Talk:Speed of light/Archive 11

Test theories of special relativity
Different conventions about the one-way speed lead to test theories of special relativity some of which differ from special relativity in their physical predictions.


 * Charvest, you seem not to understand the purpose of Test theories of special relativity. I suggest that you read the article before making any suggestions for change to this article. Martin Hogbin (talk) 10:48, 8 September 2009 (UTC)

I am surprised that this topic has not come up before. Contrary to what some might suggest, there is nothing new or controversial in this subject. If anyone really wants to understand this topic then I recommend that they obtain a copy of 'Special Relativity and its Experimental Foundations' by Yuan Zhong Zhang and read it. Otherwise, the situation is summarized in the relevant section of this article.

The two-way speed of light, that is to say from a source to a mirror and back again has been experimentally confirmed to be constant within tight limits (as always, when measured in a vacuum and an inertial frame). Also the one-way sped of light has also been shown experimentally to be independent of the motion of the source. However, the one-way speed, for example, from a source to a distant detector, cannot be measured without some convention as to how the clocks at each end should be synchronized. There was a short time when some reputable scientists may have believed that there were experiments that could measure the one-way speed of light independently of any clock synchronization scheme (for example it might at first sight appear that Romer's original measurement does this) but Zhang carefully re-analyzed these experiments and showed that, as Einstein has originally supposed, the determination of the one-way speed always requires a convention on how to synchronize distant clocks. His book is now widely recognized as the definitive work on the subject.

So, what are the choices when synchronizing clocks. One is to use Einstein's scheme, however this uses the postulated constant one-way speed of light to synchronize the clocks. Clearly, using clocks synchronized in this way to measure the one-way speed of light would be a pointless exercise as it would obviously give the expected value. Alternatively we could use slowly transported clocks (we know that moving them quickly produces strange results). The only problem is that we have know way of knowing that the clocks do not lose their synchronization whilst they are apart. However, experimentally and according to Einstein's theory this produces the same results as Einstein's scheme.

It is possible devise a theory using an additional parameter relating to the way that distant clocks are synchronized. This is essentially what was done by Lorentz in his aether theory, which was published the year before Einstein published his paper. There is an additional parameter that appeared in the equations but it could be set to any value (less than c) without affecting any experimental prediction. Zhang refers to this theory as Edwards theory in his book. One could liken this position the one that pertains to electric potential. There is no way to measure an absolute electric potential, only a potential difference, although we can set an arbitrary zero if we wish, such as the potential of the Earth. Einstein's solution was to drop the parameter completely, knowing that this would make no difference to anything physically measurable or detectable, and that is the solution universally adopted today. Anyone who wishes can add in a parameter that would, in principle, affect the one-way speed of light but as there is no way to measure the value of this parameter it serves no real purpose. In any real experiment the value makes no difference whatsoever and thus, by convention, it is ignored.

Finally, we might ask what the relevance of all this is to things like the realization of the metre. The answer is 'none'. All delineations of this kind are done using interference techniques which essentially rely on the two-way speed of light. If we really wanted to be fussy, we could point out that the two way speed of light is fixed by the definition of the metre and the one-way speed in fixed by this definition and a further convention. However, we already have a section that explains this.

If have no objection to saying more about this subject in the relevant section of the article (please not the lead), provided that absolutely no suggestion is made that there is any confusion, conspiracy, lack of understanding, cover-up, or uncertainty amongst physicists or other scientists. Martin Hogbin (talk) 10:02, 8 September 2009 (UTC)


 * Martin: Thanks for the explanation. When you have time, please consider doing some work on the Test theories of special relativity article, which is clearly in need of expert help. —Finell (Talk) 15:38, 8 September 2009 (UTC)

On this talk page I've split off the discussion of Test theories into its own section as it is a separate topic from the other one. I only used Test theories in my reply to Timothy Rias to emphasize that using a different postulate about light has been done before. It was not to suggest that the purpose of test theories was the same as the purpose of "common relativity".

When I added "Different conventions about the one-way speed lead to test theories of special relativity some of which differ from special relativity in their physical predictions." to the article, I did not mean to imply that these were potential alternative theories, and when I said some I was also thinking of Edward's theory which isn't actually a test theory at all, so I accept that the wording of that sentence is not suitable. In fact the article test theories of special relativity is rather short and should really be greatly expanded before linking to it. Charvest (talk) 20:04, 8 September 2009 (UTC)


 * OK but in that case we have to ask whether those test theories have any relevance to this article. They would be more relevant to any pages discussing the experimental verification of special relativity. Martin Hogbin (talk) 21:37, 8 September 2009 (UTC)

More thoughts on the introduction
A few possibly interesting points for the introduction: Suggestions and comments are expected, of course. LouScheffer (talk) 03:07, 7 September 2009 (UTC)
 * The use of c in connecting space and time into spacetime is not limited to special relativity. See What is General Relativity for an explicit example of how C figures into the metrics of general relativity.  So I think it was better when it just said relativity, not special relativity.
 * Showing E=mc2 does not need relativity at all. Given that the momentum of a photon is E/c, then E=mc2 follows.  The proof is simple - start with a box in free space, emit a photon from one end and absorb it at the other.  The recoil shifts the box; figure the photon weight that keeps the center of mass unchanged.
 * The so-called problem that you cannot measure the speed of light in SI did not begin in 1983; the same problem existed since 1960. Suppose the real, physical speed of light changes, with no other physical constant changing.  Post 1983, this cannot be expressed in SI since the meter changes to match.  But this was true from 1960 on - the changed speed of light will affect the wavelength of the krypton transition, and in exactly the manner that will cancel the change in SI units.  (If the speed doubles, the wavelength doubles, the meter is twice as long, and the numerical value of c is unchanged).  You'd need to go back to the old scratches-on-a-metal bar to express this (rather unlikely) change in SI units.
 * Of the readers of this article, more will be interested in *why* the speed of light is important than will be interested in the history. So the "why" part should come before the history part.


 * I have no disagreement with anything you said. I have some questions on your third point: how important is this in terms of the article? and the lead? In other words how strongly should this be explained? Do you think that using the word exactly in the lead is necessary? Advisable? Or do you think it begs more questions than can sensibly be answered in the lead? Abtract (talk) 07:34, 7 September 2009 (UTC)
 * Sorry, this was just an observation. In my opinion, this should not be in the lead at all, except a brief mention that the meter is defined in terms of the speed of light.  LouScheffer (talk) 12:19, 7 September 2009 (UTC)


 * LouScheffer, Regarding your second point, this was indeed demonstrated by the American scientist Gilbert N. Lewis in 1908. The derivation involves Maxwell's equation for radiation pressure. David Tombe (talk) 08:36, 7 September 2009 (UTC)


 * Lou, I agree with point 1. In fact there is no real need to mention relativity at all.  The whole concept of spacetime is what relativity is all about.  So we might just say that it is a constant of spacetime.


 * Regarding point 2, it was really relativity that put E=mc2 into a sound theoretical context although the formula had been proposed by other physicists before Einstein.


 * Regarding point 3, I have been trying to persuade Brews of that fact for some time but he seems to have lost interest.


 * I agree with point 4, that is how it used to be.


 * I suggest that we wait to see the results of arbitration before we make any major changes now. Martin Hogbin (talk) 16:45, 7 September 2009 (UTC)
 * Regarding point 3, since the particles making up the metal bar are held together by electromagnetic forces, I s'pose that a change in c would resize the bar, too (provided that no dimensionless constant such as &alpha; changes). IIRC, something like that was the standard interpretation of the Lorentz–FitzGerald contraction before Einstein's 1905 paper. -- _ _ _ A. di M. 22:04, 7 September 2009 (UTC)
 * (BTW, I think we could link Planck_units from somewhere in this article, but I'm not sure of where. -- _ _ _ A. di M. 22:12, 7 September 2009 (UTC))
 * I disagree with points 1–3 as well, but I think the most important for the planning of the article is point 4. While we most definitely shouldn't make the article simplistic, we cannot forget that we are dealing with a topic which is introduced in junior high school. The first point that needs to be made is that the speed of light is finite and not infinite: the history section demonstrates how we know that it is finite. I would then go on to say that the speed of light – given a few disclaimers like "perfect vacuum" – is constant, at least over the period of human experience: again, it is important to show that historical measurements converged on the same value. Only then can you really explain why the majority physical theory says that it must be constant. Finally, to be fair, one should mention that there is a minority cosmological theory which doesn't assume that c has always had the same value that it has today. All the same, I think the history section is at least as important as the high brow theoretical physics and philosophy, if not more so for the majority of readers. Physchim62 (talk) 02:11, 8 September 2009 (UTC)

Hi Lou: I am interested in your statement repeated below:
 * (If the speed doubles, the wavelength doubles, the meter is twice as long, and the numerical value of c is unchanged). You'd need to go back to the old scratches-on-a-metal bar to express this (rather unlikely) change in SI units.

If ℓ = c t assuming the time of transit is fixed, then ℓ doubles, which I think is what you are pointing at. Of course, the time of transit actually will half when c doubles, so ℓ = 2c t/2 does not change. However, the situation post-1983 is, I think, a bit different. Then ℓ = c0 t with c0 =  299 792 458 m/s exactly. (That is the BIPM definition of length). The doubling of the speed of light means the transit time halves, so ℓ = c0 (t/2) and the length halves too, even though the real length is exactly the same. Of course, those interested in GR will say that the doubling of the speed of light is necessarily accompanied by other changes that make this analysis simplistic. That is why I prefer the example where the measured speed of light changes due to improvements in measurement technique, rather than postulating changes in the actual speed of light. Any comments? Brews ohare (talk) 23:07, 7 September 2009 (UTC)


 * Is there any proof the speed of light has changed in the last billions of years?Wdl1961 (talk) 00:53, 8 September 2009 (UTC)

Wdl1961: Whether it has changed, or might change, or might be impossible to change, one can speculate about what would happen if it did change, just in order to test one's ideas. Brews ohare (talk) 02:05, 8 September 2009 (UTC)


 * There seems to be little evidence the speed of light has changed. How many times have the human definition of length and speed changed ? For the little time i got left i will bet on a steady speed of light. We can only do the best with what we know reasonably well now. I am all for thinking exercises but they should be kept in perspective. Personally i have no problem to turn back thirty or more years ago  and remember my understanding at that time.  On government inspections i only was required to know the calibration equipment was about ten times  more accurate than the measuring equipment. It is always a question what is relative to what. The most available, accurate and repeatable will win for the time  being, sometimes i go from fingertip to fingertip for length (not for sol.). Wdl1961 (talk) 03:24, 8 September 2009 (UTC)
 * Brews ohare: What would happen if the metre is defined in terms of c and c changed is not only the same thing that would happen is the metre were defined in terms of the length of a piece of metal and the the length of that piece of metal changed (i.e. the metre would change); but also, since particles in your fingertips are held together by electromagnetic forces, probably their size would change, too, so that light would still take the same time to travel across your fingertips. A varying speed of light would have much funnier consequences to worry about than changing the length of metre. My own bet is that stuff would resize so that we would have no way to even realize that it happened. -- _ _ _ A. di M. 10:14, 8 September 2009 (UTC)
 * To expand on that an alternative way of describing the expansion of the universe is not that the universe is expanding, but that the speed of light is decreasing (this simply is the redefinition c = c/a(t) with a(t) the scale factor, and using comoving position and momentum as 'true' position and momentum). This would also describe the universe equally well, the interpretation would not be that the universe is expanding but that everything in the universe is shrinking. In the end it comes down to the samething.(TimothyRias (talk) 10:36, 8 September 2009 (UTC))

I think the point here is not about deep things. It is about the functional dependence upon c of the equation for length ℓ0 in the SI system ℓ0 = c0 t, where the time of transit is   t = ℓ / c, with ℓ the hypothesized actual spatial separation of two points. One could ask whether these equations are embedded in a more complex system of equations that cause the dependence of the time of transit upon c, say t = t (c) to be different. However, maybe the simpler case already is too complicated? Brews ohare (talk) 14:41, 8 September 2009 (UTC).


 * May i suggest the following addition in the introduction: "1⁄299,792,458 of a second. As a result, c is fixed at exactly 299,792,458 metres per second.[1]" This value falls between the upper  and lower  accuracy limits of the Paris standard meter bar.Wdl1961 (talk) 15:44, 8 September 2009 (UTC)

I don't understand LouScheffer's point 3:
 * "The so-called problem that you cannot measure the speed of light in SI did not begin in 1983; the same problem existed since 1960. Suppose the real, physical speed of light changes, with no other physical constant changing.  Post 1983, this cannot be expressed in SI since the meter changes to match.  But this was true from 1960 on - the changed speed of light will affect the wavelength of the krypton transition, and in exactly the manner that will cancel the change in SI units.  (If the speed doubles, the wavelength doubles, the meter is twice as long, and the numerical value of c is unchanged)."

But why would it be the wavelength rather than the frequency (or both) which would change? If the frequency of the radiation were to change in proportion to the change in the speed of light, then the wavelength (and therefore length of the metre) would remain the same. If the period of the radiation from the same krypton transition had been used in 1960 as the basis of the time standard, then in that case the numerical value of the speed of that particular radiation would have been precisely fixed by virtue of the definitions of the length and time standards. But even then, this would not have been true for the speed of radiation of any other frequency, and in any case the time standard wasn't based on the period of any electromagnetic radiation at all in 1960&mdash;it was the astronomically based ephemeris second.

Even the redefinition of the time standard in 1967 to be based on the period of radiation corresponding to a certain transition of the cesium atom still doesn't seem to me to entail that the speed of light of any frequency was then fixed by virtue of the definitions of the standards, since it is logically possible for the speeds, c1, c2, wavelengths, &lambda;1, &lambda;2, and frequencies, &nu;1, &nu;2, of the two radiations to vary independently of each other, subject only to the constraints c1 = &lambda;1 &nu;1 and c2 = &lambda;2 &nu;2. &mdash;David Wilson (talk · cont) 15:14, 9 September 2009 (UTC)


 * Incidentally, for much the same reason, I don't understand why a change in the "real" speed of light (whatever that means) would necessarily imply a corresponding change in the length of the SI metre. Talk of the SI metre changing can only make sense if one has in mind some other standard of length with respect to which it changes, even if this is only an imaginary canonised sacred length unit (cslu) in some idealised Platonic heaven.
 * Brews ohare's discussion seems to presume that the wavelength of the radiation from some atomic transition is less likely to change with respect to such a cslu than the SI metre. But why should we presume that this is so?  And if we assume that the SI metre might be changing with respect to the cslu, why should we not also assume that the SI second might be changing with respect to its own canonised sacred time unit (cstu)?  Ultimately, the "real" speed of light would have to be expressed in cslu/cstu.  If it were to be changing, then one or both of the SI metre and the SI second would also have to be changing with respect to their corresponding canonised sacred units, but I don't see why we should presume that it would necessarily be the SI metre (or only the SI metre) that was changing.  If the SI second were to be changing (relative to the cstu) in exact inverse proportion to the change in the "real" speed of light (as measured in cslu/cstu), for example, then the length of the SI metre would in fact remain fixed (relative to the cslu).
 * &mdash;David Wilson (talk · cont) 23:12, 9 September 2009 (UTC)


 * Since writing the above, I have realised that if one uses radiation from the transition beteeen the two hyperfine levels of the cesium 133 atom's ground state to establish the length of your SI metre (which it seems you are permitted to do by the standard, since it doesn't specify that radiation of any specific frequency must be used) then it will be exactly 9,192,631,770/299,792,458 wavelengths of that radiation long, by definition. So if you were to choose the wavelenth of that radiation as the basis for your alternative standard of length, then it would be impossible for the SI metre to change relative to it (although its definition would of course give rise to inconsistencies if there were radiation of another frequency whose speed was different relative to that alternative length standard and the SI time standard).
 * &mdash;David Wilson (talk · cont) 10:38, 10 September 2009 (UTC)

Constancy of speed of light is just a convention
Can this subsection be made to say something? The way it is, I'd say it suggests there is a topic there, but provides so little one cannot fathom it. Is about generalizing relativity? Is it about quantum gravity? Brews ohare (talk) 05:27, 8 September 2009 (UTC)


 * It is not about those things. It is about alternative mathematical descriptions of special relativity which have different postulates concerning the speed of light, but which nonetheless make exactly the same predictions as special relativity. The existence of alternative ways of describing reality demonstrates that the the speed of light is not a fundamental constant in any way, but a human convention which depends on the way we describe things. Charvest (talk) 05:46, 8 September 2009 (UTC)


 * Chapter 8 of the cited source describes what the authors call "common relativity". They use this version because it is mathematically more convenient when dealing with certain calculations. In this formulation the speed of light is not constant in all inertial frames, which one might think is a problem, but not in the framework of "common relativity" it isn't. Charvest (talk) 05:51, 8 September 2009 (UTC)
 * The work you are citing gives an extremely fringe view. I've been unable to find any major publication in independent journals that support it. This means that including it in the wording you gave would and best a major case of WP:UNDUE. I'm not excluding the possibility of the article mentioning the existence of this view in some way, but it would need some kind of secondary or tertiary source to gauge its place in scientific discussion. (TimothyRias (talk) 08:33, 8 September 2009 (UTC))


 * Well the section has been deleted by two different editors, so I won't re-add it without getting agreement. For the record it said: Concepts such as force, mass and speed are just conventions. Just as epicycles or ellipses are descriptions of the solar system from different points of view, so the whole of relativity is just a mathematical description from a particular point of view. There can be other equally valid descriptions starting from different conventions. One description might be more convenient in certain circumstances, but other descriptions can be more convenient in other situations. In some descriptions the speed of light is not constant . It was deleted with the reason that it was a fringe view. I would say "common relativity" is more acccurately described as a not very well known formulation rather than a fringe view. Charvest (talk) 08:30, 8 September 2009 (UTC)


 * I would add that making changes to the postulate about the speed of light is far from fringe - Test theories of special relativity is all about making such changes. And the "Constant speed in inertial frames" section does say that Einstein was aware that his postulate was just a convention. Charvest (talk) 08:41, 8 September 2009 (UTC)


 * My remarks were mainly aimed at the fringeness of the source used (and the section title). The paragraph in the article itself was so vague it was meanless, and appeared in an odd context for that discussion. (TimothyRias (talk) 10:21, 8 September 2009 (UTC))


 * Fair enough. I've split the discussion of test theories into its own section as that is a separate topic. Charvest (talk) 20:07, 8 September 2009 (UTC)

Charvest, I understand the point that you are making. Some people are surprised that the (one-way) speed of light should be merely a convention but ,as you have pointed out, so is much of physics. This is a subject on which we need to tread very carefully so as not to stir up pointless heated debate here and confuse the general public. See the heated, and in my opinion, rather pointless debate at Evolution_as_theory_and_fact.

I think that there is a case for expanding carefully on the constant speed of light but we should probably tackle the more general philosophical issues by directing the reader to other articles where they can be discussed more fully. Martin Hogbin (talk) 09:31, 10 September 2009 (UTC)


 * Probably best to have a separate article on One-way speed of light. Charvest (talk) 16:19, 12 September 2009 (UTC)

Possible dependence on frequency
This section is unclear. Perhaps it can be combined with "Cosmology and quantum gravity" which certainly looks at frequency dependence? Brews ohare (talk)


 * I question whether this material is sufficiently well established to belong in this encyclopedia. If it is, I agree that it requires a more encyclopedic treatment. —Finell (Talk) 21:48, 8 September 2009 (UTC)

A Google scholar search turns up papers like W. Bednarek and R. M. Wagner that do not attribute this time delay to different speeds of propagation of light, but to a delayed emission model. Therefore, this sub-section is not pertinent to this article. See also Martinez & Errando. Brews ohare (talk) 22:26, 8 September 2009 (UTC)


 * Brews's sources show an alternate explanation of the observations. That, plus the speculative nature of the hypothesis itself, which bets against relativity, and the lack of quality secondary or tertiary sources, persuades me that the observations are not sufficient for an encyclopedia (as distinguished from a paper) to include this material as support for the proposition that c is frequency dependent. I deleted it. —Finell (Talk) 00:43, 9 September 2009 (UTC)


 * I think we need to make a much stronger distinction between established and experimentally verified theories on this subject, which are classical EM and relativity (with QED at the quantum level) and current physics research, experiments, and observations. In the currently accepted and verified theories, c is fixed and an important constant.


 * There is much ongoing research into cosmology and quantum theories of gravity but we should not try to write a general review of such theories or pick out specific ones for mention, unless we have a real expert on such matters amongst us. We should either make a simple statement along the lines of 'There is much ongoing research into... in some of these theories the speed of light may play a less important role or may vary with time, frequency etc...', or we should look for a reputable and accessible general review book or paper on the subject to base our comments on.  In my opinion, it is far better to say nothing at all than to say something wrong, misleading, or even correct but not particularly important.  We should not base our article on the odd paper plucked from arxiv or quotation from a popular science magazine. Martin Hogbin (talk) 09:18, 10 September 2009 (UTC)

Martin: What have your remarks to do with deletion of this section? Are you supporting it or not? Did you put your remarks in the wrong place? What are they about? Brews ohare (talk) 13:11, 10 September 2009 (UTC)


 * Brews, it is rather worrying that you cannot see what my remarks are about. Variation of speed with frequency does not form part of any established and verified theory of physics. Martin Hogbin (talk) 08:44, 11 September 2009 (UTC)

Hypothetical Questions
something i don't get; if earth was moving through space at 99% the speed of light, and someone emitted a beam of light in the opposite direction, would it appear as though the beam is moving twice the speed of light to us, or would it appear still to any bystanders?Chocolog (talk) 13:22, 9 September 2009 (UTC)


 * This is not really the place to ask questions of that nature but the short answer is 'no'. There is no such thing as motion 'relative to space' but the earth could be moving relative to some other object.  Have a look at the section 'Composition of velocities'. Martin Hogbin (talk) 13:30, 9 September 2009 (UTC)
 * Neither, it would appear to move with the same speed for both. I know that sounds ridiculous but it can be verified experimentally. (TimothyRias (talk) 13:32, 9 September 2009 (UTC))

Constant speed in inertial frames.
I am not sure why it was necessary completely rewrite this section, a copyedit is one thing but this was more than that. The sense of the original has now been lost and the text contains a number of errors in the physics. Where now? Martin Hogbin (talk) 19:13, 9 September 2009 (UTC)


 * I've reverted it. Charvest (talk) 19:31, 9 September 2009 (UTC)


 * Corrections to the modifications that the copyediting introduced:
 * Non-inertial frames result from acceleration, not just gravity.
 * Observers do not have motion with respect to their inertial frame.
 * The speed of light has been shown experimentally to be independent of the motion of the source, full stop. Charvest (talk) 20:08, 9 September 2009 (UTC)


 * Thanks, I think this was the best option. If there are problems with the English or style I am sure they can be addressed but at least the physics is now correct. Martin Hogbin (talk) 21:37, 9 September 2009 (UTC)


 * Sorry, I obviously didn't mean to muddle the physics. The statement about spacetime being curved in a non-inertial frame due to gravity, without mention of acceleration, was there before me. I knew it was incorrect, but I didn't change it and didn't look at the cited source. There was a recent awkward edit that drew me to the section, and my only intent was to try to improve the way it read (replace passive with active voice, etc.), eliminate some redundancy, and eliminate a few things that were a bit off the topic of that section. I may try a more limited copy edit, but not now. —Finell (Talk) 00:01, 10 September 2009 (UTC)


 * Finell, I am very happy to work with you to improve that section if you think in is not clear of if you feel the style could be improved. I also wonder if the section should be expanded, it is quite central to the subject of the article. Martin Hogbin (talk) 08:30, 10 September 2009 (UTC)

Metre bar in lead?
There is a new, unsourced sentence in the lead, at the end of the second paragraph: "This value falls between the upper and lower manufacturing accuracy limits of the Paris standard metre bar." "This value" referred to the 1983 redefinition of the metre. I don't believe that the metre bar had anything to do with this, and had been abandoned as the standard metre long before 1983. Should this sentence be deleted? —Finell (Talk) 00:16, 10 September 2009 (UTC)
 * Whether true or not (and given the 'distance' between the 1983 definition and the old platinum standard metre, I'm quite wary of this unsourced statement), this seems like too minor a point to need to be in the article lede. If a suitable source is found, the fact might belong elsewhere within the article. TenOfAllTrades(talk) 01:01, 10 September 2009 (UTC)
 * Wdl1961, who inserted the sentence, now added a web reference (although with an incorrect URL). The source is a web page on an anonymous site, so it is not a WP:RS. Further, the source says nothing about the "manufacturing accuracy limits of the Paris standard metre bar." I deleted the sentence until there is some consensus that (1) it is accurate, (2) it is supported by a WP:RS, and, most importantly, (3) it is sufficient importance to be in the lead. —Finell (Talk) 02:18, 10 September 2009 (UTC)
 * I agree with removing this sentence for the reasons given but I have generally stayed away from editing the lead as I think it needs completely rewriting. I originally suggest that this was done after the content of the main body of the article was stable.  The reason for this is that the lead should be a summary of the article.  It is not the place to introduce new content, or discussions that are not dealt with more fully elsewhere.


 * When things have settled down after the arbitration decision I am going to suggest that we get the content of the article stable then rewrite the lead from scratch to make it a clear and concise summary of the overall article in a good style of English. It should then only be changed to reflect changes to the article. Martin Hogbin (talk) 08:22, 10 September 2009 (UTC)
 * I think this is a clear case of the editor not quite understanding the source: the editor seems to be confusing the old International Prototype Metre with the national copies: the national copies couldn't be exactly the same length because of manufacturing difficulties, so there was a correction applied to each of them (eg, 1 m + 0.2 µm, a hypothetical example in its numbers). There's also a confusion between precision and accuracy: one can be precise without being accurate, and vice versa. The International Prototype Metre was defined as being exactly one metre long, so there were no manufacturing defects! Well, actually, there were, but these should be discussed in Metre on in an even more specialized metrology article, not here IMHO. I agree that the comment should be removed. Physchim62 (talk) 10:49, 10 September 2009 (UTC)

It is merely an attempt to relate the real physical meter used by everybody in their daily lives versus an item that the physicist  in this article have not agreed on how to measure without the Paris standard bar .This puts a limit on    fundamental circular confusion.Wdl1961 (talk) 15:50, 10 September 2009 (UTC)


 * There hasn't been a Paris standard metre bar for many decades. When most people think of the metre, they think of metric rulers; some Americans may think of a yardstick plus (approximately) 3 inches. These physical embodiments of the metre and its subdivisions (decimetre, centimetre, millimetre) conform as closely as any human can perceive to the current, 26-year-old definition of the metre based on c (and vice versa); there is nothing unreal or unphysical about it. I suspect that Wdl1961 (and also Brews ohare, and possibly Abtract) may have been confused, understandably, by Tombe's fallacious "tautology" arguments. That is all the more reason that the article must be expunged of any traces of this confusion, and that the article should dispel any confusion that might arise from the apparent tautology (which is also apparent in other modern measurement standards). —Finell (Talk) 16:26, 10 September 2009 (UTC)


 * Sorry i have lived with meters ,yards, el, nautical miles ,miles, figertip to fingertip and have no problems. Just switch to different definitions ,discipline ,years and areas and languages. All i wanted to do is show relative precision ,(.)period. For confusion look above and below. Pls do not categorize me.Wdl1961 (talk) 19:30, 10 September 2009 (UTC) err:reproducible relative precision.Wdl1961 (talk)


 * Can someone explain some basics . We have an  measurement and call it a def, no problem except  we have adjustments and  problems  a,b,c d etc. generating a little slack. We get a new superior  measurement def, no problem except we have  adjustments  p, r, s and t  generating a little slack .  Someone has the audacity  to ask how  the new slack  boundaries fall relative to the old slack boundaries does not understand the problem? Wdl1961 (talk) 01:53, 11 September 2009 (UTC)

Doppler
In the box under "Doppler Effect", the words "right" and "left" seem to be mistakes for "top" and "bottom". —Preceding unsigned comment added by 217.44.109.137 (talk) 08:56, 10 September 2009 (UTC)
 * Sorry, forgot to adjust that as well when I introduced the more compact version of this picture. (TimothyRias (talk) 09:51, 10 September 2009 (UTC))

Speed of light set by definition section
Currently, this section drifts somewhat off-topic in to the finer points of metrology concerning the metre, which is not the subject of this article. What is it that we really need say in this section. The first paragraph no covers the when, how and why of the redefinition of the metre. (Should the why part be expanded more upon?)

(The gist being that we could measure the speed of light more accurately than we could measure the then current metre and thus decided to use it as a new (more precise) definition.)

What else should we discuss? There should be at least something about the value of the speed of light become exact when expressed in m/s (and any units derived from those). IMHO there should also be a few lines explaining that even though the value of the speed of light is fixed, that this does not change anything of the involved fixes. It does not change anything about the physical act of measuring the speed of light (= comparing it to some other speed). All this, of course, with an adequate source.

Is the tag on paragraph about the word vacuum in the definition really needed? This pertains much more to the definition of the metre than to the speed of light. If it is thought that this is really needed, it may be much more functional in a footnote to the definition, instead of at the end of the section where it really feels like something that was tagged on. {Unsigned contribution by User:TimothyRias 10:16, 10 September 2009}


 * Hi Timothy: Your statement that "we could measure the speed of light more accurately than we could measure the then current metre and thus decided to use it as a new (more precise) definition.)" is not quite so. The difficulty in measuring the meter was the error in fringe counting. By switching from a length (fringe count) to a time-of-flight definition, fringe counting was eliminated and only time had to be measured. Naturally, with only time to measure, the error was only that of time measurement. The switch in definition is not about improved accuracy in measuring the speed of light. In fact, the speed of light became a defined not a measured quantity. Take a look at Sydenham and Jespersen & Randolph. Brews ohare (talk) 13:28, 10 September 2009 (UTC)


 * (PLEASE READ THIS INTERJECTION): Brews: This is the heart of the issue. What you are forgetting, or omitting, is that what you call the real, physical speed of light was measured to a high degree of accuracy using the prior standard for the metre (one that you apparently prefer), that real the physical measurement (actually, many measurements) of the speed of light was used to define the metre in 1983, and the real speed of light (which can be and is measured against standards that are not based on c) has not changed since the 1983 definition. The current speed of light is a measured quantity, that measured quantity was used to redefine the metre, and that redefinition fixed the speed of light at its previously measured value. The constancy of the speed of light was fundamental to using it as the basis for the 1983 definition. Brews, you ask us to read and understand what you write. Please read and understand this paragraph, because in explains away the problem that you think needs to be addressed. If others here think that this explanation is incorrect, please explain and demonstrate where I have erred. —Finell (Talk) 17:04, 10 September 2009 (UTC)


 * Finell: Hi. I do not prefer the older definition - it has its pros and cons. The "old" value of c = 299 792 458 ± 1.2 m/s, which means it is uncertain what the exact value is, it is bracketed. The 1983 decision did not make measurements suddenly exact. What it did was to sidestep the exact determination of c by defining it as c = 299 792 458 m/s and switching to time-of-flight as a way to compare lengths. What that means is that the metre is now uncertain, rather than c. Thus, we still have uncertainty in the speed of light because we don't know "exactly" what the metre is: it is how far light goes in 1/299 792 458 s, but how far is that? It is 299 792 458 m/s × 1/299 792 458 s = 1 m. Yeah, but aren't we going in circles? We have to introduce a differently defined length to answer that question. Brews ohare (talk) 17:16, 10 September 2009 (UTC)
 * The metre was already uncertain before 1983, the relative uncertainty was $4$, which meant we could not measure the speed of light with any more certainty than the value you give, simply because we did not know exactly how long a metre was. After the 1983 definition we still cannot determine it exactly but the precision with which we can has increased at least by an order of magnitude. (probably more but I can't be arsed to go look for the current exact limit.) (TimothyRias (talk) 17:31, 10 September 2009 (UTC))


 * You seem to have misread what I said. All I said was that the new definition was chosen to obtain a more precisely realizable definition of the metre. (TimothyRias (talk) 15:54, 10 September 2009 (UTC))


 * Timothy what you said exactly was "we could measure the speed of light more accurately than we could measure the then current metre and thus decided to use it as a new (more precise) definition.)" That says more accurate measurement of c was the critical factor. If you wish to rephrase that, that's fine, but I did not misread you. What is your more carefully worded version of what you wish to say? Brews ohare (talk) 17:22, 10 September 2009 (UTC)
 * The critical factor was indeed that we could measure the speed of light more accurately than the then metre (i.e. uncertainty in the metre was a limiting factor on the accuracy of the speed of light measurements.) This meant that the change of definition immediately led to a more precisely defined metre. Which is what I said before, but you completely failed to understand. (TimothyRias (talk) 17:49, 10 September 2009 (UTC))


 * Timothy: Please read Sydenham and Jespersen & Randolph. The increased accuracy in the metre is entirely due to the switch to time-of-transit and away from fringe counting and has absolutely nothing to do with increased accuracy in the measurement of the speed of light. Brews ohare (talk) 18:00, 10 September 2009 (UTC)
 * You are just failing to make the connection. Time-of-transit measurements where used to measure the speed of light in the past. The limit on the accuracy with which that could be done was in the way length was defined. In principle the accuracy of the time-of-transit measurements were much higher, which is why it was a good idea to use them as the base definition for length instead. This is the general practice in metrology if you find something that you can measure (consistently) more accurately than your current measurement standard, then you adopt that as your new standard. The concept that you seem to be opposing is that the time-of-transit of light is a measure for length, which is just contrary to all expert opinion on the matter.(TimothyRias (talk) 07:39, 11 September 2009 (UTC))


 * Timothy: Your remark "even though the value of the speed of light is fixed, that this does not change anything of the involved fixes. It does not change anything about the physical act of measuring the speed of light (= comparing it to some other speed)." Is worth some discussion here. Of course, you are right that one can compare the speed of light with any other speed just as before, because that is a dimensionless ratio and is independent of units. However, that does not measure the speed of light except as a multiple of some other speed. The speed of light in units of m/s cannot be measured any longer, because it is a defined value of 299 792 458 m/s and that is it. Brews ohare (talk) 13:44, 10 September 2009 (UTC)


 * Question: in 1900 was it possible to measure the length of the the international prototype? Sure it was, (with an interferometer for example) the result was however called 1 metre per definition. My point here is that the physical act of measuring is not related to units at all, the units only enter in how we then express the result. (TimothyRias (talk) 15:54, 10 September 2009 (UTC))


 * Of course interferometry was used to define lengths until 1983. Then time-of-transit was introduced, and length is no longer a measurement separate from time. I'd say that is a very different physical act of measurement, and that difference is why the comparison of lengths is now more accurate. Lengths are now in seconds. That means speeds cannot be measured in m/s, really. For example, if we want to tell how fast a car moves, we mark out a distance, find how long the car takes (tcar) and find out how long light takes (tc) Then how far the car went is cotc, and the speed of the car is co tc/tcar. We know the speed of the car only as some fraction of co = 299 792 458 m/s (an arbitrarily defined value). Do you disagree? Brews ohare (talk) 17:26, 10 September 2009 (UTC)
 * You failed to answer my question. In 1900, could you measure the length of the international metre prototype?(TimothyRias (talk) 17:51, 10 September 2009 (UTC))
 * Yes, indeed, using fringe counting. Today you could still do that, but instead it is done by time-of-transit. This last provides a measure that can be compared more accurately with other lengths than a fringe count comparison. Brews ohare (talk) 18:04, 10 September 2009 (UTC)
 * But in 1900 the metre was per definition the length of the international prototype. Even though its length would always be exactly 1 metre you could still measure its length. Maybe relevant to your confusion on this subject: You could measure the length of the international prototype today using the present day definition of the metre and you could find that now it is not exactly 1 metre (although since the length of a metal bar is actually not all that well defined you would probably that it is 1 metre within experimental bounds). Any change in definition will bring with it a change in the length of the "actual metre". The important part is that the new definition should be equal to the previous definition within the experimental bounds of the time. (In this way the measurement error on previous experiments is not increased.) (TimothyRias (talk) 07:49, 11 September 2009 (UTC))


 * Timothy, That's it in a nutshell. The new SI units speed of light is beyond measurement. It is a definition with an arbitrarily chosen number. The speed of light in all other systems of units is a physical quantity that can be measured. This article is about the speed of light, and so it should be concentrating on the latter concept. Physics as a topic is bigger than any particular system of units. We cannot subordinate the physical speed of light to the SI system. David Tombe (talk) 13:58, 10 September 2009 (UTC)


 * That is in no way a response to any of my questions. (TimothyRias (talk) 15:54, 10 September 2009 (UTC))

Timothy: Your statements: "Is the tag on paragraph about the word vacuum in the definition really needed? This pertains much more to the definition of the metre than to the speed of light." also deserve some discussion here. This section is about "the speed of light by definition". That topic is inextricably bound to the definition of the metre, as that is how the defined speed of light arises. Thus, it is difficult to support the notion that somehow the two definitions (metre and SoL) are separable. In the same vein, the definitions of both of these are inextricably related to the 'vacuum', and it seems to me pertinent to point out that the 'vacuum' in the sense meant here of 'classical vacuum' is not your garden variety terrestrial vacuum. In effect, the "defined value" of the speed of light applies nowhere in the universe but measured values of the speed of light in this universe can be referred to 'vacuum' by implementing the proper corrections of standard practice (as approved by the BIPM). Brews ohare (talk) 13:53, 10 September 2009 (UTC)
 * No, the section is (should be) about the historical episode in which due to extraordinary precision with which it could be measured the speed of light was chosen as basis for the definition of the metre. The exact details of that definition are somewhat secondary. The article already treats somewhere else that light only travels with the speed of light is free space/absolute vacuum/whateveryouwanna call it. (TimothyRias (talk) 15:54, 10 September 2009 (UTC))


 * Timothy: You are back to due to extraordinary precision with which it could be measured the speed of light was chosen as basis for the definition of the metre, which as discussed just above, is a misconception. But that is not the point here. The point here is that the section is about "the speed of light by definition and the definition includes 'vacuum' and so is properly part of this subsection. Brews ohare (talk) 17:54, 10 September 2009 (UTC)

Timothy: Another point you have raised is "the physical act of measuring the speed of light". Inasmuch as the speed of light is defined in SI units, such a measurement must go beyond SI units. That is, a unit of length must be introduced that, unlike the metre, is not defined in terms of the numerical value of the speed of light, but that is independent of the numerical value of the speed of light. So, for example, one could use as a unit the wavelength of some transition, and count fringes. (The wavelength approach assumes the speed of light is a constant, but does not prejudge its exact value.) That would produce the speed of light in wavelengths/s, and would introduce the errors of fringe counting. So the measured speed of light in wavelengths/s would have similar error bars to that using the old pre-1983 metre, which after all was used in exactly this way. Perhaps something about this should be added? Brews ohare (talk) 14:17, 10 September 2009 (UTC)


 * Just as with the measuring the international prototype in 1900, it is still possible to measure the speed of light in whatever way you want, we just choose to call the result 299 792 458 m/s per definition. (That is the speed of light has become the SI basic unit of velocity).(TimothyRias (talk) 15:54, 10 September 2009 (UTC))

Timothy: Not to be cute about it, but if one imagines the physical act of measuring the speed of light, and one entertains that the 'exact' speed of light is 299 792 458 m/s, what exactly is one trying to measure? My point is that there are two concepts here (a point that Dicklyon will fight to his dying breath), namely the SI units defined value and the value one seeks when doing "a physical act of measuring". Brews ohare (talk) 14:32, 10 September 2009 (UTC)
 * There is only one concept here. The universe (well at least the local string vacuum ;))knows a basic velocity which we call the speed of light, a dimensionfull concept which at least a certain portion of the physicist consider fundamental. It just so happens that the SI has chosen this physical quantity as one of the seven base quantities in term of which it defines its 7 basic units. (TimothyRias (talk) 15:54, 10 September 2009 (UTC))


 * Well, yet again, we have three comments from and one comment from  that serve only to promote their own idosyncratic views and not – in any way – to answer the editorial question posed at the top of this section. Physchim62 (talk) 14:37, 10 September 2009 (UTC)
 * Physchim62: On what basis do you say these views are "idiosyncratic"? They are sourced to Sydenham and Jespersen & Randolph. They are not pronouncements, but matters that can be discussed further by those who are willing to put aside violations of WP:NPA and address issues, not authors. Contrary to your unsupported pronouncement, these comments are directly related to exactly the matters raised by Timothy, and in fact each comment is prefaced by exactly the point raised by Timothy to which a response is offered. Brews ohare (talk) 14:48, 10 September 2009 (UTC)

E = m c2 or E = m c02
And a final topic, not so far addressed in this subsection, is that of further implications of the definition. For if the metre is defined in terms of time-of-transit, making c exactly 299 792 458 m/s, then that has implications for things like E = m c2, for example. Is this now E = m (299 792 458 m/s)2 exactly? Brews ohare (talk) 14:42, 10 September 2009 (UTC)
 * Yes, it is exactly. (unless the velocity of light in absolute vacuum is not equal to the spacetime constant c, which means that a)SR is wrong or b)the photon has mass.)(TimothyRias (talk) 15:54, 10 September 2009 (UTC))
 * Why wouldn't it be? c is exactly 299792458 m/s, any formula involving c will be "exact" if you change one for there other. Headbomb {{{sup|ταλκ}}κοντριβς – WP Physics} 16:37, 10 September 2009 (UTC)

Timothy, a discussion of the kind that you propose would be a welcome change after the endless nonsense (I admit to not even reading it this time) that Brews and David talk. Martin Hogbin (talk) 15:43, 10 September 2009 (UTC)


 * It's exactly the same question as I raised regarding c2 = 1/εμ. Do we read it from right to left, or from left to right? Which side is the measured variable? Maxwell's equation (132) (1861 paper), which is Newton's equation for the speed of sound, is actually both  c2 = 1/εμ and  E = mc2. Therefore as regards E = m c2, the c2 side would have to follow from the measured value of E. It can't refer to any defined value of c2 David Tombe (talk) 16:23, 10 September 2009 (UTC)


 * David Tombe: It is something that puzzles me too. The standard NIST answer is that c is fixed by the definition of the metre at a definite numerical value, and μ also is fixed. Hence, c2 = 1/εμ determines ε exactly as well. I think the answer to your question about the connection to measurement is that these quantities apply to 'vacuum', and 'vacuum' cannot be measured. So the question pushes back to understanding what effect the 1983 definition has upon the "standard good practice" corrections that bring measurements to refer to 'vacuum'. Could that be it?  Brews ohare (talk) 19:08, 10 September 2009 (UTC)


 * The question of what c to put in E=mc2 is fairly complicated. In pre-1983 SI units, one would put in 299 792 458 ± 1.2 m/s, and so would obtain an uncertainty in the value of E corresponding to m. Inasmuch as the 1983 decision simply replaced c by its most probable value of 299 792 458 m/s, on the face of it putting E = m (299 792 458 m/s)2 exactly appears to have swept the uncertainty under the rug. I don't think anyone would argue that the variation in values of E pre-1983 went away because of a decision of the BIPM committee? Brews ohare (talk) 17:03, 10 September 2009 (UTC)


 * Brews: Scientists aren't that dumb. If there were any application where the ± 1.2 m/s were significant, they would take that margin of measurement error (or a reduced margin of error based on more refined measurements) into account, notwithstanding the 1983 definition of the metre. —Finell (Talk) 17:19, 10 September 2009 (UTC)


 * Finell: I am not suggesting that scientists are dumb, and I am not suggesting that E = m (299 792 458 m/s)2 exactly (assuming m is known exactly) is incorrect. My point is that one has to interpret the relation E = m (299 792 458 m/s)2 exactly in a different manner than one interprets the pre-1983 relation E = m (299 792 458 ± 1.2 m/s)2, because the new E is measured in the new SI units which places the error bar in a different way. Brews ohare (talk) 17:47, 10 September 2009 (UTC)


 * Under the 1983 definition of the metre measuring the rest mass or rest energy are completely equivalent. So, yes it should be E = m (299 792 458 m/s)2 exactly. (With the very big caveat that the unit for mass, the kg, is not very precisely known) (TimothyRias (talk) 07:30, 11 September 2009 (UTC))
 * Addedendum: The inaccuracy that you think was swept under the rug bu the new definition was simply the in accuracy with which the Joule was defined in terms of the kg. Since that uncertainty has been lifted with the new definition, the inaccuracy has indeed disappeared (not entirely since there will always be errors in actually realizing a reference Joule which will be included in the error of any energy measurement).(TimothyRias (talk) 07:57, 11 September 2009 (UTC))


 * If we put aside the errors associated with the mass, there is still a remaining question of where the error bar in E = m (299 792 458 ± 1.2 m/s)2 (pre-1983) went to on the day in 1983 when this equation was changed to E = m (299 792 458)2. We know that the error in the speed of light ± 1.2 m/s reappears in the definition of the metre (Any refinement in measurement does not change the speed of light, but only the metre.), so I'd assume that error bar in the metre also is buried in E in the same way. Brews ohare (talk) 14:13, 11 September 2009 (UTC)


 * Actually, much of that ± 1.2 m/s disappeared on that day in 1983, since it was mostly due to the uncertainty with which the length of the old-metre was known. The length of the new metre could however be realized much precisely. Of course, the error bar remains if you try to convert new metres to old metres since the conversion between these two was not known as precisely. The same thing holds for the old and the new Joule. (and that is where the error has gone.) (TimothyRias (talk) 14:43, 11 September 2009 (UTC))


 * Timothy: I think we have distilled the source of difference between us. My take is that indeed the change in 1983 to transit-times for lengths improved the accuracy of length comparisons because only times had to be measured. However, we disagree that this improvement in length comparisons translates to an improvement in accuracy in the measured numerical value of the actual speed of light: instead, my understanding is that the speed of light was simply defined and placed beyond measurement in SI units. A number of sources make this kind of statement; for example Sullivan says "the speed of light need never be measured again" and Adams says "any improvement in measurement affects the meter, and doesn't change the speed of light". I'm not quoting exactly, but I'm sure you will agree that is what these sources say.


 * The definition of the metre says it is the length traveled by light in 1/299 792 458 s, which definition does not require any knowledge of the numerical value of the real speed of light. And of course, if light travels a metre in 1/299 792 458 s, then the speed of light in SI units is 299 792 458 m/s, no matter what. Brews ohare (talk) 15:59, 11 September 2009 (UTC)


 * FWIW, as for joule, the uncertainty in the kilogram is by many orders of magnitude larger than the uncertainty in the 1960 metre, so the 1983 re-definition of metre practically didn't affect it. -- _ _ _ A. di M. 14:48, 11 September 2009 (UTC)

Brews, Regarding your query above, the equation c2 = 1/εμ is only meant to be read from right to left. It originates with a capacitor experiment in 1856. The significance of it is that it relates the electric and magnetic constants to the measured speed of light and hence leads to an important convergence in experimental physics. Maxwell then used theoretical physics to show that these experimental results demonstrate that light is an electromagnetic wave. Now we all know that μ has been a defined quantity for some time. In Maxwell's time, μ was the density of the luminiferous medium, and as such you can see the link with E=mc2 when we consider that ε is the inverse of the transverse elasticity (a Young's Modulus type concept). So long as ε remained a measured quantity, which it was up until my teaching days, then it wasn't a major problem that μ is a defined quantity. The equation still holds, and it yields a number that is very close to the measured speed of light. My main problem with all of this is, that by using this equation in reverse with the defined speed of light, we end up with a defined value for ε. And I was very concerned to see that the SI definition of the metre has led to the removal of the associated capacitor experiment from the textbooks. So your original point is correct. The defined value of c does indeed have repercussions for the equation E=mc2, which are being stringently denied by your opponents here. That is why it is impotant that the article is segregated into (1) the physical speed of light and (2) the defined speed of light. David Tombe (talk) 08:49, 11 September 2009 (UTC)

Oops!
I've made two or three little edits which I don't think are contentious in their context, but I've just realised that I've used ν (Greek nu) for frequency, while elsewhere other editors have used f. It doesn't really matter to me which we use, as each one is used almost exclusively in some contexts and almost never in others, we will never please everyone! What do other editors think? I guess the "default" position is to standardise on f, but I wanted to check before changing (or not changing) anything. Physchim62 (talk) 10:24, 10 September 2009 (UTC)
 * I prefer f for non-technical audiences. As a bonus it looks nicer inline.TStein (talk) 17:27, 10 September 2009 (UTC)
 * I agree. —Finell (Talk) 17:43, 10 September 2009 (UTC)

Scientists in the 11th century?
My understanding is that there was not a such thing as a scientist or "science" until the 17th century or so. Looking at the science article, the people studying light's nature in the 11th c. through the period before the scientific revolution would have been called "natural philosophers." I guess my point is that to call the pre-scientific revolution people scientists is a bit misleading, or even inaccurate. Hires an editor (talk) 00:05, 12 September 2009 (UTC)

The elephant in the room (c vs v)
The real problem with all the 'debate' lately is that it distracts from the real problems with this article. The biggest problem that I think we need to resolve now is whether this article is about c--the physical constant corresponding to the speed of light in a perfect vacuum--or v--the speed at which light travels.

Currently we mix the two concepts too much; easily confusing non-technical readers. For example as the lead stand the first paragraph talks about the physical constant behind space-time, etc.. (c). The second paragraph talks about the speed at which light travels (v). (Incidentally it also includes the value for c for the third or fourth time.) The third paragraph repeats much of what is in the first paragraph (c) but then quickly changes gears to talk about (v).

This would not be so bad except that the transitions are not made clear. More importantly a non-technical reader reading about a constant c called the speed of light can easily assume that when the speed that light travels is reduced that c is also reduced; in other words that the properties of space-time depend on the speed that light travels, and not vice versa. (After all, c is not called the space-time constant.) We need to clearly distinguish from the very first paragraph that we are talking about two separate concepts in this article.

This is not just a problem with the lead either but infects the entire article.

As it currently stands the Second section is entitled 'Fundamental importance in physics' and mostly covers c. (As a side note, I absolutely detest these 'in physics' sections since it doesn't explain what c is; it is just a hodgepodge of stuff that happens to be taught in physics with no other self-consistent reason to be grouped together.)  The first subsection is about electromagnetic radiation traveling at c in Maxwell's equations. There are good reasons to put this first; the problem is that it encourages the reader to think that c is primarily about the speed that light travels and that the properties of spacetime follow from this relation; they are after all introduced next.

The light as EM radiation subsection is followed by a number of subsections that should be placed as subsections of a section called spacetime constant. (Subsections such as causality, etc are a consequence of c being the spacetime constant.) Thsee subsections are followed by light and photons which has nothing to do with the speed of light. The section is capped off by variations in the speed of light whose first paragraph is about possible variations in the space-time constant c, and the second is about deviations of the speed that light travels v from c confusing v and c further.

For those following along the second section started of with a subsection that could be interpreted as covering c or v followed by a number of subsections dealing strictly with c, followed by a subsection that has nothing to do with either, and capped off with a section that covers c in one paragraph and v in the next.

The third section Light in a Transparent Media covers exclusively v which is good. It could use a better description of the differences between group, phase, and front velocities. (The phase velocity is often greater then c but even the group velocity can be greater then c in certain cases.) I don't like the title of this section though because the title is about light and not about the speed of light, though.

The Fourth section Faster-than-light... is one of the most confusing in this aspect because it doesn't distinguish well enough in my opinion between faster-than-v (Cherenkov radiation) and faster then c everything else.

The fifth section Practical effects of the finite speed of light is mostly about v since they don't really depend on the properties of space time only the speed at which light moves. (I am uncertain how practical Terrell rotation is.)

Conclusion: Speed of light has to very separate meaning (c and v) and we don't help the readers by mixing them. If I had my way I would have two separate articles named speed of light (physical constant) and speed of light. Each article would have an about section at top such as

Thoughts.

TStein (talk) 21:11, 10 September 2009 (UTC)
 * Indeed; the constant c is usually referred to as the "speed of light", even if it's significance is much, much wider. The fact that it is the speed at which massless particles travel is just one aspect of it. (In addition to that, if the photon had a non-zero mass, the speed of propagation of electromagnetic radiation would depend on its frequency and would always be less than that.) Unfortunately the name "speed of light" has a literal meaning which is logically different, so we get apparently nonsensical statements such as "light in media travels slower than the speed of light". In an ideal world, c would be called Lorentz constant (or Minkowski constant, or whatever), "speed of light" would mean exactly what it seems to, and we could say "in free space the speed of light equals the Lorentz constant, but in transparent media it is slower". --  _ _ _ A. di M. 22:56, 10 September 2009 (UTC)


 * What TStein refers to as c versus v has been discussed before on this Talk page. Someone (Dicklyon?) quoted Encyclopaedia Britannica, which explained both senses of speed of light, as part of the discussion, but the consensus was that speed of light means c (or, as some notate it, c0), and that this is what most people think of as the speed of light. I don't believe that most people are aware of the distinction. Saying that in matter "light travels more slowly than c" sounds less awkward than saying "light travels more slowly than the speed of light", which is a head-scratcher; but the underlying problem is still there. To deal with it requires revisions throughout the article, and more discussion of the already contentious lead. I suggest that we save this issue until the arbitration is over, when we may have a better environment for discussion and editing. I'm not completely happy with using vacuum instead of free space, but c versus v is a more important issue. —Finell (Talk) 01:32, 11 September 2009 (UTC)
 * My suggestion:
 * Avoid using the exact phrase "speed of light" to refer to anything other than c; say "the speed at which light travels in ..." or reword the sentence instead.
 * Refer to the constant as c in any context where "speed of light" might be misunderstood (i.e. whenever talking about transparent media or hypothesizing what would happen if the photon were massive).
 * -- _ _ _ A. di M. 08:33, 11 September 2009 (UTC)
 * Use c rather th
 * A. di M.,I think that is an excellent idea. Unfortunately there is little point attending to such important details whilst the page is still in a state of flux due to apparently never ending discussions about the definition of the meter.  This is the problem that I have been fighting for over a year.  We need to get the major content stable before we can attend to improvements and clarification like you suggest.  If we were to change the wording now it would likely be swept away when the next well-intentioned editor rewrites the the article to deal with Brews' latest musings. Martin Hogbin (talk) 08:59, 11 September 2009 (UTC)

Tstein, regarding your proposals that the article be split into two different concepts of the speed of light, see my reply to Brews three sections up. David Tombe (talk) 08:50, 11 September 2009 (UTC)
 * I think it is important to point out that empirically the speed of light in a vacuum (as near as that can be approximated, which is pretty well for astronomical measurements) is equal to the "Lorentz constant" (as A. di M. has christened it). The agreement is at least as good as a few parts in 109. We would be able to tell if it wasn't, because length measurement by interferometry depends linearly on the speed of light, while the equations of SR are quadratic in c0.
 * Einstein certainly considered that the speed of light in his equations was exactly that. He actually used the symbol V in his annus mirabilis papers, E = mc2 wasn't written that way until a couple of years later. Nobody has actually given it another name, it's not our place to do so now (except for the purposes of avoiding confusion in this discussion). To say that light travels at the Lorentz constant in a perfect vacuum, or something similar, is well-intentioned nonsense: we'd save ourselves a lot of time and effort if we concentrated on mainstream physics instead of our own speculations. Physchim62 (talk) 10:22, 11 September 2009 (UTC)


 * Physchim62: This discussion would be easier to follow if words were used more carefully, particularly the word 'vacuum', which may refer to the BIPM 'vacuum', often called classical vacuum, or may refer to some actual real medium, like outer space, or some theoretical vacuum that is realizable in principle but not yet verifiable, like quantum vacuum or perhaps QCD vacuum. To which does the "Lorentz constant" apply? Brews ohare (talk) 13:47, 11 September 2009 (UTC)
 * The Lorentz transformations and the constant c which appears in them are properties of space-time and hold in any medium (provided you can neglect gravity). -- _ _ _ A. di M. 14:39, 11 September 2009 (UTC)


 * Good point. So as the Maxwell relations in any isotropic dispersionless medium exhibit Lorentz symmetry with c = speed of EM radiation in that medium (regardless of units chosen), the underlying question is which medium to choose to obtain the c in SR? Brews ohare (talk) 15:05, 11 September 2009 (UTC)


 * No they don't. c is always the same, in any medium. Speed of EM radiation in that medium = c/n, where n = 1 in free space and n < 1 in material media; n ≈ 1 in any medium which is rarefied enough. (And the macroscopic Maxwell's equations in a medium aren't Lorentz-invariant because there's a privileged frame, namely the one in which the medium is stationary.) Some people call "c0" what I called "c" and "c" what I called "c/n", but photons don't give a damn about which notation you use. -- _ _ _ A. di M. 15:16, 11 September 2009 (UTC)


 * OK, educate me a bit here. It was my impression that within SR at least the laws of physics are Lorentz invariant for any observers in constant relative motion. That applies to Maxwell's equations too. So if we suppose the universe to be pervaded by some 'vacuum' with some particular properties, for example, ε0, μ0, will we not find that all inertial observers agree that EM radiation travels at c02 = 1/ (ε0 μ0) ? And mightn't we decide that the c in the Lorentz transformation is the same as c0? Historically, isn't that where SR came from? Brews ohare (talk) 15:33, 11 September 2009 (UTC)
 * Physical laws are Lorentz invariant, and so are the microscopic Maxwell's equations (those with E and B). On the other hand the equations with D and H are macroscopic approximations in which you don't care about the microscopic currents and charges, and so you "average them out"; that is done assuming that the atoms which make up your medium are stationary w.r.t. your frame of reference, and so it doesn't apply to other frames unless you use some corrections. (This is like the fact that the laws directing the motion of air molecules are Lorentz-invariant, but equations describing their collective behaviour such as Navier-Stokes' aren't.) In free space there's nothing which you have to be stationary with respect to. As for your first question, I don't think it can be answered unless you specify what those "particular properties" are. -- _ _ _ A. di M. 15:47, 11 September 2009 (UTC)

OK, so I can write down the free-space Maxwell equations with whatever values I like for ε0, μ0 and will find EM waves travel at c0. Moreover, I can follow the historical path and determine that my SR applies theoretically with c0. Now I have to ask whether this theory applies to nature. So I now am faced with a measurement in a real medium, as no other medium is measurable. Maybe I choose outer space. I find that within some error bars and for a particular set of units a particular value of c0 works well (e.g. ± 4 × 10−9 for EM observations). How do I distinguish between the EM value of c0 and the SR value of c0? That cannot be done unless I go outside EM, right? For example, look at elementary particle dynamics. That is a missing section in this article, eh? Brews ohare (talk) 16:12, 11 September 2009 (UTC)

Responses from TStein
Finell, despite trying to keep an eye on this page I missed the previous discussion. *Sigh* More evidence, if we need any, that this dispute needs to be resolved. Any number of important discussions can be buried under all this other chatter. I am beginning to agree that we should postpone this until after arbitration and the page stabilizes. I was hoping (probably naively) that talking about and fixing something that is important would give those fighting over something of lesser importance something better to do with their time and energy.

A. di M., I agree with Martin, your ideas are a good start.

Martin, If we have any hope that this will end soon then we can wait. If not, then we have two choices: either give up on this article and focus on the many other articles that need improvement or forge ahead anyway.

David, if I interpreted your responses correctly, I think that I am discussing a different topic then you and Brews are. My concern has nothing to do with how the speed of light is defined. My concern is two-fold: that c refers to something broader then the 'speed of light in a perfect vacuum' and worse that readers can easily confuse c and v for light inside a material and think that the properties of spacetime depend on v.

Physchim62, I wasn't aware that the empirical evidence that the SR c is the same as the speed of light in a vacuum-c was so good. Out of curiosity, do you have a reference to that? Maybe I shouldn't be surprised since science is able to measure time to a crazy precision. I don't think anybody is suggesting that we give c another name.

TStein (talk) 14:04, 11 September 2009 (UTC)


 * The relative uncertainty in the speed of light pre-1983 is stated by the BIPM as ±4 × 10−9, and is cited in the section on speed of light by definition. Brews ohare (talk) 14:28, 11 September 2009 (UTC)
 * Re. "empirical evidence", see this, in particular section 6. It cites a paper by G.L. Greene et al. which is "An analysis combining the results of several experiments gives the result that the Lorentz limiting velocity must be equal to the speed of light to within 12 parts per million." (To Brews: I think you've misunderstood the point. They were talking about experimental evidence that the constant c appearing in the Lorentz transformations and the speed of light in free space are equal; that is a meaningful concept even without reference to the metre or any other unit. (For example, I can determine whether the lengths of the two sheets of papers which are on my desk right now are equal to within much greater precision than I could measure their length in Egyptian cubits.)) -- _ _ _ A. di M. 14:34, 11 September 2009 (UTC)
 * You're right; I missed the point. Brews ohare (talk) 14:58, 11 September 2009 (UTC)
 * And I have naïvely missed a point as well, going beyond my sources, I apologise. I was working from quoted uncertainties in astronomical measurements, but I failed to take account of the fact that the relativistic correction is much smaller than the measurement itself, and that any errors propagate in different ways depending on what value you want out at the end of your calculation. I'm glad to see that the paper of Greene et al. (1991) gives a relative uncertainty of 10−5. Independently, with back of the envelope calculations but working with more recent reports, I'd reckon that a difference of a few parts in 107 between c0 and cspace would now be clearly visible, and that you could probably shave an order of magnitude off that (down to parts in 108) if you re-examined all the data actually looking for a difference (ie, that we don't know at present that there's not a difference at parts in 108, because we haven't actually looked for it). You couldn't get down to the four parts in 109 that is the current best non-SI determination of the speed of light, nor the parts in 1010 that some groups have claimed for the astronomical (non-SI) speed of light (out as research papers, but not yet included in compendia). I'll discuss this further if people want, and there are relevant references at the astronomical unit article, but I'll keep this post brief to let others comment. Physchim62 (talk) 22:47, 11 September 2009 (UTC)

To get the ball rolling I've started the article Speed of light in a medium. Charvest (talk) 07:59, 12 September 2009 (UTC)
 * Ok, I've taken the liberty of splitting the article. What do you think? Sorting out the history section between the two articles will take some work. Charvest (talk) 08:39, 12 September 2009 (UTC)

I have just checked the date and it is not 1 April. Is someone suggesting that we have just discovered that the term 'speed of light' can have several meanings including the spacetime constant, the actual speed light would go in free space, and the speed light goes in real media, or is it that we have just discovered that outer space is not exactly the same as the hypothetical free space? Martin Hogbin (talk) 09:05, 12 September 2009 (UTC)

No consensus to split this article in two
Out of the blue, this article has been slit in two? Martin Hogbin (talk) 09:07, 12 September 2009 (UTC)
 * Maybe it was an attempt to answer your question above… Physchim62 (talk) 09:11, 12 September 2009 (UTC)
 * Contrary to what some may think, the points I made above have been know for years. The place to make the situation clear is in this article. Martin Hogbin (talk) 09:16, 12 September 2009 (UTC)
 * I meant your rhetorical question 'How much of how many people's time can one person waste?' Physchim62 (talk) 10:03, 12 September 2009 (UTC)
 * Sorry. Yes it is amazing. Martin Hogbin (talk) 16:44, 12 September 2009 (UTC)

Oppose splitting article as suggested above. No other encyclopedia has 2 articles on the speed of light, for good reasons, and neither should Wikipedia. —Finell (Talk) 09:34, 12 September 2009 (UTC) I can just see the dab header now:

Physchim62 (talk) 10:01, 12 September 2009 (UTC)


 * We should all be able to agree that light does not travel at c in media such as air and water etc. So do we want one article or two articles ? Keeping one article makes for a very long article. Charvest (talk) 10:38, 12 September 2009 (UTC)
 * I've no problem with the fact that light doesn't travel at c0 in material media, don't worry! But I don't think splitting the article is a solution to the problem you pose. It the article is too long, there are plenty of other solutions which we could take. The section on "Faster-than-light observations and experiments" has always been one of my bug-bears. It is mostly typical Wikipedia: "nothing can travel faster than the speed of light, so we're going to discuss things that travel 'faster than the speed of light' to end up saying that nothing travels faster than the speed of light." (fortunately the section is now less than half the length it used to be)
 * But if we split the article as you proposed, then we're separating concepts that really ought to be together. Is it coincidence that the speed of light in a vacuum is exactly the same (as far as we can experimentally determine) as the limiting speed in SR? Is it also coincidence that this speed coincides with the disappearance of frequency dependence (dispersion)? Mainstream physics would say no. How do we discuss the limiting speed in SR without pointing out that it is the same as the speed of light in a vacuum? Are we supposed to say that, well, there's this limiting speed which nothing can exceed, we give it the symbol c and it's about 300,000 km/s, and then just leave it at that? Physchim62 (talk) 12:18, 12 September 2009 (UTC)


 * Ok, strong opinions to keep one article and noone defending two articles - I guess the other one will be soon deleted. Charvest (talk) 16:24, 12 September 2009 (UTC)

A = BC
Continuation from above subsection


 * The question is fairly simple, and is not different than "what to plug for C in A = BC" if C = 2. The answer is 2. There's no sweeping under the rug, whatever uncertainty there was in c before is now contained in both E and m. Headbomb {{{sup|ταλκ}}κοντριβς – WP Physics} 17:22, 10 September 2009 (UTC)

I didn't follow the A=BC example? Brews ohare (talk) 17:47, 10 September 2009 (UTC)


 * First put c = 1 and then decide to use incompatible units for A and B which leads to a conversion factor c different from 1 in your equations. Then think about what would happen have happened had people measured A and B separately using different experiments in different units from the start. Then c would have had an uncertainty, but what does that really mean? Physically, this is equivalent to measuring e.g. the same length twice (e.g. using different experimental techinques and expressed in different units). The ratio will have an experimental error, even though it is some fixed constant. Count Iblis (talk) 18:13, 10 September 2009 (UTC)


 * This is interesting. So the equation in more recognizable notation is ℓ = c t. The suggestion is that we measure time of transit t and use fringe counts to get ℓ. Then c = ℓ / t = so many fringe counts / s ± ε, with ε = measurement error. Now we know c has an exact value, but here it has been measured only to within measurement error.
 * I hope I;'m following your discussion up to this point. Am I?


 * Now what is next?
 * It seems to me that the real meat of the 1983 decision is that pre-1983 lengths were compared by comparing fringe counts. Post-1983 they are compared using times-of-flight. Because we all are willing to accept that c is constant of nature whatever else it is, we suppose that these two comparisons are entirely equivalent, except the time measurement provides a more accurate comparison. So we adopt the 1983 definition.
 * My view of this development is that in fact it has zero to do with what the numerical value of the speed of light really is, or how accurately we can measure it. It has to do only with fringe counting being a less accurate basis for length comparisons than times-of-transit. Do you agree with some of this?? Brews ohare (talk) 18:35, 10 September 2009 (UTC)


 * Yes, if you have some experimental technique for measuring lengths (that is normalized in some arbitrary wayway) and and a different experimental techique for measuring time (which is normalized in some other arbitrary way) then the speed of light will be some arbitry number. You can say that the value of the speed has some inherent physical meansing, but then that meaning depends also on the ratio of the lengths and the time intervals that are numerically equal to 1 in these units. Count Iblis (talk) 23:59, 10 September 2009 (UTC)

So, more specifically, I'd say that the pre-1983 approach counted fringes to get lengths (the metre was so many fringes), and that this method is different from a time-of-transit measurement, in your sense of different experimental techniques for length and time. Then c = # of fringes /s is exactly what you mean by a speed that depends upon the units: what wavelength you picked for the basic fringe count and what unit you picked for the second. That is my understanding of how 299 792 458 ± 1.2 m/s was arrived at. Would you agree to that, I wonder? Brews ohare (talk) 00:34, 11 September 2009 (UTC)
 * Yes, I agree. Count Iblis (talk) 02:49, 11 September 2009 (UTC)


 * Count Iblis, thanks for your patience. To shift topics a bit, comparing length by comparing fringe counts is a different experimental technique than comparing lengths by comparing transit times, although the two comparisons should provide the same result. For example, if one comparison says the large length is twice the smaller, so should the other.


 * However, as counting fringes is more error prone than measuring transit times, the comparison based upon fringes is more error prone than that based upon transit times. I take that as the reasoning described by BIPM that led to the switch to transit-time comparisons in 1983. Would you agree with me on that one?


 * I hope you do, and that because of this, you also agree that the switch in 1983 is not based upon accuracy in measuring the speed of light, which actually has nothing to do with the switch at all. The switch is all about transit-time comparisons being more accurate than fringe comparisons. What do you say? Brews ohare (talk) 05:41, 11 September 2009 (UTC)


 * Yes, I think this is correct. Count Iblis (talk) 14:49, 12 September 2009 (UTC)
 * Yes and no. You're right that the change in the definition of the metre was simply so that lengths could be measured more accurately. You're also right that the time measurement doesn't pose a practical problem. Since at least 1973, by far the largest part of the uncertainty in the measured speed of light was the uncertainty in how long a metre actually was: the uncertainty in the "time" part of the measurement was orders of magnitude lower, and hence negligeable. If you can't measure length to more than 4 parts per billion, then you can't measure speed to more than 4 parts per billion, I think we agree on that.
 * To increase the accuracy in length measurements required a change in the wavelength used to (practically) define the metre. But the wavelength had only been chosen in 1960, were people going to have to change the definition every ten or fifteen years or so? How would we know in the future that one wavelength was better than another? The solution was to turn the equation round: instead of measuring speed in terms of an arbitrary fixed wavelength and a precise frequency, it was decided to measure wavelength in terms of an arbitrary fixed speed and a precise frequency. The practical problem is still to measure length, but now you can do it with lots of different light sources and compare the results. Physchim62 (talk) 15:57, 12 September 2009 (UTC)


 * I got the "yes" part. What is your "no" part? Your discussion describes the evolution of length measurement and how it caused the change to time-of-transit as a basis for length comparison. I think we're in agreement on that. The next step is the definition of length by the BIPM as ℓ = 299 792 458 m/s × t. I imagine we will agree that that is what BIPM did. With that in front of us we can put in the time of transit for the metre as 1/299 792 458 s, and recover the metre. Moreover, we can solve ℓ = c t for c and get c = 299 792 458 m/s for any choice of ℓ and its corresponding transit time t. Do we disagree about any of this??? Brews ohare (talk) 21:57, 12 September 2009 (UTC)
 * My hesitation is that high-precision measurements of lengths in the order of one metre are not usually done by measuring a transit time for that length. They're still done by counting fringes. so you get ℓ = nλ, n is your wavelength count. You then measure a "transit time", but for one wavelength, that is the inverse of the frequency of the light. The transit time for the length you trying to measure, the t in your comments, is measured as n/f.
 * Before 1983, you had c = λf as by far the most precise measurement of the speed of light, ℓ = nλ (with λ fixed) as the definition of the metre. The limit of precision in the value of the speed of light was how well you could compare the wavelength of the test frequency with the wavelength of the krypton transition chosen as the standard. Since 1983, you have ℓ = cn/f as the lab-scale measurement of length with c as a defined value; the precision in the measurement of length is limited by the measurement of f, not because it is difficult to measure time but because the sources are not perfectly monochromatic. A practical measurement of length will also have an uncertainty in the value of n, which may well be considerable, but nothing changed in 1983 in that respect.
 * You can measure length in terms of a transit time over that length, of course, just think of radar. However, the method is not very accurate for lengths which are the same order of magnitude as your test equipment because it's difficult to know whether you've eliminated your systematic errors. You could also measure transit time over a longish terrestrial distance that you've calibrated against some other standard – that's effectively the Michelson–Morley experiment.
 * You might also be interested in and, which are readable introductions to some of these points. Physchim62 (talk) 23:03, 12 September 2009 (UTC)


 * Physchim62: I take your point that the practical determination of length may vary depending upon how long a length you are looking at, and possibly some other factors as well. However, the main discussion here, in my view, is what the underlying definitions are about, and I guess we agree upon that. I wonder if you would go so far as to say the adoption of the time-of-transit definition of the metre in preference to the fringe-count definition used before is simply a matter of accuracy in length comparisons, and whatever the actual speed of light might be, it really doesn't enter into these accuracy considerations, which are all about lengths, not speeds? Brews ohare (talk) 23:32, 12 September 2009 (UTC)
 * "all about lengths, not speeds"? Are you sure that you can separate the two as neatly as you seem to be suggesting? Let me give you an example. You have just come up with a brand new measurement technique that implicitly relies on the speed of light being constant: let's call it "interferometry", and not just for the sake of argument ;) You want to calibrate your new technique, so you go and use it to measure the national length standard, which at the time is a bronze bar. Fair enough, you find that the bronze bar is a certain number of wavelengths long, and so you have the conversion factor between your new method and older measurements. A few years later, you go back to measure the same bronze bar, and you find that the conversion factor is very slightly lower: there are fewer wavelengths to the length of the bar. What has happened between the two measurements? Has the bar gotten shorter or has the speed of light increased?
 * The answer is that the bronze bars, which formed the standards for length measurements throughout the English-speaking world, were very slowly getting shorter. Fortunately, the platinum–iridium bar used in the metric system didn't seem to have this problem (or at least, the problem wasn't large enough to measure), which led the United States to switch to metric-based standards in… 1893.
 * The British were rather slower to make the legal change and so, throughout the late nineteenth and early twentieth centuries, the value of the speed of light in Imperial yards per second was slowly but surely increasing. Did this mean that the speed of light was increasing, but only in the British Empire? Would have made for some interesting refraction effects at the borders!
 * Of course, the speed of light that was being measured was the same, whether it was measured in London, Paris or Washington, DC. There was no "Imperial speed of light" and "metric speed of light", simply a physical constant that had different values in different systems of units. Similarly, today, there is no "actual speed of light" vs. a "BIPM speed of light".
 * But, you say, the "actual speed of light" must be measured! Fine, go find me a length standard. Now measure the length in metres of your chosen length standard using interferometry: you will find you have a measurement uncertainty. This uncertainty is exactly the same as the measurement uncertainty in the speed of light relative to your chosen length standard… Physchim62 (talk) 11:01, 13 September 2009 (UTC)

Physchim62: We seem to have become disrailed. I did not say the "actual speed of light must be measured". I said: "whatever the actual speed of light might be, it really doesn't enter into these accuracy considerations, which are all about lengths, not speeds." In other words, the actual speed of light does not have to measured; its numerical value is irrelevant to the decision to switch from fringe count measures of length to time-of-transit measures of length. The decision to switch definitions is based upon fringe-count measurements being less accurate than time-of-transit measurements. Do you agree with me on this? Brews ohare (talk) 15:12, 13 September 2009 (UTC)
 * The problem is that you are still speaking of the "actual speed of light" as it were not the same as the speed of light in the 1983 definition of the metre. In which case, what you speak of as the "actual speed of light" is a very different beast from the one the rest of us are trying to describe. Why I speak of the speed of light I'm talking about the one that was measured by Rømer, Fizeau and Michaelson; also about the one that enters into the Lorentz transformations and E =mc2, at least to within a difference which has completely escaped measurement so far, despite many efforts; and also the one used to define the SI metre since 1983. What are you talking about when you refer to the "actual speed of light"? Physchim62 (talk) 17:08, 13 September 2009 (UTC)


 * I think I refer to the same "actual speed of light" that you mention in connection with Fizeau and with the Lorentz transformation. I wish to distinguish that from the integer number 299,792,458 m/s that occurs in the SI units. This number is used to define lengths in terms of transit times. It also is called "the speed of light" but has a somewhat different meaning than the actual speed of light, because in fact, this number could be chosen to be anything the CIPM committee wanted it to be, unlike the actual speed of light that is a constant independent of man's definitions. This ability to make 299,792,458 m/s any number the CIPM wanted is made possible because the metre changes to adapt to whatever number they might pick. I think you know all that. So the point here is just that there is this difference, and I believe baldly stating without further comment that 299,792,458 m/s is "the exact" speed of light violates WP:Astonish. A tempest in a tea pot, I'd say. Do you agree?


 * The other point is that the switch from fringe comparisons to time-of-transit comparisons for length has nothing to do with either the actual speed of light or the number 299,792,458 m/s. It has to do with the greater accuracy of the time-of-transit comparisons. Again, a simple point, eh? Brews ohare (talk) 05:12, 14 September 2009 (UTC)


 * Brews, I think I've asked a few times for sources, and that I have stated that you have provided any, that promote this point of view that the measurable speed of light is not the same thing as the speed of light referred to in the 1983 definition. I think this is just a mental juggling trick that you've made up; if I'm wrong, show us the source that says these two things are different.  Dicklyon (talk) 05:39, 14 September 2009 (UTC)

Dicklyon: Let me begin at the beginning this time around. Is it necessary to persuade you of these points personally, or is this just you with your official WP hat on asking for sources? If the former, arguments from sourced precepts should suffice. If the latter, a contribution to the main page is under consideration, a situation that I very much find unlikely. Brews ohare (talk) 13:23, 14 September 2009 (UTC)

Another argument
Given the known laws of physics, you can pretend that time intervals and spatial lengths are physically of the same nature (in the same eay that lengths in the x and y direction are physically the same). That is how you can interpret the known laws of physics, regardless of if it is actually correct or not. No experiment can prove this interpretation wrong if such an experiment does not also contradict the known laws of physics.

Compare this to discussons about the Many World interpretation of quantum mechanics. Barring some exotic thought experiments involving artificial intelligence implemented by a quantum computer as an observer, MWI will yield the same predictions as the Copenhagen Interpretation, so you cannot point to some experimental result and then argue that the wavefunction does in fact collapse (unless that experient is in conflict with some basic postulate of quantum mechanics itself).

This then closes the argument about c as the space-time constant. The fact that c is also the speed of light follows from the validity of the theories that describe light as an electromagnetic wave (Maxwell's equations). Count Iblis (talk) 15:35, 13 September 2009 (UTC)


 * Count Iblis: I am unsure which argument you are trying to settle. The argument I've been engaged in is really inseparable from the SI Units, because the questions are "What is the role of the number 299,792,458 m/s called the "speed of light" in the SI units?" and "How does this number relate to the fact that light does travel at some speed, regardless of the units you use to describe it?" Brews ohare (talk) 16:36, 13 September 2009 (UTC)


 * I think to fully understand my argument, it would be helpful to derive all the equations in which c appears, including the correct classical limit of special relativity, starting from the equations in which c has been set to 1. I have done this on another forum, but I made a few small mistakes there. If I have the time, I'll rewrite the correct argment here. I.m.o., this is the only correct derivation of classical mechanics from special relativity, the (simple) derivation given in most textbooks is misleading.


 * You can then see that the role of c is that of a scaling constant. Given some model, you can study some (singular) limit in which one variable is very small or very large by rescaling the variables in some way. Then, in the limit that the rescaling constant tends to infinity, you can get new independent variables that did not exist (as independent) variables in the original model, because of the singular mature of the limiting case (you lose relations between variables).


 * Then, in the classical limit, you have independent quantities (e.g. mass and energy, or space and time) that were not independent if c is still finite. But because of the way physics has progressed historically, energy and mass or space and time are still considered to be physically inequivalent quantities, despite the fact that we know that c is not infinite. We take into account the view that the different quantities are physically distict, by assinging different dimensions to the physical quantities, causing c to get dimensions as well. But this means that the SI unit for c is still consistent with

c = 1. Equating 299,792,458 metre/second to 1 yields that one second = 299,792,458 metres. This thus indicates by how much the unit of time is rescaled relative to the unit of space in SI units. Count Iblis (talk) 18:22, 13 September 2009 (UTC)


 * Note that what I wrote in the last paragraph did not involve the actual SI definition of the metre in terms of the speed of light ad the second. So, if the metre were still defined by the length of some bar stored in Paris, you would still have that one second is speed of light times the length of onemeter. The only difference would be that we would not know precisely the value of the speed of light expressed in these units. Count Iblis (talk) 18:28, 13 September 2009 (UTC)

Count Iblis: I may be a bit dense here (not deliberately so, please). I understand that c t is equivalent to a length in SR, and one could use x4 = c t just like x1 x2 x3. I don't see how that helps me understand a switch from fringe counts to transit times for length comparisons. They involve different physical measurements, and it is exactly this difference, which is accompanied by different error bars, that led to the 1983 decision. It is this last statement that seems to bring out crazy allegations about fringe science, rewriting history, opposition to all of physics, and so forth. Brews ohare (talk) 17:27, 14 September 2009 (UTC)

Meter defined in terms of the speed of light
The section Meter defined in terms of the speed of light is incorrectly named and contradicts both its contents and the underlying BIPM documents. The BIPM SI Units brochure § 2.1.1.1, p. 112 clearly defines the metre in terms of the transit time of light as the length traversed in 1/299 792 458 of a second. It then says "It follows that the speed of light in 'vacuum' is c0 = 299 792 458 m/s" (Italics mine).

In short, the value of c0 does not define the metre, but the reverse is true. In SI units, the metre is the fundamental unit.
 * Indeed it is not the value of c0 that defines the metre. It is the physical magnitude of this dimensionfull quantity that defines the metre. Just like it is the physical magnitude of the mass of the international prototype kilo that defines the kilo. It also follows from the definition of the kilo that the mass of the international prototype kilo is exactly 1 kg. Also talking about fundamental units after the 1983 definition the (relevant) fundamental units of the SI are actual the second and the 'metre per second', the metre being derived from these two. (Hopefully putting it this way, you will finally understand.) (TimothyRias (talk) 15:03, 11 September 2009 (UTC))


 * I have your assertions on this point, but no sources. On the other hand, my statements about transit time as the fundamental measurement are sourced, and appear to contradict your remarks. Can you source your statements?


 * At a logical level, of course the real physical speed of light determines the metre, inasmuch as the metre is how far light actually travels in 1/299 792 458 s . However, this definition says nothing about what the numerical magnitude of the real physical speed of light might be, and actually it doesn't matter. All that does matter is what transit time you set to decide how big the metre is: it is a transit time issue. Brews ohare (talk) 15:15, 11 September 2009 (UTC)

It's pretty simple math to see that if one metre is the distance traveled by light in 1/299 792 458 s, then the speed of light follows as 299 792 458 m/s.

The key to the 1983 definition is the introduction of transit times as a basis for length comparisons. It is transit times that matter and that is why the transit time 1/299 792 458 s is the basic starting point. It is transit times that can be measured accurately, not the speed of light. See Sydenham "Time measurements are more reproducible (parts in 1014 uncertainty) than length (parts in 109 uncertainty)" Of course, the error in length is not reduced all the way down to the error in time because other uncertainties enter, among them the accuracy of realizing the 'vacuum'. Brews ohare (talk) 14:19, 11 September 2009 (UTC)
 * You can define the inch as 2.54 cm and the yard as 36 inches, or you can define the yard as 91.44 cm and the inch as 1/36 of a yard. What on Earth would be the difference between the two? -- _ _ _ A. di M. 14:45, 11 September 2009 (UTC)

I believe your point is that if A = 1/B, then B = 1/A? The difference here is that it is transit time that is measured, not speed of light. The number 299 792 458 m/s is a defined, not a measured, value. The metre is the standard that is realized. Brews ohare (talk) 14:49, 11 September 2009 (UTC)


 * That's right. But it still doesn't mean that c0 is defined that way, any more than "the temperature of the triple point of water" is defined as "273.16 K"; its numerical value (the number 299,792,458) is defined that way. -- _ _ _ A. di M. 15:03, 11 September 2009 (UTC)


 * OK, pardon me if I am a bit slow here. I agree that the real physical speed of light has some actual value and that value can be expressed in various units as various actual numerical numbers of those units; e.g. wavelengths/s for some transition. If you choose a different wavelength, or a different second, you get a different speed of light. But I don't think that is what this discussion is about, do you? Brews ohare (talk) 15:21, 11 September 2009 (UTC)

The real question here is, How much of how many people's time can one person waste?'. Martin Hogbin (talk) 20:55, 11 September 2009 (UTC)


 * Another real question here is whether 'tis nobler to simply opt out of discussions that don't interest you, or to make catcalls that interrupt ongoing discussion between parties that wish to engage? Brews ohare (talk) 15:37, 13 September 2009 (UTC)

Measuring
Some musings based on reading Brews, David's and other's posts:

1) Measurement of a quantity means relating it to something else, but more than that, the definition of a quantity is a comparison with something else. The definition of speed is a relation between distance and time. Measurement of speed means comparing it with a ruler and a clock.

2) The speed of light is postulated to be constant therefore it is assumed to always takes the same time to travel a given distance, or equivalently it always travels the same distance in a given time, therefore an ideal ruler can be defined as the distance traveled by light in a given time.

3) Actual rulers manufactured according to this definition will differ due to the accuracy of the manufacturing equipment and the accuracy of the clock. The numerical value of the speed of light given in these ideal rulers is a defined value but using one of these manufactured rulers to measure the speed of light will give a relation between the speed of light and the manufactured ruler. Since the manufactured ruler will have limited accuracy it may well give a different value than the defined value. This measured value will be a relation between the speed of light and the manufactured ruler. Since the speed of light is postulated to be constant, any discrepancy between the defined value and the measured value will be attributed to the inaccuracy of the ruler, i.e. the difference between the defined value and the measured value will tell you the difference between the ideal ruler and your actual ruler - which is a measurement of the actual ruler in terms of the ideal ruler or equivalently a measurement of the ideal ruler in terms of the actual ruler.

4) Suppose you want to measure the speed of light (either because it's hundreds of years ago and you've no idea how fast it is or that it is constant, or you don't really believe it is constant), then you must compare the light with some ruler and some clock and you will get some idea of its speed subject to inaccuracies in your ruler and clock. The act of measuring the speed of light in this way is a different concept from the act of defining the speed of light, i.e. the act of comparing the speed of light to something else is a different concept to the act of not comparing the speed of light to something else. However this doesn't matter because the speed of light is postulated to be constant therefore you use the defined value to define a ruler with which to measure the distances and speeds of everything else - so everything else is ultimately measured in relation to the speed of light.

5) The equation relating c,ε0 and μ0 may have originally been discovered as a result of experiment, but assuming the equation is true then whatever value and units are given for two of the quantities, the third quantity is fully determined. A=B implies B=A so an equation can be read either way, assuming the equation is true, regardless of how the equation was originally discovered. Charvest (talk) 18:14, 12 September 2009 (UTC)


 * I think there's another point that you miss from your list: that's that nobody is forcing you to use SI units. I'll assume that we're talking about scientific research here and not everyday commerce. Imagine, for example, that you wanted to see if the speed of light varies with frequency. You can't use SI units, because the definition of the SI metre assumes that the speed of light doesn't vary with frequency. That doesn't stop you from using some other length standard, for example an old standard metre bar: you then measure the speed of light at different wavelengths against your chosen length standard, and see if they differ. Similarly if you want to see if the speed of light changes over time: you can't use the SI metre, but you could use some other length standard which you don't think has changed over time, such as some function of the mass of the earth, the Newtonian gravitational constant and a time standard. The way that the metre is defined doesn't stop you from doing these experiments, nor does it mean that they are not physically worthwhile. But it doesn't mean that the speed of light isn't 299,792,458 m/s either. Physchim62 (talk) 19:19, 12 September 2009 (UTC)


 * As long as the new "perfect" definition results are within  the upper and lower limits of the old  "perfect "  results  there  is no problem for  everybody  having to use the result of all these (nonexistent )difficulties.Wdl1961 (talk) 00:46, 13 September 2009 (UTC)

I'd say everybody is a bit right here. For example, Physchim62 points out that you don't need to use SI units. An example might be to go back to the pre-1983 SI units based upon fringe counts of wavelengths. Then you can measure the speed of light in these wavelength-based units, and of course, as with all measurements an error bar of observation will arise (c = 299,792,458 ± 1.2 m/s). Where I think some problems with language show up is in connecting such a measured speed of light with the use of the term "speed of light" to describe the defined value 299,792,458 m/s in today's SI units. These usages are separate. There can be no argument that the speed of light in today's SI units "is" 299,792,458 m/s, but what is its relation to measurement? As the BIPM and others point out, measurement uncertainty has been transferred to the metre itself. Thus, the number 299,792,458 m/s is exact, but the unit m/s is not known. It is the unit that is the experimental quantity now. Brews ohare (talk) 15:52, 13 September 2009 (UTC)


 * Speed in natural units is expressed as a fraction of lightspeed so:
 * speed of light = 1 x speed of light
 * also speed of light = 299,792,458 m/s
 * therefore 1 = 299,792,458 m/s
 * or equivalently 1 = 299,792,458 x 1 m/s
 * i.e. the number 299,792,458 is a conversion factor between natural units and m/s
 * turning the equation around: 1 m/s = 1/299,792,458 in natural units, that is as a fraction of lightspeed.


 * The use of the term speed of light for a defined value is basically the same thing as using the term speed of light for the natural unit 1 in which we simply relate the speed of light to itself.
 * Measuring the speed of light means relating it to something else.
 * So the different uses of the phrase that you talk about are either 1) relating the speed of light to itself and then relating other speeds to light, or 2) relating it to something else straight away. Charvest (talk) 16:16, 13 September 2009 (UTC)


 * Brews, what you say is right, but the unit has always been an experimental quantity, not just "now". There's no fundamental difference between using the speed of light, the wavelength of a particular atomic transition, or the distance between two tacks on a particular piece of metal in Paris. In any case, you would know exactly the value of said speed/wavelength/distance in your unit by definition, but you could have uncertainties in the measurement of said speed/wavelength/distance which essentially become uncertainties in the unit. -- _ _ _ A. di M. 16:40, 13 September 2009 (UTC)


 * Charvest: I agree that today's SI Units use the speed of light as a unit of speed. So for example, the speed of sound can be expressed as a multiple of the speed of light. Moreover, this multiple depends in no way upon knowing the numerical value of the speed of light. So in that sense, the number 299,792,458 m/s is simply an artifact, or as Jespersen says, a defined and arbitrary value. I think we agree about that. Brews ohare (talk) 16:47, 13 September 2009 (UTC)


 * User:A. di M.: I think your description muddles me. I think I can agree that there is no difference in kind between the metal bar and counting fringes: they both are length measurements. However "the speed of light" is a speed, not a length. It can be related to a length measurement by introducing a transit time, which most probably you would agree. So then, is a transit time measure of length the same in kind as a wavelength determination of length? I'd say not, for this reason: When length is determined using a length measurement and time is determined using a time measurement, then speed can be determined as the ratio of these measurements. However, when length is determined as a time-of-transit measurement and related to length by a defined constant with the dimensions of speed, it no longer is possible to measure speed as length / time, because only the defined conversion factor can result, and it provides no physical information, only the defined value, which is arbitrary. Brews ohare (talk) 16:56, 13 September 2009 (UTC)
 * So, according to you, when the litre was defined as the volume occupied by one kilogram of water in such-and-such conditions, a measurement of the density of water could "provide no physical information"? -- _ _ _ A. di M. 17:06, 13 September 2009 (UTC)


 * I am unfamiliar with your example. You seem to suggest, however, that saying the litre = volume occupied by 1 kg. of water → the density of water is 1kg/ litre, appears to be inescapable and of no content. It looks that way. If one measured the dimensions of a litre of water as so many cubic metres, then the density in kg./m3 would have a meaning. Thus, there appears to be a parallel between time-of-transit length cf. wavelength length compared to density in kg/litre cf. density as kg./m3 Brews ohare (talk) 17:12, 13 September 2009 (UTC)
 * So what would you propose as a "non-arbitrary" standard for speed, that would give the "physical information" you're looking for? Physchim62 (talk) 17:30, 13 September 2009 (UTC)
 * I haven't raised that question, which is a different matter. I simply wished to clarify that the number 299,792,458 m/s has no physical content within the new SI units. It's a hang-over from the pre-1983 units, where it actually was a speed measurement, and was chosen for the new SI Units only to minimize dislocation with prevailing practice. I also wished to point out that the experimental error bar in the new SI Units is now in the unit m/s, having been transferred there by introduction of the now defined value of the speed of light. As the BIPM points out, any advance in the precision of measurement changes the metre, not the number 299,792,458 m/s. Brews ohare (talk) 17:39, 13 September 2009 (UTC)

[redent] Erm, no, any advance in the precision of the measurement gives us a more precise metre, which is not the same as changing it. The problem is that we would all really like to believe that you simply don't understand what a unit of measurement is: so, please, give us you proposition for a standard against which to measure the speed of light which would give you the "physical information" you're looking for, instead of just saying that everything is FUBAR since 1983 and nobody else has noticed. Because there, you're making a big claim against pretty much the whole of physics: it might just be that it's you who has the misunderstanding. Physchim62 (talk) 18:00, 13 September 2009 (UTC)


 * Physchim62: Roger, the metre becomes more precise, and the number 299,792,458 m/s is unaffected, as measurement precision improves. Your summary of my position "saying that everything is FUBAR since 1983 and nobody else has noticed" is an invention of your construction, as I have never said, suggested, or thought anything of this kind. Likewise, "you're making a big claim against pretty much the whole of physics" is completely incorrect: please explain where this crazy notion comes from. Brews ohare (talk) 18:04, 13 September 2009 (UTC)

Charvest, First of all, I'm glad that you have actually acknowledged that the equation relating c,ε0 and μ0 was originally discovered as a result of experiment. I got reported at AN/I for disruptive behaviour for bringing that matter up at WT:PHYS. That equation has got nothing to do with the measured speed of light. It arises exclusively from the ratio between the electromagnetic units of charge and the electrostatic units of charge. That ratio will always exist. There is no system of units that can get rid of that ratio out of Maxwell's equations. If we have a defined speed of light such as in the new SI system, or in the system in which we define 'c' to be equal to 1, we cannot put it into that equation. The only thing that we can do is draw attention to the closeness in value between the defined speed of light and the value that arises from the experimentally measured values in this equation.

Now let's not lose sight of what the main argument is here. The main argument is not even related to what I have just said above. The main argument is about the fact that the measured speed of light is used to define the new metre. It then follows that if we express the speed of light in terms of that new metre that is defined in terms of the speed of light, then we merely end up with an arbitrarily defined number. This defined number is beyond measurement, and it is a different concept to the measured speed of light that was used to define that metre in the first place. In SI units, the speed of light then becomes 299,792,458 times the distance that light travels in 1/299,792,458 seconds, per second. We could have chosen any number. The physical speed of light as a concept cannot therefore be sacrificed in the article for a system of units. The article introduction must clearly explain both concepts. David Tombe (talk) 18:53, 13 September 2009 (UTC)
 * "The only thing that we can do is draw attention to the closeness in value between the defined speed of light and the value that arises from the experimentally measured values in this equation." – which is all Weber and Kohlrausch could do, until there was sufficient theory to show that the relation will always hold. Some fifty of so years later, the Weber–Kohlrausch experiment had been turned round (new measurements, of course, by Rosa and Dorsey at the U.S. National Bureau of Standards) to provide a measure of the speed of light. These days, it's more likely to be used as a measure of capacitance or as a lab demonstration.
 * As for the "physical speed of light as a concept" being "sacrificed" for a system of units, the "speed of light an a vacuum" is exactly what it says it is: we could always link speed, light and vacuum if there was any risk of confusion. There should be no apology for quoting its value in the systems of units used by the overwhelming majority of our readers. Or perhaps you believe that the "speed of light" is something completely different? Physchim62 (talk) 19:31, 13 September 2009 (UTC)


 * Physchim62: As you say: "There should be no apology for quoting its value in the systems of units used by the overwhelming majority of our readers." Undoubtedly so, provided the context is provided explaining the switch from length measurement to time-of-transit measurement, which is a departure from the approach used for many centuries prior to 1983. Brews ohare (talk) 01:22, 14 September 2009 (UTC)

Physchim62: You have not responded to me as to the origin of the ridiculous statements you attribute to me here. Brews ohare (talk) 01:09, 14 September 2009 (UTC)

Physchim62, Nothing has changed as regards the Weber/Kohlrausch experiment. Maxwell's work in 1861 demonstrated a convergence of two measured results. There was the direct measurement of the speed of light by Fizeau and there was the electromagnetic/electrostatic ratio as measured by Weber and Kohlrausch. From that convergence of measured results, Maxwell was able to demonstrate that light is an electromagnetic wave. Nothing has changed to this day. The equation c^2 = 1/(εμ) always has and always will read from right to left. It is the equation which links the speed of light to the measured value of ε. Neither the measured value of the speed of light nor the defined value of the speed of light should be used in that equation. If we use that equation from left to right, we are cooking the books with the benefit of hindsight. In maths, equations may work in both directions, but you as a chemist should know that they don't necessarily work in both directions in chemistry. Likewise in physics. There are issues of cause and effect to be considered as well as the physical scenario that is being described. David Tombe (talk) 05:15, 14 September 2009 (UTC)
 * When you speak of "cause and effect", you're getting onto philosophical ground about the "nature of science". This is usually described as the "philosophy of science", although some prefer the term "sociology of science" (and, personally, I'd say they're not entirely wrong, but who am I to judge).
 * That aside, when you say: "The equation c^2 = 1/(εμ) always has and always will read from right to left." you are making a claim that is evidently false. Weber and Kohlrausch suggested it, in reading from right to left. Maxwell proposed a theory (which has been largely supported by experiment, at least in its descriptive value) that the relation will be true whichever way you read it. Rosa and Dorsay (1907) read the equation from left to right to determine the speed of light. Physchim62 (talk) 12:03, 14 September 2009 (UTC)

No Physchim62, That equation links the measured value of electric permittivity to a number that is very close to the measured (or defined) speed of light. That is all there is to it. We can certainly use it in reverse, with the benefit of hindsight, as a lazy way of obtaining a practical working value for electric permittivity. But in doing so, we are cooking the books and working against the spirit of the equation. Where it becomes really ridiculous is when we use the defined value of the speed of light to obtain a defined value of electric permittivity, and then purge the discharging capacitor experiment from the texbooks. David Tombe (talk) 12:14, 14 September 2009 (UTC)


 * David, when you invoke the discharging capacitor experiment, you also need to explain how you want to define the unit for electric charge and electric potential. Count Iblis (talk) 13:34, 14 September 2009 (UTC)

Count Iblis, When the experiment was originally done in 1856, it involved two distinct units of charge. There was an electromagnetic system of charge and an electrostatic system of charge and the experimental result yielded the ratio of these two units which was related to the measured speed of light. That puts the speed of light firmly into Maxwell's equations, irrespective of what system of units we use. David Tombe (talk) 13:43, 14 September 2009 (UTC)

Planck units" some phycisist use length/time dimension
It seems that if we do not use "Planck units" some phycisist use length/time dimension. Rather than an edit war can we get an educational discussion about this? (see art. history).Wdl1961 (talk) 15:40, 23 September 2009 (UTC)


 * I think this deserves a separate section in the article. I'm ok with removing "dimesionless" in the table (not mentioning this doesn't mean the opposite POV is taken). If I have time I'll start the new section later today. Count Iblis (talk) 15:49, 23 September 2009 (UTC)
 * Fair enough, I'll wait and see what you propose ;) I agree that the "spacetime dimensionality" needs to be mentioned somewhere, but I would also put my voice towards the opposite site, which is that most readers won't care about spacetime, and will only want to know about the classical approximation. For me that means something a brief as possible while remaining correct, and good links to other articles. How does that sound to you? Physchim62 (talk) 21:41, 24 September 2009 (UTC)


 * Assuming that "'spacetime dimensionality needs to be mentioned somewhere", is it clear that it needs to be mentioned in this article? Finell (Talk) 21:50, 24 September 2009 (UTC)


 * What I mean is dimensions in the context of unit systems. Count Iblis (talk) 14:23, 28 September 2009 (UTC)

Clarifying the distinction between the two concepts of the speed of light
This prolonged dispute has come about because of attempts to deny an important point that has been raised by Brews ohare. The matter has now gone to arbitration and the arbitrators will now be watching this page carefully. I think that it's only fair to the arbitrators, most of whom are probably not physicists, to make an attempt to explain to them, and eveybody else here, exactly what the distinction is that Brews has brought to our attention.

Everybody, whether a physicist or not, is familiar with the concept of the speed of light. It is the speed that light travels at, and it is generally known to be extremely fast and unreachable by any existing technology. Now let's imagine that I went unto a stage to give a speech on the speed of light. Imagine that I went unto a stage in front of 10,000 people and said that I am going to tell you all what the speed of light is. And then imagine that I stated "The speed of light is the speed of light". And with the speech ending at that, a loud clapping and stamping of feet erupts and lasts for the next twenty minutes. That sounds like a pretty ridiculous scenario. But in fact it is no more ridiculous than if I went unto the stage and stated the speed of light in modern SI units. If I were to go unto the stage and announce the speed of light in modern SI units, I would be stating "The speed of light is 299,792,458 times the distance that light travels in 1/299,792,458 seconds, every second". I could then expect the twenty minute clapping session to be no less sarcastic for me having just stated the obvious.

Brews has pointed out that it is not satisfactory to state the speed of light in modern SI units without some kind of extended elaboration, because the metre itself is defined in terms of the speed of light. Hence any statement of the speed of light in terms of that metre is merely a statement of the speed of light in terms of itself.

Now if we were to already accept the old classical concepts of length, I could go unto the stage and tell the crowd of 10,000 that I had performed an experiment to measure the speed of light using a Michelson interferometer on top of Mount Wilson, California. I could announce, that after performing some difficult calculations that I have found the speed of light to be in the order of 299,792,458 metres per second with an error bar of 0.04%. That would be news worth hearing. I would have given the audience a useful piece of information that had a physical meaning.

It is this latter measurememnt that Brews and I have been referring to as the physical speed of light that can be measured. It is clearly a different concept from the defined speed of light that I described further up, and which tells us nothing that we don't know already, and which is beyond measurement.

This edit war came about because Martin Hogbin wanted to only include the new SI speed of light in the introduction. His argument was that since the SI system is the internationally established system of units, then it follows that we must exclusively use that system in the introduction. Martin has of course overlooked the fact that in the special case of the speed of light, where one of the staple SI units has itself been defined in terms of the speed of light, then it is not good enough to state the speed of light exclusively in SI units without any kind of elaboration.

Brews on the other hand wanted to make that elaboration for the benefit of the readers. Martin was determined to frustrate Brews in his efforts. A crowd then descended upon the article and tried to accuse Brews of being wrong, and of advocating fringe views and pseudoscience. These allegations against Brews, and also against myself, will simply not stand up even against the mildest standards of probity. David Tombe (talk) 05:42, 14 September 2009 (UTC)


 * As has been explained to David many times, defining the speed of light as 299,792,458 metres per second is not tautological, since measuring the speed of light is equivalent to measuring a metre (i.e. a recalibration of our instruments). It is analogous to saying that a foot is twelve inches.  That doesn't stop us from measuring how long a foot is, which tells us how long an inch is.  Similarly, measuring the speed of light tells us how long a metre is. --Michael C. Price talk 05:58, 14 September 2009 (UTC)


 * When you say "It is clearly a different concept" and use that to support the POV that there are two different concepts called speed of light, you are aligning yourself with Brews, but not with any source that I have seen cited. The lack of citation to a source supporting the point of view is why it can't stay in the article.  The fact that you and Brews push an idiosyncratic point of view is the source of the problem.  Brews has at least shown us which sources he thinks are closest to representing the POV he wants to push, and I for one welcome the representation of the points of view expressed in those sources -- but I don't think any of them said anything about there being two different concepts called speed of light.  If I got that wrong, just give us the source and the quote that contradicts what I just said. Dicklyon (talk) 06:05, 14 September 2009 (UTC)
 * Dick: Are we persuading you personally? If so, sourced precepts and a logical argument should suffice. The whole matter is explained with care at User:Brews ohare/Speed of light (Example). Brews ohare (talk) 13:30, 14 September 2009 (UTC)

Dick, Brews provided the sources, and I am backing Brews up on the point that he has made. The arbitrators can decide on whether or not Brews and I have a legitimate point, or whether we need to be topic banned for having advocated this point of view. David Tombe (talk) 06:32, 14 September 2009 (UTC)


 * David, the sources that Brews cites to justify his POV are typically these: Wheeler; Jespersen; Sydenham.  If you want to support him meaningfully, just show where they support the idea of two different concepts of the speed of light.  The arbitrators are more like to be swayed by whether you argue with reference to sources than by anything about the physics, which it's not their job to understand.  If you keep pushing a POV by insistence, rather than by showing it in sources, you'll just help them see that our complaints about your behavior are well founded.  Dicklyon (talk) 07:17, 14 September 2009 (UTC)
 * But many of the 10,000 people in the audience roughly know how long a metre is (even if they have no idea of how it is officially defined) and how long a second is, in relation with everyday quantities. To such people, telling that the speed of light is 299,792,458 m/s, although tautological for those who do know the definition of the metre, is not useless; they'll know that the light travels roughly 300 million times the distance from their hips to the ground in a time roughly equal to that between two consecutive heartbeats of theirs. And as for the Wikipedia article, per WP:MTAA, WP:NOT PAPERS, WP:LEAD and all that, we should not assume that readers will know how the metre is defined, at least not in the lead section. -- _ _ _ A. di M. 09:39, 14 September 2009 (UTC)

Dick, I clicked on Jespersen and the first thing I saw was "One fall out of this new definition of the metre was that the speed of light is now a defined quantity and no longer a measured quantity". What more do you want? David Tombe (talk) 07:40, 14 September 2009 (UTC)
 * Well, just above, it says "This task [to link optical freqencies with the caesium microwave frequency] has proved to be about as much art as science." This was an art first demonstrated in 1972, and repeated by many laboratories. The value obtained was recommended from 1976, and officially adopted in 1983, after the same procedure had been applied to other light sources and found to give the same result (within experimental error). The fact that the author of an introductory book about the concept of time wishes to distinguish it from science in 1999 (the date of the quoted edition) is of little consequence here. Physchim62 (talk) 12:39, 14 September 2009 (UTC)


 * Physchim62: This book is written by several scientists at NIST. SO they have some authority. Your denigration of sources is the next step in refusing to engage in this discussion. You are simply running a debate, with the normal rules of debate, which are to obfuscate, distort to score points, and entertain with les bons mots. There is no point holding discussion with those ground rules. Brews ohare (talk) 13:14, 14 September 2009 (UTC)

Physchim62, My argument above stands on its own merits irrespective of sources. Brews has given sources for good measure and you are now trying to belittle one of those sources. David Tombe (talk) 13:31, 14 September 2009 (UTC)


 * Dear David,
 * my puny mind needs your mighty intellect's guidance:
 * Is the kilometre
 * just a defined value?
 * just a measured value?
 * both a defined and a measured value?
 * Eagerly awaiting your clarification,
 * --Michael C. Price talk 10:12, 14 September 2009 (UTC)
 * Another smart-alecky comment by Michael C. Price, champion debater and master of the snarky remark. Brews ohare (talk) 13:14, 14 September 2009 (UTC)
 * A question, not a comment. And a question that Brews and David have both avoided answering.  I wonder why? --Michael C. Price talk 14:55, 14 September 2009 (UTC)

A. di M., No. To say that "The speed of light is 299,792,458 times the distance that light travels in 1/299,792,458 seconds, every second" is a meaningless tautology that tells us absolutely nothing about the speed of light. It is no different to saying "The speed of light is k times the distance that light travels in 1/k seconds, every second". The case has been unequivocally proven in Brews's favour along with supporting sources. I suggest that the arbitration committee take note of this and swiftly fold up the case, because there is absolutely nothing more that can be said regarding the dispute. I suggest that Brews ohare is owed a major apology. David Tombe (talk) 12:06, 14 September 2009 (UTC)
 * Except that many readers won't mentally substitute "the distance that light travels in 1/299,792,458 seconds" for "metre"; most of them have an idea of how long the metre is in relation to everyday stuff (e.g. "slightly more than the width of my bed" or something), but no idea of how it is formally defined. -- _ _ _ A. di M. 12:27, 14 September 2009 (UTC)


 * Yes, Which is exactly why it is a less informative concept than the measured speed of light. Do you now agree that this matter needs to be elaborated upon in the introduction? You are turning the argument upside down. You are now saying the same thing as me, but doing so in a manner as if you are disagreeing with me. David Tombe (talk) 12:37, 14 September 2009 (UTC)


 * But what neither Brews nor yourself have been willing to tell us is: "what is this real speed of light?" Can we measure it? The measurement of the speed of light (as we all seem to have agreed on its definition) presents no problem at all, so long as you can provide a length standard that is sufficiently precise. The speed of light is still measured, at inner solar system scales at least, and to admirable precision. All of this after 1983. Physchim62 (talk) 13:00, 14 September 2009 (UTC)

I have attempted real engagement with you on these points time and again. You have many careful explanations above, which you abandon when convergence is approached, and re-open again later. Brews ohare (talk) 13:17, 14 September 2009 (UTC)


 * Physchim62, Of course we can measure the speed of light. But we have to use a system of units other than the modern SI system, because the defined speed of light in SI units is fixed by definition, and therefore cannot be measured. I've stated the argument clearly above and I intend to take that argument to the arbitration committee. It has now become patently clear that you don't understand this issue, yet you have gone to AN/I and successfully persuaded an administrator to ban me from explaining it to you. David Tombe (talk) 13:37, 14 September 2009 (UTC)


 * I agree that this explanation has been offered again and again. Dialog only goes so far, and then the questions simply are repeated days later without reference to explanations provided. The fact is, these matters are extremely simple and straightforward, and the resistance cannot be understood as a failure to grasp the issues. An extensive discussion with sources is found at User:Brews ohare/Speed of light (Example). Brews ohare (talk) 15:58, 14 September 2009 (UTC)

Mathematical model vs Experimental observations
One way of looking at this argument is to say that physics as a whole is a mathematical model of reality consisting of various constants, various variables and various equations relating the constants and variables. The model is deemed useful if it corresponds closely to experimental observations. There is an element of the model called c. The model relates c to various other elements of the model, and experiments can try to invalidate the model by measuring the real world counterparts of elements of the model and seeing whether within statistical error bounds the model reflects the real world. In this respect, the difference between c as part of the model, and real world measurements of the speed of light, is no different from how every part of the model can be contrasted with real world measurements. Does that mean every article about a concept in physics should explain the difference between a model and the real world in the lead paragraph ? There's no reason why the speed of light should be singled out for such treatment. Charvest (talk) 17:16, 14 September 2009 (UTC)


 * Your discussion of model equations involving c is fine. However, it is not related directly to the question of the status of the number 299,792,458 m/s. In the pre-1983 system of units, measurements of the c you talk about were made with the result c = 299,792,458 ± 1.2 m/s. In the post-1983 a speed called the "speed of light" and given by c0 = 299,792,458 m/s exactly is introduced. The connection between c and c0 is the subject of discussion. Because of the switch to times-of-transit for length comparisons, a means to convert such times to lengths was needed. For that conversion the number c0 = 299,792,458 m/s exactly was selected. Possibly, if the number 500,000,000 m/s exactly had been chosen instead it woudl have avoided the confusion between c = 299,792,458 ± 1.2 m/s and c0 = 500,000,000 m/s exactly and made more clear the arbitrary nature of this number. It is the distinction between c = 299,792,458 ± 1.2 m/s with its error bar and c0 = 299,792,458 m/s exactly that I would like to see clearly explained in the article. Brews ohare (talk) 17:43, 14 September 2009 (UTC)


 * And my point is that "299,792,458 exactly" is part of today's model, but 299,792,458 ± 1.2 m/s was a real world measurement based on the then-used units. This part of the model was chosen to reflect that real world measurement, just as all parts of the model should closely reflect the real world measurements. Charvest (talk) 17:54, 14 September 2009 (UTC)


 * Why do you say that the exact value is part of a model. There is no physics in the fact that the speed of light is exact when expressed in SI units.  It is just a choice of units. Martin Hogbin (talk) 18:03, 14 September 2009 (UTC)


 * I'm simply allowing the possibility of having a model which also models units. You One might say that there is no interesting physics in including units but it seems to me that a numerically complete model should allow for the modelling of units. Charvest (talk) 18:07, 14 September 2009 (UTC) (modified Charvest (talk) 05:26, 15 September 2009 (UTC))


 * I see what you are saying but I do not think it has anything to do with the Brews' perceived problem. Martin Hogbin (talk) 09:43, 15 September 2009 (UTC)


 * I think it does. It is my take on how to deal with the issue of defined values vs measured values. Charvest (talk) 05:47, 16 September 2009 (UTC)

I agree with Charvest's argument about the physical world being equivalent to a model in which c appears. However, it then has to be recognized that there exists a one parameter family of equivalent models that is obtained by rescaling the time variable relative to the spatial variables. This rescaling constant can then be absorbed into c, so the set of equivalent models is parametrized by c. Count Iblis (talk) 18:16, 14 September 2009 (UTC)
 * To Count Iblis:I didn't say the model was "equivalent" to reality though. I would say the model has been built up to reflect reality to the best of our abilities at building models. But the model known as physics isn't complete or even necessarily as accurate as we might one day make it. But anyway, you mention that we have different models parametrized by c which are equivalent to each other. What is the conclusion that you want us to draw from that statement? Charvest (talk) 18:50, 14 September 2009 (UTC)
 * The conclusion is that c has the same status as the constant 1.609344 kilometers/mile :) Count Iblis (talk) 23:21, 16 September 2009 (UTC)


 * my point is that "299,792,458 exactly" is part of today's model is not true if one means somehow that this particular number is a demand that must be met if the model is to fit nature. The measurement c = 299,792,458 ± 1.2 m/s was an evaluation of a model parameter. The number "299,792,458 exactly" is part of a definition and contains no physical information. It happens to be that a different choice, say c0 = 500,000,000 m/s would result in a 1983 metre so different from the previous metre as to cause great dislocation during its adoption, but of course, one could elect to do that if, for example, one were really hung up on easy arithmetic and didn't care about scrapping all exiting metre sticks. Brews ohare (talk) 18:36, 14 September 2009 (UTC)
 * To Brews ohare: I don't mean that nature demands it. I mean that it is a defined value therefore it is part of what I consider to be a model which includes numerical values and units. Charvest (talk) 18:50, 14 September 2009 (UTC)
 * Charvest: A model, like SR, does not require a specific value for c. But to fit nature using the pre-1983 SI units, one value of c will be optimal. However, whatever that value is, it's got nothing to do with the number 299,792,458 m/s in the modern SI Units. That number may be chosen arbitrarily to be any real number whatsoever without affecting in any way how SR fits nature. Brews ohare (talk) 20:02, 14 September 2009 (UTC)
 * Of course we don't have to put the particular number 299,792,458 in our model, but if we write down a model that includes all the units and numerical values then one way of doing this is to use 299,792,458. I don't see how you can say with a straight face that 299,792,458 has nothing to do with 299,792,458 ± 1.2 Charvest (talk) 04:56, 15 September 2009 (UTC)
 * Of course 299,792,458 ± 1.2 was an evaluation of a model parameter. But that parameter was 1,650,763.73/9,192,631,770 times the ratio between a particular transition of the krypton-86 atom and another particular transition of the caesium-133 atom. Hardly a fundamental parameter. -- _ _ _ A. di M. 19:22, 14 September 2009 (UTC)


 * I have not said it was fundamental. Brews ohare (talk) 20:02, 14 September 2009 (UTC)

Alternative (mainstream) view
Just in case the arbitrators are interested in the content of the page and none of them happen to be physicists I have written my version of what I believe to be the standard view of this subject in my user space. Please do not edit this page, it is my personal opinion. The views of other physicists and experts in metrology are welcome on the associated talk page. Martin Hogbin (talk) 17:31, 14 September 2009 (UTC)
 * Just in the same case, the Usenet Physics FAQ contain a very decent explanation of those issues. -- _ _ _ A. di M. 19:31, 14 September 2009 (UTC)
 * I've also produced a rather more flippant reply to the idea that a fixed speed of light has no physical significance… Physchim62 (talk) 09:10, 16 September 2009 (UTC)
 * Brilliant. You should be debunking crackpots on Usenet ;-) - DVdm (talk) 11:35, 16 September 2009 (UTC)
 * I'm old enough (in RL) to have cut my teeth on Usenet, that much is true ;) — Physchim62 (talk) 12:43, 16 September 2009 (UTC)

Two different concepts of the "speed limit"
I think I've figured this out. If you asked "what is the speed limit?" you could answer "the speed faster than which it is illegal to drive". Or else you could answer "65 miles per hour". Just like the answer to "what is the speed of light?" could be "the speed at which light travels in a vaccum, a fundamental physical constant" or else "299,792,458 m/s". The word "speed" is ambiguous. To say that there are two distinct concepts is misleading, though, since in each case both have to be true about the same thing.

To Brews and David, I gather, the SI's "speed of light" is a number. The number itself, since it is "defined", doesn't depend on c, although a measured value would. Drawing a distinction between the number and the physical constant sounds like a claim that c!=299,792,458 m/s. But that isn't what Brews and David are saying. This is not a dispute over a fringe theory; it's just a matter of semantics. 140.247.103.158 (talk) 14:17, 16 September 2009 (UTC)


 * 140.247.103.158, That's pretty well it. It's got nothing to do with fringe science. It's a simple case of pointing out that the speed of light, when expressed in terms of a metre that is itself defined in terms of the speed of light, is merely an uninformative tautology that should not be confused with the actual physical speed of light itself. And those who haven't grasped this point are making malice out of what Brews and I have been saying, because what we are saying can sound superficially ridiculous to those who haven't grasped the subtlety of the argument. Imagine we defined a new unit of length as being the height of the Eiffel Tower and that we called it an 'Eiffel Tower'. Then imagine somebody asking what height is the Eiffel Tower, and the reply comes that it is one 'Eiffel Tower' high. The person then asks "how high is an 'Eiffel Tower'?" The reply comes that an 'Eiffel Tower' is the height of the Eiffel Tower. So does the person now know how high the Eiffel Tower is? This would be no more ridiculous than stating the speed of light in modern SI units. David Tombe (talk) 01:44, 17 September 2009 (UTC)
 * It's not a tautology at all, it's a point on a scale. On your hypothetical scale, the Washington Monument is 0.523 Et and the Empire State Building is 1.176 Et. Are you trying to claim that there's no information in those relations? The answer to your question "How high is the Eiffel Tower?" could very well be "Just under twice as high as the Washington Monument." or "Not quite as high at the Empire State Building, but nearly." Physchim62 (talk) 10:36, 17 September 2009 (UTC)
 * To IP 140.247.103.158: Not quite. Speed is the distance traveled in a unit of time. The units of distance and time are invented by people based on standards that people choose to define the units. Speed is a real phenomenon, and you can use any units of distance and time that you choose to measure it; the number will be different, but the speed won't be. With modern technology, scientists can measure the speed of light very accurately (but, of course, not perfectly). Because the speed of light in a vacuum is constant, and because it is relatively easy to measure the speed of light very accurately in a laboratory, in 1983 the organizations that define units of measurements decided to redefine the metre (the basic unit of length in the International System of Units, abbreviated SI), based on the speed of light, as the distance light travels in 1/299,792,458 of a second. This conformed to the speed of light as measured with the pre-1983 metres, within narrow limits of accuracy. Everyone here agrees on what I have said up to this point, I believe. However, David and Brews contend that using the speed of light as the standard to define the unit of length caused big problems. I hesitate to describe their positions (which are similar in many ways but not identical), because David and Brews seem to object to everyone else's attempts to summarize succinctly what they say, but I'll try to do the best I can with some of the key points. They both contend that it changed the speed of light from something that real that can be be measured into something that is merely a "convention" (without real physical meaning) or a "tautology". David argues that this 26-year-old definition of the metre undermined part of the foundation of physics. Brews contends that the "real, physical speed of light" is now decoupled from any statement of its value (or at least from the statement of its value in SI metres). Professional physicists, which David and Brews admittedly are not, don't agree, and the professional literature on the subject doesn't support these views (although David and Brews, unlike the professional physicists here, contend that some passages in the professional literature do support them). That is the essence of the dispute, as I understand it. —Finell (Talk) 16:50, 16 September 2009 (UTC)


 * Well put. That is the problem as I understand it too. Martin Hogbin (talk) 10:42, 17 September 2009 (UTC)


 * There are many aspects to answering "what is X?"; and we should give them all. But that's not the same as saying that "X is really two different concepts"; if no source says that, then neither should we. Dicklyon (talk) 18:23, 16 September 2009 (UTC)


 * The "two different concepts" idea is definitively described here and Dicklyon's argument is disposed of here. Brews ohare (talk) 16:19, 9 October 2009 (UTC)

Has it been considered that - from our limited view amid the process of incomplete Universe - the structure of space extends more rapidly with distance, carrying its contents with it at the same faster rate, that space conducts light in the same way that a cable conducts electricity, and that therefore, by bodies travelling away from us, distant light is emitted at velocities relatively different from that in our position in space? Those bodies would apparently be static in their position in receding space, so light there would be conducted by the space there 'at the speed of light', and at all the positions between here and there, and also as it passes us here, but the light is simply increasingly 'red-shifted'. Your comments are welcome.Absolutelyamazin (talk) 07:51, 19 September 2009 (UTC)


 * Many things have been considered by many people but this page is about physics that has a sound theoretical basis and which has been experimentally verified. Martin Hogbin (talk) 08:01, 19 September 2009 (UTC)

Ah. Then would you be so kind as to explain how the proposal is 'theoretically unsound'? Absolutelyamazin (talk) 09:21, 20 September 2009 (UTC)
 * Light in vacuum appears to travel always at the same speed, regardless of the relative motion of the source and the observer. See Introduction to special relativity. -- _ _ _ A. di M. 09:33, 20 September 2009 (UTC)

Yes, I agree. And where a block of space at a distance of ten billion light years is moving away from us now at, say, half the speed of light, a galaxy it carries within it will emit light there 'at the speed of light' into its local space which is immobile relative to the galaxy itself. That light, being conducted 'at the speed of light' by space now in an outward direction through space accelerating away from us will be travelling at faster than the speed of light relative to us here and now. No? Absolutelyamazin (talk) 18:54, 20 September 2009 (UTC)


 * No indeed. Abtract (talk) 19:47, 20 September 2009 (UTC)
 * If the light is travelling away from us, we wouldn't be able to "know" anything about it until it is reflected back towards us. In principle, say if there's a random variation in the intensity of the galaxy, we can measure the distance between the galaxy and whatever is doing the reflecting by measuring the time lag between the signals (attention: this is just a thought experiment, it isn't anything that's practically feasible and the distances you're talking about). What we would see – in our frame of reference – is the reflecting object at a distance x. If we were in the galaxy itself, travelling away from the Earth at half the speed of light, we would see the giant reflecting object at a distance 3x/2. It is an example of Lorentz contraction.
 * But, you say, surely that means that the light being emitted away from us is travelling away from us at 3c/2? Well, you can imagine that if you like, but special relativity (which is a pretty well-tested theory) says that you will never be able to to an experiment to measure a speed of light that is different from c. Special relativity doesn't put a limit on your imagination, simply what you're able to observe. Warp factor 5, Scotty! Physchim62 (talk) 09:48, 21 September 2009 (UTC)

We do understand, don't we, that the Universe is already complete, that it already contains both its 'beginnings' and its 'end', whatever both may be. Only, it makes a difference if you appreciate this, as you will understand that 'time' as such does not exist, but only the relative position in the process, of which we are simply a part, and our observation gives us the impression of an incomplete Universe which is in action and with the perception of 'time'. Absolutelyamazin (talk) 17:00, 22 September 2009 (UTC)

Er. . . Hello? - I have just seen the following on 'the expansion of space' - "While special relativity constrains objects in the universe from moving faster than the speed of light with respect to each other, there is no such theoretical constraint when space itself is expanding. It is thus possible for two very distant objects to be moving away from each other at a speed greater than the speed of light (meaning that one cannot be observed from the other)."

Right, that is what I am saying here, and have been saying for fifty years, except that the objects are not constrained by anything except the nature of the Universe rather than by any 'theory' or its associated mathematical formula. I am only describing what Universe does, and if you have reservations, perhaps you should take the matter up with the 'expansion of space' page - Or, indeed, with Universe itself.Absolutelyamazin (talk) 10:08, 23 September 2009 (UTC)

So suppose you say that light travels at c through a vacuum of uniform space, but where space is distorted by extension or compression, this velocity may vary.Absolutelyamazin (talk) 05:32, 24 September 2009 (UTC)


 * We don't make stuff up here; show us a source. Dicklyon (talk) 06:20, 24 September 2009 (UTC)

Neither do I. Just write 'extension of space' in Wikipedia, and read the second paragraph.213.60.135.75 (talk) 14:15, 24 September 2009 (UTC)  Sorry, that should have been 'expansion of space'. It's still there. And, by the way, I am a source. But you still have to see it said by 'someone else'? Brother.


 * Statements on Wikipedia must be supported by what our policies and guidelines define as Reliable sources. Finell (Talk) 21:47, 24 September 2009 (UTC)
 * In terms of expanding space, you need to think of the tiny distance that light travels in an "instant", that is dx/dt in differential calculus. That speed stays constant, according to mainstream physics, although there is a minority view that things might have been different in the early Universe. Physchim62 (talk) 21:56, 24 September 2009 (UTC)

Let's make sure we're not talking at cross-purposes
Is anyone who disagree on any of the following points? My hunch says that some of the disagreements around here might actually be misunderstanding.
 * 1) "A physical quantity is expressed as the product of a numerical value (i.e., a pure number) and a unit". (From the IUPAP Red Book.) For example, in 9.109  382  15(45), me is a physical quantity, 9.109  382  15(45) is a numerical value and the kilogram is a unit.
 * 2) The numerical value of a physical quantity normally depends on the unit used, even if the physical quantity normally doesn't. For example, my choice of the unit I use to measure height has no effect on how tall I am, but the numerical value of my height is about 1.87 if I use the metre, and about 187 if I use the centimetre. In other words, the same physical quantity can be expressed with different units, but the numerical values will be different, too. For this reason, numerical values of dimensionful quantities are artefacts of the choice of units.
 * 3) Any dimensionful unit of measurement must be defined in terms of a physical quantity of the same dimension; such physical quantity can be expressed as the product of one or more physical quantities and pure numbers. For example, for the kilogram it is the mass of a piece of metal in France, and for the kelvin it is the product of the triple point temperature of water (with a certain isotopic composition) by the pure number 1/273.16.
 * 4) You can never directly measure a dimensionful physical quantity: any such measurement is inherently a measurement of a pure number, the ratio of the quantity being measured and a quantity of the same kind being used as a reference standard. For example, what I measure when I put my ruler along a line on a piece of paper is the dimensionless ratio between the length of the line and the distance between consecutive ticks on the ruler. For this reason, if two measurements of the same quantity yield different values, there's no way to determine whether the quantity has changed, the reference standard has changed, or both; to do that, we have to measure the quantity and the reference standard with respect to some other reference standard which is assumed to be constant.
 * 5) Once you have measured such a ratio, you convert it into the form "numerical value times unit" by multiplying it by the reference standard. Then there are two kinds of uncertainties in the numerical value you get: the one in determining the ratio and the one with which you know the numerical value of the reference standard; but in practice, usually one of these two kinds of error will largely dominate. For example, if I measure a time around 10 seconds with my digital stopwatch, the first kind of error (due to my reflex times when pressing the start and stop buttons) will be significant, whereas the second won't, because I can trust the time between two consecutive updates of the display to be 0.01 s to a very great accuracy. On the other hand, if I measured a time around two months, the error in the number of clock ticks in the period being measured would be negligible, but I could not be sure that the clock isn't too fast or too slow.
 * 6) When the reference standard is the unit itself, or an exact number of times the unit itself, the second kind of error is zero. For example, if I had a caesium-133 atomic clock at 0 K, I'd be sure that 9,192,631,770 of its ticks are one second exactly.
 * 7) When the quantity being measured is the unit itself, the ratio which is measured can be inverted to get the numerical value of the reference standard in the unit used, to use it for subsequent usages; this is called calibrating the measurement apparatus. For example, if I weighed the piece of metal used to define the kilogram and I got 1.024 kg, it'd mean that my scale's reference value is not 1 kg but $0.977 kg$, to within the error with which the ratio was measured. I can now multiply all subsequent measurements by this value, and the second kind of error in these measurements will be the first kind of error in the calibration. Another way of stating this is that it is pointless to measure the numerical value of the IPK mass in kilograms, as it is exactly 1 by definition; but it still makes sense to measure the ratio of that mass and other masses, for example to use the latter as reference standards, or (assuming that we can somehow be sure that the latter mass stays constant) to determine whether the IPK mass has changed.
 * 8) In some cases, the ratio between two physical quantities can be determined to within much greater accuracy than the numerical value of either of them in a particular unit; this usually happens when the physical quantity used to define the unit is such that it's hard to precisely measure the ratio between it and other quantities, and so the second type of errors will be large. For example, I can determine the ratio between the lengths of two sheets of paper on my desk to be $1$; but I can't determine the numerical values of those lengths in ancient Egyptian cubits with any decent accuracy, because I can't determine the ratio between any reference standard I could use and an ancient Egyptian cubic with any decent accuracy.
 * 9) To minimize the second kind of errors, one should use units of measurement which can be accurately compared with other reference standards. That's why the meridian definition of the metre didn't last long, and why they are thinking of replacing the International Prototype Kilogram with another definition: for example, we are able to measure the ratio of the electron mass and the IPK mass to within 50 parts per billion, and the ratio of the electron mass and the carbon-12 atom mass to within 0.42 ppb; ditto for many other subatomic particles. So, defining (for example) the unit of mass in terms of the carbon-12 atom rather than the IPK would allow for errors of the second kind about 120 times as small. Also, for the reason given at the end of point 4. above, it's useful to use reference standards with are assumed to be unable to change with time or circumstances.
 * 10) According to special relativity, which is by far the most widespread accepted description of kinematics in absence of gravity among the scientific community and is backed up by very solid experimental evidence, the speed of light in vacuum is a universal constant; also, ratios of lengths to the path traveled by light in one second can be determined to within excellent precision, better than any other reference standard.
 * 11) Everyone is free to call things whatever the f*** they want; for example, if I want to call t the quantity you call ct, E the quantity you call E/c2, v the quantity you call v/c, and so on, I am perfectly free to do so, as long as it's clear what I am doing; since in SR quantities like v/c show up far more often than quantities like v, it makes perfectly sense to use the shorter symbol for the more common quantity. This is colloquially referred to "using units in which c = 1". The philosophical reason why one would do that are irrelevant, and different people could do that having different ideas (or no ideas at all) in their mind of the philosophical reason why they do that. --  _ _ _ A. di M. 10:49, 28 September 2009 (UTC)


 * I agree with all of that and always have done. I have stopped discussion of the subject and editing the article here until we get a response from the arbitrators.  I am hoping that the arbitrators' response will let us get on with discussing the subject and article rationally without the madcap contributions and arguments we have had here in the past. Martin Hogbin (talk) 12:36, 28 September 2009 (UTC)
 * I only really disagree on one point that is minor (probably completely irrelevant) for this article, of which a quick summary in a second (however you define it). I also think there are a couple of "basic principles" point that you hint at but don't mention explicitly.
 * PC1. Every measurement is based on some theory. For example, if we measure length relative to the length of a given metal bar, we assume that the length of the metal bar (under given conditions of storage and measurement) is constant. If we measure length relative to the distance travelled by light in a vacuum in a given time, we assume that the speed of light is constant. The current definition of the metre also assumes that the speed of light is independent of frequency.
 * PC2. We can usually make different measurements based on different aspects of physical theory, although rarely to the same precision at any given moment. If one of the underlying theories is "wrong" (to within the precision of the measurements), the measurements won't agree. A priori, we don't know which of the theories is at fault, but we can then test them independently against other measurements: the one that is only an approximation will always be simply an approximation. In practice, the constancy of length of metal bars was held to be an approximation.
 * As for the slight point of disagreement, your example in point 9 will only work for the electron at our current level of theory, and your statement assumes that E = mc2 is correct (I don't dispute that it's correct, but it's an additional assumption to those inherent in the definition of the metre). Also, we cannot practically redefine the kilogram in terms of a number of carbon atom-masses, nor even with more amenable nuclides, because of the problems of accurately measuring the number of atoms: several groups are spending huge amounts of money to try overcome these problems at the present time, but they're not there yet! Physchim62 (talk) 13:11, 28 September 2009 (UTC)
 * What I meant is that currently (i.e. as of CODATA 2006), the value of the electron mass in kilograms is known with a relative standard uncertainty of 5.0 and the value of the electron mass in amu is known with a r.s.u. of 4.2. BTW, that was intended to be an example, so the fact that the latter measurement assumes that E = mc2 is only marginally relevant.

A slight clarification, particularly on point 3, is needed I think. The problem here is that different systems of units do not have to be dimensionally compatible. E.g. the cgs system is not compatible with the SI system as far as elecromagnetism is concerned. In SI units the electric charge is assigned an independent dimension but in cgs units it can be expressed in Length, Time and Mass.

So, the problem with point 3 is that the whole notion if "dimensions" is not well defined. Point 3 must actually be understood in reverse. I.e. different quantities were originally assigned different dimensions simply because when they were first measured there was no known universal way to compare the different quantities. Then, in the SI system of units, one introduced extra dimensionful quantities for metrological reasons. Even if you can do with only a few independent physical standards, that may not be the most accurate way to perform measurements. Count Iblis (talk) 14:17, 28 September 2009 (UTC)
 * Indeed that was part of the follow-up I was going to post after everyone said "yes, all of those points are valid". These are two points which in my mind are logical extensions of the ones above, but for some unfathomable reason appear to be more controversial; I wrote them shortly after posting the list above, and before reading your replies.
 * The numerical value of the speed of light in vacuum in metres per second is fixed by definition as 299,792,458 and so it's pointless to measure it. Nevertheless, you can measure the ratio of the speed of light in vacuum to any other reference standard for speeds; but if you found such a ratio to have changed, you couldn't say whether it's the speed of light which changed, your reference standard which changed, or both, short of comparing them both to another reference standard you assume to be unable to change. In the framework of special relativity, the speed of light in vacuum is a constant, so it'd be your reference standard for speeds which changed; in another framework, you could find another reference standard constant in that framework and compare both the speed of light and your reference standard.
 * If you know on theoretical grounds that two quantities are always proportional, Whether they have the same or different dimensions depends on the system of units used, and hence is partly arbitrary. For example, assuming that the first law of thermodynamics holds, you may consider heat to have the dimensions of an energy, and then the first law of thermodynamics is dU = &delta;Q &minus; &delta;W ; or you may consider them to have a different dimension, and the law is dU = k&delta;Q &minus; &delta;W, where k is the mechanical equivalent of heat, a constant with the dimension of energy/heat equal to 4184 J/kcalth. Assuming that Newton's first law holds, you may consider force to have the dimension of a mass times an acceleration, and then Newton's first law is F = ma; or you may consider it to have a different dimension, and then it's gnF = ma, where gn is a constant of the dimension of mass&times;acceleration/force equal to 9.80665 kg m/(s2 kgf). Likewise, assuming that special relativity holds, you can consider time to have the dimension of a length, and then the metric in Minkowski spacetime is ds2 = dt2 &minus; dx2 &minus; dy2 &minus; dz2, or to have a different dimension, and it's ds2 = c2dt2 &minus; dx2 &minus; dy2 &minus; dz2, where c is a constant of the dimension of length/time equal to 299,792,458 m/s. This may be viewed as the former person calling Q, F, and t the quantities the latter person calls kQ, gnF and ct.
 * Let's see whether Tombe and Ohare can find a way to claim it makes sense to agree with the points I posted earlier but not with these last two. -- _ _ _ A. di M. 14:28, 28 September 2009 (UTC)


 * A. di M, with these additions, I now fully agree. I think the "unfathomable reason" is this: If something has been treated in a certain way, this tends to stick. In high school, people are still taught that the dimensions of the unit system corresponds to a fundamental physical incompatibility. But this is not something that can be supported from within physics itself. Duff writes in the Trialogue article that he himself was taughed to believe this and only later did he realize that there isn't a shed of evidence to support this view. Count Iblis (talk) 15:20, 28 September 2009 (UTC)

upper bound on the speed at which matter and information can travel
Should the article also mention that the speed of light places the lower limit on time at which two spatially separated systems can get entangled? In other words it is also the maximum speed of the cascade of quantum entanglement. (Do not confuse with the "speed" of the quantum non-local connection!). --Dc987 (talk) 08:48, 3 October 2009 (UTC)


 * In my opinion, that is better left for the quantum mechanics articles. We should not try to make an article like Speed of light more complicated than it needs to be, just as we shouldn't make quantum mechanics articles any less difficult than they need to be. Finell (Talk) 18:08, 9 October 2009 (UTC)

Subsection: Meter defined in terms of the speed of light
In the light of these facts:


 * 1) The CGPM defines the metre as 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. I don't see speed of light mentioned in this definition. I see a time-of-transit.
 * 2) Morevover, the CGPM says that wavelengths determined from frequency measurements and a given value for the speed of light have a reproducibility superior to that which can be obtained by comparison with the wavelength of the standard radiation of krypton 86 and this superior reproducibility of frequency measurement compared to comparison of lengths is one reason for the change in definition of the metre to refer to time of transit. Again, no mention of speed of light.
 * 3) Finally, the choice of a time interval of 1/299 792 458 s is arbitrary and has been selected at this value because there is an advantage, notably for astronomy and geodesy, in maintaining unchanged the value of the speed of light recommended in 1975 by the 15th CGPM in its Resolution 2 (c = 299 792 458 m/s). Here is where the "speed of light" crops up: as a matter of convenience, not necessity.

Given these points, the above sub-section title appears inappropriate. Shouldn't it be replaced with something like
 * Meter defined in terms of time of transit,

or, because this is a speed-of-light article,
 * Speed of light set by definition of the metre.

This last title seems to put the burden of explanation upon the metre, where it belongs. Brews ohare (talk) 19:08, 28 September 2009 (UTC)
 * Here we go again! Martin Hogbin (talk) 20:14, 28 September 2009 (UTC)
 * Brews, #1 defines the metre in terms of the distance that light travels in a certain time a.k.a. the speed of light. How hard can this be to understand.(TimothyRias (talk) 21:46, 28 September 2009 (UTC))
 * It's the numerical value of the speed of light that is set by definition of the metre, not the speed of light itself. They are not the same concept. See point 1. at the top of the section above for the distinction between the two. _ _ _ A. di M. 22:46, 28 September 2009 (UTC)


 * In response to Martin Hogbin: Catcalls aimed simply at interruption don't aid discussion. They should be severely dealt with, but so far have not been. Brews ohare (talk) 23:53, 28 September 2009 (UTC)


 * In response to TimothyRias: Rhetoric ("How hard can this be to understand") doesn't aid discussion, and is inflammatory. It should be severely dealt with, but so far has not been. In my mind and contrary to your point, #1 defines the metre in terms of a transit time. That is, I believe, simply a straightforward read of the exact wording of the definition, and therefore beyond controversy. The value chosen 1/299… s is an artifact of compatibility with the earlier definition of the metre, and this arbitrary choice of value is, as a matter of reasoning, logically separate from the notion of using a transit time. With that in mind, do you now agree that the title of this subsection should be changed? Brews ohare (talk) 23:53, 28 September 2009 (UTC)


 * The exact wording does not anywhere mention "transit time" it does mention "length ... travelled ... per (unit of time)", which is a speed. The problem you have here seems more linguistic than anything else so lets just write this to a commonly understood language, math. In terms of equations the 1983 definition of the metre reads:
 * m := c s/299 792 458.
 * That is the metre is defined in terms of the physical quantities "speed of light" and "second" and the convenient/arbitrary number "299 792 458". The logical result of this definition is that the expression of the speed of light in terms of the metre and the second becomes:
 * c = 299 792 458 m/s.
 * I think it is clear to everbody here that this numerical value could in principle have been set to any value, but the chosen one kept the metre as close to the old definition as possible. (TimothyRias (talk) 06:46, 29 September 2009 (UTC))


 * In response to A. di M.: I agree with this remark. Do you support a change in title, or not? Brews ohare (talk) 23:53, 28 September 2009 (UTC)


 * "Metre defined in terms of the speed of light" (I just conformed the spelling of metre to the rest of the article) is an accurate statement and an accurate heading for the section. The section explains the consequence of this definition of the metre to the value of the speed of light. Further, the heading accurately reflects the significance of the definition of the metre to the speed of light, which is the topic of the article. Plus, the documentation of the definition explains that the value chosen for the metre was not "arbitrary", as you keep insisting, but was chosen based on the measurement of the speed of light. Finell (Talk) 02:35, 29 September 2009 (UTC)


 * Response to Finell: But the metre is not defined in terms of the speed of light at all. The metre is defined in terms of a definite transit time of 1/299 … s. The consequence of this transit time is that the SI units "speed of light" is 299 … m/s by definition. Brews ohare (talk) 05:18, 29 September 2009 (UTC)


 * Meter defined in terms of the speed of light :Agree  Wdl1961 (talk) 03:05, 29 September 2009 (UTC)


 * Response to Wdl1961: Your one-word insertion into this discussion does not, of course, further the discussion, but constitutes cheerleading, an activity tending to polarize the discussion into armed camps. That activity should be severely discouraged as an undesirable encouragement of incivility, leading to emotional rather than reasoned responses from others, but so far has not been adequately disciplined. Brews ohare (talk) 15:18, 29 September 2009 (UTC)


 * Brews, this is just you trying again to tell us that the speed of light that has a defined value in SI ("the SI units speed of light" as you call it) is not the same speed of light that is a physical constant, right? Or is there some other POV you want us to get out of these rants? Dicklyon (talk) 05:59, 29 September 2009 (UTC)


 * Response to Dicklyon: You have characterized my effort at discussion as a "rant", which is uncivil and not germane. Such inflammatory terms should be dealt with harshly, but so far have not been. You also rephrase my argument incorrectly to make it sound bizarre, to justify your scorn. Such distortion (whether deliberate, or due to various forms of intellectual limitation) also should be dealt with harshly. I suggest you read the proposal and respond to what actually was said in a civil manner. Brews ohare (talk) 14:27, 29 September 2009 (UTC)


 * If by "the SI units speed of light" he means "the numerical value of the speed of light in SI units", surely it "is not the same speed of light that is a physical constant", for the former is a dimensionless number equal to 299,792,458 and the latter is, depending on your philosophy, either a dimensionful constant or a dimensionless one equal to 1. To avoid miscommunication, I suggest that from now on we only use the phrase "speed of light" to refer to the ''physical quantity", and always explicitly say "numerical value" when we mean that (see above for the explanation from the IUPAP Red Book). _ _ _ A. di M. 17:07, 29 September 2009 (UTC)

I was the one who changed it. I changed it because the previous heading could imply that the SI "defined" the "speed of light" to be something that did not equal the actual speed of light, in much the same way as the Indiana Pi Bill "defined" pi to be 3.2. Just to make it clear that the SI has no power over the speed of light nor are they going to force everyone to use an incorrect value of c. The current heading is unambiguous. 140.247.242.101 (talk) 03:33, 29 September 2009 (UTC)


 * Response to 140.247.242.101: There is a quagmire here that you may not be familiar with: the transit time corresponding to the metre can be selected to be any fraction of a second you wish. Shorter times lead to shorter metres, for example. The SI units defined value for the speed of light simply adjusts accordingly in response to that choice.  Thus, unlike defining pi to be a convenient value, which would involve a logical error, the numerical value of the SI units speed of light can be in fact any number whatsoever. That is the gist of points 2 & 3 above, and the reference to Jespersen. Accordingly, the view that the CGPM has the power to force a numerical value upon us is not a misinterpretation at all, as they do have the authority to do exactly that. Of course, this authority does not change what you call "the actual speed of light", because all they can do is define the metre, not the actual speed of light. That is why I think the title "Speed of light set by definition of the metre" is apt; maybe "SI Units speed of light set by definition of the metre" is better? . Brews ohare (talk) 05:18, 29 September 2009 (UTC)

I have changed the section title to "Redefinition of the Metre" since by the MoS section title should not contain any reference to the article title. This is also solves any possible dispute about the previous section title. (TimothyRias (talk) 06:30, 29 September 2009 (UTC))


 * TimothyRias: Thank you. Brews ohare (talk) 14:28, 29 September 2009 (UTC)

A. di M., If I am not mistaken, is that not what Brews and I have been saying all along? The article should be about the physical speed of light and not the "SI units speed of light" which is a numerical value. David Tombe (talk) 06:21, 30 September 2009 (UTC)


 * The article is and always has been about the physical speed of light; that is not an innovation attributable to you. Stating the value of the physical speed of light in various measurement systems, the history of knowledge about physical speed of light and its measurement, its role in physics, especially in special and general relativity, its role in metrology, and the rest of the content, have also been part of the article long before you arrived at this article or talk page, and long before you discovered the 26-year-old "news", which was 25 years old when you learned about it. Adding the adjective physical to speed of light, as the two of you do, does not change the latter's meaning, so physical is an unnecessary word. Shortening the title of one subsection of the article, to conform to the MOS by eliminating repetition of the article's title, was TimothyRias's correction; it is of relatively minor importance (although we do like to conform to the MOS because it is a guideline), does not change or compromise the article's content, does not resemble the titles that Brews proposed above (which no other editor agreed with), and had nothing to do with any of your lengthy and repetitious arguments here. That Brews evidently derives satisfaction from this change in the subsection's title is a bonus. Finell (Talk) 07:32, 30 September 2009 (UTC)


 * Response to Finell: The reason for introducing the "extra word" physical, orreal, or maybe actual, on occasion as an adjective describing speed of light sometimes is necessary to distinguish it from the numerical value of the speed of light in SI units, which numerical value is referred to on the NIST website and in the CGPM documentation regarding the metre as the "speed of light". (For example, Sullivan says "a consequence of the definition is that the speed of light is now a defined constant, not to be measured again." Emphasis mine.) Where clarity is at stake, an extra word is not amiss. Brews ohare (talk) 13:02, 30 September 2009 (UTC)


 * There is no difference between the two concepts. Physchim62 (talk) 13:07, 30 September 2009 (UTC)
 * I think you missed my point. Quoting the IUPAP Red Book: "A physical quantity is expressed as the product of a numerical value (i.e., a pure number) and a unit". For example, in "1.87 m", 1.87 is the numerical value and the metre is the unit; and the physical quantity 1.87 m equals 187 cm although their numerical values are different. In "299,792,458 m/s", 299,792,458 is the numerical value and the metre per second is the unit. Of course the speed of light and the pure number 299,792,458 are different concepts: "the population of the United States on 30 September 2006 at 17:58 was 299,792,458" is a meaningful statement and might be true, "the population of the United States on 30 September 2006 at 17:58 was the speed of light" is nonsense. _ _ _ A. di M. 15:36, 30 September 2009 (UTC)


 * Response to Physchim62: I take it that you subscribe to the notion that no adjective is necessary? That seems an odd position to me as the quote from Sullivan above with addition of an implied "physical" adjective would be taken to mean that the real, physical speed of light never need be measured again, which would be absurd because a number of tests of theories like quantum gravity involve very precise measurements of this speed. That is, it is significant to separate these two concepts. Brews ohare (talk) 13:16, 30 September 2009 (UTC)
 * The use of different terms also was recommended by A. di M. above. Brews ohare (talk) 14:07, 30 September 2009 (UTC)
 * I did not recommend to call the physical quantity speed of light anything else than "speed of light"; only to call its numerical value differently. _ _ _ A. di M. 15:36, 30 September 2009 (UTC)


 * Response to A. di M.: I do not think Sullivan meant that the physical "speed of light is now a defined constant, not to be measured again.", but rather that the numerical value 299… m/s is now a defined constant, not to be measured again (as the term "defined constant" requires, in fact). Do you agree with this interpretation of Sullivan? Brews ohare (talk) 19:16, 30 September 2009 (UTC)


 * I guess he indeed mean that, or he might not be thinking about the distinction. So what? The fact that someone used a term with a confusing meaning doesn't mean that's it a good idea for us to do the same. _ _ _ A. di M. 00:03, 1 October 2009 (UTC)


 * Response to A. di M.: Thanks, just checking. If you look over my remarks, I am in support of your view that confusion be avoided. Brews ohare (talk) 14:23, 1 October 2009 (UTC)


 * Finell: You might avoid intemperance of tone in your responses, and non-contributory nasty asides like (which no other editor agreed with). Such asides do not add to the discussion, are in this case incorrect, and serve only to raise the temperature of discourse. Brews ohare (talk) 13:17, 30 September 2009 (UTC)

Finell, I was actually responding to A. di M.'s observation that the physical speed of light is a different concept to the SI units speed of light which is a fixed number by definition. A. di M. is therefore making the same point that Brews, and I, and others have been making. Many sources have been supplied, and the one that Brews has just quoted was actually supplied by Physchim62. No sources have as yet been supplied that state a contrary position. David Tombe (talk) 13:50, 30 September 2009 (UTC)


 * Response to Brews ohare: I red  and think understood  "Resolution 1 of the 17th CGPM". Conférence Générale des Poids et Mesures. BIPM. 1983. http://www.bipm.org/en/CGPM/db/17/1/. Retrieved 2009-08-23." . Now I use my hand to measure things  but in the past used lasers ,miles,submarine sound profiles  and whatever ,most details i do not remember (or care about ) anymore. Wdl1961 (talk) 14:14, 30 September 2009 (UTC)


 * This is a classic example of how this talk page has degenerated and why everyone, with the notable exception of 2 stalwarts, is so tired and frustrated:
 * The heading of a subjection was changed, although to one that only TimothyRias proposed and that did not resemble any heading that any other editor proposed or supported.
 * Brews is satisfied with the new subjection heading.
 * Tombe is satisfied with the new subjection heading.
 * Everyone else is satisfied with the new subjection heading.
 * TimothyRias should be very satisfied that he managed to satisfy everyone with his the new subjection heading.
 * Tombe is still arguing.
 * Brews is still arguing.
 * Finell (Talk) 00:03, 1 October 2009 (UTC)


 * Response to Finell: As to your last Point 7 : not arguing Finell - asking for input from A. di M., just to be sure we're on the same page. Your remarks about a "classic" example and about my "still arguing" are snippy, invalid, and counterproductive, tending promote polarization and create an atmosphere of contention. Your Point 1 processing of history to minimize my role in the change in title also is completely unnecessary, as it has no bearing upon development of discussion, and is simply argumentative.
 * I don't find your other points contribute anything to your thesis that discussion has degenerated - it seems to have led to a very satisfactory result: a sensible title for the subsection that everybody likes well enough. However, as my comments above indicate, a great many incivilities arose along the way that could be omitted with benefit to the general atmosphere. Brews ohare (talk) 14:23, 1 October 2009 (UTC)


 * I am waiting for the results of the arbitration before contributing further. I think there are still may remnants of the earlier problems that need to be dealt with before we can make progress.  This will be much easier of the arbitrators take the right action. Martin Hogbin (talk) 18:13, 9 October 2009 (UTC)


 * I'm coming into this argument a bit late, but I understand the point that describing the speed of light in terms of a unit which itself is defined by the speed of light is technically meaningless. For the purposes of the average layperson in understanding and visualizing the concept of the speed of light, I don't think the argument is particularly important.  However, for those who believe the argument is important, I haven't seen any suggestions on how to change this sentence in the article to something more acceptable.  Would you rather the sentence read "Its value is 670,616,629 miles per hour" or "Its value is 2.06923193 × 10^19 cubits per millennium?"  Let's see some suggestions on how to solve the problem from the people that are calling this a problem.  Snottywong (talk) 13:18, 11 October 2009 (UTC)
 * If I understand what you are saying, I agree with you. For most editors here this is really a non-issue.  Unfortunately we have had one editor who has continued to try to make some obscure point for nearly a year now.  The matter is being addressed by arbcom.  After their decision I hope things will be easier. Martin Hogbin (talk) 14:01, 11 October 2009 (UTC)
 * As for the former, the mile is defined in terms of the metre and the hour is defined in terms of the second, so that wouldn't solve the "problem". (In the latter you'd need to specify which kind of cubit and of millennium you want to use.) My opinion is that that's not a "problem" anyway. _ _ _ A. di M. 14:42, 11 October 2009 (UTC)

Response to Snottywong: Unfortunately this topic of the "technically meaningless" use of the term "speed of light" is taking place in the midst of an unseemly and ill-tempered quarrel. My view of the topic can be found here. It's not so much a question of "meaninglessness" as a question of what is meant: that is, what is the connection of the term "speed of light" in a context where lengths are measured as transit times compared to a context where lengths are a physical measurement (fringe counts) quite distinct from time measurement? An example can be found here. Brews ohare (talk) 15:54, 11 October 2009 (UTC)


 * What you're saying means nothing to me, and frankly I don't understand the links to your point of view and examples. Maybe I'm just too dumb, but they don't make sense to me.  I don't know what fringe counts or transit times are.  Neither do 99.9999% of the people that have read or will read the Wikipedia article on the speed of light.  This seems like a convoluted philosophical discussion about nothing in particular.  The speed of light is a universal constant.  The meter is a human invention.  Just because you define the meter based on the speed of light doesn't mean that you're now barred from talking about the speed of light in terms of meters or anything related to meters.  The speed of light is 299,792,458 meters per second.  If that statement is false, let us know.  Otherwise, I really don't understand the relevance of your argument.  Even if we further refine our measurement of the speed of light, it will continue to be 299,792,458 meters per second.
 * I think one of your points is that as our measurement of the speed of light gets more and more refined, the length of a meter might actually change by a couple of femtometers. That is a valid point, but I think it would be more relevant to discuss that in the meter article rather than the speed of light article.  The speed of light isn't going to change no matter how you define the meter, so this whole discussion is irrelevant in this context.  I still haven't seen a discussion of which part of the article needs to be changed (in your opinion) and what you would like it to be changed to. Snottywong (talk) 00:43, 12 October 2009 (UTC)

Difficult section
Hi - I'm not a scientist, I'm just approaching this interesting article with a view to its readability and impact on an relavity uneducated readership. Can anyone explain this part to me? "It is possible for the group velocity of light to exceed c,[63][64] and in an experiment in 2000, laser beams traveled for extremely short distances through caesium atoms with a group velocity of 300 times c.[65] It is not, however, possible to use this technique to transfer information faster than c since the velocity of information transfer depends on the front velocity, which is always less than c.[66]". The Rationalist (talk) 17:42, 10 October 2009 (UTC)


 * This definitely needs a rewrite. Thanks for pointing it out. Finell (Talk) 17:51, 10 October 2009 (UTC)
 * I am not sure which part you do not understand. Have a look at the group velocity article.  Martin Hogbin (talk) 18:04, 10 October 2009 (UTC)

Also does anyone have a reference for the claim that the speed of light was conjectured by 'scientists' to be finite in the 11th century? That seems highly implausible. The Rationalist (talk) 17:45, 10 October 2009 (UTC)


 * There are sources for that statement, with additional details, in the Speed of light section.


 * Your copy edit was a helpful clarification. Thanks! This article could benefit from attention by interested non-specialists, since we want the article to explain the science to non-specialists. I hope you will stay around and help out here. Finell (Talk) 17:59, 10 October 2009 (UTC)


 * I agree but suggest that you wait until after results of the arbitration on this article as there may be considerable changes to the content after the decision. Martin Hogbin (talk) 18:04, 10 October 2009 (UTC)

(1) 11th century, thanks for pointing out the section. Some of the sources I can't locate, e.g. Aristotle. I will look later. (2) On what part I don't understand what a 'group velocity' is. I did click on the link, although in principle a link should not be used if a difficult concept is to be explained. The article group velocity begins "The group velocity of a wave is the velocity with which the overall shape of the wave's amplitudes — known as the modulation or envelope of the wave — propagates through space." How does that help me with the difficult idea that something can travel faster than x but information not be carried faster than x. Is it that the object travelling faster than x cannot be observed? This would resolve the paradox, but then if it cannot be observed, how do we know this? If conversely it can be observed, and if observation implies information, then the statement is not true. I'm not objecting to any of the science here, I'm just saying that to a dummy like me, the logic is not clear. In a general purpose reference work, this should be clear to dummies. The Rationalist (talk) 21:45, 10 October 2009 (UTC)

This article helps a little though not much. The Rationalist (talk) 21:47, 10 October 2009 (UTC)


 * To answer the question of why the group velocity cannot be used to send signals faster than light you might like to look at this [] simulation. Change k1 to 8.2 and w1 to 6.5.  You will then see a group velocity, the speed at which the bulges travel across the screen that exceeds the phase velocity.  However this bulge could not be used to send a message across the screen from a transmitter on the left to a receiver on the right because it is caused by the interaction of waves which have already left the transmitter.  I am not sure if this will make sense to you.  There is a better simulation somewhere where you can attempt to send a signal and see why it does not travel at the group velocity.


 * I think that explaining why a group velocity exceeding c cannot be used to send superluminal signals is beyond the scope of this article. You will see that there are several other cases (such as light spots) where it is simply stated that the method cannot be used to send information faster than light. I think that if we attempted to give a clear explanation of why each method cannot be used to send superluminal signals the article would become too long.  perhaps the Faster than light article is the place to do it.  Martin Hogbin (talk) 13:57, 11 October 2009 (UTC)
 * Thanks - I sort of follow your explanation. I could not get the simulation to work however. As a general rule if something is too difficult or complex to explain in an article, then leave out any explanation, and either simply state the fact, or don't even state it at all.  Or perhaps say that it's very difficult, and link to a proper article on the subject.  Faster than light sounds the right place. The Rationalist (talk) 16:02, 11 October 2009 (UTC)
 * I agree that it is probably best just to say that group velocity of greater than c cannot be used to send information and leave it at that, like we do with light spots and shadows.
 * Some of us are waiting for the dust to settle from the current arbcom case before doing and significant editing here. Martin Hogbin (talk) 21:11, 11 October 2009 (UTC)

Is the speed of light same for a human being, a dog and a plant?
I am not a scientist, but yes I can possibly call myself a philosopher interested in meta physics.

Can some one clarify my doubht on Physics which goes as follows: Question : Is there any method / way to test and prove (or disprove) the following hypothesis? "The speed of light is same for a human being, a dog and a plant" That is "the "processing speed" of the "cognitive apparatus" of the "observing organism" has "no effect" on the "measured speed of light"

Has any one proved this hypthoses already? If Yes please give the link to the details. If not, is it physically possible to prove this hypothesis using current day technology. If not why? If yes How?

Thanks

Rsrinnovations (talk) 03:29, 12 October 2009 (UTC)


 * Such philosophical or metaphysical questions are not within the scope of what physicists address by theory and experiment. In physics, the speed of light is not related in any way to any cognitive apparatus or observing organism, so there's nothing that can be proven about such a relationship.  Dicklyon (talk) 05:00, 12 October 2009 (UTC)
 * The response of A.di M. and the explanation given in the link Why is the speed of light so high, is more close to the true spirit of my question. Rsrinnovations (talk) 17:22, 12 October 2009 (UTC)
 * The speed of light itself is a scale factor of spacetime and as such doesn't depend on anything; but if dogs and plants could measure, dogs would use shorter units of time and length than men do, and plants would use longer ones, because they have different perception of time and space; so they'd get a different numerical value for it. Take a look at Usenet Physics FAQ, "Why is the speed of light so high?". _ _ _ A. di M. 09:26, 12 October 2009 (UTC)
 * Most scientists practice a form of philosophical realism (or metaphysical objectivism), that is they consider that there is a reality which is independent of our perceptions – or of the perceptions of dogs or plants, for that matter. There are many different shades of this metaphysical objectivism, and their study forms one part of the philosophy of science. This also goes to heart of what is considered a "proof" for any given purpose, although that is strictly a branch of epistemology rather than metaphysics. For a readable introduction to the philosophy of science, I would recommend What is this thing called Science? by Alan F. Chalmers (ISBN 9780335201099). Physchim62 (talk) 10:58, 12 October 2009 (UTC)
 * Proving statements is not really what science is about. Statements only deserve the label scientific if they can in principle be disproven or falsified. So the statement clearly has no place in science, and therefore neither in physics. DVdm (talk) 12:02, 12 October 2009 (UTC)


 * Rsrinnovations I agree with the comments above, particularly that by Physchim62 but wonder if your question was about measurement practice. In case it was: practical measurements of the speed of light do not depend in any way on the speed of human cognition. Martin Hogbin (talk) 17:00, 12 October 2009 (UTC)
 * While the practical measurement of speed of light does not directly depend on the speed of human cognition, those gadgets are calibrated to some traceability standards and in turn these standards are linked to what humans perceive as "One metre" and "One second". (Refer answer by user A. Di. M.) So after all there is a "human cognition and perception related relativity" to the concept of space, time and speed of light. Rsrinnovations (talk) 17:22, 12 October 2009 (UTC)

"Here the speed of light enters the theory via the dimensionless fine structure constant"
Well, it is true that α can be defined in terms of c; but the statement I quoted in the title sounds so strange to me. α is essentially the square of the elementary charge in natural units, so normally I'd find it more useful to say that the elementary charge enters the theory of QED through the dimensionless constant α. You could as well define it as Z0e2/2h, so that c wouldn't show up in its definition. Can anyone think of a way of rewording this statement so that it doesn't sound this weird while staying on-topic? _ _ _ A. di M. 19:47, 12 October 2009 (UTC)


 * I agree. I wrote that sentence some time ago, replacing the claim that c palys a fundamental role in QED. I think it is better to start all over with a section in which we explain how in theoretical physics natural units in which c = 1 is used. We don't necessarily need to focus on the fine structure constant, any example will do. Count Iblis (talk) 19:55, 12 October 2009 (UTC)
 * A better idea would be improving the now-poor natural units and Planck units articles, and having a one-paragraph introduction to those ideas here with a link to there. _ _ _ A. di M. 20:12, 12 October 2009 (UTC)
 * Yes, a far better idea. --Michael C. Price talk 09:54, 13 October 2009 (UTC)

"Light and photons": do we need that section?
The only property of photons relevant to their speed is their mass, so what's this doing here: "Quantum field theory reflects not only the wave aspects of light but also the wave-particle duality of light. According to this theory, electromagnetic forces between charged particles are seen as mediated by an exchange of photons, which act as carriers of the electromagnetic force or field. One early argument for the existence of photons was Einstein's Nobel Prize–winning explanation of the photoelectric effect, which was based on a model of light as particles.[35]"

I propose we delete the first paragraph of that section, and integrate the last two paragraphs in the following section. _ _ _ A. di M. 10:27, 13 October 2009 (UTC)


 * Agree, that paragraph seems to stray off-topic. On a related note, I can't find an explanantion in in the main text that massless particles and fields must travel with the speed of light. Besides being a very relevant fact, it is also necessary to make the connection photon mass -> dispersion.
 * Before integrating anything into the next section, that section needs to be split into two very distinct topics: variations in the "spacetime constant" c and dispersion of light in the vacuum. (TimothyRias (talk) 06:45, 14 October 2009 (UTC))
 * I've given it a try; but for now the subsubsection header for dispersion of light in vacuum is commented out, because I'm not convinced of how it looks with it. _ _ _ A. di M. 10:25, 14 October 2009 (UTC)

Rearrangement of material
I'm trying to reply to both the previous sections here… At the moment we have a subsection "Light as electromagnetic radiation", which doesn't discuss photons, and a subsection "Variations in the speed of light", which seems badly named given that most (although not all) physical theory assumes that the speed of light in vacuum is constant.

My suggestion is:
 * 1) that we add a paragraph about the particle-like nature of electromagnetic radiation to the subsection "Light as electromagnetic radiation", which would state clearly and simply that the theory implies that photons have zero mass;
 * 2) that we create a new section further down the article (I would suggest at the end, for the moment) entitle "Constancy of the speed of light", where we can discuss the various experimental tests of the constancy of c0, including its implication in the fine structure constant, and also VSL theory.

It wouldn't be too difficult to put the suggestion into practice, even if it might require writing some new content and then improving that content by the normal WikiProcess, but I would like to have feedback from other editors before making the change. Physchim62 (talk) 13:27, 14 October 2009 (UTC)
 * Agree with point 1. Point 2 would be merging the sections which now are titled "Constant speed in inertial frames" and "Variations in the speed of light" into a section near the end, right? I dunno; I think that the current version, with the mainstream view (c0 is an absolute, unchanging constant, the photon is massless, etc.) first and speculations (c0 might have changed, the photon might be massive, etc.) at the end, has its merit. But then, maybe that's just because I know what I looks like while I don't know how your proposal would look like. You might try to implement that in a sandbox, first; maybe it will be better than I expect.  _ _ _ A. di M. 14:36, 14 October 2009 (UTC)


 * Remark, per MoS (sub)section titles should not contain the article title. The current section headers violate this at times, but we should try to at least avoid new violations. (TimothyRias (talk) 14:48, 14 October 2009 (UTC))
 * That recommendation says "unless doing so is shorter or clearer", and just titling a section "Variations" or "Constancy", IMO, would be unclear enough that having "speed of light" in these headers isn't totally unjustified. _ _ _ A. di M. 18:02, 14 October 2009 (UTC)
 * My experience is that usually there is a better more descriptive title not involving the article title. The fact that taking away "speed of light" makes the section titles sound vague actually show a vagueness in the original titles since "speed of light" is implied anyway. For example, "Variation with space and time" would be more descriptive then "variation of the speed of light".
 * The main point is that having the article title in the section headers should be kept to a minimum, if only to minimize the bitching it will provoke if the article goes to GAN/FAC. (TimothyRias (talk) 18:24, 14 October 2009 (UTC))
 * "Variation with time and frequency" would more closely match the content of the section. I'm going to do that myself; feel free to revert if you have an objection. 19:49, 14 October 2009 (UTC)
 * That's a good heading. Nice improvements in the article in recent days, too! Finell (Talk) 20:10, 14 October 2009 (UTC)

I've made an adjustment as per my point 1: please improve that section as need be! I've left in the final sentence about the fine structure constant for the moment, as I think α needs to be mentioned somewhere and it might as well be there for now. I'm collecting my thoughts on the "constancy" issue in a sandbox as A. di M. suggests, and I'll post back when I have something presentable. Physchim62 (talk) 12:10, 15 October 2009 (UTC)

DL rv: why?
DL reverted my change Both the "units of" and the reason not supported by the source seem wrong to me; but thanks for being bold; now let's discuss. However, I don't see any subsequent discussion of either point by him. William M. Connolley (talk) 07:39, 16 October 2009 (UTC)


 * Sorry for the delay; I missed your question in all the noise.


 * You wrote "Since theory and observation both said that the speed of light is a constant,..." but according to the source cited there, that was not a part of their reason for the change. I think you're right, pretty much, but if you want to say that you'll need a source.


 * And I didn't understand the point of the "units of" change, so maybe you can explain here why you prefer "c is an important constant connecting space and time in the unified structure of spacetime, defines the units of conversion between mass and energy..." to "c is an important constant connecting space and time in the unified structure of spacetime, defines the conversion between mass and energy...". Seems me that introducting units here will completely muddle the concept that's being presented, which should be true in any system of units, old, new, or otherwise.


 * By "now let's discuss," I meant your turn, per WP:BRD. Dicklyon (talk) 02:59, 17 October 2009 (UTC)


 * Theory and Obs: I think that is just well known. Reverting that, if you agree, rather than adding cn seems unnecessarily combative William M. Connolley (talk) 14:46, 17 October 2009 (UTC)


 * It's not well known what motivated the change; this was probably part of it. I didn't mean to seem "combative", but perhaps that's the way it came off.  It seems to me that it is clear that accepted theoretical constancy of the speed of light is a necessary condition for the new definition of the speed of light to make sense, but to be making any changes at all to this over-contentious material, we need to make sure we're reporting what sources say, not our own analysis.  Especially adding material to a sentence that cites a source, when the source does not support it, is a no-no; especially when you're adding a reason, and the source is about reasons, and it's not among those listed, and there's an ongoing editing argument about such things.  Dicklyon (talk) 16:14, 17 October 2009 (UTC)


 * Units: c doesn't "define" the conversion: a whole pile of theory and experiment "define" the conversion. c is a constant that appears in the formula William M. Connolley (talk) 14:46, 17 October 2009 (UTC)


 * I don't see what units has to do with it here. It would seem to imply that by c you mean the numerical value of c.  Can you find a source that does it this way, more or less?  Dicklyon (talk) 16:14, 17 October 2009 (UTC)

William, since you're obviously a sensible guy who listens, I'll leave it to you to try again, and we'll see if anyone else objects. I like the way A. di M. fixed up what Count Iblis was up to, and maybe he or someone will be similarly helpful with your bit if needed. Dicklyon (talk) 16:57, 17 October 2009 (UTC)
 * Here's my take, but someone might come up with something better than this. _ _ _ A. di M. 19:23, 17 October 2009 (UTC)


 * Well, "appears in the formula" is hard to argue with, but we can probably say something more powerful than that. Dicklyon (talk) 19:43, 17 October 2009 (UTC)


 * Perhaps "relates the mass of an object to its total energy content via E = m c^2". But this is more an English language exercise than physics, so its not worth wasting much time over. Count Iblis (talk) 19:56, 17 October 2009 (UTC)


 * Personally, I think it would be wise to spend less time on esoteric physics and more on writing good encyclopedia articles. Dicklyon (talk) 20:04, 17 October 2009 (UTC)
 * You can't write good articles on physics without taking the time to carefully read all relevant sources to gain a deep understanding of the topic. Count Iblis (talk) 20:16, 17 October 2009 (UTC)


 * OK, that explains why this article has problems; not enough people have studied the sources on the Scharnhorst effect. Dicklyon (talk) 20:22, 17 October 2009 (UTC)

Redefinition of the metre
OK, there are piles of talk, and an arbcomm case, and no I can't be bothered to read it all. I have just read the article section, and it appears to me to miss, or to fail to lay enough stress, on the bleedin' obvious, which I offer for your delectation. Feel free to shoot me down in flames, briefly.


 * Once upon a time, we had a defined standard for the metre, and one for the unit of time (the second item, inconveniently called the second), and thus the third item, the speed of light, was susceptible to measurement, by measuring how many metres light travelled in a second. Easy. At that point it made no sense to say "how long is a metre" because the answer was, "its a metre long, guv". And ditto for a second.
 * Then people (correctly) decided that the metre could not be measured with precision [*], but were by then confident that the speed of light was a constant. Hence the speed of light is now defined, and the second is defined as before, and thus the third item (which is now the metre) becomes susceptible of measurement (in exactly the same way as before, by measuring how many metres light goes in a second). So just as before you couldn't ask how long a metre was, now it makes no sense to ask, "how fast does light go" because the answer is "the speed of light".

There you go, that's it William M. Connolley (talk) 22:27, 15 October 2009 (UTC)

[*] This is wrong, as TR noticed, so needs correction: "Then people (correctly) decided that using the metre-bar as a definition of distance was flawed (see metre)" will do. I don't think there is any point going into the why; the metre article covers this (actually it doesn't much, but is good - even rising CO2 conc is considered!) William M. Connolley (talk) 10:40, 16 October 2009 (UTC)


 * I thought that was all covered in the article already, both in the lead and in a section. Why do you think it needs more stress?    Dicklyon (talk) 01:40, 16 October 2009 (UTC)


 * because if you can not pick it up and throw it against a wall and does not generate "bonk" sound it aint real. Wdl1961 (talk) 02:05, 16 October 2009 (UTC)


 * Well, yes, w.r.t. bonkers, I would agree. Dicklyon (talk) 02:07, 16 October 2009 (UTC)

To actually reply to your points. The question "how long is a metre?" makes perfect sense with any definition of the metre. It has perfectly good answers like "about half my length", "approxiamtely the distance travelled by a walking man in a heartbeat", "the length of the international prototype" (old definition) "the distance travelled by light in about 1/3 10^-8 seconds" (modern definition). All these express the length of the metre. Note that you in your second point you are self contradictory, you both claim that "the metre could not be measured with (enough) precision" (implying that it could be measured) and that it "becomes susceptible to measurement" (implying that it could not be measure before). Also it is not correct to say that the 1983 definition of the metre (re)defined of the speed of light, which is impossible. It only fixed the value of the speed of light in SI units. It is akin to saying that if you measure charge in elementary charges you define the charge of the electron. (TimothyRias (talk) 08:51, 16 October 2009 (UTC))


 * It has perfectly good answers like "about half my length" - no, these are not measuring the metre, they are measuring your length, etc, not measuring the metre. *If* you use the metre as your absolute standard then it cannot be measured. Of course, during the period when it was the international standard people became well aware that you could in practice "measure" it with interfereometry, but that is a different matter. Note that you in your second point you are self contradictory - not quite, but I agree that my phrasing there was poor, due to not wanting to write too much. But what I've written there is technically incorrect, so score one point to you (and -1 to DL for not spotting it). I've had another go William M. Connolley (talk) 10:40, 16 October 2009 (UTC)
 * What grates a bit for me with relation to your ideas about the redefinition of the metre is that they seem to imply that, one fine morning, the CGPM got together and said "Hey, we all agree that the speed of light is constant, so lets use it to define the metre!" There are two very real advantages in the 1983 definition over its predecessors (neither of which are really relevant to the speed of light)
 * by having a defined value for the speed of light, you automatically halve the uncertainties in length measurements by interferometry, because you no longer have to consider the uncertainty in e speed of light separately from the uncertainty in your fringe count.
 * by defining the speed of light, rather than, say, the wavelength of light from an iodine stabilized laser of the type that everyone actually uses for interferometry, you don't need to have a new definition each time there is an advance in experimental technique. The CGPM is a diplomatic conference which meets once every four years: it is a very unwieldy structure for the needs of modern science, and moves very slowly. New recommended wavelengths can be issued by the CIPM, which meets once a year and is composed of scientists rather than diplomats.
 * The 1983 definition assumes that the speed of light in vacuum is constant over time, against frequency etc., but the redefinition was a practical measure, not a philosophical statement. Physchim62 (talk) 11:11, 16 October 2009 (UTC)


 * I'm OK with that (and no I didn't mean to imply it was done out of the blue); but these are peripheral matters to the question of what the redefinition actually *meant*. Why it was done is an interesting question, which you know more about than me, and you've raised some points that I didn't know, but they aren't the issue at hand. The issue at hand is, that you can arbitrarily define 2 out of the 3 (time, distance, speed) but some definitions are more useful than others. DL says that this is already covered in the article. I say that while it is indeed implicitly covered it is not made explicit in an easy-to-understand way, which is why you're having this tedious pointless arbcomm case William M. Connolley (talk) 11:23, 16 October 2009 (UTC)

I think that most people now understand the issue. But the main problem with the lead in the article as it now stands, is that it doesn't immediately clarify the fact that the exact speed of light that is stated is a consequence of the 1983 definition of the metre. The speed of light when expressed in SI units is merely a statement of a chosen number that was chosen for the definition of the metre. It is therefore misleading for a non-physics readership, because it doesn't give any physical information about the speed of light. It is exactly the same as saying that the speed of light is one light year per year.

Checking again, I can see that the explanation is actually present further down in the lead. However, it would be better if the explanation immediately followed the statement of fact. And it would be better if the SI speed of light was not used as the means of introducing the physical speed of light at the very beginning of the article. David Tombe (talk) 02:47, 17 October 2009 (UTC)


 * What would "be better" would be to recognize the consensus of editors who have been working on the details of how to present this tidbit for months, if not years. Dicklyon (talk) 02:54, 17 October 2009 (UTC)
 * Again. The non-physics readership have no idea of the way the metre is officially defined, but a large part of them know that it is about 60% of their height. To them, saying that the speed of light is 299,792,458 metres per seconds tells them that in a time roughly equal to their heartbeat when they are calm, light travels a distance roughly equal to 180 million times their height, which is something they were likely not to know before. Or do you believe that there's someone out there who knows that one metre equals one 299,792,458th of a light-second but doesn't have the faintest idea of how many metres the typical human being is tall, the typical road is wide, or the typical car is long? (Maybe yes, but there are conversions to miles per hour and miles per second in the table for them.) _ _ _ A. di M. 10:11, 17 October 2009 (UTC)
 * I agree with A. di M. Anyone who reads the first paragraph of this article and is immediately struck by the irony that c is being described in terms of meters is, firstly, probably a theoretical physicist who is already well-versed on the topic, and secondly, probably represents 0.000001% of the population.  The speed of light is fixed and constant.  The meter is a human invention.  The speed of light doesn't change because of how we define our units of length.  Therefore, I think this discussion is much more relevant to the meter article.  It is silly to argue that an obtuse philosophical argument should be fully explained in the first paragraph of the article.  I believe the explanation given further down in the article more than suffices. Snottywong (talk) 13:57, 18 October 2009 (UTC)