One and Two Way Speed of Light

Firstly apologies for bumping up an old thread and secondly I must confess to having no particular expertise in this area. However, I think you are incorrect on Lisle being a casualty here. The observations would be the same with ESC as they would with ASC. The ‘C’ at the end of each is for ‘Convention’.

@ABC Create a new thread and present your arguments. We will engage, discuss and investigate. That is what PS is all about.

1 Like

I’m sorry… I was trying to help and I think split it when someone else did. Let me know if it is not okay now… or if you’d like the title changed and I’ll fix then delete this message… @ABC @Patrick

1 Like

@Abc, can you please flesh out your point a little more?

Ok. My point is that the ESC (Einstein Synchrony Convention) and ASC (Anisotropic Synchrony Convention) will agree on the observed results. The difference is in how events are time stamped (and time dilation is more significant under ASC). Ultimately, they are based on the same equations with epsilon (ε) used to denote the one-way speed of light. For ESC the value of ε = 1/2, with light travelling at the same speed in all directions. In Lisle’s ASC ε = 1 and light travels at infinite speed towards the observer and c/2 away (for ε = 0 light would travel towards the observer at c/2 and at infinite speed away).

The implications are that it’s taken light from the most distant galaxies around 13.8 billion years to reach earth when ε = 1/2. For ε = 0 this doubles to 27.6 billion years. And under ASC we can’t tell - light would appear instantly. Lisle argues this is consistent with the Bible (Gen 1:3 “And God said, Let there be light: and there was light.”).

Weird as it may seem, it appears we are free to set ε to whatever value we choose (although values outside the range 0 to 1 have implications that suggest they are probably not valid). The only thing the universe seems to care about is the round-trip speed of light. In fact, it cares so much that time will dilate and lengths will contract to keep it constant! To me at least, that’s weirder still.

So unless we can measure the one-way speed of light (and it seems we can’t do this, even in principle) the ASC is still very much alive. You may wish to read Lisle’s book, “The Physics of Einstein” for a far better treatment of the topic than I could give.

Now add in gravitational waves initiating from the collapse of neutron stars. The one way speed of light from the neutron stars equaled the speed of gravitational waves to within 10-15 seconds. Lisle’s ASC ε = 1 is falsified, gravational waves prove ε = 1/2 .

1 Like

If it has been proven that the one-way speed of light is c, someone better inform Wikipedia!

The ASC is just a convention, it can’t be falsified in the same way the metric system can’t be falsified. Any effects interpreted to be light travel time using ESC are due to time dilation under ASC. Observationally, we can’t tell the difference.

Except we can tell the difference between the two using gravity waves. The @physicists can confirm, but it seems @Patrick is correct.

Also, the equivalence between ESC and ASC only holds for direct two-way measurements. As soon as you move to astronomy, where we see all the important tests of general relativity, the idea that e does not equal 1/2 is quickly falsified.

If the idea that ε does not equal 1/2 is falsified then the one-way speed of light is c and Wikipedia should be updated.

The fact is that this has not been proven, and it appears to be impossible to measure it even in principle.

I’m not sure of this. AFAIK, one-way vs. two-way speed of light are just conventions. No experiment has been universally acknowledged to have been able to measure the difference. Given that GW detectors are just giant interferometers, they cannot differentiate between one-way vs. two-way speed of light. @PdotdQ, can you comment on this? If tests of GR using gravitational waves are also a form of testing the one-way speed of light, that would be a breakthrough of sorts.

1 Like

I’m pretty sure @dga471 is right about this. That fact that the arrival times of gravitational waves and light are the same really just means that both obey the lorentz transformations - and you can use the lorentz transformations with whatever synchrony convention that you want. (It’s just easier to use the ESC.)


The detection of the neutron star inspiral GW170817 in 2017, detected through both gravitational waves and gamma rays, provides the so far by far best limit on the difference between the speed of light and that of gravity. Photons were detected 1.7 seconds after peak gravitational wave emission; assuming a delay of zero to ten seconds, the difference between the speeds of gravitational and electromagnetic waves, vGW − vEM , is constrained to between −3×10−15 and +7×10−16 times the speed of light.[28]

These were one way measurements coming from the neutron star inspiral to detectors on Earth.

Yes, Patrick, but there is also the complication that in GW detection (or rather, multi-messenger detection, where you can compare the speed of GWs to the speed of light such as gamma rays), you also must use a timing synchronization convention between the detectors which assumes a two-way speed of light at several points. For example, you’re assuming that the speed of the electronic signal in the cables is the same as the speed of light from the stars. (Or something like that.)

I’m only starting to read up on this, but this short rebuttal to an experiment claiming to measure the one-way speed of light captures this intuition.

In the experiment they describe, a light beam is sent from a laser to a photosensor, and then the signal from the photosensor is transmitted through a coaxial cable back to the vicinity of the laser. The length of the cable is 23.73 m, and it is asserted that transmission through the cable “introduces a fixed time delay of 79 ns”. The authors point out that all timing is performed in a single place (the vicinity of the laser) so no convention for the synchronization of distant clocks seems to be necessary. However, the assertion of a known time delay through the cable is only meaningful if one imagines having synchronized clocks at the two ends of the cable. Therefore this assertion constitutes an implicit adoption of a convention for distant synchronization.

What the experiment of Greaves, Rodriguez and Ruiz-Camacho actually measures is the time for a round trip; the first leg of this round trip is the light propagating from the laser to the photosensor, and the second leg is the signal going through the coaxial cable from the photosensor back to the vicinity of the laser. It is the assumption that the second leg is accomplished with a known speed (in particular, the round-trip speed of light) that allows the speed of the first leg to be determined.

I would not be surprised if you can apply a very similar argument to the case of multimessenger detection of GWs and light to test GR.

1 Like

One-way speed of light experiments are notoriously subtle, so I have to think more about this before saying anything with confidence. However, at first glance I agree with @dga471 and @structureoftruth.

This just says that the speed of light and speed of gravity is ~the same, i.e. the two-way speed of light is ~equal to the two way speed of gravity, and does not constitute a one-way speed of light measurement.


If the speed of gravity were different in different directions, would there not be observable consequences for orbital mechanics?

Given the types of modification to the directional speed of gravity in this discussion:

No, there is no observable consequences for orbital mechanics at least up to order ~1/c^2 (or at least we don’t yet know of an experiment that can test it).

The reason is this: up to order ~1/c^2, gravity is exactly equal to electromagnetism (modulo a negative sign in the Lenz’s Law). Thus, if there is no experiment that can detect the one-way speed of light, there is no experiment that can detect the one way speed of gravitons up to this order.

I don’t know the answer for the full theory of gravity. But most orbital motion are order ~1/c^2 or lower (typically much lower).

1 Like

I don’t understand this. If we have two objects orbiting each other, the object at one end, let’s call it A, sends out its attraction at infinite speed to the object at the other end, call it B, while B sends its attraction at c/2. So A is attracted to where B was a while ago, while B is attracted to where A is right now. That seems quite different from both being attracted to where the other was half a while ago. But perhaps it all works out.

Actually, A is not attracted to where B was a while ago, but where B is now (if B is moving with ~constant velocity). This non-intuitive result comes from the fact that the gravitational potential has to be evaluated at retarded time where c is now c/2.

This is the same case as in electromagnetism. You can read more about it here.

1 Like

I do not understand what either of those articles mean.

1 Like

In broad strokes: in the case of when the speed of gravity is finite,

This has to be satisfied for relativity to not be violated. The point is that the situation is not as one would intuitively think: