Monday, October 10, 2005

The Others

I sneaked into the Loops '05 Conference at the AEI at Potsdam. So, I will be able to give you live blogging for today and tomorrow. After some remarks by Nicolai and Thiemann and the usual impedence mismatch between laptops and projectors, Carlo Rovelli has started the first talk. He is on slide 2, and still reviews recent and not so recent devellopments of LQG.

9:55
Rovelli talked about his paper on the graviton propagator. If you like he wants to recover Newton's law from his model. The obvious problem of course is that any propagator g(x,y) cannot depend on x or y if everything is diffeomorphism invariant (at least in these people's logic). So he had to include also a dependence on a box around the 'detector' and introduce the metric on the box as boundary values. He seems to get out of this problem by in fact using a relational notion as you would of course have to in any interesting background independent theory (measure not with respect to coordinates but with respect to physical rulers). Then there was a technical part which I didn't quite get and in the end he had something like g(x,y)=1/|x-y|^2 on his slide. This could be interesting. I will download the paper and read it on the train.

11:15
Next is Smolin. Again computer problems, this time causing an unscheduled coffee break. Smolin started out talking about problems of bckground independent approaches including unification and the nature of anomalies. Then, however, he decided to focus on another one: How does macroscopic causality arise? He doesn't really know, but looked at some simple models where macro causality is usually destroyed my some non-local edges (like in a small world network). Surprisingly, he claims, these non-local connection do not change macroscopic physics (critical behaviour) a lot and thus they are not really detectable.

Even more, these non-local "defects" could, according to Smolin, play the role of matter. Then he showed another model where instead of a spin network, the physics is in twisted braided ribbon graphs. There, he called some configurations "quarks" and asigned the usual quantum numbers and ribbon transformations for C, P and T. Then it got even better, next slides mentioned the problem of small power in low l modes in the CMB ("scales larger than 1/Lambda"), the Poineer anomly and the Tully Fisher relation that is the empirical observation behind MOND. I have no idea what his theory as to do with all these fancy open probelms. Stefan Theissen next to me makes interesting noises of astonishment.

12:15
Next speaker is John Barrett. This talk sounds very solid. He presents a 3+0 dimensional model which to me looks much like a variant of a spin network (a graph with spin labels and certain weight factors for vertices, links, and tetrahedra). He can do Feynman graph like calculations in this model. Further plus: A native speaker of British English.

13:00
Last speaker of the forenoon is Stefan Theissen. He tries to explain how gravity arises from string theory to the LQG crowd. Many have left before he started and so far he has only presented string theory as one could have done this already 20 years ago: Einstein's equation as consistency requirement for the sigma model and scattering amplitudes producing the vertices of the Einstein Hilbert action. Solid but not really exciting.

14:55
In the afternoon, there are parallel sessions. I chose the "seminar room". Here, Markopoulou presents her idesa that dynamics in some (quantum gravity?) theory has formal similiarities to quantum information processing. In some Ising type model she looks at the block spin transformation and reformulates the fact that low energy fields only talk to the block spins and not to the high frequency fields. With some fancy mathematical machinery, she relates this to error correction where the high frequency fields play the role of noise.

15:25
Next is Olaf Dreyer. Very strange. He proposes that quantum mechnics should be deterministic and non-linear. Most of what he says are philosophical statements (and I do by far not agree with all of them) but what seems to be at the core of it is that he does not want macroscopic states that are superpositions of elementary states. I thought that was solved by decoherence long ago...

15:35
At least Rovelli asks "[long pause] maybe I didn't understand it. you make very general statements. But where is the physics?"

15:56
The next speaker is Wang who expands a bit on what Smolin said in the morning. It's really about Small World Networks (TM). If you have such a network with gauge flux along the edges then in fact a non-local random link looks locally as a charged particle. This is just like in Wheeler's geometrodynamics. The bulk of the talk is about the Ising model on a lattice with a small number of additional random links. The upshot is that the critical temperature and the heat capacity as well as the correltations at criticality do not much depend on the existence of the additional random links.

16:25
Martinetti reminds us that time evolution might have a connection with temperature. Concretely, he wants to take the Tomito-Takesaki unitary evolution as time evolution and build a KMS-state out of it. There is a version of the Unruh effect in the language of KMS states and Martinetti works out the corretion to the Unruh temperature from the fact that the observer might have a finite life time. This correction turns out to be so small that by uncertainty, one would have to measure longer than the life time to detect the difference in tempaerture.

18:15
I stopped reporting on the afternoon talks as I did not get much out of those. Currently, Rüdiger Vaas, a science journalist, is the last speaker of the day. He at least admits that his talk is on philosophy rather than physics. His topic are the philosophical foundations of big bang physics.

5 comments:

Anonymous said...

it's wonderful having your live-blog report, Robert. thank you. On Tuesday, I am especially interested in the morning talk of Laurent Freidel. Of the afternoon parallel sessions, I would go to session I.
(the one with Dan Christensen) if I could be there.

Whatever talks you attend it will be great to hear your report! I am looking forward to your comment on the morning session with Robbert Dijkgraaf.

thanks again for being there for the rest of us.

Anonymous said...

Thank you for your frank and undiluted comments. Also with the breaks in your reports caused by too technical parts in a talk, accoustically only half-understood questions, or ill chosen topics in some talks it feels a little like really being there. Please keep going.
F.

Anonymous said...

Hi Robert,

thanks for the blogreport!

You wrote:

"Smolin started out talking about [...] the nature of anomalies."


Can you say more about what precisely he said about anomalies?


Best,
Urs

Robert said...

Urs,

this was really only a side remark. Remember, in the discussion about the LQG string, the Virasoro algebra, and its central charge, some people from the loopy camp made the statement that they could quantize any theory with symmetry using their group averaging procedure.

Then Jacques Distler asked what would happen to theories with chiral fermions where the triangle anomaly is not canceled. At the time, there was no good answer to this challenge.

Now, Smolin mentioned two possible "solutions": First, the chiral anomaly as we compute it is found perturbatively (it's a one loop effect and it was shown not to be corrected at higher loop orders). But the loopy approach is non-perturbative. Thus it might be that it does not see this artefact of pertubation theory.

This response of course is very weak, as the chiral anomaly seems to be an extremely robust property of IR physics and you can as well geometrically express it in terms of a mismatch of bundles. So it is highly unlikely that it is just a perturbative artefact. In string theory we know that anomaly cancelation often comes about by many different ingerdients conspiring in just the right way and in fact often tell us about non-perturbative physics. One example that comes to my mind is M-Theory on G2 manifolds with singularities where anomaly cancelation tells us about the non-perturbative existence of non-abelian gauge groups (which of course have to be there to make contact with the IIA situation with intersecting D6 branes).

The second possible solution offered by Smolin is that on spin networks it is like on the lattice and you have a doubling of chiral fermions. This I am much more willing to believe however that would be really bad for the loop approach as we all know that this would mean that effectively they cannot have chiral fermions which would make this whole approach unfit for the real world.

Let me add that there is of course a possibility three that somewhere in the details their formalism (at least when describing anomalous theories) is inconsistent when you work out the details.

Urs said...

Robert wrote:

"The second possible solution offered by Smolin is that on spin networks it is like on the lattice and you have a doubling of chiral fermions."

I cannot see how this could be true. This seems to be identifying the lattice-like choice of spin-network basis for the Hilbert space with an actual physical lattice. But there is no latticized spacetime in LQG. Just a popular choice of basis which makes people think of lattices, as far as I can see.