Way to measure one-way speed of light? [duplicate]

The speed of light, c, is the same for all inertial observers but no one has ever actually measured it in one direction. So taking an normal emitter, reflector, detector setup, can we do the following? Make the first measurement, and log the time. Put a medium that allows light through but "slows" it down slightly, put in front of the emitter side, take another measurement. Then take that same medium and put it in front of the detector side, take the last measurement, and compare. Test a should give us the same we always get. If light moves the same velocity in every direction, test b and c should give the same result, if not then they will be different, meaning light changes velocity depending on the direction it is going.

If I am wrong, please tell me what is wrong and why. What are your thoughts on trying to measure the one-way speed of light, and do you think it is possible at all?

• Suppose light travels in one direction at $c$ and returns instantly (infinite velocity). If velocity is infinite in the reverse direction, the dampening medium wouldn't do anything. Jan 6 '21 at 11:23
• @Aravind Suresh exactly! if it doesn't change with the medium on the reverse direction relative to the control with out the medium, it would confirm light moves 1/2c in one direction and at infinite velocity in the other. Jan 6 '21 at 12:11

If the one way speed of light in vacuum varies in such a way that the 2 way speed of light is c, then the one way speed of light in a medium will necessarily vary in such a way that this measurement will get a null result.

The one way speed of light is not a physical fact. It is a convention. We choose the one way speed of light by choosing our synchronization convention. Choosing that convention sets both the one way speed of light in vacuum and in media.

• Then is what your saying If the emitter trip is 49% and return trip is 51% (or vice versa), and the medium slows down light by 1% on either trip (an arbitrary value), in both tests b and c will always show a two way value of 99% except when a trip takes zero time.so the only thing we could ever confirm is a one way trip of light is not instantaneous, but can never measure a one-way trip if it has finite velocity. Is that right? Jan 7 '21 at 3:55
• Let's say the medium has length L and that light travels $\frac 23$c one way and 2c (3x faster) the opposite way. So if it takes 1 ns to travel a distance of L through a vacuum and 2 ns to travel through the medium in the fast direction, it would take 3 ns and 6 ns, respectively, in the slow direction. This would mean that when the light passes through the medium in the fast direction, the round trip would be delayed by 3 - 1 = 2 ns but when it passes through the medium in the slow direction, the delay would be 6 - 2 = 4 ns. Wouldn't this difference in delay reveal a difference in one-way c? Jan 18 at 2:45

To expand the explanation of Dale's answer a little: To measure a one-directional speed of anything you need to be able to define what 'at the same time' means between two locations because you need to agree what the start time was in order to measure a duration. Unfortunately, in a relativistic universe, there is no such thing as 'at the same time' between two locations.

Of course you could synchronise two atomic clocks and move one of them to a second location, however as soon as it is moved any concept of simultaneity is gone. A difference in velocity causes a difference in time dilation and the difference in the observed time dilation between these two clocks will depend on the observers frame of reference. It would not even be possible to agree if one clock were ahead or behind the other.

It is not an easy concept to accept that the one thing we take as a fixed constant, the speed of light in any given direction, is not actually directly measurable, however that is because all other frames of reference are relative so there is nothing to measure it against.

• Is it possible to synchronise two clocks, lets call them a and b. We then take clock c and synchronise it with b, sets b is synched with a then c will also be synched with a and without moving it there is no time dilation. We then do this for the entire length of the testing area. I'm most likely missing something fundamental to this, but why does this sequential synching of clocks not work? Feb 26 '21 at 22:43
• to synchronise them they need to be in the same location. each additional clock you are syncing all have to be in the same location. to get to another location you need to move one and you are back to the same problem. Observing one clock from a nearby position involves the time delay due to light travelling from one to the other - and while small as the aim here is to measure the speed of light this, then, matters. Feb 27 '21 at 14:01
• But an experiment like this doesn't require any distance between the emitter and detector. The light could bounce off two different mirrors to form a triangle before coming back to the detector and the medium could be placed on each of the three legs to see if they all cause the same delay. What would be wrong with that? Jan 18 at 2:03
• True but then you are no longer measuring the one-way speed of light, which was the question. Or at least to get the one way speed you need to 'see if they all cause the same delay' as you say, but to find out if the light took the same time on each leg then you need to measure the time of each leg and you have the same problem. Jan 19 at 9:18

There is no way to measure one way speed of light using clock/ruler method, relativity don't allow this, so we need search for solutions in area where relativity does not apply - quantum universe ...

Imagine an hypothetical experiment, where we build hadron collider, in universe space, that is able to turn and hold whatever stable orientation of collision against observable universe. We setup system to have a same constant energy for each collision in one system direction, for some longer period ... Then we change direction of collision for same period and maybe we see some differences.

For example, if we count how many Higs boson we saw each direction/period, or we deeper study collisions where time symmetry breaks...

All our hadron colliders are bounded to Earth, and Earth spins, moves ... But they already produced tons of data, so maybe there are some markings or dependencies that points to at least direction of velocity against universe...