Relativity without the aether: pseudoscience?

JesseM

When you say "yet", are you suggesting you see no reason why some future phenomenon might not respect local Lorentz-invariance? If so, you are underscoring point #6 from the sci.physics.relativity post by Tom Roberts I quoted in my last post:

Hurkyl

"True" and "false" aren't really (internally) applicable to science. What is true is that each of the postulates of SR are empirically verifiable, whereas the same cannot be said of LET.


No it doesn't.

To even begin to say something like "the one-way speed of light is isotropic", it requires one to specify a coordinate chart.

Coordinare charts are nonphysical choices.

You're thinking about the fact that, among all possible rectilinear coordinate charts we could use, SR chooses to define the orthogonal ones as the inertial reference frames.

This choice is exactly analogous to the fact that we generally like, when studying 3-space, for our x, y, and z axes to be perpendicular.

Special Relativity is generally done in these orthogonal, linear coordinate charts for exactly the same reason that coordinate geometry is generally done with perpendicular axes.


Another good reason to use the orthogonal, linear coordinate charts is that such things can be defined experimentally. (Of course, so can other types of coordinate charts)

Contrast to the choice of charts used by aether theories which invoke some principle of absolute simultaneity which cannot be experimentally determined.


(Since I'm talking about "orthogonal" in the above, that means I'm using some sort of "metric". Of course, I'm using the "metric" of Minowski 4-space, because that's the one that appears in all the physical laws)

Aether

Here's the first two lines from a paper from Kostelecky & Mewes for example: http://www.citebase.org/cgi-bin/cita...hep-ph/0111026 "Lorentz violation is a promising candidate signal for Planck-scale physics. For instance, it could arise in string theory and is a basic feature of noncommutative field theories...". So, when I say "yet" I simply mean that I am aware of many physicists who expect to find violations eventually.

I know that there are some good points in Tom Robert's articles, and I have read them in the past, but he's talking about a whole infinity of ether theories and I'm just talking about one; LET. So, kindly extract the point from the article that you think applies here.

He (and you by quoting it) seems to think that local Lorentz invariance is synonymous with SR to the exclusion of LET:

"Note that all phenomena discovered since 1905 do indeed exhibit the local Lorentz invariance of SR -- what is happenstance in ether theory was directly predicted by SR."

Aether

Let's get empirical then. Show me how the constancy of the one-way speed of light is empirically verifiable; the round-trip speed of light is constant in both theories.

Precisely the same metric that is used by SR is also used by LET. The difference is that the speed of light is invariant while simultaneity is relative in the Lorentz transformation, but absolute simultaneity is maintained with a variable speed of light in LET transformation. Both are equally valid.

JesseM

Since I'm not well-versed in quantum gravity I can't say much about this, but I wonder if by "Lorentz violation" they actually mean a preferred reference frame as in aether theories, and if so I wonder how mainstream this point of view is. A quick google search turned up this paper which seems to say that noncommutative theories can be compatible with Lorentz symmetry. The points in his article would apply to any theory that posits a special reference frame (for example, a theory that says there is only one frame where the speed of light is 'really' c in all directions) and yet does not make any predictions about the results of actual experiments which are different from those of SR (so all observers will measure the speed of light to be c in all directions, even if this is explained as a faulty measurement because their clocks are not ticking the 'correct' time and their rulers are not reading the 'correct' length). I don't know exactly what you mean by "LET", does it fit both these criteria? That quote doesn't say that aether theories don't exhibit local Lorentz-invariance, it just says that it is an unexplained "happenstance" if they do. In aether theories you need a multitude of separate coincidences to explain why every new phenomena happens to exhibit lorentz-invariance, and you have no reason to predict that new phenomena will exhibit it, whereas SR makes a clear prediction that all phenomena must exhibit local lorentz-invariance, and gives a single conceptual explanation for why they all do.

Hurkyl

Given the Lorentz metric, the one-way speed of light is a mathematical consequence of the Special Relativistic definition of an inertial reference frame.

You agree that the Lorentz metric is empirically verifiable, correct? Then so must the one-way speed of light in a coordinate chart that is one of Special Relativity's inertial reference frames.

(I erred slightly in my previous post -- I should have said orthonormal, rather than orthogonal)


Maybe this is a clue to the psychology of those who cling to aether theories? Do they believe that coordinate charts are physical things, rather than simply a choice we make when modelling a problem?

pervect

I think this is a good observation.

Basically, using for simplicity units where c=1, the coordinate transformation

t1 = [itex]t + \beta x[/itex]
x1 = x
y1 = y
z1 = z

transforms the Minkowski metric

ds^2 = dt^2 -dx^2 - dy^2 - dz^2

into a different metric

ds^2 = dt1^2 - 2 [itex]\beta[/itex] dx1 dt1 - (1 - [itex]\beta^2[/itex]) dx1^2 - dy1^2 - dz1^2

[itex] \beta [/itex] plays the role of a velocity here and is equal to v/c, a number between zero and 1.

So the isotropy of the speed of light comes from using the Minkowski coordinate chart, by using a different coordinate chart one can work in the LET coordinates where the "coordinate speed" of light is not constant.

Both coordinate charts are flat in that they have a zero curvature tensor.

The convention that velocities are reported in a Minkowskian metric is very common and useful, for reasons regarding momentum that have been and are being discussed in another thread (no need to repeat them here).

The main "feature" of LET that I see is that people who are philosophically comitted to the Gallilean transform have a method that should theoretically enable them to work problems in relativity, in spite of their philosophical blinders.

εllipse

I've got a question for Aether.

By "absolute simultaneity", do you mean that in LET two events which happen simultaneously in one reference frame happen simultaneously in all, or just that there is a prefered reference frame (the ether) and this reference frame is the one in which simultaneous events are called absolutely simultaneous?

Perspicacious

The one-way speed of light is the same when computed by the Lorentz transformation or by absolute frame coordinates. Here's the experiment.

In an arbitrary frame of reference, start with two synchronized clocks side-by-side and slowly transport one of them to any convenient distance D. When the ultraslow clock transport is finished, send a light pulse from the stationary clock at local clock time t1. Let t2 be the time when the light arrives as measured by the slowly transported clock. Take the limit of ultraslow transport in the expression D/(t2-t1) for a perfect answer of c.

Are you asking to be tutored in relativity or do you disbelieve the mathematics?

Aether

In LET two events which happen simultaneously in one reference frame happen simultaneously in all. The hypothetical preferred reference frame is a common synchronization reference.

Aether

I like your point about slow clock transport, and agree that it is equivalent to Einstein synchronization. However, in LET those clocks are re-synchronized according to [itex]f(v,x)=-vx/c_0^2[/itex].

Aether

I'm not sure what you mean by the statement in parenthesis.

I am using "LET" as a label the ether transformation equations from M&S-I (see my post #92 in the "consistency of the speed of light" thread for details); this may not be exactly what anyone else, particularly H.A. Lorentz, means by LET.

Is this how you would explain the essential differences between the SR and LET transformation equations (from post #92 referenced above)? It seems like a very simple choice of synchronization convention to me when I compare those two sets of equations.

Perspicacious

If your only point is that every inertial observer in SR can choose absolute frame coordinates, then I have no objection to that. If you insist on the axiom that says that there is only one frame that can use absolute frame coordinates, then I object. I object to any axiom that has no useful consequences.

The freedom to reset clocks in the universe with arbitrary synchronizations is widely regarded as legitimate physics.

http://arxiv.org/abs/gr-qc/0409105
http://www.everythingimportant.org/relativity/

Aether

OK, it is a postulate, not determined by experiment. If that is what you mean by this, then I agree.

Essentially yes, up to the point that many physicists are still looking for violations.

Empirically verifiable, Hurkyl? No, I don't see that yet. I can see that it's in the Lorentz transform [itex]\Lambda_\nu^\mu[/itex], not in the LET transform [itex]T_\nu^\mu[/itex], but what has that to do with the metric [itex]\eta_{\mu \nu}[/itex] per se?

And maybe this is a clue to the psychology of those who cling to SR as well? That was roughly my point in starting this thread.

Aether

I don't insist on such an axiom, no. I am saying that such a postulate is not empirically different from the SR postulate that the speed of light is a constant in all inertial frames; that an impartial observer, not from our culture, would have no reason (today) to prefer one over the other; and that being able to see the world from both perspectives is better than choosing either one arbitrary extreme or the other. However, if someone here actually proves that SR is empirically right and LET is empirically wrong then that's that.

Hurkyl

If you recall, I said very distinctly:

(Though I've modified the emphasis)



No, it is not a postulate, it is a definition.

I repeat that a coordinate chart is not a physical object. One does not need to make postulates, one can write down a precise definition and say "in my theory, the words 'inertial reference frame' absolutely, positively, unequivocally means this:".

Now, one might ask the question if there is any such thing satisfying the definition, but let me remind you that even in a LET, you can take this definition to produce a coordinate chart in which the one-way speed of light is isotropic.

Let me say that again:

In a Lorenz Ether theory,
the Special Relativistic definition of "inertial reference frame"
yields a coordinate chart
in which the one-way speed of light is constant.

Of course, since this chart was chosen to satisfy the SR definition of "inertial reference frame" and not the LET definition of "inertial reference frame", there is no reason to expect this coordinate chart to be an LET inertial reference frame (and generally it will not be).

The point stands -- even in a LET, you can prove that SR-inertial reference frames have a constant one-way speed of light.


You focus too much on coordinates, as if they're a fundamental thing -- it's the geometry that matters.

Both theories let you work in whatever coordinates you like. The difference between SR and LET is that SR postulates nothing more than the geometry. Things like "inertial reference frame" are simply defined from the geometry.

Whereas a LET has to make an additional postulate about absolute simultaneity: it cannot be defined from the Minowski geometry alone.

Let me repeat this:

Both theories postulate Minowski geometry. However, SR makes no additional postulates, defining everything else from the geometry.

However, a LET requires at least one additional postulate about absolute simultaneity, since that cannot be defined from the geometry.


Actually, as I understand it, LETs do not postulate Minowski geometry -- they postulate some other background geometry, postulate some properties of an aether, and then derive the Minowski geometry, or alternatively, let the experimental verification of Minowski geometry constrain the properties of their aether.

Aether

What is the geometrical difference between SR and LET? Or, do you operate in a coordinate independent geometry where there is no SR & LET per se? If so, how do I get there from here?