# Unable to see how light's one-way speed can be c experimentally

1. Aug 31, 2013

### BluMuun

It has been said that special relativity calls for light's one-way speed to be exactly c per all inertial coordinate systems. I have been trying to picture how this can happen experimentally, but cannot come up with a scenario. Perhaps someone here can help?

2. Aug 31, 2013

### ghwellsjr

Nope, no one can help. Light's one-way speed of exactly c is Einstein's second postulate on which Special Relativity is based. It's consistent with all experiments but it cannot come from an experiment.

3. Sep 1, 2013

### Staff: Mentor

Can you turn this around? Give us a scenario in which the experimental results would not be consistent with the speed of light being exactly c in all inertial coordinate systems?

If you're saying that you can't come up with an experiment that shows that one-way speed of light is c in all inertial coordinate systems, you're absolutely right. The one-way speed of light is a postulate and cannot be proven by experiment (of course it could be disproven, bu that hasn't happened).

4. Sep 1, 2013

5. Sep 3, 2013

### BluMuun

[2nd try at posting this]

Thanks for the replies, guys.

During the creation of relativity (SR), Einstein noted that prior to SR it was possible for light's one-way speed to be c for one observer but c - v for another. It was this scenario that gave Einstein a major pain and caused him to create SR on the spot. http://www.bartleby.com/173/7.html

Einstein's cause for concern was his notion that the above one-way results conflicted with the principle of relativity. (Einstein's w = c - v is not a closing velocity - such a velocity would not even apparently conflict with the PR.)

As for the postulate part, it seems to me that both Nugatory and Wells are slightly confusing a math axiom or postulate with a physical or scientific postulate.

A scientific postulate is a guess, a supposition, a hunch, or a hypothesis about the nature of nature.

Indeed, as Nugatory noted, all scientific postulates must be experimentally testable.

So this leads to the question How can SR's one-way invariance be tested?

6. Sep 3, 2013

### Staff: Mentor

See the reference in my post above.

7. Sep 4, 2013

### BluMuun

Oddly enough, given all the info at that site that you cited and I sighted, one would think that the direct experiment would be mentioned somewhere. By "direct" I mean using two mutually-at-rest clocks that are not rotating and have not been transported (because such clocks run slow).

SR says that light's one-way, two-clock speed (with the caveats above) must be c in all inertial frames, but I am still unable to see how it can be c in any inertial frame.

Indeed, according to John A. Wheeler (in his book _Spacetime Physics_), clocks are simply set by definition to get the chosen value "c." There is really no way to falsify a definition or to even test one.

Color me confused.

8. Sep 4, 2013

### ghwellsjr

If you are "unable to see how it can be c in any inertial frame", then it should be easy for you to falsify the definition, shouldn't it?

If light propagated at c only in a single absolute ether rest state, and if the Principle of Relativity were not true, then it would be very easy to falsify Einstein's second postulate, wouldn't it?

9. Sep 4, 2013

Staff Emeritus
Be careful - most articles about the one-way speed of light are written by crackpots.

Professionals don't worry about this because it's an underconstrained system. There are two things you need to know - the clock synchronization convention and the one-way speed of light. You can't determine two things with a single measurement.

A similar example: the sun is directly overhead. What is your longitude?

10. Sep 4, 2013

### Staff: Mentor

How would you suggest synchronizing the clocks in such an experiment?

11. Sep 4, 2013

### Staff: Mentor

I asked a while back in this thread, and I'm asking again... What is a situation in which you find this assumption problematic? That is, can you show us an example of how it's hard to see that the one-way speed of light could be c in all inertial frames? (It's easy to find situations in which if you assume that the one-way speed of light is not c in all inertial frames you get horrible complications).

That's true. We do that because if we set clocks any other way, we get various horrible complications, or internal contradictions, or results that don't match experiment. Thus, even though it is not possible to measure the one-way speed of light (dig deep into any experiment that appears to measure the constancy of the one-way speed of light and you'll find that the clocks used to measure the travel time were synchronized using some procedure that assumes what we're trying to prove) it is very convenient to assume that the one-way speed of light is constant.

If the definition leads to a contradiction it's falsified. If the definition leads to predictions that don't match experimental results, it's falsified. The one-way speed of light assumption has held up just fine against challenges of this sort and there are no alternatives that do so as well.

Last edited: Sep 4, 2013
12. Sep 5, 2013

### ghwellsjr

Could you please supply the page number from which you took this quote?

13. Sep 5, 2013

### harrylin

A simple, direct test is completely useless, insofar as two-way light speed has been tested already. That's due to the synchronization convention which makes the one-way speed of light equal to the (average) two-way speed of light. As a matter of fact, that's already implied in the first section of Einstein's 1905 paper : http://www.fourmilab.ch/etexts/einstein/specrel/www/
Thus the one-way speed is made equal to the two-way speed.

The self-consistency can be tested by indirect measurements, such as for example a measurement of the speed of light after clock transport. For example one synchronizes two clocks that one next moves apart very slowly or at equal speeds relative to the lab. A light signal that is next sent one way should be recorded with those clocks as taking the same time as a signal that is sent the other way. Similarly, one could rotate the setup (or wait for the lab to rotate), and again the measured speeds should be c in both directions.

ADDENDUM: those are tests of the constancy of the one-speed of light - that is, using the same reference system. For such a test one doesn't need to re-synchronize the clocks, as they should remain in sync with each other wrt the lab. However, in order to find that the one-way speed of light is invariant (between inertial frames that are accelerated to different velocities) one must in general first re-synchronize each pair of clocks after acceleration.

Last edited: Sep 5, 2013
14. Sep 5, 2013

### stevendaryl

Staff Emeritus
I don't have Wheeler's book, but I think that the original poster is talking about the Einstein synchronization convention: You set two distant, comoving clocks so that the one-way time for a light-signal is half the round-trip time. This makes light-speed isotropic by definition.

This doesn't make SR into a tautology, because there is more than one way to synchronize clocks. For example, slow clock transport: You synchronize two clocks, then slowly (at a speed much less than the speed of light) separate them. SR says that two clocks synchronized by this definition will also be synchronized by the Einstein synchronization convention.

15. Sep 5, 2013

### stevendaryl

Staff Emeritus
Conceptually, the empirical meaning of the constancy of the one-way speed of light is given by:

1. Pick an inertial reference frame F. Assume that we have a large, flat platform at rest in F to do our experiments on.
2. Draw a straight line on the platform, and mark two spots on the line a distance L apart (as measured by meter sticks at rest in F).
3. Take two identical clocks, bring them together at one of the marked spots and synchronize them.
4. Move one of the clocks at constant speed to the other spot, and then bring it back to rest in frame F. Let $v$ be the speed that the clock was transported (as computed by $v=L/\delta t$, where $\delta t$ is the elapsed time on the transported clock.
5. At time $t_0$ according to the first clock, send a light signal toward the second clock.
6. Let $t_1$ be the time the signal arrives, according to the second clock.
7. In the limit of slow clock transport ($v \rightarrow 0$), the ratio $L/(t_1 - t_0)$ will approach a constant c.
8. This constant is found to be the same in every inertial reference frame F.

I think it's clear how it could work this way in a single frame. Then the question is: How could work this way in EVERY frame?

Well, the Lorentz transformations relating inertial coordinates in different rest frames show how it could work in every frame.

16. Sep 5, 2013

### stevendaryl

Staff Emeritus
Note that this operational meaning of lightspeed does NOT make it true by definition that light has one-way speed c.

17. Sep 5, 2013

### Staff: Mentor

If you synchronize the two clocks using Einstein's synchronization convention then it is guaranteed to be c. Do you understand why?

I agree. In the reference I posted above, that is what they are talking about when they say:

18. Sep 5, 2013

### BluMuun

GHWells noted:
>If you are "unable to see how it can be c in any inertial frame", then it should be easy >for you to falsify the definition, shouldn't it?

>If light propagated at c only in a single absolute ether rest state, and if the Principle of >Relativity were not true, then it would be very easy to falsify Einstein's second >postulate, wouldn't it?

You cannot falsify a definition because it is a given (given by man, not nature). All two-clock measurements in relativity are given by definition (the synchronization definition), including light's one-way speed and the transformation equations. As long as we are given Einstein's clocks, we are stuck with his "c invariance."

Also note that the principle of relativity does not call for c invariance, but only for the same laws. It does not say that frames are indistinguishable, but that their laws are indistinguishable. For example, if the one-way, two-clock light speed law is c ± v (as in Einstein's own example ref'd above), then all frames share the same law, but they are distinguishable (and can detect their absolute motion).

But not only is c invariance "given" by definition, it cannot happen even on paper as far as I can see as long as two mutually-at-rest clocks are used. (No transported or rotated clocks allowed because they run slow.)

For Wells: Wheeler's stuff came from Wheeler & Taylor's book
_Spacetime Physics_, 1963 edition, page 18.

"How would you suggest synchronizing the clocks in such an experiment?"

The same way clock were set on paper for 100's of years prior to Einstein. Can you tell me why Einstein decided to get rid of such clocks?

Nugatory asked re no experiment showing c invariance:
"I asked a while back in this thread, and I'm asking again... What is a situation in which you find this assumption problematic? That is, can you show us an example of how it's hard to see that the one-way speed of light could be c in all inertial frames? (It's easy to find situations in which if you assume that the one-way speed of light is not c in all inertial frames you get horrible complications)."

To put the ball back in your court, please show on paper any frame getting c for the one-way, two-clock speed of light (not using slowed clocks as in clock transport or in the case of rotated clocks). As I see it, this must be possible or the c invariance of SR is not scientific.

And I am very interested in the "horrible complications" of not getting c? Can you cite one please?

Nug also wrote:
"Thus, even though it is not possible to measure the one-way speed of light (dig deep into any experiment that appears to measure the constancy of the one-way speed of light and you'll find that the clocks used to measure the travel time were synchronized using some procedure that assumes what we're trying to prove) it is very convenient to assume that the one-way speed of light is constant."

What does "the one-way light speed is constant" mean if it is not an experimental (or a possible experimental) result? Of what use is it to physics?

Why is it convenient to assume that which cannot happen?

DaleSpam quoted:
"In all of these theories the effects of slow clock transport exactly offset the effects of the anisotropic one-way speed of light (in any inertial frame), and all are experimentally indistinguishable from SR."

Sorry, but this is not true. As Einstein said, a theory that contains truly synchronous clocks predicts a variable one-way, two-clock light speed, unlike SR.
http://www.bartleby.com/173/7.html

For all of those who mentioned clock transport: I said up front that this is not allowed because moving clocks run slow, no matter how slowly they move. (I also disallowed rotating clocks for the same reason.)

In other words, I am looking for a legit experiment that shows a two-clock one-way light speed of c in any inertial frame, much less many.

DaleSpam wrote:
>If you synchronize the two clocks using Einstein's synchronization convention then it is >guaranteed to be c. Do you understand why?

But it is not via experiment; the clocks are merely manually forced to get "c," as Wheeler said.

Nugatory noted:
"If the definition leads to a contradiction it's falsified. If the definition leads to predictions that don't match experimental results, it's falsified. The one-way speed of light assumption has held up just fine against challenges of this sort and there are no alternatives that do so as well."

How can a definition lead to any contradictions if it has been assumed that any and all results of the definition are acceptable? For example, if two same-frame SR clocks are
compared with a passing clock, then the result is that "the passing clock ran slow." This may or may not be true when it comes to the intrinsic atomic rhythm of the clock, but it is accepted in SR as a valid and meaningful result. (However, it does not seem to be all that meaningful when you consider the fact that SR also has the other frame's clock running slower so that two clocks both run slower than each other. Is this really physics?)

I am not trying to be argumentative, but just trying to see how any frame's observers could use two mutually-at-rest clocks to get a one-way light speed of c experimentally. Of course, one of you said this:

stevendaryl noted:
"This doesn't make SR into a tautology, because there is more than one way to synchronize clocks. For example, slow clock transport: You synchronize two clocks, then slowly (at a speed much less than the speed of light) separate them. SR says that two clocks synchronized by this definition will also be synchronized by the Einstein synchronization convention."

Everyone agrees that using slowed clocks can result in approximate c invariance, but what I am wondering about is what happens using two mutually-at-rest clocks because SR says that this will also result in c (exactly c) experimentally. To me, these are two entirely different things (slow transport & true c invariance).

Harrylin wrote:
>A simple, direct test is completely useless, insofar as two-way light speed has been >tested already.

But SR did not predict two-way invariance - this was given prior to SR via a direct experiment. This left only the one-way case for SR to consider, and SR claims that light's one-way speed is c in all frames. Is this via experiment or merely given by definition? Can it even happen on paper?

If c invariance cannot happen, then it would seem to be of no real use to physics. This is the only reasonable conclusion that I see.

19. Sep 5, 2013

### Staff: Mentor

One such problem is that we can calculate the speed of light from Maxwell's laws of electricity and magnetism, and Maxwell's equations make no allowance whatsoever for the velocity of the observer. Thus, if we don't accept that the one-way speed of light is c for all observers, we are obliged to construct an internally consistent theory in which observers moving relative to one another experience different laws of electricity and magnetism, or to reject Maxwell's equations.

Rejecting Maxwell's equations doesn't work. There's just too much experimental evidence from the past few centuries of studying electricity and magnetism.

Different laws of electricity and magnetism for people moving at different speeds would imply that our observations on earth would change with the seasons because of the earth's motion around the sun. They don't, to the limits of accuracy of some very precise experiments.

Some history: Maxwell's laws were discovered in 1861. Reconciling them with the intuitive notion that if you're moving at speed $u$ relative to me and I measure the speed of a light signal moving in the same direction as $c$, you will measure the speed as $c-u$ was one of the great challenges of the second half of the 19th century. That's why when Einstein published his classic paper in 1905 resolving this problem by proposing a consistent and experimentally confirmed theory that assumed a constant one-way speed of light he titled it "On the electrodynamics of moving bodies".

20. Sep 6, 2013

### ghwellsjr

I have the 1966 edition which you can see the first part of here.

I cannot find anything even remotely resembling your quote. Could you please find about where it got edited out?