Talk:Aether drag hypothesis/Archive 1

Archive 1

Wet telescopes

• The Fizeau experiment supposedly shows that a particulate medium (in this case, water) does drag light along with it.

• The Airey experiment supposedly shows that a particulate medium (in this case, water) does not drag light along with it.

Funny, that.

Did they buy their water from different shops?

ErkDemon 00:59, 17 July 2005 (UTC)

No, in the Fizeau experiment water is pumped through the apparatus at high velocity, in the Airy experiment a telescope is filled with water. The telescope does not drag the light, the fast flowing water produces an effect that could be thought to be due to drag but is actually a relativistic effect (see reference at end of article and Special relativity for beginners). Fresnel's work is like that of Newton - almost right for all the wrong reasons. loxley 17:53, 17 July 2005 (UTC)

Both experiments use nominally-moving water, in the Fizeau experiment I think I've read somewhere that the speed in the lab frame was something like [~7 metres per second?], whereas in the Airey experiment, the water's not supposed to be moving wrt the telescope, but the seasonal variation in speed is presumably the Earth's changing orbital velocity wrt the background stars, of (I think) +/- ~30 km/s.

Modern textbooks on optics seem to agree that lightspeeds are dragged by moving transparent particulate media, so if we then come across another information source that suggests otherwise, I think we have to do some head-scratching as to why there seem to be two conflicting experimental results being quoted. ErkDemon 01:37, 20 July 2005 (UTC) ...

1. Relativity is not a theory about the propagation of light. It does, however, allow the explanation of almost all classical optical effects that involve relative velocities.

2. Relativity does not maintain that motion has no effect on the propagation of light. It predicts all the classical effects as well as the Fresnel Drag coefficient.

3. The Fresnel Drag Coefficient does indeed explain Airy and Fizeau.

4. The ONTOLOGY used by Fresnel is incorrect. The effects he predicted are better explained by a theory that accounts for a wide range of other effects, not just optical, and is consistent with mathematical theory. See Special relativity for beginners.

5. Fresnel's analysis is only consistent with physical reality to a first order of approximation See http://arxiv.org/abs/physics/0412055.

loxley

re: [1] There are a several different theories of relativity. Einstein's special theory of relativity does derive its particular relationships by making certain key assumptions about the propagation of light, namely, (a) that the speed of light is constant for the observer (Einstein, "Relativity", ch.7), (b) that lightspeed is isotropic, that the one-way velocity of light A->B is the same as that in the opposite direction, B->A (ch.8), and (c), that the local speed of light at the observer's position extends without change throughout the experimental region being studied - that spacetime is "flat" and "absolute" in the sense that lightspeeds are assumed to be wholly unaffected by the motion of matter. All three of these key conditions used to derive SR's equations are explicitly violated in the Fizeau experiment, we can hold onto (a) in a more limited form by saying that perhaps the principle of relativity only really requires lightspeed to be locally constant, but this does not exclude the possibility of a "relativistic" light-dragging model, where the speed of light is constant for every physical observer due to their matter dragging local light, and where the resulting equations would presumably have to be different to those of SR.

The 1905 presentation of special relativity is not used any more and has scarcely been used by any theoretical physicists since 1908. It is used by opponents of special relativity who have an agenda for being opposed to SR 'in principle' (on religious and other grounds). See Special relativity for beginners for a sample of the modern approach.loxley 09:39, 22 July 2005 (UTC)

Re: [2], special relativity does presume that the motion of matter has no effect on the propagation of light - this assumption is used when we use Einsteinian arguments about arrays of clocks and rods, or trains and embankments (Ch.9). The assumption is used to justify "flat" lightbeam geometry. If moving objects drag light, the resulting lightbeam geometry is curved, and special relativity is known not to work consistently in curved spacetime. A "flat" coordinate system is effectively specified by Einstein when he talks about using "rigid" reference systems and bodies, in dragged-light models, the lightbeam references are not rigid. Einstein's book "Relativity: The special and the general theory" covers much or all of this.

No, this is untrue, SR predicts that moving systems will indeed demonstrate optical effects due to the movement between observers. SR predicts the Fresnel result, it predicts doppler shift etc. The predictions are numerically almost identical with the old aether based predictions which had been refined to match experimental results. loxley 09:39, 22 July 2005 (UTC)

Special relativity does not "predict" all classical effects without external help, it does not, for instance, "predict" that light in a region containing water molecules will have a different speed to light in vacuo, or that light will travel faster in one direction than the other in the region due to the influence of moving material. We stipulate these non-SR behaviours as a starting point when modelling Fizeau because we already know them to be how physics really operates in real life, not because they agree with how SR would predict that light should behave (see: Einstein's "train&embankment" thought experiment).

No one has made any claim that SR predicts all classical effects without other theories. SR predicts "almost all classical optical effects that involve relative velocities". The velocity of light in transparent media is given by Quantum Electrodynamics (QED).loxley 09:39, 22 July 2005 (UTC)

SR does not claim validity for predicting the behaviour of light in a physical particulate medium. However, if we take the known non-SR behaviour of light in water as a "given", SR can be applied to the moving block of water and light as if it was a single object, and can be used to correctly predict that the maximum speed of light in flowing water should (in this exercise) be less than (the speed of light in stationary water in the lab) + (the speed of the moving water in the tube). But the "prediction" is then partly due to the theory, and partly due to the skill and judgement of the person who first worked out how to apply SR to the problem and still get a reasonable answer. There's a certain amount of skill and artistry and experience involved in the decision to do the calculation this way. There's an element of "wrangling" involved, it's not just a question of applying SR equations deterministically, according to the principles of the theory, it's also partly about fiddling about with the methodology until you finally get an answer that you accept as being reasonable.

SR predicts the effects that occur as a result of the relative velocities.loxley 09:39, 22 July 2005 (UTC)

re: [4] I'm not saying that other aspects of Fresnel's model might not suck, but in this context, it couldn't really be faulted, AFAIK.

SR and Fresnel give almost the same predictions for Airy and Fizeau. Fresnel's model is incredibly vague, he gives no indication of the nature of the aether that supposedly piles up in moving objects and his theory amounts to no more than a mathematical adjustment with density terms (mathematical weightings) to make the equations correspond to the experimental results. Fresnel's ONTOLOGY is faulty.loxley 09:39, 22 July 2005 (UTC)

re: [5] Perhaps it's more accurate to say that Fresnel's analysis and SR's are only consistent with each other to a first degree of approximation. In the context of the Fizeau experiment and velocity-addition, both seem to be equally consistent with reality - a footnote in chapter 13 of the Einstein book cedes that although SR's equation is described as "very exact", it's actually not any more "exact" than Fizeau's prediction, and in fact, the degree of "inexactness" is enough for Fizeau's equation to be used as a more convenient stand-in for the SR version.

Please see the reference at the foot of the article for a modern account.

So while we might decide that the SR relationships are the correct description of physical reality, and that Fresnel's are only an approximation, we apparently can't invoke the Fizeau experiment to support that decision. Fresnel's model gets "two out of two" for correctly predicting both the lightspeed offset dragging effect and the reduced maximum velocity, while SR gets "one out of two" for predicting just the reduced maximum velocity, when we decide to treat the velocity of light in water as if it refers ot somehting other than a light-signal.

Interpreting the Airey experiment is not all that straightforward. If a transparent medium such as water (or glass, or an astronomer's eyeball) drags light, and the relative motion of an object causes a change in the angle at which light strikes it, then presumably if we place a block of transparent material in the path of the astronomer's eyeball, the same speed-dependent angle-change will happen at the new surface, and then be transmitted faithfully back to the eyeball. There's then not an obvious reason to expect that the final change in the the angle of the light due to its encountering a dragging medium would be any different for "full" and "empty" telescopes (although perhaps the distance at which angle-change is supposed to occur within the scope might conceivably modify some of the optics).

If our telescope is an extremely long, very thin tube, the position at which the angle changes might become more important, since it might affect whether an lightray is capable of physically entering the instrument's aperture and reaching the far end without running into the walls. But telescopes are typically not an extremely narrow shape, because they need a wide primary aperture to pull in as much light as possible, and if they were, the enhanced geometrical intimacy of a long thin tube with a lightray, might be expected to increase the dragging influence of the tube. It would perhaps be "odd" if lightdragging happened only between the molecules of particulate media and not at all around the particles at the object's edge, so some sort of "edge effect" probably ought to be expected by default in these models, and if a reflecting mirror is used to focus, it would also perhaps be "odd" if the surface of glass dragged light but not the surface of brass, etc.

Once we allow "edge effects", we run into the problem that the "matter drags light" idea does not immediately tell us how far from an object the supposed light-dragging effect would be expected to extend. If it follows the inverse square law, then perhaps the Earth's moving gravitational field drags light, in which case the angle transition (or some of it) would have already happened before the light reaches the telescope, regardless of whether the scope is filled with water or air.

And more modern theory suggests that perhaps moving gravitational sources ought to drag light. If gravitational coupling can allow a moving star or planet to drag along nearby matter and allow momentum exchange between separated objects (slingshot effect), then if matter and light are affected similarly by a moving gravitaitonal field, a moving gravitational source perhaps ought to drag nearby lightspeeds, too. This sort of effect seems to be predicted (and supposedly verified) for rotating stars, where the receding edge of a rotating star pulls light more strongly than the approaching one (frame-dragging).

Rotational frame-dragging is a consequence of "Mach's principle" and is implemented in Einstein's general theory of relativity, and Mach/Einstein arguments also lead to the prediction of light-dragging effects around an object forced to accelerate in a straight line (see: Einstein's 1921 Princeton lectures for a mention of both effects).

This gets us onto the subject of "gravitomagnetism", or "gravitoelectromagnetism" (GEM).

The extension of gravitational dragging from the "rotational" case to the case of circling double stars and then on to the case of simple inertially-moving stars seems straightfoward, but since GR is supposed to reduce to SR, perhaps its not easy for people to take that last step. It's not unusual to find areas of GR where we have to bend or break general principles in order to accomodate special relativity's flat-spacetime logic, and I think this might be one of those.

So it is a little difficult, looking at quick descriptions of Airey's experiment, to see exactly what it really proved or disproved about light-dragging and what lessons we should take from it. The physics of an "SR universe" certainly requires that these complicating light-dragging effects don't exist, but experimental evidence doesn't seem to back that up. ErkDemon 01:37, 20 July 2005 (UTC)

Science is a process of increasingly accurate mathematical approximation to the state of the universe. Most scientists would be very upset to find that Modern Relativity was not an approximation, no one wants to be redundant! The only issue here is whether a theory that involves the drag of a mysterious substance is a better approximation to reality that SR. It is not. loxley 09:04, 21 July 2005 (UTC)

Well, once we get past the hurdle of special relativity and on to general relativity, the "mysterious draggable substance" rejected by SR arguably reappears, in the guise of of the gravitational field parameters (which Einstein once referred to in an aether lecture as "the aether of general relativity"). In a Machian theory (which GR was intended to be), an accelerated or rotating object distorts lightbeams and drags nearby matter and light along, and there also seem to be velocity-dependent dragging effects (slingshot effect), that certainly look like the gravitational counterpart of Fizeau-type particulate-matter dragging effects.

Whether one chooses to refer to these effects as "dragging the substance of spacetime", "dragging inertial frames", or "dragging the aether" seems to be largely down to personal preference. ErkDemon 20:30, 21 July 2005 (UTC)

No, Special Relativity is consistent with differential geometry from about 1840 and physical experiment. This is not a matter of personal preference for physicists. SR works, it predicts QM, spin, QED etc. and is the foundation of modern physics. The 'aethers' that can be postulated in GR (Dirac Ether etc.) are not at all like Fresnel's, this article is about the 'aether drag hypothesis' and there should be no doubt amongst readers that this hypothesis is part of the history of science. loxley 09:39, 22 July 2005 (UTC)

Fizeau and SR

The Fizeau experiment supposedly showed that the one-way velocity of light in a particulate medium is physically offset by the motion of the medium (ie, that the moving particles "drag" the light). A transverse dragging experiment was done by R.V. Jones in 1971 with light aimed through a spinning perpex disk, reporting a sideways deflection in the registered position of emerging light, that that depended on the speed of the intervening material. This effect appears to run counter to special relativity's assumption that the motion of objects (observers, clocks, rods, etc.) has no effect on the propagation of light.

1. Relativity is not a theory about the propagation of light. It does, however, allow the explanation of almost all classical optical effects that involve relative velocities.

2. Relativity does not maintain that motion has no effect on the propagation of light. It predicts almost all the classical effects as well as the Fresnel Drag coefficient.

3. The Fresnel Drag Coefficient does indeed explain Airy and Fizeau.

4. The ONTOLOGY used by Fresnel is incorrect. The effects he predicted are better explained by a theory that accounts for a wide range of other effects, not just optical, and is consistent with mathematical theory. See Special relativity for beginners.

5. Fresnel's analysis is only consistent with physical reality to a first order of approximation See http://arxiv.org/abs/physics/0412055.

loxley

However, the existence of this apparent dragging effect is specifically excluded from tests of special relativity, on the grounds that special relativity only claims validity for modelling the behavior of light in vacuo, in situations where the speed of light seems to be constant. The argument is then that these light-dragging effects do not invalidate SR because the fact that light has a different speed in a particulate medium was already known: this situation is then declared to be outside SR's domain of applicability (i.e. as not being a "fair" test of special relativity).

Special relativity is not a theory about the propagation of light. See Special relativity for beginners. I am deeply puzzled about why you believe that Special Relativity is specifically about light propagation. loxley 09:04, 21 July 2005 (UTC)

Consequently, the apparent dragging effect of matter on light tends to be better documented in modern "optics" textbooks (where students need to know about the effect) than books teaching special relativity (where they don't). Typically, dragging effects seem to be explained by the extinction theorem, which describes an incoming wavefront being absorbed (and "extinguished") by atoms in the medium, which act as signal transponders, generating a new wavefront that then moves at an absolute speed of cMEDIUM wrt the medium's state of motion, the bulk motion of the medium providing a local preferred frame for the speed of light.

Special relativity has a limited amount to say about the Fizeau experiment, since the primary effect demonstrated in the experiment is not predicted by the special theory, and is not compatible with it. We can't really call the main Fizeau result "relativistic" in the SR sense, because it assumes profoundly non-SR behaviour for light.

1. Special Relativity is not specifically a theory about the propagation of light in various media. It interrelates the velocity of things between reference frames so will predict adjustments for velocity related phenomena.

2. The Fizeau result is predicted by Special Relativity, which gives the identical formula to Fresnel to a first approximation. (See main article).loxley 09:04, 21 July 2005 (UTC)

So, the "success" of the special theory in this scenario is instead limited to its ability to correctly predict that, when we measure the speed of light progressing through a sample of the particulate material stationary in the lab frame "v1" (and treat this as a "conventional" velocity, not a light velocity), and we then move the material through the lab in the same direction, at velocity "v2", the final velocity of that signal in the new situation is found to less than the simple sum "v1+v2", and this is then interpreted as validation of the SR velocity addition formula.

Yes, and it is a better explanation that Fresnel's Theory because it also applies to the whole of physics and incorporates modern mathematical results.loxley 09:04, 21 July 2005 (UTC)

In the earlier "light-dragging" explanation, the final speed of the signal was also expected to be less than "v1+v2", because once the material was moving through the lab, the signal was expected to be dragged partly by the atoms in the material, but also partly by the material of the tube, the laboratory, and perhaps also the adjacent planetary mass of 6*10^24 kg that the laboratory happens to be sitting on.

Even in the limited context of the velocity-addition formula, special relativity's predictions for the Fizeau experiment apparently aren't any better than Fresnel's, even though the SR equations are longer - it seems that Fresnel's experiment used too low a water velocity for us to be able to tell the two predictions apart. This has led to a slightly odd (IMO) application of Occam's razor - we favour the SR explanation as simpler, but favour the Fresnel math because it's shorter. We say that the Fresnel explanation is more scientifically concise because it means that we don't have to hypothesise complicating dragging effects, but in this situation, the dragging effects aren't purely hypothetical, but physically real. ErkDemon 03:46, 20 July 2005 (UTC)

The SR results are indeed 'better', in the sense of more complete and wide ranging, than Fresnel's. See http://arxiv.org/abs/physics/0412055 loxley 09:04, 21 July 2005 (UTC)

Flat or dragged?: different branches of relativity theory

Special relativity is not a theory about the propagation of light. See Special relativity for beginners. I am deeply puzzled about why you believe that Special Relativity is specifically about light propagation. loxley 09:04, 21 July 2005 (UTC)

It's true that special relativity also modifies most aspects of Newtonian physics. But those modifications take the form that they do because the special theory chooses to combine the principle of relativity with a particular assumption about how light should propagate. SR bases a whole stack of definitions and derivations on observation, and all those observations hinge on how light is supposed to behave as it moves from an object to the observer. Lose the simplified propagation model and you seem to lose the whole theory. ErkDemon 01:33, 27 July 2005 (UTC)

Modern SR predicts that there will be a universal constant velocity and that this will be the speed of propagation of light. It does not assume this. Please see Special relativity for beginners. loxley 08:38, 27 July 2005 (UTC)

Modern SR “predicts” this based on the attributes of Minkowski’s geometry, but that geometry and our modern treatment of it were selected because they correctly reproduced the relationships of special relativity, and those SR relationships were derived by assuming that c was a universal constant, for all legal inertial observers. So it’s slightly circular: this “prediction” has already been hard-coded into the theory by a human selection process, even if the final geometry itself appears to come directly from Mount Olympus. Minkowski geometry has this "desired" attribute, and this is part of why Minkowski geometry is considered to be the correct geometry for SR. The model is “predicting” the behaviour that we deliberately designed into it. The existence of geometry does not make a physical theory, one also has to have a justification for choosing that particular geometry, and a set of rules for how it is to be "correctly" applied and interpreted. ErkDemon 01:59, 3 August 2005 (UTC)

history

The principle of relativity seems to require that lightspeeds be locally constant for each observer.
Special relativity went further and declared that lightspeeds were globally constant for every observer, as a way of simplifying the problem. In 1905, Einstein took the idea that "we know that the speed of light is constant" to mean that we should observe c to be constant in our own inertial frame through an extended region of spacetime that could include different objects with different states of motion. The motion of those objects (which we might be observing) had to be assumed not to affect the speed (or the velocity) of light in any way for this to work. So, with SR, "we know that c is constant" was used to mean, "we know that lightbeam geometry is flat", and therefore, that, "we know that objects that we observe, moving though our region, are not affecting the lightbeam geometry of the region".

Modern SR is a theory of a particular, flat, space-time. Please see Special relativity for beginners. The theory is a real theory that has defensible axioms and makes predictions about light propagation given these axioms. It is limited to flat space-time. GR is the theory that encompasses non-flat space-times. GR is different from SR and has a different ontology. (like SR has a different obtology from Newtonian physics) loxley 08:38, 27 July 2005 (UTC)

"Global lightspeed constancy" + "the principle of relativity" led to an apparent paradox: if scientist "A" in their lab is "moving" at v m/s, and can presume that lightspeed is fixed in their' frame, how does scientist "B", with a different state of motion, manage to describe the behaviour of the same lightbeam by declaring that it is travelling at c in their frame? That's the problem posed by Einstein in his relativity book, Chapter 7, "The Apparent Incompatibility of the Law of Propagation of Light with the Principle of Relativity", and the proposed solution was Lorentz corrections, tilted planes of simultaneity, velocity addition formulae, and all the other definitional infrastructure used by special relativity to redefine distances and times in different frames using the round-trip speed of light.

Minkowski and Hilbert were treading a different path, realising that SR was in fact part of geometry. Einstein finally accepted the geometrical approach in the 1920's. You can see his change to geometrical reasoning actually occurring in "Relativity".

Alternative route

But since the principle of relativity only required local lightspeed constancy, there was a second way of satisfying it. In a light-dragging model, scientists "A" and "B" can both agree that lightspeed has the right value in their own respective labs, but can also agree that the speed of light in the "other guy"'s lab is offset and "correctly wrong" because of the motion of the other lab's hardware, dragging it along.

With that alternative approach, instead of worrying about the details of how two different observers in Einstein's "train&embankment" thought experiment could manage to observe the same lightbeam to be moving at the same fixed speed in both frames (and redefining everything else to produce that outcome), we shrug and implement the PoR by saying that both observers agree that a lightbeam skimming both train and track moves not at c in either frame, but at an intermediate speed, with the train observer reasoning that the beam advances "too slowly" because it is being dragged back by the proximity of the moving embankment, and the embankment observer reasoning that the beam progresses "too quickly" because it is being dragged forwards by the proximity of the moving train. This explanation seems to broadly agree with the sort of thing that we see happening in the Fizeau experiment.

Minkowski came up with a much simpler answer that was already known to mathematicians of the time and provides a deep insight into reality. see Special relativity for beginners

While the Lorentz-Einstein approach gives flat spacetime and Einstein's special theory, the second generates a curved-spacetime description, and since the SR relationships are a "special" solution that is only supposed to work when spacetime is perfectly flat, a hypothetical "light-dragging" theory of relativity would seem to be compelled to use a different set of equations to SR.

That's the biggie. The math of special relativity and LET isn't a general solution or (apparently) a necessary part of a general solution unless we postulate that that velocity effects don't drag light.

This is slightly awkward. Although it was probably more convenient to try to write a full theory of relativity by starting out with "easy" flat spacetime, getting SR as a unique solution, and then build up to a more ambitious "general" theory that used curved spacetime and still reduced to SR (i.e. GR) ... if light-dragging is a "real" physics effect (as it seems to be), and obeys the principle of relativity (as relativists would hope, otherwise we are in trouble), the correct form of equations for a full description of mechanics, which includes the dragging effects that we omitted to simplify the problem, seemingly can't be the form provided by the special theory. And if SR is using the "wrong set" of relativistic equations, then if general relativity is designed to reduce to special relativity, at least part of its design may be have been corrupted, to get things to fit.

General relativity is different from special relativity. In SR the space-time is absolute, a background. In GR the space-time is a product of the manifold of masses and energy (a modified form of Mach's principle). GR predicts that the distribution of mass and energy in the universe will give the metric of SR in our local part of the universe. The space-time of SR can host preferred frames and an aether but then the idea of general covariance and relativity would need to be abandoned.

options

At this point we seem to have two main options, modify or rewrite.
With the "modify" option, we can hope that SR is only "slightly wrong, and derive at least one other set of relativistic equations (perhaps involving additional "Lorentzlike" terms) whose purpose is to describe corrections to SR that might become significant at high velocities or high densities, or with intense gravitational fields.

SR cannot easily be 'wrong', being a theorem in differential geometry. It might however not apply to the universe. We could, for instance, suggest as you have done that the distribution of material in the universe is finely adjusted to appear like a hyperbolic geometry. It seems a bit odd however to favour a 3D Euclidean space to such a degree that we are prepared to suggest effects that are specifically adjusted to make it look like hyperbolic space. Please note that SR does not apply at ultra-high energies and densities, this is an inherent limitation of the theory, we must switch to GR for a theory of these effects. loxley 08:38, 27 July 2005 (UTC)

• high-gravity limit? - Since "moving" gravitational sources should drag light along with them (as they drag matter - slingshot effect), light passing through a "moving" gravitational field should have its a momentum (and energy, and frequency) altered by the motion of the field, so that we could expect a shift effect associated with the light-dragging effect. We then have to ask whether or not the dragging shift also affects light emitted from the star, and there are a further two options, neither of which look good for special relativity: Either the basic motion shift of the star is predicted correctly by SR, and the dragging-shift effect acts on top, so that the star's total motion-shift predictions disagree with SR ... or, if we insist that the gravitational field is just an extension of the object's mass and momentum, so that the dragging shift and the "conventional" velocity shift are one and the same, and shouldn't be counted twice, then this "conventional" velocity shift description has to be compatible with velocity-dependent curvature, which seemingly rules out the SR equations for our basic equations of motion, and again, the motion shift of the star does not follow the SR math.

So either SR is nominally right about the underlying motion shifts of stars (but wrong about the shift that an observer actually sees), or its not even nominally right.

If we try to cling on to SR here and say that it's still got the right equations of motion for low-gravity objects, we end up with different "laws of mechanics" for "gravitational" and "non-gravitaitonal" objects, and presumably some way of fading between the two depending on surface gravity, which is starting to get a bit messy.

It is not messy at all. SR is specifically stated to be a theory of flat space-time. Some other theory is needed for non-flat space-times. GR is that theory at present and supercedes SR. loxley 08:38, 27 July 2005 (UTC)

• high KE limit? - If we say that energy warps spacetime, then perhaps we should expect a very high-energy particle to create a certain amount of curvature around it. Nowadays we hear talk of modern experiments being able to generate such high energy-concentrations that some are even expected to spawn mini-blackholes - if that's accurate, I think we have to have already crossed over well into the "curved-spacetime" range, we can't be looking at gravitational event horizons and still be claiming that curvature plays no part in the physics. So at some point, accepting that velocity-dependent curvature is real means that we have to expect that SR's geometry becomes progessively more unrealistic at higher velocities, and if the geometry is wrong, the physical predictions are liable to be wrong, too (unless we apply some sort of QM "fix" to bring things back into line with reality). Perhaps when a significant fraction of the particle's total energy (or even most of it) is tied up in its relative motion (rather than its rest mass), it becomes natural to suggest that the particle's momenergy exists physically as curvature between the "moving" particle and its "stationary" neighbours. But since the SR equations (again) seemingly can't apply in this sort of model we again have to suggest an additional set of equations - either the particle gets modelled by non-SR equations all the way up from the Newtonian range to the ultrarelativistic, without going via SR, or we have three nominal velocity ranges to deal with: the "Newtonian" range, the SR range, and the "significant curvature" range, where SR is only a first approximation. Again, this starts to get very untidy.

As we work through the exceptions to SR (how about another one for QM effects, where the fuzziness of a particle appears as a quantum field that smudges out its momentum and deflects nearby light like a gravitational field does?), the range of situations where we expect velocity-dependent curvature gets larger, and the range of situations where SR might have the correct equations gets smaller (Hubble recession shift? Apparently non-SR), and we then have to go to more and more extreme lengths to justify there being any range in which SR might legally apply (classical radius of an electron? Suspiciously similar to its gravitational event horizon radius).

de Broglie predicted QM on the basis of SR (not GR). Feynman predicted QED on the basis of SR. loxley 08:38, 27 July 2005 (UTC)

Unfortunately, mainstream research on this class of model seems to be ruled out by the current classification schemes that list "reduction to SR" as being an essential feature of any gravitational model that is to be considered "credible" (C.M Will, "Theory and experiment in gravitational physics").

The problem with much of the critique of SR is that appears like a love affair with 3D Euclidean geometry. Many of the critics seem unaware of the problems of an idealised Euclidean description of the universe. The better critics do not approach the problem as a binary rejection of SR in favour of school physics, they suggest other metrics. loxley 08:38, 27 July 2005 (UTC)

So for now we seem to be stalled. Einstein told us fifty years ago (Scientific American, April 1950 issue) that he now considered the adoption of SR to be a historical accident, and no longer justifiable with hindsight as the basis of proper gravitational theory, but then he died, and instead of doing the work, we seem to have spent the next half century coming up with new ways to justify not tackling the problem (e.g., "We know that the speed of light is constant, so there's no case to answer"). ErkDemon 01:33, 27 July 2005 (UTC)

Einstein was telling the readers that GR supercedes SR. SR has an absolute space-time, something that Einstein abhorred. GR gives the same result as SR if the gravitational field is flat but proposes that space-time is dynamic and identical to the gravitational field, SR on the other hand starts by proposing that there is an absolute space-time, a container with fields inside it. SR is the big brother of Galilean physics/relativity, an amended absolute space theory that takes into account modern maths and experiment. The reconciliation of GR with QM is stalled but is the subject of intense activity. In many ways this is a battle between SR (which is QM compatible) and GR. The modern theory of SR, created by mathematicians after 1908, is actually the enemy of Einstein's relationalist conception of the universe. loxley 08:38, 27 July 2005 (UTC)