A new paper by Steve Giddings, “Universal Quantum Mechanics,” arxiv:0711.0757. Here’s the abstract:
If gravity respects quantum mechanics, it is important to identify the essential postulates of a quantum framework capable of incorporating gravitational phenomena. Such a construct likely requires elimination or modification of some of the “standard” postulates of quantum mechanics, in particular those involving time and measurement. This paper proposes a framework that appears sufficiently general to incorporate some expected features of quantum gravity. These include the statement that space and time may only emerge approximately and relationally. One perspective on such a framework is as a sort of generalization of the S-matrix approach to dynamics. Within this framework, more dynamical structure is required to fully specify a theory; this structure is expected to lack some of the elements of local quantum field theory. Some aspects of this structure are discussed, both in the context of scattering of perturbations about a flat background, and in the context of cosmology.
Part of the problem in reconciling gravity with quantum mechanics is “technical” — GR is not renormalizable, by the lights of ordinary quantum field theory. But part is “conceptual” — ordinary QM takes a spacetime background as given, not as part of the wavefunction. The role of time, in particular, is a bit hazy, especially because the Wheeler-deWitt equation (the quantum-gravity version of the Schrodinger equation) doesn’t contain any explicit time parameter. Most likely, our notion of “time” makes sense only in a semi-classical context, not as part of the fundamental dynamics. Similarly, our notions of “locality” are going to have to be broadened if spacetime itself is part of the quantum picture. But the truth is that we don’t really know for sure. So it’s worth digging into the underlying principles of quantum mechanics to understand which of them rely crucially on our standard understanding of spacetime, and which are likely to survive in any sensible theory of quantum gravity. Giddings’s paper follows in the footsteps of previous work such as Jim Hartle’s Spacetime Quantum Mechanics and the Quantum Mechanics of Spacetime. (The Santa Barbara air must be conducive to thinking such deep thoughts.)
I was so excited by this abstract, after hearing Gidding’s excellent talks at GRG18, but I simply couldn’t get past Postulate 1.
Oops, sorry about the misplaced apostrophe.
Serendipitously, i just finished reading some Laszlo and wanted some science to go along with it…thanks for the link.
Hi Sean, I advise that you look at the paper below, with a much more promising-looking theory than the one you cite.
Gravity, Gauge Theories and Geometric Algebra Anthony Lasenby, Chris Doran, Stephen Gull
A new gauge theory of gravity is presented. The theory is constructed in a flat background spacetime and employs gauge fields to ensure that all relations between physical quantities are independent of the positions and orientations of the matter fields. In this manner all properties of the background spacetime are removed from physics, and what remains are a set of `intrinsic’ relations between physical fields. The properties of the gravitational gauge fields are derived from both classical and quantum viewpoints. Field equations are then derived from an action principle, and consistency with the minimal coupling procedure selects an action that is unique up to the possible inclusion of a cosmological constant. This in turn singles out a unique form of spin-torsion interaction. A new method for solving the field equations is outlined and applied to the case of a time-dependent, spherically-symmetric perfect fluid. A gauge is found which reduces the physics to a set of essentially Newtonian equations. These equations are then applied to the study of cosmology, and to the formation and properties of black holes. The existence of global solutions enables one to discuss the properties of field lines inside the horizon due to a point charge held outside it. The Dirac equation is studied in a black hole background and provides a quick derivation of the Hawking temperature.
I haven’t read Giddings paper yet, but I’m skeptical of what seems to be an approach that requires quantum theory to accommodate gravitation (rather than the other way around).
As you correctly point out, the Wheeler-DeWitt equation does not require an explicit spacetime background, which would seem to weaken an approach that assumes that such a background exists quantum mechanically and thereby poses problems for measurement.
[One could make the case the WD equation is a quantum theory of gravity in the same sense that the Schroedinger equation is a quantum theory of kangaroos. It’s just that in both cases the equations are far, far too difficult to solve.]
Well, it’s not just that the Wheeler-deWitt equation is too difficult to solve. It’s also that we don’t know what the right Hamiltonian is, so we don’t even know what equation we are solving. Thus string theory or whatever.
fascinating, and the subject meshes perfectly with all the Smolin in my current reading list. a much appreciated post, I hope for a strong discussion.
tyler: “the subject meshes perfectly with all the Smolin in my current reading list”
Giddings is a well known string theorist. The paper hasn’t got a single LQG reference so I don’t see how Smolin is relevant here.
Actually, Giddings and Smolin are colleagues and not everything Giddings does is purely string theoretic.
The latest issue of Gnu Scientist (2007-11-17 #2630) has quite a complimentary two page article on Giddings’ work:
http://www.newscientist.com/channel/fundamentals/mg19626303.900-is-mathematical-pattern-the-theory-of-everything.html
Oops, my apologies. When checking the link I realized the article was about a different guy entirely!
A bit off topic, but what do you think of Garret Lisi’s work
We can’t incorporate quantum mechanics into gravity (or vice versa!) very well unless we have any kinks ironed out of gravitational theory itself. Consider the following issue, which I just posted to sci.physics etc. for quick input:
Gravitomagnetic field does not obey superposition, what are the implications?
Moving streams of matter should produce a “Gravitomagnetic field” analogous to the magnetic field from moving charges. This is orthodox GR, even if some details are in contention. (See for example, http://en.wikipedia.org/wiki/Gravitomagnetism.)
I played around with equivalent matter streams moving in opposite directions. That should cancel out their gravitomagnetic fields B* (usually just called “B” after context given), but I found an inconsistency. The problem was, superposition of B* from two sources as if a vector field did not work. (Sure, gravity is more complicated, but that part should approximate vector fields and emulate EM at low mass levels – ? – and I expect even non-linear superpositions to cancel out opposite fields.) For example, let’s have adjacent streams going 0.5c in opposite directions, very low mass density to provide high expected linearity (albeit at relativistic speeds.) The proper value of g seen in our frame K is gamma squared times the value g_s at rest relative to either single stream (hence g = 2gamma^2(0.5c)g_s = (2*4/3)g_s = (8/3)g_s), due to the multiplied effects of Lorentz contraction and greater relativistic mass-energy density (thus field-producing power) per proper length unit in the mass flow.
We will send a unit mass M moving at 0.5c, frame K’, along the streams’ path, in either direction. We, expecting to see only “g” since the B* has ostensibly been canceled, expect M to experience gamma squared the value of proper acceleration towards the stream that we find in K (lateral acceleration transformation, from shorter proper time to fall.) Since effective mass-energy of M is gamma*M_0 (one of the cases where “relativistic mass” is still relevant), that is equivalent to a force in K increased to gamma times rest value and thus in K’, gamma squared times the force in K (due to force transformation.) Hence by that consistent-seeming reckoning, the acceleration of M in K’ should be gamma^2(0.5c)g = 2gamma^4(0.5c)g_s = (32/9)g_s.
However, in M’s frame, one stream is at rest and the other one goes at 0.8c. The combined effect is therefore [1 + gamma^2(0.8c)]g_s = (34/9)g_s =(17/16)*(32/9)g_s. It is easy to verify (using the additive gamma factor being gamma(v1 + v2) = gamma1*gamma2*(1 + v1*v2/c^2)) that the ratio in general of the second prediction to the first is (1 + v^4/c^4).
That is an odd contradiction, and I just don’t know what to make of it. Sure, gravitation is not like EM and with curved space etc., but would anyone expect low-gravity fields of any kind not to cancel out if apparently of opposite sign? Has anyone found good rules for B*? Has this issue been talked about before, and where? Thanks.
…It is easy to verify (using the additive gamma factor being gamma(v1 + v2) = gamma1*gamma2*(1 + v1*v2/c^2)) that the ratio in general of the second prediction to the first is (1 + v^4/c^4).
I mean, relativistic velocity addition for gamma(v1 + v2), but don’t have a good symbol to use (the oft-used dot on top of plus is real hard to find in fonts.
“A bit off topic, but what do you think of Garret Lisi’s work”
Complete crackpotism! The dude doesn’t even know the difference between a ghost field and a fermion, not to mention the other bizarre statements.
Neil (#13)
At the top of the Wikipedia article you cite, it notes that:
If you want a formula describing the weak-field effect of relativistic transverse velocity, starting with the Schwarzschild solution it’s pretty easy to show that a test particle with radial velocity zero and transverse velocity v is seen to accelerate in a frame stationary wrt the source mass at:
a = -M (1+2v^2)/r^2
(in units where G=1, c=1) and in the test particle’s frame this becomes:
a’ = -M (1+2v^2)/[(1-v^2)r^2]
to take account of time dilation.
Note that these are not what you’d get by substituting M/sqrt(1-v^2) in place of M.
Greg:
Thanks, I will think about that. I find your formula odd, since consider the accelerating elevator in the equivalence principle (a along y.) Just like a bullet fired along x from a gun, the moving mass must hit the floor at the same time as the dropped test body. I can’t see how the resulting transverse acceleration of the bullet in our elevator frame K’ could be anything other than *the same* in simple terms as for the dropped body. The transformation to the bullet frame is given by SRT, by time dilation as you note (but applied to your lab frame formula, not mine here.) I suspect there is an issue with how red shift (time standards in effect) affect the definition of rate of acceleration, or that when you refer to “a mass”, the field around the concentrated mass involves enough curvature to not be equivalent to the parallel field of the elevator? Remember that I was considering “streams” of matter, and we could use moving sheets as well – not point-like masses.
In any case, since you said, “…is seen to accelerate in a frame stationary wrt the source mass…” I presume you are not referring to gravitomagnetism proper (where we use a frame where both sources and test bodies are moving.) My main interest here is what happens when there are two streams that should (?) cancel out each other’s gravitomagnetic effect. So do you still think that would happen (or should it?) What are the formulas for the most general case, still with weak masses at long separations but arbitrary speeds? Most important, do the fields from different sources obey superposition (in the basic sense, not counting slight adjustments due to curvature etc.)
That’s not true, he does know about these very elementary concepts. He just has an unconventional point of view.
I guess that ‘t Hooft would have been called a crackpot too if he had published his ideas about deterministic models without first becoming famous for his work on the Standard Model.
OK, now I am rather sure Greg is considering the curved space around a concentrated mass – this is the same as the deflection of light around a body being twice the Newtonian result (as per the famous eclipse experiments, etc.) But even then, why is it “a = -M (1+2v^2)/r^2” instead of a = -M (1+v^2)/r^2, since the former makes for three times the standard acceleration when v –> c instead of twice (but I may be using a wrong intuitive assessment.) Either way, I am supposing that we don’t travel around a ST dimple, but along parallel field lines as along mass currents, or better; sliding planes. Then I can hardly imagine that the simple EP does not apply, and my paradox seems to be real then.
Iblis: Can you or someone explain “ghost field” in a good way? I heard of ghost particles, but it seemed like a clunky and unreal conceit.
Neil, in QM the superposition principle applies to the state vectors (probability amplitudes), not to the fields as the field equations may be nonlinear. Even in electromagnetism, if you take into account the effect of virtual electron loops, you get nonlinear corrections.
You can do without ghost fields, it is just a mathematical trick. What happens is that you want to do the computations using the gauge fields instead of physical fields. Since the same physical field configuration corresponds to many different guage field configurations, a path integral over the gauge fields would not lead to the correct result. It is possible to fix this problem by introducing ghost fields.
“Giddings is a well known string theorist. The paper hasn’t got a single LQG reference so I don’t see how Smolin is relevant here.”
From the abstract:
“These include the statement that space and time may only emerge approximately and relationally.”
This is a primary concern of Smolin’s; from my reading it appears that his conviction that spacetime emerges relationally predates the development of LQG and was one of his main motivations for that effort.
from Sean’s post:
“GR is not renormalizable, by the lights of ordinary quantum field theory. But part is “conceptual” — ordinary QM takes a spacetime background as given, not as part of the wavefunction.”
the non-renormalizability of GR, the (as he calls it) “background dependence” of QM and string theory, and how to reconcile all of these things in a theory of quantum gravity, are also fundamental topics of his writing, as should be fairly obvious to anyone who has read him.
If you think LCQ in and of itself is the only topic on which Lee Smolin is relevant you are pretty far afield from reality. I have yet to read his most recent book, which I know has irked a lot of people…it’s next on my list though…but I’m not saying “he’s right, you all are wrong about XX,” I hold no such opinion.
But at least thematically it’s certainly relevant, and mostly I was just interested in the synchronicity of the post with my reading of the moment.
Neil
The equivalence principle implies that all kinds of matter with the same initial position and velocity will free-fall side by side. It does not imply that a relativistic bullet racing past a falling stone will hit the ground at the same time as the stone — not even if the experiment is being conducted above an infinite sheet of mass, removing tidal effects.
The effect of transverse velocity on acceleration is related to the fact that geodesics at different velocities slice through spacetime differently and “see” different curvature.
The formula I gave for an instantaneous acceleration when the velocity is purely transverse is not going to give you an immediate comparison between Newtonian and relativistic formulas for the deflection of light, which comes from an integral over an entire trajectory.
The version of linearised gravity that is directly analogous to electromagnetism involves sources and test particles moving at non-relativistic speeds. In that approximation, a test particle mid-way between two parallel counter-flowing mass streams ought to experience an acceleration (of zero) that is independent of its velocity to first order.
I don’t know what happens if you go to the next order and use fully relativistic weak-field linearised gravity; it’s conceivable to me that a moving test particle could swerve away from the midline with an acceleration that’s second order in v, but I’m really not sure. In any case, a proper treatment will not lead to anything paradoxical.
Linearised gravity is a subtle business; even deriving Newtonian gravity from GR (as Wald notes in section 4.4 of his General Relativity), involves a careful use of terms of different order. The bottom line is that if you come up with a thought experiment that gives paradoxical results, you’re using the wrong approach. Certainly, thought experiments that mix special relativity with the idea that gravity resembles electromagnetism will be useless.
Greg, I am surprised at your take on the equivalence principle. I was always rather sure, as I gathered, that a “uniform” (well, have to adjust to the Rindler coordinates of g = -c^2/X) parallel gravitational field was indeed just like a vast accelerating elevator. Isn’t that the equivalent change to the space time? That means equal fall time for transverse bullet and dropped test body. Maybe you’re right, but isn’t it so that the particular transverse acceleration you got is for the radial field around a small compact mass, and not necessarily for the field from a sheet? (“geodesics at different velocities slice through spacetime differently and “see” different curvature.” – isn’t “curvature” proper only around dimples in the space-time, not sheets? Rem. that a cylinder for example isn’t “really” curved according to the Gaussian definition IIUC.) Have you seen or calculated that per se? Just wondering.
Finally, I am wary of the sentiment that paradoxical results must mean a wrong approach. It may be wrong in the real world, but we shouldn’t assume that our theories won’t yield such paradoxes.
Neil
The specific formula I gave was for the weak-field approximation to the Schwarzschild solution, so it only applies to a spherically symmetrical source, but there certainly would be velocity effects (at some order in v) even for an infinite planar mass.
If spacetime isn’t curved, there is no gravity. If you want to use the notion of an acceleration due to a “gravitational field”, rather than spacetime curvature, then you are limited to either a purely Newtonian approach, or the first-order approximation of gravitomagnetism. Neither of those models are valid for relativistic test particles.
If you are serious about wanting to know what happens to relativistic test particles in the vicinity of linear or planar mass flows, you need to learn about the weak-field approximation. This is linear in the perturbation to the metric, so you can add the fields from different sources, but what you are adding is not an acceleration, it’s a tensor field. You can then compute geodesics (or the proper accelerations along non-geodesic world lines, which is sometimes easier) for the metric you get from superimposing all the perturbations from all the sources. Any good textbook (e.g. Wald, Misner Thorne & Wheeler, and no doubt that of our host Sean Carroll — though I’m too ancient to have read that one myself) will get you started.