# Relativity without the aether: pseudoscience?

Gold Member
JesseM said:
(actually I'm not sure about this last one--you wrote the equation $$t_1=(1-v^2/c_0^2)^{1/2}T_1$$ for the time transformation in the ether equation, but then for the Lorentz transform you wrote $$t_1=(1-v^2/c_0^2)^{1/2}T_1-vx_1/c_0^2$$ when it's actually supposed to be $$t_1=(T_1-vx_1/c_0^2)/(1-v^2/c_0^2)^{1/2}$$, so should we also divide by $$(1-v^2/c_0^2)^{1/2}$$ rather than multiply by it in the ether transform equation?). I don't have a copy, so if you could provide one that would be helpful, thanks.
If anyone else would like to download a copy of the Mansouri-Sexl papers in the future, please let me know and I'll make them available. Let's confirm that you approve of their version of the Lorentz transform equations before going any further.

Last edited:
Aether said:
You can download a copy (8MB) from ... for the next few hours. After that I'll erase this link, and if anyone wants it at a later time all they'll have to do is ask me for it. Let's confirm that you approve of their version of the Lorentz transform equations before going any further.
OK, I've downloaded it. The time transformation in the Lorentz transformation they give in equation 3.4 seems incorrect to me--they do indeed write $$t = (1 - v^2)^{1/2} T - vx$$, but if you look anywhere else (like here or here or http://www.bun.kyoto-u.ac.jp/~suchii/lorentz.tr.html [Broken] or here) you'll see that the usual time transformation is $$t = \gamma (T - vx/c^2) = (T - vx/c^2)/\sqrt(1 - v^2/c^2)$$.

Last edited by a moderator:
Gold Member
JesseM said:
OK, I've downloaded it. The time transformation in the Lorentz transformation they give in equation 3.4 seems incorrect to me--they do indeed write $$t = (1 - v^2)^{1/2} T - vx$$, but if you look anywhere else (like here or here or http://www.bun.kyoto-u.ac.jp/~suchii/lorentz.tr.html [Broken] or here) you'll see that the usual time transformation is $$t = \gamma (T - vx/c^2) = (T - vx/c^2)/\sqrt(1 - v^2/c^2)$$.
Have a look at post #42 in the "Einstein's clock synchronization convention" thread. DrGreg says that it's OK. There are other confirmed typos in M&S, and it would be good to know for sure if this was yet another one.

Quick poll: What do you think that the metric represents if not a physical thing aka "the aether"?

Last edited by a moderator:
Aether said:
Have a look at post #42 in the "Einstein's clock synchronization convention" thread. DrGreg says that it's OK. There are other confirmed typos in M&S, and it would be good to know for sure if this was yet another one.
OK, I see what's going on. Instead of showing the formula for t in terms of the (X,T) coordinate system, they're giving the formula for t in terms of the T-coordinate of the (X,T) system and the x-coordinate of the (x,t) system. That seems overcomplicated, but it isn't a mistake. So I was mistaken when I said earlier that time transforms like $$t' = \gamma t$$ in the ether transform, it should actually be $$t' = t/\gamma$$. But this doesn't affect the numerical example I provided in my last long post (#70), because I did use $$t' = t \sqrt{1 - v^2/c^}$$ in coming up with those numbers.

Last edited:
Gold Member
JesseM said:
OK, I see what's going on. Instead of showing the formula for t in terms of the (X,T) coordinate system, they're giving the formula for t in terms of the T-coordinate of the (X,T) system and the x-coordinate of the (x,t) system. That seems overcomplicated, but it isn't a mistake.
Perhaps they did it that way to make it clear that the only difference between LET and SR is that in one this $-vx/c_0^2$ term is embedded in the transformation equation, and in the other this term is plugged into a clock?

Aether said:
Perhaps they did it that way to make it clear that the only difference between LET and SR is that in one this $-vx/c_0^2$ term is embedded in the transformation equation, and in the other this term is plugged into a clock?
Seems like a plausible reason. Anyway, now that we've got this cleared up, do you have any comments on the rest of my post #70?

Gold Member
Special Relativity (SR) has two postulates: not only that the speed of light is invariant but also that the laws of physics take the same form in all inertial frames.

The "LET" transformations quoted by Aether provide an alternative to the first postulate. But they do not deal with particle dynamics and the concepts of mass, momentum, energy and force, which, in SR, depend on the second postulate. The relativistic formulas for momentum and energy are based on the assumptions that momentum and energy are both conserved relative to any choice of inertial observer, and the same formulas apply in all frames. In particular, momentum and energy are assumed to be isotropic.

For example, the magnitude of momentum depends only on the magnitude of velocity relative to the observer, and not on the direction. This has a practical consequence: if you take two particles of equal mass and fire them towards each other at equal and opposite velocities, and they collide and stick together, the resultant mass will be stationary. This result is true, according to SR, relative to any inertial observer. I assume that there is plenty of experimental evidence to confirm this. Perhaps other readers can confirm my assumption.

If the same observer uses LET co-ordinates instead of SR co-ordinates, distance is the same but time is different, so the two particles no longer have the same speed, before collision, in LET co-ordinates, yet still they end up stationary. This establishes that, in LET, momentum cannot be isotropic. (See this thread Einstein's Clock Synchronization Convention if you need more details.)

For LET to successfully compete with SR, it needs its own definitions of momentum and energy, and, for these to be conserved, they cannot be isotropic. I don't know if there are formulations of LET which include additional postulates that will cope with momentum and energy relative to a moving observer. The formulas quoted by Aether do not answer this question.

Gold Member
JesseM said:
Seems like a plausible reason. Anyway, now that we've got this cleared up, do you have any comments on the rest of my post #70?
It is ambiguous to define a velocity as $0.8c$ because c is a variable in LET. You can define it unambiguously as $0.8c_0$, and then it should be the same in both SR & LET. We need to develop a simple but specific example where all of the coordinates are calculated using the Lorentz transform, and then see if the LET transform can't match it. You can't compute the Lorentz transform without providing the $-vx/c_0^2$ term for each clock, and I'm just going to take it and put it in the clock.

Gold Member
DrGreg said:
Special Relativity (SR) has two postulates: not only that the speed of light is invariant but also that the laws of physics take the same form in all inertial frames.

The "LET" transformations quoted by Aether provide an alternative to the first postulate. But they do not deal with particle dynamics and the concepts of mass, momentum, energy and force, which, in SR, depend on the second postulate. The relativistic formulas for momentum and energy are based on the assumptions that momentum and energy are both conserved relative to any choice of inertial observer, and the same formulas apply in all frames. In particular, momentum and energy are assumed to be isotropic.

For example, the magnitude of momentum depends only on the magnitude of velocity relative to the observer, and not on the direction. This has a practical consequence: if you take two particles of equal mass and fire them towards each other at equal and opposite velocities, and they collide and stick together, the resultant mass will be stationary. This result is true, according to SR, relative to any inertial observer. I assume that there is plenty of experimental evidence to confirm this. Perhaps other readers can confirm my assumption.

If the same observer uses LET co-ordinates instead of SR co-ordinates, distance is the same but time is different, so the two particles no longer have the same speed, before collision, in LET co-ordinates, yet still they end up stationary. This establishes that, in LET, momentum cannot be isotropic. (See this thread Einstein's Clock Synchronization Convention if you need more details.)

For LET to successfully compete with SR, it needs its own definitions of momentum and energy, and, for these to be conserved, they cannot be isotropic. I don't know if there are formulations of LET which include additional postulates that will cope with momentum and energy relative to a moving observer. The formulas quoted by Aether do not answer this question.
The second postulate is automatically satisfied because the LET coordinates transform as a tensor. M&S give the general transformation on page 508, and then whittle it down for convenience sake. I will stipulate that energy and momentum need to be conserved in LET, but suppose that they are since the coordinates transform as a tensor.

Last edited:
Gold Member
Aether said:
The second postulate is automatically satisfied because the LET coordinates transform as a tensor. M&S give the general transformation on page 508, and then whittle it down for convenience sake. I will stipulate that energy and momentum need to be conserved in LET, but suppose that they are since the coordinates transform as a tensor.
If you assume SR is correct, you can come up with some expressions for momentum and energy relative to LET co-ordinates by application of the transform equations. But if you ignore SR and try to calculate momentum and energy some other way, how do you find the answer? How, for example, do you prove that, relative to the ether, the relativistic form $p=\gamma_u m u$ should be used instead of $p= m u$?

Aether said:
It is ambiguous to define a velocity as $0.8c$ because c is a variable in LET.
The speed of light varies depending on your reference frame, but I would think that in the ether transform the symbol "c" no longer refers to the actual speed of light in an arbitrary frame, but only to the speed of light in the preferred frame where light travels at the same speed in all directions. After all, doesn't "c" appear in the transformation equations themselves? Mansouri and Sexl just set it equal to 1, but I think if you include it the ether transform equations would look like:

$$x' = (x - vt)/\sqrt{1 - v^2/c^2}$$
$$t' = t \sqrt{1 - v^2/c^2}$$

The c that appears here is a constant, no?
Aether said:
You can define it unambiguously as $0.8c_0$
OK, if $$c_0$$ means the speed of light in the preferred frame, that's what I meant. Also, I did specify that when I said the speed of the ship was 0.8c and the speed of the second observer was 0.6c, these speeds were relative to the preferred frame of the ether transform, which Mansouri and Sexl mean to represent the ether's rest frame (although you do not make this assumption, apparently).
Aether said:
and then it should be the same in both SR & LET. We need to develop a simple but specific example where all of the coordinates are calculated using the Lorentz transform, and then see if the LET transform can't match it.
I don't know what you mean by "can't match it"--match what, exactly? The point of my example in post #70 was not to show that the ether transform is incorrect or that it can't do something that the Lorentz transform can do, but just to show that the speed of a sublight object such as a spaceship or a car depends on whether you choose to use the coordinate systems given by the ether transform or the Lorentz transform, that your earlier statement that "The speed of anything can be measured to have the same value regardless of direction or clock synchronization using a round-trip radar pulse for example" is incorrect because the speed of an object does not always have the same value in an ether transform frame and the corresponding Lorentz transform frame.
Aether said:
You can't compute the Lorentz transform without providing the $-vx/c_0^2$ term for each clock, and I'm just going to take it and put it in the clock.
I don't know what you mean by "I'm just going to take it and put it in the clock". Put what in the clock? What does this have to do with my point that all velocity measurements depend on your choice of coordinate system? (you can substitute 'depend on your choice of clock synchronization' if you like, although it would be possible to invent wacky coordinate systems which differ in ways that go beyond clock synchronization, like one where the ratio of ruler length to coordinate length is different in different regions of space)

Also, do you have any response to the other points I made in post #70? Do you agree now that the "outside observer" I referred to plays no role in the construction of the different measuring devices that different observers use to assign coordinates to events, for example? What about my question about what you could possibly mean when you say there is "no mathematical difference" between the Lorentz transform and the ether transform?

Last edited:
Gold Member
JesseM said:
The speed of light varies depending on your reference frame, but I would think that in the ether transform the symbol "c" no longer refers to the actual speed of light in an arbitrary frame, but only to the speed of light in the preferred frame where light travels at the same speed in all directions. After all, doesn't "c" appear in the transformation equations themselves? Mansouri and Sexl just set it equal to 1, but I think if you include it the ether transform equations would look like:

$$x' = (x - vt)/\sqrt{1 - v^2/c^2}$$
$$t' = t \sqrt{1 - v^2/c^2}$$

The c that appears here is a constant, no?
No. $$c_0[/itex] is the round-trip speed of light, and [tex]c=c(v,\theta)$$.

JesseM said:
OK, if $$c_0$$ means the speed of light in the preferred frame, that's what I meant. Also, I did specify that when I said the speed of the ship was 0.8c and the speed of the second observer was 0.6c, these speeds were relative to the preferred frame of the ether transform, which Mansouri and Sexl mean to represent the ether's rest frame (although you do not make this assumption, apparently). I don't know what you mean by "can't match it"--match what, exactly? The point of my example in post #70 was not to show that the ether transform is incorrect or that it can't do something that the Lorentz transform can do, but just to show that the speed of a sublight object such as a spaceship or a car depends on whether you choose to use the coordinate systems given by the ether transform or the Lorentz transform, that your earlier statement that "The speed of anything can be measured to have the same value regardless of direction or clock synchronization using a round-trip radar pulse for example" is incorrect because the speed of an object does not always have the same value in an ether transform frame and the corresponding Lorentz transform frame. I don't know what you mean by "I'm just going to take it and put it in the clock". Put what in the clock? What does this have to do with my point that all velocity measurements depend on your choice of coordinate system? (you can substitute 'depend on your choice of clock synchronization' if you like, although it would be possible to invent wacky coordinate systems which differ in ways that go beyond clock synchronization, like one where the ratio of ruler length to coordinate length is different in different regions of space)

Also, do you have any response to the other points I made in post #70? Do you agree now that the "outside observer" I referred to plays no role in the construction of the different measuring devices that different observers use to assign coordinates to events, for example? What about my question about what you could possibly mean when you say there is "no mathematical difference" between the Lorentz transform and the ether transform?
This is why I'm asking for a simple example. I would rather develop a simple example so that we can model one simple problem and compare notes on it rather than have too big of a word problem to deal with.

Aether said:
No. $$c_0[/itex] is the round-trip speed of light, and [tex]c=c(v,\theta)$$.
OK, but I didn't even use $$c_0$$ in my version of the ether transform equations. If I define c as the speed of light in the ether frame (which I think is equal to the round-trip speed anyway, assuming each observer uses his own ruler and clock to measure the length and time of the round trip), and then I write the ether transformation as I did in my previous post, are you saying that the equations I wrote there are incorrect?
Aether said:
This is why I'm asking for a simple example. I would rather develop a simple example so that we can model one simple problem and compare notes on it rather than have too big of a word problem to deal with.
How will an example make any difference for the issue of the "outside observer" who has no effect on any physical system and thus cannot possibly make a difference in any numerical example? Do you agree or disagree that observers in windowless boxes can set up measuring devices to assign coordinates to different events such that the coordinates two observers assign to the same event will transform according to the Lorentz transform, but that it is impossible to build measuring devices in windowless boxes whose coordinates transform according to the ether transform?

As for the issue of whether speed depends on your coordinate system, I already provided a numerical example. Again, assume that when I use the symbol "c" it refers only to the speed of light as measured in the preferred ether frame, even if you would use the symbol differently.

Last edited:
Gold Member
JesseM said:
OK, but I didn't even use $$c_0$$ in my version of the ether transform equations. If I define c as the speed of light in the ether frame (which I think is equal to the round-trip speed anyway, assuming each observer uses his own ruler and clock to measure the length and time of the round trip), and then I write the ether transformation as I did in my previous post, are you saying that the equations I wrote there are incorrect?
Not necessarily, but there is an ambiguity with specifying a velocity as 0.8c in LET because that is a function of v and \theta.

JesseM said:
How will an example make any difference for the issue of the "outside observer" who has no effect on any physical system and thus cannot possibly make a difference in any numerical example?
It may not make any difference in the end. But you said that the outside observer is the one who computes the Lorentz transform, and that's getting way more complicated than just a windowless box.

JesseM said:
Do you agree or disagree that observers in windowless boxes can set up measuring devices to assign coordinates to different events such that the coordinates two observers assign to the same event will transform according to the Lorentz transform, but that it is impossible to build measuring devices in windowless boxes whose coordinates transform according to the ether transform?
Two observers in two different windowless boxes each assigning coordinates to an event that is not within either box...I'm not sure exactly what you're scenario is, it could be anything. I'm looking for an explicit example like: "Two windowless boxes A and B each contain an observer who manufactures his own measuring devices to assign coordinates to all of spacetime including the other windowless box, so that the two coordinate systems overlap, and this is possible using the Lorentz transform, but it is not possible using the LET transform unless both observers can somehow sense their velocity with respect to a preferred frame (or this information is provided to them by a third party). A puts coordinates x=1meter t=1second on an event that B assigns coordinates x=1E8 meters and t=1E10 seconds. Windows form in both windowless boxes, and A and B both measure their velocity realtive to the other to be v=1E6 m/s, and their distance is x_B-x_A=1E12 meters apart (motion and distances taken to be along the x-axis). Something like that. It seems to me that the two coordinate systems can't be reconciled without knowing v. Nothing happens when the windows form in the boxes? A third observer C from outside measured the relative velocities of the A and B and uses the Lorentz transform to compare events within the coordinate systems of A and B. Is that example as good as any other for you? Implicit in the coordinate systems of both A and B is the assumtion of v=0 for the observer, so let's use that same assumption for two LET observers Alpha and Beta. They will construct exactly the same coordinate systems as A and B. The only difference is that they will compute $t=T/\gamma+f(x,v)$ where $f(x,v)=-vx/c_0^2$ is the synchronization of the clock at x, but the LorentzTransformers will compute $t=T/\gamma-vx/c_0^2+f(x,v)$ where $f(x,v)=0$. There is no difference, and that's no accident; it is by design. That's why I'm quite confident that all of this talk about LET and SR not being empirically equivalent is missing the simple point that they ARE. My point is "look look, there is no real difference", it is only an illusion created by a choice of clock synchronization convention. The speed of light isn't really something that has been measured to be constant, and the relativity of simultaneity isn't some magical truth about physical reality; it's a consequence of one's choice of clock synchronization convention.

JesseM said:
As for the issue of whether speed depends on your coordinate system, I already provided a numerical example. Again, assume that when I use the symbol "c" it refers only to the speed of light as measured in the preferred ether frame, even if you would use the symbol differently.
OK, I'll go back and look at that.

Last edited:
Aether said:
Not necessarily, but there is an ambiguity with specifying a velocity as 0.8c in LET because that is a function of v and \theta.
Again, not if you use "c" the way that I said I was using it.
Aether said:
It may not make any difference in the end. But you said that the outside observer is the one who computes the Lorentz transform, and that's getting way more complicated than just a windowless box.
But it doesn't ultimately matter whether anyone "computes" the Lorentz transform, it will still be true that the coordinates the two observers assign to a given event will be related by the Lorentz transform, even if no one notices this fact.
Aether said:
]Two observers in two different windowless boxes each assigning coordinates to an event that is not within either box...
It's pretty common in discussions of SR to talk about rulers sliding arbitrarily close to each other with clocks placed at each marking, and then you can ask questions like "when the clock at the 5-meter mark of the first ruler reads 10 seconds, what marking on the second ruler is next to it at that moment, and what does the clock at that mark read"? So you can imagine something like that, with arbitarily thin boxes sliding alongside each other, and the "events" just being a clock on a marking on a ruler in one box passing arbitrarily close to a clock on a marking on a ruler in another box. Alternately, you could imagine large ghostly boxes that can pass through each other without colliding, and then the ruler markings/clocks in different boxes could occupy the same point in space. It doesn't really matter, these sorts of details aren't important since it's just a thought-experiment.
Aether said:
I'm not sure exactly what you're scenario is, it could be anything. I'm looking for an explicit example like: "Two windowless boxes A and B each contain an observer who manufactures his own measuring devices to assign coordinates to all of spacetime including the other windowless box, so that the two coordinate systems overlap
Yes, this will do fine.
Aether said:
and this is possible using the Lorentz transform
"This is possible using the Lorentz transform" is ambiguous, it's not like the observers use the Lorentz transform when constructing their measuring devices, it's just that once the devices are completed, it will be true that the coordinates that different observers assign to the same point in spacetime will be related by the Lorentz transformation.
Aether said:
but it is not possible using the LET transform unless both observers can somehow sense their velocity with respect to a preferred frame (or this information is provided to them by a third party).
Right, unless the observers have this knowledge of their velocity relative to the preferred frame, there's no technique they can use in constructing the measuring devices that will insure that the coordinates that different observers assign to given points in spacetime will be related by the LET transform.
Aether said:
A puts coordinates x=1meter t=1second on an event that B assigns coordinates x=1E8 meters and t=1E10 seconds. Windows form in both windowless boxes, and A and B both measure their velocity realtive to the other to be v=1E6 m/s, and their distance is x_B-x_A=1E12 meters apart (motion and distances taken to be along the x-axis). Something like that. It seems to me that the two coordinate systems can't be reconciled without knowing v. Nothing happens when the windows form in the boxes?
No, and like I said it's not even necessary for windows to form, it's still a true fact about nature that their coordinates are related by the Lorentz transform even if none of them are able to verify this.
Aether said:
A third observer C from outside measured the relative velocities of the A and B and uses the Lorentz transform to compare events within the coordinate systems of A and B. Is that example as good as any other for you? Implicit in the coordinate systems of both A and B is the assumtion of v=0 for the observer
Huh? v=0 relative to what? And are you talking about the observer who notes the coordinates that A and B assign to a particular event? How could it possibly matter what his velocity is? Questions about whether two events take place at a single point in spacetime or at different points in spacetime must have a single objective answer that is the same for all observers regardless of their velocity--otherwise you'd have different reference frames making different predictions about objective physical events like whether two asteroids will collide or miss each other! So if one observer says "at the moment the clock on the 12-meter mark of ruler A ticked 13 seconds, it occupied the same position as the clock at the 15-meter mark of ruler B which at that moment ticked 8 seconds", then every observer, regardless of velocity, regardless of what coordinate system they're using, must agree on this fact.
Aether said:
so let's use that same assumption for two LET observers Alpha and Beta. They will construct exactly the same coordinate systems as A and B. The only difference is that they will compute $t=T/\gamma+f(x,v)$
But they can't compute this if they are constructing their measuring devices in windowless boxes!
Aether said:
That's why I'm quite confident that all of this talk about LET and SR not being empirically equivalent is missing the simple point that they ARE.
Aether, you seem to be very confused about the meaning of "empirically equivalent" and also your other phrase "mathematically equivalent"--the way you are using these phrases seems completely incoherent, and I think you are badly misunderstanding what people like Mansouri and Sexl mean when they say the LET is empirically equivalent to SR. It's clear from the paper you sent me that they are not talking just about a different coordinate system, but a theory with some actual physical assumptions that are different from those of SR, namely the assumption of a physical substance called ether which has its own rest frame and which causes rulers to shrink and clocks to slow down when they move relative to it. To say this theory is "empirically equivalent" to SR is not to make any statements about the coordinate systems being equivalent or the measuring devices used to assign these coordinates being equivalent, it's just to note that the theory doesn't make any physical predictions which are different from those of SR. Do you understand what the difference is between an actual physical prediction and a statement which depends on your coordinate system? A lot of your previous statements, including the one I discussed earlier in this post, suggest you're pretty fuzzy on this point. Keep in mind that both the SR coordinate systems and the LET coordinate systems can be used to analyze the functioning of a set of physical measuring devices of either type--you can use the coordinate systems allowed by the LET transform to analyze the physical situation of observers who synchronize their clocks using the Einstein synchronization convention, and you will correctly predict that the matchup between physical ruler-markings and clocks of different observers will be the same as that given by the Lorentz transform; likewise, you can use the coordinate systems allowed by the Lorentz transform to analyze the physical situation of observers who all synchronize their clocks to a certain preferred frame, and you will correctly predict that the matchup between physical ruler-markings and clocks of different observers will be the one given by the LET transform. You can analyze either physical situation from the point of view of LET or SR, and you will get the same physical predictions regardless of what coordinate system you use, showing that the two theories are empirically equivalent; but the two physical situations (and the two sets of physical measuring devices they involve) themselves are different, they are not "equivalent" in any way.

Mansouri and Sexl themselves seem to understand the fact that the physical construction of SR coordinates can be done without any information exchange between observers, but the physical construction of LET coordinates cannot. That's what I think they mean when they distinguish between "system-internal synchronization" and "system-external synchronization" on pp. 499-500:
Both the Einstein procedure and the transportation-synchronization will be called system-internal synchronization. There are other such procedures, such as shaft synchronization [23-26], and the problem to be solved here is the equivalence of the various synchronization procedures. This problem will be solved in part in this paper.

System-internal methods of synchronization are not the only conceivable ones. In section 3 we shall discuss in detail an alternative procedure belonging to the class of system-external synchronization methods. Here one system of reference is singled out ("the ether system") and clocks in all systems are synchronized by comparing them with standard clocks in the preferred system of reference. Infinitely many inequivalent system-external procedures are possible. Among these, one is of special interest: A convention about clock synchronization can be chosen that does maintain absolute simultaneity. Based on this convention an ether theory can be constructed that is, as far as kinematics is concerned (dynamics will be studied in a later paper in this series) equivalent to special relativity. In this theory measuring rods show the standard Fitzgerald-Lorentz contraction and clocks the standard time dilation when moving relative to the ether. Such a theory would have been the logical consequence of the development along the lines of Lorentz-Larmor-Poincaré. That the actual development went along different lines was due to the fact that "local time" was introduced at the early stage in considering the covariance of the Maxwell equations.
This quote also shows, as I said before, that Mansouri and Sexl are considering a theory which actually involves different physical assumptions than SR--the existence of an ether--rather than just a different coordinate system for analyzing the same physical theory. And again, to say this is "empirically equivalent" is just to note that this additional assumption doesn't lead to any new predictions about coordinate-invariant physical facts--if you analyze a given physical scenario, you'll get the exact same physical predictions.

Last edited:
pmb_phy
This thread has certainly run on for a while.

Aether - You wrote a lot of things and then implied they were valid definitions that came from somewhere. Is the source of these definitions Aether himself?

To be honest I'm not 100% sure of what this thing is that you call Lorentz theory.

You claimed
..are empirically equivalent systems for interpreting local Lorentz symmetry.
You'll have to explain to me what Lorentz theory" is before I can address that appropriately. However, that said, if "Lorentz theory" is what I think it is (MMX experiment null results) then this Lorentz theory fails at attempting to describe all phenomena. If it fails at anything then its a bad theory. Nobody cares that it can work for a subset of observerations (Lorentz symmetry? What is that an observation of?)

You also stated the following
Why isn't "relativity without the aether" fairly described by the term "pseudoscience"?
Just because theory A does not assume the existence of "stuff Q" it doesn't make the theory A a pseudoscience. I see no logic which would allow one to make this jump. I see it being quite possible to have an ether as it was originally defined (that which fills all of space). It was just and then hijacked by EM later on as "ether = the medium through which EM waves propagate." But with no ether required there is no reason to assume that it doesn't exist. Such a jump in logic would be pseudoscience. However if one insists that it stays in relativity because it is required then it violates Occam's razor.
Jesseme said:
Unlike the ether, though, the sea of virtual particles is not thought to have its own natural rest frame, so it doesn't violate Lorentz symmetry even in an unobserved way.
Did you, for some reason, assume that I thought that there was a rest frame for this sea? If so then how did you reach the conclusion that this is what I assumed? I do not assume that. I never did in fact. In fact I like this example because of the fact that there is no natural rest frame for such a sea of virtual particles.

Pete

Last edited:
pmb_phy said:
Did you, for some reason, assume that I thought that there was a rest frame for this sea? If so then how did you reach the conclusion that this is what I assumed? I do not assume that. I never did in fact. In fact I like this example because of the fact that there is no natural rest frame for such a sea of virtual particles.
No, I didn't assume you thought that, but since you compared this sea to the ether I thought it was important to point out this very important difference between them, so no one would get confused and think that the virtual particle sea lends support to the ether-based idea of a preferred frame.

Gold Member
pmb_phy said:
Aether - You wrote a lot of things and then implied they were valid definitions that came from somewhere. Is the source of these definitions Aether himself?

To be honest I'm not 100% sure of what this thing is that you call Lorentz theory.

You claimed
You'll have to explain to me what Lorentz theory" is before I can address that appropriately. However, that said, if "Lorentz theory" is what I think it is (MMX experiment null results) then this Lorentz theory fails at attempting to describe all phenomena. If it fails at anything then its a bad theory. Nobody cares that it can work for a subset of observerations
Mansouri-Sexl define an ether theory (as a popular test theory of local Lorentz invariance), and I'm working from that. My questions, interpretations of what they are saying, and attempts to work through the math are coming from me. I expect half of that to get swept away, but Mansouri-Sexl should be reliable. MMX experiments verify the rotation invariance component of Lorentz symmetry, and that sweeps away some ether theories, but LET is what is left after all of that. It fails at nothing.

pmb_phy said:
(Lorentz symmetry? What is that an observation of?)
That's what most people think that SR is an observation of. I've been collecting lots of papers on experiments to test local Lorentz invariance, and I don't recall finding statements in any of them saying "and this proves SR once again" because they are scientific papers, and (I am beginning to suspect) SR is merely pseudoscientific.

pmb_phy said:
Just because theory A does not assume the existence of "stuff Q" it doesn't make the theory A a pseudoscience. I see no logic which would allow one to make this jump. I see it being quite possible to have an ether as it was originally defined (that which fills all of space). It was just and then hijacked by EM later on as "ether = the medium through which EM waves propagate." But with no ether required there is no reason to assume that it doesn't exist. Such a jump in logic would be pseudoscience.
Coordinate independent geometry is scientific, but SR and LET are empirically equivalent coordinate systems which, when taken separately, typically lead people to make false claims such as "the constancy of the speed of light is proven by experiments". The speed of light is constant in SR, but it is variable in LET.

pmb_phy said:
However if one insists that it stays in relativity because it is required then it violates Occam's razor
I don't think that it is required for doing engineering work so long as a locally preferred frame remains undetectable, but where its absence leads to wrong answers then it should be restored (or SR and LET can both be thrown out, and coordinate independent geometry will be what remains).

p.s. If I'm wrong about any of this, then I expect you guys/girls to beat it out of me. That's why I come here, because I have faith in you.

Last edited:
pmb_phy
Aether said:
Mansouri-Sexl define an ether theory (as a popular test theory of local Lorentz invariance), and I'm working from that. My questions, interpretations of what they are saying, and attempts to work through the math are coming from me. I expect half of that to get swept away, but Mansouri-Sexl should be reliable. You can download their papers here for the next few hours (8MB): http://69.13.172.13/Mansouri&Sexl.pdf. MMX experiments verify the rotation invariance component of Lorentz symmetry, and that sweeps away some ether theories, but LET is what is left after all of that. It fails at nothing.[/qupte]Are you saying that you can't tell me what this "Lorentz Ether Theory" is in one paragraph??
That's what most people think that SR is an observation of...
I wasn't suggesting anything by that question since I've yet to know what this theory you're speaking of is.
Coordinate independent geometry is scientific, but SR and LET are empirically equivalent coordinate systems which, when taken separately, typically lead people to make false claims such as "the constancy of the speed of light is proven by experiments". The speed of light is constant in SR, but it is variable in LET.
That is the most incorrect statement that I've seen posted on the internent in months. Where did you get this idea from?
I don't think that it is required for doing engineering work ...
Occam's razor is more of a philosophy than anything else and it is not always clear how it should be taken in a given application. That's why you'll never see me invoke it.
p.s. If I'm wrong about any of this, then I expect you guys/girls to beat it out of me. That's why I come here, because I have faith in you.
That is a false assumption. It takes years of scientific training to learn to become a scientist. Sometimes people will come here and get hung up or confused on the most simplest of points and assume that there is something wrong with science today because they have "bewildered" us in some way. Assuming that is wrong because some people just don't care enough to get a point through to you. There could be a million false assumptions in that person's mind that we are just too tired to want to weed through when there are just so many better ways to spend our times.

E.g. This is not a topic I'd want to continue with. I'm into something much more interesting to me. I.e. the language of the Hopi Indian's has no word for time, i.e. it has been said by an expert in linguistics that --
In particular, he has no general notion or intuition of time as a smooth flowing continuum in which everything proceeds at an equal rate, out of the future, through a present, into a past; or, in which, to reverse the picture, the observer is being carried in the stream of duration continuosly away from a past and into a future.
So I'd love to spend all my time now on trying to answer this question "How does one convert a special relativity text from English to the Hopi language?".

Pete

Pete

Gold Member
pmb_phy said:
That is the most incorrect statement that I've seen posted on the internent in months. Where did you get this idea from?
The idea that the constancy of the speed of light can't be proven by experiment?

pmb_phy said:
That is a false assumption. It takes years of scientific training to learn to become a scientist. Sometimes people will come here and get hung up or confused on the most simplest of points and assume that there is something wrong with science today because they have "bewildered" us in some way. Assuming that is wrong because some people just don't care enough to get a point through to you. There could be a million false assumptions in that person's mind that we are just too tired to want to weed through when there are just so many better ways to spend our times.

E.g. This is not a topic I'd want to continue with. I'm into something much more interesting to me. I.e. the language of the Hopi Indian's has no word for time, i.e. it has been said by an expert in linguistics that --
So I'd love to spend all my time now on trying to answer this question "How does one convert a special relativity text from English to the Hopi language?".
OK, but I didn't say that there was anything wrong with science, only with relativity.

pmb_phy said:
I wasn't suggesting anything by that question since I've yet to know what this theory you're speaking of is.
Aether seems not to be describing an alternate theory at all, but just a different set of coordinate systems for describing a spacetime which obeys exactly the same laws as the one in SR. As in relativity, each observer can assign coordinates to events using a network of rulers and clocks, but instead of each observer synchronizing their clocks using the assumption that light travels at c in their own rest frame, only one observer synchronizes his clocks this way, and all other observers synchronize their clocks in such a way that their definition of simultaneity agrees with that preferred observer. If x and t are the coordinates assigned to an event by the preferred observer, then another observer moving at v along his access will assign the same event coordinates x' and t', with the coordinates related by the following "LET transformation":

$$x' = (x - vt)/\sqrt{1 - v^2/c^2}$$
$$t' = t \sqrt{1 - v^2/c^2}$$

You can compare this with the Lorentz transformation:

$$x' = (x - vt)/\sqrt{1 - v^2/c^2}$$
$$t' = (t - vx/c^2) / \sqrt{1 - v^2/c^2}$$

Most people who use the term "Lorentz ether theory" would define the theory as saying there's an actual physical substance called "ether" and that the preferred observer should be at rest with respect to this ether, but Aether seems not to think this assumption is important, so he isn't making any new physical assumptions at all, he's just using a different set of coordinate systems. You can see, though, that if there was such a thing as ether, and all rulers moving relative to that ether shrunk by $$\sqrt{1 - v^2/c^2}$$ while all clocks moving relative to that ether had their ticks extended by $$1/\sqrt{1 - v^2/c^2}$$, then if all observers synchronized their clocks using the Einstein synchronization procedure, different observers' coordinate systems would be related by the Lorentz transform and there'd be no way to actually detect which frame was the ether's rest frame, so such a universe would be empirically equivalent to one where there is no ether but the laws of physics exhibit Lorentz-symmetry. I elaborated on this empirical equivalence a little more in the first two paragraphs of post #36.

Last edited:
Gold Member
JesseM said:
Aether seems not to be describing an alternate theory at all, but just a different set of coordinate systems for describing a spacetime which obeys exactly the same laws as the one in SR. As in relativity, each observer can assign coordinates to events using a network of rulers and clocks, but instead of each observer synchronizing their clocks using the assumption that light travels at c in their own rest frame, only one observer synchronizes his clocks this way, and all other observers synchronize their clocks in such a way that their definition of simultaneity agrees with that preferred observer. If x and t are the coordinates assigned to an event by the preferred observer, then another observer moving at v along his access will assign the same event coordinates x' and t', with the coordinates related by the following "LET transformation":

$$x' = (x - vt)/\sqrt{1 - v^2/c^2}$$
$$t' = t \sqrt{1 - v^2/c^2}$$

You can compare this with the Lorentz transformation:

$$x' = (x - vt)/\sqrt{1 - v^2/c^2}$$
$$t' = (t - vx/c^2) / \sqrt{1 - v^2/c^2}$$

Most people who use the term "Lorentz ether theory" would define the theory as saying there's an actual physical substance called "ether" and that the preferred observer should be at rest with respect to this ether, but Aether seems not to think this assumption is important, so he isn't making any new physical assumptions at all, he's just using a different set of coordinate systems. You can see, though, that if there was such a thing as ether, and all rulers moving relative to that ether shrunk by $$\sqrt{1 - v^2/c^2}$$ while all clocks moving relative to that ether had their ticks extended by $$1/\sqrt{1 - v^2/c^2}$$, then if all observers synchronized their clocks using the Einstein synchronization procedure, different observers' coordinate systems would be related by the Lorentz transform and there'd be no way to actually detect which frame was the ether's rest frame, so such a universe would be empirically equivalent to one there is no ether but the laws of physics exhibit Lorentz-symmetry. I elaborated on this empirical equivalence a little more in the first two paragraphs of post #36.
That seems to be a fair summary, thanks.

pmb_phy
JesseM said:
... but instead of each observer synchronizing their clocks using the assumption that light travels at c in their own rest frame, only one observer synchronizes his clocks this way, and all other observers synchronize their clocks in such a way that their definition of simultaneity agrees with that preferred observer. ...
That's nuts! There is no meaning to synchronizing one clock! I knew there was a good reason I dropped out of this thread!

Pete

JesseM said:
... but instead of each observer synchronizing their clocks using the assumption that light travels at c in their own rest frame, only one observer synchronizes his clocks this way, and all other observers synchronize their clocks in such a way that their definition of simultaneity agrees with that preferred observer. ...
pmb_phy said:
That's nuts! There is no meaning to synchronizing one clock! I knew there was a good reason I dropped out of this thread!

Pete
What are you talking about? I said clocks plural, in all 3 instances that you quoted above. The whole basis of SR is the idea of each observer synchronizing their clocks using light signals, it's right there in Einstein's original 1905 paper.

Gold Member
pmb_phy said:
That is the most incorrect statement that I've seen posted on the internent in months. Where did you get this idea from?
I would like to see you try and refute that. However, the main purpose for this post is to ask that you delete the link to my website from your post. I only intended for that to remain visible long enough for you to download the paper if you were interested.

Gold Member
Hurkyl said:
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.
This sounds like a potentially convincing argument, Hurkyl. I have never seen SR described that way anywhere else though; do you know of a source that teaches SR from that perspective? My objection to "SR without the aether" may arise entirely from the difference between what you wrote and the "two postulates" of SR. Am I likely to find statements like "experiments prove that the speed of light is a constant" in such a text?

Last edited:
Gold Member
DrGreg said:
If you assume SR is correct, you can come up with some expressions for momentum and energy relative to LET co-ordinates by application of the transform equations. But if you ignore SR and try to calculate momentum and energy some other way, how do you find the answer? How, for example, do you prove that, relative to the ether, the relativistic form $p=\gamma_u m u$ should be used instead of $p= m u$?
I have today at long last been able to view of copy of the Mansouri-Sexl papers. The conclusion of paper I is

"A theory maintaining the concept of absolute simultaneity can be obtained ... which is ... empirically equivalent to special relativity, as least as far as kinematics is concerned." (My emphasis)

I think that supports my point about the difference between kinematics and dynamics.

They also make the point that Einsteinian clock synchronization is the same as synchronization via ultraslow clock transport, in the context of SR and what we have been calling "LET".

By the way, I'm no longer sure whether it is correct to describe their transform as "Lorentz Ether Theory" -- there may be several competing ether theories in circulation.

I think I will have more to say once I've read all three papers in more detail.

Last edited:
Staff Emeritus
Gold Member
I have never seen SR described that way anywhere else though; do you know of a source that teaches SR from that perspective?

No -- my perspective is pretty much a synthesis of everything I've leared (from a wide variety of sources), and my working out details on my own as well.

There has to be a textbook on Minowski geometry around someplace, though -- I don't know of any, though.

pmb_phy
Aether said:
I would like to see you try and refute that. However, the main purpose for this post is to ask that you delete the link to my website from your post. I only intended for that to remain visible long enough for you to download the paper if you were interested.
At the time I wrote that assertion it was at a time when I made a false assumption in that the term "LET" of yours meant something other than what I assuimed it meant. I therefore retract my assertion since I don't want to read that article on LET that you linked to. It may be more widely used than that paper

When you explained to me what you meant by LET I then chose to bow out of this conversation since I'm not in the mood for looking into what appears to be bad physics. Please don't try to analyze my statement here because I have nothing against the notion of looking more deeply into what "appears" to be wrong, since it could very well be right. But I have to make choices on how I spend my time. The problem is sitting in this chair with my back in so much pain. If you're curious as to what my back looks like after the removal of the herniated disk then see the photo in the first post at - http://ubb-lls.leukemia-lymphoma.org/ubb/Forum14/HTML/001089.html

I still do physics now but only in those areas I love. At this time I have zero interest into looking into this LET thing. Hence the reason I bowed out of this conversation. I only came back because I neglected to state that because something is not science it doesn't mean that it should be called pseudoscience. Religion is not science but it'd be wrong to call religion pseudoscience. A wrong theory is not pseudoscience simplyt because its wrong. It must satisfy other criteria which lies beyond the science itself and lies within the minds of the holders of the theory. Its a weird thing and too drenched in debateable terminology to want to get into. There's a book called "Science and Unreason" by Radner and Radner. It was required reading in my philosophy of science course in college. I highly recommend this text for those who wish to learn what pseudoscience is.

As far as removing that link - No can do. First off you shouldn't post in open forum something which you don't want to keep there permenently. It requires me to do things like this where you want me to dig through old posts and delete something I quoted. I did look into removing the link but its too late. A post can only be edited for a day or two.

Pete

Last edited by a moderator:
Gold Member
To anyone who is interested: I have found a one-page summary of the Mansouri-Sexl framework that Aether refers to here: http://relativity.livingreviews.org/Articles/lrr-2005-5/articlesu9.html [Broken], on the "Living Reviews in Relativity" website. (The notation used in the summary is not quite the same as the notation that M&S use themselves.)

The framework is to interpret the results of experiments testing the accuracy of Special Relativity. It assumes there is at least one frame (the "ether frame") in which the speed of light is isotropic and takes the view that the transformation to other frames is an unknown linear transform, with velocity-dependent coefficients that are to be determined experimentally. The results of experiments that had been performed before publication in 1977 are then analysed to determine how close the coefficients must be to the Lorentz transform.

In the course of their analysis, M&S make the point that there is an arbitrary choice of clock synchronization to be made. Their method effectively ignores any effects that are due to the choice of synchronization.

The particular transformation that Aether has been discussing in this forum (which we have been describing as "LET") is one that can be implemented as "synchronization to the ether" (if you have chosen an ether) but which is, essentially, mathematically equivalent to Special Relativity. This is really the point - the two formulations come from different sets of assumptions but come to essentially the same conclusion, in the sense that one formulation can be mathematically transformed into the other. The two "theories" stand and fall together - they're either both true or both false.

If Aether is hoping to find something to favour an ether theory over SR, it would have to take the more generalised form discussed by Mansouri-Sexl rather than the particular form that has been quoted.

References:

Mansouri, R., and Sexl, R.U., “A test theory of special relativity. I - Simultaneity and clock synchronization”, Gen. Relativ. Gravit., 8, 497-513, (1977).

Mansouri, R., and Sexl, R.U., “A test theory of special relativity. II - First Order Tests”, Gen. Relativ. Gravit., 8, 515-524, (1977).

Mansouri, R., and Sexl, R.U., “A test theory of special relativity. III - Second Order Tests”, Gen. Relativ. Gravit., 8, 809-814, (1977).

Bluhm, R., "Breaking Lorentz Symmetry", Physics World, March 2004.

Last edited by a moderator:
Gold Member
pmb_phy said:
...I'm not in the mood for looking into what appears to be bad physics. Please don't try to analyze my statement here because I have nothing against the notion of looking more deeply into what "appears" to be wrong, since it could very well be right. But I have to make choices on how I spend my time. The problem is sitting in this chair with my back in so much pain.
OK, I hope to run into you again sometime on a topic that is of more interest to you. Best wishes on a speedy recovery.

Juan R.
Aether said:
Special relativity (SR) SR and Lorentz ether theory (LET) are empirically equivalent systems for interpreting local Lorentz symmetry. These two theories are equally valid, but it is not possible (so far) to demonstrate the truth or falseness of the postulates of either theory over the other by experimentation. Still, a superstition persists in the minds of many that somehow "SR is true, and LET is false". Why isn't "relativity without the aether" fairly described by the term "pseudoscience"?

pseudoscience - Refers to anybody of knowledge or practice which purports to be scientific or supported by science but which is judged by the mainstream scientific community to fail to comply with the scientific method.

scientific method n - The principles and empirical processes of discovery and demonstration considered characteristic of or necessary for scientific investigation, generally involving the observation of phenomena, the formulation of a hypothesis concerning the phenomena, experimentation to demonstrate the truth or falseness of the hypothesis, and a conclusion that validates or modifies the hypothesis.

su·per·sti·tion n An irrational belief that an object, action, or circumstance not logically related to a course of events influences its outcome.

1) A belief, practice, or rite irrationally maintained by ignorance of the laws of nature or by faith in magic or chance.

2) A fearful or abject state of mind resulting from such ignorance or irrationality.

3) Idolatry.

I do not argue in favor of or against the aether. In fact, i am a bit confused about the need or not of an aether. Simply i will cite a bit of history that has great physical consequences but that is ignored in typical textbooks.

In a letter to Lorentz dated 17 June 1916, Einstein wrote:

I agree with you that the general relativity theory admits of an ether hypothesis as does the special relativity theory.

According to Galina Granek, in 1920 at a lecture in Leiden [Einstein, A. (1920) Äther und Relativitätstheorie. Lecture Presented on 5th May, 1920 in the University of Leyden (Berlin: Springer).], Einstein explained why a revised notion of the ETHER was required in physics. He REPEATED Poincare's claims of 1900, according to which AETHER is required in order that movements do not take place with respect to empty space.

does the Aether exist?

I only know that MM experiments and similar cannot detect Poincaré aether (which is not the same that the Lorentz aether)?

Last edited:
derz
The mathematics of SR work without the assumption of an aether, so why would it be pseudoscience?

I believe that with an "aether" Einstein meant the dynamic spacetime of GR that has properties hidden from us, i.e. geometry.

Jimmy Snyder
Aether said:
scientific method n - The principles and empirical processes of discovery and demonstration considered characteristic of or necessary for scientific investigation, generally involving the observation of phenomena, the formulation of a hypothesis concerning the phenomena, experimentation to demonstrate the truth or falseness of the hypothesis, and a conclusion that validates or modifies the hypothesis.

Even though I attributed this quote to Aether, I assume that it was meant to be a quote from some other source. I don't buy this description of the scientific method. It is overly enthusiastic concerning the ability to demonstrate truth. My view conforms to that of this site:

http://www.philosophypages.com/lg/e15.htm

Here is a quote from that site:

philosophypages said:
Although it always remains impossible in principle to prove the truth of a scientific hypothesis, it is possible to compare the distinct hypotheses involved in rival explanations of the same event.

I think discussions of the relative merits of differing hypotheses would go smoother if this last quote were kept in mind. Calling your opponent supersticious just because they don't share your own certainty seems counterproductive to me.

Gold Member
derz said:
The mathematics of SR work without the assumption of an aether, so why would it be pseudoscience?

I believe that with an "aether" Einstein meant the dynamic spacetime of GR that has properties hidden from us, i.e. geometry.
It is not valid to claim that the constancy of the speed of light is proven "by experiment", or that "experiments prove" that SR is right and that LET is wrong. The "Consistency of the speed of light" thread is where examples of such claims can be seen, and where my question of "pseudoscience" comes from. Take note of where the speed of light is claimed to be constant in other than an "inertial reference frame" (some people say "all frames of reference" for example, which is wrong). Also take note where people claim that "experiments prove" that the speed of light is a constant.

I am concerned by such claims because they appear to be flat wrong, yet just about everyone else in the "constancy of the speed of light" thread seemed to agree (at that time at least) that these claims were clearly right and that the claims that I was making to the contrary (from Mansouri-Sexl) were clearly wrong.

SR is valid only within the context of intertial reference systems. To establish an inertial reference system one must start out by synchronizing all of the clocks at rest in the system so that experiments will measure the same speed of light in all directions. This particular clock synchronization convention is where the relativity of simultaneity comes from. There are other equally valid ways to syncrhonize clocks; in particular the clock syncrhonization of LET maintains absolute simultaneity.

Relativity per se is not pseudoscience when it is kept within the context of inertial reference systems.

Last edited: