Talk:Speed of light/Archive 6

US/UK English
I have just changed all the UK 'travelling' to US 'traveling' as the article seems to mainly use US English, however, I notice that the earliest versions of the article appear to have been written in UK English. We need to chose one and stick to it. Martin Hogbin (talk) 09:32, 13 January 2009 (UTC)

Rash statement
Isn't it a bit bold to say that the speed of light "is the speed of not just visible light, but of all electromagnetic radiation, as well as gravitational waves"? After all, no one has ever detected a gravitational wave.Lestrade (talk) 00:01, 18 January 2009 (UTC)Lestrade
 * Yes, good point, perhaps we should say something like, '...and is the speed that gravitational waves are expected to travel at'. Martin Hogbin (talk) 00:22, 18 January 2009 (UTC)
 * I disagree, I think that's injecting doubt when really there is none. There's extremely strong indirect evidence that gravitational waves exist and behave the way we expect them to, and no serious expert in this field (to my knowledge) doubts this. I think the appropriate five-word summary in the introduction is "as well as gravitational waves". If it was a couple sentences in the body of the article, one could be more specific.
 * After all, it's also true that no one has ever measured EM radiation whose wavelength is one lightyear, let alone its speed, but the article rightly claims that this radiation travels at the speed of light. --Steve (talk) 00:57, 18 January 2009 (UTC)
 * That was my first thought, but there really is a considerable difference between EM and gravitational waves. EM waves are the most observed phenomenon in existence, and their speed (at several frequencies) has been directly measured. Gravitational waves are predicted by the well-established theory of GR but have only been indirectly observed and there are no direct measurements of their speed.


 * In the end it boils down to the question of the level of evidence that we choose to required to say 'is' rather than 'is believed to be'. I do not have a strong opinion but suggest that the matter is worthy of some thought. I guess we should look to see how reliable sources put it. Martin Hogbin (talk) 11:29, 18 January 2009 (UTC)


 * I have looked for some sources. Online, the VIRGO site says,'gravitational waves can travel through space...', and the LIGO site says, 'gravitational waves can travel through space...', but also ,'Actually, after 30 years of active research, we only have an indirect proof of their existence'. I cannot find mention of their speed in either site.  Text books, such as 'Gravity' by Hartle, say that 'Linearized gravitational waves ..propagate at the speed of light', but this is clearly in theoretical context.  I have found this paper [] which seems to set some experimental constraints on the speed of gravitational waves.  Overall, with say the Hartle and the arxiv references, I agree that we can leave it as it is. Martin Hogbin (talk) 13:17, 18 January 2009 (UTC)

Direct detection is so passé. The luminiferous æther was real for a while, why can't gravitational waves and black holes be real for us, too?Lestrade (talk) 01:07, 18 January 2009 (UTC)Lestrade
 * I concur with those recommending change. Gravitational waves are undetected as yet. Although models suggest they will travel at the speed of light, detectors based on that model have yet to register anything. I don't think it would be out of line to say "...and the speed at which models predict gravitational waves travel."  The web link added discusses "epsilon", the amount by which gravity waves vary from the speed of light. The observational data aren't strong enough to fix the value at the speed of light.Novangelis (talk) 14:08, 18 January 2009 (UTC)


 * It is a difficult call, we should really only say what reliable sources say on the subject rather than state our opinions. As I said above it boils down to what level of confidence is required to say 'is' rather than '...models predict...' bearing in mind that GR is a very well established and experimentally verified model.


 * I think it might be quite hard to find a really reliable source that literally says 'gravitational waves travel at the speed of light' in a context that is not obviously purely theoretical. I think we should look for one but, if one cannot be found, then we probably should change the wording along the lines that you suggest. Martin Hogbin (talk) 15:05, 18 January 2009 (UTC)


 * Well argued and researched, Martin. I concur. --Steve (talk) 17:10, 18 January 2009 (UTC)

Joseph H. Taylor and Joel M. Weisberg have observed the binary star system PSR B1913+16 for several decades. They claim that their observations indirectly prove the reality of gravitational waves and are within 0.2 % agreement with the prediction of general relativity. This claim is based on the logical conclusion that the orbits of the two stars should become smaller due to the emission of gravitational waves. Any other possible cause of the orbital decay is not considered. On the basis of this logical conclusion, they claim indirect proof of gravitational wave radiation. The orbits shrunk so there are gravitational waves. The scientific community has seized on this indirect proof and has been content to agree that gravitational waves are real. Due to the effect of the acceleration of the galaxy, Taylor and Weisberg do not think that the results of the test will improve. So, this is as close as we can get to proof of the existence of gravitational waves. Unlike electromagnetic waves, gravitational waves will have to remain a strong belief rather than an observed fact. Many sources of information, however, simply claim the reality of these waves and do not communicate their actual basis to the public. The Wikipedia article follows this trend.Lestrade (talk) 16:13, 19 January 2009 (UTC)Lestrade
 * There is no clear distinction between belief and fact in physics, it is just a matter of degree. The question is whether there is sufficient evidence or not to say that gravitational waves travel at the speed of light.  This should be answered not by our own opinions but by reference to reliable sources.  If we cannot find a reliable source to directly support the statement that gravitational waves travel at the speed of light then I agree that we should say something weaker. Martin Hogbin (talk) 22:16, 19 January 2009 (UTC)

The sentence "gravitational waves travel at the speed of light" is a categorical, unconditional, unqualified, explicit, declarative assertion. Can such a direct statement be made about something that has, and possible never will be, directly experienced?Lestrade (talk) 12:59, 20 January 2009 (UTC)Lestrade


 * There is no clear distinction between belief and fact in physics, it is just a matter of degree. There is a rash statement. I'd guess experimental physics people say measurements are facts, and the theoretical people would say that a theory that fits the known experiments is more than a "belief". The real issue here is whether one has to carefully state a theory to explicitly discount all areas where it is yet to be tested. I'd say that is an unlikely stance, particularly as it is hard to imagine every untested case, never mind list them all every time you state the theory. It suffices to state the theory (e.g. general relativity) and simply explore events that might disprove it. On that basis there is no call to invent caution over the speed of gravitational waves, or light waves of parsec wavelengths, that far exceeds any caution expressed in other areas of physics.
 * The speed of light is of fundamental importance in physics. It is the speed of not just visible light, but of all electromagnetic radiation, as well as gravitational waves [3] [4] and anything having zero rest mass. What has to be done is to rewrite this sentence to avoid the appearance of stating an observed measurement, when it really states a prediction based upon the best available theory. I've made a proposed change based upon this attitude. Brews ohare (talk) 15:15, 20 January 2009 (UTC)


 * There is a subtle point involved with the speed of light: it is a defined value not an observed one. See the discussion of meter. Brews ohare (talk) 16:23, 20 January 2009 (UTC)


 * The speed of light is defined, based on observations. All physics is constrained by the measurable speed of light.  The meter is defined by the observation of light, not the reverse.  The speed of light is not just a concern in theoretical physics.  Relativistic aspects related to the very measurable speed of light have numerous applications including computer circuitry and particle colliders.  There is no point restricting the scope to the theoretical. Novangelis (talk) 16:39, 20 January 2009 (UTC)


 * Novangelis: I see a storm brewing here. A bit of care in terminology might be a good idea. For example, the speed of light is "exactly 299,792,458 metres per second (m/s)". Obviously, this being a defined value, nothing can change it except international agreement; it is not changed by measurements. Do you agree? Brews ohare (talk) 18:18, 20 January 2009 (UTC)


 * A second observation: the word "theoretical" need not be pejorative, inasmuch as the theory agrees with the known observations. The theory is in effect a shorthand that encapsulates the observations. There is no implication that the predictions are "only a theory". Rather, the implication is that all considerations point to the prediction, including the experimental facts. Saying the theory predicts something thus does not imply a "restriction in scope". In fact, a prediction is weightier than "just the facts, ma'am." It implies that critique, correlation and sifting of all pertinent data has been exercised. Brews ohare (talk) 18:23, 20 January 2009 (UTC)


 * There is no storm brewing. We are both working towards the same goal.  We are taking a clunky passage and trying to improve it.  The fact that we are debating a few words does not indicate a conflict of goals.  I am not suggesting theoretical has any pejorative meaning. Einstein was a theoretical physicist. There is a subdivision of physics between theoretical and experimental, and the use of the word theoretical creates an artificial constraint.  More broadly, there is a contrast between the broader use of theoretical (which encompasses both) and applied physics which is also concerned with the speed of light (e.g. GPS). That should not be excluded, either.
 * We are working in territory that is all about technical definitions, and trying to write towards a lay audience while maintaining accuracy. This is not easy. I don't see any broad problems with what you wrote. It is better than where things started, but there is still vast room for improvement. We need to untangle observation from prediction in a readable manner. I've gone through more revisions with co-authors in the same room.  The speed of light in vacuo has been measured over great expanses of the electromagnetic spectrum (technically, I admit, not all). Since photons and gluons are the only observed massless particles, and, to my knowledge, the speed of gluons has not been measured, there are a lot of fine details to cover. The limits to which we have pushed particles with mass provides a measurable asymptote. Since the article describes measurement, the lead should be congruent. The details of observed with speed measured, vs. observed directly but not clocked vs. observed indirectly in particle detectors vs. predicted by a model and inferred from astronomical phenomena is a lot to cover. Expect me to disagree with what you write, but not what you are doing. Keep it up. Novangelis (talk) 19:40, 20 January 2009 (UTC)

Thanks for your conciliatory tone. A useful set of guidelines is Lead section. In particular, it suggests the lead should provide an overview of the article, not just a few points that might interest 80% of the readers. That objective might be a handful here. Brews ohare (talk) 20:31, 20 January 2009 (UTC)


 * I would say that there is sufficient evidence to say that all EM radiation travels at the speed of light and that this statement is supported by reliable sources. For example Feynman Vol2 18-5 pretty much says it. As always, it possible that one day somebody might discover that this is all wrong but at the moment I think we can state it as fact.  On the other hand most reliable sources seem to use more cautious language about gravitational waves.  There is no need to make a meal of it in the lead section so I suggest that we write something like, 'It is the speed of not just visible light, but of all electromagnetic radiation and it is also believed to be the speed of gravitational waves and anything having zero rest mass'. We can say more in the body of the article. Martin Hogbin (talk) 22:43, 20 January 2009 (UTC)

c2
If c is the fastest speed in the universe, then c2 must be p r e t t y, p r e t t y fast.Lestrade (talk) 00:06, 18 January 2009 (UTC)Lestrade


 * c2 isn't a speed. Speeds have units of distance/time, while c2 has units of distance2/time2. --Steve (talk) 05:51, 18 January 2009 (UTC)

Speed is a quantity that has the dimension length/time. Isn't it true that when dimensions of quantity are multiplied by each other, the product is of the same dimension of quantity, squared? For example, three feet times two feet are six square feet, but it is still feet. So, speed times speed is speed squared, but it is still speed. What else could it be that makes sense mathematically and physically?Lestrade (talk) 14:17, 18 January 2009 (UTC)Lestrade


 * No, that is wrong. The speed of light squared in units where c is 1 is also 1, the same value as the speed itself, but in all other units this is not the case. What exactly is the point that you are trying to make? Martin Hogbin (talk) 15:11, 18 January 2009 (UTC)


 * So you're saying that six square feet is a distance? Have you ever met someone who's six square feet tall? --Steve (talk) 17:02, 18 January 2009 (UTC)

Height or distance is a one dimensional quantity, measured as a line. A square is a two dimensional quantity, measured as a surface or area. Some people say that c2 is to be considered as a proportionality or conversion factor. As such, only the absolute number of approximately 34,704,709,264 (if the English miles/second dimensions are used) or 89,875,543,056,250,000 (if the Metric meters/second dimensions are used) would be considered as the factor. In either case, it seems almost silly to say that we know that this number indicates precisely the ratio of energy to a unit of mass. Rather, it probably merely means "a really big quantity."Lestrade (talk) 20:51, 18 January 2009 (UTC)Lestrade


 * If Joule/amu units are used, then c2 has the value 0.0000000001492. Not such a big quantity now, eh? :-) --Steve (talk) 06:43, 19 January 2009 (UTC)

Using those units, and assuming the literal truth of the equation e = mc2, one part of mass would convert to 0.0000000001492 parts of energy when a uranium nucleus is split. That's not much of a fission bomb.Lestrade (talk) 14:05, 19 January 2009 (UTC)Lestrade


 * That's right, 1 amu of mass converts to 0.0000000001492 joules of energy. In a big fission bomb, there's ~10000000000000000000000000 amu's of mass that get converted to ~1000000000000000 joules of energy. That's enough energy to vaporize a large city. --Steve (talk) 18:44, 19 January 2009 (UTC)

Proposed reorganisation
The section entitled 'overview' is not really an overview, and anyway this is the function of the lead section. I therefore propose that we merge the 'overview' section with the physics section and make it the first section of the article after the lead.

I also suggest that we remove references from the lead, which should be a brief overview of the article as a whole, and add them to the relevant parts of the body. Martin Hogbin (talk) 10:45, 21 January 2009 (UTC)


 * By WP:LEAD, the lead should be the intro, and a section called Introduction should not be present. Overview is about the same.  There are three paragraphs in Overview; the first belongs in the lead, although the two sentences might be separated.  The second belongs somewhere in the Measurement section.  Following the Michaelson-Morley experiment seems appropriate.  In that case, the third paragraph belongs with it.  The lead should conform to verifiability.  While references should remain, some of the details should move to the appropriate sections. Double citing may be overkill. That should reduce the density.  When Overview is gone, perhaps History should come to the top.  After that, Place a Terminology that differentiates in vacuo from in materials. The adoption of c can remain as a subsection.  The last subsection of the article "Speed of light set by definition" might follow.
 * As the first paragraph of the overview says, the speed of light is a fundamental constant. While including everything that relates to it would mushroom it out of control, we need to decide what is incidental and what is fundamental. I think we should introduce the Electromagnetic wave equation into the article. Magnetic permeability and electrical permitivity are fundamental to the speed of light. This article has a cobbled together feel to it.  A head to toe reorganization might be called for. Novangelis (talk) 13:00, 21 January 2009 (UTC)

I have just done a proposed rearrangement of the first half of the article, which you can find here. Have a look and see what you think Martin Hogbin (talk) 13:26, 21 January 2009 (UTC)


 * Martin: I think your lead looks good. I am glad it still has references in it, because there are so many Wiki articles that have unsubstantiated statements in them that I find it reassuring to see some citations indicating that the statements of the main points have some support beyond the whim of the author.
 * I think it is recommended that that there should not be references in the lead as this should be a summary of the body of article, which is where all the points in the lead are expanded upon and fully referenced. I will have a look at the MOS etc. Martin Hogbin (talk) 22:46, 21 January 2009 (UTC)


 * Somewhere in the article the various statements bearing upon using a definition for the speed instead of measuring it should be put together and clarified. That is a non-intuitive notion, and has to be made simple. Speed_of_light and the tantalizing statement: "In branches of physics in which the speed of light plays an important part, such as in relativity, it is common to use a system of units known as natural units in which c is 1; thus no symbol for the speed of light is required." could be part of this discussion. Brews ohare (talk) 13:47, 21 January 2009 (UTC)


 * Very nice. You broke it up into units that can be managed.  You introduced Doppler effect, which might serve as a segue to astronomy as we organize, and spacetime, which was mentioned in the lead and appeared nowhere else. A  might be in order. Novangelis (talk) 13:55, 21 January 2009 (UTC)
 * Thanks. We should also add a section on EM radiation as you suggest. Martin Hogbin (talk) 22:46, 21 January 2009 (UTC)

If no one objects I will past the new bit in place. We can then work on improving it. Martin Hogbin (talk) 22:46, 21 January 2009 (UTC)

Light as EM radiation
I have added a section on light as EM radiation. At the end, it says that the speed of EM radiation is independent of frequency. I know this has been confirmed by observations on pulsars but I cannot find a good reference. Any offers? Martin Hogbin (talk) 10:35, 24 January 2009 (UTC)

Speed fixed by definition
Brews, you have raised some points about this subject, but I am not quite sure what you are saying. Are you suggesting that further explanation of this topic is in required in the article? Martin Hogbin (talk) 10:37, 24 January 2009 (UTC)


 * Yes, I find the notion of defining the speed of light rather than measuring it is counterintuitive. It seems circular without more explanation. Brews ohare (talk) 15:21, 24 January 2009 (UTC)


 * OK, I will try to write something on the subject. Martin Hogbin (talk) 16:47, 24 January 2009 (UTC)

Speed of gravitational waves
According to general relativity, gravitational waves travel at c, this is well known, as indicated by the reference from Hartle that I gave. There may be other, unconfirmed and not generally accepted theories, in which this is not the case.

The paper by Carlip that you quote refers to a recent attempt to measure the 'speed of gravity', and what is says is that the experiment did not actually measure the speed of gravity but the speed of light - there was some controversy over this at the time. It does not suggest in any way that they are different, just that the particular experiment failed in its objective to actually measure the speed of gravity.

It is true that the speed of gravity has not been confirmed experimentally but the only accepted, and well well-tested, theory of the subject (GR) predicts that it will travel at c. I believe that it is therefore justified to make the stronger statement that I made. Martin Hogbin (talk) 16:27, 24 January 2009 (UTC)


 * Hi Martin: I will leave you to your own devices. However, I find your overriding of my changes to be a bit muddy-minded, failing to clearly distinguish between observations (which necessarily refer to some physical medium, like outer space) and theoretical predictions, which in the case of classical electromagnetism have two aspects: behavior in free space (a hypothetical reference state, which is unobtainable like absolute zero) and predictions for a particular medium (like outer space) where constitutive relations for the relative permittivities and permeabilities are required. Brews ohare (talk) 16:55, 24 January 2009 (UTC)


 * There is nothing muddy-minded about my statement that gravitational waves are believed to travel at c. This is a clear prediction of GR, which the currently accepted theory of such matters. Thus 'believed to' is correct in he sense that it is what physicists expect to be the case.


 * My reason for removing you reference to outer space was as follows: The question is whether all EM radiation travels at the same speed in free space.  Experiments have confirmed that a wide frequency range of EM radiation travels at the same speed in outer space.  Now it would be startling if this same result did not occur in the theoretical free space.  The reverse, of course is not true; a measurement made in free space would not necessarily predict the results of a measurement made in outer space, but the fact that the a speed has been confirmed in outer space very strongly implies that the result is true in free space.  Thus the mention of outer space is somewhat superfluous. Martin Hogbin (talk) 17:24, 24 January 2009 (UTC)


 * I have just had a look and I agree that the statement was not clear as I left it, I hope it is now. Martin Hogbin (talk) 17:30, 24 January 2009 (UTC)

No one has ever detected gravitational waves, but waves of gravitation travel at the speed of light.Lestrade (talk) 20:38, 12 February 2009 (UTC)Lestrade

Distinction between theory and observation
Brews, are these quotes from the FA review or your comments? Martin Hogbin (talk) 17:00, 24 January 2009 (UTC)

The article frequently stumbles over whether statements like "all EM radiation travels at speed c" is (i) a definition, or (ii) a physical observation or (iii) a posit of some physical theories (e.g. relativity theory, quantum gravity theory etc.).


 * Agreed, this is now made clear in the EM section.Martin Hogbin (talk) 17:00, 24 January 2009 (UTC)

The article also frequently fails to point out what medium is under discussion: for example: is it the ideal free space which is the unobtainable reference state where c = c0, or is it outer space where the speed of light may not be c0 but possibly any measured value whatsoever?


 * Agreed also, I have left in your reference in the EM section.Martin Hogbin (talk) 17:00, 24 January 2009 (UTC)

I've changed some of these occurrences, but more are out there. Brews ohare (talk) 16:41, 24 January 2009 (UTC)


 * Yes, rather than always add 'in free space' everywhere, which may get rather boring and repetitive, we could use c in some cases to avoid that ambiguity. Martin Hogbin (talk) 17:00, 24 January 2009 (UTC)


 * The statement in the article reads:
 * According to classical electromagnetism, the speed of electromagnetic radiation in free space is the same for all frequencies. Observations confirm that any variation of the this speed with frequency is extremely small.
 * Here is the problem: free space is unobtainable; by definition no observation of the speed of light in free space is possible. By definition in free space this speed is c0 and is (i)independent of frequency (ii)independent of direction (iii) independent of field strength, (iv) independent of polarization. Therefore, it is impossible for observations to "confirm" anything having to do with free space (e.g it's linearity, isotropy or dichroism). All that observations can do is establish to what degree a particular medium (like outer space or the QCD vacuum) behaves in a manner like free space. In particular, observations can determine whether the relative permittivity and permeability of outer space are very nearly unity, and very nearly independent of frequency. Brews ohare (talk) 17:54, 24 January 2009 (UTC)


 * That is an interesting point. The actual definition says 'The metre is the length of the path travelled by light in vacuum during a time interval of 1/299 792 458 of a second', exactly what is meant by light in this context is not clear.  However, the 17th CGPM also made the iodine-stabilised helium-neon laser 'a recommended radiation for realising the metre'. So it would seem that it is visible light that they have in mind.


 * Now if it transpires that there is some small variation in the speed of EM radiation in free space with frequency (and no definition could prevent this), the BIPM would need to make clear exactly what they meant by 'light' in the definition of the metre. They might well choose the iodine-stabilised helium-neon laser but they could in fact choose some other standard, possible based on some notional frequency of EM radiation.  This would depend on the theory that was used to explain the speed variation.


 * At present, however, the standard says 'light' and recommends the use of a particular frequency of visible light so I would asy that it is theoretically possible for gamma rays to travel at a different speed from (visible)light. Martin Hogbin (talk) 18:11, 24 January 2009 (UTC)


 * Also see my comments in the section above. Martin Hogbin (talk) 18:11, 24 January 2009 (UTC)

Hi Martin: Your comment: "Now if it transpires that there is some small variation in the speed of EM radiation in free space with frequency (and no definition could prevent this)" is, in my opinion, based upon a misconception that measurement can impinge on the defined properties of "vacuum". In practice, any measurement must be in a medium. Hence, according to BIPM, the measurement must be corrected for its non-ideal behavior. Changes in accuracy of the measurement will affect the corrections to be made in this medium in order to refer it to the defined state of "vacuum", but do not change the defined "vacuum". It appears to me that you see free space as something real, while it is only an unobtainable reference state, like absolute zero. Brews ohare


 * Yes, I understand that free space is an ideal concept, and I agree that the speed of EM radiation in free space is not expected to vary with frequency, but experiment is the final arbiter of all things in physics. If it was found that there was a variation with frequency after all allowances had been made for any medium (such as interstellar space) then physics would have to account for this.


 * The point is that the experiment you quote does indeed set limits on the variation of the speed of EM radiation in outer space but this in turn sets experimental limits on its variation in free space. Martin Hogbin (talk) 18:41, 24 January 2009 (UTC)

Hi Martin: Our editing session overlapped. Where the "vacuum" of BIPM or "Free space" is concerned, experiment is not an arbiter; rather it is simply a matter of convention and in turn a matter of convenience. Just like any other standard.

In particular, if the meter established in air at two different light frequencies were to differ, the difference would be attributed to the medium "air" and not to the vacuum of free space. If some "universal dispersion correction" showed up in all media, and if the international community decided that making this correction all the time was a nuisance the definition of "vacuum" could be changed by international agreement to incorporate "the dispersion of the vacuum". The properties of the "vacuum" are not a reflection of nature, but of convention. The properties of nature are expressed in permeabilities and permittivities relative to the "vacuum". Instead, a 'standard medium" could be adopted, but that would require tests to see of the standard were met. Apparently, corrections to refer to a "standard medium" is viewed as a more cumbersome approach. Brews ohare (talk) 18:27, 24 January 2009 (UTC)
 * Of course if the speed of EM radiation were found to vary with frequency in air then the obvious explanation for this would be an effect of the air, and is exactly the case air is dispersive. If any variation of speed with frequency were found in the case of outer space the most likely explanation would be dispersion caused by the interstellar medium, this is not disputed. But, there is nothing in physics which says that the speed of EM radiation in free space cannot possibly vary with frequency.  One might say that you can never prove that it does, but you can set limits on how much it doesn't. Martin Hogbin (talk) 18:56, 24 January 2009 (UTC)


 * Martin: We are a bit at cross purposes here. The idea of setting limits on properties of free space suggests that one can measure its properties. But that is not so. We say a dollar is 100 cents. You cannot observe dollars and say that the dollar is somewhere around 100 cents. Likewise you cannot observe free space and say the speed of light is around c0. You also cannot measure its linearity, dichroism, dispersion or any other property. By definition, free space exhibits no dispersion, no anisotropy, no nonlinearity and no dichroism. Any measurement in any medium can be compared to this "ideal" vacuum, and may, of course, exhibit a different value for c, dichroism, nonlinearity, anisotropy and whatever.  Brews ohare (talk) 19:02, 24 January 2009 (UTC)
 * The standard refers to light not EM radiation in general. If the standard did just refer to EM radiation the we could have the bizarre situation that the metre varied with frequency!  Clearly this is not the object of the standard and such a situation would be absurd. For the moment at least let us take the standard to be The metre is the length of the path travelled by light from a iodine-stabilised helium-neon laser in vacuum during a time interval of 1/299 792 458 of a second.  Are you happy to do that just to make discussion easier? Martin Hogbin (talk) 19:12, 24 January 2009 (UTC)


 * Hi Martin: "we could have the bizarre situation that the metre varied with frequency!" I assume you mean that it could happen that measured meters were different at different frequencies. I don't see that as resulting in a bizarre situation: it would be attributed to the medium in which the measurement was made, and a correction implemented. The selection of a particular frequency is more about getting the second right in the measurement than about dispersion.


 * The standard refers to a particular frequency, and to a distance in something called "vacuum". However, it is empty as regards identifying whether a "vacuum" is a physically realizable medium, and if it is, just how it might be identified as having been realized. I'd suggest that to insure a good realization of the vacuum is available in the lab (or in outer space), the properties of the lab medium would be measured (isotropy, dispersion, linearity etc.) and the departures from "ideal" would be be used as corrections. According to BIPM, measurements in any medium are "to be corrected for nonideality" according to "good practice". "Good practice" is a moving target, and various corrections have been added over the years, for example, gravitational dilation corrections to the second. The standard also is empty as regards how the measurement using a different frequency might be corrected to refer to the standard.


 * So where do we go from here? It appears that BIPM is not very helpful. Do you subscribe to free space? Brews ohare (talk) 19:45, 24 January 2009 (UTC)


 * So what you seem to be saying is that the speed of all frequencies of EM radiation through free space is the same by definition? Martin Hogbin (talk) 10:11, 25 January 2009 (UTC)

Hi Martin: Yes. By definition of "free space", rather than by definition of the speed of light. Brews ohare (talk) 14:09, 25 January 2009 (UTC)

When I look at the article you link to above I get this (my emphasis),'In the reference state of free space, according to Maxwell's equations, electromagnetic waves, such as radio waves and visible light (among other electromagnetic spectrum frequencies) propagate at the defined speed of light'. Is that what you are referring to? Martin Hogbin (talk) 14:55, 25 January 2009 (UTC)


 * I made some changes to Speed of light. See what you think, Martin. Brews ohare (talk) 15:04, 25 January 2009 (UTC)


 * Firstly, can you confirm that the quote above is what you believe defines the speed of all EM radiation to be c. If not can you pleas refer me to where this definition is made clear.


 * Regarding your additions, although what you write is generally correct, most of it is, in my opinion, unnecessary. All physical measurements are approximations.  When standards laboratories use the speed of light to delineate the metre they take account of all the possible sources of error, including the fact that the medium used is not precisely free space. Any uncertainty in measurement or any necessary corrections will be quoted in the result.  There is no reason to pick on one particular potential source of error and not mention all the others.  We should assume that when a standards laboratory determines the length of the metre that they do it properly.  The medium used may be far from being the greatest source of uncertainty.  Do you, for example, know at what pressure measurements are made and what the reduction in the speed of light would be expected at that pressure? Martin Hogbin (talk) 15:24, 25 January 2009 (UTC)


 * The gist of the changes is not that "All physical measurements are approximations", but to clarify that free space is not a real medium, but an idealization like absolute zero. This status for free space is a source of unending confusion throughout the article. Recognition of the role for free space makes it common sense that one can define the speed of light in free space to be any number you like (of course, the expense of converting to a value like 1 m/s would preclude using a nice round number). Otherwise it seems convoluted.  Hence, your remark "There is no reason to pick on one particular potential source of error and not mention all the others" is not germane. It also is wide of the mark in suggesting that only one source of error has been raised: it refers to "all necessary corrections". If the meaning is this unclear, my additions should be rewritten to make the requisite points. Brews ohare (talk) 15:58, 25 January 2009 (UTC)


 * The sentence  Observations confirm that any variation of the this speed with frequency is extremely small. referring to free space and not to a material medium, is indicative of a complete misconception and must be changed. Brews ohare (talk) 16:09, 25 January 2009 (UTC)


 * You seem to be avoiding the question of where you think that the speed of all frequencies of EM radiation is defined to be the same.


 * If you look here [] you will see how the refractive index of air varies with pressure. It is within 3 parts in 10^5 of unity at a pressure of 3 in of Hg.  A UHV system will get down to around 10^-12 Torr (1 Torr is 1mm of Hg), so you can see that the effect of residual gas is going to be quite small, and that is before any correction is made.


 * Regarding the point that free space is not an achievable medium but an idealization this is well known and is made quite clear and discussed at some length in the article on the subject; we do not need to repeat it here. Can you explain why, as you put it, 'This status for free space is a source of unending confusion throughout the article'.Martin Hogbin (talk) 16:25, 25 January 2009 (UTC)


 * The answer to one question is: yes, I do believe all EM radiation travels at speed c in free space . The response to your observations on partial vacuum is: yes, departures from free space behavior may be small in some media. It also appears to be true that experimental error in terrestrial measurements are quite large, and resolving the distinctions between contenders for "best" realization of free space is beyond experiment at the moment. None of that bears upon the logical status of free space as an unobtainable reference state that cannot be measured. Brews ohare (talk) 16:32, 25 January 2009 (UTC)

Speed of all EM radiation is c
Firstly, can you confirm that the quote above is what you believe defines the speed of all EM radiation to be c?

I am unsure what quote is referred to here. However, the speed of all EM radiation in free space is c by definition, because, by definition, free space has no dispersion, no dichroism, no nonlinearity, and no anisotropy. The only issue in any actual measurement is whether it actually was done in a medium that approximates "free space", and just what corrections have to be made. Present theory on the QCD vacuum and the quantum vacuum suggest that realization of one of these vacuums is not tantamount to realization of free space, as these vaccua in theory are not isotropic, do exhibit dichroism and are nonlinear at large field strengths.Brews ohare (talk) 16:26, 25 January 2009 (UTC)


 * The statement I was referring to is from the article on free space where it says [my emphasis], 'In the reference state of free space, according to Maxwell's equations, electromagnetic waves, such as radio waves and visible light (among other electromagnetic spectrum frequencies) propagate at the defined speed of light'.


 * The speed of light is defined to be a certain value by international agreement, thus it is fixed by definition. Your statement is not a recognized definition but it is your opinion, based on what you think free space is.  The two are not the same. Martin Hogbin (talk) 16:42, 25 January 2009 (UTC)

The speed of all EM radiation in free space is c by definition, because, by definition, free space has no dispersion, no dichroism, no nonlinearity, and no anisotropy. The proof of these statements is the defined values of electric constant, magnetic constant, impedance of free space and speed of light in free space. All these properties of free space by definition are dispersionless, field independent, and scalars (not tensors). If they are plugged into Maxwell's equations, they result in dispersionless, isotropic, field-independent propagation of EM waves in free space.

Of course, every realizable medium, be it outer space, partial terrestrial vacuum, QCD vacuum, or whatever, has a permittivity and permeability that can be measured to some level of accuracy and will exhibit (to a degree determined theoretically and/or experimentally) anisotropy, dichroism, nonlinearity and dispersion. Brews ohare (talk) 16:49, 25 January 2009 (UTC)


 * That is not what by definition means, you are talking about a calculation using Maxwell's equations. I fully agree that to the best of our knowledge Maxwell's equations are accurate (on a macroscopic scale) and no deviations from them have been observed, but, this does not mean that such deviations, and in particular a variation in the speed of light with frequency, could never be found to exist. Only experiment will tell.  The speed of all frequencies of EM radiation is certainly not fixed by definition. Martin Hogbin (talk) 17:07, 25 January 2009 (UTC)


 * Martin: If you have the permittivity and permeability of free space established by definition, then Maxwell's equations inevitably lead to the dispersionless propagation of EM waves. So it appears that your quibble amounts to the suggestion that Maxwell's equations themselves may be incorrect. A modified Maxwell theory might predict that a free space with dispersionless isotropic field-independent permittivity and permeability does exhibit dispersion. I'm not sure that this hypothetical case clarifies anything. The status of free space as an unmeasurable ideal state remains unimpaired. The establishment of the correctness of the "new, modified" Maxwell theory would be based on observations in real media, not by observations of free space. Statements like Observations confirm that any variation of the this speed with frequency is extremely small. suggesting that experiment can be applied to free space should be purged. The most that can be said is that media exist (maybe partial vacuum or outer space) where the dispersion is very small.  Brews ohare (talk) 17:16, 25 January 2009 (UTC)


 * What the hypothetical case I described does establish is that the the speed of all frequencies of EM radiation is not fixed by definition. It is however fixed according to our current best theory, but that is not the same thing.  I am not sure how I can make this any clearer.  Once we have established this point, we can go on to discuss the content of the article. Martin Hogbin (talk) 00:00, 26 January 2009 (UTC)

Granted; not by definition; also, "speed of all frequencies of EM radiation" is meaningless without stating whether "free space" is meant, or some real medium. If "free space" is meant, then experimental observation is impossible and irrelevant. Brews ohare (talk) 00:46, 26 January 2009 (UTC)


 * When I've seen the term "free space", it usually means the limit of a better and better physical vacuum. Obviously experiments can shed light on this. You seem to be using the term differently. What would you call the limit of a better and better physical vacuum? Whatever you call it, that's what we should be talking about, since that's what's actually relevant to our universe. :-) --Steve (talk) 06:09, 26 January 2009 (UTC)


 * Brews, as Steve has said, free space can be considered as a limit of a physical vacuum. As far as we can tell, the effects of a gas on EM radiation reduce with pressure, as would be expected.  Based on the data I presented earlier, on refractive index vs pressure, the effect of the medium of outer space on the speed of light would be expected to be very small indeed.  Experiment has shown that the speed of EM radiation in the medium of outer space is not affected by its frequency within the measurement uncertainty of the experiment.  From everything we know about EM radiation, from both theory and experiment, we would expect any variation of speed with frequency to be less in the medium of ideal free space that it is in the medium of outer space.  Thus it is justifiable to say that experimental evidence sets limits on the variation of speed with frequency in free space.


 * Of course it is always possible to imagine that in true free space something very strange would happen to light and the extrapolation from increasingly better physical vacua would break down, but physics is full of idealisations and, unless there is good reason not to, it is usual to assume that the ideal case can be calculated an extrapolation of real experiments. Martin Hogbin (talk) 09:38, 26 January 2009 (UTC)

"Free space" can be viewed a limiting case, I agree. It can be approached by some real media, but it cannot be reached. Theoretically, there is no "vacuum" with its properties. Experimentally, we can establish a medium resembles free space within experimental error, but of course there is always the possibility that further refinements in technique will reveal differences.

The EM properties of free space are defined by ε0 and μ0 (or by some pair of these two and Z0 and c0). These values constitute a reference case. It might be possible that a real medium comes close to the reference. Unfortunately, because of experimental error, it never can be established that any real medium is free space, only that it is close to free space.

That is why the logical impossibility of "measuring" free space has to be recognized. All we can do is measure a real medium and point out that it is as close to "free space" as we can tell within our experimental limitations.

Therefore, no statement in the article should suggest that we can measure free space. Nor, can we say measurement puts "limits" on the free space parameters: they are defined; they are not measured. They could be totally different, they could be tensors, etc etc. There is no underlying philosophical "meaning" to free space having the parameters it does. The values of ε0, μ0, Z0 and c0 are simply historical accidents.

We could choose instead a real standard medium like air that could be realized. That would not prove to be an advantage, because we could never be sure that our reference sample of air was exactly the same as the standard. We'd have to maintain a standard air sample for comparison and reassure ourselves that it was not changing with time because of improper storage, etc.

So choosing an arbitrary ideal reference is easier, especially if we can maintain a vigilant list of "best practices" to relate any measurement in real media to the standard. Hopefully, the standard has the merit that these "best practices" are not too difficult to apply in practice. (You tell me if preparation of an atomic clock is simple?) The BIPM and its associates decide by international agreement whether the reference is appropriate, and will change it if there appears to be a simpler or a more accurate standard that would make for "best practices" that were easier to use, or accessible to greater accuracy experimentally.

A statement like: ''we would expect any variation of speed with frequency to be less in the medium of ideal free space that it is in the medium of outer space. Thus it is justifiable to say that experimental evidence sets limits on the variation of speed with frequency in free space.'' is misleading: given Maxwell's equations we know that "free space" has absolutely zero dispersion. And we know that the validity of Maxwell's equations has nothing to do with the values of the parameters in free space. Maxwell's equations are established by measurements in the real world, on real media. Whatever real media behave like has no bearing upon the parameters of free space, neither upon the actual values nor their scalar dispersionless nature.

We could imagine a hypothetical Universe where every real medium we measured exhibited dispersion. These physical facts need not impact the parameters of free space. It might be decided that reference to free space was still the simplest methodology. Or, the BIPM might decide that free space would be modified. That is a decision based upon practicality and politics.

Again: because of its defined properties, no statement in the article should suggest that we can measure free space or that experiment "constrains" the parameters of free space to have particular values, or to be scalars, or to be dispersion free, or … .Brews ohare (talk) 13:00, 26 January 2009 (UTC)


 * Imagine: scientists discover a tiny dispersion in some yet-unexplored frequency range, and find that the better they make their vacuum, the strength of the dispersion stays constant. Someone comes up with a theory where Maxwell's equations have a small modification where the modified Maxwell's equations have a dispersion in an perfect vacuum, and further experiments confirm these modified equations until it is universally accepted by scientists. I would describe this as "it turns out there's dispersion in free space". How would you describe it? --Steve (talk) 15:59, 26 January 2009 (UTC)


 * The scientists' "vacuum" exhibits dispersion in a certain frequency range. Is this "vacuum" like free space? No, definitely not, because free space has no dispersion at any frequency (according to Maxwell's equations as we know them, coupled with ε0, μ0, Z0 and c0). Does that mean Maxwell's equations are wrong, or does it mean that this "vacuum" has some interesting properties (apparently previously unknown), and are these properties shared by other candidates for "vacuum"?? A great deal of thought will go into settling the matter.


 * Will this thought include modification of the properties of free space? No. When the dust settles, it might be decided that today's free space is more cumbersome to use than necessary (given the new version of Maxwell's equations), and should be changed. Or, it may be decided, like the decision to stick with electrons having negative charge instead of positive, that the change is just more trouble than it's worth. Or, it may be decided that "vacuum" as prepared by these scientists has led to discovery of some new fundamental particle, the "vacuum polaron"(?) that requires a new "best practices" correction when the metre is measured in this frequency range in the presence of "vacuum polarons".


 * In fact, a new "best practices" correction is exactly what will happen should experimental accuracy advance to the point that the properties of the quantum vacuum can be observed. In the case of the second, the addition of new "best practices" has included the addition of the caution that the clock has to be corrected to account for gravitational time dilation and to account for temperatures above T=0 kelvins. These corrections became a requirement as advances in technique made them necessary. Brews ohare (talk) 17:46, 26 January 2009 (UTC)


 * Brews, I am having increasing difficulty in understanding your problem with free space. It is an unobtainable ideal just like many other things in physics.  As you say above, the time standard refers to caesium atoms at absolute zero - this is known to be impossible.  School physics questions often referred to massless and frictionless pulleys - they do not exist.  Free space is no different, it a conceptual medium whose properties we choose to extrapolate from real media. Martin Hogbin (talk) 18:27, 26 January 2009 (UTC)

Martin: I have no trouble with free space. I have trouble with statements in the article that suggest measurements have something to do with free space. Measurements are corrected to refer to free space, but do not impinge upon the defined properties of free space. Brews ohare (talk) 18:29, 26 January 2009 (UTC)
 * It is no different from the examples I gave above. Although they do not actually exist we can ascribe properties to massless pulleys and we can determine what we expect those properties to be from measurements made using real pulleys. Martin Hogbin (talk) 18:41, 26 January 2009 (UTC)


 * Inasmuch as your thinking has evolved to the point that these distinctions appear obvious, I have modified the article accordingly. Brews ohare (talk) 18:57, 26 January 2009 (UTC)


 * There has been no evolution in my thinking and, it would seem, there has been none in yours. Real measurements can be used to set limits for free space and the article should indicate this. I believe that your modifications should be reverted, but let us wait to see what others say. Martin Hogbin (talk) 19:12, 26 January 2009 (UTC)

Thank you, Martin. Brews ohare (talk) 19:30, 26 January 2009 (UTC)


 * Brews, you need to decide whether: (1) Free space is the extrapolated limit of a better and better vacuum, analogous to absolute zero, frictionless tabletops, etc.; or (2) Free space is where Maxwell's equations have (blah) form. You can't have it both ways, because (1) can be investigated by careful experimental and theoretical physics, while (2) cannot. You evidently have chosen (2): When I proposed a hypothetical case where (1) and (2) disagree, you responded by calling (2) free space and calling (1) a "vacuum" in scare-quotes. You agree so far? Questions: [a] Is there literature support for saying that (1) is not actually the true definition of free space? [b] In the context of this part of the article, I believe it makes more sense to discuss (1) than (2), since (1) is the basis for statements about the universe in which we live while (2) is not. Supposing that we wanted to discuss (1), what would you propose calling it? "Vacuum" in scare-quotes isn't a good option. :-) --Steve (talk) 19:44, 26 January 2009 (UTC)

Hi Steve: An interesting formulation of the discussion. Under (2), Maxwell's equation can be written for any medium: you just need the constitutive equations. The relevance of free space under (2) is that the constitutive equations are simply specification of ε0 and μ0. Evidently this specification does not need to apply to any real material (although it might). Free space is just this hypothetical material. As regards (1) there may be a sequence of steps (e.g. evacuation of a flask) that cause ε → ε0 and μ → μ0). Such a sequence might suggest free space is most nearly to be realized by pursuing this sequence. But so what? You can measure c in any medium and make corrections so your results are referred to free space, whether or not media approximating free space actually exist. There is zero information in free space and its constants.

Is there literature support for saying that [some limiting sequence] is not actually the true definition of free space? I'd say the NIST web site where ε0 and μ0 a stated as defined values is the true definition, not the possibility of some limiting sequence. Maybe there is some historical commentary on this point, but it has nothing to do with our final definition. And, just to be nasty, is there any literature supporting the limiting sequence idea?

In the context of this part of the article, I believe it makes more sense to discuss (1) than (2), since (1) is the basis for statements about the universe. Free space is not about the universe. Measurements of real constitutive relations for real media, like outer space or terrestrial vacuum, is about the real universe, quite independent of what we adopt for free space.

Historically, the yard may be related to the arm-length of monarchs and the metre to the dimensions of the planet Earth, but neither has any relation to the Universe as observed in science. Brews ohare (talk) 20:40, 26 January 2009 (UTC)

To put things differently: there is a continuing effort experimentally to determine if the speed of light in space is isotropic, and to see if it varies with frequency. Evidently, Maxwell's equations using ε0 and μ0 predict no success. Would a success prove that space is not a good realization of "free space", but a medium with its own constitutive equations, or would it prove Maxwell's equations are wrong, or something else (maybe some weird aspects of general relativity: varying speed of light theories)? None of these three categories of conclusion would change ε0 and μ0. Brews ohare (talk) 20:57, 26 January 2009 (UTC)


 * You say: "Free space is not about the universe. Measurements of real constitutive relations for real media, like outer space or terrestrial vacuum, is about the real universe, quite independent of what we adopt for free space." Precisely. A statement about free space (as you understand the term) is not a statement about our universe or its laws. So why would we ever make a statement about free space in this article? I say we replace "free space" with "vacuum" everywhere in the article.


 * Also, here's a question: If you really believe NIST thinks of free space as something other than the limit of a better and better vacuum, then how do you propose that a careful scientist checks that her meter-stick is the right length? After all, if the meter is defined in relation to the speed of light in free space, and no real-universe measurement could ever give information regarding the speed of light in free space, then there isn't any way to confirm that a meter isn't the size of an atom, right? --Steve (talk) 22:01, 26 January 2009 (UTC)

Hi Steve: Changing terms to "vacuum" we still have cO as the speed of light in a hypothetical medium. Only now "vacuum" is vague and subject to misinterpretation as a "real" medium. if the meter is defined in relation to the speed of light in free space, and no real-universe measurement could ever give information regarding the speed of light in free space, then there isn't any way to confirm that a meter isn't the size of an atom

The first condition is met: the meter is defined in relation to the speed of light in free space

The second condition is met: no real-universe measurement could ever give information regarding the speed of light in free space
 * The speed of light in free space is cO.

The conclusion is false: there isn't any way to confirm that a meter isn't the size of an atom

Here is how it could be done. Measure the meter in air, assuming the definition applies in air, not free space. Measure the refractive index of the air. Adjust the size of your meter according to the index. Brews ohare (talk) 22:46, 26 January 2009 (UTC)


 * The article you link talks about measuring the index of refraction of air using an interference refractometer. I've used one of these before. The way it works is you measure the ratio of the index of refraction of air, on the one hand, to the (literally) extrapolated limit of a better and better vacuum, on the other. Not the ratio to free space. --Steve (talk) 00:35, 27 January 2009 (UTC)

Hi Steve: The issue here is the usual one that occurs in setting up a standard, namely the "accumulated wisdom" effect. For example, how do we determine that one clock is better than another clock? It is partly accumulated experience on how well supposedly identical clocks agree, and partly a theoretical assessment as to what influences might disturb the clocks and how big these disturbances are expected to be. When experience shows the clocks don't all agree, we look around and find that it is gravitational dilation, an effect within theoretical estimation, one we should have seen coming, but one that simply was overlooked. Or, maybe its something else.

As technique improves, discrepancies in indices will show up. Sample vacuums that are supposedly identical turn out to be different. In accounting for the difference, maybe some of the observed discrepancy is due to quantum fluctuations of the vacuum, or whatever. In conjunction with theory we decide what these contributions are, learn how to avoid them, or account for their effects in the sample vacuums, and subtract them from measurements to get the "free space" value.

In dealing with air, present practice seems to be based upon some formulas that people feel have been well tested against different air samples. You just measure the partial pressures of water, CO2 etc. in your sample and calculate the index for your case. Brews ohare (talk) 03:30, 27 January 2009 (UTC)


 * The accumulated wisdom of all the metrology experiments measuring a meter moves towards the goal of defining a meter based on what the speed of light is in an extrapolated better and better vacuum. This isn't always a direct measurement, for example as you suggest you can first measure the meter-using-air and then multiply it by the measured quantity (nair/nextrapolated better and better vacuum). But the extrapolated-better-and-better-vacuum is what's being measured, either way. On the other hand, there's no accumulated wisdom whatsoever for measuring a meter based on the speed of light in "free space" (as you define it), because no aspect of "free space" is measurable. If you're willing to accept that a legitimate measurement of the meter is to use refractive interferometry along with measurement of light in air, then you've in fact already agreed that the meter is actually defined in terms of the speed of light in the extrapolated limit of a better and better vacuum. --Steve (talk) 04:24, 27 January 2009 (UTC)

Light as electromagnetic radiation
Rather than continue the long discussion on the philosophical nature of free space, I would like to suggest the reversion of the wording of the section above from Brews' version:

'According to classical electromagnetism, the speed of electromagnetic radiation in free space is the same for all frequencies. Observations in outer space and in good terrestrial vacuums confirm that in such mediums any variation of the this speed with frequency is extremely small, suggesting that, in this regard, they are good approximations to free space'.

to my original version:

'According to classical electromagnetism, the speed of electromagnetic radiation in free space is the same for all frequencies. Observations confirm that any variation of the this speed with frequency is extremely small'.

I believe that my version is clearer, fully justified, and says all that needs to be said. I am not against adding more detail but I think that Brews' attempt to make some kind of philosophical point in the article is wrong and unhelpful to the reader. Martin Hogbin (talk) 09:37, 27 January 2009 (UTC)


 * Hi Martin: I agree that discussion strayed from topic.
 * You have inserted the sentence Observations confirm that any variation of the this speed with frequency is extremely small.


 * What observations are meant? Observations in outer space or terrestrial vacuum, perhaps, or observations in a medium with parameters ε0 and μ0? Probably you mean the former.


 * A medium with ε0 and μ0 necessarily exhibits no dispersion. (Unless we digress upon the validity of Maxwell's equations, which is an interpretation of your sentence that no reader would discover without further discussion.) By definition, free space has ε0 and μ0, and therefore exhibits no dispersion.. The sentence should be revised to indicate that observations meant are observations in outer space or terrestrial vacuum, not free space. Brews ohare (talk) 16:31, 27 January 2009 (UTC)

I suggest:

'According to classical electromagnetism, the speed of electromagnetic radiation in a perfect vacuum is the same for all frequencies. Observations confirm that any variation of this speed with frequency is extremely small'.

I think everyone will understand what a perfect vacuum is; it's the extrapolation of a better and better physical vacuum; while OTOH a lot of readers will not have heard of free space. The phrase "speed of light in vacuum" is an order-of-magnitude more common on Google than "speed of light in free space". NIST and CODATA both use "vacuum", not "free space". And moreover, to the extent that that "free space" means something different from "perfect vacuum" (and I'm still not sure it does in common usage), "perfect vacuum" is what we want to talk about, since "free space" (as Brews uses the term) has no necessary relationship to anything in our universe, while this is a physics article about our universe and its laws. --Steve (talk) 17:18, 27 January 2009 (UTC)


 * I am happy with that although I have a couple of minor points. I think we should decide on what term we are going to use (free space, vacuum, perfect vacuum) and use it consistently throughout the article; I have no strong opinion on which.  My second point, I have just noticed, is actually with my own wording, '...any variation of this speed with frequency is extremely small', suggests that a small variation has been actually observed whereas all the experimental evidence shows no variation within the experimental uncertainty.  Perhaps, '...observations and experiments to date confirm this'. Martin Hogbin (talk) 19:07, 27 January 2009 (UTC)


 * The terms "free space" or "vacuum of free space" are not uncommon. Another advantage of the term "free space" is that there is another article free space where considerable discussion of various meanings of the term "vacuum" and other links are provided. Brews ohare (talk) 19:33, 27 January 2009 (UTC)


 * How about,

'Measurements based on the arrival of electromagnetic radiation from distant astrophysical events puts severe limits on the possible variation in the speed of light with frequency'. This is much closer to what is actually stated in our reliable source and it shows the context of radiation travelling through outer space. Martin Hogbin (talk) 19:37, 27 January 2009 (UTC)


 * How about, 'Measurements based on the arrival of electromagnetic radiation from distant astrophysical events puts severe limits on the possible variation with frequency of the speed of light traveling through outer space'. Brews ohare (talk) 19:46, 27 January 2009 (UTC)

I prefer Martin's version. It's overly modest to say that they merely set limits on dispersion in outer space; with a bit of simple analysis, it actually sets quantitative limits on the dispersion in the "extrapolated better and better vacuum" (call it what you will). --Steve (talk) 22:00, 27 January 2009 (UTC)


 * My version is closer to what the source actually says, or at least the abstract. It does not mention, free space, vacuum, or outer space.  I guess the reader can work out for themselves what the light travels through to reach us from a distant astrophysical event. Martin Hogbin (talk) 22:44, 27 January 2009 (UTC)


 * The so called "extrapolated better and better vacuum" is exactly free space and has exactly zero dispersion. There is no argument about the dispersion of free space and it has no limits or error bars placed upon it. The free space parameters (or those of the "extrapolated better and better vacuum" ) are defined to be the frequency independent scalars ε0 and μ0. This definition makes no reference to a limiting process.
 * Rather than say something clear, we will "leave it to the reader" or to "a bit of simple [but apparently too complex to put into the article] analysis". Great. Brews ohare (talk) 23:33, 27 January 2009 (UTC)
 * The sentence Measurements based on the arrival of electromagnetic radiation from distant astrophysical events puts severe limits on the possible variation with frequency of the speed of light traveling through outer space is not only perfectly clear, it obviously describes the observations and exactly what they mean. Brews ohare (talk) 23:46, 27 January 2009 (UTC)


 * Brews, it's an open experimental question what happens to light dispersion if you take a series of measurements in imperfect vacuums and then extrapolate them to zero pressure and zero temperature. Neither you, nor NIST, nor anyone, can define what the results of such experiments will be. This is what I mean by "extrapolated better and better vacuum". An "extrapolated better and better vacuum" is what experimentalists actually use to measure the meter, verify the isotropy of the speed of light, and so forth. These people are encouraged (and sometimes employed) by NIST, and when NIST talks about "the speed of light in free space" they're certainly talking about measurements taken in this way, since there's no other way to do it. So if you don't want to call these experiments "measurements of properties of free space", and you don't want to call them measurements of properties of an "extrapolated better and better vacuum", what do you want to call them? Name your term. Surely they're measuring something, right? --Steve (talk) 01:39, 28 January 2009 (UTC)


 * Steve: Here's the point, which your remarks do not address:
 * The sentence Measurements based on the arrival of electromagnetic radiation from distant astrophysical events puts severe limits on the possible variation with frequency of the speed of light traveling through outer space is not only perfectly clear, it obviously describes the observations and exactly what they mean. Brews ohare (talk) 02:10, 28 January 2009 (UTC)

The reason I dislike this, as I said above, is that it makes it sound like this is a measurement of a certain property of the interstellar medium. It's more than that, it's an experimental constraint on the basic laws that govern our universe. It's as if you were describing the Millikan oil-drop experiment as a "demonstration that small droplets of oil suspended in air have a quantized charge". It's accurate but gives readers the wrong idea. :-) --Steve (talk) 04:51, 28 January 2009 (UTC)


 * I would like to put, '...light traveling through free space ', Brews would like to put, '...light traveling through outer space '. We could argue over this for ever but the source itself says, '...light traveling through space', without specifying the kind of space.  My proposal is therefore a compromise solution that is supported by a reliable source, which it is hard to argue against.


 * Personally, I believe that the authors meant 'free space'. In physics, the term 'speed of light' invariably means 'speed of light in free space'. The authors were not trying to measure the properties of the interstellar medium but to put limits on a fundamental property of EM radiation. However, they do not mention 'free space' explicitly so I am happy to compromise and leave this term out out.  On the other hand they do not use the words 'speed of light through outer space' or similar so we should not say that either.  Martin Hogbin (talk) 10:09, 28 January 2009 (UTC)

Brews, all you have written above and in the section below is your opinion, which you are entitled to, as I am entitled to mine, but articles in WP should be based on reliable sources, not opinion. Now I am not fan of quoting sources verbatim, I am happy with a degree of interpretation for the benefit of non-technical readers, but only when there is general agreement about the facts. In this case there is a difference of opinion, which I am happy to discuss, but until there is a consensus to do otherwise we should say what what our source says, in the words that it says it in. Martin Hogbin (talk) 13:57, 28 January 2009 (UTC)

Deeper meaning of the dispersionless nature of outer space
Steve and Martin: You both seem to believe that the measurement of of the properties of the interstellar medium has ramifications that go far beyond the properties of this particular medium. However, all the measurement actually does is establish the properties. Whatever you two want to read into it requires that this fact be placed within a theory where the theory provides a significance beyond the simple results of the observation. So bite the bullet and dig up the sources and write the summary paragraph that shows that the existence of some media with ε and μ as close to ε0 and μ0 as present accuracy can determine is a much more meaningful thing than the simple fact.

As far as Maxwell's equations go, they are totally indifferent to whether outer space approximates ε0 and μ0 or has some wild material properties. EM waves can propagate in any medium, and at any speed up to c0. The measurement showing some media exhibit ε ≈ ε0 and μ ≈ μ0 doesn't do more than establish the existence of media with these properties.

It's like the M - M measurement of the speed of light as a function of the movement of the observer. : you need relativity to interpret the facts before the implications become clear. And it seems all special relativity demands is that an upper limit on speed of light exists, regardless of whether c0 applies to outer space.

Maybe an approach to your discussion could be to show that if outer space exhibited dispersion it would have amazing repercussions according to some theory? Brews ohare (talk) 13:17, 28 January 2009 (UTC)


 * It may be possible to find a source that meets the requirements of your first paragraph but this will not be easy, nor in my opinion, is it necessary. There are many rarely discussed assumptions in the philosophy of physics and one is the essentially pragmatic view that, until proven otherwise, the world behaves 'sensibly'.  In other words if we measure the properties of gases at reducing pressures we can extrapolate these properties to zero pressure to get the properties of free space.  Of course there have been surprises before and there will, no doubt, be some in the future; they are dealt with as they arise.


 * Actual, philosophically pure, free space is of no interest to physicists because it can never be realized or measured. It exits only as an idealization, as the end point of a graph if you like.  Free space is a concept only, it has no actual reality, thus we are able to treat it in any way that suits us, usually as the limit of a series of real measurements. Martin Hogbin (talk) 14:14, 28 January 2009 (UTC)


 * Martin: These ponderings of yours are a poor substitute for a documented argument. Brews ohare (talk) 17:18, 28 January 2009 (UTC)


 * It is the content of the articles that should be based on reliable sources. This is the page for discussion. Martin Hogbin (talk) 17:38, 28 January 2009 (UTC)

Brew, I still cannot work out what your point is. Perhaps it would help if we see what are the things that we agree on. Do we agree the following?

Free space is an idealization that is physically unobtainable.

The speed of light is defined to be a certain value in free space.

Actual measurements of the speed of light are made in media other than free space.

According to Maxwell's equations, all frequencies of EM radiation travel at the same speed in free space.

Martin Hogbin (talk) 18:11, 28 January 2009 (UTC)


 * Hi Martin: I agree on all these points. However, I believe no measurements can be made on free space, only upon media that are known to some accuracy to approximate free space. I do not find your sentence to agree with these points. It implies free space is amenable to measurement (has measurable properties, or is realizable), or that outer space (possibly, within the vagueness you advocate) is free space. Brews ohare (talk) 18:41, 28 January 2009 (UTC)


 * You say, '...no measurements can be made on free space, only upon media that are known to some accuracy to approximate free space' and I agree. But, accepting that we can never be absolutely sure about anything in physics, we can infer what we would expect the properties of free space to be based on real measurements, can't we?  Martin Hogbin (talk) 18:54, 28 January 2009 (UTC)

We don't have to infer or speculate about the properties of free space: they are defined, and therefore certain. We can speculate whether outer space behaves like free space, and attempt to support these ideas by measuring outer space. Brews ohare (talk) 05:49, 29 January 2009 (UTC)

Where are the properties of free space defined? Martin Hogbin (talk) 21:27, 29 January 2009 (UTC)


 * At ε0 and μ0. I suspect you of being deliberately obtuse. Next you will say that NIST uses the term "vacuum" and that their "vacuum" is different from "free space"? So now we have two media with defined values for ε0 and μ0? If so, how are they distinguished one from the other? Brews ohare (talk) 01:08, 1 February 2009 (UTC)

Newton's equation for the speed of sound
Maxwell used Newton's equation for the speed of sound in a long solid rod in order to obtain the speed of light on the basis of an elastic medium with a density related to the magnetic permeability, and a transverse elasticity related to the dielectric constant.

This suggests to me that light is a coherent dispersionless wave in an elastic solid.

Maxwell obtained the concept of displacement current from his own postulated elastic medium, and the concept was used in his derivation of the EM wave equation in 1864, three years after he did the analyis above with Newton's equation.

The modern derivation of displacement current bears no relationship to Maxwell's derivation of displacement current, and the modern derivation doesn't have the correct divergence properties to allow it to connect to the electromagnetic wave equation.

This is all telling us something about the nature of the so-called vacuum. David Tombe (talk) 16:07, 28 January 2009 (UTC)


 * The aether has no place in modern physics and should not be discussed in this article. Please do not add sections on aether to this article.  There are several articles on the aether where your views might be relevant. Martin Hogbin (talk) 16:47, 28 January 2009 (UTC)

Martin, All I did was corrected the facts in the already existing section about the aether. You are claiming that the Michelson-Morley experiment discredited the aether. That is not what the historical record tells us. The Michelson-Morley experiment could have confirmed the Stokes aether drag model. But Lorentz believed that the Stokes aether drag model contradicted stellar aberration. Lorentz himself believed that the aether blew right through the Earth and he devised his transformation equations on the basis that the aether wind contracted the Michelson interferometer. The aether was dropped from physics when Einstein gave a new interpretation to the Lorentz transformations. But the Michelson-Gale-Pearson experiment of 1925 detected an aether wind due to the Earth's diurnal motion.

Is there anything that I have said above that is not true? You have undone my revisions and told me not to add sections on the aether. I did not add any section on the aether. There was already such a section there and it is inaccurate. All I was doing was correcting the historical record.You are obviously intent on upholding the inaccuracy. You are pushing your own point of view here. David Tombe (talk) 06:48, 29 January 2009 (UTC)


 * Your latest changes do indeed seem to do what you claim, including you comments on Lorentz' theory. The position that the aether is no longer of use or interest is not mine but the current scientific consensus. I have no objection to accurate an historical commentary on the aether, including a reference to Lorentz theory.


 * There are no experiments that show a movement of the Earth through the aether. What do you mean by, 'the Earth's diurnal motion'?  Are you referring to its rotation?  Martin Hogbin (talk) 09:37, 29 January 2009 (UTC)

Martin. Yes, diurnal refers to the rotation of the Earth. The Michelson-Gale-Pearson experiment in 1925 got interference fringes in that regard. David Tombe (talk) 04:33, 30 January 2009 (UTC)


 * David, 'diurnal motion' is a rather curious term for what most people call 'rotation'. The MGP experiment detected the rotation of the Earth and the result is indeed consistent with an aether, however the MM experiment had already ruled out the simple rigid aether.  The only aether that is consistent with both experiments is Lorentz' aether.  As I am sure you know, Lorentz' aether theory (LET) is what gave rise to the Lorentz transformations and it is mathematically and experimentally equivalent to SR and thus a perfectly valid theory.


 * Scientists at the time generally preferred SR to LET for mainly philosophical reasons and LET began to fall into disuse. When Einstein expanded his theory to include gravitation, there was no equivalent expansion of LET and the aether concept was dropped, being considered neither useful, necessary nor complete.


 * Today, relativity has be used to explain a wide range of observations that cannot be explained with Lorentz aether theory. So, the simple aether has been discredited by the MMX, and the Lorentz aether fails to explain things that GR does.  Maybe discredited is slightly too strong a term to use for the Lorentz aether but the article should make clear that aether is a term with no scientific currency. Martin Hogbin (talk) 20:30, 30 January 2009 (UTC)


 * Jackson pages 519-522 has a lengthy discussion on modern ether-drift experiments. Not surprisingly, between 1925 and the present, people have gotten a lot better at measuring any potential diurnal-variation in ether-drift, for example using Mossbauer spectroscopy. It's been all null-results, including a 1970 experiment that would have detected a drift as small as 5 cm/second. Do you think this is worth putting into the article? --Steve (talk) 07:37, 30 January 2009 (UTC)

Steve, there have been so many contradictory experiments on this issue, that I don't know who to believe. We must also remember that we need to scrutinize the theory behind the interpretation of an experiment to ensure that we are not dealing with any tautologies. I had a look at Mossbauer spectroscopy in this regard and I could see that it was using the Lorentz transformation equations in the interpretation. One paper that I dragged up on google then left the interpretation open as between Einstein and Lorentz aether theory.

I don't mind whether or not you include Michelson-Gale in the main article. I was merely objecting to the outright assertion that Michelson-Morley in 1887 had discredited the idea of a luminiferous aether altogether. I was pointing out that the historical sequence of events since 1887 never led to grounds for any such absolute assertion.

Do you have any links to the 1970 experiment which you have referred to? I would like to read it. If not, does it involve the use of Einstein's special theory of relativity in the interpretation? David Tombe (talk) 08:24, 30 January 2009 (UTC)


 * Jackson = Jackson's Classical Electrodynamics, and it has all the references. --Steve (talk) 16:50, 30 January 2009 (UTC)

Steve, I don't have the book immediately to hand. Can you please tell me if the experiments in question use Einstein's special theory of relativity in the interpretation.

Also, in the other experiments which you mention in your recent edits to the main article, in which the speed of light has been measured at 'c' relative to sources that are moving at 99% c, is the measured value of 'c' in that case both relative to the source and to the frame of reference in which the source is moving at 99% c? David Tombe (talk) 17:10, 30 January 2009 (UTC)


 * David, please read WP:RS. Jackson is a reliable source. As far as this article is concerned, that's all that matters. For your own personal hobbies and pursuits unrelated to Wikipedia, I guess you want to know exactly how all these experiments were done. I'm not going to help you do that, sorry. I'm here to improve Wikipedia, not help your personal hobbies and pursuits.


 * The experiment measured the velocity of light (in the rest frame of the laboratory) emitted by particles traveling at 99.9% the speed of light (in the rest frame of the laboratory), which agreed with the standard, well-known speed of light. Do you think the description isn't clear enough? --Steve (talk) 17:27, 30 January 2009 (UTC)

Newton didn't write equations. He wrote quasi–geometrical proportions of ratios.Lestrade (talk) 20:26, 30 January 2009 (UTC)Lestrade


 * Lestrade, I don't get your point. How does it undermine Maxwell's use of equation (132) in his 1861 paper 'On Physical Lines of Force', which he used to deduce the speed of light by linking permeability to density and dielectric constant to transverse elasticity. Equation (132) is Newton's equation in the format that was customary by the time we had advanced to the 19th century. David Tombe (talk) 06:33, 31 January 2009 (UTC)

My point is merely that there can be no "Newton's equation" because Newton never published an equation. All of his published mathematics were in the form of proportions. You may consider this to be insignificant. Someone else may consider it to be significant. However, the world is so full of fictitious fabrications that, I guess, there is always room for one more. I had thought that it might be better if we only spoke about what we know to be true.Lestrade (talk) 14:19, 31 January 2009 (UTC)Lestrade


 * Lestrade, this is just a quibble about nothing. Equation (132) in Maxwell's 1861 paper, is to all intents and purposes Newton's equation for the speed of sound. If we are not allowed to call it that, then it doesn't make any difference to the point which I am making. Go to the talk page of Kepler's laws of planetary motion just to see an example of how genuine scientific debate can be totally stifled by people throwing out trip wires over names and terminologies. David Tombe (talk) 15:34, 31 January 2009 (UTC)
 * I'm not tying to stifle debate or oppose your assertion. I am merely stating that no equations can be found in Newton's writings. He expressed his mathematics in the form of proportions. This is a fact and can be declared without having any effect on your statements. You may think that it is a quibble about nothing, but someone else may find it interesting or even important.Lestrade (talk) 17:59, 31 January 2009 (UTC)Lestrade

Lestrade, OK I take your point. I admit that it is an interesting historical piece of information in its own right which I had not been previously aware of. I do intend to study Newton in detail eventually. David Tombe (talk) 05:04, 1 February 2009 (UTC)

Aether experiments
I think that there may be some confusion concerning experiments to detect the aether. The MMX was intended to detect the Earth's motion through the aether due to its orbit round the sun. It gave a null result as did all following experiments designed to detect the same thing; thus the simple rigid aether is ruled out.

The Earth's rotation (spin on its axis), however, can be detected by more sensitive interferometric apparatus, this effect is the basis of the laser gyro, and it is consistent with both SR and a simple aether theory. However the simple aether cannot be considered a valid explanation for these observations as it had already been ruled out as a possibility. Martin Hogbin (talk) 20:51, 30 January 2009 (UTC)


 * I think we're talking about two different things. A point on the equator moves 1000 miles/hr due to the earth's rotation. If there were a rigid aether, a point on the equator would be moving at speed X through the aether during one part of the day, and speed (roughly) (X + 2000 miles/hr) 12 hours later. So you continually measure some precise quantity over the course of 24 hours, to see if there's a sinusoidal variation. That's the "diurnal" test of the aether theory. That's what Jackson is talking about, and also David I think.


 * OTOH, obviously the fact of the earth's rotation can be measured by gyros, pendulums, etc. That's unrelated. Are we on the same page now? --Steve (talk) 21:48, 30 January 2009 (UTC)


 * David Tombe was talking about the Michelson-Gale-Pearson experiment, which detected the Earth's rotation. I do not know about the experiments described in Jackson, but it seems odd to me that they would try to detect aether drift based in the rotation of the Earth rather than its orbital motion, which would be expected to produce a much larger effect. Martin Hogbin (talk) 22:30, 30 January 2009 (UTC)

I'll try to clarify a few points here. I am suspicious about the experiments referred to in Jackson, and as to what they exactly mean. But I haven't fully investigated them yet, so I can't properly comment.

In my view, the Michelson-Gale experiment showed up the aether drift in relation to the Earth's diurnal motion, just as Steve has described above.

Martin has claimed that the 1887 Michelson-Morley experiment has ruled out a rigid aether. How exactly has it ruled out a rigid aether if the aether were to be entrained in the Earth's orbital motion right up to a cut off line where the Earth's gravitational field gives way to the gravitational field of a neighbour? Such entrainment would fully account for Michelson-Morley. Also, the Earth could be rotating within that entrained aether, accounting for Michelson-Gale.

I am not insisting on mentioning Michelson-Gale in the main article. I am not insisting on pushing the aether entrainment model either. I was merely neutralizing the wording to match the historical facts that related to why the aether was abandoned in modern physics. It is a matter of point of view to categorically state that Michelson-Morley actually discredited aether theory. It didn't do that. It began a series of postulates and investigations which ultimately led to the aether being abandoned.

Meanwhile, I would like to know more about those Jackson experiments. I am very suspicious of claims of highly accurate experiments involving measuring the speed of light coming from tiny particles that are travelling at 99.9% of the speed of light.

I would also like to know why they were doing such experiments in relation to the Earth's diurnal motion. Did they have some doubts about a certain matter? And do they use the special theory of relativity in the interpretation, because if they do, then it is a tautology to use a theory which depends on the constancy of the speed of light to prove the constancy of the speed of light.David Tombe (talk) 06:15, 31 January 2009 (UTC)
 * Steve, on reading your description of the Jackson experiments in the section above, I now realize that they do not address the issue that is raised by Michelson-Gale. Even within a rigid aether theory, I would expect those results. I would always expect light to be measured to have speed 'c' within a stationary pocket of rigid aether, relative to that aether, irrespective of the speed of the source.


 * Michelson-Gale was concerned with the perceived speed of light by a detector that was moving in the aether. That is a different matter altogether. I'm inclined now to remove those Jackson references from the main article because they are irrelevant as regards the issue of whether or not the aether exists. And just like Martin, I am suspicious as to why these experiments were being conducted with specific reference to the Earth's diurnal motion when they weren't even addressing the issue of the motion of the detector towards the source in relation to the Earth's diurnal motion. I don't think that these references should be included at all as they are totally irrelevant. David Tombe (talk) 06:25, 31 January 2009 (UTC)


 * Martin: I would guess that if you're a precision experimental physicist, you'd rather look for rotation effects than orbital effects because for the latter, you need to keep the apparatus up and running for months and months; if there's gradual drift in some component of your apparatus it would take 6 months to confirm that it's drift and not an aether signal; if the power goes out and your laser turns back on at a slightly different frequency, you've just wasted months of work. It's a pain, it's expensive, it's prone to larger systematic errors, etc. The diurnal signal is 10X weaker, but so what? Measure it for ten days, take the Fourier transform, and your signal-to-noise is just as good. Measure for 30 days, and your signal-to-noise is three times better.


 * David: Jackson describes these experiments in a section entitled "ether drift", calling them "ether drift experiments". According to Wikipedia rules, they're relevant and important, whatever your opinion. --Steve (talk) 08:30, 31 January 2009 (UTC)

Suggestion
At present the article mentions aether and the MMX twice, once under 'Luminfiferous aether' and again under 'Michelson-Morley experiment'. I suggest that we combine these two sections into one, giving a clear historical perspective to the aether and the MMX and showing current mainstream scientific though on the subject. I have created a page and copied the current two sections there. We can combine them into one to produce a new section and then put the completed section back here. Martin Hogbin (talk) 10:31, 31 January 2009 (UTC)


 * Martin, the aether has definitely been dropped from modern physics. I don't really mind how you write the history up. I gave a few guidelines. But the main article cannot make categorical statements to the extent that Michelson-Morley discredited the luminiferous aether idea. A situation evolved following that experiment in which Einstein's theories sat comfortably with the Michelson-Morley experiment, and without the involvement of an aether.


 * But Michelson-Morley does not rule out an entrained rigid aether. A rigid aether that is entrained by gravity right up to the cut-off line where the gravitational field of a neighbouring planet takes over, is perfectly commensurate with the Michelson-Morley experiment. That doesn't have to be included in the article. But you need to bear it in mind before making statements to the effect that Michelson-Morley discredited the luminiferous aether idea. David Tombe (talk) 11:08, 31 January 2009 (UTC)


 * That idea was ruled out even before the MMX, in fact as you state it it does not even make sense. A rigid aether is rigid and therefore cannot be entrained.  It is not even a theory, just an idea, that has never formed any serious part of physics. Martin Hogbin (talk) 11:30, 31 January 2009 (UTC)


 * Martin, that depends on what the particles are and how they are bonded together. Who ruled it out before MMX, and why, and what exact model did they rule out? I can conceive of a rigid particulate medium that is held static within the Earth's gravitational field, and so could Stokes in 1845. It's a question of identifying the bonding mechanism between the constituent particles, and how such a bonding mechanism will lubricate the interface regions. It may not be a topic for this article, but just because it is not part of mainstream physics doesn't mean that you can claim that such an idea has been discredited by Michelson-Morley. The aether fizzled out because of a sequence of events. It was never formally discredited, and indeed even today in quantum mechanics they recognize a polarized vacuum which is uncannily similar to the Dirac sea.David Tombe (talk) 14:08, 31 January 2009 (UTC)


 * The aether was never discredited because it was never properly described or defined. The simple rigid aether has been discredited but the word can be use with many meanings today, some having no real relation to the original concept. Today there are no accepted theories of the aether that can compete with relativity in terms of agreement with experiment. Martin Hogbin (talk) 14:20, 31 January 2009 (UTC)

Steve, I removed one of your references because nobody is disputing the point that it makes and it is not related to the issue in question. I left the other reference because it would appear to be relevant. I'd certainly like to check that reference out. David Tombe (talk) 11:08, 31 January 2009 (UTC)

I had done a draft of my proposed section which I think gives a fair view of the history and current status of the aether. Martin Hogbin (talk) 13:53, 31 January 2009 (UTC)

Practically instantaneous
The sentence stating that for many practical purposes the speed of light can be considered infinite has been changed many times, practically going full circle.

The concept that we are trying to convey is that for many real purposes the speed of light is so high that we can assume that it takes no time at all to get from A to B without causing any significant errors or problems. Examples would be, counting seconds to get the range of lightning, or starting signals for Olympic athletes.

The problem is, how to convey this message clearly, succinctly, and in good English. Martin Hogbin (talk) 20:56, 30 January 2009 (UTC)

The Mossbauer effect and aether drift
Steve, I am going to take it that your Jackson reference refers to a 1960 Mossbauer effect experiment by Ruderfer. It is a null result experiment with a highly far fetched interpretation. It has been widely criticized, including by Ronald Hatch who was heavily involved in the Global Positioning system.

Here is a web link listing some of that criticism,

I don't think that your Jackson/Ruderfer reference is a balanced reference. You certainly have no basis whatsoever to claim its superiority over the Michelson-Gale-Pearson experiment. David Tombe (talk) 15:22, 31 January 2009 (UTC)


 * First, I claim its superiority on the basis of reliable sources. A web post is not a reliable source. The definitive electromagnetism textbook is. Second, I think it's clear from context in the article that Lorentz's theory (the ether wind contracts the apparatus) explains the null result of the follow-up experiments in exactly the same way as it explains the null result of the Michaelson-Morley experiment. Do you think that that's not made clear enough? --Steve (talk) 16:37, 31 January 2009 (UTC)

Steve, was the experiment carried out indoors or outdoors? And ultimately, does the experiment disprove the existence of the aether or not? David Tombe (talk) 17:11, 31 January 2009 (UTC)


 * Perhaps, if you were to explain to us what you are using the word 'aether' to mean and what you take its properties to be, we would be able to answer that question. Martin Hogbin (talk) 17:27, 31 January 2009 (UTC)

Martin, In 1845,Stokes advocated an aether theory in which the aether was entrained with the Earth. It drew criticism on the grounds that the material of the aether would have to be such that it would behave like a solid at low pressure, such as to allow the transmission of transverse electromagnetic waves, yet it would have to behave like a liquid when under high pressure at the shear lines where it tapered off into distant space. Stokes didn't know the solution to this problem, but he was convinced that such a solution might eventually be found.

That was Stokes' opinion. Lorentz opposed the Stokes model because he believed that it contradicted the observed measurements of stellar aberration. That was Lorentz's opinion, based on no greater a knowledge than Stokes had about the nature of the material of the luminiferous aether.

Lorentz devised his own aether wind and aether contraction model to account for the Michelson-Morley experiment. The Lorentz theory was modified by Einstein and the aether was dropped from the physics books.

It is important to record this sequence of events correctly. There is no basis to categorically state that Michelson-Morley discredited the aether.

The aether has been creeping in to modern physics again through the back door. First there was the electron-positron Dirac sea, and more recently the polarized vacuum in quantum mechanics. David Tombe (talk) 04:59, 1 February 2009 (UTC)

You still have not defined what you mean by aether, thus it is impossible to agree or disagree with you. If you mean by aether 'everything that is not yet understood in physics' then I guess it is creeping back. Martin Hogbin (talk) 10:56, 1 February 2009 (UTC)

In what medium are measurements in outer space made?
I have provided a reference that places limits upon the dispersion of light using astronomical observations. This reference refers to these limits as a limit upon propagation in "space" or "vacuum". Martin insists that by use of these terms, the reference means to refer to "free space" where ε = ε0 and μ = μ0 by definition. That is, Martin suggests that in fact definitions are constrained by experimental measurements.
 * Everything in physics is constrained by experiment, that is how physics works. Martin Hogbin (talk) 10:51, 1 February 2009 (UTC)

It is obvious that these astronomical measurements refer to transmission through outer space (that is the medium pervading the nearly empty regions of the universe). It is furthermore obvious that although outer space may be well approximated by ε = ε0 and μ = μ0, the accuracy of such approximation can be established only by measurement of the medium of outer space itself (which is, of course, exactly the purpose of the cited paper). It is further obvious that there is absolutely zero dispersion if the medium has an ε = ε0 and μ = μ0. Thus, the reference has succeeded in placing limits upon how far outer space strays from such a medium with ε = ε0 and μ = μ0. I simply cannot understand Martin's objections to my statement of this fact, which he now no longer bases upon reason, but upon exactly parroting the words of the reference, and moreover, parroting these words in a misleading context, in a fashion that makes the ridiculous and unsourced implication that measurement of ε and μ in the medium of outer space somehow limits the definitions of the defined values for  ε0 and  μ0 Brews ohare (talk) 01:45, 1 February 2009 (UTC)


 * I continue to think that you're using a definition of "free space" which doesn't make sense and is inconsistent with obviously-true propositions such as "We know for sure that a meter is larger than an atom", as I describe on your talk page. With the correct definition of free space, I think we could make a lot more progress. :-) --Steve (talk) 02:40, 1 February 2009 (UTC)

So, propose a definition and support it with a reference. Personally I think an adequate definition of free space is a medium in which ε = ε0 and μ = μ0 exactly. Brews ohare (talk) 02:43, 1 February 2009 (UTC)


 * Brews, It's an interesting way to define free space. But what do permittivity and permeability actually mean to you in real terms? Maxwell sheds alot of light on the meaning of those two terms through Newton's equation for the speed of sound. Permittivity comes out to be related to the inverse of transverse elasticity, and permeability is a density. On knowing this and having followed through Maxwell's analysis of this problem, how can you sit comfortably with the idea that the vacuum is plain nothing with a couple of associated numerical constants? I thought that by now you might have been starting to see a pattern since the time when you were editing on Faraday's law and the Lorentz force. You once asked me what the magnetic vector A meant. The textbooks won't tell you that, but Dirac said that it must be a velocity. That's a start.David Tombe (talk) 05:15, 1 February 2009 (UTC)

In real terms, permittivity and permeability are EM properties of real measurable media. ε0 and μ0 are defined permittivity and permeability of a hypothetical medium free space. Some real media have ε≈ε0 and μ≈μ0. Of course, measurement error means we can never say of any real medium that its ε=ε0 and μ=μ0, only that equality exists to within experimental error. More than that, the existence of media with ε≈ε0 and μ≈μ0 is not a prerequisite for choosing to define a hypothetical medium with these defined values, although it may add to the convenience of said definition if there are in fact some real media that do exhibit ε≈ε0 and μ≈μ0. Brews ohare (talk) 05:37, 1 February 2009 (UTC)


 * Brews, the point is whether or not you are happy with that view of reality. Does Maxwell not shed a clearer light on the matter for you? David Tombe (talk) 06:00, 1 February 2009 (UTC)

In my view, and not being trained in interpreting historical documents, not a sophisticated view, Maxwell viewed the aether as just another medium, as real as quartz, say, but with its own permittivity and permeability explained by him in what is to me a very complex fashion involving vortices etc. etc. Thus, in my unversed interpretation, one could ask the question today, does the aether have ε≈ε0 and μ≈μ0? Apart from various practical difficulties in arranging to assuredly have an example of aether to measure, the matter would be settled by appeal to experiment. Whatever the outcome of this exercise, it would have absolutely no bearing upon whether free space, a very technical term with a very specific meaning, has ε=ε0 and μ=μ0. In fact, because free space is hypothetical and its properties based upon definition alone, these properties are untouched by any experimental observation. The only question one can reasonably ask about free space is whether it is at all useful, for example, as a reference state. In support of its utility one might advance that outer space or extremely good terrestrial vacuum approaches the behavior of free space, so it serves as some kind of idealization of some real media. Or, one could argue that use of a an idealized free space is more practical than maintaining a "standard medium" with "standard ε, μ" in some lab in Paris to which one would have to refer all measurement of ε, μ for any other sample of a real medium. Brews ohare (talk) 06:21, 1 February 2009 (UTC)
 * Brews your definition is the same as defining all EM radiation to travel at the same speed in free space. You cannot define something an a way that is inconsistent with experiment. If the current internationally accepted definition of the speed of light were to be found to be inconsistent with experiment then it would have to change.  Similarly, if your proposed personal definition of free space were found to be inconsistent withe experiment (and that does not necessarily mean that an experiment would need to be carried out in free space) then that definition would need to change. Martin Hogbin (talk) 11:05, 1 February 2009 (UTC)

Martin: you have not got the point here. I think it boils down to a belief on your part that "free space" is real. Is that your belief? If you do believe that, my next question is: how can the ε and μ of some real medium be established by definition (for free space, ε=ε0 and μ=μ0 by definition as you can find on the NIST website)? It seems to me that ε and μ of some real medium must be established by measurement (or by calculation based on related measured properties such as the density of constituent polarizable components with established EM properties of their own). Such values never could be set by definition but always would be accompanied by a ± error bar due to experimental uncertainty. In contrast, free space has ε=ε0 and μ=μ0 exactly. Can you fill me in on the origin of your ideas?

The purpose and objective in setting up free space is to supply a standard reference state to which measurements in all real media may be referred. It is unnecessary to establish these free space values by experiment. What is necessary is a set of "best practices" to correct measurements so as to refer to the reference state.

Steve seems to suggest that a sequence of measurements on vacuums made more and more "perfect" (that is, by pumping down further and further to eliminate polarizable constituents) can be extrapolated to the value of "free space". That approach probably could be adopted as a best practice for present technology. A different "best practice" would be needed to correct the pulsar observations of the cited article on dispersion, inasmuch as we cannot tinker with outer space to make it a more perfect vacuum. (I'd guess that observations would be screened to eliminate those whose light went too close to distorting celestial objects. In any event, the observations lead to "limits" that is ± error bars, according to the article, not to exact values for the dispersion.) We have yet to develop best practices that anticipate the advance in techniques that will allow observation of the nonlinear and dichroic behavior of quantum vacuum, which will lead to departures of quantum vacuum from "free space" despite following the extrapolation procedure Steve suggests. That discrepancy might lead to changes in the definition of "free space". However, it equally might not, and instead just lead to adding some more corrections to the list of "best practices". Which course of action is taken will be a judgment based upon the entire metrology, and international consensus on what is the simplest and most accurate way to define the reference state. It is a stretch to call this kind of deliberation over best practices a "measurement" of the properties of free space. Brews ohare (talk) 16:31, 1 February 2009 (UTC)


 * I have said many times that free space is an unobtainable ideal. As Steve has said we can approximate ever closer to free space but we can never get there.  We can extrapolate real data from increasing levels of vacuum to get a figure for free space.  Whether values we get by that process actually represent what we would get if we could make measurements on free space we shall never know as we agree that free space is an unattainable ideal.  When the standards authorities talk of free space measurements they mean real measurements, made under conditions as close to free space as we can muster, and then corrected for all the effects that we know of due to the fact that the experiment was not actually done in free space.  Now there could be some effects that we do not know of, and if they are ever discovered, we would have to change the correction process.


 * Suppose some significant value for dispersion had been measured in EM radiation from a distant object. Would that mean that free space was dispersive?  No, not for sure, it could always be due to the intervening medium.  If after calculating the expected effect of the interstellar medium, based on all that we know, this was found not to be the expected cause, then we would at least have to ask ourselves the question about dispersion in free space.


 * However, what is observed is no dispersion. It still could be argued that free space did have some dispersion but somehow the interstellar medium reduced this to zero, but that would be much harder to take.  What we can say, however, is that there is no reason to suspect that free space might be dispersive. Martin Hogbin (talk) 18:12, 1 February 2009 (UTC)


 * Martin: I'll approach your remarks individually below:
 * Suppose some significant value for dispersion had been measured in EM radiation from a distant object. Would that mean that free space was dispersive? No, not for sure, it could always be due to the intervening medium.  If after calculating the expected effect of the interstellar medium, based on all that we know, this was found not to be the expected cause, then we would at least have to ask ourselves the question about dispersion in free space.
 * In fact, measurement of dispersion in light from a distant object bears upon the dispersion of the medium "outer space". As you point out, it could be due to the intervening medium (maybe a dust cloud, for example). You then make the suggestion that if all known contributions to dispersion failed to account for the observations, "we would have to ask ourselves about dispersion in free space". Absolutely not. We would have to ask ourselves about existing ideas about the origin of dispersion, but that is a question about the physical laws of the real universe, not about hypothetical "free space".
 * However, what is observed is no dispersion. It still could be argued that free space did have some dispersion but somehow the interstellar medium reduced this to zero, but that would be much harder to take.  What we can say, however, is that there is no reason to suspect that free space might be dispersive.
 * What is observed is an upper limit on dispersion, with accuracy set by the observational error, not "no dispersion". It cannot be argued that "free space" does have some dispersion, because, by definition it does not, independent of any observation. There is no such thing as "real" free space. We can go much further than saying "there is no reason to suspect" free space might be dispersive. We can state unequivocally that free space can never exhibit any dispersion whatsoever by definition.
 * Martin, your statements above make no sense unless one believes that "free space" is a synonym for some real medium, perhaps outer space. That belief is erroneous. Don't take my statements as a personal prejudice. Try to explain for yourself how it is that the EM properties of free space are defined by BIPM, NIST etc. and are not measured values. Try to address the notion of correction according to best practices. These ideas are not mine, or not mine alone, let's say. Brews ohare (talk) 20:21, 1 February 2009 (UTC)
 * You agreed some way above that free space can never exhibit any dispersion whatsoever is not true by definition. Can you refer me to exactly where this definition is stated. Martin Hogbin (talk) 20:41, 1 February 2009 (UTC)
 * Martin: we've been around this circuit before: I mean that given the present form of Maxwell's equations and the parameters of free space, which are exactly ε=ε0 and μ=μ0, there can be no dispersion. If you wish to digress on the topic of the accuracy of Maxwell's equations and the possible effect of a departure form their present form upon some future version of the definition of free space as might be decided upon by future deliberation of the BIPM and such, I consider that an extended speculation on a different topic. The topic here is what free space means today using the present form of Maxwell's equations. Brews ohare (talk) 21:35, 1 February 2009 (UTC)

A departure in the form of Maxwell's equations due to observation of various media, of which quantum vacuum and outer space are examples (as already explained earlier in this discussion) might lead to a revision of the definition of free space, but that revision of the standard of "free space" is not based upon "agreement" of ε0 and μ0 with experimental data, but is based upon the practicality and accuracy of a revised standard versus the present standard. (In this context, "accuracy" refers not to how closely ε0 and μ0 agree with experimental data, but rather refers to how revised measurement procedures compare in terms of the ultimate ± errors of the entire procedure that corrects observations to refer to "free space".)   Those considerations of metrology are widely based (involving possibly the standards for time, length, and other issues of accuracy, convenience and what not etc.) and may simply lead to a new set of "best practices" without any change in the definition of free space. How the cookie will crumble cannot be assessed until the nature of the departures is understood. I believe this kind of discussion about how metrology selects a standard is somewhat outside the present scope, which I would take to be a discussion of how free space as used today fits in with the presently accepted form of Maxwell's equations. Brews ohare (talk) 21:44, 1 February 2009 (UTC)


 * We have been round this circuit before but I was rather hoping it would not be a circuit. Can we agree once and for all that free space can never exhibit any dispersion whatsoever is not true by definition but that it is true according to Maxwell's equations?


 * So the question now becomes, are we sure that Maxwell's equations are exactly correct? Are we sure that they could not be ever so slightly wrong? You might ask why would we suspect that anything could possible be wrong with a set of equations that has served us so well for over a century?  Can one of the most verified and respected theories of physics be incorrect?  Well, it has happened before!  Newton's laws were once held up as the exact.  They had been used and verified for centuries and no one could possibly doubt them, yet at the start of the twentieth century it became apparent that they were slightly in error in cases involving high speeds, and relativity was born.  For many purposes the errors are completely unimportant, the moon landings were calculated using Newtonian physics, but on the other hand, without the corrections of relativity the GPS system would not work at all.  We live and learn and we may yet learn that some corrections are need to Maxwell's equations, but the experiment on distant EM radiation tells us that they are still good for now. Martin Hogbin (talk) 22:18, 1 February 2009 (UTC)


 * Martin: Yes, yes. There is no dispersion of free space given the present form of Maxwell's equations.
 * The question is not whether Maxwell;'s equations are correct, as I have explained in great detail above. Brews ohare (talk) 22:23, 1 February 2009 (UTC)
 * Yes it is because, free space can never exhibit any dispersion whatsoever is true according to Maxwell's equations. Thus if Maxwell's equations are wrong then free space can exhibit dispersion. Martin Hogbin (talk) 23:09, 1 February 2009 (UTC)

The sentence in the Wiki article that I dispute is
 * "Measurements based on the arrival of electromagnetic radiation from distant astrophysical events put severe limits on the possible variation in the speed of light with frequency."

There are two problems with this sentence. Next, let us ask whether a modified form of Maxwell's equations could impact the above points were it to evolve that free space exhibited dispersion when used with these updated Maxwell equations.
 * 1) It does not specify the medium in which the limits apply, which is outer space. Therefore, it should be amended to specify the medium. I hope you do not suggest that a different medium was observed. In particular, I hope you do not suggest that "free space" was the medium observed, inasmuch as one cannot observe "free space", as it is only a hypothetical medium with defined (not measured) properties.
 * 2) This sentence is positioned directly following a comment about free space, which seemingly suggests that this sentence also has some bearing upon free space, which it does not.

I would suggest not. The observations suggest that outer space does not exhibit dispersion (to within the accuracy of the observations). Were the updated Maxwell's equations to be adopted, we would then have to explain why outer space does not exhibit dispersion.

We would not have to explain why free space does exhibit dispersion, because that would be a logical deduction using ε=ε0 and μ=μ0 and the updated Maxwell equations, and not an observation.

We then might open an inquiry into whether a new definition of free space were desirable, for example, a new definition that would lead to zero dispersion in free space despite the propensity of the new Maxwell equations to predict dispersion. The considerations entering this deliberation could be the subject of discussion in this article, but do not bear upon the numbered objections above. Brews ohare (talk) 00:13, 2 February 2009 (UTC)


 * Firstly, just let me say that these are the words that the quoted source uses. To change the wording to say 'free space' or 'outer space' would require a good reason supported by a reliable source. Martin Hogbin (talk) 00:29, 2 February 2009 (UTC)

Martin: You are simply ducking the issue. What medium other than "outer space" could possibly be meant???? Please make a suggestion for an alternative medium to pervade the Universe. Brews ohare (talk) 00:30, 2 February 2009 (UTC)
 * I am not ducking anything, you are insisting on adding your own interpretation to what the source actually says. WE both have different interpretations of what it means so it is best to leave exactly as it is. Martin Hogbin (talk) 00:37, 2 February 2009 (UTC)

Go ahead and state your interpretation of observations in outer space that are independent of the material properties of the interstellar medium. Brews ohare (talk) 00:39, 2 February 2009 (UTC)


 * I am giving up this argument, neither I nor Steve seem able to convince you, perhaps someone else will. Until there is a consensus to do otherwise we must use the words of the quoted source. Martin Hogbin (talk) 00:43, 2 February 2009 (UTC)

Martin: You have made no attempt to convince: you do not respond to the arguments presented, but wander about in your own world. What medium other than "outer space" could possibly be meant????Brews ohare (talk) 00:49, 2 February 2009 (UTC)

BTW, the source actually never mentions "space" or "vacuum" or anything at all except the "speed of light" without any reference as to the medium referred to, rather a large gaff, I'd say. Brews ohare (talk) 01:03, 2 February 2009 (UTC)


 * My view is that people understand well what's in the interstellar medium: So many photons, so many hydrogen atoms, etc. People understand very well how these photons, hydrogen atoms, etc. affect the propegation of light. So it's a very easy matter to infer from your astronomical data what the dispersion would be if there were no hydrogen atoms, photons, etc. in the interstellar medium. Therefore, the pulsar observations shed light on the dispersion in (the hypothetical medium that you get by extrapolating real-world observations to zero temperature and pressure, whatever you want to call it.) If you want a reliable source, Jackson page 523-4 describes the pulsar observations in a section called "Frequency Dependence of the Speed of Light in Vacuum". :-) --Steve (talk) 05:21, 2 February 2009 (UTC)


 * Steve: I understand these observations as support for your proposed approach for referring a measurement to "free space" by extrapolation of measurements made on a sequence of real media thought to be closer and closer to free space. That all seems fine to me. But what is your opinion of the disputed statement and my two objections to it? Brews ohare (talk) 07:00, 2 February 2009 (UTC)


 * I don't want to make any statements about specific phrasings until we can agree on what "free space" is. No rush! My belief is that your definition of free space is ultimately circular and implies the unmeasureability of the meter. Under the definition of free space that I believe, dispersion in free space is a conceivable possibility. I put a post on Talk:Free space. :-) --Steve (talk) 04:13, 4 February 2009 (UTC)

Scope of the article
As the error bars of EM measurements are reduced, for example, we might be able to measure a polarization dependence of the speed of light in quantum vacuum. Free space as presently defined with Maxwell's equations as presently understood does not exhibit such a dependence. So what do we do? Do we revise the definition of free space so it also exhibits dichroism using the accepted Maxwell equations, or not? Will we want to modify Maxwell's equations so free space as presently defined also exhibits dichroism? Are these questions ones we want to explore in this Wiki article? Brews ohare (talk) 00:30, 2 February 2009 (UTC)


 * Sorry but I just cannot understand your argument, Brews. You seem utterly convinced that 'free space can never exhibit any dispersion whatsoever' but you cannot provide any reason for this belief.  We have agreed that it is not defined to be so. What exactly is it that means that your claim must be true? Martin Hogbin (talk) 00:37, 2 February 2009 (UTC)

Martin: You are indulging in deliberate distortion. I have said nothing of the kind. I have explicitly considered the possibility that dispersion in free space could occur, and demonstrated the irrelevance of this possibility. I have merely suggested that the statement in the article is inadequately phrased and positioned. See my numbered points 1 and 2 above:There are two problems with this sentence.
 * 1) It does not specify the medium in which the limits apply, which is outer space. Therefore, it should be amended to specify the medium. I hope you do not suggest that a different medium was observed. In particular, I hope you do not suggest that "free space" was the medium observed, inasmuch as one cannot observe "free space", as it is only a hypothetical medium with defined (not measured) properties.
 * 2) This sentence is positioned directly following a comment about free space, which seemingly suggests that this sentence also has some bearing upon free space, which it does not. Brews ohare (talk) 00:41, 2 February 2009 (UTC)


 * What deliberate distortion? I have quoted your words exactly, they are, 'free space can never exhibit any dispersion whatsoever'.  If you are going to make statement like that you need to justify it. So far I have seen none except your continual assertion that it is so. Martin Hogbin (talk) 09:34, 2 February 2009 (UTC)

Martin: Below is my earlier presentation of the role of dispersion, which you have ignored. Brews ohare (talk) 15:16, 2 February 2009 (UTC)
 * Let us ask whether a modified form of Maxwell's equations could impact the above points were it to evolve that free space exhibited dispersion when used with these updated Maxwell equations.


 * I would suggest not. The observations suggest that outer space does not exhibit dispersion (to within the accuracy of the observations). Were the updated Maxwell's equations to be adopted, we would then have to explain why outer space does not exhibit dispersion.


 * We would not have to explain why free space does exhibit dispersion, because that would be a logical deduction using ε=ε0 and μ=μ0 and the updated Maxwell equations, and not an observation.


 * We then might open an inquiry into whether a new definition of free space were desirable, for example, a new definition that would lead to zero dispersion in free space despite the propensity of the new Maxwell equations to predict dispersion. The considerations entering this deliberation could be the subject of discussion in this article, but do not bear upon the numbered objections above. Brews ohare (talk) 00:13, 2 February 2009 (UTC)

Speculation about aether, vortices etc

 * Brews, on first reading, I was totally confused by Maxwell's model. He talked about vortices. Vortices in what? And he talked about centrifugal pressure in the equatorial plane of these vortices. I had always been taught that centrifugal force wasn't a real force.


 * But he was the one that introduced displacement current. And from it he derived the EM wave equation. And it was all in connection with his sea of molecular vortices. And it gave physical meaning to modern day concepts such as permittivity and permeability, because even without displacement current and the EM wave equation, he was able to use Newton's equation for the speed of sound in a solid to deduce that transverse waves moved in his sea of molecular vortices at the speed of light.


 * And the modern vacuum based derivation of displacement current as in the textbooks doesn't tie in with the EM wave equation because it involves the Gauss's law E and not the Faraday's law E.


 * After many readings, I finally saw his picture. The vortices are aligned solenoidally with their rotation axis tracing out the magnetic lines of force. When two north poles come face to face, magnetic field lines spread outwards in the space between them. The repulsive force comes from centrifugal pressure in the equatorial plane of the vortices.


 * His sea of molecular vortices is a solid. But it's not exactly a solid like quartz. The molecules retain fixed positions relative to each other. But they are constantly changing their orientations and angular accelerations in line with changing magnetic fields. David Tombe (talk) 07:57, 1 February 2009 (UTC)


 * Please do not confuse physics with speculation. Seas of vortices have been historically speculated on but they do not form any part of physics today. Martin Hogbin (talk) 10:51, 1 February 2009 (UTC)
 * The purpose of this page is to discuss ways of improving the article 'Speed of light', based on currently accepted theories of physics. It is not to speculate on new and unproven theories or proposed new versions of old ones that were abandoned long ago..Martin Hogbin (talk) 11:10, 1 February 2009 (UTC)

Martin, it's got everything to do with the topic in question. Aren't you talking about the physical meaning of permittivity and permeability? Maxwell gave good evidence that these terms refer to elasticity and density in a sea of molecular vortices. How can you discuss a subject like this and block the key points from the discussion by upholding wikipedia's rules and regulations on speculation. Could we discuss the second world war without mentioning the Germans? Maxwell's papers were alot more than speculation. They introduced important concepts and equations which we still use today.

The point that I was making here is that Maxwell's luminiferous medium was never discredited. A series of events led to it being abandoned, and that's how you need to write it up in the article. You cannot make rash and inaccurate statements to the extent that Michelson-Morley discredited the aether.

And now you ask me what is the aether. Well Maxwell's vortex sea was not the aether as such, but it involved the aether. His luminiferous medium was an elastic solid composed of electric particles that existed around the edge of aethereal vortices. It was an incomplete theory. There were no sinks or sources in his vortices.

A polarized vacuum is creeping back into physics again through the back door, on the back of Dirac's electron-positron sea. Make the electrons into aether sinks and the positrons into aether sources and consider an electron and a positron in mutual orbit. That dipole would be pretty close to a Maxwellian vortex, and a sea of such dipoles would be pretty close to the Dirac sea. And what is the aether itself? It is space. But not the rigid static space that you have in mind. It is a dynamic compressible and stretchable space with field momentum A, equivalent to the magnetic vector potential. That's how Maxwell saw the A vector, and Dirac is also on record as having said that A must be a velocity. So space as you know it is a sea of tiny whirlpools with their axes aligned along the magnetic field lines. The fine-grained angular momentum density is B = curl A.

But as far as the main article is concerned, I think that all you can say is what Brews is saying, and that is that the permittivity and the permeability, as in free space, are constants of free space. The textbooks probably wouldn't permit any deeper knowledge on the matter. There is an equation which links these two constants to the speed of light. That equation is a skeleton version of Newton's equation for the speed of sound. Perhaps you could put the two equations side by side and make the inference that permeability is a density and that permittity is the inverse of transverse elasticity. David Tombe (talk) 13:21, 1 February 2009 (UTC)


 * Hi David: What seems to be going on with the aether is an example of abandoning a concept not because it is wrong, but because people have decided that other ideas are more fruitful in suggesting advances in experiment and understanding.


 * My reading of Maxwell suggests that he thought of the aether as the ultimate source of EM fields, and matter simply responded to the aether to a degree determined by constitutive relations. In a way, this notion is a precursor of today dividing matter from fields by making matter leptons and baryons and the fields all bosons. Brews ohare (talk) 16:48, 1 February 2009 (UTC)

The Spinning Mossbauer Effect
Steve, As you say, the spinning Mossbauer effect experiment is published in a peer reviewed journal. But it is also disputed by quite a few physics professionals in other peer reviewed journals. I have therefore left your reference in place but re-worded it slightly in order to give a more accurate balance. The experiment did take place and it did get a null result. But there is dispute over the interpretation of that null result. David Tombe (talk) 05:24, 1 February 2009 (UTC)


 * The Mossbauer experiment of Champeney, Isaak, and Khan showed an aether drift of less than 3.4 m/s. This result is not disputed.


 * If you are looking for evidence that the Earth is not moving through the aether then one of the most sensitive tests would be the experiment of Hils and Hall in 1990, who improved on the experiment by Kennedy and Thorndike of 1932. This experiment found no variation in the speed of light above 5 parts in 10^14. Martin Hogbin (talk) 18:26, 1 February 2009 (UTC)
 * Also in 1972 Cialdea put an upper limit on aether drift of 0.9 m/s using two lasers. This is probably the most sensitive test to date (up to 1996). Martin Hogbin (talk) 20:13, 1 February 2009 (UTC)

Martin, I know of a few professors who are disputing Mossbauer and I could probably get books and publications if I tried. But we don't really need to go down that road. What I want to know is, were the above experiments all carried on outdoors? Aether entrainment by a wall would shield indoor experiments from any aether wind. Also, aether entrainment in the sensory material would also negate any aether wind effects. It's not all as cut and dried as some people like to think. I don't want to be involved any more in this article, as regards the aether wind issue because it would only end up in endless counter citations. Just bear in mind that it's not all cut and dried, and so try and word it so as not to give the impression that the aether is definitely a discredited concept.

I might however reword some of the things that you say about Maxwell, because I have read his 1861 paper many times. David Tombe (talk) 08:11, 2 February 2009 (UTC)


 * I trust you rewording will be in line with our current understanding of the subject. Martin Hogbin (talk) 09:28, 2 February 2009 (UTC)

Martin, I was specifically thinking in terms of the words permittivity and permeability. Maxwell, didn't use those words. So if you are doing a historical section and talking about how Maxwell related the aether to the speed of light in terms of permittivity and permeability, I might be inclined to change the wordings to 'a sea of tiny aethereal vortices', dielectric constant/transverse elasticity, and density. One of the features of Maxwell's 1861 paper was a lengthy analysis in which he linked dielectric constant to transverse elasticity. That of course involved his famous link up of mechanical stress with electric displacement current. Maxwell never claimed that light propagated in the pure aether itself. It is important to make that clear. David Tombe (talk) 10:52, 2 February 2009 (UTC)

Propose new aether section completed
As stated above, in order to consolidate the two original sections and address some of the criticisms recently made, I have done a draft of a proposed new aether section. This, I now maintain, gives a fair and balanced view of the historical and current status of the aether. It could do with a few more references.

Please take a look as, if no one objects, I intend to paste it into the article to replace the current two sections. Martin Hogbin (talk) 11:31, 1 February 2009 (UTC)


 * Martin, I've read it and it looks OK. By all means put it in, but I may make a few minor amendments on technical issues regarding Maxwell's sea of molecular vortices. Based on what I have said above, you might already be able to make those minor amendments. Try to use 'sea of molecular vortices' instead of aether because Maxwell seldom uses the word aether. Off the top of the head, I think the only reference to the word 'aether' in Maxwell's 1861 paper is on the sixth line down, part II, page 345 (page 34 in the pdf link). The web link is available at quite a few of the Maxwell related sites. If you can't find it, let me know and I'll get you a link. David Tombe (talk) 13:27, 1 February 2009 (UTC)


 * David, thanks for looking at the proposed section and for your comments. This article is intended to show how the speed of light is regarded according to current mainstream physics, with a short section on the historically important but long ago abandoned concept of the aether.   Maxwell's 'sea of molecular vortices' was never fully developed by him and has never been a historically important in the way that the concept of the aether is.  It may have a place in other WP articles but it has no place here. Martin Hogbin (talk) 15:00, 1 February 2009 (UTC)


 * Steve, I have removed the original sections from my proposed page as it confuses the references. I have added what I think are the two most sensitive tests of aether drift which are Cialdea (two lasers) and Champeney et al (the Mossbauer test).  The other test referred to by Jackson, which I believe to be Cedarholm et al (two masers), is not as sensitive as those two. What do you think now?  Martin Hogbin (talk) 20:35, 1 February 2009 (UTC)


 * I'm sure it's fine. :-) --Steve (talk) 05:00, 2 February 2009 (UTC)

If the current internationally accepted definition of the speed of light were to be found to be inconsistent with experiment, then it would have to change
The title of this section expresses a misconception. It suggests that a definition is subject to experimental observation. Perhaps one might change a definition were it to prove too much of a Procrustean bed, encumbering discussion and analysis. However, that means only that the definition proves impractical, not that it fails an experimental test.

What makes a definition impractical is hard to define outside of a particular setting.

For example, let us suppose that an operational approach to realization of free space is to make an extrapolation of measurements on a sequence of samples prepared using higher and higher pumping down, so the sequence of samples contain fewer and fewer atoms or molecules. Within some assumptions as to how such an extrapolation should be conducted, the extrapolation extends the measurements to the case of zero atoms or molecules, a case not actually realizable by any technique known at the moment. Then the speed of light resulting from this extrapolation is taken to be the speed of light in free space, a hypothetical medium where ε = ε0, μ = μ0, both defined constant values.

This extrapolation works fine until measurement technique advances to a point where the fluctuations of quantum vacuum become measurable. Theory predicts, for example, that quantum vacuum may exhibit a speed of light that depends upon the polarization of the light. Supposing this prediction to be accurate, application of the extrapolation technique will lead to different speeds of light depending upon the polarization used in the measurements.

On the other hand, application of Maxwell's equations as we know them to free space (a hypothetical medium where ε = ε0, μ = μ0) predicts no polarization dependence.

Does this disagreement mean that the definition of "free space" must be changed? That is a possible choice, but not a necessary one. One could refine the operational approach using extrapolation to free space. Or, instead, one could simply implement a "best practices" correction to the extrapolation that corrects for the polarization effect, so measurements referred to free space agree regardless of the polarizations used.

Whether a change in operational approach to realization of free space, or a correction using the "best practices" approach, or a redefinition of free space is the better route will be decided by various standards organizations based upon a wide ranging view of metrology, and the best solution from the standpoint of accuracy and convenience will be selected. These considerations extend far beyond whether a particular extrapolation technique leads to ε ≈ ε0, μ ≈ μ0.

The conclusion is that the title of this subsection of discussion is mistaken. Brews ohare (talk) 16:07, 2 February 2009 (UTC)


 * Can you please define ε0 and μ0? If free space is defined in terms of ε0 and μ0, then you must be able to give a definition of ε0 and μ0 that doesn't mention the word "free space"...otherwise it's a circular definition. Alternatively, please give a definition of free space that doesn't mention ε0, μ0, or c0. --Steve (talk) 17:21, 2 February 2009 (UTC)


 * My definition of ε0 and μ0 can be found at the NIST or BIPM web sites. Brews ohare (talk) 20:20, 2 February 2009 (UTC)

Brews, I'm not sure what the purpose of this discussion is within the confines of wikipedia's rules. The textbooks say that free space is nothing, and that there is no aether. What else could you expect to find out from this discussion? Those two constants, permittivity and permeability are constants of space related to magnetism and electricity, which are also related to the speed of light waves through the vacuum.

I noticed your discussion taking place last week, and so I decided to draw your attention to Maxwell's use of Newton's equation of the speed for sound, for the purpose of pointing out that the two constants in question might be a density and a Young's modulus. But I was soon ruled out of order under the rules and regulations. Maxwell is old fashioned and out of date and doesn't apply anymore. So under the rules and regulations, what kind of an answer would you be expecting to get? What modern up to date kind of answer would satisfy you regarding the physical nature of those two constants?

I fail to understand what all this could be leading up to. David Tombe (talk) 12:18, 4 February 2009 (UTC)

Measurements based on the arrival of electromagnetic radiation from distant astrophysical events put severe limits on the possible variation in the speed of light with frequency
This statement is restricted to the medium to which it refers. We know that the speed of light is dependent upon the ε(ω) and μ(ω) of the medium through which it passes. Because the statement refers to "distant astrophysical events", which are observable only through the medium of outer space, it would appear that a more correct statement would be:
 * Measurements based on the arrival of electromagnetic radiation from distant astrophysical events put severe limits on the possible variation in the speed of light with frequency in outer space.

This modification of the sentence is almost trite. It assumes more importance in the context of the article, however, because this sentence appears directly following the sentence "According to classical electromagnetism, the speed of electromagnetic radiation in free space is the same for all frequencies."

This juxtaposition, when done without inclusion of the limiting phrase "in outer space", has the seeming implication that the astronomical observation somehow has relevance to the properties of free space, which is clearly untrue, as these are defined, not measured. See the discussion above. Brews ohare (talk) 17:10, 2 February 2009 (UTC)


 * Perhaps you had better tell the authors of the paper to change their wording. Martin Hogbin (talk) 17:58, 2 February 2009 (UTC)


 * Brews, you continually make this assertion, 'free space can never exhibit any dispersion whatsoever', which you then use to prove your point. What is your justification for the above statement.


 * On the other, I said this, which you quoted above, "According to classical electromagnetism, the speed of electromagnetic radiation in free space is the same for all frequencies." Which clearly true. The question is, is classical electromagnetism exactly correct?

Martin Hogbin (talk) 18:02, 2 February 2009 (UTC)


 * Martin: Your comments above are not responsive.  Brews ohare (talk) 20:24, 2 February 2009 (UTC)
 * That is because your arguments always seem to come back to back the statement I quoted above. What is your justification for the above statement.Martin Hogbin (talk) 20:43, 2 February 2009 (UTC)

Martin: My statements in this subsection stand alone, and do not hark back to anything previous. If some statement made in this subsection requires amplification, please explain further. Brews ohare (talk) 20:57, 2 February 2009 (UTC)


 * What do you mean by this then, '...properties of free space, ... as these are defined, not measured' ?


 * The article makes the situation clear:
 * 'In SI units the speed of all electromagnetic radiation in free space is related to the electric constant ε0 (also called the permittivity of free space) and magnetic constant μ0 (also called the permeability of free space) by the equation c2=1/ε0 μ0. As speed of light in free space is now fixed by definition and the value of the magnetic constant is defined to be 4&pi;&times;10&minus;7 H/m the value of the electric constant is now also fixed'.


 * So c has a defined value as does μ0 and this fixes the value of ε0 according to the classical electromagnetism relation c2=1/ε0 μ0.  It is quite clear from this that if classical electromagnetism is not exactly correct then the value of ε0 may change. Martin Hogbin (talk) 22:15, 2 February 2009 (UTC)


 * The question of whether classical electromagnetism is correct could be discussed, but not in this subsection. Here what is under discussion is whether the words "in outer space" should be added to the sentence in the article. Inasmuch as "astronomical observations" must be viewed using EM radiation transmitted through outer space, these words seem a minor addition that could not possibly conflict with the author's meaning.


 * To roughly summarize the content of that paper, the author observes two events considered to be coincident at their common source, and resulting in EM disturbances of different wavelengths. Observation determines whether on Earth they appear to have different times of arrival. Any such difference is attributed to a variable delay of the EM signals due to traversal of the intervening medium at a different speed at different EM frequencies. Brews ohare (talk) 23:23, 2 February 2009 (UTC)


 * The purpose of the paper was to check the equations of classical EM not to investigate the intervening medium. Martin Hogbin (talk) 09:18, 3 February 2009 (UTC)


 * Can speed be measured without traversal of a distance? The author begins: The question of whether the speed of light varies with frequency is of fundamental and current interest. A search of the paper shows no mention is made of Maxwell's equations. The abstract states in part: to place severe limits on the fractional variation in the speed of light Δc/c < 6.3 × 10−21 based on the simultaneous arrival of photons of 30 keV and 200 keV.
 * A more important question is: Does your understanding of this article have bearing upon the subject under discussion? Brews ohare (talk) 16:36, 3 February 2009 (UTC)


 * Your first quote from the article makes it quite clear,'The question of whether the speed of light varies with frequency is of fundamental and current interest, note fundamental interest. This is a study of light itself not outer space. Martin Hogbin (talk) 19:36, 3 February 2009 (UTC)

Martin: You can pick any few words you like out of the article and say they are the gist. How about "speed of light" or "dispersion"? What is the bearing of your remarks upon including "in outer space" at the end of the sentence? Brews ohare (talk) 06:43, 4 February 2009 (UTC)

Should we continue this discussion elsewhere?
This topic is about the article but it has gone on for a long time, monopolizing this page. Should we take it to a dedicated page?
 * I have summarized the points to be made to avoid unnecessary digressions that have occurred before. If the discussion is focused on the statements at hand, it can be resolved here. Brews ohare (talk) 20:24, 2 February 2009 (UTC)
 * That is fine with me. Martin Hogbin (talk) 20:40, 2 February 2009 (UTC)

Aether
This article is about the speed of light and how it is understood by physicists today. It has a brief mention of the aether because it was of historical significance. The simple rigid fixed aether was the mainstream theory at the time of the Michelson Morley and the only theory that need be mentioned in this article. Many excellent physicist indulged in musings about the aether and proposed various theories but none of these ever came to anything and they have no place in this article. Martin Hogbin (talk) 19:54, 3 February 2009 (UTC)


 * I Agree. In this context anything beyond briefly mentioning the aether as an historical curiosum would be a case of wp:undue weight. As it is now, we already have more aether in Speed of light than we have phlogiston in the oxygen article. DVdm (talk) 20:43, 3 February 2009 (UTC)

Martin, do you mean by fixed aether, that the aether is fixed in space relative to the universe and that the Earth moves through it, causing an aether wind?

For the record, I don't actually support Lorentz's theory, but my understanding of the facts are that Lorentz's theory followed from Michelson-Morley and that the aether wasn't actually abandoned until Einstein's theories superseded Lorentz's theories. It was on that basis that I removed your reference to fact that Michelson-Morley directly disproved the fixed aether theory.

By all means keep your references to the aether short, but I was only trying to make your references to Maxwell's aether more accurate. Why give a reference to Larmor's aether in relation to a paragraph on Maxwell's aether?

On second thoughts, you're the one that wrote the article on the aether. It wasn't me. I corrected some details based on my knowledge of Maxwell's original work. You undid those corrections and then went on to complain that there was too much in the article about the aether. Well so there still is, but it's worse now because it is inaccurate. Would it not just be better to delete the whole section altogether? David Tombe (talk) 05:57, 4 February 2009 (UTC)

Martin, On third thoughts, there was already a section about the aether in this article when I came here last week. I made some corrections to it. You immediately told me that I should not be putting articles about the aether into this article, even though the article was there already. You undid my edits, but nevertheless let the article remain in place, when by your statement, it would have been more logical if you had simply deleted the entire aether article.

You then proceeded to replace it with a larger article on the aether which contains factual inaccuracies. I corrected those inaccuracies. I put quite a bit of work into it, getting quotes, and links, and dates. You undid the corrrections and went to somebody else's talk page requesting them for some input. And they came and made a statement to the extent there was too much about aether in this article.

It seems to me that it not so much a question of there being too much aether in the article, as it's a question of what point of view is being emphasized. You have just happily supplied a large section on aether yourself, which contains a strong point of view that the aether has been disproven. Tell that to quantum mechanics experts about their polarized vacuum. It seems to me that the issue is that you will not be happy about an aether article if it is not critical enough about the concept. I think that has been what's it's been about all along.

You now have some badly written information about Maxwell's aether, with copious references to Larmor's aether to back it all up. The one aether theory which you mention above in relation to Michelson-Morley was Maxwell's aether theory. I fixed up the details of Maxwell's theory for you. So what do you want? It seems that you want to draw attention to aether in a negative light, make it vague with inappropriate terminologies such as permittivity and permeability which Maxwell didn't use, and give references to Larmor's aether theory, and complain that correct information about Maxwell's theory is not needed because there is only one aether theory that is relevant? It doesn't make sense. David Tombe (talk) 12:30, 4 February 2009 (UTC)


 * Again, aether relates to physics like phlogiston to chemistry, and perhaps even more so like Flat Earth to geology/geography - it isn't even mentioned there. Let's just give this curiosum from the past at most a brief mention in the context of the MMX, and move on to more important stuff. DVdm (talk) 16:58, 4 February 2009 (UTC)

I'd suggest that a brief mention without prejudice be made here and a link provided to Aether theories, which should be the definitive article on this matter. Brews ohare (talk) 18:41, 4 February 2009 (UTC)


 * Well I'm happy enough if you remove the section completely. David Tombe (talk) 20:50, 4 February 2009 (UTC)
 * I would be happy to have a link if Aether theories were the definitive article on this matter but unfortunately it is not, it is full of pseudo scientific nonsense which gives the impression that the aether is considered a useful concept in current physics. Martin Hogbin (talk) 22:36, 4 February 2009 (UTC)
 * Having read the Aether theories article again more carefully I think that a link would be OK and perhaps the best answer. Martin Hogbin (talk) 00:03, 5 February 2009 (UTC)
 * Perhaps David could tell me what he considers to be the factual inaccuracies in the current version. Martin Hogbin (talk) 22:37, 4 February 2009 (UTC)
 * I was not the author of the bit about permittivity and permeability but I do agree that it is not particularly useful and should be removed (but not replaced with something about aether vortices).

Martin, my entire focus in this debate was on density and transverse elasticity. That's what brought me into all this. Brews and some others, including yourself were debating the vacuum and what its physical characteristics might be. You were all focused on permeability and permittivity. I drew attention to the fact that Maxwell had shed more light on this subject than anybody else. He had firmly connected the equation that links permeability and permittivity to the speed of light, with Newton's equation for the speed of sound.

That's all I was saying. Then I noticed that Maxwell's aether had been written up carelessly in terms of permeability and permittivity. Those words don't give it its true significance. They are the very words that you guys were debating. So we came full circle.

I decided to make the bit on Maxwell's aether factually accurate in as little space as possible, and also in relation to Lorentz and how the aether was eventually abandoned. It was not abandoned instantly after Michelson-Morley as you have implied.

But what I detected was a reluctance to acknowledge the essence of what Maxwell had said. That's why I asked Brews yesterday what he expected to be the final outcome of his enquiry into free space. If we are going to restrict the terms and conditions of the discussion to what is written in modern textbooks, then there is nothing to discuss. Space is nothing, and it was a conversation about nothing.

I was finally suggesting that we therefore remove all references to the aether, or else, if we want a short section on the aether in relation to Michelson-Morley, then we can at least describe Maxwell's aether correctly and not back it up with lots of references to Larmor's aether. Ultimatley Maxwell's luminiferous medium was not the aether. He never called it the aether. It was a sea of molecular vortices. If you want to mention it in connection with Michelson-Morley, then why not use the correct name. If you don't like the name, then why bother writing about the subject at all? David Tombe (talk) 08:42, 5 February 2009 (UTC)


 * I do not say or imply that the aether was abandoned instantly after Michelson-Morley. I say it rapidly fell into disuse after the publication of GR.


 * There were many (incomplete) aether theories but this article is not the place to go into any of them but the aether that the MMX disproves, I have described as the original rigid fixed aether. If you know of a more accurate name for this the please change my description. Martin Hogbin (talk) 18:41, 5 February 2009 (UTC)


 * I agree with you that mention of permittivity and permeability and agree the terms should be removed. Martin Hogbin (talk) 18:43, 5 February 2009 (UTC)

Martin, the Michelson-Morley experiment was done specifically in connection with Maxwell's model. Maxwell's model was that very fixed aether model that you keep referring to. But anyhow, what about putting those details back again and shifting the whole section to some other article which you may feel is more appropriate for the content? David Tombe (talk) 18:57, 5 February 2009 (UTC)


 * To be more precise: The MM-experiment should distinguisch between the stationary aether model of Fresnel and the dragged-along aether model of Stokes (see History of special relativity). Michelson at first (1881) believed that Stokes model was confirmed, but in 1887 he already knew that Stokes mode violates the law of aberration. That was the starting point for Lorentz, Poincare, etc. However, it was shown by Einstein, that all those developments makes the ether and the classical concepts of space and time useless. (BTW: Maxwell's himself left only a few statements on the relative motion of aether and matter, see his article Ether from 1878). --D.H (talk) 19:15, 5 February 2009 (UTC)


 * Fine, so it is the stationary model of Fresnel that is disproved by the MMX. Martin Hogbin (talk) 22:07, 5 February 2009 (UTC)

It certainly didn't disprove the Stokes aether entrainment model. It was only Lorentz's claim that the Stokes model disagreed with stellar aberration that caused Lorentz to look at an aether wind contraction phenomenon. But in the absence of any details of the material of the Stokes model, Lorentz had no basis upon which to object to the Stokes model on the grounds of stellar aberration. David Tombe (talk) 13:53, 6 February 2009 (UTC)

Dispersion
Here is another try at this paragraph:


 * According to Maxwell's equations, the speed of electromagnetic radiation in any medium with frequency independent permittivity ε and permeability μ (in particular, free space) is the same for all frequencies. Measurements have been made of differences in arrival time on Earth of electromagnetic radiations of various frequencies, radiations that originate simultaneously in distant astrophysical events.[23] The upper bounds placed upon the observed differences in delay set severe limits on any possible variation in the speed of light with frequency, suggesting that in reaching Earth this radiation traversed a medium with a very nearly frequency-independent permittivity ε and permeability μ.

I believe the merits of this paragraph are:
 * 1) It explains what classical EM says
 * 2) It explains the data used
 * 3) It makes exactly the only limited assertion that this work can claim

I do not think there is any debatable point in this paragraph other than its literary merits. Brews ohare (talk) 20:01, 4 February 2009 (UTC)


 * This is confusing, confused, and fails to make the simple point made by the authors of the reference, which is that experiment has put severe limits on the variation of the speed of light with frequency. Please somebody put it back how it was. Martin Hogbin (talk) 22:29, 4 February 2009 (UTC)


 * Is there no one else here who is willing to give an opinion in this quite simple matter? We have a reliable source stating quite clearly that severe limits have been put on the variation in the speed of light with frequency but Brews wants to replace it with this unnecessary complication?  Do we need and RFC? Martin Hogbin (talk) 22:51, 4 February 2009 (UTC)

Martin: Please explain what is wrong with this paragraph. Your statement: "severe limits have been put on the variation in the speed of light with frequency " is elliptic because it does not explain: (i) the implications of classical EM upon such variation, or (ii) how it was done, and most importantly (iii) to what medium the limitation applies. Brews ohare (talk) 23:11, 4 February 2009 (UTC)
 * It is not my statement it is a quote from our reliable source and its meaning is perfectly clear. Martin Hogbin (talk) 23:35, 4 February 2009 (UTC)

Question: Is its "perfectly clear meaning" the same as that of my paragraph (so far as you can understand my paragraph in view of its "confusion")? If not, why not? Brews ohare (talk) 07:20, 5 February 2009 (UTC)


 * No, your paragraph reaches a different conclusion from the one reached by the authors of our source. You reach a conclusion about  permittivity ε and permeability μ, the authors reach one about the speed of light. Martin Hogbin (talk) 18:04, 5 February 2009 (UTC)

You two are arguing about nothing. You are literally arguing about nothing, as in pure empty vacuum. As per the textbooks, the only thing that you can know about this 'nothing' is that c^2 = 1/με. There is nothing more to be said about it. There is nothing that can be added to this discussion that would be permitted by the modern textbooks. The summary is that we have a 'nothing' in which c^2 = 1/με. David Tombe (talk) 19:01, 5 February 2009 (UTC)


 * Martin: as David says: c2 = 1/με. The article is not about speed of light, it is about "Variations of the Speed of Light with Frequency" as the title states. That variation does not occur if με is independent of frequency, according to Maxwell's equations as presently formulated. You know that. You are just "being difficult". Brews ohare (talk) 19:12, 5 February 2009 (UTC)


 * Firstly, shall we keep the discussion to the subject rather than the people. I agree that Maxwell's equations predict no dispersion in free space but we need to do experiments to see if the theory is exactly correct. Martin Hogbin (talk) 22:16, 5 February 2009 (UTC)

Well, Martin I see you have simply restored your version. I am sorry you cannot come to grips with this topic. I do not understand how you arrive at the view that somehow this arrival time data confirms Maxwell's equations in some degree. First of all, if dispersion were observed, as you have admitted earlier, we would not discredit Maxwell's equations, but attribute the dispersion to the medium of outer space, which is not "vacuum". Hence, there is no "test". Second, the observations do not establish zero dispersion, but only set limits upon it. Thus they cannot establish that outer space is free space, but only that it approximates free space to some degree. Your statement of the "content" of the article is nothing but words out of context that can be misconstrued. Brews ohare (talk) 22:37, 5 February 2009 (UTC)


 * Yes I returned the article to a version that uses the exact words of our cited source because that is the correct thing to do in cases of dispute like this. We should only change from that if there is a consensus to do so. I appreciate that you do not agree with what I have done but I really not see your argument.


 * The only way forward that I see is to get some more editors to give their opinions on the subject. I suggest that, rather that have an RfC, we maybe ask some active editors of related articles to give their opinions here.  This is a specialist topic and the  musings of random editors are unlikely to be helpful.  What do you think? Martin Hogbin (talk) 09:37, 6 February 2009 (UTC)


 * Yes Martin. Let's do that. Brews ohare (talk) 12:10, 6 February 2009 (UTC)


 * I have put a request for comment on a few of the articles linked from the lead section of this one. Let us see if we get anyone. Martin Hogbin (talk) 18:30, 6 February 2009 (UTC)

Informal RfC
A different, formal RfC is found Here.

The dispute concerns the wording at the end of the 'Light as electromagnetic radiation section. Martin Hogbin (talk) 18:12, 6 February 2009 (UTC)

The current text is:

'According to classical electromagnetism, the speed of electromagnetic radiation in free space is the same for all frequencies. Measurements based on the arrival of electromagnetic radiation from distant astrophysical events put severe limits on the possible variation in the speed of light with frequency. {{cite journal |journal=Phys. Rev. Lett. |volume=82 |pages=4964-4966 |year=1999 |url=http://arxiv.org/abs/astro-ph/9810479v1 '

One editor would like to keep this text. Another would like to replace it with:

'According to Maxwell's equations, the speed of electromagnetic radiation in any medium with frequency independent permittivity ε and permeability μ (in particular, free space) is the same for all frequencies. Measurements have been made of differences in arrival time on Earth of electromagnetic radiations of various frequencies, radiations that originate simultaneously in distant astrophysical events.{{cite journal |journal=Phys. Rev. Lett. |volume=82 |pages=4964-4966 |year=1999 |url=http://arxiv.org/abs/astro-ph/9810479v1
 * title=Severe limits on variations of the speed of light with frequency |author=Bradley E Shaefer}} The upper bounds placed upon the observed differences in delay set severe limits on any possible variation in the speed of light with frequency, suggesting that in reaching Earth this radiation traversed a medium with a very nearly frequency-independent permittivity ε and permeability μ'.

A third alternative is :

‘According to classical electromagnetism, the speed of electromagnetic radiation in free space is the same for all frequencies. Measurements based on the arrival of electromagnetic radiation from distant astrophysical events put severe limits on the possible variation with frequency of the speed of light transmitted through outer space.’{{cite journal |journal=Phys. Rev. Lett. |volume=82 |pages=4964-4966 |year=1999 |url=http://arxiv.org/abs/astro-ph/9810479v1
 * title=Severe limits on variations of the speed of light with frequency |author=Bradley E Shaefer}}

There has been considerable discussion (see above on this page) but no consensus has been reached. Please comment below:

Comments
My thoughts - observed limits on speed variation with frequency in outer space imply (given a model) limits on speed variation with frequency in free space (or indeed in any other medium). So the first option seems to me to be clear, correct, and close to the source. The second option seems to be importing too much theoretical and philosophical baggage into what should be a straightforward observation. I dislike the (unsignposted) segue from free space to outer space in the third option, and the implicit assumption that experiment can have nothing to say about the properties of free space (or presumably therefore any other idealisation). EdwardLockhart (talk) 19:44, 6 February 2009 (UTC)

Further discussion
I have added this for discussion of the comments to keep things tidy. Martin Hogbin (talk) 20:03, 6 February 2009 (UTC)


 * Edward Lockhart comments:I dislike the (unsignposted) segue from free space to outer space in the third option, and the implicit assumption that experiment can have nothing to say about the properties of free space


 * No segue to outer space is involved here: the reported observations are made upon outer space, not "free space", and the experiment speaks to that medium. Adding three or four words to indicate the medium is a clarification at little expense. If there is an implication of these observations upon outer space that impacts the definitions used to set up "free space", the basis for such impact requires elucidation. Brews ohare (talk) 22:02, 6 February 2009 (UTC)


 * Where can we find these definitions that you talk of? Martin Hogbin (talk) 14:54, 7 February 2009 (UTC)

The speed of light in the vacuum of free space c0 is not measured. It has an exact fixed value when given in standard units. Since 1983 the metre has been defined by international agreement as the length of the path traveled by light in vacuum during a time interval of 1/299,792,458 of a second. This makes the speed of light exactly 299,792.458 km/s. This exact speed of light is found at | c0. The exact permeability is found at | μ0. The electric and magnetic fields in EM waves are related by the value of the characteristic impedance of vacuum, found at | Z0 =√(μ0/ε0). The permittivity is at | ε0=1/(μ0c02). The "standard uncertainties" of all four variables is labeled by NIST as exact (i.e. zero) because they are not measured; zero measurement error means, ipso facto, that free space is not a real, measurable medium.


 * Come off it Brews, I think I should know about the speed of light in free space being defined by now. And I agree that μ0 is fixed by definition but ε0 is fixed at an exact value not by definition but by means of a relationship from classical electromagnetism.  If this relationship should turn out not to be exact then under the current definitions ε0 could vary with frequency.  Martin Hogbin (talk) 00:30, 8 February 2009 (UTC)
 * As pointed out by the use of equations above, c0 is defined, and has no dispersion by definition. The point here is that Shaefer discusses dispersion in c for outer space. There is no point in discussing dispersion in c0; there isn't any by definition. He does not discuss ε. Brews ohare (talk) 00:44, 8 February 2009 (UTC)
 * You say that ' c0 is defined, and has no dispersion by definition' but this is not true, the definition uses the word 'light', it does not say 'all EM radiation'. Furthermore BIPM go on to a recommend specific frequency of light for the realization of the metre, allowing for the possibility that the speed of EM radiation might be found to vary with frequency in the future.


 * Martin: BIPM definition of the meter refers to a specific frequency. However, that is a choice predicated by accuracy, and as you know, is in the microwave region. A variety of optical standards also are in use in many metrology centers (read the articles on second and atomic clock), and BIPM recommends any transitions listed in a long table of frequencies. So your argument doesn't hold water. The definition of the speed of light does not specify any frequency and, in contrast, has zero dispersion. Brews ohare (talk) 16:50, 8 February 2009 (UTC)


 * The definition of the metre uses the word 'light' this implies a restriction to optical frequencies. For realization of the metre visble light is recommended.


 * Among the various recommended frequencies, one is in the infrared. See 1.54μm. However, the point is not what wavelength is recommended for the meter. The point is that | c0 has no frequency dependence. Brews ohare (talk) 16:34, 9 February 2009 (UTC)

Martin: the BIPM definition is: The meter is the length of the path traveled by light in vacuum during a time interval of 1/299 792 458 of a second. One therefore needs a second. The standard for the second is in the microwave region using a Cesium fountain. I suppose one could go into a lot of detail here. You are fixated upon light being optical. What about the Doppler shift: would you suggest that relativity is also part of the discussion? It suggests that a moving source is just red shifted, but still propagates light at the speed of light. What about radio astronomy? Brews ohare (talk) 19:08, 8 February 2009 (UTC)
 * I am not talking about the definition of the second, I am talking about the realization of the metre, for which a HeNe laser is recommended.


 * It should also be pointed out that experiments, like the one we are talking about, have established that speed does not vary with frequency to well beyond our current limits of metrology, so the issue is not important to metrologists today. Martin Hogbin (talk) 10:30, 8 February 2009 (UTC)


 * Martin: The statement "speed does not vary with frequency" is very imprecise. In particular, it does not mention the medium used in these experiments, which of course, cannot be the hypothetical medium of free space, which is unrealizable, and has a speed of light defined to be | c0, with zero dispersion. Brews ohare (talk) 16:50, 8 February 2009 (UTC)


 * You continue to claim that free space must have zero dispersion but you give no evidence for that. 'Dispersion'is not a good word to use anyway since it does tend to suggest some real medium.  What I do not understand is why you claim that the speed of light in free space cannot vary with frequency.  Martin Hogbin (talk) 18:12, 8 February 2009 (UTC)


 * Martin: The exact speed of light in free space is found at | c0. It is frequency independent.
 * The definition does not refer to gamma rays or radio waves, it refers to light, perhaps gamma rays travel more quickly than light in free space. Martin Hogbin (talk) 20:20, 8 February 2009 (UTC)
 * I'd argue that, like relativity theory, "light" in the BIPM context here refers to all EM radiation and would suggest that BIPM is quite prepared to accept Maxwell's equations for the moment, and intends to live within the assumption that they are valid. That means it is not an accident that | c0 is frequency independent. BIPM's treatment of all frequencies as exhibiting similar rates of propagation seems unequivocal. I can find no evidence to the contrary. Brews ohare (talk) 21:22, 8 February 2009 (UTC)
 * The term "dispersion" does not suggest a real medium, it suggests a variation of EM parameters with frequency. The only way that free space can exhibit dispersion is by changing its definition to something different than | c0. We are not discussing possible definitions here: we are discussing the actually adopted definition. The | c0 definition has no dispersion. Brews ohare (talk) 18:50, 8 February 2009 (UTC)

Consequently, Shaefer's paper does not refer to a delay measured in free space, but to a delay in transmission through the medium of outer space, a real measurable medium. If measurement of zero dispersion in outer space has any implication for free space, the logic behind this implication must be explained. The present wording of the article seemingly implies some connection without explanation. Brews ohare (talk) 23:33, 7 February 2009 (UTC)


 * I have explained this connection many times before. Martin Hogbin (talk) 00:30, 8 February 2009 (UTC)
 * Please repeat yourself. Brews ohare (talk) 00:44, 8 February 2009 (UTC)
 * See Edward's reply below. Martin Hogbin (talk) 10:30, 8 February 2009 (UTC)
 * I've replied carefully to Edward below, and assume his argument expresses your notions. That so, my reply should apply to you too. Edward says that any presentation of the implications of Shafer's work upon the definition of free space is too complicated to go into. I disagree: there is no implication. Brews ohare (talk) 17:02, 8 February 2009 (UTC)

Hello again. It seems to me that this is a commonplace situation in the interpretation of experiment. There are some electrical signals in a telescope of some sort. Via a whole bunch of physical theory (much of it embodied in hardware and software), we interpret these as light of multiple frequencies from a single explosive astronomical event. With the employment of some more physical theory, we interpret this to mean that there are limits on the variation of the speed of light with frequency in outer space. And after a bit more theory, we deduce limits on the variation of the speed of light in free space in a Maxwell-like theory.

There are large amounts of theory involved in going from the actual physical observations made to deducing the implied limits on physical theories, of which the fact that outer space is a close approximation to free space is just one (and perhaps one of the simplest bits).

EdwardLockhart (talk) 09:34, 8 February 2009 (UTC)


 * Hi Edward: I understand you to say there is an implied limitation on dispersion in free space stemming from measurements of near zero dispersion in outer space, but the connection is too complicated to go into detail.
 * I suspect that there is no source that makes this argument. It is just part of what you might call our zeitgeist.


 * My view is this: whatever considerations lead to the properties of free space, they have resulted in the definitions | c0, | μ0. These definitions have no dispersion, and that is a consequence of the definition: it requires no theory at all.
 * Do you agree?


 * Free space cannot be realized, because these two properties have defined values. Real measurement has error bars. I can offer you any number of sources that say exactly that.
 * Do you agree? Do you want some sources?


 * The results reported in Shaefer's paper could not possible change the definition of free space, whatever the outcome, because any dispersion found would be attributed to the medium of outer space, which of course is a measurable medium, unlike free space. All that these results suggest is that the dispersion of outer space approximates free space to within certain error bars.
 * Do you agree?


 * That is not to say that there exist no considerations that could change the definitions. But, Shaefer's results simply are not in that category. Therefore, the article should not suggest, as it now does suggest, that the results reported in Shaefer's paper do have bearing upon the definition of free space.
 * That is all I am driving at. Brews ohare (talk) 16:31, 8 February 2009 (UTC)

Shaefer's paper
I disagree with your statement "any dispersion found would be attributed to the medium of outer space". We would have a choice between ascribing significant dispersion to the physical properties of outer space, or to the theoretical properties of free space. Determining which is the best explanation would require further experimentation and theorising. But Shaefer's negative result shows that (at least for now) that we do not need to revise either one. EdwardLockhart (talk) 17:25, 8 February 2009 (UTC)


 * I agree there is a judgment call here; however, I doubt that modification of free space would be (or ever was) entertained as a possible consequence. Here's my judgment: The set-up of free space is an outgrowth of many, many different cases. The theory of constitutive equations is well advanced. To impact the definition of free space one would have to have (i)experimental results far outside the error bars of the theoretical dispersion of outer space based upon its polarizable constituents (ii) no possible alternative physical explanation (e.g. some crazy notion like dark matter or invisible aether) and (iii) a meeting of minds that no practical considerations about accuracy of standards, maintaining standards and using standards were adversely affected by altering the reference state.
 * Bottom line: no result of Shaefer's observations was likely to affect free space. It is incontrovertible that he has established limits on the dispersion of outer space; it is very controversial and difficult to establish that he did anything more than that. To make such a claim in the Wiki article requires at least a paragraph of explanation and some (very hard to find) citations.


 * My proposal is to state clearly what is incontrovertible and avoid making murky implications that cannot be supported, whatever one's gut feelings may be. Brews ohare (talk) 18:20, 8 February 2009 (UTC)

I have added a Cosmology section. It appears Shaefer's paper is of interest in this context. That suggests a possibility for compromise, and I have added a sentence to the disputed paragraph. Brews ohare (talk) 22:16, 8 February 2009 (UTC)


 * Martin deleted my addition, so I added to it, rewrote it, and added citations. I don't understand this desire to remove exactly the sort of material you need to buttress your position. Brews ohare (talk) 04:14, 9 February 2009 (UTC)


 * I have deleted your addition to the electromagnetism again because, even if it were correct, it is in the wrong place. You mention this same paper under your cosmology section which I will discuss below. Martin Hogbin (talk) 20:03, 9 February 2009 (UTC)

Speed of light and cosmology
I have just looked at this section. In my opinion it makes a number of spurious claims that are not supported by the quoted references. Martin Hogbin (talk) 23:19, 8 February 2009 (UTC)


 * Please be specific. Brews ohare (talk) 02:37, 9 February 2009 (UTC)


 * To start with, the article by Schaefer makes no mention of a change in the speed of light over time. Martin Hogbin (talk) 20:01, 9 February 2009 (UTC)


 * You cite a number of papers on quantum gravity which you do not understand. I make no claim to understand them either, in fact I doubt that any of the regular editors of this article have a good understanding of the subject.  In these circumstances, you cannot attempt to summarize or draw conclusions from the papers and add them to this article in a form suitable for the general reader.  You clearly do not have the knowledge to do that, and neither do I.  The whole cosmology section seems to be based on your own particular and inadequate understanding of a number of very advanced physics papers.  I could not do any better myself, that is why I do not even attempt it. Martin Hogbin (talk) 20:10, 9 February 2009 (UTC)

My understanding is at the level of the verbal introductions and conclusions of the various papers, and does not extend to doing the math. That approach can be risky if there is controversy, because each author wants to put their slant on things. However, no statements in the Wiki subsection favor one or another model. The predictions of dispersion etc. seem to be accepted as theoretical predictions worth testing, and some tests have been made. The accuracy that can be claimed by these tests is debated, but most are null results. No attempt has been made in the Wiki article to sort this out in any detail. I don't think there are any huge gaffs here, but I agree that an expert could write something more cogent. It is (by Wiki standards) very well documented, so the more sophisticated reader can pursue matters themselves. Possibly, having an attempt here will encourage expert action: I'm often told it is easier to edit a proposal than to start from scratch. Put up a template requesting assistance if you like. Brews ohare (talk) 03:00, 10 February 2009 (UTC)


 * The problem is that you have drawn conclusions from papers that neither of us understand. It is not acceptable in wikipedia to do that.  I have to say that I think it is better not to have the section at all than to put up a speculative interpretations and then ask for the opinion of an expert.  Sorry to be so negative about your contributions lately Brews but I am trying to kep up the quality of the article and speculation does not help with that, even with loads of references. Martin Hogbin (talk) 22:27, 10 February 2009 (UTC)


 * There are no conclusions "drawn by me" from the papers. Every statement in there is corroborated directly and/or is a paraphrase of statements in the literature from cited sources. The article is a guide to the sources. There is no basis for your comments or position. Brews ohare (talk) 13:38, 11 February 2009 (UTC)

Clarity
An improvement in clarity of this article would result if a clear distinction were made throughout between the speed of light c in various media (such as that of outer space, or quantum vacuum, or the vacuum of quantum gravity) and the defined speed of light in free space c0 = 299 792 458 m s–1. The speed of light c is a matter for theory and experiment. The defined c0 = 299 792 458 m s–1 is useful for metrology, but doesn't play a role in various theories and measurements of c (in particular, that of outer space, or quantum vacuum, or the vacuum of quantum gravity), which mainly are interested in departures of c from c0 = 299 792 458 m s–1, departures not only in numerical value, but also functional differences such as anisotropy and dispersion and variation with the age of the Universe. Brews ohare (talk) 16:43, 9 February 2009 (UTC)


 * The article itself is quite clear, in the section 'Use of the symbol ‛c’ for the speed of light' we say that the symbol 'c' is used exclusively throughout the article to represent the speed of light in free space. Martin Hogbin (talk) 19:54, 9 February 2009 (UTC)


 * Thinking a little more about that subject, we do possibly need to make the distinction between the fundamental constant of spacetime, which is sometimes known as Einstein's constant and the speed of electromagnetic radiation. Martin Hogbin (talk) 20:26, 9 February 2009 (UTC)


 * Maybe that helps. According to Einstein's constant, the constant is:

$$\kappa \, = \, - { 8 \, \pi \, G \over c^2 }~$$
 * According to Kane and Witten, Einstein's constant is the speed of light in vacuum or the charm quark. This sounds more like what you mean. Then one has to decide whether Einstein's constant is c0 = 299 792 458 m s–1 or a physical definition, for instance, "the maximum speed of transfer of information" or "the connection of space to time in space-time". If it evolves that the quantum vacuum is not a medium with the properties of free space, does that alter our notion of space-time? There is work out there about violation of Lorentz invariance. Brews ohare (talk) 16:40, 10 February 2009 (UTC)

Intro
I have modified the Intro as follows:
 * The speed of light in vacuum is an important physical constant usually denoted by the symbol c. An idealization of realizable vacuum is the vacuum of free space in which the velocity of light is denoted by c0. The metre is defined such that the speed of light in free space is exactly  299,792,458 metres per second (m/s). International metrology organizations determine the methodology needed to correct measurements in realizable vacuum to refer to the ideal of free space. The speed of light in the best realizable vacuums (such as terrestrial vacuum at very low pressure, or outer space) at present is experimentally indistinguishable from c0.

In my mind this description better distinguishes between measurement and corrected measurement, and more carefully explains the role of c0. Brews ohare (talk) 18:23, 10 February 2009 (UTC)


 * I have reverted this to some version from before with a slight modification. You make it sound like c is the defined speed, whereas c0 would be the real ideal vacuum speed. It isn't. Both refer to the speed in vacuum. The version without the subscript is more common. The remarks about "realizable" vacuums does not belong here. I suggest you find a place for it in the subsection Speed of light. Try to avoid using the (awkward) word "realizable" 3 times. DVdm (talk) 19:48, 10 February 2009 (UTC)

What I wanted to say is that c is extracted from a measurement by application of CIPM corrections that refer the measurements to free space, while c0 is a number attached to the idealization of free space (and is defined, so no measurement is implied). Do you agree with this objective in principle? Brews ohare (talk) 20:46, 10 February 2009 (UTC)


 * No. In 99.99% of the publications c0 and c denote light speed in vacuum. Few authors use c0, whereas most use c now - for the same thing. You probably are referring to one particular marginal technical publication where they make a distinction. This does not belong in the introduction. I.m.o. it doesn't even belong in the article. DVdm (talk) 20:59, 10 February 2009 (UTC)


 * Brews, there is absolutely no consensus for the changes that you continue to make throughout the article. They seem to represent your own unique view on the subject.  Martin Hogbin (talk) 20:32, 13 February 2009 (UTC)

Problems with the article
I agree that c is the common symbol, and that "vacuum" is the common description of where c can be found. However, it is clear that the speed of light has multiple meanings, all of which should enter the intro:


 * 1) it is the maximum rate of transmission of information or physical interactions. See causal contact.
 * 2) it is the fastest speed any particle can attain
 * 3) it enters the connection between space and time spacetime
 * 4) it enters the fine structure constant setting the frequencies of atomic transitions

These important roles transcend setting the speed of light at any particular value, such as 299,792,458 m/s, and these roles would be unaffected if this were not the case.

Therefore, I would support a different intro to this article, which right now puts way too much emphasis on 299,792,458 m/s. Brews ohare (talk) 21:47, 10 February 2009 (UTC)

In addition, any observation of the speed of light takes place in a realizable vacuum (such as terrestrial vacuum at very low pressure, or outer space) and then is corrected to refer to free space. These real vacuums are not perfect, of course, and were technique up to it, would reveal nonlinearities, dichroism and anisotropy. In contrast, the reference state of free space has none of these behaviors. In addition it has a defined speed of light 299,792,458 m/s, and the use of a defined value (zero error bars) sets it apart from real media that are known only to within an error bar.

None of these fundamental facts are expressed in the present article, which remains very deficient in relating standards, measurements, and ongoing research into the properties of realizable vacuums.

No recognition is given to the fact that free space is a construct, and no vacuum extant in our Universe can be proven to be exactly like free space. Nor can corrections intended to extrapolate measurement in these media to free space values be shown to be exact.

In addition, the above numbered properties of the speed of light are under scrutiny, and may have to be generalized. No indication of the tentative nature of these properties and how they are being re-examined enters the article, with the exception of the Cosmology section I have just added. Brews ohare (talk) 21:47, 10 February 2009 (UTC)


 * Brews, I think the opening paragraph is OK; it says it's an important physical constant and gives its value. The second paragraph incorporates some aspects of your items 1, 2, and 3, and could easily be extended to cover all three in a unified way, I think.  I'm less clear on how the fine-structure constant relates, so have no comment presently on how that might be integrated as well.  If I were you, I'd find a good source that covers this topic to your satisfaction, and write a proposed new paragraph based on it, citing  it, and let us talk about that proposal here.  Dicklyon (talk) 22:28, 10 February 2009 (UTC)


 * OK. Thanks Dick. I am going to drop this attempt and do something useful. Brews ohare (talk) 22:33, 10 February 2009 (UTC)

The point is that theoretical physics is always a long way ahead of metrology and engineering. Metrologists are only interested in very well established physics that has been tried an tested for years. The constancy of the speed of light in free space is just that, it was first proposed over a century ago and to date no evidence against it has been found. For that reason (with others) when it comes to the delineation of the metre, the a definition based on the speed of light (under defined conditions and of preferred wavelength) has been chosen by the standards authorities as the most stable and practical way of doing things.

Theoretical physicists, on the other hand, have speculated that there might be variations in the speed of light in free space but these variations are many orders of magnitude below those that would affect metrology. No experimental evidence of any of these theories has been found to date. If some large and unexpected effect were to be found then metrology might need to reconsider some of its definitions but that has not happened yet so there is no need to change anything.

As far as anyone can tell so far, the different meanings of the speed of light that you give above are all identical. Martin Hogbin (talk) 22:46, 10 February 2009 (UTC)


 * Excuse me, Martin, but what is significant is that the importance of the speed of light in physics is tied to the four numbered points above, which should be stressed at the outset, and not tied to the numerical value you assign. The intro to the article fails to emphasize the physical role of the speed of light and waves the flag of 299,792,458 m/s.


 * According to currently accepted theories of physics, all they are all the same thing and have exactly the same value. There are theories that some of them might be different but there is no experimental evidence for these.  Wikipedia should first state the accepted point of view.  We can include more speculative theories later.


 * It is beside the point to compare theory and metrology. The issue is that free space is an ideal space or model of space (and no model can be measured), and outer space is a real and measurable space, not a model. Logically speaking, the model and reality are different, whether they agree or not. Free space is nothing more than an idealization, and one cannot imbue 299,792,458 m/s with fundamental significance. It is the behavior of outer space that has significance via the numbered points above, and possibly other points that are yet to be raised. Brews ohare (talk) 05:25, 11 February 2009 (UTC)


 * Brews, three people have argued the point about free space with you, you must at least consider the possibility that you might be wrong. Martin Hogbin (talk) 22:01, 11 February 2009 (UTC)

Let's appeal to the literature
Martin: Maybe you are mistaken in your thoughts about free space: check this out:

In free space the principle of linear superposition of potentials and fields holds: for example, the electric potential generated by two charges is the simple addition of the potentials generated by each charge in isolation. Superposition is a defined property of free space even though the electric field near a point charge can become extremely large. See, for example:

Therefore, the ideal vacuum of free space is not the same as any physically obtainable vacuum that exhibits nonlinearity. Quantum vacuum is predicted to do so. See, for example:
 * 1) F.Moulin
 * 2) D. H. Delphenich
 * 3) G. Mourou, T. Tajima, S. Bulanov
 * 4) Christopher C. Davis, Joseph Harris, Robert W. Gammon, Igor I. Smolyaninov, Kyuman Cho.

So, Martin, maybe this work will not come up with departures from free space, but it is possible in principle, eh?. Whether differences are measured or not, the bottom line is that free space is an idealization (with an exact, linear, ε0 μ0) that may or may not apply to any realizable vacuum. That is not your belief, as I understand you.

Similar references can be advanced that predict anisotropy and dispersion of quantum vacuum, which also are not properties of a medium with the exact EM properties ε0 μ0. Brews ohare (talk) 01:41, 12 February 2009 (UTC)


 * Brews, it seems to me that if you've got something well-sourced to add, then a small section on it might be OK. But so far I have trouble understanding exactly what you're point is.  If there's well-sourced speculation that an ideal vacuum might turn out to be nonlinear and affect the speed of light, just mention it and cite it.  Not in the lead. Dicklyon (talk) 07:12, 12 February 2009 (UTC)

Hi Dick: That is not exactly the point. To follow along with your sentence "well-sourced speculation that an ideal vacuum might turn out to be nonlinear and affect the speed of light", there is no need to speculate about ideal vacuum in the sense of free space because it has an exact speed of light set by definition and is exactly linear. The issue is that what often is called "vacuum" is outer space or ultra-high vacuum, which is not necessarily ideal, and may indeed exhibit nonlinearity. There is tendency to confuse the ideal, model vacuum with an actually realizable vacuum, and that tendency is exacerbated by using "vacuum" in a variety of meanings, and referring to "the speed of light" as though it existed in some Platonic universe, instead of stating "speed of light in ideal vacuum", or "speed of light in outer space". Remarkably, in describing Shaefer's observations of light from distant astronomical objects, Martin has repeatedly removed my addition of the limiting words "in outer space" and demands to state only "the speed of light". It seems trite to say that light that travels to Earth through the interstellar medium characterizes that medium, and does not characterize the defined c of free space, but apparently not. Personally I just don't grasp this resistance, which seems to me irrational and has been supported only by saying we cannot depart from a paraphrase of the source (not even represented as a direct quote) to make this distinction. We can't even add "it seems probable that these observations based on light traveling through outer space refer to the speed of light in that medium." Brews ohare (talk) 15:25, 12 February 2009 (UTC)


 * OK, I went back and looked at some diffs, and it's clear why he's reverting you. Not that it's wrong, but that it doesn't belong in the lead paragraph; minor points like that we might someday be able to detect the difference between c and some very-nearly-c velocity in a near vacuum, or via some quantum effects, can be covered in a section where relevant.  In the lead, it's pretty much an irrelevant distraction. Dicklyon (talk) 17:30, 12 February 2009 (UTC)


 * It is worse that that. The main purpose of the investigations of EM radiation from distant objects is not to study the properties of the interstellar medium (althought something interesting could turn up) but to set limits on the variation of speed with frequency of EM radiation in free space.  That is made quite clear in the quoted references. Martin Hogbin (talk) 20:41, 12 February 2009 (UTC)

Dick: This dispute arose originally in the last paragraph of the subsection Light as electromagnetic radiation, not in the Introduction, so this version of Martin's reasoning doesn't apply. I now have added the following well-documented lines in a later subsection to present this viewpoint, rather than continue the battle to add the three simple words "in outer space" to the earlier sentence:


 * Outer space and ultra high vacuum approximate free space, but may have a non-trivial refractive index (that is, an index different from one). Ongoing experimental and theoretical work continues to explore the possibility of small departures of these mediums from free space, which could prove or disprove some theories of quantum gravity, or provide further corroboration of the predictions of quantum electrodynamics. See, for example,



This added paragraph is not a complete antidote to the logical failure to discriminate between definitions and reality, which permeates the speed of light presentation. Brews ohare (talk) 17:32, 12 February 2009 (UTC)


 * you are misinterpreting those articles and I have removed what you have added. Martin Hogbin (talk) 20:41, 12 February 2009 (UTC)

Hi Martin: I find your actions unsupported by any argument or discussion; they are high-handed, arbitrary, and based on your unsupported opinion. Brews ohare (talk) 20:52, 12 February 2009 (UTC)

We agreed to an informal RfC and we got one person to take an interest. They agreed with me. If you want to we can have a full RfC but I am not sure how many it will take to convince you that you are pursuing a nonsensical point. Martin Hogbin (talk) 22:07, 12 February 2009 (UTC)


 * Typically a statement supported by citation is acceptable. You have taken the view that somehow my citations do not support the point. I have added a verbatim quote that in my mind shows you are mistaken. Your simple statement from on high that there is some misconception is insufficient. Brews ohare (talk) 23:02, 12 February 2009 (UTC)

We need some more editors here to stop this nonsense. Is there anybody there? Martin Hogbin (talk) 23:23, 12 February 2009 (UTC)


 * Maybe you could start by pointing out what is wrong with the use of the citations? Brews ohare (talk) 23:27, 12 February 2009 (UTC)

RfC: Is the following proposed insertion accurate?
Comment is invited on the accuracy of the following proposed insertion in the subsection of speed of light discussing transparent media. A subsection to collect your observations is Here.


 * Statements by editors previously involved in dispute
 * you are misinterpreting those articles and I have removed what you have added. Martin Hogbin (talk) 20:41, 12 February 2009 (UTC)


 * Comments:
 * I fail to see any misinterpretation of these sources. Requests for clarification of this claim have been ignored. Brews ohare (talk) 01:12, 13 February 2009 (UTC)


 * Proposed insertion

Here is the suggested paragraph followed at bottom with the supporting references:


 * Outer space and ultra high vacuum approximate a medium with electric permittivity and magnetic permeability that are simple constants, ε0 and μ0, but more accurately may have a non-trivial refractive index (for example, an index different from one). Ongoing experimental and theoretical work continues to explore the possibility of small departures of these mediums from free space (for example, a refractive index exhibiting nonlinearity or dispersion), which could prove or disprove some theories of quantum gravity, or provide further tests for the predictions of quantum electrodynamics. For example, according to Delphenich:
 * “…Now, by the term "electromagnetic vacuum", what we really intend is not a region of space in [which] there is no energy present, whether in the form of mass or photons, but a region of space in which only an electromagnetic field is present. Hence, there is some justification for treating the electromagnetic vacuum as a polarizable medium in the optical sense, which suggests that treating the electric permittivity and the magnetic permeability of the vacuum as simply constants, ε0 and μ0, is basically a pre-quantum approximation, as well as the constancy of the speed of propagation of electromagnetic waves, c0 = 1/√(ε0μ0), or, equivalently, the index of refraction of the vacuum. We shall regard these constants as asymptotic zero-field values of field dependent functions. The fact that c0 itself might vary with the strength of the field suggests that quantum electrodynamics might even have something deep and subtle to say about causality itself that goes beyond the familiar concepts of special relativity.”


 * References:


 * This paragraph could be shortened by omitting the quote from Delphenich: it is included just to give a notion of what the cited references have to say for those who don't want to look it up. Brews ohare (talk) 23:47, 12 February 2009 (UTC)

Comments on RfC
This section is to collect your observations:

The quotation given does not support the proposed insertion. The authors are discussing new theoretical models for an idealised 'free space', in part via experiments on physically realisable vacua. As they explicitly say, "constancy of the speed" is a "pre-quantum approximation" - in their models, free space would behave (slightly) differently from classical free space. The work they are doing is to improve our model of free space, not to better understand the difference between physically realisable vacua and an idealised vacuum, which is uncontroversially considered to be small but non-zero. EdwardLockhart (talk) 08:57, 13 February 2009 (UTC)


 * That sounds right to me. Can you edit the paragraph to better reflect that, and put it back in a form that might be acceptable to all? Dicklyon (talk) 16:36, 13 February 2009 (UTC)


 * I have done that by removing it. I am not sure of the point Brews is trying to make.  Until that is clear there is nothing to say. Martin Hogbin (talk) 20:28, 13 February 2009 (UTC)

Hi Folks: The problem here as I see it is the use of "vacuum" and "free space" in multiple meanings. The cited literature refers variously to "electromagnetic vacuum", quantum vacuum" and just plain "vacuum". All these terms are meant to refer to a vacuum (or vacuums) where ε≠ε0 and μ≠μ0.
 * Response

In the suggested Wiki paragraph the term "free space" is introduced to represent a idealized medium, in which ε=ε0 and  μ=μ0, and therefore definitely not the medium meant by the authors. None of the authors uses the term "free space", but all refer to a baseline medium from which their vacuums depart. The baseline medium is one with ε=ε0 and μ=μ0. For example, the verbatim quote from Delphenich  says he wishes to describe the departure of (in his case) "electromagnetic vacuum" from a medium where  ε=ε0 and  μ=μ0.

That is, I think the only reason to suggest that the sources are misrepresented is because you two bring your own meaning of "free space" into the discussion, and whatever you two have in mind, it is different from the meaning stated in the proposed paragraph. So for Edward, he reinterprets the term "free space" as some more general concept for which ε≠ε0 and μ≠μ0 are possible. That contradicts the stated usage in the proposed paragraph.

I could coin a new phrase, like "baseline vacuum" to represent ε=ε0 and μ=μ0 medium. Would that make you all happy? Then I could discuss how "baseline vacuum" oddly has exactly the same ε=ε0 and μ=μ0 as the "vacuum" used by BIPM and NIST, and oddly has the same properties of linearity and zero dispersion attributed to "free space" in text books. Would this circumlocution serve the purpose of separating unwanted connotations and then putting only the desired connotations back in? Brews ohare (talk) 17:55, 13 February 2009 (UTC)


 * Comment

To me, the important idea for this article is that the very idea of *the* speed of light is only true in the "classical free space" with constant u and h. But a chunk of space with nothing in it (also confusingly called free space) does not have these properties, due to at least QM, GR, and perhaps other explanations. In empty space as it really exists, EM fields do not behave linearly. In particular, perhaps we do not have just c, but maybe c(lamda) or c(lambda, energy) or something even more complex. This article is not the right point to go into the details of what the correct model for a physical vacuum is, but I think it is reasonable to point out that the idea of "the speed of light" is an approximation (though a very good one in normal circumstances), and that physicists agree that if you look closely enough, even empty space needs to be considered as a transparent medium with non-trivial optical properties. LouScheffer (talk) 18:30, 13 February 2009 (UTC)

Revised proposal
Drawing upon all three comments above, how does this look:
 * Outer space and ultra high vacuum can be approximated by a medium ( herein called "classical free space") with electric permittivity and magnetic permeability that are simple constants, ε0 and μ0. Because classical free space employs ε0 and μ0, Maxwell's equations (as presently understood) predict it also exhibits a speed of light c0 (a refractive index of one), as well as zero dispersion, zero anisotropy, perfect linearity and no polarization dependence. But a more accurate model of outer space and ultra high vacuum may have different electromagnetic behavior. Ongoing experimental and theoretical work continues to explore the possibility of small departures of these mediums from classical free space (for example, a refractive index exhibiting nonlinearity or dispersion), which could prove or disprove some theories of quantum gravity, provide insight into the short-distance structure of spacetime, or test further the predictions of quantum electrodynamics.[References follow here] Brews ohare (talk) 19:17, 13 February 2009 (UTC)


 * You have still completely missed the point of what these refernces are saying. Read Edward's comment again. Martin Hogbin (talk) 20:29, 13 February 2009 (UTC)
 * Martin: I am trying very, very hard to accomplish something here. I've read all the comments. I cannot get anything more out of them. Please get to the point and say specifically what is on your mind rather than suggesting Edward speaks for you. At this point I have absolutely no idea what you are objecting to, and am of the opinion that you don't either. Brews ohare (talk) 21:02, 13 February 2009 (UTC)


 * We have already had a very long discussion but to no avail. I do not need anyone to speak for me I am just pointing out that several editors disagree with you. You seem to be trying to make some kind of point in the article itself.


 * If there is disagreement about the content of the article, which there clearly is, the way to deal with it is to discuss the matter here to reach a consensus. So far, at least three editors have disagreed with your stance on this subject but you have continued to try to make your point in the article itself.  That is not the way to do it.  First, explain your point clearly, then try to get people to accept it.  Martin Hogbin (talk) 22:19, 13 February 2009 (UTC)

Martin: The above proposed paragraph is not the same as its predecessors: it has evolved. However, there is no fresh input from you as to what direction would bring it closer to your desires. I see nothing wrong with it; we'll have to see what the response of others is. It has been amended to account for their comments. I have not made any posts about this topic on the article page since this RfC has been posted. Brews ohare (talk) 22:28, 13 February 2009 (UTC)


 * As I understand it, you believe that the speed of all frequencies of EM radiation in free space is absolutely fixed and that experimental evidence to the contrary is impossible. Is that right? Martin Hogbin (talk) 22:39, 13 February 2009 (UTC)


 * My view, which I believe to be the general view of physicists on the matter, is that according to current established theories on the subject, the speed of all EM radiation in free space is the same, but this is subject to experimental verification and could possibly be falsified by some future experiment. Martin Hogbin (talk) 22:44, 13 February 2009 (UTC)


 * Martin: I have introduced as a special term "classical free space" to distinguish it from Edward's "free space" (which has his meaning). By "classical free space" I mean nothing more nor less than a medium with ε = ε0 and μ = μ0. As such it is a model medium, it is what it is, and all one can inquire experimentally is whether such and such real, measurable medium is well-modeled by "classical free space" or not. Experimental evidence can decide if the model does or does not apply, but the model is the model. If you invent a better model, give it a different name.


 * So the direct answer to your question is: I will not comment upon "free space". I am restricting my discussion to "classical free space" as defined, and to the question of what ongoing theory and research have to say about how well "classical free space" models various vacuums, e.g. quantum vacuum, outer space, ultra high vacuum etc. Brews ohare (talk) 22:53, 13 February 2009 (UTC)


 * If by 'classical free space' you mean a space that obeys Maxwell's equations (for free space) exactly then there is no disagreement and there never has been. However, you clearly say, 'I have introduced as a special term...', so we cannot use that term in the article.


 * Apart from that, what is then is the point that you are trying to make? Martin Hogbin (talk) 23:03, 13 February 2009 (UTC)

There is no reason one cannot use a term to avoid continuous repetition of a cumbersome phrase. Maybe you would like to see "a medium with ε = ε0 and μ = μ0" repeated over and over? Brews ohare (talk) 23:07, 13 February 2009 (UTC)
 * The article already says, 'According to classical electromagnetism, the speed of electromagnetic radiation in free space is the same for all frequencies'. What else is there to say? Martin Hogbin (talk) 23:15, 13 February 2009 (UTC)


 * I don't understand. The point is to discuss the existence of literature describing studies on the departure of various measurable vacuums from "a medium with ε = ε0 and μ = μ0, similar in every respect to the medium called "vacuum" by BIPM with the exact properties ε = ε0 and μ = μ0" Brews ohare (talk) 23:22, 13 February 2009 (UTC)


 * But as several others have already pointed out, that is not the purpose of the papers you cite. Martin Hogbin (talk) 23:49, 13 February 2009 (UTC)

Martin, that very simply is nonsense. There is a simply overwhelming amount of literature on this subject. Take the quotes below and show me how you can possibly stretch them to contradict the proposed paragraph of Revision 2 below.
 * Here's a quote from Walter Dittrich and Holger Gies:
 * “vacuum polarization allows the photon to exist as a virtual e+e− -pair on which the various vacuum modifications can act. Under certain assumptions, this influence on the loop process can effectively be described by an immediate influence of a (generally non-linear) medium on the photon itself, e.g., by refractive indices.”


 * Here's a quote from F.Moulin & D.Bernard:
 * “It has been known for some time now that quantum electrodynamics (QED) predicts the existence of a nonlinear interaction between electromagnetic fields in vacuum [1-7]. The effects caused by the vacuum polarization are various. Electric or magnetic anisotropy of vacuum, are the subject of interesting theoretical and experimental research.”


 * Here's a quote from A. M. Ignatov & V.P.Poponin:
 * “Of particular interest are the nonlinear corrections to the linear electrodynamics arising due to vacuum polarization in the strong electromagnetic field. In the ultimate case of slowly varying fields this results in Heisenberg- Euler electrodynamics, which is discussed in many textbooks (e.g. [1]). The main point of this paper is to describe the simplest, in a sense, nonlinear vacuum process: the interaction of two electromagnetic waveforms propagating in opposite directions.”


 * Here's a quote from D Bakalov et al.:
 * “Among the novel phenomena described by QED and as yet never observed experimentally one has to include the so-called vacuum magnetic birefringence [1]. Vacuum, under the influence of a magnetic field, should show a different refractive index for light polarized parallel or orthogonal to the magnetic field direction. The purpose of the PVLAS [2] (Polarizzazione del Vuoto con Laser) experiment is to measure for the first time this small anisotropy.”

Brews ohare (talk) 00:30, 14 February 2009 (UTC)

Revised proposal number 2
In the following, all mention of "free space" in any form is paraphrased to eliminate any direct reference:
 * Outer space and ultra high vacuum can be approximated by a medium with electric permittivity and magnetic permeability that are simple constants, ε0 and μ0. In a medium with ε = ε0 and μ = μ0, Maxwell's equations (as presently understood) predict a speed of light c0 (a refractive index of one), as well as zero dispersion, zero anisotropy, perfect linearity and no polarization dependence. But a more accurate model of outer space and ultra high vacuum may have different electromagnetic behavior. Ongoing experimental and theoretical work continues to explore the possibility of small departures of ε and μ in these media from ε = ε0 and μ = μ0. For example, outer space and ultra high vacuum may exhibit a refractive index exhibiting nonlinearity or dispersion, which could prove or disprove some theories of quantum gravity, provide insight into the short-distance structure of spacetime, or test further the predictions of quantum electrodynamics.[References follow here]

I hope against hope that this version is so smooth that Martin can swallow it. Brews ohare (talk) 00:03, 14 February 2009 (UTC)

Revised proposal number 3
Following suggestions of Dick Lyon:


 * The simplest model of an electromagnetic medium is one with electric permittivity and magnetic permeability that are simple constants, ε0 and μ0. In a medium with ε = ε0 and μ = μ0, Maxwell's equations (as presently understood) predict a speed of light c0 (a refractive index of one), as well as zero dispersion, zero anisotropy, perfect linearity and no polarization dependence.


 * Although some real measurable media behave a lot like this simplest model (for example outer space and ultra high vacuum), ongoing experimental and theoretical work continues to explore for possible differences. For example, outer space and ultra high vacuum might exhibit a refractive index that is somewhat nonlinear or dispersive.


 * Establishing departures of outer space or ultra high vacuum from the simplest model could be profound. (That is, departures that remain after accounting for known imperfections, like cosmic dust or residual atoms.) For example, depending upon what was discovered, results might prove or disprove some theories of quantum gravity, provide insight into the short-distance structure of spacetime, impact our notion of how the Universe is evolving, or test further the predictions of quantum electrodynamics.[References follow here] Brews ohare (talk) 05:25, 14 February 2009 (UTC)


 * I think that still mixes the issues in a confusing way. "Departures from the simplest model" means what?  Departures from the free space model for space containing charged particles is not profound, but straightforward.  If the departures you mean are the speculative quantum departures, say so, and don't mix that up with the case of the imperfect vacuum. Dicklyon (talk) 05:36, 14 February 2009 (UTC)


 * I added a couple of words to say "departures of outer space or ultra high vacuum from the simplest model. (That is, departures that remain after accounting for known imperfections, like cosmic dust or residual atoms.)" - does that help? The relevant observations are not restricted to "speculations", but are made upon actual astronomical events and using lasers in terrestrial vacuum. I guess you can call the theory "speculation" but some is based on stuff like QED with a good track record, not just stuff like quantum gravity. Brews ohare (talk) 06:06, 14 February 2009 (UTC)


 * Maybe it helps, though it's still confusing. Am I right that you're mentioning "outer space" and "ultra high vacuum" because you'd have to account for the classical deviations in these imperfect vacuums, which is experimentally very hard already, before you'd be likely to see the theoretical quantum effects?  Dicklyon (talk) 07:18, 14 February 2009 (UTC)

Yes, that's right unless the anomalies were so large that they exceeded these contributions, which, in that case, would become a mere "background noise". Can you put your finger on what needs change? Brews ohare (talk) 12:24, 14 February 2009 (UTC)

Revised proposal number 4
This version adds a footnote to flesh out the sort of implications that might arise from observations of light:
 * The simplest model of an electromagnetic medium is one with electric permittivity and magnetic permeability that are simple constants, ε0 and μ0. In a medium with ε = ε0 and μ = μ0, Maxwell's equations (as presently understood) predict a speed of light c0 (a refractive index of one), as well as zero dispersion, zero anisotropy, perfect linearity and no polarization dependence.


 * Although some real measurable media behave a lot like this simplest model (for example outer space and ultra high vacuum), ongoing experimental and theoretical work continues to explore for possible differences. For example, outer space and ultra high vacuum might exhibit a refractive index that is somewhat nonlinear or dispersive.


 * Establishing departures of outer space or ultra high vacuum from the simplest model could be profound. (That is, departures that remain after accounting for known imperfections, like cosmic dust or residual atoms.) For example, depending upon what was discovered, results might prove or disprove some theories of quantum gravity, provide insight into the short-distance structure of spacetime, impact our notion of how the Universe is evolving, or test further the predictions of quantum electrodynamics. [Followed by the footnote below]

Proposed footnote Brews ohare (talk) 14:29, 14 February 2009 (UTC)
 * In the case of quantum gravity, as stated by Ellis et al.: “In some models of quantum gravity, space-time is thought to have a foamy structure with non-trivial optical properties. We probe the possibility that photons propagating in vacuum may exhibit a non-trivial refractive index, by analyzing the times of flight of radiation from gamma-ray bursters ….”
 * In the case of small-scale spacetime structure, as examined by Magueijo and Smolin: “The hypothesis that the Lorentz transformations may be modified at Planck scale energies is further explored. … Several examples are discussed in which the speed of light varies with energy and elementary particles have a maximum momenta and/or energy.”
 * In the case of quantum electrodynamics, as stated by Delphenich: “… quantum electrodynamics describes possible nonlinear interactions of photons with each other and with external electric or magnetic fields and the … origin of these quantum corrections to linear electrodynamics seems to be in the polarizability of the electromagnetic vacuum”

Proposed references:
This list is now an assembly of articles of which perhaps three or four would be selected. Brews ohare (talk) 20:05, 13 February 2009 (UTC)
 * 40th Winter School on Theoretical Physics
 * 40th Winter School on Theoretical Physics
 * 40th Winter School on Theoretical Physics
 * 40th Winter School on Theoretical Physics
 * 40th Winter School on Theoretical Physics
 * 40th Winter School on Theoretical Physics
 * 40th Winter School on Theoretical Physics
 * 40th Winter School on Theoretical Physics
 * 40th Winter School on Theoretical Physics
 * 40th Winter School on Theoretical Physics
 * 40th Winter School on Theoretical Physics
 * 40th Winter School on Theoretical Physics
 * 40th Winter School on Theoretical Physics

Yet Another Proposal
Brews, I salute your indefatigability. How about the following?

In the classical model of electromagnetism, the speed of light is the same for all wavelengths in free space. Modern models of small-scale physics predict some variation - the constancy of the speed of light with wavelength is considered to be a "pre-quantum approximation". Experimental results, such as analysis of arrival times of light of different wavelengths from distant astronomical events, show that the divergence from classical behaviour must be very small.

EdwardLockhart (talk) 09:15, 14 February 2009 (UTC)


 * I much prefer that but think that, 'Modern models of small-scale physics predict some variation...' is a bit too strong and somewhat vague. Which modern models?  Are they the best accepted ones?  Should we use terms like 'pre-quantum approximation'? Martin Hogbin (talk) 09:29, 14 February 2009 (UTC)


 * How about saying "some modern models", together with some references (Brews seems to have gathered some suitable ones, plus perhaps a link to Quantum electrodynamics. Incidentally, "pre-quantum approximation" is a direct quote from the source. I rather like the echoes it has of the example of Newtonian mechanics as an approximation to relativistic mechanics. EdwardLockhart (talk) 12:39, 14 February 2009 (UTC)

Hi Edward: I also made "yet another proposal" (number 4 above); can they be merged? I am particularly interested in retaining the first paragraph, which has relevance to the overall article. Brews ohare (talk) 14:33, 14 February 2009 (UTC)


 * Firstly, can I just check where we are proposing to put this. I would be much happier to put it in the new section Brews has created which he has  called 'cosmology'.  If we changed the name to something like 'current theories' it might fit there, although I am still not sure about the rest of it.   'Some modern models' is OK but my main concern is that we are dabbling in something we do not fully understand, sure we have references, but how well does the selection of papers that Brews has found represent current thinking and research on the subject.  Also 'pre-quantum approximation' may well be a direct quote from the source, but what does it mean?  The fact that quantum physics should match classical physics on the appropriate scale is usually called the correspondence principle.  Martin Hogbin (talk) 17:08, 14 February 2009 (UTC)

I propose it to go in the section on "transparent media" as it is about transparent media and that section introduces the notion that the speed of light is affected by refractive index. Whether these few lines represent the cutting edge of research in four or five prolific areas may be not so important, because the gist is simply to point out that questions in these areas are under investigation and to present some representative links that can lead the reader further, if they want to go into the literature. In any event, a deeper discussion would have to go into nuance that is a bit more than this article needs. Brews ohare (talk) 17:36, 14 February 2009 (UTC)

Although it is a laudable goal to make speed of light a superhumanly impeccable article that expresses the universe in a grain of sand, it might be noted that the standards being applied to this short section are more stringent than those applied anywhere else in this article, and way beyond most Wiki articles, which often are lucky to get the definition of the topic accurate, and may be full of unreferenced opinons. (I'm thinking of my experience with matter acceleration frame of reference centrifugal force and others.)Brews ohare (talk) 17:48, 14 February 2009 (UTC)

Merged proposal
An attempt to combine the two proposals; it partially answers Martin's questions about "which models" and "are they accepted". I believe the reference to Delphenich covers the "pre-quantum approximation" issue. Here's the proposal :
 * A simple model of an electromagnetic medium is one with electric permittivity and magnetic permeability that are constants with the values ε0 and μ0. In this classical model, the speed of light c0 is the same for all wavelengths, and there exists no anisotropy, no dispersion, no nonlinearity and no dichroism.

Brews ohare (talk) 16:13, 14 February 2009 (UTC)
 * However, modern models of a region of space in which only an electromagnetic field is present predict some variation - the constancy of the speed of light with wavelength is considered to be a "pre-quantum approximation". Experimental results, such as analysis of arrival times of light of different wavelengths from distant astronomical events, show that the divergences from the above classical model must be very small. Nonetheless, establishing departures from the classical model could be profound. (That is, departures not attributable to known imperfections, like cosmic dust or residual atoms.) Depending upon what was discovered, results might  prove or disprove some theories of quantum gravity, provide insight into the short-distance structure of spacetime, impact our notion of how the universe is evolving, or test further the predictions of quantum electrodynamics.

Sorry Brews but I still cannot see what point you are trying to make. All I can say is that I would object less if it were put in your new section, which I would suggest calling something like 'current theories' or 'new theories'. Martin Hogbin (talk) 18:42, 14 February 2009 (UTC)

To be a little more constructive I have no objection to pointing out in the 'Light as EM radiation' section that classical electromagnetism (or Maxwell's equations if you prefer) does not predict any anisotropy, dispersion, nonlinearity, or dichroism in free space.
 * I propose it to go in the section on transparent media as it is about transparent media and that section introduces the notion that the speed of light is affected by refractive index. The point of this addition fundamentally is just another example of this topic. Brews ohare (talk) 19:42, 14 February 2009 (UTC)


 * As an introduction the the properties of real media that is fine with me. Martin Hogbin (talk) 20:09, 14 February 2009 (UTC)

Beyond that all, we are really saying is that some theories under current consideration may predict some of the above effects in free space but no experimental evidence of any of them has been found to date. I am not convinced that any of the current editors have the necessary knowledge and understanding to give sensible summaries of current research papers, but if you want to give it a go, who am I to stop you. Again I strongly suggest that this goes in a separate section. Martin Hogbin (talk) 19:02, 14 February 2009 (UTC)


 * Hi Martin: Let me edit your last paragraph to say what I'd like it to say:
 * Beyond that, all we are really saying is that some theories under current consideration may predict some of the above effects in outer space or ultra high vacuum (when measurements are appropriately corrected), but no experimental evidence of any of them has been found to date. I am not convinced that any of the current editors have the necessary knowledge and understanding to give sensible summaries of current research papers, but that isn't really necessary because only some links to indicative papers are needed. Maybe more extensive specialized articles will be contributed by somebody later. . Brews ohare (talk) 19:32, 14 February 2009 (UTC)


 * This is where we disagree. Mention of outer space or UHV is a diversion and I still cannot understand why you want to do it.  Perhaps someone else can explain. Martin Hogbin (talk) 20:07, 14 February 2009 (UTC)


 * Hi Martin: I'm glad we are narrowing in on the problem here. Here is my reasoning: the appropriately corrected measurements of outer space or terrestrial vacuum still are experimental quantities, and refer to a real medium, and still have error bars attached to them. Obviously all the point of the experiment in the first place is the hope that the corrected measurements will show up something different from the predictions using the classical model with ε0 and μ0. Consequently, it is confusing to me to look upon the measurement as a measurement of ε0 and μ0, when in my mind it is actually a measurement to see if there is a departure from ε0 and μ0 . I take "free space" to refer to the classical model where ε0 and μ0 are fixed numbers, while I believe you look upon "free space" as a medium to be measured, one where ε0 and μ0 may be discovered to have various values. Hence, reference to corrected measurements is necessary for me, because I do not use "free space" in your sense of the word, and must use a substitute. I also believe that your usage of "free space" might refer to more than one entity, supposing that there are a number of real media out there that will ultimately be found to be different. Brews ohare (talk) 21:08, 14 February 2009 (UTC)


 * Brews, can you point out one good source that approaches this topic the way you do? That might help me understand your approach better.  Dicklyon (talk) 21:20, 14 February 2009 (UTC)

Hi Dick: A possibility is Delphenich, who says in part:
 * Hence, there is some justification for treating the electromagnetic vacuum as a polarizable medium in the optical sense, which suggests that treating the electric permittivity and the magnetic permeability of the vacuum as simply constants, ε0 and μ0, is basically a pre-quantum approximation, as well as the constancy of the speed of propagation of electromagnetic waves, c0 = 1/ ε0 μ0, or, equivalently, the index of refraction of the vacuum. We shall regard these constants as asymptotic zero-field values of field dependent functions.
 * From the standpoint of optics in general, it is then necessary to establish the electromagnetic properties of the vacuum when electromagnetic fields are present; at the very least, its constitutive law.

I'd say all the references in the reference list are aiming to find departures from the customarily assumed values for ε0 and μ0. I'd say none of them use the term "free space", so they won't settle the meaning of that terminology. You will find many textbooks do refer to free space, and they all cite the usual values for ε0 and μ0. However, one is left wondering if they have even thought about a vacuum that wasn't one with ε0 and μ0. Thus, there is an issue here as to what one wants to do with the term "free space" as time marches on: I prefer to keep it back at square one with ε0 and μ0; Martin prefers to let the term evolve with the times. The advantage of the first course is (i) there is continuity with the past if "free space" is taken to mean ε0 and μ0, and (ii) one has a handy term to use in place of an awkward circumlocution, and (iii) one has consistency with NIST & BIPM ε0 and μ0. The drawback is that there are those who follow Martin's path, and also suggest the symbols ε0 and μ0 evolve with the times, in which case the NIST BIPM defined values for ε0 and μ0 must be viewed as anachronisms in the making. Another hang-up of the view that ε0 and μ0 evolve with the times is that the theory suggests that a scalar constant is insufficient: real vacuums are described by εij(E, B, ω) and μij(E, B, ω). Brews ohare (talk) 21:45, 14 February 2009 (UTC)


 * If you can base a section on that source, and leave out mixing it up with imperfect vacuums, that seems like it should be OK. Dicklyon (talk) 23:13, 14 February 2009 (UTC)

Hi Dick: I don't see that there is any "mixing up" going on here; the paragraph is "remarkably lucid and straightforward" as it stands. However, nothing is beyond improvement: see the next version below.Brews ohare (talk) 23:23, 15 February 2009 (UTC)

Merged proposal II

 * A simple model of an electromagnetic medium is one with electric permittivity and magnetic permeability that are constants with the values ε0 and μ0. In this classical model, the speed of light c0 is the same for all wavelengths, and there exists no anisotropy, no dispersion, no nonlinearity and no dichroism.


 * However, modern models of a region of space in which only an electromagnetic field is present predict some variation - the constancy of the speed of light with wavelength is considered to be a "pre-quantum approximation". Experiments testing these new models, such as analysis of arrival times of light of different wavelengths from distant astronomical events, show that the divergences from the above classical model must be very small. Nonetheless, establishing departures from the classical model could be profound. (That is, departures not attributable to known imperfections, like cosmic dust or residual atoms.) Depending upon what was discovered, results might  prove or disprove some theories of quantum gravity, provide insight into the short-distance structure of spacetime, impact our notion of how the universe is evolving, or test further the predictions of quantum electrodynamics.

All vestiges of imperfect vacuum are removed; there is no mention of free space. Are we ready for prime time? Brews ohare (talk) 00:00, 16 February 2009 (UTC)
 * Which section is this for? Martin Hogbin (talk) 00:07, 16 February 2009 (UTC)

How about a sub-subsection of "transparent media"?? Brews ohare (talk) 00:11, 16 February 2009 (UTC)
 * I would be happier for it to be in the 'cosmology' section, perhaps re-named. I would also like to try and re-word some of it at some time but with your general agreement if possible.

I have tried blending it into the transparent media section. If you don't like it, try moving it to Cosmology, but I think some reworking will be needed. Brews ohare (talk) 00:34, 16 February 2009 (UTC)

Brews' point: what is free space anyway?
Brews I am still trying to understand the point that you are tying to make about free space. Do you take the words 'free space' to automatically imply the free space of classical electromagnetism - what you earlier called 'classical free space'? Martin Hogbin (talk) 10:46, 15 February 2009 (UTC)

Brews ohare (talk) 15:05, 15 February 2009 (UTC)
 * I take "free space" to mean a model medium with defined, fixed ε0 and μ0. The advantage in doing this is (i) there is continuity with standard texts where "free space" is taken to mean ε0 and μ0, and (ii) one defines constants the same as the exact values posted by NIST & BIPM for ε0 and μ0, (iii) the properties of linearity, zero dispersion, isotropy, etc. apply with no uncertainty (assuming Maxwell's equations), and (iv) one has a handy term to use in place of an awkward circumlocution. This meaning avoids any implication of a real measurable medium, in particular it avoids implication of necessarily representing, even asymptotically, a realizable vacuum, for which theory suggests that a scalar constant is insufficient: real vacuums (even asymptotically) might be described by εij(E, B, ω) and μij(E, B, ω). Of course, free space might be a good model for some realizable vacuum; experiment can decide that. The papers on this page are trying to make comparisons with ε0 and μ0, although they do not use the term "free space".
 * This is where there is a misunderstanding. In electronic engineering free space is indeed defined the way you quote above, because electronics is based on classical EM.  In free space classical EM predicts exactly what you describe above,  and no one has found any experimental evidence to the contrary.  Obviously, books on classical EM will have the same view.


 * Theoretical physicists on the other hand may mean something different by the term 'free space'. Martin Hogbin (talk) 18:00, 15 February 2009 (UTC)

Well, different groups will use terms differently; I find physicists among themselves are not consistent either (see Jackson, for instance). In any event, I have avoided this term in the proposed paragraph, which therefore is not subject to any reservations about usage. Brews ohare (talk) 23:20, 15 February 2009 (UTC)


 * Yes different groups do use terms differently but please note that Jackson is a book on classical EM. My point is that we cannot impose the classical EM definition of free space on everybody.  According to current, and maybe as yet unknown, models the speed of EM radiation in free space may very with frequency .  That is why experiments were done. Martin Hogbin (talk) 00:02, 16 February 2009 (UTC)

Well Martin, stick with whatever version of free space you like: it does not appear in the proposed paragraph (Merger II) and so it is a non-issue at the moment. Brews ohare (talk) 00:07, 16 February 2009 (UTC)