Lorentz violating severely restricted: Mqg/Mplank > 1200

[lumidek]

Lubos,

Thanks for your response.

As far as I understand, the limits found in that paper are specially worrisome for the n=1 (linear) models. On what grounds do you claim that the whole programme of LQG has been falsified by the Fermi observations? See the question #6 by Ashtekar's FAQ paper arxiv:0705.2222: "Will Lorentz invariance be violated in the low energy limit of LQG dynamics?". Please, if possible, state your counter-arguments according to the exposition presented in that paper by Ashtekar.

I would like to invert the question the other way around. Would it be correct to affirm that *if* Lorentz violations were observed, string theory would be promptly falsified?

Thanks.
Christine

Two sentences I didn't address. Yes, if the spacetime were found to deviate from Lorentz symmetry by order-one terms at the Planck scale, string theory - as understood by real string theorists and taught by Polchinski or GSW or Becker or other textbooks would be instantly falsified.

Second point. Ashtekar arguments that he would love to have Lorentz invariance in LQG are nothing else than a wishful thinking, and all his detailed statements - especially those in between the lines - are just plain wrong. It is not true that the split of dimensions to 3+1 is the only or main feature that makes LQG violate Lorentz symmetry. It is not enough to be able to define generators on a Hilbert space if one wants the dynamical laws to be symmetric - because the former condition is kinematic and knows nothing about the dynamics, while LI invariance is a dynamical question.

Also, it is not true that one can actually define proper generators on the spin network Hilbert space. Also, it is not true that discrete area spectrum may be compatible with the Lorentz symmetry. If there is any formula for the areas that is a manifestly a sum of real discrete numbers, the theory automatically violates the Lorentz invariance - for example because areas in Lorentz-invariant theories can be both real and imaginary (spacelike vs timelike).

So all his verbal proclamations seem to be wrong and there's no calculation. So what should I do with that? It's just rubbish. The other papers at least try to calculate something, and of course, they end up with the only possible answer they can: LQG much like any other theory with a naive mechanistic discrete picture of space at the Planck scale violates the Lorentz symmetry. I am sure that you know very well that Ashtekar's paragraph is pure babbling and there exists not a single paper that would make a single calculation supporting the wishful thinking in the paragraph.

Best
Lubos

 

[atyy]

For the same reason, many other classes of theories have also been falsified, including causal dynamical triangulations, emergent condensed matter-like spacetimes, Horava-Lifgarbagez gravity, and many others.

Maybe not CDT - I believe CDT is more like computational asymptotic safety - CDT itself is not a complete theory - and asymptotic safety, although it may have other problems, surely respects Lorentz invariance?

 

[ccdantas]

Ashtekar arguments that he would love to have Lorentz invariance in LQG(...)

Hmm, I'm not here in defense of anyone, but I do not find evidences for that claim in his paper, only that

"In full non-perturbative quantum gravity there is no background metric whence some
care is needed to speak of Lorentz invariance. The question can only refer either to asymp-
totic symmetries in the asymptotically flat context or effective low energy descriptions. I
would expect LQG will have the first type of Lorentz invariance generated by global charges
corresponding to asymptotic symmetries. But unfortunately so far global issues related to
asymptotic flatness have received very little attention." p. 13.

Also, it is not true that one can actually define proper generators on the spin network Hilbert space.

Would you please elaborate on that?

So all his verbal proclamations seem to be wrong and there's no calculation. So what should I do with that? It's just rubbish.

The paper is a summary for a general audience at the 11th Marcel Grossmann meeting, so there are no detailed calculations, as expected, but I suppose some can be found in the list of references that he provides (see also his footnote #8). I would have to check on that, though.

 

[lumidek]

Maybe not CDT - I believe CDT is more like computational asymptotic safety - CDT itself is not a complete theory - and asymptotic safety, although it may have other problems, surely respects Lorentz invariance?

There is no asymptotically safe theory of gravity, because of technical RG reasons and because of wrong scaling for the entropy at high energies that should be dominated by black holes. And even if there were one, CDT couldn't be its approximation.

 

[lumidek]

Hmm, I'm not here in defense of anyone, but I do not find evidences for that claim in his paper, only that

"In full non-perturbative quantum gravity there is no background metric whence some
care is needed to speak of Lorentz invariance. The question can only refer either to asymp-
totic symmetries in the asymptotically flat context or effective low energy descriptions. I
would expect LQG will have the first type of Lorentz invariance generated by global charges
corresponding to asymptotic symmetries. But unfortunately so far global issues related to
asymptotic flatness have received very little attention." p. 13.



Would you please elaborate on that?



The paper is a summary for a general audience at the 11th Marcel Grossmann meeting, so there are no detailed calculations, as expected, but I suppose some can be found in the list of references that he provides (see also his footnote #8). I would have to check on that, though.

Dear Christine, sure, I will happily elaborate on that. One link is enough. The most cited loop quantum gravity paper in 2005

http://arxiv.org/abs/hep-th/0501114

shows that the algebra of constraints, including the Hamiltonian, doesn't close in loop quantum gravity (besides dozens of other lethal problems). So this particular Ashtekar's statement has been shown incorrect.

 

[atyy]

There is no asymptotically safe theory of gravity, because of technical RG reasons and because of wrong scaling for the entropy at high energies that should be dominated by black holes. And even if there were one, CDT couldn't be its approximation.

I understand the plausibility of the first two statements - but why can't CDT be an approximation to an asymptotically safe gravity, if such a thing existed?

 

[Finbar]

There is no asymptotically safe theory of gravity, because of technical RG reasons and because of wrong scaling for the entropy at high energies that should be dominated by black holes. And even if there were one, CDT couldn't be its approximation.

Hi Lubos,

Can you give some references to back your claims about asymptotic safety in gravity? What are these technical RG reasons?

Thanks.

 

[Hans de Vries]

Every single model marketed as loop quantum gravity, spinfoam, causal dynamical triangulation, Horava-Lifgarbagez gravity, and dozens of other names violates the Lorentz symmetry by first-order terms, with a coefficient of order one, and is simply safely dead after this paper.

Well, even without this paper, it seems very improbable that the photon
propagator could emerge from any theory with path/geometry randomness
at the Planck scale.

Being on the light-cone with such an extreme precision , what mechanism
could cancel out all the contributions from the random geometry paths
which are not on the large-scale lightcone?

Now, while agreeing with you, how would you explain that your favorite theory
doesn't exhibit the same problem? Why doesn't the photon propagator become
"fuzzy" with all these complicated geometry paths at the Planck scale?

The (not so well known) "photon self-propagator" which has the photon
field itself as a source, rather than the current j, does a wonderful job in
canceling out the contributions on all paths other than the light-cone path
(see sect 1.19 of: http://physics-quest.org/Book_Chapter_EM_basic.pdf )
but it needs a flat geometry at Planck's scale.Regards, Hans

 

[ccdantas]

http://arxiv.org/abs/hep-th/0501114

Yes, that paper is well known, perhaps as much as this one:

http://arxiv.org/abs/hep-th/0608210

There are highly nontrivial material on these papers, and I'd rather not comment further, but only link here for reference.

 

[ensabah6]

Now, while agreeing with you, how would you explain that your favorite theory
doesn't exhibit the same problem? Why doesn't the photon propagator become
"fuzzy" with all these complicated geomet

Regards, Hans

I was just going to ask Lubos this,

how do you know that the compatification of the additional 6 dimensions and the landscape and various mechanisms such as KKLT doesn't break lorentz invariance, or in some way affect the speed of light at the plank scale, or SUSY breaking mechanism?

 

[humanino]

Throw a list of reference and drown the fish. It is amusing that Lubos would repeatedly quote

http://arxiv.org/abs/gr-qc/0411101[/QUOTE]where it is explained (or wished) that the breaking is an artifact.

Just like with Pauli bashing Yang because Pauli "knew" that nonabelian gauge theories were "sick", very little discussion is possible against no-go theorems until loopholes are found. And just like Yang, LQG people are not blind but quite aware of those difficulties. Yes, LQG has difficulties, and is much less attractive than string theory, especially considering how much the latter is developed. By itself it does not justify calling people names.

I failed to find a published reference to http://arxiv.org/abs/gr-qc/0411101 answering negatively to the (putative ?) hopes conveyed there. If Lubos has such an obvious answer, he would contribute positively to public money saving by publishing a Letter instead of making short statements on a blog.

 

[lumidek]

I understand the plausibility of the first two statements - but why can't CDT be an approximation to an asymptotically safe gravity, if such a thing existed?

For example, because the asymptotically safe (and other) field theories have a unique vacuum while the vacuum in (Minkowskian) triangulated or otherwise discretized models of gravity is highly non-unique, creating an entropy density that goes to infinity in the continuum limit.

This situation differs from normal lattices for QFTs because the shape of the lattices is fixed and the fluctuations of the degrees of freedom living on the lattice sites are universal in the UV. For triangulations, there's no real "UV", the metric is dynamical, and one always sums about all kinds of stuff.

Second, the "causal" in the causal triangulations refers to an artificial truncation of the configurations to a subset that satisfies a "causal" global condition on the geometry. Such a truncation can never generate a field theory because almost all (in the measure sense) individual configurations that are summed in the path integral of any quantum theory are acausal. For example, a point-like electron is moving along trajectories that are superluminal almost everywhere, and causality is only restored when all these paths are summed over.

Truncating paths that are superluminal anywhere (i.e. almost all of them) would completely damage the short-time behavior (the power laws etc.) and it would really break the uncertainty principle because in the path integral formalism, the uncertainty principle is only possible because almost all trajectories contributing to the path integral are non-differentiable.

 

[lumidek]

Throw a list of reference and drown the fish. It is amusing that Lubos would repeatedly quote where it is explained (or wished) that the breaking is an artifact.

Just like with Pauli bashing Yang because Pauli "knew" that nonabelian gauge theories were "sick", very little discussion is possible against no-go theorems until loopholes are found. And just like Yang, LQG people are not blind but quite aware of those difficulties. Yes, LQG has difficulties, and is much less attractive than string theory, especially considering how much the latter is developed. By itself it does not justify calling people names.

I failed to find a published reference to http://arxiv.org/abs/gr-qc/0411101 answering negatively to the (putative ?) hopes conveyed there. If Lubos has such an obvious answer, he would contribute positively to public money saving by publishing a Letter instead of making short statements on a blog.

Could you please stop emitting this noise and lies? The paper, much like all others, calculates Lorentz violation in the dispersion relations. Open

http://arxiv.org/PS_cache/gr-qc/pdf/0411/0411101v1.pdf

Look at pages 8-13 where the calculation is hidden. The conclusion is at the end of the section on page 13 and the conclusion is that the result "does break Lorentz invariance".

There have been many papers showing that LQG is worthless crap in detail and they have even become the most cited LQG papers of the year (like Nicolai et al. 2005). But no one really cares about it because for every correct result and sensible scientist, it has become politically correct to promote one (or two) wrong result and fund one (or two) crackpots.

 

[lumidek]

Hi Lubos,

Can you give some references to back your claims about asymptotic safety in gravity? What are these technical RG reasons?

Thanks.

Dear Finbar, see, for example
http://golem.ph.utexas.edu/~distler/blog/archives/001585.html

 

[heinz]

Lubos,

is the following reasoning correct?

1) There are no deviations from special relativity's Lorentz symmetry in nature.
1b) All proposals which state such deviations are wrong.

2) There probably are no deviations from general relativity's diffeomorphism invariance and other symmetries.
(For essentially the same reasons that 1) is correct: continuity of space-time holds.)
2b) All proposals that state such deviations are wrong.

Heinz

 

[lumidek]

Lubos,

is the following reasoning correct?

1) There are no deviations from special relativity's Lorentz symmetry in nature.
1b) All proposals which state such deviations are wrong.

2) There probably are no deviations from general relativity's diffeomorphism invariance and other symmetries.
(For essentially the same reasons that 1) is correct: continuity of space-time holds.)
2b) All proposals that state such deviations are wrong.

Heinz

Dear Heinz, it is somewhat strange to call these propositions "reasoning" because they seem to be rather isolated propositions and don't follow from each other in any way (and are not proved in your comment). But all propositions you wrote are correct. ;-)

In the case of diffeomorphisms, it is even more obvious. Whenever the dynamical metric tensor is a part of the description, there must exist an exact diffeomorphism symmetry, otherwise the negative-normed (negative probabilities!) unphysical modes of the metric tensor wouldn't be decoupled, and probabilities could become negative.

With the diffeomorphism symmetry being a fact, the Lorentz symmetry is there, too (because it is a subgroup of diffeomorphisms, at least on the sectors with the right topology). The nontrivial part of the statement is that the flat vacuum is invariant under such a Lorentz group, or at least it is "locally" invariant so that the Lorentz group can only be at longer distance scales L by positive-energy structures/objects whose typical length is L.

These facts also imply that the spectrum of positions or lengths or areas, whenever it may be definable, must be continuous. In realistic theories of quantum gravity, however, the lengths or areas are good observables only in the long-distance approximating theories. The true calculable quantities are scattering amplitudes for particles with fixed momenta etc.

 

[humanino]

Could you please stop emitting this noise and lies? The paper, much like all others, calculates Lorentz violation in the dispersion relations. Open

Actually, I printed it the other day. As written in the abstract

Furthermore, by contrasting Hamiltonian and Lagrangian descriptions we show that possible Lorentz symmetry violations may be blurred as an artifact of the approximation scheme. Whether this is the case in a purely Hamiltonian analysis can be resolved by an improvement in the effective semiclassical analysis.

the paper questions whether it is possible that the violations are an artifact of the approximation scheme. It is also illustrated with kindergarden examples on the harmonic oscillator how a the discretization can yield such artifacts and miss nonperturbative terms in the corrections.

These examples have important hints for the calculation of corrected dispersion relations and the issue of Lorentz covariance. Since only higher order corrections will be seen when a Hamiltonian is perturbed, Lorentz violations are bound to appear as a consequence of this way of doing the calculation. Space and time derivatives of the classical fields have to be related in the Lagrangian in a way dictated by the symmetry. If those terms are torn apart, because one computes the Lagrangian from a perturbed Hamiltonian which only sees higher space derivatives but not higher time derivatives in its corrections, Lorentz invariance will be violated. This kind of violation of Lorentz symmetry is not a consequence of the theory but of the way to perform perturbative calculations.

It seems like you do not even expect your readers to check your references. Again, please note that the reason I am asking is because of the possibility of you saving my time by convincing me (and everybody) that reading LQG literature is a loss of time. This is your own claim, so I hope you can back it up.

 

[Finbar]

Dear Finbar, see, for example
http://golem.ph.utexas.edu/~distler/blog/archives/001585.html

I've read that sometime ago. But in light of this paper

http://arxiv.org/abs/0902.4630


"Yes, that would also serve as a good test"
Distler would have to admit the AS program past this test i.e. finding a fixed point when non-perturbativly renormalizable terms are included in the truncation.

Do you have any other references? If there are good RG reasons why AS cannot work it would be interesting to see if these could be formulated into some "no-go" theorems.

 

[MTd2]

Look at pages 8-13 where the calculation is hidden. The conclusion is at the end of the section on page 13 and the conclusion is that the result "does break Lorentz invariance".

Wrong. P. 12 "This kind of violation of Lorentz symmetry is not a consequence of the theory
but of the way to perform perturbative calculations."p.13 :"It should
however be kept in mind that the calculations done up to now (including the model of
the previous section) can only yield preliminary results and that a definite answer to the
question of Lorentz violation by loop quantum gravity definitely has to await a more complete
treatment, possibly along the lines sketched above."

So, there is no certainty. And by the time the calculations were done, it was a problem with the hamiltonian perturbative expansion.

 

[heinz]

... But all propositions you wrote are correct...

Lubos,

do I understand you correctly that if the symmetries of general relativity
are correct at all scales, as you stated, then also general relativity itself
is correct at all length and energy scales?

heinz

 

[Finbar]

Lubos,

do I understand you correctly that if the symmetries of general relativity
are correct at all scales, as you stated, then also general relativity itself
is correct at all length and energy scales?

heinz

what do you mean by general relativity? general covariance is probably correct at all scales(in string theory?) but not the dynamics of gravity i.e the Einstein equations.

Lubos,

Would I be correct to say that in perturbative string theory one allows strings to "live" on a background spacetime such that the physics is generally covariant. Nonetheless exact non-perurbative string theory is fully background independent?

 

[turbo]

It is interesting to see how such a small observational sample set can be touted as "proving" or "falsifying" anything about LQG, String, etc. Why not wait for more observational data and see what trends (if any) evidence themselves? The GLAST project was pushed back over and over - let's see what we get now that the probe is functional.

 

[Fra]

Give me a break with the arrogance. I am just alarmed that some people want to dilute this experimental result and its consequences on physics. But physics is all about direct and indirect comparisons of observations with theories. And this observation happens to be extremely clean and settles the question. It proves that people like me have always been right and people around loop quantum gravity have always been wrong, using their poor education, weak intelligence, and lacking intuition to study questions that go well beyond their abilities. The result proves that all sponsors and foundations who have funded theories building on the assumption that Lorentz symmetry will have to be broken have wasted the money, and as soon as they care about the empirical data, they should learn a lesson and fire all these people.

I will not allow anyone to create fog about this very clear situation.

I find this type of judging and aggresive reasoning which apparent lack to humbleness to also be more alarming. It doesn't contribute to a good creative atmosphere.

It seems to me (my personal impression from your writings) an overall quality of your reasoning here and elsewhere to often make, from your point of view, very probable inferences, appear as bulletproof and unquestionable deductions, thereby clearing the fog that may exists by those who doesn't make the same inferences as you. It seems you often suggest that anyone that doesn't see it has inferior intelligence and should step aside.

I get the impression that you you think everyone who does see that string theory is the only reasonable way, must by conclusion, have inferior intelligence? :)

/Fredrik

 

[lumidek]

Lubos,

do I understand you correctly that if the symmetries of general relativity
are correct at all scales, as you stated, then also general relativity itself
is correct at all length and energy scales?

heinz

Dear Heinz, I can only subscribe to Finbar's answer.

If by general relativity, you mean a theory with a metric tensor and diffeomorphism-invariant action, i.e. one composed out of polynomials of the Riemann tensor (and perhaps its non-polynomial i.e. nonlocal i.e. nonperturbative extensions; and from the gauge-theoretical field strength and other matter), then yes, the (effective) action has to have this form.

However, it's not true that the Einstein-Hilbert action "R" is the whole story. The higher-derivative terms, such as R^n, are really the rule and are included (and have to be included) with appropriate coefficients whose magnitude may be guessed from dimensional analysis.

Cheers, LM

 

[lumidek]

I get the impression that you you think everyone who does see that string theory is the only reasonable way, must by conclusion, have inferior intelligence? :)

/Fredrik

Fra, in this particular case, the right conclusion about the intelligence of the writer can be obtained without any measurement of anyone's knowledge of string theory: grammar is enough. ;-)

 

[lumidek]

It is interesting to see how such a small observational sample set can be touted as "proving" or "falsifying" anything about LQG, String, etc. Why not wait for more observational data and see what trends (if any) evidence themselves? The GLAST project was pushed back over and over - let's see what we get now that the probe is functional.

Dear Turbo, I think you are very confused. This whole thread is about the newest result of GLAST that was renamed to Fermi one year ago:

http://motls.blogspot.com/2008/08/glast-first-results.html

In some sense, this is the ultimate result of Fermi, the culmination of its ability to measure and decide things: the future measurements will be qualitatively less important because they will be essentially repeating what we can see in this paper.

Also, I want to emphasize that in science, one properly done observation is enough to falsify theories and whole frameworks, and we're just seeing a good example here.

Cheers, LM

 

[lumidek]

Wrong. P. 12 "This kind of violation of Lorentz symmetry is not a consequence of the theory
but of the way to perform perturbative calculations."p.13 :"It should
however be kept in mind that the calculations done up to now (including the model of
the previous section) can only yield preliminary results and that a definite answer to the
question of Lorentz violation by loop quantum gravity definitely has to await a more complete
treatment, possibly along the lines sketched above."

So, there is no certainty. And by the time the calculations were done, it was a problem with the hamiltonian perturbative expansion.

This comment of yours is ludicrous.

If a symmetry is violated even perturbatively, it is pretty much guaranteed that it is also violated nonperturbatively, unless there is a cancellation of perturbative and nonperturbative terms which would imply that the whole perturbative expansion is impossible - and in this case, it would also mean that it is impossible to define the theory from any classical starting point.

Perturbative expansions remain one of the main tools to gather the information about theories and your hostility towards this very method shows that you have no clue about physics. See also http://motls.blogspot.com/2009/08/why-perturbation-theory-remains.html

Cheers, LM

 

[lumidek]

I've read that sometime ago. But in light of this paper

http://arxiv.org/abs/0902.4630


"Yes, that would also serve as a good test"
Distler would have to admit the AS program past this test i.e. finding a fixed point when non-perturbativly renormalizable terms are included in the truncation.

Do you have any other references? If there are good RG reasons why AS cannot work it would be interesting to see if these could be formulated into some "no-go" theorems.

Dear Finbar, there are several other illuminating posts on Jacques' website, e.g.

http://golem.ph.utexas.edu/~distler/blog/archives/000648.html
http://golem.ph.utexas.edu/~distler/blog/archives/001585.html
http://golem.ph.utexas.edu/~distler/blog/archives/001609.html

I think that these insights are shared by virtually all the sane people who have thought about this issue but it's not being published by anyone because it's considered a part of the general lore. See e.g. page 4 of Polchinski's book where he explains why this guess about the UV fixed point is not pursued there.

The assumption is, of course, that the terms that behave nicely only behave nicely because they're either removable by field redefinitions, renormalizable, or topological, and the true difficult contractions of powers of the Riemann tensor, i.e. those arising from higher-loop divergences, would falsify the safety - and add infinitely many new parameters in the UV.

These are technical reasons and there may exist a simple proof that this doesn't work. But I personally have very different primary reasons to be sure that gravity can't be described by asymptotically safe UV theory - namely black hole thermodynaimics, holography etc. Field theory just doesn't reproduce the right high-center-of-mass spectrum (which should be dominated by black hole microstates). Also, the black hole information loss paradox requires some nonlocality for the information to get out of the hole, so a field theory with an exactly definable metric tensor and the corresponding causal structure can't be right.

 

[MTd2]

This comment of yours is ludicrous.

I merely quoted and summarized part of the conclusion. So, you mean that paper is ludicrous. So, why did you even bother coming up with that paper?

 

[turbo]

Dear Turbo, I think you are very confused. This whole thread is about the newest result of GLAST that was renamed to Fermi one year ago:

http://motls.blogspot.com/2008/08/glast-first-results.html

In some sense, this is the ultimate result of Fermi, the culmination of its ability to measure and decide things: the future measurements will be qualitatively less important because they will be essentially repeating what we can see in this paper.

Also, I want to emphasize that in science, one properly done observation is enough to falsify theories and whole frameworks, and we're just seeing a good example here.

Cheers, LM

I am not confused. I know that the probe/instrumentation was renamed in honor of Fermi. I also remember that Fotini Markopoulou suggested that the highest-energy gamma rays might be slowed (energy-dependent time dispersion) by interacting with the fine-scale structure of the vacuum, AND suggested that with a large enough spread in energies over very long distances, Glast might be able to detect such dispersion. Let's see what happens when more bursts are analyzed. Collecting a few photons at a time and analyzing their energies and arrival times is not a trivial exercise and much can rest on the way that the data are analyzed.

As to the bolded text: you are assuming that all future observations will be similar (in contrast to the possible dispersion found by the MAGIC consortium). That may or may not be true, and it is in bad taste (IMO) to trash the careers of others who are keeping an open mind about this subject. Rarely do we get the opportunity to test cosmological theories with direct observation. Fermi may allow us to do just that, and we should make many observations and look for trends in the data.