Jump to content

Talk:One-way speed of light/Archive 1

Page contents not supported in other languages.
From Wikipedia, the free encyclopedia
Archive 1Archive 2

Changes

Hi Charvest. I have made a few changes. I note that the overview of your second reference starts with, 'Michael Tooley presents a major new philosophical theory of the nature of time'. It seems that this is just a philosophical proposal. Do you have any details of the proposed experiment? Martin Hogbin (talk) 12:51, 17 September 2009 (UTC)

Good work on the changes Martin. I just used that reference because page 350 says there has been an "unbroken series of sometimes quite ingenious but unsuccessful attempts to design an experiment to determine the one-way speed of light". Charvest (talk) 14:18, 17 September 2009 (UTC)

Is there anything here that isn't covered elsewhere?

I.o.w. do we need this article? DVdm (talk) 16:01, 17 September 2009 (UTC)

I think give it time. Google shows that there is plenty to write about. Charvest (talk) 17:08, 17 September 2009 (UTC)
Indeed. Mostly ether crackpots. This is going to be a magnet. DVdm (talk) 17:43, 17 September 2009 (UTC)
Yes. I'm considering nominating for deletion based on WP:FORK. Headbomb {ταλκκοντριβς – WP Physics} 17:47, 17 September 2009 (UTC)
I think information about the various attempts at testing the one-way speed would belong in this article rather than elsewhere. As per discussion at talk:speed of light, it's better to have the info here rather than confusing readers at the speed of light article with too much information about one-way speed conventions. Charvest (talk) 17:53, 17 September 2009 (UTC)
Per Wikipedia:Content forking this article would be a spinout of the paragraph in the speed of light article that links here. Charvest (talk) 18:11, 17 September 2009 (UTC)

I agree that this page will be a magnet for aether crackpots which is why I added material to cover most of the subject before someone else did. With regard to aether crackpots I think it is better to have an open explanation of the subject rather than ignore it which will be treated by them as though physicists have something to hide. You know the story, 'Shock horror, Einstein's theory is all based on an assumption'. Martin Hogbin (talk) 19:06, 17 September 2009 (UTC)

1983 definition of the metre

Page 71 of the book The arguments of time By Jeremy Butterfield says that the 1983 definition of the metre has implications for testing the one-way speed of light by virtue of changing the meaning of speed. Charvest (talk) 17:08, 17 September 2009 (UTC)

The problem seems to be resolved by the section just after the one you quote:

On the other hand, conventionalist philosophers can relish the though that no international agreement has restrained their freedom to drop the standard definition of distance as a symmetric real-valued function on point-pairs, […]

In other words, if you are looking to show that the one-way speed of light is different from the two-way speed of light, you need a new definition of distance. That's not nearly as outlandish as it sounds: indeed it's a practical problem in solar system ephemerides. The CIPM certainly won't make the current definition of the metre stop you! Referring to one well-known case where distance is not necessarily a "symmetric real-valued function on point-pairs":

Method (a) [time-of-flight measurements] should therefore be restricted to lengths l which are sufficiently short for the effects predicted by general relativity to be negligible with respect to the uncertainties of realization.

Or, to paraphrase, only use the SI metre when GR is negligible; otherwise, use something else. Physchim62 (talk) 18:25, 17 September 2009 (UTC)
I think it does rather support Brews position that there is an important difference in defining speed of light in terms of transit-time vs a distance based version. Charvest (talk) 18:35, 17 September 2009 (UTC)
Then again the speed of light is only constant locally, so any definition of speed of light does not apply over distances in non-inertial frames, so maybe it doesn't say anything about the recent disputes at all. Charvest (talk) 18:48, 17 September 2009 (UTC)
I think you got closer the second time Charvest. Martin Hogbin (talk) 22:41, 17 September 2009 (UTC)

Charvest, Can you please elaborate on your final sentence regarding the defined speed of light not applying over distances in non-inertial frames. Does a measured speed of light apply over distances in non-inertial frames? If the answer to the first case is 'no', and the answer to the second case is 'yes', then we have a difference. David Tombe (talk) 08:14, 18 September 2009 (UTC)

Just as I predicted, here we go! DVdm (talk) 08:17, 18 September 2009 (UTC)
I hope not. In the case of non-inertial frames and GR I strongly believe that we should say as little as possible. The only only points that we should make are that the physics is well understood, and that all intuitive notions of distance, time, and speed fail and there is little point even using these words. You must either describe exactly the method you would use to make the measurement or you use the full maths and jargon of GR. Martin Hogbin (talk) 09:07, 18 September 2009 (UTC)
This paper gives an account of the problems and suggests some fixes: since it appeared in 1995, fixes of this type (not exactly the same ones, but very similar) have indeed been applied to the astronomical system of units. So the answer to David's first point is "yes". The answer to his second point ("no") should be obvious from the CIPM quote I posted above. Physchim62 (talk) 09:41, 18 September 2009 (UTC)
I think that paper illustrates my point perfectly. Once you get into GR life gets very complicated and it is impossible to use terms such as 'distance' or 'time' in the hope that everyone knows what is meant. Best say very little. Martin Hogbin (talk) 10:18, 18 September 2009 (UTC)
Another, perhaps more practical illustration is this table of astronomical constants taken from the Astronomical Almanac. Line 6, "Light-time for unit distance", is the inverse of the speed of light in astronomical units: you'll notice that it has two values, one labelled "TDB" and the other labelled "SI". Both are correct (to within the stated measurement uncertainty, which corresponds to ur = 4×10−11). Both are actually the same measured value… they just use different definitions of the second! There are a few more details at astronomical unit, if anyone's really interested ;) Physchim62 (talk) 11:38, 18 September 2009 (UTC)
The real problems with the speed of light in GR are not with the physics, which is perfectly well understood, but with the English language which is not able to describe the situation using ordinary words (unless you use lots of them). Martin Hogbin (talk) 17:01, 18 September 2009 (UTC)

David, I know very little about GR. Better, instead, to read Physchim62's posts. Charvest (talk) 10:31, 18 September 2009 (UTC)

Charvest, Your first point above was correct. Don't be fooled into thinking that you need to understand General relativity in order to understand the basic simplicity of Brews's argument. Brews's argument, which I support, has got absolutely nothing to do with relativity. David Tombe (talk) 12:59, 18 September 2009 (UTC)
For once I agree with David. The one-way vs two-way speed of light question is not related to Brews' argument, or general relativity, or the difference between SI and astronomical units. So I suggest that we drop discussion of these subjects on this page. Martin Hogbin (talk) 16:57, 18 September 2009 (UTC)

Yes, I can't see that it is connected to the 'one-way/two way' speed of light controversy either. Charvest's post above would have assisted Brews best if it had been put on the talk:speed of light page, even though the paper in question was specifically talking about the one way speed of light. David Tombe (talk) 09:21, 19 September 2009 (UTC)

Page ix of the preface of the Jong-Ping Hsu and Yuan-zhong Zhang book Lorentz and Poincaré invariance: 100 years of relativity in their discussion of alternatives to the one-way speed convention say: "we must revert from using the current definition of the meter ... to a previous definition based on the length of a physical bar. The reason for this is that the current definition of the meter implicitly assumes the universality of the one-way speed of light". So the defined value of the speed of light is not compatible with alternative one-way speed conventions. Charvest (talk) 16:25, 20 September 2009 (UTC)

Charvest, Fair enough. This would be a good source to produce at the arbitration hearing. David Tombe (talk) 19:33, 21 September 2009 (UTC)

Analysis of rotating frame implies there is only one possible convention

This paper: TOWARD A CONSISTENT THEORY OF RELATIVISTIC ROTATION seems to say that there is no choice in what to use as a convention. Charvest (talk) 18:00, 17 September 2009 (UTC)

There are many papers saying many things, even this one [[1]] by a well respected physicist. This is one reason why I think this page should stay. Although the constant one-way speed of light was originally stated as a postulate (axion or convention) by Einstein, there have been times quite recently when experts thought that the one-way speed could actually be experimentally measured, as the linked paper shows. It was not until 1997 that Zhang's careful re-analysis of these experiments showed that, in fact, they only measured the two-way speed. As far as I know his book is still regarded as the definitive work on the subject. Martin Hogbin (talk) 23:15, 17 September 2009 (UTC)

A Broader View of Relativity

Martin, you removed the Hsu reference from the lead, with the edit summary that maybe a comment on the body. If you mean mention the reference somewhere else in the article, that's fine by me, I don't mind where it's mentioned but I do think it should be mentioned as the whole book is basically an elaboration of the possibilities of using an alternative convention for the one-way speed and it goes into a level of detail about this that isn't found anywhere else. P.S. I saw your message at sci.physics.research about gyrovectors, so I've added an explantion to gyrovector space that I hope will be more to your liking (I've also responded on usenet but it hasn't been modded yet.) Charvest (talk) 17:29, 18 September 2009 (UTC)

I did mean that we might have something in the article but it must explain that there are no such theories in common use. Also we must make the clear distinction between theories which are equivalent to SR and ones which are not. I am not sure what is being referred to here. Anyone is free to use LET (or Edwards theory) if they like because it gives exactly the same results as SR. As far as I know nobody does.
If these theories are not equivalent to SR then I would ask what experimental predictions they make and whether they have been tested.
I am extremely sceptical about gyrovectors, as were a couple of respondents at sci.physics.research. I was hoping that someone like John Baez would reply with a definitive answer. Martin Hogbin (talk) 17:50, 18 September 2009 (UTC)
The formulations by Hsu are equivalent to SR. Hsu proposes that they may find use for being easier to perform some calculations than using SR. Charvest (talk) 17:53, 18 September 2009 (UTC)
Fine, but does anyone actually use them? Martin Hogbin (talk) 17:56, 18 September 2009 (UTC)

I am also not happy with the last sentence of the lead. As your source intimates, there have been many crackpot attempts to measure the one-way speed of light as this was generally considered impossible. There was a period when some experts (notably Will) thought that some experiments might measure the one-way speed of light but after Zhang's re-analysis I understand that Will, to his great credit, accepted his error.

Anything we say here must make clear the fringe nature of any further attempts to measure the one-way speed. Martin Hogbin (talk) 18:02, 18 September 2009 (UTC)

Well the original wording was "This hasn't stopped people from trying to design experiments ..." which I think got the point across better. Charvest (talk) 13:04, 20 September 2009 (UTC)
The point is that anyone who tries to design such an experiment today will be regarded similarly to someone trying to design an experiment to disprove the second law of thermodynamics, maybe not a complete crackpot but certainly someone with little chance of success. Martin Hogbin (talk) 20:34, 20 September 2009 (UTC)

In defence of this article.

I understand, as some of you will know, that relativity (with related topics) comes second only to evolution in attracting science crackpots, but that is one reason why this page may be useful. I think there are three main groups of possible readers that we should consider. Firstly, the crackpots themselves. They will continue to hold their crackpot views regardless of anything we say here or indeed anyone else says anywhere. They can be ignored, unless they try to edit the article. At the other extreme we have the physicists who already know everything said here. There is nothing for them here either.

In between we have everything from the interested amateur to the physicist who is not aware of the exact situation regarding the speed of light, and I suspect that many physicists believe that the constant speed of light has been experimentally verified in all cases. These people are our target audience. Some of them may read and be persuaded by one of the many crackpot web sites on the subject. It should also be noted that until quite recently (1997) some eminent experts on the subject got it wrong and suggested that there were experiments that did measure the one-way speed of light.

It is therefore important that we state our case confidently, but it is equally important that we do overstate the case or make unsubstantiated claims. The way to convince people is to state the facts plainly and clearly, supported by good quality reliable references. In many cases these will come from Zhang's book, 'Special Relativity and its Experimental Foundations', which devotes considerable space to this topic. In other cases, I suggest well known relativity text books.

In short, I believe that this could be a useful article, provided that we keep the crackpots at bay. Martin Hogbin (talk) 21:17, 20 September 2009 (UTC)

It's a specific and fairly self-contained topic that deserves to be treated properly, which would be impossible in the Speed of light article without making SoL unreadably long. I would say that it serves a purpose, yes, and its existence hardly needs defending. Physchim62 (talk) 21:25, 20 September 2009 (UTC)
I wrote this in response to the suggestion above that the article should be nominated for deletion. Martin Hogbin (talk) 22:03, 20 September 2009 (UTC)

Aberration

What about aberration? Is that also generally accepted[by whom?] to be a two-way measurement? NOrbeck (talk) 05:36, 26 July 2010 (UTC)

There are no claims that aberration measures the one-way speed of light that I know of. Martin Hogbin (talk) 09:56, 26 July 2010 (UTC)
Yes, everyone accepts that experiments have shown[which?] that the one-way speed of light cannot be measured. I was suggesting that the explanation for aberration be added to the article.
It should describe how the c in this inverted form of Einstein's aberration equation is actually the a two way average light speed between the star and the observer:
NOrbeck (talk) 06:41, 27 July 2010 (UTC)

At the moment we have no explanations of why any supposed one-way measurements are in fact two-way measurements. If you want to add explanations as to why measurements which were at one time thought to be one-way were on fact two-way measurements, that would be a good idea. The logical experiments to start with are those that have been claimed at some time to measure the one-way speed of light. Martin Hogbin (talk) 08:40, 27 July 2010 (UTC)


[2] "In 1728 James Bradley, an English physicist, estimated the speed of light in vacuum to be around 301,000 km/s. He used stellar aberration to calculate the speed of light." I suggest that more relativists believe aberration is a measurement of the one-way speed of light, than believe the one-way speed of light in "undefined".

There are two great books on on this topic: "Relativity in Rotating Frames", and "Concepts of Simultaneity" by Max Jammer. It's been years, but as I recall the theme of the first book was that these issues were still being actively debated. I do know there were legitimate opponents to the views expressed in this article.

It's clear that relative simultaneity makes measuring one-way latencies virtually impossible without synchronization, but the following claim, as stated, seems untenable: "It is generally accepted that the "one-way" speed of light from a source to a detector, cannot be measured independently of a convention as to how to synchronise the clocks". The WP:Weasel words don't help it's credibility. If you replace the word "speed" with "latency" the statement becomes much more reasonable, even tautological.

I've seen this bizarre claim a few times in the literature, but I doubt this is the majority view: "the one-way speed ... is undefined (and not simply unknown) because it does not make sense to speak of one-way speed prior to the synchronization of distant clocks"

This is undoubtedly the current standard view on this matter. There was some doubt about the subject, promoted to some degree by the well-respected[peacock prose] scientist Clifford Will, before the publication of 'Special Relativity and its Experimental Foundations' by Yuan Zhong Zhang in 1997. This book was very well received by the physic community and Will has accepted that he was wrong in some claims he had made that the one-way speed of light could be measured. Martin Hogbin (talk) 13:52, 27 July 2010 (UTC)

Does anyone actually believe that aberration is not a measure of the one way speed of light? —Preceding unsigned comment added by NOrbeck (talkcontribs) 11:17, 27 July 2010 (UTC)

Yes. It is covered on pages 94-95 of Zhang. Martin Hogbin (talk) 13:34, 27 July 2010 (UTC)
If you have the book in hand, you'd be doing a very good Wikideed if you added a detailed explanation to the article of what he discusses in those two pages. ;) Wnt (talk) 16:01, 3 August 2010 (UTC)
The curiosity is killing me. Per Wikipedia guidelines I request that pages 94-95 be posted here for verification. I'm looking forward to adding the explanation to the article. NOrbeck (talk) 22:01, 6 August 2010 (UTC)
I have no means of posting the pages except by typing them in by hand. However there is no need to post them because you have not provided a source claiming that aberration measures the one-way speed of light. Martin Hogbin (talk) 17:00, 11 June 2011 (UTC)

How general is the problem?

I think this article should make it clear whether this is a theoretical or a practical problem (can you measure the one-way speed of light to some really low accuracy), and also whether it applies to the speed of light in vacuum or under all conditions.

For example, what comes to mind for me is finding a medium where (two-way) slow light propagation is equivalent to the speed of sound; by simultaneously triggering light and sound at one end, they should arrive simultaneously at the other if the one-way speed is the same. Though perhaps sound propagation is a slowly transported clock? (I'm not quite sure what isn't a slowly transported clock) Wnt (talk) 15:40, 3 August 2010 (UTC)

Well, the opening sentence says "the one-way speed of light ... cannot be measured independently of a convention...". It doesn't say it can be measured approximately. In fact, under the standard convention, it is defined to be equal to the two-way speed c, and under a non-standard convention it could take any value you want it to take between ½c and ∞, varying by direction. The value is determined by human convention, not by experiment. Note that changing the convention would change the values of all speeds, not just the speed of light. Is there any way the article could be rephrased to make this clearer? -- Dr Greg  talk  20:43, 3 August 2010 (UTC)
I take it then that you're describing the speed in vacuum only, since slow light has been clocked doing 38 mph, which isn't between 1/2c and infinity. Wnt (talk) 20:51, 3 August 2010 (UTC)
Yes that is correct. As Dr Greg all measured speeds would be affected by a change in clock synchronisation convention. Maybe I will add something to this effect. Martin Hogbin (talk) 21:00, 3 August 2010 (UTC)
Thanks. While you're at it... to tell you the truth, I can't quite figure out how you derive the 1/2 c and infinity values. I realize that such weird alternate conventions don't warrant a lot of space in the article, but a link to a good source or two could be helpful. Wnt (talk) 21:38, 3 August 2010 (UTC)
The Zhang source already cited shows the method. You can download the first chapter from here. In equations 1.3.13 the parameter qr can be set to ±1, giving the c/2 and infinity values. Zhang calls the equations "Edwards simultaneity", which I'm sure is what the mention of "Edwards theory" removed as potential vandalism referred to (just as well, as it was unexplained and out of place). It would be a tremendous improvement to this article to provide a summary of the general simultaneity equations of Edwards, since they are fairly straightforward and are at the heart of the theoretical impossibility of measuring one-way light speed without a simultaneity convention. Tim Shuba (talk) 21:34, 4 August 2010 (UTC)
Edwards theory is pretty well the same as LET mathematically, it does not mention an aether but it allows different clock synchronisation schemes in the same way that LET does. Tom Roberts always claimed the two were the same and I see no reason why he is wrong. Martin Hogbin (talk) 22:16, 4 August 2010 (UTC)
I disagree with your assessment, but in order to do so effectively I would have to know precisely what "Edwards theory" means. I tried to search for it and found basically nothing other than Zhang's book (which I've only read the first chapter). I don't think it's a standard term and would recommend leaving it out unless it can be properly referenced. In particular, LET simultaneity is not the same as the more general Edwards simultaneity given above, as the former is tied to an ostensibly preferred frame in which qr must be isotropically 1/2 zero. Tim Shuba (talk) 22:58, 4 August 2010 (UTC)

My understanding is that Edwards' theory is a framework which allows for non-standard synchronisations and which encompasses both (Einstein-synchronised) special relativity and (absolutely synchronised) ether theory as two special cases of families of coordinate systems. (So it is really a subset of general relativity – in flat spacetime – which allows for arbitrary coordinate systems.) Therefore I suspect it probably doesn't make sense to say "(Lorentz ether theory) has later been referred to as Edwards theory". Maybe someone has referred to it this way, but it would be nice to have reference. Edwards' original paper was: Edwards, W.F. (1963), "Special Relativity in Anisotropic Space", American Journal of Physics 31, Issue 7, pp. 482–489. -- Dr Greg  talk  23:14, 4 August 2010 (UTC)

The term 'Edwards theory' is used in Zhang's book, in fact it is the title of chapter 6. My reason for using the term here is that Tom Roberts, whom Tim knows as a reliable physicist, used the term consistently for the same thing as LET. My understanding of what it is, is just as Dr Greg describes. Maybe 'an Edwards theory' would have been better.
I think the best plan is to remove the Edwards reference from the LET section and add a short section on Edwards theory to the article. This could then be expanded to show its exact relationship to LET when this has been agreed. It has a different philosophical basis in any case. Martin Hogbin (talk) 08:41, 5 August 2010 (UTC)
I have started a section. It could do with rewriting to improve the flow. Martin Hogbin (talk) 09:03, 5 August 2010 (UTC)
I have now expanded the section based on what Zhang says about it.
It does look to me that Edwards' theory is mathematically pretty much the same as LET in that it differs from SR only in having a parameter q which seems to represent an arbitrary velocity. This would correspond to the motion of the aether in LET, although the language and philosophy are different in the two theories. I would like to make this point in the article but I am not sufficiently confident of my conclusion to do so yet. Any suggestions? Does anyone know how to contact Tom Roberts? Martin Hogbin (talk) 13:21, 5 August 2010 (UTC)

American or British spelling?

We seem to have a mixture of both American and British spelling: synchronise and synchronize (and -ation); metre and meter. -- Dr Greg  talk  23:41, 4 August 2010 (UTC)

There's a policy WP:ENGVAR, which I've tried to avoid reading, but skimming it just now it looks like either we can honor a particular physicist (but Einstein, Lorentz, and Zhong Zhang all seem like they might be neutral ;) or we can dig through the history to see what the earliest choice was... but who wants to bother? It's about an even split. Hmmm, maybe we can write a "flip a coin" template... Wnt (talk) 06:25, 5 August 2010 (UTC)
Sigh... looks like mw:Extension:Random isn't implemented here. What a tragedy for science... ;-) Wnt (talk) 06:32, 5 August 2010 (UTC)
Charvest started the page with 'synchronize'. Zhang's book uses that spelling. I guess it should be US then.
I have changed all to 'synchronize'. Martin Hogbin (talk) 17:41, 5 August 2010 (UTC)

Incorrect statement?

Is this statement a typo or intentional? "Observations of the arrival of radiation from distant astronomical events have shown that the one-way speed of light does not vary with frequency." NOrbeck (talk) 08:38, 8 August 2010 (UTC)

I cannot see a typo. What is the problem? Martin Hogbin (talk) 11:37, 8 August 2010 (UTC)
"Have have". I made the correction. DVdm (talk) 11:40, 8 August 2010 (UTC)
Sorry I didn't even notice that typo. Actually I was wondering if the intended statement was "the speed of energy propagation in the ISM is frequency dependent"? NOrbeck (talk) 11:07, 10 August 2010 (UTC)
No. Maybe I should add, 'in vacuum' or a statement somewhere that this should be assumed unless stated otherwise.Martin Hogbin (talk) 13:36, 10 August 2010 (UTC)
I see there already is such a statement. Martin Hogbin (talk) 13:37, 10 August 2010 (UTC)

Preferred Reference Frame in SR

This discussion has been closed. Please do not modify it.
The following discussion has been closed. Please do not modify it.

The article states that there is no preferred reference frame in SR but this is incorrect. Anyone who cares to consult Einstein's original "Electrodynamics of moving bodies" paper (http://www.fourmilab.ch/etexts/einstein/specrel/www/) can see in section 3 that a moving system "k" is compared with a "stationary" system "K". "K" is therefore a special or preferred system.

Moreover, a light ray is propagated forward from the origin in the moving system "k" and is supposed to have speed (c-v) in that inertial frame. It is further supposed to have speed (c+v) on its return trip after reflection. All this is clearly set down by Einstein in the paper, (He uses x'/(c-v) and x'/(c+v) for the times to go forward and back).

Meanwhile, the same light ray is taken to travel at just c in the "stationary" system "K". Einstein is thus making "K" an implicit preferred reference frame.

Whilst it is perfectly true that the means of clock synchronisation mean that the "stationary" preferred system is always undetectable, there is an important distinction between something not being detectable and not existing.

Because light signals are used to synchronise clocks, two separate clocks will always give c as the measured one-way speed, even though the true speed may be c+v or c-v.

The implicit preferred frame is essential to Special Relativity as without it and the consequent c-v and c+v "true" light speeds, Einstein would not have been able to derive the Lorentz transformations - which he does in section 3 precisely using just such c-v and c+v speeds for light. Zazby (talk) 16:40, 21 September 2010 (UTC)

It has been explained to you several times that this is not the place to discuss your philosophical musings on SR. Martin Hogbin (talk) 17:43, 22 September 2010 (UTC)
User was blocked for disruptive editing. DVdm (talk) 21:00, 22 September 2010 (UTC)

Appropriate addition?

I wonder if this addition, citing this source is appropriate:

  1. Is this a published, peer reviewed article?
  2. "Submitted on 11 Oct 2010 (v1), last revised 13 Nov 2010 (this version, v2))" that is, submitted last month, and revised last week.
  3. Possible conflict of interest: the paper originates from School of Physics, Institute for Research in Fundamental Sciences, Tehran, Iran and the user who made the edit is at Institute for Studies in Theoretical in Physics and Mathematics, Tehran.
  4. Wouldn't this kind of edit need at least a secondary source?

DVdm (talk) 14:45, 16 November 2010 (UTC)

Edits by 194.225.71.139

194.225.71.139, Your edits appear to be OR, please discuss here first. Martin Hogbin (talk) 19:42, 16 November 2010 (UTC)

Some suggestions

  • Zhang (2001, p. 542-547) referred to the De Sitter double star experiment as proving the source-independence of the one-way speed, while the Alväger Experiment only proves that the two-way speed (not the one-way speed) of light is independent of the source velocity. So the section "Experiments that can be done on the one-way speed of light" should be corrected.
Yes, it looks like you are right. It is odd that in the table on p 173 Zhang seems to have both one-way and two-way experiments. I think Alvager, Nilsson and Kjellman would be a better experiment to cite. It is made clear at the bottom of page 170 that this measures the one-way speed. The de Sitter experiment has been criticized because of extinction effects in the interstellar medium. Martin Hogbin (talk) 19:26, 18 April 2011 (UTC)
It was meant the other way round, i.e., Desitter should be citet. And Brecher already solved the extinction problem by observing high-frequent radiation. --D.H (talk) 19:48, 18 April 2011 (UTC)
Would it not therefore be better to cite Brecher? And why not add Alvager, Nilsson and Kjellman as a terrestrial experiment? Martin Hogbin (talk) 21:20, 18 April 2011 (UTC)
As I said above, Alväger measured the two-way velocity (at least according to Zhang, 2001, p. 547).--D.H (talk) 16:27, 19 April 2011 (UTC)
There are several experiments by Alvager et al. The Alvager, Farley, Kjellman and Wallin experiment does not according to Zhang measure variation of the one-way speed of light with source velocity but the Alvager, Nilsson and Kjellman experiment does. That is the experiment that I suggest we cite. Martin Hogbin (talk) 18:05, 19 April 2011 (UTC)
My mistake, I will include it again.--D.H (talk) 18:28, 19 April 2011 (UTC)
  • The article has three sections on LET. Maybe they should be merged into a single section, and the part concerning the 19th century "aether theory", should be moved into the empty "other theories" section.
I agree some consolidation of LET would be a good idea. Martin Hogbin (talk) 19:26, 18 April 2011 (UTC)
  • Maybe also the Greaves measurement, and the response by Finkelstein (both published in Am. J. Phys.) should be mentioned arXiv:0911.3616.
Yes, I presume you mean under the 'Experiments that appear to measure the one-way speed of light' section. Martin Hogbin (talk) 21:20, 18 April 2011 (UTC)
  • Also note, that there is still a discussion concerning the conventionality of synchronization, see Zalta, Edward N. (ed.). "Conventionality of Simultaneity". Stanford Encyclopedia of Philosophy.. --D.H (talk) 19:02, 18 April 2011 (UTC)
I would suggest that we do ad something on that subject but not too strong and maybe stating that the discussion is mainly philosophical. What do you think? Martin Hogbin (talk) 21:28, 18 April 2011 (UTC)
Yes, I agree. (Although something is already mentioned in Einstein synchronization --D.H (talk) 16:38, 19 April 2011 (UTC)

Another issue: The article says that Will later agreed with Zhang's analysis, regarding the JPL experiment. What is the source for this statement? --D.H (talk) 16:38, 19 April 2011 (UTC)

Although I am pretty sure that this is correct but I could not find a source. I do not think that we should remove the statement though, it is not contentious just a reflection of Will's integrity. Martin Hogbin (talk) 18:11, 19 April 2011 (UTC)

Slow transport

I do not much like the recent changes to this section. The fact that clocks do not remain sychronised when rapidly moved apart and together again is an experimental fact, which had two supporting references. Saying this will happen, due to time dilation involves the assumption of a theory. Martin Hogbin (talk) 21:06, 19 September 2011 (UTC)

Because the new section above on time dilation already refers to Hafele-Keating, therefore I thought it is redundant. Does it now sounds better? - if not, please make changes. --D.H (talk) 21:13, 19 September 2011 (UTC)
To my way of thinking, the phrase "time dilation" refers to a simultaneous comparison of two clock rates (and therefore, of course, depends on the definition of simultaneity you have chosen to adopt). I wouldn't describe a comparison of accumulated proper times along two worldlines as "time dilation", although I can't think of a simple short phrase to use instead. It might make some sense to refer to the twin paradox, the most well-known example of this phenomenon. -- Dr Greg  talk  01:34, 20 September 2011 (UTC)
Yes, that was more or less my point. The current wording mixes up experimental observations (such as the HK experiment) the scientific theories which explain the results. As Dr Greg points out and the article now states 'time dilation' is dependent on the assumptions made regarding clock synchronisation but the observations regarding the closed path movement of clocks do not dependent on the synchonisation scheme.
I am not sure that the time dilation section belongs in this article. It is fairly specifically related to the one-way speed of light rather than general issues surrounding definitions of simultaneity. Maybe the subject should be briefly mentioned in a special section called something like 'Other simultaneity issues'. Martin Hogbin (talk) 09:32, 20 September 2011 (UTC)
The dilation stuff is moved to "Other simultaneity issues", and the previous version of "slow clock transport" is restored. However, since there is no other article dealing with the consequences of alternative synchronizations, this is in my view the correct article to include this (since one- and two-way time dilation is analogous to one- and two-way speed of light.) --D.H (talk) 09:59, 20 September 2011 (UTC)
Thanks for making those changes, I think it is much better now. As there is no other article that deals with consequences of alternative synchronizations perhaps we should expand the "Other simultaneity issues" section. Martin Hogbin (talk) 18:23, 20 September 2011 (UTC)

Proposer of Lorentz Aether Theory

In the current article someone has added Poincare as on of the initial proposers of this theory. It is my understanding that this theory, in it complete form, was originally published by Lorentz in his 1904 paper. As with relativity, there are many other contributors to the theory, such as FitGerald. Does Poincarre deserve special mention here? Martin Hogbin (talk) 09:38, 27 February 2012 (UTC)

The current density in Lorentz's 1904 was not exactly relativistic. This was corrected by Poincare in his 1905/6 Palermo paper, so people often say that Poincare perfected and completed Lorentz's theory. Poincare also clarified the basis of Lorentz's theory.Fiddlefofum (talk) 02:02, 25 May 2013 (UTC)

Edward's "Theory" - redundant to Lorentz

The article presents Edwards' discussion as if it was distinct from Lorentz's interpretation. It wasn't. All Edwards did was to explicitly write down one member of the family of directionally dependent speeds of light implicit in Lorentz's interpretation, based on the choice of the "true" ether frame. Edwards covered just a subset of the entire set of possibilities within Lorentz's theory. So, the separate section on "Edward's theory" should be removed. I wouldn't even recommend citing Edwards in the Lorentz section, because it is incomplete.Fiddlefofum (talk) 02:15, 25 May 2013 (UTC)

I do not think that Edwards covered a subset of Lorentz's aether theory although I would say that the two are mathematically equivalent. What parts of LET are not covered by Edwards' theory? Martin Hogbin (talk) 17:04, 25 May 2013 (UTC)
Edwards didn't have a theory (or even an interpretation), he was simply describing (in his 1963 pedegogical write-up) some features of the Lorentzian interpretation. As for what he didn't cover, I don't see any explicit statement in Edwards of the simple general expression for the directionally dependent speed of light that is entailed by Lorentz's theory and interpretation. He gives the expression for one extreme case (infinite speed in one direction), but not the general case. (This Wikipedia article is even worse, since it gives c/(1+q) without even saying what q is for any value of the simultaneity parameter, so it's not even up to the level of Edwards.) Also, even if you extrapolate what Edwards was trying to say, to the point where it fully presents Lorentz's theory and interpretation, it will only be a presentation of Lorentz's theory and interpretation - which is my point, i.e., Edwards is not presenting anything different from Lorentz's theory and interpretation, so it's wrong to have separate sections for "Lorentz's theory" and "Edwards' Theory". At best, Edwards is just describing Lorentz. So you could delete the separate Edwards section and simply add a citation to the Lorentz section for an exposition - although as I said it isn't the greatest exposition. There are better, simpler, more clear and complete explanations.Fiddlefofum (talk) 21:59, 25 May 2013 (UTC)
"Edwards' theory" is the term used in reliable sources to describe a generalisation (by Edwards in 1963) of SR to include all possibilities in which the two-way speed of light is constant. The postulate on which this theory is based is included in the article. It is indeed only a mathematical theory in that, unlike LET, it does not attempt any metaphysical interpretation. It does, however, include a vector parameter P that is analogous to the reference frame's velocity through the aether in LET. Like LET Edwards' theory is experimentally indistinguishable from SR. Martin Hogbin (talk) 10:25, 26 May 2013 (UTC)
Edwards is not a generalization of special relativity, it simply expresses the results of special relativity in different coordinate systems. And those different coordinate systems are exactly the same ones that Lorentz called the "true" coordinates. The expression for the non-isotropic speed of light in Exwards' exposition is nothing but the non-isotropic speed of light in Lorentz's interpretation. The parameter P is not just analogous, it is identical. It might help to clarify this if you would revise the article to include the explicit expressions for the non-isotropic speed of light in Lorentz and in Edwards... which you would see are identical. (And of course they are also identical to the expression in special relativity for the speed of light in terms of those specific coordinates.)Fiddlefofum (talk) 16:46, 26 May 2013 (UTC)
In addition, the Lorentzian interpretation (in which a "preferred" frame is assumed) is only a special case of all possible, SR-equivalent reformulations:
a) LET related synchronizations ("absolute simultaneity"): One assumes a "preferred" or aether frame Σ, in which the Poincaré-Einstein convention is used in order to make the one-way speed of light isotropic, and in which time dilation and length contraction assume their standard values. Then by using external of absolute synchronization, all other frames such as Σ' are employing the Σ time-coordinates, making the one-way speed of light anisotropic and simultaneity is absolute. This special case was described by Grünbaum, Edwards, Winnie, Mansouri-Sexl, etc.
b) Generalized Lorentz transformations: It was pointed out (see the review article by Anderson et al.) that even the LET assumption of isotropic one-way speed of light in an "aether" frame is conventional, and is only a subset of a wider formulation in which the one-way speed is conventional in all frames of reference (there is no aether frame anymore). Those formulations rely on the assumption of isotropic two-way speed of light in all frames. Such generalizations were first attempted by Edwards, Winnie, and more generally by Anderson and Stedman (1977), Giannoni (1978), or Zhang (1994). --D.H (talk) 11:09, 26 May 2013 (UTC)
There's a long history of controversy, confusion, and misunderstanding of this subject in "reliable sources", some of which you've listed, so I concede that it may not be possible (within the rules of Wikipedia) to fix the article. When you (and your sources) talk about "all possible SR-equivalent reformulations", you (and they) are really just describing the infinitely many different systems of coordinates we can choose to describe phenomena. Einstein originally took the expedient path of adopting standard inertial coordinate systems, i.e., coordinates in terms of which the laws of mechanics hold good in their isotropic and homogeneous form, and expressed his principles and results in terms of those coordinates (e.g., the "speed of light" has the invariant value c). Ever since then, every witless newbie who discovers afresh that special relativity can also be described in terms of different coordinate systems, and that the principles and propositions sound verbally different when expressed in such terms, thinks that he has discovered a "new" or "generalized" theory or interpretation, and rushes into print with his discovery - becoming yet another "reliable source" - of misunderstanding. What I'm saying is not new, and it does have support in the reputable literature, but unfortunately it is drowned out by all the witless reputable literature, so it's probably not possible to craft a sensible Wikipedia article on this subject.
At best all you could probably say is that there are schools of thought that contend that these things are generalizations of special relativity, and other schools of thought that contend that they are simply expressing special relativity in terms of different coordinate systems, and hence do not represent a different theory or interpretation in any meaningful sense. Then, as I mentioned above, it might also be helpful to actually give the explicit expressions for the non-isotropic light speeds in the purported "generalizations", and let the reader note for himself that they are identical.Fiddlefofum (talk) 16:45, 26 May 2013 (UTC)
I only responded to your initial statement that formulations such as those of Edwards are only a variation of LET, while I (and others) think the LET type synchronization is only a subset of the, so to speak, Edwards-Winnie-Anderson-Stedman type transformations. The LET stuff is related to the single case in which one denotes a specific frame as the "aether" or a "preferred frame", and then accordingly adopts the coordinates in all other frames.
Regarding the conventionalist thesis, I actually agree with most of the things you said. Yes, it's trivial that one can choose different coordinate systems and thus by definition the one-way speed of light becomes different. (That's indeed also the opinion of the conventionalists, even though one can argue about its importance). However, despite of its trivial nature, it surprises many people when they first here about the relation of differently defined coordinate systems to the one-way and two-way speed of light, and it's certainly useful to think about it.
Regarding "coordinates in terms of which the laws of mechanics hold good in their isotropic and homogeneous form". This point is closely related to the experimentally established equivalence between slow-clock-transport synchronization and Poincaré-Einstein synchronization, which shows that the same conventional one-way speed anisotropies must be incorporated into the concept of "inertial" motion and momentum per se (see the article by Janis). In this paper, also a similar recent discussion (Ohanian against Macdonald and Martinez) in "American Journal of Physics" was mentioned. Ohanian criticized conventionalism by arguing that Newton's laws of motion require isotropic inertial motion, which is intimately connected with the concept of inertial reference frames, and any anisotropic reformulation would be absurd. Defending the conventionalist view, however, Macdonald (paper) and Martinez (paper) said that it's not necessary to define inertial reference frames in terms of Newton's laws of motion, even though it makes physics much more complicated.
So one side argues that giving up useful concepts such as (isotropic) inertial frames is absurd, therefore the isotropy of the one-way speed of light is as certain as the validity of Newton's laws of motion in the limiting case of small velocities (e.g. Ohanian). While the other side maintains, that the mere possibility of using non-inertial and anisotropic coordinate systems in order to produce an empirically equivalent transformation with anisotropic one-way speed, supports the conventionalist thesis and shows what is factual and what is conventional in relativity (e.g. Salmon, Macdonald, Martinez).
Regarding the terms "theory" versus "reformulation": Zhang employed the word "theory" for the respective formulations of Einstein and Edwards, while Anderson and Stedman say that all of those empirically equivalent formulations are manifestations of one and the same theory, namely special relativity. --D.H (talk) 19:32, 26 May 2013 (UTC)

How to improve the article

We, and the sources, agree that Edwards' theory and LET are experimentally indistinguishable from SR and have the same scope. LET and Edwards' theory are both in a section entitled, 'Theories equivalent to special relativity', so I do not think we need fret too much over whether they are generalisations, interpretations, or coordinate changes; physically they are all the same theory. Metaphysically the are different. LET considers a real aether, Edwards' theory considers space to be anisotropic.

What changes to the article does anyone think necessary? Martin Hogbin (talk) 19:25, 26 May 2013 (UTC)

Some suggestions: a list of experiments proving the equivalence of slow-clock transport synchronziation and Einstein synchronziation. Then a section on the relation between slow-clock transport, inertial reference frames, Newton's laws of motion, and one-way light speed using Janis and the AJP sources as described above. --D.H (talk) 19:32, 26 May 2013 (UTC)
I do not think that would address the concerns of Fiddlefofum, who I think is looking for changes to the 'Theories equivalent to special relativity' section. Martin Hogbin (talk) 19:46, 26 May 2013 (UTC)
I've made some suggestions to expand the article. See User:D.H/sandbox, in particular the introduction and the sections on "Inertial frames and dynamics", "Experiments comparing Einstein synchronization with slow-clock transport synchronization", and "Generalizations of Lorentz transformations with anisotropic one-way speeds". --D.H (talk) 19:04, 4 June 2013 (UTC)
I think it is important to distinguish between experiments that claimed to measure the one-way speed of light independently of any clock synchronisation scheme, or which were claimed by others to measure this, and experiments which were intended to test the equivalence of the constancy of the measured one-way speed of light and clock synchronisation by slow transport. The former are a possible sources of confusion and misinformation, the latter are somewhat esoteric experiments to verify current theory. Martin Hogbin (talk) 07:18, 5 June 2013 (UTC)
As explained by Mansouri-Sexl as well as Anderson et al., practically all of them can be interpreted as synchronization comparison experiments, irrespective of their intention to directly measure the one-way speed. --D.H (talk) 08:23, 5 June 2013 (UTC)
I am not questioning how the experiments can be interpreted but how they may be interpreted by some of our readers.
Greaves, Rodriguez and Ruiz-Camacho, and Krishner both claimed that they had measured the one-way speed of light independently of any clock synchronisation scheme although they were later proved wrong in this respect. These claims and, to a lesser degree, the claim that Romer did the same, are still found in many sources. I think it is important to let our readers know that we are aware of these experiments and that their original conclusions have been shown to be incorrect.
I think the other experiments should be listed separately under a different heading, not because they actually are different but because different claims are associated with them. Martin Hogbin (talk) 08:38, 5 June 2013 (UTC)
The problem is to properly distinguish all of them. For instance, Krisher et al. as well as Wolf et al. argued that they are considering both clock-transport as well as the isotropy of the one-way speed of light. Anyway, I think this is explained in the section by saying "Sometimes they have been denoted by their authors and others as being direct measurements of the isotropy of the one-way speed of light independent of a clock synchronization scheme...etc.". --D.H (talk) 10:03, 5 June 2013 (UTC)
I've included the changes now. Please review or correct them. --D.H (talk) 11:43, 5 June 2013 (UTC)
I think that GRR and K et al should both be in a separate section. Whatever the authors stated, these papers are often quoted as being measurements of the one-way speed of light, without reference to any clock synchronising scheme. You would be surprised how many people, including physicists, think that speed of light (one and two-way) has been experimentally confirmed to be isotropic, without any consideration of a clock synchronising scheme. Martin Hogbin (talk) 12:07, 5 June 2013 (UTC)
The JPL experiment was, I believe, mentioned in Will and Haugan's book, 'Modern Tests of Special Relativity' published in 1987 (I do not have a copy). That book (or possibly another by Will) was regarded by many physicists as the definitive work on experimental tests of SR. In the book Will said that the one-way speed of light had been measured independently of clock synchronisation. I think many physicists had their private doubts about this but it was not until the publication of Zhang's book that the issue was fully resolved. I think this is worth making clear in the article. Martin Hogbin (talk) 12:41, 5 June 2013 (UTC)
Obviously it must not be in the book I mention as the experiment was done in 1990. Martin Hogbin (talk) 13:18, 5 June 2013 (UTC)
Yes, I've re-inserted the remarks on Krisher et al. by Will and Zhang into the table. Though I don't think that another separate section is need, because it seems to me that this issue is now sufficiently discussed in the article (though it's not so clear regarding GRR). --D.H (talk) 19:09, 5 June 2013 (UTC)
I think you are hiding one of the important points about this subject in a an impenetrable technical section. There will be many readers who want to know the answer to a simple question, 'Has the one-way speed of light been measured (independently of any clock sychronising conventions)?'. We need to make clear that, although there are several sources which have made the claim that the one way speed of light has been measured, they are incorrect, insofar as they all depend on a conventional clock synchronisation scheme. It is, maybe, possible to work this out from your version but it is far from clear. I think we should revert to the earlier version but add something along the lines that there are additional experiments that set out,not to measure the one-way speed of light but to verify the equivalence between two conventions. Martin Hogbin (talk) 08:04, 6 June 2013 (UTC)

How about this?

I have separated out three experiments into their original section but retained the rest in your table. None of your text has been lost. Martin Hogbin (talk) 08:25, 6 June 2013 (UTC)

Exactly the same explanation in the clock-synch-comparison section also applies to Römer, Krisher (it they intended it or not). So the real importance of those three experiment is not mentioned anymore . It think at least the explanations should be merged with yours and included at the beginning of the section. --D.H (talk) 08:51, 6 June 2013 (UTC)
I have separated the three experiments not because of what they show but because of how they are often perceived. These three experiments are often considered, even by some physicists, to have measured the one-way speed of light independently of clock sychronisation. We need to make it as clear as we can to the general non-technical reader that this is not the case.
On further reflection about the subject it seems to me that attempts at one-way speed of light measurements can fail in two different ways. Either they are, in fact, two-way experiments (essentially equivalent to the Michelson-Morley experiment) and thus require no clock sychronisation scheme at all or they rely on slow transport of clocks. Should we not separate these two cases. Martin Hogbin (talk)
According to Anderson et al., practically all of those experiments (including Krisher et.) are clock-synch-comparison measurements. However, as shown by Finkelstein and others, Grieves et al. is an actual two-way experiment.--D.H (talk) 09:06, 6 June 2013 (UTC)
I think Anderson et al (I think John Anderson and Tom Roberts) might be wrong about some of them. Zhang, in his book, says about Riis and Krishner, 'Therefore these so-called one-way experiments, just as the closed path experiments, are only the tests of the isotropy of the two-way speed of light'.Martin Hogbin (talk) 09:19, 6 June 2013 (UTC)
There is no contradiction in those interpretations. All of the so-called one-way-clock-transport experiments are consistent with two-way speed of light. What is more important is the distinction between light following a closed path (no synchronization) or light following a unidirectional path (with Einstein or slow clock-transport synchronization).
Yes, that is exactly the distinction that I want to make but Anderson appears to say that Riis and Krishner rely on slow transport but Zhang seems to me to say that they are closed-path tests. In the case of Romer, Zhang specifically mentions slow transport. Martin Hogbin (talk) 09:38, 6 June 2013 (UTC)
I haven't access to Zhang's book at the moment, but I possess Zhang's 1995 paper on "test theories", where he wrote in relation to Will's paper: "Therefore, a test of the Manmuri—Sexl transformation is just a test of anisotropy af the two—way speed ot light (and a test or the parameter d), but not a test of anisotropy af the one-way speed of light. (The new type measurements reported in [17, Will 1992]) are tests involving unidirectional propagation along several baselines together with clock transport connecting the ends of each baseline. These kinds of experiments are related to the problem of slow transport of clocks. (Zhang, Yuan Zhong (1995). "Test theories of special relativity". General Relativity and Gravitation 27 (5): 475–493.)
Since Krisher et al. explicitly relies on slow clock-transport, I cannot imagine that Zhang said something in his book that directly contradicts his previous statement about the involvement of slow clock transport in such experiments. (Of course, all of those experiments are also a measurement of two-way speed of light, since this is the only synchronization independent speed that can be described in the Edwards- and Robertson-Mansouri-Sexl scheme, which was used by Zhang.). --D.H (talk) 10:56, 6 June 2013 (UTC)
That is interesting. The statement 'Therefore, a test of the Manmuri—Sexl transformation is just a test of anisotropy of the two—way speed of light (and a test or the parameter d), but not a test of anisotropy of the one-way speed of light' agrees with Zhang's book (which I do have a copy of). The second part, '(The new type measurements reported in [17, Will 1992]) are tests involving unidirectional propagation along several baselines together with clock transport connecting the ends of each baseline', seems not to agree with his book, where Zhang says of the JPL experiment, '...light signals were propagated simultaneously in both directions and phase comparisons made at both ends of the fiberoptic link...'. We need to find out more. Is there any way I can get a copy of Zhang's 1995 paper?. Martin Hogbin (talk) 15:25, 6 June 2013 (UTC)
See GBS preview on p. 395 where the paper was reprinted. And why do you think that this disagrees with what Zhang says above? It still don't see the problem. The expression "unidirectional propagation" only refers to the fact that signals from one clock are sent into the direction of the other clock (no return), while the other clock sends its signals into the opposite direction (no return as well). Of course, Zhang who used RMS came to the correct conclusion that this is still a measurement of the two-way speed. --D.H (talk) 16:06, 6 June 2013 (UTC)
See also the original paper by Krisher et al., where the role of slow-clock-transport due to Earth's rotation is explicitly described in terms of the RMS test theory (p. 733). Though there is a major difference between Will and Zhang: Will (and Krisher et al.) interpreted the RMS test theory as a means to measure the one-way speed of light, while Zhang (as well as Anderson et al.) pointed out that this is a misinterpretation of RMS. They showed that, if correctly applied, RMS cannot be used to determine the one-way speed of light. RMS can only used to determine the two-way speed as well as the equivalence between Einstein synchronization and slow-clock-transport. --D.H (talk) 16:42, 6 June 2013 (UTC)
There is no doubt that the JPL experiment used a travelling clock but this does not necessarily mean that the experiment actually measured the equivalence between Einstein synchronization and slow-clock-transport. Two beams of light travelling simultaneously in opposite directions is essentially the same as a closed path experiment. Martin Hogbin (talk) 22:55, 6 June 2013 (UTC)
On reflection, for the time being at least, I am going to agree with you. Roberts certainly knows about this kind of thing and he clearly says that this is a clock-sych test. Zhang, who I hoped would give a definitive answer, is not so clear but when I re-read what he wrote in his book, 'Therefore these so-called one-way experiments, just as the closed path experiments, are only the tests of the isotropy of the two-way speed of light', I see that using the words 'just as' can be taken to imply that the JPL experiment is not a closed-path one. Martin Hogbin (talk) 08:08, 7 June 2013 (UTC)

The GRR experiment

Having agreed on the JPL experiment I would like to change the heading wording for the GRR experiment. I have changed it to, 'Round trip experiments which claim to measure the one-way speed of light' but I think it would be better to change it to 'Experiments which claimed to use a one-way light signal but which actually made a round trip measurement'. What do you think? I have made the change and changed the order of the two sections. Are you happy with that? I think it makes things clearer. Martin Hogbin (talk) 22:47, 7 June 2013 (UTC)

Gagnon et al.

There is a "one-way" experiment that I have not seen satisfactorily explained as being either a closed path test of the isotropy of the speed of light in disguise or relying on slow-transport:

Gagnon, D.R.; Torr, D.G.; Kolen, P.T.; Chang, T. Guided-wave measurement of the one-way speed of light. Phys. Rev. A 38, 1767–1772 (1988)
"A new experiment has been performed as a step toward settling an unresolved issue in the testing of the special theory of relativity. The experiment is a test for velocity or direction-dependent variation of the one-way speed of light due to particular motion of a light source and an observer. The apparatus consists of a novel type of radio-frequency bridge, in the configuration of a Mach-Zender interferometer, one arm of which is a waveguide operated near its lowest cutoff frequency (at 40 GHz) to give a very large phase velocity. Clock-transport effects have been avoided by employing only one clock that effectively remains at rest in the laboratory. Our results have not yielded a measurable direction-dependent variation of the one-way speed of light. A clear null result is obtained for a hypothesis in which anisotropy of the cosmic background radiation is used to define a preferred reference frame."

Although the analysis given in the paper seems more than a little suspect to me, the experimental results are unambiguous. (Illegally scanned copies of the paper are available online with a little google searching.) I do not see where closed path or slow transport come into play. Stigmatella aurantiaca (talk) 10:44, 6 June 2013 (UTC)

Tom Roberts says about this paper in the physics FAQ'A guided-wave test of isotropy. Their null result is consistent with SR.'. Martin Hogbin (talk) 15:51, 6 June 2013 (UTC)
I agree with Roberts that this experiment belongs to experiments which "clearly use a one-way light path" but are "inherently unable to rule out a large class of theories in which the one-way speed of light is anisotropic ... In all of these theories the effects of slow clock transport exactly offset the effects of the anisotropic one-way speed of light". As Gagnon et al. said (p. 1768) "it measures interference between unidirectional continuous waves". So by definition slow-clock transport is necessary in order to understand the result in an assumed "preferred" frame. (Unfortunately, their own analysis (p. 1768) concerning clock transport and "absolute simultaneity" with respect to LET (or what they call GGT) is wrong, because conventionality of simultaneity practically applies to both models exactly the same way - the difference between SRT and LET is only metaphysical).
In other words: In terms of a test theory such as RMS/Zhang, they of course measured the isotropy of the two-way speed of light. --D.H (talk) 17:09, 6 June 2013 (UTC)
Their claim to be able to distinguish between LET and SRT was obviously wrong, but I still don't see anything resembling a clock synchronization procedure between the source and detectors. This is because the exact value of the phases relative to the source of the two outputs of wave guides was of no importance. The only thing important was that the phase difference between the two outputs was measured as constant with the apparatus oriented in different directions relative to Earth's motion in space or relative to the CMBR. Stigmatella aurantiaca (talk) 23:46, 6 June 2013 (UTC)
The claimed measurement is, of course, not that of the one way speed of light per se, but rather of possible direction-dependent variations in the one way speed of light. Measurement of OWLS necessarily requires a clock synchronization convention between the endpoints of the path. Detection of possible direction-dependent variations in OWLS does not necessarily require a clock synchronization convention or a closed signalling loop, just as detection of the winner of a horse race does not require accurate knowledge of the times that it took each horse to run the track, merely knowledge of who crossed the finish line first. Stigmatella aurantiaca (talk) 01:21, 7 June 2013 (UTC)
Measurement of OWLS in two different directions requires either synchronised spatially-separated clocks or clocks which move. Martin Hogbin (talk) 08:44, 7 June 2013 (UTC)
(1) They are not measuring OWLS per se, which obviously requires two clocks which must be synchronized by some convention. They are measuring direction dependent variations in OWLS. (2) Please identify the second clock in this experiment. (By "clock" I mean an autonomous, free-running oscillator having stability acceptable within the parameters of its intended usage. A "slave clock" in a master-slave clock system is not a "clock" by this definition, but merely a remote readout.) Stigmatella aurantiaca (talk) 09:12, 7 June 2013 (UTC)
First, "Possible direction-dependent variations in the one way speed of light" is simply another expression for anisotropy of the OWLS". Anything that has to do with one-way (an)isotropy is conventional by definition, because one has to define "direction dependence" of a velocity first in terms of certain time and space coordinates (which is implicitly always the case when one applies the laws of optics and electrodynamics), and one can always find a coordinate system in which any direction dependence of a specific phenomenon or velocity can be defined away.
Anyway, you are discussing the case when such an experiment gives a positive result. Of course, such a result (if no errors were made) would indicate Lorentz symmetry violation. However, saying that there is a Lorentz violation must be separated from its interpretation in terms of an anisotropy of the OWLS (as explained, the anisotropy of the latter can always be defined away). Of course, the Lorentz violation itself doesn't vanish by that, but is only moved into another part of a more general transformation, which indicates anisotropy of the two-way speed of light or nonequivalence between Einstein-transport-clock-synch-methods (such as in RMS test theory), or a more specific Lorentz violation in the Standard model (such as in SME test theory). --D.H (talk) 09:25, 7 June 2013 (UTC)
I would agree with that. Measurable anisotropy in one-way speed of light measurements must necessarily manifest itself as measurable anisotropy in two-way measurements as well. There is no escape from that direct mathematical consequence of the ensuing Lorentz violation, and so efforts directed towards making one-way measurements are mostly wasted effort when two-way measurements are, in general, far more precise.
The only thing I have objected to is attempting to discern a closed-loop or a synchronized second clock in the experiment under discussion. Stigmatella aurantiaca (talk) 09:49, 7 June 2013 (UTC)
As I have said above, Roberts put a lot of care and knowledge into his analysis of these matters and if he said it is a slow transport issue I tend to accept that. If you want to see Roberts' reasoning it might be on a web site somewhere, or you might try to contact him; let me know if you do. Martin Hogbin (talk) 12:09, 7 June 2013 (UTC)
I do not recollect him making that statement anywhere about the Gagnon experiment. However, I will state that long ago, I established to my own satisfaction that the Gagnon experiment, like other purported "one-way" experiments, is incapable of detecting the one-way light speed anisotropies of the Edwards' theory or of other theories experimentally indistinguishable from special relativity (TEIFSR). The inability of these "one-way" experiments to detect OWLS anisotropy cannot be universally ascribed to hidden signalling loops or to hidden assumptions about the dissemination of time (whether by slow transport or otherwise), but is a demonstrable consequence of working out the math. It is not necessarily a defect in how the experiments were set up (although that was certainly the case with Krisher et al. and of Greaves et al.). Measurable OWLS anisotropy necessarily manifests itself as measurable TWLS anisotropy. Given that TWLS anisotropies have never been measured, the OWLS anisotropy of TEIFSR theories is likewise, simply not measurable. Stigmatella aurantiaca (talk) 12:49, 7 June 2013 (UTC)
So what exactly is the problem? Roberts says, 'Their null result is consistent with SR.', which is correct. Any problems can only be in the claims of Gagnon. Their bizarre claim, 'A clear null result is obtained for a hypothesis in which anisotropy of the cosmic background radiation is used to define a preferred reference frame' makes no sense to me and unless they explain their reasoning somewhere can be ignored. Martin Hogbin (talk) 15:30, 7 June 2013 (UTC)
We are agreed that Gagnon's analysis is screwy and is best ignored. The experimental results, however, appear sound.
My problem is with the Wiki article statement: "Experiments that attempted to directly probe the one-way speed of light independent of synchronization have been proposed, but none has succeeded in doing so." Well, of course that would be true for OWLS per se, but the article implies that measurements of OWLS anisotropy cannot be made without assumptions of synchronization.
I agree that to measure OWLS as being 299792.45807±0.00013 km/s in different directions requires multiple clocks that have been synchronized according to some convention. However, I assert that to measure OWLS as being constant to within ±0.00013 km/s in different directions does not necessarily require multiple clocks.
The great majority of attempts to measure OWLS anisotropy have been flawed by overlooked round-trip loops or hidden assumptions concerning clock synchronization. However, when I examine the Gagnon experiment, I see no round-trip loops or multiple clocks. Where do you see a second clock? Stigmatella aurantiaca (talk) 16:51, 7 June 2013 (UTC)
Words such as "synchronization" and "clocks" are equivalent to assumptions about "simultaneity" and "space-time coordinates" implicit in standard laws of physics, which are used in any conceivable experiment. In particular, concepts such as "isotropy" or "anisotropy" of one-way speeds make actually no sense without such concepts, because any of such anisotropies can appear or disappear by a suitable choice of coordinate system. So yes, the article is correct when it says that is impossible to measure the isotropy of one-way light speed without prior (implicit or explict) "synchronization" and "clocks", neither by Gagnon et al. nor anybody else. --D.H (talk) 17:12, 7 June 2013 (UTC)
"Synchronization" requires a plurality of "clocks". I just need to be told where the second clock resides. Stigmatella aurantiaca (talk) 17:21, 7 June 2013 (UTC)
By analogy: Measurement of two-way light speed requires a clock. Yet Michelson and Morley placed limits on anisotropies in the two-way speed of light without use of any clocks at all. Stigmatella aurantiaca (talk) 17:26, 7 June 2013 (UTC)
It (the second clock) resides in the laws of optics and electrodynamics they used and the implicit one-way conventions those laws necessarily contain when one applies them. I'm repeating myself so I have not more to say on this subject. --D.H (talk) 17:33, 7 June 2013 (UTC)
We will need to agree to disagree then. Stigmatella aurantiaca (talk) 17:51, 7 June 2013 (UTC)

What is the problem?

Stigmatella, you agree, I presume, that the OWLS cannot be measured independently of a clock sychronisation scheme and that variations in OWLS with direction (OWLS isotropy) should also not be detectable according to SR. The experiment did not measure any anisotropy in the OWLS so it is in agreement with SR and there is no problem to explain.

The only issue is Gagnon's claim to have experimentally confirmed OWLS isotropy. I think there are three ways people can think that they have measured OWLs independently of clock synchronisation when in fact they have not. These are: they use slowly transported clocks; the path is actually a closed one; and, as I think is the case here, they (incorrectly) believe that they have managed to send information (in this case the phase of a signal) faster than light. It is well known that certain EM waves travel faster than light but the catch is that such waves cannot be used to carry information. This, I think, is where the error is in claims made for the the Gagnon experiment lie. Martin Hogbin (talk) 19:18, 7 June 2013 (UTC)

In a waveguide, phase velocity times group velocity equals c2. See microwaves101.com. Information travels at the speed of the group velocity, while the phase velocity is faster than light. As one lowers the frequency of a microwave beam approaching the cutoff frequency, the faster the phase velocity, which can reach many times the speed of light. Simultaneously, the group velocity, which represents the actual speed at which information can propagate through the waveguide, gets slower and slower.
A continuous carrier wave transmits essentially no information. They were using the high phase velocity carrier wave solely as a phase reference against which they could compare the phase of the low phase velocity carrier wave. They certainly did not believe that they were transmitting information faster than light.
They hypothesized that the phase of the high phase velocity carrier wave should be affected by "ether wind" much less than the phase of the low phase velocity carrier wave. This would be true for a classical, pre-Lorentzian ether.
It is, however, not true for a Lorentzian ether, nor is it true for SRT, nor is it true for theories experimentally indistinguishable from special relativity. Their analysis in terms of their homebrew GGT test theory was erroneous.
As I have emphasized before, any detection of anisotropy in OWLS at the level of sensitivity achievable in their experiment would have implied measurable anisotropies in TWLS at magnitudes that have already have been ruled out.
Stigmatella aurantiaca (talk) 20:54, 7 June 2013 (UTC)
They were not attempting to send information by modulating the phase. Instead, they sent a continuous carrier signal through each waveguide (near cutoff and far from cutoff), and measured the relative phases of the two outputs using a phase comparator. If motion through the hypothetical ether affected the one way velocity of light, they expected the relative phases of the two outputs to shift. Stigmatella aurantiaca (talk) 21:19, 7 June 2013 (UTC)
You say, 'They were using the high phase velocity carrier wave solely as a phase reference'. That is trying to send timing information faster than light, or, if you prefer, trying to synchronise two separated clocks via a faster than light signal. Martin Hogbin (talk) 21:43, 7 June 2013 (UTC)
I'm afraid that is a misconception of what it means to transmit information. Suppose they measured a constant phase difference of 30.142° between two pure carrier waves. To get five significant figures of accuracy, the phase comparator would have needed to integrate data over, say, a millisecond or so. That is 15 bits of information extracted over the course of a millisecond from a pair of waveguides maybe 10 meters long. That's not a very fast transmission rate of information.
An hour later, and they STILL measure a phase difference of 30.142° between the two beams. So all they have over the course of an hour is 15 bits of information.
A day later, and they STILL measure a phase difference of 30.142° between the two beams. So all they have over the course of a day is 15 lousy bits of information. That's worse than transmitting datagrams by carrier pigeon [3].
To transmit information, you have to modulate one beam relative to the other beam. Let us presume that you decide to modulate the high phase velocity beam which is going maybe 10 times the velocity of light. Guess what? Even though the phase velocity is 10x the speed of light, you discover that the modulations reach you at only 0.1x the speed of light.
But sending information by modulating beams isn't what Gagnon et al. were trying to do.
They were NOT trying to send information at faster-than-light speeds. They were merely trying to measure the phase difference between two constant carrier signals. Stigmatella aurantiaca (talk) 22:33, 7 June 2013 (UTC)
Also note, this is NOT a clock synchronization procedure, and certainly does NOT involve signals traveling at faster than the speed of light. Stigmatella aurantiaca (talk) 22:38, 7 June 2013 (UTC)
You cannot synchronize two clocks using a constant carrier wave. Stigmatella aurantiaca (talk) 22:45, 7 June 2013 (UTC)
I do not want to continue this discussion further here. The results were in agreement with SR. There is nothing more to report. I believe that Tom Roberts wrote a detailed description of why this experiment tells us nothing new. You might be able to find it somewhere. Martin Hogbin (talk) 22:54, 7 June 2013 (UTC)
I never claimed that Gagnon et al. told us anything new. I merely claimed that this was a measurement of OWLS anisotropy that had no obvious closed paths, and which apparently used only one clock and therefore was independent of clock synchronization conventions. If you can identify a closed path, or if you can identify a second clock, then I will admit my mistake. (Remember that by "clock" I mean an autonomous, free-running oscillator having stability acceptable within the parameters of its intended usage. A forced oscillator is not an independent clock.) Stigmatella aurantiaca (talk) 23:15, 7 June 2013 (UTC)
So, are you starting from the point that you agree the Gagnon experiment does not actually measure OWLS isotropy but you cannot find the error in its claim to do so? Martin Hogbin (talk) 07:56, 8 June 2013 (UTC)
  1. Measurements of OWLS isotropy tell us nothing new that measurements of TWLS isotropy have not already told us.
  2. Measurable anisotropy in OWLS implies measurable anisotropy in TWLS. I worked that out to my personal satisfaction a long time ago. I believe that I have seen supporting statements to that effect by reputable scholars, but I don't remember where.
  3. If I am correct about (2), then, since TWLS isotropy measurements are, in general, far more precise than OWLS isotropy measurements, OWLS isotropy measurements represent wasted effort. The only reason for performing them would be to appease the crackpot crowd.
  4. Here is a way to think about the Gagnon experiment. Two monochromatic beams derived from a single source travel divergent paths towards a common target, where they are combined to measure their phase difference.
  5. In other words, the Gagnon setup is nothing but a form of interferometer.
  6. Gagnon's analysis of their experiment was messed up. Their claim to be able to distinguish between LET and SRT was wrong. In general, neither OWLS nor TWLS anisotropy measurements are capable of distinguishing TEIFSR theories from SRT.
  7. If a pre-Lorentzian, absolute lumeniferous aether existed, the Gagnon setup would have detected OWLS anisotropy. But then, so would MMX.
  8. The OWLS anisotropies of LET and TEIFSR are not measurable anisotropies. See point (2).
  9. In summary, the Gagnon experiment was, in fact, valid in the sense of being capable of detecting the measurable OWLS anisotropy of, say, the pre-Lorentzian absolute aether, but it was incapable of detecting the nonmeasurable OWLS anisotropies of LET and TEIFSR. Like all other OWLS anisotropy experiments, the Gagnon experiment was wasted effort. Stigmatella aurantiaca (talk) 09:07, 8 June 2013 (UTC)
So we back to the original question, 'What is the problem?'. If the G setup is a form of interferometer then it is a closed path, two-way measurement. Martin Hogbin (talk) 09:44, 8 June 2013 (UTC)
It is an open path, one way measurement. See Gagnon et al.'s Figure 1. The klystron source at the right and the phase detector at the left are separated by the length of the waveguides, about 8 feet. Stigmatella aurantiaca (talk) 09:58, 8 June 2013 (UTC)
The basic topology of their setup could be likened to that of a Mach–Zehnder interferometer. Stigmatella aurantiaca (talk) 10:05, 8 June 2013 (UTC)
Either it is an interferometer or it is not. If it is an interferometer it is essentially a closed path experiment. Which are you saying that it is. Martin Hogbin (talk) 10:35, 8 June 2013 (UTC)
There is no return path of light or any other form of synchronization information to the source. There are many counterexamples to your statement. Look up the wiki article Mach–Zehnder interferometer. You can also look at the illustrations in the article Interferometry for the point diffraction interferometer, the lateral shearing interferometer, Young's experiment, Lloyd's mirror, and the Mach-Zehnder interferometer, as well as the article Common path interferometer for Fresnel's biprism. These are all open path. Stigmatella aurantiaca (talk) 10:56, 8 June 2013 (UTC)
I am still not sure what your problem is. Is it just that you know in principle that the experiment cannot actually be measuring OWLS isotropy but that you cannot see the flaw in the authors reasoning? Martin Hogbin (talk) 11:03, 8 June 2013 (UTC)
This all gets back to the article, which starts by making the perfectly true statement:
The "one-way" speed of light from a source to a detector, cannot be measured independently of a convention as to how to synchronize the clocks at the source and the detector.
The article, however, fails to distinguish between measurements of OWLS, versus measurements of OWLS isotropy.
All attempts to measure OWLS have been demonstrated defective due to the presence either of an unrecognized closed loop, or of an unrecognized dependence on a clock synchronization convention such as slow transport.
It is not at all clear that measurement of OWLS isotropy is subject to the same restrictions as the measurement of OWLS.
The Gagnon experiment appears to represent a counterexample. There is no recognizable closed loop in the experiment, nor does it appear to depend on a clock synchronization convention such as slow transport.
The Gagnon experiment is valid in the sense that it is perfectly capable of detecting the measurable OWLS anisotropy of a pre-Lorentzian absolute aether. Whoop-de-do. No OWLS or OWLS isotropy experiment is capable of detecting the nonmeasurable OWLS anisotropies of LET or TEIFSR. Stigmatella aurantiaca (talk) 11:35, 8 June 2013 (UTC)
So we agree that, according to the well verified and currently accepted theory of SR, no experiment can measure OWLS or OWLS isotropy independently of a clock synchronisation scheme (such as that inherent in SR).
I think that we must also agree that the Gagnon experimenters did claim to measure (a null value of) OWLS anisotropy independently of a clock synchronisation scheme. Are we in complete agreement so far? Martin Hogbin (talk) 12:01, 8 June 2013 (UTC)
We agree that according to the well verified and currently accepted theory of SR, no experiment can measure OWLS independently of a clock synchronisation scheme. OWLS isotropy is a different story.
We agree that the Gagnon experimenters did claim to measure (a null value of) OWLS anisotropy independently of a clock synchronisation scheme.
I assert that the claim appears to be valid, given that the setup was, in essence, a unidirectional interferometer experiment. Stigmatella aurantiaca (talk) 12:12, 8 June 2013 (UTC)
So are you saying that, according to SR, it is possible to measure OWLS anisotropy independently of a clock synchronisation scheme? Martin Hogbin (talk) 12:15, 8 June 2013 (UTC)
According to SR, there is no OWLS anisotropy to measure.
In LET, the OWLS anisotropy is "transformed away" so as not to be measurable by any experiment.
The same holds true for TEIFSR theories. The OWLS anisotropies in these theories are nonmeasurable, otherwise they would not be experimentally indistinguishable from SR.
In the case of classic, pre-Lorentzian aether theory, it is possible to measure OWLS anisotropy independently of a clock synchronization scheme.
In the case of other theories not equivalent to special relativity, I suppose it would depend on the theory. But most theories not equivalent to special relativity have been excluded to a high level of precision by other experiments.
Stigmatella aurantiaca (talk) 12:39, 8 June 2013 (UTC)
So why do you not agree that, according to the well verified and currently accepted theory of SR, no experiment can measure OWLS or OWLS isotropy independently of a clock synchronisation scheme (such as that inherent in SR)? Martin Hogbin (talk) 12:59, 8 June 2013 (UTC)
Let me rephrase that statement. No experiment can measure OWLS in a SR test theory independently of a clock synchronization scheme. However, multiple examples exist (Gagnon is certainly not the only one) of experiments that appear capable of measuring the magnitude of OWLS anisotropies in the said SR test theory without assumptions of clock synchronization.
Within the context of SR itself, it makes no sense to write about whether or not it it possible to measure OWLS anisotropy, since it simply doesn't exist within the context of that theory. Stigmatella aurantiaca (talk) 13:17, 8 June 2013 (UTC)

I understand that OWLS anisotropy does not exist (by definition) in SR. That means that according to SR, no experiment would be able to measure any OWLS anisotropy otherwise it would be in disagreement with SR. Agreed?Martin Hogbin (talk) 13:34, 8 June 2013 (UTC)

Agreed. Stigmatella aurantiaca (talk) 13:41, 8 June 2013 (UTC)
So now we need to talk about Gaydon's claims. AS I understand it, they have claimed to have measured OWLS anisotropy but to have measured a null result. In other words, their claim is that their experiment was capable of measuring OWLS anisotropy had there been any. Do you agree with that? Martin Hogbin (talk) 13:48, 8 June 2013 (UTC)
That was their claim, which had a certain range of validity. The OWLS anisotropy of LET is not measurable, since it is transformed away. The OWLS anisotropies of TEIFSR theories in general are not measurable, since they are all transformed away. Gagnon et al. were wrong in thinking that they could distinguish between LET and SRT on the basis of OWLS anisotropy. However, Gagnon et al. should have been able to detect the OWLS anisotropies predicted by various theories not equivalent to SRT, including classic aether theory. Stigmatella aurantiaca (talk) 14:20, 8 June 2013 (UTC)
What is the point of that? The MMX would have detected OWLS anisotropies predicted by various theories not equivalent to SRT, including classic aether theory, but it did not. That is why such theories are no longer used, they are not in agreement with experiment. The Gaydon result was in agreement with SR.
THAT IS WHAT I HAVE BEEN SAYING!!! Please re-read my entire chain of argument. You have consistently been missing my point. There is no point to OWLS anisotropy experiments, since OWLS anisotropy implies TWLS anisotropy, and TWLS anisotropy experiments are generally more precise than OWLS anisotropy measurements. Stigmatella aurantiaca (talk) 14:47, 8 June 2013 (UTC)
So what exactly was the point of Gaydon's experiment and what do you believe is its significance?
The Gagnon experiment is an example of an OWLS anisotropy experiment that does not obviously contain any hidden closed loops and which does not appear to use any hidden clock synchronization assumptions. It is capable of measuring the extent of OWLS anisotropy in various theories not equivalent to SR. Stigmatella aurantiaca (talk) 17:52, 8 June 2013 (UTC)
If I have been missing your point that is because you have not explained very clearly what it is. Martin Hogbin (talk) 16:36, 8 June 2013 (UTC)
On 09:49, 7 June 2013, I wrote: "Measurable anisotropy in one-way speed of light measurements must necessarily manifest itself as measurable anisotropy in two-way measurements as well. There is no escape from that direct mathematical consequence of the ensuing Lorentz violation, and so efforts directed towards making one-way measurements are mostly wasted effort when two-way measurements are, in general, far more precise. The only thing I have objected to is attempting to discern a closed-loop or a synchronized second clock in the experiment under discussion."
On 09:07, 8 June 2013, I wrote: "1. Measurements of OWLS isotropy tell us nothing new that measurements of TWLS isotropy have not already told us. 2. Measurable anisotropy in OWLS implies measurable anisotropy in TWLS. I worked that out to my personal satisfaction a long time ago. I believe that I have seen supporting statements to that effect by reputable scholars, but I don't remember where. 3. If I am correct about (2), then, since TWLS isotropy measurements are, in general, far more precise than OWLS isotropy measurements, OWLS isotropy measurements represent wasted effort. The only reason for performing them would be to appease the crackpot crowd."
How could I have been more clear??? Stigmatella aurantiaca (talk) 17:45, 8 June 2013 (UTC)
So now to Gaydon's claim that they measured OWLS anisotropy independently of any clock synchronisation scheme. Do you believe that that claim is justified? What about the experiment that I describe below. Would that measure OWLS anisotropy independently of any clock synchronisation scheme? Martin Hogbin (talk) 14:29, 8 June 2013 (UTC)
You have problems with relativity of simultaneity, among other things. By the way, you could just as well have used a flash of light rather than a rotating laser.
Are you claiming that a Mach-Zehnder interferometer requires clock synchronization to work? Are there hidden clock synchronization assumptions in Young's double-slit experiment?
Where is the second clock in Young's double-slit experiment? Stigmatella aurantiaca (talk) 14:47, 8 June 2013 (UTC)
I only asked some simple questions. Do you believe that that Gaydon's claim that they measured OWLS anisotropy independently of any clock synchronisation scheme is justified? Martin Hogbin (talk) 16:36, 8 June 2013 (UTC)
Yes. If you wish to dispute that, you need to tell me what clocks are being synchronized. The Gagnon experiment is nothing more than a microwave variant of the Mach-Zehnder interferometer. An unmodulated, monochromatic carrier wave is split into two beams that travel divergent paths towards a common, remote target, and the phase difference between the two beams is measured at the said remote target. Stigmatella aurantiaca (talk) 17:45, 8 June 2013 (UTC)

Gentlemen, please, this is not a forum. --D.H (talk) 17:53, 8 June 2013 (UTC)

Thank you. Let's just agree to disagree. Hopefully this discussion will help in the writing of the article, having brought up points that I think were neglected. Please remember that there are other OWLS anisotropy experiments listed in Roberts' article that do not obviously make use of clock synchronization assumptions.
By the way, I really like your latest additions! Stigmatella aurantiaca (talk) 18:06, 8 June 2013 (UTC)

How to not measure the one-way speed of light (and not to determine if it is isotropic)

All that is needed is a single clock, a pulse of light, and something that travels faster than light. Luckily there are plenty of things that travel faster than light, the motion of a spot of light, for example. So, set up a laser on the Moon to sweep a spot of light very rapidly across the Earth's surface, say from London to New York City. The spot can easily be made to travel at 1000 c. Start a light pulse at London when the light spot passes by. In NY start your clock when the light spot passes (which will be very nearly the time the pulse left London) and stop it when the light pulse reaches you. OWLS measured!

To measure OWLS isotropy the experiment works even better. Repeat the experiment as the Earth rotates and when the Moon is conveniently located. Because the time delay for the spot will not change with direction we can ignore it altogether and just look at the difference in times measured by the clock in NY.

Of course, the above experiment is just a conjuring trick. It actually measures nothing but the round trip light speed. Martin Hogbin (talk) 10:31, 8 June 2013 (UTC)

SME tests

Let's not forget that since 2002, dozens of tests of Lorentz invariance (all of them published in leading physics journals) have been performed using the Standard-Model Extension (SME) and some other dynamical test theories, which practically apply to the whole of particle physics (see also modern searches for Lorentz violation). In particular, SME includes coefficients related to two-way and one-way light speed variations. Fortunately, Kostelecky (the founder of SME) was careful in pointing out the conventional character of some of those light speed variations, without diminishing the value of the experiments as tests of Lorentz invariance symmetry. Therefore I've included a new section on "Experiments on two-way and one-way speeds using the Standard-Model Extension", and also included some details into "Experiments that can be done on the one-way speed of light" regarding "vacuum dispersion" and "vacuum birefringence", for some speed variations that are not conventional. --D.H (talk) 11:02, 8 June 2013 (UTC)

You are on your own with this one as far as I am concerned; I am not familiar with SME although I think you should make clear that these experiments relate to an, as yet, unconfirmed theory.

Are you happy with my slight changes to the article? I would like to try and make some of the wording more accessible to general readers. Martin Hogbin (talk) 11:09, 8 June 2013 (UTC)

SME is used as a test theory for the Standard Model including Lorentz violations. No Lorentz violations have been found, thus of course it's not "confirmed". But this lies in the nature of a "test theory".
Regarding your changes: I agree with them, only the section titles are probably getting too long. --D.H (talk) 11:19, 8 June 2013 (UTC)
Yes, I understand. All I am trying to do is distinguish between experiments which claim to disagree with the well-established theory of SR and those that are attempting to move physics forward. I appreciate though that there may be no clear distinction between these two concepts except in the intent of the experimenters. Martin Hogbin (talk) 11:53, 8 June 2013 (UTC)

The two-way speed

This section states:

"The two-way speed of light is the average speed of light from one point, such as a source, to a mirror and back again. Because the light starts and finishes in the same place only one clock is needed to measure the total time, thus this speed can be experimentally determined independently of any clock synchronization scheme. Any measurement in which the light follows a closed path is considered a two-way speed measurement."
"Many tests of special relativity such as the Michelson–Morley experiment and the Kennedy–Thorndike experiment have shown within tight limits that in an inertial frame the two-way speed of light is isotropic and independent of the closed path considered."

Juxtaposing these two paragraphs results in the apparent implication that since the light starts and finishes in the same place in the MMX and KT experiments, one clock is needed to measure the total time. Please clarify in the article text whether that is your intended meaning. Stigmatella aurantiaca (talk) 21:14, 12 June 2013 (UTC)

MMX and KT are two-way isotropy experiments comparing the rates of two clocks. (see Will, arXiv:gr-qc/0504085, p. 8, "clock anisotropy experiment"). Is it better now? --D.H (talk) 21:19, 12 June 2013 (UTC)
Using this method of counting, Fizeau's measurement of the speed of light also used two clocks (1) the light clock defined by the two-way light path, and (2) the clock used to monitor and control the speed of the rotating sprocket assembly. Therefore both two-way speed of light measurements and two-way isotropy measurements require two clocks, in disagreement with the first paragraph that only one clock is needed. Stigmatella aurantiaca (talk) 21:36, 12 June 2013 (UTC)
Well, think about the case of only one light-clock, counting the pings on a mirror: Is there a second clock required to count the pings? I don't think so. --D.H (talk) 21:44, 12 June 2013 (UTC)
For measuring the speed of light, you need (1) a light clock and (2) some means of counting the pings per unit of time, i.e. a second clock. Just counting the pings doesn't get you a speed-of-light measurement. Stigmatella aurantiaca (talk) 21:50, 12 June 2013 (UTC)
The pings themselves can used by me to constitute a unit of time which also defines a velocity when the light-travel-distance is considered, (even this distance can be defined using this single clock alone), I don't need another clock for this purpose. Of course, to compare this result with other measurements, or to find out irregularities or anisotropies in-between the pings, we need another observer/clock who counts the pings, or compare it with another light-clock as in MMX. --D.H (talk) 22:17, 12 June 2013 (UTC)
Therefore, a second clock is needed to compare speed of light measurements obtained with a light clock against other measurements, for a total of two clocks, and MMX uses two clocks in the measurement of light speed anisotropy. How may we compare this with the statement, "Because the light starts and finishes in the same place only one clock is needed to measure the total time, thus this speed can be experimentally determined independently of any clock synchronization scheme." Stigmatella aurantiaca (talk) 22:41, 12 June 2013 (UTC)
The section you quote is about the measurement of the speed of light itself (which needs only one clock as explained above), while the next paragraph (with MMX and KT) is about it's isotropy, which evidently needs a second clock-standard to compare with. What's the problem? --D.H (talk) 22:51, 12 June 2013 (UTC)
Why does a round trip light path in MMX count as a light-clock, but a round trip light path in a speed of light measurement NOT count as a light-clock? Stigmatella aurantiaca (talk) 22:56, 12 June 2013 (UTC)
I can't follow you. Above I talked about a minimum (very impractical but possible) requirement for measuring/defining the speed of light with one clock. Of course, most (practical) round-trip measurements employ more clocks, i.e. at least one "ordinary" clock together with some sort of light-clock for comparison. --D.H (talk) 23:06, 12 June 2013 (UTC)
You have hit upon exactly my point. A speed of light measurement requires one "ordinary" clock capable of relating the measurements obtained with the light-clock against the common time standard. MMX requires zero "ordinary" clocks, since the measurements obtained with its two light-clocks do not have to be related to the common time standard.
The statement, "Because the light starts and finishes in the same place only one clock is needed to measure the total time, thus this speed can be experimentally determined independently of any clock synchronization scheme," refers to the number of "ordinary" clocks required in the measurement. Stigmatella aurantiaca (talk) 23:16, 12 June 2013 (UTC)
Nowhere in the article it is stated that this definition is only related to ordinary clocks as opposed to light clocks - a clock is a clock (maybe a more detailed account for the complexity of clock definitions would be useful). As explained above, a speed of light measurement in principle doesn't require an ordinary clock because the measurement/definition can be done with one light clock alone (even though practically we mostly use more than one). I have the feeling that our discussion is there again where it started.
However, I have enormous respect and admiration for your work in Wikipedia, so it's hard to believe that we cannot come to an agreement. So please try to formulate an (in your view) improved reformulation of this passage here on the talk page, so that we can talk about modifications and probably its inclusion. --D.H (talk) 23:43, 12 June 2013 (UTC)

Trial wording

The two-way speed of light is the average speed of light from one point, such as a source, to a mirror and back again. Because the light starts and finishes in the same place, only one clock is needed to measure the total time, thus this speed can be experimentally determined independently of any clock synchronization scheme. Any measurement in which the light follows a closed path is considered a two-way speed measurement.

It should be noted, however, that experiments measuring the two-way speed of light represent a distinctly different class of experiment than experiments measuring directional anisotropies in the two-way speed of light.

Instead of measuring the time that it takes a modulated beam of light to traverse a single round-trip path, an experiment designed to measure two-way light speed anisotropy may compare the times that it takes an unmodulated beam of light to travel from a common origin out to two different destinations, and then back again to the common origin. Such was the principle underlying the Michelson–Morley and the Kennedy–Thorndike experiment and repetitions of these experiments up through the 1930s. These experiments measured the relative difference in time that it took light to travel two diverse round-trip paths, and an external clock was unnecessary to detect deviations of the experiments from a null result or to measure the total travel times of the unmodulated light beams. These and similar tests can also be interpreted as being "clock anisotropy experiments", since every arm of a Michelson interferometer can be seen as a light clock having a specific rate.[1] Beginning in the 1950s and up to the present, the most sensitive repetitions of the Michelson-Morley experiment have measured the relative beat frequencies between two differently oriented optical resonators.[2][3][4] Likewise, the most sensitive repetitions of the Kennedy-Thorndike experiment have measured the relative beat frequencies between an optical resonator, representing a length standard, versus a time standard such as an iodine vapor frequency standard, or a hydrogen maser traceable to cesium and rubidium fountain clocks.[5][6][7]

Repetitions of the Michelson–Morley experiment and the Kennedy–Thorndike experiment have shown within tight limits that in an inertial frame, the two-way speed of light is isotropic and independent of the closed path considered, in complete agreement with the predictions of special relativity.

Stigmatella aurantiaca (talk) 02:33, 13 June 2013 (UTC)

I've included some suggestions. --D.H (talk) 08:02, 13 June 2013 (UTC)
It's looking good overall. It will need a few added references, of course. Martin, what do you say so far? Stigmatella aurantiaca (talk) 09:52, 13 June 2013 (UTC)
My list of added references is a bit idiosyncratic, based much more on my familiarity with them rather than their relative importance. Feel free to modify the list, which I believe should be representative, not exhaustive. Stigmatella aurantiaca (talk) 12:23, 13 June 2013 (UTC)
Where are you proposing this text for? Martin Hogbin (talk) 13:02, 13 June 2013 (UTC)
Section "The two-way speed". The paragraph on the meter definition needs to be in a separate section. Stigmatella aurantiaca (talk) 13:58, 13 June 2013 (UTC)
Suggested version

Below is my suggested version with some simplification of wording and removal of some detail.

The two-way speed of light is the average speed of light from one point, such as a source, to a mirror and back again. Because the light starts and finishes in the same place, only one clock is needed to measure the total time, thus this speed can be experimentally determined independently of any clock synchronization scheme. Any measurement in which the light follows a closed path is considered a two-way speed measurement. To measure the two-way speed of light, early experimenters such as Fizeau (1849), Foucault (1862), and Michelson (1926) reflected modulated beams of light off a remote mirror,[8] while the latest determinations made prior to the 1983 redefinition of the meter employed direct frequency and wavelength measurements at terahertz frequencies.[9]

Two-way experiments can also measure only the variation in the speed of light with direction (directional anisotropy). Experiments measuring directional anisotropies in the two-way speed of light represent a distinctly different class of experiment from use different methods than experiments measuring the two-way speed of light. Experiments to measure directional anisotropies may compare the times that it takes an unmodulated a beam of light to travel from a common origin to two different destinations, and then back again to the common origin. This was the principle underlying the Michelson–Morley and Kennedy–Thorndike experiments and repetitions of these experiments up through the 1930s. These experiments measured only the difference in time that it took light to travel two diverse round-trip paths, making a clock for measuring total time unnecessary. These and similar tests can also be interpreted as being "clock anisotropy experiments", since every arm of a Michelson interferometer can be seen as a light clock having a specific rate.[1] Beginning in the 1950s and up to the present, the most sensitive repetitions of the Michelson-Morley experiment have measured beat frequencies between two differently oriented optical resonators.[2][3][4] Likewise, the most sensitive repetitions of the Kennedy-Thorndike experiment have measured beat frequencies between an optical resonator, which represents a length standard, against an accurate time standard.[5][6][7]

These experiment have shown within tight limits that in an inertial frame, the two-way speed of light is isotropic and independent of the closed path considered, in complete agreement with the predictions of special relativity. Martin Hogbin (talk) 16:20, 13 June 2013 (UTC)

I'm OK with most of your wording simplifications, but suggest putting back the wording in dark red. Stigmatella aurantiaca (talk) 19:35, 13 June 2013 (UTC)
I think you are maybe carrying on our discussion/argument or maybe some other discussion in the text. My text makes clear that they are different experiments with different objectives. I think, 'distinctly different class' is unnecessary; who has ever suggested that they are the same.
Your text does not make clear that they are different, which amounts to the same thing.
Currently, your text reads:
"The two-way speed of light is the average speed of light from one point, such as a source, to a mirror and back again. Because the light starts and finishes in the same place only one clock is needed to measure the total time, thus this speed can be experimentally determined independently of any clock synchronization scheme. Any measurement in which the light follows a closed path is considered a two-way speed measurement."
"Many tests of special relativity such as the Michelson–Morley experiment and the Kennedy–Thorndike experiment have shown within tight limits that in an inertial frame the two-way speed of light is isotropic and independent of the closed path considered."
Juxtaposing these two paragraphs results in the apparent implication that since the light starts and finishes in the same place in the MMX and KT experiments, one clock is needed to measure the total time. Stigmatella aurantiaca (talk) 03:27, 14 June 2013 (UTC)
Similarly, 'unmodulated'; I have never heard any suggestion that MM modulated their beam, and it would not actually matter if they had. Martin Hogbin (talk) 21:30, 13 June 2013 (UTC)
Classic two-way speed of light determinations such as Fizeau, Foucault, Michelson etc. used modulated beams as distinct from the unmodulated beams used in MMX, KT etc. The types of measurements performed in TWLS determinations are distinctly different from the types of measurement performed in TWLS anisotropy measurements. Stigmatella aurantiaca (talk) 22:06, 13 June 2013 (UTC)
I imagine that you are also going to propose something for the 'One-way' section. Maybe we will not agree there but let us keep our disagreement to the talk pages. Martin Hogbin (talk) 21:32, 13 June 2013 (UTC)
No, I'm not going to bother to. It is sufficient if I've gotten you to ponder what the difference may be between an OWLS measurement versus an OWLS anisotropy measurement. Your current text jumbles the two together. OWLS anisotropy measurements are rather heterogeneous in concept. The most accurate OWLS anisotropy measurement that I am aware of is Ragulsky (1997), even though it is superficially two-way. On the other hand, I consider Roberts' inclusion of Mössbauer rotor experiments among the OWLS experiments to be somewhat problematical. I consider them tests of Lorentz violation, but don't exactly see how they could be classified as tests of OWLS isotropy. Stigmatella aurantiaca (talk) 22:06, 13 June 2013 (UTC)
Further additions in teal. Stigmatella aurantiaca (talk) 02:42, 14 June 2013 (UTC)
Could you try to word your red sentence in a way that does not reflect a perceived argument with me. Maybe something like,' there are two possible types of two-way measurement of the speed of light; some measure the speed of light whilst other attempt only to detect only a change in speed with direction'.
Your distinction between modulated and unmodulated beams is somewhat arbitrary and, as far as I know, there is no distinction between two classes of experiment made in the literature. Better to describe on an experiment by experiment basis. Much of what you talk about is mentioned in the Speed of light article. This article is intended to be specifically about the one-way speed and the issues surrounding it. Obviously mention of the two way speed is required but details are not needed. Martin Hogbin (talk) 08:38, 14 June 2013 (UTC)
Latest changes in blue. Details about the two-way measurements are needed, since they do not all make use of clocks for timing comparisons against an external standard. For a balanced treatment, I encourage you to add detail to the following section about one-way measurements, since one-way measurements do not all have hidden loops or make use of hidden synchronization assumptions. I suggest that you provide a brief discussion of how the Isaak et al.(1970), Turner and Hill (1988), and Ragulsky (1997) experiments worked, which Roberts cited as providing limits on one-way anisotropy of 3 m/s, 10 m/s, and 0.13 m/s, respectively. These three would be OWLS anisotropy measurements, not OWLS measurements. None of these three have obvious hidden synchronization assumptions, and the loop in Ragulsky is asymmetric, with the outgoing and return rays traveling in different media.
I note that in a separate section with the verbose title, "Experiments which claimed to use a one-way light signal but which actually made a round trip measurement", you go into specific detail about Greaves et al (2009), Wills (1990), and Romer (1676). Stigmatella aurantiaca (talk) 11:27, 14 June 2013 (UTC)
You are most welcome to propose a more concise title that describes the content of the section accurately, that wa the best that I could do at the time. Martin Hogbin (talk) 09:21, 15 June 2013 (UTC)
The Mössbauer-Rotor experiments obviously use slow-clock transport of the clocks (the rotating clock at the rim as opposed to clocks at the emitter, or vice versa). This was pointed out by Mansouri-Sexl and more recently by by Anderson et al., thus by definition, they rest on the same one-way convention as Einstein synchronization. --D.H (talk) 11:54, 14 June 2013 (UTC)
Slow clock transport is relevant when clock synchronization is used for timing purposes. However, the Mössbauer "clocks" are used in testing resonance, and are not used for timing. As I have stated previously, I have always been uncomfortable with classification of the Mössbauer-Rotor experiments as light speed anisotropy experiments. In my personal opinion, they do no more than test the principles of translational and rotational invariance. However, that would be my disallowed "original research" opinion. Stigmatella aurantiaca (talk) 12:23, 14 June 2013 (UTC)

Getting back to the question of the proposed text, I do not think mention of modulated/unmodulated serves any real purpose and no class distinction is made in any secondary sources that I am aware of. I therefore think we should drop this wording.

I have deemphasized the distinction, but the fact remains that the classic measurements by Fizeau, Foucault and Michelson made use of toothed wheels or moving mirrors to chop/pulse/sweep the beams.

Regarding speed vs speed anisotropy measurements, I do agree that there is a clear distinction but we do not need to overstate it.

I have merged paragraphs rather than setting them off as individual paragraphs. Is this OK?

How about my revised proposal above, 'There are two possible types of two-way measurement of the speed of light; some measure the speed of light whilst other attempt only to detect only a change in speed with direction'. Martin Hogbin (talk) 09:21, 15 June 2013 (UTC)

I see no improvement in the proposed wording.

Stigmatella, I do not much like your latest proposal, 'Experiments measuring directional anisotropies in the two-way speed of light use different methods than experiments measuring the two-way speed of light'. You still seem to me trying to make some point that I do not understand. Obviously, experiments with different purposes are designed differently from one another. Martin Hogbin (talk) 09:33, 15 June 2013 (UTC)

The original wording as it currently stands in the article clearly implies that ALL two-way speed of light measurements, including MMX and KT, are dependent on use of a clock to measure the total time. Please read over the section carefully. "The two-way speed of light is the average speed of light from one point, such as a source, to a mirror and back again. Because the light starts and finishes in the same place only one clock is needed to measure the total time, thus this speed can be experimentally determined independently of any clock synchronization scheme. Any measurement in which the light follows a closed path is considered a two-way speed measurement. Many tests of special relativity such as the Michelson–Morley experiment and the Kennedy–Thorndike experiment have shown within tight limits that in an inertial frame the two-way speed of light is isotropic and independent of the closed path considered." Stigmatella aurantiaca (talk) 11:25, 15 June 2013 (UTC)
The next sentence clearly explains, Michelson-Morley type experiments are sometimes called "clock anisotropy experiments", since every arm of a Michelson interferometer can be seen as a light clock having a specific rate, whose orientation dependence can be tested.[6]. The distinction that you make is therefore not quite so clear cut as you may think, especially bearing in mind the definition of the metre. In SI units the speed of light cannot be measured at all. There are experiment in which an 'independent' clock is used and those in which a light clock is used. There are also experiments in which a physical object was used as a length standard. The two-way speed is covered in the Speed of light article let us concentrate on the one-way speed here. Martin Hogbin (talk) 11:44, 15 June 2013 (UTC)
If you will re-read the dialog between D.H and myself, you will see that we needed to resort to distinguishing between "ordinary" clocks used to relate time-of-flight measurements to external standards, versus the light-clocks represented by the interferometer arms. The distinction was very clear. TWLS measurements require an "ordinary" clock to measure time-of-flight, while MMX-style TWLS anisotropy measurements do not make use of "ordinary" clocks. The latest KT-style experiments use frequency standards, but not for the purpose of relating time-of-flight to external standards. The SI definition of the meter has no bearing on the ongoing efforts to detect the TWLS anisotropies suggested by the latest theories attempting to unify gravitation with the rest of physics.
"The two-way speed is covered in the Speed of light article let us concentrate on the one-way speed here." The fact that this is an OWLS article is no excuse for sloppy coverage of TWLS and TWLS-anisotropy measurements. The following section on OWLS needs to be upgraded to match the proposed coverage of TWLS, rather than the TWLS section being left in a downgraded state. Stigmatella aurantiaca (talk) 12:19, 15 June 2013 (UTC)
I still do not see the purpose of the statement, 'Experiments measuring directional anisotropies in the two-way speed of light use different methods than experiments measuring the two-way speed of light' It seems to, rather inaccurately, state the obvious. It simply says that experiments to measure different things use different methods.
Also, there can no longer be experiments to measure the speed of light, experiments of the type that now are actually measurements of the length of a a physical object using the two-way speed of light.
What about having two sub headings under which we group experiments, 'Experiments to measure TWLS anisotropy' and something like 'Measurements of TWLS/Measurements of the length of a a physical object using the two-way speed of light'? Martin Hogbin (talk) 17:03, 15 June 2013 (UTC)

References

  1. ^ Will, C.M (2006) [2005]. "Special Relativity: A Centenary Perspective". In T. Damour, O. Darrigol, B. Duplantier und V. Rivasseau (ed.). Poincare Seminar 2005. Basel: Birkhauser. pp. 33–58. arXiv:gr-qc/0504085.{{cite book}}: CS1 maint: multiple names: editors list (link)
  2. ^ Wolf, P.; Bize, S.; Clairon, A.; Santarelli, G.; Tobar, M. E.; Luiten, A. N. (2004). "Improved test of Lorentz invariance in electrodynamics". Physical Review D. 70 (5): 051902. arXiv:hep-ph/0407232. Bibcode:2004PhRvD..70e1902W. doi:10.1103/PhysRevD.70.051902.{{cite journal}}: CS1 maint: multiple names: authors list (link)
  3. ^ Eisele, Ch.; Nevsky, A. Yu.; Schiller, S. (2009). "Laboratory Test of the Isotropy of Light Propagation at the 10−17 level" (PDF). Physical Review Letters. 103 (9): 090401. Bibcode:2009PhRvL.103i0401E. doi:10.1103/PhysRevLett.103.090401. PMID 19792767.{{cite journal}}: CS1 maint: multiple names: authors list (link)
  4. ^ Herrmann, S.; Senger, A.; Möhle, K.; Nagel, M.; Kovalchuk, E. V.; Peters, A. (2009). "Rotating optical cavity experiment testing Lorentz invariance at the 10−17 level". Physical Review D. 80 (100): 105011. arXiv:1002.1284. Bibcode:2009PhRvD..80j5011H. doi:10.1103/PhysRevD.80.105011.{{cite journal}}: CS1 maint: multiple names: authors list (link)
  5. ^ Hils, Dieter; Hall, J. L. (1990). "Improved Kennedy–Thorndike experiment to test special relativity". Phys. Rev. Lett. 64 (15): 1697–1700. Bibcode:1990PhRvL..64.1697H. doi:10.1103/PhysRevLett.64.1697. PMID 10041466.{{cite journal}}: CS1 maint: multiple names: authors list (link)
  6. ^ Braxmaier, C.; Müller, H.; Pradl, O.; Mlynek, J.; Peters, A.; Schiller, S. (2002). "Tests of Relativity Using a Cryogenic Optical Resonator" (PDF). Phys. Rev. Lett. 88 (1): 010401. Bibcode:2002PhRvL..88a0401B. doi:10.1103/PhysRevLett.88.010401. PMID 11800924.{{cite journal}}: CS1 maint: multiple names: authors list (link)
  7. ^ Tobar, M. E.; Wolf, P.; Bize, S.; Santarelli, G.; Flambaum, V. (2010). "Testing local Lorentz and position invariance and variation of fundamental constants by searching the derivative of the comparison frequency between a cryogenic sapphire oscillator and hydrogen maser". Physical Review D. 81 (2): 022003. arXiv:0912.2803. Bibcode:2010PhRvD..81b2003T. doi:10.1103/PhysRevD.81.022003.{{cite journal}}: CS1 maint: multiple names: authors list (link)
  8. ^ Gibbs, P (1997). "How is the speed of light measured?". Usenet Physics FAQ. University of California, Riverside. Retrieved 2010-01-13.
  9. ^ Evenson, KM (1972). "Speed of Light from Direct Frequency and Wavelength Measurements of the Methane-Stabilized Laser". Physical Review Letters. 29 (19): 1346–49. Bibcode:1972PhRvL..29.1346E. doi:10.1103/PhysRevLett.29.1346. {{cite journal}}: Unknown parameter |coauthors= ignored (|author= suggested) (help)

More on speed and isotropy

  • Preliminarily, before you come to an agreement for the two-way section, I've included the wording: "Isotropy experiments of the Michelson-Morley type do not use an external clock to directly measure the speed of light, but rather compare two internal frequencies or clocks."
  • I've included the one-way speed versus isotropy arguments of Will and Anderson in the one-way section.
  • The start of the synchronization section was expanded, as well as for slow-clock transport.
  • The related content of non-standard synch., slow-clock transport and time dilation was merged. --D.H (talk) 10:09, 16 June 2013 (UTC)
Your new additions accomplish practically all of what I wanted to accomplish. Thank you!
Thank you also for including details of the Will vs. Anderson debate. I had been preparing a review of this debate as part of my response to Martin. Your addition makes this unnecessary. Thank you again!
To Martin: I think we can close out our discussion at this point. Stigmatella aurantiaca (talk) 10:43, 16 June 2013 (UTC)
As you will see below I agree with DL's, strictly correct, statement but I think it gives a very misleading impression. Martin Hogbin (talk) 11:39, 16 June 2013 (UTC)

Experiments in which light follows a unidirectional path

I think this section gives too much emphasis to the claims of Mansouri and Sexl (1977) and Will (1992). Whereas it is correct to say that, 'only if inertial frames and coordinates are defined so that space and time coordinates as well as slow clock-transport are described isotropically, the null result of these experiments can be interpreted in terms of isotropy of the one-way speed of light' this statement can, in my opinion, only be described as perversely pointless. If we allow arbitrarily complex coordinate systems then the speed of light and indeed any measurement relating to it can be pretty much anything that we wish.

If we must make a statement of this nature it would, in my opinion, be much better to say it in reverse so to speak:

It has been shown that experiments cannot, in any meaningful way, measure the anisotropy of the one way speed of light unless inertial frames and coordinates are defined so that space and time coordinates as well as slow clock-transport are described isotropically. Martin Hogbin (talk) 11:31, 16 June 2013 (UTC)

Ok, I've included your proposal. --D.H (talk) 11:42, 16 June 2013 (UTC)
I'm not participating in this debate. The sections on The two-way speed and The one-way speed have turned out pretty much as I would have wanted them, so it is between you and D.H to figure out how to make the rest of the article consistent. :-) Stigmatella aurantiaca (talk) 11:56, 16 June 2013 (UTC)
An inconvenience remaining is probably the circumstance, that some things are repeated in different sections. But maybe that's good for readers who only look for specific headlines. I think the whole question concerning the conventionality of the one-way speed of light in different coordinate systems is not whether it's true or not, but whether it's trivial/unimportant or not. But I don't think that there ever will be a clear decision. Overall, I'm fine with the article as it currently stands. --D.H (talk) 12:21, 16 June 2013 (UTC)
PS: I've included an auto-archiver for this talk page, set for 90 days. Hopefully this is acceptable. --D.H (talk) 12:21, 16 June 2013 (UTC)
While I have some remaining issues with the article, I am content with leaving the article as it stands, since the debate between the Will and the Anderson points of view has been clearly delineated in the The one-way speed section. Stigmatella aurantiaca (talk) 12:50, 16 June 2013 (UTC)

One-way speed

Here we now have a slightly conflicting view with that expressed in 'Experiments in which light follows a unidirectional path'. I believe that the current opinion on this subject is as I said above, It has been shown that experiments cannot, in any meaningful way, measure the anisotropy of the one way speed of light...'. Martin Hogbin (talk) 15:57, 16 June 2013 (UTC)

It continues "...unless inertial frames and coordinates are defined so that space and time coordinates as well as slow clock-transport are described isotropically". In addition, both sections end the same way, namely that OWLS is a matter of convention. Therefore I see no conflicting view. Maybe you can describe it better. --D.H (talk) 16:06, 16 June 2013 (UTC)
Actually I think I misread that bit as '...slow clock-transport are described anisotropically'.
Measurement of OWLS isotropy

The real point that I am making is that I believe that current expert opinion is that it is not possible to measure isotropy of the one-way speed of light. In his book, on page 150, zhang says, '...the TPA and JPL experiments as well as other one-way experiments would not determine the directional parameter q. Therefore, these so-called one-way experiments, just as the closed path experiments, are only the tests of the isotropy of the two-way speed of light'. Martin Hogbin (talk) 18:15, 16 June 2013 (UTC)

It is correct that the two-photon absorption and JPL experiments did not truly test the isotropy of the one-way speed of light. The jury is still out as to whether there could have been some flaw in the experimental design of Gagnon et al. (1988) and Ragulsky (1997) such that these experiments could not really have tested for one-way isotropy. It is absolutely true that the analysis in Gagnon et al. was defective, but I have seen no refutation of their experimental design. I have seen a flawed so-called "refutation" of Ragulsky. So no, I would not state that current expert opinion is united in stating that it is not possible to measure isotropy of the one-way speed of light. The debate between Will and Anderson is unresolved. Stigmatella aurantiaca (talk) 18:40, 16 June 2013 (UTC)
I have to disagree with you there. As Anderson says in D.H.'s quote below, it is impossible in principle to measure the isotropy of OWLS. Whilst it might be interesting to find the flaw in every OWLSI experiment there is no need to do so, just as it is not necessary to find the flaw in every perpetual motion machine design. Martin Hogbin (talk) 17:49, 17 June 2013 (UTC)
I think that as long as the conventionality of simultaneity debate is still open, it follows that the one-way speed or isotropy debate is not fully closed as well.
However, there seems to be a tendency among more recent papers that the one-way speed and isotropy is conventional, for instance: Zhang (1995, 1997), Anderson (1998), Roberts (2006), Sonego (2008), Iyer et al. (2010), see article for references.
On the other hand, Will wrote as late as in 2005 (arXiv:gr-qc/0504085) still about "a test of the isotropy of the speed of light using one-way propagation of light between hydrogen maser atomic clocks at the Jet Propulsion Laboratory (JPL)". Recently, there also was the GRAAL experiment claiming "From the above analysis a one-way isotropy limit for the light speed is obtained" (arXiv:1004.2867). This experiment was also interpreted using SME (arXiv:1005.5230, published in Physical Review Letters, doi:10.1103/PhysRevLett.104.241601, though it's interesting that the word "one-way isotropy" was avoided in the PRL publication. Of course, from a conventionalist view the experiment relies on assumptions about momentum conservation etc., but this is only my opinion....
Also SME (Standard model extension) experiments are testing Lorentz violating "one-way speed" limits such as and , even though they are explicitly defined as partially conventional and can be moved from the photon into the matter sector.
At the end, I don't think that there is a definite decision accepted by the whole physics community. But even to draw this conclusion we need some secondary sources. --D.H (talk) 19:13, 16 June 2013 (UTC)
Agreed. Best to leave the section as it stands, neither drawing any conclusions nor making a special point about not drawing any conclusions. Stigmatella aurantiaca (talk) 20:09, 16 June 2013 (UTC)
As I understand it Zhang's book is a very well respected source and is regarded as the definitive work on the subject by many. The later sources you mention are only from arXiv. I am excluding SME from this discussion as I think it is a separate subject, to be addressed in its own section as it is now.
I am not suggesting that we completely leave out the claims of OWLS isotropy measurement but I think it is reasonable to give the 'conventional' argument as the principle one and the 'can be measured' on as a minority view. Do you mind if I try to reword the 'The one-way speed' section to that effect? Martin Hogbin (talk) 22:16, 16 June 2013 (UTC)
I would not appreciate that unless I can be shown correct refutations of Ragulsky or Gagnon et al. The ones that I have seen for Ragulsky have trivial refutations. Several years ago, when I corresponded privately with a proponent of the 'conventional' argument, he was unable to identify the hidden synchronization assumptions that he insisted had to exist in these two experiments, but he still refused to budge from his position. Meanwhile, the use of asymmetric Mach-Zehnder interferometers considerably more advanced than Gagnon's apparatus are being explored for testing Lorentz violation. See Tobar, Wolf, Fowler, and Hartnett, "New methods of testing Lorentz violation in electrodynamics", Phys. Rev. D 71, 025004 (2005) and the erratum published in Phys. Rev. D 75, 049902(E) (2007). Stigmatella aurantiaca (talk) 02:10, 17 June 2013 (UTC)
I think it is agreed that, within the well-established and experimentally verified theory of SR, isotropy of OWLS cannot be measured. Experiments to do so fall, in my opinion, in the same category as experiments to show violations of the second law of thermodynamics or to design perpetual motion machines; within current theory such things are impossible.
Within the well-established and experimentally verified theory of SR, isotropy of TWLS cannot be measured, either. So why are TWLS experiments continuing to be performed? Within current theory, such a measurement is impossible. Instead of thinking of such experiments as Gagnon or Ragulsky as OWLS anisotropy experiments, why not simply consider them as experiments potentially capable of measuring parameters of SME that are not measurable by MMX and KT-style experiments? Stigmatella aurantiaca (talk) 10:27, 17 June 2013 (UTC)
Of course, new theories of physics are being continually developed, and such theories may predict various degrees of Lorentz violation. These, very small, effects will be experimentally investigated from time to time and there is no reason that such experiments should not be mentioned here, but I believe, in separate sections. Experiments to investigate SME, for example, fall into that category.
I think it is important to distinguish between experiments intended to circumvent in some clever way the measurement limitations imposed by SR and those intended to investigate the current state of knowledge in physics. Gagnon's experiment, for example, seems to fall squarely into the category of perpetual motion machines since it claims to be identifying some preferred reference frame, related to the CMBR rather than exploring any currently proposed new theory. Martin Hogbin (talk) 08:40, 17 June 2013 (UTC)
To both of you:
  • Actually I thought that this article is already written in a "conventionalist" manner. I don't know whether it's necessary or even possible (secondary sources) to strengthen this further.
  • Regarding Tobar et al: This is a SME test of the one-way speed values and , which are somewhat conventional as pointed out by the SME founder Kostelecky and in the SME section of our article, that is, those Lorentz violations can be "redefined" and moved from the photon into the matter sector. Tobar write: "although some caution is required when physically interpreting such statements (in particular one needs to unambiguously define the meaning of “one-way light velocity”)" Yes, one has to define its meaning... (Actually it would be interesting if someone would bother to interpret the null results of Gagnon or Ragulsky in terms of SME).
  • My opinion is based on Anderson et al.: Before analyzing every single experiment, one has to understand the principle behind those issues. As Anderson et al. points out: "In particular, it is a matter of principle that one cannot “test for the isotropy of the speed of light”, for to achieve this would be to contradict the synchrony dependence of a one-way speed. Rather, the insights afforded by the existence of this freedom in coordinate choice are valuable in all contexts...". So everyone who claims that it can be directly measured must also be of the opinion that physical experiments can be described without using coordinates at all, which is absurd. No one can deny the freedom of coordinate choice, which by definition affects everything which is directly or indirectly related to one-way velocities - the only problem of conventionality is its apparent triviality....
  • Regarding "definite" works: We should be careful, because this whole "one-way speed" issue is a minority topic.... Google scholar gives 90 citations for Zhangs 1997 book (Search) and 18 citations for his 1995 paper (search). Will's paper is cited 106 times (search). Anderson et al. 103 times (search). Unfortunately, many of those citations to all of those papers are fringe sources or only arxiv preprints. So the whole one-way issue (pro and contra) is rather a minory topic. --D.H (talk) 09:00, 17 June 2013 (UTC)
As I have stated before, I am quite content with the current state of "one-way light speed" section. The topology of Gagnon et al. is essentially that of an asymmetric Mach-Zehnder interferometer, and as such may be understood under the Tobar et al. analysis. Stigmatella aurantiaca (talk) 10:27, 17 June 2013 (UTC)
D.H.,my view is exactly that of Anderson and Zhang and, it would seem, you too. The problem is that this article does not reflect that view; it gives the impression that, within the context of SR, it might be possible to “test for the isotropy of the speed of light”. SME tests are a different matter and should remain in a separate section. I propose to rewrite some of the 'The one-way speed' to make this clearer. Stigmata and yourself can the comment if you have any objections
The whole article has become much too technical for most readers We need to have some simpler introductions in the lead and for many of the sections. We are writing mainly for the general reader.Martin Hogbin (talk) 17:49, 17 June 2013 (UTC)
A wiki article is ideally written to serve the needs of a wide range of sophistication in its target audience, beginning with an accessible introduction, and working its way up to technical topics of interest to moderately advanced users. The early organization of this article fits this pattern admirably. The introduction is clear and understandable. "The two-way speed" concisely defines what constitutes a two-way speed measurement, while noting the distinction between a two-way speed measurement and a two-way speed isotropy measurement such as MMX. Up through "The one-way speed", there is not a single equation. In "The one-way" speed, D.H has provided an admirable summary of the differing viewpoints concerning what it means to be performing a one-way speed isotropy measurement. It is clear from D.H's summary that Will holds a minority viewpoint, and that is perfectly fine by me. I see no reason to change what D.H has written.
Remember that readers of this article are not likely to be complete novices to the issues covered. In all likelihood, they will already be familiar with speed of light measurements and MMX, and have heard or read something about how one-way speed of light is a slippery issue.
The remainder of the article could stand some rearrangement. For example, "Experiments that can be done on the one-way speed of light" is a highly accessible, non-mathematical section that could possibly be moved forwards. The section on "Synchronization conventions" requires a rather sophisticated understanding of the issues, and could possibly be moved back.
But the first three sections are probably as good as we can get for now. "As good as we can get" is actually quite good. I'm quite content with how D.H has handled these issues. Stigmatella aurantiaca (talk) 22:32, 17 June 2013 (UTC)
Will's 2005 'argument'

Will's 2005 'argument' does not really exist. In his review paper he does not even state that he has measured OWLSI. He refers to, 'a test of the isotropy of the speed of light using one-way propagation of light'. This is the JPL experiment which Zhang has shown to measure TWLSI. Will has never challenged this analysis which was widely accepted at the time to be correct. I think claims that OWLSI has been measured must be classified as 'fringe' rather than 'minority' and that the article must reflect this fact. Martin Hogbin (talk) 07:52, 18 June 2013 (UTC)

There is still this problem: Will (who is an eminent expert on relativity tests) never mentioned Zhang's papers, instead he kept referring to his own 1992 paper. The only difference I see is that Will in 1992 used the term "isotropy of the one-way speed of light", while in 2005 and 2006 he used "isotropy of light speed using one-way propagation". For instance, see Will's well known standard work on tests of general relativity, updated in 2006:
  • Will, C.M. (2006). "The Confrontation between General Relativity and Experiment". Living Reviews Relativity. 9. (Free online version)
p. 12: "The Michelson–Morley, Joos, Brillet–Hall and cavity experiments test the isotropy of the round-trip speed of light. The centrifuge, two-photon absorption (TPA) and JPL experiments test the isotropy of light speed using one-way propagation."
p. 21: "Such kinematical approaches have been discussed by Robertson, Mansouri and Sexl, and Will (see [279])... For example, in the “JPL” experiment, in which the phases of two hydrogen masers connected by a fiberoptic link were compared as a function of the Earth’s orientation, the predicted phase difference as a function of direction is, to first order in V, the velocity of the Earth through the cosmic background.... The observed limit on a diurnal variation in the relative phase resulted in the bound .... Tighter bounds were obtained from a “two-photon absorption” (TPA) experiment, and a 1960s series of “Mossbauer-rotor” experiments, which tested the isotropy of time dilation between a gamma ray emitter on the rim of a rotating disk and an absorber placed at the center [279]"
Note that Ref. 279 to which Will referred is his own 1992 paper. No mention of Zhang's reanalysis or any other subsequent paper. If Will really changed his opinion - why in 2006 he still referred to his 1992 analysis only? So while I agree with Anderson and Zhang, it doesn't appear that Will ever abandoned his 1992 argument. --D.H (talk) 10:09, 18 June 2013 (UTC)
He has changed his wording though to be the much more vague, "isotropy of light speed using one-way propagation". His JPL experiment comments are even more puzzling, is he now claiming a non-null result. Martin Hogbin (talk) 12:28, 18 June 2013 (UTC)

1) Will is an acknowledged expert on tests of relativity, not a fringe writer.
Agreed. Martin Hogbin (talk) 12:28, 18 June 2013 (UTC)
2) In his recent works, he prefers not to mention Zhang, although he is obviously aware of Zhang's work.
3) I see nowhere in his writings where he has retracted his 1992 statement, "the synchronization of clocks played no role in the interpretation of experiments provided that one is careful to express the results in terms of physically measurable quantities", i.e. he believes that observables exist in one-way and two-way isotropy experiments that depend on test functions but do not depend on the particular synchronization procedure. So there is clearly some sort of unresolved disagreement between Will and Zhang.
Because he has not publicly retracted his original claim does not necessarily mean that there is an unresolved disagreement. Martin Hogbin (talk) 12:28, 18 June 2013 (UTC)
4) Papers continue to appear in refereed publications which assume (with appropriate caveats) that measurements of the isotropy of one-way speed of light are a valid measurement. For example, from the last couple of years (and making no attempt to evaluate their merit), we have, among others:
Spavieri, G. On measuring the one-way speed of light. Eur. Phys. J. D 66:76 (2012)
Exirifard, Qasem. Cosmological birefringence constraints on light. Physics Letters B Volume 699, Issues 1–2, 2 May 2011, Pages 1–4
Chang, Zhe and Wang, S. Constraints on spacetime anisotropy and Lorentz violation from the GRAAL experiment. Eur. Phys. J. C (2013) 73:2337
5) "OWLSI" is therefore not a fringe topic. Stigmatella aurantiaca (talk) 11:32, 18 June 2013 (UTC)

Regarding papers, I agree that there is ongoing research into new theories, such as SME, which may predict some form of Lorentz violation. I am perfectly happy to have those in appropriate sections, as is the case for SME. My objection is to those who are still looking for some kind of aether drift (preferred frame, CMBR frame), manifesting itself as OWLSI. Martin Hogbin (talk) 12:28, 18 June 2013 (UTC)

I do not understand. You mean that you would a priori dismiss as fringe the preliminary arxiv report by Gurzadyan et al. (2010) which expressed GRAAL results in terms of Δc/c, while accepting as mainstream the refereed report by (almost) the same list of authors (but in different order) by Bocquet J.-P. et al, Phys. Rev. Lett. (2010) simply because the same data was analyzed in terms of SME? Stigmatella aurantiaca (talk) 20:47, 18 June 2013 (UTC)

Do you agree that according to SR measurement of the OWLSI is impossible in principle?.

Likewise, according to SR, measurement of deviations from TWLSI is impossible in principle. What is your point?
The purpose of tests of SR are to discover the limits of where SR may possibly break down. We already know that SR is not applicable in strong gravitational fields. And it is certain that SR would break down at the Planck scale. The question is, are deviations from SR within its domain of applicability detectable at scales accessible to experiment or observation? The trouble is, we have no theory that makes clear-cut predictions on when SR might break down, or how the breakdown will manifest itself.
The situation is quite different for GR. With GR, a variety of heuristic arguments suggest that either the equivalence principle should break down in the range from 10-14 to 10-19, or alternatively, if no breakdown of the equivalence principle is observed in this range, something must be desperately wrong with our best candidate theories for quantum gravity. Overduin et al. (2009) Either way, whether deviations from EP are found or not, the result should be interesting new physics.
Just because we have no such heuristic arguments or guiding principles which should lead us to suspect that deviations from SR should ever be experimentally detectable within its domain of applicability, is no reason for us to stop trying.
Referring back to a previous point: A couple of paragraphs ago, you wrote, "I agree that there is ongoing research into new theories, such as SME, which may predict some form of Lorentz violation." SME is not a theory that predicts anything.
Rather, SME is a test theory whose purpose is to provide a framework in which experimentally discovered deviations from SM may be categorized and quantified. In contrast to RMS, where deviation from SR is encoded only in TWLS, SME includes observable parity-odd parameters that can be used to quantify one-way deviations from SR. Stigmatella aurantiaca (talk) 21:03, 19 June 2013 (UTC)
Statement attributed to Will in the text

I the article we currently have:

On the other hand, some authors such as Clifford Will (1992) claimed that this problem doesn't affect measurements of the isotropy of the one-way speed of light, for instance, direction dependent changes due to relative motion of a "preferred" (aether) frame Σ. Using such a preferred frame test theory, Will argued that it is impossible to measure the one-way speed between two clocks using a time-of-flight method without synchronization scheme, but by referring to experiments in which light follows a unidirectional path he argued

"...a test of the isotropy of the speed between the same two clocks as the orientation of the propagation path varies relative to Σ should not depend on how they were synchronized...".[9]

This statement was shown to be incorrect by Zhang in 1997. It is not clear from Will's later publications that he still supports the he view stated in 1992. I therefore think this section is too strongly worded. Can we look at changing it please. Martin Hogbin (talk) 07:41, 20 June 2013 (UTC)

It is clear from the variety of publications that both D.H and I have cited that a clear consensus has not been adopted within the physics community on this point. I think that the way that D.H has phrased this section is fine. Stigmatella aurantiaca (talk) 07:54, 20 June 2013 (UTC)
In other words, this not a matter of just Will versus Zhang. Stigmatella aurantiaca (talk) 08:05, 20 June 2013 (UTC)
Within the context of the JPL experiment, Zhang has shown Will's analysis and conclusion to be incorrect. Will has not challenged this and has changed his wording to no longer explicitly say that OWLSI had been measured. No other refernces are given for the section that I want to change. SME is covered elsewhere. Martin Hogbin (talk) 10:16, 20 June 2013 (UTC)
The JPL experiment is not the only experiment to consider. A variety of other experiments exist whose results may be interpreted (with appropriate caveats) as measurements of OWSLI, and which have no obvious closed loops or synchronization assumptions. Your preferred language, which you wrote in bold letters above, implies that all of these efforts are misguided. You wrote, "It has been shown that experiments cannot, in any meaningful way, measure the anisotropy of the one way speed of light," and this serves as a blanket condemnation of the work of multiple mainstream scientists, which neither you nor I have the necessary qualifications to do. Stigmatella aurantiaca (talk) 11:19, 20 June 2013 (UTC)
Yes, I have been thinking about this disagreement. I am referring to a world of classical physics as it is understood today, with no significant gravitation. So I believe that if we take all EM radiation to travel (two-way) at c, in vacuo, where c is Einstein's constant and the maximum speed at which matter and information can travel, with no quantum effects, and with length being defined using the two-way speed of light and a single clock, then measurement of OWLS and OWLSI is a mathematical impossibility. Would you agree with that? Martin Hogbin (talk) 16:26, 20 June 2013 (UTC)
One we allow gravitation then we have to define what we mean by length and time. If we allow the possibility that EM radiation does not always travel at c in vacuo, then things change. Add quantum effects and life gets even more complicated. Martin Hogbin (talk) 16:26, 20 June 2013 (UTC)
I do not see how continuing along this line of discussion would contribute constructively to the article. Remember that these talk pages are not a forum for general discussion of the article's subject, and the WP:NOR policy. Stigmatella aurantiaca (talk) 03:55, 21 June 2013 (UTC)
This is not a general discussion, it is a discussion about a specific section of text. There is no OR involved here it is a matter of following the sources and mainstream science.
The conditions that I describe above are not something that I have just made up; they represent how current, experimentally verified, theories of mainstream science and metrology apply to macroscopic experiments in inertial frames. As DH has shown this is also the opinion of the majority of sources on the subject. The contrary opinion stated by Will in 1992 has been shown by Zhang to be erroneous and this has not been challenged by Will. Later writings of Will are rather vague and cannot be considered as clearly stating a different opinion from that of Zhang and the majority of scientists. Other sources on the subject are generally referring to experimental investigations into new theories of the subject. What sources are there stating that, under the conditions that I specified above, OWLSI can be experimentally measured? I believe such view must be regarded as fringe.Martin Hogbin (talk) 09:12, 21 June 2013 (UTC)
The purpose of tests of special relativity are to seek out potential deviations from Lorentz invariance, and there are multiple authors who agree with Will's viewpoint. For example, see Mattingly's Living Reviews article. Is Mattingly a fringe author? If you describe a universe that is perfectly Lorentz invariant, then of course neither TWLSI nor OWLSI will give any measurement other than zero. If there are multiple opinions on a subject, Wikipedia needs to reflect the range of opinion without giving any one opinion undue weight. You wish to place yourself in the position of the final judge of this complex issue, against the opinion of multiple mainstream authors. Stigmatella aurantiaca (talk) 12:08, 21 June 2013 (UTC)
Changes made

I have made some changes to better reflect sources and to separate the current state of knowledge from ongoing research. I have left in Will's 1992 dissenting view. Martin Hogbin (talk) 09:31, 21 June 2013 (UTC)

Your rewordings give the impression that Will is alone in being the only mainstream scientist who believes that OWLSI represents a legitimate measurement, while I have given you multiple counterexamples. Is Mattingly a fringe author? Is the entire GRAAL group composed of fringe personalities?
I repeat that I believe that D.H has already given proper balance to the differing opinions on this subject. It is clear that D.H's own sympathies are towards the conventionalist point of view, and this is perfectly OK with me, since he acknowledges that a significant minority exists composed of reputable scientists. Stigmatella aurantiaca (talk) 12:34, 21 June 2013 (UTC)
Of course I am not claiming that all those you mentioned are fringe authors but I do want to make the distinction between the currently verified theories of physics and unverified new theories.
Perhaps you could explain to me what you think Will's current position now is and how it differs from that of, say, Zhang. Martin Hogbin (talk) 15:34, 21 June 2013 (UTC)
Mattingly (2005) obviously does not believe that Zhang (1997) was correct. He wrote, "Many experiments, such as those that measure the isotropy of the one way speed of light [275] or propagation of light around closed loops, have observables that depend on a, b, d but not on the synchronization procedure," with reference [275] referring to Will 1992.
What we have is a significant minority of mainstream scientists who believe that OWLSI is measurable.
D.H's compromise wording best accommodates the differing points of view. Stigmatella aurantiaca (talk) 16:07, 21 June 2013 (UTC)
Should we perhaps explore the dispute resolution process rather than fighting out our differences? Stigmatella aurantiaca (talk) 16:38, 21 June 2013 (UTC)
Civil discussion is the preferred dispute resolution process on WP. You are welcome to try a more formal process such as an RfC but I doubt that there are many editors who understand the subject and are willing to help. I am still trying to find out exactly what we do disagree about, we may not actually disagree about that much, perhaps only how the article is laid out. I am talking about the currently accepted laws of classical (non-quantum) physics without gravitation, where 'c' is the speed in-vacuo if all EM radiation and the maximum speed at which matter or information can travel. Zhang could easily be wrong if we allow FTL information transfer for example.
So, to avoid, unnecessary argument perhaps you could make clear whether you think OWLSI can is measurable under the conditions that I have described and what you believe Will's opinion is on the matter. Martin Hogbin (talk) 17:14, 21 June 2013 (UTC)
  1. If you postulate a universe in which deviations from TWLSI and OWLSI are not measurable, then deviations from TWLSI and OWLSI will not be measurable. To date, all experimental test points towards our universe being such a universe, but this is a matter for experimental test, not assertion. So I don't see what the point is of you asking whether I think deviations from OWLSI can be measurable under the conditions that you have described, since you are simply setting up a straw universe argument.
  2. A significant minority of mainstream scientists believe that meaningful measurements of OWLSI are possible that do not depend on the synchronization procedure. Mattingly stated as such in 2005, and Will continues to cite his own work. I have cited several recent refereed publications that treat OWLSI as a legitimate target for measurement, and I can easily dig up more. Hence I object to any sort of statement to the effect that Zhang definitively disproved the possibility of meaningful OWLSI measurements in 1997. The differing opinions of mainstream scientists on this must be discussed and given due weight.
  3. You are trying to make this into a Will versus The Rest of the World argument. It is not. It is a significant minority supporting a position that differs from that of the majority.
  4. I believe that D.H's handling of this subject treats it properly. His sympathies are obvious, but he respects the existence of dissenting opinion. Stigmatella aurantiaca (talk) 04:02, 22 June 2013 (UTC)
1) I am not talking about a straw universe but one which is, to the best of our current knowledge, the universe that we live in. The speed of light (against a matter length standard) can be measured and TWLSI has been measured (independently of any clock synchronisation scheme) and found to be null. Theories that are consistent with current experimental results are SR and LET (or Edwards' theory as Zhang calls it). These theories could be distinguished by OWLSI measurements (independent of any clock synchronisation scheme) but it has been shown that, within the above context, such measurements are impossible, unless some way of sending signals faster than light is found.
There are other theories that fit within current experimental limits that are not isomorphic with SR or LET.
Do you agree with the above?
Do you think that Will currently agrees with the above?
Read Mattingly. You keep on trying to make this a Will versus the rest of the world issue, and you keep on distorting his position. NOBODY has ever claimed that theories indistinguishable from SR are, in fact, distinguishable from SR.
String theory, warped brane worlds, non-commutative field theory, emergent gauge bosons, varying moduli, axion-Wess-Zumino models, analogues of emergent gravity in condensed matter, ghost condensate, space-time varying couplings, and varying speed of light cosmologies all allow for the possibility of producing predictions that are distinguishable from SR within conceivable limits of experimental sensitivity or cosmological observation.
A significant minority of scientists believe that OWLSI represents a meaningful measurement that may be employed to distinguish such theories from SR.
2) I do not understand what this statement is intended to mean. Do you mean that a significant minority of mainstream scientists believe that there are experiments that will distinguish Edwards theory from SR? If that is so then we have a disagreement about physics.
NO! Why do you keep on misunderstanding this point?
Or are you just saying that there may be, as yet undiscovered, effects that might lead to positive results for experiments that could be considered to measure OWLSI measurements independently of any clock synchronisation scheme? If that is the case then we have no disagreement about physics but maybe we do about the structure of the article.
That is about as close as you've managed to get to understanding my position.
Try this thought experiment for size. Suppose the properties of the universe represent a linear combination of a Lorentz invariant component (99.9999999999%) mixed with 0.0000000001% of a classical aether component. I imagine that such a universe will have passed all TWLSI and OWLSI tests to date. If not, just add a few 9s and 0s.
Continuing along the lines of this thought experiment, consider a futuristic detector consisting of a light bulb of ultra-constant intensity, and an ultra-sensitive blackened thermometer spaced ten centimeters from the light bulb. The temperature registered by the thermometer should depend on the relative orientation of the light bulb and thermometer with respect to the residual aether wind. Yet the operation of this detector does not depend on any clock synchronization scheme.
(This scheme may resemble Mössbauer rotor experiments in certain respects, except that I disagree with the notion that alterations in resonance between the Mössbauer source and Mössbauer detector should necessarily have been measurable. Although a hypothetical aether wind may affect photon speed, wavelength, momentum and energy, it cannot be expected to change the frequency at the detector, and resonance is a frequency-dependent phenomenon.)
In principle, given sufficient sensitivity and precision in its design, this detector should be able to measure OWLSI effects in the 99.9999999999%/0.0000000001% universe described above.
Alternatively, consider a highly monochromatic light source and a high precision diffraction grating. In a Lorentz violating universe, directionally dependent one-way wavelength variations should, in principle, be measurable without making any synchronization assumptions concerning the source and detector. Stigmatella aurantiaca (talk) 22:33, 23 June 2013 (UTC)
3) Not even that. It is no longer clear exactly what Will believes. He now refers to the rather ambiguous, 'test of the isotropy of the speed of light using one-way propagation of light'. It seems to me that Will is just leaving his options open a little.
I repeat. You keep on trying to make this a Will versus the rest of the world issue, and you keep on distorting his position. NOBODY has ever claimed that theories indistinguishable from SR are, in fact, distinguishable from SR. Stigmatella aurantiaca (talk) 05:39, 23 June 2013 (UTC)
4) It would be helpful if D.H. were to give his opinion on the above discussion. Martin Hogbin (talk) 09:08, 22 June 2013 (UTC)
Stigmatella, it looks like we do disagree on too much then. In the universe that you describe above, I agree that OWLSI would be detectable but I would point out the following. There is no suggestion that we might live in a universe exactly as you describe. The appropriate test theory for such an effect would be Robinson's (or MS). The best tests for such a universe would be two-way tests (as I think you have said).
All I am trying to do is to separate statements based on experimentally verified theories from those referring to proposed respectable academic new, but as yet unverified, theories.
As it happens I do believe that Will's statements were intended suggest that theories indistinguishable from SR are, in fact, distinguishable from SR. That is what Zhang's analysis implies to me. But my opinion is, of course, not what counts in WP. I will give reasons why we should separate the propose from the current below. Martin Hogbin (talk) 10:21, 24 June 2013 (UTC)

Why we should separate tests of new theories

1) It is important to separate in our readers' minds what we know from what we are trying to find out.

2) Different test theories (SME) would be appropriate for tests of new physics from those required for tests of classical physics (Robertson).

I do not understand your distinction here. Tests of classical physics are not intended to uphold classical physics, but rather are intended to explore where classical physics may break down. Any breakdown of classical physics automatically implies "new physics".
mSME subsumes RMS and provides, in exchange for much greater complexity, a much more complete classification scheme for tests of new physics. Its structure provides a consistent framework for classification and quantification from such Lorentz violation experiments as Penning traps, clock comparison (i.e. Hughes-Drever) experiments, cavity experiments (i.e. modern MMX), spin polarized torsion balances, neutral meson measurements, storage ring experiments, and so forth. Nevertheless, I do not understand what you mean about the simpler RMS being suited only for testing classical physics. Stigmatella aurantiaca (talk) 11:10, 24 June 2013 (UTC)
In classical physics all that is needed as a test theory is Robertson's theory. Classically and within reason, everything that can be tested (like the example you gave) can be tested using Roberson's test theory. Edwards' theory, in particular, is pointless because it tests nothing physical that cannot be tested using Roberson's theory. Once you speculate about effects that would require SME to detect, you start to fall foul of 3) below. What does 'speed' even mean? Martin Hogbin (talk) 12:39, 24 June 2013 (UTC)

3) In new theories of physics the intuitive meaning of 'the speed of light' breaks down. If a positive result were obtained for an SME test what would it mean? We could probably call it a variation in the speed of light, some effect on matter, some effect on space , some effect on time. It would initially be a matter of personal preference until maybe it was decided by convention. Whatever the decision, the connection to intuitive space and time would be tenuous.

4) Partial aether drift type theories are likely to predict larger effects that quantum gravity. Martin Hogbin (talk) 10:21, 24 June 2013 (UTC)

Existing bounds on Lorentz violation are already sufficiently tight to have excluded various quantum gravity candidates, or at least, the simplest forms of these quantum gravity candidates. See Mattingly (2005). "We currently have bounds on Lorentz violation strong enough that there is no easy way to put Lorentz violating operators of dimension ≤ 6 coming solely from Planck scale physics into our field theories. It therefore seems hard to believe that Lorentz invariance could be violated in a simple way."
Exactly. Some of the experiments performed can only be intended to distinguish between LET and SR. They would not be sensitive enough as QG tests. Martin Hogbin (talk) 12:39, 24 June 2013 (UTC)
Will

Whatever Will meant it is not appropriate to have his view where it is now. If he is talking about theories indistinguishable from SR the we agree that he is wrong. If he is talking about quantum gravity then the comment is in the wrong place. Martin Hogbin (talk) 10:21, 24 June 2013 (UTC)

That would be a misreading of Will. Mattingly (2005) restates Will in what you might consider a more appropriate section.
Fringers may believe that it should be possible to distinguish LET from SR using one-way experiments. Will is not fringe. Stigmatella aurantiaca (talk) 11:10, 24 June 2013 (UTC)
I am reading Will either way. Either he is saying that experiment can distinguish LET from SR, which I think would have to be a distinct minority view at the very least, or he is referring to quantum gravity or the like and his quote is in the wrong section. Martin Hogbin (talk) 12:27, 24 June 2013 (UTC)
Mattingly's paper is about quantum gravity and I would have no objection to mentioning it in the SME section. I doubt he gave much thought to exactly what Will meant in his paper.
You have not mentioned Fringers before. LET (Edwards' theory) is generally considered to be experimentally indistinguishable from SR, I though that you had agreed that. Martin Hogbin (talk) 12:27, 24 June 2013 (UTC)
I am quite certain that Will did not mean that experiment can distinguish LET from SR. Certainly that is not Mattingly's interpretation of Will's words.
We agree that LET is not experimentally distinguishable from SR.
My interpretation of Will's words is OWLSI represents a measurable property (in a Lorentz-violating world) that can be used to detect certain forms of Lorentz-violation. A LET universe is not a Lorentz-violating universe, and deviations from OWLSI are intrinsically impossible to measure in such a universe. Stigmatella aurantiaca (talk) 12:40, 24 June 2013 (UTC)
A general comment to the preceding discussions: If Lorentz invariance is satisfied in standard synchrony, then it is also satisfied in non-standard synchrony. If Lorentz invariance is violated in standard synchrony, then it is also violated in non-standard synchrony. It's about the interpretation of those results in terms of OWLSI, and it certainly belongs into the current position in the article. For instance, Will used a form of RMS and concluded (like before him Mansouri and Sexl), that it allows for measurements of OWLSI independent of synchronization scheme (it's notable that at the same time, both Mansouri-Sexl as well as Will argued that this doesn't exclude the possibility of an aether theory empirically equivalent to relativity, where certain effects hide real OW anisotropies). However, as explained by Zhang and Anderson et al., Will's OWLSI statement is not a direct consequence of RMS, which in their opinion doesn't contain quantities that are independent of synchrony choice. As Anderson et al. (1998) pointed out:
"The assumption of isotropy in Σ is responsible for the lack of generality of the Mansouri-Sexl formalism and underlies this common deficiency of interpretation. If this is borne in mind, there is no problem with the Mansouri-Sexl theory; the isotropy assumption in Σ is economical in helping to reduce the parameters of the theory. A simple resolution of the matter is to accept the Mansouri-Sexl formalism with this caution ... The function of the Mansouri-Sexl type of test theory is not so much as a test for a preferred frame as a test of Lorentz invariance ... Mansouri and Sexl’s erroneous belief that the one-way speed of light could be measured empirically in applications of their theory has been inherited by a surprising number of good physicists. We mention Vargas [217], also Will, Krisher and coworkers [ 108,110,229,230]. Riis et al. [180] title their paper “Test of the isotropy of the one-way speed of light...” and Gabriel and Haugan [64] retain similar terminology. Such literature has played a considerable part in publicity within the last decade (see [ 184,78,229,173,199,171]) of modern tests of relativity."
So this whole Will-Zhang-Anderson discussion is about the point: How do we interpret the RMS parameters? Of course, this is somewhat superseded by the development of SME, in which we have a complex interaction between photon and matter sectors and the possible conventionality of their definition. But bear in mind: The empirical confirmation of Lorentz invariance or its violation is not conventional in both RMS and SME, only their interpretation or description in terms of coordinates and velocities. --D.H (talk) 15:44, 24 June 2013 (UTC)
Firstly, let us leave SME out of this discussion, it is a much more complex theory and any connection with intuitive notions of speed are long gone. We, quite rightly, have a separate section it in the article.
Regarding interpretation of MS parameters (Robertson's was a simpler theory) the point is that there is a directional parameter in MS theory that cannot be determined by experiment; it essentially determines to the conventional and unmeasureable synchronicity choice.
To put it simply, I think we must say that Will was wrong about measurement of OWLSI. Zhang and Anderson clearly say this and no one has challenged them. Will has since significantly watered down his wording. Although Mattingly cites Will, this is only in passing, in an article on quantum gravity, I do not think that constitutes an endorsement of Will's original claim. Martin Hogbin (talk) 16:35, 24 June 2013 (UTC)
As far as I can see, any mention of a specific paper without clarification is some sort of an endorsement of the claims in that paper. Will himself keeps referring to it, also others such as Mattingly do so in connection with the "isotropy of the OWLS". I agree with you that this is unfortunate, but we cannot change it. The only thing that we can do is writing that there are many recent papers and authors objecting to this position - which we already did in the article. PS: Don't forget that as a test theory for Lorentz invariance violations, Will's RMS paper is indeed very valuable, independent of his views on synchronization independent OWLSI. --D.H (talk) 16:48, 24 June 2013 (UTC)
I have nothing against Will or his paper but I think we must give due weight to what the sources actually say. We have at least two sources which clearly and unequivocally say that Will was wrong and give clearly reasoned arguments why this was so.
Neither Will not anybody else has challenged Anderson and Zhang's arguments. Will has changed his wording and no longer claims to measure OWLSI. Citations are, of course, significant but they do not have the same weight as a clear and unchallenged statement that Will was wrong. These errors propagate through the system as Anderson point out above. Martin Hogbin (talk) 17:10, 24 June 2013 (UTC)
1) You claim that there are no arguments against Anderson and Zhang? On the contrary, the entire conventionality thesis has been under fierce and sometimes bitter debate. As Janis (1998) wrote, "The debate about conventionality of simultaneity seems far from settled, although some proponents on both sides of the argument might disagree with that statement." You are one of those taking the position that the matter is settled, when it is in fact not.
2) Above, I described a hypothetical "99.9999999999%/0.0000000001%" universe and a measurement protocol that you were unable to refute, writing "in the universe that you describe above, I agree that OWLSI would be detectable," nitpicking that there was no evidence that we were living in such a universe. That was not the point. The point is that universes are conceivable that are not currently distinguishable from the current universe, but in which OWLSI represents a meaningful measurement. You need to properly refute my counterexamples to Anderson and Zhang, rather than ignoring them and dismissing them as irrelevant.
3) Multiple articles by what appear to be mainstream authors have appeared since the appearance of Zhang's and Anderson's arguments which treat either OWLSI, or the odd-parity components of SMS, or both as valid targets of measurement. Here are a few peer-reviewed ones; I find also multiple arxiv articles by authors with good publication histories in respectable journals:
V. G. Gurzadyan et al. Probing the light speed anisotropy with respect to the cosmic microwave background radiation dipole. Mod. Phys. Lett. A 20, 19 (2005).
ME Tobar, P Wolf, A Fowler, JG Hartnett. New methods of testing Lorentz violation in electrodynamics. Physical Review D, 2005
Q Exirifard. Cosmological birefringence constraints on light. Physics Letters B, 2011 - Elsevier
FN Baynes, AN Luiten, ME Tobar. Testing Lorentz invariance using an odd-parity asymmetric optical resonator. Physical Review D, 2011
4) So far as I can see, the debate is still open, and D.H's wording represents a good compromise that provides due consideration to the opposing points of view that exist. Stigmatella aurantiaca (talk) 09:09, 25 June 2013 (UTC)
RfC?

I think that WP:weight dictates that we should not give Will's opinion its current weight in the article. We could possible ask a wider section of editors about this. No specialist subject knowledge would be required if we can agree on a neutral statement of the facts.

On the other hand it might be easier to try to agree on something between ourselves. Martin Hogbin (talk) 08:59, 25 June 2013 (UTC)

I believe that D.H already provides a reasonable compromise. I repeat, you keep trying to turn this into a "Will versus the rest of the world" debate, when in fact I have provided multiple examples of mainstream authors who believe that OWLSI represents a meaningful measurement. The entire conventionality thesis has been under fierce and sometimes bitter debate. As Janis (1998) wrote, "The debate about conventionality of simultaneity seems far from settled, although some proponents on both sides of the argument might disagree with that statement." You are one of those taking the position that the matter is settled, when it is in fact not.
Indeed, it has become evident that your main argument against Will is that you believe that he claims that theories experimentally indistinguishable from SR are, in fact, experimentally distinguishable from SR. This is a clear misrepresentation of Will's position, and is not supported by authors such as Mattingly (2005). Stigmatella aurantiaca (talk) 09:16, 25 June 2013 (UTC)
Martin, we both agree with conventionalism, but we cannot impose our views on others. As long as the conventionalist thesis is still disputed, there will never be a universally agreed resolution of this OWLSI problem.
Since Will is one of the leading experts on tests (and test theories) of relativity, WP:weight requires us to discuss his views. In particular, Will's paper is still cited in peer reviewed journals in relation to OWLSI (maybe more often than any other paper). As pointed out above several times: He himself refers to his paper, never directly abandoned anything he said, and refuses to cite Zhang or Anderson or anyone else who disagreed with his interpretation. Actually, his "Confrontation between General Relativity and Experiment" is the standard work for tests of general relativity.
The only thing we saw is the slight change from "isotropy of the one-way speed of light", to "isotropy of light speed using one-way propagation". I've included this in the article now.
Regarding RfC: I really don't see the problem - everything is well sourced, and the conventionalist view has the last word in every section. Maybe Wikipedia talk:WikiProject Physics would be a better place to invite other editors to discuss here. --D.H (talk) 09:56, 25 June 2013 (UTC)
As I have stated before, I am perfectly OK with the conventionalist view having the last word in every section. That is, after all, the majority opinion of the major contributors to this article. Stigmatella aurantiaca (talk) 10:29, 25 June 2013 (UTC)
Points of view do not really come in to it, not mine, not yours, not anyone else's. Will used MS theory to show that OWLSI could be measured. Zhang showed the he was wrong. Zhang shows in his book that MS theory cannot be used to analyse one-way light speed experiments; this is a simple algebraic fact. Neither Will not anyone else has even claimed that there is an error in Zhang's work let alone found one and, unless and until someone does, it is an indisputable fact that Will was wrong. The fact that others may have cited Will's original paper cannot change this; maybe they have not read Zhang's analysis. Martin Hogbin (talk) 21:41, 25 June 2013 (UTC)
A significant number of mainstream researchers would deny that Zhang succeeded in proving anything at all. Bear in mind that a proof is only as good as its assumptions, and Zhang's "proof" is firmly grounded in the conventionalist thesis, an assumption that such an eminent mainstream authority as Ohanian (2004) has called a "failure". So long as the conventionalist thesis is disputed, the issue must be considered open. Stigmatella aurantiaca (talk) 23:43, 25 June 2013 (UTC)
I can only see the abstract. Do the authors specifically mention Zhang and find fault with his argument? Martin Hogbin (talk) 11:48, 26 June 2013 (UTC)
No, but Ohanian cites many other papers directly relevant to the debate. He wrote.

        By adopting an inertial reference frame, we break the vicious circles that interfere with attempts to establish an unambiguous synchronization of the clocks of the reference frame, and we defeat the conventionalist thesis. A synchronization convention is needed only if physics does not give us sufficient guidance to determine a unique synchronization. But when we adopt an inertial reference frame, the requirement of absence of pseudoforces determines the synchronization uniquely, and no further synchronization convention can be imposed. The fundamental error of Reichenbach and his conventionalist followers was their neglect of dynamics. They failed to appreciate that the time variable must be chosen in such a way that the laws of dynamics keep their standard form. As Wheeler said in a related context, "Time is defined so that motion looks simple." 26
        The defeat of the conventionalist thesis leads to fatal consequences for some modifications of Special Relativity based on this thesis. Winnie27 formulated relativity in the nonstandard Reichenbach synchronization, and he constructed a "Lorentz transformation" between two R frames in motion with respect to each other. Tangherlini,28 Grünbaum,29 and Mansouri and Sexl30 exploited the Reichenbach synchronization to modify the Lorentz transformation so as to eliminate the relativity of synchronization, that is, the term 2Vx/c2 that appears in the standard Lorentz transformation equation for the time coordinate. In view of the failure of the conventionalist thesis, all of these attempts must be dismissed. They rely on the adoption of reference frames that are not inertial, that is, they rely on the adoption of general coordinates. None of these authors examined the implications of their transformations for dynamics, and they remained unaware that their transformations give rise to pseudoforces.31
        In an inertial reference frame, with clocks synchronized by clock transport or by mechanical methods, we can perform a one-way measurement of the speed of light by time-of-flight. Conventionalists have objected that these measurements rest on a vicious circle, but the adoption of an inertial reference frame overrules such objections. In principle, if Special Relativity is violated, the measurements can give results that vary with direction. For instance, such a dependence of the velocity on direction is predicted by the pre-relativistic ether theory whenever the laboratory is in motion relative to the ether. The experiments of Krisher et al.32 and of Wolf and Petit33 performed in the 1990s relied on clocks synchronized by clock transport; in an inertial reference frame, they are therefore valid one-way measurements of the speed of light. Earlier experiments on the one-way speed of light relied on measurements of Doppler shifts of laser light or gamma rays.34 According to the ether theory, the Doppler shift depends not only on the speed of the receiver relative to the source, but also on the speed of the source relative to the ether. Anisotropies of the one-way speed of light can therefore be detected through anisotropies in the Doppler shift. Because the atoms or nuclei that act as receivers are effectively moving clocks, the interpretation of such measurements again hinges on clock transport as a valid synchronization procedure in an inertial reference frame.

I'm fortunate enough to live only a few minutes from a medium-sized university where I can download papers for free, but what is your objection to spending less money than the cost of a McDonalds meal to rent Ohanian? Stigmatella aurantiaca (talk) 00:49, 27 June 2013 (UTC)
Yes, this important problem is discussed in the section "Inertial frames and dynamics", with the responses of Martinez and Macdonald. See also at the end of Janis (1998), and the paper by Salmon (1977) referring to similar discussions. --D.H (talk) 07:46, 27 June 2013 (UTC)
Martin, on the contrary, that Will's paper is still cited by some mainstream sources is the only thing that counts, because Wikipedia is based on sources and citations. And the same authors choose to ignore Zhang or Anderson. What we think about "indisputable facts" doesn't matter - the mainstream scienfific community has to decide by referring to it in their papers. --D.H (talk) 08:10, 26 June 2013 (UTC)
Wikipedia is indeed based on sources but not, as far as I know, on citations. I do not think that it is our job to try to rank sources that way. The point is that we have a source clearly saying that Will's use of MS theory to analyse the JPL experiment meant that OWLSI had not been measured, and clearly explaining why this is so. As far as I know, we do not have any sources saying that Zhang was incorrect and pointing out the mistake in his argument. Had this been the case we would have had to consider how to give due weight to two clearly conflicting views, however, that is not the case; we have no sources specifically saying that Zhang's argument was incorrect. The presumption that, because some sources cite Will's original paper, they are saying that Zhang was wrong is unjustified. Martin Hogbin (talk) 11:48, 26 June 2013 (UTC)
We have sources long after the publication of Zhang in 1997, which say conventionality of OWLS is wrong and some of them are referring to Krisher, Will etc.. It's not our business to say that these recent sources are wrong, only because they don't incorporate Zhang or Anderson in their papers. In order to do that, we need secondary sources that show that Will was wrong and which confirm that Zhang and Anderson provided definite answers which are accepted by mainstream. I'm not aware of any such source published in peer reviewed journals - please cite them.
In addition, saying that Zhang or anyone else was never directly refuted is a bad argument, because the same argument is used by almost all fringe authors who claim that none of their "theories" was "refuted" by anyone (of course, Zhang is not a fringe author). We need secondary sources. --D.H (talk) 12:36, 26 June 2013 (UTC)
Well, of course, I think you are wrong in the way that you evaluate sources. You cannot equate passing citations, with no detail or explanation, with a couple of chapters in a book, complete with detailed mathematical analysis. As you say Zhang is a respected mainstream author who has found a mistake in the work by another respected mainstream author. In those circumstances it would be surprising if any errors considered to exist in Zhang's work were not specifically pointed out. In many ways Zhang's book is a secondary review source, reviewing and citing over 200 primary sources but unfortunately, on this particular subject, it does contain some original work.
By the way, the reason that I am confining discussion to Will and Zhang is that there would seem to be a very clear and well defined disagreement between them, particularly over analysis of the JPL experiment using MS theory.
It still would be possible to take his discussion to a wider audience through an RfC. The disagreement is not about physics or our own personal opinions but about how to evaluate apparently conflicting sources of various types. There must be many editor her who could help with that. Martin Hogbin (talk) 08:28, 28 June 2013 (UTC)

Well, I think it is mistake to try to confine the discussion to Will and Zhang. The issues involved in the debate on the conventialist thesis are highly complex. We have multiple sources which go into considerable detail on why, in the respective authors' opinions, the conventionalist thesis, on which Zhang bases his argument, is misguided. A fair assessment of the literature on this relatively obscure, minority topic leads to the conclusion that no definitive judgement can be made. I am humble enough to admit that I could be wrong in my beliefs, and you should be as well. D.H's wording on this section provides, in my opinion, an excellent summary of the issues involved, one which is fair to all sides.

Some issues like SME are indeed highly complex but other, like Zhang's re-analysis of Will's analysis of the JPL experiment using MS theory are not so complex. You say 'We have multiple sources which go into considerable detail on why, in the respective authors' opinions, the conventionalist thesis, on which Zhang bases his argument, is misguided'. Which sources are these and are they sources which actually show mistakes in Zhang's calulations? Martin Hogbin (talk) 11:02, 29 June 2013 (UTC)
To start with, Ohanian (2004) from which I extensively quoted above. Even if you don't care to purchase the paper, you can rent the paper for $3.99. If Zhang's basic assumptions are flawed, then Zhang's "proof" is flawed. You need to realize that there is an extensive literature going back and forth on this issue. Your insistence that Zhang must be right and Will must be wrong is misguided. Stigmatella aurantiaca (talk) 11:57, 29 June 2013 (UTC)

I might mention that I am rather annoyed at the amount of time that debating this issue has taken away from my ability to work on a far more important wikipedia article, accessed nearly 100x as often as One-way speed of light, which had a considerable number of basic inaccuracies that I've needed to correct. Stigmatella aurantiaca (talk) 15:04, 28 June 2013 (UTC)

How you choose to apportion your time is up to you. Martin Hogbin (talk) 11:02, 29 June 2013 (UTC)
What I am saying is that I am getting pretty sick and tired of this conversation, and want to get on to other topics that I'm interested in. If you don't care to understand the relevance of the extensive literature debating the pros and cons of the conventionalist thesis and how it affects Zhang's analysis, then there isn't any point to my continuing. Stigmatella aurantiaca (talk) 11:57, 29 June 2013 (UTC)
You are perfectly entitled to tire of this conversation. That does not mean that other must. Martin Hogbin (talk) 12:01, 29 June 2013 (UTC)