered these things in earlier papers (some by Smolin, Starodubtsev, Freidel, and extensively by John Baez in the case of BF theory) but for me the first prolonged exposure was from the January 2005 paper of Freidel and Starodubtsev.
    http://arxiv.org/abs/hep-th/0501191

    The story goes back to MacDowell and Mansouri (1977) who discovered that classical General Relativity can be written as a perturbed BF theory where the A and B differential forms are valued in so(4,1).

    So we jump into this new Freidel paper at the bottom of page 3 where equation (2.4) gives a Gen Rel action S expressed as a perturbed BeeF action.
    If you set alpha and beta equal zero, then the perturbing terms drop out and you get a pure BeeF action.

    I guess one point to make is that it is known how to quantize a BF theory by the spinfoam method. So if they can get a comprehensive theory of gravity and matter in the form of a (perturbed) BF theory then
    1. there will be things they can find out (like what is the FLAT limit, setting alpha equal zero, is it DSR? and
    2. it is a known proceedure to quantize it.

    I could use some help transcribing a few of the equations like (2.4) in Latex. have to go out, will do some more on this in an hour or so when I get back.
     
    Last edited: Jul 7, 2006
  4. Jul 7, 2006 #3

    Kea

    User Avatar

    [tex]S = \int ( B_{IJ} \wedge F^{IJ} - \frac{\alpha}{4} B_{IJ} \wedge B_{KL} \varepsilon^{IJKL}
    - \frac{\beta}{2} B^{IJ} \wedge B_{IJ} )[/tex]
     
  5. Jul 7, 2006 #4

    marcus

    User Avatar
    Science Advisor
    Gold Member
    2015 Award
    Dearly Missed

    thanks Kea!
    I just got back and was girding for Latex.

    so this alpha is in a certain way a disguised version of the newton G. when they want to have a FLAT or a "zero gravity" version of the theory they can turn gravity off by setting alpha = 0

    and if they set both alpha and beta equal zero then they get
    THE MOST VANILLA OF ALL THEORIES, the generic jack-of-all-trades action which John Baez teaches us to associate with spinfoam, namely

    [tex]S = \int ( B_{IJ} \wedge F^{IJ} )[/tex]
     
    Last edited: Jul 7, 2006
  6. Jul 7, 2006 #5

    marcus

    User Avatar
    Science Advisor
    Gold Member
    2015 Award
    Dearly Missed

    we should look at equation (2.12) and get some hint about the physical significance of alpha. It has the immirzi number in it, but that is just some order-one number like 0.2 or 0.3, so it is not what matters.

    [tex]\alpha = \frac{G\Lambda}{3}\frac{1}{(1+\text{immirzi}^2)}[/tex]

    So for example if immirzi = 0.2, then 1+ immirzi2 = 1.04
    and the second fraction [tex]\frac{1}{(1+\text{immirzi}^2)}[/tex]
    is approximately equal to one and we don't worry about it.

    The main term is newton's G times the cosmological constant Lambda.

    Alpha is going to be made zero either in the case we want to turn gravity off, or we want to have zero cosmological constant (no acceleration in the expansion of the universe.)

    What is nice though. What I really like about this, is that this "most vanilla of all theories" this harmless nondescript BF is MANIFESTLY DIFFEOMORPHISM INVARIANT and manifestly does not require specifying any BACKGROUND GEOMETRY to start with.

    Perturbation approaches (which are the standard proceedure for almost everything) almost always seem to have an explicit requirement of some definite (usually flat) background geometry. For example perturbative string theory. For example quantum field theories. They all make you start with flat Minkowski space, or the moral equivalent of that, and perturb around it. You get to just slightly wrinkle it.

    But here we see them start with BF theory, which looks like a lot blanker slate, and perturb around THAT.
    So it HAS the diffeoinvariant explict independence of background geometry that Quantum Gravitists customarily insist on.
    ============
    there is still a long ways to go, but they sure are making a good start.
     
    Last edited: Jul 7, 2006
  7. Jul 7, 2006 #6

    marcus

    User Avatar
    Science Advisor
    Gold Member
    2015 Award
    Dearly Missed

    I will put a 4 in Kea's superscript. this is where they break SO(4,1) symmetry down to SO(3,1). It is part of the 1977 routine of MacDowell and Mansouri, who used SO(4,1) even though they always intended to do 4D Gen Rel.

    the notation is explained in Freidel Starodubtsev January 2005 paper.
    http://arxiv.org/abs/hep-th/0501191

    ===========================
    another thing we need, for understanding the paper, and maybe someone will supply is
    a handle on what could be the EIGENVALUES OF THE QUADRATIC AND QUARTIC CASIMIR OPERATORS of the so(4,1) algebra.

    this is jumping to page 8. The two eigenvalues they call C2 and C4.

    it will turn out that C2, the eigenvalue of the quadratic Casimir operator, helps determine the MASS of the particle. And C4, or rather the ratio C4/C2, determines the SPIN of the particle.

    this is going to page 9, and looking at equations 3.13 and 3.15

    the cosmological constant Lambda also enters into determining the mass.
     
    Last edited: Jul 8, 2006
  8. Jul 7, 2006 #7

    selfAdjoint

    User Avatar
    Staff Emeritus
    Gold Member
    Dearly Missed


    Marcus you are aware of course, that you can have a BF theory basd on just about any Lie Group. If G is the group and [itex]\mathfrac{g}[/itex] is its Lie Algebra, then F is the curvature of G and B is a [itex]\mathfrac{g}[/itex] valued 2-form. the seemingly magical properties here are not due so much to the BF as to the de Sitter group, which gives the all-important cosmo constant.
     
  9. Jul 7, 2006 #8

    Kea

    User Avatar

    Exactly. And the dependence of mass on [itex]\Lambda[/itex] follows from the magical parity cube. So we are beginning to look at a framework in which the gauge groups of the SM (and gravity) are derived.
     
    Last edited: Jul 8, 2006
  10. Jul 8, 2006 #9

    marcus

    User Avatar
    Science Advisor
    Gold Member
    2015 Award
    Dearly Missed

    the Casimirs C2 and C4 are important in this paper, so here is some review
    Wiki on the Lie algebra symmetric bilinear form B(x,y)
    http://en.wikipedia.org/wiki/Killing_form
    you get it using Trace on the the compose of two ad(.) maps and B(. , .) behaves nicely with bracket.
    =============
    the Casimir operator or Casimir invariant is defined using this bilinear form
    http://planetmath.org/encyclopedia/CasimirOperator.html
    one can pick any basis xi of the Lie algebra and then, using B(. , .), construct a DUAL basis xj
    with B(xi, xj) = 1 if i=j and zero otherwise.
    Intuitively, this is THE NEXT BEST THING TO AN ORTHONORMAL basis, which you can't get because the symmetric bilinear form on the Lie algebra is not positive definite (not an "inner product"). So instead of making an orthornormal basis, you construct two baseez (two basisses)

    And then one forms the Sum over i of these basis and dual basis thingees multiplied together formally in the "universal enveloping algebra" and summing
    Sumi xi:cool: xi
    Multiplying things in the "universal enveloping algebra" is rather like tensoring----you do a formal multiplication and then factor out to validate the Lie bracket.

    Wiki has an alternative definition, which may need clarification, but also an example and some discussion.
    http://en.wikipedia.org/wiki/Casimir_operator
    =================
    Does anyone have a good online source about Casimirs?
    =================
    the rank of the Lie algebra is the number of independent Casimirs. Suppose we fix a representation. Then since a Casimir commutes with everything it is some number times the identity. So it is also a NUMBER namely its unique eigenvalue.
    ==================
    Physicists are familiar with the Casimir of the rotations algebra so(3), as the squared angular momentum operator.
    If we decide on some representation---corresponding to a particle---then this operator becomes a NUMBER, as indicated above, and that number is the spin of the particle.
    ==================
    maybe this will provoke someone more expert to provide a GOOD tutorial about Casimirs:smile: and some GOOD online resources. But anyway I am doing a slapdash treatment because we need it right now.
    ===================
    WHO CAN TELL US ABOUT THE TWO CASIMIRS OF so(4,1), the deSitter algebra?

    apparently it is rank 2, and it has two Casimirs, which are the quadratic and the quartic. And if you pick a definite representation---as if one were specifying a particle---then one gets MASS AND SPIN numbers from the two Casimirs. Aiiyeee!
    ==================
    maybe a nice online tutorial on so(4,1) and SO(4,1) will show up.
    this smilie is intended to denote hopefulness:tongue2:
     
    Last edited: Jul 8, 2006
  11. Jul 8, 2006 #10

    marcus

    User Avatar
    Science Advisor
    Gold Member
    2015 Award
    Dearly Missed

    Yes! I agree that BF can be done with any of the usual groups and what makes this special is SO(4,1). I would like to know more about the deSitter group SO(4,1) and its algebra so(4,1). Just about anything you or Kea can say would be a help---either to me or to other readers of the thread.

    that sounds quite promising Kea. Please elaborate, if you feel so inclined. It is a puzzle why the Standard Model gauge should be a mix of SU(3) x SU(2) x U(1).
     
    Last edited: Jul 8, 2006
  12. Jul 8, 2006 #11
    Ah, the old idea to put the universe in a box (and put in a IR cutoff) - and then eventually take the infinite volume limit (Lambda --> 0). But the m^2 is merely just the p_ap^a no, the mass due the cosmological constant and spin being negligible. I don't know what you mean with the gauge groups of the standard model being derived here :uhh: That the
    SO(4,1) gives also spin (which contributes to the mass) and particle momentum hardly comes as a surprise: we knew that already for the Poincare group in 4 dimensions (which has exactly the same dimensionality) and this has been extensively studied in Einstein Cartan theory. Let's wait and see whether such split can cure the IR divergences in the quantum theory - there is no evidence presented for that at all !
     
    Last edited: Jul 8, 2006
  13. Jul 8, 2006 #12

    Kea

    User Avatar

    Careful

    You appear to be jumping to the most dreadful conclusions about what we are talking about. Please read a little more of the literature before making ill-informed remarks.
     
  14. Jul 8, 2006 #13
    Dear Kea,

    If you add a cosmological constant, this serves as a volume regulator for the spacelike universe, and is crucial for renormalizability even for simple toy models such as 1+1 quantum gravity in the CDT approach (here is your IR cutoff ). Having finite volume on a compact spatial universe (de Sitter), is like putting in reflecting mirrors and it effectively boils down to ``putting the universe in a box'' - any child knows this. Frankly speaking, I was asking questions and not drawing any conclusions. You made some claim of seeing the standard model arise, so where is it ???

    Now, in the thread about stroop theory, I was wondering why I should take this categorical relationship between nCob and Hilb seriously in any way. One of the reasons offered there was that making punctures destroys it all. Of course Hurkyl simply told that in TQFT this is forbidden, but quantum gravity is of course no TQFT. In going over to 3+1 gravity you will have to put on gravitational waves, that is you need quantized versions of the Riemann, Weyl, ricci scalar ... on the Hilbert space of (particle) states of this BF theory. The question is of course whether these operators will be bounded in this representation (and that the full states are normalizable !), and I see no good reason why this should be so. It is the same with the radiation problem in QFT, you start out with the coulomb states and then try to treat the radiation backreaction problem by adding photon per photon, it turns out that you need to put in a UV cutoff by hand to achieve this. Now people easily say, ohw but all operators in quantum gravity have bounded spectrum, and even the length, area and volume operators have discrete spectrum. But is this really so? None of these operators have any physical significance yet since the Hamiltonian constraint hasn't been implemented (and doing so might drastically change the spectral properties of the operators), and known work shows that the curvature operators on the FULL Hilbertspace of spin networks aren't bounded at all.

    So, it might turn out that punctures need to happen... or that you still have to put in some regulator by hand - and why to take all this trouble then ?

    I think these are all legitimate *questions* concerning those issues you believe in (I see this just as the next fad); it would be better to offer some comments/explanations than to guess about how well informed I am (on basis of a sentence where you seem to express your own ignorance).

    Careful
     
    Last edited: Jul 8, 2006
  15. Jul 8, 2006 #14

    Kea

    User Avatar

    Careful

    Tell us something we don't know...and do some reading.

    Kea :grumpy: (this is the FIRST time I have ever used grumpy)
     
  16. Jul 8, 2006 #15

    arivero

    User Avatar
    Gold Member


    Amusingly, the Casimirs were also important in our scheme to formalise Hans' model of electroweak parameters,
    http://arxiv.org/abs/hep-ph/0606171
    Section 1 of this article was sent for peer review, and the referee did the following final coment:

    "As a final comment, I note that there is nothing about Poincare invariance of quantum
    field theory that uniquely chooses a gauge group and the corresponding gauge symmetry
    breaking dynamics. I can write down an infinite number of models with different gauge
    groups, different Higgs sectors and different spectra of gauge bosons and matter fields.
    So without any further theoretical principle, the Casimir operators of the Poincare group
    have nothing to say about the mass spectra of particles. The authors propose no additional
    theoretical principle. Consequently, this paper offers nothing beyond a numerological co-
    incidence."

    I'd like to add that if we are into LQG or any other theory with a Planck Length, we are not interested anymore on plain Poincare but also in deformations of it. It has been an underground lore that a deformation of SU(2)xSU(2) could be the adequate substitute of Lorentz group.
     
    Last edited: Jul 8, 2006
  17. Jul 8, 2006 #16
    Kea,

    I am asking some legitimate questions about this approach (which you simply refuse to answer). As I remember some claims were made that category theory offers a deeper understanding between QM and GR, so it is my good right to ask where it is (since I don't see it). I do not claim to be a specialist in this kind of approach (far from), you have that position. Therefore it would be only normal that you answer my questions briefly, telling someone to read specific literature is not the way to sell your idea's (and it usually indicates that the ideas are pretty poor). If it is a good one, ``you should be able to explain it to the barmaid''. I am myself reading lots of different stuff (concerning local realism) and I really do not have the time to check all these latest fads into the details, that is why I ask (and if people ask me about the things I read, then I always offer the idea/explanation) since this one seems to excite people.

    The reason why I am asking these similar questions over and over again when I hear these cries is because I believe a deeper failure in our theories to be responsible for the problems of QG while such lines of thought merely reflect a technical issue. Therefore, I would welcome any insight which shows me otherwise ....

    Careful
     
    Last edited: Jul 8, 2006
  18. Jul 8, 2006 #17
    **I'd like to add that if we are into LQG or any other theory with a Planck Length, we are not interested anymore on plain Poincare but also in deformations of it. It has been an underground lore that a deformation of SU(2)xSU(2) could be the adequate substitute of Lorentz group. **

    Ok, never really looked into this but I have some (naive) questions :
    (a) are these deformations, like doubly special relativity, merely not a fancy way to sneak in a(n ``observer independent'') cutoff ?
    (b) the modified dispersion relations are ``non-quadratic'' so that the Minkowski metric becomes a cumbersome notion with respect to the inertial observers and only appears to be a good low energy approximation (in an observer independent way). So, how to ``geometrize'' such theory when there is no obvious geometric object present ?
    Finsler geometry perhaps ?

    Careful
     
  19. Jul 8, 2006 #18

    selfAdjoint

    User Avatar
    Staff Emeritus
    Gold Member
    Dearly Missed

    The reason no-one is discussing your questions is because they are OFF-TOPIC. This thread is devoted to the RESULT, not to generalizing or big new fad theories or any such thing. As for Einstein-Cartan theory, did you or anyone else derive from it this RESULT connecting the de Sitter group to the Poincare group and GR, VIA WILSON LINES THAT ACT LIKE WORLD LINES OF PARTICLES? This is as far as I know new, and nothing in your posts on this thread suggests that you are prepared to refute it.
    Just showing off your knowledge and being snarky isn't contributing.
     
  20. Jul 8, 2006 #19

    arivero

    User Avatar
    Gold Member

    Perpahs, but it is not the purpose. The idea is to introduce representations of SU(2)_q with q a root of the unity and then relate them to particles and fields. It is lore, not published papers on it.
     
  21. Jul 8, 2006 #20
    Ah doesn't matter (I am not the person to judge ideas to the phys rev qualifications it has). I also am not opposed to Lorentz symmetry being a low energy approximation (as is clear from many posts). I simply wonder whether there could not be an easier way to do all of this (without conflicting experiment), also one could wonder about the meaning of causality (and especially the high energy content of the ``no signalling faster than light theorem'' (I take c fixed here obviously)) in such context. Concerning (b), the Finsler suggestion might not be that bad, it reminds me at unification attempts between GR and EM which inevitably lead to violations of (usual) relativistic causality. Also, such type of geometry could lead to many appearantly non-local effects in QM.

    Careful
     
Know someone interested in this topic? Share this thread via Reddit, Google+, Twitter, or Facebook

Have something to add?



Similar Discussions: Freidel does it: BeeF, it's what's for dinner!
Loading...