Relativity of simultaneity and one-way speed of light

[Andromacus]

Since the distant clock is at rest it means there is no motion, which I guess technically is a state of motion, but one that is time-independent.

Exactly. A time-independent rest frame is clearly an absolute(rest) motion or preferred frame So Einstein's synchronization is just a condition to use the constancy of c as premise of the transformations , starting from the time-independent galilean inertial coordinates and transforming to another frame keeping c constant wich logically implies changing lengths and times in accordance.

Of course in the new frame one can once again use the equivalence between inertial frames to Einstein synchronize again, since an observer in that new frame can consider himself at rest and all distant clocks can be assumed at rest with him again.

It is hard not to see here that the second postulate is basically used to change length and time units(so that c remains constant after the transformation), and the Einstein synchronization is just readjusting to the original time-independent or absolute units.

 

Basically making c constant in all frames makes gallilean trnasformations equivalent to Lorentz transformations in the caseof light. It doesn¡t look a very physical way to keep the theory formally contradiction free.

[swansont]

Exactly. A time-independent rest frame is clearly an absolute(rest) motion or preferred frame

 

It's not at all clear to me.

It is hard not to see here that the second postulate is basically used to change length and time units(so that c remains constant after the transformation), and the Einstein synchronization is just readjusting to the original time-independent or absolute units.

 

But if you recover the clock synchronization in the new frame, then the original frame can't be absolute. The moving frame has the same relation: any spatial interval you choose is a constant with (and therefore independent of) time.

Basically making c constant in all frames makes gallilean trnasformations equivalent to Lorentz transformations in the caseof light. It doesn¡t look a very physical way to keep the theory formally contradiction free.

 

How did Lorentz-transforming light get into the conversation?

[Andromacus]

 

It's not at all clear to me.

Ok, strike the word preferred, it just follows galilean relativity. But exchanging absolute time and velocities for an absolute limiting speed c of light signals that define transformed times and lengths

But if you recover the clock synchronization in the new frame, then the original frame can't be absolute. The moving frame has the same relation: any spatial interval you choose is a constant with (and therefore independent of) time.

Yes, as I corrected above the frame follows galilean relativity and in that sense is not absolute. But I do not think I wrote literally "absolute frame" anywhere. I did say absolute motion, in reference to the fact that in newtonian physics velocity is considered absolute, but it also admited galilean relativity.

How did Lorentz-transforming light get into the conversation?

Well, I have been making the distinctin between the nul interval transformation and timelike intervals, But you are right that actually my point refers to the Lorentz transformation in general, t's only that it is more obvious the similarity to the galilean transformation for light since it doesn't involve proper time but arbitrary affine parameters for time given the time-independence. In the case of massive particles, there is a nominal change of units for time and length to keep c constant.

Edited September 14, 2015 by Andromacus

[swansont]

Ok, strike the word preferred, it just follows galilean relativity. But exchanging absolute time and velocities for an absolute limiting speed c of light signals that define transformed times and lengths

 

An invariant c does not define an absolute frame. I don't see what your objection is.

 

But I do not think I wrote literally "absolute frame" anywhere.

 

No, you wrote "A time-independent rest frame is clearly an absolute(rest) motion or preferred frame" I chose the wrong word to copy. Does that have that much of an effect on the argument? There is no absolute motion or preferred frame involved here.

 

I have been making the distinctin between the nul interval transformation and timelike intervals, But you are right that actually my point refers to the Lorentz transformation in general, t's only that it is more obvious the similarity to the galilean transformation for light since it doesn't involve proper time but arbitrary affine parameters for time given the time-independence. In the case of massive particles, there is a nominal change of units for time and length to keep c constant.

 

What is a null transformation? What is a Galilean transformation for light? (The follow up may be "what's the relevance", depending on the answer)

 

"In the case of massive particles" covers 100% of what the Lorentz transform addresses, and technically the units are not changed, it is the magnitude of the values that change as a result of c being invariant.

[Andromacus]

 

 

 

No, you wrote "A time-independent rest frame is clearly an absolute(rest) motion or preferred frame" I chose the wrong word to copy. Does that have that much of an effect on the argument? There is no absolute motion or preferred frame involved here.

Agreed.

What is a null transformation? What is a Galilean transformation for light? (The follow up may be "what's the relevance", depending on the answer)

Not a null transformation, but a Lorentz transformation of an EM wave that leaves the null interval c2t2-x2-y2-z2=0 invariant. This is the first step in many SR books when deriving the LT.

Edited: Actually the idea of such a galilean transformation is quite irrelevant to the discussion.

 

"In the case of massive particles" covers 100% of what the Lorentz transform addresses,

Actually the LT can be applied to photons in two inertial frames K and K' according to Einstein's "Relativity:the special and the general" and these are the corresponding transformations, they follow simply from substituting x by ct in the usual LTs

 

and technically the units are not changed, it is the magnitude of the values that change as a result of c being invariant.

 

This is debatable. physically it is usually argued as you say, but formally it is clear by looking at the transformations that all there is is a rescaling of the time and length units by contrancting the latter and dilating the former. Geometrically the "squeeze mapping" (see wikipedia about it) is the equivalent of a rotation but using hyperbolic or imaginary angles, it is a linear transformation that preserves the areas formed by the x and t coordinates(hyperbolic sectors) in a LT. In other words one uses c as measuring unit and change x and t units accordingly(compensating one with the other:x contracts what t dilates), see picture taken from wikipedia.

Edited September 15, 2015 by Andromacus

[Mordred]

As this is relevant, though not to the specific conversation currently under way. Rather the earlier conversion on one way speed of light.

 

Here is a recent article discussing the precision of one of the latest tests

 

https://briankoberlein.com/2015/09/15/relativity-wins-again/

Full article is here, as well as Nature

 

http://arxiv.org/abs/1412.6954

[swansont]

Not a null transformation, but a Lorentz transformation of an EM wave that leaves the null interval c2t2-x2-y2-z2=0 invariant. This is the first step in many SR books when deriving the LT.

Do you have a link or two? I don't recall ever seeing that.

 

"In the case of massive particles" covers 100% of what the Lorentz transform addresses,

Actually the LT can be applied to photons in two inertial frames K and K' according to Einstein's "Relativity:the special and the general" and these are the corresponding transformations, they follow simply from substituting x by ct in the usual LTs

Are you talking about chapter 11? That's not what he's doing. He specifically states "the law of the transmission of light in vacuo is satisfied both for the reference-body K and for the reference-body K1" (bodies — these have mass) The frames in question are inertial. He's transforming a distance, which is perfectly valid. He never mentions a frame for light. It is not a transformation for photons. The LT is a transform from one inertial frame (valid only for massive bodies where we can say they are at rest) to another.

This is debatable.

 

Again, technically true, because you can debate anything (1+1=2? That's debatable), but the units of time actually remain units of time. If you are willing to debate this, then it's pretty clear that you aren't going to be swayed by the rules of math. Contraction and dilation change the magnitude, not the units.

[Andromacus]

Do you have a link or two? I don't recall ever seeing that.

 

You don't recall seeing that c must be constant under a lorentz transformation? That's what leaving the null interval invariant means.

Are you talking about chapter 11? That's not what he's doing. He specifically states "the law of the transmission of light in vacuo is satisfied both for the reference-body K and for the reference-body K1" (bodies — these have mass) The frames in question are inertial. He's transforming a distance, which is perfectly valid. He never mentions a frame for light. It is not a transformation for photons. The LT is a transform from one inertial frame (valid only for massive bodies where we can say they are at rest) to another.

We seem to be talking at cross-purposes. I am not talking about light frames, we already discussed this. It is a transformation between inertial frames for light. It's the LTs for time and distance, but since what is travelling is light x is changed by ct.

 

Again, technically true, because you can debate anything (1+1=2? That's debatable), but the units of time actually remain units of time. If you are willing to debate this, then it's pretty clear that you aren't going to be swayed by the rules of math. Contraction and dilation change the magnitude, not the units.

 

Again I don't know what you are talking about, it seems you are interpreting that by changing the units I mean something like switching from meters to volts. Sorry but it isn't that. I'm talking about the scaling. Have you never worked with fundamental constants like c change to unit=1, if you have you should know that afterwards if you want to apply in practice the result you have to rescale to the real value of c so that the all the units come out right.

 

 

 

Well, I think it is enough for the moment.

[swansont]

You don't recall seeing that c must be constant under a lorentz transformation? That's what leaving the null interval invariant means.

 

No, as I have repeatedly said, my objection is to your claim that an EM wave is being transformed ("transformation of an EM wave"), or anything to due with a photon having a frame of reference. You keep saying that you aren't doing this, but then you keep saying it. There is no transformation of a photon going on. The transform is between inertial frames.

 

And while c is the speed of light, using it does not mean you have a photon. Determining what the length ct is in another frame does not imply light is involved, in any way, in the problem.

 

Again I don't know what you are talking about, it seems you are interpreting that by changing the units I mean something like switching from meters to volts. Sorry but it isn't that. I'm talking about the scaling. Have you never worked with fundamental constants like c change to unit=1, if you have you should know that afterwards if you want to apply in practice the result you have to rescale to the real value of c so that the all the units come out right.

 

If you mean scaling, you should say scaling, and not say the units are changing. If you mean transforming using a certain length, then you should say that, and not "inertial frames for light", since light doesn't have an inertial frame.

 

 

In all this you say there is some inconsistency, but you still haven't shown what it is.

[Andromacus]

 

In all this you say there is some inconsistency, but you still haven't shown what it is.

 

 

Ok, here it is in the simplest terms I can:

 

The relativity of simultaneity is the demand that simultaneity and therefore synchronzation of clocks must be conventional between frames in a transformation. This is necessary to comply with both the SR postulates, and it is equivalent to saying that Einstein's synchronization is not the only one valid. Different values for c in one direction and its opposite are allowed in the theory for the formula of synchronization of clocks, not just the standard 1/2.

 

On the other hand the definition of the null interval in SR, wich is an invariant, meaning that it must be preserved under any change of frames, is done using the equation ct=x that is only compatible with the Einstein synchronization, any other synch will not give the null interval c2t2-x2-y2-z2=0. It is not relevant if the derivation is done in one frame because the interval is meant to be invariant to change transformations.

 

 

[swansont]

 

 

 

Ok, here it is in the simplest terms I can:

 

The relativity of simultaneity is the demand that simultaneity and therefore synchronzation of clocks must be conventional between frames in a transformation. This is necessary to comply with both the SR postulates, and it is equivalent to saying that Einstein's synchronization is not the only one valid. Different values for c in one direction and its opposite are allowed in the theory for the formula of synchronization of clocks, not just the standard 1/2.

 

On the other hand the definition of the null interval in SR, wich is an invariant, meaning that it must be preserved under any change of frames, is done using the equation ct=x that is only compatible with the Einstein synchronization, any other synch will not give the null interval c2t2-x2-y2-z2=0. It is not relevant if the derivation is done in one frame because the interval is meant to be invariant to change transformations.

 

It's not obvious to me that relativity of simultaneity demands that the clock synchronization be conventional, but I also don't see the inconsistency here. If you had a different clock synchronization protocol, you would need a different transformation between the frames.

[Andromacus]

 

It's not obvious to me that relativity of simultaneity demands that the clock synchronization be conventional, but I also don't see the inconsistency here. If you had a different clock synchronization protocol, you would need a different transformation between the frames.

I'm under the impression that the term causing problems here is "conventional". And it is in part my fault as I'm not using the usual meaning of conventioal synchronization as used in the conventionalist versus non-conventionalists debates. For them conventional synchronization refers to different synchronizations within one frame,but I'm not using that meaning I'm using the word conventional referring to the different simultaneities allowed by definition by the relativity of simultaneity, but I see that this invites confusion.

So let me rephrase(wich makes the demand almost a tautologic redefinition): the relativity of simultaneity obviously admits different simultaneities for different inertial frames related by change of frames, and different simultaneities means different synchronizations for different frames are valid in the theory. So the theory is obviously compatible with different clock synchronizations by its central feature:the relativity of simultaneity.

 

At the same time the invariant null interval, wich is also a key part of the theory because it is the mathematical statement of the second postulate and is also required to hold thru for the definition of minkowski space line element, only allows the standard Einstein synchronization, that is only agrees with one simultaneity, the one in wich the 3 dimensional euclidean distance equals the constant c times the time that takes a light ray in vacuum to traverse that distance measured by clocks at rest with each other in the euclidean space.

 

So the theory allows the contradiction of being compatible with different simultaneities and with only one. There is of course a way to dodge the contradiction, wich is making all simultaneities equivalent, but this takes away the physical content of the Lorentz transformations and time dilations by making the meters and second change their scale in the change of frame.

Experiments seems to indicate that the Lorentz transformations and the time dilation corresponds to something physical, not a simple rescaling of the time and length units, so I don't buy that way out of the inconsistency.

 

Note that Lorentz theory didn't have this inconsistency as the length contraction was considered both real and to ocurr only in the ether absolute frame of the direction of motion. Of course the explanation was clearly hard to justify since it was not measurable and completely ad hoc.

 

So we have the math of the Lorentz transformations wich works to the best our experiments can tell, but i wonder how the Einstein explanation is an improvement to the already faulty Lorentz theory explanation if it just introduces inconsistency.

Edited September 16, 2015 by Andromacus

[swansont]

It's not obvious to me that if it were possible to have a different clock synchronization protocol that you would end up with the Lorentz transform. IOW, the definition of the null interval might be different.

[Andromacus]

It's not obvious to me that if it were possible to have a different clock synchronization protocol that you would end up with the Lorentz transform. IOW, the definition of the null interval might be different.

I don't quite follow what you mean. The synchronization protocol is always the same: tB=tA+e(tA'-tA) where only e varies being 1/2 in the Einstein synch, and taking any value between 0 and 1 for the different simultaneities in the transformed frames(for intervals different from zero) depending on the relative velocity bewteen the frames..

 

The invariance of the null interval means that the only simultaneity compatible with it for any inertial frame is the standard Einsteinian where the time it takes light to go from A to B is the same it takes to go from B to A (the simultaneity is absolute here). While the relativity of simultaneity is the assertion that any simultaneity, that is, any combination of times from A to B and from B to A are allowed as long as the round-trip is constant. Would you say it is possible to have both relative and absolute simultaneity in SR(other than by the unphysical way of re-scaling the units I mentioned above)?

Edited September 16, 2015 by Andromacus

[swansont]

I don't quite follow what you mean. The synchronization protocol is always the same: tB=tA+e(tA'-tA) where only e varies being 1/2 in the Einstein synch, and taking any value between 0 and 1 for the different simultaneities in the transformed frames(for intervals different from zero) depending on the relative velocity bewteen the frames..

 

 

So then we aren't talking about an arbitrary synchronization protocol?

 

The invariance of the null interval means that the only simultaneity compatible with it for any inertial frame is the standard Einsteinian where the time it takes light to go from A to B is the same it takes to go from B to A (the simultaneity is absolute here). While the relativity of simultaneity is the assertion that any simultaneity, that is, any combination of times from A to B and from B to A are allowed as long as the round-trip is constant. Would you say it is possible to have both relative and absolute simultaneity in SR(other than by the unphysical way of re-scaling the units I mentioned above)?

 

I am not aware of any absolute simultaneity in SR.

 

 

Scaling units? Again?

[Andromacus]

 

 

 

So then we aren't talking about an arbitrary synchronization protocol?

 

I wrote the formula that defines the synchronization for a given simultaneity in SR. I thought it was simple enough to understand Wouldn't you have to define what you mean by "synchronization protocol" for me to answer?

 

 

 

 

I am not aware of any absolute simultaneity in SR

 

But that is because you apparently claim you ignore the existence of null intervals and their definition in SR and the fact they require their simultaneity defined in terms of the synchronization tB=tA+1/2(tA'-tA), and that it doesn't allow any other simultaneity with a factor different from 1/2 as would be required by the different simultaneities postulated by the relativity of simultnaneity , no matter how I write it, or the references to Einstein's papers or books I give.

 

Could you maybe try giving an argument from SR that shows what I say about the null interval in different frames is wrong, instead of just replying "I am not aware of it", or "it is not obvious to me"?. I tend to think you should be aware of it or it should be obvious to you a way to counter what I say if it is wrong.

Edited September 17, 2015 by Andromacus

[swansont]

Could you maybe try giving an argument from SR that shows what I say about the null interval in different frames is wrong, instead of just replying "I am not aware of it", or "it is not obvious to me"?. I tend to think you should be aware of it or it should be obvious to you a way to counter what I say if it is wrong.

 

It's your claim, are you unwilling to defend/explain it? This is your thread, the burden of proof is on you. Things aren't right just because nobody has countered the claim.

 

I've had specific objection to things you've said, some of which is due to you using non-standard terminology (as you have admitted, this causes confusion), but I'm not required to defend a position just because you want me to.

[Andromacus]

 

It's your claim, are you unwilling to defend/explain it? This is your thread, the burden of proof is on you. Things aren't right just because nobody has countered the claim.

 

I've had specific objection to things you've said, some of which is due to you using non-standard terminology (as you have admitted, this causes confusion), but I'm not required to defend a position just because you want me to.

You are right.

It is just frustrating not to be able to get what I mean across.

 

I'll give it another try:

 

a) This thread is about the relativity of simultaneity, let's agree about its definition as the assignment of different simultaneities for different inertial frames in relative motion with respect to each other related by Lorentz transformations, so that what is simultaneous in one is not in the other and viceversa.

 

b) The different simultaneities for each frame(let's call them frame A and frame B) are defined by a synchronization for frame A and a different(as seen from A) one for frame B and viceversa. Of course from its own frame both A and B think their clocks are the ones correctly synchronized according to Einstein synch and the other as seen from their frame to be the incorrect one. .

 

c)The general formula the 2 synchronization followed within each frame is tQ=tP+e(tP' - tP) where tP is the time at which a light signal is sent from point P in a frame, tQ is the time it is recieved at point Q in the same frame, tP' is the time the light signaled reflected at Q arrives back at P and e(epsilon) is the coefficient from 0 to 1 that gives the different times from P to Q and from Q to P(e=0.5 corresponds to Einstein synch).

 

d)The different synchronizations between the frames that give rise to different simultaneities(relativity of simultaneity) vary in the value of e(the value of e in the moving frame depends on the relative speed between the 2 frames:e in moving frame=rest frame e times(v/c), and it's usually represented by tilting axes(or plane depending on the dimensionality of the representation) in spacetime diagrams.

 

This description of the relativity of simultaneity is consistent with mainstream SR AFAIK, before I go on I would like to have some consensus that it is so.

 

Any comments are welcome.

Edited September 18, 2015 by Andromacus